Chemical-shift X-ray standing wavefield determination of the local structure of methanethiolate phases on Ni($\text{1}\text{1}\text{1}$)

Abstract

By monitoring the X-ray absorption through the chemically-shifted components of the S 1s photoemission signal, normal-incidence X-ray standing wavefield absorption at the (1 1 1) and ($\text{1}\text{̄}$ 1 1) scatterer planes has been used to determine the local adsorption geometry of the two distinct methanethiolate (CH₃S–) species which occur on Ni(1 1 1) following exposure to methanethiol. The species which is favoured at low temperatures is found to occupy either mixed hollow or bridge sites on a non-reconstructed Ni(1 1 1) surface, whereas that seen at higher temperatures is shown to involve Ni surface layer reconstruction and the data are consistent with hollow site adsorption on a reduced density outermost Ni layer. The relative merits of alternative reconstruction models based on that which occurs due to methanethiolate adsorption on Cu(1 1 1), or the (5√3×2)rect. phase formed by atomic S on Ni(1 1 1), are discussed. Both of these models are based on local square or `pseudo-(1 0 0)' outermost Ni layers. Co-adsorbed atomic sulphur, to which the methanethiolate species decompose at higher temperatures, appears to occupy mainly fcc hollow sites at low temperatures, but is partially converted to the local geometry of the ordered reconstructed (5√3×2)rect.-S phase after higher temperature annealing.

Introduction

The interaction of alkane thiols, CH₃(CH₂)_nSH with surfaces has been studied extensively, mainly due to their ability to form self-assembled monolayers on many surfaces. While this self-assembly is usually regarded as arising from the interaction of the alkane chains, in cases where the interaction of the S head-group and the substrate is strong ordered phases commensurate with the substrate clearly imply that some aspects of this order are related to the S-substrate interaction rather than true self-assembly. In the case of metal substrates most work has been performed on the relatively chemically inert Au surface, but there have also been studies on Ag and Cu surfaces as well as on several transition metals. In order to concentrate on the role of the S-metal interaction, much of this work has been conducted on thiols with short alkane chains, of which the simplest is methanethiol, CH₃SH. Using a variety of spectroscopic methods it seems to be well established that interaction with such surfaces leads to deprotonation and the creation of an adsorbed thiolate such a methanethiolate, CH₃S–, even at temperatures as low as 100–150 K.

In the case of the interaction of methanethiol with Ni(1 1 1), high-resolution (synchrotron radiation) soft X-ray photoelectron spectroscopy (SXPS) of the S 2p level by the group at Oak Ridge National Laboratory (ORNL) [1] showed two distinct states with different chemical shifts of the photoelectron binding energy, both of which were attributed to methanethiolate species. One of these states was dominant at low temperatures (up to about 200 K) but transformed into the other as the temperature was raised. For convenience, we refer to these two states as the low temperature (LT) and high temperature (HT) thiolates, respectively. A third chemically-shifted component, dominant at higher temperatures was assigned to atomic sulphur. The LT and HT thiolates were originally assigned to species bonded in bridging and hollow sites respectively [1].

A closely related system which has been studied in detail by our group is the interaction of methanethiol with Cu(1 1 1). Following initial characterisation by ultraviolet photoelectron spectroscopy [2], investigations of the thiolate species produced on this surface by room temperature exposure to methanethiol were conducted using X-ray photoelectron diffraction (XPD) [3], near-edge and surface extended X-ray absorption fine structure (NEXAFS and SEXAFS [4]) and normal-incidence X-ray standing wavefield absorption (NIXSW) [5]. These showed that the thiolate bonds to the surface via the S atom with the S–C bond perpendicular to the surface, but the S–Cu bond angle inferred by both the SEXAFS and NIXSW could only be reconciled with the S atom occupying an enlarged hollow site, implying that there must be an adsorbate-induced reconstruction of the outermost Cu atom layer to one of lower atomic density. Subsequently, a S 2p SXPS study of the interaction of methanethiol with Cu(1 1 1) as a function of temperature [6] showed that, as on Ni(1 1 1), two distinct thiolate species also occur on this surface. By combining the chemical-state specificity of (S 1s) core level photoemission with the NIXSW technique, it was possible to show that the LT thiolate is consistent with adsorption on the unreconstructed Cu(1 1 1) surface, occupying either mixed hollow or bridge sites, while the HT thiolate was confirmed to occupy hollow sites on a reconstructed surface which was either incommensurate or had a large coincidence mesh [7]. Finally, a very recent scanning tunnelling microscopy (STM) investigation of this surface has shown that this reconstruction involves the formation of an outermost Cu layer which has a near-square unit mesh with sides of $\text{2.88}\text{A}\text{̊}\text{×2.95}$ Å [8] leading to the structural model of Fig. 1. A second thiolate phase seen in STM, which was metastable at room temperature, was identified as the LT species, and the arrangement of the molecules in this phase as seen in STM appears to favour bridge site occupation.

The objective of the present study was to apply the same chemical-shift NIXSW (CS-NIXSW) method used in the Cu(1 1 1)/CH₃SH study mentioned above to elucidate the local structure of the Ni(1 1 1)/CH₃SH system. Two partial structural studies of this system have been performed previously. In the first [9], a combination of XPD, NEXAFS, SEXAFS and conventional NISXW was applied to a surface prepared and maintained at temperatures around 155 K or less. The thiolate species studied had a S–C axis which was significantly tilted away from the surface normal with the S atoms occupying mixed hollow sites on an unreconstructed surface. On the basis of the parallel ORNL SXPS study it seems likely that this surface comprised mainly LT thiolate. More recently, the ORNL group have used a form of photoelectron diffraction (PhD), in which they resolve the chemically-shifted components of the S 2p photoemission, to investigate the local structure of the two co-adsorbed thiolate phases [10]. The azimuthal emission angle dependence of the S 2p signal was measured at a kinetic energy of ≈200 eV. The kinetic energy favours the role of backscattering and may thus provide adsorbate–substrate registry information in the same way as for the well-established method of scanned-energy mode PhD [11]. However, the extremely limited data set used in this study was much too small to ensure that the structure could be determined uniquely, and must also limit precision. Nevertheless, the results showed that the data from the LT thiolate could be reconciled with methanethiolate having a tilted S–C bond, the S atom occupying either the fcc or hcp (or both) hollow sites, while bridge site occupation was deemed to be `unlikely'. In the case of the HT thiolate, a comparison with data from the ordered (5√3×2)rect. (or $\text{2}\text{0}\text{−5}\text{10}$ in the matrix notation using the unit mesh with the acute angle) reconstruction formed by atomic sulphur led them to propose that the local geometries in these two cases are similar.

In studies of atomic sulphur adsorption on Ni(1 1 1), the best-known phase is Ni(1 1 1)(5√3×2)rect.-S, and its structure has been the subject of some controversy. The core ingredient of the favoured model, originally based on the low energy electron diffraction pattern [12], was that the Ni(1 1 1) surface reconstructs to adopt a surface layer very similar to that on the Ni(1 0 0)c(2×2)-S phase in which a square array of Ni atoms has S atoms in alternate fourfold co-ordinated hollow sites. This type of `pseudo-(1 0 0)' reconstruction appears to occur in quite a number of adsorption structures on (1 1 1) and (1 1 0) fcc metal surfaces [13], and while there was some controversy over the interpretation of STM images from this phase [14], [15], a later surface X-ray diffraction investigation [16] showed that the reconstruction is based on this type of square array of Ni atoms, but with lateral distortions similar to the `clock' reconstructions of the Ni(1 0 0)/C and Ni(1 0 0)/N (2×2) phases [17]. This model is illustrated in Fig. 2. A later NIXSW investigation of this surface [18] proved consistent with this structure.

The fact that the recent STM study of the Cu(1 1 1)/CH₃S– system shows a near-square outermost layer reconstruction lends credence to the idea that the HT thiolate phase on Ni(1 1 1) does involve a pseudo-(1 0 0) type of reconstruction. In view of the co-existence of the LT and HT thiolates and chemisorbed S, any further structural study requires both elemental and chemical-state specificity, and we have applied the CS-NIXSW method to this problem.