Submolecular Electronic Mapping of Single Cysteine Molecules by in

420015 Kazan, Republic of Tatarstan, Russian Federation. ReceiVed October 14, 2008. ReVised Manuscript ReceiVed NoVember 20, 2008. We have used ...
0 downloads 0 Views 2MB Size
2232

Langmuir 2009, 25, 2232-2240

Submolecular Electronic Mapping of Single Cysteine Molecules by in Situ Scanning Tunneling Imaging Jingdong Zhang,*,† Qijin Chi,† Renat R. Nazmutdinov,‡ Tamara T. Zinkicheva,‡ and Michael D. Bronshtein‡ Department of Chemistry, NanoDTU, Building 207, Technical UniVersity of Denmark, 2800 Kgs. Lyngby, Denmark, and Kazan State Technological UniVersity, 420015 Kazan, Republic of Tatarstan, Russian Federation ReceiVed October 14, 2008. ReVised Manuscript ReceiVed NoVember 20, 2008 We have used L-cysteine (Cys) as a model system to study the surface electronic structures of single molecules at the submolecular leVel in aqueous buffer solution by a combination of electrochemical scanning tunneling microscopy (in situ STM), electrochemistry including voltammetry and chronocoulometry, and density functional theory (DFT) computations. Cys molecules were assembled on single-crystal Au(110) surfaces to form a highly ordered monolayer with a periodic lattice structure of c(2 × 2) in which each unit contains two molecules; this conclusion is confirmed by the results of calculations based on a slab model for the metal surface. The ordered monolayer offers a platform for submolecular scale electronic mapping that is an issue of fundamental interest but remains a challenge in STM imaging science and surface chemistry. Single Cys molecules were mapped as three electronic subunits contributed mainly from three chemical moieties: thiol (-SH), carboxylic (-COOH), and amine (-NH2) groups. The contrasts of the three subunits depend on the environment (e.g., pH), which affects the electronic structure of adsorbed species. From the DFT computations focused on single molecules, rational analysis of the electronic structures is achieved to delineate the main factors that determine electronic contrasts in the STM images. These factors include the molecular orientation, the chemical nature of the elements or groups in the molecule, and the interaction of the elements with the substrate and tip. The computational images recast as constant-current-height profiles show that the most favorable molecular orientation is the adsorption of cysteine as a radical in zwitterionic form located on the bridge between the Au(110) atomic rows and with the amine and carboxyl group toward the solution bulk. The correlation between physical location and electronic contrast of the adsorbed molecules was also revealed by the computational data. The present study shows that cysteine packing in the adlayer on Au(110) from the liquid environment is in contrast to that from the ultrahigh-vacuum environment, suggesting solvent plays a role during molecular assembly.

1. Introduction Molecular electronics, as a crucial part of nanoscale science and technology, is a rapidly growing area.1 Among several issues of fundamental interest in this area, high-resolution mapping of the electronic structure and transport of organic thin films on solid surfaces has been increasingly in demand for optimal design of molecular electronic devices. Advances of both self-assembled monolayer (SAM) chemistry and surface-ultrasensitive techniques such as scanning tunneling microscopy (STM) currently offer effective tools to deal with this challenging task.2 STM has been widely used to measure molecular conductivity by either imaging or tunneling spectroscopy.3 In pioneering works by G. W. Flynn and associates,4-6 functional chemical groups of organic molecules with a long hydrocarbon chain were identified due to their different contrasts in STM images. To obtain submolecular * To whom correspondence should be addressed. E-mail: [email protected]. Phone: +45 45252352. Fax: +45 45883136. † Technical University of Denmark. ‡ Kazan State Technological University. (1) (a) Jortner, J., Rather, M. A., Eds. Molecular Electronics; Blackwell: Oxford, 1997. (b) Hush, N. S Ann. N.Y. Acad. Sci. 2003, 1006, 1–20. (c) Nitzan, A.; Ratner, M. A. Science 2003, 300, 1384–1389. (d) Tao, N. J. Nat. Nanotechnol. 2006, 1, 173–181. (2) (a) Love, C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chem. ReV. 2005, 105, 1103–1172. (b) Fendler, J. H. Chem. Mater. 2001, 13, 3196–3210. (c) Itaya, K. Prog. Surf. Sci. 1998, 58, 121–248. (d) Kolb, D. M. Angew. Chem., Int. Ed. 2001, 40, 1162–1181. (e) Vericat, C.; Vela, M. E.; Beniter, G. A.; Gago, J. A. M.; Torreles, X.; Salvarezza, R. C. J. Phys.: Condens. Matter 2006, 18, R867–900. (g) Otero, R.; Rosei, F.; Besenbacher, F. Annu. ReV. Phys. Chem. 2006, 57, 497–525. (3) (a) Xiao, X. Y.; Xu, B. Q.; Tao, N. J. Nano Lett. 2004, 4, 267–271. (b) Li, X. L.; He, J.; Hihath, J.; Xu, B. Q.; Lindsay, S. M.; Tao, N. J. J. Am. Chem. Soc. 2006, 128, 2135–2141. (c) Tao, N. J. Phys. ReV. Lett. 1996, 76, 4066–4069.

resolution information from STM, a key issue is that such molecules with long alkyl chains assemble into highly ordered monolayers in a lying-down mode by physical adsorption on graphite surfaces.7 This is in contrast to the upright adsorption mode in many SAMs from thiol-containing molecules on metal surfaces.8 Few reports therefore deal with electronic mapping of single thiol molecules at submolecular resolution though reports on monolayer lattices, dynamics, and reactivity have appeared abundantly due to their prominent properties as well as applications.8,9 The present work has used L-cysteine as a target molecule, mainly due to the presence of three structurally different chemical groups, -COOH, -NH2, and -SH, that make Cys ideally suited as a target for submolecular electronic mapping. L-Cysteine is the only amino acid among the 20 natural amino acids with a free thiol (-SH) group. Cys serves as a ligand (4) Cyr, D. M.; Venkataraman, B.; Flynn, G. W.; Black, A.; Whitesides, G. M. J. Phys. Chem. 1996, 100, 13747–13759. (5) Wintgens, D.; Yablon, D. G.; Flynn, G. W. J. Phys. Chem. B 2003, 107, 173–179. (6) Mu¨ller, T.; Werblowsky, T. L.; Florio, G. M.; Berne, B. J.; Flynn, G. W. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 5315–5322. (7) Mourran, A.; Ziener, U.; Mo¨ller, M.; Suarez, M.; Lehn, J. M. Langmuir 2006, 22, 7579–7586. (8) (a) Poirier, G. E. Chem. ReV. 1997, 97, 1117–1128. (b) Schreiber, F. Prog. Surf. Sci. 2000, 65, 151–256. (c) Love, C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chem. ReV. 2005, 105, 1103. (d) Vericat, C.; Vela, M. E.; Beniter, G. A.; Gago, J. A. M.; Torreles, X.; Salvarezza, R. C. J. Phys.: Condens. Matter 2006, 18, R867–R900. (9) (a) Chi, Q.; Zhang, J.; Ulstrup, J. J. Phys. Chem. B 2006, 110, 1102–1106. (b) Poirier, G. E.; Pylant, E. D. Science 1996, 272, 1145–1148. (c) Zhang, J.; Chi, Q.; Ulstrup, J. Langmuir 2006, 22, 6203–6213.

10.1021/la8034006 CCC: $40.75  2009 American Chemical Society Published on Web 01/22/2009

STM Contrasts of Single Cys Molecules

Langmuir, Vol. 25, No. 4, 2009 2233

Figure 1. Cyclic voltammograms (A) and capacitance curves (B) of bare Au(110) (dotted lines) and Cys-modified Au(110) (solid lines) in 20 mM NH4Ac (pH 4.6). Scan rate 20 mV s-1 for voltammograms; 100 Hz frequency, 5 mV amplitude and 20 mV step potential were carried out for capacitance measurements.

coordinated to a number of transition metals in metalloproteins10 such as iron-sulfur clusters, which are cores for iron-sulfur proteins found in all life forms.11,12 Cys forms adlayers with variable structures on noble metal surfaces13-17 as well as on Si.18 This has attracted much attention and thorough investigations by various advanced surface-sensitive techniques. Intriguingly, close packing (3 × 3)R30°, cluster-like network (33 × 6)R30°, and (4 × 7)R19° structures are all found as surface lattices in Cys monolayers on Au(111) surfaces in H2SO4,15a NH4Ac (pH 4.6),15b and HClO4,15c respectively. This suggests that solvation effects19 are crucial during Cys assembly on the surfaces. These observations are in contrast to Cys adsorption from the gas phase, since only one lattice, i.e., (3 × 3)R30°, was discovered in the monolyers on Au(111) in ultrahigh vacuum (UHV).17 Large numbers of SAMs have been studied on the (111) surface,8 but a low-index surface (110) has recently been employed as a new platform for studying chemical adsorption of chiral molecules20,16 as well as local and global chirality21 at the surfaces. The atomic structure on a (110) substrate is more open than that on a (111) surface and provides high adsorption efficiency. Several interesting chirally organized monolayers of organic molecules and amino acids have been discovered mainly (10) Saysell, D. M.; Sokolov, M. N.; Sykes, A. G. ACS Symp. Ser. 1996, 563, 216. (11) Beinert, H.; Holm, R. H.; Mu¨nck, E. Science 1997, 277, 653–659. (12) Sticht, H.; Ro¨sch, P. Prog. Biophys. Mol. Biol. 1998, 70, 95–136. (13) (a) Fawcett, W. R.; Fedurco, M.; Kovacova, Z.; Borkowska, Z. J. Electroanal. Chem. 1994, 368, 265. (b) Fawcett, W. R.; Fedurco, M.; Kovacova, Z.; Borkowska, Z. J. Electroanal. Chem. 1994, 368, 275. (c) Brolo, A. G.; Germain, P.; Hager, G. J. Phys. Chem. B 2002, 106, 5982–5987. (d) Yang, W. R.; Gooding, J. J.; Hibbert, D. B. J. Electroanal. Chem. 2001, 516, 10–16. (14) Dodero, G.; De Michieli, L.; Caveller, O.; Rolandi, R.; Oliveri, L.; Dacca, A.; Parodi, R. Colloids Surf., A 2000, 175, 121–128. (15) (a) Dakkouri, A. S.; Kolb, D. M.; Edelstein-Shima, R.; Mandler, D. Langmuir 1996, 12, 2849–2852. (b) Zhang, J.; Chi, Q.; Nielsen, J. U.; Friis, E. P.; T. Andersen, J. E.; Ulstrup, J. Langmuir 2000, 16, 7229. (c) Xu, Q.; Wan, L.; Wang, C.; Bai, C.; Wang, Z.; Nozawa, T. Langmuir 2001, 17, 6203–6206. (16) (a) Ku¨hnle, A.; Linderoth, T. R.; Hammer, B.; Besenbacher, F. Nature 2002, 415, 891–893. (b) Ku¨hnle, A.; Linderoth, T. R.; Besenbacher, F. J. Am. Chem. Soc. 2003, 125, 14680–14681. (c) Ku¨hnle, A.; Linderoth, T. R.; Besenbacher, F. J. Am. Chem. Soc. 2006, 128, 1076–1077. (17) Ku¨hnle, A.; Linderoth, T. R.; Schunack, M.; Besenbacher, F. Langmuir 2006, 22, 2156–2160. (18) Honda, M.; Matsui, F.; Daimon, H. Surf. ReV. Lett. 2006, 13, 197–200. (19) Mandouh, W.; Uji-i, H.; Ladislaw, J. S.; Dulcey, A. E.; Percec, V.; Schryver, F. C. D.; Feyter, S. D. J. Am. Chem. Soc. 2006, 128, 317–325. (20) (a) Barlow, S. M.; Louafi, S.; Roux, D. L.; Williams, J.; Muryn, C.; Haq, S.; Raval, R. Langmuir 2004, 20, 7171–7176. (b) Barlow, S. M.; Louafi, S.; Roux, D. L.; Williams, J.; Muryn, C.; Haq, S.; Raval, R. Surf. Sci. 2005, 590, 243–263. (21) (a) Humblot, V.; Lorenzo, M. O.; Baddeley, L. C.; Haq, S.; Raval, R. J. Am. Chem. Soc. 2004, 126, 6460–6469. (b) Schock, M.; Otero, R.; Stojkovic, S.; Hummelink, F.; Gourdon, A.; Lægsgaard, E.; Stensgaard, I.; Joachim, C.; Besenbacher, F. J. Phys. Chem. B 2006, 110, 12835–12838.

on (110) surfaces by UHV-STM studies. Kuhnle and associates studied Cys adsorption on Au(110) in UHV by STM and found various assembly patterns on the reconstructed Au(110) surface with 1 × 2 missing rows.16b,c On the basis of STM images with molecular resolution, they concluded that Cys adsorption depends on temperature. Dimers in pairs in a diagonal or beam-shaped organization in the rows and even clusters with eight molecules have been visualized.16 In the present work we investigate Cys monolayer assembly on Au(110) from aqueous buffer solution by electrochemistry and in situ STM. The adsorption pattern has been traced down to the submolecular level. This warrants a further study based on quantum chemistry to understand the physical meaning of each subfeature. There are few reports on quantum chemical calculations of the geometry, electronic structure, and energetics of Cys species adsorbed at single-crystal surfaces of gold and silver. The main pertinent results have been obtained in the framework of periodical (slab) calculations using density functional theory (DFT).22-25 Felice et al.22,23 have investigated the adsorption of a single Cys molecule and radical at the Au(111) surface. A comprehensive theoretical study on the adsorption behavior of different Cys species (including their zwitterionic form) at Au(111) has been broadly addressed on the basis of a cluster model for the metal surface.26 Combined with experimental data from UHV-STM, formation of a cysteinate dimer at low adsorption coverage on the reconstructed and unreconstructed Au(110) surface was proposed.24 In the present study we investigate the adsorption of a cysteine radical, SCH2CH(NH2)COOH (Cys Rad), at the uncharged Au(110) surface and disclose contributions from the atomic orbitals of each group to the STM contrast which agree well with the submolecular resolution of the experimental observations.

2. Experimental and Modeling Methods 2.1. Materials and Reagents. Cys solutions were freshly prepared from L-cysteine (>98%, Sigma) before each measurement. Fresh NH4Ac (20 mM, pH 4.6 and 6.5) solutions were prepared from ultrapure solution as described previously.15b Millipore water (Milli-Q Housing, 18.2 MΩ) was used throughout. (22) Felice, R.; Selloni, A. J. Chem. Phys. 2004, 120, 4906–4914. (23) Felice, R.; Selloni, A.; Molinari, E. J. Phys. Chem. B 2003, 107, 1151– 1156. (24) Ku¨hle, A.; Molina, L. M.; Linderoth, T. R.; Hammer, B.; Besenbacher, F. Phys. ReV. Lett. 2004, 93, 086101–1. (25) Santos, E.; Avalle, L.; Po¨tting, K.; Ve´lez, P.; Jones, H. Electrochim. Acta 2008, 53, 6807–6817. (26) Nazmutdinov, R. R.; Zhang, J.; Zinkicheva, T. T.; Manyurov, I. R.; Ulstrup, J. Langmuir 2006, 22, 7556–7567.

2234 Langmuir, Vol. 25, No. 4, 2009

Figure 2. Reductive desorption of Cys-modified Au(110) in 0.5 M NaOH (scan rate 20 mV s-1).

2.2. Preparation of Cysteine Monolayers. Homemade Au(110) electrodes were used for both electrochemistry and in situ STM measurements. The orientation of Au(110) was adjusted by reflection spots from both (111) and (100) facets on single-crystal beads which were prepared by melting the end of a gold wire (1.0 mm diameter, >99.99%).27 The quality of homemade Au(110) surfaces was checked in 0.1 M HClO4, 0.1 M H2SO4 by both electrochemistry and in situ STM, and both main and detailed (finger printer) features accord well with literature reports.28,29 Commercial Au(110) substrates from the Surface Preparation Laboratory (The Netherlands) were measured by the same methods as a comparison and accord with the quality of the homemade Au(110) surfaces. Cys monolayers on Au(110) were prepared in a similarly to SAMs on Au(111) as described in our previous paper.15b In brief, the Au(110) electrode was annealed in a hydrogen flame, quenched in dihydrogensaturated Millipore water, and transferred to the cell with NH4Ac solution or Cys-containing solution for adsorption. 2.3. Electrochemical and STM Measurements. Cyclic voltammetry, capacitance measurements, and chronocoulometry were performed using an Autolab system (Eco Chemie, The Netherlands) controlled by general-purpose electrochemical system software. An electrochemical setup including a cell and three-electrode configuration was employed as previously described.30 The reference electrode was a freshly prepared reversible hydrogen electrode (RHE), checked against a saturated calomel electrode (SCE) after each measurement. All potentials reported are vs SCE. Monolayers were formed by soaking the Au(110) electrodes in 0.1-5.0 mM Cys solution for 0.5-5 h. Electrolyte solutions were deoxygenated for several hours by Ar (5 N) purified by Chrompack (oxygen 0.1 s, potential step from a potential in the double-layer region to -0.822 V vs SCE; the charge reaches a plateau in a time window of 50 s).

Figure 4. In situ STM images of bare Au(110) in 20 mM NH4Ac (pH 4.6) (Ew ) -0.07 to -0.3 V vs SCE, Et ) -0.07 to -0.3 V vs SCE, It ) 0.15-0.5 nA, scan area (A) 100 × 100 nm2 and (B) 50 × 50 nm2).

Au(110) was represented as a single Cys radical adsorbed on a two-layer Au21 cluster. The quantum chemical calculations were performed at the DFT level with the hybrid exchange-correlation functional B3PW91 as implemented in the Gaussian 03 program suite.32 The valence orbitals of the Au atom were described on a basis set of double-ζ (DZ) quality, while the effect of inner electrons was included in the effective core potential (ECP).33 The standard basis set 4-31g was used to describe the electrons in the S, O, N, C, and H atoms.26 The basis set was augmented by a polarization d orbital for the sulfur atom. The ground spin state of the Au21-Cys Rad adsorption complex is a singlet. The open shell Au21 and Cys Rad systems were treated in terms of unrestricted formalism. The geometry of adsorbed species was fully optimized. The natural population analysis (NPA) was employed to calculate the atomic charges. The formation of a self-assembled layer of Cys Rad species at the Au(110) surface was explored with the help of periodical DFT (32) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision B.04; Gaussian, Inc.: Pittsburgh, PA, 2003. (33) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270–283.

STM Contrasts of Single Cys Molecules

Langmuir, Vol. 25, No. 4, 2009 2235

Figure 5. STM images of a Cys monolayer on Au(110) in 20 mM NH4Ac (pH 4.6) (Ew ) -0.3 to 0.0 V vs SCE, Et ) -0.5 to -0.3 V vs SCE, It ) 0.20-1.0 nA, scan area (A) 60 × 60 nm2, (B) 20 × 20 nm2, and (C) 3.1 × 3.1 nm2). (D) A schematic model of Cys molecules adsorbed on the Au(110) surface. Violet balls represent gold atoms; a lighter color means it is closer to the top surface. A yellow dotted triangle represents the location of one Cys.

calculations using the spin-unpolarized formalism. The metal surface was modeled by repeated supercells consisting of four (110) atomic layers, a Cys radical adsorbed on the surface, and a vacuum thickness of 57.6 Å. The generalized gradient approximation (GGA) in the version of Perdew-Burke-Ernzerhof (PBE) was employed as implemented in the SIESTA code.34,35 Numerical basis sets of DZ type (augmented by polarization orbitals) and Troullier-Martins norm-conserving soft ECPs36 were used as well. The Brillouin zone integration was accomplished using a 12 × 14 × 1 k-point Monkhorst-Pack grid with respect to the surface unit cell. We have taken an energy of 200 Ry as the cutoff in all calculations. All atoms of adsorbed species were allowed to relax without any symmetry restrictions, while the positions of the Au atoms were kept fixed. The adsorbate geometry was optimized until the residual force on each atom was below 0.04 eV/Å.

3. Results and Discussion 3.1. Voltammetric Evidence of Cysteine Adsorption. Figure 1A shows voltammograms of bare Au(110) (dotted line) and Au(110) in the presence of a Cys adlayer (solid line) and 20 mM NH4Ac (pH 4.6). In the absence of Cys, one strong anodic peak and one weak anodic peak appear at potentials of -0.05 and 0.15 V vs SCE, respectively, accompanied by a large and sharp cathodic peak at -0.15 V vs SCE. Two corresponding peaks appear in the same potential region on a capacitance curve (dotted line), Figure 1B. Though the origin of the peaks is not totally clear, we believe that it is most likely due to adsorption of acetate and a lift of reconstruction on Au(110) since similar peaks are found in the same supporting electrolyte on the Au(111) surface.15b,30 (34) Soler, J. M.; Artacho, E.; Gale, J. D.; Garcia, A.; Junquera, J.; Ordejon, P.; Sa´nchez-Portal, D. J. Phys.: Condens. Matter 2002, 14, 2745. (35) Soler, J. M.; Artacho, E.; Gale, J. D.; Garcia, A.; Junquera, J.; Ordejon, P.; Sa´nchez-Portal, D. htpp://uam.es/departamentos/fismateriac/siesta/. (36) Troullier, N.; Martins, J. L. Phys. ReV. B 1991, 43, 1993.

The presence of a Cys monolayer results in featureless voltammetry and capacitance in the potential range of -0.2 to +0.6 V, suggesting that acetate adsorption has been replaced by a strongly adsorbed Cys layer and the reconstruction lifted in the same potential window. This is also supported by in situ STM observation and discussed later. Adsorption of Cys could be described by the following equation:

2HCOO(CH)NH2-CH2SH + Au a Au 2HCOO(CH)NH2-CH2Sads-Au + H2

(1)

Assuming the adsorbed Cys is in the radical form and the amine and sulfur are close to the Au(110) surface, the estimated heat of the above reaction (on the basis of DFT cluster calculations) is -0.36 eV, meaning that reaction 1 could occur spontaneously. The coverage of the Cys monolayer could therefore be measured quantitatively by reductive desorption according to the following equation:

HCOO(CH)NH2-CH2Sads+Au + e a HCOO(CH)NH2-CH2S- + Au

(2)

Figure 2 illustrates a voltammogram of a Cys monolayer on Au(110) in 0.5 M NaOH solution. The reductive desorption should be carried out in a sufficient basic solution (pH13); otherwise a large charge could be aroused due to the influence of hydrogen evolution in a low-pH solution. The cathodic peak at -0.71 V corresponds to reductive desorption, e.g., eq 2. The coverage of Cys is therefore estimated as (6.74 ( 0.3) × 10-10 mol · cm-2 according to Faraday’s law. Three features should be addressed: (1) The peak potential at -0.71 V is consistent with -0.70 V for a Cys monolayer on the Au(111) surface,15b meaning the

2236 Langmuir, Vol. 25, No. 4, 2009

Figure 6. STM images of a Cys monolayer on Au(110) in 20 mM NH4Ac (pH 6.5) (Ew ) -0.25 to +0.1 V vs SCE, Et ) -0.45 to - 0.2 V vs SCE, It ) 0.15-0.30 nA, scan area (A) 58 × 58 nm2, (B) 20.5 × 20.5 nm2, and (C) 4.7 × 4.7 nm2).

S-Au strength is dominated by the adsorbed molecular rather than surface orientation. (2) The peak half-width is ca. 41 mV for the Cys monolayer on Au(110), which is significantly larger than 15 mV for the Cys on Au(111), implying the molecular packing in the monolayer is different. The narrower and sharper the peak, the more interaction among molecules is involved. A similar peak half-width is also observed for a monolayer of thiols of a similar size such as propanethiol on Au(111) with a peak half-width of 38 mV.9c (3) The coverage of Cys on Au(110) is (6.74 ( 0.3) × 10-10 mol · cm-2, notably larger than 3.83 × 10-10 mol · cm-2 for Cys on Au(111).15b This presumes that (110) face provides more adsorption sites for Cys, meaning that Cys molecules assemble more densely on Au(110) than on Au(111). A detailed surface structure study of the Cys monolayer by in situ STM will be reported below. Due to the zwitterionic state of Cys in contact with solution, it is interesting to know the charge of the Cys monolayer in the same potential window. Potential-scan chronocoulometry is an efficient technique to characterize the charge density of both chemical and physical adsorption.37-40 The method and quantitative data analysis were introduced and developed by Lipkowski et al.41-43 The potential was held at Ei in the double-layer region for a few minutes, which is positive compared to a desorption potential such as -0.5 V (vs SCE), to ascertain that the Cys monolayer has reached equilibrium adsorption. The potential was then stepped to -0.86 V (vs SCE), where the monolayer (37) Niece, B. K.; Gewirth, A. A. Langmuir 1996, 12, 4909–4913. (38) Shi, Z.; Lipkowski, J. J. Phys. Chem. 1995, 99, 4170–4175. (39) Wandlowski, Th.; Ho¨lzle, M. H. Langmuir 1996, 12, 6597–6603. (40) Wandlowski, Th.; Wang, J. X.; Ocko, B. M. J. Electroanal. Chem. 2001, 500, 418–434. (41) Richer, J.; Lipkowski, J. J. Electrochem. Soc. 1986, 133, 133. (42) Stolberg, L.; Richer, J.; Lipkowski, J. J. Electroanal. Chem. 1986, 207, 213. (43) Stolberg, L.; Lipkowski, J.; Irish, D. E. J. Electroanal. Chem. 1990, 296, 171.

Zhang et al.

was totally desorbed from the surface. During the potential stepping, the charge flow was recorded vs time and reached a plateau in a time window of 50 s. The charge value was calculated into charge density by the consideration of the electrode area and plotted vs each potential as shown in Figure 3. The charge density depends on both the potential and coverage of adsorbed molecules. Overall, the presence of Cys increases the charge density in the whole potential window from -0.6 to +0.6 V (vs SCE) with a smooth range from -0.2 to +0.3 V, at which a stable Cys monolayer exists. Compared with bare Au(110), the presence of the Cys monolayer increases the charge density from 1.0-3.0 to 6.5-8.5 µC · cm-2 in the potential range of -0.2 to +0.3 V (vs SCE), whereas the charge density decreases at a negative potential (-0.5 V vs SCE) mainly due to partial desorption. At a high potential, >+0.4 V vs SCE, the charge density increases drastically since Cys monolayer oxidation begins. The overview pattern is broadly consistent with the adsorption of organic molecules at gold surfaces39,44,45 and data for chemical adsorption of bromide on Ag(100).40 Quantitatively, the charge density increment due to Cys monolayer adsorption on Au(110) is less than the value of ca. 50 µC · cm-2 for a bromide monolayer on Ag(100),49 but closer to the value of 5 µC · cm-2 for weak chemical adsorption of 1,3-dimethyluracil on Au(111).39 This means that coverage of the adsorbed molecules is a dominating factor for charges on surfaces. A more detailed mapping of the Cys monolayers is explored by in situ STM and described below. 3.2. Surface Lattice Structures and STM Tunneling Contrasts. Figure 4A shows long-range ordered strips covering most of the area of a bare Au(110) surface in 20 mM NH4Ac (pH 4.6). The strip direction is along the atomic rows of Au(110), i.e., [11j0], with a periodic distance of 12.6 ( 0.3 Å, Figure 4B. This feature is therefore assigned to the 1 × 3 missing row reconstruction, which is similar to a reconstructed surface in H2SO4 solution.46 Such a surface structure is obviously different from UHV observation, in which a reconstruction surface with 1 × 2 missing rows is routinely obtained.16 The observation on Au(110) is in contrast to that on Au(111), for which herringbone structures, i.e., reconstruction of (3 × 23)R30°, are found in both liquid and UHV environments, meaning that solvent effects are less important for Au(111) than Au(110). In the presence of the Cys monolayer in 20 mM NH4Ac (pH 4.6), the strip feature disappears and the surface is covered by large terraces along the [11j0] direction, suggesting that the 1 × 3 missing row reconstruction is lifted by the adsorption of Cys, Figure 5A. Highly ordered monolayers are found to cover both the terrace and edge uniformly. Cys molecules shown as single spots in Figure 5B are aligned along the [110] direction. No cluster features are obtained. Higher resolution images disclose subfeatures of the adsorbed Cys monolayer in the form of triangles with three protrusions (white spots) in each corner and a dark area in the middle, Figure 5C. The periodic distance along the [11j0] direction is 5.7 ( 0.3 Å, which is twice the gold atomic distance of 2.88 Å. The periodic distance of every second row in the perpendicular direction (i.e., the [001] direction) is 8.3 ( 0.4 Å, corresponding to twice the smallest periodic distance of 4.3 Å. The unit cell of the Cys adlayer, as shown by a black box in Figure 5C, could therefore be described as a c(2 × 2) lattice, giving a coverage of 3.5 × 10-10 mol · cm-2. Comparing this value with the coverage of (44) Li, N.; Zamlynny, V.; Lipkowski, J.; Henglein, F.; Pettinger, B. J. Electroanal. Chem. 2002, 524-525, 43–53. (45) Richer, J.; Iannelli, A.; Lipkowski, J. J. Electroanal. Chem. 1992, 324, 339–358. (49) Schmickler, W. Surf. Sci. 1993, 295, 43–56. (46) Magnussen, O. M.; Wiechers, J.; Behm, R. J. Surf. Sci. 1993, 289, 139– 151.

STM Contrasts of Single Cys Molecules

Langmuir, Vol. 25, No. 4, 2009 2237

Figure 7. Optimized structures of a cysteine radical adsorbed at the model Au21 cluster in different positions: OS bridge 1 (A); NS bridge 1 (B); NS bridge 2 (C); S bridge 1 (D); OS on-top (E); NS on-top (F). A fragment of the cluster illustrating two bridge sites and the position of the sulfur atom is also shown (G).

(6.74 ( 0.3) × 10-10 mol · cm-2 measured by reductive desorption (Figure 2) indicates that each unit cell contains two Cys molecules; i.e., each Cys occupies 24.5 Å2. This strongly suggests that each triangle with three protrusions represents one Cys molecule. The area per Cys molecule on Au(110) is comparable to the value of 27 Å2 from an alanine monolayer on Cu(110).20a A simple model is shown in Figure 5D to illustrate such a lattice structure. Balls colored with violet represent gold atoms aligned in the (110) surface. A lighter color means that it is closer to the surface. A schematic model of the Cys molecule emphasizes a possible location on Au(110), while a detailed model will be discussed in the computation part. Here two issues should be addressed: (1) In contrast to Cys on Au(111) in the same buffer solution, Cys monolayers on Au(110) do not give any cluster structures. This observation is in contrast to the same molecule assembling on Au(111) in the same medium. In the latter case, highly ordered clusters composed of six Cys molecules in each cluster form Cys monolayers on the Au(111) surface.15b In addition, the coverage

of Cys on Au(110) (i.e., (6.74 ( 0.3) × 10-10 mol · cm-2) is larger than on Au(111) (i.e., 3.83 × 10-10 mol · cm-2); Cys packs more densely on Au(110) than on Au(111), meaning the Au(110) surface provides higher adsorption affinity. (2) Compared with Cys assembling on Au(110) in UHV, adsorption of Cys efficiently lifts the reconstruction of Au(110) in a liquid, and Cys homogenously adsorbs on the surface. This is different from the UHV condition where either cluster structures with eight Cys molecules in each cluster or dimers assembled on the 1 × 2 reconstructed Au(110) surface in UHV.16 Moreover, similar and smooth adsorption of Cys is found on both the inside of the terraces and on the edge or kinks on Au(110) in a liquid environment, whereas terraces and kinks gave different adsorption patterns for Cys in UHV.16 The most conspicuous feature of a Cys monolayer on Au(110) in a liquid environment is that STM imaging gives submolecular resolution of a single Cys, with three protrusions in a triangle.

2238 Langmuir, Vol. 25, No. 4, 2009

Zhang et al.

Table 1. Closest Au-S Distances, r(Au-S), Angle Characterizing the Position of the S Atom, r (See Figure 7G), Angle between the S-C1 Bond and the Metal Surface, ∠(Me-S-C1), Charge of the Model Cluster, q(Au21), and Adsorption Energya, ∆Eads, Calculated for the Cysteine Radical Adsorbed in Different Positions r(Au-S)/Å R/grad ∠(Me-S-C1)/grad q(Au21)/au -∆E ads/eV

OS bridge 1

NS bridge 1

NS bridge 2

S bridge 1

OS on-topb

NS on-topb

2.53-2.55 55 58 0.28 2.26

2.52-2.55 90 14 0.26 2.76

2.61-2.64 47 36 0.22 2.44

2.49-2.51 70 34 0.28 2.25

2.42 90 27 0.32 1.7

2.43 90 22 0.35 1.65

a ∆Eads ) Etot(Au21-Cys Rad) - Etot(Au21) - Etot(Cys Rad), where Etot is the total energy of the system. b In the quantum chemical calculations the projection of the S atom was fixed in the on-top position.

Table 2. Closest Au-S Distances, r(Au-S), Angle Characterizing the Position of the S Atom, r (See Figure 7G), Selected Valence and Torsion Angles, Normal (µ⊥) and Lateral (µ|) Projections of the Electrical Dipole Moment of the Adsorbate, and Adsorption Energy,a ∆Eads, As Obtained from Periodical DFT Calculations for a c(2 × 2) Lattice Consisting of Cys Radicals (Cys Zwitterions) Adsorbed in Two Different Bridge Positions

r(Au-S)/Å R/grad ∠(Me-S-C1)b/grad ∠(S-C1-C2-C3)/grad ∠(S-C1-C2-N) /grad µ⊥/D µ|/D -∆Eads/eV

Cys Radc S bridge 1

Cys Zwit S bridge 1

Cys Zwit S bridge 2

2.50 ÷ 2.54 77 33 70 164 2.06 1.95 2.58

2.50-2.56 79 37 46 163 1.35 11.1 3.35

2.47-3.98 57 51 72 169 2.06 10.7 3.11

a The adsorption energy is reckoned from the total energy of the Cys radical, ∆Eads ) Etot(Au slab-Cys) - Etot(Au slab) - Etot(Cys), where Cys means either Cys Rad or Cys zwitterions. A 29 Å × 29 Å × 58 Å unit cell was used to calculate the total energy of a single Cys radical. b Angle between the S-C1 bond and the metal surface. c We have not found any stable c(2 × 2) lattice consisting of Cys radicals adsorbed in the S bridge 2 site. The calculations started with such initial conditions predict the migration of adsorbed particles to the more stable position (S bridge 1).

This is in contrast to the beam shape of the dimer and round shape of Cys in a cluster of eight, six, or four molecules for Cys in UHV16 and Cys on Au(111) in a liquid.15 All this information suggests that both the orientation of the crystal and the solvent play an essential role in assembly of Cys on Au(110). The three protrusions in each triangle are most likely due to the contributions from functional groups of Cys, which is further supported both by STM observation in solutions of different pH values and by the DFT computation below. A similar observation of the Cys monolayer was found at pH 3.0, while higher pH does give different features in the monolayer structure. Figure 6A is an overview of a large scan area covered by a well-ordered Cys monolayer at pH 6.5 in 20 mM NH4Ac. Domains in a range of 6-15 nm were uniformly distributed over the whole area, Figure 6B. Interestingly, some small defects in a range of 0.7-1 nm are observed as black holes inside the highly ordered domains, Figure 6B, which is in contrast to the defect-free monolayer at pH 4.6, Figure 5B. A careful data analysis from a high-resolution STM image indicates a c(2 × 2) lattice with two Cys molecules in each unit cell, Figure 6C. This indicates that the pH does not change the lattice during Cys packing. Intriguingly, triangles with three protrusions (white spots) are again clearly distinguished, and one spot even gives higher contrast than the other two spots in each Cys, Figure 6C. Since the pKa of Cys is 8.33 in aqueous solution,47 with a reservation in the SAMs, neither the thiol nor carboxylate group could be (47) Heyrovsky, M.; Mader, P.; Varicka, S.; Vasela, V.; Fedurco, M. J. Electroanal. Chem. 1997, 430, 103–117.

deprotonated at both pH 6.5 and pH 4.6; therefore, the acid amino group is the only group which can be protonated at pH 6.5. Therefore, the high-contrast spot in the triangle of STM images might be attributed to the amino group in the cysteine monolayer. A deep understanding of the original STM contrast is desired from theoretical approach6 and will be discussed in the following part. 3.3. DFT Modeling of Cys Adsorption on a Au(110) Surface. 3.3.1. Geometric Optimization of a Single Cysteine Molecule on a Au21 Cluster. We have examined comprehensively several orientations of Cys Rad on the Au(110) surface shown in Figure 7. The species are bound to the metal surface by the sulfur headgroup (S), by both S and the carboxylic group (OS), or by both S and the amino group (NS). Some computational data are presented in Table 1. The adsorption energy (∆Eads) values range between 1.65 and 1.7 eV (on-top site) and between 2.25 and 2.76 eV (bridge sites), demonstrating a fairly strong Au-S covalent binding. The bridge 1 site provides the deepest point in the adsorption energy landscape according to results reported previously.25 We note that the calculated ∆Eads values are practically the same for the S and OS orientations. The Au-S bond lengths for the position bridge 1 lie in the interval of 2.49-2.55Å and coincide with those for Au(111).22,23 The S atom of Cys Rad in the most favorable orientations occupies a slightly distorted bridge position, shifted toward the hollow site (Figure 7G). The adsorption bond was found to be strongest for position NS (bridge 1), which supports a similar previous conclusion.22,23 A value of 2.36 Å was obtained for the Au-N bond length, which agrees well with that (2.35 Å) reported by Felice and Selloni23 for Cys Rad adsorbed at the Au(111) surface. For all positions the S-C1 bond is tilted relative to the surface plane and the corresponding angles range from 22 to 58 grad. The adsorption of Cys Rad results in some lengthening of the S-C1 bond (∼0.05 Å), while the change of lengths of the other chemical bonds does not exceed 0.01 Å. The partial transfer of electronic density from the Au cluster to the adsorbed Cys radical (0.22-0.35 au) is observed for all positions (Table 1). 3.3.2. Formation and Stabilization of Cys Rad and Zwitterion Monolayers. Since the adsorption on a bridge site was predicted as the most preferable from our cluster calculations, we addressed only this orientation in further periodical slab modeling. The Cys radical and zwitterion were considered; according to our results, the latter becomes stable within a monolayer. Only orientation S (Figure 7D) was found to favor the formation of a c(2 × 2) layer observed in the in situ experiment, whereas both the Cys radicals and zwitterions adsorbed in the other positions (OS and NS) form a less close-packed 2 × 2 lattice. The calculations therefore started with such initial conditions from which migration of adsorbed particles to the more stable position (bridge 1) is expected. A c(2 × 2) lattice composed of the zwitterions reveals stability for both bridge positions. Some results on the geometry of adsorbed species and the energy of adsorption bonds are illustrated in Table 2. The monolayer effects do not

STM Contrasts of Single Cys Molecules

Langmuir, Vol. 25, No. 4, 2009 2239

Figure 8. A view of the most stable c(2 × 2) lattice consisting of Cys zwitterions adsorbed at the Au(110) surface in the S bridge 1 position (as obtained from the periodical DFT calculations). Only the first layer of Au atoms is shown.

lead to a noticeable change of the optimal position of the S headgroup as compared to the data for adsorption of a single Cys species, although they show a slight lengthening of the Au-S bonds (Tables 1 and 2). It can be seen that the adsorbed zwitterions form a more stable monolayer as compared with Cys radicals; the most stable lattice structure is shown in Figure 8. In the absence of solvent the stabilization of Cys zwitterions in the adsorption layer originates mainly from intermolecular lateral interactions. Some additional evidence in favor of the crucial role of electrostatic interactions between the adsorbed Cys species is shown in Table 2. The normal and lateral projections of the Cys Rad electric dipole moment are comparable; attractive and repulsive electrostatic interactions within the monolayer compensate each other and hardly provide a significant energy gain. Therefore, Cys adsorption is governed in this case most likely by the direct metal-adsorbate bond. On the other hand, the lateral component of the total dipole moment of the adsorbed zwitterions exceeds several times the normal projection, and attractive interactions between the particles prevail. Since we have observed c(2 × 2) layers for both the Cys radicals and their zwitterions, it can be assumed that an energy barrier hampers proton transfer from the carboxyl group of the adsorbed cysteine radical to its amino group. This hypothetic barrier might be most likely readily overcome through proton tunneling. The c(2 × 2) layer of Cys radicals at the Au(110) surface is thus a significantly metastable state. Computational predictions have also supported the formation of stable SAMs at the Ag(111) surface in the form of Cys zwitterions.25 The energy values given in Table 2 rest on the adsorption from the gas phase. Solvent effects on the adsorption energy were addressed by a simple estimation. First, we assume that the adsorption of Cys species in a bridge position is accompanied by desorption of two water molecules (see Figure 8). A similar computation of the adsorption energy of a water molecule on a model cluster describing the Au(110) surface has been reported as 0.3 eV.48 Second, the formation of a zwitterion from Cys Rad in aqueous solution is known to be favorable.26 According to our calculations based on the polarized continuum model (PCM),32 the hydration energy of the Cys zwitterion is 1.9 eV. Then assuming full loss of the solvation sheath of Cys species in the adsorption monolayer (which might be considered as an upper (48) Ignaczak, A.; Gomes, J. A. N. F. J. Electroanal. Chem. 1997, 430, 209– 218.

limit of our estimation), we should subtract this value and the water desorption energy cost (0.6 eV) from the “gas phase” ∆Eads values. Despite a significantly reduced adsorption energy, the formation of a c(2 × 2) monolayer consisting of the Cys zwitterions at a Au(110)/water interface remains still apparently feasible. 3.4. Computation of STM Tunneling Contrasts. The STM configuration includes adsorbed species (strongly coupled to the metal surface) with fixed geometry and a tip with variable distance from the adsorbed molecule. The tunneling current i can then be recast in the form49

e i ≈ F(εF)|Hif|2 p

∫0φ Fads(ε) dε

(3)

where Hif is the electronic resonance integral (i refers to the electronic state on the tip, and f notes the electronic state on the adsorbate; see ref 26 for more details), φ is the bias voltage, F(εF) is the density of electronic states of the tip, Fads(ε) is the density of electronic states of the adsorbate, and εF is the Fermi level of the tip. The resonance integral in eq 3 is the only quantity which depends strongly on the distance. Calculation of a model STM image (constant-current mode) is therefore equivalent to the construction of isodensity contours Hif ) const.26 Neglecting a |∫φiφf dV|2 term, we can estimate the resonance integral as

Hif =

∫ φiVˆφf dV - ∫ φiVˆφi dV ∫ φiφf dV

(4)

where φi is the wave function of the tip, φf is the acceptor orbital of the adsorbed species, and Vˆ is the perturbation operator.26 As a difference from our previous studies,26 the present work has additionally addressed the second term in eq 4 (according to our test calculations, this term may play an important role in construction of model STM images). We have modeled the tip as a tungsten atom; its 5dz2 orbital was included explicitly in the calculations of Hif. A model STM contrast built at a certain Hif value for the Cys zwitterion adsorbed at the Au(110) surface in the most stable position (bridge 1, c(2 × 2) lattice) is displayed in Figure 9. The image indeed shows resolution at the submolecular level and demonstrates clearly three spots of different sizes, in good agreement with the experimental data (Figure 5). To clarify the origin of the spots, we have performed an additional analysis (see ref 26 for details). As shown by the analysis, the

2240 Langmuir, Vol. 25, No. 4, 2009

Zhang et al.

Figure 9. Model STM contrast (Hif ) 2.5 × 10-4 eV) built for the Cys zwitterion adsorbed at the Au(110) surface in the most stable position (S bridge 1, c(2 × 2) lattice). The distance scale is given in atomic units.

atomic orbitals of the COO group of the zwitterion contribute mainly to the large and small spots. On the other hand, the orbitals of the protonated amino group contribute solely to the spot of middle size. The contributions from both the sulfur headgroup and the -CH2-CH- residue were found to be negligibly small. Our computational results thus complement the experimental data and provide a deeper and transparent insight into the structure of the Cys SAM at the Au(110) electrode surface.

4. Concluding Summary Adsorption of Cys on Au(110) in ammonium acetate solution has been investigated by electrochemistry techniques including voltammetry, capacitance, chronocoulometry, and in situ STM as well as a theoretical study resting on model DFT calculations and quantum mechanical charge transfer theory. The adsorption depends on the potential and is accompanied by a lift of reconstruction of the Au(110) surface followed by formation of a highly ordered Cys monolayer. Cys molecules in their zwitterion form assemble into a c(2 × 2) lattice with two Cys molecules in each unit cell and offer a platform for high-resolution STM. The Cys zwitterion is bound to the metal surface by the sulfur headgroup in the “bridge” site, and the -NH3 and -COO groups are oriented toward the solution bulk. Three pH-dependent spots observed by STM represent a submolecule feature for a single Cys molecule. The Cys orientation, adsorption energy, adsorption bond, and tunneling contrast have been studied in the frame of a quantum chemical approach. The Cys bond with the gold substrate is a dominating controlling factor in the formulary of a c(2 × 2) lattice, while three functional groups in each Cys contribute mainly to the subfeatures of the STM images. The strongest contribution originates from the -COO atomic orbitals. This conclusion is robust and follows notably the most favorable bonding and orientation behavior of Cys in the zwitterions form, which is also the prevalent Cys species in aqueous buffer in the appropriate pH range.

5. Perspective In the present work we have initialized investigations of Cys on Au(110) in a liquid environment. Several challenges will be addressed in the future: (1) Investigation of chiral molecules on

(110) is an interesting topic for molecular reorganization, for example, interfacial binding with metal ions such as Cu(II) on surfaces, and holds a perspective for biological applications. Comparison of D- and L-Cys adsorption on (110) is therefore an obvious target of attention. (2) Both the solvent and orientation of the crystal play decisive roles in Cys assembly. The coverage of Cys on Au(110) is significantly larger than that on Au(111) in the same medium. It is thus of interest to explore adsorption of the same molecule on other crystalline facets and overview the three main low-index crystal facets. (3) Submolecular features of Cys on Au(110) are visualized by STM imaging. As shown by DFT studies in this work the origin of these features is clearly caused by the -NH3 and COO functional groups of Cys. Since contrasts in STM imaging reflect electronic conductivity of the molecules, the three protrusions in the Cys triangle represent the electronic conductivity of the functional groups of Cys. STM at such resolution allows construction of submolecular electronic circuits by a combination of designing suitable molecules and using input from theoretical study to predict the corresponding conductivity. Such a level of resolution, in a condensed matter environment at room temperature, and with accompanying theoretical support offers row openings in molecular design and surface immobilization. This would hold other perspective for supermolecular architecture and function characterized to a similar degree of detail and with perspective for molecular scale electronic and other functions in a real environment. Further modeling of both adsorption and STM contrasts will be aimed first to address the most important environmental effects (solvent, pH, electrode charge excess) as well as the electronic structure of the tip in a more rigorous and realistic way. Acknowledgment. We gratefully acknowledge financial support from the Danish Research Council for Technology and Production Sciences (Contract No. 26-00-0034) and Russian Foundation of Basic Research (Poject No. 08-03-00769-a). We appreciate Jens Ulstrup and W. Schmickler for helpful discussions. We are indebted to Dmitrii V. Glukhov for help with the periodical slab calculations. LA8034006