University of Pardubice

UNIVERSITY OF PARDUBICE
Faculty of Chemical Technology

and
EUROPEAN OFFICE
OF AEROSPACE RESEARCH AND DEVELOPMENT
NEW TRENDS IN RESEARCH

OF ENERGETIC MATERIALS
PROCEEDINGS OF THE VI. SEMINAR
Pardubice, Czech Republic

April 22 - 24, 2003
Faculty of Chemical Technology
Department of Theory & Technology of Explosives
CZ-532 10 Pardubice
EUROPEAN OFFICE OF AEROSPACE RESEARCH AND DEVELOPMENT
London NW1 5TH United Kingdom
PROCEEDINGS
of the sixth Seminar
NEW TRENDS IN RESEARCH OF ENERGETIC MATERIALS
held at the University of Pardubice,

Pardubice, the Czech Republic
April 22 24, 2003
intended as a meeting of students, postgraduate students, university

teachers and young research and development workers concerned
from the whole world
NOTICE
This publication has not been submitted to language corrections
and contributions have not been reviewed.
The only distributor of the present publication is the
Department of Theory & Technology of Explosives, University of Pardubice,
CZ-532 10 Pardubice,
where the publication can be ordered or gained by exchange of similar publications.
Contributions of the Proceedings will be quoted in the Chemical Abstracts.
March 21th, 2003

Close
Editor:
Ji Vgenknecht
Edition 1st Limited, 50 pcs OriginalCopLHV, 8QSrotected CD-R

Published by the University of Pardubice
Document designed by J. Vigenknecht at DTTX by Adobe Acrobat 5.0.5.0 CZ
University of Pardubice, 2003
Seminar is supported by:

European Office of Aerospace Research and Development of the USAF,
Austin Detonator, Inc., Vsetn,
Indet Safety Systems, Inc., Vsetn,
Dr. Oldrich Machacek, president of UTeC Corp., LLC, Dallas,
OZM Research, Ltd., Hrochv Tnec
Chairman of the Seminar:

Prof. Svatopluk Zeman, D.Sc.
Scientific Committee:
Chairman: Dr. Adam Cumming
Members: Prof. Ang How-Ghee
Dr. Stanislaw Cudzilo
Prof. Michel Lefevre
Prof. Andrzej Maranda, D.Sc.
Assoc. Prof. Shu Yuanjie
Dr. Muhamed Sueska, D.Sc.
Assoc. Prof. Pavel Vvra, Ph.D.
Dr. Fred Volk,
Dr. Woodward Waesche
(DSTL, Sevenoaks, U.K.)

(National Univ. of Singapore)
(Military Univ. Technol., Warsaw, Poland)
(Royal Military Academy, Belgium)
(Military Univ. Technol., Warsaw, Poland)
(Inst. of Chem. Materials, CAEP, Sichuan, China)
(Brodarski Inst., Zagreb, Croatia)
(Univ. of Pardubice)
(ICT Pfinztal, Germany)
(Office of Naval Res. Int. Field Office, USA)
Organizing Committee:
Chairman:
Ji Vgenknecht, Ph.D.
(Univ. Pardubice)
Members:
Zdenk Jalov, Ph.D.

Marcela Jungov, M.Sc.
Martin Kouba, M.Sc.
Miloslav Krupka, Ph.D.
Robert Maty, M.Sc.
Jakub eleovsk, M.Sc.
Pavel Valenta, MSc.
Dr. Jan Jakubko
(Univ. Pardubice)
(Univ. Pardubice)
(Univ. Pardubice)
(Univ. Pardubice)
(Univ. Pardubice)
(Univ. Pardubice)
(Austin Detonator)
(Indet Safety Systems)
INFORMATION ABOUT THE STUDY PROGRAM

OF THE DEPARTMENT
OF THEORY & TECHNOLOGY OF EXPLOSIVES
http://www.upce.cz/~kttv
Phone: (+420 46) 603 8023
Fax:
(+420 46) 603 8024
E-mail:
kttv.fcht@upce.cz
The University of Pardubice is a public university comprising four faculties and two
institutes, namely: Faculty of Chemical Technology, Faculty of Economy and Administration,
Jan Perner Transport Faculty, Faculty of Humanities, Institute of Informatics, and Institute of
Health Studies.
The Faculty of Chemical Technology is accredited for providing 7 MS and Bc study

programs: Chemical and Process Engineering (Bc level), Graphical Arts (Bc level
and subsequent MS program), Chemistry and Technological Chemistry (5-year MS
program with the possibility to sit Bc examinations after the 3rd year), Chemistry and
Technology of Foodstuffs (5-year MS studies) and Special Chemical and
Biological Programs (Bc level with subsequent 2-year MS studies). The study program
of Chemistry and Technological Chemistry involves 16 accredited lines of study; the other
programs have one line of study each.
The Bc-level studies are organized as full-time activity and last 3 years; they are finished
by state final examinations and viva voce of Bc Thesis. The graduates obtain the title of
Bachelor of Science (in Czech bakal abbreviated as Bc).
The MS-level studies last two years and are meant for graduates from the corresponding
Bc studies. The 5-year MS program is divided into two parts each of which is finished by a
state examination. After passing the 1st examination, the student enrols in the chosen line of
the 2nd part of MS studies. The whole MS study program is finished by finals involving viva
voce of the submitted Thesis. The graduate obtains the academic degree Engineer (in
Czech inenr, abbreviated as Ing.), and in the Special Chemical and Biological Program
the title Master of Science (Czech Magister, abbreviated as Mgr.).
Any student in the program Chemistry and Technological Chemistry can obtain the title
Bachelor of Science (in Czech bakal, abbreviated as Bc.) if he/she passes the
prescribed examination in optional subject and, as the case may be, can finish his/her
studies.
Graduates from MS-level studies who have the makings of scientific workers can continue
in PhD-level studies, which are finished by passing the PhD examinations and viva voce of
PhD dissertation
Forms of study: full-time form and combined type (full-time plus distance learning). In the
full-time form, the students attend the prescribed forms of education, whereas the combined
type is based on individual students learning and their attending the laboratory courses and
consultations.
Syllabuses contain the subjects of Bc- and MS-level studies divided into individual terms
and years. They involve obligatory subjects, obligatory-optional and recommended ones.
The syllabuses specify the credit evaluation of the individual subjects. The credits reflect the
extent and difficulty of the subject. The credits for the given subjects are obtained by
passing the respective examination.
Students advancement in study follows from the credits obtained. The study load in each
year expressed by the number of credits for obligatory and obligatory-optional subjects is
60.
Accommodation of foreign students and doctorands can be provided by the

Administration of Student Hostels and Canteensof the University of Pardubice in the
Universitys hostels (dormitories) at a price of ~$ 7 per night. The prices of accommodation
at hotels in Pardubice range between $ 18 and $ 50 per night.
Catering for foreign students and doctorands can also be provided by the
Administration of Student Hostels and Refectory of the University of Pardubice in the
Universitys refectory providing two warm meals each day. For lunch are served four sorts
of meal and a fast meal made to order at prices of $ 1.5 to &1.7 (depending on the meal
chosen from the four) and $ 2.5, respectively.
The urban transport of Pardubice offers to passengers tickets at the price of CZK 8
($ 0.25) for one journey without change, or quarterly pass tickets for all town lines at CZK
765 ($ 24.7).
Study program Chemistry and Technological Chemistry

Study Courses THEORY AND TECHNOLOGY OF EXPLOSIVES
The education in the field of science and technology of explosives started in 1920 at Prague
Institute of Chemical-Technological Engineering (now VCHT Prague). In 1953 it was
transferred to the present Faculty of Chemical Technology, The University of Pardubice,
later called Department of Special Production, at present called the Department of Theory
and Technology of Explosives (DTTX). Its facilities were unique on the territory of the former
Czechoslovakia, and they still are unique in the Czech Republic. The numbers of students
who have graduated so far are: 312 in MS-level studies, over 250 graduates from two fourterm license courses (technology and blasting techniques), and 51 graduates of PhD-level
studies. Nowadays, 52 to 65 students are enrolled in the above specified forms of study
each year, out of whom 15 are citizens of Slovak Republic. At present, the Department is
one of the two workplaces of this kind among the European NATO countries. The following
European institutions provide the same extent of education as DTTX:The Military University
of Technology in Warsaw, Poland, The Mendeleev University of Chemical Technology in
Moscow, Russia, The Kazan State University of Technology in Kazan, and Sankt
Petersburg State Institute of Technology, Russia.
The study line Theory and Technology of Explosives represents a multidisciplinary line
comprising two branches: Technology of Explosives and Safety Engineering. It involves
the fields of chemical technology, technology of explosives, explosion physics, and safety
engineering (see below the descriptions of the respective subjects).
GRADUATES PROFILE AND POSSIBILITIES

A graduate from the branch of technology of explosives is an engineer specialized in
explosives, who is able to independently solve problems and control processes of research
& development and production & processing of energetic materials; he/she can also work in
cognate areas such as blasting operations, military forces and the respective technologies,
police, government administration (mine authorities etc.) and NATO organs.
A graduate from the branch of safety engineering is able to independently solve problems
connected with prevention of losses in industry, carry out risk analyses, and draw safety
documents within a factory in accordance with legislature of the Czech Republic and the
European Union. He/she will find application particularly in chemical factories, governmental
administration, and projecting-consulting companies.
MS-LEVEL STUDY
For successful mastering of the study load in the line Theory and Technology of
Explosives the student must have sound knowledge in chemistry (especially organic),
physics, mathematics, physical chemistry, chemical informatics, and the English language;
also welcome is knowledge of chemical engineering, macromolecular chemistry, processing
of polymers, technology of composite materials, and the Russian language. The
precondition for enrolling in MS-level study at DTTX is a successful completion of the first
three years of study (passing of the first state examination) at the Faculty of Chemical
Technology, the University of Pardubice, or VCHT Prague, or FCHPT STU Bratislava.
Obligatory subjects in the 1st step of MS-level study (given in brackets is the
number of credits, WT = winter term, ST = summer term)
I.
WT
Mathematics I (8)
General chemistry (8)
Labs in general chemistry (5)
Toxicology (2)
Computational techniques I (4)
II.
WT
ST
Mathematics II (8)
Inorganic chemistry (8)
Labs in inorganic chemistry (5)
Physics I (7)
Labs in physics (3)
Ecology (2)
ST
Analytical chemistry II (6)

Labs in analytical chemistry II (5)
Organic chemistry II (7)
Labs in organic chemistry II (5)
Physical chemistry I (8)
Foreign language (4)
year
Physics II (6)
Labs in physics II (2)
Analytical chemistry I (4)
Labs in analytical chemistry I (3)
Organic chemistry I (6)
Labs in organic chemsitry I (4)
Foreign language
III.
WT
year
year
Physical chemistry II (7)
ST
Labs in physical chemistry I (3)
Chemical engineering I (6)
Measuring techniques and process control
(4)
Inorganic technology (3)
Company economy and management,
Part I (4)
Company economy and management,

Part II (4)
Social communication (2)
Organic Technology (3)
Labs in physical chemistry II (3)
Chemical engineering II (7)
Labs in chemical engineering II (3)
Labs in automation of chemical production
(3)
Obligatory-optional subjects (8)

SUBJECTS OF THE 1ST STATE EXAMINATION
Physical chemistry
Chemical engineering
Inorganic chemistry/ Organic chemistry/ Analytical chemistry
Obligatory subjects in the 2nd step of MS-level study (given in brackets is the
number of credits, WT = winter term, ST = summer term)
IV.
WT
Theory of explosives (7)

Technology of basic explosives (6)
Technology of propellants (4)
Labs of the specialisation (10)
Obligatory-optional subject (3)
V.
WT
year
ST
Technology and safety of blasting

operations (3)
Theory and technology of pyrotechnical
products and compositions (4)
Technology of composite & primary
explosives (4)
Safety engineering (7)
Specialised excursion
year
Theory of action of explosion (7)
Special technology of explosives (6)
Basic applied ballistics (4)
ST
Thesis (30)
Obligatory-optional subjects in the 2nd step of MS-level studies:

WT
Special analysis of explosives (3)

Basics of ammunition construction (3)
ST
Basics of gun construction (2)

Exploitation of explosion energy (2)
Basics of internal ballistics and barrel
guns (3)
SUBJECTS OF THE 2ND STATE EXAMINATION

branch Technology of explosives
Theory of explosives
Theory of action of explosion
Technology of explosives /pyrotechnical products /propellants
branch Safety engineering

Safety engineering
Annotation of special subjects

The subject is a continuation of physics, physical chemistry, inorganic and organic
chemistries. It involves definition terms, classification of explosives and processes,
energetic aspects of explosives and explosion, findings about burning, thermal explosion,
theory and practice of detonating and initiating of explosives, transition from burning to
detonation, aspects of sensitivity and stability of energetic materials in general.
Technology and safety of blasting operations

The subject is a continuation of physics, physical chemistry, inorganic and organic
chemistries, theory of explosives and action of explosion, technology of secondary
explosives. It deals with the processes in the charge and its environment, acoustic
approach, consequences of similarity at detonation, effect of secondary explosive in rock,
zone of action of charge, its fixation, disintegration and fragmentation of rock. A part of the
subject also deals with regulations for technically correct and safe performing of blasting
operations. After finishing this subject and subsequent passing a 10-day practical training
as blasters assistant, the student can apply for sitting blaster exams at the respective
regional mining authority and thereby obtainthe certificate of blaster.
Technology of basic explosives
The subject is a continuation of inorganic chemistry and particularly organic chemistry,
chemical engineering and toxicology. It deals with properties and ionic equilibria of nitric
acid, sulphuric acid, and their mixtures (i.e. nitrating mixtures), structure, reactivity and
properties of nitro arenes, nitro paraffins, N-nitro compounds (nitramines), N-nitroso
compounds (nitrosamines) and nitrate esters; mechanism of basic types of nitration and Nnitrosation, technology of production of technically attractive polynitro arenes, nitramines,
nitrosamines and nitrate esters. Also emphasized are some procedures of syntheses of raw
materials and intermediates, such as N-Mannich condensation, some formosation
reactions, Michael addition etc.
Theory and technology of pyrotechnical compositions and products
The subject is a continuation of theory of explosives, and technology of secondary and
primary explosives. It deals with physical and chemical principles of pyrotechnical reaction,
characteristics and basic methods of calculation of burning wave in a pyrotechnical
composition, classification of pyrotechnical compositions, principles of construction and
arrangement of various kinds of pyrotechnical compositions. It involves principles of
technology of production, processing and incorporation of pyrotechnical compositions into
final products and specific safety problems connected with production and processing of
pyrotechnical compositions.
Technology of composite and primary explosives
The subject is a continuation of organic and inorganic chemistries, and theory of explosives.
Primary explosives are presented in a classification in groups according to molecular
structure, chemistry and technology of their production. Composite explosives are divided
into civil and military ones and in both cases their basic classification is presented along
with requirements for their performance and stability characteristics. Both types of
explosives are dealt with taking into account their types, technologies, and technological
equipment used in their production, and, in the case of military explosives, also the way of
incorporation into ammunition.
Technology of propellants
The subject is a continuation of physical chemistry, theory of explosives, technology of
basic explosives, and ecology. It deals with the individual types of propellants, theory of
burning of their various types, their properties and applications, technology of production,
and their testing. Also presented are general requirements for chemical stability, storability
and environmental problems connected with their production and use.
Safety engineering
The subject is a continuation of physics, physical chemistry, chemical engineering, theory of
explosives, and process engineering. It deals with the problems of technique of risk
evaluation, basic concepts of risk analysis and evaluation of consequences of accidents in
industry, prevention of accidental explosion in industry and in the industrial processes that
do not involve declared explosives and, last but not least, improving the safety of chemical
plants.

The subject is a continuation of physics, chemical engineering, theory of explosives,
technology of basic explosives, and technology of propellants. It deals with the shock waves
in air, shock waves in condensed media, basics of seismology, inner explosion, effect of
shock waves on objects and persons, accumulation, and fragmentation. Attention is also
given to explosive pressing, forming and cladding of materials.
Special technology of explosives
The subject is a continuation of technology of basic explosives, and technology of
propellants. It deals with low vulnerability explosives, energetic polymers, technology of
assembling and dismantling ammunition, liquidation of demilitarised explosives and
ammunition, and detection of explosives.
Basics of applied ballistics
The subject is a continuation of physics, physical chemistry, theory of explosives,
technology of propellants, and technology of explosives. It deals with classification of
propellants, methods of their testing, basic applications of propellants in the field of barrel
arms and rockets, theory of ignition, pyrostatics and internal ballistics, basics of construction
of small arms ammunition, gun charges, and rocket engines. The subject also involves
practical examples of typical solution of ammunition objects from the point of view of
propellants.
Labs of the specialisation IVth year
The subject is a continuation of organic chemistry, analytical chemistry, technology of basic
explosives, and theory of explosives.
In the winter term it involves laboratory practice in special organic chemistry involving
preparation of basic, chemically individual explosives such as crystalline nitrate esters,
nitramines, and nitro aromatics along with analytical checking of quality of these products
and measuring their physical characteristics. At the end, skilled students perform
demanding multi-step syntheses of substances such as 2,4,6,8,10,12-hexanitro2,4,6,8,10,12-hexaaza-isowurtzitate etc.
In the summer term the students carry out measurements of explosion characteristics of
secondary explosives (fall sensitivity, sensitivity to friction, detonation velocity, TNT
equivalent, brisance, detonation transfer), measurements of characteristics of shock waves
generated by explosion, and deflagration. These lab exercises also include a 16-hour
practice in preparation of pyrotechnical compositions.
Labs of the specialisation Vth year
The subject is a continuation of inorganic and organic chemistry, physics, physical and
analytical chemistry, theory of explosives, technology of basic explosives, technology of
propellants, safety engineering, and theory of action of explosion. The students activities in
winter term are directed individually and almost entirely in the framework of scientificresearch tasks formulated by external bodies (grants, cooperation with military research
institutes of Czech and Slovak armies, police bodies, inclusive of the cooperation with the
industry).
The above activitieslead to formulation and subsequent solution of the Thesis project, which
mostly takes place at the workplaces of the above-mentioned cooperating organisations.
Obligatory-optional subjects
Special analysis of explosives
The subject is a continuation of analytical chemistry, physical chemistry, technology of basic
explosives, and technology of primary and secondary explosives. It includes detection of
traces of explosives at various surfaces, methods of taking and concentrating samples, the
detection methods proper, the analytical methods used, their principles, special analysis of
gunshot products from firearms, and single-purpose stationary and portable instruments for
detection of explosives.
Basics of construction of ammunition

The subject is a continuation of physics, theory of explosives, and technology of secondary
explosives. It involves the classification and basic terms of construction of ammunition:
small and middle arms ammunition, charges for in-shot and out-shot devices, ammunition
for hand anti-tank arms and trench mortars, signal, illuminating and imitation devices,
artillery ammunition, pioneer work ammunition, hand-grenades. Also presented are
perspectives of further development of small and middle arm ammunition and artillery
ammunition.
Basics of construction of arms
The subject is a continuation of physics, theory of explosives, and theory of action of
explosion. It includes the principles of barrel firearms, classification of barrel firearms and
their main characteristics, main properties of firearms, the functional cycle and main
components of a firearm, parts of gun-carriage and its purpose, guns, and mortars, classical
automatic firearms and highly quick-firing automatic firearms
Basics of internal ballistics of barrel firearms
The subject is a continuation of physics, theory of explosives, and technology of propellants.
It involves the processes taking place in a barrel during gunshot, parameters of internal
ballistics, their notation, course of internal-ballistic characteristic quantities, the equation of
state for constant volume and its applications, basic equations of internal ballistics, systems
of equations of internal ballistics and their solution, measurements in internal ballistics.
PhD-LEVEL STUDIES
Graduates from the MS-level study having necessary prerequisites for scientific work can
continue their education in the 3-year postgraduate study in the study program of
Chemistry and chemical technology, line Organic technology.
The PhD-level studies have a form of full-time type or a combination of full-time and
distance learning. In the former case, the PhD student (doctorand) fully participates in the
scientific work of the respective workplace. The standard form of full-time study lasts three
years, the combined form lasts five years. The study follows an individual study plan under
the guidance of a supervisor. The study is completed by a state examination and viva voce
of the submitted dissertation Thesis, which certify the ability and state of readiness for an
independent activity in the field of research and development. The PhD state examination
consists of 3-4 subjects chosen in accordance with the doctorands study line from among
the subjects of the study line of theory and technology of explosives (see above in MS-level
studies) or, as the case may be, from other subjects provided by the Faculty (e.g. physics of
polymers, metallic materials etc.).
The graduate obtains the title of Doctor (in Czech doktor abbreviated as Ph.D.).
LICENCE STUDY
This study (retraining and further increasing qualification of participants) takes place on the
basis of economic contracts with the applicants and/or their employers. The price of this
form of study for Czech and Slovak citizens (i. e., in the Czech language) is $1300, and the
study lasts four terms, its extent being 400 lesson periods; it is organized in the form of
three-day workshops each month. In the licence study the Department offers two courses:
Theory and technology of explosives
This course is designed for graduates (exceptionally undergraduates) working in plants and
factories producing, processing or dismantling explosives and ammunition, as well as those
dealing with storage and sale of explosives and explosion-risky substances. By completing
this course, the student will acquire a specialist ability for realization, control, and
management activities in the field of research and development, production and processing,
testing, storage, transport and sale of energetic materials. The course provides basic
information in the area of theory, chemistry and technology of energetic materials, and also
in the field of protection of various objects against explosions of gases, vapours or
dispersions of inflammable powders/ dusts, as well as in the field of testing and special
analysis of explosives, basics of ballistics, construction of ammunition and arms.
Rock disintegration by explosion (Theory and practise of blasting)
The applicants to this course must have A levels and certificate of blaster; the study is at
Bc-level. After a successful completion, the student takes the examination of technical
manager of blasting operations supervised by the committee of the Czech Mining Authority.
The study is realised in cooperation with the Czech Mining Authority and is acknowledged
as a supplementary study to the requirements for qualification and expert ability of
applicants for the examination of technical manager of blasting operations. The lectures in
this course are directed to the theory of explosives, action of explosion on adjacent media,
means of blasting technique, technology of drilling, safety of work, legal aspects of carrying
out blasting operations, technology of blasting operations in opencut mining, surface and
underground destructions, blasting techniques in opencut and underground coal mining.
INTERNATIONAL SEMINAR
Each year the Department organises an international seminar called New Trends in
Research of Energetic Materials, which is designed particularly for students, doctorands
and young scientific-technological workers and university teachers working in the field of
energetic materials. These seminars are free for the participants. The aim of the seminar is
to teach the coming young generation scientific-research and for the technological workers
to present their results to international audience. In April 2002 the seminar was attended by
116 participants from 19 countries.
FOREIGN STUDENTS STUDIES

Foreigners can study both the undergraduate and postgraduate programs at DTTX
in two ways:
a) Free of charge if they are able to study in the Czech language (they have to pass an
examination in Czech).
b) For the price of $ 6,000 per year if the education is in English or another world
language. However, the foreign language study is being prepared at present and
students can contact management of the Faculty of Chemical Technology for further
information (E-mail: dekanat.fcht@upce.cz).
THE LOCATION OF THE DEPARTMENT

The facilities of DTTX are located in Technological Hall of Faculty of Chemical Technology
at village Doubravice near Pardubice. The location of DTTX can be reached from main
railway station in Pardubice by means of trolleybus No. 3 (get off at trolleybus stop Semtnzastavka).
CONTENT
PREFACE
10
B.P. Aduev, E.D. Aluker and A.G. Krechetov

Kemerovo State University, Krasnaya 6, Kemerovo, 650043, Russia
PRE-EXPLOSIVE PROCESSES IN HEAVY METAL AZIDES
12
H. G. Ang
National University of Singapore Faculty of Science HEDM Research Laboratories 3 Science Drive 3 Singapore 117543
SYNTHESIS AND REACTIVITY OF (CF3)2PN3, (CF3)2ASN3, CF3ASCIN3

AND CF3AS(N3)2 AZIDES
18
Eduard D. Aluker*, Boris P. Aduev** and Alexander G. Krechetov*

* Kemerovo State University, Krasnaya, 6, Kemerovo, 650043, Russia
** Kemerovo Branch ICSSM SB RAS, Sovetcky, 40, Kemerovo, 650099, Russia
PRE-DETONATION PHENOMENA IN HEAVY METAL AZIDES
30
Alexander M. Astachov*, Ludmila A. Kruglyakova*, Irina V. Gelemurzina*, Alexander D. Vasiliev**,

and Rudolf S. Stepanov*
* Siberian State Technological University, Prosp. Mira 82, 660049 Krasnoyarsk, Russia
** Institute of Physics RAS (Sib. branch),Akademgorodok, 660036 Krasnoyarsk, Russia
SIMPLE METHOD OF SYNTHESIS AND CHARACTERIZATION OF SOME NITROCYANAMIDE

SALTS
36
Daniel Buczkowski* and Bogdan Zygmunt**,

* Institute of Industrial Organic Chemistry, 6 Annopol St, 03-236 Warszawa, PL
** WIFAMA-PREXER, 45 Niciarniana St, 92-320 d, PL
INFLUENCE OF AMMONIUM NITRATE PRILLS POROSITY AND DIMENSIONS

ON DETONATION VELOCITY OF ANFO EXPLOSIVES
45
Pawe Buraczewski*, Marek Sutkowski*, Andrzej Teodorczyk*and Shengjun Zhong**

*Institute of Heat Engeeniring, Warsaw University of Technology, Nowowiejska 21/25, 00-665 Warsaw, Poland
**Industrial Explosion Protection Institute, Northeastern University, Shenyang, China
DETO2D THE COMPUTER CODE FOR SIMULATIONS OF GASEOUS DETONATIONS

IN COMPLEX GEOMETRIES
52
L. ai*, R. Halle**and Z. Ester**

*Ministry of Interior, Zagreb, Republic of Croatia
**Faculty of Mining, Geology and Petroleum engineering, University of Zagreb, Republic of Croatia
EMULSIONS EXPLOSIVES ELMULEX AND ELMULEXAL

Technical characteristics - testing and standards
63
S. Cudzio*, S. Gachet**, A. Huczko***, M. Monthioux** and W. A. Trzciski*

*Military University of Technology Kaliskiego 2, 00-908 Warsaw 49, POLAND
**Carbons and Carbon-Containing Materials, CEMES, Rue Jeanne Marvig 29, 31055 Toulouse, FRANCE
***Warsaw University, Department of Chemistry, Pasteura 1, 02-093 Warsaw, POLAND
SYNTHESIS OF CERAMIC AND CARBON NANOSTRUCTURES BY SELF-SUSTAINING

COMBUSTION OF MIXTURES OF HALOGENATED HYDROCARBONS WITH REDUCERS
69
Radka Czeczotkov and Betislav Janovsk

Dept. of Theory and Technology of Explosives, University of Pardubice, 532 10 Pardubice, Czech Republic
THE INFLUENCE OF THE TYPE AND INTENSITY OF IGNITION SOURCES ON EXPLOSION

PARAMETERS OF THE GASEOUS, DUSTS AND HYBRID MIXTURES
76
Andrzej Dbkowski, Artur Kozak and Andrzej Teodorczyk

Warsaw University of Technology, ul. Nowowiejska 21/25, 00-665 Warszawa,Poland
THE INITIATION OF GASEOUS DETONATIONS IN H2-O2 MIXTURES

BY INCIDENT SHOCK WAVE
90
Mario Dobrilovi, Darko Vrkljan and Zvonimir Ester

University Of Zagreb, Faculty of Mining, Geology and Petroleum Engineering, 10000 Zagreb, Pierottijeva 6, Croatia
DETERMINATION OF THERMAL STABILITY, RESISTANCE TO WATER AND RESISTANCE

TO HYDROSTATIC PRESSURE OF ELECTRIC DETONATORS ACCORDING TO DRAFTS
OF EUROPEAN STANDARD
101
Eisner A.* , Adam M.*, Mikulkov P*, Ventura K.* and Kouba M.**
* University of Pardubice, Department of Analytical Chemistry, Nm. s. Legi 565, 532 10 Pardubice, CZ
** University of Pardubice, Department of Theory and Technology of Explosives, Doubravice 41, 532 10 Pardubice, CZ
FIRST EXPERIENCE WITH ANALYSIS AFTER LIQUIDATION OF SMOKELESS POWDERS
109
Darina Fierov *,**, Amer Hameed*, Timothy A. Rose*, John G. Hetherington*

and Stanislav Prochzka**
* Cranfield University, Shrivenham, SN6 8LA SWINDON, United Kingdom
** Military Academy, Kounicova 65, 612 00 BRNO, Czech Republic
SYSTEMATIC STUDY OF SIMULATED MINE EXPLOSIONS USING AUTODYN
113
Levi Gottlieb and Gady Korogodsky

Rafael, Department M1,
POB 2250 Haifa 31021, Israel
DEGRADATION, REARRANGEMENT AND REDUCTIVE ACETYLATION

OF HEXABENZYLHEXAAZAISOWURZITANE
121
Martina Chovancov, Peter Oko, Jozef Lopch, Milo Lazar and Albeta Pechov
VTS Zhorie (Military Technical and Testing Institute), 905 24 Senica, SK
CHEMICAL AND THERMAL STABILITY OF FLEXIBLE PBXs
128
Martina Chovancov, Peter Oko, Jozef Lopch, Rastislav evk and ubo avojsk
VTS Zhorie (Military Technical and Testing Institute), 905 24 Senica, SK
AGEING INFLUENCE ON PLASTIC EXPLOSIVES ON THE BASE OF RDX
137
M.A. Ilyushin, I.V. Tselinsky, I.A. Ugryumov, A.Yu. Zhilin and A.S. Kozlov
Saint-Petersburg State Institute of Technology, Russia,
COORDINATION COMPLEXES AS INORGANIC PRIMARY EXPLOSIVES
146
H. S. Jadhav*, D. D. Dhavale*, M. B. Talawar*** S. N. Asthana*** and V. N. Krishnamurthy**

*Department of Chemistry, University of Pune, Pune- 411 007, India
**DRDO/ISRO Cells, University of Pune, Pune- 411 007, India
***High energy Materials Research Laboratory, Pashan, Pune, India
1-(3,5-DINITROPHENYL), -3,3-DINITROAZETIDINE: A NEW ENERGETIC MATERIALS
153
Zdenk Jalov*, Pavel Mareek**, Kamil Dudek**

* Department of Theory and Technology of Explosives, University of Pardubice, Czech Republic,
**Explosia, a.s. VPCH, Pardubice, Czech Republic
CRYSTALLIZATION OF HEXANITROSTILBENE
160
Martin Kouba, Svatopluk Zeman and Eva Zemanov

Departmet of Theory & Technology of Explosives, University of Pardubice, CZ-532 10 Pardubice, Czech Republic
STUDY OF DECOMPOSITION OF TNT BY HEAT AND SHOCK
166
Georgy.D. Kozak, Nikolai.I. Akinin, Vlada.M. Raikova and Svetlana.V. Arinina

Mendeleev University of Chemical Technology, Miusskaya sq. 9, Moscow A-47, Russia
EXPLOSION HAZARD OF SOME ORGANIC PEROXIDES
173
Queenie S.M. Kwok*, David E.G. Jones* and Peeter Kruus**

*Canadian Explosives Research Laboratory, 555 Booth St., Ottawa, ON K1A 0G1 Canada
**Carleton University, 1125 Colonel By Dr., Ottawa ON K1S 5B6 Canada
INVESTIGATION OF THE WETTABILITY OF AMMONIUM NITRATE PRILLS
183
Michel H. Lefebvre
Dept Chemistry RMA, 30 Av de la Renaissance, 1000 Brussels, Belgium
DETERMINATION OF THE POWER OUTPUT OF DETONATORS
198
Katarzyna Lipiska *, Marek Lipiski * and Andrzej Maranda **

** Military University of Technology, 2 Kaliskiego St, 00-908 Warszawa 49, PL
INFLUENCE OF DEMILITARIZED DOUBLE BASE PROPELLANTS ON DETONATION

PARAMETERS OF AMMONALS
205
Ji Strnad and Ji Majzlk

Department of Theory & Technology of Explosives, University of Pardubice, CZ 532 10 Pardubice, Czech Republic
METHODOLOGY OF TESTS OF SENSITIVITY OF ENERGETIC MATERIALS

TO ELECTROSTATIC DISCHARGE
212
Viktor A. Malchevsky and Natalia A. Zarytovskaya

D.I. Mendeleev University of Chemical Technology, Miusskaya pl. 9, Moscow, 125047, Russia
ANALYSIS OF THE DEFORMATION AND STRENGTH PROPERTIES

OF ENERGY POLYMERIC COMPOSITES (EPC)
221
Didier Mathieu and Eric Germaneau

Commissariat lEnergie Atomique, BP 16, 37260 Monts, France
A FAST NON-SELFCONSISTENT ELECTRONEGATIVITY EQUALIZATION METHOD

WITH APPLICATIONS IN THE FIELD OF ENERGETIC MATERIALS
229
Robert Maty
Department of Theory and Technology of Explosives, University of Pardubice, 532 10 Pardubice, Czech Republic
CHEMICAL DECOMPOSITION OF TRIACETONE TRIPEROXIDE

AND HEXAMETHYLENETRIPEROXIDEDIAMIDE
241
Maciej Miszczak, Ewa Szymanowska, Beata migelska and Bogumia Bachnio

Research Department of Combat Means, Military Institute of Armament Technology,
Wyszyskiego 7, Str., 05-220 Zielonka, POLAND
QUANTITATIVE ANALYSIS OF SOME TECHNOLOGICAL COMPONENTS

USED IN INHIBITED HOMOGENEOUS ROCKET PROPELLANTS
BY MEANS OF THIN LAYER CHROMATOGRAPHY
248
Ale Nepovm*, Anja Hebner**, Andre Gerth**, Hartmut Thomas*** and Tom Vank*
* Institute of Organic Chemistry and Biochemistry, Flemingovo nm. 2, 166-10 Praha 6, CZ
** Bioplanta GmbH,Benndorfer Landstrae 2, D-04509 Delitzsch
*** Wasag Decon GmbH, Werkstrae 111, 45721 Haltern, GE
BIODEGRADATION OF TNT BY IN VITRO CULTIVATED PLANTS:

A MODEL SYSTEM FOR STUDY OF DEGRADATION PROCESSES
255
Andrzej Orzechowski*, Andrzej Maranda**, Dorota Powaa* and Jacek Borkowski***

** Military University of Technology, 2 Kaliskiego St, 00-908 Warszawa 49, PL
*** Military Institute of Armament Technology, 7 Wyszyskiego St, 05-220 Zielonka, PL
DETERMINATION OF SENSITIVITY OF PLASTIC EXPLOSIVE CONTAINING NTO
259
Jzef Paszula*, Andrzej Maranda*, Andrzej Papliski*, Barbara Gobek* and Johann Kasperski **
* Military University of Technology, 00-908 Warszawa, ul. Kaliskiego 2, Poland
** Blastexpol, 59-145 Duninw, Poland
A COMPARATIVE ANALYSIS OF BLAST WAVES PARAMETERS GENERATED BY EXPLOSION

266
OF EMULSION EXPLOSIVES AND DYNAMITES
Miroslav Pospil*, Pavla apkov*, Pavel Vvra** and Svatopluk Zeman**

* Department of Chemical Physics and Optics, Faculty of Mathematics and Physics, Charles University Prague,
Ke Karlovu 3, 12116 Prague 2, Czech Republic.
** Department of Theory and Technology of Explosives, University of Pardubice, 53210 Pardubice, Czech Republic.
CLASSICAL MOLECULAR DYNAMICS SIMULATIONS OF RDX DECOMPOSITION

UNDER HIGH PRESSURE
274
Vlada M. Raikova, Georgy D. Kozak and Evgeny A. Likholatov

Mendeleev University of Chemical Technology, Miusskaya sq. 9, Moscow A-47, Russia
THE CHEMICAL KINETIC AT DETONATION OF NITROESTERS SOLUTIONS
284
Traian Rotariu and Octavian Orban

Military Technical Academy, Department of Ammunitions, Missiles and Explosives,
81-83 George Cosbuc Boulevard, Bucharest, RO
SOME REGARDS ABOUT COLD POWDERS PLASTICIZERS
295
Karl P. Rudolf
Diehl Munitionssysteme GmbH & Co KG, PBX-Center Maasberg, Karl-Diehl-Strae 1, D-66620 Nonnweiler, Germany
IMPROVED INSENSITIVE HYTEMP/DOA BONDED HMX AND RDX MIXTURES

BY PASTE PROCESS
300
Jakub eleovsk*, Ji Pachm*and Marcel Hanus**

* Department of Theory and Technology of Explosives, University of Pardubice, 532 10 Pardubice, Czech Republic
** Military Institute for Weapon and Ammunition Technology, Dlouh 300, 763 21 Slavin, Czech Republic
CHANGES IN SENSITIVENESS OF FLEGMATIZED HIGH EXPLOSIVES

AFTER ARTIFICIAL AGING
309
Gurdip Singh and S. Prem Felix,

Chemistry Department, DDU Gorakhpur University,Gorakhpur 273 009
THERMOLYSIS OF KETO-RDX AND ITS PLASTIC BONDED EXPLOSIVES

WITH FLUORO POLYMERS
322
Gurdip Singh* and Durgesh Kumar Pandey

Chemistry Department, DDU Gorakhpur University,Gorakhpur - 273 009, India
THERMOLYSIS AND KINETICS OF SOME BIS(PROPYLENEDIAMINE) METAL NITRATE

COMPLEXES
329
C. R. Siviour, S.M. Walley, W.G. Proud and J.E. Field

PCS Group, Cavendish Laboratory, Cambridge CB3 0HE, UK
HOPKINSON BAR STUDIES ON POLYMER BONDED EXPLOSIVES
339
Radovan Skacel, Bretislav Janovsky, and Jaroslav Svihovsky

University of Pardubice, Department of Theory and Technology of Explosives,
Studentska 95, Pardubice 532 10, Czech Republic
PROCESSES IN THE VESSEL DURING BLEVE EFFECT
350
Rudolf S. Stepanov, Ludmila A. Kruglyakova and Alexander M. Astachov

Siberian State Technological University, 660049, Krasnoyarsk, prosp.Mira, 82, Russia
GEMINAL TRINITROCOMPOUNDS THERMAL DECOMPOSITION

UNDER NON-ISOTHERMIC CONDITIONS
362
J. Stockinger and T. Reed

Naval Surface Warfare Center, Crane Division, 300 Highway 361, Crane, Indiana 47522 USA
ELIMINATION OF LEAD AND LEAD CONTAMINANTS IN THE MK 141 MOD 0

DIVERSIONARY CHARGE
367
Muhamed Sueska, Sanja Matei Muani and Maa Raji

Brodarski Institute, Av. V. Holjevca 20, 10000 Zagreb, Croatia
DETERMINATION OF ARRHENIUS KINETIC CONSTANTS FOR DOUBLE BASE PROPELLANTS

BY NON-ISOTHERMAL DSC MEASUREMENTS.
374
INFLUENCE OF SAMPLE SELF-HEATING.
Muhamed Sueska, Maa Raji, Sanja Matei Muani, Sanko Bakija*, Ruica uljak*,
Vladimir Jagui* and Slavko urak*
Brodarski Institute, Av. V. Holjevca 20, 10000 Zagreb, Croatia
* Ministarstvo obrane RH, Bauerova 33, 10000 Zagreb, Croatia
APPLICATION OF DSC IN STABILITY STUDIES OF DOUBLE BASE PROPELLANTS
392
Jaroslav vihovsk, Betislav Janovsk and Radovan Skcel

Department of Theory & Technology of Explosives, University of Pardubice, CZ-532 10 Pardubice, Czech Republic
MEASURING OF THE BLEVE INTERNAL PROCESSES
405
A. Teodorczyk* and G.O. Thomas**

*Warsaw University of Technology, Nowowiejska 21/25 00-665 Warszawa, Poland
**Centre for Explosion Studies,University of Wales, Aberystwyth,Ceredigion, UK SY23 3BZ
FLAME AND DETONATION ARRESTERS EUROPEAN NORMATIVE STANDARD

FOR TESTING
411
Jasmin Terzic, Alija Lekic and Berko Zecevic

Mechanical Engineering Faculty, Defense Technologies Department,
Vilsonovo setaliste 9, 71000 Sarajevo, Bosnia and Herzegovina
PREDICTION THE THEORETICAL INTERIOR BALLISTIC PROPERTIES

OF SOLID PROPELLANT ROCKET MOTORS
420
W. A. Trzciski, L. Szymaczyk, and S. Cudzio

Military University of Technology, Kaliskiego 2, 00 980 Warsaw, Poland
DETONATION CHARACTERISTICS OF LOW-SENSITIVITY NTO-BASED EXPLOSIVES
436
J. Vgenknecht, G. Hurtosov and L. Adamk

Department of Theory and Technology of Explosives, University of Pardubice, 532 10 Pardubice, Czech Republic
UTILIZATION OF NUMERICAL SIMULATION METHODS TO STUDY POSSIBLE IFLUENCE

OF INITIATION ON THE PERFORMANCE OF LINEAR SHAPED CHARGES
445
R. Varga * and Ulbrich P. **

* Department of Theory and Technology of Explosives, Univerzity of Pardubice,532 10 Pardubice, Czech Republic
** Institute of Forensic Science of the Slovak Police Corps, Sklabinsk 1, Bratislava 812 72, Slovak Republic
TRACE ANALYSIS OF POST EXPLOSION RESIDUES OF INDUSTRIAL EXPLOSIVES

BY MEANS OF GC-ECD AND IC
462
Pavel Vvra* and Miroslav Pospil**

* Department of Theory and Technology of Explosives, Faculty of Chemical Technology, University of Pardubice,
53210 Pardubice, Czech Republic
** Department of Chemical Physics and Optics, Faculty of Mathematics and Physics, Charles University,
12116 Prague, Czech Republic
EFFECT OF INTERMOLECULAR FORCES ON SOME PROPERTIES OF EXPLOSIVES

II. INFLUENCE OF NON-COVALENT HYDROGEN BOND
474
Fred Volk, Helmut Bathelt, Richard Jakob and Dietmar Mueller

Fraunhofer Institut Chemische Technologie (ICT),
Joseph-von Fraunhofer-Str. 7, 76327 Pfinztal-Berghausen, Germany
EVALUATION OF THE EFFICIENCY OF IGNITION PROCESSES

BY COMBUSTION UNDER DIFFERENT PRESSURES
483
Richard Wild
Diehl Munitionssysteme GmbH & Co KG, PBX-Center Maasberg,
Karl-Diehl-Strae 1, D-66620 Nonnweiler, Germany
LONGTERM BEHAVIOUR OF THE HARDNESS MEASURED

ON CURE CAST SAMPLES
497
Stephan Wilker, Gabriele Pantel and Lutz Stottmeister

WIWEB ASt Heimerzheim, Groes Cent, 53913 Swisttal (DE)
STABILITY ANALYSES OF POROUS PROPELLANTS
507
Waldemar Witkowski*, Karol Buchalik*, Radoslaw Trebinski**, Andrzej Maranda**

and Andrzej Teodorczyk***
* Institute of Industrial Organic Chemistry, Annopol 6, 03-236 Warsaw, Poland
** Military University of Technology, Kaliskiego 2, 00-908 Warsaw, Poland;
*** Warsaw University of Technology, ITC, Nowowiejska 25, 00-665 Warsaw, Poland
OVERPRESSURE GENERATED BY DETONATIONS OF ELONGATED GAS CLOUDS
525
T. Wolszakiewicz* and A. Ksiczak**

* Institute of Organic Industrial Chemistry, Annopol 6, 03 236 Warsaw, Poland
** Department of Chemistry, Warsaw University of Technology, Noakowskiego 3, 00 - 664 Warsaw, Poland
PHYSICOCHEMICAL PROPERTIES OF NITROCELLULOSE MIXTURES WITH LOW MOLECULAR

COMPOUNDS IN THE SHAPE OF FOILS AND IN WATER SUSPENSIONS
535
Shu Yuanjie, Li Hongzhen, Huang Yigang and Liu Shijun
Institute of Chemical Materials CAEP, 621900, Mianyang, Sichuan, China
SYNTHESIS OF N-ACETYL-3,3-DINITROAZETIDINE
544
Berko Zecevic, Jasmin Terzic and Mario Bakarad

INFLUENCE OF THE SOLID PROPELLANT GRAINS PROCESSING ON BURNING RATE

OF DOUBLE BASE ROCKET PROPELLANTS
548
Igor Zori, Dalibor Kuhinek and Mario Dobrilovi

University of Zagreb, Faculty of Mining, Geology and Petroleum Engineering,
Pierottijeva 6, 10000 Zagreb, CROATIA
THE INFLUENCE OF MEASUREMENT METHODE ON THE INITIATION TIME MEASUREMENT

560
NECESSARY FOR PRIMER MIXTURE INITIATION OF ELECTRICAL DETONATOR
Product Education Session

SPARK TEST APPARATUS - ESZ DTTE
570
The apparatus for determination of sensitivity of various energetic materials

to high-voltage electrostatic discharge
OZM Research, Ltd.

Development and Production of Devices and Apparatus for Testing of Energetic Materials
571
PREFACE
This sixth seminar takes place in the year of 50th anniversary of the beginning
of education in the field of science and technology of explosives at the University
of Pardubice. Therefore, I consider it appropriate to briefly mention the history of this
education in Bohemia in general.
The education in the above-mentioned field started in Bohemia in 1920. Its centre was
Institute of explosives, essential oils, resins, rubbers, varnishes and blasting techniques
whose head was Prof. Dr. Cyril Krauz. At the time, the said Institute formed a part of faculty
of Chemical-technological Engineering in Prague (at present Prague Inst. of Chemical
Technology). The activities in this filed were interrupted by the German occupation of
Bohemia and Moravia in 1939-1945. After the post-war renewal of education at Czech
universities, the Department of Organic Technology at the Institute of Chemical Technology
(VCHT) in Prague established a subsection called Technology of Special Production
in the academic year 1952/1953. Dr. Ing. Josef Seifert, who was appointed the first head of
this subsection, prepared creation of similar Department of Technology of Special
Production at the Institute of Chemical Technology in Pardubice.
Ahead of the above-mentioned activities in Prague and in Pardubice, a Department of
Explosives was established at Military Technical College Brno in 1951. However, this
Department was dissolved as early as 1958. Many teachers of this Department, e.g.
a ballistic expert Prof. Frantiek Polansk, a chemist Prof. Jaromr imeek and Col. Prof.
Jindich Foltn engaged themselves in establishing and operation of Department
of Technology of Special Production in Pardubice (even after the dissolution of the
Department at the Military College).
Institute of Chemistry Pardubice was founded in 1950, renamed to Institute of Chemical
Technology in 1953, and became a part of University of Pardubice in 1994. As already
mentioned, the "Department of Technology of Special Production" began its activities within
this Institute from September 1953, and it was renamed to "Department of Technology
of Explosives" in 1962. At that time, Czechoslovakia belonged among leading exporters
of arms and ammunition, and the mining of inorganic raw materials was intensively
developed. Therefore, it is somewhat paradoxical that just at this time the Institutes
management stimulated dissolution of the Department as independent body. Starting from
January 1966 this Department thus became mere two subsections of Department of Organic
Technology and Department of Chemical Engineering of the Institute. Its independent status
was renewed in the autumn of 1986 under the name "Department of Theory and Technology
of Explosives" (DTTX). At present, this Department operates within Faculty of Chemical
Technology, the University of Pardubice.
So far the Department has given education to 312 graduates in MSc study plans, over
250 graduates in two four-term license courses (technology and blasting techniques) and 51
PhD graduates. In the mentioned types of education, altogether 52 65 students have been
enrolled each year recently, out of whom 10 15 being citizens of the Slovak Republic.
At present, the Department is one of only two workplaces of this kind within the territory
of European NATO countries. The same extent of studies as that provided by DTTX is also
available in Europe at Wojskowa akademia techniczna (Military University of Technology)
in Warsaw, Mendeleevs University of Chemical Technology in Moscow, Kazan State
University of Technology at Kazan, and Saint-Petersburg State Institute of Technology
at Saint Petersburg
10
The international seminars entitled New Trends in Research of Explosives belong

among the most significant activities of DTTX at present. The original purpose of these
meetings was to teach young research workers how to present their results in front
of scientific audience. This original aim was later complemented by not only the exchange
of findings and experience but in particular making friendly contacts between the members
of the beginning young generation of experts in the area of energetic materials coming from
many countries all over the world.
With regard to economic situation of the middle European countries, Balkan and former
Soviet Union, no fee was and is asked from the participants. The main load of financing was
always on DTTX. However, in the last two years a very significant financial support was
regularly provided by European Office of Aerospace Research & Development of the USAF,
and in this year also by Austin Detonator Comp. (Vsetn), Indet Safety Systems Comp.
(Vsetn) and OZM, Inc. Hrochv Tnec. In the case of the 5th and 6th Seminars, some
financial support was also provided by Dr. Oldrich Machacek, president of the Universal
Tech. Corporation, Dallas, who (as chance would have it) is the first PhD graduate from
DTTX. The efficient help in ensuring smooth and successful course of the meeting obtained
from all these institutions and individuals is gratefully acknowledged. We greatly appreciate
it that thanks to this sponsoring all the above-mentioned specifics of earlier seminars could
be maintained.
Finally, I wish to thank the members of the Scientific Committee, the authors of all the
seminar papers and, last but not least, you, the participants of this seminar, for its success
and its influence on the continued success and growth of all future meetings at our
University of young people and university teachers working in the field of teaching,
research, development, processing, analysing and application of all kinds of energetic
materials.
Allow me to use this opportunity for inviting you in the name of my co-workers and
myself: we are looking forward to meet you at the seventh seminar in the second half
of April 2004 in the Aula Magna of our University.
Pardubice, March 10th, 2003

Svatopluk Z e m a n
11
SPACE-TIME CHARACTERISTICS OF PRE-DETONATION

LUMINISCENCE ORIGIN IN HEAVY METAL AZIDES
Eduard D. Aluker*, Boris P. Aduev**, Alexander G. Krechetov*,
Anatoly Yu. Mitrofamov* and Evgeny V. Tupitsin*
* Kemerovo State University,
Krasnaya, 6, Kemerovo, 650043, Russia
** Kemerovo Branch ICSSM SB RAS,
Sovetcky, 40, Kemerovo, 650099, Russia
Abstract:
The topography of the origin and propagation of pre-explosive luminescence in AgN3
has been investigated for the first time. The origin of the explosive chain reaction was
found to be of hot spot nature. The propagation rate of the chain reaction was 1500300
m/s, leading to the conclusion that propagation results from hole diffusion.
Keyword:
1.
heavy metal azides, hot spots, origin, chain reaction
INTRODUCTION
For 1990s an intrigued trend in physics of explosive decomposition of energetic

materials was appeared. Simple and clear considerations lay the foundation for this trend:
initial stages of explosive decomposition are provided by the processes occurring in the
virgin (undestructed yet) crystal lattice and, hence, to investigate these stages one have to use
the solid-state physics approach. The first theoretical works in this field appeared in the
USA.[1-4] As to experiment, the possibility to use the above approach experimentally arised
from discovery of pre-explosive phenomena, such as pre-explosive luminescence and preexplosive conductivity, in heavy metal azides (HMAs),[5, 6]. Pre-explosive phenomena were
observed during induction (~100200 ns), i.e. during the time between an initiating pulse (a
laser pulse or an electron accelerator pulse) and the onset of mechanical fracture (or rather
the onset of deformation) of the sample. The explosion process appeared to be concantenated
at this stage, and the kinetics of the observed pre-explosive phenomena mirrors the kinetics
of the branching of an explosive decomposition chain reaction. Thus, pre-explosive
luminescence could be used to visualize the origin and propagation of the chain reaction of
explosive decomposition.
This paper focuses on two long-standing problems: the hot-spot origin of explosive
decomposition[5] (the so-called hot-spot model) and the propagation rate of an explosive
chain reaction.[6]
2.
EXPERIMENTAL APPARATUS
Silver azide whiskers having characteristic dimensions 100 100 m and length
12 mm were investigated.
A diagram of the experimental setup is shown in Fig. 1. A YAG:Nd3+ pulsed laser
( = 1064 nm, = 30 ps, W 5150 mJ/cm2) was used to irradiate the sample. The wave-
12
length of the laser radiation corresponds to the transmission band of the AgN3 sample. To
ensure uniform initiation, a focusing lens was used to compress the beam to 2 mm diameter
on a diffuse disperser, which was made from cloudy glass. The scattered radiation was
recollected with a second lens, which focused a 2-mm-diameter spot on the sample.
prism
prism prism
prism
mirror
diffuse
focusing
disperser
lenses
sample
optic
block
streakca
mera
filters
sweep mirror
triggering
shutter
triggering
pulse
generat
or
prism
YAG:
Nd3+
laser
Fig 1.
prism
Layout of experimental apparatus for

investigating spatial and temporal
characteristics of pre-explosive
luminescence.
The luminescence of the sample was recorded by a View-2 streak-camera having a

critical time resolution of 10 ps and a sensitivity of 1001000 photons and the resulting
image was captured by a CCD Image Sensor (made by the Ross Ltd, Russia). An optic block
projected an enlarged image of the sample to the photoelectric cathode of the streak-camera.
Space resolution was 50 m, as determined by a test object.
The axis of the sample image on the output screen of the streak-camera was
perpendicular to the sweep direction. Therefore, the image recorded represents the time
dependence of the spatial distribution of light emission through the crystal length. The image
is read from the screen of the CCD Image Sensor and is then interfaced to a computer.
The real time resolution of the images obtained is determined by the value
t = d/V, where d is the width of the sample image on the input screen and V is the sweep
rate.
An electronic circuit selects an initial monopulse from the picosecond pulse group of
the laser to initiate the sweep trigger. In order to compensate for sweep trigger delay, the
initiating laser pulse passes through an optical delay line (150 ns) consisting of a system of
prisms and mirrors, as shown in Fig. 1.
As shown in Fig. 2, to measure the glow propagation rate through the sample a
lightproof mask was added to the experimental apparatus to shield part of the sample from
the initial pulse of laser irradiation. The propagation rate was determined by the rate of the
glow extension into an area of the sample originally shielded from laser irradiation.
13

(laser beam)
sample
part of the sample
initiated by laser
irradiation
area of the sample

originally shielded
from laser irradiation
mask
3.
Fig 2.
Layout of initiation of the sample to

determine the glow propagation rate.
RESULTS
For pulse energies near the initiation threshold (5 mJ/cm2), pre-explosive

luminescence always occurs in discrete areas of the sample (Fig. 3a). The dimensions of
those luminescent areas increases with elapsed time.
With increased excitation energies (up to 15 mJ/cm2), more luminescent spots, as a
rule, were observed (Fig. 3b). After about 100 ns, the intensity of the light emission
increased abruptly, due to the presence of explosion products.[5]
Further increases in excitation energy (up to 150 mJ/cm2) caused a quasihomogeneous glow at the sweep onset (Fig. 3c), and the glow of explosion products is
observed at about 30 ns. In addition, under such excitation conditions, a short peak of
homogenous light emission is observed at the sweep onset (Fig. 3c). Special investigations
have shown that the duration of this peak coincides with the duration of the excitation pulse.
We conclude that this short peak is due to photoluminescence of the sample excited by the
laser pulse. It should be emphasized that the homogeneity of the light emission at the first
peak indicates that the hot-spot origin of pre-explosive luminescence is not related to
possible inhomogeneities in the laser beam.
A characteristic image of the propagation of light emission resulting from excitation
of only part of the sample is shown in Fig. 4. It is clear that the luminescent area seen in the
sample expands with time into the area originally shielded from laser irradiation. The rate of
propagation is constant. Its value, averaged over 20 samples, is 1500300 m/s. As the glow
is only the pre-explosive luminescence accompanying a chain reaction of the explosive
decomposition of the HMAs,[6] the rate measured is the rate of the propagation of the front of
the chain reaction through the sample length.
14
Fig 3.
Fig 4.
Time dependence of the

distribution of light
emission through the length
of a sample when the entire
sample is irradiated. The
initiation began at
t = 0; the luminous circle is
an artifact arising from the
background glow of the
output screen of the streakcamera used in the
experiment. t is the time
resolution for the specific
experiment.
Time dependence of the

distribution of light
emission through the length
of the crystal when only
part of the sample is
irradiated. Note that with
elapsed time, luminescence
extends into an area of the
sample originally shielded
from laser irradiation.
15
4.
DISCUSSION
4.1
Hot spot origin of a chain reaction
A hypothesis of the hot-spot nature of the origin of a chain reaction in explosives was
suggested in the fundamental work by Bowden and Yoffe.[7] This idea is well-known in the
thermal theory of explosions as the hot-spot model.[8]
However, although this model is almost universally used, it has never been directly
proven experimentally.[5,6] Therefore, the results shown in Fig. 3 seem to us to be significant.
These results clearly indicate that when the initiation threshold is slightly exceeded,
pre-explosive luminescence occurs in discrete spots. As soon as the initiation energy reaches
higher values, homogenous light emission occurs. This homogeneous light emission has
been previously investigated in detail.[5]
Since pre-explosive luminescence kinetics allows the visualization of the kinetics
of a chain reaction of explosive decomposition,[6] we can state that the hot-spot origin of this
chain reaction has been observed.
To date, the physical nature of hot spots is unknown. Previously, dislocations have been
suggested as the origin of hot spots. Thus, the spatial location of dislocations in a sample
should coincide with the hot spots revealed by pre-explosive luminescence. Experiments to
confirm this hypothesis are currently underway.
4.2
The rate of propagation of the explosive chain reaction
As has been stated earlier,[5,6] a necessary condition for the spatial distribution of a
chain reaction in HMAs is hole migration to an unexcited area of the sample. The simplest
mechanism for such a migration is hole diffusion.
We would like to stress the qualitative difference between hole diffusion and
conventional diffusion.[6] In conventional diffusion from a single source, the concentration of
diffusing particles decreases with distance from the source (the so-called diffusion profile).
In hole diffusion, the concentration of diffusing particles increases with distance as a result
of chain reactions, leading to a sharpening of the diffusion profile and the appearance of a
more-or-less visible reaction front, which separates the area occupied by the chain reactions
path from the undisturbed area of the sample. Such a process was considered as early as the
1940s by Zeldovich and Semenov[9] to estimate the contribution of diffusion to flame
propagation. To estimate the rate of the reaction front V, they proposed the following
equation:
V = D / ,
(1)
where D is the diffusion coefficient and is the characteristic time of a chain reaction, i.e.
the average time between two discrete events of the reaction (chain links).
Using Eq. (1), let us evaluate the reasonableness of the hypothesis of the diffusive nature
of the propagation of a chain reaction through a sample.
According to previous papers,[5,6] the accepted value of is 5109 s. If V = 1.5 km/s, D
100 cm2/s. This value is characteristic for electronic excitations in crystals.[10]
Thus, the rate of the propagation of the chain reaction front of AgN3 explosive
decomposition (1.5 km/s) that was measured experimentally can clearly be associated with
the rate of movement of the diffusion hole front.
16
5.
CONCLUSIONS
1. Hot spot origin of pre-explosive luminescence in silver azide under laser initiation has
been found. This fact give an evidence for hot spot nature of origin of an explosive chain
reaction.
2. The propagation rate was determined by the rate of the glow extension into an area of
the sample originally shielded from laser irradiation. The propagation rate of the chain
reaction was 1500300 m/s, leading to the conclusion that propagation results from hole
diffusion.
Acknowledgments
The authors would like to thank Celia M. Elliott for help in the paper design.
The work is supported by ISTC (grant No. 2180).
REFERENCES
[1]
A. TOKMAKOFF, M. D. FAYER, D. D. DLOTT: Chemical Reaction Initiation and Hot-Spot

Formation in Shocked Energetic Molecular Materials, Journal of Physical Chemistry , 97(9),
1901, 1993
[2] L. E. FRIED, A. J. RUGGERIO: Energy Transfer Rates in Primary, Secondary, and Insensitive
Explosives, Journal of Physical Chemistry, 98(39), 9786, 1994
[3] M. R. MANAA, L. E. FRIED: DFT and ab Initio Study of the Unimolecular Decomposition of
the Lowest Singlet and Triplet States of Nitromethane, Journal of Physical Chemistry,
102(48), 9884, 1998
[4] J. J. GILMAN: Chemical Reactions at Detonation Fronts in Solids, Philos. Mag. B, 71(6), 1057,
1995
[5] B.P. ADUEV, E.D. ALUKER, G.M. BELOKUROV, YU.A. ZAKHAROV, A.G. KRECHETOV:
Explosive Decomposition of Heavy Metal Azides, Journal of Experimental and Theoretical
Physics, 89(5), 906, 1999
[6] M.M. KUKLJA, B.P. ADUEV, E.D. ALUKER, V.I. KRASHENININ, A.G. KRECHETOV AND A.YU.
MITROFANOV: Role of Electronic Excitations in Explosive De-composition of Solids, Journal of
Applied Physics, 89(7), 4156 4166, 2000
[7] F.P. BOWDEN, A.D. YOFFE: Fast reaction in Solids, Butterworths Scientific Publications,
London, 1958
[8] Energetic Materials (H.D.Fair, R.F.Walker Eds), 1 , Plenum Press, New York, 1977
[9] Y.B. Zeldovich, N.N. Semenov, Journal of Experimental and Theoretical Physics, 10, 1116,
1940 (in Russian)
[10] J.R. MANNING: Diffusion kinetics for atoms in crystals, Princeton, Toronto, 1968
17
SYNTHESIS AND REACTIVITY

OF (CF3)2PN3, (CF3)2AsN3, CF3AsCIN3 AND CF3As(N3)2 AZIDES
H. G. ANG
National University of Singapore Faculty of Science
HEDM Research Laboratories 3 Science Drive 3 Singapore 117543
Abstract
The trifluoromethyl-phosphinous and -arsinous azides , namely (CF3)2PN3 , (CF3)2AsN3 ,
CF3As(N3)2 , and CF3AsCl(N3) , have been synthesized readily and safely with sodium
azide reagent. Their distinct reactivity can be attributed to the presence of
electronegative trifluoromethyl substituents. The reactions of the CF3-arsinous azides
with reagents such as hydrogen chloride and (CF3)2NO radical will also be
discussed..The Staudinger reactions of (CF3)2MN3 (M = P and As) with tertiary
phosphines yield phosphazene and the arsenic analogue, which display unusual
structures and bonding modes. Chain propagation behaviour of the diphosphazene
ligand has been shown to be possible.
Keyword:
1.
trifluoromethyl-phosphinous azide, trifluoromethyl-arsinous azides

synthesis, reactivity
SYNTHESIS & STRUCTURE: AZIDO-P(III) AND -AS(III)

COMPOUNDS
The reaction of iodobis(trifluoromethyl)phosphine with excess sodium azide at room

temperature overnight afforded bis(trifluoromethyl)phosphinous azide, (CF3)2PN3 as
a colorless gas at room temperature in 41% yield, according to the equation [1]:
(CF3)2PI + NaN3
(CF3)2PN3 + NaI
(CF3)2AsN3 , on the other hand, was obtained in quantitative yield from the reaction of
excess sodium azide with bis(trifluoromethyl)chloroarsine.[2]
(CF3)2AsCl + NaN3
(CF3)2AsN3 + NaCl
Similar method was used to synthesise trifluoromethylchloroarsinous azide and

trifluoromethylarsinous diazide from stepwise reactions of trifluoromethyldichloroarsine
with stechiometric amount of sodium azide, as shown below:
CF3AsCl2 + NaN3
CF3AsClN3 + NaCl
CF3AsClN3 + NaN3
CF3As(N3)2 + NaCl
18
CF3AsClN3 was obtained in 8% yield as white solid at -60 C fractionation trap (passed
-30 C). However, the arsinous diazide, CF3As(N3)2, was obtained in quantitative yield. It
appeared as a colorless crystalline solid at -45C fractionation trap (passed 0C).
The gas phase structures of (CF3)2AsN3 and CF3As(N3)2 are illustrated in Figure 1
below.[3,4]
Fig 1.
Gas phase Structure of CF3As(N3)2 (L-2) & (CF3)2AsN3 (L-2a)
All bond lengths in CF3As(N3)2 are equal to those in (CF3)2AsN3, except for the AsN
distance, which is slightly longer in the diazide (1.862 ) than in the azide (1.849 ). All
three bond angles around arsenic are larger in the diazide (CAsN = 102.7 and NAsN =
106.3 ) than in the azide (CAsN = 99.3 and CAsC = 97.8 ). This increase in the sum of the
arsenic bond angles As from 296 in (CF3)2AsN3 to 312 in CF3As-(N3)2 is most likely due
to a change in hybridization of the central arsenic atom and not to steric repulsion between
the substituents.
2.
STABILITY OF AZIDO-PHOSPHINES & ARSINES
The phosphinous azide (CF3)2PN3 is very sensitive to air, moisture, heat and light.
When placed at room temperature for 4 days in an evacuated ampoule, slow decomposition
took place to produce dinitrogen gas and a white solid which was identified as a oligomeric
compound containing (CF3)2PN units.[5] The reaction can thus be represented as follows:
n(CF3)2PN3
r.t.
[(CF3)2PN]n + nN2
The oligomer has a melting point of 53-59 oC, much lower than the waxy material
reported by Tesi G. et al.[6] It is insoluble in common organic solvents such as
dichloromethane, chloroform, acetone, methanol and benzene, and is only slightly soluble in
hot 1,1, 2-trifluoro-2,2,1-trichloroethane.
The phosphinous azide (CF3)2PN3 was reported to be explosive.[6] In our hands, we
found that the phosphinous azide is stable if stored in an evacuated glass ampoule, in the
absence of air.
We have prepared and handled the phosphinous azide in an evacuated system on
numerous occasions without any incident of detonation. However, in a controlled experiment
when air was introduced into a 50ml glass ampoule containing 30mg of the phosphinous
azide, (CF3)2PN3 , an explosion occurred after 30 seconds.[5]
19
Thermal redistribution of bis(trifluoromethyl)arsinous azide has been observed at about

80 C, resulting in the isolation of (CF3)3As and CF3As(N3)2 , as shown by the equation:
2(CF3)2AsN3
CF3As(N3)2 + (CF3)3As
Thus, it is clear that for the substituted-arsenic azides, the mode of decomposition is
strongly dependent on the nature of substituents. Redistribution of substituents is observed
for substituted-arsenic azides containing electronegative groups such as CF3, Cl or Br.
Special precaution has to be taken when handling the diazide as it was found to be
thermally sensitive and explodes when the diazide is in contact with heat.
3.
PHOTOCHEMICAL REACTIONS OF (CF3)2ASN3
Photolysis of (CF3)2AsN3 gives CF3As(N3)2, (CF3)3As, trifluoromethylazide, an

incondensable gas presumably molecular nitrogen and a white polymeric solid which is
insoluble in common organic solvents. The white polymeric solid was found to contain CF3
groups from its infrared spectrum.
The reaction could initially involve redistribution of (CF3)2AsN3 , similar to that
observed in its thermolysis reactions.
h
(CF3)2AsN3
CF3As(N3)2 + (CF3)3As + CF3N3

+ CF2=NF + (CF3AsN)n + N2
On the other hand, photolysis of substituted-arsenic azides containing electron-donating

alky substituents will result in the formation of nitrogen-free diarsines. [2]
4.
REACTIONS OF (CF3)NAS(N3)3-N (WHERE N =1, 2) WITH

ANHYDROUS HCL
Reaction of CF3-arsinous azides with anhydrous hydrogen chloride result in the

cleavage of As-N bonds to form the corresponding arsenic chloride and hydrogen azide
quantitatively, as shown below:
(CF3)2AsN3 + HCl
(CF3)2AsCl + HN3
CF3As(N3)2 + HCl
CF3As(N3)Cl
CF3As(N3)Cl + HCl
CF3AsCl2 + HN3
HN3
The above reactions with anhydrous hydrogen chloride in high yields provide an easy
route to HN3.[2]
20
5.
REACTIONS OF (CF3)2NO WITH (CF3)nAs(N3)3-n (n =1, 2)
Our study has shown that the radical reaction of bis(trifluoromethyl)nitroxyl and
bis(trifluoromethyl)arsinous azide in a 2:1 molar ratio affords the hitherto unreported
pentavalent arsenic azide, [(CF3)2NO]2As(CF3)2N3, in about 10% yield, together with other
products isolated by trap-to-trap fractionation, as shown by the equation :
(CF3)2AsN3 + 2(CF3)2NO
[(CF3)2NO]2(CF3)2AsN3 + (CF3)2NOAs(CF3)2
+ (CF3)2NON(CF3)2 + (CF3)2NONO
+ (CF3)2NNO + CF3N=CF2 + N2
On the other hand, the reaction of (CF3)2NO radical with trifluoromethylarsinous diazide
in 2:1 molar ratio afforded two three azides, namely [(CF3)2NO]2(CF3)2AsN3,
(CF3)2NOAs(N3)2 and (CF3)2NOAs(CF3)N3 .The overall reactions can be summarised by the
following equation:
CF3As(N3)2 + 2(CF3)2NO
[(CF3)2NO]2(CF3)2AsN3 + (CF3)2NOAs(N3)2
+ (CF3)2NOAs(CF3)N3 + (CF3)2NON(CF3)2
+ (CF3)2NOCF3 + (CF3)2NONO
+ (CF3)2NNO + CF3N=CF2 + N2
The products can be accounted for by invoking radical mechanisms [2].
6.
CLEAVAGE REACTIONS WITH ANHYDROUS HYDROGEN

CHLORIDE
The reactions of bis(trifluoromethyl)nitroxyarsenic azides with anhydrous hydrogen

chloride have also enabled us to confirm the identities of the compounds.
The reactions of (CF3)2NOAs(N3)2 and (CF3)2NOAs(CF3)N3 with excess anhydrous
hydrogen chloride result in the complete cleavage of both the As-O and As-N3 bonds, as
shown by the following equations:
(CF3)2NOAs(N3)2 + 3HCl
(CF3)2NOH + AsCl3 + H2 + 3N2
(CF3)2NOAs(CF3)N3 + 2HCl
(CF3)2NOH + CF3AsCl2 + 1/2N2 + 1/2H2
The above reactions illustrate the ease of cleavage of the As-N and the As-O bonds to
afford the hydroxylamine and the decomposition products of HN3.
21
7.
PHOSPHAZENES & ARSAZENES: UNUSUAL STRUCTURE &

BONDING MODES
7.1
(CF3)2P=N=PPh3 & Cluster Derivatives
The diphosphazene (CF3)2P=N=PPh3 (L-1) was obtained in 91% yield by the reaction of
(CF3)2PN3 with triphenylphosphine in dichloromethane at room temperature for 4 days. [7]
(CF3)2PN3 + PPh3
(CF3)2P=N=PPh3 (L-1)
The ligand L-1 was found to react with Fe2(CO)9 in hexane at room temperature to
afford C-1 and C-2 in 60% and 28% yield respectively.
(CF3)2P=N=PPh3 + [Fe2(CO)9]
Fe(CO)4{(CF3)2P=N=PPh3 (C-1)
+ Fe(CO)3{(CF3)2P=N=PPh3}2 (C-2)
The molecular structure of L-1, C-1 and C-2 are give below in Figures 2 to 4. In the
carbonyl complexes C-1 and C-2, the Fe atom assumes a nearly trigonal bipyramidal
structure with the ligand moiety taking up axial positions. The molecular structure of L-1 is
nearly the same in both the free and complexed forms.
Fig 2.
Molecular structure of (CF3)2P=N=PPh3 (L-1)
22
Fig 3.
Molecular structure of [Fe(CO)4.(CF3)2P=N=PPh3] (C-1)
Fig 4.
Molecular Structure of [Fe(CO)3{(CF3)2P=N=PPh3}2] (C-2)
The most striking and unusual structural feature is that the P=N and N=P bonds found in
L-1, C-1 and C-2 are equidistant. Furthermore, all these lie within the range of values of
1.55 1.62 , expected for the P=N double bond [8, 9], as summarized in Table 1. This
suggests that the unsymmetrical PNP segment is highly delocalised presumably owing to the
effect of the strongly electron-withdrawing CF3 moiety.
23
Table 1. Bond lengths of P=N and N=P bonds of ligand L-1 and complexes
P(III) = N
N= P(V)
L-1 : (CF3)2P=N=PPh3
1.612
1.576
C-1 :[Fe(CO)4{(CF3)2P=N=PPh3}]
1.568
1.560
C-2 : [Fe(CO)3{(CF3)2P=N=PPh3}2]
1.576
1.565
Compound
7.2
(CF3)2As=N=PPh3 with Unusual As=N=P Structure
The arsenic ligand L-2, analogous to the phosphazene L-1, was synthesized by the
Staudinger reaction from the CF3-arsinous azide (CF3)2AsN3 , according to the equation: [10]
(CF3)2AsN3
+ PPh3
(CF3)2As=N=PPh3 (L-2)
L-2 forms cluster derivatives with Os3(CO)11(CH3CN) and Os3(-H)2(CO)10 where the
terminal tertiary arsenic atom is linked to the Os atom, as shown below:
(CF3)2As=N=PPh3 + Os3(CO)11(CH3CN)
(CF3)2As=N=PPh3 + Os3(-H)2(CO)10
Os3(CO)11(CF3)2As=N=PPh3 (C-3)
Os3(-H)2(CO)9(CF3)2=As=N=PPh3 (C-4)
The molecular structures of the arsenic ligand L-2 and its cluster derivatives C-3 and
C-4 are shown in Figures 5 7.
Thus, it is clear that the neutral molecule (CF3)2As-N=PPh3 upon coordination assumes a
highly delocalised electronic structure over the unsymmetrical As=N=P segment. The
enhanced delocalization upon complexation over the As=N=P segment represents the first
examples of a delocalised non-linear structure involving an As(IV) atom. Enhanced
delocalisation upon co-ordination of (CF3)2As-N=PPh3, is reflected not only in the
shortening of the As-N bond length but also in the increase in As-N-P angles from 126.6 in
(CF3)2As-N=PPh3 to 135.4 and 133.0 in the cluster derivatives C-3 and (C-4) respectively.
24
Fig 5.
The molecular structure of (CF3)2As=N=PPh3 (L-2)
Fig 6.
The molecular structure of [Os3(CO)11(CF3)2As=N=PPh3] (C-3).
Fig 7.
The molecular structure of [Os3(-H)2(CO)9(CF3)2-As=N=PPh3] (C-4)
25
A comparison of As-N bond length of some selected compounds are tabulated in Table
2. The most significant feature of these structures is that the As-N bond distance in
(CF3)2As=N=PPh3 [1.783 ], Os3(CO)11{(CF3)2As=N=PPh3} [1.716 ] and [Os3(H)2(CO)9{(CF3)2=As=N=PPh3} [1.739 ] is significantly shorter than expected for a As-N
single bond [1.87 ] found in As4(NCH3)6 [11].
Table 2. Comparison of As-N bond lengths in selected compounds
Compounds
As-N ()
As=N ()
L-2 : (CF3)2As=N=PPh3
1.783(10)
C-3 :[Os3(CO)11{(CF3)2As=N=PPh3}]
1.716(13)
C-4 : [Os3(-H)2(CO)9{(CF3)2=As=N=PPh3}]
1.739(16)
As4(NCH3)6
7.3
1.87 mean
(CF3)2P=N=P(CF3)2=N=PPh3 and Cluster Derivative
The triphosphazene (L-3) was obtained in 78% yield by the reaction of (CF3)2PN3 with
the diphosphazene (CF3)2P=N=PPh3 in dichloromethane at room temperature, according to
the following equations:[12]
(CF3)2PN3 + PPh3
Ph3PNP(CF3)2 + N2+ + N2
(CF3)2PN3
CF3
Ph3PNPNP(CF3)2 (L-3) + N2
CF3
These reactions demonstrate the feasibility of chain propagation through the terminal
reactive phosphorus(IV) atom.
26
Fig 8.
Ortep Diagram of (CF3)2P=N=P(CF3)2=N=PPh3 (L-3)
Similar to the diphosphazene, the triphosphazene (L-3) has been found to react with
[Os3(CO)11(MeCN)] in dichloromethane at room temperature for to afford (C-5) in 71%
yield. [13]
(CF3)2P=N=P(CF3)2=N=PPh3 + Os3(CO)11(CH3CN)
Os3(CO)11P(CF3)2N=P(CF3)2N=PPh3 (C-5)
Fig 9.
Molecular Structure of Os3(CO)11[(CF3)2P=N=P(CF3)2=N=PPh3] (C-5)
27
The molecular structure of C-5, as shown in Fig.9, comprises an approximate isosceles

triangle of osmium atoms, each of which exhibits distorted octahedral coordination
geometry. The bulky triphosphazene ligand C-5 bonds to Os(2), taking up an equatorial site.
The molecular structure of the free ligand L-3 as displayed in Fig.8 is essentially similar to
that of the co-ordinated form whose bond lengths are summarized in Fig.10. The most
significant feature is that the four phosphorus-nitrogen bond lengths within C-4 (1.523-1.584
) as well as in C-5 (1.518-1.52 ) are not only close to being equidistant but also typical of
the P=N double bond (1.55-1.62 ). In addition, the two P=N=P segments within the
triphosphazene moieties in L-3 and C-5 are neatly linear, the P-N-P angles being 153.15
and 148.1 in L-3 and 162.4 and 154.1 in C-5. However, the PIV=N=PVI=N=PV ligand is
bent at the central PVI atom, with the N=P=N angle being 125.1 in L-3 and 122.7 in C-5.
1.572
1.551
P(III)==N==P(VI)==N==P(V)
1.518
1.565
Fig 10. Bond lengths of P=N bonds of cluster derivative C-5

Our findings have established that the highly delocalized electronic structures displayed
by the diphosphazene as well as its complexes can be extended to the triphosphazene L-3
and its derivative C-5 which also represents the first examples of a P=N=P=N=P non-linear
structure containing terminal quadrivalent phosphorous atom.
The most unusual feature in this class of compounds that we have synthesized is that the
terminal phosphorus aatoms of L-1 and L-3 and the arsenic atom of L-2 are quadrivalent,
and that they all contain a pair of electrons that are capable of forming co-ordinate bonds
with metal carbonyl compounds. The P=N and As=N structures and bonding modes of the
ligands are retained upon co-ordination to transition metals.
Acknowledgement
It is my pleasure to acknowledge the contributions of many colleagues whose names are
given in the references. I wish also to thank the National University of Singapore for the
generous financial support for all the work mentioned in this paper.
28
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
H.G. ANG, W.L. KWIK, Y.M.CAI AND A. RHEINGOLD: J.Chem. Soc., Chem.Commun., 1990,
1580
H.G.ANG, Y.M.CAI AND W.L.KWIK: J Organometal. Chem., 1992, 1-5.
Y.W. LEE, Ph.D. Thesis, National University of Singapore, 1994
H.G. ANG, W L KWIK, Y.W. LEE AND H.OBERHAMMER: Inorgan.Chem.,1994, 33
H.G. ANG, W.L. KWIK, Y.W. LEE, S.LIEDE AND H.OBERHAMMER: J.Mol.Struct., 1992, 268,
389.
Y.M. CAI: Ph.D. Thesis, National University of Singapore, 1994
G. TESI, C.P. HABER, C.M. DOUGLAS: Proc.R.Soc., London 1960, 219
H G ANG, Y M CAI, L L KOH AND W L KWIK: J Chem Soc., Chem. Commun. 1991, 850.
A. F. CAMERON, N. S. HAIR AND D. G. NORRIS: Acta. Crystallogr., Sect. B, 1974, 30, 221.
G.W. ADAMSON AND J.C. BART, J. Chem. Soc. (A), 1970,1452.
H G ANG, W L KWIK, Y W LEE AND A L RHEINGOLD: J Chem Soc., Dalton Trans. 1993, 663667.
J WEISS AND W. EISENHUTH: Z. Anorg. Allg. Chem., 1967, 350, 9.
H.G.ANG, Y.M.CAI AND W.L.KWIK: J.Organonetal.Chem.,1993, 448, 219.
29
PRE-DETONATION PHENOMENA IN HEAVY METAL AZIDES*

Eduard D. Aluker*, Boris P. Aduev** and Alexander G. Krechetov*
* Kemerovo State University,
Krasnaya, 6, Kemerovo, 650043, Russia
** Kemerovo Branch ICSSM SB RAS,
Sovetcky, 40, Kemerovo, 650099, Russia
Abstract:
A possible model for initiation of heavy metal azides (HMAs) is described, considering
reaction centers as associations of cationic and anionic vacancies,. Premises for the
model suggested and its physical groundwork and mathematical apparatus are
presented. From the analysis of the model it has been shown that the available wealth of
experimental data is qualitatively described by the model suggested.
Keyword:
1.
Keywords: heavy metal azides, chain reaction, initiation, model,

divacancy
INTRODUCTION
The pre-explosive phenomena (i.e. the processes occurring in the sample before its
mechanical fracture resulting from explosion) were stated to date to develop by the chain
mechanism [1-4]. It has been shown that the kinetics of a chain reaction before the mechanical
fracture of a sample is described by the following equation:
n& = n n 2
(1)
where n is the concentration of free charge carriers (holes and electrons) and and are
the rate constants.
The solution of Eq. (1) is as follows.
[ (
n = e t n1 e t 1 + n01
(2)
where n0 is the initial concentration, i.e. the concentration at the finish of an initiating
pulse, n = / - stationary concentration derived from Eq. (1) when n& = 0.
The simple kinetics presented by Eq. (1) and Eq. (2) describes the developed chain
process. However, the kinetics of early stages when the concentration of free electrons and
holes is less than 1017 cm-3 is more complicated, apparently, due to the peculiarities of
initiation [5]. In addition, the Eq. (1) does not evidently include the reaction initiation
threshold. This fact contradicts with numerous experimental results [6, 7].
In the present paper, we propose the model for HMAs initiation, which supplements
with the model for development of explosive chain reaction suggested earlier [1] and
eliminates the above-stated contradictions.
*
The work is supported by ISTC (project No. 2180).
30
2.
PREMISES OF THE MODEL
It is well-known [8] that the association of contrary charged defects into neutral pairs in
crystals is energy-wise advantageous leading to association of the most of cationic and
anionic vacancies into the so-called divacancies in a number of ionic crystals under
equilibrium conditions [9].
Under the assumption that the main defects in the initial HMA are divacancies it can be
concluded that not large cross-section value (neutral center) results in the less probability of
the trapping of a hole by divacancy than the probability of disappearance of a hole by
competitive ways (the trapping by other defects, surface recombination, etc.). However, if
the initiating pulse creates excessive concentration of free electrons, their trapping by
divacancies results in changes of charge state of a center, leading to increased by two or
three orders of magnitude cross-sections of the hole trapping ( - ) [8].
This fact can lead to sharp redistribution of holes in favor of working way and,
consequently, to origin of a chain reaction.
3.
PHYSICAL GROUNDWORK OF THE MODEL
The scheme of the processes considered by the divacancy model proposed is presented
in Fig. 1. The following denotations used in Solid State Physics are applied.
+
initiation ( 0 )
a
1
+
a
+
Fig 1.
+
recombination ( 0 )
+
+
+ + + multiplication (according to model (1))
+ + + multiplication (according to model (1))
The scheme for divacancy model for initiation of the explosive chain
reaction in AgN3. Denotations are presented in the paper.
+ anionic vacancy (the vacancy of negative ion, positively charged respectively to

the crystal lattice)
cationic vacancy (the vacancy of positive ion, negatively charged respectively to
the crystal lattice)
+
divacancy (assotiation of anionic vacancy and cationic vacancy located at not
large distances limited by adjoined points of the lattice)

+
an electron trapped by anionic vacancy (the so-called F-center neutral
respectively to the crystal lattice [8])
31
a hole trapped by cationic vacancy.

When a hole is trapped by
+
the probability 1-a and
the hole recombination with an electron occurs with
is reconstructed. The probability of the trapping of a hole by
is a. This center reconstruction results in hole multiplication [1,4], i.e. to a chain reaction.
The chain of the processes considered can be presented by the following quasi-chemical
reactions.
(V
,V A+ + e VC , F
(V
,F + h |
(V
,V A+ + h VC0 ,V A+
(
(
(4)
VC ,V A+
VC0 , F VC , F + 2h + e
) (
) (V
)
,V ) + 2 h + e ,
+
A
(5,6)
(7)
where VC and V A+ are cationic and anionic vacancies, respectively, F is F-center (an
electron trapped by anionic vacancy), e and h are an electron and hole, respectively.
The reaction (7) is the trapping of a hole by divacancy (cross-section 0), which also
results in hole multiplication [1].
The Eq. (4) represents initiation; the Eq. (6) and Eq. (7) represent hole multiplication
providing for branching of the chain and the Eq. (5) represents hole disappearance, i.e. a
competitive way. Other possible competitive ways (the hole trapping by impurity or
structural defects, surface recombination, etc.) are ignored in the above scheme.
4.
MATHEMATICAL APPARATUS OF THE MODEL
The processes considered can be taken into account if the value is presented as
follows.
= V [a - N + o ( N N ) (1 a) - N g ]
(8)
where V is thermal velocity of a hole, - and 0 are cross-sections of the hole trapping
by a cationic vacancy and divacancy, respectively, N and N are the initial concentrations of
divacancies and cationic vacancies located near F-center, respectively and g is a
characteristic of competitive ways. As there are no any reliable data on competitive ways by
now we accept g = const.
Va -N is the rate of hole generation as a result of the hole trapping by cationic
vacancies (multiplication of holes on charged centers), V o ( N N ) is the same for
divacancies (multiplication of holes on neutral centers) V (1 a) - N is the hole
disappearance rate, resulting from recombination with F-centers, Vg is the same for other
competitive ways.
The following equation is given for N:
32
[ (
N& = V 0 N N ne (1 a ) N nh
(9)
where ne and nh are the concentrations of band electrons and holes, V o ( N N ) ne is

the rate of N generation resulting from the trapping of electrons by divacancies (Eq. 4),
V (1 a) N - nh is the rate of N disappearance as a result of hole recombination with F-centers
(restoration of divacancies according to Eq. 6).
An analysis of Eq. (8) and Eq. (9) is feasible under the following conditions:
ne nh = n
(10)
N = const
(11)
g = const
(12)
- >> 0
(13)
a 0.1
(14)
The Eq. (10) is applied at the high levels of injection when the concentration of band
charge carriers exceeds the concentration of trapping centers [9]. Taking into account the
facts that the concentration of defects in the AgN3 samples investigated did not exceed
1015 cm-3 and experiments on pre-explosive phenomena were carried out when the
concentration of free charge carriers was 1016 cm-3 [5] the condition represented by Eq. 10
seem to us very realistic.
The condition represented by Eq. 11 means that the initial concentration of divacancies
is not changed in the process considered and only their charge state is altered (Eqs. (4) (7)).
This condition seem to us reasonable for early stages of explosive decomposition.
The validity of conditions represented by Eq. (11) and Eq. (12) was considered above
and the validity of condition represented by Eq. (14) will be obvious from the further
arguments.
From the Eq. (8) and Eq. (9) taking into account Eqs. (10) (14) we can derive the
following equations.
N =
0N
(1 a )
(15)
N& = V n (1 a)( N N )
(16)
= V [(2a 1) N ( g o N )]
(17)
& = V (2a 1) N&
(18)
= V (a - N g )
(19)
0 = V [(2a 1) N o- ( g o N )]
(20)
where N and are stationary values derived from Eq. (8) and Eq. (9) when t .
33
5.
THE ANALYSIS OF THE MODEL

5.1
The condition for the sample stability
If there are no initiation N = 0 and = o N g according to Eq. 17. If > 0 the

appearance of even a single hole according to (1) will result in the triggering of a chain
reaction and, consequently, result in explosion of a sample. Such the spontaneous
explosions, including explosions during synthesis, are well-known [6, 7]. Thus, the condition
for the sample stability is < 0 (without initiation!) and can be presented as follows.
0N g < 0
5.2
(21)
Initiation threshold
The condition for triggering the chain reaction resulting from initiation is 0 > 0, where
0 is after the finish of an initiating pulse (at zero time). From the Eq. 20 when 0 > 0 the
Eq. (22) is derived.
(2a 1) N 0 > g 0 N
(22)
where N 0 is initial concentration of charged centers generated by an initiating pulse as a

result of trapping of electrons by divacancies (Eq. 4). Threshold energy of an initiating pulse
is determined by the value of N 0 .
As shows Eq. (21) and Eq. (22) a chain reaction can be triggered only under the
condition > 0.5, i.e. the condition represented by Eq. (14) is automatically feasible.
5.3
Over threshold initiation
The conditions for branching of a chain reaction are obvious 0 > 0 and > 0.
The value of 0 according to Eq. (20) is determined by the value of N 0 , i.e. it is
determined by the energy of an initiating pulse. As to > 0 according to Eq. (19) and Eq.
a 0 N
(15) it is feasible when
>g.
1 a
In that case the kinetics of changes in values of (t) and N(t) according to Eqs. (15)(19) is presented in Fig.2.
As shown in Fig. 2 the rate of the reaction in initial part (0) can be less or more than the
rate of developed process, depending on initiation energy ().
Thus, the simple kinetics represented by Eq. (2) and observed at n 1017 cm-3 [1-4]
apparently corresponds with the stationary value of ( = ). For initial stages of the
process considered (when ) the kinetics is not described by Eq. (2) and as shown in
Fig.2 it depends on the energy of an initiating pulse (by values of 0 and N 0 ). The influence
of the energy of an initiating pulse on the initial stages of pre-explosive kinetics is observed
34
N
N
Fig 2.
Kinetics of changes in concentrations of charged centers (N) (Eq. 15 and

Eq. 16) and the rate constants of the branching of a chain () under over
threshold initiation (Eqs. 17 19).
during the experiments [5]. It is typical that the value n = / (Eq. (1) and Eq. (2)) in those
experiments does not depend on the energy of an initiating pulse. This result also agrees with
the model proposed as the developed process ( ) does not depend on the values of 0
and N 0 (Eq. (19) and Eq. (15).
Thus, the model suggested qualitatively agrees with the experimental data obtained.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
B.P. ADUEV, E.D. ALUKER, G.M. BELOKUROV, YU.A. ZAKHAROV, A.G. KRECHETOV:
Explosive Decomposition of Heavy Metal Azides, Journal of Experimental and Theoretical
Physics, 89(5), 906, 1999
B.P. ADUEV, E.D. ALUKER, G.M. BELOKUROV, YU.A. ZAKHAROV, A.G. KRECHETOV:
Predvzryvnye javlenija v azidakh tjazhelykh metallov, CEI Khimmash, Moskva, 2002 (in
Russian)
M.M. KUKLJA, B.P. ADUEV, E.D. ALUKER, V.I. KRASHENININ, A.G. KRECHETOV AND A.YU.
MITROFANOV: Role of Electronic Excitations in Explosive De-composition of Solids, Journal of
Applied Physics, 89(7), 4156 4166, 2000
B.P. ADUEV, E.D. ALUKER, G.M. BELOKUROV, A.N. DROBCHIK, YU.A. ZAKHAROV, A.G.
KRECHETOV, AND A.YU. MITROFANOV: Preexplosion Phenomena in Heavy Metal Azides,
Combustion, Explosion, and Shock Waves, 36(5), 622-632, 2000
B.P. ADUEV, E.D. ALUKER, G.M. BELOKUROV, A.G. KRECHETOV, AND A.YU. MITROFANOV:
Kinetics of the Early Stage of Preexplosion Conduction in Silver Azide, Combustion,
Explosion, and Shock Waves, 38(3), 378-380, 2002
F.P. BOWDEN, A.D. YOFFE: Fast reaction in Solids, Butterworths Scientific Publications,
London, 1958
Energetic Materials (H.D.Fair, R.F.Walker Eds), 1 , Plenum Press, New York, 1977
A. M. STONEHAM,: Theory of Defects in Solids, Clarendon, Oxford, 1975
F. J. BLATT,: Physics of Electronic Conduction in Solids, McGraw-Hill Company, 1968
35
SIMPLE METHOD OF SYNTHESIS AND CHARACTERIZATION

OF SOME NITROCYANAMIDE SALTS
Alexander M. Astachov*, Ludmila A. Kruglyakova*,
Irina V. Gelemurzina*, Alexander D. Vasiliev**, and Rudolf S. Stepanov*
* Siberian State Technological University,
Prosp. Mira 82, 660049 Krasnoyarsk, Russia
** Institute of Physics RAS (Sib. branch),
Akademgorodok, 660036 Krasnoyarsk, Russia
Abstract:
A simple method has been offered to obtain alkali salts of nitrocyanamide (NC
NHNO2) by means of interaction between potassium (or sodium) hydroxide and Smethyl-N-nitroisothiourea. Ammonium salt of nitrocyanamide can be obtained with
quantitative yield through the exchange reaction between potassium salt of
nitrocyanamide and ammonium sulphate. UV-, FTIR-spectroscopy, X-ray powder
diffraction and thermal characterization are all used to characterize the salts obtained.
Keyword:
1.
Keywords: Nitrocyanamide salts, S-methyl-N-nitroisothiourea
INTRODUCTION
Nitrocyanamide (NCA) salts were obtained for the first time in fifties of last century [1-3].
Nevertheless, known at the moment methods for the obtaining of the salts one cannot call
convenient. The methods are based on the reaction of alkaline splitting of N-nitro- or Nnitroso-N-alkyl-N-nitroguanidines, which synthesize a nitration/nitrosation of corresponding
N-alkyl-N-nitroguanidines [1-5]. In ones turn, the latter one can obtain from nitroguanidine
(NQ) [4, 5] or S-methyl-N-nitroisothiourea (MNITU) [6]. In latter case the yield is a little more,
and the synthesis is simpler and demands less time. In all cases, the biggest yield is observed
if the alkyl is the methyl:
NNO2
NNO2
H2N
NH2
NQ
85%
NH
3
NNO2
CH3S
CH N
3 H
NH2
MNITU
NH 2
CH 3
SH
CH 3
CH3NH
95%
NNO2
NH2
/H
NaNO 2
CH3N
NH2
NO
91%
HNO
3 / Ac O
2
NNO2
CH3N
NH2
NO2
KOH
CH N
2
H
KO
CH
O2
HN
3N
90%
NNO2
+
C N K
76%
52%
The synthesis of NCA salts through the N-methyl-N,N-dinitroguanidine provides an

yield less than 37% over the MNITU or up to 33% over the NQ, mainly because of low yield
on the stage of nitration. The synthesis through the N-methyl-N-nitroso-N-nitroguanidine
allows to obtain salts of NCA with good yield (7078%). However, the nitrosoderivative
itself is a strongest cancerogenic compound and any work with it is very dangerous [7, 8]. So,
defects of known methods of NCA salts synthesis restrict the possibility to study and use
them in practice.
36
During investigation of MNITU alkaline hydrolysis, we unexpectedly detected that the

NCA salts are generated instead of expected nitrourea salts. The study of the reaction
allowed us to work out simple laboratory method of alkali NCA salts synthesis directly from
MNITU avoiding intermediate steps.
[N CNNO2]
NNO2
CH3S
80-90%
OH
NH2
CH3SH
NNO2
O
NH2
The synthesis of NCA by way of cyanamide nitration has been recently suggested [9].
The direct nitration, doubtless, is more preferable in industrial production. But method of [9]
is less useful in a laboratory practice, than proposed in this paper one.
2.
SYNTHESIS
General procedure of alkali NCA salts synthesis comes to refluxing of water-alcohol

solution of MNITU and potassium (or sodium) hydroxide. After the reaction termination, the
alkali surplus was neutralized with strong muriatic acid and a solvent was evaporated. The
alkali salt of NCA was extracted from dry residue with boiling ethanol or acetone. One can
quantitatively precipitated this salt from hot ethanol (acetone) solution with diethyl ether.
The ethanol was used for recrystallization.
Unfortunately, the absorption maximum of MNITU is shifted in the UV-area from
285 nm to 260-265 nm in an alkaline solution. The absorption maximum of nitrocyanamide
anion is observed in the same area; this fact does not allow using UV spectroscopy for
reaction kinetics definition. Therefore, an optimization of the method was defined via the
preparative yield of alkali NCA salts.
Next parameters were changed: reaction time from 0.5 to 18 h, temperature from room
to solvent boiling temperature, the ratio alkali/MNITU from 1 to 6 (mol/mol), the solvent
(ethanol) amount and water contents in it from 5 to 100%. On the one hand, the water
presence diminishes the reaction duration; on the other hand, the less solubility of MNITU in
water, in comparison with ethanol, demands greater amount of solvent. In addition, in the
presence of large amount of water in the system, collateral reactions take place to a
considerable extent, which diminish the yield of the desired product. Optimum process
conditions were defined, which allow to receive the potassium salt of NCA with yield of
90%. It was found, that the optimum conditions for potassium salt synthesis are not good for
sodium NCA salt synthesis. The sodium salt of NCA is obtained as monohydrate. Drying in
a cabinet drier at 80 allows to obtain anhydrous sodium salt of NCA. However it is very
hygroscopic and this fact makes difficult to characterize the anhydrous salt. We have
recorded its IR-spectrum but could not fulfil thermal analysis and record X-ray powder
pattern. The optimization of synthesis sodium NCA salt method allowed receiving maximal
yield of 80%.
IR-spectrum analysis of unsolved residue after NCA salts extraction showed, that
isocyanates and partly carbonates of corresponding metals are the collateral reaction
products. They, probably, are generated in the process of alkaline hydrolysis of NCA salts
but not of initial MNITU. A yield of isocyanates is increased with temperature and ratio
alkali/MNITU increasing. The alkali NCA salts mechanism of hydrolysis up to isocyanates
37
obviously passes through the intermediate formation of nitrourea anion, i.e. begins from
hydrolysis of the nitrile group. It is known that the nitrourea is quickly decomposed in
alkaline solutions with formation of isocyanates [10]. To avoid completely the formation of
isocyanates with help of conditions change is not succeeded. Thus, the process of MNITU
interaction with alkalies can be represented with next scheme:
k1
NNO2
CH3S
OH
NH2
k2
Slow
[N CNNO2]
Slow
k3 Fast
NNO2
H2O / OH
O
CH3SH
NH2
[NCO]
N2O
The formation of carbonates one can to connect with well-known reaction of isocyanates
hydrolysis. If the environment contents an alkali, the hydrolysis will proceed more intensive.
2NCO + 3H2O CO32 + 2NH3 + CO2
Water solutions of NCA salts are more stable in alkali absence. No changes detected in
the UV-spectrum after boiling of water solution of potassium NCA salt during 2 hours. The
intensity of peak at 265 nm has decreased only by a few percents in 4 hours.
The ammonium NCA salt was obtained with 100% yield by means of exchange reaction
between the potassium NCA salt and the ammonium sulphate.
2K+ [NCNNO2] + (NH4)2SO4 2NH4+ [NCNNO2] + K2SO4
The procedure includes the preparation of two solutions of salts (potassium NCA and
ammonium sulphate) in minimal amount of water. Only a minor part of potassium sulphate
did precipitate by the solutions mixing. In order to completely precipitate the potassium
sulphate, it needs in adding of 5-10 multiple alcohol volume. The mixture was filtrated to
remove potassium sulphate and the solution was then evaporated. The solid residue is
recrystallized from ethanol or acetone. Analogous way can be used for synthesis of NCA
salts with other organic or inorganic cations.
3.
UV AND FTIR SPECTROSCOPY
The UV absorption spectra of NCA salts in water solution were recorded on the
Shimadzu UV-1601 spectrophotometer. The UV spectrum of all NCA salts shows an
absorption at wavelength max = 265 nm and extinction coefficient = 9000100 M1cm1.
A typical absorption spectrum is shown in Figure 1.
A
Fig 1. UV spectrum of 104 M NCA

anion-water solution
(thickness of the UV cell is 1cm)
0 .8
0 .6
0 .4
0 .2
0
220
240
260
280
300
320
340
nm
The infrared spectra (4000650 cm1) were obtained on the Nicolet IMPACT-400D
FTIR spectrometer with the IR microscope SpectraTech INSPECT IR+. IR spectra were
38
recorded at room temperature and 4 cm1 resolution. Figure 2 shows the IR spectrums of
NCA salts. Table 1 summarizes the frequencies and proposed assignments. Tentative
frequencies assignments has been done by analogy with the assignments for salts of primary
nitramines [11].
0.8
Potassium Nitrocyanamide
0.6
A
b
s
0.4
0.2
0.0
0.6
A
b
s
Sodium Nitrocyanamide (anhydrous)
0.4
0.2
0.0
Sodium Nitrocyanamide (monohydrate)
0.6
A
b
s
0.4
0.2
0.0
0.6
A
b
s
Ammonium nitrocyanamide
0.4
0.2
0.0
3600
3400
3200
3000
2800
2600
2400
2200
2000
1800
1600
1400
1200
1000
800
Wavenumbers (cm-1)
Fig 1.
The FTIR spectrums of nitrocyanamide salts: K(NCA), Na(NCA),

Na(NCA)H2O and NH4(NCA).
Table 1.
Observed IR frequencies and tentative assignments of nitrocyanamide salts.
Assignment
(NNO2)
s(NO2)
(NCN)
as(NO2)
(NNO2)
(OH)
(CN)
s(NH4+)
as(NH4+)
s(OH)
as(OH)
K(NCA)
773
973
1160
1284
1427
2194
Wavenumber, cm1
Na(NCA)
Na(NCA)H2O
775
764
982
965
1167
1177
1298
1285
1448
1450
1630
2212
2194
NH4(NCA)
770
967
1163
1266
1402
2190
3053
3242
3530
3598
39
4.
X-RAY POWDER DIFFRACTION
Crystal structures of K(NCA), Na(NCA)H2O and NH4(NCA) have been characterized

using the Bruker D8 ADVANCE X-ray powder diffractometer. We could not obtain the
powder pattern of anhydrous Na(NCA) because it is strong hygroscopic. The X-ray powder
patterns of NCA salts are shown in Figure 3.
Space groups of symmetry and unit cell parameters of the crystal NCA salts (Table 2)
were determined with powder methods. Earlier, the structures of K(NCA) [12,13], Ag(NCA)
[13]
and some complex NCA salts [14-19] were solved with single-crystal X-ray methods.
Table 2. X-ray powder crystallographic data for K(NCA), Na(NCA)H2O
and NH4(NCA).
Crystal data
Space Group
a ()
b ()
c ()
()
()
()
Z
Dx-ray (gsm3)
5.
K(NCA)
P21/c
6.9216(8)
7.575(1)
8.498(1)
90
108.94(1)
90
4
1.968
Na(NCA)H2O
P21/c
3.5608(5)
10.294(1)
13.104(2)
90
95.70(1)
90
4
1.766
NH4(NCA)
P21/c
7.169(1)
7.7495(7)
8.6757(8)
90
109.53(1)
90
4
1.522
THERMAL DECOMPOSITION
The thermolysis of NCA salts was measured with thermogravimetric analysis (TG) and
differential thermal analysis (DTA), and the solid decomposition residues were characterized
by FTIR spectroscopy. TG and DTA measurements were fulfilled on a Paulik-Paulik-Erdey
Q-1500D thermal analyzer with 50 mg of sample. A heating rate of 5 Kmin1. TG and DTA
curves of the thermal decomposition of K(NCA), Na(NCA)H2O and NH4(NCA) under
conditions of linear temperature increasing are shows in Figure 4.
IR-spectrums of condensed residue of NCA salts thermal decomposition are shown in
Figure 5. After full thermal decomposition of K(NCA) the residue is the practically pure
potassium isocyanate. The fact allows to write the equation of thermal decomposition
process as:
K+ [NCNNO2] KNCO + N2O
In a case of Na(NCA)H2O, side by side with the isocyanate, the sodium carbonate
presents additionally in the residue after decomposition. The formation, probably, connected
with partial thermal decomposition or oxidation of potassium isocyanate [20]:
5NaNCO 3NaCN + Na2CO3 + CO2 + N2
The kinetics and mechanism of isothermal and none isothermal decomposition of
NH4(NCA) will be published in other communication.
40
100
K(NCA)
80
Irel
60
40
20
0
5
15
25
35
45
55
65
75
85
Two Theta (Degrees)
Na(NCA)H2O
100
80
Irel
60
40
20
0
5
15
25
35
45
55
65
75
85
Two Theta (Degrees)
100
NH4(NCA)
80
Irel
60
40
20
0
5
15
25
35
45
55
65
75
85
Two Theta (Degrees)
Fig 2.
X-ray powder patterns of NCA salts.
41
K(NCA)
100
DTA
m.p.138C
Mass, %
80
Start of mass loss: 187C
TG
60
40
20
DTA peak: 299C
0
20
60
100
140
180
220
260
300
340
Temperature, C
Na(NCA)H2O
m.p.200C
DTA
Mass, %
100
80
Start of mass loss:
64C (H2O)
195C (decomposition)
60
40
TG
DTA peak: 255C
20
0
20
60
100
140
180
220
260
300
340
Temperature, C
NH4(NCA)
Mass, %
100
DTA
m.p.95C
80
Start of mass loss: 139C
60
DTA peak: 203C and 234C
40
20
TG
0
20
60
100
140
180
220
260
300
Temperature, C
Fig 3.
TGDTA curves of NCA salts at a heating rate of 5 Kmin1.
42
A
b
s
o
r
b
a
n
c
e
0.4 K(NCA) decomposition residue at 350

0.3
0.2
0.1
0.0
A
b
s
o
r
b
a
n
c
e
A
b
s
o
r
b
a
n
c
e
Na(NCA)*H2O decomposition residue at 250C

0.2
0.1
0.0
0.5 NH4(NCA) decomposition residue at 200C
0.4
0.3
0.2
0.1
4000
3800
3600
3400
3200
3000
2800
2600
2400
2200
2000
1800
1600
1400
1200
1000
800
Wavenumbers (cm-1)
Fig 4.
6.
The FTIR spectrum of residues after thermal decomposition of NCA salts:

K(NCA) (350C), Na(NCA)H2O (250C) and NH4(NCA) (200C).
CONCLUSION
The worked out method of alkali NCA salts synthesis makes them easy to access. The
salts with other metals or organic cations should be easily obtained from the alkali NCA salts
via exchange reactions.
IR-spectroscopy study showed that the observed vibration frequencies of NCA anion are
practically identical in the investigated salts. UV-spectrums of the salts water solutions are
identical too.
Like other alkali salts of such energetic compounds as 5-nitraminotetrazole [19], 3-nitro1,2,4-triazol-5-one [20], nitrourea [21], the alkali salts of nitrocyanamide are decomposed with
the formation of isocyanates of corresponding metals.
Acknowledgment
We are grateful to the Krasnoyarsk Region Scientific Fund for support of this work.
43
REFERENCES
[1] A. F. MCKAY, W. L. OTT, G. W. TAYLOR, M. N. BUCHANAN, AND J. F. CROOKER:
Can. J. Res. 28B (1950) 683688.
[2] R. H. MEEN AND G. F. WRIGHT: J. Am. Chem. Soc. 74 (1952) 20772079.
[3] S. R. HARRIS: J. Am. Chem. Soc. 80 (1958) 23022305.
[4] A. F. MCKAY AND G. F. WRIGHT: J. Am. Chem. Soc. 69 (1947) 30283030.
[5] A. F. MCKAY: J. Am. Chem. Soc. 71 (1949) 19681970.
[6] L. FISHBEIN AND J. A. GALLAGHAN: J. Am. Chem. Soc. 76 (1954) 18771879.
[7] S. RICE, M. Y. CHENG, R. E. CRAMER, M. MANDEL, H. F. MOWER, AND K. SEFF: J. Am. Chem.
Soc. 106 (1984) 239243.
[8] C. L. GALTRESS, P. R. MORROW, S. NAG, T. L. SMALLEY, M. F. TSCHANTZ, J. S. VAUGHN, D.
N. WICHEMS, S. K. ZIGLAR, AND J. C. FISHBEIN: J. Am. Chem. Soc. 114 (1992) 14061411.
[9] A. A. ASTRATIEV AND L. L. KUSNETZOV: Russ. J. Org. Chem. 38 (2002) 13081315 (In
Russian).
[10] B. BOOPSINGH AND J. M. BRIODY: J. Chem. Soc. Perkin Trans. II. (1972) 14871488.
[11] V. A. SHLYAPOCHNIKOV: Vibration spectra of aliphatic nitro compounds, Nauka, Moscow,
1989 (In Russian).
[12] H. JENCES, B. KLEWE, AND E. TJELTA: Acta Chem. Scand. A31 (1977) 151154.
[13] L. JGER, H. D. SCHAEDLER, U. GROBE, H. KOEHLER, AND V. I. NEFEDOV: Z. Anorg. und Allg.
Chem. 624 (1998) 15581562.
[14] L. JGER, C. TRETNER, M. BIEDERMANN, AND H. HARTUNG: J. Organomet. Chem. 530 (1997)
1317.
[15] L. JGER, C. TRETNER, H. HARTUNG, AND M. BIEDERMANN: Z. Anorg. und Allg. Chem. 623
(1997) 12991305.
[16] J. KOZISEK, J. G. DIAZ, M. HVASTIJOVA, AND L. JGER: Acta Cryst. C53 (1997) P. 703705.
[17] J. KOHOUT, M. HVASTIJOVA, J. KOZISEK, J. G. DIAZ, M. VALKO, L. JGER, AND I. SVOBODA:
Inorg. Chim. Acta 287 (1999) 186192.
[18] M. HVASTIJOVA, J. KOHOUT, J. KOZISEK, L. JGER, AND J. G. DIAZ: Polyhedron 19 (2000)
10211027.
[19] T. B. BRILL, B. C. TAPPAN, AND R. W. BEAL: IV International Seminar New Trends in
Research of Energetic Materials . 2001. Pardubice. Czech Republic. 1728.
[20] Z. TONGLAI, H. RONGZU, L. YANJUN, AND L. FUPING: J. Thermal. Analysis 39 (1993) 827
847.
[21] X. HEMING, Y. JINHUA, X. XINQUAN, AND D. ANBANG: Propellants, Explosives, Pyrotechnics
14 (1989) 1830.
44
INFLUENCE OF AMMONIUM NITRATE PRILLS POROSITY

AND DIMENSIONS ON DETONATION VELOCITY
OF ANFO EXPLOSIVES
Daniel Buczkowski* and Bogdan Zygmunt**,
* Institute of Industrial Organic Chemistry,
6 Annopol St, 03-236 Warszawa, PL
** WIFAMA-PREXER,
45 Niciarniana St, 92-320 d, PL
Abstract:
Prilled/granulated ammonium nitrate is commonly used as a fertiliser and basic
ingredients of industrial explosives, especially ANFO. One of the important factor
affects on explosive properties of ANFO explosives, is prills/granules porosity. In this
paper an attempt of such steering of ammonium nitrate characteristics, that
manufactured ANFO has high detonation properties, is presented. The method of
manufacturing of porous ammonium nitrate, which has high oil absorption, has been
elaborated. Relations between porosity and granulometric distribution of ammonium
nitrate versus detonation velocity of ANFO have been examined. It has been
investigated, that detonation velocity of ANFO significantly increases, if increases
porosity and decreases dimensions of ammonium nitrate prills/granules.
Keyword:
1.
Keywords:ANFO explosives, porous ammonium nitrate, detonation

velocity
INTRODUCTION
Prilled or granulated ammonium nitrate (AN) is commonly used both as a fertiliser in

agriculture and a basic ingredient of industrial explosives, especially for manufacturing
ANFO explosives: dynamites, ammonals, slurry blasting agents, etc. Ammonium nitrate has
been used for many years as the basic component for the manufacture of explosives for the
mining industry. The history of the development of modern mining explosives is in fact
connected with the practical application of various useful forms of ammonium nitrate, i.e.
crystalline, granulated or porous products or a saturated aqueous ammonium nitrate solution.
Thanks to its low cost, simple technology and accessibility of raw materials required for
its manufacture the AN containing explosives competes very effectively with other improved
types of explosives available commercially. In the USA, more than 80% of explosives used
in the mining industry are constituted by a mixture of granulated ammonium nitrate with fuel
oil (it is known in literature as ANFO, i.e. ammonium nitrate - fuel oil). Much more than
half of the world production of explosives used in civil applications is covered by ANFO
mixtures.
The main advantages of the granulated explosives include: low manufacturing costs,
readily available raw materials, simple manufacturing technology, safety of production, and
during transportation and storage, possibility of mechanising the blast hole charging in open
and underground mines. Laboratory and field tests performed in the mining industry have
45
shown that the replacement of the agriculture nonporous AN with a porous material
improves considerably the efficiency of action of the ANFO; it eliminates oil oozing and
reduces the emission of toxic gases to the atmosphere. On the other hand, all positive
features of the granulated form of the ANFO explosives are retained [1].
Typical ANFO consists of 94,5 % of AN and 5,5 % of fuel oil. The characteristic of fuel
oil are strictly determined, while the physical and mechanical features of AN may vary in a
wide range, depending on technology of production and content of additives. The decreasing
of prilled AN ability to detonation by addition of different inorganic salts are described by
the authors [2]. It has been stated, that charges of AN of small porosity do not detonate, while
after thermal treatment in order to increase porosity of prills, the same product is able to
stable detonation. It is well known that crucial features influencing on detonation parameters
of explosives are: density of the charge, dimensions of particles of explosives and content of
mixture ingredients. The ANFO mixtures detonation is regarded as the example of a nonideal detonation regime and one can expect that the change of basic physical parameters of
AN prills/granules may influence in high degree on detonation parameters the explosive [3].
The goal of the work was to determine what is the influence of porosity of granules (in the
wide range from 0 to 15 % of volume) and dimension of granules on the velocity of
detonation of ANFO explosives.
2.
A METHOD OF MAKING POROUS AMMONIUM NITRATE
The most suitable for making porous prills is AN with no inorganic additives. But such
type of AN is produced very rarely. Almost all AN contain inorganic additives which
improve their quality and make a process of manufacturing porous AN more difficult.
Relatively a small influence on this process has ammonium sulfate. Another additives like
calcium or magnesium nitrate cause, that it is very hard to produce AN about high porosity.
Used in this work AN contained about 1 % of ammonium sulfate.
The process of making porous prills was based on a polymorphic transition of AN,
which runs at + 32oC.
32C
3
IV (1,72 g/cm )
32C
3
III (1,66 g/cm )
IV (1,72 g/cm3)
During the transition IV III IV density of AN crystals significantly changes, which

causes creating porous structure and cracking of prills. Porous product has bigger oil
absorption and ability to detonation measured by tube test [4].
A technology relied on heating AN above 32 C and cooling; an additive which
increased porosity was also used. Due to different number of thermal cycles heating/cooling
and amount of mentioned additive, AN about broad scope of physical characteristics were
produced. Examinations of physical properties of produced AN were done; next we made
ANFO and determined their detonation velocity.
46
3.
INFLUENCE OF PRILLS POROSITY ON DENSITY

OF AMMONIUM NITRATE
It is common in military technique that charges of brisant explosives should have the
highest densities. According to increasing density, increase velocity and pressure of
detonation and influence on surroundings. But if density of the charge increase, decrease
sensitivity to initiation stimuli and the charge is more difficult to initiate to detonation. This
problem is especially important in the case of industrial explosives, which are in general low
sensitive. Industrial explosives are mainly multiphase mixtures and are difficult to initiate [5].
If density of such explosives increase, their detonation abilities reduce and finally they may
loss this property.
A large influence on density of materials has their porosity. The rise of porosity is
accompanied by growing of voids in materials, which caused a drop of density. In a diagram
below (Fig.1) a relationship between bulk density of prilled AN and porosity of prills is
showed. AN has been made porous in a manner presented in par. 2. An oil absorption of has
been approved as a measure of AN porosity AN (volume of fuel oil absorbed by 100 g of
AN [cm3/100g]).
1
Density [g/cm3]
0,95
0,9
0,85
0,8
0,75
0,7
0,65
0,6
0,55
0
10
12
14
16
Oil absorption [cm3/100g]
Fig 1.
Density of prilled AN versus oil absorption
As we may see from the diagram according to increasing porosity, AN density drops.
AN which hasnt been made porous has no oil absorption (0,0 cm3/100g) and their density
was 0,95 g/cm3. A few times heated and cooled AN, with the additive which rises porosity,
has oil absorption about 15 cm3/100g and bulk density 0,59 g/cm3.
4.
INFLUENCE OF PRILLS POROSITY ON MECHANICAL

ENDURANCE
Mechanical endurance has an important influence on quality of prilled AN and ANFO.

Prills about low mechanical resistance crush and split during storage, transport etc. and when
ANFO are manufactured. Splitted and crushed prills easy cake; during mixing with oil they
47
adhere to walls of apparatus, mixing equipment etc., which makes technological process
more difficult.
Endurance has a strong relation with a structure of prills. Prills about small porosity have
compact structure and are more resistant than prills about high porosity, which characterised
looser constitution.
Below in a diagram (Fig. 2), a relationship between prills resistant to press and porosity
is presented (prills diameter 1,5 - 2,0 mm). The oil absorption of AN has been approved as a
measure of AN porosity.
Resistance to press [G]
650
600
550
500
450
400
350
300
250
200
0
10
12
14
16
Fig 2.
Dependence among AN prills endurance to press and oil absorption
Conclusions from the diagram is that according to increasing of porosity, prills resistant
to press falls. Non porous AN prills about oil absorption 0,0 cm3/100g have press resistant
over 500 G, but high porous prills, about oil absorption above 10,0 cm3/100g have the
resistant two times lower.
5.
INFLUENCE OF PRILLS POROSITY ON DETONATION

VELOCITY OF ANFO EXPLOSIVES
Detonation velocity is one of the most important and very often determined properties of
explosives. In this work ANFO were made by mechanical stirring of AN prills (about
various physical characteristics) and fuel oil. The proportion was 94,5 % AN and 5,5 % fuel
oil - oxygen balance equals zero. The explosives were poured into steel tubes about internal
diameter 36 mm and wall thickness 3 mm. Detonations of the charges were initiated by
boosters (so called HT-14; mass 14 g, RDX 90 %, TNT 10%).
A relation among detonation velocity of ANFO and porosity of AN prills is presented in
a Fig. 3. Like former, the oil absorption of AN has been approved as a measure of AN
porosity.
48
2,8
2,6
2,4
2,2
2
1,6
1,8
Detation velocity [km/s]
10
12
14
16
Fig 3.
Detonation velocities of ANFO versus oil absorption of AN prills
As one may see on the diagram, detonation velocity of ANFO increases, if prills
porosity grows. Charges of ANFO, made of AN about oil absorption below 2,5 cm3/100g,
didnt detonate in test condition. Minimal oil absorption of AN prills, that detonation of full
charges were observed, was about 2,5-3,0 cm3/100g. Measured detonation velocity was than
1,6-1,7 km/s. If oil absorptions were about 12-15 cm3/100g, detonation velocities were about
2,7 km/s. The tendency to further growth wasnt observed. If we see on a Fig. 1. we may
notice, that according to porosity rise from 0,0 cm3/100g to 15 cm3/100g, density falls at
about 35 %. Although this drop, detonation velocity increases more than 70 % (range of oil
absorption 2,5-15,0 cm3/100g). This relationship between porosity and detonation velocity is
another, than dependence observed for typical brisant explosives. For that type of explosives,
if density falls, also falls detonation velocity.
Points on Fig. 3. are scattered, because of various physical characteristics of AN prills
(density, mechanical endurance, content of water, amount of dust etc.), uncertainty of
determining oil absorption and pour physical stability of ANFO made of AN about low
porosity.
6.
INFLUENCE OF PRILLS DIMENSIONS ON DETONATION

VELOCITY OF ANFO EXPLOSIVES
Dimensions of crystals, grains, granules, prills etc. have large influence on explosive
properties of explosives, especially mixtures of oxidiser-fuel [6]. For this type of materials,
according to diminish dimensions of materials pieces, grow sensitivities to external stimuli
and increase detonations parameters. Taking into consideration typical brisant explosives,
their detonation velocity depends in a smaller degree from granulometric composition.
In this paragraph a prills about small dimensions were examined. They were so called
underprills, which are, in a factories, sieved from trade product and move back to a former
part of technological line.
49
Underprills were sieved to different fractions and next used for making ANFO, about
oxygen balance equals zero. Detonation velocities of such materials were next determined, in
a way described in par. 5.
There are presented in Table 1 results of determining detonation velocities of ANFO
made of various fractions of AN prills. Oil absorption of this prills were near zero, so
physical stability of ANFO, especially made of coarser prills, was poor.
Table 1. Detonation velocity of ANFO explosives versus dimensions of AN prills
Prills
dimension
[mm]
1,00 - 1,20
0,63 - 1,00
0,50 - 0,63
0,20 - 0,50
Density of ANFO Detonation velocity

[g/cm3]
[km/s]
1,02
1,01
1,02
0,98
Detonation decay
2,33
2,50
2,96
Results collected in the table indicate that according to diminishing dimensions of AN

prills, detonation velocities of ANFO increase. ANFO made of prills about dimensions
1,00-1,20 mm didnt detonate in test condition. Material manufactured from the smallest
prills (0,20-0,50 mm) fully detonated with velocity about 3,0 km/s.
Next we wanted to confirm the influence of AN size to detonation velocities of ANFO.
Materials were produced from different fractions of milled AN. ANFO were poured into
tubes and in this state tested (without tamping). Results of experiments are showed
in Table 2.
Table 2. Detonation velocity of ANFO explosives versus dimensions of AN grains
Grains
dimension
[mm]
0,50 - 1,00
0,20 - 0,50
below 0,20
Density of ANFO
[g/cm3]
0,92
0,86
1,04
Detonation
velocity
[km/s]
2,58
3,44
4,38
As we can see from the table size of AN grains has very important influence on
detonation velocity of ANFO. ANFO made of AN grains about dimensions 0,5-1,0 mm,
detonated with velocity about 2,6 km/s; made of grains about dimensions below 0,2 mm had
detonation velocity about 4,4 km/s.
The influence of fineness on detonation velocity is so large, that predominance over the
influence of density of the charge.. Although changes of the density - fall and growth detonation velocity increases according to decreasing of AN grains.
If we compare detonations velocities of ANFO made of prilled and milled AN, we may
conclude that ANFO made of milled AN detonated with a greater velocity. For dimensions
0,2- 0,5 mm explosive made of prills has detonation velocity 2,96 km/s made of milled AN
3,44 km/s
50
7.
CONCLUSIONS
The porosity of AN prills influence in high degree on detonation parameters of ANFO

explosives. The increasing of prills porosity results a remarkable increase of detonation
velocity of ANFO mixtures in condition of conducted experiments. Charges of ANFO
prepared from AN granules of a porosity below 2,5 cm3/100g, did not detonate at all. The
threshold of AN porosity from which one could observe a stable detonation of ANFO was
2,5-3,0 cm3/100g and measured detonation velocity was 1,6-1,7 km/s. Along with increasing
of AN porosity, the detonation velocity of ANFO raised up to 2,7 km/s for the level of AN
porosity 12-15 cm3/100g. The described effect took place in spite of the fact that the increase
of AN porosity caused the sharp drop of the ANFO mixtures density from 0,9 kg/l (for 2,5
cm3/100g porosity) to below 0,7 kg/l (for 15 cm3/100g porosity).
The dependence of ANFO detonation velocity and its density observed in the paper is
quite different from that existed in the case of typical high explosives where detonation
velocity of explosives is always proportional to its density. As the explanation of the
observed effect one can point out a dominant influence of the AN prills structure which
predominant the influence of density. Such a behaviour is observed also for other explosive
mixtures consist of two ingredients an oxidisers and a fuel.
From described experiments result that detonation velocity of ANFO mixtures depend in
a high degree on the particle size of AN. Charges of ANFO prepared from fine grade
nonporous granules are capable to detonation (in the condition of experiment) only for
granules below 1,0 mm of dimension. In that case the dependence of detonation velocity and
dimensions of particles is regular, as for other types of explosives. Charges of ANFO
mixtures prepared from nonporous, milled prills detonate with much higher velocity in
comparison with charges prepared from granulated AN of similar dimension particles.
Conducted experiments confirm an importance of prills physical structure (porosity and
particle dimensions) on detonation parameters of ANFO mixtures.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
W. PGOWSKI, B. ZYGMUNT, Porous ammonium nitrate, Polish Technical Review No 2-3,

p. 13-14, 1993
W. PGOWSKI, D. BUCZKOWSKI, B. SUBOCZ, Adaptation of quality of domestic
ammonium nitrate to EU requirements, Research Works of Institute of Inorganic Technology
and Mineral Fertilisers of Wrocaw Technical University, No 39, p. 35- 42, 1994 (in polish)
C. FAUQUIGNON, The two-steps reaction zone structure: a way to interpret non-ideal
detonations observations, Propellants, Explosives, Pyrotechnics, No 3, p. 123-125
D. BUCZKOWSKI, W. PGOWSKI, B. ZYGMUNT, Evaluation of the influence of
modificating substances on resistance to detonation of fertiliser grade ammonium nitrate,
Proc. of the Third Seminar New Trends in Research of Energetic Materials, Pardubice, p. 2939, 2000
B. ZYGMUNT, Influence of physical structure on initiation to detonation non-homogenous
explosives, Organika - Research Works of Institute of Industrial Organic Chemistry, Warsaw,
p. 7-19, 1999 (in polish)
B. ZYGMUNT, Detonation properties of ammonals and hydroammonals, The 3rd Int. Autumn
Seminar on Propellant, Explosives and Pyrotechnics. Chengdu, China, p.341-347, 1999
51
DETO2D THE COMPUTER CODE FOR SIMULATIONS

OF GASEOUS DETONATIONS IN COMPLEX GEOMETRIES
Pawe Buraczewski*, Marek Sutkowski*, Andrzej Teodorczyk*and Shengjun Zhong**
*Institute of Heat Engeeniring, Warsaw University of Technology,
Nowowiejska 21/25, 00-665 Warsaw, Poland
**Industrial Explosion Protection Institute, Northeastern University,
Shenyang, China
Abstract
Mathematical model and two-dimensional computer code were developed for
simulations of propagation of detonation waves in tubes, channels and complex
geometry. The code is based on FCT numerical algorithm. It includes detailed schemes
for chemistry. The code is still under development. It is currently used for simulations of
detonation propagation, quenching and reinitiation in detonation arresters.
Keyword:
detonation, numerical simulations, algorithm
1. INTRODUCTION
A Fortran code DETO2D is dedicated to create the simulations of the two-dimensional
gas detonations controlled by multidimensional conservation equations of fluid dynamics
and chemical reactions kinetics. Still special cases of both zero-dimensional and onedimensional problem can be supported by changing corresponding parameters in the input
case-file. Since the governing equations of deflagration (Navier-Stokes equations) and
detonation (Euler equation) are similar, DETO2D code is expected to be extended to gas
deflagration as well as dust explosion and detonation.
DETO2D code is based on the LCPFCT (to solve Euler equations) and CHEMKIN (to
model chemical reactions) packages. Up to now many models of chemical reactions kinetics
are used [1,2,3,5,6].
The models used in the code can be classified by either thermodynamic properties or
chemical reactions models. To evaluate the thermodynamic properties two methods are
introduced:
simple method relating pressure and internal energy by polytropic constant ,
Newtons iteration method, which is highly accurate.
Paralelly, two chemical reactions models are implemented:
progress variable model,
elementary chemical reactions model.
According to the above rules, some typical models are catalogued in the following
classes:
A simple model relating pressure and internal energy or enthalpy by polytropic

constant. There are two subcases of reaction model. The first subcase uses
irreversible one-step Arrhenius expression to evaluate the chemical reaction rate[2,3];
52
the second one uses an expression for the reaction rate as an experimental function of
gas density and progress variable[1],
A simple model to relate the pressure and internal energy or enthalpy (Cp as a
function of temperature) with the two step induction-reaction model[6],
A detailed model to relate pressure and internal energy (enthalpy as a function of

mass fraction and temperature) with detailed elementary reactions model[5].
The code includes four main parts:
the main program to specify the problem and organize the calling sequences,
the LCPFCT interface,
the CHEMKIN interface,
IO subroutines.
The whole package must be supported by LCPFCT and CHEMKIN library subroutines.
Finally, the algorithms to solve the conservation equations are as follows:
Operator-splitting method (other names: the fractional method, time-splitting

scheme) to deal with multidimensional problem and chemical reactions,
FCT or MacCormack scheme to solve Euler equations,
An ODE solver or Newtons method to solve the chemical reactions problem.
2. MATHEMATICAL MODELS FOR GAS DETONATION

Two models are introduced in this chapter, one model is progress variable model, and
the other is elementary reactions model.
2.1 Simple model - Progress Variable Model

This model does not consider the detailed reaction mechanism. Instead a reaction
progress variable (0,1) is used in the conservation equations. The equation system is
given as follows:
v
+ ( u ) = 0
t
(2.1)
( uv )
+ (uvuv ) = p
t
(2.2)
( e )
v
v
+ ( ue ) = ( pu ) + Qw&
t
(2.3)
( )
v
+ ( u ) = w&
t
(2.4)
where specific total energy equals

v
p
u2
e=
+
( 1) 2
(2.5)
53
and
w& =
d
Ea
= A(1 ) exp
dt
RT
(2.6)
The pressure is calculated from (2.5) directly, and the temperature T is given by:
T=
pW
R
(2.7)
T=
p
Rg
(2.8)
or
where Rg is a constant for both reactants and products, W - molecular weight of the
mixture.
The model has the following assumptions and limitations:
It does not consider the detailed reaction mechanism. Chemical reaction can be
considered as a one-step reaction. Thus the capability to deal with different chemical
reactions is not flexible;
The one-step reaction is not valid when temperature is high;
The temperature calculation is not accurate. It is calculated from inaccurate pressure

and constant Rg , which in fact is not constant for reactants and products;
The reaction takes place in the whole grid cell; the reaction rate is expressed by
Arrhenius law;
Polytropic exponent is a constant. This is not accurate when applied for reacting
mixture. Eg., -exponent may range from 1,27 to 1,4 during reaction for
stoichiometric mixture of H2 and O2;
The use of the polytropic constant to connect energy and pressure is not accurate
when applied for reacting mixture;
Diffusions of momentum and energy are negligible;
Turbulence generation and dissipation are not considered;
Gas mixtures are ideal gases.
2.2
Elementary reactions model
2.2.1. Governing equations

This model considers the gas reactions as elementary reactions. Conservation equations
of the elemental species are solved.
The conservation equation for species s is:
( Yk )
v
+ ( Yk u ) = wk
t
(2.9)
54
where
wk =
dYk
= & kWk ,
dt
Yk is mass fraction of the kth species.

The total energy conservation equation is modified to be:
( e )
v
v
+ ( u e ) = ( pu )
t
(2.10)
where the formation of enthalpy is included in the specific total energy:

T
e = CvT + h0 +
v2
u
T0
(2.11)
Note that there is no source term in (2.10) because chemical energy is included in the
specific total energy. The expression of mole reaction rate & k can be found in CHEMKIN
manual[4].
Equations (2.1), (2.2), (2.9) and (2.10) form the equation system of the elementary
reactions model.
Another form of energy equation is:
( e )
v
v
+ ( u e ) = ( pu ) + ein ,k w& k
t
(2.12)
where
T
e = C vT +
T0
v2
u
2
(2.13)
The difference of (2.10) and (2.12) is that in (2.12) the standard heat formation is
included in the specific internal energy ein.
2.2.2. Evaluation of temperature and pressure
It is possible to use (2.5) to evaluate pressure and (2.7) to evaluate temperature. In this
case, equation system (2.1), (2.2), (2.9) and (2.12) can be used. As mentioned above, using
constant and gas constant Rg to evaluate pressure and temperature is not accurate. This can
be improved by solving a nonlinear equation that relates internal energy and temperature.
Accurate direct equation to connect pressure and internal energy is not available.
However, since internal energy is the function of temperature and composition of the
mixture the temperature is the base parameter and pressure can be calculated from
temperature using state equation of ideal gas. Now the problem is how to evaluate
temperature T when total energy is obtained or when the enthalpy is known.
In CHEMKIN, standard state enthalpy and standard molar heat capacity at constant
pressure of the kth species are evaluated by polynomial fit (N=5 in current):
o
C pk
= ank T n1
(2.14)
n =1
55
H ko =
298
C dt + H ko (298)
pk
N
H ko
a T n1 a N +1,k
= nk
+
RT n1 n
T
U ko = H ko RT .
(2.15)
(2.16)
(2.17)
U ko is a function of temperature T only. For ideal gases the superscript o can be

omitted:
uk =
Uk
Wk
(2.18)
u = u k Yk
(2.19)
k =1
uk is the specific internal energy of the kth species.

For a given internal energy u , to obtain the temperature a nonlinear equation must be
solved. It is time consuming, but must be done. The relation of u and T is:
N ank T n1 a N +1,k
u = RT
+
1 Wk Yk
T
k =1
n1 n
(2.20)
The nonlinear equation can be solved by Newtons method and the derivation of u
should be given to the equation solver:
du
= cv
dt v
(2.21)
Then pressure can be obtained from the state equation:

p=
RT
W
(2.22)
Applying the elementary reactions model to the numerical simulations result in high
accuracy but time-consuming calculations. It is possible to use more basic reactions instead
of elementary ones at the cost of accuracy. The fastest is the progress variable model still
accuracy of it can become unsatisfactory in some cases.
56
3. NUMERICAL METHOD FOR ELEMENTARY REACTIONS MODEL

Transfer of the conservation equations system given above to a general form gives:

+ E+ F =S
t x
y
(3.1)
where
u
v
uYk
vYk
Yk
w& k
u 2 + p
vu
F =
S = 0
= u E =
2
uv
v + p
0
v

p
p
e
u e +
v e +
(3.2)
To solve the above system of equations, operator split method is adopted. The original
conservation equations can be split into the following equations:

+ E=0
t x
(3.3)

+
F =0
t y
(3.4)
The equations describing chemical reactions form a set of the ordinary differential
equations
dT
1
=
dt cv
u
k =1
dYk
dt
d ( Yk )
= w& k
dt
(3.5)
(3.6)
The above split includes coordinate split and process split.

The simplest operator sequence is:
t
t
L(t ) = LR Lx (t )LR L y (t )
2
2
(3.7)
The general operator L( t ) is involved to represent the operator transferring the vector
n to n+1 in the form:
t t t t t t t t
L(t ) = LR Lx LR L y LR L y LR Lx
4 2 4 2 4 2 4 2
or
57
t t t t t t t t
L(t ) = LR L y LR Lx LR Lx LR L y
4 2 4 2 4 2 4 2
(3.8)
The Lx and Ly can be solved using FCT algorithm, LR is an ordinary differential equation
(ODE) which can be solved by an ODE solver. FCT algorithm is implemented by LCPFCT.
The chemical reaction equation is solved by VODE, an ODE solver provided by CHEMKIN.
For progress variable model numerical method is similar to the presented above, but
equation (3.6) is:
d ( )
= w&
dt
(3.9)
4. SIMULATIONS
A number of results were obtained with the use of DETO2D code. The simulations
included propagation of the detonation wave in the pipes of complex geometry (like Lshaped or rounded pipe) as well as difraction of detonation from the confinement into an
open space. A case of the propagation of detonation through the pipe with the barriers placed
inside was also considered. Paralelly, the mixture gradients to simulate the process of
quenching and reinitiation of detonation was adopted.
The models used in the code assume uniform conditions along the pipe. Thus, the
reflected waves from the walls of the pipe must be produced they interfere creating the cell
structure in the detonation wave. There exist two methods, which are usually implemented to
get the reflected waves: to use a short sharp change in the width of the channel or to place a
set of a few small circle barriers at the onset of the pipe.
The simulations were created with the support of a detailed model relating pressure and
internal energy (enthalpy as a function of mass fraction and temperature) with more basic
(instead of elementary) reaction model. The results of the cell structure simulations were
compared with the experimental data gathered in the Detonation Database of the Explosion
Dynamics Laboratory at the California Institute of Technology [7]. According to the data
concerning detonation in the stoichiometric H2 + O2 mixture at temperature T = 293 K and
pressure p = 1 bar the cell size is about 1 mm while the simulated result reaches 2 mm.
Taking into account a simplified model of chemical reactions and influence of the size of the
mesh cells the results of the simulations can be assumed satisfactory.
The simulations presented below include propagation of the detonation wave along an Lshaped pipe and along a pipe of more complex geometry. Every figure consists of both
temperature and pressure fields at given time.
58
Fig 1.
Detonation wave approaching a right-angle corner. A fully developed

cellular stucture is visible (stoichiometric mixture of H2 + O2, T = 293 K,
p = 1 bar; channel width is 20 mm).
Fig 2.
Detonation wave directly after passing a right-angle corner.
59
Fig 3.
Reflected waves developed in detonation wave after passing a right-angle

corner. A slight turbulence caused by decompression waves is clearly
visible.
Fig 4.
Propagation of detonation wave along the pipe with complex geometry.

Since no reflected transverse waves are produced the cellular structure is
not recognized. Velocity vectors are additionally displayed (stoichiometric
mixture of H2 + O2, T = 293 K, p = 1 bar).
60
Fig 5.
Further propagation of the detonation wave after reflection from the first
barrier. A bright area represents the high temperature zone caused by the
passage of the reflected shock wave through the products of combustion.
Fig 6.
Detonation wave after passing through barriers. A huge turbulence is

clearly visible.
61
5. CONCLUSIONS
DETO2D code proved to be a usefull numerical tool for detonation simulations. It is
highly flexible when choosing between the rate and accuracy of calculations. The code was
validated against the experimental data and is presently used for simulations of detonation
propagation, quenching and reinitiation in detonation arresters. Paralelly, the code is being
developed to gas deflagration as well as dust explosion and detonation.
REFERENCES
[1] T. D. ASLAM, D. S. STEWART D. S.: Detonation Shock Dynamics and Comparisons with Direct
Numerical Simulation, Combustion Theory & Modeling, p.77-101,1999
[2] D. N. WILLIAMS, L. BAUWENS, E. S. ORAN: Detailed Structure and Propagation of Three
Dimensional Detonations, 26th Symposium (International) on Combustion, p.2991-2998, 1996
[3] V. N. GAMEZO, D. DESBORDES, E. S. ORAN: Two-dimensional Reactive Flow Dynamics in
Cellular Detonation Waves, Shock Waves, No.9, p.11-17, 1999
[4] J. R. KEE, M. F. RUPLEY, E. MEEKS, J. A. MILLER: CHEMKIN-III: A Fortran Chemical
Kinetics Package for the Analysis of Gas-Phase Chemical and Plasma Kinetics, SAND96-8216,
1996
[5] E. S. ORAN, J. W. WEBER, E. I. STEFANIW, J. D. ANDERSON: A Numerical Study of a TwoDimensional H2-O2-Ar Detonation Using a Detailed Chemical Reaction model, Combustion
and Flame, No.113, p.147-163, 1998
[6] E. G. PANTOW, M. FISCHER, T. KRATZEL: Non-stationary Detonation Program in Complex
Geometry, 26th International Symposium on Combustion, 1996
[7] www.galcit.caltech.edu/detn_db/html/
62
EMULSIONS EXPLOSIVES ELMULEX AND ELMULEXAL

Technical characteristics - testing and standards
L. ai*, R. Halle**and Z. Ester**
*Ministry of Interior,
Zagreb, Republic of Croatia
**Faculty of Mining, Geology and Petroleum engineering,
University of Zagreb, Republic of Croatia
Abstract:
If the ammonium nitrate (NH4NO3) we sensitive with the supplement of flammable
substance (put and alloy powdered, mineral oils, waxen, etc.) in quantities sufficient during
attaining the equal balance of oxygen, we will get explosive which will be simple in the
making, low price, sure by the handling and very brizant (for example brizant of mixtures
94,5% NH4NO3 and 5,5% of the mineral oil, which is measured with the ballistic pendulum
swinger, amounts 120% brizant of TNT).
Other kind of explosives on the base of the ammonium nitrate presents ammonium nitrate
sensitized with detonation sensitize (5-12% of the TNT or the larger percentage of the DNT
or 4-6% of glycerin nitrate beside the supplement of flammable substances for example the
coal powder) in quantities which are necessary during attaining the equal balance of
oxygen.
The third kind of ammonium nitrate explosive are called emulsions or mushy explosives
which introduce suspension of non dissolution nitrate, organic fuel and insoluble sensitive
factors in the satiated mixture of ammonium nitrate and sodium nitrate.
Water in emulsions explosives (11-20%) diminishes warmth of the explosion, what is
compensated with the larger volume of developed gases and large density of fillings.
Emulsions explosives have been developed with the aim of enlargement of efficiency up to
then the known explosive beside the simultaneous decrease their sensitivity this attaining
ideal technical characteristics.
In the emulsions explosive it is possible to achieve very large contact surface

between the ammonium nitrate and emulsions factor and a very short distance
between the oxidant and reducer, then it is therefore possible make the explosive
with technical characteristics of brizant explosive and with the minimal sensitivity,
in other words the large safety by the handling and uses.
We can say that the emulsion explosives unite characteristics of nitroglycerine explosives
and ammonium nitrate explosives (the sensitivity, brizant, reliability and efficiency in the
use) in a rate which is necessary, in other words in a rate which manufacturer has reaped.
By the emulsion explosives, how forward is quoted, it is possible to achieve different
technical characteristics (the ideal relation of consumption of explosive, quantity of
executed work on the drilling of mine bores and expected respective given effects of mining quantities and granulation of mined material, areas of throwing about of material, air
striking and seismic waves as a result the activation of explosive charge).
63
1.
SAPPER-TECHNICAL CHARACTERISTICS
Check sapper-technical characteristics of emulsions explosives does consistently following

standards:
2.
HRN.H.D8.002 - density of patroness;
HRN.H.D8.003 - sensitivity on initiating;
HRN.H.D8.005 - transfer of detonations;
HRN.H.D8.006 - velocity of detonation;
HRN.H.D8.012 - establishing of theoretical characteristics.
DENSITY OF PATRONESS
Density of patroness determines the standard HRN. H.D8.002.
Standard prescribes the method of testing of density patroness with mass less than 200 g.
3.
SENSITIVITY OF INITIATING
Sensitivity of explosive determines the standard HRN. H.D8.003.
Standard prescribes the method of testing sensitivity of explosive on initiating.
4.
TRANSFER OF DETONATIONS
Transfer of detonation of explosive determines the standard HRN. H.D8.005
Standard prescribes the method for the determining of transfer detonations of explosive for
the purpose of uniform establishing quality of explosive.
5.
VELOCITY OF DETONATION
Transfer of Speed detonations of explosive determines consistently the standard HRN.

H.D8.006.
Standard prescribes the method for the determining of speed of detonation for the purpose
of uniform establishing of quality of explosive.
6.
THEORETICAL CHARACTERISTICS
The theoretical characteristic of explosive we calculate consistently the standard HRN.

H.D8.011
Standard prescribes the way of calculation theoretical characteristic of explosive and this:
products of explosion, gas volume of products of explosion, warmth and temperature of
products of explosion for explosives with balanced, positive or the negative balance of oxygen.
64
7.
COMPARISON OF CHEMICAL STRUCTURE SOME AMMONIANITRATES EXPLOSIVES

-
Structure from the technical specification of manufacturer;
Structure from the chemical analysis of manufacturer;
Structure from chemical analyses Laboratory of Faculty of Mining, Geology and

Petroleum, University of Zagreb.
In observation have taken following explosives:

ELMULEX, ELMULEXAL
the emulsions explosive, manufacturer ELMECH-RAZVOJ d.o.o. Budinina, Gotalovec,
R. Croatia;
ELMEX, ELMEXAL
granulated the explosive, manufacturer ELMECH-RAZVOJ d.o.o. Budinina, Gotalovec,
R. Croatia;
PERUNIT 20, PERUNIT 28
plastic explosive, manufacturer ALIACHEM DIV SYNTHESIA, Pardubice, Semtn,
R. Czech;
GELAMON 30
gelatinous explosive, manufacturer SPRENGSTOFFWERK, Gnaschwitz GmbH,
Schlungwitz, R. Germany;
AMONEX 2
the powdery explosive, manufacturer VITEZIT PS, R. BeH;
PERMONEX V 19
the powdery explosive, manufacturer: ALIACHEM ands. the DIV. SYNTHESIA,
Pardubice, Semtn, R. Bohemia;
KAMEX M-15
water-plastic explosive manufacturer Chemical industry KAMNIK, Kamnik, Fuine 9, R.
Slovenia;
GELATINE DONARIT 2
plastic explosive, manufacturer DYNAMIT NOBEL WIEN, Werk ST. Lambrecht, R.
Austria;
ELMULEX
Technical specification of manufacturer
Structure
(%)
NH4NO3
NaNO3
Mineral oil and emulgator
Water
Glass micro ballons
70,0 74,0
7,4 - 10
4,0 6,5
10,5 13,0
1,5 3,5
65
ELMULEXAL
Structure
(%)
NH4NO3
NaNO3
Water
Al pauder
Glass micro ballons
70,0 74,0
7,4 - 10
4,0 6,5
10,5 13,0
5,0
1,5 3,5
ELMEX
Structure
(%)
NH4NO3
Mineral oil
93,0-95,0
5,0-7,0
ELMEXAL
Structure
(%)
NH4NO3
Mineral oil
Al pauder
91,0-93,0
3,0-5,0
3,0-5,0
EMSIT 1
Structure
(%)
NH4NO3
NaNO3
Water
Glass micro ballons
60,3-66,3
10,6-15,6
4,6 6,6
11,0-17,0
3,0-5,5
PERUNIT 28
Structure
Humidity
Nitrogliceroglikol
Dinitrotoluen
Meltabl in toluen
Meltabl in water
Non-meltabl rest
do 3,0
22,0+/-5,0
9,6+/-5,0
31,5+/-5,0
65,5+/-5,0
3,0+/-1,0
Structure
Structure from chemical Analyses of

Faculty (%)
NH4NO3
TNT
80,67
14,43
(%)
AMONEX 2
66
PERMONEX V 19
Structure

(%)
NH4NO3
TNT
Humidity
Iron oxide
80,1+/-4
18,8+/-3
max. 1
0,1+/-0,1
KAMEX M-15
Structure
(%)
NH4NO3
NaNO3
Trinitrotoluen
Water
Al pauder
Organically supplement
28,5
23,2
23,0
12,7
9,1
3,5
GELAMON 30 U
Structure
Structure from chemical

Analyses (BAM-025) - Dortmund
NH4NO3
Ethylenglykolnitrat
Dinitrotoluen
60,0
29,7
5,0
GELATINE DONARIT 2
Structure
NH4NO3
NaNO3
Trinitrotoluen
Nitroceluloza
Pentrit
61,8
13,9
7,7
13,0
1,0
(%)
By the doing of testing of explosive substance in the Laboratory of Faculty of Mining,

Geology and Petroleum, University of Zagreb, L. Pierottia 6. are established are the following
sapper technical characteristic of tested explosives - see Table 1:
67
Table 1. technical characteristic of tested explosives

EXPLOSIVE
ELMULEX
ELMULEXAL
ELMEX
ELMEXAL
PERUNIT 20
PERUNIT 28
AMONEX 2
PERMONEX
V 19
KAMEX
M-15
GELAMON 30 U
GELATINE
DONARIT 2
8.
DENSITY
OF
PATRON.
(kg/l)
VELOCITY
OF
DETON.
(m/s)
TRANSFER
OF
DETON.
(cm)
VOLUME
OF
GAS
l/kg
ENERGIE
OF
EXPL.
(kJ/kg)
TEMP.
EXPL.
OXYGEN
BALANCE
(%)
1,16
1,16
0,88
0,87
1,46
1,46
1,05
1,13
4650
5100
2700
3550
2270
5800
3480
3300
3
3
contact
contact
4
6
5
4
925
890
1050
965
820
820
954
900
3120
3520
3590
4310
4630
4630
3860
4360
2050
2150
2190
2470
3180
3180
2700
2800
+0,7
+0,7
+1,5
+1,9
+1,15
+1,15
-0,43
-0,81
1,45
4500
contact
784
4338
3120
-1,3
2600
1600
6
3
+2,2
CONCLUSION:
From forward quoted is obviously that emulsions explosives according to sapper technical
characteristics (regardless what have been made from the non-explosive component) not do lag
for other explosives in other words that have determined comparative advantages which them
enable the wide range uses for the purpose of going for a ride the sapper work.
Wrap round explosives comply requirements of sure handling, warehousing and transport
and not do introduce themselves dangers of ecological risks.
Have been intended for the economic use on superficial mining and underground mining, in
the circumstances mining without appearances of methane or comes as well as in mine bores
with the appearance waters, because appears the exceptional waterproof
REFERENCES:
[1] Miljenje i Izvjee o ispitivanju No. 100/999, 101/999, 152/000/A, 381/001, 434/002, 435/002,
442/002, 470/002, 482/002, 486/002, 487/002, 153/000/B, 154/000/A, Laboratory of Faculty of
Mining, Geology and Petroleum, University of Zagreb, L. Pierottia 6.
68
SYNTHESIS OF CERAMIC AND CARBON NANOSTRUCTURES

BY SELF-SUSTAINING COMBUSTION OF MIXTURES
OF HALOGENATED HYDROCARBONS WITH REDUCERS
S. Cudzio*, S. Gachet**, A. Huczko***, M. Monthioux** and W. A. Trzciski*
*Military University of Technology
Kaliskiego 2, 00-908 Warsaw 49, POLAND
**Carbons and Carbon-Containing Materials
CEMES, Rue Jeanne Marvig 29, 31055 Toulouse, FRANCE
***Warsaw University, Department of Chemistry
Pasteura 1, 02-093 Warsaw, POLAND
Abstract:
This paper is aimed at a study of reactions of some flouro- or chloro- hydrocarbons
(like PTFE, heksachlorobenzene, heksachloroethane) with different reducers, such as
Al, Mg, Zr, Ti, Zn, Cr, Si, alloys (ZrTi, FeSi) or compounds (CaSi2, Al3Mg4, NaN3).
Combustion process was carried out in a calorimetric bomb under argon (0.5 MPa).
The released heat of reaction was measured and solid reaction products were analysed
using ultimate analysis, XRD, SEM, TEM and HRTEM techniques to identify their
composition and morphology. We found out that the mixtures are able to self-sustaining
highly exothermic reactions (heat of reaction equals up to 8500 kJ/kg). The XRD
examinations showed that in addition to turbostratic graphite the combustion products
include crystalline phases of AlF3, MgF2, ZrF4, TiF3, CrF3, ZnF4, B4C, B13C2, Si, Fe3Si,
Fe5Si3, SiC, CaF2, and NaF depending on the starting material. The microscopic
observations revealed polyaromatic carbon and carbon nanoparticles (3040 nm in
diameter) in each sample and very interesting nanofibres in the combustion products of
silicon containing mixtures. They were identified as 1D single SiC crystals covered with
thin layers of amorphous carbon and SiO2.
Keyword:
1.
combustion synthesis, carbon and ceramic nanostructures
INTORODUCTION
Combustion synthesis is a fast and energy-efficient method of formation of many

valuable materials, such as carbides, borides, silicides, nitrides, titanides and oxides that can
be used as abrasives, cutting tools, ceramic reinforcement, nuclear safety shields, ball
bearings, high-temperature superconductors and so on [1]. Its advantages result from specific
conditions in the combustion wave, namely very high temperatures (exceeding 3000 K) and
heating rates (approaching 106 K/s). Thanks to this the chemical reactions are ultra-fast and
yield new compounds and structures that are difficult to produce using other methods.
A few years ago, it was reported that chemical and electrochemical defluorination of
fluorocarbon polymers (or other perhalogenated hydrocarbons) leads to unique structurally
carbon materials [24]. Electrochemical processes occurring at an interface of an alkali metal
amalgams and polytetrafluoroethene (PTFE) as well as reactions of PTFE powder with alkali
metal suspensions in indifferent organic solvents are promising ways of formation of novel
carbon nanostructures including mainly linear sp-bonded carbon chains (called carbyne) and
69
graphenes. Carbons obtained in such processes also contain a small amount of carbon
nanoparticles, like nanotubes and carbon onions in addition to fullerene C60. Electrochemical
carbonization and reactions of PTFE with alkali metal suspension in an organic solvent
proceed at mild conditions since the rate of the reactions is controlled by mass transport.
Consequently the CC bonds do not crack and the carbonaceous product structure is
considerably precursor-dependent. This allows the synthesis of polyyne-like skeleton from a
linear halogenated hydrocarbons or graphite from an aromatic precursor [2].
This paper is aimed at a study of the reactions between fine powders of halgenated
hydrocarbons and different reducers. The reactions were ignited in pressed samples with a
heat pulse and then they propagated along a sample in the form of a self-sustaining
combustion wave, as the reactions were fast and exothermic enough to assure such a regime
of propagation. Because of high temperatures in the combustion wave the synthesis
conditions were completely different from that of described above. To the best of our
knowledge, such reaction products were generally a matter of no interest [5, 6], except for
Ref. [7]. This paper presents some results of chemical and structural analyses of the solid
products resulting from the reactions. Elemental analysis as well as XRD, SEM and HRTEM
techniques were employed to characterize the chemical composition and structure of the
products.
2.
SYNTHESIS CONDITIONS
Fine powders of aluminum, magnesium, zirconium, titanium, zinc, chromium, silicon,

boron, their alloys (ZrTi, FeSi) and compounds (CaSi2, Al3Mg4, NaN3) were used as
reducers of polytetrafluoroethylene (PTFE), chlorinated polyvinylchloride (SPVC),
heksachlorobenzene (HCB) and heksachloroethylene (HCE). Binary mixtures of the
reactants were prepared by rubbing components together, taken in an adequate proportion [7].
Then they were pelleted into cylindrical samples of 10 g in mass. A sample was placed in a
graphite crucible and ignited by an electrically heated resistance wire. Argon at a pressure of
0.5 MPa was used as an atmosphere in each experiment. At least two measurements were
performed for each mixture. Figure 1 shows the calorimetric bomb used for the syntheses
and simultaneous measurements of heat effect accompanying the reactions.
5
4
3
1
Fig. 1. Calorimeter bomb prepared for an experiment,

1 pressed sample, 2 graphite crucible,
3 resistance wire, 4, 5 electrical terminals
and valves
70
Heats of reactions are collected in Table 1. They range from quite low (less than
2000 kJ/kg) for CaSi2/SPVC mixture to very high (more than 8500 kJ/kg) for the
composition containing Al3Mg4 and PTFE. Of course alongside the exothermicity changes
caused both by choosing different reducers and their contents, the temperature in combustion
wave and the synthesis condition inside it also dramatically change.
After finishing the calorimetric measurement, the solid combustion products were
removed from the bomb with water. The suspension was filtered off and the deposit obtained
was washed with acetone. After drying, the reaction products were analysed to identify their
composition and morphology.
3.
ELEMENTAL AND XRD ANAYSES
The elemental analyses of recovered solid products were carried out in order to
determine the total carbon content. All the result reported here were obtained using PerkinElmer Analyser CHNS/O Model 2400. The runs were performed applying an elongated time
of combustion. X-ray diffraction (XRD) spectra were measured using a Simens
Diffractometer D500 in conjunction with CuK radiation. Measurements were done in the
range of 2 from 20o to 70o with a step of 0.02o. The XRD data were calculated and the
phase analysis performed utilizing an in-house program, XRAYAN. Results of the analyses
are presented in Table1.
Table 1. Heats of reactions and results of elemental and XRD analyses
Composition (wt %)
Al3Mg4/PTFE = 34.5/65.5
ZrTi/PTFE = 49.3/50.7
Zn/PTFE = 56.7/ 43.3
Cr/PTFE = 65.9/34.1
B/PTFE = 22.6/ 77.4
Si/PTFE = 26.0/74.0
Si/PTFE = 56.0/44.0
FeSi/PTFE = 35.7/64.3
CoSi2/PTFE = 22.8/77.2
CaSi2/PTFE = 57.8/42.2
NaN3/PTFE = 62.2/37.8
NaN3/C6Cl6 = 47.8/52.2
NaN3/C2Cl6 = 52.2/47.8
NaN3/SPVC = 55.0/45.0
CaSi2/SPVC = 43.2/56.8
CaSi2/C6Cl6 = 16.85/83.15
CaSi2/C2Cl6 = 39.6/60.4
Heat of
Reaction
(kJ/kg)
8520
4730
2775
2260
5130
5590
3880
5030
4700
4900
3460
2648
2590
2700
1970
2430
3100
Carbon Content
(wt %)
17.1
13.8
10.7
11.8
58.9
79.6
21.8
44.2
43.3
22.1
60.3
80.1
77.4
79.0
22.6
64.3
17.2
Phases Identified
C, MgF2, AlF3, Al
C, ZrF4, TiF3
C, ZnF4
C, CrF2, CrF3
C, B4C, B13C2
C, Si, SiC
C, Si, SiC
C, Si, SiC, Fe3Si, Fe5Si3
C, SiC, Si
C, SiC, Si
C, NaF
C
C
-
Depending on the initial composition of mixtures, the carbon content changes from
about 10 % to about 80 %. The highest yields of carbon were obtained from PTFE and Si- or
B-containing precursors owing to the volatility of SiF4 and BF3 and from mixtures of sodium
azide with chlorinated hydrocarbons because of NaCl good solubility in water. The fact that
71
carbon is not the only combustion product of NaN3/HCB and NaN3/HCE mixtures implies
that the dechlorination is not complete. Residual chlorine constitutes about 20 % of the
samples.
XRD analyses showed that alongside expected products, i.e. carbon and fluorides or
chlorides of a reducer, the tested samples contained a considerable amount of carbides and
silicides. Turbostratic graphite phase was identified in all products with the strongest peak at
2 in the 26.5026,65 range. The highest content of graphite-like carbon was again found in
Si/PTFE and B/PTFE products, that confirms the above conclusion. In some cases, unreacted
reducers were also detected in the reaction products, e.g. aluminum or silicon, Table 1. More
phases appeared when the reducer was not an element but an alloy or a compound. In the
presence of iron, silicides Fe3Si and Fe5Si3 were created, and all combustion products of
precursors containing silicon included a considerable quantity of silicon carbide, Fig. 2.
Fig. 2. Comparison of XRD spectra of products resulting from combustion of

following mixtures: CaSi2/PTFE = 22,8/77,2 (curve 1), CaSi2/PTFE =
57,8/42,2 (curve 5), and after calcination of the latter one (curve 6).
From Fig. 2 it follows that an increase in calcium silicide mass fraction in starting
mixture reduces quantity of carbon in the combustion products and increases silicon carbide
content (spectra 1 and 5). Calcination in air (for 1 h at 900K) of the reaction products of
CaSi2/PTFE = 57,8/42,2 mixture removes free carbon from the sample. Silicon carbide phase
remains due to its resistance against oxidation at the heating temperature.
The phase composition of reaction products of PTFE with boron is especially worth
mentioning. Their diffraction pattern is shown in Fig. 4. It indicates that the products include
two dominant crystalline phases, namely turbostratic graphite and boron carbide, B4C. The
formation of the compound in a quite big amount in such a fast process cannot be explain
assuming reaction between solid boron (melting point: 2673 K) and carbon (melting point:
3973 K). Probably the synthesis is facilitated by the presence of volatile boron and carbon
fluorides which provide an efficient mass transport of the reactants.
72
Fig. 3. Diffraction pattern of solid combustion products of B/PTFE = 22,6/77,4 mixture
4.
STRUCTURAL ANALYSIS
Structural features of the solid reaction products were examined using scanning electron
microscopy (SEM, LEO 5000, operating at 20 kV) and transmission electron microscopy at
low (TEM, 120 kV) and high resolution (HRTEM, 150 kV) with an X-EDS equipment
attachment, operating at 200 kV.
SEM observations showed that most of the samples are very heterogeneous. Various
types of structures, including amorphous carbon, graphens, uniform carbon nanoparticles of
3040 nm in diameter, spherical metal micro-particles and ceramic fibers of 10100 nm in
diameter and up to 10 m in length were detected. Nanofibers were present in combustion
products of all precursors containing silicon, however CaSi2/PTFE mixtures rich in calcium
silicide provide the highest yields of fiber-like structures. A typical morphology of the
elongated crystals is shown in Fig. 4.
(b)
Fig. 4. SEM images of nanofibers found in combustion products of FeSi/PTFE (a) and
CaSi2/PTFE (b) mixtures
73
Further structural examinations employing transmission electron microscopy revealed

more details of the sample morphologies. Diameters of fibers resulting from the Sicontaining precursors range within 10160 nm, but the most frequent ones are 2050 nm in
diameter. Most of the fibers are well-crystallized ones, with atom planes set-up
perpendicular to the longer axis of the crystal, Fig. 5a. Another frequent display was a
herringbone type, corresponding to the association of twinned, faulted crystals, Fig. 5b.
The fibers are gathered into bunches or isolated, and all of them are covered with a thin layer
(a few nm) of amorphous material.
Figure 2
(a)
(b)
15 nm
Fig. 5. HRTEM images of nanofibers, a well-crstallized 1D SiC crystal covered with

2-nm layer of amorphous material, b twinned SiC crystal
X-EDS analysis showed that these nanofibers are made from silicon and carbon as the
main elements and oxygen as a minor element. Taking into consideration results of XRD
analysis, it can be stated that they are silicon carbide mono-crystals covered most probably
with a mixture of amorphous silicon dioxide and carbon.
Figure 7
(a)
6 nm
5 nm
(b)
Figure 6
Fig. 6. HRTEM images of carbons found in the combustion products of CaSi2/PTFE

mixture, a polyaromatic carbon, b carbon nanoparticles
74
Polyaromatic carbon was found in all samples as a micro-porous material. Its structure is
usually turbostratic with a short range of perfect graphens, Fig. 6a. Carbon nanoparticles
were also present in addition to the graphite-like carbon. The particles are 20100 nm in
diameter with a statistical mode at 3040 nm, and they are in the form of empty, often
polyhedral carbon shells, Fig. 6b. The particle walls consist of several crystallized layers,
although severe layer defects were often observed. The inner open cavities are of ca 10 nm
in diameter.
5.
CONCLUSION
The obtained results of chemical and structural analyses show that the self-sustaining
combustion process of mixtures containing halogenated hydrocarbons and different reducers
offers an effective and low cost method of synthesis of interesting ceramic and carbon
nanostructures. Polyaromatic, micro-porous carbons and well-crystallized, empty and
polyhedral carbon nanoparticles are present in all samples. Reaction products of boron with
PTFE include a big quantity of boron carbide, B4C, in addition to turbostratic graphite. All
Si-containing precursors provide a high yield of silicon carbide, SiC, nanofibers. They are in
the form of well-crystallized, free of impurities, 1D mono-crystals, sheathed with an
amorphous C/SiO2 layer. Most probably the required fast mass transport to create
morphologies, so different from the reactants, is provided by volatile fluorides and chlorides,
but the formation mechanism of such compounds and structures calls for a more detailed
study.
Acknowledgment:
This research was partially supported by the State Committee of Scientific Research of
Poland, Grant 0 T00A 021 20
The authors would like to thank Dr Yan Qiu Zhu (University of Sussex, Brighton, UK)
for carrying out some of the HRTEM observations.
REFERENCES
[1] A. VARMA, S. ROGACHEV, A. MUKASYAN, S. HWANG, Combustion synthesis of advanced
materials: Principles and applications, Advances in Chemical Engineering, 24, p. 79, 1998.
[2] L. KAVAN, Electrochemical carbon, Chemical Reviews, 97, p. 3061, 1997.
[3] J. HLAVATY, L. KAVAN, The preparation of polymeric carbon from PTFE in the suspension of
alkali metals, Carbon, 37, p. 1029, 1999.
[4] S. SOSHI, K. HIDEYUKI, T. HIROKI, O. ASAO, S. YASUSHI, Y. YOSHIO, Electric double layer
capacitance of highly porous carbon derived from lithium metal and polytetrafluoroethylene,
Electrochemistry of Solid-State Letters, 41, A5, 2001.
[5] S. CUDZIO, Combustion of mixtures of metals with fluorocarbon polymers, (in Polish),
Wiadomoci Chemiczne, 55, p. 716, 2001.
[6] S. CUDZIO, W. A. TRZCISKI, Calorimetric studies of metal/PTFE pyrolants, Polish Journal of
Applied Chemistry 26, p. 25, 2001.
[7] A. HUCZKO, H. LANGE, G. CHOJECKI, S. CUDZIO, Y. QIU ZHU, H. W. KROTO, D. R. M.
WALTON, Synthesis of novel nanostructures by metal-PTFE thermolysis, Journal of Physical
Chemistry, 2003 (accepted for publication).
75
THE INFLUENCE OF THE TYPE AND INTENSITY OF IGNITION

SOURCES ON EXPLOSION PARAMETERS OF THE GASEOUS,
DUSTS AND HYBRID MIXTURES
Radka Czeczotkov and Betislav Janovsk
Dept. of Theory and Technology of Explosives, University of Pardubice, 532 10 Pardubice,
Czech Republic
Abstract
Explosion parameters e.g. maximum explosion pressure - pm, maximum rate of pressure
rise - (dp/dt)m and times connected with these variables: ti induction time (time from
the ignition to the first rise of pressure), tr reaction time (time when (dp/dt)m is got)
and the tm time when pm is got, are influenced by the type and intensity of ignition
sources at the gaseous, dusts and hybrid explosions. It means that variables, which are
calculated or evaluated from measured ones, e.g. flammability limits lower
flammability limit and upper flammability limit, are influenced too. Following types of
ignition sources were chosen for comparison: electric fuse, Vesuvit TN and
nitrocellulose. The intensity of ignition sources was varied. Attention was fixed on
gaseous mixture methane air and hybrid mixture black coal methane air.
Keyword:
1.
gaseous mixtures, hybrid mixtures, explosion parameters, ignition

sources
INTRODUCTION
Combustible dusts and gases or combustible vapours may coexist in air during the
processing and manufacturing of solvent-containing products or through carbonising of
overheated dusty products. The mixtures resulting from dust and gas are called hybrid
mixtures. The most serious coal mine explosions are hybrid explosions, involving
mixtures of both methane and coal dust with air. These very violent explosions, and others in
other industries, highlight the enhanced hazards for this class of explosion, Bartknecht
(1989).
Because accidental gas, dust and hybrid explosions can only occur within enclosures,
and because the initial combustion and pressure rises during the explosions are relatively
slow processes, various passive and active means of explosion control are possible, e.g.
venting (small, intermediate and large scale), inerting (nitrogen, carbon dioxide, steam,
argon, helium, flue gas, vacuum, solids), deluge (water deluge, sprinklers or other water
deluge systems), suppression (fire suppressant material), Baker (1991). Behaviour of
gaseous, dusts and hybrid mixtures and its explosion parameters, e.g. maximum explosion
pressure, maximum rate of pressure rise, times connected with these variables etc. are
necessary to know, to be able to develop some new control methods. But first, the suitable
experimental apparatus must be designed to measure explosion parameters. Many
experimental apparatuses were developed and constructed since, e.g. Hartmann apparatus,
Bartknecht 20 liter spherical explosion bomb, Eckhoff (1987), U.S. Bureau of Mines 20 liter
76
laboratory explosibility test chamber, NFPA68 (1988), experimental cylindrical 236 cubic
meter steel silo at Sotra in Norway, explosion chambers, apparatus for measurement of
minimum ignition temperature of dust clouds, hot-plate test apparatus, flame ignition
apparatus, the Nordtest apparatus for the measurement of minimum ignition energy of dust
layers, Eckhoff (1987), vertical dust detonation tube, Lee, et al (1981), experimental
apparatus of Campos, Campos, et al (1986), cylindrical dust bombs, Gaug, et al (1986), etc.
All of these apparatuses are different and also differently constructed and so various
types and intensities of ignition sources are used. Type and intensity of ignition source are
very important factors, which can influence measured explosion parameters and so variables,
which are calculated or evaluated from measured ones, e.g. flammability limits lower
flammability limit and upper flammability limit, Janovsk (2001a), Janovsk (2001b).
2.
EXPERIMENTAL
The mixture of the combustible gas or vapour with dust and air is started in the closed
and constant vessel by the enough strong ignition sources with known intensity. Course of
the explosion process is taken off by the sensing unit of explosion pressure and stored to the
memory of the PC. Maximum explosion pressure - pm, maximum rate of pressure rise (dp/dt)m and times connected with these variables: ti induction time (time from the ignition
to the first rise of pressure), tr reaction time (time when (dp/dt)m is got) and the tm time
when pm is got, are evaluated from the graph of the explosion after its software processing.
2.1
Testing apparatus
Used testing apparatus is composed of explosion autoclave and dosing device, Fig 1.
and Fig 2.
2.1.1. Explosion autoclave
Spherical steel vessel (Fig 3.) of volume 0,25 m-3 is composed of two hemispheres (1),
closed by closing stirrups (2) and with the top hinged part, which is operated by hydraulics.
Vessel is equipped with valves for filling with gases or vapours (10) and for exhaust
explosion products (11). Air valve (6) is used for arrangement of pressure inside the vessel
(underpressure, overpressure) and for mixture withdrawal used for check analysis.
The sensing unit of explosion pressure (4) and the temperature sensor for the measurement
of the temperature of the shell wrapper are built-in in the vessel.
There are two electrodes (5) in the geometrical centre of the explosion vessel, which are
used for clamping of the ignition sources. The ignition sources can be various.
77
Fig 1.
Closed explosion autoclave
Whirling device (3) is composed of small vessel for the location of the sample,
electromagnetic whirling valve, which is controlled by the time signals for setting-up
whirling time and whirling jet.
Fig 2.
Opened explosion autoclave
78
Fig 3.
Explosion autoclave 0,25 m-3, chart of the apparatus; 1 the body of the
autoclave; 2 closing stirrup; 3 whirling device; 4 piezoelectric
sensing unit of explosion pressure; 5 feed electrode; 6 air valve;
7 control desk; 8 evaluating device; 9 recording device; 10 gas
valve; 11 discharge device
Control desk (7 and Fig 4.) contains elements for control of hydraulics, adjustment
whirling pressure and timing circuit for regulation of composition and atmosphere pressure
and check of ignition circuit.
Fig 4.
Control desk
Recording device (9) is composed of double-channel A/D converter, personal

computer (8) and measuring and evaluating programme UnkelScope.
79
2.1.2. Dosing device of combustible gases

Dosing device is composed of storage of combustible gas, gas meter and dosing
balloon. Dosing through the gas meter is possible with the accuracy of 0,1 l and it means
accuracy of 0,04 vol. %, max. 0,65 vol. % for the apparatus with volume 0,25 m-3.
The influence of the temperature and atmospheric pressure is implicated in the calculation of
concentration of the combustible gas or vapour inside the autoclave.
2.1.3. Dosing device of combustible liquids
Dosing device is composed of a pipette or a burette, which are divided by the tenths of
millilitres and vapour generator, which is interconnected with the autoclave through
the valve (10), Smrnice SZ 214 . 71/92, Metodika SZ 214 . 94/94.
3.
RESULTS AND DISCUSSION
The influence of the type and intensity of ignition sources on explosion parameters of
the gaseous (methane air), dusts (black coal air) and hybrid mixtures (black coal
methane air) was measured. Experiments were provided in the explosion autoclave.
Electric fuse, nitrocellulose and Vesuvit TN (black blasting powder) were used as ignition
sources. Intensity of ignition sources has been changed according to needs during
experiments.
3.2
Hybrid mixtures
The dependence of explosion pressure of hybrid mixture with concentration 9 vol. % of

methane on dust concentration, when different values of ignition energy were used, is
imaged at Fig 5. In general, measured concentration of dust in the hybrid mixture at upper
flammability limit rises with the increasing value of the ignition energy, but only in the case
of the same type of ignition source. The exception is electric fuse, but it is not possible to
compare it with the ignition by nitrocellulose, because of the different mechanism of energy
liberation. Electric fuse (made of pyrotechnic mixture) bursts apart to several burning parts.
Ignition is then multi-point. Nitrocellulose has a negative balance of oxygen and that is why
its ideal burning depends on the content of oxygen in the mixture, compared to the electric
fuse containing balanced pyrotechnic composition, which is not dependent on the
atmospheric oxygen and therefore produces given quantity of energy regardless of
concentration of matters in the mixture. So, electric fuse moves the concentration of dust at
upper flammability limit higher.
The dust concentration 400 g.m-3 in the hybrid mixture gives at the intensity of ignition
energy 34,5 kJ (nitrocellulose) positive exam, while the exam at ignition energy 0,1 kJ
(electric fuse) is positive even at the dust concentration 500 g.m-3.It is probably caused by
the difference in the mechanism of energy liberation. Concentration, which can be regarded
as an upper flammability limit for nitrocellulose ignition source, is ignited by the electric
fuse with value of ignition energy of order 100 less. In practice, e.g. in mines, it is more
possible to meet mechanism of ignition similar to which the electric fuse provides short
electrical discharge or mechanical spark. The question remains, which ignition source would
be necessary to use not to influence flammability limits by its behaviour, to approximate the
practice by its mechanism of ignition and which would not be too financially exacting. Using
the capacitive sparks or squibs about the known value of the ignition energy is shown as the
best, as also Senecal (1988) mentioned.
80
0,8
0,7
0,6
pm [MPa]
0,5
0,4
0,3
Ignition by NC Ei = 34,5 kJ
0,2
Ignition by NC Ei = 1 kJ
Ignition by EF Ei = 0,1 kJ
0,1
0,0
0
100
200
300
400
500
600
700
-3
cp [g.m ]
Fig 5.
The dependence of explosion pressure (pm) of hybrid mixture with methane

concentration of 9 vol. % on the dust concentration (cp), when different
types and values of ignition energy were used (NC = nitrocellulose,
EF = electric fuse, Ei = the value of ignition energy)
1,0
0,0 vol. % of methane (black coal)
1,5 vol. % of methane
0,9
0,8
pm [MPa]
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0,0
0
250
500
750
1000
1250
1500
1750
2000
2250
2500
-3
cp [g.m ]
Fig 6.
The dependence of explosion pressure (pm) of hybrid mixtures with

concentrations 0, 1,5, 4,5, 7,5 and 11,5 vol. % of methane on the dust
concentration (cp)
81
1,0
0,9
3.0 vol. % of methane

0,8
pm [MPa]
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0,0
0
250
500
750
1000
1250
1500
1750
2000
2250
2500
-3
cp [g.m ]
Fig 7.
The dependence of explosion pressure (pm) of hybrid mixtures with

methane concentrations of 0, 3, 6 and 9 vol. % on the dust concentration
(cp)
The influence of the type of the ignition sources has been also examined by comparing
the explosion parameters of hybrid mixture black coal methane air with the different
concentrations of methane in the mixture, Fig 6. and Fig 7. The dust concentration at upper
flammability limit started to fall at hybrid mixtures from the concentration of methane
6 vol. % to 11,5 vol. %. The ignition by nitrocellulose with ignition energy 34,5 kJ was
negative at high concentrations of methane, e.g. 11,5 vol. %. Nitrocellulose was not able to
ignite these mixtures at high dust concentrations, because nitrocellulose did not have enough
oxygen to its ideal burning and therefore did not release enough energy for the development
of oxidation reactions.
The dependence of the explosion indicators ti and tm on coal-dust concentration at
different concentrations of methane is plotted at Fig 8. and Fig 9. Nitrocellulose with the
ignition energy 34,5 kJ was used as the ignition source. These times protract with the
approaching upper flammability limit and the dependence turns steeper with increasing
concentration of methane. Induction time (period) depends on energy content of mixture
the shortest induction time (period) will be at the stoichiometric concentration of mixture.
The values of induction times should be higher in the region around the lower flammability
limit of hybrid mixture, but the experiments did not verify present assumption. However, the
influence of results by used type of ignition sources was expected. Ignition by nitrocellulose
increases the rate of burning during initial period and therefore values of ti are falsified. On
the contrary, its ignition power is suppressed by lack of oxygen in mixture in the energetic
rich region and the initial phase is inhibited. Curves of the dependence tm on coal-dust
concentration, when the concentration of methane is different, are qualitatively in
accordance with preconditions, Zvalov (2002).
82
250
200
ti [ms]
150
100
50
0
0
250
500
750
1000
1250
1500
1750
2000
2250
2500
-3
cp [g.m ]
Fig 8.
The dependence of induction time (ti) of hybrid mixtures (concentrations 0,

1,5, 4,5, 7,5 and 11,5 vol. % of methane) on the dust concentration (cp)
1400
3.0 vol. % of methane
1200
tm [ms]
1000
800
600
400
200
0
0
250
500
750
1000
1250
1500
1750
2000
2250
2500
-3
cp [g.m ]
Fig 9.
The dependence of the time when pm is reached (tm) of hybrid mixtures

(concentrations 0, 3, 6 and 9 vol. % of methane) on the dust concentration
(cp)
83
3.3
Gaseous mixtures
0,8
Ignition by Vesuvit TN Ei = 34,5 kJ
0,7
0,6
pm [MPa]
0,5
0,4
0,3
0,2
0,1
0,0
0
Concentration of the methane [vol. %]
Fig 10. The dependence of explosion pressure (pm) of the gaseous mixture methane
air on the concentration of methane, when different types and values of
ignition energy were used (NC = nitrocellulose, EF = electric fuse,
Ei = the value of ignition energy)
0,9
0,8
0,7
pm [MPa]
0,6
0,5
0,4
0,3
0,2
0,1
0,0
0
10
12
14
16
Fig 11. The dependence of explosion pressure (pm) of the gaseous mixture methane
air on the concentration of methane, when different types and values of
ignition energy were used (NC = nitrocellulose, EF = electric fuse, Ei = the
value of ignition energy)
84
Tabulated values of flammability limits of methane are 5 and 15 vol. % for lower
flammability limit and upper flammability limit respectively, Baker (1991). Literary data,
however, do not mention experimental set-up and kind and intensity of ignition source is not
also known. Value of the lower flammability limit agrees with theory only using electric fuse
as ignition source, Fig 10. Flammability limits are 2,5 vol. % (lower flammability limit) and
13 vol. % (upper flammability limit) of methane using nitrocellulose as ignition source with
the value of ignition energy 34,5 kJ, Fig 11. The explanation can be following: nitrocellulose
has not enough oxygen to its own burning at the range from 13 vol. % to 15 vol. % of
methane and therefore it cannot generate enough heat for the propagation of the reaction and
that is why mixture is not initiated.
The dependence of maximum rate of pressure rise (dp/dt)m on the concentration of
methane at different values of ignition energy is imaged at Fig 12. and Fig 13. Maximum
rate of pressure rise increases with increasing concentration of methane till the optimum
(stoichiometric) concentration of gas is reached and then falls with another increase of the
concentration. It is possible to see, that intensity of ignition source influences the maximum
rate of pressure rise. Values of (dp/dt)m are the highest using nitrocellulose as ignition source
in measured concentration zone.
70
60
(dp/dt)m [MPa.s-1]
50
40
30
20
10
0
0
Fig 12. The dependence of maximum rate of pressure rise (dp/dt)m on the
concentration of methane at different types and values of ignition energy
(NC = nitrocellulose, EF = electric fuse, Ei = the value of ignition energy)
85
100
90
Ignition by Pi Ei = 0,1 kJ
80
(dp/dt)m [MPa.s-1]
70
60
50
40
30
20
10
0
0
10
12
14
16
Fig 13. The dependence of maximum rate of pressure rise (dp/dt)m on the
concentration of methane at different types and values of ignition energy
600
500
ti [ms]
400
300
200
100
0
0
Fig 14. The dependence of induction time (ti) on the concentration of methane at
different types and values of ignition energy (NC = nitrocellulose,
EF = electric fuse, Ei = the value of ignition energy)
86
The dependence of induction time on the concentration of methane at different types and
values of ignition energy is shown at Fig 14. The influence of the type of ignition is best
shown on this graph. In the case when the mixture is not explosive, induction time has its
limit in infinitude and it abbreviates with rising concentration of combustible in mixture and
it obtains its minimum at the stoichiometric concentration. Then the induction time increases
again with increasing concentration of combustible in mixture and it achieves its limit in
infinitude again above upper flammability limit. This process could be seen (Fig 14.) only on
the curve for electric fuse used as ignition source. Using both Vesuvit TN and nitrocellulose
as ignition source stifles the dependency of induction time on the concentration of methane.
Values of induction time equal to 0 were not measured ones, but tabular editor used 0 instead
of infinitude.
The dependence of the time when pm is got (tm) on the concentration of methane at
different types and values of ignition energy is plotted at Fig 15. This dependence is
generally the same as in the case of induction time. However, it has connection with the
maximum rate of pressure rise (dp/dt)m, which is influenced by the size of ignition and that is
why the measured dependencies are in coincidence with the theory, Zvalov (2002).
900
800
700
tm [ms]
600
500
400
300
200
100
0
0
Fig 15. The dependence of the time when pm is got (tm) on the concentration of
methane at different types and values of ignition energy
87
4.
CONCLUSIONS
The influence of ignition energy on upper flammability limit was acknowledged for
hybrid mixtures. In general, concentration of dust in the hybrid mixture at upper
flammability limit at constant concentration of methane rises with the increasing value of the
ignition energy, but only in the case of the same type of ignition source. Electric fuse (low
ignition energy) is able to initiate the mixture with higher concentration of dust in the hybrid
mixture than another ignition source (nitrocellulose). The explanation is possible to find in
different mechanisms of energy liberation infinitude.
Measured flammability limits for gaseous mixture methane air was compared with
literary data and the influence of these parameters was validated by used ignition sources.
The identity with literary data was validated only for electric fuse.
The influence of explosion parameters was also validated by used types and values of
ignition energy for gaseous and hybrid mixtures.
Acknowledgment:
The authors thank VVU, a. s., Ostrava Radvanice for supply of experimental
apparatus.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
Baker (1991) - Baker, W. E.; Ming Jun Tang; "Gas, dust and hybrid explosions"; Amsterdam,
Elsevier, 1991
Bartknecht (1989) - Bartknecht, W.; "Dust explosions: course, prevention, protection"; Berlin,
Springer-Verlag, 1989
Campos et al (1986) - Campos, J., Janeiro Borges, A. R., Bouriannes, R.; "Influence of heat
transfer on the explosion pressure of lean mixtures of methane-air-graphite"; Dynamics
of Explosions, AIAA, New York, NY, 1986
Eckhoff (1987) - Eckhoff, R. K.; "Assessing ignitibility and explosibility of dust clouds in
industry by laboratory tests"; Proc. of Shengyang Int. Symp. on Dust Explosions, Shengyang,
China, 1987
Gaug et al (1986) - Gaug, M., Knystautas, R., Lee, J. H. S., Nelson, L. S., Benedick, W. B.,
Shepherd, J. E.; "Lean flammability limits of hybrid mixtures"; Dynamics of Reactive Systems,
Part II: Medeling and Heterogeneous Combustion, AIAA, New York, NY, 1986
Janovsk (2001a) - Janovsk B.; "Vbuch ve vnitnm prostoru objekt, vnitn vbuch";
uebn texty k pedmtu teorie psoben vbuchu, Univerzita Pardubice, Katedra teorie a
technologie vbuin, 2002
Janovsk (2001b) - Janovsk B.; "Vbuchy a jejich nsledky"; uebn texty k pedmtu
bezpenostn inenrstv, Univerzita Pardubice, Katedra teorie a technologie vbuin, 2002
Lee et al (1981) - Lee, J. H. S., Guirao, C. M., Grierson, D. E.; "Fuel-air explosions";
Proceedings of the International Conference on Fuel-Air Explosions held at McGill University,
Montreal, Canada, 1981
Metodika SZ 214 . 94/94 - Podstawka, T., Vdeckovzkumn uheln stav, a. s.; "Metodika
zkouen technicko-bezpenostnch hodnot holavch plyn a par dle poadavk ISO 6184";
Ostrava - Radvanice, 1995
88
[10] NFPA68 (1988) - Anonymous; "NFPA 68, Venting of deflagrations, 1988 Edition"; Nat. Fire
Prot. Assoc., Quincy, MA, 1988
[11] Senecal (1988) - Senecal, J. A.; "Explosion suppression of high KSt dusts"; Flammable Dust
Explosion, Int. Bus. Comm., St. Louis, MO, 1988
[12] Smrnice SZ 214 . 71/92 - Podstawka, T.; Vdeckovzkumn uheln stav, a. s.; "Souhrn
metodik pro zkouen technicko-bezpenostnch parametr"; Ostrava - Radvanice, 1995
[13] Zvalov (2002) - Zvalov, R.; "Men vbuchovch parametr sms holavch plyn
a prach"; Diplomov prce, Univerzita Pardubice, Katedra teorie a technologie vbuin, 2002
89
THE INITIATION OF GASEOUS DETONATIONS

IN H2-O2 MIXTURES BY INCIDENT SHOCK WAVE
Andrzej Dbkowski, Artur Kozak and Andrzej Teodorczyk
Warsaw University of Technology,
ul. Nowowiejska 21/25, 00-665 Warszawa,Poland
Abstract:
Detonations are an extremely efficient means of burning a fuel-air mixtures and
releasing its chemical energy content. Recently there has been renewed interest in the
application of intermittent or pulsed detonation to propulsion The application of
detonations involves many difficulties in rapidly mixing the fuel and air, and initiating
and sustaining a detonation in a controlled manner.
One method of initiating detonation is by using a shock wave to heat the mixture to
a temperature, which is above the autoignition temperature.
The initiation of detonation by planar shock in constant area tubes has been studied in
the past by several authors. In most of these studies the conventional single-diaphragm
shock tube technique was used to initiate detonation. This method suffered from several
disadvantages, notably those arising from the nonideality of the diaphragm rupture
introducing very nonuniform conditions at the interface separating driver gas from the
combustible test mixture. For this reason this technique was displaced by removable
sliding plate method which was unfortunately also causing nonuniformity in the
interface zone.
The main objective of the present experiments was to perform preliminary studies of the
influence of disturbances on autoignition behind planar shock and initiation of
detonation in H2 - O2 mixture. The disturbances at the interface separating buffer gas
from the test mixture were introduced by means of single wire, multiple parallel wires
and wire meshes of various density.
Keyword:
1.
H2 - O2 mixture, autoignition temperature, initiation, detonation
INTRODUCTION
There are numerous examples of industrial accidents involving gaseous detonations that
have resulted in severe destruction of capital equipment and loss of life. Conditions that lead
to detonative combustion are more prevalent today in large modern installations with long
pipe runs, large vessels, and high-velocity vapor lines than those in the smaller refineries in
the past. Unfortunately, the majority of industrial installations is not designed to withstand
detonations, this is simply too expensive and not recommended by risk analysis. So, the only
way to prevent serious disasters and losses in human lives and property is to prevent
detonation formation.
Detonations are, on the other hand, an extremely efficient means of burning a fuel-air
mixtures and releasing its chemical energy content. Recently there has been renewed interest
in the application of intermittent or pulsed detonation to propulsion The application of
detonations involves many difficulties in rapidly mixing the fuel and air, and initiating and
sustaining a detonation in a controlled manner. One method of initiating detonation is by
using a shock wave to heat the mixture to a temperature, which is above the autoignition
temperature.
90
Shock-induced ignition in gaseous hydrogen-air mixtures may occur in one of two

distinct modes, depending on the thermodynamic state in the shocked material. At lower
temperatures the ignition is weak, or mild, with the gradual development of the gas dynamic
explosion. At higher temperatures it is strong, or sharp, with an abrupt appearance of a
secondary shock induced by the explosive reaction. Furthermore, at low temperatures the
formation of distinct flame kernels appears to be an essential precursor of ignition. Meyer
and Oppenheim[1] point out that the intrinsically turbulent flow field behind a reflected shock
in a shock tube results in a nonuniform temperature distribution which creates distinct
reaction centers and leads to weak ignition. Although the nonuniformity of temperature
caused by turbulence is one mechanism which gives rise to reaction centers, the
nonsteadiness in the velocity of the causal shock can also produce them. This latter effect has
been demonstrated by Strehlow et al. [2] in their studies of shock propagation in a slowly
converging channel. It has also been shown that when an incident shock begins to accelerate,
the hot spots occur before the transition to detonation [3,4].
The initiation of detonation by planar shock in constant area tubes has been studied in
the past by several authors. In most of these studies the conventional single-diaphragm shock
tube technique was used to initiate detonation. This method suffered from several
disadvantages, notably those arising from the nonideality of the diaphragm rupture
introducing very nonuniform conditions at the interface separating driver gas from the
combustible test mixture. For this reason this technique was displaced by removable sliding
plate method, which was unfortunately also causing nonuniformity in the interface zone.
The main objective of the present experiments was to perform preliminary studies of the
influence of disturbances at the inert-combustible interface on autoignition behind planar
shock and initiation of detonation in H2-O2 mixture. The disturbances at the interface
separating buffer gas from the test mixture were introduced by means of single wire,
multiple parallel wires and wire meshes of various density.
2.
EXPERIMENTAL
The detonation tube shown in Fig.1 was used in the experiments. It consisted of a 1 m
long booster and 8 m long square cross-section channel with internal dimensions 110110
mm. The booster was filled with the oxy-acetylene stoichiometric mixture, which ignited, by
a 1 J electric spark rapidly detonated initiating in turn detonation in the acceptor mixture in
the main channel.
The incident shock wave was generated with the use of driver section by initiation of
detonation in stoichiometric acetylene-oxygen mixture. This detonation entered buffer
section filled with air at the pressure of 0.01 MPa, where detonation was attenuated and
transformed into shock wave. The shock wave hits next the membrane made of plastic foil
and enters test section containing stoichiometric hydrogen-oxygen mixture at the same
pressure of 0.01 MPa. The strength of incident shock wave was varied by the change of
initial pressure of acetylene-oxygen mixture in driver section within the range of 0.01-0.05
MPa.
Disturbances at the plane of interface between buffer and test section were generated
with the use of frames made of steel wire of the thickness of 3 mm. Four types of frames
were used in the study (Fig.2): #0 no flow disturbance, #1 one wire, #2 two crossed
wires, #3 six wires.
91
Fig 1.
Schematic of experimental apparatus.
First series of experiments was done with frame #0 in order to get reference results.
Next, the studies were performed with other frames making records of pressure from five
transducers and the signal from microwave radar. These data were further used for
determination of velocity of shock wave in buffer section and shock wave or detonation (in
case of initiation) in test section.
#0
Fig 2.
3.
#1
#2
#3
Frames used in the study for generation of disturbances.
RESULTS
Table 1 summarizes results of all experiments. The third column of the table shows
shock velocity in buffer section, fourth column shows average shock (or detonation) velocity
in the test section, the fifth column contains symbol used in the plot in Fig.3, and finally the
sixth column contains comment on the nature of phenomena occurring in the experiment.
92
Table 1. The summary of experimental results on detonation initiation by incident

shock wave
Frame
No of
experiment
Incident shock
velocity (m/s)
Shock velocity in
H2-O2 (m/s)
Symbol
Comments
#0
3a
1050
2750
Detonation
4a
900
2740
Detonation
5a
751
1360
No detonation
6a
840
1800
Detonation
(DDT)
7a
870
2500
Detonation
8a
800
2900
Detonation
9a
775
1900
Detonation
(DDT)
10a
900
2300
Detonation
(DDT)
11a
900
2800
Detonation
12a
790
1000
No detonation
13a
750
1300
No detonation
14a
775
1350
No detonation
15a
730
1850
Detonation
(DDT)
16a
650
2800
Detonation
17a
750
1600
Detonation
(DDT)
18a
700
2700
Detonation
19a
760
1750
Detonation
(DDT)
20a
790
2000
Detonation
(DDT)
21a
740
1150
No detonation
22a
760
2900
Detonation
23a
750
1150
No detonation
24a
775
2700
Detonation
25a
775
1400
No detonation
26a
730
2600
Detonation
(DDT)
#1
#2
#3
93
27a
825
2650
Detonation
28a
720
1900
Detonation
(DDT)
29a
775
1050
No detonation
30a
740
2600
Detonation
31a
740
1100
No detonation
32a
915
2500
Detonation
33a
750
1000
No detonation
Figure 3 shows the dependence of shock velocity in test section on shock velocity in
buffer section. Three distinct groups of points are clearly visible in this plot. Empty symbols
at the bottom show the events without detonation. Solid points at the top show the initiation
of detonation. In the middle of the plot there is a group of solid points for which the velocity
is too large for turbulent combustion but too small for detonation. This is the zone of limiting
states of deflagration to detonation transition (DDT).
3000
Shock velocity in H 2-O2 (m/s)
2500
2000
1500
1000
frame 0
frame 1
frame 2
frame 3
500
600
700
800
900
1000
1100
Shock velocity in air (m/s)
Fig 3.
Shock (detonation) velocity in combustible mixture versus incident shock

velocity.
94
For the reference case without disturbance the limiting shock velocity for detonation
initiation was 800 m/s. When the frame #1 was used detonation wave was initiated for much
smaller velocity of 650 m/s. With frame #2 used the limiting shock velocity of 750 m/s for
detonation initiation was greater then for the frame #1 but smaller than for reference case.
When the frame #3 was used the limiting shock velocity was 850 m/s, the larger value than
for reference case.
The results of experiments show that frames #1 and #2 introduce disturbances, which
result in transition to detonation at lower velocity of incident shock wave. Frame #3 causes
larger energy and momentum losses that overcome the benefits of turbulization by wires and
in result the velocity is larger than in reference case.
Figures 4-8 present the variation of shock and detonation velocity along the tube
calculated from the microwave radar signal and from the pressure transducers as well as
pressure profiles in time from five transducers. Vertical broken lines indicate the location of
pressure transducers. The short horizontal lines between broken lines indicate shock
(detonation) velocity calculated from pressure records.
Figure 4 presents reference case of the initiation of detonation by shock wave of the
velocity of 1050 m/s. Figure 5 shows the reference case of no initiation of detonation by the
shock velocity of 750 m/s. Figure 6 shows the case of detonation initiation by the shock
wave of the velocity of 650 m/s when the frame #1 with one wire was placed at the interface.
Figure 7 shows the case of detonation initiation by the shock velocity of 760 m/s for the
frame #2. Finally, fig.8 presents the case of failure to initiate detonation by the shock of the
velocity of 780 m/s for the frame #3.
4.
CONCLUSIONS
The experimental studies of the initiation of detonation by incident shock wave have
shown that disturbances generated at the interface between buffer and test section cause the
reduction of critical shock velocity, which initiates detonation. These disturbances generate
local zones of increased temperature and higher rate of chemical reactions leading to
transition to detonation. When the number of disturbing elements is too large then the
increased energy and momentum losses overcome the favorable conditions for detonation
initiation.
REFERENCES
[1] J.W. MEYER, A.K. OPPENHEIM, On the shock-induced ignition of explosive gases,
Thirteenth Symposium (Int.) on Combustion, The Combustion Institute, Pittsburgh, p.1153,
1971
[2] R.A. STREHLOW, A.J. CROOKER, R.E. CUSY, Detonation initiation behind an
accelerating shock wave, Combustion and Flame 11, p. 339, 1967
[3] T.V. BAZHENOVA, R.I. SOLOUKHINS, Gas ignition behind the shock wave, Seventh
Symposium (Int.) on Combustion, The Combustion Institute, Pittsburgh, p.866, 1959
[4] D.H. EDWARDS, G.O. THOMAS, T.L. WILLIAMS, Initiation of detonation by steady
planar incident shock waves, Combustion and Flame, 43, p. 187, 1981
95
FIGURES
003a
1
a)
AIR
H2+O2
3000
Velocity (m/s)
2000
1000
1000
100
b)
3000 4000 5000

Distance (mm)
6000
7000
C1
C2
C4
C5
C3
90
80
70
Pressure (bar)
2000
60
50
40
30
20
10
0
Fig 4.
3
4
Time (ms)
Shock wave (detonation) profiles along the tube calculated from radar
signal and pressure records (a); pressure records from five transducers (b);
incident shock velocity 1050 m/s; frame #0
96
013a
1
a)
AIR
H2+O2
3000
Velocity (m/s)
2000
1000
100
90
b)
80
Pressure (bar)
70
1000
2000
3000 4000 5000

Distance (mm)
6000
7000
C1
C2
C3
C4
C5
60
50
40
30
20
10
0
2.500
Fig 5.
3.375
4.250
Time (ms)
5.125
6.000
incident shock velocity 750 m/s; frame #0.
97
016a
1
a)
AIR
Velocity (m/s)
H2+O2
3000
2000
1000
1000
2000
130.0
3000 4000 5000

Distance (mm)
6000
7000
C1
C2
C3
C4
C5
b)
97.5
Pressure (bar)
65.0
32.5
0.0
2.500
Fig 6.
3.375
4.250
Time (ms)
5.125
6.000
98
021a
1
a)
AIR
H2+O2
3000
Velocity (m/s)
2000
1000
1000
2000
80
b)
3000 4000 5000

Distance (mm)
6000
7000
C1
C2
C3
C4
C5
70
60
Pressure (bar)
50
40
30
20
10
0
2.500
3.375
4.250
5.125
6.000
Time (ms)
Fig 7.
99
029a
1
a)
AIR
Velocity (m/s)
2000
1000
1000
2000
100
3000 4000 5000

Distance (mm)
6000
7000
C1
C2
C3
C4
C5
90
80
70
Pressure (bar)
H2+O2
3000
b)
60
50
40
30
20
10
0
2.500
Fig 8.
3.375
4.250
Time (ms)
5.125
6.000
100
DETERMINATION OF THERMAL STABILITY, RESISTANCE

TO WATER AND RESISTANCE TO HYDROSTATIC PRESSURE
OF ELECTRIC DETONATORS ACCORDING TO DRAFTS
OF EUROPEAN STANDARD
Mario Dobrilovi, Darko Vrkljan and Zvonimir Ester
University Of Zagreb, Faculty of Mining, Geology and Petroleum Engineering
10000 Zagreb, Pierottijeva 6, Croatia
Abstract:
In the paper are presented results of testing electric detonators according to New
European Standards.
In order to establish real, marginal values and values directed by the standard, testing
have been performed on suggested devices with extreme parameters applied.
Consequently, thermal stability, resistance to water and resistance to hydrostatic
pressure of electrical detonators have been measured with wider range that proscribed
by the standard.
The results obtained by the research were used to evaluate reality of proposed values in
New European Standards.
Keyword:
1.
testing of electric detonators, thermal stability, hydrostatic pressure,

European standards
INTRODUCTION
Modern methods for attesting and control examinations of explosives, detonators,

electrical detonators and pyrotechnical materials are executed according to Croatian
regulations and standards, while new European standards are in process of approval.
If Republic of Croatia becomes a member of CEN (European Committee for
standardization), examination of explosive materials will be done according to methods
regulated by European standards.
The technical board CEN/TC 321 gave suggestions to EN about field of examination of
explosive materials.
This paper brings a critical review on suggested European standard, Explosives for civil
uses-Detonators and relays:
prEN 13763-12 Determination of resistance to hydrostatic pressure, European standard
(doa 2002-12-31) EN 13763-2: 2002 Determination of thermal stability.
The testing instruments were constructed and examinations took place in the
"Laboratory for testing of explosives for civil uses, detonators, electrical detonators and
pyrotechnical materials" at the institute for mining and geotechnics of the Faculty of mining,
geology and petroleum engineering, University of Zagreb.
101
Respecting the standard examination methods, critical values of testing parameters were
defined. According to the results, one of the methods will be recommended as well as test
parameters that correspond to real values of detonator sensitivity.
2.
DESCRIPTION OF INSTRUMENT AND TESTING PROCEDURE

2.1
Instrument for testing detonators to hydrostatic pressure
It consist of water tank (with 0,5 m depth water), which is closed with steel cap. On the
cap are valve, pressure gauge and exhaust pipe. Water tank and compressor are connected
with air gut for rising of pressure in water tank. (fig. 1.)
1.
2.
3.
4.
5.
6.
7.
Fig 1.
2.2
Compressor
Valve
Water tank
Valve and exhaust pipe
Cap
Pressure gauge
Sealing plug
Parts of the instrument for testing detonators on hydrostatic pressure
Testing of electric detonators to hydrostatic pressure
Seventy-nine detonators have been prepared for hydrostatic pressure sensitivity testing.
Half second detonators with aluminum cover were tested, no. 2, PSED-Al, Pobjeda
Gorade, BiH. Total charge of detonator is 1 g, composed of 200 mg of initiation charge-led
azide and 800 mg (in two levels of 300 and 500mg) of high explosive charge-pentrite.
Detonators are prepared, numerated and photographed; testing procedure is carried out
according the procedure described in Croatian standards and European standard.
Croatian standard demand hydrostatic pressure which is equal to the pressure caused
with 2 m water column (eq.~19 613 Pa or~ 0.2 Bar) during 6 h. According to EN pressure is
0.3 MPa during 48 h.
Values of testing parameters were changed from standard prescribed values up to
extreme hydrostatic pressures, limited with construction properties of testing device.
102
2.3
Instrument for testing thermal stability of electrical detonators
Testing was performed in heating cabinet with temperature range from environmental up
to 220 C-MEMMERT TYP UP 400 capable for maintained temperature in range of 2 C.
In the main heating cabinet was chamber made of steel plate (15-20 mm). That chamber
has protection purpose for main heating cabinet against mechanical damages of detonator
parts and gases released in moment of detonation of tested detonator.
Chamber is consisted of two sub chambers with same volume. Left sub chamber is for
testing detonator and right is for temperature resistance sensor.(fig 2.)
Fig 2.
Chamber
Fig 3.
Testing equipment
1-Chamber with two sub chambers

2-Detonator
3-Temperature resistance sensor
Because if different heating conditions in cabinet and chamber there is need for
measuring actual temperature in chamber in order to established real temperature in moment
of detonation. After the calibration of temperature resistance sensor the output electrical
resistance is functionally depend on inner chamber temperature.
Temperature in main heating cabinet is controlled manually or by the computer program
Celsius for windows. Testing equipment is given on fig. 3. Fig. 4. shown screen plot of
Celsius for windows with parameters that are basic for testing.
Fig 4.
Screen plot "Celsius for Windows"

103
2.4
Thermal stability testing of electrical detonators
Half second detonators with aluminum cover were tested, no. 2, PSED-Al, Pobjeda
Gorade, BiH. Total charge of detonator is 1 g, composed of 200 mg of initiation charge-led
azide and 800 mg (in two levels of 300 and 500 mg) of high explosive charge-pentrite, four
items. Also, four fuse head of same detonators have been tested and one fuse head in empty
detonator shell.
The aim of testing was determination of critical parts of detonators in subjection of heat
and temperature in the moment of detonation.
Basic testing procedure is carried out according the procedure described in Croatian
standards and European standard with extreme parameters, limited with construction
properties of testing device. Croatian standard demand temperature of 60C during 24 h.
Demands of EN are: temperature 75C and 105C for products intended for use in hot
emulsion for a period of 48 h.
3.
RESULTS OF TESTING
3.1
Hydrostatic pressure
First sixty electric detonators are divided in twelve groups with five detonators. Each
group of detonators is brought into subjection of different pressures in range from 0,5 up to 6
Bar,eq. 0,05-0,6 MPa. Hydrostatic pressure of 0,5 m high water column was constant. Total
pressure values are shown in table 1.
Extra pressure up to 6 Bar was reached wit compressed air.
Hydrostatic pressure of water:
pH =* g * h
[Pa]
(1)
[Pa]
(2)
p H -hydrostatic water pressure [Pa]

water density 1000 [kg/m3 ]
g earth acceleration - 9,80665 [ m/s 2 ]
h water column height [m]
Detonator is in subjection of entirely pressure
p UK = p H +p
p UK - entirely pressure [Pa]

p H - hydrostatic water pressure [Pa]
p Z -compressed air pressure [Pa]
After 30 min detonators are taken out of water tank. In order to examine function of
detonators and possible damages they are fired.
Same procedure is carried out for next five detonators with difference in employment
time and pressure. According to demands of standard detonators were 48 hours in water
exposed to pressure of 6 Bar.
104
Water pressure did not caused damage and initiation of all 65 detonators was carried out
successful. During the firing of detonators, whistling sound was notified (with raising of air
pressure in water tank above 2,5 Bar entirely pressure on detonator was 257 400 Pa). That
sound was imputed to water that may have penetrated in detonator shell, fuse head or under
the isolation of leading wires.
Table 1. Pressure values
Detonator
No.
1-5
61116212631364146515661-
Electrical
resistance
()
4,1
4,0
4,1
4,1
3,9
4,2
3,9
4,2
4,2
4,0
4,2
4,2
4,1
Water
column
hight
(m)
0,
0,
0,5
0,5
0,5
0,5
0,5
0,5
0,5
0,5
0,5
0,5
0,5
Water
pressure
(Pa)
Air
pressure
(Pa)
49
50
10
15
20
25
30
35
40
45
50
55
60
60
4900,3
4900,3
4900,3
4900,3
4900,3
4900,3
4900,3
4900,3
4900,3
4900,3
4900,3
4900,3
Entirely
pressure
(Pa)
55
10
15
20
25
30
35
40
45
50
56
61
61
Eq. water
column
height
(m)
5,7
10,
16,
21,
26,
31,
36,
41,
46,
52,
57,
62,
62,
Time
(min)
3
3
30
30
30
30
30
30
30
30
30
30
Initiation
Notice
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
whistling
whistling
whistling
whistling
whistling
Initiation
Notice
yes
wh
Initiation
Notice
wh
whistling
whistling
wh
Table 2. Test results detonators without isolation on leading wires

Detonator
No.
1-5
Electrical
resistance
()
3,9
Water
column
hight
(m)
Water
pressure
(Pa)
Air
pressure
(Pa)
0,
49
60
Entirely
pressure
(Pa)
61
Eq. water
column
height
(m)
Time
(min)
62,
Table 3. Test results-fuse head

Fuse
head.
1-3
4-6
7-9
Electrical
resistance
()
3,3
2,7
2,5
Water
column
hight
(m)
0,
0,
0,5
Water
pressure
(Pa)
Air
pressure
(Pa)
49
20
40
60
4900,3
4900,3
Entirely
pressure
(Pa)
20
40
61
Eq. water
column
height
(m)
21,
41,
62,
Time
(min)
3
3
30
yes
yes
yes
That because, five new detonators are tested without isolation on leading wires.(fig.5.)
They were fired too, with whistling sound. Test results are given in table 2.
105
Initiation of tested detonator pointing on conclusion that penetrated water had not caused
serious damage for detonator function
Nine fuse head are tested according the same procedure (6 Bar, 48 h) and they were
initiated normally (fig. 6.). Test results are given in table 3.
Initiated detonator is shown on fig. 7.
Fig 5.
Fig 7.
3.2
Detonators without
isolation on leading wires
Initiated detonator
Fig 6.
Fig 8.
Initiated fuse head
Damages on fuse head
Thermal stability
Fuse heads were hold in heat cabinet without sub chamber during 3.5, 4,24 hours.
Reached temperatures were 120, 180, 200 and 220C. After the heating period, temperature
was maintained constant for period of 2 and 24 h. During that time fuse head has not
initiated. Demands of Croatian standard for thermal stability of electric detonator are
temperature of 60C for the period of 24 h. According to that demands one fuse head was
tested on double temperature value (120 C ) during the 24 hours. There was no initiation of
fuse head. Damages that had occurred on sealing plug and electrical conduits were melted
plastic isolators (fig.8.). One of fuse head was set in empty detonator shell in purpose of
simulating thermodynamic condition in detonator, also initiation was missed.
In other case when detonators were tested, temperature was rising from room
temperature up to 220C in two hours. Each detonator was fired on temperature above 160C
(table 4).
106
Table 4. Temperature
4.
Detonator No.
Temperature (C)
164,5
163,5
176
163,5
CONCLUSION
The starting points for determination of limiting parameters at both testing procedures
have been taken from relevant Croatian standards and suggested European standards in order
to recommend a methodology of laboratory testing, execution and usage of test instruments,
and to define parameters of influencing factors (hydrostatic pressure, temperature, time of
exposition).
4.1
Determination of detonators resistance to hydrostatic pressure
Test results are pointing on next conclusion:

Required pressure in European norm was chosen good. Pressure in Croatian norm is too
low. Customary deep of blast holes in quarries are approximate 20-25m. In case that the
whole hole is saturated with water, water pressure is 2 Bar. Testing of detonators on water
pressure of 3 Bar give 1.5 safety coefficient for ordinary electric detonators. In case of under
water blasting with highest water pressure detonators are specialty constructed.
Duration of pressure influence is too long in booth standards. After testing with pressure
of 6 Bar all detonators were detonated no matter how long they were exposed. In ordinary
conditions on blast field, 12 hours is maximal time of detonator exposition in hole with
water (working shift time)
4.2
Determination of thermal stability of electrical detonators
In ordinary use, detonators may be subjected to temperatures up to 70C (temperature in

vehicle parked on sun influence) or 80C (in hot pumped emulsion explosives). Exposure
period shall not be more then 12 h (working shift time). According to that 24 h testing period
is acceptable. Safety coefficient in that case is two time. European norm is in this matter too
rigor. EN makes difference in testing ordinary detonators and detonators intended for use in
hot emulsions. Values of employment temperature are adequately for laboratory testing with
enough safety coefficient.
As a matter of transport security and usage of detonators on a blast field, it can be
noticed that only strict obeying of the rules in manipulation and usage of detonators can
guarantee safety.
In order to give recommendation regarding testing parameters regulated by standards, it
is necessary to test more samples of different producers and that way determine the
variations in resistively to various types of loads, and thereafter regulate limits of testing
values by obeying the critical and tested values of the same, in other words by obeying the
safety requirements and required quality of products.
107
100
REFERENCES:
[1]
[2]
[3]
[4]
European Committee for Standardization, Brussels, European standard, drafts: prEN 13763-12:
October 2000, EN 13763-2: November 2002
State Office for Standardization and Metrology/DZNM, Zagreb, Croatian standard:
HRN.HD8.110, HRN.HD8.112.
SUESKA, M.: Test Methods for Explosives. New York: Springer-Verlag,1995, 225 pp
ISBN 0-387-94555-5, 1995.
Celsius for Windows, Operating instructions, Memmert;2000.
101
108
FIRST EXPERIENCE WITH ANALYSIS

AFTER LIQUIDATION OF SMOKELESS POWDERS
Eisner A.* , Adam M.*, Mikulkov P*, Ventura K.* and Kouba M.**
* University of Pardubice, Department of Analytical Chemistry,
Nm. s. Legi 565, 532 10 Pardubice, CZ
** University of Pardubice, Department of Theory and Technology of Explosives,
Doubravice 41, 532 10 Pardubice, CZ
Abstract:
The aim of this study is an identification of products after combustion of smokeless
powders. At first a sample of smokeless powder was burned on the clean sand and than
an isolation of compounds from this sand was used for further identification. The
accelerated solvent extraction and ultrasonication were used for the isolation of these
compounds. Gas chromatography GC 17A with mass spectrometry detector QP505A
was used for an identification of compounds.
Keyword:
1.
smokeless powder, GC/MS, extraction
INTRODUCTION
Old smokeless powders must be liquidated because content of stabilisator is low. These
smokeless powders are unstable and dangerous for people. The liquidation of smokeless
powders could be executed realise with their burning. Knowledge of compounds arising out
of smokeless powders is necessary for specification their effect to the environment.
The first step is an isolation of these compounds from sample. The classical [1] extraction
methods (soxhlet, ultrasonication) and new extraction techniques (accelerated [2] solvent
extraction, supercritical fluid [3] extraction) are used for it.
In principle, the accelerated solvent extraction (ASE) is an extraction process taking
place in the solid / liquid phase system and being performed at increased temperature (50200C), increased pressure (5-20 MPa) and within a relatively short time interval (max. 20
min.)
In order to perform ASE, the same solvents as for Soxhlet extraction can be used. More
over, it is possible to choose even such solvents that are less effective when used in classical
techniques. This is due to specially enhanced extraction kinetics of the extraction process in
ASE. It is very advantageous that extraction work with ASE can be made under conditions
of liquid phase. This allows one to use even mixed solvents as there is no risk of distilling
one component separatel.
109
2.
EXPERIMENTAL PART
The single-base smokeless powder was used for the first experiments. This one
contained a nitro-cellulose, a diphenylamine as stabiliser and a phthalates as plasticizers.
The sample of the smokeless powder (about 5 g) was burned on sand. Before
combustion the sand was cleaned in the water eight times and than was dried in the oven
(150C) for 8 hours. The sample of the sand (about 50 g) was extracted to 50ml of solvent
in ultrasonicate bath for 20 minutes, filtered and most of the solvent was evaporated to
volume 1 ml.
Also accelerated solvent extraction was used as a comparative extraction technique.
After pre-treatment, weight amount of the sample was placed into extraction cell with a
portion of glass wool on the bottom. Next portion of the wool was laid on the surface of the
sample and the remaining volume of the cell was filled up with glass balls. After inserting a
prepared cell into the extractor-heating oven, the requested value of both temperature and
pressure were adjusted by means of extraction solvent.
Two extraction steps were performed when the extract was being entrapped into the
same collecting vessel. The conditions for extraction were as following: weight amount
about 15 g of sand and extracted at temperature 110 C and pressure 10 MPa end extraction
time was 2 x 10 minutes. Acetonitrile was used as a solvent for both extraction methods.
The obtained extracts were analysed by using of the gas chromatograph GC 17A
coupled with mass spectrometry detector QP 5050A (EI, NCI, both Shimadzu) and GC/MS
solution data system (Shimadzu). The helium (grade 5.0, Linde) was used as carrier gas.
Separations were performed on a 30m x 25m i.d. capillary column coated with a 0,25 m
film of polymethylsiloxane (DB-5 MS). Split injection 1:10 was used. The column oven was
isothermally maintained at 45C for 5 minutes and than temperature increase for 20C/min
to 280C that was kept for 5 minutes. The temperature of injector was 220C and the
temperature of interface was 230C. The identification of compounds was based on the
comparison their mass spectrum with the spectrum in the library (NIST 62 and NIST 12,
Shimadzu).
3.
The single-base smokeless powders that contained a nitro-cellulose, a diphenylamine

and a dibutylphthalates were burned on the clean sand. Next it was extracted to the
acetonitrile and it was analysed with gas chromatography with mass detector. The
chromatogram of ultrasonication extract is shown in figure 1. There were only three peaks in
this chromatogram. Mass spectrums in the peaks were compared with library. Peak 1 was
identified as diphenylamine and peak 3 as dibutylphthalate. These compounds are from
original smokeless powders. Peak 3 was identified as N,N-diethyl-3-methyl-benzamine. It
could be any degradation products of the diphenylamine.
110
Fig 1.
The chromatogram of ultrasonication extract of sand
The chromatogram of ASE extract of sand is shown in figure 2. More compounds were
extracted with accelerated solvent extraction than classical ultrasonication extraction. Peaks
1, 2 and 4 were identified as aliphatic hydrocarbons (undecane, tridecane and hexadecane),
peak 7 as diphenylamine, peak 10 as N,N-diethyl-3-methyl-benzamine and peak 11 as
dibutylphthalate. The peaks 3, 5, 6, 8, 9 were not identified because their mass spectrums
were not in the mass library.
Fig 2.
4.
The chromatogram of ASE extract of sand
CONCLUSION
Performed experiments have been confirmed as the fact that accelerated solvent
extraction is better than ultrasonication extraction. The first experiment was realized. In
future we will try identifying most of compounds.
111
Acknowledgement:
Experiments were performed thanks to financial support from the Ministry of Education,
Youth and Sports of the Czech Republic (Project MSM 253100002), from the Grant Agency
of the Czech Republic (Project 203/02/0023) and from Grant Agency of the Czech Republic
(203/02/001).
REFERENCES
[1] J. ZINON, C. ZITRIN: Modern methods and applications in analysis of explosives. ed.:
John Wiley & Sons Chichester (1993)
[2] J.R.DEAN: Extraction Methods for Environmental Analysis, John Wiley & Sons,
Chichester 1998.
[3] A.A.CLIFFORD: Solvent Extraction - Supercritical Fluid Extraction, Encyclopedia of Analytical
Science 8, 4724, A.Townshend, editor, Academic press ltd., London 1995.
112
SYSTEMATIC STUDY OF SIMULATED MINE EXPLOSIONS

USING AUTODYN
Darina Fierov *,**, Amer Hameed*, Timothy A. Rose*,
John G. Hetherington* and Stanislav Prochzka**
* Cranfield University, Shrivenham, SN6 8LA SWINDON, United Kingdom
** Military Academy, Kounicova 65, 612 00 BRNO, Czech Republic
Abstract:
Within the mine-blast research community, there is an increasing desire to enhance the
efficiency and efficacy of mine-resistant vehicles, albeit in a cost-effective way. The
explosion of a surrogate mine is studied and modelled by using the GUI based nonlinear dynamics analysis software, AUTODYN. A simple model has been developed to
validate numerically obtained results with accessible experimental data. The simulation
setup consists of a hemispherical charge, laid on a rigid surface, with the blast waves
propagating into the surrounding air. This model setting covers our focus of studying
the explosion at close vicinity, which is currently not well laid out in available
literature. The crucial feature is an equation of state (EOS) for materials. The
surrounding air is described as an ideal gas. Two EOS of detonation products were
studied, with the first one being the commonly used empirical equation of state, JonesWilkins-Lee (JWL). The second area of study applied the ideal gas form equation of
state, often used for simplification in complicated models. The experimental data is
obtained from code CONWEP. Two parameters of blast waves, namely, maximum
pressure and specific impulse, are compared and evaluated. Further approaches have
been proposed for modelling the explosion in air close to the charge surface.
Keyword:
1.
Explosion, mine blast, numerical simulation, AUTODYN
INTRODUCTION
The Mine Ban Treaty of 1999 has not put a stop in any way to the global landmine
crisis, as an estimated 60 million mines lay strewn in around 70 countries all over the world.
Approximately 26,000 civilians are killed or maimed every year in minefields[4]. The
medical, psychosocial, environmental and economic impacts inevitably scale up further
tension and armed conflict in the affected areas. In mine-affected countries, scarcity of
adequate medical services, safe water and food has led to dependency on the international
community for humanitarian and development assistance. Mine contamination of
infrastructure disrupts relief supplies from reaching their intended destination. Therefore,
appropriate transport means for successfully negotiating minefields must be found in order to
make international assistance more far-reaching and cost-effective.
Using computer simulation, responses of a vehicle subjected to various blast loading can
be effectively and iteratively modelled, thereby reducing the cost of expensive field trials.
The first step in designing mine-resistant vehicles is to determine the loading on the vehicle
caused by a mine explosion. Although soil-blast interaction makes a significant contribution
to the total loading, the first task is to simulate the blast wave propagation in the surrounding
air. This was undertaken using the procedures provided by the commercial software package,
AUTODYN.
113
2.
BLAST WAVES IN AIR
Upon initiation, the detonation wave propagates through the explosive material, and high
pressures and temperatures are generated in the explosive products. These products expand
violently, force the surrounding air out of the occupied volume and create a compressive
shock wave in the surrounding air. The explosive products expand to a scaled radius of about
1/ 3
0.8 m.kg-1/3, where scaled distance is defined as Z = R /W , R [m] is the distance from
charge centre and W [kg] is the TNT equivalent charge mass. At a scaled radius of 1.6 m.kg1/3
, the whole of the compressive wave is propagating in air. With this in mind, analysis or
simulation focusing up to this region must take into account both explosive products and air.
The equations of motion, which describe a blast wave, are complex. They consist of
laws of conservation of mass, momentum, energy, and equations of state (EOS) for the
explosive products and surrounding air. An equation of state, which expresses a relationship
between the pressure p, specific volume v and temperature T, plays an important role. There
are two types of equation of state for explosive products:
a) EOS without explicit chemistry are based on experimental data for a particular
composition from which the formulae are derived.
b) EOS with explicit chemistry contain individual EOS for component molecules and rules
for their combining to give an EOS for any composition.
The particular application, the required accuracy and the method of solution are the main
factors that influence the decision of the choice of appropriate EOS for explosive products.
The cylinder expansion test was developed to derive the Jones-Wilkins-Lee empirical
equation of state for explosive products. In this test, a copper tube containing explosive is
detonated, and the cylinder wall acceleration caused by the explosive products expansion is
recorded with a high speed camera until the cylinder has expanded to about three times the
original diameter. The resulting empirical equation is a pressure volume relationship which
is independent of temperature. The JWL EOS is widely used in mine blast calculations,
because it is easy to program. It is implemented in many codes, such as LS-DYNA and
AUTODYN. It is argued that JWL fit to a cylinder test may be useful for describing that
particular cylinder test, but it will not be useful for describing anything else. [3] It is because
of this that the current work was conducted to evaluate the validity of the JWL EOS in mine
blast calculations. The JWL equation is implemented in the AUTODYN software in the
form given below
R1V
R2V .E
.e
.e
p = A.1
+
+ B.1
V
R2 .V
R1 .V
where A, B, R1, R2 and coefficients depend upon the composition of the explosive. The
variable V = v/v0 is the expansion of the explosive products and E [J.m-3] is the detonation
energy per unit volume. The surrounding air was assumed to be an ideal gas whose equation
of state is in the form p = (-1)..e, where is the adiabatic exponent.
114
3.
NUMERICAL SIMULATIONS
The explosive content of an anti-tank mine varies between 1.5 10 kg [5]. Most vehicles
have a ground clearance which ranges between 200 600 mm [6]. Therefore a numerical
model was established which used a hemispherical TNT charge of mass 10.19 kg, and the
resulting blast wave parameters at stand-off distances between 200 800 mm were
examined.
A two-dimensional axi-symmetrical model was created using a multi-material flow
solver in AUTODYN. Cell sizes of 5, 3, 1, 0.5, 0.1 and 0.05 mm were used to study the
mesh sensitivity. Boundary conditions were applied to the computational domain such that
no reflections were permitted. Initial material parameters are listed in Table 1.
Table 1. Material parameters[7]
AIR: the ideal gas EOS
[kg.m-3]
EINT [kJ.kg-1]
1.225
192 x 10
[-]
1.4
TNT: the Jones-Wilkins-Lee EOS
pCJ
-3
-1
[kg.m ]
[m.s ]
[GPa]
1 630
6 930
21.0
R1
R2
CJ
[GJ.m ]
[GPa]
[GPa]
[-]
[-]
[-]
[-]
373.77
3.7471
4.15
0.9
0.35
2.727
-3
TNT: the ideal gas EOS

[kg.m-3]
1 630
4.
EINT [kJ.kg-1]
3.681 x 10
[-]
1.4
RESULTS
Two blast wave parameters: maximum overpressure and specific impulse, were
evaluated for different stand-off distances using both the above EOS for different mesh sizes.
The numerical results were compared with empirical data from CONWEP [10]. CONWEP
generates overpressure and specific impulse data for a particular explosive and given scaled
distance. It is noted that CONWEP is unable to provide data for scaled distances below 0.18
m.kg-1/3. Maximum overpressures and specific impulses at scaled distances below this value
are evaluated by extrapolation.
Overpressure-time histories extracted from the analysis of the ideal gas are shown in Fig.
1, and those from the JWL analysis are plotted in Fig. 2. The overpressure-time histories
were recorded at distances from 200 mm to 800 mm with an increment of 100 mm.
Maximum overpressures at a particular scaled distance for the various mesh sizes considered
and for both EOS are presented in Table 2, which is graphically interpreted in Fig. 3.
115
Table 2. Maximum incident overpressure

Scaled
Distance
Distance
R
CONWEP
Z
-3
[mm]
[m.kg ]
200
0.092
Maximum incident overpressure

JWL
JWL
JWL
5 mm
3 mm
1 mm
JWL
JWL
JWL Ideal Gas Ideal Gas Ideal Gas
0.5 mm 0.1 mm 0.05 mm 5 mm
0.5 mm
[kPa]
[kPa]
300
0.138
[kPa]
[kPa]
[kPa]
[kPa]
[kPa]
[kPa] [kPa]
41
6501) 333 914 241 910 160 573 128 930 44 349 41 738
27
0901) 71 756 68 005 36 148 31 747 27 392 27 498
20 747
20 083
20 999
400
0.184
19 150
42 028
25 158
22 543
21 088 20 315 20 656
9 912
9 802
9 512
500
0.231
14 420
18 571
18 265
15 634
16 256 16 059 16 314
8 289
8 168
8 160
12 247
---
2)
7 112
7 018
7 228
---
2)
6 240
6 225
6 494
0.369
7 667
8 549
8 189
8 310
9 221 8 866 --5 528
Scaled distance is out of range for duration calculation. Waveform may not be accurate.
2)
These distances were not modelled due to limitations of hydrocode.
5 581
5 876
600
700
800
0.277
0.323
11 360
9 230
14 752
11 137
12 933
10 114
9 990
13 196 13 005
10 963 10 680
2)
[kPa]
3 mm
98 584 103 429 100 213
1)
For the models using the ideal gas EOS, it was observed that cell size had little influence
on the magnitude and shape of the resulting blast waves. Comparison with CONWEP
showed that maximum overpressures were underestimated by 50 30 % at distances
exceeding 400 mm, as shown in Fig. 3. The CONWEP data below 400 mm were
extrapolated and therefore were ignored in this comparison. The specific impulse results
over the same range were in closer agreement with CONWEP. Plots of specific impulse are
presented in Fig. 4.
Results obtained using the JWL EOS were observed to be highly dependent upon the
mesh size. For the largest cell size, maximum overpressures were up to 8 times those of the
fine mesh analysis at distance of 200 mm. It was also observed that for cell sizes of 0.1 and
0.05 mm, maximum overpressures were convergent and comparable with CONWEP data
over the whole range considered, including the extrapolated region. Interestingly, opposite
trends were observed with respect to the specific impulse. Fine meshes resulted in the
greatest underestimation of specific impulse (up to 70%), while coarser meshes predicted
specific impulses only 25% lower at distances exceeding 400 mm.
Examination of the ideal gas and JWL pressure records (Figs 1 and 2) shows that the
compressive shocks were sharp in the case of JWL, while they were flat and rounded in the
case of ideal gas.
116
120
Mesh size:
3 mm
0.5 mm
Overpressure [MPa]
100
80
60
40
20
0
200 mm
400 mm
800 mm
600 mm
-20
0
30
60
90
120
150
180
210
240
270
300
Time [microsecond]
Fig 1.
Pressure-time histories for model using the ideal gas EOS
150
Mesh size:
0.05 mm
0.1 mm
0.5 mm
Overpressure [MPa]
120
90
60
30
0
200 mm
300 mm
400 mm
500 mm
-30
0
15
30
45
60
75
90
105
120
Time [microsecond]
Fig 2.
Pressure-time histories the for model using the JWL EOS
117
1,000
Dash line indicates extrapolated CONWEP data
CONWEP
JWL, 3 mm
JWL, 1 mm
100
JWL, 0.5 mm
JWL, 0.1 mm
IG, 3 mm
10
1
100
200
300
400
500
600
700
800
900
Distance [mm]
Fig 3.
Maximum overpressure versus distance
100,000
Dash line indicates extrapolated CONWEP data
CONWEP
JWL, 5 mm
JWL, 3 mm
Specific impulse [Pa.s]
Maximum overpressure [MPa]
JWL, 5 mm
JWL, 1 mm
10,000
JWL, 0.1 mm
IG, 3 mm
1,000
100
100
200
300
400
500
600
700
800
900
Distance [mm]
Fig 4.
Specific impulse versus distance
118
5.
DISCUSSION AND FUTURE WORK
Convergence of the numerical results with CONWEP data occurs with increased range
from charge centre, for both the equations of state investigated. The model, which used the
ideal gas EOS, does not agree with the blast wave parameters of CONWEP at close-in
distance, where the blast wave parameters are highly influenced by the explosive products.
Therefore the ideal gas representation will not be considered in future studies.
Results obtained from the JWL EOS model are not satisfactory within all the region
observed. Fine meshes give comparable maximum overpressures and the most
underestimated impulses, while the large cell sizes result in pressure overshots at close-in
distances and specific impulses 25% lower. Huntington-Thresher[8] concluded from his
experimental analysis that CONWEP overestimated the specific impulse by 20-25%. Taking
this into consideration, specific impulses predicted by the model using JWL with a coarse
mesh were in good agreement with experiments.
Further study of numerical methods, used for modelling reactive flows, will be carried
out to find more appropriate modelling tools. The following approaches will be considered:
a) Numerical methods for multi-material flow.
b) Implementation of static mesh refinement which is possible in AUTODYN
presents the opportunity of resolving the rapid changes occurring in the shock waves.
This would allow the pressure overshot of the coarse mesh to be avoided while
preserving reasonable agreement with specific impulse. Such a calculation will not be
as computationally expensive as one with a constant fine mesh.
c) Alternative equations of state for explosive products will be investigated with the aim
of modelling more accurately the propagation of both the detonation wave in the
explosive and blast wave forming in the surrounding air close to the charge. It is
reported that the Williamsburg EOS[9] (illustrated in Fig.5) gives better agreement
with experimental data for small reduced volume than the JWL EOS. The reduced
volume is defined as the ratio v/v0, while the reduced pressure is the ratio p/p0. This
thought is supported also in the AUTODYN manual where it is stated that Care
must be taken in using JWL EOS if pressures go significantly above the CJ value
since the exponential terms can give an unrealistic behaviour at high pressure.
Fig 5.
Williamsburg and JWL adiabat (Square indicates experimental data.

Reproduced from Byers et al. [9])
119
CONCLUSION
Incident overpressures and impulses were calculated for a surrogate mine using a range
of cell sizes and two different equations of state for the explosive products. Simulations led
to the conclusion that the ideal gas equation of state is not suitable for modelling an
explosion at close range. The JWL equation of state, although superior in some respects,
also fails to give satisfactory results in the whole range of interest of this study.
Further numerical and experimental investigations are planned in order to study close-in
explosions and the interaction of blast with the surrounding soil. These investigations should
reveal the influence of soil and blast loading on target structures from a detonating mine.
Acknowledgment:
The authors wish to acknowledge technical support from Prof. Braithwaite. They also
wish to thank to Cranfield University and Mr. Machen of Alvis Vickers Plc., UK for
financially supporting this work.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
J. HENRYCH: The dynamics of explosion and its use, Prague: Elsevier Scientific
Publishing Company, 1979.
W. FICKET and C. W. DAVIS: Detonation, University of California Press, 1979.
CH. L. MADER: Numerical modeling of explosives and propellants, 2nd ed., CRC Press
LLC, p. 256, 1998.
O. SCHLEIN and A. LAGALE: Global landmine crisis: The problem,
http://www.landmines.org/GlobalCrisis/TheProblem/TheProblem-all.htm, 2000.
Janes mines and mine clearance 1997-98, Coulsdon, UK, 1997.
P. GUDGIN: Armour 2000, London, 1990.
AUTODYN-2D User documentation, Century Dynamics, UK, 1998.
W. K. E. HUNTINGTON-THRESHER and I. G. CULLIS: TNT blast scaling for small charges,
19th International Symposium of Ballistics, Interlaken, p.647-654, 2001
W. B. BYERS and M. BRAITHWAITE: Williamsburg equation of state for detonation
product fluid, High Pressure Science and Technology, AIP Press, New York, p.73-73,
1994.
CONWEP, US Army Engineer Waterways Experiment Station, Vicksburg, December
1991.
120
DEGRADATION, REARRANGEMENT AND REDUCTIVE

ACETYLATION OF HEXABENZYLHEXAAZAISOWURZITANE
Levi Gottlieb and Gady Korogodsky
Rafael, Department M1,
POB 2250 Haifa 31021, Israel
Abstract:
In the course of searching for an efficient route for reductive acylation of 1, several
degradation products were characterized. For instance we have found that refluxing 1
in a solvent of the type R'COX in non-reductive conditions, a semicrystaline,
functionalized, isomeric product 2 is formed. The new isomer has lost its high symmetry
and turned to be highly complicated for complete NMR assignment. However, synthesis
of para-Bromo derivative afforded crystals suitable for for X-ray analysis. X-ray
crystallography showed that the new HBIW isomer had structure 2.
Keyword:
1.
Polyaza cage compounds, hexabenzylhexaazaisowurtzitane
INTRODUCTION
Hexanitrohexaazaisowurzitane (HNIW) 1 is one of the most powerful energetic

materials known to date [1]. Direct nitration of the parent cage compound
hexabenzylhexaazaisowurzitane (HBIW) 2 is impossible [2], since the nitration of one of the
cage's carbon-nitrogen bonds is more fecile than the nitration of N-benzyl bond [3]. As
a result a different route has to be adopted [4], stepwise reductive debezylation/acylation
produces a polyamide which is now suitable for nitrolysis under forcing conditions (Scheme
1).
Despite intensive recent research, little is known on the chemistry of poly aza cage
compounds and in particular on the chemistry of HBIW [5]. For the best of our knowledge,
except to the well explored reductive debenzylation reaction the only known type of reaction
of this system is the oxidative cleavage of the strained carbon carbon C1-C7 bond using
different oxidation reagents [6]. Our study of catalytic hydrogenolysis/acylation of HBIW
resulted in several surprises, which we wish to report in this paper.
121
1
Bn
Bn
N
Bn
NO2
Bn
Bn
O2N
O2N
O2N
Bn
NO2
2
HBIW
NO2
NO2
HNIW
Debenzylation
Acetilation
NO2
O
O
N
Me
Me
Me
O
N
Me
X = NO, CHO, H
Fig 1.
2.
Scheme 1
RESULT AND DISCUSSION
Catalytic hydrogenolysis of 2 in acetic anhydride resulted in a several products. As

expected mono, di and tri acetyl substituted isowurzitanes 3 were formed albeit in relatively
low yield and were always accoupanied by a substantial amount of N-benzylacetamide
(Scheme 2). From the reaction mixture triacetyl tribenzyl isowurzitane was isolated and
characterized by 1H and 13C NMR spectroscopy. Hindered rotation around the amide bond
lead to complication of NMR spectra (Fig. 3) and the final assignment of the structure
become difficult. Even heating of the sample to 150 oC in bromobenzene did not result in
complete coalescence of the acetyl group signals even though it simplified significanlly the
spectra. Finally X-ray crystallography was used for determination the exact structure of 3.
(The ORTEP drawing of triacetyl isowurzitane is shown in Figure 4). Moreover, among the
variety of unidentified highly polar products we managed to isolate 1,2-di-[Nbenzylacetylamino]ethylene 4. (The ORTEP drawing of 4 is shown in Figure 5).
122
R3 N
R4 N
Ac2O/H2
HBIW
N R1
N R2
Pd/C
Bn
Fig 2.
Bn
Ac
+
4
N
Bn
Bn
Additional
products
Ac
Scheme 2
Fig 3.
The 1H NMR spectra of 3.

A) The spectra at room temperature
B) The spectra at 150 oC.
Fig 4.
The ORTEP drawing of triacetyl isowurzitane 3. The unit cell contains a

molecule of acetonitrile (a solvent used in the crystallization) and two
molecules of 3.
123
Fig 5.
The ORTEP drawing of 4.
In the course of searching for non acidic, mild conditions for reductive acylation of 2,
we tested the use of ethylformate as the acylating agent. Surprisingly we have found that
catalytic hydrogenolysis in presence of ethylformate, HBIW undergoes a reaction to give a
new cage compound but without losing its benzyl groups. Further investigation of this new
reaction has shown that refluxing 2 in ethylformate in a non reductive atmosphere, a
semicrystaline, functionalized (see Fig 4), isomeric product 5 is formed (Scheme 3). The
new isomer has lost its high symmetry and turned to be highly complicated for complete
NMR assignment. Attempts to crystallize the product in order to assign its structure were
unsuccessful. At this point we decide to synthesize the previously unknown p-bromo
derivatives of HBIW. It was hoped that the compound will undergo a similar reaction as
HBIW and give rise to a crystalline material. Thus, synthesis of para-bromo substituted
HBIW and rearrangement afforded a crystalline material suitable for X-ray analysis.
Spectroscopic data, has shown this product to be identical to parent unsubstituted system. Xray crystallography showed that the new HBIW isomer had structure 5 (Figure 5).
N
R
N
N
R
R
NHR
EtOCHO
N
N
R
R = Ph, p-Bromophenyl
Fig 6.
Scheme 3
124
17
16
181 0
187 6
15
14
195 0
13
12
329 6
%Tran smittance
11
10
9
8
7
92 9
289 6
5
306 1
4
3
134 1
284 7
302 7
149 4
1
4 000
116 8 102 8
163 9
3 500
3 000
Fig 7.
The IR spectra of 5.
Fig 8.
The ORTEP drawing of 5.
2 500
2 000
Waven umb ers (cm-1 )
1 500
145 3
1 000
5 00
One can speculate that proximity effect as a result of the cage structure and relief of
strain[7] play an important role in this internal oxidation-reduction rearrangement, and it
undergoes by the mechanism shown in Scheme 4. Thus, we believe that the first step of the
reaction is formylation of one of the nitrogens, which followed by ring contraction of five
membered ring with simultaneous elimination of a proton to obtain a four membered ring.
The next step is the cleavage of carbon nitrogen bond which after hydride transfer and
elimination of the formyl gives the product 5.
125
NHR
EtOCHO
reflax 3 days
N
R
N
R
- EtOCHO
HN
OEt
N
EtOCHO
H
CHO
N
CHO
- EtOH
R
CHO
Hydride Transfer
HN
N
H
N
R
R
CHO
N
R
R
CHO
N
R
Fig 9.
3.
R
CHO
N
R
Scheme 4
CONCLUSIONS
A new and interesting rearrangement of the polyaza cage compound HBIW was found.
The reaction opens new opportunities for energetic materials based on this skeleton.
126
REFERENCES
[1] R. Meyer, J. Khler, A. Homburg: Explosives, 5th Ed., Wiley, New York 2002.
[2] A. T. Nielsen: Polycyclic Amine Chemistry, (Chemistry Of Energetic Materials, G.A. Olah,
D.R. Squire Eds.), Academic Press, Inc., San Diego, 1991.
[3] A. J. Bellamy: Tetrahedron. 1995, 51, 4711-4722.
[4] T. Komada: Haxaazaisowurtzitane derivatives and process for producing the same, European
patent 753519A1.
[5] D. A. Levens, C. D. Bedford, S. J. Stoats: Propellants, Explosives and Pyrotechnics, 1983, 8,
74.
[6] K. H. Chung, S. H. Chun: Bull Korean Chem. Soc., 1997, 18, 1317-1318.
[7] A T. Nielson, A. P. Chafin, S. L. Christian, D. W. Moor, M. P. Nadler, R. A. Nissan, D. J.
Vanderah, Tetrahedron. 1998, 54, 11793-11812.
[8] R. Wardle, W. Edwards Inter. Patent WO 97/20785, 1997.
[9] T. M. Klaptke, B. Krumm, H. Piotrowski, K. Polborn, G. Holl Chem. Eur. J. 2003, 9, 687-694.
[10] X. P. Guan, H. Yan, J. G. Sun, Y. Z. Yu, Molecules 1999, 4, 69-72.
[11] X. P. Guan, H. Yan, J. G. Sun, Y. Z. Yu, Chin. Chem. Letters 1996, 7, 511-512.
[12] W. G. Qiu, S. S. Chen, Y. Z. Yu, Chin. J. Chem., 1999, 17, 554-556.
[13] A. Hassner, S. Naidorf-Meir, H. E. Gottlieb J. Org. Chem., 1993, 58, 5699-5701.
127
CHEMICAL AND THERMAL STABILITY OF FLEXIBLE PBXs

Martina Chovancov, Peter Oko,
Jozef Lopch, Milo Lazar and Albeta Pechov
VTS Zhorie (Military Technical and Testing Institute),
905 24 Senica, SK
Abstract:
In this paper, the flexible PBXs chemical and thermal stability are observed before,
during and after their ageing. The following three flexible PBXs were observed:
explosive E (RDX/polymer binder), explosive H (RDX/PENT/polymer binder) and
explosive N (PETN/polymer binder). All explosives have been introduced into service of
the Slovak Armed Forces.
Keyword:
1.
PBX, ageing, chemical a thermal stability, NBR, DOF.
INTRODUCTION
The artificial ageing is a very significant and effective way to foresee the life cycle of
explosives. The analysis of explosive properties before, during and after artificial ageing also
significantly contributes to determination of the chemical, thermal and physical stability and
also to handling safety. The scope has been to investigate all the flexible plastic explosives
introduced into Slovak Armed Forces armament.
Flexible plastic explosives FPE play an important role in the military activity and
industrial demolition technique due to their easy formability, high efficiency and high
resistance to outer stimulus as well. In modern flexible plastic explosives, there are brisance
crystal explosive as RDX or PETN dispersed in the plastic binder that is usually explosionproof and serves as a carrier of the crystal explosive and gives the mass a plastic nature.
There are 3 types of FPE in the Slovak armed forces armament.
1. Flexible plastic explosive N (explosive N) based on PETN and butadieneacrylonitrile rubber (NBR) with dioctylphthalate (DOF),
2. Flexible plastic explosive H (explosive H) based on PETN, RDX, NBR
and DOF,
3. Flexible plastic explosive E (explosive E) based on RDX, NBR and DOF.
Each type of the explosive above mentioned is marked by 2,3-dimehtyl-2,3dinitrobutane (DMDNB) contents of which should be min.0,1% and both explosives E and H
contain aluminium.
Plastic explosives were exposed to the artificial ageing and to the chemical and thermal
stability tests before, during and after ageing. Test results show a notable chemical and
thermal stability at each type of the explosives all but one. Chemical stability test results of
E-explosive have suggested its instability already during early ageing tests.
128
Resulting from the part mentioned above the E and N explosives was subjected to the
same artificial ageing and tests as well. Further compatibility tests was done relating to the
RDX, PETN compatibility with the rubber binder.
2.
EXPERIMENTAL
2.1
Artificial ageing and stability test
Artificial ageing conditions was same for all the explosive samples: 65C in closed glass
tubes. Explosive samples was taken each month and exposed to the following stability tests:
1. Vacuum chemical stability test by use STABIL 15-Ex at 110C, 140C, 20 hours
according to STN 26910
2. Differential thermal analysis by use DTA 550-Ex at heating speed 5C/min
according to STANAG 4515
3. Ignition temperature determination at speed 5C/min according to STANAG
4491 (B1).
E explosive has appeared the chemical stability already in the first month of ageing by
using the vacuum test on the acceptability limit value while stability has decreased each
month. After 8 months, the rate of gas release was 6 times greater than the permissible value
mentioned in the technical conditions.
Table 1. Results of the stability tests
Test
Sample
Ageing
(month)
VST
DTA
(ml/g/20)
Start (C)
E
Before a.
0,9867
191
1
1,9475
192
2
2,4677
185
3
5,5104
190
4
6,3765
190
5
5,3072
193
6
7,4825
185
7
8,9079
190
8
12,5997
184
H
Before a.
0,4035
145
2
0,6441
142
6
0,7122
118
8
0,7436
122
N
Before a.
0,5671
149
2
0,7760
145
6
0,7079
146
8
0,8730
142
VST - Vacuum Stability Test, TI - Temperature of Ignition
TI
Max (C)
215
211
209
212
209
211
208
213
212
195
188
179
182
163
163
163
161
(C)
213
204,5
206
205
202
203
202
203
204
204
187
169
168
183
187
161
160
129
2.2
Compatibility test
Owing to the investigation of possible reasons of the E-plastic explosive instability

during artificial ageing there was carried out the mutual compatibility test of an individual
component in the explosive. Compatibility tests were done by using the vacuum stability test
at first at working temperature (140C) at which the E-explosive test is to be performed by
specification. At this relatively high temperature hexogene was compatible with the DOFplasticizer and also with aluminium however with the NBR rubber it was highly
incompatible. We had assumed that at lower working temperatures, RDX was compatible
with rubber. Test was done at 120C, 110C, 100C and 90C too. At the first three
temperatures the results was similar to those of at 140C, RDX- rubber, incompatible, RDXother components, compatible. There was found compatibility of the mixture of hexogene
rubber as late as at 90C. We were interested in the behaviour of the mixture of PETN-NBR
and PETN-RDX-NBR at the compatibility test. The test was done at 90C, 100C a 110C.
At the lowest temperature all the components considered was compatible each other. At
100C the mixture of PETN-NBR wasnt compatible, at 110C was incompatible not only
the mixture of RDX-NBR, but also the mixture of PETN-RDX-NBR. Some results of the
compatibility test, which was done by the vacuum stability test, are listed below in the table
2. In table 3, compatibility test results are listed by DTA and ignition temperature
determination as well.
Table 2. Results of the compatibility tests by vacuum stability test

Sample
RDX
NBR
Mixture
RDX+NBR
RDX
NBR
Mixture
RDX+NBR
PETN
Mixture
PETN+NBR
Mixture
RDX+PETN+
+NBR
DOF
Mixture
RDX+DOF
Mixture
RDX+DOF+
+NBR
RDX
NBR
Mixture
Temperature
C
90
90
Vacuum stability test Sum

(ml/g/20)
0,0467
0,0258
90
100
100
0,0633
0,0365
0,0617
Comparison
Evaluation
0,0725 Mixt < Sum
Compatibility
Mixt > Sum

0,0982 about 31%
Incompatibility
Mixt > Sum

0,0972 about 42%
Incompatibility
100
100
0,1290
0,0355
100
0,1378
100
100
0,1079
0,0125
0,1337 Mixt < Sum
Compatibility
100
0,0317
0,049
Mixt < Sum
Compatibility
100
110
110
0,0626
0,0502
0,0660
0,1107 Mixt < Sum
Compatibility
Mixt > Sum
130
Sample
Temperature
C
110
110
RDX+NBR
PETN
Mixture
PETN+NBR 110
Mixture
RDX+PETN+
+NBR
110
Vacuum stability test Sum

Comparison
(ml/g/20)
0,1826
0,1162 about 57%
0,0925
Mixt > Sum
0,7781
0,1585 about 490%
Mixt > Sum
0,2087 about 400%
0,8344
Evaluation
Incompatibility
Incompatibility
Incompatibility
Table 3. Results of the compatibility tests by DTA and Temperature of ignition

Sample
Melting point
C
250RDX
Mixt.RDX+NBR
PETN
Mixt. PETN+NBR
Mixt. RDX+PETN+
+NBR
203
142
140
Start of decomp. Maximum

C
decomp.
C
211
218
188
207
157
188
149
173
137
126
3.
177
of Temperature
of ignition
C
218
198
192
164
165
DISCUSSION
The results obtained proved E-explosive to be instable after 2 months ageing, while after
the first month period it was on the limit of stability but as the time passed the stability was
getting worse during further period of months. However this instability was significant
enough from the point of view of chemical stability. On the other side thermal stability was
kept constant during the whole period of artificial ageing, DTA and ignition temperature
results prove only an insignificant drop of the ignition temperature in comparison of the
sample unloaded which shows that there is no conflict with related technical condition
requirements. Start temperatures and maximum temperatures of decomposition, by DTA,
remained almost unchanged when compared with the sample unloaded.
The results obtained during H, N explosives measuring proved a contrary tendency.
Chemical stability lightly decreased during ageing, but the thermal stability dropped
considerably what is proved by either decrease of ignition temperature or start temperatures
and maximum temperatures of decomposition obtained by DTA. This partial decrease of
thermal stability could have been caused by the presence of PETN that is less stable than
RDX and at the greater temperatures it becomes early the subject to exothermic
decomposition. However, the H and N explosives, excepting the E, maintained their stability
during ageing. Chemical and thermal stability of the explosives mentioned during ageing
process are shown in both figure 1 and 2.
Different behaviour of such structural similar explosives became amazing. From the
point of view of using the same polymer binder at each type of explosive and also RDX or
PETN-RDX ultimate base we expected very similar results related to their chemical and
thermal stability. However the one thing is amazing namely that such highly instable
explosive is just type E made in 2000 and stability of explosives N and H made in 1980
and 1990 kept unchanged. It is out of question that during artificial ageing the stability
alteration is considerably affected by the explosive composition and structure.
131
We can rule out a bad effect of the DMDNB marker on the explosive chemical stability.
Nevertheless the marker is involved in each type of explosive and it is known that DMDNB
is a very volatile substance. Thus, during ageing process, the marker contents declines in the
explosive. 70% depletion of DMDNB was found in type E after 3 months ageing, but the gas
release from the sample was rapidly increasing even after a few ageing months passed when
the explosive had not held any DMDNB.
Polymer binder is the component, which considerably affects the reological, mechanical
and to a certain extent it affects physical properties of explosives too. Whether physicalmechanical properties are significantly influenced or not, relating to the stability fade, we
attempted to find out by using the compatibility tests of the ultimate bases as (RDX, PETN)
with both NBR and DOF plasticizer. The results obtained shows us that DOF is clearly
compatible with all the other explosive components even with aluminium involved in types
of E and H. What is amazing is that the incompatibility NBR with RDX and PETN already
at 100C. If this is the reason of instability of E-explosive then we want to know how it is
possible that the H, N-explosives were not affected so badly. In case of the polymer binder,
an important role is played probably not only by a sort of the rubber used, but also
polymerisation rate, individual synthetic rubber elements percentage (what relates to NBR it
is acrylonitrile contents in NBR) and needless to say some role plays also the ratio of NBR
to plasticizer. From this point of view we cannot imply that explosives E, H and N, despite
each of them contains NBR, each one includes just the same polymer binder. The following
difference, relating to the polymer binder, was found out from the available sources:
In the explosives H, N there was used NBR containing about 30 % acrylonitrile
(thereinafter AN), in explosive E containing approximately 33 % AN.
In the polymer binder explosives H and N there is ratio NBR: DOF 20:80 and in
the polymer binder explosive E the ratio is NBR: DOF 26:74, in the explosive N it was
a plasticizer DBF (dibutylphthalate) used.
Butadiene-acrylonitrile caoutchuc belongs to the oil proof stuff and it is an interpolymer
butadiene - acrylonitrile that contents goes to 18 50 %. The whole product properties are
depend on the AN contents. By an increasing AN contents: the glass transition temperature
and hardness go up, high temperature resistance goes up and the elasticity and frost
resistance decreases. It follows probably that even though the AN contents differences are
insignificant these ones may affect various properties. NBR vary in the AN contents,
polymerisation temperature (hot or cold types), dyeing (or non) stabiliser presence and also
viscosity of Mooney. Further differences can appear in the contents of volatile substances or
in the total and water-soluble ash. Following this there are many different types of NBR and
any of them may not be suitable into an individual plastic explosive. Different responds of
the investigated explosive to the artificial ageing might be caused by different quality of the
NBR applied to. Following this, different reactions of the investigated explosives on the
ageing process have been caused by the polymer binder quality. This has been confirmed
even by the compatibility test. From the availability reasons the tests was only carried out
using NBR that has been used for making the explosive E. The fact to a certain extent proves
a discrepancy between H and N explosives stability kept unchanged and NBR
incompatibility with their ultimate bases. Therefore the NBR-PETN compatibility results
should be considered as informative.
The figures 3 5 present the compatibility of the components investigated by using
DTA and Temperature of ignition determination. In the pictures it is seen that thermal
characteristics of the mixtures tend to the lower values that has proved incompatibility of
individual components.
132
4.
CONCLUSIONS
Following the results obtained we declare the E-explosive instability was caused by
incompatibility between RDX and NBR. The type of NBR that was used and also its
contents in polymer binder probably affects not only the properties of polymer binder but
even an entire plastic explosive including its stability. We can presume that when using other
type of NBR (other maker maybe) it might be achieved more favourable results at
compatibility tests et inde a better stability of the plastic explosive. Addition suitable
stabiliser to polymer binder should solve e-explosive instability concern. At higher
temperature, NBR maybe becomes the object of chemical degradation processes, which
would be successfully attenuated by an extra stabiliser.
In the next period of time, we need to investigate the DMDNB contents in the explosive
considered during their ageing and then it could be interesting to find out any changes
related to the plasticizer contents in PB during the ageing. Compatibility RDX -PETN
scanning will continue from now on but individual components of the explosives will be
separately subjected to the artificial ageing procedure and the compatibility tests will be
doing regularly.
Acknowledgement
Authors would like to thank Mrs. Bohumila Stankov from VTS Zhorie for providing
some VST analyses.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
J. MLEZIVA: Polymery vroba, struktura, vlastnosti a pouit, Praha, 62, 1993

V. DUCHEK: Polymery vroba, vlastnosti, spracovn a pouit, Praha, 148, 1995
I. FRANTA a kol.: Gumrensk suroviny, Praha,170-178, 1979
S. BREBERA: Vojensk trhaviny II, Trhaviny rzn konzistence, Univerzita Pardubice,
134-141, 2002
M. CHOVANCOVA P. OCKO, R. EVK, L. AVOJSK, L. LOPCH: Stability investigation of
plastic explosives during their ageing process, Proc. of the 5-th seminar New trends in research
of energetic material University of Pardubice, p.86-91, 2002
STN 26 910
Slovak specification 1080
AOP-7
STANAG 4170, 4515, 4491
133
FIGURES
13
12
H
11
y = 1,2806x + 0,6182
R2 = 0,9137
10
H
9
ml/g/20
8
7
6
5
4
y = 0,0273x + 0,5641
R = 0,3903
y = 0,0122x + 0,6767
R = 0,1179
1
0
0
Time (month)
Fig 1.
Chemical stability of explosives with NBR during their ageing
225
E
H
N
E
H
N
y = -0,8417x + 208,09
2
215
R = 0,4642
Tign (C)
205
195
y = -4,8167x + 200,32
2
R = 0,8106
185
175
y = -3,8333x + 188,44
2
R = 0,8326
165
155
0
Time (month)
Fig 2.
Thermal stability of explosives with NBR during their ageing
134
DTA of mixture RDX-NBR

RDXNBR
RDX
Diference of temperature C
70
60
50
40
30
20
10
0
-10
50
100
150
200
Temperature C
250
300
Meavy
Fig 3.
Compatibility of mixture RDX and NBR by DTA.

DTA of mixture PETN-NBR
80
PETN-NBR
PETN
70
60
50
40
30
20
10
0
-10
60
80
100
120
140
160
Temperature C
180
200
220
Meavy
Fig 4.
Compatibility of mixture PETN and NBR by DTA.
135
DTA of mixture RDX-PETN-NBR

90
RDX-PETN-NBR
PETN
RDX
80
70
60
50
40
30
20
10
0
50
100
150
Temperature C
200
250
Meavy
Fig 5.
Compatibility of mixture RDX, PETN and NBR by DTA.
136
AGEING INFLUENCE ON PLASTIC EXPLOSIVES

ON THE BASE OF RDX
Martina Chovancov, Peter Oko, Jozef Lopch
Rastislav evk and ubo avojsk
VTS Zhorie (Military Technical and Testing Institute),
905 24 Senica, SK
Abstract:
This paper presents results of testing of plastic explosives on base RDX during eight
months ageing. 5 kind of PBX (plastic bonded explosive) with RDX as their base were
investigated.
It was discovered one from them (E) is chemically unstable already after 2 months
of ageing.
The other tested plastic explosives are chemically and heat stable.
Keyword:
1.
PBX, RDX, ageing, chemical and thermal stability
INTRODUCTION
Since Slovak republic is going to come into NATO we have to test the ammunition and
the explosives according to NATO standards. We want to achieve the compatibility with
NATO countries in the investigation of properties of military explosives, because this is the
one from very important areas. The investigation of properties of military explosives during
their artificial ageing belongs in this area.
Every national authority selects singly own ageing conditions and the tests during
artificial ageing according to own reflection. Document AOP-7 (Manual of a Data
Requirements and Tests for the Qualification of Explosive Materials for Military Use )
provides main rules of artificial ageing tests and the methodologies of some NATO
countries. We chose our ageing conditions and the approach of testing of the explosives
during their ageing. We tested chosen approach with the samples of 5 kinds of plastic
explosives (next only PBX), which have one similar mark; their energetic base is RDX. We
were interested in the changes of sample properties during their ageing. We wondered, if
these changes will be analogical for all samples according to their similar base. We wanted
to know when we will be able to see the first changes.
2.
EXPERIMENTAL
2.1
Tested PBX
We used following PBX samples for the tests of artificial ageing:

PBX A
RDX / polymer binder with rubber Dutral + APP (polypropylen)
PBX C RDX / polymer binder with rubber Dutral + APP (polypropylen) + mineral +
ceresine / sillicone oxide
137
PBX E RDX / polymer binder with butadienacrylonitril rubber (next only NBR) +
diocthylphtalate (next only DOF) / Al
PBX F
RDX / polymer binder with rubber Dutral + APP (polypropylene)
PBX G RDX / polymer binder with polyizobutylen + oil Delwac + di(2ethylhexyl)sebacate

All tested PBX contain 2,3-dimethyl-2,3-dinitrobutane (next only DMDNB) as the
marking agent, PBX A, C, E and F minimal 0,1 %, PBX G minimal 0,25 %.
2.2
The tests of artificial ageing
We chose following ageing conditions and the approach of testing according to AOP-7
and our technical possibilities:
Ageing conditions:
65C 1, 2, ..., 8 months in standardized closed test-tube according to STANAG

4527
Approach of testing:
Chemical stability test vacuum stability test (next only VST) by STABIL 16Ex (manufactured by OZM Research), according to Slovak technical standard
(next only STS) 26 910
Thermal stability by Differential thermal analysis (next only DTA) by DTA 550Ex (manufactured by OZM Research), according to Slovak specifications
Temperature of ignition (next only TI) by apparatus manufactured by OZM

Research, according to STANAG
DMDNB content by HPLC apparatus, according to Slovak specifications
All present tests were performed with every PBX sample before ageing, then
after every month of ageing, DMDNB content was performed only with 3 PBX
samples (C, E, F) and only during the first 3 months duo to our technical
problems
Because PBX E shows as very unstable explosive already after 2 months of

ageing, 2 others series of this sample were tested by artificial ageing, so we
tested three series of PBX E: E-1, E-2 and E-3
Testing conditions:
VST temperature 140C, result is rate of evolution of gaseous decomposition

products for 20hours from 1 gram of sample
DTA linear increase of temperature with 5C/min. rate in air atmosphere results
are the start temperature of exothermal decomposition and the maximum
temperature of exothermal decomposition
TI linear increase of temperature with 5C/min. rate in air atmosphere
DMDNB eluent: Methanol / water: 48 / 52, Temperature: 40C, Flow: 1ml /

min., Detection: UV detector, = 205 nm
138
The tables 1 and 2 present the results of chemical and thermal stability of tested PBX
before and during their ageing and the table 3 gives the values of tested characteristics which
are required according to Slovak specifications for individual PBX.
Table 1. Results of the stability tests of explosive E
Test
Sample
E-1
E-1
E-1
E-1
E-1
E-1
E-1
E-1
E-1
E-2
E-2
E-2
E-2
E-3
E-3
E-3
E-3
Ageing
VST
(month)
before a.
1
2
3
4
5
6
7
8
before a.
2
6
8
before a.
2
6
8
(ml/g/20)
0,9867
1,9475
2,4677
5,5104
6,3765
5,3072
7,4825
8,9079
12,5997
0,7457
7,0125
10,6701
13,3586
0,7265
10,1675
14,2210
15,7448
Start
C
191
192
185
190
190
193
185
190
184
195
182
180
185
195
186
186
185
DTA
Maximum
C
215
211
209
212
209
211
208
213
212
214
206
208
210
214
209
209
210
TI
C
213
204
206
205
202
203
202
203
204
213
203
200
202
214
202
200
200
Table 2. Results of the stability tests of other explosives with RDX

Test
Ageing
VST
Sample
(month)
(ml/g/20)
A
A
A
A
A
A
A
A
A
C
C
C
C
C
C
C
C
C
before a.
1
2
3
4
5
6
7
8
before a.
1
2
3
4
5
6
7
8
0,1329
0,1331
0,0664
0,1106
0,1656
0,2176
0,1002
0,1061
0,1054
0,5564
0,5734
0,3831
0,4177
0,5940
0,3745
0,3526
0,4663
0,3676
DTA
Start
C
199
192
194
196
188
187
194
190
192
192
199
190
187
193
192
195
195
194
TI
Maxim
C
217
214
221
216
215
213
215
218
218
217
215
220
217
216
217
217
218
219
209
206
207
208
207
205
207
209
208
208
207
205
208
208
206
207
208
208
139
F
F
F
F
F
F
F
F
F
G
G
G
G
G
G
G
G
G
before a.
1
2
3
4
5
6
7
8
before a.
1
2
3
4
5
6
7
8
0,1145
0,1581
0,0627
0,0666
0,1180
0,1321
0,1008
0,1032
0,1006
0,1079
0,1327
0,0618
0,0778
0,0930
0,0927
0,0943
0,0925
0,0975
197
201
196
187
191
193
195
181
191
199
198
203
201
193
195
193
190
193
218
215
220
216
215
213
216
217
218
215
214
220
215
213
213
214
216
216
210
208
209
208
210
208
209
211
210
207
205
207
206
206
204
204
207
207
Table 3. Explosive values required by Slovak specifications

Sample
TI
( C )
Min. 200
Min. 185
Min. 200
Min. 200
Min. 200
VST
(ml/g/20)
Max. 2
Max. 2
Max. 2
Max. 2
Max. 2
A
C
E
F
G
DMDNB
(%)
Min. 0,1
Min. 0,1
Min. 0,1
Min. 0,1
DMDNB content was investigated only in three samples of PBX (E-1, C, F) during the
first three months. These results are in the table 4.
Table 4. DMDNB contents during ageing
Ageing (month)
Sample
E-1
C
F
3.
before a.
0,180
0,1
0,1
0,132
0,074
0,068
0,082
0,059
0,036
0,042
0,053
0,021
DISCUSSION
The results of tests of all three series of PBX E during their ageing showed this PBX is
chemical unstable. Two from them ( E-2, E-3 ) already after the first month and E-1 after the
second month of ageing. Rapid increasing of rate of gaseous decomposition products
evolution is six to seven times higher than the value in Slovak specifications. This increasing
is evident also from figure 1. The reason of this instability according to articles (1, 4) is
incompatibility RDX with NBR. We can retire nearly with certitude negative influence of
140
DOF and aluminum. DMDNB also is not responsible for rapid decreasing of chemical
stability. The investigate of DMDNB content showed PBX E-1 contained after three months
o ageing only some centenas % DMDNB. In spite of it the rate of gaseous decomposition
products evolution rapid increased also after following months when DMDNB probably
vaporized entirely from PBX. Since 1997 we have investigated in the long term the influence
of DMDNB for the stability of PBX and so far any negative influence was not achieved. In
area of thermal stability the start temperature of exothermic decomposition and TI of all
three series PBX E was moved towards lower temperatures, in sample E-2 and E-3 it was till
on the limit of stability. The decreasing of TI of PBX E against others PBX is showed on
figure 3. The decreasing of maximum temperature of exothermal decomposition of PBX E is
showed on figure 4. Figure 5 presents PBX E as least thermal stable PBX from all tested
PBX.
Others tested PBX accounted during their ageing the chemical and thermal stability.
Their behaviour was different from the behavior of PBX E. Despite of retention of their
stability during full eight months we can notice according to table 2 some changes especially
in VST results. The fact is really interesting in PBX F a G was remarked respectable
increasing of rate of gaseous decomposition products evolution after the first month and in
all four PBX was remarked surprise decreasing of this rate after second month. The values of
rate were in PBX A, C, F and G so low that their stability stayed untouched also after these
changes, it is showed on figure 3. These changes were caused probably by DMDNB, which
was after the first month contained in PBX even in sufficient quantity to occasion the
increasing of rate. Because DMDNB content in PBX is enough low already before ageing,
the changes of rate were compared to PBX E really very small. During following month
DMDNB content in PBX receded enough rapid so the rate after second month could be
lower. It has to remind these changes due to DMDNB are so remote that they cant have the
influence on chemical stability of PBX. Basically chemical stability of tested PBX (except
E) remained unchanged and on figures 3 and 5 we can see thermal stability is observed too.
4.
CONCLUSION
The results of ageing tests of PBX on base RDX during 8 month showed different
Behavior of PBX E in consequence of its chemical instability. They also showed some
analogy in the behavior of PBX A, C, F and G. This analogy relates with the content of
similar rubber in PBX A, C and F. PBX A, C, F and G observed their chemical and thermal
stability so we can actually expect they are going to stay stable minimal 10 till 15 years.
Next we can suppose pursuant to our results the evolution of gaseous decomposition
products and exothermal decomposition of PBX are hard influenced by their polymer binder,
by products of decomposition of polymer binder and of others ingredients in PBX.
Next future we have to investigate also others properties of PBX during their ageing, e.g.
plasticity, mechanical and sensitive properties. Equally it is need to test in detail the
decrement of DMDNB in PBX during ageing.
Acknowledgement
Authors would like to thank Mrs. Bohumila Stankov from VTS Zhorie for providing
some VST analyses.
141
REFERENCES
[1]
M. CHOVANCOVA P. OCKO, R. EVK, L. AVOJSK, L. LOPCH:
Stability investigation of
plastic explosives during their ageing process, Proc. of the 5-th seminar New trends in research
of energetic material University of Pardubice, p.86-91, 2002
[2] M. KRUPKA: Devicesand equipment for testing of energetic materials, Proc. of the 4-th seminar
New trends in research of energetic material University of Pardubice, p.222-227, 2001
[3] M. CHOVANCOVA P. OCKO, L. AVOJSK, L. LOPCH: Stability investigation of marked plastic
explosives, Proc. of the 4-th seminar New trends in research of energetic material University
of Pardubice, p.144-150, 2001
[4] M. CHOVANCOVA P. OCKO, L. LOPCH, M. LAZAR, A. PECHOVA: Chemical and thermal stability
of flexible PBXs, Proc. of the 6-th seminar New trends in research of energetic material
University of Pardubice, 2003
[5] M. HANUS,: Zklady analytick chemie vbuin, Skript pro inenrske stadium oboru Terie a
technolgie vbuin, VTVM Slavin, 1999
[6] STN 26910
[7] Slovak specications 1080-00
[8] Slovak specications 1060-98
[9] AOP-7
[10] STANAG 4170, 4515, 4491, 4527
142
FIGURES
14
12
A
[E-1]
y = 1,3724x - 0,3614
10
R = 0,8236
E-1
F
8
ml/g/20
G
[A]
[C]
[E-1]
4
2
[A]
y = 0,0563x - 0,0045
[C]
y = -0,0197x + 0,5327
R2 = 0,3068
R2 = 0,3102
[F]
y = -0,0013x + 0,1115
[G]
y = -0,0014x + 0,1
R2 = 0,0145
R2 = 0,0376
[F]
[G]
0
0
-2
Time
Fig 1.
Chemical stability of explosives with RDX during their ageing
1
0,9
0,8
0,7
ml/g/20
0,6
A
C
0,5
F
G
0,4
0,3
0,2
0,1
0
0
Time
Fig 2.
Process of chemical stability of PPX during ageing
143
214
212
210
TI (C)
208
A
C
E-1
206
F
G
204
202
200
0
Time (month)
Fig 3.
Process of thermal stability of PBX with RDX during ageing
0,2
y = -0,0464x + 0,1786
R2 = 0,9978
C
E
0,16
F
E
%DMDNB
F
0,12
y = -0,0156x + 0,0949
R2 = 0,9239
0,08
y = -0,0269x + 0,0966
R2 = 0,9766
0,04
0
0
0,5
1,5
2,5
3,5
Time (month)
Fig 4.
DMDNB content in PBX with RDX during ageing
144
DTA of sample E-1

E-1before a.
E-1after2m.
E-1after4m.
E-1after6m.
E-1after8m.
Dif erence of temperature C
70
60
50
40
30
20
10
0
195
200
205
210
215
Temperature C
220
225
230
Meavy
Fig 5.
Thermal stability of E explosive by DTA

DTA of PBX after ageing
65
A-after8m.
C-after8m.
E-1-after8m.
F-after8m.
G-after8m.
60
55
50
45
40
35
30
25
20
15
10
5
0
190
195
200
205
210
215
220
Temperature C
225
230
235
240
Meavy
Fig 6.
Thermal stability of PBXs by DTA
145
COORDINATION COMPLEXES
AS INORGANIC PRIMARY EXPLOSIVES
M.A. Ilyushin, I.V. Tselinsky, I.A. Ugryumov,
A.Yu. Zhilin and A.S. Kozlov
Saint-Petersburg State Institute of Technology, Russia,
Abstract:
The report deals with the synthesis of coordination compounds in the series of
perchlorates of d-metals containing polynitrogen heterocyclic ligands. Considerable
attantion is also devoted to the characterization of the compounds obtained and their
applications in safe electric and laser detonators.
Keyword:
1.
Coordination Complexes; Inorganic explosives; Initiation systems; Laser,

Pulse.
INTRODUCTION
Laser initiation is a new method for activating the charges of energetic materials (EM).
This method ensures high safety of explosive works because it provides a high level
of isolation of the light detonator from wrong impulses (Fig. 1)
1
3
2
Fig 1.
Schematic of the optical initiation system: 1 pulsed laser,

2 optical fiber (light guide), 3 optical blasting cap.
ETC
Light detonators are non-susceptible to electromagnetic stimuli and discharges of static
electricity.
Laser initiation can be introduced in to many explosive technologies which require
individual approach in developing explosive systems:
explosive welding, stamping, compacting, synthesis of new materials;
explosive mining works, hazardous in respect to gas and dust;
automated technologies with a pulsed periodic input of materials;
explosive technologies of a single action used, for example, in the systems

of pyroautomatics of space shuttles or in protection systems for dangerous
technological processes;
explosive technologies used in perforation of deep bore-holes;
technologies of extreme operating conditions using high electromagnetic fields.
146
Light blasting caps or film charges of light sensitive EM can be used for such
technologies. Initiation of these charges may be effected by laser pulses transmitted through
optical fibers (light guides) or directly in air.
Coordination complexes of d-metals represent an interesting class of explosives for laser
detonators wich are more safe than traditional devices1. Co(III) tetraamminates containing 5R-tetrazoles as ligands are unique explosives that combine properties of primary and high
explosives. As the result, investigations in sensitivity of energetic complex salts to laser pulse
radiation are of great theoretical and general interest. Energetic coordination compounds
possessing high sensitivity to laser pulse are used in medicine, processing of metal surfaces,
generation of profiled shock waves, mining and airspace technologies2, 3. They have been
used both in military and industrial safe detonators4, 5
2.
EXPERIMENTAL
The coordination complexes were synthesized by the reaction of perchlorates of dmetals with substituted tetrazoles, 1,2,4-triazoles and pyrazoles in diluted acid media at
elevated temperatures.
Their composition and structures were proved by elemental analyses; IR-, UV and NMR
H1 spectra. Detonation velocities for the salts were determined by ionization technique. The
derivatograph MOM "E.Paulic, J.Paulic, L.Erdey" (Hungary) was used to perform
differential thermal analysis of the salts. The heating rates of the samples amounted to 5
K/min, masses of the salts were 10-20 mg. IR spectra of the complexes were recorded in
Nujol on a "BrukerIFS-113" spectrometer (Germany). UV spectra of the complexes were
recorded on a SF-46 spectrometer (USSR) and on a Lisker spectrometer6. NMR H1 spectra
of the complexes were recorded on a Perkin-Elmer R-12 (UK) spectrometer (60 MHz), the
samples of the complex salts were dissolved in DMSO-d6, HMDS was used as standard.
3.
Tetraamine-cis-bis(5-nitro-2-tetrazolate-N2)cobalt(III) perchlorate (BNCP) (I) was

clained in 1986 as one of the most efficient primary explosives. It was used in laser
detonators for the system of pyroautomatics of rocket vehiles in USA
N
H 3N
NH3 N
N
Co
H 3N
NH3
+
N
C
NO2
C lO 4
N
N
N
NO2
(I)
BNCP charges in blasting caps feature fast transition from ignition to detonation playing
the role of both primary and secondary explosives (HE). The density of BNCP monocrystals
is 2,05 g/cm3, the calculated detonation velocity at density 1,97 g/cm3 approaches 8,1 km/s,
the temperature of the onset of intensive decomposition is 269 0, the minimum initiation
charge for RDX in blasting cap 8 is about 50 mg, the time of transition from ignition to
147
detonation is 10 s, the heat of thermal decomposition equals to 3319 J/g. We synthesized

BNCP complex by the following reaction:
N
2Na+ N
NO2
+ [Co(NH3)4CO3]+
HClO 4 / H2O
N
80- 90oC
2
N
Co(NH3)4 N
ClO4
NO2
BNCP
Compound (I) was obtained in 6070% yield. To prepare complex (I), the sodium salt of
5-nitrotetrazole synthesied by the Sandmeyer reaction from 5-aminotetrazole was used
H N
NaNO2
H2SO4
N
N
NH2 Na CO
2
3
N
Na+ N
N
N
NO2
. 4H O
2
The quality of BNCP depends on the purity of 5-nitrotetrazole sodium salt which can
contain sodium nitrite (NaNO2) as an admixture. In case of the incomplete removal of NO2it may be present as a ligand competing with the nitrotetrazolato-ion. Hence special attention
was payd to elaborate the purification procedure and, as a result, the purity of the sodium salt
of 5-nitrotetrazole exceeded 98%.
The effective means of thermal activation of chemical reactions is microwave heating.
Synthesis of (I) under conditions of microwave heating at reflux (102 0) during 1 hour
resulted in 80% yield of the product. So this method allows to obtain compound (I) with a
higher yield during a shorter period of time.
To investigate the properties of other tetraaminates complexes of cobalt(III)
perchlorates, 5-substituted and 1,5-disubstituted tetrazoles were used as ligands:
[Co(NH3)4(H2O)2](ClO4)3 + n Tz
HClO4/H2O
pH<2
95 0C
4
[Co(NH3)4(Tz)n](ClO4)m + 2 H2O
(XXIII-XXIX)
where Tz is a sabstituted tetrazole.
148
Characteristics of metallocomplexes are given in Table.

Table 1.
Characteristics of tetraamminates of cobalt(III) perchlorates.

N
N
Co(NH3)4
R 2
Co(NH3)4 N
ClO4
(I-III)
Entry
I
II
III
IV
V
VI
VII
N NO2
Co(NH3)4
ClO4
(IV)
R, R1, R2
R=NO2
R=CH3
R=H
R1=H, R2=NH2
R1=R2=NH2
R1=CH3, R2=NH2
*-experimental
N
(ClO4)3
2
(V-VII)
Tdecomp.
C
kJ/mol
calc.
g/cm3
D calc.
km/s
234
252
239
238
238
233
234
212,2
204,7
228,6
231,6
293,6
204,6
218,3
2.03*
1,75
1,86
1,90
1,85
1.90*
8,1
6,8
6,9
7,1
7,3
7,5
data.
The thermal decomposition behavior of these salts is a problem of general importance.

The thermal decomposition of Co (III) tetraamminates with substituted tetrazoles as ligands
is found to be subject to metal ion catalysis. Derivatographic investigations show that the
thermal decomposition of perchlorate complexes (II-VI) occurs in several steps. The
decomposition starts at temperatures exceeding 230 0C. The number of macrokinetic steps of
decomposition depends on the nature of substituents in the tetrazole ring. The values of
are given in Table for the first step of the decomposition.
The minimum charge testing in blasting cap 8 showed that all of the above complexes
possess initiating ability. Also, compounds (I), (III), (V), (VI) display certain sensitivity to
laser irrariation (pulse time ~2 ms, E~1,5 J, ~1,06 m, ray diameter ~0,5 mm). Complexes
(III) and (V) are ignited by the laser beam, whereas tre other complexes explode.
We investigated absorbtion spectra of pressed charges of metalcomplexes I-III and VI in
the range of wavelengths 600 1200 nm (see Fig.2 ). These wavelenghth values are usual
for pulse lasers.
149
Fig 2.
Absorbtion spectra of pressed charges of metalcomplexes I-III and VI on

copper plates(1 VI, 2 - II, 3 III, 4 I), absorbtion coefficient
Fig. 2 demonstrated that all metalcomplexes have almost identical absorbtion spectra.
There are not any dependence of sensitivity to laser irradiation of the metalcomplexes from
their optical spectra. We believe the problem of relation between sensitivity to laser
irradiation of the metalcomplexes and their structure needs further investigations.
Complex perchlorates of hydrazinoazoles can be regarded as a novel class of compounds
possessing high sensitivity to laser irradiation. This assertion is based on the properties of
hydrazinoazoles, viz., the azole ring is featured by high positive enthalpy of formation,
besides the hydrazine moiety has low ionization potential. This suggestion has been
experimentally tried. We have investigated the following hydrazinoazoles as ligands of
coordinated mercury perchlorates:
H2N
NHNH 2
HN
NHNH 2
HN
H2N
|
N
N
3-hydrazino-5aminopyrazole
H 2N
|
H 3C
N
NHNH 2
N
3-hydrazino-4-amino-5methyl-1,2,4-triazole
3-hydrazino-5aminopyrazolon
H2N
|
HS
N
NHNH 2
N
NHNH 2
N
3(5)-hydrazino-4-amino1,2,4-triazole
H2NHN
HN
N
N
3-hydrazino-4-amino-5mercapto-1,2,4-triazole
5-hydrazinotetrazole
Mercury (II) complex perchlorate with 3-hydrazino-5-aminopyrazole as a ligand was

synthesized:by the following reaction
H2N
HN
Hg(ClO4 ) + 2
[Hg(
)](ClO4)2
2
NHNH2
HN
NHNH 2
HN
N
N
VIII
150
Salt (VIII) showed high sensitivity to laser irradiation (pulse time 30 ns, the threshold
of initiation energy about 2.8.10-4 J). 5-Hydrazinotetrazole was taken as a ligand of choice.
According to the above reaction, the corresponding mercury (II) perchlorate complex was
obtained:
N N
Hg N \
N
H
NHNH2 (ClO4)2
(IX)
This complex demonstrated the highest sensitivity to laser irradiation combined with an
extremly low initiation threshold (pulsetime 30 ns, Einit. 1.10-5 J). It has detonation
velocity about 6 km/s. Mercury (II) complex perchlorates with other hydrazinoazoles have
lower sensitivity to laser pulse compared with mercury (II) complex perchlorate with 3hydrazino-5-aminopyrazole (pulse time ~1 s, E~0,2 J, ~1,06 m, ray diameter ~0,5 mm).
Basing on this result we synthesized a range of complexes of d-metals using 3(5)hydrazino-4-amino-1,2,4-triazole as ligand:
Me(ClO 4)2 + 2 N
NHNH 2
N
|
NH 2
Me = Cu, Ni, Co, Cd
N
[Me (
NHNH 2 ) 2] (ClO )
N
|
NH 2
4 2
(X-XIII)
Complexes (X-XIII) also showed high sensitivity to laser irradiation. Moreover,

initiation thresholds were shown in our experiments to increase in the sequence:
Cu<Cd<Ni<Co (the copper salt is the most light-sensitive compound). This is identical to the
row of the oxidative ability of central ions characterized by the sum of ionization potentials
of cations (I1+I2). It implies that the first stage of the laser initiation of perchlorate
complexes of 3(5)-hydrazino-4-amino-1,2,4-triazole is the oxidation of ClO4- to ClO4* by dmetal ions, ClO4* radicals being the species that undergo destruction and oxidize the organic
ligands.
4.
PRACTICAL APPLICATIONS
Depending on particular tasks, varions energetic metal complexes with varying values of
initiation thresholds to laser Q-switch pulse (pulsetime 10-8 s) or a non-Q-switch pulse
(pulsetime 10-3 s) have been suggested. For perforators used in oil-gas wells in Russia, a
system of laser initiation of cumulative charges was elaborated. The main light-sensitive
elements of this system were complex perchlorates of hydrazinoazoles. Practical application
of explosion energy for studying the development of deformations in materials or
constructions or the process of initiation of primary explosive charges necessitates
generation of detonation waves whose front would have a strictly specified shape. The
existing methods for obtaining plane detonation waves require intricate experimental
equipment and dont allow to simultaneously load large areas. Even more complicated is the
problem of creating cylindrical and spherical converging detonation waves.
151
A common shortcoming of the conventional means of initiation such as blasting caps or

detonation cords is that they excite detonation in the main explosive charge only locally, at
places of their location. These shortcomings can be eliminated by using a capless initiation
method based on laser ignition of explosives. This method makes it possible to obtain a
smooth front of profiled detonation waves on a surface of irregular shape. The method uses a
film charge of an explosive highly sensitive to laser pulse deposited onto the surface of a
material to be loaded or on the main explosive charge. The film charge is exposed to a laser
beam with the energy density exceeding the critical value of the ignition threshold. The
profile of the detonation wave is set by the shape of the initiated surface of the light-sensitive
explosive.
The proposed method enables formation of any detonation wave profiles without
additional explosive charges. An important advantage of the method is that the geometry of
the laser beam can be easily controlled (expanded, deflected, narrowed) and automation is
possible. In all cases the explosive is exploded remotely, with no essential limitations on the
distance from which initiation is done.
The capless laser initiation has been successfully applied to study the possibility of
hardening of steel surfaces by a plane shock wave formed by products appearing in a
explosion of film charges of light-sensitive explosives. The strength of sample surface was
improved as a result of loading by approximately 40-50%.
Consequently, energetic metal complexes are the perspective class of EM for the
development of safety initiation systems and elaborating new laser technologies of blasting.
Acknowledgments
The work was financially supported by the Ministry of Education of the Russian
Federation, in the framework of Intercollegiate Scientific and Technical Program Scientific
Research of High School in the Field of Chemistry and Chemical Products.
REFERENCES
[1] ILYUSHIN M.A., TSELINSKY I.V.: Primary explosives.//Ross. Khim. Zh. 1997, v.41, N 4,
p. 3 - 13 (in Russian)
[2] ILYUSHIN M.A., TSELINSKY I.V, Chernay A.V.: Light-sensitive explosives and compositions
and there laser initiation. Ross. Khim. Zh. 1997, v.41, N 4, p. 81 - 88 (in Russian)
[3] ILYUSHIN M.A., TSELINSKY I.V.: Laser initiation of high-energy-capacity compounds
in science and technology.//Russ. J. Appl. Chem. 2000, v.73, N 8, p. 1305 - 1312
[4] ILYUSHIN M.A., TSELINSKY I.V.: Energetic complexes of metals for initiation systems.//Ross.
Khim. Zh. 2001, v.41, N 1, p. 72 78 (in Russian)
[5] DANILOV YU.N., ILYUSHIN M.A., TSELINSKY I.V.: Industrial explosives. Part I. Primary
explosives. Text-book. Saint-Petersburg. Synthesis. 2001. 110 p.(in Russian)
[6] LISKER I.S.: Patent of Russian Federation 1673928 (1991) // B. I. 1992. 3.
152
1-(3,5-DINITROPHENYL), -3,3-DINITROAZETIDINE:
A NEW ENERGETIC MATERIALS
H. S. Jadhav*, D. D. Dhavale*, M. B. Talawar*** S. N. Asthana***
and V. N. Krishnamurthy**
*Department of Chemistry, University of Pune, Pune- 411 007, India
**DRDO/ISRO Cells, University of Pune, Pune- 411 007, India
***High energy Materials Research Laboratory, Pashan, Pune, India
Abstract:
Strained polynitro cyclic compounds are at the forefront of the search for more powerful
and less sensitive energetic materials. Examples in this regard are polynitrobicyclooctane,
polynitrocubanes, polynitropolycyclododecane, and various polynitromono- and
polycylicpolyazanitramines. Such materials are potentially useful as explosives, propellants,
fuels, and oxidizers in applications requiring substances which combining high energy, high
density and reduced sensitivity properties. An important new member of this class of
energetic materials is 1,3,3-Trinitroazetidine. Based on this analogy, we have synthesized
1-(3,5-Dinitrophenyl), 3,3-dinitroazetidine and fully characterized the new derivative.
The thermal, safety and explosive properties of the new compound compared with TNAZ.,
shows the new derivatives having acceptable energetic performance.
Keyword:
1.
Synthesis of TNAZ, Thermal Characterization, Energetic material
INTRODUCTION
The oxidizer, being the major constituent, plays a vital role in the physical and chemical
properties of the propellant system. The energy used in rocket propulsion is generally obtained
from an exothermic chemical reaction involving two molecular species, viz the oxidizer and to
fuel. Ideally, the oxidizer should have high oxygen content and highly exothermic reaction with
the fuel to increase the flame temperature. If the oxidizer potential is low, the fuel will be
starving for oxygen and hence will burn inefficiently.
An oxidizer is one having high oxidation potential and high electro negativity. The periodic
classification of the elements can be used to distinguish the oxidizer (high electro negativity)
and the fuel (high electro positivity). In other words the atoms on the right side of the neutral
substances like nitrogen are oxidizers whilst those on the left are fuels. Thus, fluorine is the best
oxidizer and F2/H2 propellant system gives maximum specific impulse among chemical
propulsion system.
In order to increase the performance level of conventional warheads, it is necessary to
increase the energy output of the explosives, which are used in the fillings. Two approaches
have emerged out successfully in this direction involving melt castable explosives and plastic
bonded explosives.
153
The one of the important high-energy oxidizer (melt castable explosives) of interest to the
military explosives is 1,3,3-trinitroazetidine. TNAZ is more powerful than widely used melt
castable explosive TNT. Low melting point of TNAZ (101 C, compared to 81 C for TNT)
than RDX despite there comparable energy potential would enable processing of formulations
on modified production lines. It has excess of oxygen available for the oxidation of fuel
ingredients in addition to its the strained energy and only half the impact sensitivity of HMX
without detonation. Successful formulation would enable a more powerful general purpose
explosive to be developed.
1,3,3-Trinitroazetidine is multifunctional compound. It has both N-NO2 and C-NO2 bonds
as well as the four membered constrained azaheterocycle. TNAZ believes to be very interesting
object for study of its unique property among the high energetic nitramines. Due to this world
over high-energy material researchers attention is towards these types of compounds. TNAZ has
been synthesized and characterized by a large number of methods in low yields [1].
It is generally reported that introduction of the picryl group increases the density and
stability of the compound and the four-member ring releases high energy due to strained ring.
Based on this analogy, we have synthesized the azetidine derivative of 3,5-dinitrobenzene
namely, 1-(3,5-Dinitrophenyl), 3,3-dinitroazetidine, which shows acceptable energetic
performance.
2.
EXPERIMENTAL
2.1
Methods and Materials
All the chemicals are purchased from Aldrich and Qualigen. The base catalyzed reaction
between nitromethane and formaldehyde gaves the diol insitu Ag+ induced oxidation of sodium
nitrite to form the nitronium ion in situ to quench the carbanion and give nitro alcohols. Which
on chlorination followed by condensation with 3,5-dinitroaniline afforded the cyclised azetidine
derivatives compound a. The melting points of the compounds were measured using Thomas
Hoover capillary melting point apparatus. The IR spectra were measured in KBr pellets on
Shimadzu FTIR spectrophotometer. Both carbon and proton NMR spectra were recorded on
Varian-Mercury 300 MHz spectrometer with tetramethylsilane as an internal standard in
solution. Elemental analysis was recorded on Perkin Elmer C, H, N elemental analyzer. The
thermal analysis (with about 2 mg samples) was performed on Mettler Toledo star system at a
heating rate of 10oC/min in nitrogen atmosphere at a flow rate of 80 ml/min.
The Impact test was conducted using an impact tester similar to that used in Naval
Ordnance Laboratory (NOL). Test specimens (30-35 mg of powder) were kept between two
hardened anvils and a 3 kg drop weight was allowed to fall freely from different heights. Both
open and aluminum foil encapsulated specimens were used. The 50% ignition height (cm) in
each case was evaluated using Bruceton method. Ten samples were tested for each compound.
The results are presented in Table 1 along with data for TNAZ.
The friction test was conducted in a test set up similar to the one employed by NOL. The
sample was kept between a fixed corrugated (rough) stainless steel plate and a movable stainless
154
steel plate. Pulling the movable plate at varying lever loads causes friction. The sample size and
evaluation procedure was the same as followed for impact testing. The data are given in Table 2.
The detonation velocity is estimated based on empirical calculations on molecular structure
developed by L.R. Rothstein [5], Detonation pressure was calculated by Chapman-Jouguet
(CJ) [5] method. Density () values were calculated by L. T. Eremenko method [5]. The results
are summarized in table 2. The oxygen balance values calculated and are given Table 2.
2.2
Synthesis
2.2.1. Synthesis of 2,2-Dinitro 1,3-propanediol.

To a mixture of 40.6 g of (0.66 mole) of nitromethane, 107.9 g of 37% formaldehyde and
91.9 g of water cooled to 0 C, a few drops of solution of 30.6 g of NaOH in 800 ml of water
was added with vigorous shaking. The cooling was stopped and was allowed to rise the
temperature to 40 C. The mixture was cooled again to 0 C and the remainder of the sodium
hydroxide solution was added in a steady stream keeping the temperature 5 to +5 C. After the
addition, the mixture was stored for 1-2 hrs, at 0 C. To this solution was added NaNO2 46 g
(0.66moles). This entire mixture was added to a solution of 2263 gm (13.3 moles) of AgNO3 in
300 ml of distilled water keeping the temperature below 25 C. The mixture was stirred for 2
hrs. The silver salt was filtered and the filtrate was extracted with three 1:1 portion of ether with
water. The ether extracts were dried on sodium sulphate, and the ether solution was evaporated
to a volume of less than 1:1. methylene chloride was added and the resulting solution was
cooled to precipitate the colorless product having mp 144-145 C. (Reported-146-147 C) [6].
1
H NMR: 4.37 (4H,s).

Elemental analysis:
Theoretical
C: 22.5
H: 3.75
N: 17.5
Observed
C: 22.67
H: 3.89
N: 17.3
2.2.2. Synthesis of 1,3-dichloro, 2,2-Dinitro propane.

2,2-Dinitropropanediol (5 gm, 0.12 mole) was dissolved in 12.5 ml of thionyl chloride.
Pyridine hydrochloride (0.750 g) was then added to the reaction mixture and refluxed for 72 hrs.
The excess thionyl chloride was removed in vacuo and the oily residue was fractionated under
reduced pressure to give 19.6 gm (81%) of product having a boiling point 58-60 C (60-62 C).
Sublimation of these liquids afforded white solid having melting point 38-39 C.
1
HNMR: 4.53 (4H,s).

Elemental analysis:
Theoretical
C: 17.73
H: 1.97
N: 13.79
Observed
C: 17.80
H: 2.0
N: 13.68
155
2.2.3. Synthesis of 1-(3,5-Dinitrophenyl), 3,3-dinitroazetidine.

3,5-Dinitroaniline (1.70 g, 9.85 mmol) was dissolved in dry methanol (30 ml), add K2CO3
and stir the reaction for 30 minutes at room temperature. Add this mixture slowly to 1,3Dichloro, 2,2-Dinitropropane (2g, 9.85 mmol) in methanol (20 ml). Reflux the reaction mixture
for 6 hrs, cooled and evaporated the methanol. Recrystallised from ethyl acetate afforded the
brown plates. (1.4 g, 35%)
IR: 1348 cm-1 (tetrtiary amine), 1535 cm-1, (aromatic region), 1585 cm-1 (geminal dinitro).
1
H NMR (CDCl3, 300 MHz, ):
4.63 (4H, s), 7.75 (1H,s), 8.35 (2H,s).

13
C NMR (CDCl3, 75 MHz, ): 58.97, 103.34, 111.79, 147.70, 149.74.

Elemental analysis:
Theoretical
C: 34.49
H: 2.25
N: 22.37
Observed
C: 34.55
H: 2.34
N: 22.30
2.2.4. Scheme
OH
HCHO + CH3NO2 + AgNO3
H2O O2N
+NaNo2
Cl
Pyridine hydrochloride O2N
SOCl2 reflux
O2N
O2N
OH
O2N
K2CO3/ Methanol
+
O2N
Cl
NO2
O2N
NH2
Cl
Cl
reflux
O2N
NO2
O2N
NO2
Compound a
156
3.
RESULTS AND DISCUSSION:

3.1
Structural determination:
The melting point of the 2,2-dinitro 1,3-propanediol 144-145 C, matches with the literature
value [6]. These compounds have been characterized by IR shows a gem dinitro band at 1575
cm-1, the absorption of alcohol shows at 3500 cm-1. 1H NMR Spectra shows two methylene
singlet at 4.37 , while the 1,3-dichloro 2,2-dinitro shows the two methylene singlet at 4.53
confirming the synthesis of the dichloro compound. Subsequent condensation gave a new
derivative of azetidine brown-colored plates it showing the aromatic region at 1535 cm 1, the
gem dinitro group shows at 1585 cm-1. Absence of the pri./sec. amine peak clearly indicates the
occurrence of condensation take place. The 1H NMR shows the 4.58 for four methylene
protons and a singlet at the 7.75 and 8.35 for the aromatic protons. 13C clearly indicates that the
condensation take place. The 13C data shows 58.97, 111.79, 147.70, 149.74 values for the title
compounds. The elemental analysis matches with the theoretical values confirmed the
compound has been synthesized.
3.2
Thermal and explosive properties
1-(3, 5-Dinitrophenyl), 3,3-dinitroazetidine (Compound a) decomposed in two steps. At

first there is a slightly weight loss at 46-90 C may be due to the moisture. The first stage starts
at 90-140 C, with 12.12% weight loss, while third stage starts at140-260 C, 52.59 % weight
loss. The DTG shows max temperature around 129, 240 C respectively for the two stages. The
DSC analysis shows an endothermic peak at 146.40 C, and an exothermic at 143.64 C
confirms the melting point of the compound. The energy of activation for two stages is, 15.06,
20.75 kcal/mole respectively. The other kinetic parameters are shown in table 2. The compound
is safer for handling as seen by the Impact and Friction data. The calculated detonation velocity
and pressure (shown in table 1) is acceptable for hazardous material. While due to the aromatic
nature of the compound it shows the high negative oxygen balance (Table1).
Table 1. Some properties of the 1-(3,5-Dinitrophenyl), 3,3-dinitroazetidine
compared with 1,3,3-Trinitroazetidine [1]:
Properties
TNAZ
Compound a
Density
1.83
1.63
Oxygen Balance %
-16.65
-68.99
Detonation Velocity
8.68
7.01
Detonation Pressure
25.99
20.02
Impact Sensitivity
28-29 [1]
>100(50%explosion ht/cm)
Energy of Activation
158.7
15.06, 21.03
157
Table 2. Thermal data for compound a and compared with TNAZ.

Decomposition (C)
Compound
Max DTG
DSC (C)
Ist stage
IInd stage
Exothermic
Endothermic
1-(3,5-Dinitrophenyl),
3,3-dinitroazetidine
90-140
140-260
128, 240.
143.64
146.40
TNAZ [1,6]
100-178
100, 178
101.97
Table 3. Kinetic data for the compound a:
Compound
Compound a
I st stage
Compound a
IInd stage
4.
Ea
Log A
-1
Kcal/mole
cm
1.8
15.06
-3.09
-1.81
-8.31
0.999
1.8
21.03
-2.29
-2.53
-6.89
0.885
CONCLUSIONS:
New derivatives of azetidine, 1-(3,5-Dinitrophenyl), 3,3-dinitroazetidine, showing

moderately powerful explosive properties, have been synthesized and fully characterized.
Furthermore, the ease with which the starting materials can be prepared from commercially
available compounds suggests that the compound would be synthesize on a large scale for
application as explosives.
Acknowledgement
Authors are thankful to Director H.E.M.R.L., Smt. Bhate, Shri Sarangan (Library incharge)
for providing the library facilities, thanks to Shri. Marimuthu for providing the thermal
analysis. One of the author HSJ thanks to H.E.M.R.L. for providing the fellowship.
158
REFERENCES
[1] ZDNEK JALOVY, SVATOPLUK ZEMAN, MUHAMED SUCESKA, PAVEL VAVRA, KAMIL DUDEK, MASA
RAJIC: 1,3,3-Trinitroazetidine, 19, J. Energetc Material, p 219-272, 2001.
[2] MEHILAL, A. K. SIKDER, R. B. SALUNKE, AND N. SIKDER: Nitroanilinodinitrofuroxans-Synthesis,
Characterization, 25, New Jour. of Chemistry, 25, 1549-1552, 2001.
[3] MEHILAL, , R. B. SALUNKE, A. K. SIKDER J. P. AGARWAL: Nitroanilinodinitrofuroxans-Synthesis,
Characterization, 25, New Jour. of Chemistry, 25, 1549-1552, 2001.
[4] R. B. KAPLAN AND H. SHECHTER: J. Am. Chem. Soc., 83, 3535, 1961.
[5] PAUL W COOPER: Explosives Engineering, VCH Publishers, Inc., New York, pp. 67-80, 1996.
[6] FRANK G. BORGARDT, ALLAN K. SELLER AND PAUL NOBLE, JR.: Aliphatic Polynitro Compounds.
I Synthesis of 1,1,1-Trinitrochloroethane and its Rearrangement to Dipotassium Tetranitroethane, J.
Org. Chem., 2806-2811, 1966.
159
CRYSTALLIZATION OF HEXANITROSTILBENE
Zdenk Jalov*, Pavel Mareek**, Kamil Dudek**
* Department of Theory and Technology of Explosives,
University of Pardubice, Czech Republic,
**Explosia, a.s. VPCH,
Pardubice, Czech Republic
Abstract:
A literature survey of methods for hexanitrostilbene (HNS) crystallization is presented.
Crystallization of HNS from N-methylpyrrolidinone, cyclohexanone, dimethylformamide
and nitric acid are part of the experimental section. Cyclic heating-cooling
crystallization from dimethylforamide produces HNS with bulk density 0,28 g/cm3.
Crystallization from nitric acid produced HNS with bulk density 0,56 g/cm3 and acidity
0,028% and 0,083%. resp. with dependence on the acidity method determination.
Keyword:
1.
crystallization, hexanitrostilbene, HNS
INTRODUCTION
Hexanitrostilbene is a heat resistant explosive with thermal stability over 300oC. It is

used as a crystal modifying additive in trinitrotoluene melt cast charges [1]. It is also used, for
example, in space technology, oil and gas exploration, detonators and boosters
for explosives.
Several methods for HNS crystallization were referred in Chemical Abstracts
(1972-2002). Syrop [2] studied a recrystallization process for HNS-II using the mixture
of acetonitrile and xylene. Quinlin et al. [3-5] developed a process for producing of HNS-II
using dimethylformamide as a solvent and entering the mixture into a heating and cooling
cycles. Stull [6] developed a process for producing of fine HNS with a mean size about 2 m.
The process involves dissolving HNS in dimethylformamide and slow addition of this
solution into precooled and rapidly agitated distilled water. Dacons [7] developed a process
for produce of finely divided HNS-I by dissolving HNS in hot 90% nitric acid and drowning
the solution into water. Kayser [8] crystallized HNS from dimethysulfoxide and methanol
to obtain finely divided particles of uniform size. Bellamy [9-11] crystallized HNS
from N-methylpyrrolidone and chlorobenzene. After sonification of HNS in suspension he
obtained crystals with higher value of bulk density. Teipel and Mikonsaari [12] reduced
the particle size of HNS by using an ultrasonic devise and a rotor stator milling device.
160
2.
EXPERIMENTAL
The aim of the work was to produce HNS with suitable crystal shape for use
in thermostable detonators manufactured by Austin Detonators, Vsetn, Czech Republic.
Fine HNS-I with particle size up to 10 m (see fig. 1) was used as a starting material
in the following examples. The speed of the stirrer during cooling, except of example 9, was
1200-1400 rpm.
Fig 1.
Starting HNS-I with particle size up to 10 m.
Example 1
10 g of fine HNS-I was added to a flask with 75 ml of N-methylpyrrolidone. The
mixture was gradually heated with mechanical stirring to 115oC. Then 35 ml of toluene was
added in one step and the mixture was gradually cooled to 15oC (cooling rate 2oC/min). HNS
was filtered and washed with acetone. The average yield of HNS was 82%. The product
crystallized in the form of yellow needles.
Example 2
10 g of fine HNS-I was added to a flask with 75 ml of N-methylpyrrolidone. The
mixture was gradually heated with mechanical stirring to 115oC. Then 35 ml of toluene was
added in one step and the mixture was rapidly cooled to 15oC. HNS was filtered and washed
with acetone. The average yield of HNS was 76%. The product crystallized in the form
of very fine crystals.
Example 3
10 g of fine HNS-I was added to a flask with 250 ml of cyclohexanone. The mixture was
gradually heated with mechanical stirring to 140oC. Then the mixture was gradually cooled
to 15oC (cooling rate 2oC/min). The product was filtered and washed with acetone. HNS
crystallized in the form of yellow needles.
Example 4
10 g of fine HNS-I was added to a flask with 180 ml of cyclohexanone. The suspension
was gradually heated with mechanical stirring to 140oC. Then the mixture was gradually
cooled to 15oC (cooling rate 2oC/min). The product was filtered and washed with acetone.
161
The product crystallized partially in the form of yellow needles and partially remains
unchanged or in the form of very fine crystals.
Example 5
75 g of fine HNS-I was added to a flask with 1500 ml of dimethylformamide.
The mixture was gradually heated with mechanical stirring to 95oC. Then the mixture was
gradually cooled to 15oC (cooling rate 20oC/hod). The product was filtered and washed
with acetone. HNS crystallized in the form of long needles.
Fig 2.
HNS crystals obtained by the method described in example 5.
Example 6
75 g of HNS-I was added to a flask with 750 ml of dimethylformamide. The suspension
was gradually heated with mechanical stirring to 95oC. Then the mixture was gradually
cooled to 15oC (cooling rate 15oC/hod). The product was filtered and washed with acetone.
The product crystallized in the form of oblong plates. Bulk density of crystallized HNS was
0,20 g/cm3.
Fig 3.
Example 7
75 g of HNS-I was added to a flask with 750 ml of dimethylformamide. The mixture
was gradually heated with mechanical stirring to 95oC. Then the suspension was submitted
to cyclic cooling and heating as follows: 95oC50oC80oC40oC60oC15oC.
162
The cooling rate was 15-20oC/hod. The product crystallized in the form of oblong plates.
The average yield of HNS was 79%. Bulk density of crystallized HNS was 0,28 g/cm3.
Fig 4.
Example 8
75 g of HNS-I was added to a flask with 750 ml of dimethylformamide. The mixture
was gradually heated with mechanical stirring to 95oC. Then the suspension was submitted
to cyclic cooling and heating as follows: 95oC50oC75oC15oC. The cooling rate was
15-20oC/hod. The product crystallized in the form of oblong plates. The average yield
of HNS was 77%.
Example 9
20 g of fine HNS-I was added in portions to a flask with 200 ml of 98% nitric acid. Then
the temperature of mixture was increased to 80oC using water bath. HNS was dissolved and
the bath was removed and the solution was stirred (500 600 rpm). Approximately
after 4 hours the temperature of mixture fell to 25oC and HNS crystallized. Suspension
was cooled to 0oC by using of ice-water bath. Recrystallized HNS was filtered and washed
with water (500 ml). To reduce the acidity in material, HNS was boiled at 300 ml of water
(80 90oC). 10 g of crystalline HNS was obtained.
The filtrate was poured into 1000 ml of ice water and precipitated HNS was filtered.
10 g of powdered HNS was obtained.
This laboratory method was successfully transferred to a pilot-plant scale. The charge
of starting raw HNS-I was 16 kg.
Bulk density of crystallized HNS was 0,56 g/cm3. Acidity of the sample after boiling
in hot water was 0,028% (titration of HNS suspension in water) and 0,083% (titration
of HNS solution in dimethylformamide) respectively.
163
Fig 5.
3.
CONCLUSIONS
The processes of HNS crystallization from its solutions in N-methylpyrrolidonene and

cyclohexanone by the manners described in examples 1 and 3 produced the final product
in the form of needles crystals.
The processes of HNS crystallization from its solution in N-methylpyrrolidone

with rapid cooling and from its suspension in cyclohexanone by the manners described
in examples 2 and 4 produced the final product in the form of powdered crystals.
The process of HNS crystallization from its solution in dimethylformamide

by the manner described in example 5 produced the final product in the form of needle
crystals, like in examples 1 and 3.
The processes of HNS crystallization from its suspensions in dimethylformamide by the

manners described in examples 6-8 produced the final product in the form of oblong
plates. Better crystal shape was obtained by introducing of some heating and cooling
cycles into the crystallization process.
The process of HNS crystallization from its solution in 98% nitric acid by the manner
described in examples 9 produced the final product in the form of oblong plates.
This process yields the best crystal shape of HNS from all mentioned examples.
Only HNS crystals produced by heating and cooling cycles of its suspension
in dimethylformamide and by crystallization of its solution in 98% nitric acid seem to be
suitable for applications as a secondary charge of detonators.
164
Acknowledgement:
The authors would like to thank to Svatopluk Zeman for useful advice during the work,
Ji Ko from Austin Detonators for advice in the field of detonators, Robert Maty
for making the pictures of HNS crystals and Jan Horkel, Milan Klusek and Luk Vejs
for very useful help during the experiments.
REFERENCES
[1] PARRY, M. A.; THORPE, B. W: Journal of Crystal Growth (1979), 47(4), 541-50.
[2] SYROP, L. J.: U.S. 3,699,176 (1972); CAN 78:32252.
[3] QUINLIN, W. T.: Evans, V. H.; Schaffer, C. L.; Osborn, A. G.; Stallings, T. L.: Report
(1976), MHSMP-76-41; CAN 88:107529.
[4] QUINLIN, W. T.: Report (1977), MHSMP-77-23; CAN 88:173078.
[5] QUINLIN, W. T.: Report (1978), MHSMP-78-54; CAN 91:93834.
[6] STULL, T. W.: Report (1979), MHSMP-79-51; CAN 93:97902.
[7] DACONS, J. C.: US Pat. Appl. 126,269 (1980); CAN 94:49756.
[8] KAYSER, E. G.: US Pat. Appl. 508,789 (1985); CAN 105:24053.
[9] BELLAMY, A.: EP Patent 277,386 (1988); CAN 109: 230521.
[10] BELLAMY, A.: EP Patent 275,607 (1988); CAN 109:202232.
[11] BELLAMY, A.: SE Patent 468,435 (1993); CAN 119:163588.
[12] TEIPEL, U; MIKONSAARI, I.: Propellants, Explosives, Pyrotechnics (2002), 27(3),
168-174.
165
STUDY OF DECOMPOSITION OF TNT BY HEAT AND SHOCK

Martin Kouba, Svatopluk Zeman and Eva Zemanov
Departmet of Theory & Technology of Explosives, University of Pardubice,
CZ-532 10 Pardubice, Czech Republic
Abstract:
Samples of technical TNT exposed to heat or to shock have been analysed
chromatographically (HPLC). It was found that the main decomposition products are
identical in the two cases. It has been stated that the chemical micro-mechanism of the
primary fragmentations of shocked TNT molecules should be the same as in the case of
their low-temperature thermal decomposition
Keyword:
1.
TNT, shock, thermal decomposition
INTRODUCTION
The homolytic fragmentations or reactions of the CNO2, NNO2, NNO, and ONO2
groupings, or other bearers of explosibility (i.e. explosophores), are common primary fission
processes of energetic materials under thermal [1-18], impact [1,3,4,6,16,19-27], shock [1,6,18,19,20,24,2632]
and electric spark stimuli [5, 32-36]. Therefore, it is natural that there are relationships
between characteristics of low-temperature thermal decomposition and impact [6,37,38] or
electric spark [35,39,47] sensitivities and also detonation characteristics [26,40-48,50] of polynitro
compounds.
From what has been said so far it follows that chemical micro-mechanism of the primary
fragmentations of explosive molecules in all the above-mentioned stimuli should be the same
as in the case of their low-temperature thermal decomposition. The similarity or identity is a
topic of numerous papers [5,6,18,26,32-35,44-50]. A thermal entity of the initiation by electric spark
[52]
appears to be thus manifested [35,39,53]. In the case of initiation of detonation the said
identity is confirmed by some striking pieces of experimental evidence. First of all they
include relationships between outputs of thermo-analytical methods and detonation
characteristics of polynitro compounds [40-47,50]. By means of XPS, the evidence of primary
fission of NNO2 bond in 1,3,5-trinitro-1,3,5-triazacyclohexane (RDX) exposed to shock
wave was obtained [51]. The same fission represents a beginning of the RDX thermal
decomposition [6,10,13,14,17,18,26,47]. On the basis of deuterium kinetic isotope effect (DKIE) it
was proved [11,54] that the rate-limiting step for the thermal decomposition of 2,4,6trinitrotoluene (TNT) in the condensed state and that for the initiation of its detonation are
identical. The presence of furoxanes and furazanes in the XPS spectrum of 1,3,5-triamino2,4,6-trinitrobenzene (TATB) exposed to shock [55-57] provides further evidence the
pyrolysis of ortho-nitroanilines is a method of synthesis of benzofurazane [58] and in the case
of 1,3-diamino-2,4,6-trinitrobenzene (DATB) this reaction leads to 4-amino-5,7dinitrobenzofurazane [59]. The identity is also reflected in a relationship between the kinetics
of the low-temperature thermal decomposition of the energetic materials and reaction rates in
the reaction zone of their detonation [60,61].
166
The above-mentioned common rate-limiting step of thermal and detonation

decomposition of TNT [11,54] made us presume the identity of the main products of
incomplete thermal and detonation decomposition of the said explosive. The present paper
deals with confirming this presumption using samples of shock- and heat-exposed TNT and
liquid chromatography as the detection method.
2.
EXPERIMENTAL
The exposition of TNT to a shock wave was realised by means of the small-scale gap
test [63,64]. The donor charges (PETN with 10 % wt. wax and RDX with 5 % wt. wax) were
separated from the acceptor TNT charge by a PMMA barrier of such a thickness that the
intensity of the resulting shock wave were at the limit of initiation ability of TNT. The set
prepared in this way was put into a round paper box serving for catching the TNT residues
after its detonation. Then these residues were transferred in a crystallisation dish, and 300 ml
distilled water was added thereto. After the extraction, the solid explosion residua were
removed by filtration, and the filtrate was subject to analytical detection by means of HPLC.
The thermal exposition of TNT sample was realised by means of a self-developed [62]
apparatus for differential thermal analysis DTA 550-Rez. At first, the measurement was
carried out with the amount of 200 mg TNT, the heating being interrupted at 318 C by
taking the test tube with TNT out of the apparatus. Then the second exposition was carried
out with an amount of 500 mg TNT, the sample being taken out of the apparatus at the
temperature of 315 C. An abrupt cooling of the exposed samples was achieved in a
crystallisation dish with 150 or 300 ml, resp., distilled water. The subsequent extraction was
carried out in dark overnight. The aqueous extract was analysed by means of liquid
chromatography.
The choice of water as the extraction agent for residues of exposed TNT samples was
not random. It represents an application of a new efficient procedure for collecting afterexplosion residua in criminology [65,66]. Water used as the solvent does not form any peak in
the resulting chromatogram, which simplifies the determination of the TNT decomposition
products (acetone produces a large response in the chromatogram which can overlap some
important peaks).
The liquid chromatography was realised on an older computerised apparatus LCD 2040
(Laboratorn pstroje, Prague). As it was not possible to measure separately the UV spectra
of each compound, the wavelength of 225 nm was chosen for the detection, which was
previously evaluated as a versatile wavelength for analyses of the most common explosives.
A mixture of acetonitrile and water 1:1 was used as the mobile phase, its flow rate being
adjusted at 1.2 ml.min1. The first part of measurements adopted an older column Separon
C18, 150 mm, grain size 7 m, whereas the next part adopted a Biospher C18 column
PSI 200, grain size 7 m. The pressure at the column was 10 MPa, the air humidity was
about 40 % and room temperature was 25 C. The chromatograms from this part are
presented in Figs 1 and 2.
The resulting values of retention time on chromatogram were compared with those of
standards, see 2,4- and 2,6-dinitrotoluenes (DNT), 2,4,6-trinitrobenzaldehyde (TNBz),
2,4,6-trinitrobenzoic acid (TNBA), 2,4,6-trinitrotoluene (TNT), and 1,3,5-trinitrobenzene
(TNB).
167
Fig 1.
A chromatogram (Biospher C18 column PSI 200) of thermally

exposed TNT.
Fig 2.
A chromatogram (Biospher C18 column PSI 200) of shock-exposed TNT.
168
3.
Dacons et al. [67] analysed a TNT sample that was exposed to 200 C 16 h, and found
that beside polymeric decomposition products it contained 4,6-dinitroanthranil (DNAt),
2,4,6-trinitrobenzaldehyde (TNBz), and 2,4,6-trinitrobenzyl alcohol (TNBol). The content of
DNAt predominated. Rogers found also 1,3,5-trinitrobenzene (TNB) and 2,4,6trintrobenzoic acid (TNBA) [68] in non-isothermally exposed TNT. At our experimental
conditions of thermal exposition of technical TNT and subsequent chromatographic analysis
we have so far safely identified TNBz, TNB and DNAt (see Fig. 1). DNT was already
present in the starting TNT. DNAt is a TNT thermolysis product formed by the so-called
trinitrotoluene mechanism [47]. TNBz and TNB are products of oxidation of methyl group in
TNT by nitrogen oxides [47].
Shock-exposed TNT has not yet been described analytically. The chromatographic
record obtained by us is shown in Fig. 2. It can be seen that the substances present in Fig. 1
are also present in Fig. 2. It must be pointed out that amount of DNAt is higher in Fig. 2 than
in Fig. 1. In near future we are going to deal with identification of the remaining peaks (they
can come from residues of donor charges, too).
DNAt, TNBz and TNB in Fig. 1 are products of thermal decomposition of TNT.
However, in the case of shock-wave-exposed TNT their formation cannot be interpreted as
resulting from action of heat. The ideas about fragmentation of energetic materials by shock
can be presented as follows [26,47]: The influence of shock on energetic materials results in
adiabatic compression of the molecular layer struck. According to Klimenko and Dremin [6972]
, the kinetic energy of the shock in this compression is accumulated, through translationalvibrational relaxation processes, by translational and vibrational modes of molecular crystals
of the material within 1013 to 1012 seconds. These causes a considerable quasi-overheating
(20 000 to 40 000 K [71,72]) especially of vibrational modes. A non-equilibrium state is
established with concomitant primary fission of the energetic material into ions and radicals
[70-72]
. Chemical reactions of these active particles cause the shock front to spread and evoke
a second equilibrium stage of detonation behind the front. This or similar ideas of
transformation of low-frequency vibrations of crystal lattice (acoustic phonons) into highfrequency vibrations (vibrons), with subsequent spontaneous localization of vibrational
energy in the explosophore groupings [78,79], have been applied by a number of authors in
their studies of shock reactivity of energetic materials (for representative papers see Refs [7379]
). The production of primary fragments after a weak stimulus (shock) is insufficient for
formation of a second equilibrium stage and the detonation goes out. These fragments are
then stabilised by producing stable products identical with those of low-temperature
thermolysis of energetic materials. In other words (see Introduction): chemical micromechanism of the primary fragmentations of shock-exposed explosive molecules should be
the same as in the case of their low-temperature thermal decomposition.
169
4.
CONCLUSION
The main intermediates of the decomposition of TNT by shock and heat are identical.
This finding excellently agrees with the experimental results obtained by Bulusu et al. [11,54]
in the field of application of DKIE to study of inititation of TNT. Our results as well as those
by Bulusu signalise that the chemical micro-mechanism of the primary fragmentations of
shock-exposed TNT molecules should be the same as in the case of their low-temperature
thermal decomposition.
Acknowledgement
The authors are indebted to Ministry of Industry and Commerce of the Czech Republic
for financial support within project FC-M2/05.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Xiao Heming, Wang Zun-Yao and Yao Jian-Min, Acta Chim. Sinica, 43 (1985) 14.
F. J. Owens, K. Jayasuriya, L. Abrahmsen and P. Politzer, Chem. Phys. Letter, 116 (1985) 434.
Fan Jianfen and Xiao Heming, J. Mol. Struct. (THEOCHEM), 365 (1996) 225.
F. J. Owens, J. Mol. Struct. (THEOCHEM), 370 (1996) 11.
V. Zeman and S. Zeman, Proc. 28th Int. Annual Conf. ICT, Karlsruhe, 1997, p 67/1.
S. Zeman, Propellants, Explos., Pyrotech., 25 (2000) 66.
V. G. Matveev, V. V. Dubikhin and G. M. Nazin, Izv. Akad. Nauk SSSR, Ser. Khim., (1978)
474.
S. Bulusu and T. Axenrod, Org. Mass Spectrom., 14 (1979) 585.
J. Wang and H.-Y. Lang, Science in China, Ser. B, 33 (1990) 257.
G. M. Nazin and G. B. Manelis, Russ. Chem. Revs., 63 (1994) 313.
S. Bulusu, D. I. Weinstein, J. R. Autera, D. A. Anderson and R. W. Velicky, in Proc. 8th Int.
Symp. on Detonation, Albuquerque, NM, July 1985.
G. J. Piermarini, S. Block and P. J. Miller, in S. N. Bulusu (Ed.), Chemistry and Physics of
Energetic Materials, Kluwer Academic Publ., Dordrecht, 1990, p. 391.
V.P. Stepanov, A.A. Fedotov, A.N. Pavlov and G.B. Manelis, in B. N. Novozhilov (Ed.),
Khimicheskaya fizika protsessov goreniya i vzryva (Chemical Physics of the Combustion and
Explosion), Proc. 9th All-union Symp. Combust. Explos, Acad. Scienses USSR,
Chernogolovka, 1989, p. 100.
M. D. Cook, J. Energ. Mater., 5 (1987) 257.
R. Behrens, Jr., and S. Bulusu, J. Phys. Chem., 95 (1991) 5838.
B. A. Lure, B. S. Svetlov and A. N. Chernyshov, Symp. Chem. Probl. Connected Stabil.
Explos., 9th (1992) 119.
G. B. Manelis, G. M. Nazin, Yu. I. Rubtsov and V. A. Strunin, Termicheskoe razlozhenie i
goreniye vzryvchatykh veschestv i porokhov (Thermal Decomposition and Combustion of
Explosives and Powders). Izdat. Nauka, Moscow, 1996.
M. M. Kuklja, J. Phys. Chem., B 105 (2001) 10159.
F. J. Owens and P. Politzer, Shock Waves Condens. Matter, Proc. 4th Am. Phys. Soc. Top.
Conf. 1985, Publ. Plenum Press, New York, 1986, pp. 857-861.
170
[20] J. S. Murray and P. Politzer, in S. N. Bulusu (Ed.): Chemistry and Physics of Energetic
Materials. Kluwer Acad. Publ., Dordrecht, 1990, pp. 175-193.
[21] P. Politzer, J. S. Murray, P. Lane, P. Sjberg and H. G. Adolph, Chem. Phys. Letters, 181
(1991) 78.
[22] J. S. Murray, P. Lane, P. Politzer and R. Bolduc, Chem. Phys. Letters, 168 (1990) 135.
[23] S. Odiot, in S. N. Bulusu (Ed.): Chemistry and Physics of Energetic Materials. Kluwer Acad.
Publ., Dordrecht, 1990, pp. 79-130.
[24] M. M. Kuklja and A. B. Kunz, J. Appl. Phys., 87 (2000) 2215.
[25] B. M. Rice and J. J. Hare, J. Phys. Chem., A 106 (2002) 1770.
[26] S. Zeman, J. Energet. Mater., 17 (1999) 305.
[27] Y. Kohno, K. Ueda and A. Imamura, J. Phys. Chem., 100 1996) 4701.
[28] S. Zeman, M. Huczala and Z. Friedl, J. Energet. Mater.,20 (2002) 53.
[29] A. Delpuech and J. Cherville, Propellants Explos., 3 (1978) 169.
[30] A. Delpuech and J. Cherville, Propellants Explos., 4 (1979) 61.
[31] M. M. Kuklja, E. V. Stefanovich and A. B. Kunz, J. Phys. Chem., 112 (2000) 3417.
[32] S. Zeman, V. Zeman and Z. Kamensk, Proc. 28th Int. Annual Conf. ICT, Karlsruhe, 1997,
pp. 66/1.
[33] V. Zeman, J. Ko and S. Zeman, HanNeng CaiLiao, 7 (1999) 127.
[34] V. Zeman, J. Ko and S. Zeman, HanNeng CaiLiao, 7 (1999) 172.
[35] J. Ko and S. Zeman, HanNeng CaiLiao, 8 (2000) 18.
[36] J. Ko, V. Zeman and S. Zeman, HanNeng CaiLiao, 9 (2001) 60.
[37] S. Zeman, in Proc. 5th Int. Seminar New Trends in Research of Energetic Materials,
University of Pardubice, April 2002, p.434.
[38] M. A. Schrader,M. W. Leeuv and A. C. Van der Steen, Proc. 9th Int. Pyrotech. Seminar,
Colorado Springs, Aug. 1984, p. 881.
[39] D. Skinner, D. Olson and A. Block-Bolten, Propellants, Explos., Pyrotech. 23 (1997) 34.
[40] S. Zeman, Thermochim. Acta, 41 (1980) 199.
[41] N. M. Bhide, S. R. Naidu, E. M. Kurian and K. R. K. Rao, J. Thermal Anal., 35 (1989) 1181.
[42] K. V. Prabhakaran, N. M. Bhide and E. M. Kurian, Thermochim. Acta, 220 (1993) 169.
[43] P. S. Makashir and E. M. Kurian, J. Thermal Anal., 46 (1996) 225.
[44] S. Zeman, M. Dimun and . Truchlk, Thermochim. Acta, 78 (1984) 181.
[45] S. Zeman, M. Dimun, . Truchlik and V. Kabtov, Thermochim. Acta, 80 (1984) 137.
[46] S. Zeman, M. Dimun, V. Kabtov and . Truchlik, Thermochim. Acta, 81 (1984) 359.
[50] S. Zeman, P. Kohlek and M. Maranda, Thermochim. Acta, - in press
[51] J. Owens and J. Sharma, J. Appl. Phys., 51 (1980) 1494.
[52] M. Auzanneau and M. Roux, Propellants, Explos., Pyrotech., 20, (1995) 99.
[53] J. Ko, S. Zeman, J. Majzlk and J. Strnad, Proc. th Int. Seminar New Trends in Research
Explosives, University of Pardubice, April 2002, p. 110.
[54] S. Bulusu and J. R. Autera, J. Energ. Mater., 1 (1983) 133.
[55] J. Sharma, J. C. Hoffsommer, D. J. Glover, C. S. Coffey, F. Santiago, A. Stolovy and S.
Yasuda, , in J. R. Asay, R. A. Graham and G. K. Straub (Eds.), Shock Waves in Condensed
Matter, Elsevier Sci. Pub. B. V., Amsterdam, 1984, p. 543.
[56] J. Sharma, J. W. Forbes, C. C. Cofey and T. P. Liddiard, J. Phys. Chem., 91 (1987) 5139.
171
[57] J. Sharma, in: S. C. Schmidt, R. D. Dick, J. W. Forbes and D. G. Tasker (Eds.), Shock
Compression of Condensed Matter, Elsevier Sci. Publ. B. V., 1992, p. 639.
[58] J. M. Prokipcak, P. A. Forte and D. D. Lennox, Canad. J. Chem., 47 (1969) 2482.
[59] T. P. Hobin, Tetrahedron, 24(1968) 6145.
[60] B. N. Kondrikov, in: J. R. Bowen, N. Manson, A. K. Openheim and R. I. Soloukhin (Eds.),
Shock Waves, Explosions, and Detonations, Vol. 87 of Progress in Astronautics and
Aeronautics, 1983, p. 426.
[61] S. Zeman, . Gazda, A. tolcov and A. Dimun, Thermochim. Acta, 247 (1994) 447.
[62] M. Krupka in: Proc. of the 4th Int. Seminar New Trends in Research of Energetic Materials,
University of Pardubice, April 2001, pp. 222-227.
[63] S. Brebera in: Proceedings of the First Internal Symposium, Res. Inst. Ind. Chemistry,
Pardubice, 1974, pp. 13/15.
[64] P. Mok and P. Vvra in: Proceedings of the 27th Int. Annual Conf. ICT, Karlsruhe, June
1996, p.37/1.
[65] J. Chldek in: Int. Symposium Ecological Liquidation of Explosives; Detection and
Identification of Explosives before and after Explosion. Sept. 1996, Hotel Labe, Pardubice,
p. 55.
[66] R. Varga and P. Ulbrich in: Proceedings of the 6th Seminar New Trends in Research of
Energetic Materials, University of Pardubice, April 2003
[67] J. C. Dacons, H. G. Adolph and M. J. Kamlet, J. Phys. Chem. 74 (1970) 3035.
[68] R. N. Rogers, Anal. Chem. 39 (1967) 730.
[69] V. Yu. Klimenko and A. N. Dremin, Proc. 6th All-Union Symp. Combust. Explos., Sept. 1980,
Alma Ata, USSR, p. 69.
[70] A. N. Dremin, . Phil. Trans. R. Soc. London, A 339 (1992) 355.
[71] V. Yu. Klimenko, M. Ya. Yakoventsev and A. N. Dremin, Khim. Fizika, 12 (1993) 671.
[72] V. Yu. Klimenko,. Khim. Fizika, 17 (1998) 11.
[73] F. J. Zerilli and E. T. Toton, Phys. Rev., 29 (1984) 5891.
[74] F. E. Walker, J. Appl. Phys., 63 (1988) 5548.
[75] M. D. Fayer, A. Tokmakoff and D. D. Dlott, in: D. H. Liebenberg, R. W. Armstrong and
J. J. Gilman (Eds.), Mat. Res. Soc. Symp. Proc., 296 (1993) 379.
[76] L. E. Fried and A. J. Ruggiero, J. Phys. Chem., 98 (1994) 9786.
[77] D. D. Dlott, J. Physique IV, Coll. C4, supplement J. Phys. III, 5 (1995) C4/337.
[78] Xiaoyu Hong, J. R. Hill and D. D. Dlott in T. B. Brill, T. P. Russell. W. C. Tao and R. B.
Wardle (Eds.), Mat. Res. Soc. Symp. Proc., 418 (1995) 357.
[79] P. J. Haskins and M. D. Cook, Proc. 1997 APS Conf. on Shock Waves in Condensed Matter,
Amherst, USA, July 1997.
172
EXPLOSION HAZARD OF SOME ORGANIC PEROXIDES

Georgy.D. Kozak, Nikolai.I. Akinin, Vlada.M. Raikova and Svetlana.V. Arinina
Mendeleev University of Chemical Technology
Miusskaya sq. 9, Moscow A-47, Russia
Abstract:
The ultimate objective of our work was to investigate the ability to heat explosion,
burning and detonation of some organic peroxides. The main attention was paid to
investigation of cumene hydroperoxide, but some properties of benzoyl peroxide and
were collected and analyzed too. For the first time the law of burning rate of CH) and
temperature of burning were measured at p=24-36 MPa, and low velocity detonation
was fixed in steel tubes by the mean of photorecording method. The explanation of the
investigation results focused on identifying the most probable occurring chemical
reactions at decomposition, heat explosion, and burning of CH and on reaction
thermochemistry data.
Keyword:
1.
peroxide, decomposition, burning, detonation.
INTRODUCTION
The investigation of process safety management of some highly hazardous energetic

materials must include adequately examination of its explosion parameters. The ultimate
objective of our work was to investigate the ability to heat explosion, burning and detonation
of some organic peroxides.
The main attention was paid to investigation of cumene hydroperoxide (CH) because of
it utilizing on a large scale in chemical industry, but some properties of benzoyl peroxide
(BP) were collected and analyzed too.
Typically, the first step in hazard evaluation of substances was a literature searches. We
used the next searches:
Scientific and technical literature,
International Chemical Safety Cards (ICSC, Internet),
Database of National Institute of Standards and Technology (NIST, Internet),
Hazardous Chemical Database (HCD, Internet).
BP and CH are an oxygenated derivatives of benzene, represented by the chemical
formulas C14H10O4 and C9H12O2 , correspondingly. The both of substances have a peroxide
group -O-O-, which is explosiphore one, and could provide them explosiveness [1]. Table 1
lists theirs characteristic properties.
The parameters and mechanism of low thermal decomposition of CH and BP were
described in literature [3,4]. The activation energies (E) and pre-exponential factors (B) of
decomposition are listed in Table 1.
BP has long been recognized as unstable chemical and explosive. International chemical
safety cards characterize CH and BP as combustible and explosive substances, however the
results of thorough investigation and description of explosion properties of CH were not
found in literature.
Burning of BP had been thoroughly investigated under head professor A.E.Fogelzang
over the pressure range 0.02 to 40 MPa in the work [4]. Not only burning law (velocity vs.
pressure) of BP was measured, but temperature of burning, and dependence of burning rate
from initial temperature were experimentally investigated.
173
Table 1. Characteristic properties of BP and CH

CHARACTERISTICS
Molecular weight, g/mol
Melting point, deg. C
Boiling point, deg. C
Density, g/ccm (water = 1)
Flash point, 0 C
Solubility in water, g/100 ml
Refractive index
Vapor pressure, at 200 C
Explosive limits, vol % in air
Activation energy, kJ/mol
Pre-exponential factor, s-1
Enthalpy of formation,
kJ/mol
Enthalpy of evaporation,
kJ/mol
Enthalpy of sublimation,
kJ/mol
BP
242.2
103 [ICSC]
103-106 [HCD]
Decomp. Explosively [HCD]
1.3 [ICSC]
1.334 [HCD]
80 [ICSC]
Poor [ICSC]
1.543 [HCD]
< 0.1 kPa [ICSC]
129.05 [4]
1014.40 [4]
-392.5 (solid) [2]
-393.06.3(solid) [NIST]
-3694.6 (solid) [NIST]
-3698.4 (solid) [NIST]
-2725.4 (gas) [NIST]
-281.76.2 (gas) [NIST]
97.92.5 [NIST]
983 [NIST]
89.74.2 [NIST]
CH
152.2
-10 [ICSC]
100.7 (at P=0.01 bar) [NIST]
1.064 1.062 [3]
1.05 [ICSC]
79 [ICSC]
1.5[ICSC]
1.5245-1.5242 [3]
32 Pa [ICSC]
0.9-6.5 [ICSC]
99.7-100 [3]
108 calc. on data [3]
-148.5 (liq) [2]
-1496.7 (liq) [NIST]
72.8 calc. on data [3]

69.90.2 [NIST]
-
There had been some incidents involving the both of peroxides, which described in
literature [5-8]. An extremely violent explosion occurred in the facility of acetone and phenol
production at distillation. The root and contributing causes of this incident were sharp rising
of temperature and thermal decomposition of CH, which were identified at the investigation
on the incident. Near-miss incident occurred at streaming of vessels contained the leavings
of CH. Much more incidents with explosion occurred at BP transportation, drying, pouring
into another container, and even at sweeping with broom on the floor. Some of these
incidents caused fatalities and subsequent to them evaluated the ability either of BP or CH to
detonate.
Had the explosion properties of peroxides been more thoroughly investigated, they could
have provided insight into the hazards associated with the operations to issue
recommendation to help prevent similar occurrences. To achieve accident prevention goal all
of the information about heat explosion, burning, and detonation of the peroxides should be
compiled, analyzed and updated before initiating design and construction some technological
processes and equipment.
We used the complex investigation methodology which had been created in Mendeleev
University of Chemical Technology, including calculation of the explosion parameters (heat
of explosion, detonation velocity etc.) and experimental measuring of some parameters
(velocities of burning and detonation, temperature of decomposition beginning, etc.) [9-12].
174
2.
EXPERIMENTAL
We had at our disposal the commercial grade product are named in Russia Hyperiz.
The product was sertificated according technical conditions of Russia 38.402-62-121-90
and contained net cumene hydroperoxide not less than 89%.
Decomposition of CP under heating was investigated in laboratory apparatus, consisting
from two test-tubs, one of them (d=13 mm) was put into another (d = 30 mm). The substance
under investigation (m=1 g) was placed into inner test-tube and was heated by electric spiral,
which was wound on the external test-tube. The temperature of the sample was measured by
thermocouple device, including the working and compensating thermocouples, variable
resister, and potentiometer for registration the curve temperature vs. time. The working
thermocouple was protected by the mean of quartz capillary. The velocity of heating (20-50
degrees/min) was regulated by mean of tension current on the spiral.
Burning of CH was investigated in the bomb of constant pressure in atmosphere of
nitrogen over the pressure range 24 to 36 MPa. The same micro-thermocouple methodology
as in [4] was applied for measuring the temperature profile at burning of CH. The burning
rate versus pressure was measured by the mean of drum photo-register device in passing
through light, because the own luminescence of CH burning was too weak to be registered.
The transparency of CH (colorless liquid) and products of its burning (light-brown liquid)
were quite enough contrast to register the board between them by the photo-register device.
Experiments investigating detonation ability of CH were carried out in steel tubes of
10 mm inner diameter, 13mm wall thickness, length l=250 mm, and 9 radial holes d=2 mm
were drilled in wall. The diameter of the tube channel in upper part was enlarged to 14 mm
to place the booster. The pressed pellets of phlegmatized RDX (d=12 mm, m=2 g, density
=1.67-1.68 g/cm3) were used as powerful initiator. Mixtures of RDX/NaCl of different
composition [13] with PMMA attenuator between the bottom of the initiating charge and the
surface of a liquid were used as donor charges for determination of critical pressure of
initiation of substance under study. The bottom of tube was covered by mean of a resin plug.
The experiments were carried out either without photoregisration or with it. In the first case
the radial holes were glued up by a scotch, and a metal witness-plate was fastened to them.
The wood plate was arranged between witness-plate and massive steel plate, on which the
construction was fastened, to prevent intense deformation of a witness-plate at detonation.
The character of deformation of a tube and perforation of a witness-plate gave an
information about stability of a process. The high velocity detonation (3-7 km/s) destroyed
the steel tube on fragments. The low velocity detonation (LVD) didnt crash the tubes, but
perforated the witness-plate uniformly, if the process propagated without damping. At
damping of low velocity detonation process a degree of perforation diminished to the end of
a tube.
Russian streak camera GFR-3 was employed to measure detonation velocity. As at
burning of CH, the luminosity at detonation was too weak for registration by means of
photography, and to develop the process propagation on streak camera record we used
modifying technique, suggested in [13,14]. The radial holes were cowered by paraffin plugs
near channel of tube, the length of plugs was the same in all holes (8 mm). Small amount
(~0.02 g) of lead azide was promptly placed into each hole on paraffin plug, and then the
holes were glued up with a scotch. Luminosity of lead azide flashes fixing in streak camera
record helped to judge about a stability of process, and to measure its velocity. As it had
been shown in works [13,14], the line connecting points-flashes of lead azide in radial holes on
a streak camera record was a straight one, if low velocity process was not damp. If a
propagating process damped the slope of line diminished to the end of a charge.
175
3.
EXPERIMENTAL RESULTS
The typical curves temperature versus time being registered at heating of both CH and
inert liquid (sulfuric acid for comparison) are shown in Fig.1. There were no any visible
qualitative changes during approximately linearly law of CH heating. The points in these
curves mark the moment of very beginning of decomposition reactions of CH producing
visible bubbles of gases or vapors. This moment coincides with acceleration of temperature
increasing. Gasification (pseudo-boiling) intensified at self-heating of CH, the droplets of a
decomposition products condensed on the wall of test-tube and flew down. At
decomposition passing primary colorless substance became a light-brown one. A flame has
never appeared at CH decomposition under these conditions. The induction periods and the
temperatures of decomposition beginning at various velocities of heating are presented in
Table 2.
Fig 1.
The experimental curves of CH heating (1,2), the points mark beginning

of decomposition. 3 and 4 are heating of inert substance (sulfuric acid).
Table 2. Temperature and time of beginning CH decomposition

at various velocities of heating.
Velocity
heating,
deg/min
of Beginning of decomposition
Temperature, 0 C
Time, min
19
145
10,3
21
159
7,2
28
142
5,2
32
141
4,2
39
160
4,4
40
157
4,2
50
160
176
Stable burning of CH propagated in quartz glasses by a diameter of 12 mm at pressure

p=24 MPa and higher. The results of runs are shown in Fig. 2, as the dependence a velocity
(u, mm/s) vs. pressure. Examples of the photoregister records (positive) are reprinted in Fig.
3 and 4. Slow oscillation of burning surface could be seen in the records at P=24 MPa
(u=1.28 mm/s) and at P=27 MPa (u=2.12 mm/s). There was no noticeable producing of
gases at burning of CH.
Fig 2. The burning rate of CH versus pressure.
Fig 3.
The photoregister record of CH burning at P=24 MPa, u=1.28 mm/s.
Fig 4.
The photoregister record of CH

burning at P=30 MPa, u=2.52 mm/s.
177
The wolfram-rhenium micro-thermocouples (W+20% Re and W+5% Re, thickness of

wires was 50 and after rolling was 20 micrometers) were used to register the profile of
temperature at burning of CH. The original curve T vs. distance, which was experimentally
fixed, is shown in Fig.5. The maximum of temperature Tm= 5550 C was found to situate
practically on the burning surface.
Fig 5.
Fig. 5. The original curve T vs. distance which was fixed at P=30 MPa.
The maximum of temperature is Tm= 5550 C.
Three runs without photoregistration were carried out at powerful initiation with liquid
CH in steel tubes described above. The middle and final parts of steel tubes were not
fragmented the witness plates were uniformly perforated. The explosion was accompanied
with a prolong sound, and the part of unreacted CH was found in explosion chamber. These
circumstances testified that the results of all of these experiments could be identified as low
velocity detonation: the same signs were observed at low velocity detonation of pressed,
cast, and liquid nitrocompounds [13-16]. The experiments with photo-registration confirmed
the first conclusion. The results of detonation velocity measurement are collected in Table 3.
Table 3. Results of measuring CH detonation velocity in steel tubes
at various initiating pressure
Initiator
Composition, %
Result of an experiment
,
g/cm3
Pin,
GPa
D, km/s
Characteristic
RDX
NaCl
100
(pressed)
1.661.67
~30
1.97
Stable LVD
100
1.34
18.2
1.87
Stable LVD
50
40
50
60
1.22
0.98
4.3
1.5
1.871.50
1.931.46
Damping LVD
Damping LVD
178
Fig 6.
The streak camera record of CH stable low velocity detonation. Pin30

GPa. D=1.97 km/s.
The initiating pressure (Pin) was verified as stated above. At the powerful and less
powerful initiating (Pin~30 GPa and Pin=18.2 GPa) detonation process propagated with
constant velocity 1.87 and 1.97 km/s correspondingly, one of the streak camera records
(positive) is preprinted in Fig.6. At the lower initiating pressure (4.3 and 1.5 GPa) low
velocity detonation damped to the end of tube, the slope of a lines connecting points-flashes
of lead azide in a records diminished to the bottom of tube. The streak camera record of
damping low velocity process is shown in Fig.7, a flashes of some samples of lead azide
sometimes in these cases were not fixed, i.e. they were not initiated. The lead azide in the
last radial hole was found out after one of the experiments at damping LVD of CH.
Fig 7.
The streak camera record of CH damping low velocity detonation.

Pin1.5 GPa.
D=1.93 km/s in the top of tube, and D=1.46 km/s in bottom part.
The flash of lead azide in the sixth hole from the top of tube is absent.
179
4.
DISCUSSION
The decomposition reactions occurring in liquid CH according to literature [3] are:

1. The principle decomposition products are acetone and phenol:
C6H5(CH3)2COOH C6H5OH + CH3COCH3 + 309.6 kJ
2. The principle decomposition products are methanol and acetophenon:
C6H5(CH3)2COOH CH3OH + C6H5CO CH3 +230.1 kJ
As one can see the both of the reactions are characterized by release of energy. The
decomposition rate constant of CH is K=0.0108 hour-1, activation energy is E=100 kJ/mol [3].
Pre-exponential factor may be estimated taking into consideration these data:
B=K/e-E/RT 108 c-1.
Taking into account an average value of decomposition reaction heat to between
reactions 1 and 2 Q=1674 kJ/kg, one can compute the self-heating temperature using the
transcendent expression of the heat explosion theory at heat exchange by convection [17]:
Ti = E/{R ln(v d B E Q/S Nu R Ti2 c)]
Here:
Ti is the self-ignition temperature,
E, B, and Q are activation energy, pre-exponential factor, and decomposition reaction
heat, correspondingly.
V and S are the volume and surface of vessel. It was suggested to be a sphere d = 1.5 cm.
and c are coefficient of temperature conductivity and heat capacity,
Nu Nusselt`s criterion.
Estimation gave the value of self-heating temperature Ti = 1580 C, which practically
coincided with beginning of temperature arising in our runs (see Table 2).
The sample of CH by several grams weight at heating decomposed with temperature
arising and self-acceleration but without appearance of flame in the laboratory experiments.
An intensity of heat explosion is well known to depend on mass of explosives under heating.
Some of nitroesters (mannitolehexanitrate, nitroglycerine) at self-heating even at weight of
0.2-0.3 grams fully crashed the glass experimental apparatus. Probably heat explosion of CH
at large amount should propagate more intensively. The incidents occurring at some facilities
involving CH, e.g. at hot streaming of vessels contained the leavings of CH, make this
assumption objective.
The computer code SD [18] was used to calculate the detonation parameters of CH, and
method [19,20] was used to estimate its explosion temperature. The reaction products
according the both of methods were determined by reactions equilibrium including
hydrogen, carbon, and products of its oxidation. The calculating parameters of CH are:
explosion heat Q = 3.13 MJ/kg, detonation temperatureTD = 1680 K, detonation velocity
D = 4.18 km/s, temperature of explosion (at P = 30 MPa) Tp = 960-980 oC. Calculated
principle products are: condense carbon, CH4, CO2, CO, H2O, H2. However we didnt
observe the any amount of carbon or gases created at burning in bomb of constant pressure,
the quartz glass to the end of burning was filled up by light brown liquid. The experimentally
measured temperature of burning was sufficiently lower than computed temperature Tp.
These facts testify to leading reaction occurs in condensed phase. The same conclusion was
made earlier at investigation of benzoyl peroxide burning [4].
180
If the proposed model is true, one could assume that overall reactions occurring at
burning of CH are the same ones at the decomposition, i.e. reactions (1) and (2). The
reaction temperature was calculated at c=2.1 kJ/kg K. At heats of reactions (1) and (2),
estimation gave the temperatures: T1=7200 C and T2=4770 C. One may conclude that
occurring of exothermic reactions (1) and (2) can ultimately increase the temperature to
555 0 C, which was fixed at burning experimentally.
The burning and low velocity detonation process of CH were observed in the
experiments for the first time. The burning was found to propagate at pressure P24 MPa.
However, luminosity of flame was too weak to be registered by mean of photography, it
could be explained taking into consideration the low temperature of burning and absence of
carbon, water, and carbon dioxide in principle burning products. These products could
provide to the flames intensive self-luminosity but they are not created as a result of reaction
(1) and (2).
CH detonation was observed to propagate in steel tubes as low velocity detonation
process D2 km/s. The minimal shock wave pressure to initiate the stable low velocity
detonation of CH was Pin=18.2 GPa, at Pin=4.3 GPa the process damped. One must note that
at D2 km/s the detonation pressure is approximately P1 GPa, in other word the destruction
effect of the explosion can be rather high.
5.
CONCLUSION
The data of our investigation in couple with literature materials characterized cumene
hydroperoxide as weak explosive, which is able to burning, low velocity detonation, and heat
explosion. The results of our investigation update the information concerning explosion
characteristics of cumene hydroperoxide and they could be used to achieve accident
prevention goal on the plant facilities at its production, processing treatment, and
transportation.
Acknowledgment
We are grateful to V.Yu. Egorshev, helped us to carry out the investigation of CH
burning.
181
REFERENCES
[1] K.K Andreev, A.F.Belyaev: Teoriya vzryvchatykh veshchestv (Rus). M., Oborongiz, 1960, 595
p.
[2] D. Stal, E.Westram, H.Zinke: Khimicheskaya termodinamika organicheskikh soedinenyi (transl.
into Rus.), M., Mir, 1971, 312 p.
[3] B.D.Kruzhalov, B.I.Golovanenko: Sovmestnoe Poluchenie fenola i atsetona (Rus.), M.,
Goskhimizdat. 1963, 187 p.
[4] A.E.Fogelzang, V.Ya.Adzhemyan, Yu.A.Pimenov, A.R.Saklantii, T.V.Doronina: The
investigation of benzoyl peroxide burning. Voprosy Teorii Cond. Vzryvchatykh System (Rus.).
Mendeleev Institute of Chemical Technology, Moscow, 1980, v.112, 67-70.
[5] B.V.Beschastnov: Avarii v khivicheskikh proizvodstvakh i mery ikh preduprezhdeniya (Rus.)
M., Khimiya, 1991, p. 445.
[6] B.V.Beschastnov: Promyshlennye vzrivi. Otsenka i preduprezhdenie (Rus.). V., Khimiya, 1991
387 p.
[7] V.L.Antonovskii: Organicheskie perekisnye initsiatory (Rus.). M., Khimiya, 1972.
[8] V.C.Marsshall: Osnovnye opasnosti khimicheskikh proizvodstv (trans. Into Rus. from Engl.),
M. Mir, 1989, p. 671.
[9] .N.Kondrikov, G. D.Kozak, V. M. Raikova. et al: Explosion hazard of some liquid
nitrocompounds (Rus), Khim. Prom., No. 5, , 1990, 26-29 .
[10] G.D.Kozak, V.M.Raikova: Energetic characteristics of amonium peroxocompounds of niobium
and tantalum. Trans. of the 26-th Int. ICT Conference, Karlsruhe,FRG, 1995
[11] G.D.Kozak, V.M.Raikova: Otsenka vzryvoopasnosti reaktsionnykh mass stadii nitrovaniya
acettoluidida (Rus.). Khim Prom. 1992, N 2, 15-18.
[12] G.D.Kozak, B.N.Kondrikov: Detonation of Nitrobenzene and Propargyl Alcohol. Combustion,
Explosion and Shock Waves v.35, 1, 1999, 80-87.
[13] V.F.Tyshevich: Investigation of low velocity detonation initiation and propagation of cast and
pressed explosive charges in strength confinement. Cand. Sci. (Ph. D.) Thesis (Rus.).
Mendeleev University of Chem. Technol. Moscow, 1992.
[14] I.V.Babaitsev,. B.N.Kondrikov,. Z.V Paukova, V.F Tyshevich: Detonation of cast explosives
with a low speed, Comb., Expl. & Shock Waves, 1969, v.5, 3, 326-330.
[15] G.D.Kozak, V.V.Kondratiev, B.N.Kondrikov, A.V.Starshinov: Low velocity detonation of
nitromethane, Detonation. Institute of Chemical Physics of Akad. Sci. USSR, Chernogolovka,
1978, 34-37.
[16] G.D. Kozak, B.N. Kondrikov, A.V. Starshinov: Critical conditions of low and high velocity
detonation regimes in liquid nitrocompounds, Proc. Of the 11th Int. Detonation Symposium,
Snowmass, Colorado, 1998, Bookcomp, Ampersand, 2000, 81-86.
[17] B.N.Kondrikov: Ignition and Combustion of Nitrocompounds (Rus.), Text-book, Mendeleev
Institute of Chem. Technology, 1985, 80 p.
[18] B.N.Kondrikov, A.I.Sumin: Equation of State of Gases at High Pressure. Comb., Expl. & Shock
Waves, 1987, vol .23, N 1, 114-123.
[19] B.N.Kondrikov: Chemical Thermodynamics of Combustion and Explosion (Rus.), Text-book,
Mendeleev Institute of Chem. Technology, Moscow, 1980, 80 p.
[20] B.N.Kondrikov, G.D Kozak, V.N Shapoval: Problems Book on Combustion and Explosion
(Rus.) , Mendeleev Institute of Chem. Technology, 1981, 48 p.
182
INVESTIGATION OF THE WETTABILITY

OF AMMONIUM NITRATE PRILLS
Queenie S.M. Kwok*, David E.G. Jones* and Peeter Kruus**
*Canadian Explosives Research Laboratory, 555 Booth St., Ottawa, ON K1A 0G1 Canada
**Carleton University, 1125 Colonel By Dr., Ottawa ON K1S 5B6 Canada
Abstract:
The wettability of various types of ammonium nitrate (AN) prills were compared using
capillary penetration measurements where the penetration rates of various alkanes and
fuel oil into packed columns of AN samples were determined. Complementary
characterization studies using scanning electron microscopy (SEM) and
thermogravimetry (TG) were performed to rationalise the observed differences in
wettability. The wettability of AN was found to be affected by several factors, including
surface tension, viscosity, density, purity, polarity and/or polarizability of liquid used, as
well as surface composition, porosity, bulk density, particle size and moisture content of
the AN samples.
High-resolution TG (HR-TG) was used to study the thermodesorption of octane from the
various AN prills. The samples were wetted by immersing in octane. Multiple steps were
obtained from the measured mass-loss curves, and they reflect the evaporation of the
excess liquid, as well as the thermodesorption of octane from the pores and the surface
of the AN prills. The quantity of octane desorbed in these steps was correlated to the
volume in the pores and the amount adsorbed on the surface, and it was used to estimate
the porosity and surface area of AN prills.
Keyword:
1.
ammonium nitrate prills, ANFO, wettability, capillary penetration,

thermodesorption, high-resolution, TG
INTRODUCTION
ANFO, an admixture of ammonium nitrate (AN) prills with fuel oil, is widely used as a
commercial explosive. The wettability of AN prills is one of the primary factors determining
the physical stability and detonation behaviour of ANFO [1]. However, there is limited
published information on the surface and interfacial properties of AN and fuel oil.
There are several strategies for quantitative assessment of the wettability of solid
materials, however, most of these techniques are not applicable to powders and porous
materials [2]. Thus, capillary penetration technique has been developed and applied to
determine the wetting characteristics of a variety of powders or porous materials, such as
pharmaceutical products, minerals and polymers [2].
The capillary penetration technique is based on the rate of liquid penetration into packed
beds of particles. The basic equation that describes the kinetics of penetration of liquids into
powder or porous bodies was developed by Washburn [3] and modified by Grundke et. al [4]:
M2 =
K 2 lv cos
t
2
(1)
183
where M is the mass of the liquid penetrated into the capillary system at time t, K is the
geometric factor of the packed beds of particles, is the density of the liquid, lv is the
surface tension of the liquid, is the contact angle and is the viscosity of the liquid. Eq. (1)
neglects the effects of slip, gravity and inertial effects, and assumes there are no external
pressure gradients.
In terms of the porous structure of AN prills, there is scarce information available in the
literature. At present, there are various methods for characterizing porous materials;
however, among these methods, only electron microscopy and mercury penetration are
suitable for describing the pore size range of AN prills (> 1 m). Mercury porosimetry has
been shown to incorrectly assess the overall pore geometry and the distribution of pores in
AN prills [5]. Scanning electron microscopy (SEM) has been used to obtain a twodimensional view of pores [6,7]. However, SEM is only able to sample surface pores and a
small fraction of the area of the sample. Although SEM is insufficient for evaluating the
percentage of the pore structure in the entire sample, it is a very important tool and gives
valuable insight into the porosity of the area sampled. 3-D nuclear magnetic resonance
(NMR) microscopy has also been applied to evaluate the porous structure of AN [5].
However, the NMR technique involves a relatively elaborate and time-consuming procedure.
High-resolution thermogravimetry (HR-TG) is useful in monitoring the stepwise
thermodesorption of adsorbed molecules from porous solids [8,9]. The technique can be used
to characterize surface and structural properties of various materials [1012]. The HR-TG
results agree with those obtained from nitrogen adsorption measurements [10,11] and
differential scanning calorimetry (DSC) [12]. In contrast to adsorption measurements, TG is
more flexible with respect to the adsorbate used, and the reagents are inexpensive and easier
to handle than mercury. Given these advantages, TG becomes a valuable alternative for the
routine measurement of the porosity of various materials.
This research explores the possible application of the capillary penetration
measurements to determine the wettability of AN prills with various hydrocarbons and fuel
oil, and subsequently the wettability of the AN prills from different sources. Complementary
characterization studies using SEM and HR-TG were performed in an attempt to rationalise
the observed differences in particle wettability.
2.
EXPERIMENTAL
2.1
Materials
2.1.1. Ammonium Nitrate

Five samples of AN prills from three different categories were used in this study. They
are identified as Types AE. Their categories, bulk density, void fraction, prill size range and
prill size distribution are summarized in Table 1. The bulk density, which was calculated
using the sample mass and volume in a cylinder, refers to the density of the prills, including
pores and inter-prill spaces. The void fraction was determined by the volume of octane
added to a cylinder packed with AN prills. The size distribution of the samples was
determined by sorting the prills through sieves with various sizes. For the HR-TG studies,
only prills within the diameter range specified in Table 2 were used. The apparent volume
(Va), which was calculated using this average prill diameter, refers to the total volume
occupied by 1 g of prill sample, excluding inter-prill spaces. All the prills used in this study
are commercial AN prills with additives and/or coating agents (e.g. kaolinite, talc). The type
and the amount of additives and/or coating agents, as well as the previous thermal history of
184
each individual sample is not known. Unless specified otherwise, the samples were used as
received. All the prills were stored under identical conditions before measurements.
2.1.2. Liquids
A series of pure organic liquids was used as wetting liquids, as listed in Table 3. Except
for fuel oil (diesel fuel), all other liquids were reagent grade with a purity of at least 98 %.
The suppliers are listed in Table 3. The fuel oil was obtained from a local gas station. All
chemicals were used without further purification. For the various wetting liquids, their
surface tension (lv), density () and viscosity () measured at 23 C are listed in Table 3.
The surface tension was measured using a CENCO-DuNOUY Interfacial Tensiometer
(Central Scientific) and the viscosity with a Cannon-Fenske Routine Viscometer (Industrial
Research Glassware Ltd.). The density was determined by weighing 10 mL of the
corresponding liquids.
2.1.3. Sample Preparation
AN powders were obtained by grinding Type A AN prills using an IKA model A10
analytical mill. Moisturized AN prills were prepared by placing Type A prills under a bell jar
with a pot of water. Samples with different moisture content were obtained by exposing the
samples to water vapor for 1.5 to 24 hours. The relative humidity (R.H.) of air in the system
was 7080 % as measured by a humidity meter. The moisture content of each sample was
verified using TG. For HR-TG studies, wetted AN prills were prepared by immersing them
in octane for at least 1 day.
2.2
Capillary Penetration Measurements
The approach is based on measuring mass as a function of time of a liquid wicking up

through a packed bed of particles contained in a tube. The experimental setup for a capillary
penetration measurement is shown in Fig. 1. About 5.0 g of AN samples are packed in a
glass tube. The inner diameter and the length of the glass tube were 1.08 cm and 15 cm,
respectively. The lower end of the tube was closed by a stainless steel filter, which was cut
from a coffee filter. The column packed with samples was then dipped in a glass dish
(volume of 50 mL and diameter of 10 cm) filled with the appropriate testing liquid. The
penetration rate of the liquid into the samples was determined by monitoring the column
mass gain as a function of time. Depending on the
nature of both sample and liquid, an experiment
Balance
took from 30 to 140 min.
An essential requirement for reproducible
measurements is the uniform and reproducible
packing of the tube with AN particles. The
packing procedure was run as follows: the tube
filled with the 5.0 g of sample was first dropped 20
times from a height of about 20 cm, and then was
vibrated for 5 min with an electric vibrator
(constructed from a shaver). The time required to
pack the tube was set at 5 min, since after 4 min,
the height of the samples was constant. Prior to
each run, the packing constant (bulk density) was
calculated, to ensure a consistent packing for each
type of sample. Temperature as well as R.H. in air
near the column were recorded to make sure that
their variations were limited, i.e., 23 3 C and 39
glass tube
sample
liquid front
dish
containing
testing
liquid
filter
lab jack
Fig 1. Setup for capillary penetration

technique
185
18 % R.H. Although the variation of R.H. was large in the whole duration of the study, it
was noted that there is no significant difference in the results for the same sample at 21 and
60 % of R.H.
2.3
Thermogravimetry measurements
The water content of the as-received and as-prepared AN prills was established by a TA
Instruments 2950 High-Resolution Thermogravimetric Analyzer (HR-TGA) accompanied
with a TA 5200 Thermal Analysis System. The sample was heated from 25 to 110 C at 5 C
min-1. A platinum pan containing about 30 mg of AN prills was purged with dry nitrogen at a
flow rate of 60 mL min-1 in the furnace and 40 mL min-1 in the balance chamber.
For the thermodesorption measurements on various AN prills, the dynamic rate high
resolution technique was used. This technique allows the heating rate to be varied between a
fixed minimum and a maximum specified, in this case 5 C min-1. The resolution (R) and
sensitivity (S) parameters were 4 and 6, respectively. These parameters were established
experimentally to provide the optimal heating rate for thermodesorption.
Thermodesorption of octane from the AN prills was monitored over a temperature range
from 25 to 120 C. An open platinum pan was used for all measurements. For each run, 3
(Sample A and E) or 15 (Sample D) wetted prills were removed from the octane and placed
in an unsealed DSC aluminium pan. This pan was used to minimize the evaporation of
octane prior to the experiments. For experiments with no excess octane, the prills were
placed in the pan using tweezers, instead of a spatula. The mass of the sample varied
between 2030 mg depending on the prill density and the amount of octane adsorbed. To
confirm the baseline of the TG, an experiment using the same heating profile was performed
on an empty pan. The TG drift was ~ 5 g, which represents 0.02 mass % for 2030 mg of
sample.
2.4
Scanning Electron Microscopy
The external and internal structure of various AN prills were examined by SEM, model
JSM 6400 (JEOL Ltd.). The external structure of the prills was examined on unbroken prills,
and the internal structure was observed in cleaved prills. These were prepared by cutting
prills into halves with a sharp blade. The samples were mounted on specimen studs using
double-sided adhesive tape, and were coated with a gold-palladium alloy. The SEM
observations were conducted at low accelerating voltage (3 or 8 kV) to minimize the beam
damage to the samples.
An energy-dispersive (ED) X-ray spectrometer attached to the SEM was used to
chemically analyze small selected areas of various AN prills. Al K X-rays were used as an
X-ray source in a vacuum of 1.3 10-7 Pa. For the X-ray microanalysis, the specimen studs
mounted with samples were coated with carbon.
186
3.
RESULTS & DISCUSSION

3.1
Wetting Rate of Various Liquids with AN
A typical result obtained from a capillary rise measurement for an AN packed column is
illustrated in Fig. 2, where M2 of the column is plotted versus t. According to the modified
Washburn equation (Eq. (1)), M is predicted to be a square root function of t. However, all
the results obtained in this study show a parabolic increase of M2 with time during the initial
period. In some experiments, downward deviations from the straight line were also observed
at longer times near the end of the runs. The non-linearities in the experimental M2 vs. t
relationship may be due to the inappropriateness of the assumptions in the Washburn
equation.
The effect of inertia is considered negligible, and laminar conditions are assumed to set
up very rapidly in the Washburn equation. However, Qur [13] has shown that at the early
stage, the capillary penetration is governed by inertial forces, and the transition time from an
inertial to viscous regime is given by = R2/4. In this study, the initial unsteady state is
observed for all the liquids tested, and the extent of this domain seems to depend on the
liquid. However, the observed deviations do not agree with the above relationship of
transition time. Apparently, the deviations at the early stage of the experiments may also
depend on other factors, e.g. mass gain data may be affected by the liquid wicking on the
internal wall of the glass tube.
Furthermore, the Washburn equation does not include the effect of gravity, which tends
to decrease the forwarding of the liquid. An increasing height of the liquid front increases the
influence of gravity, leading to a deviation of the experimental behaviour. According to
Washburn [3], the gravity effect can be neglected until h 0.1 Heq. The maximum height (Heq
= 2lvcos/Rg) of the liquid front reached in infinite time, when capillary and
hydrodynamic forces are equal at equilibrium is controlled by gravity [3]. In this work,
deviations were not observed in most of the experiments, in which the height of the
penetrated liquid at the end of each run was usually < 4 cm. Small deviations downwards
from the straight line were observed near the end of some experiments, which imply that
gravity needs to be taken into account when the liquid height is greater.
Mass loss due to the evaporation of
the adsorbed liquids on AN samples
may also cause deviation from the
Washburn equation. However, in this
study, the samples were contained in a
glass tube and the liquid vapour can
only escape from the system through
the open end of the tube (Fig. 1). Since
the vapour of the most volatile liquids
could be re-adsorbed by the prills on
the top, the evaporation effect of the
testing liquid should be negligible.
In each experiment, the wetting
rate (M2/t) is estimated from the slope
of a linear regression on the selected
(M2, t) data domain (Fig. 2). This
domain is selected from the region of
Fig 2. Capillary penetration results for AN

prills with heptane
187
the curve where its tangent has the least slope variation, i.e., where the effects of inertia and
gravity seem to be negligible.
Capillary penetration measurements on Type A AN prills were conducted with various
testing liquids, including a series of alkanes and fuel oil. The average values of the wetting
rates determined from two to three replicates are summarized in Table 3. The error term of
the values represents one standard deviation determined from the repeated experiments.
As compared in Table 3, the wetting rates of the alkanes decrease with increasing lv,
and of the liquids, in agreement with the Washburn equation. However, the results also
suggest that fuel oil has a higher wetting rate on AN, while its lv, and are also higher
than dodecane. Since fuel oil is a mixture of alkanes and aromatic hydrocarbons, the higher
wettability of fuel oil with AN may be due to the higher attractive interaction of some of its
components with the AN surface. On the other hand, a poorer reproducibility of wetting rates
was also obtained from fuel oil (error 28 %) than from alkanes (error 10 %). The
variations in the wetting rate data may be due to non-uniform and preferential penetration of
fuel oil. Each of the components has different penetration rate, and the composition of the
fuel oil may vary as it rises. The poor reproducibility and variation in wetting rate indicate
the potential problems during wettability testing with mixtures. It is thus important to use
pure probe liquids in the capillary penetration measurements.
The main deficiency of the model derived from the Washburn equation is that only a
single cylindrical capillary is used to characterize a complex porous network. Hence, the
method cannot be used to predict the exact process of the liquid penetration in AN prills.
However, given the simplicity of the capillary penetration technique, it can be used as a
valuable alternative to estimate the wetting rate. The obtained wetting rates serve as an
empirical measure to compare the relative wettability of various AN prills.
3.2
Wettability of various AN Prills
The wetting rates with heptane of the as-received AN prills were measured using the
capillary penetration technique. As seen in Table 1, there are significant variations in the
wettability of the various AN samples. Among the five samples, Type A AN prills have the
highest wetting rate with heptane. For Types B and C prills, the wetting rates are similar, and
lower than that for Type A prills by about 13 %. The wetting rates for Types D and E prills
are much lower than those for porous prills (Types AC); they are respectively 55 % and 75
% lower than that for Type A prills.
The wetting process for AN is a function of the solid and liquid characteristics, as well
as the solid-liquid interactions. Since the wetting rates of the five samples were measured
using the same wetting liquid, the variations in their wettability are mainly due to the
differences in their characteristics.
3.3
Effect of Morphology and Porosity
3.3.1. SEM
SEM was used to determine the external and internal surface morphology for the various
AN prills. As presented in Fig. 3, the external surfaces of Types A, B and C AN prills are not
smooth, and they have a non-uniform distribution of pores over their external surface.
Comparing the three porous prills (Fig. 3.ac), there are slight differences in their surface
openings. Type A prill has a channelled pore structure (Fig. 3.a). Types B and C have
characteristic growth on the external surface that covers the pore structure (Fig. 3.b and c).
188
As depicted in Fig. 4.a, all the AN porous prills (Types AC) have an internal central
hollow. Their internal structure, as shown in high magnification micrographs Fig. 5, indicate
that there are differences in the particle shape. Type A prill has a pore structure in which the
grains are interconnected with themselves (Fig. 5.a). Type B prill has an internal structure
composed of nodular shape (spheroid) particles with an irregular distribution of pores (Fig.
5.b), whereas Type C prill consists of oriented dendritic particles with channel pores (Fig.
5.c). The SEM micrographs suggest that the external and internal pore sizes of porous prills
range from 10 to 80 m.
(a)
(b)
(d)
(e)
(c)
Fig 3. SEM micrographs on the external surface of AN prills (all 150): (a) A; (b) B; (c)
C; (d) D; (e) E.
(a)
(b)
(c)
Fig 4. SEM micrographs on the internal surface of AN prills: (a) AC (35); (b) D
(70); (c) E (30).
(a)
(b)
(d)
(e)
(c)
Fig 5. SEM micrographs on the internal surface of AN prills (all 150): (a) A; (b) B; (c)
C; (d) D; (e) E.
189
Typical SEM micrographs of Types D and E AN prills are also presented in Figs. 35.
The external morphology of Types D and E prills is shown as a highly dense structure with
very few surface pores in Fig. 3.d and e. For Types D and E prills, although the central
hollow is distinctly formed (Fig. 4.b and c), there are no channels linking it to the external
surface. From the general internal micrographs of Types D and E prills (Fig. 5.d and e), it
can be observed that the prills have a continuous solid mass which is interrupted by a few
discrete pores. The approximate maximum pore sizes for Types D and E prills are 8 and 10
m, respectively.
Both the external and internal surfaces of Types D and E prills are much finer than the
porous prills (Types AC). The pore sizes for Types A, B and C prills are 10 times larger
than those for Types D and E prills. The relatively high porosity of Types A, B and C prills
contribute to their enhanced wettability (see Table 1). Among the porous prills, Type A prills
have the highest wettability, which may be the result of a higher porosity. However, SEM is
only able to sample surface pores within a small viewing area, and it is insufficient for
evaluating the percentage of pore structure.
3.3.2. HR-TG
The HR-TG mass loss curves were measured for AN prills A, D and E, which had been
immersed in octane. Fig. 6, where the ratio of sample mass to the mass of neat AN prills are
plotted versus temperature, shows the typical mass loss behaviour for AN Sample A. Fig. 7
shows the derivative of mass loss with respect to temperature. The TG results show that
octane desorbs from the AN samples via a multi-step process. In order to identify each step,
experiments were conducted using different conditions. In Figs. 6 and 7, the three TG (or
DTG) curves represent the results obtained for: (a) as-received prills, (b) prills wetted with
octane, (c) prills wetted with excess octane.
Comparison of the various TG curves suggests that the initial step is the mass loss due to
the evaporation of the bulk liquid. The subsequent step represents the thermodesorption of
adsorbed octane on the surface of AN remote from the monolayer, which is in direct contact
with the AN surface. The last step reflects a gradual thermodesorption of adsorbate in the
pores and of the monolayer of octane on the AN surface.
The properties of surface liquid differ significantly from those observed for bulk liquid
Fig 6. TG curves for desorption of octane

from AN (R = 4; S = 6). (a) asreceived prills; (b) prills wetted
with octane; (c) prills wetted with
excess octane.
Fig 7. DTG curves for desorption of octane

from AN (R = 4; S = 6). (a) asreceived prills; (b) prills wetted with
octane; (c) prills wetted with excess
octane.
190
because adsorbate molecules interact with the solid surface. The adsorbate requires higher
energies to thermodesorb from the solid surface and the pores. These properties change with
distance and approach those of bulk liquid at relatively long range.
From the TG results for Samples A, D and E, the total adsorption capacity (nt) was
estimated from the total mass loss of the wetted samples. The values of nt for the various AN
prills are summarized in Table 2.
To evaluate the pore volume and the specific surface area of the various AN prills, a
higher HR-TG setting (R = 5 and S = 8) was applied to further study the thermodesorption of
octane from wetted AN prills.
As shown in Fig. 8, the TG curve has a sharp initial step (reflecting the
thermodesorption of surface liquid) and two subsequent steps for the thermodesorptions of
the pore liquid and of the monolayer. However, the transition between thermodesorption of
the pore liquid and monolayer is not easily distinguished. The minimum of the derivative of
the time-dependent TG curve (see Fig. 9) can be used to localize the characteristic point on
the mass loss curve [10].
As illustrated in Fig. 9, the differential
DTG curve plotted against time shows two
local minima (denoted by I and II). The
mass loss estimated between the projection
of the points I and II on the TG curve
represents the thermodesorption of octane
inside the pore. In Table 2, the results are
expressed as pore capacity (np), in
millimoles of thermodesorbed octane per
gram of the dried AN sample.
The total volume of pores (Vp) is the
product of pore capacity and the molar
volume of liquid octane. The porosity (), Fig 8.
which is defined as the ratio of the total pore
volume (Vp) to the apparent volume (Va) for
1 g of AN prills, is also included in Table 2.
TG results for AN prills wetted with

octane (R = 5; S = 8). Solid line: mass
ratio (m to mass of AN); dashed line:
derivative of mass with temperature.
The projection of the point II and the

residue point III on the TG curve gives the
monolayer capacity (nm), which can be
converted to the specific surface area (a) of
the sample assuming that the cross-sectional
area () of the octane molecule is known:
a = n m L
(2)
where L is Avogadros constant and is the

average area occupied by each molecule in
the completed monolayer.
A molar volume of 163.6 cm3 mol-1 and Fig 9.
a cross-sectional area () of 4.2 10-19 m2
molecule-1 were used for octane to calculate
the volume and specific surface area of the
various AN prills. The molecular cross-
TG results for AN prills wetted octane

(R = 5; S = 8). Solid line mass ratio
(m to mass of AN); dashed line
derivative of mass with temperature.
191
sectional area was calculated using the equation proposed by Emmett and Brunauer [14]:
M
= f
L
23
(3)
where f is a packing factor, which for hexagonal close-packing becomes 1.091, is the
density of the liquid, and M is the molar mass of the adsorbate.
Prior to comparing the surface and structure of the various AN prills, it is essential to
consider the reproducibility of the HR-TG results. The results in Table 2 are the average
value of at least two replicates. The uncertainty of the values represents one standard
deviation of the repeated measurements. For prills A and E, the adsorption capacity and
porosity data (error 10 %) are more reproducible than the surface area data (error 23 %).
For prill D, there is a large uncertainty in the values of nm and a because the specific surface
area of the prill D is smaller than that for prills A and E. Therefore the mass loss resulting
from the thermodesorption of the monolayer (~ 0.15 mass %) is affected significantly by the
baseline drift (0.02 mass %, see the experimental section). Variations in the TG and
subsequently the derived data are believed to be caused by variations in the structure of the
prills. AN prills are highly irregular [6,7], the differences in the surface roughness and pore
structure among the prills significantly affect the reproducibility of the results.
Note that for the specific surface area of AN prills, the values were obtained by
assuming the mass loss of the last step was due to the monolayer of octane adsorbed on AN.
However, the mass loss may be due to adsorbed octane extending beyond the monolayer and
consequently, the surface area may be over estimated. On the other hand, the monolayer
capacity (nm) in Table 2 may still be able to provide information about the relative surface
area of each sample.
In comparison with prills D and E, AN prill A has significantly higher values of
adsorption capacity (nt), porosity () and specific surface area (a), as shown in Table 2. Thus,
the porosity and surface roughness, which increase the surface area, could be the main
factors affecting the adsorption of octane.
For AN prills D and E, there are no significant differences in the porosity. The HR-TG
results suggest that the structural properties of these two samples are similar. However, prill
D has a higher adsorption capacity than prill E, while its surface area is lower. Their
differences in adsorption capacity may be due to other factors such as their moisture content
and surface impurity. The porosity of the various AN prills obtained from HR-TG results
agrees with the general observations from SEM (Section 3.3.1), but the latter technique does
not give any quantitative information about the pore parameters. The relatively high porosity
and surface area of prill A contribute to enhanced wettability with fuel oil, which leads to its
effective application in making ANFO.
3.4
Effect of Surface Composition
X-ray microanalysis was used to characterize the surface chemistry of the individual
samples of AN prills. Spectra of the AN samples contain signals due to the elements of
carbon, nitrogen, oxygen, phosphorous, sulfur, aluminum, silicon and calcium. The signal of
carbon was mainly from the carbon coating of the samples prior to the measurements. No
quantitative analysis was performed; however, the relative quantity of each element was
determined by comparing the intensities of the lines. The results for the sample are
summarized in Table 1; the elements are listed in the order from highest to the lowest
intensities. Evidently, beside the major constituents of AN (oxygen and nitrogen), some
192
other elements are also present on the surface of the prills. It is not surprising to find these
elements, because different coatings and additives (e.g. kaolinite, talc) are added to prevent
the adsorption of moisture by AN.
The X-ray microanalysis results in Table 1 suggest that qualitatively, Types D and E
prills seem to have less impurity on the surface. However, the amount of the impurity on
these two prills may not necessarily be less than on the others, since a strong signal of
phosphorous was obtained from Type E prills. Additionally, since the surface chemistry was
analyzed by a small spot X-ray microanalysis on a selected area of AN samples, the results
may not be representative of the complete sample.
The internal surfaces of the cleaved AN prills were analyzed using X-ray microanalysis,
with the option of area scan. The area analyzed was about 0.16 mm2, which was
approximately 100 times larger than that for the external surface analysis, but the depth of
samples analyzed was also lower. The results in Table 1 suggest that there may be less
impurity on the internal surface of the prills. Moreover, there was no nitrogen detected from
the internal surface of Types AD prills. Apparently, oxygen occupies the most outer layer
of cleaved surface of the prills.
Surface impurities affect surface wettability of AN prills. To determine the effect of
coatings or additives on the wetting of AN prills, comparison of AN prills with and without
coating is required. However, AN is very hygroscopic without the coating. Hence, it would
be difficult to eliminate the effect of moisture and to obtain valid results from capillary
penetration measurement. Alternatively, wettability studies on the coating and additives used
per se may provide hints on their effects on the wettability of AN.
3.5
Effect of Bulk Density
The bulk density, void fraction and prill size distribution for various AN prills are
summarized in Table 1. The bulk density of AN prills depends upon the shape of the prills,
their density, the geometrical arrangement, as well as the void fraction. Theoretically, for
uniform sized spheres, the void fraction varies from about 0.26 to 0.48 for different types of
packing from 100 % rhombohedral to 100 % cubic. The void fractions of Types A, D and E
prills fall within these two values, but towards the higher value because of random
arrangement of the samples. Types B and C prills have a higher void fraction, which may
result from the non-uniform spherical shape of the porous prills.
Among the porous prills, sample A has the lower void fraction and higher bulk density,
due to its wider distribution in prills size (Table 1). Type A prills have about 25 mass %
prills with smaller size. As discussed by Dallavalle (15), for a binary or multiple component
system (of varied sizes of spherical particles), the smaller component can fit in the interstices
of the larger component, hence increasing the bulk density. Systems of one size component
are not amenable to as high a bulk density as are multiple component systems, and binary
systems do not reduce the void space as much as three or more component systems.
The higher bulk density enhances the wettability of Type A prills, in comparison with
Types B and C prills (Table 1), while they have similar surface and structural properties. For
Types D and E prills, the bulk density is higher than for Type A prills. However, their high
bulk densities are due to their higher prill densities, i.e. low porosities. The low porosity of
Types D and E prills are the limiting factors for their wettability.
193
3.6
Effect of Particle Size
The surface area available for adsorption is dependent upon the particle size of the
samples. The specific surface area for spherically shaped particles is inversely proportional
to the particle diameters, and the surface area per unit volume increases very rapidly with a
decrease in particle size, especially at smaller diameters. This results in a much larger
surface area per unit volume of materials accessible for liquid adsorption.
To study the effect of particle size on wettability of AN prills, Type A prills were
screened into four size ranges, and their wettability with heptane was determined using the
capillary penetration measurements. As compared in Table 4, there are no significant
differences in the wetting rates of the samples. The wettability of AN prills seems to be
weakly dependent on the prill size in this limited range.
The effect of particle size was further studied by using AN sample in powder form. The
powder sample was ground from Type A AN prills. The wetting rate of AN powder with
hexadecane is 75 8 g2 min-1, which is about a factor of 180 greater than Type A prills (0.42
0.04 m2 min-1). With finer particles, the total surface area of the samples for adsorption is
higher. Hence, the wettability of AN increases as the particle size decreases. Finer powders
have substantially higher wettability than AN in prills form. However, they also have a
greater tendencies to absorb moisture and to cake, which prevents the adsorption of fuel oil.
Moreover, X-ray microanalysis shows that the surfaces of the ground powders contain
oxygen, nitrogen and sulfur. Compared to Type A AN prills (Table 1), the ground powders
have less impurity on their surface, since the prills interior is exposed after grinding. The
substantial increase in wetting rates may also be caused by the difference in surface
composition.
3.7
Wettability of Moisturized AN Prills
Studies were carried out on the effect of moisture content on the wettability of AN with
alkanes. The moisture content of the as-received and moisturized prills was determined by
measuring water desorption from the samples using TG. The moisture contents of the
samples were compared using the total mass loss at 100 C (Table 5). The error terms
represent one standard deviation obtained from at least two replicates. The moisture content
is a function of the time exposed to water vapour and the R.H. in the system.
The wetting rate of the various
moisturized AN samples with heptane was
determined using the capillary penetration
technique. As shown in Fig. 10, the results
from the capillary rise measurement show
that the wettability of AN prills decreases
linearly with increasing moisture content.
The lower wettability may be caused by
blockage of the pores and the surface of AN
prills by water, or even by dissolving the prill
surface. Prills moisturized for a shorter
period may not be seriously affected.
However, the adsorbed water may still
increase the hydrophilic property of AN
prills, and increase the contact angle of AN
with alkanes, thereby preventing the alkanes
from entering the pores of AN prills.
Fig 10.Effect of moisture content on the

wettability of AN prills. The error
bars show standard deviation of at
least 2 replicates.
194
4.
CONCLUSIONS
The capillary penetration technique involves application of the Washburn equation to

experimentally determine the rate of penetration of a liquid into a particle bed. The main
deficiency of the model on which the Washburn equation is based is that only a single
cylindrical capillary is used to characterize a complex porous network. Hence, the method
cannot properly model the exact pathway of the liquid penetration in AN prills. However, the
wetting rate data can be used as an empirical measure to compare the relative wettability of
various AN prills. The sensitivity and simplicity of the capillary penetration technique makes
it useful for comparing the wettability of samples.
The wettability of various AN prills with alkanes and fuel oil were compared. Factors
influencing the wettability of AN prills include: surface tension, viscosity, density, purity,
polarity and/or polarizability of the wetting liquid, as well as surface composition, porosity,
bulk density, particle size, size distribution and moisture content of AN samples. The
explosive-grade (porous) AN prills show a better wettability due to combined effects of the
above factors.
HR-TG was applied to study the thermodesorption of octane from AN prills. The
adsorption capacity, porosity and specific surface area of three different AN samples were
evaluated from the resulting HR-TG curves. The estimated porosities of the AN prills agree
with scanning electron microscopy observations. It appears that HR-TG is a simple and
valuable technique to characterize AN porosity and adsorption capacity.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
A. MARANDA, D. GALEZOWSKI AND M. SWIETLIK: Przegld Grniczy, 54-11, p. 33, 1998.

F. ROUQUEROL, J. ROUQUEROL AND K. SING: Adsorption by Powders and Porous Solids,
Academic Press, San Diego, 1999.
E.W. WASHBURN: Phys. Rev., 17, p. 273, 1921.
K. GRUNDLE, T. BOGUMIL, T. GIETZELT, H.-J. JACOBASCH, D.Y. KWOK AND A.W.
NEUMANN: Progr Colloid Polym. Sci., 101, p. 58, 1996.
A.K. CHATTOPADHYAY, J. KNIGHT, R.D. KAPADIA AND S.K. SARKAR: Journal of Materials
Science Letters, 13, p. 983, 1994.
S. NIE: Journal of the Industrial Explosives Society (Japan), 49(5), p. 305, 1988.
K.V. RAMANA RAO, P.L. HARIHARAN, K. JAGANNATHAN AND S.R. YOGANARASIMHAN: Fuel,
68, p. 1118, 1989.
P. STASZCZUK, R. DOBROWOLSKI AND M. JARONIEC: Langmuir, 8, p. 229, 1992.
M. JARONIEC, R.K. GILPIN, P. STASZCZUK AND J. CHOMA IN: Characterization of Porous
Solids III (ed. J. Rouquerol, F. Rodrguez-Reinoso, K.S.W. Sing and K.K. Unger), p. 613,
1994.
M. JARONIEC, R.K. GILPIN, J. RAMLER AND J. CHOMA: Thermochimica Acta, 272, p. 65,
1996.
Z. LI, M. JARONIEC AND J. CHOMA: Thermochimica Acta, 345, p. 165, 2000.
P. STASZCZUK, M. JARONIEC AND R.K. GILPIN: Thermochimica Acta, 287, p. 225, 1996.
D. QUR: Europhys. Lett., 39, p. 533, 1997.
P.H. EMMETT AND S. BRUNAUER: J. Am. Chem. Soc., 59, 1553, 1937.
J.M. DALLAVALLE: Micromeritices, Pitman Publishing Co., New York, 1948.
Acknowledgment:
The authors like to thank Ms. Aida Kaldas of Orica Canada Inc. and Mr. Gordon
Chalmers of ETI Canada Inc. for their kind donation of samples.
195
Table 1. Characteristics of various AN prills and their wetting rate with heptane
Type
Category
Porous
Porous
Porous
Mini
Dense
Mass % with prill

diameter/mm*
1.7 1.4 1.0
2.0
1.7
1.4
Bulk
density
/g cm-3
Void
fraction
Prill size
/mm
0.83
0.01
0.75
0.01
0.76
0.01
0.97
0.01
0.91
0.01
0.48
0.01
0.55
0.01
0.52
0.01
0.43
0.01
0.43
0.01
0.9 3.0
34
41
10
14
1.2 3.0
52
46
1.2 3.0
95
0.7 1.3
100
1.8 3.4
98
>2.0
Internal
surface
103 Wetting
rate/g2 min-1
2.68 0.05
2.26 0.22
2.39 0.18
O, N, P
1.20 0.15
O, N, P
O, N, P
0.67 0.02
External
surface
<1.0
O, N, S,
Al, Si, P
O, N, Ca,
S, P, Al
O, N, S,
Ca, Al
*Note: Values correspond to mass % of prills within the specific sizes range, which is expressed in mm.
Table 2. HR-TG Results for porosity analysis of AN prills

Sample
A
D
E
Mass per
prill/mg
6.3 0.6
1.0 0.2
7.0 1.0
Diameter
/mm
2.0 0.2
1.0 0.1
2.0 0.1
Va/cm3 g-1
nt/mass %
np/mmol g-1
Vp/cm3 g-1
nm/mmol g-1
a/m2 g-1
0.66 0.06
0.52 0.05
0.59 0.03
15.7 0.3
9.1 0.5
7.7 0.5
0.83 0.06
0.29 0.05
0.31 0.03
0.14 0.01
0.05 0.01
0.050 0.004
0.21 0.01
0.09 0.02
0.08 0.01
0.14 0.03
0.02 0.01
0.08 0.02
39 9
42
23 5
196
Table 3. Capillary penetration results for Type A AN prills with various liquids
Liquids
Heptane
Octane
Nonane
Dodecane
Hexadecane
Fuel oil
Supplier* lv/mJ m-2 /g cm-3 /mPas

F
A
A
A
A
G
21.4
22.6
23.4
26.4
27.8
30.5
0.678
0.698
0.715
0.747
0.770
0.820
0.405
0.528
0.687
1.403
3.159
1.765
103 Wetting rate

/g2 min-1
2.83 0.10
1.97 0.18
1.24 0.04
0.70 0.04
0.42 0.04
1.16 0.33
*Note: A, Aldrich Chemical Inc.; F, Fisher Scientific Co.; C, Cenco Instruments Co.; G,
local gas station.
Table 4. Wetting rate of AN prills of

different sizes with heptane
Prill size/mm
1.0 1.4
1.41.7
1.72.0
2.03.0
103 Wetting rate

/g2 min-1
2.22 0.16
2.65 0.22
2.65 0.26
2.61 0.15
Table 5. TG Results for moisture analysis

of AN prills
Time of
moisturizing/h
0
1.5
3
5
6
12
Moisture content
/mass %
0
0.46 0.03
2.18 0.04
1.10 0.10
2.42 0.11
3.79 0.13
197
DETERMINATION OF THE POWER OUTPUT OF DETONATORS

Michel H. Lefebvre
Dept Chemistry RMA,
30 Av de la Renaissance, 1000 Brussels, Belgium
Abstract:
The need to assess the power of detonators led recently to the development of new
European standard tests to determine their initiating capability. In an effort to
standardize the methods, European Union experts have selected an underwater
explosion test to determine the basic power of detonators and a witness plate test to
assess the influence of the ambient temperature on their performance.
The underwater explosion test method, commonly used to calculate the explosive
performances, is used to evaluate the initiation power of electric detonators. The
investigation of the influence of selected parameters (distance detonator/pressure
gauge, integration limit in the post-processing procedure, relative position of the
detonator, the type of sensor, sampling rate of data acquisition) made possible to
draw up a satisfactory standard experimental procedure according to the EU-directive
93/15.
Keyword:
1.
Keywords: detonators, power output, initiating capability, underwater

explosion, witness plate
INTRODUCTION
The power and the energy of a detonator must be sufficient to assure proper initiation of
an explosive train or a bulk explosive charge under various situations. To satisfy reliable
initiation, the shock produced by the detonator must be higher than the initiation pressure of
the explosive. Weak initiation pressure leads to misfire or to low-order detonation of the
main charge.
Various tests exist to assess the power/brisance of explosive charges, for example the
lead (or aluminium) plate test, the sand test, or the ballistic mortar. These are often
characterized by their simplicity but also by the lack of technical information they provide.
Another typical technique is underwater explosion, commonly used for the characterization
of high explosive charges. It has been well documented for large charges but has not yet
been applied to detonators [1-5] and has been proposed in the United States, the Office of
Mines. [5]
It is generally assumed that the initiating power of an exploding detonator is related to
the energy of the produced shock wave and, to a lesser extent, to the pressure generated by
the combustion gases. These both aspects of the initiation can be observed in the underwater
experiments.
In an effort to standardize test methods, European experts have suggested to develop:
-
an underwater explosion test for the assessment of the standard initiating

performance of detonators and,
a specific witness plate test for the assessment of the performance at various initial
temperatures.
198
2.
UNDERWATER TEST
2.1
Principle of the test method
The underwater explosion shows specific characteristics: first, the formation of a shock
wave that propagates through the water; secondly, the confinement of the detonation
products by the water and the formation of a gas bubble. The work of the pressure forces is
the main interaction between water and the explosion products. Various theories tried to
establish the relationship between the intensity of the primary shock or/and the first bubble
period and the total energy of the exploding charge (Kirkwood, Brinkley, Penney, Cole [1]).
One generally assumes that:
a) the impulse of the primary shock is directly related to the shock energy (Ec) of the
explosion,
b) the first period of the gas bubble is directly related to the pressure work (Eb) of the
combustion products.
The shock energy is thus proportional to the impulse of the primary shock and the
theoretical relation may be written as
Ec = K
t0 + x
t0
P( t )2 dt
(1)
where t0 + x is the integration time interval (calculated from the initial pressure rise, t0,
to a certain decay value of the pressure), P the pressure and K is a factor of proportionality
depending on the actual experimental parameters (density of water, loss of energy by
diffusion, speed of the sound in water) and on the environmental conditions (air pressure).
K should be determined by testing the experimental setup with a benchmark explosive.
Consequently, it is difficult to obtain the absolute value of the shock energy Ec. Moreover,
the use of Eq. (1) requires the knowledge of the integral boundary (x). The integration
interval must be sufficiently large to include the total impulse of the frontal shock. The lower
limit of the integral is the start of the pressure rise and can be easily determined, but the
upper boundary is more difficult to assess. Indeed, the recorded tail of the shock does not
necessarily reflect the properties of the expansion, is partly dependent on the technical
characteristics of the pressure sensor[6], and consequently, the quantitative interpretation of
this expansion is difficult. Nevertheless, it is necessary to take part of the tail of the shock
into account for the calculation of the shock impulse. To characterize the integration interval,
we use a characteristic time of the pressure decay, , defined as the time needed for the
pressure to drop from Pmax to Pmax/e. Using the same parameter , the Bureau of Mines
suggests not to integrate Eq. (1) over 10, and mentions that integration beyond 6 does
not significantly change the calculated shock energy. In his work on underwater explosions,
Cole recommends a value of 5.[1] Other authors recommend 6.7. In this study, we vary
this parameter in order to observe its influence on the calculation of the shock energy Ec.
As mentioned above, the bubble or gas energy is proportional to the first period of the
bubble, also called pseudo-period of the gas bubble, tb. This period has to be correlated with
the internal energy and the pressure of the combustion products. For an experimental setup
with constant water depth and constant atmospheric pressure, the bubble energy can be
written as:
Eb = K ' t b3
(2)
199
2.2
Experimental setup
The experimental setup consists of a water reservoir of sufficient dimension to contain

the pressure measurement system and the detonator (Fig. 1). The distance between the tested
detonator and the sensor is typically 40 cm but, in this study, we investigated the effect of
this distance ranging from 10 to 60 cm. The sensor must be sufficiently far from the wall not
to be influenced by the shock reflections. By covering the wall with shock absorbing
materials, we minimized these reflections. All experiments presented in this work have been
carried out at a depth of 40 cm.
The sensor must be capable to measure underwater shocks and is typically a piezocrystal
of the tourmaline type, equipped with an insulated protection; the rise time is about 1 to 2 s.
The data acquisition must have a very high sampling rate for recording the shock
pressure (10 MHz or more) and a low sampling rate for the measurement of the bubble time.
40 cm
shock
absorbing
walls
20 cm
minimum
deto
40 cm
sensor
sensor
weigth
Fig 1.
2.3
Schematic drawing and picture of the experimental setup.
Experimental Results
The detonators we tested in this study are calibrated detonators whose primary charge is
lead azide (200 mg) and the secondary charge is made of 0, 250, 450, 600 or 800 mg of
PETN. They have been specifically manufactured by Davey Bickford for this investigation
and have an accuracy of the explosive masses (primary and secondary) better than 1%.
Figure 2 shows a typical record of the shock wave. The signal of the shock wave is presented
with two different time bases (Fig. 2a). Figure 2a is suitable for the integration according to
Eq. (1). The bubble period is read directly from the record shown in Fig. 2b.
Comparative testing among different European laboratories has been performed.[7]
2.4
Influence of the distance Detonator-Gauge and the integration

limit
The influence of the distance detonator/pressure sensor on the pressure impulse is

shown in Fig. 3, with both linear and logarithmic scales. The results represented in this
figure have been obtained with detonators of medium strength. If one wants to limit the
measured pressure of the shock below 100 bar, a minimum safety distance of 40 cm between
the detonator and the pressure sensor is necessary. Figure 3b shows the good exponential
dependence of the impulse vs. distance. Figure 3 gives also the values of the integral (see Eq.
(1)) for different upper boundaries of the integral. When the upper integration boundary is
equal to Pmax (i.e. time to reach Pmax), large scattering of the results is observed. When the
boundary is greater or equal than 5, the curves are identical. The analysis of the curves
shows that the optimum choice of the upper integral boundary lies around 5. An
200
integration over a time period longer than 5 may introduce phenomena more related to the
technology of the sensor (capacitance effect, discharge characteristics) than to the actual
shock wave. This observation is consistent with results discussed by the Bureau of Mines for
other applications of underwater explosions. [5]
2.5
Classification of detonators
The underwater test described above can be used to classify the detonators according to
their power output. Detonators of several strengths have been tested and Fig. 4 gives a
summary of the resulting records. The detonators are classified according to the mass of their
initial secondary explosive charge, ranging from no secondary charge (250 mg of PbN6) to a
800 mg of PETN. In addition, Fig. 4 shows the good reproducibility of the measurements.
0
10
20
30
40
50
10
15
20
25
8000
7000
7000
6000
6000
5000
P re ssu re (kP a
Pressure (kPa)
8000
(1)
4000
3000
5000
4000
3000
Bubble period
2000
2000
(2)
1000
1000
0
0
0
50
100
150
200
250
10
(a)
Fig 2.
15
20
25
30
T im e (m s)
Tim e (s)
(b)
Typical pressure records:
(a) Shock record for two different time bases and (b) bubble period.
1.0E+11
1.2E+10
Pmax
Pm/e
1.0E+10
e*theta
8.0E+09
5*theta
6.7*theta
6.0E+09
10*theta
Pressure Integral (kPa 2 ms2)
Pressure Integral (kPa 2 ms2)
1.4E+10
4.0E+09
2.0E+09
0.0E+00
0
10
20
30
40
1.0E+10
1.0E+09
e*theta
5*theta
1.0E+07
6.7*theta
10*theta
1.0E+06
50
Distance (cm)
10
100
Distance (cm)
(a)
Fig 3.
Pmax
Pm/e
1.0E+08
(b)
Shock impulse as a function of the distance between the detonator and the
gauges. (a) linear axis system, and (b) log-log diagram. The label (theta)
refers to the integration boundary used for the calculation of the impulse
(see Eq. 1).
201
Pressure (kPa)
10000
8000
800
6000
650
4000
450
250
2000
0
0
10
15
20
25
30
35
40
Time (s)
Fig 4.
3.
Shock measurement for various calibrated detonators. The labels refer

to the explosive secondary charge of the detonators, expressed in mg
of PETN. Each type of detonators has been tested 7 times and records have
been superimposed to emphasize the reproducibility of the test method.
WITNESS PLATE TEST
The witness plate test is a function test designed to ensure that the energy output does
not vary much when the detonator is used at high and low temperatures. The plate is made of
an aluminum/magnesium alloy, 50x50 mm, 10 mm thick. The figure 5 represents a
schematic draw of the standard experimental set up. The detonator/witness plate assembly is
conditioned at ambient, low and high temperatures (typically 20, -40 and 60C, respectively)
and is fired with an air gap underneath the plate. An example of the imprint o the aluminum
witness plate is shown in Fig. 6.
The dent depth resulting from the test performed at low and high temperatures is
compared to the one of the test carried out at ambient temperature. The figure 7 shows
typical results obtained with detonators of different strengths.
Detonator
Expanded foam for

insulation
Witness plate
Steel cylinder
Metal anvil
Fig 5.
Schematic of the experimental

set up of the witness plate test.
Fig 6.
Fig 6. Typical imprint after firing

of a N8 detonator.
202
2.2
Dent depth (mm)
1.8
240 mg
240_average
445 mg
445_average
650 mg
650_average
1.4
-40C
20C
0.6
Temperature
4.
55C
Fig 7. Depth of the imprint in the Al witness

plate, as function of the initial
temperature of different detonators.
The label (in mg of PETN) refers to
mass of the secondary charge of the
detonator.
DISCUSSION
Many theoretical discussions could be carried out on the relation existing between the
gas energy, the shock energy, and the actual initiation output of a detonator. Many notions,
sometimes contradictory, should be taken into consideration, but question remains:
-
Is the initiation output of a given detonator may be dependent on the initiated

explosive?
Is the initiation of an explosive only a function of the delivered energy and the shock
pressure?
Does the gas energy, i.e. the characteristics of the bubble, also contribute
significantly to the initiating capability, specifically for initiation of slurries and
emulsions?
No straight answers exist to these questions. Moreover, the presented results do not
assess quantitatively the different forms of energy produced by the detonator. But the use of
underwater experiments for the characterization of the initiation output characteristics of
detonators has many advantages:
-
simplicity of the experimental setup and easiness of the experimental procedure,
good reproducibility of all experimental results, for a given experimental setup,
good reproducibility of the experimental results for the bubble period,
good qualitative physical output (gas energy, shock energy, shock pressure), much
more meaningful than the output provided by other simple tests (lead plate, sand
test),
easy and relevant comparison between detonator of various strengths.
203
In addition to classification and comparison purpose, the good reproducibility of the test
results mean that this method might be considered as a promising technique for quality
controls during manufacturing process. However, a certain number of issues require
additional investigation, the most important one being the theoretical relationship between
the gas energy, the shock energy and the actual initiation capability of detonators.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
R.H. COLE: Underwater Explosions, Princeton University Press, 1948.

M. SUCESKA: Test Methods for Explosives, Springer, 1995.
T. YOSHIDA, Y. WADA, AND N. FOSTER: Safety of Reactive Chemicals and Pyrotechnics,
Elsevier, 1995.
P. WOLLERT-JOHANSEN: Underwater Testing of Slurry Explosives, Propellants and
Explosives, 1980.
G. BJARNHOLT: Suggestions on Standards for Measurement and Data Evaluation in the
Underwater Explosion Test, Propellant and Explosives, 5, pp. 67-74, 1980.
Characteristics of Delay Detonators, US Bureau of Mines, RI8989, 1985.
K. MURATA, K. TAKAHASHI, AND Y. KATO: Measurement of Underwater Shock Wave by
Fluoropolymer Sensor, 23rd Intl Pyrotechnics Seminar, Tsukuba, Japan, pp. 548-560, 1997.
M. BROCHIER, M.H. LEFEBVRE, A. CRINON, M. VAN ROOIJEN: Puissance dAmorage des
Detonateurs, 25th Pyrotechnics Seminar, Brest, France, 1999.
204
INFLUENCE OF DEMILITARIZED DOUBLE BASE PROPELLANTS

ON DETONATION PARAMETERS OF AMMONALS
Katarzyna Lipiska *, Marek Lipiski * and Andrzej Maranda **
** Military University of Technology,
2 Kaliskiego St, 00-908 Warszawa 49, PL
Abstract:
The incorporation of double base propellants into mining blasting agents offers an
optimal method of disposing military propellants. The influence of double base
propellants on detonation velocity of ammonals mixtures containing ammonium
nitrate and aluminum powder - was investigated. Initial results suggest that double base
rocket propellants used as ingredients of ammonals can increase their detonation
parameters.
Keyword:
1.
waste energetic materials, ammonal, double base propellant
INTRODUCTION
Waste military solid rocket propellants are a growing disposal problem as a result of
world demilitarization. Double base propellant is a basic propellant commonly used for
small arms ammunition, artillery shells, and the like. It is basically a cellulose nitrate but
may contain nitroglycerine and other ingredients. Double base propellant no longer suitable
for its original purposes may be economically and efficiently utilized in safe mining blasting
agents[1-5]. There is a variety of blasting agents, which are suited for incorporation of double
base propellants. One of them is ammonal a mixture of ammonium nitrate and aluminum.
2.
EXPERIMENTAL
As a part of our researches we investigated the influence of double base propellants on

detonation velocity of some blasting agents containing aluminum powder. We used two
different kinds of double base propellants. A double base propellant designated as propellant
A [Fig. 1] was in the form of squares of 22 mm, having a thickness of 0,25 mm and a
composition given in Table 1. Another double base propellant designated as propellant B
[Fig.2] was in the form of small rings: OD 5,7 mm, ID 2,2 mm, thickness 1,7-2,7 (average
2,1 mm) and a composition given in Table 2. Detonation velocities of both propellants are
given in Table 3.
205
Table 1. Composition of double-base propellant A

component
[%]
nitrocellulose
57,0
nitroglycerine
40,0
centralite I (1,3-diethyl-1,3-diphenylurea)
2,5
vaseline
0,5
Table 2. Composition of double-base propellant B

component
[%]
nitrocellulose
56,0
nitroglycerine
26,5
dinitrotoluene
9,0
centralite I (1,3-diethyl-1,3-diphenylurea) 3,0

dibuthyl phthalate
4,5
vaseline
1,0
Table 3. Detonation velocities of propellant A and propellant B

density
[kg/m3]
oxygen balance
[%]
detonation
velocity [m/s]
double base propellant A
746
-26,77
4290
double base propellant B
721
-47,17
1400
206
Fig 1.
Double base propellant A
Fig 2.
Double base propellant B
Ammonals were prepared by mixing crystalline ammonium nitrate (AN) with fine
metallic aluminum (Al) and double base propellant in the ratio given in Tables 4-6. Two
different kinds of aluminum powder were applied. Their refinement is shown [Fig. 3 and
Fig. 4].
207
100
7
80
6
60
[%]
[%]
40
3
2
20
1
0
Fig 3.
20
40
60
80
Particleziarna
size [m][mm]
Rozmiar
100
0
120
Refinement of flaked aluminum
100
7
80
6
60
[%]
[%]
40
3
2
20
1
0
100
200
300
400
500
0
600
hgfhfsize [m]
Particle
Fig 4.
Refinement of aluminum powder
Then steel tubes (ID36 mm, OD 42 mm and length 210 mm) were filled with the
mixtures. Detonation velocity was determined by the measurement of the time of detonation
wave transition through three sections; each of them was 40 mm long. Investigated
explosives were initiated by RDX/Al boosters. Measurements of time were done by short
circuit sensor method.
208
3.
RESULTS
Compositions of investigated mixtures, their oxygen balances and results of detonation

velocity measurements are presented in Table 4, 5 and 6.
Table 4. Detonation velocities of ammonal (with flaked aluminum) and
ammonal/double base propellant A
density
[kg/m3]
oxygen
balance [%]
detonation
velocity [m/s]
Al/AN 3/97
856
+16,72
3290
Al/AN/double base propellant 3/77/20
855
+8,59
4030
867
+0,47
4440
Al/AN 6/94
843
+13,45
3450
880
+5,32
4100
881
-2,80
4460
Al/AN 10/90
844
+9,10
3470
854
+0,97
3950
842
-7,16
4260
Table 5. Detonation velocities of ammonal (with flaked aluminum) and

ammonal/double base propellant B
density
[kg/m3]
oxygen
balance [%]
detonation
velocity [m/s]
Al/AN 3/97
856
+16,72
3290
1096
+3,29
3410
1104
-10,14
3400
Al/AN 6/94
843
+13,45
3450
1050
+0,02
3600
1059
-13,41
3250
Al/AN 10/90
844
+9,10
3470
947
-3,34
3310
949
-17,77
3090
209
Table 6. Detonation velocities of ammonal (with aluminium powder) and

ammonal/double base propellant B
density
[kg/m3]
oxygen
balance [%]
detonation
velocity [m/s]
Al/AN 3/97
1034
+16,72
2590
1095
+3,29
2840
1120
-10,14
2840
Al/AN 6/94
1038
+13,45
3090
1089
+0,02
3280
1093
-13,41
3230
Al/AN 10/90
1039
+9,10
3600
Al/AN/double base propellant

10/70/20
1091
-3,34
3380
Al/AN/double base propellant

10/50/40
1094
-17,77
3270
The results show that the addition of double base propellant A increases detonation
velocities of ammonals. The bigger propellant content is the higher detonation velocity of
investigated explosives. That indicates detonation process in investigated mixtures
propagates mainly between propellant A grains. Ammonals without double base propellant
have the highest positive oxygen balances. The explosives with 3 % and 6 % of flaked
aluminum and 20 % of double base propellant A have the most favorable detonation
parameters and positive oxygen balance. There is no clear detonation velocity increase in the
case of double base propellant B despite the increase of densities of charges. Some mixtures
detonate at lower velocities than initial ammonals. That shows the detonation process
proceeds between ammonium nitrate and aluminum.
4.
CONCLUSIONS
Double base rocket propellants can be used as ingredients of ammonals, increasing their
detonation parameters. The incorporation of double base propellants into mining blasting
agents offers the most controlled, safe and relatively environmentally friendly method of
disposing military propellants. Moreover, the use of these propellants as ingredients of
commercial blasting agents would be an economical means of their disposing and in the
process, producing industrial explosives exhibiting a favorable cost for performance. It is
necessary to stress that technological processes should be adapted if possible to existing
production in order to reduce economic problems.
210
REFERENCES
[1]
[2]
[3]
[4]
[5]
G. ECK, O. MACHACEK, K. TALLENT: The Use of Surplus Smokeless Powder Propellants

As Ingredients In Commercial Explosive Products in the United States, (Application of
Demilitarized Gun and Rocket Propellants in Commercial Explosives, Kluwer Academic
Publishers), Dordrecht, 2000
E. F. OKHRIMENKO, N. G. IBRAGIMOV, E. KH. AFIATULLOV, L. S. LUKIN, YU. M.
YUKOV, I. P. IVANOVA: Use of Powders in National Economy, (Conversion Concepts for
Commercial Applications and Disposal Technologies of Energetic Systems, Kluwer Academic
Publishers), Dordrecht, 1997
US Patent 3,331,717 1967
B. VETLICK, T. DOSOUDIL: Utilisation of Double-based Powders and Rocket Propellants
for Production of Dry Explosives of Slurry Type, (Application of Demilitarized Gun and
Rocket Propellants in Commercial Explosives, Kluwer Academic Publishers), Dordrecht, 2000
M. BALOJAN, V.V. GALKIN, D. F. MAROV, N. I. KONEV, V. V. FALKO, Industrial
Explosives Based on Double Base Propellants (Conversion Concepts for Commercial
Applications and Disposal Technologies of Energetic Systems, Kluwer Academic Publishers),
Dordrecht, 1997
211
METHODOLOGY OF TESTS OF SENSITIVITY

OF ENERGETIC MATERIALS TO ELECTROSTATIC DISCHARGE
Ji Strnad and Ji Majzlk
Department of Theory & Technology of Explosives, University of Pardubice

CZ 532 10 Pardubice, Czech Republic
Abstract:
The literature describes concepts of test devices designed for evaluation of sensitivity of
energetic materials (EM) to spark discharge. However, mutual comparability of the
resulting activation energies is relatively low, which among others is due to
- differences in construction of the respective devices,
- type of the test electrodes and container,
- differences in the amount of sample exposed,
- sometimes differences in interpretation of the sample activation symptoms,
- a wide variety of the activation symptoms found in the whole large palette of EM.
The aim of this communication is to describe the tests performed on a self-constructed
electrostatic device (ESZ) and to suggest a unique approach to evaluation of the
activation measure of EM. Acceptance of the regulations suggested could reduce the
personal view of the test course and the resulting condition of the sample tested.
The apparatus works at the conditions of both oscillation and damped discharge of a
condenser charge into a circuit containing besides parasitic members also the spark gap
(JZ) where the sample of EM is exposed.
Keywords: electric spark, sensitivity, explosives, primers
1.
SENSITIVITY TESTS OF SECONDARY EXPLOSIVES
Secondary explosives are predominantly sensitive to intense short-term initiation stimuli,

which can be brought about e.g. by a discharge of an electrostatic charge. The sensitivity of
explosives to this factor is evaluated by means of an electrostatic device whose construction
and operation is described in Ref. [1]. The present article deals with the principles to be
followed during preparation of the test sample and during carrying out and evaluating a test
series.
1.1
Preparation of Device
A sandwich set is included into the ESZ circuit; the set includes a meander shunt of the
value Rb = 0.1 , which simultaneously serves for exposition of test sample according to
Fig. 1. In this test variant, the time course of discharging current and voltage is usually not
measured. In laboratory, the temperature and relative air humidity are monitored.
Permissible working conditions are given in Ref. [2].
212
1.2
Preparation of Sample
The characteristic dimensions of sample particles of energetic material should be much

smaller than those of the container = 2.5 mm 1 mm. In the context of the aim or set
conditions of test, the substance is recrystallised or ground and sieve tested, the appearance
of particles is checked etc. The sample must be dried and brought to, and kept at, the
required temperature in usual way.
4,2
Needle 1,6
JS
1,5
Anode with
upper tube
Upper tube
(Plastic)
Sample
Bottom
tube
Metal hat
Bottom
tube
3,8 (Plastic)
1
2,2
2,5
Rb
Steel disk
Fig 1.
Test spark gap (SZ) assembled
Using the same ESZ after slight modification it is also possible to evaluate the dielectric
strength and reactivity of compact EM samples. In this case the test spark gap and the
sample dimensions must be chosen in such a way that the surface trajectory of discharge
should be much higher than the sample thickness.
1.3
Sensitivity Evaluation
1.3.1. Sample Deposition

Safety goggles must be worn when handling sensitive EM.
The sample is filled into the space formed by the lower plastic tube slid (with an overlap
of 1 mm) on the mandrel of 2.5 mm. The material that remained above the upper level of
the flexible tubing is gently pressed into the container by means of a squeegee. The density
increase above the previous (loose) condition is modified in accordance with the particle
dimensions and their properties in bulk. The material that fell over on the metal cap is
removed.
213
The movable part of the device is centred with the axis of the vertically adjustable
anode. The contact between the anode and sample surface is adjusted. The error of
adjustment of electrode distance h = 1 mm should not be greater than ca 0.1 mm.
1.3.2. The Test Proper
Capacity of the battery of condensers can be chosen in the range from 40 pF to 350 nF.
After carrying out all the necessary safety operations, the feeding of the HV source is
switched on. The mechanical change-over switch is transferred into its unstable position by
means of a draw rod, which connects the condenser battery to the HV source. The charging
voltage can be adjusted in the range from 4 to 10 kV. The draw rod of spark gap change-over
switch is abruptly set free, which causes the charge to be discharged into the sample.
Simultaneously with the discharge it is necessary to monitor the light, mechanical, acoustic
and smoke symptoms of initiation or inertness of the sample. After the test, necessary safety
operations are carried out and the sample is submitted to inspection.
1.3.3. Symptoms of Decomposition Processes
The materials evaluated in practice possess various mechanical and rheological
properties and differ in appearance. Similarly varied also are the symptoms of initiation of
test sample. It is necessary to make the statement of decomposition (+) or inertness ()
on the basis of two or many symptoms. This requirement follows from the fact that in the
given modification of the test regime it is impossible to differentiate between the light and
acoustic effects due to the electrostatic discharge and the increments of the two effects
caused by decomposition of a part of the exposed sample. The tests of secondary explosives
at the level of threshold initiation energies are usually only accompanied by indistinct
evolution of smoke and smell. The changes in initial colour and appearance of the sample
tend to be indistinct too. Reliable symptoms of the decomposition threshold of secondary
explosives include
- remains of several darkened particles of the substance
- throw-out of considerable amount of material from the container on opening at least
a part of its bottom
- a shift of both isolation tubes out of their initial position, which is also caused by the
pressure of gases formed on decomposition of the substance.
The above-mentioned activation symptoms usually begin to make themselves felt
together within a deviation interval: sometimes the throw-out takes place at energies
somewhat lower than the activation energy. The reliability of the combined criterion of
throw-out and shift of tubes was verified in such a way that we compared the discharge
energy levels needed for activation of several types of EM with the energy needed for throwout of a part of crushed glass. In Table 1 it can be seen that the energy needed for attaining
distinct symptoms of activation of EM (denoted as E+) iswith the test chamber arranged
according to Fig. 1roughly more than one order of magnitude lower than the energy
needed for attaining comparable effects with crushed glass. Hence, the comparison
procedure has confirmed that the throw-out of a part of sample and the shift of tubes during
testing EM in the vicinity of the threshold energies (ca 20 to 800 mJ) are caused exclusively
by the pressure of the gases generated.
214
Table 1. The energy needed for throw-out of a considerable part of sample

Gunpowder
grated, fine
particles,
< 0.5 mm
Substance
Sufficient energy for

throw-out
E+ [mJ]
Threshold energy
E~ [mJ]
Picture number
Insufficient energy
E
[mJ]
PETNk
fine
particles,
< 0.1 mm
TEX
RDX
KTTV, fine
Particles
particles,
< 0.1 mm 0.1 1 mm
Crushed
glass,
particles
min. 70 m
av. 150 m
max. 400 m
787
169
506
178
8100
762
160
387
140
7500
Fig. 2
Fig. 3
Fig. 4
Fig. 5
756
139
284
128
7225
Table 1 also gives numbers of the pictures belonging to the samples that were tested by
the threshold activation energy E~. If the energy drops below this value, the EM activation
symptoms tend to become distinctly lower.
With some types of secondary explosives, in addition, the beginning of activation can
also be detected by smell or by change in appearance of the sample exposed.
Fig 2.
Grated gunpowder,
test energy E ~ 762 mJ
Fig 3.
PETNk,
215
Fig 4.
TEX KTTV,
Fig 5.
Crushed glass,
test energy E ~ 7500 mJ.
After exceeding this value, the amount

of the crushed glass thrown out is
distinctly increased.
1.4
Test Evaluation
The evaluation of tests in the oscillation regime does not take into account the losses in
the circuit, and the Joule energy set free in the test space is given by relationship (1):
E = 0,5.CU
[J; F; V]
(1)
The value E50 is stated after previous determination of reliable overcritical (En) and subcritical energy Ep, given by equation (1). It is recommended to divide the interval (Ea - Ep)
into ten equidistant energy steps, whose difference lies within three to five per cent of value
Ep. The corresponding voltage levels Ui are determined from equation (2), where C = const.
Ui = (Ei /0,5C)
[V; J; F]
(2)
Then using the Up-and-Down method, 20 tests are carried out, and by applying the
respective mathematical procedure the charge energy value on the working condenser E50 is
determined, which corresponds to a fifty per cent probability of initiation of the material.
Another result of the calculation is the standard deviation .
If the sum of all capacitances connected to the HV source is comparable with the
capacity of the coaxial cable between the switch-over and test spark gaps (which is Ck = 50
pF for ESZ), then the E50 value must be corrected with the help of relationship (3), which is
derived from the law of conservation of the system charge:
E50opr = E50 .C/(C + Ck)
[J; F]
(3)
The obtained standard deviations must be corrected in similar way.
216
1.5
Electrical Parameters of Discharging Circuit
The measurement and calculation of the inductance proper, loss resistance of circuit,
energy efficiency of discharge and correction of the i(t) function in the oscillation regime of
ESZ can be carried out according to the procedure given in Ref. [2]. The parameters of the
circuit are measured by means of a high-voltage HF probe P6015A, a shunt with the value of
Rb = 0.1 , and a suitable memory oscilloscope.
2.
SENSITIVITY TESTS OF OTHER EM
The initiation ability of materials that are predominantly sensitive to effects of long-term
thermal impulse can be tested with our instrument at conditions of damped low-output
discharge.
The necessary change in the working regime of ESZ is achieved by inserting a series
resistance of standard value RP = 10 k into the circuit according to [1], which can prolong
the discharge duration of exponential course to 12 ms at the most (at C = 350 nF). If the
voltage is adjusted at U = 10 kV and the initial resistance of test sample is RVZ 1 k, then
the output consumed at the sample at the moment of its breakdown is Pmax 1 kW.
2.1
Preparation of Device
The circuit contains a set of resistances RP = 10 k. The voltage-time course at the

sample is measured by means of a HV HF probe P6015A. The course of current in the circuit
is measured by means of a shunt with Rb = 10 . The passive electrodes of both probes must
be reliably connected to the earth potential of ESZ. The outputs of both probes are connected
to the input of the TDS 3012 oscilloscope see Ref. [3].
2.2
Preparation of Sample and Carrying Out of Test
In principle, this is performed in accordance to the instructions given in paragraphs 1.2.,

1.3.1. and 1.3.2. Special attention must be paid to preparation, filling and testing of primary
explosives.
2.2.1. Symptoms of Decomposition Processes
At the conditions of damped discharge, both the acoustic and the light effects of the
discharge proper are relatively indistinct. On this background, the decision is facilitated
about the activation of sensitive substances or those whose chemical transformation at the
initial phase is accompanied by distinct effects.
- The activation of highly sensitive primary explosives usually makes itself felt by
unambiguous acoustic, light and smoke effects with potential shift of both tubes. If the
energy of condenser charge is decreased below the critical value and no breakdown takes
place, then as a rule no sample activation takes place either. The statement about its reaction
is usually unambiguous.
- The activation of less sensitive primary explosives and initiator mixtures usually makes
itself felt by evolution of smoke, smell and by burning out of a considerable part of the
sample. Acoustic effect can be observed. The statement about the reaction of sample is
usually relatively unambiguous.
217
- The activation of sensitive powders usually makes itself felt by evolution of smoke,
smell, and burning out of a part of the sample. The statement about reaction of this and
similar types of substances usually is relatively unambiguous.
- The activation of less sensitive powders and compositions usually makes itself felt by
partial throw-out, burning out of a part of the sample or other symptoms of beginning of its
burning. However, these phenomena usually are not accompanied by a shift of the tubes due
to the lower level of energy transformation.
In the case of less sensitive EM, only a hole is formed, which is due to a breakdown of
the sample. Figures 6 and 7 show the appearance of activated and non-activated PETNk,
respectively. It is recommended that it should be accepted that a hole of 0.5 mm
diameter, whose edges show traces of burning nuclei, were taken as the minimum symptom
of activation.
Fig 6.
PETNk,
tested in damped regime
C = 47 nF, U = 5 kV, E = 31 mJ Symptoms

of inertness of tested sample
- without blacken particles
Fig 7.
PETNk
tested in damped regime.
C = 145 nF, U = 8 kV, E = 140 mJ.

Tresholds symptoms of activation of tested
sample
- The statement about activation/inertness of less sensitive substances need not always
be quite unambiguous. Therefore, it is recommended to compare (like in paragraph 1.3.3) the
effects of discharge in the EM with those in crushed glass. For a discharge of condenser with
C = 350 nF/10 kV, the energy released in the sample space was determined in the interval
from 200 to 380 mJ. Very slight cases of throw-out of material and formation of a thin
breakdown channel were observed. At the inner wall of the tube, there were usually present
several black particles of the crushed material. These traces can then be compared with the
unambiguous symptoms of activation or inertness of relatively insensitive EM, e.g.
secondary explosives at comparable test conditions.
218
- The records of time course of measured voltage and current exhibit, in some
cases, superposed pulses caused by an interruption of current in the circuit and/or a
marked increase in resistance of the sample. However, it is difficult to find a general
relationship between the irregularities of the discharge course and the activation
degree of the test sample.
2.2.2. Procedure of Tests
The test procedure comprises two phases. In the first phase, according to the
instructions given, the lowest attainable activation energy of the substance and the
highest safe, i.e. non-activating, energy are determined (Emin+ and Emax ,
respectively), both the values being continuously evaluated from oscilloscopic
recording of p(t).
In the second phase, twenty tests are carried out with the ESZ apparatus on which
the parameters are adjusted at which the highest value of safe energy Emax has been
reached. If in these tests activation of sample takes place, it is necessary to repeat the
whole first phase and to determine the real, i.e. the lowest Emax value, whereupon
also the second phase of the test process is repeated.
Both phases must be performed at constant condenser capacitance.
The distance between the two parameters, Emin+ and Emax must be adequate, and
it is Emin+ > Emax. Precise values of the two parameters are determined by means of
computer program SCOPE 5.5 see manual [4].
2.2.3. Processing of Primary Data
The recordings of u(t) and i(t) serve for determination of the energy consumed in
the area of exposed sample. The data set recording the output-time course is created
by the mathematical operation p(t) = u(t).i(t), which can be performed by means of
the oscilloscope accessories or subsequently during evaluation of the primary data
u(t); i(t) by the PC program SCOPE 5.5.
The energy set free in the area of exposed sample is defined by equation (4):
EVZ
t
tk
1 k
2
=
(uVN (t ) u Rb (t )) d t u Rb (t ) d t
Rb t0
t0
[J; V; s; ]
(4)
A correction for voltage loss in the shunt Rb is expressed by the second member
of equation (4) see Ref. [2]. Its calculation by the SCOPE 5.5 program makes use of
the Joule integral of voltage uRb(t).
219
3.
CONCLUSION
The suggested working procedure was created on the basis of the requirement of
defining the way of preparation, performing and evaluating the sensitivity tests of
various types of EM at the testing workplace of DTTE. It can be claimed that
differentiation between excitation/inertness of sensitive or readily inflammable
substances is relatively easy. However, problems are encountered with estimation of
the threshold conditions of initiation of less sensitive EM in the cases of both
oscillation and damped discharge. Literature presents suggestions for solving this
problem by means of monitoring technique. However, introduction of additional
probes into the test area lowers the operativeness of tests, and what is more, the
mediated decision process is always adjusted, controlled and guaranteed by the
operators of the apparatus.
The modification of test spark gap according to Fig. 1 enables an additional
evaluation of sample by its appearance and smell, and the evaluation of the whole test
course. When one tests substances that on increasing energy or output load slowly
without conspicuous breaksincrease the symptoms of their own chemical
transformation, one must connect the amount of material thrown out with the
beginning of shift of one or both tubes of container out of their rest position. If
slightly sub-critical discharge energy is set free, then a slight amount of the test
sample is thrown out without the above-mentioned shift. The acceptability of the
above-described way of deciding about the (+)/() statement was verified in practical
tests.
REFERENCES
[1]
[2]
[3]
[4]
J. STRNAD and J. MAJZLK: Determination of Electrostatic Spark Sensitivity of

Energetic Materials, 4th Seminar DTTE, Pardubice, p. 303 307, 2001
J. STRNAD, J. MAJZLK and J. KO : Measurement of Parameters of Discharging
Circuit (In Czech: Men parametr vybjecho obvodu). UPa, DTTE, 2002
TDS 3000 Series Digital Phosphor Oscilloscopes 071-0274-00. (In Czech)
TM Direct Prague: Scope 5.5. A Program for Computer Support of Oscilloscopes
Tektronix. (In Czech: Scope 5.5. Program pro potaovou podporu osciloskop
Tektronix)
220
ANALYSIS OF THE DEFORMATION AND STRENGTH PROPERTIES

OF ENERGY POLYMERIC COMPOSITES (EPC)
Viktor A. Malchevsky and Natalia A. Zarytovskaya
D.I. Mendeleev University of Chemical Technology
Miusskaya pl. 9, Moscow, 125047, Russia
Abstract:
This study covers one of the most important problems of the physicochemical mechanics
of polymeric materials, i.e., the problem of kinetic and critical approaches to the
evaluation of the mechanical destruction of polymeric composites. It follows the studies
presented at the previous (5th) Seminar held in 2002 in Pardubice. This study is meant to
check the adequacy of the method we are developing for calculation of kinetic
parameters of mechanical destruction of energy polymeric composites (EPC) based on
tests performed using standard tensile testing machines at constant deformation rates
(=const). In this study, the nominal deformation and strength characteristics of a
number of standard compositions of double-base powder and high-filled energy
composites are analyzed using interrelated formulas obtained previously. The results of
this analysis show that the method we are developing is adequate and can be used
successfully, not only to predict the kinetics of destruction of newly developed
compositions but also to reach their technological perfection.
Keyword:
1.
mechanical destruction, energy polymeric composites,

physicochemical mechanics
INTRODUCTION
The prediction of performance and guaranteed storage period of energy polymeric

composites in stressed state is based on evaluation of their damaging during storage,
transportation, and under the operation conditions in propulsion systems. On the other hand,
quantitative evaluation of EPC damaging is feasible without a detailed study of their
destruction kinetics. There are a number of equations that describe the mechanical
destruction of EPC. These include equations (1-5) and some others.
= 0 * exp [U0 + exp * (a b) * (1/T 1/T)]
(1)
= * exp [ + exp * ( f)]
(2)
= B * m
(3)
= C * (ln /)
(4)
= 0 (, ) * exp [U0 (, ) /RT]
(5)
where 0,U0, , a, b, , f, , are constants.

As these equations show, the specific features of the kinetics of EPC destruction consist
in the fact that their strength vs. temperature and strength vs. time curves are non-linear
along the range of working stress and temperature. For different EPC types, this nonlinearity manifests itself in different ways and depends considerably on the EPC
221
composition; this complicates the prediction of kinetics of their destruction, especially in the
life time range of the material under stress within 105-108 sec. This range is of most
practical importance for determination of performance and guaranteed storage period.
Attempts to plot the strength of EPC versus temperature and time in the entire range of their
working temperature and stress upon the results of accelerated tests have been carried out
repeatedly [1-5].
2.
THEORY
In a series of publications [3-5] we have set forth the main physical principles for
prediction of EPC destruction kinetics based on tests performed using standard tensile
testing machines at = const. These approaches are mainly based on an analysis of EPC
strength characteristics.
In particular, considering the dependence of strength on temperature (6)
= U0/ 4.63 (lg lgo) / (1/T 1/T)
(6)
where lg, U0, , 1/ are factors, at =0 or =0 and under the condition that lg
and 1/ are constants, it is possible to calculate U0 and , i. e., the kinetic parameters
of the process:
U0 = 4.63 (lg lg0) / (1/T 1/T)
(7)
= U0/0
(8)
After that, it is easy to plot the linear parts of lg(, T) curves. This method of prediction
may be applicable for certain types of EPC, but it becomes much more complicated if 1/ is
not a constant for a certain material.
Yet another prediction method is based on the use of the damaging accumulation
additivity method [6]
dt/ (, ) = 1
(9)
and equation [7]

= C [ln (K/)]
(4)
where and >>1 are empirical factors, and is the critical strength. It can be
demonstrated that equation (4) can be transformed into equation (3) in the low stress
field (3)
= Bm
(3)
where B and m are empirical factors. In the field of moderate stress it can be
transformed into the well-known equation:
=
(10)
where and are constants. Solving together equations (9) and (3) and then
equations (9) and (10), it is easy to derive equations (11) and (12)
lg = [1/(m + 1)] lg [B (m + 1)] + [1/(m + 1)] lgV
(11)
= (2.3/) lg(A) + (2.3/) lgV
(12)
222
that relate the material strength (, lg) to the overall loading rate V=/tp, where and
tp are the tensile strength of the material and the time to failure, respectively.
The main advantage of this approach is that it does not require that the and 1/
constants be known, and that it makes it possible to predict both linear and non-linear parts
of lg(, T) curves.
The relationships indicated above were checked for a number of standard and model
EPC [3-5]. The check showed that this method allows a rather reliable prediction of
destruction kinetics. However, we undertook additional studies, considering the vast number
of EPC compositions of various types created for various purposes, being of various nature,
and produced with various processes, and that for some of these it may be difficult (and
sometimes impossible) to choose a reasonable value of 1/. These studies were aimed at
analysis of adequacy of the method being developed using not only the strength, but also the
deformation characteristics of composites.
We selected an analysis scheme based on published data on the performance
characteristics of the composites of interest. The following relationships are known: strength
vs. temperature dependencies measured at a standard loading rate of 10-3 sec-1; the values of
lgB and m in the working temperature range; the dependencies of strength and deformation
on the deformation rate over a range of 4-6 orders at =20. The available data do not
allow a direct calculation of destruction kinetics of the compositions being analyzed in order
to compare the calculated and experimental values of lgB and m. This requires that
relationships between the logarithm of strength and the logarithm of loading rate,
lg=f(lgV), be known. Since such relationships are unavailable, the following scheme was
chosen for checking the adequacy of the previously established relationships between test
results obtained at =const and =const:
1. Determine the relationship between the deformation rate () and the loading rate (V)
for each of the compositions analyzed.
2. Using known lgB and m values for a series of temperature values and equation (1),
calculate the strength values corresponding to the standard deformation rate.
3. Compare these calculated values with the available strength vs. temperature curve.
If the method that we are developing for predicting the destruction kinetics is applicable
to the compositions being studied, then the calculated and experimental strength values
should match satisfactorily at all temperatures.
In addition, the scheme that we chose allows us to use not only strength but also
deformation parameters of EPC as a basis for predictions.
3.
EXPERIMENTAL
The relationships between the deformation and loading rates were studied for a number
of standard and model compositions. Examples of such plots are shown in Figures 1-3 for a
high-filled composition with an elastomer matrix, for a double-base composite, and for the
NC-NHC (50/50) model system. It is evident that the loading rate vs. deformation
relationship depends on temperature rather weakly. In the temperature range studied, lgV
changes by no more than one order of magnitude at lg=const. If the factors B and m are
known for T=20C, we can use equation (11) to calculate the dependence of strength on
loading rate lg=f(lgV) in the entire temperature range. By comparing this relationship with
the experimental dependence of strength on deformation rate, lg=f(lg), it is easy to find
223
lgV,gf/2*s
the dependence between the deformation and destruction rates for each of the compositions
studied.
1
1
2
-1
-2
-3
-4
-4
Fig.1
-3
-2
-1
lg,s-1
Experimental dependence of loading rate from deformation rate of the composition NC/NHC
(/=50/50) at temperatures: 1 - -20, 2 - 20, 3 - 40.
lgV,gf/2*s
1
2
3
1
0
-1
-2
-3
-4
-4
Fig.2
-3
-2
-1
lg,s-1
Experimental dependence of loading rate from deformation rate of the double-base

composition at temperatures: 1 - 15, 2 - 40, 3 - 50.
lgV,gf/c2*s
1
0
-1
Fig.3
-2
-3
-4
-4
-3
-2
-1
lg,s
-1
Experimental dependence of loading

rate from deformation rate
of the high-filled composition with
elastomer matrix at temperatures:
- =20
- =50
- =-30
- =-11
224
2 3
0,9
lg. kg/sm2
1,4
lg. kg/sm2
1,3
1,2
1,1
2
1
3
0,8
0,7
0,6
1
0,9
0,5
0,8
0,4
0,7
0,3
lg,s-1
0,6
-5
-4
-3
-2
-1
lg,s-1
0,2
-5
lgV,kg/sm s
2,3
2,1
lg. kg/sm
lg. kg/sm2
Fig 4. Dependence of strength from deformation rate Fig 5.

(1) - and loading rate
(2) - based on characteristics of lgB and m,
(3) - dependence of strength from loading
rate, based on deformation characteristics for
the high-filled copmposition 1 with
elastomer matrix at =20.
Composite at = 20.
1,9
2,6
-4
-3
-2
-1
lgV,kg/sm2s
Dependence of strength from

deformation rate
(2) - based on characteristics of lgB
and m,
(3) - dependence of strength from
loading rate, based on deformation
characteristics for the high-filled
copmposition 1
2,4
3
2
2,2
1,7
1,5
1,8
1,6
1,3
lg,s
1,1
-5
-4
-3
-2
-1
-1
lgV,kg/sm2s
lg,s-1
1,4
Fig 6. Dependence of strength from deformation rate Fig.7

(2) - based on characteristics of lgB and m,
(3) - dependence of strength from loading
rate, based on deformation characteristics for
the double-base compositon 1 at =20.
-5
-4
-3
-2
-1
lgV,kg/sm2s
Dependence of strength from

deformation rate
(2) - based on characteristics of lgB
and m,
(3) - dependence of strength from
loading rate, based on deformation
characteristics for the double-base
compositon 1 at =20.
225
, kg/sm2
, kg/sm2
80
60
40
0
200
Fig.8
250
300
350
Temperature dependence of strength

of the high-filled composition 1.
- experimental dependence,
- based on characteristics
(lgB and m) at temperatures
=50, = 35, = 5, = -50,
- based on deformatin characteristics.
2
250
Fig.9
350
300
,0
350
Temperature dependence of strength

of the high-filled composition 2.
- experimental dependence,
(lgB and m) at temperatures =35,
and =0,
- based on deformatin characteristics
400
, kg/sm2
, kg/sm2
20
300
250
300
200
200
150
100
100
50
0
200
250
300
,0
350
Fig.10 Temperature dependence of strength of the

double-base composition 1.
- experimental dependence
(lgB and m),
at temperature =50,=-20,=-50.
0
200
250
300
,0
350
Fig.11 Temperature dependence of strength

of the double-base composition 2.
- experimental dependence
(lgB and m)
at temperature =50,=-20,
=-50.
226
It is also possible to determine the relationship between loading rate and deformation
rate using the dependence of the rupturing deformation on the deformation rate at =20.
In this case, the time to failure corresponding to each deformation rate is calculated. Once
the dependence of strength on the deformation rate is known, it becomes easy to determine
the loading rate V=/t, where is the tensile strength and t is the time to failure.
Examples of such plots are shown in Figures 4-7 for two high-filled compositions with
elastomer matrix and for two double-base compositions. As expected, the plots are congruent
(see curves 2 and 3, Figures 4-7); these plots make it possible to determine the loading rate
corresponding to the standard deformation rate for each composition. Assuming that the
effect of temperature on this dependence is weak (see Figures 1-3), one can calculate the
dependence of strength on temperature using equation (11), provided that and m values are
known for the working temperature range. If the and m factors taken from literature have
been determined correctly and if the above assumption is correct, good correlation between
calculated and experimental dependences of strength should be observed for each of the
compositions studied. Figures 8-11 display calculated and experimental dependences of
strength for each of the compositions studied (see Figures 4-7). One can see that the
congruency is fairly good. A total of 12 high-filled compositions with elastomer matrices of
different nature and eight double-base compositions were analyzed. The good congruency of
calculated and experimental dependences of strength further support the adequacy of the
suggested method for calculation of EPC destruction kinetics based on tests performed using
standard tensile testing machines.
4.
CONCLUSION
A method has been worked out and suggested, which makes it possible to predict the
destruction kinetics of EPC of different nature. The method is based on tests performed
using standard tensile testing machines at =const, with no need of direct tests at =const. In
addition, the method allows inverse problems to be solved. The results of the tests indicated
above do not contradict but complement each other. They make it possible to consider
certified deformation and strength characteristics of compositions as interrelated complex
parameters, and to regard the method described above as a way of monitoring the process
technology and the validity of studies of EPC performance.
227
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
FALKOVSKY M. G., MALCHEVSKY V. A., ZHELEZNOV V. I.: Time and temperature-time

dependence of polymer strength in the small-load region, Proceedings of USSR Science and
Engineering Symposium of Polymeric Materials and Constructions. Rostov-on-Don, 1971, p.
97-103
SAVITSKY A. V., MALCHEVSKY V. A., SANFIROVA T. P., ZOSIN L. P.: Temperature
dependence of polymer strength, High Mol. Comp., Ser. A, 1974, v. 16, p. 2130-2135
V. A. MALCHEVSKY, N. A. ZARYTOVKAYA: To the problem of kinetic and critical evaluations
of mechanical destruction of energetic polymeric composites (EPC), Theory and Practice of
Energetic Materials. V.IV. China Science and Technology Press. Beijing, 2001, s. 472-480
V. A. MALCHEVSKY, N. A. ZARYTOVKAYA AND T. A. MIKHAILOVSKAYA: Mechanical
destruction of energetic polymeric composites (EPC), New trends in research of energetic
materials. Proceedings of the V. seminar. Pardubice, Czech Republic, 2002, 150-157s.
V. A. MALCHEVSKY, N. A. ZARYTOVKAYA: Mechanical destruction of energetic polymeric
composites (EPC), Contemporary problems of technical chemistry., Kazan, 2002, p. 65-73
BAILEY J: Glass Industry, 1939, vol. 20, N1, p. 21-25; N2, p. 59-65; N3, p. 95-99; N4, p. 143147.
GOIKHMAN B. D.: On one equation for the time dependence of polymer strength, High Mol.
Comp. Ser. B, 1969, 11, p. 244-245
228
A FAST NON-SELFCONSISTENT ELECTRONEGATIVITY

EQUALIZATION METHOD WITH APPLICATIONS IN THE FIELD
OF ENERGETIC MATERIALS
Didier Mathieu and Eric Germaneau
Commissariat lEnergie Atomique, BP 16, 37260 Monts, France
Abstract:
The needs for simple models of electron distribution in energetic materials and molecules
are pointed out. The derivation of linear equations for atomic charges from the
electronegativity equalization principle is outlined, with an emphasis on the assumptions
and limitations of such schemes. Then, a simple non-selfconsistent model involving a
minimum number of empirical parameters is presented and parameterized against charges
derived from first-principles calculations. It is found that Mulliken charges are more
amenable to a description in terms of electronegativity equalization than potential-derived
or even Hirshfeld atomic charges. This result suggests that charges that reproduce
electrostatic potentials satisfactorily while remaining close to Mulliken values - such as
those defined by Levy and Enescu - would be especially useful to describe electrostatic
interactions without resorting to quantum computations. Finally, applications to the
derivation of materials properties - such as sublimation enthalpies - are presented.
Keyword:
1.
atomic charges, density functional theory, sublimation.enthalpy
INTRODUCTION
1.1
Motivations
The design of new materials is made more efficient if relevant properties may be predicted a
priori, i.e. before the actual synthesis of the material. In the field of energetic materials,
densities and heats of formation are of primary interest. Impact sensitivity, which is a very
important aspect, depends on physical features, such as the concentration of defects or the
morphology of the material. In the early stages of materials design, such properties cannot be
properly estimated. However, there depend on the chemical structure of the compound through
simpler properties. For instance, impact sensitivity depends on the decomposition temperature,
which is essentially determined by molecular structure. All those properties - density, heat of
formation or decomposition temperature - depend on the electron distribution on the atoms.
While the latter can be easily derived from quantum chemical calculations, there are at least two
reasons why more efficient schemes would be useful.
First, the probability to identify new energetic materials of interest is increased as more
extensive investigations of new compounds and formulations are carried out. Therefore, in
recent years, combinatorial methods have been used to generate potentially interesting candidate
compounds.[1,2] This approach is gaining more and more interest in the field of materials in
229
general.[3] However, because huge numbers of molecules are considered, very efficient schemes
are required to estimate the relevant properties from the molecular structure. [4] In particular,
electronic structure calculations are ruled out.
On the other hand, whenever chemical reactions are involved, e.g. in studies concerned with
the initiation of decomposition in energetic materials, the electron distribution is dramatically
altered. Therefore, atomic charges calculated on isolated molecules are not useful any more.
Such processes involving chemical reactions in condensed materials may be studied with the
help of large-scale atomistic simulations. Thus, highly efficient models are needed to calculate
atomic charges on the fly. The latter should not depend on molecular topologies (i.e. on the way
atoms are bonded together) since atoms connectivities are changing in the course of chemical
reactions.
1.2
Available methods
Presently, most efficient schemes to generate atomic charges rely on the electronegativity
equalization (EE) principle. The so-called EEM method has been used with ad hoc
parameterization suitable for specific systems studied against ab initio atomic charges, and
unscreened coulomb interactions.[5] A similar model has been introduced in the field of
energetic materials - albeit with different parameters - and applied successfully to estimate
impact sensitivities of nitroaromatic compounds.[6] Presumably, such predictions were made
possible by the fact that the sensitivities differences within this class of compounds mostly
depend on decomposition temperatures. Surprisingly, in some of the early models, a dielectric
constant was sometimes reported to improve the fit, while in principle the EEM equations are
not affected, as discussed below. More recent mehods use no dielectric constant, but introduce a
short-range screening of coulomb interactions to describe interpenetration effects. Among such
approaches, the QEq model is especially popular.[7] However, the quality of the charges thus
obtained is not well documented. In addition, this method is relatively costly in computer time,
as it is self-consistent. Most other available models for atomic charges rely on semi-empirical
calculations of the electronic structure, at least for systems. In addition, they depend on the
molecular topology and therefore, cannot be used for atomistic simulations affecting chemical
bonds. [8-12]
The model presented here focusses on efficiency, economy of adjustable parameters and
independence on chemical bonds. It is quite similar to the method implemented as part of a
reactive force field for hydrocarbons, named ReaxFF.[13] However, it is parameterized against
atomic charges from quantum chemical calculations, and does not require empirical parameters
for electrostatic interactions, in contrast to ReaxFF. Moreover, it may be applied to some
heteroatoms. A simple derivation of the model equations is outlined. Then, molecular
electrostatic potentials generated with this model are compared with those obtained using the
charge model associated with the MMFF force field.[14] Finally, an application relevant to the
prediction of solid-state heats of formation of energetic materials is presented.
230
2.
THEORY
2.1
Assumptions
The present model may be obtained through simple approximations to the density functional
E[] which yields the ground-state total energy of the system with electronic density (r). The
basic step consists in replacing the density function (r) by a set of electron populations rk at
sites k, such that (r) is given by :
(r) =
k k(r)
(1)
The sites k available for electrons may be associated with bonds, lone pairs, localized
orbitals or atoms. In view of deriving a simple practical scheme, only atoms sites are considered
in the sequel. In that case, it is convenient to introduce the net atomic charge qk = rk - Zk for
every atom k, where Zk is the charge of the nucleus. To obtain an approximate charge
distribution, the density which minimize the functional E[] must be sought, under the
requirement that it integrates to the total number of electrons in the system. In principle, with
only one site per atom, both the net charges qk and the site distributions k(r) should be varied to
describe the trial densities (r). However, for the sake of simplicity, only variations of the
charges will be considered. This is of course a severe approximation, but it is consistent with the
aim to obtain atomic charges, that could for instance mimic charges derived from quantum
chemical calculations. In that case, since the k(r) are fixed, the functional E[] may be viewed
as a genuine funtion of the charges, i.e. E(q). At this stage, the following decomposition is
assumed :
E(q)=
Ek +
k,k
Vk,k
(2)
There are clearly many possible choices for such a decomposition. Ideally, Ek and Vk,k
should be allowed to depend on the precise state and environment of the atoms (e.g. on their
hybridization). However, since we want to avoid any explicit dependence on the bonding state,
the parameters involved in those energy terms are assumed to be transferable for the same
chemical elements in different molecular environments.
To proceed, we focus on the dependence of Ek on the atomic charge qk on site k, through a
development up to second order :
Ek = EK(0) + K qk+ (1/2) K qk2
(3)
The electronegativities K and chemical hardnesses K are labelled with an upper index K in
order to emphasize that they do not depend on the specific atom at site k, but only on its
chemical identity K. In addition to this expression for Ek, a guess at the interaction energy Vk,k
is needed. For large interatomic distances rk,k it must reduce to the coulomb expression
qkqk/rk,k. This coulomb energy is damped as the charges come in close contact owing to
penetration effects. Therefore, a damping function f(r) is introduced, with f(r) 1 as r :
Vk,k (qk, qk) = qkqk/r k,k f(rk,k)
(4)
Although so far no assumption was made on the local charge distributions k describing the
electron density, the fact that Vk,k depends only on the interatomic distance implies spherical
atoms, in line with the use of atomic charges. In the original EEM method, no damping function
231
is used . Thus the dependence of interatomic energies on the charges is simply given by qkqk/r
k,k even for chemically bonded atoms. In the QEq method, the damping function is expressed in
terms of the overlap of the interpenetrating electron clouds, while in the more recent ReaxFF
scheme, an empirical damping function is used, involving ad hoc parameters. In the present
scheme, the Klopman interpolation is used, i.e. in atomic units :
Vk,k = qkqk [r k,k2 + (1/4)(1/UK+1/UK) 2] -(1/2)
(5)
This expression involves the Hubbard parameters UK defined as the differences between
first ionization potentials and electron affinities. In our experience, within the self-consistent
charges density functional tight binding scheme[15], such an interpolation provides electrostatic
energies in close agreement with experiment for molecular ionic crystals, despite the use of the
Mulliken definition for the charges. In the next sections, the proportionality constant Jk,k =
Vk,k/(qkqk) is introduced for convenience.
2.2
Model equations
The working equations are obtained through minimization of the energy wrt the atomic
charges qk. The constraint that the qk sum to the total charge Q of the system may be associated
with a Lagrange multiplier . It is straightforward to show that for every atom k :
K + K qk +
k?k
Jk,kqk =
(6)
Thus, may be defined as the electronegativity of every atom in the system, independent of
the atom considered at equilibrium. Some simple models do not include explicitly the coulomb
interaction of atom k with the others: to make up for this approximation, only a partial
electronegativity equalization is carried out. For a system with N atoms, electronegativity
equalization leads to N-1 equations :
(K - Jk+1,k)qk - (K+1 - Jk,k+1) qk+1 +
k?k,k+1
(Jk,k - Jk+1,k)qk = K+1 - K
(7)
Taking into account the additional constraint on the total charge Q of the system :
qk = Q
(8)
a system of N linear equations for the N unknown charges qk is obtained. From those
equations it is clear that introducing a dielectric constant >1 in the ansatz for the interaction
energy between atoms amounts to a scaling of the electronegativities and hardnesses. Therefore,
there is no point to take >1.
On the other hand, no precise definition has been introduced so far for the three ingredients
involved in the model, namely the atomic parameters (K, K), the interatomic interaction Vk,k
and the charges qk. Assuming a given definition for the atomic parameters - such as theoretical
values for free atoms - and a reasonable expression for Vk,k such as the Klopman expression
used here, the atomic charges obtained will have no reason to match any of the many charges
definitions already available. An alternative consists in selecting target atomic charges to be
reproduced through electronegativity equalization, after optimization of K and K. This
procedure is described in the following section.
232
3.
PARAMETERIZATION
A panel of 200 molecules was considered for the calibration and validation of parameters in
view of deriving empirical atomic charges good enough for an approximate description of
electrostatics. In principle, potential-derived atomic charges - by definition - are the best
possible values for this purpose. Unfortunately, they are somewhat ill-defined as many different
sets of values can yield potentials of similar quality. Therefore, they are unlikely candidates for
reference charges to be estimated through electronegativity equalization. Mulliken charges are
especially straighforward to derive from molecular orbitals, and do not suffer from this
numerical problem. However, they are very dependent on the basis set employed. Finally, the
Hirshfeld charges are commonly considered as especially significant, and are the preferred
values for advanced applications of atomic charges, such as the estimation of localized Fukui
functions. In this work, three parameters sets were derived in a attempt to estimate the three
charges definitions : potential-derived (CHELPG), Mulliken and Hirshfeld. Reference charges
were calculated using density functional theory, either at the B3LYP/6-31G* (for CHELPG and
Mulliken charges) or VWN/DNP (for Hirshfeld charges) level. A least-square adjustement was
carried out through minimization of the following quantity involving the differences between
charges derived from the EE formalism and reference charges:
F=
(qk(EEM) - qk(ref.))2
(9)
where the index m stands for the molecules in the calibration panel and k runs over atoms in
molecule m. Provided physically reasonable values were used for the initial guess, the algorithm
systematically converges to the same minimum. To estimate the significance of the regression, a
cross-validation was carried out. Namely, random regression panels were generated by
extracting 80% of the whole panel. The parameters obtained from the optimizations on those
restricted panels were then applied to the 20% remaining molecules. The charges predicted for
the molecules outside the regression panel were recorded with the associated reference charges
for subsequent assessment of the model. The optimized values of the model parameters are
reported in Table 1 below (in fact, the parameters calibrated against potential-derived charges
had been reported previously in a different context[16]). The expected trends may be noted. For
instance, electronegativities are largest for oxygen, fluorine and nitrogen atoms. Moreover, the
relatively smaller differences between the atomic electronegativies obtained from Hirshfeld
charges are consistent with the small magnitude observed for such charges.
Table 1. Optimized parameters (electronegativities and chemical hardnesses)
for chemical elements encountered in the calibration panel, in atomic units.
Potential-derived
H
C
N
O
F
Cl
K
0.206
0.219
0.270
0.670
0.259
0.180
K
0.492
0.334
0.460
1.386
0.588
1.354
Mulliken
K
0.108
0.170
0.242
0.510
0.596
0.133
K
0.637
0.370
0.481
0.981
1.811
0.385
Hirshfeld
K
0.359
0.378
0.404
0.546
0.427
0.383
K
0.585
0.395
0.500
1.301
0.823
0.328
233
Fig 1.
MEPs for acetamide (see text for details)
234
The correlations coefficients between reference and predicted (during the cross-validation
step) charges are 0.900 for potential-derived charges, 0.990 for Mulliken charges and 0.936 for
Hirshfeld charges. As expected, potential-derived charges are difficult to mimic as their precise
value is numerically ill-defined. In contrast, Mulliken charges may be very well approximated
on the basis of the EE principle, thus allowing considerable computer time savings.
Unfortunately, the Hirshfeld charges much more useful in practice than Mulliken charges
are not so well described. The fact Mulliken charges derived from molecular orbitals are much
better described by the present scheme than Hirshfeld charges derived from the electron density
was unexpected. In the next sections, the use of the parameters optimized against potentialderived charges to generate electrostatic potentials and associated descriptors is discussed, in
view of applications for energetic materials.
RESULTS AND APPLICATIONS
4.
4.1
Molecular electrostatic potentials
Despite the underdetermination inherent to the derivation of charges from electrostatic

potentials, their counterparts derived from the present scheme are able to reproduce the main
features of molecular electrostatic potentials (MEPs). Indeed, the electrostatic potential they
provide appear to be of a quality comparable to that calculated from MMFF charges, for
instance. This is illustrated on Figures 1 and 2 where MEPs are presented for acetamide and
formyl fluoride. For each figure, the MEPs presented were obtained respectively from a fit
against B3LYP/6-31G* densities (top), from bond parameters as implemented in the MMFF
force field (middle) and from EE using the first set of parameters reported in Table 1 (bottom).
The advantages of the present approach over MMFF charges stem from its much reduced
number of empirical parameters and more significantly from its independence on bonds
network, which make it able in principle to describe bond breaking situations. On the other
hand, the ability of the present model to mimic Mulliken charges suggests that its calibration
against Levy-Enescu (LE) charges[17] might provide even more accurate MEPs. Indeed, in the
present work, the quality of the EE atomic charges suffer from the fact they were calibrated
against potential-derived charges whose physical significance is questionable.
4.2
Electrostatic descriptors
It has been emphasized by Politzer and co-authors that MEPs can be used to define
electrostatic descriptors suitable for prediction of complex materials properties[18]. For this
purpose, they define the average V+ of the electrostatic potential values on the molecular
surface, restricted to regions where this potential is positive, and a parameter + defined as the
variance of these values. For practical purposes, those quantities are evaluated though averages
based on values of the potential computed at n+ points sampling the positive part of the surface:
+2 = (1/n+)
[V(r )-V ]
i
+ 2
(10)
-
Similar definitions are introduced for V and derived from the negative regions of the
molecular surface. Then, symmetric descriptors and are defined:
2 = +2 + 2
(11)
235
Fig 2.
MEPs for formyl fluoride (see text for details)
236
= +2 2/ 2
(12)
The values of and reflect the significance of electrostatic interactions. The descriptor
depend on the presence of regions with very positive or negative values of the potential. The
value of characterizes the degree of balance between positive and negative regions of the
surface. It may vary between zero and 0.25. Values close to 0.25 reflect the ability of the
compounds to interact through both its positive and negative surfaces. To illustrate the meaning
of those descriptors, some values are reported in table 2 for simple molecules. The Van der
Waals definition was adopted for the molecular surface.
Table 2. Values of electrostatic descriptors for simple molecules. The value of is
expressed in kJ/mol for convenience, through multiplication of the potential by
the electron charge.
Compound
Benzene
Aniline
Nitrobenzene
Nitroaniline
TATB
0.069
0.245
0.118
0.219
0.232
(kJ/mol)
16.3
47.3
71.5
93.2
87.4
This table illustrates the small values of for hydrocarbons, such as benzene, and the
influence of electron withdrawing and repelling substituents leading to values close to the
theoretical maximum of 0.25 when complementary moieties such and the amine and nitro
groups are present. The value of is small for benzene deprived of polar groups. In contrast
to , it would be small also for a large hydrocarbon backbone with few substituents.
4.3
Sublimation enthalpies
The electrostatic descriptors derived from EE charges are designed to express the
dependence of complex materials properties on the electrostatic interactions within the material.
Indeed, the influence of the latter on properties is difficult to express analytically owing to their
long-range character. In the lack of a known crystal structure, no lattice sums can be carried out
and one has to be satisfied with simple descriptors such as and . The first application we
made of this approach was for the calculation of standard sublimation enthalpies (at 298K). A
preliminary account of this work has been mentioned previously in relation with the evaluation
of solid state formation enthalpies[16].
To get a working equation for estimating sublimation enthalpies, the latter are written as the
sum of three contributions :
subHo = Evdw + Eelec + Ehb
(13)
where the first and third terms on the rhs represent short-range contributions to subHo,
namely the Van der Waals and hydrogen bonds contributions. Explicit expressions for these two
terms Evdw and Ehb are derived from simple considerations[16]. Of special interest here is the
237
second term Eelec which represents the electrostatic contribution. To be significant, the
electrostatic potential at the molecular surfaces must be large and the molecule should
preferably exhibit a good balance between positive and negative regions. Therefore, Eelec is
simply assumed to be proportional to the product of and . Since kJ/mol units are used for ,
the proportionality coefficient a is dimensionless.
Fig 3.
Calculated versus observed sublimation enthalpies (kJ/mol).
The value of a was fitted together with the parameters involved in Evdw and Ehb. The
adjustment using a panel of 245 compounds with sublimation enthalpies recorded in the NIST
WebBook database[19] yielded an optimal value a=813. With regard to subHo, the calculated
values are plotted against the experimental ones in figure 3. The outlier on this plot is p-Xylene,
for which an experimental value of 153 kJ/mol is reported in the database. The calculated value,
almost three times smaller, is more consistent with the data usually observed for hydrocarbons
compounds of this size[20]. Indeed, the original reference cited in the WebBook contains no
mention of p-Xylene[21]. Therefore, the value of 153 kL/mol seems to have been recorded in the
database by error.
The filled circles in figure 3 denote hydrogen-bonded crystals. It is apparent from this plot
that hydrogen bonds contribute significantly to subHo. The four datapoints with subHo > 250
kJ/mol are associated with porphyrins derivatives. It is worth pointing out that calculated values
of subHo are less accurate for hydrogens-bond crystals (average absolute error d=17 kJ/mol)
than for the others (with =12 kJ/mol). This suggests that the errors on calculated subHo arise
essentially because of Ehb, despite the more empirical character of Eelec. This is understandable
as all kinds of hydrogen bonds are treated on the same footings in the present model for Ehb[16].
238
5.
CONCLUSION
Preliminary investigations of a simple electronegativity equalization model are reported.

With regard to the evaluation of electrostatic potentials, it compares fairly well with heavily
parameterized methods that depend explicitly on chemical bonds. While it can in principle
handle situations where bonds are formed or broken, the usefulness of the present model in such
situations remains to be investigated. An interesting finding concerns the fact that Mulliken
charges are specially easy to reproduce within the present approach, suggesting the use of LevyEnescu charges to improve the electrostatic potentials thus obtained. Further work will be
needed to overcome some limitations common to every electronegativity equalization scheme,
such as its inability to distinguish localized Fukui functions for electrophile or nucleophile
reactions, and the fact that isolated molecules dissociate into fragments with non-integer charges
determined by electronegativity and hardnesses only.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
T. S. PIVINA, V. S. SHCHERBUKHIN, M. S. MOLCHANOVA AND N. S. ZEFIROV:

Computer-Assisted prediction of novel target high-energy compounds, Propellants Explos.
Pyrotech. 20, p.144-146, 1995
P. BOUGRAT: Dveloppement dun algorithme gntique pour la recherche de molcules
explosives nouvelles, report, university of Nancy, 1997
V. VENKATASUBRAMANIAN, K. CHAN and J. M. CARUTHERS: Computer-auded
molecular design using genetic algorithms Computers Chem. Engng 18, p.833-844, 1994
D. MATHIEU and P. SIMONETTI: PIMM as an efficient scheme to estimate gas-phase
formation enthalpies of energetic molecules Molecular Engineering 8, p.121-134, 1999
W. J. MORTIER, S. K. GHOST and S. SHANKAR: Electronegativity equalization method for
the calculation of atomic charges in molecules J. Am. Chem. Soc., 108, p.4315-4320, 1986
D.-C. ZHANG and Y.-Q. ZHANG : Partial atomic charges calculations in nitrobenzenes J.
Energetic Materials, 15, p.269-281, 1997
A. K. RAPPE and W. A. GODDARD III: Charge Equilibration for Molecular Dynamics
Simulations, 95, p.3358-3363, 1991
J. GASTEIGER and M. MARSILI: Iterative partial equalization of orbital electronegativity - a
rapid access to atomic charges Tetrahedron, 36, p.3219-3288, 1980
K. T. NO, J. A. GRANT and H. A. SHERAGA: Determination of net atomic charges using a
method of partial equalization of orbital electronegativity method. 1. Application to neutral
molecules as models for polypeptides J. Phys. Chem. 94, p.4732-4739, 1990
P. R. GERBER: Charge distribution from a simple molecular-orbital type calculation and nonbonding interaction terms in the force field MAB J. Comput.-Aid. Mol. Des. 12, p.37-51, 1998
Y. CONG and Z.-Z. YANG : General atom-bond electronegativity equalization method and its
application in prediction of charge distribution in polypeptides Chem. Phys. Lett., 316, p.324329, 2000
A. JAKALLIAN, B.L. BUSH, D. B. JACK and C. I. BAYLY: Fast , efficient generation of highquality atomic charges. AM1-BCC Model: I. Method J. Comp. Chem., 21, p.132-146, 2000
A. C. T. VAN DUIN, S. DASGUPTA, L. LORANT and W. A. GODDARD III: ReaxFF: a
reactive force field for hydrocarbons J. Phys. Chem. A 105, p.9396-9409, 2001
239
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
T. A. HALGREN: Merck Molecular Force Field. II. MMFF94 van der Waals and electrostatic
parameters for Intermolecular Interactions J. Comp. Chem., 17, p.520-552, 1996
M. ELTSNER, D. POREZAG, G. JUNGNICKEL, T. FRAUENHEIM, S. SUHAI AND G.
SEIFERT: A selfconsistent-charge density-functional tight binding scheme Mat. Res. Soc. Symp.
Proc. 491, p.131-136, 1998
D. MATHIEU and P. SIMONETTI: Evaluation of solid-state formation enthalpies for energetic
materials and related compounds Thermochimica Acta, 384, p.369-375, 2002
B. LEVY and M. ENESCU, Theoretical study of methylene blue: a new method to determine
partial atomic charges; investigation of the interaction with guanine J. Mol. Struct. (Theochem),
432, p.235-245, 1998
P. POLITZER, J. S. MURRAY, M. E. GRICE, M. DESALVO and E. MILLER: Calculation of
heats of sublimation and solid phase heats of formation Mol. Phys. 91, p.923-928, 1997
P J. LINSTROM: NIST WebBook system, http://webbook.nist.gov/chemistry
D. MATHIEU, P. BOUGRAT: Model equations for estimating sublimation enthalpies of organic
compounds Chem. Phys. Lett. 303, p.601-606, 1999
P. M. BURKINSHAW and C. T. MORTIMER, Enthalpies of sublimation of transition metal
complexes J. Chem. Soc. Dalton Trans. p.75-77, 1984
240
CHEMICAL DECOMPOSITION OF TRIACETONE TRIPEROXIDE

AND HEXAMETHYLENETRIPEROXIDEDIAMIDE
Robert Maty
Department of Theory and Technology of Explosives, University of Pardubice
532 10 Pardubice, Czech Republic
Abstract:
Triacetone triperoxide (TATP) and hexamethylenetriperoxidediamine (HMTD) are well
known primary explosives often called improvised that are frequently prepared in
improvised conditions. These organic peroxides are very often components of
pyrotechnist's findings. This paper summarizes experimental results of chemical
decompositions of these peroxides.
Keyword:
1.
chemical decomposition, improvised explosives, organic peroxides,

triacetone triperoxide, TATP, hexamethylenetriperoxidediamine, HMTD.
INTRODUCTION
Triacetone triperoxide (TATP, 3,3,6,6,9,9-hexamethyl-[1,2,4,5,7,8]hexaoxonane) and

hexamethylenetriperoxidediamine (HMTD, 3,4,8,9,12,13-hexaoxa-1,6-diaza-bicyklo[4,4,4]
tetradecane) are very powerful primary explosives[1,2]. TATP and HMTD are improvised
explosives that are prepared very frequently in Czech Republic [3,4]. The know-how of their
production is very often mentioned on the web sites that deal with illegal production of
explosives. The synthesis of these organic peroxides is very simple and it can be carried out
even without deeper knowledge of chemistry or proper experience with explosives. Their
preparations are relatively safe and can be carried out practicaly anywhere. The production
apparatus is undemanding and starting materials are substances that can be purchased in
many drugstores.
The simplicity of the production of TATP and HMTD allows the preparation of both of
these organic peroxides in large quantities in home conditions. They were both already used
by terrorist organizations in some countries (Israel [5], USA [6,7], United Kingdom [8]).
241
1.1
Triacetone triperoxide (TATP)
Triacetone triperoxide (I) can be prepared by the reaction of acetone with hydrogen
peroxide in the presence of acid catalyst [1,9]. Acetone is being sold as a paint thinner,
hydrogen peroxide as a bleaching agent or antiseptic. Hydrochloric acid and sulfuric acid are
also easily accessible in drugstores.
TATP is very volatile substance. At room temperature TATP loses about 6.5 wt% in
24 hours [1], 68.6 wt% in 14 days [10]. At 50C TATP loses 1.5 wt% in 2 hours [1].
TATP is extremely sensitive to friction and impact. It was also reported [1] that TATP
can detonate even when it contains up to 25% water or under water. Some solvents were
tested to reduce sensitivity of TATP to mechanic stimulus in order to reduce risk of
accidental initiation during handling. It was observed that saturated solutions of TATP in
toluene (approximately 28 wt %) can not be initiated by detonator. This solution is
chemically stable and it may be stored in a closed vessel [11].
Decomposition of TATP
Decomposition of TATP samples can be possibly achieved in three ways [11]: (a) burning
of TATP solution, (b) thermal degradation or (c) chemical decomposition. The burning of
TATP solutions or thermal degradation can be hazardous.
TATP is practically insoluble in water. No hydrolysis of TATP in water has been
observed. It was reported that dilute sulfuric acid decomposes TATP to produce acetone and
hydrogen peroxide. It was found that after 1 hour of boiling this substance decomposes
quantitatively.
TATP is relatively resistant to chemical reduction. Little or no reaction was observed
with many of reducing agents commonly used for decomposition of peroxides (acidified
solution of the potassium iodide with 18-Crown-6, FeSO4, (NH4)2SO4.FeSO4, Ph3P)[11].
However TATP was found to be readily reduced by an excess of tin (II) chloride in ethanol
(less than 1 hour reflux or 6 hours at 65C). The reaction does not take place at room
temperature. The details of this method are given in reference [11].
H 3C
CH 3
H 3C
CH 3
O
O
O
CH2
OO
C H2
C H2
OO
C H2
CH2
OO
C H2
C
CH 3
(I)
CH 3
(II)
242
1.2
Hexamethylenetriperoxidediamine
Hexamethylenetriperoxidediamine (II) can be prepared by the reaction of hydrogen

peroxide with hexamethylenetetramine in presence of acid (most often citric acid)[2]. All of
these substances are accessible in drugstores. Hexamethylenetetramine is the main
component of the fuel for tourists cooker, citric acid is used as a relish and hydrogen
peroxide is used as a germicide and bleach agent.
HMTD is very sensitive to friction and impact. HMTD is not volatile substance. Dry
HMTD is not compatible with most of the common metals.
Decomposition of HMTD
HMTD is slightly soluble in water (0.01 g/100 g water at 22C) as well as in organic
solvents [2]. It is unstable in contact with water, it complete decomposes after 4 months.
HMTD decomposes in contact with alkalis with liberation of ammonia and it decomposes
with warm water and acids to formaldehyde, with boiling water relatively quickly [12].
2.
EXPERIMENTAL
Triacetone triperoxide was prepared by the reaction of acetone with 30% hydrogen
peroxide in presence of sulfuric acid at the temperature of 10C [1]. HMTD was prepared by
the reaction of hexamethylenetetramine and 30% hydrogen peroxide in presence of citric
acid at temperatures below 15C [2].
Relatively little information was found about chemical decomposition of TATP and
HMTD in open literature. On the basis of bibliographic search three groups of chemical
agents for chemical decomposition were selected: acids, alkalis and reduction agents.
The tests of decomposition were carried out with samples of TATP and HMTD in
quantity 0,5 g. The samples were placed into Petri dish or test tube and poured with
decomposition reagent. Organic peroxides were under the surface of decomposition reagents
during all reaction process. After the addition of the reagent the reaction course was
observed.
Water solutions (25 %) of the sulfuric acid, hydrochloric acid and acetate acid were used
for decomposition of TATP and HMTD samples. The acid solution was used in the amount
of approx. 5 ml was poured over the organic peroxide and the test tube was sharked. The
tests were made at temperatures 22C, 50C and 75C. The results are summarized in table
1.
The same conditions were used for reaction TATP and HMTD with basis. The 25%
water solution of the sodium hydroxide, ammonia and sodium carbonate were used. The
decomposition was carried out at 22C, 50C and 75C. The results are shown in table 2.
The following reducing agents were used for decomposition of the organic peroxides:
acidified solutions of the KI and KBr, ethanol and water acidified solution of SnCl2, zinc in
acids, HBr and HI. The reactions were carried out at 22C.
243
3.

3.1
Decomposition of TATP and HMTD by acids and bases
The results of the decompositions by using of 25% solutions of the sulfuric acid,
hydrochloric acid and acetic acid are shown in the table 1.
Table 1. Reaction of TATP and HMTD with acids.
Acid
22C
without
TATP decomposition
Sulfuric
within 24 hours
acid
decomposes within
HMTD
4 hours
without
TATP decomposition 24
Hydrochloric
hours
acid
decomposes within
HMTD
20 minutes
without
TATP decomposition
within 1 week
Acetic
acid
without
HMTD decomposition
within 1 week
50C
decomposes within
6 hours
75C
decomposes within
8 minutes
decomposes within
1 hour
decomposes within
6.30 hours
decomposes within
3 minutes
decomposes within
8 minutes
decomposes within
1 hour
without
decomposition
within 1 week
decomposes within
4 hours
decomposes within
3 minutes
decomposes within
4 hours
decomposes within
10 minutes
Decomposition of TATP with acids proceeds only at higher temperatures. It takes

approx. 6-7 hours at 50C and 8-10 minutes at 75C. Process of the decomposition of
HMTD is similar to the decomposition of TATP, but faster. It takes a few hours at 22C,
approx. 1 hour at 50C and approx. only 3 minutes at 75C with H2SO4 and HCl.
Decomposition of HMTD with acetic acid takes approx. 4 hours at 50C and approx.
10 minutes at 75C.
The 25% water solution of sodium hydroxide did not decompose TATP even at 75C
after 10 hours. HMTD was decomposed within 3 hours at 75C. 25% water solution of
ammonia and sodium carbonate did not decompose TATP and HMTD even at 75C after
5 hours. The same results were obtained with 25% water solution of ammonia.
3.2
Reduction of TATP and HMTD
Little reaction was observed with acidified solution of potassium bromide

(0,5 g KBr + 3 ml water solution (1:1) H2SO4 ). There was not a visible decomposition of
TATP by this solution, the reaction mixture only turned yellow in vicinity of TATP crystals.
HMTD reacted very slowly, reaction mixture turned brown. The same results were observed
with acidified solution of the potassium iodide.
The results of decomposition of TATP and HMTD with water and ethanol solutions of
the tin (II) chloride are shown in table 2.
244
Table 2. Reaction of TATP and HMTD with SnCl2 solutions.

2 ml saturated water solution
water solution of SnCl2
of SnCl2
of SnCl2+ 2 ml conc. HCl
HMTD
2 ml saturated ethanol
ethanol
solution of SnCl2
solution of
1 ml ethanol + 1 ml saturated
SnCl2
ethanol solution of SnCl2
water solution of SnCl2
of SnCl2
of SnCl2+ 2 ml conc. HCl
TATP
2 ml saturated ethanol
ethanol
solution of solution of SnCl2
SnCl2
violent reaction, HMTD

decomposed immediately
decomposition of HMTD
rapid reaction, HMTD
decomposed immediately
slow decomposition of
HMTD within 1 hour
without reaction after
1 week
1 week
1 week
Zinc with water solution of hydrochloric acid was used for decomposition of TATP and
HMTD. Zinc reacts with hydrochloric acid and liberated hydrogen reduces organic
peroxides. TATP and HMTD rapidly decompose by zinc with concentric hydrochloric acid
within a few seconds. Zinc with dilute HCl (to 1:3) decomposes HMTD easily but TATP
decomposes more slowly than HMTD and certain concentration of HCl was required. A lot
of the heat is liberated during the reaction due to exothermic reaction of zinc with HCl. The
higher concentration acid liberated more heat.
Good results were obtained for decomposition of TATP with concentric hydroidic acid.
During this reduction iodine is formed. HMTD reacts with HI very violently. The
hydrobromic acid appears to be more suitable for decomposition of HMTD.Unfortunately
TATP reduces with concentric HBr after 3 days. Crystals size has a considerable influence
on the decomposition time. A fine TATP decomposes in concentric HI after a few minutes.
Decomposition of larger size crystals of TATP can take even a few hours (crystal of TATP
of size 3 mm decomposes after a few hours or days). The results of decompositions of TATP
and HMTD with HI and HBr are shown in table 3.
245
Table 3. Reaction of TATP and HMTD with HBr and HI

wt. ratio
HX : water
progress of the reaction
1:0
fine crystals decomposed within 4 days, higher

crystals without decomposition within 1 week
1:1
without decomposition
1:0
decomposition within a few sec., Br2 formed
1:1
decomposition within 15-45 min., Br2 formed
1:0
fine crystals decomposed within 1-3 min. total

decomposition of higher crystals within 2 hours
1:1
without decomposition within 1 week
1:0
instant violent decomposition
1:1
decomposition within -1 min., I2 formed
TATP
HBr
HMTD
TATP
HI
HMTD
4.
CONCLUSION
The method of chemical decomposition of two very common improvised explosives,

triacetone triperoxide (3,3,6,6,9,9-hexamethyl-[1,2,4,5,7,8]hexaoxonane) and hexamethylenetriperoxidediamine (3,4,8,9,12,13-hexaoxa-1,6-diaza-bicyklo[4,4,4]tetradecane)
was tested. Acids, bases and reduction agents were chosen for chemical decomposition.
Sulfuric and hydrochloric acids as 25 % water solutions completely decomposed TATP
only at higher temperatures (50C, 75C). HMTD decomposed at normal temperature.
25% solution of sodium hydroxide did not decomposed TATP even at 75C after 10 hours.
HMTD was easy decomposed at 75 C. 25% water solution of sodium hydroxide, ammonia
and sodium carbonate did not decomposed TATP and HMTD (sodium hydroxide
decomposed only HMTD at 75C).
Reductions of TATP and HMTD were carried out at room temperatures. Acidified
solutions of the potassium bromide and iodide did not decompose TATP and HMTD. Water
or ethanol solution of tin (II) chloride did not decompose TATP, but easily decomposed
HMTD. Zinc with water solution of the hydrochloric acid decomposed TATP and HMTD,
but this reaction is very exotermic. It could be dangerous due to possible initiation of the
organic peroxide. Hydrobromic acid is suitable for decomposition of HMTD, but not for
TATP. Hydroiodic acid decomposes TATP, HMTD reacts with HI too violently.
Variety of chemical agents was examined for TATP and HMTD decomposition. Several
chemical agents (H2SO4, Zn+acid, HBr, HI, SnCl2) were found to be suitable for successful
decomposition of TATP and HMTD.
246
REFERENCES
[1]
FEDOROFF T.: Encyclopedia of explosives and related items, Volume 1, A 41, Picatinny
Arsenal, New Jersey, USA (1960).
[2] FEDOROFF T.: Encyclopedia of explosives and related items, Volume 7, H 83, Picatinny
Arsenal, New Jersey, USA (1975).
[3] http://pxd.zde.cz/ [09.11.2002]
[4] http://kattarit.vyrobce.cz/ [24.2.2003]
[5] ZININ S., KRAUS S., GLATTSTEIN B.: Proc. Int. Sym. on the analysis and detection of
explosives. Quantico USA, 137 (1983).
[6] EVANS H. E., TULLENERS F. A. J., SANCHES B. L., RASMUSSEN C. A.: J. Forensic. Sci.; 31,
1119 (1986).
[7] WHITE G. M.: J. Foresic. Sci.; 37; 652 (1992).
[8] Middle-class woman planted embassy bomb. The independent, London, 1996 Oct 8:7.
[9] MILAS N. A., GOLUBOVIC A.: J. Amer. Chem. Soc.; 81, 6461 (1959).
[10] PHILLIPS A. J.: Picatinny Arsenal Technical Report 1202 (1942).
[11] BELLAMY, ANTHONY J: J. Forensic Sci., 44(3), 603 (1999).

[12] FICHEROULLE H., KOVACHE A.: Memorial pouders, 31, 18 (1949)
247
QUANTITATIVE ANALYSIS OF SOME TECHNOLOGICAL

COMPONENTS USED IN INHIBITED HOMOGENEOUS ROCKET
PROPELLANTS BY MEANS OF THIN LAYER CHROMATOGRAPHY
Maciej Miszczak, Ewa Szymanowska, Beata migelska
and Bogumia Bachnio
Research Department of Combat Means, Military Institute of Armament Technology,
Wyszyskiego 7, Str., 05-220 Zielonka, POLAND
Abstract:
In this paper new original methods of thin layer chromatography (TLC) analysis
for quantitation of plasticizers, energetic additives and stabilizers being components
of inhibited double-base propellants used in rocket motors assigned for launching
of explosive long line changes to clearance of mine fields are presented.
Analytically measured substances were as follows: nitroglycerine (NG),
dibutylphthalate (DBP), 2,4-dinitrotoluene (2,4-DNT), 2,4,6-trinitrotoluene (TNT),
ethylcentralite (EC) and 4-nitro-ethylcentralite (4-NO2-EC). Above mentioned
substances were separated by TLC and next they were quantitatively measured with use
of UV/VIS spectral densitometry. On the basis of TLC some migration and nitration
processes of analyzed substances were observed in the propellant and in its inhibitor.
Keyword:
1.
TLC analysis, plasticizers, additives, double-base propellants
INTRODUCTION
Problem of analysis of composition changes in inhibited homogeneous rocket

propellants used to launch explosive long line charges for clearance of mine fields,
especially analysis in the area of contact layer between the propellant grain and its inhibitor,
appeared in Polish Army after recording of failure of burning processes of this type
propellants in rocket chambers. This malfunction of rocket motors operation was manifested
by too high maximum pressure of burning products, by too large differences in burning
temperatures and in burning durations. These observations were confirmed by technical
condition (state) of rocket chambers after propellant burning. For example, in several
chambers their lacquer layer was completely or partly burnt or it was not burnt in any
degree.
In some cases rocket motors blasted during static test firing. Inspections of rocket
chambers before and after their test firing showed that failure work of rocket motors was
caused because of propellant inhibitor adhesion to inner walls of rocket chamber and
because of inhibitor stratification from propellant grain surface. So it was concluded that for
disturbances in burning of inhibited propellants in rocket motors can be responsible
migration processes of plasticizers, energetic additives and stabilizers going on in the area of
contact layer between propellant and inhibitor materials. Relatively wide variety of world
literature describes problem of substances migration in propellants, in their inhibitors and
between propellant and inhibitor [1-17].
In this work, authors tried to estimate the possibility of TLC use for quantitative
determinations of technological ingredients as plasticizers, energetic additives and stabilizers
in aspect of their migration in materials being in contact and creating propulsive rocket
system i.e. solid homogeneous propellant and its inhibitor.
248
2.
EXPERIMENTS
2.1
Materials and Apparatus
For analytical measurements double-base inhibited propellant produced in 1986,

withdrawn from rocket motors assigned for launching of explosive long line charges to
clearance of mine fields were selected. The propellants grains contained nitrocellulose (NC),
NG, DNT, TNT, EC, 4-NO2-EC and the inhibitor contained DBP.
Samples of the same lot were divided into two parts. In all samples inhibitor
was in contact with propellant. One part of samples was heated for 1h at 120C. All samples
were analyzed to observe difference of their composition especially caused by migration
processes going in contact layers of propellant and inhibitor.
Standards of analyzed substances:
NG, EC, 4-NO2-EC - pure for analysis (Plastics Works - Pionki (POLAND)),
2,4-DNT, TNT - pure for chromatographic analysis (Institute of Organic
Industrial Chemistry - Warsaw (POLAND),
DBP - pure for analysis (Merck).
Solvents: acetone, hexane, benzene, ethyl acetate -pure for analysis.
Agent for visualization of NG - N,N,N',N'-tetramethyl-1,4-phenylenediamine (so called
blue Wrsters salt) pure for analysis (Merck).
Stationary phase: HPTLC plates (Merck, No: 5548) with fluorescent agent F-254.
Apparatus: CS-9000 scanning densitometry (Shimadzu), autosampler Linomat IV-Y
(Camag) for spray-on linear sampling of analytes on TLC plate, TLC saturated vertical
chambers for chromatograms developing (Camag), 254/366 nm UV chamber for
chromatograms visualization (Camag).
2.2
Preparation of Inhibited Propellants Samples
Lacquer and paint was removed from the outside surface of inhibitor and its layer
contacted with the propellant grain was removed, as well.
From propellant grain were taken samples situated at the distance of 3 mm and 10 mm
from its contact surface with inhibitor. Separated inhibitor and propellant samples of mass
0,1 g each, after natural or accelerated ageing were dissolved in 10 ml of acetone.
2.3
Separation of Substances in Propellants and in Inhibitor Samples
Type of absorbent, elution strength, polarity and composition of developing phases were
changed, tending to isolate in propellant NC matrix and to separate as many as possible
analyzed substances at the shortest distance of development.
After dissolving of propellant and inhibitor in acetone, their solutions of 50l
(for inhibitor) and of 30l (for propellant) were sprayed on the start line of HPTLC plate.
Variable parameters of separations were: distance of elution and composition of mobile
phase. Chromatograms were developed in the saturated vertical chambers. In this step of
chromatographic analysis i.e during optimization of separation, chromatograms were
visualized by method of fluorescence quenching.
249
The best separation of NG, 4-NO2-EC, EC, 2,4-DNT was obtained under the following
conditions:
mobile phase: hexane/benzene/ethyl acetate (8:3:2),
distance of development: 80mm.
Retention factors values enlarged 100 times i.e. hRf for separated components were as
follows:
4-NO2-EC 26
EC 46
NG 54
2,4-DNT 64
For separation of DBP and TNT it was used two-step chromatographic development at
the same direction. In the first step of development it was used mobile phase
hexane/benzene/ethyl acetate (8:3:2) at the distance of 80mm and in the second step at the
distance of 40mm it was applied development phase: hexane/acetone (5:1) after pervious
cut off of the rest analyzed substances from first step of chromatographic development. In
this case TNT had hRf=33 and DBP had hRf=50.
Chromatogram of above separated substances are shown in Figure 1. Analyzed
substances are visible in this Figure as distinctly separated densitographic peaks.
(A)
d(mm)
Fig 1.
Chromatograms of separation of EC, 2,4-DNT and DBP with TNT;

Densitometric peaks: 1 EC, 2 2,4-DNT, 3 TNT+DBP. Scanning UV
wave length - 255 nm, d - distance of chromatogram development,
A - absorbance of measured signal, mobile phase: hexane/benzene/ethyl
acetate (8:3:2)
250
2.4
Quantitation of Substances in Propellant and in Inhibitor Samples
Quantitative determinations of analyzed substances spots (bands) in situ are possible

in TLC in the following cases:
analysis of fluorescent substances (measurement of fluorescence intensity);
combination of area measurement of chromatographic spot (band) and intensity
of its colour characteristic one or caused by chemical reaction;
combined measurement of chromatographic spot (band) area and its light
absorption (transmission) at characteristic wave length for analyzed substance.
Because analyzed substances are not fluorescent, they do not have intensive colours and
in concentrations used for analysis they are practically colourless, we decided to measure
their absorption during quantitative analysis. Such measurements are applied for quantitation
when analyzed substances have got characteristic absorption spectra. All analyzed
substances besides NG have such relatively strong characteristic spectra in UV region. So, it
was necessary to find a wave (max) in the UV or VIS region, at which maximum absorption
for NG occurs.
In the case of NG we used N,N,N,N-tetramethyl -1,4-phenylenediamine for strength its
absorption signal (NG visualization). So, NG spot (band) was sprinkled by 1% methanol
solution of N,N,N,N-tetramethyl -1,4-phenylenediamine, then it was dried for 1 minute
with use of cold air stream and finally it was subjected to UV radiation (254nm) for five
minutes. To obtain max for each of analyzed substances they were spotted on TLC plates to
be as close as possible to conditions for final quantitative measurements. Next the spots were
scanned by UV or VIS light beam and signals of reflected light were measured. Absorption
spectra were recorded by the densitometer. Characteristic wave lengths of analyzed
substances determined by the maxima of absorption peaks were as follow: EC 255nm, 4NO2-EC 340nm, 2,4-DNT 265nm, TNT 250nm, DBP 240nm, NG 600nm.
As example in Figure 2 and 3 the spectra of TNT and EC are presented respectively. These
max values were used for quantitation.
A
Fig 2.
UV absorption spectrum of TNT,
[nm]
A - absorbance of measured signal, - length of UV wave
251
[nm]
Fig 3.
UV absorption spectrum of EC
A - absorbance of measured signal, - length of UV wave
In order to obtain calibration curves, after separation of standards of analyzed

substances of various masses according to the above chromatographic conditions, spots
(bands) of each substance were scanned by a zig-zag method using light wave length of
max in the direction of movement of head of mobile phase. Measured signals were
represented by areas of densitometric absorption peaks.
In every case linear relation A=f(c) was obtained, where: A densitometric peak
area, c - amount of the analysed substance in the spot (band). Parameters of quantitation for
analyzed substances are gathered in the Table 1.
Table 1. Parameters of quantitation of analyzed standard substances
Analyzed
substance
max
Calibration curves
A=f(c)
EC
255
A=26.625c+9820.5
0.9973
Range of
linearity
[ng]
300 3000
4-NO2-EC
340
A=35.552c+12280
0.9999
200 3000
100
NG
600
A=18.323c+10895
0.9997
400 4400
200
DBP
240
A=19.979c+4474.1
0.9980
200 3000
100
TNT
250
A=44.643c+9091.6
0.9998
500 2600
100
2,4-DNT
265
A=31.44c+25109
0.9689
550 3300
100
[nm]
Correlation
coefficient
Detection
limit
[ng]
100
252
2.5
Investigation Results and Their Discussion in Aspect of Migration

Processes Going on in Inhibitor and Propellant Samples
Results on compositions changes of inhibitor and propellant samples after natural ageing
for ca. 17 years under ambient conditions and after accelerating ageing for 1h at 120C are
gathered in Table 2.
Table 2. Compositions changes of inhibitor and propellant aged under natural

and artificial conditions
Component
Inhibitor
Propellant
aged under ambient
aged at 120C for 1h
conditions for 17 years
Content of component
Content of
[%m/m]
component
[%m/m]
Sample taken at the
Sample taken at the
distance of 3mm from
distance of 3mm
inhibitor
from inhibitor
aged under ambient

conditions for 17 years
Content of component
[%m/m]
aged at 120C for

1h
EC
0.96
0.80
2.34
2.07
4-NO2-EC
0.00
trace amounts
trace amounts
trace amounts
NG
6.62
9.57
21.73
19.13
DBP
3.46
4.21
4.81
5.84
TNT
0.64
0.93
1.16
1.40
2,4-DNT
2.98
3.85
7.29
6.34
Content of
component
[%m/m]
These results show that after natural storage in inhibitor material appeared EC (ca. 1%),
NG (ca. 6,6%), 2,4-DNT (ca. 3%) and TNT (ca. 0,6%). So into inhibitor migrated 11% of
new chemical compounds from propellant.
Simultaneously in propellant esp. in its outer layer contacting with inhibitor (at the
distance from 0 to 3mm), contents of EC, NG, 2,4-DNT decreased disappeared respectively.
Additionally in propellant DBP content increased. It means that DBP migrated from
inhibitor, but DBP does not reach depth in propellant less than 10mm.
As a result of artificial ageing content of EC in propellant decreased and appeared trace
amount of 4-NO2-EC which points that nitration reaction occurred during this ageing. At the
same time migration processes of NG, 2,4-DNT, and TNT, from propellant to inhibitor and
DBP migration from inhibitor to propellant were continued with relatively high rate what is
characteristic at heightened temperature. At the distance of 10mm in propellant from contact
layer with inhibitor changes of propellant compositions are practically not distinct. It seems
also possible nitration of 2,4-DNT into 2,4,6-TNT as accompanying reaction during above
described migration processes.
253
3.
CONCLUSIONS
Naturally and artificially aged inhibitor and propellant used in explosive long line
charges for clearance of mine fields show considerable changes in their composition esp.
inhibitor what can lead to undesirable physico-chemical processes. These processes seem to
be main cause of failure operation of above mentioned rocket throwing systems.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
W.F MABER: Symp. Chem. Probl. Stabil. Explos. (Proc.), Sweden, 1971, 2nd , 89;
A. SCHWARTZ: Symp. Chem. Probl. Stabil. Explos. (Proc.), Sweden, 1976, 4th, 65;
H.J. REITSMA: Symp. Chem. Probl. Stabil. Explos. (Proc.), Sweden, 1980, 5th, 287;
B. BOTMA, H. SCHIMANSKY: Symp. Chem. Probl. Stabil. Explos. (Proc.), Sweden, 1989, 8th,
245;
N.VAN DER MEER, W.DE KLERK, R. EERLIGH: Symp. Chem. Probl. Stabil. Explos. (Proc.),
Sweden, 10th, 139;
B. VOGELSANGER, B. OSSOLA, E. BRNNIMANN: Symp. Chem. Probl. Stabil. Explos. (Proc.),
Sweden, 10th, 305;
J. WILKEN: 24th Annual Conference of ITC 1993, Karlsruhe, Germany, 44-1;
M. RAT, G. LACROIX: 25th Annual Conference of ITC 1994, Karlsruhe, Germany, 32-1;
B. OSSOLA, B. VOGELSANGER, E. BRNNIMANN: 25th Annual Conference of ITC 1994,
Karlsruhe, Germany, 38-1;
W.DE KLERK, N.VAN DER MEER, R. EERLIGH: 25th Annual Conference of ITC 1994, Karlsruhe,
Germany, 44-1;
M. MISZCZAK, J. BDEK: 25th Annual Conference of ITC 1994, Karlsruhe, Germany, 57-1;
L. GOTTLIEB: 25th Annual Conference of ITC 1994, Karlsruhe, Germany, 90-1;
J.P. AGRAWAL, H. SINGH: Propellants, Explosives, Pyrotechnics 18, 106-110 (1993);
B. VOGELSANGER, B. OSSOLA, E. BRNNIMANN: Propellants, Explosives, Pyrotechnics 21,
330-336 (1996);
R A. PESCE-RODRIGUEZ, R.A. FIFER: Propellants, Explosives, Pyrotechnics 21, 337-343
(1996);
S.Y. HO: Propellants, Explosives, Pyrotechnics 20, 206-214 (1995);
J.P. AGRAWAL, N.T. AGAWANE, R.P. DIWAKAR: Propellants, Explosives, Pyrotechnics 24, 371378 (1999).
254
BIODEGRADATION OF TNT BY IN VITRO CULTIVATED PLANTS:

A MODEL SYSTEM FOR STUDY OF DEGRADATION PROCESSES
Ale Nepovm*, Anja Hebner**, Andre Gerth**, Hartmut Thomas***,
and Tom Vank*
* Institute of Organic Chemistry and Biochemistry,
Flemingovo nm. 2, 166-10 Praha 6, CZ
** Bioplanta GmbH,
Benndorfer Landstrae 2, D-04509 Delitzsch
*** Wasag Decon GmbH,
Werkstrae 111, 45721 Haltern, GE
Abstract:
Removing of explosives and nitroaromatic compounds from contaminated soil and
ground water belongs to one of the most important environmental tasks.
Phytoremediation appears to be a promising technique for treatment of water or soil
contaminated by low concentration of these compounds.
We have used in vitro cultivated water plants - halophytes for removing of TNT from
cultivation medium. The efficiency of four selected plant species of reed (Phragmatis
australlis), rush (Juncus glucus), reed mace (Typha latifolia) and sedge (Carex
gracillis) were compared. The highest efficiency in removing of TNT in initial
concentration of 100 mg/l was reached by reed and reed mace. Reed showed as well as
high tolerance to TNT, however, more tolerant of TNT was sedge, which showed
induction of growth in comparison to untreated plants.
Keyword:
1.
TNT, halophytes, water plants, phytoremediation. explosives
INTRODUCTION
The analysis of soil contaminated by TNT shows a presence of 2,4,6-TNT, 2,4-DNT and
several metabolites involving oxidized forms as 2,4,6-trinitrobenzoic acid (TNBA) and 2amino-4,6-dinitrobenzoic acid (2-ADNBA), and reduced forms 2- and 4- aminodinitrotoluenes.[1] The fate of nitroaromatic compounds and their aminodederivates in soil is
associated with formation of conjugates with humic acid. They are preferentially bound to
quinoid and phenolic moieties of soil humic substances.[2] The humic monomer like catechol
is initially transformed into semi-quinone radical by a phenoloxidase (laccase, Trametes
villosa or horseradish peroxidase) and Wang et al. further postulated a subsequent oxidative
coupling involving reactions with additional humic monomers or anilinic products derived
from TNT, forming an anilinoquinone via nucleophilic addition or a benzoquinone-imine
through condensation.[2]
Removing of explosives from contaminated soil and ground water belongs to one of
most important environmental tasks. Possible ways of removing of explosives like TNT
from heavily contaminated sites are traditional excavation and subsequently incineration
255
techniques but the treatment is not applicable in case of low contaminant concentration due
to low economical efficiency. In spite of the real concentration of contaminant is low, the
concentration is high enough being toxic and dangerous. Therefore an alternative ways of
removing of contaminant is necessary develop and study. A combination of bioremediation
techniques offers such solution. The remediation made by plants and associated
microorganisms is called phytoremediation.
In our study we compared plant species of water plants (Halophytes) cultivated under in
vitro conditions on the concentration of TNT in the medium. In vitro experimental set up
allows characterize metabolism only in plants i.e. with exclusion of the effect of
microorganisms.
2.
MATERIALS AND METHODS

2.1
Plant material
Plant species of reedmace (Typha latifolia), reed (Phragmites australis), rush (Juncus
glaucus) and sedge (Carex gracillis) were cultivated on liquid medium prepared according to
Murashige and Skoog [3] and supplemented with phytohormone 6-benzylaminopurine (BAP,
5 mg/l). The formation of roots was initiated on a hormone-free medium. The plants were
subcultured every 4-5 weeks on a fresh medium.
2.2
HPLC analysis
The content of TNT and its degradation products was analyzed by High Performance
Liquid Chromatography (HPLC). The sample of 20 l was separated on stainless steel
column (250x4 mm) packed by reverse phase Si-C18 (size:7 m; Biosphere, Labio Praha,
CZ) using a linear gradient of mobile phases (10100% methanol in 40 min). The
concentration of known compounds was calculated from particular peak areas integrated
from chromatogram at 230 nm.[4]
3.
The growth characteristics of tested plants in dependence on different concentration of

TNT are summarized at the Table 1. The growth of untreated plants was quite balanced with
the growth value about 0,86 except Typha, which grew slower. Whereas Carex showed the
best resistance among the others because was able growth and propagate in medium with
high concentration of TNT plant species Phragmites and Typha showed higher uptake of
TNT from the medium than Juncus and Carex. The values of residual TNT concentration are
shown at the Figure 1 end of the experiment. The concentrations of TNT are expressed in
Figure 1 at 10th day of cultivation. Carex showed the best resistance to presence of TNT in
the medium. The growth values increased with growing concentration of TNT in medium,
therefore inductive effect of TNT was observed. Only TNT the concentration of 5 mg/l
caused small induction effect on growth of Phragmites did not affect the growth. Presented
data illustrated that halophytes, esp. Carex and Phragmites can be successfully used for
treatment of water or soil contaminated by TNT.
256
Table 1. Comparison of growth values (GV) for particular species in dependence

of different concentrations of TNT.
The growth value is expressed by equation GV=w0/(wt-w0), where w0 [g]
is weight in time t=0, whereas wt is at time t.
TNT
concentration
[mg/l]
Carex gracillis
Juncus glaucus
GV
std. dev.
GV
std. dev.
0,85
0,11
0,85
0,11
1,02
0,12
0,95
20
1,19
0,13
50
1,28
100
1,14
Phragmites
australis
Typha latifolia
std. dev.
GV
std. dev.
0,87
0,11
0,60
0,03
0,24
1,00
0,15
0,29
0,06
1,03
0,17
0,76
0,12
0,42
0,18
0,35
0,86
0,08
0,89
0,13
0,12
0,14
0,15
0,64
0,05
0,81
0,11
0,11
0,04
GV
120
Carex
Juncus
Phragmites
100
TNT [mg/]
80
Typha
60
40
residual TNT
in the medium
16,24
20
0
0
Fig 1.
4.
6
cultivation [day]
10
8,93
4,96
2,30
12
Decrease of TNT concentration in the MS medium supplemented

with 5 mg/l BAP during cultivation of halophytes Carex gracillis,
Juncus glaucus, Phragmites australis and Typha latifolia
under in vitro conditions.
CONCLUSION
Plant species of rush (Juncus glaucus), reed (Phragmites australis), reed mace (Typha
latifolia) and sedge (Carex gracillis) have been found as suitable adepts for
phytoremediation of TNT. Further experiments applied in situ in larger scale will verify the
efficiency of these plant species and will be compared with results obtained from in vitro
conditions.
257
Acknowledgment:
Thank to Prof. Ing. Svatopluk Zeman, DrSc. and Ing. Zdenk Jalov, Ph.D. for synthesis
and preparation of degradation products of TNT.
The work was supported by GAR project no. 206/02/P065, COST 837.1, MMT project
Kontrakt no: CZ 01/024 and Research project no: Z4 055 905.
REFERENCES
[1] D.BRUNS-NAGEL, T.C.SCHMIDT, O.DRZYZGA, E. VAN LW and K. STEIBACH: Identification of
oxidized TNT metabolites in soil samples of a former ammunition plant. Environ. Sci. Pollut.
Res., 6, 7-10, 1999
[2] C.-J. WANG, S. THIELE and J. M. BOLLAG: Interaction of 2,4,6 trinitrotoluene (TNT) and 4amino-2,6-dinitrotoluene with humic monomers in the presence of oxidative enzymes. Arch.
Environ. Contam. Toxicol. 42, 1-8, 2002
[3] T. MURASHIGE and F. SKOOG: A revised medium for rapid growth and bio assays with tobacco
tissue cultures. Physiol. Plant., 15, 473-497, 1962
[4] T. VANK, A. NEPOVM, R. PODLIPN, S. ZEMAN and M. VGNER (2003) Phytoremediation of
selected explosives. Water, Air & Soil Pollution, in press, 2003
258
DETERMINATION OF SENSITIVITY
OF PLASTIC EXPLOSIVE CONTAINING NTO
Andrzej Orzechowski*, Andrzej Maranda**,
Dorota Powaa* and Jacek Borkowski***
2 Kaliskiego St, 00-908 Warszawa 49, PL
*** Military Institute of Armament Technology,
7 Wyszyskiego St, 05-220 Zielonka, PL
Abstract:
We tried to obtain insensitive high explosive. We investigated plastic explosive based on
hexogen, octogen.. As an additive we used NTO. We researched sensitivity to impact
and friction. This additive generated the decrease of sensitivity without worsening
another parameters.
Keyword:
1.
insensitive high explosive, plastic explosive
INTRODUCTION
Military technology development forces the necessity of new explosive elaboration.

The research workers evolve low sensitivity explosives for military and civil application.
We investigated low sensitivity plastic explosives because of a profitable technological
form and low toxicity of these explosives. A major plastic explosive component, which
decides about explosive properties, is an explosive itself. Hexogen (RDX), pentrit (PETN)
and octogen (HMX) are used very often. These are characterized by high explosive energy
and low sensitivity to impact, friction etc. Sensitivity bullet impact test is one of the most
important parameters. This test standard is described in ONZ transport regulations [1]. Its
a shot testing of samples by bullet caliber 12,7 mm of defined energy. Research procedure
and comparative shot bullet sensitivity tests are presented in [2]. Comparing research result of
samples put in paper tubes we could notice some correlation between sensitivity
to mechanical stress (to impact, to friction) and sensitivity to bullet shot. Researches show
enlarging resistance on shot bullet with desensitized explosive in comparison with crystalline
one. But shot sensitivity depends on applied measurement methods. Results depend on the
kind of tube material and relative bullet energy on area unit (caliber, speed, mass and shape
of bullet). The result of shot researches show that samples of traditional high explosives just
like TNT, HX-desensitized put in a metal tube are too sensitive and detonate.
We made a decision to investigate the influence of NTO additive, in order to attain
the increase of sensitivity to mechanical stress without worsening their detonation
parameters.
NTO is crystalline substance with high density and velocity detonation, melting point
(with decomposition) at 260 Co [3,4]. NTO is suitable and thermal resistant. Simultaneously
NTO are low sensitive to mechanical stress [5].
259
Sensitivity to impact:
Sensitivity to friction:
0,5 kgm (51 Nm) NTO,

0,75 kgm (7,35 Nm) RDX,
1,5 kgm (14,7 Nm) TNT.
>36 kg ( >353 N) NTO,

12 kg (118 N) RDX,
>36 kg ( >353 N) TNT.
Simultaneously critical diameter is high and depends on grain size characteristic and
compression ratio range from 13 to 25 mm [6-7].
2.
MATERIALS AND METHODS

2.1
Characteristics of explosives
We tested following explosives:

hexogen (RDX), shake bulk density 0,96 g/cm3, grain size analysis is presented in [Fig.1],
octogen (HMX), shake bulk density 1,04 g/cm3, grain size analysis is presented [Fig.2],
nitrotriazol (NTO), shake bulk density 0,95 g/cm3, grain size analysis is presented
in [Fig.3]. Crystals of NTO are shown in [Fig.4].
Plastic binder consists of butadiene-styrene copolymer, dioctyl adipate and oil.
We checked two following methods of preparing of plastic explosives:
mixing of before prepared binder with dry explosives,
mixing of dry or wet explosives with water suspension of copolymer, then drying and
plasticizing.
Due to safety and mixing quality the best method is a mixing of wet explosive with
butadiene-styrene copolymer water suspension.
Volume [%]
Hexogene was not crystallized from acetone but was received by long mixing with
water. Octogene was crystallized from acetone.
5
4,5
4
3,5
3
2,5
2
1,5
1
0,5
0
1,8
31
59
88
116
145
173
Particle diameter [m]

Fig 1.
Particle size distribution RDX
260
2
1,8
Volume [%]
1,6
1,4
1,2
1
0,8
0,6
0,4
0,2
0
1,8
31
59
88
116
145
176

Fig 2.
Particle size distribution HMX
8
7
Volume [%]
6
5
4
3
2
1
0
5,7
137
270
403
535
668
801

Fig 3.
Particle size distribution NTO
261
Fig 4.
2.2
Cristals of NTO
Methods
Bullet impact tests are expensive so we carried out researches on impact and friction
plastic explosive sensitivity, which we will use to preliminary verification of plastic
explosive components.
The impact sensitivity of explosives was studied using the drop weight apparatus.
The friction sensitivity of explosives was studied using the Peters apparatus.
3.
RESULTS
3.1
Research result for plastic explosive with RDX
Samples contain 84% and 90% explosive. NTO content was changed, when NTO
content increases explosive (RDX or HMX) decrease. Binder content was stable for all
series.
butadiene-styrene copolymer 30%,
dioctyladipate
60%,
oil
10%.
Samples density is presented in [Tab.1] and [Tab.3]. Result sensitivities to impact and
to friction is presented in [Tab.2] and [Tab.4] The lowest sensitivities was obtained
for samples content 84% explosives. However density was less then in samples content
90% explosives.
NTO additive reduces sample sensitivity to impact to 34,3 Nm for explosive content
84%, to 29,4 Nm for explosive content 90% (they contain 20-60% of NTO). A high decrease
of sample sensitivity to impact appears for 80% NTO additives.
262
Table 1. Density of plastic explosive containing 84% explosive

Content 84%
Explosive
16% binder
Density [g/cm3]
20% NTO
80% RDX
40% NTO
60% RDX
60% NTO
40% RDX
80% NTO
20% RDX
1,42
1,48
1,50
1,53
Table 2. Test result of sensitivity to impact and friction for plastic explosive
containing 84% explosive
Content
84% Explosive
16% binder
0% NTO
100% RDX
20% NTO
80% RDX
40% NTO
60% RDX
60% NTO
40% RDX
80% NTO
20% RDX
Impact
Sensitivity*
[Nm]
Impact
Sensitivity**
[Nm]
Friction
Sensitivity *
[N]
Friction
Sensitivity **
[N]
29,4
24,5
>353
353
34,3
29,4
>353
353
34,3
29,4
>353
353
34,3
29,4
>353
353
>49,0
39,2
>353
353
*min. 1 decomposition at 6 samples

**no decomposition at 6 samples
Content
90% Explosive
10% binder
Density [g/cm3]
20% NTO
80% RDX
40% NTO
60% RDX
60% NTO
40% RDX
80% NTO
20% RDX
100%
NTO
1,51
1,56
1,59
1,59
1.56
Table 4. Test result of sensitivity to impact and friction plastic explosive containing
90% explosive
Content
90% Explosive
10% binder
0% NTO
100% RDX
20% NTO
80% RDX
40% NTO
60% RDX
60% NTO
40% RDX
80% NTO
20% RDX
Impact
Sensitivity *
[Nm]
Impact
Sensitivity **
[Nm]
Friction
Sensitivity *
[N]
Friction
Sensitivity **
[N]
24,5
19,6
353
317
29,4
24,5
>353
353
29,4
24,5
>353
353
29,4
24,5
>353
353
49,1
39,2
>353
353

263
3.2
Research result for plastic explosive with HMX
Samples contain 90% explosive. NTO content was changed. Binder content was stable
for all series.
Samples density is presented in [Tab.5] Result sensitivities to impact and to friction
is presented in [Tab. 6].
NTO additive reduces sample sensitivity to impact to 29,4 Nm for explosive content
90% (they contain 20-60% of NTO).
Content
90% Explosive
10% binder
Density [g/cm3]
0% NTO
100% HMX
20% NTO
80% HMX
40% NTO
60% HMX
60% NTO
40% HMX
80% NTO
20% HMX
1,56
1,64
1,62
1,58
1,57
Table 6. Test result of sensitivity to impact and friction plastic explosive containing
90% explosive
Content
90% Explosive
10% binder
0% NTO
100% HMX
20% NTO
80% HMX
40% NTO
60% HMX
60% NTO
40% HMX
80% NTO
20% HMX
Impact
Sensitivity *
[Nm]
Impact
Sensitivity **
[Nm]
Friction
Sensitivity *
[N]
Friction
Sensitivity **
[N]
19,6
14,7
353
317
29,4
24,5
>353
353
29,4
24,5
>353
353
29,4
24,5
>353
353
39,2
34,3
>353
353

4.
CONCLUSIONS
Crystalline form and grain size characteristic used NTO enable technological processing
and ensures for good plasticity.
NTO additives generate the decrease of sensitivity to impact explosive. The samples
containing 84 % of explosive have the least sensitivity to impact. For samples, which contain
20-60% of NTO, sensitivity is the same. The high decrease of sample sensitivity to impact
appears for 80% of NTO additive. The NTO additive of 20% can decrease plastic explosive
sensitivity to impact.
264
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
Recommendations on the transport of dangerous goods Test and Criteria, ST/SG/AC.

10/11/Rev. 7, Genewa 1990.
E. MILEWSKI, J. SZYMANOWSKI, M MISZCZAK, B. BACHNIO, Wraliwo
materiaw wybuchowych kruszcych na uderzenie (przestrzelenie) pociskiem. Weryfikacja
metod badawczych w wietle procedur NATO, Problemy techniki uzbrojenia i radiolokacji (68)
1999.
MANCHOT UND NOLL, Derivate des triazols Justus Liebigs Annalen der Chemie,
1906, 343.
A. MARANDA, S. CUDZIO, J. NOWACZEWSKI, A. PAPLISKI, Podstawy chemii
materiaw wybuchowych, WAT Warszawa, 1997.
F.G. FOUCHE, G. C. VON SCHALKWYK, TNT based insensitive munitions, Proc. of 27th
International Annual Conference of ICT, Karsruhe, Germany 1996.
L.B. CHAPMAN, NTO develpoment at Los Alamos, Proc. of 9th Symposium (International)
on Detonation, Portland, Oregon, USA, pp. 1001-1013, 1989.
P. GIMENEZ, P. CHABIN, J. MALA, C. SPYCKERELLE, An extensive experimental study of
pressed NTO, Proc. Of 10th International Symposium on Detonation, Boston, Massachusetts,
pp 273-285, 1993.
265
A COMPARATIVE ANALYSIS OF BLAST WAVES PARAMETERS

GENERATED BY EXPLOSION
OF EMULSION EXPLOSIVES AND DYNAMITES
Jzef Paszula*, Andrzej Maranda*, Andrzej Papliski*,
Barbara Gobek* and Johann Kasperski **
* Military University of Technology,
00-908 Warszawa, ul. Kaliskiego 2, Poland
** Blastexpol,
59-145 Duninw, Poland
Abstract:
In the paper results of experimental investigations as well as numerical analyses of blast
waves parameters generated by explosion of water-in-oil emulsion explosives and
dynamites is presented. In the experiments the overpressure magnitude in transient
shock wave generated by explosion of investigated explosive charges was measured.
Also numerical evaluation of blast field in the space surrounding the explosive charge is
performed. Comparison of the blast wave intensity generated by various kinds of
explosives is performed. It is concluded, that the blast wave overpressure can be
considered as an optional characteristic of the explosive performance, likewise
detonation velocity, detonation (explosion) energy.
Keyword:
1.
emulsion explosives, blast waves, explosive performance.
INTRODUCTION
Many explosives used in blasting industry are accomplished as multi-ingredient mixtures

in which the detonation process occurs in a nonideal regime. Differences in physical and
chemical characteristics of the components of the mixture results in differences in their
reaction susceptibility and overall kinetic mechanisms. In such explosive compositions the
chemical reactions in Chapman Jouguet point, as usually, are still not completed. Same
additional reactions continue to proceed in the first stage of expansion of detonation
products. As the result, additional energy is liberated after the detonation zone. Therefore, in
such explosives the value of detonation velocity which is immediately connected to the
detonation energy i.e. energy evolved before the C-J point, cannot be considered as the
adequate measure of the energetic performance of the explosives [1].
Also other classic method like lead block test or ballistic mortar test used for estimation
of work capacity do not assure the correct illustration of real energetic potential of nonideal
explosives, mainly due to limited mass of samples used in this test [2].
In the paper, which is continuation of earlier works [3,4], the method of overpressure
amplitude estimation in the blast wave as the method of appraising of the work performance
of an investigated explosive charge is proposed. The experiments are performed according to
the methodic presented in AASTP-4 Report [5].
266
2.
EXPERIMENTAL
2.1
Explosives tested
Emulsion explosive ALAN taken at various mass densities as well as two dynamites
were used in experiments. Base detonative parameters of the considered explosives are set
up in Table 1.
Energetic parameters of investigated explosives were evaluated along with the Polish
Norm BN-80-6091-42.
Table 1. Composition and energetic parameters of tested explosives.
Explosive
Explosive Parameters
Dynamites
ALAN 3
10G5H
20G5H
Nitrates
84,48
62,0
68,8
Nitroesters
25,0
22,0
Aluminium
5,00
Water
3,27
Others
7,25
13,0
9,2
Mass density, [kg/m3]
1100
1620
1550
Work capacity, [cm3]
310
360
Heat of explosion, [kJ/kg]
4307,29
Explosion temperature, [K]
3006
Explosion pressure, [MPa]
1021,13
Ideal work of explosion, [kJ/kg]
3581,13
Composition [cg/g]
2.2
Emulsion
explosive
Blast overpressure measurements and detonation velocity

estimation
Time-overpressure distribution in the blast wave generated by investigated explosive

charges was measured by pressure gauges of PCB Piezotronics, Inc, Serie 137A21. The
gauges register overpressure in the transient shock wave, sliding on the working surface of
the device. It works as an electromechanical transducer that converts the pressure (ratio of
the force acting on the working surface of the active element to magnitude of this surface)
into the proportional voltage signal. In the Table 2 technical characteristics of the gauge
used by overpressure measurements are presented.
Pressure gauges were mounted in sockets fitted to the walls of the testing shelter. The
gauges as well as the explosive charge were so disposed to avoid of influence of reflected
waves upon the registered pressure records.
267
Table 2. Technical parameters of used PCB pressure gauge.

Characteristic
Units
PCB 137A21
Measurement range
[MPa]
34,50
Pressure resolution
[kPa]
0,69
Sensitivity
[mV/kPa]
0,145
Resonance frequency
kHz
500
Temperature sensitivity coefficient.
[%/C]
0,054
Active element
Diaphragm
quartz
invar
The sketch of the experimental set is presented in Fig. 1. Emulsion explosive charges
were initiated by ERG electric igniters while dynamite charges were boosted by ERG
igniters and aluminium hexogene detonators.
Fig 1.
Experimental set up: 1 explosive charge; 2 pressure gauge; 3 PCB

signal conditioner; 4 digital oscilloscope; r distance from explosive
charge to pressure gauge
Detonation velocity was measured by use of short-circuit gauges. Explosive samples of

outer diameter of 36 mm were investigated. The results of detonation velocity measurements
are comprised in Table 3.
3.
RESULTS OF EXPERIMENTS AND DISCUSSION
In the Fig. 2 an exemplary registration of overpressure course for ALAN explosive is

presented. The 450 g charge of mass density of 1115 kg/m3 was tested. The registration is
made at the distance of 1 m. Upon the registered overpressure history the maximal pressure
value is estimated (0,87 0,05 Mpa in the considered case, see Table 3). In the Table 3 the
results of overpressure measurements and measured detonation velocities for investigated
explosives are presented. As the reaction mechanisms in detonation wave depend upon the
268
physical structure of the mixture, various mass densities of emulsion explosive were
considered. Magnitude of mass density is connected with porosity degree. As the gas
comprised in micropores acts as a sensitizer, it influences upon the overall kinetics of the
reactions occurring in detonation zone.
0,9
0,8
0,7
Overpressure, MPa
0,6
0,5
0,4
0,3
0,2
0,1
0
-0,1
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
-0,1
Time, ms
Fig 2.
Overpressure vs time diagram for ALAN explosive

mexp = 0,45 kg; exp = 1,15 g/cm3
Table 3. Measured detonation velocity and overpressure values for tested

explosives
Explosive
ALAN 3
Charge
diameter
Mass
density
Overpressure
[mm]
[kg/m3]
[m/s]
[MPa]
1115
3840
0,87 0,05
1200
3780
0,56 0,05
1250
3520
0,45 0,05
36
10G5H
36
1620
6115
0,94 0,01
20G5H
36
1550
5550
0,67 0,05
As an important factor influencing upon the detonation parameters aluminium particles

reaction degree inside the detonation zone can be indicated. In first stage aluminum particles
absorb energy and act as an inert additive. The process of aluminum activation occurs on
behalf of energy liberated in other chemical reactions. Absorption of energy evolved in
parallel chemical reactions by aluminum alters the kinetic of this process. Therefore,
aluminium lowers detonation parameters, in particular the detonation velocity.
The main output of energy produced in reactions of aluminium combustion occurs after
the as the C-J point. The liberated energy influences the process of expansion of detonation
products. Due to relatively large energy efficiency of aluminium oxidation the energy
269
liberated in aluminium combustion can bring a considerable contribution to the work

capacity of expanding gases. Therefore, value of blast overpressure can by an important
indicator of energetic performance of explosive charge. The comparison of detonation
velocities and overpressure magnitude attained in blast wave of considered explosives is
depicted in Fig. 3 and 4.
7000
Prdko detonacji, m/s
6000
ALAN 3 1,15
ALAN 3 1,2
10G5H
20G5H
ALAN 3 1,25
5000
4000
3000
2000
1000
Fig 3.
Detonation velocity of investigated emulsion explosives and dynamites

Considered mass densities of the emulsion explosives marked in the
picture
1,2
ALAN 3 1,15
ALAN 3 1,2
10G5H
20G5H
ALAN 3 1,25
Nadcinienie, MPa
1,0
0,8
0,6
0,4
0,2
0,0
Fig 4.
Maximun overpressure values
As can be seen, despite that detonation velocity of dynamites surpass those attained by
emulsion explosives the working efficiency of emulsion charges, accordingly to value of
overpressure in blast wave, is comparable to that of dynamites. Therefore one can conclude
that the value of overpressure can be a useful characteristic of the working potential of
nonideal explosives.
270
4.
NUMERICAL ANALYSIS OF THE BLAST FIELD
A comparison of blast field parameters can be also performed in a theoretical way. Upon
the models developed, influence of the particular parameters characterizing the explosive can
be appraised. In the paper, numerical evaluation of the pressure wave generated by
expansion of detonation products is s performed. Air properties are described by the real-gas
equation of state. Therefore only the explosion energy and the detonation products properties
influence on evaluated parameters of the pressure wave.
In Fig. 7 the exemplary space distribution of pressure in a blast wave is depicted. The
diagram shows the pressure distribution at the time moment when front of shock wave reach
the radii of 1 m. The explosion energy (Qv = 3,581 MJ/kg) correspond to the ALAN
emulsion explosive.
p [Mpa]
0,60
pf =0.49 MPa
0,40
t =758,7 s
0,20
0,00
0
Fig 5.
200
400
600
800
1000
r [mm]
Numerically estimated pressure time diagram at the distance of 1m;

mexp = 0,45 kg; Qv = 3,581 MJ/kg
Blast wave propagation was evaluated by the program package elaborated for numerical
analysis of shock and detonation wave propagation in reactive gaseous suspensions [6].
Three-step numerical algorithm is employed. A second-order Godunov type scheme is
implemented, spatial interpolation is used as a higher-order extension to Godunovs method.
Along with the exact incorporation of real-gas effects, the algorithm enables a reliable remapping of the flow-field caused by expanding products of explosion.
The evaluated flow-field corresponds to the assumed explosion energy. In the case of
heterogeneous explosives energy supply can occur both in detonation zone and due to
reaction taking place after the C-J point. In numerical simulation the full energetic effect of
explosion is accounted. This can differ from the detonation energy that corresponds to the
detonation velocity. By comparison of the numerically evaluated parameters of the blast
271
wave to the experimentally measured values an appraisal of the effective explosion energy
can be obtained.
As is seen from the results obtained for ALAN type explosives a general agreement
between the value of ideal explosion energy and the numerically evaluated blast wave
intensity can be observed. However, further theoretical and experimental investigations are
necessary to correlate the energy liberated in explosion with the energy deposited to the
loading wave generated in material continuum surrounding the explosive charge. These
problems will be the subject of further investigations of authors.
5.
CONCLUSIONS
In the paper analysis of blast waves parameters of mining explosives is performed.

Emulsion explosives of water-in-oil type and dynamites were investigated. Values of
detonation velocity and maximal overpressure magnitude in blast wave generated by
explosion of investigated explosive are considered as characteristics of energetical
performance of explosive. Detonation velocities were estimated experimentally. Both
experimental measurements and numerical evaluation of blast wave field is performed.
Explosive characteristics of heterogeneous explosive mixtures, like emulsion explosives
are considered. An attention is paid to the influence of additional energy output, which can
proceed after the detonation zone. In this case, the detonation velocity cannot be the measure
of the global explosion energy as it is strictly coupled with the energy liberated before the
C-J point, through the detonative adiabate. The additional energy that is produced during
expansion of detonation products will enhance intensity of loading wave generated in
material surrounding the explosive charge. Therefore, by measuring the blast wave intensity,
a more integral characteristic of energy output occurring in explosion can be obtained.
In the paper, experimental measurements as well as numerical evaluation of blast field
produced by explosive charges is performed. As can be seen from obtained results, in the
case of aluminized explosives when the detonation velocity is lowered due to inertial effects
of aluminum particles in the first stage of the detonation process, the value of blast
overpressure can give a more realistic estimation of the energy performance.
The proposed method can be used as the optional characteristic of the explosive
performance. However, the more comprehensive methodic should be elaborated, to
introduce this approach as the criterion for estimation energetical efficiency of explosives.
272
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
ORNELLAS D. L.: Calorimetric determinations of the heat and products of detonation for
explosives: October 1961 to April 1982. Report UCRL 52821, Lawrence Livermore National
Laboratory.
PERSSON P-A., HOLMBERG R., Lee J.: Rock Blasting and explosives engineering, Ch. 4,
Boca Raton, Florida, 1994.
PASZULA J., MARANDA A., GOBEK B., KASPERSKI J. 2002, Badanie nadcinienia fal
podmuchowych w powietrzu generowanych wybuchem materiaw wybuchowych emulsyjnych
typu LWC, Prace Naukowe Gwnego Instytutu Grnictwa. Seria Konferencje Nr 43,
Bezpieczestwo Robt Strzaowych w Grnictwie, Katowice 2002, 157.
PASZULA J., MARANDA A., GOBEK B., KASPERSKI J. 2002: Badanie nadcinienia fal
podmuchowych w powietrzu generowanych wybuchem materiaw wybuchowych nieidealnych,
Mat. Konferencji Naukowej IPO2002, Warszawa 21-22.11.2002, 34.
Report AASTP-4 No: DEOS/MH-B/1, Department of Enviromental and Ordnance Systems,
1998.
A. PAPLISKI, Evaluation of Nonstationary Detonation Initiation in Combustible Gaseous
Suspensions, 16-th International Symposium on Combustion Processes, Kazimierz,
1 - 5.08.1999, pp. 63 66.
273
CLASSICAL MOLECULAR DYNAMICS SIMULATIONS

OF RDX DECOMPOSITION UNDER HIGH PRESSURE
Miroslav Pospil*, Pavla apkov*, Pavel Vvra** and Svatopluk Zeman**
* Department of Chemical Physics and Optics, Faculty of Mathematics and Physics,

Charles University Prague, Ke Karlovu 3, 12116 Prague 2, Czech Republic.
** Department of Theory and Technology of Explosives, University of Pardubice,
53210 Pardubice, Czech Republic.
Abstract:
Classical molecular dynamics simulations in NPT ensemble have been performed to
investigate the decomposition of cyclotrimethylene-trinitramine C3H6N6O6, more
commonly known as RDX, under high pressure. Dynamics simulations were carried out
in 3D periodic systems of RDX crystal at 300 K under the high hydrostatic pressure.
Dynamic trajectories were calculated for a series of hydrostatic pressure in the range
46 - 500 GPa. The detailed analysis of dynamic trajectories revealed the mechanism
and kinetics of decomposition under high pressure. Increase of the hydrostatic pressure
changes the course of dynamic trajectory and results in increase of the kinetic energy
and decrease of the time corresponding to the rupture of the first N-NO2 bond.
Keyword:
1.
energetic materials, molecular decomposition, high pressure, explosives,

molecular dynamics, RDX.
INTRODUCTION
The work presented here is a continuation of studies focused to unraveling of molecular

decomposition process in explosive materials [1- 10]. We use the classical molecular dynamics
in Cerius2/Materials Studio modeling environment [11] for the simulation of molecular
decomposition. This work is an extension of our previous study [10] of unimolecular
decomposition at high temperatures studied by molecular dynamics simulations. The aim of
the present work is to reveal the mechanism and kinetics of molecular decomposition in ideal
crystals of RDX under high pressure.
Classical molecular dynamics simulations of molecular decomposition have been
performed for cyclotrimethylene-trinitramine (C3H6N6O6), more commonly known as RDX.
Dynamics simulations in NPT ensemble were carried out in 3D periodic system of RDX
crystal at 300 K under the high pressure. Dynamic trajectories were calculated for a series of
hydrostatic pressure in the range 46 - 500 GPa. The analysis of dynamic trajectories revealed
the mechanism of decomposition and enabled us to characterize the decomposition kinetics
under high pressure.
274
2.
STRATEGY OF MODELING
Initial models of RDX crystals for dynamic simulations were built using crystal structure
data of Choi & Prince [12]. Orthorhombic, primitive lattice: a = 13.182 , b = 11.574 , c =
10.709 with 8 molecules in the unit cell, space group Pbca, density d = 1.80598 g/cm3.
Structure of RDX molecule and arrangement of molecules in the unit cell is in the figures 1
and 2.
Classical molecular dynamics in NPT ensemble was carried out at 300 K with the time
step 0.0001 ps for the series of hydrostatic pressure values: 46 GPa, 100 GPa, 150 GPa, 210
GPa, 220 GPa, 225 GPa, 230 GPa, 350 GPa, 400 GPa, 500 GPa. For dynamics simulations
the crystal structure was treated as space group P1 that means no symmetry constraints were
set up. The lower limit for the hydrostatic pressure 220 GPa was determined as the lowest
value at which the decomposition starts within the time interval 0-100 ps. Berendsen method
of temperature-bath coupling was used with the relaxation time 0.1 ps [13] and the
hydrostatic pressure was controlled using Andersen method [14]. Force field cvff_950 [15] was
used to parameterize the models. This force field uses the Morse potential to describe the
bond stretching and consequently allows us to describe the molecular decomposition.
Series of dynamics trajectories was calculated starting at 500 GPa and then the pressure
was gradually decreased to estimate the lower limit of pressure, at which the decomposition
occur at 300 K before 100 ps. It has been found that for pressure lower than 220 GPa no
decomposition was observed before 150 ps. Dynamic trajectories were analyzed using the
time dependence of kinetic energy and animation of the molecular motion in the unit cell.
This enables to estimate the time and kinetic energy corresponding to the break of the first
C-N and N-NO2 bond in the unit cell.
3.
Figure 3 shows the time dependence of kinetic energy related to one unit cell for
dynamic trajectories calculated for 220, 225, 230, 350, 400 and 500 GPa. As one can see all
the dynamic trajectories exhibit similar character, which can be described by three stages: (1)
first stage with starting increase of kinetic energy and consequently increase of temperature
of the system, (2) the middle part with nearly constant (or fluctuating) kinetic energy and (3)
the third stage with the dramatic increase of the kinetic energy. The first stage is terminated
with the break of the first C-N bond in the unit cell. The second stage starting with the break
up of the first C-N bond is terminated with the rupture of the first N-NO2 bond and release of
the first NO2 group in the unit cell. Both these events (break up of the C-N and N-NO2
bonds) are marked in the graphs corresponding to the individual trajectories for 220, 225,
230, 350, 400 and 500 GPa in the figures 4 - 9. It is also evident from the figure 3, that
increase of the hydrostatic pressure results in the increase of kinetic energy of the system and
shortening of the time necessary for the rupture of the first N-NO2 bond.
Figures 4-9 show in details the course of dynamic trajectory for the above mentioned
values of the hydrostatic pressure (220, 225, 230, 350, 400 and 500 GPa). The lowest value
of the hydrostatic pressure for which the decomposition was observed was estimated as
220GPa. The corresponding dynamic trajectory i.e. the time dependence of kinetic energy of
the system is in the figure 4. Arrows show the point of break up the first C-N bond and the
first N-NO2 bond in the unit cell of the crystal. The range between these two events is the
second stage with only very slight increase of the kinetic energy. During this stage the
system obtains the energy, which is equivalent to the applied hydrostatic pressure and
animation of the dynamic trajectory revealed the mechanism of the decomposition process.
275
After the rupture of the first C-N bond in the unit cell one can see gradual rupture of C-N
bond in other molecules in the unit cell and this process is accompanied by the slight
increase of the kinetic energy of the system. This stage is terminated at the moment when the
first N-NO2 bond in the unit cell is ruptured. The third stage of dynamic trajectory, which
starts at the rupture of the first N-NO2 bond is characterized with the dramatic increase of the
kinetic energy of the system accompanied with the total decomposition of RDX molecules
into fragments.
Dynamic trajectories were analyzed using following characteristic values describing the
decomposition process in dependence on hydrostatic pressure:
<Ek(2st)> - the average kinetic energy at the second stage of dynamic trajectory,
averaged over the time interval between the break up of the first C-N and N-NO2
bond in the unit cell. Energy value is related to one molecule.
t(N-NO2) corresponding to the rupture of the first N-NO2 bond in the unit cell
Ek(NO2) - instantaneous kinetic energy corresponding to the rupture of the first NNO2 bond in the unit cell related to one molecule.
Table 1. Hydrostatic pressure, average kinetic energy at the second stage of
dynamic trajectory <Ek(2st)>, t(N-NO2) - time of the rupture of the first
N-NO2 bond in the unit cell and Ek(NO2) - instantaneous value of the
kinetic energy corresponding to the rupture of the first N-NO2 bond. All
the values of energy are related to one molecule (not to unit cell).
Hydrostatic pressure
<Ek(2st)>
t(N-NO2)
Ek(NO2)
(GPa)
(kcal/molec)
(ps)
(kcal/molec)
220
400
2.5491
555
225
440
1.907
551
230
460
1.520
556
350
850
0.997
888
400
1000
0.947
1243
500
1380
0.540
1549
As one can see in the table 1 and figures 3-9, the time of N-NO2 bond rupture decrease
with increasing hydrostatic pressure, that means the shortening of the second stage of
dynamic trajectory with increasing hydrostatic pressure. On the other hand the kinetic energy
at all stages of dynamic trajectory increases with increasing hydrostatic pressure.
Let's compare present results of molecular dynamics at high pressure with our recent
study [10] of uni-molecular decomposition of RDX using molecular dynamics at high

temperature. The lower limit of temperature necessary for the initiation of the RDX unimolecular decomposition before 100 ps was found ~ 1500 K, the kinetic energy
corresponding to the N-NO2 bond rupture for this dynamic trajectory was ~ 290 kcal/molec
for the reaction R1 and ~ 400 kcal/molec for the reaction R2. (For more details see [10].) The
lower limit of the hydrostatic pressure necessary for the initiation of the RDX decomposition
in perfect crystal before 100 ps is ~ 220 GPa. The kinetic energy corresponding to the NNO2 bond rupture for the corresponding dynamic trajectory is 555 kcal/molec.
276
4.
CONCLUSION
Present investigation of the molecular decomposition under high hydrostatic pressure

was carried out in the perfect RDX crystals, without any structural defects. Recent work of
Kuklja & Kunz [16,17] showed the influence of the crystal defects like vacancies and
dislocations on the electronic structure of the RDX crystals and the effect of the edge
dislocation on the impact sensitivity to detonation. It should be noted that in such a case the
crystal compression near dislocations greatly increases the probability of electronic
excitations. That means more favorable conditions for molecular decomposition (especially
the rupture of N-NO2 bond).
Taking into account present results and results of [17] we can briefly summarize the effect
of the hydrostatic pressure on the molecular decomposition in RDX crystals in the following
items:
Hydrostatic pressure supplies the energy to the molecular crystal, which can initiate
the molecular decomposition.
The time and energy corresponding to the initiation of molecular decomposition

depends on the value of the hydrostatic pressure.
The lower limit of the hydrostatic pressure necessary for the initiation of the
decomposition ~ 220 GPa was estimated for the perfect crystal and this value in nonperfect crystal will be dependent on the character and concentration of defects.
Acknowledgements:
The authors express their gratitude to the Ministry of Industry and Commerce
of the Czech Republic for supporting the work within the framework of
research project No. FC-M2/05 and to Grant Agency of the Czech Republic for support of
an access to the crystallography database in Cambridge (within the framework of the project
No. 203/02/0436).
277
5.
FIGURES
Fig 1.
Molecule of RDX.
Fig 2.
Arrangement of 8 RDX molecules in orthorhombic unit cell.
278
Fig 3.
NPT dynamics trajectories for hydrostatic pressure (a) 220 GPa, (b) 225
GPa, (c) 230 GPa, (d) 350 GPa, (e) 400 GPa, (f) 500 GPa.
Fig 4.
NPT dynamics trajectory for the hydrostatic pressure 220 GPa.
279
Fig 5.
Fig 6.
280
Fig 7.
Fig 8.
281
Fig 9.
282
REFERENCES:
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
P. Politzer, J. S. Murray, J. M. Seminario, P. Lane, M. E. Grice and M. C. Concha:

Computational Characterization of Energetic Materials, J. Mol. Struct. (Theochem), 573 p. 110, 2001.
M. M. Kuklja: Thermal Decomposition of Solid Cyclotrimethylene Trinitramine, J. Phys.
Chem. B, 105, 10159-10162, 2001.
D. V. Shalashilin and D. L. Thompson: Monte Carlo Variational Transition-State Theory Study
of the Unimolecular Dissociations of RDX, J. Phys. Chem. A, 101, p. 961-966, 1997.
Y. Kohno, K. Ueda and A. Imamura: Molecular Dynamics Simulations of Initial
Decomposition Process on the Unique N-N Bond in Nitramines in the Crystalline State, J.
Phys. Chem., 100, 4701-4712, 1996.
P. Vvra: Electronic Density of Molecule and Some Properties of High Explosives, Proc. 4th
Seminar New Trends in Research of Energetic Materials Univ. of Pardubice, Pardubice,
Czech Republic, p. 345-351, April 2001.
S. Zeman, R. Huczala and Z. Friedl: The Study of Chemical Micromechanism Governing
Detonation Initiation of Some m-Dinitrobenzopolyazaarenes, J. Energet. Mater., 20, p. 53- ,
2002.
B. M. Rice, G. F. Adams, M. Page and D. L. Thompson: Classical Dynamics Simulations of
Unimolecular Decomposition of CH2NNO2: HONO Elimination vs N-N Bond Scission, J.
Phys. Chem., 99, p. 5016-5028, 1995.
C. J. Wu and L. E. Fried: First-Principles Study of High Explosive Decomposition Energetics,
Proc. 11th International Symposium on Detonation, Colorado, p. 490-497, 1998.
H. Dorsett: Computational Studies of FOX-7, a New Insensitive Explosive, DSTO Aeronautical
and Maritime Research Laboratory, Salisbury, South Australia, p. 1-19, 2000.
M. Pospil, P. apkov, P. Vvra and S. Zeman: Charackterisation of Explosive Materials
Using Molecular Dynamics Simulations, Proc. of the V. Seminar "New Trends in Reserch of
Energetic Materials", Univ. of Pardubice, R, p. 262-270, April 2002
Cerius2 documentation, Molecular Simulations Inc, San Diego, 2000.
C. S. Choi and E. Prince: The Crystal Structure of Cyclotrimethylene-Trinitramine, Acta Cryst.
B, 28, p. 2857-2862, 1972.
H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, A. DiNola, J. R. Haak: Molecular
Dynamics with Coupling to an External Bath, J. Chem. Phys., 81, p. 3684-3690, 1984.
H. C. Andersen: Molecular Dynamics Simulations at Constant Pressure and/or Temperature, J.
Chem. Phys., 72, p. 2384-2393, 1980.
A. T. Hagler, E. Huler and S. Lifson: Energy Functions for Peptides and Proteins. I. Derivation
of a Consistent Force Field Including the Hydrogen Bond from the Amide Crystals, J. Am.
Chem. Soc., 96, p. 5319-5327, 1974.
M. M. Kuklja and A. B. Kunz: Ab initio Simulation of Defects in Energetic Materials:
Hydrostatic Compresion of Cyclotrimethylene Trinitramine, J. Appl. Phys., 86, p. 4428-4434,
1999.
M. M. Kuklja and A. B. Kunz: Compresion-Induced Effect on the Electronic Structure of
Cyclotrimethylene Trinitramine Containing an Edge Dislocation, J. Appl. Phys., 87, p. 22152218, 2000.
283
THE CHEMICAL KINETIC AT DETONATION

OF NITROESTERS SOLUTIONS
Vlada M. Raikova, Georgy D. Kozak and Evgeny A. Likholatov
Mendeleev University of Chemical Technology,
Miusskaya sq. 9, Moscow A-47, Russia
Abstract:
The main objective of this work is measurement of failure diameters (df) of some
nitroesters (NE) solutions and estimation of overall kinetic parameters of reactions in
detonation wave. The dependencies df vs. content of components were measured for
mixtures on a base of nitroglycol (NGL) and nitroglycerin (NG) with
diethyleneglycoldinitrate (DGDN) propylnitrate (PN), methanol and acetone.
A minimum of df is fixed at a small quantity of additive. Failure and reinitiating of
a chemical reaction at detonation in the reaction zone according to the Failure diameter
theory by Dremin were chosen to be the base for calculation of temperature () and
pressure (p) in this zone. According to our calculations maximum of values was fixed
at the concentrations of the additives corresponding to the minimum of df. It was
proposed, that detonation wave decomposition of NE solutions includes only
monomolecular decomposition of a nitrocompound. The results of the calculations
revealed that values of E=140-160 kJ/mol and lg k0= 13,5-14,5 (s-1) at detonation of the
systems based on nitroesters is closed to the data obtained previously in the course of
investigation of slow decomposition of these substances.
Keyword:
1.
kinetic, detonation, shock wave, nitroester
INTRODUCTION
The failure diameter of detonation (df) is one of the important parameters for estimation
of an explosion hazard of every energetic material from the one hand, and from the second
one the experimental values of df can be a source for calculating kinetic parameters of
chemical reactions at detonation. Detonation failure diameter data of the nitroesters (NE) and
their mixtures have been obtained in earlier works [1-5].
The failure and reinitiating of the chemical reaction at detonation occur in the so-called
zone three according to the Failure diameter theory of Dremin [6, 7], which is the foundation
for calculation of the temperature (T3), pressure (p3) reaction time (3) in this zone. Earlier
this theory has been used to calculate the df values of liquid substances: nitromethane,
trinitrotoluene, and nitroesters [7-12]. Later the kinetic parameters of detonation reactions in
sulfuric acid solutions of NM and TNT have been defined [13]. The kinetics and mechanisms
of
chemical
reactions
in
detonation
wave
of
ethyleneglycoldinitrate,
diethyleneglycoldinitrate, and acetic anhydride solutions in nitric acid have been studied in
works [14, 15].
The main purpose of this paper was to use the methods developed in previous works to
calculate overall kinetics of chemical reactions in detonation wave of NE solutions. The
failure diameter versus content of addition (cd) dependency was the main source of kinetic
284
information in detonation wave. The dependencies df(cd) for the mixtures of

ethyleneglycoldinitrate (NGL) and nitroglycerine (NG) with diethyleneglycoldinitrate
(DGDN), propylnitrate (PN), methanol, acetone were measured in experimental part of our
work.
Shock and detonation parameters as well as characteristics of the reaction zone defining
the process of reinitiating of detonation wave in the zone of instability were calculated by
means of SGKR computer code (Shock and Detonation General Kinetic and
Thermodynamics in Reactive Systems computer package) developed in our Laboratory
previously [16].
It was proposed that decomposition of NE solutions in detonation wave included only
one monomolecular reaction. The effective values of kinetic parameters of detonation
reactions in every solution were defined and compared with low temperature decomposition
kinetic parameters of nitroesters.
2.
EXPERIMENTAL RESULTS
Empirical formula, oxygen balance, density, sound speed and formation enthalpy of
solution components are listed in the Table 1.
Table 1. Empirical formula, oxygen balance (), density (0), sound speed (U0), and
formation enthalpy (Hf0) of solution components
Substance
Ethyleneglycoldinitrate
Empirical
formula
, %
0,
U0,
Hf0,
g/cm3
m/s
kJ/mol
C2H4N2O6
1.489
1416
-242.8
C3H5N3O9
3.5
1.596
1480
-378.3
Diethyleneglycoldinitrate DGDN)
C4H8N2O7
-40.8
1.389
1570
-418.7
Propylnitrate (PN)
C3H7NO3
-99.04
1.054
1277
-174.1
Methanol
CH4O
-150
0.792
1123
-238.6
Acetone
C3H6O
-220.7
0.790
1192
-217.4
(NGL)
Nitroglycerine (NG)
Nitroesters were synthesized in the Laboratory. The identity of the NE was confirmed by
volumetric analysis of content of nitrogen in nitroester.
The detonation failure diameter of solutions was measured in glass tubes (the length was
120 150 mm, the wall thickness was 1-1,5 mm) by method go-no-go. If df of solution was
less 2 mm, the thin-walled conical capillaries were used. The tubes and capillaries had
smooth widening (the diameter was 15 mm) in the upper part for booster setting and for the
continuous transition to the steady-state detonation. Two pellets of phlegmatazed RDX (the
mass was 2 g, the diameter was 12 mm, the density was 1,67 g/cm3) were used as a booster.
The result of experiment was determined by means of metal witness-plate attached directly
to tube wall. Russian streak cameras SFR-2 and GFR-3 were used to register the detonation
velocity of solutions.
285
The detonation failure diameter of net NGL and NG are 1.9 and 2 mm correspondingly.
The df of high velocity detonation of DGDN is 16 mm. High velocity detonation of PN was
not founded, it did not propagate in glass tube. Low velocity detonation of PN in steel tube
(the diameter was 10 mm, the wall thickness was 13 mm) was registered.
The dependencies df(cd) of mixtures of NGL with methanol and PN are represented in
Fig 1. Addition of 20% methanol increased df of NGL from 1.9 mm to 8 mm. PN caused less
diluted effect, then methanol. The df of NGL mixture with 40% PN was 7 mm. However, it
was interesting to find that introduction of small quantity either of methanol or of PN
(cd=3-5%) into NGL slightly decreased df from 1.9 to 1.3-1.5 mm.
10
df, mm
8
6
4
2
cd, %
0
0
Fig 1.
10
20
30
40
The detonation failure diameter of mixtures of NGL with methanol (1)

and propylnitrate (2). Points are experiment data.
Curves are calculated dependencies of df (cd).
The detonation failure diameter of mixtures NGL-DGDN and NG-DGDN was varied
from 2 to 16 mm (Fig 2). For mixtures contained 10 % DGDN slaw reduction of df was
observed.
One can see that addition small quantity of inert substance or of nitroester with negative
oxygen balance to NGL or NG decreases its df.
3.
RESULTS OF CALCULATION
The detonation parameters: velocity (D), pressure (pcj) and explosion heat (Qv) of net
nitroesters and mixtures were computed by means of the Steady State Detonation (SD) Code
[16]
.
Shock parameters and characteristics of the reaction zone defining the process of
reinitiating of detonation wave in the zone of instability were calculated by means of SWDK
Code. Logical flow diagram of SWDK computer code is presented in Fig 3. SWDK code
consists of data base, input block, calculation block and output block. The data base contains
empirical formulas of the substances, numbers of chemical bonds in molecule, dependencies
of density (o), sound speed (Uo), specific heat (cp) of the substances versus initial
temperature. Hugoniot equation of liquid, as well as dependency of detonation velocity on
density, pressure pcj, df and velocity of reaction failure wave (v) for substances or mixtures
are also included in data base. The calculation block consists of two parts. In first part shock
parameters of the substance or mixture are computed and transmitted into second part. In
286
second part characteristics of reaction zone (p3, T3 and 3) near the limit of detonation
propagation and relationships between df and 3 are calculated. First part of code operates
independently from second one.
df, mm
16
12
4
c d, %
0
0
Fig 2.
20
40
60
80
100
The detonation failure diameter of mixtures of NGL (1) and NG (2) with
DGDN. Points are experiment data.
Curves are calculated dependencies df (cd).
Research previously reported [17] has shown that calculation D values were in good
agreement with the experimental data for NGL, NG and DGDN. Oxygen balance of NGL is
equaled zero. Addition of DGDN, or PN, or methanol into NGL decreases oxygen balance
and density of mixtures. Influence of diluent content on calculated detonation velocity of
NGL and NG is shown in Fig.4. The velocity is almost constant or slightly arises up to cd =
5-10% and then decreases. The effect of influence of diluent content on pressure pcj is
similar.
Calculated dependencies df(3) of the mixtures under investigation are described by the
general linear functions (df in mm, 3 in sec):
df = 42.24 3 0.20
(1)
The pressure p3 arises with increasing of the pressure pcj. The correlation between pcj
and p3 is found to be a common one for all systems under investigation:
p3 = 0.533 pcj + 1.178
(2)
The T3 value is defined by shock temperature dependency T(p) at correspondent

value of p3 for each mixture. The dependencies T(p) and T3(p3) for the mixtures of NGL
with methanol are shown in Fig. 5. The maximum of T3 is found to be at cd=4%.
287
Input block
T0, (T0), cp(T0),
df, D(0), pcj, v
U0(T0), U(u)
Calculation Block
Calculation of shock
parameters of the substance or
mixture
Calculation of characteristics of
reaction zone near the limit of
detonation propagation
Output block
T(p), cv(T)
p3, T3, 3,
df(3)
Fig 3.
Logical flow diagram of SWDK computer code.
D, m/s
7500
4
7000
1
2
cd, %
6500
0
Fig 4.
20
40
60
80
100
The calculated detonation velocity vs. cd for mixtures NGL-methanol (1),

NGL-PN (2), NGL-DGDN (3), and NG-DGDN (4).
288
950
4
10
900
15
p,
20
850
8
Fig 5.
10
11
12
Pressure dependencies of shock temperature (dash lines) and temperature

T3 (curve). Numbers correspond to concentration of methanol in solution.
Table 2. The failure detonation diameter (df), detonation velocity (D), and
characteristics of the reaction zone defining the process of reinitiating
of detonation wave (p3, T3, 3) of nitroesters and its mixtures
Substance or
mixture
cd , %
Experiment
Calculation
df, mm
D, m/s
p3, GPa
3,
3, mcs
NGL
1.9
7395
11.4
925
0.0538
NG
2.0
7647
13.2
940
0.0505
NGL - DGDN
10
1.73
7421
11.5
931
0.0461
NG - DGDN
15
1.56
7671
13.2
944
0.0393
NGL - PN
1.38
7404
11.4
947
0.0368
NGL-methanol
1.52
7390
11.2
933
0.0402
NGL-acetone
1.40
7287
10.9
938
0.0379
The experimental data of df and calculated values of D, p3, T3 and 3 for all systems
under investigation are collected in Table 2. The data of Table 2 show a common tendency:
decreasing of experimental df values coincides with increasing of calculated T3 values at
weak dilution (cd =415%) of NGL and NG. Dilution of NGL and NG with a large quantity
of inert substances or other NE with negative oxygen balance decreases detonation
parameters (D, pcj, Qv ), reaction zone temperature T3 and increases df of solutions
conformably.
289
4.
DISCUSSION
The classic works on the failure detonation diameter of liquid substances[6-8] had shown
that experimental results could be interpreted using homogeneous thermal explosion theory.
The most important parameters in thermal explosion theory are reaction time (3), shock
temperature (T3), and chemical kinetics of heat releasing reactions. The interrelation of these
parameters for monomolecular reaction should be described with the equation:
3 =
cv R T32
exp ( Ea RT3 )
Qv Ea k 0
(3)
where Ea is activation energy, k0 is pre-exponent factor, cv is specific heat of solution,

R is universal gas constant. Equation (3) could be transform to another expression:
ln Z = ln A +
Ea
RT3
where Z = 3Qv (T32 cv ),
(4)
A = R ( Ea k0 ) .
If mono-molecular reaction model is true, the straight line must be obtained in Arrhenius
coordinates (Z -1/T3) and the values of Ea and k0 could be defined.
Calculated values of Z and T3 of NGL mixtures with methanol, PN and DGDN are
plotted in Arrhenius coordinates (Fig.7). Points are fitted by a linear function. The slop of
these lines gives activation energy E0 = 140 160 kJ/mol. Calculated values E0 are in good
agreement with those obtained at sub-atmospheric pressure in usual kinetic experiments [18].
Approximately the same result (144 kJ/mol) was fulfilled earlier in the work [11] for some
nitroesters.
Calculated kinetic parameters of detonation reactions in NGL solutions are represented
in the Table.3. The activation energy of decomposition of the mixtures NGL with methanol
and PN is about 146 KJ/mol and corresponds to activation energy of thermal decomposition
of net NGL [18]. There is little difference between the pre-exponent factors for those
mixtures. The activation energy of decomposition of the mixtures NGL with DGDN is 161
KJ/mol, which is close to middle between Ea values of NGL (146 KJ/mol) and of DGDN
(171 KJ/mol)[18]. The pre-exponent factor is 2.6 1014 sec-1.
The dependencies df(cd) for mixtures on a base of NGL calculated by equations (2) and
(3) at values of Ea and k0 (Table 3) are compared with experimental data in Fig.1 and Fig.2.
The calculation results (curves) are in good agreement with the experimental data (points).
290
Table 3. Calculated kinetic parameters of detonation reactions in NGL solutions

Solution
Cd, %
a, KJ/mol
lgk0
NGL
146
13.67
NGL-methanol
0-20
146
13.69
NGL - PN
10-40
146
13.49
NGL - DGDN
0-100
161
14.41
-12.5
ln Z
2
-13.5
-14.5
-15.5
1.0
Fig 6.
1.1
103/T3, K-1 1.2
Treatment of experimental data df of NGL mixtures with PN (1), DGDN

(2), and methanol (3).
The effective values of rate constant of detonation reaction in the mixture NG with
DGDN was defined by additive function:
k = k11 + k22,
(5)
where k1 k2 are the rate constants of decomposition reaction of NG and DGDN corr.,
1 2 are their molar fractions in solution .
The values of activation energy of thermal decomposition of NG and DGDN[18] and preexponent factors obtained from experimental df data were used for calculation of k1 and k2:
and
k1 = 1,7 1014 exp(-159000/RT),

k2 = 1,1 1015 exp(-171000/RT),
291
The dependencies of kinetic constant k1, k2, and k versus temperature T3 are shown in
Fig.7. The kinetic constant k was used for calculation of df(cd) for NG-DGDN mixture. The
calculation results correspond to the experimental data very well (Fig.2).
13
ln(k)
3
12
2
1
11
10
1.05
1.10
1.15
1.20
103/3, -1
Fig 7.
The kinetic constants of decomposition of NG (1), DGDN (2), and the

overall reaction at detonation of NG-DGDN mixtures (3).
Dremins failure diameter theory has been used to predict values of df for liquid
substances [6-9]. For this purpose the kinetic parameters of reaction, shock temperature T(p)
and detonation parameters (D, pcj, Qv) of substance were required. In our work the df values
of some NE and their mixtures with inert substances are estimated using correlative
functions (1) and (2) and reaction time equation (3). For example, from experiments [1],
detonation of mixture of NGL with 35% methanol propagates only in steel tube (the
diameter is 10 mm and the wall thickness is 13 mm). Our calculation using kinetic
parameters of decomposition of NGL-methanol mixture (Tabl.3) shows that df of this
mixture is about 90 mm. The df of net PN up to now was not measured. Now one can
compute it using activation energy (158 KJ/mol) and pre-exponent factor (7.94 1013 s-1) of
thermal decomposition of PN [18]. Our estimation gives the value df 80 mm.
5.
CONCLUSIONS
The detonation failure diameters of the several nitroesters mixtures were investigated.
Detonation reactions in these mixtures were considered in terms of A. Dremins theory and
SGKR computer code calculations. All the results are explained in the light of homogeneous
thermal explosion theory for monomolekular reaction with kinetic parameters obtained in
usual kinetic experiments. The df values of some nitroesters and their mixtures with inert
substances were estimated by chemical kinetic and correlative equations.
292
Acknowledgements
V.V. Kondratiev took part in the carrying out of the experiments.
We thank professor B.N. Kondrikov for helpful discussions and Russian Foundation of
Basic Research (RFBR) for partial financial support of this work (Grant 01-03-32610(Grant 01-03-32610-).
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
B.N.KONDRIKOV AND V.M.RAIKOVA: Detonation limits of explosive solutions, Combustion,

Explosion and Shock Waves, v.13, 1, p. 55 62.1977.
B.N.KONDRIKOV, G.D.KOZAK AND A.V.STARSHINOV: Quenching detonation diameter of
nitroesters, Chemical Physics of Condensed Explosive Systems, Mendeleev Institute of
Chemical Technology, Moscow, v.104, p. 87-91, 1979 (Russ.).
B.N.KONDRIKOV, G.D.KOZAK, A.V.STARSHINOV AND V.A.POLIKARPOV: Kinetics of chemical
reactions at detonation of nitroesters, Problems of Theory of Condensed Explosive Systems,
Mendeleev Institute of Chemical Technology, Moscow, v.112, p.91-96, 1980 (Russ.).
B.N.KONDRIKOV, G.D.KOZAK, A.V.STARSHINOV AND A.V.POLIKARPOV: Kinetics of chemical
reactions at the detonation of nitroesters, Detonation, IV National Symposium on Combustion
and Explosion, Akad. Sci. USSR, Alma-Ata, p.18-21, 1980 (Russ.).
B.N.KONDRIKOV, V.V.KONDRATIEV AND G.D.KOZAK: Effect of additives on quenching
diameter of nitroesters, Combustion, Explosion and Shock Waves, v.19, 4, p.127-130, 1983.
DREMIN A.N: The failure detonation diameter of liquid explosives, DAN SSSR, v.147, p. 870873, 1962 (Russ.).
DREMIN A.N. AND TROFIMOV V.S: Calculation of failure detonation diameter of liquid
explosives, PMTF, 1, p. 126-131, 1964 (Russ.).
J.W.ENIG AND F.W. PETRONE: The failure diameter theory of Dremin, 5th Symposium (Int.) on
Detonation, Pasadena, Ca. Aug. 18-21, p. 99-104, 1970.
C. N.TARVER, R. SHAW AND M.COWPERTHWAITE: Detonation failure diameter studies of four
liquid nitrialkanes, J.Chem.Phys. v.64, 6, p. 2666-2673, 1976.
V.N.GAMEZO AND B.N. KONDRIKOV: Calculation of kinetic constants for decomposition of
niromethane and TNT from data on dependence of failure diameter on initial temperature,
Proc. of IV Conf. (USSR) on Detonation, Telavi. v.1, p. 111-117,1988 (Russ).
V.N.GAMESO: The heat release kinetics in the reaction zone of detonation wave, Cand. Sci.
(Ph. D.) Thesis, Mendeleev University of Chemical Technology, Moscow, 1992 (Russ.).
V.N. GAMEZO, B.N. KONDRIKOV AND G.D. KOZAK: Dependence of detonation failure
diameter of nitroesters on initial temperature, Chemical Physics (Russia), v.12, 5, p. 719722,1993.
B.N. KONDRIKOV, V.N.GAMEZO, G.D.KOZAK AND S.M. KHOROSHEV: Sulfuric acid influence
on the nitrocompounds detonation reactions, J. de Physique IV, Colloque C4, supplement au
J. de Physique III., v. 5, p. C4-395 - C4-405,1995.
V.M.RAIKOVA, A.HALAK AND B.N.KONDRIKOV: Chemical kinetics at detonation of organic
compounds nitric acid mixtures, Proc. of the 30th Int. Annual Conf. of ICT, Karlsruhe, FRG,
June 29 - July 2 1999, p. 95-1 95-12.
V.M.RAIKOVA, B.N.KONDRIKOV AND A.HALAK: Kinetics of chemical reactions at detonation
of mixtures of organic compounds with nitric acid, Combustion, Explosion and Shock Waves,
v.37, 3, p. 106-114, 2001.
293
[16] A.I.SUMIN, V.N.GAMEZO, B.N.KONDRIKOV AND V.M.RAIKOVA: Shock and detonation

general kinetics and thermodynamics in reactive systems computer package, Trans. of the 11th (Int.) Detonation Symp., Snowmass, Colorado, USA. August 31-September 4, 1998.
Bookcomp, Ampersand, p. 30-35, 2000.
[17] G.D.KOZAK: Measurement and calculation of the ideal detonation velocity for liquid
nitrocompounds, Combustion, Explosion and Shock Waves, v.34, 5, p. 581-586, 1998.
[18] B.A.LURIE AND B.S. SVETLOV: The kinetic parameters of the primary stage of thermal
decomposition of organic nitrate, Kinetics and Catalysis (Russia), v.35, 2, p.145-154, 1994.
294
SOME REGARDS ABOUT COLD POWDERS PLASTICIZERS

Traian Rotariu and Octavian Orban
Military Technical Academy, Department of Ammunitions, Missiles and Explosives,
81-83 George Cosbuc Boulevard, Bucharest, RO
Abstract:
This article presents experimental tests results done by authors in order to determine
explosive and ballistic properties of spherical gun powder modified by adding a cooling
agent, a plasticizer like glycidylazide polymer, having a low molecular mass. Assuming
that by adding such a gas generator the combustion temperature of the powder can
be lowered, by keeping the same ballistic properties, the authors developed a laboratory
technology for the doped powder manufacture and they made a series of tests in order to
characterize it.
Keyword:
1.
glycidylazide polymer (GAP), spherical powder, cold powder
INTRODUCTION
It is well known that the high values of the explosion heat produced by gun powders in
an armature system have an important influence over the damage rate of the barrel. This
observation started the proceedings for the implementation of triple base powders
(nitrocellulose, nitroglycerine, nitroguanidine), which are used nowadays on large scale in
high caliber armament ammunitions. The cooling agent used in this situation was nitro
guanidine added in pulverulent condition in nitrocellulose paste, plasticized with
nitroglycerin. The adaptation of such powder to a reduced caliber is practically impossible
due to the technological difficulties encountered at the extrusion of the paste containing solid
nitro guanidine. Moreover the vivacity of such powder would be much too low in the
specific temperature and pressure situation encountered at the reduced caliber armament.
In this context the authors decided to cool a powder calculated for the reduced caliber
armament by adding a liquid gas generator agent, having nitrocellulose plasticizing
properties. Such a plasticizer is glycidylazide polymer obtained by complete azidifcation of
a low mass polyepichlorohydrine (M1000 g/mol) according to the reaction:
CH2 - Cl
R (CH2 - CH - O)n H
CH2 - N3
+
nNaN3
R (CH2 - CH - O)n H
nNaCl
The authors starting assumption was that using a GAP type plasticizer in a spherical
powder can lead to the decreasing of the explosion heat and to the increasing of the specific
volume, the ballistic properties being at least similar to those of the usual spherical powder.
For experimental purposes was chosen the obtaining technology of the spherical powder,
because it is easy to adjust it to the laboratory specific conditions, it is reducible, safe and
allows the adding of new operations in the technological flux. Actually, these advantages are
distinguished also by the fact that in the last decade, the spherical powders have imposed
themselves in almost all the reduced caliber armament system.
295
2.
EXPERIMENTS
2.1. GAP preparation
GAP was preparated by treating 92 g polyepychlorhydrine having molecular mass of

1000, with 71.5 g sodium azide, in polar solvent (DMF, DMSO) in presence of a surfaceactive agent, for 12 hours, at 100oC. The reaction product was washed with hot water and
then with methylene chloride. After drying with anhydrous sodium sulphite and solvent
removal, 80.5 g GAP were obtained according to a 87% yield.
The complete azidification was demonstrated by IR spectrometry, observing the
complete disappearance of the C-Cl group absorption (650 cm-1) and the appearance of a
prominent absorption around 2100 cm-1, corresponding to the azide group.
The obtained GAP was a viscous liquid, amber like color, 1.3 g/cm3 density, having the
following elementary composition: C-36.1%; H-5.15%; N-40.5% and 980 average molecular
mass.
2.2. Spherical powder preparation
After a short experimental investigation, the authors drew the conclusion that the
powders containing nitrocellulose and GAP have special advantages for 9 mm caliber
ammunition armament system. Based on this conclusion the obtained powder samples were
destinated to this ammunition type and were produced through a classical spherical powder
manufacture technology [1].
The advantages of such a technology are: facility in preparing row materials, a reduced
number of steps during the process, reduced work time for each step, elasticity in time and
production volume, good reducibility of the final product, safe exploitation.
The steps for GAP and nitroglycerine plasticized spherical powder manufacture were:
Apparatus preparing and row materials dosage
Apparatus: 4 necks 5 l glass reactor; 2 necks 1.5-2 l flask; variable speed stirrer;
vacuum pump or aspirator pump; manometer; refrigerator; thermometer 0-150oC; 0.5 l
funnel; Nuce filter; 1 assay flask; sieves; adjustable stove; graphitization drum.
Raw materials: nitrocellulose (alcoholic or moisture); nitrocellulose pulp with
nitroglycerine moisture; filtered and dried glycidylazide polymer; diphenylamine (stabilizer);
detergent and stabilizer; aqueous solution sodium sulphate 10%; dibuthylphthalate
(regulator).
Organic lacquer preparation
79 g nitrocellulose with 10 g nitroglycerine, 10g GAP and 1 g diphenylamine are being
added in a stirring vessel and dissolved in ethyl acetate by vigorous stirring.
Lacquer dispersal in water
The dispersal is realized by adding under vigorous stirring of the organic lacqer into the
main vessel, where were poured the stabilizer and the detergent in aqueous solution and a
part of ethyl acetate is vacuum distillated.
Solvent removal
In vessel is added a sodium sulphate solution in order to stabilize the suspension. The
heating and vacuum distillation (0.1-0.3 atm) are being continued, keeping an vigorous
stirring until the complete solvent removal.
296
Impregnation with regulator

Powder beads separation and drying
Flattening and sorting
Drying
Graphitization
Following the previous procedure, the authors produced 100 g batches of nitrocellulose
glycidylazide polymer and nitroglycerine based powder. After testing the physical-chemical
characteristics, it came out that the powder corresponds from volatiles content,
hygroscopicity, temperature sensitivity and stability point of view.
2.3. Explosion heat and specific volume determination
In order to certify the cold property of the powder, the authors made explosion heat
and specific volume determination tests for the GAP doped powder. With this purpose in
view they used an IKA adiabatic calorimeter (fig. 1) and a Julius Peters gasometer (fig. 2).
According with standard procedures for these instruments, the authors made 10
measurements using about 1.5 g powder samples.
By disregarding the condensed water volume and the yield of vaporization heat, the
authors obtained the following averages: explosion heat 906 kcal/kg and specific volume
794 l/kg.
The authors tested, following same procedures, a classical spherical powder containing
10% nitroglycerine moderated with dibuthylphthalate. They obtained an average of 1070
kcal/kg for the explosion heat and an average of 668 l/kg for the specific volume.
Fig. 1. Adiabatic calorimeter
Fig. 2. Gasometer
2.4. Ballistic properties determinations

In order to identify the ballistic properties of the nitrocellulose based powder plasticized
with glycidylazide polymer, the authors made 100 cartridges of 9x19 mm caliber, having a
FMJ 7 g bullet with lead core, containing 0.34 g GAP powder.
297
Fig. 3.
Powder seen at the
microscope
Fig. 4.
Cartridge assembling
Fig. 5.
9 mm ballistic barrel
For testing, the cartridges were shot through a 9 mm ballistic barrel designed for 9x19
mm ammunition system, provided with piezoelectric pressure transducers and B2000 MAS
(AVL - Austria) bullet speed measure system. The identifications were made using 0.34 g
GAP powder and Sinoxid fuses.
The experimental shooting results are shown in table 1, 2 and drawing 6
Table 1.
Statistical data
Maximum value
Minimum values
Average
Maximum variation
Standard deviation
Pmax
[bar]
1880.3
1704.4
1807.4
175.9
52.6
t
[s]
0.525
0.384
0.431
0.141
0.046
v25
[m/s]
381.67
374.62
378.73
7.06
2.15
W
[J]
542.6
522.8
534.3
19.9
6.1
Table 2.
Shooting
no.
1
2
3
4
5
6
7
8
9
10
Pmax
[bar]
1869.6
1849.6
1880.3
1755.5
1796.9
1803.5
1795.9
1823.2
1704.4
1794.7
t
[s]
0.484
0.384
0.396
0.425
0.385
0.525
0.430
0.461
0.423
0.399
v25
[m/s]
374.62
380.54
379.47
377.64
380.03
376.30
379.83
381.67
377.72
379.62
W
[J]
522.8
539.4
536.4
531.2
538.0
527.2
537.4
542.6
531.5
536.8
298
Fig. 6. Presure curves by time
3.
DISCUSSION
The authors consider that the final results are at least interesting as they indicate the fact
that through adding in the spherical powders 10% of gas generator compound like liquid
glycidyl azide polymer, there can be obtained substantial reduction (15%) of the explosion
heat and important increasing (18%) of the specific volume.
Ballistical testing of GAP doped powder indicates highly rigorous values for 25 m speed
(v25 med= 379 m/s with standard deviation of 2.5) while maximum pressure values (Pmax max=
1880 bar and Pmax med=1807 bar) are significant smaller then those mentioned in NATO
standards.
The authors also consider that GAP addition brings technological advantages by
improving the product homogeneity. It could be an good explanation for the results obtained
in v25 mesurement.
4.
CONCLUSION
The present experimental study did not intend to be an exhaustive approach of the
subject. It can be considered as a step towards the producing of a cold powder intended for
the infantry armament, as it is obvious that for the implementation of a new powder in the
reduced caliber ammunition fabrication are required much more experimental investigations.
Anyway this contribution could represent the first step in the process of implementing
new kinds of spherical gun powders specialy designed for reduced explosion heat and
pressure values. These properties could easily recommend such a cold powder for automatic
high shooting cadence armament, where heat and temperature of explosion values might
have a particularly importance.
REFERENCES:
[1]
[2]
FRANKEL, B.MILTON, FLANAGAN, E.JOSEPH: Energetic hidroxi-terminated azido polymer,

U.S. Patent 4,268,450, 1981
J.QUINCHON, J.TRANCHANT, M.NICOLAS: Les poudres, propergols et explosifs, vol. 3, p. 87,
Technique et Documentation, Paris,1986
299
IMPROVED INSENSITIVE HYTEMP/DOA BONDED

HMX AND RDX MIXTURES BY PASTE PROCESS
Karl P. Rudolf
Diehl Munitionssysteme GmbH & Co KG, PBX-Center Maasberg
1.
INTRODUCTION:
A study on miniaturized electric detonators and insensitive booster high explosives

showing two initiation shock levels for No-Go in the BICT-Small Scale Gap Test (SSGT)
26 kbar (TL 1376-800) and
42,2 kbar (STANAG 4170)
In addition passing of Fast Cook Off and Bullet Impact Tests with 12,7 mm mild core
munition at reaction level type IV is mandatory.
The intended use of an European made PBXW11 and PBXN11 at that time failed
already showing already to low sensitivity at SSGT, however an additional treatment in a
Jahnke and Kunkel high shear mixer could improve the shock insensitivity to the threshold
of TL 1376-800 as we expected on our experience on HTPB bonded castable insensitive
high explosives.
Above mentioned experiences led to the decision to develop pressable IHEs using
standard RDX and HMX crystals quality A and a solvent paste process suitable for a DRAIS
vertical and High-Shear Mixers.
2.
WORK PERFORMED:
First step was to find a proper solvent for the Hytemp 4454/DOA lacquer capable to
achieve a full transparent mixture and in addition to achieve sufficient separation of either
HMX or RDX impurities from the main content and to smooth rough surfaces of the
standard HE crystals. Ethyl acetate only from original slurry process cannot sufficiently
perform this in the paste process.
Some few 5 kg batch mixings using coarse grain HMX class and RDX class and fine
grain HMX class 5/ class and RRDX class 2 (all type 2) as received from crystal producer
and different solvents resulted in:
-
Acetone/ ethyl acetate solvent mix shows improved SSGT results
Use of 45 m fine grain crystals instead of 15m material shows improved

pressing behavior, slightly increased densities and improved insensitivity
both at 8 % and 4 % binder systems content.
STANAG 4170 qualification at WIWEB/BWB of these first mixtures at the TL 1376800 with all HMX and RDX mixtures with 8 % binder system were passed and HMX with 4
% binder system just did not meet SSGT and Fast Cook Off criteria.
300
During a second step of mixings the morphology of the larger coarse grain crystals was
considered by limiting crystal sizes in a first step to 630m and in a second step to 500 m.
These measures resulted in:
-
HMX/8 % binder system meets in addition SSGT requirement

of STANAG 4170.
HMX/4 % binder system meets all requirements of TL 1376-800.
RDX/8 % binder system did not show significant improvement in SSGT testing.
RDX type A/8 % binder system did show some improvement in SSGT testing but
could not meet STANAG 4170 threshold and failed at Fast Cook Off. Addition of
at least 5 % HMX led to passing of Fast Cook Off.
RDX type A, fine grain RDX milled from coarse shows reduced sensitivity just
meeting TL 1376-800 requirement.
Fist SSGT results of RDX/6% binder system mixtures are indicating no

difference of shock insensitivity levels.
In addition all mixtures require a specific pressing pressure of 1,2 kbar achieve > 98 %
t.m.d at pellet diameter 50 mm.
This extremely good pressing behavior causes however DOA exudation during
temperature shock testing of RDX/8 % binder system in heavy steel casings. First step of
trimodal mixtures either inert or high explosive thus increasing crystals surfaces are
showing significant improvements and it looks to be independent of adding all components
in the beginning or adding additional components later in a repair mode. Some more details
on this ongoing effort will be given during presentation.
3.
CONCLUSIONS:
During this effort RDX and HMX crystals with significant crystal defects where used
and sufficient insensitivities of pressable RDX and HMX mixtures achieved. Further
improvements of RDX and HMX crystals quality will result in a reduced Hytemp/DOA (1:3)
binder system and probably in an additional reduced DOA-plasticiser content.
The paste process does not require new specific equipment (except oilless vacuum
pumps) and offers a later repair possibility of the mixtures if any criteria of the molding
powder is not met.
Acknowledgements:
Many thanks to my coworkers especially Mr. H. Hofmann and J. Licha, the cooperation
of Mr. Roland Wild, WIWEB and the contributions of Mr. Richard Wild.
301
Technology Symposium
Improved Insensitive Hytemp/DOA Bonded

HMX and RDX Mixtures by Paste Process
Karl P. Rudolf
NDIA
2003 Insensitive Munitions and Energetic Materials
March 10 - 13, 2003
Orlando, FL, USA
A Company of
Diehl VA Systeme
Setting of Effort
German DoD on miniaturized electric detonator and insensitive boosters
required pressable insensitive high explosives
at two initiation shock levels for NoGo (BICT-SSGT)
26 kbar (TL 1376-0800)
42,2 kbar (STANAG 4170)
type V reaction at
Bullet Impact Test 12,7 mm mild core
Fast Cook Off Test
Intention: Use of an European made PBX W11 and PBX N9
A Company of
Diehl VA Systeme
Technology-symposium-10-03-03
2
25.3.2003
Setting of Effort (cont.)

Inspection analyses show not sufficient shock insensitivity
Additional treatment of this PBXN9 in our high shear mixer show
improvement of insensitivity close to the threshold of TL 1376-0800
Conclusion:
Install own capabilities to develop and produce small batches of
insensitive pressable high explosives
Develop process capable to use standard RDX and HMX crystals type B
Reduce required specific pressure for pressing
Use of the Hytemp 4454/DOA 1/3 binder system
Use of in-house DRAIS vertical finger and J+K high shear mixer
Paste process using organic solvents only for additional treatment of the
HE crystals during mixing
A Company of
Diehl VA Systeme
3
25.3.2003
302
HE Crystals
HMX / RDX Type B
A Company of
Diehl VA Systeme
4
25.3.2003
Comparison
Slurry Process
Paste Process
Hytemp
DOA
Solvent
Vacuum
HE-Crystals
Water
A Company of
Diehl VA Systeme
5
25.3.2003
Work Performed, First Step
Definition of a solvent (mixture) capable
Smoothing HE type B crystal surfaces
achieving a fully transparent lacquer of Hytemp/DOA
having proper evacuation rates to separate HMX and RDX crystals

and keep them separated during coating of the crystals
Different solvents tested
a proper candidate is a mix of Ethylacetate/Acetone
A Company of
Diehl VA Systeme
6
25.3.2003
303
Work Performed, First Step (cont.)

Reduction of specific pressing pressures
Coarse grain: HMX or RDX class 3
Fine grain: 45 m crystals allow additional
reduction of specific pressure
compared to 15 m crystals
to achieve t.m.d. of more than 98 %
Result:
Required specific pressures, fine grain 45 m, vacuum < 1 mbar; l/d ration up to 2
kbar
diameter
22 mm
0,85
50 mm
0,70
105 mm
0,55
Note:
Some values valid for harmonic trimodal mixtures, thus significant

weight reduction of pressing tools possible symposium report shows
values for non harmonic bimodal mixtures
A Company of
Diehl VA Systeme
7
25.3.2003
Work Performed, First Step (cont.)

Example of results from WIWEB/BWB STANAG 4170 Analyses
Versuch/Test
DXP 1380, Charge 1
Vakuumstabilitt
Schlagempfindlichkeit
Reibempfindlichkeit
DXP 1380, Charge 2
0,05 ml/2,5 g
0,08m1/2,5 g
7,5 Joule schlagempfindlich
7,5 Joule schlagempfindlich
160 N reibempfindlich
240 N reibempfindlich
thermische Empfindlichkeit
d=2 mm, D, wenig empfindlich
Stowellenempfindlichkeit
R 38,80 kbar (10 mm WS), kR 30,40 kbar (13

mm WS)
Typ 5 (Abbrand) unempfindlich
R 42,20 kbar (9 mm WS), kR 35,70 kbar (11

mm WS)
Cook-Off-Verhalten
Beschuss mit 12,7 mm Munition
Verpuffungstemperatur
279C
278C
leicht entzndlich
leicht entzndlich
elektrostat. Empfindlichkeit
unempfindlich
unempfindlich
Detonationsgeschwindigkeit
- 8453 m/s
- 8402 m/s
Gurneykonstante
2,38 mm/ps
2,34 mm/ps
E-Modul (aus Druckversuchen)
41,53 MPa (20C); 116,80 MPa (-40C)
22,41 MPa (20C); 83,13 MPa (-40C)
long. Schallgeschwindigkeit
vs
spez. Wrmekapazitt
cp [Joule/(g*K)]
thermische Lngenausdehnung
2165 m/s (20C); 2512 m/s (-30C)
2121 m/s (20C); 2493 m/s (-30C)
Wrmeleitfhigkeit
Entzndlichkeit
0,90
0,90
im Mittel 6*10"5 C1
im Mittel 6,5*10- C"
0,47 W/m*K (bei Raumtemperatur)
0,47 W/m*K (bei Raumtemperatur)
A Company of
Diehl VA Systeme
8
25.3.2003
Qualified Pressable Insensitive High Explosives of Diehl;

Qualification in accordance with STANAG 4170 and TL 1376-0800
Status: 11.2001
DXP 1380/1 DXP 1380/21)
92
92
8
8
Type
Composition
%
Density
Det. Velocity.
Sensitivity GAP
HMX
RDX
Al
Bindersystem
(g/cm3)
Vo (m/s)
(K bar)
go/no go
1,76
1,8
1,68
1,7
8400
8600
8200
8350
42/36
22/20
33/28
28/24
2)
qualified
410/25830/01
1)
2)
DXP 2340
96
4
1,74
qualified
WIWEB Doc.-No.
DXP 2380
92
8
8450
39/31
Status
DXP 1340
96
4
410/25860/01 410/25861/0
1
2)
410/25818/01
differs from DXP 1380/1 by grain size distribution

fulfills at present only TL 1376-0800 standard requirements
A Company of
Diehl VA Systeme
9
25.3.2003
304
A Company of
Diehl VA Systeme
10
25.3.2003
A Company of
Diehl VA Systeme
11
25.3.2003
Work Performed, First Step

Results:
HMX and RDX mixtures with 8 % binder meet criteria of TL 1376-0800
HMX mixtures show reduced SSGT shock sensitivity compared to RDX
Conclusion:
Additional reduction of porosities required to improve insensitivity
HMX: remove crystals with high risk of increased internal porosities
RDX: as above and use of improved HE crystals type A
A Company of
Diehl VA Systeme
12
25.3.2003
305
Work Performed, Second Step
Remove high risk crystals
1.)
maximum grain sizes 630 m
2.)
maximum grain sizes 500 m
Results:
no additional improvements of RDX / 8 % binder
HMX 4 % binder meet all criteria of TL 1376-0800 at densities 1,81 - 1,82 g/cm3
(WIWEB-BWB results)
HMX / 8 % binder meets less sensitive status of STANAG 4170

(WIWEB-BWB results)
A Company of
Diehl VA Systeme
13
25.3.2003
Work Performed , Second Step (cont.)
Improved RDX Crystals Type A
iRDX / 8 % binder failed Fast Cook Off, no significant improvement of shock

insensitivity
Fast Cook Off passed after addition of 5 % (+) HMX
Coarse grain iRDX with milled fine grain iRDX show no improvement
No different of SSGT results at 8 % or 6 % binder system
Insensitive RDX analyses are planned
A Company of
Diehl VA Systeme
14
25.3.2003
Work Performed , Second Step -- Results:
Further improvements on Hytemp / DOA bonded RDX and HMX by solvents

process led to following results:
GAPtest
HMX /
HMX /
RDX /
iRDX /
4% binder
8% binder
4% binder
4% binder
32,9 / 26
45,2 / 42,2
32,9 / 28
32,9 / 30,4
Type V
Type V
Type V
Failed
(kbar)
FCO
* Slurry process
t.m.d
min 98%
Failure of iRDX / 8% binder was not expected due to improved crystallization.

A Company of
Diehl VA Systeme
15
25.3.2003
306
Paste Process
Advantages
Use of type B crystals sufficient
Significantly reduced sensitivities
Significantly reduced specific pressures
Complete recycling of solvents possible
Quick improvement rework of mixtures possible
Use of existing high shear or other composite propellant mixers

possible thus avoiding additional buildings and equipment
A Company of
Diehl VA Systeme
16
25.3.2003
Shaped Charge
Shaped Charge 80mm
DXP 1340/2, 0.9 kbar, 1.81 g/cm3
A Company of
Diehl VA Systeme
17
25.3.2003
Disadvantages
Agglomeration of coated crystals more difficult to control
Agglomerates tend to stick together during transportation and

storage
possible cure: dusting with Aerosil
Standard primers e.g. K1, Binder 14 to achieve sufficient wall

adhesion fail
cure: use WENKO Antirutschbeschichtung, qualification planned at
WIWEB-BWB
A Company of
Diehl VA Systeme
18
25.3.2003
307
Bomblet M77 Impact into 155 mm IHE
A Company of
Diehl VA Systeme
19
25.3.2003
Conclusions
Today's HMX crystals Type B are sufficient for acceptable

insensitive HE mixtures
Improvement of HMX crystals will result in additional insensitivity
against shock stimuli
Shock insensitivity improvement is expected to be achieved by use

of Insensitive RDX crystals only
In both cases an additional reduction of binder system seems to be

possible to achieve sufficient insensitive pressable mixtures by
Paste Process using organic solvents
A Company of
Diehl VA Systeme
20
25.3.2003
308
CHANGES IN SENSITIVENESS OF FLEGMATIZED HIGH

EXPLOSIVES AFTER ARTIFICIAL AGING
Jakub eleovsk*, Ji Pachm*and Marcel Hanus**
* Department of Theory and Technology of Explosives, University of Pardubice,
** Military Institute for Weapon and Ammunition Technology,
Dlouh 300, 763 21 Slavin, Czech Republic
Abstract:
An isothermal heating of four samples of wax flegmatized high explosives at 70C for
113 days was performed to simulate artificial aging. This research was done as a part of
an extensive program dealing with characterization of aging process of military
explosives. Changes in sensitiveness to impact, friction and electrostatic discharge due
to artificial aging were evaluated. Sensitiveness increase to all stimuli was observed
despite the thermal characteristics (from DTA, VST) remained same. The tests were as
well performed on pure RDX, PETN and PETN/TNT (90/100) for comparison. The
changes in sensitiveness were not observed in case of these samples.
Keyword:
1.
sensitiveness to external stimuli, high explosives, artificial aging,

thermal stability
INTRODUCTION
There are many requirements demanded on explosives used for military or ammunition
applications[1]. Beside high performance, simple preparation and manipulation safety, which
is today more and more alluded to, it is foremost their longtime stability and shelf life. This
demand results logically from the necessity of their longtime storage.
Explosives are in general unstable substances and they are decomposing (faster or
slower) in time. This chemical decomposing is usualy accompanied by changes of various
other (physical, mechanical, performance, safety, ) parameters. These changes can affect
the functionality of whole weapon/ammunition system and it is necessary to pay attention to
them. Mere monitoring of various properties is in most cases not sufficient for prediction of
system surveilance. Artificial aging of explosives (or whole ammunition object) must be
used in this case.
The simplest method of artificial aging is isothermal heating at elevated temperature.
This method assumes validity of Arrhenius equation for chemical decomposition reactions.
This means that elevated temperature increases decomposition rate and speeds up the total
rate of aging. This method allows relatively simple prediction of decomposition progress
based on determination of temperature coefficient of given reaction. The simplicity of shelflife prediction is however at the cost of insufficient embodiment of another influences
(thermal changes and shocks, moisture, mechanical loading,). The decomposition
reactions can also change their mechanism at higher temperatures making the extrapolation
to lower temperatures very inaccurate.
309
Other methods of artificial aging, briefly mentioned in[2], include for example thermal
shocks and cycles, mechanical loading, influence of moisture and taking into account
cumulative damage principe. These methods are certainly more believable, but prediction
based on them is somewhat more difficult.
The biggest part of attention of experiments dealing with aging of explosives is paid to
their chemical or thermal stability. Another part, mainly in latest years, is given to changes
of mechanical and structural properties. Very little attention is however given to changes of
sensitiveness of explosives to accidental initiation stimuli during aging, natural or artificial.
The sensitiveness of high explosives to various external stimuli is summarized for example
in[3, 4], but the association with aging is not discussed.
Narang et al.[5] studied sensitiveness of high explosives to external ignition stimuli in
connection with their aging and found out that sensitiveness of high explosives (TNT and
Dentex, unspecified aluminized HE) to impact after exposition to elevated temperature
(120C, till 672 hours) slightly increased. Sensitiveness to friction did not change.
Sensitiveness of Composition B to impact[6], impact, friction, electrostatic discharge,
gaptest[7] and sensitiveness of TNT/RDX/Al mixtures to impact[8, 9] did not change after
long-term storage.
Studies of PBXs[10, 11, 12] gave similar results. Examined explosives did not change their
sensitiveness (Steven Test, Small Scale Gap Test) after aging as well.
Changes in sensitiveness of explosives during aging have been not sufficiently examined
yet. Published works dealed only with small number of HE types (based on TNT,
TNT/RDX, specific PBX), and only one kind of test is usually applied in most cases (rarely
two[5] or more exams[7]).
2.
EXPERIMENTAL
2.1
Used Samples
A set of samples, representing some of the explosives established in Czech Army, was
aged artificially[13] and changes of sensitiveness to impact, friction and electrostatic
discharge(ESD) were evaluated[14]. The obtained results were compared with the results of
unaged samples. In this paper, the changes of sensitiveness of wax flegmatized high
explosives (AIX-1, A-IX-2, OKFOL, Np 10) will be discussed. The results for pure RDX
and PETN are given for comparison. The results of differential thermal analysis (DTA) and
vacuum stability test for all samples (aged and unaged) are assigned too. The encoding and
composition of samples is in table 1.
310
Table 1. Labelling and composition of samples

Label
A
B
C
D
E
F
G
2.2
Type
A-IX-1
Composition
95% RDX
5% ceresin/stearin (60/40)
resinol orange
A-IX-2 20% Al
4% ceresine/stearine (60/40)
resinol orange
RDX
100% RDX, recrystallized from acetone, p. a.
OKFOL 97% HMX
3% wax
Sudan red
Np 10 90 % PETN
10 % wax
Np 10T 90% PETN
10% TNT
PETN 100% PETN, recrystallized from acetone p.a.
Artificial Aging
Artificial aging was performed as isothermal at temperature 70C for 113 days and zero
air moisture. The powdered material was placed (in amount of 30 g) into a glass bowl (in
layer about 1 cm high) and covered by aluminum foil. Samples arranged in this way were
placed in climatic boxes and examined without any other adjustment after expiring the
predeterminated time. The unaged samples were marked as X-0, the aged ones as X-113.
2.3
Determination of Sensitiveness
Sensitiveness to impact was determined on the BAM apparatus, in accordance with

STANAG 4489 [15], implemented in OS 137601[16]. The E50 value, corresponding to impact
energy needed for initiation of the sample with the probability of 50 %, and its standard
deviation, , was calculated.
Sensitiveness to friction was determined on the BAM apparatus, in accordance with
STANAG 4487[17], implemented in OS 137601[16]. The F50 value, corresponding to load of
friction stick needed for initiation of the sample with the probability of 50 %, and its
standard deviation, , was calculated.
Sensitiveness to ESD was determined on apparatus developed at the Dept. of Theory and
Technology of Explosives[18, 19], in accordance with OS 137601[16]. The E50 value,
corresponding to electric energy needful for initiation of the sample with the probability of
50 %, and its standard deviation, , was computed. The undamped discharge was used for
the test.
311
2.4
Determination of Thermal Characteristics
Thermal characteristics of samples were determined using DTA 550 Ex apparatus.

Thermal stability was evaluated using the vacuum stability test performed in accordance
with STANAG 4556[20], implemented in OS 137601[16]. The samples were stored at
temperature of 120C for 40 hours.
3.
RESULTS
3.5
Sensitiveness
3.5.1. Sensitiveness to Impact

The results of impact sensitiveness test (the E50 value and its standard deviation ) are
stated in table 2 and changes of E50 are demonstrated graphicaly in the figure 1.
3.5.2. Sensitiveness to Friction
The results of friction sensitiveness test (the F50 value and its standard deviation ) are
stated in table 3 and changes of F50 are demonstrated graphicaly in figure 2.
3.5.3. Sensitiveness to Electrostatic Discharge
The results of ESD sensitiveness test (the E50 value and its standard deviation) are stated
in table 4 and changes of E50 are demonstrated graphicaly in the figure 3.
3.6
Thermal Characteristics
The results of vacuum stability test are stated in table 5. DTA thermograms are situated
at the end of this article (figures 4-10).
312
Table 2. Sensitiveness to impact

Sample
A
B
C
D
E
F
G
Fig 1.
Virgin
E50 [J]
23.2
0.048
24.6
0.047
8.0
0.035
29.8
0.099
10.0
0.050
6.3
0.091
4.1
0.097
Aged
E50 [J]
14.6
0.098
5.9
0.100
8.3
0.028
9.0
0.026
14.0
0.054
5.7
0.087
3.6
0.087
Changes in the sensitiveness to impact after artificial aging.
313
Table 3. Sensitiveness to friction.

Sample
A
B
C
D
E
F
G
Fig 2.
Virgin
F50 [N]
[N]
252
70
279
41
182
47
222
24
174
34
96
47
64
15
Aged
F50 [N]
[N]
240
39
293
36
172
36
189
41
138
28
88
11
55
8
Changes in the sensitiveness to friction after artificial aging.
314
Table 4. Sensitiveness to electrostatic discharge.

Sample
A
B
C
D
E
F
G
Fig 3.
Virgin
E50 [mJ]
[mJ]
173
25
465
48
55
8
234
19
146
17
128
10
30
5
Aged
E50 [mJ]
[mJ]
158
12
218
35
67
6
173
16
158
10
75
7
20
4
Changes in the sensitiveness to electrostatic discharge after artificial aging.
Table 5. Vacuum stability test.

Sample
A
B
C
D
E
F
G
Virgin
V [ml]
0.86
0.82
0.22
0.64
0.42
1.23
0.79
Aged
V[ml]
0.62
0.84
0.44
0.84
0.77
1.51
0.22
315
4.
DISCUSSION
4.7
Sample A (A-IX-1)
The artificial aging of sample A caused increase of sensitiveness in all three tests,
mainly in the impact sensitivity test. The changes of sensitiveness to friction and electrostatic
discharge are relatively smaller. The thermal characteristics were not influenced.
4.8
Sample B (A-IX-2)
This sample showed remarkable increase of sensitiveness to impact and electrostatic

discharge after aging. The sensitiveness to friction slightly decreased. The DTA and VST
tests indicate, that thermal characteristics and stability were similar before and after the
aging.
4.9
Sample C (RDX)
The sample C did not show any significant changes neither in sensitiveness nor in
thermal characteristics.
4.10
Sample D (OKFOL)
The sensitiveness of this sample underwent to the biggest changes of all the tested
samples. The significant increase of impact sensitiveness, somewhat lower but still important
increase of friction and ESD sensitiveness. The thermal characteristics was not influenced
with the artificial aging.
4.11
Sample E (Np 10)
This sample showed increase of sensitiveness to friction and in reverse decrease of

impact and ESD sensitiveness. The maximum of decomposition peak (DTA) has slightly
moved to higher values in case of aged sample.
4.12
Sample F (Np 10T)
In case of this sample increase of ESD sensitiveness was observed. As far as thermal
stability concerned, this sample showed highest volume of gases, envolved in vacuum
stability test. This behavior (lower stability of PETN/TNT mixtures as against of TNT, or
PETN itselves) is well known, and the value of envolved gas volume did not increase
significantly after aging.
4.13
Sample G (PETN)
PETN was the most sensitive sample, as it was assumed. Its sensitiveness to all stimuli
increased with aging, but this change was not too fatal. The thermal stability did not change.
5.
CONCLUSIONS
Based on previous experiments it is believed that chemical changes in tested explosives

were negligible. Increase of sensitiveness observed during above mentioned tests is most
probably caused by baring of HE crystals. Used type of artificial aging does not appear to be
suitable for monitoring of sensitiveness changes of wax flegmatized explosives stored in
mild climate. The high temperature causing flowing down of the wax and baring of the high
explosives crystals is unlikely to occure in central Europe. This phenomenon however has to
be taken into consideration when planing deployment in very hot climate locations.
316
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
S. BREBERA: Vojensk trhaviny a technologie vroby trhavinovch nlo, University of

Pardubice, Pardubice 2001 (Dept. of Theory and Technology of Explosives textbook).
J. LEENDERS; C. A. VAN DRIEL ET AL: Proc. Int. Pyrotech. Seminar 27, p. 738, 2000.
J. A. ZUKAS; W. P. WALTERS: Explosive Effect and Applications, Springer Verlag, New York
1998.
M. SUESKA: Test Methods for Explosives, Springer Verlag, New York 1995.
R. NARANG; S. R. NAYAK: Def. Sci. J. 43(3), p. 263-267 (1993).
J. S. LEE; C. K. HSU: Thermochimica Acta 367-368, p. 371-374 (2001).
M. S. KIRSHENBAUM; L. AVRAMI: Report ARLCD-TR-84001, AD-A139611 (1984), abstract.
W. LI; S. JI; Q. SONG: Huozhayao Xuebao 19, p. 33- (1997), abstract.
L. WEN; N. FUSHUI. ET AL.: Proc. Beijing Int. Symp. Pyrotech. Explos. 3, p. 193- (1995).
S. K. CHIDESTER; C. M. TARVER ET AL.: CP505, Shock Compression of Condensed Matter,
p. 663-666 (1999).
D. J. IDAR; J. V. STRAIGHT ET AL.: CP505, Shock Compression of Condensed Matter,
p. 655-658 (1999).
M. T. WU; S. I. CHEN ET AL.: Proc. Int. Annu. Conf. ICT 20, p. 14 (1989).
J. PACHMN: New Methods of Stability and Shelf Life Evaluation of Military Explosives,
University of Pardubice (thesis), 2000.
J. ELEOVSK: Evaluation of Stability and Shelf Life of Military Explosives,
University of Pardubice (thesis), 2002.
STANAG 4489.
OS 137601 (Czech Defence Standard).
STANAG 4487.
J. STRNAD; J. MAJZLK: Proc. 4th Sem. on New Trends in Research of Energetic Materials,
University of Pardubice, p. 303-307, Apr. 2001.
J. KO; S. ZEMAN ET AL.: Proc. 5th Sem. on New Trends in Research of Energetic Materials,
University of Pardubice, p. 111-127, Apr. 2002.
STANAG 4556
317
6.
THE DTA THERMOGRAMS OF UNAGED AND AGED SAMPLES
Fig 4.
Thermogram (DTA) of sample A (A-IX-1).
Fig 5.
Thermogram (DTA) of sample B (A-IX-2).
318
Fig 6.
Thermogram (DTA) of sample C (RDX).
Fig 7.
Thermogram (DTA) of sample D (OKFOL).
319
Fig 8.
Thermogram (DTA) of sample E (Np 10).
Fig 9.
Thermogram (DTA) of sample F (Np 10T).
320
Fig 10. Thermogram (DTA) of sample G (PETN).
321
THERMOLYSIS OF KETO-RDX AND ITS PLASTIC BONDED

EXPLOSIVES WITH FLUORO POLYMERS
Gurdip Singh and S. Prem Felix,
Chemistry Department, DDU Gorakhpur University,
Gorakhpur 273 009
Abstract:
Thermolysis of keto-RDX and its plastic bonded explosives (PBXs) containing Kel F and
Viton A was studied using various thermo-analytical techniques. The PBXs decompose
at slightly lower temperatures and with higher heat release than K-6. Kinetic analysis of
isothermal data was made using a model free isoconversional method. The activation
energy is dependent on extent of conversion and different values were obtained for
thermolysis of K-6 and the PBXs. Thus even thermally stable, inert binders such as
Viton A and Kel F, alter the thermochemistry of K-6 in their PBXs.
Keyword:
1.
Keto-RDX, plastic bonded explosives, fluoro polymers, thermolysis,

isoconversional method.
INTRODUCTION
2-oxo-1,3,5-trinitro-1,3,5-triazacyclohexane (K-6) is the keto-derivative of 1,3,5-trinitro1,3,5-triazacyclohexane (RDX). K-6 is an interesting high explosive as it is having high
density ( = 1.93 g cm-3)[1] and high heat of formation (-101.0 kcal mol-1)[2]. The compound
can be prepared from cheap starting materials and is having 4% more energy than 1,3,5,7tetranitro-1,3,5,7-tetraazacyclooctane (HMX) at early volume expansions[2]. It also finds
application as a fuel component in inflatable vehicle occupant protection device with
reduced smoke[3]. However, the compound is more sensitive to impact than RDX[4], HMX[2]
and is thermally less stable than RDX[4]. The sensitivity of energetic compounds to
hazardous stimuli may be reduced by thoroughly coating them with thermoplastic polymers
to make PBXs[5]. Thus, Mitchell et al.[2] have coated K-6 with 5% Viton A to form a PBX
named RX-41-AB and small scale safety tests revealed that slight reduction in sensitivity to
drop hammer impact and electric spark was achieved by coating.
Our interest in thermolysis of PBXs arises from the fact that a systematic study on this
aspect is found rarely in open literature. Most of the studies reported on PBXs are either
routine characterizations or are not aimed at understanding the thermo-chemistry. Our recent
studies showed that the binder (hydroxyl terminated polybutadiene-HTPB) influence the
thermolysis of RDX, in their PBX[6]. The thermolysis of HMX takes place at a lower
temperature in its PBXs with Estane as binder and the extent of lowering of decomposition
temperature increases as percentage of binder increases[7]. The role of binder was found to be
in facilitating the condensed phase reactions and reducing the contribution of competing gas
phase processes in PBXs of RDX with Estane and Viton A[8]. In continuation of these
studies, thermolysis of K-6 and its four PBX formulations has been undertaken using various
thermo-analytical techniques and the results are presented here. Isoconversional kinetic
analysis of thermolysis of K-6 has not been reported yet and hence we used the above said
method to evaluate kinetic parameters for thermolysis of K-6 and its PBXs.
322
2.
EXPERIMENTAL
Samples of K-6, its PBXs with Viton A in the weight (%) ratios 95:05 (K-6V 9505),
90:10 (K-6V 9010) and with Kel F in the ratios 95:05 (K-6K 9505) and 90:10 (K-6K 9010)
were supplied by HEMRL, Pune. These samples were used as received. All the nonisothermal analyses were performed at a heating rate of 10 C min-1. TG-DTA experiments
on the samples were done on a Mettler Toledo Star system under inert atmosphere (flowing
N2 at a rate of 100 ml min-1, sample mass ~1-2 mg). DTA thermograms on the samples were
recorded using an instrument supplied by Universal Thermal Analysis Instruments, Mumbai.
Isothermal TG thermograms in the temperature range 170-180 C were recorded using
indigenously fabricated TG apparatus under static air atmosphere (sample mass ~25mg).
3.
RESULTS
3.1
TG-DTA
The TG-DTA thermograms of all the samples are shown in Figure 1 and corresponding
phenomenological data are summarized in Table 1. Thermolysis of K-6 occurs in a single
step with ~90% mass loss. No further mass loss was observed for K-6 up to 500 C. It is
evident from Table 1 that the thermolysis of PBXs occurs in multi steps. The first step in the
thermolysis of PBXs is invariably the decomposition of K-6. Interestingly, the onset
temperature (Ti), inflection temperature (Ts) and endset temperature (Tf) for the first step in
the PBXs that are summarized in Table 1, have lower values than that for pure K-6. The
extent of lowering increases as the percentage of binder increases in the PBXs. Kel F and
Viton A are fluoro polymers and hence their thermolysis occurs at high temperatures. Thus
the mass losses above 400 C for the PBXs are due to the thermolysis of binders in them.
Thermolysis of K-6 was observed in a single exothermic peak in DTA at ~202 C. The
exothermic peaks corresponding to thermolysis of K-6 in the PBXs (Table 1) were observed
at lower temperatures than that for pure K-6. A second exothermic peak was observed above
480 C for PBXs with Viton A. However for the PBXs with Kel F, there were no observable
processes in DTA, after the first exothermic peak. Integral and normalized peak area of the
samples for the DTA thermograms are also reported in Table 1. For the PBXs, the integral
area of the peaks were normalized according to the theoretical mass (%) of K-6 and the
binder for their respective thermolysis processes.
3.2
Isothermal TG
Isothermal TG thermograms on the samples were recorded at a narrow temperature

range of 170-180 C.
3.3
Kinetic Analysis of Isothermal TG Data
Kinetic analysis of isothermal TG data has been made by using an isoconversional

method as described in our previous[6-8]. The dependencies of activation energy (E) on
extent of conversion () were determined and are given in Figure 2.
323
4.
DISCUSSION
The results of our thermal analyses on K-6 was similar to that are reported in
literature[2,4]. Mitchel et al.[2] have reported an exotherm at 204.9 C for K-6 as per DSC. Our
result (202 C) is close to this value. Ostmark et al.[4] have reported the occurrence of a
strong exotherm having maximum at 195 C according to DSC. This lower value is
evidently due to the lower heating rate in their study, i.e. 5 C min-1. They[4] made an ocular
inspection, and reported that K-6 decomposes without melting. Mitchel et al.[2] have reported
that crystalline K-6 melts at 184-185 C, without mentioning the experimental method and
there was no mention of any corresponding endothermic peak in their DSC study also. The
ocular inspection of Ostmark et al.[4] reveled that a gaseous bubble has interacted with the
solid and formed a liquid phase. In order to assess, the occurrence of any competing reaction
channels, other than solid-state decomposition, DTA analyses of K-6 were done under
various conditions. No endothermic process was observed in the thermograms shown in
Figure 3, even under vacuum atmosphere (0.5 atm). Thus, we could not observe sublimation
or melting and hence it is reasonable to conclude that thermolysis of K-6 occurs in solid state
or in the condensed phase. However, when K-6 was confined by using a sealed DTA cup,
the exothermic peak was observed at slightly higher temperature.
It is interesting to note that thermally stable polymeric binders lower thermal stability of
K-6. Such a lowering was observed during DSC analysis of RX-41-AB by Mitchel et al.[2].
However sensitivity of RX-41-AB was lower as compared to pure K-6, even when there was
a slight decrease in thermal stability. Since, binders in the present samples are highly stable,
inert fluoropolymers, the lowering in thermal stability cannot be attributed to any chemical
interaction between the binder and K-6 or its decomposition products. In K-6, cumulation of
strong electronegative groups in the 1,2,3-positions leads to crowding and consequently a
dipole is generated[1]. The dipole leads to stronger orientation interactions and thus the
crystal lattice is stabilised more and thus physical thermal stability is enhanced. In such cases
where stabilizing influence of the crystal lattice is strong, even very small admixture of a
solvent can markedly decrease thermal stability. It is known that Viton A and Kel F clearly
dissolve nitramines at higher temperatures. Thus, in the case of PBX samples, the lower
thermal stability compared to K-6 may be due to marked dissolution of K-6 in the binder
matrix.
The integral normalized peak areas of DTA exotherms reported in Table 1 are in the
following order, K-6K 9010 < K-6 < K-6V 9010 < K-6V 9505 < K-6K 9505. Higher heat
release for the PBXs (except K-6K 9010) than K-6, show that the reaction pathways may be
different. Lesser heat release for K-6K 9010 is associated with lesser mass loss (%) in the
first step. Thus the actual mass (%) of K-6 may be lower in K-6K 9010 than 90 %.
4.1
Kinetic Analysis
The isoconversional analysis shows that thermal decomposition reactions of the samples
are not simple. K-6 has E values strongly dependent on values of . E decreases from an
initial value of ~245 kJ mol-1 to ~150 kJ mol-1 at =1. However, the variation of E with
is not so drastic for the PBXs. All the four PBX samples have slightly different values of E,
but the nature of variation in E with respect to is essentially the same. We can see from
Figure 2, that values of E are different for K-6 and its PBXs initially, but it is approximately
same above 0.75. Hence it is reasonable to conclude that the mechanism of thermolysis
at the initial stage of reaction is different in K-6 and its PBXs whereas all the samples follow
the same mechanism above 0.75.
324
Since there is an induction period and the thermolysis is taking place in the solid-state, it
may be possible that initial stages (induction period) are controlled by nucleation. The
acceleration period may be controlled by nuclei growth. The process of nucleation is
extremely sensitive to the prehistory of a solid. Thus, pure K-6 and K-6 in the PBXs may
have different prehistory and hence the difference in activation energy for nucleation. In the
molecular level there are many reaction pathways possible for K-6, as does RDX, which is a
closely related compound in structure. Ostmark et al.[4] studied the decomposition of K-6 and
RDX, using mass spectrometry (MS), laser induced mass spectrometry (LI/MS),
chemiluminescence and DSC. They have concluded that concerted symmetric fission of C-N
bonds is less important for K-6 than for RDX. They suggested NO2 elimination followed by
a break down of the ring as the probable mechanism of decomposition of K-6. For the
production of NO from K-6 by chemiluminescence in the temperature range 90-120 C,
activation energy of 140 kJ mol-1 and a frequency factor of 9 x 109 s-1 was obtained. In our
study, E was ~150 kJ mol-1 above > 0.75 for K-6 and the PBXs and hence final stages of
thermolysis may be limited by breaking of N-NO2 bond. Ostmark et al.[4] have calculated
bond breaking energies for N-NO2 and concerted triple fission (C-N) of RDX and K-6, using
modified neglect of diatomic overlap (MNDO) using PM3 parameter set at the unrestricted
Hartree-Fock (UHF) level of theory. They have reported Hbond1 = 26.4 kcal mol-1 (110.5 kJ
mol-1) and Hbond2 = 21.5 kcal mol-1 (90 kJ mol-1) as bond scission energies for the two types
of (non-equivalent) N-N bonds in K-6. The bond scission energy for concerted triple fission
was calculated to be Hbond = 67.1 kcal mol-1 (280 kJ mol-1). But the values calculated for
RDX, i.e. Hbond for N-N = 22 kcal mol-1 (92.1 kJ mol-1) and Hbond for concerted triple
fission = 68.2 kcal mol-1 (285.4 kJ mol-1) were quite different from the experimental values
(~ 200 kJ mol-1) for the former and 150 kJ mol-1 for the latter processes respectively. If the
calculated values of Ostmark et al.[4] are correct, then the initial stage of K-6 decomposition
may be controlled by concerted triple fission and as the reaction progresses, the contribution
from N-NO2 bond breaking increases. In the PBXs the contribution from concerted triple
fission is considerably less and the other process dominates prevalently at the final stages.
Moreover as stated previously dissolution of K-6 may take place in the binder matrix and
thus in the solution phase N-NO2 bond cleavage is promoted which has lower activation
energy. However, as there is an insufficiency of reliable data on kinetic parameters for K-6,
the interpretations in the present study must be treated as indicative and not exhaustive.
5.
CONCLUSIONS
Even thermally stable, inert polymers such as Viton A and Kel F affect thermal stability
of K-6 to some extent, during a dynamic heating programme, when they are used as binders
to make the PBXs. The reason for this seems to be the marginal solubility of K-6 in the
binder matrix. Isoconversional analysis shows that the mechanism of thermolysis changes as
the reaction proceeds, for both K-6 and the PBXs and it is different for the pure compound
and the PBXs. However, further studies are required to explain these changes.
325
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
R. GILARDI, J. L. FLIPPEN-ANDERSON AND C. GEORGE: Acta Crystallogr. Sect. C. 4, 706,

1990.
A. R. MITCHELL, P. F. PAGORIA, C. L. COON, E. S. JESSOP, J. F. POCO, C. M. TRAVER,
R. D. BREITHAUPT AND G. F. MOODY: Propellants, Explos., Pyrotech. 19, 232, 1994.
H. R. BLOMOUIST: U.S. Patent Appl. No. 203113, 1998.
H. OSTMARK, H. BERGMAN, K. EKVALL AND A. LANGLET: Thermochim. Acta 260, 201, 1995.
H. SINGH: in Propellants and Explosives Technology; (S. Krishnan, S. R. Chakravarthy and S.
K. Athithan.) Allied Publishers Ltd., Chennai, p245, 1998.
G. SINGH AND S. P. FELIX: Proc. 5th Semin. New Trends in Research of Energetic Materials
p. 68, 2002.
G. SINGH, S. P. FELIX AND P. SONI: Thermochim. Acta 2003 (In press).
G. SINGH, S. P. FELIX AND P. SONI: Thermochim. Acta 2003 (Communicated).
Table 1. TG-DTA Phenomenological Data on K-6 and the PBXs

Under Inert atmosphere
TG
Sample
Name
DTA
Ti
Ts
Tf /
Mass
Loss
Exothermic
Peak Temperature
Integral Normalized
Peak Area
[C]
[C]
[C]
[%]
[C]
[C 2 mg-1]
K-6
191
199
203
89.6
202
183.2
K-6V 9505
186
193
197
87.9
195
229.2
434
440
449
1.5
--
--
464
477
486
3.7
483
450.0
185
193
195
83.2
193
214.5
438
443
450
2.0
--
--
470
483
492
7.8
488
397.6
187
194
197
87.8
196
280.8
426
441
448
5.7
--
--
186
193
196
79.6
194
133.6
413
418
426
3.6
--
--
439
445
451
6.8
--
--
K-6V 9010
K-6K 9505
K-6K 9010
Ti Onset temperature, Ts inflection temperature, Tf endset temperature
326
Fig 1.
TG-DTA thermograms recorded under inert atmosphere at a heating rate

of 10 C min-1.
327
Fig 2.
Dependencies of activation energy (E) on extent of conversion ()

obtained from isoconversional method.
Fig 3.
DTA thermograms for K-6 under various conditions.
328
THERMOLYSIS AND KINETICS

OF SOME BIS(PROPYLENEDIAMINE) METAL NITRATE COMPLEXES
Gurdip Singh* and Durgesh Kumar Pandey
Chemistry Department, DDU Gorakhpur University,
Gorakhpur - 273 009, India
Abstract:
Five bis(propylenediamine)metal nitrate (BPMN) complexes like [M(pn)2](NO3)2 (where M
= Cr, Mn, Ni, Cu, Zn and pn = propylenediamine) have been prepared and characterized.
Thermal decomposition studies have been done using TG-DTA in an inert atmosphere of
nitrogen. TG and DTA studies have also been carried out separately in air atmosphere to
examine the effect of the atmosphere change on thermolysis trend of these complexes.
Several measurable changes were recorded on decomposition mode of all these complexes
when atmosphere was changed. However, as indicated by thermoanalytical techniques,
thermal stability of all these complexes decreases in the order
[Cr(pn)2](NO3)2 >[Mn(pn)2](NO3)2 >[Zn(pn)2](NO3)2 >[Ni(pn)2](NO3)2 > [Cu(pn)2](NO3)2
Isothermal TG, over the temperature range of deamination (a step corresponding to the loss
of amine ligand) of all the complexes, has been done. An analysis for kinetics of
deamination of these complexes was made on the basis of isoconversional method.
Keyword:
1.
Metal complex, TG, DTA, kinetics, isoconversional method
INTRODUCTION
Propylenediamine is a primary amine, which acts as a bidentate ligand in bonding to metal

ions. It has two similar types of neutral donors i.e. nitrogen atoms of the amine groups. This
compound plays an important role as reducing agent in complexes when complexed
simultaneously with a metal ion and oxidizing groups like nitrate (NO3). The resulting
compound exhibits characteristics of high energetic one, which undergoes highly exothermic
decomposition reaction due to simultaneous presence of both oxidizing and reducing groups in
the same complex molecule. Such types of energetic compounds may find applications in
explosives, propellants and pyrotechnic compositions. The explosivety of such type of energetic
compounds is reported to lie between those of primary and secondary explosives[1,2]. Such
transition metal complexes are also used as precursors to obtain simple ultrafine metal oxides,
which may have interesting electrical, magnetic and catalytic properties[3]. Thermal analysis is
proved to be an indispensable analytical tool for the research and development of novel
energetic materials. Recently we have undertaken studies on the thermolysis and kinetics of
some hexammine metal perchlorates[4], bis(ethylenediamine)metal nitrates[5] and
bis(diethylenetriamine)metal nitrates[6]. In continuation to these investigations, we are reporting
here the preparation, characterization and decomposition results on bis(propylenediamine)metal
nitrate complexes. Kinetics of thermolysis has been evaluated applying isoconversional method,
which is a novel attempt for kinetic analysis of these complexes.
329
2.
EXPERIMENTAL
2.1
Materials
Following AR grade commercially available chemicals (obtained from the sources given in
parentheses) were used as received without any further purification.
Carbonates of copper, chromium, manganese (BDH), nickel, zinc (Thomas Baker), 70%
nitric acid, methanol (Ranbaxy), ethanol (Hayman), propylenediamine, silica gel, iodine (s.d.
fine chemicals) and petroleum ether (Merck).
Firstly, the metal carbonates were treated with 70% nitric acid to obtain the respective metal
nitrates. The metal nitrates thus obtained were washed with petroleum ether, recrystallized from
distilled water and dried over fused CaCl2 in a desiccator. BPMN complexes were prepared by
reacting ethanolic solution of various metal nitrates with propylenediamine following the
standard procedures reported in the literature[7,8]. The precipitated complexes of different
colours were separated from mother liquor by decantation and washed with methanol. All of
these nitrates were recrystallized from aqueous solution, dried over fused CaCl2 in a desiccator
and their purity was assessed by thin layer chromatography (TLC).
2.2
Thermal Decomposition and Impact Studies
All the non-isothermal analyses were undertaken at a heating rate of 10oCmin-1. The
simultaneous TG-DTA thermograms on the complexes (weighing 1-2 mg) were obtained on
Mettler Toledo Star system applying inert atmosphere (flowing N2 at a rate of 100 mLmin-1).
Non-isothermal TG thermograms were recorded on our indigenously fabricated TGA
apparatus[9] in static air (sample mass 20 mg, 100-200 mesh). Non-isothermal DTA
thermograms were taken under flowing air atmosphere (100 mLmin-1) by making use of an
instrument supplied by Universal Thermal Analysis Instruments, Mumbai (sample mass 10
mg). Isothermal TG on the complexes (weighing 20 mg) was done in static air using above
said indigenously fabricated TGA apparatus[9] at appropriate temperatures of decomposition.
3.
RESULTS
3.1
Characterization
The complexes were characterized by gravimetric methods[10], infrared[11,12] and elemental

analysis (Fission Instruments DP 200 series 2). The deduced molecular formulae of the
complexes and their physical and spectral data are reported in Table 1. The present complexes
exhibits 1:2 (metal:ligand) stoichiometry and these all are anhydrous.
3.2
TG-DTA
The simultaneous TG-DTA thermograms are demonstrated in Figure 1 and the

corresponding phenomenological data are listed in Table 2. It is evident that all these complexes
(except copper complex) undergo decomposition in multistep. For the copper complex, which
shows an exceptional decomposition pattern of this series of complexes, single step
decomposition was observed. Thermolysis of these complexes broadly involves deamination,
deanionation and oxidation of metals to respective metal oxides. The first weight loss in TGA is
330
due to deamination and second corresponds to simultaneous deanionation and oxidation of

metal. Thus no separate weight loss step was observed for the last reaction i.e. oxidation. The
complex of chromium metal shows weight loss about 62% (Table 2) in the first step, which is in
good agreement (65%) with the percent losses of two molecules of amine ligand along with one
molecule of counter anion (NO3-). In the next step, the loss of second molecule of counter anion
takes place. This step is also consistent with oxidation of chromium to chromium oxide, which
is an ultimate decomposition product. The deamination of manganese complex takes place in
two steps and complete decomposition of the complex involves three steps. First step
corresponds to removal of one molecule of ligand, which is immediately followed by second
step occurring due to loss of second molecule of amine ligand. Finally manganese nitrate,
formed during the course of thermolysis of parent complex, undergoes simultaneous
decomposition and oxidation processes leaving behind manganese oxide as decomposition
product. In the first step of decomposition of both nickel and zinc complexes, two ligand
molecules are expelled out from the respective complex molecule and corresponding metal
nitrates are formed that decompose in final step to their metal oxides. However, in the case of
copper complex thermolysis involves single and rapid step that corresponds to simultaneous
losses of ligand molecules, counter anion and oxidation of copper to copper oxide.
Both the above said processes viz. deamination and deanionation are accompanied by
derivatograms showing the thermal effect of respective decomposition processes. All the steps
cause exoeffect in DTA. The sharp nature of first exothermic peak in all the cases is indicative
of quite steep deamination process occurring within a few degrees of temperatures. The last step
of decomposition is found to be very broad in nature resulting broad exotherm. A hump at about
210oC can be seen in the exothermic peak corresponding to the thermolysis of copper complex
(Figure 1). This hump indicates that, though the decomposition occurs in single step, at least two
competitive processes are occurring over a very short interval of temperature.
3.3
Non-isothermal TG and DTA in Air Atmosphere
To record the effect of atmospheric change on thermolysis trend of the complexes, TG and
DTA techniques were used separately in static air and flowing air atmosphere respectively. Nonisothermal TG thermograms are shown in Figure 2 and corresponding data profiles are collected
in Table 3. Some measurable changes in decomposition pattern of the complexes were recorded
during TG analysis in air. The complexes of manganese, nickel and zinc essentially decompose
in two steps whilst those of chromium and copper in only one-step. One step of mass loss each
of chromium and manganese complexes was found to be disappeared (Figure 2, Table 3) when
atmosphere was changed from nitrogen to air. Derivatograms recorded in air are presented in
Figure 3 and data are given in Table 3. Multistep thermolysis of all these complexes can be seen
here. All the steps were found to be exothermic in nature.
3.4
Isothermal TG
Isothermal TG thermograms were obtained over a short temperature range. Thermal

decomposition in chosen step for analysis i.e. deamination is quite faster at higher temperatures
and slow at lower temperatures. Therefore, an optimum temperature range was selected for
analysis.
331
3.5
Kinetic Analysis of Isothermal TG Data
Kinetic analysis of isothermal TG data has been made by using isoconversional method as
described in our earlier paper[13]. The dependencies of activation energy (E) on extent of
conversion () were determined and presented in Figure 4
4.
DISCUSSION
As can be seen from analytical data listed in Table 1, the estimated values of percentage of
each element are quite comparable to those of calculated values. This result confirms the
formation of BPMN complexes, beyond doubt. In addition, the values of characteristic
absorption frequencies of different groups, present in the complexes are in good agreement with
those reported in standard texts[11,12]. It is obvious from TG-DTA thermograms (Figure 1) and
corresponding data (Table 2) that the complexes of chromium, nickel and zinc essentially show
two weight loss regions in the temperature range studied. First step is due to removal of ligand
(pn) and second corresponds to expulsion of counter anion (NO3). The observed percentage
weight losses are in agreement with the calculated values. Deamination of manganese complex
takes place in two steps which may be due to higher thermal stability of
mono(propylenediamine)manganese nitrate formed during thermolysis of pure complex and
thus stage separation of deamination process is observed. In the case of copper complex, single
step decomposition was observed which is attributable to a chain type highly explosive reaction
because of simultaneous presence of an oxidizing group (NO3) and a reducing group (pn) in the
same molecule. Although, the same ligand (pn) and counter anion (NO3) are present in all the
complexes, the explosive oxidation of copper complex seems to be due to better activity of
copper ion as oxidation catalyst in comparison to other metal ions at experimental conditions[14].
In all these amine metal complexes, except that of copper, respective metal nitrates are formed
as intermediates that are stable to a considerable temperature range. As described in results,
deamination of all these complexes is quite steep and final step of decomposition is spread over
a wide range of temperatures (335-463oC) in TG followed by a broad exotherm in DTA. Such
features of thermograms is attributable to simultaneous occurrence of two processes in the final
step i.e. decomposition of respective metal nitrates and oxidation of metals to metal oxides.
Thus, the corresponding metal oxides are final products of thermal decomposition of all the
complexes.
Atmospheric change (from N2 to air) shows remarkable change on decomposition mode of
the complexes. A comparison of the TG data for various complexes from Table 2 to that of
Table 3 clearly indicates the following differences i.e. firstly a comparatively high SDT values
of TG in air than those of Ti values of TG in nitrogen and secondly a considerable lower values
of FDT (Table 3) than those of Tf values (Table 2). Thus, it can be inferred that, though the mass
losses in TG (air) of various complexes start at comparatively high temperatures, but when once
the decomposition reactions start, achieve the completion at considerable lower temperatures
under air than that in nitrogen atmosphere. This change in decomposition temperatures seems to
be due to oxidative nature of the atmosphere (air), which facilitates oxidative decomposition.
Moreover, one step of mass loss each of chromium and manganese complexes has not been
observed experimentally and these were found to decompose in single and double step
respectively in air. This changed mode of thermolysis is also due to change in atmosphere from
332
nitrogen to air. Thus, the thermal stability of the complexes as indicated by TG decreases in the
order
[Cr(pn)2](NO3)2 >[Mn(pn)2](NO3)2
[Cu(pn)2](NO3)2
> [Zn(pn)2](NO3)2
> [Ni(pn)2](NO3)2
>
It was recorded that exothermic peak temperatures in DTA decreases in the same order.
Trustworthy, kinetic parameters can only be extracted in a way that is independent of the
reaction model. Isoconversional method is known to allow for model independent estimation of
activation energy. There are several isoconversional methods reported in literature we have
adopted the method reported by Vyazovkin[15]. This approach indicates that the decomposition
of these complexes is not simple as indicated by model fitting approach. Initial activation energy
values of all these complexes are found to be higher (Figure 5) that are decreasing gradually in
the range 0.5 to 1. This nature of graph may be due to two ligand molecules expelled
simultaneously during deamination, which are actually associated with two different values of
E. A decreasing value of E on increasing is indicative of an increased rate of deamination for
second ligand molecule than that of first one. These features are also attributable to an induction
period (at lower values) and an acceleration period (at higher values) that are controlled by
nucleation and nuclei growth respectively. Similar a common and general mechanism has been
found to operate during thermolysis throughout this series of complexes.
5.
CONCLUSIONS
The present TG and DTA studies under nitrogen and air atmosphere thus enable to
understand the nature of thermal decomposition pattern. Thermolysis involves evolution of
amine ligand in the first step and corresponding metal nitrates, formed as intermediates, were
found to decompose at considerable high temperatures, leaving behind respective metal oxides
as product of decomposition. The entire thermolysis step exhibit exoeffect in corresponding
DTA. The rate of decomposition is relatively high in air as compared to that in nitrogen. The
order of thermal stability of the complexes was proposed in which complex of copper is at lower
stability end and that of chromium at higher stability end. Kinetic parameters were obtained by
the use of standard isoconversional method and though a single step deamination of the
complexes, E varies as the degree of conversion () increases. Nucleation followed by nucleus
growth seems to operate as general mechanism throughout thermolysis of this series of
complexes.
333
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
W. FRIENDRICH AND P. VERVOOST: C.A., 21,1184, 1924.

K.C. PATIL AND V.R. PAI VERNEKER: Combust Flame., 25, 387, 1975.
S.Y. SAWANT, K.R. KANNAN AND V.M.S. VERENKAR: Thirteenth Symposium (National) on
Thermal Analysis, BARC, Mumbai, 2002.
G. SINGH AND D.K. PANDEY: J. Energ. Mater., 20,135, 2002.
G. SINGH AND D.K. PANDEY: Prop. Explos. Pyro., In Press, 2002 (MS No. 2001/060).
G. SINGH AND D.K. PANDEY: J. Therm. Anal., Communicated, 2002 (MS No. JTA/MSS/578).
H. NAKAI AND Y. DEGUCHI: Bull. Chem. Soc. Jpn., 48, 2557, 1975.
S. PAL AND L. JANSONNE: Magy. Kem.Foly., 81,70, 1975; CA 83,21077j.
G. SINGH AND R.R. SINGH: Res. Ind., 23,92, 1978.
A.I. VOGAL: Text Book of Quantitative Inorganic Analysis (J. Bassett, R.C. Denny, G.H. Jeffery
and J. Mendham Eds.) Longman, London, 1985, 4th Ed.
K. NAKAMOTO: Infrared and Raman Spectra of Inorganic and Coordination Compounds Wiley
and Sons, New York, 1978, p. 206.
C.W. ROBERT: Handbook of Physics and Chemistry CRC Press, Florida, vol. 66, 1996, p. 197.
G. SINGH AND D.K. PANDEY: Thermochimica Acta 2002 (Communicated).
G. SINGH, I.P.S. KAPOOR, S.K. TIWARI, AND S.P. FELIX: Indian J. Engg. Mater. Sci., 7,167,
2000.
S. VYAZOVKIN: Thermochim. Acta., 355, 155, 2000.
334
Table 1. Physical, elemental and spectral parameters of BPMN complexes

Complex
Colour
Element %/observed(calculated)
TLC*
Eluent
Rf
2a:b:c
0.79
[Mn(pn)2](NO3)2
Light
green
Brown
2a:b:c
0.68
[Ni(pn)2](NO3)2
Violet
2a:b:c
0.70
[Cu(pn)2](NO3)2
Dark
blue
White
2a:b:c
0.80
2a:b:c
0.75
[Cr(pn)2](NO3)2
[Zn(pn)2](NO3)2
IR**
Metal
pn
M-N
H2N-CH2 NO3
21.8
(22.3)
21.5
(22.1)
21.2
(21.8)
20.6
(21.5)
20.5
(21.4)
5.7
(6.2)
5.7
(6.1)
5.8
(6.0)
5.7
(6.0)
5.3
(5.9)
25.4
(26.0)
25.5
(25.8)
24.5
(25.3)
24.2
(25.0)
23.4
(24.9)
14.8
(15.7)
15.9
(16.5)
16.9
(17.8)
18.0
(18.9)
18.7
(19.3)
1098m
430s
1055m
1365s
1460m
480s
1060m
445s
1058m
1358s
1478m
460s
1085m
438s
1108m
1400s
1464m
449s
1110m
432s
1070m
1376s
1482m
455s
1090m
440s
1045m
1388s
1480m
458s
CH2
NH2
*a = n-butanol, b = chloroform, c = gl. acetic acid, Rf = Retention factor, Locating reagent = iodine
**m = medium, s = sharp
335
Table 2.
TG-DTA Phenomenological Data on BPMN Complexes

under Inert Atmosphere
Complex
TG
[Cr(pn)2](NO3)2
[Mn(pn)2](NO3)2
[Ni(pn)2](NO3)2
[Cu(pn)2](NO3)2
[Zn(pn)2](NO3)2
DTA
Ti/oC
Ts/oC
Tf/oC
%Mass
loss
Exothermic peak
temperature/oC
280
340
262
297
395
230
335
220
230
355
318
360
276
318
420
251
380
226
270
380
337
379
295
337
463
290
415
229
290
405
62
78
21
44
77
43
76
76
43
75
321
370
282
315
424
254
388
227
274
384
Integral peak
area/mgsoC
2051
2667
142
202
2630
2110
2670
112
215
1220
Ti = onset temperature, Ts = inflection temperature, Tf = endset temperature
Table 3. TG and DTA Phenomenological Data of BPMN Complexes

in Air Atmosphere
Complex
SDT/oC
[Cr(pn)2](NO3)2
[Mn(pn)2](NO3)2
[Ni(pn)2](NO3)2
[Cu(pn)2](NO3)2
[Zn(pn)2](NO3)2
DTA (Exothermic
peak temperature)/oC
TG
305
FDT/oC
345
%Mass
loss
78
248
335
292
325
44
298
335
357
75
375
250
267
45
259
285
330
76
365
225
245
75
200
225
250
270
44
250
290
320
74
357
SDT = starting decomposition temperature,

FDT = final decomposition temperature
336
Fig 1.
TG-DTA thermograms of BPMN complexes

under nitrogen atmosphere.
Fig 2.
Non-isothermal TG thermograms of complexes recorded

in air atmosphere
337
Fig 3.
DTA thermograms obtained in air atmosphere.
Fig 4.
Dependencies of activation energy on extent of conversion

for these complexes (each atomic symbol with a marker represents
the graph of respective metal complex).
338
HOPKINSON BAR STUDIES ON POLYMER BONDED EXPLOSIVES

C. R. Siviour, S.M. Walley, W.G. Proud and J.E. Field
PCS Group, Cavendish Laboratory,
Cambridge CB3 0HE, UK
Abstract:
This paper describes the Split Hopkinson Pressure Bar (SHPB) method for obtaining
mechanical properties at high strain rate, and its applicability to Polymer Bonded
Explosives (PBXs). The advantages and abilities of the technique are discussed, as are
the difficulties of testing soft materials. Recent results are presented to demonstrate
typical properties of PBXs in these experiments.
Keyword:
1.
polymer bonded explosive, Hopkinson bar, mechanical testing
INTRODUCTION
Polymer bonded explosives (PBXs) consist of explosive crystals dispersed in a polymer

matrix, the binder. Typically the percentage by mass of crystals is high, up to 95%, however
the small amount of binder has a very large effect on the mechanical properties, because its
modulus is many orders of magnitude lower than that of the crystals [1]. The overall material
is easier to handle than a pure explosive, for example, it is possible to cast the PBX into
moulds, or to machine it. PBXs also have a reduced sensitivity compared to the pure
explosive. Initiation takes place due to hot spot formation, from conversion of mechanical
energy into heat by friction or fracture [2]. The PBX is more ductile than a pure explosive
crystal, so the system can absorb mechanical energy by deformation without initiation.
It has long been known that the determination of the mechanical properties of explosive
composites is of importance. These properties have a number of uses in, for example,
validation of constitutive models, finding parameters for finite element modeling and as data
for safety assessment.
The sensitivity of the PBX is strongly dependent on its mechanical properties. For
example it has been found during drop-weight experiments that a PBX below the glass
transition temperature of its binder is much more sensitive than that above. As would be
expected, soft PBXs are less sensitive than hard PBXs. It will be seen later than the
properties of PBXs are strong functions of particle size and temperature.
1.1
The Split Hopkinson bar
The Split Hopkinson Pressure bar is a device for determining the mechanical properties
of materials at strain rates from 500 to 10,000 s 1. A schematic diagram of the bar is shown
in figure 1. The specimen is a right cylinder of material, which is placed between the input
and output bars. A striker bar is propelled into the input bar, introducing an elastic stress
wave. This wave travels down the input bar to the specimen, which deforms under the
stress. Some of the wave is reflected back down the input bar while the rest is transmitted.
Effectively the bars act as nearly-one-dimensional wave guides. The magnitudes of the
waves in the two bars are measured using strain gauges. The three waves of interest are
denoted incident, reflected and transmitted, figure 2. Because the bars are not truly one
dimensional there is some dispersion of the initially rectangular pulse, which introduces
339
Pochammer-Chree oscillations. The voltage traces from the strain gauges, which are
normally incorporated in some sort of bridge circuit, are converted into force, and the force
pulses used as follows;
The transmitted force gives directly the force on the output end of the specimen
The reflected wave gives a representation of the movement of the end of the input
bar.
The sum of the incident and reflected waves gives directly the force on the input end
of the specimen.
Striker Bar
Input Bar
Input Gauges
Gas Gun
Fig 1.
Sample
Output Bar
Output Gauges
Momentum Trap
Schematic diagram of the Split Hopkinson Pressure Bar (SHPB)
0.3
Output Voltage / V
0.2
Incident pulse
0.1
0
-0.1
Input gauge
Tansmitted gauge
-0.2
-0.3
Fig 2.
Transmitted pulse
Reflected pulse
0
50
100
Time / s
150
200
Wave outputs from the SHPB, Pochammer-Chree oscillations can be seen at

the start of the incident and reflected pulses.
340
140
True stress / MPa
120
100
80
60
40
20
0
Fig 3.
0.1
0.2 0.3
True strain
0.4
0.5
A stress strain curve for polycarbonate,

at a strain rate of 4500 s-1
The reflected and transmitted waves can be used to construct a stress-strain curve for the
specimen. An excellent derivation of the equations required is given by Gray III [3]. An
example of a typical stress strain curve is shown in figure 3. True stress and true strain are
defined, taking compressive strains as positive, as follows;
True stress =
Force supported
Current area of sample
Original length of sample
Current length of sample
True strain = ln
(1)
(2)
The advantages of these measures over the more traditional engineering stress and strain
are that they give material properties, rather than specimen specific properties, and the strain
is additive over repeated loading cycles. However, calculation of these properties does
assume that it is possible to calculate the changing area of the sample from its changing
length. This means that the sample must conserve volume, the standard analysis cannot be
used for foams, or brittle materials. In these cases engineering stress may be used, or the
sample area can be determined from other measurements, such as photographs from highspeed cameras.
In order for the results to give a meaningful representation of the specimen properties it
is necessary that the strain rate of the specimen be constant and the specimen in equilibrium.
This means that the mechanical state of the specimen (i.e. the stress and the strain) is the
same throughout its length. In practice equilibrium is confirmed by comparing the
transmitted force pulse to the sum of the incident and reflected force pulses, figure 4. This
compares the forces on the front and back surfaces of the specimen, which are the same if
the specimen is in equilibrium. Because the stress wave has a finite speed it takes time for
this to occur, and hence the initial results from the bars are functions of the specimen plus
bar. This means that it is not possible to measure elastic material properties using the
SHPB [4].
341
5
Incident + Reflected
Transmitted
Force / kN
4
3
2
1
0
Fig 4.
1.2
40
60
Time / s
80
100
Testing equilibrium in a polycarbonate specimen
Advantages of the Hopkinson Bar system
1.3
All instrumentation is on the bars, which remain elastic throughout the test,
unlike the specimen. Gauges can be accurately calibrated, and then used for a
large number of experiments.
Specimen preparation is simple, only small specimens are required, which is an
important safety issue for explosives testing.
There are few consumables, so testing is cheap.
Disadvantages
1.4
20
It is not possible to measure elastic properties of the specimen.

Equilibrium must always be checked [5].
Bar Materials
The bars can be made from rods of metal, or polymer. The choice of material used
depends on the specimen properties. The bars must be hard enough not to yield at the point
where they exert stress on the specimen. Further discussion of the bar properties continues
below. The two materials used in these tests were Inconel, and magnesium. Relevant
properties of these materials are given in table 1.
Table 1. Mechanical properties of Inconel and magnesium
Material
Sound speed (m s-1)
Density (kg m-3)
Impedance (kg m-2 s-1)
Magnesium
4920
1800
8.8 x 106
Inconel
4980
8270
41 x 106
342
1.5
Temperature
As well as testing at different strain rates, specimens can be tested at different

temperatures. This can be done in a number of ways. A common method is to enclose the
specimen and the ends of the bars in a chamber, and then to heat or cool the chamber to the
required temperature by introducing hot or cold helium. The chamber temperature can be
measured using a thermocouple, allowing time for the specimen to reach the same
temperature as the chamber. Usually Hopkinson bars made of Inconel are used for this
purpose, because the mechanical properties of Inconel are a relatively weak function of
temperature [6]. It is also possible to measure the changing temperature of the specimen
throughout the tests, either using a thermocouple embedded in the specimen, or a thermal
imaging camera [7].
1.6
Lubrication
Friction between the bars and the specimen is an important consideration. It will tend to
increase the force required to deform the specimen, and also to cause barrelling, where the
radius of the specimen increases more in the centre than at the ends, figure 5. This can be
prevented for polymer specimens by using paraffin wax lubricant [8,9]: Molybdenum
disulphide is also commonly used [10].
(a)
Fig 5.
1.7
(b)
An example of barrelling in quasi-static testing of an epoxy,
(a) before testing, and (b) during testing [11].
Specimen considerations
The specimen is a right cylinder, large enough that it is representative of the bulk
medium. Ideally this means that each of the linear dimensions is about 10 times larger then
the linear dimension of the explosive grains, so that the specimen contains about 1000
crystals [12].
The specimen length is determined by the strain rate desired, and the requirement to
reach equilibrium within the timescale of the loading. Longer specimens take longer to
reach equilibrium. The initial strain rate is approximately equal to the striker speed divided
by the specimen length. Depending on the range of the gas gun used large changes in
desired strain rate often require the specimen length to be changed.
The specimen diameter is determined by the strain rate, and also by the bar diameter. It
is necessary that at full strain the diameter of the specimen does not exceed that of the bars.
A large diameter specimen is also more difficult to deform, due to its radial inertia [13, 14].
The added stress due to this can be approximately calculated, so that a specimen geometry
can be chosen whose inertia is not significant during the deformation. This stress is
dependent on the strain rate, specimen length, specimen diameter, and specimen density, as
343
shown in equation 3, where l represents the specimen length and r the specimen radius.
Smaller specimens can be tested to higher strain rates.
l 2 r2 2 l 2 r 2
i = + +
16 6
6 8
1.8
(3)
Soft specimens
Testing soft specimens in Hopkinson bars is complicated by two problems. The first is
the low sound speed. This means that they take longer to reach equilibrium. Secondly, they
have low mechanical impedances, giving a large impedance mismatch between the specimen
and the bars. This results in a large reflected wave, and small transmitted wave. Whilst it is
possible to increase the transmitted wave by increasing the specimen diameter, this is limited
by the diameter of the Hopkinson bars themselves, and the strain rate required. To get round
this problem low impedance metal, or even polymer, bars can be used [15, 16, 17]. By matching
the impedance of the bar more closely to that of the specimen the magnitude of the
transmitted wave is increased. There is a further advantage that low impedance materials
also have low elastic moduli, so there is more strain in the bar for a given force pulse, which
makes the strain gauge measurements more accurate. The disadvantages of these bars are as
follows:
Low impedance bars are usually more dispersive. In low impedance metal bars initially
square stress pulses develop very large Pochammer-Chree oscillations. The dispersion in
visco-elastic (polymer) bars is very large, and complex analysis is needed to correct for it
[18,19, 20]
.
It takes longer for a specimen to reach equilibrium [21].
The mechanical properties of the bars change significantly with temperature.
The problem of equilibrium is intimately related to bar selection. The small output pulse
when testing soft specimens means that it is very difficult to check equilibrium in the normal
way, by looking at the sum of the pulses, since the difference between the incident and
reflected pulses will be very small compared to the value of these pulses.
The results presented later in this report address these issues directly. In order to check
specimen equilibrium PVDF stress gauges were placed between the ends of the bars and the
specimen.
2.
EXPERIMENTS AND RESULTS

2.1
HTPB
Figure 6 shows a stress-strain curve obtained for pure HPTB at 5000 s-1. As expected
for a rubbery polymer there is no distinct yield point. Because the transmitted signal is very
low compared to the incident signal, figure 7, the traditional method of checking equilibrium
could not be used. Instead, PVDF stress gauges were used to measure the stress on the front
and back faces of the specimen, this is shown in figure 8. The sample reached equilibrium
approximately 10 s after the loading started. It is generally the case that softer materials
have lower sound speeds, since c = E , and the yield stress is related to the modulus, E.
Since all the polymers discussed later in this paper are stiffer than pure HTPB we can
assume that they will reach equilibrium in about 10s.
344
0.6
0.4
Voltage output / V
True stress / MPa
2.5
1.5
1
0.5
0
0
Gauge output / arbitrary units
Fig 6.
0.1
0.2
0.3
True strain
0.4
Stress-strain curve for HTPB at

5000 s-1, showing average (solid
line) and standard deviation
(dashed lines) from 6 tests.
Input gauge
Transmitted gauge
0.2
0
-0.2
-0.4
-0.6
0
50
100
150
Time / s
200
Fig 7.
Voltage signals for HTPB

in magnesium bars.
Fig 8.
Traces from PVDF gauges

on the front and rear faces of
an HTPB specimen.
0.1
Input face
Output face
0.08
0.06
0.04
0.02
0
2.2
20
40
60
Time / s
80
100
PBXs
The following section presents results from experiments on two PBXs: the first an
HTPB-based PBX available in four different grain sizes (termed PBX 1). The second PBX
will be termed PBX 2.
Figure 9 shows stress strain curves for PBX 1 with crystals of 30m diameter.
A comparison of data for four different grain size materials is shown in figure 10. These
curves are averages over a number of experiments on each grain size. At this temperature
HTPB is a rubbery material, and in common with other rubbers does not yield. At low
temperatures the PBX shows a distinct yield, and there is also more separation between the
stress strain curves for PBXs with different crystal sizes, figure 11. The effect of raising the
temperature is to weaken the polymer. This makes the properties more difficult to measure,
345
because the transmitted signal is smaller. This problem is made even worse by the
requirement to use Inconel bars for the high temperature tests. Figure 12 shows stress strain
curves for PBX 2 at room temperature and also at +70 C; the material at room temperature
yielded. It is also possible that the PBX will show brittle behaviour. This happens for PBX 2
at -80 C, figure 13.
10
100
True stress / MPa
True stress / MPa
8
6
4
2
0
Fig 9.
0.1
0.2 0.3 0.4

True strain
0.5
PBX 1 with 30m diameter

crystals at around 4000 s-1
40
3 m
8 m
30 m
250 m
20
0
0
0.05
0.1 0.15 0.2

True strain
0.25
Fig 11. PBX 1 at -60C, 4000 s-1

6
3 m
8 m
30 m
250 m
5
True stress / MPa
8
True stress / MPa
60
0.6
10
6
4
2
0
80
0.05
0.1 0.15 0.2

True strain
0.25
Fig 10. PBX 1 with different grain sizes

at around 4000 s-1
4
3
2
70 deg C
25 deg C
1
0
0
0.1
0.2 0.3
True strain
0.4
0.5
Fig 12. PBX 2 at two different

temperatures, at 5000 s-1
346
Engineering stress / MPa
50
40
30
20
10
0
100
3.
150
200
Time / s
250
Fig 13. PBX 2 at 80 C and 5000 s-1
DISCUSSION
The Hopkinson bar allows the response of PBXs to an applied stress to be measured
over a range of temperatures and strain rates. In the regions observed the PBXs show a
number of different behaviours.
The results from placing PVDF gauges between an HTPB specimen and the bars showed
that the specimens reach equilibrium during the test. Results from experiments on PBXs in
magnesium bars demonstrated that PBX testing in these bars is possible with a good
transmitted signal. Experiments at low temperatures in Inconel were also successful: the
polymer is able to support large stresses at low temperature. Unfortunately at high
temperature this is not the case, the output signal is very low, because the high impedance
Inconel bars were used to test a much softer specimen.
HTPB is a rubbery polymer, which displays hyperelastic behaviour over the range of the
tests. When crystals are added at room temperature this behaviour may remain elastic, or
yielding-type behaviour may result. This is not due to yielding of the polymer, but rather the
added crystals. The material yields because of debonding of the crystals from the polymer
matrix [22 24]. This debonding has been observed in specimens deformed in low strain rate
tests, and then examined in the ESEM. The grain size effect on strength was fist noticed in
drop-weight testing by Field, and is again seen here [26].
The effect of temperature on the properties of the PBX is dramatic, and is similar to that
oserved in other polymer materials. Increasing the temperature reduces the strength of the
material, whilst reducing it makes it stronger. Reducing the temperature can also make the
polymer brittle. These are important considerations for the use of PBXs in low temperature
environments, because increasing the strength of a PBX makes it more sensitive.
At low temperatures the previously elastic PBX exhibits yielding behaviour. The flow
stress of the material decreases with increasing particle size, similar to the behaviour of
metals. However, this is due to a different mechanism from that in metals. In metals the
grain boundaries inhibit the movement of dislocations, in PBXs the crystal surfaces allow
deformation to occur.
347
4.
CONCLUSION
The Hopkinson bar technique is able to determine mechanical properties of PBXs over a
range of strain rates and temperatures. It is able to distinguish both qualitative, and
quantitative, differences in material behaviour. By suitable choice of a low impedence
Hopkinson bar material very soft PBXs can be tested (transmitting stresses of a few MPa),
although this choice may not be available for all temperature ranges. Rubber samples in the
Hopkinson bar do reach equilibrium.
PBXs can display a wide range of high strain rate properties. The general behaviour is
that of the binder material, but this is modified by the presence of the explosive crystals. For
example the non-yielding rubber can become a yielding PBX. The properties also depend on
the crystal size, smaller crystals give a stronger material.
Hopkinson bar research on PBXs is ongoing. As well as characterising new materials,
continuing work on the effect of grain size, loading density and binder properties will
hopefully allow development of complete constitutive models of these interesting materials.
Acknowledgements
The authors would like to thank QinetiQ, and in particular Dr D Mullenger and Dr P
Church for their continuing support of the PCS group, and the PBX work presented. We
would also like to thank Professor R Armstrong for his support of PBX research in the
Cavendish. The help of R Murrah with the Hopkinson Bar is greatly appreciated, as is the
expertise of P Flaxman with cryogenic, and high pressure, gases. S Palmer is thanked for
his assistance with samples of HTPB, and Dr M Gifford in similar regard to the PBX.
C Siviour would also like to thank EPSRC and the Worshipful Company of
Leathersellers for their financial support.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
S.M. WALLEY, J.E. FIELD, and S.J.P. PALMER. Impact sensitivity of propellants, Proc. R.
Soc. Lond. A Vol 438, p 571, 1992.
J.E. FIELD, N.K. BOURNE, S.J.P. PALMER, and S.M. WALLEY. Hot-spot ignition
mechanisms for explosives and propellants, Phil. Trans. R. Soc. Lond. A Vol 339, p 269, 1992.
G.T. GRAY III, in "ASM Handbook. Vol. 8: Mechanical Testing and Evaluation", (ASM
International, Materials Park, Ohio,2000) p. 462.
G. GARY, in "New Experimental Methods in Material Dynamics and Impact", (Institute of
Fundamental Technological Research, Warsaw, Poland,2001) p. 179.
G.T. GRAY III and W.R. BLUMENTHAL, in "ASM Handbook. Vol. 8: Mechanical Testing
and Evaluation", (ASM International, Materials Park, Ohio,2000) p. 488.
R. KANDASAMY and N.S. BRAR, in "High Pressure Science and Technology 1993",
(American Institute of Physics, New York,1994) p. 1031.
S.M. WALLEY, W.G. PROUD, P.J. RAE and J.E. FIELD. Comparison of two methods of
measuring the rapid temoerature rise in split Hopkinson bar specimens, Rev. Sci. Instrum.
Vol. 7 No. 4 p 1766, 2000.
S.M. WALLEY, J.E. FIELD, P.H. POPE, and N.A. SAFFORD. The rapid deformation
behaviour of various polymers, J. Phys. III France Vol 1, p 1889, 1991.
348
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
B.J. BRISCOE and R.W. NOSKER. The influence of interfacial friction on the deformation of
high density polyethylene in a split Hopkinson pressure bar, Wear Vol 95, p 241, 1984.
D.J. IDAR, D.G. THOMPSON, G.T. GRAY III, W.R. BLUMENTHAL, C.M. CADY, P.D.
PETERSON, E.L. ROEMER, W.J. WRIGHT, and B.L. JACQUEZ, in "Shock Compression of
Condensed Matter - 2001", (American Institute of Physics, Melville, NY,2002) p. 821.
A.D. BERNATSKII and A.L. RABINOVICH. Strain in certain crosslinked polymers, Polym.
Sci. USSR Vol 6, p 1166, 1964.
R.W. ARMSTRONG. Strength and ductility of metals, Trans. Indian Inst. Metals Vol 50, p
521, 1997.
D.A. GORHAM. Specimen inertia in high strain-rate compression, J. Phys. D: Appl. Phys. Vol
22, p 1888, 1989.
D.A. GORHAM. The effect of specimen dimensions on high strain rate compression
measurements of copper, J. Phys. D: Appl. Phys. Vol 24, p 1489, 1991.
G. GARY, J.R. KLEPACZKO, and H. ZHAO. Generalization of split Hopkinson bar technique
to use viscoelastic materials, Int. J. Impact Engng Vol 16, p 529, 1995.
G.T. GRAY III, W.R. BLUMENTHAL, C.P. TRUJILLO, and R.W. CARPENTER II.
Influence of temperature and strain rate on the mechanical behavior of Adiprene L-100, J.
Phys. IV France Vol 7 Colloq. C3 (EURODYMAT 97), p 523, 1997.
H. ZHAO, G. GARY, and J.R. KLEPACZKO. On the use of a viscoelastic split Hopkinson
pressure bar, Int. J. Impact Engng Vol 19, p 319, 1997.
C. BACON. Separation of waves propagating in an elastic or viscoelastic Hopkinson pressure
bar with three-dimensional effects, Int. J. Impact Engng Vol 22, p 55, 1999.
C. BACON and A. BRUN. Methodology for a Hopkinson bar test with a non-uniform
viscoelastic bar, Int. J. Impact Engng Vol 24, p 219, 2000.
C. BACON, E. GUILIORIT, B. HOSTEN, and D.E. CHIMENTI. Acoustic waves generated by
pulsed microwaves in viscoelastic rods: Modeling and experimental verification, J. Acoust.
Soc. Amer. Vol 110, p 1398, 2001.
D.J. FREW, M.J. FORRESTAL, and W. CHEN. A split Hopkinson Bar Method Technique to
Determine Compressive Stress-Strain data for Rock Materials, Exp. Mech. Vol 41, p 40, 2001.
P.J. RAE, H.T. GOLDREIN, S.J.P. PALMER, J.E. FIELD, and A.L. LEWIS. Quasistatic
studies of the deformation and failure of -HMX based polymer bonded explosives, Proc. R.
Soc. Lond. A Vol 458, p 743, 2002.
S.J.P. PALMER, J.E. FIELD, and J.M. HUNTLEY. Deformation, strengths and strains to
failure of polymer bonded explosives, Proc. R. Soc. Lond. A Vol 440, p 399, 1993.
J.E. FIELD, S.J.P. PALMER, P.H. POPE, R. SUNDARARAJAN, and G.M. SWALLOWE, in
"Proc. Eighth Symposium (Int.) on Detonation", (Naval Surface Weapons Center, White Oak,
Maryland, USA,1985) p. 635.
349
PROCESSES IN THE VESSEL DURING BLEVE EFFECT

Radovan Skacel, Bretislav Janovsky, and Jaroslav Svihovsky
University of Pardubice, Department of Theory and Technology of Explosives
Studentska 95, Pardubice 532 10, Czech Republic
Abstract
In the last fifty years, more than twenty great accidents with storage tanks, rail tank cars
and road tankers filled with pressurized gases with death toll more than one hundred
people, passed there. BLEVE boiling liquid expanding vapour explosion - is a type of
this incident, when containment (the cover of pressurized gas) is loosed. The paper
contains results of literature recherch, which displays some important findings about
recent research of BLEVE: 1) Now it is known that there are several types of accidents,
which are very different in consequences and magnitude of claims in dependence on
circumstances before rupture of the vessel like temperature and pressure, degree of
filling, rigidity of cover and creation of leak or hole before rupturing. 2) However, the
mechanism of last processes before rupture passes inside the vessel is not still exactly
clear. This process was summarized and specified on grounds of literary sources. In the
text, there is a suggestion of a possible scenario of the processes proceeding in the
vessel and its description. Definition of this mechanism is important for explanation and
prediction of blast and pressure effects in different starting conditions. For contribution
to resolve this problem, a special apparatus was designed that can measure liquid
behaviour in conditions of high pressures and temperatures similar to those in real
propane and butane vessels engulfed in fire, and can visualize the processes inside the
vessel with the help of endoscope-camera system.
Keyword:
1.
BLEVE, blast wave, superheat temperature limit
INTRODUCTION
In the whole world, many gases are used in the process industries and as the fuel for
building heating. These gases such as propane or butane are stored and transported often as
liquid pressurized gases (LPG) in metal vessels. It has been accompanied with accidents,
when this vessels exploded as a consequence of engulfing in fire, derailment etc. BLEVE
boiling liquid expanding vapour explosion is used as an acronym for the loss of containment
process.
What is BLEVE? It is one of the most serious hazards in the process industries.
Generally, this occurs when a pressure vessel, usually containing flammable liquid, is
exposed to fire, the metal loses its strength, internal pressure increases and vessel ruptures in
the end. It may be accompanied with some effects blast-wave, fireball and fragments but
all effects are not present every time. For example, when a non-flammable liquid is
superheated, no fireball occurs. So, in the present context the term BLEVE is used for any
sudden loss of containment of a liquid above its normal boiling point at atmospheric
conditions at the moment of its failure. However, some authors have distinct BLEVE
350
definitions and propose new terms. The typical process scheme before BLEVE is
following.
When a vessel containing liquid under pressure is exposed to fire, the liquid heats up and
the vapour pressure rises, increasing pressure in the vessel. When this pressure reaches the
set pressure of the pressure relief valve, the valve operates. The liquid level in the vessel falls
as the vapour is released to the atmosphere. The liquid is effective in cooling that part of the
vessel wall, which is in contact with it, but the vapour is not. The proportion of the vessel
wall, which has the benefit of liquid cooling, falls as the liquid vaporizes. After a time, when
a metal, which is not cooled by liquid, becomes exposed to the fire, the metal becomes hot
and weakens and may then rupture. This can occur even though the pressure valve is
operating correctly. A pressure vessel is designed to withstand the relief valve set pressure,
but only at the design temperature conditions. If the metal has its temperature raised, it may
lose strength sufficiently to rupture.
If the vessel bursts then blast wave to the surrounding area is released. The starting
overpressure has the same value, which was the pressure in the tank before rupture but
sometimes may be greater. It depends on temperature and pressure in the vessel, function of
pressure relief valves, material and shape of the vessel and other circumstances. The most
important kinds of scenarios are discussed later on.
If blasting separate the containment on two or more pieces then fragments will be
presented. Fragments have great kinetic energy and can travel to distances sometimes more
than one kilometre with weight of several tons in dependence on properties of tank, scenario
of accident and shape of the fragment.
If pressurized medium is flammable, fireball is the most dangerous effect of BLEVE,
which is produced by now gasified fuel released from vessel. The heat flux from fireballs are
sizable - more than 300 kW/m2 and majority of results with death tolls attached with BLEVE
rise from fireball radiation.
Special case happens when BLEVE is not fire induced. Then the released fuel may
evoke vapour cloud explosion, but this case is not very frequent. Most of accidents happen
when the vessel is engulfed by fire or the flame comes from any other tank with fuel or
sometimes jet flame from the pressure relief valve of one tank plays to another.
Compilations of BLEVE incidents are less numerous and comprehensive than those for
example for vapour cloud explosions. Anyway, there are more then twenty big BLEVE
accidents in the period 1950-1980, involving plants and transport documented in MIHDA
databank of the SRD. Summary death toll was more than 100 people. Usually rail tank cars,
road tankers or storage vessels exploded. Main effort in recent research is to determine all
scenarios of BLEVE, evaluate consequences of accidents like blast effects, fireball and
scattering of fragments at various conditions.
2.
THEORY
2.1
Mechanism of a BLEVE
In this section, the phenomenon of BLEVE is discussed according to the theories

proposed by Reid [1, 2, 3], Board [6], and Venart [5].
Reid [1, 2, 3] based a theory about the BLEVE mechanism on the phenomenon of
superheated liquids. The event described in this chapter concerns process that immediately
351
follow the initial failure of the tank metal. Soon after metal failure there is usually an
explosion accompanied by a shock wave that destroys and often fragments the tank. The
very rapid release of fuel is manifested in an impressive fire - even a fireball - and portions
of the tank can be scattered over a wide area. In some cases there may be a secondary fuelair explosion not dissimilar to those used by the military to produce damaging shock waves
over wide areas. But this is not standard situation. The question of primary interest is: What
is the source of the initial powerful explosion?
The explanation described here is related to the behaviour of the hot liquid, as the tank is
very rapidly depressurised after failure of its metal wall. Just before failure, the liquid was
saturated that is the temperature and pressure were as expected for a system where the
liquid was in equilibrium with vapour phase. With a rapid decay in tank pressure due to
failure of the metal wall the bulk liquid should boil and reduce a temperature to a value
compatible with a lower pressure. But initiation of the boiling process requires efficient
nucleating sites. There are no such nucleating sites in the bulk of the liquid and for a brief
period after depressurisation. This bulk liquid is superheated that is, at a temperature
higher than the boiling temperature predicted at the existing pressure. If the bulk temperature
were sufficiently above the boiling point, a superheated liquid-vapour explosion would be
expected. Such explosions are known to occur in the microsecond time domain with
concomitant shock waves.
Superheated liquid-vapour explosions occur frequently in some industrial operations.
Usually, however, they occur when two liquids one hot and non-volatile and the other cold
and volatile come in contact under appropriate conditions. The cold, volatile liquid cannot
boil on the hot liquid since there are no suitable nucleation sites, and it therefore superheats.
There is a theoretical limit to the attainable degree of superheat, and if this limit is reached
spontaneous homogenous nucleation occurs in the cold liquid with the formation of vapour
in a very brief period of time. To an observer, this vapour formation resembles an explosion.
Fig 1.
Vapour pressure and superheat curves for propane
Application of the superheated liquid-vapour explosion theory to the rapid

depressurisation of tanks containing a liquid may be illustrated by the following example.
Consider a tank of liquid propane, which is engulfed in a fire due to some accident. Before
the accident, the liquid propane was at ambient temperature and the tank pressure was 8 to 9
352
atm. This initial equilibrium state is shown as point A in Fig 1. Because of the fire, the
propane warms and the pressure increase follows the vapour pressure curve (Fig 1).
(Although in an actual case there may be some stratification of the warmer liquid near the
vapour-liquid interface, the assumption of a well-mixed liquid is not unreasonable.) A line
termed the superheat limit locus is also shown in Fig 1. This represents the limits to which
liquid propane may be heated before spontaneous nucleation occurs with a vapour explosion.
It can be calculated from theory or measured experimentally in the laboratory.
Suppose the propane failed when the pressure and temperature were those corresponding
to point B in Fig 1. Rapid depressurisation would occur as depicted by a vertical line
terminating at 1 atm (point C). Violent boiling would take place within a few seconds [1], but
since the superheat limit locus was not attained, no vapour explosion would be predicted.
However, if failure occurred under the conditions denoted by point D (55 oC), rapid
depressurisation to point E (3,3 atm) could lead to a vapour explosion. In fact, any tank
temperature exceeding about 53 oC (tank pressure > 16,3 atm) could produce vapour
explosion but if failure occurs before this point, only rapid boiling would be expected. This
definitive cut off suggests that safety valves might be designed to prevent any pressure
exceeding this critical limit.
Also, small-scale tests with single drops of superheat liquids indicate that the difference
in pressure between the vapour pressure curve and the superheated limit locus is related to
the intensity of the initial explosion. Thus tank failure near the critical point may result in a
lower explosive potential even though there is significant energy stored in the liquid.
2.2
Estimating the superheat temperature limit (STL)
Determination of the STL value for compounds, which are stored in pressure vessels, is
necessary for design of safety valves as explained in the chapter above.
Opschoor applied Van der Waals equation of state to estimate the maximum superheat
temperature for atmospheric pressure (Tsl) from the critical temperature (Tc) (i.e.,
temperature above which a gas cannot be liquefied by pressure alone) as follows:
Tsl = 0,84 Tc
Reid [2] used the Redlich-Kwong equation-of-state, which predicts a superheat

temperature limit as:
Tsl = 0,895 Tc
Reid ]2] further determined that the superheat temperature limit for a wide range of
industrial compounds falls within a narrow interval of 0,89 Tc to 0,90 Tc. Reid and many
other authors give pure propane a superheat temperature limit of 53 oC at atmospheric
pressure. The superheat temperature limit calculated from Van der Waals equation is 38 oC,
whereas the value calculated from the Redlich-Kwong equation is 58 oC. This value
indicates that, though an exact equation among P, V and T in the superheat liquid region is
not known, the Ridlich-Kwong equation of state is reasonable alternative.
353
2.3
Theoretical evaluation of STL:
There are two basic principles how to theoretically estimate the direct value of STL for
any compound. The first is thermodynamic approach, which is good for founding of STL for
any material in given pressure from temperature-pressure-volume diagram. The second
(kinetic approach) is based on principle how to find amount of potential nucleation sites,
which can rise to visible bubbles.
2.3.1. Thermodynamic prediction:
A two-dimensional projection of the pure liquid phase behaviour is shown schematically
in Fig 2. Pressure is plotted as a function of specific volume and a few isotherms are shown.
A liquid at Tl is in equilibrium with a vapour phase at Pl. For any volume less than V1, the
liquid is sub-cooled. The vapour phase in equilibrium with a liquid at Tl and Pl has a specific
volume of Vl. Similar comments apply to higher temperatures, e.g. if T2 > Tl then P2 > Pl.
T3 is drawn to represent the critical temperature isotherm, i.e. T3 = Tc. It is tangent to the
saturation envelope at the critical point and also has an inflection point. At any higher
temperature, such as T4, no liquid phase is possible, and the material approaches ideal gas
behaviour at low pressures. This is the equilibrium case.
The non-equilibrium situation, which encompasses superheated liquids, is quite
different. Consider again the basic structure of Fig 2, as redrawn in Fig 3. If a liquid is at Tl
at some pressure below P1, vapour may not form if the liquid is not in contact with suitable
nucleation surfaces and if no vapour is a priori present. That is the superheated liquid state.
As describe earlier, superheated liquids (dashed curves) are commonly said to have
transgressed, or crossed, saturation liquid phase boundary [20]. Fig 3 also shows the
thermodynamic superheat limit temperature (TSL). The locus of all TSL values is called the
spinodal curve. Thus, for example, T2 = TSL at P2. For Tl = TSL, the pressure P1 is negative.
Thermodynamic stability analysis indicates that for a pure liquid the superheat limit results
when (P/V)T = O [10]. For any isotherm, the slope on a P-V plot is zero at the intersection
between the isotherm and the spinodal curve. There is a complementary branch of the
spinodal curve on the vapour side of the saturation envelope, which defines the extent to
which a vapour phase may be subcooled, but it is outside the subject of this paper.
Based on the previous analysis, it is now possible to calculate rigorously the superheatlimit temperature for any pure liquid provided that we have pressure-volume-temperature
(PVT) data or an analytical function relating these three variables. Then, by simple
differentiation with the criterion (P/V)T = O. It is possible to delineate the limit of
superheat for any specified pressure. Many have tried to use this approach [11, 14, 18] - but all
results are in doubt since no satisfactory correlations now exist to relate P, V and T in the
superheated-liquid region. Some typical results obtained by selecting the simple RedlichKwong equation [16], which follows, are demonstrated:
1
a
P = R T
3/ 2
V b R T V (V + b )
where a and b are parameters, which are simple functions of the critical temperature and
pressure.
With this PVT correlation and the thermodynamic criterion, it is relatively simple to estimate
the ratio Tsl/Tc as a function of P/Pc and find the value Tc for given pressure.
354
Fig 2.
Pressure-volume-temperature
presentation of typical material
Fig 3.
Pressure-volume-temperature
presentation with spinodal curve
2.3.2. Kinetic approach:

Whereas the thermodynamic treatment of superheated liquids is rigorous and necessary,
but it is also cold and impersonal. There is not a glimmer of the exciting events taking place
on a microscopic scale! Vapour embryos are born, grow, break up, condense, reform, etc.
The kinetic approach is concerned with describing these events. The ultimate question is,
how rapidly do visible vapour bubbles form in a superheated liquid? The theory is much too
detailed and the mathematics too involved describing here, but the reader can refer to several
review papers and theses [11, 13, 15, 17, 19, 20]. There will be described model briefly and results
will be illustrated in this paper. In the beginning, a volume of superheated liquid isolated for
many influence of the environment, which is readily accomplished in any real case, will be
selected by choosing short time scales and small sample sizes. Random density variations are
always present and these create monomer, dimer, etc., cavities (or vapour holes) that give
birth to vapour embryos.
These cavities exist in a dynamic local environment and may grow by further
vaporization, molecule by molecule, or they may diminish inside and perhaps disappear by
condensation or vapour molecules on the bubble-embryo walls. These embryos are invisible
to the naked eye, but theory indicates that, on a molecular level, there is an enormous
population of them distributed in size in an exponential manner. The number of embryos
with g vapour molecules in a Unit volume of liquids is proportional to e-a/kT, where k is
Boltzmann's constant, T is the absolute temperature, and a is analogous to the work
necessary to form the g-mer. Because increases significantly with g, most embryos contain
only a few vapour molecules.
The most logical question is why don't some of these vapour embryos grow to become a
visible bubble? The answer here is necessarily simplistic because the true mechanism is
subtle. All events in our experience occur because some potential or a driving force exists
there. In superheated liquid-vapour embryo system, there is a driving force that encourages
further vaporization and embryo growth. But there is, in addition, an opposing force exerted
by intermolecular interactions at the vapour-embryo interface. In effect, a surface force
exists which squeezes the embryo (bubbles do tend to be spheres!). In very small embryos,
355
this surface effect predominates, and the embryo is doomed to eventual destruction unless,
by a series of fortuitous additions of molecules, it can grow to the critical size, at which the
driving force that favours vaporization balances this destructive surface force.
Growth larger than the critical size then continues unabated until a visible bubble
appears. Reviewing this simplified picture, it is possible to find that many microscopic
vapour cavities are created by fluctuations in density or by some other initiating process
(radiation). These embryos valiantly try to grow and, depending on their fortune, a few may
reach large size. Most, however, recondense and disappear. Rarely does one attain the
critical size, after which further growth is assured. Kinetic theory allows estimating
numerical values of the rate of formation of critical-size embryos (and thus vapour bubbles)
from a given volume of liquid. Inconceivably low rates are predicted unless the liquid is
highly superheated.
Relating these data to the bubble-column experiments, as a liquid drop rises in to the
warmer, immiscible host liquid, it becomes progressively more superheated. Yet for most of
the drop's life as a superheated liquid, kinetic theory predicts that virtually no vapour embryo
can expect a fate other than eventual condensation. The n-pentane drop in the sulphuric acid
system was superheated at all temperatures in excess of 309 K, but not until it reached a
temperature of 419 420 K did the rate of formation of critical-size vapour embryos attain a
value where vapour appeared. Then, over a very small temperature range, the rate increased
so drastically that there was an apparent vapour explosion. In order to calculate this
superheat-limit, or vapour-explosion, temperature, using kinetic theory, it is necessary to
decide what rate of formation of critical-size embryos constitutes an explosion. Different
values have been chosen, but fortunately the results are relatively insensitive to the choice.
For example vapour explosion results if a million critical-sized vapour embryos are created
every millisecond in the cubic millimetre of superheated liquid. Applying this criterion to
n-pentane, it is found that the vapour explosion temperature is 421 K (148 oC) - close to the
experimental value of 419 420 K. As predicted nucleation rates value by several orders of
magnitude per degree, estimated values of the rates are, at best, only an approximation.
The important result is that kinetic theory indicates no significant bulk nucleation until a
definite temperature is reached. Then, in a range of a few degrees, the rate changes from a
value, which is negligibly small to a very large value.
Predicted value of Tsl/Tc can be graphed as a function of P/Pc, as calculated from the
kinetic approach. If these results are compared with these calculated from the
thermodynamic approach, the curves are in qualitative agreement.
2.4
Boiling liquid collapsing bubble explosion (BLCBE)
Venart and co-workers [4, 5] further investigated the interaction between homogenous
nucleation and the vessel itself and developed the concept of the boiling liquid collapsing
bubble explosion BLCBE.
The BLCBE has three aspects, involving the behaviour of liquid, a crack in the vessel
wall and the liquid-vessel system. If the vessel wall in the vapour space is subject to heating
by a flame, a crack may develop. The behaviour of this crack than depends on the degree of
fill. If the fill is high, the crack may be arrested in the vapour space section of the wall by
plastic blunting. If the fill is lower, there may be sufficient energy in the vapour to cause the
crack to propagate plastically to the cold metal, where it can than propagate elastically
without restraint and with only minimal energy requirement.
356
This second case, with the lower fill, corresponds to the regular BLEVE. The first case
has the potential for a much greater release physical energy associated with homogenous
nucleation. If the fill is high, one of two effects may occur: (1) massive two-phase jet or (2) a
violent vessel failure with total loss of containment caused by a homogenously nucleated
liquid.
It is this second type of event in a vessel with high fill, which constitutes a BLCBE. It
occurs with a liquid homogenous in temperature and depressurising. The sequence of events
is envisaged to be typically as follows. The liquid is depressurised, say by operation of the
pressure relief valve, and due to this becomes nucleated. When the crack occurs, there is an
almost instantaneous swell of low void liquid, which chokes in the crack. There is a rapid
depressurisation of the vessel contents. The resultant liquid shock pressures are very high
(1 - 3 kbar). The crack propagates elastically through now-cooled wall of the vapour space
and on the wall of the liquid space. The vessel fails and the highly compressed two-phase
contents are again depressurised. Expansion of the high-pressure vapour bubbles shatters the
liquid and results in ejection of the whole of the superheated liquid as a fine aerosol. If
flammable, this aerosol may than be ignited.
Venart et al. [4] describes tests in which prenucleated R112 and R113 were used end
events of the type just described were observed, with destruction of the vessel in time as
short as about 200 s. They give details of the pressure effects in the vessel, including
measurement of the rapid pressure transient bellow the liquid in excess of the static pressure
in the vessel and an estimate of the instantaneous overpressure at the crack. The vessel static
pressure was some 790 kPa, the average excess pressure 490 kPa and the estimated
overpressure at the crack itself some 6200 kPa. These authors also give estimates of the
diameter of the bubbles (1 3 mm), the diameter of the resultant aerosol droplets (100 m)
and the evaporation time of the droplets (40 ms). They state that there were significant blast
overpressures.
What is very interesting, these effects were obtained with a liquid temperature more than
50 C below the superheat limit temperature. Venart et al. suggests that all three types of
event described (regular BLEVE, two-phase jet and BLCBE) occurred at Mexico City. They
draw attention to the fact that a BLCBE may be caused by a leak, and discuss the
implications of this for the leak-before-break philosophy.
o
2.5
Summary of Venarts conclusions about BLEVE and BLCBE:
Venart separates containment accidents to three basic types: two-phase jet (release of
material as a jet from the vessel; this type does not lead to loss of containment), less violent
regular BLEVE (less filling containers) and very violent type, which designated BLCBE
(more filled containers). However, according to definition of BLEVE (BLEVE is a sudden
loss of containment filled with superheated liquid) both last types satisfy definition of effect
BLEVE.
Problem is more or less phenomenological, because BLCBE is one of the possible
scenarios for physical effect called BLEVE.
Nevertheless, it is certain there are two different types of accidents (as Venart
described), which results in a loss of containment (for our approach we eliminate doublephase jet, which does not lead to a serious spread of leak and loss of the vessel). The first
one is loss of containment accompanied with flash of liquid and vapour phase as a
consequence heating of the vessel (energy of vapour phase can significantly participate).
This flash is not very violent and sometimes does not separate the vessel to more pieces (but
it devastates containment completely). The second type occurs when conditions allow the
357
onset of homogenous nucleation in satisfactory part of liquid that creates a strong blast wave
and the consequences are more dangerous (total loss of the containment and the strong blast
wave). This type of accidents always become when liquid is at the moment of failure above
its superheat temperature limit. However, in special circumstances, it can happen although if
the liquid is below superheat temperature limit as Venart described for his BLCBE and
consequences can be similarly very serious.
3.
ASSUMPTIONS AND INTENTIONS
It is assumed that scenario described by Venart for BLCBE is basically true for
superheated liquids (pressurized gases), which are below or above its superheated
temperature limit (STL) (all Venart's experiments were performed below STL). Although
this is possible, then pressurized gases, which are situated in the moment of the accident
above its STL, undergo different scenario. There are confusions of authors, what is the
reason of the final bursts of vessels, if it is the water hammer, the instantaneous excess of
pressure in liquid phase based on collapsing bubbles or if it is only the energy coming from
vapour phase.
Our assumption is as follows (Janovsky [21]):
In the vessel creates a leak as a consequence, for example, of the vessel heating. Vapour
phase escapes and pressure drops hardly. Refraction wave passes through the vessel and
implicates formatting two-phase liquid-bubble system from the liquid in the vessel. This
system increases its volume and cokes a leak as Venart described. However, there is a
possibility that the explosion does not occur immediately, but the process is morephased.
This assumption is based on the facts of pressure-time graphs obtained from Venart's and
Mac Devit's experiments (Fig 4 and 5). When the vessel is penetrated, than vapour phase
releases and the pressure drops. Liquid stands in metastable state (the pressure is lower at
given temperature than at the saturation pressure) in time scale about 100 ms and than
pressure rise occurs probably as a consequence of the bubble rise. However, the pressure
does not need to rise to its maximum value continuously as was possible to see in several
cycles, which support our assumptions.
Fig 4.
Pressure response during a BLEVE

of freon R22 in a 1 litre tank
(Mc Dewit) [7]
Fig 5.
Pressure response during a BLEVE

of freon R11 in a 0,4 litre tank
(Venart) [4, 5]
358
We propose the following possible process before a rupture of the vessel:

1. The vessel filled with pressurized gas is heat stressed. Pressure attains the value of
setting pressure relief valve, which operates, and the pressure slightly drops. As a
consequence of the pressure drop, bubbles are created from miniature bubbles
presented in the liquid (see chapter theory, kinetic approach). This first phase could
not run over always, but when pressure relief valve operates, accident often ends like
a BLCBE upon Venart's experiments.
2. A leak creates and the pressure drops the growth occurs in the top layers of liquid
bubbles and very volatile vapour-liquid phase swells. When reached double-phased
interface (between the vapour-liquid system and the gas phase) the top of the vessel
then bubble break down and phases are separated (liquid falls down).
3. Now the top layer of liquid is fine nucleated. These bubbles grow in the liquid and
the liquid-vapour phase level rises to the top of the vessel. When this liquid-rich
phase reaches the top of the vessel, and then cokes a leak and the pressure rises
again. The bubbles have not the right size to be able to collapse in a consequence of
the pressure in liquid. The liquid and the vapour are separated again but the liquid is
now more nucleated.
4. Deeper layers are nucleated and the process is repeated. In the next cycles, more and
less bubbles are created and the density of the prenucleated phase increases. When a
sufficient amount of the liquid and the bubbles achieve the right size then after
compression serious amount of bubbles will collapse and great pressure wave in
liquid is generated and the vessel ruptures.
4.
APPARATUS:
Svihovsky showed the scheme of proposed apparatus in his paper [22]. The idea is to
create a horizontally placed apparatus with the middle part, which can be filled with the
volatile liquid of volume 1,5 2 l and heated. After membrane penetration the medium will
release to expanding tubes.
359
The requirements to the apparatus measurement ability are:

1. Pressure effects (in liquid and vapour phase in the middle pressure part and in
expanding tubes)
2. Temperature in liquid and vapour phase
3. Visualization of processes in pressure part (endoscope + camera system)
Two basic aims could be outlined:
1. To explain how really the processes come in the vessel and to describe these
processes.
2. To evaluate the pressure effects in different scenarios (temperature above and
bellow STL, prenucleation, different shapes of primarily lake) in 1D geometry
(tunnels).
5.
CONCLUSIONS
When the vessel containing liquefied gas (superheated liquid) ruptures, the
consequences and magnitude of the accident depends primarily on the temperature and
pressure inside the vessel. When the temperature is above the superheat temperature limit,
than after the leak or hole creation in the vessel, homogenous nucleation accompanied with
bubble collapse occurs and a total loss of the containment with serious air pressure wave
would be expect.
When the temperature is bellow STL, than often only a relatively slow flash of liquid
occurs without pressure wave but in some special circumstances bubble collapse, like in the
first case, can occur. Then violent pressure growth burst vessel and strong pressure wave is
released again.
We want to explain and display how these processes in the vessel come on and to
estimate the pressure effects in different circumstances and scenarios of proceeding
processes. To attain this aim, the special apparatus with the possibility to measure the
pressure effects inside and outside the vessel and to visualize the processes in the internal
vessel part has been designed.
360
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
R. C. REID: Possible mechanism for pressurized-liquid tank explosion or BLEVEs, Science,

203(3), 1979
R. C. REID: Superheated liquids, American Scientist, 64:146-156, 1976
R. C. REID: Some theories on boiling liquid expanding vapor explosions: Fire, 525-526, March
1980
J.E.S. VENART, K. SOLLOWS, G.A. RUTLEDGE, K.SUMATHIPALA AND X. JIAN: Boiling liquid
compressed bubble explosion, Experiments/Models. FED-Vol. 165, ASME, Gas-liquid flows,
pp. 55-60, 1993,
J.E.S. VENART: In Vessel Transient Termophydraulics, Heat Transfer and Major technological
Hazard proceedings (Vol.2), Eurotherm Seminar Nr. 14, Universite Catholiquid de Louvain,
Louvain-la-Neuve, Belgium, pp I1.1-I1.12, May 15-18,1990
S. J. BOARD, R.W. HALL AND R.S. HALL: Detonation of fuel coolant explosions, Nature 254:
319-320, 1975
C.A. MCDEWITT, C.K. CHAN, F.R. STEWARD, K.N. TENNAKORE: Initiation step of boiling
liquid expanding vapour explosions, Journal of hazardous materials, 25, 169-180, 1990
CCPS: Guidelines for Evaluating the Characteristics of Vapor Cloud Explosions, Flash Fires,
and BLEVEs, American Institute of Chemical Engineers, New York, NY, 1994
F.P. LEES: Loss Prevention in the Process Industries, London: Butterworths, Second Edition,
1996
B. BEEGLE, M. MODELL AND R.C. REID: Thermodynamic stability criterion for pure
substances and mixtures, AIChE J. 20: 1200-06, 1974
M. BLANDER AND J.L. KATZ: Bubble nucleation in liquids, AIChE J. 21:833-48, 1975
S.J. BOARD, C.L. FARMER AND D.H. POOLE: Fragmentation in thermal explosions, Int. J. Heat
Mass Trans. 17:331-39, 1974
J.J.CHEN: Nucleation in a metastable liquid, Ph.D. Thesis, Rice University, Huston, Texas,
1970
G.V. ERMAKOV AND V.P. SKRIPKOV: Equation of state and stability boundary of a metastable
liquid, Tep. Vysokikh Temp. 8(5):976-82, 1970
Y. KAGAN: The kinetics of boiling of a pure liquid, Zh.Fiz. Kchim. 34:92-96, 1960
O. REDLICH AND J.N.S. KWONG: On the thermodynamics of solutions, V. Chem. Rev. 44:44043, 1949
E.N. SINITSYN AND V.P. SKRIPOV: Kinetics of nucleation in superheated liquids, Zh. Fiz.
Kchim. 42:440-43, 1968
V.P. SKRIPOV: The shape of the spinodal curve in liquid-vapor systems and the rectilinear
diameter rule, Zh. Fiz. Kchim, 39:438-40,1965
V. P. SKRIPOV: Metastable liquids, N.Y: John Wiley and Sons, 1974
M. VOLMER: Kinetic der Phasenbildung. In, Die Chemische Reaktrion, Band IV. Dresden:
Theodore Steinkopff, 1939
B. JANOVSKY: BLEVE blast wave generation process, TNO Prins Maurits Laboratory, report
No. 2001- UPCE-TNO, 2001
J. SVIHOVSKY, B. JANOVSKY, R. SKACEL: Measuring of the BLEVE internal processes,
6th International Seminar New Trends in Research of Energetics Materials, Univ. Pardubice,
April 2003
361
GEMINAL TRINITROCOMPOUNDS THERMAL DECOMPOSITION

UNDER NON-ISOTHERMIC CONDITIONS
Rudolf S. Stepanov, Ludmila A. Kruglyakova and Alexander M. Astachov
Siberian State Technological University
660049, Krasnoyarsk, prosp.Mira, 82, Russia
Abstract:
The influence of chemical structure on thermostability of the polyfunctional geminal
trinitrocompounds is shown. Characteristic temperatures and activation parameters of
thermal decomposition are determined. It is shown that compounds thermostability may
change by more than two orders of magnitude depending on the compounds structure.
This is caused not only by the influence of substituents, but also by difference in the
mechanism of final decomposition stage.
Keyword:
1.
thermal decomposition, geminal trinitrocompounds, liquid phase,

mechanism, compensation effect, kinetics, characteristic temperatures.
INTRODUCTION
Thermal decomposition of the polynitrocompounds of general formula R-C(NO2)3,

where R is: 1H-Tetr*- (I), 2CH3-Tr**- (II), CH3CONHCH2- (III), C6H5-CONHCH2- (IV),
(O2N)3C6H2N(NO2)CH2- (V), CH3N(NO2)CH2- (VI), C2H5-N(NO2)CH2- (VII), C2H5OOCCH2N(NO2)CH2- (VIII), H2C[OCH2-]2 (IX), (O2N)2C[CH2O-]2CHCH2CH2- (X),
C6H3[COOCH2-]3 (XI), (O2N)3C6H2COOCH2-(XII), -CH2CH2N(NO2)CH2COOCH2- (XIII),
HOOCCH2CH2N(NO2)CH2CH2(XIV),
HOOCCH2N(NO2)CH2(XV),
and
HOOCCH2CH2N(NO2)CH2- (XVI) is of interest for the following two reasons. First, in
order to predict how synthesis of a compound proceeds it is important to know how its
structure influences thermostability. And second, thermal decomposition allows estimating
conditions of thermal combustion and burning for these compounds.
2.
EXPERIMENTAL
Decomposition and thermal analysis were carried out using derivatograph in the
temperature range of 20-500OC. The heating rate was 2.5OC/min and sample mass was 0.1g.
TG curves were treated using splines [1]. Immediate values for the mass of a sample were
derived from the immediate values for the change of mass (mi): mi=m0*(1-mi/ mmax).
Smoothing cubic splines approximated how sample mass immediate values depend on time
(temperature). Since a derivative can be taken for any value of t for function m=f(t), one may
use the most general and simplest equations of formal kinetics to calculate activation
parameters for decomposition, that is, -dm/dt = k*mn, where k=A*exp(-E/RT). A typical
thermogram is shown on Fig.1.
1H-Tetrazolyl2-Metyl-1,2,3-Triazolyl-
**
362
300
DTG
T, OC
T
200
DTA
Time
100
0
G, mg
TG
25
Fig 1.
3.
Thermogram for compound IX
TG and DTG curve analysis showed that mass loss proceeds either in one (compounds
III, VI-X, XIII-XVI), two (compounds II, IV, XI) or three (compounds V, XII) stages. At the
first stage, thermal decomposition of trinitromethyl group occurs. Intensive oxidation
destruction of the molecules fragments down to simplest compounds also occurs at the first
stage. This is proven by the IR-spectra analysis data for gas products of compounds IV, IX,
XV and XVI near Tmax of decomposition. Independently of the compounds structure, the
main gas products of the decomposition proved to be NO2, NO, N2O, N2, CO2, CO. The
exothermic peak accompanies decomposition. The peaks size depends on the compounds
structure. Destruction dynamics may be judged by TG and DTG curves, which appear to be
the steepest for compounds I-V, XIV-XVI. Slow destruction of the more thermally stable
molecular fragments (triazol and benzol substituents) proceeds at the second stage at higher
temperatures (200-250 OC). Cross-connecting bonds are formed in the structure of condensed
residue at the third stage. The last two stages are thermoneutral. One may judge by
endothermic peak on DTA curve that melting precedes decomposition for II, IV and VI-XIII.
The endothermic peak on DTA curve is absent for the other compounds so their
decomposition starts in the solid state.
The kinetic compensation effect shows in the polynitrocompounds (Fig.2) This is
characteristic for catalytic reactions.
363
22
21
20
13
lgA = (0.114 0 .006)E-(2.051 0.923)

r=0.992
S lgA = 0.035.
19
lgA
18
10 9
17
16
11
12
15
4
14
13
12
120
130
140
150
160
170
180
190
200
E, kJ/mol
Fig 2.
Plot of all Arrhenius parameters for polynitrocompounds thermal

decomposition showing that a kinetic compensation effect exists. Points
denominators correspond to the Table 1 compounds denominators.
lgA = (0.1140.006)E-(2.0510.923) r=0.992 SlgA = 0.035
Judging by velocity constant, one may notice from Table 1 that the least thermally stable
compounds are the derivatives of amid acids (compounds III, IV). The most thermally stable
are esters and acetals of trinitroethanol (compounds IX-XII). Trinitroalkylnitroamines
(compounds VI-VIII, XIII) are of medium thermal stability. The other thermal stability
indicator, Tmax, increases with velocity constant decreasing. Its value for esters and acetals is
187-200 OC, for trinitroalkylnitroamines is 135-156 OC, and for trinitromethylazols is
78-105 OC.
One may expect that thermal decomposition of nitroaminoacids (compounds XIV-XVI)
in a solid phase would show Tmax higher than that of nitroamines. This is because
nitroaminoacids have higher melting points due to the braking effect of the crystalline
structure. Contrary to the expectations, Tmax for nitroaminoacids is in the range of 145-152
O
C and is close to the nitroamines liquid phase Tmax figures. Possibly, this may be a
manifestation of the stronger intramolecular interactions between functional groups
C(NO2)3, =N(NO2) and COOH. Also, one may not exclude intermolecular acid catalysis in
addition to oxidation-reduction autocatalysis. Intermolecular acid catalysis was also
observed before at decomposition of the 3-oxy-2,4,6-trinitrophenyl-N-methylnitroamine [2]
364
Table 1. Characteristic temperatures and Arrhenius parameters for the first stage
of thermal decomposition of the polynitrocompounds of general formula RC(NO2)3. Tm is a melting point temperature, Tb is a temperature at which
decomposition begins, Tmax is a temperature at exothermic peak maximum
N
1
Substituent R
1H-Tetr
Tm
Tb
kJ/mol
65
78
Tmax
Log A
K160
*104 S-1
2CH3-Tr
26
90
105
CH3CONHCH2
92
95
108
123.5
13.65
728.6
C6H5CONHCH2
105
114
132
128.5
14.25
570.1
(O2N)3C6H2N(NO2)CH2
120
135
H3CN(NO2) CH2
85
123
148
136.9
13.85
22.0
H5C2N(NO2)CH2
64
110
137
134.9
13.39
13.7
H5C2OCOCH2N(NO2)CH2
78
122
144
152.5
15.58
16.0
H2C[OCH2]2
66
160
195
180.2
18.22
3.3
10
(O2N)2C[CH2O-]2CH(CH2 )2
145
158
187
178.1
18.11
4.5
11
C6H3[COOCH2]3
152
162
195
164.5
16.52
4.8
12
(O2N)3C6H2COOCH2
138
165
200
168.3
16.90
4.0
13
CH2OCOCH2N(NO2)CH2CH2
107
127
156
192.7
20.42
16.2
14
HOOC(CH2)2N(NO2)CH2CH2
117
145
15
HOOCCH2N(NO2)CH2
118
150
16
HOOC(CH2)2N(NO2)CH2
120
152
On the whole, change in thermal stability for compounds under investigation ranges by
two orders of magnitude. Not only inductive and steric effects of substituents and
autocatalysis cause this difference among compounds, but also difference in mechanism of
the limiting decomposition phase plays a role.
Thus, decomposition of geminal trinitrocompounds in a condensed phase proceeds
homolitically and is limited by a break in C-NO2 bond [3]. Homolitical mechanism changes
for heterolitical decomposition on bond C-C with formation of carbkation and anion
trinitromethan when NH- or NHCO- functions are found in -position in respect to
trinitromethyl group [4].
XCONHCH2C(NO2)3 XCONHC+ H2 + C -(NO2)3
where X = CH3, C6H5.
365
This is indirectly confirmed by the fact that the values of lgA for compounds III and IV
fall out of fit (See Fig.2). Heterolitical mechanism for decomposition in condensed phase
was assumed also for compounds of type ROCH2C(NO2)3 [5].
4.
CONCLUSION
The kinetic regularities discovered may be used to predict thermal stability for new
compounds based on their chemical structure. They may also prove useful when constructing
or numerically simulating combustion wave.
REFERENCES
[1]
[2]
[3]
[4]
[5]
A.M. ASTACHOV, A.A. NEFEDOV, L.A. KRUGLYAKOVA, AND R.S. STEPANOV: Siberian
Conference Methods of Spline-Functions. Novosibirsk, Sobolev Institute of Mathemtics (In
Russian), SORAN, Russia, 2001
C.N. HINSHELWOOD: J. Chem. Soc., 1921, Vol 119-120, 72
G.M. NAZIN, W.G. PROKUDIN, G.B. MANELIS: Russ. Chem. Bulletin, 2000, Vol.2, 231
R.S. STEPANOV, V.I. VLASENKO, M.N. PETRASHKEVICH, P.A. IVANOV: Izvestiya Vysshih
Uchebnyh Zavedenii Khimii I Khimicheskoi Tehnologii (In Russian), 1975, Vol 18, 733
T.B. BRILL, R. SUBRAMANAN: Combustion and Flame, 1990, 80, 150
366
ELIMINATION OF LEAD AND LEAD CONTAMINANTS

IN THE MK 141 MOD 0 DIVERSIONARY CHARGE
J. Stockinger and T. Reed
Naval Surface Warfare Center, Crane Division,
300 Highway 361, Crane, Indiana 47522 USA
Abstract:
The Charge, Diversionary, MK 141 Mod 0 is a pyrotechnic device used for diversionary
purposes. The device is used and operated like a hand grenade without its lethal effects.
The fuze system contains a pyrotechnic delay column intended to delay the function of the
item until after a user has thrown it. This delay is crucial for mission effectiveness and user
safety. The original design contained a lead-sheathed delay column where a boron-barium
chromate delay composition was compacted in a lead tube through a progressive
rolling/swaging process. The Naval Surface Warfare Center Crane Division sought
replacement of lead in the delay holder during early production of the MK 141
Diversionary Charge. A first attempt, in the early 1990s, was incomplete due to exhaustion
of funding. In 1999, the Naval Surface Warfare Center Crane Division and the prime
contractor, Pyrotechnics Specialties, Incorporated (PSI), Byron, GA, began a limited
product improvement teaming effort to improve environmental impact and producibility of
the boron-barium chromate pyrotechnic delay design by using an aluminum housing to
replace the lead sheath of the existing design. The change also included the incorporation
of a pressed boron-barium chromate delay column. The change resulted in environmental
improvements at all stages of the life cycle of the item, while generating a recurring
manufacturing cost savings, with no adverse effects on performance or reliability. The end
result was that lead was removed from manufacturing, usage, and disposal aspects of the
items life cycle. This mitigated expensive environmental problems on test and training
ranges. Savings achieved were both tangible and intangible.
Keyword:
1.
MK 141 Mod 0, pyrotechnic device, elimination of lead
ITEM DESCRIPTION
The MK 141 Mod 0 Diversionary Charge, Figure 1, is a pyrotechnic device used to confuse
and disorient an enemy by producing bright light and loud sound. It is a low hazard device,
which produces an intense flash and sound report, with a minimal amount of smoke. Sometimes
referred to as a "flash bang", it is intended to provide a non-lethal means of temporarily
dazzling, dazing, or disorienting targeted individuals or groups before they can injure
themselves or others.
Made primarily of molded polyethylene and polyurethane foam, the diversionary charge
consists of two major subassemblies, the fuze system and the pyrotechnic charge. Upon removal
of the safety pin and release of the lever, the spring-loaded striker initiates the primer. After a
delay of approximately 1.4 seconds, the separation charge initiates.
DISTRIBUTION STATEMENT A. Approved for public release; distribution is unlimited.
367
Fig 1.
MK 141 Mod 0 Diversionary Charge - scheme
A closely controlled burn time is crucial for mission effectiveness and user safety. The
separation charge ejects the fuze from the main body to preclude the fuze becoming a missile
hazard due to pressure of the main charge function and ignites a second delay of 100
milliseconds. The second delay initiates an ignition charge, in turn, setting off the output
charge of aluminum powder, carbon, and potassium perchlorate. The device is 1.750 inches in
diameter and 5.0 inches in height. The main body of the device is made of polyurethane foam
and is consumed during function of the main charge. The fuze body is of 10% glass filled
polyethylene. Since the fuze is ejected and the main body is foam, fragmentation is not a
significant hazard during function of the main charge. Figure 2 is a photograph of the item.
Fig 2.
MK 141 Mod 0 Diversionary Charge original view
368
The operating parameters of the device are nominally 2,000,000 candela luminous intensity
and 182 dB sound output.
2.
BACKGROUND
The design was the product of a series of government contracts between the government and
Ensign-Bickford Aerospace Company (EBAC) in Simsbury, Connecticut in the late 1980s.
The original design consisted of a lead sheathed delay column, containing a boron-barium
chromate mix, where the composition was compacted through a progressive rolling/swaging
process. This delay was fabricated by modifying a 2.70 second burn time per inch delay cord,
manufactured in-house by EBAC. The cord was cut to a fixed length with a small saw and
crimped first into an aluminum ferrule, Figure 3, and then into a .30 caliber brass cartridge case,
Figure 4. This T1 Delay Module Assembly was then molded into the plastic fuze body.
Figure 3.
Figure 4.
3.
IMPROVEMENT OPPORTUNITY
Three competitive procurements of the device were awarded by NSWC Crane, to

Pyrotechnic Specialties, Incorporated (PSI) in Byron, Georgia. The first was awarded in 1996.
Follow-on contracts were awarded in 1998 and 2002. Several events occurred during 1999, in
the latter stages of production on the 1996 contract, leading to interest, both on the part of the
government and the contractor, in changing the design of the T1 Delay Subassembly.
369
Long-standing necessity to eliminate lead hazards from manufacturing, training, and

operational environments
Diminished availability of 2.70 second per inch lead sheathed delay cord
Expiration of an interim hazard classification for the delay cord
Significant dud rates (approximately 4%) due to failure to transfer from the T1 to T2
delay subassemblies
Process control difficulties in cutting delays, with accurate burn times, from the delay
cord
Concern, on the part of the prime contractor, to perform delay fabrication work in-house
Selection and use of a pressed delay column to match the process to the prime
contractors capabilities
After several shipments of lead sheathed delay from their vendor, McCormick Selph
Incorporated, in Hollister, California, PSI was faced with the expiration of the interim hazard
classification. Renewal or extension was not an option. For subsequent interplant shipments of
delay cord, PSI was required to bear the expense and delay of obtaining a hazard classification
assignment (EX number) from the U.S. Department of Transportation. Also, PSI was having
problems maintaining control of delay burn times in the cutting of the cord lengths. The Navy
recommended a direction for the effort that included elimination of lead sheath. Following
several discussions between the Navy and PSI, the contractor proposed an alternate design of
the delay using the same delay composition pressed into a machined aluminum housing. In
order to avoid an extensive and expensive repeat of numerous safety tests, the composition and
physical envelope of the delay subassembly were left unchanged.
4.
APPROACH
Significant changes in an item containing energetics must be thoroughly tested, reviewed,

and approved by the Weapons System Explosives Safety Review Board (WSESRB). The
described course of action was chosen to minimize the impact of the change and allow a
minimal amount of qualification testing to satisfy the WSESRB. The Navy, in 1992 and 1993,
as an adjunct to the last contract with EBAC, contracted with EBAC, in an attempt to replace
the lead with aluminum or pewter. These efforts were incomplete and no change was adopted
as a result of the effort.
Pyrotechnic Specialties, Incorporated and the Navy used an iterative review process to
finalize the physical configuration of the alternate delay subassembly. Since there were no
funds available to separately fund the effort, the Navy and contractor agreed to an allocation of
tasks, risks, responsibilities, and rewards.
The contractor conceptualized the pressed delay element, including the aluminum housing.
The Navy reviewed the proposed changes and consensus was reached. The contractor built 100
prototype T1 delays and performed burn time tests at ambient, hot, and cold environments. The
delay consolidation process was finalized and documented with 82 MK 141 Diversionary
Charge units tested as a first article. Test results and supporting documentation were sent to the
WSESRB and the change was approved.
370
Four increments of the composition, with the addition of a binder, are pressed into the tube,
on top of the separation charge. The design of the housing allows a more focussed jet of
burning gas from the separation charge to ignite the second delay column. This is another
element of improvement in the reliability of the transfer from T1 to T2.
The aluminum tube, as a replacement for lead, eliminates environmental concerns in three
distinct areas. First, lead particles, residue and waste generated during the manufacturing
process are eliminated at the source. Second, lead vapor emitted during item function is
eliminated from training facilities. Third, residual lead scrap is eliminated in the field by
replacement of lead sheath with aluminum.
The aluminum delay module assembly, Figure 5, contains the same amounts and
percentages of boron/barium chromate delay composition as the T1 Delay manufactured by
McCormick Selph Incorporated.
Figure 5.
The focussed transfer of ejection charge to the T2 delay column has resulted in very reliable
transfer while requiring a lower ejection charge weight. The improved reliability of the transfer
between the T1 and T2 has reduced the incidence of duds. Over 2200 test firings have been
performed on 22 lots produced with no duds experienced due to failure to transfer between T1
and T2. The reduced ejection charge weight has resulted in lower ejection charge velocities and
the associated hazard due to fuze ejection.
371
5.
ALLOCATION OF SAVINGS
The savings of this initiative accrue both to the government and the contractor.
5.1
Contractor savings
To date, over 220,000 devices have been manufactured with the aluminum delay
housing as a replacement for lead. With the elimination of the cutting process for the
lead sheath, the environmental impact and disposal cost of nearly ten kilograms of waste
in the form of lead particles has been avoided.
Employees at the manufacturing facility are no longer exposed to the lead in the product
they are processing.
The cost of obtaining a Final Hazard Classification was avoided for the lead-sheathed
delay since interplant component material shipments would no longer be needed.
The change to an in-house processing effort resulted in manufacturing cost savings.
The pressing process is more easily controlled and requires less in-process testing than
the roll/swaging and cutting processes.
Delivery schedule slippage has been reduced since moving delay fabrication in-house.
The contractor has also reduced the risk of lot acceptance test failures due to erratic
delay burn times.
Total tangible cost savings for the contractor, to date, is estimated at $110,000.
Intangible savings in health and environmental impacts are significant.
5.2
Government savings
Elimination of the lead sheath in over 220,000 units has eliminated environmental
impact of over 1045 kilograms of lead in the spent fuze assemblies.
Need for personnel breathing protection on training ranges due to airborne lead
particulates has been eliminated.
Use of units in training has mitigated the incremental increase in cost of range cleanup.
Improved competition since a pressing procedure rather than a swaging procedure is a

more wide spread method and can be easily tooled at various sites.
Improved competition since the Navy acquired unlimited data rights in a new delay
design for the cost of preparing the drawings, validating tests, and obtaining WSESRB
safety concurrence.
The Navy will avoid the incremental cost of future range cleanup with the elimination of
the source of lead contaminants.
Improved reliability and material availability.
Total tangible cost savings on current and future contracts is estimated at $175,000.
Intangible savings in health and environmental impacts are significant.
372
6.
CONCLUSION
Through a joint effort, the contractor contributed the expertise to develop, test, and validate
the environmental improvement and the Navy led the effort to document the change, validate
test results, and gain WSESRB concurrence following successful test and evaluation.
Teaming of contractor and government resolved a multifaceted problem. The resolution
addressed environmental, performance, and economic elements with a single solution. The
change was conceived, tested and adopted on a limited budget, with limited risk, and in a
compressed time frame of less than six months, due to synergy achieved in this combined effort
of contractor and government.
373
DETERMINATION OF ARRHENIUS KINETIC CONSTANTS

FOR DOUBLE BASE PROPELLANTS BY NON-ISOTHERMAL DSC
MEASUREMENTS. INFLUENCE OF SAMPLE SELF-HEATING.
Muhamed Sueska, Sanja Matei Muani and Maa Raji
Brodarski Institute,
Av. V. Holjevca 20, 10000 Zagreb, Croatia
Abstract:
In order to predict with required accuracy the shelf-life and thermal hazard potential of
an explosive material a true decomposition mechanism and true kinetic constants should
be known. Various experimental techniques and experimental conditions, various kinetic
approaches and data treatments procedures have been applied in order to evaluate
Arrhenius kinetic constants from experimental data as accurately as possible.
Consequently, the kinetic constants for an energetic material reported in literature may
be in considerable disagreement.
In this paper we studied the influence of propellant sample self-heating degree on the
values of Arrhenius kinetic constants, using non-isothermal DSC measurements and the
isoconversional kinetic approach described by Ozawa, and Flynn and Wall. The results
have shown that Ozawa method is capable of producing valid results for double base
propellants, provided that self-heating is minimised by adjustment of the experimental
conditions.
The activation energy was calculated to be ~173 kJ/mol in the case when sample selfeating has been avoided, while in the case of samples self-heating existence the
activation energy was calculated to be ~140 kJ/mol.
Keyword:
1.
double base propellants, differential scanning calorimetry,

kinetic constants, self-heating
INTRODUCTION
There are many reasons why the explosive community is so interested in the mechanism
and kinetics of thermal decomposition of energetic materials. From a practical point of view,
the most important are that the rate of thermal decomposition affects the quality of an
energetic material and its shelf life, as well as its thermal hazard potential, i.e. human safety
[1]
.
Some investigators have asserted that a true decomposition mechanism and true kinetic
constants should be known in order to be able to predict with required accuracy the shelf-life
and thermal hazard potential of an explosive material [2-9]. The application of an inaccurate
decomposition model and inaccurate values of Arrhenius kinetic constants in the calculation
of the thermal hazard potential of an explosive may result in highly uncertain and unusable
data on critical conditions of the self-ignition and on the shelf-life [10,11].
In order to determine Arrhenius kinetic constants as accurately as possible various
experimental techniques and testing conditions have been used. Also, various kinetic
approaches and data treatment procedures have been applied, resulting sometimes
374
in considerable disagreement in the values of the kinetic parameters reported in literature.

In some cases this disagreement is by orders of magnitude [12].
Thermal methods, particularly thermogravimetry (TG) and differential scanning
calorimetry (DSC), are experimental techniques that are frequently used in kinetic studies.
However, it should be noted that these techniques, like all other experimental techniques,
have some limitations. One of the limitations is due to the fact that thermal methods use
small samples (0.5-20 mg), which may not be representative in the case of inhomogeneous
samples [2]. On the other hand, the dynamic nature of the processes taking place during the
experiment (heat flow, gas flow, progress of reactions, interaction of the surroundings with
the sample, etc.) have the consequence that the results obtained by thermal methods are
much less material specific and dependent on the conditions used during the experiment
[13]
. Thus it must be borne in mind when performing DSC and TG experiments that the shape
and positions of DSC and TG curves depend on the sample mass and shape, heating rate,
surrounding atmosphere, sample cup, etc. [13,14]. Obviously, all of these will affect the kinetic
result [15].
However, not only the testing conditions but also the kinetic approach may play
a decisive role in evaluation of the kinetic constants. The traditional methodology of kinetic
analysis is based on isothermal experiments and on fitting of experimental data to reaction
models. The model fitting approach suffers from an inability to determine the reaction model
uniquely, and often this does not allow a reliable conclusion on the reaction mechanism to be
drawn even from isothermal data. Another serious problem in isothermal experiments is that
a sample requires some time to reach the experimental temperature. During this period
(which is non-isothermal) the sample undergoes some transformations that are likely to
affect the results of the kinetic analysis. This problem limits the use of high temperatures in
isothermal experiments [3].
A single non-isothermal experiment, typically run at a constant rate of temperature
increase, contains information on the temperature dependence of the reaction rate, and this is
the reason why it is widely believed that such an experiment would be sufficient to derive
the Arrhenius constants and reaction model of the process studied. However, the use of
traditional methodology in non-isothermal kinetics may result in highly uncertain values of
Arrhenius constants that cannot be compared with isothermal values in a simple way.
Vyazovkin and Wight discussed these problems in detail [3,16].
Another approach is based on methods that reduce the thermal analysis data (DSC
curves obtained at varying heating rates) to a set of isoconversional temperatures. Thanks to
some its of advantages, this methodology is frequently used within the explosive community
- unfortunately in some cases uncritically, i.e. not taking into account limitations of the
method and possible dependence of experimental data on testing conditions. Another serious
problem with the use of isoconversional methods is that variation of Arrhenius constants
with the extent of reaction poses difficulties in the interpretation of the kinetic data [3,16,17,18].
In our previous paper [19] we have studied influence of some testing conditions and data
treatment method on results of double base propellant Arrhenius constants determination by
Ozawas non-isothermal isoconversional methods. The results obtained have shown that
sample self-heating occurs at very slow heating rates and small sample masses, resulting in
20-30 % lover value of the activation energy. We have continued this study in order to
obtained more reliable results extending heating rates range: from very slow heating rate of
0.2 oC/min to heating rate of 30 oC/min. This paper summarises kinetic results obtained from
more than 50 DSC non-isothermal measurements.
375
1.1
Non-isothermal isoconversional method
In order to clarify what might be the cause of limitations in the application of the
isoconversional method, as well as what are the advantages of this method, one should start
from the theory of the method, laid down independently by Ozawa [20] and Flynn and Wall
[21, 22]
.
Taking that the temperature dependence of the rate constant is given by the Arrhenius
equation [ k (T ) = A exp( E / RT ) ], and that in non-isothermal experiments samples are
heated at a constant heating rate ( = dT / dt ), the basic kinetic equation may be written it the
following differential form:
d A
= exp( E / RT ) f ( )
dT
(1)
where T is temperature, k(T) is temperature-dependent rate constant, is heating rate, A

is pre-exponential factor, E is activation energy, is extent of reaction (i.e. conversion), and
f() is a function called the reaction model (i.e. the function that describes the dependence of
the reaction rate on the extent of reaction, e.g. for the first order reaction f() = (1-), etc).
After integration over the variables and T, Eq.1 becomes:
T
A
A
E
g ( ) = exp
dT = I ( E , T )
T0
RT
(2)
where I(E,T) is temperature integral.

A single non-isothermal experiment simultaneously provides the information on both
k(T) and f() [or g()], but not in a separate form. The model fitting method attempts to
determine E, A, and f() simultaneously. For this reason almost any f() can satisfactorily fit
the data (for sufficiently low values of ) at the cost of a dramatic variation in the Arrhenius
parameters [3].
Another approach to non-isothermal kinetics is the so-called isoconversional
approach. This approach rests upon the principle according to which the reaction rate at a
constant conversion is only a function of temperature [3, 16]. Several isoconversional methods
were suggested in the 1960s. The method developed by Ozawa, and Flynn and Wall starts
from the fact that the temperature integral in Eq. 2 (which does not have an analytical
solution), may be calculated by Doyles equation. Doyle [23] was the first to point out the
linearity of the logarithm of p(x) as a function of x. He proposed the approximate
relationship:
log[ p ( x )] 2.315 0.4567 x
(-20 > x > -60)
(3)
Combining Eqs. 2 and 3 one obtains:

AE
E
log() 2.315 0.4567
(4)
R
RT
It follows from Eq.4 that if series of experiments are performed at heating rates 1, 2,
3,j and if Tk,j is the temperature at which the fraction of conversion k was reached at
heating rate j, then a plot of log(j) vs. (1/ Tk,j) for each of k fractions of conversion 1, 2,
3, k will give k isoconversional lines whose slopes are:
g ( ) log
376
E
(5)
R
An accurate value of the activation energy is obtained through an iterative procedure
described elsewhere [23, 24]. Also, the temperature integral may be nowadays calculated using
methods based on numerical integration, or by using more accurate approximations such is,
for example, the Senum-Yang approximation [3, 25, 26].
slope 0.4567
Taking that the conversion at the DSC (or DTA) peak maximum temperature (Tm) is
constant (m=const.) and independent on the heating rate, Eq. 4 may be written it the
following form:
log() = 0.4567
E
AE
2.315 + log
log[ g ( m )]
RTm
R
(6)
where m is conversion at the peak maximum temperature, and Tm is the peak maximum
temperature.
It follows from the above equation that log() vs. 1/Tm will give straight line slope of
which yields the activation energy. This is the basis of the Ozawa method for the derivation
of kinetic parameters from non-isothermal DSC measurements. The Ozawas method was
the basis for the ASTM E 698-79 standard for the determination of Arrhenius kinetic
constants for thermally unstable materials [24, 27].
From the above mentioned one may note that some preconditions should be fulfilled in
order to apply the non-isothermal DSC measurements and isoconversional method,. The
most important are:
the extent of reaction at the peak maximum is constant and independent on the
heating rate
the temperature dependence of the reaction rate constant obeys the Arrhenius
equation
the reaction model should be known in order to calculate the pre-exponential factor.
There are also some general and some specific limitations regarding thermal methods
that can influence kinetic results. It is known that true kinetic results can be derived only if
there is no temperature gradient within the sample [29]. To avoid the temperature gradient
(which is connected with the heating rate, sample mass and shape, sample self-heating
during exothermal reactions, etc.), small and thin samples should be used, as well as slower
heating rates.
While many studies confirm that the non-isothermal DSC method is capable of yielding
excellent kinetic results (under optimum conditions a set of five data points over at least
five-fold range of heating rates has resulted in a standard deviation within 2 % for the
calculated activation energy [24]), it must be used with care. Just as classical analytical
methods can give erroneous kinetic results, so can the DSC method, notwithstanding its easy
use [1].
The Achilles heel of this method turns to be excess self-heating [1,19]; that is, the
tendency of energetic materials to increase the rate of heating of the sample to a greater
degree than that of the programmed rate. The main consequences of self-heating are the
following:
377
the actual (i.e. instantaneous) heating rate experienced by the sample is greater than
the programmed heating rate (see Fig. 1)
the peak maximum temperature at the same programmed heating rate does not have
the same value as the temperature obtained with no self-heating.
Temperature
T = f (t )
dT (t )
dt
Heating rate
The effect of self-heating during the exothermal decomposition is substantial deviation

of the T = f(t) curve from linearity, i.e. existence of a pronounced blip visible in Fig. 1. A
sufficiently small sample will give an essentially straight line with no evidence of selfheating, while a large sample will show behaviour given in Fig. 2. The larger the sample size
(at a given heating rate), the greater self-heating, and consequently, the error in the heating
rate (the greater deviation of the actual sample heating rate from the programmed heating
rate).
Time
Fig 1.
Effect of self-heating on actual heating rate and temperature
The final result of self-heating is a lower value of the calculated activation energy [1, 19,
28]
. Energetic materials are capable of considerable self-heating in a DSC instrument.
Although the reaction kinetics and enthalpy of reaction are obviously the root cause of selfheating, for a given material the degree of self-heating is also influenced by the heating rate
and sample size.
It is common in practice to use the non-isothermal isoconversional DSC method
applying a constant sample size method. The consequence of using a constant sample size
for all heating rates may be considerable self-heating at faster heating rates, and
consequently in an incorrect value of the activation energy calculated.
To avoid self-heating in the non-isothermal DSC method it is necessary to find a way to
measure the degree of sample self-heating and its influence on the kinetic results. In modern
DSC instruments this is possible by following the deviation of the actual heating rate from
the programmed heating rate.
To avoid self-heating, the constant peak size method was used in some studies.
According to this method, an almost equivalent degree of self-heating is present in all
measurements. To obtain a constant peak size at various heating rates, the sample size must
be scaled inversely with the heating rate, which makes this method slightly complicated.
378
2.
EXPERIMENTAL
For the study we used a double base (DB) propellant containing ~40 % of nitroglycerine.
The samples weighing 0.5-2.5 mg were cut from the strip-like propellant grains. In order to
reduce the influence of sample shape on testing results, the thickness of all samples was
maintained constant (0.15 mm).
The DSC experiments were carried out using the TA instruments DSC 2910 apparatus
that is based on the heat flux type of the cell. The measurements were done using aluminium
sample pans with perforated aluminium cover, and under nitrogen purging with 100 ml/min.
3.
In order to analyse the influence of the sample mass and heating rates range on the
degree of sample self-heating during the exothermal decomposition, i.e. on the kinetic
results, a series of non-isothermal DSC experiments were carried out.
The mass of samples in these experiments was varied from 0.5 mg to 2.5 mg, and the
heating rates from 0.2 oC/min to 30 o/min (Table 1).
A typical non-isothermal DSC curve of the tested DB propellant is shown in Fig. 2.
Along with the heat flow curve, the heating rate curve, which is important for further study,
is given. The parameters necessary for the kinetic study are assigned on the curve.
Fig 2.
Typical non-isothermal DSC curve of tested DB propellant showing heat

flow and heating rate curves (testing conditions: sample mass 2.5 mg,
programmed heating rate 10 oC/min)
Summarised experimental data necessary for studying of the kinetics in accordance with
the isoconversional Ozawa method are given in Table 1.
379
Table 1.
Sample
mass,
mg
0.5 0.02
1.0 0.02
1.5 0.02
2.0 0.04
2.50.05
Summarised experimental data obtained from non-isothermal DSC

measurements
programmed,
C/min
0.2
0.5
1
2
3
5
7
10
15
20
25
30
0.2
0.5
1
2
3
5
7
10
15
20
25
30
0.2
0.5
1
3
5
7
10
15
20
25
0.2
0.5
1
3
5
7
10
15
20
25
0.2
0.5
1
3
5
7
10
15
20
25
Parameters derived from non-isothermal DSC

measurements
Tm, oC
165.63
178.22
185.25
192.11
195.80
197.37
199.96
203.96
210.09
210.08
214.75
216.54
169.95
178.57
184.90
189.27
189.44
194.94
197.86
202.03
208.13
212.22
214.89
217.59
170.17
179.25
183.48
188.89
194.83
198.69
201.99
208.19
212.06
215.11
170.30
178.52
181.08
190.35
193.88
198.34
202.89
208.13
211.38
214.62
169.22
174.68
180.21
190.12
195.13
198.84
202.83
208.20
211.93
214.20
hm, mW
0.20
0.32
0.70
1.10
1.53
2.70
3.65
4.71
6.25
8.35
11.73
10.76
0.30
0.66
1.37
2.28
3.52
5.54
7.01
9.65
14.90
19.24
25.07
26.85
0.49
0.93
1.73
5.34
8.53
11.73
12.93
21.23
29.43
34.74
0.57
1.24
2.43
7.68
11.41
15.19
22.94
30.28
38.01
48.72
0.70
1.60
4.01
10.06
15.58
20.67
26.87
38.80
51.79
58.66
m, %
48.92
46.67
55.35
55.59
57.48
50.56
50.01
48.74
49.08
45.08
52.42
49.65
47.03
50.48
52.71
55.05
45.47
46.28
44.54
46.78
50.60
53.55
51.33
52.26
57.14
55.07
46.49
49.80
49.43
45.40
51.15
52.27
55.90
56.89
51.04
51.65
47.58
48.55
51.18
53.08
51.38
52.93
45.92
50.13
50.10
51.15
50.40
51.25
50.42
52.43
50.77
49.39
max,C/min
0.2
0.50
1.00
2.00
3.00
5.05
7.09
10.15
15.39
20.67
25.87
31.24
0.20
0.50
1.00
2.02
3.04
5.08
7.17
10.33
15.81
21.08
26.53
32.15
0.2
0.50
1.00
3.05
5.15
7.24
10.41
15.94
21.70
27.40
0.2
0.50
1.00
3.08
5.19
7.43
10.67
16.38
22.30
29.02
0.2
0.50
1.01
3.10
5.22
7.44
10.84
16.86
23.36
30.22
prog.,C/min
0.2
0.50
1
2
3
5
7
10
15
20
25
30
0.19
0.50
1.00
1.98
2.95
4.90
6.85
9.66
14.27
18.75
23.04
27.48
0.2
0.50
0.99
2.95
4.86
6.72
9.56
13.81
17.92
22.0
0.2
0.50
0.99
2.90
4.81
6.60
9.21
13.22
17.26
20.41
0.2
0.50
0.98
2.87
4.76
6.51
9.04
12.73
16.28
18.64
Legend: Tm uncorrected peak maximum temperature; hm peak height; m degree of conversion at peak
maximum temperature; max maximum heating rate at the peak region; prog. - programmed
heating rate at the peak region.
380
3.1
Influence of heating rate and sample mass on conversion at the

peak maximum
Because the first precondition for the applicability of non-isothermal DSC measurements
for the derivation of kinetic parameters according to the iso-conversional Ozawa method is
that the degree of conversion at the peak maximum temperature (m at Tm, in Eq. 2) is
constant and does not depend on the heating rate, the first step in this work was to check if
this precondition is fulfilled in the case of the studied DB propellant.
The conversion at the peak maximum was determined from the DSC thermogram as a
ratio between the peak area until peak maximum (Am) and the total peak area (Atot):
am =
Am
100
Atot
(7)
The influence of the heating rate and sample mass on conversion at the peak maximum
(m) is shown graphically in Fig. 3.
Conversion at peak maximum, %
58
m = 0.5 mg
m = 1 mg
m = 1.5 mg
m = 2 mg
m = 2.5 mg
56
54
52
50
48
46
44
42
0
Fig 3.
10
15
20
Programmed heating rate, oC/min
25
30
Dependence of conversion at the peak maximum on the heating rate for

DB propellant samples having different mass
It is obvious from Fig. 3 that the results obtained are considerably scattered - the
maximum value of conversion at the peak maximum reaches 58 %, while the minimum
value equals about 45 %. The majority of values at the heating rates between 5 and 30
o
C/min lie between 45 % and 52%, which means that the deviation from the mean value
(app. 49 %) is about 8 %.
Although it is difficult to make a reliable conclusion that there is a systematic change of
the conversion at the peak maximum with heating rates, it seems that sample mass and
heating rates do not affect significantly the conversion at the peak maximum.
However, it is obvious that the influence of the heating rate and sample mass on the
conversion at the peak maximum cannot be explained in a simple way. First of all, the fact
that small samples were used may be one of the main reasons while the results are so
scattered. This is a well-known disadvantage of thermal methods. Secondly, in interpreting
the influence of the sample mass and the heating rate one should bear in mind that two
processes affect the results - one is self-heating during the exothermal decomposition, which
381
is more pronounced for larger samples and at faster heating rates, and another is thermal lag,
which is also higher at faster heating rates and for larger samples. Although the effects of
these two processes may be opposite, both of them cause a temperature gradient within the
sample.
Another reason for the data scatter is that in all probability the DB propellant sample
decomposes in a complex manner since it is a mixture of two materials, nitrocellulose and
nitroglycerine (as well as small amounts of stabiliser, etc.). Because concurrent reactions are
not resolved, using a single set of Arrhenius constants to characterise the global kinetics of
this material is in fact an approximation to the true kinetics.
Also, with regard to the measurements of areas to obtain conversion values at the peak,
it its apparent that these conversion values are affected by the selection of beginning and
ending points of the decomposition peak, a problem that is further complicated by the fact
that these points must vary with the heating rate. This judgement factor, together with the
baseline flatness, will in turn affect the data scatter.
3.2
Influence of testing conditions on degree of self-heating
The thermal decomposition of DB propellant sample during the DSC measurements may
cause considerable sample self-heating if the sample mass is sufficiently large and the
heating rate sufficiently high. Since one of the consequences of self-heating is an increase of
the instantaneous sample heating rate to a greater degree than that of the programmed ones,
the degree of self-heating may be measured, i.e. quantified, by following an increase of the
heating rate at the peak region (see Figs. 1, 2 and 4), in respect to the programmed heating
rate. As an example two non-isothermal DSC curves of the DB propellant are given in Fig.
4, along with corresponding heating rates.
Fig 4.
A way of quantification of self-heating of DB propellant sample by

increase of heating rate at DSC peak region
382
In the case of a 2.5 mg sample and the 10 oC/min heating rate, considerable self-heating
exists, causing an increase of the heating rate at the peak region to 10.84 oC/min, while in the
case of the small sample (0.5 mg) and a slow heating rate (2 oC/min) there is no evidence of
self-heating, i.e. there is no evidence of the heating rate increase in the peak region.
Fig. 6 shows relationship between programmed heating rates and the maximal value of
actual heating rates for samples having different size. The large deviation of the maximal
value of actual heating rate from the programmed ones is for 2.5 mg sample: e.g. at 25
o
C/min programmed heating rate maximal value of actual heating rate equals 30.22 oC/min
(difference 5.22 oC/min, i.e. 21%).
35
m=0.5 mg
m=1.0 mg
m=1.5 mg
m=2.0 mg
m=2.5 mg
30
20
actual, C/min
25
15
10
5
0
0
10
15
20
25
30
35
programmed , C/min
Fig 5.
Relationship between programmed heating rates and maximal value of

actual heating rates
To quantify the degree of self-heating we used the difference between the maximum
value of the heating rate (max) in the peak region and the programmed heating rate
(programmed):
= max programmed
(8)
The values of , i.e. the degrees of self-heating, as a functions of the heating rate and
sample size are presented graphically in Fig. 6. This figure shows that at the same sample
mass, the degree of self-heating increases exponentially with the heating rate, while at the
same heating rate the degree of self-heating increases almost linearly with the sample mass.
At the same time, Fig. 6 and results given in Table 1 show that a measurable degree of selfheating exists at higher heating rates, even when a small sample size is used. For example, a
sample having 0.5 mg in mass exhibits a measurable degree of self-heating at the heating
rates above 3 oC/min, while a sample having 2.5 mg in mass exhibits a measurable selfheating even at the 1 oC/min heating rate. Such data indicate that self-heating in the case of
the DB propellant is quite large. It implies that great care should be taken in the evaluation of
the true kinetic results for DB propellants.
383
6
m = 0.5 mg
m = 1 mg
m = 1.5 mg
m = 2 mg
m = 2.5 mg
, C/min
4
3
2
1
0
0
10
15
20
25
30
programmed , C/min
Fig 6.
3.3
Influence of heating rate on degree of self-heating for DB propellant

samples having different mass
Kinetic results
In order to evaluate the Arrhenius kinetic constants in accordance with the Ozawa
icoconversional method, the data obtained experimentally by the non-isothermal DSC
measurements (given in Table 1) were treated according to the common procedure described
elsewhere [1, 11, 12]. Since the evaluation of the kinetic constants by the Ozawa method
requires precise measurement of the peak maximum temperature, the uncorrected peak
maximum temperatures given in Table 1 (Tm) are corrected for the thermal lag and heating
rate in accordance with [12].
According to the common practice in many studies, the non-isothermal DSC
measurements that are carried out at several different heating rates using samples having the
same mass were used for the evaluation of the kinetic parameters (so-called constant sample
mass method). Five series of experiments were carried out using the same sample mass and
variable heating rates. Mass of samples was in the range 0.5 mg to 2.5 mg, while the heating
rates ranges from 0.2 oC/min to 30 oC/min (Table 1). Typical DSC curves obtained in this
way are given in Fig. 7.
The consequence of using a constant sample mass for all heating rates is apparent from
Fig. 7: faster heating rates yield higher peak temperatures, while the peak height increases
proportionally with the heating rate.
The Ozawa plots, i.e. the log() vs. 1/T curves, for different sample masses are given in
Fig. 8. In order follow the degree of self-heating in each measurement, along with the Ozawa
plots the vs. 1/T curves are given.
384
Sample mass = 1 mg
Fig 7.
Non-isothermal DSC curves of DB propellant obtained at several different

heating rates using constant sample mass method (m = 1 mg)
5.9
m=0.5
m=1
m=1.5
m=2
m=2.5
1.4000
E=141,58 kJ/mol
, C/min
2.9
0.2000
1.9
-0.2000
-0.6000
-1.0000
2.0000
4.9
3.9
0.6000
ln( ), C/min
1.0000
0.9
E=172,94 kJ/mol
2.0500
2.1000
2.1500
2.2000
2.2500
-0.1
2.3000
1000/T, 1/K
Fig 8.
Ozawa plots and curves for different mass of DB propellant sample

obtained applying constant sample mass method
385
200.0
Heating rate, C/min
195.0
190.0
Point of discontinuity,oC
One may observe from Fig. 9 that the log() vs. 1/T curves for the samples weighing 0.5
and 2 mg have a pronounced discontinuity. For sample weighing 2.5 mg the discontinuity
occurs at very slow heating rate. A parallel analysis of Ozawa plots and corresponding
vs. 1/T, one may note that the discontinuity occurs at the point at which self-heating becomes
measurable ( > 0.02 oC/min). For larger samples the point of discontinuity shifts to lower
temperature (Fig. 9a). Also, the heating rate at which discontinuity occurs shifts to slower
heating rates with sample mass increase (Fig. 9b).
185.0
180.0
175.0
170.0
0
0.5
1.5
2.5
4
3
2
1
0
0
0.5
Sample mass, mg
Fig 9.
1.5
2.5
Sample mass, mg
Change of discontinuity point with sample mass (a) and heating rate at
which discontinuity occurs at different sample mass (b)
The question rises, from the Ozawa plots given in Fig. 8, which data points to use for the
evaluation of the activation energy. Although from Eq. 1 follows that only those points for
which = const., i.e. = 0 oC/min, should be taken into account for the calculation of the
true activation energies, in order to analyse influence of sample self-heating on the kinetic
results we have calculated activation energies for the following three regions,:
the region at which self-heating does not exist ( = 0 oC/min)
the region at which self-heating exists ( > 0 oC/min)
the whole region (all data points), which includes all data and the regions with and
without self-heating.
The calculated values of activation energies and pre-exponential factors are given in
Table 2. The results 2 show that the activation energies calculated for the whole region (all
data points) lie between 178.67 and 186.39 kJ/mol; the activation energies calculated for the
region without self-heating equal 160.91-178.00 kJ/mol, while the activation energies
calculated for the region with self-heating equal 137.47-166.14 kJ/mol. It should be noted
that the activation energies calculated from data points without self-heating should be, in
accordance with the theory given earlier, the true values of activation energy.
386
Table 2. Activation energies and pre-exponential factor for DB propellant

calculated applying constant sample mass method
Sample
mass,
mg
0.5 0.02
1.0 0.02
1.5 0.02
2.0 0.04
2.5 0.05
Whole range
all data points
(0.2 30 oC/min,
0 oC/min)
Range without self-heating

( = 0 oC/min)
Etot,
kJ/mol
179.51
186.39
186.38
182.40
178.67
Ewo,
kJ/mol
171.31
173.88
160.91
178.00
-
Atot,
1/s
6.351017
4.201018
4.031015
1.421018
6.101017
Awo,
1/s
5.381016
1.111017
2.901015
3.251017
-
,
C/min
0.5 - 2.0
0.2-1.0
0.2-0.5
0.2-0.5
-
Range in which self-heating

exists
( > 0 oC/min)
Ew,
kJ/mol
166.14
143.66
139.12
142.25
137.47
Aw,
1/s
2.221016
7.261013
2.091013
4.591013
1.251013
,
C/min
5-30
3-25
3-25
3-25
5-25
One may note from Fig. 8 that all data points at which self-heating exist lie on the same
straight line slope of which yields an average value of the activation energy of 141.58
kJ/mol. Similarly, all data points at which self-heating was avoided lie on the other straight
line slope of which yields an average value of the activation energy of 172.94 kJ/mol. These
data clearly show that the calculated value of activation energy of studied DB propellants
decreases for about 20 % if sample self-heating during the non-isothermal DSC
measurements is not avoided.
The fact that self-heating of tested DB propellant sample exists at relatively slow heating
rates and small sample masses, limits possible range of heating rates and sample mass that
can be applied in DSC experiments heating rate up to 3 oC/min and sample mass up to 1
mg. On the other hand, the consequence of having narrow range of heating rates is less
reliable value of the activation energy and pre-exponential factor calculated. These are some
of the reasons while the constant sample mass may not be good experimental practice in the
case of variable heating rated DSC kinetics [1].
It is obvious from the results mentioned that the calculated value of the activation energy
is greatly influenced by testing conditions first of all depends on whether or not sample
self-heating is avoided. On the other hand, sample self-heating can be avoided by decreasing
sample mass and lowering the heating rates. To check existence of self-heating one may plot
actual heating rate vs. temperature curve for each non-isothermal DSC measurement.
In order to analyse influence of self-heating on the kinetic results we introduced in this
work one more method for the evaluation of kinetic constant from the non-isothermal DSC
measurements constant degree of self-heating method. According to this method, degree of
self-heating (expressed by ) is maintained constant at different heating rates. To obtain
constant degree of self-heating at different heating rates the sample mass was scaled
inversely with the heating rate. Typical non-isothermal DSC curves of tested DB propellant
obtained according to this method are given in Fig. 10.
387
m=2.5 mg, =3 oC/min

m=2 mg, =3 oC/min
m=1 mg, =5 oC/min

m=0.5 mg, =7 oC/min
Fig 10. Non-isothermal DSC curves of DB propellant obtained applying constant

degree of self-heating method ( = 0.09 oC/min, while sample mass
and/or heating rates are different)
The Ozawa plots for obtained applying the constant degrees of self-heating method are
given in Fig. 11.
0.0000
E=162.7 kJ/mol
E=150.5 kJ/mol
log( ), C/min
1.5000
1.0000
4.8
4.3
3.8
3.3
2.8
0.5000
dHR=0
dHR=0,05-0,5
dHR=0,5-2,0
dHR=2,0-5,0
dHR=0
dHR=0,05-0,5
dHR=0,5-2,0
dHR=2,0-5,0
, C
2.0000
2.3
1.8
1.3
0.8
-0.5000
0.3
-1.0000
2.0000
2.0500
2.1000
2.1500
2.2000
2.2500
-0.2
2.3000
1000/T, 1/K
Fig 11. Ozawa plots for DB propellant obtained applying constant degree of selfheating method
388
The activation energies and pre-exponential factors were calculated for each of degrees
of self-heating. The results of calculation are given in Table 3.
Table 3. The activation energies and pre-exponential factor calculated applying
constant degree of self-heating method
Degree of selfheating
( ), oC/min
0
0.05 0.5
0.5 - 2.0
2.0 5.0
Heating rate
range,
o
C/min
0.2 3.0
5 10
10 - 30
20 - 30
Sample mass
range,
mg
0.5 - 2.5
1.0 2.5
0.5 2.5
1.0 2.5
E,
kJ/mol
A,
1/s
162.73
149.71
154.09
147.76
5.721015
3.281014
9.881014
1.791014
Table 3 shows that the activation energy obtained for = 0 oC/min (162.73 kJ/mol) is
very close to the value calculated by the constant sample mass method from data points
without self-heating (170.05 kJ/mol). The values of activation energies calculated for
different degrees of self-heating (in the range from 0 up to 5 oC/min) are very close each
other 147.76 154.09 kJ/mol. This data and the fact that all data points at which different
degree of self-heating exist lie on the same straight line on the Ozawa plots (Fig. 8) with an
average value of the activation energy of 150.5 kJ/mol, indicates that two regions can be
separated on Ozawa plot region without self-heating existence, and region with existence
of self-heating with lower value of the activation energy.
It follows from Figs. 8 and 11 and data given in Table 2 and 3, that the calculated value
of the activation energy is almost constant in the region without self-heating, afterwards it
decrease suddenly with appearance of self-heating and remains unchanged in the region with
self-heating, in spite of an increase of degree of sample self-heating up to 20 %.
4.
CONCLUSIONS
The results presented in this paper show that the non-isothermal DSC measurements and
the Ozawa isoconversional kinetic model may be used for the evaluation of kinetic
parameters of DB propellants. However, great care should be taken in order to avoid errors
that can be caused by temperature gradient within the sample. The temperature gradient
caused by sample self-heating can be avoided using proper testing conditions - a sufficiently
small sample size and sufficiently slow heating rate. In the case of tested DB propellant
sample mass should be less than 1 mg, and heating rates must be in less than 3 oC/min.
An excellent way to follow and to quantify degree of sample self-heating is by an
increase of the sample instantaneous heating rate in respect to the programmed ones. The
modern DSC instruments usually provide the ability to observe this by plotting the derivative
of the sample temperature.
Summarising all data presented, one may conclude from the Ozawa plots given in Figs.
8 and 11 that two regions with different slopes, i.e. different values of the activation energies
exist in the case of DB propellant tested. The first is region at which all data points without
self-heating belong, and for which an average activation energy of 172.94 kJ/mol was
calculated; the second is region to which all data points at which self-heating exists and for
which an average activation energy of 141.58 kJ/mol was calculated. This data show that
existence of sample self-heating during the non-isothermal DSC measurements results in
20 % lower value of the activation energy.
389
REFERENCES
[1] MCCARTY, J.: Introduction to Differential Scanning Calorimetry (effects of self-heating on
kinetics), TLN Systems Inc., Phoenix, Arizona, USA, URL:
http://www.TechLearningNow.com, 2002.
[2] BOHN, M. A.: Kinetic modelling of the ageing of gun and rocket propellants for the improved
and time-extended prediction of their service lifetime, Proc. of 1998 Life Cycles of Energetic
Materials, Fullerton, California, USAA 29 March-1 April, 1998, pp.1-38.
[3] VYAZOVKIN, S., WIGHT, W.: International Reviews in Physical Chemistry 17 (3) (1998),
407.
[4] MERZHANOV, A. G., ABRAMOV, V. G.: Propellants and Explosives 6 (1981) 130.
[5] ISLER, J., KAYSER, D.: Correlation between kinetic properties and self-ignition of
nitrocelulose, Proc. of 6th Symp. Chem. Probl. Connected Stab. Explos, Kungalav, Sweden,
1982, pp. 217.
[6] SUESKA, M.: J. Therm. Anal. Cal., 68 (2002) 865.
[7] TICMANIS, U., PANTEL, G., WILD, R., EICH, T., WILKER, S.: Simulation and verification
of exothermically reacting systems, Proc. of 33rd Int. Annual Conference of ICT, Karlsruhe
(Germany), June 25-June 28, 2002, pp.111.1.
[8] SUESKA, M.: Influence of thermal decomposition kinetic model on results of propellants selfignition numerical modeling, Proc. of. 5th Seinar New Trensd in Research of Energetic
Materials, Pardubice (Czech Republik), 2002, pp. 308.
[9] MCQUIRE, R. R., TARVER, C. M.: Chemical decomposition models for thermal explosion of
confined HMX, RDX, and TNT explosives, Report UCRL-84986, Lawrence Livermore
Laboratory, Livermore, 1981.
[10] TICMANIS, U., PANTEL, G., WILKER, S., KAISER, M.: Precision required for parameters
in thermal safety simulation, Proc. of 32nd Int. Annual Conference of ICT, Karlsruhe
(Germany), July 3 -6, 2001, pp.135.1.
[11] SUESKA, M., RAJI, M.: Change of critical condition of propellants self-ignition under
conditions of artificial ageing at elevated temperature (in Croatian), Report No. 9-2-668,
Brodarski institut, Zagreb, 2002.
[12] STANKOVI, M., KAPOR, V., PETROVI, S.: The thermal decomposition of triple base
propellants, Proc. of 7th European Symposium on Thermal Analysis and Calorimetry,
Balatonfured (Hungary), Aug. 30-Sept. 4, 1998, pp. 196.
[13] HAINES, P. J.: Thermal methods of analysis. Principle, applications and problems, Blackie
Academic & Professional, Glasgow, 1995.
[14] VAN MILTENBURG, J. C., CUEVAS-DIARTE, M. A.: Thermochimica Acta 156 (1989) 291.
[15] ESTAK, J.: Thermichimica Acta 3 (1971) 1.
[16] VYAZOVKIN, S., WIGHT, C. A.: Annu. Rev. Phys. Chem. 48 (1997) 125.
[17] SANTHOSH, G., VENKATACHALAM, S., FRANCIS, A. U., KRISHNAN, K.,
CATHERINE, K. B., NINAN, K. N.: Thermal decomposition kinetic studies on ammonium
dinitramide (AND) - glycidyl azide polymer (GAP system), Proc. of 33rd Int. Annual Conference
of ICT, Karlsruhe (Germany), June 25-June 28, 2002, pp. 64.1.
[18] SUESKA, M., MIHALI, ., RAJI, M.: Applicability of non-isothermal methods and
different kinetic approaches for description of thermal decomposition of homogeneous
propellants (in Croatian), Report. No. 9-2-250, Brodarski institut, Zagreb, 2000.
390
[19] SUESKA, M., MCCARTY, J., MATEI-MUANI, S., RAJI, M.: Influence of testing
conditions on results of Arrhenius constants determination by non-isothermal isoconversional
methods, Forum Explosivstoffe 2002, 3rd International workshop Thermoanalyse des
WIWEB, October 29-30, 2002, pp.65-87
[20] OZAWA, T.: J. Thermal Anal. 2 (1970) 301.
[21] FLYNN, J. H., WALL, L. A.: Polym. Lett., 4 (1966) 323.
[22] FLYNN, J. H.: J. Thermal Anal. 27 (1983) 95.
[23] DOYLE, C. D.: J. Appl. Polymer Sci., 6 (1962) 639.
[24] Arrhenius kinetic constants for thermally unstable materials, American Society for Testing and
Materials (ASTM) standard, E 698-79, 1979.
[25] VYAZOVKIN, S.: J. Comp. Chem., 18 (1997) 393.
[26] VYAZOVKIN, S.: J. Thermal. Anal. 49 (1997) 1493.
[27] DUSWALT, A. A.: Thermochimica Acta 8 (1974) 57.
[28] MCCARTY, J.: Self-heating errors in using ASTM method E698 for the determination
of reaction kinetics, TA Hotlinks, 1984.
[29] ROGERS, R. N.: Private communications, 2002.
391
APPLICATION OF DSC IN STABILITY STUDIES

OF DOUBLE BASE PROPELLANTS
Muhamed Sueska, Maa Raji, Sanja Matei Muani, Sanko Bakija*,
Ruica uljak*, Vladimir Jagui* and Slavko urak*
Brodarski Institute,
Av. V. Holjevca 20, 10000 Zagreb, Croatia
* Ministarstvo obrane RH,
Bauerova 33, 10000 Zagreb, Croatia
Abstract:
Thermal decomposition of propellants limits their safe and reliable service life. Under
certain conditions propellants decomposition may become autocatalytic and can lead to
the well-known phenomena of self-ignition. It is therefore of great importance to
determine accurately propellant stability at a given moment of storage time, as well as
to predict with sufficient accuracy whether or not a propellant will ignite under given
conditions.
Decomposition causes decrease of relevant propellants properties: decrease of
nitrocellulose molar mass, decrease of stabiliser content, decrease of specimen mass,
change of mechanical and ballistic performances, etc. Thus, the quality of propellants,
i.e. their stability, may be quantified and judged by measuring the changes of these
properties.
In this paper we have studied thermal properties of several stable and unstable double
base propellants in order to see if there are such changes in thermal properties that can
be used in propellants stability assessment. Thermal properties were studied applying
the non-isothermal differential scanning calorimetry (DSC) measurements.
It was found out that DSC curves of stable and unstable double base propellants,
obtained under the same experimental conditions, differ considerable. Also, it was found
out that the information obtained from DSC measurements are in reasonable agreement
with the results obtained by some other tests for propellants stability determination
the heat test at 100 oC and the stabiliser concentration determination test. These
suggest that DSC results can help in propellants stability judgment.
Keyword:
1.
thermal analysis, thermal properties, double base propellant, stability
INTRODUCTION
The homogeneous propellants are composed from substances containing nitrate esters
groups. Because of a relatively low bond energies (155-163 kJ/mol) and low value of the
activation energy (90-200 kJ/mol), such substances are capable of thermal degradation even
at moderately high temperatures [1]. Although the primary decomposition of nitrate esters
cannot be prevented, the effects of chemical degradation can be partly reduced by adding
substances called stabilisers. The stabilisers are capable of catching reactive nitrogen
oxides originated during the decomposition process.
Apart from the ageing due to the chemical reactions, there is also ageing due to physical
and physical-chemical processes. They include diffusion or migration of low molecular mass
constituents such as nitroglycerine, phlegmatisers, plasticisers, burning catalyst, etc. The
ageing of propellants due to the chemical reaction as well as through the physical effects,
392
change some relevant performances of propellants. Therefore, in order to predict propellants

service life with the required reliability, it is necessary to recognise precisely the ageing
processes, to determine their rates and to quantify them as accurately as possible.
A large number of tests have been used to assess relative propellants stability at a given
moment of time. In addition, some quantitative methods based on accurate measurements of
propellants parameters (such as stabiliser concentration, decrease of mean molar mass of
nitrocellulose, specimen mass loss, gas formation, heat generation, etc.) can be used to
predict the safe service life time [1, 2, 3, 4, 5].
All of the earliest tests for the determination of relative propellants stability (e.g. Abel
test, Vieille test, Bergmann-Junk test, Hansen test, etc.) are based on the propellants
decomposition products detection or analysis. Such tests, however, may lead to erroneous
conclusion when results are used to compare stabilities of explosive materials that
decompose to produce different products. Change in products composition may confuse
these tests [5]. On the other hand, all of these stability tests are useful and suitable to
determine the momentary quality of a propellant, but they are not suitable for prediction of
the safe service life time or the ageing behaviour at the time-temperature situations during
service. As mentioned earlier, the prediction of ageing behaviour and propellants safe service
life time require temperature-time data of performance determining quantities of the
propellants formulation, which cover a representative degree of property change [1].
The procedure for propellants safe service life time determination and stability
judgement may be different in different countries [6,7, 8, 9]. The procedure may depend also on
propellant type, presumed service safe time, storage history, etc. Typically, the first level of
propellants surveillance testing may include some simple tests (e.g. heat test and remaining
chemical stabiliser concentration determination) [6,7]. The sentencing criteria established on
the basis of such simple tests enable prediction of propellant safe service life time up to a
limited period of time [6].
The second level of propellants surveillance testing may involve the sample heating at a
relatively low temperature, during some predetermined period of time (e.g. in accordance to
NATO STANAG 4117 the sample is heated at 65.5 oC for 60 or 120 days [10]), and
determination of difference in the residual effective stabiliser concentration before and after
the ageing. This procedure enables one to assign the safe service life time up to a limited
period of time for example up to 10 years.
More definitive determination of propellant safe service life time includes multitemperature ageing studies. This procedure may be necessary for propellants having new
composition, for propellants having unknown history and ageing characteristics. In such
cases it is essential to establish the ageing characteristics in order to be able to judge on the
safe chemical service life and the frequency of periodic testing.
However, it should keep in mind that all initiations of explosives are thermal in nature.
On the other hand, none of chemical tests observe the product of chemical decomposition
reaction that causes the self-heating accidents heat. To understanding of explosives, their
ageing and their stability, requires an understanding of their thermochemical properties [5].
This explains while the Heat Flow Calorimetry and Thermal Analysis tests seem only logical
tests when the self-heating seems to be the problem.
Following this idea we have studied thermal properties of several stable and several
unstable double base propellants in order to see if there are differences between them which
can be used in the stability estimation. The thermal properties were studied by using
differential scanning calorimetry.
393
2.
EXPERIMENTAL
Five DB propellants that are identified as unstable according to the surveillance

program and sentencing criteria [8], and five DB propellants which are identified as stable,
were selected for the study. All these propellants were aged naturally under different ambient
conditions.
The stability of selected propellants was judged on the basis of the heat test at 100 oC
and stabiliser concentration test, in accordance with the procedure and sentencing criteria
described in Ref. [8].
Along with the propellants selected for the study a DB propellant of known origin,
composition and stability was tested. This propellant is assigned as reference.
The heat test at 100 oC is carried out in accordance with the procedure described
elsewhere [5, 8, 9].
The concentration of stabiliser (centralite I) was determined using UV-VIS
spectrophotometer (Varian), at wavelength of 505 nm. The stabiliser was extracted by
methyl chloride [9].
The differential scanning calorimetry (DSC) measurements were carried out using TA
Instruments DSC, Model 2910. The DSC measurements were carried out using samples
weighing 1.0 0.2 mg. The samples were tested in aluminium sample pans covered by
perforated cap, with heating rate of 5 C/min, and under nitrogen atmosphere with flow rate
of 50 ml/min.
3.

3.1
Stability tests
The stability of tested propellant was determined applying the heat test at 100 oC and the
stabiliser concentration test. The results of these tests are given in Table 1.
Table 1. Results of heat test at 100 oC and stabiliser concentration test
Propellant
US-1
US-2
US-3
US-4
US-5
ST-1
ST-2
ST-3
ST-4
ST-5
Reference
propellant
Heat test at
100 oC, hour
24
24
16
30
24
72
84
98
84
72
Concentration
of stabilizer, %
0.50
0.50
0.50
0.60
0.80
1.60
1.80
1.60
1.90
1.90
90
1.74
394
According to the heat test at 100 oC, a number of hours required to cause the liberation
of red-brown fumes of the nitrogen oxides is taken to assess propellants stability. Taking that
the stability criterion is that the time to appearance of reddish fumes is minimum 60 hours,
than the propellants assigned as US-1 to US-5 will be assigned as unstable (they did not pass
the test), while the propellants assigned as ST-1 to ST-5 will be assigned as stable.
According to the stabiliser concentration test a propellant will be assigned as unstable if
a stabiliser concentration at a given moment of time is less than 50 % of the initial stabiliser
concentration. The initial stabiliser concentration for DB propellants studied was unknown
(origin of propellants is unknown), thus this sentencing criterion cannot not be applied
reliable in this case. However, supposing that the initial stabiliser concentration was
approximately 1.6 1.9 % (typical concentration for DB propellants produced in former
Yugoslavia), than 50 % of the initial stabiliser concentration will be 0.80-0.95 %.
Consequently, propellants for which stabiliser concentration is less than 0.75 % are assigned
as unstable.
Fig. 1 shows that the stabiliser concentration test gives results that are consistent with
the heat test at 100 oC the propellants US-1 to Us-5 are identified as unstable, and the
propellants ST-1 to ST-5 are identified as stable according to both tests.
2.00
160
HT 100 oC
1.80
Stabiliser content, %
1.60
120
1.40
100
1.20
80
1.00
STABLE
0.80
60
UNSTABLE
0.60
40
0.40
20
0.20
3.2
Ref
.
ST5
ST4
ST3
ST2
ST1
US
-5
US
-4
US
-3
US
-2
0.00
US
-1
Fig 1.
Concentration of stabiliser, %
Heat test at 100 C, hour
140
Comparison of heat test at 100 oC and stabiliser concentration test
DSC characterisation
The non-isothermal DSC measurements were carried out using ~1 mg samples placed in
aluminium sample pans with perforated cover. The experiments were performed under
nitrogen atmosphere. Summarised DSC curves for all propellants tested are given in Fig. 2.
Since experimental conditions may significantly influence results of DSC analysis, the
great care was taken to keep almost identical experimental condition during all measurement
(e.g. mass of samples was held within 0.2 mg).
395
Fig 2.
Summarised DSC curves of tested DB propellants
It is evident from Fig. 2 that the shapes of DSC curves of stable and unstable propellants
are quite different. No significant difference can be observed between the curves in the early
decomposition stage (up to ~170 oC), however at a certain point the thermal decomposition
of unstable propellants becomes very rapid, resulting in significant changes of a number of
DSC curve parameters: the maximum value of heat flow (i.e. peak height) increases, the rate
of heat flow (derivative DSC curve) in the peak area increases, the peak maximum
temperature and the peak end set temperature shift to lower values, etc.
Several characteristic parameters of DSC curves were followed and analysed in order to
quantify differences in thermal behaviour of stable and unstable propellants. Along with
commonly used parameters such as the peak onset, end set and maximum temperature, the
maximum heat flow (i.e. peak height), the peak width at the half of height, and the heat of
reaction, we have analysed derivative DSC curves in order to quantify the rate of heat flow
(Fig. 3), the DSC curve integral in order to determine conversion at the peak maximum
temperature and at the point corresponding to the maximum rate of heat flow (Fig. 3), and
the actual (instantaneous) heating rate in order to quantify the degree of sample self-heating
(Fig 4). The results obtained are summarised in Table 2.
A detailed analysis of the results given in Table 2 shows that some parameters have
almost the same values for stable and for unstable propellants (i.e. extrapolated DSC peak
onset temperature), while some other parameters values may differ drastically (e.g.
maximum heat flow rates differ for about 5 times). Regarding the degree of difference, the
DSC parameters studied may be (arbitrary) divided into several groups:
1. group: there is no differences between stable and unstable propellants (Tie)

2. group: difference is up to 10 % (Tfe, Tp and max)
3. group: difference is between 25 and 100 % (Hr, hp, wp, (Tp) and (d/dt)max )
4. group: difference is more than 100 % ((d/dt)max)
396
Table 2. Summarised results of non-isothermal DSC analysis of DB propellants

Propellant
US-1
US-2
US-3
US-4
US-5
ST-1
ST-2
ST-3
ST-4
ST-5
Reference
propellant
Sample
mass,
mg
0.90
0.86
1.04
1.00
1.12
0.95
0.80
0.95
1.15
1.15
Tie,
C
Tp,
C
Tfe,
C
Hr,
J/g
hp ,
W/g
wp,
C
177.11
178.54
177.83
176.22
178.85
177.47
175.15
176.40
176.01
179.19
183.34
183.40
185.44
183.55
182.92
195.91
194.88
195.51
193.60
195.66
204.98
204.41
205.19
212.84
203.66
217.27
219.07
213.94
216.60
215.78
2038
2156
2008
2206
2292
1747
1648
1704
1571
1801
7.39
9.91
7.42
7.20
11.98
5.29
4.60
5.32
4.67
5.76
18.80
14.16
18.36
22.26
12.73
22.40
24.74
21.39
23.63
20.80
0.99
176.82
195.80
216.33
1711
5,45
22.36
Legend: Tie extrapolated peak onset temperature

Tfe extrapolated peak end set temperature
hp peak height
(d/dt)max conversion at maximum heat flow rate
max maximum actual heating rate at the peak region
(d/dt)max,
%
(Tp),
%
max,
C/min
12.04
16.94
14.26
11.64
15.61
21.99
23.00
24.43
24.40
24.60
24.40
23.99
30.00
24.98
22.21
45.65
42.22
47.61
43.81
45.24
5.31
5.43
5.48
5.40
5.94
5.08
5.06
5.09
5.11
5.14
7.67
11.12
8.32
5.57
15.67
1.69
1.33
1.61
1.55
2.14
24.68
44.81
5.11
1.70
(d/dt)max,
W/g min
Tp peak maximum temperature

Hr heat of reaction
wp peak width at the half height
(Tp) conversion at peak maximum temperature
(d/dt)max maximum heat flow rate
397
Fig 3.
Heat flow curve, derivative heat flow curve and DSC peak integral
(parameters used for analysis are assigned)
Fig 4.
Heat flow curve, temperature-time curve and actual heating rate-time curve
(parameters used for analysis are assigned)
A detailed analysis of individual parameters is given below.

The extrapolated DSC peak onset temperature (Tie) is a parameter that remains almost
unchanged with the ageing of DB propellants the mean value of Tie for unstable propellants
equals 177.7 1.0 oC and 176.8 1.6 oC for stables propellants. Because of that Tie cannot
be used in propellants stability estimation.
The peak maximum temperature (Tp), the peak extrapolated end set temperature (Tfe) and
the maximum value of the actual heating rate (max) for stable and unstable propellants differ
between 5 % and 10 %. For example, the peak maximum temperature for stable propellants
398
equals 195.1 0.9 oC, and 183.7 0.9 oC for unstable ones (difference is ~6 %), (Fig. 5),
while the peak end set temperature equals 206.2 3.7 oC for unstable, and 216.5 1.8 oC for
stable propellants (difference is ~ 5%).
Peak maximum temperature,C
198.00
196.00
194.00
192.00
190.00
188.00
186.00
184.00
182.00
180.00
Fig 5.
ST
-1
ST
-2
ST
-3
ST
-4
ST
-5
Re
f.
-5
-4
US
-3
US
-2
US
US
US
-1
178.00
Comparison of DSC peak maximum temperatures for stable and unstable

DB propellants
The actual (or instantaneous) heating rate is the heating rate experienced by the sample
(Fig. 4). The exothermal decomposition of an energetic material during the DSC
measurements may cause considerable sample self-heating if the sample mass is sufficiently
large and the heating rate sufficiently high. The consequence of self heating will be an
increase of the actual heating rate to a greater degree than that of the programmed ones [11].
The degree of sample self-heating can be controlled by the sample mass and the programmed
heating rate.
The experimental conditions which have resulted in a certain degree of the sample selfheating have been applied in DSC experiments (sample mass ~1 mg and heating rate 5
o
C/min). Table 2 and Fig. 6 show that the maximal values of actual heating rate at the peak
area (max) are larger for unstable than for stable propellants. For example, at 5 oC/min
programmed heating rate, the maximal actual heating rate (max) for unstable propellants
equals 5.51 0.25 oC/min, and 5.09 0.03 oC/min for stable propellants (difference is 6.7
%). Since the degree of sample-self heating under identical experimental conditions is
connected with the rate of sample exothermal decomposition, this means that unstable
propellants decompose more rapidly than stable ones. Although the difference between the
programmed and actual heating rates for stable and unstable propellants is relatively small,
the actual heating rate is very suitable in propellants thermal behaviour and stability
estimations since it is directly related with the rate of sample decomposition, i.e. the heat
generation rate.
399
6.20
6.00
max ,
C/min
5.80
5.60
5.40
5.20
5.00
4.80
Fig 6.
R
ef
.
ST
-5
ST
-4
ST
-3
ST
-2
ST
-1
U
S5
U
S4
U
S3
U
S2
U
S1
4.60
Comparison of maximal values of actual heating rates for stable and

unstable DB propellants (programmed heating rate for all samples
was 5 oC/min)
The majority of DSC parameters for stable and unstable propellants differ between 25
and 85 %. For example, the mean value of the heat of reaction for unstable propellants
equals 2140 118 J/g and 1694 89 J/g for stable ones (difference is 26 %) (Fig. 7); the
mean value of conversion at the point of maximum heat generation rate equals 14.1 2.3 %
for unstable propellants and 22.6 1.6 % for stable ones (difference is 68 %); the mean
value of conversion at the peak maximum temperature equals 25.1 2.9 % for unstable
propellants and 44.9 2.0 % for stable ones (difference is 79 %) (Fig. 8).
2500
Heat of reaction, J/g
2300
2100
1900
1700
Fig 7.
ef
.
R
ST
-5
ST
-4
ST
-3
ST
-2
ST
-1
S5
U
S4
U
S3
U
S2
U
S1
1500
Comparison of heats of reaction for stable and unstable DB propellants
400
50.00
45.00
(Tp ), %
40.00
35.00
30.00
25.00
Fig 8.
ef
.
R
ST
-5
ST
-4
ST
-3
ST
-2
ST
-1
S5
U
S4
U
S3
U
S2
U
S1
20.00
Comparison of conversions at the peak maximum temperature for stable

and unstable DB propellants
Fig. 2 clearly shows that the DSC peak heights (i.e. maximum value of heat flow) are
considerable higher for unstable propellants than for stable ones. This is the consequence of
faster decomposition of unstable propellants under the same experimental conditions. Also,
it is visible from Fig. 2 that the peak width decreases for unstable propellants. This means
that the peak height and peak width can be also used to differ unstable from stable DB
propellants. The results from Table 2 show that the mean value of the peak height equals
8.78 2.11 W/g for unstable propellants and 5.13 0.49 W/g for stable propellants
(difference is 71 %). Approximately the same level of difference obtains for the peak width
at the half height.
13.000
12.000
Peak height, W/g
11.000
10.000
9.000
8.000
7.000
6.000
5.000
Fig 9.
ef
.
R
S5
ST
-1
ST
-2
ST
-3
ST
-4
ST
-5
S4
U
S3
U
S2
U
S1
4.000
Comparison of peaks height for stable and unstable DB propellants
401
The DSC parameter which changes the most significantly in the case of unstable
propellants is the maximum rate of heat flow (maximum value of derivative DSC curve),
Fig. 10. It follows from Table 2 that the mean value of the maximum heat flow rate equals
9.67 3.89 J/Wmin for unstable propellants, and 1.66 0.30 J/Wmin for stable ones
difference is almost 6 times), Fig. 10.
18.000
(d /dt)max, mW/g min
16.000
14.000
12.000
10.000
8.000
6.000
4.000
2.000
ef
.
R
ST
-5
ST
-4
ST
-3
ST
-2
ST
-1
S5
U
S4
U
S3
U
S2
U
S1
0.000
Fig 10. Comparison of maximum rates of heat flow for stable and unstable DB
propellants
3.3
Correlations between DSC parameters, heat test at 100 oC and

stabiliser concentration test
If one plots difference in DSC parameters values for the reference DB propellant and for
other propellants tested, it will become visible that some general accord exists between the
results of DSC analysis, the results of heat test at 100 oC and the results of stabiliser
concentration test (Fig. 11).
The DSC parameters and the times to appearance of reddish fumes in the heat test at 100
C are very close each other for the reference and for stable propellants. On the other hand,
all of these parameters for unstable propellants differ considerable. This is visible from Fig.
11 which shows that both stability tests applied, as well as the results of DSC analysis,
identified propellants assigned as US-1 to US-5 as unstable, and propellants assigned as ST1 to ST-5 as stable.
o
Fig. 12 shows that a rough correlation may be fond between the heat test at 100 oC and
some of DSC parameters (the correlation coefficient was 0.70-0.85). However, in should be
mentioned that an establishment of possible correlation between different stability tests for
DB propellants requires a lot of work to be done on samples with different stability.
402
Peak maximum temp.
800.00
Maximum actual heating rate
700.00
Peak height
Conversion at max. Heat flow
600.00
Conversion at peak maximum

Heating test 100 oC
500.00
Stabiliser concentration test
400.00
Maximum heat flow rate
300.00
200.00
100.00
f.
Re
-5
ST
-4
ST
ST
ST
-3
-2
-1
ST
US
US
-5
-4
-3
US
US
US
-2
0.00
-1
Difference in respect to reference sample, %
900.00
Fig 11. Correlation between DSC parameters, heat test at 100 oC and stabiliser
concentration
18.000
14.000
Max. heat flow rate
16.000
Tp
196.00
% CI
2.00
194.00
1.50
192.00
190.00
1.00
188.00
186.00
0.50
184.00
Peak height
14.000
Peak height, W/g
198.00
10.000
12.000
10.000
8.000
8.000
6.000
6.000
4.000
4.000
2.000
2.000
182.00
0.00
0
20
40
60
80
Heat test at 100 oC, hour
100
12.000
0.000
0
20
40
60
80
Max. rate of heat flow, W/g min
2.50
Stabiliser concentration, %
Peak maximum temperature, oC
200.00
0.000
100
Heat test at 100 oC, hour
Fig 12. Correlation between 100 oC test and some DSC parameters
4.
CONCLUSIONS
The results presented in this paper show that DSC curve shapes and values of DSC
parameters for unstable double base propellants differ significantly compared with those for
stable propellants. The most sensitive DSC parameters to propellant age are: maximum heat
flow rate (maximum value of DSC curve derivative), maximum heat flow (peak height), and
conversion at the peak maximum. Less sensitive DSC parameters are: the maximum value of
actual heating rate, the peak maximum temperature, and the peak end set temperature, while
the peak extrapolated onset temperature remains almost unchanged with the propellant
ageing.
Thanks to that, DSC parameters obtained by the non-isothermal measurements can be
used to determine DB propellants relative stability at a given moment of time. However, it
403
should be noted that a great care must be taken in order to keep the same experimental
conditions during the DSC measurements.
It seems that a rough correlation exists between the heat test at 100 oC and DSC
parameters, although a lot of work should be done to provide reliable evidence of that
statement.
REFERENCE
[1] BOHN, M. A, Kinetic Modelling of the Ageing of Gun and Rocket Propellants for the Improved
and Time-Extended Prediction of Their Service Lifetime, Proc of the 4th Life Cycles of Energetic
Materials Conference, March 29 April 1, Fullerton, USA, 1998, p. 1-38.
[2] BOHN, M. A., VOLK, F., Ageing Behaviour of Propellants Investigated by Heat Generation,
Stabiliser Consumption, and Molar Mass Degradation, Propellants, Explos. Pyrotech., No. 17,
(1992), p. 171-178.
[3] BOHN, M. A, Kinetic Modelling of the Concentrations of the Stabiliser DPA and some of its
Consecutive Products as Function of Time and Temperature, J. of Therm. Anal. Cal., No. 65,
(2001), p. 103-120.
[4] BOHN, M. A, Prediction of Life Times of Propellants Improved Kinetic Description of the
Stabiliser Consumption, Propellants, Explos. Pyrotech., No. 19, (1994), p. 266-269.
[5] ROGERS, R. R., ROGERS, J. L., Explosive science, Los Alamos, NM,
http://home.att.net/~mrogers, 2002.
[6] DODDS, J. S., Current UK procedure for propellant shelf-life determination, Proc. of the 9th
Symp. Chem. Probl. Connected Stabil. Explos., Margretetorp, August 23-27, 1992, pp. 247-258.
[7] NYBERG, H., Stability studies of NC propellants in the Finnish defence forces, Proc. of the 10th
Symp. Chem. Probl. Connected Stabil. Explos., Margretetorp, May 25-June 1, 1995, pp. 205210.
[8] Cjeloviti nadzor kvalitete ubojnih sreddstava (prijedlog), MORH, 2001.
[9] Praenje kemijske stabilnosti baruta i raketnih goriva, HRVN 8069/91, 1992.
[10] Stability test procedures and requirements for propellants stabilised with diphenylamine, ethyl
centralite, or mixture of both, NATO STANAG 4117.
[11] SUESKA, M., MATEI MUANI, S., RAJI, M., Determination of Arrhenius kinetic
constants for DB propellants by non-isothermal DSC measurements. Influence of some testing
conditions, Proc. Of 6th Int. Seminar New trends in research of energetic materials, Pardubice,
2003 (to be published).
404
MEASURING OF THE BLEVE INTERNAL PROCESSES

Jaroslav vihovsk, Betislav Janovsk and Radovan Skcel
Department of Theory & Technology of Explosives,
University of Pardubice, CZ-532 10 Pardubice, Czech Republic
Abstract:
In the present context, the term BLEVE is used for a sudden release of liquid with
temperature higher than its boiling point from the container at the failure time, with a
total collapse of the container. Most theories about the formation of BLEVE are based
on evaporation dynamics and pressure changes inside the container. In the available
theories, however, there is a gap in clarifying the initiatory process mechanism, which is
going on during and immediately after the collapse of the container. The upcoming
experimental works are conducted with the intention to unify the notions of this
initiatory process. In the text of this contribution are described apparatuses, which were
used during researches on BLEVE problems and which became a basis for the proposed
apparatus. The proposed apparatus, together with the matrix of the experiment, is
mentioned at the end of this contribution.
Keyword:
1.
depressurization, metastable state, blast wave propagation, BLEVE
INTRODUCTION
BLEVE (Boiling Liquid Expanding Vapour Explosion) is one of the most serious types
of physical demonstrations, which can be encountered not only in the chemical industry, but
mainly during the transportation of dangerous substances and potentially also in common
life when using household tanks and bottles with propane-butane. In the present context, the
term BLEVE is used for a sudden release of liquid with temperature higher than its boiling
point from the container at the failure time, with a total collapse of the container. Most
theories about the formation of BLEVE are based on evaporation dynamics and pressure
changes inside the container. These theories agree on the general description of the process,
when, after a sudden depressurization inside the container with superheated liquid, its
intensive evaporation occurs. The consequent fast expansion of this metastable phase
(liquid/vapour), enabled by the collapse of the container, has the main share in the formation
of BLEVE. In the available theories, however, there is a gap in clarifying the initiatory
process mechanism, which is going on during and immediately after the collapse of the
container.
One of the most compact theories was created by Venart[1,2] on the basis of the
conducted experiments. In his works he suggested a division of this kind of accident into
three basic types (diphasic jet, BLEVE, BLCBE). He described the significance of STL
(Superheat Temperature Limit) and stages of fulfillment for this type of accident. In case of
BLCBE he assumed a formation of homogeneous nucleation and a consequent collapse of
the formed bubbles inside an essential part of the liquid capacity. The experiments he
conducted proved that under certain conditions, with the collapse of the container, BLCBE
can occur even if, under the given pressure, the liquid temperature is below the value of
critical superheat (STL). Another interesting findings about this problem were introduced by
405
experiments conducted by Bartk[3], who gave his attention mainly to the inner mechanism
of the depressurization process. He assumed a scheme where the depressurization of the
container is stopped when the crack is filled with growing bubbles, which arise from the
homogeneous nucleation. He discovered the significance of the initiatory temperature and of
the depressurization speed for the whole process of degradation of the container. He also
gave attention to the behaviour of the dilution wave and to the influence of the liquid purity.
Gelfand[4] was also engaged in studying the behaviour of shock waves with rapid leakage of
boiling liquid. From the pressure histories that he obtained during his experiments he found
out a difference between cases when the failure of the container occured under or above the
liquid level. He described a dependence of the generated impulse and pressure on the
medium characteristics (mass, molecular weight) and conditions (pressure, temperature)
inside the container at the moment of its failure.
2.
BASIC APPARATUS USED AT BLEVE EXPERIMETS
As mentioned above, Venart suggested a division of this kind of accidents into three
types, according to different schemes of the process going on inside the container. The first
and, with regard to physical demonstrations the weakest one, is the diphasic JET. It is in
essence a spurt of the container capacity through a bounded aperture without a split or a
larger deformation of the whole container. The second type, termed as BLEVE, is a failure
of the container, caused by a sudden expansion of the superheated container capacity. The
expansion of the medium is not too rapid and sometimes does not even split the container
into more parts, yet it totally damages it. The term BLCBE (Boiling Liquid Compressed
Bubble Explosion) was suggested by Venart to name the process, which leads to the
degradation of the container with the most intensive demonstrations (total lost of the
container and generation of shock wave with high overpressures).
The experiments were carried out on the Pop Can apparatus, whose scheme is in fig. 1. It
consisted of a pressure tank with an inner pressure sensing, heat (or mechanical) installation
for the violation of the container and instruments for recording pressure, temperature and
visualisation. After the filling and pressurization, the process was initiated by short PRV
(Pressure Relief Valve) action and after repressurization the container was broken up either
by mechanical or heat impulse (creep) and consequent demonstrations were recorded. For
the experiments, three different media (Ar, water and R11) were gradually used.
On the basis of experiments conducted in that way, Venart compiled the following
scheme of the process going on during BLEVE. At the moment of the failure of the
container above the liquid level, there occurs an immediate leakage of the steam phase into
the surroundings and pressure inside the container decreases. The liquid phase is for a very
short period of time in a metastable state (pressure inside the container is at the given
temperature lower than the corresponding pressure according to the saturation curve). The
pressure continues to increase by reason of bubble growing, but its progress is not continual
and reaches the maximal value in cycles, as can be seen in fig. 2. He attributes the
consequent rapid expansion to the collapse of bubbles (diameter 1-3 mm, formation in 80
s). In his further work[2], Venart gave attention to the decompression wave, which runs
through the liquid during the failure of the container and spreads at the speed of sound at the
given space. Venart considered it to be the cause of the formation of bubbles and their
propagation into the whole container capacity. Finally, he compiled a mathematical model of
the bubble growing speed and described its relation to process conditions (pressure,
temperature).
406
Fig 1.
Detailed diagrammatic view of

Pop-Can BLEVE and piece and
instrumentation connection[1]
Fig 2.
Expanded time scale pressure

trace [1]
Bartk[3] adopted for his measurings conclusions of Alamgir and Leinhard[5,6] (1978-81):
Depressurization is stopped abruptly by an explosive bubble nucleation process

causing, in its turn, a certain increase in pressure. After that, pressure remains
almost constant for a relatively long period of time (100ms)
The pressure undershoot shows a strong dependence on the initial temperature of

the coolant and weaker dependence on the depressurization rate.
The propagation velocity of the depressurization wave does not differ from the
local sound velocity in the liquid.
The purity of the liquid does not significantly affect the pressure undershoot.
He carried out his experiments on an apparatus, which consisted of a horizontally placed

tube with dynamic pressure sensors and thermocouples (fig. 3). Water was used as a
medium. Initiatory temperatures and pressures were used in the range 130-300C, resp.
8-12,5 MPa.
Fig 3.
The horizontal experimental channel[3]
The experiments he conducted confirmed the expected scheme (filling of aperture with
bubbles, formation of the metastable phase and the consequent relatively long increase of
pressure) for temperatures above 240C. He discovered a different behaviour of the reflected
407
wave with temperaturess below 220C and above 260C. He admits that the process going
on inside the container is by its behaviour very similar to homogeneous nucleation. The
author further states the primary significance of the initial temperature and the
depressurization speed for the whole process, and the fact that with the increase in the initial
temperature during depressurization the number of nucleuses in liquid also increases.
Gelfand in his work pursued mainly the behaviour of shock waves, which are generated
during the rapid expansion of the boiling liquid. He carried out his experiments on an
apparatus (fig. 4) consisting of a vertically oriented shock tube with a heated HPC (high
pressure chamber) and equipped with pressure sensors all over the LPV (low pressure
chamber) length. Two basic experiments were carried out under the same conditions and
with the same liquid (F-113). The experiments varied only in the position of HPV (high
pressure volume), i.e. in the second experiment the liquid phase was diretcly in contact with
the diaphragm. Thus, this setup, in contrast to the first case, led to the possibility of a
diaphragm disruption under the surface of the liquid phase.
Fig 4.
The scheme of experimental setup[4]

. 1 - HPV (2 - highting bands, 3 thermocouple, 4 HC heating
chamber), 5 - LPV (6 - bursting
diaphragm, 7-10 pressure gauges,
11 photo gauge).
Fig 5.
Pressure histories[4] , Liquid- Freon113, m = 12g, V = 26 cm3.
In case of the diaphragm disruption above the liquid surface, an initial shock and the
propagating blast wave without shock front were recorded (fig. 5A). The author explains that
the formation of the main shock is caused by the rapid expansion of the steam-saturated
metastable phase, and the consequent notably slower increase of pressure as an expansion of
the erupted cloud of droplets of the superheated liquid in LPV.
In the second experiment, the diaphragm was disrupted under the liquid surface. As can
be seen in fig. 5B, the pressure behaviour differs in the initial stage from the first case. The
expansion of the superheated liquid was accompanied by a blast wave without the shock
front. So, it means that only a quick eruption of the superheated liquid occured.
From the difference between these two cases can be clearly seen how important and
significant role, at this event, has the liquid evaporation and how this gaseous phase
contributes to the whole process. In further series of experiments he gave his attention
mainly to the evaluation of amplitude and impulse of the evoked shock wave and its relation
to the initiatory conditions. He discovered that with an increasing amount of substance
408
taking part in the event the wave impulse (operating period) increases, while the amplitude
remains the same. But the amplitude increases with an increasing initial pressure, yet the
liquid characteristics do not influence it markedly.
3.
DESIGN OF THE APPARATUS
From the above mentioned works it is clear that so far there is no compact opinion of
what process leads to the container disruption. Our intention is to describe and represent the
processes going on inside the container before and after the container disruption and to
measure pressure effects under different conditions and schemes of the process. To achieve
this goal, we want to construct a special apparatus, whose design is shown in Fig. 6, with the
aid of which we could measure pressure effects inside and outside the high pressure part and
observe the processes going on in the inner part of the container.
Fig 6.
Design of the experimental setup: 1 HPC- high pressure chamber,

2 bursting diaphragm, 3 LPV low pressure chamber, 4 electrical
heating band, 5 piezo-electric pressure gauge, 6 pressure relief device
(solenoid valve), 7 endoscope, 8 filling opening with screw and reling,
9 thermocouple.
The apparatus consists of three basic parts of cylindrical shape: the middle part labelled
as HPC (1 - high pressure chamber) and then two identical parts at both sides of HPC
labelled as LPV (3 - low pressure volume). HPC and LPV are separated from each other by
diaphragms (2 - bursting diaphragm). HPC is equipped with electrical heating band (4),
solenoid valve (6), filling opening with screw and reling (8), endoscope with a recording
device and thermocouples for measuring temperatures of both steam and liquid phase. All
three basic parts contain pressure sensors (5 - piezo) in their configurations. As is already
clear from the apparatus construction, there should be three basic outputs from the
measurings: pressure and temperature measured in different parts of the apparatus and a
visual record of the inner part of HPC. From these pressure and temperature behaviours we
want to, by combining the measurings, reach as accurate notions of the whole process going
on inside HPC as possible.
409
4.
TEST MATRIX
The set of measurements was suggeted by Janovsky[7] for these experiments.

Fundamentally, it is based on variation of following initial conditions:
1. Size and shape of initial opening;
2. Initial temperature ~ initial pressure
3. Pre-nukleation
4. Type of liquid
5. Volume of vapour space alone
The intention of phase 1 is to make the first step to obtaining adequate experimental data
to prove basic assumption and equations used for simulation of blast wave generation during
BLEVE effect in 1-D geometry described in Skacel[8]. The intention of phase 2 is to confirm
the results obtained during the phase 1 and extent the database created by small-scale
experiments. The evaluation of the experimental data will reveal whether the theory of
collapsing bubbles is useful for simulation of BLEVE blast wave generation or not.
5.
CONCLUSIONS
Process of BLEVE was generally understood. Literature search shows equations solving
particular problems but deeper understanding of the homogeneous nucleation and cavitation
processes is necessary. It seems to be possible to simulate blast wave generation process
with help of homogenous nucleation theory and theory of collapsing bubbles. The general
aim is to understand the processes before and after crack development.
6.
REFERENCES
[1] VENART, J.E.S., RUTLEDGE, G.A., SUMATHIPALA K., AND SOLLOWS, K.,: To BLEVE
or not to BLEVE: Anatomy of a Boling Liquid Expanding Vapor Explosion, Proceedings of 7th
International Symposion on Loss Prevention and Safety Promotion in the Process Industries,
1992, Vol. 4, pp. 115/1-21
[2] YU, C.M., VENART, J.E.S.: The boiling liquid collapsed bubble explosion (BLCBE): A
preliminary models, Journal of Hazardous Materials, 1996, Vol. 46, 197-213
[3] BARTK, J.: A Study of the Rapid Depressurization of Hot Water and the Dynamics of Vapour
Bubble Generation in Superheated Water, Int1 Journal of Multiphase Flow, 1990, Vol. 16:5,
pp. 789-798
[4] GELFAND, B., MEDVEDEV, S., POLENOV, A., TSYGANOV S.: Shock waves by sudden
expansion of hot liquid, Proceedings of 7th International Symposion on Loss Prevention and
Safety Promotion in the Process Industries, Taormina, Italy, 1992, Vol.2. 89-1/89-16
[5] ALAMGIR, M. AND LIENHARD, J. H.,: Correltion of pressure undershoot during hot water
depressurization, ASME Jl Heat Transfer, 1981, Vol. 103, 52-55
[6] ALAMGIR, M. AND LIENHARD, J. H.,: An experimental study of the rapid
depressuratization of hot water, ASME Jl Heat Transfer, 1980, Vol. 102, 433-438
[7] JANOVSKY B.: BLEVE blast wave generation process, TNO Prins Maurits Laboratory, 2001,
report No. 2001- UPCE-TNO
[8] SKACEL, R., JANOVSKY, B., SVIHOVSKY, J.: Processes in the vessel during BLEVE effect,
6th International Seminar New Trends in Research of Energetics Materials, Univ. Pardubice,
April 2003
410
FLAME AND DETONATION ARRESTERS EUROPEAN

NORMATIVE STANDARD FOR TESTING
A. Teodorczyk* and G.O. Thomas**
*Warsaw University of Technology,
Nowowiejska 21/25 00-665 Warszawa, Poland
**Centre for Explosion Studies,University of Wales,
Aberystwyth,Ceredigion, UK SY23 3BZ
Abstract:
This paper describes the European normative standard developed on flame and
explosion protection devices destined for use in the European community. The new
standard covers the design, constructional and performance aspects, test methods and
limits of use of flame and explosion arrester devices. In the present paper, to illustrate
why a range of tests are required under such a standard, the various stages of explosion
pressure events that can arise in pipelines containing potentially explosive mixtures are
first summarized. This is followed by a discussion of the tests included in the standard to
ensure that an explosion arrester protection device is fit for purpose, with particular
attention paid to detonation testing. A brief comparison with other existing national
standards is also presented.
Keyword:
1.
normative standard, flame and explosion protection, testing
INTRODUCTION
Flame and detonation arresters provide a low cost and widely used method for the
suppression of explosions in pipelines that transport reactive gases or vapors or where
explosive mixtures may be generated accidentally. They are also attractive from an
operational standpoint as they are passive devices and do not require sophisticated control
and activation electronics. Applications are widespread, especially in pipelines that transport
or may accidentally contain reactive gases or vapors. Arresters are effective because each
element is usually comprised of many narrow channels, whose widths are so small as to
extinguish combustion as heat energy is lost through the element.
They are already widely used in the petrochemical industry to protect plant against
accidental explosions propagating along pipelines between chemical plant. In recent years
there has been renewed interest in their use, due to the introduction of environmental
legislation that limits the discharge of hydrocarbon vapors to the atmosphere. The
introduction of vapor recovery systems, initially for marine ship-to shore application, has
increased the risk of developing flammable mixtures.
Surprisingly, little detailed attention has been given to the mechanisms by which
arresters operate. Similarly, it is somewhat surprising, given their widespread use, that the
procedures for evaluating arresters are not that technologically well advanced. Current
procedures for testing flame and detonation arresters performance differ in many countries.
411
The European Standard EN 12874 directive for Flame arresters performance

requirements, test methods and limits for use has introduced standard normative testing
procedures. The outcome of applying these standard procedures contributes to determining
whether protection devices designed, manufactured and tested in accordance with the
standard are, when sold commercially, fit for purpose.
The wide range of test procedures required to cover the diverse range of devices
included under the terms of the directive were developed by a number of working groups of
experts in the various technical and scientific fields of interest. In the present paper the main
scientific issues raised during the development of the new normative standard for the
experimental testing of flame and explosion arrester protection devices for use in flammable
atmospheres are discussed. However, before outlining the key elements of the test
procedures adopted by the working group, it is first instructive to review some of the
processes that can arise during pipeline explosions and that could thus threaten practical
explosion arrester devices.
2.
EXPLOSIONS IN PIPELINES
Explosion events in pipelines, initiated by low energy sources, first propagate as slow
deflagrations where the flame velocity relative to the pipe is small (<< 300m/s). During this
stage the gas flow displacements ahead of the exothermic flame or reaction front are
relatively small, the pressure increase ahead of the flame is correspondingly low, less than
one bar, and no shock wave is formed.
As the flame front velocity and rates of energy release increase, due to increased rates of
turbulent burning resulting from the interaction of the flowing gas with the pipe walls, a
shock wave is formed ahead of the flame front and the overall explosion process accelerates
further. This acceleration is caused by the positive feedback between the combustion front
and the gas flow ahead of the flame. The feedback arises via flow induced turbulence and
subsequent increased turbulent mass burning rates. It is generally acknowledged, but not
fully quantified, that further escalation of the explosion process can occur as the velocity of
the shock-flame complex approaches ca. 1000 m/s. Localized explosions now develop, with
significant transient overpressures in some cases. This is termed the deflagration to
detonation transition or DDT phase. Following this transition the detonation wave moves
faster than that predicted from simple steady state theory but is always decelerating and
eventually reaches the steady or Chapman-Jouguet state. For an explosion in a pipeline, each
of these stages could potentially develop at a location where a detonation arrester is located.
Steady state detonation are a special form of propagating supersonic combustion waves
characterized by the constancy of their velocity, whose magnitude can be readily computed
from basic thermodynamic data pertaining to the initial gas mixture. Detonations are
potentially damaging as the pressures can increase to at least 18 times the ambient pressure.
During the process of transition from a deflagration to a detonation the transient pressures
can reach 100 times the ambient pressure, so the potential for very damaging localized
overpressures is significant. During the period immediately following transition, and before
the detonation wave relaxes to its steady state velocity and pressure, the wave is described as
being overdriven. During this phase, both peak pressures and local wave propagation
velocity are greater than the theoretical steady state values.
The various phases during an accelerating explosion event in a pipeline explosion
initiated by a low energy ignition source may therefore be categorized in general terms as
follows:
412
1. Slow deflagration (SDEF)

2. Fast deflagration (FDEF)
3. Overdriven or transition to detonation (DDT)
4. Steady CJ detonation (CJDET)
Examples of pressure histories corresponding to each of these phases are shown in Fig. 1
(a)-(d), obtained during flame acceleration tests, some of which were performed with an
explosion arrester device in position. Explosion development in pipelines is known to be a
stochastic process and the pressure evolution is seldom repeated in multiple tests. However,
Fig. 1 does serve to illustrate the general forms of the pressures developed and their
respective magnitudes.
Figure 1(a) shows a slow deflagration which exhibits a gradual pressure rise with a
maximum over pressure below 1 bar. Figure 1(b) is a pressure record from a fast deflagration
test where a pre-cursor shock has formed. The flame front arrival at the pressure gauge is
coincident with the second pressure rise at ca. 55.5 ms. The third pressure change, at some
56.5 ms, is due to the entry of the flame into a flame arrester housing, where a localized
explosion sends a pressure wave back towards the ignition point. The pressure gauge in this
test was located some 100 mm before the arrester flange. Figure 1(c) shows the pressure
histories typical of an overdriven detonation immediately following a transition to detonation
event. The peak pressures are several times greater than the pressures observed for a steady
detonation wave, shown in Figure 1(d).
3.
CURRENT PROCEDURES FOR EXPLOSION ARRESTER TESTING
A number of test standards exist world-wide. The following lists some existing
standards: UK BS7244 [1], US Coast Guard [2], International Maritime Organisation (IMO)
[3]
, Canadian Standard [4]. They outline procedures for assessing the response of arresters
when subject to flames and detonations. To qualify, these require that the arrester withstand
a set number of flame or detonation impacts without combustion propagating beyond the
arrester. Recent studies have indicated that the procedures outlined in the current standard do
not result in as reproducible a test as one might expect.
For flame arresters, the basis of any test evaluation is the determination of the maximum
length of pipe, placed before the arrester, for which a propagating flame does not pass
through the arrester over ten repeated tests.
The UK BS7244 standard assumes implicitly that a transition to detonation event is the
most threatening and the standard involves an initial test procedure where the transition point
along the test pipe is identified. The pipe length is then adjusted until the arrester is located
at the transition point. The arrester must then survive ten repeated tests at that location. The
basic approach adopted is to test against the violent events that arise during transition to
detonation. The weakness of the standard is that it does not stipulate explicitly that these
conditions are reproduced during the subsequent testing and it is quite possible that the
actual conditions are different, a steady detonation or a fast deflagration. This can arise
because the flame acceleration process that gives rise to the transition interacts with the
arrester. Moving the arrester (or changing the pipe length) can thus move the transition point
as well.
The US Coast Guard standard requires that five steady detonations are tested, as well as
five overdriven detonations, although as with BS7244 the definition of overdriven is far
from specific. The US standard also requires that five each of both slow and fast
413
1.0
15
0.8
12
Overressure (barg)
Overressure (barg)
deflagrations are tested, and these are specified by applying pressure bounds. The ratio of
explosion pressure to the initial pressure must be less than 1.0 for slow deflagrations and
between 1 and 10 for fast deflagrations. The standard is in fact more stringent in that it
requires that the fast and slow deflagrations are repeated with a significant flow restriction
on the protected side. The use of a restriction simulates a partial obstruction on the protected
side. In practice the effect is that, in some tests, this leads to an additional increase in the
local pressure and temperature at the element, and this may influence the arresters ability to
function as the effective MESG of the mixture changes as the local physical conditions of
pressure and temperature vary.
0.6
0.4
0.2
0.0
15
25
35
45
55
65
75
Time (ms)
85
95
0
50
51
52
53
54
55
56
57
58
59
60
61
62
10
11
12
13
Time (ms)
b)
a)
75
20
60
16
Overressure (barg)
Overressure (barg)
45
30
12
15
0
9.0
10.5
c)
12.0
13.5
15.0
16.5
18.0
Time (ms)
19.5
21.0
22
Time (ms)
d)
Fig 1.
Typical pressure histories showing a) slow deflagration, b) fast

deflagration, c) overdriven detonation and d) steady detonation.
414
4.
THE NEW EUROPEAN STANDARD

4.1
Scope of the standard
The standard specifies requirements for flame and explosion arresters designed to
prevent flame transmission in and between sections of pipes or vessels when flammable
gases and vapours are present. It specifies test methods to verify the proper operation of the
devices.
The standard further specifies maximum operating pressures and temperatures where
devices tested under the standard may be used. For example, arresters are only deemed to be
suitable for use at pressures less than or equal to the pressure at which they are tested.
4.2
Classification of explosion groups
Explosion devices destined for use in operating process plant may be required to operate
with a range of gas and vapour mixtures. The reactivity or potential hazard afforded by any
individual mixture may be quantified by protection device manufacturers by establishing the
Maximum Experimental Safe Gap (MESG) of the process mixture. The MESG of a mixture
is the maximum gap of the joint between two chambers that prevents transmission of an
explosion from one chamber to the second. The length of the joint is set at 25 mm. The
MESG standard then provides an explosion group classification, with different groups
delineated by MESG. In general terms, the more reactive a mixture the lower is its MESG.
Standard test mixtures, of propane, ethylene and hydrogen are used as specific
representative test gas mixtures for each explosion group in the new flame and arrester
standard, as shown in Table 1. Once tested with the appropriate test mixture for an explosion
group, a device may be used with any mixture (in any explosion group) whose MESG is
greater than the lower MESG bound specified for the explosion group tested.
Table 1. Classification of gas explosion groups for deflagration and detonation
tests in terms of MESG of the mixture together with specified test gas
mixtures for each explosion group.
Explosion Group
MESG of
mixture (mm)
Test gas
Test gas in air

(%)
II A
> 0.9
Propane
4.2 0.2
II B1
0.85
Ethylene
4.8 0.2
II B2
0.75
Ethylene
5.4 0.2
II B3
0.65
Ethylene
6.5 0.2
II B
0.50
Hydrogen
45.0 0.2
II C
< 0.50
Hydrogen
28.5 0.2
One practical testing problem recognized in the standard arises because it becomes more
difficult to generate explosions and detonations in mixtures as the mixture MESG increases.
This is increasingly problematic as the nominal pipe diameter is decreased (the propensity
for explosion acceleration and onset of detonation is linked to reactivity and pipe size). The
415
standard thus allows a mixture from another explosion group to be substituted if any
difficulties are encountered when testing with a low reactive mixture.
The standard also tabulates specific gas mixtures that are to be used for each explosion
group application for both short term and endurance burning tests.
4.3
Devices covered
The standard includes specific test procedures for a range of devices. The most common
devices covered in the new standard and the protection they offer are listed in Table 2. In
addition, the standard also provides a procedure for testing flame arresters fitted to
compressors, fans and vacuum pumps as well as high velocity vents, designed to limit flame
propagation by controlling the magnitude of the vented gas velocity. The standard also
recognized devices manufactured with elements that are measurable as well as those formed
from random structures that cannot be reproduced in technical drawings.
Table 2. Definitions of some common explosion arrester devices and the
explosions they are designed to protect against.
Device
Device designed to protect against:
End of line deflagration
Unconfined deflagration propagating into an enclosure or

vessel
In-line deflagration
Confined deflagration propagating further along a pipe
Pre-volume
Deflagration propagation from a volume (length/diameter < 5)

to the outside atmosphere or connected apparatus
In-line stable detonation
An established steady detonation or residual combustion from

propagating further along a pipe
In-line unstable detonation
A deflagration, or an overdriven detonation, or a steady

detonation, as well as any residual combustion, from
End-of line detonation
An established steady detonation or residual combustion from

Liquid product arresters
Explosion propagation through a liquid seal, formed by the

liquid product normally flowing through the device
Hydraulic flame arresters
Explosion propagation through a water seal that breaks normal

product flow into discrete bubbles
Excluded from the standard are devices for use with carbon disulphide and gases or
vapours that tend to self-decomposition, or which are chemically unstable. The standard does
not apply to flame arrester devices for use on internal combustion engines. Finally, the
standard does not cover devices intended for use in atmospheres where the oxygen
concentration is greater than that found in fuel-air mixtures, i.e. it is not applicable for
oxygen enriched gases and vapours.
As can be seen from Table 2, the standard covers a number of devices and applications
and it is not surprising that a range of tests are detailed in the standard. These tests can
however be grouped within one of four general classes. Two are concerned with preventing
the propagation of two explosion waves, deflagration and detonation waves. The third is
concerned with tests of the effectiveness in providing protection against flame transmission
in stabilized or endurance burning tests. Finally, procedures for determining flow
measurement and pressure drop characteristics are described. Of these four, it is the testing
416
for performance against unstable detonation which proved the most contentious and difficult
to formulate. This is a consequence of the variability in the temporally and spatially resolved
explosion pre-cursor events that lead to transition to detonation and the resulting difficulty in
quantifying key parameters such pressure and velocity.
4.4
Deflagration and detonation test procedures
4.4.1. Deflagration
The various test procedures used for deflagration testing are relatively straightforward
and in essence they all describe constraints on the pipe lengths or volumes before or after a
device. The test mixture is ignited in one of these volumes, termed the unprotected volume,
and the ability of the device to limit flame propagation into the second or protected volume
is observed. Certain parameters pertaining to the explosion behaviour in the unprotected
volume are then reported. For example, for an in-line deflagration device, six repeat tests
must be performed where no flame transmission occurs. The flame speed and maximum
pressure in the unprotected side under the final conditions must be reported.
4.4.2. Steady detonation
A steady detonation is defined within the standard as one whose velocity lies within
limits derived from theoretical calculations of detonation velocities, and that also exhibits a
very rapid pressure increase at the leading wave front.
One complication does however arises as the nominal pipe diameter to which the device
is to be fitted is decreased. It is known that there are certain limits below which steady
detonation cannot be sustained in smaller pipe diameters. Close to these propagation limits
the detonation wave may be decelerating slowly. To provide an additional check that the
wave used in testing is a true steady detonation the standard requires that the pressure at the
wave front, averaged over a period of 200 s after the arrival of the main waveform, Pmd,
must lead to a ratio of Pmd/Pi that exceeds tabulated values, Ptd/Pi. These tabulated threshold
values of pressure ratio, Ptd/Pi, vary with nominal pipe/device size and the threshold
pressures increase with increasing pipe size, reflecting the decreasing energy and momentum
losses suffered by the detonation at larger nominal pipe diameters. Pi is the initial pressure of
the test mixture before it is ignited. The specified values of Ptd/ Pi are shown in Table 3 for a
range of nominal pipe diameters. Figure 2 (a) shows a typical detonation wave, with the 200
s time marked by a vertical bar for a wave that just meets the condition Pmd/Pi >14 from a
test in a 150 mm diameter pipe. The device must then prevent flame propagation in three
tests that meets these conditions.
In addition to the steady detonation tests, the device must also successfully prevent
flame transmission in a further three deflagration tests. In these specific tests the pressure, Pe,
at the time the flame arrives at a designated pressure transducer location must lie within the
limits 2.0 < Pe/Pi < 0.8Ptd/ Pi.
417
Table 3. Ratio of Ptd/Pi for various explosion groups and pipe diameters D.
Explosion Group
D 80
80 < D 150
150 < D 400
D 400
II A
10
12
14
16
II B1
10
12
14
16
II B2
10
12
14
16
II B3
10
12
14
16
II B
10
10
12
II C
20
75
16
60
Overressure (barg)
Overressure (barg)
The choice of a 200 s as the period over which the pressure is averaged is to some
extent arbitrary but this specific value was chosen such that, when combined with the
specification of the pressure gauge location, it ensures that no reflected pressure waves from
an explosion within the arrester housing are included in the averaging process.
12
45
30
15
0
2.0
2.1
a)
2.2
2.3
2.4
2.5
2.6
Time (ms)
2.7
0
14.8
2.8
14.9
15.0
15.1
15.2
15.3
15.4
Time (ms)
15.5
15.6
15.7
b)
Fig 2.
Typical overpressure histories showing a) a steady detonation and b) an

overdriven detonation. The solid vertical markers denote an elapsed time
of 200 s following onset of the main pressure event
4.4.3. Unstable detonations

Further difficulties arise in quantifying unstable or overdriven detonations. The pressure
averaging approach described above for steady detonation has therefore been further
developed to provide threshold conditions for acceptance as valid unstable detonation tests.
An average pressure, Pmu, is again obtained from test pressure records, averaged over 200 s
after the wave front arrival. The test is deemed to represent an unstable or overdriven
detonation if the ratio of Pmu/Ptd exceeds 2.5 and 3 for pipes diameters that are less than 100
mm or greater than 100 mm respectively. An overdriven wave that meets this condition is
shown in Figure 2 (b) from a test in a 150 mm pipe. The device must then prevent flame
propagation in five tests at these conditions.
In addition, unstable detonations must also be subjected to a further three deflagration
tests, as described above for stable detonation device tests
418
5.
CONCLUSIONS
In comparison with key elements in other existing standards, the new European standard,
thus requires significantly fewer tests than the US Coast Guard standard procedures. It does
however include far more precise definitions of detonation events and includes crucial tests
against accelerating flames that are not considered as part of the BS7244. The mechanisms
by which explosion arresters operate if far from understood and may be different for
different phases of explosion events. For example, the author is aware of instances where an
arrester was optimized to pass an overdriven test but failed when subject to a deflagration
test.
REFERENCES
[1] Flame arresters for general use, BS7244, British Standards Institution, 1990
[2] Marine vapour control systems, U.S.Federal Register 25396-25451, United States Coast Guard,
1990
[3] Revised standards for the design, testing and location of devices to prevent the passage of flame
into cargo tanks in tankers, MS/CIRC 373 Rev 1,8, International Maritime Organization, 1988
[4] N. ROUSSAKIS, K. LAPP, A comprehensive test method for inline flame arresters,
Plant/Operations Progress 10, p.85-92, 1991
419
PREDICTION THE THEORETICAL INTERIOR BALLISTIC

PROPERTIES OF SOLID PROPELLANT ROCKET MOTORS
Jasmin Terzic, Alija Lekic and Berko Zecevic
Abstract
Process for prediction the interior ballistic properties of solid propellant rocket motors
is based on identification and quantitative estimation of idealized performances
deviations. Solid propellants combustion and expansion of combustion gases within a
rocket motor and nozzle represent a complex problem, which includes influences of
chemical reactions, erosive burning, transport phenomena etc.
The computer program TCPSP (Thermo-Chemical Properties Solid Propellant), which
ensures calculation the equilibrium contents of combustion gases, transport properties
of combustion gases and theoretical performances of rocket motors with specified
propellant grains, has been developed.
Comparative analyses of results from the TCPSP computer code with referring codes
Ophelie and CEA have been carried out. A very good agreement has obtained.
Keyword:
1.
rocket motors, solid propellant, combustion chamber, nozzle, combustion,

expansion, combustion products, thermodynamic properties, transport
properties, rocket performances
INTRODUCTION
Theoretical model, which describes complex phenomena appearing within rocket motor,
is based on several hypotheses. These are referred to conditions of combustion process
taking place in rocket motor, to expansion of exhaust products through nozzle and to
properties of combustion products.
Calculation of energetic performances of solid propellants is directed to determination of
combustion products mixture, in conditions of energetic, phase and chemical equilibrium.
This calculation includes setting and solving the equation system describing phase and
chemical equilibrium.
Necessary data for calculation of energetic characteristics of solid propellants are:
chemical constituents and their enthalpies of formation, thermodynamic properties of
prospective combustion products within a wide temperature diapason and parameters of
inter-molecular potentials for all compounds within corresponding temperature diapason.
Basic sources for these data are data bases and publications [1, 2, 3, 12]. Calculation, based on
the model as above specified, should:
provide optimization a solid propellant composition relating to energetic

properties, and
provide necessary data for design of rocket motor.
420
2.
COMPUTATIONAL MODEL FOR CALCULATION

THE THEORETICAL PROPERTIES OF SOLID ROCKET
PROPELLANTS
2.1
Assumptions
A basic assumption, which was also used in different models [10,14,4,16,9,13,7], is that solid
propellant combustion and expansion of exhaust combustion products represent two separate
processes located in the combustion chamber and in the nozzle.
Combustion and expansion through the nozzle have an adiabatic character and they
occur as steady-state processes. Combustion gaseous mixture is homogeneous across any
cross-section of the combustion chamber and the nozzle normal to the nozzle axis.
The gaseous combustion products mixture behaves as a mixture of perfect gases, which
is not soluble into condensed phases, so that such a mixture obey the perfect gas law
[10,14,4,16,9,13,7]
. Condensed-phase particles within the combustion products mixture are small
so that their total volume is negligible when compared to the gases volume. Combustion
products mixture is homogeneous, and condensed-phase particles are in thermal and kinetic
equilibrium with gaseous phase.
Nonisentropic and irreversible process occur within the combustion chamber. Flow
through the nozzle is considered as steady-state and adiabatic, which gives it an isentropic
character. Velocity of gases at inlet nozzle area is neglected when compared to one which is
obtained throughout the nozzle. During expansion throughout the nozzle, composition of
combustion products, can be considered as:
Gaseous mixture, generated by the burning of solid propellant, stays frozen

frozen equilibrium expansion, or
Gaseous mixture stays in chemical equilibrium, which depends on local pressure
and temperature in any moment of expansion shifting equilibrium expansion.
Parameters of the real process are among values obtained by calculations under both
above conditions.
2.2
Combustion products mixture under conditions

of shifting equilibrium
Calculation of combustion product constituents represents an essence of procedure for

prediction the solid propellant properties.
Regarding the assumption in which the total thermodynamic equilibrium is frozen,
contents of the combustion product mixture will be calculated at shifting equilibrium
conditions. For calculation of shifting equilibrium combustion products mixture, a method
of minimum Gibbs energy is used in this paper.
Generally, combustion products of solid propellants generate a mixture consisting of
multi-phase constituents, although in most cases there are usually two phases. The basic
condition for equilibrium state of mixture at constant pressure and temperature is [15]:
dGT , P = 0
(1)
where; G Gibbs free energy.

Gibbs free energy for two-phase system is equal to sum of energies of particular phases,
which are assumed to be homogeneous:
421
NG
NS
n~
n
G = ni i0 + RT ln i + RT ln P + n~i i0 + RT ln ~i
n
n
i = NG +1
i =1
(2)
where, NG number of gaseous combustion products, NS total number of combustion

products, ni - number of mols , i0 - chemical potential at the standard conditions, P
pressure referred to the standard pressure (pressure of p0 = 0.10132472 MPa is chosen as the
standard state) [15]. Sign is referred to condensed phase.
During calculation of the composition such a two-phase system in conditions of
chemical equilibrium, besides the condition (1), conditions of mass balance for all chemical
elements, consisting in combustion products, must be fulfilled:
NG
aij ni +
i =1
NS
a n~ a
NG +1
ij
0j
= 0,
for j = 1, 2, . . ., l
(3)
where: aij stechiometric coefficients of components, or number of atoms j consisted in

component i, a0j total number of atoms of the element j consisted in combustion products,
and l number of elements consisted in combustion products.
In order to be calculated composition of combustion products mixture, conditions of
mols balance for particular phases, must be also fulfilled.
NG
NS
ni n = 0
and
i =1
n~ n~ = 0
i = NG +1
(4)
2.2.1. Procedure
It can be seen from equations (1) to (4) that a system of NS + l + 2 non-linear equations
should be solved in order to be predicted composition of combustion products. Number of
equations for calculation of combustion products can be decreased to l + 2 if a geometric
programming is used [15]. Let the function v be defined as:
G
v = e RT
(5)
and after substitution of the equation (2) for Gibbs energy

ni
c
~
v = i n n n~ n
i =1 ni
NS
(6)
where:
0
e i RT
ci =
p
~0
c~i = e i
RT
and
(7)
Minimum of Gibbs free energy with limitations (3) and (4) can be presented through
maximum of the function v, which is defined by the equation (6) under same limitations.
Using theory of geometric programming [15], it can be seen that function v, with
limitations (3) and (4), represents a bipartite function, whose direct criterion function has a
following form:
422
g0 = t j 0 j
a
(8)
j =1
where: t independent variable of the criterion function. Then, limitations (3) and (4) are
reduced to two limitations:
NG
i =1
j =1
g1 = ci t j 0 j = 1
g2 =
NS
i = NG +1
and
a
c~i t j 0 j = 1
(9)
j =1
Maximum of the function v with limitations (3) and (4) corresponds to minimum of the
function g0 with limitations (9). Assignment of minimum of the function g0 with given
limitations leads to assignment of mathematical minimum of function
f = g 0 + 1 g1 + 2 g 2 ,
(10)
where 1 and 2 Lagranges multipliers.

In accordance with above considerations, problem of assignment of minimum of Gibbs
free energy with limitations (3) and (4) is reduced to assignment of minimum of the function
(10) with l + 2 variables.
Since among optimal values of independent variables of the direct criterion and the
bipartite function exists a mutual relation, molar constituents are as follows:
l
i = ci t j
aij
and
j =1
l
~
a
i = c~i t j
(11)
ij
j =1
Each of two equations of chemical elements balance (3) might be used for assignment of
condensed phase constituents. Taking into consideration following:
ni = i n
~
n~i = i n~
and
(12)
from the balance each of two chemical elements one gets:

NG
n aij1 i + n~
i =1
NG
NS
i = NG +1
NS
n aij 2 i + n~
i =1
i = NG +1
~
a 0 j1 = 0
ij1 i
ij 2
~
i a0 j 2 = 0
and
(13)
Minimum of the direct criterion function with corresponding limitations, equations (3)
and (4), is assigned by application of the Newton-Raphson method. As criterion for
assignment of values for independent variables at minimum of the direct criterion function,
the following condition is used:
423
v g0
10 16
c
(14)
Initial values of Lagrange multipliers 1=-2,45 and 2=-0,0003 are taken on the basis of
number of iterations, which are needed for determination of minimum of the direct criterion
function g0 [15].
Convergence and its rate depend on selection of the initial values for all independent
variables of direct criterion function g0. Initial values for all independent variables of direct
criterion function g0 in the equation (8) are same and equal to 10-5,5. These values are used
for calculation of gaseous combustion products, for all initial temperatures and selected
pressure at each cross-sectional area of rocket motor. A possible case that there is no
convergence, when combustion products mixture which has condensed phase products is
being calculated. In such a case, calculated data for gaseous phase of each independent
variable, should be used as initial values of condensed phase.
Thermodynamic properties of combustion products are assigned by known relations for
properties of ideal-gas mixture [10,15,18]:
molar mass
NG
n i M i + n~
M=
i =1
NS
i = NG +1
Mw =
n i M i + n~
i =1
and
NS
~ ~
i = NG +1
(15)
n + n~
enthalpy
NG
h=
n
NG
~ ~
n H i i + n~
i =1
NS
~ ~
i = NG +1
M (n + n~ )
entropy
NG
NS
NG
NS ~ ~
~ ~
n S i i R i ln i R ln P + n~ S i i R i ln i
i =1
i = NG +1
i = NG +1
;
s = i =1
M (n + n~ )
(16)
(17)
specific heat at constant pressure

o no reacting mixture
NG
cpf =
n C p i i + n~
i =1
NS
~ ~
i = NG +1
M (n + n~ )
pi
i
(18)
o reacting mixture
cp =
1 h
M T P
(19)
Gas constant and density of gaseous phase are assigned in the same way as for ideal gas:
424
R
,
M
(20)
P
.
Rg T
(21)
Rg =
The ratio of specific heats and speed of sound for no reacting mixture are assigned by
relations [15]:
R
f = 1 g
cp
f
(22)
a f = RgT ,
(23)
and for reacting mixture:
R
= 1 g
cp
a=
T M

1
M T P
P M

1 +
M P T
Rg T
.
P M
1+
M P T
(24)
(25)
Partial derivations from equations (19), (24) and (25) are assigned by numerical
procedure with increments T = 0,01 K and P = 0,1 Pa.
Properties of components are calculated by means of reduced thermodynamic potential,
which is determined by the following relation:
) (10 T )
0 = ln 10 4 T +
= 2
(26)
and relation
= f H T 0
(27)
where fH enthalpy of formation at temperature of 298,15 K

Data for coefficients in the equation (26), enthalpy of formation and molar mass of
particular components were taken from ref. [1,2,3]. Relations for calculation thermodynamic
properties by means of thermodynamic potentials and their values for particular components
are given in ref. [18].
425
2.3
Relations for calculation of transport properties
Transport properties of combustion products mixture are assigned only for gaseous
phase. They are calculated by means of following relations:
Kinematic viscosity of the mixture [6]:
g =
i
i i
ij j
j
(28)
where:
i M j
4
1 +
j Mi
1
ij =
8
M
1+ i
Mj
(29)
Kinematic viscosity of a component [4]:
= 2,7627 10 6
MT
*2, 2
2
(30)
Reduced integral of crash
l ,s
*
= d i ln T *
i =0
diameter of the effective cross-section of crash [10-10 m], T * = T

constant, and maximal energy of attraction between two molecules.
(31)
, k Boltzman
Thermal conductivity coefficient for combustion products mixture of reacting

gases[15,10,4]:
= fr + rr ,
(32)
were:
thermal conductivity coefficient of no reacting gases mixture [6,15,4]:
fr = +
(33)
thermal conductivity coefficient of one-atom gases mixture
=
i
i i
ij j
j
(34)
thermal conductivity coefficient of the one-atom gas [6,15,4]:
426
i =
15 R
i ,
4 Mi
(35)
1 2 M 1 4
1 + i j
(
M i M j )(M i 0,142M j ) j M i
ij = 1 + 2,41
12
(M i + M j )2
M
i
2 2 1 +
M
j
(36)
share of the thermal conductivity coefficient caused by energy transfer of atom interior
degrees of freedom.
=
i
(i i) i
Dii
Dij
(37)
thermal conductivity coefficient of multi-atom gas [6,15]:
i M i i Di 3 R 5 i Di

=
+
i cvi
i
2 cvi 2
i
(38)
self-diffusion and binary-diffusion coefficients, D, Dii and Dij in the equations (37) and
(38) are assigned by the relation [4]:
Dij = 0,2628 10 6
T T (M i + M j ) 2 M i M j
p ij2 1*,ij1
(39)
share of the thermal conductivity originated due to chemical reactions within the mixture
:
[4,15,13]
A11
A12
M
A1
re =
A12
A22
K
K
H1
H 2
A2
K
K
K
0
A1
A2
1 H1 H 2
A11 A12
RT 2
A12 A22
M
A1 A2
(40)
heat effect of independent reaction at temperature of the mixture is

H j = aij ( H i H j ) ,
j=1, 2, ,
(41)
elements of determinants in the equation (40) are assigned by means of relations:
427
Aij =
NG 1 NG
(aik k ail l )(a jk l a jl k )
k =1 l = k +1
k l Dkl
(42)
Stechiometric coefficients in independent reactions, are components of a vector of basic

coefficients from homogeneous system of linear equations, representing balance of mass of
elements [15].
2.4
Theoretical rocket performances
Besides the cited assumptions, calculation of theoretical performances of rocket motors

is based on assumption of the Infinite-area combustion chamber (IAC) model [10].
For calculation of theoretical performances of a rocket motor, following simplified
equations are used [10]:
1. Equation of continuity:
2 A2 u 2 = 1 A1u1
(43)
2. Equation of momentum:
P2 + 2 u 22 = P1 + 1u12
(44)
3. Equation of energy:
h2 +
u 22
u2
= h1 + 1
2
2
(45)
Equation (43) describes condition of constant mass flow rate m& , given by the following
relation:
m& = Au
(46)
where: A referent nozzle cross-sectional area (throat or exit area), u gas stream velocity.
If the combustion chamber is marked by subscript c, and the exit area by subscript e,
then relation for gas stream velocity can be obtained from equation of energy (45) [10,13]:
u e = 2(hc he )
(47)
where: h enthalpy at given cross-sectional area, expressed in J/kg, u gas stream velocity,
expressed in m/s.
Rocket motor thrust for one-dimensional and steady-state flow can be defined by the
relation:
F = m& u e + (Pe Pa )Ae
(48)
where, Pa external (atmospheric) pressure.

Specific impulse is defined as the ratio of thrust to mass flow rate:
I sp =
(P Pa )Ae
F
= ue + e
m&
m&
.
(49)
Mach number is defined as the ratio of the flow velocity to the local sonic velocity:
428
M =
u
a
(50)
The flow velocity is given by the relation (47), and the sonic velocity is given by
relations (23) or (25).
Characteristic velocity, marked as C*, is defined as
C* =
Pc At
m&
(51)
where subscript t is referred to the nozzle throat region.

Using the equation of continuity (43) one gets the relation for ratio of the nozzle exit
cross-sectional area to the nozzle throat cross-sectional area:
Ae t u t
=
At e u e
(52)
The thrust coefficient is defined by:

CF =
2.5
I sp
C* .
(53)
Characteristics of the combustion process
Determination of characteristics of the combustion process is based on assumptions in

which the combustion chamber pressure and the propellant enthalpy of formation are
considered as constants. Input data for the procedure are composition and enthalpy of
formation of the propellant.
According to hypothesis referring to the combustion process, basic equation of energy
conservation at the combustion chamber (pc = const) is given by:
hc h f pm = 0
(54)
where: hc enthalpy of combustion products in the combustion chamber, hfpm enthalpy of

formation of the solid propellant.
The combustion temperature is assigned from the following relation by means of
Newton method:
T = hi h f pm
hi h f pm
h
=
cp
T P
.
(55)
Temperature for each further iteration is:

T ( r +1) = T ( r ) + T ( r +1)
(56)
where, r number of iterations.

Iteration procedure is repeated until condition T 0,01 being satisfied.
429
2.6
Expansion characteristics in the nozzle
The basic equation for assignment of characteristics of gas expansion in the nozzle is the
equation of isentropic process:
s sc = 0
(57)
where: sc entropy of combustion products in the combustion chamber, s entropy of

combustion products at a cross-sectional area of the nozzle.
For both frozen and shifting equilibrium conditions, three cases are considered as
follows [10,4,9,13]:
expansion to given pressure at the nozzle exit,

expansion to given Mach number,
expansion to given expansion ratio ( ratio of cross-sectional area at the nozzle exit
to the throat area).
For the expansion up to given pressure:

s
ln T = s
ln T P
,
(58)
where: s = S S c and
s
s
ln T = s
=
cp
ln T P
(59)
Iteration procedure is repeated until the condition ln T < 0,1 10 4 being satisfied.
For the expansion up to given Mach number, the equations (59), (47), (23) or (25), are
solved simultaneously, until the following condition being satisfied:
ut2 at2
0,4 10 4
2
ut
(60)
Pressure at the nozzle throat, is obtained by the following relation:
2 1
Pt = Pc
+1
(61)
For the expansion up to given expansion ratio At/Ae , relation (59) is solved, until the
condition ln T < 0,1 10 4 being satisfied.
Initial pressure estimation at the nozzle exit is predicted on the basis of the following
empirical relation [10]:
ln
Pc
A
= + 1,4 ln e
Pe
At
za e 2
At
(62)
430
where, value of is taken from the previous cross-sectional nozzle area. For further
iterations, pressure is calculated from the equation of state.
In case of expansion at frozen equilibrium conditions, combustion products mixture is
considered as non-reacting, while an expanding gas mixture at shifting conditions is
chemically reacting mixture.
3.
VERIFICATION OF THE MODEL FOR CALCULATION

OF THEORETICAL PERFORMANCES OF ROCKET MOTORS
On the basis of the presented mathematical model, the computer code TCPSP (ThermoChemical Properties Solid Propellant) has been developed. This program enables
calculation of the combustion products composition at chemical equilibrium, transport
properties of gaseous combustion products and theoretical performances of rocket motors.
This model describes procedures for obtaining theoretical performances of rocket motors for
both cases of expansion, at frozen equilibrium and shifting equilibrium conditions.
The computer program TCPSP enables calculation of theoretical performances of rocket
motors with propellants consisting of the following chemical elements: Al, C, Ca, H, K, Mg,
N, Na, O, P, S, Si, Ti, F, Fe, Cl, Pb. The data-base consisting of 100 propellant ingredients
based on available data published by MARTIN MARIETA [8] and STANAG 4400 [5] has
been established. This program is capable to predict properties of combustion products
mixture with 156 gaseous and 39 phase-condensed ingredients. The data bases, which
consist of propellant ingredients and combustion products, can be upgraded by new
ingredients.
A comparative analysis of results obtained by the TCPSP program with results from
French computer code Ophelia, version 3, for a composite solid propellant BUTALANE
82/4 [13] and with American computer code CEA (Chemical Equilibrium with Applications)
for a composite solid propellant with the following composition: ammonium perchlorate,
CHOS-Binder, Al, MgO, H2O (with percentage contents of 72.06/18.58/9/0.2/0.16) [11].
Besides, a detailed comparative analysis of several double base rocket propellants has
been performed, and an analysis of some theoretical performances of DB propellants NGRA and NGR-B compared to experimental results, has been considered particularly.
By analyzing the values of molar parts of combustion products for composite
propellants, a small deviation of results obtained by TCPSP program compared to results
obtained by programs Ophelia and CEA for all variants of calculation, has been noted.
Maximum deviations of calculated results for combustion products mixture, obtained by
TCPSP program [18] are:
compared to results from Ophelia V3 about 1,2 %,
compared to results from CEA about 1,5 %.
A good agreement of calculated thermodynamic properties of combustion products
mixtures with results from the program Ophelia V3 (Fig. 1 and 2), for both cases of
expansion, has been obtained. Maximum deviation of results obtained by TCPSP program
compared to those from Ophelia V3 is 4.5 % (pressure at nozzle exit for Ae/At=30).
A good agreement of thermodynamic properties of combustion products mixture in the
combustion chamber, calculated by TCPSP compared to those from CEA code has also been
obtained (Table 1).
431
11%
10%
9%
8%
7%
6%
5%
4%
3%
2%
1%
0%
Combustion chamber
Throat area
Fig 1.
p)
t
t)p
(d
v
(d
v
Dv
c*
Ae
/A
Is t
p,
ad
Vz
v
ac
h
M
C
G p
A
M
A
s
Exit area for the expansion ratio of Ae/At=30
DEVIATION
The program TCPSP gives a good accuracy of prediction for combustion products
viscosity and thermal conduction coefficient in the combustion chamber and at the nozzle
throat when compared to results from Ophelia V3 program. Maximum deviations of 7.5% to
10.5% have appeared only at thermal conductivity coefficient prediction, at the nozzle exit
and for the expansion ratio of Ae/At=30.
Comparative analysis of some properties for solid propellant BUTALANE

82/4 in case of shifting equilibrium expansion when compared to results
from Ophelia version 3
7%
Combustion chamber
6%
Thort area
5%
Exit area for the expansion ratio of Ae/At=30
4%
3%
2%
1%
Fig 2.
Dv
t
Is
p,
ad
Ae
/A
c*
ac
h
M
Vz
v
s
M
A
G
A
Cp
0%
DEVIATION
8%
Comparative analysis of some properties for solid propellant BUTALANE

82/4 in case of frozen equilibrium expansion when compared to results
from Ophelia version 3
432
Table 1. Comparative analysis of some properties in the combustion chamber for

the solid propellant AP/CHOS-Binder/Al/MgO/H2O
3.447
P [MPa]
1.724
[11]
[11]
TCPSP CEA
DEVIATION TCPSP CEA
T [K]
2716.8 2724.46
-0.28% 2700.2 2708.02
Cp [J/gK]
2.4185 2.40789
0.44% 2.5407 2.531738
1.1969
1.1945
0.20% 1.1926
1.189
s [J/gK]
10.529 10.57506
-0.44% 10.788 10.82443
h [J/g]
-2028.3 -2028.24
0.00% -2028.3 -2028.24
3
3527
3520.9
0.17%
1772
1768.1
[g/m ]
M (1/n)
23.112
23.136
-0.10% 23.071
23.096
MW [g/mol] 22.262
22.282
-0.09% 22.225
22.246
a [m/s]
1080.1
1081.4
-0.12% 1075.4
1076.6
(dvt)p
1.0457
1.0518
-0.58%
1.069
1.0686
(dvp)t
-1.0026 -1.00263
0.00% -1.0035 -1.00342
DEVIATION
-0.29%
0.35%
0.30%
-0.34%
0.00%
0.22%
-0.11%
-0.09%
-0.11%
0.04%
0.01%
Stand tests of 128mm rocket motors with two types of DB propellants (NGR-A and
NGR-B) were performed. Measured values of specific impulse of the tested rocket motors
and those calculated by TCPSP is shown in Table 2. Deviations of the measured specific
impulse compared to those calculated by TCPSP are: for propellant NGR-A is 7.98% and for
propellant NGR-B is 9.96%. These deviations of specific impulse are in accordance with the
published results given by other authors, since usual specific impulse looses are between 2%
to10 %.
Table 2. Compared results of the measured and the calculated values of specific
impulse for DB solid rocket propellants NGR-A and NGR-B
Reference
TCPSP
Measured [17]
Deviation [%]
TCPSP
Measured [17]
Deviation [%]
Isp
[Ns/kg]
2213.6
2050
7.98
2243.3
2040
9.96
Type of solid propellant and rocket

motor working conditions
NGR-A, pc=130 bar, pa=1.01325 bar,
Ae/At=7.51
NGR-B, pc=130 bar, pa=1.01325 bar,
Ae/At=7.51
433
4.
CONCLUSION
The program TCPSP enables:
Calculation of the chemical equilibrium composition for a multi-phase, multiingredient combustion products mixture, its thermodynamic and thermo-physical
properties, as well as performances of the process within a rocket motor.
Applied algorithm enables calculation of theoretical performances for double base
and composite rocket propellants as well.
Further extension of the data base:
o includes thermo-chemical properties of different ingredients as possible
constituents of wide spectrum of solid rocket propellants, and
o coefficients of reduced thermodynamic and thermo-physical potentials for
calculation of corresponding properties of possible combustion products and
parameters of potentials as well.
Very good agreement of calculated theoretical performances of rocket motors with

referent programs Ophelie and CEA.
434
REFERENCES:
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
AKADEMIJA NAUK SSSR: Termodinamieskie svojstva individualnih veestv: Tom I, Nauka,

Moskva, 1978.
AKADEMIJA NAUK SSSR: Termodinamieskie svojstva individualnih veestv: Tom II, Nauka,
Moskva, 1979.
AKADEMIJA NAUK SSSR: Termodinamieskie svojstva individualnih veestv: Tom III, Nauka,
Moskva, 1981.
ALEMASOV V.E., DREGALIN A.F., TICHIN A.P. AND KHUDYAKOV V.A.: Thermodinamic and
Thermochemical Properies of Combustion Products, Volume I, Editor Glushko V.P.,
Akademija nauk SSSR, Israel Program for Scientific Translation, Jerusalem 1974
ANON: Derivation of Thermochemical Values for Interior Ballistic Calculation, NATO
Military Agency for Standardization, STANAG 4400, Brussels, 1993
BIRD, R.B., STEWART W.E., LIGHFOOT E.N.: Transport phenomena, New York, 1960.
BOISSON, J.: La propulsion par fusee (tome I), ENSTA, Paris, 1972
CUNNINGHAM J.: Ingradients of Propellant data, Martin Marieta, Orlando Florida, 1986
FILIPOVIC M., KILIBARDA N.: The calculation of thermochemical enegetic performances of
composite rocket propellants, J. Serb. Chem. Soc., Military Technical Institute, Beograd, 2000.
GORDON S. AND MCBRIDE B.: Computer Program for Calculation of Complex Chemical
Equilibrium Compositions and Applications, I. Analysis, NASA Reference Publication 1311,
october 1994
GORDON S. AND MCBRIDE B.: Computer Program for Calculation of Complex Chemical
Equilibrium Compositions and Applications, II. Users Manual and Program Description,
NASA Reference Publication 1311, June 1996.
STULL D.R AND PROPHET H.: JANAF Termochemical Tables, Second Edition, NSRDS-BBS
37, Issued, June1971
KILIBARDA N.: Calculation of theoretical performances of rocket propellants, Proceedings of
papers on research referred to combustible cases and rocket propellants, Military Technical
Institute JNA, Belgrade, 1991.
KUO K. K.: Principeles of Combustion, John Wiley&Sons, 1986, New York.
LEKIC A., POPOVI A., VUJOVI V., TADI V.: Program for assignment of composition and
properties of combustion products of coals, ENERGOINVEST, RO ITEN, Sarajevo 1984.
SORKIN R.E.: Teorija vnutrikamernih processov v raketnih sistemah na tverdom toplive,
Moskva, 1983.
ZECEVIC B.: Influence of the variable radial acceleration to internal ballistics of rocket motors
with DB propellants, Disertacion, Faculty of Mechanical Engineering, Sarajevo 1999.
TERZI J.: Prediction of idealized internal ballistic properties of a rocket motor with DB solid
propellant, Faculty of Mechanical Engineering, Sarajevo 2002.
435
DETONATION CHARACTERISTICS
OF LOW-SENSITIVITY NTO-BASED EXPLOSIVES
W. A. Trzciski, L. Szymaczyk, and S. Cudzio
Military University of Technology,
Kaliskiego 2, 00 980 Warsaw, Poland
Abstract:
Detonating performances of new explosive compositions containing NTO, TNT and
RDX are investigated in this work. Detonation velocity, pressure and energy of the
mixtures tested and acceleration ability and equation of state of their detonation
products are determined. Shock and impact sensitivity is evaluated in the gap test and
heavy hammer test.
The compositions investigated can be considered as a suitable insensitive explosive to
replace TNT/RDX mixtures in munitions applications.
Keyword:
1.
insensitive munitions, detonation performance
INTRODUCTION
NTO (3-nitro-1,2,4-triazol-5-on) is commonly accepted as an ingredient of explosive

compositions due to its relatively high performance and low sensitivity to mechanical
stimuli [1]. The wide review of compositions containing NTO is made in Ref. [2]. From the
review it follows that the low sensitivity mixtures of NTO, TNT and RDX have been
manufactured and tested. However, their detonation performance is lower that those of
Composition B (TNT/RDX 36/64). That is why the problem of manufacturing insensitive
TNT/RDX/NTO compositions with detonation parameters comparable with those of
Comp. B is still opened.
In this work, an attempt is undertaken to obtain low sensitivity mixture of NTO, TNT
and RDX of high detonation and energetic performances. To select the composition contents
the results of thermochemical calculations of detonation parameters are used. Form two
chosen mixtures the detonation and energetic characteristics are determined. Shock and
impact sensitivity of the compositions is also evaluated.
2.
COMPOSITIONS
Theoretical estimation of detonation characteristics of TNT/RDX/NTO mixtures was

performed by the use of the thermochemical code CHEETAH [3]. For a chosen amount of
TNT the detonation velocity was calculated as a function of the mass contents of NTO. It
was assumed that the density of the mixtures was 95 % of the theoretical maximal density.
The dependence of the detonation velocity D on the contents of NTO xNTO is shown in
Fig. 1.
The detonation parameters of TNT/RDX 40/60 mixture (hereafter denoted by CB) are
assumed to correspond to the parameters of Comp. B. The detonation velocity of CB is
indicated by a triangle symbol in Fig. 1. From curves presented in Fig. 1 it follows that the
436
contents of TNT in a mixture TNT/RDX/NTO should be reduced to obtain a detonation

velocity comparable with that of CB. The composition TNT/RDX/NTO 30/20/50 (denoted
by C50) should have parameters similar to Comp. B. This mixture and the composition
TNT/RDX/NTO 30/40/30 (denoted by C30) were chosen for investigations.
8000
30 % TNT
7900
C30
7800 CB
C50
D [m/s]
7700
40 % TNT
7600
7500
50 % TNT
7400
7300
7200
7100
10
20
30
40
50
60
70
xNTO [%]
Fig 1.
Dependence of the detonation velocity of TNT/RDX/NTO mixtures on the

contents of NTO for differnt amount of TNT
The procedure of preparing the compositions was as follow. NTO and RDX were added
to melted TNT and they were mixed. The mixture was frozen with continuous mixing and
finally the granulated product was obtained.
3.
DETONATION PERFORMANCE
3.1
Detonation pressure and velocity
To determine the detonation pressure of the mixtures, a variant of the aquarium test was
applied [4]. In this method, profiles of an oblique shock wave propagating in a cylindrical
layer of water during detonation of a cylindrical charge of an explosive tested is recorded
with a X-ray set. The experimental profiles are then compared with results of numerical
modelling of the expansion process which are in a form of relation between the position of
the front of oblique shock wave in water and the exponent of isentrope () of detonation
products. The value of corresponding to the solution that overlaps the experimental profile
is accepted as the exponent sought. The detonation pressure is calculated according to the
following equation:
p CJ =
0 D 2
+ 1
(1)
where pCJ denotes the detonation pressure and 0 is a density of the explosive tested.
The scheme of the experimental arrangement, used in the aquarium test, is shown in
Fig. 2. A cylindrical charge of 23.7 mm in diameter and 250 mm in length was placed inside
437
a PCV tube with an inner diameter of 71 mm and wall thickness of 2 mm. The tube was
filled with water. Short-circuit sensors were located in the charge to measure detonation
velocity.
Fig 2.
Aquarium test experimental arrangement: 1 booster, 2 charge tested,

3 water, 4 sensor to measure detonation velocity, 5 sensor to trigger
the X-ray set
The density and detonation velocity of the mixtures used in the water test and the results
are given in Tab. 1. The exponent of isentrope of detonation products was determined by
comparison of measured and calculated positions of the shock wave front in a plane section
located at a distance of one charge radius from the front of detonation wave. The detonation
pressure was calculated from the Eq. (1). Table 1 also contains some theoretical values of the
parameters (with an index t) calculated with the thermochemical code CHEETAH.
Table 1. Experimental and calculated detonation properties of the mixtures tested
Explosive
CB
C30
C50
0
[kg/m3]
1674
1717
1738
D
[m/s]
7830
7850
7730
pCJ
Dt
pCJ,t
2.97
3.21
3.22
25.85
25.1
24.6
7810
7885
7797
3.02
3.14
3.22
25.40
25.78
25.05
From the results obtained it follows that the detonation parameters of the composition
C30 are close to those of the composition CB. The detonation velocity and pressure of the
composition C50 are about 1 and 5 % lower than those of CB, respectively.
3.2
Cylinder test
The cylinder test results were the basis for determination of acceleration abilities and
energetic characteristics of the detonation products of mixtures investigated. The process of
acceleration of a copper tube by detonation products was recorded with the impulse X-ray
apparatus. The tube was 300 mm long with internal diameter of 25 mm and wall thickness of
2.5 mm. The results of the test are presented in Fig. 2. in the form of the dependence of the
radius of external surface of the tube on the axis co-ordinate x. To determine the radial
velocity of the copper tube, the data obtained from the cylinder test were recalculated by the
use of the method described in Ref. [5].
438
40
TNT/RDX 40/60
TNT/RDX/NTO 30/40/30
re [mm]
35
30
25
20
15
20
40
60
80
100
120
140
x [mm]
Fig 3.
3.3
Dependence of the external radius of the tube on the axis co-ordinate
Gurney and detonation energy
The acceleration ability of explosive can be described by so-called Gurney energy. For
cylindrical envelopes the Gurney energy is expressed by the following relation
EG = +
1 u L2
,
2 2
(2)
where uL is the velocity of the tube and denotes the ratio of tube mass to explosive mass.
Results of the cylinder test enable us to analyze the dependence of the Gurney energy,
described by formulae (2), on the degree of tube expansion. This dependence also describes
the acceleration ability of the explosive. The relationship between the Gurney energy and the
relative volume of detonation products for the mixtures tested is given in Fig. 4.
3500
EG [kJ/kg]
3000
2500
TNT/RDX 40/60
2000
1500
v/v0
Fig 4.
Acceleration ability of the explosive tested
439
The results of cylinder test can also be used to estimate the detonation energy. In Ref. [5]
it was shown that there was a correlation between the velocity of driven tube at the infinite
volume of the detonation products and the detonation energy of an explosive. The relation
can be written as follows
1
+
e0
2
=
s
e0 s + 1
uL
us
L
(3)
where e0 and e0s are the detonation energy of a given explosive and a standard explosive,
respectively uL and uLs denote the tube velocity determined at the infinitive volume of
detonation products of the explosives. These velocities were estimatid by the method
described in Ref. [5]. Using phlegmatized RDX as a standard explosive the detonation
energies of the explosive tested was calculated from (3). The values obtained are given in
Table 2.
Table 2. Detonation energies estimated of the basis of cylinder test data
Explosive
RDXph
CB
C30
C50
e0
[kJ/kg]
5263
5126
4911
4700
E0 = e0 0
[MPa]
8.7
8.6
8.4
8.2
From the Fig. 4 and Table 2 it follows that the Gurney energy and detonation energy of
compositions with NTO are a bit lower that those of TNT/RDX mixture.
3.4
Isentropes of the detonation products
So-called effective exponent of isentrope is determined on the basis of the real isentrope
of the detonation products. It can be estimated from results of the cylinder test. In Ref. [6] the
effective exponent is determined by comparison of the experimental profile of the copper
tube with that obtained from numerical modelling of the expansion process. The detonation
products, driving the tube, are described by the constant- equation of state. This method was
applied in the present work to estimate the effective exponent of isentrope for detonation
products of explosive tested. The values of are as follows: = 3.01, 3.10, and 3.15 for CB,
C30 and C50, respectively.
In reality, the exponent of the isentrope greatly changes during the expansion of the
detonation products from a value of about 3 at the Chapman-Jouguet point to
approximately 1.2 at low pressure. Therefore the physical properties of the expanding
gaseous detonation products are more precisely described by the isentrope which was
proposed by Jones, Wilkins and Lee. Cylinder test results are commonly employed in most
methods of determination of the JWL constants. In one of them [7], some connections
between JWL coefficient are used. Using the model described in Ref. [8], as well as the
values of detonation velocity, detonation pressure and detonation energy of the explosive
tested, the constant of the JWL isentrope were estimated. Results of calculations are given in
Table 3.
440
Table 3. The detonation pressure and the constants of JWL isentrope for the
detonation products of explosives tested
Explosive
CB
C30
C50
pCJ
[GPA] [GPA]
25.85 705.261
25.3
986.158
24.7
951.452
B
[GPA]
11.9033
13.5494
12.6088
R1
C
[GPA]
1.31774
1.16431
1.13379
R2
4.80
5.21
5.16
1.31
1.30
1.33
0.33
0.32
0.30
Fig. 5 displays how different JWL and constant- isentropes for the explosives
investigated are.
TNT/RDX 40/60
lnp [GPa]
1
0
-1
JWL
-2
-3
-4
0.0
0.5
1.0
1.5
2.0
2.5
ln v/v0
Fig 5. The constnat- and JWL isentropes for detonation products of explosives tested
3.5
Expansion work
After determining an isentrope, the expansion work of detonation products can be

calculated from the equation:
v
W (v ) = ec + pi d v ,
(4)
vCJ
where pi is the pressure on the isentrope starting from the CJ point, and ec = (pCJ-p0) (v0
vCJ)/2 is the energy of compression of the explosive at the detonation front.
Dependence of the expansion work on the relative volume of the detonation products of
the explosives tested is shown in Fig. 6.
441
8500
8000
JWL
7500
W [MJ/m3]
7000
6500
6000
5500
TNT/RDX 40/60
5000
4500
4000
10
20
30
40
50
60
70
80
90 100
v/v0
Fig 6. Expansion work as a function of the relative volume of detonation products
4.
SENSITIVITY TO MECHANICAL STIMULI

4.1
Gap test
The gap test enables the determination of the shock sensitivity of explosives. The charge
configuration used in experiments is shown in Fig. 7. The explosive tested was placed inside
a cooper tube similar to that used in cylinder test. A booster made of phlegmatized RDX
served as a shock wave generator. From shot to shot, the length of plexiglass attenuator was
changer with 1-mm step. The highest and the lowest gap values were appointed for which
the complete detonation and failure of explosion process were observed. The complete
detonation of explosive charge was indicated when a clean hole was cut in the steel witness
plate.
60
CB
h [mm]
C30
50
4
TNT
C50
Fig 7. Scheme of the charge configuration to

determine the shock sensitivity: 1 detonator
and holder, 2 booster, 3 plexiglass gap,
4 - explosive tested in cooper tube, 5 witness
plate
40
Fig 8. Thickness of plexiglass gap for

complete and incomplete detonation
in the explosive tested
442
The shock sensitivity of explosive tested as the usual gap results in the form detonation
no detonation is presented in Fig. 8. The sensitivity of the mixtures with NTO is lower
than that of composition CB.
The mean shock wave velocity of explosion process propagation in the acceptor charge
was measured by short-circuit sensors. The results obtained for composition C50 are shown
in Fig. 9.
10000
mean wave velocity [m/s]
8000
h = 48 mm
h = 50 mm
h = 54 mm
6000
4000
2000
10
20
30
40
distance [mm]
50
60
70
Fig 9. Mean velocities of wave in the charge C50 at the consecutive measuring
distances as the function of distance from the end of plexiglas gap for different
lengths of the gap
4.2
Hammer test
For the impact sensitivity determination a fallhammer apparatus was used. A 0.03 g
sample of explosive tested was placed in a piston device. The weight of hammer applied was
5 kg. For each drop height 10 trials were conducted. The minimum drop height at which 10
initiations occurred was 9, 25 and 30 cm for compositions CB, C30 and C50, respectively. It
means that NTO admixture decreases significantly the sensitivity of TNT/RDX mixtures.
5.
SUMMARY
New TNT/RDX/NTO formulations were successfully manufactured and characterized.

Detonation performance and energetic characteristics of these mixtures are comparable with
those of TNT/RDX (40/60) composition. The shock and impact sensitivity of tested
formulations is lower than the sensitivity of the TNT/RDX mixture. Manufactured
TNT/RDX/NTO compositions can be considered as a suitable main charge filling to
substitute for existing Composition B.
443
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
K. Y. LEE, M. D. COBURN, 3-Nitro-1,2,4-triazole-5-one a Less Sensitive Explosive, Journal of

Energetic Materials, 1987, vol. 5, pp. 27-33.
W. A. TRZCISKI, L. SZYMACZYK, Explosive Composition of Low Sensitivity Containing
NTO (in Polish), Biul. WAT, 2, 52, 2003.
L. E. FRIED, CHEETAH 1.39 Users Manual, Lawrence Livermore National Laboratory
1996.
W. A. TRZCINSKI, S. CUDZIO, L. SZYMACZYK, Determination of the Detonation Pressure
From a Water test, Engineering Transactions, 49, 4, 2001.
W. A. TRZCISKI, Application of the Cylinder Test for determining Energetic Characteristics
of Explosives, J. Techn. Phys., 42, 2, 2001.
S. CUDZIO, R. TRBISKI, W. TRZCISKI, Determination of the effective exponent of
isentrope for the detonation products of high explosives, Chem. Phys. Reports, 1997, vol.
16(9), p. 1719-1732.
R. TRBISKI, W. A. TRZCISKI, Determination of an expansion isentrope for detonation
products of condensed explosives, J. Techn. Phys., 40, 4, 1999.
R. TRBISKI, W. A. TRZCISKI, Modelling of the process of driving a cylindrical tube by the
detonation products described by the JWL equation of state, J. Techn. Phys., 38, 4, 1997.
444
UTILIZATION OF NUMERICAL SIMULATION METHODS

TO STUDY POSSIBLE IFLUENCE OF INITIATION
ON THE PERFORMANCE OF LINEAR SHAPED CHARGES
J. Vgenknecht, G. Hurtosov and L. Adamk
Department of Theory and Technology of Explosives,
University of Pardubice, 532 10 Pardubice, Czech Republic
Abstract:
There was realised an analysis of problems studying a possible influence of initiation on
performance of linear shaped charges without cumulative liner. A model system was
designed to study this type of charge. The behaviour of designed system was examined
by three different ways of initiation by means of numerical simulation using method of
final elements. There was discovered a significant improvement of observed
performance parameters of model system at the application of anticipated initiatory
source on the back of the charge.
Keyword:
1.
linear shaped charges, LSC, detonation, symmetry, initiation,

numerical simulation, FEM
INTRODUCTION:
Linear shaped charges (LSC) are practically used in a lot of series of very useful
applications and their importance has been increasing in recent times. One of the main things
is intensification of the blasting works in special cases, especially at destruction [1], where
LSC enables to reach required effect with minimisation of the explosive use. Subsequently
there is reached an increase of total safeties of these works. Mentioned increase of safeties is
connected with manipulation of smaller explosive volumes than at classical "sappers"
procedures and subsequently in that connection there is not such threatening and burdening
of environment
In comparison to classical - rotary symmetrical - shaped charges, which are applied first
of all at point penetration of resistant - strong thickness - materials [2] (especially in military
rank to perforate armoured and otherwise resistant targets) the cutting effect of LSC finds
much extensive use also in special destructive and building technologies in civil sector [3, 4] .
Rather considerable perspectives and therefrom-resulted distribution of similar
applications led to complete elaboration of these problems. There was mentioned above all
application-oriented research with emphasis on the practical use of obtained records.
Theoretic, generally valid and in wide criteria transmittable knowledge are not in this area so
currently publicised in the needed range. Acquired data are often of empirical character,
usually hardly generalised and they are also the object of productive or business
confidentiality in addition. It complicates to a great extent the study of these problems out of
commercially interested specialised workplaces. But potential of LSC area is so considerable
and interesting, that it becomes the object of study of unprofitable - academic workplaces.
Recently there have been provided relatively extensive studies of cumulation of gaseous
products of detonation at small rotary symmetrical charges at KTTV [5,6,7]. Obtained
knowledge and practical output of these works indicated a possibility of perspective
445
detachment of study of products of detonation cumulation of the explosives from the more
complex and difficult study of common cumulation systems with cumulative liners. Limited
practical use of systems without cumulative liners brings complications connected with
markedly limited volume of available (referenced) theoretic and also practical knowledge
and outputs, though the meaning of these studies, now rather marginal ways of research,
according to all indicia will evidently increase in future.
That is why there was designed a realisation of experiments using model system of
hollow LSC, i. e. LSC without cumulative liner, to study some fundamental factors
influencing practical use or utility value, of final realised cutting effects of these and also
similar charges.
2.
THEORY
2.1
Fundamental factors influencing quality of cumulation at LSC
At brief recapitulation of cumulations problems, it is possible to divide main factors

influencing resulting effect of resulting cumulative jet into three basic ones: energetic,
geometric and other factors [8].
2.1.1. Energetic factor
Energy contained in the explosive of the shaped charge is an absolutely determinative
factor for obtained resulting effect. Without any sufficient performance (impulse)
transmitted to accelerating mass it is impossible to reach corresponding, preferably
maximum effects. As there is used a principle of local concentrated energy, it is usual to
concentrate the energy as much as possible directly in the explosive itself.
2.1.2. Geometrical factor
Design of the charge certainly influences the quantity of energy, which is efficiently
possible to use for formation of the cumulative jet. In this aspect a phenomenon of so called
"active part of the charge. It is the same part of the explosive mass, which forms (in case of
charges with liner it shapes) the cumulative jet. However the size of charges active part isn't
determinated only by the geometry of the charge. Used casing of the charge, which to a
certain extent (to the moment of its disintegration) operates as a tamping and it retards
expansion of products of detonation, has also a definite indispensable influence. Whereby it
enables better utilisation of energy of products of detonation for deformation of cumulative
liner.
Evidently the most important influence of charge geometry on the resulting effect is
connected with the shape of cumulative hollow. The most often there are used hollows of
conical (with different vertex angle) or hemispherical shape. Shape of cumulative hollow is
above all influenced by the velocity of cumulative jet and the distribution of the mass inside.
It is important to include characteristics of cumulative hollow surface e.g. its smoothness.
2.1.3. Other factors
Other characteristics of the charges, which directly affect exploitation of explosion
energy and/or characteristics of cumulative jet, are very important for the formation of
cumulative jet, which is not formed from the cumulative liner (whose characteristics have
otherwise absolutely fundamental influence). Here belong above all characteristics of the
explosive - especially its homogenity. Basic meaning also has a kind and a way of initiation.
The stand-off distance has rather different meaning at hollow charges than at classical
446
charges with liner, which is connected with the character of cumulated medium, which is
significantly influenced by expansion.
Economic limitations and possible other - specific - requirements are considered here
marginally and only in some cases.
If the meaning of particular factors for resulting qualitative indicators of resulting effects
would be generally evaluated, it is possible to state, that besides the requirement on
preferably maximalised performance, which indeed in case of linear charges must often
submit to demands on necessary mechanical characteristics - for example formability of
explosive mass, is absolutely fundamental requirement of global symmetry of cumulations
process. If we take into account a perfect symmetry of complete charge, absolute smoothness
of its surface and absolute homogenity of the explosive, a determinative factor of final effect
of the charge is the way and quality of initiation.
That is why our further efforts here were concentrated on study of possible influence of
location and the way of LSC model system initiation on the character of formation of the
cumulative jet of products of detonation and its supposed effect.
2.2
The problems of hollow LSC
As it was already mentioned in introduction, there wasnt paid much attention to study
of cumulation jet of products of detonation and to their possible use. Most of available
studies concern common and generally valid descriptions of characteristics and behaviour of
products of detonation except specific area of their cumulation [9]. However complex, wave
and dynamic characters of this process dont allow its exact analyses and practically
exploitable, uncomplicated mathematical description. At present there are not available
simple and effective tools, which would help and/or allow solving these problems within
necessary range without any use of specialised computer programmes.
Escalation development of efficient computer technology and corresponding software
applications in last decades stood for a meaningful advancement in possibilities, which are
connected with implementing of numerical simulation methods. Use of these methods and
procedures, despite of their unexceptionable technical and educational challenge, brings a lot
of possibilities, which dont offer in corresponding range any common experimental practice
in the meantime. It means above all hardly appreciated possibilities of visualisation of
studied processes and significant elimination of financial (transferred material, instrumental,
area, hygienic and other) severity. Problem of numerical simulation is very difficult and
specific for area of engineering activities, which has limited witness ability without any
relevant relationship to some experimental results. Indeed when correct processes and
corresponding verification are applied, it provides really valuable results with corresponding
practical use.
2.3
Explosives for LSC
High-performance plastic explosives on HMX base or more frequently (because of

economical reasons) on RDX base are mostly used for preparation of linear shaped charges.
The characteristic sign of these explosive compositions is a high filling with brisant
component (up to 90%) in reference to reach needed performance, adequate physical and
chemical stability, but while keeping requisite plasticity degree. Use of casted or pressed
explosives is minimal just with reference to impossibility of operative shape adjustment.
These explosives are applied mainly at large volume (usually military) applications.
447
2.4
NUMERICAL SIMULATION
Current modern methodics of numerical simulation use in most cases a final element
method (FEM). FEM is a computational method, whose basic aim is an approximate
determination of response of studied object on outer stress of static or dynamic character.
This method is based on replacement of continuous mediums of material of the object continuum by finite number of discreet elements. They are among each other bonded in a
finite number of contact nodes, which are good for loading outer discreet forces, which
corresponds to model stress and subsequently it enables to determine from changes of their
co-ordinates required response on this stress.
FEM mathematically consists in replacement of exact analytical solution of certain
physical problem by system of differential equations. At a differential solving there are
searched continuous functions for certain area, which are suitable for these equations and
which come to specified values. FEM is an approximate solution, based on discretisation of
given problem [10].
Practical solution of general problem by means of computer programmes, which use
FEM, is formally possible to divide into three basic steps:
1. preprocessing - complex preparation of computational model (discreet of the area of
the solved problem, physical jointing of the elements of the mesh, application of the
system of outer forces and the system of marginal conditions, choice of simple base
functions, choice of a suitable physical law, determination of the condition of the
balance of power in every node). It is done by the user interface of appropriate
programmes -pre-processors.
2. processing - the solving of the problem itself by specialised modules - solvers
3. postprocessing - presentation of the computing results in form of graphic outputs is
solved by other specialised modules - postprocessors
2.4.1. Determination of basic description of model problem

One of the absolutely basic tasks is a creation of geometrical correct model and its
division into suitable dimensioned mesh of finite elements. Other important problem is a
correct choice of equation of state, which describes a relevant model material and a process
description, which proceeds in this material
Material i.e. high explosive is for needs of simulation described mostly with use of
relevant equations of state, in this case there is used JWL (Jonson-Wilkins-Lee) [11] equation
of state in following form:
p S = A e R1 V + B e R2 V + C V (1+ )
(1)
, , C ..... calibration constants (GPA)

1, 2, ..... calibration constants (dimensionless)
V = V/V0 ..... relative volume
Propagation of detonation wave in material - explosive - is realised by two ways with

reference to the real character of initiation, either parameterised according to the detonation
velocity or to pre-defined initiation points.
448
3.
EXPERIMENTAL PART
3.1
Simulation of HLSC detonation with different geometries

of initiation
3.1.1. Pre-processing
In ANSYS programme system there was designed a simplified geometrical model of
hollow linear shaped charge (HLSC) of given dimensional parameters. There was used
charge with temporary marking B450 with dimensions seen from the draft in picture No. 1:
supposed filling 450 g.m-1
19,5 mm
8,4 mm
13,7 mm
200 mm
100
Fig 1.
Shape and basic dimensions of model HLSC
Usual procedure of model creation in ANSYS is well described in the set of pictures in
Fig. 2, where there is also documented a chronology of particular steps.
a)
b)
c)
d)
e)
f)
Fig 2.
a)
b)
c)
d)
e)
f)
Model creation in ANSYS
definition of key points of models front

definition of lines determining surface of models front
definition of surface of models front
extrusion of surface of models front into required volume
division - meshing - of model into elements, including definition of marginal conditions
detail of finished mesh - model
449
Regarding high requirements of modelling on the computational output, it is necessary

to optimise (distribute) the model with respect of adequate severity of the model. The
modelling is usually done in more steps - from simpler units to more complicated details.
Symmetrical problems are solved on divided entity (1/2, 1/4), when possible.
In this particular case the model was processed in 1/2 along symmetry plain. The model
was optimised in several steps. The model was composed from two separately divided
volumes for minimisation of inhomogenities in detonation wave propagation in the model
itself. Division coefficient was 10 (evident for example from lines in picture 2b and from
final mesh 2e,f) - with compression in critical areas at the cavity top. Mesh created by the
programme contained 180 000 final elements of a "brick" type (with 193 161 nodes).
Defined marginal conditions (arrows in picture 2e) eliminated the possibility of elements
expansion across the symmetry plain.
3.1.2. Parameters of used explosive
Explosives, which usually suit to specific requirements to explosive characteristics for
LSC, are developed and framed especially for use in these applications.
Producer of these special explosives is the Experimental institution of industrial
chemistry in Semtn in Czech republic, which produces besides others also an explosive
indicated as Pl SEM (plastic-sheet-explosive-military), which was projected and set in such
way to correspond with mentioned conditions.
Approximate composition of this explosive is: 88% of RDX, 10% of polybutadien
rubber and 2% of fined additives.
JWL parameters of this, to a large extent developing, explosive have not been
experimentally found yet. Absence of these data have been solved in co-operation with
Military Technical Academy in Warsaw, where the needed parameters have been calculated
from the known initial data by a help of CHEETAH code [13]. Result is summarised in table
No. 1.
Table 1. Calculated JWL parameters for Pl SEM explosive
JWL constants
detonation characteristics
[GPa]
636,77
[GPa]
7,61
density (TMD)
0 [kg.m-3]
1570
R1
[-1-]
4,81
detonation pressure
pC-J [GPa]
21,64
R2
[-1-]
1,05
detonation energy
E0 [J.m-3]
8,378
[-1-]
0,34
detonation velocity
vD [m.s-1]
7495
Calculated JWL parameters have limited witness ability, but in case of general study of
the problems, which we have been occupied with in this solution, are more than suitable.
3.1.3. Initiation points
Object of the research was the influence of localisation and the way of initiation of
model LSC system on character of cumulative jet of products of detonation formation and its
supposed effect.
450
There were solved three basic ways from all of the possible alternatives of localisation of
initiation place, which are illustrated in fig. 3a-c. Two alternatives of initiation way preleading initiation source (equivalent of the detonating cord with higher detonation velocity
than the nominal D of used explosive) are illustrated in figures 3d,e.
Since a certain real practical benefit have only three ways - a, b and d, there were
submitted for further investigation mentioned possibilities.
b)
a)
c)
e)
d)
Fig 3.
Possible places and ways of initiation
3.1.4. Processing
Simulation of detonation of a hollow linear shaped charge was solved by ANSYS - LS
Dyna3D solver [12] in version 6.1 for PC with Windows 2000 system. Average time of
solving of a simulated problem was cca 7 hours at used computer (Athlon XP1600+,
256 MB SDRAM). The simulation was realised to reaction time 5,5 sec of the initiation.
This time is sufficient to form a cumulative jet, in order to enable to find its final parameters.
Simulation for further products of detonation expansion (longer times) was connected with
numerical instabilities, which it would be necessary to treat and this would be very
complicated. Nevertheless this contribution of such efforts would be irrelevant for the results
of research.
Fig 4.
absolutely rigid
wall
Idealised model for

study of dynamic
response of cumulative
jet of products of
detonation at LSC
451
For research of value, which could give us needed information about important
characteristics (e.g. above all its kinetic energy, eventually a power, by which it can effect on
materials in its surrounding, and which it is not possible to determine by common
experimental procedures at present) of formed cumulative jet, it was necessary to use a
modified model for the simulation. Therefore there was simulated an impact of expanding
products of detonation on idealised, absolutely rigid wall in place of the base of cumulative
hollow - see fig. 4.
For this specific case severity of the model had to be reduced by lower division factor
(n=2) in critical area. Interaction of the cumulative jet at fine division of the mesh
disproportionately increased time consumption of the calculation without distinct
improvement of the results accuracy.
3.1.5. Postprocessing
Processing of result simulation files was realised by the form of graphic outputs across
marginal line of postprocessor LSTC/ LSPOST ver. 2.0 - 17JUL2002(1622) [14].
Possibilities of outputs processing of the simulation itself by the help of particular tools,
which the postprocessor offers, are extremely large. Besides others they enable to observe
the whole process not only from global view, when various marginal conditions and limits
are defined, we can observe characteristics and display of observed process from outer view,
but thanks to possibilities of cut systems, it is possible to watch also various changes of
profiles of important values in particular places inside the charge. Absolutely basic meaning
is then the possibility of division of the observed process to needed small (or big)
geometrical and time sections, which allow us to focus directly on important places and
times of studies processes, including the most illustration animations.
From the view of qualitative importance of big set of particular values, which it was
possible to observe and to evaluate, for checking resulting performance characteristics, as the
most important appeared the following ones:
at common model:
pressure for checking a possible general destruction effect of particular
parts of cumulative jet in time
velocity in direction of cumulative jet propagation - regarding idealised
conditions of studied process (condition of surroundings vacuum - no
interaction with the surrounding and unlimited expansion of products of
detonation) have the absolute values of the velocity the minimum witness
ability, but orientation of compression and expansion of the products of
detonation in given profile of the charge allows relatively good to review
the efficient size of active part of the charge products of detonation
at the model with interaction on absolutely rigid wall:
the force effected on the interface - enables to review ideal power effect of
influencing products of detonation and according to this parameter to
evaluate a hypothetical performance of the charge
3.1.6. Problem of practical use of outputs
A definite limit factor for quality and real utility of the results is the relationship with
available experimental data, which would guarantee a verification of used models,
procedures and simulation outputs. Also without this possibility it is real, if respecting of
basic mathematical and physical rules and axioms and with binding on practical experience
with related problems, to express certain, though infinite, conclusions with sufficient witness
value.
452
4.

4.1
HLSC with initiation in the middle of charge
Simulation of detonation of HLSC of given parameters at the initiation according to the

scheme in fig. 3b provided result values of observed values (according to chap. 3.5.6), which
can the best document following graphic records - fig. 5,6,7, 8 and 9.
Fig 5.
Final shape of formed products

of detonation at HLSC initiated
in the middle of charge +
indicated position
of an analytical cut
Fig 6. Formation of cumulative jet at HLSC cuts

fringes of pressure at charge initiated
in the middle
- interval of particular cuts ~ 6 mm
453
Fig 7.
Fig 8.
Indicated position of analytical

cut for study the profile of
cumulative jet velocity at the
charge initiated in the middle
Formation of cumulative jet of HLSC cuts

fringes of velocities in profile of cumulative jet
at the charge initiated in the middle
- interval of particular cuts ~ 4 mm
454
Fig 9.
4.2
Graph of history of the force of cumulative jet effected on the interface on the ideal
rigid wall at the charge initiated in the middle.
HLSC with initiation at the edge of charge
Simulation of HLSC detonation of given parameters at initiation according to scheme in

fig. 3a provided result values of observed values (according to chapter 3.5.6), which are
documented in pictures 10,11, 12 and 13.
Fig 10. Final shape of formed products
of detonation at HLSC initiated
at the edge of charge + indicated
position of analytical cut
455
Fig 11. Formation of cumulative jet of HLSC cut

fringes of pressure at charge initiated at the edge interval of particular cuts ~ 4 mm
Fig 12. Indicated position of analytical

cut for following a profile
of cumulative jets velocities
at charge initiated at the edge
456
Fig 13. Formation of cumulative jet of HLSC cuts

fringes of velocities in profile of cumulative jet at charge initiated at the edge
interval of particular cuts ~ 4 mm
Fig 14. Graph of history of the force of cumulative jet effected on the interface on the ideal
rigid wall at the charge initiated at the edge
As it is evident from the pictures, but especially from graphs of filtered (filter was SAE
600Hz*1000) data in pictures 9 and 14, behavior of cumulative jets of products of detonation
at charges, which were initiated in the middle or at the edge of the charge, is very similar
including supposed force effects in the given surface - max. effecting forces 564 kN at the
initiation from the edge and 1147/2 i.e. 574 kN at the initiation from the middle of charge.
Then it is possible to give a competent presumption, that localization at point initiation is not
important from the view of energy utilization. It is possible to expect a similar situation also
at reached effects. But we have to remember specifications of the mass acceleration, which is
realized at speeding up of materials of cumulative liners in practical applications, where
there is more suitable a side initiation with respect to necessary stabilization of cumulation
process. At the side initiation it is therefore possible to take into account a smaller volume of
ineffectively utilized explosive for stabilization of formation of cumulative blade.
457
4.3
HLSC with pre-initiation
Simulation of HLSC detonation at initiation according to fig. 3c, i.e. with initiation
equivalent use of detonating cord with higher detonation velocity than nominal detonation
velocity of used explosive is, provided result values of observed values (according to chap.
3.5.6), in graphical illustration in pictures 15 - 18.
Fig 15. Final shape of formed products
of detonation at HLSC with
initiation in advance from the
edge of the charge + indicated
position of analytical cut

fringes of pressure at charge with initiation in advance (from the edge of charge)
458
Fig 17. Indicated position of analytical

cut for observing the profile of
velocities of cumulative jet at
charge with initiation in advance
459

fringes of velocities in profile of cumulative jet
at charge with initiation in advance
Fig 19. Graph of history of the force of cumulative jet effected on the interface
on the ideal rigid wall at the charge with pre-initiation
As it is evident from pictures 16 and 17 the initiation in advance causes different
formations of cumulative jet of products of detonation. Everything indicates, that this way
leads to the most effective utilisation of cumulative jets energy in given model setting,
probably by a positive influence of the way of expansion of products of detonation. This is
evident above all from graph in picture 19, where it's possible to read maximum values of
effecting forces 653,9 kN, which is cca more than 15% higher value in comparison to forces,
which are provided at point initiations (fig. 9 and 14). This difference it is possible to follow
in fig. 20, where there are side by side carried out dependencies of effecting forces at the
charge point initiated from the side and at the charge with initiation in advance.
Fig 20. Comparison of history of the forces of cumulative jets effected on the interface
on the ideal rigid wall at the charge with pre-initiation and charge initiated in the
middle
460
It is then possible to presuppose that at the utilisation of similar way of initiation there
will be reached higher performance characteristics at linear shaped charges on the whole.
5.
CONCLUSION
There were successfully realised simulation experiments with use of FEM in ANSYS
system, which should help to determine at least approximately to which there can prove
an influence of initiation on detonation process, formation of cumulative jet of products
of detonation and its supposed final effect at model linear shaped charges without liner.
Results of simulations indicate, that localisation of point initiation is not important in
terms of utilisation of energy at studied system. On the other hand at system with preinitiation there it is evidently reached a more effective utilisation, probably by positive
influence of the way of expansion of products of detonation, of cumulative jet energy.
It is then possible to suppose that use of the system with pre-initiation will signify
a positive contribution for final performance characteristics also in real, practical
applications.
Acknowledgement:
We would like to thank to prof. Waldemar A. Trzcinsky from Military Technical
Academy in Warsaw (WAT) for his help with calculation of the theoretical values of JWL
parameters for the used explosive by the help of the computational CHEETAH code.
REFERENCES
[1] WEICKERT, CH.A.: (1998) Demolitions (Explosive Effects and Applications, Zukas, A.
and Walters, W.P., Eds.), 10, Springer-Verlag, New York
[2] WALTERS, W.P. AND ZUKAS, J.A.: (1989) Fundamentals of Shaped Charges, Wiley, New York
[3] Special Explosive Products, publicity matter, Research Institute for Industrial Chemistry,
Synthesia Corp. Pardubice-Semtn
[4] BLADE The Cuting Edge, publicity matter, British Aerospace Royal Ordnance
[5] VGENKNECHT, J.: (1995) Classification of effects of hollow charge, diploma work, University
of Pardubice (in Czech)
[6] VGENKNECHT, J.: (2000) Using of hollow charges for determination the detonating pressure
of high explosives, Ph.D thesis, University of Pardubice (in Czech)
[7] VGENKNECHT J. AND ADAMK L.: Determination of the characteristics of a cumulative jet
of products of the detonation, 4th International Seminar New Trends in Research of Energetics
Materials, Univ. Pardubice, April 2001, p. 329.
[8] VVRA, P., VGENKNECHT, J.: Teorie psoben vbuchu, teaching texts,
Univerzity of Pardubice, 1. publication 2002. (in Czech)
[9] BAUM, F.A., ORLENKO, L.I. STANJUKOVI, K.P., CHELISHEV, V.I., SHEHTER B.I.: (1975)
Fizika gorenja i vzryva, Nauka, Moskva
[10] SCHLOSBAUER V.: FEM Finite Element Method 1999, review, DFJP,
University of Pardubice (1999) (in Czech)
[11] DOBRATZ, B.M. AND CRAWFORD, P.C.: (1985) LLNL Handbook of Explosives,
UCRL-52997, Lawrence Livermore National Laboratory, Change 2, 8-21 - 8-23
[12] ANSYS: (1996) Ansys/LS-Dyna3D, theoretical manual
[13] L. FRIED, P.C. SOUERS, W.M. HOWARD, AND P. VITELLO: Cheetah a thermochemicalkinetics code, software, LLNL (2000)
[14] LSTC/ LSPOST ver. 2.0, postprocessor for LS DYNA, software, ANSYS, 2002
461
TRACE ANALYSIS OF POST EXPLOSION RESIDUES

OF INDUSTRIAL EXPLOSIVES BY MEANS OF GC-ECD AND IC
R. Varga * and Ulbrich P. **
* Department of Theory and Technology of Explosives, Univerzity of Pardubice,
** Institute of Forensic Science of the Slovak Police Corps,
Sklabinsk 1, Bratislava 812 72, Slovak Republic
Abstract:
The paper deals with possible applications of GC-ECD and IC to specification of post
explosion residues of secondary explosives. Technical TNT and dynamite Danubit 2
have been used as real secondary explosives: at present they are the most frequently
abused explosives on the territory of the Slovak Republic and Czech Republic.It has
been shown that distilled water is a good medium for collecting the post explosion
traces of nitro compounds (TNT and DNT) in testing ground. The separation,
purification and simultaneous concentrating of aqueous extracts of post explosion
residues was realised by means of SPE (the extraction discs type C18). The procedure is
simple and allows obtaining extracts for organic and inorganic analyses in a single
operation, at a considerably preconcentration the traces of organic explosives. The
ECD detector was shown to be highly selective and its combination with GC makes it an
efficient tool in trace analysis of organic components of post explosion residues.
Similarly, also ion chromatography (IC), thanks to its sufficient sensitivity, is well
applicable to trace analysis of inorganic components in these residua.
Keyword:
1.
TNT, dynamites, post explosion residues, trace analysis, GC-ECD, IC.
INTRODUCTION
The aim of analysis of post explosion products is to determine the type of explosive
material that exploded on the scene of the crime (bomb attack, non-professional production
etc.). This task is considerably complicated by the number of unfavourable factors: trace
quanta of non-reacted explosives are analysed which are dispersed to a radius of several
meters and are accompanied by interfering impurities in amounts that are orders of agnitude
higher. High demands is the collecting of the traces, their extraction and purification of the
extracts as well as the highly sensitive and relevant analytical evidence [1]. One of the best
media used for collecting of the traces is preferably water [2], which in contrast to organic
solvents dissolves only a negligible part of the organic component of the explosive
depending on its solubility at the given temperature. Another advantage of water as
compared with organic solvent is the fact that it is environmentally friendly [3].
One of the simplest but also most efficient pre-separation techniques, which are used
in analyses of aqueous solutions, is solid phase extraction (SPE) [4]. It can be used also
in the case of analyses of the post explosion extracts [5]. An experiment of this type was
described
462
by THOMPSON et al. [3]. TONTARSKI et al. [6] found out that organic explosives (RDX,
PETN and TNT) are best sorbed on non-polar sorbents such as C18.
A procedure of purification of post explosion products (removing polar impurities that
catalyse thermal decomposition of PETN) for gas chromatography was described by KOLLA
[7]
. The above-mentioned complexity of analysis of post explosion products necessitates
application of at least two independent analytical methods in order to get relevant results.
These methods-used for organic component-involve, e.g., HPLC/TEA, LC/MS, GC/ECD,
GC/TEA, GC/MS. For the inorganic components it is possible to use colour reactions or ion
chromatography [1].
Gas chromatography (GC) is frequently applied to investigation of traces of explosives
in the post explosion extracts. Improvements of capillary columns have brought an increase
in applicability of GC to analyses of thermally instable explosives (liquid nitrate esters
and nitramines) [8].
The majority of industrial explosives (e.g. dynamites) contains liquid nitrate esters
belonging among thermally instable secondary explosives [9]. Besides the active basis of
dynamites (which is blasting gelatine: 8 % NC and 92 % NG), they contain oxidizing agents
(e.g. ammonium nitrate) nitroaromatic compounds (DNT and TNT), wood flour,
nitrocellulose, dyestuff, trace additives that are specific for individual manufacturers.
Determination of wood flour and dyestuff by gas chromatography is out of question because
their transformation into vapours is impossible [10].
For dynamites it is characteristic the presence of NG and EGDN; however, these
components are hardly to analysed because they are evaporated due to the high explosion
heat. The components having higher vapour pressures partially disappear when the post
explosion gases are condensed on cold surfaces: therefore, it is better to compare only
compounds with similar vapour pressures [10].
At present, dynamite explosives can be detected very selectively thanks
to the nitroaromatic compounds (most frequently DNT and TNT) presented in slight
amounts within the complex composition of extracts of post blast residues. The
carcinogenicity of nitroaromatics (e.g. 2,6-dinitrotoluene) has led to their gradual
replacement by less toxic but harder analysed explosives [10].
The applicability of GC-ECD to analyses of post explosion products is reported
by MCCORD & BENDER [11]. Practical applications in this area are described by KAPLAN
& ZITRIN [12]. As the ECD detector becomes easily contaminated, the GC-ECD method can
only be used for analyses of relatively pure extracts. Its selectivity makes it possible
to analyse organic secondary explosives containing nitro groups. The application of ECD is
also very important in the identification of secondary explosives presented in matrices
contaminated with hydrocarbons [13].
Ion chromatography (IC) possesses sufficient sensitivity to be favourably applicable
to trace analysis of post explosion product of industrial secondary explosives existing
in ionic forms [1]. Its complex utilisation in analyses of cations and anions in samples
of industrial secondary explosives as well as their post explosion residues was reported
by REUTTER et al. [14].
The verification of GC-ECD and IC in the applications for specification of post
explosion residues of secondary explosives most frequently abused on the territory of the
Slovak Republic and Czech Republic (TNT and Danubit 2) is dealt with in M.Sc. Thesis [15].
This present paper reports the most important findings of the said Thesis.
463
2.
EXPERIMENTAL
The most frequently misused secondary explosives on the territory of CZ and SR

involve TNT and the dynamite explosive Danubit 2. These two products were used in Thesis
[15]
for verification of analytical techniques of the post explosion residues. The secondary
explosive Danubit 2 is composed of ethylene glycol dinitrate (EGDN), glycerol trinitrate
(NG), cellulose nitrate (NC), dinitrotoluene (DNT), ammonium nitrate (AN), sodium nitrate
(SN), wood flour, carboxymethylcellulose, micro-ground limestone and ferric oxide (as a
dyestuff).
2.1
Used Apparatus and Instruments
2.1.1. Apparatus for SPE

For the purification and pre-concentrating of post explosion products were used Empore
3M octadecyl C18 extraction discs, which are suitable for environmental analyses. The SPE
disc was attached to the grid of extraction apparatus, and the sample was sieved through the
SPE disc, the necessary underpressure was created by a water-jet pump.
2.1.2. Apparatus for Analyses of Extracts
Gas Chromatograph with Electron Capture Detector (GC-ECD)
The standards of explosives (TNT, DNT, NG a EGDN) and the SPE extracts obtained
were analysed on a Hewlett Packard 5890 chromatograph, using a capillary column HP-5
of 5 m length, 0.53 mm inner diameter, and 0.88 m film thickness. An electron capture
detector (ECD) was used for the detection, and the carrier gas was nitrogen of 20 kPa
overpressure. The sample was injected directly (on-column), the injector temperature being
70 C. The temperature regime was as follows: the column was kept at 70 C for 1 min,
whereupon the temperature was increased at a rate of 15 C/min to 150 C, and 5 C/min
to 230 C, and this final temperature was kept 8 min.
All the GC-ECD chromatograms
of CHEMSTATION program.
measured
were
evaluated
by
means
Ion Chromatograph
The aqueous extracts filtered through solid phase (SPE) were submitted to analysis
by means of an ion chromatograph Dionex DX-20. The analysis of cations took place
in a system consisting of a pre-column CG 12A (4 50 mm), a column Ionpac CS 12A (4
250 mm), and a suppressor CSRS-Ultra 4 mm. The system for analysis of anions consisted
of a pre-column AG 9-HC 4 mm, a column Ionpac AS 9-HC 4 mm, and a suppressor ASRSUltra 4 mm. The elution solvents for cations and anions were 2.2 mN H2SO4 (1g H2SO4 per
1l H2O) and 9.0 mM Na2CO3 (18 ml Na2CO3 per 1l H2O) with flow-rates of 1.00 and 1.07
ml/min at the overpressure of 2300 - 2400 and 1600 psi (1 psi = 6.89 kPa).
464
2.2
Field Work and Sample Collecting
All the experimental explosions were carried out in a stone quarry in Mal Karpaty
(Mountains) near Bratislava in Slovakia. The samples used were low-diameter charges
of dynamite (Danubit 2) of total masses 200, 400, and 600 g, and tritol charges of total
masses 200 and 400 g. The initiation was accomplished by means of Initiator No. 8.
The individual charges were placed on the terrain in such a way as to facilitate collecting
the post explosion samples. Also the particular spots of explosions were chosen carefully so
as to avoid any contamination from previous explosions. After the detonation, samples were
taken from the explosion epicentre and surrounding surfaces of stones near the place
of explosion (roughly one and two meters from the explosion epicentre). The traces of post
explosion products were collected by wiping off c. 1 m2 stone surface area with a wet cloth.
The extraction was carried out using 700 ml distilled water. The wet cloth with collected
traces of explosives was wrung out into a clean bottle. The aqueous extracts of post
explosion products obtained in this way had the volume of c. 200 ml each. The same
procedure was used for all the samples.
2.3
Preparation of Samples for Solid Phase Extraction (SPE)
At first, all the investigated aqueous samples of post explosion residues from dynamites
and TNT were left to sediment. Before proper concentrating the analyte and its simultaneous
purification by the solid phase extraction (SPE), the extracts were filtered through a paper
filter (blue ribbon) in order to remove mechanical impurities.
2.4
Solid Phase Extraction (SPE)
After preliminary purification of the samples by filtration they were subjected to

extraction with solid sorbent C18: 50 ml aqueous extract was left to flow through the
extraction disc. Water soluble components of the explosive flowed through the disc while the
organic ones remained trapped on it and were subsequently washed out with a small amount
of organic solvent (10 ml n-hexane). Details of this separation technique are given
elsewhere [6, 7]. The extractions provided organic and aqueous extracts: the former
(containing n-hexane) were used for analysis of explosives of organic origin (TNT, DNT,
EGDN and NG), while the latter were used for analysis of inorganic water-soluble
components of explosives (Na+, NH4+ a NO3 ions).
3.

3.1
Solid Phase Extraction
Considerable attention was paid to the choice of purification procedure since the electron
capture detector (ECD) used is easily contaminated.
As already stated, the separation, purification and pre-concentrating of the analytes from
the post explosion extracts were realised with the help of extraction discs C18. This
purification technique was time consuming, and moreover, it was impossible to filtrate equal
volumes through the discs. Fine colloid particles collected together with the post explosion
residues during sample collecting could not be perfectly removed by the paper filter.
Therefore, the pores of extraction disc get blocked after some time, which also affects the
flow rate of n-hexane during washing out the organic extracts from the extraction disc.
At a slow flow rate and large underpressure, the n-hexane tends to evaporate rapidly,
and the purified organic extracts obtained have then various volumes.
465
Hence, the drawback of SPE is low rate of processing of analyte solutions if the samples
are considerably contaminated. Purification of 50 ml aqueous solution of post explosion
products can take as long as 30 - 60 min. Advantages of this procedure are its simplicity
and the possibility to obtain the extracts for both organic and inorganic analyses
simultaneously, the traces of organic explosives being considerably pre-concentrated.
3.2
GC-ECD Analysis of TNT and Danubit 2 in Organic Extracts

of Post Explosion Residues
On the basis of literature data and structure of the substances determined, gas
chromatography with ECD was chosen for the given analysis. The reason of this choice lies
in the high sensitivity of the electron capture detector (ECD) to electronegative atoms
(nitrogen compounds) and its low sensitivity to contaminants. The said detector was
calibrated by means of standard solutions (2,4,6-TNT, technical DNT, EGDN and NG)
with subsequent specification of detection limit for selected types of explosives (see Table
1). The amount of sample injected was low (1l). Thermally unstable explosives (EGDN,
NG) were analysed with the application of direct injection (on-column) and a short capillary
column (5 m).
Table 1. The detection limits for the chosen types of explosives
Standard of explosives
TNT
DNT
EGDN
NG
Limits of detection of ECD (mg/l)

1.60
1.62
2.38
---
At the above-given conditions of GC-ECD analysis, the analysis time

for the investigated group of compounds was up to 6 min. All the purified and concentrated
extracts of post explosion residues from TNT and Danubite 2 were measured at identical
conditions of the analysis. The qualitative analysis was based on comparison of retention
times of the analytes and standards. The quantitative analysis (concentration)
of the individual components in the post explosion extracts was based on the calibration
of detector (the surface areas of peaks). The results of measurements are summarised
in Tables 2 and 3.
Table 2. Amount of TNT in different distances
Type of
charge
Amount of
charge (g)
Place of
sampling (m)
Peak Area of
TNT
TNT
200
200
200
400
400
400
0
1
2
0
1
2
35053620
40562530
23883650
6219390
60223020
28728040
Concentration
of TNT in post
blast residues
(mg/l)
28,88
33,16
20,19
6,44
48,46
23,96
466
Table 3. Amounts of DNT and EGDN in different distances from explosion

of Danubit 2
Amount
Place of
Type of
of
Peak Area
sampling
charge charge
of DNT
(m)
(g)
Danubit
2
200
200
200
400
400
400
600
600
600
0
1
2
0
1
2
0
1
2
8464321
51186175
31112250
67955900
42370730
43039080
23014090
2647196
5401747
Peak
Area of
EGDN
457185
73960
47453600
1369140
1708240
250146
4198780
U
U
Concentration Concentration
of DNT in
of EGDN in
post blast
post blast
residues
residues
(mg/l)
(mg/l)
14,55
2,69
79,83
2,43
49,16
35,29
105,45
3,33
66,36
3,56
67,38
2,55
36,78
5,29
U
5,67
U
9,87
The GC-ECD chromatograms of post explosion products from TNT (Fig 1.) showed
three peaks identifiable by comparison with the retention times of standards. In the case
of TNT charge, the organic extracts always contained TNT in each sample irrespective
of the charge mass and the distance of sample collecting from the explosion epicentre (Table
2). The TNT concentration in the post explosion extracts varied from 6 to 50 mg/l. Besides
TNT, also the presence of DNT could be detected, which is typical of technical tritol (DNTs
are intermediates in production of TNT). It must be pointed out that TNT for civilian
applications usually is of lower purity than military TNT.
The GC-ECD chromatograms of post explosion products from Danubite 2 exhibited four
dominant peaks (Fig. 2). The first was identified as the solvent (n-hexane), and the other
three were EGDN, 2,4- and 2,6-DNT. In each sample of the organic extracts both the somers
of dinitrotoluene were present. In the case of the 600 g charge of Danubite 2, the result was
four times negative for EGDN, see in the samples taken 1 and 2 m from the explosion
epicentre (Table 3). The detection limits were in the range from 2 to 100 mg/l
for the dinitrotoluenes, and from 2 to 35 mg/l for EGDN. Using its characteristic retention
time, NG was not detected at all although it forms-along with EGDN-a component of
industrial secondary explosives of dynamite type produced on the territories of CZ and SR.
The analysis was not designed to identify other components of Danubit 2 (NC, wood flour,
carboxymethylcellulose, dyestuff). Partial failure of identification of EGDN and NG was
probably due to their thermal decomposition during explosion or to hydrolysis during sample
collecting.
Splitting of organic explosives can proceed in the shock wave or at the contact
with surrounding environment (rock) and air moisture. The samples of aqueous extracts
of post explosion residues before SPE were kept in refrigerator at lowered temperature
in order to minimise the losses due to hydrolysis. Since the extraction of traces of explosives
was performed with water, the results of analysis of the components followed could be
partially affected by hydrolysis during the period of time between sample collecting
and subsequent sample purification by SPE.
467
TNT
2,6-DNT
2,4-DNT
Fig 1.
GC-ECD chromatogram of post explosion analysis of TNT

2,6-DNT
EGDN
2,4-DNT
Fig 2.
GC-ECD chromatogram of post explosion analysis of Danubit 2
468
3.3
Analysis of Cations and Anions in Aqueous extracts of Post

Explosion Residues from Danubit 2
The portion of aqueous extracts from Danubit 2 obtained from SPE was analysed
for the presence of inorganic components by means of IC. The results are summarised
inTable 4.
Table 4. Amounts of cations (Na+and NH4+) and anion NO3- in different distances
from Danubit 2
Type of
charge
Danubit 2
Amount
of charge
(g)
200
200
200
400
400
400
600
600
600
Place of
Contentration Contentration
Contentration
sampling
of NH4+
of NO3+
of Na (mg/l)
(m)
(mg/l)
(mg/l)
0
1,94
3,59
194,48
1
7,06
5,14
35,91
2
7,55
2,12
15,38
0
4,10
8,49
439,93
1
21,39
14,33
159,08
2
9,62
10,63
89,18
0
2,46
6,58
488,84
1
8,00
8,30
76,25
2
12,93
20,17
201,71
Since ammonium nitrate forms a basic component of Danubit 2, predominant attention

was paid to the NO3 anion. Besides Na+, NH4+, Ca2+ and NO3 ions (see Figs 3 and 4),
which are components of this dynamite, also other ions were found in the samples. Among
the most significant ones there were K+, Cl, PO42 and SO4 (see Figs 3 and 4). In the cases
of potassium, chloride, phosphate and sulphate, the presence is due to contamination of real
samples by ions from environment. The presence of nitrites can be interpreted by reduction
of the oxidising agents (NH4NO3 and NaNO3) during the explosion. The ions detected
in the post explosion extracts were Na+, NH4+ and NO3. The detection limits varied from 2
to 20 mg/l for Na+ and NH4+ cations and from 15 to 490 mg/l for the NO3 anion.
Retention Time (min)
Fig 3.
IC chromatogram of cations at real samples of post blast residues

of Danubit 2 Peaks: (1) Na+, (2) NH4+, (3) K+, (4) Mg2+ and (5) Ca2+.
469
Retention Time (min)
Fig 4.
3.4
IC chromatogram of anions at real samples of post blast residues

of Danubit 2. Peaks: (1) F, (2) Cl, (3) NO2, (4) NO3, (5) PO42 and (6)
SO4.
Distribution of Post Explosion Residues
A very important factor is correct determination of the place with the most probable
occurrence of traces of explosives. The knowledge of distribution of post explosion residues
forms a basis of criminological expertise of attacks with explosives. Therefore, samples were
taken at various distances from the epicentre of explosion, the trapping surfaces being
located perpendicularly to the direction of expansion of gases. In order to map
the distribution of particles, suitable experimental arrangement was adopted with metal
plates of 1 m2 surface area located at various distances from explosion epicentre, allowing
determination of the amount of explosive per 1 m2 at a given distance. This experiment was
designed to find approximate distribution of non-reacted particles in the determination
of sensitivity of GC-ECD and IC for trace analysis of post explosion products.
The values measured (see Tables 2, 3 and 4) were plotted in diagrams showing
the dependence of amount of post explosion residues upon the distance from the explosion
epicentre (see Fig. 5a - DNT, Fig. 5b - TNT, Fig. 6a - NH4+, Fig. 6b - NO3). All
the diagrams show the distribution of non-reacted particles of original explosives (TNT
and Danubit 2) at various distances.
The dispersion of post explosion residues is a very complicated problem, because
the distribution of these particles during explosion is affected by a number of unfavourable
factors (first of all mass, type of the explosive used, and direction of initiation).
The probability of occurrence of non-reacted particles increases with decreasing efficiency
of the secondary explosive, with its decreasing mass, and with decreasing diameter of its
charge.
In the case of detonating secondary explosives, the non-reacted particles are accelerated
by the expansion of the gases formed, in the final phase of the flight they could to be
affected by wind. The particles are heated by the shock wave, and in some cases they can
have the form of droplets of melt, which subsequently solidify on cold surfaces of the
surrounding environment. The non-reacted particles of explosives are present in very slight
quanta in the vicinity of the explosion, being irregularly dispersed on a large surface from
the explosion epicentre.
470
60
100
50
80
200 g
60
400 g
600 g
40
20
Amounts (ppm)
Amounts (ppm)
120
40
200 g
30
400 g
20
10
0
0
Place of sampling (m)
Fig 5a
Fig 5.
Fig 5b
Amounts of DNT (Fig 5a) and TNT (Fig 5b) in different distances (mg/l)
25
20
Amounts 15
(ppm)
10
5
0
0,0 m 1,0 m
2,0 m
Place of sampling
(m)
Fig 6a
Fig 6.
4.
Place of sampling (m)
600 g
400 g Amount of
200 g
charge (g)
500
450
400
350
Amounts 300
250
(ppm)
200
150
100
50
0
0,0 m 1,0 m
2,0 m
600 g
400 g Amount of
200 g
charge (g)
Place of sampling
(m)
Fig 6b
Amounts of cations NH4+ (Fig 6a) and anions NO3- (Fig 6b) in post
explosion extracs of Danubit 2
CONCLUSION
The study results have shown that distilled water is a good medium for collecting traces
of nitro compounds (TNT and DNT) because it does not extract any large amounts
of undesirable impurities (it provides the purest extracts). Another advantage of water is
the possibility to avoid organic solvents with concomitant posing a lower ecological load
on the environment. However, the application of water can also bring considerable
complications in analyses of nitrate esters (NG, EGDN) due to their hydrolysis, which can be
partially eliminated by keeping the extracts at low temperatures or in dry state or-best of allby quick processing of the extract.
471
The separation, purification and simultaneous pre-concentrating of the explosives

collected in water from the post explosion extracts were carried out by solid phase extraction
(extraction discs of C18 type). This procedure is advantageous in being simple and allowing
extracts to be obtained for both organic and inorganic analyses at the same time, the traces
of organic explosives being considerably pre-concentrated. At present, SPE represents a
dynamically developing separation technique and is more and more often adopted in
analyses of post explosion products.
The results of measurements confirmed the high selectivity of ECD detector, its
combination with GC being effective in trace analysis of organic components of post
explosion products. The advantages of GC are in the speed of analysis and simplicity
of measurements.
The analysis of aqueous extracts was carried out by ion chromatography (IC), which
managed to prove the presence of the cations (Na+, NH4+) and anion (NO3) followed. It was
confirmed that IC, thanks to its sufficient sensitivity, is well applicable to trace analysis
of inorganic components of post explosion products.
The experiments carried out showed that the tested dynamite (Danubit 2) can be detected
in post explosion products by the presence of DNT (GC-ECD) and Na+, NH4+ and NO3
ions. A more precise identification of this type of dynamite can also make use of the
occurrence of EGDN in its post blast residues. The monitoring of distribution of the nonreacted particles showed that collecting of traces of explosives for reliable detection is
sufficient as far as the distance of 2 m from the explosion epicentre.
With regard to the complexity of both the matrix and the analyses it is impossible
to develope a simple method for analysis of traces of explosives in post blast residues.
Acknowledgement
The authors wish to thank the management of Institute of Criminology of Police Forces
of the Slovak Republic in Bratislava for making possible and supporting the realisation
of this work in the Institute.
472
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
M. HANUS: Analytick chmia vbunn pre kriminalistickou expertzu, Disertan prce,

FCHT Pardubice, 1999.
J. CHLDEK: Separation and Preconcentration of Postblast Residues by Salting-Out Solvent
Extraction, International Symposium, Pardubice, 55-58, 1996.
ROBERT Q. THOMPSON, DEAN D. FETTEROL, MARK L. MILLER and ROBERT F.
MOTHERSHEAD II: Aqueos Recovery from Cotton Swabs of Organic Explosives Residue
Followed by Solid Phase extraction, Vol. 44(4), Journal Forensic Sci., 795-804, 1999.
J. CHUREK A KOL: Nov trendy v teorii a instrumentacii vybranch analytickch metod,
Academia Praha, 1993.
M. HANUS: Zklady analytickej chemie vbuin (skripta pro inenrske studium oboru
Teorie a technologie vbuin ), FCHT Pardubice, 1999.
R. A. STROBEL, R. E. TONTARSKI: Proc. 1 st Int. Symp. On Analysis and Detection
of Explosives, FBI Academy, Quantico USA, 67, 1983.
P. KOLLA: J. Forensic Sci. 36, 1342, 1991.
J. YINON, S. ZITRIN: Modern Methods and Applications in Analysis of explosives,
Chichester, England: John Wiley & Sons Ltd., 1996.
S. BREBERA: Trhaviny rzn konzistence, Vojensk trhaviny II, Uebn texty, Pardubice,
2002.
P. KOLLA, A. SPRUNLER: Identification of Dynamite Explosives in Post Explosion Residues,
Vol. 40, Journal of Forensic Sciences, JFCA, No. 3, 406-411, 1995.
B. McCORD, E. C. BENDER: Forensic Invest. Explos. 231, 1998.
M. A. KAPLAN, S. ZITRIN: J. Assoc. Off. Anal. Chem. 60, 619, 1977.
J. CONNOR: J. Chromatogr. 200, 15, 1980.
D. J. REUTTER, R. C. BUECHELE: Ion chromatography of explosives and explosive residues,
Proc. 1 st Int. Symp. On Analysis and Detection of Explosives, FBI Academy, Quantico USA,
199-207, 1983.
R. VARGA: Prspevok k analze povbuchovch splodn, Diplomov prca, FCHT Pardubice,
2002.
473
EFFECT OF INTERMOLECULAR FORCES

ON SOME PROPERTIES OF EXPLOSIVES
II. INFLUENCE OF NON-COVALENT HYDROGEN BOND
Pavel Vvra* and Miroslav Pospil**
* Department of Theory and Technology of Explosives
Faculty of Chemical Technology, University of Pardubice
53210 Pardubice, Czech Republic
** Department of Chemical Physics and Optics,
Faculty of Mathematics and Physics, Charles University
12116 Prague, Czech Republic
Abstract
On the basis of known structural data, the values of total energy of crystals have been
calculated for a selected set of explosives, and their non-negligible effect on sensitivity
parameters, melting temperatures and densities has been found. The presence of noncovalent hydrogen bonds and/or van der Waals forces in crystalline individual
explosives distinctly lowers the impact sensitivity, increases their melting temperatures
and contributestogether with other factorsto their enhanced density. Possibility of
prediction of these parameters for proposed structures of new explosives is discussed.
Keyword:
1.
intermolecular forces, calculations, high explosives, HE, hydrogen bond,

properties of molecules
INTRODUCTION
From the calculated values of total sublimation energy and their contributions (van der
Waals, Coulomb, and Hydrogen bonds) obtained for a selected set of explosives, a nonnegligible effect of inter-forces on sensitivity, density and melting temperature has been
found [1]. The set of the substances tested included both linear and cyclic C-, N-, and O-nitro
compounds whose molecular structure is known [2].
In order to verify the above-mentioned effects on the said properties, crystal energies
related to the basic crystal cell were calculated for the same set of substances with the aim of
more precise specification of the effects of the individual components of total crystal energy
on energetic materials.
2.
CALCULATIONS
The calculations of total crystal energies of the explosives were carried out in Cerius2
modelling environment [3]. The crystal structures were built in Crystal Builder module on the
basis of published X-ray diffraction data. These structures were not anyway changed after
the building.
The total crystal energy of these investigated structures was calculated in the Minimizer
module with using the universal force field (UFF) [4]. This empirical force field contains
474
description of all atoms in the periodic table. The choice of UFF allows us to compare the
total sublimation energy [1] and its components: van der Waals, Coulomb and Hydrogen
bond energy with the total crystal energy, especially with its non-bond interactions. The
valence and non-bond energt components of the total crystal energy were determined. The
valence components are bond, angle, torsion and inversion energy terms and the non-bond
components are Coulomb and van der Waals energy terms. The Hydrogen bond energy
terms are calculated as a part of non-bond components [3,5]. The charges were calculated by
charge equilibration method (Qeq) [6].
The results of calculation of total energy (Es), bond energy (Eb), angle energy (Ea),
torsion energy (Et), inversion energy (Ei), van der Waals forces (Ew), and Coulomb energy
(Ec) along with numbers of molecules in crystal unit (Z) and values of detonation energy [1]
are presented in Table 1 for the selected set of explosives.
Table 2 presents the values of impact sensitivity (h), melting points (m.p.), densities (d),
values of ratio of total energy, intermolecular energy and van der Waals energy for the
individual molecule (Es/Z, (Ew+Ec)/Z, Ew/Z) of the same set of substances along with the
values taken from Ref. [1] for non-covalent hydrogen bonds (Hb/Z). In both tables, the
investigated substances are arranged in the order of decreasing values of detonation energy
(Edet).
3.
DISCUSSION
When evaluating the effect of energy content of crystal on selected parameters of

explosives, attention was predominantly directed to the impact sensitivity expressed by the
height of hammer fall (in cm). From Table 2, and better still from Fig 1 A, B and C, it can be
seen that a low sensitivity (a higher value of h) is exhibited by compounds having noncovalent hydrogen inter-bonds (these compounds are given in underlined italics).
Exceptional cases are TNT, TNB and HNS; however, these substances possess the lowest
content of releasable energy out of the set of the explosives investigated. All other
compounds areirrespective of the magnitude of total or inter-bond energyconsiderably
sensitive, and yet do not contain hydrogen bonds.
The presence of such Hb bonds between the molecules in a crystal distinctly increases
the melting points (see Fig 2 A, B and C); however, compounds without such bonds of
course exhibit different manifestations. It can be seen that with decreasing total energy of
crystal (Es) the melting point increases within a broad interval (Fig 2 A), and generally the
same is true for the sum of van der Waals and Coulomb energies (Fig 2 B). Obviously, the
values of energies of the van der Waals forces are in this case most affected by the presence
of Hb bonds. If substances do not possess such bonds, the melting point can be affected also
by other properties (composition, configuration, symmetry, cycles etc.).
In contrast to the above-mentioned trends, the effect of various components of crystal
energy on density of the tested set of compounds is ambiguous (Fig 3 A, B and C). For
instance, from among the six compounds possessing the density 1.9 g.cm3 or higher, two
substances (TATB and HMX) reach the maximum and the minimum of total energy Es, but
the highest density is encountered with HNIW, which has a medium Es value. Of course, this
compound lacks Hb bonds, and its considerably big molecule possesses a cage structure.
Analogous situation is encountered with the sum of energies of inter-forces Ew+Ec
(Fig 3 B) and also for Ew (Fig 3 C). Another example is nitroglycerin NG with almost zero
values of inter-forces energy. This substance is liquid at ambient temperature (melting
temperature is 13 C) with the density of 1.59 g.cm3, and the value of crystal density is only
475
valid at lower temperatures. It would be possible to proceed in this way also in the cases of
other substances, but attention is attracted predominantly to those of them that possess
hydrogen bonds. Besides TATB, whose Ew energy is high and density corresponds with it,
the highest contribution of Hb energy was observed in the case of DADNE, where the value
of density is considerably high with respect to the relatively small molecule. A similar effect,
even though not so distinct, is also present in other substances with hydrogen bonds (see
Fig 3 B, C).
4.
CONCLUSIONS
It has turned out that physical and explosion properties of the tested set of explosives are
affected by a number of parameters (composition, configuration etc.), but the values of
sensitivity, melting temperature and density can be markedly affected by the presence of
hydrogen bonds in crystal together with the magnitude of total, van der Waals, and Coulomb
energies. The calculations allow not only interpretation of various magnitudes of the
parameters monitored but can also serve for a certain extent of their prediction in the case of
proposed new structures. Further work in this direction will presumably involve the
calculation of total crystal and sublimation energies for other energetic materials with the
aim of both extension of the existing statistical set and more precise determination of the
values so far known. Such a body of data will probably make it possible to predict properties
of new structures of explosives.
Acknowledgements:
The authors acknowledge the Czech MPO (Project FC M2/05) and GA R
(Project 203/02/0436 CSD) for financial support.
REFERENCES:
[1] VVRA P., POSPIL M., REPKOV J.: Effect of intermolecular forces on some properties of
explosives, Proc. 5th Seminar New trends in Research of Energetic materials, Univ. Pardubice,
April 24 25, pp. 357 368, ISBN 80-7194-435-1.
[2] VVRA P. Electronic density of molecule and some properties of high explosives, Proc. 4th
Seminar Proc New Trends in Research of Energetic Materials, Univ. Parduubice, April 11 12,
2001, pp. 345-351, ISBN 80-7194-333-9.
[3] Cerius2 documentation, June 2000, San Diego: Molecular Simulations Inc, 2000.
[4] RAPP A.K., CASEWIT C.J., COLWELL K.S., GODDARD W.A.III, SKIFF W.M.: UFF, a Full
Periodic Table Force Field for Molecular Mechanics and Molecular Dynamics Simulations,
J. Amer. Chem. Soc., 114, 10024, 1992.
[5] COMBA P., HAMBLEY T.W.: Molecular Modeling of Inorganic Compounds, Weinheim,
New York, Basel, Cambridge, Tokyo, VCH, 1995.
[6] RAPP A.K., GODDARD W.A.III: Charge Equilibration for Molecular Dynamics Simulations,
J. Phys. Chem., 95, 3358, 1991.
476
5.
TABLES
Table 1. Calculated energy of crystals of selected high explosives
No.
Name
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
HNIW
HNB
HMX
BTNEN
BTNEU
TNAZ
RDX
PETN
DADNE
DINA
BDNPN
NG
ADNBF
2,4-DNI
TATB
NQ
TE
DATB
TNA
PA
HNS
TNB
TNT
Es
[kcal/mol]
600,7
126,5
-120,4
936
375,6
1297,6
633,2
586,2
359,5
219,1
310,4
807
1440,8
1302,7
961,5
-68,6
165,8
429,2
339,3
702,8
113,7
83,4
385,5
Eb
[kcal/mol]
554,6
187,6
108,9
533,3
203,2
693,4
656,2
448,5
297,9
303,5
280,2
656,2
871,7
232,5
240,1
1092,8
427,8
203,3
285,1
587,1
326,1
886,1
576
Ea
[kcal/mol]
287,5
72,7
43,6
197,3
79,2
811,7
392,1
302,3
46,5
145,3
77,1
206,9
356,3
418,3
7,3
166,1
80,3
23,4
45,3
70,5
88,5
178,5
134,8
Et
[kcal/mol]
78,2
151,6
17,3
23
7,5
80,5
46,6
23,7
44,8
8,7
8,3
6,7
16
7,9
1,6
1,3
71,1
1,1
7,4
21
129,5
26,9
117
Ei
[kcal/mol]
22,3
0,7
1,8
0,7
0,1
14,2
19,8
0,4
20,4
1,3
0
0
5
0,8
0,3
0,2
0,8
0,1
0,1
1
0,2
0
0,2
Ew
[kcal/mol]
190,4
127,9
75,7
135,9
41,1
-1,8
254,4
17,9
103
34,8
45
27,6
296,4
451,4
753,8
567,1
64,8
270,2
227,1
368,7
128,1
235,3
36,9
Ec
[kcal/mol]
-523,4
-414
-367,6
45,8
43,5
-300,5
-735,9
-206,6
-153
-274,5
-100,2
-90,4
-104,6
191,7
-41,7
-1896,1
-478,9
-68,8
-225,6
-354,3
-558,7
-1244,2
-479,3
Z
4
4
2
4
2
8
8
2
4
4
2
4
4
8
2
16
4
2
4
8
4
16
8
Edet
[kcal/cm3]
2,313
2,238
2,116
2,114
2,068
2,057
1,987
1,948
1,884
1,755
1,754
1,725
1,690
1,626
1,603
1,589
1,574
1,527
1,464
1,459
1,369
1,345
1,274
477
Table 2. Some properties of selected high explosives
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Name
HNIW
HNB
HMX
BTNEN
BTNEU
TNAZ
RDX
PETN
DADNE
DINA
BDNPN
NG
ADNBF
2,4-DNI
TATB
NQ
TE
DATB
TNA
PA
HNS
TNB
TNT
M.hm.
m.p.
Es/Z
(Ew+Ec)/Z
Ew/Z
Hb/Z
[g/mol]
438
348
296
388
386
192
222
316
148
240
326
227
241
158
258
104
287
243
228
229
450
213
227
[g/cm3]
2,03
2,00
1,90
1,96
1,86
1,84
1,80
1,77
1,88
1,67
1,73
1,84
1,90
1,76
1,94
1,77
1,73
1,84
1,77
1,76
1,74
1,68
1,65
[oC]
212
258
265
95
101
100
204
141
215
51
187
13
267
274
350
246
131
286
188
122
316
123
80
[cm]
23
12
26
5
17
21
24
12
100
23
29
10
160
100
320
177
32
320
170
87
54
100
160
[kcal/mol]
150,175
31,625
-60,200
234,000
187,800
162,200
79,150
293,100
89,875
54,775
155,200
201,750
360,200
162,838
480,750
-4,288
41,450
214,600
84,825
87,850
28,425
5,213
48,188
[kcal/mol]
-83,250
-71,525
-145,950
45,425
42,300
-37,788
-60,188
-94,350
-12,500
-59,925
-27,600
-15,700
47,950
80,388
356,050
-83,063
-103,525
100,700
0,375
1,800
-107,650
-63,056
-55,300
[kcal/mol]
47,600
31,975
37,850
33,975
20,550
-0,225
31,800
8,950
25,750
8,700
22,500
6,900
74,100
56,425
376,900
35,444
16,200
135,100
56,775
46,088
32,025
14,706
4,613
[%]
0,00
0,00
0,00
0,00
0,00
0,00
0,00
0,00
32,60
0,00
0,00
0,00
9,75
9,95
3,31
27,34
0,00
5,54
7,50
1,50
0,00
0,00
0,00
478
Used Abbreviations of the Substances:

HNIW 2,4,6,8,10,12 hexanitro 2,4,6,8,10,12 hexaazatetracyclo [5.5.0.05,9.03,11] dodecane
HNB hexanitrobenzene
HMX 1,3,5,7 tetranitro 1,3,5,7 tetraazacyclooctane
BTNEN di (2,2,2 trinitroethyl) nitramine
BTNEU di (2,2,2 trinitroethyl) urea
TNAZ 1,3,3 trinitroazetidine
RDX 1,3,5 trinitro 1,3,5 triazacyclohexane
PETN pentaerythritol tetranitrate
DADNE 1,1 diamino 2,2 dinitroethylene
DINA di (2 nitratoethyl) nitramine
BDNPN di (2,2 dinitropropyl) nitramine
NG glycerine trinitrate
ADNBF 7 amino 4,6 dinitrobenzofuroxane
2,4 DNI 2,4 dinitroimidazole
TATB 1,3,5 triamino 2,4,6 trinitrobenzene
NQ nitroguanidine
TE N methyl 2,4,6,N tetranitroaniline
DATB 1,3 diamino 2,4,6 trinitrobenzene
TNA 1 amino 2,4,6 trinitrobenzene
PA 1 hydroxy 2,4,6 - trinitrobenzene
HNS 2,2,4,4,6,6- hexanitrostilbene
TNB 1,3,5 trinitrobenzene
TNT 2,4,6 trinitrotoluene
479
FIGURES
A)
B)
C)
Fig 1.
Relationship impact sensitivity and
A) total energy
B) intermolecular energy
C) van der Waals energy
480
A)
B)
C)
Fig 2.
Relationship melting point and
A) total energy
481
A)
B)
C)
Fig 3.
Relationship density and A) total energy

482
EVALUATION OF THE EFFICIENCY OF IGNITION PROCESSES

BY COMBUSTION UNDER DIFFERENT PRESSURES
Fred Volk, Helmut Bathelt, Richard Jakob and Dietmar Mueller
Fraunhofer Institut Chemische Technologie (ICT),
Joseph-von Fraunhofer-Str. 7, 76327 Pfinztal-Berghausen, Germany
Abstract:
For the optimization of ignition materials a better insight into the burning behavior is
necessary. Especially incomplete and none-equilibrium reactions accompanying the
combustion process must be found out. This could be achieved by the analysis of
reaction products as well of propellants as of the igniter materials as a function of
pressure or loading density. It was found that it is important for an excellent ignition
process that the combustion of the igniter material is as far as possible independent of
the pressure. In this case the full exothermic reaction heat of the igniter can be
transferred to the propellant bed. Very good results were recorded with high energetic
ignition materials such as boron / potassium nitrate, B/KNO3, and zirconium/barium
nitrate, Zr/Ba(NO3)2 , containing ignition charges or benite strands.
Keyword:
1.
Igniter, Propellant, Combustion, Reaction products
INTRODUCTION
During the past few years, great efforts have been undertaken to increase the
performance of solid propellants and gas generators for airbags and belt restraint systems [1].
Despite the many improvements achieved during the past decades in the field of
propellant chemistry, it is still black powder that was used in many cases as ignition
material. On the other hand, lova propellants based on nitramines are difficult to ignite and
require different ignition stimuli than conventional nitrocellulose based propellants to yield
optimal ballistic performance [2] . In order to improve the ignition efficiency, the purpose of
this paper is to investigate higher energetic igniter materials than black powder, such as
boron / potassium nitrate, zirconium / barium nitrate (Zr / Ba-nitrate), zirconium hydrogen /
barium nitrate (ZrH2 / Ba-nitrate) and other materials.
2.
EXPERIMENTS
In order to study the possible arising of none-equilibrium reactions and the pressure or
loading density dependence of reactions during the ignition process, the gaseous reaction
products of different propellant and igniter materials were analysed by gas chromatography
(GC) and mass spectrometry (MS). These examinations were done in a calorimetric vessel at
different loading densities in order to simulate the influence of the ignition pressure. The
burning volume of the vessel was 25 cm3. The combustion residue of B/KNO3 was
determined by ESCA method and by induced neutron activation analysis. For the
thermodynamic calculations the ICT-Thermodynamic Code was used [3] .
483
RESULTS
3.
3.1
None equilibrium and equilibrium burning of propellants
Optimum ignition performance has been reached when so much heat is transferred to the
surface of the propellant bed in the shortest possible time so that the propellant charge will
burn by itself in the sense of an equilibrium burning. This means that during the short
duration of the reaction, the ignition material has to transfer to the propellant not only an
amount of heat that far exceeds the activation energy of the latter, but that on an equal scale
a simultaneous pressure rise in the propellant bed is caused, which is sufficiently great to
overcome the so-called fizz-burning zone and also the dark burning zone, which is
responsible for the none equilibrium burning with the resulting nitric oxides and the adverse
effect on the performance [4, 5, 6] .
3.1.1. Solid rocket double base propellant D 714
Table 1 compares the result of a none equilibrium burning of a double base propellant
achieved at varying pressure rise with an equilibrium burning. It refers to the GC and MS
analysis of the reaction products of the double base rocket propellant D 714. This propellant
with an O2 balance of 43.2% and a heat of explosion of 3657 J/g was burned under
different loading densities in the closed vessel in gram/liter (g/l) such as 10, 20, 50, 100, 200
and 400 g/l.
We see that the first two investigations with 1.0 g/l and 10 g/l produced typical none
equilibrium gases containing 20.3 and 13.2 percent of volume of nitric oxide (NO). At first
with 50 g/l and 100 g/l typical equilibrium burning gases were analysed. It is of interest to
recognise that the formation of methane (CH4) is increasing very clearly from the loading
density 100 g/l to 400 g/l.
As a reason for this, the following reaction is responsible, especially in the cooling
period of the gas mixture in the closed vessel:
CO + 3 H2
CH4 + H2O; H = -206.4 kJ
This reaction is very strong exothermal and proceeds under a volume decrease. In
addition, Table 1 contains the result obtained by calculation of the freezing temperature of
the water gas reaction; this was done via determination of the equilibrium constant Kp
according to:
By using the water gas reaction:
CO + H2O CO2 + H2;
Kp =
H = -41.03 kJ
PCO PH 2O
PCO 2 PH 2
As freeze out temperatures, values of 1590 K, 1450 K and 1600 K have been found.
We need such kind of freezing temperatures for the correct calculation of the heat of
explosion of propellants using the ICT-Thermodynamic Code [3] . Many analyses of the
reaction products of different propellants based on nitrate esters have shown that a mean
value of the freezing temperature of 1500 K was evaluated. This temperature corresponds to
a Kp value of 2.6.
484
In this connection it should be mentioned that most of the combustion reactions of

CHON containing explosives are determined by the water gas reaction. On the other side,
detonation reactions are determined by the Boudouard-Reaction
2 CO CO2 + C;
H = -127.7 kJ
because carbon monoxide decomposes under the very high detonation pressure into CO2
and carbon with a strong exothermal reaction [7] .
3.1.2. Single base propellant GP A5020
The gun propellant A5020, which exhibits a heat of explosion of about 3360 J/g and an
oxygen balance of 42.2 g O2 /100g propellant (-42.2%) was tested under a very low loading
density of 0.33 g/l. The propellant was burned in a glass vessel of 1.5 l containing argon (Ar)
as an inert gas.
In addition the same propellant was ignited in a calorimetric bomb of 25 cm3 free
volume. By using 2.5 g of the propellant, the loading density was 0.1 g/cm3 or 100 g/l. The
result of both tests can be seen in Table 2. For the two loading densities very different
reaction gases were found by mass spectrometric analysis: very large amounts of nitric oxide
(NO) at the lower loading density and nearly no NO for the higher loading density,
respectively, the higher combustion pressure. As we see, the combustion products at 100 g/l
loading density agree well with the gases calculated thermodynamically using the ICTThermodynamic Code [3] .
3.1.3. Double base propellant GP JA-2
The gun propellant JA-2 which contains in addition to ca. 59.5% nitrocellulose the two
liquid energetic systems nitroglycerine (ca. 15%) and diglycoldinitrate (ca. 25%) leads to a
heat of explosion of 4550 J/g. The O2 balance is 30.2%. It was tested under the same
conditions as the propellant A5020: At first it was burned in an argon atmosphere at about 1
bar; secondly, heavily confined in a closed bomb of 25 cm3 volume. The results are depicted
in Table 3. Compared with the lower energy containing propellant A5020, the nitric oxide
(NO) content of JA-2 is much higher: 18.3 mol% compared with 11.1 mol%. Also in this
case, the reaction products of the closed bomb agree well with those of the thermodynamic
calculation. On the other side, the heat of explosion from the closed vessel is 4550 J/g
compared with 2488 J/g of the low pressure combustion.
3.1.4. Triple base propellant with 50% Nitroguanidine: M30 or KN-6540
The triple base gun propellant M30 contains 28% nitrocellulose, 22% nitroglycerine and
ca. 50% nitroguanidine (Nigu). The heat of explosion measured in the calorimetric bomb is
4010 J/g, and the O2-balance -27.4% .
It is of interest that the NO concentration of the low pressure combustion is much
smaller (3.4%) than in the case of the single base A5020 (11.1%) and the double base
propellant JA-2 (18.9%). The reason for this may be the formation of ammonia (NH3), which
is reacting with nitric oxides by a reduction reaction, see Table 4.
Therefore, the low pressure products of M30 contain only 3.4 mol% of NO compared
with the much higher concentration of the single and double base propellants. The reason for
this are the many NH2 radicals formed during the combustion which reduce most of the
NO molecules [8] .
485
Also in this case, the comparison of the closed bomb products with the calculated ones is
very good, whereas the low pressure combustion products are quite different including the
heats of explosion.
3.1.5. Nitramine propellant KHP 232 with RDX and TAGN
In the same way as before we analysed also the reaction products of a nitramine
propellant consisting of
12% polybutadiene binder (HTPB)
78% RDX
10% TAGN
with an O2 -balance of 55.0% and a heat of explosion of 3684 J/g.
For the low pressure combustion under a loading density of 0.67 g/l we found high
concentrations as well for NO (12.3%), as for hydrogen cyanide (HCN) and nitrous oxide
(N2O), see Table 5. In addition, a quite high amount of solid carbon (24.6%) was found,
typical for incomplete combustion processes. The formation of HCN and NH3 is also due to
the none equilibrium combustion of RDX and TAGN [8] .
It is of interest to see that the agreement between measured and calculated products for
the high loading density of 0.1 g/cm3 is not so good as it was for all the investigations before
using the nitrate ester components. The explanation for this behaviour may be the formation
of 1.5 mol% of methane. For the calculations with the exclusion of the formation of CH4, we
found about 7.8% of solid carbon and for the calculation under which the CH4 formation was
allowed, too much methane (6.4 mol%).
The investigation of all the single base, double base, triple base and nitramine
propellants has shown that the combustion is very strong dependent on the pressure. This
means that only a good ignition process can overcome the none equilibrium reactions
leading to an energy output with an excellent performance.
3.2
Combustion of ignition materials
3.2.1. Black powder

The classic black powder, the so-called 3-component (3K) black powder consists of 75%
potassium nitrate, 15% charcoal (alder with 79% C), 10% sulfur and burns basically
according to the following empirical formula [9]
16 KNO3 + 21 C + 5 S 5 K2CO3 + K2SO4 + 2 K2S2 + 13 CO2 + 3 CO + 8 N2
About 43 percent in weight of gases and 57 percent of condensed reaction products are
developed during this burning process.
As an alternative for the 3K-black powder, there is the so-called sulfur-free 2K-black
powder, which is normally composed of 80% potassium nitrate and 20% charcoal (alder
with 79% C). When using alder charcoal with a carbon content of 79%, the burning of this
2K- back powder will yield about 45 wt% of gas.
For a qualification of ignition material, it is important to know whether the burning is
strongly dependent on pressure respectively loading density, for this will involve a change in
the heat of reaction.
486
Another important factor is the reproducibility of results, because for black powder and
other ignition mixtures, the reproducibility is sometimes noticeably affected. This may be
attributed to the fact that simultaneously with the gas reactions additionally solid reactions
and gas/solid reactions occur. For example, 3K- black powder develops more than 50 wt.%
of solid reaction products, among them K2CO3, K2SO4, K2S, C and other products.
Depending on the development of the burning and on test conditions, the below stated
reactions will take place in a more or less complete form:
K2CO3 + H2S K2S + CO2 + H2O; H = +62.8 kJ
2 K2CO3 + 2 H2S K2SO4 + K2S + CH4 + CO2; H = +7.1 kJ
An incomplete equilibrium reaction will have no strong effect on the heat balance
because of the low reaction enthalpy, however it gives rise to a change of the reaction
products.
The composition of the reaction products of black powder containing sulfur and without
sulfur may be seen from Table 6. Both the reaction gases and the most important solid
reaction products were analysed after burning in a closed vessel at three different loading
densities: 0.1, 0.2, and 0.3 g/cm3. As the most important result, it should be stressed that
under the test conditions selected the burning of either black powder type is almost
independent of the loading density, and therefore also independent of the pressure.
In contrast with the afore-described result, however, it is found that in the lower range of
loading densities, i.e. between 0.025 and 0.15 g/cm3 , the composition of the gas is clearly
dependent of the loading density: The content of CO is noticeably increasing as it is shown
in Table 7 by the example of the 3K-black powder Y593.
Table 7 contains moreover the result obtained by calculation of the freezing temperature
of the water gas reaction; this was done via determination of the equilibrium constant Kp
according to
Kp =
PCO PH 2O
PCO 2 PH 2
Besides, the total gas volume of combustion products has been recorded. The gas
volumes without water vapour are between 485 and 509 cm3/g at 20C. The freeze out
temperature of the water gas reaction is between 700 K and 955 K.
3.2.2. Boron/potassium nitrate AZM 953 and other ignition materials
Another ignition material which has been employed very often is boron/potassium
nitrate. Compared to black powder, this ignition mixture offers the advantage of having a
substantially greater heat of explosion and consists of precisely known basic materials,
whereas the charcoal of black powder does not have a precise composition. Besides boron
and KNO3, about 5% of a plastic bonding agent is normally applied in the mixture of
AZM 953.
The reaction products obtained by burning of AZM 953 contain a substantially smaller
amount of gaseous products than black powder. Only about 10 wt.% of the reaction products
are to be found in the gas produced; however, this applies to the cooled-down state of
reaction products only. At the adiabatic flame temperature, however, which is about 2900 K,
487
only 10 to 15% of condensed reaction products are present, since the boron oxides B2O3 and
B2O2 as well as KBO2 are obtained in the gaseous state on this precondition.
Table 8 indicates both the composition and the volume of the reaction gases of four
different ignition materials. Besides, it contains the heat of explosion measured in the
calorimetric vessel and the freezing temperature of the water gas reaction. These four
ignition materials are substances that have been the object of investigation within the
framework of a study on ignition in Germany:
Y 593:
NKP 536:
MV 7308:
AZM 953:
3K- black powder

single base propellant containing 2K- black powder
porous nitrocellulose powder
boron / potassium nitrate
As can be seen from Table 8, boron/ potassium nitrate (AZM 953) is among these four
ignition mixtures the material that distinguishes itself by the greatest heat of explosion and
the smallest amount of gas produced.
Table 9 shows the dependence of gas composition and gas formation of the boron /
potassium nitrate with 5% of plastic binder on the loading density. These results are
compared with the theoretically calculated composition. When considering them, attention
should be given to the fact that, for reasons of better comparability of theory and experiment,
all gas compositions are stated in percent by volume whereas the composition of condensed
reaction products is recorded in percent by weight.
It can be seen that the theoretical computation suggesting a production of boron nitride
(BN) in proportion to the equilibrium yields much to great a heat of explosion when
compared with practical experience, i.e. 7901 J/g instead of the measured amount of
6611 J/g. Since boron nitride is produced out of the elements through an exothermic
reaction, the thermodynamic calculation was repeated on the condition that BN was not
admitted within the reaction products. However, the heat of explosion computed in this case
is 6165 J/g and thus below the experimental amount. Assuming a partial formation of BN,
which is still far away from the state of equilibrium, this difference between theory and
experiment can be accounted for.
In this context, it should also be stressed that this ignition mixture contains in its
combustion residue approximately 4.5 % of unreacted boron even after a formation of boron
nitride corresponding to the equilibrium. It is due to this excess in boron that no free nitrogen
appears in the thermodynamic calculation. The fact that the gas chromatographic and mass
spectrometric analyses of the combustion gases yielded a content in N2 of approximately
30% by volume for all the loading densities examined, clearly indicates that BN can be
produced only partially during the burning process. However, the presence of BN could be
proved for sure within the residues from combustion: Both the induced neutron activation
analysis and the analysis by ESCA method[10] are clearly indicative of the presence of BN
beside elementary boron, KBO2 and B2O3.
When studying the production of gas also shown in table 9, it is found that the formation
of gas first increases with rising loading density, but then decreases noticeably above a
loading density of 0.1 g/cm3. Moreover, a decrease in the N2 content becomes apparent in the
same direction. This permits the conclusion that with increasing loading density and the
consequent rise of the burning pressure, an ever growing amount of N2 is converted into
boron nitride (BN).
488
3.2.3. Modified benite strands based on B/KNO3 and Zr/Ba(NO3)2

In order to improve the energy output of the conventional benite strands manufactured of
nitrocellulose and the components of black powder, as an alternative, benite strands have
been synthesised containing higher energetic basic mixtures of boron/potassium nitrate,
(B/KNO3), and zirconium/bariumnitrate, Zr/Ba(NO3)2, compared with black powder.
For the manufacture, 30% of nitrocellulose was used. The modified benite strands were
synthetisised by extruding the propellant dough into sticks of 254 mm length and 2.0 mm in
diameter. For the interior ballistic investigation of the ignition efficiency, the sticks were cut
into pieces of 40 mm length. The compositions were as follows:
1) Basic mixtures:
AZM-1
26.3% B
73.7% KNO3
AZM-2
34.5% Zr
43.2% Ba(NO3)2
13.7% Nitroguanidine
8.6% metal oxides
2) Modified benite strands:
a)
70% AZM-1
29% nitrocellulose (12.2% N)
1% acardite II (stabilizer)
b)
70% AZM-2
29% nitrocellulose (12.2% N)
1% acardite II
Contrary to the modified benite strands, the original strands containing the components
of black powder exhibit the following composition:
c)
29.2% nitrocellulose (12.2% N)

68.8% black powder Y593
1.0% dibutylphthalate
1.0% acardite II (stabilizer)
3.2.4. Measuring the ignition efficiency by determination

of the ignition delay time
Internal ballistics investigations have been achieved with the aim of measuring the
ignition delay time by registration of the pressure / time plot during the ignition of a surface
coated gun propellant in a closed vessel with a volume of 40 cm3. The ignition delay time is
the time which is needed until 10% of the maximum pressure is obtained. For these tests,
0.3g of the igniter charge has to ignite 4g of the propellant A5020, which is surface coated
by dibutylphthalate and ethylcentralite. The pressure rise was measured using a piezoelectric
pressure gauge.
In Table 10 the thermochemical data as well as the ignition delay times (t2 - times) can
be compared. We see that the benite strands containing B/KNO3 show the best ignition
efficiency with 24 milliseconds (ms), compared with 36 ms for Zr/Ba(NO3)2 and the black
powder containing strands with 46 ms.
489
It is also of interest that the B/KNO3 benite strands produce the highest heat of explosion
of 7009 J/g, which was measured in a calorimetric vessel. On the other side, the basic
ignition mixture of Zr/Ba(NO3)2 shows the highest adiabatic flame temperature of 4335 K.
4.
DISCUSSION
It is shown that the combustion of nitrate ester and nitramine containing propellants
depends very strongly on the pressure. Equilibrium burning with the highest energy output is
found only above a distinct minimum pressure. Therefore it is important to optimize the
ignition process in such a way that the heat transfer of the igniter to the propellant starts
immediately a complete combustion reaction. This is possible by using an igniter system
with a high energy formation not only by hot gases, but also by hot solid and liquid reaction
products [11] .
In this connection, it is important that igniters and ignition charges burn without strong
pressure dependence. For black powder, this burning characteristic is fulfilled, but with the
disadvantage that the heat output is relatively low. Therefore, a much better initiation of
ignition processes have been recorded with the use of boron/potassium nitrate or
zirconium/bariumnitrate containing igniter charges, because of the very much higher heat of
explosion. As an example, see the ignition delay times of the modified benite strands in
Table 10.
5.
CONCLUSION
The use of boron/potassium nitrate, B/KNO3, and zirconium/barium nitrate,

Zr/Ba(NO3)2, exhibiting a high energy output, are promising for the initiation of a successful
ignition of propellants which are difficult to ignite. These components are also successful
when incorporated in modified benite strands or similar ignition charges.
490
REFERENCES
[1]
F. VOLK, H. BATHELT, M. HUND, D. MUELLER: Innenballistische Bewertung der

Wirkung von Anzuendmitteln, Proceedings 16.Internationale ICT-Jahrestagung 1985,
Fraunhofer-Institut fuer Treib- und Explosivstoffe, July 2-5, 1985, Karlsruhe, p. 6-1 to 6-19
[2] S.T. PETERS: Ex-98, an igniter material designed for Lova gun propellants, Proceedings 33rd
Intern. Annual Conference of ICT, June 25-28, 2002, Fraunhofer Institut Chemische
Technologie, p. 27-1 to 27-8.
[3] H. BATHELT, F. VOLK, M. WEINDEL: Der ICT-Code 2000, Fraunhofer Institut Chemische
Technologie, 76327 Pfinztal, e-mail: dvolk@t-online.de
[4] O.K. RICE, R. GINELL: The Theory of the Burning of Double Base Rocket Powders, J. Phys.
Coll. Chem. 54 (1950), p. 885 917
[5] B.L. CRAWFORD, C. HUGGETT, J.J. MCBRADY: The Mechanism of the Burning of
Double-Base Propellants, J. Phys. Coll. Chem. 54 (1950), p. 854 862
[6] N. KUBOTA: Smokeless igniters for solid rockets, Proceedings 16. Int. Jahrestagung
Fraunhofer Institut, July 2-5, 1985, Karlsruhe, p. 21-1 to 21-13
[7] F. VOLK, F. SCHEDLBAUER: Analysis of post detonation products of different explosive
charges, Prop. Expl. Pyr. 24 (1999), p. 182-188
[8] Y. OYUMI, T.B. BRILL: Thermal decomposition of energetic materials. 6. Solid phase
transitions and the decomposition of 1,2,3-Triaminoguanidinium Nitrate, J. Phys. Chem. 89
(1985), p. 4325-4329
[9] R. ESCALES: Schwarzpulver und Sprengsalpeter, Verlag von Veit und Comp., Leipzig 1914
[10] H. MEIER: Staatl. Forschungsinstitut fuer Geochemie, Bamberg, Private communication,
March 31th, 1980
[11] M. HUND, F. VOLK, M.L. LENJI: Influence of the reaction products of ignition powders on
the ignition behavior of a surface coated gun propellant, 4th Int. Symp. on Ballistics, October
17-19, Monterey, USA
491
6.
TABLES
Table 1: Reaction gases, equilibrium constant and freeze out temperature
of the double base rocket propellant D 714 as a function of the loading density
Loading density
(g/l)
Gas analysis (Vol %)
H2
N2
NO
N2O
CO
CO2
CH4
C2H4
H2O
10
50
100
200
400
3.6
2.9
20.3
0.3
28.2
8.9
1.5
1.7
32.6
6.1
5.5
13.2
0.1
33.4
13.4
1.0
1.3
26.0
18.3
10.9
0.7
42.5
10.9
0.1
0.1
16.5
18.8
11.1
0.3
41.4
12.2
0.2
16.3
17.9
11.4
0.1
38.9
14.5
1.5
15.8
9.6
0.08
13.5
29.7
20.6
7.0
19.6
Kp
29.6
10.2
3.57
2.92
2.37
2.97
1750
1590
1450
1600
Temperature ( K)
Table 2: Reaction Products of Single Base GP A 5020

Experiment
Loading Density [g/l]
Combustion Condition
Products [Mol %]:
H2
CH4
CO
CO2
N2
NO
HCN
NH3
H2O
C solid
QEX [J/g]
Kp (T)
0.33
1 bar
100
Closed Vessel
1.8
0.6
17.0
7.0
4.4
11.1
0.2
30.3
27.6
2696
40.88
18.4
0.1
42.7
12.6
10.1
0.85
15.2
3360
2.799
Calculation
without CH4
including
CH4
100
100
19.2
42.4
12.5
10.4
0.0018
0.013
0.14
15.4
3472
2.72
15.8
2.1
40.2
14.7
10.9
0.01
0.10
16.2
3679
2.80
492
Table 3: Reaction Products of Double Base GP JA-2

Experiment
Products [Mol %]:
H2
CH4
CO
CO2
N2
NO
HCN
NH3
H2O
C solid
QEX [J/g]
Kp (T)
0.66
1 bar
100
Closed Vessel
Calculation
100
-
3.5
1.0
24.9
7.0
1.3
18.9
1.1
0.8
28.0
13.5
2488
28.45
14.0
32.7
17.0
12.4
0.025
23.4
0.5
4550
3.215
13.6
0.7
31.2
19.0
12.6
0.004
0.08
22.9
4719
2.765
Table 4: Reaction Products of Triple Base Propellant GP M30 (KN-6540)

Experiment
Products [Mol %]:
H2
CH4
CO
CO2
N2
N2O
NO
HCN
NH3
H2O
C solid
QEX [J/g]
Kp (T)
0.66
1 bar
100
Closed Vessel
Calculation
100
-
5.9
0.1
16.9
8.3
23.7
1.4
3.4
0.4
3.4
27.2
8.2
2818
8.34
15.8
23.6
10.7
28.0
0.018
0.3
21.5
0.1
4010
3.00
14.9
0.8
22.5
11.6
28.5
0.2
21.5
4142
2.80
493
Table 5: Reaction products of the nitramine propellant KHP 232

Experiment
Loading density [g/l]
Products [Mol %]:
H2
CH4
CO
CO2
N2
N2O
NO
HCN
C2H2
C2H4
NH3
H2O
C solid
QEX [J/g]
Kp (T)
Calculation
Without CH4
including CH4
0.67
1 bar
100
Closed Vessel
100
-
100
-
1.48
0.58
4.68
2.13
7.66
6.11
12.31
9.07
0.13
0.33
7.24
23.70
24.56
2297
35.1
27.99
1.53
33.44
2.80
25.71
0.11
7.76
0.68
3684
3.31
25.50
25.43
4.18
24.84
0.03
0.01
0.41
11.80
7.79
4254
2.82
19.44
6.42
30.15
5.56
28.08
0.03
0.01
0.26
10.06
4277
2.81
Table 6: Reaction products of black powder with (3K-) and without (2K-) sulfur
Igniter material
Loading density g/cm3
Gas analysis (Vol.%):
H2
N2
CO
CO2
H2S
SO2
COS
CS2
CH4
Residue analysis (wt.%)
Carbon
K2CO3
K2SO4
K2S + others
Other products
3K-black powder
Charcoal: 79% C (Alder)
0.1
0.2
0.3
2K-black powder
Charcoal: 79% C (Alder)
0.1
0.2
0.3
3.5
27.7
10.6
52.5
5.5
0.1
0.2
-
2.9
30.7
10.2
50.4
5.9
0.1
0.2
-
4.1
28.9
9.2
53.3
5.6
0.1
0.2
-
12.0
29.2
35.5
23.2
0.1
13.4
28.4
33.6
24.3
0.3
14.4
28.0
31.4
25.4
0.8
3.6
56.6
16.5
23.3
-
3.6
56.1
17.4
22.9
-
3.5
54.5
17.1
24.9
-
4.8
78.2
17.0
5.3
77.9
16.8
5.4
75.7
18.9
494
Table 7: Combustion gases of 3K-black powder Y593

as a function of the loading density
Loading density g/cm
0.025
0.05
0.1
H2
0.7
0.8
1.2
N2
32.9
31.6
31.5
CO
4.6
3.8
6.4
CO2
52.4
53.5
53.3
H2S
8.2
9.2
6.9
SO2
0.1
0.1
0.06
COS
1.1
0.9
0.5
CS2
0.04
0.06
0.05
Kp Water gas reaction
0.165
0.107
0.245
Freeze out temperature (K)
745
700
805
3
Total gas volume cm /g
485
496
506
3
0.15
0.2
2.0
31.5
8.7
51.5
4.7
0.06
0.5
0.05
0.405
885
509
2.1
33.6
9.2
50.2
4.2
0.07
0.5
0.05
0.577
955
488
Table 8: Reaction gas of different ignition materials

with a loading density of 0.1 g/cm3
Ignition material
Y 593
NKP-S 536
MV-7308
AZM 953

H2
1.2
18.7
19.4
34.3
N2
31.5
17.2
12.2
35.6
CO
6.4
45.3
45.2
27.0
CO2
53.3
18.7
23.2
0.2
H2S
6.9
SO2
0.06
COS
0.5
0.1
2.5
Total gas volume cm /g
506
702
1002
180
Kp Water gas reaction
0.245
0.682
1.379
Freeze out temperature (K)
805
995
1205
Heat of explosion (kJ/kg)
3052
3299
4287
6611
CH4
3
495
Table 9: Reaction products of B/KNO3 with 5% binder (AZM 953)

with different loading densities: experiment and calculation
Experiment
Theory
3
Loading density g/cm
0.05
0.1
0.15
0.2
0.1
0.1
H2
36.9
36.6
36.5
40.6
19.0
11.3
N2
35.3
32.9
31.9
30.9
52.7
CO
24.6
27.3
27.5
23.5
57.5
25.2
CO2
0.2
0.3
0.45
0.45
CH4
3.0
2.9
3.4
4.6
23.5
10.8
3
Gas volume cm /g
164
182
155
132
81
153
Condensed reaction
products (wt.%)
KBO2
BN
B2O3
B
K
Carbon
Heat of explosion (KJ/kg)
78
11
?
10
?
1
6611
7342
60.5
19.3
15.1
4.5
0.6
7901
67.2
18.2
13.9
0.7
6165
Table 10: Thermochemical data and ignition delay time of different benite strands
Oxygen
balance
%
Density
g/cm3
Adiabatic
temperature
K
Heat of explosion
J/g
Calc./ measured
Ignition
delay time
ms
-29.8
1.85
3128
8822 / 7009
24 +/- 2
-11.1
2.47
3917
4188 / 3802
36 +/- 3
-15.9
1.70
2500
3333 / 2913
46 +/- 3
AZM-1
Basic mixture
-29.2
2900
8558 / -
14 +/- 2
AZM-2
Basic mixture
-2.5
4335
4141 / -
10 +/- 1
Ignition charge
AZM-1
Modified benite
With B/KNO3
AZM-2
Modified benite
with Zr/Ba(NO3)2
Original benite
with black powder
496
LONGTERM BEHAVIOUR OF THE HARDNESS

MEASURED ON CURE CAST SAMPLES
Richard Wild
Diehl Munitionssysteme GmbH & Co KG, PBX-Center Maasberg

Abstract:
Shore A measurement of PBX cure cast samples is usually used as a method to discover
the polymerisation conditions of the resin.
Other applications for hardness measurements are material selection, material
comparison and quality control.
When the test specimen are stored under higher temperature conditions for a longer
time, hardness can also be one of the data to predict aging.
Keyword:
1.
PBX, Shore A, Hardness, Aging.
INTRODUCTION
1.1
Definition
Shore hardness is the resistance of a plastic material against the penetration of a steel
cone pressed with a constant force (Figure 2).
1.2
Test set up
A calibrated spring in the Shore A hardness tester is pressing the truncated cone of 35
with a test loyad of 12.5 N for 15 seconds against the sample surface (Figure 3).
A pointer on dimensionless scale shows
100 for no penetration, Shore A = 100
0 for full penetration, Shore A = 0
1.3
Test specimen
The PBX sample should have a minimum diameter of 35 mm and a thickness of more
than 6 mm (Figure 4).
The sample surface has to be flat, parallel and smooth.
497
1.4
PBX hardness
Main applicyations for PBX hardness measurement is control of the polymerization time
(Figure 5).
Other applications are raw material and additive selection, material comparison as well
as quality control.
When the test specimen are stored under higher temperature conditions for a longer time,
hardness can also be one of the data to predict aging.
2.
EXPERIMENTS
2.1
PBX samples
All samples have HTPB/DOA binder and are cured with diisocyanate (Figure 6).
The PBX samples with 85% HMX are cast without vacuum.
Variables were the quantity and type of catalyst and the polymerization / storage
temperature.
The samples were taken from production, from test batches and from curing tests.
2.2
Dwell time
The course of the cone penetration as a function of time (Figure 7), shows that after
about 12 seconds Shore A becomes constant. So 15 seconds were chosen as dwelling time
for all measurements of these kind of materials.
3.
RESULTS
3.1
Samples in aluminium foil
To avoid oxidation with air and loss of plasticizer the samples are usually covered with
Al-foil.
Compared to unwrapped samples the aluminium covered PBX specimen with an Fecatalyst showed so small increase in hardness over 200 days, that for our purposes we did the
tests without Al-covering (Figure 8).
3.2
Ambient temperature storage
At ambient temperature the 60C cured samples with an Fe-catalyst dont change in
hardness over a 200 days storage
(Figure 9).
Storage at 60C within a 200 days period shows a continuous rise, but no significant
change in the hardness curve.
3.3
Catalyst concentration
To get at 40C acceptable polymerization times, the concentration of the iron catalyst
has to be about 5 times higher than for 60C.
Further increase of the catalyst concentration has only small influence on the curing
time.
498
Faster curing results in a higher final hardness of the PBX.
3.4
Temperature variation, Fe-catalyst
Lower polymerization temperatures than 40C with the iron catalyst strongly are
increasing the curing times.
Nethertheless curing with this Fe-catalyst is possible, but it takes about two months.
3.5
Temperature variation, Sn-catalyst
The PBX samples with the Sn-catalyst become much softer than with the Fe-catalyst.
The hardness for all curves has a maximum.
4.
SUMMARY
A 200 days storage at 60C shows no significant change in the hardness curves.
The constant increase of hardness is first of all caused by plasticizer migration effects.
Faster curing at higher temperature or higher catalyst concentration causes a harder

product.
For the Fe-catalyst there is an optimum of the catalyst concentration, where more
catalyst has only small influence on the curing time.
At elevated temperatures samples with Sn-catalyst have shorter curing times than with
Fe-catalyst.
The temperature has the strongest influence on the curing time.
499
PBX
University of Pardubice
Sixth International Seminar
New Trends in Research of
Energetic Materials
April 23-25, 2003
Longterm Behaviour of the Hardness

measured on Cure Cast Samples
Richard Wild
Diehl Munitionssysteme GmbH & Co KG
Werk Maasberg
WILD
03/03
observe protection mark i. a. w. DIN 34
Fig.1
A Company of
Diehl VA Systeme
Cover
PBX
Shore A
Penetration of a
pressed with a
for a
at a
WILD
03/03
Fig.2
truncated cone
defined force
specified period of time
constant temperature
A Company of
Diehl VA Systeme
Definition
500
PBX
Hardness
Shore A
Test Set Up
WILD
03/03
Fig.3
A Company of
Diehl VA Systeme
Test set up
PBX
Test Specimen
WILD
03/03
Fig.4
Diameter
> 35mm
Thickness
> 6mm
Surface
flat, parallel and smooth
A Company of
Diehl VA Systeme
Sample requirements
501
PBX
PBX Hardness
Compare material properties
Polymerisation time
Quality control
Aging
WILD
03/03
Fig.5
A Company of
Diehl VA Systeme
Application
PBX
PBX Samples
HTPB / DOA binder cured with diisocyanate
85 weight % HMX as explosive filler
No vacuum cast
Catalyst and temperature varied
WILD
03/03
Fig.6
A Company of
Diehl VA Systeme
Sample properties
502
PBX
Indentor extension as a function of elapsed time

50
45
40
35
Shore A
30
25
20
15
10
5
0
0s
2s
4s
6s
WILD
03/03
8s
10s
12s
14s
16s
A Company of
Diehl VA Systeme
Fig.7
time
Dwell time
PBX
35
30
Shore A
25
20
Fe / 60C-Alu
Fe / 60C
15
10
5
0
5
25
50
75 100 125 150 175 200
WILD
03/03
Fig.8
Hardness, wrapped and unwrapped samples
Storage
Time
in days
A Company of
Diehl VA Systeme
503
PBX
35
30
Shore A
25
20
Fe / 60C-RT
Fe / 60C
15
10
5
0
5
25
50
WILD
03/03
75 100 125 150 175 200
Storage
Time
in days
A Company of
Diehl VA Systeme
Fig.9
Cured samples at ambient temperature and at 60C
PBX
30
25
Shore A
20
Fe / 40C
4Fe / 40C
8Fe / 40C
15
10
5
0
5
WILD
03/03
Fig.10
25
50
75
100
Storage
Time
in days
A Company of
Diehl VA Systeme
Catalyst concentration, samples cured at 40C
504
PBX
30
25
Shore A
20
4Fe / RT
4Fe / 30C
4Fe / 40C
15
10
5
0
5
25
50
WILD
03/03
75
100
A Company of
Diehl VA Systeme
Fig.11
Storage
Time
in days
Temperature variation, samples with Fe-catalyst
PBX
25
Shore A
20
15
Sn_60C-40C
Sn_RT
Sn_40C
10
5
0
5
WILD
03/03
Fig.12
25
50
75
100
125
150
Storage
Time
in days
A Company of
Diehl VA Systeme
Temperature variation, samples with Sn-catalyst
505
PBX
Summary
Only slight differences in hardness covered
uncovered
No significant change in hardness within 200 days at 60C

Faster curing causes a higher final hardness
Fe-catalyst:
- no further increase of the curing time at a

certain maximum concentration (T=40C)
Sn-catalyst:
- faster curing than Fe at higher temperatures

- hardness maximum
WILD
03/03
Fig.13
A Company of
Diehl VA Systeme
Summary
506
STABILITY ANALYSES OF POROUS PROPELLANTS

Stephan Wilker, Gabriele Pantel and Lutz Stottmeister
WIWEB ASt Heimerzheim, Groes Cent, 53913 Swisttal (DE)
Abstract
Porous propellants are used in some special applications. Their main feature is a high
surface to volume ratio. This affects mainly their burning behaviour. But this feature has
a negative impact on the chemical stability.
As the oxidation of nitrocellulose is an elementary part of the decomposition process of
all propellants and starts on the surface the chemical stability of porous propellants is
markedly lower than in comparable solid propellant blocks of the same chemical
composition. This can be easily demonstrated by microcalorimetry. Whereas standard
single base propellants do not show any autocatalysis at 89C within 60 or 80 days,
a porous single base DPA stabilized propellant has a strong autocatalysis after 10-11
days.
The paper also describes in detail the decomposition of the stabilizer DPA into its
daughter products and shows the difference between this reaction in porous and in solid
propellants (there is a completely different distribution of daughter products in
dependence of the available oxygen in the ageing vessel). However, the time to
autocatalysis changes independently of a pre-ageing of the porous propellant,
demonstrating that other effects (probably mainly the composition of the atmosphere in
the ampoule) play the decisive role.
Finally, a concept to determine the low temperature (40-30C) activation energy and
including first results of this study on two single base propellants is presented. The
determination of the activation energy at these temperatures is extremely important to
correctly extrapolate results from high temperature ageing tests to ambient temperature.
Keyword:
1.
single base propellants, porous propellants, chemical stability, stabilizer

depletion, HPLC, microcalorimetry, activation energy
INTRODUCTION
Nitrocellulose based propellants decompose under liberation of nitrogen oxides, mainly

NO and NO2 [1]. These gases further react with the nitrocellulose chain to generate more
nitrogen oxides. Finally the reaction turns into an autocatalytic decomposition.
To prevent this autocatalysis stabilizers are added to a propellant [2]. They are usually
based on aromatic amines like diphenylamine or on aromatic urea compounds like centralite
I or akardite II. These stabilizers react with nitrogen oxides under the formation of nitrosated
and/or nitrated species. Thus an autocatalytic decomposition usually can be avoided at least
within the usual service life time.
A second process involved must not be forgotten. It is the oxidation of the nitrocellulose.
Oxygen usually is present in the samples. Oxygen has three major effects on the sample [3].
At first the nitrocellulose itself is oxidised, mainly the side-chains of the nitrocellulose are
involved. Then the stabilizers (which usually are also anti-oxidants) can be oxidised. A third
507
effect is the reaction of NO (which usually does not react with nitrocellulose or the
stabilizers) back to NO2, which is a strong oxidising agent.
Whereas normal propellants have a bulk density of about 1 g/cm porous propellants
only show a bulk density of 0.4 to 0.6 g/cm. And, due to their manufacturing process, they
show a high porosity which leads to a high surface to volume ratio. This means that oxygen
is present along the large surface of the sample which leads to a more rapid and severe
oxidation of the grain.
2.
SAMPLES
The main subject of this paper is a single base porous propellant V 6090 (produced
according to UK specification P387). For the purpose of comparison results of a compact
single base propellant A 5020 and of a double base ball propellant (K 5810) are presented [4].
All three propellants are stabilized with diphenylamine (DPA). Table 1 shows the main
ingredients and properties of the propellants.
Table 1. Specification of the propellants discussed in this paper
Propellant
V 6090
A 5020
K 5810
Production year
1988
1981
2002
Nitroglycerin content 0 %
0%
9,5 1,5 %
DPA content
1,0 0,2 %
0,8 %
1,2 0,3 %
Other ingredients
KNO3 0,1 %
DBP 3 %
DBP 4,5 1,5 %
C I 1,25 %
KNO3 0,25 %
Moisture
1,5 %
1,15 %
0,55 %
Bulk density
0,46 g/cm
0,95 g/cm
0,98 g/cm
DBP = dibutylphthalate, C I = ethyl centralite
3.
EXPERIMENTAL PART
3.1
Microcalorimetry
Heat flow measurements were conducted with a Thermal Activity Monitor TAM 2277
(Thermometrics AB, Sverige). The measurements were performed in 3 mL glass ampoules.
Usually the ampoules are completely filled and sealed. For the determination of the influence
of different measuring conditions we varied the standard method by
filling the ampoule to only 5% to 80% (details see in chapter 4.2)
replacing the air in the ampoule by argon
3.2
Stabilizer analyses
The stabilizer content was monitored by ageing propellant samples in the microcalorimeter under different conditions. Prior to HPLC analyses around 120 mg of the propellants
were dissolved in 10 mL of acetonitrile, and then 40 mL methanol are added. The
nitrocellulose was precipitated by adding 50 mL of water. The solution was filtered through
508
a filter syringe and directly passed into the HPLC autosampler flasks. The stabilizers were
detected with a Gynkotek HPLC system consisting of a pump M480S, an automatic sampler
Gina 50, a column oven and a diode array detector UVD 320S ( = 200-356 nm). The
detection wavelength was 225 nm. The column (Lichrospher 100RP18 - 5m; 250x4 mm
with a pre-column 20x4 mm) was tempered at 25C. A methanol/water mixture (67/33) was
used as mobile phase and pumped with a flow rate of 1.2 mL/min.
3.3
Synthesis
The stabiliser depletion products N-NO-2-NO2-DPA, N-NO-4-NO2-DPA, N-NO-2,4DNDPA and N-NO-4,4-DNDPA were synthetised according to literature methods from 2NO2-DPA (resp. 4-NO2-DPA, 2,4-DNDPA or 4,4-DNDPA) and NaNO2 [6].
4.
RESULTS
4.1
Microcalorimetry
4.1.1. Microcalorimetry (standard procedure)

Microcalorimetric measurements were performed mainly at 89C and at 70C. Some
measurements were also made at other temperatures (80C, 60C). Figures 1 and 2 show the
results of heat generation rate of unaged samples in completely filled and sealed ampoules.
P, W/g
T = 89C
severe
autocatalysis
900
600
300
V6090
0
0
K5810
Fig 1.
A5020
V6090
A5020 (lot
from 1993)
9
Time, day
Microcalorimetry of unaged samples at 89C. A second lot of A 5020 type

propellant is added to show that this propellant does not show an
autocatalytic decomposition
509
P, W/g
T = 70C
90
60
30
K5810
A5020
V6090
0
Time, day
0
30
Fig 2.
60
90
120
Microcalorimetry of unaged samples at 70C 1
4.1.2. Activation energy determination

The activation energy of the decomposition of the three propellants was calculated by a
model free iso- evaluation (see e.g. [7]). For this purpose the times when a certain
decomposition degree is reached and the corresponding heat generation rates at these points
are listed and the activation parameters according to Arrhenius are calculated. The results are
presented in table 2 and in figs. 3 to 5.
Table 2. Activation energies of propellants V 6090, A 5020 and K 5810 between
89C and 70C. Values taken at the highest decomposition degree reached.
All values in kJ/mole.
Energy release [J/g]
V 6090
A 5020
K 5810
Corresponding figure
23,5
140
32,6
135
55,0
145
At the time of collecting the results for the manuscript (Jan. 30, 2003) the 70C measurement of V 6090 was
not yet finished. The complete plot will be presented at the seminar.
510
155
Ea [kJ/mole]
150
145
140
Ea (t70/t89)
135
Ea (P89/P70)
130
0
10
20
30
40
50
60
Q [J/g]
Fig 3.
Activation energies calculated from iso--points of the measurements at

89C, and 70C of propellant V 6090. Ea (P) refers to the ratio of heat
generation rates at the iso--points, Ea (t) refers to the time when an iso-point is reached.
160
Ea (t70/t89)
Ea [kJ/mole]
150
Ea (P89/P70)
140
130
120
0
10
15
20
25
30
35
Q [J/g]
Fig 4.

89C, and 70C of propellant A 5020. Ea (P) refers to the ratio of heat
generation rates at the iso--points, Ea (t) refers to the time when an iso-point is reached.
511
170
160
Ea [kJ/mole]
150
140
130
120
Ea t70/t89
Ea t80/t89
110
Ea P89/P80
100
0
10
20
30
40
50
60
Q [J/g]
Fig 5.

89C, and 70C of propellant K 5810. Ea (P) refers to the ratio of heat
generation rates at the iso--points, Ea (t) refers to the time when an
iso--point is reached.
4.1.3. Microcalorimetry (measurements under different conditions)

Two different types of experiments were performed to determine the dependence of the
decomposition pathway from the measuring conditions: At first the loading density (LD)
was varied, thus the ratio between oxygen and propellant changes. A second experimental
series was performed under argon at lower loading density. For this purpose the measuring
ampoules were filled with propellant and before closing a slight stream of argon was flushed
through the ampoule to replace the air inside. After that the samples were measured by
microcalorimetry. Figures 6 (overview) and 7 (details) show the curves. More information is
presented in table 3.
Table 3. Evaluation of HFC curves of propellant V 6090 measured under different
conditions (measuring temperature 89C)
measuring conditions
LD 100%, air
LD 50%, air
LD 50%, argon
1,99 J/g
1,66 J/g
1,36 J/g
2,32 J/g
0,53 J/g
5,78 J/g
103,5 W/g
100,9 W/g
78,4 W/g
time of 2nd maximum
1,36 d
1,38 d
1,36 d
time to autocatalysis (time at

1500 W/g taken)
11,10 d
18,23 d
17,83 d
energy released by oxidation

energy released in 2
maximum
nd
Heat generation rate in 1st

minimum
512
P, W/g
T = 89C
4000
3000
LD 50%, argon
LD 50%, air
2000
1000
LD 100%, air
enlarged area
see figure 7
0
0
Fig 6.
10
15 Time, day
Dependence of HFC signal on the measuring conditions

(complete measurements). Propellant V 6090
P, W/g
T = 89C
200
Oxidation of NC
LD 100%, air
150
LD 50%, air
100
50
DPA depleted
LD 50%, argon
0
0
Fig 7.
Time, hour
20
40
60
80
Dependence of HFC signal on the measuring conditions (detail).

Propellant V 6090
513
P, W/g
T = 89C
300
LD 100%, air
LD 60%, air
200
Oxidation of NC
DPA depleted
100
LD 60%, argon
0
0
Fig 8.
Time, day
Dependence of HFC signal on the measuring conditions.

Propellant K 5810
P, W/g
T = 89C
300
200
K5810; LD 1.0
K5810; LD 0.32
100
0
0
Fig 9.
V6090; LD 1.0
2nd maximum
(oxidation)
2
Time, day
Comparison of HFC curves of K 5810 and V 6090; K 5810 with

comparable loading mass to completely filled V 6090 (LD 0.32)
514
Table 4. Evaluation of HFC curves of propellant K 5810 measured

under different conditions (measuring temperature 89C)
measuring conditions
LD 100%, air
LD 60%, air
LD 60%, argon
0,93 J/g
3,01 J/g
0,13 J/g
43,8 W/g
58,2 W/g
39,8 W/g
2,65 d
2,70 d
3,22 d
228,5 W/g
204,2 W/g
194,8 W/g
energy released by oxidation

st
Heat generation rate in 1

minimum
time of 2nd maximum
Heat generation rate in 2nd

maximum
4.2
Stabilizer depletion
Stabilizer analyses were performed after HFC measurements. All HFC experiments
referred to in this chapter have been performed in completely filled and hermetically sealed
ampoules under air. Because up to 13 different stabilizer depletion products could be
identified and quantified we have collected them in different groups (e.g. all mononitroDPAs
together). Which single components stand behind every column is noted in the explanations
(below tables 5, 6 and 7). Figures 10 and 11 show the stabilizer depletion.
Table 5. Stabiliser depletion of V 6090 after ageing at 89C. All values in weight-%
Ageing Ageing
t [d] Q [J/g]
DPA
N-NO- N-DPA N-NODPA

NDPAs
DNDPA
a)
b)
c)
d)
e)
f)
g)
N-NO- TNDN- DPAs

DPAs
TeN- ConverDPAs
sion
0,000
0,0
0,643
0,216
0,133
0,000
0,000
0,000
0,000
0,000
0,932
0,167
2,4
0,306
0,395
0,271
0,000
0,000
0,000
0,000
0,000
0,856
0,326
4,4
0,206
0,484
0,286
0,000
0,000
0,000
0,000
0,000
0,846
0,535
6,6
0,069
0,591
0,322
0,005
0,000
0,000
0,000
0,000
0,831
0,721
8,6
0,025
0,604
0,312
0,013
0,000
0,000
0,000
0,000
0,797
1,330
15,2
0,000
0,455
0,231
0,198
0,052
0,000
0,000
0,000
0,742
1,888
20,4
0,000
0,305
0,155
0,329
0,241
0,000
0,000
0,000
0,769
3,200
29,6
0,000
0,173
0,017
0,324
0,366
0,079
0,010
0,000
0,678
5,130
42,8
0,000
0,073
0,000
0,343
0,273
0,297
0,047
0,000
0,680
7,490
58,1
0,000
0,000
0,000
0,182
0,247
0,378
0,184
0,000
0,612
10,000
82,0
0,000
0,000
0,000
0,000
0,173
0,103
0,523
0,046
0,486
11,111
115,2
0,000
0,000
0,000
0,000
0,000
0,000
0,310
0,330
0,410
Explanations to table 5: a) sum of 2-N-DPA and 4-N-DPA, b) sum of N-NO-2-N-DPA and N-NO-4-N-DPA, c)
sum of 2,4-DN-DPA, 2,2-DNDPA and 4,4-DN-DPA, d) sum of N-NO-2,4-DN-DPA and N-NO-4,4-DNDPA, e) sum of 2,2,4-TN-DPA, and 2,4,4-TN-DPA, f) 2,2,4,4-TeNDPA; g) conversion = molecular-weight
normalized sum of all diphenylamines
515
1,0
DPA
N-NO-DPA
0,8
[wgt-%]
MononitroDPAs
0,6
N-NO-mononitro-DPAs
Dinitro-DPAs
0,4
N-NO-dinitroDPAs
Trinitro-DPAs
0,2
TetranitroDPAs
Konversion
0,0
0
25
50
75
100
125
Exponentiell
(Konversion)
Q [J/g]
Fig 10. Stabiliser depletion in propellant V 6090 after microcalorimetry at 89C:

Bold lines: nitrosamines; dashed lines: amines. Konversion: molecularweight normalized sum of all diphenylamines
Table 6. Stabiliser depletion of propellant A 5020 after ageing at 89C.

All values in weight-%.
Ageing Ageing
t [d] Q [J/g]
DPA
N-NO- N-DPA N-NODPA

NDPAs
DNDPA
a)
b)
c)
d)
N-NO- Centra- N-NO- ConverDNlit I

N-EA
sion
DPAs
e)
f)
0,00
0,0
0,681
0,180
0,036
0,000
0,000
0,000
1,348
0,000
0,868
8,81
77,4
0,000
0,670
0,082
0,090
0,008
0,000
1,059
0,017
0,727
2,4-DN-DPA, d) sum of N-NO-2,4-DN-DPA and N-NO-4,4-DN-DPA, e) N-NO-N-EA = N-nitroso-N-ethyl
aniline, f) conversion = molecular-weight normalized sum of all diphenylamines
516
Table 7. Stabiliser depletion of propellant K 5810 after ageing

at different temperatures. All values in weight-%.
Ageing Ageing Ageing
T [C] t [d] Q [J/g]
DPA
N-NO- N-DPA N-NODPA

NDPAs
DNDPA
N-NO- TN- ConverDN- DPAs

sion
DPAs
a)
b)
c)
d)
e)
f)
0,00
0,0
0,500
0,486
0,073
0,000
0,000
0,000
0,000
0,972
89
0,30
2,3
0,353
0,550
0,079
0,000
0,000
0,000
0,000
0,885
89
0,77
4,2
0,244
0,693
0,094
0,000
0,000
0,000
0,000
0,910
89
1,09
5,7
0,247
0,785
0,109
0,000
0,000
0,000
0,000
1,003
70
35,80
23,4
0,000
0,959
0,021
0,144
0,000
0,000
0,000
0,935
89
4,06
47,6
0,000
0,860
0,011
0,260
0,055
0,016
0,000
0,978
80
14,00
50,8
0,000
0,873
0,000
0,270
0,022
0,018
0,000
0,968
80
14,00
51,7
0,000
0,839
0,000
0,260
0,034
0,015
0,000
0,937
89
5,70
72,4
0,000
0,679
0,000
0,318
0,077
0,039
0,000
0,897
89
6,76
89,5
0,000
0,482
0,000
0,356
0,191
0,062
0,014
0,864
89
6,80
90,0
0,000
0,502
0,012
0,347
0,192
0,045
0,010
0,863
sum of 2,4-DN-DPA, 2,2-DNDPA and 4,4-DN-DPA, d) sum of N-NO-2,4-DN-DPA and N-NO-4,4-DNDPA, e) 2,4,4-TN-DPA, f) molecular-weight normalized sum of all diphenylamines
1,2
DPA
[wgt-%]
N-NO-DPA
0,8
Mononitro-DPAs
0,6
N-NO-mono-nitroDPAs
Dinitro-DPAs
0,4
N-NO-dinitroDPAs
Trinitro-DPAs
0,2
Konversion
0
0
25
50
75
100
Polynomisch
(Konversion)
Q [J/g]
Fig 11. Stabiliser depletion in propellant K 5810 after microcalorimetry at 89C:
Bold lines: nitrosamines; dashed lines: amines. Konversion: molecularweight normalized sum of all diphenylamines
517
The following figures 12a-12d show a comparison of the distribution of stabilizer

depletion products of V 6090 (aged 5,1 days at 89C in completely filled and sealed
ampoules, total energy release = 42,8 J/g), of K 5810 (aged 4,1 days at 89C in completely
filled and sealed ampoules, total energy release = 47,6 J/g), of K 5810 (aged 40 days at 70C
in 12% filled and sealed ampoules, total energy release = 31,5 J/g) and of A 5020 (aged 9
days at 89C in completely filled and sealed ampoules, total energy release = 77,4 J/g).
Comparing the energy release they all show roughly a comparable decomposition degree.
Trinitro-DPAs
N-NO-dinitroDPAs
N-NO-DPA
MononitroDPAs
N-NO-mononitro-DPAs
V 6090, 5 d 89C, 43 J/g
Dinitro-DPAs
Fig. 12 a
Distribution of stabilizer depletion products in V 6090 after ageing
N-NO-dinitroN-NO-mono-nitroDPAs
DPAs Dinitro-DPAs
Trinitro-DPAs
N-NO-DPA
Mononitro-DPAs
K 5810, 4 d 89C, 48 J/g
Fig. 12 b. Distribution of stabilizer depletion products in K 5810 after ageing

K 5810, 40 d 70C, 31,5 J/g, low LD
N-NO-dinitroDPAs
Trinitro-DPAs
N-NO-DPA
Mononitro-DPAs
Dinitro-DPAs
N-NO-mono-nitroDPAs
Fig. 12 c. Distribution of stabilizer depletion products in K 5810 after ageing

(low LD)
518
N-NO-mono-nitro- Dinitro-DPAs N-NO-dinitroDPAs

DPAs
Mononitro-DPAs
N-NO-DPA
A 5020, 9 d 89C, 77 J/g
Fig. 12 d. Distribution of stabilizer depletion products in A 5020 after ageing
4.3
Microcalorimetry of pre-aged porous propellant
At the moment WIWEB runs a project to determine the activation parameters of the
decomposition reactions of propellants at 30-40C. In order to do this, samples are preaged
at these temperatures and after ageing they are measured by microcalorimetry. Characteristic
points of the HFC signals (maxima, minima etc.) are shifted towards earlier times. From the
time difference of the signals between unaged and aged samples at 89C (measuring
temperature) and the storage time and temperature (40C, 35C, or 30C) mean activation
parameters can be calculated by application of the Arrhenius equation [8].
P, W/g
T = 89C
2400
1800
Sample aged 345 days at 40C (1)
1200

unaged sample
600
0.0
0
enlarged area
see figure 14
Time, day
Fig 13. HFC curves of unaged and pre-aged V 6090 samples

(complete measurements)
519
P, W/g
T = 89C
200
150
unaged sample
100

50
Time, hour
0
0
20
40
60
80
Fig 14. HFC curves of unaged and pre-aged V 6090 samples (detail)
The program is performed with six different (gun and rocket) propellants, both A 5020
and V 6090 are a part of this study. The preceding figures show the HFC curves of unaged
and of aged V 6090 propellant, all measured in completely filled and sealed ampoules at
89C.
5.
DISCUSSION
5.1
Microcalorimetry of unaged samples under standard conditions
The samples show a very typical behaviour. Both A 5020 and K 5810 have the typical
shape of DPA stabilized propellants with a first maximum (this is the oxidation peak of the
nitrocellulose), a broad first minimum and a sharp increase up to a second maximum. This is
the point where the DPA is completely consumed. After the second maximum K 5810 shows
a slight decrease of the HFC curve whereas A 5020, which contains ethyl centralite in
addition to DPA has nearly no decrease of the heat generation rate after DPA has been
completely consumed. The endothermic peaks at the end of the measurements are due to gas
evolution from the ampoule because the decomposition gases build up a pressure which the
PTFE cap of the ampoules can stand only up to about 8 bar.
In contrast to these two compact propellants the porous propellant V 6090 shows a
completely different HFC pattern. After a first maximum (this is again the oxidation peak of
the nitrocellulose) and a short minimum a second maximum is observed. The HPLC
analyses of the aged propellant proved that this 2nd maximum is equivalent to the time until
the DPA is completely consumed. After that point the HFC curve decreases to a comparably
low level and stays there for a long time until after about 11 days a severe autocatalysis
520
occurs. Thus this propellant is one of very few qualified propellants that shows a severe
autocatalysis at 89C after a relatively short time period.
Nevertheless, all three propellants pass the stability criterion according to STANAG
4582 [9], where within 3,83 days at 89C (equivalent to ten years at 25C) a heat flow of
314 W/g should not be exceeded.
All three propellants show a comparable activation energy although the value for V
6090 is slightly higher than of other propellants (the average over all propellants is 130-142
kJ/mole [evaluation by iso- microcalorimetry in the temperature range between 70 and
89C]) but this is the first porous material that ever has been evaluated by this method.
5.2
Stabilizer depletion of unaged samples under standard conditions
V 6090 demonstrates a behaviour that is usually found when DPA stabilized samples are
aged in not completely filled and/or not hermetically closed vessels. The N-NO-DPA
reaches a maximum when DPA is completely consumed. In contrast to K 5810 (see below)
N-NO-DPA itself is readily consumed after that maximum and its nitration products as well
as di- and trinitro derivatives are formed. The stability criterion according to AOP 48 [10] is
not fufilled, because after 1,33 days at 89C or according to [9] 3,5 years at 25C the effective
stabilizer concentration is below 0,5 %. In the state of autocatalysis (see last entry in table 5)
still 0,41 % stabilizer depletion products are present, composed of a mixture of 2,2,4- and
2,4,4-TNDPA and of 2,2,4,4-TeNDPA. In this stage no nitrosamines could be detected.
K 5810 shows a stabilizer depletion which is practically identical to that of the closely
related ball powder K 6210 [7]. In closed and completely filled ampoules N-NO-DPA
dominates the depletion products UNTIL a decomposition degree of 90 J/g (see table 7). The
stability criterion according to AOP 48 [10] is fufilled until at least 72,4 J/g (5,7 days at 89C
or according to [9] 14,9 years at 25C). The second major fraction are the N-NO-mononitro
species, probably formed by a slow nitration of N-NO-DPA. N-H compounds play only a
minor role.
The stabilizer depletion of A 5020 is comparable to that of K 5810. After about 4.5 days
the DPA is completely consumed and N-NO-DPA reaches its maximum, followed by a slow
decrease. The stability criterion according to AOP 48 [10] is fufilled until at least 9 days at
89C or according to [9] 23.5 years at 25C. The small C I amount in this propellant starts to
be depleted after DPA is gone. Its major decomposition product is N-NO-N-ethyl aniline.
In all cases the sum of stabilizers (conversion) is decreasing with increasing ageing of
the propellant. Whereas this is likewise moderate in cases of A 5020 and K 5810 propellant,
the conversion is below 50 % of the original value in V 6090. This is in accordance with
literature findings and allows the assumption that part of the stabilizer has reacted with the
nitrocellulose chain [11].
5.3
Oxygen influence
The influence of oxygen on the decomposition of the three propellants was monitored by
variation of the loading density, by the application of inert gas (argon) and by the
comparison of bulk (A 5020, K 5810) and porous samples (V 6090).
If loading density is varied heat flow calorimetry curves of propellant V 6090 show a
very comparable (time and level) first maximum and first minimum. The second maximum
521
which corresponds to oxidation processes is much longer and has a higher energy release
when the loading density is reduced (see table 3). The time to autocatalysis is much longer
than in the case of a completely filled ampoule. This can easily be explained by a lower
concentration of NO2 in the atmosphere (less propellant which can release this molecule and
a larger gas volume which leads to a certain dilution).
K 5810 heat flow curves show a very strong dependence on the loading density. With
decreasing sample mass the energy release of the first maximum increases. If the loading
density is lower than 40 % a second oxidation peak appears after about one day at 89C. At a
LD of 32 % the HFC curve appears to be somewhat similar to the V 6090 curve (see fig. 9).
A detailed study of this propellant will be presented in [4].
Measurements under argon (there are only traces of oxygen present) with K 5810 prove
that the first maximum is due to oxidation processes, because they are nearly absent in this
case (see fig. 8 and table 4). Because oxygen is absent in this experiment the decomposition
of the stabilizer DPA (it also reacts as an antioxidant [12]) is depleted 13 hours (or 20 %) later
than the corresponding sample which had been measured under air.
In case of the V 6090 propellant there is a big difference in the HFC pattern in the 2nd
maximum (which is much lower than under air). That leads to the conclusion that this peak
is also connected to oxidation processes. But finally there is no difference in the time when
the strong autocatalysis occurs. If this time of autocatalysis is connected with a certain NO2
concentration this behaviour can easily be understood. Whether air or argon is used as gas in
the beginning the oxidation of the nitrocellulose and of the DPA has taken out all of the
oxygen in the sample and the self-decomposition of the nitrocellulose has delivered
sufficient NO2 into the gas phase to start the autocatalysis.
5.4
Influence of pre-ageing of samples
As mentioned in chapter 4.3 the results presented here are only preliminary because the
study will go on in the next years. The effect of pre-ageing is clearly visible in the oxidation
reactions in the beginning (see figures 13 and 14) and in the 2nd maximum of A 5020 (no
figure shown). The energy release of the oxidation reaction decreases, which indicates that
this reaction is going on at storage temperatures (30C, 35C and 40C). Table 8 collects the
evaluation of time shifts of pre-aged samples of propellants V 6090 and A 5020. Please note
that these values may change during the program, because higher degradation degrees could
not be realized by now.
It seems that the activation energy evaluated by the pre-ageing method is about
10 kJ/mole lower than the one calculated by the iso- procedure. Whether this is an effect of
the different approach, of the different temperature ranges or of the fact that the ageing has
not yet been finished will be discussed at the end of this study.
Within the (little) decomposition degree achieved by now the time to reach the main
autocatalysis does not decrease with pre-ageing (see fig. 13). There seems to be no correlation between the pre-ageing temperature and time and the time to autocatalysis. Therefore
this method is not applicable for the prediction of the time of the start of the autocatalytic
reaction. This once again shows that this point depends on the individual conditions in the
ampoule, probably the NO2 concentration. This may have been not fully controlled in the
experiments, although all the samples were weighed in at the same day with exactly the same
mass.
522
Table 8. Activation parameters from time shifts in HFC curves of pre-aged samples
(time of minima and maxima taken from V 6090 curve; point of inflection
taken from A 5020 curve)
Propellant Pre-ageing
V 6090
A 5020
6.
Activation energy 89C-35C Activation energy 89C-70C

(pre-aged samples) [kJ/mole]
(iso-) [kJ/mole] (see table 2)
158 d, 40C
130
145
345 d, 40C
135
145
367 d, 35C
132
145
166 d, 40C
124
135
341 d, 40C
125
135
370 d, 35C
123
135
CONCLUSION
This work demonstrates the influence of the sample type and the ageing conditions on
the decomposition processes of propellants. Oxygen plays an important role in the
decomposition of propellants. It has three major impacts on this reaction. It oxidises the
nitrocellulose, it oxidises the stabilizer and it changes non-reactive NO back into highly
corrosive NO2. If oxygen is avoided the propellants chemical stability is improved although
the self-decomposition reactions of the propellants deliver enough nitrogen oxides to cause a
severe autocatalysis in the case of porous propellants. If propellants are aged it is important
to know the oxygen influence on the ageing parameters. So it is recommended to do ageing
studies in complete ammunition articles or, at least, under storage conditions as closely as
possible related to the conditions in ammunition.
The activation parameters of the decomposition reactions are comparable to those found
in other propellants. In single base propellants there seems to be a smaller activation energy
at lower temperature (30-40C) compared to higher temperatures (70-89C) which has to be
proven (see also [13]).
All three propellants show a sufficient chemical stability according to HFC criteria; but
the porous propellant does not fulfil the stability criterion for stabilizer depletion [10]. This
will not be the only case with a discrepancy between these two completely different
STANAG stability evaluation methods [9, 10]. Therefore discussions amongst the experts and
additional work will be necessary to allow a decision which method is the superior one.
523
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
HENRI MURAOUR: La Dcomposition des poudres et explosifs, Chimie & Industrie, 20, 110117 (1928); JEAN TRANCHANT, Internal Mechanism of the Chemical Evolution of
Nitrocellulose Propellants Hypotheses and Consequences; Symp. Chem. Probl. Connected
Stabil. Explos. 6, 1-20 (1982).
OSCAR GUTTMANN: Zwanzig Jahre Fortschritte in Explosivstoffen, Julius Springer Verlag
Berlin 1909, p. 55.
STEPHAN WILKER, ULDIS TICMANIS, LUTZ STOTTMEISTER, PIERRE GUILLAUME, BEAT
VOGELSANGER, CORINNE BALS, PASQUALE BISOGNI, WIM DE KLERK, DAVID TUCKER,
DAVID J. WOOD: Influence of atmospheric conditions on the ageing of single, double and
triple base propellants, Symp. Chem. Probl. Connected Stabil. Explos. 12 (2001), in press.
The K 5810 decomposition under different ageing conditions will be presented in detail in:
PIERRE GUILLAUME, LAURENCE JEUNIEAU, MICHEL LEFBVRE, SAMMY CHEVALIER,
STEPHAN WILKER, GABRIELE PANTEL, ULDIS TICMANIS: Stability analyses of rolled ball
propellants Part I , Int. Annu. Conf. ICT 34 (2003).
S. WILKER, U. TICMANIS, G. PANTEL, P. GUILLAUME: Detailed investigation of sensitivity
and reproducibility of HFC measurements, Symp.Chem.Probl.Conn. Stabil. Explos., Bstad
11, 383ff. 1998.
STOERMER, HOFFMANN: Chem.Ber. 31, 2535 (1898) [Beilstein XII, page 728].
PIERRE GUILLAUME, MAURICETTE RAT, GABRIELE PANTEL, STEPHAN WILKER: Heat Flow
Calorimetry of propellants Effects of sample preparation and Measuring Conditions,
Propellants, Explosives, Pyrotechnics 26, 51-57 (2001).
A full paper about this study will be presented probably in 2005 when the ageing of the
samples will be finished. Preliminary results are collected in: ULDIS TICMANIS, STEPHAN
WILKER, GABRIELE PANTEL: Tieftemperatur-Aktivierungsenergie von Treibladungspulvern
Machbarkeitsstudie, WIWEB Report 710/26374 (2002).
STANAG 4582, Explosives, Nitrocellulose Based Propellants, Stability Test Procedures and
Requirements Using Heat Flow Calorimetry, 1st Edition, March 2003; ULDIS TICMANIS,
STEPHAN WILKER, GABRIELE PANTEL, MANFRED KAISER, PIERRE GUILLAUME, CORINNE
BALS, NIELS V.D. MEER: Principles of a STANAG for the estimation of the chemical
stability of propellants by heat flow calorimetry, Int. Annu. Conf. ICT 31, 2 (2000).
AOP-48, Explosives, Nitrocellulose Based Propellants, Stability Test Procedures and
Requirements Using Stabilizer Depletion, 1st Edition, November 2002.
TORBJRN LINDBLOM: Reactions in Stabilizer and Between Stabilizer and Nitrocellulose
in Propellants, Propellants, Explosives, Pyrotechnics 27, 197-208 (2002).
A.N. PANKRATOV, Structure of the Products of the Oxidation of Diphenylamine, the first
member in a Family of Analytical Redox Reagents, J. Anal. Chem. 56, 140-142 (2001)
and references cited therein; JAN PETRLEK, STEPHAN WILKER, ULDIS TICMANIS, GABRIELE
PANTEL, LUTZ STOTTMEISTER, JAN SKLDAL, Stability analyses of propellants containing
new stabilizers Part II , Int. Annu. Conf. ICT 32, 16 (2001).
ULDIS TICMANIS, GABRIELE PANTEL, LUTZ STOTTMEISTER: Stabilittsuntersuchungen
einbasiger Treibladungspulver Mikrokalorimetrie im Grenzbereich, Int. Annu. Conf. ICT
29, 27 (1998).
524
OVERPRESSURE GENERATED BY DETONATIONS

OF ELONGATED GAS CLOUDS
Waldemar Witkowski*, Karol Buchalik*, Radoslaw Trebinski**,
Andrzej Maranda** and Andrzej Teodorczyk***
Annopol 6, 03-236 Warsaw, Poland
Kaliskiego 2, 00-908 Warsaw, Poland;
*** Warsaw University of Technology, ITC,
Nowowiejska 25, 00-665 Warsaw, Poland
Abstract:
Experiments were performed to study overpressure generated in air by detonations of
acetylene-air mixtures in a form of elongated clouds. In the first series of experiments
the influence of cloud diameter on overpressures was studied. Next, the series of
experiments were performed with the clouds of the same volume but different diameter
to length ratio. To examine the influence of ignition location on blast wave intensity the
experiments were performed for five different locations along main axis of cylindrical
cloud. Experimental results were compared with the predictions of Dorofeevs empirical
formula for spherical clouds and TNT equivalent.
The results were used for formulating the expressions, which describeb overpressure
along two main directions.
Keyword:
1.
gaseous detonation, blast wave, TNT equivalent
INTRODUCTION
Most often, risk estimation from fuel-air explosions (FAE) is based on a model of
explosion in spherical or hemispherical shape. Whereas, analysis of results of explosions
caused by flammable gases or vapours shows that the source of the blast waves are elongated
fuel-air clouds. Blast waves generated by such sources of disturbances have a little or
medium intensity, but they act on a big area and possess a significant asymmetry relating to
the spherical explosion [1].
One of the first pieces of information about the asymmetrical blast waves come from the
sources [2, 3, 4]. In paper [1] the results of modeling for the wave generated by scattering of an
ellipsoidal gas cloud, under proportion 2:1 for ellipse axes are shown. It appears that for a
low initial value of overpressure in the cloud, the asymmetry of overpressure field remains
even in a long distance from the cloud.
In paper [5] the results of numerical calculations for free clouds are shown. The shape of
such clouds is symmetrical in relation to the center of the Cartesian coordinates.
There is comparatively a little number of literature data relating to the explosions of
other than spherical clouds.
525
2.
EXPERIMENT
The work presented here is experimental. Our experiments were based on measurements
of intensity of the blast wave, which was generated by the explosions of acetylene-air
mixtures containing 15 % vol. of fuel. Investigated fuel-air mixtures were not free gas
clouds. They were contained in containers, which were made from thin-walled polyethylene
foil. This foil was 40 m thick. The containers had similar cylindrical shapes but different
diameters: 0.31, 0.36 and 0.49 m. Eight different gas clouds with volume ranged from 115 to
371 m3 and length ranged from 1.05 to 4.72 m were investigated. Four filled containers are
shown in photograph - Fig. 1.
Fig 1.
A view of containers, described as L105D36, L220D49, L345D36,

L472D31 (L and D mean length and diameter of container, in cm).
The containers were hung up at the height of 1.5 m above the ground and were initiated
by electric blasting cup no. 8. There was slight overpressure of 500 Pa of the explosion
mixture which was inside, to assure a proper shape of the container.
The intensity of the blast wave, outside the exploding cloud, was measured by PCB
sensors. The sensors were located in two directions: along main axis of cylindrical cloud
p(r) and perpendicularly towards the main axis p(p). The sensors were placed at the
height of 1.5 m above the ground in such a way, to assure the measurement of positive
overpressure and impulse in blast wave. Overpressure was measured in 8 different places
outside cloud, located at the distance of 0.25 to 3 m from the cloud. The places of
measurement of overpressure according to geometry of the cloud are shown in Fig. 2.
The detonation velocity of the applied acetylene-air mixture was 204645 m/s.
3.
RESULTS
In the first series of experiments the influence of cloud diameter on overpressures was
studied. The gas clouds were the same in length (about 2.26 m) but different diameters: 0.31
m, 0.36 m and 0.49 m, respectively Pressures were measured at six locations outside the
cloud at the distance of 0.25 m to 2.0 m from cloud border. It was found that up to some
526
distance the overpressure along cloud main axis exceeded the value in perpendicular
direction. Distance at which these overpressures become similar depends on cloud diameter.
r(r)
a(p)
p(r)
p(p)
PRESSURE
GAUGE
Fig 2.
Directions of measuring of overpressure.
For three clouds, mentioned above, this distance is about 1.4 times bigger than diameters
of the clouds. At longer distances from the cloud border, overpressure in perpendicular
direction was greater and difference in blast wave intensity in both directions was the larger
as the ratio of cloud diameter to its length was smaller.
Figure 3 shows typical plot of overpressure variation vs distance from the cloud border in
two directions.
5,5
4,5
(p-p0)/p0
3,5
2,5
1,5
0,5
-0,5
0,0
0,4
0,8
1,2
1,6
2,0
2,4
a(p)=a(r), m
Fig 3.
Blast wave intensity vs distance from cloud border; gas cloud

2.32 m long and 0.31 m in diameter.
The obtained changes of overpressure in both directions had similar characteristics for
all investigated clouds.
Positive phases of impulse of blast wave were measured in settled distances from cloud
border. They had similar profiles to this one which is shown in Fig. 4.
527
3.5
p(p)=36.4 kPa/div
p(r)=33.2 kPa/div
3
2.5
p(p)
p(r)
1.5
1
0.5
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
time, ms
Fig 4.
Intensity and profile of the blast wave at a distance of 0.5 m

from border of cloud marked L472D31.
Next, series of experiments were performed with the clouds of the same volume of 371
dm but different diameter to length ratio, equal to 1:15.2, 1:8.9 and 1:3.9. Detonation was
initiated at the cloud end or in the center of symmetry. The overpressure was measured at the
distance of 3 m from the center of symmetry in both directions. An asymmetry in blast
intensity was detected. Results of the measurements are presented in Table 1.
3
Table 1. Overpressure [kPa] for clouds with different geometry and equal
volume of 371 dm3.
Symbol
of bag
Central ignition
Side ignition
p(p)
p(r)
p(p)
p(r)
L472D31
24.0
113.9
32.0
113.9
L322D36
31.4
35.1
33.7
36.4
L190D49
38.2
27.3
40.3
29.2
To perform the detailed studies of the influence of ignition location on blast wave
intensity the experiments were performed in gas cloud 4.72 m long and 0.31 m in diameter.
The explosions were initiated in five different locations along main axis of cylindrical cloud
fig. 5. Initiation point was changed along main axis of the cloud from the one end of the
cloud (A point) to the other (E point). The larger influence of ignition location on blast wave
intensity was clearly determined in the direction of main axis. Graphic presentation of the
results is shown in Fig. 6
528
A - E: IGNITION
LOCATION
PRESSURE
TRANSDUCERS
Fig 5.
Location of initiation places of the cloud

120
p(r)
100
p, [kPa]
80
60
40
p(p)
20
0
ignition location
Fig 6.
4.
Influence of the initiation places on distribution of overpressure
RESULTS ANALYSIS.
Experimental data were worked out with the use of a non-linear estimation procedure.
Approximating functions are shown in a following form:
p + a1
a
a
= b1 + b22 + b33
p0
For analysis the Sachs dimensionless variables were used:

=
r p1/3
0
,
E1/3
p=
p0
,
p0
I=
Ic
E p
0
1/3 1/3
0
The comparison of the measured blast parameters was carried out using Dorofeevs
formulae [6]:
p + 0.34 0.062 0.033
= 4/3 + 2 + 3 , [0.21 3.7]
p0
Ic
E p
0
1/3 2/3
0
= 0.0353 0.968
529
One stated, that for all investigated clouds, shown above Dorofeevs formula does not
describe enough precisely the changes of overpressure in the whole range of distances.
Therefore, a trying to find the function for approximating the changes of overpressure vs
distance was undertaken. We were looking for an expression of the formula in a form as
similar to Dorofeevs polynomial as possible.
And finally, for perpendicular direction towards main axis of cloud, the changes of
overpressure vs reduced distance may be described by the formulae mentioned below:
a) for the cloud with diameter to length ratio equal 1 : 7.5
p + 0.620 0.078 0.002

= 4/3 2 +
p0
3
b) for the cloud with diameter to length ratio equal 1 : 6.3
p + 0.516 0.0004 0.001

= 4/3 +
3
p0
c) for the cloud with diameter to length ratio equal 1 : 4.5
p + 1.257 0.661 0.104

= 4/3 2 +
p0
3
All experimental data for perpendicular direction towards main axis of cloud and they
were collected and compared with results of numerical modeling. The scattering of
detonation products of an infinite cylindrical charge, initiated in its axis was modeled [7].
The change of overpressure in a blast wave is connected with a mentioned below parameter:
p r r0
=
p0
r0
where:
r - distance from main axis of the cloud,

r0 - radius of the cloud.
The numerical modeling was carried out with use of mentioned below expression:
p
= a 1 4/3 + a 2 2 + a 3 3 + a 4 4
p0
where:
=(r-r0)/r0
a1=16.4, a2=-22.4, a3=15.28, a4=-8.32
for 1.8<<25
Experimental data and data obtained from calculations for perpendicular direction
towards main axis of the cloud are shown in Fig. 7.
530
(p-p0)/p0
1,0
a
0,5
a- calculated data
b - experimental data
0,1
8 10
20
(r-r0)/r0
Fig 7.
Comparison of the data obtained from
experiments and from calculations.
It is easy to observe a good conformability of results in a wide range of distance from

main axis of the cloud.
It is possible to approximate experimental data, which are presented in Fig. 7 (curve b),
as far as a necessity for the first estimation is taken into consideration. It can be shown by
following equation:
where:
Y=1.272-1.002*X
Y=Ln((p-p0)/p0)
X=Ln((r-r0)/r0).
A straight line as graphic presentation of this approximation is shown in Fig. 8.
2,0
1,5
Ln((p-p 0)/p0)
1,0
0,5
0,0
-0,5
-1,0
-1,5
-2,0
-2,5
0,5
1,0
1,5
2,0
Ln((r-r0)/r0)
2,5
3,0
Fig 8.
Changes of overpressure in perpendicular
direction towards main axis of the cloud vs
diameter of the cloud and distance from a
main axis.
Analysis of the results collected along longitudinal axis of the cloud direction confirmed
the specificity of overpressure field generated by the elongated charges of exploding gases.
The Dorofeevs formula does not describe this direction with good enough accuracy in a
wide range of the reduced distance.
This is why a attempt of description of the overpressure field in this direction by the use
of dimensionless parameter was undertaken. The parameter was defined as the ratio of
distance from a cloud border to diameter of this cloud, equal = a(r)/D.
Based on experimental data for different clouds, changes of intensity of wave vs
dimensionless parameter can be shown as in Fig. 9.
531
5,0
1:7.5
(p-p0)/p0
1:6.3
1,0
1:4.5
0,5
1 : 15.2
0,1
5
a(r)/D
Fig 9.
Changes of overpressure at main axis direction of clouds.
The intensity of wave in this direction was described by a simple expression :

Y=1.2025-1.6065*X
Y=Ln((p-p0)/p0)
where:
X=Ln()
Presented above considerations complete calculations of positive phase of impulse of
blast wave in both investigated directions. The value of impulse for the cloud with the
biggest elongation ratio (1:15.2) are shown in figures below. The changes of impulse for this
same volume of gas contained in a shape of sphere were calculated according to the
Dorofeevs formulae. They are presented in Fig. 10.
For given range of the distances, the obtained results permit to describe the impulse of
wave as a function of the reduced distance in such a way:
perpendicular direction
Ic
E p
0
1/3 2/3
0
= 0.0412 0.446
parallel direction
Ic
E p
0
1/3 2/3
0
= 0.0552 2.284
532
60
Dorofeev's formula
impuls,[Pa*s]
50
40
30
parallel direction
20
10
0,6
0,8
1,0
1,2
1,4
1,6
1,8
2,0
2,2
reduced distance
Fig 10. Comparison of impulses generated by a cloud with sphere shape

and with elongated shape.
The results presented in this work refer only to the one explosive mixture. Therefore,
they have not an energetic calibration character. But, they are a practical advice for
estimation of a disappearance of intensity of a blast wave as a function of the explosion
arrangements that was described. It is very useful, because nor Dorofeevs formulae for
spherical explosions nor trinitrotoluene (TNT) equivalent formulae for point explosions, do
not describe the investigated arrangements, exactly good enough. This is confirmed by
nearby presented graph - Figure 11. Experimental data regarding to the cloud with
proportion of diameter to length equal 1:7.5 were compared with data obtained from
calculation with Dorofeevs formulae regarding to this same volume of the mixture and with
data from calculation for estimation for shock waves generated by TNT charges [8].
20
TNT charge
Dorofeev's formula
(p-p 0)/p0
15
10
parallel direction
0,0
0,4
0,8
1,2
reduced distance
Fig 11. Comparison of overpressure generated by different sources.
533
5.
CONCLUSION
Based on performed experiments it can be affirmed that:
explosion of an elongated gas cloud generates asymmetrical field of overpressure
in environments; parameters of this overpressure depend mostly on volume and
shape of the cloud and on distance from the cloud border.
for elongated clouds, with a shape similar to the cylindrical, a scale of asymmetry
appears mostly in two directions: parallel and perpendicular towards the main
axis of the cloud.
if a distance from a border of the cloud does not exceed 130 - 140 % of diameter
of the cloud, the intensity of wave is bigger at parallel direction towards the main
axis than at perpendicular direction.
if a distance from a border of the cloud exceeds 130 - 140 % of diameter of the
cloud, the intensity at perpendicular direction is bigger than in parallel direction;
for the cloud with the greatest proportion of length to diameter (15.2 : 1) and at
distance 16 times bigger than diameter, the differences in intensity reach a few
kilopascal.
in a close to explosion zone, a valuation of the intensity of wave based on
spherical symmetry principles gives a big error.
Dorofeevs formulae for spherical explosions can be used for explosions of
elongated charges in a far explosion zone; for all investigated clouds this
formulae describes perpendicular towards main axis direction better than parallel
direction.
perceptible influence of the place of initiation of the cloud on the generated
overpressure field suggests that it is necessary to take into account, so called
effective length of the cloud as far as the estimation of intensity of the wave along
main axis of the cloud direction is taken into consideration.
REFERENCES
[1] LEE JHS: (1984), Physics of Explosions, McGill University, Montreal
[2] PANARELLA E ET AL: (1968), Blast waves from a laser-induced spark in air, Canadian Journal
of Physics, 46, 183
[3] LAUMBACH DD ET AL (1969), A point explosion in a cold exponential atmosphere, Journal of
Fluid Mechanics, 35, 1
[4] CHIU KW ET AL (1977), The Blast Waves from Asymmetrical Explosions, Journal of Fluid
Mechanics, 82, 1
[5] :BAKER WE (1983), Explosion Hazards and Evaluation, Elsevier Scientific Publishing
Company, New York
[6] DOROFEEV SB: (1995), Blast effects of confined and unconfined explosions, Proceedings of the
20th International Symposium on Shock Waves
[7] KUCZAJ A: (1999), Numerical analysis of detonation products scattering for an infinite
cylindrical charge, Military University of Technology, Warsaw (in Polish)
[8] STANJUKOVICH KP ET AL.: (1975), Physics of Explosion, Moscow (in Russian)
534
PHYSICOCHEMICAL PROPERTIES OF NITROCELLULOSE

MIXTURES WITH LOW MOLECULAR COMPOUNDS
IN THE SHAPE OF FOILS AND IN WATER SUSPENSIONS
T. Wolszakiewicz* and A. Ksiczak**
* Institute of Organic Industrial Chemistry,
Annopol 6, 03 236 Warsaw, Poland
** Department of Chemistry, Warsaw University of Technology
Noakowskiego 3, 00 - 664 Warsaw, Poland
Abstract:
Samples of nitrocellulose mixed with the low molecular compounds, prepared as foils
and water suspensions were stored and the long - therm changes were observed. Some
their parameters were determined - the glass transition temperatures and the difference
between the maximum of DSC peak due to relaxation effect. The analysis of DSC
curves after different storage time showed that the migration undergoes into the
nitrocellulose structure. For the samples prepared as a foil thermal effects of phase
changes were very low. It proves that low molecular component is fully bonded with
nitrocellulose chains.
1.
INTRODUCTION
The aim of this work is prediction of thermal behaviour of two component mixtures
of nitrocellulose (NC) and low molecular components such as 2,4 dinitrotoluene
(2,4-DNT), diphenyloamine (DPA) and 2-nitrodiphenyloamine (2nDPA) during prolonged
storage. This low molecular compounds are used as modifiers and plasticisers for NC.
2.
EXPERIMENTAL
The wooden NC produced in ZPS Pronit in Pionki, with nitrogen content 13.2% and
average molecular mass Mn = 54 103 [g/mol] was used for experiments. Before experiments
NC was rinsed by distilled water during 24 h to remove the impurities. The measurements
were conducted using Perkin-Elmer differential scanning calorimeter, model Pyris 1. The
two component samples was proposed to determine the properties of grain gunpowder for
possibility of application the obtained results in real technological processes of production.
In the first described method of sample preparation the components of mixture with certain
amounts of component were added to sufficient volume of water for forming proper
mixture. The mixture was heated in the water bath at temperature higher than 10 0C than the
melting point of low molecular component and afterwards mixed by 2 hours. The liquid low
molecular component non-miscible with water was absorbed completely by nitrocellulose
fiber. Than nitrocellulose with low molecular component was filtered and dried. This process
was similar as in the case of technological operations applied for production of gunpowder.
The structure of nitrocellulose was modified not only by the chemical interactions but also
by using of mechanical forces. In the second method the mixture with established batch
535
content was dissolved in large amounts of volatile solvent (acetone or ethyl acetate). The
homogenic solution was poured into the surface of mercury to obtain the thin layer foil. This
foil was dried at about 40 0C in vacuum dryer for 2 weeks. The fibrous structure of
nitrocellulose was destroyed by action of volatile solvent in this method.
3.
The thermal effects after prolonged storage were examined for the samples prepared by
both methods. The temperature of glass transition of mixtures and the shift of the glass
transition temperature responding to relaxation [1, 2] were established.
Sygna DSC
The Figure 1 presents DSC curve of mixture xw24DNT = 0.065 obtained from water
suspension after 21 days from time of preparation. We can see the peak responding
relaxation at temperature T = 324.9 K with enthalpy H = 0.49 J/g, and next the
exothermal peak responding to the penetration of 24DNT particles on the surface of
nitrocellulose fibres.
280
305
330
355
380
T em p eratu ra [K ]
Fig 1.
The DSC curve for mixture containing xw24DNT = 0.065 derived from
the water suspension after 21 days from time of preparation.
The relaxation has the great influence for calorimetric signal changes in the case of
mixtures obtained in a shape of foils. The figure 2 presents the DSC curves of mixture
containing xw24DNT = 0.030 after one heating (thick line) and another heating made on the
same sample (thin line), for the same cycle of measurement after 20 days after time of
sample preparation. The peak corresponding the relaxation effect at temperature T = 337.8
K with heat effect H = 1.07 J/g was visible after the first heating. The deviation of
calorimetric signal is toward the exothermal processes due to probability of lowering the
decrease of heat capacity and exothermal effects of thermal decomposition or mixing of
components.
536
Sygna DSC
200
260
320
380
Temperatura [K]
Fig 2.
The DSC curves of mixture xw24DNT = 0.030 obtained in the shape of foil
20 days after time of sample preparation. The first heating (thick line),
second heating (thin line) for the same cycle.
In the second heating the calorimetric changes at this temperature range were low. It
proves that there were no decomposition process the first and second heating. The low
concentration of 2,4-DNT causes that there were no exothermic effect for mixing of
components. It draws to the conclusion that at this temperature range the structural changes
of nitrocelluloes appears. They are caused by association of nitrocellulose chains. After 412
days the same sample was examined again twice. In this measurements temperature of
relaxation was higher and was T = 349.2 K (it was the highest temperature observed), and
enthalpy was lower H = 0.77 J/g. The decrease of the thermal effect suggested that there
were decrease of structural changes with ageing of sample. The figure 3 presents the DSC
curve for mixture xw24DNT = 0.155 obtained as a foil. The measurement was performed 6 days
after time of preparation of the sample. The peak corresponding to relaxation appeared at
temperature T = 335.2 K and enthalpy was H = 0.23 J/g. The baseline after peak of
relaxation sharply goes down and it can be attributed to great and rapid structural changes of
nitrocellulose. In next measurement performed after 528 days, temperature of relaxation
increased only 2C, and enthalpy decreased significantly and was H = 0.03 J/g. In spite of
such a small thermal effect, baseline changes were also great and showing for similar
structural changes of polymer matrix.
537
Sygna DSC
200
240
280
320
360
Temperatura [K]
Fig 3.
The DSC curve for mixture xw24DNT = 0.155 obtained as a foil.

Measurement was performed 6 days after time of preparation.
Due to acidic properties of DPA the mixtures with nitrocellulose showed lower thermal
stability. Figure 4. presents the DSC curve of mixture xwDPA = 0.105 measured 354 days after
preparation of water suspension. Thermal effects of changes were very small and it can be
seen the changes of calorimetric signal due to work of compensation system. The small
endothermic peak with enthalpy H = 1.17 J/g with maximum Tp = 321.6 K could be
attributed to relaxation or melting process in small diameter pores. This effect was not
analysed. The lack of relation between mixture content and value of endothermal effect
suggest that it appears due to relaxation. But at 360 K is visible the small exotherm as a
deviation of baseline as an effect of beginning of mixture decomposition. For mixtures with
range of content from xwDPA = 0.05 to 0.25 temperature of this effect was nearly the same.
Because this effects were small the heating rate was high ( = 20C/min). Figure 5 presents
DSC curve for mixture xwDPA = 0.105 using heating rate = 20C/min. It can be seen on
DSC curve not only endothermic peak, previously observed in figure 4 but also endothermic
peaks in high temperatures above 343 K, confirming the new structures of nitrocellulose or
interactions between different molecules.
538
Sygna DSC
300
320
340
360
Temperatura [K]
The DCS curve of mixture xwDPA = 0.105 with heating rate = 2C/min
354 days after time of preparation of water suspension.
Sygnal DSC
Fig 4.
280
305
330
355
380
Temperatura [K]
Fig 5.
The DSC curve for mixture xwDPA = 0.105 obtained from water suspension
for heating rate
= 20C/min
539
The Table 1. contains the glass transition temperatures for mixtures obtained from water
suspension.
Table 1. The glass transition temperature (Tgm) of mixtures NC + DPA obtained
from water suspensions.
xwDPA
Tgm [K]
0.062
225.1
0.102
226.2
0.183
226.6
Sygna DSC
The similar shape of DSC curves in high temperatures were observed for mixtures
obtained in a shape of foils. Figure 6 presents the DSC curve for mixture xwDPA = 0.200
formed as a foil . The measurement was performed with hetaing rate = 20C/min.
300
320
340
360
380
Temperatura [K]
Fig 6.
The DCS curve of mixture xwDPA = 0.200 formed as a foil.

The measurement was performed with heating rate = 20C/min.
In this case the complex peak was observed before the beginning of decomposition
process. The similar effect was observed for the mixture with lower content of DPA xwDPA =
0.048 and it suggest that observed thermal changes are attributed to structural changes of
nitrocellulose. Information about placement of DPA in polymeric matrix obtained as a foil
are clearlu seen from the DSC curves presented in figure 7. The measurement was performed
in cycle of two subsequent heating for the mixture xwDPA = 0.200 after several days after
preparation. After first heating there were no glass transition process visible. The weak
540
Sygna DSC
endothermic effect at temperature Tp = 318.8 K and H = 0.37 J/g was converting in further
stage into exothermal effect attributed to penetration of DPA into deeper structures of
nitrocellulose. The process seems to be complicated as visible as calorimetric signal. The
lack of glass transition peak and melting peak proves that DPA particles are completely
spread, mainly on the nitrocellulose surface. The particles can penetrate deeper structures of
nitrocellulose at temperature above relaxation because of perturbations of arrangement at
the surface. The presence of glass transition in second heating and melting in limited surface
is the proof that particles of DPA are released from sorting action of nitrocellulose. In this
case it can be attributed to melting, but in such a short time, the relaxation is not appearing.
The shift of peak maximum is Tm = 7.07 C. The glass transition temperature is Tgm =
215.8 K and is significantly lower than temperatures placed in Table 1 .
200
260
320
380
Temperatura [K]
Fig 7.
The DSC curve for mixture xwDPA = 0.200 several days after time of
sample preparation. First heating (thick line), the second heating (thin line)
for the same measurement for one sample.
Figure 8. presents the beginning parts of DSC curves before melting peak for mixture
xw2nDPA = 0.200 in first heating (thick line) and second heating (thin line) for the same
measurement cycle. The grass transition was not observed despite of crystallization before
melting in first heating. It means the presence of liquid phase when molecules were
arranged on the surface of nitrocellulose and dont take part in glass transition. The
movements of particles caused by temperature destroyed the arrangement in higher
temperature and it allows for crystallization of hydrated particles. Thermal motions in high
temperatures are great and causes the release of particles form the influence of interactions
with nitrocellulose chains. Only glass transitions are observed in second heating. The
agreement of maximum main melting peak of 2nDPA for freshly prepared mixture and for
pure component shows that introduction of low molecular component to nitrocellulose fibres
from suspension water and low molecular component is mainly on the surface of fiber.
541
Sygna DSC
Probably water from nitrocellulose pores effectively blocks admission into deeper parts of
nitrocellulose. During drying the particles of 2nDPA are shoved forward by bladders of
water vapours evacuated form pores. The analysis of DSC curves of samples after different
time of storage calculated after drying showed the presence of migration into nitrocellulose
structures. The value of melting enthalpy diminished at about 50% and it means that 2nDPA
particles was placed under the surface as the thin layers which arrangement is determined by
interaction of nitrocellulose surface and they dont undergo by phase transitions. The peak of
relaxation shift towards the higher temperatures with the growth of storage time. The
maximum temperature after 4 days was T = 316.5 K, and after 23 days - T = 327.2 K.
After 23 days the endothermal peak appeared at maximum temperature Tp = 368.6 K and
enthalpy of changes H = 0.08 J/g. For mixtures xw2nDPA > 0.200 the weak exothermic effect
was observed after melting for freshly prepared samples and it proves the progressive mixing
of components. It leads to the conclusion that during mixing of components homogenisation
is not completed.
210
245
280
315
350
Temperatura [K]
Fig 8.
The beginning of DSC curves before melting peak for mixture

xw2nDPA=0.200 obtained from water suspension in first heating (thick line )
and second heating (thin line) for the same cycle of measurements.
542
4.
CONCLUSION
The results of investigations presents of significant structural changes during storage of

samples prepared as foils and water emulsions. The glass transition temperature was
determined for mixtures and the shifts of peaks connected with relaxation. The analysis of
DSC curves for examined mixtures after different storage times showed the migration of low
molecular compounds into nitrocellulose structure.
REFERENCES
[1] C. ASH, T. J. LEWIS: Polymer Papers, 26 (1985) 643.
[2] R. C. WARREN: Polymer, 31 (1990) 861.
543
SYNTHESIS OF N-ACETYL-3,3-DINITROAZETIDINE *
Shu Yuanjie, Li Hongzhen, Huang Yigang and Liu Shijun
Institute of Chemical Materials CAEP,
621900, Mianyang, Sichuan, China
Abstract:
N-Acetyl-3,3-dinitroazetidine(ADNAZ) has such advantages as low melting point,
insensitive to detonation on impact and forming a eutectic with TNAZ which has many
applications in melt-cast explosive production. ADNAZ was synthesized by a five-step
reaction, using nitromethane, paraform aldehyde and tert-butylamine etc. as starting
materials with overall yield of 42.8%. The factors affecting the preparation were
discussed in detail. Structures of intermediates and ADNAZ were confirmed by
FTIR,HNMR,CNMR and MS.
Keyword:
synthesis N-Acetyl-3,3-dinitroazetidine(ADNAZ), energetic material
INTRODUCTION
1.
Multinitroazetidines and their derivatives, as important high energetic density materials,

have excellent detonation performance because of their ring-tension strength. Among them,
TNAZ[1] is a good example and widely used. N-acetyl-3,3 dinitroazetide ( ADNAZ) is
another one, it is white, soluble in organic solvents as acetone and methyl chloride. ADNAZ
has such advantages as low melting point (114.8C), relative high crystal density
(1.55g/cm3) and insensitive to detonation on impact etc.. So, ADNAZ is suitable as an
ingredient in energetic materials, either by itself or with other compounds. Especially
ADNAZ can form a eutectic with TNAZ (TNAZ/ADNAZ =66/34) with melting point about
78.6C [2]. So it can be melt and cast or be used as an energetic carrier for higher melting
energetic materials such as HMX and the like. On the other hand, in ADNAZ, the acetyl
group attached to the nitrogen atom can easily be replaced by other group (such as nitro
group), ADNAZ can be used as intermediate to prepare other energetic compounds (such as
TNAZ[3]). In this paper, ADNAZ is synsthesized by a five-steps reaction, starting from
nitromethane, paraformaldehyde and tert-butylamine etc., with overall yield of 42.8%.
2.
EXPERIMENTAL
2.1
Equipment and reaction reagents
Reagents: CH3NO2, (CH2O)n, t-BuNH2,NaOH, HCl, DIAD, Ph3P, NaNO2, BF3Et2O are
all chemical pure.
Equipment for analysis: FTIR, Nicolet-800; NMR, Varian INOVA400 MHz;
Magnetic-Mass Spectroscopy, Finnigan MAT-95S etc.
This project was supported by fund of CAEP (20020539)
544
2.2
Reaction
OH
NaOH
CH3NO2 + CH2O
HCl
O
t-BuNH2
t-Bu
NH
NO2
N
Ac
BF3. Et2O
Ac2O
(ADNAZ,1)
2.3
NO2
O2N
N
t-Bu
(5)
Ph3P
NaOH ,NaNO2
K3Fe(CN)6 , Na2S2O8
OH
t-Bu. HCl
(2)
O2N
OH
O2N
O2N
O2N
(3)
DIAD
OH
N
t-Bu . HCl
(4)
Preparation
2.3.1. 3-Tert-butyl-5-hydroxymethyl-5-nitrotetrahydro-1,3-oxazine(2)
To a solution of paraformaldehyde(12.12g,0.4mol) and NaOH (cat.) in water at 40-50C
was slowly added nitromethane (5.38ml,0.1mol). The solution was heated to 60-80C and
tert-butylaming (10.66mlm 0.1mol) was added dropwise. The mixture was then stirred for a
further hour, cooled to ambient temperature and stirred for 16 hours. After filtration, washing
and drying to constant mass, the precipitate was collected to give (2) as a white powder, m.p.
138.8-139.4C.
2.3.2. 2-Tert-butylaminomethyl-2-nitro-1,3-propanediol hydrochloride(3)
To an mixture of con. HCl in water was added 2 (17.46g, 80mmol) and the solution was
heated to 40-60C. Some CH3OH was added and the mixture was stirred at 40-60C for one
hour followed by another 16 hours stirring at ambient temperature. The solvent was removed
in vacuum,. The residue was filtered, washed with cold iso-propanol and dried to give (3) as
white powder. mp.177-178C.
2.3.3. 1-Tert-butyl-3-hydroxymethyl-3-nitroazetidine hydrochloride(4)
To a stirred mixture of 3 (12.14g,50mmol) and diisopropylozodicarboxylate
(16.68g,75mmol) in butone at 50-65C was added dropwise a solution of triphenylphoshine
(19.66g,75mmol) in butane. The solution was stirred for 5 hours. Precipitate was collected
via filtration , washing with butone and drying to give 4 (8.73g , 77.7%) as white powder ,
m.p. 169.2-171.0C.
2.3.4. 1-Tert-butyl-3,3-dinitroazetidine (5)
To a solution of 4 (6.74g , 30mol) in water was added aqueous NaOH. It was stirred for
3 hours at ambient temperature. Then solution was cooled down and a solution of sodium
nitrite(33.46g, 120mmol), potassium ferricyanide(0.99g, 3mmol) and sodium persulfate
(12.64g, 40mmol) in water slowly added to it. Then It was warmed to ambient temperature
and stirred for 1 hour, extracted with dichloromethane. The organic phase was dried
(MgSO4) and removed the solvent under vacuum to give 5 (4.98g, 81.8%) as a yellow oil,
m.p: 16C.
545
2.3.5. N-acetyl-3.3-dinitroazetide(1)
To a stirred solution of 5(2.03g, 10mmol) in acetic anhydride was added BF3.OEt2
dropwise. The solution was heated to 120-124C and stirred for 13 hours at nitrogen
atmosphere. A dark colored solid was obtained via vacuum distillation, cooled in ice-water,
the solid was recrystallized from methylene chloride/hexanes to give 1(1.60g, 84.6%) as
white crystal, m.p:113.1-114.5C.
3.

3.4
Structure analysis
Fig 1.
FTIR-spectrum
Fig 2.
H-NMR spectrum
The IR-spectrum of ADNAZ was shown in Fig1. The extending frequency decreased to
O
||
C

-1
1669cm because the uncommon electron pair of nitrogen atom conjugated with
in
amino-group. The extending vibration absorbance peak of C-H, NO2 is 3000cm-1, 1582cm-1,
and 1434cm-1 respectively. `H-NMR spectrum of ADNAZ was shown in Fig2. Because of
O
||
C
the effect
of (keto-group), the resonance peak2 of hydrogen atom in CH2- are
seperated with =4.79ppm and =4.95ppm, while the resonance of CH3-is 1.99ppm,
=7.25ppm is the resonance of the solvent CDCl3.
3.5
Factors affecting the yield of intermediate (2)
The first step reaction is a Mannich condensation reaction. It is discovered the amount of
catalyst (NaOH) has a great effect on the yield of intermediate (2). When the usage of NaOH
is under 3%(wt), the yield of 2 can be very high (94.8%), but when it is above 5% (wt), the
yield of 2 is only 8.5%. This is because nitromethane is very sensitive to alkali. In alkali
solution, it will proceed the following reaction:
O
CH =N
+ CH3NO2
2
OH
CH2
N
CH2 NO2
OH
OH
CH CH2NO2
2
N= O
CH2NO2
CH
N
OH
(1)
H O O C CH2NO2
(2)
546
3.6
Yield of intermediate (4)
The third step rcaction is a mitsunobu reaction. There are many factos such as mole ratio
of materials, reaction temperature and the type and amount of solvent etc, to affect the yield
of intermediate (4). It is found that the yield of 4 is higher with the following conditions:
ratio of intermedinte (3) /DIAD/ph3p 1/1.2-2/1.2-2, temperature 50~65C, the reaction time
3~6 hours, solvent- butone.
3.7
Factors affecting overall yield of ADNAZ
Apart from the above factors, there are some other factors affecting overall yield of
ADNAZ. The second step is a conventional reaction in which the aldehyde can be prepared
from 1,3-oxazine, so intermediate (2) decomposes to formaldehyde and quaternary
ammonium salt in acid solution. The yield of ADNAZ in nitrogen atmosphere is higher
about 5% than that in air atmosphere in the fifth step reaction.
4.
CONCLUSION
ADNAZ was synthesized by a five-steps reaction with overall yield of 42.8%. The
factors affecting yield were analysized. During the preparation, water was used as the
reaction media (butanone used in the third step and acetic anhydride used in the fifth step).
This decreased not only the production cost but also the pollution to environment
remarkably. Structures of intermediates and ADNAZ were confirmed by IR, 1HNMR,
13
CNMR and MS.
REFERENCES
[1]
[2]
[3]
BOZHOU WANG, CHUNHUA ZHU, QIUHE LONG, Hanneng Cailiao (Energetic Materials), 1995.
S. DANIEL, R. D. PARITOSH: ADNAZ, Composition and Processes, US5808099, 1998.
M. A. HISKEY, M. D. COBURN: Synthesis of 1,3,3-Trinitroazetidine, US5336784, 1994.
547
INFLUENCE OF THE SOLID PROPELLANT GRAINS PROCESSING

ON BURNING RATE OF DOUBLE BASE ROCKET PROPELLANTS
Berko Zecevic, Jasmin Terzic and Mario Bakarad
Abstract
Besides a basic value of burning rate measured in standard test rocket motors, actual
burning rate within a real rocket motor includes other influences as rocket motor
dimensions, combustion gases flow within the chamber, chamber pressure,
environmental temperature, angular acceleration, processing technology etc. The basic
burning rate of solid rocket propellants is measured by standard ballistic test motor and
is expressed by Saint Roberts law in form of r = a p n . Actual burning rate can be
greater or less than the basic burning rate. Determination of different impacts on the
actual burning rate is a complex task, which includes introduction of several
assumptions in order to estimate an influence of each factor on the total burning rate.
Impacts of geometric shape of propellant grains (star grain and hollow cylindrical grain
with interior burning surface) and grain processing technology on variance of the
basic burning law have been considered. Propellant grains, which were used in this
research, had been manufactured by pressing and extrusion.
Significant deviation between burning rate measured in standard ballistic test rocket
motors and the actual burning rate within the real rocket motors has been observed.
Keyword:
1.
rocket motors, double base propellant, burning rat, HUMP effect,

pressing, extrusion manufacturing, radial acceleration
BURNING RATE
Propellant burning rate is mostly influenced by the combustion chamber pressure and is
expressed by Saint Robert's or Vielle's law within a limited pressure range:
r = a pn
(1)
The pressure exponent n and the burn rate coefficient a are dependent on chemical
composition of a solid propellant and initial temperature of the propellant grain. These
coefficients are usually determined by ballistic evaluation motors [1,5,6,11,13].
Applied shapes of solid propellant grains for standard ballistic evaluation motors should
ensure a low flow velocity over the burning surface, or mass flux of combustion products
through the internal flow channel. The pressure exponent n should be independent on the
combustion chamber pressure at a defined pressure range, and should be valid for a defined
initial grain temperature.
Burning rate measured by ballistic evaluation motors must be corrected for real rocket
motors, which depends on rocket motor size and conditions of its application. In order to
obtain real values of burning rates within a rocket motor, previous measured values should
be fitted for a real rocket motor. Typical fitting coefficient of burning rates, which is
applicable to real rocket motors, lies between values of 1.01 to 1.05 [1].
548
For typical rocket propellants, at a pressure range of 3 to 15 MPa, pressure exponent is

usually from 0.2 to 0.7 [2]. At double base rocket propellants, which contain burning rate
catalysts, the pressure exponent is an indicator for catalytic efficiency to be obtained
plateau and mesa effects (Fig. 1). If the pressure exponent approximates zero, plateau
effect appears, and for a negative one, mesa effect. The pressure exponent n at extreme
high burning rates, plateau or mesa effects, depends on physical and chemical properties
of lead catalysts and their particle properties, as well as chemical composition of a double
base propellant.
Fig 1.
Dependence of burning rate as a function of combustion chamber

pressure [5,6]
Actual burning rate within real rocket motors is under other influences and because of
that the burning rate is one of ballistic properties, which is determined with difficulty.
An actual burning rate in a rocket motor, except the basic value measured in standard
ballistic evaluation motors, consists of several components. Determination of these
components is a very complex task so many assumptions must be included, in order to be
estimated their influence on the total actual burning rate.
Actual burning rate can be expressed as:
ri = a pcn ri
(2)
where: ri i-th component of influence on a basic burning rate, a and n burn rate coefficient
and pressure exponent measured in ballistic evaluation motors.
From the equation of mass conservation, which establishes balance between mass
generation rate generated from the propellant burning and the sum of the gaseous mass rates
& n , actual
accumulated in the chamber dM/dt and exhausted gases rate through the nozzle m
[13,14,15]
burning rate can be determined
:
r=
1
p Ab
p
dp
p A
dV
V
c
+
c + c th
c*
R Tc dt R Tc dt
(3)
where: p solid propellant mass density [kg/m3]; Ab grain burning area [m2]; r solid
propellant burning rate [m/s], g instantaneous gas density in the chamber; V free volume
within the chamber [m3]; pc chamber pressure [Pa]; R combustion gas constant; Tc
combustion chamber temperature [K], Ath cross-sectional throat area [m2] and c*
characteristic velocity.
Estimation of variation of the basic burning rate due to influence of several factors can
be made by appropriate separation each of influence components. Research of the influence
of gaseous mass flux on the basic burning rate, which were performed by many authors [9,1],
549
has shown that combustion products flow over the burning surface affects the effect of
erosive burning.
When a solid propellant burning takes place under conditions of high radial acceleration,
variation of the basic burning rate rises more intensive [13]. The newest research shows that
the solid propellant grain processing affects the variation of basic burning rate [10], which has
been confirmed through authors research work [13].
1.1
Influence of the solid propellant grain processing
Influence of the solid propellant grain processing, which is called effect the HUMP [4],
consider a radial variation of burning rate as a function of the web burned. This effect has
been noted when chamber pressure-time and thrust-time curves, for simple standardized
propellant grains, were analyzed.
The greatest influence of the HUMP effect appears at the half of the web thickness, and
calculations have shown the burning rates were increased in 3 to 7% [2], compared to the
basic burning rate. Analysis of burning rates, obtained as a consequence of the HUMP effect,
has shown that the burning rate varied with web burned. This influence is caused by
rheology of the solid propellant grain processing, or actually by propellant mass flow during
the pressing determined by pressing tool. The HUMP effect can be taken into consideration
for internal ballistic prediction, if an empirical dependence of the burning rate of the web
burned being established. Such dependence can be established for cylindrical and star grains,
while a finocyl grain (with burning direction in different directions) this dependence has a
sinusoidal character [5,6].
According to [7], participation of the HUMP effect in variation of the burning rate is about
5% and it can be reflected on the burning rate increase or decrease. Fig. 2 shows the HUMP
factor as a function of the relative web ratio burned for different rocket motors.
Fig 2.
1.2
HUMP factor as a function of the relative web for different rocket

motors[10]
Influence of solid propellant grain processing method
A standardized production method for double base propellant grains is discontinuous

process (pressing technology). High rate serial production of limited size solid propellant
grains requires a different way of solid propellant processing continuous process
(extrusion). Differences between these two processing methods, lies in way of moisture
550
separation from the strong mixture, gelatinization procedure and forming processing of
solid propellant grains. As a consequence of these technological procedures, percent
ingredients of some additives and ballistic catalysts in the tough mixture are changed.
These ingredients variations should not affect the character of the burning process essentially
(curve pressure-time should be approximate same).
2.
EXPERIMENTAL RESEARCH
Basic burning rate law for solid propellants manufactured by the discontinuous process
has been determined for three initial grain temperatures 243 K, 293 K and 323 K. The
burning rate law was determined by means of the standard ballistic evaluation motor 32/16
(cylindrical solid propellant grain with external diameter of 32 mm, internal diameter of 15.9
mm and length of 125 mm). The solid propellant grain ensures a neutral burning character.
Each point of the burning rate plot corresponds to the test, which is determined by the nozzle
throat diameter of the test motor. Burning rate laws for double base propellant NGR-A, at
three different grain temperatures, are shown in Fig. 3.
0,030
0,028
323 K
Burning rate [m/s]
0,026
0,024
293 K
0,022
243 K
0,020
0,018
0,016
0,014
0,012
0,010
6
Fig 3.
2.1
10
12
14
16
Pressure [MPa]
18
20
22
Kinetic characteristics of the solid propellant NGR-A
Research of the Hump effect influence
Experimental research with two types of double base solid propellant rocket motor was
performed. Fourteen rocket motors of 57 mm (RM-3) diameter with cylindrical (CP) grain
and rocket motors with 128 mm diameter with star grain and two types of nozzles (two
motors with a central nozzle RM-2 and five motors with multiple peripheral nozzles RM-1).
The basic burning rate laws for both DB solid propellants were measured in the ballistic
evaluation motor 32/16 previously.
During testing of the rocket motors RM-2 with the central nozzle and star grains (Fig.4),
combustion chamber pressure were measured at both ends of the combustion chamber
(Fig.5).
551
Fig 4.
Test rocket motor RM-2
The pressure-time curve measured by piezo-transducer system, and another one, which
was determined by means of the computer program SPPMEF [12], deviate from each other.
This deviation was caused by variation of the burning rate measured in standardized ballistic
evaluation motors from the actual burning rate within a real rocket motor. Burning rate in a
real rocket motor comprises a component of the HUMP effect.
16
Pressure [MPa]
14
12
10
8
Measuring point - bottom
Measuring point - nozzle inlet
Prediction
2
0
0
0,2
0,4
0,6
0,8
1,2
1,4
1,6
Time [s]
Fig 5.
Pressure-time curve for the test rocket motor RM-2 [12]
Variation of the basic burning rate, caused by HUMP effect, is defined as a ratio of
actual burning rate (within a real rocket motor) to the basic burning rate (measured by
standardized ballistic evaluation rocket motor) under same chamber pressures and the
definite web thickness burned (w). Influence of gaseous mass flux to the actual burning rate
is neglected [12]:
HUMP (w) = ri ( pc , w) r0 ( pc )
(4)
Actual burning rate is determined by the relation (3), and the basic burning rate law was
determined by tests with standardized ballistic evaluation rocket motors (Fig. 3).
Component of the burning rate caused by the HUMP effect for the rocket motor RM-2,
was determined by analysis of the pressure-time curve, and it is shown in Fig. 6 as a function
of relative web thickness.
552
1,10
HUMP factor
1,00
0,90
0,80
0,70
0
Fig 6.
0,2
0,4
0,6
Web fraction
0,8
Influence of the HUMP effect for the rocket motor RM-2 [12]
Basic burning rate was corrected by HUMP effect influence, and the pressure-time curve
was predicted by means of the computer program SPPMEF. The prediction has shown a
good agreement with test results (Fig. 7).
14
Pressure [MPa]
12
10
8
Measured
Predicted
4
2
0
0
0,2
Fig 7.
0,4
0,6
0,8
Time [s]
1,2
1,4
1,6
Pressure-time curve for the motor RM-2, after correction due to HUMP
effect [12]
Rocket motors RM-1 with multiple peripheral nozzles were tested in order to be
evaluated an influence of nozzle configuration on the actual burning rate. Character of
variation of the component due to HUMP effect, as a function of web thickness, is shown in
Fig. 8.
Variation of the component due to HUMP effect for rocket motors RM-3 (57 mm) is
also shown in Fig. 8, where a significant variance can be noted. This fact points to the
assumption, in which mass flux influence was neglected, is not correct. Results of
calculations of the mass flux variation within the rocket motors 57 mm and RM-2, obtained
from the computer program SPPMEF, is shown in Fig. 9. It can be seen that the mass flux in
the rocket motor 57mm is considerable greater, and its influence on the basic burning rate is
essential.
553
1,3
HUMP factor
1,2
1,1
1,0
0,9
RM-1
RM-2
0,8
RM-3
0,7
0
0,2
Fig 8.
0,4
0,6
Web fraction
0,8
Variation of the basic burning rate under influence of HUMP effect as a

function of web thickness for rocket motors 57 mm, RM-1 and RM-2 (with
central and multiple peripheral nozzles)
6000
Mass flux, [kg/s/m2]
5000
RM-2
RM-3
4000
3000
2000
1000
0
0
0,2
Fig 9.
0,4
0,6
Web fraction
0,8
Variation of the mass flux as a function of the web thickness for rocket
motors 57 mm and RM-2
Character of variation and values of the burning rate component caused by HUMP effect
is agreed with results obtained from other authors [5,6,10], and so that method, which was used
in this paper, can be considered as correct. However, results obtained for 57 mm rocket
motors, has shown a considerable influence of mass flux, so it should be careful when an
assumption of its neglect being introduced.
2.2
Research of an influence of the solid propellant processing method
In order to raise the productivity, Vitezit- company (Bosnia and Herzegovina)

introduced a continuous method of DB rocket propellant manufacturing. This production
procedure comprises a process of gelatinization nitrocellulose by nitroglycerin and addition
ballistic catalysts [3, 8] during the production sequences. Gelatinized mixture, which is used in
this manufacturing process, has viscosity of 8-15 mPas, while discontinuous process requires
(pressing) requires a mixture with greater viscosities of 25-35 mPas.
554
Ballistic evaluation rocket motors 32/16 with solid propellant grains manufactured by
continuous and discontinuous process were tested at initial grain temperatures 293 K, 243 K
and 323 K (Fig. 10. to Fig. 12.).
24
24
Pressing
21
18
18
Pressure, MPa
Pressure, MPa
Extrusion
21
15
12
9
15
12
9
0
0
0,05
0,1
0,15
0,2
0,25
0,05
0,1
0,15
0,2
0,25
Time, s
Time, s
Fig 10. Standardized ballistic evaluation motor 32/16 at temperature of 293 K,

propellant NGR-A manufactured by pressing and manufactured by
extrusion [13]
24
24
Pressing
21
18
Pressure, MPa
Pressure, MPa
18
15
12
9
15
12
9
0
0,00
Extrusion
21
0
0,05
0,10
0,15
Time, s
0,20
0,25
0,05
0,1
0,15
0,2
0,25
Time, s

extrusion [13]
555
24
24
Pressing
21
18
Pressure, MPa
18
Pressure, MPa
Extrusion
21
15
12
9
15
12
9
0
0,00
0
0,05
0,10
0,15
0,20
0,25
0,05
Time, s
0,1
0,15
0,2
0,25
Time, s

extrusion [13]
The basic burning rate for DB propellants manufactured by extrusion (Fig. 13) was
determined on the basis of plots shown in Fig. 10 to Fig. 12.
0,028
323 K
Burning rate [m/s]
0,026
293 K
0,024
0,022
243 K
0,020
0,018
0,016
6
10
12
14
16
Pressure [MPa]
18
20
22
24
Fig 13. Kinetic characteristics of the propellant NGR-A (manufactured by

continuous process )
Unstable burning of solid propellant grains manufactured by continuous process, at low
combustion chamber pressures, was observed. It was particularly distinguished at
temperature of 323 K. Character of the basic burning rate law, for all three initial test
temperatures (Fig. 13), is significantly different than in case of the rocket propellant
manufactured by discontinuous process (Fig. 3). At initial temperatures of 323 K and 243K,
the mesa effect appeared, while plateau effect disappeared. It was also observed that the
start point of the plateau effect moved in 0,5 MPa to the left direction, at initial grain
temperature of 293 K [13].
556
Real rocket motors with DB rocket propellants NGR-A, manufactured by continuous

process, were tested at initial grain temperatures of 243 K and 323 K in order to research an
influence of manufacturing process variation, composition and conditions of high radial
acceleration, on the burning process (Fig. 14).
Fig 14. Test rocket motor RM-1

Unstable burning was observed in standardized ballistic evaluation motor 32/16 at initial
temperature of 323 K and at combustion chamber pressure less than 15 MPa (which
determines a zone of super burning rate). There was no unstable burning when real rocket
motors were tested, at same initial temperature and under variable radial acceleration (Fig.
15).
21
3200
18
2800
Pressure, MPa
2000
12
1600
9
1200
6
Spin rate, rad/s
2400
15
800
400
0
0
0,2
0,4
0,6
0,8
1,2
1,4
Time, s
Fig 15. Influence of variation of solid propellant processing on the curve pressuretime form for DB propellant NGR-A tested at temperature of 323 K [13]
A distinguished unstable burning was observed when real rocket motors were tested at
initial temperature of 243 K (Fig. 16). It was more intensive while rocket motors were
approaching to maximum radial acceleration.
557
21
18
18
15
15
12
9
6
3
3200
With radial acceleration
2800
2400
2000
12
1600
9
1200
800
400
0,5
1
Time, s
1,5
Spin rate [rad/s]
No radial acceleration
Pressure, MPa
Pressure, MPa
21
0
0
0,5
1,5
Time, s
Fig 16. Influence of variation of solid propellant processing on the curve pressuretime form for DB propellant NGR-A tested at temperature of 243K,
without and with radial acceleration [13]
Pressure-time curves of two groups real rocket motors, which were exposed to influence
of variable radial acceleration, are shown in Fig. 15 and Fig. 16. Continuous line refers to
solid propellant grain, manufactured by discontinuous process (pressing), while other dashed
or dash dot lines refer to solid propellant grains manufactured by continuous process
(extrusion).
Unstable burning, observed at extruded DB propellants in standardized ballistic
evaluation motors 32/16, becomes more distinguished at real rocket motors, which are
exposed to variable radial acceleration.
3.
CONCLUSION
The presented method enables determination of the HUMP effect influence on the basic
burning rate.
Accuracy of the method depends on mass flux intensity, which flows over the burning
surface. For extreme intensities of mass flux, this method is not applicable.
Measured components of burning rates caused by HUMP effect, obtained for different
types of rocket motors during the presented research, are agreed with results obtained from
other authors.
Comparative analysis of the pressure-time plots points to conclusion that variation of
solid propellant processing has an influence on DB propellant burning process, but also that
burning behavior depends on radial acceleration.
The most distinguished unstable burnings were observed for propellant grains tested in
standardized ballistic evaluation motors 32/16 at initial grain temperature of 323 K.
However, the most intensive unstable burning at real rocket motors, which were exposed to
variable radial acceleration, appeared at initial grain temperature of 243 K.
Burning instability caused by variation of solid propellant grain processing,
corresponding percentage ingredients and types of additives has shown that burning
mechanism had been changed.
558
Dispersion of ballistic additives within the basic DB propellant structure, manufactured

by extrusion, is different than dispersion from structure obtained by pressing. This affects
catalytic effects of different additives, which cause the mesa effect in the basic burning
rate law of the extruded solid propellant NGR-A.
REFERENCES:
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
ANON: Solid Rocket Motor Performance Analysis and Prediction, NASA Report SP 8039,
May 1971, N72-18785.
DAVENAS A.: Solid Rocket Motor Design, Tactical Missile propulsion, Edited by Jensen E. G.
and Netzer W.D., Progress in Astronautics and Aeronautics, Vol. 170, AIAA, 1996.
FILIPOVIC M.: Influence of poly-dispersion of nitrocellulose on the manufacturing process and
characteristics of DB propellant, Military Technical Institute KoV JNA, Belgrade, 1991.
FRIEDLANDER M AND JORDAN F.V.: Radial Variation of Burning Rate in Center Perforated
Grains, AAIA Paper 84-1442, July 1984.
FRY R.S.: Solid propellant subscale burning rate analysis methods for U.S. and selected NATO
facilities, The Johns Hopkins University, Chemical propulsion information agency, Columbia,
Maryland, Januar 2002
FRY R.S.: Solid propellant test motor scaling, The Johns Hopkins University, Chemical
propulsion information agency, Columbia, Maryland, Septembar 2001
GOSSANT B.: Combustin des propergols solides et balistique interieure des propulseurs,
Davenas A. et collabrateurs : Technologie des propergols solides, Masson,Paris, 1989.
KRIVIC K.: Variations of DB propellant processing, Private correspondention, 1996-1997.
RAZDAN K.M. AND KUO K.K.: Erosive Burning of Solid Propellants, Fundamentals of SolidPropellant Combustion, Edited by Kuo K.K. and Summerfield, Progress in Astronautics and
Aeronautics, Vol. 90,AIAA, 1984.
SCIPPA S.: Propellant grain design, AGARD-LS-150, Design Method in Solid Rocket Motors,
Paris, 1988
SUTTON P. G.: Rocket Propulsion Elements - An Introduction to the Engineering of Rockets John Wiley&Sons, Fith Edition,1986, New York.
TERZIC J.: Prediction of idealized internal ballistic properties of a rocket motor with DB solid
propellant, Faculty of Mechanical Engineering, Sarajevo 2002.
ZECEVIC B.: Influence of the variable radial acceleration to internal ballistics of rocket motors
with DB propellants, Disertacion, Faculty of Mechanical Engineering, Sarajevo 1999.
ZEEVI B., TERZIC J.: Estimating current burning rate in standard ballistic rocket motor, 2nd
International Symposium, Revitalization and Modernization of Production RIM 99, ISBN
9958-624-06-0, Biha, October 1999.
ZECEVIC B., TERZIC J. AND SERDAREVIC-KADIC S.: Effects of kinematics and mass
characteristics of gas flow to basic rate burning of double-base rocket propellants, 3rd
International Symposium, Revitalization and Modernization of Production RIM 2001, ISBN
9958-624-10-9, Biha, September 2001
559
THE INFLUENCE OF MEASUREMENT METHODE ON THE

INITIATION TIME MEASUREMENT NECESSARY FOR PRIMER
MIXTURE INITIATION OF ELECTRICAL DETONATOR
Igor Zori, Dalibor Kuhinek and Mario Dobrilovi
University of Zagreb, Faculty of Mining, Geology and Petroleum Engineering,
Pierottijeva 6, 10000 Zagreb, CROATIA
Abstract:
This paper considers the differences in measurement of time necessary for primer
mixture initiation of electrical detonator when measured with two different methods.
First method is electrooptical where time measurement is electrically triggered but
stopped by light of primer mixture flame. Second method is electrical and measures time
of current flow through bridge wire. There was significant difference in attained results.
Reasons, as well as the influence on application are discussed in this paper.
Keyword:
1.
primer mixture, initiation time measurement, electric detonator
INTRODUCTION
Electric detonators have two big advantages over other systems of explosives
initiation: first being the possibility of explosive initiation from a fairly distant and sheltered
place, and the second one is a complete control of the moment of explosive initiation.
Due to the above, the electric detonators are used to activate nonel systems, which initiate
explosive. Either direct current, or electric current pulse is used for electric detonators firing.
Current ignited electrical detonator is a hermetically sealed capsule (made of aluminum
or copper) with two line cords (leading wires), containing a resistive bridge wire surrounded
with primer mixture (ignition head), which is in contact with delay element (if any) or
primary charge of an initiating explosive used to initiate secondary charge of a high brisance
explosive, relatively insensitive to heat and shock, as can be seen in Fig.1.
Fig 1.
Elements of electrical detonator (with delay): 1- leading wires; 2- sealing

plug; 3- ignition head; 4- delay element; 5- primary charge; 6- secondary
charge; 7- capsule [5,6]
560
Current from an external source is applied via leading wires ends to a resistive bridge
wire, heats the bridge wire, which in turn heats the primer mixture, causing it to ignite.
If there is no delay element, then we talk about instantaneously fired detonator; if the
delay element is present, firing time may be from several milliseconds to several seconds
after primer mixture is initiated.
Primer mixture, besides heat of bridge wire, can be activated by spark, flame, impact,
friction or other primary heat sources of appropriate magnitude.
It is considered that the time of firing is the time elapsed from the moment of electrical
circuit closing to the moment of detonator firing. That very time, in spite of being short, is
found at instantaneously fired detonators too. Following the standards, the delay time of the
instantaneously fired detonators must be shorter than 10 ms for standard detonators and
shorter than 1ms for seismic detonators. The total delay consists of ignition head firing
delay, primary charge firing delay and the delay of secondary charge firing. This paper
considers the measurement of delay time of ignition head firing; which is the only part of the
whole detonation process depending on the electric current.
2.
THEORY
2.1
Electric detonator ignition head
Electrical detonators ignition head usually uses fuse head, flat filament or thin-film
technology to produce temperature necessary to initiate primer mixture. Schematized crosssection on Fig 2. shows a fuse head version of a detonator ignition head.
Fig 2.
Ignition head of electrical detonator: 1- leading wires; 2- isolator;

3- leading wires ends; 4- resistive bridge wire; 5- primer mixture [2,4]
Primer mixture has two layers. The first layer of an inflammable material ensures
uniformity of ignition, which is very important for serial blasting circuits. The second layer
of an inflammable material gives heat impulse for ignition of delay element or primary
charge. [6] Delay in head ignition is due to time necessary for temperature rise to the level
appropriate to ignite primer mixture, and that is depending on energy feed into the bridge
wire. [3] In order to achieve the relevance of the delay measurement it is necessary to ensure
constant current feed into the bridge wire.
The measurement of the initiation time must start at the same moment when current
starts to flow through the bridge wire. The measurement of initiation time must stop at the
same moment when primer mixture is fired. As a moment of firing we can consider primer
mixture light of flame occurring or stopping of current flow caused by burning out the bridge
wire. Therefore we can use two methods to measure delay time.
561
2.2
Electro-optical method of primer mixture initiation time

measurement
The first method used for measurement is electro-optical. Time measurement is

electrically triggered and stopped by light of primer mixture flame. Principles of
measurement can be seen in Fig. 3. When current starts to flow by triggering the thyristor,
current rise is converted to optical signal trigger and is sent through one fiber to time counter
as Start counting. A second fiber is physically attached on the outer side of ignition head
and when primer mixture is ignited, light of flame of ignition sends optical Stop counting
signal to time counter. DC source must ensure constant current feed through bridge wire. [1]
Thyristor
Electrical
trigger
to optical
Fire
Fiber
Start
DC
Source
Stop
RDET
Counter
Fiber
Ignition head
Fig 3.
2.3
Electro-optical method of time measurement
Electrical method of primer mixture initiation time measurement
In the second method time of current flow through bridge wire is measured.
Measurement can be done by monitoring voltage drop on an appropriate resistor of known
value caused by current flow. If the resistor is connected in series with resistive bridge wire,
as shown in Fig. 4., the voltage drop is proportional to current strength. For these
measurements oscilloscope with memory can be used.
Thyristor
Fire
DC
Source
RDET
Ignition head
RSHUNT
Oscilloscope
Fig 4.
Electrical method of time measurement
562
If we use oscilloscope with memory in single sweep mode with auto trigger, then
triggering of current flow will trigger oscilloscope. [8] For triggering the current flow, the
same as in previous method, it is recommended to use thyristor. In that way we can
overcome the problem of switch contact resistance instability, which may occur during first
few milliseconds after switch is closed. This is important because the expected measured
times are few milliseconds. Using time delay (prehistory) it is possible to monitor current
flow even before it started. Using time and voltage markers, precise measurement of time
necessary for bridge wire burn out the same as current strength can be done. Used DC source
must ensure constant current feed through bridge wire. Idealized view of current flow can be
seen in Fig. 5.
Current
I
Safety
current
of firing
tSTART
Fig 5.
3.
Time t
tSTOP
Current flow through bridge wire
EXPERIMENTS
In order to compare the measured results of the above described measurement methods,
the measurement using both methods simultaneously has been carried out. The measurement
scheme is given in Fig. 6. It is shown that the time measurement of both measurement
methods is initiated at the same moment by thyristor triggering, whereas the end of time
measurement is independent for each measurement method.
Thyristor
Electrical
trigger
to optical
Fire
DC
Source
Fiber
Start
Stop
RDET
RSHUNT
Counter
Fiber
Ignition head
Oscilloscope
Fig 6.
Simultaneously measurement by electro-optical and electrical method
563
Constant current of desired value, as given in the specifications for ignition heads, is
insured by adjusting total resistance of electric circuit loop and stabilized DC source with
low internal resistance.
Measurement was carried out on three bridge wire sets, the characteristics of which are given
in Table 1. The instrument EXPLOMET-FO-MULTIMODE was used for measurement the
initiation time by electro-optical method. The digital Hewlett Packard HP 54501A
oscilloscope was used for measurement by electrical method. To adjust total resistance of
circuit on appropriate value necessary to achieve wanted value of current through bridge
wire milliohm meter Hameg HM 8014 was used. [7]
Table 1. Types of bridge wires used for measurement:
Ignition head for detonator type
Specifications
Pobeda Gorade: EU-S1

- instantaneous electric detonators
- Blasting cap No 8 (aluminum)
Leading wire: 2x0,5 m

Bridge wire resistance: 1,4 2,0 Ohm
Maximum resistance: 1,3 Ohm
Safety current of firing for t<10ms, I1002A
Do not fire with 0,18 A
Pobeda Gorade: EZG-S2

- instantaneous electric detonators

Bridge wire resistance: 0,6 0,8 Ohm
Safety current of firing for t<10ms, I1002A
Pobeda Gorade: EU-VN (Type C)

- 30 millisecond high nonsensitive
electric detonators

Bridge wire resistance: ~0,07 Ohm
Safety current of firing I 100 min. 25 A
The electric current of 2A was applied for measurement the bridge wire initiation time
for the detonators EU-S1 and EZG-S2; and the current of 25A was selected for measurement
the bridge wire initiation time for the detonators EU-VN (Type C), which together meet the
requirements of sure initiation of one hundred serially connected detonators.
The results of the measurements are given in Table 2. The table shows the value of total
resistance RTOT which is the sum (total) of bridge wire resistance and the leading wires
resistance plus the times of initiation obtained through both measurement methods: electrooptical and electrical method. Average values, standard deviation of the measured resistance
value and the initiation time are given in Table 2.
Figure 7 shows average values and measured values of the ignition head initiation, and
Figure 8 shows a typical oscillogram attained while using the electrical method for
measurement the ignition head for detonator type EU-S1. Figure 9 gives average values and
measured values of ignition head initiation, while Figure 10 shows a typical oscillogram
attained at measurement when using electrical method for ignition head for detonator type
EZG-S2. Figure 11 gives the average value and measured average values of ignition head
initiation while using electrical method. Figure 12 gives the average value and measured
average values of ignition head initiation while using electro-optical method. Figure 13 is a
typical oscillogram attained at measurement when using electrical method for ignition head
for detonator type EU-VN (Type C).
564
Table 2. Results of measurement

bridge wire for detonator type
EU-S1
EZG-S2
EU-VN (Type C)
RTOT toptical telectrical RTOT toptical telectrical
RTOT
()
(ms)
(ms)
()
(ms)
(m)
(ms)
(ms)
2,140
1,0378
2,0202
1,10
2,7487
10,4800
89
1,8219
1,849
2,180
0,8999
2,1016
1,09
2,7772
10,1680
93
2,5374
1,762
1,960
0.8873
2,1016
1,15
2,8300
7,0800
83
1,9237
1,752
2,060
0.9448
2,1867
1,13
2,8665
10,7582
85
1,8626
1,988
2,030
1,0201
2,1534
1,11
2,8335
7,4053
79
2,0195
1,842
2,200
0,9323
1,8611
1,27
2,9021
6,3587
90
1,7028
1,848
2,100
0,9220
1,8426
1,03
2,7940
10,4755
70
2,4407
1,844
2,500
0,9121
1,9166
1,09
3,0951
8,9230
80
1,6211
1,844
2,180
0,8958
2,1386
1,15
2,8648
10,8475
87
2,7545
1,864
Average 2,150 0,9391
2,0359
1,124
2,8569
9,1662
83,1
2,1763
1,8325
St. dev.
0,1315
0,066
0,1014
1,7748
7,2
0,4938
0,0727
sample
0,153
0,0542
(ms)
toptical telectrical
2,3000
2,1867
2,1016 2,1016
2,1000
2,1544
2,1386
2,0359
2,0202
1,9166
1,9000
1,8611 1,8426
time (ms)
1,7000
Time electrical
Average electrical
Time optical
Average optical
1,5000
1,3000
1,1000
1,0378
1,0201
0,8999 0,8873 0,9448
0,9000
0,9391
0,9323 0,9220 0,9121
0,8955
0,7000
0
10
sample
Fig 7.
Initiation time for EU-S1 using the current of 2A
565
Fig 8.
Typical oscillogram for electrical method of time measurement for type

EU-S1 (I=2 A)
11,0000
10,4800
10,1680
10,4755
10,7582
10,8475
10,0000
9,1662
9,0000
8,9230
time (ms)
8,0000
7,4053
7,0800
Time electrical
Average electrical
Time optical
Average optical
7,0000
6,3587
6,0000
5,0000
4,0000
3,0951 2,8648
2,7772 2,8300 2,8665 2,8335 2,9021
3,0000
2,8569
2,7940
2,7487
2,0000
0
10
sample
Fig 9.
Initiation time for EZGS2 using the current of 2A
Fig 10. Typical oscillogram for electrical method of time measurement for type
EZG-S2 (I=2 A)
566
2,9000
2,7545
2,7000
2,5374
2,5000
time (ms)
2,4407
2,3000
Time optical
Average optical
2,1000
2,0760
2,0195
1,9237
1,9000
1,8219
1,8626
1,7000
1,7028
1,6211
1,5000
0
10
sample
Fig 11. Initiation time measured by electro-optical method for UE-VN (Type C)
using the current of 25A
2,9000
2,7545
2,7000
2,5374
2,5000
time (ms)
2,4407
2,3000
Time optical
Average optical
2,1000
2,0760
2,0195
1,9237
1,9000
1,8219
1,8626
1,7000
1,7028
1,6211
1,5000
0
10
sample
Fig 12. Initiation time measured by electrical method for UE-VN (Type C) using
the current of 25A
567
Fig 13. Typical oscillogram for electrical method of time measurement for type
EU-VN (Type C) (I=25 A)
4.
DISCUSSION
The results of initiation time measurements show that the ignition heads for normally
sensitive detonators type EU-S1 have shorter time obtained by electro-optical method than
the time by electrical method, i.e. shorter for 1,0968 ms. Furthermore, measuring initiation
head type EZG-S2, the time obtained by electro-optical method is even shorter than the time
by electrical method, namely for 6,3093 ms. It means that with both types of initiation heads
the primer mixture light appears earlier than the bridge wire is burned out. This phenomenon
is possible provided that only if the higher temperature is needed for the bridge wire burning
out than the temperature at which the primer mixture light appears.
The results of initiation time measurements show that the ignition heads for highly
nonsensitive detonators type EU-VN (Type C) S1 have longer time obtained by electrooptical method than the time by electrical method, i.e. longer for 0,3438 ms. It means that
that the bridge wire is burned down earlier than the primer mixture reaches the temperature
of light. This phenomenon is possible only if the speed of bridge wire temperature rising is
considerably greater than the speed of temperature rising of the outer primer mixture layer.
5.
CONCLUSION
The differences of the obtained measurement results of initiation by electro-optical and

electrical methods depend on the type of detonator as well as the initiation head. Carried out
measurements lead us to a conclusion that the bridge wire selection in a combination with
used primer mixture has a direct impact on difference in initiation times obtained by the
above compared methods. In terms of reliability of initiation of serially connected electrical
detonators, it is more reliable to use normally sensitive detonators than highly non sensitive
detonators. This advantage has been specially evident at EZG-S2 where the current circuit
stays closed three times longer than it is necessary to initiate primer mixture. The results of
the carried out measurements direct to a possibility of comparing the influence of bridge
wire and primer mixture sets construction on the relation of initiation time on the time
necessary for bridge wire burning out (the duration of close current circuit). The longer the
time necessary for bridge wire burning out in relation to primer mixture firing, the higher
the reliability of detonator firing. The relation of times mentioned above could be determined
by simultaneous initiation time measurement using electro-optical and electrical methods.
568
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
HRN. H.D8.111 Odreivanje zakanjenja detonacije (Determination of detonation delay),

Zagreb, 1999.
Z. ESTER: Miniranje I (Mining 1), Zagreb, 2001.
B. T. NEYER: Bridgewire heating, Proceedings of the sixteenth Symposium on Explosives
and Pyrotechnics, Essington, PA, April 1997.
PRAXIS DER ELEKTRISCHEN ZNDUNG, SCHAFFLER & CO. GES.M.B.H.,
Neuauflage, Wien, 1982.
G. B. HEMPHILL, P.E.: Blasting operations, McGraw-Hill Book Company, 1981.
Initiating devices and accessories, Catalog of products, Gorade factory, Sarajevo, 1970.
Hameg Instruments, Manual for Miliohmmeter HM8014, Hameg GmbH, Frankfurt
Hewlett - Packard, Manual for HP 54501A Digitazing Oscilloscope, USA, 1988.
569
SPARK TEST APPARATUS ESZ DTTE

The apparatus for determination of sensitivity of various energetic materials to high-voltage electrostatic discharge
1.5 kV voltage source
Voltage multiplier
Sample testing box
Battery
of HV condensers
Basic technical data
Capacitance of battery installed:

Regulation range of HV:
Overall resistivity of discharge circuit
including serial shunt 0.1 :
Overall inductance of discharge circuit:
Peak current at max. parameters
of device adjusted:
Time to the discharge of C = 350 nF
when serial resistor Rs = 10 k is included:
Power supply:
Mechanical
HV switch
C = 40 pF 350 nF
U = 4 - 10 kV
R = 200 300 m
L = 1.2 1.5 H
Imax = 3 kA
tmax = cca 12 ms
P = 27 VA
Testing and current

measurement assembly
General view on the testing apparatus ESZ

Measurement of HV and current time courses:
Optimally, via dual channel digital memory oscilloscope, whose sampling frequency fs is approximately 1 GS/s and length of the acquired record is
10 000 points.
570
Development and Production of Devices

and Apparatus for Testing of Energetic Materials
OZM Research Ltd.
DTA 551-Ex
DTA is a technique in which a change of a temperature of a sample is observed while
the sample is subjected to a temperature alteration (heating or cooling). The results can be
effectively used for quality and purity analysis. Some compatibility and chemical stability
evaluations are also possible.
DTA551-Ex is a differential thermal analyzer designed
for testing of explosives. It has very robust design and is able
to withstand shocks accompanied with decomposition of up
to 50 mg samples of regular secondary explosives like RDX,
or 10 mg of primary explosive like Lead Azide in repeating
everyday testing. Unlike some other commercially available
devices it has been designed for explosives and can therefore
withstand violent decomposition of much higher amount of
sample. This happens to be a very important feature since
measurements of small amounts are usually not reproducible
even on much more sensitive machines. Problems are
especially apparent when measuring non-homogenous
samples like pyrotechnic mixtures. While everybody else is
looking for higher sensitivity to pick even a very small amounts of heat flux (necessary for
polymers, biological applications or glass scientists) we are developing robust design for
explosives where the amount of heat has never been a problem.
The sample is measured in glass test tube placed in aluminum block of furnace. The
temperature is measured with thermocouples placed inside the sample. A corrosion of
thermocouple is prevented using
DTA Solution of HMX in HNO3
stainless steel sheet and glass
capillary. Higher amounts of
measured samples successfully
eliminate influence of a possible
inhomogeneity.
The
higher
volume of test tubes also enables
measurement of liquids, solutions
and foaming substances.
34
32
30
28
26
24
22
20
T [C]
18
16
14
12
10
8
6
4
The equipment consists of a

heating block with thermocouples,
temperature
controller,
data
acquisition and evaluation system,
software MEAVY and thermogram database of elementary explosives
2
0
-2
-4
-6
20
40
60
80
100
120
140
160
180
200
220
240
260
280
300
320
Temperature [C]
Meavy
Specifications:
Temperature range: 30-550C
Heating rate:
0,01-40C/min
Sample weight:
10 mg - primary explosives
10-500 mg (default 30mg) - other materials including liquids
571
DTA 551-Rez
This is a modification of previously
mentioned device DTA551-Ex developed for
measurements of samples with a mass of several
grams. It is intended for substances that
decompose less violently. It is however possible
to use this device for measurement of propellant
specimens, pyrotechnic mixtures or explosive
materials especially when studying effect of
thermal explosion. It is as well as it little sister
capable of measurements of liquids or foaming
substances. In addition easy to use software
MEAVY makes all of your measurements and
evaluation of results as simple as possible.
STABIL - Modernized Vacuum Stability Tester Evolved Gas Detection Apparatus

STABIL is a controlling, measuring and evaluating system, enabling evaluation of
temperature stability from an increase of preassure of gases evolved from a sample during
long term isothermal heating. It provides an objective results from the entire process of
stability evaluation and may be performed in vacuum, in presence of gasses (inert nitrogen,
oxidative oxygen) or under the air at atmospheric pressure at the beginning of the test. It
permits a completely automated determination of chemical stability of smokeless powders,
explosives, nitrocellulose and a number
of further raw materials, mixtures and
material
substances
which
are
decomposed during a long term
temperature treatment under generated
gas products. Unlike the classical VST it
eliminates
the
influence
of
decomposition products
or mercury
vapours on the operator. System
STABIL comes in several modifications
with differences in measuring units and
software.
STABIL 16-Ex - this model is based on modernized
STABIL system (fulfills requirements of Czech military
standard CSVN 26910). The results are provided as a
continual dependence of volume on time.
STABIL 14- Ex - this model is designed to fulfill
requirements of STANAG 4556, Annex C-3, transducers
method.
STABIL-HP this model is modified for testing under
higher pressures up to 20 Bar.
The equipment consists of main unit for data acquisition
and temperature control, 1 or 2 heating blocks, 1 20 pressure
transducers, barometric transducer, software WINSTAB.
572
Specifications:
Sampling rate: 1 sec-1
Pressure range of transducers: 0 160 kPa
Pressure range of barometer: 80 120 kPa
Accuracy of transducers: 0,25 %
Number of transducers: from 6 up to 20
Number of heating blocks: 1 or 2
Number of holes for tubes: 10 per block
Temperature range: from 30 up to 180 0,2C
Communication via RS232 or Ethernet
Volume [ml/g sample]
1
200
400
600
800
1000
1200
1400
1600
1800
2000
2200
2400
2600
2800
Time [min]
ESZ KTTV Electrostatic Discharge Sensitivity Tester

The device is produced in two
modifications either for small or for large
scale testing depending on the the spark
energy. It fulfils the requirements of
STANAG 4490. For this device is a broad
variety of measuring system configurations,
spark gaps and discharge energies. The
device is of a very compact design. It differs
from other devices of this kind with its
ability to measure very preciously the
ammount of energy used for initiation. So it
is not the overal energy of the spark, but
only the effective part of it used for
initiation. The oscilloscopic evaluation of
voltage and current time dependence assures
that.
VZB 501 - Apparatus for Automatic Determination of the Explosion Temperature

The apparatus VZB 501 is applicable for determination of the explosion temperature in
accordance with different standards SN, STANAG or MIL-STD 650. It is equipped with a
trigger for test tubes for the time-to explosion determination method. Digital temperature
controller allows more comfortable reading of
measured temperatures. Unit for automatic
Test
registering of the explosion had been incorporated.
This means that you do not have to sit in front of
your device during the whole test. All you have to
do is start the test and set alarm clock to remind
you that the test is over. Our system will recognize
the decomposition temperature even if it is not
accompanied with light emission. Data acquisition
and measurement control are performed by ADET
as [s]
c
d
e
f
g
b
c
d
e
f
g
c
d
e
f
g
software which again is ridiculously simple to use.
300
280
260
240
220
Teplota [C]
200
180
160
140
120
100
80
60
20
40
60
80
100 120 140 160
180 200 220 240
260 280 300 320

Teplota 1
340
Teplota 2
Teplota 3
The equipment consists of main unit for data acquisition and temperature control,
heating block, unit for the automatic registration of samples explosion and ADET software.
573
DMA DX04T Dynamic Thermomechanical Spectrometer

Dynamic mechanical analysis (DMA) or Dynamic mechanical thermal analysis (DMTA)
measures mechanical properties, such as modulus (elasticity) and viscosity (damping) as a
function of temperature, time, frequency, stress or combinations of these parameters. DMA
characterizes the analyzed material by applying oscillating (sinusoidal or other waveform)
stress (strain) and measuring the resulting strain (stress), which is developed in the sample.
The phase difference and the amplitude
of the stress and strain waves are used
to determine a fundamental material
parameters as a storage and loss
modulus, tan d, complex and dynamic
viscosity or storage and loss
compliance. DMA can determine
characteristics such as glass transition
and softening temperatures, heat
resistance, rate and degree of cure,
damping intensity, stress relaxation,
creep, gel point, degree of crystallinity
and orientation of various materials. It
is used for testing of rocket propellants
and fulfils the requirements of
STANAG 4540 and 4525.
CRP03 Creep Apparatus

The creep apparatus employs lever action to apply a dead-weight load to a specimen. A
linear variable displacement transducer (LVDT) is used to measure the deformation.
Measured values are acquired from up to 8 devices with use of one computer. Heating is
present to measure at higher temperatures.
OZM Research Ltd.

Ndran 266
538 62 Hrochv Tnec
Czech Republic
Phone: +420 608 742 777
Telefax: +420 469 692 882
E-mail: info@ozm.cz
http://www.ozm.cz
We also provide some other

traditional devices for stability and
performance testing of energetic materials
such as: Bergmann Junk test, Abel test,
Heat storage test 100 C, Holland test,
Ballistic mortar, Slow cook-off test,
Heating blocks for accelerated aging of
explosives etc.
574

University of Pardubice - 6th International Seminar NTREM (2003)

Enviado por

Dados do documento

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

University of Pardubice - 6th International Seminar NTREM (2003)

Enviado por

Direitos autorais:

Formatos disponíveis

Faculty of Chemical Technology

NEW TRENDS IN RESEARCH

Pardubice, Czech Republic

NEW TRENDS IN RESEARCH OF ENERGETIC MATERIALS

held at the University of Pardubice,

intended as a meeting of students, postgraduate students, university

March 21th, 2003

Edition 1st Limited, 50 pcs OriginalCopLHV, 8QSrotected CD-R

University of Pardubice, 2003

Seminar is supported by:

Chairman of the Seminar:

(DSTL, Sevenoaks, U.K.)

Zdenk Jalov, Ph.D.

INFORMATION ABOUT THE STUDY PROGRAM

The Faculty of Chemical Technology is accredited for providing 7 MS and Bc study

Accommodation of foreign students and doctorands can be provided by the

Study program Chemistry and Technological Chemistry

GRADUATES PROFILE AND POSSIBILITIES

Analytical chemistry II (6)

Company economy and management,

Obligatory-optional subjects (8)

Theory of explosives (7)

Technology and safety of blasting

Obligatory-optional subjects in the 2nd step of MS-level studies:

Special analysis of explosives (3)

Basics of gun construction (2)

SUBJECTS OF THE 2ND STATE EXAMINATION

branch Safety engineering

Annotation of special subjects

Technology and safety of blasting operations

Theory of action of explosion

Basics of construction of ammunition

FOREIGN STUDENTS STUDIES

THE LOCATION OF THE DEPARTMENT

B.P. Aduev, E.D. Aluker and A.G. Krechetov

PRE-EXPLOSIVE PROCESSES IN HEAVY METAL AZIDES

SYNTHESIS AND REACTIVITY OF (CF3)2PN3, (CF3)2ASN3, CF3ASCIN3

Eduard D. Aluker*, Boris P. Aduev** and Alexander G. Krechetov*

PRE-DETONATION PHENOMENA IN HEAVY METAL AZIDES

Alexander M. Astachov*, Ludmila A. Kruglyakova*, Irina V. Gelemurzina*, Alexander D. Vasiliev**,

SIMPLE METHOD OF SYNTHESIS AND CHARACTERIZATION OF SOME NITROCYANAMIDE

Daniel Buczkowski* and Bogdan Zygmunt**,

INFLUENCE OF AMMONIUM NITRATE PRILLS POROSITY AND DIMENSIONS

Pawe Buraczewski*, Marek Sutkowski*, Andrzej Teodorczyk*and Shengjun Zhong**

DETO2D THE COMPUTER CODE FOR SIMULATIONS OF GASEOUS DETONATIONS

L. ai*, R. Halle**and Z. Ester**

EMULSIONS EXPLOSIVES ELMULEX AND ELMULEXAL

S. Cudzio*, S. Gachet**, A. Huczko***, M. Monthioux** and W. A. Trzciski*

SYNTHESIS OF CERAMIC AND CARBON NANOSTRUCTURES BY SELF-SUSTAINING

Radka Czeczotkov and Betislav Janovsk

THE INFLUENCE OF THE TYPE AND INTENSITY OF IGNITION SOURCES ON EXPLOSION

Andrzej Dbkowski, Artur Kozak and Andrzej Teodorczyk

THE INITIATION OF GASEOUS DETONATIONS IN H2-O2 MIXTURES

Mario Dobrilovi, Darko Vrkljan and Zvonimir Ester

DETERMINATION OF THERMAL STABILITY, RESISTANCE TO WATER AND RESISTANCE

FIRST EXPERIENCE WITH ANALYSIS AFTER LIQUIDATION OF SMOKELESS POWDERS

Darina Fierov *,**, Amer Hameed*, Timothy A. Rose*, John G. Hetherington*

SYSTEMATIC STUDY OF SIMULATED MINE EXPLOSIONS USING AUTODYN

Levi Gottlieb and Gady Korogodsky

DEGRADATION, REARRANGEMENT AND REDUCTIVE ACETYLATION

CHEMICAL AND THERMAL STABILITY OF FLEXIBLE PBXs

AGEING INFLUENCE ON PLASTIC EXPLOSIVES ON THE BASE OF RDX

COORDINATION COMPLEXES AS INORGANIC PRIMARY EXPLOSIVES

Alexander M. Astachov, Ludmila A. Kruglyakova, Irina V. Gelemurzina*, Alexander D. Vasiliev**,

Pawe Buraczewski, Marek Sutkowski, Andrzej Teodorczyk*and Shengjun Zhong**

L. ai*, R. Halleand Z. Ester

S. Cudzio, S. Gachet, A. Huczko, M. Monthioux** and W. A. Trzciski*

Darina Fierov *,**, Amer Hameed, Timothy A. Rose, John G. Hetherington*

H. S. Jadhav, D. D. Dhavale, M. B. Talawar* S. N. Asthana* and V. N. Krishnamurthy**

Zdenk Jalov*, Pavel Mareek, Kamil Dudek

Queenie S.M. Kwok, David E.G. Jones and Peeter Kruus**

Katarzyna Lipiska , Marek Lipiski and Andrzej Maranda **

Miroslav Pospil, Pavla apkov, Pavel Vvra and Svatopluk Zeman

Jakub eleovsk, Ji Pachmand Marcel Hanus**

Waldemar Witkowski, Karol Buchalik, Radoslaw Trebinski, Andrzej Maranda