Professional Documents
Culture Documents
Controlled Drug Release of Oral Dosage Forms
Controlled Drug Release of Oral Dosage Forms
DOSAGE FORMS
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
ELLIS HORWOOD
NEWYORK
Table of contents
PREFACE . . . . . . . . . . . . . . . . . 0. . . . . . . . . . . . m. . . . . . . . . . . . m. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ACKNOWLEDGEMENTS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
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
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
Constant concentration
on the surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2
Finite coefficient of matter transfer on the surface . . . . . . . . . . . . . .
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
101
101
102
105
105
106
107
110
113
113
114
118
118
119
122
122
125
126
126
127
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
Infinite coefficient of matter transfer on the surface . . . . . . . . . . . .
8.2.2
Finite coefficient of matter transfer on the surface . . . . . . . . . . . . . .
Radial diffusion through a sphere with concentration-dependent
diffusivity . . . . . . . . . . . . . . . . . . e. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.3.1
Infinite coefficient of matter transfer on the surface . . . . . . . . . . . .
8.3.2
Finite coefficient of matter transfer on the surface . . . . . . . . . . . . . .
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 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ,...... . . . . . . . . . . . . . . . . . . . . . . . . . . . .
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
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
313
313
314
314
317
318
320
320
322
327
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
345
345
346
352
352
353
354
354
357
358
13.5
363
363
364
364
365
366
373
373
373
376
380
380
382
383
390
393
393
394
395
397
404
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..~...........................................................
409
PREFACE
.
VIII
Preface
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
Preface
xi
xii
Preface
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
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
[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)
fi
Sec. 1.11
Introduction
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
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
1 .1.5 BOUNDARY
CONDITIONS
-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
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
(1.11)
c, < text
(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
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
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
Fx+dx
X
Sec. 1.21
Equations
of diffusion
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)
aF, ac
-ax=Ji
$j
a%
with constantD
FOR A RECTANGULAR
MATERIALS
The diffusion
equations
[Ch. 1
dx
(1.19)
ac aF, aF,
-Ji=~+~+az
a&
$=&ID.
$$+$[D.$)+=$.
$1
ANISOTROPIC
MATERIALS
Sec. 1.21
Equations
(1.22) -F,=D,,.
E+D,*.
of diffusion
$+DIS.$
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 :
1.2.3
CYLINDER
CYLINDER
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-
CYLINDER
OF FINITE
LENGTH
The diffusion
10
equations
ICh. 1
(1.27) ?$;.;(r.D.~j+~(D.$)
When the diffusivity is constant,this equationreducesto :
+- a2c
az2
(1.28)
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
KINDS
OF
IS
SOLUTION
Sec. 1.31
Methods
of solution
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
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
(1.35) $
equations
[Ch. 1
. KS + X2. D=O
dx2
and
(1.37)
C, = A. sin hx + B . cos hx
C,,,= %[A.
n=O
t=O
O<x<L
C = Gin
initial conditions
Methods
Sec. 1.31
of solution
is constant
13
in
(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
and integratingfrom 0 to L it
The diffusion
14
equations
(1.46)
C,.,
[Ch. 1
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
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
is constant
15
TRANSFORM
(1.49)
f(P) = ;
ea
f(t)=exP(-at)
f(p)=
f(p)=---&
uexp-(p+a)t.dt
t= 0
-L<x<L
c=o
(1.51)
t>O
x=fL
c = C-J
surfaces
t 2 0
$0
x= 0
midplane
16
The diffusion
equations
[Ch. 1
= D.
O 2
ac exp ( - pt) dt
O ax2
00
-ac
ox.exp(-pt)dt=[C.exp(-pt)]i+p
oC.exp(-pt)dt=p.C
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
(1.56) $=O
The solution of eqn.(1.53) satisfying the conditions (1.55) and (1.56) is given by
eqn. (1.1)
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
~=~(-l)n.erfc(2n~~-Xj+~(-l)n.erfc/~2n~~+x)
n=O
n=O
METHOD
PLANE
OF REFLECTION
SOURCE
AND
SUPERPOSITION
The diffusion
18
equations
[Ch. 1
(1.60) C = $exp
is a solution of the equationof diffusion in one dimensionwith the constantdiffusivity
+-
(1.61)
M = -co
Upon putting
x2
v =4
2
dx=2fidx
(1.62) M = 2AI%
(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
exp
?$ 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
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&
ci*
C x,t = lr;;
J
m
x+2fiexP (-v)dv
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
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
= 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
IN THE SHEET
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
coefficient
of matter transfer
23
= 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
n=O
necessitatesthat
A,,=0
(2.6)
and h,=12n2+L1z
for
-L<x<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
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
sin%
2
oddvaluesofn
B, = (2n+41,x(-
sin-3n = - 1= sin--(2n+l) z
2
2
1 = &I?!
1)"
dx + . . . . .
Sec. 2.21
Non-steady
coefficient
of matter transfer
25
- (- lIn
-CC4
X+t
= ;. C~.COS
2n+ 1
tin- cca
n=O
2.11
for
(2.13)
ac
I2-F=
I
2(cin-cJ
x=fL
and
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
(2.18) x=
00
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
- 1) n. erfc
+
C(
27
of matter transfer
(2n+ l)L+x
n=O
2l-D-t
(2.21)
$+2.
= 2fE.
ca
I
z(n=l
l)n. ierfcs
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 !
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
00
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
- Constant diffusivity
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
REMARK 5
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
REMARK
The dimensionless number a can also be expressed in terms of the partition factor K,
Non-steady
Sec. 2.21
of matter transfer
31
(2.34)
DIFFERENT
MEDIA
t=O
OCXCL
(2.36)
t>o
x=0
x=L
($0 = f(x)
sheet
surface0
surfaceL
CO*CL.-
(2.371
Cc.,Co.,
=(CL,.,Co,,);
+;. c
n=l
+2.
zsinFexp(-FDt).
L
Jlfcxl,.
sti%dxl
Mathematical
32
treatment
0.1
of diffusion
0.2
in a plane sheet
0.3
[Ch. 2
0.1
0.5
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
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
(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
h 02, - C,,,)
Non-steady
Sec. 2.21
of matter transfer
35
(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
(2.43)
= 1 - B.
x
for x = 0
36
Mathematical
treatment
of diffusion
in a plane sheet
[Ch. 2
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
Sec. 2.31
Non-steady
of matter transfer
37
REMARK
'
Dc.t
(2.47) M,=+
LC.
- y--c
2
7c2
(- l)n
exp
n2
2.3 NON-STEADY
OF MATTER
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
(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
of matter transfer
39
(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
Mathematical
40
treatment
of diffusion
in a plane sheet
(2.55)
c,
cca
(2.56)
c,
< co3
[Ch. 2
Roots of eqn (2.52) are given in Table 2.1, for various values of the dimensionless number M = T,
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
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
c=o
C = Ci~
C = Ci~
sheet
solution
solution
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
(2.65)
cos9nX
L
(2.66)
2 2
= 1 + 2n=l 1 +2(1+a)
CI+ 01qn
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
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
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
(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
(2.68)
L=
l+a
M2L*cin
(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
Sec. 2.51
Steady-state
with a membrane
45
MzMt
= ?.-.IF
.exp -f2n+1)2n2Dt
(2n+ l)*
4L2
as
when
(2.70) qn = n + i
2.5 STEADY-STATE
a+=
WITH A MEMBRANE
with constantD
OF MATTER TRANSFER
= constant
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.
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
(2.76)
C,-
Ci
cOeq-
ci
h.x
= D#+h.
(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,=
(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
(2.83)
Ra=
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
Per = D.S.
REFERENCES
1
Mathematical
treatment
an isotropic rectangular
3.1
of diffusion in
parallelepiped
INTRODUCTION
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.
WITH A
h(C,- Ces)
surface
Sec. 3.21
Isotropic
rectangular
parallelepiped
with a constant
diffusivity
51
c,
= K. c,,
SERIES
(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
Dt]
(2p+l) xx
2c
. exp[-2p:iF2
IX]
Pa
52
Mathematical
treatment
of diffusion
tCh. 3
correspondingwith a one-dimensionaltransport.
O (-l)m
*jzl
(2m+ 1)~
cos(2m+
l)nx
. exp 2a
! 4a2
Dt
,:I]
(3*8)cx.y.z.t
- c.2= f (x,a). f (y,b) . f (z,c)
tin-
Coo
Sec. 3.21
Isotropic
rectangular
parallelepiped
with a constant
diffusivity
53
l)merfc(2m~~-x
m=O
+ C(
-l*. )
erfc
m4
(2m+l)a+x
21Dt
(3.10) 2
1(Dt 1+2w-lm.
C(
a 2/31 mll
I
ma
) .1erfc7s
REMARK
By plotting M,fM, as a function of the square root of time, the slope of the
tangentattheorigini{~+~+~~~,
REMARK 2
54
Mathematical
treatment
0.25
[Ch. 3
of diffusion
0.75
0.5
Q,
(3.13) MzwMt
= y.
exp[-($+-$+$]+I
Isotropic
Sec. 3.21
rectangular
parallelepiped
with a constant
diffusivity
55
(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-
2x.
co4
f(x, a, h) = c
(3.15)
. exp
m=l (p~+x2+x)cosp,
p.tanp=x
x = g
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)
(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
REMARK
0
0
025
05
075
58
Mathematical
treatment
of diffusion
[Ch. 3
G>cw
desorption or evaporation
(3.20)
c,<c,
absorption,or condensation
OF A SHEET
REFERENCES
1
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
or
60
(4.2)
Mathematical
treatment
ac
2 ac
= D. ar2+T.
ar
of radial diffusion
in a sphere
LCh.4
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.
t=O
or
t=O
(4.3)
t>O
r<R
r<R
r=R
U = C.r
C = f(r)
C = Ci~
c = c,
solid sphere
solid sphere
surface
Sec. 4.21
diffusivity
61
(4.5)
a*u
= D. arz
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).
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
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
(4.11)
We-M,
- =Dt
R2
= 4.
rc
21'Dt
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
diffusivity
63
A
Cr,t
a707
-[in
Cal-C
in
75-
Fig. 4.1.
(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
Sec. 4.21
diffusivity
65
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,.
Mathematical
66
treatment
of radial diffusion
LCh.4
in a sphere
0.25
0.5
a75
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
diffusivity
67
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
3 Mce
4n:R3Cin
(4.22)
= -?1+a
The surrounding atmosphere is initially at the uniform concentration C, in, and the
amount of substance transferred after infinite time is K.
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
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
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
Mathematical
70
treatment
of radial diffusion
Kh.4
in a sphere
t=O
(4.29)
t>O
Ri < r < R,
r = Ri
r= Re
c=o
C = Ci = constant
c=o
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
= 6 (Fig. 4.5)
Ri)2
= 4nRi.D.Ci
R,- Ri
Sec. 4.41
0,
a5
I
0.75
, IR,-R$*
71
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
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
c = K,+?
t>o
(4.35)
C = Ci
c = c,
(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
73
Dt
(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, =
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
Mathematical
cylinders
5.1
treatment
of diffusion
in
INTRODUCTION
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.
Sec. 5.21
5.2
Non-steady
NON-STEADY
INFINITE
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
t= 0
t>O
0 <r< R
r= R
C = Ci
c = c, = c,,
78
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
(5.6) C,,- C, = 2
n=l
= 0
(5.8) Gin- C, =
n-l
(5.10)
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
erfc($$5)+(Rm~R~
+[9R2-7r2-2Rr)Dt
ierfc($&)
. [ierfc[$L]J,....
32 Ri. rf
Some roots hs of
Jo (CUR) = 0
Table
5.1
Rootsof Jo(x) = 0
Xl
X2
x3
2.4048
5.5201
8.6537
=4
=s
=6
=7
=8
=9
x10
Mathematical
treatment
of diffusion
in cylinders
al
0.6
[Ch. 5
0.8
(5.15)2 00= g
-2
21 7T
for 2
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)
82
Mathematical
treatment
of diffusion
[Ch. 5
in cylinders
t = 0
0 < r < R
r>R
(5.19) t > 0
C = Ci,
c = text
r= R
-D.g
= h(C,-C,)
(5.20)
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(-$&)
+ ...
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.
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
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
t=O
r<R
r
(5.29) t > 0
Cin.s
R
A.%=
Cin.ext
D.
solid
surrounding
27rR
atR
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
(5.35) A!l+a
~ R2. Cin,s
Mathematical
90
treatment
of diffusion
[Ch. 5
in cylinders
5.3 NON-STEADY
FINITE LENGTH
53.1
CONSTANT
STATE
WITH
CONCENTRATION
A SOLID
ON THE
CYLINDER
OF
SURFACE
(5.36)
(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
(5.39)
cnt - cca
tin-
coa
. -.4 c- (-1)
7T @ 2p+l
J2P+
22lhz
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,
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&,)
94
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
p. tan p = M,
(5.44)
M, = $+r
P. J,(P)
= M, Job)
Mr = R$
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.
Sec. 5.41
Non-steady
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
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
cylinder of infinite
CONSTANT
AND
97
length
EQUAL
ON
t = 0
Ri < r < R,
C = Ci,
cylinder
(5.48)
t>o
r=R
r= R,
c = c,
c = c,
internal surface
externalsurface
(5.49)
(5.52) t
Mt
0
(-a:.
Dt)
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
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.
The initial and boundaryconditionsare :
(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
Re
ci, LnT+Ce
Lng
(5.55)
c,, =
Re
+ IF. Ci,
hrr;
- 5t. c
II=1
Sec. 5.51
of infinite
length
99
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,
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
-&(I.%)
for
= 0
Ri < r < R,
(5.59)
internal surface
r=R;
Ci
r = R,
C, = ct
Ct
external surface
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
cylinder of infinite
length
101
t>O
Ci
r=R
-D.g
r= Re
internal surface
Ct
= h(C,-CW)
externalsurface
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
102
Mathematical
treatment
of diffusion
in cylinders
[Ch. 5
(5.67)
2dCrCz)
Rate =
.. . =
2rcDnIC- Cn+l]
R n+l
hR
Rl
Ln
R*
(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 -$Dt
n=l tex
( nR2) = t-;y/y
(5.72)
exp[-2p:ir*Dt!
2
= {[l-$c]
= 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
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
6.1
a plane
INTRODUCTION
t=j.At
x = n.Ax
time
space
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
(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
Sec. 6.21
Diffusion
through
diffusivity
107
--L
n+O.S
Ax
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 :
108
(6.3)
Numerical
analysis
with one-dimensional
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
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,]
(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
4<M<8
110
Numerical
diffusion
[Ch. 6
CN, = A[C,-,+(M-2)C,+C,+t]
(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
D.
= h. (C,- Cq)
surface
Sec. 6.21
Diffusion
through
diffusivity
111
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)
the new concentration on the surface after elapse of time At, CN,, is thus given by the
relation
(6.15)
112
Numerical
diffusion
[Ch. 6
h. Ax
P = -D-
CN,.,,-
Co.25 = z (CN,-C~)+~~CN,-C,)
CNo = -f
(CN,-Cl]+&
8
[ C 1+ (s-1
-p)co+pc~]
M-2P > 2
3M-8P
> 8
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
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,
(6.23)
h. Ax
D
is not very high.
co;f c,
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
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,+$[-
Jn-o.s+ Jn+o.s]
with
n=O,orn=N
Resulting from the very high value of the coeffkient of matter transferon the surface
h, the concentration of the substanceon the surface is constantly equal to the value at
equilibrium.
(6.7)
co = CN = c,,
The equation (6.25) can thus be used for the whole sheet, and especially for the
abscissaen = 1 and n = N - 1
CONDITIONS OF STABILITY FOR CALCULATION
M > 2
6.3.2 FINITE
SURFACE
WITHINTHE SHEET.
COEFFICIENT
with :
OF
lln<N-1
MATTER
TRANSFER
ON
THE
Diffusion
Sec. 6.31
through
a sheet
115
The problem is the sameas that studiedwith an infinite coefficient of matter transfer
on the surface,and eqn (6.25) can also be used :
(6.25)
CN, = C+g[Ax
Jn-o.s+ Jn+o,j
n = O,orn = N
with
The slice of thickness Ax / 2 located next to the surface, as shown in Fig. 6.3, is
considered.The matter balance during the increment of time At is evaluated within this
slice by taking into account the diffusion of substancethrough the plane 0.5 and the
transferof substancethrough the surface0 with a finite coefficient of matter transferh.
A -(D$&-h(Co-Ccq)
At = q[CNo.ts-Co.25j
I
By making the simple assumption
(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)
Caj]
s = Do.,. y
CNo.25~
Co.25
= 43 (CN,-
Co)++(CN,-
C,)
116
Numerical
analysis
with one-dimensional
(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
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)
c, ce,
c, < G,
evaporation
condensation
Numerical
118
analysis
with one-dimensional
diffusion
ICh. 6
6.4 MEMBRANE
SEPARATING
TWO DIFFERENT
MEDIA
Ci >
Co
CONSTANT DIFFUSIVITY
Sec. 6.41
(6.6)
Membrane
separating
two different
media
119
CN, = &[C,-,+(M-2k,+C,+,I
(6.5)
= constant
co = constant
inlet surface
outlet surface
ci
CONCENTRATION-DEPENDENT
DIFFUSIVITY
CN, = C, + e[-
(6.23
Jn-o.s+Jn+o.s]
= constant
co = constant
ci
inlet surface
outlet surface
Numerical
120
diffusion
[Ch. 6
Gq, i
Gq.0
The problem is studied is section 6.2.2. The concentrationsare thus given without
calculation.
Within the sheet,the concentrationis given by :
(6.6)
(6.5)
(Ad
M = D. At
CNO = b 2C1+(M-2-2Pi)Co+2PiCeq,i
(6.16)
hi. AX
Pi = D
CNo = A
M-2-2Po)Co+2P,,.
Ceq,O
I
Sec. 6.41
Membrane
separating
two different
media
121
(6.16)
ho Ax
P, = D
CONCENTRATION-DEPENDENT
DIFFUSIVITY
This caseis studiedin section6.3.2, and the results are given without calculation.
Within the sheet,the concentrationis given by :
(6.25)
(6.24)
J, = D. $&
i
I
On the inlet surface,the concentrationis expressedby the equation:
(6.38)
CNi = Ci-k(CNi+,-
Ax
Cinl
. . LL
Cl
Cin2
X
*
122
Numerical
(6.39)
CN, = Co- f
diffusion
[Ch. 6
where Ci and Co are the concentrations of substanceon the internal and the external
surface,respectively,
ceq, i md ceq, 0 are the concentrations on the internal and external surface required to
maintain equilibrium with the media i and 0, respectively,
and Ci+, and Co-1 arethe concentrationsat the position hx apart from the internal and the
external surfaces,respectively.
6.5 DIFFUSION
BETWEEN
TWO DIFFERENT
SHEETS
Two sheetsof different materials 1 and 2 are in perfect contact at their junction, and
the problem of diffusion betweenthesetwo sheetsis consideredin this section. (Fig. 6.5).
The diffusivities are D, and D, and the concentrationsare written Ct and CZ.
Each medium is characterizedby two parameters: the diffisivity and the amount of
substancewhich can be absorbedat equilibrium, C,,.
6.5.1
CONSTANT
DIFFUSIVITIES
In the medium 1 with x > 0, the initial concentration is C,$ r, and in the medium 2
with x < 0, the initial concentrationis zero.
The initial and boundaryconditions are :
(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 :
(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
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
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
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
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
CONCENTRATION-DEPENDENT
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
Sec. 6.61
Transfer
127
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
concentration-dependent.
(7.1)
= &[D.g)+=$[~.g]+=j-[D.%)
(7.2)
= D.
is
Numerical
130
parallelepiped
[Ch. 7
-D.g
= h(C,-C,)
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
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)
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
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
concentration
on the surface
133
i,j.k]
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
(7.8)
(Ad2
M, = D. At
MY - (Ay)2
D. At
M, = ()
D. At
ON THE SURFACE
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
0I i I N
N-l
(7.10)
pi
= ~mi,o+~MK,~l+2C
MKi,j+MKi,,
j-2
(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)
With a constant
Sec. 7.21
7.2.2
concentration
on the surface
135
CONCENTRATION-DEPENDENT
DIFFUSIVITY
The concentrationis calculatedwithin the solid and on the surface.
[-(D*E)i-0.5,j,k+~DEli+0.5,j.*l
Ay*AzAt
i, j -0.5.k
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
DIFFUSIVITY
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)
Sec. 7.31
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
concentration-dependent.
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
138
Numerical
[Ch. 7
parallelepiped
2ci,j,k+ci,
j, k+l
(7.8)
bx)
M, = D. At
M - (Ay)2
Y
D. At
M, = (Az2
D. At
C~ . Ay. AZ. At
I
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
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
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
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
(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
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
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~
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
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 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
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)
with
(7.17)
Ci-0.5,j.k
2 (Ci-l,j,k+Ci.j.k)
Numerical
144
[Ch. 7
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]
(7.30)
DO,j,k-0.5
= k DO,j,k-l+%.j,k
1
sec. 7.31
of matter transfer
on the surface
145
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
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
146
Numerical
analysis
with a rectangular
parallelepiped
[Ch. 7
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
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
147
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.
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)
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
within
a sphere
[Ch. 8
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).
When the 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)
Sec. 8.21
Radial diffusion
through
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,,]
Iz.I
152
Numerical
within
a sphere
[Ch. 8
(8.5)
(Ar)(n*+A)[CN,-
. At =
I
c,J
(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
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.
+ D. (A. $)o,i
(8.7)
By replacing the area of the spherical surface Ao.5by its value, this equation leads
to :
CN, = C,,-&-
(8.8)
C,)
where M is a dimensionlessnumber
(Ar I2
M = D. At
(8.9)
with
n = N,orr = R
c, = c, = c,,
AMOUNT OF SUBSTANCE LOCATED IN THE SPHERE
(8.11)
M, =
04rcr2. C,,,. dr
154
Numerical
Mt
(8.12)
analysis
= c-+y
24
4 IT(Ar)3
8.2.2
FINITE
(n2+g
within
a sphere
C,,,+[N3-(N-0.5)3].
[Ch. 8
[&CN+&-l]
n=l
COEFFICIENT
OF
MATTER
TRANSFER
ON
THE
SURFACE
The problem within the sphere and at the middle of the sphereis the same as that
studiedin the previous section (8.2.1), and the equationsare the same.
with
n=N,orr=R
The sphericalmembraneof thicknessAr / 2 located next to the surfaceof the sphere
is considered (Fig. 8.3).
The matterbalancewithin this sphericalmembraneduring the incrementof time At is
evaluatedby consideringthe diffusion through the sphericalsurfaceof radius (R - Ar / 2)
andthe transferthrough the externalsurfacewith the coeffkient of matter transfer.
Ar
4%
h$-Ceq)
D
Ccxt
N -0.5
Sec. 8.31
Radial diffusion
h- AN (CN- Cq) =
3
through
vN-0.25
a sphere
ICNN-0.25-
155
CN-0.2S
(8.14)
CN-0.25
= CN,-
CK
N - 0.5
CNN = CN +
(CN-,-CN]
N3-(N-0,5)3A
N
N3-(N-0.5)3
h. At
-.
Ar
(cN-
CN~-o.zs-
CN-0.25
= i(CN,-
CN]+;(CNN-l-CN-l)
WITH
Two cases are considered, depending on the value of the coefficient of matter
transfer h on the surface.
ceq]
Numerical
156
within
a sphere
[Ch. 8
with a difference due to the fact that the diffusivity dependson position at any time.
Upon putting the function J, as follows (1) :
(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
(8.19)
CN, = C,+
- Jn-0.5+Jn+o.5
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
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
(8.23)
V = 3~
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
(8.24)
3.At
CNN = CN +
JN-0.5
3 h. N2. At
kN-
ceq)
Ar.[N-(N-0.J13]
CNN-O.~~CN-0.25
= 43 (CNN-cN]+ $(CNN-l- cN- 1)
(8.16)
The amount of matter transferred can be calculated by integrating the rate of matter
transferred on the surface with respect to time.
(8.25)
M, = 4 IF N2. (Ar)2. 2
0
h (CN- Cq). At
Sec. 8.41
Constant
HOLLOW
SPHERE
WITH
8.4
CONCENTRATION
ON THE INTERNAL
CONSTANT
DIFFUSIVITY
diffusivity
159
CONSTANT
SURFACE,
AND
This case is of high interest for dosage forms, especially for those where the drug is
surrounded by a spherical membrane.
The spherical membrane of internal and external radii Ri and R, is considered (Fig.
8.4). This spherical membrane of thickness Ro - Ri can be divided into N spherical
membranes of thickness Ar, in such a way that :
NC!
Ro
Fig. 8.4.
(8.26)
In this case, the equations are very simple. However there is a drawback since the
thickness of the membrane Ro - R, as well as the internal radius R must be divisible by
Ar.
A better way for selecting the value of Ar consists of dividing the thickness of the
membrane as follows :
(8.26)
R, = R + N. Ar
160
Numerical
analysis
within
with
r = & + n. Ar
a sphere
[Ch. 8
OSnlN
Two cases are considered, depending on the value of the coefficient of matter
transfer on the external surface.
OF MATTER
TRANSFER
ON THE
The problem is studied in the same way as for the solid sphere, under the same
condition with constant diffusivity.
with
llnlN-1
(8.6)
CN, = C,+
(n-0.5)2(C,-1-C,]-(n+0.S)2(C,-C,+1)
(8.9)
br I2
M = D. At
with
n = 0,orn
= N
r=F$
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,
(8.30)
Ci < C,
Sec. 8.41
Constant
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
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
Ro, with constant concentration Ci and a finite coefficient of matter transfer on the
external surface.
Numerical
162
within
a sphere
[Ch. 8
N - 0.5
CN, = CN +
I
N. h. At
lc~-rcrJ
N3-(N-0.5)3.
[N3-(N-0.5)3]Ar
(c,-cd
CNN-O.Z~-CN-O.~~
= ~(CNN-CN)+~~CNN-~-CN-I)
8.5.1
INFINITE
EXTERNAL
COEFFICIENT
OF
MATTER
TRANSFER
ON
SURFACE
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
(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)
(8.19)
n = 0 , r = Ri
with
15 n I N- 1
[- J, -OS+Jr,+0.5
CN, = C,+
EXTERNAL SURFACE,
n = N , r = R,
CNN
- 0.25
cN - 0.25
CNN
- CN
164
Numerical
within
a sphere
[Ch. 8
(8.24)
3.At
CN, = CN +
CN-0.25 = ~[CNN-CNj+~(CNN-l-CN-l)
CN,-,.,,-
8.6 CONCLUSION
CONDITIONS OF STABILITY
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
1 -&
> 0
or
M>6
BOUNDARY CONDITIONS
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.
CONCENTRATION-DEPENDENT
DlFFUSlVll-Y
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)
REFERENCES
J.M. Vergnaud, in Liquid Transport Processesin Polymeric Materials. Modelling
1
and Industrial Applications, Prentice Hall ed., 1991, Chapter 6.
2
9
Numerical
9.1
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
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
CONSTANT DIFFUSIVITY
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:
(9.2)
[-D./A.~~j-~~~+D.~A.~~j+o,5].At
= Aj. Ar.(CNj-Cj)
I2I
j-O.5
By replacing the volume of the annulus and the areaof the two cylindrical surfaces
by their values,the matter balancebecomes:
(9.3)
5+ij+0.5)(~/j+o~~At
= 2nj.(Ar)2[CNj-CJ
(9.4)
(g)j+o
5 = b;
cj
(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
(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)
11+ &[cj+l-cj-l]
with j = 0
The small cylinder of radius Ar / 2 with the sameaxis asthat of the cylinder of radius
(9.8)
D. (A. ?&,.
Sec. 9.21
Solid cylinder
of infinite
length
171
C,)
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 :
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
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
the new concentration after elapse of time At, CNN, is thus expressed in terms of the
previous concentrations :
(9.14)
(CNTl-CN) -
2N
N-0.25
N - 0.25
@N-
ceq)
(Ad2
M = D. At
(9.6)
(9.13)
h. At
P = Ar
cNN-o.25-
CN-0.25 =
%(CNN-CN)+~(CNN-~-
cNal)
(9.16)
CNN = CN-~(CNN-~-CN-~)+
N-05.
&.&,_,)N - 0.25
N
N -0.25
(c,
3p
The amount of substance located in the cylinder at time t, per unit length, is obtained
Sec. 9.21
173
(9.17)
M, = 2x
r. C,,,. dr
(9.18)
M, = II
. Co+2n(*r)2.
CONDITIONS OF STABILITY
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
CONCENTRATION-DEPENDENT
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)
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
I,., = LD
2 0.9 b
Cl)
cN
= ce, = constant
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)
cN~-o.25
- CN-0.25 = CNN - CN
(9.28)
CNN = CN+
2. At
2 IN-O.S(N- 0.25)(A.r)
N 2 ;25.
-
p. @N- ceq)
(9.29)
8. At
CNN = CN-~(CNN-~-CNJ+
2 IN-O.,3 (N - 0.25)(Ar)
8N
. $NN - 0.25
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 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
Dj-0.5 = D.exp(A.Cj-o.s)
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
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.
(9.33)
[-D.[A.~)j-o.5+D.[A.~)j+o,~.At
= Vj. [CNj-CJ
As the areaof the cylindrical surface$4.5 and the volume of the annulusVj are :
(9.34)
Aj-0.5 = 2x[Ri+(j-O.S)Ar]
(9.35)
Vj = 2x(Ri+j.Ar).
Ar
the new concentrationafter elapseof time At, CNj, at position j is thus expressedby :
178
Numerical
analysis
(9.36)
CNj = Cj+~
with cylinders
[Ch. 9
[Ri+(j-0.5)Ar](Cj-r-Cj)+[Ri+(j+0.5)~r].
I
(Cj+l-Cj)
with r = R
and
j=O
The concentrationon the internal surface is constantly equal to the concentrationin
the solid which is at equilibrium with the liquid, when a liquid is locatedwithin the hollow
cylinder.
(9.37)
CR = C,, i = constant
The annulus of thickness Ar / 2 located next to the external surface of the hollow
cylinder is considered,with a unit length (Fig. 9.5).
Cext
Cext
-0%
- lk--
Ni
Nit05
-D!E
-L
h&y&)
b
r
Ne-OS Ne
Fig. 9.5. A~ulus of thicknessAr / 2 next to the external surfaceof the hollow
cylinder,- with a fmite coefficient of matter transferon this surface.
Hollow
Sec. 9.31
cylinder of infinite
length
179
AN h-(&r
&.)I
= h-,.&,-ox
%-0.25)
By replacing the areas AN _ o.s and AN, as well as the volume V, _ o.25,by their
values,the new concentrationis thus obtained:
(9.40)
CN,-0.25
= CN-0.25+M
4 2R,-Ar
4R -Ar~cN-I-CN)-48~e~r+N-C.q.e)
c-
The new concentrationon the surfaceCN can easily be obtainedeither by making the
simple assumption
(9.12)
C&i - 0.25 -
CN - 0.25 = CNN - CN
9.3.2
CN,-,.2s-C~-o.25
= 34 (CNN-CN)+$(CNN-l-CN-I)
CONCENTRATION-DEPENDENT
DIFFUSIVITY
(9.41)
[-(A.D.~)j-o,~+(A.D.~)j+o.~At
= Vj.[CNj-Cjl
Numerical
180
[Ch. 9
the volume of the ~IIIIU~US Vj. (9.3.11,the new concentration CNj is thus
(9.42)
At. Ij-0.5
CNj = Cj+
At. Ij +0.5
Ar[Ri+j.Ar]-Ar[Ri+j.Ar]
(9.22)
Ri+(j+0.5)Ar
Ij+O.5
D,
J+0.9 (cj-cj+l)
Ar
with
r = l$ and j = 0
CRi
Gq.
= constant
(9.38)
= Ceq.e
= constant
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)
CN,-,.25
= CN-0.25
Ar[4Re-Ar]
4 f ReAr* p* (CNe-
&)
Sec. 9.41
Solid cylinder
of finite length
181
(9.22)
IN-O.5 =
2R,-Ar
2 Ar
DN-0.9
I cN-
l-
CNI
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
CONSTANT DIFFUSIVITY
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. 9.7). 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]
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*
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.
(9.48)
CNi,j = Ci,j+&[Ci,j-l-2ci,j+Ci,j+l
I
-Ci,j-1
(9.49)
bd*
M, = D.At
1
1+& 2[
Ci-l,j-2Ci,j+ci+l,j
184
Numerical
[Ch. 9
(9.50)
(AZ)
M, = D.At
(9.51)
At.
radial
longitudinal
Sec. 9.41
Solid cylinder
of finite length
185
V*[CNi,o-Ci.01
iI
Aif0.5,0
= n 4
Ai.o.5=
2rc . E2 * AZ
I I
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
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
Vi.Nr-0.25 = 2n(Nr-0.25).
= 2
Aif0.J,Nr-0.2S
A.z,Nr-0.5
A,.$.
AZ
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.
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
(9.59)
ANz-0.2S,j
-0.5
ANz-o.5,j
v=
2rr(j-0.5)Ar.T
2rr. j.Ar.Ar.2
(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
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
(9.63)
ANZ-0.25.0.5
= 2n
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
Cd
191
sec. 9.41
With the following valuesof the volume andthe areasof the various surfaces
(9.66)
ANZ-o.D,N~-0.25
ANZ-o.z,~r
ANZ-os,~r-0.25
2rrNr.Ar.g
ANz.Nr-0.25
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-
&]
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
(9.72)
M, = 2 TC
dr* dZ
(9.73)
Mt,i = TZ
2 Nr-1
j. Ci,j
+2x(Nr-0.25)w2
ZC.
+C.
2 . I4 1.Nr 4
l,Nr-
(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.
Sec. 9.41
193
(9.75)
4
2
1 > M,+K
9.4.2
CONCENTRATION-DEPENDENT
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)
(9.78)
0.5,
j - Ci + 0.5, j)
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
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.
(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.
(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.
(9.85)
-.*025
. .
j-05+(D*A'$)Nz-025
j+05'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 [
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)
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
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
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.
200
Drug delivery
[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
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
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
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
began to demand
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
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
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
Overdosaqe
I
)i
[Ch. 10
Overdosage
I-
1
l
Time
hdcrdos-aqe
Fig, 10.2.
10.1.4
ORAL
CONCEPT
THERAPEUTIC
SYSTEMS
Therapeutic systems must release the drug for a certain period of time at a constant
rate, and offers important advantages overconventional
Introduction
Sec. 10.11
and definitions
207
SYSTEM
Drug delivery
208
[Ch. 10
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
(10.1) 2
= ; F
00
i i
(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
[Ch. 10
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
212
Drug delivery
[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).
Sec. 10.11
Introduction
and definitions
213
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-
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
[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
Space
*
Cd,0
cn,o
Cn+l,O
-m___
i.A+
(i+l).At
I
I
I
I
I
I
I
I
I
Time
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,
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
[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
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
t=O
t>o
O<x<L
x= 0
x=
cd = constant
cd=C%q
Cl = Cleq
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
Drug delivery
218
[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,
(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
CN; = -!4
C~-,+(M;-2)C;+C,d+l
(10.10)
(Ax)
M,d = -.
At
1
D;
(10.11)
Dz = exp !
Ad
-Bd
cf+
a.
c;
Drug-Eudragit
Sec. 10.21
219
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
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
Kh.
SAL
O/O
+-+--+
-+-ii2
)Os5
l
tlhl
,.
15
[Sh)
2.
LO
Time(h)
10
Sec. 10.21
Drug-Eudragit
10.2.4 RESULTS
MATRIX
OBTAINED
WITH
EUDRAGIT
AS
221
POLYMER
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
(time)u.5, for various initial concentrations of drug (15-25-50 weight %). In vitro
test with Eudragit RS.
222
Drug delivery
[Ch. 10
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
Drug-Eudragit
Sec. 10.21
223
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
Dd = 16.5 x lo-*
-11
002
I
(cm2/s)
001
1
006
I
+
-L
C
-15,
Ln(D)
'I
224
Drug delivery
[Ch. 10
10.2.5 CONCLUSIONS
WITH EUDRAGIT
AS POLYMER
MATRIX
Fig. 10.9.
,-,,,
o/o LIQ
within the sheet of thickness 0.022 cm and the initial drug concentration of 50 %.
Sec. 10.31
Drug-Carbopol
F F
225
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.
10.3.1
SHEET
IN GASTRIC
LIQUID
INTRODUCTION
Drug delivery
226
[Ch. 10
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
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
4 SAL
O/O
m-s15
300 cso-s-50
C75-S25
(Time)'.'
1
20
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
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
The validity of the numerical model as well as the accuracy of the data are tested by
230
Drug delivery
[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.
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
concentration-dependent.
connected with each other, the diffusivity of the drug increasing with the concentration of
liquid in the dosage form.
Sec. 10.41
231
600
ces-s15
Time(h)
*
40
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
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
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
10.4.3
EXPERIMENTAL
Drug delivery
234
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 %).
RESULTS
WITH
DRUG-EUDRAGIT
SHEETS
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
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
10
15
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
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.
Drug delivery
238
Kh.
~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
= 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
Drug delivery
240
[Ch. 10
I
X
X
3836\
31_
32_
30,
1231567
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
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
The following assumptions are made on the sample and on the process :
some
Spherical
Sec. 10.51
Drug-Eudragit
243
The equation of radial transient diffusion through the spherical bead is given by :
(10.19)
= $.
ac
& D. r . a
t = 0
r < R
C = C$
(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
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:
-n-m+
n+OJ
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
Jcr5 = ;.
Do,,. (Co-C,)
n=N
liquid
CL = C;
d
cN=
c,
d
dJ-%
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
EXPERIMENTAL
MATERIALS
OF THE PARAMETERS
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
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
[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
Spherical
Sec. 10.51
Drug-Eudragit
249
4P
Licl A
O/O
O/O
30,
Time(h)
LO
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
[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
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
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
252
Drug delivery
[Ch. 10
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
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.
254
Drug delivery
[Ch. 10
Sec. 10.51
Spherical
Drug-Eudragit
255
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
256
Drug delivery
[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
A.F. Stockwell, S.S. Davis and S.E. Walker. In vitro evaluation of alginate gel
delivrance des principes actifs mtdicamenteux. IV. Systemes a gonflement control&. STP
Pharma. 2, (1986), 38 - 46.
29
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
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
delivrance
des principes
actifs medicamentaux.
Modelisation
des mecanismes
plasticized PVC and of plasticizer from polymer into liquid. J. Appl. Polym. Sci., 26,
Sec. 10.51
Spherical
Drug-Eudragit
257
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
plasticized PVC and liquids in case of a maximum for liquid-time curves. J. Appl.
Polym. Sci., 31, (1986), 111 - 122.
39
plasticized PVC previously immersed in n-hexane. Europ. Polym. J., 25, (1989),
1013 - 1018.
40
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
and
258
Drug delivery
[Ch. 10
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
Y. Khatir,
desorption of liquid by a rubber annulus using a model and short-term tests. Plast.
Rubber. Process. Applic. 9, (1988), 53 - 58.
48
Model of matter transfers between sodium salicylate-Eudragit matrix and gastric liquid.
Int. J. Pharmac., 27, (1985), 233 - 243.
49
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
of matter
oral dosage forms by whole body liquid scintillometry. J. Pharm. Sci., 55, (1966), 678.
53
Chapt. 4.
Sec. 10.51
55
Spherical
Drug-Eudragit
259
56
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
plasticized PVC and liquids in case of a maximum for liquid-times curves J. Appl.
Polym. Sci., 31 (1986), 111-122.
59
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
[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
THEORETICAL
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
(11.2)
-D.(g),
= h(CR- Cd
(11.3)
cr.,-ceq
tin-
eq
(11.4)
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
Drying in a surrounding
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
1 . (n- 0.512.(C,_,-C,J-(n+0.5)2.
+12 i
1
and
(11.9)
r = n.Ar
R= N. Ar
[-n-0.5+
n+m I
(Cn-Cn+r)
266
[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,)
N - 0.5
l. (cN-+N)-
N3-(N-0.5)3M
N
N3-(N-0.5)3
h. At
.-(CN-C
Ar
3. At
(Ar)*. [N3- (N - o,5)3] JN-0.5- Ar. I:-
. .
(: fL.5)3] (-
cq1
Sec. 11.21
Drying in a surrounding
atmosphere
of infinite
volume
267
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
(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)
h(CN- Cq). A t
11.2.2
EXPERIMENTAL
OF DOSAGE
MATRIX.
FORMS
acrylate and
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
Drying in a surrounding
atmosphere
of infinite
volume
269
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.
The validity of the model is tested by comparing the kinetics of drying of various
Sec. 11.21
Drying in a surrounding
atmosphere
of infinite
volume
271
D (cm* /s)
4x 10-s
2x 104
2.5 x 10-4
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
Drying in a surrounding
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
DlFFUSlVllY
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
(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
evaporation of a liquid depends also on the rate of stirring of the surrounding atmosphere.
Sec. 11.21
Drying in a surrounding
atmosphere
of infinite
volume
275
;x103tk-'I
-6
-15,
-16,
+LnDT
Ln k,
OF THE PARAMETERS
[Ch. 11
276
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
temperature
277
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
11.3 DRYING
WITH A PROGRAMMED
TEMPERATURE
(24)
278
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
of
temperature :
(i) The diffusivity
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,
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 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
,
80
TimeIh 1
I
100
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
Sec. 11.31
temperature
285
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
THEORETICAL
OF
11.4.1
ATMOSPHERE OF FINITE VOLUME
DRYING
IN
SURROUNDING
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.
MATHEMATICAL TREATMENT
As shown in Chapter 4, the general equation of diffusion within the spherical bead
is :
(11.28)
= D.
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
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
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
Drying a surrounding
atmosphere
of finite volume
289
(11.8)
CN, = C,+
[(n-0.5j2.
(C,-,-C,)-(n+0.512.
(C,-C,,,il
(11.13)
(Ad2
M = D. At
(11.12)
CN, = Grub
Cl)
(11.16)
CNN = CN+
N - 0.5
3 A. bw-
c,)-
N3w(,_,.5)3.
y%-
N3- (N - 0.5)
by making the simple assumption :
(11.17)
- CN
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)
(11.38)
C,
= k$&
where k, k and b are constant depending on the vapour, the nature of the polymer and
temperature.
The amount of liquid evaporated at tune t is given by :
(11.40)
= nt. MW
C, = &$.
M, = K,. M,
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
Drying a surrounding
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
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
RESULTS
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
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
Drying a surrounding
Sec. 11.41
atmosphere
of finite volume
293
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
Drying a surrounding
atmosphere
of finite volume
295
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
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
Drying a surrounding
atmosphere
of finite volume
297
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
Drying a surrounding
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.
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
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
vapour pressure
301
11.5.1
OF THE
PROCESS
WITH
CONTROLLED
VAPOUR
PRESSURE
ASSUMPTIONS
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)
(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
O<r<R
r=O
r=
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
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
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
- R=0.38cm
- h=19.8xlO-scm/s.
rl
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
(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
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
REFERENCES
1
Sec. 11.61
Conclusions
309
dependence of matrix formulations with low drug content. J. Pharrn. Sci. 71, 1982,
749-752.
5
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
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.,
dosage forms made of drug dispersed in a polymer. Int. J. Pharm., 67 (1991) 51-57.
23
the drying process of dosage forms prepared by a humidity technique. Int. J. Pharm., 67
(1991)163-168.
24
N. Laghoueg-Derriche
surrounding atmosphere on the drying process of dosage forms made of polymer and
drug. J. Polym. Engng., 10 (1991) 313-331.
26
M. Malizewicz
Modelling of the drying process of coatings with various layers. J. Coat. Technol., 59
(1987)27
28
- 32.
PVC
12
Drug delivery from dosage forms
consisting of a drug dispersed in an
erodible polymer
12.1
INTRODUCTION
Drug delivery
314
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.
MATHEMATICAL TREATMENT
(12.1)
= D.
Sec. 12.21
Theoretical
aspects
315
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
(12.4)
,zMt
n2. 7r2
----D.t
R2
= 4.
II
where n is an integer.
FINITE VOLUME OF LIQUID
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&
Drug delivery
316
[Ch.
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
(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
(12.10)
= ;[;].5
Sec. 12.21
Theoretical
aspects
317
12.2.2 POLYMER
ASSUMPTIONS
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 :
voumeatt)r l = $t
As the volume of the bead is proportional to its weight, this equation is transformed
as follows :
Drug delivery
318
[Ch. 12
Mofbeadatt
I Mofbeadatt=O
(12.14)
3 = 1-k
I
In the present study, measurements are made in the liquid. The above equation is
thus expressed by :
(12.15)
(1-g$
(1
Mdissolved at t F= 1 k t
-Mdissolved at = =
(12.16)
12.3
EXPERIMENTS
MATERIALS
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
Radius (cm)
Volume
of
liquid (ml)
1
264
0.36
200
187
0.33
200
92
0.268
100
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
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
[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
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.
< 0.9.
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
[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
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
324
Drug delivery
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
Fig. 12.5 Kinetics of release of drug with the erosion process. Weight of bead :
95 mg - volume of liquid : 100 ml.
:+
Theoretical : -Experiments
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
[Ch. 12
Fig. 12.6 Kinetics of release of drug with the erosion process. Weight of bead :
266 mg - volume of liquid : 100 ml.
:+
Theoretical : -Experiments
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
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
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
[Ch. 12
Gattefosst Gelucire puts liquid formulations into hard gelatin capsules Private
paper (1983).
5
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
330
[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
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
= 1.;
r
D.r*.$
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
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.
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.
Diameter = 0.85 cm. Thickness of Gelucire = 0.16 mm. 410 mg.
10
20
30
LO
Fig. 13.5 . Kinetics of releaseof sodium salicylate from the core and shell device.
Diameter = 0.85 cm. Thickness of Gelucire = 0.17 mm. 414 mg.
13
Sec. 13.41
Results
337
Fig. 13.6 . Kinetics of releaseof sodium salicylate from the core and shell device.
Diameter = 0.85 cm. Thickness of Gelucire = 0.20 mm. 409 mg.
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
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.
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.
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
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.
matter transfers between sodium salicylate-Eudragit matrix and gastric liquid. Int. J.
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
346
Druq delivered
Time
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 sym- metrical 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 equation of radial diffusion under transient conditions with a concentrationdependentdifusivity is written as follows :
(14.1) a;-=ysi3T
ac
l
a
2 ac
Der.,
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
(14.4) 2
= ; (Y)
with
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)
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
previous concentrations
(14.7) CN, = C, +
At
[-J
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 :
(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 :
(14.10)
CN, = C, +
At
*Il (A %I2
. [-
JnmOe5
+ Jn+0,5]
349
[Ch. 14
350
- 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
(1 - $-i,.
[l + $.
8. At. Ar2
3B.R;
[CN&wl-
C&ml]
[CN;,+t-C;,+tj
JIN,+1
-
4.At.Ar2
3B.R;
,
JN
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,
Sec. 14.21
Theoretical
351
(14.15)
CN;, =
4At.
3 (A rt + A r2) R:
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
(14.18)
K = M - M
M:
352
[Ch. 14
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
polymer
is
dimethylaminoethylacrylate
Eudragit
RS
(RGhm
Pharma),
a copolymer
of
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
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
14.4.1 DATA
Experiments are carried out under the operational conditions described in the
experimental part with
various
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
55
1.4 x lo-
29
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.
Thickness of the shell = 0.003 cm.
10
20
28 0
10
20
Fig. 14.5. Kinetics of transfer of the drug (left) and liquid (right).Rl = 0.424cm.
Thickness of the shell = 0.005 cm.
356
[Ch. 14
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.
Thickness of the shell = 0.0055 cm.
Fig. 14.7. Kinetics of transfer of the drug (left) and liquid (right).Rl = 0.441cm.
Thickness of the shell = 0.0105 cm.
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.
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
360
3
[Ch. 14
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
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
of a liquid by a polymer device made of a core and shell. Europ. Polym. J. 25 (1989)
89-94.
10
an EVA polymer device composed of a core and shell. Plast. Rubber. Proces. Applic. 11
(1989) 9-16.
11
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
plasticized PVC and of plasticizer from polymer into liquid. J. Appl. Polym. Sci. 26
(1981) 2315-2324.
Sec. 14.51
Conclusions
361
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.
15.2.1
WITH
(4)
A POLYMER
MATRIX
AND
THEORETICAL
ASSUMPTIONS
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
matrix
365
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
low number of quatemary ammonium terminal groups.The ratio between the numbers of
quaternary and ester terminal is around 1 : 20.
DOSAGE 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
RESULTS
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
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)
Sec. 15.21
371
matrix
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
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
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.
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 %.
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)
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
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
IN
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
15.4 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
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.
MATHEMATICAL TREATMENT
(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
r=
c=o
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
(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
DOSAGE FORMS
Sec. 15.41
Polymer-Drug-Eudragit)
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
RESULTS
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
Polymer-Drug-Eudragit)
385
6-
2-
Rl.R2
R,.Rl (cm)
&-RI
D
6
R2-Rl
delivery obtained from various dosage forms II is plotted against R, . R2 / (R2 - R,),
exhibiting a straight line within
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
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
Polymer-Drug-Eudragit)
387
Table 15.3Characteristics
Gastric liquid
Intestine liquid
5.6 x lo-*
5.7 x 10-s
0.9
4.1
30-
20-
10,
Rl
* R2
R2-Rl
liquid at 37C.
JML
W4
kl-il)
I
388
Controlled
[Ch. 15
rate of delivery
It
TO/O
75L
3
2
50-
b
0
12
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.
Polymer-Drug-Eudragit)
389
(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
15.5
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
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
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
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]
(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
16.3
EXPERIMENTAL
396
to a polymer
tCh. 16
in a mixture of
chloroform and hexane. After drying, the amount of anhydride is 5.3Og, with a yield of
52%.
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).
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
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
Results
Sec. 16.41
I
0
100
200
300
thh
I
399
LOO
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
400
[Ch. 16
to a polymer
6-
1-
2-
10
fi
lminl
m
12
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
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
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
16.7 Kinetics of matter transfers(drug : right, liquid : left) with dosageform made
of Eudragit-branchedpolymer 50-50, at pH 1.2 at 37 C.
+ experimental: --------theoretical
404
to a polymer
[Ch. 16
$70
20
40
60
t(h)
I
80
16.8 Kinetics of matter transfers(drug : right, liquid : left) with dosageform made
of Eudragit-branchedpolymer 50-50, at pH 8 at 37 C.
+ experimental: --------theoretical
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
16.9 Kinetics of matter transfers(drug : right, liquid : left) with dosageform made
of Eudragit-branchedpolymer 70-30, at pH 8 at 37 C.
+ experimental: --------theoretical
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
406
3
to a polymer
[Ch. 16
10
11
12
Sec. 16.51
Conclusions
407
14
15
16
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
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
22
408
to a polymer
[Ch. 16
(1975) 89-92.
23
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-Carbopol sheet, 225
Drug-Eudragit sheet, 215
Drug-Eudragit sphere, 242
dosage forms with core and shell,
345
drying, 261
parallelepiped,
135, 142
sheet, 113, 125
sphere, 155, 162
parallelepiped
equations of diffusion, 7
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