Nanoscopic Evidence for Dissociative Adsorption of Asymmetric

© Copyright 2000 American Chemical Society

MARCH 7, 2000 VOLUME 16, NUMBER 5

Letters Nanoscopic Evidence for Dissociative Adsorption of Asymmetric Disulfide Self-Assembled Monolayers on Au(111) Jaegeun Noh and Masahiko Hara*,† Frontier Research System, RIKEN (The Institute of Physical and Chemical Research), Wako, Saitama 351-0198, Japan Received October 28, 1999 We have monitored the adsorption process of 11-hydroxyundecyl octadecyl disulfide (CH3(CH2)17SS(CH2)11OH, HUOD) self-assembled monolayers (SAMs) on the Au(111) surface during the initial SAM growth stage using scanning tunneling microscopy (STM). STM imaging clearly exhibits two types of phase-separated domains having different corrugation periodicities which are consistent with the lengths of CH3(CH2)17S and HO(CH2)11S molecules, respectively. This is the first direct observation of the dissociative adsorption of disulfides on the nanometer scale. The self-assembly process of HUOD molecules physisorbed on graphite, on the other hand, is mainly governed by a hydrogen bond with a hydroxyl group facing another hydroxyl group of adjacent molecules without any S-S bond cleavage of the disulfide group unlike the adsorption process of HUOD molecules on gold. We have demonstrated a new and simple insightful method of comparing monolayers chemisorbed on Au(111) with those physisorbed on graphite.

Self-assembled monolayers (SAMs) formed by organic molecules on metal surfaces are of considerable importance for use in various technical applications such as biosensing, nonlinear optics, nanolithography, molecular recognition, corrosion inhibition, and nanoparticle formation.1-5 In particular, SAMs derived from alkanethiols and dialkyl disulfides on gold have been extensively studied because of their tendency to form highly ordered and densely packed monolayers having high stability.6-8 It is generally * To whom correspondence should be addressed. E-mail: [email protected]. † Also at: Department of Electronic Chemistry, Tokyo Institute of Technology, Nagatsuta, Midori-ku, Yokohama 226-8502, Japan. (1) Fuchs, H.; Ohst, H.; Werner, P. Adv. Mater. 1991, 3, 10. (2) Piner, R. D.; Zhu, J.; Xu, F.; Hong, S.; Mirkin, C. A. Science 1999, 283, 661. (3) Motesharei, K.; Myles, D. C. J. Am. Chem. Soc. 1998, 120, 7328. (4) Chailapakul, O.; Sun, L.; Xu, C.; Crooks, R. M. J. Am. Chem. Soc. 1993, 115, 12459. (5) Ohgi, T.; Sheng, H.-Y.; Nejoh, H. Appl. Surf. Sci. 1998, 130-132, 919. (6) Poirier, G. E.; Tarlov, M. J. Langmuir 1994, 10, 2853.

believed that SAMs made from both precursors form identical species which are adsorbed as alkanethiolates at the 3-fold hollow sites of the Au(111) surface.9-15 It has also been reported that SAMs formed on gold may exist as a dimer-like form of sulfur headgroups rather than as gold-bound alkanethiolates.16-20 (7) Alves, C. A.; Smith, E. L.; Porter, M. D. J. Am. Chem. Soc. 1992, 114, 1222. (8) Takami, T.; Deramarche, E.; Gerber, C.; Wolf, H.; Ringsdrof, H. Langmuir 1995, 11, 3876. (9) Nuzzo, R. G.; Zegarski, B. R.; Dubois, L. H. J. Am. Chem. Soc. 1987, 109, 733. (10) Bain, C. D.; Biebuyck, H. A.; Whitesides, G. M. Langmuir 1989, 5, 723. (11) Hagenhof, B.; Benninghoven, A.; Spinke, J.; Liley, M.; Knoll, W. Langmuir 1993, 9, 1622. (12) Biebuyck, H. A.; Whitesides, G. M. Langmuir 1993, 9, 1766. (13) Biebuyck, H. A.; Bain, C. D.; Whitesides, G. M. Langmuir 1994, 10, 1825. (14) Zhong, C. J.; Brush, R. C.; Anderegg, Porter, M. D. Langmuir 1999, 15, 518. (15) Carron, K. T.; Hurley, G. J. Phys. Chem. 1991, 95, 9979. (16) Fenter, P.; Eberhardt, A.; Eiesenberger, P. Science 1994, 266, 1216.

10.1021/la991423l CCC: $19.00 © 2000 American Chemical Society Published on Web 02/04/2000

2046

Langmuir, Vol. 16, No. 5, 2000

Despite a number of such studies, there remain some intrinsic questions regarding adsorption processes and states of dialkyl disulfide SAMs on the gold surface. For example, does the S-S bond cleavage of disulfide in the course of the formation of monolayers actually occur, and thereafter, do monolayers consist of gold-bound thiolates or a dimer-like form of sulfur headgroups, or if the dissociative adsorption of disulfides on gold occurs, what is the adsorption behavior of thiol moieties after the S-S bond cleavage on the gold surface? To date, the bond cleavage adsorption of disulfide on the gold surface has conventionally been studied using macroscopic techniques such as contact angle measurements, X-ray photoelectron spectroscopy (XPS), high-resolution electron energy loss spectroscopy, time-of-flight secondary ion mass spectroscopy, and Fourier transform infrared spectroscopy (FTIR). As a typical example, Biebuyck and Whitesides showed that the reaction of an unsymmetrical disulfide (HO(CH2)10SS(CH2)10CF3) with a gold surface yields SAMs containing approximately equal proportions of the two thiolate groups, and the S(CH2)10CF3 group in these mixed SAMs is replaced by S(CH2)10CN about 103 times faster than the HO(CH2)10S group by XPS study.12 This is strong evidence supporting the bond cleavage adsorption of disulfide and the formation of gold-bound thiolates. On the other hand, although scanning probe microscopy (SPM) is a very powerful technique for obtaining information on the molecular behavior and surface structures of adsorbates on substrates, to date there is still no clear evidence revealed by SPM concerning this matter. The main objective of our study is to elucidate the adsorption processes of dialkyl disulfide from the nanoscopic viewpoint using scanning tunneling microscopy (STM). A number of earlier studies have been performed with the aim of solving this problem using asymmetric or unsymmetric disulfides with STM and atomic force microscopy (AFM). These studies were mainly performed on the fully covered monolayers derived from compounds on the gold surface. This is a limitation for clarifying this problem because the exchange process, from the shorter alkyl chain thiolates to the longer alkyl chain ones during increasing surface coverage, lateral diffusion rate of adsorbed molecules on the surface, and interaction differences between adsorbing molecules and the solvent induced by two different alkyl parts, can affect the SAM formation.21-23 Therefore, the most important objective is to overcome this experimental restraint. In this study, we used an 11-hydroxyundecyl octadecyl disulfide (CH3(CH2)17SS(CH2)11OH, HUOD), which has different alkyl chain lengths and terminal groups. It is well-known that striped phases where molecules are oriented parallel to the surface have usually been observed in the low surface coverage region during the initial SAM growth stage, or from the surface which is prepared using an extremely dilute solution of organosulfurs.24-28 The striped phases depend (17) Fenter, P.; Schreiber, F.; Berman, L.; Scoles, G.; Eisenberger, P.; Bedzyk, M. J. Surf. Sci. 1998, 412/413, 213. (18) Nishida, N.; Hara, M.; Sasabe, H.; Knoll, W. Jpn. J. Appl. Phys., Part 1 1996, 35, 5866. (19) Nishida, N.; Hara, M.; Sasabe, H.; Knoll, W. Jpn. J. Appl. Phys., Part 2 1996, 35, L799. (20) Wan, L.; Hara, Y.; Noda, H.; Osawa, M. J. Phys. Chem. B 1998, 102, 5943. (21) Kajikawa, K.; Hara, M.; Sasabe, H.; Knoll, W. Jpn. J. Appl. Phys., Part 2 1997, 36, L1116. (22) Schlenoff, J. B.; Li, M.; Ly, H. J. Am. Chem. Soc. 1995, 117, 12528. (23) Folkers, J. P.; Laibinis, P. E.; Whitesides, G. M.; Deutch, J. J. Phys. Chem. 1994, 98, 563. (24) Poirier, G. E.; Pylant, E. D. Science 1996, 272, 1145. (25) Camillone, N., III; Leung, T. Y. B.; Schwartz, P.; Eisenberger, P.; Scoles, G. Langmuir 1996, 12, 2737.

Letters

strongly on the alkyl chain length of organosulfur compounds used for the formation of SAMs on metal surfaces. In this context, we investigated striped phases to be formed during the process of the SAM growth stage from solution containing asymmetric HUOD molecules. The gold substrates were prepared by the vacuum evaporation of gold onto freshly cleaved mica sheets prebaked to 300 °C for 2 h under a vacuum pressure of 10-7-10-8 Torr. After deposition, the substrates were annealed at 420 °C in a vacuum chamber for 2 h to obtain a large flat single-crystal surface. SAMs were prepared by dipping the gold substrates in 0.25 µM ethanol solutions of HUOD for appropriate lengths of time. After the substrates were removed from the solutions, the SAM samples were immediately rinsed with absolute ethanol to remove weakly adsorbed multilayers on the monolayer before drying the surfaces with a stream of N2. All STM imaging was performed within 6 h after sample preparation. There was no evidence of further surface reaction proceeding during the time period between removing the sample from the solution and STM imaging within 6 h. Moreover, we observed surface structures of HUOD molecules physisorbed on highly oriented pyrolytic graphite (HOPG) to compare with those of the molecules chemisorbed on the gold surface. In this case, the imaging process is slightly different from that for SAMs on gold and was carried out as reported in previous papers.29,30 A nearly saturated solution of this compound was made in phenyloctane solvent, which was purged with N2 for 20 min to remove any residual oxygen dissolved in the solvent prior to preparation of the solution. After a drop of solution was applied to the freshly cleaved surface of HOPG, the STM tip was immersed into the deposited droplet and images were obtained under the solution. In addition, it should be noted that the HUOD compound is highly stable without any changes of chemical composition, even in solutions which are placed longer than 2 days in the pure ethanol or phenyloctane solvents used in this study. The stability of this compound in these solvents was confirmed by mass spectroscopy in solution. Furthermore, we also examine a self-assembly process and the effect of the hydroxyl group during the SAM formation of these molecules by introducing different substrates. STM images in Figure 1 show the surface morphology of HUOD SAMs on Au(111) depending on the deposition time at the initial growth stage. After dipping the gold substrate in 0.25 µM HUOD solution for 1 min, there appeared an ordered island nucleation, as shown in Figure 1a. The formation of the well-ordered nucleation has revealed that individual molecules diffusing on the surface can bind at periodic herringbone dislocations, which are induced to relieve the surface strain created by the 22 × x3 reconstruction on Au(111), and grow up as nucleation by way of further aggregation.31 In addition, it has also been reported that these regular and periodic herringbone dislocations can be induced by the deposition of organic molecules on the surface reconstruction of zigzag dislocations and straight parallel lines.32 This reconstruction (26) Yamada, R.; Uosaki, K. Langmuir 1998, 14, 855. (27) Poirier, G. E. Langmuir 1999, 15, 1167. (28) Xu, S.; Cruchon-Dupeyrat, S. J. N.; Garno, J. C.; Liu, G.-Y.; Jennings, G. K.; Yong, T.-H.; Laibinis, P. E. J. Chem. Phys. 1998, 108, 5002. (29) Venkataraman, B.; Breen, J. J.; Flynn, G. W. J. Phys. Chem. 1995, 99,, 6608. (30) Venkataraman, B.; Flynn, G. W.; Wilbur, J. L.; Folkers, J. P.; Whitesides, G. M. J. Phys. Chem. 1995, 99, 8684. (31) Chambliss, D. D.; Wilson, R. J.; Chang, S. Phys. Rev. Lett. 1991, 66, 1721. (32) Hara, M.; Sasabe, H.; Knoll, W. Thin Solid Films 1996, 273, 66.

Letters

Langmuir, Vol. 16, No. 5, 2000 2047

Figure 1. STM images of 11-hydroxyundecyl octadecyl disulfide (HUOD) SAMs on Au(111) obtained after (a) 1 min and (b) 3 min of deposition in 0.25 µM ethanol solutions of HUOD. (a) The STM image showed an ordered nucleation of these molecules on the gold surface during the initial SAM growth stage. The height of the nuclei reached about 0.5 nm, while the monatomic step height of Au(111) was about 0.25 nm. The scan size was 205 nm × 205 nm, and imaging conditions were 500 mV (sample positive) and 0.3 nA in the constant current mode. (b) Two phase-separated domains (A and B), formed as the result of the S-S bond cleavage of disulfide, were clearly observed in this image. The scan size was 130 nm × 130 nm, and imaging conditions were 300 mV (sample positive) and 0.2 nA in the constant current mode.

Figure 2. STM images exhibiting the striped phase of 11-hydroxyundecyl octadecyl disulfide (HUOD) physisorbed on graphite. Bright rows correspond to the disulfide groups. Molecular arrangements were strongly dependent on the hydrogen bond derived from the hydroxyl group positioned at the end of the shorter alkyl chain among the two different alkyl chain lengths attached to the disulfide group. (a) The scan size was 160 nm × 160 nm, and imaging conditions were 1500 mV (sample positive) and 0.12 nA in the constant current mode. (b) The scan size was 16 nm × 16 nm, and imaging conditions were 1300 mV (sample positive) and 0.12 nA in the constant current mode.

process results in an ordered nucleation of these molecules on the gold surface during the adsorption process. The STM image in Figure 1b clearly exhibits the disordered intermediate phase and two types of stripedphase domains with phase separation of the SAM sample obtained after 3 min of deposition. The corrugation periodicity values of the striped phases in regions A and B are 1.80 and 2.53 nm, which show good agreement with the length of HO(CH2)11S and CH3(CH2)17S molecules, respectively. To interpret the origin of these two striped phases more precisely, the orientation of the two alkyl chains attached to the disulfide group on the surface must be examined. In a previous report, it was revealed that the molecular axis of didococyl disulfide physisorbed on the graphite surface is linear and the periodicity of the striped phase corresponds to the length of the molecules.30 This straight orientation of the two alkyl chains is energetically much more favorable than the U-shape orientation in a molecule because the latter orientation

can induce the steric repulsion of the two alkyl chains and a large strain in the C-S-S-C bond angle. As shown in Figure 2, the observation of paired lines in molecular arrangements of HUOD molecules on the graphite surface reveals that the molecular axis is close to linear, as in the case of didococyl disulfide. Moreover, two striped phases observed on Au(111), which showed that the two corrugation periodicities are nearly consistent with the length of HO(CH2)11S and CH3(CH2)17S molecules, could not be formed without the bond cleavage of the disulfide group. This finding strongly implies that the S-S bond cleavage of asymmetric disulfide has already occurred when the HUOD molecules adsorb on the surface in the reaction with gold at room temperature and that the direct adsorption of the original disulfide cannot occur on the surface during the SAM formation. This result can also be drawn from the fact that if there is no S-S bond cleavage during the adsorption process of disulfide, the striped periodicity should show the value of an entire molecular

2048

Langmuir, Vol. 16, No. 5, 2000

length of HUOD molecule. However, we could not find any striped phases that were formed by the original HUOD molecules without the bond cleavage of disulfide. From this aspect, this STM imaging is the first direct observation of the dissociative adsorption of disulfides on Au(111) on the nanometer scale. The disordered phase between striped phases, on the other hand, can be attributed to the phase transition from the striped phase to the standing-up phase as surface coverage increases. In addition, the observation of these two striped phases, whose periodicities are nearly equal to the single molecular length of HO(CH2)11S and CH3(CH2)17S, suggests that the sulfur atoms bound to the gold remain not in dimer-like form of sulfur headgroups but as gold-bound monomers at low surface coverage. A recent study by Kondoh et al. using thermal desorption spectroscopy (TDS) showed that the desorption of dimers is observed only for the closepacked surface and not for the striped phase.33 It can also be assumed based on our previous TDS study that the dimerization of sulfur headgroups proceeds in the standing-up phase as the surface coverage of monolayers increases.18 Why can we observe and identify two kinds of stripedphase domains composed of a single component among the two different thiolates formed after the bond cleavage of disulfide in this SAM system? It is most likely due to the larger stabilization energy gained from the optimized lateral interaction between the same molecules than that gained from the lateral interaction between different molecules. In the case of alkane molecules on graphite, they orient at 90° with respect to the troughs and, hence, have complete overlap along the entire length of molecules.29 This results in optimum van der Waals interactions between neighboring molecules. This molecular behavior on graphite is applicable in the initial SAM adsorption stage on gold as well. To confirm our result of the S-S bond breaking on Au(111), we have examined the molecular arrangements of HUOD molecules physisorbed on graphite. The main difference in the self-assembly process of HUOD molecules on gold and graphite is the interactions (chemisorption or physisorption) between the molecule and the substrate, as discussed above. Organic molecules on the graphite surface usually lie flat on the surface similar to striped phases observed on gold. The STM image of a large scan (33) Kondoh, H.; Kodama, C.; Nozoye, H. J. Phys. Chem. B 1998, 102, 2310. (34) Nejoh, H. Appl. Phys. Lett. 1990, 57, 2907.

Letters

area in Figure 2a shows the well-ordered molecular arrangements of HUOD molecules that orient parallel to the graphite surface. The high-resolution STM image in Figure 2b displays the individual disulfide groups, indicated by spots in bright lines, and about 23° tilted orientation of alkyl chains. The bright lines corresponding to the disulfide group in this STM image may be due to the large molecular polarizability and the increase of the local density of states near the Fermi level of the surface.30,34 These bright lines are paired by a hydrogen bond through a hydroxyl group of the alkyl chain, facing another hydroxyl group of adjacent molecules. Here, two longer alkyl moieties are located in the large dark area (region A) between two bright lines, and two shorter hydroxylated alkyl moieties are located in the small dark area (region B). The three bright lines shown in region C, on the other hand, may be due to the incorporation of bis(11-hydroxyundecyl) disulfide ((HO(CH2)11S)2) molecules as an impurity into the monolayers during the monolayer formation on graphite. These results suggest that the hydrogen bond between molecules is an important factor in determining the molecular arrangements in this system. The STM image supports the finding that the self-assembly process on graphite proceeds without any disulfide bond rupture as expected. On the basis of our STM results, we confirmed for the first time from a nanoscopic viewpoint that the selfassembly of disulfide on gold proceeds along with the bond breaking of disulfide by chemical reaction between the disulfide group and gold without the direct adsorption of the disulfide group on the gold surface. Furthermore, we suggest that the SAM formation on the gold surface was not strongly influenced by a hydrogen bond, at least in the initial growth stage, because we did not find any striped phases induced by a hydrogen bond between molecules, while phase separation had already begun in the striped phase by surface diffusion. We believe that the observation of striped phases with a well-designed model compound will provide a new guideline to probe and interpret new fundamental aspects, such as surface reactions, interactions, mobilities, and orientation mechanisms of molecules, and to monitor target molecules on a surface as shown in this study. In addition, we demonstrated that comparing the monolayer ordering on Au(111) with that on graphite is a simple and insightful method which is widely applicable to various SAM systems. LA991423L