Multiple-phase behavior and its microscopic implication for 4-acrylamidosalicylic acid gel

Article abstract of DOI:10.1063/1.1357201

The microscopic structure of PASA gel was studied by means of small-angle neutron scattering (SANS) to elucidate the nature of the multiple phases. The gel exhibited not only multiple phases characterized by different degrees of swelling but also a marked

Full text of DOI:10.1063/1.1357201

Multiple-phase behavior and its microscopic implication for 4-
acrylamidosalicylic acid gel
Masahiko Annaka, Mitsuhiro Shibayama, Fumiyoshi Ikkai, Masaaki Sugiyama, Kazuhiro Hara et al.
Citation: J. Chem. Phys. 114, 6906 (2001); doi: 10.1063/1.1357201
View online: http://dx.doi.org/10.1063/1.1357201
View Table of Contents: http://jcp.aip.org/resource/1/JCPSA6/v114/i15
Published by the AIP Publishing LLC.
Additional information on J. Chem. Phys.
Journal Homepage: http://jcp.aip.org/
Journal Information: http://jcp.aip.org/about/about_the_journal
Top downloads: http://jcp.aip.org/features/most_downloaded
Information for Authors: http://jcp.aip.org/authors
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

JOURNAL OF CHEMICAL PHYSICS
VOLUME 114, NUMBER 15
15 APRIL 2001
Multiple-phase behavior and its microscopic implication
for 4-acrylamidosalicylic acid gel
Masahiko Annakaa)
Department of Materials Technology, Chiba University, Inage-ku, Chiba 263-8522, Japan
Mitsuhiro Shibayama
Neutron Scattering Laboratory, The Institute for Solid State Physics, The University of Tokyo,
Tokai, Naka-gun, Ibaraki 319-1106, Japan
Fumiyoshi Ikkai
´
L’OREAL RECHERCHE, Tsukuba, Ibaraki 300-2635, Japan
Masaaki Sugiyama
Division of Chemistry and Physics of Condensed Matter, Faculty of Science, Kyushu University,
Higashi-ku, Fukuoka 812-8581, Japan
Kazuhiro Hara
Institute of Environmental System, Faculty of Engineering, Kyushu University, Higashi-ku,
Fukuoka 812-8581, Japan
Takayuki Nakahira
Department of Materials Technology, Chiba University, Inage-ku, Chiba 263-8522, Japan
Toyoichi Tanakab)
Department of Physics and Center for Materials Sciences and Engineering,
Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
͑Received 22 June 2000; accepted 30 January 2001͒
Poly͑4-acrylamidosalicylic acid͒ gels exhibited multiple-phase behavior depending on their histories
in the parameter space of pH and temperature. Four different phases, denoted as phase091
͑as-prepared͒, phase244 ͑swollen at high pH͒, phase064 ͑heat-treated͒, and phase233 ͑swollen at
high pH after heat treatment͒ were clearly resolved, where the three digits denote their linear
swelling ratios in percentage with respect to their sizes at preparation. Each phase was stable and did
not change its swelling ratio with pH or temperature as long as the values of pH or temperature were
within limited ranges. Transitions among different phases were discrete with hysteresis loops. The
structure factors corresponding to these four phases were obtained by small-angle neutron
scattering, which indicated the presence of characteristic structures depending on pH and
temperature, particularly in the shrunken state ͑i.e., phase064͒. © 2001 American Institute of
Physics. ͓DOI: 10.1063/1.1357201͔
I. INTRODUCTION
interact with each other through randomly distributed repul-
sive and attractive interactions. The latter should be hydro-
gen bonding plus either hydrophobic or electrostatic interac-
tions. Among these interactions, it has been demonstrated
experimentally and theoretically that hydrogen bonding
plays an essential role in inducing a multiple-phase behavior.
In this study, we will show that such a condition, so far
demonstrated only in copolymer gels,4–6 can also be
achieved in homopolymers as in the case of poly͑4-
acrylamidosalicylic acid͒ gels ͑PASA gels͒ examined in this
work.
The multiple-phase behavior in polymer gels has so far
been studied by simple swelling ratio measurements as a
function of pH, temperature, solvent quality, and so on. Al-
though these studies demonstrated essential roles of the fun-
damental interactions for triggering volume phase transitions
in gels, they have not disclosed any microscopic view of the
structure of these kinds of gels. In this paper, we will inves-
tigate the microscopic structure of PASA gel by means of
Polymer gels are known to have two phases: swollen and
collapsed phases. Volume phase transition occurs between
the two phases either continuously or discontinuously.1–3 Re-
cently, more than two phases were found in copolymer gels
consisting of copolymers with randomly distributed posi-
tively and negatively charged groups. These phases were
characterized by different degrees of swelling at a given pH:
the gel can take different degrees of swelling depending on
the route taken in the pH–temperature coordinate system.
The number of multiple phases appearing in the coordinate
system depends on the monomer composition and pH.4–6
By studying gels that exhibit such a multiple-phase be-
havior, the criterion for a gel to show the multiple-phase
behavior is considered. It is deduced that polymer molecules
a͒
Author to whom correspondence should be addressed. Electronic mail:
annaka@galaxy.tc.chiba-u.ac.jp
b͒
Deceased.
0021-9606/2001/114(15)/6906/7/$18.00
6906
© 2001 American Institute of Physics
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Multiple-phase behavior
6907
B. Determination of phase diagram
The cylindrical gel was placed in a glass cell whose
temperature was controlled within Ϯ0.1 °C. The gel was
continuously flushed with water of controlled pH from a res-
ervoir. pH was controlled by adding aqueous HCl solution
͑to lower pH͒ or water ͑to increase pH͒ below pH 7. Aque-
ous NaOH solution was used above pH 7. Gel diameter at
swelling equilibrium, d, was measured under a microscope.
To avoid the effect of carbon dioxide in the air, all experi-
ments were carried out under nitrogen gas atmosphere. The
temperature of the gel was controlled to within Ϯ0.1 °C by
circulating water from a Lauda RM-6B, a temperature-
controlled water circulation system.
C. Small-angle neutron scattering
Small-angle neutron scattering ͑SANS͒ experiments
were carried out using the research reactor, SANS-U, of the
Institute for Solid State Physics, the University of Tokyo,
which is located at the Japan Atomic Energy Research Insti-
tute, Tokai, Japan. Cold neutrons from the reactor, mono-
chromatized with a velocity selector to a flux of neutrons
with a wavelength of ␭ϭ7 Å and a wavelength distribution
of ⌬␭/␭ϭ0.1, were used as the incident beam. Sample gels
were introduced into a quartz cell having an optical length of
4 mm. The observed scattering intensity was corrected for
cell and solvent scattering, incoherent scattering, and trans-
mission, and then rescaled to the absolute intensity. Incoher-
ent scattering from a Lupolen standard sample was used for
intensity calibration. SANS experiments were also per-
formed with a 12-m SANS camera installed at Oak Ridge
National Laboratory ͑ORNL͒. The incident neutron wave-
length was collimated to be ␭ϭ4.75 Å with a graphite crys-
tal. An area detector was located 5.8 m from the sample. The
detected intensity was corrected for transmission, sample
thickness, measuring time, and background, and was then
circularly averaged.
FIG. 1. Chemical structure of poly͑4-acrylamidosalicylic acid͒ ͑PASA͒ gel.
The molecules can interact with each other through ͑a͒ hydrogen bonding;
͑b͒ ionic repulsive interaction; and/or ͑c͒ hydrophobic stacking between
benzene rings. Temperature influences the hydrogen bonding and hydropho-
bic interaction, whereas pH affects the degree of ionization and thus the
electrostatic interaction. The monomer undergoes ionization via two steps:
one for the carboxyl group and the other for the hydroxyl group.
small-angle neutron scattering ͑SANS͒ in order to elucidate
the nature of the multiple phases.
II. EXPERIMENT
A. Sample preparation
4-acrylamidosalicylic acid ͑Fig. 1͒ was synthesized from
acryloyl chloride and sodium 4-aminosalicylate.7 Sodium
4-aminosalicylate ͑40 g, 0.24 mol͒ was dissolved in distilled
water ͑250 ml͒ and stirred for 1 h. Acryloyl chloride ͑total
0.37 mol͒ was added twice, once at the beginning ͑20 ml,
0.25 mol͒, and then after the solution was stirred for 1 h ͑10
ml, 0.12 mol͒. Acidification to pH 4.0 with 10 N hydrochlo-
ric acid solution yielded a gray precipitate which was filtered
off and washed with distilled water ͑500 ml͒ to give
4-acrylamidosalicylic acid ͑27 g, yield 57%͒.
III. RESULTS AND DISCUSSION
A. Phase diagram
PASA gels were prepared by radical polymerization:
When pH or temperature was varied, PASA gel changed
its volume discontinuously. As shown in Fig. 2͑a͒, the as-
prepared gel took phase091 (d/d0ϭ0.91), where d/d0 is the
linear swelling ratio expressed as the gel diameter, d, nor-
malized by the original diameter d0. At low pH, the gel was
in a shrunken state ͑phase091͒. Here, the three digits follow-
ing the term ‘‘phase’’ denote the linear swelling ratio in per-
centage with respect to the size at preparation. The gel in
phase091 swelled discontinuously at pH 7.6 to phase244. As
the pH was lowered from 7.6, it shrank back to phase091 at
pH 5.8. This process had good reproducibility. The hyster-
esis indicates that the transition is of the first order.
After the loop, the temperature of the gel in phase091
immersed in deionized water was raised from 25 to 60 °C.
The gel shrank to phase064 (d/d0ϭ0.64͒ at 50 °C, as shown
in Fig. 2͑b͒. This phase remained unchanged even if the
temperature was lowered back to 25 °C. This temperature
scan was followed by another experiment at 25 °C by vary-
1.04
g of 4-acrylamidosalicylic acid, 0.0773 g of
N,NЈ-methylenebisacrylamide ͑cross linker͒, and 8.0 mg of
azobisisobutyronitrile ͑initiator͒ were dissolved in 5.0 ml of
dimethyl sulfoxide. The solution was polymerized at 70 °C
for 8 h. Two types of gels were prepared, one for swelling
measurement and the other for SANS. In the case of the
former, a cylindrical gel was prepared in a capillary with an
inner diameter 140 ␮m ͑ϭd0). After completion of gelation,
the cylindrical gel was removed from the capillary mold and
was washed with distilled water. On the other hand, gels for
SANS were prepared in petri dishes and then passed through
a sieve. The sieved gel was immersed in a large amount of
dimethyl sulfoxide to remove unreacted chemicals, and then
in water. Water in the gel was replaced by heavy water by
keeping its swelling equilibrium. This was done by repeat-
edly exchanging the water surrounding the gel with heavy
water.
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

6908
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Annaka et al.
both hydrogen bond donor and acceptor in the monomer unit
and undergoes two-step ionization. As a result, due to a com-
bination of hydrogen bonding ͓Fig. 1͑a͔͒, electrostatic inter-
action ͓Fig. 1͑b͔͒, and hydrophobic interaction ͓Fig. 1͑c͔͒,
PASA gels can have several different phases depending on
the route of pH and/or temperature conditioning. Tempera-
ture influences the hydrophobic interaction, whereas changes
in the degree of ionization the gel by pH variation affect the
electrostatic interaction. The ionization process of the mono-
mer units consists of two steps: one for the carboxyl group
and the other for the hydroxyl group at higher pH. Hydrogen
bonding, on the other hand, is governed by both temperature
and pH. Hence, the unique phase behavior of PASA gel is
deduced to result from its unique chemical structure.
Figure 2 also shows that the swelling properties of
PASA gel depend on its past history: the phase behavior of
the gel depends on whether the gel had experienced the most
swollen phase ͑phase244͒ or the most collapsed phase
͑phase064͒ in the past. When a pH loop is started from
phase091, the gel remains in a hysteresis loop, as shown in
Fig. 2͑a͒. On the other hand, once the gel experiences
phase064, the phase transition loop is changed to a different
loop, as shown in Fig. 2͑c͒. The most collapsed state ͑i.e.,
phase064͒ is formed by hydrophobic interaction, followed by
interlocking with hydrogen bonding ͓see Fig. 1͑a͔͒. The
memory can be easily erased by high pH conditioning ͓see
Fig. 2͑c͔͒. Therefore, information about its history, i.e.,
whether the gel has experienced the most swollen or the
most shrunken state, can be stored as the degree of swelling.
These results strongly suggest that the most collapsed
phase is interlocked with another kind of physical interaction
that is created when the gel is in phase064. Recent studies
have shown that the incorporation of a hydrophobic moiety
into polyelectrolytes leads to a decrease in acidity or basic-
ity. This decrease in acidity or basicity is related to the di-
electric constant surrounding ionizable groups.8–10 In the
case of PASA gel, since benzene rings are less polar than
water, the dielectric constant surrounding the charges of car-
boxyl groups and hydroxyl groups is considered to be low-
ered with an increase in temperature. This will lower the
value of pKa and is expected to lead to the formation of
hydrogen bonds. Therefore, hydrogen bonding may be a rea-
sonable interpretation of the observed interaction.
The swelling behavior of the most collapsed gel
͑phase064͒ was examined as a function of urea concentration
at 25 °C to confirm the effect of hydrogen bonds. As shown
in Fig. 3͑a͒, the gel swelled discontinuously at the urea con-
centration of 0.75 mol/L to phase233. As the urea concentra-
tion was reduced from 6 mol/L to pure water ͑0 mol/L͒, the
gel shrank gradually and finally went back to its original
shrunken phase ͑phase091͒. After that, the phase transition
loop remained in the loop shown in Fig. 2͑a͒ ͓Fig. 3͑b͔͒.
Urea is a protein denaturant and is considered to break pro-
tein hydrogen bonds. From these facts, it is most likely that
the multiple-phase behavior and the memory effect will
appear when hydrogen-bonding interlocking effects are
significant.
FIG. 2. Diameter, d, of the PASA gel normalized by the mold size, d0, as a
function of pH and temperature. Measurements were carried out on a single
gel under a microscope. The time course of the variation of d was carefully
monitored whenever pH or temperature was changed in order to ensure that
equilibrium was reached at each measurement. The sequence of each set of
measurements is given in alphabetical order ͑a͒ to ͑d͒. The starting point of
each diagram is the last point of the previous diagram. The detailed descrip-
tion of the paths is given in the text. Open circles in the phase diagram
indicate the phases of the samples used for SANS experiment.
ing pH, the results of which are shown in Fig. 2͑c͒. The gel
underwent another phase transition from phase064 to a new
phase, i.e., phase233 ͑d/d0ϭ2.33͒. This cycle also had good
reproducibility.
When pH was further increased to 12 and then de-
creased, the gel went back to its original swollen phase
͑phase244͒, as shown by solid squares in Fig. 2͑c͒. This ex-
periment was followed by another pH-loop experiment for 4
ϽpHϽ8, the results of which are shown in Fig. 2͑d͒. It is
shown that the transitions at pH 7.5 and pH 5.8 are the same
as that observed in the first loop in Fig. 2͑a͒. Therefore, this
series of d/d0 measurements discloses the following features
of the gel: ͑1͒ As the starting point, the gel is in phase091.
͑2͒ It swells to phase244 at higher pH ͑i.e., 8 ϽpHϽ10͒. ͑3͒
A temperature-induced transition from phase091 to phase064
takes place at 50 °C. ͑4͒ However, the memory of phase064
can be erased by immersing the gel in a reservoir of pH 12.
͑5͒ In all of these transitions, there exists a clear hysteresis
loop.
Although PASA gel is a homopolymer gel consisting of
a single kind of monomer, it exhibits a multiple-phase be-
havior with four discrete phases. This multiple-phase behav-
ior is a result of its unique chemical structure. PASA has
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Multiple-phase behavior
6909
FIG. 3. Diameter of PASA gel as a function of concentration of aqueous
urea solution. Open squares indicate the process of increasing concentration
of urea. Solid squares indicate the process of decreasing concentration of
urea ͑a͒. Equilibrium swelling degree d/d0 of the PASA gel in water as a
function of pH at 25 °C ͑b͒. The starting point of the diagram is the last
point of ͑a͒. Temperature was fixed at 25 °C and was controlled to within
Ϯ0.1 °C by circulating water from LAUDA RC-3 during the measurement.
Open circles in the phase diagram indicate the phases of the samples used
for SANS experiment.
FIG. 4. Swelling curves of the gel for various values of h, the number of
hydrogen bonds on an active chain, calculated using Eq. ͑2͒, 2 ␯ /N
v
0
A
ϭ0.25, ␾0ϭ0.1, ⌬FHB /kTϭ10, and fNϭ10.
͑1͒ Hydrogen bonding gives rise to additional physical
cross-linking points in the chains comprising the gel.
The formation of these cross-linking points divides a
chain into short, connected chains of equal length. The
number of elastically active chains per unit volume at ␾
0
B. Theoretical consideration
changes to (hϩ1)␯ due to the formation of h cross-
0
linking points.
According to the mean-field theory, the swelling behav-
ior of a gel is described by the following equation:2,3,11
͑2͒ Physical cross-linking points are randomly distributed in
the network, and no cooperative interactions are consid-
ered.
͑3͒ Solvent molecules and monomers composing the gel
system have the same molar volume.
Ϫ5/3
Ϫ1
⌬F 2 ␯
␾
1
2
␾
v
0
2
0
␶ϵ1Ϫ
ϭ
Ϫ fϪ
ϩ1
ͫ
ͩ ͪ
ͩ
ͪͩ ͪ
ͬ
kT
␾
␾
NA␾
0
0
2
2 ln 1Ϫ␾͒
͑
1/3
ϩ
ϩ
,
͑1͒
⌬F 2 ␯
␾
v
2
0
2
0
1/2
Ϫ3/2
␾
␾
␶ϵ1Ϫ
ϭ
hϩ1͒
Ϫ hϩ1͒
͑
͑
ͫ
ͩ ͪ
kT
␾
NA␾
0
where NA is Avogadro’s number, k the Boltzmann constant,
Ϫ1
T the temperature, the molar volume of the solvent, ␾ the
v
1Ϫh/f ͒
hϩ1
1
␾
͑
ϫ
fϩ
ͩ ͪ
ͬ
ͭ
ͮ
volume fraction of the polymer network, ␾ the volume frac-
0
2
␾
0
tion of the polymer network at the reference state, ⌬F the
excess free energy for the association between polymer seg-
2 ␯ h ⌬F
v
2
2 ln 1Ϫ␾͒
͑
0
ϩ
HB ϩ1ϩ
ϩ
,
͑2͒
2
ment and solvent, ␯ the number of elastically active chains
NA
2
kT
␾
␾
0
per unit volume at ␾0, and f the number of dissociated coun-
terions per effective chain. The equation of state of a gel
uniquely determines the equilibrium swelling degree of the
gel, V/V0ϭ␾0 /␾ at a given value of the reduced tempera-
ture. Three values of ␾, which correspond to two minima
and one maximum value of the free energy, satisfy Eq. ͑1͒
for a certain value of ␶. Equation ͑1͒, however, does not
predict the experimental results shown in Figs. 2 and 3 be-
cause these equations apply to only non-hydrogen-bonded
networks. These results suggest that a new, attractive, and
short-range interaction arises in the polymer network of
PASA gel after the gel experiences the most collapsed phase.
In the case of PASA gel, carboxylic acid groups can lead to
the formation of both ionic and hydrogen bonds. When a
hydrogen bond is formed, it influences the equation of state:
it adds a new cross linking between polymers and effectively
shortens the polymer chain length, and the energy is lowered
due to hydrogen bonding. Therefore, we obtain the equation
of state of a gel by considering hydrogen bonds that form a
new cross linking under the following three assumptions:12,13
where ⌬FHB is the free-energy change of the system due to
the formation of hydrogen bonds in an active chain.
It is worthy to note that Eq. ͑2͒ coincides with the equa-
tion of state for a non-hydrogen-bonded network ͑1͒ when h
ϭ 0. In the case of hydrogen-bonded networks, Eq. ͑2͒ de-
termines the equilibrium volume fraction of the polymer net-
work, ␾, as a function of the reduced temperature and the
number of hydrogen bonds h. Thus, a set of swelling curves
corresponding to different numbers of hydrogen bonds is ob-
tained. The calculated results are shown in Fig. 4. Phase
behaviors of the PASA gel shown in Fig. 2͑b͒ are considered
to correspond to the hydrogen-bonded network. The discrete
volume phase transition temperature of the gel decreases as
the number of hydrogen bonds in the gel increases, and then
it disappears at a critical value of h, i.e., hϭhc . The volume
change at the phase transition temperature becomes smaller
as the value of h is increased. The swelling curves of the gel
become monotonic above hc . The swelling ratio of the gel in
the swollen state becomes smaller as the number of hydrogen
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

6910
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Annaka et al.
TABLE I. Sample characteristics.
phase064–phase233–phase064–phase244. Phase091urea is
the revisited phase of phase091 after experiencing the urea
aqueous solution. The magnitudes of I(q) for phase091,
phase064, and phase091r and phase091urea are much larger
than those for phase244 and phase233. This is due to the
difference in the polymer concentration in the gel by a factor
of approximately 20 ͑between phase244 and phase091, for
example͒. I(q) for the gel in phase091 seems to be a mono-
tonically decreasing function of q. I(q) of phase064, ob-
tained by heating the gel in phase091, has a scattering maxi-
mum around qϭ0.02 ÅϪ1. This scattering maximum
indicates the presence of charged groups in a poor solvent
moiety whose repeating distance is about 400 Å.
Code
␾
d/d0
␾0 /␾
Phase091
Phase244
Phase064
Phase233
Phase0.91r
Phase091urea
Dry gel
0.294
0.0152
0.832
0.0170
0.294
0.294
1.00
0.91
2.44
0.64
2.33
0.91
0.91
•••
0.754
14.5
0.262
13.0
0.754
0.754
•••
bonds increases in the gel, while that of the collapsed state is
insensitive to the number of hydrogen bonds. Numerical cal-
culation reveals that the critical number of hydrogen bonds
hc is 2.8 under the present conditions. It indicates that the
formation of only three hydrogen bonds out of ten acrylic
acid units in an intact active chain causes a drastic change in
the swelling behavior of the gel. These results imply that the
hydrogen bond in the polymer network affects the phase be-
havior of the gel.
As predicted by Borue and Erukhimovich,15 and by
Joanny and Leibler,16 weakly charged polymer solutions
͑and polymer gels͒ in a poor solvent undergo microphase
separation in order to compromise for a huge penalty of lo-
calization of charges caused by collapse in a poor solvent by
forming microclusters ͑or microphase separation͒. Therefore,
this peak is a strong indication that PASA gel is
charged.17–20
In order to analyze the structure factor of PASA gels
more quantitatively, we employed a theory proposed by
Borue and Erukhimovich ͑the BE theory͒, which deals with
the structure factor of weakly charged polymers in a poor
solvent. The scattering intensity is given in the form
C. SANS
Open circles in Figs. 2 and 3 indicate the phases of the
samples used for the SANS experiment. SANS samples are
coded as phase091, phase244, etc., according to the results of
swelling experiments. Table I lists the characteristics of
SANS samples, such as the volume fraction of the polymer,
␾, d/d0, and the volume swelling ratio, V/V0, where V/V0
was simply obtained by (V/V0)ϵ(d/d0)3. The volume frac-
tion ␾ was assumed to be equal to the weight fraction of the
polymer because the mass density of heavy water is 1.1 and
is close to that of the solute.14
x2ϩs
I q͒ϳ
͑
,
͑3͒
x2ϩt͒ x2ϩs͒ϩ1
͑
͑
where x and s are the reduced scattering vector and the re-
duced salt concentration, respectively, as defined by
xϭr0q, sϵ␬2r20 .
͑4͒
Figure 5 shows the SANS intensity functions, I(q), for
PASA gels in phases 091, 244, 064, 233, phase091r, and
phase091urea, where q is the scattering vector. Numbers in
parentheses indicate the order of sample treatment in the
pH–temperature coordinate system. Phase091r is the revis-
ited phase of phase091 after passing through the loop of
The parameter r0 is the characteristic length of screening
in a salt-free polyelectrolyte solution. ␬ is the Debye screen-
ing length. The variable t is the reduced temperature related
to Flory’s interaction parameter ␹ and is defined by
2
r0
tХ12
1Ϫ2␹͒,
͑5͒
͑
ͩ ͪ
a
where a is the segment length of the chain molecules. Solid
lines in Fig. 5 are the results of curve fitting with Eq. ͑3͒,
where the parameters, x, s, and t were floated. The theoret-
ical curves reasonably reproduce the observed scattering in-
tensity functions. Since the essential parameters are s and x,
an s-t phase diagram is constructed in Fig. 6. This phase
diagram shown is divided into eight regimes ͑denoted by
Roman numerals͒ depending on the values of s and t. The
borders of these regimes are given by simple relations, such
as tϭsϩ2 and stϭ1, as denoted in the figure. The classifi-
cation of these regions and their characteristics are described
in the original paper15 and elsewhere;18–20 thus, we focus on
the case of PASA gels studied here. The gels in phase091,
phase091r, and phase091urea are located very close to the
spinodal line with tϽ0 and sϾ1 (sϭ1.062 and tϭϪ0.936
for phase091, sϭ1.078 and tϭϪ0.921 for phase091r, and
sϭ1.097 and tϭϪ0.908 for phase091urea͒. However, be-
cause of tϾϪ1/s, phase091, phase091r, and phase091urea
are in regime V, where Coulomb interactions are practically
FIG. 5. SANS intensity profiles of PASA gels obtained at 25 °C. Solid lines
are fitted curves using Eq. ͑3͒.
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Multiple-phase behavior
6911
FIG. 6. Phase diagram showing the states of weakly charged polymer gels.
t and s are the reduced temperature and the reduced salt concentration,
respectively. The thick solid curve indicates the spinodal line. Dashed lines
and the spinodal line divide different regimes coded by I to VIII. The states
of PASA gels are indicated by open circles in the diagram.
FIG. 7. Comparison of scattering intensity functions among phase091 ͑open
circles͒, bulk PASA gel (␾ϭ1; open diamonds͒, and the gel with ␾
ϭ0.294 prepared by swelling from dried PASA gel ͑solid circles͒.
The dashed curve is obtained by multiplying ␾ϭ0.294 with I(q) of the
dried gel.
neglected. This is why I(q) values for phase091, phase091r,
and phase091urea are monotonically decreasing functions, as
is often observed for neutral polymer solutions in the semi-
dilute regime. On the other hand, phase064 is classified un-
der regime VI since tϾϪ1/s (sϭ0.987 and tϭϪ1.012),
where Coulomb interactions are practically neglected but mi-
croscopic fluctuations are amplified. Note that at stϭϪ1,
I(qϭ0) diverges ͑macrophase transition͒ for sϾ1 and I(q
Þ0) diverges ͑microphase transition͒ for sϽ1. Hence, I(q)
for phase064 starts to exhibit a scattering maximum near its
spinodal line. According to the above discussion, it can be
deduced that phase064 and phase091 are very close to mi-
crophase separation transition and macrophase transition, re-
spectively. Phase244 and phase233, on the other hand, are
located in regime III, where the excluded volume effects
dominate over Coulomb interactions. In this regime, a gel
behaves like a neutral polymer solution in the semidilute
regime. It should be noted here that s becomes indeterminant
for phase244 and phase233 and the evaluated values of s do
not have any strong physical meaning. The essential feature
to be noted here is that the states of PASA gel differ quite
significantly in the s-t phase diagram due not only to pH or
temperature but also to the history of sample treatment.
Now, we discuss why PASA gels can be found in many
regimes in the s-t diagram even though pH and T are fixed.
One possibility is that the gel in phase064 is strongly inter-
locked via hydrogen bonds. This possibility is deduced from
the fact that phase064 is stable with temperature change ͓see
Fig. 2͑b͔͒ and the chain molecules are packed compactly
(␾ϭ0.89͒. In order to break hydrogen bonds, an aqueous
medium with high pH ͑Ͼ10͒ is required, which leads to
strong ionization of the gel and breaking of hydrogen bonds.
This is why the gel undergoes swelling transition from
phase064 to phase233. Note that ͑1͒ phase233 is different
from phase244 ͓see Fig. 2͑c͔͒ and ͑2͒ phase233 does not go
to phase091 but to phase064 upon lowering pH. Unfortu-
nately, any difference in the structure factor between
phase244 and phase233 could not be resolved by SANS be-
cause the scattering intensity functions for these phases were
very weak and highly scattered. Phase233 went back to
phase244 by increasing pH to 12. The resultant phase244
could finally go back to phase091r, whose scattering inten-
sity is shown in Fig. 5. Although I(q) values for phase091r
and phase091urea are somewhat larger than that for
phase091, they are roughly the same. Therefore, it can be
concluded that not only macroscopically but also micro-
scopically, PASA gels return to their original state by trav-
eling along the pH-temperature loop applied here.
Figure 7 shows a comparison of I(q) values among
phase091 ͑open circles; ␾ϭ0.294͒, dried PASA gel (␾ϭ1;
open diamonds͒, and the gel with ␾ϭ0.294 prepared by
swelling from dried bulk PASA gel ͑solid circles͒. Note that
bulk PASA gel is porous and I(q) originates from void scat-
tering. This comparison allows us to judge the difference in
structure between as-prepared gel in phase091 ͑open circles͒
and the gel having a history being completely dried ͑filled
circles͒. As is clearly shown here, I(q) for the gel prepared
from the dry state ͑solid circles͒ is quite different from that
for the gel in phase091 ͑open circles͒. The dashed curve
obtained by multiplying ␾ϭ0.294 with I(q) of the dried gel
͑open diamonds͒ nicely superimposes on the results for the
gel prepared by swelling from the dry state ͑filled circles͒.
Therefore, the microscopic structure of the gel prepared by
swelling from the dry state is similar to the dried gel itself.
This indicates that the structure of the gel changes irrevers-
ibly by complete drying, probably due to the formation of
strong hydrogen bonds.
This transformation reminds us of the denaturation of
protein by heat and or acid/base treatment.21 It is well-known
that most proteins can fold spontaneously into their desired
shapes. Proteins can be unfolded, i.e., denatured, by treat-
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions

6912
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Annaka et al.
ment with certain solvents, to form flexible polypeptide
chains. When the denaturing solvent is removed, the proteins
usually refold into their original conformations. In the case
of PASA gel, however, the ‘‘denatured’’ gel can revert back
to its original conformation ͑or size͒ by treatment with a high
pH aqueous solution. Hence, PASA gel can be used as a
model system for studying the processes of denaturation and
renaturation of proteins. It should be noted that Takata et al.
recently observed the temperature-induced denaturation pro-
cess of a globular protein.22 A scattering maximum in the
SANS intensity profile for ␤-lactoglobulin aqueous solution
was detected, which is very similar to that observed in PASA
gel. This observation supports the hypothesis on the analogy
between the denaturation of proteins and the multiple-phase
behavior of polymer gels.
Sports and Culture ͑No. 11740382 for M.A., No. 10875199
for M.S., and No. 11650286 for KH͒. This work was par-
tially supported by the US–Japan Cooperative Research Pro-
gram on Neutron Scattering. The authors wish to express
their gratitude to Dr. T. Norisuye and Dr. G. Wignall for
their help in the SANS experiment at ISSP-NSL and ORNL.
M.A. and M.S. are, respectively, grateful to the Yazaki Me-
morial Foundation for Science Technology and to the Cos-
metology Research Foundation, Tokyo, for financial support.
1 T. Tanaka, Phys. Rev. Lett. 40, 820 ͑1978͒.
2 T. Tanaka, D. J. Fillmore, S.-T. Sun, I. Nisho, G. Swislow, and A. Shah,
Phys. Rev. Lett. 45, 1636 ͑1980͒.
3 M. Shibayama and T. Tanaka, Adv. Polym. Sci. 109, 1 ͑1993͒.
4 M. Annaka and T. Tanaka, Nature ͑London͒ 355, 430 ͑1992͒.
5 M. Annaka and T. Tanaka, Phase Transit. 47, 143 ͑1994͒.
6 T. Tanaka and M. Annaka, J. Intell. Mater. Syst. Struct. 4, 548 ͑1993͒.
7 J. F. Kennedy, S. A. Baker, J. Epton, and G. R. Kennedy, J. Chem. Soc.,
Perkin Trans. I 1, 488 ͑1973͒.
IV. CONCLUSION
PASA gel exhibits multiple phases that are characterized
by different degrees of swelling; the gel can take one of four
different values of the degree of swelling, but none of the
intermediate values. The gel exhibited a marked memory ca-
pability: its phase behavior depends on whether it has expe-
rienced the most swollen phase or the most collapsed phase
in the immediate past. The information is stored during a pH-
T cycle, and can be easily erased by high- pH or urea treat-
ment.
The structure factors obtained by small-angle neutron
scattering also indicate the presence of characteristic struc-
tures corresponding to the four phases, particularly the
shrunken state ͑i.e., phase064͒. The structure factors also in-
dicate that PASA gel reverts to its original state not only
macroscopically but also microscopically via the pH–
temperature loop or urea treatment applied here. It is also
found that the microscopic structure of the gel prepared by
swelling from the dry state is similar to that of the dried gel
itself.
8 R. A. Siegel and A. Firestone, Macromolecules 21, 3254 ͑1988͒.
9
´
A. P. Sassi, S. Beltran, H. H. Hooper, H. W. Blanch, J. Prausnitz, and R.
A. Siegel, J. Chem. Phys. 97, 8767 ͑1992͒.
10 H. Kawasaki, S. Sasaki, and H. Maeda, J. Phys. Chem. 101, 5089 ͑1997͒.
11 P. J. Flory, Principles of Polymer Chemistry ͑Cornell University Press,
Ithaca, 1953͒.
12 M. Annaka, M. Tokita, T. Tanaka, S. Tanaka, and T. Nakahira, J. Chem.
Phys. 112, 471 ͑2000͒.
13 K. Otake, H. Inomata, M. Konno, and S. Saito, J. Chem. Phys. 91, 1345
͑1989͒.
14 M. Shibayama, T. Tanaka, and C. C. Han, J. Chem. Phys. 97, 6829
͑1992͒.
15 V. Borue and I. Erukhimovich, Macromolecules 21, 3240 ͑1988͒.
16 J. F. Joanny and L. Leibler, J. Phys. ͑France͒ 51, 545 ͑1991͒.
17 F. Schosseler, A. Moussaid, J. P. Munch, and S. J. Candau, Macromol-
ecules 24, 225 ͑1991͒.
18 M. Shibayama, T. Tanaka, and C. C. Han, J. Chem. Phys. 97, 6842
͑1992͒.
19 M. Shibayama, T. Tanaka, and C. C. Han, J. Phys. ͑Paris͒, Colloq. C8, 25
͑1993͒.
20 M. Shibayama, Multiphases in Polymeric Gels, edited by T. Araki, Q.
Tran-Cong, and M. Shibayama ͑Dekker, New York, 1998͒, Chap. 6, p.
195.
21 T. E. Creighton, Protein, Structures and Molecular Principles ͑Freeman,
New York, 1984͒.
ACKNOWLEDGMENTS
22 S. Takata, T. Norisuye, N. Tanaka, M. Shibayama, Macromolecules 33,
5470 ͑2000͒.
This work has been supported by Grants-in-Aid for Sci-
entific Research from the Ministry of Education, Science,
Downloaded 12 Oct 2013 to 128.143.23.241. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions