Subscriber access provided by UNIV OF CONNECTICUT
Article
Wetting Behavior of Lightly Sulfonated
Polystyrene Ionomers on Silica Surfaces
Xiaowen Zhai, and R. A. Weiss
Langmuir, 2008, 24 (22), 12928-12935 • DOI: 10.1021/la802195j • Publication Date (Web): 23 October 2008
Downloaded from http://pubs.acs.org on November 20, 2008
More About This Article
Additional resources and features associated with this article are available within the HTML version:
•
•
•
•
Supporting Information
Access to high resolution figures
Links to articles and content related to this article
Copyright permission to reproduce figures and/or text from this article
Langmuir is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
12928
Langmuir 2008, 24, 12928-12935
Wetting Behavior of Lightly Sulfonated Polystyrene Ionomers on Silica
Surfaces
Xiaowen Zhai and R. A. Weiss*
Department of Chemical Engineering and Polymer Science Program, UniVersity of Connecticut, Storrs,
Connecticut 06269-3136
ReceiVed July 10, 2008. ReVised Manuscript ReceiVed September 18, 2008
The wetting/dewetting behavior of thin films of lightly sulfonated low molecular weight polystyrene (SPS) ionomers
spin-coated onto silica surfaces were studied using atomic force microscopy (AFM), contact angle measurements, and
electron microscopy. The effects of the sulfonation level, the choice of the cation, the solvent used to spin-coat the
films, and the molecular weight of the ionomer were investigated. Small angle X-ray scattering was used to determine
the bulk microstructure of the films. The addition of the sulfonate groups suppressed the dewetting behavior of the
PS above its glass transition temperature, e.g. no dewetting occurred even after 240 h of annealing at 120 °C. Increasing
the sulfonation level led to more homogeneous and smoother surfaces. The choice of the cation used affected the
wetting properties, but not in a predictable manner. When tetrahydrofuran (THF) or a THF/methanol mixed solvent
was used for spin-casting, a submicron-textured surface morphology was produced, which may be a consequence of
spinodal decomposition of the film surface during casting. Upon annealing for long times, the particles coalesced into
a coherent, nonwetted film.
Introduction
Contemporary coatings technologies require increasingly
thinner polymeric films, and the ability to produce useful and
uniform coatings depends on the wettability and adhesion of the
polymer to the substrate.1,2 Dewetting of a coating exposes the
underlying substrate and compromises the barrier properties of
the coating, as well as producing surface roughness and defects
that may be deleterious to the performance of the polymer film.3
Dewetting of polymeric films may be prevented by modifying
the equilibrium polymer-surface interactions.4
Controlling film stability is crucial for obtaining uniform,
continuous,defect-free,andstablecoatings.Thepolymer-substrate
and polymer-air interfacial energies and kinetic barriers play
important roles in determining the morphology and dynamics of
thin films. Confinement influences conformation and interaction
of the molecules.5 Structural changes induced by the interaction
with a solid substrate can also affect the dynamics of polymers,
as well as the adhesion and wetting characteristics.6,7
The wetting/dewetting of polystyrene (PS) has received
considerable attention.8-16 Thin PS films are metastable when
the spreading coefficient, S, is less than zero, where
S ) Γs - (Γsp + Γp)
(1)
Γs and Γp are the surface free energies of the substrate and polymer,
respectively, and Γsp is the interfacial energy between the polymer
and the substrate. For glassy polymers such as PS, wetting can
be metastable below the glass transition temperature (Tg), but
* To whom correspondence should be addressed. E-mail: rweiss@
ims.uconn.edu.
(1) Geoghegan, M.; Krausch, G. Prof. Polym. Sci. 2003, 28, 261–302.
(2) Green, P. F. J. Polym. Sci., Part B: Polym. Phys. 2003, 41, 2219–2235.
(3) Wu, S. Polymer Interfaces and Adhesion; Marcel Dekker, Inc.: New York,
1982.
(4) Koberstein, J. T. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 2942–
2956.
(5) Yerushalmi-Rozen, R.; Klein, J.; Fetters, L. J. Science 1994, 263, 793–
795.
(6) Kerle, T.; Yerushalmi-Rozen, R.; Klein, J.; Fetters, L. J. Europhys. Lett.
1998, 44, 484–490.
(7) Henn, G.; Bucknall, D. G.; Stamm, M.; Vanhoorne, P.; Jerome, P.
Macromolecules 1996, 29, 305–4313.
dewetting occurs above Tg. The dewetting process is driven by
long-range interactions and proceeds in three basic steps: rupture,
expansion, and coalescence of dewetted holes to form a network,
and, finally, break-up of the network system into droplets.17,18
Strategies to stabilize thin polymer films against dewetting
include chemical and physical modifications of the substrate,
introduction of reactive end-groups onto the polymer, and the
addition of nanoparticle or dendrimer fillers to the polymer.8,19-23
Previous work in our group showed that randomly sulfonated
polystyrene (SPS) ionomers were much more resistant to
dewetting from silica than PS with the same molecular weight.24
This paper addresses the effects of sulfonation level, molecular
weight, the choice of the cation, and the solvent history of the
SPS ionomer on its wetting and dewetting behavior on silica.
Experimental Details
Materials and Sample Preparation. Toluene (99.9%), tetrahydrofuran (THF, 99.9%) methanol (99.9%), and 1,2-dichloroethane
(reagent grade) were obtained from Fisher Scientific Co. and used
(8) Brochard-Wyart, F.; Martin, P.; Redon, C. Langmuir 1993, 9, 3682–3690.
(9) Reiter, G. Phys. ReV. Lett. 1992, 68, 75–78.
(10) Brochard-Wyart, F.; Daillant, J. Can. J. Phys 1990, 68, 1084–1088.
(11) Brochard-Wyart, F.; Redon, C.; Sykes, C. C. R. Acad. Sci. 1992, 314,
19–24.
(12) Jones, R. A. L. Polymer 1994, 35, 2160–2166.
(13) Martin, P.; Buguin, A.; Brochard-Wyart, F. Europhys. Lett. 1994, 28,
421–426.
(14) Lambooy, P.; Phelan, K. C.; Haugg, O.; Krausch, G. Phys. ReV. Lett.
1996, 76, 1110–1113.
(15) Qu, S.; Clark, C. J.; Liu, Y.; Rafaillovich, M. H.; Sokolov, J.; Phelan,
K. C.; Krausch, G. Macromolecules 1997, 30, 3640–3645.
(16) Karapanagiotis, I.; Gerberich, W. W. Surf. Sci. 2005, 594, 192–202.
(17) Reiter, G. Langmuir 1993, 9, 1344–1351.
(18) Sharma, A.; Reiter, G. J. Colloid Interface Sci. 1996, 178, 383–399.
(19) Yerushalmi-Rozen, R.; Klein, J.; Fetters, L. J. Science 1994, 263, 793–
795.
(20) Kerle, T.; Yerushalmi-Rozen, R.; Klein, J.; Fetters, L. J. Europhys. Lett.
1998, 44, 484–490.
(21) Sharma, S.; Rafailovich, M. H.; Peiffer, D.; Sokolov, J. Nano Lett. 2001,
1, 511–514.
(22) Mackay, M. E.; Hong, Y.; Jeong, M.; Hong, S.; Russell, T. P.; Hawker,
C. J.; Vestberg, R.; Douglas, J. F. Langmuir 2002, 18, 1877–1882.
(23) Barnes, K. A.; Karim, A.; Douglas, J. F.; Nakatani, A. I.; Gruell, H.;
Amis, E. J. Macromolecules 2000, 33, 4177–4177.
(24) Feng, Y.; Karim, A.; Weiss, R. A.; Douglas, J. F.; Han, C. C.
Macromolecules 1998, 31, 484–493.
10.1021/la802195j CCC: $40.75  2008 American Chemical Society
Published on Web 10/23/2008
Wetting BehaVior of SPS Ionomers on Silica
Langmuir, Vol. 24, No. 22, 2008 12929
Table 1. Sulfonated PS Ionomers (SPS) Used in This Study
sample
sulfonation
level
(mol %)a
cation
Mw (PS)
(kg/mol)
Nb
xavec
Tg
(°C)
2.5LiSPS4
6.5LiSPS4
2.5NaSPS4
2.5KSPS4
2.5RbSPS4
2.5CsSPS4
0.76LiSPS13.5
2.7LiSPS13.5
2.5
6.5
2.5
2.5
2.5
2.5
0.76
2.7
Li
Li
Na
K
Rb
Cs
Li
Li
4.0
4.0
4.0
4.0
4.0
4.0
13.5
13.5
39
39
39
39
39
39
130
130
1.0
2.5
1.0
1.0
1.0
1.0
1.0
3.5
87
95
88
87
88
88
102
104
a
p ) sulfonation level/100. b N ) degree of polymerization. c xave ) p · N.
as received. Two narrow molecular weight distribution PS precursors
were obtained from Pressure Chemicals: (1) Mw ) 4.0 kg/mol,
polydispersity index (PDI) < 1.06, Tg ) 82 °C and (2) Mw ) 13.5
kg/mol, PDI < 1.06, Tg ) 98 °C, denoted as PS4 and PS13.5,
respectively. These molecular weights are far below the chain
entanglement molecular weight (Me) of PS, 16.6 kg/mol.25 The PS’s
were sulfonated at 50 °C in 1,2-dichloroethane using acetyl sulfate,
prepared by the reaction of concentrated sulfuric acid (Fisher
Scientific Co., 95%) with acetic anhydride (Aldrich Chemical Co,
reagent grade) at 0 °C, as the sulfonating agent.24,26 The sulfonic
acid derivative of the sulfonated polystyrene ionomer (HSPS) was
recovered by flashing the solvent in boiling water and then drying
under vacuum at 120 °C. HSPS was neutralized to form alkali metal
salt (Li+, Na+, K+, Rb+, and Cs+) ionomers using the corresponding
alkali metal hydroxide (Aldrich Chemical Co.). The sulfonation
procedure randomly sulfonates the styrl ring, primarily at the paraposition. The nomenclature used for the ionomers was x.yMSPSz,
where x.y is the sulfonation level in mol% of styrene substituted,
M is the alkali cation, and z is the molecular weight of the parent
PS in kilograms per mole. The ionomer samples used in this study
are summarized in Table 1. The column labeled xave corresponds to
the average number of sulfonate groups per chain, calculated by
multiplying the sulfonation level by the degree of polymerization,
N. The glass transition temperatures (Tg) of the polymers were
determined with a differential scanning calorimeter (DSC), TA2920,
TA Instruments, using a heating rate of 20 °C/min.
Atomically smooth silicon wafers with {100} orientation and one
side polished were obtained from Wafer World, Inc. Prior to sample
deposition, the silicon wafers were treated in an ultrasonic bath for
10 min followed by immersion in a “piranha” solution (30%
concentrated hydrogen peroxide and 70% concentrated sulfuric acid)
for 1 h to convert the surface to silica. Afterward, the wafers were
rinsed thoroughly with deionized water and dried under a stream of
dry nitrogen.
Ionomer films were deposited onto silica surfaces from 1 wt %
solution by spin-coating using an SCS Corp. spin-coater, model
P6700, with a rotor speed of 3000 rpm at room temperature. Four
different solvents were used: (1) toluene, (2) THF, (3) a mixture of
toluene and methanol (9/1, v/v) and (4) a mixture of THF and
methanol (9/1, v/v). The dielectric constants, ε, of toluene, THF, and
methanol are 2.4, 7.5, and 33, respectively. The dielectric constants
of the solvent mixtures, estimated as the volume average of the
dielectric constants of the individual components of the mixture,
were 5.5 for toluene/methanol (9/1, v/v) and 10.1 for THF/methanol
(9/1, v/v). The solubility parameters, δ, of toluene, THF, methanol,
and PS are 18.2, 19.6, 29.6, and 19.1 J1/2/cm3/2, respectively, and the
solubility parameters of the solvent mixtures, also obtained using
a volume average of the solubility parameters of the individual
components, were 19.3 J1/2/cm3/2 for toluene/methanol (9/1, v/v)
and 20.6 J1/2/cm3/2 for THF/methanol (9/1,v/v).
The films on silica were annealed at constant temperature in a
vacuum oven for various times after spin-coating. The dewetting
(25) Graessley, W. W. Polymeric Liquids and Networks: Structure and
Properties; Garland Science Publishing: New York, 2004, p 454.
(26) Maowski, H. S. ; Lundberg, R. D.; Singhal, G. H. US Patent 3,870,841,
1975.
Figure 1. Probability P(x) distribution as a function of the number of
ionic groups per chain, x, for SPS ionomers with molecular weights of
4.0 and 13.5 kg/mol and four sulfonation levels.
behavior of the films on silica was assessed after removing them
from the oven and allowing them to equilibrate to room temperature.
Characterization. The surface morphology of the films was
imaged with an Asylum Research, Inc. atomic force microscope
(AFM; model MFP-3D) using the tapping mode and silicon tips (tip
curvature radii ∼ 5-10 nm) with a spring constant of about 2.5 N/m.
A variable angle spectroscopic ellipsometer (J.A. Woollam Co.,
Inc., model WVASE32) was used to measure film thicknesses
(∼30-40 nm) and refractive index (∼1.57-1.59) of the ionomer
films. A Cauchy model was used in the calculations.
Intrinsic viscosities of the SPS ionomers in solution at 25 °C were
measured with a Ubbelohde viscometer (Fisher Scientific Co, model1). Water contact angles on the films were measured with a custombuilt goniometer. Five measurements were made for each sample.
Small-angle X-ray scattering (SAXS) was carried out using a BrukerAnton-Paar small-angle scattering instrument with a Rigaku Ru300 rotating anode generator (Cu KR radiation) and a HiStar 2-D
detector. Samples were prepared by solvent casting a 5 wt % ionomer
solution onto a silica surface at ambient conditions and then scraping
off the material from the silica substrate after complete evaporation
of the solvent. The SAXS specimen consisted of an ionomer powder
that was encased in a 0.25 mm thick aluminum washer between
polyimide films (Kapton; DuPont de Nemours Co.). A background
correction was made by subtracting the scattering from two polyimide
films.
Results and Discussion
Sulfonation Distribution. Sulfonation in a homogeneous
solution is a random reaction. As a result, it is not possible to
control the precise number of sulfonate groups on any chain, so
the best characterization possible with these materials is an average
number of sulfonate groups per chain. A binomial distribution
was used to calculate the probability, P(x), that a chain with a
degree of polymerization, N, has x sulfonate groups on a chain
and where p is the probability of a monomer being sulfonated
(see eq 1). Note that p ) mole fraction of substituted styrene
units ) mol% sulfonation divided by 100.
P(x) )
N!
px(1 - p)(N-x)
(N - x) ! x!
Figure 1 shows the distribution function P(x) for the four
ionomers studied. Although the average number of sulfonate
groups per chain varied from 1 to 4 for the various ionomers
produced (see Table 1), the distributions show that each sample
will also have a finite number of chains with no sulfonation and
with more sulfonate groups than the average. The latter is
particularly important with regard to the experimental results
that are discussed below, in that the samples with an average of
12930 Langmuir, Vol. 24, No. 22, 2008
Figure 2. Schematic of chain structure for simple ion-dipole associations
in MSPS oligomers: (a) x ) 1; (b) x ) 2; (c) x ) 3, and (d) for an ionic
aggregate structure where xave g 3. The small circles represent the
sulfonate groups, and the large circles are ionic aggregates.
one sulfonate group per chain (x ) 1) will also have a finite
concentration of multiple substituted chains. For simple ion-dipole
associations, one would expect that chains with one or two
sulfonate groups (x ) 1 and x ) 2) only produce chain extension
of the oligomers (see Figure 2a,b). When x > 2, network formation
is possible. According to the distributions P(x) for the different
ionomers, one might reasonably expect that network formation
is possible in all of the samples studied here, since all should
contain some chains with x > 2 (see Figure 2c). If ionic
aggregation occurs, as SAXS data discussed later in this paper
indicates, more complicated structures are possible, including
networks with micelle-like structures, as shown in Figure 2d.
Dewetting of PS from Silica. The dewetting behavior of PS
from silica has been studied extensively by a number of
authors8-16 and is only discussed briefly here. Figure 3 shows
optical micrographs of spin-coated PS films with Mw ) 4 kg/mol
and 13.5 kg/mol annealed above Tg at 120 °C for 20 h. The
dewetting process for both PS films followed the typical pattern
described by Reiter,9 which involves the following steps: (1)
The initially flat film ruptures spontaneously, resulting in the
formation of holes at random sites; (2) The holes then grow
laterally and coalesce, resulting in a polygon pattern of holes
with polymer ridges; (3) The polymer ridges finally break up
into spherical droplets. The dewetting kinetics of the higher
molecular weight PS film (Figure 3b) was slower than that of
the lower molecular weight PS film (Figure 3a). After 20 h of
annealing, the lower molecular weight film had completely broken
up into droplets, while the higher molecular weight film still
exhibited only dewetted holes typical of early stage dewetting,
i.e., after 20 h. The slower dynamics of the dewetting process
for the higher molecular weight PS is a consequence of decreased
molecular mobility due to the higher viscosity of the polymer
melt.
Effect of Sulfonation Level. Figure 4 shows AFM topographical images of 2.5LiSPS4 (Figure 4a-d) and 6.5LiSPS4
(Figure 4e-h) films spin-coated onto silica from toluene (Figure
4a,b,e,f) and THF (Figure 4c,d,g,h) and annealed under vacuum
Zhai and Weiss
Figure 3. Optical micrographs of spin-coated PS films annealed at 120
°C for 20 h on a silica surface: (a) Mw ) 4.0 kg/mol; the darker regions
are the polymer drops and the lighter regions are the silica surface; (b)
Mw ) 13.5 kg/mol; the continuous region is polymer and the lighter
regions within the darker polymer rim are holes (i.e., silica surface). The
inset is an AFM of one of the dewetted holes.
Figure 4. AFM topographical images of 2.5LiSPS4 (a-d) and 6.5LiSPS4
(e-h) on the silica surface spin coated from toluene (a,b,e,f) and THF
(c,d,g,h) annealed at 120 °C for 0 h (a,c,e,g) and 240 h (b,d,f,h). Scan
size 50 × 50 µm2 (a,b,e,f,g,h) and 20 × 20 µm2 (c,d).
at 120 °C for 0 and 240 h. The 2.5LiSPS4 did not dewet after
24 h and actually resisted dewetting even after 240 h of annealing
at 120 °C. This result is remarkable in that Figure 1 suggests that
Wetting BehaVior of SPS Ionomers on Silica
Langmuir, Vol. 24, No. 22, 2008 12931
Figure 5. SEM images of 2.5LiSPS4 spin-coated onto a silica surface
from a THF/Methanol (9:1) solution. (a) Top view and (b) side view.
∼37% of the chains in the 2.5LiSPS4 were unsulfonated and the
molecular weight was sufficiently low that the chains should not
be entangled. Figure 1 also indicates that ∼10% of the chains
have three or more sulfonate groups, which can produce network
formation (see Figure 2c). Interactions of the sulfonate groups
with the silica surface, e.g., ion-dipole interactions, should
improve wetting, and the network formation should suppress the
kinetics of dewetting. However, it is puzzling that unentangled,
unsulfonated chains, if present at the concentrations shown by
Figure 1, would not dewet sufficiently to be apparent in the
morphologies shown in Figure 4.
The 2.5LiSPS4 spin-coated from THF (Figure 4c) and a solvent
of (9/1 v/v) THF/methanol exhibited an unusual morphology
(see Figure 5) that was not caused by dewetting. This morphology
of well-wetted, submicron particles is believed to involve a
spinodal decomposition of the surface.27 When annealed at
elevated temperature, that morphology coalesced into a rough,
but wetted film (Figure 4d).
Increasing the ionic concentration of the SPS produced a higher
effective density of physical cross-links, which increased Tg and
the melt viscosity. The 6.5LiSPS4 films spin-coated from toluene
(Figure 4e,f) and THF (Figure 4g,h) were featureless (i.e.,
smoother and more homogeneous than the 2.5LiSPS4 films),
even after annealing 240 h at 120 °C. Figure 1 predicts that the
unsulfonated chain concentration in the 6.5LiSPS4 is only ∼8%,
and the concentration of chains with x g 3 is ∼50%. Thus, the
probability that a single chain has a functional group that can
complex with the silica is greater than that for the lower sulfonated
ionomer, and the higher physical cross-link density provides a
much higher viscosity that help suppress dewetting of these films
above Tg.
Effect of Solvent. The ionic groups of SPS ionomers tend to
aggregate in relatively nonpolar solvents due to the incompatibility
of the polar sulfonated groups with the solvent. The ionic
aggregation influences the properties and conformation of the
ionomer molecules in solutions.28 Therefore, the nature of the
solvent used in the spin-coating process was expected to affect
the conformation and size of the polymer coils in solution and
the location of the sulfonate groups and the polymer conformation
of the cast films, since the solvent is removed very rapidly during
spin-coating. On the basis of the differences in the dielectric
constants (2.4 and 5.5) and solubility parameters (18.2 and 19.6
J1/2/cm3/2), toluene was judged to be less polar than THF and
probably a poorer solvent for the sulfonate groups.
The polarity of the SPS ionomer surface did depend upon the
solvent used to spin-coat the films. Contact angle data for
2.5LiSPS4 films spin-coated from toluene and THF solution and
6.5LiSPS spin-coated from toluene are shown in Figure 6. For
the 2.5LiSPS4, the as-cast toluene solutions were more hydro(27) Weiss, R. A.; Zhai, X.; Dobrynin, A. V. Langmuir 2008, 24, 5218–5220.
(28) Hara, M.; Wu, J.; Lee, A. H. Macromolecules 1989, 22, 754–757.
Figure 6. Water contact angle on 2.5LiSPS4 (open symbols) and
6.5LiSPS4 (filled symbols) samples spin-cast unto silica from toluene
and THF as a function of annealing time at 120 °C.
Table 2. Intrinsic Viscosities of Ionomers at 25°C
intrinsic viscosity (ml/g)
sulfonate
conc.
ionomer (mol%) cation toluene
THF
2.5SPS4
′′
′′
′′
′′
6.5SPS4
8.68
4.60
4.40
ns
ns
14.58
a
2.5
′′
′′
′′
′′
6.5
Li
Na
K
Rb
Cs
Li
6.67
3.47
3.60
nsa
ns
7.81
toluene/
MeOH THF/MeOH
(9/1)
(9/1)
12.29
21.10
16.86
17.00
12.50
8.04
8.89
26.04
41.78
9.91
4.57
82.50
ns ) not soluble.
phobic, as judged by the higher water-air contact angle. This
result suggests that the THF evaporation during spin-coating
promotes the location of the sulfonate groups to the air-polymer
surface, at least to a greater extent than does toluene. Both solvents
are good solvents for PS, but the THF is probably a better solvent
for the polar, sulfonate groups. That is, in solution, one might
expect that the sulfonate groups are more likely to bury themselves
inside the polymer coil to avoid contact with the toluene solvent
than THF. That hypothesis is supported by the intrinsic viscosity
data for toluene and THF solutions given in Table 2. The higher
intrinsic viscosity for the THF solution indicates a more expanded
ionomer coil, which is expected if THF is more favorable to the
sulfonate groups than is toluene. The hydrophilic nature of
the films cast from toluene did, however, increase when the
sulfonation level of the ionomer increased (see Figure 6). In this
case, one might expect that the higher sulfonate group concentration and restrictions of the chain conformation of the polymer
film as it dries traps some sulfonate groups at the air-film
interface.
The 2.5LiSPS4 samples spin-coated from either toluene or
THF became more hydrophobic, i.e., the contact angle of water
increased, when annealed above the Tg of the ionomer. The
toluene-cast films, however, remained more hydrophobic than
the THF-cast films for a fixed annealing time. The increase in
the hydrophobicity with time may be explained by the lower
energy hydrocarbon part of the polymer displacing the more
polar sulfonate groups at the surface.29 Surprisingly, the higher
sulfonated, 6.5LiSPS4 sample spin-coated from toluene exhibited
(29) Bigelow, B. W.; Bailey, F. C.; Salaneck, W. R.; Pochan, J. M.; Pochan,
D. F.; Thomas, H. R.; Gibson, H. W. In Ions in Polymers; Advances in Chemistry Series 187; American Chemical Society: Washington, DC, 1980; pp
296-308.
12932 Langmuir, Vol. 24, No. 22, 2008
the opposite behavior. The as-cast sample had a water contact
angle similar to that of the 2.5LiSPS4 cast from toluene, but
decreased, i.e., the sample became more hydrophilic, upon
annealing at 120 °C (see Figure 6).
The differences in the contact angles for the 2.5LiSPS4 samples
spin-coated from the two different solvents may also be due to
the differences in the film morphology, as shown in Figure 4.
The films spin-coated from toluene (Figure 4a) were relatively
smooth and covered the silica, while those cast from THF
exhibited a submicron particle morphology (see Figures 4c and
5). Upon annealing, both films formed coherent films that covered
the silica, but were rougher than the toluene as-cast film. The
6.5Li-SPS4 spin-coated from either toluene or THF appeared to
be similar to the 2.5Li-SPS4 samples spin-coated from toluene,
but the annealed films were much smoother. Those morphologies,
however, do not appear to explain the increase in the hydrophilicity
of the 6.5Li-SPS4 film when annealed at 120 °C.
The addition of methanol to either toluene or THF increases
the average polarity of the solvent, and mixed polar-nonpolar
solvents can have a dramatic effect on the ionic aggregation and
chain conformation of ionomers in solution.30 The intrinsic
viscosity data in Table 2 show that the addition of methanol
expands the chain conformation of the Li-SPS4 ionomers,
especially in the case of the 6.5Li-SPS4 in a 9/1 (v/v) THF/
methanol solution. Previous solution viscosity studies by Lundberg and co-workers31,32 indicate that the methanol preferentially
solvates the ionic associations in toluene/methanol and THF/
methanol solutions.
Whereas the surface of the 2.5LiSPS4 spin-coated from toluene
was smooth and homogeneous (Figure 4a), the addition of
methanol to the solvent produced films with a small amount of
discrete particles on the surface of the film (Figure 7a). The
addition of methanol to the THF produced a wetted, submicron
particle morphology (c.f. Figures 4 and 5 and Figure 8), as did
the ionomers cast from THF alone, but the addition of methanol
produced larger particles (c.f. Figures 4c and 8a). The particles
in both samples eventually coalesced into relatively smooth films
after annealing at 120 °C (c.f. Figures 4d and 8c), even smoother
than the films spin-coated from the toluene solutions.
Effect of Counterion. The effect of the cation on the
morphology of the SPS ionomers, 2.5MSPS4 (where M represents
an alkali metal cation: Li, Na, K, Rb or Cs) is shown in Figures
7 and 8 for samples spin-coated onto silica from 9/1 (v/v) toluene/
methanol and THF/methanol solutions, respectively. Changing
the cation changes the strength of the ion-pair, which affects the
ion-dipole associations in the ionomer and with a solid surface.
The strength of the ion-pair scales as the Coulomb energy, which
for a fixed anion (i.e., for the sulfonate, charge ) 1), is q/a,
where q is the charge of the ion and a is the ionic radius.33 The
values of q/a for the alkali metal salts are given in Table 3. For
the alkali metal salts, the strength of the ion-pair decreases with
increasing ionic radius. The distribution of particle sizes for the
preannealed samples of the Li, Na, and Rb salts are shown in
Figure 9. The particles were not well resolved in the micrographs
for the K and Cs salts. In general, there does not appear to be
any trend in the data with respect to the strength of the ion-pair.
The average size, Dm, of the particles is also shown in Figure
8. These values are similar for each salt.
(30) Chakrabarty, K.; Shao, P.; Weiss, R. A. In Ionomers: Synthesis, Structure,
Properties and Applications; Mauritz, K. A., Wilkes, G. L., Tant, M. R., Eds.;
Blackie Academic & Professional: London/New York, 1997; pp 158-207.
(31) Lundberg, R. D.; Makowski, H. S. J. Polym. Sci., Polym. Phys. Ed. 1980,
18, 1821–1836.
(32) Lundberg, R. D.; Phillips, R. R. J. Polym. Sci., Polym. Phys. Ed. 1982,
20, 1143–1154.
(33) Eisenberg, A. Macromolecules 1971, 4, 125–128.
Zhai and Weiss
Figure 7. AFM topographical images of 2.5MSPS4 on silica, spincoated from toluene/methanol and annealed at 120 °C for 0 h (a,c,e,g,i)
and 240 h (b,d,f,h,j). M is the alkali cation: Li (a,b), Na (c,d), K (e,f),
Rb (g,h) and Cs (i,j). Scan size is 50 × 50 µm2.
The morphology of the as-cast samples for all of the salts was
similar to the results described above for 2.5LiSPS4. Coherent,
but rough films with some holes were produced from the toluene/
methanol solutions. Annealing the toluene/methanol films
roughened the surface, but no dewetting was observed for any
of the salts (see Figure 7). The submicron particles produced
from THF/methanol casting became larger and anisotropic in
shape upon annealing. After 240 h at 120 °C, coherent films
were formed, although the surfaces were rough, with the notable
exception of the 2.5LiSPS4 sample (Figure 8). None of the salts,
however, exhibited any dewetting after 240 h of annealing.
The transition kinetics of the particle morphology was affected
by the choice of the cation (see Figure 8), but the quantitative
nature of the results was not easily correlated to q/a. In all cases,
the particles coalesced and flowed into a film, but the process
generally took at least 20 h or more. The Cs salt, which has the
weakest ion-pair, appeared to lose the particle morphology the
quickest and formed a rough film after 20 h of annealing at 120
°C. The Li salt, which has the strongest ion-pair, formed the
most homogeneous and smoothest surface after 240 h. A variety
of morphologies were displayed by the different salts, although,
for the most part, annealing produced very rough films, and it
was clear that the origin of the film formation began with
coalescence and growth of the particles.
The addition of methanol to toluene or THF preferentially
solvates the ion-dipole interactions, which expands the polymer
coils. As a result, [η] increased upon the addition of methanol
for any of the alkali metal salts (see Table 2). However, one
might expect that [η] should decrease with increasing strength
and concentration of the ion-pairs, i.e., increasing q/a and the
number of interacting groups per chain, since the interacting
ion-pairs in those cases should be less amenable to solvation by
the methanol cosolvent. The data in Table 2 show that, for the
Wetting BehaVior of SPS Ionomers on Silica
Langmuir, Vol. 24, No. 22, 2008 12933
Figure 9. Particle size distribution measured from AFM micrographs
in Figure 8 of various salts of 2.5MSPS4 spin-coated onto silica from
THF/methanol. The inset shows the average particle size, Dm, for each
salt.
Figure 8. AFM topographical images of 2.5MSPS4 on the silica surface
spin-coated from THF/methanol annealed at 120 °C for 0 h (a,d,g,j,m),
20 h (b,e,h,k,n), and 240 h (c,f,i,l,o). Here, M represents alkali cation:
Li (a-c), Na (d-f), K (g-i), Rb (j-l) and Cs (m-o). Scan size 50 ×
50 µm2 (f,i,l,m-o) and 20 × 20 µm2 (a-c,d,e,g,h,j,k).
Table 3. Ion-Pair Strength (q/a) of the Alkali Metal Salts
(q ) 1) of SPS
cation
ionic radius, a (nm)
q/a (nm-1)
Li
Na
K
Rb
Cs
0.050
0.102
0.138
0.149
0.170
20.00
9.80
7.25
6.71
5.88
2.5SPS4 ionomers, [η] of the Li salt in the mixed solvents was
relatively low, as expected since that salt is expected to have the
strongest ion-pair. [η] increased for the Na salt, which again is
consistent with the idea that the methanol is able to more efficiently
solvate the weaker Na-sulfonate interactions.
However, as q/a decreased more, the [η] appeared to pass
through a maximum, for the K salt in toluene/MeOH and the Na
salt in THF/methanol solutions. Lower [η]’s were observed for
the Rb and Cs salts, even though those salts had the weakest
ion-pairs. While inconsistent with the idea that methanol can
more easily solvate lower salts with lower q/a, the decrease in
the intrinsic viscosity of the Rb and Cs salts compared with the
Li, Na, and K salts in the mixed solvents are consistent with the
solubility of the different salts in THF and toluene. The stronger
ion-pairs (Li, Na, and K salts) dissolved, but the weaker ionpairs (Rb and Cs) did not (see Table 2). The reason for the
unexpected cation effect is not yet known, but it would seem
apparent that there is some specific unfavorable interaction of
the larger cations with the nonpolar solvent that has an opposite
effect on solubility than the ionic group solvation by the methanol.
In that case, the lowering of the [η] when methanol was added
to the Rb and Cs salts (and K salt for the toluene/methanol
solutions) may be more related to unfavorable interactions of the
methanol with the PS backbone than favorable interactions with
the sulfonate groups. These complicated mixed solvent effects
deserve more scrutiny than was possible in this study.
The [η] data for the 6.5LiSPS4 sample are equally puzzling.
For toluene and THF solutions, the [η]’s are higher than those
for 2.5LiSPS4. Note that the degree of polymerization was the
same for both ionomers. Then, all else equal, one would expect
the two nonpolar solvents to become poorer solvents for the
ionomer as the concentration of ionic groups increased. Yet, the
data in Table 2 suggest otherwise. When methanol was added
to the toluene solvent, the [η] of 6.5LiSPS4 did not change
appreciably. Yet, the addition of methanol to THF produced a
large increase in [η], indicating a more expanded chain, or more
likely, a longer effective chain length due to the multiple chain
associations as shown in Figure 2d. Thus, even for these low
molecular weight ionomers, the intrinsic viscosity measurement
is not representative of individual chains, but of soluble aggregates
of multiple chains.
Initially, it was thought that the particle texture shown in Figure
5 might be related to the size of the polymer in solution. However,
the data in Table 2 indicate that this is not the case, or at least
it is not that simple. The particle texture occurred only when
THF or THF/MeOH solutions were used for the spin-casting,
but there is no consistent correlation between [η] and the
occurrence of the particle texture. The solvent effect, however,
can be explained if the particle texture results from a spinodal
decomposition of the surface during spin-coating (see ref 28).
For the ionomers with higher sulfonation level, 6.5LiSPS4
(Figure 4e-h and Figure 10) films with smooth surfaces were
formed after spin-coating, and annealing did not significantly
affect the surface smoothness or texture. The films formed from
THF/methanol were, however, a bit less homogeneous. In those
cases, it appears that the more robust network of associations
that formed at the higher sulfonation level dominated the film
morphology.
SAXS characterization of the microstructure of the ionomer
films is shown in Figures 11 and 12 for samples spin-coated onto
silica from single-component solvents (toluene and THF) and
from mixed solvents containing methanol, respectively. Ionomers
generally exhibit a SAXS peak corresponding to nanophase
separation of ionic aggregates with a characteristic size of d ∼
2-5 nm and high scattering intensity at very low scattering vector,
q ) 4π sin 2θ/λ (where λ ) the X-ray wavelength, and 2θ )
12934 Langmuir, Vol. 24, No. 22, 2008
Zhai and Weiss
Figure 10. AFM topographical images of 6.5MSPS4 on the silica surface
spin coated from toluene/methanol (a,b) and THF/methanol (c,d) annealed
at 120 °C for 0 h (a,c) and 240 h (c,d). Scan size 50 × 50 µm2.
Figure 12. SAXS of 2.5NaSPS4 ionomers spin-coated onto silica from
toluene, toluene/methanol (9/1 v/v), THF, and THF/methanol (9/1 v/v).
Figure 11. SAXS of SPS ionomers spin-coated onto silica from toluene
and THF solutions.
the scattering angle) that has been attributed to a heterogeneous
distribution of nonaggregated ion-pairs.34 Both features are
observed in the SAXS profiles of the 2.5MSPS4 ionomers shown
in Figures 11 and 12. The peak is weak and diffuse, but
qualitatively, it appears that the peak is better defined in the
samples spin-coated from toluene and toluene/MeOH mixtures.
The better developed peak from those samples is probably a
consequence of the lower polarity of toluene compared to THF
and, as a result, stronger aggregation of the ionic groups and
(34) Grady, B.; Cooper, S. L. In Ionomers: Synthesis, Structure, Properties
and Applications; Mauritz, K. A., Wilkes, G. L., Tant, M. R., Eds.; Blackie
Academic & Professional: London/New York, 1997, pp. 41-87.
better development of the nanophase separation of ion-rich
aggregates. As one might expect, the intensity of the scattering
for the 2.5MSPS4 increased as the electron density contrast
increased (i.e., as the atomic number of the cation increased and
as the ionic group concentration increased).
Assuming that the origin of the scattering is interparticle
interference between ionic aggregates, the average distance
between ionic aggregates (d ) 2π/q) was estimated from the
approximate position of the maximum of the scattering peak to
be 3.9-5.0 nm for the different samples. The better resolution
of the peak in the samples prepared from the toluene-based
solvents indicates that the particle texture formation, which
occurred in samples cast from THF-based solutions, is favored
by a poorer developed ionic nanostructure in these ionomers.
Effect of Molecular Weight. Figure 13 shows AFM topographical images of higher molecular weight LiSPS ionomers
spin-coated onto silica. The average number of sulfonate groups
per chain for 0.76LiSPS13.5 and 2.7LiSPS13.5 correspond to 1
and 3.5, respectively, which are comparable to the lower molecular
weight ionomers 2.5LiSPS4 and 6.5LiSPS4. The Tg’s of the
higher molecular weight ionomers were higher (102-104 °C),
which is mainly due to the higher Tg of the parent PS. With the
exception of the 0.76LiSPS13.5 sample spin-coated from THF,
the spin-coated higher molecular weight ionomers were smooth
and relatively featureless, and the surface morphology did not
change even after annealing for 240 h at 140 °C.
The 0.76LiSPS13.5 sample spin-coated from THF exhibited
a surface morphology of densely packed nanoparticles similar
to that of the 2.5LiSPS4 sample (c.f. Figures 4 and 13). When
annealed for 240 h at 140 °C, the particles coalesced into a
generally homogeneous film with a rough surface. There was,
however, some formation of very small holes in the annealed
sample (see the dark dots on the AFM photo in Figure 4a and
the topographical profile in Figure 4b of the dotted line drawn
in Figure 4a). These holes did not appear to grow or coalesce
as would be expected if they were due to dewetting of the film,
and they may have simply been imperfections arising from the
coalescence of the submicron particles.
Wetting BehaVior of SPS Ionomers on Silica
Figure 13. (a) AFM topographical images of 0.76LiSPS13.5 and
2.7LiSPS13.5 spin-coated onto silica from toluene and THF and annealed
at 140 °C. Scan size 50 × 50 µm2. (b) Topographical profile of the white
dotted line drawn on the AFM of the 0.76LiSPS13.5 sample cast from
THF and annealed at 140 °C for 240 h.
Conclusions
SPS ionomers can greatly improve the wetting behavior of PS
films on a silica surface, and ionomer films showed excellent
resistance to dewetting. The latter was true for unentangled ionomer
chains with any alkali metal sulfonate and sulfonation levels sufficient
to provide at least some chains with multiple sulfonate groups. The
choice of the solvent used to prepare the films affected the surface
morphology, but did not have a significant effect on dewetting. The
surface was smoother and more homogeneous for higher sulfonation
levels. The probability that a chain has a functional group that can
complex with the silica increases with increasing the sulfonation
level. Increasing the ionic concentration of the SPS also produced
a higher effective density of physical cross-links which provides a
much higher viscosity that also helps suppress dewetting of these
films above Tg.
The effect of the cation on the morphology of the SPS ionomer
films did not have a substantial effect on the surface morphology.
Annealing films roughened the surface, but no dewetting was
observed for any of the salts. Increasing the polarity of the solvent
used for spin-coating affected the morphologies of the 2.5MSPS
ionomer as-cast films from the homogeneous to a nanoparticle-
Langmuir, Vol. 24, No. 22, 2008 12935
textured surface, and the particles coalesced into a homogeneous
film upon annealing. The SPS ionomers with higher molecular
weight resulted in an even smoother and relatively featureless
surface on the silica surface.
A remarkable result was that it was not necessary that all the
chains be sulfonated to achieve the dewetting resistance. That
is, the homogeneous sulfonation reaction used to prepare the
sulfonate concentrations used in this study is expected to leave
a high fraction of chains unsulfonated. Since the chain length of
these polymers was far below the entanglement molecular weight
for PS, one would have expected that the unsulfonated chains
would dewet under the conditions used to anneal the samples.
But, no evidence of dewetting of any part of the sample was
observed in any of the experiments conducted in this study.
That observation is consistent with a study of the phase behavior
of SPS polymers where it was found impossible to fractionate
chains with different sulfonation concentrations, including no
sulfonate groups, by cooling a solution slowly from elevated
temperature.35 Even though the polymer samples consisted of a
distribution of sulfonate groups/chains, a distinct upper critical
solution temperature (UCST), which varied with the average
sulfonate concentration, was observed, and at the UCST all chains,
with and without sulfonate groups, precipitated. That phenomenon
was attributed to aggregation of the sulfonated chains, which
presumably imbibed the unsulfonated chains at about 0.5 °C
above the UCST. In that study, however, all of the molecular
weights used were above the entanglement molecular weight of
PS, so the inclusion of the unsulfonated chains could have been
the result of entanglements.
In the present study, where the molecular weight of the
ionomers was below the entanglement molecular weight of PS,
it is not obvious why the unsulfonated polymer does not dewet
during the annealing experiments. This may, perhaps, be a
consequence of miscibility of the PS with the SPS ionomer,
although Beck Tan et al.36 reported that high molecular weight
(>105 g/mol) SPS and PS were immiscible because of the strong
repulsive interactions between the sulfonated species and
styrene.37 However, there is no reported data on the phase behavior
of PS and SPS for low molecular weight blends, where the
combinatorial entropy, which promotes mixing, may be important.
If, in fact, a single PS chain is content to be in an SPS environment,
there may be no driving force for phase separation as temperature
is increased. The combinatorial entropy term in the free energy
expression will actually decrease with increasing temperature
(note that ∆Gmix ∝ -T∆Scomb), which favors mixing, and if it
decreases faster with temperature than the increase in the enthalpy
of mixing or the equation of state terms in the free energy
expression, the miscibility would persist. To dewet, the PS chains
must first demix from the blend, which in the scenario just stated
would be unfavorable, so that even though there is a distribution
of sulfonate groups/chains in the ionomer samples, the mixture
is stable to dewetting. This, of course, is a very speculative
argument, but may provide some incentive to revisit the phase
behavior of PS/SPS mixtures.
Acknowledgment. This research was supported by the Polymer
Division of the National Science Foundation, Grant DMR
0304803.
LA802195J
(35) Chakrabarty, K.; Seery, T. A. P.; Weiss, R. A. Macromolecules 1998, 31,
7385–7389.
(36) Beck Tan, N. C.; Liu, X.; Briber, R. M.; Peiffer, D. G. Polymer 1995,
36, 1969–1973.
(37) Tucker, R. T.; Han, C. C.; Dobrynin, A. V.; Weiss, R. A. Macromolecules
2003, 36, 404–410.