Effects of Ca2+ Ion Condensation on the Molecular Structure of Polystyrene Sulfonate at Air-Water Interfaces.

The structure of poly(sodium 4-styrenesulfonate) (NaPSS) polyelectrolytes at air-water interfaces was investigated with tensiometry, ellipsometry, and vibrational sum-frequency generation (SFG) in the presence of low and high CaCl2 concentrations. In addition, we have studied the foaming behavior of 20 mM NaPSS solutions to relate the PSS molecular structure at air-water interfaces to foam properties. PSS polyelectrolytes without additional salt exhibited significant surface activity, which can be tuned further by additions of CaCl2. The hydrophobicity of the backbone due to incomplete sulfonation during synthesis is one origin, whereas the effective charge of the polyelectrolyte chain is shown to play another major role. At low salt concentrations, we propose that the polyelectrolyte is forming a layered structure. The hydrophobic parts are likely to be located directly at the interface in loops, whereas the hydrophilic parts are at low concentrations stretched out into near-interface regions in tails. Increasing the Ca2+ concentration leads to ion condensation, a collapse of the tails, and likely to Ca2+ intra- and intermolecular bridges between polyelectrolytes at the interface. The increase in both surface excess and foam stability originates from changes in the polyelectrolyte's hydrophobicity due to Ca2+ condensation onto the PSS polyanions. Consequently, charge screening at the interface is enhanced and repulsive electrostatic interactions are reduced. Furthermore, SFG spectra of O-H stretching bands reveal a decrease in intensity of the low-frequency branch when c(Ca2+) is increased whereas the high-frequency branch of O-H stretching modes persists even for 1 M CaCl2. This originates from the remaining net charge of the PSS polyanions at the air-water interface that is not fully compensated by condensation of Ca2+ ions and leads to electric-field-induced contributions to the SFG spectra of interfacial H2O. A charge reversal of the PSS net charge at the air-water interface is not observed and is consistent with bulk electrophoretic mobility measurements.

Introduction

Polyelectrolytes are well known as flocculants and stabilizers of colloidal dispersions due to electrostatic and entropic interactions, which makes them ideal candidates for stabilization of thin liquid films.[1] Due to a variety of applications, e.g., in foams,[2−4] polyelectrolytes and especially their mixtures with surfactants have been studied intensively.[5−7] Prerequisite for the formation of aqueous foams are amphiphilic molecules that tend to adsorb at the gas–water interface with sufficient surface excess. The stability of foams depends on several aspects such as hydrophobicity of the surface active molecules, concentration, ionic strength, interface charging, bulk and surface rheology, or in some cases, also the presence of aggregates.[8,9]

Sodium polystyrene sulfonate (NaPSS) is a well-known and much examined polyelectrolyte. It is not only used in industry but also in medicine because of its ion exchange properties for treatment of hyperkalemia.[10] PSS is considered a flexible polyelectrolyte with a hydrophobic backbone and charged sulfonate groups that make it partially hydrophilic.[11,12] Previous works have also reported on different polyelectrolyte conformations that depend on the solvent, ionic strength, and degree of sulfonation.[12−14] Differences of the latter are attributed to different synthesis routes.

Particularly, mixtures of polyelectrolytes and oppositely charged surfactants have recently seen great interest,[8,15−22] where it was shown, e.g., with neutron reflectivity (NR), that the mixing ratio and the mixing protocol as well as the ionic strength can play a major role in the formation of polyelectrolyte/surfactant complexes as well as aggregates in the bulk, at interfaces, and in foam films.[23] The driving forces for the latter are electrostatic, van der Waals, and hydrophobic interactions. In case of small ions, compared with molecular ions with long alkyl chains, the ion–polyelectrolyte interactions are necessarily different and the question arises if small cations have a similar effect like oppositely charged surfactants on the polyelectrolytes’ surface activity and foaming behavior. Hu et al.[24] applied vibrational sum-frequency generation (SFG) to investigate the influence of NaCl and CaCl2 on the ordering of partially hydrolyzed polyacrylamide (HPAM) polyelectrolytes at air–water interfaces and reported that the addition of divalent cations leads to highly disordered HPAM polyelectrolytes at the interface.

The partial hydrophobicity of NaPSS polyelectrolytes has been reported in several studies[11,12,25] and is most commonly explained with the hydrophobicity of the backbone, but also the structure seems to play a major role since intramolecular sulfone linkages are likely to be formed when postsulfonation of polystyrene takes place.[25] Addition of salt to the system is assumed to screen charges and to enhance counterion condensation, which lowers the effective charge and leads to a more compact conformation of the polyelectrolyte. In fact, the latter was demonstrated by electrophoretic NMR[13] and molecular dynamic simulations.[14] The extent of counterion condensation depends not only on both the Bjerrum and Debye lengths in the system but also on the length of the polyelectrolyte chain and the choice of the counterion.[13,26−28]

Many different polymers have been investigated using SFG spectroscopy;[29,30] however, there is no study that investigates the effect of ionic strength on the surface properties and structure of partially hydrophobic NaPSS. To the best of our knowledge, this is also the first study reporting about polyelectrolyte foams without the addition of any surfactant or nanoparticles. In our work, we use a unique combination of bulk specific methods such as electrophoretic mobility measurements and interfacial sensitive methods such as tensiometry, ellipsometry, and SFG spectroscopy. The combination of our results increases the understanding of structure–property relationship between the molecular structure of a polyelectrolyte at the liquid–gas interface and the properties of macroscopic foams produced from these solutions.

Materials and Methods

Sample Preparation

All glassware was cleaned with Alconox detergent solution (Sigma-Aldrich) and ultrapure water (18.2 MΩ cm, total oxidizable carbon <5 ppb) and stored in concentrated sulfuric acid (98% p.a., Carl Roth) with oxidizer NOCHROMIX (Sigma-Aldrich) for at least 12 h. Prior to usage, the glassware was rinsed thoroughly with ultrapure water to remove the acid and dried with nitrogen.

Poly(sodium 4-styrenesulfonate) (NaPSS) with an average molecular weight of 70 kDa (PDI < 1.2, batch no. BCBP3081V) and calcium chloride dihydrate (purity ≥99%) were purchased both from Sigma-Aldrich and were used as received. Stock solutions were prepared by dissolving the powders of NaPSS and CaCl2 in ultrapure water, and mixtures were prepared by adding the necessary amount of CaCl2 stock solutions to NaPSS. All NaPSS concentrations reported below refer to the total concentration of monomers. All experiments were performed at a room temperature of 297 K.

Electrophoretic Mobility Measurements

Electrophoretic mobilities were determined with a Zetasizer Nano ZSP (Malvern Panalytical, U.K.). Three measurements per concentration were performed at a scattering angle of 173°, and the results were averaged.

Surface Tension Measurements

Surface tension measurements were performed using the pendant drop method with a drop shape analyzer DSA100 (Krüss, Germany). Using a syringe pump, drops were created at the end of a cannula with a diameter of 1.83 mm. To avoid evaporation, we have measured the pendant drops in a water-saturated atmosphere inside a vessel, which had a liquid reservoir at its bottom. Using image analysis and the Young–Laplace equation, the surface tension was calculated from the drop shape as a function of time. Each concentration was measured at least three times, and the resulting values for the surface tension after 30 min were averaged. At this point, we emphasize that 30 min are not sufficient for the interface to reach an equilibrium state and thus the data represent nonequilibrium conditions. This rather short time scale in our study can be justified by its relevance for foam formation and decay, which takes place at the same time scale. After 30 min, the drops were subjected to sinoidal oscillations with a constant frequency of 0.1 Hz and a mean change in relative drop surface area of ΔA/A = 0.04 for 100 s. Through subsequent Fourier transformation, the dilatational storage modulus E′ and the dilatational loss modulus E″ were obtained and represent the elastic and viscous parts of the interfacial layer.

Ellipsometry

The thickness of the adsorbed polyelectrolyte layer at the air–water interface was determined with a phase-modulated ellipsometer (Picometer, Beaglehole Instruments, New Zealand). Prior to every measurement, the sample was equilibrated for 30 min. To achieve sufficient sensitivity to thin layers, we performed angle scans around the Brewster angle of water (53.1°) from 51 to 55°, with a step width of 0.5° at five different sample positions. Three measurements per position were performed, which yielded in total 15 measurements per concentration of which we have averaged the results. The phase-modulated ellipsometer measures the quantities x and y, which are closely related to the real and imaginary parts of the reflectivity r. Therefore, this accounts for the phase shift Δ and the ratio of amplitudes tan Ψ. More details can be found in the Supporting Information.A three-layer model was assumed for fitting the data with a refractive index of 1.33 for the subphase, 1.40 for the polyelectrolyte adsorbate layer, and 1.00 for the gas phase, which were held constant for all CaCl2 concentrations. The choice of 1.40 for the adsorbate layer is supported by the good agreement of our results for salt-free NaPSS solutions with previously reported measurements of PSS with a molecular weight of 200 kDa and 55% sulfonate groups on the chains.[31] Here, we have to state clearly that the index of the adsorbate layer depends on the conformation and composition of the layer at the interface, which is a priori unknown. Therefore, ellipsometry is not able to provide quantitative information for such thin layers; however, a comparison with results from complementary methods provides good agreement with our data and is discussed in detail below. Independent of the choice of refractive indices, we can make qualitative statements on the change in relative layer thickness with ionic strength; we expect both the refractive index and the layer thickness to change with ionic strength. However, we are limited with ellipsometry because we have to keep either the refractive index or the layer thickness constant in our analysis.

Foaming Experiments

Foaming experiments were performed with a Dynamic Foam Analyzer DFA100 (Krüss, Germany). For foam formation, air was streamed for 30 s through a porous glass frit with pore sizes of 16–40 μm at a volumetric flow rate of 5 mL/s. The foam and the liquid heights as a function of time were recorded by determining the intensity of transmitted light through the column using two columns of photodiodes and light-emitting diodes (LEDs) at opposite sides of the foam column. The foam stability at a certain time is defined as the foam height at that time divided by the initial foam height before the foam starts to collapse. For each concentration, at least three foams were analyzed and the results were averaged. Additionally, images of the foam were taken to resolve structure and bubble size distribution. For the latter, we have used a foam structure module attached to the DFA100, which consists of a glass column with a prism, an LED, and a charge-coupled device camera to analyze a cross section of up to 285 mm2.

Sum-Frequency Generation (SFG) Spectroscopy

SFG spectroscopy is an interfacial sensitive nonlinear optical method that is useful for studying the structure of molecules at an interface with respect to composition, coverage, orientation, and surface charge. For SFG spectroscopy, two laser beams are overlapped spatially and temporally at the interface of interest. A picosecond visible pulse with a narrow line shape and frequency fixed to ωvis and a second broadband femtosecond infrared (IR) pulse that is tunable and has the frequency ωIR serve as laser beams. In a second-order nonlinear optical process that takes place where both beams overlap, a third beam with the sum frequencyis generated. The SFG intensity depends on the intensity of the two impinging laser beams Ivis and IIR as well as on the resonant and nonresonant second-order electric susceptibilities χR(2) and χNR(2). The nonresonant contribution is caused mainly by electronic excitations at the interface, whereas the resonant contribution arises from molecular vibrations. In the case of charged surfaces, an additional static electric field leads to a third-order contribution χ(3) due to the polarization and polar ordering of water molecules near and at the charged surface.[32−35] The Debye length κ–1 and the surface potential ϕ0 are decisive factors for the magnitude of this contribution.The resonant part depends on the resonance frequency ω, the bandwidth Γ of the vibrational mode k, and the oscillator strength A ∝ N⟨αμ⟩, which is determined by the molecular number density N and the orientational average of the Raman polarizability α and the dynamic dipole moment μ:In the case of centrosymmetric materials or isotropic bulk liquids, this orientational average and consequently the oscillator strength equals zero. However, the symmetry break at interfaces gives rise to nonzero contributions to χ(2) that originate purely from interfacial molecules. In this case, vibrational SFG spectroscopy is inherently interface specific.

SFG measurements were performed with a home-built broadband sum-frequency generation spectrometer that is described in detail in the Supporting Information. The basic principle is the temporal and spatial overlap of a femtosecond infrared beam (full width at half-maximum (FWHM) bandwidth: 300 cm–1) and an etalon-filtered visible beam with 804.1 nm wavelength (FWHM bandwidth: 4 cm–1) at the air–liquid interface at incidence angles of 60 and 54°, respectively. The IR wavelength was varied in the frequency range of 2800–3800 cm–1 in five steps and in the frequency range of 950–1250 cm–1 in three steps to cover the whole frequency range. The acquisition times for measurement of C–H and O–H stretching vibrations (2800–3800 cm–1) ranged from 15 to 120 s per IR wavelength and for measurement of S–O stretching vibrations (950–1250 cm–1) from 10 to 15 s depending on the SF signal strength. All measurements were referenced to the nonresonant SFG signal of a polycrystalline and air plasma-cleaned gold film. The polarizations used for investigations of air–water interfaces were ssp (SF/vis/IR), whereas ppp polarizations were used for the reference spectrum from a clean Au surface. Before SFG spectra were taken, the samples were equilibrated for 30 min.

Experimental Results

Electrophoretic Mobility

We have determined the electrophoretic mobility uζ of 20 mM NaPSS in aqueous solutions for different bulk concentrations of Ca2+ and present the results in Figure . Figure shows that Ca2+ concentration as low as 0.2 mM can already change the mobility to some extent. However, analysis of the data at low concentrations is difficult due to the relatively high experimental scatter. Concentrations higher than 1 mM lead to a distinct increase in the mobility with a significantly reduced scatter in our data. We attribute this change in mobility to the association of Ca2+ cations to the negatively charged sulfonate groups of the PSS polyanions, which reduces the negative net charge of the PSS chain. Even at high salt concentrations such as 1 M, not all sulfonate groups are modified by Ca2+, which is indicated by a residual uζ of about −0.3  ×  10–8 m2/( V s) at 1 M CaCl2. This observation is consistent with an incomplete screening of the negative charges at the PSS polyelectrolyte chain.

Figure 1

Electrophoretic mobility uζ of 20 mM NaPSS solutions as a function of CaCl2 concentration. The solid line is a guide to the eye.

Tensiometry

Since the adsorption of PSS polyelectrolytes at the air–water interface can take up to several days, the results presented in Figure a show surface tensions 30 min after the surface was created. Addition of 20 mM NaPSS already lowers the surface tension of water to a value of 63 mN/m (Figure a), whereas adding CaCl2 leads to a further change in surface tension. Compared to the surface tension of NaPSS solutions without additional salt, the surface tension with salt first increases weakly with increasing salt concentration and reaches a local maximum of 65 mN/m at 1 mM CaCl2. Since this small increase lies within the experimental error, the interpretation of this maximum has to be treated with caution. Note, however, that the concentration region around 1 mM CaCl2 will be of special interest in the discussions below, where we address the role of the PSS molecular structure on the properties of air–water interfaces and aqueous foam. CaCl2 concentrations >1 mM cause a steep decrease of the surface tension from 65 to 50 mN/m at 1 M CaCl2.

Figure 2

(a) Surface tension γ of 20 mM NaPSS solutions as a function of Ca2+ concentration (■) as well as γ for CaCl2 electrolytes without polyelectrolyte additions (○). (b) Thickness h of the interfacial layer as determined from ellipsometry (see also experimental details). (c) Surface dilatational storage and loss modulus E′ and E″ and (d) foam stability (FS) after 90 s of 20 mM NaPSS solutions as a function of Ca2+ concentration. Solid lines are a guide to the eye.

Such low surface tensions can be associated with a high surface excess of PSS polyelectrolytes at the air–water interface. Because the measured surface tensions represent a nonequilibrium state, the kinetic barriers for adsorption might change with concentration, which has to be kept in mind when associating low surface tension with a high surface excess. Nevertheless, further justifications for our conclusion can be found below.

Figure b presents the results from ellipsometry at the air–water interface, which are consistent with our observations from surface tension measurements. Without salt, we determine a thickness of the NaPSS adsorbate layer of around 1.8 nm. Although there is only little literature about the thickness of partially hydrophobic polyelectrolyte layers at air–water interfaces available, the measured thickness is consistent with earlier reports by Théodoly et al.[31] using ellipsometry and X-ray reflectivity to study randomly sulfonated polystyrene (50 mM) with charge fractions between 0.3 and 0.9. Théodoly et al. reported layer thicknesses <3 nm.[31]

At concentrations <1 mM where the surface tension shows little change, the layer thickness is quite constant (1.6–1.8 nm), whereas a further increase in salt concentration leads to thicker layers of ∼3.7, 7.1, and 9.1 nm for CaCl2 concentrations of 0.01, 0.1, and 1 M, respectively.

Surface Dilatational Rheology

The elastic and viscous properties of an interfacial layer are important factors for foam properties.[36]Figure c shows the dilatational storage E′ and loss modulus E″ as a function of calcium concentrations. From an inspection of Figure c, it can be clearly seen that the elastic portion E′ (∼30–50 mN/m) is for all salt concentrations significantly higher than the viscous portion E″ (∼2–13 mN/m). For that reason, the viscoelastic properties of the interfacial layer under dilatational surface area changes are dominated by the elastic component E′ of the complex modulus E = E′ + iE″. Analogous to the complementary results in Figure a,b, only small changes in E′ and E″ are observed at CaCl2 concentrations <1 mM. For concentrations >1 mM, both increase and reach a local maximum at ∼4 mM CaCl2 that is followed by a plateau at the highest concentrations that we have investigated.

Foam Characterization

The foam stability after 90 s is shown in Figure d as a function of CaCl2 concentration. It is obvious that addition of Ca2+ leads first to destabilization of the foams and to a shallow minimum in foam stability at CaCl2 concentrations of 1–2 mM. This is in good agreement with the low E′ and E″ of the interfacial layers at this concentration regime (Figure c). CaCl2 concentrations >5 mM lead to more stable NaPSS foams with a stability of 40% after 90 s. The highest stability of 60% is observed for the highest salt concentration.

Figure shows cross sections of the foams, which were taken at different calcium concentrations and at foam ages of 60 and 180 s. Particularly noteworthy is the foam structure at low Ca2+ concentration: As can be seen very well for 0.5 mM CaCl2, a bimodal foam is observed directly after foaming. The bigger bubbles are growing fast and collapse once they reach a critical size and are thus responsible for the poor foam stability.

Figure 3

Structure of foams from 20 mM NaPSS solutions for different Ca2+ concentrations and foam ages of 60 and 180 s. Columns indicate liquid (dark blue) and foam (light blue) heights. The scale bar is for all images, as shown in the left hand side of the second row.

SFG Spectroscopy

The recorded SFG spectra from PSS-modified air–water interfaces are shown in Figure . In the fingerprint region from 950 to 1250 cm–1 (Figure a), we observe three distinct vibrational modes: The bands centered at 1011 and 1134 cm–1 can be assigned to the in-plane bending and in-plane skeleton vibrations of the aromatic ring, whereas the band centered at 1042 cm–1 is attributed to S–O symmetric stretching vibrations of the sulfonate group.[8,37] In the second and third frequency region, which are presented in Figure b,c, we observe C–H stretching vibrations (2800–3070 cm–1) and O–H stretching bands (3000–3800 cm–1). Specifically, the band at 2850 cm–1 is caused by CH2 symmetric stretching vibrations, whereas CH3 symmetric stretching vibrations give rise to the vibrational band at 2878 cm–1. Additional vibrational bands are centered at 2920, 2946, 2975, and 3065 cm–1 and can be attributed to CH2 antisymmetric stretching vibrations, the CH3 Fermi resonance, antisymmetric CH3 stretching, and aromatic C–H stretching vibrations, respectively.[38−40] The broad bands centered around 3200 and 3450 cm–1 are due to O–H stretching vibrations of hydrogen-bonded interfacial water molecules. These bands are often termed as icelike and liquidlike water bands originating from tetrahedrally and asymmetrically coordinated water molecules.[41]

Figure 4

SFG spectra of the liquid–air interface using 20 mM NaPSS solutions with different CaCl2 concentrations, as indicated (in mM) in the figure (concentrations in (a), (b), and (c) are identical). The SFG spectra are stacked with a constant offset for better comparison of the spectral changes: (a) shows the fingerprint region with vibrational bands at 1011 and 1134 cm–1 due to the in-plane bending (1011 cm–1) and skeleton vibrations (1134 cm–1) of polyelectrolyte’s aromatic ring as well as S–O symmetric stretching vibrations (1042 cm–1); (b) shows the frequency region where C–H stretching vibrations are observed; and (c) shows the region of aromatic C–H and O–H stretching (for details, see the main text).

Figure clearly shows that even at low and zero salt concentrations, strong bands are observed in the fingerprint region. An increase in the intensity of all three bands of the fingerprint region is observed when the salt concentration is increased from 0 to 5 mM. This increase continues up to 20 mM CaCl2, while higher concentrations result in constant intensities of all three bands. To perform a more quantitative analysis of our SFG spectra, fitting of the spectra was done using model functions according to eqs and 4.

From the fits to the spectra, we obtain the SFG amplitude of the contributing vibrational bands. At this point, we recall that the latter is directly dependent on the surface coverage of the respective molecular moiety and the orientational average of its hyperpolarizability. In Figure a, the SF amplitudes are shown as a function of CaCl2 concentration. Figure a clearly indicates a strong increase in SF amplitude of all three bands in the fingerprint region until a plateau value is reached at ∼20 mM.

Figure 5

(a) SFG amplitudes of vibrational bands in the fingerprint region (Figure a) and (b) SFG amplitudes of O–H stretching vibrations. Solid lines are a guide to the eye.

In addition to the latter change of the bands in the fingerprint region, we find that C–H and O–H stretching bands in Figure b do change both in shape and in intensity with increasing CaCl2 concentration. In particular, the low-frequency branch of the O–H stretching bands at 3200 cm–1 decreases to negligible values at high salt concentrations, whereas the high-frequency part at 3450 cm–1 persists even for the highest studied concentration of 1 M CaCl2.

The shape of the C–H stretching bands does also change dramatically and can be partially explained by weaker interference (according to eqs and 3) with the low-frequency branch of the O–H stretching bands when the latter gets reduced to negligible values at high CaCl2 concentrations.

At this point, we recall that the SFG intensity of O–H stretching vibrations is not only a function of the interfacial water structure that gives rise to second-order contributions χ(2) but in the presence of a static electric field at the interface is also dependent on the surface potential ϕ0 via a third-order contribution χ(3) to the electric susceptibility (eq ). Because the lowest ionic strength in our system is with 20 mM much higher than the concentrations where possible interference effects could modulate our SF signals and impair their analysis in terms of relative change in surface potentials or in surface net charge, we can neglect interference effects in our analysis. For further information where such effects can be important, we refer to the previous works by Gonella et al.[32] and Ohno et al.[33,42]

From our analysis of the polyelectrolytes in the bulk solution, we have already seen that the absolute value of bulk electrophoretic mobility decreases with CaCl2 concentration. For that reason, the decrease in O–H intensity is consistent with the observations in the bulk and can be rationalized by a reduction of the polyanion’s net charge that is caused by binding of Ca2+ ions to the polyelectrolyte. This reduces the net charge on the polyelectrolyte as well as the net charge at the interface. Similar behavior was previously reported for mixtures of NaPSS with CTAB cationic surfactants: for air–water interfaces that were modified by PSS–/CTA+ complexes, the SFG intensity of water bands was negligible for the mixing ratio where also the electrophoretic mobility in the bulk was close to zero.[8]

We now return to our earlier observation that for our system (NaPSS polyelectrolytes in CaCl2 solutions), the high-frequency part of the water bands at 3450 cm–1 is persistent even at 1 M CaCl2 (Figure c). This indicates that for solutions with 1 M CaCl2, there are still sufficiently ordered but rather asymmetrically coordinated water molecules at the interface. The latter give rise to the observed strong O–H contributions at high ionic strengths (Figure c).

The intensity of CH3 vibrations at 2878 (symm. stretch) and 2946 cm–1 (Fermi resonance) is decreasing when small amounts of salt are added. When the salt concentration is increasing, however, an increase in CH2 vibrations is notable. Since we have observed a change of surface tension at Ca2+ concentrations slightly above 1 mM along with an increase in the layer thickness and a distinct minimum in foam stability (Figure ), the above described change in C–H intensities brings some evidence to a structural change of the polyelectrolyte. Therefore, SFG measurements in this concentration range have been additionally performed where we have concentrated specifically on the spectral region of C–H stretching bands (2800–3150 cm–1). To be able to evaluate the intensity of C–H vibrations without the interfering broad bands from O–H stretching vibrations, we have made use of the properties of the time-asymmetric visible pulse in our SFG spectrometer.[43,44] The short-lived O–H vibrations outlast less than 100 fs,[45] whereas the dephasing and life times of C–H vibrations are typically in the order of several picoseconds.

This means that we can suppress the water bands by inducing an appropriate time delay of the visible pulse. Similar to previous works by Dlott and co-workers,[43,46] who used this method to suppress nonresonant contributions to the second-order susceptibility, we could suppress the broad water bands using an IR–vis time delay of 300 fs. As a result, we cannot compare the intensity of C–H vibrations to the intensity of those with zero time delay in Figure but we can compare the changes in intensity and spectral shape of C–H stretching bands for different CaCl2 concentrations without the strong interference with the low-frequency branch of O–H stretching vibrations (Figure ). From a close inspection of SFG spectra with suppressed water bands, it is obvious from Figure that the SFG bands from CH2 and CH3 bands are less intense at low CaCl2 concentrations between 1 and 1.5 mM, whereas the aromatic C–H band (3060 cm–1) increases in intensity. However, for 2.5 mM CaCl2, the trend observed for smaller concentrations is reversed and a distinct increase in CH2 as well as CH3 contributions and a decrease in the aromatic C–H contributions is found. Since the CH2 bands represent the backbone and the aromatic bands the side group of NaPSS, these trends are helpful to draw conclusions about the interfacial conformation of the polyelectrolyte at different ion concentrations, which we will elaborate further in the discussion.

Figure 6

SFG spectra taken at the liquid–gas interface from 20 mM NaPSS solutions with different Ca2+ concentrations. CaCl2 concentrations in mM were as indicated in the figure. Red lines represent the experimental results, whereas black solid lines are fits to the experimental data.

Discussion

Structure of PSS-Modified Air–Water Interface for Ca2+ Concentration ≫1 mM

Below, we will address structure–property relations between liquid–gas interfaces and foams that are modified by polystyrene sulfonate polyelectrolytes as a function of bulk Ca2+ concentration.

In Figure , we have demonstrated that the surface tension of NaPSS solutions in contact with air is decreasing while simultaneously, the measured layer thickness is increasing. Taking into account that the surface tension is closely related to the surface excess of adsorbed molecules, the measurements indicate that the amount of adsorbed PSS and also the measured thickness increases significantly with increasing ionic strength. Although care must be taken when considering absolute values of the layer thickness from ellipsometry (see above), this conclusion is still possible for a qualitative treatment of the data.

The discussed increase in surface excess is not surprising since the effective charge and thus the hydrophobicity of PSS is a function of the ionic strength in the system. With increasing ionic strength, more counterions can be associated to the polyelectrolyte and decrease its net charge (Figure ). Using a combination of pulsed-field gradient NMR and electrophoresis NMR,[13,26,47] Scheler and co-workers have shown experimentally that ion condensation results in weaker repulsive electrostatic interactions along the polyelectrolyte chain and thus in a more compact conformation. The decrease in net charge as well as in the hydrodynamic radius of polyelectrolyte chains is also predicted by calculations from Kundagrami et al.[28] using an adsorption model and by Fahrenberger et al.[27] using MD simulations. In addition to the effects of monovalent ions such as Na+, even more pronounced effects from divalent counterions have been suggested.[26,48] Similar ion-specific effects were also observed for sodium polyacrylate polyelectrolytes with small-angle X-ray scattering and small-angle neutron scattering measurements.[49−51]

The increased hydrophobicity of the polyelectrolyte at higher salt concentrations leads to a higher surface excess of NaPSS, which is consistent with the increased thickness of the adsorbate layers and lower surface tensions (Figure ). Similar effects with increasing salt concentration are previously reported for different polyelectrolytes.[52] In our study, we have investigated a mixture of 20 mM Na+ and Ca2+ ions; however, considering the study by Kundagrami et al.,[28] it is likely that the Na+ counterions are being replaced completely by the divalent calcium counterions. This assumption is also corroborated by our experimental results, which we discuss below. SFG spectra in Figure a show a substantial increase of the sulfonate and aromatic bands, which is a direct cause of the above discussed counterion condensation. The latter renders the PSS polyelectrolyte more hydrophobic and thus leads to an increase of the polyelectrolytes’ surface excess and confirms our earlier conclusion. At low CaCl2 concentrations ≪1 mM, we propose that mostly hydrophobic parts of the backbone due to intramolecular sulfone linkages and resulting loops from the synthesis route (see the Supporting Information) are present at the interface. In this case, the remaining charged sulfonated chains and thus the aromatic parts of the molecule are likely to be spread from the interface into regions near the interface like tails, as described by Noskov et al.[53] A similar organization of polymer layers at the air–water interface is well known for block copolymers where the hydrophobic parts collapse at the interface and form a “carpet layer”, whereas the more hydrophilic parts stick into liquid phase and form a brush-like layer.[54,55] When the salt concentration is increased, Ca2+ ion condensation occurs at the charged sulfonate groups of PSS and causes a local reduction of the polyelectrolytes’ net charge and consequently a collapse of the tails. This is because these sites become more hydrophobic and thus also more eager to enter the air–water interface. For this reason, the whole backbone is brought to the air–water interface when the tails collapse and the previous layer, consisting of hydrophobic loops, increases in thickness. These conclusions are consistent with our observation that the thickness of the interfacial layer (Figure ) and the CH2 contributions to the SFG spectra (Figure b) increase. In addition to this change in hydrophilicity, specific ion effects have to be considered. Hu et al.[24] observed in their SFG study of partially hydrolyzed polyacrylamide at the air–water interface a similar effect of Ca2+ ions. They concluded that inter- and intramolecular interactions between carboxylate groups were induced by the presence of Ca2+ ions and lead to a change in the polyelectrolyte’s conformation as well as to a formation of polymer networks or aggregates at the interface. The latter caused a more disordered structure of water molecules at the interface.

Here, we propose that similar effects of calcium ions on PSS polyelectrolytes are also possible and Ca2+ ions are likely to cause inter- and intramolecular bridges between the polyelectrolyte’s sulfonate groups at the air–water interface. One unique feature of the O–H stretching bands from SFG spectroscopy is the persistence of the water band at 3450 cm–1 even for extreme CaCl2 concentrations of 1 M. It is generally assumed that the band at 3450 cm–1 is attributable to asymmetric coordinated water molecules whereas the branch at lower frequencies is attributable to more tetrahedrally coordinated water molecules.[24,42] From the existence of strong O–H contributions, we conclude that even at the highest salt concentrations, some interfacial water molecules are polar ordered. Assuming that the surface is covered by a slightly negatively charged layer of NaPSS polyelectrolytes, interfacial water molecules can be locally polarized and electric-field-induced contributions can be induced by the remaining (net) charge of the polyelectrolyte. In fact, our electrophoretic mobility measurements (Figure a) corroborate the latter assumption on the remaining net charge because small but negative uζ are found at high ionic strength and overcharging in the bulk solution as well as at the interface can be ruled out for the investigated concentrations.

In addition to the increased hydrophobicity and surface excess that is accompanied by a reduced net charge at PSS-modified air–water interfaces, we observe an increase in surface dilatational elasticity E′ to a relatively high value of 40 mN/m. This increase is a direct consequence of the above discussed reduction of net charges that drives the lateral interactions at the interface in a more attractive regime. Furthermore, the formation of Ca2+ bridges (as discussed above) between and within the collapsed polyelectrolyte chains at the interfaces could be an additional origin of the observed higher E′. Interfacial PSS layers with higher E′ and surface excess can modify the foam stability on macroscopic length scales through structure–property relations favorably, an effect that has been previously reported for other even complementary systems.[8,9]

Structure of a PSS-Modified Air–Water Interface for Ca2+ Concentration of ∼1 mM

To address the effects at low CaCl2 concentrations, we now discuss the origin of the surface activity of NaPSS without added salt. All methods applied within this study show a significant surface activity of PSS, which is consistent with previous reports.[12,31,56] Although the increase of surface excess with high salt concentration can be explained as discussed above, the changes we see at relatively low Ca2+ concentrations (∼1.5 mM) cannot be rationalized in the same way. At this point, we want to emphasize that for tensiometry and ellipsometry, quite small changes that lie within the experimental errors are being interpreted since they support our findings from other methods. However, the concentration region around 1 mM clearly marks the onset of drastic changes in the conformation of NaPSS (see discussion for concentrations ≫1 mM). Additions of 0.5–1 mM CaCl2 lead to a slight increase of surface tension that is accompanied by a small but noticeable decrease in layer thickness compared with the salt-free 20 mM NaPSS solutions and a minimum in surface dilatational elasticity E′ (Figure ). These results are consistent with a reduction of the polyelectrolyte’s surface excess that is confirmed by the increase in surface tension and a decrease in layer thickness for concentrations around 1 mM (Figure ). Furthermore, foaming experiments and SFG spectroscopy (Figure ) show that the amplitudes of CH2 and CH3 vibrational bands decrease and reach a minimum (Figure ). This minimum is accompanied by a minimum in foam stability at 0.5–2.5 mM Ca2+ concentrations (Figure d), and we conclude that a structural change, different from what we have discussed in the previous sections, takes place already at low concentration. This structure change at the interface renders PSS more hydrophilic until at higher CaCl2 concentrations >5 mM, the effects described in the previous section start to dominate both bulk and interfacial properties. From the latter observations and the bimodal structure of foams in this concentration region, we hypothesize that at ∼1 mM CaCl2, two fractions of PSS polyelectrolytes with different conformations can coexist and possess different surface activities. Some support for our conclusion concerning a substantial change in the structure of PSS polyanions within a very narrow concentration range comes from reports by Böhme et al.[26] as well as Drifford et al.[57] who have shown that the bulk diffusion coefficient and thus the hydrodynamic radius change dramatically at CaCl2 concentrations of ∼1 mM. In fact, Böhme et al.[26] reported a reduction in the hydrodynamic radius from 14 to 6 nm of PSS polyelectrolytes (77 kDa) by crossing 1 mM CaCl2 from low to high concentrations.

Conclusions

In this study, we have investigated the effects of Ca2+ ions on the structure of polystyrene sulfonate at the air–water interface, which can cause different foam properties through structure–property relations. Compared to CaCl2-free solutions, we observe at concentrations c(Ca2+) <5 mM a weak but noticeable decrease in the PSS surface excess when c(Ca2+) increases. This decrease is accompanied by a reduced surface dilatational elasticity and foam stability. In addition, SFG spectra of C–H stretching bands as well as the foam structure change substantially at Ca2+ concentrations of ∼1 mM and are caused by modifications in the PSS molecular structure at the air–water interface. Increasing the Ca2+ concentration well above 5 mM leads to an increase of the PSS surface excess and the formation of a layer with loops at the interface and tails stretching into the bulk. The latter collapses to a dense closed-packed layer at the highest CaCl2 concentration of 1 M, which results in an overall interaction potential at the interface that is more attractive compared to that of salt-free solutions and gives rise to a substantial increase in the surface dilatational elasticity and foam stability. SFG spectra of O–H stretching bands show a decrease in intensity of the low-frequency branch that is attributable to tetrahedrally coordinated water molecules when the salt concentration is increased. However, the high-frequency branch of O–H stretching modes due to asymmetrically bonded interfacial H2O persists even for 1 M CaCl2. This persistence originates from the remaining net charge of the PSS polyanions at the air–water interface that is not fully compensated by condensation of Ca2+ ions even at 1 M and leads to electric-field-induced contributions to the SFG spectra of interfacial H2O. For this reason, overcharging of the polyelectrolyte at the air–water interface is not observed in our experiments, which is in good accordance with bulk electrophoretic mobility measurements.