Miniemulsion Polymerization-A Comparative Study

of Preparative Variables

P. 1. TANG, E. D. SUDOL, C. A. SILEBI, and M. S. EL-AASSER*

Emulsion Polymers Institute and Department of Chemical Engineering,
Lehigh University, Bethlehem, Pennsylvania 18015

SYNOPSIS

Variations in the conditions used to prepare miniemulsions of monomers in water have

resulted in substantial differences in polymerization kinetics and final particle sizes and
distributions. These variations have included ( 1) surfactant amount and type; ( 2 ) cosur-
factant amount and type; ( 3 ) monomer( s ) type; ( 4 ) temperature of preparation and poly-
merization; ( 5 ) means and conditions of homogenization; and ( 6 ) degree of aging of the
emulsion. In an effort to unify some of the disparate information of previous work, a sys-
tematic study of some of the above variables was undertaken using styrene as the oil phase
with sodium lauryl sulfate as surfactant. Based on the polymerization kinetics and particle
sizes obtained, the following conclusions are drawn: The finest droplet size miniemulsions
are obtained by ( 1 ) using a cosurfactant; ( 2 ) homogenizing at elevated temperature; ( 3 )
homogenizing using a uniform high shear device (Microfluidizer) ; and ( 4 ) limiting the
aging time prior to polymerization.

INTRODUCTION are not stable and usually phase separate upon

standing for a short time. However, in miniemul-
Miniemulsions are relatively stable submicron ( 50- sions, the presence of a low molecular weight and
500 nm) dispersions of oil (such as monomer) in relatively water-insoluble compound such as hexa-
water prepared by shearing a system containing oil, decane or cetyl alcohol in the monomer droplets re-
water, surfactant, and, additionally, a “cosurfac- tards the diffusion of monomer out of the droplets.
tant.” The principle behind the making of stable Higuchi and Misra2 used diffusion theory to show
miniemulsions is the introduction of a low molecular that the diffusion of the more water-soluble com-
weight and relatively water-insoluble compound (the pound in the droplets is governed by the diffusion
“cosurfactant” ) inside the monomer droplets to re- of the water-insoluble one. Initially, when the
tard substantially the diffusion of monomer out of monomer is broken up into droplets, some monomer
the monomer droplets. It is well known that the would diffuse out of the smaller droplets, thus the
chemical potential of a phase is a function of the concentration of the water-insoluble compound
molal surface area of that phase, so a small crystal would be increased in the smaller droplets, until the
dissolves much faster than does a larger one and concentration of the water-insoluble compound is
small droplets would have a higher vapor pressure high enough to offset the chemical potential of the
than do the larger ones. In conventional emulsions, smaller droplets and significantly retard the diffu-
the smaller monomer droplets that have a higher sion process. Some miniemulsions can be prepared
chemical potential dissolve while the larger droplets that are stable upon standing for months.
grow a t their expense; thus, conventional emulsions For the cetyl alcohol system, the enhanced sta-
bility is also attributed to the formation of “inter-
* To whom correspondence should be addressed.
molecular complexes” at the oil/water interface.
Journal of Applied Polymer Science, Vol. 43, 1059-1066 (1991)
These complexes would be liquid condensed and
0 1991 John Wiley & Sons, Inc. CCC 0021-8995/91/06l059-08$04.00 electrically charged, creating a low interfacial ten-
1059
1060 TANG ET AL.

sion and high resistance to droplet ~oalescence.~-~ weeks before use. Potassium persulfate (FMC) was
In conventional emulsion polymerization, the recrystallized from distilled deionized water and then
principal locus of particle nucleation is the aqueous dried a t room temperature under vacuum. Sodium
phase or the monomer swollen micelles depending lauryl sulfate (BDH Chemicals), hexadecane (Ald-
on the degree of water solubility of the monomers rich), cetyl alcohol (Aldrich) , and sodium bicar-
and the amount of surfactant used; lower water sol- bonate (Fisher) were used as received. The water
ubility monomer and higher amounts of surfactant was distilled and deionized (DDI) .
would favor particle nucleation in monomer swollen
micelles.' Monomer droplets are considered only to
The Polymerization Process
act as monomer reservoirs supplying monomer to
the growing polymer particles. However, in the The recipe used in this study consists of 80 wt %
miniemulsion polymerization, the small size of the DDI water, 20 wt % styrene, with 2.66 m M sodium
monomer droplets enables them to become the bicarbonate, 2.66 m M potassium persulfate, 5 or 10
principal locus of particle nucleation. Thus, different m M sodium lauryl sulfate, and 15-40 m M (all based
conditions used in their preparation produce mini- on water) cosurfactant (cetyl alcohol or hexadec-
emulsions with different average droplet sizes and ane) . The reaction temperature was 70°C.
size distributions, which, in turn, exhibit different There are some differences in the procedures for
polymerization kinetics and final particle size dis- preparing styrene miniemulsions with hexadecane
tributions. versus cetyl alcohol (Fig. 1) . In the preparation of
Since the introduction of miniemulsion polymer- miniemulsions with hexadecane, sodium lauryl sul-
ization in the early 1970s,' many investigators have
studied the subject and have used many different
methods to prepare mini emulsion^.^-'^ The impor- PREPARATION OF MINIEMULSIONS WITH HEXADECANE
tant parameters in preparing miniemulsions are the
type and amount of cosurfactant (usually hexade-
cane or cetyl alcohol) and the means of carrying out
Fl Dissolve

the homogenization. Hansen and Ugelstad l1 used Homogenize

styrene as monomer, hexadecane as cosurfactant,
and a two-stage homogenizer (Manton Gaulin hexadecane
1 I
A
S. A.) as the means of homogenization. Chamberlain instyrene
et aLl4 used styrene, cetyl alcohol, and a sonifier.
Choi et aL15used styrene, cetyl alcohol, and the Mi- (Miniemulsions)
crofluidizer (Microfluidics Corp.) . Delgado et a1.16
used vinyl acetate and butyl acrylate as monomers,
hexadecane as cosurfactant, and both the Omni
mixer (Ivan Sorvall, Inc.) and a sonifier (Branson PREPARATION OF MINIEMULSIONS WITH CETYL ALCOHOL
Sonic Power Co.) ,and Rodriguez l7 used styrene and
methyl methacrylate as monomers, both cetyl al- & water at
cohol and hexadecane as cosurfactants, and a son-
ifier. This lack of consistency in the preparation of
miniemulsions necessitates the evaluation of the ef-
fect of the different methods of preparing mini- Cool to
emulsions, particularly the effect on the polymer-
ization kinetics and the final particle size distribu-
Stir for
tion. I0 minutes Homogenize
r
1 1
I

EXPERIMENTAL
Add
styrene
1
Materials
,j
(Miniemulsion

Styrene ( Polysciences ) was distilled under reduced

pressure (20 mmHg) of nitrogen. The purified Figure 1 Method of preparation of stable miniemul-
monomer was stored at -2°C for no more than 4 sions with hexadecane (top) and with cetyl alcohol.
 MINIEMULSION POLYMERIZATION 1061

fate is first dissolved in the water and the hexadecane

I
+
+ + +

in the styrene monomer. The two solutions are

mixed with a magnetic stirrer for 10 min followed
by homogenization to break the monomer into the 0.8

small droplets forming a miniemulsion. A few extra

steps are involved in the preparation of miniemul-
sions with cetyl alcohol. Sodium lauryl sulfate, cetyl 'cB 0.e-

alcohol, and water are mixed a t 65'C for 2 h, cooled 8

to room temperature, and sonified to break up the
gel phase.17Styrene is then mixed with the gel phase a3 0.4-

for 10 min with a magnetic stirrer followed by ho-

mogenization to form the miniemulsion. E
The Microfluidizer-110 ( Microfluidics Corp.) and 0.2-
Sonifier Disruptor W-350 ( Branson Sonic Power
Co.) were used to provide the high shear required to
produce submicron monomer droplets. In the Mi- 04*+. , . , . ~ . , . , . 1
crofluidizer, the inlet stream is pressurized and 0 30 80 80 120 150 I80
Reaction Time (Minutes)
forced into an interaction chamber. There it is di-
rected into precisely defined microchannels, causing Figure 3 Conversion versus time curve of run M10 ob-
the streams to instantly accelerate to high velocities tained by dilatometry.
and, when recombined in the interaction region, to
produce an emulsion with fine droplets and a narrow breaking up the oil droplets. In this work, the Mi-
size distribution as a result of shear, turbulence, and crofluidizer was operated at 80 psi inlet pressure and
cavitation forces." The sonifier produces rapid local about 7000 psi outlet pressure with the smallest or-
pressure variations below the probe tip, bringing ifice size available (A10) and 10 passes of the emul-
about cavitation and shear. The Microfluidizer is sion. The sonifier was operated at 50% duty cycle
expected to produce more uniform monomer drop- and power 7 for 60 s of sonification time.
lets than does the sonifier, because of its more uni- Dilatometry is applied to monitor the conversion-
form exposure of the entire fluid to the shear forces time behavior of the polymerizations. Its principle
is based on the difference in density between mono-
- mer and polymer. Polystyrene has a density of 1.037
g/cc a t 70°C, which is higher than the 0.860 g/cc
30C130030030000Q3 i of styrene at the same temperat~re.'~ As the poly-
merization proceeds, styrene is converted to poly-
styrene, and this, in turn, causes the contraction of
Constant the volume of the mixture inside the dilatometer.
temperature Therefore, the conversion at time t can be calculated
ivater bath from the initial amount of monomer and the volume
2 i n i n inner contraction at time t . Because the conversion de-
/-q
~

diameter pends on the volume change of the reaction mixture,

capillary with ~
the emulsion is degassed to avoid any gas bubbles
40 C l l l swle i from forming inside the dilatometer during the
polymerization. The bath was controlled at a tem-
perature of 70 f 0.02"C.
The apparatus used in this study is shown in
Figure 2. The dilatometer is immersed in a constant
temperature water bath that is controlled by a heat-
Stirring ing and circulating device. The mixture inside the
bar dilatometer is kept uniform by a stirring bar and a
[T$- Magnetic
stirrer
magnetic stirrer. After the emulsion is degassed at
a pressure of about 20 mmHg for 20 min using an
aspirator, it is loaded into the 25 cc flask, and the
Figure 2 Schematic representation of the dilatometer capillary is quickly inserted to avoid entrapment of
used for monitoring polymerization kinetics. air between the flask and the capillary. It takes about
1062 TANG ET AL.

5 min for the mixture to reach 7OoC. To start the

polymerization, an initiator solution, also at 70”C,
is injected into the dilatometer with a microsyringe
connected to a small polyethylene tube. Much data
can be obtained with frequent readings. Typical
conversion-time behavior is shown in Figure 3,
which contains about 100 data points.
Calibrated transmission electron microscopy
(TEM) is used to determine the final latex particle
size distribution. A monodisperse polystyrene
standard with diameter of 824 nm is added to a latex
sample before the sample is placed on a TEM grid.
For samples with a wide particle size distribution,
at least 1,000 particles are counted, and for narrow
particle size distributions, at least 500 particles are
counted.
0
Reaction Time (Minutes)

RESULTS AND DISCUSSION Figure 4 Conversion versus time curves for runs C2
( Microfluidizer, 10 mM SLS) , C8 (Microfluidizer, 5 mM
The conditions for all experiments are described in SLS), M6 ( Microfluidizer, 10 mM SLS/30 mM CA) , and
M10 (Microfluidizer, 5 mM SLS/ 15 mM CA) .
Table I. The variables in these experiments are the
concentration of sodium lauryl sulfate, the type and

amount of cosurfactant, and the device used for ho-

mogenization.
Table I Conditions of Polymerization of Delgado et a1.16found that for the same surfactant
Emulsions Prepared by Different Methods
concentration the overall rate of polymerization of
Run Description a conventional emulsion system is usually faster
than that of a miniemulsion system. Most of these
c2 “Conventional,” Microfluidizer, 10 mM comparisons were carried out with a total concen-
sodium lauryl sulfate (SLS) tration of surfactant above its critical micelle con-
centration (cmc). The monomer swollen micelles
C6 “Conventional,” Microfluidizer, 5 mM
are an order of magnitude smaller than the mini-
SLS, 2 h unstirred at 25°C before
initiation emulsion droplets, which means that the number of
monomer swollen micelles is much larger than is
c7 “Conventional,” Microfluidizer 5 mM SLS, the number of monomer droplets. Thus, more par-
1 h with stirring a t 70°C before ticles are nucleated in the conventional system than
initiation in the miniemulsion system. However, when the
C8 “Conventional,” Microfluidizer, 5 mM SLS concentration of surfactant is below the cmc, the
rate of polymerization is greater for the miniemul-
M6 Microfluidizer, 10 mM SLS, 30 mM cetyl
alcohol (CA), monomer added to SLS- sion system because there are no monomer swollen
CA solution at 65°C micelles present in the conventional emulsion sys-
tem. This phenomenon can be seen from the results
M10 Microfluidizer, 5 mM SLS, 15 mM CA, given in Figure 4 in which the variables studied were
monomer added to SLS-CA solution at the concentrations of sodium lauryl sulfate and cetyl
65°C
alcohol. Curves C2 and M6 were obtained using rec-
M11 Microfluidizer, 5 mM SLS, 15 mM CA, ipes containing 10 m M sodium lauryl sulfate,
monomer added to SLS-CA solution at whereas curves C8 and M10 contained 5 m M (the
25°C cmc of this surfactant was determined to be 7.4
M12 Microfluidizer, 5 mM SLS, 20 mM m M ) . The recipe used to obtain the solid curves
hexadecane (C2 and C8) did not contain any cosurfactant,
whereas those represented by the dashed curves
M13 Sonifier, 5 mM SLS, 20 mM hexadecane
contained cetyl alcohol. All these emulsions were
 MINIEMULSION POLYMERIZATION 1063

subjected to high shear with the Microfluidizer prior

to polymerization. The results in Figure 4 show that
for emulsions prepared with 10 m M sodium lauryl
sulfate the polymerization kinetics of the “conven-
tional” case ( C 2 ) is faster than that of the mini-
emulsion case (M6). However, for those prepared
with 5 m M sodium lauryl sulfate, the polymerization
kinetics of the “conventional” case ( C 8 ) is slower
than that of the miniemulsion case (M10). These
results also show that the polymerization rates of
miniemulsions (M6 and M10) prepared with varying
surfactant and cosurfactant levels do not differ as
much as those (C2 and C8) prepared without co-
surfactant. The final particle numbers (sizes) are
consistent with the kinetic results (Table 11) ; the
faster polymerizations produced greater numbers of
particles. O !
0 0.2 0.4 0.1) 0.8
From the conversion-time data of Figure 4, the Fractional Conversion
polymerization rate versus conversion is obtained Figure 5 Polymerization rate versus conversion curves
as shown in Figure 5. These results indicate that for runs C2 (Microfluidizer, 10 mM SLS) , C8 (Micro-
the polymerization rates of all the runs exhibit sim- fluidizer, 5 mM SLS) , M6 (Microfluidizer, 10 mM SLS/
ilar behavior with respect to conversion; the poly- 30 mM CA) , and M10 (Microfluidizer, 5 m M SLS/ 15
merization rates increase to a maximum at about mM CA).
30% conversion followed by a decrease. No signifi-
cant constant rate period is observed. This phenom-
enon has been reported p r e v i ~ u s l y ’and
~ would in- brings about a decrease in the rate of termination
dicate that when the nucleation is predominantly in inside the particle, causes the rate to increase. The
the monomer droplets it continues until all droplets polymerization rate continues to increase to about
disappear (by either nucleation or consumption by 85% when the particles reach their glass transition
growing particles) ; thus, no constant rate period is point, bringing about a reduction in the propaga-
seen. This postulate is supported by results found tion rate and thus reducing the polymerization rate.
in a study of semicontinuous miniemulsion poly- Another set of experiments was carried out to
merizations by Tang et al.,” which showed that evaluate the effect of emulsion stability on the poly-
particle nucleation takes place whenever mini- merization kinetics. The results are shown in Figure
emulsion droplets are present in the reaction system. 6. The emulsions used in these experiments were
The polymerization rate continues to decrease until prepared using the Microfluidizer but without cetyl
about 60% conversion when the gel effect, which alcohol, and they were polymerized after aging under
varying conditions. These emulsions were relatively
unstable without the cosurfactant, the droplets be-
Table I1 Volume Average Diameter, Coefficient coming larger with time due to their instability
of Variation, and Final Particle Concentration against diffusion degradation. The emulsion in run
Obtained by Polymerization of Emulsions
C8 was initiated just after degassing, whereas C7
Prepared by Different Methods
was stirred a t 70°C for 1h and C6 was left unstirred
for 2 h at room temperature before initiation. The
results from Figure 6 show that the polymerization
c2 115 0.13 26.8 rate decreased with increased aging of the emulsion.
C6 169 0.06 8.75 These results are also consistent with the final par-
c7 168 0.05 9.34 ticle numbers (Table I1 ).
C8 160 0.08 10.8 Different methods of preparing miniemulsions
M6 134 0.12 18.4 with the same recipe can also result in different ki-
M10 147 0.11 13.4 netics and final particle concentrations. This can be
M11 168 0.14 9.36 seen from the results in Figure 7 in which mini-
M12 145 0.06 15.7
emulsions were prepared with cetyl alcohol (see Fig.
M13 271 0.18 2.5
1) and the only difference is that M10 was prepared
1064 TANG ET AL.

40
I

Reaction Time (Minutes) 0

Diameter (nm)
Figure 6 Conversion versus time curves for runs C8
Figure 8 Particle size distributions for runs M10 (Mi-
( Microfluidizer,5 m M SLS and no aging time), C7 (Mi-
crofluidizer,5 mM SLS/15 mM CA, styrene added to the
crofluidizer, 5 m M SLS and one hour at 70°C) and C6
gel phase at 65°C) and M11 (Microfluidizer,5 mM SLS/
( Microfluidizer,5 mM SLS and two hours at 25°C prior
15 mM CA, styrene added to the gel phase at 25°C).
to polymerization).

without cooling the gel phase. The styrene was added rene followed by homogenization. The kinetics of
to the gel solution (sodium lauryl sulfate, cetyl al- run M10 is significantly faster than that of run M11.
cohol, and water) at 65°C before being homogenized This indicates that the emulsion droplets of run M10
with the Microfluidizer, whereas in run M11, the gel are smaller than those of run M11. This is probably
solution was cooled to room temperature and soni- because the viscosity of the emulsion prepared at
fied to break up the gel phase before adding the sty- higher temperature (M10) is lower than that at room
temperature (M11), allowing more effective ho-
mogenization. The final particle numbers are con-

I
0.8
M10:
__--
,_-- ,A11
sistent with the kinetic results (Table 11). The cor-
responding particle size distributions (Fig. 8) exhibit
similar shapes and show a significant number of
small particles. This indicates that some particle
nucleation occurs continuously throughout the
polymerizations.
Another study was carried out to evaluate the ef-
fect of the type of homogenization on the polymer-
ization kinetics and final particle size distribution.
The results are shown in Figure 9 in which the only
difference is that the emulsion in run M12 was ho-
mogenized using the Microfluidizer,and in run M13,
by the sonifier ;both contained 20 m M hexadecane.
The polymerization rate in run M12 was much faster
than that of run M13, which agrees with the final
,;:'
,;.' particle numbers (Table 11). Figure 10 shows that
0 ' . , . , . , . , . , . , . the particle size distribution of run M12 is much
0 30 80 00 120 IS0 180 210
Reaction Time (Minutes) narrower than that of run M13 (which has a con-
Figure 7 Conversion versus time curves for runs M10 siderable number of large particles), which is con-
( Microfluidizer,5 mM SLS/ 15 mM CA, styrene added to sistent with the claim that the Microfluidizer pro-
the gel phase at 65°C) and M11 (Microfluidizer, 5 mM vides a greater and more uniform shear than does
SLS/15 mMCA, styrene added to the gel phase at 25"C), the sonifier.
 MINIEMULSION POLYMERIZATION 1065

Figure 9 Conversion versus time curve for runs M12, Figure 11 Conversion versus time curves for runs C6,
homogenized with the Microfluidizer and M13, homoge- C7, and C8, “conventional” cases; M10 and M11, mini-
nized with the sonifier. emulsions with cetyl alcohol; and M12, miniemulsion with
hexadecane.

A comparison of the various methods of preparing slower than those prepared with hexadecane (M12)
miniemulsions can be made from the combined re- and those without cosurfactant (C6, C7, and C 8 ) .
sults illustrated in Figure 11.At conversions higher However, the rate in run M10 increases faster and
than about 30%, all runs exhibit similar kinetics, surpasses those of the “conventional” cases. This
but a t lower conversions, some differences can be may indicate that particle nucleation is slowed by
noted. This region is magnified in Figure 12. The the presence of cetyl alcohol on the surface of the
initial polymerization rates of the miniemulsions monomer droplets. This is the subject of further in-
prepared with cetyl alcohol (M10 and M11) are vestigations.

0.004 . , . I . I . I ’ I

0 10 20 30 4u 50
Reaction Time (Minutes)
Diameter (nm)
Figure 12 Conversion versus time curves magnified for
Figure 10 Particle size distributions for runs M12, ho- runs C6, C7, and C8, “conventional” cases; M10 and M11,
mogenized with the Microfluidizer and M13, homogenized miniemulsions with cetyl alcohol; and M12, miniemulsion
with the sonifier. with hexadecane.
1066 TANG ET AL.

SUMMARY 3. Y. T. Choi, Ph.D. Dissertation, Lehigh University,

1986.
4. E. G. Cockbain and T. S. McRoberts, J. Colloid In-
A comparative study was carried out to evaluate the
terface Sci., 8 , 440 ( 1953).
effect of the variation in parameters important to
5. D. C. Blackley and A. C. S. Lawrence, Discuss. Faraday
the preparation of styrene emulsions on the poly- SOC.,2 8 , 2 6 8 (1954).
merization kinetics and the resulting particle size 6. P. Pithayanukul and N. Pipel, J. Colloid Interface Sci.,
distributions. The variables studied were the type 8 9 , 4 9 4 ( 1982).
of cosurfactant and the means of homogenization. 7. J. W. Vanderhoff, J. Polym. Sci. Polym. Symp., 7 2 ,
The results show that when the concentration of 161 (1985).
sodium lauryl sulfate is above its cmc, the overall 8. J. Ugelstad, M. S. El-Aasser, and J. W. Vanderhoff,
rate of the “conventional” polymerization is faster Polym. Lett., 11, 503 (1973).
than that of the miniemulsion system. However, it 9. J. Ugelstad, F. K. Hansen, and S. Lange, Makromol.
is the opposite when the concentration of sodium Chem., 175,507 (1974).
lauryl sulfate is below its cmc. When the emulsion 10. F. K. Hansen, E. Baumann Ofstad, and J. Ugelstad,
is allowed to age, thereby increasing the monomer in Theory and Practice of Emulsion Polymerization,
A. L. Smith, Ed., Academic Press, London, 1976, p.
droplet size, the polymerization was found to de-
13.
crease with aging time (or with increasing droplet 11. F. K. Hansen and J. Ugelstad, J. Polym. Sci. Polym.
size). Different temperatures used in preparing cetyl Chem. Ed., 17,3069 (1979).
alcohol miniemulsions with the same recipe can also 12. J. Ugelstad, P. C. Mork, K. Herder Kaggerud, T. El-
result in different kinetics; the miniemulsion pre- lingsen, and A. Berge, Adv. Colloids Interface Sci., 1 3 ,
pared with styrene added to the gel solution at 65°C 101 (1980).
had a faster reaction rate than did the one prepared 13. J. Ugelstad, T. Ellingsen, and K. Herder Kaggerud,
with styrene added to the gel solution at 25°C. The Advances in Organic Coating Science & Technology,
polymerization rates agree with the final particle Vol. 2, Technomic, Westport, CT, 1980, p. 1.
size; the faster rates correspond to the smaller final 14. B. J. Chamberlain, D. H. Napper, and R. G. Gilbert,
particle size. The results indicate that, under the J. Chem. SOC.Faraday Trans. I, 78,591 (1982).
15. Y. T. Choi, M. S. El-Aasser, E. D. Sudol, and J. W.
reported conditions, the Microfluidizer provides a
Vanderhoff, J. Polym. Sci. Polym. Chem. Ed., 23,2973
greater and more uniform shear than does the son-
( 1985).
ifier, thus resulting in smaller and more uniform 16. J. Delgado, M. S. El-Aasser, and J. W. Vanderhoff,
final particle size distributions. The results also in- J. Polym. Sci. Polym. C h m . Ed., 24, 861 (1986).
dicate that particle nucleation is slowed by the pres- 17. V. S. Rodriguez, Ph.D. Dissertation, Lehigh Univer-
ence of cetyl alcohol on the surface of the monomer sity, 1988.
droplets. 18. Technical Bulletin on Microfluidizer, Microfluidics
Corp., Newton, MA, 1989.
We would like to thank the National Science Foundation 19. E. D. Sudol, Ph.D. Dissertation, Lehigh University,
for financial support under Grant CTS-8907600. 1983.
20. P. L. Tang, E. D. Sudol, M. E. Adams, J. M. Asua,
and M. S. El-Aasser, J. Appl. Polym. Sci., to appear.
REFERENCES 21. M. Ballard, R. G. Gilbert, and D. H. Napper, Mac-
romolecules, 17 ( 3 ), 504 ( 1984 ).
1. C. N. Lewis and M. Randall, Thermodynamics and
the Free Energy of Chemical Substances, McGraw-Hill,
New York, 1923, p. 252.
2. W. I. Higuchi and J. Misra, J. Pharm. Sci., 5 1 , 459 Received September 27, I990
(1962). Accepted December 20, 1990