Você está na página 1de 8

Thermochimica Acta 637 (2016) 102109

Contents lists available at ScienceDirect

Thermochimica Acta
journal homepage: www.elsevier.com/locate/tca

Short communication

Chemical kinetics on thermal decompositions of cumene


hydroperoxide in cumene studied by calorimetry: An overview
Yih-Shing Duh a,b,
a
Department of Occupation Safety and Health, Jen-Teh Junior College of Medicine, Nursing and Management, Miaoli, 35664, Taiwan, ROC
b
Department of Safety, Health and Environmental Engineering, National United University, No. 1 Lien-Da, Miaoli, 36052, Taiwan, ROC

a r t i c l e i n f o a b s t r a c t

Article history: Study on chemical kinetics related to the thermal decomposition of cumene hydoperoxide (CHP) in
Received 18 February 2016 cumene is summarized in this work. It is of great importance to gather and compare the differences
Received in revised form 3 June 2016 between these kinetic parameters for further substantial applications in the chemical industry and pro-
Accepted 5 June 2016
cess safety. CHP has been veried to possess an autocatalytic behavior by using microcalorimetry (such
Available online 6 June 2016
as TAM and C-80) operated at isothermal mode in the temperature range from 70 C to 120 C. However,
it exhibits a reaction of n-th order detected by non-isothermal DSC scanning and adiabatic calorime-
Keywords:
ter. By the isothermal aging tests, activation energy and frequency factor in logA(s1 ) were averaged to
Organic peroxide
Cumene hydroperoxide
be (117.3 5.9) kJ mol1 and (11.4 0.3), respectively. Kinetic parameters acquired from data of inter-
Chemical kinetics laboratories by using heat-ow calorimetry, the averaged activation energy and frequency factor in
Thermal decomposition logA(s1 ) were (119.3 11.3) kJ mol1 and (12.0 0.2), respectively. On the analogy of results from adia-
Calorimetry batic calorimetry, the activation energy and frequency factor in logA(s1 ) were respectively averaged to
be (122.4 9.2) kJ mol1 and (11.8 0.8). Five sets of kinetic models in relation to autocatalytic reactions
are collected and discussed as well.
2016 Elsevier B.V. All rights reserved.

1. Introduction possible mechanisms in various solvents [5]. In these early years,


researchers seldom devoted to the kinetic parameters related to
Cumene hydroperoxide (CHP) is generally made from the thermal decompositions of CHP due to the chemical analysis of
cumene oxidation process using air or pure oxygen as the oxi- transient intermediates being difcult and time consuming. In
dant in a batch reactor. Two CHP molecules can further dimerize by 1970, Hattori et al. reported that the reaction order on the thermal
the condensation reaction to produce a dicumyl peroxide (DCPO) decomposition of CHP in bubble column was determined to be near
and to give off a water molecule. In the phenol process, CHP is 1.0 by using iodometry [6]. However, in these twenty years, some
an intermediate used for producing phenol and acetone by acidic major incidents caused by the thermal runaway in the process ves-
cleavage. Besides, CHP is also widely used as an initiator for mak- sels containing CHP had ever been reported [1,2]. These incidents
ing the acrylonitrile-butadiene-styrene (ABS) copolymer in a ABS recalled and emphasized the noticeable relationships between the
plant. Fires and Explosions occurred in cumene oxidation reactor, phenomena of thermal runaway on CHP associated with chemical
CHP condensation reactor or CHP storage tank were the common kinetics. Duh et al. heighlighted this viewpoint and these incidents
types of incidents which were caused by the violent decomposition occurred in Taiwan [1,2]. Hou et al. rst proposed the autocat-
of CHP [1,2]. CHP 80 mass% has been classied to the class III (Fire alytic behavior on thermal decomposition of CHP 80 mass% at
hazards) by the NFPA, however, runaway incidents related to the temperature below 90 C detected by the Thermal Activity Monitor
reactive hazard of CHP in the lower concentration such as 35 mass% (TAM) [7]. Miyake et al. and Di Somma et al. reported the autocat-
have been reported [3,4]. alytic exothermic reactions of CHP in the concentration range of
In 1951, Kharasch et al. studied the chemistry of thermal 30 mass% to 80 mass% by spectroscopy, calorimetry, chromatog-
decomposition of CHP by analyzing products and establishing the raphy, equations of material balance and thermal balance [810].
Thereafter, ve models related to the characteristics of autocat-
alytic decomposition on CHP were consequentially propounded by
four laboratories [8,9,1114].
Correspondence to: National United University, Department of Safety, Health
The ASTM International Committee E27 on Hazard Potential
and Environmental Engineering, No. 1 Lien-Da, Miaoli, 36052, Taiwan, ROC.
E-mail address: yihshingduh@yahoo.com.tw of Chemicals had performed the search of calorimetric reference

http://dx.doi.org/10.1016/j.tca.2016.06.003
0040-6031/ 2016 Elsevier B.V. All rights reserved.
Y.-S. Duh / Thermochimica Acta 637 (2016) 102109 103

(4) Emergency response base on the temperature with respect to


Nomenclature time curve such as emergency cooling and bottom dumping
(self-heat rate versus rate constant).
A, A01 , A02 Frequency factor [sec1 M1n ] (5) Linkage between lumped kinetics and stepwise kinetics of ele-
dT/dt Self-heat rate [K min 1 ] mentary reactions.
Ea , Ea1 , Ea2 Activation energy [kJ mol1 ] (6) Relationship between n-th order reaction and autocatalytic
k, k1, k2 Rate constant [s1 M1n ] reaction.
k* Pseudo zero-order rate constant [s1 M1n ] (7) A possible candidate of kinetic reference materials for thermal
n,n1 ,n2 ,n11 ,n21 ,n22 Order of reaction analysis and adiabatic calorimetry.
R Gas constant [8.314 J g1 K1 ]
Degree of conversion In order to make better uses of the kinetic parameters in the
above mentioned applications, it is cardinal to have the kinetic
triplets within acceptable uncertainty. Consequently, a summary
work on searching precise kinetic parameters is imperative for the
materials and issued the E2781 [15]. Blaine had examined six safety engineering to avoid oversizing or underestimating designs.
candidate materials, including di-tertiary-butyl peroxide (DTBP),
trityl azide, azobenzene, azobisisbutyronitrile (AIBN), cumene 2. Experimental
hydroperoxide (CHP), and phenytetrazolthiol to identify possible
calorimetric reference materials for evaluating kinetic parameters, 2.1. Chemicals
CHP was one of the chosen compounds that ever been discussed
[16]. Phenomena of macroscopic kinetics in relation to thermal Cumene hydroperoxide (CHP) with a purity of about 82 mass%
decompositions of CHP/cumene solutions performed by DSC, RSST was provided from local company in Taiwan. Cumene with the
and ARC were published [3,4,17]. Nevertheless, the reported acti- purity higher than 99.9% was purchased from Merck Company and
vation energies (Ea ) and frequency factor (logA, A in s1 M1n ) used as a compatible diluent. CHP was stored in a refrigerator at
displayed the quite broad distinction, Ea varied from 100.0 to the 4 C environment for maintaining the purity. Samples of CHP
176 kJ mol1 and logA(s1 M1n ) distributed from 11.25 to 14.89, in diluted concentrations were made of cumene and CHP in 82
respectively [16]. In recent years, some few astonishing logA mass%. Both chemicals were commercially available and were used
(sec1 ) around 20 and Ea as high as 196 kJ mol1 were successively without further purication.
reported [11,18,19]. Besides, several similar studies on the ther-
mal decomposition of dicumyl peroxide (DCPO) which possessing 2.2. Differential scanning calorimeter (DSC)
the same cumyloxy radical and dissipating mechanism have been
reported [2023]. Though calorimetry offers a very high efcient Dynamic scanning tests were performed by a Mettler TA-4000
approach to determine the chemical kinetics of a typical reac- system coupled with a DSC822e measuring cell [26]. CHP samples
tion with one-step global kinetic mechanism, the kinetic triplets were sealed in the disposable crucibles (ME-26732) which were
differ between interlaboratory tests and sometimes the between- used for highly exothermic reactions or decomposition reactions.
laboratory uncertainty larger than 20%. However, being existed The data acquisition system recorded the thermal curves and stored
signicant discrepancy among these frequency factors and activa- for further analysis. Scanning rates were selected to be 2, 4, and
tion energies, the microscopic pinpoints or evidences have not been 10 C min1 in programmatic ramps from 30 to 350 C.
exaggerated to resolve the existing paradox. Obviously, the biased
quantities of these activation energies and frequency factors in the 2.3. Accelerating rate calorimeter (ARC)
published literature are too huge to be convincible so far forth. The
relative standard deviation between laboratories recommended by An accelerating rate calorimeter (ARC), manufactured by
the ASTM E2041 for activation energy, frequency factor, and reac- Columbia Scientic Industries in Austin, Texas [27], was used to
tion order have to be less than 9.8, 9.8 and 22%, respectively [24]. conduct the unrestrained CHP runaway experiments. Three types
For this reason, the International Conference on Thermal analysis of test bomb made of titanium, Hastlloy C, and stainless 316 with a
and Calorimetry (ICTAC) had taken a mathematical model to verify volume of 10 ml were utilized. The thermal inertia were about from
kinetic methods selected in worldwide laboratories. By using the 1.2 to 4. The detailed performance and theory of the ARC instru-
thermal decomposition of calcium carbonate as a model compound, ment was proposed by Townsend or by ASTM E1981 [28,29]. Once
the inter-method reliability should be less than the simulated stan- the ARC system detected the temperature rate exceeding the sen-
dard deviation a 39% for reaction order, a 16% for Ea, and a 25% for sitivity of 0.02 C min1 , then turned off the radiating heater and
log (frequency factor) [25]. switched on the guard heaters for tracing the accelerating reactions.
Based on emphasizing the mentioned reasons and to solve the The temperature and pressure behaviors related to the violent ther-
disputes for long views, an objective for declaring the gigantic mal runaway of CHP samples can thus be explored by using an
differences in activation energy and frequency factor is ardently accelerating rate calorimeter.
desired to be executed now. Especially, the accurate kinetic param-
eters are of great importance in the following aspects: 2.4. VSPII(vent sizing packageII)

For conducting the highly exothermic reaction in the ARC, mass


(1) Optimization for the cumene oxidation process with control- of sample was generally less than 3 g, the thermal inertia thus had
lable CHP concentrations (rate constant of oxidation versus a value larger than 2. For a larger phi-factor transcends a value
thermal decomposition). of 2, the collected data will be distorted by the thermal inertia.
(2) Emergency relief system design of cumene oxidation reactor This is the so-called phi effect. Even by using Townsends meth-
such as vent sizing and mass ow rate (analytical equation for ods for deconvoluting the effects of high phi-factor, data of the
vent sizing versus rate constant). adiabatic temperature rise can be corrected accurately. However,
(3) Evaluation of a safe reaction region for cumene oxidation reac- the kinetics-related data such as self-heat rate and adiabatic time-
tor (reactor parameters versus rate constant). to-maximum-rate cannot be recovered as precise as the real ones
104 Y.-S. Duh / Thermochimica Acta 637 (2016) 102109

Table 1
Kinetic parameters on thermal decompositions of CHP by isothermal methods.

Concentration (mass%)/diluent Calorimeter Order (n) Ea (kJmol1 ) Log A (s1 ) Reference

35/cumene DSC 0.5 120.6 3.0 11.5 [3]


2080/cumene DSC 0.50 0.02 122.0 3.0 11.3 0.5 [4]
80/cumene DSC NA 110.3 NA [8]
80/cumene in N2 C-80 NA 128 NA [10]
80/cumene in air C-80 NA 119 NA [10]
80/cumene in air TAM NA 104 NA [10]

Note: activation energy and frequency factor in logA are averaged to be 117.3 5.9 kJ mol1 and 11.4 0.3, respectively.
NA: Not available.

Table 2
Kinetic data for n-th order thermal decomposition of CHP by heat-ow calorimetry.

Concentration (mass%)/diluent Calorimeter/Methodology Scanning rate (K min1 ) Order (n) Eaa (kJ mol1 ) Log Ab (s1 ) Reference

35/cumene DSC 4 0.5 122.0 11.7 [3]


35/cumene DSC 4 0.5 126.0 12.2 [3]
NA Iodometry NA 1.0 100.0 NA [5]
80/cumene C-80 0.11.0 NA 129.0 NA [10]
88/cumene DSC 1 NA 173.0 18.7 [11]
88/cumene DSC 2 NA 197.0 20.8 [11]
88/cumene DSC 4 NA 193.0 20.4 [11]
88/cumene DSC 10 NA 186.0 19.1 [11]
NA DSC NA NA 176.0 NA [18]
80/cumene DSC 4 NA 190.0 NA [19]

Note: activation energy and frequency factor expressed in logA are averaged 119.3 11.3 kJmol1 and 12.0 0.2, respectively.
NA: Not available
a
Ea > 140.0 kJ mol and Ea < 00.0 kJ mol 1 are disregarded in averaged value.
b
logA > 14.0 and logA < 10.0 are disregarded in averaged value.

Table 3
Kinetic parameters on autocatalytic thermal decomposition of CHP determined by heat-ow calorimetry.

Concentration (mass%)/diluent LogA01 (s1 ) n1 Ea1 (kJ mol1 ) LogA02 (s1 M1n ) n2 Ea2 (kJ mol1 ) Reference

NA NA NA 120 12 NA NA NA [8]
1.85.4 M 9.25 0.1 1 108.5 0.1 8.67 0.1 1 97.2 0.0 [9]
88/cumene 6.086 0.638 48.682 19.2 n21 = 0.547, n22 = 1.295 180.176 [11]
80/cumene NA NA NA NA n21 = 0.5, n22 = 0.5 NA [12]
80/cumene 8.87 1.8 102.47 9.98 n21 = 4.02, n22 = 3.00 95.81 [13]
80/cumene 9.3 0.1 1 102.0 2.0 8.2 0.1 1 89.0 2.0 [14]

Note: the divergency in chemical kinetics of autocatalytic decompositions is too immense to be good for taking mean values.
NA: Not available.

Table 4
Rate equations proposed on autocatalytic thermal decomposition of CHP.

Calorimeter /Methodology Observation/Rate equation Reference

TAM Observation [7]


C-80 and TAM Observation [8]
DSC and ARC d/dt = k1 (1 )1 + k2 (1 )1 ()1 [CHP]0 1 [9]
DSC d/dt = k1 (1 )0.638 + k2 (1 )0.547 ()1.295 [CHP]0 1.295 [11]
TAM/Avrami-Erofeev rate law d/dt = k(1 )0.5 ()0.5 [CHP]0 1 [12]
DSC d/dt = k1 (1 )1.8 + k2 (1 )4.02 ()3.00 [CHP]0 3.00 [13]
DSC d/dt = k1 (1 )1 + k2 (1 )1 ()1 [CHP]0 1 [14]

under the phi-factor approaching unity. Therefore, scientists con- the thermal runaway of CHP under the thermal inertia less than 1.2.
secrated to the following three approaches to solve the hindrances The thermal inertia of the reaction system has been improved to be
distorted by high thermal inertia. First, using large quantities of as low as 1.1 by a thin can made of stainless steel with a volume
diluted sample in ARC bomb was commonly adopted for lowering of 110 ml. The lower heat capacity of the cell assured the excellent
the phi value and overcoming the defect of phi-factor. The second adiabatic condition to perform the real runaway behavior within
way was to use the VSP or Phi-TECH instrument, in which a thin the test sample. Temperature, pressure and derivative data such as
cell with a maximum volume of 112 ml for holding sample about self-heat rate or adiabatic temperature rise under excellent ther-
50 ml, which cell had been capable of improving the phi-factor mal inertia can therefore be applied directly to process scale. Even
as low as 1.1. Third, some researchers established the theoretical realizing the importance of reducing the phi effect on the thermal
approaches to simulate thermal runaway for the energetic materi- decomposition of CHP, only a few exothermic decompositions of
als which could not be either diluted with solvent or used in large pure CHP were carried out by using adiabatic calorimeters.
quantity in adiabatic calorimeters [3032].
An adiabatic calorimeter with the enhanced capability of
improving thermal inertia, the Vent Sizing Package II (VSP II) man-
ufactured by Fauske and Associates Corp. [33] was used to measure
Y.-S. Duh / Thermochimica Acta 637 (2016) 102109 105

Table 5
Kinetic parameters on thermal decomposition of CHP determined by adiabatic calorimetry.

Concentration (mass%)/diluent Adiabatic calorimeter Thermal inertia Ordera (n) Ea b (kJ mol1 ) Log Ac (s1 M1n ) Reference

35/cumene ARC 1.28 0.5 119.8 11.1 [3]


35/cumene ARC 1.28 0.5 116.9 10.7 [3]
35/cumene ARC 1.28 0.5 118.1 10.9 [3]
15/cumene VSP II 1.19 0.5 125.0 11.3 0.4 [4]
12/cumene RSST 1.119 0.005 NA 111.4 5.4 NA [17]
20/cumene RSST 1.119 0.005 NA 108.6 7.2 NA [17]
30/cumene RSST 1.119 0.005 NA 109.7 1.4 NA [17]
40/cumene RSST 1.119 0.005 NA 113.4 2.9 NA [17]
50/cumene RSST 1.119 0.005 NA 136.6 8.2 NA [17]
60/cumene RSST 1.119 0.005 NA 136.4 5.3 NA [17]
70/cumene RSST 1.119 0.005 NA 138.3 7.5 NA [17]
80/cumene RSST 1.119 0.005 NA 138.7 5.4 NA [15]
25/cumene VSP II NA 0.5 122.3 13.0 [41]
80/cumene VSP II 1.1 0.5 132.0 12.1 [42]
88/cumene VSP II 1.1 0.5 134.0 12.6 [42]
15/cumene VSP II NA 0.5 120.2 12.5 [43]
12/cumene VSP II NA 1.0 127.0 NA [44]
28/cumene VSP II NA 1.0 124.5 NA [44]
48/cumene VSP II NA 1.0 109.8 NA [44]
16/DIBd Phi-TEC I 1.75 NA 155.5 0.4 NA [45]
24/DIBd Phi-TEC I 1.65 NA 142.0 0.9 NA [45]
32/DIBd Phi-TEC I 1.77 NA 127.5 NA [45]
16/DIBd Phi-TEC II 1.12 NA 155.5 0.4 NA [45]
24/DIBd Phi-TEC II 1.12 NA 142.0 0.9 NA [45]

Note: activation energy and frequency factor in logA are averaged to be 122.4 9.2 kJ mol1 and 11.8 0.8, respectively.
NA: Not available
a
Order has been determined by the Arrhenius plot of ln k* versus 1/T using ARC theory.
b
Ea > 140.0 kJmol1 and Ea < 100.0 kJmol1 are disregarded in averaged value.
c
logA > 14.0 and logA < 10.0 are disregarded in averaged value.
d
DIB is the abbreviation of 2,2,4-trimethyl-1,3-pentaanediol diisobutyrate. Due to the different dynamics of H-abstraction and C C scission of cumyloxy radical in
addition to the induced decomposition of CHP proceeding in diverse solvents, these activation energies have not been adopted in the averaged activation energy.

2.5. Methodologies for acquiring kinetic parameters 2.6. Search for kinetic parameters on thermal decomposition of
CHP
2.5.1. Differential thermal analysis
By the dynamic scanning using DSC, the thermal curve of CHP Thermal curves identied by isothermal, non-isothermal and
decomposition appeared to be a simple curve. As a rule of thumb, adiabatic modes have been implemented in diversely interna-
this means that the thermal curve upon thermal decomposition tional laboratories. Concentrations of CHP solutions in cumene
of CHP must be a smoothly well-behaved one with nice conversion distributed from 12 mass% to about 82 mass%. Kinetic param-
versus temperature relation. Interferences of shoulders, overlapped eters searched from calorimetric studies were collected and
multiple peaks, shifts in baselines or discontinuous steps have not summarized. These kinetic results mainly determined by the com-
ever observed in thermal curves of CHP. Kinetic parameters in rela- mercial calorimeters such as DSC, TAM, ARC, RSST, C-80 and
tion to the thermal curve can be extracted by using a suitable kinetic VSP2. These laboratories had chosen methodology or software
model. Because of its simplest feature in thermal curve, method- package for analyzing the kineic parameters such as the method
ology of thermal safety software supplied by CISP (Cheminform St. of Flynn/Wall/Ozawa (ASTM E698) [34], method of Borchardt &
Petersburg Ltd., Russian) or thermal analysis methods (such as the Daniels (ASTM E2041) [24], method of model-free, Townsends the-
method of Flynn/Wall/Ozawa or the method of Borchardt & Daniels) ory (ASTM E1981) [29], Thermal safety software (TSS) [36] and
of DSC can be applied for determining the chemical kinetics on ther- Thermokinetics software (Advanced Kinetics and Technology Solu-
mal decomposition of CHP [24,3436]. These approaches provided tions, AKTS) [37]. Owing to the existed large discrepancy among
the efcacious procedures for determining activation energy, fre- kinetic parameters, a detailed inspection of these data was val-
quency factor and order of reaction by using differential thermal idated to discriminate the reasonable data from the published
analysis with excellent precision. kinetic parameters. A data judged to be acceptable or not was based
on the bias of individual parameter with an uncertainty compared
to the averaged value less than that recommended by ASTM E2041
2.5.2. Adiabatic calorimetry and townsends theory [24].
By using the adiabatic thermal curve, data analysis has based on
the assumption of the simple reaction scheme of one reactant. By 3. Results and discussion
obeying the Townsends theory in conjunction with the application
of adiabatic calorimeter, kinetic parameters on thermal decompo- 3.1. Chemical kinetics
sition of CHP can be deduced from self-heat rate (dT/dt) versus
reciprocal of the absolute temperature. Intrinsically, k is propor- 3.1.1. Isothermal methods
tional to self-heat rate (dT/dt) in rate equation. By the method of Applications of isothermal methodology for exploring the
trial and error, the plot of ln(k*) as a function of the reciprocal of exothermically autocatalytic reactions have been emphasized by
absolute temperature was used to identify the reaction order. A lin- the guidelines in ASTM E487 [3,4]. Autocatalytic decomposition of
ear tted line can thus be obtained by selecting a correct n, thereon energetic compound can be thus detected by the programmatic
Ea (kJ mol1 ) and A(s1 M1n ) will be explicitly determined by the isothermal mode within the heat-ow calorimeter. Isothermal
straight line of ln (dT/dt) vs. 1/T [28]. experiment keeping the excellent thermal equilibrium is of benet
106 Y.-S. Duh / Thermochimica Acta 637 (2016) 102109

Table 6
The proposed mechanism on thermal decomposition of CHP in cumene.

Number of elementary reaction Method/Instrument Important results Reference

11 DSC, ARC, GC and HPLC CHP/cumene at 3080 mass% exhibited [9]


autucatalytic behavios under thermal
decomposition. Acetophenone and carbinol are
the main products inentied in the reacting
solution during the CHP decomposition process
with a minor occurrence of -methylstyrene,
DCPO and phenol. Gas chromatographic
analyses of the gases evolved and collected
during the experimental runs indicated the
presence of methane and oxygen as the main
gaseous products. The effect on the catalytic
decomposions by carbinol or acetophenone or
-methylstyrene was observed.
NA Super-CRC and IR spectrometry CHP/cumene 40 mass% decompose directly [10]
into acetophenone and organic acids with a
small heat of reaction. CHP/cumene 80 mass%
decompose in stages, rst to acetone and
-methylstyrene via an autocatalytic reaction
with a large heat of reaction, followed by a
mild reaction to acetophenone and acids.
7 DSC and GC-Mass The rate equations related to autocatalytic [14]
mechanism of the CHP decomposition was
described. In elementary Reaction (7), oxygen
can be produced by the dimerization of cumyl
biradicals.
8 Density function theory and CHP 40 mass% is a critical concentration where [17]
Gaussian-03 program package/RSST the reaction pathways begins to change. Once
the concentration exceeds 40 mass%,
elementary Reactions (3)(5) are no longer
kinetically competitive because of the decrease
of cumene; Reactions (6) and (7) become
dominant because radicals of OH need to be
consumed. The dramatical increase of
maximum pressure and pressure-rising rate
when the CHP above 40 mass% will need to be
further studied.
4 Gas-liquid chromatography A quantitative analysis of the products of CHP [39]
in heptane and benzene was used to study the
kinetics of the induced reaction due to the
abstraction of hydroperoxy hydrogen by
cumyloxy radical. Branch ratios of k4 /k3
(induced decomposition/ C C scission) and
k2 [SH]/k3 (hydrogen abstraction from solvent
SH/C C scission) were determined.
51 Equations of thermal balance and A detailed model had been proposed to explain [40]
material balance/DSC, ARC, GCand the thermal decomposition of CHP coupling a
HPLC thermal balance and a material balance
equation. The decomposition of CHP was
successfully explained by quantitative analysis
in conjunction with kinetic model, calorimetry
and chromatography.
12 DSC, ARC, and VSP The elementary reaction of the induced This work
decomposition of CHP by cumyloxy radical is
added in Eq. (9). Huge pressure hazard in
maximum pressure and pressure-rising rate on
the thermal decomposition of CHP has been
deduced to be the generation of gaseous
products CH4 and O2 . The phenomena imply
the increased branch ratios of elementary
Reactions (2), (3) and (10) will dominate when
the CHP above 35 mass% under thermal
decomposition.

to eliminate the thermal lag effect. Thermal decomposition of CHP 3.1.2. n-th order reaction determined by heat-ow calorimetry
at temperature above 100 C behaved as a n-th order reaction. The Kinetic parameters on thermal decompositions of CHP in various
fact was judged from the thermal curve detected in the isothermal solutions studied by DSC or related methodology are depicted in
mode in DSC, i.e. the maximum peak power appeared at the ini- Table 2 [3,6,10,11,18,19]. Order of reaction equals 0.5 0.02 which
tial 0% conversion at the instant of isothermal test. Table 1 lists the has been delicately veried by means of calorimetry operated
kinetic parameters on thermal decompositions of CHP determined in isotherml and non-isothermal modes [3,4]. Activation ener-
by isothermal methods [3,4,8,10]. gies exhibit excellent agreement in the range between 120 and
126 kJ mol1 . Frequency factor expressed in logA is validated to
be about 11.7 0.5. Due to the analogous dissipating pathway of
cumyloxy radical liberated by the decomposition of CHP and DCPO,
Y.-S. Duh / Thermochimica Acta 637 (2016) 102109 107

logA should present a reasonable value in the range of 1114 [35]. Table 7
Kinetic parameters of CHP on n-th order thermal decompositions.
Recently, Reyes Valdes et al. studied the thermal decompositions
of DCPO solvated both in 2,2,4-trimethyl-1,3-pentanediol diisobu- CHP mass%/cumene Ea logA Methodology/Calorimetry
tyrate (DIB) and cumene with concentrations of 20, 30 and 40 (kJ mol1 ) (s1 M1n )
mass% by using adiabatic calorimetry [20,21]. The comparable fre- 2080 mass% 117.3 5.9 11.4 0.3 Isothermal tests
quency factors (in logA) were reported to be within the range of 35, 80 and 88 mass% 119.3 11.3 12.0 0.2 Heat-ow calorimetry
14.015.33 in addition to the activation energies were reported to 1280 mass% 122.4 9.2 11.8 0.8 Adiabatic calorimetery

be from 140 to 156 kJ mol1 [20,21]. One of the decisive reason is


that the logA and Ea is related to the exit channels of the cumy-
3.1.4. Kinetic parameters of n-th order reaction determined by
loxy radical by C C scission or bimolecular reactions. The logA
adiabatic calorimetry
and activation energy for C C scission of cumyloxy are approxi-
Even realizing the importance of reducing the phi effect on
mately 1.4 log units and 6.7 kcal mol1 less than these parameters
thermal decomposition of CHP, only a few exothermic decompo-
estimated for decomposition of tert-butoxy radical(log A = 14.1
sitions of pure CHP were carried out by using ARC or adiabatic
and E = 15.3 kcal mol1 ) [38]. The decrease in Ea can be attributed
calorimeters with a thermal inertia less than 1.2 in the past decade.
to a difference in the conguration of transition states between
Table 5 displays the kinetic data on thermal decomposition of vari-
cumyloxy and tert-butoxy radicals, in the case of cumyloxy radi-
ous CHP solutions determined by adiabatic calorimetry and ARSST
cal which involving some resonance stabilization in the transition
(Advanced Reactive Systems Screening Tool) [3,4,17,4145]. Duh
state. Due to containing benzyl group with higher steric hindrance,
et al. rst proposed the order is equal to 0.5 on the thermal decom-
this stiffened complex also resulted in a reduction of the frequency
position of CHP in the range of 1535 mass% from the Arrhenius plot
factor owing to the loss of entropy of activation resulting from
of ln k* versus 1/T by applying ARC theory. Lu et al. conducted the
an increase in the barrier to internal rotation or pseudo-rotation
thermal decomposition of CHP with the concentrations from 12 to
[38,39]. Some drastically paradoxical results concerning the ther-
80 mass% by Reactive System Screening Tool (RSST) [17]. Activation
mal decomposition of CHP were reported. Six sets of tremendous
energies presented manifestly the lower values in the range from
activation energies or log A without reaction order were announced
108.6 to 111.4 kJ mol1 , when the concentrations of CHP below 40
[11,18,19]. Besides, such astonishing deviations in activation ener-
mass%. Once the concentrations of CHP being above 40 mass%, acti-
gies and log A are far away from the suggested uncertainties in
vation energies held the higher values distributing from 136.4 to
ASTM E2041 [24]. Once the bias of the Ea or logA exceed 10%, which
138.7 kJ mol1 which were suspected to be the changes in reaction
was the upper uncertainty recommended by ASTM E2041, then
mechanisms [17].
these highly deviant data were abandoned and not included in the
average value.
3.2. Chemical kinetics and reaction mechanism
3.1.3. Autocatalytic reaction determined by heat-ow calorimetry
Elementary reactions can only be directly veried by crossed
DSC and adiabatic calorimetry failed to discriminate the
molecular beams under an ultra high vacuum condition. Radi-
detectable signal from the thermal decomposition of CHP below
cals, intermediates and transient species with the existing time
90 C. A novel observation of the exothermic phenomenon with
scale of sec in elementary reactions cannot be discriminated by
autocatalytic feature was veried by Hou et al. and Miyake et al.
enthalpy changes. Thermal curve and self-heat rate acquired from
using TAM [6,8]. Chemical kinetics associated with autocatalytic
calorimetry are both a representation of the overall reaction. Reac-
models were proposed [9,1114]. By combining both the calori-
tion mechanism is a bridge to link the rate of reaction and its
metric data and decomposition products detected by GC/MS,
hypothetical elementary reactions proposed. Traditionally, reac-
mechanism on the decomposition of the CHP was claimed explic-
tion mechanism is supported by the products analysis by using
itly [9]. Furthermore, a more detailed study proposed by Di Somma
spectrometry, chromatography coupled with mass analyzer and
et al. by the conformation on concentrations of products deter-
electrochemical analysis. Thermal curve thus detected by calorime-
mined by GC and HPLC [40]. A common rate equation related to
ter is an exothermic behavior of the bulk reaction system. It is now
parallel reaction with autocatalytic behavior was suggested by Di
still a challenge to link reaction mechanism to thermal curve. Even
Somma et al. and Talouba et al., which equation combined a rst
so, researchers devote themselves to propose the mechanism on
order decomposition and an autocatalytic reaction of second order
thermal decomposition of CHP for explaining the lumped kinetics
[9,14].
determined by calorimetry.
They proposed an autocatalytic model to develop chemical
kinetics as the following.
3.2.1. Reaction mechanism on thermal decomposition of CHP in
AB k1 = A01e Ea1 /RT (1) cumene
Solvated CHP adopted the cleavage of an O O bond to go
A + B 2B k2 = A02e Ea2 /RT (2)
through the pathways of decomposition after thermal-assisted
The corresponding rate equation can thus be represented as collisions or photo-excited states. The decay routes of cumyloxy
radical depend on the molecules or radicals encountered in the
d/dt = k1 (1 )1 + k2 (1 )1 ()1 (3) solution. Cleavage or elimination of cumyloxy radical via tertiary
carbon and C C scission is one of the ways to penetrate the
Another simulation by software regarding the detailed elemen-
reaction scheme. In hydrocarbon solutions, the competition of
tary reactions which composite the overall decomposition of the
hydrogen abstraction has been well-known in the exit channels
CHP was proposed by Chen et al. [11]. The rate equation was
of cumyloxy radical. The cumyloxy radical can proceed an elimina-
expressed as the following.
tion by giving up a methyl or phenyl group in breaking tertiary C C
d/dt = k1 (1 )n 1 + k2 (1 )n 21 ()n 22 (4) bonds. Both the acetone and acetophenone is therefore predicted
to be dominated in the nal products. Two similar mechanisms
Table 3 depicts the kinetic parameters related to autocatalytic with respective 8 and 11 elementary reactions for explaining the
decompositions [9,1114]. Rate equations in relation to the auto- thermal decomposition of CHP were proposed by Lu et al. and Di
catalytic models are collected and expressed in Table 4. Somma et al. [9,17]. An improved mechanism for explaining the
108 Y.-S. Duh / Thermochimica Acta 637 (2016) 102109

chemistry on thermal decomposition of CHP in cumene is presented reactions. After excluding those less competitive elementary reac-
as follows: tions, a dominant reaction pathway was determined based on
ve elementary Reactions (1)(4) and (8). The overall reaction
C6 H5 C(CH3 )2 OOH C6 H5 C(CH3 )2 O + OH (1)
was balanced by the dominant elementary reactions as follows:
C6 H5 C(CH3 )2 O C6 H5 COCH3 + CH 3 (2) CHP + 2 cumene C6 H5 COCH3 + CH4 + H2 O + [C6 H5 C(CH3 )2 ]2 [17].
CH
The predicted reaction pathway also agreed with the available
C6 H5 C(CH3 )2 H + 3 C6 H5 C(CH3 )2 + CH4 (3) experimental results of methane and acetophenone. When the CHP
C6 H5 C(CH3 )2 H + OH
C6 H5 C(CH3 )2 + H2 O (4) concentration is lower than 40 mass%, the reaction followed the
proposed reaction pathway since cumene was sufcient in the
reaction system. Once the concentration exceeded this 40 mass%,
C6 H5 C(CH3 )2 O + C6 H5 C(CH3 )2 Reactions (3)(5) in the dominant pathway were no longer kineti-
cally competitive because of lack of cumene: Reactions (6) and (7)
H C6 H5 C(CH3 )2 OH + C6 H5 C(CH3 )2 (5) become active because radicals OH needed to be consumed [17].
Rado et al. had studied the kinetics of the induced reaction due
to the abstraction of hydroperoxy hydrogen atoms by cumyloxy
C6 H5 C(CH3 )2 O + OH CH3 COCH3 + C6 H5 OH (6a) radicals [39]. Eqs. (9) and (10) presented that the cumylperoxy
radical can be produced by the induced decomposition of CHP
C6 H5 C(CH3 )2 O CH3 COCH3 + C6 H5 (6b)
and cumyloxy radical. Two cumylperoxy radicals formed a dimeric
C6 H5 C(CH3 )2 + OH C6 H5 CH3 C = CH2 + H2 O (7) intermediate then completing the induced reaction by generating

a DCPO molecule and releasing an oxygen [46].
C6 H5 C(CH3 )2 + C6 H5 C(CH3 )2 [C6 H5 C(CH3 )2 ]2 (8a)

C6 H5 + C6 H5 [C6 H5 ]2 (8b) 3.3. Kinetic parameters on n-th order thermal decomposition of
CHP
CH3 + CH3 C2 H6 (8c)

C6 H5 + CH3 C7 H8 (8d) Table 7 compiles the well-conditioned kinetic parameters of


the n-th order thermal decompositions of CHP gathered by various
methodologies. The data listed in Tables 1, 2 and 5 tell the results
C6 H5 C(CH3 )2 O + C6 H5 C(CH3 )2 OOH C6 H5 that the activation energies in the thermal decomposition of CHP
in cumene are determined to be the reasonable values between
C(CH3 )2 OH + C6 H5 C(CH3 )2 OO (9) 100 and 138.7 kJ mol1 . In Table 7, the overall averaged activation
energy is determined to be 119.7 11.3 kJ mol1 with a standard
deviation about 9.4%. The overall determined logA (s1 ) has a pre-
cise value of 11.7 0.8 with a bias less than 7%. It is clear that
2C6 H5 C(CH3 )2 OO [C6 H5 C(CH3 ) the relative standard deviation of the above kinetic parameters in
Table 7 is acceptable by obeying the practice of ASTM E2041 [24]
2 OO] [C6 H5 C(CH3 )2 O]2 + O2 (10)
and the recommendation of ICTAC [25].

4. Conclusions
[C6 H5 C(CH3 )2 O]2 2C6 H5 C(CH3 )2 O (11)

In this study published kinetic parameters on the thermal


decompositions of CHP in cumene are summarized and discussed.
C6 H5 C(CH3 )2 OOH + catalyst C6 H5
It is of great importance to obtain a reliable and precise range of
C(CH3 )2 O + OH + catalyst (12) chemical kinetics for the safety and engineering applications in
CHP-related processes. For n-th order decomposition of CHP, the
averaged activation energy and logA in possession of the value
The proposed mechanism on thermal decomposition of CHP in of (120.1 11.3) kJ mol1 and (11.7 0.8) are determined world-
cumene from literature is listed in Table 6 [9,10,14,17,39,40]. wide by heat-ow calorimetry and adiabatic calorimetry. The
above uncertainties are a 9.4% in activation energy and a 6.8% in
3.2.2. Reaction pathway analysis [5,17,39,40,46] logA which meet the both recommendations of ICTAC and ASTM.
Dimethylphenylcarbinol, DCPO, methane and phenol had been The paradoxical model between the n-th order and autocatalytic
veried as the major products on thermal decomposition prod- decomposition is existing and has not been exactly resolved. Espe-
ucts in conjunction with the solvent effects in the early literature cially, some large deviations in the activation energy fell in 100
[5]. As seen in the above mechanism, Reaction (1) is the initia- and 197 kJ mol1 and logA(s1 M1n ) varied from 11.25 to 14.89,
tion step and the rate-determining step of the overall reaction. respectively, which facts need more advanced studies to solve the
The cumyloxy radical (C6 H5 C(CH3 )2 O) produced by the rst step problems existed. More future efforts are suggested to aim at the
could proceed in three steps of Reactions (2), (5) and (6). The induced decomposition of CHP and autocatalytic decomposition by
formation of acetophenone through C C scission by the cumy- diverse methodologies.
loxy radical in the Reaction (2) accompanying a methyl radical
given off. The consequential reaction of methyl radical then
References
abstracts a hydrogen atom from cumene or hydrogen donors to
produce the methane, which is the major non-condensable com- [1] T.-C. Ho, Y.-S. Duh, J.R. Chen, Case studies of incidents in runaway reactions
ponent of pressure hazard. For the OH radical generated by the and emergency relief, Process Saf. Prog. 17 (1998) 259262.
homolysis of O O bond in the rst step, there are also three pos- [2] Yih-Shing Duh, Xiao-Hui Wu, Chen-Shan Kao, Hazard ratings for organic
peroxides, Process Saf. Prog. 27 (2008) 8999.
sible reaction routes including reacted with cumene, cumyloxy [3] Y.-S. Duh, C.-S. Kao, C. Lee, S.W. Yu, Runaway hazard assessment of cumene
radical or cumyl radical via Reactions (4), (6a) and (7), respec- hydroperoxide from cumene oxidation process, Trans. IChemE Part B Process
tively. Reactions (8a), (8b), (8c) and (8d) are possible termination Saf. Environ. Prot. 75 (1997) 7380.
Y.-S. Duh / Thermochimica Acta 637 (2016) 102109 109

[4] Y.-S. Duh, C.-S. Kao, H.-H. Hwang, W.W.-L. Lee, Thermal decomposition project-data, methods and results, Thermochim. Acta 355 (2000) 125143.
kinetics of cumene hydroperoxide, Trans. IChemE Process Saf. Environ. Prot. [26] [4000YA] operation instructions, Mettler Company Switzerland, 1993.
Part B 76 (1998) 271276. [27] A.R.C. Manual Columbia Scientic Industries Corp. Austin Texas, 1993.
[5] M.S. Kharasch, A. Fono, W. Nudenberg, The chemistry of hydroperoxide. VI. [28] D.I. Twonsend, J.C. Tou, Thermal hazard evaluation by an accelerating rate
The thermal decomposition of -cumyl hydroperoxide, J. Organ. Chem. 16 calorimeter, Thermochim. Acta 37 (1980) 130.
(1951) 113127. [29] ASTM E198112, Standard Guide for Assessing Thermal Stability of Materials
[6] K. Hattori, Y. Tanaka, H. Suzuri, T. Ikawa, H. Kubota, Kinetics of liquid phase by Methods of Accelerating Rate Calorimetry.
oxidation of cumene in bubble column, J. Chem. Eng. Jpn. 3 (1970) 7278. [30] E. Gimzewski, G. Audley, Thermal hazards: calculating adiabatic behavior
[7] H.-Y. Hou, C.-M. Shu, Y.-S. Duh, Exothermic decomposition of cumene from DSC data, Thermochim. Acta 214 (1993) 129140.
hydroperoxide at low temperature conditions, AIChE J. 47 (2001) 18931896. [31] W.Y. Wang, Y.S. Duh, C.M. Shu, Evaluation of adiabatic runaway reaction and
[8] A. Miyake, K. Nomura, Y. Mizuta, M. Sumino, Thermal decomposition analysis vent sizing for emergency relief from DSC, J. Therm. Anal. Calorim. 85 (2006)
of organic peroxides using model-free simulation, J. Therm. Anal. Calorim. 92 225234.
(2008) 407411. [32] A.A. Kossoy, J. Singh, E.Y. Koludarova, Mathematical methods for application
[9] I. Di Somma, R. Marotta, R. Andreozzi, V. Caprio, Thermal decomposition of of experimental adiabatic dataan update and extension, J. Loss Prev. Proc.
cumene hydroperoxide: chemical and kinetic characterization, AIChE J. 54 Ind. 33 (2015) 88100.
(2008) 15791584. [33] VSP II Manual and Methodology, Fauske and Associates, Burr Ridge, IL 2002.
[10] A. Miyake, Y. Ohama, Thermal hazard analysis of cumene hydroperoxide [34] ASTM E698-11 Standard Test Method for Arrhenius Kinetic Constants for
using calorimetry and spectroscopy, J. Therm. Anal. Calorim. 93 (2008) 5357. Thermally Unstable Materials Using Differential Scanning Calorimetry and
[11] K.-Y. Chen, S.-H. Wu, Y.-W. Wang, C.-M. Shu, Runaway reaction and thermal the Flynn/Wall/Ozawa Method.
hazards simulation of cumene hydroperoxide by DSC, J. Loss Prev. Proc. Ind. [35] Y.-S. Duh, J.-M. Yo, W.-L. Lee, C.-S. Kao, J.-M. Hsu, Thermal decompositions of
21 (2008) 101109. dialkyl peroxides studied by DSC, J. Them. Anal. Calorim. 118 (2014) 339347.
[12] J.-R. Chen, S.-H. Wu, S.-Y. Lin, H.-Y. Hou, C.-M. Shu, Utilization of [36] Thermal safety software by Cheminform St Petersburg Ltd., Russian.
microcalorimetry for an assessment of the potential for a runaway [37] Advanced Kinetics and Technology AG, Switzerland.
decomposition of cumene hydroperoxide at low temperatures, J. Therm. Anal. [38] A. Baignee, J.A. Howard, J.C. Scaiano, L.C. Stewart, Absolute rate constants for
Calorim. 93 (2008) 127133. reactions of cumyloxy in solution, J. Am. Chem. Soc. 105 (1983) 61206123.
[13] J.-R. Chen, S.-Y. Cheng, M.-H. Yuan, A.A. Kossoy, C.-M. Shu, Hierarchical kinetic [39] R. Rado, I. Chodak, V. Durdovic, Contribution to the kinetics of induced
simulation for autocatalytic decomposition of cumene hydroperoxide at low decomposition of cumyl hydroperoxide, React. Kinet. Catal. Lett. 2 (1975)
temperature, J. Therm. Anal. Calorim. 96 (2009) 751758. 7379.
[14] I. Ben Talouba, L. Balland, N. Mouhab, M.A. Abdelghani-Idrissi, Kinetic [40] I. Di Somma, R. Marotta, R. Andreozzi, V. Caprio, Detailed thermal and kinetic
parameter estimation for decomposition of organic peroxides by means of modeling of cumene hydroperoxide decomposition in cumene, Process Saf.
DSC measurements, J. Loss Prev. Proc. Ind. 24 (2011) 391396. Environ. Prot. 91 (2013) 262268.
[15] ASTM E2781-11 Standard Practice for Evaluation of Methods for [41] H.-C. Chou, S.-H. Wu, C.-C. Chiang, J.-J. Horng, J.-H. Chi, C.-M. Shu, Effects of
Determination of Kinetic Parameters by Thermal Analysis. stirring rate for thermal runaway reaction in cumene hydroperoxide
[16] R.L. Blaine, The search for kinetic reference materials for adiabatic and manufacturing process using calorimetric techniques, J. Therm. Anal. Calorim.
differential calorimetry, J. Therm. Anal. Calorim. 106 (2011) 2531. 106 (2011) 238243.
[17] Y. Lu, D. Ng, L. Miao, M. Sam Mannan, Key observations of cumene [42] S.-H. Wu, J.-T. Wu, Y.-T. Wu, J.-C. Lee, Y.-H. Huang, C.-M. Shu, Explosion
hydroperoxide concentration on runaway reaction parameters, Thermochim. evaluation and safety storage analyses of cumene hydroperoxide using
Acta 501 (2010) 6571. various calorimeters, J. Therm. Anal. Calorim. 111 (2013) 669675.
[18] Q. Wang, W.J. Rogers, M.S. Mannan, Thermal risk assessment and rankings for [43] S.-H. Wu, Thermal hazard investigation of cumene hydroperoxide in the rst
reaction hazards in process safety, J. Therm. Anal. Calorim. 98 (2009) 225233. oxidation tower, J. Therm. Anal. Calorim. 109 (2012) 921926.
[19] S.-H. Wu, C.-M. Shu, Reactive hazard analysis of cumene hydroperoxide and [44] F. Sun, F. Zhang, M.-P. Jin, N. Shi, W. Xu, Vent sizing of cumene hydroperoxide
dicumyl hydroperoxide, Process Saf. Prog. 29 (2010) 162165. (CHP) system under re scenario considering emergency ooding measure, J.
[20] O.J. Reyes Valdes, V. Casson Moreno, S.P. Waldram, L.N. Vechot, M.S. Mannan, Loss Prev. Proc. Ind. 32 (2014) 230237.
Experimental sensitivity analysis of the runaway severity of dicumyl peroxide [45] O.J. Reyes Valdes, V. Casson Moreno, M.S. Mannan, L.N. Vechot, Evaluation of
decomposition using adiabatic calorimetry, Thermochim. Acta 617 (2015) the thermal runaway decomposition of cumene hydroperoxide by adiabatic
2837. calorimetry, Chem. Eng. Trans. 43 (2015) 10091014.
[21] O.J. Reyes Valdes, V. Casson Moreno, S.P. Waldram, L.N. Vechot, M.S. Mannan, [46] J.J. Zwolenik, Photolytic generation and kinetic electron spin resonance
Runaway decomposition of dicumyl peroxide by open cell adiabatic testing at spectrometry of cumylperoxy radicals, J. Phys. Chem. 71 (1967) 24642469.
different initial conditions, Process Saf. Environ. Prot. 102 (2016) 251262.
[22] J. Lv, L. Chen, W. Chen, H. Gao, M. Peng, Kinetic analysis and self-accelerating
decomposition temperature (SADT) of dicumyl peroxide, Thermochim. Acta
571 (2013) 6063.
[23] R. Sanchirico, Adiabatic behavior of thermal unstable compounds evaluated
by means of dynamic scanning calorimetric (DSC) techniques, AIChE J. 59
(2013) 38063815.
[24] ASTM E2041-13, Standard test method for estimating kinetic parameters by
differential scanning calorimeter using the Borchardt and Daniels method.
[25] M.E. Brown, M. Mciejewski, S. Vyazovkin, R. Nomen, J. Sempere, A. Burnham,
J. Opfermann, R. Srrey, H.L. Anderson, A. Kemmler, R. Keuleers, J.J. Hanssens,
H.O. Desseyn, C.-R. Lee, T.B. Tang, B. Roduit, J. Malek, T. Mitsuhashi,
Computational aspects of kinetic analysis Part A: the ICTAC kinetics

Você também pode gostar