ARTICLE pubs.acs.org/JPCC
Adsorption Configuration of Serine on Ge(100): Competition between the Hydroxymethyl and Carboxyl Groups of Serine During the Adsorption Reaction Sena Yang,† Yaewon Kim,† Sunmin Park, Heesun Lim, and Hangil Lee* Department of Chemistry, Sookmyung Women's University, Seoul 140-742, Republic of Korea ABSTRACT: We investigated the adsorption structures of serine on a Ge(100) surface by core-level photoemission spectroscopy (CLPES) in conjunction with density functional theory (DFT) calculations. The adsorption energies calculated using DFT methods suggested that four of six adsorption structures were plausible. These structures were the OH dissociated-N dative bonded structure, the OH dissociation bonded structure, the OmH dissociated-N dative bonded structure, and the OmH dissociation bonded structure (where Om indicates the hydroxymethyl oxygen). These structures are equally likely, according to the adsorption energies alone. The core-level C 1s, N 1s, and O 1s CLPES spectra confirmed that the carboxyl oxygen competed more strongly with the hydroxymethyl oxygen during the adsorption reaction, thereby favoring formation of the OH dissociated-N dative bonded and OH dissociation bonded structures at 0.30 and 0.60 ML, respectively. The experimental results were corroborated theoretically by calculating the reaction pathways leading to the two adsorption geometries. The reaction pathways indicated that the OH dissociated-N dative bonded structure is the major product of serine adsorption on Ge(100) due to comparably stable adsorption energy.
I. INTRODUCTION Over the past several decades, the chemical functionalization of group IV semiconductor materials has been actively pursued for applications in a variety of industrial technologies.1,2 Many groups have examined the deposition of organic molecules on semiconductor surfaces, including Si and Ge.3,4 Organic materials may be easily tailored to tune their properties via substitution and addition reactions, so that organic functionalization may be readily exploited in applications such as biosensors, biochips, and electronic devices.5,6 Recently, several studies have examined the adsorption of amino acids on surfaces.3,710 Our group has studied the adsorption structures and behavior of amino acids specifically on Ge(100) to lay the groundwork for fabricating new bio-electronic hybrid devices.1115 In this article, the adsorption geometry of serine on a Ge(100) surface was modeled using density functional theory (DFT) calculations, and the electronic structure was characterized experimentally using core-level photoemission spectroscopy (CLPES). Previous adsorption studies of alanine, in which the side chain is a methyl group, demonstrated that the adsorption structure was the OH dissociated-N dative bonded structure at low coverage and the OH dissociation bonded structure at higher coverage.11 Interestingly, in addition to carboxyl and amino groups, serine molecules also contain an OH group attached to the methyl group, which can be possible to donate electron density to the Ge down-atom since the hydroxyl group of methanol was shown by Kim et al. to participate in the adsorption process on a Ge(100) surface via OH dissociation.16 r 2011 American Chemical Society
The goal of this study was to investigate the competition between hydroxyl groups from hydroxymethyl and carboxyl groups during the adsorption of serine on Ge(100). The relative stability of the adsorption products was also examined. This work predicts the behavior of amino acids that are structurally similar to serine, such as tyrosine. To our knowledge, this is the first experimental and theoretical characterization of the adsorption structures of serine on Ge(100) surfaces using DFT calculation and CLPES techniques.
II. EXPERIMENTAL SECTION The Ge(100) surface (p-type, R = 0.100.39 Ω) was cleaned by several cycles of sputtering with 1 keV Arþ ions for 20 min at 700 K, followed by annealing at 900 K for 10 min. The cleanness of the Ge(100)-2 1 surface was checked using low-energy electron diffraction (LEED). Serine (HO2CCH(NH2)CH2OH, 99% purity) were purchased from Aldrich and further purified through several sublimation and pumping cycles to remove all dissolved gases prior to exposure to the Ge(100) surface. We defined 1.0 monolayer (ML) serine coverage as the maximum density of serine that could be deposited on the Ge(100) surface without yielding the N 1s CLPES peak characteristic of multilayer formation. DFT calculations were used to predict the energetic of the reaction pathways for serine adsorption onto Ge(100). All DFT Received: January 26, 2011 Revised: March 30, 2011 Published: April 19, 2011 9131
dx.doi.org/10.1021/jp200866f | J. Phys. Chem. C 2011, 115, 9131–9135
The Journal of Physical Chemistry C
ARTICLE
Figure 1. Structure of the serine molecule and six possible adsorption structures of serine on Ge(100), with their calculated adsorption energies. (O = carboxyl oxygen, Om = hydroxymethyl oxygen).
calculations of the adsorption energies were performed using the JAGUAR 9.1 software package, which applied a hybrid density functional method and included Becke’s three-parameter nonlocal-exchange functional with the correlation functional of LeeYangParr (B3LYP).17 In these calculations, we considered two-dimer (Ge35H32) cluster models. The geometries corresponding to the important local minima on the potential energy surface were determined at the B3LYP/LACVP** level of theory. The LACVP** basis set is a mixed basis set that uses the LACVP basis set to describe Ge atoms and the 6-31G basis set for the remaining atoms. Moreover, the LACVP basis set describes atoms beyond Ar in the periodic table using the Los Alamos effective core potentials developed by Hay and Wadt.18 For each cluster, optimization was performed by fixing the bottom two layers of the Ge atoms in ideal Ge crystal positions while allowing the top layer of the Ge atoms (including the dimer atoms) and the atoms of the chemisorbed adsorbate to relax. Moreover, geometries of important local minima and transition states on each energy diagram were calculated. Local minima and transition states were verified as being optimized structures using the same set of basis sets.19 CLPES experiments were performed at the 7B1 beamline of the Pohang Accelerator Laboratory. The C 1s, N 1s, and O 1s core-level spectra were obtained using a PHOIBOS 150 electron energy analyzer equipped with a 2D charge-coupled device (2D CCD) detector (Specs GmbH) using photon energies of 340, 460, and 588 eV to enhance the surface sensitivity. Four binding energies of the core-level spectra were calibrated with respect to the binding energy of the clean Au 4f core-level spectrum (84.0 eV) at the same photon energy. The base pressure of the chamber was maintained below 9.5 1011 Torr. All spectra were recorded in the normal emission mode. The core-level spectra were carefully analyzed using a standard nonlinear leastsquares fitting procedure with Voigt functions.20
III. RESULTS AND DISCUSSION We first proposed six potential adsorption structures of serine on a Ge(100) and calculated the adsorption energies using DFT
methods to identify the most stable adsorption configuration, as shown in Figure 1. Parts a and d of Figure show that the nitrogen atom of the amine group or the oxygen atom of the carboxyl group may bond to the Ge(100) surface through a dative bonding configuration. The OH of the carboxyl group may also potentially be bound to the Ge(100) surface through OH dissociation (part f of Figure 1) or the OH dissociated-N dative bonded structure (part c of Figure 1). In addition to these four possible structures, we considered two more structures involving the OH of the hydroxymethyl group in the adsorption process, the OmH dissociation bonded structure (part b of Figure 1) and the OmH dissociated-N dative bonded structure (part e of Figure 1). Interestingly, the adsorption energies of the six potential adsorption structures suggested that four structures were likely to adsorb on the Ge(100) surface. Figure 2 shows the four most favorable adsorption structures viewed from the front and top, along with their adsorption energies. The four structures were divided according to their configuration into two distinct types: type I contains the structures in which the carboxyl oxygen is adsorbed onto the Ge(100) surface (parts a and b of Figure 2) and type II contains the structures in which the hydroxymethyl oxygen is adsorbed onto the Ge(100) surface (parts c and d of Figure2). An earlier study indicated that at high alanine coverage, the OH dissociated-N dative bonded structure reconfigured into the OH dissociation bonded structure”, in which the amine group was altered.11 Evaluation of the adsorption energies suggested that the OH dissociated-N dative bonded structure may be the dominant product because its adsorption energy was the lowest (Eads = 60.523 kcal/mol). Moreover, the other configuration of type I, the OH dissociated bonded structure was expected to develop (Eads = 47.025 kcal/mol). Type II was also expected because the adsorption energy of the OmH dissociated-N dative bonded structure was low (Eads = 56.204 kcal/mol). A negligibly small difference in adsorption energies between the OH dissociated-N dative bonded structure and the OmH dissociated-N dative bonded structures supports such a possibility as well. Hence, we performed CLPES measurement to identify 9132
dx.doi.org/10.1021/jp200866f |J. Phys. Chem. C 2011, 115, 9131–9135
The Journal of Physical Chemistry C
ARTICLE
Figure 2. Diagrams of the lowest energy adsorption structures of serine on Ge(100).
Figure 3. C 1s, N 1s, and O 1s core-level spectra for (a, b, c) 0.30 ML or (d, e, f) 0.60 ML serine on Ge(100). The dots indicate experimental values and the solid lines represent the results of peak fitting.
which the adsorption structure is more dominant experimentally because the dominant product indicated the relative reactivity of the carboxyl or hydroxymethyl OH groups as electron donors to the Ge(100) surface during adsorption. Figure 3 shows a series of C 1s, N 1s, and O 1s core-level spectra as a function of serine concentration. We first determined the C 1s core-level spectra for two different serine coverage levels (0.30 and 0.60 ML) to corroborate our DFT calculation results. Part a of Figure 3 shows the C 1s core-level spectrum obtained after exposure of 0.30 ML serine at 300 K. This spectrum contains three distinct bonding features corresponding to the three carbons of serine. Considering the relationship between electronegativity and binding energy, we assigned the bonding features C1, C2, and C3 to CNH2 (285.74 eV), CH2OH (286.95 eV), and deprotonated carbonyl carbon (287.47 eV), respectively. These results demonstrated that serine adsorbed to the Ge(100) surface without bond breaking at this coverage level. Increasing the coverage of serine to 0.60 ML (part d of Figure 3, greater than a half monolayer) produced features that differed from those found at 0.30 ML. A new peak (marked C4)
Table 1. Binding Energies Determined from the Three CoreLevel Spectra of Serine Adsorbed on Ge(100) at 300 K as a Function of Coverage C 1s (eV) coverage (ML) 0.30
C1/C2/C3/C4 285.74/286.95/
0.60
285.74/286.95/
N 1s (eV)
O 1s (eV)
N1/N2 O1/O2/O3 400.77/(N/A) 530.84/532.03/533.35
287.47/(N/A) 400.77/398.28 530.84/532.03/533.35
287.47/286.31
emerged at 286.31 eV, indicating a change in bonding due to formation of a new adsorption structure. This observation suggested that the electronic state (or electric charge) of the amine group changed from positive to neutral. If the electric charge of the amine group became neutral upon formation of the adsorption structure, the electric charge of the adjacent R-carbon would also become less positive because of the reduced withdrawing ability of the neutral nitrogen atom. Therefore, the peak would be expected to shift to higher binding energies. 9133
dx.doi.org/10.1021/jp200866f |J. Phys. Chem. C 2011, 115, 9131–9135
The Journal of Physical Chemistry C
ARTICLE
Figure 4. Theoretically calculated potential energy surfaces for (a) OH dissociated-N dative bonded structure and (b) OH dissociation bonded structure of serine on Ge(100).
The changes in the N and O 1s core-level spectra between the two coverage levels suggested an explanation for this interesting phenomenon. Using the procedure described above, we first obtained a N 1s core-level spectrum for 0.30 ML serine (part b of Figure 3) adsorbed onto Ge(100) at 300 K. This spectrum clearly contained a single N 1s peak (400.77 eV), which corresponded to the charged N atom in the serine molecule. Part e of Figure 3 shows the N 1s core-level spectrum obtained for a 0.60 ML coverage, which corresponded to another adsorption structure. This spectrum showed that a new peak emerged at 398.28 eV (marked N2). This peak provided critical proof that a new adsorption structure was present at this coverage level. As mentioned above, in the interpretation of the C 1s spectra, the new N 1s peak (N2) arose from changes in the charge on the amine group, from positive to neutral. The peak assignments of previous studies involving N dative bonding adsorption structures on Ge(100) surfaces clearly indicate that at a low initial coverage, the nitrogen atom in the amine group adsorbs onto Ge(100) surfaces in a dative bonding configuration.11 Thus, the nitrogen atom appears to break its initial dative bonding with a Ge up-atom as the coverage of serine increases. Finally, we examined the changes in the O 1s peaks as a function of exposure to serine molecules. Part c of Figure 3 shows the O 1s core-level spectrum obtained after deposition of 0.30 ML serine. With consideration for the experimental resolution and the chemically inequivalent oxygen atoms in serine, we divided the O 1s peaks into three components, assigned O1, O2, and O3, to OGe (530.84 eV), CdO (532.03 eV), and CH2OH (533.35 eV). As mentioned in the Introduction, the focus of this experiment was to determine whether the oxygen of the hydroxymethyl group bonded to the Ge(100) surface. If this occurred, the hydroxyl peak (OH) typically located at 533 eV would not have been observed. Because the hydroxyl peak actually was observed, as shown in part c of Figure 3, we could exclude the possibility that the hydroxymethyl group bonded to the Ge(100) surface during adsorption. Therefore, the oxygen of the hydroxymethyl group did not participate in adsorption on the Ge(100) surface. Interestingly, increasing the coverage from 0.30 to 0.60 ML did not produce changes in the O 1s peaks. Even at higher coverage levels, all adsorption structures underwent OH dissociation.
Analysis of the C 1s, N 1s, and O 1s core-level spectra demonstrated that the N1 peak arose from the OH dissociated-N dative bonded structure, and the N2 peak arose from an OH dissociation bonded structure. Furthermore, it was deduced that the OH dissociated-N dative bonded structure was dominant at low initial coverage levels, whereas the OH dissociation bonded structure emerged at higher coverage levels. Finally, we confirmed that the OH of the hydroxymethyl group did not participate in adsorption on the Ge(100) surface. Table 1 summarizes the binding energies obtained from the three corelevel spectra at various serine coverage levels on the Ge(100) surface at 300 K. The experimental results were corroborated theoretically using DFT calculations. Model reaction pathways for formation of the OH dissociated-N dative bonded structure and OH dissociation bonded structure were obtained. Part a of Figure 4 shows the theoretically calculated adsorption pathway of the OH dissociated-N dative bonded structure derived from the OH dissociation bonded structure. The transition state of the OH dissociated-N dative bonded structure occurred at 28.382 kcal/mol, indicating that this reaction was experimentally possible. The reaction pathway for the OH dissociation bonded structure, shown in part b of Figure 4, was calculated from the initial O dative bonded structure. The transition state energy of this structure was 19.767 kcal/mol, which is also attainable. The activation barriers for the OH dissociated-N dative bonded and the OH dissociation bonded structures were 11.925 and 2.816 kcal/mol respectively, and these barriers may be easily overcome at room temperature. Thus, two adsorption structures will be present on the Ge(100) surfaces at room temperature. The OH dissociated-N dative bonded structure is likely to be dominant initially because its adsorption energy is 60.523 kcal/mol, which is 13.50 kcal/mol lower than that of the OH dissociation bonded structure (Eads = 47.025 kcal/mol). Because of relatively stable adsorption energy, it is the OH dissociated-N dative bonded structure that is the major product of serine adsorption on Ge(100). These theoretical results support the analysis of the CLPES data, demonstrating that the OH dissociated-N dative bonded structure is preferred at low serine coverage (0.30 ML) and the OH dissociation bonded structure occurs at high serine coverage (0.60 ML). 9134
dx.doi.org/10.1021/jp200866f |J. Phys. Chem. C 2011, 115, 9131–9135
The Journal of Physical Chemistry C
ARTICLE
IV. CONCLUSIONS We investigated the adsorption structures of serine adsorbed on the Ge(100) surface by performing DFT calculations and CLPES experiments. DFT calculations of the adsorption energies allowed prediction of the four most energetically favorable adsorption structures among the six possible structures. These were the OH dissociated-N dative bonded structure, the OH dissociation bonded structure, the OmH dissociated-N dative structure, and the OmH dissociation bonded structure. These structures are equally likely on the basis of the adsorption energy alone. The core-level C 1s, N 1s, and O 1s CLPES spectra confirmed that the carboxyl oxygen is more competitive against the hydroxymethyl oxygen in the adsorption reaction. Therefore, the OH dissociated-N dative bonded structure and the OH dissociation bonded structure are preferred at 0.30 and 0.60 ML, respectively. The experimental results were modeled by calculating the reaction pathways for preparing the energetically favorable adsorption geometries. The reaction pathways indicated that the OH dissociated-N dative bonded structure is the major product of serine adsorption on Ge(100) due to comparably stable adsorption energy. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected], Tel.: þ82-2-710-9409, Fax: þ82-2-2077-7321. Author Contributions †
Equally first author.
’ ACKNOWLEDGMENT This work was supported by the Sookmyung Women’s University Research Grants 2010. ’ REFERENCES (1) Loscutoff, P. W.; Bent, S. F. Annu. Rev. Phys. Chem. 2006, 57, 467. (2) Yates, J. T., Jr. Science 1998, 279, 335. (3) Filler, M. A.; Bent, S. F. Prog. Surf. Sci. 2003, 73, 1. (4) Wolkow, R. A. Annu. Rev. Phys. Chem. 1999, 50, 413. (5) Filler, M. A.; Van Deventer, J. A.; Keung, A. J.; Ben, S. F. J. Am. Chem. Soc. 2006, 128, 770. (6) Kasemo, B. Surf. Sci. 2002, 500, 656. (7) Smith, R. K.; Lewis, P. A.; Weiss, P. S. Prog. Surf. Sci. 2004, 75, 1. (8) Whaley, S. R.; English, D. S.; Hu, E. L.; Barbara, P. F.; Belcher, A. M. Nature 2000, 405, 665. (9) Goede, K.; Busch, P.; Grundmann, M. Nano Lett. 2004, 4, 2115. (10) Iwai, H.; Emori, A.; Egawa, C. Surf. Sci. 2006, 600, 1670. (11) Youn, Y.; Kim, K.; Kim, B.; Kim, D.; Lee, H.; Kim, S. J. Phys. Chem. C 2011, 115 (3), 710. (12) Lee, H.; Youn, Y.; Kim, S. Langmuir 2009, 25 (21), 12574. (13) Youn, Y.; Jung, S. J.; Lee, H.; Kim, S. Langmuir 2009, 25 (13), 7438. (14) Youn, Y.; Lee, H.; Kim., S. ChemPhysChem 2010, 11, 354. (15) Park, S.; Yang, S.; Shin, N.; Lee, E.; Lee, H. J. Phys. Chem. C 2010, 114 (34), 14528. (16) Kim, D. H.; Bae, S.; Hong, S.; Kim, S. Surf. Sci. 2010, 604, 129. (17) Kachian, J. S.; Tannaci, J.; Wright, R. J.; Tilley, T. D.; Bent, S. F. Langmuir 2011, 27, 179. (18) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (19) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299. (20) Schreier, F. J. Quant. Spectrosc. Radiat. Transfer. 1992, 48, 743. 9135
dx.doi.org/10.1021/jp200866f |J. Phys. Chem. C 2011, 115, 9131–9135