Water Interface

Langmuir 2004, 20, 4051-4058

4051

Adsorption of Alkyl Polyglucosides on the Solid/Water Interface: Equilibrium Effects of Alkyl Chain Length and Head Group Polymerization Maria K. Matsson,*,†,‡ Bengt Kronberg,† and Per M. Claesson†,‡ YKI, Institute for Surface Chemistry, Box 5607, SE-114 86 Stockholm, Sweden, and Department of Chemistry, Surface Chemistry, Royal Institute of Technology, Drottning Kristinas va¨ g 51, SE-100 44 Stockholm, Sweden Received October 21, 2003. In Final Form: March 1, 2004 The equilibrium adsorption behavior of two n-alkyl-β-D-glucosides (octyl (C8G1) and decyl (C10G1)) and four n-alkyl-β-D-maltosides (octyl (C8G2), decyl (C10G2), dodecyl (C12G2), and tetradecyl (C14G2)) from aqueous solution on a titania surface, as measured by ellipsometry, has been investigated. The main focus has been on the effect of changes in the alkyl chain length and headgroup polymerization, but a comparison with their adsorption on the silica/water and air/water interfaces is also presented. Some comparison with the corresponding adsorption of ethylene oxide surfactants, in particular C10E6 and C12E6, is given as well. For all alkyl polyglucosides, the maximum adsorbed amount on titania is reached slightly below the critical micelle concentration (cmc), where it levels off to a plateau and the amount adsorbed corresponds roughly to a bilayer. However, there is no evidence that this is the actual conformation of the surfactant assemblies on the surface, but the surfactants could also be arranged in a micellar network. On hydrophilic silica, the adsorbed amount is a magnitude lower than on titania, corresponding roughly to a layer of surfactants lying flat on the surface. A change in the alkyl chain length does not result in any change in the plateau molar adsorbed amount at equilibrium; however, the isotherm slope for the alkyl maltosides increases with increasing chain length. Headgroup polymerization on the other hand affects the adsorbed amount. The alkyl glucosides start adsorbing at lower bulk concentrations than the maltosides and equilibrate at higher adsorbed amounts above the cmc. When compared with the ethylene oxide (EO) surfactants, it is confirmed that the EO surfactants hardly adsorb on titania, since the measured changes in the ellipsometric angles are within the noise level. They do, however, adsorb strongly on silica.

Introduction Alkyl polyglucosides, CnGm, are a class of surfactants that has become of increasing interest during the past few years, for which environmental concern is one reason. The alkyl polyglucosides can be synthesized from renewable raw materials and are biodegradable and nontoxic.1 Furthermore, they show good detergency properties, while being mild to the skin,2 and are currently used in some personal care products. The surfactants consist of a number of glucose units linked to an alkyl chain via an ether bond. This study focuses on alkyl glucosides and maltosides; i.e., the headgroup is built from one or two glucose units. Knowledge of the adsorption behavior of a surfactant is crucial for understanding the underlying mechanisms for applications such as wetting, detergency, and lubrication. The aim of this work was to determine and rationalize trends in the adsorption behavior of alkyl polyglucosides due to structural changes, i.e., head and tail group sizes, and to correlate this to similar relationships for the more known ethylene oxide (EO) surfactants. Another important issue is to explain the differences in adsorption of alkyl maltosides on the metal oxide/water and silica/water interfaces respectively, as well as to elucidate the differences in adsorption behavior between alkyl polyglucosides and ethylene oxide surfactants on these surfaces. * To whom correspondence should be addressed. E-mail: [email protected]. † YKI, Institute for Surface Chemistry. ‡ Royal Institute of Technology. (1) Garcı´a, M. T.; Ribosa, E.; Campos, J.; Sanchez Leal, J. Chemisphere 1997, 35, 545. (2) Schmid, K.; Tesmann, H. In Detergency of Specialty Surfactants; Friedli, F. E., Ed.; Dekker: New York, 2001; Vol. 98, p 1.

The bulk behavior of alkyl polyglucosides has been investigated in some detail,3-7 as has the adsorption behavior at the air/liquid interface.8-14 However, studies on the adsorption process at the solid/liquid interface are still limited. Investigations of the dispersion of titania pigments in aqueous solutions using technical grade surfactants show that alkyl polyglucosides adsorb on titania and that the adsorbed amount exceeds that of a monolayer.15 It was suggested that the adsorption could be explained by hydrogen bonding of the slightly acidic hydroxyl groups on the surfactant with the basic OH groups on the oxide surface. On alumina, this has been further investigated, and again the authors concluded that hydrogen bonding is the main driving force for adsorp(3) Boyd, B. J.; Drummond, C. J.; Krodkiewska, I.; Grieser, F. Langmuir 2000, 16, 739. (4) Do¨rfler, H.-D.; Go¨pfert, A. J. Dispersion Sci. Technol. 1999, 20, 35. (5) Nilsson, F.; So¨derman, O.; Hansson, P.; Johansson, I. Langmuir 1998, 14, 4050. (6) Nilsson, F.; So¨derman, O.; Johansson, I. Langmuir 1996, 12, 902. (7) Shinoda, K.; Carlsson, A.; Lindman, B. Adv. Colloid Interface Sci. 1996, 64, 253. (8) Bo¨cker, T.; Thiem, J. Tenside, Surfactants, Deterg. 1989, 26, 318. (9) Drummond, C. J.; Warr, G. G.; Grieser, F.; Ninham, B. W.; Evans, D. F. J. Phys. Chem. 1985, 89, 2103. (10) Kahl, H.; Enders, S.; Quitzsch, K. Colloids Surf., A 2001, 183185, 661. (11) Kjellin, U. R. M.; Claesson, P. M.; Vulfson, E. N. Langmuir 2001, 17, 1941. (12) Rosen, M., J.; Sulthana, S., B. J. Colloid Interface Sci. 2001, 239, 528. (13) Aveyard, R.; Binks, B. P.; Chen, J.; Esquena, J.; Fletcher, P. D. I. Langmuir 1998, 14, 4699. (14) Shinoda, K.; Yamaguchi, T.; Hori, R. Bull. Chem. Soc. Jpn. 1961, 34, 337. (15) Smith, G. A.; Zulli, A. L.; Grieser, M. D.; Counts, M. C. Colloids Surf., A 1994, 88, 67.

10.1021/la035959p CCC: $27.50 © 2004 American Chemical Society Published on Web 04/14/2004

4052

Langmuir, Vol. 20, No. 10, 2004

tion.16 Comparisons between adsorption on different solid surfaces show that the adsorption is high on metal oxides while low on silica surfaces.17,18 Ethylene oxide based surfactants on the other hand adsorb strongly on silica, but scarcely on metal oxides. This difference is especially interesting since this latter adsorption process also has been explained by invoking hydrogen bonding.19 We note that the standard explanation put forward to explain the adsorption of nonionic surfactants on hydrophilic surfaces from aqueous solutions is hydrogen bonding. The main difficulty with this hypothesis is that water is able to form very strong hydrogen bonds with both the surface and the surfactant headgroup and it is hard to see how replacements of such hydrogen bonds with surfactant headgroupsurface hydrogen bonds will result in an energy gain. Nevertheless, the enthalpy of adsorption for poly(ethylene oxide) on silica is negative.20 We conclude that the driving force for adsorption is not well-understood; however, one may perceive a situation where water hydrogen bonded both to the surface and the surfactant headgroup mediates the surfactant-surface interactions. On hydrocarbon and at air/water interfaces, both surfactants adsorb to an approximate monolayer.21-23 So far, the techniques used to study the adsorption behavior of alkyl polyglucosides have been solution/depletion and calorimetry, but ellipsometry, the main technique used here, has not previously been employed.

Matsson et al. Methods. The main experimental method used to investigate the adsorption properties of the surfactants was null ellipsometry. The instrument used is a Multiskop (Optrel GdBR, Berlin, Germany), with horizontally aligned measurement arms with an angle of incidence of 68°, equipped with a 532 nm laser and set up according to the PCSA arrangement. The instrument has been extensively described elsewhere.24 In principle the ellipsometer measures the ellipticity of light before and after reflection at a substrate. From these measurements, the ellipsometric angles Ψ and ∆, which reflect the difference in polarization between the incident and the reflected light, are obtained. Ψ is related to the change in the amplitude upon reflection, and ∆ describes the phase shift. The ellipsometric angles are related to the optical parameters of the substrate as

tan(Ψ)ei∆ ) f(λ,φ0,optical params)

where λ is the wavelength of the light and φ0 the angle of incidence. The optical parameters are the refractive indices of the substrate and adsorbent and their respective layer thickness as well as the refractive index of the ambient media.25 By studying the bare substrate in two ambient media, air and water, both the complex refractive index of the bulk material and the thickness and the refractive index of the upper oxide layer can be calculated.26 From the change in polarization upon reflection during adsorption, the thickness and the refractive index of the surfactant layer are calculated, and from those the adsorbed amount, using the de Feijter’s formula:

Γ)

Experimental Section Materials. A series of monodisperse alkyl β-D-polyglucosides, CnGm (C8G1, C8G2, C10G1, C10G2, C12G2, C14G2) g 99% pure, purchased from Anatrace, were used without further purification. For comparative measurements two ethylene oxide surfactants, C10EO6 and C12EO6 g 99% pure, purchased from Nikko Chemicals, were used. Polished silicon wafers, thermally oxidized to produce a SiO2 layer thickness of approximately 300 Å and then cut into slides with a width of 12.5 mm, were provided by Dr. Stefan Klintstro¨m, University of Linko¨ping, Linko¨ping, Sweden. Titanium slides, prepared by evaporating an approximately 1500 Å thick layer titanium on polished silicon wafers, which then were cut into slides with a width of 8 mm, were provided by Bo Thune´r, University of Linko¨ping. The silica-covered silicon wafers were cleaned for 10 min at 80 °C in a mixture of NH4OH (29%), H2O2 (30%), and H2O at a 1:1:5 ratio, followed by a cleaning for 10 min at 80 °C in a mixture of HCl (37%), H2O2 (30%), and H2O at a 1:1:6 ratio. They were then stored in ethanol until used. Just prior to the measurements the wafers were plasma-cleaned in low-pressure air for 5 min using a radio frequency glow discharge apparatus (Harrick PDC-3XG, Harrick Scientific Corp., Ossining, NY). The titanium-covered silica wafers were flushed in ethanol and dried with nitrogen to remove particles and then cleaned using only the plasma cleaner in the same way as the silica slides. All glassware used in the experiments, for storing and preparation of surfactant solutions, were cleaned using surfactant-free Deconex and then thoroughly rinsed consecutively with water, ethanol, and water. The water used in all experiments was treated by a Milli-Q Plus unit (Millipore, Bedford, MA) including ion exchange, active carbon adsorption, and reverse osmosis before the final 0.22 mm filtration step, yielding ultrapure reagent-grade water of resistivity 18.2 MΩ cm (at 25 °C). (16) Zhang, L.; Somasundaran, P.; Mielczarski, J.; Mielczarski, E. J. Colloid Interface Sci. 2002, 256, 16. (17) Kira´ly, Z.; Bo¨rner, R. H. K.; Findenegg, G. H. Langmuir 1997, 13, 3308. (18) Zhang, L.; Somasundaran, P.; Maltesh, C. J. Colloid Interface Sci. 1997, 191, 202. (19) Partyka, S.; Zaini, S.; Lindeheimer, M.; Brun, B. Colloids Surf. 1984, 12, 255. (20) Trens, P.; Denoyel, R. Langmuir 1993, 9, 519. (21) Somasundaran, P.; Snell, E. D.; Xu, Q. J. Colloid Interface Sci. 1991, 144, 165. (22) Tiberg, F.; Landgren, M. Langmuir 1993, 9, 927. (23) Kira´ly, Z.; Findenegg, G. H. Langmuir 2000, 16, 8842.

(1)

(n - n0) d dn/dc

(2)

where n0 is the refractive index of the surrounding media, n and d are the refractive index and the thickness of the adsorbed layer, and dn/dc is the refractive index increment of the surfactant solution. In the calculations we have invoked a model of the adsorbed layer as being a homogeneous planar film characterized by its thickness and refractive index. We note that the errors obtained for n and d during evaluation are significantly larger than the errors in the adsorbed amount. This is due to the errors in n and d being coupled, and they largely cancel out when the adsorbed amount is calculated.27 During the measurements the substrate was placed vertically in a 5 mL quartz cuvette equipped with tubing for inlet and outlet liquid and stirred at a constant rate of 300 rpm. The cuvette was temperature-controlled to within (0.1 °C. The adsorption process was studied as the bulk concentration was increased stepwise, either by adding surfactant from a concentrated stock solution or by pumping a solution at the desired concentration through the tubing. All measurements have been performed at pH 5.8 and 25.0 °C unless otherwise stated. The refractive index increment, dn/dc has been determined for each surfactant using an interferometric refractometer (Optilab DSP, Wyatt Technology Corp., Santa Barbara, CA) The instrument is based on a wave front shearing technique and measures the difference in refractive index between a stored reference liquid and a liquid sample stream. The refractive index is measured at a fixed wavelength of 450 nm, and the instrument is equipped with a temperature-control system, here set to 25 °C. The difference in refractive index was measured for five different concentrations of surfactant, with water as the reference solution. ESCA28 was performed to determine the chemical composition of the titanium slide surface. This is partly to verify that the upper surface layer is titania and partly to show that the chosen cleaning procedure leaves a clean surface as result, without damaging the original surface. With ESCA, the upper 5 nm of (24) Harke, M.; Teppner, R.; Schulz, O.; Motschmann, H. Rev. Sci. Instrum. 1997, 68, 3130. (25) Ellipsometry and polarized light; Azzam, R. M. A., Bashara, N. M., Eds.; North-Holland: Amsterdam, 1977. (26) Landgren, M.; Jo¨nsson, B. J. Phys. Chem. 1993, 97, 1656. (27) de Feijter, J. A.; Benjamins, J.; Veer, F. A. Biopolymers 1978, 17, 1759. (28) Practical surface analysis: By Auger and X-ray photoelectron spectroscopy; Briggs, D., Shea, M. P., Eds.; Wiley: Chicester, U.K., 1983.

Adsorption of Alkyl Polyglucoside

Langmuir, Vol. 20, No. 10, 2004 4053

the surface is studied in great detail, investigating the relative amount of different elements and their chemical environment on the surface. First a wide spectrum was run to detect elements present in the surface layer. This was followed by detailed spectra for all elements detected in the wide spectrum, from which the relative surface compositions were obtained. The spectra were recorded using a Kratos AXIS HS X-ray photoelectron spectrometer (Kratos Analytical, Manchester, U.K.). The samples were analyzed using a monochromator (Al X-ray source). The analysis area was below about 1 mm2. AFM29,30 can be used for imaging, chemical mapping and force spectroscopy. Here, the surface was investigated using the tapping-mode technique, giving a visual description of the surface roughness and structure. With AFM, details in the nanometer range are easily detectable. Some measures of the roughness can also be calculated; here, the surface roughness is described by the Ra- and Rq-values and the surface area difference. The Ra-value is the average surface roughness, and the Rq-value is the root-mean-square average of the deviation from a plane fit to the surface; they are calculated according to the following equations:

Ra ) Rq )

1 L

∫ |z(x)| L

0

xL1 ∫ z(x) dx L

0

2

(3)

(4)

where z(x) is the vertical deviation from a plane fit of the surface in the point x and L is the length of the distance measured. The surface area difference is the difference between the real area of a rough surface, Areal, which the surfactant recognizes, and the measured projected area, Aproj. The instrument used was a multimode nanoscope IIIa AFM (Digital Instruments, Santa Barbara, CA). The projected surface area imaged was 1 and 0.25 µm2, and each image was scanned line by line with 512 scan lines. During the measurement, errors causing height changes can occur. These errors have, before calculating the roughness parameter, been smoothed out using the flatten function of the AFM. To determine macroscopic hydrophobicity of the substrate, the contact angle of pure water on the surface was measured using a contact-angle goniometer, A 1100 (Rame´ Hart, Inc., NJ), after plasma cleaning, after adsorption from a solution at a concentration above the critical micelle concentration (cmc) followed by drying with N2, and after rinsing with pure water followed by drying with N2.

Figure 1. Overview ESCA spectrum for the Ti substrate, showing peaks from C at 280-290 eV, Ti at 450-470 eV, and O around 530 eV. The insert shows a detail 2p spectrum for the Ti peak, where the peaks at 459 and 465 eV correspond to titanium in TiO2, while the small peak at 454 eV corresponds to titanium in Ti metal form.

Figure 2. Image of the titanium surface topography, as seen by AFM including the calculated Ra- and Rq-values. The surface was investigated using the tapping-mode technique.

Results Surface Characterization. From ESCA measurements it can be concluded that the upper 5 nm of the titanium surface consists mainly of Ti and O, although some carbon is still left at the surface, Figure 1. The bond energy spectrum for Ti shows peaks at values corresponding to pure Ti and TiO2, but no peaks corresponding to other titanium oxides. The ratio nTi/nO for the layer is approximately 0.45. The carbon content in the upper surface layer has been significantly lowered after treatment with the plasma cleaner, compared with a slide that only was flushed with ethanol and dried with nitrogen. The titanium content is about the same both before and after plasma treatment. The surface characterization gives Ra ) 4.349 nm and Rq ) 3.463 nm for the titania surfaces when measured over 1 mm2, giving Ra/Rq ) 1.26, which corresponds approximately to a Gaussian distribution of the topological height features. Calculation of the real surface area using (29) Binnig, G.; Quate, C. F.; Gerber, C. Phys. Rev. Lett. 1986, 56, 930. (30) Atomic force microscopy/scanning tunneling microscopy; Cohen, S. H., Bray, M. T., Lightbody, M. L., Eds.; Plenum Press: New York, 1994.

Figure 3. Adsorption isotherm for C10G2, showing the adsorbed amount (open circles) and the corresponding headgroup area per surfactant (filled circles), assuming that the surfactants assemble in a bilayer. The solid line is drawn to guide the eye.

an internal AFM routine gives a surface area ratio, (Areal - Aproj)/Aproj ) 0.092. Figure 2 shows an AFM image of the titania surface, where it also can be observed that the upper layer of the substrate seems to have a granular structure. After cleaning with plasma, water completely wets both the silica and the titania slides (θ ) 0); i.e., the surface is clearly hydrophilic.

4054


Matsson et al.

Table 1. Presentation of Literature Values and Measured Results for the Different Surfactantsa surfactant

M (g mol-1)

cmc (mM)

dn/dc (mL g-1)

Γcmc* a (mg m-2)

Γcmc*/** a (mmol m-2)

A/molecule*/** a (nm2)

A/molecule at air/liquid (nm2)

C8G1 C8G2 C10G1 C10G2 C12G2 C14G2

292.4 454.4 320.4 482.6 510.6 538.6

22b 19c 2.0d 2.0d 0.15e 0.015f

0.146 0.146 0.153 0.146 0.142 0.142

2.6 3.4 2.8 3.6 3.8 4.1

8.7/7.9 7.5/6.8 8.8/8.0 7.6/6.9 7.4/6.7 7.6/6.9

0.38/0.42 0.44/0.49 0.38/0.42 0.44/0.48 0.45/0.50 0.44/0.48

0.38b 0.42c 0.40d 0.49d 0.50e

a Adsorbed amount and area per molecule, assuming bilayer structure, are presented both for an ideally smooth surface (*) and corrected with surface roughness (**). b Reference 11. c Reference 3. d Reference 12. e Reference 9. f Reference 8.

Figure 4. Adsorption isotherms for C10G1 at (a, left) 25 and (b, right) 30 °C. Each individual measurement series reaches a plateau value around cmc, but the scatter around the average is almost a magnitude larger for the lower temperature. The averages for the two temperatures are also different; while the adsorption at 30 °C corresponds to that expected for a double layer, the adsorption at 25 °C significantly exceeds it. The solid lines represent average values, drawn to guide the eye.

Adsorption at the Solid/Liquid Interface. The ellipsometer has been used to measure equilibrium data at different concentrations to obtain adsorption isotherms for the surfactants. The measured refractive index increments, dn/dc, necessary to calculate the adsorbed amounts from the optical characteristics of the surfactant film obtained from the ellipsometer, are listed in Table 1, together with a summary of the plateau adsorbed amounts for all studied surfactants and other properties of the surfactants. Starting with the maltosides, Figure 3 shows the adsorption isotherm for C10G2, both the adsorbed amount and the calculated headgroup area per surfactant. From this it is seen that the adsorption, as registered by the ellipsometer, commences at concentrations around half the cmc and that maximum adsorption is reached at concentrations slightly below the cmc of the surfactant. Above cmc no additional adsorption takes place, but the adsorption levels out at a plateau value of about 7.6 µmol/ m2, which corresponds to an apparent headgroup area of approximately 0.22 nm2. This is clearly a too-small value for being a monolayer, which indicates that at the adsorption plateau bilayer structures are formed. The average area per molecule in the bilayer is thus 0.44 nm2 in each layer. The glucoside with a C10 hydrocarbon chain, C10G1, shows different adsorption behavior compared to C10G2. At 25 °C, the reproducibility between measurements is low, as can be seen from Figure 4a. The isotherm levels out to a stable plateau for all individual measurement, but at values varying between 9.1 and 13.1 µmol/m2. However, when increasing the temperature to 30 °C, this wide distribution disappears and repeated measurements give an average plateau value of 8.8 µmol/m2, shown in Figure 4b. Combining the isotherms of C10G1 and C10G2, Figure 5, it is easily observed that the adsorption begins at a lower concentration relative to the cmc for the

Figure 5. Adsorption isotherms for C10G1 at 30 °C (filled circles) and C10G2 at 25 °C (open circles), showing the effect of headgroup polymerization. The glucoside starts adsorbing at lower surfactant concentration. The solid lines are drawn to guide the eye.

glucoside and levels out at around the same concentration, slightly below the cmc, but at a higher adsorbed amount. The plateau value corresponds to an apparent headgroup area per surfactant of 0.18 nm2, i.e., 0.36 nm2 in each layer. To investigate the effect of the alkyl chain length on adsorption, measurements with four alkyl maltosides and two alkyl glucosides with different hydrophobic chain lengths were carried out. As illustrated in Figure 6a,b, the adsorption starts around the same concentration relative to the cmc and has approximately the same equilibrium values at different concentrations for all four maltosides. The slope of the ascending section of the isotherm increases slightly with increasing chain length. The isotherms for the two glucosides also show similarities, and they also level out at the same concentration, as can be seen in Table 1, and by comparing Figure 4b and Figure


Langmuir, Vol. 20, No. 10, 2004 4055

Figure 6. Isotherms for four different alkyl maltosides C8G2 (filled diamonds), C10G2 (open circles), C12G2 (filled squares), and C14G2 (open triangles) (a, left) as a function of the concentration and (b, right) as a function of the concentration scaled with cmc. In a, the arrows indicate the cmcs of the different maltosides. For C8G2 a linear extrapolation of the slope, whose intercept with the x-axis is used as a definition of csac, is also indicated. The solid lines are drawn to guide the eye.

Figure 7. Typical adsorption isotherm of C8G1 (filled diamonds), showing the effect of small amounts of impurities in the surfactant. The arrow indicates the cmc. The solid line is drawn to guide the eye.

7. However, at concentrations slightly below the cmc, the C8G1 shows a maximum, presumably due to the presence of a low amount of impurities with higher affinity to the surface. This is confirmed by surface tension measurements, showing a slight dip in surface tension just before reaching the cmc, Figure 8a, which is not observed for the other surfactants, Figure 8b. It can be observed, by comparing Figures 8a and 7 that the minimum in the surface tension curve is smaller than the maximum in the adsorption isotherm curve; i.e., ellipsometry is more sensitive to impurities than surface tension measurements. The same observation has previously been made for impure SDS solutions.31 We suggest that the sensitivity difference is mainly due to ellipsometry data being directly proportional to the adsorbed amount, while the reduction in surface tension is given by the integral of the adsorbed amount as a function of the chemical potential of the surfactant. This sensitivity difference is also affected by the difference in the surface/volume ratio between the techniques, which for the surface tension measurements is 0.6 m-1, while being 0.4 m-1 for the ellipsometry measurements. To verify that the adsorbed amount is independent of the route of adsorption, additional measurements were made. For instance, the maximum adsorbed amount has been obtained by adding highly concentrated surfactant solutions to the cuvette and compared with the above(31) Arnebrant, T.; Ba¨ckstro¨m, K.; Jo¨nsson, B.; Nylander, T. J. Colloid Interface Sci. 1989, 128, 303.

described measurements where the bulk concentration was increased stepwise. Further, measurements where the concentrations below the cmc have been obtained from dilution of the bulk solution have been conducted. In Figure 9, a typical adsorption/desorption curve for one of the surfactants is shown and the maximum adsorbed amount reached here is the same as that obtained when determining the adsorption isotherm by the stepwise addition method. Hence, the adsorbed amount is independent of the path, as it should be for an equilibrium situation. The adsorption of C10G2 on silica surfaces and of ethylene oxide surfactants, C10EO6 and C12EO6 on titania was also measured. No adsorption of C10EO6 and C12EO6 on titania could be detected. For C10G2 on silica, Figure 10, a small adsorption, about an order of magnitude lower than on titania, could be detected. Discussion Surface Characterization. Three main conclusions can be drawn from the ESCA measurements. First, the only titanium oxide present in the upper 5-10 nm is TiO2, supported both by the location of the peaks and the Ti/O ratio. Second, in the upper layer, titanium is also present as pure Ti, indicating either that the TiO2 layer is less than 5 nm thick or that there is some Ti close to the immediate surface. Since Ti oxidizes spontaneously, the most probable reason for the peak corresponding to pure Ti is that the oxide layer is less than 5 nm. This is also in agreement with ellipsometry measurements on these surfaces, which yield an oxide thickness of 3.5-4.5 nm. Third, the cleaning process clearly removes carbon from the surfaces. The slightly rough profile observed in the AFM measurements is still smooth enough to allow for successful ellipsometry measurements using the TiO2 substrates. However, the TiO2 surface is not as smooth as the oxidized hydrophilic silica, which usually has an Ra-value of less than 0.5 nm. We note that the roughness is so high that it requires us to consider that the real surface area is significantly larger than the projected one. This has to be taken into account when evaluating the adsorbed amount per unit area. Adsorption at the Solid/Liquid Interface. Generally speaking, the isotherms on titania for all surfactants have a, for nonionic surfactants on hydrophilic surfaces, typical sigmoidal shape, with a fairly steep slope between the critical surface aggregation concentration (csac) and the cmc, which indicates cooperative adsorption. Since there is not a step change in the adsorbed amount, the adsorption

4056


Matsson et al.

Figure 8. Surface tension measurements of (a, left) C8G1 and (b, right) C10G2. The solid lines are polynomial fits of the second order of ln(c) for the surface tension, and the dotted lines represent the equilibrium surface tension above the cmc.

Figure 9. Adsorption/desorption cycle for C14G2 obtained when adding a concentrated stock solution to the cuvette to obtain a bulk concentration well above the cmc, equilibrating and finally rinsing with water.

Figure 10. Comparison of the adsorbed amount of C10G2 on titania (open circles) with that on silica (filled circles), showing a difference of around 1 order of magnitude in adsorption. The solid lines are drawn to guide the eye.

process cannot be considered as a phase separation, but the growth of the adsorbed layer is continuous. This means that the bilayer thickness increases, or that the surface micelle structure changes during the adsorption process. That no significant adsorption for the maltosides is registered in the low adsorption region does not automatically give the conclusion that no adsorption occurs at these concentrations, but it can also be attributed to that the relative error of the ellipsometry measurements increases with decreasing registered adsorbed amounts; e.g. a measured value of Γ ) 0.2 µmol/m2 would have a relative error of around 15%.32

Figure 11. Adsorption isotherms for C12G2 at 25 (open squares) and 30 °C (filled diamonds), showing the temperature dependence for the maltosides. The solid line is drawn to guide the eye.

For C10G1, the first measurements at 25 °C were performed as for the other surfactants, by adding a small amount of concentrated solution to the cuvette. However, nonreproducible results were obtained. Phase studies have shown that this particular surfactant has a two-phase region between 0.1 and 19 wt% of the surfactant in water (cmc ) 0.07 wt%),5 so the concentrated stock solution would then be in the two-phase region. The first measure taken was to change the addition process to pumping solution through the cuvette at the desired concentrations. This did nevertheless not change the large distribution between measurements. Another study of the surfactant bulk properties shows that as the temperature decreases to below room temperature, C10G1 crystallizes in water.3 Therefore adsorption measurements at higher temperatures were performed, assuming that if the reason for the unreliable results obtained at 25 °C was surfactant crystals adsorbing at the interface, the reproducibility would be significantly better at 30 °C. This was, as seen above, also the case. To verify that increasing the temperature would not give other effects as well, the isotherm for C12G2 was also measured at 30 °C. For this surfactant, no significant difference in adsorbed amounts between the temperatures is observed, as can be seen in Figure 11. In particular it should be noted that, unlike C10G1, C14G2 gives reproducible results, though it also is measured below its Krafft point. This is most likely due to the very slow rate of crystallization for this surfactant. Since the conversion into crystals in solution was observed to take roughly 2 (32) Tiberg, F. J. Chem. Soc., Faraday Trans. 1996, 92, 531.


Langmuir, Vol. 20, No. 10, 2004 4057

Table 2. Measures of Cooperativity for the Alkyl Maltosides (csac Estimated from Adsorption Isotherms) surfactant C8G2 C10G2 C12G2 C14G2

cmc (mM) 19a 2.0b 0.15c 0.015d

csac (mM)

∆µ0 (kT)

7.5 0.92 0.076 0.011

-0.93 -0.78 -0.67 -0.35

a The cmc values are from reference 3. b Reference 12. c Reference 9. d Reference 8.

days, reproducible results could be obtained from freshly prepared solutions. The observation that the difference between the maximum adsorbed amount within the groups of maltosides and glucosides respectively is minimal leads to the conclusion that the maximal adsorption of alkyl polyglucosides is independent of the alkyl chain length. The headgroup polymerization on the other hand seems to affect the adsorption process and in particular the plateau adsorbed amount significantly, indicating that the maximum adsorption is limited by the size of the headgroup. The alkyl chain length does nevertheless affect the slope of the isotherms, and from the adsorption isotherms, some measure of cooperativity can be obtained. The free energy difference between a surfactant in a micellar-like aggregate at the surface and a surfactant in a free micelle can be estimated as

µ0ads - µ0mic ) kT ln

(csac cmc )

(5)

To estimate csac, the slope of the ascending section of the isotherm is extrapolated to a straight line; see the isotherm for C8G2 in Figure 6a. Then the csac is defined as the concentration at the intercept with the x-axis. The estimated csacs and calculated ∆µ0 are listed in Table 2. Csac is 0.4-0.7 cmc, and the free energy difference is 0.350.93 kT; hence, there is a rather weak interaction between the surfactant molecules and the surface. The magnitude of ∆µ0, and thereby the extra driving force for selfassemblying at the surface rather than in bulk solution, is larger for the short-chain surfactants. A plausible explanation is that the hydrophobic interaction between the tails increases with the surfactant chain length, and thus the extra contribution due to interactions between the headgroup and the surface becomes less noticeable. These measures can be compared with the interaction between EO surfactants and silica surfaces, which amounts to

Water Interface

Recommend Documents