7360
Langmuir 2004, 20, 7360-7364
Adsorption and Coordination of Tartaric Acid Enantiomers on Cu(111) in Aqueous Solution Hui-Juan Yan, Dong Wang, Mei-Juan Han, Li-Jun Wan,* and Chun-Li Bai* Institute of Chemistry, the Chinese Academy of Sciences, Beijing 100080, People’s Republic of China Received July 16, 2003. In Final Form: June 27, 2004 Chiral modifiers have gained much attention because they can induce high enantioselectivity on reactive metal surface in heterogeneous enantioselective catalysis. The high enantioselectivity is attributed to that the chirality of modifiers is bestowed onto the metal surface upon adsorption. Much study on the adsorption of modifiers on metal surface has been performed in an ultrahigh vacuum. In this paper, the adsorption of tartaric acid on Cu(111) has been studied by electrochemical scanning tunneling microscopy (STM) in aqueous solution. It is found that (R,R)-tartaric acid and (S,S)-tartaric acid can form a well-ordered adlayer on the Cu(111) surface with a (4 × 4) symmetry. A dimeric structure is proposed in the temporary model from STM observation.
Introduction With the rapid development of nanotechnology and single molecular science, the investigation of chemical reaction and catalysis at the molecular level becomes possible and gives us unprecedented insight into chemical science.1-3 Recently, much attention has been paid on the heterogeneous enantioselective catalysis because of its potential application in pharmaceutical and chemical industries.4-7 Two typical heterogeneous enantioselective reactions are R-ketoester hydrogenation over supported Pt in the presence of different cinchona alkaloids and asymmetry hydrogenation of β-ketoesters in solution over supported Ni modified by tartaric acid (TA). It is generally accepted that the chirality of the modifier is bestowed onto achiral metal surfaces upon adsorption, leading to highly enantioselective catalysts. Therefore, to improve enantioselectivity of catalysts and develop a new chiral heterogeneous system, it is necessary to understand the adsorption conformation of chiral modifiers on metal surfaces and elucidate the enantioselective mechanism at the molecular level. Various analytical techniques have been applied to investigate the structures of modifiers on metal surfaces, and several mechanisms have been proposed to explain the enantioselective catalysis process.7-15 * Corresponding authors. E-mail:
[email protected]. Fax: +86-10-62558934. (1) Silva, S. L.; Patel, A. A.; Pham, T. M.; Leibsle, F. M. Surf. Sci. 1999, 441, 351-365. (2) Cudia, C. C.; Hla, S. W.; Comelli, G.; Sljivancanin, Z.; Hammer, B.; Baraldi, A.; Prince, K. C. Rosei, R. Phys. Rev. Lett. 2001, 87, 19610411961044. (3) De Feyter, S.; Gesquiere, A.; Abdel-Mottaleb, M. M.; Grim, P. C. M.; De Schryver, F. C. Acc. Chem. Res. 2000, 33, 520-531. (4) Studer, M.; Blaser, H. U.; Exner, C. Adv. Synth. Catal. 2003, 345, 45-65. (5) Baiker, A.; Blaser, H. U. In Handbook of Heterogeneous Catalysis; Ertl, G., Knozinger, H., Weitkamp, J., Eds.; Verlag Chemie: Weinheim, 1997. (6) Blaser, H. U. Tetrahedron: Asymmetry 1991, 2, 843-866. (7) Keane, M. A. Langmuir 1997, 13, 41-50. (8) Xu, Q.-M.; Wang, D.; Wan, L.-J.; Wang, C.; Bai, C.-L. J. Am. Chem. Soc. 2002, 124, 14300-14301. (9) Bonello, J. M.; Williams, F. J.; Lambert, R. M. J. Am. Chem. Soc. 2003, 125, 2723-2729. (10) Lorenzo, M. O.; Baddeley, C. J.; Muryn, C.; Raval, R. Nature 2000, 404, 376-378. (11) Lorenzo, M. O.; Haq, S.; Bertrams, T.; Murray, P.; Raval, R.; Baddeley, C. J. J. Phys. Chem. B 1999, 103, 10661-10669. (12) Barbosa, L. A. M. M.; Sautet, P. J. Am. Chem. Soc. 2001, 123, 6639-6648.
Particularly, the unique power of scanning tunneling microscopy (STM) makes it possible to reveal the structural details of the modifier on solid surfaces. For examples, Xu et al. reported a direct evidence on the adsorption mode of cinchonidine on the Cu(111) substrate with STM in solution.8 Lambert et al. demonstrated that the 1:1 chiral induction model between modifiers and reactants is applicable for cinchona/Pt catalysts by ultrahigh vacuum (UHV) surface characterization techniques.9 On the (R,R)TA-based catalysis system, Lorenzo et al. studied the adsorption of (R,R)-TA on Cu(110) with low-energy electron diffraction, STM, and Fourier transform reflection-absorption infrared spectroscopy techniques.10,11 The chiral spaces are proposed as a “template” and are responsible for imparting enantioselectivity. Barbosa and Sautet conducted theoretical simulation and proposed that the adsorption of bitartrate on Cu(110) results in the relaxation of Cu atoms, which accounts for the formation of chiral channels.12 The similar study is also carried out on Ni(110) in an UHV by experimental and theoretical methods. It is proposed that the enantioselectivity of catalysts is attributed to the chiral reconstruction of the substrate induced by the adsorption of (R,R)-TA molecules.13 Recently, Jones and Baddeley found that (R,R)TA produced three distinct ordered structures on the Ni(111) surface at different temperatures in an UHV.14 When methylacetoacetate (MAA) was introduced to the (R,R)-TA-modified Ni(111), MAA locally rearranged the TA modifiers to produce a two-dimensional cocrystal.15 The previous results obtained in an UHV really provide reasonable interpretations for the enantioselectivity of heterogeneous catalysts. However, elucidating the adsorption structure of TA on the metal surface in an aqueous solution is also important. In the present study, we report the adsorption of TA on well-defined Cu(111) by electrochemical STM in an aqueous solution. In heterogeneous catalysis, the optimal Ni particle size was estimated to be 10-20 nm.5 The {111} surface is the most thermodynamically stable face of a face-centered cubic (fcc) metal and is expected to be the dominant crystal face on Ni particles with nanoscale size. Cu and Ni are all fcc metals, and their {111} surfaces (13) Humblot, V.; Haq, S.; Muryn, C.; Hofer, W. A. Raval, R. J. Am. Chem. Soc. 2002, 124, 503-510. (14) Jones, T. E.; Baddeley, C. J. Surf. Sci. 2002, 513, 453-467. (15) Jones, T. E.; Baddeley, C. J. Surf. Sci. 2002, 519, 237-249.
10.1021/la035283y CCC: $27.50 © 2004 American Chemical Society Published on Web 08/06/2004
Letters
Langmuir, Vol. 20, No. 18, 2004 7361
Figure 1. Chemical structures of the (R,R)-TA and (S,S)-TA molecules.
have similar surface structure. It is expected that the results obtained from this model system can provide a useful reference for the real enantioselective catalysis mechanism. Figure 1 shows the molecular structures of TA enantiomers generated by hyperchem software. The chirality of TA molecules arises from the different orientations of two hydroxyl groups on C2 and C3. Each TA molecule includes two chiral centers, as indicated by arrows. Two carboxyl groups are at the ends of the molecule, respectively. In the present study, we found that both (R,R)-TA and (S,S)-TA form ordered adlayer structures in a (4 × 4) symmetry on the Cu(111) surface. A tentative model of the (R,R)-TA adlayer is proposed in this paper. Experimental Section A bulk Cu(111) single-crystal disk with a diameter of 10 mm (from MaTeck) was used as a working electrode for both electrochemical measurement and in situ STM observation. The atomically flat Cu(111) single-crystal surface was prepared as described in the literature.16 Briefly, after having been mechanically polished with a diamond paste down to 0.05 µm in particle diameter, the Cu(111) disk was further treated by electropolishing in a phosphoric acid solution (50 mL of 85% H3PO4 and 50 mL of water). The Cu crystal was then rinsed repeatedly with Ultrapure water (Millipore Milli-Q). Cyclic voltammetric experiments were performed by using an EG&G PAR (Princeton Applied Research) Basic Electrochemical System controlled by an IBM microcomputer with EG&G PARC Powersuite software. A homemade electrochemical cell contains a reversible hydrogen electrode (RHE) and a Pt counter electrode in a 0.1 M HClO4 solution. The working electrode was the Cu(111) single crystal. The solutions were deaerated with high-purity nitrogen. The potential scan rate was 50 mV/s. All the potentials were reported with respect to the RHE. The in situ STM apparatus used was a Nanoscope E (Digital Instrument, CA). The tunneling tips were prepared by electrochemically etching a tungsten wire (0.25 mm in diameter) in 0.6 M KOH. The alternating current voltage of 12-15 V was applied until the etching process stopped. Then the W tips were coated with a clear nail polish to minimize the faradic current. The STM images shown here were obtained in the constant-current mode to evaluate corrugation heights of adsorbed molecules. The STM operating parameters are reported in the figure captions. All electrolyte solutions were prepared by diluting ultrapure HClO4 and reagent grade (R,R)-TA and (S,S)-TA (Acros Organics, USA) with Millipore water.
Results and Discussion Cyclic Voltammetry. Steady-state cyclic voltammograms (CVs) of Cu(111) were obtained by using the socalled hanging meniscus methods in the absence and (16) Wan, L. J.; Itaya, K. Langmuir 1997, 13, 7173-7179.
Figure 2. CVs of Cu(111) in (a) 0.1 M HClO4, (b) 0.1 M HClO4 + 1 mM (R,R)-TA, and (c) 0.1 M HClO4 + 1 mM (S,S)-TA. The potential scan rate was 50 mV/s.
presence of TA in 0.1 M HClO4. The first scan of each CV was made in the negative direction from the open circuit potential. Figure 2a shows a CV of bare Cu(111) in 0.1 M HClO4. It can be seen that the double layer region extends from -0.35 V to +0.15 V. The CV of Cu(111) in the presence of (R,R)-TA is shown in Figure 2b. A featureless CV similar to that in Figure 2a is recorded. Only the electric charge involved in the double layer potential region becomes smaller because of adsorption of organic molecules. A similar CV is also observed in the electrolyte solution containing 1 mM (S,S)-TA shown in Figure 2c. The electrochemical results indicate that there is no phase transition of TA adlayers in the double layer potential region. According to the electrochemical results reported by Attard,17 high-symmetry single-crystal surface has no enantioselectivity toward chiral molecules. Thus, similar CVs are obtained on the adlayers of enantiomers. STM Measurement. (R,R)-TA Molecules. The atomic image of a Cu(111)-(1 × 1) structure was routinely discerned on the atomically flat Cu(111) surface in the absence of the molecules for the ease of determining the registry of the molecular adlayers to the underlying Cu(111) lattice. After the atomic resolution image of the Cu(111)-(1 × 1) structure was observed, a small amount of (R,R)-TA solution was directly injected into the electrochemical STM cell. The average concentration of (R,R)TA in 0.1 M HClO4 solution was 1 mM. After several minutes from the addition of (R,R)-TA, a well-ordered adlayer of molecules can be observed to extend over a wide atomically flat terrace, different from the underlying Cu(111) lattice. A typical large scale STM image acquired at -0.20 V is shown in Figure 3a. TA molecules appear in ordered bright spots with some defects as indicated by arrows in Figure 3a. The molecular rows cross each other at an angle of either 60 or 120° within an experimental error of (2°. From a comparison with the orientation of 〈110〉 determined by the Cu(111) atomic image as indicated by a set of arrows, it is concluded that all molecular rows are parallel to the underlying Cu(111) lattice. The distance between neighboring bright spots along the close-packed direction of Cu(111) is measured to be about 0.98 ( 0.02 nm, about four times the lattice parameter of Cu(111). On the basis of the intermolecular distance and orientation of molecular rows, a (4 × 4) structure for the adlayer can be concluded. A unit cell is superimposed in the higher resolution image of Figure 3b. Figure 3b shows the details of the (R,R)-TA adlayer. It is clearly seen in the image (17) Attard, G. A. J. Phys. Chem. B 2001, 105, 3158-3167.
7362
Langmuir, Vol. 20, No. 18, 2004
Figure 3. (a) STM top view of the (R,R)-TA adlayer. The tunneling current was 3 nA. The scanning rate was 10 Hz. (b) High-resolution STM image of the (R,R)-TA adlayer. The tunneling current was 6 nA. The scanning rate was 13 Hz.
that each bright spot in Figure 3a appears in a pair of lobed spots as outlined by two ellipses in Figure 3b. The length of a pair of lobed bright spots is about 0.6 nm, which is larger than the maximum length of TA molecules (∼0.5 nm).11 It could be assumed that each bright spot may arise from a molecule. Two individual lobed spots in a pair are separated by about 0.4 nm. The distance admits the formation of intermolecular H bonds between the acid groups and OH groups in the pair molecules. The structural details will be discussed in next part. (S,S)-TA. With a similar procedure, the (S,S)-TA adlayer was prepared on Cu(111). A well-defined adlayer similar to (R,R)-TA is observed in large-scale STM images of Figure 4a. The molecular rows cross each other at an angle of either 60 or 120° within an experimental error of (2°. Compared with the orientation of the Cu(111) substrate, the molecular rows are determined to be parallel to the 〈110〉 direction of the Cu(111) substrate. The same (4 × 4) array can be determined by analyzing the intermolecular distance and symmetry of the adlayer with the Cu(111) substrate. A unit cell is outlined in a typical high-resolution
Letters
Figure 4. (a) Large-scale STM image of the (S,S)-TA adlayer. The tunneling current was 3 nA. The scanning rate was 13 Hz. (b) High-resolution STM image of the (S,S)-TA adlayer. The tunneling current was 6 nA. The scanning rate was 15 Hz.
STM image of Figure 4b. It is surprising that the features of (S,S)-TA molecules in Figure 4b are very similar to those of (R,R)-TA in Figure 3b. (S,S)-TA molecules also show two-lobed bright spots, indicated by two ellipses in Figure 4b. It is temporarily proposed that each bright spot may arise from a molecule. Two individual lobed spots form a molecular pair separated by about 0.4 nm. The results described above demonstrate that both (R,R)-TA and (S,S)-TA molecules form ordered structures on Cu(111) in aqueous solution. To further understand the adsorption of the TA molecule on the Cu(111) substrate such as adsorption site and symmetry, a composite STM imaging was performed by stepping potentials. Various composite STM images were acquired. Figure 5a is a typical composite STM image in the solution containing (R,R)-TA molecules. The lower part was imaged at -0.20 V, and the upper part was imaged at -0.39 V, when the imaging direction was from bottom to top. The upper part of the STM image clearly
Letters
Langmuir, Vol. 20, No. 18, 2004 7363
Figure 5. (a) Composite STM image acquired at different potentials in 0.1 M HClO4 + 1 mM (R,R)-TA. The upper part shows the atomic image of the substrate, and the lower part shows the TA adlayer. The scan rate was 15 Hz. The tunneling current was 3 nA. The proposed models for the (R,R)-TA in (b) two molecules in a dimer locate on the same top sites and (c) one molecule locates on the top sites and another on the bridge sites.
shows the (1 × 1) atomic structure of the Cu(111) substrate. The lower part shows the two-lobed bright spots of the (R,R)-TA molecule. A geometric network is imposed on the image based on the Cu(111) lattice. The (4 × 4) symmetry for the adlayer is further demonstrated by the extended lines in the network along the close-packed atomic rows of the underlying Cu(111) substrate. From the composite STM image, we try to propose a structural model for the adsorption of (R,R)-TA molecules on Cu(111). It is clear that if the two bright spots in a pair were treated as a single (R,R)-TA molecule, there is more space in the (4 × 4) structure. Therefore, we propose that the two bright spots in a pair are two individual (R,R)-TA molecules, which form a dimeric structure. A careful observation found that the two bright spots in the STM image of Figure 5a show different contrasts. This implies that two molecules of each dimer may adsorb on different adsorption sites of Cu(111). The intermolecular distances of two bright spots can be illustrated in Figure 5a with about 0.40 nm in the AC direction, about 0.42 nm in the AB direction, and and 0.18 nm in the BC direction. On the basis of the STM data and the chemical structure of (R,R)-TA molecules, two possible models for the (R,R)TA adlayer are proposed in Figure 5b,c. According to the
results of IR spectroscopic studies, carboxylic acids, such as formic,18,19 benzoic,20,21 citric acid,22,23 and others, adsorb on the noble metal surface with the carboxylate form. Previous results have shown that the oxygen atoms of the carboxylate groups favorably adsorb on the top sites in a short brige.12,24-27 Furthermore, a bend geometry for the adsorption of (R,R)-TA molecules on Cu(110) surface has (18) Bowker, M.; Haq, S.; Holroyd, R.; Parlett, P. M.; Poulston, S.; Richardson, N. J. Chem. Soc., Faraday Trans. 1996, 92, 4683-4686. (19) Hahn, F.; Beden, B.; Lamy, C. J. Electroanal. Chem. 1986, 204, 315-327. (20) Li, H. Q.; Roscoe, S. G.; Lipkowski, J. J. Electroanal. Chem. 1999, 478, 67-75. (21) Frederick, B. G.; Chen, Q.; Leibsle, F. M.; Lee, M. B.; Kitching, K. J.; Richardson, N. V. Surf. Sci. 1997, 394, 1-25. (22) Floate, S.; Hosseini, M.; Arshadi, M. R.; Ritson, D.; Young, K. L.; Nichols, R. J. J. Electroanal. Chem. 2003, 542, 67-74. (23) Nichols, R. J.; Burgess, I.; Young, K. L.; Zamlynny, V.; Lipkowski, J. J. Electroanal. Chem. 2004, 563, 33-39. (24) Woodruff, D.; McConville, C. F.; Kilcoyne, A. L. D.; Linder, Th.; Somers, J.; Somers, M.; Surman, M.; Paolucci, G.; Bradshaw, A. M. Surf. Sci. 1988, 201, 228-244. (25) Somers, J.; Robinson, A. W.; Linder, Th.; Bradshaw, A. M. Phys. Rev. B 1989, 40, 2053-2059. (26) Gomes, J. R. B.; Gomes, J. A. N. F. Surf. Sci. 1999, 432, 279290. (27) Bao, S.; Liu, G.; Woodruff, D. P. Surf. Sci. 1988, 203, 89-100.
7364
Langmuir, Vol. 20, No. 18, 2004
been proposed.10-12 Hence, on the basis of the abovementioned results, one possibility for the adsorption of (R,R)-TA is assumed that each TA molecule coordinated with Cu surface deprotonates fully and locates on the top sites of Cu(111) substrate with a bend geometry as shown in Figure 5b. In this configuration, two molecules in a dimer should show an identical appearance in STM images. However, the right bright spot is longer than the left one in Figure 5a, which indicates different adsorption models between the two molecules. If the oxygen atoms in the carboxyl groups in the dimer adsorb on the different sites of the Cu atoms, the STM image will show a difference. Therefore, we propose a structural model in Figure 5c. In this model, the oxygen atoms of carboxyl end groups in the left (R,R)-TA molecule are located on the bridge sites, and those in the right (R,R)-TA molecule are located on the top sites of the Cu(111) lattice. The TA is in a bicarboxylate form. The intermolecular distances in the model are almost consistent with the measured distances in STM images. On the other hand, two (R,R)TA molecules in a dimer can form H bonds between the OH groups and carboxyl groups, as shown in Figure 5c in dashed lines. Each (R,R)-TA molecule can leave a hydroxyl group protruding into solution to form H bonds with other reactant species. A similar adsorption model for (S,S)-TA molecules is also proposed (not shown here). Note that there are three forms in existence for TA according to the study in an UHV: a bicarboxylate, monocarboxylate, or biacid.11 Although the TA molecules in the model of Figure 5c are in the form of bicarboxylate, the real form of the TA species on the Cu(111) surface cannot be determined by only STM. More techniques such as IR and theoretical calculation are needed to clarify this issue. It is seen that the ordering of TA adlayers can provide an ordered array of vacancies on the metal surface available for β-ketoester molecules to adsorb on, similar to that of TA on the Ni(111) surface in an UHV.14 Because of the chirality of TA molecules surrounding vacancies, ketoester molecules only can dock onto vacancies in one direction to form H bonds with OH groups of neighboring TA molecules. The H bonds between the reactants and TA will define the adsorption geometry of ketoester molecules. If all ketoester species are adsorbed on vacan-
Letters
cies with the same face attached to the surface, the dominant product produced by the hydrogenation of reactant molecules will be the one of two possible enantiomers. The enantioselectivity is created. However, it is possible that some relaxations of the surrounding TA molecules may occur according to the size of the reactant molecules, when MAA and other reactant species are introduced into the system.12,15 Therefore, (R,R)-TA can offer an ordered chiral template on the Cu(111) surface in the liquid phase, which is consistent with the theory of the template in heterogeneous catalysis. In fact, Jones and Baddeley proposed a model that MAA and TA can coadsorb on the surface.15 It is known that the enantioselectivity of a real catalyst depends on the metal dispersion, modifier coverage, temperature, solvents, and so forth.5,6 These factors will partially change the adsorption conformation of the TA molecules and, thus, affect the enantioselectivity. For example, the adlayer structure of TA on the Ni(111), (110) surface is significantly different.13-15 Our work is only a preliminary step toward the understanding the enantioselectivity of TA-based heterogeneous catalysts. To get insight into the catalyst mechanism of (R,R)-TA-based catalytic systems, the detailed understanding of the interaction of the reactants with the TA-modified Ni surface will be important in further investigations. Conclusion In summary, (R,R)-TA and (S,S)-TA can adsorb on Cu(111) and form long-range ordered adlayers on the Cu(111) surface in an aqueous solution with a (4 × 4) unit cell. According to the STM data, it could be assumed that two TA molecules form a dimer through intermolecular H bonds in the adlayer on the Cu(111) substrate. Each TA molecule can leave one hydroxyl group to form H bonds with other reactant species. A possible structural model for the TA adlayer is proposed. Acknowledgment. Financial support from National Natural Science Foundation of China (Nos. 20025308 and 20177025), National Key Project on Basic Research (Grants G2000077501 & 2002 CCA03100), and Chinese Academy of Sciences is gratefully acknowledged. LA035283Y
