Partial Electron Transfer in Octadecanethiol Binding to Gold

Langmuir 1994,10, 4286-4294

4286

Partial Electron Transfer in Octadecanethiol Binding to Gold? Pawel Krysinski,* Richard V. Chamberlain 11, and Marcin Majda* Department of Chemistry, University of California in Berkeley, Berkeley, California 94720 Received July 12, 1994@ Langmuir-Blodgett transfer of octanedecanethiol(Cl&H) containingmonolayers onto gold-coatedglass slides under potentiostatic conditions leads to a quantitative characterization of Cl&H chemisorption on gold. The thiol-gold bond involves a partial electrontransfer from sulfur to the gold substrate. The CL$H partial charge number was determined to vary linearly with the potential of gold substrates from 0.30 at -0.30 V to 0.45 at 0.70 V vs SCE. Only ca. 81%of Cl&H molecules appear to chemisorb upon LB transfer at 20 mNlm. This is likely a result of some orientational restrictions (e.g. tilt and registry with the gold surface lattice) imposed on the system by the lateral pressure during LB transfer. The requirement of monolayer charge neutrality, stemming from a negligible capacitance of the Au/LB monolayer interface, results in a dissociation of thiol's proton to an extent equal to the charge transferred due to C&H chemisorption.

Introduction The well-ordered structure, simplicity of preparation, and wide ranging applications of self-assembled alkanethiol monolayers on gold surfaces have attracted the considerable interest of scientists from a spectrum of backgrounds from chemical p h y s i ~ i s t s l - ~and elec~ - ~analytica120-24 ~ chemists. trochemists5-13to ~ r g a n i c l and This relatively new area is already a subject of several review^.^^-^^ Research in this area has been concerned,

* To whom all correspondence should be addressed. +Dedicated to Professor Dr. Hab. Zbigniew Galus, Warsaw University, on the occasion of his 60th birthday. Permanent address: Department of Chemistry, Warsaw University, ul Pasteura 1, 02-093 Warsaw, Poland. Abstract published inAdvance ACSAbstracts, October 15,1994. (1) Camillone, N., 111; Chidsey, C. E. D.; Eisenberg, P.; Fenter, P.; Li, J.; Liang, K. S.; Liu, G.Y.; Scoles, G. J. Chem. Phys. 1993,99,744. (2)Camillone, N., 111; Chidsey, C. E. D.; Liu, G. Y.; Scoles, G. J . Chem. Phys. 1993,98, 3503. (3) Camillone, N., 111; Chidsey, C. E. D.; Liu, G.; Scoles, G. J . Chem. Phys. 1993, 98, 4234. (4) Camillone, N., 111; Chidsey, C. E. D.; Liu, G.; Putvinski, T. M.; Scoles, G. J . Chem. Phys. 1991, 94, 8493. (5) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J . Am. Chem. SOC.1987,109,3559. (6) Sabatini, E.; Rubinstein, I. J . Phys. Chem. 1987, 91, 6663. (7) Chidsey, C. E. D.; Loiacano, D. N. Langmuir 1990, 6,682. (8) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M. J . Am. Chem. SOC.1990,112,4301. (9) Creager, C. E.; Collard, D. M.; Fox, M. A. Langmuir 1990, 6,

*

@

1617. (10) Miller, C. J.; Cuendet, P.; Gratzel, M. J . Phys. Chem. 1991,95, 877. (11) Redepenning, J.; Tunison, H. M.; Finklea, H. 0.Langmuir 1993, 9, 1404. (12) Sun, L.; Crooks, R. M. Langmuir 1993,9, 1951. (13) Chailapakul, 0.;Sun, L.; Xu, C. J.; Crooks, R. M. J.Am. Chem. SOC.1993,115, 12459. (14) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evail, J.;Whitesides, G. M.; Nuzzo, R. G. J . Am. Chen. SOC.1989,111, 321. (15) Bain, C. D.; Evall, J.; Whitesides, G. M. J.Am. Chem.SOC. 1989, 111, 7155. (16) Bain, C. D.; Whitesides, G. M. J . Am. Chem. SOC.1989, 111, 7164. (17) Bain, C. D.; Whitesides, G. M. Angew. Chem. 1989,101, 506. (18)Biebuyck, H. A.; Whitesides, G. M. Langmuir 1993,9, 1766. (19) Folkers, J . P.; Laibinis, P. E.; Whitesides, G. M. Langmuir 1992, 8, 1330. (20) Spinke, J.; Liley, M.; Gudav, H.-J.; Angermaier, L.; Knoll, W. Langmuir 1993, 9, 1821. (21) Chailapakal, 0.; Crooks, R. M. Langmuir 1993, 9, 884. (22) Sun, L.; Crooks, R. M.; Ricco, A. J . Langmuir 1993, 9, 1775. (23) Kepley, L. J.; Crooks, R. M.; Ricco, A. J . Anal. Chem. 1992,64, 3191. (24) Rubinstein, I.; Steinberg, S.;Tor, Y.; Shanzer,A.;Sagiv,J.Nature 1988,332,426. (25) Whitesides, G. M.; Laibinis, P. E. Langmuir 1990, 6, 87.

0743-7463/94/2410-4286$04.50l0

to a large extent, with structural characterization of the monolayers of n-alkanethiols, disulfides, and their w-derivatives on gold, silver, and other metal Related research involves earlier studies of self-assembled alkyltrichlorosilane monolayer^^^,^^!^' and LangmuirBlodgett These monolayer structureshave been used to model and to investigate interfacial properties of biological membranes and organic materials in processes and phenomena such as molecular recognition, iontransport, wettability, adhesion, friction, biocompatibility, and other^.^^-^^ Structural stability and molecular order of alkanethiol monolayers on gold allowed systematic studies of long range electron t r a r ~ s f e r ' ~ and ! ~ ~opened -~~ a number of possibilities in designing chemical sensors.21-24,44,46 Formation of the sulfur-gold bond in alkanethiol selfassembly is, naturally, the process of fundamental importance to a large fraction of research mentioned above. Surprisingly, then, the understanding of some key mechanistic elements involved in this reaction still remains incomplete and contr~versial.'~,~~,~~-~~ Specifically, XPS (26) Ulman, A. An Zntroductwn to Ultrathin Organic Films; From Langmuir-Blodgett to Self-Assembly;Academic Press: San Diego, CA, 1991. (27) Dubois, L.; Nuzzo, R. G. Annu. Rev. Phys. Chem. 1992,43,437. (28) Bard, A. J.; Abrufia, H. D.; Chidsey, C. R.; Faulkner, L. R.; Feldberg, S. W.; Itaya, IC;Majda, M.; Melroy, 0.;Murray, R. W.; Porter, M. D.; Soriaga, M. P.; White, H. S. J . Phys. Chem. 1993,97, 7147. (29) Nuzzo, R. G.; Dubois, L. H.; Allara, D. A. J. Am. Chem. SOC. 1990, 112, 558. (30) Popenoe, D. D.; Deinhammer, R. S.; Porter, M. D. Lungmuir 1992, 8, 2521. (31) Bryant, M. A.; Pemberton, J . E. J . Am. Chem. SOC.1991,113, 8284. (32) Carron, K. T.; Hurley, L. G. J . Phys. Chem. 1991,95, 9979. (33) Strong, L.; Whitesides, G. M. Langmuir 1988, 4, 546. (34) Widrig, C. A.;Alves, C. A.; Porter, M. D. J.Am. Chem.SOC. 1991, 113, 2805. (35) Alves, C. A.; Smith, E. L.; Porter, M. D. J.Am. Chem. SOC.1992, 114, 1222. (36) Sagiv, J . J . Am. Chem. SOC.1980, 102, 92. (37) Maoz, R.; Sagiv, J . J. Colloid Interface Sei. 1984, 100, 465. (38) Langmuir-BlodgettF i l m ;Roberts, G., Ed.; Plenum Press: New York, 1990. (39) Chidsey, C. E. D. Science 1991, 314, 13. (40) Miller, C. J.; Gratzel, M. J . Phys. Chem. 1991, 95, 5225. (41) Becka, A. M.; Miller, C. J . J . Phys. Chem. 1992, 96, 2657. (42) Finklea, H. 0.; Hanshew, D. D. J . Am. Chem. Soc. 1992,114, 3137. (43) Becka, A. M.; Miller, C. J. J . Phys. Chem. 1993, 97, 6233. (44) Hughes, R. C.; Ricco, A. J.; Butler, M. A.; Martin, S. J. Science 1991,254, 74. (45) Hickman, J . J.; Offer, D.; Laibinis, P. E.; Whitesides, G . M.; Wrighton, M. S. Science 1991,252, 688.

0 1994 American Chemical Society

Cl&H Chemisorption on Gold

Langmuir, Vol. 10,No. 11, 1994 4287

data supported by IR spectroscopy were used to suggest that chemisorption of alkanethiols on gold(0) surface results in the formation ofgold(1) thiolate denoted as either RS-Au(1) or RS-Au+.14,31z47*49 It has then been postulated that this general mechanism requires loss of hydrogen as H2 (eq l ) , formation of water (eq 2), or of some other products of the surface oxidation such as hydrogen peroxide (eq

3).25947149950

-

+ l/,H, BRSH + 2Au(O), + 1/20, - 2RS-Au(I), + H,O BRSH + 2Au(O), + 0, - 2RS-Au(I), + H,O, RSH + Au(O),

RS-Au(I),

(1) (2)

(3)

Indeed, H2Oz was detected in one study involving selfassembly of propanethiol on gold powder in an oxygensaturated aqueous solution.50 Its release was postulated to be a result of 0 2 and H2reaction following reaction 1. The mechanistic schemes presented above seem to be supported by the measurements of surface potential4*and by the electrochemical studies of alkanethiol reductive d e s ~ r p t i o n . ~Recent ~ - ~ ~ laser desorption mass spectrometric studies point at the chemisorption of alkanethiolates rather than a l k a n e t h i ~ l s .Temperature-programmed ~~ desorption spectra of monolayers of methanethiol and dimethyl disulfide on gold(ll1) allowed estimation of the binding energy of the former to be 12-14 kcal/mol (indicating physisorption) while the S-Au bond strength in chemisorption of the latter was estimated to be ca. 45 k c d m 0 1 . ~Overall ~ then, as much as the thiolate-gold bond formation has been well established, the exact character of this bond and the identity, or even presence, of the products of the surface oxidation reaction (eqs 1-3) a r e not well known. In this report, we present a novel electrochemical method which combines Langmuir-Blodgett transfer technique with either potential (at open circuit) or current measurements (under potentiostatic conditions). This method allows us to determine that the primary step of alkanethiol chemisorption involves a partial electron transfer from thiol to gold. The partial charge number (charge per thiol molecule) is a linear function of the potential of gold substrates and varies from 0.30 electron at -0.300 V to 0.45 electron at 0.700 V vs SCE. Consideration of a negligible capacitance of the gold/ClsSH interface following LB transfer and an associated charge neutrality requirement leads to a postulate of partial thiol

I

T

1

Initial

0.50 cm

final

Figure 1. Schematic diagram illustrating Langmuir-Blodgett transfer of a monolayer from the aidwater interface onto a gold-coated substrate with concurrent monitoring of the substrate's potential. Below, an inset shows the pattern of the vapor-deposited gold film. The central rectangular area (A = 0.20 cm2)is coated with an LB monolayer as the substrate is withdrawn from the subphase. The two lines mark the initial and the final positions of the triple-phase line in the LB experiments.

grade, Aldrich),NaC104, KC1, and KOH (all from Fisher Scientific, ACS grade). House distilled water was passed through a fourcartridge Barnstead Nanopure I1 purification train consisting of Macropure pretreatment, Organics Free, two ion-exchange columns, and 0.2-pm hollow fiber final filter for removing particles. The final resistivity of water used in the experiments was in the range 18.1-18.3 MQ cm. Electrode Fabrication Procedures. Electrodes were produced by vapor deposition of about 100 nm thick gold films (99.95%Au, Lawrence Berkeley Laboratory) on microscopeglass slides (Corning Plain Micro Slides). The rate of deposition was ca. 5-10 k s . A 6-10 nm thick chromium underlayer was evaporated first to provide good adhesion and stability of the gold layer. The geometrical pattern of the vapor-deposited electrodes was specially designed for the LB electrochemical experiments and imprinted with the use of thin aluminum masks. The pattern is shown in Figure 1(for a full description see text in Results and Discussion). The surface area of its central rectangular fragment was 0.20 cm2. Prior to the vacuum depositionof Cr and Au films, glass slides were washed in chromic dissociation upon chemisorption. acid, rinsed thoroughly in Nanopure water, and then dried. Subsequently, they were boiled in 50:30:20 2-propanoVmethanoV Experimental Section chloroform mixture, dried again, and transferred to the evaporator (Veeco 7700). After fabrication, vapor-deposited electrodes Chemicals. 1-Octadecanol (Aldrich, 99%) and n-octadewere transferred and stored in a vacuum desiccator. They were canethiol (Aldrich, 98%)were triply recrystallized from ethanol. used in the Langmuir-Blodgett (LB) experiments typically All other reagents were used as received from the supplier: within 1 week from their fabrication. Immediately before the chloroform (Fisher ACS certified), perchloric acid (70%reagent LB experiments, gold electrodes were visually examined t o select those free of any scratches or blemishes. The cleaningprocedure (46) Nuzzo, R. G.; Zegarski, B. R.; Dubois, L. H. J.Am. Chem. SOC. involved rinsing with organic solvents (methanol, acetone, 1987,109, 733. (47) Bain, C. D.; Biebuyck, H. A.; Whitesides, G. M. Langmuir 1989, 2-propanol,chloroform)and Nanopure water. Subsequently, gold 5, 723. electrodes were immersed into a fresh, hot (ca. 80 "C) chromic (48) Evans, S. D.; Ulman, A. Chem. Phys. Lett. 1990,170, 462. acid solution for about 25 s. The chromic acid treatment step (49) Laibinis,P.E.;Whitesides,G.M.;Allara,D.L.;Tao,Y.-T.;Parikh, was followed by an extensive rinsing with Nanopure water. Upon A. N.; Nuzzo, R. G. J.Am. Chem. SOC. 1991,113, 7152. completion of this step, the electrodes were dried in a stream of (50) Widria, C. A.; Chung, - C.; Porter, M. D. J. Electroanal. Chem. argon and transferred immediately (in order to minimize their 1991, 310, 3$5. (51) Walczak, M. M.: PoDenoe. D. D.: Deinhammer. R. S.:. LamD, contamination) into the Plexiglas enclosure ofthe LB instrument. - B. D.; Chung, C.; Porter, M.D.Laigmuir 1991, 7, 2687. Before each LB transfer experiment, the surface of a gold (52) Weisshaar, D. E.; Walczak, M. M.; Porter,M. D.Langmuir 1993, substrate was electrochemically reduced and conditioned by 9, 329. voltammetric scanning in the potential range from -0.300 to (53) Wiesshaar, D. E.; Lamp, B. D.; Porter, M. D. J.Am. Chem. SOC. 0.600 V at 50 mV/s directly in the subphase in the LB trough. 1992,114, 5860. In the case of potentiometric experiments, the voltammetric (54) Li, Y . ;Huang, J.; McIver, R. T., Jr.; Hemminger, J. C. J.Am. Chem. Soc. 1992,114, 2428. cycling was stopped at 0.0 V during an anodic scan, and then the

Krysinski et al.

4288 Langmuir, Vol. 10, No. 11, 1994 potentiostat was switched to open circuit conditions. The gold substrates were allowed typically 10 min to attain a stable potential value. Similarly, in the case of the potentiostatic experiments, each potential value selected for a particular LB transfer experiment was applied by interrupting an anodic scan. Langmuir-Blodgett Procedures. The Langmuir-Blodgett experiments were carried out with a KSVMinitrough instrument operated under computer control. Routine Langmuir-Blodgett procedures and the LB transfer protocol were described previously.55 C~~SWClsOH monolayers were spread as chloroform solutions at a surface concentration corresponding to 35 Az/ molecule. Following solvent evaporation, the monolayers were compressed at a rate of ca. 4 &/(molecule/min) to a pressure of 20 mN/m. The LB transfers were carried out at that pressure at a rate of 10 mm/min unless specified otherwise. Instrumentation. The potential control of the gold electrode substrates and current measurements during LB transfer experiments were accomplished with a three-electrode potentiostat (CV 27 by BAS, Inc., West Lafayette, IN). The potentiometric measurements were done with the same instrument operated at open circuit. The counter and reference (SCE) electrodeswere placed in the LB trough behind the moving barrier to prevent their interference with the monolayer compression and transfer. In order to avoid C1- leakage, the reference electrode was isolated from the subphase in the LB trough via a salt bridge filled with the same electrolyte as that in the trough. The current'potential, current'time and potentidtime data were recorded on a X-Y recorder.

Results and Discussion Langmuir-Blodgett Transfer under Potentiometric Conditions. The experiments described in this and in the following sections concern standard LB transfers of octadecane alcohol (CISOH) and mixed octadecane alcohol and octadecanethiol (C18SH) monolayers from the water surface onto gold substrates fabricated by vapor deposition of 100nm thick gold films on glass slides (see Experimental Section). Behavior of Langmuir monolayers of ClBOH on the water surface is well-known in the literature.26~2s Pure ClaSH cannot be compressed on the aqueous subphase above ca. 12 mN/m due to insufficient polarity of the thiol We showed earlier that CISSH and CISOH are fully miscible in their mixed monolayers which combine the properties of the individual compone n t ~ Specifically, . ~ ~ the collapse pressure of the mixed system increases linearly with the increased mole fraction of ClaOH. All LB transfer experiments discussed below were carried out a t 20 mN/m. To accomplish this, the ClSSH mole fraction must not exceed 70%.56 Figure 1shows schematically a n LB experiment with potentiometric monitoring of the gold substrate. In a n LB experiment, a solid substrate is slowly pulled vertically up from the aqueous subphase across the water/air interface. This results in a transfer of a monolayer from the water surface onto the substrate's surface. During the LB transfer, the monolayer on the water surface is maintained a t a selected constant surface pressure by the forward motion of the barrier of the Langmuir trough controlled by the feedback mechanism of the instrument. As we show below, the transfer of both CISOH and CISSWClsOH monolayers can be classified as, so-called, reactive transfer57during which the surfactant molecules interact directly with the substrate's surface at the moment of transfer. The nonreactive transfer involves a transfer of a monolayer onto a thin film of water which subsequently is squeezed out as the monolayer is deposited (55) Lindholm-Sethson, B.; Orr, J. T.; Majda, M. Langmuir 1993,9, 2161. (56)Bilewicz, R.;Majda, M. Langmuir 1991, 7 , 2794. (57)Gaines, G.L. Insoluble Monolayers at Liquid Gas Interfaces; Interscience Publishers, J. Wiley C Sons: New York, 1966;Chapter 8, pp 326-333.

0.150

-I

1

0.0 0

20

40

60

t/S

Figure 2. Typical potential-time transients recorded during LB transfer of CISOHand C&H/ClsOH (70:30mol %) monolayers at 20 mN/m, from 0.05 M HC104 subphase onto goldcoated glass slides. LB transfer rate was 10 mm/min.

on the surface of the substrate. In the latter case, the dynamic contact angle observed during the LB transfer, 8, is essentially zero while its value is significantly greater than zero during a reactive transfer leading to a lower water meniscus.57 The values of 6 observed on the gold surface in our experiment were approximately 70" and 80"for ClsOH and C1BSWCl8OHmonolayers, respectively. Since the dynamic contact angles on the glass surface are close to zero, it is difficult to make a precise assessment of 8 on gold. Nevertheless, the high values reported above clearly demonstrate the reactive character of the monolayer transfer of gold. The geometric pattern of the vapor-deposited gold films on glass slides used in our experiments is shown in Figure 1. The lines coinciding with the top and the bottom edge of the central rectangular part of the electrode mark the initial and the final positions of the water meniscus in all experiments. Thus, the top circular area was always kept above the water surface and used to make the electrical contact while the bottom circular area was always submerged in the aqueous subphase. In the course of each LB experiment, the rectangular part was withdrawn above the water surface a t a constant rate (typically 10 m d m i n ) while coating it with a monolayer from the water surface. Figure 2 shows two typical recordings of the potential of the gold substrate monitored during the LB transfer experiments carried out on 0.05 M HC104 (70: subphase. In both cases of CisOH and C~~SWClaOH 30 mol %) monolayers, the initial open circuit potential of the gold substrates was ca. 0.13 V vs SCE. Gold substrates approximately reach this value in the course of a 10 min equilibration in the subphase. As we show in the next section, this value of a stable open circuit potential is approximately 0.21 Vnegative ofthe potential of zero charge (PZC) of the gold substrate. Consider first the potential-time trace obtained during LB transfer of CISOHmonolayer. Independent measurements showed that the transfer ratio of CISOH under these conditions is 1.0 f0.05. A linear decrease ofthe electrode potential is observed (Figure 2) during the withdrawal of

Langmuir, Vol. 10, No. 11, 1994 4289

CI&H Chemisorption on Gold Scheme 1

1

alkanethiols is naturally expected to result in much stronger interactions. Hence, in this case, we postulate that a much larger decrease ofthe gold substrate potential is due to two factors: (1) the double-layer charge “squeezing” as above and (2) electron transfer from C&H molecules to gold substrate a s they bind to its surface. Thus, we postulate that the primary event in the CISSH-gold bond formation can be represented by the reaction R-S

I

+ Au,

-

the rectangular section,58followed by a gradual increase of the electrode potential aRer the completion of the LB transfer when the entire rectangular area of the electrode was brought above the water surface. Since the initial potential of the gold substrate is negative of PZC, the observed decrease of the potential is consistent with a n increase of the negative excess charge density on the gold surface remaining under water caused by a shift of the excess negative charge from the upper part ofthe electrode. In other words, as the CISOH monolayer is deposited on the rectangular section of the gold substrate and the latter is emersed dry from the electrolyte solution, the excess negative charge stored at that interface is “squeezed”down to the interface remaining under water and leads to the observed decrease of the potential. The double-layer charge “squeezing“ observed in this experiment merely reflects the large difference between the interfacial capacitance of the clean gold electrode immersed in 0.05 M HClO4 (ca. 20pF/cm2)and that of the gold surface coated with C180H monolayer in the air. The latter can be assumed to be negligibly small. This assumption is well justified since the interfacial capacitance of the gold surface coated with octadecanethiol or similar monolayer assemblies in aqueous electrolytes is only ca. 1 . 0 p F / ~ m ~ .Thus ~ J ~the * ~ situation ~ at the end of the experiment, when only the lower part of the electrode remains under water and stores all the charge, can be represented in terms of a n equivalent circuit in Scheme 1with C1 lo9 S2) eliminates the possibility of instrument “leakage” current as a reason for the observed potential increase. Let us consider now the second potential-time trace in Figure 2 corresponding to the LB transfer of a mixed CISSWClsOH (70:30 mol %) monolayer. Unlike aliphatic alcohols, for which weak physisorption on gold does not involve partial electron t r a n ~ f e r , chemisorption ~~,~~ of ~~

(58)The magnitude of the potential decrease observed in such experiment is independent of the presence or absence of oxygen in the subphase electrolyte. (59) Richter, J.; Lipkowski,J. J.Electroanul. Chem. 1988,251,217.

I

H

H

c, >> c1

+

R - S + ~ * * A U ~ 6e(4)

where 6 is thepartial charge number expressing the order and the strength of this bond per alkanethiol molecule.61 Additional mechanistic steps accompanying the partial electron transfer are discussed below. In view of the notation used in eqs 1-3, it is perhaps worth pointing out that no distinction is made between gold atoms on the surface, as far as electron energetics are concerned, whether they engage or do not engage in thiol-gold bond formation. Electron energy at the gold surface is expressed by the electrode potential. As explained above, charge stemming from both doublelayer charge “squeezing”and C d H binding accumulates at the clean goldelectrolyte interface resulting in a more significant negative shift of the potential as shown in Figure 2. Langmuir-Blodgett T r a n s f e r under C o n s t a n t Potential Conditions: ClsOH Monolayers. Quantitative interpretation of the changes of the potential of the gold substrate in order to obtain the partial charge number of C&H on gold requires the exact value of the doublelayer capacitance of the goldelectrolyte interface and its dependence on the electrode potential. It is therefore more convenient to carry out the experiments of Figure 2 under constant potential conditions. In those experiments, 2 three-electrode potentiostat was used. The gold substrate was connected as a working electrode. As previously, LB transfer of the CISOH monolayer results in “squeezing”of the double-layer charge upon emersion of a dry gold substrate. In this case, however, excess charge removed from the emersed gold surface cannot accumulate at the part of the electrode under water since its potential is kept constant by the potentiostat. Instead, the doublelayer charge is collected and recorded as discharging current, idis. A typical current-time trace recorded at 0.10 V vs SCE in the course of LB transfer of CISOH is shown in Figure 3. The observed current can be simply expressed by62

,i

=# W d t

(5)

where uM is the excess charge density on the gold surface immersed in the electrolyte solution (before the LB transfer) and W d t is the rate of surface emersion (in cm2/s) during the LB transfer. Since the latter is a n adjustable parameter, the observed current gives the double-layer charge density. (In practice, uM was obtained by integration of the current transient above background.) As expected, idis is linearly proportional to the rate of withdrawal of the electrode surface, dA/dt (see Figure 41, as long as the LB transfer rate is less then 50 m d m i n . (60) Lipkowski,J.; Stolberg, L. In Adsorption of Molecules at Metal Electrodes, Frontiers in Electrochemistry Series; Lipkowski,J., Ross, P. N., Eds.; VCH Publishers, Inc.: New York, 1992; Vol. I, Chapter 4, pp 171-238. (61) Schultze, J. W.; Vetter, K J. J.Electroanal. Chem. 1973,44,63. (62) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; John Wiley & Sons: New York, 1980;pp 145-157.

Krysinski et al.

4290 Langmuir, Vol. 10,No. 11, 1994 S.D.

a

a

00

%OH

l##

J

120flA

30 25

--

0

-15 --

0 0

.H

5.0 u.u

t I

0

-0.2

1

I 1

1 1

0.1

0.4

0.7

,-.. 1

I

1.3

E [VJvs. SCE

Figure 5. Plots of charge densities vs substrate potential recorded in the course of the LB transfer experiments: open circles, total charge recorded with ClsSWCl8OH (70:30mol %I monolayers;squares, charge recorded with ClsOH monolayers corresponding to the double-layer charge density at the gold/ solution (0.05M HC101)interface; closed circles, charge due to ClsSH chemisorption and the apparent ClsSH partial charge number. The former was obtained as a difference between the total charge and the double-layer charge. The bar indicates the average standard deviation, for all the measurements, of 1.25 NC/cm2.

0

20

2

-10 -0.5

I

I

10

20

I

30

40

50

60

LB Transfer Rate [mdminl Figure 4. Plot of current vs transfer rate recorded during LB transfer of ClsOH monolayers from 0.05 M HC104 onto 0.2 cm2 gold-coated glass substrates at 0.20 V vs SCE.

A negative deviation observed a t 50 m d m i n marks a breakdown of this experiment, as some electrolyte is entrapped on the gold surface during LB transfer.63 This results in an incomplete removal of the double-layer charge. The plot of the double-layer charge density vs potential obtained in these experiments is shown in Figure 5 (closed squares). It is possible to approximate these data by a straight line. The crossing point of the linear regression line with the potential axis gives an estimate of PZC of 0.33 V vs SCE. This value is in a good agreement with the literature data obtained for gold(ll1) surfaces in nonadsorbing electrolyte^.^^^^^ This is perhaps unexpected but not surprising, since vapor-deposited gold films on mica, glass, and other substrates have been known to Our vaporexhibit predominantly (111) (63) Srinivasan, M.P.;Higgins, B. G.; Strove, P.; Kowel, s.T. Thin Solid Films 1988, 159, 191. (64)Lecoeur, J.;Andro, J.; Parsons, R. Surf. Sci. 1982, 114, 320. (65) Hamelin, A.In Modern Aspects ofElectrochemistry; Conway, B. E., White, R. E., Bockris, J. OM., Eds.; Plenum Press: New York, 1985; Vol. 16, Chapter 1, pp 1-101. (66) Kolb, D.M.;Schneider, J. Electrochim. Acta 1986, 31, 929. (67) Zei, M.S.;Naka, Y.; Lehmpfuhl, G.;Kolb, D. M. J . Electroanal. Chem. 1983,150, 201.

deposited films are likely to be microcrystalline because they were deposited a t rather fast rates (ca. 10 k s ) a t room temperature on chromium-coated glass slides. STM images ofour gold substrates show, indeed, about 10times smaller crystal domains (ca. 20 nm in diameter) compared to those found in Au films on mica.34 It is worth pointing out that the uMvalues were impossible to obtain a t E > 0.4 V in the presence of even small quantities of C1- ions in the electrolyte solution. Chloride adsorption in this potential range apparently prevents close interactions of C&H with the gold surface during LB transfer required to observe the double-layer charge squeezing effect. Although special measures were taken to eliminate chloride ions from the system, it is, nevertheless, possible that the uMvalues a t E > PZC are low. This could have resulted in a positive error in our estimate of the PZC value. Charge density at potentials more positive than 0.6 V was essentially impossible to measure accurately due to a relatively high background current. In that potential range, we approximated the double-layer charge density by the linear regression line through the data a t more negative potentials as shown in Figure 5. The slope of the surface excess charge density vs potential plot in Figure 5 provides the integral capacitance value of ca. 14 pF/cm2. This value appears to be low compared to ca. 20pF/cm2that could be obtained over the same potential range from the literature data for single crystal gold(111)surfaces.65 Knowing that double-layer capacitance is very sensitive to the structure of the electrode surface, traces of impurities, and the type and concentration of the electrolyte, we believe that this discrepancy is not unusual, considering all the differences in the experimental conditions between our system and the reference systems in the l i t e r a t ~ r e .Overall ~~ then, our charge density data in Figure 5 are reasonably consistent with the literature values. This, together with the proportionality of with the LB transfer rate (Figure 4), assures us that the LB transfer of C180H results in a n approximately complete removal of the double-layer charge as gold substrates are emersed dry from the electrolyte solution. (68) Hallmark, V. N.; Chiang, S.;Rabolt, J. F.; Swalene, J. D.; Wilson, 1987, 59, 2879.

R. J. Phys. Rev. Lett.

Langmuir, Vol. 10, No. 11, 1994 4291

CI8SH Chemisorption on Gold

LB Transfer of Mixed ClsSWCisOH Monolayer under Controlled Potential Conditions. Consider now the results of the potentiostatic LB transfer experiments obtained with mixed Cd3-YC18OHmonolayers (see Figures 3 and 5). As expected on the basis ofthe discussion in the previous section, when these experiments are carried out a t E < PZC, a substantially higher current is observed. At constant potential, current due to ClsSH binding to gold is added to the double-layer discharging current. Naturally, a t the potentials positive of PZC, the two current components are of the opposite sign. Using the same procedure of data analysis as above, we obtained the total charge collected during the LB transfer experiments (see Figure 5 open circles). By subtracting from the latter series the values of the double layer charge density, we obtained the dependence of the charge transferred due to ClsSH binding to gold vs the potential of the gold substrate (Figure 5, closed circles). In order to obtain the apparent partial charge number (d), i.e. the fraction of electron charge transferred by each ClsSH molecule (assuming their 100% activity-see discussion below), we measured the geometric transfer ratio (the ratio of the surface area of the monolayer removed from the water surface to the geometric surface area of the gold substrate coated with the monolayer). Large (3 in. x 1in.) glass slides coated with gold on both sides were used in these experiments. A transfer ratio of 0.98 f0.06 was obtained by averaging data obtained with four different substrates. Knowing the transfer ratio and the mean molecular area ofthe molecules on the water surface a t the LB transfer pressure (20.1 f 0.3 AVmolecule), we obtained the apparent C&H partial charge number (see Figure 5, closed circles and the right-hand side ordinate). The latter exhibits a weak, roughly linear dependence on the potential of gold substrates increasing from 0.25 a t -0.3 V to 0.36 at 0.7 V vs SCE. The data a t potentials greater than 0.7 V are strongly affected by the formation of gold oxide species and by the direct electro-oxidation of ClsSH. Thiol derivatives are known to be electro-oxidized from a n adsorbed state on gold electrodes in this range of potential^.^^,'^ Oxidative desorption of propanethiol from gold yielding propanesulfinic acid was observed recently by Widrig and cow o r k e r ~ .Some ~ ~ additional experiments carried out in this potential range will be presented in a separate report. We consider now some mechanistic implications that can be derived from the data presented above. The net charge recorded in the potential range from -0.3 to 0.7 V (closed circles in Figure 5) reflects the partial electron transfer from thiol to gold as stated in eq 4 above. Thus we can say that under our experimental conditions ClsSH chemisorption leads to the formation of a weak coordination type bond with the surface gold atoms, involving a transfer of approximately 0.24 to 0.36 electron (depending on the potential of the gold substrate, and assuming that all ClsSH transferred chemisorb) per alkanethiol molecule. It becomes important to realize that this would leave approximately +0.3 charge on each thiol group as the ClsSH monolayer is transferred onto the gold surface. In view of the fact that the capacitance of the interface above water is essentially zero (see Scheme 1 and the related discussion above), a charge of this magnitude clearly cannot exist a t the interface under these conditions. Explicitly, the condition of negligible capacitance of the interface means that no charge can be stored on either the metal or the monolayer side of the interface. Since ~

(69)Koryta, J.;F’radac, J. J.Electroanal. Chem. 1968,17, 177. (70)Reynaud, J.A.;Malfoy, B.; Canesson, P. J.Electroanal. Chem. 1980,114,195.

no counterions can be transferred with the ClsSWCleOH monolayer to match this charge (perchlorate ions are the only anions in the subphase and their transfer as counterions is sterically impossible and inconsistent with the transfer ratio reported above), we postulate that a n equivalent quantity of protons is released from the monolayer as ClsSH molecules are transferred and bound to the gold substrate. This is the only way that the system can maintain charge neutrality of the ClsSH monolayer above the water surface. We note that the postulated proton dissociation is consistent with the expected weakening of the -S-H bond upon thiol binding t o gold. It is also consistent with several Raman,31”1IR,29and mass54 spectroscopic studies of self-assembled alkanethiol monolayer on gold showing absence of the thiol proton. In a related Raman investigation, absence of the thiol proton was also reported in chemisorption of 6-mercaptopurine on gold electrode^.^^ We should stress again that in our experiments the proton dissociation takes place primarily to maintain the required charge neutrality of the monolayer, and, thus, is a consequence of, rather than a prerequisite to, the thiol-gold bond formation. Its extent is linked directly to the magnitude of the partial charge number. The extent of proton dissociation during alkanethiol self-assembly experiments may be different or, indeed, complete due to the differences in the prevailing conditions of the interfacial capacitance, presence of traces of ionic species, etc. Recent ab initio geometry optimization of CH3S on Au(ll1) and Au(100) surfaces by Ulman and co-workers modeling alkanethiol chemisorption predicts -0.7 electron charge on sulfur atom bound to on-top sites and -0.4 charge for those bound to 3-fold hollow sites on g ~ l d ( l l l ) . ’ ~ Extent of Cl&H Chemisorptionupon LB Transfer of ClsSWCleOH Monolayers on Gold. Considering the mechanics ofthe potentiostatic LB experiments, it is only fair to ask whether all C18SH molecules transferred from the water surface bind to the gold surface on the time scale allowed by the LB transfer rate (530s). In other words: how well do we approximate the equilibrium conditions of self-assembly? To answer these questions, it is interesting to point out that LB transfer (under “reactive”conditions)can be considered as a 2-D equivalent of self-assembly. In the LB transfer, a one-dimensional “section” of a surface or a “line” of gold is exposed to a two-dimensional condensed system of the assembling molecules. That exposure time is admittedly only ca. 6 ps (for a 10 wide ”line” with the transfer rate of 10 m d m i n used in these experiments). However, there are two important characteristics of LB transfer that weight heavily toward acceleration of the “2-D self-assembly”. The first one is the concentration of ClsSH in a monolayer a t 20 mN/m which is effectively several orders of magnitude larger than those used typically in solution selfassembly experiments. The second, even more important factor, is the fact that in the LB process, the surface is exposed to a n already well-ordered array of ClsSH molecules. In the conventional self-assembly experiments carried out in 1mM ethanol solutions of octadecanethiol, Bain and co-workers observed close to the limiting contact angles on gold surfaces in less than 1min and proposed that the self-assembly process proceeds in two distinct phases.14 This was recently confirmed by near-edge XAFS

a

(71)Garrell, R. L.,private communication, 1994. (72) Taniguchi, I.; Higo, N.; Umekita, K.; Yasukouchi, K. J. Electroanal. Chem. 1986,206,341. (73)Sellers, H.; Ulman, A.; Shnidman, Y.; Eilers, J. E. J.Am. Chem. SOC.1993,115,9389.

4292 Langmuir, Vol. 10,No.11, 1994 500

400

1

e

t

Krysinski et al. 0

Table 1. Dependence of the C&H Partial Charge Number and the Cl~SWClsOHLB Transfer Ratio on the Roughness of Gold Substrates

120

v-

Qo*dep o*de(n)/ LB transfer QLB? substrate type fiC/cmZ%,dl) ratio &/cmZ 1. Aulglassd 352 1.00 0.98f 0.06 15.7f 0.8 0.29i 0.02

2.Au foild

100

t

385 unroughened 3. Au foil' 428 roughened 4. Au foil' 619 roughened 5. Au foil' 1150 roughened

5 0

0.00

10

0

20

30

40

50

60

LB Transfer Rate [mdmin] Figure 6. Plots of current and the corresponding chargedensity vs transfer rate recorded during LB transfer of ClsSWClsOH (70:30mol %) monolayers from 0.05 M HClO4 at 0.0 Vvs SCE.

I S.D.

1.09

0.99i 0.04 17.5f 0.8 0.32i 0.02

1.22

1.13f 0.08 22.0f 1.5 0.34f 0.03

1.76

1.35f 0.08 29.02c 1.5 0.39 0.03

3.27

1.38f 0.08 28.0 f 1.5 0.36i 0.03

a Charge due to electroreduction of the surface gold oxide film produced in an anodic voltammetric scan extended to +1.3 V vs SCE in 0.05 M HC104. Charge due to C&H chemisorption obtained during potentiostatic LB experiments at +0.30 Vvs SCE. Apparent value of ClsSH partial charge number. Data in this row are the averages of four experiments done with different substrates. e Roughening of the gold plate substrates involved voltammetric cycling between -0.3 V and 1.2V vs SCE in 0.1 M

~ ~ 1 . 7 5

15

s p e c t r o ~ c o p y .The ~ ~ arguments presented so far would suggest that all C18SH molecules could bind to gold. To examine this issue further, we have carried out three additional sets of experiments. Figure 6 shows the dependence of current and charge recorded during the LB transfers of C18SWClsOH monolayers as a function of the substrate withdrawal rate. Linearity ofthis plot for rates below 50 m d m i n (where one begins to transfer molecules on a thin film of water rather than directly onto gold surface), shows that the time scale of the experiment does not affect the efficiency of Cl8SH bonding to gold. In the second set of experiments, we measured the total charge collected during LB transfer a t 0.0 V vs SCE as a function of Cl8SH mole fraction in mixed Cl~SWCl8OHmonolayers. The linearity of the observed charge, shown in Figure 7, proves that all or a constant fraction of C&H molecules transferred bind to gold. In the third set of experiments, we have examined the effect of roughness of the gold substrateson the LB transfer ratio and on the charge due to C&H binding to gold. The latter issue directly addresses the question whether all C18SH molecules transferred onto gold in the LB experiments bind to its surface. Self-assemblyof octadecanethiol on gold(ll1) surfaces is known to result in surface monolayers exhibiting a ca. 30" tilt and a specific registry with respect to the surface gold atoms.27 A different monolayer structure may be obtained in the case of the LB deposition experiments since the ClsSH molecules are

transferred in a compressed state where their projected area per molecule on the water surface is only 20.1 molecule. The following potentiostatic LB transfer experiments were carried out at 1.5cm2thin gold foil substrates. Their original surface was smooth but not optically flat. To further increase the extent of surface roughness, we followed a procedure involving repetitive voltammetric cycling of the gold substrates between +1.2 V and -0.3 V vs SCE in 0.1 M KC1 solution.75 (The same procedure is used to produce substrates for surface enhanced Raman spectroscopic studies.) Roughened and unroughened substrates were then pretreated chemically and electrochemically using the same procedures as those used for the vapor-deposited gold on glass substrates. Subsequently, the relative extent of surface roughness was determined by integrating current due to reduction of gold oxide formed anodically by extending voltammetric scan to 1.3V vs SCE in 0.05 M HC104. The transfer ratio and the potentiostatic C1&3WClsOH LB transfer experiments were done a t 0.30 V as described above. The substrate potential was selected to coincidewith the experimentally obtained value of the PZC to avoid double-layer charge corrections. The results are collected in Table 1. The gold oxide charge (Qodae) in the second column illustrates a n increase of true surface roughness of the gold substrates. The third column lists surface roughness relative to Adglass. As expected, the LB transfer ratios and the charge due to C&H chemisorption (QLB)increase with the substrate roughness. This is consistent with the reactive nature of the LB transfer of the ClsSH/Cl@H monolayers. However, comparison of the transfer ratios and the QLB values for the two most roughened substrates (entries 4 and 5 ) indicates that monolayer transfer on excessively rough surfaces no longer follows the microscopic contour of the surface. The variation of the apparent partial charge number ( 6 )with surface roughness is the key feature of the data in Table 1. It is clear that 6 is higher on rough surfaces compared to relatively smooth Adglass substrates for which roughness factor is known to be ca. 1.2.60 In view of the finite precision of these measurements, we believe that there are no significant differences between the 6 values for the entries 3, 4, and 5 in Table 1. Therefore, we will use their average of0.36 & 0.03. Since this average

(74)Hahmer, G.; Woll, C.; Bach, M.; Grunze, M. Langmuir 1993,9, 1955.

(75)Gao, P.; Gosztola, D.; Leung, Electroanal. Chem. 1987,233,211.

0.0

r !

I

0

20

I

I

40

60

I

I

80

Mol % of CWSH Figure 7. Plot of the total charge collected during LB transfer (10 d m i n ) of ClsSWC18OH monolayers from 0.05 M HClO4 on gold substrates at 0.0 V vs SCE as a function of ClsSH mole fraction. The average standard deviation bar is 1.0 pC/cm2.

L.-W. H.; Weaver, M. J. J.

CI&H Chemisorption on Gold

Langmuir, Vol. 10,No. 11, 1994 4293

99 n cc no longer depends on the increasing roughness of the gold S.D. surface, it can be considered to be the true partial charge number of C18SH chemisorption at 0.30 V. This value of 20 6 is 24%higher than the apparent partial charge number obtained on the vapor deposited Adglass substrates. Since the 6 values were calculated under the assumption of 100% E 0.45 fi reactivity of the transferred Cl8SH molecules, the lower value of 6 on Adglass substrates indicates that only ca. 0 81%of C&H molecules actually chemisorb on gold upon 0.40 2 16 LB transfer. The true partial charge number can, therefore, be obtained by multiplying the values in Figure B e e e B 5 by 1.24. Thus, they increase from 0.3 at -0.3 V to 0.45 0 . 3 12 14 5 q 14 a at 0.7 V vs SCE. 0 2 4 6 8 1 0 : This important conclusion requires some discussion. PH Since it is apparent that the higher surface roughness Figure 8. Plot of the charge-transferred due to ClsSH leads to a n increase of the fraction of the thiol molecules chemisorption on gold, and the ClsSH partial charge number capable of chemisorption, we postulate that the LB transfer (correctedfor incomplete chemisorption)as a function of pH of the subphase in the LB transfer experiments. Data were of a monolayer a t high pressures is followed by pressure obtained at -0.20 V vs SCE in HC104, NaC104, and KOH relaxation and partial respreading of the monolayer before electrolytes. These data represent the difference between the chemisorption of Cl8SH. These events may, indeed, be total charge and the double-layer charge (see Figure 5). The necessary to allow the molecules to adopt appropriate bar indicates the average standard deviation of 0.80 pClcm2. surface orientation andor registry with respect to the surface gold atoms. Such conditions for chemisorption Schneider and B ~ t t r ythat , ~ ~Porter’s interpretation is would be met more readily on substrates with certain burdened with a positive error since a significant portion minimum microscopic surface roughness. Conversely, of the voltammetric “desorption”in Porter’s experiments these conditions may be lacking on surfaces composed, is undoubtedly due to a double-layer charge that has to partially, of atomically flat terraces over which monolayer flow to the interface, since desorption of a surface relaxation and respreading may be difficult. It appears monolayer results in a large increase of its capacitance. that the vapor-deposited Adglass substrates represent Nevertheless, we attempted to carry out our measurethe latter type of surface. The dominance of the 111 ments in more alkaline electrolytes (see data in Figure 8). domains on vapor-deposited Au films on glass and mica Unfortunately, our experiments could not be done a t pH’s surfaces is well-known in the l i t e r a t ~ r e (see ~ ~ ,also ~ ~the , ~ ~ above 13 due to partial solubility of dissociated octadediscussion of the experimentally determined PZC value canethiol from Cl8SWCl8OHmonolayers on very alkaline above). subphases. However, considering the rate of the increase of the partial charge number in Figure 8 (corrected for The preceding discussion leads to a prediction that a incomplete chemisorption of ClsSH as described above), higher extent of CleSH chemisorption might also be we do not think that its magnitude a t pH 14 would be observed when C18SWCl8OH LB films are transferred a t much above its largest value obtained a t pH 13. It is a lower surface pressure. Under such conditions, surface interesting to point out that the observed small increase monolayers are in a slightly more expanded state; for of the partial charge number at and above pH 12 coincides example, C&H mean molecular area increases from 20.1 with a n expected value of C18SH pKa77,78and could A21moleculeat 20 mN/m to 21.3 A21moleculeat 8 mN1m. therefore be explained by a n induction effect that allows We have repeated the potentiostatic LB transfer experifor a larger fraction of a n electron to be donated to gold. ments on Adglass substrates at 0.3 V at 8 mN1m and This is not inconsistent with our earlier postulate of the found, indeed, that a higher charge was transferred coupling via charge-neutrality requirement between the compared to 20 mN1m (19.2 f 1.2 pClcm2 was obtained partial electron transfer and the extent of proton disas an average of six runs compared to 15.7pClcm2in Table sociation. 1). In view of the fact that the LB transfer ratio remained Overall, the mechanism of C18SH chemisorption on gold unchanged, this yields a partial charge number of 0.37 f presented in this report is consistent (at least qualitatively) 0.02, a value essentially identical to that in Table 1 afier with Porter’s electrodesorption experiments. Our results correction for partial chemisorption. This result gives could also be compared with the detailed studies of strong support to our explanation of the partial chemichemisorption of sulfide ions by Schultze and c o - w ~ r k e r s ~ ~ sorption of C&H molecules in LB films transferred at and hydrosulfide ions carried out by Briceno and Chandhigh pressures. We should point out that the higher er.80 Schultze et al. reported the electrosorption valency surface pressure conditions are, nevertheless, more adof S2- on gold to be -2 in the potential range -0.6 to 0.4 vantageous as far as precision and reproducibility of the V vs SCE. The same value was obtained for HS- on gold results are concerned. by Briceno and Chander who also postulated a concurrent We move now to a discussion of a pH effect on alkanethiol dissociation of the hydrosulfide’s proton. The transfer of chemisorption. In recent reports, Porter and co-workers two electrons from sulfide to gold (observed in 1 M NaOH described a series of electrochemical experiments aimed solution) compared with the 0.3 to 0.45 electron transfer a t desorption of alkanethiols from gold e l e ~ t r o d e s . ~ O - ~ ~accompanying ClsSH chemisorption reflects reasonably Their experiments conducted in 0.5 M aqueous or ethanolic (76)Schneider, T.W.; Buttry, D. A. J.Am. Chem. SOC.1993,115, KOH solutions showed well-developed surface voltam12391. metric peaks a t negative potentials (e.g. -1.35 V vs Agl (77)Dalman, G.;Gorin, G. J. O g . Chem. 1961,26,4682. AgCl for octadecanethiol). The authors assigned the (78)Irving, R.J.; Nelander, L.; Wadso, I. Acta Chem. Scund. 1964, 18. 769. charge associated with these peaks to be due to breaking (79)Wierse, D.G.;Lohrengel, M. M.; Schultze, J. W. J . Electroanul. of the thiolate-gold bond. Its value was 1.0 electron per Chem. 1978,92,121. alkanethiol, a value substantially larger than that re(80) Briceno,A.; Chander,S. J.Appl. Electrochem. 1990,20,506 and 512. ported above. It is almost certain, as shown recently by

--

1

I

t

F 3

4294 Langmuir, Vol. 10,No.11, 1994 well the chemical differences between these two sulfur species. Proton dissociation (and not HZrelease) observed in the chemisorption of HS- 8o is fully consistent with our postulate of partial proton dissociation. Partial Charge Number and Electrosorption Valency. Chemisorption and physisorption at the electrode/ solution interface has been an area of considerable importance and a subject of extensive investigations in electrochemistry.81 The parameter describing the charge supplied to the electrode surface from the external circuit upon adsorption of a species is electrosorption valency61 or, accordingto a recent IUF'AC recommendation,82formal partial charge number, ya, a quantity defined as

where uM is the excess charge density at the electrode surface and rais the surface concentration ofthe adsorbed species. In general, adsorption of a species a t the electrode surface may result in a change of the excess charge density for two reasons: one stems from a n inevitable change (usually a decrease) of differential capacitance of the electrodeholution interface, and the other related to charge transfer between the adsorbing molecule and the electrodeS6l The IUPAC recommended term for the latter, which may be negligible (as in physisorption), is partial charge number.82 It is important to note that the formal partial charge number is the only quantity accessible experimentally. This is because chemisorption of molecules at the electrodeholution interface simultaneously causes both effects mentioned above. In other words, the partial charge number can be measured only when adsorption does not affect the structure of the doublelayer-a virtually impossible scenario.82 We believe that the LB transfer experiment carried out under potentiostatic conditions described in this report creates experimental conditions that allow, for the first time, a direct determination of thepartial charge number independently of the double-layer capacitance component. This is accomplished by allowing chemisorption to occur only at the triple-phase contact line as the electrode surface is emersed dry from the electrode solution (see Figure 1). Thus the structure and the capacitance of the electrode/ solution interface are never perturbed. The double-layer charge ofthe electrodeholution interface, as demonstrated above, can also be measured independently in a parallel experiment involving physisorbed species (Cl8OH). The measurements of the double-layer charge density by the LB transfer of Cl8OH monolayer, followed by the total charge measurements with mixed CleSWCl8OH monolayers, and a calculation ofthepartial charge number for Cl8SH adsorption on gold as a difference of the two constitute a n illustration of this approach. The only limitation of this approach, besides its precision which is admittedly not as high as impedance or chronocoulometric methods,60 is the requirement that the adsorbent be insoluble in water. This would require one to work, for example, with long alkyl chain derivatives of adsorbents of interest.

Conclusions We have developed a coulometric technique that combines Langmuir-Blodgett monolayer transfer with constant potential current measurements. This technique (81)See, for example, a series of recent reviews in Adsorption of Molecules at Metal Electrodes. Frontiers in Electrochemistry Series; Lipkowski, J.,Ross, P. N., Eds.;VCH Publishers, Inc.: NewYork, 1992; VOl. 1. (82) Trasatti, S.; Parsons, R. J.Electroanal. Chem. 1986,205, 359.

Krysinski et al.

can be used to measure charge associated with a transfer of a monolayer from the aidwater interface onto a conducting substrate under potentiostatic conditions. In cases of weak monolayerlsubstrate interactions (physisorption), this charge corresponds to the double-layer charge density on the substrateholution interface before the LB transfer. In cases of chemisorption, a sum of the double-layer charge and the charge transferred between the molecules in the monolayer and the substrate is obtained. The latter is illustrated in this report by LB transfer of C~~SWC18OH. In this case, the difference between the total charge and the double-layer charge, which can be obtained independently in the experiments involving weakly interacting C18OH, gives the partial charge number of the chemisorbed species, C&H. Thus this technique allows one, for the first time, to obtain partial charge number under conditions where chemisorption and the measurements of the related charge transfer are not perturbed by the otherwise inevitable changes of the double-layer capacitance that normally accompany adsorption at the electrodeholution interface. We have investigated the potential dependence of Cl8SH chemisorption on vapor-deposited gold films on glass. The C&H partial charge number increases from 0.30 at -0.30 V to 0.45 a t 0.70 V vs SCE. Thus C&H binding to gold involves formation of a coordination type bond whose character depends weakly on the potential of the gold substrate. Clearly, this mechanism does not involve gold oxidation, as postulated in the l i t e r a t ~ r e ,nor ~~~~~,~~ does it postulate reduction of proton or oxygen (eqs 1-3). I t is well-known that in the potential range investigated in this study the gold surface is in a reduced state.6O~6~-6~ To maintain electroneutrality of the surface monolayer upon its transfer on gold and above the water surface, we postulated that the charge transferred due to chemisorption is coupled to partial proton dissociation of the thiol groups. This is due to the negligible capacitance of the gold-Cl8SWCl8OH interface above water. The release of protons is also consistent with weakening of the S-H bond caused by thiol binding to gold but its extent is governed quantitatively by the magnitude of the partial charge number via the electroneutrality requirement. Investigations of the effect of the surfaces roughness of the gold substrates on the LB transfer ratios and the apparent partial charge number revealed that only approximately 81% of the Cl8SH chemisorb to gold when a compressed Ci&WC180H LB monolayer is transferred onto vapor-deposited Aufglass substrates. We postulate that inability ofthe Cl8SH molecules that were transferred in a n ordered array onto a flat gold substrate to assume appropriate orientation and/or registry with the surface gold atoms may be responsible for this incomplete chemisorption. Electrochemical roughening of the surface apparently alleviates this problem by allowing the molecules to respread and readjust their registry and tilt with respect to the substrate surface due to its microscopic roughness. Consistent with this postulate, we determined that LB transfer done a t a lower surface pressure also results in a n increase of the partial charge number and in the full extent of chemisorption. The values of the partial charge number given above include the correction for the incomplete chemisorption on Aufglass substrates.

Acknowledgment. We thank the National Science Foundation for supporting this research under Grant CHE-9108378. We also acknowledge and thank Dr. Steven Feldberg for a very insightful discussion of some double-layer problems related to this project.