10474
Langmuir 2004, 20, 10474-10480
Time Evolution Studies of the H2O/Quartz Interface Using Sum Frequency Generation, Atomic Force Microscopy, and Molecular Dynamics Irene Li, Jayasundera Bandara, and Mary Jane Shultz* Department of Chemistry, Tufts University, 62 Talbot Avenue, Medford, Massachusetts 02155 Received June 2, 2004. In Final Form: September 2, 2004 Many interfacial studies on solid surfaces, for example, quartz/water, assume that a standard cleaning procedure regenerates the surface reproducibly. In the reported work, the results of two surface specific techniques, sum frequency generation (SFG) spectroscopy and atomic force microscopy, show that the effects of prolonged exposure to Nanopure water and to pH 10 NaOH are distinctly different. In conjunction with the experimental data, molecular mechanics is used to correlate the SFG spectral frequencies to the hydrogen stretching vibrations of the surface-bound water molecules. It is found that after 17 days of soaking in water, water molecules penetrate into the SiO2 matrix to produce a swollen and amorphous layer; it is likely that broken Si-O bonds from the polishing process serve as nucleation sites for hydration and swelling. Disorder introduced in the interfacial water layer is detected by the rising intensity of the weakly hydrogen-bonded SFG peak at 3450 cm-1. Dominance of the 3450 cm-1 is absent in a pH 10, NaOH-soaked quartz disk, indicating that the strong hydrogen-bonded network in water remains intact.
* To whom correspondence should be addressed. E-mail:
[email protected].
for micro- or nanometer-scale applications where surface features become extremely important, for example, as process technology reaches nanometer dimensions. Quartz is one of the most abundant natural minerals on earth and much is known about its bulk structure as a result of modeling and experimental work.17 In the bulk, quartz is represented as a network of SiO4 corner-sharing tetrahedra with a computed mean Si-O bond length of 1.61 Å and O-Si-O angle equal to 109.3° for an amorphous bulk.18 On the surface, quartz may simply be considered as a SiO2 network with neutral or charged silanol groups (SiOH or SiO-). Experimentally probing the surface is more difficult than probing the bulk solid, so modeling and simulations have been performed to try to elucidate the structure.7,13-17,19-22 Using isotopic labeling, Lu et al. found no oxygen exchange between the bulk SiO2 and 18O tagged oxygen, concluding that the Si-O-Si bonds in the network are stable enough not to react with interstitial oxygen molecules.23 The role of water molecules, on the other hand, is more complicated as highlighted by Bakos et al.20 in a first principles calculation for amorphous SiO2. Although molecular water is the most stable form of water during diffusion into bulk amorphous SiO2, ionization of water
(1) Iiyama, T.; Ruike, M.; Kaneko, K. Chem. Phys. Lett. 2000, 331, 359-364. (2) Pant, D.; Riter, R. E.; Levinger, N. E. J. Chem. Phys. 1998, 109, 9995-10003. (3) Bhattacharyya, K.; Bagchi, B. J. Phys. Chem. A 2000, 104, 1060310613. (4) Loughnane, B. J.; Scodinu, A.; Fourkas, J. T. J. Phys. Chem. B 1999, 103, 6061-6068. (5) Willard, D. M.; Levinger, N. E. J. Phys. Chem. B 2000, 104, 1107511080. (6) Baldelli, S.; Schnitzer, C.; Shultz, M. J.; Campbell, D. J. J. Phys. Chem. B 1997, 101, 10435-10441. (7) Nihonyanagi, S.; Ye, S.; Uosaki, K. Electrochim. Acta 2001, 46, 3057-3061. (8) Benderskii, A. V.; Eisenthal, K. B. J. Phys. Chem. B 2000, 104, 11723-11728. (9) Zimdars, D.; Eisenthal, K. B. J. Phys. Chem. A 1999, 103, 1056710570. (10) Baldelli, S.; Schnitzer, C.; Shultz, M. J.; Campbell, D. J. Chem. Phys. Lett. 1999, 302, 157-163. (11) Schnitzer, C.; Baldelli, S.; Shultz, M. J. J. Phys. Chem. B 2000, 104, 585-590. (12) Simonelli, D.; Shultz, M. J. J. Chem. Phys 2000, 112, 68046816.
(13) Du, Q.; Freysz, E.; Shen, Y. R. Phys. Rev. Lett. 1994, 72, 238241. (14) Duval, Y.; Mielczarski, J. A.; Pokrovsky, O. S.; Mielczarski, E.; Ehrhardt, J. J. J. Phys. Chem. B 2002, 106, 2937-2945. (15) Kim, J.; Kim, G.; Cremer, P. S. J. Am. Chem. Soc. 2002, 124, 8751-8756. (16) Yeganeh, M. S.; Dougal, S. M.; Pink, H. S. Phys. Rev. Lett. 1999, 83, 1179-1182. (17) Iler, R. K. The Chemistry of Silica : Solubility, Polymerization, Colloid and Surface Properties and Biochemistry of Silica; John Wiley & Sons: New York, 1979. (18) Du, M.-H.; Kolchin, H.-P.; Cheng, H.-P. J. Chem. Phys. 2004, 120, 1044-1054. (19) Koudriachova, M. V.; Beckers, J. V. L.; de Leeuw, S. W. Comput. Mater. Sci. 2001, 20, 381-386. (20) Bakos, T.; Rashkeev, S. N.; Pantelides, S. T. Phys. Rev. Lett. 2002, 88, 055508, 1-4. (21) Xiao, Y.; Lasaga, A. C. Geochim. Cosmochim. Acta 1996, 60, 2273-2486. (22) Goss, K.-U.; Schwarzenbach, R. P. J. Colloid Interface Sci. 2002, 252, 31-41. (23) Lu, H. C.; Gustafsson, T.; Gusev, E. P.; Garfunkel, E. Appl. Phys. Lett. 1995, 67, 1742-1744.
1. Introduction Water plays a significant role in many natural, biological, chemical, and industrial systems. Various processes involving water take place at the interface rather than in the bulk. These include cloud formation and ozone depletion in the atmosphere, cross-membrane transport in cell biology, heterogeneous catalytic reactions, waterrock interactions in geochemistry, and contamination and oxidation on silicon wafer surfaces, just to name a few.1-5 Many studies focus on the molecular arrangement at solid surfaces,6,7 the air/water interface,8-10 and the water interface for ionic solutions.6,11,12 Of the studies of water on quartz, few noted the storage conditions prior to the initial surface preparation procedure.7,13-16 Reproducible regeneration of the quartz surface by cleaning is assumed. However, sum frequency generation (SFG) spectra and atomic force microscopy (AFM) images presented here show that the surface can be altered during storage, and its history must be considered. Lack of knowledge of the surface history and resultant modifications prior to analysis may lead to erroneous conclusions, particularly
10.1021/la048639u CCC: $27.50 © 2004 American Chemical Society Published on Web 10/20/2004
Time Evolution of the H2O/Quartz Interface
into H+ and OH- may also occur at “low” temperatures (e.g., 400 °C). The ions, reacting with the surface, can form adjacent SiOH groups and possibly exchange oxygen atoms with the network.20 Xiao et al. studied quartz dissolution under a high pH (12) using disilicic acid (HO)3Si-O-Si-(OH)3 as the molecular model.21 In alkaline conditions, dissolution occurs by an OH- catalytic mechanism via fivefold coordinated Si species. In the mid-pH range under ambient conditions, the most likely interaction is hydrogen bonding of H2O to the surface silanol groups. In the present report, timed exposure of quartz to two aqueous environments is investigated: pure water and a NaOH solution of pH 10. Water was selected to simulate laboratory conditions of high humidity because it has been shown that moisture in the air can change the behavior of a substrate’s surface.22,24 The alkaline solution was selected for its technological importance in the semiconductor industry where much of the planarization process25-27 of the Si(100) substrate is carried out at pH 10. This paper is organized as follows. Section 2 indicates the materials and methods. Section 3 begins with a molecular mechanics simulation that identifies the SFG spectral frequencies with molecular vibrations of the surface-bound water molecules. The simulation highlights the importance of including certain surface species in the spectral interpretation. The calculation is followed by presentation of the SFG and AFM experimental results. Section 4 presents a discussion of the three components: the simulation, the SFG, and the AFM experimental results. Section 5 concludes with an overview interweaving the three components. 2. Materials and Methods 2.1. Molecular Mechanics Calculations. The molecular mechanics calculations used the SYBYL force field and the MMFF94 force field module of PC Spartan Pro for the calculation of the equilibrium geometry and the normal-mode vibrational frequencies. 2.2. Materials. Laser polished polycrystalline infrared quartz disks were purchased from Heraeus Optics (2 in. diameter, 0.315 in. thick). Distilled and deionized water (Millipore, 18 MΩ) was used for preparation of all solutions and static soaks. The pH of the alkaline aqueous NaOH solution was verified with a pH meter both prior to and following the SFG experiments. Quartz disks were cleaned [3-day bath soak in concentrated sulfuric acid, NoChromix (Godax)], then rinsed with copious quantities of 18 MΩ water, and left in water overnight. The glass cell was similarly cleaned except the acid soak was only for 1 day. 2.3. SFG. Periodic SFG experiments were performed on two IR quartz disks: water-soaked (A) and NaOH pH 10 solutionsoaked (B) over 24 days. To uniformly compare the surfaces, all SFG spectra are collected in contact with a pH 10 NaOH solution freshly prepared before each run. The water and basic solution were changed daily during the soaking time. Data on the effect of evacuation was collected at the end of the timed soaks. The cell was first drained and heated with a heat gun to drive off excess moisture and then connected to a vacuum line, and the pressure was reduced to ∼30 mTorr for 30 min. All SFG spectra were collected using ssp polarization (spolarized SF, s-polarized visible, p-polarized IR). The tunable IR beam is generated by pumping a KTP-based OPO/OPA (LaserVision) with horizontally polarized light from the 10-Hz Nd: (24) Kuramochi, H.; Ando, K.; Yokoyama, H. Surf. Sci. 2003, 542, 56-63. (25) Steigerwald, J. M.; Murarka, S. P.; Gutmann, R. Chemical Mechanical Planarization of Microelectronic Materials; J. John Wiley & Sons: New York, 1997; pp 40-43, 48-126. (26) Li, I.; Forsthoefel, K. M.; Richardson, K. A.; Obeng, Y. S.; Easter, W. G.; Maury, A. Mater. Res. Soc. Symp. Proc. 2000, 613, E7.3.1E7.3.10. (27) Pietsch, G. J.; Chabal, Y. J.; Higashi, G. S. Surf. Sci. 1995, 331333, 395-401.
Langmuir, Vol. 20, No. 24, 2004 10475 YAG laser (Spectra-Physics GCR 150). Changing the angle of the nonlinear KTP crystals allows continuous tuning in the desired range of 3000-3800 cm-1 with a bandwidth of 4 cm-1. Doubling the 1064-nm fundamental of the laser produces the 532-nm visible input beam. The two beams are then spatially and temporally overlapped at 52° and 46° to the surface for the 532 nm and IR, respectively. The power densities of the IR and visible beams are ∼130 and ∼ 540 mJ/cm2, respectively. The SFG signal is spatially filtered from the reflected visible beam and directed through a polarizer, a 532-nm filter to minimize scattered light, and a 0.25-m monochromator (JarrellAsh) before being detected by a photomultiplier tube (Hamamatsu R3443). The signal is collected by a gated boxcar averager and analyzed by a computer. Data points are collected every 10 cm-1 at 400 shots per point. The spectra are referenced to the SFG signal of a Si(100) wafer. Normalizing each spectrum to the IR power spectrum eliminates fluctuations caused by the frequencydependent intensity of the IR source. More specific details of the SFG experimental setup are described elsewhere.6,28 2.4. AFM. AFM imaging was performed in air on treated quartz (0.003 in. thick) with a Nanoscope III (Digital Instruments) operating in contact mode with a 100-µm wide-legged cantilever (spring constant, 0.58 N/m). A scan rate of 4.36 Hz was used for the reference, and 15.26 Hz was used for the samples. The samples were air-dried to remove excess solutions immediately prior to imaging. All imaged areas are 100 nm × 100 nm.
3. Results 3.1. Molecular Mechanics Calculations. Cyclic (SiO2)6 structures saturated with water were used to simulate the quartz network on the surface. At low pH, water associates to some of the silanol groups via hydrogen bonding.7,14,16,17,29 The atoms for all the simulated structures are represented as follows: Si atoms by the large spheres, O atoms by the medium spheres, and H atoms by the small spheres. Structure I has a water molecule associated with each silanol group, that is, six water molecules for the six-membered ring.
At high pH, the deprotonated silanol groups are negatively charged and either singly hydrogen-bonded to a water molecule (structure II) or bridged as shown in structure III. 3.2. SFG. Figure 1a shows a SFG spectrum of quartz after the initial 3-day acid and water soak. It consists of a featureless broad band in the region between 3000 and 3350 cm-1 and a smaller band centered at 3450 cm-1. After 17 days of exposure to water, the peak dominance is reversed (Figure 1b). The 3450 cm-1 peak is sharper and gains in intensity, whereas the broad band remains approximately the same. After 20 days (Figure 1c), the two bands merge into one very broad peak. These findings imply that the quartz surface has been significantly modified upon long-term contact with water. After 24 days the cell was evacuated. Figure 2 is the spectrum collected following a further 1-day water soak. (28) Baldelli, S.; Campbell, D. J.; Schnitzer, C.; Shultz, M. J. J. Phys. Chem. B 1997, 101, 4607-4612. (29) Thiel, P. A.; Madey, T. E. Surf. Sci. Rep. 1987, 7, 211-385.
10476
Langmuir, Vol. 20, No. 24, 2004
Li et al.
Quartz disk B, treated with pH 10 NaOH, does not show the same reversibility in peak dominance as quartz A. Parts a-c of Figure 3 are, respectively, the SFG spectra taken after 6, 17, and 20 days. The only observable change is the increase in the SFG signal intensity for both peaks after prolonged exposure. 3.3. AFM. An AFM image of a quartz sample without any treatment, Figure 4a, is shown for comparison. After soaking in water for 20 days, the quartz surface (Figure 4b) is markedly different from that of the reference. Distinct structures with dimensions of approximately 5.4 nm wide by 8.3 nm long are clearly observable. There are also smaller structures that appear to be in the process of “growing”. All the structures are aligned with the long axis in the same direction. When the cell is evacuated (Figure 4c), the structures in Figure 4b collapse. There are still visible structures that appear to have shrunk, but overall, the 100 nm × 100 nm image area is significantly smoother than that in Figure 4b. A quartz disk soaked in pH 10 NaOH for 20 days shows no distinct structures (Figure 4d), and in fact, the alkaline-soaked surface is slightly more glazed and smoothed than that of the reference. 4. Discussion 4.1. Molecular Mechanics Calculations. Figure 5a-c shows the computational results for structures I-III. The vibrational frequencies, calculated with MMFF94, are plotted as Gaussian peaks,
y ) Ae-(x-x0) /(2σ ); A ) 2
2
1 σx2π
(1)
where A is the amplitude, x - x0 is the vibrational frequency difference, and σ is the bandwidth. The calculated frequencies indicate that the quartz surface encompasses the three structures to varying degrees. The choice of structures I- III was guided by studies of the surface OH group density on amorphous silica, models
Figure 1. SFG spectra of the quartz/NaOH interface with a surface history of soaking in water for (a) 1, (b) 17, and (c) >20 days. The data (9) is fitted with Gaussian peaks and is illustrated by the solid line.
for describing ionization at oxide-water interfaces, and the pK’s of the ionization reactions. The surface concentration of OH groups30 or the silanol number, ROH, is expressed by eq 2
ROH ) δOH(s)NA × 10-21S-1
(2)
where δOH(s) is the concentration of OH groups (mmol OH/g SiO2), NA is Avogadro’s number, and S is the specific surface area with respect to krypton adsorption (m2/g). (30) Zhuravlev, L. T. Langmuir 1987, 3, 316-318.
Time Evolution of the H2O/Quartz Interface
Langmuir, Vol. 20, No. 24, 2004 10477
Figure 2. SFG spectra of the quartz/NaOH interface with a surface history of soaking in water for 24 days and then evacuated and soaked for 1 more day. The data (9) is fitted with Gaussian peaks and is illustrated by the solid line.
The units for ROH are OH groups/nm2, and ROH is referred to as the hydroxyl number density. In a fully hydroxylated (silanol-covered) amorphous silica surface, Zhuravlev reports an average value of ROH ) 5.0 OH groups/nm2 regardless of other structural characteristics (i.e., pore size or distribution, SiO2 skeletal structure).30 The most probable species on the fully hydroxylated surface is one OH group per Si atom (SiOH) in this state.30 A principal model to describe ionization at the quartzwater interface is a 2-pK model.14 It assumes two consecutive protonations. The surface reactions and pK values are as follows: +H
SiOH 98 SiOH2+ -H
SiOH 98 SiO-
pK ) 3.0 ( 1.0 pK ) 7.0 ( 1.0
(3) (4)
At pH 10, the surface is sufficiently ionized to strongly orient nearby water molecules. Neutral silanol groups are also present, and these can interact with water to give structure I. Structure III, with a water bridging two neighboring Si-O- groups, is included because there are broken Si-O bonds from the polishing process. Comparing any of the experimental spectra to the component spectra of Figure 5a-c supports the presence of silanol groups even in a relatively basic environment. As seen in Figures 1 and 3, both initial soak environments have a major contribution from structure I. The importance of including structures II and III becomes clear as the quartz-water interaction evolves. Figure 6 compares the 17-day water spectrum to its simulated spectrum. The spectrum results from a convolution of Gaussian peaks from each mode with the frequency corrected by a multiplicative factor of 1.03. The major species is no longer predominately structure I but includes an equal weighting of structures I and II. There is a smaller contribution from structure III to broaden the spectrum on the red side. The agreement between the experimental and simulated spectra is very good, indicating the increasing influence of structures II and III with time. 4.2. SFG. The SFG spectra reflect the species at and near the quartz surface. The surface properties of silica, for example, surface electric charge, dissolution, and adsorption/desorption, have been attributed to the pres-
Figure 3. SFG spectra of the quartz/NaOH interface with a surface history of soaking in pH 10 NaOH for (a) 6, (b) 17, (c) >20 days. The data (9) is fitted with Gaussian peaks and is illustrated by the solid line.
ence of three surface species:14 SiOH, SiOH2+, and SiO-. Because the isoelectric point of quartz is around pH 2, SiOH2+ is present only in very acidic environments, that is, pH < 2.13,14,16,25 In weakly acidic or weakly basic solutions, silanol groups dominate the quartz surface. SiObecomes significant in high pH environments. In a neutral solution such as water, interfacial water molecules loosely associate with the surface (structure I), forming a relatively disordered layer. In a basic environment, pH 10, the negatively charged SiO- groups on the hydrophilic quartz surface produce a
10478
Langmuir, Vol. 20, No. 24, 2004
Li et al.
Figure 4. AFM images of quartz following (a) 1 day of water soak, (b) >20 days of water soak, (c) 1 day of water soak post-drying and evacuation of the cell after 24 days of water soak, and (d) >20 days of pH 10 NaOH soak.
strong electric field that orients several monolayers of water molecules near the interface.7,13-17,20,21,23,29 Under the influence of the negative charge, near-surface water molecules align with their hydrogen atoms pointing toward the Si-O- groups on the surface (structure II). The oriented water molecules produce a signal intensity that is directly related to the electric field at the interface, which in turn is proportional to the hydroxyl number density.16 When the surface is covered with silanol groups as in the case of a pH 7 water environment, the surface charge density is very low, and the SFG signal intensity is expected to be low. The effect of surface charging can be observed through the intensity variation of the strongly hydrogen-bonded 3200 cm-1 feature, which is often attributed to strong intermolecular coupling of the symmetric O-H stretch of tetrahedrally coordi-
nated water molecules.6,7,13,15,16 The 3450 cm-1 peak is attributed to water in a weaker hydrogen-bonding environment and is a reflection of water disorder at the surface.6,7,13,15 Under a short-term water soak (Figure 1a), water molecules seem to only “wet” the surface with little chance of penetrating the network. Upon contact with NaOH at pH 10 for the SFG experiment, the surface becomes negatively charged, and the 3200 cm-1 broad band of the SFG spectrum is the most intense. When enough time is allowed for water molecules to diffuse into the matrix (17 days, Figure 1b) a noticeable effect is observed in the SFG spectrum. There is an increase in the weakly bonded water as seen by the dominance of the 3450 cm-1 peak. In contrast, the 3200 cm-1 peak intensity in Figure 1a,b is unchanged; hence, the amount of strongly bonded water appears unaffected by the water soak.
Time Evolution of the H2O/Quartz Interface
Langmuir, Vol. 20, No. 24, 2004 10479
Figure 6. Comparison of the 17-day water SFG spectrum (9 with line) to its simulation (9). The simulation is a convolution of equal weightings of I and II with a small contribution from III.
Figure 5. Plots of hydrogen stretch vibrations calculated from molecular modeling for (a) -SiOH‚‚‚ HOH, (b) -SiO‚‚‚HOH, and (c) -SiO‚‚‚OH‚‚‚OSi-. The bandwidths are 25 cm-1. The peak amplitudes are set to 1.
It is believed that water penetration occurs by nucleating at the broken Si-O bonds arising from the polishing process. As the water molecules slowly hydrate the broken bonds, the matrix forms a swelled (Si-O)n structure that disturbs the molecularly flat and strong hydrogen-bonded water network on the surface. The surface then adopts a more amorphous structure as some of the Si atoms are freed from the lattice. Once nucleated, water molecules
continue to hydrate the lattice because of local strain. In an amorphous layer, there is a mixture of surface sites ranging from two- to six-membered Si-O rings that water molecules may act upon to produce a variety of water clustering, adsorption, and penetration.18 Generally, smaller rings have more strain and are more easily attacked by water. Two-membered Si-O rings (concentration of 0.2-0.4/nm2) are particularly susceptible to attack by water at room temperature, and water molecules adsorbed to the surface can hydrolytically weaken the local Si-O network.31-33 Water molecules associated with these Si sites have a disrupted hydrogen-bonded network. The high intensity of the 3400 cm-1 peak in Figure 1b is consistent with this description of the disruptive nature of the structures on the hydrogen-bonded water network. As the “softening” continues, the disruption should become more pronounced. This is observed after 20 days of soaking in water (Figure 1c), where the two, once distinct, SFG peaks now merge into one very broad band. When the cell is evacuated (Figure 2), the dominance of the strongly bonded peak is restored. Under ultrahigh vacuum conditions, physically adsorbed water on the surface is less than a monolayer.31 Under less severe evacuation conditions, physisorbed water is partially removed. However, it is highly improbable that the SiO2 network completely reforms; the surface still retains a more amorphous structure than the original surface.14,30 The AFM images taken in tandem with the SFG spectra show that features from the prolonged water soak substantially shrink upon evacuation of the cell (Figure 4c). For quartz soaked in NaOH at pH 10, neither the peak position nor the peak dominance varies over time (Figures 3a-c). The SFG intensity increases over time but the ratio between the 3200 and the 3450 cm-1 peaks remains constant. The rise in intensity over time implies that the number density of oriented water at the surface is increased, which is likely due to water confined in the electrostatic double layer between the cations and the negatively charged surface. Water in the double layer is strongly oriented with its hydrogen atoms pointing toward (31) Mitsui, T.; Rose, M. K.; Fomin, E.; Ogletree, D. F.; Salmeron, M. Science 2002, 297, 1850-1852. (32) Hu, J.; Xiao, X.-D.; Salmeron, M. Science 1995, 268, 267-269. (33) Diez-Perez, I.; Luna, M.; Teheran, F.; Ogletree, D. F.; Sanz, F.; Salmeron, M. Langmuir 2004, 20, 1284-1290.
10480
Langmuir, Vol. 20, No. 24, 2004
the surface and its partially negative oxygen atom facing the sodium cation. In Figure 3b,c the effects of the NaOH solution appear to stabilize because the peak intensities in parts b and c are essentially the same. 4.3. AFM. The AFM images show that the interaction of water and that of the pH 10 NaOH solution with quartz are not the same. In Figure 4b, and to a lesser extent Figure 4c, the most striking feature is the aligned structures. As explained in the SFG discussion, it is believed that the orientation is attributed to tracks formed as a polishing pad rotates across the surface. These tracks are sufficiently shallow so that the quality of the quartz satisfies “laser polished” specifications. Upon evacuation, the structures collapse as some of the water is removed. Although the surface morphology is still rough (Figure 4c), the hydrogen-bonding disruption of the water network is diminished and the 3200 cm-1 peak is the strongest in the SFG spectrum (Figure 2). Under alkaline conditions (Figure 4d), the surface has a smoother, glazed appearance. The hydroxide presence appears to prevent water from penetrating into the matrix, probably by associating with the partially positive Si of the broken Si-O bonds. The sodium ion may also serve as a barrier to water attack due to its electrostatic attraction to the charged surface. In essence, a basic solution serves to fill gaps in the Si-O-Si network, producing a more molecularly flat surface. 4.4. Overview. The three techniques together suggest a molecular model of the quartz surface interaction with the surrounding liquid. In the presence of pH ∼7 water, water molecules slowly generate an amorphous layer, which weakens the strong hydrogen bond network of the near-surface water molecules. This is experimentally observed in the SFG and AFM results by the spectral peak intensity changes and the formation of features. In a basic solution, the experimental results are distinctly different from that of water-exposed quartz. The absence of the spectral shift in peak dominance, as well as any detectable features in the AFM image, points toward a barrier against water penetration. From the calculations, the presence of other species, specifically SiOH, cannot be discounted although charging on the surface from NaOH is likely the primary means of preserving the hydrogenbonded network.
Li et al.
An appreciable change in the SFG spectrum and AFM image is found after 17 days of water soak. On that time scale, water appears to penetrate into the SiO2 matrix, yielding a swelled and amorphous layer. The disruption of the surface hydrogen-bonded network is supported by the rising dominance of the 3450 cm-1 SFG peak. The AFM image provides further insight into the intercalation of the water molecules, primarily via broken Si-O bonds on the surface. When a drying process (evacuation of the cell) is applied, there is a return to a flatter surface. On an interesting side note, the switch in dominance in the water SFG spectrum is also observed at the water-air interface as a function of the anion present.34 Spectra of the various species are simulated using molecular modeling. The association of water to a silanol group, structure I, results in a broad band with a range of 3025-3325 cm-1. The other two structures (water hydrogen bonding to SiO- and water bridging neighboring SiO- groups) generate narrower bands. Fitting the experimental data shows that SiOH groups on the surface cannot be ignored, even in the presence of pH 10 where the surface is known to be negatively charged. The interaction of quartz with a pH 10 NaOH solution is quite different from its interaction with water. The AFM image shows that the surface flattens following prolonged exposure to base. The flattening is attributed to an electrostatic barrier protecting the surface from water attack. By associating to the surface, both the hydroxide anion and the sodium cation serve to increase water ordering. Increased orientation is observed as an increased intensity in the SFG spectra over time. The electrostatic field generates a protective covering on the surface. Experiments to distinguish between purely electrostatic effects and pH are underway. Chemical interactions occurring at the quartz surface are important to the integrated circuits industry. The commercial process of chemical mechanical polishing in fabrication exploits the solubility of the hydrated surface at pH 10. The addition of polishing beads lends to more effective planarization, likely due to defects generated at the surface.18,25 The limitations of this chemical-mechanical method for planarization of the surface have been reached, so the impact of being able to chemically etch the surface with precision is enormous and greatly in demand.
5. Conclusion SFG spectroscopy and AFM techniques are used to experimentally study the effects of prolonged aqueous exposure (water and pH 10 NaOH) on quartz. Short-term water exposure results in a more intense broad peak at 3200 cm-1, reflecting the order of surrounding water molecules in response to the strong electric field generated by the negatively charged surface under a pH 10 solution.
Acknowledgment. The authors thank Dr. Kuang Pang Li for his advice on the molecular modeling aspects of this work. LA048639U (34) Liu, D.; Ma, G.; Levering, L. M.; Allen, H. C. J. Phys. Chem. B 2004, 108, 2252-2260.