Features
Science Physics Surface Science

Surface Science & the Electronic Structure of Solids

 

by D Phil Woodruff


Surface science is the name given to the investigation of the structural, electronic and chemical properties of the outermost few atomic layers of solids. Investigations in the field are motivated by the need to understand important phenomena that are determined by these properties. For example, heterogeneous catalysis – the process in which molecules can interact at a high rate only when transiently adsorbed on the surface of a special material (the catalyst) - have huge importance in the chemical industry but also, for example, in automobile exhaust systems in which toxic carbon monoxide and nitrogen oxides (known collectively as NOX) are converted into harmless carbon dioxide and nitrogen gas. Corrosion of materials (including rusting or ferrous metals) and its prevention, as in stainless steel, are also important surface processes. In addition the properties of semiconductor electronic devices are largely determined by the properties of the few atomic layers at the interfaces between the different constituent materials.

Synchrotron radiation has proved to be an invaluable tool in these investigations.  In order to investigate the few atomic layers at a surface in the presence of the vastly larger number of underlying atomic layers of the bulk it is essential that one uses investigative techniques that are surface-specific, i.e. that provide information only on these few surface layers [1]. A key technique in the surface scientists’ armoury with this property is photoemission – the process in which electrons are ejected from atoms and solids by incident photons with a kinetic energy characteristic of the photon energy and the binding energy of the electron in the atom or solid. In photoemission, surface specificity is achieved by ensuring that the photoelectrons emitted have an energy in the range of ~50-200 eV, for which the mean-free-path for inelastic scattering is only about 2-3 atomic spacings. To achieve this condition over a wide range of binding energies of different materials it is essential that the photon energy can be tuned within the soft X-ray region, something that is only possible with synchrotron radiation. At higher ‘hard’ X-ray energies, the energy tunability and very high brilliance of synchrotron radiation are also essential for several techniques used to determine the structure of surfaces.

Photoemission investigations of surface chemistry

One way of identifying the elemental species in the near-surface region is by using photoemission to eject electrons from the strongly-bound ‘core’ levels of the constituent atoms. The binding energies of these core states are characteristic of the atomic species, so if the incident photon energy is known, the emitted kinetic energy spectrum provides a ‘spectral signature’ of the constituent atoms. An important detail of this approach, however, is that small changes in these core level binding energies occur when the emitting atom is in different ‘chemical states’, such as when the atom is in a bulk metal or in an oxide. The recognition of this effect led to the development of the technique of ESCA – Electron Spectroscopy for Chemical Analysis by Kai Siegbahn for which he was awarded the Nobel Prize in Physics in 1981. The application of this technique to study surfaces (for which it is now more commonly referred to as XPS – X-ray Photoelectron Spectroscopy) has proved to be a valuable method of surface analysis, but using soft-X-ray synchrotron radiation at much higher spectral resolution and with the photon energy tuned to achieve high surface specificity, the technique can be used to follow surface chemical reactions, such as molecular dissociation, the identification of molecular fragments during a reaction, and the progress of surface oxidation. For example, it is well-known from experiments using laboratory X-ray sources that there is a large ‘chemical shift’ of approximately 2.6 eV between the photoelectrons emitted from core levels of aluminium atoms in Al metal and in the Al2O3 oxide. However, by using soft X-ray synchrotron radiation and studying the uptake of oxygen atoms from exposure to oxygen gas (O2) on a single crystal Al surface (Al(111)) the resulting Al 2p photoemission spectrum recorded at the SRS (Fig. 1) clearly shows three intermediate chemically-shifted states (with shifts of approximately 0.5, 1.0 and 1.5 eV) that were filled sequentially prior to true oxidation, these states corresponding to surface Al atoms bonded to one, two or three O atoms on the surface [2].

Fig. 1 Photoelectron energy spectrum recorded in the energy range of the Al 2p emission from an Al(111) surface following exposure to oxygen gas, showing the fitted components of Al2O3 oxide, metallic Al and three intermediate peaks corresponding to different numbers of O atoms bonded to individual Al atoms. Note each state consists of a spin-orbit split doublet, although in some cases broadening smears out the two components.

Photoemission studies of electronic bands in solids and at surfaces

While photoemission from core levels, with binding energies characteristic of the atomic species and its chemical state, are invaluable in studies of surface composition and surface chemical reactions, photoemission from shallowly-bound valence electronic states of solids can provide information of the electronic structure of these solids. In particular, the interaction of the electronic valence states of adjacent atoms in a solid leads to delocalised electronic bands, and the full electronic structure is characterised by the energy-momentum (E-k) band structure of the solid.

Fig. 2 Schematic diagram of a simple solid band structure showing the k-conserving photoemission transitions at two different photon energies, 1 and 2. Different photon energies are required to access different values of the electron wavevector (momentum), k

Angle-resolved photoelectron spectroscopy (ARPES), in which measurement of both the energy and direction of the emitted electrons determines both E and k for the emitted electrons, leads through conservation laws to a determination of the E and k values of the states from which the photoelectrons have been emitted. This technique has proved to be an exceptionally powerful tool in understanding the electronic properties of complex solids. An important feature of these conservation laws, illustrated in Fig. 2, is that the full range of band states can only be probed if one has access to a fully-tunable range of photon energies, and thus to synchrotron radiation. Notice that the figure shows that between the two (initial state and final state) bands is a band gap, corresponding to an energy range in which no electron band states exist in the solid. Within this energy range, however, surface states can exist; these are electronic states that are localised at the surface but cannot penetrate into the solid because there are no bulk states at this energy to couple to. Synchrotron radiation ARPES also allows one to clearly distinguish between photoemission from these surface states and photoemission from bulk states, because while changing the photon energy changes the energy of the bulk initial states that are probed (see the diagram), it does not change the initial-state energy of the surface state emission. Pioneering experiments of this type conducted at the SRS, include investigations of the surface and bulk electronic structure of metal alloys, by the group of Robin Jordan at the University of Birmingham (e.g. [3]), and of bulk metallic states by the group of Colin Norris at the University of Leicester (e.g. [4]). Subsequent major advances in instrumentation for these experiments have rendered this technique exceptionally important in recent years in understanding the electronic structure of materials such as graphene, complex oxides, and topological insulators.

Surface Structure Determination

A number of techniques to achieve quantitative determination of atomic positions in the surface rely on the use of synchrotron radiation. Perhaps the most obvious one is surface X-ray diffraction (SXRD). The difficulty with using conventional X-ray diffraction to determine the structure of a surface, however, is that X-ray scattering is extremely weak. It is for this reason that X-rays penetrate far into (or through) most materials, making the technique ideal for bulk structural studies. To investigate a surface structure, however, this presents two problems, namely surface sensitivity (the scattered signal from a few atomic layers is very weak) and surface specificity (under most conditions the detected signal is dominated by scattering from the underlying bulk). The extremely high spectral brilliance of synchrotron radiation means that a high flux of X-rays can be delivered to the sample in a narrow beam, offering a solution to the surface sensitivity problem. Surface specificity can be achieved in two ways; firstly by using extreme grazing incidence below the critical angle for total reflection (the refractive index of solids in the X-ray region is very close to unity), preventing the X-rays to penetrate into the bulk, and secondly by making measurements in directions that do not correspond to bulk diffraction peaks but do receive surface scattering. The extreme demands of the SXRD technique are such that on the SRS these experiments were performed on the first wiggler insertion device installed on the storage ring, where a facility to perform experiments on semiconductor surfaces during growth from the vapour phase was pioneered (e.g. [5]).

A quite different approach to the determination of surface structure is through the use of a surface version of the EXAFS (extended X-ray absorption fine structure) technique, known as SEXAFS. The EXAFS technique, widely used in materials science (including biological materials) exploits the coherent interference of the outgoing photoelectron wave from an atom, following impact by photons with an energy above a (core level) absorption edge, with components of the same electron wavefield backscattering from nearby atoms. By scanning the photon energy (only possible with synchrotron radiation) the photoelectron wavelength changes, causing the scattering paths to switch in and out of phase, leading to the modulations (the ‘fine structure’) in the absorption scan. The periodicity of these modulations provides information on interatomic distances.  As such it is a means of determining the local structural environment around the photo-absorbing atom. A conventional absorption measurement (determining the intensity of the X-ray flux passing through a sample) clearly lacks the surface specificity required for SEXAFS, so instead the absorption at the surface is detected by monitoring the processes leading to the decay (rather than the creation) of the core hole, typically through the emission of Auger electrons that provide surface specificity through the short mean-free-path for inelastic scattering of these low energy electrons. A beamline designed for these SEXAFS experiments (e.g. [6]) was one of the first to be established on the SRS.

 

Fig. 3 Schematic diagram, showing the effects of the interference of the incident and scattered X-rays (shown by the arrows and the plane wavefronts) at a Bragg scattering condition leading to an X-ray standing wave in the crystal.

 

However, this beamline also led to the pioneering development of a new technique for surface structure determination, namely normal incident X-ray standing waves (NIXSW) [7]. X-ray standing wave experiments exploit the fact that at a Bragg reflection condition the incident and scattered X-rays interfere to produce a standing wave with the periodicity of the crystal lattice (Fig. 3).  As the X-ray energy is scanned through the narrow range of this diffraction condition, the phase of the standing wave shifts in a clearly predictable fashion, and by measuring the absorption this X-ray wavefield at specific atoms, the location of the absorbing atoms either inside or on the solid surface can be determined relative to the position of the scattering atoms in the bulk. The XSW technique at more grazing incidence angles had used in other experiments for some years, but in its standard form it required extremely perfect crystals, meaning that it had been almost entirely restricted to studies of silicon. Using normal incidence renders the experiment far less sensitive to crystal perfection (the standard Bragg condition, 2dsinθ=nλ is at a turning point with θ=90°, and so is insensitive to the exact value of θ), which means that NIXSW can be applied to a wide range of metal and oxide crystals. The lower energy (longer wavelength, λ) also makes detecting the absorption by monitoring the core hole decay by Auger electron emission, or the absorption process itself by photoemission, more surface specific. Indeed, by monitoring the photoemission and exploiting the same chemical shifts that form the basis of the ESCA/XPS technique, it was possible to show that chemical-state-specific surface structures could be extracted [8]. This technique is now yielding important results on the understanding of interfaces in organic semiconductor materials through experiments at the new Diamond facility in the UK.

 


[1] D.P. Woodruff & T.A. Delchar, Modern Techniques of Surface Science, 2nd Edition, Cambridge University Press, 1994.

[2] C.F.McConville,  D.L.Seymour,  D.P.Woodruff  and S.Bao 'Synchrotron  radiation  core  level photoemission investigations of the initial stages of oxidation of Al(111)' Surf.Sci. 188 (1987) 1-14

[3] R.G. Jordan, G.S. Sohal, and P.J.Durham, Photon-energy dependence of photoemission from the Schockley-type surface state on a Cu-Pd(111) alloy, J. Phys. F: Met. Phys, 16 (1986) L135-139.

[4] H.A. Padmore, C.Norris, G.C. Smith, C.-G. Larsson, and D. Norman, Comparison of theory and experiment for angularly-resolved photoemission from Ag(001) in the photon energy range 40 to 100 eV J. Phys. C: Solid State Phys 15 (1982) L155-158.

[5] A. Vlieg, A.W. Dernier van der Gon, J.F. van der Veen, J.E. MacDonald and C. Norris, Surface x-ray diffraction during crystal growth: Ge on Ge(111), Phys. Rev. Lett. 61 (1988) 2241-2244.

[6] R.G.Jones,  S.Ainsworth,  M.D.Crapper,  C.Somerton, D.P.Woodruff,  R.S.Brooks, J.C.Campuzano, D.A.King, G.M.Lamble and M.Prutton 'A surface EXAFS study of a  surface iodide phase on Ni{100}' Surf.Sci. 152/153 (1985) 443-52

[7] D.P.Woodruff, D.L.Seymour, C.F.McConville, C.E.Riley, M.D.Crapper, N.P.Prince and      R.G.Jones  'A simple X-ray standing wave technique and its application to the  investigation  of the Cu(111)(Ö3xÖ3)R30°-Cl structure' Phys.Rev.Lett. 58 (1987) 1460-2

[8] G.J.Jackson, D.P.Woodruff, R.G.Jones, N.K.Singh, A.S.Y.Chan, B.C.C.Cowie, V.Formoso, Following Local Adsorption Sites through a Surface Chemical Reaction: CH3SH on Cu(111) Phys.Rev.Lett. 84 (2000) 119-122