Chemistry of sulfur-containing molecules on Au(111 ): thiophene, sulfur dioxide, and methanethiol adsorption

Chemistry of sulfur-containing molecules on Au(111 ): thiophene, sulfur dioxide, and methanethiol adsorption

Surface Science 505 (2002) 295–307 www.elsevier.com/locate/susc Chemistry of sulfur-containing molecules on Au(1 1 1): thiophene, sulfur dioxide, and...

255KB Sizes 0 Downloads 11 Views

Surface Science 505 (2002) 295–307 www.elsevier.com/locate/susc

Chemistry of sulfur-containing molecules on Au(1 1 1): thiophene, sulfur dioxide, and methanethiol adsorption Gang Liu, Jose A. Rodriguez *, Joseph Dvorak, Jan Hrbek *, Tomas Jirsak Department of Chemistry, Brookhaven National Laboratory, Upton, NY 11973-5000, USA Received 7 December 2001; accepted for publication 29 January 2002

Abstract The interactions of three sulfur-containing molecules (C4 H4 S, SO2 , CH3 SH) with a clean Au(1 1 1) surface have been studied with a combination of thermal desorption spectroscopy (TDS) and synchrotron-based high-resolution soft Xray photoelectron spectroscopy. The adsorption and reactivity of the three molecules on Au(1 1 1) are very different. Thiophene adsorbs molecularly on Au(1 1 1) at 100 K and desorbs completely below 330 K without further decomposition. In the submonolayer range, three different adsorption states for chemisorbed thiophene are identified in TDS. It is suggested that thiophene preferably adsorbs on the defect sites at the lowest exposure. After the defect sites are saturated, the change from a flat-lying geometry to a tilted adsorption configuration follows as the exposure increases. Sulfur dioxide also does not decompose on Au(1 1 1). For SO2 adsorption at 100 K, in addition to the multilayer desorption feature (130 K), only one distinct monolayer peak with a tail extending to higher temperature appears in TDS. The desorption temperature difference between the SO2 monolayer and multilayer is only 15 K, indicating a weak binding between SO2 and Au. For methanethiol adsorption on Au(1 1 1) at 100 K, three desorption states appear in the submonolayer range for the parent thiol. All of them appear below 300 K. The only desorption products at higher temperature are methane or methyl radicals (540 K), and dimethyl disulfide (470 K). Apart from the intact methyl thiol molecule, which exists at low temperatures ( 6 150 K), two inequivalent intermediate thiolates, are seen to coexist on Au(1 1 1) in the 150–400 K temperature range, with one of them existing as low as 100 K. Atomic sulfur is present on the surface from 200 to 950 K. Ó 2002 Elsevier Science B.V. All rights reserved. Keywords: Sulphur; Gold; Surface chemical reaction; Thermal desorption spectroscopy; Synchrotron radiation photoelectron spectroscopy

1. Introduction In general, the surface chemistry of S-containing molecules is an important subject in many

*

Corresponding authors. Fax: +1-631-344-5815. E-mail addresses: [email protected] (J.A. Rodriguez), [email protected] (J. Hrbek).

technologies [1–3]. Gold surfaces are frequently employed as templates to grow self-assembled monolayers (SAMs) [1]. In this respect, it is important to understand how different sulfurcontaining groups bond to gold [1]. In this work, we study the interactions of methanethiol, thiophene and sulfur dioxide with Au(1 1 1) by thermal desorption spectroscopy (TDS) and XPS, as a function of adsorbate coverage and heating

0039-6028/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 6 0 2 8 ( 0 2 ) 0 1 3 7 7 - 8

296

G. Liu et al. / Surface Science 505 (2002) 295–307

temperature. Methanethiol is the simplest anchor sulfur group of the CH3 (CH2 )n SH (1 6 n 6 36) systems that are used for preparing SAMs [1]. N-alkanethiol SAMs on solid surfaces have a lot of potential applications in the fabrication of biosensors, corrosion inhibition, catalysis, lubrication, and production of microelectronic devices [1,4]. In the past two decades, considerable theoretical [5–9] and experimental [10–16] efforts have been made to explore the molecules’ adsorption sites and geometry, desorption kinetics and thermal decomposition mechanisms. However, the interactions between the substrates and molecular headgroups in the monolayer interface, including molecular adsorption and dissociation sites, and the possible formation of S–S bonds, still remain uncertain and a matter of debate in the literature. In general, the consensus is that the bonding in the interface is via a Au-thiolate complex [15], but Xray diffraction [13,14] and SPM [16] studies suggest the possible dimerization of the sulfur headgroups by S–S bonding. The adsorption and structure of thiophene on metal surfaces have been of considerable interest to the surface science community, since thiophene can adopt either flat or tilted adsorption geometries via its p clouds or S lone-pair of electrons, respectively [2,17–19]. On a noble metal like Cu(1 1 1) [20], using normal incidence X-ray standing wavefield absorption (NIXSW) and Xray absorption fine structure (NEXAFS), it was found that thiophene undergoes a structural phase transition as a function of coverage, manifested by ) and an increase of the Cu–S bond length (0.2 A the orientation angle (18° with respect to the surface). On defect sites, thiophene molecules have a shorter Cu–S bond and orientation angles between 0° and 7°. Similarly, for thiophene adsorption on Ag(1 1 1) [21], NEXAFS and theoretical calculations show that both a compressed monolayer with inclined geometry and a relaxed monolayer species parallel to the surface could exist on the surface. It is not clear how thiophene (or its derivatives) will behave on a Au(1 1 1) surface since this substrate is much less reactive than Cu(1 1 1) or Ag(1 1 1). The behavior of SO2 on metal surfaces has been the subject of many studies in recent years [3,22,23]

due to the negative effects of this molecule in the atmosphere and catalysts poisoning [24]. For noble metal surfaces such as Cu(1 1 1) [25] and Cu(1 0 0) [25], both TDS and high-resolution XPS show that at 170 K, molecular SO2 adsorption dominates. Upon heating up to room temperature, SO2 was decomposed into SO, O, and S species. However, most recent experimental and theoretical works [26] show SO3 instead of SO as a stable intermediate on Cu(1 0 0). On other noble metal surfaces such as Ag(1 0 0) [27], Ag(1 1 0) [28], and Ag(1 1 1) [29], SO2 has been reported to adsorb and desorb molecularly without decomposition. Surprisingly, Au nanoparticles supported on TiO2 (1 1 0) readily dissociate SO2 [30], making necessary a detailed investigation of the SO2 /Au(1 1 1) system.

2. Experimental This work was carried out in two separate ultrahigh vacuum (UHV) chambers. The photoemission experiments were performed in a UHV chamber [31] at the U7A station in the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory. This chamber (with a base pressure of 5  1010 Torr) is fitted with a hemispherical electron energy analyzer with multichannel detection, optics for low-energy electron diffraction (LEED), a quadruple mass spectrometer (QMS), and a twin (Mg Ka and Al Ka) XPS source. The combined energy resolution in the synchrotron experiments was 0.3–0.4 eV. The binding energy (BE) values were determined with respect to the Fermi energy and by checking the position of the Au 4f7=2 peak at 84.0 eV. The TDS experiments were done in a second UHV chamber (with a base pressure of 5  1010 Torr) [32] which is equipped with a hemispherical electron energy analyzer with single channel detection, a twin (Mg Ka and Al Ka) XPS source and a residual gas analyzer (SRS-RGA). The residual gas analyzer was surrounded by a stainless steel jacket with a 10-mm aperture for the TDS measurements. All TDS spectra were collected at a heating rate of 3 K/s. For the TDS measurements, the crystal was positioned 1–2 mm away from the aperture to

G. Liu et al. / Surface Science 505 (2002) 295–307

prevent contributions to TDS signals from surfaces other than the sample. The Au(1 1 1) single crystal (Monocrystals Ltd.) was hold by one 0.5 mm diameter tantalum wire wrapped along a groove machined in the edge of the crystal. The sample could be cooled to as low as 100 K by thermal contact with a liquid nitrogen reservoir and resistively heated to 1200 K. The temperature was monitored by a type C thermocouple inserted in a hole at the sample edge. Prior to each experiment the Au(1 1 1) crystal was cleaned by repeated cycles of 1 keV Neþ ion bombardment followed by heating at 900 K until no impurities were detected by XPS. Very minor traces (<0.01 ML) were sometimes detected in the S 2p spectra with the high sensitivity of photoemission. High-purity (99: þ %) sulfur dioxide (Matheson), thiophene (Aldrich), and methanethiol (Aldrich) were further purified by several freepump-thaw cycles with liquid nitrogen prior to dosing. The gas dosing was performed by backfilling the chamber through a leak valve and all reported exposures are based on the ion gauge reading in Langmuir units (1 L ¼ 106 Torr s). Mass spectrometry indicated that the reactants did not decompose on the gas handling system or the wall of the UHV system.

3. Results 3.1. Adsorption of thiophene Fig. 1 displays a series of typical TDS spectra of thiophene adsorbed on Au(1 1 1) at 100 K as a function of increasing exposure. The 84 amu signal (a major cracking fragment of the thiophene molecule) was monitored. No other masses besides thiophene were found, indicating that no dissociation occurred. For a very small amount of thiophene exposure (0.004 L), only one desorption peak (labeled a1 ) with a peak maximum (Tmax ) at 297 K is found, and it is rapidly saturated with increasing thiophene exposure. It can be attributed to a defect adsorption state, as seen on Cu(1 1 1) [20]. By the exposure of 0.025 L, a second peak (a2 ) with Tmax at 255 K develops. The a2 desorp-

297

Fig. 1. TDS spectra for thiophene adsorption on Au(1 1 1) at 100 K as a function of exposure. The parent ion signal (84 amu) was monitored. The heating rate is 3 K/s.

tion peak shifts to lower temperatures with further thiophene exposures. Generally, the peak position shifting to lower temperatures with increasing exposures implies repulsive interactions between the adsorbed C4 H4 S molecules. This a2 desorption state is assigned to desorption from a chemisorbed monolayer on terraces of the surface. By the exposure of 1.8 L, a third additional peak (a3 ) appears, centered at 186 K. The a3 peak is not shifted much compared to the a2 peak, but it can be assigned to another desorption state, as discussed in the Section 4. At 2.4 L, a fourth additional desorption peak at 142 K appears, labeled b. The b peak is relatively sharp and does not saturate with further thiophene exposure. Such behavior is consistent with desorption from a physisorbed multilayer of thiophene. Very similar desorption temperatures for condensed multilayers are reported for thiophene adsorption on Cu(1 1 1) [20] and Ag(1 1 1) [21] surfaces. The overlapping desorption peaks in Fig. 1 make very difficult a complete analysis of the TDS data [33]. In this

298

G. Liu et al. / Surface Science 505 (2002) 295–307

situation approximated values to the desorption energies can be obtained using the Redhead method [33,34]. The estimated desorption energies, according to the Redhead method [34] assuming a first-order desorption kinetics and a pre-exponential factor of 1013 s1 , are 18, 16, and 11 kcal/mol for a1 , a2 , and a3 states, respectively. Fig. 2 displays the S 2p spectra for C4 H4 S adsorption at 100 K and subsequent heating. Only one doublet appears after an exposure as low as 0.2 L, with a pronounced S 2p3=2 peak at 163.4 eV. A very weak signal is seen around 161 eV. The doublet in the 165–163 eV region matches the BE found in previous studies for chemisorbed thiophene [35], while the weak features near 161 eV are at the position expected for atomic S on Au(1 1 1) [10,36]. The weak features probably arise from the decomposition of a negligible ( 6 0.01 ML) amount of thiophene on defect sites of the surface. Such features were more pronounced when thiophene was

adsorbed on polycrystalline Au, and disappeared upon adsorption of the molecule on Au(1 1 1) surfaces that were pre-annealed at 900 K for long periods of time [36]. Further C4 H4 S exposure (0.6 L) increases the S 2p intensity in Fig. 2, but the S 2p3=2 binding energies remain the same. Heating to 180 K does not drastically reduce the S 2p intensity or change the BE values. By 250 K, the S 2p signal disappeared. Since the TDS data (Fig. 1) show that a condensed C4 H4 S multilayer should desorb at 142 K, the photoemission results in Fig. 2 are within the submonolayer range. In addition, the S 2p3=2 BE for physisorbed C4 H4 S on Mo(1 1 0) [35] and Ag(1 1 1) [37] is reported at 164.5, and 164.8 eV, respectively. Fig. 3 shows the corresponding C 1s core level spectra for C4 H4 S adsorption on Au(1 1 1). A single C 1s BE appears at 284.0 eV, which is different from the reported C 1s values of 284.8, and 285.2 eV for multilayer C4 H4 S on Mo(1 1 0) [35], Ag(1 1 1) [37], respectively. Similar

Fig. 2. S 2p spectra (hm ¼ 380 eV) for the adsorption of C4 H4 S on Au(1 1 1). C4 H4 S was dosed at 100 K, and the surface progressively annealed to the indicated temperature.

Fig. 3. C 1s spectra (hm ¼ 380 eV) for the adsorption of C4 H4 S on Au(1 1 1). C4 H4 S was dosed at 100 K, and the surface progressively annealed to the indicated temperature.

G. Liu et al. / Surface Science 505 (2002) 295–307

299

to the behavior seen for the S 2p features, the C 1s signal essentially disappears after heating to 250 K. 3.2. Adsorption of sulfur dioxide Fig. 4 shows representative TDS spectra for SO2 adsorption on Au(1 1 1) at 100 K. The desorption process is exclusively molecular. At an exposure of 0.05 L, one desorption peak (a) with Tmax ¼ 142 K was observed. The a state shows a very slight shift (4 K) to lower temperature with the increasing dosing amount. Using a leading edge analysis [33], the desorption activation energy is 7.8 kcal/mol at low coverages. For medium and large coverages, a clear desorption tail extending toward higher temperatures is seen, probably due to attractive interactions between the adsorbed molecules or complex interactions in the desorption process [33]. With the increase of SO2 exposure, at 0.4 L, an additional desorption feature (b) developed with Tmax ¼ 130 K. The (b) peak is relFig. 5. S 2p spectra (hm ¼ 380 eV) for the adsorption of SO2 on Au(1 1 1). SO2 was dosed at 100 K, and the surface was progressively annealed to the indicated temperature.

atively sharp and does not saturate with further SO2 exposure. Thus, it can be assigned to the desorption feature of condensed multilayers of SO2 . Similar desorption temperature values for multilayer SO2 were reported on other metal surfaces [25,27–29]. Fig. 5 displays S 2p core level spectra for SO2 adsorption on Au(1 1 1) at 100 K followed by heating to the indicated temperatures. The spectrum at 100 K is a typical one for a physisorbed multilayer of SO2 with the S 2p3=2 peak at 167.4 eV. Heating to 120 K drastically reduces the multilayer feature. By 150 K, most of the S species have disappeared from the surface, consistent with the TDS results in Fig. 4. No significant atomic S signal is observed at 180 K. 3.3. Adsorption of methanethiol Fig. 4. TDS spectra for sulfur dioxide adsorption on Au(1 1 1) at 100 K as a function of exposure. The parent ion signal (64 amu) was monitored. The heating rate is 3 K/s.

Fig. 6 shows typical TDS spectra for methanethiol adsorption on a Au(1 1 1) surface after a

300

G. Liu et al. / Surface Science 505 (2002) 295–307

Fig. 6. TDS spectra for methanethiol adsorption on Au(1 1 1) at 100 K as a function of exposure. The parent ion signal (67 amu) was monitored. The heating rate is 3 K/s.

Fig. 7. Evolution of hydrogen (2 amu), methyl radical (15 amu) and methane (16 amu) during TDS for 1.5 L methanethiol exposure on Au(1 1 1) at 100 K. The heating rate is K/s.

series of exposures at 100 K. The largest cracking fragment (47 amu) of the methanethiol molecule was monitored. For a 0.01 L exposure, only one desorption peak (Tmax ¼ 200 K) appears, labeled a1 . As the exposure increases to 0.03 L, a second peak (a2 ) with Tmax ¼ 176 K develops [38]. At 0.3 L, a third additional desorption peak appears, labeled a3 with Tmax ¼ 139 K. All a1 , a2 , and a3 peak positions are not significantly shifted with the exposure. The sharp desorption peak (b) centered at 120 K, growing without saturation at exposures larger than 0.3 L, is clearly due to the condensed multilayer desorption. The desorption temperature for the multilayer peak is consistent with the values reported previously [39]. All a1 , a2 , and a3 peaks are saturable. The overlapping of these desorption peaks (as in the case of the C4 H4 S/Au (1 1 1) system) makes very difficult a complete analysis of the thermal desorption spectra [33]. Based on first-order kinetics and the Redhead method, the desorption activation energy for the

a1 , a2 , and a3 states are estimated to be 12, 11, and 8 kcal/mol, respectively. Fig. 7 shows the TDS profiles of products for a multilayer CH3 SH exposure on Au(1 1 1) at 100 K. Hydrogen (H2 ) desorbs completely below 250 K. Methane (16 amu) and the methyl radical (15 amu) desorb at higher temperatures with Tmax ¼ 540 K. The peaks for 15 amu and 16 amu have very similar shape and peak temperature, suggesting a single desorption product [40,41]. Assuming first-order kinetics, the activation energy for desorption at 540 K is 34 kcal/mol. However, at 540 K, with a hydrogen absence on Au(1 1 1), it is possible that the CH4 formation is due to H-atom abstraction from the mass spectrometer shield as a result of collisions of methyl radicals with the walls [40]. We also observed a very weak peak for desorption of dimethyl disulfide (CH3 S–SCH3 ) at 470 K (not shown), consistent well with TDS data (450–500 K) of

G. Liu et al. / Surface Science 505 (2002) 295–307

Fig. 8. S 2p spectra (hm ¼ 380 eV) for the adsorption of CH3 SH on Au(1 1 1). CH3 SH was dosed at 300 K, and the surface progressively annealed to the indicated temperature.

dimethyl disulfide on Au(1 1 1) [38]. No other hydrocarbons were detected in the present study. Fig. 8 displays S 2p core level photoemission spectra for methanethiol adsorption on Au(1 1 1) at 300 K. Only one doublet is seen, with the S 2p3=2 BE at 162.0 eV, implying the existence of one sulfur-containing species. The BE value of the S 2p features in the present work agrees well with a chemisorbed thiolate species on gold reported in the literature [10,42–44]. From 0.05 to 0.35 L, the S 2p intensity is gradually increased and saturated finally. Upon dosing of 1.35 L, the S 2p intensity does not increase, indicating that the surface is saturated at room temperature. Upon heating to 420 K, the S 2p intensity is drastically decreased, but another peak at 161 eV develops. The decrease of the S peaks is partly due to the desorption of dimethyl disulfide (2CH3 Sad ! CH3 S–SCH3g ), as seen in TDS. The BE of chemisorbed atomic sulfur on Au(1 1 1) was previously reported at 161 eV [ 10,36]. The results suggest that

301

Fig. 9. C 1s spectra (hm ¼ 380 eV) for the adsorption of CH3 SH on a Au(1 1 1) surface. CH3 SH was dosed at 300 K, and the surface was progressively annealed to the indicated temperature.

the atomic sulfur species is present after heating at higher temperatures. After additional heating to 600 K, the S signal decreased. It totally disappeared by heating above 900 K. Fig. 9 shows the corresponding C 1s signal for methanethiol molecules. Only one C 1s peak is found, with the BE located at 284.2 eV. Similar to the behavior of the S 2p features, the C 1s intensity raises with the CH3 SH exposure up to 0.35 L. Heating to 420 K significantly decreases the C 1s intensity. Upon heating to 520 K, there is no C 1s signal on the surface, suggesting complete S–C bond cleavage or desorption/recombination of the CH3 S groups. According to the TDS results, the only species that desorb at high temperature (>400 K) containing C are CH3 , CH4 and CH3 S–SCH3 . The left panel of Fig. 10 displays the changes in the S 2p spectra for 0.85 L CH3 SH adsorption on Au(1 1 1) at 100 K, and sequentially heating to high temperatures. Figures a and b in the right

302

G. Liu et al. / Surface Science 505 (2002) 295–307

Fig. 10. S 2p spectra (hm ¼ 380 eV) for the adsorption of CH3 SH on Au(1 1 1). CH3 SH was dosed at 100 K, and the surface progressively annealed to the indicated temperature.The spectra for the S 2p core level at 150 and 300 K were curve-fitted following the procedure described in Ref. [45]. The dots represent the raw data and the lines connecting them are the sum curves resulting from the fitting.

panel of Fig. 10 show the curve fitting [45] results for the sample heating to 150 and 300 K, respectively. At least two doublets are seen in the left panel at 100 K: physisorbed CH3 SH and chemisorbed CH3 S are clear. Heating to 150 K causes a rapid decrease in the intensity of the multilayer features. The fitting results of spectrum a (top right panel of Fig. 10) reveal the existence of three sulfur components at 150 K, with the S 2p3=2 BE values at 161.9, 163.5 and 166.1 eV, respectively. The doublet with the S 2p3=2 BE at 166.1 eV is probably from physisorbed CH3 SH on the surface. The S 2p3=2 peak at 161.9 eV, displayed at both 100 and 150 K, is attributed to the formation of a gold– thiolate bond [10], suggesting deprotonation via cleavage of the S–H bond at a temperature as low as 100 K. The peak at 163.5 eV could be assigned as an unbound thiol/disulfide [10], but is probably

related to a second thiolate in a different adsorption site [36] or to chemisorbed CH3 SH, as discussed in Section 4.3. The TDS spectra in Fig. 6 show that upon heating to 150 K, there exist two CH3 SH desorption states. After additional heating to 200 K, a shoulder centered at the high-BE side (161.0 eV) develops. In addition, the tail at the lower binding side disappears, implying the complete desorption of the physisorbed CH3 SH. The curve fitting [45] of spectrum b (bottom, right panel of Fig. 10) obtained upon heating to 300 K, shows again there are still three sulfur species on the surface, with the S 2p3=2 BE at 163.4, 162.1 and 161.0 eV, respectively. The peak at 161.0 eV is assigned as atomic sulfur on gold [10,36]. The S 2p3=2 features at 162.1 eV for adsorbed CH3 S are now dominant, whereas the S 2p3=2 peak at 163.4 eV has become relatively weak. The S signal from

G. Liu et al. / Surface Science 505 (2002) 295–307

the adsorbed CHx Sy compounds gradually decreases and the intensity of the atomic S peaks gradually increases with the raise of the temperature. By 600 K, only the atomic sulfur doublet exists on the surface. Sulfur was removed entirely from the surface by 950 K (not shown). The corresponding C 1s core level XPS spectra for multilayer CH3 SH adsorbed on Au(1 1 1) are shown in Fig. 11. For the condensed phase at 100 K, the C 1s BE is 285.5 eV, which is in good agreement with the C 1s values for multilayer CH3 SH on other metal surfaces such as Ni(1 1 1) [39], Pt(1 1 1) [46] and Mo(1 1 0) [47]. Heating to 150 K removes the multilayer CH3 SH feature and the C 1s BE shifts to a lower value (284.7 eV). This shift is due to the surface binding configuration difference between the unbound physisorbed thiol and chemisorbed methyl thiolate or relaxation effects in the multilayer [39]. The C 1s BE slightly shifts to lower BE and the C 1s intensity decreases as a function of heating. By 400 K, the

Fig. 11. C 1s spectra (hm ¼ 380 eV) for the adsorption of CH3 SH on Au(1 1 1). CH3 SH was dosed at 100 K, and the surface progressively annealed to the indicated temperature.

303

C 1s is located a 284.2 eV. Heating to 600 K completely removes the C 1s feature. No atomic C feature resulted from the C–S bond cleavage.

4. Discussion In industrial applications and academic studies, gold surfaces are frequently employed as templates to grow SAMs of organo sulfur molecules [1]. These molecules can contain sulfur in several types of chemical groups. Our results show important differences in the behavior of thiophene, sulfur dioxide and methanethiol on Au(1 1 1). 4.1. Thiophene adsorption The TDS and photoemission data show that thiophene does not dissociate on Au(1 1 1). Thiophene is an aromatic molecule with clouds of pelectrons above and below the four carbon atoms and a lone-pair of electrons on sulfur [20,49]. It is interesting to compare thiophene thermal desorption behavior on gold with that on other sp metal surfaces. The present TDS result, with three desorption states in the submonolayer range, is very similar to that of thiophene on Cu(1 1 1) [20]. For Cu(1 1 1) [20], on the defect sites, thiophene adsorbs with a shorter Cu–S bond at inclined angles between 0° and 7°. With the increasing coverage, the thiophene occupies the atop sites with the orientation of 26  5°. A compression layer occurs with angles of 26–44° with respect to the surface at higher coverages. In the case of thiophene adsorption on Ag(1 1 1) [21], no adsorption state from the defects was reported although there are obvious desorption features on the high-temperature desorption side. However, depending on the coverages, both flat lying and compressed configuration with tilted angles of 40–45° are proposed [21]. STM [48] shows that thiophene molecules adsorb preferentially at step edge defect sites, from the upper step edges at lowest coverage to the lower step edge at higher coverage. As the coverage increases, a phase transition from a planar to a tilted geometry was observed [48]. A theoretical investigation [49] predicts that the chemical interactions between thiophene and gold are weaker

304

G. Liu et al. / Surface Science 505 (2002) 295–307

than for thiophene and copper. This does not agree with the trends seen in our results and the results in Ref. [20]. Furthermore, in the submonolayer range, for both Au(1 1 1) and Cu(1 1 1) [20], three desorption states can be seen in TDS, and all a desorption temperatures for thiophene on Au(1 1 1) are almost identical to those on Cu(1 1 1) [20]. The theoretical predictions in Ref. [49] were based on a single-metal atom model, and must be revised to include cooperative bonding on extended Au(1 1 1) and Cu(1 1 1) surfaces. 4.2. Sulfur dioxide adsorption Au nanoparticles supported on TiO2 (1 1 0) readily dissociate SO2 at room temperature [30]. In contrast, SO2 adsorbs and desorbs molecularly on Au(1 1 1) (Figs. 4 and 5). Previously, only on Ag surfaces, it was reported that SO2 does not decompose [21]. On transition metals, full dissociation of SO2 is observed [50]. In Fig. 4, the monolayer desorption temperature of SO2 is only 15 K higher than that of the mutilayer, indicating a SO2 –Au(1 1 1) bond slightly stronger than the van der Waals interactions between molecules within the multilayer. On a Ag(1 1 1) surface, the desorption temperature difference for the monolayer and that of multilayer is about 50 K [29]. In addition, three submonolayer desorption peaks were seen in TDS for SO2 adsorption on Ag(1 1 1) [29]: a1 (204 K), a2 (176 K), a3 (155 K) peaks were assigned as the desorption from the defect sites, terraces, and compressed layer, respectively. It was suggested that in a dense submonolayer of SO2 on Ag(1 1 1), the interplay between the SO2 – SO2 repulsive interactions and SO2 -Ag attractive interactions results in a less strongly bound SO2 with a low-activation energy for desorption. In the present study, only one desorption peak with a tail extending to higher temperature was observed for SO2 on Au(1 1 1) in the monolayer range, possibly due to the extremely weak BE of Au–SO2 compared to Ag–SO2 . Sulfur dioxide decomposes on copper surfaces at temperatures above 200 K with SO or SO3 as intermediates [26,51]. The different behavior of SO2 on the noble metals can be rationalized in terms of differences in the electronic properties of Au, Ag, and Cu as we will see below.

The nature of SO2 binding with metal surfaces is very complicated, with the possibility of binding via the S atom, the two oxygen atoms and a combination of S and O [26]. The first theoretical study of SO2 with an extended metal surface, Cu(1 0 0) [26], shows that at low coverages, SO2 adsorbs preferentially with its molecular plane parallel to the surface (g3 -S, O, O bonding). As the coverage increases, SO2 adopts g2 -O, O or g2 -S, O bonding configurations to minimize adsorbate– adsorbate repulsions. In the bond between SO2 and metals, a transfer of electrons from the metal into the LUMO of SO2 plays a dominant role in the bonding energy of the molecule [26,52]. This leads to a weakening of the S–O bonds, since the LUMO of SO2 is S–O antibonding. Copper is able to transfer electrons into the LUMO of SO2 [26]. In the case of Ag and Au, the valence d levels appear deeper in energy than in Cu [53]. This leads to poor electron donation into the LUMO of SO2 , and the molecule chemisorbs weakly and does not dissociate on extended Au or Ag surfaces. Gold nanoparticles supported on TiO2 (1 1 0) could become chemically active due to interactions with the oxide substrate that allow them a better hybridization with the LUMO of SO2 [30]. 4.3. Methanethiol adsorption Our TDS and photoemission studies indicate that CH3 SH deprotonates on Au(1 1 1) to form adsorbed CH3 S. These results are consistent with recent sophisticated theoretical calculations [7], but different from early experimental studies [38], where the desorption process of methanethiol on Au(1 1 1) is reported to be exclusively molecular. Using ab initio slab calculations based on DFT, Andreoni and coworkers reported [7] that for methanethiol adsorption on Au(1 1 1), the SCH3 thiolate due to S–H bond cleavage is thermodynamically favored over pure thiol adsorption by 3 kcal/ mol, and both thiol and thiolate could coexist on the surface, as seen in the data of Figs. 7 and 10. Deprotonation and formation of RS groups is the normal reaction pathway for the adsorption of thiols on gold [7,11]. Two reasons can lead to the discrepancy between our results and the lack of dissociation reported in Ref. [38] for CH3 SH on

G. Liu et al. / Surface Science 505 (2002) 295–307

Au(1 1 1). In the previous work, the authors stopped the temperature ramp in their TDS spectra at temperatures (<450 K) well below those seen in Fig. 7 for evolution of the decomposition products of CH3 SH. In addition, a weak but clear signal in the S 2p XPS spectra, after heating to 300 K (Fig. 2 in Ref. [38]), was neglected. This is the kind of signal that one could expect if the synchrotron data in Fig. 10 were taken with a commercial XPS source (much less sensitivity [50,54]). Surface defects could certainly facilitate the dissociation of CH3 SH on Au(1 1 1) [3]. Woodruff and co-workers [55,56] have studied CH3 SH adsorption on Cu(1 1 1) using X-ray standing wave (XSW) and high-resolution XPS. At 140 K, four distinct S-containing co-adsorbates were present on the surface, including intact CH3 SH, two thiolate intermediates and atomic S. Reconstruction of the Cu(1 1 1) surface with temperature is a key to induce different local adsorption geometries for two thiolate species. Relative to the BE of atomic S species, the BEs of the low temperature and high-temperature thiolates are 0.55 and 1.46 eV, respectively. In the present study, the equivalent values for two S species on Au(1 1 1) are 0.9 and 2.5 eV, respectively. Similarly, on Ni(1 1 1), the corresponding values are reported at 1.17 and 1.75 eV [39]. Based on similar behavior for CH3 SH on Au(1 1 1), Cu(1 1 1) [55,56] and Ni(1 1 1) [39], it is possible that two types of thiolates are formed on Au(1 1 1) in the temperature range of 150–400 K. It is interesting to see that at room temperature, only one CH3 S species is observed for methanethiol adsorption, whereas at 100 K, two CH3 S species are present after heating to 300 K. The difference is probably related to a more crowded adsorption arrangement for CH3 S from 100 to 300 K. Previous works indicate that the S 2p core levels are very sensitive to changes in the adsorption site of sulfur on a metal [36,57]. For example, a change in the adsorption site from a-top to hollow can induce a shift of 1–1:5 eV in the position of the S 2p levels [57]. Thus, in the surface of Fig. 10, one could easily have thiolate species on two different adsorption sites. One important issue regarding the SAM studies is whether the thiol molecules are adsorbed in the

305

form of thiolates or disulfides. Most recent DFT slab calculations by Vargas et al. [5] show that for the molecular and dissociative adsorption of dimethyl disulfide on Au(1 1 1) as a function of coverage, the dissociative adsorption of dimethyl disulfide is favorable over the adsorbed molecular state, e.g., with a thermodynamic gain of 13 kcal/ mol at h ¼ 1=3. Similarly, combined DFT slab calculations and high-resolution electron-energyloss spectroscopy (HREELS) experimental results [6] show the methylthiolate adsorption is more stable than the dimethyl disulfide adsorption. Kluth et al. [58] studied octadecanethiol on Au(1 1 1) using HREELS and suggested that the molecules are adsorbed as thiolate at room temperature and the dimerization is thermally activated at 375 K. In our TDS spectra, since only a very small amount of dimethyl disulfide is observed, CH3 S–SCH3 is likely formed by the recombination of some adsorbed thiolate species near the desorption temperature (420–520 K). It is clear that in Fig. 6, three different desorption states exist on the surface in the submonolayer range at temperature below 200 K. These states could be related to different adsorbate phases in terms of different molecular-substrate bonding. There are other possibilities, such as recombination of a small amount of thiolate with hydrogen to produce and desorb CH3 SH [39]. Generally, the adsorption configuration of n-alkanethiol on Au is a very complex issue [59,60]. SPM, TDS, LEED and AES [11] results revealed that for hexanethiol adsorption on Au(1 1 1), at room temperature, there are at least four different phases during the growth of the monolayer, including a 2D liquid phase (low coverage), striped phase, disordered phase, and c(4  2) phase (high coverage). At low coverages, the thiolate molecules adsorb as flatlying species, while they adsorb with standing-up configurations in the latter two phases. From the above hexanethiol studies, one can expect coverage-dependent changes of the desorption states for methanethiol adsorption on Au(1 1 1) at 100 K. Indeed, our TDS results indicate the existence of three-submonolayer desorption states for this system. Based on the TDS and photoemission results, we propose the following mechanism for the

306

G. Liu et al. / Surface Science 505 (2002) 295–307

adsorption and desorption process of CH3 SH on Au(1 1 1): 100 K

CH3 SHðgÞ ƒƒ ƒ! CH3 SHðadÞ 100 K

CH3 SHðadÞ ƒƒƒ! CH3 SðadÞ þ HðadÞ 6 200 K

2HðadÞ ƒƒƒƒ! H2 ðgÞ 140–200 K

CH3 SHðadÞ ƒƒƒƒƒ! CH3 SHðgÞ 140–200 K

CH3 SðadÞ þ HðadÞ ƒƒƒƒƒ! CH3 SHðgÞ

ð1Þ ð2Þ ð3Þ ð4Þ ð5Þ

The high-desorption temperature peak (540 K) for CH3 radicals is exclusively due to the C–S bond scission during heating, with atomic sulfur left on the surface. These radicals directly desorb from the surface without the formation of other hydrocarbons. A trace of dimethyl disulfide species seen in TDS spectra is probably formed during the desorption process: 480–580 K

CH3 SðadÞ ƒƒƒƒƒ! CH3 ðgÞ þ SðadÞ 420–520 K

2CH3 SðadÞ ƒƒƒƒƒ! ðCH3 SÞ2 ðgÞ

states exist apart from the multilayer feature in TDS. In addition, desorption of CH4 , CH3 , (CH3 S)2 and H2 is seen upon CH3 SH adsorption. The high-desorption temperature for both methane and methyl radicals is up to 540 K, and 470 K for dimethyl disulfide. In addition, SXPS data provide evidence that there are four distinct Scontaining adsorbates on the surface, including intact CH3 SH ( 6 150 K), two thiolate species (150–400 K, one as low as 100 K), and atomic sulfur (200–950 K). Atomic sulfur is the only adsorbate on the surface at high temperature (>600 K).

Acknowledgements The authors thank the support of the U.S. Department of Energy (DOE), Division of Chemical Sciences under the contract of DE-AC0298CH10886. The NSLS is supported by the Division of Materials and Chemical Sciences of DOE.

ð6Þ ð7Þ

5. Conclusions The adsorption and desorption of three sulfurcontaining molecules (C4 H4 S, SO2 , and CH3 SH) on Au(1 1 1) have been studied with TDS and XPS. Thiophene adsorbs molecularly on Au(1 1 1) at 100 K and desorb completely below 350 K without significant decomposition. In the submonolayer region, three adsorption states for thiophene are identified by TDS, probably resulting from a coverage-driven phase transition. For SO2 adsorption onto Au(1 1 1) at 100 K, only one monolayer feature with a tail extending to high temperature is seen in TDS. Sulfur dioxide does not decompose on the Au(1 1 1) surface as well. The temperature difference for SO2 adsorption on Au and the condensed multilayer is only 15 K, indicating an extremely weak interaction between SO2 and metallic Au. For methanethiol adsorption on Au(1 1 1) at 100 K, three molecular desorption

References [1] A. Ulman, Self-Assembled Monolayer of Thiols, Academic Press, New York, 1998, and references therein. [2] C.M. Friend, D.A. Chen, Polyhedron 16 (1997) 3165. [3] J.A. Rodriguez, J. Hrbek, Acc. Chem. Res. 32 (1999) 719. [4] J.D. Swalen, D.L. Allara, J.D. Andrade, E.A. Chandross, S. Garoff, J. Israelachvili, T.J. McCarthy, R. Murray, R.F. Pease, J.F. Rabolt, K.J. Wynne, H. Yu, Langmuir 3 (1987) 932. [5] M.C. Vargas, P. Giannozzi, A. Selloni, G. Scoles, J. Phys. Chem. B 105 (2001) 9509. [6] T. Hayashi, Y. Morikawa, H. Nozoye, J. Chem. Phys. 114 (2001) 7615. [7] H. Gr€ onbeck, A. Curioni, W. Andreoni, J. Am. Chem. Soc. 122 (2000) 3839. [8] H. H€akkinen, R.N. Barnett, U. Landman, Phys. Rev. Lett. 82 (1999) 3264. [9] H. Sellers, A. Ulman, Y. Shnidman, J.E. Eilers, J. Am. Chem. Soc. 115 (1993) 9389. [10] D.G. Castner, K. Hinds, D.W. Grainger, Langmuir 12 (1996) 5083. [11] H. Kondoh, C. Kodama, H. Sumida, H. Nozoye, J. Chem. Phys. 111 (1999) 1175. [12] D.J. Lavrich, S.M. Wetterer, S.L. Bernasek, G. Scoles, J. Phys. Chem. B 102 (1998) 3456. [13] P. Fenter, A. Eberhardt, P. Eisenberger, Science 266 (1994) 1216.

G. Liu et al. / Surface Science 505 (2002) 295–307 [14] N. Camillone III, C.E.D. Chidsey, J. Li, G. Scoles, J. Chem. Phys. 97 (1993) 3503. [15] H.A. Biebuyck, G.M. Whitesides, Langmuir 9 (1993) 1766. [16] L.-J. Wan, Y. Hara, H. Noda, M. Osawa, J. Phys. Chem. B. 102 (1998) 5943. [17] J.P. Fulmer, F. Zaera, W.T. Tysoe, J. Phys. Chem. 92 (1988) 4147. [18] F. Zaera, E.B. Kollin, J.L. Gland, Surf. Sci. 184 (1987) 75. [19] J.T. Roberts, C.M. Friend, Surf. Sci. 186 (1987) 201. [20] P.K. Milligan, B. Murphy, D. Lennon, B.C.C. Cowie, M. Kadodwala, J. Phys. Chem. B 105 (2001) 140. [21] P. V€ aterlein, M. Schmelzer, J. Taborski, T. Krause, F. Viczian, M. B€ aler, R. Fink, E. Umbach, W. Wurth, Surf. Sci. 452 (2000) 20. [22] H. Sellers, E. Shustorovich, Surf. Sci. 356 (1996) 209. [23] J. Haase, J. Phys.: Condens. Matter 9 (1997) 3647. [24] A.C. Stern, R.W. Boubel, D.B. Turner, D.L. Fox, Fundamentals of Air Pollution, 2nd ed., Academic Press, Orlando, 1984. [25] M. Polcik, L. Wilde, J. Haase, B. Brena, D. Cocco, G. Comelli, G. Paolucci, Phys. Rev. B 53 (1996) 13720. [26] J.A. Rodriguez, J.M. Ricart, A. Clotet, F. Illas, J. Chem. Phys. 115 (2001) 454, and references therein. [27] M. H€ ofer, H. Stolz, H.-W. Wassmuth, Surf. Sci. 272 (1992) 342. [28] J. Ahner, A. Effendy, K. Vajen, H.-W. Wassmuth, Vacuum 41 (1990) 98. [29] Z.-J. Sun, S. Gravelle, R.S. Mackay, X.-Y. Zhu, J.M. White, J. Chem. Phys. 99 (1993) 10021. [30] J.A. Rodriguez, G. Liu, T. Jirsak, J. Hrbek, Z. Chang, J. Dvorak, A. Maiti, J. Am. Chem. Soc. 124 (2002), in press. [31] G. Liu, J.A. Rodriguez, J. Hrbek, J. Dvorak, C.H.F. Peden, J. Phys. Chem. B 105 (2001) 7762. [32] J.A. Rodriguez, T. Jirsak, M. Perez, L. Gonzalez, A. Maiti, J. Chem. Phys. 114 (2001) 4186. [33] A.M. de Jong, J.W. Niemantsverdriet, Surf. Sci. 233 (1990) 355. [34] P.A. Redhead, Vacuum 12 (1962) 302. [35] J.A. Rodriguez, J. Dvorak, T. Jirsak, Surf. Sci. 457 (2000) L413. [36] J.A. Rodriguez, J. Dvorak, T. Jirsak, J. Hrbek, Surf. Sci. 490 (2001) 315. [37] K.M. Baumg€ artner, M. Volmer-Uebing, J. Taborski, P. B€ auerle, E. Umbach, Ber. Bunsenges. Phys. Chem. 95 (1991) 1488.

307

[38] R.G. Nuzzo, B.R. Zegarski, L.H. Dubois, J. Am. Chem. Soc. 109 (1987) 733. [39] T.S. Rufael, D.R. Huntley, D.R. Mullins, J.L. Gland, J. Phys. Chem. 99 (1995) 11472. [40] A. Paul, B.E. Bent, J. Catal. 147 (1994) 264. [41] S.H. Kim, P.C. Stair, J. Phys. Chem. B 104 (2000) 3035. [42] F. Sun, D.G. Castner, G. Mao, P. McKeown, D.W. Grainger, J. Am. Chem. Soc. 118 (1996) 1856. [43] P.E. Laibinis, G.M. Whitesides, D.L. Allara, Y.-T. Tao, A.N. Parikh, R.G. Nuzzo, J. Am. Chem. Soc. 113 (1991) 7152. [44] W. Fabianowski, L.C. Coyle, B.A. Weber, R.D. Granata, D.G. Castner, A. Sadownik, S.L. Regen, Langmuir 5 (1989) 35. [45] J.A. Rodriguez, M. Kuhn, J. Hrbek, Chem. Phys. Lett. 251 (1996) 13. [46] T.S. Rufael, R.J. Koestner, E.B. Kollin, M. Salmeron, J.L. Gland, Surf. Sci. 297 (1993) 272. [47] B.C. Wiegand, P. Uvdal, C.M. Friend, Surf. Sci. 279 (1992) 105. [48] X. Chen, E.R. Frank, R.J. Hamers, J. Vac. Sci. Technol. B 14 (1996) 1136. [49] F. Elfeninat, C. Fredriksson, E. Sacher, A. Selmani, J. Chem. Phys. 102 (1995) 6135. [50] T. Jirsak, J.A. Rodriguez, J. Hrbek, Surf. Sci. 426 (1999) 319. [51] G.J. Jackson, S.M. Driver, D.P. Woodruff, N. Abrams, R.G. Jones, M.T. Butterfield, M.D. Crapper, B.C.C. Cowie, V. Formoso, Surf. Sci. 459 (2000) 231. [52] S. Sakaki, H. Sato, Y. Imai, K. Morokuma, K. Ohkubo, Inorg. Chem. 24 (1985) 4358. [53] Y. Baer, P.F. Heden, J. Hedman, M. Klassen, C. Nordling, K. Siegbahn, Solid State Commun. 8 (1970) 517. [54] J.A. Rodriguez, M. Kuhn, J. Phys. Chem. 99 (1995) 9567. [55] G.J. Jackson, D.P. Woodruff, R.G. Jones, N.K. Singh, A.S.Y. Chan, B.C.C. Cowie, V. Formoso, Phys. Rev. Lett. 84 (2000) 119. [56] M.S. Kariapper, G.F. Grom, G.J. Jackson, C.F. McConville, D.P. Woodruff, J. Phys. Condens. Matter 10 (1998) 8661. [57] Z. Yang, R. Wu, J.A. Rodriguez, Phys. Rev. B, in press. [58] G.J. Kluth, C. Carraro, R. Maboudian, Phys. Rev. B. 59 (1999) R10449. [59] G.E. Poirier, M.J. Tarlov, H.E. Rushmeier, Langmuir 10 (1994) 3383. [60] G.E. Poirier, E.D. Pylant, Science 272 (1996) 1145.