Bonding and Structure of Glycine on Ordered Al2O3 Film Surfaces

Langmuir 2004, 20, 10551-10559

10551

Bonding and Structure of Glycine on Ordered Al2O3 Film Surfaces G. Tzvetkov,*,† G. Koller,† Y. Zubavichus,‡ O. Fuchs,§ M. B. Casu,§ C. Heske,§,| E. Umbach,§ M. Grunze,‡ M. G. Ramsey,† and F. P. Netzer† Institut fu¨ r Experimentalphysik, Karl-Franzens-Universita¨ t Graz, Universita¨ tsplatz 5, A-8010 Graz, Austria, Angewandte Physikalische Chemie, Universita¨ t Heidelberg, Im Neuenheimer Feld 253, D-69120 Heidelberg, Germany, Experimentelle Physik II, Universita¨ t Wu¨ rzburg, Am Hubland, D-97074 Wu¨ rzburg, Germany, and Department of Chemistry, University of Nevada, Las Vegas, Nevada 89154-4003 Received May 23, 2004. In Final Form: August 2, 2004 The interaction between glycine (NH2CH2COOH) layers and an ultrathin Al2O3 film grown epitaxially onto NiAl(110) was studied by temperature-programmed desorption, X-ray photoelectron spectroscopy, ultraviolet photoelectron spectroscopy, work function measurements, and near-edge X-ray absorption fine structure (NEXAFS) spectroscopy. At monolayer coverages at 110 K, there are two coexisting molecular forms: the anionic (NH2CH2COO-) and the zwitterionic form (NH3+CH2COO-) of glycine. As deduced from the photoemission data, the buildup of multilayers at 110 K leads to a condensed phase predominantly in the zwitterionic state. In contrast to the monolayer at 110 K, the monolayer formed at 300 K consists primarily of glycine molecules in the anionic state. The latter species is adsorbed with the oxygen atoms of the carboxylic group pointing toward the substrate. The polarization-dependent C K- and O K-edge NEXAFS spectra indicate that the glycinate species in the monolayer at 300 K is oriented nearly perpendicular to the surface, with the amino group pointing away from the surface.

1. Introduction There is a growing interest in the adsorption of amino acids (NH2CHRCOOH, with R representing the side group which is different for each amino acid) on well-defined surfaces under ultrahigh vacuum (UHV) conditions.1-11 This is due to the fact that this subject is of prime importance for the development and the fundamental understanding of organic/inorganic interfaces. The organic functionalization of inorganic substrates has important applications, for example, in biomedical and newly developed biomaterials fields,12,13 catalysis,14 sensors,15,16 * To whom correspondence should be addressed. E-mail: [email protected]. † Karl-Franzens-Universita ¨ t Graz. ‡ Universita ¨ t Heidelberg. § Universita ¨ t Wu¨rzburg. | University of Nevada. (1) Lo¨fgren, P.; Krozer, A.; Lausmaa, J.; Kasemo, B. Surf. Sci. 1997, 277, 370. (2) Booth, N. A.; Woodruff, D. P.; Schaff, O.; Giessel, T.; Lindsay, R.; Baumga¨rtel, P.; Bradshaw, A. M. Surf. Sci. 1998, 397, 258. (3) Barlow, S. M.; Kitching, K. J.; Haq, S.; Richardson, N. V. Surf. Sci. 1998, 401, 322. (4) Williams, J.; Haq, S.; Raval, R. Surf. Sci. 1996, 368, 303. (5) Hasselstro¨m, J.; Karis, O.; Weinelt, M.; Wassdahl, N.; Nilsson, A.; Nyberg, M.; Pettersson, L. G. M.; Samant, M. G.; Sto¨hr, J. Surf. Sci. 1998, 407, 221. (6) Hasselstro¨m, J.; Karis, O.; Nyberg, M.; Pettersson, L. G. M.; Weinelt, M.; Wassdahl, N.; Nilsson, A. J. Phys. Chem. B 2000, 104, 11480. (7) Chen, Q.; Frankel, D. J.; Richardson, N. V. Surf. Sci. 2002, 497, 37. (8) Efstathiou, V.; Woodruff, D. P. Surf. Sci. 2003, 531, 304. (9) Tzvetkov, G.; Ramsey, M. G.; Netzer, F. P. Surf. Sci. 2003, 526, 383. (10) Zhao, X.; Yan, H.; Zhao, R. G.; Yang, W. S. Langmuir 2002, 18, 3910. (11) Ernst, K. H.; Christmann, K. Surf. Sci. 1989, 224, 277. (12) Kasemo, B. Curr. Opin. Solid State Mater. Sci. 1998, 3, 5. (13) Jones, F. H. Surf. Sci. Rep. 2001, 43, 75. (14) Raval, R. J. Phys.: Condens. Matter 2002, 14, 4119. (15) Kong, J.; Franklin, N. R.; Zhou, Ch.; Chapline, M. G.; Peng, S.; Cho, K.; Dai, H. Science 2000, 287, 622. (16) Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Science 2001, 293, 1289.

and molecular electronic devices.17,18 Within such a technological context, the role of amino acids is twofold. First, as the simplest biologically active molecules and building blocks of proteins, adsorption studies of amino acids are the first steps toward the understanding of surface phenomena of more complicated biomolecules (peptides, proteins, etc.). Second, amino acids are able to form ordered superstructures on metal surfaces, which offers an opportunity for the formation of extended chiral surfaces.7 The latter is of significance for the fundamental mechanism of heterogeneous enantioselective catalytic systems. The main objectives of studying amino acid adsorption on well-defined surfaces are to characterize the chemical form, bonding, and orientation of adsorbed molecules. An important property of amino acids is that they are characterized by two active functional groups (NHx and COO(H)), which define the chemical form of the molecules and can be involved in the bonding to the substrate. The amino acids exist in zwitterionic form (NH3+CHRCOO-) in the solid state and in the neutral or acidic form (NH2CHRCOOH) in the gas phase. Depending on the pH, the chemical form of amino acids in solution varies from cationic (NH3+CHRCOOH) through zwitterionic to anionic (NH2CHRCOO-). Glycine (NH2CH2COOH, also denoted R-amino acetic acid), the simplest amino acid, has attracted particular attention as a basic adsorption model system. Using a range of surface science techniques, the interaction of glycine with Pt(111),1,11,19 Pt(100),19 Cu(110),2,3,5,6 and Cu(100)8,20,21 surfaces has been extensively investigated in recent years. (17) Chen, J.; Reed, M. A.; Rawlett, A. M.; Tour, J. M. Science 1999, 286, 1550. (18) Collier, C. P.; Wong, E. W.; Belohradsky, M.; Raymo, F. M.; Stoddart, J. F.; Kuekes, P. J.; Williams, R. S.; Heath, J. R. Science 1999, 285, 391. (19) Stern, D. A.; Salaita, G. N.; Lu, F.; McCargar, J. W.; Batina, N.; Frank, D. G.; Laguren-Davidson, L.; Lin, C.-H.; Walton, N.; Gui, J. Y.; Hubbard, A. T. Langmuir 1988, 4, 711.

10.1021/la048726m CCC: $27.50 © 2004 American Chemical Society Published on Web 10/20/2004

10552

Langmuir, Vol. 20, No. 24, 2004

In contrast to metal surfaces, amino acid adsorption studies on well-defined oxide surfaces are quite rare. Soria et al.22 have examined multilayers of glycine in zwitterionic form, adsorbed on TiO2(110) at room temperature, by photoelectron spectroscopy of the valence band, the O 2s band, and the Ti 3p levels using synchrotron radiation in the energy range from 15 to 150 eV. The authors found that photon damage occurs fast (cross section of 5.4 × 10-16 cm2, comparable with the electron excitation cross section in the gas phase). After photon exposure of glycine multilayers, the coverage was reduced to less than two monolayers; the remaining surface species contained C and NHx (x ) 1, 2), but their chemical state was not identified. Furthermore, dissociative adsorption of glycine on a TiO2(110)-(1 × 2) reconstructed surface at the submonolayer stage has been observed.23 In the latter case, carboxylic fragments of glycine dominated at the surface, while most of the amino groups were released into the gas phase. Kasemo and co-workers24 studied the adsorption of glycine on a TiO2 surface by means of temperature-programmed desorption (TPD). They reported that heating of the glycine multilayer results in the desorption of both intact molecules and dissociation products. The molecules from the glycine multilayer desorb at ∼310 K, while the saturated monolayer bound directly to the surface partly (∼40%) desorbs at ∼400 K as intact molecules; at higher temperatures, only dissociation fragments were detected. In addition, the coadsorption of water and glycine, a system of relevance in biomaterials research, has been studied.24 It was found that the water molecules do not affect the glycine TPD spectra but that the bonding of H2O to the surface is changed by the presence of glycine. In this study, we have investigated the adsorption of glycine on an ultrathin Al2O3 film epitaxially grown on NiAl(110) under UHV conditions. This oxide film is wellordered and purely oxygen-terminated; the sequence of layers from bulk to surface is most likely NiAl-Al-OAl-O, that is, two oxygen/aluminum bilayers with an overall thickness of ∼5 Å.25 As suggested in ref 25, the oxide film is considered to mimic the γ-Al2O3(111) surface. Prepared in this way, the thin alumina film has several important advantages over bulk single crystal surfaces. In particular, charging of the sample during ion and electron spectroscopic measurements is completely avoided. This allows most of the surface science techniques to be used in order to elucidate the elementary steps of the interaction between adsorbates and oxide surface. The goal of this contribution is to investigate the adsorption of glycine on an Al2O3/NiAl(110) surface at 110 K and at room temperature, examining the chemical state of the molecules in both the mono- and multilayer range. The molecule-substrate interaction and bonding geometry of the glycine species in monolayers, adsorbed at different conditions, have been analyzed. The studies were performed using a variety of surface science techniques including TPD, ultraviolet photoelectron spectroscopy (UPS), work function (∆φ) measurements, X-ray photoelectron spectroscopy (XPS), and near-edge X-ray (20) Atanasoska, L. L.; Buchholtz, J. C.; Somorjai, G. A. Surf. Sci. 1978, 72, 189. (21) Zhao, X.; Gai, Z.; Zhao, R. G.; Yang, W. S.; Sakurai, T. Surf. Sci. 1999, 424, L347. (22) Soria, E.; Roman, E.; Williams, E. M.; de Segovia, J. L. Surf. Sci. 1999, 433-435, 543. (23) Soria, E.; Colera, I.; Roman, E.; Williams, E. M.; de Segovia, J. L. Surf. Sci. 2000, 451, 188. (24) Lausmaa, J.; Lo¨fgren, P.; Kasemo, B. J. Biomed. Mater. Res. 1999, 44, 227. (25) Jaeger, R. M.; Kuhlenbeck, H.; Freund, H.-J.; Wuttig, M.; Hoffmann, W.; Franchy, R.; Ibach, H. Surf. Sci. 1991, 259, 235.

Tzvetkov et al.

absorption fine structure (NEXAFS) spectroscopy. To our knowledge, this is the first comprehensive study of amino acid adsorption on an oxide surface under UHV conditions. 2. Experimental Section The TPD, UPS, ∆φ, and XPS measurements were performed in a custom-designed µ-metal UHV chamber with a base pressure of ∼1 × 10-10 mbar. The details of the apparatus have been described elsewhere.9 The UPS (XPS) spectra in normal emission geometry were collected with photon incidence angles of 40° (65°) with respect to the surface normal. The UPS spectra were excited with He II radiation, and the hemispherical electron energy analyzer (Omicron, EA 125) was operated with a fixed pass energy of 5 eV, corresponding to an energy resolution of 0.12 eV. The XPS spectra were taken using an unmonochromatized Al KR source (1486.6 eV) operated at a power of 180 W and a pass energy of 20 eV (resolution of 0.85 eV). The binding energy scale was referenced to the Ni 2p3/2 peak at 852.7 eV.26 To minimize the X-ray-induced damage in the adsorbed glycine layers, each XPS spectrum was taken from a freshly prepared glycine-covered surface with an overall acquisition time of about 20 min. The fitting of the XPS spectra was performed by using Lorentzian line shapes convoluted with a major Gaussian contribution as model functions for the individual components. In several cases, XPS and UPS difference spectra, (glycine covered)-minus-(clean substrate), are presented to emphasize the adsorbate-induced features. For the differencing procedure, the clean surface spectrum was attenuated so that negative “overshooting” in the difference spectra is avoided or minimized. The work function was determined by the low-energy onset of the valence band spectra using He I radiation. The sample was biased (-9 V) to improve the secondary electron cutoff. TPD experiments were performed with a line-of-sight detection quadrupole mass spectrometer, which is surrounded by a liquid nitrogen cooled stainless steel shield in order to suppress the detection of gases desorbing from the crystal holder and chamber walls; spectra were collected at a linear heating rate of 1 K/s. The NEXAFS measurements were performed at the PM-1 and PM-3 bending magnet beamlines at the BESSY II synchrotron radiation laboratory in Berlin. The NEXAFS spectra (C K- and O K-edges) were recorded in the Auger yield detection mode by a spherical sector analyzer (Vacuum Generators, CLAM 2) with an angle between the electric field vector of the photon beam and the normal of the crystal surface of 90° (normal incidence) and 20° (grazing incidence). The raw spectra of the glycine multilayers were normalized to the incident photon flux by division through the current collected on a gold grid. The monolayer spectra were obtained by subtracting the signal of the clean substrate from the signal of the glycine-covered surface. The experimental endstation, from the University of Wu¨rzburg, has been described elsewhere.27 The sample could be cooled to ∼100 K and heated to ∼1200 K in both UHV systems used in the present work. Alumina films were prepared by oxidation of a NiAl(110) single crystal surface, as established earlier.25,28 The NiAl(110) crystal was cleaned by several cycles of ion sputtering (1.5 kV Ar+ for 15 min) and subsequent short time annealing at ∼1150 K. The oxide was formed by backfilling the chamber to 7 × 10-6 mbar of oxygen for 20 min at a sample temperature of 550 K. The crystal was finally briefly annealed at ∼1150 K. The structure and cleanliness of the surface were checked by low-energy electron diffraction (LEED) and XPS. Vapor deposition of glycine was achieved from a specially designed Knudsen cell-type molecule evaporator.9 The glycine powder (>99%, Fluka) was introduced into a small copper cell heated by a button heater attached to its backside. This evaporation source was separated from the UHV chambers by a gate valve and pumped by a turbomolecular pump. Before (26) Handbook of X-ray Photoelectron Spectroscopy; Chastain, J., Ed.; Physical Electronics Division, Perkin-Elmer Corp.: Eden Prairie, MN, 1992. (27) Taborski, J.; Va¨terlein, P.; Dietz, H.; Zimmermann, U.; Umbach, E. J. Electron Spectrosc. Relat. Phenom. 1999, 75, 129. (28) Va¨terlein, P.; Schmelzer, M.; Taborski, J.; Krause, T.; Vicsian, F.; Ba¨ssler, M.; Fink, R.; Umbach, E.; Wurth, W. Surf. Sci. 2000, 452, 20.

Bonding and Structure of Glycine

Figure 1. Temperature-programmed desorption spectra of intact glycine molecules (represented by m/z ) 30) from various glycine exposures adsorbed on an Al2O3/NiAl(110) surface at 110 K. The insert shows an expanded region of the spectra together with the TPD curve from the saturation coverage (16 DU of glycine) of glycine on Al2O3/NiAl(110) at 300 K (room temperature). evaporation, the glycine was carefully outgassed for several hours at 370 K with the gate valve closed. During deposition, the evaporator was moved toward the sample surface and the glycine powder was kept at 400 K. Exposures are quoted in arbitrary doser units (DU), where 1 DU corresponds to ∼1 min glycine exposure. Since the evaporation rate is not linear for large evaporation times, it has been calibrated against the TPD peak areas. The quoted DU values are the corrected linear values.

3. Results and Discussion 3.1. TPD. In the TPD experiments, the m/z ) 30 signal was chosen to represent the intact molecules. In the mass spectroscopic fragmentation pattern of glycine (Gly), the parent ion (m/z ) 75) has an intensity of only ∼2% of the m/z ) 30 signal.1 The latter can be assigned to the positive molecular ion produced by removing the COOH group from the Gly molecule, that is, NH2CH2+. Figure 1 shows a series of TPD spectra obtained as a function of Gly exposure on Al2O3/NiAl(110) at 110 K. For the lowest exposure (2 DU), a broad feature is observed (see the insert of Figure 1), centered at ∼350 K (labeled β), and after 4 DU of Gly exposure, a second, sharper peak at 300 K (labeled R) is detected. The R peak shows typical zeroorder kinetics behavior, that is, it grows without saturation and has a common leading edge with increasing exposures. This is a strong indication for multilayer formation. The β TPD peak grows with exposure but eventually saturates at ∼6 DU exposure without a detectable shift. A close inspection of the TPD spectra (see the insert) reveals that the R peak starts growing before saturation of the higher temperature desorption state. We suggest that the β TPD peak is related to desorption of intact Gly molecules from the first monolayer. The presence of both point and line defects has been reported previously on the Al2O3 film.29 Since the β peak is broad, we presume that this peak is (29) Ba¨umer, M.; Freund, H.-J. Prog. Surf. Sci. 1999, 61, 127.

Langmuir, Vol. 20, No. 24, 2004 10553

due to the Gly desorption from both defect sites and regular oxide sites. The presence of the R TPD peak at low coverages indicates that the growth of the first monolayer is accompanied by the formation of Gly clusters, presumably nucleating at defect sites. An Arrhenius plot of the leading edge of the R peak at high coverages gives a desorption energy of 0.8 eV molecule-1, which is in good agreement with the previously reported estimate of the sublimation energy of Gly.1,24 In the insert of Figure 1, the TPD spectrum of m/z ) 30 from the saturation coverage of Gly (16 DU of Gly, see below) on Al2O3/NiAl(110) at room temperature (bottom curve) is compared with the spectra from the coverages at 110 K. Since 300 K is at the onset of the multilayer desorption peak, it is not possible to grow multilayers of Gly on Al2O3/NiAl(110) at room temperature. The TPD spectrum from the saturation coverage at room temperature shows a broad peak centered at ∼350 K. The similarity of this peak with the β TPD peak is clearly seen. Based on the TPD results, we propose that Gly forms only a monolayer on the oxide film upon exposure at room temperature. 3.2. XPS. Figure 2 shows a series of C 1s and N 1s XPS spectra displaying the buildup of Gly multilayers on Al2O3/ NiAl(110) at 110 K. The XPS sequences start with spectra obtained after a dose of 6 DU of Gly, that is, at the completion of the first monolayer of Gly as judged from the TPD measurements (see the previous paragraph). The spectra for the submonolayer coverage (