UC Berkeley Research

LBNL-48573
Preprint
ERNEST ORLANDO LAWRENCE
BERKELEY NATIONAL LABORATORY
Distribution Coefficients
and Diffusivities in Three
Polymers for Nineteen
Aqueous Nonvolatile Solutes
F. Fornasiero, M.M. Olaya, B. Esprester,
V. Nguyen, and ].M. Prausnitz
Chemical Sciences Division
June 2001
Submitted.,to'_.__._....;.-.:.................. ..... "
Journal"oJApplied .'--.
Polymer Science' .

DISCLAIMER
This document was prepared ·as an account of work sponsored by the United States
Government. While this document is believed to contain correct information, neither the
United States Government nor any agency thereof, nor the Regents of the University of
California, nor any of their employees, makes any warranty, express or implied, or
assumes any legal responsibility for the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or represents that its use would not
infringe privately owned rights. Reference herein to any specific commercial product,
process, or service by its trade name, trademark, manufacturer, or otherwise, does not
necessarily constitute or imply its endorsement, recommendation, or favoring by the
United States Government or any agency thereof, or the Regents of the University of
California. The views and opinions of authors expressed herein do not necessarily state or
reflect those of the United States Government or any agency thereof or the Regents of the
University of California.

Distribution Coefficients and Diffusivities in Three
Polymers for NineteeRAqueous Nonvolatile Solutes
LBNL-48573
F, Fornasiero, M.M. Olaya, B. Esprester, V. Nguyen, and J.M. Prausnitz
Chemical Engineering Department
University of California, Berkeley
and
Ernest Orlando Lawrence Berkeley National Laboratory
University of California
Berkeley, California 94720
June 2001
This work was supported by the Director, Office of Science, Office of Basic Energy Sciences, Chemical
Sciences Division, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

Distribution Coefficients and Diffusivities in Three Polymers
for Nineteen Aqueous Nonvolatile Solutes
F. Fomasiero, M. M. Olaya1
, B. Esprester
2
, V. Nguyen3
, and J. M. Prausnitz4
Chemical Engineering Department
University of California, Berkeley
and
Lawrence Berkeley National Laboratory
Berkeley, CA 94720, USA
1 Present address: Departamento de Ingenieria Quimica, Universidad de Alicante, Alicante, Spain
2 Present address: Lehrstuhl fUr Feststoff- und GrenzfHichenverfahrenstechnik, Technische Universitat, Miinchen,
85748 Garching, Germany
3 Present address: Chemical Engineering Department, University of Califomia, Los Angeles, CA 90095-1592
4 To whom correspondence should be addressed. Email address: prausnit@cchem.berkeley.edu
1

Synopsis
Experimental results at 25°C are reported for infinite-dilution distribution coefficients for
nineteen non-volatile solutes between aqueous solution and three kinds ofpolymer films, and for
their diffusion coefficients in the polymer matrix. The experiments were performed by coupling
UV-spectroscopy and gravimetric measurements with mass balances. The solutes are aromatic
non-volatile compounds that are ofinterest in environmental technology and may serve as model
compounds for drug-delivery systems. The polymers are ethylene-vinyl acetate copolymers with
33 (EVAc33) and 45 (EVAc45) weight percent vinylacetate, and poly(vinylacetate) (PVAc),
widely used in drug delivery devices.
Infinite-dilution distribution coefficients Ks range from near unity to several hundred. Ks
tends to increase with vinyl acetate content in the polymer because the vinyl acetate group
interacts favorably with the polar functional groups in the solute. The nature and position of
functional groups in the solute affect Ks; ifthey augment lipophilicity, the distribution coefficient
tends to increase in EVAc copolymers and viceversa. In PVAc, the effect appears to be more
complex: polar moieties can raise Ks because of favorable interaction, but non-polar groups may
give the same effect because of increased lipophilicity.
In EVAc copolymers, diffusion coefficients D range from 10-10
to 10-8
cm2
/sec; they are
weakly dependent on VAc and usually larger for the solute with lower distribution coefficient. In
PVAc, the diffusion process is much slower (D_1O-12
cm2
/sec). As a consequence, the time to
equilibrate the polymer with the aqueous solution of the solute is quite long (more than 4-5
weeks). An extrapolation procedure with a suitable diffusion model has been used to obtain the
equilibrium distribution coefficient and the diffusion coefficient.
2

Experimental results and calculations with the Langmuir isotherm show that the
contribution ofsurface adsorption is negli¥ible.
3

Abstract
Experimental results at 25 DC are reported for Infinite-dilution distribution coefficients for
nineteen non-volatile solutes between aqueous solution and three kinds ofpolymer films, and for
their diffusion coefficients in the polymer matrix. The experiments were performed by coupling
UV-spectroscopy and gravimetric measurements with mass balances. The solutes are aromatic
non-volatile compounds that are ofinterest in environmental technology and may serve as model
compounds for drug-delivery systems. The polymers are ethylene-vinyl acetate copolymers with
33 (EVAc33) and 45 (EVAc45) weight percent vinylacetate, and poly(vinylacetate) (PVAc),
widely used in drug-delivery devices.
For PVAc, a long time is required to reach equilibrium. Because the required time is too
long for reasonable experimental studies, equilibrium distribution coefficients were calculated
from finite-time data using a diffusion model. The contribution ofsurface adsorption is shown to
be negligible. Infinite-dilution distribution coefficients Ks, defined as the volume fraction of
solute in the polymer divided by that in water, tend to increase with vinyl acetate content; they
range from near unity to several hundred. Diffusion coefficients, determined from time-
dependent sorption data, are significantly larget in EVAc copolymers (10-10
to 10-8
cm2/sec) than
in PVAc (10-12 cm2/sec).
These data may be useful for design of membrane processes, for controlled delivery of
drugs and for application in packaging and storage offood, chemicals and pharmaceuticals.
Keywords: diffusion, distribution coefficient, nonvolatile solutes, ethylene-vinyl. acetate,
poly(vinyl acetate).
4

Introduction
Numerous applications of polymeric materials concern polymer films in contact with an
aqueous solution of heavy organic solutes, i.e. those with a negligible vapor pressure at room
temperature. For example, in wastewater purification, polymer membranes enable the removal of
small amounts ofpollutants from water streams and the potential recovery ofuseful compounds.
Polymeric bags, containers and wrapping films are extensively used for 'packaging of food and
storage of pharmaceuticals or other chemicals. Leaching of polymer-processing chemicals could
contaminate the stored product. On the other hand, sorption of compounds from a stored
pharmaceutical solution could compromise its effectiveness by lowering the compound's
concentration.1
-
3
Distribution-coefficient data are helpful for selection of suitable polymers for
separation processes and for packaging.
Most pharmaceuticals are organic compounds with very low volatility at room
temperature. In controlled drug-delivery systems, polymers serve as a supporting matrix for a
uniformly dissolved or dispersed solute (matrix systems) or as permeable barriers surrounding
the stored drug (reservoir systems) as discussed, for example, by Langer.4
The solubility of the
solute in the polymer and the distribution coefficient between the aqueous phase and the
polymer, together with the diffusion coefficient are controlling properties in drug-release to the
surrounding media.5
These properties affect the ability of a drug to reach significant
concentration in the blood stream for distribution into tissues.
The polymer-water distribution coefficient and the diffusion coefficient in the polymer
are essential parameters for rational design ofappropriate devices for specific purposes.
In this work, we report distribution coefficients for nineteen non-volatile organic solutes .
between a dilute aqueous solution and a polymer film at 25°C. For each solute, the diffusion
5

coefficient in the polymer matrix is also obtained at the same temperature. The solutes are
aromatic compounds with different functional groups with molecular weights ranging from 79 to
182 g/mol; most of them are solids at 25°C. They may be considered as models for typical
pharmaceuticals.
Ten ofthe solutes are aromatic nitro-compounds, an important class of substances with a
variety of applications; they are used as solvents, fuel additives, intermediates in manufacture of
dyes and explosives, etc... Nitroaromatics present a serious environmental hazard because of
their high stability and tendency to leach into groundwater; they frequently appear on lists of
high-priority pollutants. Under environmental conditions, they are encountered at very low
concentrations.6
Naphthols are used directly i~ manufacture of dyes and are converted into
tanning agents, antioxidants, and antiseptics; acetophenone is used in perfumery industry, as
solvent and as intermediate in organic synthesis; benzophenone is used for pesticides, perfumes,
pharmaceuticals, and as organic ultraviolet (UV) absorber. Benzoic acid is mainly used in dye
and paint industries; benzyl alcohol is used in the soap, perfume, and flavor industries "and as
solvent for inks and lacquers.
The polymers are two ethylene-vinyl acetate copolymers (33 and 45 weight % vinyl
acetate) and poly(vinylacetate). Ethylene-vinyl acetate copolymers (EVAc) find a variety of
industrial applications thanks to their chemical resistance, flexibility and excellent processability.
Their biocompatibility and biological inertness make them appealing for biomedical
applications. EVAc copolymers with 30-50% VAc content are FDA-approved materials
employed as non-degradable polymers for drug-delivery systems.7
-
10
6

Materials
Tetrahydrofuran (THF) was purchased from EM Science and used as a solvent to cast the
polymer films. Ultrapure water (Barnstead NANOpure system) was used to prepare the aqueous
solutions.
Table 1 lists the solutes; for each solute, Table 1 gives chemical formula, light sensitivity,
purity, and supplier. Table 2 presents the most relevant solute properties: molecular weights,
densities and melting points Tf eC). All solutes were used without further purification.
EVAc copolymers were. obtained from Scientific Polymer Product, while PVAc was
purchased from Aldrich. Table 3 gives the average molecular weight Mw, purity, composition
and glass-transition temperature. The glass-transition temperature for EVAc has been estimated
using the Flory-Fox equation:!!
1 wpl wp2
-=--+--
~ Tg'PI ~,p2
(1)
where Tg is the glass-transition temperature of the copolymer, while Tg,p! and Tg,p2 are the glass-
transition temperatures of the parent homopolymers, and wp1 and wP2 are the mass factions of the
parent monomers. Possible uncertainties in Equation (l) are not relevant for our measurements
because the glass-transition temperature is much lower than 25°C. The glass-transition
temperature of PVAc has been measured by modulated differential scanning calorimetry (DSC
29~0, MQdulated DSC TA Instrument) at a heating rate of 3 °C/min, with modulation amplitude
of±1 °c and period of60 sec.
The polymers are polydisperse. Experimental studies by Nguyen!2 have shown that the
effect of Mw on Ks is negligible for rubbery polymers at high degree of polymerization
(MW>80000). Polydispersity may affect the solubility and the distribution coefficient of a solute
7

in a glassy polymer by increasing the glass-transition temperature and reducing polymer chain
mobility. However, the EVAc copolymers and the water-presaturated PVAc (see the next
section) used in this work are rubbery at room temperature;13 therefore, polydispersity is not
expected significantly to affect our results.
Polymer-film densities have been determined using a 10-ml pycnometer. The densities
reported are averages over several measurements; typical uncertainty (standard deviation) is less
than 2 weight percent. Measured densities agree within the experimental error with those
reported by the producer (the latter are 2 % higher for EVAc copolymers, the same for PVAc).
Distribution coefficient
In the following section, we present the experimental procedure to determine the
distribution coefficient in EVAc33. For PVAc and EVAc45 we report only the variations applied
to the procedure for EVAc33. Table 4 summarizes the characteristic features of preparation of
the distribution-coefficient experiment for each polymer.
a) Experimentalprocedure
EVAc33
To obtain the infinite-dilution distribution coefficient K; ,the weighed polymer samples
were put into contact with a weighed dilute aqueous solution ofthe selected solute.
To reduce the time to reach equilibrium, the polymer samples were prepared as thin
films. A liquid polymer solution was obtained by dissolving the polymer in tetrahydrofurart
(THF). The polymer concentration in THF should be low to avoid problems with liquid transfer
due to high viscosity, but not too low, lest the time for evaporating THF is too long. Typical
polymer concentration was 6% (gr polymer/ml THF). Consistent with the desired film thickness,
a small amount of polymer/THF solution was poured into a 63-mm-diameter aluminum dish. 20
8

ml/pan were used for EVAc33. Each aluminum pan was then covered and put into an oven,
fluxed with nitrogen, for a heating time of40 hours at 100°C. The drying temperature is chosen
well above the normal boiling point of THF (66°C) and well above the glass-transition
temperature of the polymer' (estimated to be -95 °C for EVAc33) to allow complete drying.14
After that, the polymer melt in the aluminum pan was allowed to cool to room temperature for at
least 8 hours. Next, a polymer film was removed from the aluminum pan and cut into strips. For
EVAc33, the average thickness of the polymer film was 0.266 mm (measured by MDC-1"Pf
Micrometer from Mitutoyo).
Polymer films should be totally free of solvent after casting. Thermogravimetric analysis,
using 5-10 mg polymer sample at heating rate 5 °C/min, showed that all solvent in the polym~r
film is absent (TG 220 SSCI5200 SII Seiko Instruments).
EVAc45
EVAc45 films were cast from an 8% THF solution. Pyrex dishes (139-mm diameter)
were filled with 84-ml polymer solution each, covered, and heated to 100°C for 40 hours under
a nitrogen flux. The dishes were then stored at -20 °C in the refrigerator for a minimum of one
day. The cold polymer films were detached from the Pyrex container as soon as possible after
removing the dishes from the refrigerator. Use of Pyrex dishes (instead of aluminum pans)
combined with the cooling procedure and the use of a higher film thickness (0.421 mm)
minimized problems in the recovery ofthe polymer films from the dishes caused by the adhesive
properties ofEVAc45. For EVAc33, it was easy to peel offthe polymer film from the aluminum
pan, but not for EVAc45.
PVAc
9

10 ml of a 4% THF polymer solution were poured in each aluminum pan to obtain 0.101
mm thick PVAc films. The temperature and the heating time are the same as those used for
EVAc33. The polymer films, removed easily from the aluminum pans, were then placed into
contact with nanopure water and allowed to equilibrate for 2 weeks. The water sorption lowers
the glass transition temperature ofthe wet PVAc films; with roughly 2.5 weight % water content
(smaller than the measured equilibrium value at 25°C) they become rubbery at room
temperature. Therefore, there is no glass transition during subsequent distribution-coefficient
experiments. Moreover, pre-equilibration with water provides additional removal of possible
impurities from the polymer films.
After the polymer films were prepared, several strips ofEVAc33 (total weight ~1 g) were
placed into an equilibrium vial, in contact 'With a weighed amount of aqueous solution of the
solute at known concentration. Because of the strong adhesion properties of EVAc45 and wet
PVAc, the polymer strips tend to stick together reducing the contact area with the aqueous
solution. This compromises the reproducibility of distribution coefficient data, especially for
those systems that require a long time to reach equilibrium. Therefore, for EVAc45 and PVAc,
only one strip was used in each vial (respectively ~ 0.5 g and ~ 0.2 g).
The optimum initial concentration ofthe solute in water was determined by trial and error
for each solute-polymer pair. The concentration ofthe solute in the aqueous solution must not be
too large lest the change in aqueous solute concentration before and after sorption become too
small for accurate measurement. On the other hand, the concentration must not be too small lest
the amount of sorbed solute be difficult to detect because of limits in analytic accuracy. To
minimize loss of aqueous solution during repeated measurements at different times, the solution
10

concentration should also allow direct UV spectroscopic measurement of absorbancy without
dilution. In this case, the sampled solution can be recovered for subsequent measurements. To
minimize any error caused by evaporation of water in the equilibrium vial, each vial was
completely filled with solution and polymer and then sealed with Teflon tape or septa and
capped tightly. The equilibrium vials prepared for the sorption experiment were maintained in a
thermostatic bath at 25 ± 0.1 °C until equilibrium was nearly obtained. Time required for near-
equilibrium was a minimum of a few days for EVAc copolymers, and more than 4-5 weeks for
PVAc. As discussed in Appendix I, the near-equilibrium data for PVAc were fit to a diffusion
model that permits rational. extrapolation with time to obtain equilibrium distribution
coefficients. For the light-sensitive solutes, precautions were taken to minimize the exposure to
light by using brown glass vials for the sorption experiment and for storing aqueous solutions. -,
Using an ultraviolet-visible spectrophotometer (Shimadzu UV-160), the solute
concentration in water was measured prior to contact with the polymer films and then, as a
function of time extending over several weeks, at regular intervals. These data, coupled with the
known mass ofsolution and mass ofpolymer, were used to determine the concentration of solute
in the polymer as a function oftime.
.Data reduction
The ratio of solute concentrations (expressed as volume fractions) in the polymer and in
the aqueous phases gives the time-dependent distribution constant Ks before extrapolation, as
shown below:
(2)
(3)
11

ms,p
W =---.....:..:....---s,p 0
ms,p + mw,p + mp
(4)
(5)
where, mi,j is the mass ofcomponent i in phase j; Wi,j is the mass fraction ofi in phase j; <1>i,j is
the volume fraction of i in phase j; i or j = s (solute), w (water), or p (polymer); sUbscript sol
stands for aqueous solution and superscript 0 stands for initial.
For data interpretation through thermodynamic analysis, it is convenient to obtain the
distribution coefficient at infinite dilution K; .For a particular polymer-solute pair, distribution-
coefficient data were therefore obtained at five levels of dilute aqueous-solution concentration.
Because the experimental distribution coefficients do not show any trend with the final solution
concentration, the experimental points are considered to be inside the infinite-dilution regime
and K; was calculated as a simple average over these five values. More precisely, this
procedure gives the ratio of concentrations <1>s,p /<1>S,w at infinite dilution as a function of time.
These data, in tum, were reduced to give the equilibrium (corresponding to infinite time)
distribution coefficient K; at infinite dilution as discussed in Appendix 1.
To convert weight fraction into volume fraction, we introduced in Equation (5) the
densities ofthe solutes in the actual state at the experimental temperature. While a more rigorous
calculation should use the density in a hypothetical liquid state for the solid solutes at 25°C,
negligible errors are introduced by this approximation. The precise value of ps in Equation (5)
has no appreciable influence on the distribution coefficient because ws,p and Ws,w are much
smaller than W w,wand wp,p'
12

Water sorption
To close the, mass balance (Equations (3)-(5)), the amount of water in the polymer at
saturation is also required. It was obtained from a separate experiment with solute-free water and
was assumed to be independent of the solute. Thermogravimetric analysis was performed to
determine the water content of water-saturated polymer films. Fifteen days of contact were
suffic~ent to reach saturation. The equilibrium water content has been determined from the loss
ofweight of the water-saturated polymer sample after heating to at least 100°C as illustrated in
Figure 1. An independent measure ·of the polymer weight before and after equilibration with
water gives consistent results as indicated in Table 5.
Diffusion coefficient
.To determine the diffusion coefficients of the solutes in the polymer films, we used a
short-time approximation to the solution ofthe diffusion problem illustrated in Appendix 1:15
(6)
where <l>s,p is the volume fraction of the solute in the polymer at time t, <l>C() is the equilibrium
volume fraction of the solute in the polymer, T = ~t and a =_a_; D is the diffusion
I I·Ks
coefficient of the solute in the polymer, a/I is the ratio of the volume of solution to that of
polymer and Ks is the equilibrium distribution coefficient.
Performing a Taylor expansion and applying the solute mass balance, Equation (6) simplifies to
(7)
13

where <l>s.w is the solute volume fraction in the aqueous solution. Equation (7) shows that for (~O
the solute concentration decreases linearly with (1/2 and that the slope is proportional to the
distribution coefficient andJi5. Once the distribution coefficient has been determined using
Equation (5), the diffusion coefficient can be calculated from the slope of the straight line when
<l>s.w(()/<l>~,w is plotted against (1/2.
The data required for Equation (7) were obtained by measuring as a function of time the
decrease of the solute concentration in the aqueous solution near the beginning of the
experiment, i.e. every 2-5 min for 30-60 min for EVAc copolymers, and every 0.5-1 hour for 8-
10 hours for PVAc. In these time intervals the experimental solute concentrations are linear
function ofthe square root ofthe time. .
Results and discussion
Contribution ofsurface adsorption
Equations (2)-(5) do not distinguish between solute adsorbed on the polymer film surface
and solute absorbed in the polymer bulk. Because our interest is in the bulk properties of the
polymer/aqueous solution system, it is necessary to estimate the contribution of surface
adsorption to the total sorption. This problem was addressed in two ways. First, sorption
experiments were performed varying the thickness of the polymer film while keeping constant
the polymer mass as well the mass and concentration of the aqueous solution. Second, the
Langmuir adsorption isotherm was used to provide an upper-bound estimate of the amount of
adsorbed solute. For a conservative estimate, we assumed a large adsorption-energy parameter
Kads = 100-1000 to favor solute adsorption, and we estimated the monolayer coverage nm as the
14

ratio of the typical surface area of the polymer films to the projected van der Waals areal6
of a
solute molecule.
The experimental approach is meaningful only if the measurements of solution
concentration are at long time, close to the equilibrium condition. Because the amount of solute
diffused into the polymer at a given time is inversely proportional to the film thickness, the time-
dependent distribution coefficient decreases with rising film thickness while the diffusion
process is proceeding. At equilibrium, for films of varying thickness, the distribution coefficient
attains a common value only if there is no surface adsorption. On the other hand, if the
equilibrium distribution coefficient depends on film thickness (surface-to-volume ratio), there is
significant surface adsorption contribution. At non-equilibrium conditions, the dependence of
<Ds,p Ie!)S,w on the film thickness cannot be discerned from the eventual equilibrium su..rface
adsorption contribution. For this reason, and to reduce the experimental effort, we used this
experimental approach only for EVAc33 but not for the other polymers. Figure 2 shows typical
results for EVAc33 with an aqueous solution of acetophenone, or 2-nitroaniline or 2-nitrophenol.
The zero slope indicates that surface adsorption for all examined EVAc33/s01ute systems is
negligible.
The calculation based on the Langmuir isotherm shows that for both EVAc copolymers
and PVAc, the contribution ofsurface adsorption is less than 0.5% ofthe total sorption.
Finally, the early-time measurements also indicate a negligible effect of surface
adsorption. The typical equilibration time for surface adsorption of a small solute on a non-
porous solid is of the order of a few miilUtes.17
Therefore, a plot of the solute concentration in
the aqueous solution as a function of the square root of the time should present a nearly
instantaneous decline at the beginning, corresponding to the amount adsorbed. None of the )
15

systems studied here shows such a decline; see, for example, Figure 8. In conclusion, Equation
(5) can be used to evaluate the bulk distribution coefficient without introducing any appreciable
error due to surface adsorption.
Water sorption
Table 5 reports the measured water sorption in EVAc33, EVAc45 and PVAc. The weight
per cent of water at saturation (at T=25 DC) increases with VAc content. The water content is
negligible in EVAc33; it rises to ~0.4 weight % in EVAc45 and reaches 7.0% in PVAc. Our
measured water contents in EVAc copolymers are in agreement with those reported by Chen and
Lostritto,18 who find less than 0.5% water content for copolymers-with VAc content between 0
and 40%. PVAc films are transparent in the dry state and become cloudy after equilibration with
water. Johnson et al. 19
associated the whitening to water clustering in the polymer matrix. They
showed that PVAc absorbs up to about 4 weight % water that is bound to the polymer and is
responsible for the decrease of !g below room temperature. Additional sorbed water (clustered
water) fills microscopic cavities within the polymer matrix and has no effect upon Tg for PVAc.
Therefore, roughly 3 weight % of clustered water is present in our PVAc films. EVAc45 is also
transparent when dry but becomes cloudy when wet.
Distribution coefficient
Figure 3 shows distribution coefficients for three different solutes in EVAc33 as a
function of the solute equilibrium volume fraction in the aqueous phase. In the experiments, the
masses of both phases and the thicknesses of the polymer films were held constant. The error
bars and the average distribution coefficient are also shown. Clearly, the measured distribution
coefficients are independent of aqueous solute concentration for the concentration range
considered here.
16

Figure 4 shows similar data for EVAc45; analogous conclusions can be drawn. In Figures
5-7 CPVAc), there is again no evident trend of the distribution coefficient with the solute
concentration. Here, however, the concentration in the horizontal axis is a non-equilibrium
concentration because of the long time necessary to reach equilibrium. Comparing the ratio of
solute volume fractions in the two phases at different times, time-dependent distribution
coefficient <1>s /<1> remains independent of dilution. Therefore, the assumption of infinite-,p s,w
dilution conditions for PVAc remains valid for all times up to the time required for equilibrium.
Table 6 lists distribution coefficients with corresponding standard deviations for all
solutes in EVAc33, EVAc45 and PVAc. Because EVAc copolymers in contact with aqueous
solutions reach equilibrium in a few days, extrapolation to infinite time is not required and the
reported Ks are simple averages of the final measurements. However, for PVAc systems, the
diffusion process is much slower and a long time is needed for equilibration. Therefore, the
reported distribution coefficients are the average of the 5 values (corresponding to the 5 different
solution concentrations) of the fitting parameter Ks appearing in the diffusion model used for
time extrapolation (Appendix I).
In general, the distribution coefficient of a solute is affected primarilyby two factors: the
presence of functional groups on the solute molecule that may interact specifically with the VAc
moiety, and the lipophilicity of the solute as a measure of how much the solute dislikes the
aqueous environment (hydrophobicity).
Highly polar molecules like pyridine and benzyl alcohol have indeed the lowest Ks (-1-
10) while benzophenone has the highest (;"'103
). For a given solute, Ks increases with VAc
content in the polymer with few exceptions: 2 and 4-nitrotoluene and benzophenone; for these
solutes, EVAc45 gives the highest distribution coefficient. The addition ofpolar VAc moieties in ..
17

the polyethylene backbone appears to increase the favorable interaction between the solute and
the polymer. Each solute studied here has a polar functional group (nitro, amino, hydroxyl,
carbonyl, carboxyl, etc...) attached to the aromatic ring; specific attractive interactions between
these groups and VAc are likely. Maurin et al. 10
observed similar behavior for monosubstituted
benzoic acids in EVAc copolymers. They explained their results on the basis of complex
formation between the carboxylic group and the vinyl acetate moiety because experimental
studies have shown that these aqueous monosubstituted benzoic acids have essentially zero
partition coefficients in polyethylene.
4-Nitroaniline in EVAc copolymers exhibits a smaller Ks than nitrobenzene (92, 145,- 172
111 EVAc33, EVAc45 and PVAc respectively). The additional amino group decreases the
lipophilicity of the solute increasing its partition into the aqueous solution. When the amino
group moves from the para to the ortho position, intramolecular interaction with the adjacent
nitro-group may reduce this effect causing an increase in lipophilicity and in the distribution
coefficient.
For all three polymers, the distribution coefficient of 2-nitrophenol'is the same as that of
nitrobenzene. The addition of a hydroxyl moiety has a parallel effect in EVAc copolymers; the
ortho-substituted isomer has th~ higher Ks while the meta and para isomers show nearly the same
Ks.
For nitrophenols, PVAc presents an opposite trend while for nitroanilines the behavior is
more complex. For both solutes, the additional polar group raises Ks above that for nitrobenzene.
The favorable interaction between polar moieties appears here to be the major driving force for
partition into the polymer.
18

For nitrotoluenes, EVAc copolymers and PVAc show similar behavior. The methyl group
favors partition into the polymer by increasing the lipophilicity ofthe solute.
Diffusion coefficient
Table 7 gives diffusion coefficients for all solutes in EVAc33, EVAc45 and PVAc as
obtained with the method presented by Equations (2)-(5). The standard deviation is also reported.
Typical experimental data are presented in Figure 8. The diffusion coefficients for solutes in
EVAc45 and EVAc33 range from 10-10
to 10-8
cm2
/sec, while for the wet PVAc they are two
orders of magnitude smaller. With few exceptions, diffusion coefficients in EVAc45 are a little
larger than those in EVAc33.
Our data are in reasonable agreement with prevIOUS results for similar aromatic
compounds. Chen and Lostrittol8
measured the diffusion coefficient of benzocaine (Mw=165.2)
in EVAc copolymers with different VAc content; their results for 28%-40% VAc are ofthe order
of 10-9
_10-8
cm2
/sec, slightly increasing with %VAc. Maurin et al. 0 report diffusion coefficients
for monosubstituted benzoic acids in EVAc copolymers at 37°C; they found D - 10-9
_10-8
cm2
/sec and little change in the diffusion coefficie~t with VAc content in the range 0-40%.
Kumar et al.2o
obtained diffusion coefficients for benzene, toluene and xylene in EVAc with
18% VAc; here, the results are ofthe order of 10-7
cm2
/sec. To Qur best knowledge, no diffusion-
coefficient data have been reported in literature for aromatic solutes in wet PVAc.
Table 7 also shows diffusion coefficients for solutes in PVAc obtained by the
extrapolation procedure described in Appendix I. Agreement between the two methods is good.
The experimental uncertainty of the diffusion coefficient obtained through the extrapolation
procedure is smaller than that obtained from the early-time measurements because in the former,
the uncertainty in the film thickness is not taken into account.
19

For a given polymer/solute system, there is a weak correlation between the diffusion
coefficient and the distribution coefficient; as the latter increases, the former decreases, as shown
in Figure 9 for EVAc33. The stronger polymer/solute interaction in the systems (indicated by a
higher distribution coefficient) retards solute motion in the polymeric matrix. Results for isomers
of the same solute show some exceptions; however, these exceptions may not be significant
because the uncertainties in the data are often quite large, perhaps as high as ±50%.
Conclusion
For nineteen aqueous non-volatile solutes, experimental results are reported for
distribution coefficients at infinite dilution and diffusion coefficients in three polymers. Solutes
are aromatics with a variety offunctional groups; the polymers are two EVAc copolymers and
PVAc. The solutes are of environmental interest and may serve as models for pharmaceuticals;
EVAc copolymers are widely used as rate-controlling membranes in drug-delivery systems.
Surface adsorption is negligible. The distribution coefficients Ks increase with vinyl
acetate content in the polymer because the vinyl acetate group interacts favorably with the polar
functional groups in the solute. The nature and position of functional groups in the solute affect
Ks; if they augment lipophilicity, the distribution coefficient tends to increase in EVAc
copolymers and viceversa. In PVAc, the effect appears to be more complex; polar moieties
(hydroxyl and amino group) can raise Ks because of favorable interaction with the vinyl acetate
group, but non-polar groups (methyl) may give the same effect because of increased
lipophilicity.
In EVAc copolymers, diffusion coefficients D are weakly dependent on VAc; they range
from 10"10 to 10-8
cm2
/sec. They are usually larger for the solute with lower distribution
coefficient. In PVAc, the diffusion process is much slower (D_1O-12
cm2
/sec). As a consequence,
20

the time to equilibrate the polymer with the aqueous solution of the solute is quite long (more
than 4-5 weeks). Therefore, for solutes in PVAc, an extrapolation procedure with a suitable
diffusion model has been used to obtain the equilibrium distribution coefficient and the diffusion
coefficient.
Acknowledgement
For financial support, the authors are grateful to the Office of Basic Energy Sciences, US
Department of Energy and to the Donors of the Petroleum Research Fund administrated by the
American Chemical Society. For research grants, MO thanks the Ministerio de Educacion y
Ciencia (Spain) and BE thanks the Solvay Foundation. For extensive assistance in experimental
work, all authors are grateful to Nancy Fung, Claire Kimball, Shirley Poon, Wilson Wong, Chris
Lee, Susan Hsiong, Jenny Fang, Lauren Bergeson, Nausheen Husain, Raymon Morcos, Alan
Chen, Sannie Wiryo, Tara Fitzgerald, Ted Hsiao, Chi Cheung, Samson Famurewa, Carlos
Hernandez, Duong Le, Dang-Khoa To, Linda Yu, Zachary O'Brien, Hyat Khan and Michael
Chiu.
21

References:
1. Hayward, D. S.; Kenley, R A.; lenke D. R. Int. 1. Phann. 1990, 59,245-253.
2. Hayward, D. S.; lenke, D. R. Int. 1. Pharn1. 1990,66,87-96.
3. Jenke, D. R; Kenley, R. A.; Hayward, D. S. l. Appl. Poly. Sci. 1991,43, 1475-1482.
4. Langer, R Science 1990,249,1527-1533.
5. Langer, R Acc. Chern. Res. 1993,26,537-542.
6. Benes M.; Dohnal, V. J. Chern. Eng. Data 1999,44, 1097-1102.
7. Brannon-Peppas, L. Medical Plastics and Biomaterials 1997, 34-45.
8. Langer, R.; Peppas, N. J. Macromol. Sci.-Rev. Macromol. Chern. Phys. 1983, C23, 61-126.
9. Govil, S. K. in Drug Delivery Devices, Tyle, P., Ed., Marcel Dekker: New York, 1988,
Chapter 13.
10. Maurin, B. M.; Dittert, L. W.; Hussian, A. A. 1. Phann. Sci. 1992,81,1, 79-84.
11. Fox, T. G. Bull. Am. Phys. Soc. 1956, 1, 123.
12. Nguyen, V. M.S. Thesis, University of California, Berkeley, California, 1999.
13. Bair, H. E.; Johnson, G. E.; Anderson, E. W.; Matsuoka, S. Polym. Eng. Sci. 1981, 21, 14,
930-935.
14. Leibler, L.; Sekimoto, K. Macromolecules 1993,26,6937-6939.
15. Crank, J. The mathematics of Diffusion, Clarendon Press: Oxford, 1975.
16. Bondi, A. Physical Properties of Molecular Crystals, Liquids, and Glasses, John Wiley &
Sons: New York, 1968.
17. Kipling, J. J. Adsorption from Solutions ofNonelectrolytes, Academic: New York, 1965.
18. Chen, S. X.; Lostritto, R. T. J. Control. ReI. 1996,38, 185-191.
22

19. Johnson, G. E.; Bair, H. E.; Matsuoka, S.; Anderson, E. W.; 1. E. Scott, r E. in Water in
Polymers, Rowland, S. P., Ed., ACS Symposium Series 127: Washington,1980, Chapter 27.
20. Kumar, S. A.; Thomas, S.; Kumaran, M. G. Polymer 1997, 38, 18,4629-4640.
21. Hamzaoui, F.; Baert, F.; Wojcik, G. Acta Crystallogr. B 1996,52, 1, 159-164.
22. Domalski, E. S.; Hearing, E. D. J. Phys. Chem. Ref. Data 1996,25, 1-526.
23. Domalski, E. S.; Hearing, E. D. 1. Phys. Chem. Ref.Data 1990, 19,881-1041.
24. NIOSH, "International Chemical Safety Card," ICSC: 0519, 1993.
25. Brandup, J.; Immergut, E. H.; Grulke, E. A. Polymer handbook, Jo1m Wiley & Sons: New
York, 1999.
23

Appendix I. Extrapolation of the Distribution Coefficient to Infinite Time.
For PVAc systems, non-negligible increases of the amount of absorbed solute were
measured even after 40 days. To obtain the equilibrium value of the distribution coefficient K; ,
extrapolation to infinite time is necessary. Toward that end, we have chosen a mathematical
model for diffusion that is consistent with our experimental conditions. The main features of this
model are:
1. The decrease of solute concentration in the aqueous phase is due only to diffusion of the
solute into the polymer bulk, i.e. the contribution of surface adsorption negligible.
2. The solute mass transfer into the polymer follows Fick's law with constant diffusion
coefficient.
3. The solution volume is limited, i.e. the concentration of the solute in the solution falls as the
solute enters the polymer film. This feature is essential because the solute uptake in the
polymer is experimentally deduced by monitoring the concentration in the solution as a
function oftime.
4. The solute mass transfer in the aqueous phase is much faster than that in the polymer phase;
the solute concentration in the aqueous phase is then uniform and depends only on
time(<Ds,w(t)) .
5. The polymer film is modeled as a plane sheet of infinite surface area with negligible border
area. The only relevant dimension x is orthogonal to the plan~ surface.
6. Equilibrium is established instantaneously at the solution-polymer interface x=±1
(<Ds,p(±l,t)=Ks .<Ds,w(t)) where 21 is the wet-film thickness.
24

Adsorption, discussed in the "Results 'and discussion" section, IS negligible. The
assumption of constant diffusion coefficient is justified at infinite-dilution condition where
The fifth feature is easily tested by calculating the ratio of the border area Ab to the
surface area As of the polymer films for a typical polymer thickness. AiAs is always less than 1%.
Finally, the time-independent relation between solute concentration in the polymer film at
the interface and that in the solution is reasonable for the highly dilute solutions used here.
The solute uptake in the polymer film as a function of time is obtained by solving the
one-dimensional diffusion problem:
(1.1 )
with the initial conditions:
<1>s,p =0
<1>S,IV =<1>~,1V
-I<x<l, (= 0,
t=O
(1.2)
stating that at time (=0 all the solute is in the aqueous solution phase with concentration <1>0 •S,w
The boundary condition equates the rate at which the solute leaves the solution to that at which it
enters the film at the surfaces x=:i: I:
a<1> a<1>
V ~= A . I =+A D-!:!!...IV at pJs,p x~±1 p ax
x~±1
(1.3a)
a a<1>s,p _ a<1>s,p
---=+D--
Ks at ax
x=±I, t >0.
One of the boundary condition in Equation (1.3) can be replaced by the symmetry
condition:
25

x= 0, t> O. (1.3b)
In Equations (1.1)-(1.3), Vw is the solution volume, Ap is the surface area of the polymer
film, <D represents the solute volume fraction in the polymer film, D is the diffusion coefficients,p " .,
in the polymer phase, x is the direction orthogonal to the polymer surface, t is time, 21 is the wet-
film thickness, <DS,IV is the solute volume fraction in the aqueous solution, 2a=V,./Ap is the
solution thickness, Ks is the distribution coefficient of the solute between polymer film and
aqueous solution, defined in Equation (5) of the text.
The solution of the diffusion equation is (Crank, 1975):
s,p(t) =1- f 2a(l+a) exp(- Dl~,t)
<Xl 0=1 l+a+a21~, " 12
(1.4)
In Equation (1.4), s,p(t) is the solute concentration in the polymer averaged over the film
" (" ) (<DO J. . " a a S I V "
thIckness at the tIme t; a = - - and <Xl = - . -'- .
I·K I l+as " "
Equation (1.4) contains two parameters: diffusion coefficient D and equilibrium
distribution coefficient Ks. We fit these parameters with a non-linear least-square algorithm
minimizing the weighted. sum S of the squares of the differences between calculated <1>calc,; and
experimental "tP.i concentrations: .
opt 2
S = L W;(calc,; - exp,; ) (1.5)
i=1
where npt is the number ofmeasurements ~t different times and Wi is the weight corresponding to
26

Because the contribution of the nth teml is decreasing exponen~ially with increasing index
71, only the first 10 tenns ofthe series in Equation (I.4) were used in the calculation.
The parameter fitting is repeated for each of the five vials used in the experiments for
detennining the equilibrium value ofKs and the diffusion coefficient D. The average distribution
coefficients and diffusion coefficients are reported respectively in Tables.6-7 of the text along
with their standard deviations.
27

Table 1: Characteristics of the solutes: name, chemical formula, light
sensitivity, purity and supplier
Solute Formula Light Purity (%) Supplier
I-Naphthol
CIOHsO
Yes 99.0 Aldrich
2-Naphthol Yes 99.0 Aldrich
2-Nitroaniline No 98.0 Aldrich
3-Nitroaniline C6H6N20 2 No 98.0 Aldrich
4-Nitroaniline No 99.0 Aldrich
2-Nitrophenol No 99.0 Acros
3-Nitrophenol C6H5N03 No 99.0 Aldrich
4-Nitrophenol No 98.0 Acros
2-Nitrotoluene No 99.0 . Acros
3-Nitrotoluene C7H7N02 No 99.0 Aldrich
4-Nitrotoluene No 99.0 Acros
Acetophenone CsHsO No 99.0 Pfaltz
Benzoic Acid C7H60 2 No 99.5 Aldrich
Benzonitrile C7H5N No 99.0 Aldrich
Benzophenone C13HIOO No 99.0 Fisher
Benzyl Alcohol C7HsO No 99.0 Fisher
Nicotine ClOH14N2 No 98.0 Aldrich
Nitrobenzene C6H5N02· No 99.9 Fisher
Pyridine C5H5N No 99.9 Fisher
28

Table 2: Solute properties: molecular weight M' (g/mol), density ps
(g/cm3), melting point T;-(0C)
Solute Mw ps
a
I-Naphthol 114 l.220
2-Naphthol 114 1.280
2-Nitroaniline 138 . 1.442
3-Nitroaniline 138 1.430
4-Nitroaniline 138 1.424
2-Nitrophenol 139 1.495
3-Nitrophenol 139 1.540b
4-Nitrophenol 139 1.479
2-Nitrotoluene 137 l.163
3-Nitrotoluene 137 l.157
4-Nitrotoluene 137 1.392
Acetophenone 120 1.055
Benzoic Acid 122 1.266
Benzonitrile 103 1.010
Benzophenone 182 1.110
Benzyl Alcohol lOS l.045
Nicotine 162 1.010
Nitrobenzene 123 1.196
Pyridine 79.1 0.978
a All solute densities are at 20 °C12
except for 3-nitrophenol (25°C)
b Harnzaoui et al. 21 .
C Benes and Dohnal6
d Domalski and Hearingll
e Domalski and Hearing23
f International Chemical Safety Card24
29
TJ
9S.0d
120.4d
69.3d
113.8d
147.5d
44:8d
96.8e
113.8d
_3.8e
16.0e
51.6d
19.6d
112.3d
-12.8d
47.Sd
-15.5d
-80.0f
5.7d
-41.6d

Table 3: Polymer properties: average molecular weight Mw (g/mol),
density Ps (g/cm composition (comonomer weight percent), glass-
transition temperatures Tg (OC) and purity
Polymer Ps %w Purity (%)
PVAc 500000 1.189 39.7 (±.5) >99
EVAc45 250000' '0.952 45% VAc >99
EVAc33 150000 0.936 33% VAc >99
a Estimated using Equation (l) and Tg (PVAc)=32 °e, Tg (PE)=-125 °e25
30

Table 4: Conditions for polymer film preparation
Conditions EVAc33 EVAc45 PVAc
THF solution
concentration" 6% 8% 4%
Container Aluminum pan Pyrex dish Aluminum pan
(diameter) (63 mm) (139 mm) (63 mm)
/Ill solution/pan 20 84 10
Heating T (<(:') 100 100 100
Heating time
40 40 40(hours)
Cooling T (<(:') room -20 room
Averagejilm
266 (±45) 421 (±94) 101 (±23)
thickness (Jl11l)
a
gr polymer/ml THF
31

Table 5: Water content at saturation (weight %).
Method
Thermogravimetric
Gravimetric
EVAc33
~O
~O
EVAc45
0.43 (±O.04)
0.42 (±0.05)
32
PVAc
7.0 (±0.3)
7.6 (±0.3)

Table 6: Equilibrium distribution coefficient 1( and corresponding
standard deviation for nineteen solutes in two EVAc copolymers and in
PVAc.
Solute EVAc33 EVAc45 PVAc
I-Naphthol 599 (±8) 623 (±2S) 2203 (±6S)
2-Naphthol 372 (±S) 743 (±8) 1693 (±27)
2-Nitroaniline 66.3 (±2.7) 140 (±2) 350 (±3)
3-Nitroaniline 36.5 (±1.3) 80.1 (±1.3) 264 (±4)
4-Nitroaniline 25.2 (±0.3) 65.9 (±o.S) 322 (±2)
2-Nitrophenol 93.7 (±4.3) 122 (±2) 185 (±S)
3-Nitrophenol 44.1 {±0.6) 94.2 (±1.1) 314 (±S)
4-Nitrophenol 33.9 (±0.2) 87.3 (±2.6) 374 (±7)
2-Nitrotoluene 336 (±17) 483 (±I) 435 (±IS)
3-Nitrotoluene 105 (±4) 249 (±2) 386 (±2)
4-Nitrotoluene 345 (±II) 471 (±IO) 423 (±3)
Acetophenone 38.7 (±1.2) 41.5 (±2.2) 55.7 (±0.7)
Benzoic Acid 5.18 (±O.4S) 13.9 (±1.9) 53.4 (±o.S)
Benzonitrile 43.9 (±0.7) 54.5 (±1.3) 77.1 (±3.1)
Benzophenone 1.356 (±30) 2231 (±12) 1782 (±37)
Benzyl Alcohol 2.79 (±0.08) 10.8 (±.7) 23.4 (±3.3)
Nicotine 1.42 (±0.08) 22 (±14) 46 (±9)
Nitrobenzene 91.5 (±0.8) 145 (±2) 172 (±7)
Pyridine 1.09 (±0.26) 1.38 (±0.41) 5.32 (±1.32)
33

Table 7: Diffusion coefficient D and corresponding standard deviation for
nineteen solutes in two EVAc copolymers (10-9
cm2
/s) and in PVAc (10-12
2 .
cm Is). For PVAc, results from both early-time measurements and from
the extrapolation procedure are compared.
Solute EVAc33 EVAc45
PVAc
early-time extrapolation
I-Naphthol 0.13 (±O.O5) 0.46 (±O.21) 1.08 (±O.50) 4.36 (±O.32)
2-Naphthol 0.44 (±O.l7) 0.66 (±O.30) 1.34 (±O.61) 6.60 (±O.77)
2-Nitroaniline 2.85 (±l.OO) 3.86 (±1.73) 3.38 (±1.54) 9.28 (±1.3I)
3-Nitroaniline 4.29 (±1.60) 5.11 (±2.31) 5.87 (±2.69) 11.3 (±l.O)
4-Nitroaniline 4.60 (±1.57) 6.21 (±2.77) 4.69 (±2.15) 12.3 (±l.l)
2-Nitrophenol 1.15 (±O.43) 10.2 (±4.7) 6.86 (±3.IS) 14.5 (±I.4)
3~Nitrophenol 3.00 (±l.O7) 2.51 (±1.l6) 2.39{±1.l1) 12.3 (±1.6)
4-Nitrophenol 4.62 (±1.5S) 1.55 (±O.69) 6.85 (±3.14) 13.2 (±1.9)
2-Nitrotoluene 2.21 (±O.SO) 1.53 (±O.69) 8.18 (±3.7S) 12.5 (±l.O)
3-Nitrotoluene 2.87 (±l.OO) 2.92 (±1.3I) 5.62 (±2.57) 15.6 (±I.4)
4-Nitrotoluene 1.25 (±O.46) 3.71 (±1.6S) 8.21 (±3.76) 15.9 (±2.4)
Acetophenone 4.61 (±1.65) 12.8 (±6.0) 10.8 (±5.0) 12.5 (±2.1)
Benzoic Acid 12.0 (±4.71) 7.2 (±3.8) 8.83 (±4.04) 6.63 (±1.3I)
Benzonitrile 4.98 (±1.74) 14.7 (±7.1) 19.9 (±9.2) 28.3 (±2.5)
Benzophenone 0.10 (±O.O3) 0.34 (±O.15) 1.36 (±O.63) 4.67 (±O.37)
Benzyl Alcohol 56.0 (±21.6) 15.9 (±S.S) .10.5 (±5.7) 145 (±27)
Nicotine .3.10 (±1.l4) .0.05 (±O.O7) 0.19 (±O.12) 1.96 (±1.70)
Nitrobenzene 2.63 (±O.97) 5.53 (±2.46) 18.7 (±8.7) 17.4 (±l.O)
Pyridine 51.3 (±30.1) 93.9 (±69.2)
34 .

Figure Captions
Figure 1. Water sorption in EVAc copolymers and PVAc. The initial decrease to the plateau
region corresponds to the water weight fraction.
Figure 2. Distribution coefficients for solutes in EVAc33. Acetophenone (after 20 days; circles),
2-nitroaniline (after 22 days; diamonds) and 2-nitrophenol (after 14 days; triangles). Ks is
independent ofthe poiymer's area/volume ratio indicating that adsorption is negligible.
Figure 3. Distribution coefficients for three solutes in EVAc33. Squares are for 3-nitroaniline
(after 23 days; Ks = 34.9 ± 1.3), diamonds are for acetophenone (after 35 days; Ks = 38.8
± 1.2), and circles are for 3-nitrophenol (after 35 days; Ks = 44.0 ± 0.7). The dotted line
represents the average distribution coefficient Ks.
Figure 4. Distribution coefficients for three solutes in EVAc45. Diamonds are for acetophenone
(after 14 days; Ks = 66.1 ± 0.5), squares are for 2-nitrophenol (after 14 days; Ks = 94.5 ±
1.1) and circles are for nitrobenzene (after 14 days; Ks = 145.6 ± 1.6). The dotted line
represents the average distribution'coefficient Ks.
Figure 5. Effect of time on the distribution coefficient of 2-naphthol in PVAc (circle: 7 days;
diamond: 21 days; triangle: 42 days). Ks is independent of solute concentration at any
time. The dotted lines represent the time-dependent average distribution coefficients.
Figure 6. Effect oftime on the distribution coefficient of4-nitroaniline in PVAc (circle: 14 days;
time. The dotted lines represent the time-dependent average distribution coefficients.
35

Figure 7. Effect oftime on the distribution coefficient of acetopheneone in PVAc (circle: 7 days;
time. The dotted lines represent the time-dependent average distribution coefficients.,
Figure 8. Short-time solute depletion in the aqueous solution. The vertical axis is the normalized
solute volume fraction in solution. Circles are for EVAc33 + 2-nitroaniline; triangles are
for EVAc33 + benzophenone; diamonds are for EVAc45 + 4-nitroaniline; squares are for
EVAc45 + nitrobenzene; stars are for PVAc + I-naphthol and crosses are for PVAc +
nitrobenzene. Open and filled symbols refer to two separate experiments for the same
polymer/solute system.
Figure 9. Double logarithmic plot of diffusion coefficients for nineteen solutes in EVAc33
versus polymer/water distribution coefficients.
36

101
~
- - - - - - - - - -,- - - - - - - - - - -,- - EVAc33
, ,
, ,
~ 1 AI ........... , ~ >':+:.v ;.....'
, ,
o'('r..,;·"",,~.,.~.... ~.,,._ ">""w'-'.'y.~'" I
,
----------~----------~
, ,
99 - - - - - - - - - - - - - - - - - - - _1- __________ 1___________I __________ _
. 'EVAc45'
97 - - - - - - - - 1- - - - - - - - - - -1- - - - - - - - - - -1- - - - - - --:- - - - -I - - - - - - - - - - ..,
-~
95 , "
--------------------------------------~----------~
................
= PVAc.....
93 -------~----------~
91
25 75 125 175 225 275
Temperature [ C ]
Figure 1.
37

120
100
A A
A
A
80
~ ~ ~ ~
~
60~
40 0 0 0 0
20
0
10 15 20 25 30 35 40 45
Area!Volume for Polymer Film (Item)
Figure 2.
38 .

50 ~----------------------------------------,
45 -. . ----I---I--y---!------~---
~ 40: _" _~) - - fT f -1-I-----
35 -- %if'-! I
30 - ...1.1-
25 ~~-~.~~~.~~~,~_~..~.~~~._~.~.,~~~~
0.01 0.03 0.05 0.07 0.09 0.11
4
10 cI>s,w
Figure 3.
39

--~ ~ -2-£- ---2 ---
--2 ---2-----g - - - -il - - - - - - - It - - -
- - - .. ~ - 0" - -0 - - - ~ - - - ~ - -
.. • . . • • _. 1-
0.050 0.060 0.070 0.080 0.090 0.100 0.110
4
10 <Ps w
. '
Figure 4.
40

1800 . - - - - - - - - - - - - - - - - - - - - - - ,
1600
I1--l-1----I---------------------------
--1--!---I---1- ---i- ----------------------
1400 -
1200
1000
-----------------2 ------'§.~- -------"Q- - -~-
800 ~~.~~~.~~~~~~~~.~.~.~~~~~
0.006 0.008 . 0.010 0.012 0.014 0.016
4
10 <!>s,w
Figure 5.
41

360 ...---------"""------------.....
-----I-1"---II---------r----.;.-----
~ 320 - - - - - - r------------If--------
:: -----I-----1------1----I-------------1
340
260 L....&...............&....IL....L...&....L.-'-'-'-'-'-....................--&....I....&....a....L...&....IL....&......................L....L...&....L...a....L......
0.070 0.080 0.090 0.100 - 0.110 0.120 0.130 0.140
4
10 <Ps,w
Figure 6.
42 !

70 ~--------------------------------------~
60 -
~ 50 :::::if::Ir::1f:~:::::Jr:::::::::
40
30 •••••••• ·1·...y.....!...J........J...
0.050 0.060 0.070 0.080 0.090 0.100 0.110 0.120
4
10 s,w
Figure 7.
43

1.0
~
- ~~A
- ~ ~~ Ai'
0.9
.~'~A :I:
.. ~ ., A t
~ i-
A ±rii
~ *'
..0
A ·t
a' of,
~
0.8 - !!!I
.. +~ .. :I:'rii
cP·a %
c
0.7 - o~ ~
ec ~
- 01;1 :K
o ' . . ~.
0.6 • . . .
0 5 10 15 20 25
t1!2 [min1!2]
Figure 8.
44

1.0E-06
~
1.0E-07
Q
. - 1.0E-08 QfIl
-M
Q~Q~e
QC.I
.......
1.0E-09 "Q
Q
1.0E-I0 ~ Q Q
r-
1.0E-ll ' , , "',. , ", , , " ",. "
1 10 100 1000 10000
Ks
Figure 9.
45

(!I;J;-I#b-nr @FJ"!.eroi@ (V;.4;J§I~I"::IC @J§j;Ja§l1..l'B ~I.I~ ~
®m ~ ~ B~j;J3"U3I?o ~·loJmA ~

UC Berkeley Research

More Related Content

Similar to UC Berkeley Research (20)

UC Berkeley Research