Adsorption of Sodium Dodecyl Sulfate at the Hydrophobic Solid

Department of Chemical and Biological Engineering, University of Maine, Orono, Maine ... Adam J. Hopkins , Cathryn L. McFearin , and Geraldine L. Rich...
1 downloads 0 Views 580KB Size
Download PDF
Langmuir 2006, 22, 3105-3111

3105

Adsorption of Sodium Dodecyl Sulfate at the Hydrophobic Solid/ Aqueous Solution Interface in the Presence of Poly(ethylene glycol): Dependence upon Polymer Molecular Weight Michael T. L. Casford and Paul B. Davies* Department of Chemistry, UniVersity of Cambridge, Lensfield Rd., Cambridge, CB2 1EW, U.K.

David J. Neivandt Department of Chemical and Biological Engineering, UniVersity of Maine, Orono, Maine 04469 ReceiVed August 21, 2005. In Final Form: January 27, 2006 The polar orientation and degree of conformational order of sodium dodecyl sulfate (SDS) adsorbed at the hydrophobic octadecanethiol/aqueous solution interface in the presence of poly(ethylene glycol) (PEG) has been investigated as a function of the surfactant concentration and the molecular weight of the polymer. Sum frequency generation (SFG) vibrational spectroscopy was employed to obtain spectra of interfacial surfactant; weak SFG signals from interfacial polymer were also detected for polymer molecular weights of 900 and above. The phase of the SFG spectra indicated that both the surfactant and polymer had a net orientation of their CH2 and/or CH3 groups toward the hydrophobic surface. Spectra of SDS in the presence of mixed polymer/surfactant solutions showed increasing conformational order as the surfactant concentration was raised. At the lowest surfactant concentrations, the spectra of SDS were weaker in the presence of the polymer than in its absence. All PEG molecular weights investigated, with the exception of PEG 400, gave rise to significant inhibition of ordered surfactant adsorption below the critical micelle concentration. The greatest inhibitory effect was noted for PEG 900. Probing interfacial PEG specifically through the use of perdeuterated SDS revealed that the polymer spectral intensity decreased monotonically as the surfactant concentration was increased for all polymer molecular weights where a PEG spectrum was apparent. These findings are interpreted in terms of the displacement of preadsorbed polymer as the surfactant concentration increases. This result is compatible with observations of adsorption from SDS/PEG solutions at solid/solution and solution/air interfaces made using other techniques.

Introduction Most of the commercial interest in low-molecular-weight poly(ethylene glycol) (PEG) has been directed toward covalent surface modification for biomedical applications and has revolved around the capacity of PEG to resist protein adsorption in biological environments and its low biological toxicity.1-7 An alternative surface application is to physisorb PEG from solution, a method used for example to deliver PEG in synthetic lung surfactant formulations.8 The majority of related research has concentrated on the macroscopic biological aspects of PEG adsorption behavior onto model interfaces, typically from multicomponent solutions containing for example surfactants, lipids, and proteins. Little work has been performed to date on the molecular conformation of PEG upon adsorption, principally due to the lack of suitable techniques that are capable of elucidating surface structural information in situ. With the advent of nonlinear optical spectroscopy and single-molecule fluorescence techniques, this situation is now changing.9-11 * To whom correspondence should be addressed.. Tel: +44 1223 336460. Fax: +44 1223 336362. E-mail: [email protected]. (1) McGall, S. J.; Davies, P. B.; Neivandt, D. J. J. Phys. Chem. B 2003, 107, 4718. (2) McGall, S. J. Ph.D. Thesis, University of Cambridge, 2004. (3) Morra, M. J. Biomater. Sci.-Polym. Ed. 2000, 11, 547. (4) Greenwald, R. B.; Conover, C. D.; Choe, Y. H. Crit. ReV. Ther. Drug Carr. Syst. 2000, 17, 101. (5) Nelson, K. D.; Eisenbaumer, R.; Pomerantz, M.; Eberhart, R. C. Asaio J. 1996, 42, M884. (6) Sheu, M. S.; Hoffman, A. S.; Ratner, B. D.; Feijen, J.; Harris, J. M. J. Adhes. Sci. Technol. 1993, 7, 1065. (7) Braatz, J. A.; Heifetz, A. H.; Kehr, C. L. J. Biomater. Sci.-Polym. Ed. 1992, 3, 451. (8) Reddy, K. R. Ann. Pharmacother. 2000, 34, 915.

The properties of higher-molecular-weight PEG, often referred to as PEO, in solution and upon adsorption are dominated by the chemistry of the ethylene oxide repeat unit and entropic considerations of long polymer chains in solution. As the PEG molecular weight decreases, the effect of the chain-terminating groups become increasingly important both for behavior in solution and in the adsorption/desorption of PEG. This is coupled with a reduction in the entropic penalty incurred through solvation of the polymer for low-molecular-weight PEGs. The principle factor determining the solubility and adsorption characteristics for different PEG molecular weights is generally considered to be the change in the latter. A common component in PEG formulations is the surfactant sodium dodecyl sulfate (SDS). Although relatively weak, the attractive interaction between the anionic surfactant SDS and the neutral polymer PEG has been very extensively studied.12,13 The interaction of the surfactant with the polymer is driven by the ion-dipole interaction between the anionic headgroup of the surfactant and the polar oxygen group of the polymer, further facilitated by the flexibility of the polymer chain allowing interactions between the surfactant hydrocarbon chain and the polymer methylene moiety14 which reduces the entropic penalty incurred through disruption of the water hydrogen-bonding (9) Zhao, J.; Granick, S. J. Am. Chem. Soc. 2004, 126, 6242. (10) Zolk, M.; Eisert, F.; Pipper, J.; Herrwerth, S.; Eck, W.; Buck, M.; Grunze, M. Langmuir 2000, 16, 5849. (11) Jayachandran, K. N.; Maiti, S.; Chatterji, P. R. Polymer 2001, 42, 6113. (12) Cabane, B. J. Phys. Chem. 1977, 81, 9. (13) Shirahama, K. Colloid Polym. Sci. 1974, 252, 978. (14) Ghoreishi, S. M.; Li, Y.; Bloor, D. M.; Warr, J.; Wyn-Jones, E. Langmuir 1999, 15, 4380.

10.1021/la052271z CCC: $33.50 © 2006 American Chemical Society Published on Web 03/03/2006

3106 Langmuir, Vol. 22, No. 7, 2006

network upon solvation of the hydrocarbon groups. The interaction begins at the critical aggregation concentration (cac ) 5.3 ( 0.2 mM at 298.1 K15), which is well below the critical micellar concentration (cmc) for pure SDS solutions, namely 8.3 mM at 298 K,16 and increases rapidly due to cooperative complexation. Above a threshold concentration (c2) corresponding to SDS saturation of the polymer, the formation of SDS micelles in solution commences. The commonly accepted structure of the PEG/SDS complex, at least for molecular weights of 1000 and above, is that of micellar-like surfactant aggregates distributed along the polymer backbone, the so-called necklace model of polymer/surfactant complexation. In the presence of electrolyte, the cac drops markedly, for example, from 4.2 ( 0.1 to 0.8 mM for 0.1% w/v PEO (Mw ) 8000) in 0.1 M NaCl, while c2 hardly changes (18 ( 1 to 21 ( 2 mM for the same system).17 A significant amount of research has been carried out on the interaction of SDS with PEG in solution. The choice of molecular weights investigated in the current work was determined in part by the conductivity measurements of Minatte et al. and Jones et al.,17,18 both of which indicated the presence of three separate regions in the SDS/PEG conductivity behavior, suggesting that two types of polymer/surfactant complex may be forming as the surfactant concentration increases. Viscosity measurement of the same molecular weight polymer solutions also led to similar conclusions.19-21 Both conductivity and viscosity studies are in satisfactory agreement with the hypothesis that the chain length for PEG oligomers with a molecular weight below 1000 is insufficient to completely wrap around a surfactant micelle in solution and furthermore would be insufficiently long to allow more than one micelle to be associated with any given oligomer strand. Conversely, multiple micelles per polymer strand are believed to occur for polymer weights above 4000. Above this weight, both the conductivity and the viscosity plots against surfactant concentration are observed to be linear. More recently, isothermal titration measurements carried out by Dai and Tam22 and by Bernazzani et al.23 confirmed this hypothesis, although the precise nature of the polymer surfactant aggregate is still under debate. Additionally, they recorded a deviation from linearity in the plot of surfactant aggregation number against PEG molecular weight, which they interpreted as a stepwise addition of surfactant micelles to a specific polymer strand as its length increased. Plateau regions corresponded to situations in which increased polymer weight (up to 8000) gave rise to no further increase in the number of surfactant molecules aggregated on the polymer strand. This article describes the adsorption from the binary polymer/ surfactant solution onto a hydrophobic surface of the anionic surfactant SDS and the neutral polymer PEG of different molecular weights. It extends earlier work24 on the interaction between SDS and PEG 12 000, which indicated the displacement of a preadsorbed polymer layer by the surfactant and the formation of an effectively complete SDS monolayer by one-tenth of the (15) Gjerde, M. I.; Nerdal, W.; Hoiland, H. J. Colloid Interface Sci. 1998, 197, 191. (16) Witte, F. M.; Buwalda, P. L.; Engberts, J. Colloid Polym. Sci. 1987, 265, 42. (17) Minatti, E.; Zanette, D. Colloids Surf., A 1996, 113, 237. (18) Jones, M. J. J. Colloid Interface Sci. 1967, 23, 36. (19) Cai, J. L.; Bo, S. Q.; Cheng, R. S.; Jiang, L. S.; Yang, Y. J. Colloid Interface Sci. 2004, 276, 174. (20) Chari, K.; Antalek, B.; Lin, M. Y.; Sinha, S. K. J. Chem. Phys. 1994, 100, 5294. (21) Chari, K.; Kowalczyk, J.; Lal, J. J. Phys. Chem. B 2004, 108, 2857. (22) Dai, S.; Tam, K. C. J. Phys. Chem. B 2001, 105, 10759. (23) Bernazzani, L.; Borsacchi, S.; Catalano, D.; Gianni, P.; Mollica, V.; Vitelli, M.; Asaro, F.; Feruglio, L. J. Phys. Chem. B 2004, 108, 8960. (24) Casford, M. T. L.; Neivandt, D. J.; Davies, P. B. Langmuir 2003, 19, 7386.

Casford et al.

pure surfactant cmc. Particular emphasis has been given in the current study to the lower-molecular-weight oligomers, namely those of Mw ) 400, 600, 900, and 1500 where significant differences have been observed between them in their solution phase behavior, compared with higher-weight polymers. Experimental Considerations Sum frequency generation (SFG) vibrational spectra were recorded on the Cambridge nanosecond laser spectrometer which has been described in detail elsewhere.24-26 A counterpropagating beam geometry was employed (60° and 65° to the surface normal for the visible and infrared beams, respectively); all spectra presented here have been recorded in the ppp (sum frequency, visible, infrared) laser beam polarization combination in order to take advantage of the nonresonant signal enhancement arising from the hydrophobic gold substrate. Spectral searches have also been made using the ssp beam polarization combination and polymer solutions containing various polymer molecular weights. No SFG signals were observed. This is believed to reflect the reduced sensitivity of the SFG spectrometer when using this polarization combination. The absolute concentration detection limit, however, cannot be readily determined mainly due to the dependence of the SFG signal intensity on the orientation of the SFG active group with respect to the surface normal. Typically, individual ppp spectra were co-added for a period of 6 h in order to obtain a final spectrum. Spectra were normalized to that of the gold background in pure water and analyzed by fitting with Lorentzian line profiles using a Levenberg-Marquardt leastsquares fitting routine.27 Samples were mounted in a stainless steel flow cell, and a thin layer of the required solution trapped between the substrate and a CaF2 prism, the latter allowing optical access to the interface. The substrates consisted of 200 nm gold layers thermally evaporated onto chromium-primed silicon wafers. The gold surfaces were subsequently rendered hydrophobic by immersion in a methanolic solution of perdeuterated octadecanethiol (d-ODT) for a period of 24 h. The quality of a sample of each batch of d-ODTcoated substrates was initially assessed by recording the SFG spectra in the C-D stretching region (not shown), which revealed only the three methyl resonances. This result confirms the presence of a virtually defect-free close-packed monolayer. Sample-to-sample checks of substrate hydrophobicity were performed via the simple expedient of measuring the contact angle of pure water on each sample immediately prior to use. The alkylated surfaces were subsequently dried and mounted in the flow cell. Typically, an equilibration period of 30 min was allowed with the sample retracted from the prism for each solution under flowing conditions. The substrate was then advanced to within a few micrometers of the prism and the SFG spectrum recorded between 2750 and 3050 cm-1 in order to provide a background baseline and, hence, permit the determination of the signal-to-noise level from regions free from known resonances. To confirm that the SFG signal originates solely from the sample surface and not from the CaF2 prism/solution interface (where it has previously been shown that SDS adsorbs28,29), spectra were recorded using a fused silica prism. Due to surface charge reversal with respect to CaF2, little ordered adsorption has been observed on this substrate compared to that detected on CaF2. No differences were apparent between the SFG spectra using the different prism materials, and it was therefore concluded that adsorption at the prism/solution interface makes no significant contribution to the SFG spectra reported here. All glassware, O-rings, and stainless steel components were cleaned following standard procedures30 and rinsed copiously with 18.2 MΩ cm-1 Milli-Q water. SDS (Aldrich) was doubly recrystal(25) Duffy, D. C.; Davies, P. B.; Bain, C. D. J. Phys. Chem. 1995, 99, 15241. (26) Lambert, A. G.; Neivandt, D. J.; Briggs, A. M.; Usadi, E. W.; Davies, P. B. J. Phys. Chem. B 2002, 106, 5461. (27) Lambert, A. G. Ph.D. Thesis, University of Cambridge, 2001. (28) Becraft, K. A.; Richmond, G. L. J. Phys. Chem. B 2005, 109, 5108. (29) Becraft, K. A.; Moore, F. G.; Richmond, G. L. PCCP Phys. Chem. Chem. Phys. 2004, 6, 1880. (30) Windsor, R. Ph.D. Thesis, University of Cambridge, 2001.

Adsorption of SDS at the Solid/Aqueous Solution Interface

Langmuir, Vol. 22, No. 7, 2006 3107

lized from ethanol prior to use. d-SDS (CDN Isotopes) was kindly supplied by Unilever R&D Port Sunlight and was used as received. PEG (Aldrich) with number average molecular weights (Mn) of 400, 600, 900, 1500, and 12 000 (Mw/Mn ) 1.1) were used as received for initial trial investigations. All subsequent experiments comprising the results reported here were performed with lower-dispersity polymers (typical Mw/Mn ) 1.00-1.05, from Polymer Laboratories plc) of the nearest equivalent molecular weights. No differences were observed in the results when the same molecular weight polymers from the different suppliers were used. SDS solutions were prepared immediately prior to use at above the pure surfactant cmc (8.3 mM at 298 K) and diluted as required with water or with stock solutions of polymer. The SDS solution concentrations used throughout this work are expressed as fractions of the cmc of the pure surfactant at 298 K, while the polymer concentration is expressed as parts per million (ppm) of the solution by weight. No turbidity was observed in any of the solutions prepared and employed in this work. For all of the results presented here, the flow cell was first equilibrated with a stock solution of PEG for 30 min prior to spectral acquisition. The cell was then drained, and a mixed solution of PEG/SDS was introduced and equilibrated for 30 min before further spectra were recorded. This procedure was then repeated for all subsequent increases in surfactant concentration.

Results and Analysis SFG spectra of the adsorbate from solutions containing the following molecular weight PEG samples, 400, 600, 900, and 1500 in the presence of SDS, in the C-H stretching region from 2800 to 3000 cm-1 are presented in this section. To facilitate comparison with earlier results24 the spectra of pure SDS and SDS with PEG 12 000 are also included. The concentrations of SDS used are given in units of the room temperature cmc of pure SDS, namely 8.3 mM, and the spectra are offset vertically from each other for clarity. The spectrum of the pure PEG sample, labeled with its molecular weight, is shown as a reference. Lorentzian fits of the experimental data are shown as solid lines and experimental points as filled circles. All spectra are shown to the same scale and are normalized to the gold nonresonant signal. Figure 1 shows the SFG spectra of the adsorbate from solutions of different concentrations of SDS in the absence of polymer. The appearance of the spectra as dips in the counterpropagating geometry and ppp polarization employed here shows that the functional groups in the hydrocarbon tails of the surfactant have a net orientation toward the d-ODT/gold surface. The SFG intensity (ISFG) can be envisaged as a combination of a resonant susceptibility contribution (χ(2) R ) arising from the molecular layer and a nonresonant susceptibility contribution (χ(2) NR) arising from the substrate, (2) 2 ISFG ∝ |χ(2) R + χNR|

(1)

which can be expanded to include the resonant (δ) and nonresonant () phases: (2) 2 (2) (2) (2) | + |χNRijk |2 + 2|χRijk ||χNRijk | cos( - δ) ) |χRijk

(2)

where i, j, and k represent the tensor components of χ(2) R in the laboratory frame. For a metal substrate such as gold, χ(2) NR is dependent on the excitation wavelength and typically is 28-31 Contributions from all comparable in magnitude to χ(2) R . terms of eq 2 must therefore be considered. The χ(2) NR term is (typically) almost constant with infrared frequency and hence provides a nearly frequency-independent background SFG signal. The third term in eq 2, a cross term of the resonant and nonresonant

Figure 1. Spectra of pure SDS adsorbed from aqueous solution as a function of SDS concentration.

susceptibilities, containing the phase angles, provides a second resonant contribution to the SFG spectrum. The line shapes of the spectral features appearing in an SFG spectrum are determined by the relationship between the phase angles δ and , provided there is a strong nonresonant background. The nonresonant phase angle, , can be considered to be infrared frequency independent over the measured frequency range. On a gold substrate, its value is approximately +90° for the counterpropagating beam geometry of the nanosecond spectrometer with an excitation wavelength of 532 nm.28,29 The absolute value of these phases is dependent upon the precise geometry, wavelength, and the polarization combination employed.28-31 The resonant phase, δ, varies as the infrared frequency tunes through the resonant frequency of a specific vibrational mode and can be modeled with a phase angle of δ ) +90° on resonance if this mode is located in a functional group oriented in a direction away from the interface.32 For present purposes the symmetric (r+) and asymmetric (r-) stretching modes of the methyl group are assumed to have the same phase. Thus, the counterpropagating geometry using 532 nm visible light will lead to constructive interference between the resonant and nonresonant phases (cos( - δ) ) 1 in eq 2), giving rise to spectral resonances of positive phase, i.e., peaks. Conversely, if the resonant functional group is oriented toward the surface, the spectral features appear with negative phase, i.e., as dips, due to destructive interference. Figure 2 contains the adsorbate SFG spectra obtained from pure PEG 400 solution and from solutions of PEG with increasing concentrations of SDS. First, it is apparent that for Mw ) 400, the lowest polymer molecular weight used in this study, there (31) Hai, M. T.; Gao, J.; Han, B. X.; Yan, H. K.; Liu, Y.; Han, Q. Y. Acta Phys.-Chim. Sin. 1998, 14, 747. (32) Dreesen, L.; Humbert, C.; Hollander, P.; Mani, A. A.; Ataka, K.; Thiry, P. A.; Peremans, A. Chem. Phys. Lett. 2001, 327.

3108 Langmuir, Vol. 22, No. 7, 2006

Figure 2. Spectra of pure PEG 400 and PEG 400/SDS, adsorbed from binary solution, with increasing SDS concentrations.

Figure 3. Comparison of the fits of the SFG spectra from pure SDS at one-hundredth of the cmc (red line) and from a polymer/SDS mixture at the same surfactant concentration (black line).

are no spectral resonances attributable to the polymer in any of the spectra in Figure 2. As the concentration of the surfactant is increased, resonances characteristic of SDS begin to appear and are clearly visible in the spectrum recorded at 1/100 cmc. However, examination of the spectrum at this surfactant concentration shows that it does not match the spectrum obtained for the adsorption of SDS from this solution concentration in the absence of polymer (Figure 1), which has been recorded in numerous other studies. To aid the comparison, the spectra are overlain in Figure 3. In particular, the intensity of the lowestfrequency feature in Figure 3, the methylene symmetric stretch, d+, which is quite well resolved in both spectra, is significantly

Casford et al.

diminished in the case of the mixture compared to the spectra observed for SDS alone. Furthermore, the expected methyl antisymmetric stretching resonance, r-, at ca. 2960 cm-1 is absent from the spectrum of the adsorbate formed from the mixed solution. There are several possible explanations for this difference. First, the polymer is adsorbing at the interface but it is entirely isotropic in its surface distribution and its ordering, or, if ordered, that all potential SFG-active functional groups are lying in the plane of the interface. The latter conclusion is a consequence of the beam polarization combination employed (ppp). Only adsorbates having a resonant susceptibility component with a z constituent (χzzz, χxxz) i.e., predominantly perpendicular to the surface plane, generate a substantial SFG response from this polarization combination on a gold substrate. Second, the reduced intensity of the surfactant SFG resonances and the absence of the methyl r- resonance is due to the formation of a polymer/ surfactant complex in bulk solution, leading to a reduction in the effective SDS concentration available for adsorption to the interface. The second hypothesis is considered to be of negligible effect in view of the published surface tension data,31 which indicate little or no interaction between PEG and SDS at concentrations below approximately half of the pure surfactant cmc. The steady increase in the SDS SFG signals from the solid/ liquid interface (Figure 2) as the SDS concentration in the solution increases occurs for both the normalized resonant intensity and for the ratio of the intensities of the methyl and methylene symmetric stretches, r+ and d+, respectively. An increase in the magnitude of the r+/d+ ratio indicates that fewer gauche defects are present in the alkyl chain, and this is taken to denote an increase in the conformational ordering of the surfactant. Comparison of the 1 cmc plots with the results obtained for the pure SDS solution shows that the r+/d+ ratio is consistent with that found for an almost complete SDS monolayer measured in other work. The picture that emerges then is that at low SDS concentrations PEG 400 is adsorbed at the interface but is insufficiently ordered to exhibit an SFG spectrum. As the concentration of SDS increases, the surfactant competes with the polymer for the interface more efficiently until by the cmc an effectively complete monolayer of SDS is formed. Due to the absence of any discernible spectral resonances attributable to PEG 400 (and PEG 600) adsorbing from their pure polymer solutions, recording spectra from PEG/d-SDS solutions was deemed not worthwhile. Increasing the molecular weight of the polymer to 600 Mw significantly alters the observed SFG spectra at different SDS solution concentrations (Figure 4). As for PEG 400, there is no spectral evidence for the ordered adsorption of the polymer from pure PEG solution. As the concentration of surfactant is increased to one-hundredth of the cmc, in contrast to that observed for PEG 400, there is little indication of ordered surfactant adsorption, although some minor resonances may be present, and the onset of significant adsorption of surfactant only begins to appear at 1/30 cmc. However, there is a significantly reduced signal intensity compared with that for the same surfactant concentration in combination with PEG 400 (Figure 2). Furthermore, there must be a reduction in conformational ordering because the methyl r resonances are completely absent, in particular the r- resonance at 2960 cm-1. From the onset of ordered adsorption as indicated by the appearance of the SFG spectra, there is a monotonic increase in the r+/d+ ratio. This ratio is, however, less than that from SDS/PEG 400. From the spectral evidence, in particular the absence of the methyl r- resonance in the 1/10 cmc spectrum,

Adsorption of SDS at the Solid/Aqueous Solution Interface

Figure 4. Spectra of PEG 600 and PEG 600/SDS adsorbed from binary solution. Dashed lines indicate the position of the spectral d and r resonances

it is reasonable to conclude that either less surfactant is adsorbing at the interface for a given surfactant concentration or the adsorbed surfactant is significantly more disordered up to at least onetenth of the cmc. Additionally, the conformational ordering of the surfactant layer is reduced in the presence of PEG 600 compared with PEG 400 throughout the spectral sequence, although by the cmc, the difference becomes slight. Specifically, the presence of strong r resonances in the spectrum from PEG 600/1 cmc SDS solution indicates that an almost complete surfactant monolayer is present at the interface. Comparison of the r+/d+ ratio and peak intensities shows that the cmc spectrum (Figure 4) is comparable to that from SDS/PEG 12 000 solution at the cmc. Increasing the polymer molecular weight to 900 Mw gives a spectral sequence similar to that observed for PEG 600 but now with the appearance of a small spectral signature attributable to adsorbed PEG 900 (Figure 5). The polymer d+ resonance at 2870 cm-1 is observable and reduces in intensity as the surfactant concentration is increased. In contrast to that observed for the PEG 600, however, no SDS resonances are observed until the concentration reaches 1/10 cmc. The r+/d+ ratio is compatible with that observed for PEG 600 at both 1/10 and 1 cmc. In neither case does the r+/d+ ratio reach that observed for pure SDS at these concentrations. Finally, SFG spectra from SDS solutions containing PEG 1500 Mw were recorded (Figure 6). The spectra for PEG 1500 are most usefully compared to those observed for SDS and PEG 12 000 (Figure 7). The resonances attributable to the adsorption of PEG at the interface are present for SDS concentrations up to one-thirtieth of the cmc. Simultaneously, SDS resonances are also visible at 1/30 cmc. The slight resonant intensity at ca. 2870 cm-1 is attributed here to residual polymer CH2 groups, as no evidence of the methyl ris present in the spectrum. As the SDS concentration is increased

Langmuir, Vol. 22, No. 7, 2006 3109

Figure 5. Spectra of PEG 900 and PEG 900/SDS adsorbed from binary solution.

Figure 6. Spectra of PEG 1500 and PEG 1500/SDS adsorbed from binary solution.

further, the r+/d+ ratio of the SDS resonances increases monotonically and the polymer resonances eventually disappear. To confirm the assignment and intensity changes in the PEG resonances, spectra were taken of PEG 900 and PEG 1500 adsorbed from solutions containing d-SDS. As expected, the spectra attributable to the PEGs decreased in intensity as the

3110 Langmuir, Vol. 22, No. 7, 2006

Casford et al. Table 1. Average r+/d+ Ratios for 1 cmc SDS Solutions for Different PEG Molecular Weightsa solution composition

r+/d+ ((0.05)

1 cmc pure SDS PEG 400 and 1 cmc SDS PEG 600 and 1 cmc SDS PEG 900 and 1 cmc SDS PEG 1500 and 1 cmc SDS PEG 12 000 and 1 cmc SDS

0.36 0.34 0.28 0.22 0.29 0.35

a Error represents the maximum experimental variation observed over all experimental runs.

Figure 7. Spectra of PEG 12 000 and PEG 12 000/SDS adsorbed from binary solution.

Figure 8. d-SDS and PEG 900 spectra as a function of d-SDS concentration.

d-SDS concentration increased. The d-SDS/PEG 900 spectra are shown in Figure 8.

Discussion With the exception of the Mw 400 and 600 samples, there is clear SFG evidence of polymer adsorption from the appearance

of the PEG methylene groups. The direct spectral evidence for polymer adsorption is first apparent in the spectrum of the adsorbate from pure PEG 900 solutions. In addition, the inhibitory effect of the PEG upon adsorption of SDS, i.e., the suppression of SDS spectra at the lower SDS concentrations compared with pure SDS solutions (Figure 1), is unambiguous at all PEG molecular weights. As the surfactant concentration is increased, resonances that can definitely be assigned to ordered adsorbed surfactant become dominant at approximately 1/10 cmc, and by the 1 cmc concentrations, the SDS spectra are similar for all polymer molecular weights within the experimental error, as shown in Table 1. Note data shown are the average of all the experimental runs carried out for each molecular weight with the error calculated to include the maximum observed variation across all runs. In all the spectra recorded, the peak intensities are negative going, i.e., they are dips. The implication of this for the counterpropagating beam geometry employed here is that the adsorbate molecules must have a net orientation with the methylene hydrogen atoms of the polymer and both the methyl and methylene hydrogen atoms of the surfactant pointing toward the hydrophobic interface. The weak SFG signals arising here from PEG adsorbed from solution onto a hydrophobic surface can be compared with two other experiments in which PEG has been observed by SFG. Dreeson et al32 recorded intense SFG signals from PEG 400 at the pure solid polymer/air interface. The signal intensity diminished significantly in the presence of water. They attributed this to the disordering of the PEG surface when in contact with water. The PEG film in these experiments was a thick, cast film on a hydrophilic CaF2 substrate. In this case, the surface density of the polymer was considerably greater than that found on adsorption from solution, as investigated here. The spin-cast deposition produces a polymer brush in which the chains are sterically unable to lie in the plane of the interface. The most likely explanation for the lack of a SFG signal here when PEG 400 is adsorbed from solution is that only partial coverage by PEG 400 occurs permitting, the polymer chains to adopt disordered conformations. Our observations are in accord with the SFG results of Chen et al33 on PEG 400. They found no signals from PEG 400 adsorbed from solution at polystyrene and poly(methyl methacrylate) interfaces, while only small SFG signals appeared when the samples were removed from solution and the SFG spectrum of the interface was recorded in air. It is reasonable to conclude then that PEG 400 does not adsorb to a sufficient extent to lead to significant ordering at the hydrophobic/ aqueous solution interface investigated here. Neither Dreeson et al nor Chen et al examined higher-molecular-weight PEG. The results obtained by changing the PEG molecular weight can be broadly interpreted along similar lines to those used previously to explain the results for PEG 12 000 solutions, (33) Chen, C.-y.; Even, M. A.; Wang, J.; Chen, Z. Macromolecules 2002, 35, 9130.

Adsorption of SDS at the Solid/Aqueous Solution Interface

inasmuch as there is clear evidence of competitive adsorption between the SDS and the PEG moieties. For PEG 1500, the recorded spectra are both qualitatively and quantitatively similar to those recorded for PEG 12 000. The SFG spectrum of the polymer adsorbed from the pure PEG 1500 solution is clearly more intense than that observed for the lower-molecular-weight PEG samples and is little different from that recorded for PEG 12 000 (and also for PEG 100 000, not shown). It is therefore likely that the adsorption of the polymer on the surface is similar in terms of orientation and conformation for polymer weights of 1500 and higher, up to molecular weights of at least 100 000. Intuitively, this would seem to be reasonable on the grounds that long-chain polymers have significantly more bonds to the surface that have to be broken before the chain can desorb from the interface. However, from a thermodynamic perspective, there is little difference between the surface coverage adsorption efficiency of long-chain polymers in comparison to shorter chains,34 and the principle difference in the desorption or surface rearrangement of longer polymer chains is thought to be due to steric considerations, such as increased pinning of the train bound fraction by polymer loops, and to the entropy of solvation for higher-weight polymers. The interpretation of the SFG spectra in terms of the polar orientation and conformational ordering of the adsorbed polymer layer is now considered. The presence of dips in the spectra of the PEG denotes, for the counterpropagating geometry and ppp polarization used here, that the methylene groups of the polymer backbone must be oriented toward the hydrophobic interface. The increasing intensity of these methylene resonances as the polymer molecular weight increases implies that the origin of the SFG signal is unlikely to be the tails of the polymer, as the fraction of the polymer comprised of tail groups decreases with increasing molecular weight for the weights and concentrations used here. The origin of the SFG signal is most likely to be the train-bound fraction. This would be consistent with the experimental observations of increasing SFG response with increased molecular weight for the shortest oligomers followed by a relatively constant SFG response for molecular weights of 1500 and above. Furthermore, it would account for the absence of any significant SFG signal from the lowest-molecular-weight polymers, where the train-bound fraction is predicted to be negligible. The expected polar orientation for a train-bound PEG fraction would be with the methylene groups oriented toward the hydrophobic surface to maximize the hydrophobic interaction, while the polar ether groups should orient toward the overlying aqueous phase in order to maximize the hydrogen bonding of the ether oxygen atoms with the neighboring water molecules. It is therefore likely that, as the surfactant concentration increases, the train-bound fraction of the adsorbed PEG decreases, either through surface reorientation or by partial desorption of the PEG, as observed at the air/aqueous solution interface,35 giving rise to the reduced SFG intensity from the polymer which was unambiguously confirmed in the perdeuterated SDS spectra. The current understanding of how SDS adsorbs onto PEG in solution is incomplete. It has been suggested36,37 that on addition of SDS the Na+ counterions of the surfactant bind directly to the oxygen atoms of the PEG, causing the strand to contract and coil (34) Fleer, G. J.; Stuart, M. A. C.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces, 1st ed.; Chapman and Hall: London, 1993. (35) Cooke, D. J.; Dong, C. C.; Lu, J. R.; Thomas, R. K.; Simister, E. A.; Penfold, J. J. Phys. Chem. B 1998, 102, 4912. (36) Kuhl, T. L.; Berman, A. D.; Hui, S. W.; Israelachvili, J. N. Macromolecules 1998, 31, 8258.

Langmuir, Vol. 22, No. 7, 2006 3111

in solution. This behavior has recently been observed in tethered PEG strands by Jayachandran et al.11 using fluorescence energy transfer (FRET). The exact structure adopted by the polymer on binding to the surfactant headgroup is still under debate. The FRET results of Jayachandran et al. are of particular relevance to this work in that the surfactant concentrations are comparable to those used here. They showed that the onset of contraction in the polymer strands starts at SDS concentrations below 8 × 10-5 M, equivalent to 1/100 cmc. The contraction increases monotonically with increasing surfactant concentration until the cmc is reached, after which the polymer strand is believed to elongate due to the electrostatic repulsion between micelles adsorbed on it. This model is only applicable for PEG with a molecular weight above 4000. Below weights of 4000, the polymer strand is believed to be incapable of adsorbing more than one surfactant micelle due to the short chain length of the polymer.23 The contraction of the polymer reduces the excluded volume of a surface-bound polymer strand, in turn reducing the steric hindrance experienced by the SDS molecules as they approach the interface. The binding of the surfactant to the polymer strand also reduces the affinity of the polymer for the surface as the entropic penalty for the solvation of the polymer is decreased. The combination of these two effects results in a reduction in the surface area occupied by the adsorbed polymer layer through partial desorption of the polymer and the reduction in the effective size of each polymer strand on the surface. This partial desorption enhances the likelihood of adsorption of SDS to the surface.

Conclusions The nonlinear optical technique of SFG vibrational spectroscopy has been employed to probe the molecular-weight dependence of the adsorption of PEG from mixed binary solutions of PEG and SDS to a model hydrophobic interface. The conformational ordering and polar orientation of the surfactant in the presence of PEG 400, 600, 900, and 1500 have been investigated as a function of the concentration of SDS at a constant concentration of the ethylene glycol repeat unit. Evidence for the adsorption of polymer and subsequent partial or complete desorption of a preadsorbed polymer layer upon addition of SDS is presented. Increasing the polymer molecular weight results in significant inhibition of the ordered adsorption of SDS, in comparison to the results obtained for PEG 400, at concentrations up to at least one-tenth of the room-temperature cmc of the pure surfactant. This effect appears to be quite pronounced for polymer weights of 600 and 900. No direct SFG spectral evidence is observed for adsorbed polymer when polymers of molecular weight below 900 are used, which has been attributed to the reduced train-bound fraction in these polymers. The polymer of weight 1500 behaves in a directly comparable way to that observed for PEG 12 000. Acknowledgment. We thank Unilever R&D Port Sunlight for providing perdeuterated SDS and Dr. S. A. Johnson for helpful discussions. M.T.L.C. thanks the Engineering and Physical Science Research Council for a studentship and the Isaac Newton Trust and Unilever Research for supplementary financial support. LA052271Z (37) Kuhl, T. L.; Berman, A. D.; Hui, S. W.; Israelachvili, J. N. Macromolecules 1998, 31, 8250.