The effect of the Hofmeister series on the deswelling isotherms of poly(N-isopropylacrylamide) and poly(N,N-diethylacrylamide)

Article abstract of DOI:10.1039/c1sm05222e

The influence of Na₂SO₄, NaCl, NaBr, and NaI was studied on the deswelling isotherms of thin films of photo-crosslinked poly(N-isopropylacrylamide) and poly(N,N-diethylacrylamide) to help elucidate the mechanisms by which ions of the Hofmeister series affect the solubility of neutral, amide-based polymers. The films were characterized with both ellipsometry and FTIR to determine both water content and the frequency of the CO, N-H, and CH₃ vibrations associated with the polymers. When compared at the same water content, the frequency of the N-H bend in poly(NIPAAm) red-shifts in the order I^- > Cl^- > Br^- > SO₄^2-. The red-shift is consistent with disrupted hydrogen bonding of the N-H moiety due to an ion-dipole attraction. The CO stretch on the other hand is insensitive to the ion, suggesting that the ion pairs primarily with the partially positive end of the HN-CO dipole. For poly(DEAAm), which lacks an N-H moiety, the CO stretch is now sensitive to the ion, with the trend following the order of the ion in the Hofmeister series. Both of these findings signify that I^- > Cl^- > Br^- > SO₄^2- with respect to the strength of the ion-dipole interaction. Moreover, the ion-dipole interaction is stronger in poly(NIPAAm) than poly(DEAAm), suggesting that the specificity of the Hofmeister ions arises from the effect of vicinal groups on the amide dipole. The Royal Society of Chemistry 2011.

Full text of DOI:10.1039/c1sm05222e

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Soft Matter
Cite this: Soft Matter, 2011, 7, 6061
www.rsc.org/softmatter
PAPER
The effect of the Hofmeister series on the deswelling isotherms of
poly(N-isopropylacrylamide) and poly(N,N-diethylacrylamide)
*
Leena Patra,† Ajay Vidyasagar† and Ryan Toomey
Received 10th February 2011, Accepted 21st April 2011
DOI: 10.1039/c1sm05222e
The influence of Na2SO4, NaCl, NaBr, and NaI was studied on the deswelling isotherms of thin films of
photo-crosslinked poly(N-isopropylacrylamide) and poly(N,N-diethylacrylamide) to help elucidate the
mechanisms by which ions of the Hofmeister series affect the solubility of neutral, amide-based
polymers. The films were characterized with both ellipsometry and FTIR to determine both water
content and the frequency of the C]O, N–H, and CH3 vibrations associated with the polymers. When
compared at the same water content, the frequency of the N–H bend in poly(NIPAAm) red-shifts in the
order Iꢀ > Clꢀ > Brꢀ > SO42ꢀ. The red-shift is consistent with disrupted hydrogen bonding of the N–H
moiety due to an ion–dipole attraction. The C]O stretch on the other hand is insensitive to the ion,
suggesting that the ion pairs primarily with the partially positive end of the HN–C]O dipole. For poly
(DEAAm), which lacks an N–H moiety, the C]O stretch is now sensitive to the ion, with the trend
following the order of the ion in the Hofmeister series. Both of these findings signify that Iꢀ > Clꢀ > Brꢀ
> SO42ꢀ with respect to the strength of the ion–dipole interaction. Moreover, the ion–dipole interaction
is stronger in poly(NIPAAm) than poly(DEAAm), suggesting that the specificity of the Hofmeister ions
arises from the effect of vicinal groups on the amide dipole.
Hofmeister salts result from ion-specific interactions with both
I. Introduction
amide and nonpolar moieties,14 where the amide interaction is
perhaps a nonspecific ion–dipole interaction that is modulated by
vicinal methyl groups.18 Cremer and co-workers proposed that
strong ‘‘salting out’’ salts act by polarizing the water molecules
adjacent to the amide groups and by increasing the surface
tension around the hydrophobic groups, while ‘‘salting in’’ salts
Lower critical solution temperature (LCST) polymers can serve
as model systems for probing the effect of ions on the stability of
biological macromolecules.1–4 It is well established that the
dissolution temperature of LCST polymers, including poly(N-
isopropylacrylamide), or poly(NIPAAm), is strongly affected by
the addition of salt, in a manner similar to how protein stability is
affected by salt.5–13 The effect of ions, however, is not purely
concentration dependent. The dissolution temperature depends
on the ‘‘structure forming’’ or ‘‘structure breaking ability’’ of the
ion and follows the Hofmeister series.14–17
bind directly to the amide groups.7 They also suggested a two-
2ꢀ
step demixing process for strong salting out anions like SO4
,
wherein amide groups are first dehydrated followed by the
dehydration of the isopropyl groups.7
Near the water/air interface, all ions of the Hofmeister series
increase surface tension due to exclusion of the ions near the
surface.19,20 Similar effects have been observed with the salting
out of benzene from aqueous solution, where ions decrease the
solubility of benzene in water.21 It may therefore be expected that
all macromolecules have decreased solubility in salt solutions.
Indeed, early members of the Hofmeister series, or kosmotropes,
do monotonically decrease solubility. Later members of the
Hofmeister series, or chaotropes, however, can form ion–dipole
pairs with hydrogen bonding moieties and increase solubility.22–24
In this paper, we explore the influence of Na2SO4, NaCl, NaBr,
and NaI on the swelling of thin layers of surface-attached
networks of poly(NIPAAm) and poly(N,N-diethylacrylamide),
or poly(DEAAm), both of which are amide-based polymers. Poly
(NIPAAm) contains a secondary amide that donates and accepts
a hydrogen bond; whereas, poly(DEAAm) contains a tertiary
LCST polymers contain a hydrogen-bonding group, such as
an amide, and a hydrophobic group. At low temperatures,
solubility of the macromolecule is engendered through favorable
mixing of the hydrogen-bonding groups and water. As temper-
ature is increased, a point is reached where mixing can no longer
stabilize hydrophobic interactions and the polymer phase-sepa-
rates from solution.
Despite a considerable body of research, however, questions
still remain concerning the exact mechanism of the interaction of
salts with macromolecules that contain both amides and
hydrophobic groups. It is generally thought that the effect of the
Department of Chemical and Biomedical Engineering, University of South
Florida, Tampa, FL, 33620, USA. E-mail: toomey@usf.edu
† Both authors contributed equally.
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 6061–6067 | 6061

View Article Online
amide that only accepts a hydrogen bond. The Hofmeister series
suggests that SO42ꢀ > Clꢀ > Brꢀ > Iꢀ in terms of the anion’s ability
to salt-out neutral, LCST polymers. Within this ordering, SO42ꢀ is
kosmotropic, Iꢀ is chaotropic, and Clꢀ and Brꢀ are intermediate.
We use both ATR-FTIR and ellipsometry to characterize the
surface-attached networks. While FTIR has been extensively used
to measure the frequency shifts of the vibrations associated with
the amide and aliphatic groups in LCST polymers and gels upon
demixing,25–29 interpretation of spectra is difficult without precise
knowledge of the water content. To that end, we use ellipsometry
to derive the average water distribution in the polymer coatings.
FTIR spectra are then compared at the same water content in
order to unambiguously determine the effect of ions on the
vibrations of the amide and aliphatic groups.
Preparation of surface-tethered network for ellipsometry
LaSFN9 prisms were cleaned with ozone to remove any organic
impurities followed by deposition in a 1% solution of 3-amino-
propyl triethoxysilane in acetone. The substrates were heated to
100 ꢁC to drive condensation of the silane groups to the substrate
surface. A solution of the appropriate polymer in cyclohexanone
was spin cast on the freshly prepared substrate. Crosslinking was
accomplished by exposing the film to UV light (365 nm) for 30
minutes. Ultraviolet radiation (l ¼ 365 nm) triggers the n,p*
transition in the benzophenone moieties leading to a biradicaloid
triplet state that abstracts a hydrogen from a neighboring
aliphatic C–H group, forming a stable C–C bond. The coatings
were subsequently solvated in water and/or salt solutions for 30
minutes, enough time to attain homogeneity of the coating
throughout its thickness.
II. Experimental section
Ellipsometry
Materials
The dry and swollen profiles of the surface-attached films were
measured with a home-built variable-angle nulling ellipsometer
in an ATR configuration. The light-source was a HeꢀNe laser
with a wavelength of l ¼ 633 nm. The experimental setup has
been described in detail elsewhere.31 The refractive index of the
LaSFN9 prism is n ¼ 1.845, and in contact with water, the critical
angle of total reflection is 46.7ꢁ. The thickness of all films was in
the optical range, giving rise to sufficient features in the recorded
spectra to infer the refractive index profile. To analyze the data,
model refractive index profiles were generated, and the ellipso-
metric parameters were numerically calculated using the matrix
optical formulation.32 The parameters of the model were
adjusted to minimize the differences between the simulation and
experimental data.
N-Isopropylacrylamide (NIPAAm), N,N-diethylacrylamide
(DEAAm), 4-hydroxybenzophenone, methacryloyl chloride,
triethylamine, acetone, D2O (99.9 atom%), Na2SO4, NaCl,
NaBr, NaI, azobisisobutyronitrile (AIBN), hexanes, diethyl
ether and 3-aminopropyl triethoxysilane were purchased from
Sigma. Acetone was distilled from calcium hydride before use
and NIPAAm was recrystallized from hexanes. All other chem-
icals were used as received.
Monomer synthesis
Methacryloyloxybenzophenone (MaBP) was synthesized from 4-
hydroxybenzophenone and methacryloyl chloride in dry acetone
at 0 ꢁC as described previously.30 The monomer yield was around
90% MaBP and characterized with an INOVA 400 NMR spec-
trometer. The spectrum had typical aromatic peaks (multiplet) at
7.2–8.0(m) ppm and a singlet at 2.1(s) ppm representing the
methyl group characteristic of MaBP.
ATR-FTIR
ATR-FTIR measurements of poly(NIPAAm-co-MaBP) or poly
(DEAAm-co-MaBP) were performed using a Nicolet 6700 FTIR
spectrometer equipped with a temperature controlled multi-
bounce ATR ZnSe plate. The polymers dissolved in acetone were
cast onto the ZnSe plate and heated to 60 ꢁC for 2 hours to remove
residual acetone. An FTIR spectrum was then performed to
determine that there was no residual acetone left in the polymer
cast on the ZnSe plate. All FTIR measurements were performed at
24 ꢁC in absorbance mode with 100 scans and a resolution of
4 cmꢀ1. The spectral ranges for all scans were in the region between
4000 and 650 cmꢀ1. The crystal background was subtracted from
the samples for each run.
Copolymer synthesis
Poly(NIPAAm-co-MaBP) or poly(DEAAm-co-MaBP) was
copolymerized with 3 mol% MaBP using 0.1% AIBN as the
initiator. Several samples were prepared. All reactions were
carried out for 18 hours at 65 ꢁC in dioxane under nitrogen. The
samples were degassed with nitrogen by freeze and thaw cycles
prior to the reaction. After completion, the polymer was precip-
itated in diethyl ether. To verify incorporation of MaBP, 1H NMR
in CDCl3 was conducted. The percentage of MaBP was calculated
from the integration of the CH peaks of the aromatic group (7.2–
8.0(m) ppm) and the integration CH peaks of the isopropyl group
of NIPAAm or the diethyl group of DEAAm in the NMR spec-
trum with an error estimate of 5%. All samples showed complete
incorporation of the MaBP. GPC on all samples were performed
on a Viscotekꢀ using RI, VIS, RALS, and LALS. Molecular
weights were determined by the triple detection method. The
number-average molecular weight of poly(NIPAAm-co-MaBP) is
42 830 g molꢀ1 with a polydispersity of 4.7. The number-average
molecular weight of poly(DEAAm-co-MaBP) is 13 342 g molꢀ1
with a polydispersity of 2.39.
III. Results
Fig. 1 shows the thermally induced deswelling transition (or
deswelling isobars) in both poly(NIPAAm) and poly(DEAAm)
networks with 3 mole% MaBP cross-linker. Both have a demix-
ꢁ
ing temperature near 30 C. Below the deswelling temperature,
poly(NIPAAm) has a higher equilibrium water content than poly
(DEAAm), which is likely the result that poly(NIPAAm) can
hydrogen bond to water at both the C]O and N–H positions
whereas poly(DEAAm) can hydrogen bond at the C]O posi-
tion. Interestingly, the two ethyl groups of poly(DEAAm) do not
6062 | Soft Matter, 2011, 7, 6061–6067
This journal is ª The Royal Society of Chemistry 2011

View Article Online
they display similar collapse profiles with each salt. If a salt
concentration C1/2 is defined for when the polymer layer reaches
1
a dimensionless thickness value of /2 , the values of C1/2 for
Na2SO4, NaCl and NaBr, are 0.25 M, 1.3 M, and 2.5 M,
reꢁspectively, for both poly(DEAAm) and poly(NIPAAm). At
5 C, a C1/2 for NaI, could not be found up to 5.0 M. NaI also
leads to slight additional swelling of both layers (by ꢂ5 to 10%)
at low salt concentrations.
The air/water surface tension increments of the individual salts
in dynes/(cm ꢃ molality) are Na2SO4 ¼ 2.77, NaCl ¼ 1.73, NaBr
¼ 1.47, and NaI ¼ 1.14.19 Hence, based on the values of C1/2
previously mentioned, it becomes readily apparent that the air/
water surface tension increments of the individual ions cannot be
used to predict deswelling behavior. An argument based solely
on air–surface tension increment would suggest that C1/2 for
NaCl would be 2.77/1.73 or 1.6 times the value C1/2 for Na2SO4.
Likewise, NaBr and NaI would be 1.9 and 2.4 times the value of
Na2SO4, respectively. The fact that each of these ratios under-
predicts the experimentally observed values suggests that ions
may be directly interacting with the polymer, perhaps through
a dipole–ion interaction that stabilizes the swollen network.
By changing the temperature of the experiment, the subtle
effects of the ions can be distinctly observed. Fig. 3 and 4 show
similar experiments conducted at 15 and 24 ꢁC. At both temper-
atures, the deswelling isotherms for poly(NIPAAm) and poly
(DEAAm)overlapinNa2SO4 andNaCl. However, theratioofC1/2
at 5 ꢁC to C1/2 at 24 ꢁC is different for the two salts. The value of
Fig. 1 Deswelling isobars of poly(NIPAAm) and poly(DEAAm)
showing a demixing temperature near 30 ꢁC.
alter the transition temperature with respect to poly(NIPAAm);
however, the deswelling isobar is more gradual than in poly
(NIPAAm).
Fig. 2 shows the deswelling isotherms, or the variation in the
dimensionless thickness of both poly(NIPAAm) and poly
(DEAAm), over a range of salt concentrations (0–2.0 M) for
Na2SO4 (Fig. 2a), NaCl (Figꢁ . 2b), NaBr (Fig. 2c), and NaI
(Fig. 2d), all carried out at 5 C. The dimensionless thickness is
defined as (H ꢀ Hcollapsed)/(Hswollen ꢀ Hcollapsed), where H is the
thickness at the respective salt concentration, Hswollen is the
thickness in water at 5 ꢁC and Hcollapsed is the thickness at 42 ꢁC.
Comparing poly(DEAAm) and poly(NIPAAm), despite the
difference in the collapse profiles with respect to temperature,
C
1/2 for Na2SO4 at 5 ꢁC (Fig. 2a) is 2.2 times the value at 24 ꢁC
(Fig. 4a); whereas, the value of C1/2 for NaCl at 5 ꢁC (Fig. 2b) is 3
Fig. 2 Variation in thickness of poly(NIPAAm) (-) and poly(DEAAm) (O) at 5 ꢁC in between 0 and 2.0 M salt concentration for (a) Na2SO4; (b)
NaCl; (c) NaBr; and (d) NaI.
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 6061–6067 | 6063

View Article Online
Fig. 3 Variation in thickness of poly(NIPAAm) (-) and poly(DEAAm) (O) at 15 ꢁC in between 0 and 2.0 M salt concentration for (a) Na2SO4; (b)
NaCl; (c) NaBr; and (d) NaI.
Fig. 4 Variation in thickness of poly(NIPAAm) (-) and poly(DEAAm) (O) at 24 ꢁC in between 0 and 2.0 M salt concentration for (a) Na2SO4; (b)
NaCl; (c) NaBr; and (d) NaI.
6064 | Soft Matter, 2011, 7, 6061–6067
This journal is ª The Royal Society of Chemistry 2011

View Article Online
times the value at 24 ꢁC (Fig. 4b). This may be due to an
The amideII mode isthe out-of-phasecombination of the N–H in-
plane bend and the C–N stretching vibration with smaller
contributions from the C]O in-plane bend and the C–C and N–C
stretching vibrations. As a general rule, hydrogen bonding lowers
the frequency of amide I since it lowers the restoring force of the
C]O stretching vibration, but increases the frequency of amide II
since it produces an additional restoring force.34
2ꢀ
increasingly stronger interaction of Clꢀ with the amide than SO4
as the temperature is increased.
For NaBr, there is a small difference in deswelling isotherms of
poly(DEAAm) and poly(NIPAAm) at 15 ꢁC, where NaBr is
perhaps slightly more effective at salting-out poly(DEAAm).
This difference becomes significant, however, at 24 ꢁC, where
C1/2 for poly(NIPAAm) is 1.0 M and for poly(DEAAm) it is
0.7 M. Moreover, a slight salting-in effect is measured for poly
(NIPAAm), which increases with temperature. At 24 ꢁC, the
maximum degree of salting-in is 8%.
At 24 ꢁC, the amide I peak consists of one band that is centered
at 1625 cmꢀ1 and the amide II consꢁists of one band at 1558 cmꢀ1
.
As temperature is increased to 50 C, both bands shift to lower
frequencies. A distinct sub-band arises, however, at 1651 cmꢀ1 in
the amide I region. It is believed that this band arises due to the
hindered ability of carbonyl moiety to hydrogen bond water,
which increases the electron density leading to a higher stiffness.
Hence, upon increasing temperature, the overall amide I peak first
decreases in frequency followed by an increase in frequency due to
the rise in the sub-band. The amide II, in contrast, does not
develop and secondary bands and shifts monotonically towards
lower frequencies. It is thought that as the layer collapses, the N–
H group, like the C]O group, can no longer adopt a conforma-
tion for optimal hydrogen bonding, and hence the frequency of
N–H bending vibrations is lowered due to a lower restoring force.
We have previously shown that in a collapsed layer, 2–3 water
molecules per segment remain and that the amide hydrogens are
readily exchanged with deuterium.35 This suggests the relative
absence of the intermolecular hydrogen-bonding between C]O
and N–H moieties.
With respect to NaI, there is considerable deviation in the
deswelling isotherms for poly(NIPAAm) and poly(DEAAm),
with the deviation becoming more prominent as the temperature
is increased. At both 15 and 24 ꢁC, NaI is more effective at
deswelling poly(DEAAm) than poly(NIPAAm). Moreover, NaI
salts-in poly(NIPAAm) at low salt concentrations, with the
degree of salting-in increasing with temperature. In contrast, the
opposite trend is seen with poly(DEAAm), where the degree of
salting-in decreases with temperature. That said, the maximum
degree of salting-in is 5% for poly(DEAAm) at 5 ꢁC and 18% for
poly(NIPAAm) at 24 ꢁC.
It was noted that that the bulk surface tension increments of
the individual salts are insufficient to quantitatively interpret the
deswelling isotherms. It is tempting to explain the deviation with
respect to a difference in the partitioning of ions between the
macromolecule surface and bulk water and the partitioning ions
between an air interface and bulk water interface. At the air/
water interface, ions that possess a large charge/volume ratio are
generally excluded from the interface, as they are less likely to
give up their hydration waters.19,33 Ions with a smaller charge/
volume ratio can more easily forfeit their hydration waters and
therefore are able to access the interfacial zone. A question is
how the presence of dipoles influences ion partitioning. Towards
this end, we use ATR-FTIR to study the amide group and
aliphatic CH3 vibrations of poly(NIPAAm) and poly(DEAAm)
upon exposure to salt.
Fig. 5 shows the FTIR and the associated 2nd derivative spec-
trum of the amide I and amide II regions of poly(NIPAAm) in
pure water at both 24 and 50 ꢁC. The amide I vibration, absorbing
near 1650 cmꢀ1, arises mainly from the C]O stretching vibration
with minor contributions from the out-of-phase CN stretching
vibration, the C–C–N deformation and the N–H in-plane bend.
As the frequency of the vibrations is clearly influenced by the
amount of water in the network, rather than comparing
the frequencies as a function of salt concentration, we compare
the vibrational frequency at the same degree of swelling for each
salt, as determined from the ellipsometry measurements.
Fig. 6 shows the frequency of the amide I of poly(NIPAAm)
versus degree of swelling in Na2SO4, NaCl, NaBr, and NaI at
ꢁ
24 C. In all cases the amide I shows an initial decrease in the
vibration frequency as salt is added followed by an increase in the
frequency as the layer collapses, which is caused by the rise of
the sub-band around 1651 cmꢀ1. The frequency trends for all
salts overlap (to within +/ꢀ 0.5 cmꢀ1); hence, the frequencies of
the vibrations appear independent of ion and depend only on the
degree of swelling. In contrast, Fig. 7 shows the amide II, where
now the frequencies no longer overlap. Most notably at a degree
of swelling (H/Hcollapsed) of 2.4, a frequency of 5 cmꢀ1 separates
Fig. 5 FTIR spectra of the amide I and amide II vibrations of poly(NIPAAm) in water at 50 ꢁC and 24 ꢁC.
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 6061–6067 | 6065

View Article Online
In poly(DEAAm) (which lacks a N–H moiety), the frequency
of the amide I vibration now depends strongly on the ion (Fig. 8).
The vibration associated with Na2SO4 displays the same U-shape
trend as seen in poly(NIPAAm), which is quite close to the
frequency trend in pure water upon collapse. In NaCl, the depth
of the minimum in the U-shape becomes shallower. In NaBr and
NaI, the minimum disappears and now the frequency increases
monotonically with temperature. We now speculate that in the
case of poly(DEAAm), ions interact primarily with the partially
positive end of C]O dipole. The interaction weakens the ability
of the C]O to hydrogen bond with water and shifts the
frequency of the vibration to higher frequencies. As in the case of
the amide II vibration in poly(NIPAAm), the strength of the
interaction follows the Hofmeister series. The proposed dipole–
ion interactions for both poly(NIPAAm) and poly(DEAAm) are
shown in Fig. 9.
Fig. 6 Frequency shift of Amide I band as a function of degree of
swelling of poly(NIPAAm) at 24 ꢁC from 0–2.0 M salt concentrations.
Finally, it is of note to examine the CH3 asymmetric stretching
vibration in both poly(DEAAm) and poly(NIPAAm). This
vibration was chosen as it is the most prominent of all the aliphatic
vibrations, it shows a relatively large shift upon collapse (8 cmꢀ1),
and it provides a direct comparison of the 2 terminal carbons of
the isopropyl group with the 2 terminal carbons of the ethyl
groups. The CH3 asymmetric vibrations are shown in Fig. 10 as
a function of salt for both poly(NIPAAm) and poly(DEAAm).
The CH3 asymmetric stretching vibration collapses to the same
curve for all the salts, with the exception of poly(DEAAm) in NaI.
The relative independence of the CH3 asymmetric stretching
vibration on the salt ion suggests negligible accumulation of ions
on the hydrophobic surface, as expected from bulk air/water
surface tension measurements. Mysteriously, however, NaI cau-
ses an increase of 3 cmꢀ1 in the CH3 vibration in poly(DEAAm),
which is not observed in poly(NIPAAm). The increase may be
NaI (1551 cmꢀ1) and Na2SO4 (1555 cmꢀ1). The corresponding
degree of swelling in pure water occurs at 30 ꢁC, where the
position of the amide II is centered at approximately 1555 cmꢀ1
,
the same position of Na2SO4. While the proximity of the two
vibrations suggests negligible interaction of Na2SO4 with the
amide, a direct comparison is probably inappropriate due to the
different temperatures of the experiment (24 ꢁC vs. 30 ꢁC). That
said, for all values of H/Hcollapsed above unity, the amide II
vibration of poly(NIPAAm) is lowest in NaI, followed by NaBr,
NaCl and then Na2SO4. Interestingly, the trend follows the order
of the Hofmeister series.
We speculate that the ions are electrostatically attracted to the
O]C–N dipole and destabilize or sterically inhibit hydrogen
bonding at the N–H moiety, decreasing the amide II frequency.
NaI, which exhibits the largest effect, thus displays the largest
ion–dipole attraction, which correlates with the salting-in effect
measured with ellipsometry. Moreover, the trend in the vibration
frequencies could explain the increase in the bulk air/water
surface tension at C1/2 for each of the salts; in other words, the
stronger the ion–dipole attraction, the greater the required
surface tension to collapse the layer. As the frequency of amide I
is unaffected by the type of ion, it is presumed that binding or
attraction is primarily between the anion and the partially posi-
tive end of the O]C–N and does not affect hydrogen bonding to
the partially negative O]C end of the dipole.
Fig. 8 Frequency shift of Amide I band as a function of degree of
swelling of poly(DEAAm) at 24 ꢁC from 0–2.0 M salt concentrations.
Fig. 7 Frequency shift of Amide II band as a function of degree of
swelling of poly(NIPAAm) at 24 ꢁC from 0–2.0 M salt concentrations.
Fig. 9 Proposed sites of the amide–ion interaction in both (a) poly(N-
isopropylacrylamide) and (b) poly(N,N-diethylacrylamide).
6066 | Soft Matter, 2011, 7, 6061–6067
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Fig. 10 CH3 stretch vibrations as a function of degree of swelling at 24 ꢁC from 0–2.0 M salt concentrations for (a) poly(DEAAm) and (b) poly
(NIPAAm).
7 Y. J. Zhang, S. Furyk, D. E. Bergbreiter and P. S. Cremer, J. Am.
Chem. Soc., 2005, 127, 14505–14510.
8 T. G. Park and A. S. Hoffman, Macromolecules, 1993, 26, 5045–5048.
9 H. Inomata, S. Goto, K. Otake and S. Saito, Langmuir, 1992, 8, 687–
690.
a consequence of the close proximity between an iodide and the
CH3 group resulting from attraction of the ion to the amide
dipole. In contrast, the amide in poly(NIPAAm) is less sterically
hindered and may better accommodate an iodide without
perturbation of the CH3 vibration.
ꢀ
10 E. Sedlak, L. Stagg and P. Wittung-Stafshede, Arch. Biochem.
Biophys., 2008, 479, 69–73.
11 H. G. Schild and D. A. Tirrell, J. Phys. Chem., 1990, 94, 4352–4356.
12 W. Kunz, P. Lo Nostro and B. W. Ninham, Curr. Opin. Colloid
Interface Sci., 2004, 9, 1–18.
IV. Conclusions
The mechanisms by which ions affect the solubility of neutral
macromolecules remain unresolved. Herein, we explored the
deswelling isotherms of both poly(NIPAAm) and poly(DEAAm)
induced through the addition of Na2SO4, NaCl, NaBr, and NaI.
In the presence of early members of the Hofmeister series, the
demixing isotherms of both poly(NIPAAm) and poly(DEAAm)
are identical. Later members of the Hofmeister series, in contrast,
impart distinct deswelling isotherms for poly(NIPAAm) and poly
(DEAAm), due to differences in accumulation of ions at the amide
dipole. The strength of the ion–dipole interaction appears to be
dictated by surrounding hydrophobicity or steric hindrance of the
amide, as later members of the Hofmeister series collapse poly
(DEAAm) more quickly than poly(NIPAAm). Finally, the rela-
tive absence of a dependence on the CH3 shift and ion suggests
that ions do not adsorb to the hydrophobic moieties. The one
exception is NaI and poly(DEAAm), which may not arise from
adsorption but close proximity of the ion to the diethyl groups due
to interaction with the amide dipole.
13 L. Vrbka, P. Jungwirth, P. Bauduin, D. Touraud and W. Kunz,
J. Phys. Chem. B, 2006, 110, 7036–7043.
14 R. L. Baldwin, Biophys. J., 1996, 71, 2056–2063.
15 L. M. Pegram and M. T. Record, J. Phys. Chem. B, 2008, 112, 9428–
9436.
16 F. Hofmeister, Arch. Exp. Pathol. Pharmakol., 1888, 24, 247–260.
17 A. Voet, Chem. Rev., 1936, 20, 169–179.
18 A. Hamabata and P. H. Vonhippel, Biochemistry, 1973, 12, 1264–
1271.
19 L. M. Pegram and M. T. Record, J. Phys. Chem. B, 2007, 111, 5411–
5417.
20 P. B. Petersen and R. J. Saykally, Annu. Rev. Phys. Chem., 2006, 57,
333–364.
21 W. F. Mcdevit and F. A. Long, J. Am. Chem. Soc., 1952, 74, 1773–
1777.
22 P. K. Nandi and D. R. Robinson, J. Am. Chem. Soc., 1972, 94, 1299.
23 O. S. Lawal, Food Chem., 2006, 95, 101–107.
24 T. H. Plumridge and R. D. Waigh, J. Pharm. Pharmacol., 2002, 54,
1155–1179.
25 Y. Maeda, T. Nakamura and I. Ikeda, Macromolecules, 2001, 34,
1391.
26 A. Percot, X. X. Zhu and M. Lafleur, J. Polym. Sci., Part B: Polym.
Phys., 2000, 38, 907.
27 E. Kesselman, O. Ramon, R. Berkovici and Y. Paz, Polym. Adv.
Technol., 2002, 13, 982–991.
28 Y. Maeda, T. Nakamura and I. Ikeda, Macromolecules, 2002, 35,
10172–10177.
29 Y. Paz, E. Kesselman, L. Fahoum, I. Portnaya and O. Ramon,
J. Polym. Sci., Part B: Polym. Phys., 2004, 42, 33–46.
30 A. Vidyasagar, J. Majewski and R. Toomey, Macromolecules, 2008,
41, 919–924.
31 J. Habicht, M. Schmidt, J. Ruhe and D. Johannsmann, Langmuir,
1999, 15, 2460–2465.
Acknowledgements
This work was supported in part by Camille and Henry Dreyfus
Foundation and NSF DMR-064557.
References
1 P. H. Von Hippel and T. Schleich, Acc. Chem. Res., 1969, 2, 257–265.
2 Y. J. Zhang and P. S. Cremer, Curr. Opin. Chem. Biol., 2006, 10, 658–
663.
32 M. W. Born and E. Wolf, Principles of Optics, University Press,
Amsterdam, 1997.
33 K. D. Collins, Biophys. J., 1997, 72, 65–76.
34 N. B. Colthup, L. H. Daly and S. E. Wiberley, Introduction to Infrared
and Raman Spectroscopy, Academic Press, New York, 1990.
35 A. Vidyasagar, H. L. Smith, J. Majewski and R. G. Toomey, Soft
Matter, 2009, 5, 4733–4738.
3 J. E. Kinsella and D. Srinivasan, J. Biol. Chem., 1981, 256, 3394–3398.
4 M. J. Hey, J. M. Clough and D. J. Taylor, Nature, 1976, 262, 807–809.
5 R. Freitag and F. Garret-Flaudy, Langmuir, 2002, 18, 3434–3440.
6 Y. Zhang, S. Furyk, L. B. Sagle, Y. Cho, D. E. Bergbreiter and
P. S. Cremer, J. Phys. Chem. C, 2007, 111, 8916–8924.
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 6061–6067 | 6067