Langmuir 2008, 24, 13393-13398
Surface Modification of ZnO Using Triethoxysilane-Based Molecules C. G. Allen,† D. J. Baker,† J. M. Albin,† H. E. Oertli,† D. T. Gillaspie,‡ D. C. Olson,‡ T. E. Furtak,† and R. T. Collins*,† Department of Physics, Colorado School of Mines, Golden, Colorado, and National Renewable Energy Laboratory, Golden, Colorado ReceiVed NoVember 26, 2007. ReVised Manuscript ReceiVed September 17, 2008 Zinc oxide (ZnO) is an important material for hybrid inorganic-organic devices in which the characteristics of the interface can dominate both the structural and electronic properties of the system. These characteristics can be modified through chemical functionalization of the ZnO surface. One of the possible strategies involves covalent bonding of the modifier using silane chemistry. Whereas a significant body of work has been published regarding silane attachments to glass and SiO2, there is less information about the efficacy of this method for controlling the surface of metal oxides. Here we report our investigation of molecular layers attached to polycrystalline ZnO through silane bonding, controlled by an amine catalyst. The catalyst enables us to use triethoxysilane precursors and thereby avoid undesirable multilayer formation. The polycrystalline surface is a practical material, grown by sol-gel processing, that is under active exploration for device applications. Our study included terminations with alkyl and phenyl groups. We used water contact angles, infrared spectroscopy, and X-ray photoemission spectroscopy to evaluate the modified surfaces. Alkyltriethoxysilane functionalization of ZnO produced molecular layers with submonolayer coverage and evidence of disorder. Nevertheless, a very stable hydrophobic surface with contact angles approaching 106° resulted. Phenyltriethoxysilane was found to deposit in a similar manner. The resulting surface, however, exhibited significantly different wetting as a result of the nature of the end group. Molecular layers of this type, with a variety of surface terminations that use the same molecular attachment scheme, should enable interface engineering that optimizes the chemical selectivity of ZnO biosensors or the charge-transfer properties of ZnO-polymer interfaces found in oxide-organic electronics.
Introduction Self-assembled and covalently bonded organic monolayers have been widely researched1-3 and can be used to manipulate surface properties for a variety of applications. For example, prior studies have shown that molecular treatments are beneficial for improving corrosion resistance,4 casting ultrathin photoresist films,5 selectively depositing metal,6 and modifying chemical sensitivity.7,8 Functionalizations of metal oxides, in particular, could be beneficial because of the importance of these materials in several hybrid inorganic-organic electronic applications. Example hybrid systems currently being explored are biological sensors,9,10 light-emitting diodes,11,12 and solar cell devices.13-17 Metal oxides also exhibit variable surface energies, which lead to controllable wetting characteristics.18,19 In nearly all of these applications, the interface between the organic and inorganic phases critically influences the electronic properties of the system. Surface molecular modifications are actively researched to control and manipulate these properties.20-23 Here, the primary interest in molecular modification stemmed from our ongoing work utilizing metal oxides interfaced with conductive, conjugated polymers to form hybrid solar cells. In * Corresponding author. E-mail: [email protected]
† Colorado School of Mines. ‡ National Renewable Energy Laboratory.
(1) Schwartz, D. K. Annu. ReV. Phys. Chem. 2001, 52, 107–137. (2) Sagiv, S. J. Am. Chem. Soc. 1980, 102, 92–98. (3) Ulman, A. Chem. ReV. 1996, 96, 1533–1554. (4) Metikos-Hukovic, M.; Babic, R.; Petrovic, Z.; Posavec, D. J. Electrochem. Soc. 2007, 154, C138–C143. (5) Hiroyuki, S.; Hikaru, S.; Kyung-Hwang, L.; Kuniaki, M. Jpn. J. Appl. Phys 2006, 45, 5456–5460. (6) Lu, P.; Walker, A. V. Langmuir 2007, 23, 12577–12582. (7) Yakimova, R.; Steinhoff, G.; Petoral, J.; Vahlberg, C.; Khranovskyy, V.; Yazdi, G. R.; Uvdal, K.; Lloyd, Biosens. Bioelectron. 2007, 22, 2780–2785. (8) Sadik, P. W.; Pearton, S. J.; Norton, D. P.; Lambers, E.; Ren, F. J. Appl. Phys. 2007, 101, 143517.
these cells, an exciton that is created in the active polymer layer must diffuse to the oxide interface in order to be separated into free carriers and collected.24 Because the diffusion length of the exciton is quite short (∼10 nm or less), it is a critical need to optimize the structural order of the polymer chains near the oxide interface without negatively impacting the charge-transfer process. ZnO, in particular, is a good candidate for hybrid solar cells.13-15,25-28 ZnO has low toxicity, can be synthesized into a wide range of nanostructures, and is synthesized from naturally (9) Flink, S.; van Veggel, F. C. J. M.; Reinhoudt, D. N. AdV. Mater. 2000, 12, 1315–1328. (10) Zhang, Z.; Emanetoglu, N. W.; Saraf, G.; Chen, Y.; Wu, P.; Zhong, J.; Lu, Y.; Chen, J.; Mirochnitchenko, O.; Inouye, M. IEEE Trans. Ultrason., Ferroelectrics, Frequency Control 2006, 53, 786–792. (11) Haque, S. A.; Koops, S.; Tokmoldin, N.; Durrant, J. R.; Huang, J.; Bradley, D. D. C.; Palomares, E. AdV. Mater. 2007, 19, 683–687. (12) Burroughes, J. H.; Bradley, D. D. C.; Brown, A. R.; Marks, R. N.; Mackay, K.; Friend, R. H.; Burns, P. L.; Holmes, A. B. Nature 1990, 347, 539–541. (13) Takanezawa, K.; Hirota, K.; Wei, Q. S.; Tajima, K.; Hashimoto, K. J. Phys. Chem. C 2007, 111, 7218–7223. (14) Olson, D. C.; Piris, J.; Collins, R. T.; Shaheen, S. E.; Ginley, D. S. Thin Solid Films 2006, 496, 26–29. (15) Beek, W. J. E.; Wienk, M. M.; Kemerink, M.; Yang, X.; Janssen, R. A. J. J. Phys. Chem. B 2005, 109, 9505–9516. (16) Coakley, K. M.; McGehee, M. D. Appl. Phys. Lett. 2003, 83, 3380–3382. (17) Huynh, W. U.; Dittmer, J. J.; Alivisatos, P. A. Science 2002, 295, 2425– 2427. (18) Zhang, Z.; Chen, H.; Zhong, J.; Saraf, J.; Lu, Y. J. Electron. Mater. 2007, 36, 895–889. (19) Feng, X.; Feng, L.; Meihua, J.; Jiang, L.; Zhu, D. J. Am. Chem. Soc. 2004, 126, 62–63. (20) Chong, L. W.; Lee, Y. L.; Wen, T. C. Thin Solid Films 2007, 515, 2833– 2841. (21) Kudo, N.; Honda, S.; Shimazaki, Y.; Ohkita, H.; Ito, S.; Benten, H. Appl. Phys. Lett. 2007, 90, 183513. (22) Goh, C.; Scully, S. R.; McGehee, M. D. J. Appl. Phys. 2007, 101, 114503. (23) Chen, Y.; Liu, W.; Ye, C.; Yu, L.; Qi, S. Mater. Res. Bull. 2001, 36, 2605–2612. (24) Gregg, B. A. Mater. Res. Soc. Bull. 2005, 30, 20–22. (25) White, M. S.; Olson, D. C.; Shaheen, S. E.; Kopidakis, N.; Ginley, D. S. Appl. Phys. Let. 2006, 89, 143517. (26) Baxter, J. B.; Aydil, E. S. Appl. Phys. Let. 2005, 86, 053114+
10.1021/la802621n CCC: $40.75 2008 American Chemical Society Published on Web 10/31/2008
13394 Langmuir, Vol. 24, No. 23, 2008
abundant zinc metal. In addition, ZnO films can be prepared by methods amenable to large-scale manufacturing at temperatures compatible with low-cost, flexible substrates. For example, ZnO synthesized from sol-gel precursors is currently being investigated for patterned deposition by ink-jet printing.29 Molecular modifications of spin-coated, sol-gel-grown ZnO films have recently been evaluated in bilayer excitonic hybrid solar cell device structures.30,31 ZnO/poly(3-hexylthiophene) devices had better illuminated short-circuit currents when the ZnO surface was modified with octadecanethiol, as compared to unmodified samples. This was attributed to more favorable polymer morphology near the ZnO interface, which improved the overlap of the polymer’s absorption spectrum with the AM 1.5 solar spectrum. It is likely that the exciton diffusion length and hole mobility in the polymer were also improved. This would be consistent with the reports of improved polymer/C-60 blend solar cells when the blend contained small percentages of alkanedithiol32 and agrees with surface modifications of silicon dioxide (SiO2) that improved polymer thin film transistors.33 Here, we investigate triethoxysilane treatments on thin sol-gelgrown films of ZnO. In addition to desirably influencing the morphology of the polymer (to be published elsewhere), our objective was to optimize and assay triethoxysilane molecular layers attached to ZnO using silane chemistry in the presence of an amine catalyst. Layers were grown from octadecyltriethoxysilane (n-C18H37Si(OCH2CH3)3, OTES), decyltriethoxysilane (n-C10H21Si(OCH2CH3)3, DTES), and phenyltriethoxysilane (C6H5Si(OCH2CH3)3, PTES). Our results suggest that the deposition scheme is useful for a variety of different molecular groups. Molecular Attachment Scheme. Many molecular treatments have been investigated on ZnO, including the following attachment schemes: chlorosilanes,34 methoxysilanes,34 phosphonic acids,34 thiols,8,30,34-36 and carboxylic acids.34,37,38 Some of these strategies that might appear to be very simple, such as those involving acid-base reactions, also lead to etching of the ZnO. This effect can be reduced by alloying ZnO with Mg.34 Here, we investigated an alternative approach using R-triethoxysilane (TES) molecules where a strong covalent Si-O attachment to the surface might be expected. With the exception of aminopropyltriethoxysilane deposited from vapor phase7 and aqueous solution,39 TES-based modifications have not been extensively evaluated as modifiers of ZnO. Managing siloxane formation to achieve monolayers on a planar substrate without the 3D growth of a polymer is very difficult. However, because ethoxysilanes hydrolyze at much slower rates than do chlorosilanes or (27) Ravirajan, P.; Peiro, A. M.; Nazeeruddin, M. K.; Graetzel, M.; Bradley, D. D. C.; Durrant, J. R.; Nelson, J. J. Phys. Chem. B 2006, 110, 7635–7639. (28) Greene, L. E.; Law, M.; Tan, D. H.; Montano, M.; Goldberger, J.; Somorjai, G.; Yang, P. Nano Lett. 2005, 5, 1231. (29) Shen, W.; Zhao, Y.; Zhang, C. Thin Solid Films 2005, 483, 382–387. (30) Monson, T. C.; Lloyd, M. T.; Olson, D. C.; Lee,Y.-J.; Hsu, J. W. P. AdV. Mater., in press (DOI: 10.1002/adam.200801082). (31) Olson, D. C.; Lee, Y. J.; White, M. S.; Kopidakis, N.; Shaheen, S. E.; Ginley, D. S.; Voigt, J. A.; Hsu, J. W. J. Phys. Chem. C 2008, 112, 9544–9547. (32) Peet, J.; Kim, J. Y.; Coates, N. E.; Ma, W. L.; Moses, D.; Heeger, A. J.; Bazan, G. C. Nat. Mater. 2007, 6, 497–500. (33) Salleo, A.; Chabinyc, M. L.; Yang, M. S.; Street, R. A. Appl. Phys. Let. 2002, 81, 4383–4385. (34) Taratula, O.; Wang, D.; Chu, D.; Galoppini, E.; Zhang, Z.; Chen, H.; Saraf, G.; Lu, Y. J. Phys. Chem. B 2006, 110, 6506–6515. (35) Ogata, K.; Hama, T.; Hama, K.; Koike, K.; Sasa, S.; Inoue, M.; Yano, M. Appl. Surf. Sci. 2005, 241, 146–149. (36) Hou, X.; Zhou, F.; Yu, B.; Liu, W. Mater. Sci 2007, 452-453, 732–736. (37) Jeon, K. A.; Son, H. J.; Kim, C. E.; Shon, M. S.; Yoo, K. H.; Choi, A. M.; Jung, H. I.; Lee, S. Y. Sensors 2006, 1265–1268. (38) Law, M.; Greene, L. E.; Johnson, J. C.; Saykally, R.; Yang, P. Nat. Mater. 2005, 4, 455–459. (39) Watts, B.; Thomsen, L.; Dastoor, P. C. Synth. Met. 2005, 152, 21–24.
Allen et al.
methoxysilanes,40 less oligomer forms in the bulk of the solution, minimizing the potential for producing multilayer films. For the same reason, the surface reaction does not proceed as readily. However, it has previously been shown41 that high-quality monolayers of octadecylsiloxane (ODS) could be produced on glass and oxidized Si from solutions of OTES in toluene when a small amount of n-butylamine catalyst is added. The addition of n-butylamine to the deposition solution promoted the reaction of OTES ethoxy groups with surface hydroxyls. The proposed reaction mechanism involved the adsorption of the amine to a surface OH, resulting in a strongly nucleophilic site that attracted the OTES.42 The organosilane is believed to form a pentacoordinated intermediate through nucleophilic attack. The ethoxysilane group would then extract a proton from the amine, which would be regenerated with a proton from the OH. The byproducts of this reaction are ethanol and butylamine, plus a single covalent bond to the surface. The amine is also believed to catalyze the formation of a siloxane network by aiding in the cross-linking of adjacent surface-anchored species. Our prior experiences with forming ODS layers on silicon demonstrated that poor monolayers were formed under completely anhydrous conditions. Trace amounts of water were found to be necessary in the deposition solution. We believe that the trace water was provided by handling the chemicals in ordinary laboratory air. This suggested that some of the OTES, hydrolyzed into octadecyldiethoxyhydroxysilane in the bulk of the solution, adsorbed onto the surface as a mobile species.41 Conversion of the species to a monolayer, through covalent linking to the surface or cross-linking to molecules already in the layer, would have then occurred by a dehydration reaction promoted by the amine. Despite the fact that subtle details of this surface-modification method are not completely understood, it has the potential to allow broadening of the range of materials on which siloxanebased molecular layers can be formed, particularly to metal oxides such as ZnO that are easily hydroxylated. It has previously been reported that the hydroxylation of ZnO is increased upon exposure to ultraviolet radiation.43,44 When illuminated with above band gap light (>3.2 eV), electron-hole pairs are generated in ZnO. Electrons and holes that do not recombine can move to the surface and react with adsorbed surface species, providing the basis for photocatalysis. Additionally, oxygen vacancies can form via hole reactions with lattice oxygen. In the presence of a moisture-containing atmosphere, these vacancies become hydroxylated.44 For this work, a commercial ultraviolet ozone (UVO) cleaner was used to remove organic contaminants and promote the hydroxylation of the ZnO surface. Hydroxylated ZnO samples were treated with a series of TES molecules and characterized using water contact angles (CA), Fourier transform infrared (FTIR) spectroscopy, and X-ray photoemmission spectroscopy (XPS). OTES-modified surfaces were compared to silicon (Si) surfaces modified by monolayers of stearic acid (n-C17H35COOH) produced by the LangmuirBlodgett technique. The efficacy of the n-butylamine-catalyzed deposition approach was established by depositions using two additional TES molecules, DTES and PTES. DTES served as a (40) Francis, R.; Louche, G.; Duran, R. S. Thin Solid Films 2006, 513, 347– 355. (41) Walba, D. M.; Liberko, C. A.; Korblova, E.; Farrow, M.; Furtak, T. E.; Chow, B. C.; Schwarz, D. K.; Freeman, A. S.; Douglas, K.; Williams, S. D.; Klittnick, A. F.; Clark, N. A. Liq. Cryst. 2004, 31, 481–489. (42) Blitz, J. P.; Shreedhara; Leyden, D. E. J. Colloid Interface Sci. 1988, 126, 387–392. (43) Asakuma, N.; Fukui, T.; Toki, M.; Awazu, K.; Imai, H. Thin Solid Films 2003, 445, 284–287. (44) Sun, R. D.; Nakajima, A.; Fujishima, A.; Watanabe, T.; Hashimoto, K. J. Phys. Chem. B 2001, 105, 1984–1990.
Surface Modification of ZnO
Langmuir, Vol. 24, No. 23, 2008 13395
Figure 1. (a) XRD angle scan from a sol-gel-grown ZnO film. (b) AFM image of the surface of a sol-gel-grown ZnO film showing the location of the line along which the height profile (shown in c) was measured.
reasonable first extension of the chemistry beyond OTES because it has the same alkyl R-group structure but with a shorter chain. PTES was investigated to determine the applicability of the aminecatalyzed surface reaction to molecules with different surface terminations and wetting properties. Our results indicated that all of the these molecules could be used to modify the surface.
Methods ZnO Layers. ZnO planar films (∼30 nm thick) were produced from sol-gel precursors. A solution of 0.75 M zinc acetate dihydrate (99.999%, Aldrich) and 0.75 M ethanol amine (99%, Aldrich) in 2-methoxyethanol (98% anhydrous, Aldrich)45 was spin-coated onto Si wafers. These were reacted at 300 °C on a hot plate for 10 min. The samples were then rinsed with deionized water, acetone, and ethanol and were blown dry with a stream of nitrogen. ZnO Pretreatment. The ZnO films were prepared for the TES modifications by ultrasonic cleaning in acetone (99.5%, Pharmco, CT) for 5 min followed by a blow drying treatment using nitrogen. For the purpose of quantitative comparison of the OTES-treated samples to an LB-deposited stearic acid layer, we initially restricted the deposition of OTES to only the ZnO side of the substrate by using a photoresist mask, which was cured at 100 °C, on the back side of the silicon wafer. However, we found that even without the photoresist, peak frequencies, line shapes, and integrated intensities in the FTIR spectra were nearly the same as when the back side of the sample was protected. We presume that the unprotected back surface was dehydroxylated during the ZnO reaction so that subsequent silane bonding was prevented on that side. Thus, to reduce the risk of contamination from the photoresist, further studies did not include this step. To promote hydroxylation on the ZnO surfaces the samples were placed, with the ZnO facing up, into a UVO cleaner (Jelight, CA) for 15 min. The power density of the 254 nm mercury emission line was measured to be 20 mW/cm2 near the location of the samples. Triethoxysilane Treatments. Solutions of OTES, DTES, and PTES (Gelest, PA) were prepared from as-received chemicals by mixing 1.50, 1.08, and 2.50 mL of the respective TES molecule with 35 mL of toluene (99.5% Pharmco, CT) and 0.50 mL of n-butylamine (Acros, NJ) in a base bath cleaned staining jar. The jars were filled to a final volume of 70 mL by the addition of more toluene.41 For each deposition, the jar of solution was heated in a 45 °C water bath for 30 min. The ZnO-coated substrates were placed into the heated TES solution, immediately after their removal from the UVO cleaner. After 90 min, the samples were removed from the solution, thoroughly rinsed with toluene and acetone, and blown dry with nitrogen. Samples were then baked in air for 60 min at 110 °C and again rinsed with acetone and blown dry. Control samples were prepared in an identical manner but were submerged in solutions of toluene and n-butylamine containing no TES molecules.
Stearic Acid Layer on Si. OTES molecular layers were compared to layers of stearic acid that were deposited on double-side-polished Si using a commercial Langmuir-Blodgett (LB) trough (NIMA 611M). An LB film of stearic acid on silicon46 is a useful calibration because it forms a close-packed, uniform monolayer with nearly the same number of methylene groups as OTES. The aqueous subphase was a 5 × 10 -5 M cadmium chloride solution buffered to a pH of 6.5 using sodium bicarbonate. The stearic acid was diluted to 2 mg/mL with chloroform, and 5 µL of the dilution was floated on the surface of the subphase. After compressing the surface at 12 cm2/min to a surface pressure of 20 mN/m, the Si was extracted at a rate of 1 mm/min. Characterization. X-ray diffraction (XRD) angle scans from sol-gel samples were collected with a 2D large-area detector and an x-y sample-positioning stage (Bruker D8). Atomic force microscopy (AFM) was used to evaluate the surface topography and roughness of the ZnO substrates. AFM images were collected in tapping mode (Veeco, Nanoscope III) using tips with high aspect ratios and 127 µm cantilevers (Nanoworld, Switzerland). Water CA averages and standard deviations for each surface type were calculated from multiple measurements on at least four identically prepared samples using water droplets that had volumes ranging from 1.5 to 2.5 µL. In each measurement, the outline of the drop in the recorded image was fit to a constrained curve using a polynomial spline.47 Normal incidence FTIR transmission spectra were obtained using a Nicolet Magna Jr. 560 equipped with a calcium fluoride beam splitter and a liquid-nitrogen-cooled mercury-cadmium-telluride detector. Each spectrum represents the average of 800 scans collected at 4 cm-1 resolution. The absorbance spectra were calculated as the logarithm of the transmission of the modified sample divided by the transmission from an appropriate control sample. For XPS measurements, a Kratos Analytical (Manchester, U.K.) HSi instrument utilizing a monochromated Al KR source (1486.7 eV) was used. Measurements were performed with a source power of 120 W and an electron takeoff angle of 0°.
Results The XRD angle scans recorded from the sol-gel-grown ZnO films are shown in Figure 1a. These results confirm that the sol-gel method produced ZnO.48 Samples were polycrystalline, as demonstrated by peaks in the XRD scan associated with the (100), (002), and (101) crystal planes of ZnO. The AFM image of a ZnO film shown in Figure 1b gives a root-mean-square (45) Ohyama, M.; Kouzuka, H.; Yoko, T. Thin Solid Films 1997, 306, 78–85. (46) Schwartz, D. K. Surf. Sci. Rep. 1997, 27, 241–334. (47) Stalder, A. F.; Kulik, G.; Barbieri, L.; Hoffmann, P. Colloids Surf. 2006, 286, 92–103. (48) Schulz, H.; Thiemann, K. H. Solid State Commun. 1979, 32, 783–785.
13396 Langmuir, Vol. 24, No. 23, 2008
surface roughness of 2.2 nm and indicates that grain sizes were typically tens of nanometers. This is confirmed by height profiles along lines across the image, such as that illustrated in Figure 1c. Images of the ZnO films before and after organic molecule depositions were not discernably different, which is an indication that the deposition process did not etch the ZnO surface. By contrast, ZnO films exposed to even very dilute acid solutions exhibit rounding of the surface features and reduced film thicknesses. In general, the roughness of a surface can effect any CA measurement.49 However, we found that the sol-gel method was very reproducible and that identically grown ZnO films had average surface roughness levels that were identical. This enabled us to use CA measurements on sessile drops as a qualitative assay of the nature of ZnO surfaces before and after treatment with the organic layers. These measurements revealed that the 15 min treatment in the UVO cleaner resulted in hydrophilic ZnO surfaces with an average CA value of 17.6 ( 3.6°. This is in agreement with previously reported values for hydroxylated surfaces (18.7-20.2°)50 and with previously observed values for ZnO exposed to light from a xenon arc lamp.51 The alkylmodified ZnO surfaces were much more hydrophobic than the surfaces of the control samples, which were exposed to treatment solutions without TES molecules. OTES and DTES CA values were found to be 106.0 ( 1.6 and 102.0 ( 2.2°, respectively, whereas the CA values on the control samples were 59.2 ( 14.8°. The CA values on the ZnO control samples were more variable, as indicated by the large standard deviation between measurements. We attribute this to the instability of the control surface because the CA values on these surfaces were observed to change as a function of time. In contrast, alkyl-treated surfaces gave essentially the same CA results immediately after treatment, after months of storage, or after aggressive rinsing with a wide variety of solvents. Annealing was performed on OTES-treated ZnO to temperatures as high as 180 °C, with no observed change in the CA. These results are consistent with the conclusion that the organic molecular layers are covalently attached to the surface. The CA values for the alkyl-based layers were less than what has previously been reported for these types of alkyl treatments on silicon.41,52 Figure 2 shows the FTIR absorbance spectra of alkyl-treated surfaces in the spectral region of the C-H stretching modes. In comparing the spectrum of the OTES-derivatized surface on sol-gel-grown ZnO to an LB stearic acid layer on Si (corrected for being double-sided), we note that the integrated intensity of the C-H stretching modes in the OTES layer is about 75% of that observed in the stearic acid layer. This result is consistent with the water CA data and indicates that the surface coverage was incomplete. This submonolayer coverage was consistently observed in many samples over a broad range of deposition conditions and was found not only on sol-gel-produced ZnO but also on ZnO produced by other techniques, such as radio frequency sputtering. Additional information about the nature of the molecular layer can be inferred from the positions of the C-H stretching peaks. It is well established that the peak frequencies of the C-H stretching modes are a good indicator of the conformation of alkyl chains. For all-trans alkyl chains, the reported frequencies for the symmetric mode (νs) are from 2846 to 2850 cm-1. The antisymmetric (νa) mode frequencies (49) Sheng, Y. J.; Jiang, S.; Tsao, H. K. J. Chem. Phys. 2007, 127, 234704. (50) Ingall, M. D. K.; Honeyman, C. H.; Mercure, J. V.; Bianconi, P. A.; Kunz, R. R. J. Am. Chem. Soc. 1999, 121, 3607–3613. (51) Rico, V.; Lopez, C.; Borras, A.; Espinos, J. P.; Gonzalez-Elipe, A. R. Sol. Energy Mater. Sol. Cells 2006, 90, 2944–2949. (52) Pamidighantam, S.; Laureyn, W.; Rusu, C.; Baert, K.; Puers, R.; Tilmans, H. A. C. Sens. Actuators, A 2003, 103, 202–212.
Allen et al.
Figure 2. FTIR absorbance spectra in the C-H stretching region: (b) OTES-treated ZnO, (2) DTES-treated ZnO, and (9) stearic acid LB layer on Si (intensity corrected to represent a single monolayer).
are from 2915 to 2920 cm-1.53,54 This is in agreement with the peak locations found for the stearic acid LB layer and previous reports of alkylsiloxane modification of SiO2 substrates.53,55 For the alkyltriethoxysilane layers on ZnO, the νa and νs peaks are located at 2924 and 2854 cm-1, respectively. These values are between those associated with densely packed layers and those measured for liquidlike alkyl chains (νs ) 2856 cm-1 and νa ) 2928 cm-1).56 This indicates that the lower CA was at least partially caused by conformational disorder, which is consistent with incomplete coverage.57 A perfect trans-extended monolayer is terminated by methyl groups and yields a CA >110° on a smooth surface. Disorder allows methylene groups to be exposed on the surface, which leads to a lower CA.3 Comparing the OTES and DTES spectra, we see that the integrated absorbance is not only smaller for DTES but also broader. The slightly lower CA and broader absorption peaks recorded from the DTES-modified samples, compared to those treated with OTES, are consistent with a higher density of gauche defects58 in the shorter alkyl chains. These results indicated that functionalization of ZnO with DTES was very similar to that of OTES, which motivated further exploration of the deposition chemistry to include molecules that have different terminal groups. PTES-treated ZnO samples had a water CA of 72.5 ( 4.3°. Although the CAs for PTES -treated surfaces were more variable than those obtained on the OTES- and DTES-treated surfaces, the average value was in reasonable agreement with previously reported values for phenyl-terminated surfaces prepared on SiO2 (79.0°).50 This value is also much larger than reported for PTES-modified indium tin oxide (ITO) used in organic lightemitting diodes.20 This suggests that the surface coverage of PTES on ITO was much less than a monolayer. (53) Parikh, A. N.; Allara, D. L.; Azouz, I. B.; Rondelez, F. J. Phys. Chem. 1994, 98, 7577–7590. (54) MacPhail, R. A.; Strauss, H. L.; Snyder, R. G.; Elliger, C. A. J. Phys. Chem. 1984, 88, 334–341. (55) Howland, M. C.; Johal, M. S.; Parikh, A. N. Langmuir 2005, 21, 10468– 10474. (56) Snyder, R. G.; Strauss, H. L.; Elliger, C. A. J. Phys. Chem. 1982, 86, 5145–5150. (57) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558–569. (58) Spori, D. M.; Venkataraman, N. V.; Tosatti, S. G. P.; Durmaz, F.; Spencer, N. D.; Zurcher, S. Langmuir 2007, 23, 8053–8060.
Surface Modification of ZnO
Figure 3. X-ray photoemission spectra showing the Si 2p emission. The binding energy of this peak is an indication that the Si was in the 3+ charge state. The sloping background in these spectra is caused by emission from Zn 3s.
Whereas infrared absorption, determined from direct transmission through PTES-treated layers, was not sufficiently sensitive to detect absorbance features in the phenyl stretching region, XPS could be used to detect the presence of the layer on ZnO. In Figure 3, similar Si 2p photoemission was observed for all three of the TES functionalizations. After the inclined background was subtracted, the peak location was found to be 102.5 eV, which is associated with Si in the 3+ charge state. This is consistent with the existence of a polysiloxane molecular layer59,60 and agrees with spectra reported by Yakimova et al., who used an evaporative APTES surface treatment to functionalize ZnO.7 Additionally, we note that integrated intensities of Si 2p are all comparable whereas Zn 3s and O 1s peak intensities showed the expected trend. That is, the XPS intensity from ZnOtreated with PTES was greater than the intensity from surfaces treated with DTES, which was greater than the intensity from surfaces treated with OTES. This trend arises because thicker carbon layers decrease the probability that electrons will escape from Zn and O, which is buried under the molecular layers. C 1s core emissions showed the reverse trend with the layers that have the largest carbon content giving the largest signal. These trends are comparable to XPS studies performed on alkylterminated layers formed from phosphonic acid molecule modification of titanium dioxide.58
Discussion and Conclusions The results of this study demonstrate that the aminecatalyzed silane attachment protocol, which was developed and studied on silicon dioxide and glass substrates, can be extended to ZnO and likely to a broad range of metal oxides. These materials are now receiving considerable attention in applications where control of surface electronic and structural properties are critical such as hybrid metal oxide-organic solar cells. With different R-triethoxysilane precursors, it should be possible to achieve a broad range of surface functionalizations through variation of the R end group while allowing considerable control of the surface condensation reaction involved in attachment. A remaining question, however, concerns the fraction of the surface covered by monolayers formed in this way. The (59) Flis, J.; Kanoza, M. Electrochim. Acta 2006, 51, 2338–2345. (60) Kallury, K. M. R.; Krull, U. J.; Thompson, M. Anal. Chem. 1988, 60, 169–172. (61) Kulkarni, S. A.; Kakade, B. A.; Mulla, I. S.; Pillai, V. K. J. Colloid Interface Sci. 2006, 299, 777–784.
Langmuir, Vol. 24, No. 23, 2008 13397
CA and infrared results do provide evidence of disorder and incomplete coverage for OTES and DTES functionalizations. Although less coverage information was obtained for PTES, it is likely that a complete monolayer was not formed with this treatment. We propose that the incomplete surface coverage observed in these treatments was primarily associated with the level of surface hydroxylation of the ZnO films. We explored this relationship by varying the time of exposure of the samples in the UVO treatment. The extent of hydroxylation should be related to the density of surface defects produced by this treatment. We observed that the coverage of OTES was very small if no UVO treatment was used. The coverage increased as the UVO treatment time was increased up to 15 min, after which additional UVO exposure did not influence the coverage. This suggests that a significant factor that limits the surface coverage in the TES-based deposition is the density of surface defects that can be produced to enable hydroxylation. We also observed a correlation between the surface coverage and the presence of trace amounts of water in the deposition solutions. When solutions were prepared using a new bottle of anhydrous n-butyamine, we observed smaller surface coverages relative to the cases when molecular layers were formed with the same n-butyamine source after the bottle had been opened several times. During use, the solutions were exposed to ambient laboratory conditions with the relative humidity near Denver averaging about 35%. These observations are similar to those made during OTES molecular layer formation on SiO2. Motivated by the possibility that water contamination might be effecting coverage, experiments were performed in an environment with a higher relative humidity (50%). The initial deposition in that environment resulted in an improvement in the ordering of the alkyl chains (inferred from the frequencies of the C-H stretch vibrational modes). This is consistent with formation, in the presence of water and n-butylamine, of a reaction intermediate that would be mobile on the surface. This supports the hydrolyzation-dehydration mechanism discussed earlier. Subsequent depositions in the high-humidity environment resulted in lower CA values and more disordered layers. The integrated intensity of the C-H stretching band absorbance also nearly doubled and was greater than that observed from the stearic acid layer. We interpret this as evidence of polymerization of the OTES molecules in the bulk solution, caused by excess water from the ambient atmosphere, with subsequent collection of the polymerized material on the surface leading to thick layers. All of these observations confirm that moisture plays a significant role in the formation of silane-coupled layers, even with the amine catalyst, which is capable of promoting a water-free mechanism. Farrow62 made a similar speculation about the potential role of trace amounts of water in the amine-assisted reaction of OTES with hydroxylated SiO2. Similar sensitivity to humidity has also been reported for the noncatalyzed synthesis of surface molecular layers from octadecyltrichlorosilane.63 Whereas it is desirable, in general, to form a complete monolayer during the molecular treatment of a surface, if submonolayer coverage can be reproducibly achieved there may be advantages in some applications. For example, whereas a complete octadecylsiloxane monolayer is electrically insulating,61 a submonolayer version of this might still improve wetting and polymer ordering on ZnO surfaces while allowing a useful amount of charge transfer. Future work will investigate (62) Farrow, M. J. Ph.D. Thesis, University of Colorado, Boulder, CO, 2004. (63) Wasserman, S. R.; Tao, Y. T.; Whitesides, G. M. Langmuir 1989, 5, 1074–1087.
13398 Langmuir, Vol. 24, No. 23, 2008
controlling the coverage through control of the level of surface hydroxylation of ZnO by manipulating the oxygen vacancy density at the surface and by controlling the moisture content of the deposition solutions. This control, combined with the ability to investigate a wide variety of R-end groups, could make amine-catalyzed TES functionalization very useful for a wide variety of hybrid organic-inorganic applications.
Allen et al.
Acknowledgment. We acknowledge valuable discussions with and assistance from George Radziszewski, Joseph Dahdah, Tracy Berman, Matt Bergren, Emily Prezekwas, Christian Weigand, David Ginley, and Don Williamson. This report is based on work supported by the National Science Foundation under grant no. DMR-0606054. LA802621N