Structure of Poly (methyl methacrylate) Chains Adsorbed on Sapphire

Langmuir 2004, 20, 7183-7188

7183

Structure of Poly(methyl methacrylate) Chains Adsorbed on Sapphire Probed Using Infrared-Visible Sum Frequency Generation Spectroscopy Ashwin Rao, Hasnain Rangwalla, Vikas Varshney, and Ali Dhinojwala* Department of Polymer Science, The University of Akron, Akron, Ohio 44325 Received March 6, 2004. In Final Form: May 24, 2004 We have studied the orientation of the train segments of a poly(methyl methacrylate) (PMMA) adsorbed layer at the CCl4-sapphire interface using surface-sensitive IR-visible sum frequency generation (SFG) spectroscopy. The SFG spectra of PMMA chains adsorbed on sapphire indicate ordered ester methyl groups. In comparison, we did not observe any significant contributions from the backbone methylene and alpha methyl groups, suggesting that these groups are disordered. No change in the structure of the adsorbed layer is observed upon cooling the solvent below the theta temperature; this is consistent with the picture of flat adsorbed chains on the surface. Interestingly, the orientation of the ester methyl groups of a spincoated PMMA film at the PMMA-sapphire interface is similar to that of the same groups in the chains adsorbed from solution.

Introduction Polymer adsorption has been the subject of numerous theoretical and experimental investigations.1 The structure of an adsorbed chain can be described in terms of “loops”, “tails”, and “trains”. In such a description, the “loops” and “tails” of the adsorbed chain are extended in the solution, while the train segments are in contact with the surface and the interactions between the train segments and the surface provide the driving force for the adsorption process. For example, as a result of strong interactions between the ester side groups of poly(methyl methacrylate) (PMMA) and surface hydroxyl groups, PMMA adsorbs very strongly on aluminum and silicon substrates.2 Moreover, theoretical calculations of the binding energy of PMMA on these substrates has been found to be a function of the conformation of the polymer segments and the orientation of the ester side groups in train segments of the adsorbed chain.2 Although there have been many studies of PMMA adsorption,2-14 there have been few measurements of the local structure and * To whom correspondence should be addressed. E-mail: ali4@ uakron.edu. (1) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Chapman and Hall: New York, 1993. (2) Shaffer, J. S.; Chakraborty, A. K.; Tirrell, M.; Davies, H. T.; Ted, H.; Martins, J. L. J. Chem. Phys. 1991, 95, 8616. (3) Frantz, P.; Granick, S. Macromolecules 1995, 28, 6915. (4) Soga, I.; Granick, S. Colloids Surf., A 2000, 170, 113. (5) Caucheteux, I.; Hervet, H.; Jerome, R.; Rondelez, F. J. Chem. Soc., Faraday Trans. 1990, 86, 1369. (6) Sakai, H.; Fujimori, T.; Imamura, Y. Bull. Chem. Soc. Jpn. 1980, 53, 3457. (7) Konstadinidis, K.; Thakkar, B.; Chakraborty, A.; Potts, L. W.; Tannenbaum, R.; Tirrell, M.; Evans, J. F. Langmuir 1992, 8, 1307. (8) Watts, J. F.; Leadley, S. R.; Castle, J. E.; Blomfield, C. J. Langmuir 2000, 16, 2292. (9) Carriere, P.; Grohens, Y.; Spevacek, J.; Schultz, J. Langmuir 2000, 16, 5051. (10) Papirer, E.; Perrin, J.-M.; Nanse, G.; Fioux, P. Eur. Polym. J. 1994, 30, 985. (11) Grohens, Y.; Brogly, M.; Labbe, C.; Schultz, J. Eur. Polym. J. 1997, 33, 691. (12) Grohens, Y.; Schultz, J.; Prud’homme, R. E. Int. J. Adhes. Adhes. 1997, 17, 163. (13) Durning, C. J.; O’Shaughnessy, B.; Sawhney, U.; Nguyen, D.; Majewski, J.; Smith, G. S. Macromolecules 1999, 32, 6772. (14) Lin, W.-Y.; Blum, F. D. Macromolecules 1997, 30, 5331.

orientation between the surface and the train segments of the adsorbed layer.3,4 In the case of PMMA adsorption on silicon substrates, hydrogen bonding interactions between the carbonyl groups of the polymer side chain and the surface hydroxyl groups cause a shift in the IR vibrational peak of the bound carbonyl groups.3,4 By resolving this shift in the IR dichroism spectrum, the bound carbonyl groups of PMMA adsorbed on silicon were selectively observed, and it was found that the orientation of this group is isotropic. In comparison, IR dichroism measurements of the asymmetric methyl stretch vibration showed a large directional preference for the orientation of methyl groups, whereas the same measurements for the symmetric methyl stretch vibration indicated an isotropic orientation. The opposite results for the same group were puzzling, and the orientation of the methyl groups was unresolved. This was due to the overlap of vibrational bands for ester methyl, alpha methyl, and backbone methylene groups. Furthermore, IR spectroscopy lacks the sensitivity to distinguish the methyl and methylene groups from those in the loops and tails. In this paper, we have employed surface-sensitive IRvisible sum frequency generation (SFG) spectroscopy15,16 to study the orientation of PMMA train segments at the CCl4-sapphire interface. SFG spectroscopy, which has been successfully used to study the structure of polymers at various interfaces,17-21 exploits the fact that the centrosymmetry in the bulk of most substances is necessarily broken at surfaces and interfaces. Within the electric-dipole approximation, centrosymmetric media are SFG-inactive; therefore, no SFG signal is generated in the bulk of such materials. However, the surface and (15) Shen, Y. R. Nature 1989, 337, 519. (16) Bain, C. D. J. Chem. Soc., Faraday Trans. 1995, 91, 1281. (17) Gautam, K. S.; Schwab, A. D.; Dhinojwala, A.; Zhang, D.; Dougal, S. M.; Yeganeh, M. S. Phys. Rev. Lett. 2000, 85, 3854. (18) Wang, J.; Chen, C.; Buck, S. M.; Chen, Z. J. Phys. Chem. B 2001, 105, 12118. (19) Harp, G. P.; Rangwalla, H.; Yeganeh, M. S.; Dhinojwala, A. J. Am. Chem. Soc. 2003, 125, 11283. (20) Chen, C.; Even, M. A.; Wang, J.; Chen, Z. Macromolecules 2002, 35, 9130. (21) Kim, J.; Kim, G.; Cremer, P. S. J. Am. Chem. Soc. 2002, 124, 8751.

10.1021/la049413u CCC: $27.50 © 2004 American Chemical Society Published on Web 07/16/2004

7184

Langmuir, Vol. 20, No. 17, 2004

Rao et al.

interfaces of such a material are SFG-active. Particularly, the SFG signal from surfaces and interfaces is greatly enhanced when the molecules within these areas are in an ordered state. In an adsorbed polymer chain, the methyl and methylene groups in the loops and tails of the adsorbed layer are expected to be randomly oriented in solution; therefore, they are centrosymmetric and SFG-inactive. However, as a result of the localization of train segments in the plane of the surface, certain chemical groups within these segments are well-ordered; hence, these groups can be selectively probed by SFG. IR-Visible SFG Spectroscopy. In a SFG setup, a pulsed visible laser beam of frequency ω1 and a pulsed IR laser beam of frequency ω2 are spatially and temporally overlapped on an area of the interface being probed. If the molecules at the point of overlap are non-centrosymmetric, this overlapping will result in the emission of light at a frequency ω3 ) ω1 + ω2. The intensity of the emitted light is proportional to the square of the second-order nonlinear susceptibility of the interface. Additionally, if ω2 is in resonance with a vibrational mode of the molecule, there will be an enhancement in intensity of the emitted light. The theory of SFG has been published elsewhere22-25 and will not be discussed here. The intensity of the SFG beam is related to the intensities of the probing visible and IR beams as follows:26

I(ω3 ) ω1 + ω2) ∝ |χeff|2I(ω1) I(ω2)

(1)

Here, χeff is an effective second-order nonlinear susceptibility of the surface which is a sum of terms, each of which contains a single component χijk (i, j, k ) x, y, z) of the second-order nonlinear susceptibility tensor χ, of the surface. For an interface with azimuthal isotropy, only 7 out of the 27 components of χijk are nonvanishing and only 4 are independent,23 namely,

χzzz, χxxz ) χyyz, χxzx ) χyzy , and χzxx ) χzyy The different components of the susceptibility tensor can be deduced by measuring the SFG spectra using different combinations of input and output beam polarizations. Each combination is represented by a sequence of three letters (each being S or P). The letters stand for the following: polarization of the SFG, visible, and IR beams. Therefore, an SSP spectrum has an s-polarized SFG output beam, an s-polarized visible excitation beam, and p-polarized IR excitation beam. For example, for the SSP combination χeff can be obtained as follows:

χSSP eff ) Ts(ω3) Lyy(ω3) Ts(ω1) Lyy(ω1) Tp(ω2) Lzz(ω2) × sin(φ2) χyyz(ω2) (2) x, y, and z are the laboratory coordinates chosen such that z is along the interface normal and x is in the incident plane, and 1-3 correspond to the visible, IR, and SFG beams. Ts(ωi) and Tp(ωi) (i ) 1-3) are the combinations of the Fresnel transmission coefficients to account for the reflection losses of the s- and p-polarized beams when (22) Shen, Y. R. The Principles of Nonlinear Optics; John Wiley & Sons, Inc.: New York, 1984. (23) Hirose, C.; Akamatsu, A.; Domen, K. Appl. Spectrosc. 1992, 46, 1051. (24) Hirose, C.; Akamatsu, N.; Domen, K. J. Chem. Phys. 1992, 96, 997. (25) Hirose, C.; Yamamoto, H.; Akamatsu, A.; Domen, K. J. Phys. Chem. 1993, 97, 10064. (26) Zhuang, X.; Miranda, P. B.; Kim, D.; Shen, Y. R. Phys. Rev. B 1999, 59, 12632.

passing through a series of interfaces.27 The terms Lii (i ) X, Y, Z) are the Fresnel coefficients,23 and φi (i ) 1-3) are the angles between the surface normal and the visible, IR, and SFG beams, respectively. Additionally, χijk(ω2) can be written as iΦ χijk(ω2) ) χnr + ijke

∑q ω

2

χijk,q - ωq + iΓq

(3)

where χijk,q, ωq, and Γq are the line strength, frequency, and line width of the resonance q and Φ is the relative phase of the nonresonant term χnr ijk with respect to the resonant terms. The line strengths’ χijk,q have their origins in the molecular hyperpolarizability tensor (β) components, which are as follows:

∫ ∑ Uijk:lmn(Ω)βlmn,q]f(Ω) dΩ l,m,n)a,b,c

χijk,q ) N [

(4)

where N is the total number of molecules; a, b, and c are the axes of the Cartesian frame-of-reference that is fixed to the molecule; and c is conventionally taken to coincide with the axis of highest symmetry. βlmn,q ) Alm,qMn,q where Alm,q is the lm component of the Raman tensor and Mn is the nth component of the dipole-moment tensor (therefore, only those modes that are both Raman and IR-active contribute to the hyperpolarizability tensor). Ω in eq 4 is the Euler angle that describes the orientation of the molecular axis (abc) with respect to the lab axes (xyz), Uijk:lmn are the 27 × 27 projection coefficients to project the βlmn on the lab axes, and f(Ω) is the probability distribution function of the molecular orientation as a function of Ω. In this work, we have treated the methyl groups as having C3ν symmetry and used the transformation coefficients presented by Hirose et al.23,28

[

]

〈cos θ〉 - 〈cos3 θ〉 + βaac〈cos θ〉 2 (5)

χyyz,s ) N (βccc - βaac)

χyyz,as ) N[-βcaa(〈cos θ〉 - 〈cos3 θ〉)]

(6)

Here, N is the number of methyl groups contributing to the SFG signal, θ is the tilt angle of the c axis from the surface normal (z axis), and βaac ≈ -βcaa and βaac ) rsβccc.24,29 The ensemble averages appearing in the above equations were obtained using the method outlined by Simpson and Rowlen (Supporting Information in ref 30).

〈cos θ〉 ) K

∫0π cos θ f ′(θ) sin θ dθ

(7)

∫0π cos3 θ f ′(θ) sin θ dθ

(8)

〈cos3 θ〉 ) K K)[

∫0πf ′(θ) sin θ dθ]-1

(9)

Here, K is a normalization constant, and f ′(θ) is the net angular distribution function defined as follows: +∞

f′(θ) )

∑ f(2πn + θ) + f(2πn - θ) n)-∞

(10)

1 -(θ-θ0)2 (2σ2) e / x2πσ

(11)

f(θ) )

Structure of PMMA Chains Adsorbed on Sapphire

Langmuir, Vol. 20, No. 17, 2004 7185

Figure 1. Schematic of the flow cell used in our SFG measurements.

f(θ) is a Gaussian function with a mean of θ0 and a standard deviation of σ. For our calculations, the sum in eq 10 only needed evaluation in a limited range (at maximum, from n ) -3 to n ) 3). Experimental Section Materials. Atactic, high-molecular-weight PMMA (Mw, 470 kg/mol; polydispersity index, PDI, 1.06); atactic, low-molecularweight PMMA (Mw, 17 kg/mol; PDI, 1.06); and partially deuterated, atactic, low-molecular-weight PMMA (Mw, 16 kg/mol; PDI, 1.10); and PMMA-CD2-CD3 with deuterated R-methyl and backbone-methylene groups) were bought from Scientific Polymer Products and used as received. The sapphire prisms used as substrates were cleaned initially by sonication in hot toluene followed by treatment in oxygen plasma to remove organic contaminants. These prisms were then rinsed in deionized water and dried in a stream of nitrogen prior to use. All of our measurements were performed in situ in a temperaturecontrolled stainless steel flow cell (Figure 1). The sapphire prism was clamped onto the cell and a Teflon spacer was used to ensure adequate sealing between the prism and the cell. The stainless steel flow cell was thoroughly cleaned (initially in concentrated nitric acid followed by thorough rinsing in deionized water and subsequent exposure to oxygen plasma to remove residual organic contaminants) prior to the measurement. Before introducing the PMMA solution, the cell was filled with pure carbon tetrachloride (HPLC grade, Aldrich) and maintained at 40 °C by a temperature controller (Lakeshore). A dilute PMMA solution was introduced into the cell for less than 1 min. The PMMA solution was then replaced by pure solvent by flushing the cell with a copious amount of warm carbon tetrachloride. Following the above procedure results in the formation of a “starved surface” where the surface coverage is incomplete.4 Care was taken to ensure that the temperature was maintained at 40 °C during the entire process of polymer adsorption. We also spin-coated PMMA (Mw ) 470 kg/mol) films from 4 wt % solutions of PMMA in methylene chloride onto sapphire prisms. These films were then annealed in a vacuum oven at 110 °C and were then used to obtain the SFG spectra of PMMA chains at the PMMA-sapphire interface. SFG Measurements. The measurements were performed using the Spectra-Physics laser system at the University of Akron with a tunable IR output (2700-3200 cm-1, 1-ps pulse width, 1-kHz repetition rate, and full width at half-maximum ∼ 18-20 cm-1) overlapped with an 800-nm visible output (1-ps pulse width and 1-kHz repetition rate). A total internal reflection geometry (as shown in Figure 1) was used to probe the polymer-solid (27) Born, M.; Wolf, E. Principles of Optics, 7th ed.; Cambridge University Press: Cambridge, 1999. (28) Watanabe, N.; Yamamoto, H.; Wada, A.; Domen, K.; Hirose, C.; Tadashi, O.; Norihisa, M. Spectrochim. Acta, Part A 1994, 50, 1529. (29) Gautam, K. S.; Dhinojwala, A. Macromolecules 2001, 34, 1137. (30) Simpson, G. J.; Rowlen, K. L. J. Am. Chem. Soc. 1999, 121, 2635.

Figure 2. SFG spectra of PMMA adsorbed from CCl4 on sapphire. The solid lines are the fits to the data using eq 3. Parts A and B are SSP and PPP spectra, respectively, of Mw ) 470 kg/mol PMMA chains adsorbed on the sapphire substrate at 40 °C (squares) and 25 °C (circles). Parts C and D are SSP and PPP spectra, respectively, for Mw ) 17 kg/mol PMMA adsorbed chains on sapphire substrates at 40 °C (squares) and 25 °C (circles), respectively. interface. The output SFG signal was passed through interference filters and measured using a photomultiplier tube (Hamamatsu R928). Data were collected at 5-cm-1 intervals using a gated photon counter (Stanford Research SRS 400). The measured photon counts were normalized by a reference signal taken by reflecting a fraction of the IR laser beam onto a photodiode. The measured spectra were normalized by the wavelength-independent SFG generated from zinc selenide.

Results and Discussion SFG Spectra of Adsorbed PMMA and PMMA Film-Sapphire Interface. The SFG spectra of adsorbed PMMA molecules at the CCl4/sapphire interface are shown in Figure 2. The 470 and 17 kg/mol PMMA were adsorbed from a 5 × 10-3 mg/mL and a 0.5 mg/mL solution, respectively. As a comparison, the SFG spectra at two different temperatures are shown in Figure 2, 40 °C (above theta temperature) and 25 °C (below theta temperature). The solid lines correspond to fits using eq 3. The main peak in the spectra is near 2951 cm-1. Additionally, our results suggest the existence of two other peaks at 3010 and 3156 cm-1 for the adsorbed layer at 40 °C while for the adsorbed layer at 25 °C these peaks are located at 3004 and 3163 cm-1. The SFG spectra of PMMA and PMMA-CD2-CD3 films at the sapphire interface are shown in Figure 3. The spectra were fit using eq 3, and the two main peaks are located at 2956 and 2995 cm-1. Peak Assignments. The peak assignments for the SFG spectra of PMMA are complicated by the overlap of the symmetric and asymmetric modes of the ester methyl (O-CH3), alpha methyl (C-CH3), and backbone methylene (CH2) groups. We have used the IR and Raman peak assignments reported in the literature (Table 1)31-33 to interpret the SFG spectra of PMMA. According to these assignments, there is significant overlap in the assignments for CH2, C-CH3, and O-CH3 groups. In our measurements, the SSP spectra of PMMA in the adsorbed (31) Schneider, B.; Sˇ tokr, J.; Schmidt, P.; Mihailov, M.; Dirlikov, S.; Peeva, N. Polymer 1979, 20, 705. (32) Dirlikov, S.; Koenig, J. L. Appl. Spectrosc. 1979, 33, 555. (33) George, W. O.; Houston, T. E.; Harris, W. C. Spectrochim. Acta, Part A 1974, 30, 1035.

7186

Langmuir, Vol. 20, No. 17, 2004

Rao et al.

Figure 3. (A) SSP (circles) and PPP (squares) spectra of PMMA-sapphire and (B) SSP (circles) and PPP (squares) spectra of PMMA-CD2-CD3-sapphire interfaces. The solid lines are the fits to the data using eq 3. Table 1. Peak Assignments from 2800 to 3200 cm-1 Based on IR and Raman Spectroscopy IR

Raman

2930-2932 cm-1 CH2(s) and C-CH3(s)

2928-2939 cm-1 CH2(s) and C-CH3(s)

2951-2958 cm-1 CH2(as), C-CH3(as), and O-CH3(s)

2951-2960 cm-1 CH2(as), C-CH3(as), and O-CH3 (s)

2993-3000 cm-1 C-CH3(as) and O-CH3(as)

2996-3004 cm-1 C-CH3(as) and O-CH3(as)

3018 cm-1 O-CH3(as)

3026 cm-1 O-CH3(as)

layer as well as in the film are dominated by a peak near 2950 cm-1 which could be attributed to contributions from the symmetric vibrations of the O-CH3 groups and the asymmetric vibrations of the C-CH3 and CH2 groups. However, we do not observe the CH2 or C-CH3 symmetric stretching vibrations located at 2930 cm-1 in our measurements. According to eqs 5 and 6, this implies that the asymmetric vibrations of the C-CH3 groups would also not have any contributions to the SFG spectra.40 This can be also shown for CH2 groups. On the basis of the above argument, we believe that the peaks near 2950 and 3000 cm-1 arise because of the ordering of the O-CH3 groups of PMMA on sapphire while the CH2 and C-CH3 groups of PMMA are disordered. This is also confirmed by measuring the SFG spectra for PMMA-CD2-CD3 (only has O-CH3 contributions) as shown in Figure 3B. The spectral features in the SSP polarization for PMMA and PMMA-CD2-CD3 are identical and support the hypothesis that the peaks near 2950 and 3000 cm-1 in Figure 2 are associated with O-CH3 groups. The difference in SSP intensity (Figure 3) is due to the difference in film thickness of PMMA and PMMA-CD2-CD3 films. For the PMMA adsorbed layer, the results of our fitting also indicate the existence of two peaks ∼ 3160 cm-1 (more pronounced for high molecular weight polymer adsorbed from very dilute solution). While the position of these peaks is beyond the spectral range accessed in this measurement, not including these peaks in our analysis leads to poor fits to the experimental data. Similar features have been recently observed in the SFG spectra of aqueous solutions on fused silica.34-36 We attribute these peaks to the

Figure 4. Theoretical predictions of Ras/s as a function of orientation angle θ for various values of σ ) 0, 10, 15, and 20 are shown. Experimental ratios of Ras/s (rs ) 2.37) for the O-CH3 group for adsorbed PMMA chains at the CCl4-sapphire interface are plotted as short dashes. The ratios for 40 and 25 °C are indistinguishable from each other. The long dashes and solid lines are Ras/s values for PMMA and PMMA-CD2-CD3 films, respectively, at PMMA-sapphire interfaces. The inset shows the effect of rs on the estimation of θ for the case σ ) 0. The error bars for the values of Ras/s were determined from the standard deviation obtained from fits. The error bars are shown scattered along the x axis for clarity; their position along the x axis has no meaning.

presence of hydroxyl groups (or adsorbed water molecules) on sapphire.37 In the case of the PMMA film, we find no evidence for the presence of hydroxyl peaks at the PMMA film-sapphire interface. Because the PMMA films were annealed for extended periods of time under a vacuum, this could possibly account for the disappearance of the surface hydroxyl (or adsorbed water) groups at the PMMA film-sapphire interface. Orientation of Methoxy Groups of PMMA Chains on the Sapphire Surface. In addition to identifying chemical groups at the surface, SFG spectroscopy also provides information regarding the orientation of these chemical groups. For each of the vibrations identified in the previous section, we have also obtained the values of χijk,q (line strength). As per eqs 5 and 6, these line strengths are related to the orientation of the molecules with respect to the surface. From the ratios of the line strengths (Ras/s ) |χYYZ,as|/|χYYZ,s|) of the asymmetric and symmetric vibrations of the O-CH3 groups in the SSP spectra of PMMA, we are able to quantify the orientation (θ) of the O-CH3 group with respect to the surface normal. We have observed no significant differences in the Ras/s values between the high- and the low-molecular-weight PMMA adsorbed layer, and in the rest of the paper no further distinction will be made between them. In Figure 4, we have plotted the measured value of Ras/s for the O-CH3 groups of the PMMA adsorbed layer in contact with sapphire. Because θ is assumed to be distributed around a mean θ0 with a standard deviation corresponding to σ, we have calculated Ras/s for different values of σ. Additionally, Ras/s also depends on the ratio rs ) βaac/βccc. For the methyl group, this ratio has been reported to vary from 1.6 to 4.3.18,29 While we have used a value of r ) 2.37 for our measurements, as observed in the inset in Figure (34) Wang, J.; Clarke, M. L.; Zhang, Y.; Chen, X.; Chen, Z. Langmuir 2003, 19, 7862. (35) Du, Q.; Freysz, E.; Shen, Y. R. Science 1994, 264, 826. (36) Ye, S.; Nihonyanagi, S.; Uosaki, K. Phys. Chem. Chem. Phys. 2001, 3, 3463. (37) Baumgarten, E.; Wagner, R.; Lentes-Wagner, C. Fresenius’ Z. Anal. Chem. 1989, 334, 246.

Structure of PMMA Chains Adsorbed on Sapphire

4 for small values of θ, the orientation is not a strong function of rs. These results indicate that the O-CH3 groups in the train segments are tilted with θ < 20°. Additionally, the orientation of these groups is not affected by the change in solvent quality. We have performed similar calculations for the orientation of the O-CH3 groups at the PMMA (or PMMA-CD2-CD3) film-sapphire interface (Figure 4). On the basis of the ratio Ras/s and the size of the error bars, we find that the orientation of the O-CH3 groups at the PMMA (or PMMA-CD2CD3) film-sapphire interface is similar to that of the adsorbed layer. Comparison with Literature. Effect of Solvent Quality. The initial stages of PMMA adsorption on oxidized silicon have been studied using techniques such as IR dichroism3,4 and electron spin resonance.6 These studies indicate that at low coverages adsorbed chains adopt a flattened conformation with most of the chain segments being distributed in trains and short loops because such a structure would maximize the enthalpic interactions between the polymer segments and the surface. Because our measurements are performed under conditions of low surface coverage and as a result of the strong interactions between PMMA and sapphire, we expect that the PMMA chains adopt a flattened conformation on sapphire. Under such conditions, a change in the solvent quality would not cause a significant change to the structure of the adsorbed layer. This expectation is confirmed from the results of our measurements when we change the solution temperature from 40 to 25 °C. Because CCl4 is a theta solvent for PMMA at 27 °C, changing the solution temperature would result in the collapse of any loops and tails present in the adsorbed layer,38 leading to an increase in the measured SFG intensity. However, as seen in Figure 2, no such increase is observed in our measurements. The small decrease in SFG intensity with temperature is within the error bars of the SFG measurements. Additionally, we also find that there are no orientational changes in the polymer segments in the adsorbed layer. This result can be attributed to the strong interactions between the chain segments and the surface, which acts as a barrier to any segmental rearrangements at the surface. Local Structure in the Adsorbed Layer. Prior IR3-4,39 studies of adsorbed PMMA chains on silicon substrates have reported dichroism measurements of the methyl and methylene groups of PMMA. At low surface coverages, the CH3 asymmetric peak centered at 3000 cm-1 (asymmetric stretches of the C-CH3 and O-CH3 groups) exhibits a strong directional preference to orient the asymmetric transition dipole moment in the plane of the surface. Because the c axis of O-CH3 is orthogonal to the asymmetric dipole moment, the IR measurements are consistent with the SFG results, where the O-CH3 groups are oriented normal to the surface. In comparison, the IR peak at 2950 cm-1 shows no anisotropic dichroism. This peak is assigned to O-CH3 symmetric stretch and the in-plane C-CH3 and CH2 asymmetric stretches. On the basis of only the ordered O-CH3 groups, the symmetric dipole transition of the O-CH3 would be oriented away (38) Karim, A.; Satija, S. K.; Douglas, J. F.; Ankner, J. F.; Fetters, L. J. Phys. Rev. Lett. 1994, 73, 3407. (39) Schneider, H. M.; Frantz, P.; Granick, S. Langmuir 1996, 12, 994. (40) According to eq 5, χyyz,s ) 0 if 〈cos θ〉 and 〈cos3 θ〉 ) 0 or when 〈cos θ〉/〈cos3 θ〉 ) (1 - rs)/(1 + rs). However, because 〈cos θ〉/〈cos3 θ〉 g 1, the latter condition is not feasible for any value of rs > 0. Because the values of rs for the methylene and methyl groups are positive,18,24,29 this implies that, when χyyz,s ) 0, 〈cos θ〉 and 〈cos3 θ〉 ) 0. According to eq 6, this in turn would result in χyyz,as being equal to 0.

Langmuir, Vol. 20, No. 17, 2004 7187

Figure 5. (A) Different torsional degrees of freedom available for the PMMA side chain. The two torsional angles ψ1 and ψ2 are shown in the figure. (B) Two representative trans-like segments of the adsorbed chains. The red, orange, green, and gray spheres correspond to oxygen atoms, hydrogen atoms of methylene backbone, hydrogen atoms of the O-CH3, and hydrogen atoms of C-CH3, respectively. In both the cases, the O-CH3 groups are oriented almost vertically with respect to the surface normal.

from the surface and would result in an IR dichroic ratio of more than unity. However, the overlap of the O-CH3 symmetric transition with the C-CH3 and CH2 asymmetric transitions and the fact that last two groups are disordered results in the measured dichroic ratio of the peak at 2950 cm-1 being equal to unity. NMR measurements of PMMA adsorbed on alumina report that isotactic and heterotactic sequences in adsorbed PMMA adopt an all-trans conformation.7 Such an all-trans conformation of the chain segments suggests a possible ordering of the CH2 and C-CH3 groups in the adsorbed layer. However, NMR measurements only probe the configuration of the adsorbed chain and are not sensitive to the orientation of the CH2 and C-CH3 groups. On the other hand, SFG measurements are sensitive to the average orientation of chemical groups in the region being probed by the SFG beam. On the basis of the results of prior NMR measurements and the results of our work, we believe that the CH2 and C-CH3 groups of PMMA are ordered on a local scale corresponding to individual adsorbed isotactic and heterotactic sequences. However, when the orientations of the CH2 and C-CH3 groups are averaged over all of the adsorbed sequences, one obtains a net isotropic orientation of these groups at the surface. The reason for such behavior can be found in the factors governing the energetics of the adsorption process. The driving force for PMMA adsorption on sapphire is the interactions between the carbonyl oxygen atoms and the surface hydroxyl groups. The strength of these interactions depends on a number of factors, such as the chain conformation, the orientation of the C-C′ bond which connects the ester side group to the main chain, and the two torsional rotational angles ψ1 and ψ2 (Figure 5A), which sterically hinder the ability of the oxygen atoms to interact with the surface. As a result of the torsional degrees of freedom available to the ester side groups (ψ1 and ψ2), the C-C′ bond can adopt any possible orientation with respect to the surface and still ensure that the carbonyl oxygen of the side groups are able to favorably interact with the substrate. Moreover, because the sequences in the train segments have an all-trans configuration, all of the C-C′ bonds within a given sequence

7188

Langmuir, Vol. 20, No. 17, 2004

would have identical orientations with respect to the surface. One can then envision a series of adsorbed sequences such that each of the sequences has a different orientation of the C-C′ bond with respect to the surface. Two such sequences are illustrated in Figure 5B. On averaging the orientation of these bonds over all the adsorbed sequences, we would obtain a net isotropic orientation of the C-CH3 while the O-CH3 can remain ordered. The ability of O-CH3 to maintain this orientation irrespective of the orientation of the groups in the main chain is because the O-CH3 group of PMMA is decoupled from the main chain because of the torsional degrees of freedom. The net outward orientation of the ester methyl group can be attributed to the torsional rotation (ψ2) of this group about the C′-O bond, which ensures that the methoxy oxygen is brought closer to the surface while the ester methyl group points away from the surface. Because the binding energy of the chain depends on the orientation of the carbonyl bond with respect to the surface,2 this implies that there are many different energetically bound states of the adsorbed PMMA chain. Such a physical picture of the adsorbed chain is consistent with the hypothesis that the energy landscape of the adsorbed chain has a number of local minima, corresponding to different conformations of the adsorbed chain. Detailed molecular

Rao et al.

dynamics simulation of adsorbed PMMA chains on sapphire surface is in progress. Conclusion We have studied the structure of a PMMA chain adsorbed at the CCl4-sapphire interface as a function of solvent quality. Our measurements on the adsorbed layer indicate that the chains adsorb in a flat conformation, which is consistent with the results reported in the literature. We find that the ester methyl groups of adsorbed PMMA segments at CCl4-sapphire interface are ordered, and on average the backbone methylene and alpha methyl groups are disordered. We find that the orientation of the train segments is not affected by a change in the solvent quality. We also find that the structure of the PMMA chain at a sparsely covered surface is similar to that of a chain at the PMMA film-sapphire interface. Acknowledgment. We gratefully acknowledge funding from NSF-DMR 9984996 and a graduate fellowship from OMNOVA, Inc. (A.R.). We would like to acknowledge Dr. Alexander D. Schwab for the development of the fitting program used to analyze the SFG spectra. LA049413U