13
Diffusion and Dissolution
Stea -State Diffusion Diffusion Principles in Biologic Systems
Procedures and Apparatus Thermodynamics of D:ffusion
Dss:!ution Fick's Second Law
Drug Release Diffusion and Ecology
Free diffusion or passive transport of substances
throu g h liquids. solids, and membranes is a process of
considerable importance in the pharmaceutical sci-
ences. Topics of mass transport phenomena applying to
pharmacy are dissolution of drugs from tablets, pow-
a tion, and
der s, anii granules: lvophilic.atton. ultrafiltr
other mechanical processes: release from ointments and
supoositry oases: passage of water vapor, gases.
drags. and dosage form additives through coatings,
packazinn. films, plastic container walls, seals. and
caps: and permeation and distribution of drug molecules
in living tissues.
Diffusion. Di1fuso,i is defined as a process of mass
iransiir of individual molecules of a substance, brought
about fl y random molecular motion and associated with
P concentration gradient. Flow of molecules through a
barrier so ch ms a polymeric membrane is a particularly
convenient \va',' to study diffusion processes. The
uf matter through a barrier (Fig. 1$- I mdv
occur b y simple molecular pertac- at on or by movement
through pores and channels. Molecular diffusion or
permeation through nonporous media depends on dis-
. lu'ion of the permeating molecules in the bull.:
ric'nbrane (Fig, 13- In), whereas a second process may
involve passage of a substance through solvent-filled
pores of a membrane (Fig. 13- 10) and is influenced by
the relative sizes of the penetrating molecules and the
diameter of the pores. The transport of a drug through
a polymeric membrane involves dissolution of the drug
in the matrix of the membrane and is an example of
simple molecular diffusion. Passage of steroidal mole-
cules, suhtituted with hydrophilic groups, through
human s:in may predominantly involve transport
through h.mir follicles, sebum ducts, and sweat pores in
the epidermis (see Fig. 13-22). Perhaps a better
representation of a membrane on the molecular scale is
a matted arrangement of polymer strands with branch-
32-1
ing and intersecting channels as shown in Figure
13- Ic. Depending on the size and sha p e of the diffusing
molecules, they may pass through the tortuous pores
formed by the overlapping strands of pol y mer. If too
large for such channel transport. the diffusant may
dissolve in the polymer matrix and pass through the
film by simple diffusion.
Dialysis. Hwang and Kammc'rmeyer 1 define diotmsts
as a separation process based on unequal rates of
- passage of solutes and solvent through nncroporous
membranes, carried out in batch or continuous mode.
Jjcoiodialusis is used in kidney malfunction to rid the
blood of metabolic waste products (small molecules)
while preserving the high-molecular-weight compo-
nents of the blood.
Osmosis. A process related to dialysis, OSfliOSi.S was
originally dc-fined as the passage of bothsolute and
solvent across a membrane, but now refers to an action
ii which On:V t nc solvent is transferred . The so] vent
passes through the semipermeable membrane to cltlutc'
the solution containing solute and solvent (sec p. 116).
The passage of solute to; ether with solvent now is
called diffusion or dial ysi.'
Ultrafiltration. Ultrafiltration is used to separate col
loidal particles and macromolecules by the use of a
membrane. Hydraulic pressure is employed to force the
solvent through the membrane while the microporous
membrane prevents the passage of large soiute mole-
cules. Uitrafl]tration is similar to a process called
reverse osmosis, but a much higher osmotic pressure is
developed in reverse osmosis, which is used in desali-
nation of brackish water. Ultrafiltration is used in the
pulp and paper industry and in research to purify
albumin and enzymes. Microfiltnthosi. a process that
employs membranes of slightly larger pore size, 100
nanometers to several micrometers, removes bacteria
from intravenous injections, foods, and drinking Wa-

Diffusion
through the
homogeneous
film.
. bitfunion
Fig. 13-1. mi Homogeneous membrane without pores. (C) alv'mbrane afdnse inaienal with ctraight-throues pores. as found in certain iUter
b,u'riers such as Yucleoriore (c) Cellulose membrane used in filtration processes. snowing inEe rtwirung nature of fibers and tortuous channels.
ter. 2 In ordinar y osmosis as well as in dial y sis, separa-
tion is spontaneous and does not involve the high
applied pressures or ultrafiltration and reverse osmosis.
Elvnn ci al.' differentiate between a membrane and a
barrier. A membrane is a film separating the phases,
and material passes b y passive, active, or facilitated
transoort across this film. The term barrier applies
in a more general sense to the region or regions
that offer resistance to passage of a diffusing mate-
rial, the total barrier being the sum of individual
resistances of membranes or the component sims of
laminae inlet-posed between a donor and a receptor
chamber.
STEADY-STATE DIFFUSION
Fick's First law. The amount .11 of material flowing
through a unit cross-section, S. ofa ban 'ierin urtittime,
is known as the flux. J.
d.11
(13-1)
= S•dt
The flux in turn is proportional to the concentration
gradient, dCldx:
dC
J = (13-2)
di
in which D is the diffusion coe fficie'nc of a penetrant
(also called the dzfJ'usartt) in cm 2isec, C its concentra-
tion in g,,* CM3 , and x the distance in cm of movement
perpendicular to the surface of the barrier. In equation
Chapeer!S-Diffusion and Dissolution 325
0
0 Diffusion
0 0 through uofvent (usually
witer).fiii5
000 Pores.
0 0 0 0
0
0 0 00
O 0
0 00
00 0 C. o
0 0 0 (h)
through and/o
between the
iibraus membrane
strands.
(13-1), the mass, M. is usuaiiv given in grams or moles.
the barrier surface, S. in crn, and the time, t, in
seconds. The units on I are g cni S ee - '. The SE units
of kilogram and meter are sometimes used, and the
time ma y be given in minutes, hours, or da y s. The
negative sign of e q uation (13-2) signifies that diffusion
occurs in a direction (the positive z direction) opposite
to that of increasing concentration. That is to say,
diffusion occurs in the direction of decreasing concen-
tration of diffusant; thus, the flux is alwa ys a positive
quantity.
The diffusion constant. D, or dtjusii'dy as it is often
called, does not ordinaril y remain constant, for it may
change in value at higher concentlations. D is also
affected by temperature, pressure, solvent properties,
anti the chemical nature of the diffusant. Therefor. D
i s referred to more orreccI y as a diffusion roe! ricier'(
rather than as a constant. Equation (13-2) is known as
lick's first late.
Fick's Second law. One often wants to examine the
rate of change of diffusant concentration at a point in
the system. An equation for mass transport that
emphasizes the change in concentration with time at a
definite location rather than the mass diffusing across a
unit area of barrier in unit time, is known as lick's
second Mit. This diffusion equation is derived as
follows. The concentration C in a particular volume
element. (see Figs. 13-2 and 13-3) changes only as a
result of riot flow of diffusing molecules into or out of the
region. A difference in concentration results from a
difference in input and output. The concentration of
diffusant in the volume element changes with time, that

326 Phy$1t411 Pharmacy
Mrmbiu oR Receptor
Receptor
compartment
Flue in Flue
Donor
compartment1
Flow of solvent
to maintain sunk
cond tin no.
hE. 13-2. D._:''v veil [mr,, c, r';.ar'-vmm, alre
is. as the flux or amount diffusing changes wth
in the x directio n , or-
distance, is different (rotc e, and v. iz diffcrcr.t from c,. C, would be CQiiiii lv
= C,,. for example. only if K cure fl//il a value of tinily.
a,
dl ('2'
Differentiating the first-lass expression. equation
(13-2), with respect to r. one obtains
aj
&.r ii.t-
Substituting C a? from equation (13-3) into equatton
(13-4) results in Fick's second law, namely
ac
= D
at
Equation (13-5) represents diffusion only in the z
direction. If one wishes to express concentration
changes of diftusant in three dimensions. Fick's second
aw is written in the general form
= D ( C , c I (13-6
at 'ax- a- a:-
TIns expression is not usually needed in pharmaceutical
problems of diffusion, however, since movement in one
direction is sufficient to describe most cases. Fick's
second law states that the change in concentration with
time in a particular region is proportional to the change
in the concentration gradient at that point in the
system.
Steady State. An important condition in diffusion is
that of the steady state. Fich's first law, equation
(13-2), gives the flux (or rate of diffusion through unit
area) in the stead' state of flow. The second law refers
in general to a change in concentration ufdiffusant with
time, at any distance, x (i.e., a nonsteady state of flow).
Steady stute ma y be described, however, in terms of
the second law, equation (13-5). Consider the diffusant
Concer,'v..tinn and flux are often wntten as C(z,i) and J(n,i),
rvzpec'ti;. in (tulph&at3Ie that these paranneLers are functions of
both diaL,',.. .r and time I.
Donor comp.arlment
compartment (sink)
High conceflttatiofl
of pentirani
molecules
(solute)
L
'I
of'
Thickness
of barter
Fig. 13-3. Concentration grathent of ,iif(usant scrubs the Oiapfiractr.
of a dIf,isior. roll. It is no-Mal for the cvnconiratlOfl curve in mr,,re;ee
Cr decicane sharply at the bouridaren of the barrier Since, iii general,
oricinally dissolved in a solvent on the left-hand corn-
partment of the cell shown in Figure 13-2. Solvent
alone is placed on the right-hand side of the barrier, and
the solute or penetrant diffuses through the central
barrier from solution to solvent side (donor to receptor
compartment). In diffusion experiments, the solution in
the receptor compartment is constantly removed and
replaced with fresh solvent to keep the concentration at
(13-5) a low level. This is referred to as "sink conditions," the
left compartment being the source and the right
compartment the sink.
Originall y , the diffusant concentration will fall in the
left compartment and rise in the right compartment
until the s y stem comes to an equilibrium, based on the
rate of removal of diffusant from the sink and the
nature of the harrier. When the system has been in
existence a sufficient time, the concentration of diffti-
cant in the solutions at the left and right of the barrier
PL-COiaC constant with respect to time, but obviousl y
not
the same in the two compartments. Then within each
diffusiorial slice perpendicular to the direction of flow,
the rate of change of cOncentration, dC/dt, will be zero,
anI b y the second law,
= 0 (13-7)
dt dr--
is the concentration of the permeant in the barrier
C
expressed in mass/cm 3 . Equation (13-7) demonstrates
that since D is not equal to zero, aC/dr = 0. When a
second derivative such as this equals zero, one con-
cludes that there is no change in dC/dr. In other words,
the concentration gradient across the membrane dCidx
is constant, signifying a linear relationship between
concentration, C, and distance, x. This is shown in
Figure 13-8 (in which the distance x is equal to h) for
drug diffusing from left to right in the cell of Figure

13-2. Concentration will not be rigidly constant but
rather is likely to vary slightly with time, and then
dC/dt is not exactly zero. The conditions are referred to
as a"quasi-stationary state, and little error is intro-
duced by assuming sleacly state under these conditions.
Fick adapted the two diffusion equations, (13-2) and
(13-5), to the transport of matter from the laws of heat
conduction. Equations of heat conduction are found in
the book by Carslaw. General solutions to these
differential equations y ield complex expressions: simple
equations are used here for the most part, and worked
examples are provided so that the careful reader will
have no difficult y in following the arguments of disso
lution and diffuswn.
If a diap hragm separates the two compartments of a
di ffusion cell of cross-sectional area S and thickness /,
and if the concentrations in the membrane on the 1(-ft
(donor) and on the right (receptor) sides are C, and
respectively (Fig. 13-3), the first law of Fick may be
written
-_Ca) concentration may vary widely depending oil so)-
(13-8)
in which (C l - C0Y/i approximates dC/dx. The gradient
(C l - C 2) 1)1 within the diaphragm must be assumed to
be constant for a quasi-stationary state to exist.
Equation (13-S) presumes that the aqueous boundary
layers (so-called static or unstirred aqueous layers) on
both sides of the membrane do not significantly affect
the total transport process.
The concentrations C, and C0 within the membrane
ordinarily are not known but can be replaced by the
partition coefficient multi p lied by the concentration C
on the donor side or C, oil receiver side, as follows.
The distribution or partition coefficient. K. is given by
K Li C2
(13-9)
Hence
dtl = DSK(Cd - C, (13-
101-1)
T h
and, if sink conditions hold in the receptor compart-
ment, C 0,
dAf= DSKCd =PSCiOS- 10b)
in which
P = (cmJsec) (13-11)
It is noteworthy that the permeability coefficient, also
called the permeabiiity, P, has units of linear velocity.
s contusion arises wben the permeability eoethent is defined by
P - DK(crn'ilee) as used when D and K are not independeniiy
known, EAMgum (13-11), induding A In the denominator, is the eon-
________ permsebility.
Chop(cr IS Daf1,..ioe and Dissolution 327
In some cases, it is not possible to determine D, K. or
li independently and thereby to calculate P. It is a
relatively simple matter, however, to 'measure the rate
of barrier p ermeation and to obtain the surface area S
and concentration Cd in the donor phase and the amount
of permeant M in the receiving sink. One can then
obtain P from the slope of a linear plot of lf versus 1:
M = PSC5 1 (13-121)
providing that C,, remains relatively constant through-
out time. If C d changes appreciably with time, one
recognizes that C,, = Md/i'd, theamount of drug in the
donor phase divided b y the donor phase volume, nd
then one obtains P from the slope of log C,, versus 1:
'log C,, = log C(0) - 2.3O3Vd (13-12)
The flux J of equation (13-8) is actually proportional
to a gradient of thermodynamic activity rather than
concentration. The activity will change in different
solvents, and the diffusion rate of a solvent at a OCntIte
vent employed. The thermodynamic activity of a drug
may be held constant (a = 1) ill delivery form by using
a saturated solution in the presence of excess solid
drug. Unit activit y ensures constant release of the drug
at a rate that depends on the membrane permeability
and the geometry of the dosage form. Figure 1:3-4
shows the rate of delivery of t\';u steroids non, a device,
providing con s tant drug activity and what is known as
'zero-oi'der release.' The reader is famiiiar with zero-
order process from a study of kinetics (Chapter 12). If
excess solid is not present in the deliver',' form, the
activity decreases as the drug diffuses out of the device,
the release rate falls exponentially, and the process is
referred to as first-order release. analogous to the
well-known reaction in chemical kinetics. First-order
release from dosage forms is discussed by Baker and
Lonsd ale.5
2.0
1.5
:i 1.0
0.5
0
0 10 20 30 40 50 60 70
Time (hours)
flg. 13-4. Drag release for two st&oids from s matrix or device
providing zero-order release. (After H. W. Baker and H. IC
L,onsdale, in ControUed )? eie.o.se of BioloØcoily Active A9en11, A. C.
Tanquary and H. E Lacey, Eds-, Plenum Press, New York, 1074, p.
30.)
 328 PAVvical PAar,,aricy

A constant-activity dosage form may not exhibit a to be 47-5 minutes. The original concentration C, i s 0.003 mmole/cna*
steady-state process from the initial time of release. The amount of steroid passing through the membrane in 4.0 bourn on
Figure 13-5 is a plot of the amount of butylparaben 3.65 a I0'° rnmole.
(a) Calculate the parameter, DK. and the permeability, P.
penetrating through guinea pig skin from a dilute
aqueous solution of the penetrant. It is observed that 3.65 a io mmole . -
Q - 0.35 a 10 mmole/cm-
the curve of Figure 13-5 is convex to the time axis in 10.36 cm0
the early stage and then becomes linear. The early (°'° msndeicm')[ 0 hr -
= DK )hrJ
stage is the nonsteadv-state condition. At later times.
the rate of diffusion is constant, the curve is essentially DK 0.0031 cm z,hr 8.6 X 10.1 cn14sec
linear, and the system is at steady state. When the P 13K/i (8.6 a 10i cm2 'necl0.085 cm 1.01 x io° err/sec
steady-state portion of the line is extrapolated to the (h) t.1umg the lag time, 1, /i,.g0 calculate the diffusion
time axis, as shown in Figure 13-5, the point of coefficient,
intersection is known as the lag time, Q. This is the
h 2 0. cm-'
time required for a p enetrant to establish a uniform D
61 & 6 :0 47.3 nun
concentration gradient within the mumbrane separat-
= 254 9-e cm0/min
ing the donor from the receptor compartments.
or
In the case of a time lag, the straight line of Figure
13-5 may be represented by a modification of equation 4.23o10 emsec
(13- 10): id) Combining the permeabilit y , Equation 113-111, with the value
of 13 from W. calculate the partition coefficient. K.
SDKC,, Ph (1.01 s to -s rrmnec)(0.018S cr0) -
(13-13)
1.23 e 10 'cmnt'sec
The lag time, 1,, is given by Partition coefficients have already been discussed in the
chapter on solubility.
IL = ( 13-14a) Diffusivity depends on the resistance to passage of a
diffusing molecule. Gas molecules diffuse rapidly
and its measurement provides a means of calculating through air and other gases. Diffusivities in liquids are
the diffusivit y D, presuming a knowledge of the smaller, and in solids still smaller. Gas molecules pass
membrane thickness it. Also, knowing P. the thickness slowly and with great difficult y through metal sheets
h can be calculated from and crystalline bai'nerv. Diffusivities are a function of
1
the molecular structure of the diffusant as well as the
I, = / (13-14b) barrier material. Diffusion coefficients for gases and
liquids passing through water, chloroform, and poly-
Exarrrp(e 13-1. A newly s ynthesized steroid is allowed to pass meric materials are found in Table 13-i, Approximate
through a oiiexaoe membrane. having a cross-sectional area S of 10.38 diffusion coefficients and permeabi[ities for drugs pass-
cm 0 and a thickness rr o(O.045 cm. i r. a diffusion ce3 at 25' C. From the
ing from a solvent in which they are dissolved (water,
horizontal intercept of a plot u(Q = .11'S vs. 1. the lag time t is found
unless otherwise specified) through natural and syn-
thetic meeribranes are found in Table 13-2.

5}-
I St 0
TABLE 13-I. Diffusion Coefficient's of Compounds in Various
4 Media

Prt,al Molar
Volume 13 a 106 Medium or Barrier
0uffusnt (cm'rnole) (cm 2.5ec) (temperature. 'C)
'.7 Stnady state
Ethanol 40.9 12.4 Water (251
2 riPeri)arol 59.5 88 Water (25')

E Ft(roramide 26 17.2 Water 25')
I Nnnnteady Giycire 42 9 10 6 Water 325')
I Sod i um laurpi Oul(ate 235 62 Walor (251
Glucose 116 6.8 Water (25')
Hexane 103 15.0
,su( time) Hexadecar'e
Chloroform (25')
255 78 Chloroform (25')
0 5 10 15 Methanol 25 26.1 Chloroform (25')
20 Acetic acid Surer 64 14.2 Chiorolorm 1251
Time (hours) Methane 22.4 1.45
Natural rubber (40')
nPenlane - 6.9 Sdore rubber (501
Fit. 13-5, But y l pa.raben diffusing through guinea pig skin from Neopentane - 0.002 Et:ivce)Iu)ose (50')
aqueous solution. Steady-state and nonsteadv-state regions are
shown. (From H. Komatsu and M. Suzuki, J. Pharm. Sd. 6$, 593, * From G. L Flynn, S. H. Yalkow5ky. and T. J. Roseman. 5, POarro. Sc. 63,
1979, reproduced with permission of the copyright owner.) 507, 1974, reproduced with permission of the Cogysglnt ownel.
 Chapter 13 Diffusion and Dissolution 329

TABLE 13 .- 2. Drug Diffusion and Permeability Coefficients

Membrane Diffusion Membrura Permeability

Coefficient Cocilcent Ten eratum
Drug (crn2Iec) Ccrn'scc) Pathway ('Cf Reference

Benzoic acid - 36C-siO' Absorption from rat je j unum 37 a

Bulyl p-aminenriole 2.7 x - From aqueous solution 37 b
through elastic membrane
Chloramphenicol I 1.87 s 10° ThrciuFtt mouse skin 25
- 5.02 n 10 Through moos,' r,kn 37
E:hyrrodol daceI.vte 3.94 n 10 - Release from s. AsIc matrix 25 d
13.4 a 10 2 crr:
Es! rose 20.7 A050'OtiSfl from rat jejunum 37
Fuocrrrolone aceicrue 1.11 x 108 From 30% propIerre glycol_ 25
(4 a 1Q cm°hr) 70% water solvent through
a polyethylene membrane
H s' Or ocOrt . SO re 0.56 (0 Absorption from rat je1urrurrr 37 a
5.8 •' t0 Absoitrtror from rabbit vaginal 37
tract
S'edrooyoroeuierune acetate 3.7 a Release from silsOc matrix 25 g
MCC,: 12 TIC 6 154 X 10" Absorption from rat jeusum 37 a
Octoror 12 10' Absorption from rat e5unum 37 a
0tsrrorc acrd 39..10" Absorption trm rat jejunum 37 a
Pr c.eSle rose 7 x 10" Arcsortroc from rabbit vaginal 37 .3
tract
Frooraglandis. I 55I-retril- - C 58 a In situ ebsorpt.cs from rat 37
,u. r jejunum
Salcylates 1.69 a 106 Diflirsios across cellulose 37 h
membrane no
Sat.cylic acid - 10.4 u Absorption from laS jejjnum 37
Testosferorre 20 x 10' Absorption from rat jcjunurra 37
Wale' 2.8 a' 2.78 a Diffusion intD human okr, 37
Ia yC's
a IN F. H Cc, J. Y. Pa r k, I'.' Mo'czowch and hh I. H.guchi, Physical model aprca:h IC Ons.gv of C'un rufh irnproed riesl,nal ahso,or.c' is Canigs of
S oc• .'ac6t.ca' PmOt'O.er Ivri'uj.' Pr-O'ugu anc' Analog y . E. B. hucir. Ed . A'v,r y n °r-a,ma-.eurc.ai Ass.ccal on, /.Ca2erny 05 Flrarrr.aceal 'ca Sciences,
DC 1977 Chaple' 8. or 154- 151
C -- anc 5 H Yao.asCy. J. Pram Sc 61, 535. 1572.
C A J Agar and C k F.o.m. rrrrn Sc. 58, 210. 1569
a. Y. W. COws, in SuxerieO a'S Co."!'u'led Rerease Dug De r .ey 5srerrrr. J. B. Robr.sov. Ea., Marcel Drinker. Nm. snea. 1978. Criapre' A
4. S Tar. D. Danelsur aria V. n.alSe'Sor,. J. Phans Sc. 6$, 275 1979.
I. N. F. H. Ha, L. Suf'ara, S i ' .uIg. C. Ciwado. A CoSoeha C L. FInn, W I. H'guch and J. Y. Park. 4, Pha,rn Sc' 65, 1575. 1976
7, J. Sonora.,. J. Prom Sc 61, uf . 1972. Y. W. Chico. 'r, J Lan,Deri amid C C Gran: .5 Fharm. Sc.. 63, 355 1974
S K. F Fern5 and K. C Neisurr, J. F'rrarrr. Sc, 66. 1611. 1977
R. J. Scrreuo;e,n. J. lr.,eSt Der,rrol 45, 334. 1565.

itt the chapter on colloids, p. 401, we soul see that
the molecular weight and the radius of a spherical
protein can be obtained from a knowledge of its
diffusivity.
PROCEDURES AND APPARATUS
A number of experimental methods and diffusion
cells have been reported in the literature. Examples of
those used mainly in pharmaceutical and biologic trans-
port studies are introduced here.
Cells of simple construction, such as the one reported
by Karth CE al. 6 (Fig. 13-6), are probably best for
duffuston work. They are made of glass or clear plastic.
are easY to assemble and clean, and allow visibility of
the liquids and rotating stirrer. They may be therrno-
stated and lend themselves to automatic sample collec-
tion and assay. The donor chamber is filled with drug
solution. Samples are collected from the receptor
compartment in an automatic fraction collector and
subsequently assayed spectrophotometrically. Experi-
ments ma y be run for hours under these controlled
conditions.
Tuber and Rhodes 7 constructed a Plexilas three-
compartment diffusion cell for use with either synthetic
or isolated biologic membranes. The drug was allowed
to diffuse from the two outer donor compartments in a
central receptor chamber. Results were reproducible
and compared favorably with those from other workers.
The three-compartment design created greater mem-
brane surface exposure and improved analytic senst-
tivity.
The permeation through plastic film c; water vapor
and of aromatic organic compounds from aqueous
solution may he investigated in two-chamber glass cells
s:rnilar in design to those used for studying drug
solutions in general. N asinu et 51.6 reported on the
permeation of 19 aromatic compounds from aqueous
solution through pol yethylene films. Higuchi and Agu-
jar9 studied the permeability of water vapor through
enteric coating materials using a glass diffusion cell and
a McLeod gauge to measure changes in pressure across
the fun. .
 330 Physe4PAo,mnacy
Stirrer
Sampling Port
'Ie^
LJ [J
Receptor
Compartment
Donor Membrane Clamped
Compartment oerween Teflon SLlOpOrl
Fig 3-6. Sim i.,Ie riaIsion cel. After M, C. xarth, tV. I.
ins .1. L. F y i. J Phanin. Si. Ti, 'itO, i955, reproduced ':rh
reion of hr cop yright owner,)
The sorption of gases and vapors may be determined
b y use of a microbalance enclosed in a temperature-
controlled and evacuated vessel that is capable of
weighing within a sensitivit y of 2 x 10 6 g. The gas or
vapor is introduced at controlled pressures into the
glass chamber containing the polymer or biologic film of
known dimensions, suspended on one arm of the
balance. The mass of diffusant sorbed at various
pressures by the film is recorded directly, tO The rate of
approach to eq uilibrium sorp tion pet-riots easy calcula-
tion of the diffusion coefficients for gases and vapors.
In studying percutaneous absorption, animal or hu-
man skin, ordinarily obtained by autopsy, is employed.
Scheuplein" described a cell for skin penetration
experiments made of Pyrex and consisting of two
halves, a donor and a receptor chamber, separated by a
sample of skin supported on a perforated plate and
securely clamped in place. The liquid in the receptor
was st i rred by a 1'o1lonäted bar magnet. The appa-
ratus was submerged in a constant-temperature bath.
and sam p les were removed periodically and assayed by
appropriate means. Fr compounds such as steroids,
penetration was slow, and radioactive methods were
found necessary to determine the low concentrations.
Wurster et al. developed a permeability cell to
study the diffusion through stratum corneurn (stripped
from the human forearm) of various permeants. includ-
ing gases, liquids, and gels. The permeability cell is
shown in Figure 13-7. During diffusion experiments it
'.1-as kept at constant temperature and gently shaken in
the plane of the membrane. Samples were withdrawn
from the receptor chamber at definite times and
analyzed for the pet-meant.
The kinetics and equilibria of liquid and solute
absorption into plastics, skin, and chemical and other
biologic materials may be determined simply by placing
sections of the film in a constant-temperature bath of
the pure liquid or solution. The sections are retrieved at
A
I
C
Fig. 13-7 Dt(fuson cell for p errneat:on iorioh s trm .oed skin lavers..
The nerrr..3rii ma y be in :1w frnt CIa r.is.. iquic. .irgni. ey. -i. gia
wpper: C. gIa.a chamber: C. aiurn:nij,-n .oitir: D. membrane and
'ample holder. i From 1) E. Wurter. J. A. Ui:i'c-imga ind L. E.
%12chee,n. Jr.. 0 Pharmi Sri. 65, 1.1041. 11 If). I579. w:mrn,dnced with
pe rmission iii :lte copynghc sorer.
various times, excess liquid is removed with absorbent
tissue, and the film samples are accurately weighed in
tared weighing bottles A radioactive-counting tech-
nique also may be used with this method to anal yze for
drug remaining in solution and, by difference, the
amount sorbed into the dim.
Partition coefficients are de terinineil simply by eq itil-
ibrating the drug between two immiscible solvents in a
suitable vessel at a constant temperature and removing
samples from both phases, if possible, for analysis.'3
Adclicks et al. ' described a new flow-through cell,
Grass and Sweetanai proposed a diffusion cell for the
study of gastrointestinal permeation, and Addicks et
al. lb designed a cell that y ields results more comparable
to the diffusion of dru gs under clinical conditions.
Equilibrium solubilities of drug solutes are also re-
quired in diffusion studies, and these are obtained as
described earlier (Chanter 10).
DISSOLUTION
Biopharmaceutics and the modern design of dosage
forms, as dealt with later in Chapter 19. are based
partly on principles of dissolution and diffusion theory.
The present chapter lava a foundation for the stud y of
these topics by wa y of p resenting concepts,
i illustra-
tions, and worked exam p les. Dissolut on is introduced
first, followed by examples of diffusion front the
literature, with applications of both subjects to phar-
maceutical problems.
Dissolution Rate. When a tablet or other solid drug
form is introduced into a beaker of water or into the
gastroinrestinal tract, the drug beg-ins to pass into
solution from the intact solid. Unless the tablet is a
continguous polymeric device, the solid matrix also
 Chopfrr ia 1i.k"; and Di.txolufion 331

disintegrates into granules, and these granules
deaggregate in turn into fine particles. Disintegration,
deaggrogation, and dissolution may occur simulta-
neouslv with the release of a drug from its delivery
form. These steps are separated for clarificaCion as
depicted in Figure 13-8.
The effectiveness of a tablet in releasing its drug for
s y stemic absorption depends somewhat on the rate of
disintegration of the dosage forms and cleaggregaticin of
the irranules. Ordinarily of more importance. however,
is the dissolution rate of the solid drug. Frequently.
dissolution is the limiting or rate-controlling step in
hioabsorption for drugs of low soltihility, because it is
often the slowest of the various stages involved in
release of the drug from its dosage form and passage
into s v.stemic circulation. Dissolution has been re-
i-ic-wed by \Vurster and Ta y lor. Wagner. and 1_cc--
IS
son and Carstensen. Release rate processes in gener:ll
are discussed b y W. Hiruchi. 20
The rate at which a solid dissolves in a solvent was
ur000a(d in quantitative terms by Noyes and Whitney
in ))I7 and elaborated subsequently by other workers.
The equation ma y be ai'ittc-n as
= (Cl- C) (13-15)
(Ii Ii
or
dC DS C,
=- (13-16
dt ii
in which At is the mass of solute dissolved in time 1,
d.iI ' dt the mass rate of dissolution(mass time). D the
diffusion coefficient of the solute in solution, S the
sat-face area of the exposed solid, It the thickness of the
diffusion la y er, C, the solubility of the solid (i.e.,
concentration of a saturated solution of the compound at
the surface of the solid and at the temperature of the
experiment). and C the concentration of solute in the
bulk solution and at time 1. The quantity dC/cit is the
disd1ut ion rate and V the volume of solution.
In dissolution or mass transfer theory. it is assorted
that an c.qucous diffusion layer or sla q ,icnt 10! uitd J ri
of thickness h exists at the surface of a solid underging
dissolution, as observed in Figure 13-1'. This tnieknsss
Ii represents a stationary layer of solvent in winch the
solute molecules exist in concontrations from C, to C.
Be y ond the static diffusion la y er, at .r greater than it,
r
nixinic occurs in the solution, and the d ug is found at
a uniform concentration. C, throughout tb- hubi phase.
At the solid surface - diffusion layer interface, i CI,
the drug in the solid is in e q uilibrium with drug in the
(iffusion layer. The gradient, or change in concentra-
tion with distance across the diffusion la y er, is con-
slant, as shown by the straight dovnward-siopiiig wic-.
This is the gradient represented in equanons (13- iS)
and (13-16) by the term (C, - C).'h. The similarity of
tte Noves-Whitney equation to Rick's first, law is
evident in equation (13-15).
When C is considerably less than the drug's solubil-
ity, C,, the s ystem is represented by sinic conditions.
III concentration C may be eliminated front equations
i13- 15) and 03- 16. Equation '13-25 1 then becomes
dMldt = DSC,/Ii (13-17)

1
TABLET
OR cP:utE:

GRANULES OR
II oeor.
DRUG IN
SOLUTION
DRUG IN
BLOOD, OTHER
AGGREGATES tin sirs or FLUIDS AND
In sirs) TISSUES

[FINE PARTICLES

Fit. 13-8. Disintegration, deaggregation, arid dissolution stages as a drug leaves a tablet or granular matrix- ' (From John C. Wagner.
4l, p.
8iophwnaceu&a and RcM'assi PAarnsacokinelic, published by Drug Intelligence Publicationa, hoc., 1241 Brosdway. Hamilton, IL
99, with permission of the copyright owner.)'
 332 PAysitoJ PIUOTm4C1 -

Aqueous
Solid dMusion Bulk
dosage ronii laerlseagnant solution
liquid ifirnf

! Sampling
tube
Matrix

xh
Fig. 13-9. DissoluOuc of a drag from a oisiil rnatr.x, oho.vr':.c
stagnant diffusion layer b'aen the •Josae form surface and wik
solution.

Paddle

In the derivation of equations (13- 15) and (13- 1(3), it

was assumed that h and S were constant, but this is not
the case. The static diffusion layer thickness is altered
by the force of agitation at the surface of the dissolving
tablet and will be referred to later. The surface area S
obviously does not remain constant as a powder,
granule, or tablet dissolves, and it is difficult to obtain
an accurate measure of S as the process continues. In
experimental studies of dissolution, the surface may be
controlled by placing a compressed pellet in a holder
that exposes a surface of constant area. Although this
ensures better adherence to the requirements of equa-
tions (13-15), (13-16), and (13-17) and provides
valuable information on the drug, it does not simulate
the actual dissolution of the material in practice.
Dissolution of Tablets, Capsules, and Granules. A
number of methods for the in vitro and in vivo testing of
dosage forms have been suggested." The purpose of
an in vitro dissolution stud y is to provide a fast and
inexpensive notFiod that correlates'.vith the p e-rfor-
mance of a d:sne form in humanubjecs, and 1
number of studies towards this end have been reported
in the literature. 23 Various dissolution apparatus and
methods are described in some detail in the Latter
work. The Hansen paddle equipment and a research
apparatus provide two convenient systems, as shown in
Firure 1:3- tOo and b. As obse rved, the apparatus are
similar except that the surface area of the tablet or
compacted material in Figure 13- lOb remains constant
as the drug dissolves. This design has advantages in
research and product formulation. Furthermore, exact
hydrod y namic conditions are maintained by the fixed
position of stirrer and sample holder. Much attention
has been paid to the rotating disk, a modification of the
apparatus in Figure 13-10b, in which the tablet in its
holder is attached to a rotating shaft of a precision
variable-speed motor. 2' The paddle of Figure 13- lOb is
not required in the rotating-disk apparatus. The paddle
equipment of Figure 13- 10cc is currently known as the
Disintegrating
and dissolving tablet
Mato(
Paddle
Circulating
constant- --
tem p erature
liquid Tablet or
compacted powder
(9)
Fig. 13— 10. I)iooluii 5 n apparatus. (a) Hansen paddle equipment for
ranules and tablets. (h) Research design to ensure a constant surface
area of tablet or compacted powder as the drug dissolves and diffuses
of the dosage form lFrom A. P. Simnonelli, S_ C. Mehta and W. I.
)lig'jchi, J. Ph.u-m. Sci. 53, 52-9. 1969, re produced with permission of
the copyright owner.)
US? Dissolution Apparatus 2, and a rotating basket
apparatus (not shown) is referred to as liSP Dissolution
Apparatus 1.
In calculating the diffusion coefficient and dissolution
rate constant, the application of equations (13- 15) to
(13-17) is demonstrated by way of the following two
examples.
&amp!a 13-2. A preparation of drug granules weighing 0.53 g and
having a total surface area of 0.28 m 2 (0.28 X 104 cm°) is allowed to
dissolve in Sv) aoL of water at 252 C. After the brat minute. 0.76 g
have passed into solution. The quantity CIA may be referred to as a
dissolution rate constant, k.

If the solubility C. of the drug is 15 mgimL at 25' C. what is k'
yrom equation (13-17). M changes linearly with t initially. and
di! 60mg
- -7 - 12.67 Mg sec
di 60 sec
12.67 mgfsec k a 028 x 104 cm t v 15 rngcm3
A = 3.02 a
10-4 cvnJsec
In this example, 0.760 g dissolved in 500 ml, after a
time of 1 minute or 760 mg 500 mL = 1.5 rngcm 2 . This
value is one tenth of the drug's solubility and may be
omitted from equation (13- 15) without tnzroducing
significant error, shown by employing the lull equation
12.67sec
/=
(0.28 x lO cn) (15 mg.cm t - 1.5 mp;cm')
k = 3.35 x 10ri cm/sec
When this result is compared with 3.02 x 101
cmsec. obtained using the less exact ex pression, it
signitleS that "sink conditions" are in effect, and the
concentration term (' may be omitted from the rate
equation.
'Example 13-3. The diffusion la yer thickness in Exnsipk i.O-' is
estimated to be 5 x iu"' cm. Calculate D. the diffusion coefficient
using the relation 1: LI
LI (3.35 10'em/s.ec) s (50 iü cm
= lotS a 15' 6 cm2,sec
The reader will find additional examples in the bool': by
Cat'stenser.
Powder Dissolution: The Hiason- Crowell Cube Root Law.
For a drug powder consisting of uniformi y sized
particles, it is possible to derive an equation that
expresses the rate of dissolution based on the cube root
of the weight of the particles. The radius of the particie
is not assumed to be constant.
The parcle (sphere) shown in Figure 13-11 has a
radius r and a surface area 4r7-2 . Through dissolution,
the radius is reduced by dr, and the ir.fin:tesimal
volume of this section lost is
w g
Fit. 13-11. Sdierruitir 0(6 particle, sho in change in surface ares
and volume as the partie4c dissolves. The volume dV dissolved ii'.
seconds is given b' X surface dr x 4n'm°, (After
J. T. Carstensen, PAo.rinocewdaca of Solids and Solid L10609e Forms,
Wiley, New York. 1977, p. 75.),,.
Cia plrr ;,t rj.on and Tho)ut,00 333
dV = 4tr2 dr (13-18)
For N such particles, the volume loss is
(13-19)
cfV = 4]'r7 dr
The surface area of N particles is
S = 4Nn (13-20)
Now, the infinitesimal mass change as represented by
the Noyes-Whitneylaw, equation (13-15), is
-dM = kSC > dt (13-21)
in which k is used for TJih as in E.roinpir 13-5. The
drugs density multiplied by tliC infinitesimal volume
change, p aT, may be set equal to dM, or
- p d\' = kSC 1 iU (13-22)
Equations (13- l9 and 03-20) are substituted into
eq uation (13-22) to yield
- )oN d = 4Nr ia kC, dl (1::;-23)
Equation (13-: ) is divided through b y 4l.V7ln to give
-p dr = kC, dt (13-24)
integration with r = r, at I = 0 produces the expression
kC, I
r = r0 - ( 13-25.)
The radius of spherical particles may be ?'eplaced b y the
n usi g the relat ionship taco inside
mass of .V particles by
front cover for volume of a sphere).
M = .\'p(6)d3 (13-26)
in which d 2r. the diameter of the particle. Taking the
cube root of equation (13-26) yields
The diameter d or equation (13-27) is substituted for 2r
into (13-25), giving
- (13-25)
in which
2k C, M0 ii 2k (7
- = —f- - (13-29)
p e p
Mo is the original mass of the drug particles. Equation
(13-28) is known as the Hixson-Crowell cube root
law, 26 and K is the cube root dissolution rate constant.
LxmpIe 13-4. A speriaiiv prepared intootnnlice of
uniformly sized narticle.c, with a diameter of itO oi'', wrghLed 75
m. Dissolution of the drug was deterir.ined in loX..! rol, of soater at
25' C as a function of time. Determine the raiue of K, the rune root
GLSSOIUtIOS rate constant, at each time lniCrra slid cokutate the
average value of v. The data and rrouii.s are Set forward in Ta-
ble 13-3.
'The symbol, Wn, stands for micrometer arid is equal to 10'
meter
 334 Physical Ph.ar-rnacy
TABLE 13-. 3. Df=latitia ol Tolbuta.Jde Powder"
0oncentra)io weigh)
Time Dissolved ljndissolvee
(mini •mgmLi (grams) M f
0 '0 M0 0.0750 0 -
10 0.01970 0.0553 0, 3.106 0.0041
20 0.0374 0.0375 33866 00013
30 0.0510 0.0240 3 :332 3.0014
40 0.0595 00155 3.1724 0.0043
0.0041
50 10650 0.0100 02063
00212 000424 g ! 'own
In the situation in which the aqueous diffusion layer
thickness about a spherical particle is comparable to or
aiger than he size the sphere, for example.
of
micron i zed particles less than 50 ivm in Ufumeter, the
change in particle radius with time becomes
?J)C
= -- ---- 13-30)
and the estimated time for com p lete dissolution, T, (i.e
When i 0)15
= . (13-31)
Example 13-5.isdirucai macbce. diaze pam :o)Oetlon a oierOe
ooxr:on ot diazeparn in a oropvleve glvcol - cihanol - water eustiver.:
sy .acrnt a often diluted macfold with normal saline inectron.
An Incipient prec;PItatlofl of dia.ze panr occurs near_ablyDori addi-
tion of saline o0oa'est by cain p iete ilissoiutioo svitiuii a fltflUtO
upon shaking. '.Vaii C, ii water " 3 rng'mL. p 1. pp....and
D 3 e )0" crC see. calculate the time for complete dissolution
shen c,, II) ivin tIll < 10-'
- 11 goi X II x 10 'cm)2
215 i 10 coi t seC l) 3 < j-J g.mLl
t
5
= 2 ti. = 108 sec.
Convective Diffusion. Convection, the transfer of heat
neriç'C and ttto presr-ce af OL'ratlon accononanyjng
the t:o'.'ometl uf a tutcl, ma y be combined vrth
diffusion to provide a cont'cctii'e dffiisioio model for the
study of dissolution. 2" The convective diffusion irlOviOl,
unlike the simpler Noves-Whitney and Nernst-Bi'Un-
nor ,sp p roachen, takes into consideration such factors as
flow rate, mix i ng (agitation), and the dimensions of the
dosage form. Nelson and Shah 2 ° have investigated the
convective diffusion model for the dis.solution of a!kvi
p-aminohenzoates as test compounds. Dc Smidt et al."'
also used a convective diffusion model in the stud y of
the dissoution kinetics of griseot'ulvin in solutions of the
solubi[ir-.ing agent sodium dodecylsulfate.
Ex.nr.plC 13-6. The above warloera" introduced a drug dasoltitivo
rate approach to 010.iel the rates of dissolution of alloy) p-arnioobcri.
O'iates in a specially designed diffusion cell. The model is based on
convective diffusion, the eosaticr1.1 of which may be used to calculate
R, the rate of diffusion or permeation rate:
R 0,3ij3 riC, a bL m (13-32)
for i rectangular tablet suilure of width b arid length 1. in the
direction of flow, and
6< - 2.137 oac.0 1n (13-33)
for a crreular ta blet surface if radius r. In thene equ.atrovu U is the
difTuziviiy or diffusion roelitcient, C is the solubility. anti 'a is the rare
of shear as the iou era is pumped over the dissoiving surface. The
rate of shear is calculated from a - 6Q11P3I', where Q vs the low rile
aid H arid (V are the height ar.d width, respectively, of a channel in
the diffusion cell to allow the tow of solvent neater) over the
dissolving tablet.
Experimontis on disso(uuon rate, R. were carried out at 37' C seth
reetarig oar tablet surfaces containing the drug model eth y l p-soiL-
nober.zoaie. The loin' axis of the rectangular surface was 25.4 mm ar,d
the short a.sjre 3.175 TIc,.
(a and 6bCo1iipue he rote of dissolution R with the long axis ft
iiLaced ocrtreoilicaiiir to the direction of low, and then with the
.ong axis placed parallel to tOe direction of tow. The Sow rate Q is
14.9 m 7 min: the dilfusivitv and solubility of the drug are 0 = (tOil
Ill ° cm a . sec and C , 7.07 '.. 10 6 mole err 0 : and nOV = 0.3600
(ci The evc,erireer:t is retested but 'sang a disk seth a circular
surface of ea et;'aal to the surface area of the rectangle referred to
abes'e. Compute 6< expressing the results in orolernia.
dl What differences do you Er-ri between this model soil the classic
stacriant or unstirred diffusion layer model' You may care to refer Ia
the artclOs" to cheek thv answers given here.
'il The rate of shear is
a 6Q-Ff2 1' e 14.9 crn°rn;n''0.3506 cm' 255,0 min1
Forthe lung axis aertierd;eular to flow. 2.54 em and!, 03175
err). Then.
= 0.306<9 l(v Ill" cra.oec 'u 00 5?c mivr 0<
c7,27 < II) tnele.cm t l o 25.5. rrin - ':'i x
2.54 carl .0 (03173 crn)°
3,10 x IS roole,'rnrn
For the long axis parallel to the flw. o " 0.1)73 coo soul C
2.34cm.
0.006(9.8)3 x i0° sec 0< 0) .vce.nrrint°' 5<
(7.27 0< lO rnolecrn's ) .0 (25-b. mirC'h1
0:3175 cm) 5< 02.01 crrii°°
1.55 x lO 7 truSs Iris
Fur the Ior4 axis eerrer.'iioia(or to the 6< iots'ice the value uteri
the long as1.. 0 traralluol to the how, as observed in lot and 16< above:
0 5)0 ' L0 - I.55 -< 10-)
TS p s;rCce area of Ire rec:ar.'ular tablet a 2.31 rio a' 0 3175
ciii c) )C cm-, which is also One suoi'ace area of the circular tablet or
vlik. Therefore, the rndiu.s. r. of the c'.rcular surface is ire o.ss;e; or
0.507 cm.. and the rule fI of drffusion or perrneaivon for a tablet
of Crrcelar oirt'ace (eqraron:r 140.31 is
R 2.13715.54 's 10) o 60 crn 1 munil2 ° a'.
7.27 .0 ii',)'' male, cm') 0< 256.rno'
0.507 cm,`
= 2.26 x 10 -7 roolerrin
i .J( The cnv.,live diffusion creole!, the CD model, which tales jr-to
account iluiul flow as well as diffusion. has several parameters sri
common with the classic diffusion rood 's). These include the solubilky.
C., diffusion coefflcieot er diffusivily. 0, and the dimensions of a
rect.ar.gutsr or circular vorfau'e, b, L. and r. In the classic 'redo). 6< so
proportional to D, while in the CD model If is proport1ofl5l 10 11uo1 in
the classic model, If is prupoirtionsi to the surface area, S of a
rectangle or disk; in the CD model 6< is proportional to a reduced
faoctisn of nor-lace area; that (a, bC°' or n'' 2. A new parameter. o, the
rate of shear over the dissolving surface, is introduc e d in the CD


Chapter 15 Dffasion and Dwt!td)iir, 235

model; it is raicajlaeed from the eats rate and the dimension of the
diffusion cell. 7.

DRUG RELEASE ut

Release from dosage fcirnis and subsequent titoab- mt eliot lisa- hcrnoteoe000
sorption are controlled bV the ph ysical chemical prop- polymer manic
erties of d rug and deliverY form, ad the physiologic
Pc'yrnei
and ph y sical chemictd properi ies of the biologic systen. rr.otr n
Drug coricertrtatior., aQUeous sri]uhili:y. muc-cular sPa',
crystal form, proteit: ixndiitg, and pK 0 ire among the \. \ To nun path
t /, a! d:i5 an ii
ph y sical chemical fec: ora that must be understood to
_'( -'•'--.,'_, lea cn ire c.aU ix
design a deliverys y stem that exhibits conu'oiic-d or
sustained-release chnrarieristics. it)
The release of a du'ui from delivery system J11VOiVCS
faeturs of both dissolution and difiusion As the readcr Liquid enters pores
has alread y observed in this chapter. the foundattons of in leach out drug
diffusion and dissolution theories ear man y resent-
blat-.rc's. D i ssolution rate has been d i scussed as it Fecedint
influences drug reeasc: the rema i nder of the charter
Ce dial SC if
emphasizes principle— of d i ffusion as rc'iated it thc
transport of drugs from dosage rnatnces, througn the
ivalis of containers and p ackages, and into the locy by Static d,llus,cn ta5r -
pathways through the gastrointestinal raucosa. s:r..
vagina. buccal cavity, and other sites of entr y Into the 'Mat-u
body. layer
Drugs in Polymer Matrices A powdered Urup Is '
homureneouslv d.s perred throughout the matrix of an ]
Peret-i.'S
credible table:. The drug is assumed to dissolve :r tOy .____ .2L1d_ 'I

(CI
pol ymer matrix and to diffuse out from the surface of S
P.
the device. As the drug is released, toe distance fcr
diffusion becomes increasingly greater. The boundary Fig. .3- 12. Relrrase of drug from homogeneous and zrs_iota.- mama
do'e forms 'a) D-ag etuiec train a uiemogeta'ous aoivrrr moire,
that forms between drug and empty matrix therefore lit-ag
(ii, leached from. or grsnui2 orot'ix I C.-
recedes into the tablet as drug is doted. .1 schematic Sc).- sane of the solid ma:re, and its receding bour,aary

illustration of such a device is shown in Figure 13-12o.
Fiurure 13-12t s:onvS a rranular matrix with intercon-
necting pores or ca p illaries. The drug is leached out of
this device by entrance of the surrounding medium.
Figure 13-12c depicts the concentration proflie and
shows the receding dp!etuo zone that moves to the
cer.ter of the table: as the drug is reeascd.
Higuzitl 32 developed an eouatton for the release of a
drug from an ointnten: base and laterna applied to
diffusion of solid dru gs d i spersed in homogeneous and
granular matrix dosage systems (Fig. 13-12).
Fick's first law,
d.J - dQ - DC,
Sdt dt h
may be applied to the case of a drug embedded in a
pol ymer matrix, in which dQidf* is the rate of drug
released per wait area of exposed surface of the matrix.
Since the boundary between the drug matrix and the
'aM is amount of drug diffusing dQ is introduced here to represent
d.M/S, in which S is surface area of the boundary.
diff -'a from the dosage form. ;Frorn T. Higch.. Phanr, Sc.. 50.
S74. PSi, reproduced with porrnissicn of the cer vrzh: corret.
dr.,, ,;-depleted matrix recedes with time, the thickness
cf :'.e empty matrix, oh. through \V:",jCh tae arug
duCt-es also increases with tIme.
V.h-iereas C, is the solubuiity or saturation concentra-
tier of drug in the matrix, A is the total concentration
(arr cant per ttnit volume), dissolved and "ndissolved. of
dr.; in the matrix.
drug passes out of a homogeneous matrix, a in
Ficure 13-12, the boundary of drug (re p resented by
the dotted vertical line in Figure 13— 12c) moves to the
left by an infinitesimal distance. dh. The inflr,itessmal
amount, dQ, of drug released because of this shift of the
frcr.t is given by the approximate linear expression:
dQ = A dh - C, dh (13-35)
Now dQ of equation (13-35) is substituted into equation
(3-34), integration is carried out, and the resulting
equation is solved for h. The steps of the derivation as
given by Higuchi are


334 Pkysicv. Phernss2Oy

The drug's solubility C. is LO a 10' Skin' in the polymer. The

(A - IC,) dh d (:13-36)
cfh=Sdt (2A - C) h ?
:3-37)
C ,..338)
= 4DC b) What is the instantaneous rate of drug release occurring at 10
The integration constant, C, may be evaluated a: t = 0
at which h = 0, giving
(2A - C,)h2 = 5.5 x 10' g cm' nun-
(13-39)
/ J,DC,L )1.2 (13-40)
-
The amount of drug de p ieted per unit area of mat: :x. Q,
at time 1. is obtained by integrating equation (13-35) to
yield
Q = hA - hC, (1:3-41)
Substituting eouation (13-40) into (13-41) pmduces
the result
12
DC
- C,) (13-42)
2A -
which is known as the Hisjuc.he equation:
Q = (D(2_4. - (- ) C]12 ('343)
The instantaneous rate of release of a drug at .ime
is obtained by different i atin g eq iation 13-43) to yield
- 1 ,D(14 - C3In (13-47) is the total porosity of the matrix after the drug
(13 -.44)
d12[ I
Ordinaril y , A s C., and equation (13-43) redues to
(l -4.1)
Q = ( 2ADC,i)' 2 from the matrix and the drug's specific volume or
and equation (13-44) becomes
- tration, A, is converted to volume fraction of drug that
- '11 -46)
21
for the release of a drug from a homogeneous polymer
matrix-type delivery system. Equation (13-15) indi-
cates that the amount of drug released is propor ional
to the square root of A. the total amount of drug is. unit
volume of matrix; [), the diffusion coefficient of the drug
in the matrix; C,, the solubility of drug in poly nenc
matrix; and 1, the time.
The rate of release, dQ/dt, can be altered by inc-eas-
ing or decreasing the drug's solubi!ity C, in the po1mer
by complexation. The total concentration A of drug that
the physician prescribes is also seen to affect the i-a :s of
drug release.
Lc.incpis 13- 7. (a) What is the RmounL of drug per unit are.
relas.nd fr,rn a tablet mains at time 121) rrthiutes? The i.e ni
concer.nution of drug in the homeg'neutis rnatrto. A, is 0.02 gI:Im.'.
diffusion coefficient I) of the drug in the polymer matrtx at 25 C is
6.0 X jo-s ='/see or 360 'C 10 em1hnkn-
Equation (I3-4) is used:
Q [2(0.02 gcTTl'060 x l0n1,mnin)
ic (1.0 5 10 g/cm3)(120 min)]'-
1.3 'C 10' g/cxn'
minutes?
r (o,y)s300 x 10'Xl.0 x
2x120
Release from Granular Matrices: Porosity and Tortuosity.
The release of a solid drug from a granular matrix (Fig.
13- 12h) involves the simultaneous penetration of the
surrounding li q uid, dissolution of drug, and leaching out
of the drug through interstitial channels or pores. A
granule is, in fact, dedred as a porous rather than a
homogeneous matrix. The volume and length of the
opening in the matrix must be accounted for in the
diffusional equation, leading to a second form of the
Higuchi equation:
Q=
( 2A - eC,)C,i (13-47)
in which e. is the porosity of the matrix and is the
tortuosity of the capillary system, both parameters
being dimensionless quantities.
Porosity, €, is the fraction of matrix that exists as
pores or channels into which the surrounding Liquid can
penetrate. The porosity term, e, found in equation
has been extracted. This is equal to the initial porosity,
co, due to pores and channels in the matrix before the
leaching process begins, and the porosity created by
extracting the drug. If A g/cm3 of drug is extracted
reciprocal density is lip cm-1 /g, then the drug's concen-
will create an additional void space or porosity in the
matrix once it is extracted. The total porosity of the
matrix, e, becomes
± A(11p) (13-48)
€ =
The initial porosity e, of a compressed tablet may be
considered to be smail (a few percent) relative to the
porosity :iJp created by the dissolution and removal of
the drug from the device. Therefore, the porosity
frequently is calculated conveniently by disregarding €0
and writing
€ Alp (13-49)
Tablet porosity and its measurement and applica-
tions in pharmacy are discussed in more detail on pages
4-12 to 4.16.
 Chapter IS 'D(fa.sicrn and Diso1uion 331

Equation (13-47) differs from equation (13-43) only lipd Hynausi

in the addition of and 'r. Equation (13-43) is applicable layer I isyer
ho
to release from a homogeneous tablet that gradually
erodes and releases the drug into the bathing medium.
Equation (13-47) a pplies instead to a drug-release
mechanism based upon entrance of the surrounding
medium into a polymer matrix, where it dissolves and Dr
leaches out the soluble drug, leaving a shell of polymer 'eh "c" on Deeper layers
a t the SOrmiS
and empty pores. in equation (13-47). diffusivity is Shin uariace
mulitphed o porosit y , a fractional quantity, to account
for the decrease in D brought about by empty pores in
the matrix. The apparent solub i lit y of the drug C, is
also reduced by the volume fraction term, which Distance, a
re presents porosiv. Fig. 13— 13. Fa.s'.age if a dr.oc or. the i., in's surfhce throzc a
Tortuositv, 'r, is introduced into equation (13-47) to tiyer. 1.. irrI a hvdriu la'.t-r, C... ard irir lii dre;ier L,'
account for an increase to the path length of diffusion rarrn:. T..':' curve rO c . rr.ectrazcn o g ieri: dj'urce oharon',
at tis Ea O bouidirie. becaUNL ISo two pa-c it ii cw :.,. ,:a
due to branching and bending of the pores, as compared atues otis-c tSar. uucv.
to the ehomest "strai g ht-throu gh pores. Tortuosity
tends to reduce the amount of drug released in a given
interval of time, and so it appears in the denominator (equation L13-111i as the diffusion cfficient D niuti-
under Ihe square root sign. A straight channel has a plied b y the uartitkr coefficient K. and divided b y the
tortuosit y of unity, and a channel through spherical membrane thickness, Ii. For a parlicutar jarnina, ,
beads of uniform size has a tortuosity of 2 or 3. At
P, =
times, an unreasonable value of. ear, 1000 is obtained
for 't, as L)csai et al. 3 have noted. When this occurs,
the pathwa y for diffusion evidently is not adequately
R 1 - I'P 1 = h.'(DK,) (13-50)
described b y the concept of tortuosity. and the system
must be studied in more detail to determine the factors in which R1 is the resistance to diffusion. The total
controlling matrix permeability. Methods for obtaining resistance R is the reciprocal of the total permeability.
diffusivit y , porosit y , tortuositv. and other quantities P. and is additive for a series of la y ers it is written in
req uired in an anal ysis of drug diffusion are given by generaL as
rJesai et
I? = R1 + Th 4- . F.. (13-51a)
Equation (13-47) has been adapted to describe the
kinetics of lyophilization, commonly called freeze- IF = 1P J., . . - liT. (13-51b)
drying. of a frozen aqueous solution containing drug
and an inert matrix-buildin g substance (e.g., niannitol F = iF = h/DK
or lactose). The process involves the simultaneous + ' ' h,,!Ds-K0 (13-51c)
change in the receding boundary with time, phase in which K. is the distribution coefficient for la yer i
transition at the ice-vapor interface governed by the relative to the next corresponding layer. I - 1. of the
Clausius - Cla p eyror. pressure- temperature relation- qvszem. 16 The tc.:ai penneahilitv for the two-ply rrodel
ship, and water vapor diffusion across the pore path of the skin is obtained by taking the reciprocal of
length of the orv matrix under iow temperature and equation (13-51c, expressed in terms of two layers, to
vacuum conditions. yield
Multilayer Diffusion. Diffusion across biologic barriers
ma y involve a number of la y ers consisting of separate DiKiD.K-
1- = _______._.__ (1'-5')
membranes, cell contents, and fluids of distribution. - /i i DK + 1iD 1 K 1 -
The passage of gaseous or liquid solutes through the
walls of containers and plastic packaging materials is The lag time to steady state for a two-layer system is
also frequently treated as a case of multilayer diffusion. 2 ( h ) \
}bguchi 32 considered the passage of a topically ap-
plied drug from its vehicle through the lipoidal and
lower hydrous la yers of the skin. Two barriers, in h 22 f h1
series, the lipoidal and the hydrous skin layers of D2 2DK 1 6D2k2
thickness h 5 and /n,, are shown in Figure 13- 13. The (io-&3)
= (h 11D 1 K1 + h2ID-2K2)
resistance R to diffusion in each layer is eoual to the
reciprocal of the permeability coefficient Pi of that When the partition coefficients Ki of the two layers are
particular layer. Permeability P was defined earlier essentially the same and one of the h/I) terms, say 1, is

338 Pkysica2 Pharnwuc '
much larger than the other, however, the time lag
equation for the bilayer skin system reduces to the
simple time lag expression
= h 1 16D 1 (13-54)
Membrane Control and Diffusion Layer Control. A
multikiver case of s pecial importance is that of a
membrane between two a q ueous phases with station-
ax-', or stagnant solvent layers in contact with the donor
and receptor sides of the membrane (Fig. 13— 13).
The permeabilit y of the total barrier, consisting
of the membrane and two static aqueous diffusion
a y ers. is -'
D.,KD5
!'( h.,D. +
ii., D,,K 2hD,
This expression is analogous to equation (13-52). In
equation :13-55), however, onl y one partition coeffi-
cient K aPpears, that giving the ratio of concentrations
of the drug in the membrane and in the aqueous
solvent, K = Ca/C. = C 1 ;C0. The dux 1 through this
three-ply barrier is s imply equal to the permeability P
multi p lied by the concentration gradient (C l - Cn), that
is. J = P(C 1 - C .5 ). The receptor serves as a sink (i.e.,
C 5 = 01, and the donor concentration C 1 is assumed to
be constant, providing a stead y -state this.5'
- I dM D,,.KD,C1
(13-56)
S dl = h.D, — 2/i9D,,K gradient that controls the flux now resides in the
In equations (13-55) and (13-56), D D, are
membrane and aqueous solvent diffusivities, is the
membrane thickness, and it, the thickness of the
aqueous diffusion laye. as observed in Figure 13— 14.
.1! is the amount of perrneant reaching the receptor, and
S121Ic Suet Aqueous
Saueous d,iiuiin Oltusion receptor
Oye .iysr Meatisue ayer (Sink)
Fig. 13-14. Schrnntic of a multilayer )Lhree-pl y ) barrier. The
rr,ernbrane is found between two Italic aqueous diffusion layers.
(From G. L. Fi)-nn. 0. S. Carpenter and S. H. Yalkon sky. J. Ph arm.
Sd. 61, 313. i9?2, reproduced with permission of the copyright
owner.)
S is the cross-sectional area of the barrier. It is
important to release that h is physically influenced by
the hydrodynamics in the bulk aqueous phases. The
higher the degree of stirring, the thinner is the
stagnant aqueous diffusion layer- the slower the stir-
ring, the thicker is this aqueous layer,
Equation (13-56) is the starting point for considering
two imuortant cases of muitilayer diffusion, nanielv,
diffusion under mnnhrarie conlrol and diffusion under
aqueous diffusion layer control-
Membrane Control. When the membrane resistance to
diffusion is much greater than the resistances of the
aqueous diffusion la y ers, that is. R,, > R by a factor of
at least 10, or correspondingly, P,, 'a P.,, the rate-
determining step (slowest step) is diffusion across the
membrane. This is reflected in e q uation (13-56) when
a> Thus, equation 13-561 reduces to
.1 = ----- C 1 (13-5)
S 0,,,
Equation (13-57) represents the sim plest case of
membrane control of flux.
Aqueous Diffusion Layer Control. When 2h 0 KD,, a>
h,,,/),,, equation (13-56) becomes
J = () C 1 (1:3-58)
and it is now said that the rate-determintr.g barriers to
diffusional transport a re the stagnant aqueous diffusion
layers. This statement means that the concentration
aqueous diffusion Ia ,, 'ers rather than in the membrane.
From the relationship, 25,,KD,,, a> h,,D0 , it is observed
that membrane control shifts to diffusion layer control
when the partition coefficient K becomes sufficiently
large—
Flynn and Yalko'.vskv 17 demonstrated a transfer
from membrane to diffusion-layer control in a homolo-
gous series of n-aikvl p-aniinobenzoates çPABA esters).
The concentration gradient is almost entirely within the
silicone rubber membrane for the short-chain PA.A
0
II
HN —S\ ,(—C-0—CH5— Cl-I 2 —Ci-0---CH3
n-Butyl p—am inobenzoate
esters. As the alkyl chain of the ester is lengthened
proceeding from butyl to pc-ntyl to hexyl, the concen-
tration no longer drops across the membrane. Rather,
the gradient is now found in the aqueous diffusion
layers, and diffusion-layer control takes over as the
dominant factor in the permeation process. -
Example 13-8. The steady-Late ttu.o I for hunyl p-aminubes.xaain
was found to be 1.60 s 101 rnrnole cm - sec'. D. is 60 x iO cm°
5ec' and he consentration of the PA[IA ester, C. is 1.0 mmcm
liter'. The system is in diffusion-Layer control, so equation (13-56)
applies. Calculate the thickness of the static diffusion layer, ii,,.

fD,\ID, \
J C or t , C
= 6.0 x 10' cm 2 sec`
Sn .iaj 1(0 mmnie crn' r-e
tie > It'_mroo)e cri'r 2 )
One observes from equations (13-57) and (13-5S)
that, under sink condition, steady-state flux is propor-
tional to concentration, C. in the donor phase whether
toe flux-determining mechanism is under membrane or
'i:ffusiun . lio.e' c'';trol. Euua:ion (1f-5 y how tuet
tue flux is n'lerenent of membrane thickness h,, and
other pro p erties of the membrane when under static
diffusion layer control.
The maximum flux onisined in a membrane nrenara-
lion depend on the oluilitv. or iin:itirig concentra-
tion. of the 1'-,'.BA hontiriug. The niaxtniurn flux may
therefore be obtatned usin g equation (13-5(t in which
C is replaced b'- C. the so]uhility of the permeating
coinnouitd
1). [.).
11
3-59)
= j,,D - 2h0KD.i
The maximum stead y -state flux. i,,,. for saturated
solutions of the PABA esters is plottc-d against the
ester chain length in Figure 13-15. The plot exhibits
teak flux between n . = 3 and ii = 4 carbons, that is
between propvi and hutvl f-aminubenzoates. The peak
in Figure 13-15 suggests in part the solubility charac-
teri s tics of the PABA esters, but primariv reflects the
change from membrane to static diffusion layer control
of flux. For the meth y l, eth y l, and pi-opyl esters, the
concentration gradient in the membrane gradually
10
3L
\ example, the larger lipuph:cicitv Increases the on-t
2
Measured by gas chromaiolraPhi\
S Spectrophniomerry p
0 1 2 3 4 5 6 7
Number of carbons in alkyl chain (n)
Fig. 13-15. Sce&dy-state flux of a series of p-nminobenzoic and
eaters. Maxirnunc flux occurs between the nsserx having three and
four carbons and to due to a change from membrane to diffiaon-laver
control, as explained in the text (From G. L. Flynn and S. 11.
Yaliiowsky, J. Phzrm. Sci. 61, 839, 1972., reproduced with permission
of the copyright owner.)
Ciwpter is Diffucion and Dissolution 339
decreases and shifts, in the ease of the longer-chain
esters, to a concentration gradient in the diffusion
layers.
By using a well-characterized mcii:braie sach as
siloxane of known thickne and a homologous series of
0.09 cm
PABA esters, these workers 3 were able ta stud': the
various factor s : solubilit y . partition coefficient. cluuFV-
it y , diffusion lag time, and the effects of membrane and
diffusion-layer control. From such carfuhv desumed
and conducted studies, it is possrbie to p redict the roles
played by various tchvsicochemicci 1rc.:s us 11ev
relate to diffuston of drone through piascic coutai:.e:'s.
as the y influence release rates from sustained deii cry
forms, and as they influence absorption and excretion
processes for drugs dktt'ibuted in the body.
Lag Time under Difiusion-Layer Control i-
showed tim: toe la g time for ultra h:=:
under conditions of ci:ffusion-layer control may be
represented as
in which Eh, is the sum of the thicknesses of tim
aqueous diffusion i'ave.rs on the donor and receptor sides
of the membrane. The correspondence of 1,, equation
(13-60). for S y stems under membrane control (equation
[13-57]) is evident. The lag time for (luck membranes
o p erating under diffusion laver conncd a-
(13-Cl)
(L.., + 1102)D0
When the diffusion la y ers, ii, and )c. are o f same
thickness, the lag tlme reduces to
(13-62)
The partition coefficient, which was shown earlier to
be instrumental i n ceiverting the flux tro:n micniLrccne
to diffusion-lay er ccn:ro. new appears n -,-;c
tor of the lag-time equation. Atarge K signifi
lipophihcity of the penetrating drug snecies. As one
ascends a homologous series of P.PA esters. fir
time for stead y -state behavior; in other words, length-
ening of the ester molecule increases the isp time, once
the svstenl is in diffusion-la y er control. The sharp
increase in lag time for PABA esters with alkvi chain
length beyond C 4 is shown jr Figure lu- 16
Soluble Drugs in Topical Vehicles and Matrices. The
original Higuchi model 3 does not provide a fit to
experimental data when the drug has a significant
solubility in the tablet or ointment base. The model can
be extended to drug release front solid or
semisolid vehicies, however, using a q uadratic expres-
sion introduced by Bott,ari et
Q2 + 2DRA t Q _. 2DACt = 0 (13-63)
 340 Phpsicü PAar,eIcv
150
70
5 50
540
30
20
10
3 i 2 3 4 5 6 7
4umser or Carbons in Alkyl Chain (a)
Fig. (3-6. Charge in lag tcrne of n-amirrobenzoic acid esters ath
ails y i 'chain ler.h. From C. .. ?)vrin and S. H. 't'alkowskv, J.
?harm. So. ii. S&(, I'Y12. c-oriroducesl with permission, of toe
Copyright ow-ner. i
in which
= .4 - (C, - C 0) (13-64)
Q is the amount of drug released per unit area of the
dosage form, Dan effective diffusivity of the drug in the
vehicle, .4 the total concentration of drug, C, the
solubility of drug in the vehicle, C0 the concentration of
drug at the vehicle-barrier interlace, and R the di.f-
fusional resistance afforded by the barrier between the
donor vehicle and the receptor phase. A° is an effective
.4 as defined in equation (13-64) and is used when A is
onl y about three or four times greater than C,.
When
Q2 or. 2.DRA 5 Q (13-65)
equation (13-613) reduces to one form of the Higuchi
equation (equation (13-451):
Q = (2ADC,L) (13-65a)
Under these conditions, resistance to diffusion R is no
longer significant at the interlace between vehicle and
receptor phase. When C, is not negligible in relation to
.4, the vehicle-controlled model of Higuchi becomes
Q [D(2A - C,)C,t] (13-65b)
as derived earlier (p. 3.36).
The quadratic expression of Bottari (equation (13-
63 J) should allow one to determine diffusion of drugs in
ointment vehicles or homogeneous polymer matrices
when C, becomes significant in relation to A. The
approach of Bottari et aI.° follows.
Being a second-degree power series in Q, equation
(13-63) may be solved using the well-known quadratic
approach. One writes
aQ2 '1<? I) (13-661
in which, with reference to equation (13-63). a is unity,
b = 2.DRA, and c -2IJ.4'Ct. Equation (13-66) has
the well-known solution
-b a Vh - 4ac
(13-67)
or
2DR 4 + \ 2DRA + (WA°C,t) (363
in which the positive root is taken for physical signifi-
cance. If a lag time occurs, t in equation (13-68) is
replaced by (t - 1 ) for the stead y-state period. Bottari
et al. obtained satisfactory values for b and c b y use of
a least-square lit of equation (13-63) involving the
release of benzocaine from sus pension-type aqueous
gels. 1?, the diffusional resistance, is determined from
steady-state permeation, and C 0 is then obtained from
the expression
C. = R(dQJdt) (13-69)
The application of equation (13-63) is demonstrated in
the following example.
E.za.'np!e 73-9. (a) Calculate Q, the amount in milligrams of
micronized benzocaine relea.sed per cm-' of surface area from an
aqueous gel after 9000 seconds (2.5 hours) in a diffusion cell.
Assuming that the total concentration .1 iii 10.9 mg'mL, the solubility
C, is 1.31 rngmL, C, = 1.05 mg/mL, the difTusional resistance R of a
silicone rubber barrier, sepsratiiig the get from the donor compart-
ment. is 3.10 x 10 secicm. and the diffusivity Li of the drug in the gel
is 9.13 e 10- 6 cm'/sec. From equation (13-64).
= 10.9 nigmL-(1.31 ± 1.05) mgmL 9,72 rngimL
Then
DRA (9.14 X 10 cm t sec) a IS.10 s 10 secfcmi(9.72 mgcmL(
0. 71 Q 6 mgcmi
DAC,t (9.14 a 10')(9,72)(1.31)(0000) = 1.047 mgicmi
Q -0. 7 196 ((0.71963)2 20.047)P mgcm2
-0.7190 + (1.6161 090 mg'cnn
The Q,,, of 0.90 mg'cm° compares well with Q00. - 0.58 mg'cm2.
A slight increase in accuracy may be obtained by replacing ( 50)
sec with 1 (9000 - 405) sec, in which the tag time 1 305 sec is
obtained from a plot of experimental Q values versus Iii. This
correction yields a Q, = 0.87 mgcrn2.
(hi Calculate Q using equation 03_6.56) and compare the result
with that obtained in lal.
Q ((9.14 a 10 '(112 5 10.9) - 1.31J(l.31X9(X'J)l
1.49 mng.'cm2
Paul and coworkers have studied cases in which A,
the rnatrtx loading of drug per unit volume in a
polymeric dosage form, may be greater than, equal to,
or less than the equilibrium solubility C, of the d rug in
a matrix. The model is a refinement of the original

Higuchi approach,' providing an accurate set of
equations that describe release rates of drugs, fertiiiz-
era, pesticides, antioxidants, and preservatives in corn-
mrcia1 and industrial applications, over the entire
range of ratios of A to C.
A Capsule-Type Device. A silastic capsule, as depicted
in Figure 13-17a, has become a popular sustained and
controlled delivery form in pharmacy and inedi-
cinc.'° 42 The release of a drug from a silastic capsule is
shown schematicall y in Figure 13- 17b. The molecules
of the crystalline drug lying against the inside wall of
the capsule leave their cry stals, pass into the polymer
wali by a dissolution process, diffuse through the wail,
and pass into the li q uid diffusion layer and the medium
surrounding the capsule. The concentration differences
across the polymer wall cf thickness 1:,, and the
stagnant diffusion la y er of thickness )t are represented
by the lines, C - Cm and C 0 - C, respectively. C is
the solubilit y of the drug in the polymer and C,,, the
concentration at the polymer-solution interface, that is,
the concentration of drug in the pol ymer in contact with
the solution. C,, on the other hand, is the concentration
of the drug in the solution at the polymer- solution
interface, and it is seen in Figure 13-17b to be
somewhat below the solubility of drug in polymer at the
interface. There is a real difference between the
solubilitof the drug in the polymer and in the solution.
although both exist at the interlace. Finally, C 5 is the
concentration of the drug in the bulk solution surround-
ing the capsule.
To express the rate of drug release under sink
conditions, Chien 4 ° used the following expression:
r KDD,.
(13-70)
D,,,hO C,,t
Th,ia)
Drug
moicculet
54.-
Pu)yme' brrie' of
siOCtic caparile
Layer of ,.vI
drug crystals ' Stagnant diffusion layer
151
Fig. 13.. IT Diffusion of drug from silasuc capsule. to) Drug Er
capsule surrounded be polymer barrier, (h) diffusson of drug through
pol ymer will and stagnant aqueous diffusion layer and into the
receptor mpart.ment at atnk'c'ondiuon.a. (After Y. W. Chieri, in
SustoAned and COntr,SUed Rekai,e Drug Delivery Sytierms, J. R.
Rotinnoto, Ed., Marcal Dekker, New York, 1978, p. 29 Chem.
Pharno. BulL 24. 147, 1976,)
Chapter 13 fl,Tuown a nd Dsoe(iOion 1111
which is an integrated form analogous to c-oualion
(13-56). In equation (13-70i, Q is the amount of drug
released per unit surface area of the capsule, and K r is
the partition coefficient, defined as
I'), = C,/C, (13_71)*
\V'nen diffusion through the capsule membrane or film
is the limiting factor in dnit' release, that is, when
K j ,r D,,/i,,. equatior. (13-70) reduces to
Q = (-j C.t (13-72:
and when the limiting factor is passage through the
diffusion la y er (D,,,li,, 5>
c' "KrF]o'
= D t( = ----) C 7.,! (12-7-4
The right-hair) expression ma y be written because
C, = K rC 5 as defined earlier, in. ecuation (12-71).
The rate of drug release, Q . . for a polymer-ct.:-
trolied process may he calculated from the sl: . pe of a
linear Q-versus-( plot, and front (13-72) is
seen to equal C,Dih m . Likewise-, Q/t, for the diffusion-
layer-controlled process, resulting from ..luting Q
versus I, is found to be C 5D0 ih,,. Furthermore, a plot of
the release rate. Q1 versus C,, the solubil;tv of the drug
in the surrounding medium, should he linear with a
slo pe of D0h,,.
Example )3- 70. T'e pat-ulico cusff,cient. K. C, (,.
erons is sum; the solution diffusivitv D. 4.954 a 1 (1 en; > ( au
the siaos'.uc membrane a funojt3 D,,, = 4.26 a IC' crn eta-.-; ite
solubiiitv of progesternne in the siiastuc membrane. C 0 is 313 p cm-';
the thickness of the capsule membrane, ).,,, is 0.1,54 cr7,, and that r.f
s
the diffu ion las-er, Ii,, as estimated be Chier., is 0.0( 15 cm.
Caiculate the rate of release- of progesterone from the capsaic- .ra1
express it In agcm per da y . Compare the calculated resuS with the
observed value, Q.i " (u.3O iag.'cnu per da3. Using equation (13- 70%'
Ce-1.%J)>D,,
Qii =
br
013 jag/er5 0.Om)(4.i'tal a tO- cmt,dav
Q-t - ________ il-USC a t ' cm-ca-
.
(0.0225%l.55-1 :. crr-.-ay)i(,,(5e cur.
- (14.26 5: 10 — cnrdav)(0.005 err.)
0.03037 - -
Q'I 6ic34 eugcnl per cay
In the example just given, u) is KZ).); s D,,.h > or (b'
is L),,,1t 6 ' KD,,h,,, (c) What conclusion can be drawn
regarding matrix or diffusion-layer control?
• C.onLrarv to convention, the partiuor. roeft,ient K, in ecruacon
(13-71) is defined as the watrr.iip,d disuibuuor. roeflicient. Teas, K.
decreases here as the compo.irid becomes more hydrophobic. The K as
used previously, for exa-'ruie, in equations (13-501. (13-5-5), and
(13-61), is the lipid water distribution coeThcier,t such that K
increases with incroueour.g hvarophobidty of the conupoeizid. C. varies
with time and evcntoaflv equals C. Tnertfore, the equilibrium
partition coefficient does not involve ,, but rather, the equilibrium
ribses C, and C..

342 Plq.cv.zL Phannacy ' . -
= 8.79 x io n; D,,,h 1.14 x iO
D,juI(K,.Dh,. + DI,hJ
= (1.14 x 10')/t(8.79 x 10)
+ (1.14 x lo_J)J = 0.93
Therefore, D,h 0 a K,.Dh,,, and the system is 93%
under aqueous diffusion-la yer control. It should thus be
possible to use the sim plified equation (13-73):
= KO D 0 05 = ( 0.022)(4.994 x 10(513)
Qit
it., 0.008
= 70.45 Jag/cm2 per day
Although D/t, is larger than K,D,h,,, by about one
order of magnitude (i.e., D,,h 0 iKDh,,, = 13), it is
evident that a considerabl y better result is obtained by
using :he full expression. eauaion (13-70).
Example 73—/i. Two new contraceptive steroid esters, A and B.
.vere synthesi z ed. and the parameters determined for release from
polymeric capsules are`
D. 1J C. glcm
K. tcmday) )cm4av) eg!cmz) h,(cm) per day
A 0. L5 2.5 x IO 2.8 a :0_ n 100 0.008 24.50
B 0.04 4.0 x 10 3.0 a 10 - -' i5 0.008 10.32
Using equation (13-70) and the quantities in the table given,
cale'jateealuca of in cm far these capsule membranes.
-
- +
(Q t)(K.Dnh, - D,.h,) CK.D,D.,
• Q n<KJ),h,,) C,X,.DD,., - D,,h(Q.t)
h - C,X.D,D., - D,,h,(Q't)
--
(Q//)K.D,
For capsule .4:
(11)0)10. 1552.5 0 10)(2.6
2.6 a 10250.000024.50l
24.500. 1.5/(25 x I0)
0.0924
CM 0.101cm
N:0 that all unita cancel exce p t cm is the equation for h.,. The reader
should carry Out the calcuhitions (Or compound B. (Answer.' 0.097 cm.)
DIFFUSION PRINCIPLES IN BIOLOGIC SYSTEMS
Gastrointestinal Absorption of Drugs. Drugs pass
:hruug'n living membranes according to two main
classes of transport, passive antI active. Passive trans-
fer involves a simple diffusion driven by differences in
drug concentration on the two sides of the membrane.
In intestinal absorption, for example, the drug travels
in most cases by passive transport from a region of high
concentration in the gastrointestinal tract to a region of
low concentration in the systemic circulation.
Active transport requires an energy source such as
an enzyme or biochemical carrier to ferry the drug
across the membrane; transport can proceed from
regions of low concentration to regions of hi9/e concen-
tration through the "pumping action" of these biologic
transport systems. Other special mechanisms include
convective and ion-pair transport. We will make limited
use of specialized carrier systems, and will concentrate
attention rnarniy on passive diffusion.
Many drugs are weakly acidic or basic, and the ionic
character of the drug and the biologic compartments
and membranes have an important influence on the
transfer process. From the Henderson-Hasselbalch
relationship (p. 169) for a weak acid,
pH p110 +
log [HA)
in which [HA] is the concentration of the nonionized
weak acid and [A) is the concentration of its conjugate
base. For a weak base, the equation is (p. 170)
pH = prC5 + log [B]
in which [B] is the concentration of the base ind [B1-U]
that of its conjugate acid. pK0 is the dissociation
exponent for the weak acid in each case. For the weak
base, pK = pK, - pK.
The percent ionization of a weak acid is the ratio of
concentration of drug in the ionic form (1) to total
concentration of drug in ionic (1) and undissociated (U
form, multipled by 100:
7c Ionized 100 (13-74)
The Henderson- HasselbalcE equation for weak acid
therefore may be written as
= I0 - = antilog ( P Kn - pH) (13-75)
or
U = I antilog (p110 - pH) (13-76)
Substituting U into the equation for percent ionization
yields:
10))
% Ionized = (13-77)
antilog (pK., - pH)
Similarly, for a weak molecular base,
100
% Ionized = (13-78)
1 + antilog (pH - pk0)
In equation (13-77), p11,, refers to the weak acid, while
in (13-78), pK,, signifies the acid that is conjugate to the
weak base.
The percentage ionization at various pH values of a
weak acid, su]fisoxazole, pK,, M 5.0, is found in Table
13-4. At a point at which the pH is equal to the drug's
pK,,, equal amounts are present in the ionic and
molecular forms.

TABLE 13-4. Percent Su1.zuIe. p - 5.0. Oissoc:ated
and Undissociated at Various pH Values
PH Omociacea Unotisociated
2.0 0.100 99.900
.L.O 9.091
90.909
5.0 00000 50.000
5.0 90.909 9.091
99900 0.1CO
0.0 99.999 0.001
pH-Partition Hypothesis. Biologic membranes are
p redominantly lipoohilic. and drugs penetrate these
barriers mainl y in their molecular. undissociated form.
Brodie and his associates° were the first workers to
sooty the p rinci p le. known is the off-oanition hy-
oothesis, that drugs are absorbed from the rastrointes-
:uiai tract by passive diffusion depending on the
action of unciissociated drug at the pH of the intes-
tines. It is reasoned that the partition coefficient
between membranes and gastrointestinal Ouids is large
for the ur.dissociated drug species and favors transport
of the molecular form from the intestine through the
mucosal wall and into the s ystemic circulation.
The p H- partition principle has been tested in a large
number of in vitro and in vivo studies, and it has been
found to be only partly applicable in real biologic
SVStQMS.13,44 In many cases, the ionized as well as the
un-ionized form partitions into, and is appreciably
transported across, lipophilic membranes. It is found
for some drugs, such as sulfathiazole, that the in vitro
permeability coefficient for the ionized form may actu-
ally exceed that for the molecular form of the drug.
Transport of a drug by diffusion across a membrane
such as the gastrointestinal mucosa is represented by
Ficks law:
dv! DSK
- Tt = —v--- (C, - C) (13-79)
in which M is the amount of drug in the gut compart-
ment at time , D.., is diffusivity in the intestinal
membrane. S the area of the membrane, K the partition
coefficient between membrane and aqueous medium in
the intestine, it the membrane thickness, C2 the
concentration of drug in the intestinal compartment,
and C the drug concentration in the plasma compart-
ment at time I. The gut compartment is kept at a high
concentration and has a large volume relative to the
plasma compartment so as to make C a constant.
being relatively small, may be omitted. Equation
(13-79) then becomes
D,,,SK CA (13-80)
- =
The left-hand side of (13-80) is converted into concen-
tration units, C (mass/unit volume) x V (volume). On
the right-hand side of (13- . 80), the diffusion constant,
Chapter 13 . Oiffu..io* and D'is*,Autoii 3.
membrane area, partition coefficient, and membrane
thickness are combined to yield a perineabilthj coeffi-
c-.e,tt. These changes lead to a pair of equations:
dC7
-v -- = p 5c7 (13-81a)
dt
dC,
-v pc3 (13-81b)
dt
in which C. and P2 of equation (13-81a) are the
concentration and permeability coefficient. espec-
tiveiy, for drug passage from intestine to piasma. :
equation i13-31b), C3 and P are corres ponding terms
for the reverse passage of drug from plasma to
intestine. Since the gut volume V and gut concentration
C7 are constant, dividing (13-81) by (13-81b) yields
dC7 dt P3
13-32)
dt =
Equation (13-82) demonstrates that the ratio of a
sorption rates in the intestine-to-plasma and the
plasma-to-intestine directions equals the ratio of per-
meability coefficients.
In a study by Turner St al. , results show that
undissociated drugs pass freely through the intestinal
membrane in either direction by sim p le diffusion, in
agreement with the pH-partition principle. Drugs that
are partly ionized show an increased permeability ratio
indicating favored penetration from intestine to plasma.
Completely ionized drugs, either negatively or posi-
tively charged, show permeability ratios P 9JP of about
1.3, that is, a greater passage from gut to plasma than
from plasma to gut. This suggests that penetration of
ions is associated with sodium ion flux. Their forward
passage P is apparently due to a coupling of the ions
with sodium transport, which mechanism then ferries
the drug ions across the membrane, in conrlict with the
simple pH-partition hypothesis.
Colaizzi and Klinki& have investigated the pH-parti-
tion behavior of the tetracyclines, a class of drugs
having three separate pK., values, which complicates
the princi p les of pH-partition. The lipid solubility and
relative amounts of the ionic forms of a tetracycline at
physiologic pH may have a bearing on the biologic
activity of the various tetracycline analogs used in
clinical practice.
Modification at the pH-Partition Principle. Ho, Higiichi
and coworkers also have shown that the DH-partition
principle is only a p proximate, assuming as it does that
drugs are absorbed through the intestinal mucosa in the
riondissociaced form alone. Absorption of relatively
small ionic and nonionic species through the aqueous
pores and the aqueous diffusion layer in front of the
membrane must be considered. 47 Other complicating
factors, such as metabolism of the drug in the gas-
trointestinal membrane, absorption in micellar form,
and enterohepatic circulatory effects, must also be

334 Ph,sw.oJ P1numno.c
accounted for in any model that is proposed to reflect in
vivo processes. -
Ho, Higuchi. and their associates" have investigated
the gastrointestinal absorp tion of drugs using diffu-
sional principles and a knowledge of the physiologic
factors involved. They employed an in situ preparation,
as shown in Figure 13-18, known as the modified
Doluisio method for in situ rat intestinal absorption
(The original rat intestinal preparation' 6 employed two
syringes without the mechanical pumping modification.)
The model used for the absorption of a drug through
the mucosal membrane of the small intestine is shown in
Figure 13-19. The aqueous boundar-v layer is in series
with the biomembrane, which is composed of lipid
regions and aqueous pores in parallel. The final reser-
voir is a sink consisting of the blood. The flux of a drug
Permeating the mucosal membrane is
J = P (C6 - C) membrane wail to the systemic circulation (sink) is marked of!' from 0
(13-85)
or. since the blood reservoir is a sink, Ce
d aE 0;
= PaPp C1 (13-84)
in which P .pp
is the apparent permeability coefficient
(cm/see) and C 6 is the total drug concentration in bulk
solution in the lumen of the intestine.
The apparent permeabilit y coefficient is given by
PaPp ( 13-85)
=
in which P is the permeability coefficient of the drug
in the aqueous boundary layer (cm/sec). and P, is the
effective permeability coefficient for the drug in the
lipoidal and polar aqueous regions of the membrane
(cm/see).
The flux may be written in terms of drug concentra-
tion C 1, in the intestinal lumen by combining with it a
term for the volume, or
Iwo-directional pump
Rubber stopper
5-ml Plaroc syringe
17 /
,Ieju,wm
connected to
fleas Cannulan
Fig. 13-18. Modthed Doluiglo technique for no ama FE ioLuaDna1
Rosorpuon. (From N. F. H. IloetaL mGoneunJAtps
erDr-epu. A. .1. Aguia.r. Ed.. Annerwaru PrmaeeuucaJ Aauoatson.
Acadeyrn of Pharrazowtical Sciences. WasAnngson. D.C.. 19SL
repreucured with pnuwon of the cop'ngtnt owner.
Aqueoui____ Gut membrane I
Intestinal
lumen
DI L i
a
I
L 'I
Sink
Fig. 13-19. Model for the absorption of a drug through the onurona of
toe enrol] intestine. The rnnesuna) lumen is on the 1er.. foliowed by
stsZir aoroeoou diffusion layer, DL The gut membrane consists of
zoueotis p01-es a and Ijoida) regions I The distance (ruin the
to -L: the distance through the diffusion la y er is to L 1 . ( From N.
F. H. lie, W. I. Higuchi and J. Turi, J. Pharr- &r 51. 193. 1972,
reproduced with permission of the copyright owner.)
V dCi,
(13-86)
dt
in which S is the surface area and V is the volume of the
intestinal segment. The first-order disappearance rate
It,. (sec'') of the drug in the intestine is found in the
expression
dC6
zi--KC (13-87)
Substituting equation (13-87) into (13-86) gives
J = L. KuC 6 (13-88)
and from equations (13-84) and (13-85). together with
yields
= , 1
Pupt. K,
P uC P,,,
or
S P
= . i-(13-90)
Faq
14
Pm
Consideration of two oases, (1) aqueou s boundari' layer
control and (2) membrane control, results in simplifica-
tion of equation (13-90).
(1) When the Dermeabibty coefficient of the intestinal
membrane (i.e. , the veiocit't' of drug passage through
the membrane in cm/sec) is much greater than that of
the aqueous layer, the aqueous la yer will cause a slower
passage of the drug and become a rate-limiting barrier.
(The slower passage is alwa ys the rate-determining
process.) Therefore, P1 JP, will be much less than
unity, and equation (13-90) reduces to
K,haru.as = (S117)P1 (13-91)

now written as ic because the maximum
possible diusionaI rate constant is determined by
passage across the aqueous boundary layer.
2) If, on the other nand, the permeability of the
aqueous boundary layer :s much greater than that of
the membrane, P . P, will become much larger than
unity , and equation (13-90) reduces to
= (SIV)P,, (13-92)
The race-der.ermmuig ste p for transport of drug across
the membrane is now under membrane control. When
neither P., nor?, is much larger than the other, the
process is controlled by the rate of drug passage
through both the stationary aqueous layer and the
membrane. Figures 13-20 and 13-21 show the absorp-
:ion studies of n-ai.kanol and n-alkanoic acid homologs
at conciseiv Illustrate ;he hiooriysicai interoiav of pH,
oK,, solute Uoo p hilicity via carbon chain length, mem-
brane permeability of the lipid and aqueous pore
pathways, and p ermeability of the aqueous diffusion
layer as influenced by the hydrodynamics of the stirred
solution.
&ampie 13- IZ Cakulate the that-order rate constant, K,,, for
transport of an atiphatic alcohol across the inucosal membrane of the
at r.ailincesuaeifS1V I1.2 cm'.P.q - L. a 10 -1 =Wsec, and
LI x 10' cm,eec,
10 lace x L0)
K. (11.2) L5 11.2 (1.5
+ 1.5 x I0 cm,sec
1.1 X 10 cmxi sec
K. = 7.1 x 10 sec'
For a weak electrolytic drug, the absorption rate
constant K, is°°
K. = (13-9:3)
1
Pox:_
_ P1
30-
0
26 ship between equations (13-95) and (13-77) and
— / Qac;ilati*
/ 0.075 mI/sec
-V.
/ Static
14 W2tef
:j/ L /-
/i____4I50see
300 see
10 - small intestine actively secretes buffer species (princi-
2 4 6 8 10 12
if limber of carbons in n-alk.eimoe
Fig. 13-20. Firat-ormier absorption race com.wt for a series of
n-aka nols under i'anous hydrodynarnoc conditions(static or low
stirring rates and oshlation or high stirring of thuid at 0.075 mnLsec)
in the Jejunum. using the mnodiSed Doluisio cecnnjque. (From N. F.
H Ho, J. Y. Park, W. Morozomich and W. LHiguchl,inDertqnof
Bicphamac.ti.L Properties Through Prodo,9z and Analog:,
B. Roche, Ed, American Pharmaceutical Association, Academy of
Phaj-xr,aceuum.j Sciences, Washington. D.C., 1077, p. 148, repro-
duced with permission of the copyright owner.)
BSOIC ACidS
- I
c.
Hyarodymaxtics
Static. ISO see—;
2.o ^SCEIW104
0.075 mui;ac
Bulk pit
Fig. 13-21. First-order absorption race constants of aikanoic acids
versus buffered pH of the bulk solution in the i-at gut lurneis, using the
roodithed Doiuisio cecitruque. Hydrodynaniic conditions are mown as
;he d gure. Fr' rn N. F. H. Ho, J. Y. Pacit. V. Morowwicn and W. L
HIF'.lcru. i n
Design ii Bum zaccuzlcaL ?covertoes Througri ?co.
7ug3 ...4 .4.rmai&gs. E. 3. Roche. Ed.. A.mei-ican ?harrnaceuccar
Assocx.auon, Aoulerrrv if ?h.urrr.aceurical Sciences. Vanhington.
D.C.. 1977. p. 150, re produced with permission of t rie comryr±ghi
owner.)
in which P,,, of the membrane is now seoarated into a
term P 0 , the permeability coefficient of the liooidal
pathway for nondissociated drug, and P the perme-
ability coefficient of the polar or aqueous pathway for
both ionic and nonionic species.
= P ox0 # P (13-94)
The fraction of nondissociated drug species, X,, at the
pH of the membrane surface in the aqueous boundary is
1
_____
K. 1 + antilog(pH, - pK.1)
(13-95)
for weak acids, and
K. 1
- + K, = 1 + antiop - pH,)
(13-96)
for weal bases. The student should note the relation-
between (13-96) and (13-78). K, is the dissociation
constant of a weak acid or of the acid conjugate to a
weak base, and [H} 0 is the hydrogen ion concentration
at the membrane surface, where s stands for surface.
The surface pH, is not necessarily equal to the p H of the
buffered drug solution°' since the membrane of the
pally CO2 and HCO3 ). It is only at a pH of about 6.5
to 7.0 that the surface pH is equal to the buffered
solution p H. One readily recognizes that for nonelectro-
lytes, X, becomes unity, and also that for large
molecules such as steroids, P is insignificant.
1.48 x 101is
E.cample 73-13. Awealdyaddic drug ha ving lIm..
placed in the duodenum in a buffered anlution of pH 5.0. Assume
E 1H1. 1 0 10° in the duodenum. P. = 5.0 x 10' /sec, P.
1.14 X 10° cirusec. P = 2.4 X 10 .'riISec, and S/V 11.20 mi.
Calculate the abeOrption race constant. K,, using equation (13-93).
Fit-at, 2r om suation (13-94).


346 Pnyeic& PJT,iaCV

Slope 4.897 v 10' = ( F' 4 - 4.836 i

Calculate the absorption rate conatant. K. using equation (13-93:. 9.733 x lo-1 cmisec
P
First, fmin equation (13-94).
II 10-51 - (cI The ratio PiP 8 ,,.- - 2 The permeability of the un-ionized form
is seen to be about twice that of the ionized form. The reader should
S
0.403
1 a 10) 1.45 x 10' now be in a position to explain the result under (ci above based on the
Then p63-partition hypothesis.
5.0 x Percutaneous Absorption. Percutaneous penetration.
ot' X 10- that is, passage through the skit-.. involves (a) dissolu-
(1. 14 X 10)0 403 2 4 s 10 tion of a drug in its vehicle. (in diffusion of soiubilized
2.75 s 10' se-c drug (solute) from the vehicle to the surface of the sian.
and (c) penetration of the drug through the layers of the
Transcorneal Permeation, in gastrointestinal absorp- skirt, principally the stratum corneum. Figure 13-22
tion (Example 15-12) the permeability coefficient is shows the various structures of the skirt involved in
divided into P0 as the lip oidal pathway for undissociated percutaneous absorption. The slowest step in the
drug and P the polar pathway for both ionic and process usually involves passage through the stratum
nonionic species. in an analogous way P can be divided corneum; therefore, this is the rate that hmn.s or
for corneal penetration of a weah base into two controls the permeation.
permeation coefficients: P 5 .for the un-ionized species Scheupleiri 5° found that the average permeability
and P 5k - for its ionized conjugated acid. The following constant, P,, for water into skin was 1.0 x iU cm!nr
example demonstrates the use of these two permeabil- and the average diffusion constant. D,. was 2.8 x
ity coefficients. crn2/sec. (The subscript. s. on D stands for "skin."l
Mitt-a and Mikkelsor. 49 studied the transcorneál Water penetration into the stratum corneum appears to
permeation of pilocarpine using an in vitro rabbit alter the barrier only slightly, primarily by its effect on
corneal preparation clamped into a special diffusion cell. the pores of the skin. The stratum corneum is consid-
The perine.abthty (permeability coefficient) P as deter- ered to be a dense homogeneous film. Small polar
mined experimentally is given at various pH values in nonelectrolytes penetrate into the built of the stratum
Table 13-5. corrieuzn and bind strongly to its components: diffusion
Example 13- 14. al Compute the un-ionized fraction. J.
of of most substances through this barrier is quite slow.
pj)ocarpine at the p}i values found in Table 13-5. using equation Diffusion, for the most p art, is transcellular rather than
(13-961. The pR. of pOocarotne (actually' the p
F, of the conjugate
occurring through channels between cells or through
add of the weak base, pliocarpine, and known as the pliocarpuwlim sebaceous p ores and sweat ducts (F)g. 13-22. A ratner
Ion) is 6.67 at 34 C.
fbI The relationablo between the permeability P
and the un-ionized than B. C. and E). Stratum corneum, normal and even
given by
fraction j5 of piocarpine oase over this range of oil values is hydrated. is a most impermeable biologic membrane:
the equation this is one of its important features in living systems.
P - Pfp,jv (13-97) It is an oversimplification to assume that one route
where 'B stands for base and BH for its ionized or cOii)Ligot.e acid prevails under all conditions.° Yet after steady-state
form. Noting thatJ-BF - I - fe . eouation (13-97) car be written conditions have been established. transdermal diffusion
P P WI P)fs tTh-fiSi through the stratum corneurn most likely predomi-
Obtain the permeability for the protonated species F 8,- and the nates. in the ear) .,.. stages of penetration. diffusion
i
unchar g ed base P 5 us ng- eac:-500ares linear rerTesswn or, ecuatlor, through the appendages (hair folbcles. sebaceous and
in wruch P. toe tots! rerrneabilit: . . is toe dependent sanabve sweat ducts) may ne significant. These sIcanif pathways
and f, the incienenoeni variable are even important in steady-state diffusion in toe case
ici Obtain toe ratio of the two permeability coefficients. F'8-/F of large polar molecules. as noted in the following.
Answer-, Scheuplein et aL investigated the percutaneous
id
The c )culatedJ values are given at the various p63 values.
absorption of a number of steroids. They found that the
7.0,4 7 47 skin's main barrier to penet r ation by steroid molecules
ph oonor 4.67 5.67 6.2' 6.4C 6.67 6.9 was the stratum corneurrv. The diffusion coefficient, Dn,
solution
0.0) 0.09 0.27 0.39 USC 0.6' 0.70 0.8' for these compounds was approximately 10 - cEnt/sec.
several orders of magnitude smaller than for most
(III Upon linear regression. eauauon (13-98)
Devomes
nonelectrolytes. This small value of D. resulted in lou
P 4 . R2f c 16 - 4.897 x ermeabill...of the steroids. The addition of poiar
Intercept 4.98 s 10" groups to the steroid moiecuie reduced the di.flusior.
constant still more. For the polar steroids. sweat and
TABLE 13-5. Permeability Coefficients at Various pH Values sebaceous duct-s appeared to play a more important part

PH. 0000! 4.67 5.67 624 6.40 6.67 6.91 7.04 7.40
Solution AppToab1e smowsia of some drugs. such as steroids. may also
P a 10k 4.72 5.44 6.11 6.81 7.06 7.56 8.79 8.85 pm.eU'Ite the ikin thrmagb sebsceous ciucts ordinarily associated with
cnv bsirs on the ilaii surface (LrinSfolbCLiiai' aDsOrptiOn).

71:^Sebac
(ri\ gland LIS
lerv.
Fig. 13-22. Son structures involved in percutaneous absorption. Thickness of layers is not drawn to scale. Key to sites of percutaneous
penetration: A, transceilutaj- B, diffusion through channels between cells; C, through sebaceous ducts; i) tranafoilicular; E, through sweat duct&
in percutaneous absorption than diffusion through the
bulk stratum corneum.
The studies of Flynn. Higuchi, Ho, and coworkers
demonstrate the methods used to characterize the
permeability of different sections of the skin. Distinct
protein and lipid domains ap ocar to have a role in the
penetration of drugs into the stratum corneuni. The
uptake of a solute may depend on the characteristics of
the protein region, the li pid pathway, or a combination
of these two domains in the stratum corneum. and
depends on the lioophilicity of the solute. The lipid
content of the stratum corneum is imoortant in the
uptake of Lipophilie solutes but Is not involved In the
attraction of hydrophilic drugs.ss
The proper choice of vehicle is important in ensuring
bloavailability of topically applied drugs. Turi cc aL
studied the effect of solvents—propylene glycol in
water and polyoxypropyjene 15 stearyl ether in mineral
oil—on the penetration of diflorasone diacetate (a
steroid ester) Into the skin. The percutaneous flux of
the drug was observed to be reduced by the presence of
excess solvent in the base. O p timum solvent concentra-
Chapter :5 Dtffu.non and Ditso iuftm 347
Sb'aOm, comeum
Sweat
gland
Subcutaneous
Blood vessels
and pIneo
I and
jtissue
tions were determined for products containing both
0.05% and 0.1% diflorasone diacetate.
The important factors influencing the penetration of a
drug into the slthi are (1) concentration of dissolved
drug C,, since penetration rate is proportional to
concentration; (2) partition coefficient K between the
skin and vehicle, which is a measure of the relative
a ffi nity of the drug for skin and vehicle; and (3) diffusion
coefficients, which represent the resistance of drug
molecule movement through vehicle D 5 and skin D,
barriers. The relative magnitude of the two diffusion
coefficients, Dv and D,, determines whether release
from vehicle or passage through the skin is the
rate-limiting St€.sss
For dillorasone diacetate in propylene glycol–water
(a highly polar base) and in polyoxyoropylene 15 steary
ether in mineral oil (a nonpolar base), the skin was
found to be the rate-limiting barrier. The diffusional
equation for this system is
dC, - SKV,DSC
(13-99)
dt Vh

345 PeySic& Pavnw.i
in which C, is the concentration of dissolved drug in the
vehicle (gIcm); S is surface area of application (cm2);
K,, the skin-vehicle partition coefficient of diorasone
diacetate; D1 the diffusion coefficient of the drug in the
skin (cm 2/sec); V the volume of the drug product applied
(cm); and h the thickness of the skin barrier (cm).
The diffusion coefficient and skin harrier thickness
may be replaced by a resistance R to diffusion in the
skin:
R. = WD, (13-100)
and equation (13-99) becomes
dC - SK5,C
(13 101)
do VR,
In a percutaneous experimental procedure, Turi et
measured the drug in the receptor rather than in the
donor compartment of an in vitro diffusion apparatus
the barrier of which consisted of hairless mouse skin. At
steady-state penetration,
_Vt = VR (13-102)
The rate of loss of drug from the vehicle in the donor
compartment is equal to the rate of gain of drug in the
receptor compartment. With this change, equation
(13-101) is integrated to yield
(SKVC,.)
MR = l )t(13-103)
in which MR is the amount of difloraaone diacetate in
the receptor solution at time t. The flux. J, is
j == K,Cr J aSa
(13-104)
The steady-state flw for a 0.05% difiorasone diacetate
formulation containing various proportions (weight
fractions) of polyoxypropyiene 15 steary; ether in
mineral oil is shown in Figure 13-23. The skir—vemcie
partition coefficient was measured for each vehicle
forznu.iation. The points represent the experimental
values obtained with the diffusion ap paratus: the line
was calculated using equation (13-104). The point at 0
weight fraction of the ether cosolvent is due to low
solubility and slow dissolution rate of the drug in
mineral oil and may be disregarded. Be yond a critical
concentration. about 0.2 weight fraction of ooiyoxyprci-
pyiene 15 stearyl ether, penetration rate decreases.
The resuits ai indicated that one application of the
topical steroidal preparation per day was adequate and
that the 0.05% concentration was as effective as the
0.1% preparation.
£xamge 13- 15. A penetrzuon study o Sb x 10 giar
difiorasone diaceiste iiolution was conducted at 27 C in the diffusion
cell of 7w, e: a!, unuif a solvent of 0.4 wtilrh. fracuon of
polvoxonlene 35 si.wy l ether in miners) oil. The rtiUOn
£.. for Use drugdisuibuted between hairless mouse skin
• 0.2 1-
g
, 0.05
C
0.02 0 0.1 0.2 0.3 0.4 0.5 0.5
Polyorypropylene 15 slearyl eTher.
weight tractions. IS mineral oil
Pig. 13-23. Steady-state 15ax of difiorosonie diacetate in a mixture of
polvoxypropyiene 15 etesrate ether in mineral alt (From I. S. Thri.
D. Danielson and W. Woltersoni, J. Pharm. Sci. 65, 275, 1979,
reproduced with permission of the copyright owner.)
and vehicle was found to be 0.625. The resistance P, of the drug in
mouse akin was determined to be 6660 brim,. The diameter of a
circular section of mouse skin used as the barrier in the diffusion cell
was 1.35 ccii Calculate (c) the flux. I M 5 1(S . t). in g/m,i(i.. and
(5) the amoimtMp in ug of diDrasone diacetate that diffused through
the hairless mouse sloc in S hours.
Using equation (13-104)
(a)
l0.625X5.0 a 10 gicm',
6666 nr;crn
1 4.69 a 10 gicmtlnr
(b:
Mr (4.69 a 10 g/mMr) a (l.35cm)' (8 hr)
5.37 x 10 g 5.37 eg
Ostrenga and his associates 56 studied the nature and
composition of to p ical vehicles as they relate to the
transport of a drug through the skin. The varied D5,
K,,n . and C,. in order to improve skin penetration of two
topical steroids. fluocinonioe and fluocinotone acetonide,
incorporated into various propylene giyco—water gels.
In vivo penetration and in vitro diffusion using abdom-
inal skin removed at autopsy were studied. It was
concluded that clinical efficacy of topical steroids can be
estimated satisfactorily from in vitro data regarding
release, diffusion, and the physical chemical properties
Of drug and vehicle.
The diffusion. D,, of the drug in the skin barrier can
be influenced by components of the vehicle Imainiy
solvents and surfactants). and an optimum paruuor
coefficient may be obtained by altering the affinity of
the vehicle for the drug.
"The area S of a circle. expressed in diameter. is S - (±}wif. (See
inside front cover.'

The in vitro rate of skin penetration of the drug,
dQldt. at 25' C is obtained exoerunentaily at definite
times, and the cumulative amount penetrating (rnea-
sured in radioactive disintegrations oer minute) is
plotted against time in minutes or hours. Alter steady
state has been attained, the slo pe of the straight line
ieids the rate. dM/dt. The iag tune is obtained by
extrapolating the stead y-state Line to the time aus.
In vitro penetration of human cadaver skin' and in
vivo penetration of fluocinolone acetorude from pro y-
ene glycol gels into living skin are com pared in Figure
13-24. it is observed that the shapes and peaks of the
two curves are approximately similar. Thus, in vitro
studies using human skin sections should serve as a
rough guide to the t'orrnulation of acce ptable bases for
these seeroidal compounds.
Ostrenga et ai. were able to show a relationship
between release of the steroid from its vehicle, in vitro
penetration through human skin obtained at autopsy,
and in vivo vasoconstrictor activity of the drug deuend-
ng on compositions of the vehicle. The correlations
t
obtained suggest hat information obtained from diffu-
sion studies can assist in the design of effective topical
dosage forms. Some useful guidelines are (a) all the
drug should be in solution in the vehicle, (b) the solvent
mixtures must maintain a favorable partition coefficient
so that the drug is soluble in the vehicle and yet has a
great affinity for the skin barrier into which it pene-
trates, and (c) the components of the vehicle should
'lao
0 70
x and polyethylene glycol 400, a constant diffusivity was
60.e
E The partition coefficient, K,, for the paraben be-
50 3'
40;
'-I
C
30 ".
20
10
0 •— '0
0 20 40 60 80 100
Propylene glycol, %
Fit- 13-24. Comparison of in vitro penetration of steroid through a
sluri aecuon and in vivo inn bLanching tesL Key: s , in vitro method:
C. in vivo method. (From J. 0trena, C Steinmetz and D. Poiiiaert,
J. Pharm. Sci. 60, IITT, 1971. reproduced with permission of the
copyright owner.)
Chapter ir Otffston and DIUOd'.thO,I :14e
favorably influence the permeability of the stratum
corneam.
Sloan and coworkers" studied the effect of vehicles
having a range of soiubility parameters, S isee pp.
4-223), on the diffusion of saiicvlic acid and theo p hyt-
Line through hairless mouse skin. The, were able to
correlate the partition coefficient K for the drugs
between the vehicie and skin calculated from soiubthty
parameters (see P"oblem 21 3-16. ru. :359) and the
Permeability coefficient P. obtained exnenmentailv
from the diffusion data. The results obtained with
salicylic acid, a soluble molecule, and with theophylline.
a poorly soluble molecule with q uite different physical
chemical properties, were practically the same.
In the studies of skin permeation described thus far,
efforts have been mane to increase percutaneous ab-
sorotiorl processes. it is important, rowever, that some
compounds not be absorbed. Pharmaceutical adjuvants
such as antimicrobial agents, antioxidants. coloring
agents, and drug solubilizers, although ideally they
should remain in the vehicle on the skin's surface, can
penetrate the stratum corneum.
?arabens, typical preservatives incoroorated into
cosmetics and topical dosage forms, may cause allergic
reactions if absorbed into the derails. Komatsu and
SuzuI6 58 studied the in vitro percutaneous absorption
of butylparaben (butyl p-hydroxybenzoate) through
guinea-pig skin. Disks of dorsal skin were placed in a
diffusion cell between a donor and receptor chamber,
and the penetration of '°C-butylparaberi was deter-
mined by the fractional collection of samples from the
cell's receptor side and measurement of radioactivity in
a liquid scintillation counter.
When butylparaben was incorporated into various
vehicles containing polysoroate 80, propylene glycol,
obtained averaging 3.63(0.47 S.D.) x 10'' cm. °/hr.
tween vehicle and skin changed markedly depending
upon the vehicle. For a 0.015% (w/u) aqueous solution or
butvl paraben. K,., was found to be 2.77. For a 0.1%
(wiv) solution of the preservative containing 2% (iv,v) of
poiysorbate 80 and 10% (oly) propylene glycol in water,
the partition coefficient drooped to 0. 18, There was no
apparent complexation between these solubilizers and
butylparaben, according to the authors.
The addition of either propylene glycol or polyethyl-
ene glycol 400 to water was found to increase the
solubiiity of paraben in the vehicle and to reduce its
partition coefficient between vehicle and skin- By this
means, skin penetration of butylraraben could be
retarded, maintaining the preservative in the topical
vehicle where it was desired.
t
In the case of polysorbate 80. Koma su and SUZukiss
found that this surfactant, too, reduced preservative
absorption, maintaining the antibacterial action of the
paraben in the vehicle. These workers concluded that
the action of polysorbate SO was a balance of complex
ass Phnco.L Phoi'noc
factors, which is difficult for the product formulator to
predict and manage.
Buccal Absorption. Using a wide range of organic
acids and bases as drug models, Beckett and Moffat59
studied the penetration of drugs into the li pid mem-
brane of the mouths of humans. in harmon y with the
pH-partition hypothesis, absorption was related to the
pK of the compound and its lipid-water partition
cocient.
o and Higuchi° applied one of the earlier mass
6insfer models"' to the anal y sis of the buccal absorp-
tion of n.-alkanoic acids. 62 They utilized the aqueous-
lipidphie model in which the weak add species are
transported across the aqueous diffusion la yer and,
subsequently. only the nanionized species pass across
the lipid membrane. Unlike the intestinal membrane,
the buccal membrane does not appear ti possess
significant aqueous pore oathways. and the surface pH
is essentially the same as the buffered drug solution
pH. Buccal absorption is assumed to be a first-order
process owing to the nonaccumu.Iation of drug on the
blood side.
= -Kt (13-105)
Cc
in which C is the aqueous concentration of the nra]-
kanoic acid in the donor or rnucosal compartment. The
absorption rate constant, K,, is
Ku _s.
1, 2C (13-106)
1 + ----
the terms of which have been previously defined. Recall
1/(1 - 10 - 1)1) or by e quation (13-95) is
that X.
1'(1 antilog (pH ,. - pK)) and is the fraction of
un-ionized weak acid at pH
ibbit. (Froni T. c Panapi itaL.J. Pcarni. Sc. 647119Th. e,ced
Fig. U-25 I pam,d rib-cape eeC in vaguii1 trs
of the Co Tip'bi owner.)
With S = 100 cm 2, V = 25 cm', P,,0 = 1.73 x
cmisec. P5 2.27 x 10 cmlsec, pK 4.84, and
pH, 4.0. equations (13-95) and (13-106) yield for
caproic acid an absorption rate constant:
100/ 1.73
K - X 10
1.73 x
2.27 x 10r 3 x 0.874'
3,7 x 10 sec-1
Buccal absorption rate constants constructed accoring
to the mode! of Ho and Higuchi agreed well with
exoei-imental values. The study shows an excellent
correspondence between diffusional theory and in viva
absorption and suggests a fruitful approach for struc-
ture-activitY studies. not only for buccal membrane
permeation but for bioabsorption in genera..
Uterine Diffusion. Drugs such as progesterone and
other therapeutic and contraceptive compounds may be
delivered in microgram amounts into the uterus by
means of diffusion-controlled forms (intrauterine de-
vice. IUD). In this way the patient is automatically and
continuously provided medication or protected from
pregnancy for days, weeks, or months'3
Ho, Fl ynn. Higuchi and their associates' performed
in situ vaginal drug absorption studies using the rabbit
doe as an animal model develop more effective
uterine drug delivery systems. A solution of a model
drug was perfused through a speciall y constructed cell
and imp lanted in the vagina of the doe (Fig. 13-25), and
the drug disappearance was monitored. The drug
release followed first-order kinetics, and the results
permitted the calculation of apparem permeability
coefficient and diffusion laver thickness.
The drug may also be implanted in the vagina in a
silicone n'iatriz (Fig. 13-26) and drug release at any
time can be calculated using a quadratic expression.'
with p1mwon


Ch.e.r 1$ Diffsirion and D,a.oAon 351

( I I
oiiac mauix
Vagina' 7 104 cmisee 7 x 10-4cm,sec
membrane = 0.0587 (dimensionless)

Lipid
= -2mhD,C, = - 2(3.1416)(6 cm)
water x (4.5 x 10 crn2/secg0.572 mgicm)
layer
-...- Aqueous poet
< 86,400 sec (t days) = -0.8384 )< I
day
How much progesterone is released in5 days?
• Blood side days? The quadratic formula tobeusd here is

W=_ii+V&Z_I
I isaue
N._ conc collation A.'
.
•- 5_ --4 ii - After 5 da y s,
Receding c = -0.8384 mgiday x 5 clays = -4.1920 ".mg,
drug Boundary
Fig. 13-26. Contraceptive drug trio wacer-insoluble silicone polymer and
rnatri.x. Dimensions and sections of the matrix are shown Logether V(0.105s2_ -(4)(0.000433)(-4.1920)
with concentration gradients across the drug release pathway. iFrom M-
S. Hwang et ii.. i. Pharrn. Sci. iS. 1573. 976. reproduced with 2(0.000438)
permission of the co pyright owner.)
=51.6mg
I \ Alter 20 days, c = -0.8384 mg/day x 20 days =
M -16.77 mg.
2mha0 2 A) O Pnq
- (21rh.D 1C34) 0 (13- L07) - -0.0587+V(0.058'7)2- ( 4)(0.000438)(- L6.77)
- -
2(0.000438)
The method of calculation may be shown, using the data
of Hwang et a1., which are found in Table 13-5. 139.8 mg
When the aqueous diffusion layer, )t,., is 100 p.m, the
4 Okada et aL have carried out detailed studies on the
aqueous permeability coefficient P is 7 x 10 cm/see; vaginal absorption of hormones, as reported in four
this value is used in the following example. The length
papers.
h of the siiastic cyclinder (see Fig. 13-26) is 6 era, its
radius, a,,, is Li cm, and the initial amount of drug per
unit volume of plastic cylinder, or loading concentra-
tion, A, is 50 mg/cm3. THERMODYNAMICS OF DIFFUSION
Equation (13 . - UYT) is of the quadratic form, aM 2 +
Permeation of gases, Liquids, and solutes through
bM ± c = 0, in which, for progesterone, membranes requires an energy of activation (p. 295) for
the small molecules to move through the matrix of the
barrier material. This fact is expressed in the Arrhe-
ruus equation:"
(2)(3.1416)(6 crn)(Li cm)(50 mg/cm3) P = Pae 6u' ' . ( 13-108)

= 0.000438 mg' lnP = lnPcs - EpJRT (13-109)

D.K. (1 1) = (45 x cm2/sea)(50.2) in which P0 is a factor independent of temperature and

a0 \F5 P LI cm proportional to the number of molecules entering the

TABLE 13-5. physical Parameters for the Relaas. at Pregestvea. and H)*scoi'1sem hem
a Silicone Matrix for Vaginal Absorption in the Rabbit "°
Progestefone Hydrocortisone

Solubility in matrix, C, (mg/cm") 0.572 0-014

Diffusion coefficient in matrix. O (cm 2fspc) 4.5 x 10 4.5 a 10
Silicone-water partition coefficient. K. 50.2 0.05
6nnneaOility, coefficient of rabbit vaginal membrane, P,, (crrvsec) 7 a 10' 5.8 a IO
P. when h, - 100 m) 7 a 10' 7 a 10'
P. 1when h,, - L000 i.m) 0.7 a 10' 0.7 )i10


352 Fiztcai P&arMaCjI

I D.enstly

0.955
N
E N
N.
10 " r
0.967

I I

3.0 3.1 312 3.3 3.4 3.5 3.6

lIT ) 10 (Temperature, K)

Fig. 13-2'L /rrhernus plot for permeability constaras of nitrogen permeating polyethylene films olvarious density. (From A. S. Michaels and
R S. Parser. J Poim. Sc. 41. St. 1855.)

film and to the probability that these molecuies have
sufficient energy to engage in the diffusion process. E,
is the energy of activation for permeation in calories
mole, and R and T have their usual meaning.
Michaels and Parker` -' determined the permeability
of oxygen, nitroger., and other gases in polyethylene
films of different densities. The Arrhemus plot (Fig.
1;2-27. demonstrates the constant energy of activation
and corresponding decrease in permeability of nitrogen
a density of the poIetnyiene film is increasec h';
greater short-chain branc±nrig of the polymer structuro.
E, values vary from ö 1.0 20 kcalimoie for permeation
through li q uids and polymeric barriers, respectively.
energies have lower permeability constants, as might
be expected. The slopes of the lines in Figure 13-25
may be used to calculate the activation energies with
the aid of equation (13-109). The uppermost line has a
break at 14.2 C. the penetration of octaiiol appearing to
show a higher activation energy at lo'ser temperatures.
Detanol. 1, = 10 enilimole
T
E 7f, -
Pentanol, 165.5 kcalimiile

Higuchi and Agwar ea obtained a value of S kcalimoie for

the permeation of water vapor into wax-like enteric
tablet coatings.
Activation energies have also beer obtained in toe
study of the diffusion of liquids across biologic mem-
branes. Blank et aL° showed that the apparent activa-
tion energies E5 for permeation through human skin 0 3.2 3.3 3.4 3.5 3.6
taut.opsy specimens) by lower alcohols (ethanol to
(1/Ti a 10'(1em perWre. l
p entanol) and higher or less polar alcohols (hexanol to
oct.anol) are 16.5 kcalimoie and 10 kcal/mole. respec- Fir. 13-26 Arrbemus plot (Tog penneab&liiy consu.n versu'
reciproca absolute temperature) for difiusion of normal Slsanot5
tivel y . The less polar alcohols penetrate the skin more through autopnv specimens of human s1r, (From 1. H. hank, R. i
rapidly and exhibit permeability similar to ethers and SclieupIein and D. J. MeFarlane. J. bive. DerinaLct. 41. 5. 196
ketones. The polar compounds with higher activation C 1967. The Williams & Wjlna Co., Baiumorn.i

that is. between 4.6 C (VT = 3.60 x 10) and 14.2 C
(l)T = 3.48 x 10K). Temperature dependency of
oermeacion can provide valuable information about the
nature of the barrier and the mechanism of transport.
Autian° reviewed the importance of p lastics in
pharmac y and dirucal p ractice and treated the theory of
diffusion and sorp tion of drugs and environmental
vapors in polymeric packaging materials. See the
chapter by Autian in L}ispenring of Medicuion t° for
refe
rence to diffusion in ia.stic materials as it ap p lies to
clinical pharmacy.
FICK'S SECOND LAW
Fick' rst .aw e q uation 13-2), p . 225) has been
used througnout this cnaoter as a starting point in the
development or equations to describe the diffusion of
arugs through natural and poiyrneric membranes. How-
ever, there are man y diffusion problems in which the
.rst law of Fick is not app iicable, and the second law
iequation (13-5), pp. 325. 326),
- D°'2
edt 3.i
must be used. Here we use Is instead of C to express
concentration. The symbol a signals that partial den y
are being used since u is a function of both t and-anves
z. The second law is used to express diffusion in cyl-
inders and spheres, as well as through flat plates. The
simplest form of the second law diffusion equation is
(13-110)
for symmetrical diffusion outward from the axis of a
cylinder, in which r is the radius of the cylinder.
For diffusion proceeding symmetrically about the
center of a sphere of radius r, the partial differential
eq uation, representing Ficks second law in its simplest
form. Is
(13-111)
at si-. 01'
The equations for diffusion in cylinders and spheres are
discussed by Crank" and by Jacobs.
Although the derivation of equations based on Fick's
second law is in most cases beyond the mathematical
scope of this book, it is of value to present some
equations and obtain their solutions. Such exercises
give the student practice in caicu::ions for diffusion
problems that axe more complicated than those derived
from Fick's first law.
Diffusion in a Closed System. Determination of D. A
simple apparatus (Fig. 13-29) was used by Graham
(1861), one of the pioneers in diffusion studies, to obtain
diffusion coefficients D for soiutes in various solvents.
C5apLm ii Diffusso,i and DzaoLyhon 353
-I
=0
'g.3-29. Stmole apoa.i-cuo rrasn tr cony L:sLon uaiel.
:F-um 4. H. ,acoos. 2iffuno't P cesoe, Springer- :rc New
'fbrk 19716. p. lt )
The coefficients for some solutes diffusing through
various media are listed in Table 13-1. In the a ppara-
tus depicted in Figure 13-29, the height of the solution
is h, the combined height of solution and solvent is H,
and the distance traversed by the solute is .r. The
concentration of solute at a position a and time z in the
solution is is and its initial concentration is u,,. From the
ex perimental values of u, x, and t, it is possible to
determine the diffusion coefficient D for the solute in
the solvent.
Initially— that is, at time t = 0 sec—the concentra-
tion u is equal to is0 (moles or grams per em') in the cell
Tom position r = 0 to r = h (cm), and is = 0 from x=
It to z = H. These statements are known as initiat
conditions. In a case in which It is taken equal to H12,
that is, both solution and solvent are of equal volume,
the equation for u ls
U, 2u,
[COS
-
3 H i (13-112)
Equation (13-112) is simplified if we choose x, the
position of sampling in the cell, to be H16; the second
cosine term in the parenthesis of equation (13-112)
becomes cos(-,,,;2) = cos 90 = zero. This leaves only the
Prat cosine term, cos(ni6) = cos 30 = 0.866. Thus,
taking z = H/ti, we have
is = - _f2 0.866 e (13-112)
The reader is reminded that with trigonometric fur-
lions such as cos(lr/61, p1 is given in degrees; that is,
= 1800 and i6 = 30*, whereas in a term such as 2 u,Jr
or the value of ir is 3.14159

354 Pflysw.oJ Phcr7rwci.
Example 13- 16. A new watir oub)e drug. coraz.ole, is placed in
it Graham dithaaion call (see Fig. 13-29) at an mjtiai concentration.
0.030 mmoleicrii 5 to determine its diffusion coefficient in
a.25C. The height ofthe soicxuonhm 2.82 =. and the total
height of aqueous solution and overlying water is if 5.64 cm. A
sample is taken at a depth of r - 1:1/6 an attune f - 4.3 hours (15481)
see) and is found by spectrophotOmeuc analysis to have a concentra-
tion it - 0.025 minole/ait 3 . Li is obtained by rearranging equation
(13- 313t
11
Li thç0866(2 n)))
i,i
(-3L8096)
- °79313(98696X15480)
- 16.41 x 30 cn2/sec
Diffusion in Systems with One Open Boundary. The
Graham cell for the determination of diffusion coeffi-
cients is an example of a closed system. in pharmaceu-
tics, physiology, and biochemistry, systems with one or
two open boundaries are of more interest than the
closed-boundary system. In 1850 Graham introduced a
sy stem with one open and one closed boundary, as
shown in Figure 13-30. Insignificant mixing occurs
between the solution and the water because of differ-
ences in density. The condir.ion at the interface between
the solution and the water layer, mown as a boundary
condition, is expressed as "IL = o when z h." A
second boundary condition states that the change in
concentration ii with the change in position z is zero, or
in mathematical notation, (u./3z) 0. This occurs at
the bottom of ti- cell. for the solute cannot pass out
through the bottom. In addition to the two boundary
conditions. it is useful to specify an initial condition, as
was done for the closed cell treated earlier. The imtlal
condition is often taken as uniformity of concentration
within the solution in the inner vessel of the cell: that is.
= tz at I = ' O.
For a system with one open and one closed surface
and in which the amount. X,, of solute escaping
- iC
Fig. 13-30. Diffusion apparains witS one opes and one 6064d
boundary . (m . H. Jacob€. DijMaiOii Prvcwu. Spnnper-Vrriag.
New 'iii. 1976. p. 47.)
between time 0 and time f is expressed by the
equation7
=u,,4 h[1 - +.
(13-114)
in which A is the cross-sectional area of the inner cell of
height h (see Fig. 13-30), and the other terms have
been defined in connection with equations (13-112) and
(13-113).
£ximpic 13- 17. Caicuiate the tots) amount 1d rn , of the new drug
corasole that escapes between limes 1 0 and 1 2.70 hours (9720
see) from the ccli with one open boundary (see Fig. 13-30). The area
A of the cell is 8.27 cir 2 . and its height is h - 2.65 cm. The ongmal
concentration of the drug in the eel is u, - 0.0437 g/em5. The total
in multiplied
amount at drug M in the cc11 is the concentraion
by A S h. the volume of toe cell 0.0437 glcci' x 8.27 eni 5 2.65
cm - 0.9577 g. Toe diffusion coefficient Li of the ortig corasole in
water at 25 Civ 16.5 x )0- 1 cra t /ser, as found in Example 13-16.
Inserting these raises into equation (13-114) yields
(0.0487 g,cir 3 x 8.27 cm' x 2.65 an) S
lries,ui..eXS7Z
-
ç_ -
15 i cvuieein9I a,,
4 i. cv' -I. .
- 0.9511 grain a 0.53805 0.51520 gram.
Thus we arrive at the result that in a cell containing 0.0437 glae 3 or
(1.9577 g of total drug. 8.5155 g diffuses out in 2.7 hours
The diffusion of macromolecules such as proteins, is discussed in
the chapter on colloids.
DIFFUSION AND ECOLOGY
Diffusions] processes have wide application not only
in physical and chemical sciences but also in biology, in
which living systems. such as animal colonies, are
subject to diffusion as well as aggregation or consolida-
tion. Diffusion is a random process by which atoms and
moiecules. colloida) and coarse particles, and even
iivi'w meinbery o( popuLo2toss diffuse or spread out
with time. in the booh Diffu..saon and Ecoi.opw&
P'roble,n.s. Okubo 73 considers such diverse subjects as
diffusion and dispersal of spores by the winds, distri-
bution of fish eggs in the sea, migration of turtles,
diffusion factors in the homing of birds, insect swarm-
ing and fish schooling, and animal movements on their
home ranges. An important means of communication
between animals is through the release of olfactor y or
gustatory signals and their transmission through the
environment. These processes may be treated in terms
of passive diffusion models.
In Figure 13-31, Okubo applies the principles of
diffusion to the ventilation of animal burrows. The main
opening of the tunnel of the prairie dog is on a mound,
and the emergency et at a lower level. This design
apparently provides the most efficient diffusion of air

Fig. 3-31. Ventilation of a p rairie dogs burrow by action of the wind, ;From A..Okubo. Drffu.non and 3cologwal Problems: Mathemaiiri.ol
Models. Springer-Verlag, Berlin, 1980. p. 43.)
and ventilation for the burrow of the prairie dog. The
difference in the height of the two openings produces a
"viscous suddng" by the wind on the stagnant air in the
burrow, according to Bernoulli's law. This is the
princi p le of the Venturi tube, used to expei liquid
medication from a spray atomizer. Diffusion theor y has
been used by Wilson and Kilgore 4 to calculate the
concentration gradient of res p iratory gases in small
animal burrows. The interested student should see
Okubo° for other exam p les of dLfusion in bioloeic
systems.
References and Notes
I. S. T. Hwang and K. K.amrnerrmeyer, Yrmhmnes in Seronons
(Techniques
Inter-science, of ChemtolrlJ). A. Weisaberger, Ed., Wiley-
New Yore. 1975, pp. 32-23.
2. V. T. Stanneti st at. Recent advances in membrane science and
tecoriology, .n .Advances in Polymer Science, Voi. .32, Springer.
Verlag Berlin, 1979, pp. 71- 1322; D. Paul and C. Morel,
Membrane technology, in Encyclopedia of Ckemicai T.ichnoloqy,
3rd Kit. Kirk & Othmer. Edo., Vol. 15, Wiley, New York. 1981.
pp. 92-131,
1 G. L. Flynn, S. Ii. Yoikowsicy and T. J. Roseman, J. Ph.arm. Sci.
53, 479. 1974.
4. H. Carsiaw, ,l(a4he,n5tc1 Then,'j of the Conduci'.on 0/ Heat.
Macriulian, New York, 1921.
5. R. W. Baker and H. K. Lansdale. in ControlLed ReLease or
&oioqiatlly Active Sprain. A. C. Tanqusry and R. K. Lacey,
Ed,a., Plenum Press, New York. 1974. pp. 30-32.
ii, M. C. Karth, W. I. Higuchi and J. L. Fox. J. Phann. Sd. 74. 612,
1985.
7. M. Z. Biber and C. T. Rhodes, J. Pharm, Sri. 55. 564, 1975.
Chapcer !S . Diffunon and Disaotithon 385
Sir stTearn
rim burrow
8. K. Nasim. N. C. Meyer and J. Autian. J. Pharm. Sd. 51, 1775,
9. T. H.igucni and A. Agusar. J. Am. Pharm. -Assoc., Sri. Ed. 48.
574, 1969-
10.J. Crank and G. S. Park. Dsd'unio,t in Poiyrnero. Academic
Press, New York. 1968, pp. 20-32.
ii: R. Srheuviein, J. Invest. DermatoL. 45, 324, 1965.
12. D. K. Wurater. J. A., Ciscreriga and L. K. Matheson. Jr.. 2.
Pharrn. Sri. 58, 1406, 1410, 1979.
13. Y. C. Martin, Qwinstlat ire Drug Design. Marcel Dekker. New
York, 1978. pp. 76-81.
14. W. .3. Addicks. C. L. Fivn.n and N.'Weiner. Pharm. Res. 4. 517,
1967.
15. C, M. Grass and S. A. Swbetar.a. Phartit. Res. 5. 372, 1988.
16. W.J. Addicks, G. L. Flynn- N. Weiner ar.d C-3d. Chiang, Pharm.
Boo. 5, 377, 1985.
17. D. K. Wurst,er and P. W. Tayior. J. ?harm. Sri. 54. 169. 1985.
18.J. C. Wagner. B opkas-maeutics and Relevond Pacoki*eO-
Ci, Drug Intelligence, Hamilton. lU. 1 971, pp. 98-.47.
19. L. J. Lesson and J. T. Cursterisen. Dihsalunon Teckiotop,
Academy of Pharmaceutical Sciences. Washington. D.C., 1974.
20. W. 1. Higuchi. J. Pharm. Sri. 56, 315, 1967.
21. J. P. Skelley, C. L. Amidoa, W. H. Barr, eiaL,,J. Pharm. Sc'. 'N-
849. l99L
22.J. McGinity, S. Stavehansky and A. Martin. Biovai1abliity in
tablet technology, in P5arr'soce'.,cical Do.sage Farms: Tablets. H.
A. Lieberman and L. Lachnsan, Kits. Vol. II. Marcel Dekkar,
New York, 1981.
El. M. K. T. 'L"au and M. C. Me yer. 3. Pharm. Sri. 70, 1017, 1981; V.
P Shah. V. K. Pra.sad. T. Aistori, B. K. Cabana. B. P. Gui-al. and
M. C. Meyer. J. Pharm. S,± 72, 306, 1983; P.2. McNama.ra, T. S.
Footer, C. A. Digerus. K. B. Pate!, W. A. Craig, P. C. Welling,
B.S. P.apaxs. V. K. Pninaci. and V. P. Shah. Ph.arm, Boo. 4. lEO.
1997.
24. C. Levy and B. A. Sahli. J. ?harm. S± 31.58. 1962J. H. Wood,
J. K. Syarto and H. Letterman. 3. Pharoi.Sd. 54, 1068, 1965: J.
H. do Smidt, el al. J. Pharm. Sri. 75, 497, 1986.
356 Pnpoirai PflW7IIACY
Solid Dosa'r
25 3 T Car 1.enaeri. Pharno,ceaLWo a' Sotui and
F(,in.c. Wile y . New York. 1977. Cnapler 111.
26. A Htxson and J. Crowell. Ind. Enç. Chem. 23. 921, 3931.
27 51 J. Miralle. Dissolution rates of Lo)but.a3Thdt coprec2pI1.aLe.
M. S. Thesis. University of Texas. August. 1980.
20 V L€vjch, Phvaicocae?flW& Jrodynamw.O. Prentice-Hail,
Englewood CLiffs, N.J.. 1982
29 K C. Nelson and A. C. Shah. J. Phartt. Sd 64, 610. 1516, 3975.
Crommelin. J.
30. J. H. de Smid. J. C. A. Offrings and D. J. 61 A. Suzuki, W. 1. Higuchi and N. F. H. No. 3. Pi,amnv. Sci. 59.644.
Pharm. Sr- 76, 711, 1957.
31. 3 K Robinson. Susootnca and Cont,'oUe4 ReLease Drug De1weay
SvsUrr'u. 1,{arcel Dekker. New 'lark.. 1976.
12. T Hirjciu. J . Soc. (loam. Chain. 11.30, 31460.
Higuchi. J. Pnarrn Sc'. U. 874. 1961.
34. lof S. J. Dexal, A.. P. Simone lli and W. 1. Higuclu. 3. Pharro. Sc,
54. 1459. 1965: (5) S. J. Des.a, P. Singh. A. P. Sunoneth and W.
1. Higuchi, J. Pnarm. Sci. 5.5. 14, 1230, 1235, 11466.
33 N. F. H. Ho and T. J. Rosemar.. J. Phartn. S. 65. 1170, 1979,
IC. C. L. Fl ynn. S. H. Ya1kowscy and T. J. Roseman3. Pharm. Sc.
63, 4714, 1974
31 C L Fiynr. and S. H. maIkoWsiy. J. Pnarrn. Sc: 61. 305, 1972
F. Bottac:. G. DiCoic'. K. Nampier. et a... 3. Pharm. Sc. €3,
1779. 197t )bid 60. 1477. 1977.
314. B. R. Paul and S. K. McSpadden.J. Mernb. Sc. 1.33. 1f' 76: S K
Chandrasekarori and D. K. Paul. .3. Pharm. Sc, 71, 13149. 3902.
40. V. W. Chien.
Svateon.s. J. 11. Robinson. Ed., Marcel Dekke r . Ne'.' York, 1971,
Chapter 4: V. 'V' Chien. Chem. Pitarm. Hal. 24, 147. 1970
41 0' K. Erickson, K 1. Koch, C. S. Metha and J. W. MeGmity,
Sceenee. 199. 1457, 1976.3. VV. McG'ni:y. L. A Hun,e and A.
C..arnbc, 3. Pharm. Sc'. 66, 662. 19714.
Systenur. B,o,,,edicaj Apvk
41. V. W. Chien. Novel L.rao Deittx7v
caLwns and TheOeeLW& Basin, Marcel Dekker. Neu York.. 19S2,
Chapter 9.
Fhs, Pnarmaeol. 9.345,
43. B. B. Brodie and C. A. H. lioghen. J.
3957; P. A, Shore. B. B. Brodie and C. A. 51. Hogber., 3,
PnarmacoL Lap. Tner. 119, 35.1. 3967: K. S. Shanker. D. 3.
'Ioc. B. B. Brodie and C. A. H. Hogben. ibid. 122. 83, 1955: C.
,'. 14,5. Hogben. D. 3. '2occo. B. B. Brodie and K. S. Snartoer. ibid.
123. 275. 19514: L. S. Shanker. 3. Sled. Coors. 2. 343. 10460: L. S.
Snanker. Ann. Rev. Poarmacol. 1. 214. 11461.
44. R. H. Turner. C. S. Y4ehta and L. Z. benet, J. Pnarrn. Sc,. 59,
590, 3970.
43. 3 L. Cothzn and P. F. Kiliti., J. Pharrr.. Sc:. 55. 1184. 1969.
40. A. Suzuki W . 1. }i,guchi and N. F. H. 330.3. Pnarit. Sc'. 59, 044.
651. 2970N. F. H Ho. 'LV. 1. Higuchi and 3. Tun, 3.P'10. Sc:.
61. 199., 1972: 34, F. 21. Ho and W. I. H,guciu. 3. Pharen. Sc'. 63,
65)7. 1974-
ho, J. V. Park. G E. A,mdor. ci a... ii'
Gastrmn1tincs' Absorp tion oILsrsps. A 3. Aginar. Ed.. Amer
ican Pharmaceutjca Association. Aca000iV of Pharmaceutical
Sciences Washisigwr.. D.C.. 1981: (6) 7V. 1. }iieuch,. N. F. H.
.3 1 Par,.. .'ha 1 komivo it'. D'cac An'i'ri.:iO':. K F.
F'yeacettvid W. S. Nonmo. Eth. Adis Press. baioos'ia.. N5\\
20542 Aan.raia, 1981ps. K5.60;1cl N. F E. 31'.. 3. 'l Park. 'LV.
Moruzowsec and W. 1. liiguch. in Deswi. of Bwvjtor,.c.csi,iiC&
Properta 70r0u90 Proaruos and Anaoos. K. B. }'.ticfle. Ed..
Amecin Pnaymaeeutical /,.s.socaljOr.. ACSdVOY of Pnannaceu
tical Som. Washingcor.. D.C.. 1977. Chapter
4. J.T,Doin.F. N. 31b p s. 1. W.Ditcerc. L.T. Sugits and Y.
J. Pharsm, Sc... 55. 1190. 11469.
45 A. K. Miu and 7 . J. 1,likaeisor,.3. Pi'.snr.. Sc. 77.771. 1955.
P. J. Sáieupleir.. 3. lnves'. Derenatal. 41. 334. 1963. F.. 3.
S.cbeupa.. .3. Invest. Derrnacol 45. 79. 1967: R. J. Soneupleir,
and 1. H. Blank. Phvaio. he'.. 51. 702. 1971
J.
51 1. J. Seupiern. 1. H. Blaju.. G. 1. Branner and B
MacFarne. .3. invest. Derma'4 52, 63. 31469
52 C D. J. K. Fox. N. F. H Ho and 'LV I. Hipiichi. J. Pnarir
Sci. 65. 1341. 1347, 1979: ibid. 60. 772. 1980'. F. Durrneim, C. L.
Fivo, W. 1. $jguchi and C. 33. hen.. J. Prism. Sc. 69. 761. 1980.
C K. F0vz. H. Dur ybesm arid 'LV. 1. Higuch..J. Ptsam. Sc. 70.
52. 1953; C. R . BehI. C. K. Fl ynn. T. Kuru'.ars. N. Harper, W
Smith. W. L 1{iguch N. F. H. ho and C. K. Pierson. 3. Invest.
taerc& 75. 346. 1980
53. F V. kzvkar. Id-C. Fung. and B. D. Anderson. Phann, Rex.. 6,
340. 1.
€4. J. S. Twi. D. Dathelsor.. and W. Wolterior.. J. PharlT.. Sc,. 6$.
275.
56 S. H. Yalkowsk'v and C. 1. Fl ynn. J. Pharir.. Sd. 63. 1276. 1974.
50. .3. betrengi. C. Steinnietz and B. Pouleen. 3. Foam. Sd. 60,
1173. 1190. 11471.
51. K. B. Sloan. K. C. Siver, and S. A. M. Koch. J. Pharm. Sci. 75.
744. 1986
55. H KornaLs'a and H. Suzuki, J. Pharnt. Sri. 66. 590, 1979,
59. A. H. Becket:. and A. C. Moffat. 3. Prism. Pnarrnacol. 20. 239S.
1965.
6)7. N. F. H. Ho and W. I. Higuchi. 3. Pharm. Sc,. 60. 537, 1971.
651, 1970.
02 C L. Flynn. N. F. H. ho. S IOwans, E Owada. A. Mocokhia. C
E. Behi. W. 1. Higuchi. T. Yotsuyan.ag'., 1. Shalt and J. Park. in
Controlled Re,eo.ae Poivmeric Formn100iofls, D. R. Paul and F.
W. Hams. Edo., American Chemical Societ y , Washington, DC.
3976, pp. 57-123
63, S K Cnandmaseiiaran, H. Benson and J. Urquhart. in Sustained
and ConLeoILed Release Briar Delsveri, Systems, 3. R. Robinson,
Ed.. Marcel Dekker, Neu York. 1970, pp. 572-574.
64. (a) T. 'l'otsuyanagi. A. Molalchia, S. Hwang. N. F. H. Ho, C. K.
Pivot', and W. 1. Higuchi, 3. Priarir. Sri. 64. 71. 3975: 16' S.
Hwang. K. Owad'..T. 1otsuyanis57. K. Jr. I' F H
HoG. L. Flynn. W. 1. Higuchi and 3. V. Park,.). Pharm. Sc,. 65,
1574. 1575, 1976.
65. R. Okada, 1. Yesnazaid, Y. Ogawi, S. Hirai. T. Yashiki and H.
Miens, J. P1isrrn. Sd. 71. 1367. 1969.: H. CIcada, I. Yaraza5c, '2.
Yashiki and H. Him;, ibid. 72. 71. 1951;)'). Okada. T. Ya.chski and
H, Miens. ibid. 72. 173, 1553:33. Oaaa .....amazaiu, T. Yashilu,
T. Shimainoto and H. Miens, ibid. 73. 298, 1904.
66. D F. Othmer and G.3. Frohiici'.. hid. Log. Chem. 47.1034.1935
67. A. S. Miclisein and R. B. Parker. 3. Polyiri. Sri. 41. 53. 1936.
65. T. Higuchi and A. Aguier, J. Are. Pharm. Moor.. Sd. Ed. 46,
574. 1956.
65. 1. H. Blank, R. J. Sc3ieuplein and D. J. McFarlane. J. invest.
DermatoL 49, 523, 2967.
70. J. Autian, Plastics medication. Chapter 35 in Dispenssno of
Medwadso'.t. K. W. Martin, Ed., Mack. Eascor,, Penn.. 1971.
71. J. Crank, The Madhen'.c.ücc of Dift'asioio. 2nd Edition, Oxlord,
Oxiord University Press. 1975. Cha pters 5 and 6.
72. Id. 31. .)aeobo. Lnffu.swn Processes. Sprmger'Veriag. New York.
2967, p 43, and p. 49, euauon 39 where we use the symbol N to
replace Jacobs' tern". Q.
73. A. Oeubo. D',fiaision and Ecolocricc" Problems: Mo.nw.tic&
Models. Springer . Verlag. Berlin. 1980.
74 N.J. Wilson and B. K. Kilgore. J. 'lInear. Biol. -,1. 73. 1976.
75. S. Esezobo. S. Zubair and 14. Pilpe!, J. Prism. Pharniacol. 41,7.
1992.
76. K. G. Nelson and A. C. Shah..). Pharn'.. Sc,. 66. 133. 1977,
77 A. C. Scan and K. C. Nelson, J. Pnarm. Sd. 64, 1516. 1975:.ibid.
76. 799. 11457.
70. Y. W. Chico el al.. J. Pharer. Sc!. 63. 365. 1974.
79. A. C. Shalt and K. C. Nelson, J. Pnarm. Sc. 61, 210. 1990.
SO. 5, borodkin and F. K. Tsreer.J. Pnarrr.. Sc:. 64, 1299. 1975.
51. K. F. Farnp and K. C. Nelson. 3. Pssrtr Sci. 66, 1631, 3977.
ki. S. S. Da"i'.. Expecienu.'.. 26. 677. 1970'
82. A Adje. 3. New'ourger. S. 53.2rchan&117' and A Martin. J.
Pnamrs. Sc. 72. 742. 1954.
84. K. To)o. C. C. Ctuang and 1. W. Chseas. J. Pnarrn. Sc!. 76, 123.
1957.
53. A..). Agwar and 3d. A. Weiner, J. Pharm. Sc. 55.210. 1969.
80. Z. Limn and S. Coner., J. FOam. Sc!. 73. 534. 11484: ibid. 73, 535.
1994.
. N. F. H. Hoe: al.. J. Fharm, SrI. 63. 1576. 1976
30.
J. Priarm. Sc. 69. 770', 1980
Problems
13.-K The diffu.cior coefficient of tetracycline in a hvaroxyet14y.
meUsacryiate- Meth Y l methacrylate copoiym er fun in a mole mon
of 2 is B S.O'.:4.7F a 10" crn,o.ec and the partition coefficient
The qusntities m parentheses are siiaidird deviations. That us.
for emm*. the difluajon coeffsat ranges from (8.0 - 4.7) a 18''
4.7) a 10" based on a varbl1g of 1 wAndard deviation
(4.7) from the mean D value of S.

K for Ietrscycae between the membrane and the reservoir
6.8(5.9) Sc 10. The membrane thickness A of the trthmsnar deyiva
is 1.40 Sc 10 1 an. and the concentration of tetracycline in the core.
C.. is 0.02 g/nsi 5 of the core material. Using equation 13-731, p. 341.
calculate the release rate Qit in units of ,g cm - ' of tetracycline
per day.
Answer: QI1 6.71 em day'
13-2. Diffusion of 8uodnolo.se acetonide occurred from a 30%
propylene glycol-water solution through a circular section of a
polyethylene membrane in a two-compartment glass eel The thick-
ness A of the membrane was 0.076 tin and its diameter was 2.21 ccii.
The oartition co efficient of the drug between the snemarane and the
solution -as 1.28 at 2Y C and C. was 0.025 g/iOO cm'. A plot of drug
amount (in grams) permeating versus time in hours produced a
straight line (after steady state was established) with a lag time it of
25.0 hours. Calculate La> the diffusion coefficient. knowing ii and t,.,
and Ib) dQ?dz in u.ç(csrrt hi'), using the expression dQldi DKC,lL.
Answers: si .85 o 10 -s crn i 1,.r (b) IQI& 1.62 x
'(cm hi'). o r )_162 eg,(cr& iu')
13-3. Esez000 et al. the effect of concentration of binder
on the dissoeuuon razes of acetaminoonen tablets using the integrated
fcrrn of the 7ioyes- Whitney equation
I C. \
where C, is the concentration of the solute at saturation (solubility).
C is the concentration at time I. and k is the dissolution rate constant.
The data are found in the following table.
Data for Problem 13-3
I C.
Ln
(mm) no binder 2.5% P'TP
5 0.125 0.145 0.10
10 0.43 0.20 0.20
15 0.645 0.453 0.30
20 0.815 0.58 0.40
30 1.25 0.87 0.60
40 - 1.16 0.80
Plot 1n(_ ._) (vertical axis) against I and compute the dissolution
rate constant Is from the slope for the acetasninophen tablets a-ills no
binder and in the presence of 2.5% and 5% PVP.
Answer With no binder. Is - 0.041 min - '; with 2.5% PVP, Is
0.029 min'; with 5% PVI', Is = 0.020 min - '
13-4. A new fast..dissolving tolbutamide tablet was prepared and
tested for rate of dissolution. The milligram amounts (in) of drug
dissolved at various times are given here:
Data for Problem IS-I
[ tuft 6 8 tO 15
(-9)
m^ 1^14.4 -^M_S j 21.3 25.9 25.2
(a) Plot the r es ults art i-ectsngalar graph paper (m versus t) and on
nemilog paper (log in versus 1) arid Snafly plot in versusC 12. Which
method yields a straight tine?
(b) Using regression analysis and the data that give a linear
relationship, calaslate the correlation coefficient, the intercept, and
the slope. Is, the dssohrtion rate constant.
CAnpter 13 . DOrIaio% and Dioàtioe 367
(c) An intereept - the value of the ordinate of a graph at the
point at which the quantity on the horinsocal am is ssro hi is it
poesible that the mil ligrams o( tolbutamide dissolved at 0 ini&"5 is
3.3972 mg Give severa' possible expiazsacm.
Partial A,jwir. (b) r1 0.992?. slope Is = 7.706 m(rnin).
usternept 3.387 mg
13-S. A 0.625-g sample of acetohexamide powder was dissolved in
1000 cm3 of water in a Paddle-type dissolucion apparatus at 30' C. The
dissolution rate data are found in the table below. They are presumed
to follow the liizsoas-Crowell law.
Data for Problem 13-5
Weight I
Time Concentrazioa Undissolved Us
(nun) (mWznL) is gram (g) M,"5-M"
o 0 M.,=0.620 Io -
0.159 0.4136 ).080 0.040
4 0.288 0.337 0.159
0.390 j 0.225
0.468 1 0.157 0.315
0.529 0.097 0.396
(a) Calculate c for each time period to fill in the last column of the
table. Then add these separate is values and divide by the number to
obtain an average a.
(b) Regress (M - M us) versus time in minutes and obtain is as
the slope of the least-squares line. Which of these methods is the
better fr obtaining ,c? Give reasons for your preference.
An.nocrv (a) a5., 0.0397 ginimin; (b) a 0.0304 17"trnin
13-6. Nelson and Shah" applied their convective diffusion model
(see Ksampl. 13-6) to the permeation rate (rate of diffusion). R. of
butamben through a dimethicone membrane under conditions of
aoueous don layer control (p. XDO. For a circular membrane;
the permeation rate is given by the convective d (CD)
equationt
2.157 D°C,&'r5 (13-115)
where D. is the diffusivity (diffusion coefficient) of the solute in the
aqueous layer. C. is the solubility, a the rate of shear over the
membrane, and ,' the radius of the circular membrane. The rate of
permeation, 1?, written as dMldt elsewhere in this chapter, was
considered the dependent variable, usung the radius o f the membrane
as the independent variable. The equation obtaineO by least-squares
regression predicts a straight line in a plot of log 1? versus log r,
holding the other parameters constant-
(a) Prepare the graph of log /f versus log r, compute the slope and
intercept, and determine how well they compare with coonparaoie
quantities predicted by equation (13-115). The values of the . 'urn
parameters are 1),, - & x 10° con2 sr-c = 3.60 x 10', con2 n js
C, 9.4 * lO mole/liter 9.4 x 10' mole cm - '; and a 58.006
min; the experimental R and r values for the plot are:
- Data for Problem 13-6
R (nioleniin) 1 1.325 C 10_ a 2.712 a 10 - ' 3.361 X
r(cm) 0.65 1 j L25 ]
(b) Choose one of the three values of r from the above table and the
known parameters given above; subetitute these into the least'
squares regression equation and into the CD equation (equation
(13-115)). Compare your results from these two methods for
ealculadaig R. What is the percentage error in the two methods
relative to the experimental value of R?

355 Pnysw& j-is,-nuicg
(c) This system is under aqueous diffusion la y er control. Compute
the thickness k. of the static aqueous aver, using a- - 1.25 an and
R 3.861 ii 10 mole mm'- 1 - 6.47 x 10 mole see.
Notes: in preparation for graphing, be sure to first convert R and
a- to logarithms Recall the when converted to logarithmic form in
equation (13-115) quantities multiplied together are added. There-
fore equation (11-115). written as log R versus a-, becomes log
(2.157) • 24 log Li, log C. - lJ3 log n i- 5!3 log and for D., Cr
and u held constant:
log R lsg(2.157 D 14C,oVi 5/3 log- 113-1I6)
j effects on the system
with the first og terms on the right side as tee intercept and 5.3
1.667 in the second log term as the siope.
Partial An.su'er; (a) The equation of the linear regression line in
log Il - 7.56_° -4 1.646 log a-. The slope. 1.646, and interceo. -7.569,
compare wet with the values 1.667 and -7.4739 predicted by the -
convective diffusion (CD) equation (13-I15) (5) For r = 1.25 cm
substituted into equation (13-116) for the linear regression line
log A -7.561' - .645 loglt.25:
A 3.90 15 - ' mole/mm. From the CD euustior,. A 4.57 s If-
molennir. The regression equation (15-116) has an e r ror of about
0.5 and tile CD equation, (11- liS). an error of about 26tti is reiatlon
to the experimental value of A 3.861 v lC' moie,mln. (c) Under
aqueous diffusion layer control, p. 33.5, 005atior.
dM1 I V dii' (S D - '1c
J A C. = C..
Ii, = 00214 can. The actual dimethicone memorane thickness is about
0.025 cm.
33-7. Shah and Nelson- studied the dissolution rate of alkyl
p-amioobenzususs using a convective diffusion model (see p. 334 and
Erarnplr Js-$), where for a circular sha p
ed tablet of surface radius
a-, the equation is
A = 2.357 DrsC,n t '2r 1-
D is the diffusivity. C, the solubility of the drug. and o the rate of
shear over the dissolving surface of the tablet. A is the tablet
dissolution rate.
(a) If C. is 9.33 a. it' - mole'crr 3 , o is 255 rTuE'. a- is 0.5012 car..
and R. tue rneavuj-ed dissolution rate, 1.533 a. )0° mole mm - ', wrist
is tue numerical value of the dsffusiou coefhcient5
(h) One may wish to stud y the change in the dissolution a-ace A
s-itt a change In the radius r. or regarding the change- of another
- controlled by time aourtiirs drifuimmun laser or by themembrane
pa±snieter. such as thE, s.olubthtv C, or the rate o f shear u. because of
time exponents in equation 113-115, this is best done b y p lotting ton
iogorttnm of P sea-ce.-' tee Isgaretnin of r, hold-n.- L'. C_ and o
rurstust Tri p ecusuc:. t,e-come./
log A = (ojr(2.157 iog Li - log C, + -3og o) - log - (1-1-117,
Using tha- dais front table (a' below, plot log A versus log m- 01
rvceanguiar coordinate graph paper. If the plot is linear with a alote
of 53 1.667. or approamate:\ se. sot- nave a good indication tea:
the convective difiusior approach is a satisfactor y model for tn,
aisn-oiUuon iQnJc asaiov WiOer invesL'gatinrt. Next. one cart p lot log A
versus ion C, for exam p le, using the data in tabje (t.. side niohiin;
D. a. and a- coaant. W'nat ssouid be the slope of this line"
Data La) for Problem fl-7
P icmr"ntun log P 1i r lent
1.25 Y. if_ i -6.90 0.391' ' -0.41'
2.34 10' -6.63 0.575 -0.24
4,17 a 10 -625 0.794 -0.11'
LIsts (h ) for Problem 13-
A crnintui: log P C, mole/car" log C,
125 5 10 -6.74 6.31 x 10 '-
2.20
225 5 1' 1 , - 7.55 9,33 a 10---
L1'04 x 10 - 9 -6.10 2.9 a 39_I
The authors show that o t her properties such as viscosity of the
solution n-isv also be tested b y the- convective diffusion model for their
Answers, is) Li = 2.39 a 10'- ' crr 1/mir.: (b) in a plot of log R versus
log a-, the slope obtained from regression analysis is 1.725. The
theoretical siope In 513 1.667 because of the final term in equation
(13-117). in a plot of log A versus log C. the slope should be 1.90
since the exponent on C. it. equation (13-115) or (13-116) IS mints
The least-square equation obtained is log A -4.556 • 0.991 log C,
The oouared correlation coefficient, r. 0935.
13-5. The release of etns-nodis diacetate- through a aibeone- dosapt
form ia-a,' be caicuiae.ed using the Eiguthsi equation.
= ID(2.4 - C )C,u5
since diffusion is found in this case to be the rate-lamiting factor for
0mg release-. A. toe amount of drug per unit volume o f the silicon
matrix. as 100 g.')IO" cm';, the s.o)ubilits' C, of the drug in the silicone
pol ymer it 1.50 gi( If" cm' ! . and D. the diffusivity of the drug in the
silicone maL-tO, is 3.4 a. 1(' -" cm°lday. Calculate the rate of drug
release from the silicone nonage form in units of g/003 cm 2 ) per day°'2.
Ansuesr 3.182 g/(]('° con 2 ) per day '°
13-9. \V'nen C, is small relative to as found in Problem If-f.
the Rigucha e q uation reduces to Q.'1° \'WAC,. Recalculate the
results of Pa-bole-ni Il-f using this abbreviated equation.
An.si,er.' 3.194 g/(1(''1 cm t ) per 0aylt. A value of about 3.) gilIff
cioi') per da°5 is obtained experimental) v.si
13-10. The permeation of methyl p-sminobensciate through
disnethicone membrane of thickness 1.,, - (1. 026-4 cm wa conducted it
a sarnanar flow cell at 25' C. At hugs fibs rates the membrane Li,,
and aqueous Solvent Li, diffusivlties are D,,, 2.72 a lt' cict
and D. 6 a crc2 sec'- '. The thickness of the a q ueous layer I:.
is 0.003 cm. the solubility C of the drug is 1.04 a. 10 moleqiter.
and the partition coefficient A' across trio drug-membrane interface is
0.2€
Is) By a pp ro p riate calculations. sugges: s--bother the 6o.s is
fbi Com p ute thefiux. .7 In-isle crn rnir.J See- page 333 for
membrane and difius,om la y er contrc'
,4"s"es-r is' 2' ,.EL /.,,,Dv: tn, p rocess one of rr,emoran,
centre... Snow this Lv 0,' tree i, caicuiating esese two titrantities in
The fiux (A'b_i,.,)C for nie,morane csrttro: is ).')-t v 15' mole
em '-- mar. '. Using the entire eQuation (13-59,. page 335. we obsisri.
1.61' a. I)"- mule ca-n' 2 mir,, which compares favorably wit,j,
I 1.7$ s _0 oriole err" nur.'. the membrane control result
13-11. A cell contains a a jiasuc tni€mbra,no- with diliusion save-ri- of
lbentic,aJ thicaness on either side and burs'l-amui-iobenzoase at conrer-
u-won C= 1.72 n'unoie/uter or 1.72 a l0' mmole,cori 2 in the dono-
comparuner,t. Calculate Inc stead y -state flux. J. through the merr..
ba-see in nnillirnoies per err.' per hour. The equation that represent.'
time procesm in which both membrane and diffusion-laser conu'o.
DDLAM0
Li_fCC',
- 2i.JCC1'
Toe data obtained from an esoerirnent at 37' C are as foleoa-a: L,_ -
27 ii )('° con°,iiec or 0.00972 cnt°3tr, D - 6.0 v 10° cni m,eec =
0.02164' an 5/nr, I,.. 0.006 err.. ii, 0.0155 cm, and K 10.3.
Aaawrr. I - 9.53 a. )(' mnrnoiellerr 5 br
13-22. B,oyodi<in and T erswdiedtJtediffualoii of aalic'vbe acid
from a polymer filin containing dispersed drug. The kineties wmade

linear with twit. i.e.,zero order, by laminating a second film,
consisting of a b.vdroxypropvkedulosepojvvmvl acetate membrane of
thickness F. 0.0164 ctr to the releasing side of the flint with the
drug as a reservoir layer.
The drug layer controlled the duration of release, while the
nondrug la y er consisting of the cellulose membrane served as a
rate-conu-oiling membrane. Diffusion through the film is the limiting
factor ins drug release, so that ecuation (13-72) on we 341 applies.
Q, the amount of drug released per unit surface ares. is given for
various tunes:
Data for Problem 13-12
(mg/crc2 ) 0.46 1.0(1 1.54 2.19 2.69 3.23
I 1 I H i j4 5 67
Equation (23-72; may also be written as
dQ D,C.
(13-IM
=
the instantaneous rate of release of drug at ttine integration of this
equation and evaluation of the integration constants yields a term for
toe lag tune.
ic,
Q t, 113-119)
Equation (13-1291 shown thata pie: of Q against t yields a straight
line of siope D,.,C,1Js, = . tee auparent zero-order rate constant.
The mterrept of equation (13-119) is (DCll,,,)t from
wojen can be computed.
La) Plot Q versus time. t, from the data given in the table. (b)
Regress Q against (amid obtain i: DC,Jn. (mg rot'i from the
s/one and t,, from the interce pt. (C) Knowing the lag time and the
treckness ., of the riondrug la y er, compute the diffusion coefficient.
T. _. You win need the lag time eouatjon 113-14(2.) from page 325, (d)
From toe slope ostmiined is par. 'b) and the diffusion coefficient D
calculated in part (c). compute toe value of Cr. the concentration of
the drug. salicylic acid, in the reservoir layer of the base flina
Answer. (b) k, = D.C,jç,. = 0.530mgcrn' t hr - :z. 0.Q539hr
or 3.23 mitt. (C) D,, = 2.31 a I( em" sec- : using 5 0.0154 cm;
(dl C, 10,46 mg rm-'
13-11. Farng and Neivon n studied the effect of tioivelectrolvtes,
such as rarboxvmethvkeljulose (Chic), on the permeation rate of
sodium aabcvlate across a cellulose membrane at 37 C. Pick's law
ma y ce written to cover Lnts case, assurmng ins existence of Litres
barriers it sener: the memnray e arc ar, uns::rrei liould diffuior,
layer or, e;trer s J i:it, The renrrors, cf :z, pvrmea,icn coefficient for
tbt tnree users is
in which 3. 5= fiS s I("'' en-, for the two static diffusion lavern'
46.6 5 10"' mit for the rellu,ose membrane tiucariess. 2),
3,33 x 10"cin 5,sec, D_ 1.11 5' 10,crnt sec. and D,, ),69 a
crit,sec- for the three diffusion cue.fF'cient,t. Toe Dam-titian coefficient .9'
for use saljcvlate between the solution us compartment I (left-hand
reservoir) and water in the membrane is 1.16. Tue volume fraction
Of 'ester in the membrane is 0.667. Calculate the permeabilit y P for
the ealirvlate and compare it with Pua. 1.86 N )f" cm,set.
Answer P,, = 203 ii 10"''
13-1i. Suiladtazmna, pK • 6,50 a: 2 C. as with at weak acids
snows 2 va-abie. perreiti dissociation as a function of pH. is) Prepare
a table snowing tue pam-remit diasociatec and Derceuv, undiasocated
aulfadmazanc at pH 2, 4, 6,5, 7. E. K. and 12. (b) Plot toe results or
rectangumar graph paper, (c) Predict the absorption of sulIs.daam,e
from the gut and from the smut intestine in terms of the phi-
partition nypethesis and f rom the results obtained in paris (a) and (b:
Portia,' Answer; Percent dmsoa5mon as p11 4 U23; at phi S
96.93
Chapter if. Diffusion and Dissotutto y, 359
13-15, Bile acids are transported across the ileum by active and
passive processes," For dilute norimicellar bile acid solutions, the
absorption rate constant is giver, by equation (13-93):
p,x
in which P5= is the permeability constant for the active transport of
We add anions: Pv the permeabiuits constant of undisscciated scies
undergoing passive diffusion: P 0 Li e permeability constant of the
aqueous diffusion layer: and X,' and X, - the fraction of undissociat.ed
and anionic species, respectively
For taurocoolic acid, pK, 2.0, the bile acid is essentiall y all in the
ionic form at pH 6 to 7. Therefore, X,' - 0, Xj" - 1.0, and A,
becomes
S P.0
A,
F,,
If SA ' . the ratio of effective surface area 'to solution vclun',e, is 10
cm", F., is2 a 10" cm,sec. and P_ 1 is I a 11" cm/se-s. what is
the rate constant for the absorption of taurc,cnolic acid across th,
ileum"
Answer: K 1.67 a o"
13-16. in situ rat gut permeation exoerinsent,s were oer(orrned a:
37' C to determine the influence of the I'iildr'orar,ii soluhihty
parameter. 6. The rat intestinal pre paration isshown in Figure 13-13
and the parameter is defined on page 23-1. Toe drug
theophyinne, with a solubilit y parameter or delta raise. 5,. of 14
(cal,anO )ts and a molar volume, 'u' of 124 em 2/moe. was dissolved in
1
the solvent mixtures of pol yeth y lene givcol 400 and water of various
polarities as expressed by their solubilij' parameters. 3,
It is postulated that inc more alike are the delta values of the
drug li, and the solvent mixture 3. tee grnss.er would be-inc attraction
of solvent and drug and toe poorer the release of the drug for
absorption Conversely, when 3, and 5, are quite di fferent arid aiss'
when the soluoij's y parameters of the drug 6 2 and the intestinal
mucoss E. are simliar, the drug is readi)' released from the sciven:
for bioabsorptioia.
A parutios parameter K is obtained in the rat gut experiment
from drug absorption 3, and desorption k rate constants, in whic'r;
A:, = k,/e.. The experimental data for this problem are found in the
table:
Data for Problem 33-10
er
PG _c Iii - Z- r
UI "are' I 11,,inicm'"°' ' - ' muir i 0,0.,
ie 11.1 -2.2 0.041 0,025 11,464 10.33:
SC 13 -o.s I0.040 0035 l.02Z
-5.2 t.osG 1 sot' -o 42'-
I 14.4 11,4 I 0.045 I 0 O'i I 0.64: -0 446
65 lOt' 1.5 I 0.036 I O(Oe , 0iii
6e H: 1.7 503 1,031 hOD I00
(a) Pin: in K versus 1k, - b, and hate from the curve obtained
how the partition parameter varies with Inc deiLa value difference
(3 - 05).
(b) Regress in A., against (, - 02) using a paraiso'ac (quadratic, 0
better, a cubic least-squares regression procedure, Plot a iufficuen:
number of points, obtained front, either the quadratic or the cubic
equatmor,, on the graph prepared for (a). This stitinZacai J'tgresaiOr.
analysis .bouid give a smooth line approxurosting the rigs anguia=
360 PktPb.ej
- o&*nsd &nI - strug - the exphesia1
polaiva in (a).
(c) Calculate a theoretical partition parameter (In K,) using the
emplrscal equation suggested by Davis't
in K, --
V7 -of ((61-S.f-(5,-&)9 (13-120)
RT
where V, the molar volume of theoayiiiae. is 124 /rrioIe and S. is
the solubility parameter of the rat gut membrane. 12.6 (calIcin3)° a
determined in the work of Adjel et aL. R is the gas constant, and
7 the absolute temoeracure (31' C or-31T K). The volume fraction
(p. 324). b, of the solvent mixture is approximately unity in this
tiroblem and may be disregarded in equation (13-120). Are the
-xperunental In K, values St better with the In K, obtained from
solving equation (13-120) or by the In K, from the quadratic or cubic
regression equation? Observe the points on the gra p h or the data in
your summary table (See below) to answer this part.
Partial Answer! (b) The squared correlation coefficient ri had a
value of onl y 0.794 when carrYing out parabolic quadratic) regres-
.ion. so the quadratic equation was droopcti mini further consider-
ation. The cubic equation is
In K,, -0.4541 - 0.3689 a- 0.2166z 4' 0.0960r (13-121)
and has a satisfactory of 0.957. The term a in the equation stands
for 18, - s,j. ':c( The results of calculating In K, from equation
13- 120) are shown in the following summary table, together with in
K, (experimental) and In K, calculated by carrying out cubic
regression (equation 13-121).
Answers for Problem 13-16
in K ' In K,
P532 400 Solvent In K, (cubic eq.) (Eq.
us water 0,(calJem°) [)8_8) (exp.) (13-121) (13-129))
90 .11.2 0.294 0.519
51) 13.1 -0.9 0.1)28 0017 -0291
75 13.7 -0.3 -0.429 -0.327 -0.316
70 14.4 0.4 -0.446 -0.561 -0.252
4 11.)) 1.0 -0574 -0.510 -0.193
5) 15.7 1.7 0.029 0.9)7 9.IKI
It is Seen n the ouznlnary table, column 5, that the In K. values
obtained by cubic re9reoaiori correspond satisfactorily to the In K,
(experirnentoi) and the cubic ecuation, (13-121), can be used to
calculate In K, and therefore K,.
The cakulated In K, values from equation (13-120) (column 6 in the
summary table) differ from the experimental In K, values, but the
principle remains that the Largest (8 1 - 9.,), Le. the largest
differences between 6 1 and 8.,, correspond to the greatest In K,
(experimental) and through the expression K, k,/k-, to the largest
absorption rate constants. Conversely, the smallest In K, (equation
(13-120)) occurs at 8 1 of 13.7 to 14.4 in which region is found the
drug's delta value, namely 14.0. Here the solubility parameter of the
mixed solvent, PGE 400-water, is sufficiently dose to that of the
drug, theophylline, to interact with it and hinder the bioabsorptlon of
the drug.
13-17. (a) Asa continuation of work on trsnscornesl permeation of
pilocarpuse (see Example 13-14, p. 346), you are asked to plot the
experimentally detennursed corneal membrane permeability P versus
pH (Table 13-5. p. 346) on ordinary rectangular graph paper.
(b) Calculate P(theoretical) (see Example 13-14, equation (13-
98)) using the values P8 = 9.733 x 10' cm/sec and Pan = 4,335 ii
10 cm/sec. The non.iortized fraction f8 of pilocarpine at each p11
value is calculated from equation (13-96), page M. Plot P(tbeoret-
ical) versus pH on the graph prepared under (&)- Note that the pH at
the inflection point of the sigmo(dsl line is equal roughly to plC. - 6.67
for the con)ugaus acid of psiocarpine at 34' C.
(c) Does this siginoidal curve suggest any relationship to the
titration curve of a weak arid, Figure 8-1. page 075, where, at
hall-neutralization. pH -
P,udiail Answer* (b) at pH 4.67. P(theoretical) - 4.885 a
cm/see, and at pH 6.67. P(theoredcaf) = 7.255 X 10" crotset. Ic) The
Henderson- Ha,sselbalch equation is used on page 176 to arrive at the
relationship pH - p /C• at the half neutralization point, arid the
Henderson- Hasselbalch equation is also used on page 342 to obtain
an equation for the percent ionization of a weak and at various p1-1
values. These remarks should assist you in arriving at your answer.
13-I8. The oteady-state penetration of progesterone and its
nydroxyl derivatives arroos the intact skin was found to be related to
the solubility of the drug in the stratum corneum. The solubility in the
stratum corneurn and the rate of Permeation decrease as the
hydrnohilic,tv of the progesterone derivatives increases. 5' The data
are found ii the following table for the progesterone derivatives
numbered througis 7:
Data for Problem 13-18
Drug no. 1 1 2 3 4 1 3 1 6 7
Solubility
(mgt,nL) 109 2.46 i* .91 2.82 12.5 32.4 47.7
I[rate
Permeability
(isgi(cm° hi-)) 0.15 0.31 0.97 0.29 0.57 4.73 2.37
(a) Plot the permeability rates dQ/df. (gs(cm' hr)), on the vertical
axis against the solubility of the p rogesterone derivatives on the
horizontal axis of three-cycle log-log paper. One may postulate a
Linear relationship between the steady-state rat-es of permeation of
progesterone and its solubility in the stratum corneurn. However, it
is found that a linear relationship is obtained only when the data are
plotted an a log-log graph, which suggests that the true relationship
is a power curve. Let permeability rate dQIdt be y and the solubility
in the stratum curnearn be a, and write the power curve relationship:
Y isa5 (13-1.32)
in which a and I are arbitrary constants.
(b) Use a hand cacuIator or a personal computer to obtain the
values of a and I using the data given above.
(c) Now assume a relationship between y and a in the linear form
Y = ar 4' 1' (13-1215)
and com p ute a and V. Which equation. (13-132) or (12-123), better
tits the data?
Id> Why did a log-log fit of the data suggest using the power
equation I, =
(e) Do you believe it would be possible to extrapolate results such
as these to the transdermal absorption of progesterone derivatives
from various ointment bases into intact and damaged or diseased
human skin? What factors would need to be taken into consideration?
Partial Anawm-s: (b) y = 0.111. 11w6, (c) y 00724 a 1- 02053
13-19. The percutaneous absorption of chloramphenicol through
mouse skin was investigated at various temperatures. Permeability
coefficients were recorded together with temperatures as follows:
Data for Problem 13-19
Temperature. 'C 25 31 31 45
P (cm/min) x 10' 1.12 1.87 3.01 6.20
Plot the Arrheruu., curve and calculate the energy of activation for
permeation. Compare your results with those of Agwiar and
Weiner," who, in a similar study, found E. to be 15,000 cal/mole.

Answer: 15945 cal/mole, or 16 kcal/mole. The answer varies
depending on the number of significant figures retained.
13-20. Liron and Coher studied the pereutaneous absorption of
the straight thairi fatty acid, propionic acid, through porcine sian.
They found that the permeability coefficient. P - KD/h (equatior.
(13-11), p. 327) increased with elevation of the temtiersture as shown
in the table
Data for Problem 13-20
TernDerature, C 15 1
P x ceisjnur, (1.8) 1.21 1.64 3.00
Prepare an Arrhenius plot of in P against reciprocal temperature
WT. '}'). Obtain the activation energy, E, in heal/mole, and P_
which is proportions) in the number of molecules entering the film and
to the probability that these molecules have sufficient energy to
engage in the diffusion process. P is comparable to the fi-eouenv
factor A in the field of Kinetics /OOU3UO (12-72j, p 295:
Answer K 11.35 atcai.maie: in P 12.62. P 3.1 5
13-21. Liroi'. and Cohen' applied Fick law of difisistari to thc
percutaneous absorption of a drg. The delivery of vaienc (penlanoic)
acid from n-heptane solution (I molar) into excised porcine skin was
studied in adiffusion cell having a I col t cross-sectional area. The
following resultS were ootained.
Data for Problem 13-21
2 3 4 5 £
Time (hr)
0l(,.51 g/crn°, the solute's diffusion coefficient in this medium is D 9.81 a
Q.iemoieicnl01.47.6119.i 34.149.164.1 1 79.3
(a) Plot Q. the cumulative penetrated mass per unit area of valeric
acid. against time.
(hi Find the flux at the steadstate. from the slope of the linear
segment of the cumulative penetrant mass per unit area Q versus
time, i.e., for time 2 through 6 hours.
Ic) Use J. and the concentration of valeric acid in the donor posse.
C 0 . 1005 eimoiemrn 5 , to calculate the permeability coefficient, P.
(dl Find the Lag time front the penetration curve by extrapolat-
ing the linear segment to zero penetration per was cross-section
(c) Find the diffusion coefficient D from 1L and thickness h of the
skin barrier (stratum corneum): (1 15 micrometers (1i.sn
tl) From the results obtained, calculate the partition coefficient,
using equation (13-li), page 227 of this chapter.
igi Calculate the partition coefficient K using the equation of
Davis..
V. d
ir.K,--1((, &18c&vDa.7.,)Jf1Z..124
in which t', is the molar volume ofuaesoiute. Jtand Thave their usuai
meaning. 8, and 8 are toe s.oiubth:y parameters for the solute larugi
and solvent. respective),. V, in the volume of the donor phase. equal
to 300 e._. and 1_. is the volume of the elan bamen eOua) tot cm
0.015 mr = 0.015 err' - 15 iii. in Problem 13-IC. the approximate
equation (13-3201 was used. disregarding the final tern'.. In 11'11'
in the present case it is nuffloexti y import.aitt and casino: be
Tnese avuaom are thanuied for providing Fi-o&enu 23-20 and
22-el.
Chapter 13 Diffusion and Dsotiolathos 361
Partial Answers: (b) J 15 usnolellcmctm hr); Ic) P 0.25 it
cm/coin; (d) (, = 40 nun; (c) from equation (13-14a), p. 325, D =
1.6 a 10 -11 cm I sec; (I) K 39.1: (g) K = 43.4
13-22. Hydrocortisone was released from a silicone matrix ins-
planted in the vaginal tract of a rabbit (eee pp. 350-351) Caiculate
the amount of hydrocortisone (mg) released from the vagina) implant
in 2.5, 5, 10, 15, 20, 25. 30, and 40 days. Plot the amount released. ns
(mg). versus time in day s. The following data are taken from the war).
of Ho es a.°' The solubibty of the drug in the polymer mssrtx. C.. wax
0.014 rag/cot3 ; the diffusion coefficient in the matrix, .1),. was 4,5 5
10-' cm'.mec: the partition coefficient for silicone/water. K. was 0 0-1:
the permeability coefficient of the rabbit vaginal membrane. P,. was
5.8 5 10" cm/sec; jo, was 7 a 30 cmisec; the loading concentration
(initial amount of drug per unit volume of plastic c ylinder), A, was 100
rsgvcm 5 the length /t of the silicone c ylinder was 6.0 cm, and its
radius, a,,.
" was 1.1 ml'..
Partial Answer: 36. mg released in 15
f days
13-22. The rate at which a solute di fuses out of a single open-
boundar y' eel). Figure 13-30. at time I is e q ual to the concentrattor.
gradient awax at the open boundary' multi p lied by ID x A) the
diffusion coefficient of the e.niute multiplied b:. toe area .4 of* he cell.
The rate R of solute escape is given h3
P (gisec) = (D - A)
ax
2u (,-2Lit .9iT°DI - -.
-ezP 5 -r -e exp;---'.- .....1(L) Ai (1 125)
The area A of the cell is 6.78 cm 5 , the cell height 8 is 2.92 cm. the
initial concentration ii,, of solute in aqueous solution at 25' C in 0.0628
10 cnr°isec, and the time of diffusion is t 8560 Sec (2.38 hr;
Calculate the rate of escape from the cell
Ass-steer: P - 2.99 5 10-e gisec = 2.39 x 104gOiuis
12-24. Anew drug is placed in a Graham closed boundary diffusion
eel) (see Fig. 13-29) to determine the cr-rig's diffunron coefficient. D.
The initial concentration of the drug it, is 0.0273 gicm° in water at 25'
C The total height of the cell is B = 2.86 cm and the height of the
drug solution in the eel) is 1.93 cm. A sample iS taken at a Oeptt
Of r 3113 at time 1 - 10.523 eec (2.925 hr) and analysed for the ds-.ig:
it, concentration ii is found to be 0.0173 g,crni, Rearrange equatioc
(13-113) so as to calculate I). the drugs diffusion coefficient. (See
Ezample 33-36.)
Answer: D - 12.4 x 10- 1 mrr°/set
13-25. Diffusion through the skin may be considerec as a multi-
layer diffusion where the rosin lasers are the stratum corneurit.
epidermis, and dermis. The stratum c'orneumn is the mayor diffusions)
barrier, and eoidermis may be signihcaistiv less permeable than
dermis. Yu ci at.' performed experiments of permeation of a orc'drug
of vidarabme to study the diffusiona) characteristics of toe differ-en:
layers of the skin. Using oermis membranes of mat-c laverage
tsickness, 35(1 11.10), the lag time for the permeation of the drug wss
152.1 sec
In a second experiment. the stratum corneum was removed with
Scotch tue. Tne permeability coefficient of the strapped skin lP'i
(without stratum corneum I was determined to ti11 1,74 a 10 - cnn/sec.
Finall y , the permeability coefficient of the whole skin (stratum
10 cnn,o.ec.
corneurls, epidermis, and dermis' was 2,95
(a) Compute the permeability coefficient of the dermis. You will
need equation 03-14/ti. We 221. for this par.
(b) Compute the permeability coefficients of the stratum corneure
and the dermis. Jdsnl' Use equation (13-51b;, page 237, as a model.
Answerz: (s) P (dermis) - 3.1 i(' cnn/eec: (hI P (stratairo
cornewrn) 3.0 a I0" anis-ec; P (epidermis' 4.2 5 10' cm/sec