710
J. Phys. Chem. C 2011, 115, 710–713
Stereoselective Attachment via N Dative Bonding: S-Proline on Ge(100) Young-Sang Youn,† Ki-jeong Kim,‡ Bongsoo Kim,‡ Do Hwan Kim,*,§ Hangil Lee,*,| and Sehun Kim*,† Molecular-LeVel Interface Research Center, Department of Chemistry, KAIST, Daejeon 305-701, Republic of Korea, Beamline Research DiVision, Pohang Accelerator Laboratory (PAL), Pohang, Kyungbuk 790-784, Republic of Korea, DiVision of Science Education, Daegu UniVersity, Gyeongbuk 712-714, Republic of Korea, and Department of Chemistry, Sookmyung Women’s UniVersity, Seoul 140-742, Republic of Korea ReceiVed: October 7, 2010; ReVised Manuscript ReceiVed: NoVember 29, 2010
The adsorption configurations of S-proline on Ge(100) were studied using scanning tunneling microscopy (STM), density functional theory (DFT) calculations, and high-resolution photoemission spectroscopy (HRPES). We identified three adsorption structures of S-proline on Ge(100) through analysis of the STM images, DFT calculations, and HRPES results: (i) an “intrarow O-H dissociated and N dative bonded structure”, (ii) an “O-H dissociation structure”, and (iii) an “N dative bonded structure”. Moreover, because adsorption through the N atom of S-proline creates a new chiral center due to symmetry reduction produced by N dative bonding with the surface, the adsorption configurations of S-proline on Ge(100) have either (R,S) or (S,S) chirality. Through DFT calculations, we clearly demonstrated that the adsorption configurations have (R,S) chirality with a preference for reaction at the Re face. This work presents a novel method for generating stereoselective attachment using S-proline molecules adsorbed on a Ge(100) surface. Introduction Heterogeneous catalysts that yield stereoselective products are intensively pursued for their utility in the development of pharmaceuticals.1,2 Chiral studies on surfaces are particularly important for investigations of stereospecific catalysts because the stereospecificity of such catalysts arises from the adsorption of chiral molecules.3,4 While metal surfaces have been extensively investigated for catalysis, few studies have focused on semiconductor surfaces, which are gaining importance with the development of group-IV semiconductor-based industries.5-13 Therefore, studies of chiral configurations on group-IV semiconductor surfaces are central to the evolution of semiconductorbased stereochemical devices. Adsorption of prochiral molecules, which are achiral in the gas phase, on Si(100) and Ge(100) group-IV semiconductor surfaces induces the formation of chiral structures through [2 + 2] and [4 + 2] cycloaddition reactions with surface dimers.11-13 However, because these reactions always yield a racemic mixture, the introduction of stereoselectivity on Si(100) and Ge(100) is required for the development of new surface chemistry. A recent study of the adsorption of glycine (the simplest amino acid) on Ge(100) reported that the adsorption proceeds via a multibonding mechanism, forming an “intrarow O-H dissociated and N dative bonded structure”.14 Among the 20 essential amino acids, proline, which is used as an organocatalyst in organic synthesis,15 is unique in that it contains only a hydrogen atom bonded to the nitrogen atom. If proline reacts with Ge(100) surface via a multibonding adsorption mechanism * Corresponding author. D.H.K.: tel., +82-53-850-6986; fax, +82-53850-6989; e-mail,
[email protected]. H.L.: tel., +82-2-710-9409; fax, +82-2-2077-7321; e-mail,
[email protected]. S.K.: tel., +82-42350-2831; fax, +82-42-350-2810; e-mail,
[email protected]. † KAIST. ‡ PAL. § Daegu University. | Sookmyung Women’s University.
(O-H dissociation and N dative bonding), in agreement with previous findings for glycine,14 the nitrogen atom should become a chiral center due to symmetry reduction produced by N dative bonding with the surface. The chirality would be dependent on the direction of attack on the proline N atom. However, the attack direction of the lone electron pair on the nitrogen atom easily interconverts via “nitrogen inversion” at room temperature,16 such that both configurations (R and S) are possible adsorption products. Thus, the chirality is determined by the relative stabilities of the two stereoisomers rather than the attack direction of the lone pair of electrons. Experimental and Theoretical Details The Ge(100) surface (n-type, Sb doped, R ≈ 0.10 Ω) was cleaned by submitting the substrates to several sputtering cycles using 1 keV Ar+ ions (20 min, 700 K), followed by annealing at 900 K for 10 min. The cleanliness of the Ge(100) surface was checked with STM. S-Proline (L-proline; C4H7NHCOOH, 98.5% purity) was purchased from Aldrich and purified through several sublimation and pumping cycles to remove all dissolved gases prior to dosing. To obtain an appropriate vapor pressure for dosing, the dosing line was heated during S-proline deposition. STM observations were performed in an ultrahigh-vacuum (UHV) chamber equipped with an OMICRON VT-STM instrument at a base pressure below 1.0 × 10-10 Torr. All STM images were recorded with electrochemically etched tungsten tip at bias voltages of Vs ) -2.0 V with a tunneling current of It ) 0.1 nA. To investigate the adsorption configurations of proline on the Ge(100) surface, we carried out ab initio calculations within gradient-corrected density-functional theory (DFT-GGA) using the Vienna ab initio simulation package (VASP).17 Plane waves with energy up to 270.2 eV were included to expand the wave functions, and the atoms were represented by ultrasoft pseudo-
10.1021/jp109633m 2011 American Chemical Society Published on Web 12/20/2010
S-Proline on Ge(100)
J. Phys. Chem. C, Vol. 115, No. 3, 2011 711
potentials, as provided by VASP.18 Our slab model of the proline-adsorbed Ge(100) surface consisted of adsorbed proline molecules, six Ge atomic layers, and an H passivating layer. The Ge atoms of the bottom layer were each passivated with two H atoms. We used a p(4 × 4) supercell with a c(4 × 2) surface symmetry. The topmost four layers of the slab and the adsorbed molecules were allowed to relax with respect to the calculated Hellmann-Feynman forces, and the two remaining Ge layers were kept frozen during the structure optimization. The surface structure was considered to be in equilibrium when the Hellmann-Feynman force was less than 0.02 eV/Å. A Gaussian broadening with a width of 0.02 eV was used to accelerate the convergence in the k-point sum. Along with the use of self-consistent Kohn-Sham eigenvalues and wave functions, the constant-current STM images were simulated with the Tersoff-Hamann scheme.19,20 The tunneling current I(r, (V) is proportional to the energy-integrated local density of states:
I(r, (V) ∝
∑ ∫E
EF(V
nk
F
|Ψnk(r)| 2δ(E - Enk) dE
where +V and -V are the sample bias voltages for the emptystate and filled-state measurements, respectively. The HRPES results were obtained using the soft X-ray beamline (8A2) at the Pohang Accelerator Laboratory. The C 1s, N 1s, and O 1s core-level spectra of the system consisting of S-proline adsorbed onto the Ge(100) surface were obtained with a high performance electron analyzer (SES-100, Gamma Data, Sweden) using photon energies of 320, 460, and 590 eV to enhance the surface sensitivity. Three binding energies of the core-level spectra were determined relative to the clean Au valence band for the same photon energy. The base pressure of the chamber was maintained below 1.2 × 10-10 Torr. All the spectra were recorded in the normal-emission mode. The photoemission spectra were carefully analyzed using a standard nonlinear least-squares fitting procedure with Voigt functions.21
Figure 1. (a) Filled-state STM image (22.0 × 16.0 nm2, Vs ) -2.0 V, It ) 0.1 nA) of the Ge(100) surface at 300 K after exposure to 0.04 ML of S-proline. (b) The average prevalence of the adsorption configurations obtained from a lot of STM images. The error bars indicate the 95% confidence interval.
Results and Discussion Proline adsorbed on a Ge(100) were imaged by scanning tunneling microscopy (STM). Figure 1a presents a STM image, in which the three distinct adsorption features [marked as A (including A1 and A2), B, and C] are exhibited. To obtain statistical data for three adsorption features, from all STM images, we have separated several areas, which are not overlapped by one another, and then we have calculated the percentage of individual features tallied from total protrusions in each area. On the basis of the data obtained in several regions, the average and the standard deviation are calculated. In particular, the average prevalence of the three features was 46% (feature A), 36% (feature B), and 18% (feature C) as shown in Figure 1b. Figure 2 shows a magnified view of each of the three distinct adsorption features observed in the large-scale STM image (Figure 1a). In Figure 2a, feature A is characterized by a bright protrusion between two dimer rows with a dark adjacent dimer. This feature is similar to the characteristic adsorption structure of glycine on a Ge(100),14 suggesting that the adsorption structure of feature A most likely corresponds to an “intrarow O-H dissociated and N dative bonded structure”. Close examination of feature A reveals that it separates A1 and A2 subset configurations. STM has the outstanding ability to determine the chirality of adsorbed molecules on surfaces at the molecular level, but not the different surface analytical
Figure 2. The STM images, theoretically simulated STM images, and schematic models of (a) feature A (A1 and A2), (b) feature B, and (c) feature C. First panels: filled-state STM images (2.5 × 2.5 nm2, Vs ) -2.0 V, It ) 0.1 nA) shown in Figure 1a. Middle panels: the corresponding simulated filled-state images. Last two panels: the schematic models, which depict a bright protrusion position (marked as a red solid circle) and a dark site (marked by a black dashed circle) in the STM images. The blue and red asterisks indicate the intrinsic chiral center of an S-proline and a newly formed chiral center upon adsorption, respectively.
712
J. Phys. Chem. C, Vol. 115, No. 3, 2011
Youn et al.
Figure 4. The optimized adsorption structures of (a) (R,S)-“N dative bonded structure” and (b) (S,S)-“N dative bonded structure”. The white, gray, blue, red, and teal balls represent the hydrogen, carbon, nitrogen, oxygen, and germanium atoms, respectively.
TABLE 1: Adsorption Energies of the Optimized Adsorption Configurations within the p(4 × 4) Unit Cell Figure 3. The optimized adsorption structures of (a) L-shaped-(R,S)“intrarow O-H dissociated and N dative bonded structure”, (b) Γ-shaped-(R,S)-“intrarow O-H dissociated and N dative bonded structure”, (c) L-shaped-(S,S)-“intrarow O-H dissociated and N dative bonded structure“, and (d) Γ-shaped-(S,S)-“intrarow O-H dissociated and N dative bonded structure”. The white, gray, blue, red, and teal balls represent the hydrogen, carbon, nitrogen, oxygen, and germanium atoms, respectively. A red solid circle and a black dashed circle depict a bright protrusion position and a dark site in the STM images, respectively.
techniques. However, despite its excellent ability, we just establish two types of features (A1 and A2) without the assignment of definite chiral forms because it is difficult to distinguish the detailed chirality of the adsorbed proline. For instance, two L-shaped-(R,S) forms (Figure 3a,c) are not easily discriminated by the analysis of STM images. To complement the limitation of STM, we have, therefore, performed theoretical calculations using the Vienna ab initio simulation package (VASP).17,18 When proline attaches to a surface via N dative bonding, the new chiral center can assume either an (R) or an (S). In other words, reaction of the N atom at either the Re or the Si face leads to (R,S) or (S,S) configurations, respectively, for feature A. As shown in the calculation results (Figure 3), in which Figure 3a,c and Figure 3b,d correspond to A1 and A2, respectively, the (R,S) form of “intrarow O-H dissociated and N dative bonded structure” is more stable than the (S,S) form for both A1 and A2, suggesting that the energetic difference between the two diastereomers resulted in the favorite for the (R,S) configuration. Additionally, the reason why the adsorption structure prefers A1 (39%, L-shaped) to A2 (7%, Γ-shaped) among 46% feature A surveyed in the STM images is also explained by the energetic discrepancy between the two (R,S) configurations; although the difference in energies is slight, the relationship between the percentages of two isomers at equilibrium and the energy difference shows that the proportion of the isomer, which is higher in energy by just 1 kcal/mol than the other, approximately is 85%.22 As a result, we suggest that A1 and A2 correspond to L-shaped-(R,S)- and Γ-shaped-(R,S)“intrarow O-H dissociated and N dative bonded structures”, respectively. Before addressing feature B, we will discuss feature C. Feature C appears as a bright protrusion without a dark dimer (Figure 2c), corresponding to a dative bonded structure in that the absence of a dark dimer implies that no dissociation reaction occurred. Thus, the “N dative bonded structure”, the “carbonyl oxygen dative bonded structure”, and the “hydroxyl oxygen dative bonded structure” are all possible candidates for feature C. In previous researches of pyrrolidine and acetic acid on a Ge(100), it is reported that the “N dative bonded structure” was well produced with an adsorption energy of 22.9 kcal/mol, while
optimized adsorption structure
Eads (kcal/mol)
O-H dissociation structure (R,S)-N dative bonded structure (S,S)-N dative bonded structure carbonyl oxygen dative bonded structure hydroxyl oxygen dative bonded structure
-31.0 -25.2 -22.1 -12.6 -6.6
the others did not easily form.23,24 Furthermore, through the density functional theory (DFT) calculations for these three structures, we found that the “N dative bonded structure” is calculated to be the most stable among them (Table 1). On the basis of the previous studies and the calculations, we predict that the adsorption geometry of feature C is an “N dative bonded structure”. Moreover, because this structure also forms via N dative bonding, both (R,S) and (S,S) forms can be produced upon adsorption. The calculations for two diastereomeric forms (Figure 4 and Table 1) show that the (R,S) form is more stable than the (S,S) form. Consequently, we propose that the adsorption structure corresponding to feature C is an (R,S)-“N dative bonded structure”. Feature B appears as a bright protrusion above a dark dimer (Figure 2b), similar to feature A, except for the relative position of the bright protrusion. The adsorption site attributable to O-H or N-H dissociation reactions appears in STM images as a depressed feature.25-27 In the case of pyrrolidine on Ge(100), the N-H dissociation did not easily occur because the energy of its transition state was only 1.0 kcal/mol below the vacuum level.24 The dark site of feature B, therefore, is related to an O-H dissociation reaction. Two adsorption structures, including the O-H dissociation reaction and excluding feature A, are possible: the “O-H dissociation structure” and the “interrow O-H dissociated and N dative bonded structure”. Because it is difficult to determine the absolute adsorption structure of feature B through analysis of the STM images alone, the sample was characterized by high-resolution photoemission spectroscopy (HRPES). Although the C 1s core-level spectrum displays three distinct peaks involved in alkyl carbons, R-carbon, and carbonyl carbon and the O 1s core-level spectrum shows the peak, which is divided two peaks related to Ge-O (or H-O) and -CdO in the carboxyl group,14 these spectra are ineffectual data because they are complex results mixed from three adsorption structures. On the other hand, the N 1s core-level spectrum provides critical information. As shown in Figure 5b, we found two individual peaks. As compared to previous studies,14,28 the peaks at 398.0 and 400.7 eV are assigned to neutral N atom and positively charged N atom, respectively. Because the adsorption structures of features A and C include N dative bonding, the observation of double peaks in the N 1s core-level spectrum, therefore, implies that the adsorption configuration of feature B corresponds to an “O-H dissociation structure”, unrelated to N dative bonding. Moreover, the ratios of the peak integrals in HRPES correspond to the relative populations of the features. The ratio
S-Proline on Ge(100)
J. Phys. Chem. C, Vol. 115, No. 3, 2011 713 suggest a novel method for generating stereoselective attachment. Furthermore, because proline is utilized as an organocatalyst in organic synthesis, we anticipate that characterization of the adsorption configurations of proline will form a basis for developing semiconductor-based stereoselective devices.
Figure 5. HRPES spectra of (a) C 1s, (b) N 1s, and (c) O 1s for 0.05 ML S-proline on Ge(100) at 300 K. The dots indicate experimental values, and the solid lines were obtained by peak fitting.
Acknowledgment. This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2010-0001950, 2010-0023313, 2009-0067346). This work was also supported by the National Research Foundation of Korea Grant, funded by the Korean Government (KRF-2008-314-C00169). Calculations were performed using the supercomputing resources of KISTI (grant no. KSC-2009-S02-0005). References and Notes
Figure 6. Sequential STM images (6.8 × 6.0 nm2, Vs ) -2.0 V, It ) 0.1 nA) at the same region of the Ge(100) surface. White arrows display that an adsorption configuration is undergoing the structural transformation (feature C f feature A1 f feature C). A yellow spot is used as an indicator for notification of the identical area.
of the 398.0 and 400.7 eV peak areas, obtained from peak fitting, is 35%:65%, which is well matched with the relative feature prevalence obtained from the STM images, 36% (feature B containing neutral N atom):64% (features A and C, including N dative bonding). On the basis of the overall results, we also carried out the STM simulations from the suggested structures discussed above.19,20 As shown in Figure 2, every STM image is similar to its corresponding simulated one. Through this comparison, we once more demonstrate that the assignment of features A, B, and C is quite reasonable. In sequential STM images, we have occasionally found the structural transformation between feature A1 and feature C (Figure 6). On the other hand, the structural change related to feature B was not observed. It means the following possible inferences; (1) feature A is not transformed from feature B, but feature C, and (2) when feature C is converted to feature A, its configuration is determined by the relative stabilities of the configurations shown in Figure 3. Herein, because we focus on the stereoselective attachment via N dative bonding, we did not execute the more detailed theoretical calculations. Thus, further theoretical studies will be necessary to comprehensively understand this issue. Conclusions In this study, we have deduced the adsorption structures of proline on a Ge(100) using STM, DFT calculations, and HRPES. Although several adsorption configurations were observed, we
(1) Rouhi, A. M. Chem. Eng. News 2004, 82, 47–62. (2) Li, C. Catal. ReV. 2004, 46, 419–492. (3) Webb, G.; Wells, P. B. Catal. Today 1992, 12, 319–337. (4) Baddeley, C. J. Top. Catal. 2003, 25, 17–28. (5) Lorenzo, M. O.; Baddeley, C. J.; Muryn, C.; Raval, R. Nature 2000, 404, 376–379. (6) Ku¨hnle, A.; Linderoth, T. R.; Hammer, B.; Besenbacher, F. Nature 2002, 415, 891–893. (7) Chen, Q.; Richardson, N. V. Nat. Mater. 2003, 2, 324–328. (8) Bu¨rgi, T.; Baiker, A. Acc. Chem. Res. 2004, 37, 909–917. (9) Linares, M.; Minoia, A.; Brocorens, P.; Beljonne, D.; Lazzaroni, R. Chem. Soc. ReV. 2009, 38, 806–816. (10) Zaera, F. Acc. Chem. Res. 2009, 42, 1152–1160. (11) Lopinski, G. P.; Moffatt, D. J.; Wayner, D. D. M.; Wolkow, R. A. Nature 1998, 392, 909–911. (12) Hwang, Y. J.; Kim, A.; Hwang, E.; Kim, S. J. Am. Chem. Soc. 2005, 127, 5016–5017. (13) Zhang, Q. J.; Liu, Z. F. J. Phys. Chem. C 2009, 113, 5263–5273. (14) Youn, Y.-S.; Jung, S. J.; Lee, H.; Kim, S. Langmuir 2009, 25, 7438– 7442. (15) Dalko, P. I.; Moisan, L. Angew. Chem., Int. Ed. 2004, 43, 5138– 5175. (16) McMurry, J. Organic Chemistry, 5th ed.; Brooks/Cole: Pacific Grove, CA, 2000; pp 980-981. (17) Kresse, G.; Hafner, J. J. Phys.: Condens. Matter 1994, 6, 8245– 8257. (18) Kresse, G.; Furthmu¨ller, J. Comput. Mater. Sci. 1996, 6, 15–50. (19) Tersoff, J.; Hamann, D. R. Phys. ReV. Lett. 1983, 50, 1998–2001. (20) Tersoff, J.; Hamann, D. R. Phys. ReV. B 1985, 31, 805–813. (21) Schreier, F. J. Quant. Spectrosc. Radiat. Transfer 1992, 48, 743– 762. (22) McMurry, J. Organic Chemistry, 5th ed.; Brooks/Cole: Pacific Grove, CA, 2000; pp 133-136. (23) Filler, M. A.; Deventer, J. A. V.; Keung, A. J.; Bent, S. F. J. Am. Chem. Soc. 2006, 128, 770–779. (24) Wang, G. T.; Mui, C.; Tannaci, J. F.; Filler, M. A.; Musgrave, C. B.; Bent, S. F. J. Phys. Chem. B 2003, 107, 4982–4996. (25) Jung, S. J.; Lee, J. Y.; Hong, S.; Kim, S. J. Phys. Chem. B 2005, 109, 24445–24449. (26) Chung, O. N.; Kim, H.; Chung, S.; Koo, J.-Y. Phys. ReV. B 2006, 73, 033303-1–03303-4. (27) Bae, S.-S.; Kim, D. H.; Kim, A.; Jung, S. J.; Hong, S.; Kim, S. J. Phys. Chem. C 2007, 111, 15013–15019. (28) Chen, X. H.; Ranke, W. Surf. Sci. 1992, 262, 294–306.
JP109633M
