Molecular Macrocluster Formation on Silica Surfaces in Phenol

(3) Korn, M.; Killmann, E.; Eisenlauer, J. J. Colloid Interface Sci. 1980, 76,. 7-18. ..... 0.5 mol % phenol and the essential role of the silanol gro...
0 downloads 0 Views 190KB Size
Download PDF
6070

Langmuir 2007, 23, 6070-6075

Molecular Macrocluster Formation on Silica Surfaces in Phenol-Cyclohexane Mixtures Neval Yilmaz, Masashi Mizukami, and Kazue Kurihara* Institute of Multidisciplinary Research for AdVanced Materials (IMRAM), Tohoku UniVersity, 2-1-1 Katahira, Aobaku, Sendai 980-8577, Japan ReceiVed January 6, 2007. In Final Form: March 7, 2007 The adsorption of phenol, an aromatic compound with a hydrogen-bonding group, onto a silica surface in cyclohexane was investigated by colloidal probe atomic force microscopy (AFM), attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR), and adsorption isotherm measurements. ATR-FTIR measurements on the silica surface indicated the formation of surface macroclusters of phenol through hydrogen bonding. The ATR-FTIR spectra were also measured on the H-terminated silicon surface to observe the effect of the silanol groups on the phenol adsorption. The comparison of the ATR-FTIR spectra for both the silicon oxide and H-terminated silicon surfaces proved that the silanol groups are necessary for the formation of phenol clusters on the surface. The surface force measurement using colloidal probe AFM showed a long-range attraction between the two silica surfaces in phenolcyclohexane mixtures. This long-range attraction resulted from the contact of the adsorbed phenol layers for the phenol concentrations below 0.6 mol %, at which no significant phenol clusters formed in the bulk solution. The attraction started to decrease at 0.6 mol % phenol due to the exchange of the phenol molecules between the clusters in the bulk phase and on the surface. The surface density of phenol in the adsorbed layer was calculated on the basis of the long-range attraction and found to be much smaller than the liquid phenol density. The plausible structure of the adsorbed phenol layer was drawn by referring to the crystal structure of the bulk phenol and orientation of the phenol molecules on the surface, estimated by the dichroic analysis of ATR-FTIR spectroscopy. The investigation of the phenol adsorption on the silica surface in a nonpolar solvent using this novel approach demonstrated the effect of the aromatic ring on the surface packing density.

Introduction It is known that the adsorption at the solid-liquid interface has a significant impact in many technological, environmental, and biological applications such as detergency, lubrication, paints, adhesives, and many more which involve the stability of colloidal dispersions and modification of the solid surfaces for the control of wetting. Therefore, it is important to understand the phenomena taking place at the solid-liquid interface. Adsorption from a solution onto solid surfaces concerning binary liquids is generally studied on the basis of the adsorption excess isotherm and microcalorimetry and the stability of colloidal dispersions.1-5 For example, in the studies by Vincent et al.4 and Dekany et al.5 the stability of the silica dispersions was examined in ethanolcyclohexane mixtures by rheology measurements, and the van der Waals interaction pair potential for two spherical particles surrounded by an adsorption layer was calculated to determine the interaction between the silica particles. It was found that, in mixtures rich in cyclohexane, the silica particles are in an aggregated state, while in ethanol or in mixtures rich in ethanol suspensions, they are stable due to the decreasing attraction between the particles. However, it is difficult to understand the behavior and the structure of the adsorbate on the surface at the molecular level by classical methods. Recently, a novel approach, * To whom correspondence should be addressed. E-mail: kurihara@ tagen.tohoku.ac.jp. (1) Dawidowicz, A. L.; Wianowska, D.; Patrykiejew, A. J. Colloid Interface Sci. 2003, 258, 213-218. (2) Dekany, I.; Farkas, A.; Kiraly, Z.; Klumpp, E.; Narres, H. D. Colloids Surf., A 1996, 119, 7-13. (3) Korn, M.; Killmann, E.; Eisenlauer, J. J. Colloid Interface Sci. 1980, 76, 7-18. (4) Vincent, B.; Kiraly, Z.; Emmett, S.; Beaver, A. Colloids Surf. 1990, 49, 121-132. (5) Dekany, I.; Haraszti, T.; Turi, L.; Kiraly, Z. Prog. Colloid Polym. Sci. 1998, 111, 65-73.

which is a combination of colloidal probe atomic force microscopy (AFM), Fourier transform infrared spectroscopy in the attenuated total reflection mode (ATR-FTIR), and adsorption isotherm measurements, was employed for studying the adsorption phenomena of some aliphatic compounds such as ethanol, methanol, 1-propanol, 2-propanol, carboxylic acid, and ethylene glycol onto a silica surface in cyclohexane.6-10 Based on these measurements, we found organized structures, which we call molecular macroclusters, of alcohols and acids through hydrogen bonding on the silica surface. The major properties of the molecular macroclusters are summarized taking the case of ethanol as an example:6 (1) The ATR-FTIR measurement indicated the necessity of the silanol groups for the formation of macroclusters on the surface. (2) At 0.1 mol % ethanol, a long-range attraction was observed between the silica surfaces in cyclohexane at a distance of 35 ( 3 nm. Half the attraction range agreed with the adsorbed layer thickness at this concentration. This is support for attributing the long-range attraction to the contact of the opposing adsorbed layers. This long-range attraction remains basically the same at ethanol concentrations below 0.5 mol %. (3) When the ethanol concentration was increased beyond 0.5 mol %, at which the hydrogen-bonded clusters of ethanol began to form in the bulk phase, half the attraction range decreased while the adsorbed layer thickness remained constant. When clusters form both in the solution and on the surface, ethanol molecules should have similar affinities for both types of clusters, which leads to the exchange of ethanol (6) Mizukami, M.; Moteki, M.; Kurihara, K. J. Am. Chem. Soc. 2002, 124, 12889-12897. (7) Mizukami, M.; Nakagawa, Y.; Kurihara, K. Langmuir 2005, 21, 94029405. (8) Mizukami, M.; Kurihara, K. Aust. J. Chem. 2003, 56, 1071-1080. (9) Mizukami, M.; Kurihara, K. Proc. Jpn. Acad. 2001, 77B, 115-120. (10) Kurihara, K.; Nakagawa, Y.; Mizukami, M. Chem. Lett. 2003, 32, 8485.

10.1021/la0700366 CCC: $37.00 © 2007 American Chemical Society Published on Web 04/18/2007

Molecular Macrocluster Formation on Silica

Langmuir, Vol. 23, No. 11, 2007 6071

molecules between the clusters on the surface and in the bulk phase. This results in a decrease in the interfacial energy and thus in the attraction range. Similar results were obtained for the adsorption of methanol and 1-propanol in cyclohexane.7,8 By using this novel method, the interaction between silica particles in a binary liquid and the structure of the adsorbed layer could be directly demonstrated. It would then be interesting to know how the layer structures and interactions are correlated with the chemical structures of the adsorbate. In this study, we investigated the adsorption of an aromatic compound, phenol, on a silica surface in cyclohexane using our novel approach. We found hydrogen-bonded macrocluster formation of phenol on the silica surface and demonstrated the effect of the aromatic ring on the adsorbed layer structure. In addition, by comparison of this study with the previous studies,6-10 the difference between the packings of the aliphatic and aromatic compounds on the surface based on the adsorbed layer density was revealed. Materials and Methods Materials. Phenol and cyclohexane with a purity of 99.5% were supplied from Nacalai Tesque. Phenol was distilled over MgSO4 under a reduced pressure (∼20 mmHg) in a nitrogen atmosphere and stored in a desiccator over silica gel in the dark. Cyclohexane was dried over sodium and distilled under atmospheric conditions just prior to use. Surface Force Measurement. A colloidal glass sphere (Polysciences, 10-30 µm in diameter), attached to the end of a rectangular cantilever (Olympus, RC-800PS, 40 µm in width and 100 µm in length) with epoxy resin, and a glass plate (Matsunami, micro cover glass) were used to examine the interaction between the two glass surfaces in phenol-cyclohexane mixtures using AFM (Seiko II, SPI3700-SPA400). The approaching speed of the colloidal probe to the substrate was 416 nm/s. The glass spheres and glass plates were cleaned in a 4:1 (v/v) sulfuric acid-hydrogen peroxide mixture, which promotes the growth of surface silanol groups,11 rinsed with pure water, and then treated with water vapor plasma for 3 min just before each measurement to ensure the full existence of silanol groups on the surface. Infrared Spectroscopy in the Transmission and Attenuated Total Reflection Modes. The FTIR spectra were recorded using a Perkin-Elmer FTIR system 2000 with a TGS detector. The transmission infrared spectra were obtained using a CaF2 cell (Nihon Bunko) with an optical path length of 25 µm, and the ATR-FTIR spectra were measured on a silicon prism (Nihon PASTEC, 60 × 16 × 4 mm trapezoid) using the ATR attachment from Grasby Specac. The oxide layer formed on the silicon prism surface was used as an adsorbent. The silicon prism was cleaned in the same way as the glass spheres and glass plate, used for the surface force measurement, and treated with water vapor plasma for 20 min. The infrared light was reflected six times with an incident angle of 45°. To investigate the mean orientation of the adsorbed phenol, the dichroic analysis of ATR-FTIR spectra was performed. The integrated peak intensities of the OH and aromatic CdC stretching absorptions were calculated for the dichroic analysis, and the orientation of the adsorbed phenol was estimated by applying a three-layer model (silicon prism, adsorption layer, and bulk solution).6 The refractive indices used for the three-layer model are 3.42, 1.55, and 1.42 for the silicon prism, adsorption layer (phenol), and bulk solution (cyclohexane), respectively. The ATR-FTIR spectra were also measured on a H-terminated silicon surface, which was prepared by immersing the ATR prism in a 0.5% (v/v) hydrofluoric acid aqueous solution for 30 min and then rinsing it with pure water. Adsorption Excess Isotherm Measurement. For the adsorption isotherm measurement, 0.5 g of glass spheres (Polysciences, 2-5 µm in diameter), treated similarly to the silicon oxide surface, was dispersed in a phenol-cyclohexane solution and equilibrated for 24 h. The solution concentration after adsorption was determined using (11) Frantz, P.; Granick, S. Langmuir 1992, 8, 1176-1182.

Figure 1. Force profiles of interaction between glass surfaces upon compression in phenol-cyclohexane mixtures at 0.0, 0.4, 0.5, 0.6, and 1.0 mol % phenol. The dashed line represents the van der Waals force calculated using the nonretarded Hamaker constant for silica/ cyclohexane/silica (3.64 × 10-21 J). a differential refractometer (Otsuka Electronics, DRM-1021). The equilibrium concentration was used to calculate the adsorption layer thickness (t) on the basis of the following equation: t ) n1σVm,1/a

(1)

where n1σ, Vm,1, and a are the adsorbed amount per gram of phenol, molar volume of phenol (0.145 nm3/molecule at 20 °C), and specific surface area of the glass spheres (0.56 m2/g), respectively. The specific surface area was calculated by determining the size distribution of the glass spheres using an optical microscope equipped with a CCD camera.

Results and Discussion Surface Force Measurement. In Figure 1, the interaction forces measured between two silica surfaces in phenolcyclohexane mixtures at various concentrations are presented together with the van der Waals force calculated for silica/ cyclohexane/silica using the Hamaker constant, 3.64 × 10-21 J.12 In pure cyclohexane, the observed attraction range agreed with the calculated van der Waals attraction. The addition of phenol changed the range and the force of attraction. At 0.3 mol % phenol, there is a slight increase in the attraction range. At 0.4 and 0.5 mol % phenol, long-range attractions extending to 20 ( 3 and 21 ( 4 nm were observed, respectively. The interaction between the silica surfaces at 0.4 and 0.5 mol % phenol turned into repulsion at a distance of around 2 nm. Beyond 0.5 mol % phenol, the attraction range showed a remarkable decrease, and at 1 mol % phenol, the range of attraction was 6 ( 1 nm. Due to the adsorption of phenol on the surface, the repulsion at a distance of around 2 nm at 0.4 and 0.5 mol % phenol can be ascribed to the steric effect arising from the overlap of the tightly adsorbed layers of phenol. The decrease in the attraction range will be explained later. In Figure 2, the changes in the attraction range and adhesion force are plotted as a function of the phenol concentration. The adhesion force showed almost the same tendency as the attraction range, which was similar to that in the case of ethanol adsorption from its mixture with cyclohexane onto silica.6 Considering the profile of long-range force at 0.4 and 0.5 mol % phenol as well (12) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: London, 1991.

6072 Langmuir, Vol. 23, No. 11, 2007

Yilmaz et al.

Figure 2. Attraction range and adhesion force plots for silica surfaces in phenol-cyclohexane mixtures at 0.0, 0.3, 0.4, 0.5, 0.6, 0.7, and 1.0 mol % phenol.

as the concentration dependencies of the attraction range and adhesion force, the long-range attraction can be attributed to the contact of the adsorbed layers of phenol on the opposing silica surfaces. It is possible to analyze the long-range attraction and obtain the interfacial energy between the adsorbed layer and the bulk solution. Previously, the long-range attraction caused by the contact of the adsorbed methanol layers on the silica surface in cyclohexane was fit using eq 2, and the interfacial energy was

2πγ[(2t - D)(2R + D)] F )R R

1/2

( DB)

+ A exp -

(2)

found to be around 7 mN/m.13 Here, the first and second terms are ascribed to the attractive and steric forces, respectively. F, R, γ, t, D, A, and B are the force, radius of the sphere, interfacial energy, adsorbed layer thickness, distance, strength factor, and decay length of the solvation force, respectively. For the case of phenol, the long-range attraction was analyzed using the same equation and the force curve could be well reproduced at 0.4 and 0.5 mol % phenol. This confirms the origin of the long-range attraction as the contact of the adsorbed phenol layers. On the basis of this analysis, the interfacial energy between the adsorbed phenol layer and the bulk solution was evaluated to be 3.6 mN/ m. Considering the molecular size, the difference between the interfacial energies of methanol and phenol is reasonable, since the density of the free hydroxyl group at the interface must be lower in the case of phenol. Infrared Spectroscopy in the Attenuated Total Reflection Mode. To characterize the phenol molecules adsorbed on the silica surface and the structure of the adsorption layer, ATRFTIR spectra were measured on a silicon oxide surface in phenolcyclohexane mixtures (Figure 3). The peaks at 3045 and 3078 cm-1 are assigned to the CH stretching vibrations of the phenol molecules.14 The peak at 3618 cm-1 is due to the OH stretching vibration of the phenol monomer (free OH).14,15 The peaks at 3597, 3627, 3700, and 3729 cm-1 could be ascribed to the presence of water vapor in the air. The parallel change in intensities of these four peaks for different measurements supports this (13) Mizukami, M.; Kurihara, K. e-J. Surf. Sci. Nanotechnol. 2006, 4, 244248. (14) Marshall, K.; Rochester, C. H. J. Chem. Soc., Faraday Trans. 1 1975, 71, 2478-2484. (15) Bellamy, L. J. The Infrared Spectra of Complex Molecules, 2nd ed.; Wiley and Sons: New York, 1958.

Figure 3. ATR-FTIR spectra measured on a silicon oxide surface in phenol-cyclohexane mixtures at (a) 0.0, (b) 0.1, (c) 0.3, (d) 0.5, (e) 1.0, and (f) 2.0 mol % phenol.

assignment. The band at 3340 cm-1, ascribed to the polymer OH,15 appeared at 0.1 mol % phenol. The FTIR transmission measurement (data not shown) indicated that phenol began forming clusters in the bulk solution at 0.5 mol % phenol. Therefore, the polymer OH absorption at concentrations lower than 0.5 mol % phenol could be ascribed to the hydrogen-bonded clusters of the phenol adsorbed on the silica surface. It is known that the OH group can interact with the aromatic ring, which gives a peak between the free OH and polymer OH bands.16,17 For the adsorption of 4-methylphenol and 2,6-dimethylphenol on silica in heptane16 and the benzene-water cluster,17 OH-π interactions were observed at 3595 and 3650 cm-1, respectively, which were located between the absorption bands assigned to the free OH (3621-3630 cm-1,16 3713 cm-1 17) and the OHOH interaction (3370-3440 cm-1,16 3550 cm-1 17). In our spectrum, there is no indication of any hydrogen bond interaction of the hydroxyl groups with the aromatic π-electron systems of the phenol molecules between the bands assigned to the free OH (3618 cm-1) and the polymer OH (3340 cm-1). Therefore, phenol must be forming the hydrogen bond primarily with the silanol groups and OH groups of other phenol molecules. The ATR-FTIR spectra were also measured on a H-terminated silicon surface (Figure 4) to investigate the effect of the surface silanol groups on the formation of surface clusters. The spectra measured on the H-terminated silicon surface exhibited a negligible polymer OH absorption at 0.1 and 0.3 mol % phenol. The polymer OH peak started to appear at 0.5 mol % phenol at 3346 cm-1 as a result of cluster formation in the bulk solution. The comparison of the spectra measured on the silicon oxide and H-terminated silicon surfaces indicated that the silanol groups are essential for the formation of clusters on the surface. In Figure 5, the integrated peak intensities of the ATR-FTIR spectra measured on both the silicon oxide (AATR-OH) and H-terminated silicon (AATR-H) surfaces and those of the transmission spectra (ATS) are presented as a function of the phenol concentration. The change in AATR-H showed a tendency similar to that of the change in ATS. This indicated that the polymer OH, which started to appear at 0.5 mol % phenol for the spectra measured on a H-terminated silicon surface, resulted from the clusters formed (16) Rochester, C. H.; Trebilco, D-A. J. Chem. Soc., Faraday Trans. 1 1978, 74, 1137-1145. (17) Gruenloh, C. J.; Carney, J. R.; Arrington, C. A.; Zwier, T. S.; Fredericks, S. Y.; Jordan, K. D. Science 1997, 276, 1678-1681.

Molecular Macrocluster Formation on Silica

Langmuir, Vol. 23, No. 11, 2007 6073

Figure 6. Comparison of the phenol concentration dependence of AATR-OH - AATR-H to the adsorbed layer thickness.

Figure 4. ATR-FTIR spectra measured on a H-terminated silicon surface in phenol-cyclohexane mixtures at (a) 0.0, (b) 0.1, (c) 0.3, (d) 0.5, (e) 1.0, and (f) 2.0 mol % phenol.

Figure 5. Phenol concentration dependencies of the integrated peak intensity of polymer OH absorption obtained in ATR spectra measured on a silicon oxide surface (AATR-OH) ([) and a H-terminated silicon surface (AATR-H) (b) and in transmission spectra (ATS) (2).

in the bulk solution. As noted, there is a sharp increase in the integrated peak intensity of the polymer OH obtained for the silicon oxide surface (AATR-OH) up to 0.5 mol % phenol, and beyond 0.5 mol % phenol, the increases in AATR-OH and AATR-H showed a parallel tendency due to the contribution of the clusters formed in the bulk solution. The comparison of the change in AATR-OH to the change in AATR-H and ATS provided a clearer understanding of the cluster formation on the silica surface below 0.5 mol % phenol and the essential role of the silanol groups in the surface cluster formation. Adsorption Excess Isotherm. The difference between the integrated peak intensities (AATR-OH - AATR-H), which corresponds to the adsorbed amount, was compared in Figure 6 with the adsorbed amount of phenol, obtained from adsorption excess isotherm measurement. The plot of AATR-OH - AATR-H versus the concentration showed the same tendency as the plot of the adsorbed amount. This shows that the adsorption phenomenon is independent of the size and shape of the substrate. The same tendency for the adsorption isotherm is observed for both the flat surface and colloidal particles. Since the ATR-FTIR spectra measured on the silicon oxide surface demonstrated that the adsorption in phenol-cyclohexane mixtures mainly takes place through the hydrogen bonding between the OH groups of phenol and of the silica surface as well as between the phenol molecules, the same tendency between the plots of AATR-OH - AATR-H and the adsorbed amount versus the concentration emphasizes the dominant contribution of the OH-OH interaction.

Figure 7. Half the attraction range (-b-) and the adsorbed layer thickness estimated using the adsorbed layer density (0.48 g/cm3) (--[--) as a function of the phenol concentration.

The adsorption isotherm measurement showed that the adsorption reached saturation at 0.8 mol % phenol. The previous studies of ethanol, methanol, and 1-propanol adsorption on silica surfaces has shown that the estimated adsorbed layer thickness agrees with half the attraction range at low concentrations of alcohol where there is no cluster formation in the bulk solution.6-8 On the basis of these results, the origin of the long-range attraction was attributed to the contact of the adsorbed layers of alcohol, and the adsorbed layer density was found to be almost the same as that in the liquid phase. This conclusion was also confirmed by the analysis of the force curve based on the interfacial energy.13 Therefore, the adsorbed layer density was calculated considering half the attraction range as the real adsorbed layer thickness at 0.4 and 0.5 mol % phenol.18 From this calculation the density of the adsorbed layer was found to be around 0.48 g/cm3, which is much smaller than the liquid phenol density (1.06 g/cm3). This can be related to the rigid structure of phenol, which might prevent its packing on the surface as densely as in the bulk phenol due to steric effects, in contrast to simple alcohols such as ethanol, methanol, and 1-propanol.6-8 Therefore, the lower packing density of the phenol clusters on the surface than that in the bulk phenol leads to a higher adsorbed layer thickness than that calculated using the liquid phenol density. In Figure 7, the adsorbed layer thickness calculated using the adsorbed layer density (0.48 g/cm3) is plotted together with half the attraction range. Although the adsorbed layer thickness continued increasing up to 0.8 mol % phenol and remained constant above this concentration, half the attraction range started to decrease beyond 0.5 mol % phenol. The decrease in the attraction range can be

6074 Langmuir, Vol. 23, No. 11, 2007

Figure 8. Dichroic ratio obtained from the ATR-FTIR spectra measured on a silica surface as a function of the phenol concentration for (a) polymer OH at 3340 cm-1 and (b) the phenyl ring at 1598 cm-1. (The dashed line represents the calculated Value for the random orientation.)

explained by the decrease in the interfacial energy resulting from the exchange of solute molecules between the adsorbed layer and clusters in the bulk solution.6,19 Dichroic Analysis and a Plausible Structure of the Adsorbed Phenol Layer. The mean orientation of the adsorbed phenol was investigated on the basis of the dichroic analysis applied for the ATR-FTIR spectra measured on the silica surface using s- and p-polarized light. The absorptions for the polymer OH at 3340 cm-1 and phenyl ring at 1598 and 1498 cm-1, corresponding to the CdC stretching vibration, measured using the s-polarized light were far less intense than those obtained using the p-polarized light. The dichroic ratios, Ap/As, were calculated for the polymer OH (3340 cm-1) and the phenyl ring (1598 cm-1) and are plotted in Figure 8 as a function of phenol concentration. The dichroic ratios for the polymer OH and the phenyl ring absorptions were significantly larger than that expected for a random orientation (1.21), shown by the dashed line in Figure 8, at phenol

Yilmaz et al.

concentrations below 0.5 mol %. This indicated that the adsorbed phenol has a preferential orientation at concentrations where there is no contribution from the bulk solution. The dichroic ratio showed different tendencies for the orientation of the polymer OH and the phenyl ring at low concentrations. While the former showed a maximum, the latter decreased monotonously. This might be due to the change in the orientation of the hydrogen bond and the phenyl ring with the increasing adsorbed amount. The maximum of the ratio at 0.3 mol % phenol indicated that the angle between the hydrogen bond and the chain axis at 0.1 mol % phenol, 50°, was larger than that at 0.3 mol % phenol, 35°. In contrast to the polymer OH, the phenyl ring changed its orientation angle from smaller (29°) to larger (40°) upon elongation of the macrocluster. For both the polymer OH and phenyl ring, at concentrations above 2 mol % phenol, the dichroic ratio decreased and became similar to that for the random orientation. This can be explained by the dominant contribution of the phenol clusters formed in the bulk solution because of the long penetration depth of an evanescent wave, which is 270 nm at 3340 cm-1 and 560 nm at 1598 cm-1. However, the dichroic ratio remained higher than that for the random orientation (1.21) even above 2 mol % phenol. This may indicate that the adsorbed phenol molecules maintained some of their orientation above 2 mol % phenol. Using the dichroic ratios obtained at 0.3 mol % phenol, the mean orientation angles of the OH group and phenyl ring from the surface normal were calculated to be 35 ( 3° and 40 ( 7°, respectively. A plausible structure of the adsorbed phenol layer, which was drawn by ChemBats3D on the basis of the crystal structure of phenol20 and the orientation angles found from the dichroic analysis, is shown in Figure 9. The drawn structure of the adsorbed phenol layer on the silica surface has mean orientation angles of 41° for the OH group and 53° for the phenyl ring. There is a free (non-hydrogen-bonded) OH group at the end of this phenol cluster, i.e., the interface between the adsorbed layer and bulk solution. This free OH group at the interface can be the cause of the interfacial energy and result in an attraction when the adsorbed layers contact each other. On the basis of the density of the adsorbed layer calculated from half the attraction range, the plausible structure of the two phenol clusters on the surface was drawn as shown on the right-hand side of Figure 9. In this structure, three phenol molecules from the top view were fit into a circle and the circles were intersected to pack them sparsely with a density of 0.48 g/cm3.

Figure 9. A plausible structure of the adsorbed layer of phenol on the silica surface.

Molecular Macrocluster Formation on Silica

Conclusions The adsorption of phenol on a silica surface was investigated in cyclohexane using a combination of ATR-FTIR, colloidal probe AFM, and adsorption isotherm measurements. The results obtained from these measurements are summarized below. (1) Phenol formed macroclusters on the silica surface in cyclohexane primarily through hydrogen bonding between the hydroxyl groups of the silica surface and phenol molecules. (2) A long-range attraction (∼20 nm) was observed between the silica surfaces in cyclohexane at 0.4 and 0.5 mol % phenol, resulting from the contact of the adsorbed phenol layers on opposing silica surfaces. (3) The clear understanding of the phenol adsorption on the silica surface in cyclohexane revealed an important difference in the adsorbed state between the aliphatic compounds studied with the same method6-10 and an aromatic compound, phenol. While aliphatic compounds pack on the solid surface as densely as in the bulk, packing of the phenol was much less dense than that in the bulk. This might be useful information for applications requiring an ordered structure and less dense packing on the surface than that in the bulk liquid or solid rather than a packing with the same density in the bulk.

Langmuir, Vol. 23, No. 11, 2007 6075

It is interesting to learn how differently the molecules can adsorb on the silica surfaces. The molecular level of study on liquid adsorption has just started. More efforts are needed to explore this rich field of interface chemistry. Acknowledgment. This work was supported by the CREST program of the Japan Science and Technology Agency (JST). LA0700366 (18) We may need to consider the effect of the approaching colloidal probe on enhancement of the macrocluster formation on the silica surface. We conclude that this effect should not be significant, and half the attraction range measured at low concentrations reasonably provides the adsorbed layer thickness on the basis of the following reasons: (1) There is significant cluster formation of phenol on the free surface as shown by ATR-FTIR data measured on silica surface (Figure 3). The adsorbed amount determined by ATR-FTIR showed a tendency similar to that of the adsorbed layer thickness obtained from the adsorption isotherm measurement (Figure 6). This indicates that the adsorption phenomenon is independent of the size and shape of the substrate. (2) Assuming half the attraction range as the adsorbed layer thickness, the force curve could be reproduced by eq 2. This shows that the adsorbed layer thickness does not change on approach. (3) As mentioned in the main text, the previous studies of the simple alcohols such as ethanol have shown that the estimated adsorbed layer thickness agrees well with half the attraction range at low concentrations of alcohol. This also shows that the adsorbed layer thickness does not change on approach. (19) Nakagawa, N.; Endo, S.; Mizukami, M.; Kurihara, K. Trans. Mater. Res. Soc. Jpn. 2005, 30, 667-670. (20) Allan, D. R.; Clark, S. J.; Dawson, A.; McGregor, P. A.; Parsons, S. Acta Crystallogr., Sect. B 2002, 58, 1018-1024.