Professional Documents
Culture Documents
www.elsevier.com/locate/jconrel
Institute of Bioengineering and Nanotechnology, 31 Biopolis Way, The Nanos, #04-01, Singapore 138669, Singapore
b
DSO National Laboratory, 27 Medical Drive, #12-00, Singapore 117510, Singapore
Received 1 June 2004; accepted 1 November 2004
Available online 8 December 2004
Abstract
A mathematical model is developed to describe the transport phenomena of a water-soluble small molecular drug
(caffeine) from highly swellable and dissoluble polyethylene oxide (PEO) cylindrical tablets. Several important aspects in
drug release kinetics were taken into account simultaneously in this theoretical model: swelling of the hydrophilic matrix and
water penetration, three-dimensional and concentration-dependent diffusion of drug and water, and polymer dissolution. The
moving boundary conditions are explicitly derived, and the resulting coupled partial differential equations are solved
numerically. In vitro study of swelling, dissolution behavior of PEOs with different molecular weights and drug release are
also carried out. When compared with experimental results, this theoretical model agrees with the water uptake, dimensional
change and polymer dissolution profiles very well for pure PEO tablets with two different molecular weights. Drug release
profiles using this model are predicted with a very good agreement with experimental data at different initial loadings. The
overall drug release process is found to be highly dependent on the matrix swelling, drug and water diffusion, polymer
dissolution and initial dimensions of the tablets. Their influences on drug release kinetics from PEO with two different
molecular weights are also investigated.
D 2004 Elsevier B.V. All rights reserved.
Keywords: Swelling; Diffusion; Dissolution; Controlled release; Polyethylene oxide (PEO) matrix; Caffeine
1. Introduction
* Corresponding author. Tel.: +65 68247106; fax: +65
64789084.
E-mail address: yyyang@ibn.a-star.edu.sg (Y.-Y. Yang).
1
Current address: The Department of Chemical Engineering,
Princeton University, USA.
0168-3659/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jconrel.2004.11.002
570
2. Mathematical model
A schematic for the model considered here is
plotted in Fig. 1. The matrix is cylindrical with initial
radius of r 0 and initial height of 2z 0. Only the upper
half part of the cylindrical matrix is considered in
mathematical modeling since it is symmetrical to the
571
z
z
n
z0
r0
t=0
t>0
r Br
Bz
Bt
Br
Bz
1
C 1 C1 =C1;e
r r=r0
C 2 C2 =C2;0
z z=r0
D1 D1 =D1;e
s D1;e t=r02
D2 D2 =D1;e
3
so that we can write Eq. (1) in dimensionless form
B Ci
1 B
Bs
r Br
r Di
B Ci
Br
B
Bz
Di
B Ci
Bz
572
0V r VR
0V z VZ
C 1 0:0
C 2 1:0
dn
dm
C1;e
dn C1;e
dm
2pmn
pm2
2pmn
dt
dt
dt
dt
q1
q
! 1
Z nZ m
1 BC1
1 BC2
2prdrdz
q
q
Bt
0
0
1
2 Bt
pm2
0V r VR
0V z VZ
z Z
r R
C 1 1:0 C 2 0:0
C 1 1:0 C 2 0:0
0V r VR
z 0
0 V z VZ
r 0
sN0
B C1
0:0
Bz
B C1
0:0
Br
B C2
0:0
Bz
B C2
0:0
Br
7
qp
0
where q i is the density of water (i=1), drug (i=2) or
polymer (i=p).
Note that the third term on the right-hand side of
Eq. (8) represents the polymer dissolution process,
which is rather complex and involves polymer chain
disentanglement and reptation [30]. To simplify the
model analysis, we use a single parameter, dissolution
rate constant K p, to characterize the surface (heterogeneous) dissolution of polymer. It is a mass rate
constant and defined as (1/A s)(dm p/dt)=K p. A s is the
total surface area of polymer matrix in contact with
the release medium and m p,0 is the initial mass of
polymer. K p can also be viewed as a product of
dissolution/mass transfer coefficient and the equilibrium polymer concentration at the dissolution
(swelling) front [18,31].
1
Kp pm2 2pmn
qp
C1;e
C1;e
dm
2 dn
1
1
pm
2pmn
dt
dt
q1
q
Z n
1
BC1
1
BC2
2pm
D1
D2
dz
q1
Br
rm q2
Br
rm
0
Z m
1
BC1
1
BC2
D1
D2
rdr
2p
q
q
Bz
Bz
zn
zn
1
Kp pm2 2pmn
qp
10
BC1
D1
dz
Br
rm
0
Z n
Kp
1
BC2
D2
dz n
11
q
Br
dm
1
n 1 f 1
dt
q1
BC1
D1
2rdr
Bz
zn
0
Z m
Kp
1
BC2
D2
2rdr m2
q
Bz
q
dn
1
m 1 f 1
dt
q1
2
rm
zn
12
where f 1 is the equilibrium volume fraction of water in
the full-swollen matrix.
(i)
sN0
0VnV1
C 1 0:0
C 2 1:0
0VnV1
0VgV1
B Ci
Bs
!
r ;z
B Ci
Bs
!
n;g
1 B C i n BR2
2 Bn R2 Bs
1 B C i g BZ 2
13
2 Bg Z 2 Bs
D
i
n Bn
Bg
Bs
Bn R2
Bg Z 2
1 B C i n BR2 1 B C i g BZ 2
2 Bn R2 Bs
2 Bg Z 2 Bs
14
0VgV1
C 1 1:0
C 1 1:0
g1
n1
15
C 2 0:0
C 2 0:0
16
1VnV1
g0
0VgV1
n0
sN0
sN0
B C1
0:0
Bg
B C1
0:0
Bn
B C2
0:0
Bg
B C2
0:0
Bn
17
n1
0VgV1
Z 1
1
dR2
B C 1
f1
1 f1
D1
dg
2
ds
Bn
n1
0
Z 1
Kp
B C 2
D2
dg R
f2;0
Bn n1
qp
0
18
sN0
g1
0VnV1
B C 1
2ndn
Bg
g1
0
Z 1
Kp
B C 2
f2;0
D2
2ndn Z
Bg g1
qp
0
1
dZ 2
f1
1 f1
2
dt
B Ci
1 B Ci
R Bn
Br
B Ci
1 B Ci
Z Bg
Bz
573
D1
19
20
(a)
100
water uptake (%)
80
60
40
Model
20
0
Experiment
0
12
18
time (hr)
24
30
36
20
24
120
(b)
100
water uptake (%)
574
80
60
40
Model
20
Experiment
0
12
time (hr)
16
3.5
(a)
n / z0
2.5
m / r0
2
1.5
15
20
2.2
Model
100
10
time (hr)
(b)
(a)
Experiment
90
80
70
n / z0
1.8
m / r0
1.6
1.4
1.2
60
1
0
10
20
time (hr)
30
40
(b)
Model
100
10
15
time (hr)
Fig. 4. Relative dimensional change of pure PEO tablets during
swelling and dissolution. Squares denote the experimental measurements. (a) PEO with M w=8106. (b) PEO with M w=4106.
120
relative dry weight of PEO (%)
2.4
110
50
575
Experiment
80
60
40
20
0
12
time (hr)
18
24
576
80
60
40
20
0
100
relative drug release (%)
100
(a)
12
time (hr)
18
40
20
100
80
60
40
20
0
60
24
(c)
12
time (hr)
18
24
(b)
80
0
0
100
12
time (hr)
18
24
(d)
80
60
40
20
0
12
time (hr)
18
24
Fig. 5. Fit of the model to the experimentally determined relative amount of caffeine released at different loadings from PEO of M w=8106.
Squares denote experimental data. (a) 8.33% w/w (R 2=0.998); (b) 16.67% w/w (R 2=0.997); (c) 33.33% w/w (R 2=0.998); (d) 50% w/w
(R 2=0.995).
(a)
80
60
40
20
0
10
time (hr)
15
60
40
20
100
(c)
80
60
40
20
10
time (hr)
15
20
10
time (hr)
10
time (hr)
15
20
(d)
80
60
40
20
0
0
(b)
80
20
100
100
100
577
15
20
Fig. 6. Fit of the model to the experimentally determined relative amount of caffeine released at different loadings from PEO of M w=4106.
Squares denote experimental data. (a) 8.33% w/w (R 2=0.994); (b) 16.67% w/w (R 2=0.997); (c) 33.33% w/w (R 2=0.991); (d) 50% w/w
(R 2=0.969).
(a)
1
0.9
0.8
0.7
8.33%w/w
16.67%w/w
33.33%w/w
0.6
0
50.00%w/w
6
12
time (hr)
18
24
(b)
8.33%w/w
relative dry weight of PEO
578
0.8
16.67%w/w
33.33%w/w
50.00%w/w
0.6
0.4
0.2
0
0
10
time (hr)
15
20
579
(a)
relative drug release
0.8
0.6
0.4
2z0 / r0 = 0.5
2z0 / r0 = 1.0
0.2
2z0 / r0 = 2.0
0
0
1
10
time (hr)
15
20
(b)
0.8
relative drug release
0.6
0.4
r0 / 2z0 = 0.5
r0 / 2z0 = 1.0
0.2
r0 / 2z0 = 2.0
0
0
10
time (hr)
15
20
580
Acknowledgements
This research was funded by Agency for Science,
Technology and Research, as well as Defence Science
and Technology Agency, Singapore.
References
5. Conclusions
In this work, a mathematical model is developed
and in vitro study is carried out for the controlled
release of caffeine from polyethylene oxide (PEO)
cylindrical tablets with two different molecular
581