Evaluation, control, and prediction of drug diffusion through polymeric membranes

MISSING IMAGE

Material Information

Title:
Evaluation, control, and prediction of drug diffusion through polymeric membranes
Physical Description:
xiii 151 leaves : ill. ; 29 cm.
Language:
English
Creator:
Chemburkar, Pramod Bhaurao, 1932-
Publication Date:

Subjects

Subjects / Keywords:
Pharmaceutical Preparations   ( mesh )
Cell Membrane Permeability   ( mesh )
Polymers   ( mesh )
Pharmacy thesis Ph.D   ( mesh )
Dissertations, Academic -- Pharmacy -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1967.
Bibliography:
Bibliography: leaves 145-150.
Statement of Responsibility:
by Pramod Bhaurao Chemburkar.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000898626
oclc - 20764388
notis - AEK7332
sobekcm - AA00006121_00001
System ID:
AA00006121:00001

Full Text















EVALUATION, CONTROL, AND PREDICTION
OF DRUG DIFFUSION THROUGH
POLYMERIC MEMBRANES


By
PRAMOD BHAURAO CHEMBURKAR


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY











UNIVERSITY OF FLORIDA
December, 1967


't1>)i"






















DEDICATION


To my wife,

Mrunal,

whose patience and help made this accomplishment possible

and to my parents,

Mr. and Mrs. Bhaurao H. Chemburkar,

whose encouragement has been a source of unending energy to me.
















ACKNOWLEDGEMENTS


The author expresses his sincere gratitude to Dr. Edward R.

Garrett, Chairman of the Supervisory Committee, for his valuable

guidance in research work and help in the preparation of this manu-

script. He extends his appreciation to Dr. Oscar E. Araujo, Dr. Robert

B. Bennett and Dr. Russell E. Phares, Jr. for serving on the super-

visory committee and for their helpful suggestions during the course

of this investigation.

The author gratefully acknowledges the following companies

for the supply of chemicals and materials: Abbott Laboratories,

North Chicago, Illinois, Avisun Corporation, Philadelphia, Pennsyl-

vania, Dow Corning Center for Aid to Medical Research, Midland,

Michigan, Eastman Chemical Products, Inc., Tennessee, E.I. duPont de

Nemours & Co. (Inc.), Wilmington, Delaware, May Industries, Inc.,

Atlanta, Georgia, Syntex Laboratories, Palo Alto, California, The

Polymer Corporation, Reading, Pennsylvania, Upjohn Company, Kalamazoo,

Michigan, Vicks Division Research and Development, New York, New York.

The author wishes to extend his appreciation to Mr. Barry

Dvorchik, Mr. Arthur H. Kibbe, Miss Michele Deckle, and Mrs. Maria

Losada for help in the experimental work, Mr. George L. Perry for

preparing illustrations, and Mrs. Sharon Cooper for typing this

dissertation. He also thanks Mr. H. J. Lambert, Mr. P. J. Mehta,











Dr. H. J. Nestler and the rest of his colleagues for their suggestions

and participation in the formal and informal seminars pertaining to

this investigation.

The author also wishes to thank the American Institute of

Biological Sciences, Washington, D. C., for a grant to defray a part

of the expenses of the preparation of this manuscript and the College

of Pharmacy and Graduate School of the University of Florida for the

financial help during his tenure as a graduate student.

Sincere thanks are extended to Mrs. Arthur H. Kibbe and Mr.

Y. Raghunathan for their help in proofreading this manuscript.
















TABLE OF CONTENTS



Page

ACKNOWLEDGEMENTS ............................................. iii

LIST OF TABLES ............................................... viii

LIST OF FIGURES .............................................. x

LIST OF APPENDICES ........................................... xiii

INTRODUCTION ................................................. 1

HISTOR ICAL ................................................... 2

MATHEMATICS OF TRANSPORT PROCESSES ........................... 7

EXPERIMENTAL

Materials ........................................... 13
Methods of Analysis ................................. 17
Screening of Polymer Films for the Permeability
of Drugs ............................................ 23
Diffusion Apparatus ................................. 24
Solubility Studies .................................. 29
Determination of Partition Coefficients .............. 30
Measurement of Thickness of Membranes ............... 32
Study of Permeability of Silastic Membrane to
Phosphate Buffer Salts and Hydrochloric Acid ........ 32
Preparation of Solutions ............................ 34
A Typical Steady State Diffusion Experiment ......... 35
A Typical Quasi-Steady State Diffusion Experiment ... 37
Diffusion of 4'-Aminopropiophenone Through
Silastic Membrane ................................... 38
Diffusion of 4'-Aminoacetophenone and 3'-Amino-
acetophenone ........................................ 44
Diffusion of Barbituric Acid Derivatives ............ 45
Diffusion of Phenylalkylamines ...................... 46
Diffusion of Dextromethorphan and Progesterone ...... 46
Diffusion of Drugs Through Silastic Capsules ........ 47

RESULTS

Screening of Permeability of Polymer Films .......... 50











TABLE OF CONTENTS (Continued)


Page

Solubility Studies .................................. 51
Partition Coefficients............................... 51
Thickness of Membranes .............................. 52
Permeability of Silastic Membranes to Phosphate
Buffer Salts and Hydrochloric Acid Solution........... 52
Treatment of Data ................................... 53
Diffusion of 4'-Aminopropiophenone Through Silastic
Membrane ..................................... .... ... 54
Diffusion of 4'-Aminoacetophenone and 3'-Amino-
acetophenone ........................................ 61
Diffusion of Barbituric Acid Derivatives ............ 62
Diffusion of Phenylalkylamines ...................... 62
Diffusion of Dextromethorphan and Progesterone ...... 63
Diffusion of Drugs from Silastic Capsules ........... 64

DISCUSSION

Permeability of Silastic Membrane ................... 65
Fick's Law of Diffusion ............................. 66
Effect of pH on Diffusion ........................... 67
Apparent Diffusion Constants from Steady
and Quasi-steady State Diffusion Experiments ........ 67
Non-dependence of Apparent Diffusion Constants of
Drugs on Molecular Weights .......................... 68
Postulated Mechanism of Transport in Silastic
Membrane ............................................ 69
Apparent Diffusion Constants and Partition
Coefficients ........................................ 70
Apparent Diffusion Constants and Solubilities ....... 71
Effect of Ethanol on Diffusion of 4'-Amino-
propiophenone through Silastic Membrane ............. 72
Temperature Dependence of Diffusion in Silastic
Membrane ............................................ 74
Pharmaceutical Application .......................... 76
Control and Prediction of Diffusion of Drugs
through Silastic Membranes .......................... 78

SUMMARY AND CONCLUSIONS ....................................... 80

APPENDIX

A. Tables .......................................... 84
B. Figures ......................................... 102
C. Derivations and Statistical Analyses ............ 127










TABLE OF CONTENTS (Continued)


Page

REFERENCES ...................................... ........ 145

BIOGRAPHICAL SKETCH ........................................... 151















LIST OF TABLES


Table Page

I WAVELENGTHS OF MEASURED ABSORBANCES, MOLAR
ABSORPTIVITIES A:. pKa VALUES OF C ,P:.. .,NDS .......... 85

II DIFFUSION OF 4'-AMINCPROPIOPHENONE FROM SATURATED
SOLUTION (2.36 x 10 M) IN pH 6.5 PHOSPHATE BUFFER
THROUGH 3 MIL SILASTIC MEMBRANE INTO 200 ML. of
0.12 N HCI AT 37.30 ................................. 86

III SOLUBILITIES, PARTITION COEFFICIENTS, AND APPF'PE;T
DIFFUSION CONSTANTS OF BARBITURIC ACID DERIVATIVES
AT 25.00 ............................................ 87

IV PARTITION COEFFICIENTS, SOLUBILITIES, ACTIVATION
E;:KRGIES OF DIFFUSION, AND APPARENT DIFFUSION
CONSTANTS OF AMINOALKYLPHENONES ..................... 88

V SOLUBILITY, AND APPARENT DIFFUSION CONSTANTS OF
4'-AMINOPROPIOPHENONE AS A FUNCTION OF ETHANOL
CONCENTRATION AT 25.00 .... .......................... 89

VI SPECIFIC RATES OF DIFFUSION OF 4'-AMINOPROPIOPHENONE
AS A FL..CTION OF :I:.RANE THICKNESS AT 24.900 ....... 90

VII SPECIFIC RATES OF DIFFUSION, AND APPARENT DIFFUSIC;;
CONSTANTS OF 4'-AMINOPROPIOPHENONE AS FUNCTION OF
E;.IE:.ATURE .:D CONCENTRATION ....................... 91

VI I APPARENT DIFFUSION CONSTANTS OF 4'-AMINOPROPI CPHEL;ONE
AS A FUNCTION OF pH AT 25.00 ........................ 92

IX APPARENT DIFFUSION CONSTANTS, AND SPECIFIC RATES
OF DIFFUSION OF 4'-AMINOPROPIOPHENONE AS A FUNCTION
OF ETHANOL CONCENTRATION AT 24.600 .................. 93

X CONSTANTS OF QUASI-STEADY STATE DIFFUSION OF 4'-
AMINOPROPICPZPH.:NE THROUGH SILASTIC MEMBRANE AS A
FUNCTION OF ETHANOL CO;IENTRATION AT 25.250 ......... 94

XI SPECIFIC RATES OF DIFFUSION OF ETHANOL AS A FL.:.TION
OF 4'-AMINOPROPIC.-.NONE CONCENTRATION AT 24.600 .... 95


viii












LIST OF TABLES (Continued)


Table Page

XII APPARENT ACTIVATION ENERGIES OF DIFFUSION, AND
APPARENT DIFFUSION CONSTANTS OF BARBITURIC ACID
DERIVATIVES ...................................... ... 96

XIII APPARENT DIFFUSION CONSTANTS OF BARBITAL AND
PENTOBARBITAL THROUGH SILASTIC MEMBRANE AT 37.30
AS A FUNCTION OF pH ............................... .. 97

XIV APPARENT DIFFUSION CONSTANTS, AND SPECIFIC RATES
OF DIFFUSION OF PHENYLALKYLAMINES THROUGH SILASTIC
'.oR.:ANE AT 25.00 FROM BORATE BUFFER SOLUTIONS INTO
200 ML. of pH 6.8 PHOSPHATE BUFFER .................. 98

XV APPARENT DIFFUSION CONSTNATS, AND RATES OF DIFFUSION
OF DEXTROMETHORPHAN FROM PE/,!T OIL, MINERAL OIL AND
pH 10.1 BORATE BUFFER THROUGH 3 MIL SILASTIC MEMBRANE
INTO pH 6.8 PHOSPHATE BUFFER AT 37.50 ............... 99

XVI APPARENT DIFFUSION CONSTANTS OF PROGESTERI FROM
PEA 5IT OIL, A 1~ pH 6.8 PHOSPHATE BUFFER T-:jLCH 3
MIL SILASTIC :.:EL:EANE INTO pH 6.8 PHOSPHATE BUFFER
AT 37.50 ............................................ 100

XVII APPARENT DIFFUSION CONSTANTS AND RATES OF DIFFUSION
OF DRUGS FROM SILASTIC CAPSULES AT 37.50 ............ 101
















LIST OF FIGURES


Figure Page

1. Typical calibration curves for spectrophotometric
analyses of various drugs at their pertinent wave-
lengths .............................................. 103

2. Sketch of an assembled vial for screening polymeric
films for their permeability to drugs ................ 104

3. Sketch of diffusion cell used for steady state
diffusion ............................................ 105

4. Sketch of diffusion cell used for quasi-steady state
diffusion ............................................ 106

5. Sketch of diffusion apparatus and pump assembly for
steady state diffusion studies ....................... 107

6. Solubility of 4'-aminopropiophenone in phosphate
buffer containing various concentrations of ethanol
at 25.00 plotted as absorbance of the filtered
solutions at 307 mp versus percentage of ethanol
in phosphate buffer .................................. 108

7. Diffusion of 4'-aminopropiophenone (1.52 X 10-3 M)
from pH 6.8 phosphate buffer through Silastic
membranes of different thicknesses at 24.900 ......... 109

8. Effect of thickness of Silastic membrane on the
specific rate of diffusion (rate of diffusion/concen-
tration x area) of 4'-aminopropiophenone from
phosphate buffer solutions at 24.900 ................. 110

9. Effect of concentration of 4'-aminopropiophenone in
phosphate buffer solution on the rate of diffusion
of the drug through 3 mil Silastic membrane at
24.900, 33.600, and 41.00 ............................ 111

10. Arrhenius plot for the apparent diffusion constants
of aminoalkylphenones through 3 mil Silastic membrane
from phosphate buffer solutions ...................... 112











LIST OF FIGURES (Continued)


Figure Page

11. A plot of log (C2 CI/CO) versus time for quasi-
steady state diffusion of 4'-aminopropiophenone from
phosphate buffer through 5 mil Silastic membrane at
25.00 ................................................ 113

12. Apparent diffusion constants of 4'-aminopropiophenone
through 3 mil Silastic membrane at 24.900 as a
function of pH of the drug solutions ................. 114

13. Effect of ethanol on the rate of diffusion of 4'-
aminopropiophenone through 3 mil Silastic membrane
from solutions in phosphate buffer containing varying
percentages of ethanol at 25.00 ...................... 115

14. Diffusion of barbituric acid derivatives through 3
mil Silastic membrane from solutions in pH 4.7
acetate buffer at 37.30 .............................. 116

15. Diffusion of barbituric acid derivatives through 3
mil Silastic membrane from solutions in pH 4.7
acetate buffer at 37.30 .............................. 117

16. Apparent diffusion constants of barbital and pento-
barbital through 3 mil Silastic membrane as a
function of the pH of the drug solutions ............. 118

17. Arrhenius plots for the apparent diffusion constants
of barbituric acid derivatives from solutions in pH
4.7 acetate buffer through 3 mil Silastic membrane ... 119

18. Plots for the diffusion of phenylalkylamines from
solutions in borate buffer through 3 mil Silastic
membrane at 25.0 ..................................... 120

19. Plots for the diffusion of dextromethorphan from
solutions in pH 10.1 borate buffer, peanut oil, and
mineral oil through 3 mil Silastic membrane at 37.50 121

20. Plots for the diffusion of progesterone from
solutions in phosphate buffer and peanut oil through
3 mil Silastic membrane at 37.50 .................... 122

21. Plots of the apparent diffusion constants of
barbituric acid derivatives through 3 mil Silastic
membrane at 25.00 versus partition coefficients for
the compounds between acetate buffer solutions and
chloroform............................................ 123










LIST OF FIGURES (Continued)


Figure Page

22. Plot of the apparent diffusion constants of amino-
alkylphenones through 3 mil Silastic membrane at
25.00 versus the partition coefficients for the
compounds between solutions in phosphate buffer
and chloroform ....................................... 124

23. Plot of the apparent diffusion constants of
barbituric acid derivatives through 3 mil Silastic
membrane at 25.00 versus reciprocal of the solubility
of the compounds in pH 4.7 acetate buffer solutions
at 25.00 ............................................. 125

24. Plot of the apparent diffusion constants of 4'-amino-
propiophenone from solutions in phosphate buffer
containing varying percentages of ethanol versus the
reciprocal of solubility in the respective buffer
solutions ............................................ 126
















LIST OF APPENDICES



Appendix Page

C I Derivation of Sutherland and Einstein Equation ..... 128

C II Derivation of Fick's Second Law of Diffusion ....... 130

C III Partial Derivation of Equation for Calculation of
Diffusion Constants by Time-lag Method ............. 131

C IV Derivation of Equations for Quasi-steady State
Diffusion .......................................... 133

C V Berthier Method for Calculation of Diffusion
Constants .......................................... 136

C VI Statistical Analysis of Thickness Measurements
Obtained on 3 mil Silastic Membranes ............... 138

C VII Regression Analysis of Raw Data for Diffusion of
4'-Aminopropiophenone .............................. 139

C VIII Analysis of Variance of Data Collected to Show the
Reproducibility of Rates of Diffusion of 4'-Amino-
propiophenone through 3 mil Silastic Membranes ..... 140

C IX Effect of Ionic Strength on Diffusion of 4'-Amino-
propiophenone through 3 mil Silastic Membrane into
200 ml. of 0.12 N HCI at 23.00 as a Function of
lonic Strength of Drug Solution in Phosphate Buffer. 143

C X Specific Rates of Diffusion of 4'-Aminopropio-
phenone Obtained as a Function of Hydrostatic
Pressure Exerted on Silastic Membrane at 25.00...... 144


x iii
















INTRODUCTION


The differential transport of substances in solution through

membranes has been used for separations based on differences in

molecular weights (1-6). The selectivity of membranes to penetrants

has been used to enrich or separate mixtures of gases (7, 8, 9) based

on the basic information obtained'on the transport of gases and water

vapors through membranes (10). The models of transport processes in

the artificial membranes are being used to explain the possible

mechanisms for the passage of nutrients and drugs across a succession

of membranes in the living organism (11-15).

However, fundamental studies on the diffusion of drugs

through artificial membranes to formulate optimal dosage forms are

scarce and mostly qualitative in nature (16, 17, 18). Basic and

quantitative information is needed to predict the transference of

drugs in solution through membranes to provide proper dosage forms

based on known values of release of drugs through these membranes.

The purposes of this investigation were to determine and

quantify the basic factors that influence the diffusion of drugs

through synthetic polymeric membranes. The rates of diffusion were

to be correlated with all measurable physical and chemical para-

meters that could be used to predict the diffusivities of drugs.

Ultimately it was planned to test the predicted in vitro diffusion

of drugs from pharmaceutically proper dosage forms.
















HISTORICAL


A membrane is an imperfect barrier separating two fluids,

whether gases or liquids. Membrane technology (19) and the applica-

tion of polymeric materials in the medical and health related

professions (20, 21) have been discussed in recent fine reviews.

Diffusion is defined as the tendency for molecules to migrate

from region of high concentration to a region of lower concentration

and is a direct result of molecular movement. Dialysis is a term

applied to the use of membranes for the separation of particles of

colloidal dimensions from the molecules of suspending liquid and is

a consequence of diffusion (22).

Dialysis, as observed in transport through cellophane and

animal membranes and parchment papers (23, 24, 25), is due to sieve

action. Such membranes may be considered as heterogeneous barriers

in the sense that they possess pores. The transport is generally a

measure of the probability of a solvated molecule entering and diffusing

through the pores. There is little selectivity in the separation of

two closely related molecules except when their size is approximately

that of the size of the pore (26). In general, the solvent as well

as the solute is transported; membranes which allow salt transport

are permeable to water (27).

Dialysis, where a membrane acts as a barrier to free diffusion

of a substance in an isotropic medium, is largely dependent on the










molecular weight of the diffusate and the viscosity of the solvent.

The Sutherland and Einstein equation (28) for spherical colloidal

particles is (Appendix C-1)


D = (RT/6/rn N) (4TIN/3 M v)1/3 (Eq. 1)


where D is the diffusion coefficient, R is the molar gas constant, T

is the absolute temperature, n is the viscosity of the solvent, N is

the Avogadro's number, M is the molecular weight and 7 is the partial

specific volume of the solute. An empirical relation correlating

diffusion constant, molecular weight and viscosity is (29)


D = 7.4 X 10-8 (XM)05 T/n 7 0.6 (Eq. 2)


where X is an association parameter defining the effective molecular

weight of the solvent with respect to the diffusing species (for water

X = 2.6).

Dialysis has been used for the separation of the components

of blood (26, 30), the transfer of macromolecules in an artificial

kidney (31), the fractionation of high polymers (25), the correlation

of molecular size and structure with transport rates (1-6.) and the

determination of the binding of drugs and chemicals to proteins and

macromolecules (32). Ion-exchange resins have been used extensively

for the separation of charged particles or molecules (33-37). Electro-

dialysis, where electromotive force is used as the driving force for

the separation of ionic solutes, has been used for the recovery of

salts from sea water (38, 39, 40), the recovery of acids from spent

acid solutions (41), and the partial demineralization of milk and

whey (42).










Many polymeric membranes such as copolymers based on poly-

oxyethylene glycols and polyethylene terephthalate act as homogeneous

barriers (26). Transport is generally dependent on the relative adsorp-

tion of the molecules diffusing to the face of the membrane and

solubility of these molecules in the membrane (26). The selectivity

of polymeric membranes such as polystyrene (43) and ethyl cellulose

(44) has been used for the enrichment of air with respect to oxygen

and attempts to make these processes industrially feasible have been

reported (45). The diffusion of gases in polymers follows the

Arrhenius equation (46) and the temperature dependence of diffusion

is given by


D = D, e -Ea/RT (Eq. 3)


where Do is a constant, and AEa is the apparent activation energy for

diffusion.

The methods to determine the diffusivity of water vapors

through membranes have been considered in great detail (10). The

uptake of moisture by hygroscopic substances could not be prevented

by encapsulating them in gelatin capsules (47). /The permeation of

water vapors decreased with increased chain length of the acid moiety

of cellulose ester membranes (48). The water vapor transmission

initially decreased and then increased asthe concentration of the

plasticizers in polymer was gradually increased (49). The solubility

of water and the Arrhenius parameters for the diffusion of water in

a polymer differ above and below the temperature at which the slope

of the volume-temperature curve for the polymer changes (50)--the

"glass temperature" (51).






5


The extensive use of polymeric materials in the pharmaceutical

industry for packaging purposes has initiated the investigation of

polymer-drug interactions. Kapadia et al. (52) showed interaction

between salicylic acid and Nylon 66 and calculated the heat of sorption

from equilibrium sorption studies at several temperatures using the

van't Hoff equation (53). The heat of sorption for this and in the

subsequent work with other weak organic acids(54, 55) was low (1-

4 Kcal./mole). The magnitude of the sorption decreased with decreases

in the polarity of the solvent. The pH-sorption studies implied inter-

action of tne unionized acids with the basic group in the polyamide.

The diffusion within the polymeric material was shown to be the rate

determining step with the heats of activation in the range of 10 20

Kcal./mole. Studies on the sorbic acid-Nylon interactions (56-60)

confirmed the results obtained with weak organic acids.

One of the possible important uses of polymeric materials in

the pharmaceutical industry is as coating material for sustained

release products (61 -64). Vinyl, acrylic and cellulosic polymers

were shown to be good for prolonged action coatings based on their

solubility in simulated gastric and intestinal fluids (62). Copolymer-

coated prednisolone tablets extended absorption of prednisolone over a

period of 10 12 hours in intact dogs and in segments of intestinal

tracts (63). Silicone rubber which has been used extensively in the

subcutaneous prosthetic devices (65) has been shown to be permeable

to steroids (16), cardiac pacers (66) and some other materials (17, 18).

Lyman and coworkers (26, 67, 68) have prepared synthetic

membranes with the express intent of endowing them with specific

characteristics which would transfer substances by an adsorption and






6


solubility mechanism and not by sieve action. When the weight percent

of polyoxyethylene glycol, the hydrophilic monomer in the copolyether-

ester membranes was increased, the rates of transfer of glucose and

urea increased. The magnitude of the increase in rates was different

for the two compounds and the authors postulated a different degree of

association or partitioning with the membrane (30).










MATHEMATICS OF TRANSPORT PROCESSES


The mathematics of transport processes depends upon the model

chosen for the particular transport phenomenon under consideration.

This has been discussed in detail by Tuwiner (69), Jost (70), Crank

(71) and Lakshminarayaniah (72). Higuchi and Higucn; (73) have giv '-

a theoretical analysis of diffusional movement through heterogeneous

barriers.

A transport process is considered to be in the steady state

when the amount of penetrant passing through a reference point in a

membrane matrix is invariant with time. When the amount passing

.'l-ruh a reference point varies with time, the process is said to be

in the non-steady state.

When a membrane is interposed between a solution of the

penetrant and a solvent, the penetrant is transported initially by

a non-steady state process. This ag period" continues until the

amount leaving the membrane is equal to the amount entering. A steady

state results when the concentrations of the solutions on either side

of the membrane are kept constant. However, if the concentration of

the solutions are allowed to equilibrate, the amount entering and

leaving may be equal for all analytical purposes even though the rate

of permeation will be changing with time as a function of the concen-

tration gradient. This may be termed a "quasi-steady state" transport.

The rate determining factor in non-steady state transport

through the membrane is the rate of diffusion in the membrane; in steady

state transport it is the constant concentration gradient alone; and

in quasi-steady transport both the concentration gradiert and the rates










of approach to equilibrium of both extra-membrane phases are rate

determining. Fick's first law of diffusion (74) states that the

rate of diffusion is proportional to the concentration gradient,


dA/dt = DSdC/dx


(Eq. 4)


where A is the amount of penetrant in moles diffusing in time t in

seconds through a membrane having a surface area of S cm.2. The

concentration gradient is dC/dx across the membrane in moles per

liter-cm. and D is the diffusion coefficient in cm.2/sec. The term

x is the distance into the membrane in cm. The concentration of the

penetrant at a particular position in a polymer at a given time is

given by Fick's second law of diffusion


dC/dt = D S d2C/dx2


(Eq. 5)


where d2C/dx2 is the change in the concentration gradient as a function

of the distance x within the membrane. The derivation of Fick's second

law from Fick's first law is given in Appendix C-Il.

For a steady state condition, the change in the concentration

at any point in the membrane is zero.


dC/dt = 0

D S d2C/dx2 = 0


Thus


(Eq. 6)

(Eq. 7)


and the concentration gradient is a constant


dC/dx = K


It follows from Eqs. 4 and 8 that

dA/dt = D S K = D S (C2 C1)/X


(Eq. 8)


(Eq. 9)









where C2 and C1 are the invariant concentrations of the concentrated

and dilute solutions in contact with the membrane surfaces respectively

and X is the thickness of the membrane. This states that the rate of

diffusion through a membrane of thickness X and surface area S is

constant with time for a constant concentration gradient. Thus the

total amount diffused can be plotted against time to obtain a straight

line with slope of D S K from which the diffusion constant D can be

calculated when S, C2, C1 and X are known.

A more rigorous mathematical treatment considers the concen-

trations of the penetrant at the membrane surfaces or in the first

monolayer of the membrane material, instead of the concentrations in

the solutions for the concentration gradient in Eq. 9. This equation

may be written as


dA/dt = D' S (Cm2 Cml)/X (Eq. 10)


where Cm2 and Cml are the concentrations of penetrant at the two

membrane surfaces and D' is the intrinsic diffusion constant. The

concentrations are related to the activity of the penetrant in the

membrane as


Cm = ami/m and Cm2 = am2/xm (Eq. 11)


where aml and am2 are the activities of the penetrant at the two surfaces

and m is the activity coefficient of the penetrant in the membrane

material.

The relation between activities and concentration of penetrants

in the solution is given by


asl = CIYs and as2 = C2~s


(Eq. 12)









where asl and as2 are the activities of the penetrant in the solutions

and s is the activity coefficient of the penetrant in the solution.

Assuming a rapid equilibration of penetrant between solutions and

membrane material at the surfaces


as1 = aml and as2 = am2 (Eq. 13)


From Eqs. 11, 12, and 13 it follows that


Cml = Ci 's/Ym and Cm2 = C2 Ys/Am (Eq. 14)


When these values of Cm1 and Cm2 are substituted in Eq. 10,

the following expression is obtained


dA/dt = (D' S s/X Ym) (C2 C1) (Eq. 15)


Since the ratio of activity coefficients is the partition

coefficient Kp


Is / m = Kp (Eq. 16)


Eq. 15 may be rewritten as


dA/dt = (D' SKp/X) (C2 C1) (Eq. 17)


The product D'Kp is the term D in the original Fick's

diffusion equation (74) and is termed the permeability constant for the

system.

The time lag method used extensively (46, 75, 76) for the

calculation of diffusion constant in the non-steady state is based on

the premise that a finite amount of time will be needed for a penetrant

to traverse the thickness of the membrane before the attainment of a









steady state (76). After this initial lag period, when the concen-

trations on both sides of the membrane are constant, the plot of the

amount diffused versus time will be a straight line (Eq. 17) which

when extrapolated will give an intercept on the time axis.


0 = X2 / 6 D' (Eq. 18)


The mathematical basis for this method is


A = D S C2 (t X2 / 6 D)/ X (Eq. 19)


derived partially in Appendix C-Ill. This equation may be written in

terms of the partition coefficient Kp and the intrinsic diffusion

coefficient D' as (73)


A = (D' S Kp C2 / X) (t X2 / 6 D') (Eq. 20)


Thus the slope of the plot of the amount diffused versus

time, for a constant concentration gradient in the membrane, will be

D' S Kp C2/X. The Eq. 20 is the more rigorous version of Eq. 17 with

C1 = 0.

In the equilibrium diffusion experiments, the concentration

gradient across the membrane decreases to zero (77). The mathematical

expression derived for this quasi-steady state is given in Appendix C-

IV and is (78)


[XV1V2/S(V1 + V2)] In (C2 C1)/Co) = -Dt (Eq. 21)


where VI and V2 are the volumes of the solutions in the compartments

with molar concentrations CI and C2. CO is the concentration in one

compartment at t = 0. When the volumes V in the two compartments






12


are equal Eq. 21 simplifies to


(XV/0.869 S) log (C2 CI/Co) = -Dt (Eq. 22)


An alternate method is the method of Berthier (79) wherein

the fractional uptake has been used for the calculation of diffusion

constants (52). This method has been explained in Appendix C-V.















EXPERIMENTAL


Materials


Polymer Films:

RILSAN (Nylon 11) (May Industries Inc., Atlanta, Ga.) is 11-amino

undecanoic acid polymer (80).

POLYPENCO (Nylon 101) (The Polymer Corporation, POLYPENCO Division,

Reading, Pa.) is a polyamide (81).

Cellulose Acetate (KODACEL A 29), Cellulose Triacetate (KODACEL TA 401),

and Cellulose Acetate Butyrate Sheets (KODACEL B 298) (Eastman Chemical

Products, Inc., Kingsport, Tennessee) are thermoplastic cellulosic

films (82).

Polyethylene Type B, Mylar Polyester Type S (E.I. DuPont De Nemours

and Co., Inc., Wilmington, Delware). Mylar Polyester is polyethylene

terephthalate (83, 84).

Polypropylene (Avisun Corporation, 215, 12th Street, Philadelphia, Pa.)

(85).

Silastic Medical Grade Sheeting (H-0169, H-0293) (Dow Corning Center

for Aid to Medical Research, Midland, Michigan) is a dimethylsiloxane

polymer (86).


Chemicals:

The following chemicals were supplied by Abbott Laboratories, North

Chicago, Illinois.









Amobarbital Equivalent weight.-Calculated for C11H18N203: 226.28.

Found: 236.64.

Barbital Equivalent weight.--Calculated for C8H12N203: 184.20.

Found: 197.03.

Butabarbital Equivalent weight.--Calculated for C10H16N203: 212.23.

Found: 215.15.

Cyclobarbital Equivalent weight.--Calculated for C12H16N203: 236.26.

Found: 245.20.

Diallylbarbituric Acid Equivalent weight.--Calculated for C10H12N203:

208.21. Found: 201.72.

Mephobarbital Equivalent weight.--Calculated for C13H14N203: 246.26.

Found: 250.0.

Metharbital Equivalent weight.-Calculated for C9H14N203: 198.23.

Found: 200.92. *

Pentobarbital Equivalent weight.--Calculated for C11H18N203: 226.26.

Found: 231.5.

Phenobarbital Equivalent weight.-Calculated for C12H12N203: 232.24.

Found: 256.76.

Secobarbital Equivalent weight.--Calculated for C12H18N203: 238.29.

Found: 229.55.

Thiamylal Equivalent weight.--Calculated for C12H17N202S: 254.34.

Found: 289.80.

Thiopental sodium Equivalent weight.-Calculated forC11H18N202S:

242.33. Found: 246.30.










The following chemicals were supplied by Smith Kline and French

Laboratories, Philadelphia, Pa.

a-methyIphenethylamine hydrochloride Equivalent weight.--Calculated

for CgH13N.HCI: 171.5. Found: 165.3.

a-ethylphenethylamine hydrochloride Equivalent weight.--Calculated*

for C10H15N.HCI: 185.5. Found: 203.1.

2-amino-4-methyl-4-phenylpentane hydrochloride Equivalent weight.-

Calculated for C12HI9N.HCI: 213.5. Found: 205.31.

3-amino-1-phenylbutane sulfate Equivalent weight.--Calculated for

(C10H15N) *H2SO4: 198. Found: 242.13.

1-methyl-5-phenylpentylamine hydrochloride Equivalent weight.-

Calculated for C12H19N-HCl: 213.5. Found: 355.93.


The following compounds were purchased from Eastman Organic Chemicals,

Rochester, 3, New York.

4'-Aminopropiophenone Melting point 139 140.50; literature value

1400 (87).

4'-Amir:i..: t.:then:.r, Melting point 105 1060; literature value

1060 (87).

3'-Aminoacetophenone Melting point 97 990; literature value 990

(87).


The following compounds were supplied by The Upjohn Company, Kalamazoo,

Michigan.

Progesterone Melting point 128 1300; literature value 127 1310

(87).

Cortisone Melting point 230 2360; literature value 236 2400 (87).

\






16


Hydrocortisone Melting point 217 2200; literature value 217 2200

(87).

Prednisolone Melting point 239 2400; literature value 240 2410

(87).


The following compounds were purchased from National Biochemical

Corporation, Cleveland, Ohio.

Sulfathiazole Melting point 200 203; literature value 200 -

(87).

Sulfisoxazole Melting point 195 1980; literature value i940

Sulfabenzamide Melting point 181 1830; literature value 181.

183.30 (87).

Sulfadiazine Melting point 256 2570; literature value 252 -

(87).


2040



(87).

2 -


2560


Tetraethylthiuram disulfide (Disulfiram) (Ayerst Laboratories, Incor-

porated, New York, N. Y.) Melting point 70 720; literature value

700 (87).

Dextromethorphan (Vick Divisions Research and Development, Richardon-

Merrell Inc., Mt. Vernon, New York).

Peanut oil and Mineral oil used were of United States Pharmacopoecial

grade.
















Methods of Analysis


All compounds used in this investigation except ethanol were

analyzed spectrophotometrically by the Cary Model 15 dual beam recording

spectrophotometer, Beckman Model DU spectrophotometer or Beckman Model

DU-2 spectrophotometer. Standard, square, silica cells of 10 mm.

light path (Sargent /# S-75730) were used. All spectrophotometric

measurements were made at 24.0 1.00.

Spectrophotometric measurement of absorbances was used for

the quantitative estimation of barbituric acid derivatives, amino-

alkylphenones, progesterone, phenylalkylamines and dextromethorphan.

The linear relationship between the absorbance of drug solutions and

their concentrations in accordance with Lambert-Beer's law (88) was

verified in all cases. A few representative calibration curves are

shown in Fig. 1.

The absorbances of solutions of barbituric acid derivatives

were measured in pH 10.1 borate buffer as the molar absorptivities

(C values) of their anionic forms are much higher than those in the

uncharged form. Absorbances of all barbituric acid derivatives except

thiamylal and thiopental were measured at a wavelength of 238 mp.

The wavelength of 238 ml is not the /max of all these compounds.

However, measurements at one wavelength were found to be convenient

and time saving. The absorbances of thiamylal and thiopental were

measured at their max of 304 mL. The wavelengths of absorbance









measurement and the E values at these wavelengths are recorded in

Table I.

The absorbances of dextromethorphan, progesterone and amino-

alkylphenones were measured in pH 6.8 phosphate buffer solutions. The

wavelengths at which the absorbances were measured and the values

at these wavelengths are shown in Table I.

The phenylaklylamines-a-methylphenethylamine, a-ethyl-

phenethylamine, 1-methyl-5-phenylpentylamine, 3-amino-l-phenyl-butane

and 2-amino-4-methyl-4-phenylpentane were analyzed colorimetrically

by the method of Gettler and Sunshine (89) modified in the following

manner: A 1.00 ml. sample of phenylalkylamine (1.5 X 10-5 1.0 X

10-4 M) in pH 6.8 phosphate buffer was transferred to a 6" x 5/8"

pyrex test tube. The solution was made alkaline with .0.1 ml. solution

of 2 N NaOH and was mixed on a Vortex Jr. Mixer. Five ml. of chloro-

form (A. R. Grade) was then added and mixed on the Vortex Jr. Mixer

for about 1 minute. The solution was centrifuged at about 3,200

r.p.m. for 3 minutes. A 4 ml. pipette was inserted through the

aqueous layer into the chloroform layer blowing lightly through the

pipette to avoid entry of the aqueous solution. Four ml. of the

chloroform was removed and transferred into a test tube of similar

dimensions. Two-tenth ml. of freshly prepared methyl orange reagent

(equal volumes of saturated solutions of methyl orange and boric

acid in water) was added and mixed for about one minute. This solu-

tion was centrifuged at 3200 r.p.m. for 3 minutes and 3 ml. of the

chloroform layer was carefully pipetted into another test tube. Then

0.2 ml. of absolute ethanol containing 2% concentrated sulfuric acid

was added. The solution was mixed and then transferred to a cuvette









for the measurement of absorbance against the chloroform layer obtained

from a phosphate buffer blank treated identically. The absorbance was

measured on a Beckman DU spectrophotometer at 520 mp. A five point

calibration curve was prepared each day the samples from the diffusion

were analyzed.

The remaining compounds cortisone, hydrocortisone, predniso-

lone, sulfadiazine, sulfathiazole, sulfisoxazole, sulfabenzamide and

disulfiram were used only in the screening part of this investigation.

Spectrophotometric methods were used only for their qualitative

analysis and no atTempt was made to verify the absorbance concentra-

tion linearity.

Since the Xmax and c values of all steroids were independent

of the pH of the solution, the spectra of their solutions were obtained

without any adjustment of pH.

The solutions of sulfadiazine, sulfathiazole, sulfisoxazole

and sulfabenzamide were made alkaline with NaOH solution before

obtaining their U.V. spectra on the Cary spectrophotometer. The \ max

and the 6 values of these compounds are given in Table 1.

Disulfiram was qualitatively analyzed by a method involving

chelation with copper ion (90). Five ml. of 0.02 M CuSO4 solution in

pH 6.5 phosphate buffer was added to 10 ml. of solution of disulfiram

in phosphate buffer. The solution was mixed on Vortex Jr. Mixer. The

copper complex formed was extracted into 5 ml. of ethylene dichloride

with vigorous mixing on the Vortex Jr. Mixer for 1 minute. The

ethylene dichloride layer was separated and its absorbance was measured

against the ethylene dichioride layer obtained from identical treat-

ment of a phosphate buffer blank. The disulfiram-copper complex showed

two peaks at wavelengths of 272 and 285 mu.









Ethanol was quantitatively determined using vapor phase

chromatography. An F & M Model 700 Gas Chromatograph with flame

ionization detector was used with a 4' x 1/4" o.d. stainless steel

column packed with 2~:' Carbowax 20 M. on 60-80 mesh Chromosorb W,

worked isothermally at 50.00. The temperature of the detector was 250.00,

and that of injection port was 150.00. The carrier gas, helium, was

used at a pressure of 30 psig. while hydrogen and air for the flame

were used at 10 and 40 psig. respectively. Samples of five microliters

of aqueous ethanol were injected without any prior treatment. The

peak heights were measured and the concentration of the ethanol in

the sample was obtained from a calibration curve prepared with ethanol

solutions of known concentrations.

Determination of pKa.-The pKa is the negative logarithm of

the acid dissociation constant. The pKa's of barbituric acid deri-

vatives, dextromethorphan and phenylalkylamines were determined by

potentiometric titration. The pKa of 4'-aminppropiophenone was

determined spectrophotometrically.

The potentiometric titrations were performed with a Sargent

Model D automatic titrator. The titrator was equipped with a syringe

with a titrant capacity of 2.5 ml. The pH scale of the titrator was

standardized with two of the pH 4.0, 7.0 and 10.0 standard Beckman

buffer solutions (Beckman Instrument, Inc., Fullerton, California)

bounding the pH range to be titrated. The accuracy of the pH measure-

ment was 0.05 pH unit. All titrations were performed at 24.0 1.00.

The barbituric acid derivatives were dissolved in 2 ml. of

0.04 N NaOH and the solution was diiuted with freshly boiled distill ed

water. A 20 ml. a iquot of this solution was titrated with 0.1 N HC!04

on the titrator.









The acid salts of phenylalkylamines were dissolved in

distilled water to contain about 5 milliequivalent of accurately

weighed compound in 30 ml. of distilled water. The solutions were

titrated on the titrator against 0.1 N NaOH.

A weighed quantity of dextromethorphan was dissolved in

2 ml. of 0.01 N HCI and the solution was diluted to 40 ml. This

solution was titrated against 0.1 N NaOH. In all these titrations

an equal volume of the blank solution was titrated under identical

conditions. The calculated equivalent weights are given in the

section on materials.

The pKa of a compound was determined from these titration

curves by the method of Parke and Davis (91) wherein the difference

between the volumes of the titrant for attaining the same pH for the

sample and blank solutions was plotted against the pH. The pH value

corresponding to the midpoint of the resultant sigmoidal curve was

the half neutralization point or the pKa of the compound.

The pKa of 4'-aminopropiophenone was determined spectrophoto-

metrically. Hundred ml. of 8.56 X 10-5 M solution was used. The pH

of the solution was measured on the Beckman Expanded Scale pH meter

standardized with pH 4.0 and 7.0 standard Beckman buffer solutions.

An ultraviolet spectrum of this solution was obtained against a

distilled water blank on the Cary Recording spectrophotometer. The

sample solution in the cuvette was returned to the bulk solution. A

drop of concentrated HCI was added and the solution was stirred on a

magnetic stirrer. The pH of the solution was measured and a U.V.

spectrum was obtained again. Thus, U.V. spectra of the solutions were

obtained for several pH-values. The absorbance of th. sc .ons at a









wavelength of 307 mp was measured from recorded spectra and plotted

against the pH of the solutions. The pH value corresponding to the

midpoint of the resultant sigmoidal curve was the half-neutralization

point or pKa of 4'-aminopropiophenone. The pKa values of the compounds

are reported in Table I.

D :- r~-.i-', --I f 6 -.-A weighed amount of barbituric

acid derivative was dissolved in 2 ml. of 0.04 N NaOH solution and

diluted to a total volume of 25 ml. with freshly boiled distilled

water. Twenty ml. of the solution was pipetted into a 30 ml. beaker

and titrated against 0.10509 N HCI04 on the Sargent Model D automatic

titrator. The molarity of the compound present in the 20 ml. solution

was calculated from the volume and normality of HC104 used in the

titration. The remaining 5 ml. of the solution was diluted with pH

10.1 borate buffer solution to obtain 5 different concentrations of

the solution whose absorbances were then measured on The Beckman DU

spectrophotometer at wavelength of 238 mu for all the barbituric acid

derivatives except thiamylai and thiopental whose absorbances were

measured at 304 mi. The absorbances so obtained were then plotted

against the molarity of the solution and E values were calculated from

the slope by the equation (88)


S= Absorbance / Concentration (Eq. 23)


The compounds 4'-aminopropiophenone, 4'-aminoacetophenone, 3'-amino-

acetophenone, progesterone, dextromethorphan were weighed accurately

into a volumetric flask. These were dissolved in distilled water and

the absorbances of the adequately diluted soluTions were measured.

The C values were then calculated by Eq. 23. The C values are reported

in *- .'. '; I .









Screening of Polymer Films for the Permeability To Drugs


The polymer films Rilsan, Polypenco, cellulose acetate,

cellulose triacetate, cellulose acetate butyrate, polyethylene, Mylar

polyester, polypropylene and Silastic were used in thicknesses of 5,

5, 5, 3, 3, 3, 3, 5, and 3 mil respectively.

The membranes were washed in running tap water and then with

distilled water. These membranes were then dried and cut into small

pieces ("1 x -").

Saturated solutions of 4'-aminopropiophenone, 4'-aminoaceto-

phenone, 3'-aminoacetophenone, sulfadiazine, sulfathiazole, sulfabenza-

mide, sulfisoxazole, disulfiram, barbital, phenobarbital, progesterone,

cortisone, prednisolone in 0.1 N HCI, 0.1 N NaOH, pH 6.8 phosphate

buffer, propylene glycol, peanut oil, ethanol, mineral oil, ethylene

glycol and polyethylene glycol 200 were used.

Glass serum vials of 10 ml. capacity wiTh rubber stoppers and

aluminum caps were thoroughly cleaned and dried. A hole was bored

into the rubber stopper with a # 3 cork-borer. About 2 ml. of a

saturated solution of a drug under study was delivered into the serum

vial without touching its lip. The vial was stoppered with a rubber

stopper with a hole in it. A small piece of the membrane under study

was laid flat on the stopper to cover it completely and an aluminum

cap was then crimped on the rubber stopper so as to sandwich the

membrane between the two. The sketch of an assembled vial is shown

in Fig. 2. The detachable circular disc in the aluminum cap was

removed to expose the membrane. The serum vial was then inverted

and placed into a two oz. ointment jar conTaining about 10 ml. of pH 6.8









phospha e buffer. The jars were capped, properly labelled and set

aside for at least one week.

After this time the phosphate buffer solution in each jar

was qualitatively analyzed on a Cary Model 15 recording spectrophoto-

meter. In the cases of the barbituric acid derivatives and sulfonamides,

the phosphate buffer solutions from the jar were made alkaline before

reading Them on the spectrophotometer. Disulfiram was chelated with

copper and the chelate was extracted into ethylene dichloride. The

spectrum of the ethylene dichloride solution was then obtained on the

spectrophotometer.

A drug was said to have permeated through a membrane when the

peaks at wavelengths corresponding to the Xmax, values of the drug

were observed in the spectrum. Experiments with positive results were

repeated to avoid the acceptance of leaking vials as evidence for

permeation. Solvents without drugs were put in the vials to test

their permeability through the membranes and their effect on the

rubber stopper and the membranes.


Diffusion Apparatus


The diffusion apparatus consisted of diffusion cells, beakers,

Durrum 12-Channel Dial-A-Pump, and a stirring device.

Diffusion cells.--Two types of diffusion cells were designed

to study the steady state and quasi-steady state diffusion.

The steady state diffusion cell designed was a modification

of the cell used by Lyman et al. (68). An exploded schematic view of

the diffusion cell is shown in Fig. 3. It consisted of two stainless

steel ;iates (2 1/2" x 2" x 1/8"), a glass T joint, two silicone









rubber gaskets (2.57 cm. i.d. x 2.95 cm. o.d. x 0.16 cm. thick) and

a set of four stainless steel; nuts and bolts.

Each of the stainless steel plates had a hole in the center,

2.57 cm. in diameter. Around this hole there was a border 0.43 cm.

in width from the perimeter of the hole and recessed 2 mm. into one

face of the stainless steel plate. The silicone gaskets fit into the

recessed borders around the holes. The two plates had four holes in

four corners for the stainless steel bolts.

The glass T joint consisted of a 3 cm. long hollow glass

cylinder 2.60 cm. in diameter joined in its center at right angles on

one side to an 11 cm. long glass tube 0.8 cm. in diameter to form a

hollow T shaped device. The two annular edges of the glass cylinder

fitted into the recessed borders against the gaskets in the plates.

A piece of membrane under study was positioned between the

recessed border of each of the plates and a silicone rubber gasket.

The open ends of the glass cylinder were then fitted into the recessed

borders on the face of the silicone rubber gaskets. The whole

assembly was held in position by a set of four nuts and bolts

passing through the holes in the corners of the plates. The volume

of the assembled cell was approximately 22 ml. and the tot-i area of

the two membranes available for diffusion was 10.4 cm.2.

The quasi-steady state diffusion cell (see Fig. 4) consisted

of two stainless steel plates (2 ,/2" x 2 1/2" x 1/8"). Eac of these

plates was welded to one end of a stainless steel tube (1 1/2" in

diameter and 7" long) at an angle of 45. Each plate had a hole in

the center (3.1 cm. in diameter) with a circular border 0.8 cm. in

width from the perimeter of the hole and recessed 2 mm. into the face






26

of.the plate. The plates had four holes in the four corners for the

stainless steel bolts. Two silicone rubber ga(ets (3.1 cm. i.d. x

3.8 cm. o.d. x 3 mm. thick) fitted into the recessed border around the

holes in the plates. A membrane under study was clamped between the

two silicone rubber gaskets placed in the recessed borders in the

plates. The plates .were then clamped together with four stainless

steel nuts and bolts. The diffusion cell when assembled formed a V

shaped device and held 150 ml. of solution in each of the tubes. The

area of the membrane available for the diffusion was 7.55 cm.2

Beakers.--A 200 ml. or 400 ml. Pyrex beaker was used with

each steady state diffusion cell. A plastic cover was fabricated for

each beaker from a disposable polyethylene Petri dish. The rim of the

Petri dish was sawed off at one position to accommodate the beak of the

beaker. A hole, 1/2" in diameter, was burned into the center of the

cover with a hot glass rod. Another hole of the same size was made

in the cover about 1" away from the center.

Durrum 12-Channel Dial-A-Pump.--This pump (Durrum Instrument

Corporation, 925, E. Meadow Drive, Palo Alto, California) was used for

pumping a solution from a reservoir into the diffusion cell and back.

The pumped fluids entered the pumping unit by way of plug-in flexible

plastic tubing, received their pumping action by a flat pressure pate

vertically cycling against a series of independently, adjustable backing

blocks, and emerged by way of another plug-in connection similar to

input. The resilient pumping tubes used were amber latex surgical

tubing (Rubber Latex Products, Inc., Cuyaholga Falls, Ohio) 5/1"" o.d.

x 3/16" i.d. Each channel used two pieces of vinyl :;L.ng (Becton,

Dickinson and Company, Rutherford, N. J.) 3/16" o.d. x 1/16" i.d.









One piece of the tubing carried the solution from a container through

the plug-in connection into the pumping tube. The exit was connected

via a plug-in connection to another piece of tubing that delivered the

solution from the pumping tube to its destination.

Stirring device.--A thermostated shaker bath (American Instru-

ment Company, Silver Spring, Maryland and Eberbach Corporation, Ann

Arbor, Michigan) shook the beakers and agitated the solutions in the

beakersbesides keeping the temperature of the solutions in the diffu-

sion cell assembly constant within 0.50. The shaking rate of the

ban was 108 strokes per minute with each stroke traveling a distance

of 1 1/2".

Two other stirring devices tried were Mag-Jet stirrers

(Will Scientific, Atlanta, Georgia, Catalog # 25212) and nitrogen

bubbling. The Mag-Jet stirrer kept underneath a beaker was operated

by water pressure and rotated a magnetic stirring rod in the beaker

and another magnetic stirring rod in the diffusion cell kept in the

beaker. The speed of the Mag-Jet stirrer decreased slowly to zero in

5 to 6 hours probably due to deposition of salts in the small clearance

between the rotor and its metal casing.

Nitrogen under pressure was bubbled through the solutions in

the diffusion cell and the beaker, using narrow bore glass tubing.

It was observed that the solutions were stirred only at the spot of

the bubbling. The bubbling of the gas was, therefore, not useful for

stirring a large volume of the solution in the beaker. Also the

bubbling of the gas over a long period of time through the solution

caused evaporation of the solution. The volume of the solution in the

beaker was ai .'s cri-icai because the amount of drug diffusing through









the membrane was calculated from the concentration and the volume of

the beaker solution. When the solution used in the diffusion cell

was a saturated solution, the bubbling of the gas was used as a

stirring device. The volume of the solution in the diffusion cell

was not critical and the evaporation of the saturated solution did

not in any way affect the concentration of the drug in the solution.

The small volume of the saturated solution in the cell could there-

fore be well agitated by nitrogen gas bubbling method.

The Durrum 12-Channel Dial-A-Pump pumped the solution of a

drug in and out of the steady state diffusion cell and was observed

to be a good method of agitating the solution in the diffusion cell

(see Fig. 5). The drug solution from a reservoir was pumped into

the cell through one channel at the rate of 25-30 ml. per minute.

The end of the delivery tube from this channel reached the diametric

center of the horizontal portion of the glass T joint of the assembled

diffusion cell. Another tube carried the solution from the cell to

the channel which pumped the solution at the rate of 35-40 ml. per

minute. The end of the tube carrying the solution from the cell to

the pumping tube, was adjusted in the stem of the cell at the level

of the solution in the beaker. Since the channel carrying the solution

out of the cell pumped the solution at a rate faster than the other

channel pumping the solution into the cell, the level of the solution

in the cell was maintained at the level of the solution in the beaker.

The pulsating action of the pump kept the solution in the cell well

agitated besides renewing its contents every few minutes because of

the high turnover of the solution. The thermostated shaker bath held

the diffusion assembly rigidly in a metal rack fixed to the shaker









tray and agitated the solution in the beaker maintaining the tempera-

ture of the solutions constant.

The solutions in the arms of the quasi-steady state diffusion

cell were agitated by the thermostated shaker bath in which the cell

was wired to the shaker tray.


Solubility Studies


Saturated solutions of barbituric acid derivatives in pH 4.7

acetate buffer (p = 0.1) were prepared at 50.00. These solutions were

then allowed to cool to 25.00 in the thermostated shaker bath and to

equilibrate at that temperature for 48 hours in presence of excess

solids. Similarly saturated solutions of aminoalkylphenones were

prepared in pH 6.8 phosphate buffer (4 = 0.3) at 50.0 and equilibrated

to 37.50 in the thermostated shaker bath. The saturated solutions of

4'-aminopropiophenone in pH 6.8 phosphate buffer containing 0, 10, 20,

30 and 40% ethanol were prepared by vigorously shaking the solutions

at room temperature in presence of excess solids and allowed to equili-

brate at 25.00 in the thermostated shaker bath for 48 hours.

The solutions were filtered by suction through electrode

isolation tubes (E. H. Sargent & Co., 4647 West Foster Avenue, Chicago,

Illinois, Catalog # S-30417). The electrode isolation tube is a tube

fitted with a finely porous fritted glass membrane and is 125 mm. long

and 13 mm. in diameter in its lower section, 10 mm. in diameter in

upper section. Aliquots of the filtered solutions of barbituric acid

derivatives were appropriately diluted with pH 10.1 borate buffer.

Aliquots of filtered solutions of aminoalkylphenones were appropriately

diluted with pH 6.8 phosphate buffer. The absorbances of these diluted









solutions were measured on a spectrophotometer at the pertinent wave-

lengths reported in Table I. The solubilities of these compounds in

respective buffer solutions were then calculated from these absorbance

values and the knowledge of their E values reported in Table I.

The solubility values of 4'-aminopropiophenone in phosphate

buffer containing 0, 10, 20, 30 and 40% ethanol were plotted against

the concentration of ethanol. in the phosphate buffer solutions as

shown ;n Fig. 6. This curve was used as a calibration curve to obtain

the solubility of 4'-aminopropiophenone in phosphate buf.ier containing

7.5, 15, 22.5, 30 and 37.5% ethanol.


Determination of Partition Coefficients


The coefficients of partition between the solutions of

barbituric acid derivatives in acetate buffer, aminoalkylphenones in

phosphate buffer, and an organic liquid were determined at room tempera-

ture. The organic liquids used were chloroform, mineral oil, cyclo-

hexane and silicone liquid 200.

The aqueous media used for the preparation of the solutions

of these compounds and the organic liquids used for the partitioning

were saturated with respect to each other by shaking them together in

large quantity and then separating them, first by separatory funnel

and then by centrifugation at 3200 r.p.m. for five minutes.

An approximate 5 x 10-4 M solution of a barbituric acid

derivative was prepared in pH 4.7 acetate buffer presaturated with

chloroform. One ml. of this solution was diluted with 10 ml. of pH

10.1 borate buffer and its absorbance was measured on Beckman DU

spectrophotometer at wavelength reported in Table J. The blank

solution for the absorbance measurement consisted of acetate buffer









medium presaturated with chloroform and diluted 10 times with pH 10.1

borate buffer. The pH of the diluted solution was observed to be 10.1.

Generally the concentration of the compound in the solution used was

such that the diluted solution would have an absorbance between 0.5

and 0.7. Five ml. of the solution in acetate buffer was transferred

to a 10 ml. glass vial. Five ml. of chloroform presaturated with acetate

buffer was added to it. The vial was closed with a rubber stopper and

sealed with an aluminum cap. The solution was mixed on a Vortex Jr.

Mixer for 3 minutes. The solution was then centrifuged at 3200 r.p.m.

for about 3 minutes. About 3 ml. of the aqueous layer was withdrawn

from the vial with a glass syringe and a needle and transferred to a

test tube. One ml. of this solution was then diluted with 10 ml. of

pH 10.1 borate buffer and its absorbance was measured against an

acetate buffer blank treated identically. The difference between the

absorbances of the aqueous solution before and after the partitioning

represented the concentration of the substance partitioned into the

chloroform layer. The ratio of the difference in absorbances to the

absorbance of the aqueous layer after partitioning was the partition

coefficient for the barbituric acid derivative between its solution

in acetate buffer and chloroform.

The same procedure was used for studying the partitioning of

barbituric acid derivatives between their solutions in acetate buffer

and the organic phases, cyclohexane and silicone liquid 200. The

amounts partitioned into these organic phases were very small and gave

small differences between the absorbances of the acetate buffer layer

before and after the partitioning with these organic liquids. A small

error in the absorbance measurement therefore introduced large error









in the calculation of the partition coefficient. Hence these values

were not used in further investigation and have not been reported.

The coefficients of partition of aminoalkyiphenones 4'-

aminopropiophenone, 4'-aminoacetophenone, and 3'-amiroacetophenone -

between their solutions in pH 6.8 phosphate buffer and chloroform

were determined by the method described for barbituric acid derivatives.


Measurement of Thickness of Membranes


The silastic membranes were available in four different

labelled thicknesses of 3, 5, 10 and 20 mil. The actual thickness

of each membrane was measured with a micrometer screw capable of

measuring a minimum thickness of 0.1 mil. A clean sheet of paper

was cut into a rectangle measuring 2 1/2" x 5". This was folded into

a square of 2 1/2" x 2 1/2". The paper was marked lightly with a

pencil in a square pattern of 1" x 1" from number 1 to 9. The thick-

ness of the paper at each number was measured with the micrometer

screw. A membrane was carefully placed flat between the folded paper

and the thickness was measured at each number again. The thickness

of the membrane was then obtained by difference. The measurements

were made on seven different pieces of the similarly labelled thick-

nesses of the membrane.


Study of Permeability of Silastic Membrane to Phosphate
Buffer Salts and Hydrochloric Acid


A steady state diffusion cell was assembled with 5 mil. thick

silastic membrane in position. It was filled with about 20 ml. of

distilled water. The cell was kept in a 400 ml. beaker containing










about 250 ml. of pH 6.5 phosphate buffer (L = 1.2). A sample of

distilled water from the cell was tested after 15 hours for the

presence of P04- by the ammonium molybdate test (92).

Another diffusion cell filled with 20 ml. distilled water

was kept in a beaker containing 250 ml. of 0.1 N HCI solution. The

distilled water inside the cell was tested for the presence of chloride

ions after 11 hours by the silver nitrate test (93). The pH of the

distilled water was also measured.

Two-tenths of a ml. of 0.6 M and 5 ml. of 4.80 x 10-4

phosphate buffer solutions were diluted to 250 ml. with distilled

water to obtain 4.80 x 10-4 M and 9.60 x 10-6 M phosphate buffer

solutions respectively. Ammonium molybdate reagent (92) was prepared

by dissolving 1.59 g. of molybdic acid in 3.5 ml. of concentrated

ammonia solution diluted with 3.5 ml. of water. This solution was

slowly added to a mixture of 8 ml. of concentrated nitric acid and

10 ml. of distilled water. Three ml. of this ammonium molybdate

reagent was added to 3 ml. of sample solution for testing the presence

of phosphate salts.

The silver nitrate solution used for testing the presence of

chloride ions was prepared by dissolving 2 g. of silver nitrate in

20 ml. of distilled water. The sample to be tested for the presence

of chloride ions was acidified with a drop of concentrated nitric acid

and an equal volume of 10% silver nitrate solution was added. A

solution of 4 x 10-5 N HCI was used as a standard solution for the test.









Preparation of Solutions


The solutions of aminoalkylphenones were prepared in pH 6.8

phosphate buffer. A weighed amount of a compound (about 2.5 g.) was

dissolved in a known volume (about 100 ml.) of 0.4 N HCl. A known

aliquot of this solution (usually 20 ml.) was then diluted with pH

6.8 phosphate buffer (p = 0.3) to 2 liters.

Aliquots of 4'-aminopropiophenone solution in 0.4 N HCI were

diluted to 2 liters with acetate buffers and phosphate buffer of

different pH values and hydrochloric acid solutions of different

normalities. These solutions were then used to study the effect of

pH on the diffusion of 4'-aminopropiophenone through Silastic membrane.

The solutions of barbituric acid derivatives amobarbital,

barbital, cyclobarbital, diallylbarbituric acid, mephobarbital,

metharbital, pentobarbital, phenobarbital, secobarbital, thiamylal

and thiopental were prepared in pH 4.7 acetate buffer. A known

amount (about 0.5 g. in the case of mephobarbital, thiamylal and

thiopental and 1.0 g. in the case of the remainder of the barbituric

acid derivatives) was dissolved in 20 ml. of 0.1 N NaOH and then

diluted with pH 4.7 acetate buffer (p = 0.1) to 2 liters. Similarly

2 liters of the solutions of barbital and pentobarbital were also

prepared in acetate, phosphate and borate buffers of different pH

values to study the effect of pH on the rate of diffusion of these

barbituric acid derivatives.

Solutions of a-methylphenethylamine hydrochloride, 2-amino-

4-phenylpentane hydrochloride, 3-amino-1-phenylbutane sulfate, 1-

meTnyi-5-pheny!pentylamine hydrochloride and a-ethyiphenethyiamine









hydrochloride were prepared in borate buffer. A known amount (about

0.5 g.) of a drug was dissolved in 20 ml. of distilled water and

diluted with 0.1 N NaOH to 2 liters. The pH of the solution was then

adjusted with boric acid solution to a value close to the pKa value

of the drug. (The pKa values of these compounds are given in Table I).

The pH values of the final solutions of the compounds in the order

listed above were 8.93, 9.45, 9.40, 9.60 and 9.28 respectively.

A saturated solution of dextromethorphan was prepared in pH

10.1 borate buffer by equilibrating the solution at the temperature

of study in presence of excess of undissolved dextromethorphan for

about 12 hours. The saturated solution of progesterone was prepared

in the same way in pH 6.8 phosphate buffer solution.


A Typical Steady State Diffusion Experiment


A strip of Silastic membrane was washed with water several

times to remove the sodium bicarbonate dusting powder from its surfaces.

The membrane was finally washed with distilled water and driuc in air.

It was then cut into 2" x 2" pieces. A steady state diffusion cell was

assembled and two such pieces of Silastic membrane were fixed in

position between the stainless steel plates and silicone gaskets.

(See Fig. 5.)

The diffusion cell was filled with the buffer solution used

for the preparation of the solution of the drug under study. The

level of this solution in the stem of the glass T was kept about 1 1/2"

above the horizontal portion of the T joint. The cell was the.

placed in a 200 or. 400 ml. capacity beaker. The beaker containeG

either 120 or 200 ml. of 0.1 N HCI, borate buffer or phosphate buffer










solution. The solvents chosen depended on the drug under study. The

beaker was then covered with a piece of Parafilm (American Can Company,

Neenah, Wisconsin) with a hole in it for the stem of the diffusion cell.

The Parafilm was then taped to the sides of the beaker. The weaker

was covered with a plastic Petri dish with two holes in it. The stem

of the diffusion cell was inserted through one hole anc the other hole

was used to remove the samples from the solution in the beaker. The

Petri dish cover was taped to the sides of the beaker. The Petri

d sh cover held the diffusion cell immobile in th,, .,,
the widening of the hole in the Parafilm by the stem of the diffusion

cell, and thus minimized the evaporation or contamination of -he

solution in the beaker.

The beaker with the diffusion cell was fitted rigidly in the

metal rack resting in the shaker bath. This assembly was allowed to

equilibrate at the temperature of the bath for about 10 hours. A two

liter solution of the drug under study was also allowed to equilibrate

in another constant temperature bath maintained at The same temperature.

This solution was used as reservoir of the drug solution for the

diffusion experiment.

The pump was started with the ends of all four vinyl tubing

from the two channels dipping in the drug solution. The pumping was

continued until the tubes were full with the drug solution. After

thermal equilibration a hole was burned into the Parafilm by passing

a hot glass rod through the opening in the Petri dish cover. A

sample of the solution from The beaker was removed and analyzed

spectrophotometrica ly. The buffer solution in the diffusion cell

was withdrawn in-;-o suc; *.-. flask.









At zero time 20 ml. of the drug solution was pipetted i>to

the diffusion cell. The delivery and the suction ends of the tubes

from the pump were positioned in the diffusion cell as described

earlier. The pump was then started.

Samples were removed from the solution in the beaker a;T

regular intervals of time. In all cases, excepting the phenyialkyl-

amines, these samples were analyzed immediately with the spectro-

photometer. Samples of the drug solution from the reservoir were

taken at various intervals during the course of the diffusion

experiment. These samples were adequately diluted and analyzed

spectrophotometrically. The temperature of the bath was monitored

throughout the course of an experiment.

At the end of an experiment, the volume and pH of the solution

in the beaker and the pH of the solution in the diffusion cell were

measured. The pH values did not change throughout the course of an

experiment for all studies.


A Typical Quasi-steady State Diffusion Experiment


A clean piece of Silastic membrane was clamped into positi.

in the V-shaped quasi-steady state (see Fig. 4) diffusion cell. Each

of the arms of the diffusion cell was filled with 100 ml. of phosphate

buffer solution. The cell was then kept in a thermostated shaker bath

for equilibration for about 8 10 hours. After this the solutions

from both arms were withdrawn into a suction flask. Into one arm of

the cell 50 to 100 ml. of phosphate buffer was added and at zero time

an equal volume of the drug solution in phosphate buffer ,,s added to

the other arm. The open ends of the arms were covered wirh Parafilm.










At regular intervals of time, samples from both the arms of

the diffusion cell were removed and analyzed for the drug spectro-

photometrically. The experiment was continued until the difference

between the absorbance measurements became very small. At the end

of the experiment the volume and pH of the solutions in both arms

were measured.


Diffusion of 4'-Aminopropiophenone Through
Silastic Membrane


The solution of 4'-aminopropiophenone (PAPP) prepared in pH

6.8 phosphate buffer was used for both the steady state and quasi-

steady state diffusion work. The solutions used on the desorption

side of the membrane were 0.12 N HCI for the steady state diffusion

and pH 6.5 phosphate buffer for-the quasi-steady state diffusion

experiments.

The steady state diffusion of PAPP from its saturated

solutions through Silastic membrane was studied to determine the

reproducibility of the results. The effects on diffusion of PAPP

of ionic strength, concentrations of PAPP solutions, thickness of

Silastic membranes and the hydrostatic pressure exerted on the

membrane were studied.

Reproducibility of results.--The diffusion of PAPP from its

saturated solution in pH 6.5 phosphate buffer through 3 mil thick

Silastic membrane into 200 ml. of 0.12 N HCI at 37.50 was studied.

The cell solution containing excess undissolved PAPP was agitated by

bubbling nitrogen gas under pressure through it. The whole assembly

was shaken in a thermostated shaker bath. Samples (1 mi.) of the









beaker solution were removed every hour for about 8 10 hours. Each

was diluted with 4 ml. of pH 6.8 phosphate buffer (p = 0.3) and the

absorbance was measured at 307 mp, on the Beckman DU spectrophotometer.

Five separate experiments were performed on each of the five days.

Typical raw data for one such day are given in Table II.

Effect of hydrostatic pressure.--Solutions of PAPP in pH 6.8

phosphate buffer having same concentration were circulated into four

diffusion cells. The suction tubes of the channels returning the

solutions from the diffusion cells to the reservoir were adjusted at

four different levels (see Fig. 5) in the stem of the diffusion cell.

The levels of the tubes were 0, 1/2", 1" and 3" above the level of the

solution in the beaker. The higher rate of pumping the solution from

the cell to the reservoir over that of pumping the solution from the

reservoir to the cell, kept the level in the cell constant at the

level of the suction tube orifice. The rate of diffusion was monitored

by spectrophotometric analysis of the samples from the beaker.

Effect of ionic strength.-A 100 ml. solution of 0.4 N HCI

containing 2.5 g. of PAPP was prepared. Twenty ml. aliquots of this

solution were added with 170, 330, 500 and 670 ml. of pH 6.8 phosphate

buffer with an ionic strength of p = 1.2. The solutions were then

diluted with distilled water to 2 liters to obtain solutions of

PAPP of the same concentrations in phosphate buffer but with ionic

strengths of 0.102, 0.198, 0.300 and 0.402 respectively. The pH

values of these solutions were 6.69, 6.73, 6.75 and 6.78 respectively.

These solutions were then circulated in four diffusion cells to study

the diffuso.c of PAPP through Silastic membrane into 200 ml. of 0.12

N HCI at 23.00.









Effect of thickness of membrane.--Silastic membrane was

available in four different thicknesses of 3, 5, 10 and 20 mil. The

diffusion of PAPP through these four thicknesses was studied at 24.900.

A solution of PAPP in pH 6.8 phosphate buffer from one reservoir

containing about 8 liters of the solution was circulated through four

cells, each one fitted with a Silastic membrane of different thickness.

The samples of 0.12 N HCI from the beakers were analyzed as a function

of time to monitor the rates of diffusion of PAPP.

Effect of temperature.-The diffusion of PAPP from its solution

in pH 6.8 phosphate buffer through 3 mil thick Silastic membrane into

200 ml. of 0.12 N HCI was studied at seven different temperatures:

24.750, 24.900, 30.400, 31.250, 33.600, 37.500 and 41.00. At each

temperature the diffusion was studied at four. or more concentrations

of PAPP in the phosphate buffer solutions. The same diffusion cells

were used without changing the membranes to avoid any variation in

membrane thickness and in area of the membrane available for diffusion.

The temperature of the bath (American Instrument Company) was monitored

throughout the diffusion experiment and was observed to hold constant

within 0.250.

Quasi-steady state diffusion of PAPP.--The diffusion of PAPP

from its solution in pH 6.8 phosphate buffer through 5 mil thick

Silastic membrane into an equal volume of pH 6.8 phosphate buffer was

studied at 25.00. The volumes of phosphate buffer solutions used -

PAPP solution and phosphate buffer without any drug in it in the two

experiments were 50 and 100 ml.










Effect of pH on diffusion of PAPP.-The solutions of PAPP

were prepared in 0.001 N, 0.01 N and 0.1 N HCI and pH 3.48, 4.38 and

5.47 acetate buffer and pH 6.70 phosphate buffer. The steady state

diffusion of PAPP from these solutions through 3 mil thick Silastic

membrane in 200 ml. of 0.12 N HCI was studied at 25.00. The pH of

the solutions were noted before and after the diffusion experiments

and were observed to be unchanged. The absorbance of PAPP in the HCI

solution in the beaker was measured as a function of time after 1:5

dilution with phosphate buffer. The absorbances of the solutions in

the reservoirs, used for the circulation into the diffusion cells, were

measured after appropriate dilutions with phosphate buffer.

Effect of ethanol on the rate of diffusion of PAPP.--The

effect of ethanol on the steady state diffusion of PAPP through 3 mil

thick Silastic membrane was studied at 25.00. In one set of experiments

identical percentages of ethanol were present in the solutions of PAPP

in pH 6.8 phosphate buffer as well as in the 0.12 N HCI on the desorp-

tion side of the membrane. In another set of experiments ethanol was

present in the phosphate buffer solution of PAPP but no ethanol was

present in the HCI solution. In the third set of experiments ethanol

was present only in the HCI solution on the desorption side of the

membrane but no ehtanol was present in the phosphate buffer solutions

of PAPP inside the diffusion cells.

Six flasks containing 500 ml. of pH 6.8 phosphate buffer

(, = 1.2) and 500 ml. of distilled water were added with 160, 320, 480,

640, 800 and 960 ml. of 95% ethanol and the volume in each flask was

made up to 2 liters with distilled water. A solution of 3 g. of PAPP










in 140 ml. of 0.3 N HCI was prepared. Twenty ml. of this solution was

transferred to each of the six flasks containing ethanolic phosphate

buffer solutions to obtain solutions of PAPP in phosphate buffer

containing 7.5, 15.0, 22.5, 30.0, 37.5 and 45.0% V/V ethanol. Six

250 ml. volumetric flasks containing 15 ml. of 2 N HCI were added

with 20, 40, 60, 80, 100 and 120 ml. of 95% ethanol and the volumes

of the solutions were made up to 250 ml. to obtain 0.12 N HCI con-

taining 7.5, 15.0, 22.5, 30.0, 37.5 and 45.0% V/V ethanol. Two

hundred ml. of these ethanolic solutions was used in the diffusion

experiment on the desorption side of the diffusion cells. The phos-

phate buffer solutions of PAPP containing corresponding concentration'

of ethanol were circulated into these diffusion cells.

In another set of experiments the 0.12 N HCI used on the

desorption side of the diffusion cell contained 0, 10, 20, 30, 40 and

50% ethanol prepared as described before. The PAPP solution in the

bulk volume of 8.5 liters was used for circulation into all six

diffusion cells and did not contain any ethanol.

In the third set of experiments the diffusion of PAPP from

its solution in phosphate buffer containing 18% ethanol through 3 mil

Silastic membrane into 200 ml. of 0.12 N HCI containing no ethanol was

studied. The concentration of PAPP in the ethanolic phosphate buffer

solutions was varied from 1.87 x 10-3 M to 3.28 x 10-3 M.

The diffusion of PAPP from its saturated solutions in O,

10, 20 and 30% ethanolic solution in water through 3 mil thick Silastic

membrane into 0.1 N HCI containing 0, 10, 20 and 30% ethanol was

studied at 24.250.









The effect of ethanol on the quasi-steady state diffusion

of PAPP was studied at 25.250. Five hundred ml. solutions of PAPP in

0, 10, 20 and 30% ethanolic phosphate buffer were circulated in steady

state diffusion cells fitted with 2.30 mil thick Silastic membrane

(Fig. 5 ). The diffusion of PAPP from each of these solutions into

120 ml. of 0, 10, 20 and 30% ethanolic phosphate buffer was studied.

The diffusion was monitored by measuring the absorbance of the samples

of the external solution from the beaker at 307 mp as a function of

time. The samples were returned to the beakers. The loss of solutions

from the beakers due to evaporation and repeated sampling was less than 2%.

Effect of ethanol on Silastic membrane.--Three steady state

diffusion cells were filled with 0, 10 and 30% ethanolic phosphate

buffer solutions and kept in 0.12 N HCI solutions containing the same

concentrations of ethanol. After 15 hours the cells were emptied and

washed with water, and the steady state diffusion of PAPP from one

solution of PAPP in phosphate buffer circulated in all three cells was

studied in absence of any ethanol, inside or outside the cell.

Diffusion of ethanol through Silastic membrane.--The steady

state diffusion of ethanol through 3 mil Silastic membrane into 200

ml. of 0.12 N HCI was studied from absolute ethanol, from phosphate

buffer solutions of PAPP containing 18% ethanol and from phosphate

buffer solutions containing 10, 20, 30 40 and 50% ethanol.

Absolute ethanol was circulated into a diffusion cell fitted

with a 3 mil Silastic membrane. About 0.5 ml. samples of 0.12 N HCI

were removed at hourly intervals and 5 microliters of these samples

were injected into a gas chromatograph for vapor phase analysis of










ethanol. The percent ethanol present in the samples were calculated

from a calibration curve.

The diffusion of ethanol from 18% ethanolic phosphate

buffer solution containing increasing concentrations of PAPP from

1.87 x 10-3 M to 3.28 x 10-3 M through 3 mil Silastic membrane was

studied at 24.600.

The diffusion of ethanol from phosphate buffer solutions

containing 10, 20, 30, 40 and 50% ethanol through a 3 mil thick

Silastic membrane into 200 ml. of 0.12 N HCI was studied at 24.600.

Effect of ethanol, propyl alcohol, isopropyl alcohol and

t-butyl alcohol on diffusion of PAPP.-Three hundred grams of isopropyl

alcohol, propyl alcohol and t-butyl alcohol were weighed and diluted

with pH 6.8 phosphate buffer to 2000 ml. Similarly 240 ml. of ethanol

was diluted to 2 liters with phosphate buffer. To each solution 20

ml. of a solution of 3.0 g. of PAPP in 120 ml. of 0.4 N HCI was added.

The solutions were mixed and used for circulation in the steady state

diffusion cells fitted with 3 mil Silastic membrane. The concentra-

tion of PAPP diffusing into 200 ml. of 0.12 N HCI was measured as a

function of time.



Diffusion of 4'-Aminoacetophenone and 3'-Aminoacetophenone.


The diffusion of these two drugs through a 3 mil Silastic

membrane from their solutions in pH 6.8 phosphate buffer into 120 ml.

of 0.12 N HCI was studied at 25.00 and 37.50. The concentrations of

the drugs diffusing into HCI were measured spectrophotometrically after

1:5 dilutions with pH 6.8 phosphate buffer.










Diffusion of Barbituric Acid Derivatives


The steady state diffusions of barbituric acid derivatives -

amobarbital, barbital, butabarbital, cyclobarbital, diallylbarbituric

acid, mephobarbital, metharbital, pentobarbital, phenobarbital,

secobarbital, thiopental and thiamylal through 3 mil Silas-i.

membrane from their solutions in pH 4.7 acetate buffer into pH 10.1

borate buffer were studied at 25.00 and 37.50.

Weighed amounts of barbituric acid derivatives were dissolved

in small volumes of 1 N NaOH and immediately diluted with pH 4.7

acetate buffer solution to 2 liters to obtain solutions which were

approximately 5 x 10-3 M in thiamylal, thiopental and mephobarbital

and 5 x 10-3 M in the remaining barbituric acid derivatives. These

were used as reservoir solutions and circulated through the diffusion

cells fitted with 3 mil Silastic membranes. The solutions used on the

desorption side of the membrane in the beakers were 200 ml. of pH 10.1

borate buffer. The samples of borate buffer from the beaker were

removed at regular intervals of time. Their absorbances were measured

on the Beckman DU spectrophotometer without any treatment. After

measuring their absorbances, the samples were returned to the beakers.

The steady state diffusion of barbital and pentobarbital

through 3 mil Silastic membrane was also studied from their solutions

in acetate, phosphate and borate buffer solutions of different pH

values at 37.50. Solutions of barbital in pH 4.70 acetate buffer,

pH 6.65, 7.30, 7.60 and 8.12 phosphate buffer and pH 9.15, 9.75 and

10.55 borate buffer solutions were used. Solutions of pentobarbital

in pH 4.60 acetate buffer, pH 6.99, 7.57, 7.90 and 8.18 phosphate










buffer and 8.65, 9.05 and 9.53 borate buffer solutions were used.



Diffusion of Phenylalkylamines


The solutions of a-methylphenethylamine hydrochloride, 2-amino-

4-methyl-4-phenylpentane hydrochloride, 3-amino-1-phenylbutane sulfate,

1-methyl-5-phenylpentylamine hydrochloride and a-ethylphenethylamine

hydrochloride in pH 8.93, 9.45, 9.40, 9.60 and 9.28 borate buffer

respectively were used for the diffusion work. The steady state

diffusion of these substances through 3 mil Silastic membrane into

120 ml. of pH 6.8 phosphate buffer was studied at 25.00. Samples of

the phosphate buffer solutions (1 ml.) were removed every 15 minutes

and stored in a refrigerator. All the samples were analyzed simultane-

ously with the standard solutions for the preparation of the calibra-

tion curve.



Diffusion of Dextromethorphan and Progesterone


The steady state diffusion of dextromethorphan from solutions

in peanut oil, mineral oil and pH 10.1 borate buffer through 3 mil

Silastic membrane into pH 6.8 phosphate buffer was studied at 37.50.

Similarly steady state diffusion of progesterone from a solution in

peanut oil and a saturated solution in phosphate buffer through 3 mil

Silastic membrane into pH 6.8 phosphate buffer was studied at 37.50.

The solutions of the two compounds in peanut oil were prepared

by dissolving an accurately weighed amount of each drug in an accurately

weighed amount of peanut oil. The density of the peanut oil was










determined by weighing 10 ml. of peanut oil in a 10 ml. volumetric

flask. The volumes of the peanut oil used for the preparation of the

solutions were then calculated from the knowledge of the weights and

the density of the peanut oil. The molarities of dextromethorphan

and progesterone in the peanut oil solutions were then calculated.

The diffusion of dextromethorphan was also studied from its

saturated solutions in peanut oil and mineral oil. The concentration

of dextromethorphan in peanut oil was determined by diluting a

weighed amount of the filtered solution with chloroform and measuring

the absorbance of the solution at 277 mp against a chloroform blank

containing an equivalent amount of peanut oil. The E value of

dextromethorphan in peanut oil diluted with chloroform was determined

the same way using the solution of dextromethorphan in peanut oil

containing a known weight of the drug. The E value of dextromethorphan

in mineral oil was determined by dissolving an accurately weighed

quantity of dextromethorphan in mineral oil and measuring its absorb-

ance against a mineral oil blank. The saturated solution of the drug

in the mineral oil was filtered. Its absorbance was measured and the

concentration in the solution was calculated from the knowledge of

the E value.



Diffusion of Drugs Through Silastic Capsules


Silastic capsules (Dow Corning Center for Aid to Medical

Research) were small Silastic pouches prepared by sealing pieces of

tubes (about 6 mm. in diameter x about 2.4 cm. in length) at both ends.

These were cleaned externally with water and dried in air.










To determine the area available for diffusion five capsules

were cut open at their sealed ends. The resultant tubes were cut

again on one side to obtain flat sheets of the capsule material. The

thickness of this capsule material was measured with a micrometer

screw as described previously. The material was then cut into small

pieces and inserted into 10 ml. graduated cylinder containing 5 ml.

of water. The volume of water displaced by the material from the five

capsules was measured. The average volume of Silastic material from

one capsule was then calculated from the volume of the water displaced

by material from five capsules. The area of the Silastic membrane of

one capsule was then obtained by dividing the volume of the Silastic

material from one capsule by the thickness of the Silastic material.

Saturated solutions of 4'-aminopropiophenone in phosphate

buffer, solutions ofbarbital, pentobarbital and phenobarbital in

acetate buffer and solutions of dextromethorphan in borate buffer,

peanut oil, and mineral oil were prepared. The Silastic capsules

were cut open at one corner and were filled with these saturated

solutions. The hole in each of the capsules was sealed by forcing a

portion of Silastic medical grade adhesive type A (Dow Corning Center

for Aid to Medical Research) into and around the hole. The capsules

were set aside for one day and were tested for leaks by pressing them

lightly between fingers. Similarly capsules containing solvents

without drugs were also prepared.

The capsules containing barbital, pentobarbital and pheno-

barbital were washed with water and put in 50 ml. of pH 10.1 borate

buffer in glass vials. These vials were preequilibrated at 37.50 for






49



about 10 hours. Similarly capsules of 4'-aminopropiophenone and

dextromethorphan were put in glass vials containing 50 ml. of 0.12 N

HCI and pH 6.8 phosphate buffer respectively. The samples from the

vials were assayed spectrophotometrically for the respective compounds.
















RESULTS


Screening of Permeability of Polymer Films


The spectrophotometric cruves obtained on the Cary recording

spectrophotometer were checked for the peaks characterizing the com-

pounds under study. This method of analysis was sensitive to 1 x 10-4

M concentration in all cases and much more sensitive in the case of

steroids (Table I).

Polyethylene membrane is permeable to4'-aminoacetophenone in

aqueous solutions and in peanut oil, and to 4'-aminopropiophenone in

ethylene glycol. Both compounds permeate Rilsan membrane from ethanolic

solutions. Polypenco membrane is permeated by 4'-aminopropiophenone

and 3'-aminoacetophenone from their solutions in 0.1 N NaOH and ethanol.

Polypropylene is nonpermeable to all the drugs tested. Substances

were leached by phosphate buffer from cellulose triacetate and cellulose

butyrate membranes and interfered with the spectrophotometric analyses

of the drugs. After compensating for the absorbances of these inter-

fering substances, it was concluded that 4'-aminopropiophenone permeated

cellulose triacetate from solutions in propylene glycol, ethanol and

polyethylene glycol 200. Positive results were recorded also for barbital

in polyethylene glycol 200 and sulfabenzamide in ethanol. Mylar poly-

ester film was available only in small quantity and was impermeable to

all the drugs tested with the available films. Silastic membrane was










permeable to 4'-aminopropiophenone from solutions in all the solvents

tested except mineral oil, to 3'-aminoacetophenone from all solvents

except HCI and mineral oil, to progesterone from ethanol, to barbital

from ethanol and ethylene glycol and to phenobarbital from peanut oil,

ethanol, polyethylene glycol 200, phosphate buffer and 0.1 N HCI.

Since the purpose of this screening was to search for a membrane

which would allow permeation of a majority of compounds in measurable

quantities only Silastic membrane was investigated further.



Solubility Studies


The solubilities of barbituric acid derivatives in pH 4.7

acetate buffer at 25.00 are reported in Table III. The solubilities

of aminoalkylphenones in pH 6.8 phosphate buffer at 37.50 are given

in Table IV. The solubility of 4'-aminopropiophenone in phosphate

buffer containing various percentages of ethanol were interpolated

from Fig. 6 showing the relation between solubility and percent

ethanol. The solubility values are given in Table V.

The solubilities of phenylalkylamines in 0.1 N NaOH could

not be determined because the insoluble liquid amines formed an

emulsion which could not be broken even after extended periods of

centrifugation. Clear solutions of these compounds for the deter-

mination of the solubilities were not obtainable.



Partition Coefficients


The partition coefficients for barbituric acid derivatives

between their solutions in acetate buffer and chloroform presaturated










with each other are reported in Table Ill. The partition coefficients

for 4'-aminopropiophenone, 4'-aminoacetophenone, and 3'-aminoaceto-

phenone between their solutions in pH 6.8 phosphate buffer and chloro-

form presaturated with each other at 25.00 are given in Table IV. The

partition coefficient between phosphate buffer solutions of 4'-amino-

propiophenone containing ethanol and organic solvents were not

obtained because of the high solubility of ethanol in most organic

solvents.



Thickness of Membranes


The results obtained from the measurements of the membrane

thickness are recorded in Table VI. The results were analyzed

statistically (94) as shown in Appendix C VI for 3 mil thick

Silastic membranes. The coefficients of variation (94) and the limits

within which the mean thicknesses vary have been reported at 95%

confidence levels (94) in Table VI.



Permeability of Silastic Membranes to

Phosphate Buffer Salts and Hydrochloric Acid Solution


The ammonium molybdate reagent solution (92) gave a yellow

colored solution when mixed with an equal volume of 4.8 x 10-4 M

solution of phosphate buffer. A faintly yellow colored solution was

obtained with 9.60 x 10-6 M phosphate buffer solution. The sample

from the diffusion cell kept for 15 hours in a beaker containing

phosphate buffer, did not yield a precipitate or a colored solution










with this reagent. It was therefore concluded that phosphate buffer

salts do .not diffuse significantly through Silastic membrane.

A white turbid solution was obtained when 4 x 10-5 M solution

of HCI was acidified with nitric acid and an equal volume of 10%

silver nitrate solution was added. A sample from the diffusion cell

kept for 11 hours in a beaker containing 0.1 N HCI did not yield a

turbid solution when treated similarly. Also, the 6.45 pH of the

solution in the cell was constant for the time interval. It was

therefore concluded that HCI does not diffuse significantly through

Silastic membrane.



Treatment of Data


Table II shows the raw data obtained on the steady state

diffusion of 4'-aminopropiophenone from solutions in pH 6.5 phosphate

buffer through 3 mil Silastic membrane into 200 ml. of 0.12 N HCI

at 37.3. This and other such data were plotted according to the

integrated form of Eq. 9


A = D S K t = D S (C2 CO) t / X (Eq. 24)


where A is the amount in moles diffused in t seconds through a

membrane of X cm. thickness and S cm.2 area. K is the concentration

gradient (C2 C1) / X. D is the apparent diffusion constant in

cm.2/sec. and C2 is the concentration of the compound in the cell.

C1 is the concentration of the compound on the desorption side of

the membrane. In the steady state diffusion experiments the con-

centration C1 was maintained zero by keeping the diffused species










charged in the solution in the beaker. Thus 0.12 N HCI in the case

of 4'-aminopropiophenone and other aminoalkylphenones, pH 10.1 borate

buffer in the case of barbituric acid derivatives and pH 6.8 phosphate

buffer in the case of phenylalkylamines and dextromethorphan kept the

respective diffusates in the charged form. Charged species do not

diffuse through Silastic membrane. Hence the concentration C1 in the

beaker was always zero with respect to the diffusing species.

When the concentration of the solution in the beaker was

plotted against time in hours, the slope of the plot was


slope = D S C2 / X V = dC/dt (Eq. 25)


where V was the volume of the solution in the beaker. Since all the

terms in Eq. 25 are known except the apparent diffusion constant D,

it can be calculated from the value of the slope.

The quasi-steady state diffusion data were treated according

to Eq. 21. Since the volumes Vi and V2 used in the two arms of the

quasi-steady state diffusion cell were equal, Eq. 22


(X V / 0.869 S) log (C2 C1 / CO) = -D t (Eq. 22)


was used. The plot of -log (C2 Cl / CO) against time yields a slope

of 0.869 S D / X V from which D, the apparent diffusion constant, can

be calculated.



Diffusion of 4'-Aminopropiophenone Through Silastic Membrane


Reproducibility of results.-The regression analysis of the

raw data on the diffusion of 4'-aminopropiophenone (PAPP) in the










replicate five runs of Table II is given in Appendix C VII. The

intercepts obtained by regression analysis were tested by the "student"

t-test (94, 95, 96) to ascertain the validity of the hypothesis that

the intercepts were zero. In all five cases the hypothesis that the

intercepts were zero could not be rejected at the 95% confidence

level. Hence it was concluded that the plots of the concentration

of the diffused material versus'time go through the origin. The

coefficient of variation for values among the slopes was 6.98%.

The slopes of the concentration versus time plots for five

replicate diffusion experiments on each of five days at 37.50 are

noted in Appendix C VIII. The concentration of PAPP in the saturated

solutions used for the experiments was kept constant at 2.35 x 10-3 M.

The slopes were analyzed statistically (94, 95, 96) (Appendix C VIII)

to test the hypothesis that there was no significant difference

between the slopes on the same day and between the slopes on different

days. An analysis of variance table was constructed and by the "F"

test it was concluded that the slopes obtained within and among days

were statistically equal. The coefficient of variation for means of

the slopes on different days was 6.09%.

lonic strength effect.-The data obtained from the steady

state diffusion experiments for the concentrations of PAPP diffusing

through Silastic membrane from phosphate buffer solutions with ionic

strengths of 0.102, 0.198, 0.300 and 0.402 were plotted against time.

The slopes of these plots were normalized by dividing by the concen-

trations of the diffusing solutions. The mean of these resultant

specific rates of diffusion was (7.13 0.23) x 10-2 (Appendix C IX).










The coefficient of variation was 3.24%. This value of the coefficient

of variation is well within the value of 6.98% calculated for the

variation in slopes from five replicates of an experiment on one day.

This indicates that the ionic strength has no significant effect on

the diffusion of PAPP through Silastic membrane.

Effect of hydrostatic pressure.--The values of the slopes of

the concentration of PAPP diffused versus time in the steady state

diffusion experiments were normalized with respect to the concentra-

tion of the drug as shown in Appendix C X. The values of the

specific rates of diffusion so obtained were very close to each other.

The coefficient of variation was 1.45%. This value being less than

the coefficient of variation calculated for replicates of a diffusion

experiment on one day, it was concluded that the hydrostatic pres-

sures exerted by the liquids within the limits of the levels studied,

have no significant effect on the rate of diffusion of PAPP through

Silastic membrane.

Effect of thickness of membrane on the specific rate of

diffusion.--The plots of the concentration of PAPP diffused in steady

state diffusion experiment through Silastic membrane versus time as

a function of membrane thickness are shown in Fig. 7. Equation 25

may be rewritten as


D / X = Slope x V / S C2 (Eq. 26)


Thus the slopes of the plots of concentration of diffused drug versus

time may be normalized as


D / X = slope x 0.200 / 10.40 x 1.517 x 10-3 x 3600
(Eq. 27)










where 0.200 liter is the volume of the solution, 10.40 cm.2 is the

area, 1.517 x 103 is the molar concentration of PAPP and 3600 is the

factor for converting hours to seconds. The normalized values of the

slopes,specific rates of diffusion, are given in Table VI. When

these specific rates of diffusion are plotted against the reciprocal

of thickness of the membranes, a straight line is obtained as shown

in Fig. 8. The slope of this line, 3.76 0.19 x 10-10 cm.2/sec.

is the apparent diffusion constant D for the diffusion of PAPP through

Silastic membrane at 24.900.

Effect of concentration on the rate of diffusion.--The slopes

of the plots of concentration of PAPP diffused through Silastic

membrane versus time are listed in Table VII as a function of concen-

tration and temperature. These rates of steady state diffusion were

plotted against the respective concentrations at each temperature.

Fig. 9 shows such plots at three temperatures. The slopes of these

plots are the specific rates of diffusion of PAPP through Silastic

membrane and are equal to D S / X V and are given in Table VII. The

apparent diffusion constants calculated from these values are also

given in Table VII. These values were calculated using the equation


D = slope x X x V / S = slope x 7.52 x 10-3 x 0.200 / 10.40 x 3600

(Eq. 28)

where 0.200 liter is the volume V, 10.40 cm.2 is the area S, 7.52 x

10-3 cm. is the thickness X and 3600 is the factor for converting

hours to seconds.

Effect of temperature on the rate of diffusion.--The effect

of temperature on the rate of steady state diffusion of PAPP through









3 mil thick Silastic membrane is shown in Fig. 9. The apparent

diffusion constants at seven temperatures are recorded in Table VII.

Figure 10 shows the plot of logarithm of the apparent diffusion

constants versus the reciprocal of the absolute temperatures. From

the slope of this plot the activation energy of diffusion AEa can be

calculated using the equation


log D = log DO 6Ea / 2.303 R T (Eq. 29)


&Ea for PAPP diffusion through Silastic membrane was calculated to be

4.90 Kcal./mole.

Quasi-steady state diffusion.-The plots of -log (C2 CI/CO)

versus time for quasi-steady state diffusion of PAPP through 5 mil

Silastic membrane are shown in Fig. 11. The apparent diffusion con-

stants were calculated from the slopes of these plots to be 3.93 x 10-10

and 3.94 x 10-10 cm.2/sec. for 50 ml. and 100 ml. volumes of the

solutions in the arms of the diffusion cell at 25.00. These values

are within the estimated interval for diffusion constants 3.76 .19 x

10-10 cm.2/sec. obtained from steady state diffusion experiments.

Effect of pH on the rates of diffusion.-The apparent diffusion

constants obtained for the diffusion of PAPP from the solutions of

various pH values through 3 mil Silastic membrane have been reported

in Table VIII. The D pH profile constructed from the values

reported in Table VIII is shown in Fig. 12. The apparent diffusion

constant approaches an asymptotic value of 3.66 at higher pH values.

The pH corresponding to half the asymptotic value of 1.83 is 2.45

which is the pKa of the drug obtained from the diffusion experiments.









The pKa of the drug obtained spectrophotometrically is 2.42. (Table I)

Effect of ethanol on the rate of diffusion of 4'-aminopropio-

phenone.--The results obtained in the steady state diffusion of PAPP

through 3 mil Silastic membrane are recorded in Table IX. Set A in

Table IX lists the rates of diffusion when ethanol is present on both

sides of the membrane in equal concentration. A decline in the

specific rate of diffusion was observed with the increase in the

concentration of ethanol present in contact with PAPP. The plots of

the absorbance of the diffused PAPP versus time for different

concentrations of ethanol in PAPP solution are shown in Fig. 13. Set

D in the Table IX shows the same results obtained with saturated

solutions of the drug in aqueous ethanol and 0.1 N HCI containing

the same concentration of ethanol. In set B the diffusion constants

and the specific rates of diffusion are invariant for the different

concentrations of ethanol present in HCI solution. Set C shows that

when the concentrations of the drug are varied in solutions with a

constant concentration of ethanol, the specific rates of diffusion

with respect to the concentration, and the apparent diffusion constants

are invariant.

The data obtained from quasi-steady state experiments were

treated according to


V2/(Kp2VI + KplV2)ln[Kp2CO/Kp2CO CI(Kp2V1 + Kpl 2)/V2] = D'St/XVI

(Eq. 30)
where subscripts 1 and 2 refer to the compartments containing solutions

without PAPP and with PAPP respectively at zero time. VI and V2 are

the volumes of the solutions with PAPP concentrations of C, and C2









respectively. Kpl and Kp2 are the apparent partition coefficients

between membrane material and aqueous solutions in the two compartments.

Equation 30 has been derived in Appendix C III.

When the numerator and the denominator of the logarithmic

term in Eq. 30 are multiplied by the intrinsic diffusion constant D'

for diffusion within the membrane alone, the resultant equation after

.rearrangement of terms becomes


In[D'Kp2COV2/D'Kp2CoV2 CI(D'Kp2V1 + D'KplV2)]

(D'Kp2V1 + D'KplV2)St/XVIV2 (Eq. 31)


If Kpl = Kp2 = Kp as is the case when the concentration of ethanol on

both sides of the membrane is the same, Eq. 30 reduces to


In[COV2/(COV2 C1(VI + V2)] = D S t(V1 + V2)/XVIV2 (Eq. 32)


where D is the apparent diffusion constant, D'Kp. The apparent

diffusion constants of PAPP for diffusion from 0, 10, 20 and 30%

ethanolic phosphate buffer into ethanolic phosphate buffer of the same

composition were calculated from the slopes of the plots of logarithmic

term in Eq. 32 versus time and their values are given in set A of Table X.

These values of D'Kp were used in places of D'Kpl and D'Kp2 in

Eq. 31 for comparable compositions of ethanolic phosphate buffers.

These values of D'Kpl and D'Kp2 were used to determine the slope

(D'Kp2VI + D'KplV2)S/XVIV2, from the plot of the calculated left hand

term of Eq. 31 versus time for various concentrations of ethanolic

phosphate buffer on the two sides of the Silastic membrane. The values

of D'KplV2 + D'Kp2VI derived from the measured slopes and calculated

from the known values of D'Kpl, D'Kp2, V, and V2 are given in Table X

for purposes of comparison.









The specific rates of diffusion of ethanol from phosphate

buffer containing 18% ethanol and different concentrations of PAPP

were found to be independent of the drug concentration (Table XI).

The specific rates of diffusion of ethanol from different concen-

trations of ethanol in phosphate buffer are also reported in Table XI.

The specific rates of diffusion of PAPP from solution in

phosphate buffer through Silastic membrane pretreated with ethanol

solutions of different concentrations was observed to be the same

in all cases indicating that the membrane is not altered physically

by ethanol; and that the alteration, if any, is reversible.

Effect of ethanol, propyl alcohol, isopropyl alcohol and

t-butyl alcohol on the diffusion of 4'-aminopropiophenone.-The con-

centrations of all the alcohols in the phosphate buffer were 2 M.

The specific rates of diffusion of PAPP from solutions containing

above listed alcohols were 5.05, 5.95, 6.41 and 5.12 x 10-3 hr.-1

respectively. It was therefore concluded that all four alcohols

modify the rate of diffusion of the drug to the same extent.



Diffusion of 4'-Aminoacetophenone and 3'-Aminoacetophenone


The apparent diffusion constants for the steady state

diffusion of these two drugs through 3 mil Silastic membrane from

solutions in pH 6.8 phosphate buffer into 0.12 N HCI have been

reported for two temperatures of 25.00 and 37.50 in Table IV. (see

also Fig. 10) Table IV also lists the apparent diffusion constants

for 4'-aminopropiophenone, the partition coefficients of these three

drugs and their solubilities together with the activation energies of

diffusion.










Diffusion of Barbituric Acid Derivatives


The apparent diffusion constants and the specific rates of

steady state diffusion of amobarbital, barbital, butabarbital,

cyclobarbital, diallylbarbituric acid, mephobarbital, metharbital,

pentobarbital, phenobarbital, secobarbital, thiamylal, and thiopental

through 3 mil Silastic membrane from acetate buffer solutions are

given in Table XII for two temperatures 24.600 and 37.50. The plots

of the concentration of diffused drugs against time have been shown

in Figs. 14 and 15.

The apparent diffusion constants for the steady state diffusion

of barbital and pentobarbital from solutions of different pH values

through Silastic membrane are reported in Table XIII. The plots of

apparent diffusion constants versus pH for the two drugs are shown

in Fig. 16. From these plots the pKa values of barbital and pento-

barbital were calculated to be 7.50 and 7.72 respectively. The pKa

values of these two drugs obtained titrimetrically were 7.45 and 7.65

respectively.

The plots of the logarithm of apparent diffusion constants

for some of these compounds against the reciprocal of the absolute

temperatures have been shown in Fig. 17. The activation energies of

diffusion aEa for barbituric acid derivatives have been reported in

Table XII.



Diffusion of Phenylalkylamines


The apparent diffusion constants and the specific rates of

diffusion from the steady state diffusion of a-methylphenethylamine,










3-amino-1-phenylbutane, 1-methyl-5-phenylpentylamine, a-ethyl-

phenethylamine and 2-amino-4-methyl-4-phenylpentane from their

solutions in borate buffer at pH values of 8.93, 9.40, 9.60, 9.28

and 9.45 through 3 mil Silastic membrane into 200 ml. of pH 6.8

phosphate buffer are reported in Table XIV. The plots of absorbance

of the drug diffused versus time are shown for three drugs in Fig.

18. The concentrations given in Table XIV are the total concentra-

tions and the concentrations of the uncharged species calculated

from the equation (97) where the pKa values are given in the table.


pH = pga + log [base] / [salt] (Eq. 33)




Diffusion of Dextromethorphan and Progesterone


The data obtained in the steady state diffusion study of

dextromethorphan were plotted as concentration of the drug diffusing

from solutions in peanut oil, mineral oil and pH 10.1 borate buffer

through 3 mil Silastic membrane into phosphate buffer versus time

as shown in Fig. 19. The apparent diffusion constants calculated

from these plots are given in Table XV.

The diffusion of progesterone was studied using a steady

state diffusion cell and the data obtained were plotted as shown in

Fig. 20. It is observed from Fig. 20 that the concentrations of

progesterone approach asymptotic values in the two plots. The pH

adjustment of the solution on the desorption side of the membrane

to force the steady state diffusion conditions was not possible

as progesterone does not have a pKa. The data obtained were










therefore treated by the quasi-steady state equation,

(Appendix C IV) using the equilibrium concentration values.

The apparent diffusion constants calculated for progesterone are

given in Table XVI.



Diffusion of Drugs from Silastic Capsules


The average area of Silastic membrane in the capsule was

found to be 5.58 cm.2. The average membrane thickness was 3.134 x

10-2 cm. The adhesive sealed the capsules well as tested for their

airtightness by pressing between fingers. However the portion of the

adhesive spread over a large part of the capsule making thickness at

those places different from average thickness of the capsule membrane.

The apparent diffusion constants and the specific rates of

diffusion obtained for the diffusion of barbital, phenobarbital,

pentobarbital from saturated solutions in acetate buffer, for 4'-

aminopropiophenone from saturated solution in phosphate buffer and

for dextromethorphan from saturated solutions in borate buffer,

peanut oil and mineral oil are given in Table XVII.
















DISCUSSION


Permeability of Silastic Membrane


Silastic membrane is impermeable to chloride ions and phosphate

buffer salts. The apparent diffusivities of 4'-aminopropiophenone,

pentobarbital and barbital (Figs. 12 and 16) decrease with decreased

concentrations of uncharged species in the solution. Increased concen-

trations of charged species on the desorption side of a Silastic

membrane had no effect on the rate of transfer of the uncharged species

through the membrane. This has-been shown for all the compounds

studied e.g. the rate of transfer of dextromethorphan (pKa = 8.25)

from saturated solution in borate buffer or mineral oil was invariant

with increasing concentrations of protonated dextromethorphan in the

phosphate buffer on the other side of the membrane (Fig. 19). In fact,

the concentrations of dextromethorphan on the desorption side of the

membrane greatly exceeded the total concentration on the diffusing side.

Progesterone does not have a pKa and cannot exist as a charged

species. The effect of decreased concentration gradient with increase

in the concentrations in the solution on the desorption side is

apparent from the approach to an asymptotic concentration (Fig. 20).

These facts demonstrate that Silastic membrane is impermeable

to charged molecules. This phenomenon was used to force steady state

conditions in the diffusion experiments throughout this investigation.










Fick's Law of Diffusion


The observed effects of the concentrations of 4'-aminopropio-

phenone (PAPP) solutions and the thickness of Silastic membrane on the

diffusion of PAPP were in accordance with Fick's law of diffusion


dA/dt = D S dC/dx (Eq. 4)


where A is the amount in moles diffused in time t through a membrane

with surface area of S cm.2. The differential dC/dx is the concentra-

tion gradient and D is the apparent diffusion constant. Under the

steady state conditions where the concentration of the diffusable

species on the desorption side of the membrane is essentially zero,

Eq. 4 becomes


D/X = dC/dt (V/SC2) (Eq. 34)


where dC/dt is the slope (Eq. 26) of the linear plot of the concen-

tration of PAPP appearing on desorption side of the membrane with time

(Fig. 7), V is the volume and C2 is the concentration of PAPP in

solution from which the diffusion is occurring. The linearity of such

plots indicate that the specific rate of diffusion, D/X, is inversely

proportional to the thickness of the membrane. The validity of this

relationship is shown by the linearity and zero intercept of the plot

in Fig. 8.

The effect of concentration of PAPP on the rate of diffusion

is shown in Fig. 9 and is in accordance with Eq. 25 rewritten as


dC/dt = DSC2/XV


(Eq. 35)










which states that the rate of diffusion is directly proportional to the

concentration of diffusing species provided that the concentration is

zero on the desorption side of the membrane.


Effect of pH on Diffusion


The plot of the apparent diffusion constant D versus pH of

the diffusing solution is shown in Fig. 12 for PAPP and in Fig. 16 for

barbital and pentobarbital. The apparent diffusion constants were

calculated by Eq. 24 where C2 was the total concentration and included

both charged and uncharged species. These apparent diffusion constants

decrease with decreased concentrations of the uncharged species in the

solution since Silastic membrane is impermeable to charged species.

The pKa values of barbital and pentobarbital were the pH at the mid-

point of the sigmoid curve of apparent diffusion constants versus pH

(Fig. 16). The pKa of PAPP was the pH at half the asymptotic value

in sigmoidal curve of apparent diffusion constants versus pH (Fig. 12).


Apparent Diffusion Constants From Steady and Quasi-steady

State Diffusion Experiments


The value of the apparent diffusion constants (D in Eq. 24)

for the steady state diffusion of uncharged 4'-aminopropiophenone

(PAPP) through Silastic membrane at 25.00 is(3.76 0.19)x 10-10 cm.2/

sec. These results have been shown to be highly reproducible

(Appendix C VIII). The apparent diffusion constant (D in Eq. 22)

obtained for the same compound at the same temperature from quasi-

steady state diffusion experiment was 3.94 x 10-10cm.2/sec. This










value lies within the confidence interval calculated for the apparent

diffusion constants for the steady state diffusion data. The coinci-

dence of the apparent diffusion constants obtained from the two

different methods shows the validity of the results.


Non-dependence of Apparent Diffusion Constants of

Drugs on Molecular Weights


The diffusion constants for the diffusion of a substance in

an isotropic medium has been related to molecular weight in several

equations e.g. Eqs. 1 and 2. Barrer and Chio (98) have simplified

these relations, lumping together the constants, to


D = a Mb (Eq. 36)


where D is the diffusion constant, a and b are arbitrary constants

and M is the molecular weight. They have claimed that this expression

holds moderately well for the diffusion of inert gases through

silicone membranes.

The diffusion of barbituric acid derivatives through Silastic

membranes was studied to test this relationship. The apparent diffusion

constants (Table III) were plotted against the cube root and square

root of the molecular weights of the respective compounds according to

Eqs. 1 and 2 respectively. Similarly, the logarithm of the apparent

diffusion constants of barbituric acid derivatives (Table III) and

those of the other compounds (Tables IV and XIV) were plotted against

the logarithm of their molecular weights in accordance with the

logarithmic transformation of Eq. 36. The points were too scattered










to permit any assumption of correlation between molecular weightsand

the apparent diffusion constants.

Barrer et al. (99) have shown that the diffusion of hydro-

carbons in silicone rubber is "less sensitive to size and shape of the

penetrant molecule, and thus silicone rubber is less selective than

natural rubber as a separation medium." Their data presented for the

diffusion of hydrocarbons in silicone rubber did not permit correla-

tion with the molecular weights of the compounds.


Postulated Mechanism of Transport in Silastic Membrane


The transport of compounds from a solution into a membrane,

through the membrane and then into another solution has been shown

to obey Fick's law of diffusion. The transfer of compounds from the

solution into the membrane must be non-rate determining in the concen-

tration range studied. Fick's law with its dependence on membrane

thickness and concentration gradient fully describes the transport

process from very low to saturation concentrations of PAPP. The

diffusion constants obtained from the steady state and the quasi-

steady state experiments are equal.

The non-dependence of the apparent diffusion constants on

the molecular weights of the diffusing species may be explained on the

basis of a partition hypothesis. In the partition hypothesis, the non-

charged diffusing species partitions from the solution into the adjacent

membrane monolayer, is transported across the membrane and then

repartitioned into the solution on the other side of the membrane.

The partitionings on either side of the membrane are rapidly effected.










Apparent Diffusion Constants and Partition Coefficients


The partition hypothesis states that the apparent diffusion

constant D is related to the partition coefficient Kp


D = D'Kp (Eq. 37)


as may be readily seen from Eqs. 17 and 20 where D' is the intrinsic

diffusion constant for a compound in Silastic membrane.

This hypothesis was tested by plotting (Fig. 21) the apparent

diffusion constants of barbituric acid derivatives against their

coefficients of partition between chloroform and acetate buffer

solutions (Table III). A fairly good straight line could be drawn

through the points. This linear relationship implies that the parti-

tion coefficients of all the barbituric acid derivatives between

Silastic membrane and the acetate buffer are linearly related to the

partition coefficients between chloroform and acetate buffer. It

assumes that the intrinsic diffusion constant D' is the same for all

the barbituric acid derivatives in the Silastic membrane. These

assumptions may be reasonably true for compounds belonging to a

homologous series. However not all the barbituric acid derivatives

studied belong to such a series. Hence the deviations of the few

compounds (Fig. 21) from the linear plot are to be expected. The

diffusion constants of thiamylal and thiopental are extremely high,

commensurate with their high partition coefficients(Table III).

The partition hypothesis is further substantiated by the

good linear plot (Fig. 22) of diffusion constants against the

coefficients of partition of aminoalkylphenones between chloroform










and phosphate buffer solutions (Table IV).


Apparent Diffusion Constants and Solubilities


The partition coefficient may be defined as the ratio of the

activities and/or solubilities of a compound in two phases (100).


Kp = Cm/Ca = Sm/Sa (Eq. 38)


where Kp is the partition coefficient, Cm and Ca are the concentrations

of a compound in membrane and the aqueous buffer solution respectively,

and Sm and Sa are the solubilities of the compound in the membrane and

the aqueous buffer solution respectively. When Eq. 38 is substituted

into Eq. 37 the resultant equation obtained is


D = D' Sm/Sa (Eq. 39)


Barbituric acid derivatives do not significantly partition

into silicone liquid 200 from acetate buffer solutions. The estimates

of partition coefficients between the two solvents was in great error.

When the solubilities and the intrinsic diffusion constant D' of

barbituric acid derivatives in Silastic membrane are assumed to be

equal, the product D'Sm in Eq. 39 will be constant. The plot of the

apparent diffusion constants against the reciprocal of the solubilities

of barbituric acid derivatives in acetate buffer (Table III) show a

reasonably linear correlation (Fig. 23). The deviations of some

points from the straight line relationship are to be expected since

the simplifying assumptions of constancy of solubility and intrinsic

diffusion constant in Silastic membrane are not rigidly true in all cases.










The phenylalkylamine series of drugs were diffused through

Silastic membrane to confirm the relationship between apparent

diffusion constants (Table XIV) and partition coefficients, and

solubilities of the drugs. Unfortunately the ease of emulsification

of the liquid amines prevented the ready estimates of the solubilities

and the partition coefficients of this series of drugs.


Effect of Ethanol on Diffusion of 4'-Aminopropiophenone

Through Silastic Membrane


The results of the steady state diffusion of PAPP from

ethanolic phosphate buffers (Table IX) demonstrate that the rate of

diffusion of PAPP decreases with increased concentration of ethanol

in contact with the drug (Table IX, Set A). The rate of diffusion

is invariant with different concentrations of ethanol in the HCI

solution on the desorption side of the membrane (Table IX, Set B).

However, the rate is slightly but significantly higher than that in

complete absence of ethanol.

The rates of diffusion of PAPP and ethanol from solutions

of different concentrations of PAPP in ethanolic phosphate buffer

show very little variation (Table IX, Set C and Table XI). This

demonstrates that the rates of diffusion of PAPP and ethanol are

independent of each other.

The data from the quasi-steady state diffusion were treated

specifically (Eqs. 31 and 32) to show the dependence of the apparent

diffusion constants on the partition coefficients for PAPP between

membrane and solutions on its two sides. The values of (D'Kp2V1 +
p2 1










D'KplV2) (Table X) where D' is the intrinsic diffusion constant, Kpl

and Kp2 are the partition coefficients and V1 and V2 are the volumes

of the solutions on two sides of the membrane, were obtained from the

graphical analysis of the data and also calculated from the known

values of the apparent diffusion constants D'Kp for comparable

conditions obtained from separate experiments (Table X, set A). In

all the cases (Table X, sets B, C, D and E), the two sets of value

are quite close to each other. It is apparent from the results

(Table X) that partition coefficients on either side of the membrane

form an integral part of the apparent diffusion constants. This

observation together with the fact that the rates of diffusion of

ethanol and PAPP are independent of each other (Table XI) indicate

that the intrinsic diffusion constant of PAPP in Silastic membrane

is not significantly dependent on the ethanol concentration on

either side of the membrane.

It was noted previously that the partition coefficient is

a ratio of the solubilities (100) in two media. In the case of

diffusion of PAPP through Silastic membrane in the presence of

ethanol, the premise of constancy of solubility in the membrane is

warranted since only one compound is involved. Consequently the

plot of the apparent diffusion constants of PAPP from ethanolic phos-

phate buffer through Silastic membrane versus the reciprocal of the

solubilities of PAPP in these solutions (Table V) should be linear

according to Eq. 39. This was found to be true as seen from linearity

of the plot in Fig. 24. The regression analysis of the plot in Fig.

24 showed that at the 5% level of significance the intercept is not

significantly different from zero.










Temperature Dependence of Diffusion in Silastic Membrane


The temperature dependence of diffusion in polymeric films

may be expressed by


D = Do e -AEa/RT (Eq. 3)


where LEa is the apparent activation energy of diffusion. Barrer

and Chio (98) have reported that the activation energy of diffusion

of inert gases through silicone membrane is about 4 Kcal./mole.

The activation energy of diffusion in solution is theoreti-

cally about one-third of the energy of vaporization (101) and is in

the range of 3-5 Kcal./mole. in many liquids. These calculations are

based on the assumption that the diffusing molecules are large

compared to the molecules of the medium into which they are diffusing

and that the movement of the solvent molecules determines the rate

of diffusion of the solute (102). It has also been pointed out that

large diffusion constants have small temperature coefficients and

conversely "more slowly diffusing substances have to form relatively

large holes for.the activated state; the activation energy and hence

the temperature coefficient of diffusion are consequently large" (102).

In the diffusion of substances through polymers, the diffusing

molecules are small compared to the molecules of the polymer. The

polymer molecules are also rigidly fixed in position with respect

to one another necessitating the diffusing species to move on its own.

This should demand higher activation energies than the 3-5 Kcal./mole.

postulated for diffusion in' liquid media. However the results for

inert gases and hydrocarbons (98, 99) demonstrate that the silicone










rubbers may be considered to be liquidlike elastomers with great

flexibility of atoms in an individual chain.

The effect of temperature on the rate of diffusion of

4'-aminopropiophenone through Silastic membrane is shown in Fig. 9.

The apparent activation energies of diffusion Ea were calculated

(Eq. 29) from the plots of the logarithm of the apparent diffusion

constants versus the reciprocal of the absolute temperature for amino-

alkylphenones (Fig. 10) and for barbituric acid derivatives (Fig. 17).

The lEa values reported (Table IV and XI) vary within the small

range of 4.9 to 7.5 Kcal./mole. except for mephobarbital and thiamylal.

The values of the apparent diffusion constants for amino-

alkylphenones and barbituric acid derivatives are smaller than those

reported (98, 99) for inert gases and the hydrocarbons. Consequently

the apparent activation energies were higher as had been anticipated.

The apparent diffusion constant D, is a product of intrinsic

diffusion constant D' in the membrane and the partition coefficient,

Kp (Eq. 37). Since Kp is the ratio (Eq. 38) of the solubilities of

the compound in the membrane Sm and the solvent for the drug solution

Sa, Eq. 3 may be rewritten as


log D = log D'Kp = log D'Sm/Sa = log DO 6Ea/2.303 RT (Eq. 40)


If log Sm = log SmO 6Hm/2.303 RT (Eq. 41)


log Sa = log SaO AHa/2.303 RT (Eq. 42)


and log 0' = log D'0 LED/2.303 RT (Eq. 43)


then log D'Sm/Sa = log SmoD'o/Sao-[E D + (tHm Ha)]/2.303 RT
(Eq. 44)










where EDI, is the true activation energy of diffusion inside the

membrane and --Im and -Ha are the heats of solution for the compound

in the membrane and the solvent phase. Smo, SaO and D'0 are the pre-

exponential factors in the Arrhenius equation, Eq. 3. Thus the

apparent activation energy of diffusion is the sum of the true activa-

tion energy of diffusion inside the membrane and the difference between

the heats of solutions of the diffusing species in the membrane and

the solvent phase.


Pharmaceutical Application


The diffusion of drugs through Silastic membrane in steady

state conditions is of zero order (Figs. 7, 9, 14, 15 and 18) when

the concentration of the drug in the solution on one side of the

membrane is kept constantly high by use of large volume of the

solution of high concentration or by using a saturated solution in

presence of large quantity of undissolved drug. The zero order

release of a drug from a dosage form at a predetermined rate to

exactly compensate the loss of drug from the body is ideal for a

sustained release formulation (103). The diffusion of drugs from

saturated solutions in Silastic capsules was studied with this view

in mind. The concentrations of the drugs inside the capsules were

thus kept constant for a long period of time.

The apparent diffusion constants for the compounds from

the capsules are given in Table XVII, together with those obtained

from the steady state diffusion studies in the diffusion cells, for the

purpose of comparison. The two sets of values are comparable to











each other. But the apparent diffusion constants calculated for

diffusion from Silastic capsules are significantly lower than those

from the diffusion cells. This is due to the fact that the area

available for diffusion in Silastic capsules was calculated from

geometrical considerations and the air entrapped in the capsules

reduced this available area. Also the thickness of the part of the

membrane changed to an unknown value because of the spread of the

glue used for sealing the holes in the capsules. The two sets of

values of apparent diffusion constants were, therefore, normalized

with respect to the apparent diffusion constant of barbital in each

set. The close agreement between these two sets of normalized values

demonstrate that the apparent diffusion constants from Silastic

capsules differed from those obtained with diffusion cells, by a

constant factor.

The apparent diffusion constants of dextromethorphan and

progesterone (Table XV, XVI and XVII) for diffusion from saturated

solutions in non-aqueous solvents are much lower than those from

saturated solutions in aqueous solvents. However because of the

higher solubilities of the drugs in the non-aqueous solvents, higher

concentrations are permissible and thus from these solvents the rates

of diffusion of these drugs are greater than those from aqueous

solvents (Figs. 19 and 20).

The specific rates of diffusion and apparent diffusion

constants of dextromethorphan from mineral oil were higher than those

from peanut oil (Table XV and XVII). This was expected as dextro-

methorphan is more soluble in peanut oil than in mineral oil and









consequently the coefficient of partition from peanut oil to the

membrane would be lower than that from mineral oil to the membrane.

However the rate of diffusion is higher from saturated solution in

peanut oil than that from saturated solution in mineral oil due to

the higher concentration of drug attainable in the peanut oil. Thus

the choice of a solvent to obtain a particular rate of diffusion from

saturated solutions may be made on the basis of the solubility of the

drug in a series of solvents.


Control and Prediction of Diffusion of Drugs

through Silastic Membrane


It has been shown conclusively that the diffusion of drugs

through Silastic membranes conform to Fick's law of diffusion. The

apparent activation energy for diffusion has been shown to vary within

a narrow range of 4.9 to 7.5 Kcal./mole. for most drugs studied.

If the diffusivity of a drug through Silastic membrane is

obtained experimentally at one temperature, the apparent diffusion

constants and the rates of diffusion at any other temperature can be

predicted from Eqs. 4 and 29


dA/dt = DS dC/dx (Eq. 4)

log D = log DO AEa/2.303 RT (Eq. 29)


using a mean value of 5.7 Kcal./mole. for the term LEa, the apparent

activation energy of diffusion. The optimum area S, and the thickness

X of the membrane to obtain a pre-determined rate of diffusion may be

calculated from Fick's law (Eq. 4) which reduces to










dA/dt = D S C/X (Eq. 45)


for a concentration C on the diffusing side of the membrane and when

the concentration of the diffusing species is negligible on the

desorbing side of the membrane.

The apparent diffusion constant has been shown to be a

linear function of the partition coefficient or the ratio of the

solubilities


D = D'Kp = D'Sm/Sa (Eq. 39)


where K is the partition coefficient and Sm and Sa are the solubilities

of a compound in the membrane material and aqueous solvents respectively.

The diffusivity of compounds belonging to a homologous series

can be predicted reasonably from the experimental value for diffu-

sivity of a compound belonging to the series, and the partition

coefficients and/or solubilities of these compounds. The linear

relationship between the solubility and the diffusivity may be used

to choose the best solvent to obtain a pre-determined rate of diffusion

of a drug by studying the solubility of the drug in several solvents

as in the case of 4'-aminopropiophenone in ethanolic phosphate buffers

or dextromethorphan in peanut oil and mineral oil.

Thus the area and the thickness of the membrane and the

solvent for the drug may be used to control the rate of diffusion.

The Fick's law of diffusion, the apparent activation energy of

diffusion and the knowledge of partition coefficients and/or solubility

of drugs may be used for predicting the rates of diffusion of homologous

series of drugs from the experimental value obtained for one of the

members of the series.















SUMMARY AND CONCLUSIONS


1. Nine polymeric films were screened for their permeability

to aminoalkylphenones, sulfonamides, steroids, barbituric acid

derivatives, and other drugs. Silastic membrane was found to be

permeable to many drugs in measurable quantities.

2. Silastic membrane was shown to be impermeable to HCI

and phosphate buffer salts.

3. The rate of diffusion of 4'-aminopropi.ophenone through

Silastic membrane was independent of ionic strength, and minor changes

in the hydrostatic pressure.

4. The effect of concentration of drug solutions and

thickness of Silastic membrane on the rate of diffusion of 4'-amino-

propiophenone was studied by steady state diffusion technique. The

zero order transport of the drug through the membrane conformed to

Fick's law of diffusion


dA/dt = D S dC/dx (Eq. 4)


where D is the apparent diffusion constant, A is the amount trans-

ported in time t across a membrane of surface area S and with a

concentration gradient of dC/dx.

5. The apparent diffusion constants D and the specific

rates of diffusion [(dA/dt)/(concentration of diffusing drug)] through

Silastic membranes were calculated for aminoalkylphenones, barbituric










acid derivatives, phenylalkylamines, dextromethorphan and progesterone.

6. The effect of temperature on the rate of diffusion of

barbituric acid derivatives and aminoalkylphenones through Silastic

membrane were studied and the activation energy of diffusion was observed

to vary between 4.9 and 7.5 Kcal./mole. for these compounds except

thiamylal (21.78) and mephobarbital.(15.22).

7. The effect of pH on the rate of diffusion of barbital,

pentobarbital and 4'-aminopropiophenone through Silastic membrane was

studied. The rate of diffusion was shown to be directly proportional

to the concentration of the uncharged species. The pKa values of these

compounds calculated from the plot of apparent diffusion constants

versus the pH of the solutions were observed to be the same as those

obtained by other methods.

8. The apparent diffusion constants in Silastic membranes

were found to be independent of the molecular weights of the diffusing

species.

9. The transport of a drug across Silastic membrane was

postulated to be due to partitioning of the drug from its solution

into the membrane, its transport across the membrane and then

repartitioning into the solution on other side of the membrane. This

postulated mechanism was substantiated by quasi-steady state diffusion

studies of 4'-aminopropiophenone from phosphate buffer solutions

containing varying concentrations of ethanol into ethanolic phosphate

buffer solutions. The intrinsic diffusion constant of 4'-aminopropio-

phenone in the membrane was independent of the composition of the

solutions, since the data could be fitted to the equations derived on

the assumption of constancy of intrinsic diffusion constant.










10. The apparent diffusion constants, D, of barbituric acid

derivatives and aminoalkylphenones in Silastic membranes were shown

to be lineraly related to their partition coefficients between

aqueous solutions and chloroform. This was shown to be in accordance

with the equation


D = D'K (Eq. 37)


where D' is the intrinsic diffusion constant of a drug inside the

Silastic membrane and Kp is the partition coefficient of the drug

between the membrane and the drug solution.

11. The apparent diffusion constants in Silastic membranes

were shown to be linearly related to the reciprocal of the solubilities

of barbituric acid derivatives in acetate buffer and 4'-aminopropio-

phenone in phosphate buffer containing varying concentrations of

ethanol. This was shown to be in accordance with equation


D = D'Sm/Sa (Eq. 39)


where Sm and Sa are the solubiliiies of a drug in the membrane and

the solvent for the drug solution.

12. The diffusion of dextromethorphan, pentobarbital,

phenobarbital, barbital, and 4'-aminopropiophenone from saturated

solutions in Silastic capsules was studied to investigate the

feasibility of using Silastic membranes as encapsulating material

for drugs. The rate of diffusion of dextromethorphan from peanut

oil was greater than that from mineral oil because of the high

concentration of the drug attainable in peanut oil. It was shown that










the rates of diffusion of a drug through Silastic membrane could be

controlled by choice of a solvent for the drug solution. If a

method of sealing the Silastic capsules, better than using a glue,

is available,Silastic membranes can be used to encapsulate a drug

solution to obtain its zero order release.

13. The area and the thickness of Silastic membranes and

solvents for the drug solutions can be chosen to control the rate of

diffusion of the drugs through Silastic membranes.

14. The apparent diffusion constants for diffusion of drugs

through Silastic membranes can be estimated from the knowledge of

partition coefficients and/or solubility of drugs belonging to a

homologous series and the experimental apparent diffusion constant

of one of the drugs from the series.






































A. Tables











TABLE I

WAVELENGTHS OF MEASURED ABSORBANCES, MOLAR ABSORPTIVITIES AND pKa

VALUES OF COMPOUNDS.



Compounds pKa Wavelength Molar
Measured Literature mp Absorptivity
value (104)

Amobarbital 7.40 238 5850
Barbital 7.45 7.91 238 8300
Butabarbital 238 10410
Cyclobarbital 7.27 7.50 238 8700
Diallyl barbi-
turic acid 7.30 7.79 238 5990
Mephobarbital 7.45 238 6050
Metharbital 7.90 238 7700
Pentobarbital 7.65 8.11 238 8900
Phenobarbital 7.40 7.41 238 6800
Secobarbital 7.45 8.08 238 7700
Thiamylal 6.80 304 21000
Thiopental 7.12 304 13240
Dextromethorphan 8.25 277 3002
4'-aminopropio-
phenone 2.42 307 16016
4'-aminoaceto-
phenone 312 13700
3'-aminoaceto-
phenone 327 1840
a-methyl phen-
ethylamine 9.07 520
a-ethyl phen-
ethylamine 9.30 520
2-amino-4-methyl-
4-phenylpentane 9.42 520
3-amino-l-phenyl-
butane 9.30 520
1-methyl-5-phenyl-
pentylamine 9.50 520
Progesterone 248 16470
Cortisone 238 16000
Hydrocortisone 242 16200
Prednisolone 242 15200
Sulfadiazine 255 5370
Sulfathiazole
Sulfisoxazole
Sulfabenzamide 255 4480










TABLE I I

DIFFUSION OF 4'-AMINOPROPIOPHENONE FROM SATURATED SOLUTION (2.36 x 10-3M)

IN pH 6.5 PHOSPHATE BUFFER THROUGH 3 MIL SILASTIC MEMBRANE INTO 200 ML.

OF 0.12 N HCI AT 37.30.



Date Time Time Since Tempera- Absorbance at 307 mpa
Zero Hrs. ture C 1 2 3 4 5

7/16 10.00 0 37.30 0.007 0.004 0.008 0.014 0.006

11.00 1.00 37.30 0.134 0.141 0.116 0.121 0.112

12.00 2.00 37.30 0.234 0.224 0.232 0.207 0.209

13.00 3.00 37.30 0.357 0.319 0.340 0.287 0.319

14.00 4.00 37.50 0.478 0.428 0.460 0.395 0.433

15.00 5.00 37.50 0.560 0.515 0.532 0.470 0.505

16.00 6.00 37.30 0.663 0.602 0.653 0.559 0.612

18.00 8.00 37.30 0.886 0.815 0.875 0.722 0.825


Volume in the cell in ml. 21.0 22.0 23.0 23.0 22.0

pH of cell solution 6.55 6.55 6.50 6.50 6.55

Volume of solution in
beaker in ml. 180 184 184 184 184

pH of solution in the
beaker 1.00 1.00 1.00 1.00 1.00


a The sample solutions were diluted 1:5
before measurement of absorbances.


with pH 6.5 phosphate buffer



















m-
LU





u
>






O






I-

0

0

ro







U-
0
Ct


U)












I--
z



o


C C


LL
LL-



0
C)
z
0











1-

-
L<

O
-




- 2
- L





O
(0



I-









O
,
U)



C )
























..-I
O --
0
Q)

2:
0\







Cu
U)



z~i
0 C
^
-J


u
a)

0
nU) E
O u

4--

1C-
4-
CC
Q -
c c

L -







Q-
< 0





-O



0-
U')
LO






0


-o a)
C -


u
4- -
4-

(U





4- Q).

L L

CL Q.
X



0
>-
t

no
II


0-
0*-
.O
3-
S-

















3

E
0
U
O
C.)


U

L
U -
( 4- (o (D (0 -
- (u 4- 4- 4- 4- (
(u 4- *- (u *- +- -
4- -O L -0 4- -- ru
- Q L (DU CO L L L 0 (C 4-
.-0 L u ( .0 (O -O u C(D L -
L 4- CO >- .- L 3 -0 (0 > (
0 -- -0 O O C O O E o0
-o n Co -C -C 4- c 0 (C 0
0 L +- U Cu 0- 4- C 0) U
E Cu >. : -- () (a) ) ) C r _-
< C 0 OC) 0 2 2 a- a- O F-


CO N 0 CO
C M N Ot 0
n o C -













Ot O O N














vao
0 0 0











N N











O C
cO 0) c0 CO

U) d U) -








NC
N CO 0 m





N 0co 00 r

O Cr- r- r-:
i-1- -


ko s in t Q n r\ in O

0 0 O 0 m Nm 01














S N 0 00

o 1 ^- ^- N













CO 4- 0










o I 0 0 00 -t I I


0) in CN L)n a)
> > >







N 0
r- 'o 0 rn NQ r

O 0 U n r-: O 0


L -
a)
C- E

0 rw


( 0





(0
-o






G) 0
+- i
+_










L +-
r L

0




C-)
-C
-+- o
(D














4-
0
-n N















|c +L


0- -
-F- "I
(a C CD
L 4-




O (5
4-
0 0)


a)
u









L 0C

C L-

O 4-0






0 (n
4- S- C E
n o C
*) -.

Cu L C



4- -0 0 (D
) 4- 4-




C
S 0 U1-


ocD


c: 4-

00 E
I C -
C u-
4- (D

C 1) 4-
0 aS I V)
3 U (D (0
C 0--
0 0 CL-
c C)
(U -0 U