Controlled Drug Release of Oral Dosage Forms

CONTROLLED DRUG RELEASE OF ORAL
DOSAGE FORMS
ELLIS HORWOOD SERIES IN PHARMACEUTICAL

TECHNOLOGY
incorporating
Pharmacological
Sciences
Series Editor: Michael H Rubinstein, Professor of Pharmaceutical Technology, School of
Liverpool John Moores Universi
Pharmac
UNDERSTANDIN 2 EKPERIMENTAL DESIGN AND INTERPRETATION IN
Armstrongklames
PHARMACEUTICS
(Available in Cloth and Paoerbaok)

MICROBIAL QUALITY ASSURANCE IN
PHARMACEUTICALS,
COSMETICS, TOILETRIES
AUTONOMIC PHARMACOLOGY
Broadley
PHARMACEUTICAL
PRODUCT LICENSING: Requirements
for Europe
Cartwright & Matthews
INTERNATIONAL PHARMACEUTICAL
PRODUCT REGISTRATION:
Cartwright 8 Matthews
Quality, Safety, Efftcacy
DRUG DISCOVERY TECHNOLOGIES
Clark & Moos
PHARMACEUTICAL
PRODUCTION FACILITIES: Design and Application
Cole
PHARMACEUTICAL
TABLET COATING TECHNOLOGY
E;;k Hogan & Aulton
POTASSIUM CHANNELS: Structure,
Classification,
Function
and Therapeutic
Potential
DIELECTRIC ANALYSIS OF PHARMACEUTICAL
SYSTEMS
Craig & Newton
PHARMACY AND PHARMACOTHERAPY
OF ASTHMA
DArcy & McElnay
GUIDE TO MICROBIOLOGICAL
CONTROL IN PHARMACEUTICS
Denyer & Baird
RECEPTOR DATA FOR BIOLOGICAL EKPERIMENTS: A Guide to DNg Selectivity
Doods 8 Van Meel
POTASSIUM CHANNEL MODULATOR DRUGS:
Evans et al.
From Synthesis to Clinical Experience
DRUG
RESISTANCE
IN VIRUSES: Principles, Mechanisms
Field & Goldthorpe
and Clinical Perspectives
Ford & Timmins
PHARMACEUTICAL
THERMAL ANALYSIS: Techniques
and Applications
DICTIONARY OF ANTIBIOTIC-PRODUCING
ORGANISMS
Glasby
PHYSICOCHEMICAL
PROPERTIES OF DRUGS: A Handbook for
Gould
Pharmaceutical
Scientists
DRUG DELIVERY TO THE GASTROINTESTINAL
TRACT
Hardy et al.
DRUGS FROM NATURAL PRODUCTS: Pharmaceuticals
and Agrochemioals
Harvey
POLYPEPTIDE AND PROTEIN DRUGS: Production,
Characterixation,
Formulation
;I;
& Barlow
DRUG CHIRALITY: Chemistry,
Biology, Regulatlon, Industrial Issues
_.
DRUGS, DIET AND DISEASE, Volume 1: Mechanistic
Approaches
to Cancer
loannrdes & Lewis
Structure,
Distribution
and Function
lzquierdo & Medina NATURALLY OCCURRING BENZODIAZEPINES:
DRUG TARGETING AND DELIVERY: Concepts in Dosage Form Design
Junginger
EXCITATORY AMINO ACID RECEPTORS:
Krogsoaard-Larsen
& Hansen
- Design of Agonists and Antagonists
ADVANCED DRUG DESIGN AND DEVELOPMENT
Kourounakis & Rekka
STEROIDS. DRUG RESPONSE AND METABOLISM:
Kourounakis & Rekka
Pharmacochemical
Approach to Defensive Steroids
HANDBOOK OF PHARYACOKINETICS:
Labaune
The Toxicity Asssessment
of Chemicals
IMMUNOASSAY PROCEDURES: A Practical Guide
Law
BIOPHARMACEUTICS
OF
ORALLY
ADMINISTERED
DRUGS
&
Dressman
Macheras, Reppas
PEPTIDE HORMONES AS PROHORMONES
Martinez
ANTI-RHEUMATIC
DRUGS: Actions and Side Effects
Rainsford
PHARMACEUTICAL
DESIGN AND DEVELOPMENT: A Molecular Biological Approach
Ramabhadran
TABLET MACHINE INSTRUMENTATION
IN PHARMACEUTICS;
Principles and Practice
Ridgway Wan
PHARMACEUTICAL
CHEMISTRY, Volume I: Drug Synthesis
Roth et al.
PHARMACEUTICAL
CHEMISTRY, Volume 2: Drug Analysis
Roth et al.
PHARMACEUTICAL
TECHNOLOGY Controlled DNg Release, Volume 1:
Rubinstein
PHARMACEUTICAL
TECHNOLOGY Tabletina Technoloav. Volume 1:
Rubinstein
PHARMACEUTICAL
TECHNOLDGY D-&g Stability
Rubinstein
UNDERSTANDING ANTIBACTERIAL
ACTION AND RESISTANCE (Cloth 8 Paper)
Russell & Chopra
PHARMACEUTICAL
SPECIFICATIONS:
Standards for Drugs
Rutherford
MODERN MEDICINAL CHEMISTRY fAvailable in Cloth and Paoerback)
Tavlor & Kennewell
RADIOPHARMACEUTICALS:
Using Radioactive Compounds
in Pharmaceutics
and Medicine
CHEMISTRY FOR PHARMACY, PHARMACOLOGY AND
Thomas & Thurston
THE HEALTH SCIENCES (Available in Cloth and Paperback)
CLINICAL TRIALS FOR THE PHARMACEUTICAL
INDUSTRY
Tweed
Van Meel, Hauel &Shelley
CARDIOTONIC AGENTS FOR THE TREATMENT OF HEART FAlLURE
CONTROLLED DRUG RELEASE OF ORAL DOSAGE FORMS
Vergnaud
Washington
PARTICLE SIZE ANALYSIS IN PHARMACEUTICS
AND OTHER INDUSTRIES
PHARMACOKINETIC
MODELLING USING STELLA
Washington et al.
ON THE APPLE MACINTOSH (TM)
PHARMACEUTICAL
PREFORMULATION
Wells
PHARMACEUTICAL
TECHNOLOGY Controlled
Drug Release, Volume 2
Wells & Rubinstein
Wells & Rubinstein
PHARMACEUTICAL
TECHNOLOGY Tableting Technology, Volume 2
PHYSIOLOGICAL PHARMACEUTICS:
Wilson &Washington
Biological Barriers to Drug Absorption
Bloomfield et al
The above is a complete list of all Ellis Horwood titles in the pharmaceutical and pharmacological sciences, both
published and in preparation. Further details can be obtained from Simon and Schuster International Group 0442 881900.
CONTROLLED DRUG
RELEASE OF ORAL
DOSAGE FORMS
Professor JEAN-MAURICE VERGNAUD

Faculty of Sciences, University of Saint-Etienne,
France
ELLIS HORWOOD
NEWYORK
LONDON TORONTO SYDNEY TOKYO SINGAPORE
First published in 1993 by

Ellis Horwood Limited
Market Cross House, Cooper Street,
Chichester, West Sussex, PO19 IEB, England
A division of
Simon & Schuster International Group
0 Ellis Horwood Limited, 1993
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form, or by any means, electronic, mechanical, photocopying, recording or
otherwise, without prior permission, in writing, from the publisher
Printed and bound in Great Britain
by Bookcraft, Midsomer Norton
British Library Cataloguing in Fublication Data
A catalogue record for this book is available from the British Library
ISBN O-13-1749564
Library of Congress Cataloging-in-Publication Dam
Available from the publisher
Table of contents
PREFACE . . . . . . . . . . . . . . . . . 0. . . . . . . . . . . . m. . . . . . . . . . . . m. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ACKNOWLEDGEMENTS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 THE DIFFUSION EQUATIONS AND BASIC CONSIDERATIONS ...............

Introduction.. ................................................................................
1.1
1 .I .I
Process of diffusion.. .......................................................
1 .I .2
Diffusion of a substance through a polymer.. ......................
Polymer in the rubbery state (Case I). ................................
Polymer in the glassy state (Case II). .................................
Absorption of liquid in Case Ill.. ........................................
1 .l .3
Steady and non-steady conditions.. ...................................
1 .I .4
Initial conditions.. ............................................................
1 .I .5
Boundary conditions.. ......................................................
1 .I .6
Volume of the surrounding atmosphere and partition factor . .
Equations of diffusion for various shapes ..........................................
i.2
1.2.1
Equations of diffusion for a thin sheet.. ..............................
1.2.2
Equations of diffusion for a rectangular parallelepiped ..........
Isotropic materials.. .........................................................
Anisotropic materials.. .....................................................
1.2.3
Cylinder of infinite and finite length.. .................................
Cylinder of infinite length.. ...............................................
Cylinder of finite length.. ..................................................
1.2.4
Radial diffusion in a sphere.. .............................................
Methods of solution when the diffusivity
is constant.. .......................
1.3
1.3.1
Kinds of solution .............................................................
1 .3.2
Method of separation of variables.. ....................................
1.3.3
Method for the Laplace transform.. ....................................
1 .3.4
Method of reflection and superposition.. ............................
Plane source ...................................................................
Reflection at a boundary.. ................................................
Extended initial distribution of the substance.. ....................
2 MATHEMATICAL
Introduction..
2.1
TREATMENT OF DIFFUSION IN A PLANE SHEET .........

................................................................................
vii
. ..
XIII
1
1
2
2
2
3
3
3
4
4
5
6
6
7
7
8
9
9
9
10
10
10
11
15
17
17
19
19
21
21
Table of contents
ii
2.2
2.3
2.4
2.5
Non-steady state with a high coefficient of matter transfer on the

surface and an infinite volume of the surrounding.. ............................
Uniform initial distribution in the sheet.. .............................
2.2.1
2.2.2
Initial distribution f(x) in the sheet of thickness L.. ..............
Two different media.. ......................................................
The two media are identical.. ............................................
2.2.3
Non-steady state with a membrane with a uniform initial
distribution and surface concentration
different.. .................
Non-steady state with a finite coefficient of matter transfer on the
surface.. .......................................................................................
Non-steady state diffusion in a stirred solution of limited volume .........
2.4.1
Absorption of diffusing substance by the sheet.. .................
2.4.2
Desorption of diffusing substance from the sheet in the
solution ..........................................................................
Steady-state
with a membrane.. ......................................................
2.5.1
High value of the coefficient of matter transfer on the
surf ace ...........................................................................
2.5.2
Case of finite value of the coefficient of matter transfer .......
2.5.3
Composite membrane ......................................................
2.5.4
Membrane separating gases or vapour.. .............................
MATHEMATICAL
TREATMENT OF DIFFUSION IN AN ISOTROPIC
RECTANGULAR PARALLELEPIPED .......................................................
Introduction.. ................................................................................
3.1
Isotropic
rectangular parallelepiped with a constant diffusivity ...........
3.2
3.2.1
Infinite coefficient of matter transfer on the surface.
Constant concentration
on the surface.. .............................
Use of trigonometrical
series.. ...........................................
Use of error function.. ......................................................
3.2.2
Finite coefficient of matter transfer on the surface.. ............
Effect of the thickness of the sheet.. ...............................................
3.3
22
22
31
31
32
34
37
40
41
43
45
45
46
47
48
49
49
50
50
51
52
55
58
4 MATHEMATICAL
TREATMENT OF RADIAL DIFFUSION IN A SPHERE.....
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1
4.2
Solid sphere in non-steady state with constant diffusivity . . . . . . . . . . . . . . . . . .
Infinite coefficient of matter transfer on the surface . . . . . . . . . . . .
4.2.1
4.2.2
Finite coefficient of matter transfer on the surface . . . . . . . . . . . . . .
4.2.3
Diffusion between a sphere and a well-stirred surrounding
atmosphere of finite volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3
Hollow sphere in non-steady state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hollow sphere in steady state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4
4.4.1
Hollow sphere with a constant concentration
on each
surface.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.2
Hollow sphere with a constant concentration
of the internal
surface and a finite coefficient of matter transfer on the
external surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
59
60
60
62
5 MATHEMATICAL
TREATMENT OF DIFFUSION IN CYLINDERS . . . . . . . . . . . . . . . .
5.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2
Non-steady state with a solid cylinder of infinite length . . . . . . . . . . . . . . . . . . . . . .
5.2.1
on the surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2
5.2.3
Solid cylinder in a well-stirred surrounding of finite volume...
75
75
77
77
81
86
66
69
71
72
73
Table of contents
5.3
5.4
5.5
5.6
6
6.1
6.2
6.3
6.4
6.5
6.6
7
7.1
7.2
7.3
7.4
...
111
Non-steady state with a solid cylinder of finite length.. ......................

5.3.1
on the surface ...............................
5.3.2
Finite coefficient of matter transfer on each surface ............
Non-steady state with a hollow cylinder of infinite length.. .................
5.4.1
Surface concentration
constant and equal on each surface ...
5.4.2
Constant concentrations
on each surface.. .........................
Steady state with a hollow cylinder of infinite length.. .......................
5.5.1
Constant concentrations
on each surface.. .........................
5.5.2
on the internal surface and a finite
coefficient of matter transfer on the external surface ..........
5.5.3
Composite hollow cylinder ................................................
Conclusions ..................................................................................
101
101
102
NUMERICAL ANALYSIS WITH ONE-DIMENSIONAL

DIFFUSION
THROUGH A PLANE SHEET ................................................................
Introduction.. .................................................................................
Diffusion through a sheet with constant diffusivity.. ..........................
Infinite coefficient of matter transfer on the surface .............
6.2.1
6.2.2
Finite coefficient of matter transfer on the surface ..............
Diffusion through a sheet with concentration-dependent
diffusivity .....
6.3.1
Infinite coefficient of matter transfer on the surface.. ..........
6.3.2
Finite coefficient of matter transfer on the surface.. .............
Membrane separating two different media.. ......................................
6.4.1
Infinite coefficient of matter transfer on each surface ..........
6.4.2
Finite coefficient of matter transfer on each surface .............
Diffusion
between two different sheets ...........................................
Constant diffusivities.. .....................................................
6.5.1
6.5.2
Concentration-dependent
diffusivities.. ..............................
Transfer with special conditions.. ....................................................
6.6.1
Programmation
of temperature.. ........................................
6.6.2
Programmation
of the concentration
in the surrounding ........
105
105
106
107
110
113
113
114
118
118
119
122
122
125
126
126
127
NUMERICAL ANALYSIS WITH A RECTANGULAR PARALLELEPIPED,

TRANSFER.. ..........................................
AND A THREE-DIMENSIONAL
Introduction.. ................................................................................
Transfer through a rectangular parallelepiped with a constant
concentration
on the surface.. .........................................................
7.2.1
Constant diffusivity.. ........................................................
7.2.2
diffusivity.. .................................
Transfer through a rectangular parallelepiped with a finite coefficient
of matter transfer on the surface.. ...................................................
7.3.1
Constant diffusivity.. ........................................................
7.3.2
diffusivities,
and finite coefficient
of matter transfer on the surface.. .....................................
Transfer with special conditions.. .....................................................
7.4.1
Anisotropic material.. .......................................................
of temperature.. ........................................
7.4.2
Programmation
7.4.3
Programmation
of the concentration
in the surrounding ........
NUMERICAL ANALYSIS WITH A RADIAL TRANSPORT WITHIN A

SPHERE.. ..........................................................................................
Introduction.. . . . . . . . . . . . . . . . . . . . . . . . ,.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . .
8.1
90
90
93
95
97
98
99
99
129
129
131
132
135
137
137
142
146
146
146
147
149
149
iv
8.2
8.3
8.4
8.5
8.6
Table of contents
Radial diffusion through a sphere with constant diffusivity ..,..............
8.2.1
8.2.2
Radial diffusion through a sphere with concentration-dependent
diffusivity . . . . . . . . . . . . . . . . . . e. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.3.1
8.3.2
Hollow sphere with constant concentration
on the internal surface,
and constant diffusivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.4.1
Infinite coefficient of matter transfer on the external
surf ace.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . .
8.4.2
Finite coefficient of matter transfer on the external surface..
Hollow sphere with constant concentration
on the internal surface,
and concentration-dependent
diffusivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.5.1
Infinite coefficient of matter transfer on the external
surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.5.2
Finite coefficient of matter transfer on the external surface..
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ,...... . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9 NUMERICAL ANALYSIS WITH CYLINDERS.. ........................................

Introduction.. ................................................................................
9.1
Solid cylinder of infinite length.. ......................................................
9.2
9.2.1
Constant diffusivity .........................................................
9.2.2
diffusivity ..................................
Hollow cylinder of infinite length.. ...................................................
9.3
9.3.1
Constant diffusivity.. .......................................................
9.3.2
diffusivity.. ................................
Solid cylinder of finite length.. .........................................................
9.4
9.4.1
Constant diffusivity.. .......................................................
9.4.2
diffusivity.. ................................
Conclusions ..................................................................................
9.5
10 DRUG DELIVERY FROM DOSAGE FORMS CONSISTING OF A DRUG
DISPERSED IN A NON-ERODIBLE POLYMER.. .......................................
and definitions.. ...........................................................
IO. 1 Introduction
10.1 .I Problems of the drug passing through the body.. .................
10.1.2
Pharmacokinetics,
pharmacodynamics
and
biopharmaceutics..
..........................................................
10.1 .3 Conventional dosage forms ..............................................
10.1 .4 Oral therapeutic systems.. ................................................
IO. 1.5 Simple monolithic devices with a polymer matrix.. ..............
10.1.6
Processes with a double matter transfer.. ...........................
10.2 Drug-Eudragit sheet in gastric liquid.. ...............................................
10.2.1
Introduction.. ..................................................................
10.2.2 Theoretical aspects.. ........................................................
10.2.3
Experimental.. .................................................................
10.2.4
Results obtained with Eudragit as polymer matrix.. ..............
10.2.5
Conclusions with Eudragit as polymer matrix.. .....................
10.3 Drug-Carbopol sheet in gastric liquid.. ..............................................
10.3.1
Introduction.. ..................................................................
10.3.2 Theoretical aspects ..........................................................
10.3.3
Experimental.. .................................................................
10.3.4
Results with Drug-Carbopol devices.. .................................
150
150
154
155
156
157
159
160
161
162
162
163
164
167
168
169
173
176
177
177
179
181
181
193
196
199
199
199
203
205
206
208
212
215
215
216
219
221
224
225
225
226
226
227
Table of contents
10.4
10.5
10.3.5
Effect of
10.4.1
10.4.2
10.4.3
10.4.4
10.4.5
Spherical
10.5.1
10.5.2
10.5.3
10.5.4
10.5.5
Conclusions with Drug-Carbopol devices.. ...........................

pH on drug release.. ..........................................................
Introduction .....................................................................
Theoretical aspects.. ........................................................
Experimental.. ..................................................................
Results with a Drug-Eudragit sheets.. ................................
Conclusions.. ...................................................................
Drug-Eudragit beads in gastric liquid.. .................................
Introduction.. ...................................................................
Theoretical aspects ..........................................................
Experimental.. ..................................................................
Results with Drug-Eudragit spheres.. ..................................
Conclusions.. ...................................................................
11 DRYING OF DOSAGE FORMS MADE OF A DRUG DISPERSED IN A

POLYMER MATRIX ............................................................................
1 1 .I Introduction.. ................................................................................
11.2 Drying dosage forms in a surrounding atmosphere of infinite volume . .
1 1 .2.1 Theoretical.. ...................................................................
1 1 .2.2 Experimental ...................................................................
11 .2.3 Results with a constant temperature.. ...............................
1 1.2.4 Effect of temperature.. .....................................................
1 1 .3 Drying with a programmed temperature .........................................
1 1 .3.1 Theoretical.. ...................................................................
1 1 .3.2 Experimental.. .................................................................
temperature.. .........................
1 1 .3.3 Results with a programmed
11.4 Drying in a surrounding atmosphere of finite volume.. ........................
1 1.4.1 Theoretical of drying in a surrounding atmosphere of finite
volume.. .........................................................................
1 1 .4.2 Experimental.. .................................................................
11 .4.3 Results.. .........................................................................
1 1 .5 Drying with a controlled vapour pressure.. ........................................
11.5.1 Theory of the process with controlled vapour pressure ........
1 1 .5.2 Simulation of the process.. ...............................................
1 1 .6 Conclusions.. ................................................................................
230
231
231
232
233
234
240
242
242
242
246
246
251
261
261
263
263
267
268
274
277
279
280
280
286
286
291
291
300
301
302
308
12 DRUG DELIVERY FROM DOSAGE FORMS CONSISTING OF A DRUG

DISPERSED IN AN ERODIBLE POLYMER.. .............................................
12.1 Introduction.. ................................................................................
12.2 Theoretical aspects ........................................................................
1 2.2.1 Diffusional process.. ........................................................
12.2.2
Polymer erosion is the driving force.. .................................
1 2.3 Experiments.. ................................................................................
1 2.4 Results.. .......................................................................................
12.4.1
Results with the diffusion process.. ...................................
12.4.2
Results with the erosion process.. .....................................
12.5 Conclusions.. ................................................................................
313
313
314
314
317
318
320
320
322
327
13 DOSAGE FORMS MADE OF A CORE AND SHELL, WITH AN ERODIBLE

SHELL. CONSTANT RATE OF DELIVERY ..............................................
13.1 Introduction.. ................................................................................
13.2 Theoretical aspects.. ......................................................................
13.3 Experimental.. ...............................................................................
329
329
330
333
vi
13.4
Table of contents
Results.. .......................................................................................
13.4.1
Results with sodium salicycate.. .......................................
13.4.2
Results with sulfanilamide.. ..............................................
Conclusions.. ................................................................................
334
334
338
341
14 DOSAGE FORMS MADE OF CORE AND SHELL, WITH A NON-ERODIBLE

POLYMER.. .......................................................................................
14.1 Introduction.. ................................................................................
14.2 Theoretical.. ..................................................................................
14.3 Experimental.. ...............................................................................
14.4 Results.. .......................................................................................
14.4.1
Data.. ............................................................................
14.4.2 Validity of the model.. ......................................................
14.4.3
Effect of parameters.. ......................................................
14.4.4
Profiles of concentration.. ................................................
14.5 Conclusions.. .................................................................................
345
345
346
352
352
353
354
354
357
358
13.5
15 CONTROLLED RATE OF DELIVERY WHEN THE SOLUBILITY OF THE

DRUG IS LOW, BY USING A SWELLING POLYMER.. ..............................
15.1 Introduction.. ................................................................................
15.2 Dosage form with a polymer matrix and a swelling polymer.. ..............
1 5.2.1 Theoretical.. ...................................................................
15.2.2
Experimental.. .................................................................
15.2.3
Results.. .........................................................................
15.3 Dosage forms with gelucire and a swelling polymer.. .........................
1 5.3.1 Theoretical.. ...................................................................
15.3.2
Experimental.. .................................................................
15.3.3
Results and discussion.. ....................................................
15.4 Dosage form with a core (swelling Polymer-Drug-Eudragit)
and an
erodible polymer.. ..........................................................................
15.4.1 Theoretical.. ...................................................................
15.4.2
Experimental.. .................................................................
15.4.3
Results.. .........................................................................
15.5 Conclusions ..................................................................................
363
363
364
364
365
366
373
373
373
376
380
380
382
383
390
16 DOSAGE FORMS WITH A DRUG ATTACHED TO A POLYMER

DISPERSED IN A NON-ERODIBLE POLYMER MATRIX ............................
16.1 Introduction.. ................................................................................
16.2 Theoretical.. ..................................................................................
16.3 Experimental.. ...............................................................................
16.4 Results.. .......................................................................................
16.5 Conclusions ..................................................................................
393
393
394
395
397
404
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..~...........................................................
409
PREFACE
Therapeutic systemsrepresent a new route for drug administration: as the drug is

delivered continuously at a controlled rate over a predeterminedperiod of time, uniform
and constantblood level is achieved,smaller amount of drug is neededreducing the side
effects, and the therapy is improved. Various people beyond the patient are concerned
with this therapy,physicians andpharmacistsin various areasof specialization,of course,
but also bio-engineersand even workers in chemicalengineering.Various oral therapeutic
systemsconsist of a polymer matrix through which the drug is dispersed,and thus good
knowledge of the matter transfers through the polymer is necessarywhen they are in
contact with the gastric or the intestine liquid.These matter transfers being controlled by
transient diffusion, the mathematical treatment of diffusion must be known when it is
feasible in simple casesand especiallyfor constantdiffusivity. Morever, in complex cases
and when the diffusivity is concentration-dependent, numerical methods with finite
differences must be used instead of the mathematical treatment. Finally, in order to
accustom the users with these ways of calculation, various mathematical or numerical
models are built andtestedin the study of different oral dosageforms. Thus a new way of
working is developed, coupling the experiments with the models of the process, these
experimentsbeing performed in so called in-vitro tests which simulate the conditions in
the body asmuch aspossible.
The drug, by using this term in the senseof a biologically active substance,is a
chemical compound administered to the patients organism, with which it develops a
reciprocal interaction for therapeuticpurposes.Generally, for many reasons,the drug is
not used in the pure state.The supply form of presentationof the drug, or dosageform, is
the completemedication.Conventionaldosageforms consistof the drug, the active agent,
.
VIII
Preface
and auxiliary substancesbiologically inert, the excipients. The role of excipients is

essentially of binding the drug, filling the dosageform in order to ensurethe consistency
and volume necessaryfor the patient use.When in contactwith the gastric liquid, the drug
is releasedfrom the dosageform used for the administration. Two factors determine the
releaseof the drug: the solubility and the rate of dissolution. When the drug is released
from the dosageform, it must pass through several barriers, before reaching the site of
action. The driving force responsible for the transport of the drug through these
membranes is the concentration gradient across the membrane, the process being
controlled by diffusion. The amount of unchangeddrug that is absorbedby the organism
in a certain time and that arrives at the target site through the circulatory system, or
bioavailability, dependson the dosageform, and thus may be alteredby this dosageform.
The drug conveyed by the bloodstream, leaving the intravascular compartment is
distributed between extracellular and intracellular compartment where it can reach the
receptorsfor drugs lying in the tissues.Finally the drug is eliminated either by chemical
alteration of the molecule with formation of metabolites or by excretion via various
organs. The time required for elimination of half the plasma content of the drug by
metabolism or excretion, the biological half-time, is of high interest for the dosage
regimen prescribed by the physician. An exact dosage regimen is of high importance
when the concentrationof the drug must be maintained constantin the tissuesover a long
period of time. Morever, the concentrationof the drug must be kept between the median
lethal dose(causing the deathsof 50% of experimental animals) and the median effective
dose (effective in 50% of cases).The therapeutic index, equal to the ratio of these above
concentrationsdefinesthe safetymargin.
All conventional dosageforms made of a drug dispersedin excipients, releasethe
drug according to the following pattern. The drug is very rapidly dissolved from the
dosagefonn and quickly builds up to a maximum high concentration, which then falls
exponentially with tune becauseof the first order absorption.The result is an undulating
concentrationof the drug in the stomachor intestine, as well as in the blood and tissues,
where high concentrations with overdosages alternate with low concentrations and
underdosages.The limitations of conventional dosageforms made of drug and excipients
appearthen since they causeproblems in maintaining therapeutic drug levels over only
brief durationsof tune :
(i) The fluctuating drug levels with conventional dosageforms lead to an insufficient
efficacy of therapy provoking an excessiveuse of the drug.
Preface
ix
(ii) Overdosage appearing after dissolution of the drug may be responsible for a high
frequencyof side effects,leading to iatrogenicdamage.
(iii) High frequency of administration of conventional dosage forms is limited by the
reliability of the patient andthe patient compliance(omission,wrong frequency)
(iv) A potent drug may largely lose its therapeuticefficacy through improper formulation,
and thus a pharmacologicallyactive substanceis not necessarilyan effective drug.
Oral dosage systems able to release the drug at a constant rate for a given time
period are thus of mterest.The result is then a constantuniform concentrationof drug in
blood and tissuesover a given period of time, with the following advantages:
(i) Significant smaller amountsof drug are generally prescribedwith a therapeuticsystem
of drug delivery.
(ii) The reduced amount of drug administered reduces the problems of side effects,
improving the safety of therapy.
(iii) The patient compliance is usually better with these types of dosage forms, as the
frequencyof administration is considerablylower.
Simple oral dosageforms capableof controlling the releaseof the drug are often and
easily obtained with monolithic devices where the drug is dispersedin a biocompatible
polymer. This polymer which cau be either biodegradableor non degradable,plays the
role of a polymer matrix. Not only the polymer brings the consistency to the dosage
form, but also it controls the releaseof the drug. The processis generally as follows: the
liquid (gastric liquid or intestineliquid) entersthe polymer, dissolvesthe drug and enables
the drug to leave out the dosageform through the liquid located in the dosageform. The
matter transfers for the liquid and for the drug are controlled by transient diffusion, with
concentration-dependent diffusivities, the diffusivity of the drug depending on the
concentrationof the liquid in the dosageform. The releaseof the drug being controlled by
transient diffusion, exhibits a rather high rate at the beginning of the process which
decreaseswith time in an exponentialway. Thesedosageforms arevery simple to prepare
and rather inexpensive, but the processof releaseis controlled by diffusion, and the rate
of releaseis far from being constant.
The drug delivery from the dosageform is studied by using in-vitro tests,theseinvitro testsbeing built up in such a way that they simulate as much aspossible the story in
the stomach or intestine of the patient. These in-vitro tests are very useful for many
Preface
reasons,and the most obvious are only given :

(i) The conditions of the in-vitro test are very well defined and standardized,enabling
comparisonsbetweenvarious results.
(ii) They are easyto perform, and the effect of eachparametercan be analysedseparately.
(iii) In contrastwith the in-vitro test, the in-vivo test is far more complex, asthis latter is
subjectto a variety of influencesthat differ greatly amongindividuals.
There are several objectives in this book devoted to the study of the process of
matter transfers in oral dosageforms with a polymer matrix able to control the releaseof
the drug. As the driving force for the matter transfers of the liquid and the drug through
the polymer is the gradient of concentration, the process is controlled by transient
diffusion. Some emphasisis thus placed upon the mathematicaltreatment of diffusion in
solids of various shapes,when the processis so simple that an analytical solution exists.
As very often the process of matter transfers is rather complex, it must be studied by
using numerical methodswith finite differences. Finally, various examplesare described
by consideringsimple oral dosageforms with either a non-erodibleor an erodible polymer
matrix, and with more complex systemsconsisting of a core and shell. These studiesare
made by using the method coupling experimentswith short tests and long real tests and
modelling of the process.
The book is divided in threeparts with sixteenchapters:
The first part presents an overview of the mathematical treatment of diffusion
through a polymer in the elastomeric state.Various shapesare consideredfor the solid :
thin plane sheets,rectangular parallelepiped, cylinders and spheres.In order to help the
readers understanding, some emphasis is placed upon the conditions in which the
mathematical treatment is feasible constant diffusivity, uniform initial concentration,
simple boundary conditions. For people wanting to improve their backgroundknowledge
of the mathematical treatment of diffusion, various examplesare describedin a didactic
way in the first five chapters.Specialconsiderationis given to the operationalconditions :
with a very high volume of the liquid in which the dosageform is immersed, or with a
finite volume of this liquid ; with a very high coefficient of matter transfer on the surface
leading to a constant concentration on the surface, or with a finite coefficient of matter
transfer on the surface.
- In chapter 1, general equations of diffusion are given for various shapesof the dosage
form. and basic considerationsare described.
Preface
xi
- In chapter 2, the mathematical treatment of diffusion is shown in various caseswith a

plane sheetand mono-directionaldiffusion.
- In chapter 3, the mathematical treatment of diffusion is given with a rectangular
parallelepipedand three-dimensionaldiffusion.
- In chapter4, radial diffusion through spheresis studied.
- In chapter 5, cylinders of infinite and finite lengths are consideredwith radial diffusion
in the first caseand radial andlongitudinal diffusion in the secondcase.
The secondpart is devotedto numerical treatmentof diffusion, in order to accustom
the readersto this new and powerful way of working. This method is very useful, as very
often no analytical solution can be obtained from the mathematicaltreatment,becauseof
the complexity of the process : double matter transfers of the liquid and drug,
concentration-dependent
diffusivity. Explicit numericalmethodswith finite differencesare
developed, becauseof their easy use with microcomputers. Four chaptersenable one to
consider various shapes.
- In chapter 6, plane thin sheets are considered and classical examples of numerical
analysis are developedin the following simple cases: the diffusivity is either constantor
concentration-dependent,while various values of the coefficient of matter transfer on the
surfaceare given.
- In chapter 7, numerical analysis is developed with a rectangular parallelepiped and a
three-dimensionaltransfer.
- In chapter8, numerical analysisfor the radial transferthrough a sphereis presented.
- In chapter 9, the matter transfers, either radial with long cylinders or radial and
longitudinal with cylinders of finite length, are studied with the help of numerical
analysis.
The third part examines various approchesto industrial problems with practical
purposes. A new method coupling experiments and modelling of the process is widely
used.
Experimentsareusedfor the following reasons:
- to get deepknowledge of the process
- to obtain the values of parameters,such as the diffusivities by using short tests
- to test the validity of the models.
Modelling of the process is widely used, either with the mathematical treatment
when the problem is simple, or with the numerical treatmentwhen the processis complex.
Each of these different casesare discussedin chapters 10 to 16, working through
the difficulties encounteredwith experimentsand calculation.
xii
Preface
- Chapter 10 concentrateson the drug delivery from simple dosageforms consisting of a

drug dispersedin a non-erodible polymer. Two matter transfers are consideredwith the
liquid entering the polymer, dissolving the drug and enabling the drug to leave the dosage
form through the liquid located in the polymer. These two transfers are connectedwith
eachother, and the diffusivity of the drug dependson the liquid concentration.
- Chapter 11 shows the complexity of the process of drying of dosage forms with a
polymer matrix, the process being controlled not only by evaporation but also by
diffusion of the liquid through the polymer. Various examplesare describedand the effect
of factors such as the temperatureor programmation of temperature,the pressureof the
vapour in the surrounding atmosphere,is evaluated.
- Chapter 12 discussesthe problem of drug delivery from dosageforms made of a drug
dispersedin an erodible polyme matrix.
- Chapter 13 focuses on the interest of preparing dosage forms with constant rate of
delivery. Typical dosageforms are presentedwith a core containing the drug dispersedin
a polymer andwith an erodible shell surroundingthe core.
- Chapter 14 deals with dosageforms made of a core and shell, where the core contains
the drug dispersedin a non-erodiblepolymer and the shell is a non-erodiblepolymer. The
effect of the relative thicknessof the shell is of high interest.
- Chapter 15 is devoted to special dosageforms able to deliver the drug from the dosage
form when the drug is poorly soluble in the liquid. A swelling polymer is thus addedin
the erodible polymer matrix which helps the disintegration of the dosageform and thus
disseminationof the drug in the liquid. As somepolymers swell differently in gastric and
intestine liquid, they allow the dosageform to deliver the drug partly in the stomach and
intestine.
- Chapter 16 examines the problem of dosage forms where the drug is attached to a
polymer, this branchedpolymer being dispersedin a polymer matrix.
ACKNOWLEDGEMENTS
A large part of this book covers various applications and industrial problems, as
many people working in industrial firms have influenced this work through industrial
contracts.I am glad to thank them for their interestingcooperation.
Many colleaguesand studentshave supportedmy efforts and brought contributions
worth noting.
Deep gratitude is extendedto my colleaguesM. Rollet who showedme round the
world of galenic pharmacy as well as J. Bardon and C. Chaumat. I am grateful for the
collaboration of my colleague J. L. Taverdet in the work concernedwith the preparation
and studies of dosageforms. I give my best thanks to my colleaguesJ. Bouzon for his
participation in numerical analysis and modelling of the process, and J. P. Montheard
who dealt with the polymerization problemsin chapter 16. I appreciatethe kind help of H.
Liu and J. Paulet.
My best appreciationis given to my students:
Y. Armand, D. Bidah, N. Chaffi, A. Droin, A. Eddine, N. Farah, M. Kolli, N.
Laghoueg, F. Magnard, P. Magron, Y. Malley, E.M. Ouriemchi, M. Saber, who did
their best for preparating their Theses.Many thanks to D. Berthet for his efficient help in
calculation and for his drawings, and to C. Cervantes, N. Fauvet, D. Ianna and M.
Novais Da Costafor their competenttyping of the manuscript.
1
The diffusion
considerations
1 .I
1 .l.l
equations
and basic
INTRODUCTION
PROCESS
OF DIFFUSION
Generally diffusion is the process through which matter is transferred from one
place to another, resulting from random molecular motions. Of course, on the average, the
matter is transferred by diffusion from the region of higher to that of lower concentration
of the matter. The example of diffusion of a drop of dye in motionless water is a good
example.
Transfer of heat by conduction is also due to random molecular motions transferring
kinetic energy, and there is some analogy between these two processes of matter and heat
transfers. The mathematical equation of heat conduction was established by Fourier in
1822. A few decades later, Fick in 1855, recognizing this analogy, put diffusion on a
quantitative basis by adopting the same equation. In an isotropic substance the rate of
transfer of diffusing substancethrough unit area of a section is proportional to the gradient
of concentration measured normal to this section.
(1.1)
F= -D
where g
ax
is the gradient of concentration C of the substance along the x-axis of diffusion,
F is the rate of transfer per unit area of the section perpendicular to the x-axis,
and the coefficient D is called the diffusion coefficient of diffusivity.
The diffusion
equations
and basic considerations
[Ch. 1
The term diffusivity will be usedin the book. The negativesign arisesbecausethe
substanceis transportedfrom higher to lower concentrationof the substance.
If the rate of transfer of substanceper unit area(or the flux) F, and the concentration
of substanceare expressedin terms of the sameunit of quantity, e.g. gram, it is clear that
the diffusivity D is independentof this unit andhas dimensions:
(1.2) (length)2.(time) or cm2/s
asthe flux F is expressedby g/cm2.s,and the concentrationC is g/cm.
1.1.2 DIFFUSION OF A SUBSTANCE THROUGH A POLYMER
When a dosageform is made of a drug dispersedin a polymer, the polymer playing
the role of a matrix, the whole processof diffusion is as follows. When the dosageform
is in contact with a liquid, e.g. the gastric liquid, this liquid entersthe polymer, dissolves
the drug, and then enablesthe drug to diffuse out of the dosageform through the liquid
located in the dosageform. Both thesetransfersarecontrolled by diffusion.
It is thus of interestto have a knowledge on the diffusion behaviour of various kinds
of polymers.
Generally a polymer is in the glassy or in the rubbery state, depending on the
temperature.Below the temperatureof glassy transition Tg, the polymer is in the glassy
state, and above this temperature it is in the rubbery state. Diffusion of a liquid differs
notably when it goes through a glassy or a rubbery polymer. Segmentsof the polymer
chain are continualy in motion, in the sameway as Brownian motion for gases,creating
voids. As the volume of thesevoids is of the samemagnitudeasthe volume of a molecule
of liquid, these motions enable the molecule of liquid to go through the polymer.
Polymers have a wide spectrum of relaxation times associated with these structural
changes.An increasein temperatureor in concentrationof liquid generally enhancesthe
motion of the polymer segmentsand decreasesthe relaxationtime.
POLYMER
IN THE RUBBERY
STATE
(CASE
I)
A polymer in the rubbery state respondsrapidly to changesin its condition. The

polymer chains adjust very quickly to the presenceof the molecule of liquid. The rate of
diffusion of the liquid is much lessthan that of relaxation of the segmentsof the polymer.
In this case,the diffusion is Fickian (case 1). The amount of diffusing substanceabsorbed
(or desorbed)at time t can be expressedin termsof time by the following relation :
(1.3) M,=k.
fi
Sec. 1.11
Introduction
when Mt is much lower than the correspondingamountafter infinite time a.

where k is a constantdependingon &, the shapeof the polymer and the diffusivity.
POLYMER
IN THE GLASSY
STATE
(CASE
II)
In a polymer in the glassy state,the stressmay be slow to decay after this polymer
has been stretched.Thus, the relaxation process is very slow compared with the rate of
diffusion. In this case, called case II, the liquid diffuses through the polymer with a
constant velocity showing an advancing front which marks the penetration limit of the
liquid. Behind this advancingfront of the liquid, the polymer may turn into swollen gel or
rubber polymer, while aheadof this front, the polymer free of liquid is in the glassy state.
The amountof liquid absorbedat time t is expressedin terms of time t by the following :
M,=k.t
(1.4)
ABSORPTION
OF LIQUID
IN CASE III
When the rates of diffusion of the liquid and of the relaxation of the polymer are of
the same order of magnitude, anomalous or non-Fickian diffusion is observed. This
systemlies betweencaseI and caseII and the amountof liquid absorbedat time t is given
in terms of time by the expression
(1.5)
Mt=k.tn
where n is between h and 1

The caseI system is characterizedby n = 0.5 and the caseII system by n = 1, by
considering eqn (1.5).
1 .I .3 STEADY AND NON-STEADY CONDITIONS
Steady or stationary conditions are reachedwhen the concentrationof diffusing
substancewithin the solid doesnot dependon time. The mathematicalcondition is :
(1.6) g=O
The concentration of substancedepends only on position, and the concentration
distribution of the substancethrough the solid is constant.
Non-steady or transient conditions are obtained when the concentration of
diffusing substancewithin the solid depends on position and time. The mathematical
condition is :
(1.7.) g,o
The diffusion
equations
[Ch. 1
Steady conditions are reached only in a few cases,when the solid through which the
liquid diffuses is considered as a membrane. Two cases are of high interest : the plane
membrane, the spherical membrane. Steady conditions can be obtained when constant
concentrations Ci and Co are maintained on each surface. At the beginning of the process,
the non-steady diffusion of the liquid takes place. After a given time, a steady state is
reached in which the concentration of diffusing substance remains constant at all points
within the membrane.
1 .1.4 INITIAL
CONDITIONS
The initial conditions represent the concentration distribution of diffusing substance

within the solid and in the surrounding, at time t = 0, before the process starts.
For a simple dosage form, the concentration distribution of the drug is generally
uniform. In special dosage forms made of a core and shell, the drug concentration may be
uniform throughout the core and zero in the shell.
1 .1.5 BOUNDARY
CONDITIONS
The boundary conditions express the concentration of diffusing substance on the

surface of the solid.
When the drug does not diffuse out of the dosage form, there is no transfer through
the external surface of the dosage form. This condition is mathematical by expressed by
writing that the rate of transfer through the surface is zero. The following is thus obtained:
(1.8)
-D.g=O
orsimply
g=O
onthesurface
When the substance (liquid, drug) is transferred through the external surface of the
dosage form, the above equation becomes :
(1.9)
%o
2X
on the surface
Two cases are of interest :

(i)
When there is a finite coefficient of matter transfer through the external surface of
the solid, h. The rate at which the diffusing substance is transferred per unit area of
the external surface is thus expressed by :
(1.10)
t>O
- J$js=h(Cs-C.x~
on the surface
where C, and C,,, are the concentration on the surface and in the surrounding,
Sec. 1.11
Introduction
respectively.
is the gradientof concentrationnext to the surface
and D is the diffusivity
Of course,the diffusing substanceenters or leaves the dosageform, depending on
the respectivevalue of C, and C,xt.
(1.11)
cs > Gxt
the substanceleavesthe dosageform
(1.11)
c, < text
the substanceentersthe dosageform
(ii)
When the coefficient of matter transfer through the external surface of the solid is
very high (comparedwith the diiusivity D), the concentrationon the surfaceC, can
be consideredasconstantduring the whole process
(1.12)
t>O
C, = constant
- Instead of using C,, in eqn (l.lO), it is better to use the concentration on the
surface which is at equilibrium with the surrounding atmosphere, C,. Eqn (1 .lO)
expressesthat the rate at which the substance goes through the external surface is
constantly equal to the rate at which the substanceis brought to this surface by internal
diffusion. This rate is also proportional to the difference betweenthe actual concentration
C, and the concentration on the surfacewhich is necessaryto maintain equilibrium with
the surrounding, C, or (C,,,).
REMARK
1.1.6 VOLUME OF THE SURROUNDING ATMOSPHERE AND

PARTITION FACTOR
It is of interest to consider the volume of the surrounding atmosphereas compared
with the volume of the dosageform.
Moreover, there is often a partition factor, K, meaning that at equilibrium the
concentration of diffusing substanceis K times in the solid than in the surrounding
atmosphere.
When the volume of the surrounding atmosphereis much higher than that of the
dosageform, this volume can be consideredas infinite. This meansthat the concentration
of diffusing substancein the surrounding atmosphere does not vary. For instance, it
The diffusion
equations
[Ch. 1
remains zero (or negligible) for the drug leaving the dosageform, and constant for the
liquid entering the dosageform.
In some cases,the volume of the surrounding atmosphereis not much higher than
that of the dosage form. This means that the concentration of drug in the surrounding
atmosphereincreases.
The ratio of the volumes of the surrounding atmosphereand the dosage form is
expressedby :
(1.13)
a = Kvsu* dos
form
where K is the partition factor for the drug.

1.2 EQUATIONS
OF DIFFUSION
FOR VARIOUS
SHAPES
Equations of diffusion are considered for various shapesof the dosageforms, and
some emphasis is placed in the case of the thin sheet by developing calculation. The
following shapes are examined successively : thin sheet, rectangular parallelepiped,
cylinder of infinite and finite length, and sphere.
1.2.1
EQUATION
OF DIFFUSION
FOR A THIN
SHEET
The thin sheet of thickness dx perpendicular to the direction of diffusion is

considered(Fig. 1.1) with the areaA through which the substancediffuses along the xaxis. The matter balance is evaluatedduring the time dt within this small volume Adx.
Fx+dx
X
Fig. 1.1. Diffusion of a substancethrough a thin sheetof thicknessdx
Sec. 1.21
Equations
of diffusion
for various shapes
The rate at which the diffusing substanceenters the sheet of area A is F,.A, and the
changein the amount of substanceduring the time dt is A (F, - Fx+& dt. This changein
the amount of substanceis responsiblefor a changein the concentration within the sheet
which can be written asfollows :
A. (Fx- Fx+,& . dt = A.dx.dC
(1.14)
The difference F, - Fx+& is given by :

(1.15) Fx- Fx+dx= andthe matterbalancebecomes:
(1.16)
aF, ac
-ax=Ji
where the concentrationC is a function of spaceand time.

As the flux F, of diffusing substanceis expressedin terms of the diffusivity D and
of the gradient of concentrationeqn (l.l), the above equation can be rewritten in the final
form :
(1.17) G=&(D.
$j
When the difusivity is constant,this equationsimplifies

(1.18) $=D.?
a%
with constantD
1.2.2 EQUATIONS OF DIFFUSION

PARALLELEPIPED
ISOTROPIC
FOR A RECTANGULAR
MATERIALS
A rectangularparallelepipedwhose sidesare parallel to the axesof coordinatesand

of length dx, dy, dz is drawn (Fig. 1.2). The matter balance within this small
parallelepiped can be evaluated in he same way as for the sheet by considering the
diffusion along the three axes.Eqn (1.16) is thus rewritten as follows :
The diffusion
equations
[Ch. 1
dx
Fig. 1.2. Diffusion of a substancethrough a small rectangularparallelepipedof

sides dx, dy, dz
(1.19)
ac aF, aF,
-Ji=~+~+az
a&
where Fx representsthe flux of diffusing substancealong the x-axis.

This equationbecomes,by replacing eachflux F by its value given in eqn (1.l) :
l1.20)
$=&ID.
$$+$[D.$)+=$.
$1
where the diffusivity D may be concentration-dependent.

In the caseof a constantdiffusivity, the equationreducesto :
ANISOTROPIC
MATERIALS
Anisotropic materials have different diffusion properties in different directions.

Someexamplesare given with crystals, wood, and polymer films in which the molecules
have been oriented. In this case,the direction of flow of diffusing substanceat any point
is not normal to the surface of constantconcentrationthrough this point. Eqn (1.1) must
be replacedby the assumption:
Sec. 1.21
Equations
(1.22) -F,=D,,.
E+D,*.
of diffusion
for various shapes
$+DIS.$
showing that F x dependsnot only on the gradient
but also on $ andg.
They are three principal axesof diffusion, and three principal diffusivities Dx, Dy,
Dz, in the casewhere the principal axesof diffusion are the sameas the x, y, z axes.
The equationof diffusion can thus be written :
or more simply when eachprincipal diffusivity is not concentration-dependent.

(1.24)
1.2.3
CYLINDER
CYLINDER
OF INFINITE AND FINITE LENGTH
OF INFINITE
LENGTH
The caseof a cylinder of infinite length is easyto study, as the transfer of substance
is radial only.
When the diffusivity is concentration-dependent,the equation of diffusion is as
follows :
(1.25)
C=f.#.D.$)
-g-
and when the diffusivity is constantit becomes:

(1.26)
CYLINDER
OF FINITE
LENGTH
The diffusion is radial andlongitudinal along the z-axis.

When the diffusivity is concentration-dependent,the equation of diffusion results
from the superpositionof the radial and longitudinal diffusion.
The diffusion
10
equations
ICh. 1
(1.27) ?$;.;(r.D.~j+~(D.$)
When the diffusivity is constant,this equationreducesto :
+- a2c
az2
(1.28)
Of course,in thesetwo cases,the material is isotropic.

1.2.4
RADIAL
DIFFUSION
IN
A SPHERE
When the diffusion is radial and the material is isotropic, the equationof diffusion is
given by :
with a concentration-dependent
diffusivity :
(1.29)
G=$.g[D.r.$]
with a constantdiffusivity :
(1.30) $=D.
a2c
ac
ar2+- r?F
1.3 METHODS OF SOLUTION WHEN THE DIFFUSIVITY

CONSTANT
1.3.1
KINDS
OF
IS
SOLUTION
When the diffusivity is constant,general solutions of the equation of diffusion can

be obtained for various initial and boundary conditions, when these latter conditions are
not too complex.
Two standardforms can generally be obtained.The one, when the concentrationof
diffusing substanceand the amount of substancetransferred are expressedin terms of
error functions and of related integrals. Becausethey converge rather quickly for small
times, they are of great interest for evaluation during the early stagesof diffusion, when
the amount of substance transferred is low. The other, when the concentration of
susbstanceand kinetics of substancetransferred are given in the form of trigonometrical
Sec. 1.31
Methods
of solution
when the diffusivity
is constant
11
series. These series can be used for large values of time, when the amount of matter
transferred as a fraction of the correspondingamount after infinite time is high, because
thesetypes of seriesconvergevery quickly in theseconditions.
Three methodsof solution of the diffusion equationcan be used.
(i) The method of separation of variables, which is widely used. It gives solutions
expressedin terms of trigonometrical series.
(ii)The Laplace transform which is an operator method through which the partial
differential equations are transferred in ordinary equations. The two kinds of
solutions can be obtained.
(iii)The method of superposition and reflection. Solutions are obtained in terms of error
functions.
In the caseof a cylinder with radial transfer,seriesof Besselfunctions are obtained
insteadof trigonometrical series.
1.3.2 METHOD OF SEPARATION OF VARIABLES
By making the assumptionthat the variable x and t are separable,an attempt can be
madeto find a solution for the partial differential equationsof diffusion.
For instance, in the case of a one-dimensional difusion through a sheet, upon
putting :
(1.31)
C&t= c,. Ct
where C, and Ct are functions of x and t, respectively, the general equation of diffusion
(1.18) becomes:
(1.32) 2.
C,=D.
Ct. s
dx2
This equationcan be rewritten after separationof variables :

1 dc,
(1.33) q. z=
D d2C,
-c, * dx2
where the left-hand side dependson time only, and the right-hand side dependson x only.
The two ordinary equationsare thus obtained
The diffusion
12
(1.34) &. ?=-A.
(1.35) $
equations
[Ch. 1
. KS + X2. D=O
dx2
the solutions of which are :

(1.36)
Ct = Constantx exp ( - h2 Dt)
and
(1.37)
C, = A. sin hx + B . cos hx
The solution (1.31) of eqn (1.18) is thus

(1.38)
C&= C,. Ct = Constant (A . sin hx + B . cos hx) . exp ( - A2 Dt)
where A and B are constants.

Eqn (1.38) being a linear equation, a general solution is obtained by summing these
solutions as follows :
(1.39)
C,,,= %[A.
sh-~h,x+B,. cosh,,x). exp(-(Dt)
n=O
where A,,, B, and h, are constants.

The constantsA,.,,B, and ?+,are determinedby the boundary and initial conditions
for eachproblem.
The simple problem of a sheet of thickness L with 0 < x < L is considered (Fig.
1.3). The diffusing substanceis initially uniform throughout the sheet,and the coefficient
of matter transfer on both surfaces x = 0 and x = L is so high that the concentrationon
thesesurfacesis zero as soon as the processstarts.
The initial andboundary conditions are as follows :
(1.40)
t=O
O<x<L
C = Gin
initial conditions
Methods
Sec. 1.31
of solution
is constant
13
in
Fig. 1.3. Diffusion of a substanceout of a sheet 0 c x c L with a uniform initial

concentrationand zero concentrationon the surfaces
t>o
(1.41)
x=0
c=o
boundaryconditions
andx=L
The boundary condition C = 0 for x = 0 necessitates
(1.42)
B, = 0
since cos 0 = 1,
while the other boundary condition C = 0 for x = L is satisfied by :
sin LL
(1.43)
= 0 = sin nx
andby:
(1.44) h,= r-r:
The initial condition can thus be written :
(1.45)
Cx,o = gA,.
sinn?
for 0 c x L
By multiplying both sidesof this equationby sin n y

becomes:
and integratingfrom 0 to L it
The diffusion
14
equations
(1.46)
C,.,
[Ch. 1

L
J
J
,,sin~.
o sinn~dx=Ar.
+A,
. prcx
L
sm-.
sin.Fdx+
. nrcx
slnLdx+
. . . . A,,
J
0
sin2~dx+
....
As
sinp. sinq=;[cos(p-q)-cos{p+q)l
sin2p=;
1l-cos2
pI
it becomesobvious that :
L
J
J
. pxx
L
0
sm-.
. n7tx
SlnLdx
= 0
forp#n
2n7cx
hlpX=;
forp = n
Moreover,
L
J
0
sin~clx
= -&[l-cosnrr]
The even terms of n make A,, vanish, and only the odd terms of n are considered: n
= 1, 3, 5
The constant A can thus be expressedin this way with odd valuesof n, or 2n + 1 :
4 cx.0
147
A=(2n+
4 tin
1)rr = (2n+ 1)7t
The final solution for the concentration of substancewithin the sheet 0 c x < L is
thus given by the trigonometrical series:
(1.48)
4Ci, O 1
Cx,t= x
c -.
+p+
l
. (2n+l)rrx
sm
L
Methods
Sec. 1.31
of solution
1.3.3 METHOD OF THE LAPLACE
is constant
15
TRANSFORM
The Laplace transform is a mathematical treatment through which a partial

differential equation expressed in terms of the variable x and t is transformed in an
ordinary equation expressedin terms of the variable x only. Application of the Laplace
transform to the diffusion equation removes the time variable, leading to an ordinary
equationwith spacevariables.
The Laplace transform f(p) of a function of time f(t) for positive values of time is
given by :
ca
T(p) =
(1.49)
o f(t) . exp (- pt) . dt
where p is a number which is chosenhigh enoughto make the integral converge.

By integrating (1.49), it is thus possible to obtain the Laplace transform of a
function. A few simple examplesare given :
f(t) = 1
f(p) = I o exp b-P4
f(P) = ;
ea
f(t)=exP(-at)
f(p)=
f(p)=---&
uexp-(p+a)t.dt
Tables of Laplacestransfonnsare availablein somebooks of mathematics

Consider a plane sheetof thickness2 L, with - L < x < L, initially free from liquid,
with a constantconcentrationon eachsurface.
The initial andboundaryconditions are :
(1.50)
t= 0
-L<x<L
c=o
(1.51)
t>O
x=fL
c = C-J
surfaces
In this case, the midplane x = 0 is a plane of symmetry, and the gradient of

concentrationis constantlyzero at this place.
(1.52)
t 2 0
$0
x= 0
midplane
Upon multiplying both sidesof the equationof diffusion
16
The diffusion
equations
[Ch. 1
by exp ( - pt) and integrating with respectto t from 0 to 00,there is :

-aC
oxexp(-pt)dt
= D.
O 2
ac exp ( - pt) dt
O ax2
Integrating by parts the left-hand side
00
-ac
ox.exp(-pt)dt=[C.exp(-pt)]i+p
oC.exp(-pt)dt=p.C
as the term in bracket vanishes at t = 0 becauseof the initial conditions and at t = 00

becauseof the exponential term.
The right-hand side is written as follows :
by interchangingthe ordersof integration anddifferentation.

The equationfor the Laplacetransformc is :
(1.53)
e=
dx2
(1.54)
q2+
q2. C
It is more convenient to use the condition at the midplane with eqn.(l.S2) and to
consider only half the thicknessof the sheet,0 I x I L, instead of the condition x = L, C
= C-J for t > 0.
The boundary condition C = Cm for x = L and t > 0 gives the following in terms of
Laplacetransform :
ca
oC, .exp(-pt)&=$
x= L
Sec. 1.31
Methods
of solution
is constant
17
The equationat the midplanebecomes:

x= 0
(1.56) $=O
The solution of eqn.(1.53) satisfying the conditions (1.55) and (1.56) is given by
eqn. (1.1)
This hyperbolic function can be expressedin terms of negativeexponentialsand can

thus be expandedis a seriesby using the binomial value.
explqxl+exp(-qxl
Gp. CO
exp kd . [ 1 + exp (- 2 9
L)]
C=:[exp(-q(L-x)+exp(-q(L+x)].
t(-
l)n. exp (- 2 n q L)
n=O
(1.58) c=$o.
g( - l)? exp(-q(2n+ljL-x)+:.
g(-l)(-q(2n+l)L+x)
n=O
n=O
As the Laplacetransformof
erfc
x
i i-75
. exp( -9x)
P
I lS
the concentrationis thus given by :

(1.59)
1.3.4
~=~(-l)n.erfc(2n~~-Xj+~(-l)n.erfc/~2n~~+x)
n=O
n=O
METHOD
PLANE
OF REFLECTION
SOURCE
The function with the constantA
AND
SUPERPOSITION
The diffusion
18
equations
[Ch. 1
(1.60) C = $exp
is a solution of the equationof diffusion in one dimensionwith the constantdiffusivity
This function is symmetrical with respectto the plane x = 0, vanisheswhen it tends

to infinity.
The amount of substancediffusing along the x-axis through a cross-sectionof unit
areais
+-
(1.61)
M = -co
Upon putting
x2
v =4
2
dx=2fidx
this integral becomes
(1.62) M = 2AI%
exp (-v 1 dv = 2Am
On substitutingfor the constantfrom the aboverelation, the function C expressedby

eqn (1.60) becomes:
(1.63) C = &exp
This equation expressesthe diffusion of the amount of substanceM located in the
plane
x=Oofunitareaattimet=O.
Methods
Sec. 1.31
REFLECTION
of solution
is constant
19
AT A BOUNDARY
The concept of reflection of the diffusion substancecan be used at an impermeable

boundary.
For instance,the diffusion of the amount M of substancefrom the plane surfacex =
0 through the cross-section of unit area is considered along positive x, with an
impermeableboundaryto prevent the substancediffusing along negativex.
The solution for the diffusion of the substancegiven by eqn (1.63) can be used.By
using the principle of reflection at the boundary and superpositionof the substancealong
the positive x, it is clear that the concentrationalong the positive x only is given by :
(1.64) C = -$&
exp
As shown above,the condition at the impermeableboundaryis given by :
?$ 0
EXTENDED
x=0
INITIAL
along negativex
DISTRIBUTION
OF THE SUBSTANCE
Generally, the substanceis not located in a plane, but it occupiesa finite region.
In the casewhere the substanceinitiaIly occupiesthe semi-infinite medium x = 0 the
initial conditionsare :
(1.65)
t = 0
xCO
C=Ci
and
x>O
C=O
By considering that the extended distribution of substanceis decomposed in an

infinite number of plane sources,the solution to this problem of transfer through the plane
x = 0 is given by superposingthe correspondinginfinite number of solutions eqn.cl.63).
The amountof diffusing substanceinitially locatedwithin the elementof thicknessd&
is Ci, de diffuses along the x-axis. The concentration at a given point P, distance e
from this element,resulting from this diffusion is given at time t by :
(1.66) ?!!!2m
exp
The solution for the concentration at point P resulting from the initial distribution
with the semi-infinite medium x < 0 full of substance eqn(1.65) is thus given by
integrating the eqn.tl.66) over the limits x and -.
20
The diffusion
(1.67)
C,,,
equations
= &.
J~~FIp(
[Ch. 1
-&)d&
Upon putting v2 = -&
ci*
C x,t = lr;;
J
m
x+2fiexP (-v)dv
asthe error function is given by :

erfcy=-
2
v5
y
exp (- v2) dv
0
the concentration at point P at position x and time t resulting from the initial uniform
distribution of substancein the semi-infinite medium x < 0, is expressedby :
(1.68) C&=$erfc(&)
The error function hasthe following properties:
erf t-y) = - erf (y)
erf(O)=O
erf(=)= 1
and the error function complementis given by :

erfc (y) = 1
1 -a-f(y)
REFERENCES
1 J. Crank, The Mathematics of Diffusion, Clarendon Press,Oxford, 1975, p.22.
Mathematical treatment
a plane sheet
2.1
of diffusion
in
INTRODUCTION
A plane sheetis a medium bounded by two parallel planes of infinite dimensions.

In this case,a one-dimensionaldiffusion of substanceis thus considered.
In fact, the plane sheetis so thin with respectto the other two dimensionsthat all the
substance diffuses along the axis perpendicular to the plane faces and the amount
transferredthrough the edgesis negligible.
Various casesare of interest, depending on the value of the coefficient of matter
transferon the substanceand on the volume of solution in which the plane sheetis placed.
Generally there is a non-steady diffusion. The steady state diffusion is only
observed in the case of a membrane, e.g., when the concentrations on each surface are
constantand different.
The diffusion must be constant for a mathematical treatment. In the case of a
concentration-dependent diffusivity, a numerical method is necessary to resolve the
problem.
The equationof one-dimensionaldiffusion with constantdiffusivity is :
(2.1) $
= IS
ax2
As shown already in Chapter 1, thene are two types of solutions, one expressedin
22
Mathematical
treatment
of diffusion
in a plane sheet
[Ch. 2
terms of trigonometrical series,and the other in terms of error functions. Thesetwo types
of solution are given for the simple problems of diffusion. Three ways of calculation exist
(the method of separation of variables, the Laplace transform, the method using
superpositionand reflexion in finite system),but the method of separationof variables is
widely used, and full calculations are given for the simple problems of diffusion. In other
more complex cases,the solutions are only given without calculation, but some emphasis
is placed upon the conditions for which they can be used, e.g., initial and boundary
conditions.
2.2
NON-STEADY
STATE WITH A HIGH COEFFICIENT
MATTER TRANSFER ON THE SURFACE AND AN
INFINITE VOLUME OF THE SURROUNDING
OF
This is the casewhere the volume of the surrounding atmosphereis so high with
respectto the volume of the sheetthat it can be consideredas infinite. A value of the ratio
of the volumes of the surrounding atmosphere and the sheet for which this case is
obtainedis determinedin Section 2.4.
The concentration on the surfacesof the sheetcan thus be considered as constant.
They are equal to the concentration of the surrounding atmosphere,when the partition
factor is 1 and to KC,,, when there is a partition factor K.
Three Subsectionsare of interest :
- when the initial distribution is uniform in the sheet
- when the initial distribution is f(x) in the sheet
- when the concentrationson eachsurfaceof the sheetare different. This is the caseof the
plane membrane.
2.2.1 UNIFORM
INITIAL
DISTRIBUTION
CASE OF A SHEET OF THICKNESS
IN THE SHEET
2L, WITH - L < x < L
The sheetin the region - L < x < L is initially at the uniform concentrationCk, and
the surfacesareat a constantconcentrationC,.
The initial and boundaryconditionsare :
(2.2)
t=O
(2.3)
t>O
-L<x<L
x=fL
C = Ci~
c=c,
sheet
surfaces
Sec. 2.21
Non-steady
state with a high
coefficient
of matter transfer
23
As the midplane x = 0 of the sheetis a plane of symmetry,there is also the condition

at any times :
(2.4) t 2 0
= 0
x= 0
midplane
The general solution for the equation of diffusion (2.1) is shown in Chapter 1, by
using the method of separationof variables:
(2.5) c,,,-
c,
= %[A*.
n=O
sin h,,x + B,. cos h,,x) . exp[ -krDt)
The condition (2.4) written asfollows :
n=O
necessitatesthat
A,,=0
(2.6)
The condition (2.3), for example x = L and CXWt

- C, = 0, for t > 0, is fulfilled
when
(2.7) cos h,L=O=cos(2n+l)t
and h,=12n2+L1z
The initial condition becomes
(2.8) Ci,- Cm = c B,.cosw

n=O
for
-L<x<L
The coeffkient B, is obtainedas follows, by multiplying both sidesof the

(2n+ ~)IFX
aboveequationby cos
2L
dx and integratingbetween - L and L.
L
(2.9) (cin-cJ
-L ~0s.~~:~~~
dx = . ..B.
-Lcos22n;~xxdx
+ . . .. .
24
Mathematical
treatment
of diffusion
in a plane sheet
Kh.
+B
-LCOS
(2n+ 1JKx
(2p+ llxx
2L
.cos
2L
Ascosn. cosp = $ [ cos (n + p) + cos (n - p)]

cos2n = ;[l + cos 2n]
it is obvious that :
L
J
J
-L
coJ2
n+ 1)xx
2L
. cos2p;;)~xdx=o
-L
Cos2(2n+ l)lcx
dx=L
2L
Moreover
L
J
-L
cos12n+ 1)x:x dx =
. (2n+ 1)n
4L
2L
(2n+ 1)7csm
2
which can be also written in the form

4L
(-1)
(2n+ 1)X
becausefor even valuesof n
sin%
2
oddvaluesofn
B, = (2n+41,x(-
sin-3n = - 1= sin--(2n+l) z
2
2
The coeffkient B, is thus equal to :

(2.10)
1 = &I?!
1)"
dx + . . . . .
Sec. 2.21
Non-steady
state with a high
coefficient
of matter transfer
25
The profile of concentrationof the diffusing substancewithin the sheetof thickness

2L, (- L < x c L), is thus expressedin terms of space x and tune t by the following
series:
C
- (- lIn
-CC4
X+t
= ;. C~.COS
2n+ 1
tin- cca
n=O
2.11
(2n+ 1)7cx .exp -(2n+ 1t21r2Dt

2L
1
4L2
The amount of substancewhich is transferredinto or out of the sheetat time t, M,, is

calculatingby integrating the flux of substancethrough the surfacewith respectto time.
for
(2.13)
ac
I2-F=
I
2(cin-cJ
x=fL
m _ 1 n .&2n+ 1)Xx. exp _ (2n+ 1)2n2Dt

C(
1.
2L
4 L2
n=O
and
The amountof matter transferredMt becomes:
or
Mathematical
26
treatment
of diffusion
in a plane sheet
[Ch. 2
(2.16) M, =
co
As the series c
=- iT2
*So (2 n + Ii2
and the amount of substancewhich has enteredof left the sheetafter infinite time, a,
given by :
is
(2.17) M,= 2 (Ci,- C,( L

the total amount of diffusing substancewhich has entered(or left) the sheetof thickness
2L (- L c x < L), at time t, M,, is expressedas a fraction of the corresponding quantity
after infinite time M.,,,by the following relation :
exp -(2n+ 1)2n2Dt
4L2
i
(2.18) x=
00
Of course,when the diffusing substanceentersthe sheet:

(2.19)
Gin < Co3
and when the diffusing substancesleavesthe sheet

(2.19)
Gin > Ccu
The corresponding solutions are also obtained in terms of error functions by using
either the Laplacetransformor the method of superpositionand reflection.
The profiles of concentration are expressedas a function of spacex and time t, for
the samesheetof thickness2 L.
(2.20)
:*t
- tin
co- tin
= z(n=O
l)n. erfc
(2n+ l)L-x
21cDi
Sec. 2.21
Non-steady
state with a high coefficient
- 1) n. erfc
+
C(
27
of matter transfer
(2n+ l)L+x
n=O
2l-D-t
where C, representsthe constantconcentrationon the two surfacesof the sheetfor t > 0.

This is also the concentrationobtained in the sheetafter infinite time, when the sheetand
the surroundingatmosphereare at equilibrium.
The kinetics of the diffusing substancewhich has entered or left the sheetis given
by :
(2.21)
$+2.
= 2fE.
ca
I
z(n=l
l)n. ierfcs
- Diffusion for long times with a sheetof thickness 2 L (- L < x c L)

The equations(2.11) and (2.18) are of better use for long times of diffusion, as
REMARKI
the seriesdoesnot convergefast for short values of time, and low values of the ratio $.
For very long times of diffusion or ratherfor high values of the ratio 2,
OD
the first
term of the seriesin eqn (2.18) is preponderent.The simple equationis thus obtained,for
Mt > 0.5 - 0.7 dependingon the accuracywhich is wanted.
M
0
(2.22)
Mt = +exp
ca
x
_ nZDt
1 4 L2 !
Expressedin the logarithmic form, a straightline is obtainedby plotting log
as a
function oft. The diffusivity D can this easily be calculatedfrom the slope of this straigth
line.
- Diffusion for small times with a sheetof thickness 2 L (- L < x < L)
The solutions expressedin terms of the error functions are of better use for small
tunes of diffusion, asthe serieconvergevery fast for short tunes and small values of the
M
ratio 2
M,
REMARK 2
Mathematical
28
treatment
of diffusion
in a plane sheet
[Ch. 2
For very small times of diffusion, the series in eqn (2.21) tends to 0, and the
Mt
very well known equation is obtained for M c 0.3 - 0.5, depending on the desired
ca
accuracy.
(2.23)
= t E
with thickness = 2L
This equation is of high interest for calculating the diffusivity D obtained by

plotting the ratio $
00
as a function of the square root of time, at the beginning of the process
when the coefficient of matter transfer is very high and the volume of the surrounding
atmosphere is much larger than that of the sheet.
3 - Half-time of sorption (desorption) with a sheet of thickness 2 L
(- L < x < L).
REMARK
The half-time of the sorption (or desorption) process is sometimes used for
w
calculating the diffusivity. This is the time necessary for M = 1.
m 2
By considering only the first term of the trigonometrical series obtained for long
time, e.g., eqn (2.22) the following relation is obtained :
(2.24)
D= 0.1958
for
ce
= k
By considering the simple relation-ship obtained for small times, e.g. eqn (2,23),
the diffusivity is expressed in terms of $
(2.25)
by :
D= 0.196
where L is half the thickness of the sheet.

REMARK 4
- Constant diffusivity
When the diffusivity
obtained by using these three methods (long times, small
tunes, half-time of transfer), is the same, the diffusivity can be considered as constant.
Otherwise, the diffusivity is concentration-dependent.
sec. 2.21
Non-steady
of matter transfer
29
- Partition factor of the diffusion substance

If there is a partition factor for the diffusing substancebetweenthe solid and
REMARK 5
the surrounding atmosphere,it is expressedby K = $,

ext
where C, is the constant concentration on the surface. This concentration is also the
concentrationrequired to maintain equilibrium with the surounding in which the constant
concentrationof diffusing substanceis C,,.
- Infinite value of the coefficient of matter transfer on the surface
The caseof a constantconcentrationon the surfaceis obtainedwhen the coefficient
of matter transfer on the surfaces of the sheet is infinite, as shown in the following
relation :
REMARK 6
F=h(C,-C,)
where C, is the concentrationon the surface,which is always equal to C,,
h is the coefficient of matter transfer on the surface(cm/s),
andF is the flux of matter through the surface.
In practical use, this caseof a constantconcentrationon the surfaceis encountered
when the coefficient of matter transfer on the surfaceis sufficiently high with regardto the
diffusivity of the substance within the sheet of thickness 2 L, or rather when the
dimensionless
L$ is high enough.
For instance.when
(2.27) +$ > 20-50
dependencyon the accuracywanted,h can be consideredasinfinite.
- Caseof a sheetof thickness L, with 0 < x < L
The solution of the equation of diffusion in this simple case is also obtained by
using the method of separationof variables.The initial and boundaryconditions are :
REMARK 7
(2.28)
t=O
O<x<L
C = Ci~
sheet
Mathematical
30
treatment
t>o
(2.29)
of diffusion
x= 0
[Ch. 2
in a plane sheet
surfaces
c=c,
andx=L
The profiles of concentration CXt at any time is given by the relation :
:*yz-
(2.30)
= 4. &-&-.
in
00
sb(2n+L1)Kx.
exp -(2n+
n=O
12A2Dt
L2
The amount of substance transferred through the surfaces of the sheet is calculated
by integrating the flux of substancethrough the surfaces with respect to time :
(2.31)
x=0
for
M, = 2D
or x=L
The total amount of diffusing substancewhich has entered (or left) the plane sheet of
thickness L, (0 < x < L), at time t, M,, is expressed in terms of time, as a fraction of the
corresponding quantity after infinite time M, :
(2.32)
Co
l-+.2
.exp
n=~Pn+
1j2
-i2n+1)2rc2D1
j
L2
a - Volume and nature of the surrounding atmosphere

The case of a constant concentration of diffusing substance on the surface of the
REMARK
sheet necessitatesthat the concentration of this substance in the surrounding atmosphere is

always constant and uniform. This fact can only be obtained when the surrounding
atmosphere is strongly stirred and of infinite volume.
In practical use, this case can be encountered when the volume of the surrounding is
high enough with respect to the volume of the sheet. This is obtained when the
dimensionless number 01which is the ratio of the volumes of the surrounding atmosphere
and of the sheet is high enough.
The dimensionless number a can also be expressed in terms of the partition factor K,
Non-steady
Sec. 2.21
of matter transfer
31
when this partition factor is not 1.

Generally, the equations for the concentration distribution and for the kinetics of
transfer of the diffising substancecan be usedwhen
a> 20-50
(2.34)
dependingon the accuracywhich is desired.

2.2.2 INITIAL DISTRIBUTION f(x) IN THE SHEET OF THICKNESS L
(0 < x < L)
The sheetwith the initial concentrationdistribution of the diffusing substancef(x) is
immersed in a surrounding atmosphereof infinite volume and an infinite coefficient of
matter transfer on the surfaces.
Two casesof interest can be considered,when the plane sheetseparatestwo media
of different concentrations, and when the sheet is immersed in the same surrounding
atmosphere.
TWO
DIFFERENT
MEDIA
In the first case,the initial and boundaryconditions are :

(2.35)
t=O
OCXCL
(2.36)
t>o
x=0
x=L
($0 = f(x)
sheet
surface0
surfaceL
CO*CL.-
The solution of the equation of diffusion can be obtained by the method of

separation of variables. The concentration distribution C,,t within the sheet is thus
expressedby the trigonometrical series:
(2.371
Cc.,Co.,
=(CL,.,Co,,);
+;. c
n=l
+2.
zsinFexp(-FDt).
L
CL,, . cos n X - Co., . nrtx

sm exp
L
n
Jlfcxl,.
sti%dxl
Mathematical
32
treatment
0.1
of diffusion
0.2
in a plane sheet
0.3
[Ch. 2
0.1
0.5
Fig.2.1. Kinetics of matter absorbed(or desorbed)by a sheet,by plotting M&v&, vs.

m.
-with
the ticlmess 2 L and an infinite rate of transferon the surfaces
2L
This caseof the sheetwith the two surfacesmaintained at a constant and different
concentration is of interest, becauseit representsthe diffusion through a membrane of
thickness L separatingtwo media, when the initial distribution of substancewithin the
sheetis not uniform.
THE TWO MEDIA
ARE IDENTICAL
Of course, the initial condition remains the same,while the boundary condition is
more simple
(2.38)
t>o
x=0
andx=L
co,, = qua = cc0
surfaces
Sec. 2.21
Non-steady
of matter transfer
33
CxJ-Cin
t Coo-Cin
1.0_
I.5
.1
LU
a3
02 0.2
wg$g
PA
rnOL
Ok------0
0.2
04
00.1
I
06
0.005
I
0.8
x/l
I
1.0
Fig.2.2. Concentrationdistribution at various times in the sheetof thickness2 L,

with uniform initial concentrationCh and constantconcentrationC, on the
surfaces,various valuesof the dimensionlessnumber E.
L2
The concentrationdistribution of the substancewithin the sheetis thus given by :
2 2
(2.39) C,,t-C,
= %.
~cosnnn-l
. sin?.
n=l
2 - .
+ -. c smy.
L
n=l
exp[ -FDt].
exp -yDt
i
Jlf(xj,
sinqdx
Mathematical
34
treatment
of diffusion
in a plane sheet
[Ch. 2
Ci
\
\
+,a
Fig.2.3. Membrane of thicknessL, with uniform initial concentrationsCh, and

constantconcentrationson each surface : Ci for x = 0, and Co for x = L.
2.2.3 NON-STEADY STATE WITH A MEMBRANE WITH A UNIFORM

INITIAL DISTRIBUTION AND SURFACE CONCENTRATIONS
DIFFERENT
This is the interesting caseof a membranewith a uniform initial distribution of the
substanceC, within the sheetand whose each surfaceis kept at a constantconcentration.
One surfaceis kept at the concentrationCi (inlet) andthe other at Co (outled) with Ci > Co
(Fig 2.3). When the surrounding atmosphereis strongly stirred, the coefficient of matter
transfer at the surfaceis infinite, and the concentrationof the substanceon each surfaceis
K times that of the surrounding, K being the partition factor. When the surrounding
atmosphereis not well stirred, the coefficient of matter transfer on the surfaceh is finite,
andthe rate of transfer is given by :
(2.26)
h 02, - C,,,)
where C, is the actual concentrationon the surface

and C, is the constantconcentrationin the surroundingfar from the membrane.
Non-steady
Sec. 2.21
of matter transfer
35
The caseof a constantconcentrationon the surfaceis only considered,becausethis

casegenerally occurswhen the membraneis in contactwith two liquids.
The initial and boundaryconditionsfor a sheetof thicknessL are given by :
IL
(2.40)
t=O
OSX
(2.41)
t>O
O<x<L
X = OCi
x = LCo
Ci
c,,,
Over a given period of time, the substanceis transferred in non-steady state, and
after this time a steady-statetransfer is setup.
The concentrationdistribution of substancewithin the membranein non-steady state
is obtainedby the method of separationof variables :
(2.42)
C,,, =
n=l
4cin w 1
+x c msm
. (2m+l)Irx
L
m=O
exp _ Pm+1i2x2 D t
L2
!
i
When t +=, the exponentialsvanish, and the linear concentrationdistribution

is attained,the substancediffusing in steadystate.
The amountof substancewhich entersthe membraneduring time t, Mt, is given as a
fraction of the correspondingquantity after infinite time, M,.
(2.43)
= 1 - B.
x
This aboveequationis obtainedby integratingwith respectto time, per unit area:
for x = 0
36
Mathematical
treatment
of diffusion
in a plane sheet
[Ch. 2
where F is the gradientof concentrationat x = 0.

X
The value of M, is obviously given by :

(2.45)
M, =
Ci+ CO)-Ci, L
1
The amountof substancewhich hasleft the membraneafter time t per unit area
M, is obtainedby integratingwith respectto time the following equation:
dt
for
x = L
This amount is thus obtainedas follows :
Fig. 2.4. Non-steadyand steadyconditions for the membrane[Gg=(~))
Sec. 2.31
Non-steady
state with a finite coefficient
of matter transfer
37
1 - Simple caseof a membraneinitially free of substance,with Co = 0.

In this simple case,C, = Co = 0, and eqn (2.46) reducesto :
REMARK
'
Dc.t
(2.47) M,=+
LC.
- y--c
2
7c2
(- l)n
exp
n2
When t +w, the exponentialvanishes,andM, is proportional to time :

DC,
(2.48) M,=L
As shown by plotting the dimensionlessnumber &
as a function of the dimensionless

1.
D.t
number T the non-steadystateis observedat the beginning of the process, and the curve
L
has an asymptoticvalue which is associatedwith the steadycondition eqn (2.4).
2.3 NON-STEADY
OF MATTER
STATE WITH A FINITE COEFFICIENT

TRANSFER ON THE SURFACE
This case is of high interest as very often the coefficient of matter transfer on the
surfaceis not infinite, evenwhen the sheetis immersedin a liquid.
This caseis also obtained when the diffusing substanceevaporatesfrom the surface
(or condensates)with a finite rate of evaporation.
The problem is resolved with a sheet ot thickness 2 L initially at a uniform
concentrationC,, immersedin a surroundingatmosphere.
The initial andboundaryconditionsare :
Mathematical
38
-D.
t > 0
of diffusion
in a plane sheet
-LCX<L
t =o
(2.49)
(2.50)
treatment
tin
$ =h(C,-CW)
I I
[Ch. 2
sheet
surfaces
The condition on the surface expressesthat the rate at which the substanceleaves
the sheetis constantly equal to the rate at which the substanceis brought to the surfaceby
internal diffusion. It is also seen that this rate, per unit area, is proportional to the
difference of the actual concentrationon the surfaceC, and the concentrationrequired C,
to maintain equilibrium with the surroundingatmosphere.This concentrationis written C,
as it is also the value obtainedfor C, at infinite time.
The concentrationdistribution within the sheetis given by : (1,2,3)
(2.51)
Pnx
x.t- tin=1_
cca- ci,
-dDt
L2
where the p,s arethe positive roots of :

p.tanP=M
(2.52)
Table 2.1
L.h
Roots of p.tan p = M with M = D
pl
!32
P3
P4
PS
P6
P7
P8
P9
PlO
0.00
0.01
0.10
0.20
0.50
1.00
2.00
5.00
10.00
100.00
900.00
OD
0.0000
0.0998
0.3111
0.4328
0.6533
0.8603
1.0769
1.3138
1.4289
1.5552
1.5691
1.5708
3.1416
3.1448
3.1731
3.2039
3.2923
3.4256
3.6436
4.0336
4.3058
4.6658
4.7072
4.7124
6.2832
6.2848
6.2991
6.3148
6.3616
6.4373
6.5783
6.9096
7.2281
7.7764
7.8453
7.8540
9.4248
9.4258
9.4354
9.4459
9.4775
9.5293
9.62%
9.8928
10.2003
10.8871
10.9834
10.9956
12.5664
12.5672
12.5743
12.5823
12.6060
12.6453
12.7223
12.9352
13.2142
13.9981
14.1215
14.1372
15.7080
15.7086
15.7143
15.7207
15.7397
15.7713
15.8336
16.0107
16.2594
17.1093
17.2596
17.2788
18.84%
18.8501
18.8549
18.8602
18.8760
18.9024
18.9547
19.1055
19.3270
20.2208
20.3977
20.4204
21.9911
21.9916
21.9957
22.0002
22.0139
22.0365
22.0815
22.2126
22.4108
23.3327
235358
23.5620
25.1327
25.1331
25.1367
25.1407
25.1526
25.1724
25.2119
25.3276
25.5064
26.4450
26.6739
26.7036
28.2743
282747
28.2779
28.2814
28.2920
28.3096
28.3448
28.4483
28.6106
29.5577
29.8120
29.8452
Sec. 2.31
Non-steady
state with a finite coefficient
of matter transfer
39
and the dimensionlessM is given by :

(2.53) M = +h
The amount of substanceentering (or leaving) the sheetup to time t, M,, is given as
a fraction of the correspondingquantity after infinite time, a, by :
(2.54)
M;Mt = 2
00 n=l
2M2
exp
p;(~~+ M2+ Ml
Of course, depending on the relative values of the concentrations C, and C,, the
substanceentersor leavesthe sheet.
i-i5
as a function of L,
ca
with the thickness2 L of the sheet
Fig. 2.5. Kinetics of sorption (or desorption)by plotting 2

for various valuesof M = g,
Mathematical
40
treatment
of diffusion
in a plane sheet
(2.55)
c,
cca
substance enters the sheet
(2.56)
c,
< co3
substanceleaves the sheet
[Ch. 2
Roots of eqn (2.52) are given in Table 2.1, for various values of the dimensionless number M = T,
with the thickness 2L of the sheet.
Sorption (or desorption) curves for the finite rate of matter transfer at the
W
surfaces are shown in Fig. 2.5 by plotting the ratio - as a function of the dimensionless
MOO
for various values of the dimensionless number M = g.
number F,
REMARK
I - The case of a finite value of the coefficient of matter transfer h on the surface
is also the case of the evaporation of the liquid out of the surface with a finite rate of
evaporation, or the case of condensation of the vapour on the surface with a finite rate of
condensation.
REMARK 2
- High values of h and M.
For very high values of the coefficient of matter on the surface h, M can be
considered as infinite, and the series in eqn (2.54) reduces to
(2.18)
as
(2.57)
for
p,,=(n+k)rr
M +=J
This equation has already been found for h + 00, and a constant concentration
on the surface of the sheet.
2.4 NON-STEADY
STATE DIFFUSION IN A STIRRED
SOLUTION OF LIMITED VOLUME
Very often, the sheet is not immersed in a volume of solution so large as compared
Sec. 2.41
Non-steady
state diffusion
in a stirred solution
of limited volume
41
with the volume of the sheet.The volume of the solution cannot be consideredas infinite,
and the concentrationof liquid on the surfaceof the sheetis not constant.
When the finite volume of solution is well stirred, the concentration through this
volume is uniform. The main condition is that the total amount of diffusing substancein
the solution and in the sheetremainsconstantduring the process.
The thicknessof the sheetis 2 L and that of the solution is 21.
Two casesare considered: when the substanceinitially in the solution diffuses the
sheet,and when the substanceleavesthe sheetand entersthe solution.
2.4.1 ABSORPTION OF DIFFUSING SUBSTANCE
The diffusion equationwith a constantdiffusivity :
(2.1)
= D.
BY THE SHEET
J2C
ax2
is resolved(2.1) with the initial conditions

-LIxIL
t =o
(2.58)
-L-LSXI-L
LSxlL+L
c=o
C = Ci~
C = Ci~
sheet
solution
solution
and the boundarycondition :

(2.59) t > 0
x = fL
L.g
= D.
The boundary condition expressesthat the rate at which the substanceenters the
sheetthrough the surfacesx = f L is constantly equal to the rate at which the substance
leavesthe solution.
The ratio of the volumes of solution andsheetis a, with the partition factor K
(2.60) a = &
where the partition factor K is given by the ratio of the concentrationsat equilibrium in the
sheetand the solution
Mathematical
42
treatment
of diffusion
[Ch. 2
in a plane sheet
C,inthe sheet
(2.61) K = C, in the solution
The ratio 01can be expressedin terms of the fraction of the substanceabsorbedby
the sheetat equilibrium The matter balanceis asfollows :
(2.62)
+ L* C, L. Ci,
where C ~ and K ate the uniform concentrationof the substanceat equilibrium in the sheetand
in the solution, respectively.
The amountof substanceabsorbedby the sheetat equilibrium, M, is given by :
M, = 2 L. C,
(2.63)
2 L . Ci,
1+ a
The amount of solution in the sheet at time t, M,, as a fraction the corresponding
amount after infinite time, W, is given by (1,2,3)
(2.64)
M;Mt
ca
O 2a(l+a)
= c
n=l l+a+a
22
exp
qn
where the q,$ are the non-zeropositive roots of

tanq,=-a.q,
(2.65)
The concentrationdistribution within the sheetis given by the relation :
cos9nX
L
(2.66)
2 2
= 1 + 2n=l 1 +2(1+a)
CI+ 01qn
cos%I . exp 1-9nDt

L2
Someroots of eqn (2.65) are given in Table 2.2 for various values of the ratio 01.
Sec. 2.41
Non-steady
state diffusion
in a stirred solution
Table
of limited volume
43
2.2
Roots of tan q,, = -a.q,

a
41
42
%l
q4
QS
46
47
48
99
410
1.5708
4.7124
4.7359
4.7648
4.8014
4.8490
4.9132
5.0037
5.1386
5.3540
5.7172
6.2832
7.8540
7.8681
7.8857
7.9081
7.9378
7.9787
8.0385
8.1334
8.3029
8.6587
9.4248
10.9956
11.0057
11.0183
11.0344
11.0558
11.0856
11.12%
11.2010
11.3349
11.6532
12.5664
14.1372
14.1451
14.1549
14.1674
14.1841
14.2075
14.2421
14.2990
14.4080
14.6870
15.7080
17.2788
17.2852
17.2933
17.3036
17.3173
17.3364
17.3649
17.4119
17.5034
17.7481
18.8496
20.4204
20.4258
20.4326
20.4413
20.4529
20.4692
20.4934
20.5335
20.6120
20.8283
21.9912
23.5620
23.5667
23.5726
23.5801
235902
23 A043
23.6253
23.6602
23.7289
23.9218
25.1328
26.7036
26.7077
26.7129
26.71%
26.7285
26.7409
26.7S95
26.7904
26.8514
27.0250
28.2744
29.8452
29.8489
29.8535
29.8595
29.8674
29.8786
29.8953
29.9229
29.9778
30.1354
31.4160
1.6385
9.Goo
4.OOoO 1.7155
2.3333
1.5000
1.0000
0.6667
0.4286
0.2500
0.1111
0.0000
1.8040
1.9071
2.0288
2.1746
2.3521
2.5704
2.8363
3.1416
The ratio of the amountsof substanceabsorbedat time t and infinite time,

03-t
Mt
is plotted as a function of the dimensionlessnumber - L , for various values of a.
M,
(Fig. 2.6).
2.4.2
DESORPTION
SHEET
IN THE
OF DIFFUSING
SUBSTANCE
FROM
THE
SOLUTION
The initial conditions are different from the oppositecasewhen the substanceenters
the sheet.
(2.58)
t =o
-LIxIL
-L-LIX-L
LSx<L+L
C = Ci*
c=o
c=o
sheet
solution
solution
The kinetics of substanceleaving the sheetis also expressedby eqn (2.54).

The amountof substancetransferredin the solution at equilibrium M, is given by :
(2.67)
M,=2LI;cm=2L(Ci,-C~)
L
and the ratio a = K.L is thus :
Mathematical
44
treatment
of diffusion
in a plane sheet
[Ch. 2
Fig. 2.6. Finite volume of solution. Kinetics of sorption (or desorption)by plotting
w as a fonction of the dimensionlessnumber Lv-n-7 for various valuesof the ratio

M
OL
a=----K.L
(2.68)
L=
l+a
M2L*cin
The concentrationdistribution within the sheetis expressedby the relation :
(2.69)
tin-
tinREMARK I
cX.t
cca
=l+
2(1+a)
c
22
n=l l+cX+a qn
cos f&!F
-exp
cos%l
- High value of the surrounding volume

When the volume of the surroundingvolume is very high, the ratio of the volume cx
Sec. 2.51
Steady-state
with a membrane
45
becomesinfinite, and the eqn (2.64) expressingthe kinetics of the substancetransferred

reducesto :
(2.18)
MzMt
= ?.-.IF
.exp -f2n+1)2n2Dt
(2n+ l)*
4L2
as
when
(2.70) qn = n + i
2.5 STEADY-STATE
a+=
WITH A MEMBRANE
As shown alreadyin the caseof a membranewith a non-steadystateof the diffusion

of the substance(section 2.2.3), after a given time, the diffusion can be considered as
achievedin steadystate.
The membraneof thicknessL hasits surfacesmaintainedat constantconcentrations:
Ci for x = 0 and Co for x = L. When the steady state is reached, the concentration of
substanceis independent of tune and remains constant at all places of the sheet. The
equationof diffusion in one dimensionreducesto :
(2.71) k
= 0
dx*
with constantD
2.5.1 HIGH VALUE OF THE COEFFICIENT

ON THE SURFACE
By integrating with respectto x, it becomes:
(2.72) g
OF MATTER TRANSFER
= constant
and by anotherintegration taking into accountthe constantconcentrationsat x = 0 andx =

L, there is :
(2.73)
cat
-- Cici=;
Mathematical
46
treatment
of diffusion
in a plane sheet
[Ch. 2
The rate at which the diffusing is transferred the section of the membrane of unit
area Ra is given by :
Ra=-D
274
dC_ o(ci-co)
dx-
Of course, in this case, the concentration Ci and Co are the same as those in each
surrounding atmosphere when the partition factor K is 1. When this partition factor is not
1, there is the obvious relation : K . Ci ext = C . Ci showing that the concentration of
substance on the surface of the membrane is K times the concentration in the surrounding
Ciext with which it is in equilibrium.
2.5.2 CASE OF FINITE VALUE OF THE COEFFICIENT

TRANSFER
OF MATTER
For instance, the coefficient of matter transfer is high at the surface x = 0 and finite
at x = L. Thus, the surface x = 0 is maintained at a constant concentration Ci, while the
condition at the surface x = L is given by :
(2.75)
-D.$=h(C,-C,d
x=L
where h is expressed in cm/s, and

where C,, is the concentration on the surface x = L required to maintain equilibrium with
the external atmosphere.
In this case, the integration gives the profile of concentration
(2.76)
C,-
Ci
cOeq-
ci
h.x
= D#+h.
and the constant rate at which the substance diffuses is
(2.77)
Ra =
D. h(Ci-C,g)
D+h.L
When the coefficients of matter transfer are finite on each surface, h at x = 0 and ho
at
x= L
Sec. 2.51
Steady-state
(2.78) -Do
= hi(C,g-
(2.78) - Dg
= ho@,- C&P)
with a membrane
47
Ci)
where Cieqand C,, are the concentrationson the surface x = 0 and x = L, required to
maintain equilibrium with eachcorrespondingexternalatmosphere.
The concentrationat position x, within the sheet,is given by :
(2.79) C,=
hi. Ci,, D+ho(L-X)

+ho. Coe4[D+hix]
D(hi+ho)+hi. ho. L
andthe rate at which the substanceis transferredthrough the membrane
(2.80) Ra=
D. hi. ho[C*q-Cd
D(hi+ho)+hi. ho. L
2.5.3
COMPOSITE MEMBRANE
In the case of a composite membrane made of n sheetsof thickness Lr, L,.....L,
with the diffusivities Dr, Dz.....D,, the drop in concentration for the whole composite
membraneis the sum of the drop in concentrationfor eachmembrane.
From the drop in concentrationexpressedby eqn (2.74), there is
(2.81)
L.
= 2 representsthe resistanceto diffusion of the membranei.
where Ri Di
Of course, this caseis obtained when the sheetsare in perfect contact, and there is
no additional resistancebetweeneachmembrane.
When there are the coefficients of matter transfers ht, h,.....h, between the layers,
the fall in concentrationfor the compositemembraneis :
(2.82) Ci-Co=Ra
z+g
. . . . .k+k+&
. . . . .k
2
48
Mathematical
treatment
of diffusion
in a plane sheet
[Ch. 2
2.5.4 MEMBRANE SEPARATING GASES OR VAPOURS

In the above systems, the membrane is separating two liquids, and the rate of
transfer is expressedin terms of the concentrations.
When the membraneseparatestwo gasesor vapours, it is easierto expressthe rate
of substancetransferredin terms of the pressurePi and Poof this gas at the inlet and outlet
of the membrane.
(2.83)
Ra=
Per (Pi - PO)

L
and Per represents the permeability of the membrane Per is expressed as cm3 gas
measuredat standardtemperatureand pressuremoving through a section of 1 cm2 of the
membrane 1 cm thick when the difference in pressure acrossthe membrane is 1 cm of
mercury.
In the simple casewhere the sorption isotherm of vapour by the membraneis linear,
the concentration of liquid at equilibrium with the surrounding atmospherewhere the
pressureis P is given by :
(2.84)
c = S.P
where S is the solubility of the vapour in the membrane.

Following these two above relations, the permeability is connected with the
solubility and diffusivity.
(2.85)
Per = D.S.
REFERENCES
1
H.S. Carslaw and J.C. Jeager,in Conduction of Heat in Solids, Clarendon

Press,Oxford. 1978, Chapter 3.
J. Crank, in The Mathematicsof Diffusion, ClarendonPress,Oxford, 1975,

Chapter4.
J. Bouzon and J.M. Vergnaud, in Mathematical and Numerical Treatment of

Diffusion PrenticeHall, in press.
Mathematical
treatment
an isotropic rectangular
3.1
of diffusion in
parallelepiped
INTRODUCTION
Rectangular parallelepiped dosageforms are sometimesused, and this chapter is

thus of practical interest. From a theoretical point of view, the caseof a sheethas been
considered as a rectangular parallelepiped whose one side (thickness) is very thin as
compared with the other two sides, and the effect of the edges on the total matter
transfer is neglected.It is of concern in this chapter to define the dimensions for which
this assumption can be made, and when a thin parallelepiped can be considered as a
sheetwith a one-directional transport of matter.
Some materials are anisotropic, e.g., wood or stretchedpolymers. But in the case
of dosage forms, the polymer can be considered as isotropic. The problem of an
anisotropic rectangular parallelepiped with three principal axes of diffusion and three
principal diffusivities can be seenin other books (1, 2).
The diffusivity is constant.
The coefficient of matter transfer on the surface is either infinite (3.2.1) or finite
(3.2.2) as comparedwith the diffusivity and dimensionsof the parallelepiped.
When the coefficient of matter transfer on the surfaceis infinite, as already shown
in Chapter 1, the concentration on the surface is constant and is equal to the value
which is at equilibrium with the surrounding. This concentration is also the concentration in the parallelepiped attained after infinite time.
When the coefficient of matter transfer on the surface is not infinite, there is a
surface condition expressingthat the rate at which the matter is transferred through the
surface is constantly equal to the rate at which the diffusing substanceis brought to the
surfaceby internal diffusion.
50
Mathematical
treatment
of diffusion
[Ch. 3
The volume of the surrounding atmosphereis so large with regard to the volume
of the parallelepiped that the surrounding volume may be considered as infinite.
Moreover the surrounding atmosphereis well stirred. The concentrationof the diffusing
substance in the surrounding atmosphere remains constant and uniform during the
process.
No changein dimensions of the parallelepiped is consideredduring the processof
absorption or desorption, whatever the amount of liquid absorbed or desorbed. This
assumptionis necessaryfor the mathematicaltreatment .
The equation of the three-dimensionaldiffusion with constantdiffusivity is given
by :
(3.1) ?$=D [$+z$+$]
where x, y and z are the coordinates taken along the three perpendicular axes parallel
with the sidesof the rectangularparallelepiped.
3.2. ISOTROPIC RECTANGULAR PARALLELEPIPED

CONSTANT DIFFUSIVITY
WITH A
The following two casesare studied :

- when the coefficient of matter transfer on the surfaceis so high that it can be
consideredas infinite
- when the coefficient of matter transfer is not infinite.
The diffusivity is constant.
The initial concentration of the diffusing substance in the parallelepiped is
uniform.
3.2.1 INFINITE COEFFICIENT OF MATTER TRANSFER ON THE SURFACE.
CONSTANT CONCENTRATION ON THE SURFACE
As the coefficient of matter transfer on the surfaceis infinite, the concentrationof
the diffusing substanceon the surface of the parallelepiped is constant as soon as the
process starts. As already shown in Chapter 1 (1.1.5), the boundary condition
expressing that the rate at which the substance is transferred through the surface is
constantly equal to the rate at which the substancediffuses within the solid next to the
surface :
(3.2) -D(g),=
h(C,- Ces)
surface
Sec. 3.21
Isotropic
rectangular
parallelepiped
with a constant
diffusivity
51
where C, is the concentrationon the surface

and Cesis the concentrationon the surfacewhich is at equilibrium with the surrounding
atmosphere,which is also noted C, as it is the value of C, after infinite tune.
When h tends to infinity, the concentration C, and C,, are equal, and the
concentration on the surface reachesthe value at equilibrium as soon as the process
starts.
The rectangular parallelepiped of sides 2a, 2b, 2c, is located between the
coordinatestaken along the x-, y-, z- axes : - a < x < a, - b < y c b , - c < z < c.
The initial concentration in the parallelepiped is Ci, and the constant
concentration on the surface is C, or C,,. This constant concentration C, is related to
the constantvalue of the surrounding atmosphereC,,.
(3.3)
c,
= K. c,,
where K is the partition factor expressedin Chapter 1.

USE OF TRIGONOMETRICAL
SERIES
The concentrationwithin the parallelepiped after time t is thus expressedin terms

of the four parametersx, y. z, t : C, ,,,=,t. This concentrationcan be obtained by making
the product of the three solutions given in the case of a sheet with a one-dimensional
transport (Chapter 2).
(3.4)
Cx.Y.z,t-
c-
Gin- C,
w I= 64 &-cos
1)
TF2m=O2m+1
p.
n=O2n+l
. g_p.
2p+l
COS
cos
(2m+llrtx . exp _ (2m+1)27t2Dt

2a
4a2
I
(2n+l) 7tx
2b
. exp[-2n~~~2
Dt]
(2p+l) xx
2c
. exp[-2p:iF2
IX]
Pa
The amount of diffusing substance transferred through the surface of the

parallelepiped after time t, M,, as a fraction of the corresponding amount after infinite
time, M,, is expressed in terms of time t by the product of the three series
52
Mathematical
treatment
of diffusion
tCh. 3
correspondingwith a one-dimensionaltransport.
In the caseof a cube of side 2a, the abovetwo equationsreduceto :

For the concentrationdistribution Cx,Y.Z,t at eachtime t :
O (-l)m
*jzl
(2m+ 1)~
cos(2m+
l)nx
. exp 2a
! 4a2
Dt
,:I]
For the amount of the diffusing substancetransferredup to time t :
USE OF ERROR FUNCTION
Other solutions are expressedin terms of the error function.

The dimensionless number expressed in terms of the concentrations of the
substanceis given by the product of the following (three solutions)
(3*8)cx.y.z.t
- c.2= f (x,a). f (y,b) . f (z,c)
tin-
Coo
eachof them being the solution for a one-dimensionaltransport
Sec. 3.21
Isotropic
rectangular
(3.9) f(x, a) = 1- z(-
parallelepiped
with a constant
diffusivity
53
l)merfc(2m~~-x
m=O
+ C(
-l*. )
erfc
m4
(2m+l)a+x
21Dt
The amount of diffusing substancetransferredafter time t, M,, is given by :
(3.10) 2
Mt = cplx,4 cply,W . cp(4
where each function is the solution obtained for a one-dimensionaltransport :
(3.11) cp(x, a) = 1-2-
1(Dt 1+2w-lm.
C(
a 2/31 mll
I
ma
) .1erfc7s
I - Small time of diffusion.

For small time of diffusion, when Mr/M, < 0,4-0,6 depending on the desired
accuracy,the seriesin eqn (3.11) vanishes,and a simple relation is obtained :
REMARK
By plotting M,fM, as a function of the square root of time, the slope of the
tangentattheorigini{~+~+~~~,
enablingone to obtain the diffusivity.
The amount of the substancetransferedthrough the surfaceof the parallelepiped,

Mt is plotted asa functionof the dimensionlesslumber -m forvariousvaluesof f and
-9
a
MO3
for a = b (Fig.3.1). The valuesof E arenoticed.
- Long tune of diffusion
For long times of diffusion, when M&M, > 0,7, the first term in the exponential
REMARK 2
54
Mathematical
treatment
0.25
[Ch. 3
of diffusion
0.75
0.5
Fig. 3.1. Amount of mattertransferred 2
Q,
numberaYEi for various valuesof c and for a = b. Constantvaluesof the

concentrationon the surface.Thevalues of E arenoticed.
seriesbecomespreponderant,and the amount of matter transferred can be expressedin

terms of tune by the simple relation :
(3.13) MzwMt
= y.
exp[-($+-$+$]+I
as a function of time, a straightline is obtained,the slopeof

which the diffusivity can be determined.
Isotropic
Sec. 3.21
rectangular
parallelepiped
with a constant
diffusivity
55
3.2.2 FINITE COEFFICIENT OF MATTER TRANSFER ON THE SURFACE

The condition on each surface is written by expressing that the rate of transfer
through the surface is constantly equal to the rate at which the diffusing substanceis
transferredby internal diffusion next to the surface.
-D.
(3.2)
g=h(C,-CW)
the x-axis being replaced by y-or z-axis, where C, is the actual concentration of the
substanceon the surface,
C, the concentration of substanceon the surface required to maintain equilibrium with
the surrounding atmosphere,
and h the coefficient of matter transfer through the surface (cm/s).
For the rectangularparallelepiped of sides 2a, 2b, 2c, with : - a < x < a, - b c y <
b, and - c < z c c, the concentration distribution and the amount of matter transferred
are thus given by the product of each solution obtained with each one-dimensional
transfer.
(3.14)
cd..
m-
= f(x,a,h) . f(y,b,h) . f(z,c,h)

00
with the function f expressedby :
2x.
co4
f(x, a, h) = c
(3.15)
. exp
m=l (p~+x2+x)cosp,
where the &,s are the positive roots of

(3.16)
p.tanp=x
and X is the dimensionlessnumber :

(3.17)
x = g
The amount of substancetransferred through the surface of the parallelepiped
Mathematical
56
treatment
of diffusion
0.25
[Ch. 3
0.5
0.75
00
a.h
number -lTTi for various values of the dimensionlessnumber X = Dand
for a
a
cube. Finite coefficient of matter transfer on the surfaceh.
during the time t, M,, is expressed as a fraction of the corresponding quantity after
infinite time, K, by the product of the functions cp:
(3.181
Ms- M,
M
= cp(x,a,W. cp(y,b,h) . qNw,h)
where each function cpis of the form :
(3.19)
cp(x,a, h) =
exp
Sec. 3.21
Isotropic
rectangular
parallelepiped
with a constant
diffusivity
57
Some roots of eqn (3.15) are given in Table 2. 1 for various values of X.
The valuesof 2
areplotted asa functionof the dimensionlessnumbera

liDt for
00
variousvahresofthe dimensionlessnumberX= ah in Fig.3.2. ForX-+w,when the
D
coefficient of matter transfer on the surface h is infinite, the curve is the same as that
obtained with a constant concentration of the substanceon the surface as soon as the
processstarts.
I - Absorption or desorption of the substance.
Depending on the relative values of the concentrations of the substanceon the
REMARK
0
0
025
05
075
as a function of a?mt for various values

00
of the thickness:c, with a = b and for X = $ = 1.
Effect of the edgeson the one-dimensionaltransfer in a sheet,with a finite

coefficient of matter transfer on the surface.
58
Mathematical
treatment
of diffusion
[Ch. 3
surfaceof the parallelepiped, the substanceentersor leaves the solid.

(3.20)
G>cw
desorption or evaporation
(3.20)
c,<c,
absorption,or condensation
3.3 EFFECT OF THE THICKNESS
OF A SHEET
A sheetis a particular rectangular parallelepiped for which the thickness is much

less than the other two-dimensions.
It is thus possible to determine the effect of the edges on the total transfer for
various values of the thickness2a with respectto the other sides2b and 2c.
The effect of the edges is thus quantitativity determined in the following two
cases : when the coefficient of matter transfer on the surface is so high that the
concentration of the substanceon the surface is constant ; and when the coefficient of
matter transfer on the surfaceis finite.
When the coefficient of matter transfer on the surface is infinite, the effect of the edges
can be seenin Fig.3.1 where the amount of matter transferredis plotted as a function of
llDt for various values of the thicknessC and for a = b. The values of t arenoticed.
-aWhen the coefficient of matter transfer on the surfaceh is finite, the effect of the edges
is shownin Fig.3.3 wherethe amountof mattertransferredsplotted asfunctionof -m
a
for various values of the thicknessc and for a = b, and for a given value of the
a.h
dimensionlessnumber X = -D- = 1, a being half the thicknessof the sheet.
REFERENCES
1
H.S. Carslaw and J.C. Jaeger,in Conduction of Heat in Solids Clarendon

Press,Oxford, 1978, chapter6.
M. Vergnaud and J. Bouzon, in Mathematical and Numerical Treatment of

Diffusion, Prentice Hall ., N.J., 1993, chapter 5.
4
Mathematical
treatment
diffusion in a sphere
4.1
of radial
INTRODUCTION
The caseof the sphereis of high interesteither from a theoretical or from a practical
point of view. Spherical oral dosageforms are widely used, especially for the controlled
delivery of the drug in the stomach.
The mathematical treatment of diffusion is only feasible when the diffusivity is
constant, and when the initial distribution in the sphere is uniform or is described by a
function which can be integratedwith respectto space.
The diffusion of the substanceis radial only.
The equationof radial diffusion in non-steadystateis :
(4.1)
5 = $. $ Dar*,
$!f
r
[ 1
when the diffusivity is not constant,

and it is reducedto :
or
60
(4.2)
Mathematical
treatment
ac
2 ac
= D. ar2+T.
ar
of radial diffusion
in a sphere
LCh.4
when the diffusivity is constant.
4.2
SOLID
CONSTANT
SPHERE
IN
DIFFUSIVITY
NON-STEADY
STATE
WITH
Dependingon the boundaty conditions and especially the value of the coefficient of
matter transfer on the surface, and on the volume of the surrounding atmosphere,three
main casesareconsidered:
(i) When the coefficient of matter transfer on the surfaceis infinite, and the concentration
on the surfaceis constant.
(ii) When the coefficient of matter transfer on the surfaceis finite, and the concentrationof
the surroundingatmosphereis constant.
(iii) When finite volume of the surroundingatmosphereis well-stirred.
4.2.1 INFINITE COEFFICIENT

OF MATTER TRANSFER ON THE
SURFACE
As already shown in Chapter 1 for the boundary conditions, when the coefficient of
matter transfer on the surfaceis infinite, the concentrationof the substanceon the surface
is constant and equal to the value which is at equilibrium with the surrounding
atmosphere.
The solid sphere of radius R with an initial concentration of substancewhich is
either uniform or function of r is considered.
(4.3)
t=O
or
t=O
(4.3)
t>O
r<R
r<R
r=R
Upon putting the function U

(4.4)
U = C.r
C = f(r)
C = Ci~
c = c,
solid sphere
solid sphere
surface
Sec. 4.21
Solid sphere in non-steady
state with constant
diffusivity
61
the equationfor U becomes:
(4.5)
a*u
= D. arz
with the following initial andboundarycondition :

t = 0
r < R
U = r.f(r)
(4.6)
t = 0
(4.7)
t>o
r < R
r= 0
r=R
U = r.Ci,,
u=o
U = R.C,
(4.6)
or
surfacer = 0
surfacer = R
The equation for U with the initial and boundary conditions correspondswith the
problem of the diffusion through a plane sheetof thicknessR with constantconcentrations
on each surface, e.g. at r = 0. U = 0, and at r = R, U = R.C,, and the initial distribution
r.Cin or r.f(r). The solution of this problem is given in the caseof the membrane(Chapter
2).
When the initial concentrationis f(r) the concentrationdistribution is thus given by :
o r[f/r)-Cdsin$dr
When the initial concentration is uniform f(r) = Ci,, the concentration distribution
is :
(4.9)
C
Cr- cin = 1 + z.
00~tin
O (-1) . nrcr
c n,
sm R
exp
n=l
At the centreof the sphere,asy
+ 1 when y + 0, the aboveequationbecomes:
62
(4.10)
Mathematical
ycin
00~tin
treatment
= 1+2
of radial diffusion
00
co j
n=l
in a sphere
LCh.4
-5.LDt
-1 n. exp
R2
The amount of substance transferred after time t, M,, as a fraction of the

corresponding quantity after infinite time, M, is expressedin terms of time by the
following relation :
(4.11)
We-M,
- =Dt
R2
= 4.
rc
Another solution is given by using the error functions.

The concentrationdistribution is :
erfc (2n+ l)R-r
2v-E
erfc (2n+ l)R+r

-
21'Dt
and the amountof substancetransferredMt asa fraction of M, is :
(4.13) 2 Do= 6~[~+2~ierfc~]-~
The concentrationdistribution is shown by plotting the dimensionlessnumber C, C, / C, - Ci, as a function of the dimensionlessnumber r / R for various values of the
dimensionlessnumbers m / R (Fig. 4.1)
The amount of the matter transferred,Mt / K, is expressedas a function of
the dimensionlessnumber Ja I R (Fig. 4.2)
4.2.2 FINITE COEFFICIENT
OF MATTER TRANSFER
ON THE
SURFACE
In the caseof the uniform initial concentrationin the sphere,the initial andboundary
Sec. 4.21
state with constant
diffusivity
63
A
Cr,t
a707
-[in
Cal-C
in
75-
Fig. 4.1.
Concentration distribution (CB - Ci, / C, - C,) as a function of r / R,
for various values of m
/ R. Constant concentration on the surface.
conditions are as follows :
(4.14)
(4.15)
t=O
t > 0
O<rcR
r=
C = Ci~
-D.g
= h(C,-CJ
where C, is the actual concentration on the surface at time t, and C, is the concentration
on the surface which is at equilibrium with the surrounding atmosphere.
The concentration distribution within the sphere is given by the relation :
64
Mathematical
treatment
of radial diffusion
=-
4.16
in a sphere
rCh.4
. exp
Mt%
Kl
15
50
Fig. 4.2. Amount of matter transferredMJMooasa function of the dimensionless

number I%? / R . Constantconcentrationon the surface.
where the dimensionlessnumber S is :
(4.17) s = h+
and the p, s are the roots of :
(4.18 ) p,. cot p,, = 1 -S
For the centre of the sphere,r = 0, and the above eqn (4.16) cannot be used. By
Sec. 4.21
state with constant
diffusivity
65
using the relation sin x / x + 1 when x + 0, there is :
(4.19) ,t-,= 2s.

Ml- Someroots of eqn. (4.18) are given in Table 4.1.
Table 4.1
Rootsof p.cotp = 1-S with S =h+
s
Pl
I32
P3
P4
P5
p6
P7
p8
fl9
PlO
3.1416
6.2832
6.2706
6.2204
6.1582
5.7172
5.0036
4.7124
4.6042
4.5601
4.5379
4.5157
4.5045
4.4956
4.4934
9.4248
9.4059
9.3308
9.2384
8.6587
8.0385
7.8540
7.7899
7.7641
7.7511
7.7382
7.7317
7.7265
7.7253
12.5664
12.5412
12.4414
12.3200
11.6532
11.1295
10.9956
10.9499
10.9316
10.9225
10.9133
10.9087
10.9050
10.9041
15.7080
15.6766
15.5521
15.4034
14.6869
14.2421
14.1372
14.1017
14.0875
14.0804
14.0733
14.0697
14.0669
14.0662
18.8496
18.8119
18.6632
18.4888
17.7481
17.3649
17.2788
17.2498
17.2382
17.2324
17.2266
17.2237
17.2213
17.2208
21.9912
21.9472
21.7746
21.5764
20.8282
20.4934
20.4203
20.3958
20.3860
20.3811
20.3762
20.3738
20.3718
20.3713
25.1328
25.0825
24.8865
24.6664
23.9218
23.6254
23.5619
23.5407
23.5322
23.5280
23.5237
23.5216
23.5199
23.5195
28.2744
28.2178
27.9987
27.7589
27.0250
26.7595
26.7035
26.6848
26.6773
26.6736
26.6698
26.6679
26.6664
26.6661
31.4160
31.3532
31.1114
30.8540
30.1354
29.8953
29.8451
29.8284
29.8217
29.8183
29.8150
29.8133
29.8119
29.8116
5f&o 3.1353
100.003.1102
50.00
10.00
2.50
1.00
0.50
0.30
0.20
0.10
0.05
0.01
0.00
3.0788
2.8363
2.1746
1.5708
1.1656
0.9208
0.7593
0.5423
0.3854
0.1730
0.0000
The amount of diffusing substancewhich has entered(or left) the sphereafter time t,
M,, is expressedin terms of time by the ratio M, / M,.
where M, is the amount of substancetransferredafter infinite time, when the equilibrium

is reached.
Mathematical
66
treatment
of radial diffusion
LCh.4
in a sphere
The amount of substancetransferred after time t as a fraction of the corresponding

amountafter i&mite time, Mt / M, is expressedin terms of the dimensionlessnumber
mi for various values of the dimensionlessnumber S = h.R
R,
D (Fig.4.3). As shown in
this figure, the kinetics obtainedfor S + 00is the sameasthat obtainedin (Fig. 4.2) for
a constantconcentrationon the surface.
0.25
0.5
a75
Fig. 4.3. Amount of matter transferredM&L asa function of the dimensionless

number 1(Dt / R, for various valuesof the dimensionlessnumber S = h . R / D.
Finite coefficient of matter transferon the surface.
4.2.3
DIFFUSION
SURROUNDING
BETWEEN
ATMOSPHERE
A SPHERE
OF FINITE
AND
A WELL-STIRRED
VOLUME
This case is very frequent, and the ratio of the volumes of the sphere and the
surroundingatmosphereis of importance.
Sec. 4.21
state with constant
diffusivity
67
The sphere of radius R is located in the surrounding atmosphere of volume V. The

concentration in the surrounding atmosphere is always uniform, as it well-stirred.
Two cases are considered, when the substance leaves or enters the sphere
THE SUBSTANCE LEAVES THE SPHERE
The initial concentration in the sphere is uniform C,. The ratio a of the volumes of
the surrounding atmosphere and sphere is given by :
(4.21)
a =
3v
4 n R3.K
where K is the partition factor defined in Chapter 1 (1.1.6).

The ratio 01 can be determined in terms of the amount of substance which is
transferred after infinite time, M,, and of the amount of substance initially in the sphere.
3 Mce
4n:R3Cin
(4.22)
= -?1+a
THE SUBSTANCE ENTERS THE SPHERE
The surrounding atmosphere is initially at the uniform concentration C, in, and the
amount of substance transferred after infinite time is K.
The ratio a is thus expressed by
the relation :
M,
(4.23)
1
=
1+a
In both cases, the concentration of the substance within the sphere C, t is expressed
in terms of time t, as a fraction :
(4.24)
tin - cr,t =
tin-
. exp
cca
where C, is the uniform initial concentration in the sphere, an C, is the concentration in

the sphere which is at equilibrium with the surrounding atmosphere.
Mathematical
68
treatment
of radial diffusion
in a sphere
ECh.4
For the centreof the sphere,as sin x I x + 1 when x + 0, this equation becomes:
(4.25)
tin-
CO.t
ci*- coa
In thesetwo cases,the amountof substancetransferredafter time t, M,, is expressed

in terms of time as a fraction of the correspondingquantity transferred after infinite time,
K, by the relation :
(4.27) tan q,, =
3 9n
3+cLqi
Someroots of eqn(4.27) are given in Table 4.2 for various values of the ratio a.
The ratio of the amountsof substancetransferredafter time t and after infiite time,
llDt
Mt
.
- is expressedin terms of the dimensionlessnumber - R , for various values of the ratio
MC0
CL(Fig. 4.4).
Table
4.2
3%
Roots of tan qn =
3+a. q;
93
q4
95
q6
q7
q8
q9
GO
9.k
4.0000
3.1416
3.2410
3.3485
3.4650
3.5909
3.7264
3.8711
4.0236
4.1811
4.3395
4.4934
6.2832
6.3350
6.3978
6.4736
6.5664
6.6814
6.8246
7.0019
7.2169
7.4645
7.7253
9.4248
9.4559
9.5029
9.5567
9.6254
9.7156
9.8369
10.0039
10.2355
10.5437
10.!4041
12.5664
12.5928
12.6254
12.6667
12.7204
12.7927
12.8940
13.0423
13.2689
13.6134
14.0662
15.7080
15.7291
15.7554
15.7888
15.8326
15.8924
15.9780
16.1082
16.3211
16.6831
17.2208
18.8496
18.8672
18.8892
18.9172
18.9541
19.0048
19.0785
19.1932
19.3898
19.7565
20.3713
21.9912
22.0063
22.0251
22.0492
22.0811
22.1251
22.1895
22.2914
22.4719
22.8350
23.5195
25.1328
25.1460
25.1625
25.1836
25.2117
25.2504
25.3075
25.3989
25.5647
25.9189
26.6661
28.2744
28.2861
28.3008
28.3 1%
28.3446
28.3793
28.4305
28.5131
28.6656
29.0082
29.8116
31.4160
3 1.4265
3 1.4398
3 1.4567
31.4792
31.5106
31.5570
31.6322
3 1.7732
32.1025
32.9564
2.3333
1.5000
1.0000
0.6666
0.4286
0.2500
0.1111
0.0000
Sec. 4.31
Hollow
sphere in non-steady
state
69
Of course,for a high value of CC,e.g. higher than 20, the curves are similar to that
obtainedwith an infinite volume of the surroundingatmosphere.
025
0.5
035
Fig. 4.4. Amount of matter transferredM&L, as a function of the dimensionless

number m / R, for various values of the ratio cr..The valuesof cxarenoticed.
4.3 HOLLOW SPHERE IN NON-STEADY STATE

The hollow sphereis of special interest for dosageforms, as it plays the role of a
spherical membraneseparatingtwo media. Generally the internal medium is full of drug,
and the surrounding atmosphereis free of drug. The concentrationon the internal surface
is thus kept constantat least for sometime, aswell as on the external surface.
The hollow spherelocatedbetweenthe radii R and %, with :
Ri < r < R,
Mathematical
70
treatment
of radial diffusion
Kh.4
in a sphere
is initially free from diffusing substanceand the surfacesr = R is maintained at Ci and r =

R, at 0.
(4.28)
t=O
(4.29)
t>O
Ri < r < R,
r = Ri
r= Re
c=o
C = Ci = constant
c=o
The concentrationat the radius r and time t, C,, is given by
C,, r =
(4.30)
Ci.Ri(R,-r)
-- 2 O a.shnn(r-RJ
Re-Ri
c
=I
exp
R,-Ri
n2n2
i
Dt
R,- Rd2
The amount of diffusing substanceleaving the external surface r = R,, M,, is thus
given by :
(4.3 1)
4nRtiRe(Re-Ri)Ci
Dt
1 2
----.
6 n2
(Re-R)Zi
For infinite time, or at least for times long enough,the seriesin the above equation
tendsto zero. The amountof substanceleaving the spherevaries thus linearly with time.
By plotting
Mt
4~cRi. R,(R,-
as a function of
Ri) Ci
Dt
a straightline is thus
(Re- Ri)2
Dt
obtainedwhich has an intercepton the time axis

(Re-
= 6 (Fig. 4.5)
Ri)2
The constantrate at which the diffusing substanceleavesthe hollow sphereis given

by :
(4.32)
= 4nRi.D.Ci
R,- Ri
Sec. 4.41
Hollow sphere in steady state
0,
a5
I
0.75
, IR,-R$*
71
Fig. 4.5. Amount of diffusing substanceleaving the hollow sphere,with a

constantconcentrationat the internal surface.Approach to steady-stateflow.
eqn.(4.31).
4.4 HOLLOW
SPHERE
IN STEADY
STATE
Generally this is the casewhere the hollow spherecontains the drug, and thus plays
the role of a spherical membranesurrounding the drug. As shown in 4.3, when this kind
of hollow sphere with drug in it is immersed in a surrounding atmosphere,the drug is
transportedthrough the hollow sphereby transient diffusion. After infinite time, or rather
after a time for which the series in eqn(4.31) tends to zero, the diffusion takes place in
steadystate.
In steady state, when the drug is located within the hollow sphere, generally the
concentrationof drug on the internal surfaceis constantat least for sometime, becausethe
amount of drug decreasesduring the process.Two casesare of interest :
(i) when the coefficient of matter transfer on the external surfaceis infinite, meaning that
Mathematical
72
treatment
of radial diffusion
ICh.4
in a sphere
the concentrationof drug is constanton the external surface.

(ii) when the coefficient of matter transferon the external surfaceis finite.
4.4.1
HOLLOW
ON EACH
SPHERE
WITH
CONSTANT
CONCENTRATION
SURFACE
The problem has already beenresolved in 4.4, when the time is long enough for the
seriesto tend to zero.
Another classicalsolution is given for the steadystate.
As in steadystate,the concentrationdoes not dependon time, the rate of changein
concentrationis zero, and eqn.(4.2) becomes:
+-(r2. $1
(4.33)
= 0
The generalsolution of this equationis of the form :

(4.34)
c = K,+?
where Kt and K2 areconstantswhich can be determinedfrom the boundary conditions.

The boundaryconditions are :
r = Ri
r= R,
t>o
(4.35)
C = Ci
c = c,
The solution is thus given by :
(4.36)
2, =
Cp Ri(R,-r)+C,
r (Re-
and it reducesto :
(4.37 ) c, =
Cp Ri (R,- r) *
r (Re-
Ri)
R,(r-Ri)
Ri)
internal surface
externalsurface
Sec. 4.41
Hollow sphere in steady state
73
when the concentrationof the substanceis zero on the external surface.

The amount of substanceMt which leavesthe sphericalmembranein time t is given
by relation (4.32).
Ri. R,
(4.32) M, = 4 x ---Cp
R,- Ri
Dt
4.4.2 HOLLOW SPHERE WITH A CONSTANT CONCENTRATION

THE INTERNAL
SURFACE
AND A FINITE COEFFICIENT
MATTER TRANSFER ON THE EXTERNAL SURFACE
The boundary conditions arewritten as follows :
t20 r = Ri
(4.38)
OF
OF
C = Ci = Ct
internal surface
DE = -h(Cs- C,,,)
dr
r = R,
where C, is the concentration on the surface and C,,t is the concentration in the external
medium and h is the coefficient of matter transfer.
The concentrationon the surfaceC, is given by :
(4.39) c, =
Rp Ci h.Rq+r(D-h-R,)
+h.Rt. C,,t(r-Ri)
r[h.R:+Ri(D-h.R.)]
The amount of substancepassing through the hollow sphere during the tune t is
given by:
(4.40) M, =
4 X. h. Rp R:(Cih. Rt+ Ri(D-h.
C,,,) D t
Re)
REFERENCES
1
H.S. Carslaw and J.C. Jeager,in Conduction of Heat in Solids, Clarendon Press,
Oxford, 1978, Chapter 9.
74
Mathematical
treatment
of radial diffusion
in a sphere
LCh.4
J.M. Vergnaudand J. Bouzon, in Mathematical and Numerical Treatmentof

Diffusion, Prentice Hall, 1993, Chapter7.
J. Crank, in The Mathematicsof Diffusion, ClarendonPress,Oxford, 1975,

Chapter 6.
Mathematical
cylinders
5.1
treatment
of diffusion
in
INTRODUCTION
The problem of diffusion through cylindrical solids is of interest either from a

theoretical or from a practical point of view. Of course,the problem of solid cylinders of
finite length is of practical interest, as some dosageforms have this shape.The hollow
cylinder of infinite length may also appearof concern, as theselong hollow cylinders are
used as tubings for the transport of liquids, e.g., blood in hemodialysis.
The following three casesare successivelyconsidered:
(i) Solid cylinder of infinite length. This is the more simple case,as the diffusion is only
radial, the effect of the edgesbeing negligible. An example is obtained with a thread, or a
very long cylinder.
(ii) Hollow cylinder of infinite length. This is the case of tubings used to transport a
liquid. In this casethe diffusion is radial only, in the sameway as in the first case.
(iii) Solid cylinder of finite length. The problem is more complex, asthe diffusion is radial
and longitudinal at the sametime. But this questionis of high interestfor dosageforms.
Two kinds of boundary conditions are considered: the one, when the coefficient of
matter transfer on the surface is so high that the concentration on the surface reachesthe
constantvalue at equilibrium as soon asthe processstarts.The other, when the coefficient
of matter transfer on the surface is finite, and the concentration on the surface varies
during the processfrom the initial value to the value at equilibrium.
The volume of the surrounding atmosphereis also of interest. In a first case,it may
76
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
be so large with regard to the volume of the dosageform that the concentration of the
diffusing substancein the surrounding remains constant during the process.This caseis
also obtained when the fluid in the surrounding is circulated. The other casewith a finite
volume of the surroundingis also considered.
The diffusivity is constantduring the whole process.
The changein dimensionsof the solid is negligible.
The concentrationof diffusing substanceis initially uniform throughout the solid.
With a long cylinder in which diffusion is radial, the concentration of diffusing
substanceis function of radius r and time. It is thus given by :
(5.1) $+ = p. &(r.g)
or
(5.1)
= D.
a*c+:
ac
[- 1
ar
r 7-F
With a cylinder of finite length, the diffusion of the substance is radial and
longitudinal. The equationof diffusion is thus given by :
orby:
(5.2)
= D.
theseequationsresulting from the superpositionof the radial and longitudinal transportsof

the diffusing substance.
The diffusing substanceis transportedin non-steadystatein most cases.A transport
in steadystate is of interest in the caseof a tubing, when the surroundingson the internal
and on the external surfacesarekept at a different constantconcentrations.
Sec. 5.21
5.2
Non-steady
NON-STEADY
INFINITE
state with a solid cylinder
STATE
WITH
of infinite
A SOLID
length
CYLINDER
77
OF
LENGTH
The cylinder being so long that the effect of the edgesis negligible, the diffusion of
the substanceis radial only.
The equationof radial diffusion is :
15.1) g
withO<rcR
= y. =j-(r.g)
or
(5.1)
a2c
= D. a+7
1 ac
ar
with the constantdiffusity D

andR being the radius of the cylinder.
5.2.1 CONSTANT CONCENTRATION

ON THE SURFACE
The coefficient of matter transfer on the surface is high enough so that the
concentrationof substanceon the surfaceis constantas soon asthe processstarts.
The initial concentrationof diffusing substanceis uniform throughoutthe solid.
The initial and boundaryconditions are :
(5.3)
(5.4)
t= 0
t>O
0 <r< R
r= R
C = Ci
c = c, = c,,
With the methodof separationof variables,the concentrationat position r and time t,

C,, is expressedby the product of the two functions :
(5.4) C,, = U,. exp I-aDt]
where U, is a function of r only.
This function C, is a solution of the equation of radial diffusion (5.1), provided that
U satisfiesthe Besselsequation of order zero :
78
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
The concentrationC, is thus expressedby :
(5.6) C,,- C, = 2
A,. Jo(anr). exp(- a:Dt)
n=l
where the ct,,sare roots of

(5.7) &,(a,$)
= 0
satisfying the boundarycondition

The valuesof 4, aredeterminedby multiplying both sidesof eqn (5.6) at time 0 :
(5.8) Gin- C, =
n-l
by r. Jo (a,, r) and integrating from 0 to R

By using the following resultsfound in books of mathematics:
r. Jo(a r). Jc(p r). dr = 0

R
(5.10)
or. Ji(cz r). dr = $.
J:(aR)
where a and p are different roots of eqn. (5.7) and Jr is the Bessel function of the first
order.
The value of the concentrationC, is thus given by the relationship :
JO(an
an.
r,
2
I -a,Dt
J,(anR)exp
Sec. 5.21
Non-steady
of infinite
79
length
The amount of diffusing substancewhich has enteredor left the cylinder in time t,
M,, is expressedas a fraction of the correspondingquantity after infinite time M, by the
equation:
(5.12)
Me,-
Mt
ca
The solution can also be found by using the error function.

The concentration,provided that r / R is not to small, is thus expressedby :
= E.
(5.13) zrt-Er
co- In
erfc($$5)+(Rm~R~
+[9R2-7r2-2Rr)Dt
ierfc($&)
. [ierfc[$L]J,....
32 Ri. rf
The amountof substancetransferredafter tune t is :

3
Some roots hs of
Jo (CUR) = 0
aregiven in Table 5.1
Table
5.1
Rootsof Jo(x) = 0
Xl
X2
x3
2.4048
5.5201
8.6537
=4
=s
=6
=7
=8
=9
x10
11.7915 14.9309 18.0711 21.2125 24.3540 27.4957 30.6375
The kinetics of the substancetransferredis drawn in Fig. 5.1, by plotting Mt / M,

as a function of the squereroot of time, or ratherof the dimensionlessnumber I(Dt / R .
Mathematical
treatment
of diffusion
in cylinders
al
0.6
[Ch. 5
0.8
Fig. 5.1 Amount of matter transferredthrough the surfaceof a solid cylinder as a

function of the squareroot of time, with a constantconcentrationon the surface.
The concentration distribution within the circular cross-sectionof the cylinder is
shown in Fig. 5.2, where the ratio of the concentrations(C, - C,) / (C, - C,) is plotted
as a function of the position along the radius r / R, for various values of the dimensionless
number
D.t / R2.
I - Short time of transfer
The equationwith the error function is very useful in this case,as it convergesfaster
than the serieswith the Besselfunctions.
For very small times, e.g., when Mt /M, < 0.2 - 0.3, the first term in this equation
becomespreponderant,and the simple equationis obtained:
REMARK
(5.15)2 00= g
-2
21 7T
for 2
< 0.2 - 0.3
Non-steady
Sec. 5.21
A
of infinite
length
Cr#t-Gin
C,Ci"
0.2
0.1
05
0.0
Fig. 5.2 Concentrationdistributions at various times (the valuesof D t / R* are noticed), with a uniform concentrationand a constantconcentrationon the surface.
REMARK 2- Long
time
Of trimfer
When the time is long enough, e.g., when Mt /M, > 0.7 - 0.8, the first term in the
series with the Bessel function becomespreponderant, and the simple equation is thus
obtained:
(5.16) ;
Mt = Aexp(m
(% 4
cx:. Dt)
5.2.2 FINITE COEFFICIENT OF MATTER TRANSFER ON THE

SURFACE
The cylinder of infinite length is immersed in a surrounding atmosphereof infinite
volume. The concentration of the diffusing substancein the surrounding atmosphereis
82
Mathematical
treatment
of diffusion
[Ch. 5
in cylinders
thus constant,and uniform if it was well stirred.

The caseof the long cylinder with an initial uniform concentration is considered,
while the constantuniform concentrationin the surrounding atmosphereis C,.
At equilibrium, the concentrationin the solid cylinder is C,, which is related to the
concentrationC, with the partition coefficient
(5.17) C, = K.C,,,
As the coefficient of matter transfer at the surfaceof the cylinder is finite, the rate at
which the substancepassesthrough the cylinder is constantlyequal to the rate at which the
substancediffuses within the cylinder next to the surface.
The initial andboundaryconditions are as follows :
(5.18)
t = 0
0 < r < R
r>R
(5.19) t > 0
C = Ci,
c = text
r= R
-D.g
= h(C,-C,)
where h is the coefficient of matter transfer,

C, is the actualconcentrationof the diffusing substanceon the surfaceof the cylinder
and C, is the concentration of the substance in the cylinder necessary to maintain
equilibrium with the surroundingatmospherekept at the constantconcentrationC,,.
The solution of the problem for the concentration of the diffusing substancein the
solid cylinder is expressedin terms of the serieswith the Besselfunctions (1 - 3).
(5.20)
where the &s arethe roots of :

(5.21)
fJ. J,(p) = M. J,(p)
with the dimensionlessnumber M :

(5.22) M = y
Sec. 5.21
Non-steady
of infinite
length
83
Table 5.2
Roots of S.J#) = M.J,-&) with M = 2
I
p1
p2
p3
P4
Ps
hi
p7
p8
PlO
0
0.01
0.1
0.2
0.5
1.0
2.0
5.0
10.0
00.0
00
0
0.1412
0.4417
0.6170
0.9408
1.2558
1.5994
1.9898
2.179s
2.3809
2.4048
3.8137
3.8343
3.8577
3.8835
3.9594
4.0795
4.2910
4.7131
5.0332
5.4652
5.5201
7.0156
7.0170
7.0298
7.0440
7.0864
7.1558
7.2884
7.6177
7.9569
8.5678
8.6537
10.173s
10.174s
10.1833
10.1931
10.222s
10.2710
10.3658
10.6223
10.9363
11.6747
11.791s
13.3237
13.3244
13.3312
13.3387
13.3611
13.3984
13.4719
13.6786
13.9580
14.7834
14.9309
16.4706
16.4712
16.4767
16.4828
16.5010
16.5312
16.5910
16.7630
17.0099
17.8931
18.0711
17.6122
19.6128
19.6183
19.6244
19.6426
19.6728
19.7326
19.9046
20.1515
21.0347
21.2127
20.7538
22.7544
22.7599
22.7661
22.7842
22.8143
22.8742
23.0462
23.2931
24.1763
24.3543
23.8954
25.8960
25.9015
25.9077
25.9258
25.9560
26.0158
26.1878
26.4347
27.3179
27.4959
27.0370
29.0376
29.0431
29.0493
29.0674
29.0976
29.1574
29.3294
295763
30.4595
30.6375
The fmt roots of this equation are given in Table 5.2 for various values of M.
The amount of diffusing substancewhich is transferredafter time t, Mt, as a fraction
of the corresponding amount after infinite time, m, is expressedin terms of time by the
equation:
00
(5.23) MG Mt = 2
II=1
4M2
f([Pi;+Ml
. exp
-!!!&
i I
R2
The solution of the problem with a finite rate of transfer at the surface is also
expressedin terms of the error functions, when the ratio r / R is not small :
i5.241
crt-
tin
cca-
ci,
= Zhgierfc(s)
+4h
A-%).
i2erfc(-$&)
+ ...
The amountof diffusing substancewhich is transferredafter time t is given by :
84
Mathematical
treatment
2M.D.t
of diffusion
8M2 Dt:
-~(R:/
R2
in cylinders
-M(~-M).
[Ch. 5
(~~+....
The profiles of concentration,expressedby the ratio (C, - C,) / (Ci, - C,), are
drawn for various values of the dimensionlessnumber D t / R*, with a given value of the
dimensionlessnumber M = h. R/D = 1. (Fig. 5.3).
0
0
0.2
0.L
a6
a8
Fig. 5.3 Concentrationdistributions at various times (the values of D t / R* arenoticed) with a given value of the dimensionlessnumber M = h. R / D. Transfer with
a finite coefficient of matter transferon the surface.
The amountof substancetransferredafter time t as a fraction of the corresponding

amountafter infinite time, M, / M, is expressedin terms of the dimensionlessnumber
Sec. 5.21
Non-steady
of infinite
length
85
A Mt
Mco
075-
0.5-
0.25-
0.6
0.8
Fig. 5.4 Amount of matter transferredthrough the surfaceof a long solid cylinder
as a function of the squareroot of time, for various values of the dimensionless
number M = h. R / D.
m/R,for
variousvaluesofthedimensionlessnumberM = h.R/D(Fig.5.4).
It is shown, of course, that the curve obtained with a high value of M, i.e., 100, is the
sameasthat obtainedwith a constantconcentrationon the surface.
REMARK I - Absorption or desorption
Dependingon the relative valuesof the concentrationsof substanceon the surfaceof
the solid, absorptionor desorptiontakesplace :
(5.26)
c, < ca
c, > cca
absorption
desorption
- Evaporation or condensation
Generally, a liquid evaporateswith a finite rate, and the problems of evaporationof a
REMARK 2
86
Mathematical
treatment
of diffusion
[Ch. 5
in cylinders
liquid and of condensationof a vapour aredescribedby theseequations.

REMARK 3 - &IIdl
tiIIXS
Of traIXfer
As the error function convergesfaster than the series with the Bessel function, the
equationswith the error functions are more suitablefor small times of transfer.
- Tangent at the origin
The coefficient of matter transfer can be determinedfrom the tangent at the origin of
the curve M, expressed as a function of time. At the beginning of the process, the
concentrationon the surfaceis still the sameasthe initial concentrationCti.
REMARK 4
(5.27)
for
= h(Ci-C~)
t + 0
5.2.3 SOLID CYLINDER IN A WELL-STIRRED

SURROUNDING OF
FINITE VOLUME
This problem is the sameasthat alreadyshown with the plane sheet.
The cross-sectionsof the cylinder and of the bath with the surroundingatmosphere
and the solid in it are considered.The cylinder occupiesthe region defined by r < R, and
A is the areaof the cross-sectionof the surroundingatmosphere.
The initial concentrationsof the diffusing substanceare uniform in the cylinder and
in the surrounding atmosphere,and are written C, extand C,,,, respectively.
On the surface of the cylinder, the rate of transfer of the substanceis the sameon
eachside of the cylinder-surroundinginterface.
(5.28)
t=O
r<R
r
(5.29) t > 0
Cin.s
R
A.%=
Cin.ext
D.
solid
surrounding
27rR
atR
per unit length
The concentration of substancewithin the cylinder at time t, C,, is thus expressed

by the following relation (5.1 - 5.3) :
Sec. 5.21
Non-steady
n=l
length
87
r
Jo qnE
I 1
4(a+ 1)
2
of infinite
4+4cc+ a .qn
Jo(qd * exp
where C,,, is the concentrationin the solid cylinder after infinite time. This concentration
C&, is at equilibrium with the concentrationC,,, in the surrounding atmosphere.
The qnsarethe positive non-zeroroots of :
(5.31) aq,. Jo(q,)+2J,(qnl
= 0
where 01is the ratio of the amount of substancein the surrounding atmosphereand the
cylinder, per unit length.
This ratio 01is expressedby :
(5.32) a = A
rrR2. K
where K is the partition factor
The amount of diffusing substancewhich hasenteredor left the cylinder after tune t,
M,, as a fraction of the corresponding quantity after infinite time, m, is expressedin
terms of tune by the series:
Rootsof
a
00
9.oooo
4.0000
2.3333
1.5000
1.0000
0.6667
0.4286
0.2500
0.1111
0
41
2.4048
2.4922
2.5888
2.6962
2.8159
2.9496
3.0989
3.2645
3.4455
3.6374
3.8317
Table 5.3
CL.qn.Jo(q,,)+2.J1(qn) = 0
92
5.5201
5.5599
5.6083
5.6682
5.7438
5.8411
5.9692
6.1407
6.3710
6.6694
7.0156
s3
8.6537
8.6793
8.7109
8.7508
8.8028
8.8727
8.9709
9.1156
9.3397
9.6907
10.1735
94
11.7915
11.8103
11.8337
11.8634
11.9026
11.9561
12.0334
12.1529
12.3543
12.7210
13.3237
Ss
14.9309
14.9458
14.9643
14.9879
15.0192
15.0623
15.1255
15.2255
15.4031
15.7646
16.4706
q6
18.0711
18.0833
18.0986
18.1183
18.1443
18.1803
18.2334
18.3188
18.4754
18.8215
19.6159
88
Mathematical
(5.33)
Mea- M,
M
treatment
of diffusion
in cylinders
[Ch. 5
4a(l+a)
2
01
4+4a+a
. q;
The ratio CLis very often expressedin terms of the final fractional uptake of the
diffusing substanceeither by the solid cylinder when the substanceentersthe solid or by
the surroundingatmospherewhen the substanceleavesthe solid.
In the caseof absorptionof the diffusing substanceby the solid cylinder, the amount
of substancetransferred in the cylinder after infinite time is M,. The initial uniform
concentrations in the cylinder and the surrounding are respectively Cin.sand Cin,ext,so
that the initial amount of substancein the surrounding atmosphereis A.C,,,. The ratio c1
is given by :
02
01
06
08
Fig. 5.5 Amount of matter transferredthrough the surfaceof a long solid cylinder
as a fonction of the squareroot of time Cl%? / R), for various valuesof the ratio 01.
(The values of a arenoticed).
Sec. 5.21
(5.34) &
Non-steady
of infinite
length
89
M-
A*Cin,ext
In the case of desoxption of the diffusing substancefrom the solid cylinder, the
concentrationof substancein the cylinder is initially uniform Cin,srwhile the surrounding
is initially free from substance.The ratio a is expressed in terms of the ratio of the
amounts of substancetransferred after infinite time M, and originally located in the
cylinder TI:R2. Cin,s*
0.6
a8
Fig. 5.6 Concentrationdistributions of diffusing substancewithin the long solid

cylinder for various valuesof m / R, and for CC= 1.
(5.35) A!l+a
~ R2. Cin,s
The amount of substancetransferred through the surface of a cylinder after time t
Mathematical
90
treatment
of diffusion
[Ch. 5
in cylinders
M,, as a fraction of the correspondingquantity after infiite time, M,, is expressedin

terms of the dimensionlessnumber m I R , for various values of the ratio 01(Fig. 5.5).
For large values of a, e.g., a > 30-50, the curves are about the same as those obtained
with a constantconcentration on the surface which correspondsto the caseof an infinite
volume of the surrounding atmosphereand a high coefficient of matter transfer on the
surface, as shown in (Fig. 5.1).
The concentrationdistributions of the substancewithin the cylinder aredrawn for
various values of the dimensionlessnumber m / R, for a = 1 (Fig. 5.6) and for a = 9
(Fig. 5.7). when the cylinder is initially free from solute, and the initial concentration of
diffusing substanceis uniform in the surroundingatmosphere.
5.3 NON-STEADY
FINITE LENGTH
53.1
CONSTANT
STATE
WITH
CONCENTRATION
A SOLID
ON THE
CYLINDER
OF
SURFACE
The solid cylinder of radius R and length 2 Z, initially at uniform concentrationCi,

is immersed in a strongly stirred surrounding atmosphere of infinite volume kept at
concentration C,,,. The coefficient of matter transfer on the surface is very high, so that
the concentrationof substanceon the surfaceis kept constantat C, as soon as the process
starts.There is of course a relationship betweenthe concentrationsC,, and C,, with the
partition factor K :
C, = K. C,,,
(5.36)

t=
(5.37)
and
t>O
(5.38)
or
0 < r < R
-z<z<z
C = C,
r= R
z = fZ
CC.0=surface
solid cylinder
of finite length
Of course, the diffusion of the substance is both radial and longitudinal. The
solution to the problem can be expressedby the product of the solutionsobtainedeither for
the radial transportonly, or for the longitudinal transportonly.
Sec. 5.31
Non-steady
0.2
06
0.1
of finite length
91
0.8
Fig. 5.7 Concentration distributions of diffusing substance within the long solid
cylinder for various values of the dimensionless number m / R , for a = 9
The concentration distribution C, at position r and z at time t is thus expressed by

(3) :
(5.39)
cnt - cca
tin-
coa
. -.4 c- (-1)
7T @ 2p+l
J2P+
22lhz
. exp ! - (2p+ l)nZD,

4z2
The amount of diffusing substance which has entered or left the solid cylinder of
finite length after time t, M,, is expressed in terms of the fraction of the corresponding
quantity after infinite tune K,
and in terms of time t by :
92
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
The values of the diffusing substancetransferred after time, M,, as a fraction of the
correspondingquantity transferredafter infhrite tune &, aredrawn as a function of the
dimensionlessnumber m / R, for various valuesof the ratio of half the length 2 and
the radius R.
The values of Z / R are noticed. (Fig. 5.8).
0.2
0.L
a6
0.8
Fig. 5.8 Cylinder of finite length with a uniform initial concentrationand a very
high coefficient of matter transferon the surface.Matter transferredMt / M, as a
function of the dimensionlessnumber dot/R, for various valuesof the ratio Z / R .
Sec. 5.31
5.3.2 FINITE
SURFACE
Non-steady
COEFFICIENT
OF MATTER
93
of finite length
TRANSFER
ON
EACH
The case of a solid cylinder of finite length with a finite coeffkient of matter transfer
on the surface is of high interest. The solid cylinder initially at the uniform concentration
Ci, is immersed in a surrounding atmosphere whose volume is so large that its
concentration remains contant at C,,,. With the partition factor between the solid and the
surrounding atmosphere, the concentration C, in the solid is thus at equilibrium with the
concentration C,, in the surrounding.
The initial and boundary conditions are :
t=
(5.41)
OSrSR
tin
cylinder
and r=R
surface
-ZlzlZ
(5.42~
t>o
for z = It Z
-D.g
= h(C&,)
where CR is the actual concentration on the surface and a C / dr is the gradient of

concentration in the cylinder next to the surface.
The solution to the problem with radial and longitudinal diffusion is also equal to the
product of the solutions obtained with the radial diffusion and with the longitudinal
diffusion.
The concentration distribution is thus given by the product of the Bessel series and
of the trigonometrical series (3) :
where the PI, s are the positive roots of
94
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
p. tan p = M,
(5.44)
with the dimensionlessnumber

(5.45)
M, = $+r
Z being half the length of the cylinder

and where the p, s are the positive roots of
15.21)
P. J,(P)
= M, Job)
with the dimensionlessnumber

(5.22)
Mr = R$
In the sameway, the amount of matter transferredafter time t, M,, can be expressed
by the product of the following two series,the first seriesrepresentingthe contribution of
the longitudinal diffusion and the secondseriethat of the radial diffusion.
The amount of substance transferred after time t, M,, as a fraction of the

correspondingquantity after infinite tune, M,, is expressedas a function of the
dimensionlessnumber V%-t/ R , for various valuesof the ratio Z / R, with the valuesD =
1O-6cm% and h = 1O-6cm/s and R = 1 cm. (Fig. 5.9). These curves show the effect of
the length of a cylinder on the matter transferred,when the coefficient of matter transfer is
Sec. 5.41
Non-steady
state with a hollow
0.2
cl.4
cylinder
of infinite
0.6
length
95
0.8
Fig. 5.9 Cylinder of finite length with a finite coefficient of matter transferon the
surface.Matter transferredMt / M, as a function of the dimensionlessnumber and
R=lcm.
finite.
The amountof substancetransferredafter tune t, M, / M, is expressedas a function
of the dimensionlessnumber 1(Dt / R , for various valuesof the coefficient of mattertransfer
h (thesevalues of h are noticed), and with the values : D = 10 cm2/sand R = Z = 1
cm. (Fig. 5.10). Thesecurvesillustrate the effect of the coefficient of matter transfer h on
the matter transferred,for a given cylinder of finite length.
5.4 NON-STEADY
STATE
INFINITE
LENGTH
WITH A HOLLOW
CYLINDER
OF
Two caseare considered:

(i) When the initial concentrationof diffusing substancein the hollow cylinder is uniform
96
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
C,,, and when the concentrationsof substanceare constantand equal on the two surfaces,
as soon as the processstarts.This means,of course,that the coefficient of matter transfer
on the surfaceis very high.
(ii) When the initial concentration of substance is uniform, Ci,, within the hollow
cylinder, andwhen the concentrationson eachsurfaceare constantand different.
The second case is of high interest, as it correspondswith the caseof a tubing in
which a liquid is circulating.
As the hollow cylinder is very long, the diffusion is radial only.
cl.?
0.6
Fig. 5.10 Cylinder of finite length with a finite coefficient of matter transferon the
the surface.Matter transferredMt 1M, asa function of the dimensionlessnumber
m / R for various valuesof the coefficient of mattertransferh (the valuesof h
are noticed), with the values : D = lO-j/ s, and R = 2 = 1 cm.
Sec. 5.41
Non-steady
state with a hollow
54.1 SURFACE CONCENTRATION

EACH SURFACE
cylinder of infinite
CONSTANT
AND
97
length
EQUAL
ON
The initial andboundaryconditions are as follows :

(5.47)
t = 0
Ri < r < R,
C = Ci,
cylinder
(5.48)
t>o
r=R
r= R,
c = c,
c = c,
internal surface
externalsurface
The concentration distribution of the diffusing substanceis expressedin terms of

time by the series(5.1 - 5.3) :
(5.49)
G- coa= Tc. m Jo(Ri. a,). Vo(r. a,)

c
n=l J&i. cc,)+J&e a,) exp
ci,- coa
where the function Vo represents:

(5.50) Vo(ra,.,) = Jo(f. %). Yo(& a,)- JIJ(% a,). Yo(r. a,,)
and the a, s arethe positive roots of
(5.51) V,(Ri. a,) = 0
The amount of diffusing substancewhich has entered or left the hollow cylinder
after time t is given by :
(5.52) t
Mt
0
Jr&,. a,)-- J&e. a,)

= ~
(-a:.
Dt)
~~. [Jo(Ri. cI,)+ Jo(R, ~.i] exp
Someroots of Vo (Ri. a,,) = 0 are given in Table 5.4 for various values of the ratio of the
radius R, 1%.
Jo and Y, are Bessel functions of the first and secondkind, respectively, of order
zero.
98
Mathematical
treatment
of diffusion
[Ch. 5
in cylinders
Table 5.4
RootsOf Jo(Ria,). Yo(Rea,)- J~(R,oz~)* Yo(Ria,)
1.2
1.5
2.0
2.5
3.0
3.5
4.0
5.4.2
CONSTANT
15.7014 31.4126 47.1217 62.8302 78.5385

6.2702 12.5598 18.8451 25.1294 31.4133
3.1230 6.2734 9.4182 12.5614 15.7040
2.0732 4.1773 6.2754 8.3717 10.4672
1.5485 3.1291 4.7038 6.2767 7.8487
1.2339 2.5002 3.7608 5.01%
6.2776
1.0244 2.0809 3.1322 4.1816 4.1816
CONCENTRATIONS
ON
EACH
SURFACE
This is the very interesting case of a tube, with a very high coefficient of matter
transferon eachsurface.The initial concentrationis uniform within the hollow cylinder.
(5.53)
t = 0
Ri < r < R,
C = Gin
hollow cylinder
(5.54)
t>O
r= Ri
r= Re
C = Ci
c = c,
internal surface
externalsurface
where Ci, C, and Ci, are constant.

The concentration distribution of the diffusing substanceis expressedin terms of
time by (5.1 - 5.3) :
Re
ci, LnT+Ce
Lng
(5.55)
c,, =
Re
+ IF. Ci,
hrr;
- 5t. c
II=1
C, J,(R,. LX,)- Ci. J&R,. a,)

J;(R;. a,)- J:(Rc a,)
. Jo(Ri. a,). Vo(ra,). exp(-a%. Dt)
Sec. 5.51
Steady state with a hollow cylinder
of infinite
length
99
where the CX,,

s are roots of Vo (Ri s)
= 0.
The amount of diffusing substancewhich has left the hollow cylinder after time t,
per unit length, when the concentrationsC, = C, = 0, is given by :
where T is the intercept with the time-axis of the straight line obtained when t becomes
very large.
RF-R;+(R;+R,Z).
(5.57) T =
4hq
In;
R,
5.5 STEADY STATE

INFINITE LENGTH
WITH
A HOLLOW
CYLINDER
OF
Three problems with the hollow cylinder of infinite length are consideredwhen the
substancediffuses in steadystate:
(i) When the concentrationsof the diffusing substantare constanton eachsurface.
(ii) When the concentrationof the substanceis constanton the internal surface,and there is
a finite coefficient of matter transferon the externalsurface.
(iii) When the hollow cylinder is madeof various cylindrical membranes.
5.5.1
CONSTANT
CONCENTRATIONS
ON
EACH
SURFACE
The problem is the same as that studied in steady state in section 5.4.2, and the
solution is obtainedfor times long enoughfor the seriesto vanish.
Another method of calculation can be used, by considering the steady state of
diffusion.
The transfer of substanceis radial within the hollow cylinder with the internal radii
Ri and Re.
100
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
The equation of diffusion in steady state is :

(5.58)
-&(I.%)
for
= 0
Ri < r < R,
as the concentration does not depend on time

The solution is of the form
C = A+B.Lnr
(5.59)
where A and B are constants to be determined from the boundary conditions.

These boundary conditions are :
(5.60)
internal surface
r=R;
Ci
r = R,
C, = ct
Ct
external surface
The concentration C, is thus :
Ci. Ln~+C,
(5.61)
C, =
Ln~
R,
Ln-ir;
and the rate of diffusing substance entering or leaving the hollow cylinder, per unit length
is :
(5.62)
= -2rrr.D.g
at any value of r.
The rate of diffusing substance per unit length is thus :
(5.63)
2TFD.(Ci-Ce)
Re
LnF
Sec. 5.51
Steady state with a hollow
length
101
5.52 CONSTANT CONCENTRATION ON THE INTERNAL SURFACE

AND A FINITE COEFFICIENT
EXTERNAL SURFACE
The diffusion is radial and in steady state. The boundary conditions are written as
follows :
(5.64)
t>O
Ci
r=R
-D.g
r= Re
internal surface
Ct
= h(C,-CW)
externalsurface
where h is the finite coefficient of matter transfer,

C, is the constantconcentrationof substanceon the external surface,
C, is the concentrationrequired to maintain equilibrium with the surroundingatmosphere
C ccl
The concentrationCr is :
Ci D+h.R,
Lrr
(5.65) C, =
Re+C,
h.R,. Ln+
i
Re
D+h.Re
LnF
The constantrate at which the substancepassesthrough the hollow cylinder is, per
unit length :
(5.66) %p = 2x(ci-cW).
and the substanceenters or leaves the hollow cylinder depending on the values of the
concentrationsCi and C,.
5.53
COMPOSITE HOLLOW CYLINDER

The composite hollow cylinder is made of the cylindrical membranesof radii RI,
R2 R,, at the respectiveconcentrationsC,, C2 ... C,.
The contact between these cylindrical surfacescan be either perfect, or imperfect
102
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
with an internal resistancebetweenthem.

When the contact betweenthe following membranesis perfect, the rate of diffusing
substancein steadystateis :
(5.67)
2dCrCz)
Rate =
.. . =
2rcDnIC- Cn+l]
R n+l
hR
Rl
Ln
R*
andthe fall in concentrationis given by :
(5.68)
cl-c+1
=z. 2 $.I+R n
1
When the contact between the following membranesis not perfect, the additional
contact resistancesRI, R, ...R. ... must be taken into account.The fall in concentration
becomesthus :
n+l
(5.69)
Cl-C,+z
= 2
2
I
&. LnF+c
$
1
5.6 CONCLUSIONS
Two casesare of high interest :
- the hollow cylinder of infinite length
- the solid cylinder of finite length.
SOLID CYLINDER OF FINITE LENGTH
In the caseof a solid cylinder of finite length, a simple solution can be useful, for
short times or rather when Mt /M, < 0.3 - 0.4.
The solution q,, of Jo (q) = 0 is approximately given by :
(5.70)
CL
= TC(n - 0.25)
Sec. 5.61
Conclusions
103
When M, / M, < 0.3 - 0.4, or when D.t / R2 is small, the series can be
approximatedby :
(5.71) 2 r
p -$Dt
n=l tex
( nR2) = t-;y/y
In the sameway, when D.t I4 Z2 is small, the seriescan be approximatedby :
(5.72)
exp[-2p:ir*Dt!
2
p=o (2p+ II2
= {[l-$c]
where 2 Z is the length of the cylinder,

and R is its radius
By replacing the series in eqn (5.40), by their approximated values, the amount of
diffusing substancetransferredis thus given by the simple relationship :
(5.73) 2
= 4
Jyk+A]
A more general solution, available not only for a cylinder of finite length, but also
for a sphereor a parahelepiped,is given by :
(5.74) 2
;.
= 2
O
21
where A is the areaof the solid

andV is its volume
provided that the amount Mt / M, < 0.2 - 0.3.
REFERENCES
1
H.S. Carslaw and 3.C. Jeager,in Conduction of Heat in Solids, 2nd ed.,
104
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
Clarendon Press,Oxford, 1959, Chapters7 and 8.

2
J. Crank, in The Mathematics of Diffusion, 2nd ed., Clarendon Press,Oxford,

1976, chapter 5,69.
J.M. Vergnaud and J. Bouzon, in Cure of ThermosettingResins, Springer

Verlag, London, 1992, chapter 5.
Numerical analysis with onedimensional diffusion through

sheet
6.1
a plane
INTRODUCTION
The principle of numerical analysisis to replacethe continuous transfer of diffusing

substancewhich is studied with the mathematical treatment by a stepwise transfer with
finite incrementsof time At and of spaceAx.
The time t as well as the abscissaalong the thickness of the plane sheet are thus
expressedin terms of theseincrements,as shown in Fig. 6.1.
(6.1)
t=j.At
x = n.Ax
time
space
wherej and n are integers.

In Fig. 6.1, at time t = 0, and j = 0, the initial conditions are obtained. They are
shown in the first line. They may be either the same in uniform initial concentration or
quite different in non-uniform initial distribution,
This two-dimensional space-timediagram in Fig. 6.1 is of help to understandthe
method of calculation. The thickness of the sheet is divided into N slices of constant
thickness Ax, and the concentration is uniform within each slice at any time. The
concentrationwithin the slice associatedwith the integer n at time t = j. At is written Cnj.
After elapseof time t, the new time being t + At = (i + 1) At, the new concentration
is written as follows :
106
Numerical
analysis
with one-dimensional
diffusion
n-l
I
h-1,0
n
I
h,O
n+l
I
h+l,O
j.At
cn-l,j
cr,j
Cn+l,j
IIIII
[Ch. 6
III
N
I Space
*
I
CNn
(j+l).At
I
v Time
Fig. 6.1. Space-timediagram for numerical analysis.Concentrationof diffusing

substancein various places and times.
(6.2)
t + At
cn,j+l
or
ml
This method of calculation with finite differencesis very powerful. It can be usedin
any cases,whatever the initial and boundary conditions. It is also easily applied to the
difficult casewith a concentration-dependentdiffusivity when the mathematicaltreatment
is not feasible.
Various casesare described,firstly with a constantdiffusivity in order to accustom
the readerto practising this techniqueof calculation, and secondlywith the concentrationdependentdiffusivity.
6.2 DIFFUSION
DIFFUSIVITY
THROUGH
A SHEET
WITH CONSTANT
Two casesare of interest,dependingon the value of the coefficient of matter transfer

on the surface,h :
- when h is so high that it may be consideredas infinite
- when h is not high, and is thus finite.
In fact, the value of the coefficient of surfaceh is expressedin terms of the thickness
Sec. 6.21
Diffusion
through
a sheet with constant
diffusivity
107
--L
n+O.S
Ax
Fig. 6.2. Diffusion of the substancewithin the plane sheet,by consideringthe

matterbalancewithin the slice of centern.
of the sheet L and of the diffusivity by the following dimensionlessnumber L.h / 2 D,

already shown in Chapter 2. (section 2.2.1 - Remark 6).
Two positions are considered:
- within the sheet
- on the parallel surfacesof the sheet.
6.2.1 INFINITE
SURFACE
COEFFICIENT
OF MATTER
TRANSFER
ON THE
with :
lSnlN-1
The cross-sectionof the sheetof cross-sectionalareaA through which the substance
diffuses, is considered in Fig. 6.2, with the planes of abscissaen - 0.5 and n + 0.5
parallel with the plane surfaces.
The matter balance is evaluatedduring the increment of time At within the slice of
thickness Ax located between the two planes of abscissaen - 0.5 and n + 0.5, by
consideringthe amountof matter which entersandleavesthis slice. The following relation
is thus obtained :
WITHIN THE SHEET,
108
(6.3)
Numerical
analysis
A.l-D(~)-O.S+D.i~)~+~.~.
[Ch. 6
diffusion
At = bd(CN,-C,,tj
where the value in the right-hand memberis the changein the amount of substancewithin
the slice during the time At.
where D. &I / dx representsthe flux of matter (expressedby the first Ficks law). (section
1.1.1 - Chapter 1)
C N, and C,,t are the new concentration after elapse of time At and the previous
concentrationat time t andposition n, respectively.
The gradient of concentrationat the plane n - 0.5 is approximatedby the chord slope
betweenthe abscissaen and n + 1 :
(6.4)
With the same assumption for the gradient of concentration at n + 0.5, the matter
balancebecomes:
(6.3)
4cn-l-cl
Ax
D(cn-cn+ll
Ax
At = Ax(CN,-C
)
n.t
after simplification by A.
Upon putting the dimensionlessnumber M
(6.5)
(A xl2
M = D. A t
the following equationis obtained:

(6.6)
CN,
= A[ C,-,+b--4C,+C,+,
The new concentration after elapse of time At is thus expressedin terms of the
previous concentrationsat the sameplace and at the adjacenttwo places.
Sec. 6.21
Diffusion
through
diffusivity
109
with
n= 0
or
n= N
As the coefficient of matter transfer through the surfacesof the sheetis very high,
the concentration on each surface reachesthe equilibrium value as soon as the process
starts:
ON THE SURFACESOF THE SHEET,
(6.7)
co = CN= c,,
The equation (6.6) is a basic equation,which can be used in all places and especially
for the positions next to the surfacesof the sheet.For instance,the new concentrationCN,
is written asfollows by using this equation :
(6.8)
CN, = k[C,+(M-2)C,+C,]
AMOUNT OF SUBSTANCE LOCATED IN THE SHEET
It is of interest to determinethe amount of substancewhich is located in the sheetat

any time, this information being in closerelation with the kinetics of matter transferred.
The total amount of diffusing substancelocated in the sheetat time t is obtained by
integratingthe concentrationsat this time t with respectto space:
N-l
(6.9)
M, = A. Ax. c C,,,
n=O
CONDITION OF STABILITY
The condition of stability for calculation is very simple in this case.It is obtainedby
writing that the coefficient of the concentration C, must be positive in eqns. (6.6) and
(6.8). This leads to :
(6.10)
M-2
> 0
In fact the best valuesfor M are asfollows :

(6.11)
4<M<8
110
Numerical
analysis with one-dimensional
diffusion
[Ch. 6
6.2.2 FINITE COEFFICIENT OF MATTER ON THE SURFACE

Two positions are of interest : within the sheetand on the surface.
with :
lInSN-1
WITHIN THE SHEET,
The problem is the sameas that shown with an infinite coeffkient of matter transfer
on the surface.The new concentrationCN, is thus also given by the relation :
(6.6)
CN, = A[C,-,+(M-2)C,+C,+t]
with the dimensionlessnumber M :
(6.5)
M = -(Ad
D. At
with
n = 0,orn = N.
The main condition on the surfacesis that the rate at which the substanceleaves (or
enters)the surfaceis always equalto the rate at which the diffusing substanceis brought to
(or from) the surface by internal diffusion. Moreover, the rate at which the substance
leaves (or enters) the surface is proportional to the difference between the actual
concentrationon the surfaceand the concentrationnecessaryto maintain equilibrium with
the surrounding atmosphere.This condition is
ON THE SURFACE OF THE SHEET,
the sameasthat written in the mathematicaltreatment(Chapter 1 - section 1.1S).

(6.12)
D.
= h. (C,- Cq)
surface
where h is the coefficient of matter transferon the surface,

C, and C,, are the actual concentrationon the surfaceand the concentrationon the surface
requiredto maintain equilibrium with the surroundingatmosphere,respectively.
is the gradientof concentrationnext to the surface.
f2
The slice of thickness Ax / 2 next to the surface 0 is considered, as shown in Fig.
6.3.
The matter balanceduring the incrementof time At is evaluatedwithin this half slice
by considering the matter diffusing through the plane 0.5 and the matter passing through
the surface 0 with the coefficient of matter transfer h, with the cross-sectional area A
Sec. 6.21
Diffusion
through
diffusivity
111
Fig. 6.3. Diffusion of the substancenext to the surface,with a finite coeffkient

of matter transfer on the surface.
through which the substancepasses:

(6.13)
A. -D
h(Cc- Cq). At
A.Ax
= 2
[ CNo.25
- co.251
where Co.25and CNo,,, are the concentrationsat the middle of this half slice, at time t and
after elapseof time At, respectively.
The gradient of concentrationat position 0.5 is given by the chord slopebetweenthe
positions 0 and 1 :
(6.4)
By making the simple assumption

(6.14)
CN0,25- Co.zs = CNo- Co
the new concentration on the surface after elapse of time At, CN,, is thus given by the
relation
(6.15)
CN, = & [2CI+(h4-2-2P)Co+2P.c~
112
Numerical
diffusion
[Ch. 6
where the dimensionlessnumber P is expressedby :

(6.16)
h. Ax
P = -D-
By making the more accurateassumption

(6.17)
CN,.,,-
Co.25 = z (CN,-C~)+~~CN,-C,)
the new concentrationon the surfaceCNc, is obtained :

(6.18)
CNo = -f
(CN,-Cl]+&
8
[ C 1+ (s-1
-p)co+pc~]
Of course,the new concentrationCN, must be calculatedbeforehandby using eqn. (6.6)

with n = 1
CONDITION OF STABILITY
The stability of calculation needspositive coefficients for the concentrationCu. This

leadsto :
(6.19)
(6.20)
M-2P > 2
3M-8P
> 8
in case6.2.2 with eqn (6.15)

in case6.2.2 with eqn (6.18)
REMARK 1 - Absorption or desorption of the substance
Depending on the relative values for the concentrations on the surface and at
equilibrium with the surroundingatmosphere,the substanceis absorbedor desorbed:
(6.21)
c, G,
desorption with C, = Co = CN
cS<ce,
absorption
REMARK 2 - Finite or infinite value of h
Depending on the relative value of the coefficient of matter transferh with respectto
the other two values of Ax and D expressed in eqn (6.16), the value of h can be
consideredasfinite or infinite.
Sec. 6.31
Diffusion
through
a sheet
113
When
(6.22)
h. Ax
D
is very high.
c, = c,
as obtained in eqn (6.12), leading to a constant concentration on the surface.

When
(6.23)
h. Ax
D
is not very high.
co;f c,
resulting from a finite coefficient of matter transfer on the surface
6.3
DIFFUSION
THROUGH
CONCENTRATION-DEPENDENT
A
SHEET
DIFFUSIVITY
WITH
In the same way as for a constant diffusivity, the following two cases are studied :
- when h and the ratio h. Ax / D is so high that it may be considered as infinite, and the
concentration on the surfaces is constant.
- when h and the ratio h. Ax / D is not very high, and it may be considered as finite.
Two positions must also be examined :
- within the sheet
- on the parallel surfaces of the sheet.
6.3.1 INFINITE
SURFACE
COEFFICIENT
WITHIN THE SHEET,
with :
OF MATTER
TRANSFER
ON THE
lln5N-1
The cross-section of the sheet with a cross-sectional area A through which the
substance diffuses is also considered (Fig. 6.2).
The matter balance evaluated during the incremental time At within the slice of
thickness Ax between the planes of abscissae n - 0,5 and n + 0,5 is written in the same
way as for a constant diffusivity :
(6.3)
A.[-iD.~]~-o.S+(D.~]+0.5j.
At = (A.Ax)(CN,-C,,,)
114
Numerical
diffusion
[Ch. 6
where the right-hand memberrepresentsthe changein the amountof substancein the slice
during the time At
and D. LX / ax is the flux of matter at the various positions n - 0.5 and n + OS, the
diffusivity D dependingon the concentrationof substanceat eachposition.
Upon putting the function 3
(6.24)
Jn+o.s = ID.gn+os
I.
= &+o.s
-G
Ax
the new concentration CN, in eqn (6.3) can be expressed in terms of J and other
parametersby :
(6.23
CN, = C,+$[-
Jn-o.s+ Jn+o.s]
with
n=O,orn=N
Resulting from the very high value of the coeffkient of matter transferon the surface
h, the concentration of the substanceon the surface is constantly equal to the value at
equilibrium.
ON THE SURFACESOFTHE SHEET,
(6.7)
co = CN = c,,
The equation (6.25) can thus be used for the whole sheet, and especially for the
abscissaen = 1 and n = N - 1
CONDITIONS OF STABILITY FOR CALCULATION
The coefficient of the concentrationC, must be positive. This condition is fulfilled

when :
(6.10)
M > 2
6.3.2 FINITE
SURFACE
WITHINTHE SHEET.
COEFFICIENT
with :
OF
lln<N-1
MATTER
TRANSFER
ON
THE
Diffusion
Sec. 6.31
through
a sheet
115
The problem is the sameas that studiedwith an infinite coefficient of matter transfer
on the surface,and eqn (6.25) can also be used :
(6.25)
CN, = C+g[Ax
Jn-o.s+ Jn+o,j
n = O,orn = N
with
The slice of thickness Ax / 2 located next to the surface, as shown in Fig. 6.3, is
considered.The matter balance during the increment of time At is evaluated within this
slice by taking into account the diffusion of substancethrough the plane 0.5 and the
transferof substancethrough the surface0 with a finite coefficient of matter transferh.
ON THE SURFACE OF THE SHEET,
A -(D$&-h(Co-Ccq)
At = q[CNo.ts-Co.25j
I
(6.13)
(6.14)
(30.25
- Co.25 =
CNo - Co
the new concentrationafter elapseof time At on the surface,CNo, is thus expressedby the
relation
(6.26)
CNo = Co- F[Jo.s+h(Co-
Caj]
where the function Jo.s,defined by eqn (6.24), is :

(6.24)
Jo.5 = [I). g),
s = Do.,. y
By making the more accurateassumption:

(6.17)
CNo.25~
Co.25
= 43 (CN,-
Co)++(CN,-
C,)
the new concentrationon the surfacen = 0, CN,, is thus given by :
116
Numerical
analysis
(6.27)
CNo = Co-f(CN,-C,)-T
diffusion
[Ch. 6
5. [Jo.~+hko- Cq)]
As the new concentration CNu is expressedin terms of the new concentration CN,, the
new concentrationCN, must be calculatedbeforehandby using eqn (6.26) with n = 1.
AMOUNT OF DIFFUSING SUBSTANCE LOCATED IN THE SHEET
The amount of diffusing substancelocated in the sheetat time t can also be obtained
by integratingthe concentrationsat time t with respectto space:
N-l
(6.9)
M, = A. Ax. c C,,,
n=o
CONDITIONS OF STABILITY
The conditions of stability for calculations are fulfilled when the coefficients of the
concentration C, in eqn (6.25) and of the concentration Co in eqn (6.26) or (6.27) are
positive. Theseconditions lead to :
(6.28)
l-2A-(Do.5+h.Ax)>0
(4
(6.29)
1 - ;. AtDo.,+
iAd2
(6.10)
M>2
h. Ax)> 0
REMARK 1 - Expression of the diffusivity in terms of concentration
There are many ways to express the diffusivity in terms of the concentration of
diffusing substance.
Someof them arepresented:
- When the diffusivity varies little (because the diffusivity varies little with the
concentration, or becauserather flat gradients of concentration are observedwithin the
sheet),the meanvalue of diffusivity can be used
(6.30)
Dn+o.5 = ;[ D, + D,+I]
Sec. 6.31
Diffusion
through
a sheet
117
- When the diffusivity highly varies (becausethe diffusivity strongly varies with the
concentration,or becausesteepgradients of concentration appearthrough the sheet),the
diffusivity can be expressedin terms of the concentrationof substancein an exponential
way :
(6.31) D,,+o.s= D. ew be - G+o.~)
In this case, it is necessaryto calculate beforehand the concentration of substanceat
position n + 0.5 by using the simple relation :
(6.32)
Cn+o.s= ; I C,+ C~+I
and then it is possibleto obtain the diffusivity at the position n + 0.5.

REMARK 2 - Drying or condensation
The problems of evaporation and condensation are identical to those studied in

sections6.2 and 6.3.
When the rate of evaporationor of condensationis very high, when the ratio h.L / D
is very high, the concentration is constant on the surface. The problem is resolved in
section 6.2.1 when the diffusivity is constant and in section 6.3.1 when the diffusivity is
concentration-dependent.
When the ratesof evaporationand condensationarenot very large, the concentration
on the surface varies during the process.The problem is resolved in section 6.2.2 when
the diffusivity is constant and in section 6.3.2 when the diffusivity is concentrationdependent.
Of course,dependingon the relative values of the concentration on the surface and
the concentrationwhich is at equilibrium with the surroundingatmosphere,evaporationor
condensationtakesplace.
(6.21)
c, ce,
c, < G,
evaporation
condensation
REMARK 3 - Case of a membrane
The interesting case of a sheetplaying the role of a membranewith two different

concentrations on each surface is also considered in sections 6.2 and 6.3. The various
equationsare written in section6.4.
Numerical
118
analysis
diffusion
ICh. 6
Fig. 6.4. Membrane: Sheetseparatingtwo media with a different concentration

of diffusing substance.
6.4 MEMBRANE
SEPARATING
TWO DIFFERENT
MEDIA
The membraneis a sheetwith parallel surfaces,eachsurfacebeing in contact with a

medium having a different concentrationof the diffusing substance.(Fig. 6.4).
Two cases are considered, depending on the value of the coefficient of matter
transfer on the surfacewhich can be either i&mite or finite. In eachcase,the diffusivity is
either constantor concentration-dependent.
6.4.1 INFINITE COEFFICIENT OF MATTER TRANSFER ON EACH

SURFACE
As the coefficient of matter transfer on the surfaceis very high, the concentrationon
eachsurfaceis equal to the concentrationat equilibrium as soon asthe processstarts.
Of course,the concentrationsat the inlet andthe outlet are different, with the obvious
relation :
(6.33)
Ci >
Co
The concentrationwithin the sheetwas shown in section6.2.1,

The concentrationwithin the sheetis given by :
Sec. 6.41
(6.6)
Membrane
separating
two different
media
119
CN, = &[C,-,+(M-2k,+C,+,I
with the dimensionlessnumber M

(Ad2
M = D. At
(6.5)
The constantconcentrationsat the inlet andoutlet surfacesare :

(6.34)
= constant
co = constant
inlet surface
outlet surface
ci
DIFFUSIVITY
The results are given in section6.3.1.

The concentrationwithin the sheetis expressedby :
CN, = C, + e[-
(6.23
Jn-o.s+Jn+o.s]
with the function J given by :
The concentrationsat the inlet andoutlet surfacesare :

(6.34)
= constant
co = constant
ci
inlet surface
outlet surface

OF MATTER TRANSFER
ON EACH
SURFACE
The concentrationsin eachmedium are different, and thus the concentrationson the
surfacesof the sheetrequired to maintain equilibrium with eachmedium are different, e.g.
Ceq.i and Ceq.09 at the inlet and the oulet of the membrane,respectively. Of course,there
Numerical
120
diffusion
[Ch. 6
is the obvious relation :

(6.35)
Gq, i
Gq.0
The coefficient of matter transfer may be different on each surface,e.g., h and ho at

the inlet and the outlet of the membrane.
The problem is studied is section 6.2.2. The concentrationsare thus given without
calculation.
Within the sheet,the concentrationis given by :
(6.6)
CN, = A[ C,-, + (M - 4 C-, + Cn+l
(6.5)
(Ad
M = D. At
On the inlet surface, the concentration is expressedby a relation similar to eqn

(6.15) :
(6.36)
CNO = b 2C1+(M-2-2Pi)Co+2PiCeq,i
with the dimensionlessnumber Pi
(6.16)
hi. AX
Pi = D
On the outlet surface,the concentrationis given by a relation similar to that shown

for the inlet surface,where \ is replacedby ho :
(6.37)
CNo = A
M-2-2Po)Co+2P,,.
Ceq,O
I
Sec. 6.41
Membrane
separating
two different
media
121
with the dimensionlessnumber Pu :
(6.16)
ho Ax
P, = D
DIFFUSIVITY
This caseis studiedin section6.3.2, and the results are given without calculation.
Within the sheet,the concentrationis given by :
(6.25)
CN, = C, + 2 [- J,,-as+ Jn+m]
with the function J :
(6.24)
J, = D. $&
i
I
On the inlet surface,the concentrationis expressedby the equation:
(6.38)
Ci+l)- i- ** [Jo.s+hi(Ci- Cqi)]
CNi = Ci-k(CNi+,-
Ax
and on the outlet surface,a similar equationis given :
Cinl
. . LL
Cl
Cin2
X
*
Fig. 6.5. Diffusion betweentwo different media 1 and 2.
122
Numerical
(6.39)
CN, = Co- f
diffusion
[Ch. 6
(CNo-I- CO-~)-3. E- [Jo.,+h,(Co- c,,,,)]
where Ci and Co are the concentrations of substanceon the internal and the external
surface,respectively,
ceq, i md ceq, 0 are the concentrations on the internal and external surface required to
maintain equilibrium with the media i and 0, respectively,
and Ci+, and Co-1 arethe concentrationsat the position hx apart from the internal and the
external surfaces,respectively.
6.5 DIFFUSION
BETWEEN
TWO DIFFERENT
SHEETS
Two sheetsof different materials 1 and 2 are in perfect contact at their junction, and
the problem of diffusion betweenthesetwo sheetsis consideredin this section. (Fig. 6.5).
The diffusivities are D, and D, and the concentrationsare written Ct and CZ.
Each medium is characterizedby two parameters: the diffisivity and the amount of
substancewhich can be absorbedat equilibrium, C,,.
6.5.1
CONSTANT
DIFFUSIVITIES
In the medium 1 with x > 0, the initial concentration is C,$ r, and in the medium 2
with x < 0, the initial concentrationis zero.
(6.40)
(6.41)
t = 0
t > 0
x>o
tin,
xc0
tin.
x=
medium 1
medium2
= 0
-=
c2
interface
Cl
(6.42)
x=
ac1
1. ax
ac2
= J,. x
where k is the partition factor for the substancebetweenthe two media 2 and 1. It is also
the ratio of the concentrationsattainedat equilibrium CesS
2 and Ces.l.
Diffusion
Sec. 6.51
between
two different
sheets
123
where the rate of the matter transferis the sameon eachside of the interfacex = 0.
Three positions are consideredsuccessively:
on the interface,within eachmedium, on the external surfaceof eachmedium.
0
For very short times, the solution to the problem of two sheetsis the same as that
obtained for two semi-infinite media separatedby an interface.The solutions for positions
next to the interfaceof two semi-infinite media are of the type :
CONCENTRATIONS NEXT TO INTERFACE
(6.43)
x > 0
Cl = A,+B,.
(6.44)
x < 0
c* =
x=
erf
where A,, A*, B, and B, are constantsto be determined from the initial and boundary
conditions.
The concentrationsCl and C2 are thus :
(6.45)
Cl =
(6.46)
C2
At the interface x = 0, the concentration on each side Cint, 1 and C&t, 2 remains
constant:
(6.47)
tin.
Cint,
1 =
124
Numerical
[Ch. 6
diffusion
As shown from eqn (6.42), the flux of substanceis the same on each face of the
interface x = 0. By approximating the gradient of concentration on each face in a
parabolic way, it is obtained :
(6.49)
D,
3 Cint.l-
4 Cint+l,l+
Cint+2,1
= D,
3c int.t-
4 Cint-1.2+
Cint-2.2
+2Ax,
-2Axr
is the concentration in the medium 1 at a distance Axi from the interface,

whereCint+l,~
is the concentrationat 2 Axr from the interface.
and where Cint-1.2is the concentration in the medium 2 at the distance Ax2 from the
and Cint+2,1
interface.
CONCENTRATION IN EACH MEDIUM
In each medium 1 and 2, the equationsfor the concentration are of the sameform.
They are the sameas that shown in the sheetwith constantdiffusivity (section 6.2.1).
(6.50)
CN, = +,+(M-2)C,+C,+,I
where the dimensionlessnumber is M, andM2 in the media 1 and 2 :
M2
(Ax2)
D,. At
Of course,the conditions of stability for calculation arethe sameasfor a sheetwith a
constantdiffusivity :
(6.52)
Mr>2
and
M2 > 2
CONCENTRATION ON THE EXTERNAL SURFACE
The external surface of each medium being in contact with the surrounding
atmosphere,three casescan be considered:
- when there is no transfer
- when the coefficient of matter transfer is so high that the concentrationon the surfaceis
constant
Sec. 6.51
Diffusion
between
two different
sheets
125
- when the coefficient of matter transferis finite.

NO TRANSFER ON THE EXTERNAL SURFACE
The rate of transfer through the external surfaceis zero. As shown in section 1.1.5,
the gradientof concentrationnext to this surfaceis zero.
(6.53)
(gjs
externalsurface
= 0
The external surfaceis thus a plane of symmetry,and the following relation is useful
for calculation:
(6.54)
CN-1
CN+l
where the positions N-l andN+l are symmetricalwith regardto the external surface.
CONSTANT CONCENTRATION ON THE EXTERNAL SURFACE
The problem is consideredin section6.2.1.

FINITE COEFFICIENT OF MAlTER TRANSFER ON THE EXTERNAL SURFACE
The problem is consideredin section6.2.2.

6.5.2
DIFFUSIVITIES
The problem is aboutthe sameasthat with constantdiffusivities, with a difference in
the media 1 and 2.
On the surface separatingthe two media, the concentrationscould be expressedby
eqn (6.49).
In each medium, the new concentration is expressed in terms of the previous
concentrationsat the sameand adjacenttwo placesby the following relation shown with a
concentration-dependentdiffusivity (section 6.3.1 or 6.3.2).
(6.55)
CN, = c,+q-
6x
Jn-O.5+Jn+0.5]
On the external surfaces,the solutions to the problems are given in section 6.3.1
126
Numerical
diffusion
[Ch. 6
when there is an infinite coefficient of matter transfer on the surface,and in section 6.3.2
when the coefficient of matter transferon the surfaceis finite.
6.6 TRANSFER
WITH SPECIAL
CONDITIONS
The numerical model is able to take into account all the known facts. Two casesare
of special interest : the one when the temperature is programmed, the other when the
concentrationof substancein the surroundingis programmed.
6.6.1 PROGRAMMATION
OF TEMPERATURE
This case is of special interest in the case of drying a dosage form. As a humid
dosageform is more plastic at high temperaturethan a dry dosageform, it is necessaryto
start the processof drying at a rather low temperature.In order not to lengthen the time of
drying too much, an increase in temperature during the process is thus capable of
increasingthe rate of drying andthus of reducing the time of the process.
The problem is rather complex, becausetemperatureactsupon various parameters:
the diffusivity, the coefficient of matter transfer at the surfaceh, the concentrationon the
surfacewhich is at equilibrium with the surrounding.
The diffusivity generally is increasedby increasingtemperature.Exponential law in
the following form can be tried :
(6.56)
Dr = Do. exp -k
I I
where Do and b are constants,

and the temperatureT is expressedin Kelvin.
The coefficient of matter transfer on the surfaceh also varies with temperaturein an
exponentialway :
(6.57)
where ho and b are constant

and the temperatureT is expressedin Kelvin.
Sec. 6.61
Transfer
with special conditions
127
The concentration C,, of the substance on the surface required to maintain

equilibrium with the surroundingdecreaseswith temperature.
6.6.2 PROGRAMMATION
OF THE CONCENTRATION
IN THE
SURROUNDING
It is sometimes of interest to programme the pressure of the vapour in the
surroundingatmosphereinsteadof eliminating it. This is especiallytrue at the beginning of
the process when the diffusivity is concentration-dependent.In this case the diffusing
substancemay
be entrappedin the solid during the evaporationprocess.The explanation of the processis
as follows :
the concentration of the substance on the surface decreasesstrongly because of the
evaporationfrom the surface ; as the diffusivity decreaseswith the concentration,the rate
of diffusion next to the surface becomesslower and slower.
The solution consists of regulating the value of the pressure of vapour P,,t in the
surrounding atmosphere.As the concentrationC,, required to maintain equilibrium with
the surrounding is related to this vapour pressure,it is thus possible to control the rate of
evaporationexpressedby :
h (C, - C,,)
aswell asthe concentrationof evaporatingsubstanceon the surface.
Numerical analysis with a rectangular

parallelepiped,
and a three-dimensional
transfer
7.1
INTRODUCTION
Depending on the nature of the material rectangular parallelepipedic in shape, which
can be isotropric or anisotropic, the problem is rather simple or more complex. In the case
of woods, it is well known that they behave as anisotropic solids with three principal axes
of diffusion and three principal difusivities. This problem is especially studied either for
the process of absorption of a liquid (1) or for the process of drying (2). Generally, bulk
polymers are isotropic materials, and only this case will be studied in this book.
When the medium is isotropic, the direction of flow of diffusing substance is normal
to the surface of constant concentration.
As it has been established in Chapter 3, the rate at which the concentration of
diffusing substance varies is given by the following equation when the diffusivity
(7.1)
= &[D.g)+=$[~.g]+=j-[D.%)
When the diffusivity is constant, the above equation reduces to :
(7.2)
= D.
is
Numerical
130
analysis with a rectangular
parallelepiped
[Ch. 7
where x, y and z are three rectangularaxes.

Various initial conditions can be used, depending on the problem. This is an
advantageof the numericalmodel over the mathematicaltreatment.
Various boundary conditions can also be consideredwith the numerical model. Two
of them are widely encountered, depending on the value of the coefficient of matter
transfer on the surfacewith regard to the value of the diffusivity and the dimension of the
sample.These values are introduced in the dimensionlessnumbers h. a / D, h. b / D and
h. c / D, where a, b and c are half the length of the three sides of the rectangular
parallelepiped. When these values are higher than 20 to 50, depending on the desired
accuracy,the coefficient of matter transfer on the surfaceis so high that the concentration
on the surface reaches the constant value at equilibrium as soon as the process starts.
When these values are lower than 20, the coefficient of matter transfer on the surface is
finite, and the boundarycondition is expressedby the relation :
(7.3)
-D.g
= h(C,-C,)
Fig. 7.1. Rectangularparallelepipedof sidesX, Y andZ built along the rectangular

axesx, y and z.
Sec. 7.21
With a constant
concentration
on the surface
131
In this relation, the rate of matter transfer through the surface is proportional to the
difference between the actual concentration on the surface C, and the concentration C,,
required to maintain equilibrium with the surrounding atmosphere ; it is also equal to the
rate at which the substance is transported to (or from) the surface, by internal diffusion.
h is the coefficient of matter transfer on the surface and n is one the three axes x, y or z.
Two cases are thus considered :
(i) with a very high coefficient
of matter transfer on the surface, and a constant
concentration on the surface.

(ii) with a finite value of the coefficient of matter transfer on the surface.
In each case, the problem is resolved with either a constant diffusivity
or a
concentration-dependent diffusivity.
The rectangular parallelepiped is built along the three rectangular axes x, y and z, and
its sides are respectively X, Y and Z. (Fig. 7.1)
7.2 TRANSFER THROUGH A RECTANGULAR

PARALLELEPIPED WITH A CONSTANT CONCENTRATION
ON THE SURFACE
The rectangular parallelepiped of sides X, Y and Z is considered (Fig. 7.1). Each

side is divided into 2 N increments of thicknesses Ax, Ay and AZ :
(7.4)
x=
2N.Ax
Y = 2N.Ay
Z = 2N.Az
and each point within the solid is defined by the three integers i, j, k, at the coordinates x,
Y, z.
(7.5)
x = i. Ax
Y = j.Ay
z = k. AZ
with
OlilN
OljSN
OlklN
132
Numerical
7.2.1
CONSTANT
parallelepiped
[Ch. 7
DIFFUSIVITY
The concentration is calculated in various positions, and especially within the solid,
and on the surface of the solid.
with
WITHIN THE SOLID,
11i12N-1,
11j12N-1
and l<k<2N-1.
The small rectangular parallelepiped of dimensions Ax, Ay and AZ, whose centre is
defined by the integers i, j, k, is considered (Fig. 7.2).
The matter balance during the increment of tune At is evaluated within this small
rectangular parallelepiped by considering the substance which enters and leaves the
parallelepiped by diffusion, along the three perpendicular axes x, y and z.
I -dx
I
I
I
I
I
I
i +l
; + 0.5
; - 0.5
i-1
Fig. 7.2. Matter balance during the increment of time Ax within the solid. Small
rectangular parallelepiped of sides Ax, Ay and AZ with the center i, j, k. Matter
balance during the increment of time At.
Sec. 7.21
With a constant
= Ax. Ay. AZ. CNi,j,k-
concentration
on the surface
133
i,j.k]
where Ci,j. k is the concentrationat the position i, j, k at tune t,

CNi, j, k is the new concentrationat the sameplace after elapseof tune At, at tune t + At.
By assumingthat the gradient of concentrationat the position (i - 0.5, j, k) is equal
to the chord slope between the positions (i - 1, j, k) and (i, j, k), as well as the other
positions, the new concentrationCNi, j, k can thus be obtainedfrom the matter balance:
(7.7)
CNi,j,k
Ci,j.k+~[ci-l.j.k-2Ci.j.k+Ci+l,j,k]
x
2Ci.j.k+Ci,j+l.k
Ci,j,k-l-2Ci.j,k+Ci,j,k+l
with the dimensionlessnumbers:
(7.8)
(Ad2
M, = D. At
MY - (Ay)2
D. At
M, = ()
D. At
ON THE SURFACE
The concentrationis constanton eachsurface,being equalto the value at equilibrium

with the surrounding atmosphere,as soon as the processstarts.
(7.9)
C0,j.k = Ci.0. k = Ci,j. 0 = C,q
AMOUNT OF SUBSTANCE LOCATED IN THE SOLID
The amount of diffusing substancelocated in the solid at time t is obtained by

integratingwith respectto spacealong eachof the three rectangularaxesthe concentrations
134
Numerical
analysis
with a rectangular
paralielepiped
tCh. 7
at time t.
The calculationis made along eachaxis successively:
Alongthez-axis(k),with
OSi12N
and OIj52N
N-l
(7.10)
MKi,j
= ~C;.j,O+$Ci,j,l+2
C ci.j.k+
ci.j,b
k-2
Along the y-axis (i), for
0I i I N
N-l
(7.10)
pi
= ~mi,o+~MK,~l+2C
MKi,j+MKi,,
j-2
Along the x-axis (i) :

N-l
(7.10)
M, = 8 a~,+$h2T,+2C
~i+~N
i=2
1Isx.Ay.A~.
STABILITY OF CALCULATION
For the stability of calculation, the coefficient of the concentrationCi,j. k in eqn (7.7)
must be positive. The following condition is thus obtained :
(7.11)
K+q+@
< 1
SYMMETRIES
By using the three planes of symmetry in the rectangular parallelepiped, only one
eighth of this parallelepipedcan be consideredfor calculation
(7.12)
ci. j. N+l = Ci,j.N-l

ci, N+l. k = ci, N-l, k
CN+I, j, k = CN-1. j. k
With a constant
Sec. 7.21
7.2.2
concentration
on the surface
135
DIFFUSIVITY
The concentrationis calculatedwithin the solid and on the surface.
WITHIN THE SOLID
The small rectangular parallelepiped of sides Ax, Ay and AZ is considered (Fig.

7.2). The matter balance within the incremental time At is evaluated within this small
parallelepiped
(7.13)
[-(D*E)i-0.5,j,k+~DEli+0.5,j.*l
Ay*AzAt
i, j -0.5.k
= Ax. AY. AZICNi,j,k- Ci,j,k]

This relation is about the sameasthat obtainedin section7.2.1 with a constantdiffusivity.
However, a difference exists, as the diffusivity varies with position and time in the same
way asthe concentration.
Upon putting the functions JX, JY, and JZ, in the same way as for the sheet
(Chapter6) :
(7.14)
Jxi-o.5,j.k =
JYi,j-0.5.k =
the new concentration after elapse of time At, CNi, j,k is thus expressedin terms of the
previous concentrationand of the functions JX, JY andJZ by the relation :
136
Numerical
(7.15)
CNi,j.k
= Ci.j.kbt[-
Ax
Jxi-0.5,j,k+
parallelepiped
[Ch. 7
Jxi+0.5,j.k]
-JZi,j,k-0.5+Jzi.j,k+0.5
ON THE SURFACE
The concentration is constant on the surface and equal to the concentration at

equilibrium with the surrounding atmosphereas soon as the processstarts.
DIFFUSIVITY
The diffusivity is expressedin terms of the concentrationin various ways.

When the diffusivity varies little with the concentration or when the profile of
concentrationis flat, the mean diffusivity may be used:
(7.16)
Di-O.s,j,k
= ibi,j,k+Di-I.j,k]
When the diffusivity varies strongly with the concentration or when the profile of
concentrationis steep,the meanconcentrationmust be evaluated:
(7.17)
ci-0.5,j.k
[ci-l,j,k+ci.j.k]
and the diffusivity at this position and time is thus expressed in terms of the actual
concentration at this position and time. For instance, the following equation could be
used :
(7.18)
Di-0.5,j.k
= A.exp(B.Ci-O.5,j,k)
The sameway of calculation asfor the constantdiffusivity is followed.
Sec. 7.31
With a finite coefficient
TRANSFER
7.3
PARALLELEPIPED
MATTER TRANSFER
of matter transfer
on the surface
137
THROUGH
A
RECTANGULAR
WITH
A FINITE
COEFFICIENT
OF
ON THE SURFACE
Two cases are considered, when the diffusivity is constant and when it is
7.3.1
CONSTANT
DIFFUSIVITY
The matter balanceduring the increment of time At is evaluatedwithin the solid and
in various placeson the surfaceof the rectangularparahelepiped.
WITHIN THE SOLID
In the same way as for the constant concentration on the surface (7.2.1), the new
concentrationafter elapseof time At, CNt j, k, is expressedby :
(7.7)
CNi,j,k
i.j.k+&
i-l,j,k-
ci,j,k+
2 i.j,k+
Ci+l,j,k
ci,
-D.!K
j+l,k
hK,-C,LJ
dX
l
X
0.5
Fig.7.3. Matter transferon the surface.Small rectangularparallelepipedof sides

Ax / 2, Ay and AZ, next to the surfaceperpendicularto the x-axis.
138
Numerical
[Ch. 7
parallelepiped
2ci,j,k+ci,
j, k+l
with the dimensionlessnumbers:
(7.8)
bx)
M, = D. At
M - (Ay)2
Y
D. At
M, = (Az2
D. At
SURFACE OF THE PARALLELEPIPED
The small parallelepipedof dimensionsAx I 2, Ay and Az located next to the surface

perpendicular to the x-axis is considered (Fig. 7.3). The matter balance during the
incrementof time At is evaluatedby consideringthe diffusion of the substancealong the yand z-axis, aswell as the diffusion and evaporation(or condensation)along the x-axis.
h. (co,j.k-
C~ . Ay. AZ. At
I

(7.20)
cN0.25.
j. k - c0.25. j. k =
CNO, j, k - 0. j, k
the new concentrationon the surfaceperpendicular to the x-axis after elapseof time At is
thus expressedin terms of the previous concentrationsin adjacentplaces.
(7.21)
CN0,j.k
= CO.j.k+~~Cl.j,k-~O.j,k)x
E(cOqj.k-Cq)
Ax
Sec. 7.31
* cO,j,k+
of matter transfer
on the surface
139
1
1
CO,j+l.k
2C0.j.k+C0,j.k+l
The new concentration on the other surfaces can be obtained by permuting the axes
and indexes in the above equation.
CONCENTRATION ON AN EDGE OF THE PARALLELEPIPED
The small rectangular parallelepiped of dimensions Ax / 2, Ay / 2, AZ, next to the

edge parallel with the z-axis is considered (Fig. 7.4). The matter balance during the
increment of tune At is evaluated within this small parallelepiped by considering the
diffusion along the z-axis and the diffusion and evaporation along the x- and y-axis.
0,5s0.25qk-
h.(C0.0.25,k-
ceq)
hlC,-&,I
A
1
$
Az.
At
y.(j)
x.19
!!!T!
2
-0 dC
dX
I
-------- l
h&-Ce,l
I
I
f-OK
I
Fig.7.4. Matter transfer on an edge. Small rectangular parallelepiped of edges

Ax / 2, Ay / 2 and AZ, next to the edge parallel with the z-axis.
Numerical
140
analysis
with a rectangular
h. (C,,25,0,k-
[Ch. 7
parallelepiped
Cd
1
$.
Az. At
Ax Ay
0.25,0.25.k+0.5
At
= AX~~~Az[cNo.25,0.25,k-c0.25.0.25,k
By making the assumptions of the kind :

CN0.2S. 0.25, k - c0.25, 0.25.k =
CNO, 0. k - co, 0. k
the new concentration after elapse of time At on the edge parallel with the z-axis is thus
expressed in terms of the previous concentration in various places :
(7.23)
CNO.O.k
= cO.O,k+&
C0,0.k-l-2C0,0,k+
- cO.O,k I
2
+ Q-
CO,l,k-
CO.O.k
--
2 h.At
Ax
1
--
cO.O,k+l 1
cO,O.k-
ceqj
[ cO,O,k-
ceq]
2 h.At
AY
The concentrations on the other two edges can be obtained by permutation of the
axes and integers in the above equation.
ON THE CORNERS
The small rectangular parallelepiped of dimensions Ax / 2, Ay / 2, Az / 2, next to the

comer is considered (Fig. 7.5). The matter balance during the increment of tune At is
evaluated by considering the diffusion and evaporation along the three perpendicular axes.
(7.24)
h. (CO, 0.25.0.25 -
C&$$Dt
Sec. 7.31
of matter transfer
on the surface
-D()C 1
&I
h(Cs-C,,
I
-.e------.-,-
--
/
i
/
dY
x(i)
w
AL
7
Fig.7.5. Matter transfer on a comer. Small rectangularparallelepipedof sides

Ax / 2, Ay / 2 and Az / 2, next to a comer.
o5 025-h.lc0.25
, . , .
o,25,0.25,0,5-
, 0 , o.2s-Gql.~.+3t
h. ~co.25.0.25.0-
C,,.$.$.Dt
I
Ax. Ay. Az
8
' CN0.25.0.25,0.25-
co.25.0.25.0.25 I
By making assumptionsof the kind :

(7.25)
CN0.25, 0.25. 0.25 = cNO. 0. 0

cl, 0.25. 0.25 = cl, 0, 0
the new concentrationafter elapseof time At on the comer is thus given by :
141
142
Numerical
parallelepiped
(7.26)
CNo,o,o
=Co,o,o+w
- c,.o,o]y[co.o.ocq]
x[c1.0.0
[Ch. 7
2
+M
Y
+&
[ co.1.0
z[
-co*o,o]-~[~o,o,o-cq]
o.o.l-Co,o,o]-~[co.o,o-c~
AMOUNT OF SUBSTANCE LOCATED IN THE PARALLELEPIPED
The amount of diffusing substancelocated in the parallelepipedis calculated at each

tune by integrating the concentrationat this time with respectto space.The calculation is
made along each axis successively,as shown above (7.2.1).
The coefficient of the concentration Co.o,o at the comer must be positive. This
condition is fulfilled when :
$+L+l+h.At
< 1
Mz
The other conditions are fulfilled when the increment of time obeys this above
equation. Within the solid, the coefficient of the concentration Ci,j.k must be positive,
leadingto :
(7.27)
(711)
1+2+2
M, My
M,
< 1
SYMMETRIES
In the sameway as for the constantconcentrationon the surface,only one eighth of

the rectangular parallelepiped is studied, and the sameequations for the symmetries are
used.
7.3.2 CONCENTRATION-DEPENDENT
DIFFUSIVITIES,
AND FINITE
COEFFICIENT OF MATTER TRANSFER ON THE SURFACE
The matter balanceduring the incrementaltime At is evaluatedin various places,in
Sec. 7.31
of matter transfer
on the surface
143
the same way as for the constant diffusivity. As the diffusivity depends on the
concentrationof substance,the function J = D. &Z / dx
where D is the concentration-dependentdiffusivity, is consideredinstead of the constant
dimensionless number IV&.The function J is the same as that already shown in section
7.2.2.
The following placesare considered.

WITHIN THE SOLID
The matter balance is evaluated in the same manner as in 7.2.2 in the small
rectangularparallelepipedshown in Fig. 7.2. The new concentrationafter elapseof time At
at the position (i, j, k), CNi, j, kr is thus given by the sameequation :
(7.15)
CNi*j,k
At
=
i,j,k+-
Ax I
-Ji-O.,.j.kJXi+O.5,j.k
1
-Jzi.j,k-0.5JZi,j,k+O.5
with the function Jx, Jy and Jz of the type :
The diffusivities at various positions are calculatedin the samemanner as in section

7.2.2, either by considering the mean value of the diffusivity for a small change in the
diffusivity with the concentration:
(7.16)
Dim0.5,j.k
= 2
Di-I.j,k+Di,j,k]
or by considering the mean value of the concentrationwhen the diffusivity highly varies
with the concentration:
(7.18)
Di-0.5.j.k = A. exP(B. Ci-0.s.j.k)
with
(7.17)
Ci-0.5,j.k
2 (Ci-l,j,k+Ci.j.k)
Numerical
144
[Ch. 7
parallelepiped
SURFACE OF THE PARALLELEPIPED
The small parallelepiped of dimensions Ax /2, Ay and AZ, next to the surface
perpendicular to the x-axis is considered (Fig. 7.3). The matter balance during the
incrementof time At evaluatedby taking into accountthe diffusion along the threeprincipal
axes and the evaporation from the surface, leads to the sameequation as for a constant
diffusivity (7.3.1). Instead of using the constantdimensionlessnumbersMx, My and Mz,
calculation is made with the functions Jx, Jy andJz.
The new concentrationafter elapseof time At is thus expressedby the relation :
CN0,j.k =
(7.28)
2 At
cO,j,k+-
Ax
- Jx O.S,j.k- h(CO,j.k-
+d -JYo.~s,j-o.5,kJYo.2S,j+O.S.k I
AY [
0.2S.j,k-0.5+JZ0.25,j.k+0.5
The diffusivity appearingin the functions J must be calculatedat the positions defined by
integers,aswell asthe concentrations.
The concentrationcan be evaluatedin the following way :
(7.29)
C0.25,j,k-0.5
(7.29)
cO,j,k-O.S
= ~c0.j,k-0.5+$c~,j.k-0.5
= ~[CO,j,k-l+CO.j.k]
The diffusivity can be evaluatedeither by the equation
(7.30)
DO,j,k-0.5
= k DO,j,k-l+%.j,k
1
when the diffusivity varieslittle with the concentration,or by calculatingthe concentrations

at the positions defined by integers and using the function of the diffusivity with
concentration:
sec. 7.31
of matter transfer
on the surface
145
EDGE OF THE PARALLELEPIPED
The small rectangular parallelepiped of dimensions Ax / 2, Ay / 2, AZ next to the

edgeparallel with the z-axis is considered(Fig. 7.4). The matter balanceduring the tune At
evaluated in the case of a constant diffusivity (7.3.1) by taking into consideration the
diffusion and evaporationleadsto the new concentrationon the edgeCNo,o,k :
CNo.0.k
(7.32)
2 At
= CO,O.k+-
JXo.5,om.k h(Co,o.25.k-
c.,i]
Ax
+ 2 At
-
- JY0.25.0.5,k-
h(C0.25.0.k-
%q)]
AY
-JZo,o,k-o.5+J-=o,o,k+o.5
CORNER OF THE PARALLELEPIPED
The small rectangularparallelepiped of dimensionsAx / 2, Ay / 2, Az / 2 next to the

comer already shown for the constantdiffusivity (Fig. 7.5) is considered.
From the matter balanceevaluatedduring the tune At within this small parallelepiped by
taking into accountthe diffusion and evaporation, the new concentration on the comer is
obtained,with the sameassumptions:
(7.33)
2 At
CNo,o.o = Co,,,,+ - Jxo.5.0.0- h(Co*o.o- ceql]
Ax
- h(co.o.o-
+-
2 At
AZ
c,
)I
- Jzo.0,o.s- h(Co.o.o- Cd]
The amount of diffusing substancelocated in the solid is calculated at eachtime by
146
Numerical
analysis
with a rectangular
parallelepiped
[Ch. 7
integrating with respectto spacethe concentrationat this time as alreadyshown (7.2.1).

CONDITIONS OF STABILITY OF CALCULATION
The coefficient of the concentrationat the comer must be positive in the samemanner
as for the constant diffusivity (7.3.1). Generally the diffusivity increases with
concentration, and the above condition can be used with the highest values of the
dimensionlessnumber Mx, My, Mz.
(7.27)
A+L+L+h.At
ti
My
m
7.4 TRANSFER
< 1
WITH
SPECIAL
CONDITIONS
As alreadyshown in Chapter6 for the plane sheet,the numerical model is capableof

taking into considerationall the facts. Three casesare of special interest for a rectangular
parallelepiped : when the material is anisotropic, when the temperatureis programmed,
when the pressureof vapour in the surrounding is programmed.
7.4.1 ANISOTROPIC MATERIAL

In the case of a polymer, generally the material is isotropic, especially when it is
pressed without an orientation of the molecules. Of course, when the polymer is
transformed into a film or a thread, an orientation of the molecules is observed, and the
material is anisotropic. Thesescasesare not considered here, as they are studied in two
specializedbooks (1,2).
Wood is a typical example of an anisotropic material with three principal axes of
diffusion : longitudinal along the tree, radial and tangential, and three principal
diffusivities. The longitudinal diffusivity is considerably higher (10 to 50 times) than the
radial and tangential diffusivities which are of the sameorder of magnitude. This caseis
also studied in the two books (1,2).
7.4.2 PROGRAMMATION
OF TEMPERATURE
In the same way as for a sheet, it may be of interest to use a programmation of
sec. 7.41
Transfer
with special conditions
147
temperature,especiallyfor drying a dosageform.

As already shown, the effect of temperature on the process is complex, as it acts
upon various parametersat the sametime : the diffusivity, the coefficient of matter transfer
on the surface h, the concentration on the surface which is necessary to maintain
equilibrium with the surroundingatmosphere.
As shown for the plane sheet,the diffusivity and the coefficient of matter transfer h
vary with temperaturein an exponentialway.
For calculation, at eachtime and eachposition, the temperatureis evaluatedfirstly,
and the diffusivity is calculated.At eachtime, the temperatureis evaluatedon the surface,
and the coefficient of matter transfer is thus calculated.In the sameway, the concentration
on the surface at equilibrium with the surrounding is calculated at each time by using a
relationshipbetweenC,, andtemperature.
IN
THE
7.4.3
PROGRAMMATION
OF THE CONCENTRATION
SURROUNDING
As shown in Chapter6 for the plane sheet,it is sometimesof interest to maintain the
vapour in the surroundinginsteadof removing it.
Two cases are generally considered : when the volume of the surrounding
atmosphereis finite, and when the pressureof the vapour is controlled.
In thesetwo cases,the concentration on the surface at equilibrium with the vapour
pressureis calculatedfirst, andthe rate of transport is thus evaluatedat eachtime by using
the following relation :
h CC,- C,,)
REFERENCES
1 J.M. Vergnaud, in Liquid transport processesin polymeric materials. Modelling and
industrial applications, Prentice Hall ed., N.Y., 1991, Chapters7 and 13.
2.J.M. Vergnaud, in Drying of polymeric and solid materials. Modelling and industrial
applications, Springer Verlag ed., 1992, Chapters8 and 13.
Numerical analysis with a radial

transport within a sphere
8.1
INTRODUCTION
The problem with a sphereis of great interest, as many dosageforms are spherical
in shape.In this case,the transfer of diffusing substanceis radial only.
The equation of radial diffusion with concentration-dependent diffusivity is as
follows :
D = qC)
and with constantdiffitsivity it becomes:

(8.2)
G = s. g[r*. $1
D = constant
or
(8.2)
= D.
where C is the concentrationof diffusing substanceat time t at a distancer from the centre
of the sphere.
Numerical
150
analysis with a radial transport
Of course, the diffusivity
within
a sphere
[Ch. 8
may be either constant or concentration-dependent.
Generally, an analytical solution exists in the case of a constant diffusivity when the initial
concentration is uniform and when the boundary conditions are simple (Chapter 4).
is concentration-dependent, no analytical solution exists, and the
problem must be resolved by using a numerical model.

However, in order to accustom the reader with numerical analysis, and also to
resolve all the problems which cannot have an analytical solution, the numerical analysis is
developed when the diffusivity is either constant or concentration-dependent.
Two cases are of great interest :
- Solid sphere with a radial transport
- Hollow sphere with a radial transport
8.2 RADIAL DIFFUSION THROUGH

CONSTANT
DIFFUSIVITY
A SPHERE
WITH
Two cases are considered, depending on the value of the rate of transfer on the
surface.
8.2.1
INFINITE
COEFFICIENT
OF
MATTER
TRANSFER
ON
THE
SURFACE
In this case, the coefficient of matter transfer h on the surface is so high that the
concentration of substance on the surface reaches the value at equilibrium as soon as the
process starts (Chapter 1 - section 1.1S).
The solid sphere of radius R is divided into N spherical membranes of thickness Ar,
and each spherical membrane of radius r is associated with an integer n :
(8.3)
R=
N. Ar
r = n.Ar
Calculation is made within the sphere and at the centre of the sphere.
WITHIN THE SPHERE,
with
IlnlN-1
(Fig. 8.1)
The circular cross-section of the spherical membrane is considered with radial

diffusion, as well as the circular membrane of radius r = n. Ar and thickness Ar. Plane
sheets are also shown for simplifying the drawing.
Sec. 8.21
Radial diffusion
through
a sphere with constant
r-b
r+br
-Hn-l
n+l
diffusivity
151
-II
I
I
I
n-l
n - 0.5
I
n+Q5
n+l
Fig. 8.1. Scheme of the circular cross-section of the sphere, with radial diffusion
of the substance within the spherical membrane of radius r. Ar.
The matter balance is evaluated during the increment of time At within this circular
membrane by considering the matter which diffuses through the circular surfaces of radii
(n -0.5) Ar and (n + 0.5) Ar.
(8.4)
At[-D./A.F)_o.~+D.jA.~)~+o.~
= V,. [CN,-C,,]
where A,, is the area of the circular surface of radius r = n. Ar

D is the constant diffusivity
Iz.I
is the gradient of concentration at position n .
V, is the volume of the circular membrane of radius r and thickness Ar.

CN, and C, are the new concentration after elapse of time At and the previous
152
Numerical
within
a sphere
[Ch. 8
concentrationwithin the circular membraneof radius n. Ar, respectively.

When n is rather high, the volume of the sphericalmembraneV,, can be given by the
simple relation 4 rr n*. (Ar)3. A more accurate value is obtained by calculating the
difference betweenthe volumes of the spheresof radii (n + 0.5) Ar and (n - 0.5) Ar. This
more precisevalue for the volume V, is thus : 4 7~(A# (n* + 1 / 12).
By replacing the areaof the circular surface4, _o.5by 4 x (n - 0.5)*. (Ar)*, and the
gradient of concentrationat position n - 0,5 by the chord slope betweenthe position n - 1
and n, it becomes:
(8.5)
D. (Ar)*. (n - 0.5)* -i;
(Ar)(n*+A)[CN,-
(Ar)*. (n+0.5)* Cn-Cn+1

Ar
. At =
I
c,J
The new concentrationCN, after elapseof tune At within the sphericalmembraneof

radius n. Ar is thus expressedin terms of the previous concentrations in the same and
adjacentplaces
(8.6)
CN, = C,+
(n-0.5)2(C,-r-C,)-(n+0.5)2(C,-C,+1
Fig. 8.2. Schemeof the small sphereof radius Ar / 2 with radial diffusion
through the sphericalsurfaceof radius Ar 12.
Sec. 8.21
Radial diffusion
through
a sphere with constant
diffusivity
153
n=O
with
The small sphereof radius Ar / 2 with the samecentre as the sphereof radius R is
considered(Fig. 8.2). The matter balanceduring the incrementof time At within this small
sphereis evaluatedby taking into accountthe diffusion through the surfaceof the circular
surfaceof radius Ar 12.
AT THE CENTRE OF THE SPHERE,
+ D. (A. $)o,i
(8.7)
At = ;rt [f)s. [CNo- Co]
By replacing the area of the spherical surface Ao.5by its value, this equation leads
to :
CN, = C,,-&-
(8.8)
C,)
where M is a dimensionlessnumber
(Ar I2
M = D. At
(8.9)
ON THE SURFACE OF THE SPHERE,
with
n = N,orr = R
The concentration on the surface of the spherereachesthe value at equilibrium as

soon as the processstarts.
c, = c, = c,,
AMOUNT OF SUBSTANCE LOCATED IN THE SPHERE
The amount of diffusing substancelocated in the sphere at time t is obtained by

integratingwith respectto spacethe concentrationwithin the sphereat this time.
R
(8.11)
M, =
04rcr2. C,,,. dr
By using the finite differences,this expressionbecomes:
154
Numerical
Mt
(8.12)
analysis
= c-+y
24
4 IT(Ar)3
8.2.2
FINITE
with a radial transport
(n2+g
within
a sphere
C,,,+[N3-(N-0.5)3].
[Ch. 8
[&CN+&-l]
n=l
COEFFICIENT
OF
MATTER
TRANSFER
ON
THE
SURFACE
The problem within the sphere and at the middle of the sphereis the same as that
studiedin the previous section (8.2.1), and the equationsare the same.
with
n=N,orr=R
The sphericalmembraneof thicknessAr / 2 located next to the surfaceof the sphere
is considered (Fig. 8.3).
The matterbalancewithin this sphericalmembraneduring the incrementof time At is
evaluatedby consideringthe diffusion through the sphericalsurfaceof radius (R - Ar / 2)
andthe transferthrough the externalsurfacewith the coeffkient of matter transfer.
CONCENTRATION ON THE SURFACE,
Ar
4%
h$-Ceq)
D
Ccxt
N -0.5
Fig. 8.3. Schemeof the sphericalmembraneof radius (N - 0.25) Ar and

thicknessAr / 2, with diffusion and convection.
Sec. 8.31
Radial diffusion
h- AN (CN- Cq) =
3
through
vN-0.25
a sphere
ICNN-0.25-
155
CN-0.2S
whereh is the coeffkient of matter transferthrough the externalsurface

CN and C,, are the concentrationon the surfaceat time t and the concentrationrequired to
maintain equilibrium with the surroundingatmosphere,respectively.
VN _ o.25is the volume of the sphericalmembraneof radius (N - 0.25) Ar.
AN _ o,5and AN are the areasof the spherical surfacesof radius (N - 0.5) Ar and N.Ar,
respectively.
CN~-o.25-
(8.14)
CN-0.25
= CN,-
CK
the new concentrationon the surfaceCNN is thus obtained :

(8.15)
N - 0.5
CNN = CN +
(CN-,-CN]
N3-(N-0,5)3A
N
N3-(N-0.5)3
h. At
-.
Ar
(cN-
Of course,a better assumptioncan be made :

(8.16)
CN~-o.zs-
CN-0.25
= i(CN,-
CN]+;(CNN-l-CN-l)
AMOUNT OF SUBSTANCE LOCATED IN THE SPHERE
The amount of substancelocated in the sphereis obtained in the sameway as for an

infinite coeffkient of matter transferon the surface.
8.3 RADIAL DIFFUSION THROUGH A SPHERE

DIFFUSIVITY
WITH
transfer h on the surface.
ceq]
Numerical
156
within
a sphere
[Ch. 8

SURFACE
The concentration on the surface reachesthe value at equilibrium as soon as the
processstarts.
The concentrationis thus determinedin two positions ; within the sphereand at the
middle.
IlnlN-1
with
The spherical membraneof radius r = n. Ar of thicknessAr is considered(Fig. 8.1).
The matter balanceevaluatedduring the incrementaltime At within this membraneleadsto
a relationship similar to eqn (8.4) :
WITHIN THE SPHERE,
with a difference due to the fact that the diffusivity dependson position at any time.
Upon putting the function J, as follows (1) :
J, = n*. D,. (G-o.~-G,+o.J)
(8.18)
the new concentration at position n after elapse of time At is thus expressed by the
equation:
(8.19)
CN, = C,+
(Ar)2. L*+ A)
[- Jr,-o.s+ Jn+o.5
n=O
with
The small sphereof radius Ar / 2 with the samecentre as the sphereof radius R is
considered(Fig. 8.2). The matter balanceduring the incrementaltime At within this small
sphereis evaluated:
CENTRE OF THE SPHERE,
(8.20)
+(D.A.$],.;
At = ;+$
[CN,-Co]
The new concentrationCNo at the centreof the sphereafter elapseof time At is thus
Sec. 8.31
Radial diffusion
through
a sphere
157
expressedby :
(8.21)
CN, = Co--.
24. At
Jo.5
(Ad
with the function Jo.5given by :
(8.18)
8.3.2
Jo.s = a. Q&,-Cl)
FINITE
COEFFICIENT
OF
MATTER
TRANSFER
ON
THE
SURFACE
With the concentration-dependentdiffusivity, the problem is the sameas in section

8.3.1 in all positions exceptfor the surface.
llnlN-1
with
The new concentrationafter elapseof time At is given by eqn (8.19) :
WITHIN THE SPHERE,
(8.19)
CN, = C,+
- Jn-0.5+Jn+o.5
with the function J, :

(8.18)
J, = n2.D,. (C,-O.S-
G+o.s)
n= 0
with
The new concentrationCNo is given by eqn (8.21) :
24. At
(8.21). CN, = Co- Jo.5
br)
n= N
with
The spherical membrane of radius (N - 0.25) Ar and thickness Ar / 2, next to the
surface, is considered (Fig. 8.3). The matter balance during the increment of time At is
SURFACE OF THE SPHERE,
Numerical
158
within
[Ch. 8
a sphere
evaluated within this spherical membrane by considering the diffusion through the surface
of radius (N - 0.5) Ar and the convection on the surface :
h. A, (C,-
Cq) . At = V. [CNN-o.,,-
C,-,.,,I
where V is the volume of this spherical membrane
(8.23)
V = 3~
C eq is the concentration on the surface required to maintain equilibrium
with the
surrounding atmosphere.
and AN _ o.s and AN are the areas of the spheres through which the spherical membrane is
located.
With the simple assumption :
(8.14)
CNN
0.25 - CN - 0.25 = CNN - CN
the new concentration on the surface after elapse of time At is given by :
(8.24)
3.At
CNN = CN +
JN-0.5
(AT). IN3 - (N - 0.5)3]
3 h. N2. At
kN-
ceq)
Ar.[N-(N-0.J13]
A better assumption is made with :
CNN-O.~~CN-0.25
= 43 (CNN-cN]+ $(CNN-l- cN- 1)
(8.16)
AMOUNT OF MAlTER TRANSFERRED
The amount of matter transferred can be calculated by integrating the rate of matter
transferred on the surface with respect to time.
(8.25)
M, = 4 IF N2. (Ar)2. 2
0
h (CN- Cq). At
Sec. 8.41
Constant
HOLLOW
SPHERE
WITH
8.4
CONCENTRATION
ON THE INTERNAL
CONSTANT
DIFFUSIVITY
diffusivity
159
CONSTANT
SURFACE,
AND
This case is of high interest for dosage forms, especially for those where the drug is
surrounded by a spherical membrane.
The spherical membrane of internal and external radii Ri and R, is considered (Fig.
8.4). This spherical membrane of thickness Ro - Ri can be divided into N spherical
membranes of thickness Ar, in such a way that :
NC!
Ro
Fig. 8.4.
Scheme of the spherical membrane of internal and external radii R and
Ro, with constant concentration on each surface Ci and Co.
(8.26)
Ro- Ri = (No- Ni) Ar = N. Ar
In this case, the equations are very simple. However there is a drawback since the
thickness of the membrane Ro - R, as well as the internal radius R must be divisible by
Ar.
A better way for selecting the value of Ar consists of dividing the thickness of the
membrane as follows :
(8.26)
R, = R + N. Ar
160
Numerical
analysis
with a radial transport
within
with
r = & + n. Ar
a sphere
[Ch. 8
OSnlN
transfer on the external surface.

EXTERNAL SURFACE
OF MATTER
TRANSFER
ON THE
The problem is studied in the same way as for the solid sphere, under the same
condition with constant diffusivity.
with
WITHIN THE SPHERICAL MEMBRANE,
llnlN-1
The new concentration after elapse of time At is expressedby the relation :
(8.6)
CN, = C,+
(n-0.5)2(C,-1-C,]-(n+0.S)2(C,-C,+1)
with the dimensionless number M
(8.9)
br I2
M = D. At
INTERNAL AND EXTERNAL SURFACES,
with
n = 0,orn
= N
The concentrations are constant on the internal and external surfaces :

(8.27)
r=F$
C = Ci
internal surface
(8.28)
r=Ro
c = c,
external surface
Of course, the substance diffuses into or out of the spherical membrane, depending
on the relative values of the concentrations on the surfaces
(8.29)
Ci > C,
the substanceleaves the membrane
(8.30)
Ci < C,
the substanceenters the membrane
Sec. 8.41
Constant

EXTERNAL SURFACE
OF
MATTER
diffusivity
TRANSFER
161
ON
THE
The spherical membrane of internal and external radii Ri and Ro, with a finite
coeffkient of matter transfer on the external surface, is considered (Fig. 8.5).
Three positions are considered : on the internal surface, within the spherical
membrane, on the external surface.
ON THE INTERNAL SURFACE,
with r = R
and n = 0
The concentration is constant :

C = Ci
r=R
(8.27)
with 1 I n I N - 1
The new concentration is given by the equation already given within the solid
sphere, and in the previous section (8.4.1).
(8.6)
CN, = C,+
No-0.5
Fig. 8 S.
No-0.25
Scheme of the spherical membrane of internal and external radii R and
Ro, with constant concentration Ci and a finite coefficient of matter transfer on the
external surface.
Numerical
162
within
a sphere
[Ch. 8

r= R,
n= N
and
The spherical membrane of radius (R, - Ar / 2) next to the external surface is
ON THE EXTERNAL SURFACE,
considered (Fig. 8.3).

The new concentrationon the external surfaceis given by the sameequation as that
obtainedfor the solid sphere(section 8.2.2) :
(8.15)
N - 0.5
CN, = CN +
I
N. h. At
lc~-rcrJ
N3-(N-0.5)3.
[N3-(N-0.5)3]Ar
(c,-cd
with the simple assumption:

(8.14)
CNN - 0.25 - CN - 0.25 = CNN - CN
Of course,the better assumptioncan be made :

(8.16)
CNN-O.Z~-CN-O.~~
= ~(CNN-CN)+~~CNN-~-CN-I)
HOLLOW SPHERE WITH CONSTANT CONCEN8.5

TATION ON THE INTERNAL SURFACE, AND CONCENTRATION-DEPENDENT
DIFFUSIVITY
The two caseswith an infinite or finite coefficient of matter transfer on the external
surfaceare considered.Theseproblems are studiedin the caseof a solid sphere.
8.5.1
INFINITE
EXTERNAL
COEFFICIENT
OF
MATTER
TRANSFER
ON
SURFACE
The new concentrationsare given in variouspositions :

with
INTERNAL SURFACE,
(8.27)
C = Ci
r=q
r=R;
and
n=O
THE
Sec. 8.51
Concentration-dependent diffusivity
with
(8.19)
163
l<nlN-1
CN, = C,+
- Jn-0.5+ Jn+o.5
with the function J
J, = n2. D,. (C,-o.s- Cn+o.s)
(8.18)
with
EXTERNAL SURFACE,
r=
c = co
(8.28)
8.5.2
FINITE
EXTERNAL
r=&
and
n=N
RO
COEFFICIENT
OF
MATTER
TRANSFER
ON
THE
SURFACE
The new concentrationsare given in various positions, as they are found with the
solid sphere (8.3.2).
with
INTERNAL SURFACE,
r=&
C = Ci
(8.27)
WITHIN THE SHERICAL MEMBRANE,
(8.19)
n = 0 , r = Ri
with
15 n I N- 1
[- J, -OS+Jr,+0.5
CN, = C,+
with the function J

with
EXTERNAL SURFACE,
n = N , r = R,
With the simple assumption

(8.14)
CNN
- 0.25
cN - 0.25
CNN
- CN
164
Numerical
within
a sphere
[Ch. 8
the new concentrationon the externalsurfaceis given by :
(8.24)
3.At
CN, = CN +
A better assumptionis madewith :

(8.16)
CN-0.25 = ~[CNN-CNj+~(CNN-l-CN-l)
CN,-,.,,-
8.6 CONCLUSION
The conditions of stability for calculation are obtained when the coefficient of the
previous concentrationat the position consideredis positive. It is thus found :
- Within the sphere,in eqn (8.6)
(8.29)
1>
Mn
I
+=
(n - O.5)2+ (n + O.J)2
1
I
or
(8.30)
M >
2n2+0.5
2 1
n +iT
Centre of the sphere,in eqn (8.8)

(8.31)
1 -&
> 0
or
M>6
BOUNDARY CONDITIONS
The concentration in the surrounding atmospherecan be changed or programmed

during the process,as well as the concentrationon the surfaceC,, which is at equilibrium
with the concentrationin this surroundingatmosphere.
Conclusions
Sec. 8.61
165
PROGRAMMATION OF TEMPERATURE
The temperature can be programmed during the process. Generally, the rate of heat
transfer is much higher than the rate of matter transfer by diffusion. The temperature can
thus be considered as uniform within the sphere and equal to that of the system of
programmation (2,3).
The diffusivity, and the coefficient of matter transfer on the surface, as well as the
concentration in the surrounding atmosphere and the concentration on the surface of the
sphere at equilibrium with this concentration are functions of temperature.
After each increment of time, these values are calculated as a function of the
temperature obtained at this time.
DlFFUSlVll-Y
The diffusivity can by expressed in terms of the concentration of the substance by

many ways.
varies little, the mean value of the diffusivity
may be
calculated :
(8.32)
Dn+o.5
= ;[%+D,+,]
When the diffusivity highly varies, the concentration is determined beforehand, and
the diffusivity is expressed in terms of this concentration, e.g., in an exponential way :
(8.33)
Dn+o.5= D. exp(ae- C, +0.5)
REFERENCES
J.M. Vergnaud, in Liquid Transport Processesin Polymeric Materials. Modelling
1
and Industrial Applications, Prentice Hall ed., 1991, Chapter 6.
2
J.M. Vergnaud and J. Bouzon, in Cure of thermosetting resins. Modelling and
Experiments, Springer Verlag ed., 1992, Chapter 13.

3
J.M. Vergnaud, in Drying of polymeric and solid materials. Modelling and
Industrial Applications, Springer Verlag ed., 1992, Chapter 15.
9
Numerical
9.1
analysis with cylinders
INTRODUCTION
The caseof cylinders is of high interest either from a theoretical point of view since
the problem may be complex or from a practical point of view as they have many
applications.
In the same way as for solids of other shapes,the mathematical treatment is only
feasible when the diffusivity is constantand when the initial and boundary conditions are
rather simple. In all other cases,the numerical treatmentmust be used.
The following three casesare studied:
(i) Solid cylinder of infinite length. The diffusion of the matter is radial only, asthe effect
of the edgesis negligible. The coefficient of matter transfer on the surface is either very
high, or finite. Moreover, the diffusivity is constantin 9.2.1 and concentration-dependent
in 9.2.2.
(ii) Hollow cylinder of infinite length. The diffusion is also radial only. The important case
of the tubing usedfor the transportof a liquid is considered.
(iii) Solid cylinder of finite length. This caseis of high interest,asmany dosageforms are
shapedin the form of pellets. The transfer of diffusing substanceis radial andlongitudinal
at the sametime.
168
Numerical
9.2 SOLID
CYLINDER
[Ch. 9
OF INFINITE
LENGTH
As the effect of the edges is negligible, the diffusion of matter is radial only, and the
circular cross-section perpendicular to the longitudinal &is is thus considered (Fig. 9.1).
jm4
j -0.5
j+QS
j+l
Fig. 9.1. Circular cross-section of the solid cylinder of infinite length.Annulus

on thickness Ar and radius j. Ar.
The circle of radius R is divided into N circular annuli of constant thicknesses Ar,
and each annulus is associated with an integer j.
(9.1)
r = j.Ar
R=
N. Ar
Two cases are considered, depending on the diffusivity which may be either constant
or concentration-dependent.
In each of these two cases, the coefficient of matter transfer on the surface is either
infinite or finite.
sec. 9.21
Solid cylinder
of infinite
length
169
9.2.1
Three places are consideredin succession: within the cylinder, on the longitudinal
axis of the cylinder, on the surfaceof the cylinder.
lljlN-1
with
The annulus of thickness Ar and unit length, located betweenthe positions (j - 0.5)
and cj + 0.5) is considered(Fig. 9.1).
The matter balanceduring the increment of time At is evaluatedwithin this a~ulus
by taking into accountthe diffusion of the substancethrough the two cylindrical surfaces:
WITHIN THE CYLINDER,
(9.2)
[-D./A.~~j-~~~+D.~A.~~j+o,5].At
= Aj. Ar.(CNj-Cj)
where Aj. Ar representsthe volume of this annulusper unit length

Aj -0.5 is the areaof the cylindrical surfaceof radius cj - 0.5) Ar and unit length
I2I
j-O.5
is the gradientof concentrationat position (i - 0.5) and D is the constantdiffusivity
By replacing the volume of the annulus and the areaof the two cylindrical surfaces
by their values,the matter balancebecomes:
(9.3)
2~. D. Ar. -(j-0.5)($)j-o
5+ij+0.5)(~/j+o~~At
= 2nj.(Ar)2[CNj-CJ
By assumingthat the gradient of concentrationis equal to the chord slope between

j-landj:
(9.4)
(g)j+o
5 = b;
cj
the following equationis obtained:
(9.5)
j(
(CNj- Cj) D
= (j+O.5)(Cj+1-Cj)-(j-O.5)(Cj-Cj-1)
170
Numerical
analysis
with cylinders
tCh. 9
Fig. 9.2. Small cylinder of radius Ar and centre0.
Upon putting the dimensionlessnumber
(9.6)
(Ad2
M = D. At
the new concentration after elapse of time At, CNj, is thus expressedin terms-of the
previous concentrationsat the sameand adjacenttwo places:
(9.7)
CNj = b[ C,+1 +(M-2)Cj+Cj-
11+ &[cj+l-cj-l]
with j = 0
The small cylinder of radius Ar / 2 with the sameaxis asthat of the cylinder of radius
ON THE AXIS OF THE CYLINDER,
R, is considered (Fig. 9.2).

The matter balance during the increment of time At is evaluated within this small
cylinder of unit length, by considering the matter which diffuses through the circular
surfaceof radius Ar / 2.
(9.8)
D. (A. ?&,.
At = rt [f12. (CN,- Co)
By replacing the gradient of concentrationat position 0.5 by the chord slopebetween
Sec. 9.21
Solid cylinder
of infinite
length
171
1 and 0, the new concentrationCNo is thus obtained :

(9.9)
CN, = Co- ;(C,-
C,)
SURFACE OF THE CYLINDER, WlTH INFINITE COEFFICIENT OF MATTER TRANSFER, j = N
The concentration of diffusing substanceis thus constantly equal to the constant

value requiredto maintain equilibrium with the surroundingatmosphere
(9.10)
CN = ce, = constant
N
The annulusof thicknessAr / 2 located next to the external surfaceof the cylinder is
considered (Fig. 9.3). The matter balance within the increment of time At is evaluated
within this annulusby taking into accountthe diffusion through the surface (N - 0.5) and
convectionout or onto the external surfaceN :
SURFACE OF THE CYLINDER, WITH FINITE COEFFICIENT OF MATTER TRANSFER, j =
AN h. icN-
c-1
= VN-0.25 I CNN-om-
CN-0.25 1
By replacing the volume V of this annulus, and the areasof the surface (N - 0.25)
and N by their value, this equationbecomes:
N- OS
I
I
I
I
;N-015
I
I
I
Fig. 9.3. Small annulus of radius (N - 0.25) Ar and thickness Ar / 2.
172
Numerical
(9.12)
analysis
with cylinders
2 x (N - 0.5) Ar. D.
cN-
1- cN]
[Ch. 9
- 2 7t N. Ar. h (C,-
Ar
27t(N-0.25).
Cq) At =
Ar.~(CNN_o.2J-C,_o,,,i

(9.13)
CNN - 0.25 - CN - 0.25 = CNN - CN
the new concentration after elapse of time At, CNN, is thus expressed in terms of the
previous concentrations :
(9.14)
CNN= CN + &. N-0'5
(CNTl-CN) -
2N
N-0.25
N - 0.25
@N-
ceq)
with the dimensionless number
(Ad2
M = D. At
(9.6)
(9.13)
h. At
P = Ar
By making the more accurate assumption

(9.15)
cNN-o.25-
CN-0.25 =
%(CNN-CN)+~(CNN-~-
cNal)
the new concentration on the surface CNN is thus obtained as follows :
(9.16)
CNN = CN-~(CNN-~-CN-~)+
N-05.
&.&,_,)N - 0.25
N
N -0.25
(c,
3p
AMOUNT OF MA-ITER LOCATED IN THE CYLINDER
The amount of substance located in the cylinder at time t, per unit length, is obtained
Solid cylinder of infinite length
Sec. 9.21
173
by integratingwith respectto spacethe concentrationat time t :
(9.17)
M, = 2x
r. C,,,. dr
where C, t is the concentrationat the radius r and time t.

With the finite differences,this equationbecomes:
2
(9.18)
M, = II
. Co+2n(*r)2.
j.C j + x (N - 0.25) (Arj2 a CN+ : CN- r
The conditions of stability for calculation are obtainedby writing that the coefficient
of Cj in eqn (9.7) and of Co in eqn (9.9) is positive.
(9.19)
(9.20)
9.2.2
M>2
M>4
for
for
Cj
Co
DIFFUSIVITY
The samethree places are also considered: within the cylinder, on the longitudinal
axis, and on the surfaceof the cylinder.
with 1.S j IN-1
The matter balanceduring the incrementof time At is evaluatedwithin the annulusof
radius j. Ar and thickness Ar shown in Fig. 9.1.
(9.21)
[(-D.A.~)j~o~r+(D.A.~)j+o~~At
= Aj. Ar.[CNj-Cj]
where D, A and aC / Jr representsthe diffusivity, the areaof the cylindrical surface and
the gradient of concentration,at positions (j - 0.5) and (j + 0.5).
By replacing A and aC / dr by their values, and upon putting the function I as
follows :
(9.22)
Ij +0.5= (j+OJ). Dj+o.F(Cj-Cj+l)
Numerical
174
[Ch. 9
the new concentration after elapse of time At, CNj, is thus expressedin terms of the
previous concentrations:
(9.23)
CNJ = Cj+ ~.
[Ij -0.5- Ij +,,I
j. (Ar)2
with j = 0
The matter balancewithin the small cylinder of radius Ar / 2 with the sameaxis as
the cylinder of radius R, is evaluatedduring the incrementof time At (Fig. 9.2) :
ON THE AXIS OF THE CYLINDER,
[D.A.$)o,;At
(9.24)
= r($j2.[CN,,-Co]
The new concentrationon the axis after elapseof time At is thus given by :
(9.25)
CN, = C,,- =.
br)
Io.5
with the function 1o.s:

(9.22)
I,., = LD
2 0.9 b
Cl)
ON THE SURFACE OF THE CYLINDER, WITH INFINITE VALUE OF THE COEFFICIENT OF

MAlTER TRANSFER
The concentration on the surface is thus constantly equal to the concentration

required to maintain equilibrium with the surroundingatmosphere.
(9.10)
cN
= ce, = constant
ON THE SURFACE OF THE CYLINDER, WITH FINITE VALUE OF THE COEFFICIENT OF

MAlTER TRANSFER
The matter balance during the increment of time At is evaluated within the small
annulusof thicknessAr / 2 locatednext to the surface.
Solid cylinder
Sec. 9.21
A, h.(CN-Ceq)
(9.26)
of infinite
175
length
1 [
At = V. CN,_,.,,-C,-o.,,I
The new concentration within the annulus of radius (N - 0.25)Ar is thus given by
(9.27)
CN,-,.,,-
C,-,.,,
2. At
2 IN-O.,-
N 2 ;,;
(N- 0.25)(Ar)
p* @N-
cq)
With the simple assumption

(9.12)
cN~-o.25
- CN-0.25 = CNN - CN
the new concentration on the surface is thus :
(9.28)
CNN = CN+
2. At
2 IN-O.S(N- 0.25)(A.r)
N 2 ;25.
-
p. @N- ceq)
With the more accurate assumption

(9.15)
cNN-o.25-cN-o.25= i(cNN- c,) + $(cNN- 1- cN- 1)
the new concentration on the surface becomes :
(9.29)
8. At
CNN = CN-~(CNN-~-CNJ+
2 IN-O.,3 (N - 0.25)(Ar)
8N
. $NN - 0.25
As the new concentration at position N, CNN, is expressed in terms of the new

concentration at position N - 1, CNN _ t, this new concentration CNN _ t must be
calculated beforehand by using the equation obtained within the cylinder.
AMOUNT OF MAlTER LOCATED IN THE CYLINDER
The amount of matter located in the cylinder is calculated by using the same equation
as that obtained with a constant diffusivity.
ceq)
176
Numerical
[Ch. 9
The problem is the same as that shown for a constant diffusivity.

CALCULATION OF THE DIFFUSIVITY
The concentration is calculated at positions defined by the integer j. For places (j OS), the diffusivity or the concentration must be approximated.
When the diffusivity varies little, the mean value of the diffusivity can be used :
(9.30)
Dj-0.5
= k(Dj+Dj-lJ
When the diffusivity highly varies with concentration, the mean concentration is
evaluated, and the diffusivity
is calculated for this mean concentration by using a
relationship of this form :

(9.31)
Dj-0.5 = D.exp(A.Cj-o.s)
where D and A are constants.
9.3 HOLLOW
CYLINDER
OF INFINITE
LENGTH
As the effect of the edges is negligible, the diffusion of the substance within the
hollow cylinder of infinite length is radial only. The circular cross-section of the hollow
cylinder is considered (Fig. 9.4), with the internal and external radii R and R,.
The hollow cylinder is divided into N annuli of constant thicknesses Ar, each
annulus being associated with an integer j in the following way :
(9.32)
r = R+j.Ar
with
OljlN
R, = R + N. Ar
Calculation is made for the hollow cylinder in the same way as for the solid cylinder
in many places, with either a constant or a concentration-dependent diffusivity.
Generally, the hollow cylinder (or tubing) is used for the circulation of a liquid, and
the internal surface is in contact with the liquid. The concentration of the liquid on the
internal surface is thus often assumed to be equal to the value attained at equilibrium, with
an infinite value of the coefficient of matter transfer on the internal surface.
Sec. 9.31
Hollow
cylinder
of infinite
length
177
Fig. 9.4. Circular cross-sectionof the hollow cylinder of infinite length.Annulus

of thickness Ar and radius j. Ar.
9.3.1
CONSTANT
DIFFUSIVITY
The concentrationof the diffusing substanceis evaluatedin many places : within the
solid, on the internal surface,and on the external surface.
with Ri<r<R,,
or
lljlN-1
The annulus of radius R + j. Ar and thickness Ar is considered (Fig. 9.4). The
matter balance during the increment of time At is evaluated within this annulus of unit
length in the sameway asfor the solid cylinder.
WITHIN THE SOLID,
(9.33)
[-D.[A.~)j-o.5+D.[A.~)j+o,~.At
= Vj. [CNj-CJ
As the areaof the cylindrical surface$4.5 and the volume of the annulusVj are :
(9.34)
Aj-0.5 = 2x[Ri+(j-O.S)Ar]
(9.35)
Vj = 2x(Ri+j.Ar).
Ar
the new concentrationafter elapseof time At, CNj, at position j is thus expressedby :
178
Numerical
analysis
(9.36)
CNj = Cj+~
with cylinders
[Ch. 9
[Ri+(j-0.5)Ar](Cj-r-Cj)+[Ri+(j+0.5)~r].
I
(Cj+l-Cj)
with r = R
and
j=O
The concentrationon the internal surface is constantly equal to the concentrationin
the solid which is at equilibrium with the liquid, when a liquid is locatedwithin the hollow
cylinder.
(9.37)
CR = C,, i = constant
ON THE EXTERNAL SURFACE, WITH AN INFlNlTE COEFFICIENT OF MAlTER TRANSFER
The concentrationon the externalsurfaceis thus constantlyequalto the concentration

in the solid required to maintain equilibrium with the surroundingatmosphere.
(9.38)
CRe = Ceq,e = constant
ON THE EXTERNAL SURFACE, WITH A FlNlTE COEFFICIENT OF MAlTER TRANSFER
The annulus of thickness Ar / 2 located next to the external surface of the hollow
cylinder is considered,with a unit length (Fig. 9.5).
Cext
Cext
-0%
- lk--
Ni
Nit05
-D!E
-L
h&y&)
b
r
Ne-OS Ne
Fig. 9.5. A~ulus of thicknessAr / 2 next to the external surfaceof the hollow
cylinder,- with a fmite coefficient of matter transferon this surface.
Hollow
Sec. 9.31
length
179
The matter balanceduring the increment of tune At is evaluatedwithin this a~ulus

by taking into account the diffusion through the surface (R, - Ar / 2) and convection
through the exemal surface%.
(9.39)
At. [- D. (A. $),-,.,-
AN h-(&r
&.)I
= h-,.&,-ox
%-0.25)
By replacing the areas AN _ o.s and AN, as well as the volume V, _ o.25,by their
values,the new concentrationis thus obtained:
(9.40)
CN,-0.25
= CN-0.25+M
4 2R,-Ar
4R -Ar~cN-I-CN)-48~e~r+N-C.q.e)
c-
The new concentrationon the surfaceCN can easily be obtainedeither by making the
simple assumption
(9.12)
C&i - 0.25 -
CN - 0.25 = CNN - CN
or the more accurateassumption:

(9.15)
9.3.2
CN,-,.2s-C~-o.25
= 34 (CNN-CN)+$(CNN-l-CN-I)
DIFFUSIVITY
The concentration of the diffusing substanceis calculated in the following three

places: within the solid, on the internal surface,on the external surface.
with Ri<r<R,,orl.SjIN1
The annulus of radius Ri + j. Ar and thickness Ar is consideredin the sameway as
for constant diffusivity (Fig. 9.4). The matter balance during the incremental time At
within this annulusof unit length is written as follows :
WITHIN THE SOLID,
(9.41)
[-(A.D.~)j-o,~+(A.D.~)j+o.~At
= Vj.[CNj-Cjl
where the diffusivity Dj _ 0.5and Dj + 0.5may be different.

With the values alreadycalculatedfor the areaof the cylindrical surfacesAj _ o,5and
Numerical
180
Aj + 0.5 and for
[Ch. 9
the volume of the ~IIIIU~US Vj. (9.3.11,the new concentration CNj is thus
expressedin terms of the previous concentrationsby the relation :
(9.42)
At. Ij-0.5
CNj = Cj+
At. Ij +0.5
Ar[Ri+j.Ar]-Ar[Ri+j.Ar]
with the function Ij +0.5
(9.22)
Ri+(j+0.5)Ar
Ij+O.5
D,
J+0.9 (cj-cj+l)
Ar
with
r = l$ and j = 0
As shown in section 9.3.1, the concentration on the surface is equal to the

concentrationrequiredto maintain equilibrium with the liquid : Ces.i.
(9.37)
CRi
Gq.
= constant
ON THE EXTERNAL SURFACE, WITH AN INFlNllE COEFFICIENT OF MA-l-l-ER TRANSFER
The concentration on the external surface is equal to the concentration which is at

equilibrium with the surrounding,C,, e :
CRC
(9.38)
= Ceq.e
= constant
ON THE EXTERNAL SURFACE, WITH A FlNlTE COEFFICIENT OF MATTER TRANSFER
The annulusof thicknessAr / 2 next to the external surfaceof the hollow cylinder is
considered(Fig. 9.5), with a unit length.
The matterbalanceduring the incrementof time At is calculatedwithin this annulus.
(9.43)
-D&
[(
. ar
N o,5I-
A,
h(GrCeq.e)
At = VN-0.2,.
(CN-ON
CN-0.2s)
The new concentrationat position (N - 0.25) is thus obtained :

8 At. IN-o.~
(9.44)
CN,-,.25
= CN-0.25
Ar[4Re-Ar]
4 f ReAr* p* (CNe-
&)
Sec. 9.41
Solid cylinder
of finite length
181
with the function 1~_ 0.5
(9.22)
IN-O.5 =
2R,-Ar
2 Ar
DN-0.9
I cN-
l-
CNI
The new concentrationon the external surfaceis easily calculated,either by making

the simple assumption
(9.12)
CNN - 0.25 - CN - 0.25 = CNN - CN
or by making the more accurateapproximation:

(9.15)
CN,-a,~-
9.4 SOLID
(&-0.25
CYLINDER
= 43 (CNN-cN~f$(c~N-l-CN-l)
OF FINITE
LENGTH
This case is certainly the most important for the cylinders, as many dosageforms
have this type of shape.
The cylinder with radius R and length 2 Z is considered(Fig. 9.6). The diffusion of
the substancewithin this cylinder is both radial and longitudinal. The cylinder is thus
divided into N, annuli of thicknessAr and of length AZ, Ah being defined as follows :
The diffusivity is either constantor concentration-dependent.
(9.45)
R = N,. Ar
z= N,. AZ
r = j.Ar
z = z.Az
9.4.1
The concentration is calculated in various places : within the cylinder, on the
longitudinal axis, on the cylindrical surface, on the plane surface, on the center of the
plane surface,on the comer.
with r=j.Ar
and z = i.Az
The small annulus of radius r = j. Ar and thickness Ar, and length AZ, is shown
182
Numerical
Kh. 9
-H
Fig. 9.6. Cylinder of radius R and length 2 H.
(Fig. 9.7). The matter balance during the increment of time At is evaluated within this
small annulusby consideringthe radial andlongitudinal diffusion.
(9.46)
+o 5 J = ~[CNi.j..
Ci.l]
The volume Vi, j and the areasAj l 0.5and 4 _ 0.5are :

(9.47)
Vi* j = 2 x. j. (Ar)*. AZ
Ai, j f 0.5 = 2 n (i f 0.5) Ar. AZ
Ai-0.5.j = A,+o,s,j = 2n.j. (Ad*
The new concentration after elapse of time At within this
annulus,
CNi, j, is thus
Sec. 9.41
Solid cylinder
of finite length
183
h
i+l
i -0.5
j+aS
j - 0.5
Fig. 9.7. Small annulus of radius r = j. AI, Ar thick and AZ long at position r and
Z.
expressedin terms of the previous concentrationsat the sameand adjacentpositions by the

following expression.
(9.48)
CNi,j = Ci,j+&[Ci,j-l-2ci,j+Ci,j+l
I
-Ci,j-1
with the dimensionlessnumbersM, and M, :
(9.49)
bd*
M, = D.At
1
1+& 2[
Ci-l,j-2Ci,j+ci+l,j
184
Numerical
[Ch. 9
Fig. 9.8. Small cylinder of radius Ar / 2 and length AZ.
(9.50)
(AZ)
M, = D.At
with r = 0 and z = i.Az

The small cylinder with the same axis as the large cylinder, of radius Ar / 2 and
length AZ is considered(Fig. 9.8). The matter balanceduring the increment of time At is
calculated within this small cylinder by accounting for the radial diffusion through the
cylindrical surface at Ar / 2 and the longitudinal diffusion through the plane surfacesat
positions (i - 0.5) AZ and (i + 0.5) AZ.
ON THE LONGITUDINAL AXIS,
(9.51)
At.
radial
longitudinal
Sec. 9.41
Solid cylinder
of finite length
185
V*[CNi,o-Ci.01
As the volume V andthe areasof the surfaceare :
iI
Aif0.5,0
= n 4
Ai.o.5=
2rc . E2 * AZ
I I
the new concentrationCNi, o is thus obtained as a function of the previous concentrations

at the sameand adjacentposition :
bzI
-D%
dr
hICN,-C,,)
D
r
Q-JC
i-05 'I
di!
Fig. 9.9. Annulus next to the cylindrical surface of thicknessAr / 2 and length k.
186
Numerical
(9.53)
CNi.0 =
[Ch. 9
ci,o+~~ci.I-ci.o)+~[ci-l,o-2Ci,o+Ci+~~*]
r
with the dimensionlessnumbersMr andM,

and i
N, ,
The annulus next to the external surface, of thickness Ar / 2 and length Ah, is
considered (Fig. 9.9). The matter balance during the increment of time At is evaluated
within this annulus by taking into account the radial and longitudinal diffusion and the
radial transferwith a finite coefficient of mattertransferh
ONTHE CYLINDRICALSURFACE,
(9.54) At -D
[(
j =
i-OS,Nr-0.*5
+ D (84
Vi,Nr-0.25.
Ii+0.5.Nr-0.*5-D
CNi.Nr-0.2S-
Ci,Nr-0.25
( A. ZC I i,Nr-0.5-
h. %Nr(Ci.Nr-
Gq
As the values of the volume and areasof the surfacesare :

(9.55)
Vi.Nr-0.25 = 2n(Nr-0.25).
= 2
Aif0.J,Nr-0.2S
A.z,Nr-0.5
A,.$.
AZ
TC(Nr - 0.25). Ar. $
= 2 x (Nr - 0.5). Ar. AZ
Ai.Nr = 2 IT.Nr. Ar. AZ

the new concentrationwithin this a~ulus is given by :
(9.56)
CNi.~r-0.25
ci,Nr-0.25+k
+2 Nr-0.5
Nr-0.25
1
Mr
Ci-1.Nr-0.25-2Ci,Nr-0.25+Ci+l.Nr-0.25
Sec. 9.41
Solid cylinder
of finite length
187
Nz ______ _
?i
Nz-0.5________
j - 0.5
4
I.
jtO5
D
J
hr
Fig. 9.10. A~ulus next to the plane surfaceof radius j. Ar, and thicknessesAr
andAzf2.
With an infinite coefficient of matter transfer on the surfaceh, the concentrationon
the surfaceis constant:
(9.57)
Ci, Nr = Ceq
with i = Nz ,
and j
The annulus next to the plane surface i = Nz, of radius j. Ar, thickness Ar and AZ /
2 is considered (Fig. 9.10). The matter balance during the increment of time At is
evaluatedwithin this annuluswith the radial and longitudinal diffusion and the coefficient
of matter transferon the plane surface.
ON THE PLANE SURFACE,
IA* % INz-0.25.j
+ At.
= V. I CNNz-O.25,j- CNZ
-0.25.j
+0.5
Numerical
188
analysis
[Ch. 9
with cylinders
The values of the volume and areas are :
ANz-0.2S.j +O.S = 2rr(j+0.5)Ar.F
(9.59)
ANz-0.2S,j
-0.5
ANz-o.5,j
v=
2rr(j-0.5)Ar.T
= AN,,j = 2 Tc.j. Ar. Ar
2rr. j.Ar.Ar.2
and the new concentration within this annulus is thus given by :

CNN,-o.25,j
(9.60)
1
+2j.Mr
cNz-0.25,jf&
~Nz-0.25,j-1-2CNz-0.25.j+CNz-0.25,j+1
1 1
CNz-0.25,j+1-CNz-0.25.j-1
-- 2h. At c
AZ
+&
cNz-l.j-cNzqj]
NZ.J .-%I
With the infinite value of the coefficient of matter transfer, the concentration on the
surface is :
(9.61)
cNz, j
= ceq
PLANE SURFACE AT THE CENTRE, i = Nz
and
j=O
The small cylinder with the same axis as the cylinder, located next to the plane
surface, of radius Ar / 2 and thickness AZ / 2, is drawn (Fig. 9.11).
The matter balance during the increment of time At by considering the radial and
longitudinal diffusion and the coefficient of matter transfer h through the plane surface is
written as follows :
Solid cylinder
Sec. 9.41
of finite length
j=O
Fig. 9.11. Small cylinder next to the plane surfacewith the sameaxis asthe
cylinder, of radius Ar / 2 and thicknessAZ / 2.
= v.I
CNNz-0.25.0-
CNz-0.25,0
With the valuesof the volume and areas:
(9.63)
ANZ-0.25.0.5
= 2n
the new concentrationwithin this small cylinder is thus given by :

(9.64)
CNNZ-0.25.0 =
2
CNz-O.ZS.O+&
CNz-0.25.
1-
cNz
-0.25.0
189
190
Numerical
analysis
with
CNz-l,O-CNz,O
cylinders
1
--
ICh. 9
2h. At
CNz,O-
&]
AZ
with i = Nz
and j = Nr
The annulus next to the comer of the cylinder, of thicknessesAr / 2 and AZ / 2, is
drawn in Fig. 9.12.
The matter balanceduring the increment of time At is evaluatedby considering the
longitudinal andradial diffusion and the matter transferthrough the radial and longitudinal
surface:
CORNER,
+At.-D. A.$
Nz-0.5sNr-0.25-AN~,~r-0.25
~.(CN~,N~-O.~~-
11)
= v.
~NNz-o.~~,N~-o.~Y
CNz-0.25,Nr-0.25
A h(C,,-C,,)
LDdc
dz
Nz-0.5*
Nr
w-05
ar
Fig. 9.12. Annulus next to the comer of thicknessAr / 2 and AZ / 2.
Cd
191
Solid cylinder of finite length
sec. 9.41
With the following valuesof the volume andthe areasof the various surfaces
(9.66)
= 2 rc(Nr - 0.25) Ar. $
ANZ-o.D,N~-0.25
ANZ-o.z,~r
ANZ-os,~r-0.25
2rrNr.Ar.g
ANz.Nr-0.25
2 II (Nr - 0.25) Ar. $
V = 2~(Nr-0.25)Ar.~.g
the new concentrationwithin this annulusis thus expressedby :
2Nr-1
(9.67)
CN~z-o.x.~r-o.25
-CNz-0.25.Nr
1
-
CNz-0.25,Nr-0.25+
Nr - 0.25
2 Nr
* =[CNz-O.ZS,NrNr-0.25
Ar
. E
&]
I-,[
+~[C~.-l,N~-0.25-~N1,Nr-0.25
CNz-0.25.Nr-
CNz.Nr-0.25-
&]
ASSUMPTIONS FOR CALCULTION ON THE SURFACES
Some assumptions must be made for evaluating the concentrations in positions

which arenot associatedwith integers.
Only a few examples are given, becausethe others can be followed in the same
manner.
On the cylindrical surface,with j = Nr, and i, the simple assumptions
(9.68)
CNi.
(9.69)
Ci-l.Nr-0.25
Nr - 0.25 -
ci. Nr - 0.25 =
ci-
1. Nr
cNi,
Nr - ci. Nr
192
Numerical
[Ch. 9
can be made.
The more accurateassumptions
(9.71)
Ci-l,Nr-0.25
43Ci-I,Nr+~Ci-l.Nr-l
give more complex solutions.

AMOUNT OF SUBSTANCE LOCATED IN THE CYLINDER OF FINITE LENGTH
The amount of substancelocated at time t in the cylinder of finite length is obtained

by integratingthe concentrationat this time with respectto space.
(9.72)
M, = 2 TC
dr* dZ
With the finite differences,the summationis made in eachcylinder of thicknessAh,

by using the following equation :
2
(9.73)
Mt,i = TZ
2 Nr-1
. AZ. Ci,o+ 2 Ir(Ar) . C
j. Ci,j
+2x(Nr-0.25)w2
ZC.
+C.
2 . I4 1.Nr 4
l,Nr-
andthe total amountis M,
(9.74)
M, = 2 ~
M,,i
i=O
CONDllTlONS OF STABILITY
Some conditions of stability for calculation exist. They are obtained by writing that
the coefficient of the previous concentrationCi,j or Ci, o is positive.
Solid cylinder of finite length
Sec. 9.41
193
Within the cylinder, it is thus obtained:

1 2+1
M, M,
(9.75)
On the axis of the cylinder

(9.76)
4
2
1 > M,+K
9.4.2
DIFFUSIVITY
The calculation is made in the same manner as for constant diffusivity, by using
similar matter balances with a difference due to the concentration-dependenceof the
diffusivity, and similar Figures.
The following two functions areused :
(9.77)
I~,j = j. Di, j. (Ci, j - 0.5- i, j + 0.5)
(9.78)
Hi, j = Di, j* (Ci -
0.5,
j - Ci + 0.5, j)
The sameplacesare studiedin succession:
r=j.Ar
and z = i. AZ
with
The small annulus of radius r = j. Ar and thickness Ar and AZ shown in Fig. 9.7,
enablesone to evaluatethe mattertransferduring the incrementof time At
and the new concentrationafter elapseof time At, CQ is thus given by :
194
Numerical
(9.80)
CNi,j = Ci,j+p
[Ch. 9
At
-o.s,j-
Hi +0.5.j
j. (AI-)~
and z = i.Az
with r = 0
The matter balanceduring the incrementof time At is evaluatedin the small cylinder
of radius Ar / 2 and length AZ shown in Fig. 9.8.
ON THE LONGITUDINAL AXIS,
The new concentrationafter elapseof time At, CN,.o, is thus given by :

8. At I,
CNi.0 = Ci.0 - - 2 1*o.5
(9.82)
+ [Hi
(Azl2
-0.5.0-
Hi +O.S.O]
br)
and i
j = Nr
with
The matter balanceduring the increment of tune At is calculatedin the annulusnext
to the cyhndrical surfaceof thicknessAr / 2 and length Ah, shown in Fig. 9.9.
ON THE CYLINDRICAL SURFACE,
(9.83)
+iDA~)i+0.5.Nr-0.25
i,Nr-0.5-
= Vi,Nr-0.25
h*Ai,Nr
CNi,Nr-0.25
i Ci.Nr-
At
f&j At
- Ci.Nr-0.25]
The new concentrationafter elapseof tune At within this annulusis thus given by
(9.84)
CNi,Nr-0.25
= Ci.Nr-0.25+
Hi-O.5.Nr-o.25-Hi
+0.5,Nr-0.25
Solid cylinder
sec. 9.41
+
2Nr
. ~(Ci,NrNr-0.25
Ar
2
. - At I i,Nr-O.SNr - 0.25 tArjz
195
of finite length
c~)
and j
i = Nz
with
The matter balanceduring the increment of time At is evaluatedwithin the annulus
next to the plane surfaceshown in Fig. 9.10.
ON THE PLANE SURFACE,
(9.85)
-.*025
. .
j-05+(D*A'$)Nz-025
j+05'At
Nz-0.5sj- h. ANz.f (CNz.i -C&At
- CNz-0.25.
= V.ICNNz-0.25.j
The new concentrationafter elapseof time At within this annulusis thus given by
(9.86)
CNN,-O.25.j = CNz-0.25. j ~
At
2 1Nz-0.25,j-0.5-
INz-0.25.j
+O.5I
j. br)
+ 2At
-HNz-O.S,j
iAd2
-- 2h. At CNz.hAZ [
ON THE PLANE SURFACE AT THE CENTRE
From the matter balanceevaluatedduring the increment of time At within the small
cylinder next to the plane surface, shown in Fig. 9.11, the new concentration is obtained
as follows :
(9.87)
~NN~-o.~s,o
= CNz-0.25.0
-qINz-0.25.0.5
b-1
+ 2At
-HNz-0.5.0
(AZ)
-~
2h. At
AZ
CNz.O-
'cq]
196
Numerical
with
ON THE CORNER,
[Ch. 9
i = Nz
and
j = Nr
From the matter balance during the incremental tune At made within the annulus next
to the comer (Fig. 9.12), the new concentration within this annulus is thus drawn :
(9.88)
CNN, - 0.25. Nr - 0.25 = cNz - 0.25, Nr - 0.25
At
. -+Nz-0.25.N~0.25Nr - 0.25
(Ad
2
2. At
HNz-0.5,Nr-0.25
-- 2h.At
AZ
2Nr
At. h
CNz-0.25,Nr-
Nr-0.257
[ CNz,Nr-0.25
-%I
ASSUMPTIONS FOR CALCULATION ON THE SURFACES
The same assumptions as those made for constant diffusivity enable one to obtain the
new concentration on the surface and in positions associated with integer values of i and j.
AMOUNT OF SUBSTANCE LOCATED IN THE CYLINDER
Calculation is made in the same manner as for a constant diffusivity.

The conditions of stability are obtained by writing that the coefficient of the previous
concentration in the position considered is positive. Values identical as those obtained for
constant diffusivity can be used.
9.5 CONCLUSIONS
The cylinder is an important case, and the hollow long cylinder and the cylinder of
finite length are of high concern.
HOLLOW CYLINDER OF FINITE LENGTH
The hollow cylinder of finite length can also be studied. Calculation can be easily
made by considering the results obtained with the solid cylinder of finite length. The
equations are the same in all places (e.g., within the solid, on the external cylindrical
surface, on the plane surfaces) except on the internal cylindrical surface.
Sec. 9.51
Conclusions
197
In the caseof a very high coefficient of matter transfer,the solution is obvious asthe
concentrationis constanteverywhereon the surface.
In the caseof a finite coefficient of matter transfer,the concentrationon the internal
cylindrical surfacecan be obtainedby consideringthe way of calculation describedfor the
hollow cylinder of infinite length and for the solid cylinder of finite length.
CYLINDER OF FINITE LENGTH LOCATED WITHIN ANOTHER CYLINDER OF FINITE LENGTH
This is the case of a cylindrical envelope. The problem is tedious and rather
complex. A solution is obtained by using the same method as for the transfer of the
substancebetweentwo different sheets(section 6.5).
10
Drug delivery from dosage forms
consisting of a drug dispersed in a
non-erodible polymer
10.1
INTRODUCTION
AND
DEFINITIONS
The main purpose in this introduction is to give general information on the problem
of drug delivery from conventional dosage forms, by following the drug through the
body. It seems useful to define certain important biopharmaceutical concepts, because it is
said sometimes that very few of these concepts are clearly defined at present. In fact there
are no sharp demarcations between pharmacology and pharmaceutical chemistry and an
interdependence exists in the development of these disciplines.
From this knowledge, the interest of controlled drug delivery appears. A literature survey
is thus made for these kinds of dosage forms in order to point out the essential results and
the most important facts which must be considered in the future.
10.1.1 PROBLEMS OF THE DRUG PASSING THROUGH THE BODY

A general diagram (Fig. 10.1) shows the path followed by the drug from absorption
to excretion from the organism. The following steps are of importance, and they are
shortly described :
- Drugs and their supply forms
- Liberation of the drug
- Absorption of the drug through the membranes of the organism, and problem of
200
Drug delivery
from dosage forms
[Ch. 10
Fig. 10.1. Path followed by the drug from a dosageform orally absorbedthrough
the organism.
bioavailability
- Distribution of the drug into compartments
- Pharmaceuticalaction andmetabolism
- Elimination of the drug out of the organism
DRUG AND THEIR SUPPLY FORMS
The term drug is used by physicians to represent the active agent of a dosage
form, as well as the dosageform itself. The World Health Organization defines the drug
as any substanceused to modify physiological systems or pathological states for the
benefit of the recipient. In scientific literature the word drug generally means a
biologically active substance.
A drug, or active material, is thus a chemical compound, obtained synthetically or
naturally, which developsa reciprocal interactionwith the organism.
The dosageform, or supply form of presentation,is the complete pharmaceutical
Sec. 10.11
Introduction
and definitions
201
preparation. It consists of active agents and excipients which are auxiliary substances.
Excipients are biologically inert, and they are used for many purposes. They are
commonly used for binding, disintegrating or even solubilizing the drug, and thus they
bring a consistency,potency and volume of supply form suitablefor the use of the dosage
form. Galenic pharmacy is the science of formulation of dosage forms by combining
drugs and excipients.
LIBERATION OF THE DRUG
The delivery or releaseof the drug from the dosageform is representedby the term
liberation. This is the initial step of interaction between the dosage form and the
organism, and it is thus of great practical importance, Depending on the formulation, the
therapeuticvalue of a drug may be reducedto a swell extent.
Two factors determine the releaseof the drug : solubihty and rate of dissolution.
Before being absorbed,the drug must be in solution. The distribution partition of the drug
between hydrophilic and lipophilic phases is an important parameter. The rate of
dissolution is also a main parameter.The rate of releaseof the drug from the dosageform
can be the rate-limiting factor in absorptionwhen the rate of dissolution is slower than the
rate of absorptionof the drug.
ABSORPTION OF THE DRUG
The drug dissolved in a liquid is thus absorbedand enters the bloodstream or the
lymphatic systemof the body. The drug must thus passthrough various barriers, such as
skin, mucous membrane. Additional substancesmay be used in the formulation of the
dosageform to enhanceabsorption.Absorption may vary with dosageforms, depending
on the pH in the stomach and upper intestinal tract which ranges from 1.2 to 8.
Absorption of a drug occursalong the gastrointestinaltract, and essentiallyin the intestine,
and dependson the pH, the degree of filling of the intestine, enzymatic activation and
bacterial decomposition.The drug has to passthrough severalmembranes,the membrane
of the epithelial cells, the basementmembrane,before reachingthe connectivetissuesand
finally the capillaries. The drug is thus transportedvia the blood-streamto the portal vein
or via lymph to the thoracic duct. Penetrationof drug within epithelial cell membranesis
the rate-limiting factor.
MENBFWNES AND TRANSPORT THROUGH THE MENBRANE
During the stage of absorption, the drug must pass through several barriers or
202
[Ch. 10
membranes.These membranesform barriers or interfaces delimiting functional entities.

Generally, a membraneis a complex structurecomposedof lipid andprotein molecules.
The driving force leading to transmembranemovement of drugs is generally the
gradient of concentration of the drug through the membrane. The process of drug
transport through a membrane is thus controlled by diffusion. The rate of diffusion
dependson the gradient of concentrationthrough the membrane,and on the diffusivity ; it
is thus related, through the diffusivity and concentration, to the nature of the membrane
and drug. The time necessaryfor the drug to diffuse through the membranewould then be
proportional to the squareof the thicknessof the membrane.
BlOAVAlLABlLlTY OF THE DRUG
The bioavailability is referred to the amount of drug which is absorbed by the

organism in a given time and which reachesthe target site. It is the percentageof drug
absorbedfrom a given dosageform. Bioavailability is the most important criterion. It is
often defined by the ratio of the amountof unchangeddrug reachingthe circulatory system
after administration of a test dose to the amount obtained for a standard dosage form
administered under the same conditions. Bioavailability is determined from the drug
concentrationin blood and in urine.
DISTRIBUTION
An organism consistsof fluid and cells. The body fluid is sometimesdivided into
threecompartments: intravascular,interstitial or extracellular,intracellular compartments.
The drug which has entered the circulation system in the bloodstream is thus
distributed between extracellular and intracellular compartments. This stage is of
importanceas most receptorsfor drug are located in the tissues.
ELIMINATION
The drug is inactivatedessentiallyby two processes:

- metabolismor chemicalalterationof the molecule.
- excretion via various organs.
The biological half-time representsthe time required for elimination of half the
plasmacontentof the drug either by excretionor by metabolism.
Metabolism is the sum of all chemicalreactionsinvolved in the biotransformationof
exogenous and endogenous substanceswhich occur in the living cells of people and
animals.The drug may be broken into simpler substances(catabolism), or it may be used
Sec. 10.11
Introduction
and definitions
203
for the synthesis of new materials (anabolism). The presence of enzymes is required.
Metabolism may occur in the blood, or more often in tissuesor in an organ where many
various enzymesare availablefor the reactionprocess.Generally,the metabolitesaremore
watersoluble than the original drug, and this transformation facilitates the excretion
through the urine.
The major route of excretion for the majority of substances, e.g., drug or its
metabolites, is through the kidneys. Only a few substances,such as gaseousanesthetics,
are excretedin appreciableamountsvia the lungs. The rate and extent of renal excretion is
controlled by glomerular filtration, tubular reabsorption and tubular secretion. When
excretion is slow, there is a risk of accumulation leading to an important increasein the
plasmalevel of the drug.
THERAPEUTIC INDEX
The optimal dose dependson the patient, and it is difficult to define the optimal
standarddosefor the averagepatient. However, it is useful to introduce the conceptof the
therapeuticindex TI.
Ehrlich defined TI as the relationship between the minimal curative dose and the
maximal tolerated dose.In pharmacology,TI is the ratio betweenthe median lethal dose,
which causesthe deathsof 50 % of experimental animals, and the median effective dose,
which is effective in 50 % of cases.
PHARMACOKINETICS,
10.1.2
BIOPHARMACEUTICS
PHARMACODYNAMICS
AND
PHARMACOKINETICS
Pharmacokineticsdeals with what the body does to the drug ; it is thus concerned
with the kinetics of absorption, distribution, metabolism and excretion of drugs or
metabolitesproducedby the body.
Phannacokineticshas two basicpurposes:
(i) it deals with the relationship between the quantity of a drug administered to an
organism and its concentration in plasma and the body fluid which is next to the
therapeutictargetregion.
(ii) it dealswith the drug concentrationin body fluids immediately after administrationof a
drug, and its decline over a given period of time. Drug delivery from a conventional
dosageform into the circulation systemfollows a typical pattern,the plasma concentration
rises discontinually, reachesa peak of concentration,and following complete absorption,
204
Drug delivery
[Ch. 10
from dosage forms
falls gradually to zero. Blood levels are important in determining a drugs tune-effect
curve. The effective action of the drug begins when a given minimum concentration has
been attained at the target site.
Most drugs are absorbed or eliminated by following a first-order reaction. The
delivery of a conventional dosage form is generally defined by a first- or second-order
kinetics. The liberation of a drug through a continuous therapeutic system as perfusion
follows a zero-order reaction with a constant rate.
The purpose is thus to know what happens to a drug after its release from the dosage
form in the body.
PHARMACODYNAMICS
Pharmacodynamics is concerned with the reactions of the biosystem to the drug, or

in short, with what the drug does to the body. The purpose of Pharmacodynamics is thus
to know what happens to the organism under the influence of the drug.
BIOPHARMACEUTICS
First advanced by Levy in 1958, the concept of biophannaceutics has been

differently interpreted according to the various cases : The study of factors that affect the
extent and speed of absorption or liberation of a drug from its various physicochemical
forms.. .I. Biophannaceuticals deals essentially with the optimal or desired availability of
a drug from the dosage form employed for the target site, ie, the influence of drug
formulation on the biological activity of the drug.
Being the study of the influence of formulation on the therapeutic efficacy of a drug,
various factors in animals or people which determine the relationship between the nature
and intensity of biological effects are considered :
- Type of drug, its physical state, and size and shape.
- Additional substances.
- Nature of supply form in which the drug is located.
Biopharmaceutics was considered as a vital branch of pharmaceutical research in the
early eighties when the FDA (Food and Drug Administration)
began to demand
biopharmaceutical documentation for the introduction of new therapeutic preparations in

the USA.
As a result of biopharmaceutical
research, it was clearly
shown that a
Sec. 10.11
Introduction
and definitions
205
pharmacologically active substance is not necessary an effective remedy, and that the
efficacy may highly depend upon the method of delivery into the organism.
10.1.3
CONVENTIONAL
DOSAGE
FORMS
The measure of the efficacy of a drug can be determined by the reduction of

symptoms of disease, accompanied by minimal complaints and side effects.
LIMITS OF ACTION
Except for continuous intravenous infusion which delivers the drug at a constant rate
(or a zero-order reaction), all other conventional dosage forms release drugs through a
typical complex concentration-time history. Initially the drug is very rapidly liberated from
its dosage form depending on the rate of disintegration and solution, and thus quickly
builds up to a high concentration. Because of the absorption which can be described by a
first-order process, the drug concentration then falls exponentially with time until the next
dose. As a result, undulating concentration patterns of the drug are observed either in the
blood or in tissues, where high concentrations alternate with low concentrations producing
an alternation of overdosing and underdosing of the drug. (Fig. 10.2). Of course, the
concentration of the drug in an organism cannot be constant, as the rate of the
discontinuous administration is not the same as the rate of the continuous elimination and
excretion.
Conventional dosage forms often cause problems to the patient because they
maintain therapeutic drug levels for only brief durations. This gives rise to : sharp
fluctuations of drug levels in plasma and in tissue, and a high frequency of administration.
The fluctuating drug levels in blood and tissues obtained with conventional dosage
forms lead to an excessive use of the drug and to an insufficient
influence on the
mechanisms of the disease. Moreover, initial overdosing is responsible for a high

frequency of side effects, leading to other illness or damage.
PATIENT COMPLIANCE
Investigations conducted in many countries, and especially in the USA and Europe,
demonstrate that around 50 7%of patients do not comply with the regimen.
Six factors are distinguished in noncompliance : simple omission, wrong dosage and
wrong frequency, impermissible delay, execution in the wrong mode and at the wrong
time.
The percentage varies from disease to disease. Generally, patients have far more
206
Drug delivery
from dosage forms
Overdosaqe
I
)i
[Ch. 10
Overdosage
I-
1
l
Time
hdcrdos-aqe
Fig, 10.2.
Concentration of drug in the gastrointestinal liquids and blood or tissues
as a function of time, with a conventional dosage forms.
difficulty in complying with treatment of a disease causing no sensible or perceptible

complaint, such as high blood pressure. Moreover, patients do not appreciate the necessity
of therapy for a disease, serious but causing no complaints, when the therapy causes
disturbing side effects. The cooperation of patients decreases continuously with the
duration of treatment and with the complexity of the regimen.
Therapeutic systems able to deliver the drug over a prolonged period of time offer
the opportunity of improving patient compliance by reducing some problems due to
conventional dosage forms. Generally, it is said that reducing the frequency of drug
administration substantially improves compliance. And increasing patient compliance may
reduce the undesirable side effects of the drug.
10.1.4
ORAL
CONCEPT
THERAPEUTIC
SYSTEMS
AND MAIN PURPOSE
Therapeutic systems must release the drug for a certain period of time at a constant
rate, and offers important advantages overconventional
dosage forms, especially in
Introduction
Sec. 10.11
and definitions
207
diseasesrequiring a constant concentration of the drug in the blood over prolonged

duration of therapy.The drug is thus brought at a constantrate to the target organ through
the blood, and uniform concentrations of drug are attained in this organ. These dosage
forms can thus decreasethe total daily amount of drug administered, and in so doing
reducethe importanceof side effects,andimprovesthe patient compliance.
The concept,its practical applications,and the term itself are given by Zaffaroni (1) :
A therapeuticsystemis a drug-containingpreparationor dosageform that releasesone or
more drugs continuously in a predetermined pattern for a fixed period of time, either
systemicallyor to a specifiedtarget organ.
Systemictreatmentis often administeredvia the gastrointestinaltract becausemany
dosageforms are easyto swallow. The principle of retardation consistsin slowing down
the process of dissolution of the drug or delaying it for a period of time, by coating the
drug with polymers having a low solubility in water, the retardationin drug delivery being
prolonged by increasingthe thicknessof the coating.
MEMBRANE-CONTROLLED
SYSTEM
The oral osmotic therapeuticsystem(OROS) consistsof a drug reservoir surrounded

by a semi-permeablemembranewith an opening in it (2-5). Becauseof the difference in
concentration on the two faces of the membrane, water from the gastrointestinal tract
diffuses through the semi-permeablemembrane.This water dilutes the drug or dissolvesit
if the drug is in solid state.The influx of water causesan increasein the pressure(osmotic
pressure)which producesa releaseof the drug in solution through the delivery opening in
the membrane.
The rate of releaseis related to the solubility of the drug. The osmotic systemis thus
only suitable for drugs having specific properties which ensure an adequate osmotic
pressure. Only a few polymers are able to resist mechanical stress and the digestive
liquids, ensuring a constantvolume for the system. Cellulose acetatehas been the most
suitablepolymer for the membrane.
The delivery opening has to comply with strict specification for the accuracyof the
rate of delivery. CO2 lasers are able to bore a hole of 100 - 250 microns in diameter
through only one part of the membrane.
The OROS systemis suitableas a therapeuticdosageform for many substances.The
rate of drug release can be adapted to the required therapy by varying the following
parameters: the surface,thicknessand specific permeability of the membrane.
Someexamplesaregiven :
Drug delivery
208
from dosage forms
[Ch. 10
- Hypertension with beta-blockers, as metropolos (Lopresor, from Ciba-Geigy) and

oxprenolol (Trasicor, Ciba-Geigy).
- Rheumatism, with indomethacin (Indos, from Merck Sharp and Dohrme).
- Glaucoma, with acetazolamide.
CONCLUSION
Oral therapeutic systems of the OROS type are very useful. Two drawbacks exist :
they can be used only when the drug is rather soluble in water. Moreover, they are not a
simple dosage form like a pill but a rather complex system.
10.1.5
SIMPLE
MONOLITHIC
DEVICES
WITH
A POLYMER
MATRIX
MAIN PRINCIPLES
Various devices have been studied and built up in order to control the release of the
drug in gastric liquid. They can be divided into four categories by considering the
mechanism which controls the release of the drug from the dosage form (6 - 8) :
- Osmosis, already seen with OROS systems.
- Polymer erosion. Two systems can be considered : the one where the drug is dispersed
in the erodible polymer, the other where the drug is surrounded by the erodible polymer.
- Diffusion and reaction, when the drug is chemically bound to a polymer playing the role
of a backbone.
- Diffusion, when the drug is dispersed in a non-erodible polymer playing the role of a
consistent matrix.
However, it appears that for each device the release of the drug can be controlled by
more than one mechanism (7). For instance, in the case of a device with an erodible
polymer, diffusion of the liquid through the polymer is generally the first step, followed
by swelling and polymer erosion. The diffusion of the drug through the liquid located in
the polymer generally takes place as well as polymer erosion.
SIMPLE MONOLITHIC DEVICES
Special attention has been given to finding a way for regulating the rate of release of
drug by means of monolithic devices where the drug in solid state is dispersed in an inert
polymer playing the role of a consistent matrix. The polymer must be biocompatible and
pass through the body without any change. Both biodegradable and non-biodegradable
polymers have been used for the matrix (9 - 12).
Introduction
Sec. 10.11
and definitions
209
THEORETICAL STUDIES ON THE TRANSFER OF THE DRUG.

DIFFUSIONAL TRANSPORT OF THE DRUG
Severaltheories have beenput forward in order to describethe processof releaseof

the drug from the inert polymer matrix. Sometimes they consider a combination of
diffusion effects and of hydrodynamic effects, the pores in the device playing a
fundamentalrole (13).
Very often, experimentaldatahave been obtainedonly for the earlier period of time,
and the diffusional processhas been proved by the squareroot law of time dependence
with the amount of drug transported.The following equationis often usedfor a thin
tablet:
0.5
(10.1) 2
= ; F
00
i i
where Mt is the amountof drug releasedin time t,

andM, the correspondingamountafter infinite time,
and D is the diffussivity and L the thicknessof the sheet.
Of course,this law can be used only when M, / M, < 0.5. (Chapter 2).
When long experiments have been carried out, the following equation has been
successfullyusedto describethe processof releasefrom a tablet of thicknessL.
(10.2) ;
Mt = .
00
x
c1
-
,,i+
lf
exp[-2n~~2Dt]
As shown in Chapter 2, this equation can be used, only when the diffusivity is
constant,and when one transport of matter is considered.
Another fact is worth noting. In all the studies, only the transport of the drug is
considered, in spite of the fact that the drug is often in solid state and necessitatesthe
presenceof a liquid.
Various studieshave been carried out with devicesmade of a soluble drug dispersed
in a polymer. A few examplesonly are given :
- Delivery of KC1 dispersedin ethylcellulose (14). An interesting review of the previous
theoretical studiesis shown in this paper. According to the authors,zero order kinetics is
observed in the latest stage of the process when M, / M, > 0.7. In fact, the kinetics
obtained are perfectly described by a diffusional process with a vertical tangent at the
210
Drug delivery
from dosage forms
[Ch. 10
origin, and a rate of delivery which decreasesexponentiallywith time.

- Various drugs and polymers, dispersedin biodegradablepolymers (9). Many results are
given in this important article. Amongst them, cyclazocine releasedfrom compositesof
poly (lactic acid) film, releaseof naltrexone.
- Sodium salicylate dispersed in a (hydroxyethyl) methylcellulose matrix (12). The
kinetics of release are determined from various tablets having a concentration of drug
ranging from 10 to 30 %
- KC1 dispersedin (hydroxypropyl) methylcellulosematrix (15). The kinetics of releaseof
KC1 in dissolution medium is determined, and described by a diffusional process with
constant diffusivity. In this paper, an important swelling is observed, due to the water
which enters the polymer. It is also seenthat the diffusivity of the drug in the swollen
polymer is a function of concentrationof water in the polymer. However, due to the high
water content in the polymer a constant average diffusivity is only considered. The
profiles of concentrationof KC1 andwater are determinedat various times.
- Diphenhydramine HCl, acetaminophen,dyphylline and theophylline are successively
dispersedin copolymers of polyvinyl acetateand polyvinyl chloride, with other additives
and calcium phosphatetribasic. (11).
- Diphylline dispersed in polyvinyl phthalate acetate,with other additives, as well as
dicalcium phosphate.(16). Various parametersare also studied, such as the hardnessof
the dosageform compressedin various ways, the concentrationof polymer. In eachcase,
diffusional curves are obtainedfor the releaseof the drug.
- Theophylline dispersed in poly (2 - hydroxyethyl methacrylate - co - methyl
methacrylate)copolymers (17). The kinetics of releaseof the drug is determined, aswell
as the kinetics of swelling due to the water which enters the polymer. No correlation is
establishedbetweenthe diffusivity of the drug andthe water concentrationin the polymer.
- Trinitrine dispersedin microbeadsand surroundedby a film of Eudragit E 30 D and L 30
D (copolymer of dirnethylaminoethylacrylateand ethyhnethacrylate)(18). Various kinetics
of releaseare observed,some of them describedby a Fickian processof diffusion, other
with a constant rate. But the dosageforms are rather complex, since the drug is firstly
dispersedin microbeadsand thesemicrobeadsare thus surroundedby a film of polymer.
This is an interesting study leading to the preparationof dosageforms with a constantrate
of drug delivery.
- Drugs dispersedin biodegradablepolymers (gelatin andpolyethylacrylate) (19).
- Various drugs or biocides such as trichlorophenol, salicylic acid, codeine, are dispersed
in crosslinked dextran or modified crosslinkeddextran. Some of thesesamplesare coated
Sec. 10.11
Introduction
and definitions
211
by acetylchitin, poly (vinyl acetate) or Eudragit. Various kinetics of release are obtained
depending on the drug and the presence or not of a coating (20).
- Piretanide dispersed in gelatin matrix surrounded or not by a layer of formalin (21).
Depending on the presence or not of the thin film surrounding the bead, various kinetics
are obtained.
- Isoniazid, rifampicin and novocain dispersed in polyurethane foams (22). Diffusional
kinetics of release are obtained in various liquids.
- Polypeptides and other macromolecules dispersed in an ethylene-vinyl acetate copolymer
(23). The slabs release the macromolecule in 0.9 5%NaCl solution, and the kinetics of
release follows a diffusional process.
- Chloramphenicol dispersed in 2 - hydroxyethyl methacrylate (24). The drug is released
in pH 7 phosphate buffer, by following a diffusional process.
- Vincamine dispersed in hydrophilic or lipidic polymers (25).
- Chlorpheniramine dispersed in Methocel K 4 M matrix tablets and hydrogels (26). The
effect of additives such as sodium alginate, sodium bicarbonate or calcium phosphate on
the kinetics of release is also studied (27).
- Theophylhne and proxyphylline dispersed in copolymers of N-vinyl- 2-pyrrolidone and
2 - hydroxyethyl methacrylate (28). Swelling-controlled
release systems are studied,
giving kinetics which are widely controlled by Fickian diffusion.

PSEUDO STEADY-STATE CONDITIONS
A complex but interesting theory, based on the existence of a pseudo steady-state

condition, has been built up. This fact may be observed with polymers in glassy state
below their transition temperature T,.
A sharp linear concentration of liquid is formed moving through the polymer at an
almost constant rate. On the sides of this moving boundary are the extracted part and the
untouched portion of the device (14,29 - 3 1).
In all cases, the following drastic assumptions are made :
- only the transport of the drug is considered.
- the diffusivity of the drug is constant.
In fact, this interesting case can be explained by the presence of the liquid in the
polymer. The liquid diffuses within the glassy polymer with a rather low and constant rate
of transport and with a low concentration of liquid. The polymer, initially in glassy state,
becomes less and less glassy and more and more rubbery, when the concentration of
liquid in the polymer increases. A Fickian diffusion of liquid with a higher concentration
212
Drug delivery
from dosage forms
[Ch. 10
is thus superimposedon the non-Fickian diffusion in the rubbery part of the polymer.
RELEASE OF DRUG WITH A CONSTANT RATE
A remarkable and surprising result was obtained for the transfer of dyestuff RedSudan from polystyrene into n-hexane, when the dyed polymer was immersed in the
liquid. The rate of releaseof the dye in the liquid is constantduring the greaterpart of the
process. This rate is also a function of the initial dye content in polymer and of the
thicknessof the polymer fiim (32). Of course,many attemptshave been made in order to
extendthis exampleto the delivery of drug with a constantrate (33).
Based on the pseudo steady-stateas well as on the rate of dissolution, especially
when it is low, calculations were made in order to find a kinetics of delivery with a
constantrate (14). In fact, the kinetics obtainedin this casefor the drug delivery follows a
diffusional process,with perhapsa variation of the diffusivity with time or rather with the
concentrationof the drug in the dosageform.
As a matter of conclusion, it was shown by various authors :
(i) it is difficult if not impossible to obtain a constantrate for the drug delivery by using
deviceswhere the drug is dispersedin a polymer (34).
(ii) a constant rate of drug delivery can be obtained with a simple drug-polymer device
only after sufficient knowledge on the processof diffusion and polymer swelling (33).
SWELLING OF THE POLYMER MATRIX
A few studies of interest have shown that a swelling of the polymer matrix takes
place when the drug-polymer device is immersed in the liquid, this swelling being due to
the fact that the rate of transport of the liquid in polymer is higher than the rate of drug
release (17). The rate of transport of liquid determined in the case of KC1 dispersedin
hydroxypropylmethyl-cellulose follows a diffusional process (15). A dimensionless
number, Deborahsnumber was introduced with interest (33).
10.1.6 PROCESSES WITH A DOUBLE MATTER TRANSFER

Experiments made with various materials, such asplasticized PVC, elastomersand
dosageforms, have shown that two matter transportstake place simultaneously,when the
material is in contactwith the liquid : the liquid entersthe solid, while the plasticizer of the
plasticizedPVC or the drug of the dosageform leavesthe solid.
Sec. 10.11
Introduction
and definitions
213
PIASTICIZED PVC IN CONTACT WITH A LIQUID
Plasticized PVC are widely used for packagings for various liquids (food, solvent,
blood) or for tubings through which the patients blood is circulated during the
hemodialysis treatment.
When a plasticized PVC with an extent of plasticizer ranging from 10 to 50 percent
is in contact with a liquid, a simultaneous transport of the liquid into, and plasticizer out of
the plasticized PVC, is generally observed. In a first attempt, each of these transports
studied separately were found to be controlled by diffusion. The following conclusions
are worth noting from these experiments (35,36) :
- the rate of transfer at the beginning of the process is higher for the liquid than for the
plasticizer.
- the diffusivities of the liquid and drug vary during the process.
- in some cases, depending on the initial amount of plasticizer in the polymer, the amount
of liquid transferred into the polymer passesthrough a maximum.
- gradients of concentrations are determined by experiments either for the liquid or
plasticizer.
Deeper and further studies have shown that the diffusivities
dependent in a complex way : both the diffusivity
are concentration-
of the plasticizer and liquid are
expressed as a function of the concentration of liquid and plasticizer (37, 38). Resulting
from this dependency with the concentrations, the diffusivity for each matter varies with
time and position in the polymer during the process.
More recent studies on the process of evaporation of the liquid which has entered the
PVC have shown how difficult it is to evaporate this liquid to dryness. This is due to the
concentration dependency of the diffusivities and to the presence of typical gradients of
concentrations of the liquids. With a low concentration of plasticizer and a decreasing
concentration of the liquid on the surface during the process of drying, it is obvious that
the rate of diffusion of the liquid next to the PVC surface becomes very low (39,40,41 :
chapter 8, (42) : chapter 12). From the knowledge obtained either by experiments or
modelling for these two matter transfers, a new method for preparing flexible plasticized
PVC with low matter transfers has been elaborated and successfully tested (37 - 45).
The process of transport of these two liquids (plasticizer, liquid) is quantitatively
described by using numerical models with finite differences. These models are very
versatile, taking into consideration all the known facts, especially the simultaneous
transport of liquid and plasticizer which are connected with each other, the concentrationdependent diffusivities, and the amount of liquid absorbed by the PVC at equilibrium. The
214
Drug delivery
from dosage forms
[Ch. 10
complex casewith a double matter transfer can be fully describedby the numerical model
(35 - 38), as well as the more complex case of transfer when the amount of liquid
absorbedby the polymer passesthrough a maximum (38,41).
ELASTOMERS IN CONTACT WITH A LIQUID
When an elastomeris in contactwith a liquid, the liquid may enter the polymer with
a subsequentswelling. The liquid enters the elastomer following a diffusional process
with constantor concentration-dependentdiffusivity (46,47). The desorption processof
the liquid is also simple, being controlled by diffusion of the liquid within the rubber and
evaporation from the surface.Apparently, absorption and desorption can be followed in
succession,without great change in the kinetics, the process being reversible. Deeper
studiesshow that some additives, such asplasticizer and other materials of low molecular
weight, are generally extractedfrom the elastomersduring the stageof absorption,leading
to irreversible damageto the rubber.
DRUG-POLYMER DEVICES IMMERSED IN GASTRIC LIQUID
Various studiesmade on drug-polymer deviceswhere sodium salicylate as a drug is

embedded in Eudragit (48) or Carbopol (49) have described the process of transports
when thesedevicesare immersedinto gastricliquid. The following results aredrawn :
- two matter transfers simultaneously take place by following this process : the liquid
Space
*
Cd,0
cn,o
Cn+l,O
-m___
i.A+
(i+l).At
I
I
I
I
I
I
I
I
I
Time
Fig. 10.3. Space-timediagramfor calculationof the concentrationof drug andliquid.
Drug-Eudragit
Sec. 10.21
sheet in gastric
liquid
215
enters the polymer matrix and dissolves the drug, enabling the drug to diffuse through the
internal liquid out of the dosage form.
- these two matter transfers are controlled by transient diffusion with concentrationdependent diffusivities.
- further and deeper studies have shown that the diffusivities of the liquid and drug can be
expressed in terms of the concentrations of liquid and drug, these two transfers being
connected with each other (50,5 1).
Numerical models based on methods with finite differences have been built up in
order to describe the whole process in the case of plane and spherical dosage forms, (48 51), and in more complex case where the dosage form is made of a core and shell. Of
course, as shown in the following sections, these numerical models take into account all
the known facts.
10.2 DRUG-EUDRAGIT
10.2.1
SHEET
IN GASTRIC
LIQUID
INTRODUCTION
In this section, a polymeric matrix sheet is considered in which the drug, sodium
salicylate, is dispersed and embedded. These sheets are produced by pressing powder
mixtures by using Eudagrit RS as the tablet binder. Eudragit RS is a high molecular
weight polymer, and tablets having Eudragit RS as the binder exhibit such advantages as
hardness, palatability,
strength and stability. Moreover, the Eudragit residue is not
absorbed in the body and passes through unchanged (52). Tablets containing various
concentrations of sodium salicylate in the polymer matrix swell without disintegrat;on or
attribution, when they are immersed in gastric liquid.
The first purpose in this section (48) is to show that the rate of release of the drug
from the whole tablet is described by the Ficks diffusion equations. Moreover two matter
transfers simultaneously take place as follows : the gastric liquid penetrates the matrix and
dissolves the drug, which then diffuses through the liquid located in the polymer and
leaves the dosage form.
The other aim in this section is to get insight into the process of matter transfers by
developing a numerical model able to describe all the known facts. Modelling liquids
transfers by using a numerically explicit method with finite differences and data concerned
with diffusivities and amounts of liquids transferred at equilibrium was proved of interest
in previous works devoted to plasticized PVC (37, 38, 53). No drastic assumption was
Drug delivery
216
from dosage forms
[Ch. IO
necessary for this method, and the diffusivities were found to be concentration-dependent.
In the present section (48), the numerical model is applied to the study of the diffusion of
the gastric liquid into and the diffusion of the drug out of the dosage form. The model is
developed and verified on tablets containing various concentrations of the drug in the
polymer ranging from 15 to 50 (weight %) by considering not only the transfer of the
drug but also that of the liquid into the polymer.
10.2.2
THEORETICAL ASPECTS
ASSUMPTIONS
The following assumptions are made :

(i) A thin plane sheet of material is considered with a one-dimensional transfer.
(ii) The drug and polymer are intimately mixed, and the initial drug concentration is
uniform.
(iii) Two simultaneous matter transfers take place, according to the following process : the
liquid penetrates the polymer and dissolves the drug, enabling the drug to leave the dosage
form through the liquid located in the polymer.
(iv) The two matter transfers are controlled by transient diffusion, with concentrationdependent diffusivities.
(v) These two matter transfers are connected with each other, and each diffusivity depends
on the concentrations of the liquid and of the drug.
(vi) The concentrations of the liquid and drug on the faces of the sheet reach the values at
equilibrium as soon as the dosage form is soaked into the liquid. The drug concentration
on sheet faces is not zero as it is often assumed in mathematical treatment (41 - chapter 1).
This is due to the finite volume of the liquid and the presence of a partition factor for the
drug between the dosage form and the liquid.
(vii) The change in dimension of the dosage form is not considered.
MATHEMATICAL TREATMENT
The equation of diffusion
in on-dimension with a concentration-dependent
diffusivity for the liquid and the drug is :

(10.3)
ac z&D.%
at
I
where C is the concentration of the liquid (or of the drug) at time t and position x
and D is the diffusivity of the diffusing substance.
Drug-Eudragit
Sec. 10.21
sheet in gastric
liquid
217

(10.4)
(10.5)
and
t=O
t>o
O<x<L
x= 0
x=
cd = constant
cd=C%q
Cl = Cleq
In the presentproblem, becauseof the concentrationdependencyof the diffusivities

and also becauseof the presenceof two matter transfers connectedwith each other, no
analytical solution can be found. (41) : chapter 1 - (54).
However, for very short times, and especially at the beginning of the process,the
concentrations of the liquid and drug are about the same as the values of the initial
concentrations. The concentrations of the matters transferred are thus uniform. The
amount of diffusing substancetransferred M, at time t can thus be expressedin terms of
time t by the well-known relation :
where M, is the amount of matter transferred after infinite time, when equilibrium is
reached.
This above equation is very useful for determining the diffusivity from the slope of
the straight line obtained by plotting M, / M, as a function of the squareroot of time. It
can be usedevenwhen the diffusivity is concentration-dependent.
NUMERICAL
ANALYSIS
The problem must be solved by using a numerically explicit method with finite
differences,availablefor micro-computers.
The principle of the method is describedin Chapter6 with a plane sheet.
In the cross-sectionof the sheet shown in Fig. 10.3, the thickness is divided in N
equal finite slices of thickness Ax by concentration-reference planes (n, i). A matter
balancewritten on the plane n between the time i.At and (i + 1) At enablesone to obtain
for the liquid and drug within the polymer matrix the new concentrationCN,.,as a function
of the previous concentrationsat the sameand adjacentplaces.
LIQUID IN THE POLYMER MATRIX
The new concentrationis given by :
Drug delivery
218
from dosage forms
[Ch. 10
with the dimensionless number Mi expressed in terms of the increments of space Ax and
time At
(A$
Mt, = -.
At
(10.7)
1
Df,
and with the diffusivity DL :
(10.8)
Df, = exp
Bl
CL and Cf being the concentrations at time t and position n of the liquid and plasticizer,
respectively.
and A, and B, constants to be determined from the short test experiments.
DRUG IN THE POLYMER MATRIX
The new concentration is :

(10.9)
CN; = -!4
C~-,+(M;-2)C;+C,d+l
with the dimensionless number Mz
(10.10)
(Ax)
M,d = -.
At
1
D;
and the diffusivity of the drug Df :
(10.11)
Dz = exp !
Ad
-Bd
cf+
a.
c;
Drug-Eudragit
Sec. 10.21
sheet in gastric liquid
219
LIQUID AND DRUG ON THE SHEET FACES
Following the assumption, the concentrations of liquid and drug attained the values
at equilibrium as soon as the process starts.
AMOUNT OF MAlTER LOCATED IN THE SHEET
The amounts of drug and liquid located in the dosage form are obtained by
integrating the concentrations with respect to space :
N-l
(10.12)
M, =
10.2.3
Co+c
C, .Ax
n=l
EXPERIMENTAL
MATERIALS
Sodium salicylate and Euragit RS (copolymer of dimethylaminoethylacrylate
and
ethylmethactylate) of molecular weight 150,000 (Riihm Pharma) are used as the drug and
polymer matrix, respectively. These two materials in powder form are intimately mixed in
a mortar. Sheets are pressed in a steel mould operated by a press at 120 C for 30 s under
a pressure of 180 bars, after a 6 mm heating. Tablets (1.5 cm in diameter and 0.024 cm
thick) are cut from the sheets. Cohesion after compression is excellent, Eudragit being an
effective tablet binder.
DETERMINATION
OF MA-ITER TRANSFERS
Experiments are carried out in a closed flask using a controlled rate of stirring. The
tablet (500 mg) inserted in a basket made of glass fiber, is soaked in synthetic gastric
liquid (100 ml) at 37 C. The composition of the liquid is : for 1,000 ml of aqueous
solution, 80 ml H Cl 1 N, 2 g Na Cl and pH = 1.2.
At intervals, a sample of liquid (0.1 ml) is taken for analysis, and the tablet is
weighed. The concentration of sodium salicylate in the liquid is determined using a
double-beam UV-spectrophotometer (Beckman DB-G) calibrated at 300 nm.
CALCULATIONS
The profiles of concentrations of the liquid and drug developed through the sheet are
calculated by using the numerical model and data concerned with the diffusivities and the
amounts of matter transferred at equilibrium.
220
Drug delivery
from dosage forms
Kh.
SAL
O/O
+-+--+
-+-ii2
)Os5
l
tlhl
,.
15
[Sh)
2.
Fig. 10.4. Amount of drug released(weight % of initial sample)vs. (tirne)0.5,for

various initial concentrationof drug (15-25-50weight %). In vitro test with Eudragit
RS.
LO
Time(h)
Fig.10.5. Kinetics of releaseof drug for various initial concentrationsof drug in

the device. In vitro test with Eudragit RS.
10
Sec. 10.21
Drug-Eudragit
10.2.4 RESULTS
MATRIX
OBTAINED
WITH
EUDRAGIT
AS
221
POLYMER
EXPERIMENTAL RESULTS AND DETERMINATION OF THE DATA
The amount of sodium salicylate released in the synthetic gastric liquid is plotted as a
function of the square-root of time, for various initial concentrations of drug in the
polymer matrix (Fig. 10.4).
The values at equilibrium are obtained for times of about 15 - 20 h.
The following conclusions can be drawn :
(i) Straight lines are obtained for the amount of drug as a function of the square-root of
time.
(ii) The higher the initial concentration of drug, the higher the rate of transfer of the drug.
(iii) The amount of drug transferred at equilibrium increases with the initial amount of
drug in the dosage form as shown in Fig. 10.5.
(iv) The diffusivity
Fig. 10.6.
for the drug is about constant and does not depend on the drug
Amount of liquid transported (weigh % of the initial sample) vs.
(time)u.5, for various initial concentrations of drug (15-25-50 weight %). In vitro
test with Eudragit RS.
222
Drug delivery
[Ch. 10
from dosage forms
concentration.
The amount of liquid transportedinto the dosageform is expressedas a function of
the square-root of time for various initial concentration of the drug in the devices (Fig.
10.6).
The amount of liquid absorbedby the devices at equilibrium is shown in Fig. 10.7
as well as in Table 10.1. The diffusivity in the logarithm form of the liquid is inversely
proportional to the initial concentration of drug in the devices (Fig. 10.8). The values of
the diffusivity and amountof liquid absorbedat equilibrium are shown in Table 10.1.
The diffusivity of the liquid can be expressedin terms of the initial concentrationof
drug in the device by :
(10.13)
SO-50
75-25
85-15
Time(h)
*
LO
Fig. 10.7. Kinetics of absorptionof liquid by deviceswith various initial

concentrationof drug (15-25-50 weight %). In vitro test with Eudragit RS.
Drug-Eudragit
Sec. 10.21
223
Table 10.1Diffusivities (cm2/ s) and amountsof matter transferredat equilibrium (%)

Eudragit-Drug
50 - 50
75 - 25
85 - 15
w.aug
20.5
12.5
4
M, liquid
48
37
29.5
Dd x 108drug
16.5
16.5
16.5
Dx 108
70.5
37.5
11
andthe diffusivity of the drug is constant:

(10.14)
Dd = 16.5 x lo-*
-11
002
I
(cm2/s)
001
1
006
I
+
-L
C
-15,
Ln(D)
'I
Fig. 10.8. Log diffusivity of the liquid vs (initial drug concentration)-1.
VALIDITY OF THE MODEL
A swelling of the dosageform is observedas the rate of liquid transferredis higher

than that for the drug. However, a frame of referencefixed with respectto the initial sheet
is considered in this numerical model. Then the thickness of the sheet is constant and
224
Drug delivery
from dosage forms
[Ch. 10
equal to the thickness of the original unswollen sheet.

The validity of the model is tested in Figs 10.5 and 10.7 by comparing the kinetics
of drug and liquid transferred obtained either by experiments or calculation. Not only a
good agreement is observed for the drug but also for the liquid. In fact, the validity of the
model is tested as well as the accuracy of the data.
Some profiles of concentration of the drug and liquid developed through the
thickness of the sheet are calculated and drawn in Fig. 10.9 when the initial concentration
of drug is 50 % in the device and in Fig. 10.10 when the initial concentration of drug is 25
%. These profiles of concentration at various times give a fuller insight on the process.
10.2.5 CONCLUSIONS
WITH EUDRAGIT
AS POLYMER
MATRIX
The mechanism of release of drug is elucidated when drug-Eudragit devices with

various concentrations are soaked into gastric liquid. Simultaneous diffusions of the liquid
%SAL
Fig. 10.9.
,-,,,
o/o LIQ
Profiles of concentration of sodium salicylate (left) and liquid (right)
within the sheet of thickness 0.022 cm and the initial drug concentration of 50 %.
Sec. 10.31
Drug-Carbopol
F F
225
Fig. 10.10. Profdesof concentrationof sodium salicylate(left) andliquid (right)

within the sheetof thickness0.022 cm and the initial drug concentrationof 25 %.
into, and the previously dispersed drug out of the polymer sheet are observed. Both
transfers are explained by transient diffusion, and the diffusivity is constantfor the drug
and concentration-dependent
for the liquid.
The numerical model is capableof describingthe simultaneoustransfersof drug and
liquid, in spite of the fact that the swelling of the polymer is neglected.
Only a part of the drug is releasedin gastric liquid, and a significant amount of
drug remains in the dosageform at equilibrium. This is due to the fact that the volume of
liquid is rather small with regardto the amountof drug in the dosageform.
10.3 DRUG - CARBOPOL
10.3.1
SHEET
IN GASTRIC
LIQUID
INTRODUCTION
The purposein this section is to work on tablets containing different concentrations
Drug delivery
226
from dosage forms
[Ch. 10
of sodium salicylate dispersed in Carbopo1934 P which is chosen as polymeric matrix.

Carbopol has shown promise in pharmaceutic tableting where it controls the release of the
medicament (49,55).
Tablets having Carbopol 934 as binder exhibit several advantages : hardness,
palatibility, strength and stability. Moreover, the Carbopol residue is not absorbed in the
body and passes through the gastrointestinal tract unchanged. Thin sheets are considered,
and soaked into gastric liquid. Both transfers of liquid into, and drug out of the sheet are
studied, and diffusivities are determined from the square-root of tune dependence with
amounts of both matters transferred. The method previously described with Eudragit in
section 10.2 is applied to these dosage forms.
THEORETICAL ASPECTS
10.3.2
The problem is the same as that shown with Eudragit, and the method used to
resolve this problem is the same.
The same numerical method as that described in section 10.2 is used.
10.3.3
EXPERIMENTAL
MATERIALS
The drug is sodium salicylate in powder fonn. Carbopo1934 P, in powder form (PB
Goodrich), a polymer of acrylic acid with a mean molecular weight around 3,000,OOOis
used for the polymer matrix.
PREPARATION OF THIN SHEETS
The drug and polymer in powder fonn are intimately mixed in a mortar. Sheets are
pressed with a steel mould operated by a press at 120 C under a pressure of 180 bars,
after a 6 min heating.
Several tablets (2 cm in diameter, and 0.016 cm for the thickness) are cut from the
sheets. The average weight of such a tablet is 0.250 g.
IN VITRO TESTS
Experiments are conducted in a closed 250 ml flask with a controlled rate of stirring.
The effect of the rate of stirring on the kinetics of transport is significant, especially at the
beginning of the process (35,56, 57).
The volume of synthetic gastric liquid is 100 ml, and its composition is described in
10.2.3.
Drug-Carbopol
Sec. 10.31
10
227
20
15
Fig. 10.11. Amount of drug released from the drug-carbopol device as a function of
the square root of tune, for various initial conditions of drug in the device : 15-25-50 9%.
Measurements are made in order to determine the kinetics of the matters transferred :
drug, liquid. At intervals, the tablet is weighed and the concentration of the drug released
from the device into the gastric liquid is determined by UV-spectrometry.
10.3.4
RESULTS
WITH
DRUG-CARBOPOL
DEVICES
EXPERIMENTAL RESULTS AND DETERMINATION OF THE DATA
Two kinds of results are of interest : the diffusivity for the liquid and the drug and
their concentration-dependency, and the concentration of liquid and drug in the dosage
form at equilibrium with the surrounding after infinite time.
DIFFUSIVITIES
The amount of drug released into the gastric liquid varies about proportionally with
the square-root of time, as shown in Fig. 10.11. An increase in the value of the slope is
observed at around 2 h, for the various initial concentration of the drug in the devices
ranging from 15 to 50 96. The diffusivity of the drug is not constant but concentration-
Drug delivery
228
[Ch. 10
from dosage forms
4 SAL
O/O
m-s15
300 cso-s-50
C75-S25
(Time)'.'
1
20
Fig. 10.12. Amount of liquid absorbed by the drug-carbopol device as a function

of the square root of tune, for various initial concentrations of drug in the device :
15-25-50 %.
dependent. This diffusivity is expressed in terms of the concentrations of drug and liquid
in eqn (10.15).
(10.15)
Dd = D;.
The amount of drug released in gastric liquid at equilibrium Mt increases with the initial drug
concentration in the sheets.
The square-root of time dependence with the amount of liquid which enters the
polymer is illustrated in Fig. 10.12. The value of the diffusivity for the liquid varies with
the initial drug concentrations. The values of the diffusivity for the liquid in the logarithm
form is found to be inversely proportional to the drug concentration in the dosage form, as
shown in Fig. 10.13 for various drug concentrations throughout the 15 - 50 % range. The
amount of liquid absorbed at equilibrium increases with the amount of Carbopol in the
device.
Sec. 10.31
Drug-Carbopol
0.01
0.02
229
0.06
l
1
C
-16.
Fig. 10.13. Log diffusivity for the liquid as a function of the (Na Salicylate
concentration)-1 in Carbopol matrix.
The diffusivity for the liquid is expressed in terms of the concentration of liquid and
drug by the eqn (10.16).
(10.16)
Dr = exp(s-
14.71
The various values of the parameters are shown in Table 10.2

Table 10.2 Diffusivities and amount of matters transferred at equilibrium
carbopo1-Drug
Dx 108(crn2/s)
Ddx 10 (cm2 / s)
da,(%)
M:(%)
50 - 50
16
3.3
235
38
75- 25
23
3.6
315
19
85- 1
30
6.7
500
7.5
VALIDITY OF THE MODEL WITH CARBOPOL
The validity of the numerical model as well as the accuracy of the data are tested by
230
Drug delivery
from dosage forms
[Ch. 10
&aAL
O/O
10
20
30
10 Time(h)
Fig. 10.14. Kinetics of release of drug for various concentration of Carbopoldrug devices in gastric liquid.
comparing the kinetics of the matters transferred obtained either by experiment or by

calculation. Good agrement between these kinetics can be appreciated either for the release
of the drug in Fig. 10.13 or for the liquid absorbed by the dosage forms in Fig. 10.14.
PROFILES OF CONCENTRATION WITHIN THE SHEET
Some profiles of concentration of drug and liquid developed through the thickness
of the sheet are calculated with the help of the model and drawn in Fig. 10.15 when the
initial drug concentration in the dosage form is 25 %.
10.3.5 CONCLUSIONS
WITH
DRUG-CARBOPOL
DEVICES
The case of a sheet is rather simple in the same way as of a sphere, because there is a
one-dimensional transport.
A double transport takes places : the liquid enters the polymer, dissolves the drug,
and then enables the dissolved drug to leave the sheet through the liquid located in the
sheet. These two transports are controlled by transient diffusion, and the diffusivities are
Moreover, the two transports of the liquid and drug are
connected with each other, the diffusivity of the drug increasing with the concentration of
liquid in the dosage form.
Effect of pH on drug release
Sec. 10.41
231
600
ces-s15
Time(h)
*
40
Fig. 10.15. Kinetics of absorptionof gastricliquid for various concentrations

of Carbopol-drugdevices.
The numerical model is able to describethe processproperly. The kinetics calculated

for the liquid and the drug are in good agreement with the experiments. Moreover,
interesting information is available, with the profiles of concentration of the liquid and
drug developedthrough the thicknessof the thin dosageform.
10.4 EFFECT
10.4.1
OF pH ON DRUG RELEASE
INTRODUCTION
The first purpose in this section is to show that both transfersof the liquid and drug
can be studied when the diffusivities are concentration-dependent,for various values of
the pH of the synthetic gastric liquid. Dosageforms composedof sodium salicylate and
Eudragit RS are used in this study.
Modelling of the process for various values of the pH of gastric liquid is another
purpose in this section. The numerical model taking into account the simultaneous
diffusion of the liquid into, and the previously disperseddrug out of the polymer matrix,
is applied to this case.This model, based on an explicit numerical method with finite
differences, describesboth thesematters transferredby transient diffusion. It used some
Drug delivery
232
[Ch. 10
from dosage forms
Fig. 10.16. Profiles of concentrationof drug (left) andliquid (right) developed

within the thicknessof the sheetof drugCarbopo1.Initial concentrationof drug
= 25 5%.
data obtained from experimentssuch asthe dependencelaws of the diffusivities with the
drug and liquid concentrations and the amounts of drug and liquid transferred at
equilibrium.
10.4.2
THEORETICAL
ASPECTS
ASSUMPTIONS
The assumptionsarethe sameasthosemade in section 10.2.

(i) A thin sheetof drug-polymer matrix is used so that only one-dimensionaltransport of
mattersis considered.
(ii) The diffusivities are determinedas a function of the drug concentrationby using short
tests and eqn.(10.l), for given values of the pH of the liquid.
(iii) The concentrationsof the drug and liquid on both facesof the sheetreach the value at
equilibrium as soon asthe sheetis immersedin the liquid.
Sec. 10.41
1
S
TimelhI
1
10
1s
233
20
Fig. 10.17. Kinetics of liquid transferredin the dosageforms for various values
of the pH-Eudragit-sodium salicylate (75 : 25 weight %).
+ experimentalcalculated
NUMERICAL ANALYSIS
The model describedin section 10.2 is used for this study.
10.4.3
EXPERIMENTAL
DRUG-POLYMER DOSAGE FORMS
Eudragit RS (already described in 10.2.3) is used as polymer matrix and sodium

sahcylate for the drug. These componentsin powder form are mixed and pressedin the
sameway as shown in 10.2.3.
Tablets of 1.9 cm in diameterand 0.024 cm in the thicknessare cut from the sheets.
The composition for the tablets is : Eudragit 75 - sodium salicylate 25, these
percentagesbeing expressedin weight %.
Drug delivery
234
from dosage forms
10
[Ch. 10
Time lh)
I
15
20
Fig. 10.18. Kinetics of drug transferred out of the dosage forms for various values
of the pH- Eudragit-sodium salicylate (75 : 25 weight %).
MEASUREMENTSOF MAllER TRANSFERRED
The tablet of around 250 mg previously inserted in a basket of fibar glass, is

immersed in synthetic gastric liquid (100 ml) maintained at 37 C with a controlled rate of
stirring.
Samples of liquid (0.1 ml) are taken at intervals for analysis and the tablet is
weighed.
Various compositions are prepared for the liquid in order to have several pH values.
The values chosen for the pH of the liquid range from 1.9 to 7.4.
10.4.4
RESULTS
WITH
DRUG-EUDRAGIT
SHEETS
EXPERIMENTAL RESULTS FOR THE MA-ITERS TRANSFERRED
The amount of liquid transferred into the dosage forms is plotted against tune in Fig.
10.17 for various values of the pH of the liquid within the 1.9 - 7.4 range. As shown in
these curves, the behaviour of the liquid is quite different according to the pH value. For
instance, the amount of liquid transferred increases regularly with time for the lower
Sec. 10.41
235
Fig. 10.19. Amount of liquid transferred as a function of the square-root of time,

for various values of the pH of the liquid. Eudragit-sodium salicylate 75 : 25.
values of the pH (1.9 - 2.25). But for the higher values, the amount of liquid transferred
increases with time and then decreases, passing through a maximum. The value of this
maximum varies with the pH of the liquid. On the whole, it is seen that the higher the pH,
the lower the maximum of the amount of liquid transferred.
As shown in Fig. 10.18, the amount of drug released from the dosage forms into the
liquid increases regularly with time, for the various values chosen for the pH of the liquid.
The value of the pH of the liquid has an effect on the drug transfer : the higher the pH, the
larger the amount of the drug released.
DIFFUSIVITIES AND THEIR pH-DEPENDENCY
In order to obtain the values of the diffusivity for the liquid and drug, the amounts of
liquid (Fig. 10.19) and drug (Fig. 10.20) transferred are expressed as a function of the
square-root of time. Straight lines are obtained in all cases, demonstrating a diffusional
process for these two matters.
236
Drug delivery
[Ch. 10
from dosage forms
10
15
Fig. 10.20. Amount of drug transferredas a function of the square-rootof time,

for various values of the pH of the liquid. Eudragit-sodiumsalicylate 75 : 25.
By using eqn.(lO.l) :
and the values of the slopesin Fig. 10.19 and 10.20, it is possible to determinethe values
of the diffusivities, if the values of the amount of matter transferred at equilibrium are
available. There is no special problem in obtaining the amount of drug at equilibrium, as
the amount of drug increasesregularly with tune with an asymptotic tendency. But the
problem for the liquid is far more complex, as the amount of liquid-time history passes
through a maximum. Three different values can be chosenfor the value of the amount of
liquid transferredat equilibrium :
(i) the value obtained for a long time, by consideringthe decreasein the amount of liquid
with tune when a maximum is observed.
Sec. 10.41
-PH Log 0 SAL

II = expl-15.2-
237
:I
10
6
/
X
1231567
Fig. 10.21. pH. Log D for the drug as a function of pH of liquid. Eudragitsodium salicylate 75 : 25.
(ii) the maximum value of the liquid transferred

(iii) the value obtainedby extrapolationfor a long time.
As the numerical model describedin the previous method taking into account this
maximum is rather complex (38), and the effect of the liquid is very important on the drug
transfer especially fort the first 6 h, the simple caseii is chosenas it gives good results in
this case(59).
The values of the diffusivities are shown in Table 10.3 for the liquid and drug.
The variation of the diffisivity of the drug with the pH of the liquid is shown in Fig.
10.21 where pH. Log D is plotted as a function of the pH. The straight line obtained
Drug delivery
238
Kh.
from dosage forms
Table 10.3Diffusivities (D x 108cm2/ s) and M, (%) with pH

D&8
~liquid
-8
1.2 37.5
1.9 37.9
15.2
21
48
13
40
18
2.25 38
22
39.3
18.2
PH
Dliquid
55.8
22.5
34
19.4
20
4.2
100
22.8
32
104
23.1
29.5
20.4
7.4
131
23.6
28
21.8
123kS67
Fig. 10.22. Diffusivity for the liquid as a function of the pH. Eudragitsodium salicylate 75 : 25.
10
Sec. 10.41
239
Fig. 10.23. Amount of drug transferredat equilibrium as a function of the pH of

the liquid.
allows one to fit the following equationfor this variation :

(10.17) D,,
= 2.5~
10-7XeXp(-~)
(a2/s)
As shown in a previous study (48) and in section 10.2, the diffusivity for the drug
doesnot dependon the concentrationof drug or liquid in the polymer matrix.
The diffusivity of the liquid transfer through the polymer increaseswith the pH of
the liquid as shown in Fig. 10.22. The dependenceof the pH of the liquid is of great
importancefor valueslower than 3, and slightly sensitivefor higher values.By taking into
account the effect of the concentration and pH, the following expression is found to
expressthe variation of the diffusivity :
(10.18) Dliquid= 1.48x 10dxexp(0.129pH)xexp
Another result of interest is concerned with the amount of matter transferred at

equilibrium, and its variation with pH. The increasein pH from 1.8 to 7 is responsiblefor
an increasein the amount of drug transferredat equilibrium (Fig 10.23) and a decreasein
Drug delivery
240
[Ch. 10
from dosage forms
I
X
X
3836\
31_
32_
30,
1231567
Fig. 10.24. Amount of liquid transferredat equilibrium as a function of the pH

of the liquid.
the correspondingamount of liquid (Fig 10.24).

MODELLING OF THE PROCESS
Modelling of the process is obtained by using the numerical model in the case of
experimentsmade with a constantvalue of the pH. As shown in Fig. 10.18 for the drug
and in Fig. 10.25 for the liquid, the calculated results are in good agreementwith the
experimental ones for various values. In each case,the pH of the liquid is kept constant
during the experiment.
10.4.5
CONCLUSIONS
The effect of the pH of the liquid on the rate of the matters transferred is studied in
the case of Eudragit-sodium salicylate devices. Not only the diffusivities, but also the
amountof matter transferredat equilibrium arefound to vary with the pH of the liquid.
The diffusivities for the drug and liquid increasewith the pH of the liquid, but the
effect of the pH is more significant for the liquid than for the drug. The effect of the pH of
the liquid on the amountof matterstransferredat equilibrium is quite different for the drug
Sec. 10.41
Time [h)
I
10
IS
241
20
Fig. 10.25. Amount of liquid transferred as a function of time, for various pHs.
x : experimental: calculated
and the liquid. While this amount increases with the pH in the case of the drug, in contrast
it decreasesfor the liquid. The result is obvious for the drug as it is the salt of an acid.
Models of both these transfers are obtained with good results, by taking into account
all the experimental results. In the case of constant pH, these values are chosen within a
large range.
Of course, these models can describe more complex processes where the pH of the
liquid varies with time according to given laws. After each increment of time, the value of
the pH of the liquid is calculated, as well as the various diffusivities
and amounts of
matters transferred at equilibrium. The concentrations of the liquid and drug are thus
calculated within the sheet ; the amounts of liquid and drug located within the sheet are
obtained by integrating these concentrations with respect to space.
Drug delivery
242
from dosage forms
10.5 SPHERICAL
LIQUID
1 OS.1
DRUG-EUDRAGIT
[Ch. 10
BEADS
IN GASTRIC
INTRODUCTION
The first purpose in this section is to describe and study the behaviour of dosage
forms, spherical in shape, prepared by dispersing the drug (sodium salicylate) in the
polymer matrix (Eudragit RS) and compressing the mixture. The mixture is previously
transformed into a homogeneous thick paste, after pulverization with a small amount of
ethanol which is a bad solvent of the drug. Spherical beads are thus easily obtained by
pressing the thick paste in a mould at a very low pressure and at room temperature. A
problem thus appearswith the final drying stage of the beads until complete evaporation of
ethanol. The problem of drying is especially studied in Chapter 11.
The choice of spherical beads is made for two reasons at least :
(i) From the theoretical point of view, these forms are of interest because of their
symmetry.
(ii) Because of the interesting effect of the size of these beads per unit of weight on the rate
of drug release, industrial applications can be found either by using these spherical dosage
forms themselves, or by surrounding these beads with the same polymer or with an
erodible polymer.
Another purpose in this Section is to develop a numerical model able to account all
the facts. By simulating the process with the help of computer, it is thus possible to
determine the effect of all parameters and to gain a fuller insight into the process of matter
transfers (50, 51).
The process is controlled by the diffusion of the liquid through the polymer matrix.
The liquid enables the drug to dissolve and to diffuse out of the dosage form through the
liquid located in the polymer. These two matters transferred are thus connected with each
other, and the diffusivity of the drug is bound to depend on the concentration of the liquid
in the polymer.
10.5.2
THEORETICAL
ASPECTS
Before describing the model based on numerical analysis calculation,

assumptions are made in order to clarify the process and mathematical problems.
ASSUMPTIONS
The following assumptions are made on the sample and on the process :
some
Spherical
Sec. 10.51
Drug-Eudragit
beads in gastric liquid
243
(i) The samples are spherical in shape

(ii) The drug is properly dispersed in the polymer matrix
(iii) Two matter transfers take place in the sample : the one concerned with the liquid
which enters the polymer and dissolves the drug, allowing the other transfer of the drug
out of the dosage form through the liquid located in the bead.
(iv) Both these transfers are controlled by transient diffusion. Moreover they are
connected with one another.
(v) The diffisivities of these matter tranfers are concentration-dependent, these laws being
determined by experiments with short tests.
(vi) As soon as the dosage form is immersed in the gastric liquid, the concentrations of the
liquid and of the drug attained their values at equilibrium on their surface, meaning that the
coefficient of matter transfer on the surface is very high.
(vii) The polymer swells, especially at the beginning of the process. However, a frame of
reference fixed with respect to the original dosage form is taken for numerical analysis and
calculation (46,47,50).
The equation of radial transient diffusion through the spherical bead is given by :
(10.19)
= $.
ac
& D. r . a
with the concentration-dependent diffusivity D.

Following the assumptions, the initial and boundary conditions are written as
follows :
(10.20)
t = 0
r < R
C = C$
(10.21)
t > 0
r = R
C = C,
sample
For very short tests, when Mt / M, < 0.3, eqn (10.22) can be used for determining
the diffusivity
0.5
(10.22)
co
= ;
(I
rc
244
Drug delivery
[Ch. 10
from dosage forms
As already said, for a short time, the amount of matter transferred is very low and
the concentration can be consideredas equal to the initial concentration.The diffusivity
obtained by using this equation is thus obtained for this initial concentration.With short
tests, the diffusivity can thus be determined even when the diffusivity is concentrationdependent.
NUMERICAL ANALYSIS
No analytical solution can be found in the present casewith the two simultaneous
matter transfers which are connectedwith each other, and the concentration-dependent
diffusivities.
The problem is solvedby using an explicit numrical method with finite differences.
As the problem is especiallystudiedin Chapter8, only the equationsare given in the
present section 10.5, with a constant concentration on the surface meaning that the
coefficient of matter transferon the surfaceis infinite.
The concentrationis thus given in the following three places : within the sphere,at
the centreof the sphere,on the surface.
at pOSitiOn n
The new concentration CN, after elapse of time At is expressedin terms of the
previousconcentrationsat the sameand adjacentplaces:
WITHIN THE SPHERE,
(10.23) CN, = C,+
-n-m+
n+OJ
where the function J, is :

(10.24)
J, = n2. D,. (C, -0.5 + C, + OS)
with n = 0
The new concentration CN, at the centre of the sphere after elapse of time At is
given by :
(10.25)
24. At
CN, = Co- ~
JO.5
(Ad2
Sec. 10.51
Spherical
Drug-Eudragit
245
with the function Jo.5:

(10.24)
Jcr5 = ;.
Do,,. (Co-C,)
n=N
The concentration on the surface is constantly equal to the value attained at

equilibrium.
(10.26)
liquid
CL = C;
d
cN=
c,
d
dJ-%
AMOUNT OF MAll-ER LOCATED IN THE SPHERE
The amount of diffusing substance (liquid, drug) located in the sphere at time t is
obtained by integrating with respect to space the concentrations within this sphere at this
time.
(10.27)
Mt
4 * (*r)3
- (N - O.5)3
CONDITIONS OF STABlLllY
The conditions of stability for calculation are obtained when the coefficient of the
previous concentration at the position considered is positive. It is thus found at the centre
of the sphere for a constant diffusivity :
(10.28)
&
>
246
Drug delivery
10.5.3
[Ch. 10
from dosage forms
EXPERIMENTAL
MATERIALS
Sodium salicylate (COPER) in powder form is chosenfor the drug.

Eudragit RS (Rohrn Pharma), a copolymer of dimethylaminoethylacrylate and
ethylmethacrylateof molecular weight 150,000is usedfor the polymer matrix.
Thesetwo materialsin powder form are intimately mixed in a mortar. The mixture is
transformed into a homogeneousthick paste after pulverization with a small amount of
ethanol. Spherical beads of various sizes are obtained by pressing the paste in a mould,
andthen drying until completeevaporationof the ethanol (2-4 days).
Severalsphereswith the samesize (radius of 0.43 cm) and various compositionsare
preparedand tested in the sameway. The percentageof drug in thesedosageforms is as
follows : 25-40-50-60-75weight percent.
Other spheres with the same composition (50 wt % of drug) and various sizes
(diametersof 0.36 - 0.62 - 0.87 cm) are also preparedand tested.
DETERMINATION
OF THE PARAMETERS
Experimentsare conductedin a closed flask with a rotating bar magnet.Thespheres

(400-500 mg), inserted in a fiber glass basket, are immersed in synthetic gastric liquid
(100 ml) at 37T, with the following compositionsfor 1,000ml of aqueoussolution :
pH = 1.2
pH = 4.7
pH = 7.2
2 gNaC1
8.2 g sodium acetate
10.8 g Naz H PO4
8Oml HCl
lN,
6 g acetic acid ,
2.74 g Na H, PO4
At internals, samplesof liquid (0.1 ml) are taken for analysiswhile the spheresare
weighed. The amount of sodium salicylate released from the polymeric device is
determined using a double-beam W-spectrophotometer (Beckman DB-G) calibrated at
300 run.
10.5.4
- RESULTS
DETERMINATION
WITH
DRUG-EUDRAGIT
SPHERES
OF THE PARAMETERS
DIFFUSIVITIES OF THE LIQUID AND DRUG
The variation of the amount of the drug releaseand of the liquid absorbedby the
polymer as a function of the square-rootof time is illustrated in Fig. 10.26 for the drug
Spherical
Sec. 10.51
A PP
O/O
Drug-Eudragit
247
x 60%
30-
( Tirne1~5
l
10
Fig. 10.26. Drug transferredas a function of the square-rootof time, for various
drug concentrationsin sphericaldosageforms. Radius = 0.435 cm.
and in Fig.10.27 for the liquid, with various initial concentrations of drug in dosage
forms. Two conclusionscan be drawn :
i - Straight lines are obtained for the liquid absorbed (Fig. 10.27), passing through the
axesorigin.
ii - In contrast witih the liquid, the slopes of the curves expressing the amount of drug
releasedversus the square-rootof time, are not constant at the beginning of the process.
Starting with a very low value when the dosageform is immersedin the liquid, the slopes
increaseregularly with time up to a constantvalue.
The diffusivity for the liquid is constant:
(10.29)
Dt = 4.3 x 10-6(cm%)
while the diffusivity for the drug is expressedin terms of the concentrationof the liquid in
the dosageform by the relationship :
(10.30) Dd= 6 x lo-x exp - 4oo (c&*/s)
i C I
248
Drug delivery
from dosage forms
[Ch. 10
Fig. 10.27. Liquid absorbedby sphericaldosageforms as a function of the squareroot of time for various initial drug concentrations.Radius = 0.435 cm.
10
15
20
25
Fig. 10.28- Kinetics of the matter transported(drug, liquid) with a drug-Eudragit

(50-50 wt %) dosageform. Radius = 0.435 cm + experimentstheoretical
Spherical
Sec. 10.51
Drug-Eudragit
249
4P
Licl A
O/O
O/O
30,
Time(h)
LO
Fig. 10.29. Kinetics of the matter transported(drug, liquid) with a drug-Eudragit

(40-60 wt %) dosageform. Radius = 0.435 cm + experimentstheoretical
whateverthe initial concentrationof the drug.

AMOUNT OF MA-l-l-El3 TRANSFERRED AT EQUILIBRIUM
The amount of matter transported at equilibrium or rather after a long time, is

obtained from the kinetic curves shown in Fig. 10.28, 10.29 for various initial
concentrationsof drug in dosageforms (50 and 40 % in weight) having the samesize and
weight, and for a dosageform with a different size and weight.
The data obtained for the diffusivities and the amounts of matter transported at
equilibrium are shown in Table 10.4.
Table 10.4Diffusivities and amountsof matter transferredat equilibrium.
d
D
Radius
Drug-Polymer
M:
DO
wt %
cm
(%o)
50-50
50-50
40-60
0.435
0.31
0.435
6 x 10-3
6 x 1O-3
6 x lo-
4.3 x 10-e
4.3 x 10-6
4.3 x 1O-6
66
70
62
d
(%)
88
90
76
250
Drug delivery
from dosage forms
[Ch. 10
b PA
LIQ b
- 10
-30
-20
-10
0
0
10
TimelhI
I
20
Fig. 10.30. Kinetics of the matter transported (drug, liquid) with a drug-Eudragit
(50-50 wt %) dosage form. Radius = 0.435 cm + experiments theoretical
Some results are of interest :

(i) The amount of liquid transferred after infinite time, expressed in terms of the polymer
weight, is about constant, ranging from 76 to 90 wt %.
(ii) The amount of drug released at equilibrium as a fraction of the initial weight of drug is
about constant, ranging from 62 to 70 wt %.
(iii) The diffusivity for the liquid is constant.
(iv) The diffttsivity for the drug can be expressed in terms of the concentration of the liquid
in the dosage form (eqn.10.30).
VALIDITY OF THE MODEL
The experimental and calculated kinetics of matters transferred (liquid, drug) are
drawn in Figs 10.28 - 10.29 - 10.30. Good agreement is obtained between these curves
not only for the drug but also for the liquid, for various concentration of the drug and
various sizes of the dosage forms. These curves illustrate the validity of the numerical
model for both matter transports.
Spherical
Sec. 10.51
Drug-Eudragit
beads in gastric
liquid
251
Fig. 10.3 1. Profiles of concentrations developed through the radius of the bead.
Left : drug - Right : liquid - Drug-Eudragit : 50-50 in weight- Radius = 0.435 cm.
PROFILES
OF CONCENTRATION
OF THE DRUG AND LIQUID
The numerical model is capable of providing one with the profiles of concentration
of drug and liquid developed through the dosage form during the process, as shown in
Figs. 10.31, 10.32 and 10.33. Some results can be drawn from these curves :
(i) Steep gradients of concentration of liquid and drug are developed within the dosage
form, especially next to the surface.
(ii) The dosage forms swell, as the rate of transport is higher for the liquid than for the
drug.
(iii) Following
the assumption, a constant value is reached on the surface for the
concentration as soon as the process starts which is equal to the concentration at

equilibrium.
(iv) After infinite time, a part of the drug remains in the dosage form. This fact is due to
the low solubility of the drug in the gastric liquid and to the ratio of the volume of liquid
and drug. When the volume of gastric liquid is increased, the amount of drug remaining in
the dosage form decreases.
10.5.5
CONCLUSIONS
The case of a sphere is of interest because many dosage forms are spherical in
shape.
252
Drug delivery
[Ch. 10
from dosage forms
LIQ
O/O
22h
A!O
,i !O
Fig. 10.32. Profiles of concentration developed through the bead. Left : drug Right : liquid-Drug-Eudragit : 40-60 in weight- Radius = 0.435 cm
b LIQ
PA
O/O
O/O
lh
7h
,O,
Fig. 10.33. Profiles of concentration developed through the bead. Left : drug Right : liquid-Drug-Eudragit : 50-50 in weight- Radius = 0.30 cm.
Sec. 10.51
Spherical
Drug-Eudragit
253
The problem is perhaps simple as there is a radial transport. However, the

contribution of each spherical slice of constantthickness to the transport is not the same,
the volume of each spherical slice being proportional to radius squared. Under these
conditions,the behaviourof the external slice next to the surfaceis more effective than that
of any internal slice.
As shown in the case of a sheet, a double transport of matters is observed. The
liquid enters the polymer, dissolves the drug, and enables the drug to leave the dosage
form through the liquid located in the dosage form. Both these matter transfers are
controlled by transientdiffusion, with a concentrationdependencefor the diffusivity of the
drug. In a simple way, the diffusivity of the drug can be expressed in terms of the
concentrationof the liquid in the dosageform.
The numerical model is able to describethe processproperly, the kinetics calculated
for the drug and liquid being in good agreementwith experiments.Moreover, information
of interest is obtained in addition, with the profiles of concentration of each matter
developedwithin the bead.
REFERENCES
A. Zaffaroni, New approachesto drug administration, 31st Intern. Congress of
1
Pharmac.Sci., Washington, DC (1971), 19 - 20.
J. Godbillon, J. Richard, A. Gerardin, J. Moppert, D. Leory, and F.Theeuwes,
2
Plasmaconcentrationprofiles of metoprolol achievedafter administrationof oral osmotic
controlled-releasesystems(OROS), in eds.J.M Aiache and J. Hirtz, 1st Europ Cong. of
Biopharmac and Pharmacokinet. Proceed, V.l Technique et Documentation,
Biophamraceutics(1981), 437 - 466.
3
Ibid., p. 163.
T. Higuchi, Osmotic dispenser with collapsible supply container, U.S. Pat,

4
3,760,805, September 15, 1973.
M.J. Kendall, D.B. Jack, K.L. Woods, S.J. Laugher, C.P. Quarter-man,and V.A.
John, Comparaisonof the phamracodynamicand pharmacokineticprofiles of single and
5
254
Drug delivery
from dosage forms
[Ch. 10
multiple dosesof a commercial slow-releaseMetoprolol formulation with a new OROS

Delivery System, Br. J. Clin. Pharmacol., 13, (1982), 393 - 398.
K. Heilman, in Therapeutic Systems: Rate-controlleddrug-delivery ; Concept and
6
Development. New York : Thibme StrattonInc. (1984), 1 - 34.
J. Feijen, Controlled drug delivery basedon the use of syntheticpolymers,
7
Rouen : XIV Meet. of French Polymer Group, November 1984.
M. Vergnaud-Trouin and J.M. Vergnaud, Modelisation de la liberation du principe
actif a partir dune matrice polymerique, SIP Pharma,(Avril - Mai 1986), 410 - 415.
J. Heller, Biodegradable polymers in controlled drug delivery, CRC Critical

Reviews in Therap. Drug Carrier Systems,1, (1984), 39 - 90.
10 B. Focher, A. Marzetti, V. Sarto, P.L. Baltrane, and P. Carmitti, Cellulosic

materials : structure and enzymatic hydrolysis relationships, J. Appl. Polym. Sci., 29,
(1984), 3329 - 3338.
11 H. Fessi, J.P. Marty, F. Puisieux, and J.T. Carstensen, Square root of time
dependenceof matrix formulations with low drug content, J. Pharma. Sci., 71, (1982),
749 - 752.
12 E. Touitou and M. Donbrow, Drug release from non-disintegrating hydrophilic

matrices sodium salicylate as model drug, Int. J. Pharm., 11, (1982), 355 - 364.
13 M. Nicklasson, A. Brodin, and L.O. Sundelof, Studies of some characteristicsof
molecular dissolution kinetics from rotating discs, Int. J. Pharm., 23, (1985), 97 - 108.
14 E. Teillaud et Y. Pourcelot-Roubeau,Validation dun nouveaumodele de liberation
in-vitro, Labo. Pharma. Probl. Tech., 32, (1984), 279 - 283.
15 N.A. Peppas, R. Gumy, E. Doekler, and P. Buri, Modelling of drug diffusion
through swellable polymeric systems,J. Membrane Sci., 7, (1980). 241 - 253.
Sec. 10.51
Spherical
Drug-Eudragit
255
16 C. Brossard, D. Lefort des Ylouses, D. Duchene,F. Puisieux, and J.T. Carstensen,

Dissolution of a soluble drug substancefrom vinyl polymer matrices, 3. Pharm. Sci.,
72,(1983),
162 - 169.
N.A. Peppasand N.M. Franson, The swelling interface number as a criterion for
17
prediction of diffusional solute releasemechanismsin swellable polymers, J. Polym. Sci.
Phys. Ed., 21, (1983), 983 - 997.
D. Wouessidjewe, C. Brossard, J.J. Goujon, F. Puisieux, and J.P. Devissaguet,
Preparationde microgranulesde trinitrine a liberation prolongeepar pelliculage en milieu
aqueux, Labo Pharma.Probl. Tech., 32, (1984), 772 - 777.
18
19 G.S. Kumar, V. Kalpagam, and U.S. Nandi, Biodegradation of gelatinpolyethylacrylate copolymers, J. Appl. Polym. Sci., 29, (1984), 3075 - 3085.
20 S. Polowinski, L. Szosland,J. Szumilewicz, A. Polowinska and A. Pierzchlewska.
Controlled-releasemodel systems.Brit. Polym. J. 23, (1990), 241 - 244.
21 S. Goto, T. Tsujuyama, M. Kawata and N. Suzuki. Preparation of
biophamraceuticalevaluation of gelatin microcapsulesof piretadine. J. Control. Release,
1, (1985), 291 - 300.
22 E.O. Batyrbekov, S.A. Moshkevicxh, L.B. Rukhima, R.A. Bogin and B.A.
Zhubanov. Some fields of biomedical application of polyurethanes.Brit. Polym. J. 23,
(1990) 273 - 276.
23 R. Bawa, R.A. Siegel, B. Marasca, M. Karel and R. Langer. An explanation for
the controlled releaseof macromoleculesfrom polymers. J. Control. Release,1, (1985),
259 - 267.
24 L. Yean, C. Bunel and J.P. Vairon. Reversible immobilization of drugs on a
hydrogel matrix. Diffusion of free chloramphenicol from poly (2- hydroxyethyl
methacrylate)hydrogels. Makrom. Chem. 191, (1990), 1119 - 1129.
25
P. Buri. Formulation, Caracteristiqueset Interet des formes galtniques orales a
256
Drug delivery
from dosage forms
[Ch. 10
liberation prolongee de type matrice. Bull. Chim. Farm. 123, (1984), 453 - 464.
26
S.S. Davis. The design and evaluation of controlled release systems for the
gastrointestinal tract. J. Control. Release. 2, (1985), 27 - 38.

27
A.F. Stockwell, S.S. Davis and S.E. Walker. In vitro evaluation of alginate gel
systems as sustained release drug delivery systems. J. Control. Release. 3, (1986),

167 - 175.
28
N.A. Peppas et C. Bindschaedler. Les dispositifs B liberation controlte pour la
delivrance des principes actifs mtdicamenteux. IV. Systemes a gonflement control&. STP
Pharma. 2, (1986), 38 - 46.
29
T. Higuchi, Rate of release of medicaments from ointment bases containing drugs
in suspension, J. Pharm. Sci., 50, (1961), 874 - 875.

30
W.I. Higuchi, N.F.A. Ho, and H.P. Merkle, Design of oral drug delivery
systems : past, present and future, Drug. Dev. Ind. Pharm., 9, (1983), 1227 - 1239.
31
P.I. Lee, Diffusional release of a solute from a polymeric matrix. Approximate
analytical solutions, J. Membr. Sci., 7, (1980), 255 - 275.

32
H.B. Hopfenberg and K.C. Hsu. Swelling-controlled
constant rate delivery
systems, Polym. Engng. Sci., 18, (1978), 1186 - 1191.

33
N.A. Peppas. Release of bioactive agents from swellable polymers : theory and
experiments, in Recent Advances in Drug Delivery Systems, eds. J.M. Anderson and
S.W. Kim, NY : Plenum Publishing Corp., (1984), pp. 279 - 289.
34
N.A. Peppas and S. Segot-Chicq. Les dispositifs a liberation controlee pour la
delivrance
des principes
actifs medicamentaux.
Modelisation
des mecanismes
diffusionels, STP Pharm. 1, (1985), 208 - 216.

35
D. Messadi and J.M. Vergnaud. Simultaneous diffusion of benzyl alcohol into
plasticized PVC and of plasticizer from polymer into liquid. J. Appl. Polym. Sci., 26,
Sec. 10.51
Spherical
Drug-Eudragit
257
(1981), 2315 - 2324.

36 D. Messadi, J.L. Taverdet, and J.M. Vergnaud. Plasticizer migration from
plasticized PVC into liquids. Effect of several parameters on the transfer. I and EC. Prod.
Res. Develop. 22, (1983), 142 - 146.
37
J.L. Taverdet and J.M. Vergnaud. Study of transfer process of liquid into and
plasticizer out of plasticized PVC in using short test. J. Appl. Polym. Sci, 29, (1984),
3391 - 3400.
38
J.L. Taverdet and J.M. Vergnaud. Modelization
of matter transfers between
plasticized PVC and liquids in case of a maximum for liquid-time curves. J. Appl.
Polym. Sci., 31, (1986), 111 - 122.
39
A. Aboutaybi, J. Bouzon, and J.M. Vergnaud. Modelling the process of drying of
plasticized PVC previously immersed in n-hexane. Europ. Polym. J., 25, (1989),
1013 - 1018.
40
A. Aboutaybi, H. David, B. Touchard, J. Bouzon, and J.M. Vergnaud. Drying of
plasticized PVC previously immersed into liquids. in Drying of Solids, ed. Mujumdar,
Sakri, (1990), chapter 13.
41 J.M. Vergnaud, in Liquid transport processes in polymeric materials. Modelling
and industrial applications, Prentice Hall ed., 1991.
42
J.M. Vergnaud, in Drying of polymeric
and solid materials. Modelling
and
industrial applications, Springer Verlag ed., 1992.

43 A. Aboutaybi, J. Bouzon and J.M. Vergnaud. Study of the preparation of
plasticized PVC with a surface with barrier properties. J. Vinyl Technol., 12, (1990),
58 - 64.
44
A. Aboutaybi, J. Bouzon and J.M. Vergnaud. Drying at room temperature of
plasticized PVC previously immersed in liquids. Modelling and experiments. Europ.

Polym. J., 26, (1990), 285 - 291.
258
Drug delivery
[Ch. 10
from dosage forms
45
A. Senoune and J.M. Vergnaud. Preparation of plasticized PVC coatings with low
matter transfers by using n-pentane. Europ. Polym. J., 28, (1992), 311 - 314.
46
Y. Khatir, J. Bouzon and J.M. Vergnaud. Liquid sorption by rubber sheets and
evaporation, models and experiments. Polym. Testing. 6, (1986), 253 - 265.

47
Y. Khatir,
J. Bouzon and J.M. Vergnaud.The
kinetics of absorption and
desorption of liquid by a rubber annulus using a model and short-term tests. Plast.
Rubber. Process. Applic. 9, (1988), 53 - 58.
48
A. Drain, C. Chaumat, M. Rollet, J.L. Taverdet, and J.M. Vergnaud.
Model of matter transfers between sodium salicylate-Eudragit matrix and gastric liquid.
Int. J. Pharmac., 27, (1985), 233 - 243.
49
I. Malley, J. Bardon, M. Rollet, J.L. Taverdet, and J.M. Vergnaud.
Modelling of controlled drug release in case of Carbopol-sodium salicylate matrix in

gastric liquid. Drug Develop. Ind. Pharm., 13, (1987), 67 - 81.
50
J.Y. Armand, F. Magnard, J. Bouzon, M. Rollet, J.L. Taverdet, and
J.M. Vergnaud. Modelling of the release of drug in gastric liquid from spheric galenic
forms with Eudragit matrix. Int. J. Pharm., 40, (1987), 33 - 41.
51
M. Saber, F. Magnard, J. Bouzon and J.M. Vergnaud. Modelling
of matter
transfers in drug-polymer device used as a galenic form. J. Polym. Engng., 8, (1988),

295 - 314.
52
G. Hecht, J.E. Christian and G.S. Banker. In vivo pharmacodynamic evaluation of
oral dosage forms by whole body liquid scintillometry. J. Pharm. Sci., 55, (1966), 678.
53
J.M. Vergnaud. Scientific aspects of plasticizer migration from plasticized PVC
into liquids. Polym. Plast. Technol. Engng., 20, (1983), 1 - 22.

54
J. Crank, in The Mathematics of Diffusion,
Chapt. 4.
Clarendon Press, Oxford, 1986,
Sec. 10.51
55
Spherical
Drug-Eudragit
259
Carbopol Resins, FB Goodrich. Information sheet.
56
D. Messadi, J.L. Taverdet and J.M. Vergnaud. Plasticizer migration from

plasticized PVC into liquids. Effect of several parameters on the transfers. I and EC.
Prod. Res. Develop. 22, (1983), 142-146.
57
J.L. Taverdet and J.M. Vergnaud. Study of transfer process of liquid into and
plasticizer out of plasticized PVC in using short test J. Appl. Polym. Sci., 29, (1984),
3391-3400.
58
J.L. Taverdet
and J.M.Vergnaud. Modelization
of matter transfers between
plasticized PVC and liquids in case of a maximum for liquid-times curves J. Appl.
Polym. Sci., 31 (1986), 111-122.
59
A. Manoussalci, J.L. Taverdet and J.M. Vergnaud. Transfer of liquids between
plasticized PVC and acetic acid. Europ. Polym. J. 21 (1985) 967-972.
11
Drying of dosage forms made of a
drug dispersed in a polymer matrix
11 .l
INTRODUCTION
As already shown in the previous chapter (chapter lo), ordinary oral dosage forms
consisting of the drug and excipients are usually dissolved rather quickly in the gastric
liquid. The drug is thus rapidly liberated and its concentration in the gastric liquid builds
up to a high concentration. Because of the absorption of the drug through the stomach
membrane, the drug concentration falls exponentially until the next dose. The result of this
process is an undulating concentration pattern of the drug firstly in the gastric liquid, and
thus in the blood and tissues. As high concentration of the drug alternates with low
concentration, the optimal therapeutic level is only briefly present (1).
In the last few years, efforts have been directed to the development of new oral
dosage forms for safer administration of drugs than the conventional dosage forms.
Special attention has been given to controlling the release of the drug by means of
monolithic devices where the drug is dispersed in a polymer matrix (2-7).
There are various techniques to prepare these monolithic devices, and three of them
are worth noting :
1 - By compressing the components in powder form at a given temperature (8,9).
2 - By melting the mixture of the drug and polymer, or at least the polymer matrix
(10,ll).
3 - By transforming the polymer into a viscous paste after addition of a liquid in which
the drug is not soluble. The paste can thus be easily shaped into beads at room temperature
(12-14).
262
Drying of dosage forms
[Ch. 11
Of course, the above techniques exhibit some advantages and drawbacks. The last
humidity technique does not necessitate high pressure and thus high temperature for
shaping the forms, but the inconvenience appearswith the drying of the dosage form.
The main purpose in this chapter is to study the process of drying of these dosage
forms, when the polymer is Eudragit, a biocompatible polymer, and the drug is sodium
salicylate. Spherical beads are used for the purpose, but the method can be applied to other
shapes. The process of drying a polymer or a solid is rather complex. As shown in
previous papers with elastomers of various shapes such as thin sheets (15), cylinders of
finite length (16), long hollow cylinders (17) or flat hollow cylinders as annuli (18), the
process is controlled by diffusion through the polymer and evaporation from the surface.
Another aim in this study is to build up mathematical and numerical models able to
describe the process. Mathematical models can be used with a constant diffusivity and for
simple initial and boundary conditions, one of them being that the initial distribution of the
liquid in the polymer is uniform (19). Numerical models are of help whatever the
conditions, e.g., with a concentration-dependent diffusivity, various initial and boundary
conditions (20,21). In all cases, the rate of evaporation is constantly equal to the rate at
which the liquid is brough to the evaporating surface by internal diffusion. This rate of
evaporation is also proportional to the difference between the actual concentration of liquid
on the surface and the concentration required to maintain equilibrium with the surrounding
atmosphere.
Some emphasis is placed upon the effect of parameters such as the radius of the
bead (22) and the temperature of drying (23). Temperature is a very important factor, as it
acts not only on the rate of evaporation of the liquid out of the surface but also on the rate
of diffusion of the liquid throughout the polymer. Generally, the diffusivity varies with
temperature in the same way as the rate of evaporation, i.e., in an exponential form by
following the Arrhenius equation. In the same way as for Gas Chromatography or for
Calorimetry, a programmation of temperature may be very useful to dry a material,
especially when an increase in temperature decreasesthe viscosity of the humid material
(24).
Very often, the drying process is carried out in a surrounding atmosphere of infinite
volume, and the vapour pressure is constant to zero. Sometimes, it may be of interest to
consider a process of drying with a surrounding atmosphere of finite volume (25).
Sec. 11.21
Drying in a surrounding
11.2 DRYING
ATMOSPHERE
DOSAGE
FORMS
IN
OF INFINITE VOLUME
atmosphere
of infinite
volume
263
SURROUNDING
It is difficult to say that the volume of the surrounding atmosphereis infinite in

physical scienceor in pharmacy.In fact, when the volume of the surrounding is so large
with regard to the volume of the dosageform, the pressureof the vapour is very low, and
the volume of the surrounding may thus be consideredas infiite. When the surrounding
atmosphereis constantly removed and replaced by a new atmosphere,the value of the
vapour pressureis constantlyvery small. Thesetwo casesarecoveredin this
Section 11.2.
11.2.1
THEORETICAL
The plan in this subsectionsis as follows : assumptions,mathematical treatment,

andnumerical analysis.
ASSUMPTIONS
The following assumptionsaremade in order to build up the models :

(i) The processof desorption of the liquid is controlled by transient diffusion within the
polymer and evaporationout of the surface.
(ii) The rate of evaporationis proportional to the difference of the actual concentrationof
liquid on the surface and the concentrationwhich is at equilibrium with the surrounding
atmosphere.
(iii) The rate of evaporation is also constantly equal to the rate at which the liquid is
brought to the evaporatingsurfaceby internal diffusion.
(iv) The beadis sphericalin shape,and its dimensionsdo not changeduring the process.
(v) The diffusivity is constant,as it is found from experiments.
(vi) The concentrationof the liquid in the beadis uniform at the beginning of the process.
This condition is necessaryfor the mathematicaltreatment.
The transfer of liquid within the beadis expressedby Ficks equationfor a sphere:
D a
(11.1) ac
at =fz
iN2
TF
1 1
r2
with a constantdiffusivity
264
[Ch. 11
The rate of evaporationon the surfaceis defined by the surfacecondition :
(11.2)
-D.(g),
= h(CR- Cd
where CRis the actual concentrationof liquid on the surface,

Cesis the concentrationrequired to maintain equilibrium with the surroundingatmosphere,
h is the coefficient of matter transferon the surface.
As the sphere is initially at the uniform concentration C, and the diffusivity is
constant,the required solution of the aboveequationsis :
(11.3)
cr.,-ceq
tin-
eq
where the &, s are the roots of

p,. cot p, + L - 1 = 0
(11.4)
with the dimensionlessnumber L

(11.5) L = g
At the centre of the sphere, as sin r / r + 1 in this position, the concentration is
given by :
(11.6)
The amountof liquid leaving the sphereafter time t, M,, is expressedas a fraction of
the correspondingquantity after infiite time M, :
(11.7)
Me-M,
M
= 2
n=l
exp
m&D,
R2
Sec. 11.21
atmosphere
of infinite
volume
265
Some roots of eqn (11.4) are given in Chapter 4, for various values of L. Other
valuesof p,, can be obtainedby interpolation.
NUMERICAL
ANALYSIS
A numerical model is also built up, in order to resolve the problem when the
diffusivity is concentration-dependent
or the initial concentrationof liquid is not uniform.
As shown in Chapter 8, the concentration after elapseof the increment of tune At,
CN, is expressedin terms of the previous concentrationsat the sameand adjacentsplaces,
by evaluatingthe matter balancein variouspositions.
The following equationsarethus obtained:
Ar I2 < r c R - Ar / 2, and a constantdiffusivity
WlTHIN THE SPHERE, WlTH
(11.8) CN, = C,+

Mn
I
1 . (n- 0.512.(C,_,-C,J-(n+0.5)2.
+12 i
1
and
(11.9)
r = n.Ar
R= N. Ar
WITHIN THE SPHERE, WlTH A CONCENTRATION-DEPENDENT DlFFUSlVlTY
(11.10) CN, = C,+ (Arf* F2+&)
[-n-0.5+
n+m I
with the function J

(11.11)
J, = n2. Dn- (Cn - 0.5 - Cn + OS)
where D, is the diffusivity at position n. Ar and time t.

AT THE CENTRE OF THE SPHERE, WlTH A CONSTANT DIFFUSIVITY
The new concentrationat the centreof the sphereis given by :
(Cn-Cn+r)
266
(11.12) CNO = Co-#,-
[Ch. 11
C,)
where M is a dimensionlessnumber
(Ad*
(11.13) M = D. At
AT THE CENTRE OF THE SPHERE, WITH A CONCENTRATION-DEPENDENT DIFFUSIVITY
The new concentration at the centre of the sphere is expressed in terms of the
function J :
24. At
(11.14) CNO = CO- JO.5
br)*
with Jo.sgiven by :
(11.15) Jo,5 = f. Do.,. (Co-C,)
ON THE SURFACE OF THE SPHERE, WITH A CONSTANT DlFFUSlVlTY
The new concentrationon the surfaceCNN is given by the following equation

(11.16) CN, = CN+
N - 0.5
l. (cN-+N)-
N3-(N-0.5)3M
N
N3-(N-0.5)3
h. At
.-(CN-C
Ar
by making the simple assumption:

CNN - 0.25 - CN - 0.25 = CNN - CN
ON THE SURFACE OF THE SPHERE, WITH A CONCENTRATION-DEPENDENT DlFFUSIVllY
The new concentrationon the surfaceafter elapseof time At is given by :
(11.18) CNN = CN+
3. At
(Ar)*. [N3- (N - o,5)3] JN-0.5- Ar. I:-
. .
(: fL.5)3] (-
cq1
Sec. 11.21
atmosphere
of infinite
volume
267
by making the simple assumption (11.17).

AMOUNT OF MAllER
LOCATED IN THE SPHERE
The amount of liquid remaining in the sphere after time t is evaluated by integrating
the concentration at time t with respect to space.
(11.19)
M, = 4x.
r2. C,,. dr
By using the finite differences, this equation becomes :
(11.20)
Mt
= $
4 7F(Ar)3
+ y
(n2+&)C,t+
n=l
[N3-(N-0.5)3.
I[&+&+
The amount of matter transferred after time t can be calculated by integrating the rate
of matter transfer on the surface with respect to time :
(11.21)
M, = 4 IT. N2. (Ar)2. 2
h(CN- Cq). A t
11.2.2
EXPERIMENTAL
Two parts are of interest : i)
the preparation of the dosage forms and the
characteristics of the beads ; ii) the determination of the kinetics of drying.

PREPARATION
POLYMER
OF DOSAGE
MATRIX.
FORMS
Eudragit RL, a copolymer of dimethylaminoethyl
acrylate and
ethylmethacrylate (Rohm Pharma) of MW = 150,000, in powder form, is the polymer

matrix.
DRUG.
The drug is sodium sacylate, in powder form.
268
[Ch. 11
The polymer and drug are intimately mixed, and the mixture is transformed in a
viscous paste by addition of ethanol. The paste can be pressed into spherical beads of
different sizes.
These humid beads can be characterizedby their size (weight and radius) and the
amountof ethanolto be evaporated.
Table 15.1 - Characteristicsof the beads.
Total weight (mg)
Radius (cm)
Alcohol (mg)
% liquid (W / W)
697.2
0.494
120
17.2
352.3
0.39
65
18.5
215.5
0.27
50
19.9
KINETICS OF DRYING
The kinetics of drying is determinedby following the weight, when the beadsarelet
dry in open air at constanttemperature(20 C).
The rate of evaporationof the liquid can be obtainedby two ways :
(i) by following the weight of the liquid which evaporatesin a flat flask of given area
under the sameconditions of temperatureand stirring of air.
(ii) by determining the initial rate of drying of the polymer when the concentrationof the
liquid is uniform and known.
Generally, the secondmethod is used, as the rate of evaporationof a liquid may be
different when it is in liquid stateor dispersedin a polymer (15).
The effect of the temperatureon the processof drying is determinedby drying the
beadsin a surrounding atmosphereof large volume, well stirred, and kept at a constant
temperature.The temperaturerangesfrom 20 to 60 C.
11.2.3 RESULTS WITH A CONSTANT TEMPERATURE
Following the method described in this book for studying the processesof matter
transfers,experimentsand modelling arecombined.
Two kinds of results are of interest :
- the one with the determination of the parameters such as the diffusivity and rate of
evaporation;
- the other is concernedwith the evaluation of the validity of the model by comparing the
kinetics of drying obtained by experimentsand by calculation with the help of the model
Sec. 11.21
atmosphere
of infinite
volume
269
and using the data.

Finally, an overview of all parameters which play a role in the process of drying is
attempted.
DETERMINATION OF THE PARAMETERS
The process of drying being controlled not only by the stage of evaporation but also
by the stage of diffusion, the diffusivity and rate of evaporation are two parameters of
interest.
DIFFUSIVITY
The diffusivity can be determined from the kinetics of drying either for short tune
experiments or for long tune experiments (Chapter 4).
The diffusivity is easily obtained by using short tune experiments and the squareroot of time dependance of the amount of matter transferred. But this method is of interest
only when the process of drying is controlled by diffusion. This means (section 4.2.2)
that the dimensionless number L = h. R / D is very high, at least higher than 10 - 20, the
coefficient of matter transfer on the surface h being very high. In this case, the tangent at
the origin of time is vertical. The diffusivity can thus be determined from the slope of the
straight line obtained by plotting the ratio M, / M, as a function of the squareroot of tune,
with the following relation :
for;
< 0.2-0.3
The diffusivity can also be determined by using long time experiments, when the
ratio M, / M, > 0.7. In this case, the series in eqn (11.7) converges very fast, and the
first term becomes preponderant, leading to :
P2
= -Dt
R2
+ Ln
By plotting the first term of this equation as a function of time, a straight line is obtained
2
whose slope is equal to the quantity PI. D/R* (Fig. 1.5.1). The constant PI being a function of the
diffusivity as well as of the coefficient of matter transfer h, as shown in eqns (11.4) and
270
50
[Ch. 11
100
150
200
Time(h)
Fig. 11.1. Log (M, - Mt / MJ as a function of time, for a bead of weight 697.2
mg, at 20 C.
(11 S), an iterative calculation must be made to obtain the diffusivity.

RATE OF EVAPORATION
Generally, the rate of evaporation can be determined by two means (20,21) :

- the one, by evaporating the pure liquid under the same conditions of temperature and
vapour pressure as those which exist during the process of drying of the bead. Of course,
special attention must be given to the shape of the flask which must be flat and full of
liquid, and to the motion of the surrounding air.
- the other, by using the initial rate of drying of the beads, when the concentration of
liquid is still constant and uniform, and therefore during this short period when the
process may be assumed to be controlled by evaporation.
Generally, the second method gives better results, as interactions may exist between
the polymer and the liquid.
The values of these parameters are shown in Table 11.2.
A better accuracy is generally obtained for the rate of evaporation than for the
diffusivity. A precision of 10 % is reasonable for the diffusivity.
VALIDITY OF THE MODEL OF DRYING DOSAGE FORMS
The validity of the model is tested by comparing the kinetics of drying of various
Sec. 11.21
atmosphere
of infinite
volume
271
Table 11.2 Diffusivity and rate of evaporation at 20 C

Diffusivity
D (cm* /s)
4x 10-s
Rate of evaporation (g / cm*. s)
2x 104
Coefficient of matter transfer h (cm / s)
2.5 x 10-4
Density of ethanol (g / cm3)
0.789
f 0.5 x 10-8
Fig 11.2. Kinetics of drying at 20 C for a bead of 697.2 mg and radius 0.494
cm.+ : experimental.- : calculated
beads when they are obtained either from experiments or calculation. This validity of the
model, as well as the accuracy of the parameters, e.g., diffusivity, rate of evaporation,
radius of the bead, is illustrated in Figs 11.2 - 11.4 for various sizes of the beads.
Some conclusions can be drawn from these curves :
(i) The diffusivity being constant during the whole process and the initial concentration of
the liquid uniform throughout the beads, the analytical solution can describe the kinetics of
dryh3.
(ii) The process is controlled by diffusion of liquid within the solid and evaporation out of
the surface. The rate of drying is thus very high at the beginning of the process and
decreasesin an exponential way with time as the operation proceeds.
272
[Ch. 11
w
0
50
100
150
200
250
Fig 11.3. Kinetics of drying at 20 C for a bead of 352.3 mg and radius 0.39
cm.+ : experimental.- : calculated
(iii) Near the end of the process of drying, the rate becomes very low, and an asymptotic
limit is attained when the operation approaches completion.
(iv) The process of drying is described very well by the model, taking into account not
only the radial diffusion of the liquid through the bead but also the evaporation from the
surface. Moreover, the diffusivity may be concentration-dependent, and the initial and
boundary conditions can be complex.
EFFECT OF THE PARAMETERS ON THE DRYING PROCESS
Three parameters are of concern, and their effect on the process is examined
successively : the radius of the bead, the diffusivity and the rate of evaporation.
RADIUS OF THE BEAD
The size of the bead plays an important role not only on the stage of evaporation but
also on the stage of diffusion. As the area of the external surface of the beads per unit
mass of solid is inversely proportional to the radius of the beads, a lower value of the
Sec. 11.21
50
100
150
atmosphere
of infinite
200
250
volume
273
Fig 11.4. Kinetics of drying at 20 C for a bead of 25 1.5 mg and radius 0.27 cm.+ :
experimental.- : calculated
radius of the bead is thus responsible for a faster rate of drying.

On the other hand, an increase in the radius of the bead is followed by an increase in the
time of diffusion. This fact comes from the existence of the dimensionless number D t / R2
in the exponentials, where the time necessary for diffusion to proceed is proportional to
the square of the radius. Some profiles of concentration of liquid expanded through a bead
are drawn for various values of the radius ranging from 0.2 to 0.6 cm, at 20 C, in Fig.
11.5.
DlFFUSlVllY
It is not possible to predict by calculation the value of the diffusivity
of a liquid
through a polymer, in spite of many attempts and interesting studies. This value must be
determined by experiments. However, general rules exist for the variation of the
diffusivity :
(i) The diffusivity depends on the nature of the liquid and polymer (20 - 22,26).
274
Oh
Oh
Oh
[Ch. 11
M'S
Fig 11S. Concentration distribution through dosage forms of various radii at 20 C :

a : 0.2 cm - b : 0.4 cm - c : 0.6 cm.
(ii) The diffusivity for a given polymer and liquid increases with temperature, following
an Arrhenius law with a constant energy of.activation above the temperature transition
when the polymer is in the elastomeric state (23,26 - 28).
RATE OF EVAPORATION
The rate of evaporation is a well-known parameter, but it is rather complex to

determine its value. In the case where the surrounding atmosphere is of large volume, the
rate of evaporation is proportional to the vapour pressure of the liquid. As this vapour
pressure is expressed in terms of temperature by the Clausius-Clapeyrons law, the
diffusivity
increases with temperature in an exponential way. However, the rate of
evaporation of a liquid depends also on the rate of stirring of the surrounding atmosphere.
11.2.4 EFFECT OF TEMPERATURE

The following results are worth noting (23) :
Sec. 11.21
atmosphere
of infinite
volume
275
;x103tk-'I
-6
-15,
-16,
+LnDT
Ln k,
Fig 11.6. Temperaturedependenceof the diffusivity andrate of evaporationof

ethanol.
- the temperature-dependence
of the diffusivity.
- the validity of the model with various temperatures.
- the effect of temperatureon the concentrationdistribution of the liquid within the beads.
TEMPERATURE-DEPENDENCE
OF THE PARAMETERS
The diffusivity is determined under isothermal conditions at various given

temperatures,by using the two methodsdescribedin section 11.2.3 : short time testswith
eqn (11.22), and long time experimentswith the eqn (11.23) when the rate of evaporation
is previously determined.
The diffusivity is often expressedin terms of temperatureby an Arrhenius equation:
(11.24) D, = Do. exp -&
I
.I
where D, is the value of the diffusivity at temperatureT expressedin Kelvins,
Do is a constant,
E is the activation energy,andR is the ideal gas constant.
[Ch. 11
276
By plotting the logarithm of diffusivity against the reciprocal temperature, a straight

line is obtained, and the activation energy can easily be determined from it slope. (Fig.
11.6).
50
100
200
150
250
Fig 11.7. Kinetics of drying of dosage forms at various temperatures : 20 - 40 60 C.- : calculated with the models+ : experiments
The rate of evaporation of ethanol out of the bead is obtained from the slope of the
kinetics of evaporation at the earlier stage of drying when t + 0, as shown in section
11.2.3.
The values of the diffusivity and the rate of evaporation are shown in Table 11.3 for
various temperatures.
Table 11.3 Temperature-dependence of diffusivity and rate of evaporation, for ethanol
with Eudragit RL beads
Temperature (C)
20
40
60
D x 10 (cm2 / s)
0.6
1.6
4.0
hx 104(cm/s)
2.5
7.8
19.8
Sec. 11.31
Drying with a programmed
temperature
277
VALIDITY OF THE MODEL AT VARIOUS TEMPERATURES
From the experimental and calculated kinetics of drying drawn in Fig. 11.7, good
superimposition of all curves can be appreciated, proving the validity of the numerical and
mathematical models. The analytical solution can also be used in this case, as the
diffusivity is constant and the initial concentration of liquid within the bead is uniform,
The effect of temperature on the rate of drying can also be clearly appreciated in Fig.
11.7. Of course, the higher the temperature, the higher the rate of evaporation. However,
a drawback appears, due to the fact that the humid bead is plastic, and that the viscosity of
the material decreaseswith temperature.
CONCENTRATION DISTRIBUTION OF THE LIQUID
The concentration distributions of the liquid are drawn in Fig. 11.8 during the
process of drying of a bead of radius 0.4 cm for the three temperatures : 20,40 and 60 C.
The following conclusions can be drawn from these curves :
(i) The effect of temperature on the kinetics of drying is of great importance as already
shown in Fig. 11.7. This is due to the fact that temperature acts not only upon the rate of
evaporation but also upon the diffusivity and the rate of diffusion.
(ii) Of course, the effect of temperature on the concentration distribution of the liquid
within the beads is also very significant. In fact, these curves give complementary
information, affording a further insight into the nature of the process.
(iii) As already shown in Table 11.3, the diffusivity
and rate of evaporation increase
exponentially with temperature.
11.3 DRYING
WITH A PROGRAMMED
TEMPERATURE
(24)
As shown in section 11.2.4, temperature plays an important role in the rate of

drying, since the rate of evaporation increases with temperature according to the ClausiusClaperons law, and since the diffusivity of the liquid also increases with temperature
exponentially. However, high temperatures cannot be used in order to reduce the time of
drying without taking precaution due to the plasticity of the humid polymer which is often
enhanced by increasing temperature, the result being a distortion in the shape of the final
product.
The first objective in this section is to develop the process of drying of dosage
forms by using a programmed temperature system, the heating rate being constant or not.
The programmed temperature system is defined in order to address the two problems of
278

r
Oh
[Ch. 11
Oh
Oh
>
Fig 11.8. Concentration distribution of the liquid within a dosage form of radius
0.4 cm at various temperatures. a : 20 C - b : 40 C - c : 60 C.
how to reduce the time of drying, and how to achieve this purpose without modifying the
shape of the material.
The second purpose is to build up a numerical model able to describe the process
and to predict the best operational conditions for a defined objective. No analytical
solution exists in this case. (19-21). The model, based on a numerical method with finite
differences, already described in section 11.2.1, takes all the facts into account : the
diffusion through the solid, the evaporation out of the surface. The temperaturedependency of the diffusivity and of the coefficient of matter transfer on the surface must
be also considered :
(12.24)
and
Sec. 11.31
temperature
279
111.25) hT = ho exp
where Do, ho, E and H are constants to be determined from the measurements of the
diffusivity and coefficient of matter transfer at the surface, at various temperatures,
and DT and hT are the values of diffusivity and coefficient of matter transfer at the surface
at temperature T.
The parameters are especially examined, the rate of heating of the heating system
coupling constant temperature and programmed temperature, as well as the radius of the
spherical dosage form.
11.3.1
THEORETICAL
ASSUMPTIONS
The same assumptions as those in section 11.2.1 are made, for the process of
diffusion and evaporation. The numerical model is thus the same, with the same
equations.
Some additional
assumptions are made to consider the programmation
of
temperature :
(i) The diffusivity
and coefficient of matter transfer on the surface are temperature-
dependent, and this dependency follows exponential behaviour. (eqns 11.24 and 11.25).
(ii) The rate of heating is constant, and the heating system follows successively isothermal
conditions and programmation of temperature.
(iii) As the rate of heating is rather low and the dimension of the dosage form is small, the
temperature is always uniform within the solid, this temperature being constantly equal to
that of the heating system (21,29).
(iv) In the present case, the diffusivity depends only on temperature, but the model can
work with a concentration-dependent diffusivity.
NUMERICAL MODEL
The numerical model described in section 11.2.1 is used, while the temperature of
the bead is a function of time with the constant heating rate b :
(11.26)
T, = Tto+b(t-to)
280
11.3.2
[Ch. 11
EXPERIMENTAL
The kinetics of drying is determined by recording the weight of the bead, as well as
the temperature. The volume of the surrounding atmosphere is so large that the pressure of
vapour is negligible.
These thermogravirnetric studies are carried out by using a Ugine Eyraud G 70
apparatus (SETARAM,
Lyon, France) equipped with a PRT 540 temperature
programming system. This apparatus can work either under dynamic conditions with
constant heating rate, or under isothermal conditions.
11.3.3 RESULTS
WITH A PROGRAMMED
TEMPERATURE
The following results are considered :

(i) Determination of the parameters of interest : the diffusivity and rate of evaporation of
the liquid, as well as their variation with temperature.
(ii) As the validity of the model has already been tested in section 11.2, only a brief
description is given here.
(iii) The effect of the rate of heating on the process of drying. Moreover, some more
complex temperature-time histories are also considered.
(iv) The effect of the radius of the beads.
DETERMINATION OF THE PARAMETERS
The parameters of interest, such as the diffusivity and coefficient of matter transfer
at the surface, are determined from experiments carried out under isothermal conditions at
various temperatures : 20,40,60 C.
The values of the diffusivity and coefficient of matter transfer at the surface are
listed in Table 11.3 for the three temperatures.
The temperature dependence of the diffusivity
is obtained by considering an
Arrhenius expression while the coefficient of matter transfer at the surface is expressed in
terms of temperature by the Clausius-Clapeyron equation, the coefficient of matter transfer
being proportional to the pressure of vapour at equilibrium with the liquid (Eqns 12.24
and 12.25).
The values of the pre-exponential term, and of the constant E and H in these
equations are shown in Table 11.4.
DRYING WITH PROGRAMMED TEMPERATURE AND A CONSTANT HEATING RATE
Temperature can be programmed in many ways, but heating at a constant rate is the
Sec. 11.31
temperature
281
Table 11.4 Values of parameters

Do = 0.46 (cm2 / s)
E = 9,300 cal /mole.
b = 3,000 (cm / s)
H = 9,500 Cal/ mole.
most usual, as well as various temperature-time histories combining in successive
programming of temperature and isothermal conditions.
The programmed temperature technique displays a great advantage over the method
run under isothermal conditions in the sense that drying may begin at a rather low
temperature and end at a temperature as high as possible. This advantage is clearly
illustrated in Fig. 11.9 where the kinetics are drawn for the three sets of conditions :
isothermal conditions at 20 and 60 C (curves 1 and 2), with programmed temperature
from 20 to 65 C and a heating rate of 0,Ol C / min. (curves 3 with the temperature-tune
history described in curve 4).
As the coefficient of the matter transfer h increases with temperature, as well as the
diffusivity of the liquid within the dosage form, the process of drying is accelerated by
,
80
TimeIh 1
I
100
Fig 11.9. Kinetics of drying either under isothermal condition or with

programmed temperature.1 : 20 C - 2 : 60 C - (3) with heating rate of 0.01 C /
min - 4: withaheating history - radius = 0.39cm-Iv&
= 60mg
282
[Ch. 11
TIOCI
i0
15
Time (h)
10
20
30
40
50
Fig 11.10. Kinetics of drying with programmed temperature and a heating rate of
0.02 C / min - radius = 0.39 cm - M, = 60 mg.
using a programmed temperature system. In Fig. 11.9, the kinetic curve of drying
obtained with the programmed temperature is located between the isothermal kinetic
curves obtained at 20 and 60 C.
The following conclusions are worth noting :
(i) It is possible to have rather low rate of drying at the beginning of the process by using
a constant heating rate. This fact is especially of interest when the material of the dosage
form becomes more and more plastic when temperature increases, because of the presence
of the liquid in the solid.
(ii) Because of the heating rate, the time necessary to evaporate to dryness is only a little
longer with the programmed temperature process than with the constant temperature at
60C.
The heating rate is a relevant factor with the programmed temperature system, as
well as the initial and final temperatures selected for the heating stage.
The effect of the heating rate on the process of drying is determined by drying
samples of same size with various values of the heating rate :
O.OlC/min(Fig.
11.9),0.02C/min(Fig.
11.10)and0.03C/min(Fig
ll.ll).The
Sec. 11.31
temperature
283
-50
- 25
Time lh)
0
10
20
Fig 11.11. Kinetics of drying with programmed temperature and a heating rate of
0.04 C /min - radius = 0.39 cm - M, = 60 mg.
three kinetic curves of drying are drawn with their temperature-time histories. Some
conclusions are worth noting :
(i) The time of drying is reduced by using a higher value of the heating rate.
(ii) The time of drying is a complex function of temperature.
(iii) The rate of heat transfer by conduction through the bead is higher than the rate of
matter transfer. As the heating rate is rather low, the temperature within the dosage form
can be assumed to be uniform and equal to that of the heating system (21,29).
DRYING WITH TEMPERATURE-TIME HISTORIES
When the material of the dosage form is highly plastic, a rapid increase in
temperature is followed by an increase in plasticity which is responsible for distortion of
the shape at the beginning of the process when the proportion of liquid is still high. In this
difficult but very common case, the other two possibilities for improving the process are
of interest, in addition to using a lower heating rate :
(i) by starting the process of drying under isothermal conditions during a period of time
long enough to make the dosage form less plastic.
284
20
40
[Ch. 11
60
80
100
Fig 11.12. Kinetics with various temperature-time histories. (1) 20 C for 5 h

followed by a heating rate of 0.01 C /min from 20 to 65 C. (2) 10 C for 5 h
followed by a heating rate of 0.01 C / min from 10 to 65 C.
(ii) by using a lower temperature at the beginning of the drying process.

The effect of these temperature-time histories on the process of drying is illustrated
in Fig. 11.12 where two heating histories are followed : curve 1, when this initial
temperature is 20 C ; and curve 2, when this initial temperature is 10 C. For these
curves, the other operational conditions are the same, with a heating rate of 0.01 C / min
starting after a period of time of 5 h at the constant initial temperature.
Of course, a lower value for the initial temperature is responsible for a longer time
of drying. But this lower initial temperature also brings with it the advantage of
evaporating the liquid at lower temperature. Thus, 25 % of liquid is evaporated up to 12
C in the second case with the initial temperature of 10 C, while the same amount is
evaporated at 25 C in the first case with the initial temperature of 20 C. In fact, the
advantage is still more important if one consider the concentration of liquid on the surface.
Sec. 11.31
temperature
285
EFFECT OF THE DIMENSION OF DOSAGE FORMS
As the process of drying is controlled not only by the rate of evaporation but also by
diffusion through the solid, the dimension of the dosage form plays an important role in
the kinetics of drying.
The effect of the radius of spherical dosage forms on the kinetics of drying is clearly
illustrated in Fig. Il. 13, with a heating rate of 0.01 C / rnin starting at 20 up to 65 C.
The effect of the radius of the bead on the process of diffusion is clearly shown by
considering the dimensionless number D t / R2 which appears in the kinetics of diffusion.
(11.27)
E
R2
As shown in this dimensionless number, the time necessary for a given amount of liquid
to diffuse is proportional to the square of the radius of the spherical bead. In the same
way, this time is inversely proportional to the value of the diffusivity.
20
GO
60
80
100
Fig 11.13. Kinetics of drying with dosage forms of various dimensions. Heating
rate = 0.01 C /min from 20 to 65 C - (1) radius = 0.1 cm - (2) = 0.2 cm (3) = 0.4 cm.
286
DRYING
IN
FINITE VOLUME
11.4
[Ch. 11
SURROUNDING
ATMOSPHERE
OF
As already shown, the process of drying of solids and especially of polymers is

controlled by diffusion of the liquid within the solid and evaporation from the surface.
Thus the diffusivity and rate of evaporation (or better the coefficient of matter transfer on
the surface) are of great concern in the kinetics of drying. However, a third parameter may
intervene sometimes, e.g., the volume of the surrounding atmosphere. This parameter can
play the role of a limiting factor in the sensethat it defines the fraction of the liquid which
is evaporated at equilibrium, after infinite tune.
The main objective in this section is to build a mathematical and a numerical model
capable of describing the process of drying of a dosage form in a surrounding atmosphere
of finite volume.
Another purpose is to determine the effect of the volume of the surrounding
atmosphere as compared with that of the dosage forms.
The effect of the coefficient of matter transfer on the surface is also examined, by
considering the dimensionless number h.R / D.
THEORETICAL
OF
11.4.1
ATMOSPHERE OF FINITE VOLUME
DRYING
IN
SURROUNDING
According to the assumptions, a mathematical or a numerical model can be built.

ASSUMPTIONS
The following assumptions are made :

(i) The process of drying is controlled by diffusion of the liquid within the solid and
evaporation out of the surface.
(ii) The volume of the surrounding atmosphere is finite, and thus the pressure of vapour in
the surrounding increases during the process of drying.
(iii) The diffusion of the vapour in the surrounding atmosphere is very high, so that the
vapour pressure is uniform.
(iv) The rate at which the liquid evaporates is constantly equal to that at which the liquid is
brought to the evaporating surface by internal diffusion.
(v) An analytical solution is obtained when : The diffusivity is constant, as well as the
coefficient of matter transfer on the surface and the initial concentration in the solid is
UIlifOlTIl
Drying a surrounding
Sec. 11.41
atmosphere
of finite volume
287
(vi) The numerical model must be used when the above assumptions shown in v do not
hold.
As shown in Chapter 4, the general equation of diffusion within the spherical bead
is :
(11.28)
= D.
with the constant diffusivity D.

The initial conditions expressing that the initial concentration of liquid in the sphere
is uniform, and that the volume V of the surrounding atmosphere is free of vapour are :
(11.29)
Ocr<R
t = 0
C = Ci
bead
c=o
surrounding
The boundary condition must express the fact that the rate of evaporation out of the
surface of the bead is constantly equal to the rate at which the liquid is brought to the
evaporating surface by internal diffusion :
(11.30)
t > 0
-47t . R2..ar
D g
v.g=
r = R surface
The solution of the problem is given in Chapter 4.

The concentration of the liquid at position r within the bead is given by :
sin %r r
(11.31)
c,,-c,
Ci,-C,
6(1+a)
R = -.
c
r n=l
9 + 9a + 4;. a
-R
2 G&y
where the G s are the non-zero roots of

(11.32)
tanq, =
3 %I
3+&q;:
and a is the ration of the evaporating substancelocated in the surrounding atmosphere and
288
[Ch. 11
in the bead
(11.33)
a =
3v
471:s. K
and K is the partition factor of the liquid in equilibrium in the sphere and the surrounding
atmosphere.
This ratio a is expressed in terms of the final fractional uptake of evaporating
substance by the surrounding atmosphere in the relation :
(11.34)
a
=------l+a
4 x R. Ci,
3M,
At the centre of the sphere, the equation of the concentration is a little more simple,
assinr/r
+ lwhenr + 0:
co
(11.35)
c,, - CM =
c
ci*- cca
n=l
6(1+4
qn
. v.
exp
2 smqn
9+9CX+ 4:. CX
The total amount of liquid which has evaporated after time t, M,, is expressed as a
fraction of the corresponding quantity after infinite time M,, by the relation :
(11.36)
Me.,- M,
M
c-3
= 2
n=l
6a(a+
1)
2. exp
9+901+ 4:. CC
Some roots of eqn (11.32) are given in Table 4.2 for various values of the ratio a.
NUMERICAL ANALYSIS (25)
There are many conditions for the analytical solution to be used : the diffusivity
must be constant, the initial concentration in the solid is uniform. In all other cases, the
process of drying of a solid bead in a surrounding atmosphere of finite volume can only
be described with the help of a numerical model taking into account all the facts.
The radius of the sphere of radius R is divided into N spherical membranes of equal
thickness Ar in the same way as shown in Chapter 8. The matter balance is evaluated at
many places during the increment of time At, by considering the transport of matter by
Sec. 11.41
atmosphere
of finite volume
289
diffusion and by evaporation.

The new concentration after time At is thus expressed in terms of the previous
concentrations by the following equations.
WITHIN THE SPHERE,
(11.8)
with a constant diffusivity
CN, = C,+
[(n-0.5j2.
(C,-,-C,)-(n+0.512.
(C,-C,,,il
with the dimensionless number M
(11.13)
(Ad2
M = D. At
AT THE CENTRE OF THE SPHERE,
(11.12)
with constant diffusivity
CN, = Grub
Cl)
(11.16)
CNN = CN+
with constant diffusivity
N - 0.5
3 A. bw-
c,)-
N3w(,_,.5)3.
y%-
N3- (N - 0.5)
by making the simple assumption :
(11.17)
CNN - 0.25 - CN - 0.25 = cNN
- CN
The concentrations at the same places with a concentration-dependent diffusivity are

shown either in Chapter 8 or in section 11.2.1.
CONCENTRATION AT EQUILIBRIUM Cq
The concentration C,, in eqn (11.16) is the concentration on the surface of the bead
required to maintain equilibrium with the surrounding atmosphere. As the vapour pressure
ceq1
290
[Ch. 11
increases during the process of evaporation, the concentration C,, increases also. The
concentration C,, can be related to the pressure of this vapour as follows.
By making the assumption that the vapour behaves as an ideal gas, the vapour
pressure P, is proportional to the number of moles of vapour nt at time t :
Pt. V = n,. R T
(11.37)
where R is the universal gas constant.

V is the volume of the surrounding atmosphere,
and T is the temperature (Kelvin).
Various relationships are used to relate the concentration C,, on the surface of the
solid at equilibrium with the vapour pressure P. Two of them are worth noting :
One represents a linear absorption isotherm :
C,, = k. P
(11.38)
The other when the absorption isotherm follows Langmuirs equation :

(11.39)
C,
= k$&
where k, k and b are constant depending on the vapour, the nature of the polymer and
temperature.
The amount of liquid evaporated at tune t is given by :
(11.40)
= nt. MW
where MW is the molecular weight of the liquid

Thus in the case of a linear absorption isotherm, the concentration at equilibrium C,,
is proportional to the amount of liquid evaporated :
(11.41)
C, = &$.
M, = K,. M,
AMOUNT OF LIQUID REMAINING IN THE SPHERE
The amount of liquid remaining in the sphere can be determined by integrating the
concentration of the liquid located in the sphere with respect to space, as shown already
Sec. 11.41
atmosphere
of finite volume
291
(11.2.1).
The amount of liquid evaporatedafter time t can be obtained by integrating the rate
of evaporationon the surfacewith respectto time, as shown in 11.2.1.
11.4.2
EXPERIMENTAL
MATERIALS
The dosageforms consist of sodium salicylate as the drug dispersedin a polymer

matrix madeof Eudragit RL (R&rn Pharma).The characteristicsof the beadsare :
radius = 0.38 cm. empty weight = 222 mg. initial weight of liquid (ethanol) = 49
mg
OPERATIONAL CONDITIONS
As the operation is carried out at 60 C, the values of the diffusivity and the
coefficient of matter transferare asfollows :
D = 4x 10-7cm2/s
11.4.3
h = 19.8 x lOA (cm/s)
RESULTS
The process is described either by the mathematical model or by the numerical

model. The simulation of the processof drying is capableof determining the role played
by eachfactor, and especiallyby the two parameters:
- the value of the coefficient of matter transferon the surface,
- the volume of the surrounding atmosphere,which is defined by the value of the constant
K, shown in eqn (11.41).
EFFECT OF THE COEFFICIENT OF MAlTER TRANSFER ON THE SURFACE
The effect of the coefficient of matter transfer at the surface h on the process is
determinedby varying the value of this parameter and keeping constantthe value of the
other parameters,the diffusivity, and the constantK, taken equal to 1.
Three kinds of information can be obtained by calculation : the kinetics of drying,
the profiles of concentration of liquid expanded through the dosage form, and the
concentrationof liquid on the surfaceduring the process.
with K, = 1
The kinetics of drying are drawn in Fig. 11.14 for various values of the coefficient
KINETICS OF DRYING,
292
[Ch. 11
Time(h)
*
0
10
20
30
LO
so
Fig 11.14. Kinetics of drying with dosage forms in a surrounding atmosphere of

finite volume. Effect of the coefficient of matter transfer on the surface.
D = 4x10-7cm2/s
- R = 0.38cm - K, = l(l)h==
- (2)h=19.8xlO-6
- (3) h = 19.8 x 1O-7 - (4)h=19.8xlO-*cm/s
of matter transfer on the surface h ranging from 19.8 x lo-* cm / s to infinity, while the
other parameters are kept constant to the values shown above. The kinetics for an infinite
coefficient of matter transfer are determined by using the mathematical model, and for
finite coefficients of matter transfer by using the numerical model. In fact, the value of
19.8 x 1OA cm / s for the coefficient of matter transfer on the surface h may be considered
as the infinite value, as the kinetics of drying obtained either with this value or with the
infinite value are very well superimposed. For this high value of the coefficient h, the
value of the dimensionless number h.R/D is around 2,000.
PROFILES OF CONCENTRATION OF LIQUID THROUGH THE DOSAGE FORM WITH K, = 1
It is often of interest to consider the profnes of concentration of the liquid expanded

within the spherical dosage form during the process of drying. The profiles of
concentration are calculated with the help of the numerical model for two values of the
Sec. 11.41
atmosphere
of finite volume
293
coefficient of matter transfer on the surface h (h = 19.8 x lo- cm / s, h = 19.8 x 10

cm / s), the first of them corresponding with the infinite value of h. (Fig. 11.15 and
11.16).
This information helps one to understand the process of drying, leading to the
following conclusions :
(i) Steep gradients of concentration are developed at the beginning of the process of
drying, with a very low value of the concentration on the surface. Comparison between
Fig 11.15. Concentration distribution through the dosage form at various times
during the process of drying, with the parameters : D = 3.5 x 10-7 cm* / s - R =
0.38cm- K,=l
- h=19.8x104cm/s.
the profiles drawn in the two Figs. 11.15 and 11.16 leads to the following statement : the
higher the rate of evaporation, the steeper the gradient of concentration during the early
stage of drying.
(ii) As drying proceeds, these gradients of concentration become less and less steep.
(iii) A typical result appears, with the concentration on the surface which decreases
strongly at the very beginning of the process and thus increases slowly up to the
equilibrium value. This result is quite different from that obtained with a very large
294
[Ch. 11
Fig 11.16. Concentration distribution through the dosage form at various times
during the process of drying, with the parameters : D = 3.5 x IO-7 cm2 / s - R =
0.38cm - K, = 1 - h=19.8x104cm/s.
volume of the surrounding atmosphere where the pressure of vapour remains very low
during the whole process.
CONCENTRATION
ON THE SURFACE-TIME
HISTORIES
WITH
K, = 1
As already shown with the profiles of concentration developed through the dosage
form, the concentration of liquid on the surface provides interesting information on the
process. It is very difficult to make measurements on the surface, as shown in a previous
study on plasticized PVC using the attenuated total reflectance technique (30), but the
concentration on the surface can be easily obtained by calculation.
The surface concentration as a fraction of the original value, is drawn during the
process of drying in a finite volume of air with K, = 1, for various values of the
coefficient of matter transfer on the surface (Fig. 11.17). Time is expressed in seconds as
the surface concentration varies very quickly, and it is given in a logarithim form in order
to expand the period over which the process is observed.
The following conclusions are of interest :
Sec. 11.41
atmosphere
of finite volume
295
Fig 11.17. Surface concentration-time histories during the process of drying in a

finite volume of air with K, = 1, with various values of the coefficient of matter
transfer on the surface h.D = 3.5 x 10-7 cm2 / s - R = 0.38 cm - K, = 1
(l)h=m
- (2) h = 19.8 x 1OA - (3) h = 19.8 x 10-5.(4) h = 19.8 x 10-e

- (5) h = 19.8 x lo- - (6) h = 19.8 x lo-8 cm / s
(i) When the coefficient
of matter transfer on the surface is infinite,
the liquid
concentration on the surface falls abruptly to zero as soon as the process of drying starts.
(ii) For a very high value of the coefficient h (19.8 x 1O-4 cm / s), the surface
concentration sharply declines during the first 10 s of the process.
(iii) When the coefficient of matter transfer on the surface is higher than the diffusivity, or
rather when the dimensionless number h.R / D is higher than 1, the surface concentrationtime histories pass through a minimum. The time at which this minimum is attained
increases when the value of the dimensionless number h.R / D is decreased.
(iv) This minimum studied in iii is not observed when the value of the dimensionless
number h.R / D is lower than 3 - 5.
(v) For long times, e.g., 10,000 s, the surface concentration reaches the value at
296
equilibrium.
[Ch. 11
In a surrounding atmosphere with Kt
= 1, the value of the surface
concentration at equilibrium is half the initial value, the liquid being equally disributed
between the solid and the surrounding atmosphere.
EFFECT OF THE VOLUME OF THE SURROUNDING ATMOSPHERE
Of course, the value of the volume of the surrounding atmosphere as compared with
the volume of the dosage form, is a parameter of great importance. This parameter is
expressed by the coefficient K, defined in the theoretical part. The effect of the coefficient
K, on the process of drying is evaluated with the help of the numerical model, by varying
the values of K, and keeping the other parameters constant at their initial values at time 0.
In the same way as for the effect of the coefficient h, the following results are
considered : the kinetics of drying, the concentration distribution through the spherical
beads, the surface concentration-time histories during the process of drying.
Time (hl
IO
20
30
LO
50
Fig 11.18. Kinetics of drying for various volumes of the surrounding atmosphere,
with K, between 0 and l.D = 3.5 x lo- cm2 / s - R = 0.38 cm - h = 19.8 x
104cm/s
Sec. 11.41
atmosphere
of finite volume
297
KINETICS OF DRYING WITH VARIOUS VALUES OF K,
It is of interest to know the effect of the volume of the surrounding atmosphere and
thus of the value of the parameter K, on the kinetics of drying. The kinetics of drying are
calculated for various values of the parameters K,. The effect of the value given to K, on
the process of drying is illustrated in Fig. 11.18, where the kinetics of drying are drawn
for various values of K, ranging from 0 to 1, when the bead is dried at 60 C.
The following conclusions are drawn :
(i) The volume of the surrounding atmosphere notably affects the kinetics of drying, and
especially the amount of liquid evaporated at equilibrium.
(ii) Of course, for K, = 0, the volume of the surrounding atmosphere is infinite. For K, =
1, the amount of evaporating substance is equally distributed at equilibrium in the bead
and the surrounding atmosphere.
(iii) In this case, the coefficient of matter transfer on the surface is very high, and the value
of the dimensionless number h.R / D is around 2,000. The process of drying in that case
is thus essentially controlled by diffusion.
Fig 11.19. Concentration distribution of liquid through the bead during the process
of drying, for K, = 0.5.D = 3.5 x lo- cm2 /s - R = 0.38 cm - h = 19.8 x
lOA cm/s.
298
[Ch. 11
CONCENTRATION DISTRIBUTION OF THE LIQUID THROUGH THE BEAD
As in many cases,the profiles of concentrationof the liquid expandedthrough the

bead are able to provide a fuller insight on the process. The profiles of concentration
developedwithin the bead during the processof drying are calculatedwith the help of the
numerical model for various values of the volume of the surrounding atmosphere.These
profiles are shown in Fig. 11.15 for Kr = 1, in Fig. 11.19 for K, = 0.5, and in Fig.
11.20 for K, = 0.1.
A few conclusionsareworth noting :
(i) Steepgradients of concentrationare developedduring the early stageof the processof
drying, with a low concentrationof liquid on the surface.
(ii) As drying proceeds, these gradients of concentration become less and less steep,
leading to a flat gradient at equilibrium.
Fig 11.20. Concentrationdistribution of liquid through the beadduring the process

of drying, for K, = O.l.D = 3.5 x 10V7cm2 /s - R = 0.38 cm - h = 19.8 x 10-4
cm/s.
(iii) Of course, the lower the value of the coefficient K,, the lower the value of the
concentration at equilibrium in the dosage form. This conclusion is obvious, if one
Sec. 11.41
atmosphere
of finite volume
50,
299
+l
t 0.5
25
Fig 11.21. Surfaceconcentration-timehistories for various volumes of the surrounding atmosphere,or rather for various values of K,.D = 3.5 x 10-Tcm2/ s
- R = 0.38 cm - h = 19.8 x 10-4 cm/s.
remembersthat the volume of the surroundingatmosphereis inversely proportional to the

value of the parameterK,.
SURFACE CONCENTRATION-TIME HISTORIES
The concentration of liquid on the surface can be calculated with the help of the
model, and this information is of great interest. The effect of the volume of the
surrounding atmosphere,or of the value of the parameterK,, on the surfaceconcentration
is determined at various times during the processof drying. The surface concentrationtime histories are shown for various values of the parameter K,, by using a logarithm
scalefor time (Fig. 11.21).
Someresults are drawn from thesecurves :
(i) For a surrounding atmosphereof infinite volume with K, = 0, the concentration of
liquid on the surfacefalls quickly to zero.
(ii) For a surrounding atmosphereof finite volume, the generalpattern is quite different.
300
[Ch. 11
The concentration at the surface drops first from the initial value to a very low value,
passesthrough a minimum, and then rises to the equilibrium value.
(iii) Of course, the value at equilibrium of the concentration on the surface is the same as
that attained within the solid.
(iv) The minimum for these surface concentration-time histories is attained between 10 and
100 s, under the selected operational conditions.
11.5 DRYING
WITH A CONTROLLED
VAPOUR
PRESSURE
The diffusivity and also to some extent the rate of evaporation or, as used in this
book, the coefficient of matter transfer on the surface, depend on the nature of the liquid
and polymer. For a given liquid-polymer combination, the diffusivity and coefficient of
matter transfer on the surface depend on temperature in an exponential way.
When the coefficient of matter transfer on the surface is very high, or in other
words, when the dimensionless number h.R / D is high, i.e., higher than 10 - 20, steep
gradients of concentration are developed within the solid next to the surface with a low
concentration of liquid on the surface. As shown above in section 11.2, the concentration
on the surface decreases very quickly as soon as the process starts. This phenomenon is
generally a real drawback, and becomes more worth noting still when the diffusivity is
concentration-dependent. As the diffusivity
decreases with concentration, the rate of
transport by diffusion next to the surface becomes lower and lower as drying proceeds.
This fact is observed by a pseudo-equilibrium of the process of drying, and a part of the
liquid is bound to be entrapped in the solid.
A solution to this difficult problem can be described with the help of the numerical
model, by acting upon the value of the vapour pressure of the substance in the
surrounding atmosphere which is related to the surface concentration at equilibrium C,,.
Instead of drying the dosage forms in a surrounding atmosphere free of vapour, the
drying process must be carried out in a surrounding atmosphere where the pressure of
vapour is controlled. An example is developed where the vapour pressure in the
surrounding is varied in such a way that
(11.42)
C,, = KZ. CR
where C,, is the concentration on the surface of the solid at equilibrium with the vapour
Sec. 11.51
Drying with a controlled
vapour pressure
301
pressure in the surrounding,

CR is the actual concentration of liquid on the surface at time t,
and K2 is a constant selected by the user.
THEORY
11.5.1
OF THE
PROCESS
WITH
CONTROLLED
VAPOUR
PRESSURE
ASSUMPTIONS
The following assumptions are made to clarify the problem :

(i) The process of drying is controlled by diffusion of the liquid within the solid and
evaporation out of the surface.
(ii) The volume of the surrounding atmosphere is finite.
(iii) The pressure of vapour in the surrounding atmosphere is controlled.
(iv) The dosage form is spherical in shape in the present case. However, a model can be
developed for other shapesby using the results given for the numerical analysis with other
shapes.
NUMERICAL MODEL
The main relationship for this problem is concerned with the vapour pressure P to
be controlled and the actual concentration on the surface of the solid CR.
By combining the linear absorption isotherm
C,, = k.P
(11.38)
where k is often considered as the solubility of the liquid in the polymer,

and eqn (11.42), the main relationship is obtained :
(11.43)
P = 2.
CR
This above relation means that the vapour pressure must be varied in such a way
that it is always kept proportional to the actual concentration on the bead surface. Of
course, this fact necessitates good knowledge of the process, and especially of the surface
concentration-time history.
The numerical model described in section 11.4 must be used, with the following
equations :
302
[Ch. 11
eqn (11.8) within the sphere, with

eqn (11.12) at the centre of the sphere, with
O<r<R
r=O
eqn (11.16) on the surface of the sphere, with
r=
eqn (11.43) for the vapour pressure in the surrounding atmosphere.
11.5.2
SIMULATION
OF THE
PROCESS
The process of drying is simulated by using the numerical model (31). The effect of
two parameters on the process are especially considered : the value of the constant K2
appearing in eqn (11.42), and the value of the coefficient of matter transfer on the surface.
EFFECT OF THE CONSTANT K2
As shown in eqn (11.42), the constant K, is the coefficient of proportionality

between the concentration of liquid on the surface of the solid at equilibrium with the
vapour pressure in the surrounding atmosphere and the actual concentration on the
surface.
Fig 11.22. Kinetics of drying for various values of the constant K, : (1) 0.9 - (2)
0.99 - (3) 0.999 - (4) 0.9999.D = 3.5 x lo-7cm2/s
x lOA cm/s.
- R = 0.38 cm - h = 19.8
Sec. 11.51
vapour pressure
303
Fig 11.23. Concentration distribution through the bead with controlled vapour
pressure and K2 = 0.9.D = 3.5 x lo- cm* / s - R = 0.38 cm - h = 19.8 x
lOA cm/s.
Calculations are made with the help of the numerical model, by keeping constant the
radius of the bead, diffusivity, coefficient of matter transfer on the surface, and initial
amount of liquid in the bead, and by varying the values of K2.
The effect of the constant K2 on the process of drying is illustrated in Fig. 11.22,
where the constant K, is varied between 0.9 and 0.999, and the other parameters are kept
constant.
The concentration distributions developed within the bead are also calculated and
drawn for K2 = 0.9 (Fig. 11.23) and for K2 = 0.999 (Fig. 11.24).
The surface concentration-time histories are also drawn for various values of K2
(Fig. 11.25), by keeping the other parameters constant.
The following conclusions can be drawn from these results.
(i) The effect of the value of the vapour pressure in the surrounding atmosphere, defined
with the help of the constant K2, is of great importance for the process of drying. The
higher the value of K2, the lower the rate of drying.
(ii) A high value of the constant K2 produces the development of almost flat gradients of
concentration within the bead, and a constant rate of drying.
304
[Ch. 11
Fig 11.24. Concentrationdistribution through the beadwith controlled vapour

pressureand K, - 0.999.D = 3.5 x lo- cm2 / s - R = 0.38 cm - h = 19.8 x
1OA cm/s.
Fig 11.25. Surfaceconcentration-timehistories for various valuesof the constant

K, : (1) 0.9 - (2) 0.99 - (3) 0.999 - (4) 0.9999.D = 3.5 x lo- cm2/ s R = 0.38 cm - h = 19.8 x 10 cm/s.
Sec. 11.51
vapour pressure
305
hyo
t
Mi
Fig 11.26. Kinetics of drying for various values of the coefficient of matter
transfer at the surface : (1) 19.8 x 10 - (2) 19.8 x 10B6 - (3) 19.8 x lOA7
cm2/sD=3.5x10-7cm2/s
- R=0.38cm
- K2=0.9
(iii) The concentration on the surface decreasesvery slowly at the beginning of the process
when the value of the constant K2 is high (0.9999).
EFFECT OF THE COEFFICIENT OF MATTER TRANSFER ON THE SURFACE
The gradients of concentration of liquid developed within the bead are very steep
when the coefficient of matter transfer on the surface is very high, or rather when the
value of the dimensionless number h.R / D is high.
The effect of the value of the coefficient of matter transfer at the surface on the
process of drying is determined using the numerical model, keeping the other parameters
constant and with K2 = 0.9. The values of this coefficient are ranging from 19.8 x lOA to
19.8 x lo- cm / s.
The effect of the coefficient of matter transfer at the surface on the kinetics of drying
is illustrated in Fig. 11.26. The profiles of concentration expanded through the bead are
drawn with K2 = 0.9 and for h = 19.8 x 10m5cm / s (Fig. 11.27) and for h = 19.8 x 10
cm / s (Fig. 11.28). The surface concentration-time histories are drawn for various values
306
[Ch. 11
Fig 11.27. Concentration distribution with controlled vapour pressure and K,

=0.9D=3.5~10-~cm~/s
- R=0.38cm
- h=19.8xlO-scm/s.
rl
Fig 11.28. Concentration distribution with controlled vapour pressure and Kz =

0.9 D = 3.5 x 10-T cm2 / s - R = 0.38 cm - h = 19.8 x 10-e cm / s.
Sec. 11.51
vapour pressure
307
Fig 11.29. Surface concentration-time histories for various values of the coefficient
of matter transfer at the surface and K2 = 0.9 : (1) 19.8 x lo- - (2) 19.8 x 1O-5 (3) 19.8 x 10
- (4) 19.8 x lo- cm / s.D = 3.5 x IO- cm* /s - R = 0.38 cm
of the coefficient of matter transfer at the surface (Fig. 11.29).

The following results are worth noting :
(i) When the constant K, is 0.9, steep gradients are developed when the rate of
evaporation is very high, i.e., when the value fo the dimensionless number h.R / D is
200.
(ii) When the coefficient of matter transfer at the surface is lower with a value of the
dimensionless number h.R / D of 20, rather flat gradients of concentration are expanded
through the bead. In this case a value of 0.9 for K2 is high enough.
(iii) When the coefficient of matter transfer at the surface is higher, as well as the values of
the dimensionless number h.R / D, higher values of the constant K2 are necessary to
reduce the steepnessof the high gradients of concentration developed within the bead.
308
11.6
[Ch. 11
CONCLUSIONS
General conclusions on the drying process of dosage forms consisting of a drug

dispersedin a polymer can be drawn, by considering the processand the effect of some
parameters.
PROCESS OF DRYING
The process of drying is controlled by diffusion of liquid within the bead and
evaporation out of the surface. The value of the dimensionless number h.R / D is of
interestfor the process.
SURROUNDING ATMOSPHERE
The effect of the volume of the surrounding atmosphereis also of high concern.
This volume acts upon the amount of liquid distributed through the bead and the
surroundingatmospherewhen equilibrium is reachedafter a time long enough.
In the casewhere the volume of the surrounding atmosphereis finite, the vapour
pressureis a very interesting parameter.By controlling this vapour pressure,for instance
by keeping it proportional to the surfaceconcentration,it is possibleto obtain an effficient
method of drying with flat profiles of concentration within the solid. This method is of
high concern when the diffusivity is concentration-dependent,as it prevents the liquid
from being entrappedin the solid.
DRYING WITH A PROGRAMMED TEMPERATURE
As the diffusivity and rate of evaporationincreaseswith temperatureexponentially,

high temperatureappearsto be of interestin order to reducethe time of drying. In the case
of dosageforms with a polymer, the plasticity of the solid increaseswith temperature,and
a high temperature can be responsible for a distortion of its shape. A programmed
temperature system can achieve drying in reducing this drawback. Two parametersof
interestappearwith the rate of heatingand the temperaturesat the beginning andthe end of
the process.
REFERENCES
1
K. Heilmann, in Therapeutic Systems;Rate-controlled drug delivery ; Concept
Sec. 11.61
Conclusions
309
and Development. Thieme Stratton ed., N.Y., 1984.

2
N.A. Peppas, R. Gumy, A. Doelker and P. Buri. Modelling of drug diffusion
through swellable polymeric systems. J. Membr. Sci., 7, 1980,241-253.

N.A. Peppas. Analysis of Fickian and non-Fickian drug release from polymers.
3
Pharm. Acta. Helv. 60, 1985, 110-111.
4
H. Fessy, J.P. Marty, F. Puisieux and J.T. Carstensen. Square-root of time
dependence of matrix formulations with low drug content. J. Pharrn. Sci. 71, 1982,
749-752.
5
E. Touitou and M. Donbrow. Drug release from non-disintegrating hydrophilic
matrices : sodium salicylate as a model drug. Int. J. Pharm, 11, 1982, 355-364.
J. Heller. Biodegradable polymers in controlled drug delivery. CRC Crit. Rev.
6
Therm. Drug Carrier Syst. 1, 1984,39-90.
B. Focher, A. Marzetti, V. Sarto, P.L. Baltrame and P. Carmetti. Cellulosic
7
materials : structure and enzymatic hydrolysis relationships. J. Appl. Polym. Sci., 29,
1984, 3329-3338.
A. Droin, C. Chaumat, M. Roller, J.L. Taverdet and J.M. Vergnaud. Model of
8
matter transfers between sodium salicylate-Eudragit matrix and gastric liquid. Int. J.
Pharm., 27, 1985, 233-243.
I. Malley, J. Bardon, M. Roller, J.L. Taverdet and J.M. Vergnaud. Modelling of
9
controlled drug release in case of Carbopol. Sodium salicylate matrix in gastric liquid.
Drug Dev. Int. Phann., 13, 1987, 67-81.
P. Magron, M. Roller, J.L. Taverdet and J.M. Vergnaud. Spherical oral polymer10
drug device with two polymers for constant drug delivery. Int. J. Pharm., 38, 1987,
91-97.
11
N. Laghoueg, J. Paulet, J.L. Taverdet and J.M. Vergnaud. Oral polymer-drug
310
[Ch. 11
devices with a core and an erodible shell for constant drug delivery. It-n.J. Pharm., 50,
1989, 133-139.
H. Liu, P. Magron, J. Bouzon and J.M. Vergnaud. Spherical dosageform with a
12
core and shell. Experiments and modelling. Int. J. Pharm., 50, 1988, 217-227.
J.Y. Armand, F. Magnard, J. Bouzon, M. Rollet and J.M. Vergnaud. Modelling
13
of the releaseof drug in gastric liquid from spherical forms with Eudragit matrix. Int. J.
Pharm., 40, 1987, 33-41.
M. Saber, F. Magnard, J. Bouzon and J.M. Vergnaud. Modelling of matter
14
transfers in drug-polymer devices used as galenic forms. J. Polym. Engng., 8, 1988,
295314.
Y. Khatir, J. Bouzon and J.M. Vergnaud. Liquid sorption by rubber sheetsand
15
evaporation.Models and experiments. Polym. Test., 6, 1986, 253-265.
Y. Khatir, J. Bouzon and J.M. Vergnaud. Non destructive testing of rubber for
16
the sorption and desorption-evaporation of liquid by modelling (cylinders of finite
length). J. Polym. Engng., 7, 1987, 149-167.
Y. Khatir, J. Bouzon and J.M. Vergnaud. Determination of processes of
17
absorption and desorption of liquids with rubber tubings by using model and short tests.
J. Polym. Engng., 7, 1987, 275-299.
Y. Khatir, J. Bouzon and J.M. Vergnaud. The kinetics of absorption and
18
desorption of liquid by a rubber annulus using a model and short term tests. Plast.
Rubber Process.Applic., 9, 1988, 53-58.
J. Crank, in Mathematics of Diffusion, Clarendon Press ed., Oxford, 1976,
19
chapter6.
J.M. Vergnaud, in Liquid transport processesin polymeric materials.Modelling
and industrial applications. Prentice Hall ed., Englewood Cliffs, N.J., USA, 1991,
chapter6.
20
Conclusions
Sec. 11.61
21
311
J.M. Vergnaud, in Drying of polymeric and solid materials, Springer Verlag ed.,
London, 1992, chapter 7.

22
N. Laghoueg-Derriche and J.M. Vergnaud. Modelling the process of drying of
dosage forms made of drug dispersed in a polymer. Int. J. Pharm., 67 (1991) 51-57.
23
N. Laghoueg-Derriche, J. Bouzon and J.M. Vergnaud. Effect of temperature of
the drying process of dosage forms prepared by a humidity technique. Int. J. Pharm., 67
(1991)163-168.
24
N. Laghoueg-Derriche and J.M. Vergnaud. Drying of dosage forms prepared by
a humidity technique using a programmed temperature system. Int. J. Pharm., 71 (1991)

229-236.
25
N. Laghoueg-Derriche
and J.M. Vergnaud. Effect of the volume of the
surrounding atmosphere on the drying process of dosage forms made of polymer and
drug. J. Polym. Engng., 10 (1991) 313-331.
26
J.M. Vergnaud. Scientific aspects of plasticizer migration from plasticized PVC
into liquids. Polymer Plast. Technol. Engng., 20 (1983) 1 - 22.

27
H.P. Blandin, J.P. David, M. Illien,
M. Malizewicz
and J.M. Vergnaud.
Modelling of the drying process of coatings with various layers. J. Coat. Technol., 59
(1987)27
28
- 32.
W. Koros and H.B. Hopfenberg. Scientific aspects of migration of indirect
additives from plastics to food. Food Technol., April (1979) 56 - 60.

29
J.M. Vergnaud and J. Bouzon in Cure of thermosetting resins. Modelling and
Experiments, Springer Verlag ed., London, 1992, Chapter 11.

30
J.L. Taverdet and J.M. Vergnaud. Surface analysis of plasticized
PVC
packagings by attenuated total reflectance. In : Instrumental analysis of foods, vol. 1.

Academic Press (1983) 367 - 377.
31
N Laghoueg-Derriche and J.M. Vergnaud, unpublished results.
12
consisting of a drug dispersed in an
erodible polymer
12.1
INTRODUCTION
Monolithic devices prepared by dispersing the drug in a biocompatible polymer

playing the role of a matrix are described in Chapter 10 when the polymler is nondegradable. Other simple devices where the drug is dispersed in an erodible polymer
matrix are also of interest(l-3).
The main purposein this chapteris to show that simple dosageforms can be built up
by dispersing the drug in an erodible material such as Gelucire. Gelucire is a mixture of
polyglycide fatty esterswith controlled hydrophilic properties.Gelucire has beenused for
preparing hard gelatin capsulesable to sustain releaseformulations of various drugs or
simulants(4,5). There is a large family of Gelucireswhich are characterizedby the melting
point ranging from 33 to 64C and by a hydrophilic-lipidic balance value. The material
selectedin this cgapter is Gelucire 46-7 becauseit is slightly soluble in gastric liquid at
37C, and then behavesas an erodible polymer.
Another aim of the presentchapteris to developtwo models ableto describethe drug
delivery. The one is built by considering a diffusional process for the drug delivery (2).
The other model is set out when the erosion of Gelurice is assumedto be the driving force
for drug delivery (3).
Drug delivery
314
from dosage forms
THEORETICAL
12.2
[Ch. 12
ASPECTS
The process of drug delivery from a dosage form where the drug is dispersed in
Gelucire is rather complex and can be described in more than one way. A first attempt at
elucidation of the process is made by considering a diffusion process. A second one is
made according to another scheme, namely, that of an erosion process.
12.2.1
DIFFUSIONAL
PROCESS
ASSUMPTIONS
The following assumptions are made in order to describe the known facts and to
simplify the process :
(i) The dosage forms are spherical in shape. The drug is thoroughly distributed in the
Gelucire matrix, and hence the dosage form is homogeneous.
(ii) A simple process of drug delivery is considered whereby only the drug release in the
liquid takes place.
(iii) The process of drug delivery is governed by transient radial diffusion with constant
diffusivity, as demonstrated by the square-root of time dependence of the amount of drug
delivery.
(iv) The radius of the beads is kept constant for calculation. This assumption is quite
distinct from that made in the process of erosion (3).
(v) The volume of liquid is greatly in excess with respect to the amount of drug in the
dosage form for the dosage forms 2 and 3. The concentration of drug in the liquid is
therefore very low. The concentration on the surface of the beads falls to a very low value,
immediately upon immersion of the dosage form in the liquid.
(vi) The volume of liquid is finite for the dosage form 1. The concentrati!on of drug in the
liquid increases uniformly with time. The rate of release of drug from the dosage form is
equal to that of dissolution of drug in the liquid throughout the entire duration of the
process.
Dosages forms l-3 are described in Table 12.1.
The process of radial diffusion under transient conditions is expressed by :
(12.1)
= D.
Sec. 12.21
Theoretical
aspects
315
where D is the constant diffusivity

and C denotes the concentration of drug at radial position r and time t.
The two cases referred to the assumptions v and vi are detailed successively.
VERY LARGE VOLUME OF LIQUID
For an infinite, i.e., very large volume of liquid, the boundary conditions are defined
according to eqn (12.3).
(12.2)
t=O
OlrlR
C = Ci*
dosage form
(12.3)
t>o
r=R
c=o
surface
where R is the radius of the dosage form,

and C, represents the uniform initial concentration of drug in the dosage form.
The amount of drug released from the dosage form up to time t, M,, is thus
expressed in terms of time t as fraction of the corresponding quantity released after inftite
time K, by the relation (Chapter 4).
(12.4)
,zMt
n2. 7r2
----D.t
R2
= 4.
II
where n is an integer.
FINITE VOLUME OF LIQUID
The initial condition is also defined by eqn(12.2), with a uniform

concentration of drug within the dosage bead.
initial
The boundary condition expresses the fact that, throughout the duration of the
process, the rate of drug delivery remains equal to that at which the drug is brought to the
surface by internal diffusion within the dosage form.
(12.5) t> 0
-D.A&
where V represents the volume of the liquid

and A is the surface area of the dosage form of radius R.
Drug delivery
316
[Ch.
from dosage forms
12
The total amount of drug M, released from the sphere after time t is evaluated in
terms of the corresponding quantity after infinite time M, :
i12,61t; W = 2
m
6a(l
+a)
n=l 9 + 9 a + q; . a
2.exp
-sDt
R2
where the q,,s are the non-zero roots of
(12.7)
3 qn
q,, =
3+a.q,
and a denotes the ration of the amounts of drug in the solution and the dosage form at
equilibrium
(12.8)
a =
3v
47tR3. K
with the partition factor K between the drug in equilibrium in the sphere and in solution.
The value of the ration a can be obtained by the following relation (Chapter 4 :
4.2.3).
(12.9)
-%
1+o!
3M,
4 X:R3.Cin. K
where M, is the amount of drug released in the solution after infinite time
and C, is the uniform initial concentration in the dosage form of radius R.
SHORT DURATION OF DRUG DELIVERY
For short periods of experimental time, for which the ratio 2
wing equation is obtained :
(12.10)
= ;[;].5
> 0.3, the follo
Sec. 12.21
Theoretical
aspects
317
whatever the volume of the solution.

This relation is very useful for determining the diffusivity of the release of drug.
12.2.2 POLYMER
EROSION IS THE DRIVING FORCE
ASSUMPTIONS
The following assumptions are made in order to clarify the problem :

(i) The samples are spherical in shape.
(ii) The drug is properly dispersed ito the dosage form.
(iii) The delivery of the drug is controlled by the erosion of the bead.
(iv) The rate of erosion of the bead is proportional to the actual area of the bead.
(v) The initial weight of the bead and the volumes of the liquid are taken into account, so
that the concentration of the drug in the solution can be established during the process.
The rate of erosion of the bead can be expressed via the rate of decrease in the
volume of the bead. It is also proportional to the actual area of the bead. This leads to :
112111
- d (volume)
dt
= k (area) = k (4 7cr2)
By considering the values of the volume and external area of the spherical bead, the
above equation can be rearranged to yield :
(12.12)
-a(v~t~e)=4rrk(~volmei
/it
= k. (volume)2/3
Integration of this evaporation between t = 0 (the beginning of the process) and t,
gives :
(12.13) (volumeatt=O)? l/I -
voumeatt)r l = $t
As the volume of the bead is proportional to its weight, this equation is transformed
as follows :
Drug delivery
318
from dosage forms
[Ch. 12
Mofbeadatt
I Mofbeadatt=O
(12.14)
3 = 1-k
I
In the present study, measurements are made in the liquid. The above equation is
thus expressed by :
(12.15)
(1-g$
(1
Mdissolved at t F= 1 k t
-Mdissolved at = =
since all the sample is dissolved at the end of the process :

(M of the bead at t = 0) = (M dissolved at t = -) = M,
(12.16)
12.3
EXPERIMENTS
MATERIALS
The following components are used :

- Sodium salicylate (COPER) in powder form, for the drug
- Gelucire 46-7 (Gattefosse, Lyon) for the erodible polymer matrix. Gelucire is a mixture
of polyglycide fitty esters with controlled hydrophilic properties. the ratio between the
number of quatemary ammonium and ester terminals is arount 1 : 20. Each member of the
Gelucire family is characterized by its melting poit and the hydrophylic balance value
(HLB). The Gelucire selected melts at 46C (drop point : Mettler), while the HLB value is
7, precisely in the middle of the 1- 14 range.

PREPARATION OF DOSAGE FORMS
The grains of sodium salicylate are dispersed in melted Gelucire heated to around
50C. The heated liquid mixture is stirred throughly in order to mix the components
properly. Spherical beads of various sizes are prepared from the paste with a ratio of drug
to Gelucire of 50 % w/w. The beads are thus cooled to room temperature
IN VITRO TESTS IN GASTRIC LIQUID
Experiments are carried out in a closed flask with the regulation of both temperature
(37OC) and rate of stirring. The bead inserted in a fiberglass basket is immersed into
Experiments
Sec. 12.31
319
synthetic gastric at 37C. The synthetic gastric liquid at pH 1.2 is prepared via the addition
of 2 g Na Cl and 80 ml of 1 N HCl to 1000 ml water to yild and aqueous solution.
At intervals, a small sample (0.1) is taken for analysis. The concentration of drug in
the liquid is determined by employing a UV spectrophotometer (Hitachi U-1000)
calibrated at 207 nm after appropriate dilution in liquid at pH 1.2.
EXPERIMENTS TESTS FOR THE DIFFUSION PROCESS
The experimental tests for studying the diffusion process are described in Table
(12.1).
Table 12.1 Experimental tests for the diffusion process.
Test No 1
Bead weight (mg)
Radius (cm)
Volume
of
liquid (ml)
1
264
0.36
200
187
0.33
200
92
0.268
100
EXPERIMENTAL TEST FOR THE EROSION PROCESS
The experiment tests for studying the erosion process are described in Table (12.2).
Table 12.2 Experimental test for the erosion process
Radius (cm)
Volume
of
Test No2
Bead weight (mg)
liquid (ml)
1
210
0.34
221
95
0.28
100
226
0.38
100
Three kinds of experiments are performed as shown in Table 12.2. The ratio
(volume of solution)/(weight of bead) is kept constant in expts 1 and 2, despite the use of
various weights of drug. For expt 3, this ratio is 2.8 - fold lower.
320
Drug delivery
from dosage forms
[Ch. 12
12.4 RESULTS
12.4.1
RESULTS
WITH
THE
DIFFUSION
PROCESS
Two types of results on the process of diffusional transport for the drug delivery
from oral dosage forms composed of Gelucire are given :
(i) those where the volume of synthetic gastric liquid is very large is relation to the amount
of drug
(ii) those in which the volume of the synthetic gastric liquid is finite.
VERY LARGE VOLUME OF LIQUID
The kinetics of drug delivery, determined either experimentally or by means of

calculation with the use of eqn, (12.4), are depicted in Fig. 12.1 for the three ensuing
cases (l-3) :
t (h1
*
?
10
Fig. 12.1 Kinetics of drug release from drug-Gelucire dosage forms in a very
large volume of synthetic gastric liquid. (1) dosage form 1- (2) Dosage form 2
- (3) Dosage form 3.
Results
Sec. 12.41
321
Fig. 12.2 Kinetics of drug release from drug-Gelucire dosage forms in a finite
volume of synthetic gastric liquid. (1) dosage form l- (2) Dosage form 2 - (3).
Dosage form 3.
(1) 264 mg dosage form in 200 ml gastric liquid

As demonstrated in the three curves shown in this figure, the experimentally
determined plots and these calculated based on theory are clearly in good agreement
at leats when2
< 0.9.
FINITE VOLUME OF LIQUID
For the case of a finite volume of liquid, eqn, (12.6), with the ration a and the qns
given in Table 4.2 (Chapter 4) for various values of a.
Calculation is performed with the consideration that the ratio of the drug that is
extractable from the dosage form amounts to 95 % for dosage form 1.
The results on the kinetics of drug delivery as observed during experimentation and
Drug delivery
322
from dosage forms
[Ch. 12
as calculated are drawn in Fig. 12.2. AS is clearly evident in this figure, the close
agreement between the two plots provides a further demonstration of the validity of the
assumption of the diffusion of drug in a finite volume of liquid.
12.4.2
RESULTS
WITH
THE
EROSION
PROCESS
The main result is obtained for the kinetics of release of the drug into synthetic
gastric liquid. The amount of drug released after time t is assumedto be proportional to the
amount of Gelucire dissolved. Thus, the eqn, (12.15) obtained for the Gelucire can also
be used for the drug.
Another feature is also considered because it appears in both in vitro and in vivo
tests : namely, the effect of the (volume of drug)/(volume of liquid) ratio on the kinetics of
drug release.
KINETICS OF RELEASE OF DRUG BY EROSION OF THE DEVICE
The kinetics of drug release being assumed to be controlled by erosion of the device,
the rate of erosion of the device as well as as that of release of drug is taken to be
proportional to the actual area of the device.The above mathematical treatment then leads to
the determination of the lew relating the time to the weight of the bead to the power
one- third, as shown ineqn. (12.14) and (12.15). The values of
obtained for
the drug released in the solution are plotted as a function of time, for the three experiments
described in Table 12.2 (Eqn 12.3).
The following conclusions are drawn from the curves in Fig. 12.3 :
(i) Straight lines are obtained from the beginning of the process to around 4 or 5 h. The
slopes of these lines are equal to the value of the constant k in eqn (12.15).
(ii) The point of intersection of each of these straight lines with the vertical axis, denoted b
in Table 12.3, is near unity.
(iii) Following the results in i and ii, the kinetics of drug release can be described by an
erosion model with a rate of release that is proportional to the actual area of the device or to
the actual weight of the device to the power two-thirds.
The values of constant k shown in eqn.(12.15) which is equal to the slopes of
curves drawn in Fig. 12.3 are listed in Table 12.3, as well as the value of b.
The straight lines observed in Fig. 12.3, as well as the values of b chose to unity,
demonstrate clearly that the process of drug release can be explained by erosion of the
Sec. 12.41
Results
323
Table 12.3 - Characteristicsof the kinetics of erosion

Expt
1
2
3
k (min-1) x 103
2.21
2.21
1.74
b
0.98
0.96
0.995
polymer matrix. However, it is always better to compare the kinetics obtained from
experimentsandcalculationwith the mathematicalmodel.
t (hl
2
plotted as a function of time. Rateof releaseis proportional to the actualsurfacearea.
324
Drug delivery
from dosage forms
ICh. 12
Fig. 12.4 Kinetics of release of drug with the erosion process. Weight of bead :
210 mg - volume of liquid : 221 ml.
:+
Theoretical : -Experiments
The kinetics of release of drug in synthetic gastric liquid are drawn in Fig 12.4, 12.5
and 12.6 for the tests listed in Table 12.2. A number of results are indicated :
(i) Good agreement is obtained between the experimental and calculated kinetics of drug
release for the entire process perhaps for test 3 in Fig. 12.6 at the process when
(ii) In test 3 depicted in Fig 12.6, the agreement between experiment and theory is perfect
at the beginning of the process, but becomes steadily worse at the end of the process.
(iii) The kinetics of release are described by the same equation in tests 1 and 2, in spite of
the fact that the weights of the device are quite different (210 and 95 mg), because the
Sec. 12.41
Results
325
:+
volume of the liquid is proportional to the weight of the device. The kinetics of release is
slightly different in test 3, with a lower value of the costant k (1.74 vs. 2.21 for test 1
and 2).
(iv) The major factor concerned in all these three cases is the ratio of the volume of liquid
to the weight of the device.
EFFECT OF THE RATIO OF THE VOLUME OF LIQUID AND WEIGHT OF DEVICE
For cases where the ratios (volume of liquid)/(weight of device) are equal, such as
tests 1 and 2, the same kinetics of drug release is obtained, meaning that the rates of
Mt
relases are.identical at any time, and particularly when M < 0.9.
Drug delivery
326
from dosage forms
[Ch. 12
:+
For lower values of this ratio, e.g. as in test 3, the kinetics of release differ slightly
from those in the above two cases, and the rate of release of the drug is lower. This is
clearly observed, especially for the initial stages shown in Fig. 12.6 and test 3.
Moreover, in test 3, with the lower value of the volume and of the (volume of
liquid)/(weight
of device) ratio, the experimental and calculated curves are no longer
superimposable after around 5 h of release, the experimental values being lower than those
obtained by calculation.
This result is of interest with respect to the kinetics of release at the end of the
process.
Sec. 12.51
12.5
Conclusions
327
CONCLUSIONS
The kinetics of releaseof drug from drug-Geluciredevicescan be explainedeither by

a diffusion processwith constantdiffusivity or by an erosionprocess.
With the diffusion process,analytical solutions for the kinetics of drug releasedexist
either when the volume of syntheticgastric is very large and when very small as compared
with the drug contentof the dosageform. This is an interestingconclusionfrom a practical
point of view, since the kinetics of drug delivery can therefore be expressedin terms of
bead radius, in addition to being defined according to the liquid/bead volumes ratio, by
simple equations. It is thus easy to determine the right size of the beadsnecessaryfor a
desiredpurpose.
However, it should be borne in mind that the process follows a more complex
course of development. Gelucire is slightly soluble in the liquid and the bead size
decreases.The rate of releaseof drug is shown to be proportional to the actual surfacearea
of the dosageform.
As a matter of conclusion, it could be said that the process is perhaps even more
complex than the description given by either of the two cases,with both diffusion and
erosion playing contributory roles in the release of drug. Nevertheless, the general
principle (6) is still applicable. The liquid diffuses into the Gelucire matrix, hence
dissolving the drug which is then able to diffuse out of the dosageform through the liquid
located therein. In addition to this classical scheme, Gelucire swells slightly when in
contact with the liquid, subsequently undergoing dissolution at a sufficiently high
concentrationof liquid in Gelucire.
REFERENCES
J. Heller Biodegradable polymers in controlled drug delivery. CRC Critical
1
Reviews in Therap. Drug Carrier Systems.1 (1984) 39-90.
D. Bidah, E.M. Ouriemchi and J.M. Vergnaud Diffusional processof drug delivery
2
from a dosageform with a Gelucire matrix. Int. J. Pharm. 80 (1992) 145-149.
D. Bidah and J.M. Vergnaud Kinetics of in vitro release of sodium salicylate
3
dispersedin Gelucire Int. J. Pharm. 58 (1990) 215-220.
328
Drug delivery
from dosage forms
[Ch. 12
Gattefosst Gelucire puts liquid formulations into hard gelatin capsules Private
paper (1983).
5
C. Doelker, E. Doelker, B. Buri, L. Wag&tire and B. Glas Incorporation of liquid,
deliquescent, and unstable active ingredients into excipients for hard gelatin. 3 rd Int.
Conf. Techol. Paris, June 1983.
J.M. Vergnaud in Liquid transport processes in polymeric materials. Modeling and
6
industrial applications Prentice Hall ed (1991) chapter 10.
13
Dosage forms made of a core and
shell, with an erodible shell. Constant
rate of delivery
13.1
INTRODUCTION
With classicaldosageforms administeredvia the gastrointestinaltract, the releaseof

drug is controlled by dissolution. The drug is then liberated very rapidly, and the drug
concentrationin the gastric liquid and blood builds up to a maximum value and then falls
exponentially with time until the following dose.As a result, an undulating concentration
history is observed, low concentrations and underdoses alternating with high
concentrations and overdoses (1.3). Optimization of conventional agent delivery to
maximize agentavailability with a minimum amountof the drug cannotbe realizedeasily.
The method is practical if one consider new types of dosageforms consisting of a
protectedsupply of drug from which the drug is automatically releasedat a controlled rate
over a long period of time. Monolithic devices where the drug is dispersed in an inert
matrix are studied in Chapter 10 and Chapter 11. Both non-degradable(Chapter 10) and
degradable(Chapter 11) polymers are used for the matrix, the purpose being that these
polymers are not absorbedin the body but passthrough. Several theories are put forward
in an attemptto model the details of the delivery of the drug. Invariably they arebuilt on a
combination of hydrodynamic and diffusion processeswhich explain the square-rootof
time dependenceof the drug delivery. Other studiesconsiderthe simultaneoustransfer of
the drug and the liquid. In all theseabovestudies,the delivery of the drug is controlled by
diffusion and the amount of drug delivered is proportional to the square-root of time
(Chapter 10 and 11).
With the diffusional process,the rate of releaseof the drug is controlled, but it is far
330
Dosage forms made of a core and shell
[Ch. 13
from being constant. Very high at the beginning of the processwhen the tangent of the
kinetic curve is vertical, the rate of releaseof drug decreasesexponentiallywith time.
A releaseof drug with a constantrate can be obtainedby using devicesbasedon the
osmotic principle (1) with a membraneof constantthickness.It can also be obtained with
the help of more simple devices. These devices are made of a core and shell, the core
consistingof a device where the drug is dispersedin a non-erodiblepolymer, and the shell
being an erodible polymer free of drug. The process is thus based on diffusion and
polymer erosion (4).
The purpose in this chapter is to show that dosageforms able to deliver the drug at
constantrate can be preparedby using simple devices. The basis of the method is to use
simustaneously :
(i) A core made of a spherical dosage form obtained by dispersing the drug in a
biocompatibleandnon-degradablepolymer
(ii) A shell with a sphericalmembranemade of an erodiblepolymer.
Another purposeis to describe a very simple model ablenot to describethe process
but to fit the results. Thus the constantrate of releaseof the drug is correlated with some
parametersof interest, such as the thicknessof the shell as comparedto the radius of the
core.
The core is made of drug dispersedin Eudragit, while the shell is made of Gelucire.
The dosageform madeof a core and shell is immersedin gastric liquid.
13.2 THEORETICAL
ASPECTS
The whole problem is rather complex, becauseof the simultaneousand successive

steps: the transfer of the liquid that enters successivelythe shell (made of Gelucire) and
the core (made of a dispersion of the drug in Eudragit) ; then the dissolution of the drug
andtransportof the drug through the liquid locatedin the core and shell. At the sametime,
a slight erosion or dissolution of the shell takesplace.
ASSUMPTIONS
In order to simplify the problem and find an analytical solution, the following
assumptionsare made :
(i) The dosageform is made of two concentric spheres,the core of radius R, and the shell
of external radius R2.
Sec. 13.21
Theoretical
aspects
331
(ii) The core is obtained by dispersing the drug in a non-degradablepolymer (Eudragit)

and the shell is madeof a pure degradablepolymer (Gelucire).
(iii) Two matter transferstake place simultaneously: the one with the liquid which enters
the dosageform and dissolves the drug, the other with the drug in solution through the
liquid absorbedinto the dosageform (4 - 6).
(iv) The transfersof the liquid and drug are controlled by diffusion.
(v) A very simple model is tested,by consideringthe shell as a sphericalmembranewith a
constantthickness and by assumingthat the diffusivity of the drug is lower in the shell
than in the core.
With the help of the assumptions,and especiallyof V, a mathematicaltreatmentcan

be made for the diffusion of the drug.
As the diffusivity of the drug is assumedto be higher in the core than through the
shell, the drug is brought very quickly to the core-shellinterface as soon asthe drug enters
the shell. The concentrationof the drug at this interface can be consideredas constant,at
least over a period of time when the drug is available in the core. In a large volume of
liquid, the concentrationof drug is very low. Moreover, for a shell of constantthickness,
the model of a sphericalmembraneworking under stationaryconditions can be used.
The transientradial diffusion through the sphereis describedby the Ficks equation
(13.1) g
= 1.;
r
D.r*.$
The initial and boundary conditions are as shown in Fig. 13.1 :

(13.2)
t=O
r I Rl
Rl < r < R2
c = c,
c=o
(13.3)
t>O
r= Rl
r 2 R2
Rl < r < R2
C = Ci~
c=o
core
shell
cr
Under these simple conditions, an analytical solution is found for eqn. 1 when the
diffusivity of the drug is constant(Chapter4).
[Ch. 13
332
Fig. 13.1 . Schemeof dosageform with core and shell.Core : drug and Eudragit
R L of radius RI. Shell : Gelucire of thicknessR2 - RI.
The amount of drug transferred after time t out of the spherical membrane is
expressedby the series:
Mt
(13.4)
4~.
Rr. R,.(R,
--.
Dt
-RI).
Ci, = (RZ-Rr)
1
-
2
2
7r
When the tune is long enough, the series vanishes, and the drug is transferred in
steadystateconditions.The amountof drug transferredvaries linearly with time :
(13.5) M, = R;;(RD.-c;;lj
[t - R2;;1].
4xR;
The slope of the straight line Mt = f (t) or the rate of matter transferred is
proportional to the diffusivity in the membrane:
Experimental
Sec. 13.31
(13.6) F
41r:.R,.R,.ci,.D
R2
13.3
333
Rl
EXPERIMENTAL
MATERIALS
The following materialsareused:

- Sodium salicylate, in powder form, is chosenfor the drug.
- Eudragit R 5, a copolymer of dimethylaminoethylacrylateand ethylmethacrylateof mol.
wt. = 150,000(Rohm Pharma) aspolymer matrix for the core of radius R,.
- Sulfanilamide, the p - aminobenzenzesulfamide, in white powder form, is selectedfor
anotherdrug.
- Eudragit RL is usedas the polymer matrix in the core of radius RI.
- Gelucire 46-7, a waxy solid made of partial glycerides andpolyglycides fatty esterswith
controlled hydrophilic properties (Gattefosse) is selected for the spherical coating
surrounding the core (shell). The drop point (Mettler) is 46C and the hydrophiliclipophilic balanceis 7, correspondingto a medium value.
Eudragit and drug (either sodium salicylateor sulfanilamide) in powder form are
intimately mixed in a mortar, and transformed into a thick paste after addition of a small
amount of ethanol. Spherical beads are prepared from this paste and dried at room
temperaturefor four days (core>.
The core is then surroundedby a sphericalmembraneof Gelucire, by dropping the
core for 2-5 s into Gelucire kept liquid at 60 C. The Gelucire membrane hardensafter
cooling at room temperature.The thicknessof the shell can be controlled by selectingthe
right valuesfor the temperatureof the liquid Gelucire and for the time of contact.
Severaldosageforms arepreparedwith various values of the thicknessof Gelucire,
by keeping a constantpercentageof drug (50 7o w I w ) in the core and about the same
weight of the core.
IN-VITRO TESTS
Experiments are carried out in a closed flask with a controlled rate of stirring. The
334
[Ch. 13
bead (around 400 mg) inserted in a glass fibre basket is soaked into synthetic gastric liquid
(100 ml) at 37 C with the classical composition at pH 1.2 : 1000 ml of aqueous solution,
80mlHC11Nand2gNaC1.
At intervals, the bead is weighed and a small sample (0.1 - 0.2 ml) of liquid is taken
for analysis. The amount of drug released in the liquid is determined by using a UVspectrophotometer (Hitachi U - 1100) calibrated at 216 nm for the sulfanilamide and at
303 MI for sodium salicylate.
10
20
Time (h1
30
CO
Fig. 13.2 . Kinetics of release of sodium salicylate from the core alone in synthetic
gastric liquid.415 mg - Diameter : 0.85 cm.
13.4
RESULTS
The kinetics are given either for sodium salicylate or for sulfanilamide.
13.4.1 RESULTS
WITH SODIUM
SALICYLATE
Some results are given for the kinetics of release of sodium salicylate into synthetic
gastric liquid from the following two drug-polymer devices
(i) The device with the core alone
Sec. 13.41
Results
335
Fig. 13.3 . Kinetics of releaseof sodium salicylate from the core and shell device.
Diameter = 0.85 cm. Thickness of Gelucire = 0.14 mm. 406 mg.
(ii) The devicewith the core and shell.

KINETICS OF DRUG RELEASE WITH THE CORE ALONE
The kinetics of releaseof the drug in synthetic gastric liquid from the core alone is
shown in Fig.13.2. Someresults can be noticed from this curve :
(i) - The amountof sodium salicylatein the liquid at the end of the process,at equilibrium,
is around34 % when the device contains50 % of drug originally
(ii) - The rate of drug delivery is very high at the beginning of the processwith a vertical
tangentat t = 0, andthen decreasesexponentiallywith time.
KINETICS OF DRUG RELEASE WITH CORE-SHELL DEVICES
The kinetics of releaseof sodium salicylate in synthetic gastric liquid from various
core-shell devices are shown in Fig.13.3 - 13.6 for different thicknesses of the shell
ranging from 0.14 to 0.20 mm, while the cores arethe samewith a weight around400 mg
and a diameterof 0.85 cm. Four resultsof interest areworth pointing out :
(i) The amount of drug delivered at equilibrium into the gastric liquid is around 33-34 %.
336
A DRUG
%
30-
Kh.
Time (h1
I
0
10
20
30
10
Fig. 13.4 Kinetics of releaseof sodium salicylate from the core and shell device.
10
20
30
LO
13
Sec. 13.41
Results
337
5,
I.
3
0.10
0.15
AR(mm1
I
0.20
Fig. 13.7 . Rate of delivery of the drug as a function of the thicknessof the layer
of Gelucire (weight of drug per second).
338
[Ch. 13
A slight decrease in this amount can be appreciated when the thickness of the shell is
increased.
(ii) In all cases, a straight line is obtained for the kinetics of drug released in the liquid.
The rate of drug delivery is constant within a very large range of the transfer. In fact, it
can be said that this rate is constant along the whole process.
(iii) The rate of drug delivery is inversely proportional to the thickness of the shell, for a
given value of the radius of the core. In fact, the constant rate is inversely proportional to
(1 /R, - 1 /R2). (Fig.13.7)
(iv) The time for which the process is conducted under transient conditions is very short,
at the beginning of the process.
Some practical results are given in Table 13.1
Table 13.1. Caracteristics of the drug release

Duration
Shell thickness
Rate of delivery
Dx 109
(md
(mg /W
(cm2 / s)
0-d
0.14
9.4
6.5
11
0.16
8.4
6.3
12
0.17
0.20
8.0
6.0
6.0
5.8
13
16
THEORETICAL CONSIDERATION
As the diffusivity of the drug within the core is of the order of magnitude of
2 x lo- cm2 / s and for the core-shell devices is around 6 x 1O-9cm2/ s, the rate of
diffusion of drug is higher in the core. The drug is thus constantly available in the core,
providing a concentration which could be considered as constant on the internal surface of
the membrane.
The most important result is certainly the constant rate of drug delivery, as well as
the relationship between this rate of delivery and the thickness of the shell or rather
(1 /RI-l
/ R2).
13.4.2
RESULTS
WITH
SULFANILAMIDE
The same results are obtained when the drug is sulfanilamide. They are thus
summarized.
Sec. 13.41
Results
339
Fig. 13.8 . Kinetics of releaseof sulfanilamide from the core alone. 337 mg.
R1 = 0.38 cm.
KINETICS OF DRUG RELEASE WITH THE CORE ALONE
The kinetics of release of sulfanilamide is shown in Fig.13.8 for the core alone
immersedin gastricliquid.
About 95 % of the drug previously dispersedin the core is releasedin gastric liquid.
The solubility of sulfanilamideis higher than that of sodium salicylatein gastricliquid.
The process is controlled by diffusion with a high rate of drug release at the
beginning.
KINETICS OF DRUG RELEASE WITH TH CORE-SHELL DEVICES
The kinetics of drug releasedobtained with the devicesmade of core and shell are
shown in Fig.13.9 - 13.12 for various values of the thickness of the erodible shell. The
following resultsare worth noting :
(i) The amount of drug delivered at equilibrium into the synthetic gastric liquid is around
90-95 %, this amountbeing aboutthe sameasthat attainedwith the core alone.
(ii) A constant rate of delivery is obtained for the drug during a very large part of the
340
10
20
30
[Ch. 13
10
Fig. 13.9 . Kinetics of releaseof sulfanilamide from the core and shell device.
368 mg. Rl = 0.39 mm. Shell thickness = 0.11 mm.
A DRUG
O/O
Fig. 13. 10 . Kinetics of releaseof sulfanilamide from the core and shell device.
396 mg. Rl = 0.40 cm. Shell thickness= 0.23 mm.
Conclusions
Sec. 13.51
341
ADRUG
%
--
10
20
30
LO
Fig. 13.11 . Kinetics of releaseof sulfanilamide from the core and shell device.
373 mg. R1 = 0.39 cm. Shell thickness= 0.31 mm.
process,the tune of transientdiffusion being rather low.

(iii) The processof drug transfer appearsas follows : a fit stagewhen the rate of release
of drug increasesup to a constantvalue ; a secondstagewith about a constantrate.
(iv) The rate of drug releasedin the liquid is inversely proportional to the thicknessof the
shell made of Gelucire, for the various deviceswhere the radius of the core is kept about
constant. (Fig.13.13).
13.5 CONCLUSIONS
Oral dosageforms madeof a core and shell aretestedin syntheticgastricliquid. The
main characteristicsof theseforms are as follows : the core is obtained by dispersingthe
drug in a non-erodible polymer such as Eudragit, and the shell is made of an erodible
polymer such as Gelucire.
Then dosage forms, tested with two kinds of drug, namely sodium salicylate or
sulfanilamide, are able to deliver the drug with a constant rate throughout the process.
342
[Ch. 13
Fig. 13.12 . Kinetics of release of sulfanilamide from the core and shell device.
328 mg. Rl = 0.38 cm. Shell thickness = 0.50 mm.
Moreover, the rate of drug delivery happens to be inversely proportional to the initial
thickness of the shell.
As a result, it is possible to prepare any dosage forms able to deliver the drug with a
desired rate of release.
Following these experimental results, a very simple model based on the presence of
a spherical membrane is shown. However it is sure that this model is of little value from a
theoretical point of view becausetwo facts are in contradistinction to this simple model :
(i) The concentration in the core, or rather on the surface of the core which is in contact
with the shell, is not constant during the process.
(ii) The thickness of the shell decreasesduring the process, because of the solubility of the
Galucire in gastric liquid.
In fact, in a qualitative way, two mains stages can be considered for the drug
transfer from these dosage forms : one with the transfer of the drug through the core, the
other through the shell. The decreasein the concentration of the drug in the core during the
process is followed by a decrease in the rate of drug released through the shell. But this
Sec. 13.51
Conclusions
343
ADRUG/h
O/O
8,
6.
L.
0
I
0.1
0.2
0.3
0.1
hR(mml
l
0.5
Fig.13.13 . Rate of drug released(% drug / h) as a function of the thicknessof the

shell (mm).
decreasein the rate of the drug is compensatedby a decreasein the thicknessof the shell.
As a result of this compensation, a constant rate of delivery out of the dosageform is
obtained.
Before building a model taking into account all the facts, good knowledge of the
processis necessary.In this case,somedata concerningthe transportof the liquid through
Gelucire are needed: the diffusivity and the amount of liquid absorbedat equilibrium. As
a stageof erosion of Gelucire is superimposedon the stageof the transport of liquid, the
processis very complex.
REFERENCES
K. Heilmann, in Therapeutic Systems.Rate-controlled. Drug Delivery ; Concept
1
and Development. Theme Stratton ed., New York (1984).
N.A. Peppas, R. Gurny, A. Doelker and P. Buri. Modelling of drug diffusion
2
through swellable polymeric systems.J. Membr. Sci. 7 (1980) 241-253.
344
[Ch. 13
N.A. Peppas. Analysis of Fickian and non-Fickian drug release from polymers.
Pharm. Acta. Helv. 60 (1985) 110-111.

P. Magron, M. Rollet, J.L. Taverdet and J.M. Vergnaud. Spherical oral polymer4
drug device with two polymers for constant drug delivery. Int. J. Pharm. 38 (1987)
91-97.
N. Laghoueg, J. Paulet, J.L. Taverdet and J.M. Vergnaud Oral polymer-drug
5
devices with a core and an erodible shell for constant drug delivery. Int. J. Pharm. 50
(1989) 133-139.
4
A. Droin, C. Chaumat, M. Rollet, J.L. Taverdet and J.M. VErgnaud. Model of
Pharm. 27 (1985) 233-243.
I. Malley, J. Bardon, M. Rollet, J.L. Taverdet and J.M. Vergnaud. Modelling of
5
controlled drug release in case of Carbopol-sodium salicylate matrix in gastric liquid.
Drug Devel. Ind. Pharm. 13 (1987) 67-81
J.M. Vergnaud. in Liquid Transport Processes in Polymeric Materials. Modelling
6
and Industrial Applications. Prentice Hall ed., N.Y. (1991) chapter 10.
14
Dosage forms made of core and shell,
with a non-erodible polymer
14.1
INTRODUCTION
Monolithic devices where the drug is an inert biocompatible polymer are able to
control the releaseof the drug in the stomach or intestine. As shown in Chapter 10, the
delivery of the drug is described by a rather complex process : the liquid enters the
polymer and dissolves the drug, enabling the drug to leave out the polymer through the
liquid located in the polymer. Two matter transfers are thus considered,the one with the
liquid and the other with the drug. Thus two matter transfers are controlled by transient
diffusion and they are connectedwith eachother, meaning that the diffusivity of the drug
dependson the concentrationof the liquid in the polymer (l-3).
As a result of this diffusionnal process,the rate of drug delivery is far from being
constant. Very high at the beginning of the process when the tangent of the curve is
vertical, the rate of drug releasedecreasesregularly with time in an exbonentialway. This
is the reason why many authors have found that the amount of the drug delivered is
proportional to the square-rootof time, at least for the earlier stageof the process,when
2
< 0.6 for a thin plane sheetor when 2 < 0.3 - 0.4 for a sphericalbead(4).
00
Some simple devices able to deliver the drug with a constantrate are prepared by
using a core and shell, the core being obtained by dispersing the drug in a non-erodible
polymer, and the shell being made of an erodible polymer free from drug (5, 6 Chapter 12)
Another way to obtain a constantrate of delivery consistsof preparing devicesable
[Ch. 14
Dosage forms made of core and shell
346
Druq delivered
Time
Fig. 14.1. Schemeof kinetics of drug de1ivery.a: diffusional processwith a core

alone consistingof a drug disperseda non-erodiblepo1ymer.b: kinetics of drug
releasedwith the device madeof a core and shell with a non-erodiblepolymer.
to deliver the drug with kinetics expressedby curve b in Fig. 14.1, and mixing these
deviceswith the devicesmade of a core alonewith a non-erodiblepolymer (1 - 4, Chapter
10). As shown in Fig.14.1, the kinetics of releaseof drug with thesenew devices shown
in curve b is about symmetrical with respectto curve a obtainedwith transientdifusion
through the core alone.
Thesenew devicesable to deliver the drug with the kinetics illustrated by curve a are
made of a core and shell, the core with the drug dispersedin a non-erodiblepolymer, and
the shell beinga non-erodiblepolymer free from drug (7,8).
The polymer used for the shell may be either the sameasthat in the core (7,8) or a
different polymer (9 - 11). In the latter case,the role of the shell consistsof controlling the
rate of releaseeither by a low diffusivity or by a low concentrationof the drug passing
through.
14.2
THEORETICAL
Two matter transportstaking place, one concerned with the liquid entering the
polymer and dissolving the drug and the other with the drug leaving the device, thesetwo
transportsare studied simultaneously.
Sec. 14.21
Theoretical
347
ASSUMPTIONS
The following assumptionsaremadefor the core and shell.

(i) The dosageform is made of a core and shell, spherical in shape.In the core, the drug
is dispersedin a non-erodible polymer. The shell surrounding the core is a non-erodible
polymer. The non-erodiblepolymer usedfor the core and shell is Eudragit.
(ii) Two matter transportstake place simultaneously,with the liquid entering the polymer
and the drug leaving the dosageform.
(iii) These two transports are controlled by transient diffusion, these diffusions being
connectedwith eachother.
(iv) The diffusivities for thesetransportsareconcentration-dependent.
(v) The concentrationof the liquid on the external surface of the dosageform attains the
value at equilibrium as soon asth processstarts.
(vi) A partition coefficient exists for the drug betweenthe external surface of the dosage
form and the liquid of finite volume in which the dosageform is immersed.
(vii) Another partition cofficient exists for the liquid on each surface of the interface
betweenthe core and shell.
(viii) The swelling is neglected, and a frame of reference is fixed with respect to the
original dosageform.
The equation of radial diffusion under transient conditions with a concentrationdependentdifusivity is written as follows :
(14.1) a;-=ysi3T
ac
l
a
2 ac
Der.,
The initial andboundaryconditions are :

(14.2)
t>o
0 I r < R,
R, < r I R,
0 I r c R2
C = Ci,
c=o
c=o
drug, core
drug, shell
liquid
(14.3)
t=O
r=
c = c,,
liquid
R2
In the case of a short test, when the amount of matter tranferred M, is low, the
following equationcan be used :
348
D&age
[Ch. 14
forms made of core and shell

0.5
(14.4) 2
= ; (Y)
with
M, = < Oe3- o.4
expressingthat the amount of matter transferredis proportional to the square-rootof time.

NUMERICAL ANALYSIS
No analytical solution can be obtained for this complex problem, and a numerical
method with finite differencesmust be used.
The dosageform with the two concentric spheres(core and shell) is shown in Fig.
14.2. The core,of radius R, and shell of radius R, are divided into concentric spherical
membranesof constantthickness : Ar, for the core and Ar, for the shell. The obvious
relations aregiven :
(14.5)
R, = N, . R,
r = n.Arr
core
OSnINr
(14.6)
R, = R,+ (N - N,) .Ar,

r = RI + (n - Nr) . Ar,
shell
N, I n I N
The matter balance is evaluated in various places, and the new concentration is
obtained(Chapter8)
1 5 n 5 Nr - 1
The new concentration CN, , after elapseof time At is expressedin terms of the
WITHIN THE CORE, with
previous concentrations
(14.7) CN, = C, +
At
[-J
with the function J :

(14.8)
J, = n2 . D, . (C,-0.5 - C,+o.,)
n-0.5
+ Jn+0.5
Sec. 14.21
Theoretical
(j+lkix
SpCt2
jhx
I
i.At-C(j+l)-
(j-lI.hx
I
CCjL C(j-1)
I
I
I
CN(j+l)-CNIjLCNij-1)
(i+l).bf-
vTime
Fig. 14.2. Dosageform with a core and shell. Schemefor numerical analysis.
with n = 0
The new concentration CNo is given by :
ATTHE MIDDLE OFTHE CORE,
(14.9) CN, = Co - y
. JO.5
(A 4
N1+ 1 S n < N-l
The new concentrationis given in the sameway asfor the core :
WITHIN THE SHELL, With
(14.10)
CN, = C, +
At
*Il (A %I2
. [-
JnmOe5
+ Jn+0,5]
349
[Ch. 14
350
with A,, given by :

2
- NI+n
A, = 2
(14.11)
1
+12
2
CORE-SHELL INTERFACE, with
n = N,
Special attention is given to the core-shell interface in order to account for the
discontinuity of the concentrationof the liquid anddrug.
- For the liquid The new concentrationof liquid CNL , at the core-shellinterfaceis given by
(14.12) CN& = C& - &

- g
(1 - $-i,.
[l + $.
8. At. Ar2
3B.R;
[CN&wl-
C&ml]
[CN;,+t-C;,+tj
JIN,+1
-
4.At.Ar2
3B.R;
,
JN
The concentrationof the liquid varies discontinuouslythrough the interface:

12
(14.13)
c;,
= c;,
. c,
CEk
meaning that the rate of the concentration on each side of the interface is constant and
equalto the value at equilibrium.
The term B is given by :
12
(14.14)
B = Art + Ar, . c,
- For the drug -
Sec. 14.21
Theoretical
351
The new concentration of drug CN:, at the core-shell interface is given by :
(14.15)
CN;, =
4At.
3 (A rt + A r2) R:
EXTERNAL SURFACE OF RADIUS
4At.
Ar2
. &+I
3 (Arl
Art
+ Ar2)
d
. JN
Rt
R2
The concentration of the liquid on the surface is constantly equal to the value
attained at equilibrium
(14.16)
CL = CL
The concentration of the drug on the surface is related to the value of the
concentration in the liquid of finite volume with a partition factor K.
(14.17)
CN; =
K. (VI. Cfdorein
- Mt)
V + i tr, Ri . A r2. K
where K is the partition factor
(14.18)
K = M - M
M:
Mz being the amount of drug located in the dosage form at time t,

Mi
the corresponding amount after infiite time
V, is the value of the core of radius R,

and V is the value of the dosage form.
352
[Ch. 14
AMOUNT OF MA-KERS LOCATED IN THE DOSAGE FORM
The amount of drug and liquid located in the dosage form is calculated at any time
by integrating the concentration of drug and liquid in the dosage form with respect to
space, as shown in Chapter 8.
14.3
EXPERIMENTAL
MATERIAL
The drug is sodium salicylate, in powder form.

The
polymer
is
dimethylaminoethylacrylate
Eudragit
RS
(RGhm
Pharma),
a copolymer
of
and ethylmethacrylate with a mean molecular weight of
150,000, in powder form.

DOSAGE FORM
The device consists of a core and shell.

The core is obtained by dispersing the drug in Eudragit with a small amount of
ethanol, and pressing the paste into spherical beads, and drying them. The core is
prepared with a mixture of drug and Eudragit 50 - 50 in weight.
The core is surrounded by a shell made of pure Eudragit. A paste of Eudragit with a
small amount of ethanol is used for this purpose. The core is thus dried at room
temperature. The thickness of the shell is determined by its weight.
CONDITIONS OF TESTING
In-vitro experiments are carried out by immersing the dosage form (400 - 500 mg)
located in a glass fibre basket in synthetic gastric liquid (100 ml) at 37 C and pH 1,2.
The device is weighed and a sample (0,l ml) of liquid is taken for analysis at
intervals. The amount of drug released is determined by using
UV-spectrometry
calibrated at 300 nm.
14.4
RESULTS
Three points are of concern in this chapter :
(i) Experiments are made in order to obtain data, such as the diffusivities and amounts of
matter transferred at equilibrium.
Sec. 14.41
Results
353
(ii) The determination of the validity of the process.

(iii) The effect of parameters of interest, and especially the effect of the thickness of the
shell on the retardation of drug delivery.
14.4.1 DATA
Experiments are carried out under the operational conditions described in the
experimental part with
various
values of the thickness of the shell.
The diffusivities are determined either with pure Eudragit samples and with samples
having the same composition as the core. From the straight lines obtained by plotting the
amount of matters transferred during short tests as a function of the square root of time,
the diffusivity is determined by using eqn.( 14.4).
The values of the matters transferred at equilibrium are calculated from the long
experiments when equilibrium is reached.
The values of the diffusivities and amount of matter transferred at equilibrium are
shown in Table 14.1.
Table 14.1.Diffusivities and matters transferred at equilibrium
D (cm 2 /s)
Moo(%)
38
Drug (in core and shell)

Liquid (in core)
4,5 x 10M4x exp
55
Liquid (in shell)
1.4 x lo-
29
Some conclusions of interest are noticed :

(i) The diffusivity of the liquid is constant in pure Eudragit, but it depends largely on the
initial concentration of the drug located previously in the core.
(ii) The diffusivity
of the drug is strongly affected by the concentration of the liquid,
increasing largely with the concentration of liquid.

(iii) The amount of liquid transferred in the polymer is of 55 %, either in the shell or in the
core when only the polymer is considered.
354
[Ch. 14
x
Time IhI
I
20
Time IhI
I
28 0
10
I
20
Fig. 14.3. Kinetics of transfer for the drug (left) and liquid (right), Rt = 0.435cm.
Thickness of the shell = 0.002 cm.
14.4.2
VALIDITY
OF THE
MODEL
Experimental and calculated kinetics of transfers obtained for the drug (left) and for
the liquid (right) are drawn in Fig.14.3 - 14.8, for various values of the thickness of the
shell ranging from 0.002 to 0.02 cm.
In all cases, good agreement between the theoretical and experimental curves can be
appreciated either for the liquid and for the drug, during the whole process.
The following conclusions can be drawn :
(i) The model describes the complex process in a perfect way.
(ii) The data obtained from short tests are of interest, and especially the diffusivities and
their concentration-dependence laws.
(iii) The model can be used under various conditions, especially with various sizes of the
samples.
14.4.3
EFFECT
OF PARAMETERS
A parameter of interest appears with the thickness of the shell with respect to the
radius of the core. The effect of the thickness of the shell on the kinetics of drug delivery
Sec. 14.41
Results
355
MSS
O/O
30,
ML
IY
0
Time(h)
10
20
Timelh)
28 0
10
20
Fig. 14.4. Kinetics of transfer of the drug (left) and liquid (right).Rl = 0.43cm.
10
20
28 0
10
20
356
[Ch. 14

AMSS
AML
O/O
O/O
30 -
20,
lO_
+t
0
10
20
I-
w
0
10
20
Fig. 14.6. Kinetics of transfer of the drug (left) and liquid (right).R1 = 0.425cm.
Sec. 14.41
Results
10
20
30
LO
357
50
Fig. 14.8. Kinetics of transfer of the drug (left) and liquid (right) Ri = 0.415cm.
Thickness of the shell = 0.02cm.
is of high concern. A small value for the thickness of the shell can already result in a
change in the shape of the kinetics curves of the drug. This fact is appreciated in Fig.14.3
or 14.4 for a thickness of only 0.02 or 0.03 mm. Of course this fact highly increases as
the thickness is increased.
14.4.4
PROFILES
OF
CONCENTRATION
The numerical model is able to give more information than experiments. For
instance, it is possible to obtain the profiles of concentration of the liquid and the drug as
they are developed through the sample as a function of time. It is difficult to make some
measurements for proving the accuracy of these profiles (12,13). However, as the
kinetics of the matters transferred are obtained by integrating these profiles of
concentration, the validity of the numerical model can be expanded in the profiles of
concentration. Some profiles of concentration are drawn for the drug (left) and the liquid
(right) in two cases, for a thickness of the shell of 0.02 mm (Fig.14.9) and of 0.2 mm
(Fig.14.10).
[Ch. 14
358
c L P/o)
I6
c5 P/o)
Middle
RI R2
Middle
RI R2
Fig. 14.9. Profiles of concentrationof the drug (left) and liquid (right) through
the sample.Rr = 0.435cm. Shell thickness= 0.002 cm.
The following conclusionscan be drawn from theseprofiles of concentration:

(i) The liquid penetratesthe sample,and a strong discontinuity is observedat the interface
betweenthe shell and the core at the beginning of the process.This is due to the various
values of the liquid at equilibrium found in pure Eudragit and in the core with 50 %
Eudragit.
(ii) Steepgradientsof concentrationof drug are developedthrough the sample.
(iii) The shell plays the role of a limitation for the drug aswell as for the liquid.
14.5
CONCLUSIONS
Some new dosage forms are built and tested in order to prepare devices able to
deliver the drug with a constant rate. With these new dosageforms made of a core and
shell, the rate of releaseis very low at the beginning of the process and then increases
progressivelywith a positive acceleration.
Sec. 14.51
Conclusions
359
CL
O/O
50
60-
30
20
10.
10 /
50
20-
0
Middle
RI
R2
Middle
RI
R2
Fig. 14.10. Profiles of concentration of the drug (left) and liquid (right) through
the sample.Rl = 0.415 cm. Shell thickness = 0.02 cm.
As the kinetics of release of the drug obtained with this new dosage form and the
kinitics of release with a simple dosage form with a core alone are symmetrical with
respect to the bisectrix, it is thus understandable that a mixture of these two dosage forms
can give kinetics of release with a constant rate.
REFERENCES
1
J.M. Vergnaud. in Liquid Transport Processes in Polymeric Materials.Modelling
and Industrial Applications. Prentice Hall ed. 1991. chapter 10.

A. Droin, C.Chaumat, M. Rollet, J.L. Taverdet and J.M. Vergnaud.Model of
2
Pharm. 27 (1985) 233-243.
360
3
[Ch. 14
I. Malley, J. Bardon, M. Rollet, J.L. Taverdet and J.M. Vergnaud. Modelling of
controlled drug-release in case of Carbopol-sodium salicylate matrix in gastric liquid.

Drug Dev. Int. phann. 13 (1987) 67-79.
J.Y. Armand, F. Magnard, J. Bouzon, M. Rollet, J.L. Taverdet and J.M
4
Vergnaud. Modelling of the release of drug in gastric liquid from spheric galenic forms
with Eudragit matrix. Int. J. Pharm. 40 (1987) 33-41.
5
P. Magron, M. Rollet, J.L. Taverdet and J.M. Vergnaud. Spherical oral polymer-
drug device with two polymers for constant drug delivery. Int. J. Pharm. 38 (1987)
91-97.
6
N. Laghoueg, J. Paulet, J.L. Taverdet and J.M. Vergnaud. Oral polymer-drug
devices with a core and an erodible shell for constant drug delivery. Int. J. Pharm. 50
(1989) 133-139.
H. Liu, P. Magron, J. Bouzon and J.M. Vergnaud. Spherical dosage form with a
7
core and shell. Experiments and modelling. Int. J. Pharm. 45 (1988) 217-227.
9
H. David, J. Bouzon and J.M. Vergnaud. Modelling of absorption and desorption
of a liquid by a polymer device made of a core and shell. Europ. Polym. J. 25 (1989)
89-94.
10
H. David, J. Bouzon and J.M. Vergnaud. Modelling of desorption of liquid from
an EVA polymer device composed of a core and shell. Plast. Rubber. Proces. Applic. 11
(1989) 9-16.
11
H. David, J. Bouzon and J.M. Vergnaud. Modelling of matter transfer with a
polymer device made of a core and shell. Effect of the capacity of absorption by the
silicone shell. Europ. Polym. J. 25 (1989) 939-945.
12
D. Messadi and J.M. Vergnaud. Simultaneous diffusion of benzyl alcohol into
plasticized PVC and of plasticizer from polymer into liquid. J. Appl. Polym. Sci. 26
(1981) 2315-2324.
Sec. 14.51
Conclusions
361
13 D. Messadi, J.L. Taverdet and J.M. Vergnaud. Plasticizer migration from

plasticized PVC into liquids. Effect of severalparameterson the transfer. I and EC Prod.
Res. Dev. 22 (1983) 142-146.
15
Controlled rate of delivery when the
solubility of the drug is low, by using
a swelling polymer
15.1
INTRODUCTION
The fit purpose in this chapter is to show that using a swelling polymer can be of
interest for controlling the releaseof a drug, especially when its solubility is very low. As
shown in earlier studies (l-3), the drug was dispersedin a swelling polymer. The aim of
this study is to use the swelling polymer as an additive which is dispersed in another
polymer matrix, in the sameway as the drug. The new device is thus made of a polymer
matrix in which the drug and a swellable polymer are dispersed,the polymer matrix being
either a non-erodiblepolymer such asEudragit (4) or an erodible polymer such as Gelucire
(5).
As shown in Chapter 13, dosage forms consisting of a core and shell, with an
erodible shell made of Gelucire, are able to deliver the drug with a constant rate.
Following this way, an attempt is also made by testing a dosage form with a core and
shell, the core being obtained by dispersing the drug and the swelling polymer in an
erodible polymer playing the role of a matrix, and the shell being made of an erodible
polymer (5).
Another purpose in this chapteris to consider a drug which is very poorly soluble in
gastric liquid and poorly soluble in intestinal liquid. Moreover, it is desired that a small
part of the drug of the dosageform is releasedin the stomach,while the larger part of the
drug is delivered in the intestine. In order to resolve this difficult problem, the swelling
364
Controlled
[Ch. 15
rate of delivery
polymer is selectedin such a way that its high capacity of absorptionof water dependson
the pH of the liquid : rather high in acid liquid, it is still higher in a liquid of pH 8.
Experiments are performed by immersing the dosage form in a synthetic gastric
liquid and then in another liquid of pH 8, successively, in order to simulate the
gastrointestinalpath of the dosageform.
15.2 DOSAGE FORM

SWELLING
POLYMER
15.2.1
WITH
(4)
A POLYMER
MATRIX
AND
THEORETICAL
ASSUMPTIONS
The following assumptionsaremade :

(i) The transportsof liquid into and drug out of the dosageform are controlled by transient
diffusion.
(ii) The drug, as well as the swelling polymer (Sumikagel) are well dispersed into the
polymer matrix .
(iii) The dimensionsof the dosageform arekept constantduring the process.
(iv) The diffusivities are constant.
(v ) The concentrationof liquid and drug on the surfacesreachesthe equilibrium values as
soon asthe dosageform is immersed.
(vi) The volume of the liquid is so large with regard to that of the dosageform that it can
be consideredas infinite.
The transient diffusion for the liquid and drug through the spherical dosageform is
describedby the equationwith constantdiffusivity :
(15.1)
D
% =
a%+;
.
&2
ac
r?F
With the above assumptions, an analytical solution exists for the problem. The
concentration C,, at position r and time t is given by the trigonometrical series for either
Sec. 15.21
Dosage form with a polymer
matrix
365
the drug or the liquid :
crJ -%l+g.cp
(15.2)
cs-
tin
= (- l)n
sin?.
n
2 2
-nlFDt
exp
1
n=l
R2
where C, is the initial concentration of the drug or liquid (C, = 0 for the liquid) and C,
denotes the concentration on the surface (C s = 0 for the drug, C, = C, for the liquid).
An easy simplification is obtained when r = 0, at rhe middle of the sphere,
nrcr n7tr
tends to 1.
since sin R I R
The total amount of diffusing substance Mt entering (liquid) or leaving (drug) the
spherical dosage form is given by :
2 2
Err:1
R2
where M, is the amount of matter transferred after infinite time.
15.2.2
EXPERIMENTAL
MATERIALS
The following materials are used for preparation of dosage forms : sodium salicylate
as the drug ; Sumikagel (Sumikagel, Osaka, Japan) in the form of a white powder is the
swelling polymer. This acrylic acid-vinyl alcohol copolymer is safe in terms of health
hazard. The grade selected in this study (SP-520) corresponds to a diameter average of 20
pm. The capacity of absorption (w/w %) ranges from 500-700 in pure water to 40-60 in
saline
(0.9
% NaCl).
dimethylaminoethylacrylate
Eudragit
RL
(Rhom
Pharma)
is a copolymer
of
and ethylmethacrylate of molecular weight 150 000, with a
low number of quatemary ammonium terminal groups.The ratio between the numbers of
quaternary and ester terminal is around 1 : 20.
DOSAGE FORMS
Two types of dosage forms are prepared and tested as follows.

Dosage fonn I : sodium salicylate and Eudragit RL (50-50 wt %), in powder forms,
366
Controlled
rate of delivery
[Ch. 15
are intimately mixed and transformed into a thick paste by adding a small amount of
ethanol which is not a solvent for thesematerials. Spherical beadsare preparedfrom this
pasteand dried to completion at room temperaturefor 4 days.
Dosageform II : sodium salicylate,Eudragit RL and Sum&age1(45-45-10 wt %), in
powder forms, are intimately mixed, and the paste obtained with a small addition of
ethanol is transformed into sphericalbeads.Thesebeadsare dried in the sameway as for
dosageform I.
IN VITRO TESTS
Experimentsare carried out in a closedflask with controlled temperature(37C) and

rate of stirring. The dosageform in a glassfibre basketis immersedin 200 ml of liquid.
Two types of liquid are used : (i) synthetic gastric liquid at pH 1.2, obtained by
adding 80 ml HCl (1N) and 2 g NaCl in 1000 ml of aqueous solution ; (ii) synthetic
intestinal liquid at pH 8, with 50 ml of 0.025 M borax solution and 20.5 ml HCl (0.1 N).
Small samples (0.2 ml) of liquid are taken at intervals, for measuring the
concentrationof drug with the help of a UV spectrophotometer(Hitachi 1100)calibratedat
207 nm, while the dosageform is weighed.
Experiments needing a change in solution are performed by immersing the dosage
beadin syntheticgastricliquid, and then in syntheticintestinal liquid, successively.
15.2.3
RESULTS
Two types of resultsare of interest :

(i) The kinetics of drug releaseaswell as of absorptionof liquid obtainedwith oral dosage
form II (composed of Eudragit-Sumikagel-drug, 45 : 10 : 45 wt %), which compare
favourably with the kinetics obtainedwith dosageform I (comprising Eudragit-drug 50-50
wt %). These kinetics are determined either in synthetic gastric liquid at pH 1.2 or in
syntheticintestinal liquid at pH 8.
(ii) In vitro tests are used to simulate the gastrointestinaltract, by immersing the dosage
form firstly in syntheticgastricliquid andthen in syntheticintestinalliquid.
KINETICS OF TRANSPORT WITH THE NEW DOSAGE FORMS
The new oral dosageforms (dosageform II) are testedby immersing them either in
synthetic gastric liquid at pH 1.2 or synthetic intestinal liquid at pH 8. From the
measurementsmade on the concentration of drug in the liquid and in the weight of the
dosageform, the kinetics of releaseof drug are determinedas well as those of absorption
Sec. 15.21
matrix
367
t(h)
15.1 Kinetics of releaseof drug (left) and absorptionof liquid (right) in synthetic
gastricliquid. Dosagefonn II (Eudragit RL - Sumikagel - sodium salicylate45 : 10 :
45 wt %). 142.7mg - R = 0.36 cm.+ experimental---- calculatedmatter transferred
vs (time)0.5
of liquid by this form.
The kinetics determined from experiments are depicted in Fig. 15.1 for the caseof
dosageform JI immersed in synthetic gastric liquid, and in Fig. 15.2 for immersion in
synthetic intestinal liquid. From the slopes of the straight lines obtained by plotting the
amount of matter transferredas a function of the squareroot of tune, constantdiffusivities
are calculatedwith the help of eqn. (15.4).
0.5
(15.4) j$
ca
= g ($)
The kinetics calculatedby using eqn (15.3) for either the liquid entering or the drug
leaving the dosageforms are also shown in Fig. 15.1 for gastric liquid and in Fig. 15.2
for intestinal liquid.
368
Controlled
[Ch. 15
rate of delivery
Atbo/O
Tco
Mto,oA
MO
75,
375
SO-
250
25-
125
1
0
t(h)
15.2 Kinetics of releaseof drug (left) and absorptionof liquid (right) in synthetic
intestinal liquid. Dosageform II (Eudragit RL - Sumikagel- sodium salicylate45 :
10 : 45 wt %). 72 mg - R = 0.39 cm.+ experimental----calculatedmatter transferred
vs (tirne)0.5
The same kinetics are also observed and calculated for dosage forms of type I,
composedof Eudragit and drug (50-50 wt %), in either synthetic gastric liquid (Fig. 15.3)
or intestinal liquid (Fig. 15.4).
A number of conclusionscan be drawn from theseresults :
(i) The processof drug delivery from dosageforms of type I (Eudragit-drug, 50-50 wt %)
is controlled by transient diffusion, as shown previously. In fact, the process is more
complex : the liquid entersthe polymer, dissolvesthe drug and enablesthe drug to diffuse
through the liquid locatedin the polymer.
(ii) The process for the two matter transports with dosageform II is also controlled by
transientdiffusion, with constantdiffusivities. In fact, the diffusivities of the liquid and of
the drug are constant,as shown in Figs. 15.1 and 15.2, but a slight difference appearsfor
Sec. 15.21
matrix
369
t (hl
15.3
Kinetics of release of drug in synthetic gastric liquid. Dosage form I (Eudragit
RL- sodium salicylate 50 : 50 wt 9%).176 mg. R = 0.32 cm.
the drug. The straight line obtained by plotting the amount of drug delivered as a function
of the square root of time does not pass through the origin of the axes. There is a slight
time delay for the release of drug at the beginning of the process. This is due to the fact
that a significant amount of liquid in the dosage form is needed to dissolve the drug and
thus permit its transport.
(iii) The diffusivity of drug in the case of dosage form I is about the same when the drug is
immersed in synthetic gastric liquid or in synthetic intestinal liquid. All the drug located in
the dosage form is released after a period of about 200 h.
(iv) The diffusivity of the drug is higher in the case of dosage form II than for dosage
form I, as shown in Table 15.1 This difference which is already significant in the liquid of
pH 1.2 is even more sensitive in the liquid of pH 8.
(v) The diffusivity in the case of dosage form II is higher for the liquid than for the drug
when there are immersed in synthetic gastric liquid. The diffusivity is a little lower for the
liquid than for the drug in the case of intestinal liquid. The amount of liquid absorbed by
dosage form II is 5-fold higher in the case of intestinal liquid (500 9%)as compared to
370
Controlled
rate of delivery
[Ch. 15
SO
tlh)
15.4 Kinetics of releaseof drug in syntheticintestinal liquid. Dosagform I (Eudragit

RL - sodium salicylate 50 : 50 wt %). 153 mg. R = 0.31 cm
gastric liquid (100 %).

(vi) The large amount of liquid absorbed by dosage form II is due to the presence of
Sumilcagelwhich is capableof swelling to a greaterextent when in contact with liquid at
pH 8 (500 wt %). Moreover, a synergistic effect is clearly obtained with the liquid at pH
8, as the increase in the amount of liquid absorbedby dosage form II is considerably
higher than the amountof liquid due to Sumikagel.
SIMULATION OF THE GASTROINTESTINAL TRACT
For simulating the gastrointestinaltract, dosageforms II are immersed successively

in synthetic gastric liquid and then in intestinal liquid. The time of immersion in synthetic
gastric liquid is a parameterof interest,and is selectedto be within the range 2-4 h.
In order to gain further insight into the process,especially at the end of the stageof
immersion in synthetic gastric liquid, the profiles of concentration developed within the
dosageform are plotted for the drug (Fig. 15.5, left) and the liquid (Fig. 15.5 right). Of
Sec. 15.21
371
matrix
Table 15.1Diffusivities D x lo8 (cm2/s)and amountof matterstransferred.

lOOxMJ&
Gastric
Intestinal
liquid
liquid
Diffusivities
Intestinal
Gastric
liquid
liquid
Dosage
form
Drug I
Drug II
5.6
28
163
Liquid II
5.7
65
99
99
99
99
100x MJrn()
100
500
48
course, these profiles are calculated by using eqn(15.2) and the values of diffusivities
obtainedin the gastricliquid.
The kinetics of drug delivery and of liquid absorptionare illustrated for various times
+%
25-
RO
15.5 Profiles of concentrationdevelopedwithin the dosageform II in synthetic

gastric liquid. Drug (left) - Liquid (right)
372
Controlled
12
rate of delivery
[Ch. 15
1S.6 Kinetics of releaseof drug (left) and absorptionof liquid (right) from dosage
form II (Eudragit RL - Sumikagel - sodium salicylate45 : 10 : 45 wt 9%).Immersion
of 2 h in synthetic gastric liquid, followed in intestine liquid. 152 mg. R = 0.366 cm.
of immersion in gastric liquid : 2 h (Fig 15.6) 3 h (Fig. 15.7) and 4 h (Fig. 15.8).
Someresultsare worth noting :
(i) The drug is, of course,deliveredpartly in gastric liquid and in intestinal liquid.
(ii) The rates of delivery of the drug are higher in intestinal than in gastric liquid, and are
responsiblefor the changein shapeof the kinetic curves at the time at which the dosage
forms are immersedin intestinal liquid.
(iii) Eqns.(15.2) and (15.3) can be used for calculating the profiles of concentration and
the kinetics of drug delivery only during the initial stageof immersion in gastric liquid, as
theseequationsare determinedwith an initially uniform concentrationwithin the dosage
form. A numerical model is thus necessaryfor calculatingthe kinetics of releaseof drug in
the secondstageof immersion in intestinal liquid.
Sec. 15.31
373
Dosage form with gelucire
Mt
-O/0
m0
100
50
t(h)
0
15.7 Kinetics of releaseof drug (left) and absorptionof liquid (right) from dosage
form II (Eudragit RL - Sumikagel - sodium salicylate45 : 10 : 45 wt %), immersed
for 3 h in synthetic gastric liquid, following in intestine liquid. 146 mg. R =
0.362 cm.
15.3 DOSAGE
POLYMER
FORMS
WITH GELUCIRE
AND A SWELLING
15.3.1 THEORETICAL
The theoretical part is about the sameas that describedin Section 15.2, especially if
the diffusion of the mattersis consideredthrough the Gelucire matrix (10).
15.3.2
EXPERIMENTAL
MATERIALS(5)
The following components have been used. Sodium salicylate in powder form
(COPER) for the drug, Gelucire 46.7 (GateffosC,1983) for the erodible polymer matrix.
374
Controlled
rate of delivery
[Ch. 15
Lo/0
in
t(h)
0
12
15
15.8 Kinetics of releaseof drug (left) and absorptionof liquid (right) from dosage
form II (Eudragit RL - Sumikagel - sodium salicylate45 : 10 : 45 wt %), immersed
for 4 h in synthetic gastric liquid, following in intestineliquid. 137mg - R =
0.355 cm.
Gelucire is a mixture of polyglycide fatty esterswith well defined hydrophilic properties.

The ratio betweenthe number of quatemaryammonium and esterterminals is about 1 : 20.
Gelucire 46.7 melts at 46 % C (drop point : Mettler) andits hydrophilic-lipid balancevalue
(HLB) is 7, in the middle of the 1-14 range.Sumikagel (Sumitomo Chemical Co. Ltd,
Japan) in white powder has been used for the swelling polymer. This is an acrylic acidvinyl alcohol copolymer. The SP-520 grade with a diameter average of 20 km for the
grain has been selected.Its absorbingcapacity in w/w % rangesbetween 500 and 700 in
pure water and 40 and 60 in salted water (0.9 % NaCl). This resin is safe in terms of
health hazard(Sumikagel).
The grains of sodium salicylate and Sumikagel intimately mixed are dispersed in
Sec. 15.31
375
t(h)
15.9 Kinetics of releaseof drug in 200 ml of syntheticgastric liquid (pH 1.2) from
the dosageform Gelucire-sodiumsalicylate 50 : 50 wt %.
melted Gelucire heatedto around5OC.The liquid mixture is stirred thoroughly in order to

have the componentsdistributed properly. Various spherical beads are prepared for this
paste,with a ratio in wt % : sodium salicylate45 %-Sumikagel, 10 %-Gelucire 45 %.
IN VITRO TESTS
Experiments are carried out in a closed flask with control of the rate of stirring and
temperature(37C). The beadin a glassfibre basketis immersedinto 200 ml of liquid.
Two kinds of liquid have beenprepared : the one simulating the acid gastric liquid,
with 1000ml of aqueoussolution with 80 ml HCl (1N) and 2 g NaCl at pH 1.2 ; the other
simulating the intestine liquid with 50 ml of 0.025 M borax solution and 20.5 ml HCl(O.1
N), at pH 8.
A small sample (0.1 ml) of liquid is extracted for analysis of the drug, by using a
W spectrophotometer(Hitachi U- 1100) calibrated at 207 nm, after dilution in 50 ml of
liquid at pH 1.2.
For the experimentsneeding a changeof solution, the bead in the fiber-glass basket
376
Controlled
rate of delivery
tCh. 15
t(h)
15.10 Kinetics of releaseof drug in 200 ml of syntheticintestineliquid (pH 8) from

the dosageform Gelucire-sodiumsalicylate 50 : 50 wt %.
is extracted ou of the first solution, and immersed in the second solution. In our
experiments,the pH of the first solution is 1.2, while the pH of the secondsolution is 8.
Great care is taken to eliminate all the liquid wetting the bead and the basket before
reimmersingit in the secondsolution.
153.3 RESULTS AND DISCUSION
Two kinds of results are of interest : (i) the study of the behaviour of the new dosage
forms with Gelucire and Sumikagel as polymer matrices when they are immersed into a
liquid of pH 1.2 or a liquid of pH 8, by considering especially the kinetics of delivery of
the drug ; (ii) the kinetics of delivery of the drug when the dosage form is immersed
successivelyin a liquid of pH 1.2 and in a liquid of pH 8.
KINETICS OF DELIVERY OF THE DRUG OUT OF THE NEW DOSAGE FORM (5)
More simple dosageforms consisting of Gelucire 46.7 (50 wt %) and Na salicylate

(50 wt %) are studied firstly by immersing the dosageforms either in synthetic gastric
Sec. 15.31
377
t(h)
15.11 Kinetics of releaseof drug in 200 ml of syntheticgastric liquid (pH 1.2) from
the dosageform Gelucire-sumikagel-drug45 : 10 : 45 wt %.
liquid (pH 1.2) or in synthetic intestinal liquid (pH 8). The dosage forms have the
characteristicsdescribedin Table 15.2. The experimental (plots) aswell as the theoretical
kinetics of delivery of the drug are drawn in Fig. 15.9 when the liquid is the synthetic
gastric liquid, and in Fig. 15.10when the pH of the liquid is 8. In each figure, the amount
of drug in the liquid is also plotted as a function of the squareroot of time. The following
conclusionscan be drawn :
(i) A straight line is obtained by plotting the amount of drug delivered as a function of the
squareroot of time, at the beginning of the processwhen M@& < 0.4.
(ii) The theoretical curves obtained by using the series in eqn (15.4) are very well
superimposedwith the experimentalkinetics.
(iii) The diffbsivities for the transportof drug are about the samewhen the pH of the liquid
is 1.2 or 8.
The behaviour of the new dosageforms containing (10 %) Sumikagel besidesthe
drug (45 %) and Gelucire (45 %), is described in Fig. 15.11 in the case of synthetic
378
Controlled
rate of delivery
[Ch. 15
15.12 Kinetics of releaseof drug in 200 ml of syntheticintestineliquid (pH 8) from

the dosageform Gelucire-sumikagel-drug: 45 : 10 : 45 wt 9%.I : 125mg II: 131mg
gastric liquid and in Eig.15.12 when the pH of the liquid is 8. The following facts are
worth noting :
(i) Transientdiffusion can describethe processof drug delivery from the new dosageform
in syntheticgastric liquid.
(ii) The diffusivity is of the sameorder of magnitudeasthat obtainedwith the dosageform
without Sumilcagel,in the caseof syntheticgastricliquid.
(iii) The kinetics for the delivery of the drug in the liquid of pH 8 are quite different. They
cannot be described by transient diffusion. The rate of drug delivery is about constant
during the whole process.
(iv) In the case of Fig. 15.12 with the liquid of pH 8 the rate of delivery is not only
constant but also higher than that with the liquid of pH 1.2, as all the drug is delivered
within a period of time shorter than 1 h, when the radius of the dosageform is around
0.35 cm.
Sec. 15.31
379
(v) The oral dosage form disintegrates slowly and regularly in the synthetic intestine
liquid. The polymer Sumikagel absorbswater to such a high extent that it makesburst out
the dosageform layer after layer.
Table 15.2Characteristicsof the oral dosageforms.
D (x 107)
(cm>
pH of
liquid
0.28
93.2
0.26
Weight
Radius
(mg)
120.1
Figures
(-as)
Moo
(after 1 h)
1.2
6.6
50%
5.7
50%
4.9
37%
15.9
15.10
15.11
15.12.1
15.12.11
15.13.1
15.13.11
15.13.111
120.5
0.35
1.2
125.2
0.354
96%
130.8
0.36
95 %
126.8
0.356
1.2-8
122.5
0.352
1.2-8
129.3
0.358
1.2-8
KINETICS OF DRUG DELIVERY
OUT OF NEW DOSAGE FORMS ON IMMERSION
IN
SYNTHETIC GASTRIC LIQUID AND THEN SYNTHETIC INTESTINAL LIQUID
Three dosagefonns with the samecomposition (as shown in Table 15.2) and with
about the same dimensions ranging from 122.5 mg to 129.3 mg, are immersed
successivelyin synthetic gastric liquid and then in synthetic intestine liquid. The time of
immersion in synthetic gastric liquid is different for each dosageform, as shown in Fig.
15.13 : 2 h for curve I, 3 h for curve II, and4 h for curve III.
From the kinetics drawn in Fig. 15.13,the following conclusionscan be drawn :
(i) The processof drug delivery is followed, when the dosageforms are immersed in the
secondsolution of pH 8.
(ii) Of course,the kinetics of drug delivery are describedby transient diffusion in the first
stageof the process,when the pH of the liquid is 1.2.
(iii) A higher rate for the drug delivery is clearly shown, when the dosage form is
immersedin the secondsolution of pH 8. This increasein the rate of delivery is especially
high when the time of contact of the dosageform with the first liquid of pH 1.2 is short,
according to the following statement: the shorterthe time of contactwith the acidic liquid,
the higher the increasein the rate of delivery just after the changein aqueoussolution.
(iv) Thesetimes of contact of the dosageform with the acidic liquid ranging from 2 to 4 h
Controlled
380
[Ch. 15
rate of delivery
t(h)
1
15.13 Kinetics of releaseof drug from the dosageform Gelucire Sumikagel-drug45 :

10 : 45 wt %Iwhen immersedsuccessivelyin gastricliquid and intestineliquid. Time
of immersion in gastric liquid : 2 h (I) - 3 h (II) - 4 h (III)
can representthe time of transit of the dosageform in the stomachand in intestine.
(v) The oral dosage forms disintegrate slowly but regularly in the synthetic intestinal
liquid. This property is of interest when the solubihty of the drug is very low, becausethe
drug is thus disposedinto a large volume of liquid.
15.4 DOSAGE
FORM WITH
POLYMER-DRUG-EUDRAGIT)
POLYMER
15.4.1
A CORE
AND
AN
(SWELLING
ERODIBLE
THEORETICAL
ASSUMPTIONS (6)
The following assumptionsare made in order to simplify the problem and to get an
Dosage form with a core (swelling
Sec. 15.41
Polymer-Drug-Eudragit)
381
analytical solution.
(i) Dosageform I (or core) is obtained by dispersing the mixture of Eudragit RL 45 %Sumikagel lO%drug 45 %. The distribution of the drug and Sumikagel in the polymer
matrix is perfect.
(ii) Dosage form II consists of the core (dosage form I) surrounded by Gelucire. The
thicknessof the shell is constant.
(iii) Only diffusion of the drug is considered,in spite of the fact that the liquid entersthe
beadandthus enablesthe drug to leave.
(iv) Transfer of the drug is controlled by transientdiffusion with constantdiffusivity.
(v) Concentrationof the drug is least at the beginning of the processconstanton eachface
of the shell, Ci, on the internal surfaceand zero on the external surface.
(vi) Thicknessof the shell remainsconstantat least at the beginning of the process.
Transient diffusion through the spherewith constantdiffusivity is expressedas :
(15.1) $+=D.
$+;.
Initial andboundaryconditionsbecomes:
(15.5)
(15.6)
t=O
t>o
r IR,
R,cr<R2
c=o
core
shell
r = R,
tin
internal surface of the
r=
c=o
external surface of the
tin
shell
R2
shell
In spite of the high complexity of the problem, an analytical solution can be found
with the assumptions.
The amountof the drug leaving the shell is obtainedby integratingthe rate of transfer
on the externalsurfacewith respectto time :
(15.7) Mr=-4nR;.
,D. %,A
s
I 1
for
r= R,
382
Controlled
rate of delivery
[Ch. 15
The amount of drug leaving the dosageform is thus given as :
(15.8)
Mt
4rR1.R2.(R2-Rl)Ci*=(R2
Dt
-R1)
----1
6
-n2rc2D
2 - -(- l)*
c
X2n=l n2 exp R2- Rl)2 t
i(
I
The series tends to zero for high values of time, and the rate of drug delivery can
thus be expressedas a function of the thicknessof the shell by the simple relation :
dM,
(15.9) 7=
4Ir.R,.R:!.Ci,.D
R2-Rl
15.4.2
EXPERIMENTAL
MATERIALS
The dosageforms necessitatethe following materials:

- Sodium salicylate as the drug.
- Eudragit RL (Rhom Pharma), a copolymer of dimethylaminoethylacrylate and
ethylmethacrylateof M = 150,000,plays the role of the polymer matrix.
- Sumikagel SP-520 (Sumitomo Chem. Co. Japan) in powder form is the swelling
polymer. This is an acrylic - vinyl alcohol copolymer which is safe in terms of health
hazard. The capacity of absorption (wt/wt %) dependslargely on the pH of the liquid,
ranging from 500-700 in pure water to 40-60 in saltedwater at pH 1.2.
- Gelucire 46.7 (Gattefosse, France) is a waxy solid made of partial glycerides and
polyglycides fatty esterswith hydrophilic properties. Gelucire 46.7 melts at 46C and its
hydrophilic-lipidic balancevalue (HLB) is 7 at the middle of the -14 range.
DOSAGE FORMS
Two types of dosageforms areprepared:

Dosage form I : Sodium salicylate, Sumikagel and Eudragit RL in powder form are
intimately mixed and then transformed into a viscous paste by adding a small amount of
ethanol. Spherical dosageforms are pressedfrom this paste and dried. The composition
(in weight) is : Sodium salicylate45 % ; Sum&age110 % Eudragit 45 %.
Dosage form II : Dosage forms II are prepared by surrounding dosage forms I with a
Sec. 15.41
383
spherical coating of Gelucire 46.7 ; this is done by immersing them in liquid Gelucire at
60C for 2-5 s. The coating becomeshard after cooling at room temperature.The thickness
of the Gelucire shell is determined by selecting immersion time or by immersing the
dosageform with coating many times in Gelucire.
IN VITRO TESTS
Experimentsare carried out in a closed flask at a constanttemperature(37C) with a

controlled rate of stirring. The dosageforms in a glassfibre basketare immersedin 200 ml
of liquid. Two kinds of liquid are used : i) Synthetic gastric liquid at pH 1.2, obtained
with 80 ml HCl 1N and 2 g NaCl in 1,000ml aqueous; ii) synthetic intestinal liquid at pH
8, obtained with 50 ml of 0.025 molar Borax solution and 20.5 ml HCl 0.1 N. The
solution 0.025 molar Borax is obtained with 9.534 g of Borax (M = 381.37 g) in 1,000
ml aqueoussolution.
A small amount of liquid (0.1 ml) is taken at intervals for analysis by UVspectrometryafter dilution, with a UV-spectrophotometercalibratedat 207 run.
Experiments needing a changein solution are performed by immersing the bead in
the glassfibre successivelyin the acid solution and then in the other solution at pH 8.
15.4.3
RESULTS
Two kinds of resultsare of interestin this study :

(i) The behaviour of the new dosageforms with a core and shell as comparedto that of the
dosage forms made of the core alone, when they are allowed to stand in contact with
gastricliquid or intestineliquid.
(ii) The simulation of the gastrointestinaltract by using in vitro tests.
KINETICS OF RELEASE OF DRUG FROM THE NEW DOSAGE FORMS
The kinetics of release of drug are determined by using the dosage forms I (core
alonewith drug 45 %-Sumikagel 10 %-Eudragit Rl45 %) or the dosageforms II (with the
core and shell), when they are immersedeither in synthetic gastric liquid at pH 1.2 or in
syntheticintestineliquid at pH 8.
IN VITRO TESTS IN GASTRIC LIQUID
The kinetics of releaseof Sodium salicylate are drawn in Fig. 15.14,when they are
obtained from dosageforms I and from the dosageforms II. The effect of the shell on the
kinetics is illustrated in these curves. The kinetics of releasefrom the dosageform I is
384
Controlled
[Ch. 15
rate of delivery
12
15.14 Kinetics of release of drug in synthetic gastric liquid at 37C. Dosage forms I
(143 mg - R, = 0.36 cm). Dosage form II (144 mg - R, = 0.362 cm - R2 - R, =
0.031 cm).
essentially controlled by diffusion, as shown in earlier studies (7,9, 10) : the liquid enters
the dosage form, dissolves the drug, and allows the drug to leave the dosage form by
diffusion through the liquid located in it. Both these transfers are controlled by transient
diffusion. The typical facts of diffusion appear here with the vertical tangent at the
beginning of the process, and the rat; which decreasesexponentially with time.
The kinetics of release of the drug from the dosage forms II are quite different from
the kinetics obtained with the forms I. The rate is about constant over a long period of
time, and the rate is not very high at the beginning of the process as it is for the dosage
forms I.
It is of interest to have a model able to describe the process, even though many
drastic assumptions are made to build it, because the results can be expressed
quantitatively by way of a relationship. With the simple model of the shell playing the role
of a spherical membrane separating two media of constant concentration, the rate of
delivery is expressed in terms of the thickness of the shell by eqn (15.9). The rate of drug
Sec. 15.41
385
6-
2-
Rl.R2
15.15 .Rate of drug delivery from a dosage form II vs -in
R,.Rl (cm)
&-RI
D
6
gastric liquid at 37C.
R2-Rl
delivery obtained from various dosage forms II is plotted against R, . R2 / (R2 - R,),
exhibiting a straight line within
a rather wide range for the thickness of the shell (Fig.
15.15). This result is of concern especially from a practical point of view, because it is
thus possible to know exactly the characteristics of the dosage forms II for a desired
purpose.
IN VITRO TESTS IN INTESTINE LIQUID
The kinetics of drug delivery are also determined with dosage forms I and II when
they are in contact with the intestinal liquid. As for the gastric liquid, the effect of the shell
is very significant on the kinetics of release.
For dosage forms I, the kinetics are controlled by transient diffusion, with a very
high rate of delivery at the beginning of the process which correlates well with the vertical
tangent, and the rate decreasesexponentially with time. The rate of delivery as well as the
diffusivity are higher for the intestinal liquid than for the gastric liquid. These differences
are due to the fact that the polymer Sumikagel swells to a higher extent with intestine
liquid, provoking also a swelling of the polymer matrix.
386
Controlled
tCh. 15
rate of delivery
15.16 Kinetics of releaseof drug in syntheticintestineliquid. I : dosageform I

mg - R, = 0.39 cm)
II (1) Dosageforms II (130 mg - R, = 0.348 cm - R, -R, = 0.0175 cm)
II (2) Dosageforms II (164 mg - R, = 0.377 cm - R, - R, = 0.353 cm)
II (3) Dosageforms II (126 mg - R, = 0.345 cm - R, _R, = 0.053 cm)
(172
Dosageforms II with the shell display a rather constantrate of delivery at least over a
large part of the process(Fig. 15.16).This rate of delivery can also be expressedin terms
of the thickness of the shell, or more precisely by the ratio R, . R2 / (R2- R,), as shown
in Fig. 15.17 where a straight line is obtained.
The effect of the shell on the decreaseof the rate of drug delivery can be determined
by the slope of the curves shown in Figs. 15.15 and 15.17. From these slopes of 4.1 and
0.9 (per hour per cm) in the intestine and the gastricliquid, respectively,it can be seenthat
the effect of the shell is more effective in the intestine liquid when the rate of delivery is
high.
SIMULATION OF THE GASTROINTESTINAL TRACT
The gastrointestinal tract can be simulated with in vitro tests by immersing the
dosage forms first in synthetic gastric liquid for a defined period of time and then in
Sec. 15.41
387
Table 15.3Characteristics
Gastric liquid
Intestine liquid
5.6 x lo-*
5.7 x 10-s
0.9
4.1
Diffusivity Form I (cm2/)

(Figs 15.14.1and 15.16.1)
SlopesForm II (h x cml)
(Figs. 15.15 and 15.17)
synthetic intestinal liquid. Two types of experimentswere performed using dosageforms

II of various thicknessesfor the shell.
Some experiments were carried out by keeping various dosageforms II in gastric
liquid for 4 hours, which correspondsto the averagetime of the material in the stomach.
Becausethe thickness of these dosageforms differed notably from one to the other, the
kinetics of release on the gastrointestinal tract were quite different (Fig. 15.18). The
30-
20-
10,
Rl
15.17. Rateof drug delivery vs -
* R2
R2-Rl
liquid at 37C.
JML
W4
kl-il)
I
for dosageforms II in syntheticintestine
388
Controlled
[Ch. 15
rate of delivery
It
TO/O
75L
3
2
50-
b
0
12
15.18 Kinetics of releasefrom dosageforms II with various shell thicknesses,

immersedin gastricliquid for 4 h and in intestineliquid at 37C.
l(143 mg - Rr = 0.361 cm - R2 - Rr = 0.085 cm)
2 (160 mg - Rr = 0.374 cm - R, - R, = 0.046 cm)
3 (172 mg - R, = 0.383 cm - Rz - R1 = 0.025 cm)
4 (162 mg - Rr = 0.376 cm - R2 - R, = 0.016 cm)
amount of drug releaseat the end of the stagein gastric liquid can, of course,be relatedto
the dimensions of the bead and thickness of the shell by a straight line as shown in Fig.
15.19
Other experiments are performed with dosage forms II by varying the time of
immersion in the gastric liquid within the 2-5 hour range (Fig. 15.20). The four dosage
forms II are about the same, except for that which is immersed for 2 hours in gastric
liquid. The effect of the time in gastric liquid on the process of drug release is of high
interest.
Some conclusions from these results obtained with dosage forms II during the
gastrointestinaltract areworth noting.
Sec. 15.41
12
f.
389
15.19 . Amount of drug releasedafter 4 h from dosageforms II in gastricliquid vs.

Rl32
R2-Rl
(i) It is possibleto predict the rate of delivery of a drug in syntheticgastric liquid, bu using
relation 15.9 as shown in Figs 15.15, 15.17 and 15.19. This result does not correspond
exactly with the theory of the membranefor at leasttwo reasons: the thicknessof the shell
decreasesduring the process, as Gelucire dissolves slowly : the concentration on the
surfaceof the core doesnot remain constant.However, from a practical point of view, this
relationship is of high interest, as it enablesthe user to preparethe right dosageform for
this pupose.
(ii) A constantrate of delivery is obtainedin gastricliquid.
(iii) The rate of delivery in intestineliquid is not exactly constant,but it varies slightly with
time.
(iv) The rate of delivery is considerablyhigher in the intestinal liquid than in gastric liquid.
This fact, due to the high extent of swelling of Sumikagel in the intestinal liquid, is of
interestin the casewhere the drug is poorly soluble in gastricliquid and a little in intestinal
liquid.
390
Controlled
rate of delivery
[Ch. 15
3
L
t lh)
0
12
15
18
15.20 Kinetics of releaseof drug from dosageforms II during gastrointestinal

histories.
1 (163 mg - RI = 0.376 cm - R2 - R, = 0.036 cm)
2 (156 mg - R, = 0.37 cm - R2 - R, = 0.042 cm)
3 (172 mg - R, = 0.383 cm - R2 - R, = 0.034 cm)
4 (160 mg - RI = 0.374 cm - R, - R, = 0.035 cm)
15.5
CONCLUSIONS
Typical dosageforms are built and testedby using a swelling polymer which swells
to a different extent in a gastric or in intestinal liquid. As the polymer (Sumikagel) swells
to a greaterextent in intestinal liquid, it is thus able to modify the kinetics of releaseof the
drug, especially in intestinal liquid.
When the polymer matrix is Eudragit, the effect of the presenceof the swelling
polymer is not very significant. When the polymer matrix is Gelucire, the swelling
polymer increasesthe rate of disintegrationof Gelucire when the dosageform is in contact
with intestinal liquid.
The presenceof an erodible shell surrounding the core made of drug and swelling
Sec. 15.51
Conclusions
391
polymer dispersed in Eudragit is responsible for a constant rate of drug delivery is gastric
liquid and in intestinal liquid.
Simulation of the gastrointestinal history of the dosage form gives information of
interest for various dosage forms.
The presence of Sum&age1 in the dosage form or in the core is of help to increase the
rate of drug release in the intestinal liquid.
Building a model able to describe the process in this case is of interest, especially
when the drug and swelling polymer are dispersed in Gelucire. This problem is very
complex, and better knowledge of the process as well as some data such those concerned
with the liquid transport through Gelucire and the effect of the swelling of Sumikagel on
the rate of disintegration of Gelucire, are needed.
REFERENCES
N.A. Peppas, R. Gumy, E. Doelker and P. Buri. Modeling of drug diffusion
through swellable polymeric systems. J. Membrane Sci 7 (1980) 241-253.
N.A. Peppas and N.M. Franson The swelling interface number as a criterion for
prediction of diffusional solute release mechanisms in swellable polymers. J.
Polym. Sci. Phys. Ed. 21 (1983) 983-997.
N.A. Peppas. Release of bioactive agents from swellable polymers. Theory and
experiments. in Recent Advances in Drug Delivery System, eds. J.M. Anderson
and S.W. Kim., N.Y. Plenum Publishing Corp. (1984) 279-289.
D. Bidah and J.M. Vergnaud Dosage forms with a polymer matrix and a swelling
polymer Int. J. Pharm. 77 (1991) 81-87.
D. Bidah and J.M. Vergnaud New oral dosage form with two polymers : Gelucire
and Sumikagel Int. J. Pharm. 72 (1991) 35-41.
D. Bidah and J.M.Vergnaud Dosage forms with a core made of a swelling polymer
matrix with an erodible shell. J. Polym. Engng., in press.
Controlled
392
rate of delivery
[Ch. 15
A. Droin, C. Chaumat,M. Rollet, J.L. Taverdet and J.M. Vergnaud Model of

matter transfer betweensodium salicylate-Eudragitmatrix and gastric liquid. Int. J.
Pharm. 27 (1985) 233-243.
N. Laghoueg-Derricheand J.M. Vergnaud Modeling the processof drying of

dosageforms made of drug dispersedin a polymer. Int. J. Pharm. 67 (1991) 5 l57.
J.Y. Armand, F. Magnard, J. Bouzon, M. Rollet, J.L. Taverdet and J.M. Vergnaud
Modeling of the releaseof drug in gastricliquid from sphericgalenic forms with
Eudragit matrix. Int. J. Pharm. 40 (1987) 33-41.
10
D. Bidah, E.M Ouriemchi and J.M. Vergnaud Diffusional processof drug delivery
from a dosageform with a Gelucire matrix. Int. J. Pharm. 80 (1992) 145149.
16
Dosage forms with a drug attached to
a polymer dispersed in a non-erodible
polymer matrix
16.1 INTRODUCTION
The retardation in the release process of a drug over a prolonged period can be
attained by using devices made of the drug dispersed in a polymer matrix. Three various
mechanisms have been examined (l), i.e. osmosis, polymer erosion (Chapter 12) and
diffusion (Chapters 10, 14). A case of interest appears with monolithic devices prepared
by dispersing the drug into a polymer, this polymer being either a biodegradable or a nonbiodegradable matrix (2-11). More recently, another method has been explored, by
attaching the drug to a polymer by a labile chemical bond (12-17) and dispersing this
polymer into another biocompatible polymer matrix (18-21).
This chapter is devoted to the preparation and study of dosage forms in which the
drug is firstly attached to a biocompatible polymer and then dispersed into a second
biopolymer and non-erodible polymer such as Eudragit (18-21). An advantage of these
dosage forms exist, when they are chewed instead of being swallowed, it takes some tune
for the liquid to diffuse into the small grains and thus to enable the drug to diffuse out of
the chewed dosage form.
Two ways are used for attaching the drug to a compatible polymer. The first
method consists of attaching the drug to the polymer (12-15). The second method is based
on the following two steps : attaching the drug to a monomer which can be synthesized,
and then polymerizing this new monomer (16-21). Some advantages are inherent to this
394
Dosage forms with a drug attached
to a polymer
[Ch. 16
latter technique e. g. a good knowledge for the polymer matrix, a degree of substitution
reaching 100 % which is necessary for a higher yield of the drug release.
The drug delivery from the dosage forms obtained by dispersing the drug in a
polymeric matrix has been studied in various cases (Chapter 10, 12, 14). The process is
as follows (7 - 9, 11) : the liquid enters the polymer and dissolves the drug which can then
diffuse out of the dosage form through the liquid located in it.
A typical example is described when the drug is sodium salicylate and the monomer
an anhydride, and the polymer matrix is Eudragit (20). Other examples exist, as shown in
the literature survey (12-21).
16.2 THEORETICAL
ASSUMPTIONS
The whole process is rather complex due to several sucessive and simultaneous
steps : diffusion of the liquid through the polymer matrix Eudragit and into the branched
polymer ; reaction between the branched polymer and the liquid ; dissolution of the drug in
the liquid ; diffusion of the drug through the liquid located in the branched polymer and
the polymer matrix.
The following assumptions are thus made in order to simplify the problem :
(i) The spherical dosage forms are homogeneous, the branched polymer being well
dispersed into Eudragit matrix.
(ii) Two matter transfers take place (7 - 9, 11) : the liquid entering the dosage form, the
drug leaving the dosage form. They are studied not simultaneously but separately.
(iii) Both these matter transfers are controlled by transient diffusion within the dosage
form, and through the branched polymer.
(iv) The rate of the reaction between the liquid and branched polymer is very fast as
compared to the rate of diffusion.
(v) The diffusivities
are nearly constant throughout the whole dosage form, The
concentration at equilibrium is attained on the surface of the dosage form as soon as it is

immersed in the liquid.
The transient radial diffusion for the liquid and the drug is described by the Ficks
equation :
Sec. 16.31
(16.1) g
Experimental
395
= D. [$+;.g]
where D is the constantdiffusivity

and r the radial abscissain the sphere.
With the above assumptions, the solution for the problem is expressed by the
following equation (22,23).
(16.4
Mea,M
M,
00
=$.
II:
+.exp
n=l n
where M, and M, are the amount of matter transferred after time t and at equilibrium
when the processis achieved.
For very small times, the well-known equation expressing the linear relationship
between the amount of matter transferred and the square-root of time, can be used
W
whenE
< 0.3
m
This equationcan be usedfor determiningthe di&sivity
16.3
EXPERIMENTAL
PREPARATION OF ANHYDRIDE MONOMER (20)
The sodium salicylate (0.05 mole ; 8 g) is dissolved in anhydrous tetrahydrofuran

(12 ml), placed in a dry 150ml three-neckedflask, equippedwith a magnetic stirrer and a
reflux condenserwith a Drierite-filled drying tube, andkept at a temperaturearound 4-6
C with an ice-bath.
The solution of methacryloyl chloride (0.05 mole ; 5.23 g) in anhydrous
tetrahydrofuran(6 ml) is addeddropwise in the flask under stirring. The temperatureof 46 C is maintainedduring the whole addition (15 - 20 min).
After this addition, the reaction medium is gently heatedfor 20 min, after which the
396
to a polymer
tCh. 16
mixture is stirred for 4 h.

The precipitate of NaCl is filtered and washed with a small amount of dry tetrahydrofuran. The solvent is removed from the titrate in a rotary evaporator, and a white
powder of anhydride is obtained and purified by recrystallization
in a mixture of
chloroform and hexane. After drying, the amount of anhydride is 5.3Og, with a yield of
52%.
PREPARATION OF THE POLYMER
The monomer (0.0194 mole, 4 g) dissolved in anhydrous tetrahydrofuran (lcm)

with 2o/oo in weight of azobisisobutyronitrie
(8 mg) is placed in a glass tube. The air
dissolved is removed by a nitrogen stream for 10 min and the tube is sealed under
vacuum. After 16.5h of heating at 65 C, the polymer is dissolved in tetrahydrofuran (4
days) under stirring to obtain of gel. The polymer is then precipitated by addition of
petroleum ether. After drying, a yield of 90.7% is attained for the polymer (3.63 g).
CHARASTERISTICS OF THE POLYMER
IR spectrum : 1690 cm-l and 1745 cm-r : anhydride function

1600 cm-r : aromatic double bond.
Microanalysis : (Theorical / Experimental, in % weight)
H = 4.85/5.12;
0 = 31.07 /30.93
C = 64.07/63.81;
Ts (DSC) : 114 C ; glass transition temperature.
The branched polymer and Eudragit RL (a copolymer of dimethylaminoethylacryIate

and ethylmethacrylate, mol. wt. 150,000 ; Rijhm Pharma), both in powder form, are
intimately dispersed, and transformed into a thick homogeneous paste after pulverization
of a small amount of ethanol. Spherical beads are then obtained from this paste, and dried
until complete evaporation of ethanol (4 days at room temperature).
IN VITRO-TEST
These experiments are carried out in a closed flask kept at 37 C, with a controlled
rate of stirring. The beads (about 400 mg), inserted in a permeable glass fibre basket, are
soaked either in simulated gastric liquid at pH 1.2 (1000 ml of aqueous solution, 80 ml
HCl 1 N and 2 g NaCl), and in a liquid at pH 8 (50 ml borax at 0.025 M ; 20.5 ml HCl
0.1 N).
Sec. 16.41
Results
16.1 Kinetics of releaseof drug from the branchedpolymer alonein synthetic

gastricliquid at 37 C.
Samples of liquid are taken at intervals for analysis and the beads weighed. The
amount of drug released from the beads is determined by using a double-beam W
spectrophotometer(Beckman).
The sameexperiments are also made with the branchedpolymer, by soaking it in
syntheticgastric liquid under the sameconditions (pH 1.2,temperature).
16.4
RESULTS
Two main purposes are considered in this chapter : the one with preparation and
investigation of a branchedpolymer able to deliver the drug ; the other with the study of
the behaviour of dosageforms obtained by dispersing the branchedpolymer into a nondegradablepolymer. Thesedosageforms are of interestbecauseof their wide possibilities:
(i) Various parametersare of help to determine the controlled rate of drug delivery : the
100
200
to a polymer
300
[Ch. 16
LOO
16.2 First-orderkinetics for the releaseof drug from the branchedpolymer in

syntheticgastricliquid.
diameterof the bead,and percentdrug (11).

(ii) Their mechanicalpropertiesareexcellent.
RELEASE OF DRUG FROM THE BRANCHED POLYMER
When the branchedpolymer in powder form is immersedin synthetic gastric liquid,

a liberation of the drug, salicylic acid, is observedwith typical kinetics as shown in Fig.
16.1.This kinetics of drug delivery cannotbe expressedby first or secondorder reactions
since no linear line is obtained in Fig. 16.2 (1st order) and Fig. 16.3 (2nd order).
However, the diffusional aspectof this drug delivery is illustrated in Fig. 16.4 where the
amount of drug liberated is plotted as a function of the square-root of time. A linear
relationship is observedfor short times, when the processis controlled by diffusion. It is
difficult to estimate the value of the diffusivity from this curve becauseof the lack of
Results
Sec. 16.41
I
0
100
200
300
thh
I
399
LOO
16.3 Second-orderkinetics for the releaseof drug from the branchedpolymer in

syntheticgastricliquid.
accuracyin measurementof the averagesize of the powder. Morever, it must be said that
the branched polymer becomes gelatinous during the process because of the liquid
transport.
From theseexperimentssomeconclusionscan be drawn :
(i) The process of matter transfers is not simple for the branched polymer itself when
contactedwith synthetic gastric liquid. Two matter transferstake place : the liquid enters
the polymer provoking an important gelling, reactsupon the sites and dissolvesthe drug
producedby the reaction ; the drug dissolvedin the liquid leavesthe polymer.
(ii) The whole processof drug delivery is controlled by diffusion throughout the grains of
the branchedpolymer, in spite of the reaction which takesplace betweenthe active part of
the branchedpolymer andthe liquid.
RELEASE OFTHE DRUG FROM DOSAGE FORMS WITH EUDFIAGIT
The branched polymer is dispersed in Eudragit RL which plays the role of a
400
[Ch. 16
to a polymer
6-
1-
2-
10
fi
lminl
m
12
16.4 Diffusional transportof the drug. Amount of drug releasedfrom the

branchedpolymer in syntheticgastricliquid as a function of the square-rootof time,
consistent matrix. As shown in previous papers concerned with the use of polymeric
matrices able to be shaped in convenient dosage forms for the drug alone (7, 8, 11 Chapter 10) or for a branchedpolymer (18 - 21), it is of interest to determinethe kinetics
of drug delivery and of liquid transport.
The kinetics of drug delivery are obtained by using various dosageforms of the
same size with different extents of branched polymer, as shown in Fig. 16.5 (30 %
branched polymer), Fig. 16.6 (40 % branched polymer) and Fig. 16.7 (50 % branched
polymer).
In contrast to the branched polymer, the kinetics for the transfer of liquid can be
determinedfrom the weight of dosageforms andthe amount of drug releasedmeasuredat
intervals. The kinetics for the liquid are drawn in Fig. 16.5 - 16.7 with the kinetics of
drug delivery.
Another parameter of interest is the pH of the liquid in which the dosage form is
immersed (9). Experiments are made with a liquid of pH 8, and the results are shown in
Results
Sec. 16.41
20
LO
60
80
16.5 Kinetics of matter transfers(drug : right, liquid : left) with dosageform made
of Eudragit-branchedpolymer 70-30, at pH 1.2 at 37 C.
+ experimental: --------theoretical
Fig. 16.8 (50 % branchedplymer) and in Fig. 16.9 (30 5%branchedpolymer).

It is often of interest to build a model, even a rough and very simple model able to
describe the process,becausesimulations are thus possible. In the caseof dosageforms
with the branched polymer dispersed in Eudragit, the diffusional model with constant
diffusivity expressedby eqn(16.2) is successfully tested either for the transport of the
drug or for that of the liquid, as shown in Fig. 16.5 - 16.9.
The following conclusionscan be drawn from thesecurves:
(i) The liquid transport in the dosageform is described by a diffusional process with a
constantdiffusivity.
(ii) The drug delivery is controlled by diffusion with a constantdiffusivity (Table 16.1).
(iii) The rate of transferaswell asthe amountof matter transferredat equilibrium is higher
for the liquid than for the drug.
402
20
LO
to a polymer
60
[Ch. 16
80
16.6 Kinetics of matter transfers (drug : right, liquid : left) with dosage form made
of Eudragit-branched polymer 60-40, at pH 1.2 at 37 C.
+ experimental : --------theoretical
(iv) In contrast with the branched polymer which exhibits the drawback of turning into gel
when immersed in a liquid, the dosage forms keep good physical properties during the
process.
(v) The rate of release of the drug is lower for the dosage forms than for the branched
polymer, the retardation effect of the matrix being superimposed on the retardation due to
the branched polymer alone.
As shown in Table 16.1, the total amount of drug branched and located in the
dosage forms is not completely delivered (between 40 and 70 %). When the dosage forms
are first immersed in synthetic gastric liquid until equilibrium is reached and then in
synthetic intestinal liquid, the total amount of the drug delivered at the end of this
gastrointestinal history is around 80 % of the initial drug.
Sec. 16.41
Results
Table 16.1 - Characteristicsof the dosageforms

Eudragit/
hquid
PH
Branched
(cm2 / s)lO*
polymer
1.2
1.2
1.2
8
8
70 : 30
60 : 40
50 : 50
50 : 50
70 : 30
20
&quid
D&e
(cm2/ s)lO*
0.67
0.67
0.47
1.0
33
8.3
8.3
8.3
11
17
LO
60
403
M%s
(%)
(76)
122
90
53
102
295
68
69
41
42
38
80
of Eudragit-branchedpolymer 50-50, at pH 1.2 at 37 C.
404
to a polymer
[Ch. 16
$70
20
40
60
t(h)
I
80
of Eudragit-branchedpolymer 50-50, at pH 8 at 37 C.
16.5
CONCLUSIONS
A new way for preparing dosage forms with a controlled release of the drug is
explored.The drug is previously attachedto a reactive monomer and this new molecule is
then polymerized. The branched polymer is dispersedin a polymer matrix, this matrix
bringing consistencyand good mechanicalproperties.
This type of dosageforms may be of interest for the patient s safety. If the dosage
form is crushed by mistake instead of being swallowed whole, another means of
controlled releaseexists involving the branchedpolymer.
Of coursethe drug must react with a monomer, and this branchedpolymer must be
ableto reactwith the liquid, either the syntheticgastricliquid or the intestinal liquid.
In spite of the fact that the processis very complex, a simple model is successfully
Sec. 16.51
Conclusions
405
of Eudragit-branchedpolymer 70-30, at pH 8 at 37 C.
tested for these dosage forms. By considering that the whole process is controlled by
transientdiffusion throughout the materialswith a constantapparentdiffusivity, the wellknown analytical expressionfor homogeneousspheresis able to describe the kinetics of
drug delivery.
REFERENCES
1
J. Feijen. 14th Meeting of French Polym. Group, Rouen, Nov. 1984.
J. Heller. Biodegradablepolymers in controlled drug delivery. CRC Crit. Rev.

TherapeuticDrug Carriers Systems,1 (1984) 39-90.
406
3
to a polymer
[Ch. 16
H. Fessy, J.P. Marty, F. Puisieux, and J.P. Carstensen. Square-root of time

dependence of matrix formulations with low drug content. J. Pharm. Sci. 71
(1982)749-752.
B. Focher, A. Marzetti, V.Sarto, P.L. Baltrame and P. Carmitti. Cellulosic

materials : structure and enzymatic hydrolysis relationships. J. Appl. Polym. Sci.
29(1984)3329-3338.
R. Gumy, E. Doelker and N.A. Peppas. Modelling of sustained release of water

soluble drugs from porous hydrophobic polymers. Biomaterials, 1 (1982) 27-32.
E. Touitou and M. Donbrow. Drug release from non-disintegrating hydrophilic
matrices : sodium salicylate as a model drug. Int. J. Phamr. 11 (1982) 355-364.
I. Malley, J. Bardon, M. RolIet, J.L. Taverdet and J.M. Vergnaud. Modelling of
controlled drug-release in case of carbopol-sodium salicylate matrix in gastric
liquid. Drug. Dev. Ind. Pharm. 13 (1987) 67-79.
8
A. Drain, C. Chaumat, M. Rollet, J.L.Taverdet and J.M. Vergnaud. Model for

matter transfers between sodium salicylate-Eudragit matrix and gastric liquid. Int.
J. Pharm. 27 (1985) 233-243.
A. Eddine, A. Droin, J.L. Taverdet and J.M. Vergnaud. Effect of pH on drug

release between sodium salicylate-Eudragit compound and gastric liquid : modelling
of the process. Int. J. Pharm. 32 (1986) 143-150.
10
C. Brossard, D. Lefort des Ylouses, D. Duchene, F. Puisieux and J.T. Cartensen.

Dissolution of a soluble drug substance from vinyl polymer matrices. J. pharm.
Sci. 72 (1983) 162-169.
11
J.Y. Amrand, F. Magnard, J. Bouzon, M. Rollet, J.L. Taverdet and J.M.

Vergnaud. Modelling of the release of drug in gastric liquid from spheric galenic
forms with Eudragit matrix. Int. J. Pharm. 40 (1987) 33-41.
12
B .Z. Weiner, A. Zilkha, G.Porath and Y. Grunfeld. Atropine attached to
Sec. 16.51
Conclusions
407
polyethylene glycols. Eur. J. Med. Chem. 11 (1976) 525-526.

13
S. Zalipsky, C. Gilon and A. Zilkha. Attachment of drugs to polyethylene

glycols. Eur. Polym. J; 19 (1983) 1177-1183.
14
G. Cramichon, P. Hemery, B. Raynal and S. Raynal. Poly(vinylchloroforate) :

application andpharmacologically active polymer. J. Polym. Sci. Ed. 20 (1982)
3255-3259.
15
C. Pinazzi, J.C. Rabadeuxand A. Pleurdeau.Polymers with potential

pharmaceuticalproperties ; studieson model molecules. J. Polym. Sci. Part C. 15
(1977) 2909-2917.
16
J.C. Meslard, L. Yean, F. Subira and J.P. Vairon. Reversible imrnobilization of

drugs in a hydrogel matrix. Makromol. Chem. 187 (1986) 787-794.
17
J.C. Brosse and J.C. Soutif. Polymers Drugs Carriers of glycerol derivatives.
Polym. Prepr. Am. Chem. Sot. 27 (1986) 7-8.
18
N. Chafi, J.P. Montheard and J.M. Vergnaud. Dosage form with drug attachedto
polymer (polyanhydride)dispersedin a Eudragit matrix : preparationand releaseof
drug in gastric liquid. Int. J. Pharm. 45 (1988) 229-236.
19
N. Chafi, J.P. Montheard and J.M. Vergnaud. Releaseof 2-aminothiazolefrom

polymer carriers. Int. J. Pharm. 67 (1991) 265-274.
20
N. Chafi, J.P. Montheard and J.M. Vergnaud. Dosageform with salicylic acid
attachedto a polyanhydride polymer dispersedin an Eudragit matrix. Int. J.
Pharm. 52 (1989) 203-211.
21
M. Kolli, J.P. Montheard and J.M. Vergnaud. Releaseof benzocaine,procaine, 2

aminothiazole and 4-amino 4 H - 1,2,4 - triazole from polymer carriers. Int. J.
Pharm. 81 (1992) 103-110.
22
J. Crank. in The Mathematicsof diffusion. ClarendonPressed. Oxford. 6
408
to a polymer
[Ch. 16
(1975) 89-92.
23
J.M. Vergnaud. in Liquid transportprocessesin polymeric materials.Modelling

and industrial applications. Prentice Hall ed., Englewood Cliffs, N.J., USA. 3
(1991) 32-44.
Index
absorption of liquid
dosage forms with a core and shell,
and an erodible shell, 329
dosage forms with a core and shell,
and a non-erodible polymer,
345
dosage forms with a drug attached to
a polymer, 393
dosage forms with a swelling
polymer, 363
drug dispersed in an erodible
polymer, 313
Drug-Carbopol sheet in gastric liquid,
225
Drug-Eudragit sheet in gastric liquid,
215
mathematical
treatment
cylinder, 75
parallelepiped, 49
sheet, 21
sphere, 59
numerical analysis
cylinder, 167
parallelepiped,
129
sheet, 105
sphere, 149
cylinder, mathematical treatment
equations of diffusion, 9
solutions with a hollow cylinder
of infinite length, 95
composite hollow cylinder, 101
solutions with a solid cylinder
of finite length, 90
of infinite
length, 77
diffusion of liquid
diffusion equations for various
shapes, 6
dosage forms with a drug and
polymer
core and shell with erodible shell,
329
core and shell with non-erodible
shell, 345
Drug-Carbopol sheet, 227
Drug-Eudragit bead, 242
Drug-Eudragit sheet, 221
effect of pH, 231
erodible polymer, 3 13
with drug attached to a polymer,
393
with swelling polymer, 363
initial and boundary conditions, 4
mathematical treatment for constant
diffusivity
cylinder, 75
parallelepiped, 49
sheet, 21
sphere, 59
methods of solution when the
diffusivity is constant
Laplace transform,
15
reflection and superposition,
17
separation of variables, 1 i
numerical analysis, for concentration
dependent diffusivity
cylinder, 176, 179, 193
410
index
parallelepiped,
135, 142
sheet, 113, 125
sphere, 155, 162
process of diffusion, 2
drying of dosage forms, 261
boundary conditions, 4
controlled pressure of vapour, 300
programmed temperature,
277
surrounding atmosphere of finite
volume, 286
surrounding atmopshere of infinite
volume, 263
mathematical
treatment with constant
diffusivity
cylinder, 75
diffusion equations, 6
methods of solutions, 10
Laplace transform,
15
reflection and superposition,
17
separation of variables, 11
parallelepiped, 49
sheet, 21
sphere, 59
numerical analysis with concentrationdependent diffusivity
cylinder, 176, 179, 193
dosage forms
Drug-Eudragit sphere, 242
dosage forms with core and shell,
345
drying, 261
parallelepiped,
135, 142
sheet, 113, 125
sphere, 155, 162
parallelepiped
mathematical
treatment, with
constant diffusivity
isotropic rectangular
parallelepiped, 49
methods of solution, 10
numerical analysis, with
concentration-dependent
diffusivity,
135, 142
polymer used as matrix
Carbopol, 225
Eudragit, 215, 242, 380, 394
Gelucire, 313, 330, 373, 380

swelling agent, 363
release of drug
dosage forms with core and shell
erodible shell, 345
non-erodible shell, 329
dosage forms with a swelling
polymer, 363
dosage forms with drug attached to a
polymer, 393
dosage forms with erodible polymer
core alone, 3 13
erodible shell, 329
dosage forms with non-erodible
polymer
Drug-Eudragit bead, 242
mathematical treatment
basic consideration,
1
cylinder, 75
methods of solution, IO
parallelepiped, 49
sheet, 21
sphere, 59
numerical analysis
cylinder, 167
parallelepiped,
129
sheet, 105
sphere, 149

Controlled Drug Release of Oral Dosage Forms

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

Controlled Drug Release of Oral Dosage Forms

Enviado por

Direitos autorais:

Formatos disponíveis

CONTROLLED DRUG RELEASE OF ORAL

ELLIS HORWOOD SERIES IN PHARMACEUTICAL

(Available in Cloth and Paoerbaok)

Professor JEAN-MAURICE VERGNAUD

LONDON TORONTO SYDNEY TOKYO SINGAPORE

First published in 1993 by

1 THE DIFFUSION EQUATIONS AND BASIC CONSIDERATIONS ...............

TREATMENT OF DIFFUSION IN A PLANE SHEET .........

Non-steady state with a high coefficient of matter transfer on the

Non-steady state with a solid cylinder of finite length.. ......................

NUMERICAL ANALYSIS WITH ONE-DIMENSIONAL

NUMERICAL ANALYSIS WITH A RECTANGULAR PARALLELEPIPED,

NUMERICAL ANALYSIS WITH A RADIAL TRANSPORT WITHIN A

9 NUMERICAL ANALYSIS WITH CYLINDERS.. ........................................

Conclusions with Drug-Carbopol devices.. ...........................

11 DRYING OF DOSAGE FORMS MADE OF A DRUG DISPERSED IN A

12 DRUG DELIVERY FROM DOSAGE FORMS CONSISTING OF A DRUG

13 DOSAGE FORMS MADE OF A CORE AND SHELL, WITH AN ERODIBLE

14 DOSAGE FORMS MADE OF CORE AND SHELL, WITH A NON-ERODIBLE

15 CONTROLLED RATE OF DELIVERY WHEN THE SOLUBILITY OF THE

16 DOSAGE FORMS WITH A DRUG ATTACHED TO A POLYMER

Therapeutic systemsrepresent a new route for drug administration: as the drug is

and auxiliary substancesbiologically inert, the excipients. The role of excipients is

reasons,and the most obvious are only given :

- In chapter 2, the mathematical treatment of diffusion is shown in various caseswith a

- Chapter 10 concentrateson the drug delivery from simple dosageforms consisting of a

is the gradient of concentration C of the substance along the x-axis of diffusion,

and basic considerations

A polymer in the rubbery state respondsrapidly to changesin its condition. The

when Mt is much lower than the correspondingamountafter infinite time a.

where n is between h and 1

and basic considerations

The initial conditions represent the concentration distribution of diffusing substance

The boundary conditions express the concentration of diffusing substance on the

Two cases are of interest :

the substanceleavesthe dosageform

the substanceentersthe dosageform

1.1.6 VOLUME OF THE SURROUNDING ATMOSPHERE AND

and basic considerations

where K is the partition factor for the drug.

The thin sheet of thickness dx perpendicular to the direction of diffusion is

Fig. 1.1. Diffusion of a substancethrough a thin sheetof thicknessdx

for various shapes

The difference F, - Fx+& is given by :

where the concentrationC is a function of spaceand time.

When the difusivity is constant,this equationsimplifies

1.2.2 EQUATIONS OF DIFFUSION

A rectangularparallelepipedwhose sidesare parallel to the axesof coordinatesand

and basic considerations

Fig. 1.2. Diffusion of a substancethrough a small rectangularparallelepipedof

where Fx representsthe flux of diffusing substancealong the x-axis.

where the diffusivity D may be concentration-dependent.

Anisotropic materials have different diffusion properties in different directions.

for various shapes

showing that F x dependsnot only on the gradient

but also on $ andg.

or more simply when eachprincipal diffusivity is not concentration-dependent.

OF INFINITE AND FINITE LENGTH

and when the diffusivity is constantit becomes:

The diffusion is radial andlongitudinal along the z-axis.

and basic considerations

Of course,in thesetwo cases,the material is isotropic.

1.3 METHODS OF SOLUTION WHEN THE DIFFUSIVITY

When the diffusivity is constant,general solutions of the equation of diffusion can