Professor Jack Yarwood is an emeritus professor at Sheffield Hallam University. Professor Simon Duckett is a research group leader at the University of York, UK. His group is mainly involved in the design, development and implementation of NMR methods, supported by the synthesis of inorganic and organometallic complexes. Dr Richard Douthwaite is at the University of York, UK. His main research interests include molecular and materials chemistry and photocatalysis. Both an EPSRC college member and fellow of the Royal Society of Chemistry, Dr Douthwaite is also on the SCI National Materials Committee.

Spectroscopic Properties of Inorganic and Organometallic Compounds Techniques, Materials and Applications Volume 42

A review of the recent literature

By J. Yarwood, R. Douthwaite, S. B. Duckett

The Royal Society of Chemistry

Copyright © 2012 The Royal Society of Chemistry
All rights reserved.
ISBN: 978-1-84973-152-2

Contents

Preface, v,
Photoelectron spectroscopy of metal surfaces for potential heterogeneous catalysis Georg Held, 1,
Vibrational spectroscopic studies of catalytic processes on oxide surfaces Olivier Marie, Philippe Bazin and Marco Daturi, 34,
Studies of water at inorganic solid surfaces Maria Antonietta Ricci, Rosaria Mancinelli and Fabio Bruni, 104,
Infrared and Raman spectroscopic studies of archaeological materials Rosemary A. Goodall and Peter M. Fredericks, 129,
Terahertz spectroscopy of inorganic glasses and carbon nanotubes Edward P. J. Parrott, J. Axel Zeitler and Lynn F. Gladden, 157,
Nuclear quadrupole resonance spectroscopy K.B. Dillon, 184,
Electrochemical impedance spectroscopy (EIS) for PEM fuel cells Mali Hunsom, 196,
NMR diffusion methods in inorganic and organometallic chemistry Paul S. Pregosin, 248,

CHAPTER 1

Photoelectron spectroscopy of metal surfaces for potential heterogeneous catalysis

Georg Held

DOI: 10.1039/9781849732833-00001

1 Introduction

The “Photoelectric Effect” was discovered in 1888 by Wilhelm Hallwachs and famously explained by Albert Einstein in 1905 as the annihilation of a photon by transferring all its energy to the excitation of a bound electron into vacuum. Einstein’s work, together with Planck’s law for black-body radiation, introduced the concept of photons as quanta of electromagnetic radiation and was pivotal for the development of quantum theory in general, however it was the development of precise electron energy spectrometers in the 1950’s by Kai Siegbahn and coworkers, which turned X-ray induced photoemission into X-ray Photoelectron Spectroscopy (XPS), one of the most powerful tools of surface analysis available to date. The availability of synchrotron radiation for “parasitic users” since the mid 1960’s and particularly the improved beam qualities of purpose-built third generation synchrotrons that started to go online in the 1990’s has marked another step-change, now enabling spectroscopic measurements at the time scale of seconds and at an energy resolution that is essentially determined by the natural line width of the samples under investigation. These improvements have led to developments of new variants of XPS, which are especially useful to research in the field of catalysis. A recent issue of the Journal “Nuclear Instruments and Methods in Physics Research”, which is dedicated to Kai Siegbahn (volume 601, issues 1–2) provides an excellent overview over recent developments in XPS. The present review aims at providing a summary of new experimental techniques and other methods related to XPS, which are relevant to the field of heterogeneous metal catalysts and their application, roughly covering the last decade since 2000.

1.1 Physical principles

The kinetic energy of a photoelectron emitted from a solid is

[MATHEMATICAL EXPRESSION OMITTED] (1)

hv is the photon energy, ΦSpectr the workfunction of the spectrometer (typically around 5 eV), and EB the electron binding energy (positive, with respect to the Fermi energy). When a monochromated UV or X-ray source and a calibrated electron spectrometer are used the kinetic energy spectrum can be converted directly into binding energies using Equation (1). The spectral region of the valence electrons (binding energies below 20 eV) is affected strongly by the chemical bonds of the emitter atoms and/or the crystal structure of a solid and cannot, therefore, be used for elemental analysis. The binding energies of core-level electrons, however, are characteristic for the chemical identities of the emitter atoms. Their photoionisation cross section depends on the photon energy and polarisation and on the nature of the core orbital but not on the chemical environment of the emitter atom if diffraction effects can be ignored (see below). This enables quantitative elemental analysis. The strong interaction between electrons and solid matter leads to a short escape depth of photoelectrons of only around 1 nm if their kinetic energy is below 1000eV. This makes XPS very surface sensitive.

The remaining core and valence electrons will react to the creation of the core hole at the same timescale as the photo-ionisation process. This relaxed electron system partially screens the core hole and leads to a reduction of the measured binding energy, EB, by several eV with respect to the binding energy of the core electron before ionisation, which is often referred to as the initial state or Koopmans energy. The relaxation energy depends on the configuration of the system after the photoionisation has taken place and is, therefore, a “final state effect”.

Eventually, the core hole is filled with an electron from a higher-lying orbital shell. The difference in binding energies is either used to create a photon or transferred to another electron which is excited into vacuum. The former core-hole decay mechanism is called X-ray fluorescence and leads to the creation of a new electron hole in the upper orbital. The latter mechanism, Auger electron emission, leads to a final state with two electron holes and to additional well-defined peaks in the electron energy spectrum, which are also characteristic for chemical nature of the emitter atom. For a particular orbital the ratio between X-ray fluorescence and Auger electron emission increases with increasing atomic number. The newly created electron holes lead to a cascade of fluorescence and Auger emission with decreasing energies.

Other final state effects include satellite peaks due to electronic excitations or incomplete relaxation of the electron system, broadening due to vibrational excitations induced by the photoionisation process and peak broadening due to short core-hole lifetimes. The latter effect is a direct consequence of Heisenberg’s uncertainty principle and is particularly noticeable for core holes that can be filled via fast Auger processes, which include electronic states from the same shell (Coster-Kronig process).

1.2 Chemical shifts

Both initial state energy and final state relaxation, are affected by changes in the valence electronic structure of the emitter atom. This leads to XPS binding energy differences of up to several eV for the same element in different chemical environments. Such chemical shifts provide the basis for analysing the chemical composition of a sample in much more detail than a simple elemental analysis.

1.2.1 Oxidation state, surface core level shifts (SCLS). Two types of chemical shifts are particularly useful to characterise metal catalyst surfaces or nanoparticles, namely shifts due to the metal oxidation state and surface core level shifts. Both are related to a change in the valence electron density near the emitter atom.

For low lying core levels of transition metals an increase in the oxidation state leads to a positive shift in the binding energy, which has been used frequently to determine the chemical state of model catalysts surfaces and nanoparticles. This shift is commonly considered as initial state effect: the reduced number of valence electrons in the metal cation leads to a lower degree of repulsion between core and valence electrons and, hence, to a higher binding energy. Consequently, also other charge-transfer mechanisms from/to metal atoms, e.g. in alloys, lead to such shifts. Rodriguez and Goodman have used binding energy shifts to determine the level of charge transfer in bimetallic model catalyst systems and could correlate this with their reactivity in terms of CO adsorption. A similar correlation has been found in systematic theoretical studies by Ruban et al.. These studies also revealed that the experimental core level shifts observed for transition metals in bimetallic systems are mainly affected by changes in the position of the d-band with respect to the Fermi energy, which, in turn, depends on the amount of charge transfer.

The lower coordination of surface atoms causes a shift in their binding energy with respect to bulk atoms, which is typically less than 1 eV. Early work has shown that the sign and magnitude of surface core-level shifts for clean transition metal surfaces depend on the degree of d-band occupation and the coordination of the surface metal atoms. Since the early 1990’s the effect of adsorbate atoms and small molecules, such as O, H, CO and NO has been studied making use of the high resolution of third generation synchrotron sources. A compilation of experimental data for clean and adsorbate-covered surfaces has recently been published by Denecke and Maårtensson. A series of recent theoretical and experimental studies by Baraldi et al. and Over et al. has shown a linear correlation between the SCLS and the coordination of surface transition metal atoms (see Figure 1). It was also found that the SCLS are usually only affected by direct interaction between adsorbates and surface atoms, which has been used to determine the coordination of adsorbate molecules such as CO, water, and thiolates.

1.2.2 Molecular bonding. Chemical shifts of the same element in different molecular environments can be as big as 10 eV. This is about the range of shifts observed in the C 1s binding energies from polymers. The lowest binding energies are measured for carbon atoms forming only non-polar bonds with hydrogen or other carbon atoms, the highest for highly electronegative bonds, e.g. -CF3. These chemical shifts are a combination of initial state effects and final state effects. In many cases the latter dominate and can be described reasonably well using the equivalent core approximation whereby the core-ionised atom within the molecule is replaced by a neutral atom in the same position with the atomic number increased by one. The such created new molecule will normally not be in its optimum configuration and, thus, have a strain energy added to its total energy. The binding energy differences between the same element in different molecular bonding configurations are due to differences in these strain energies. For example, the equivalent core atom for core-ionised oxygen is fluorine. Comparing the final states of OH and H2O one finds that the hypothetical H2F molecule with the atoms in same positions as in water has a much higher strain energy than FH with the atoms in the same positions as in OH. Therefore the experimental O 1s binding energy of H2O is significantly higher, by 1–2 eV, than that of OH. Similar considerations can also explain the relatively large binding energy shifts when small molecules are adsorbed on different adsorption sites of metal surfaces. CO adsorption on transition metal surfaces is the best-studied example in this context. In all cases the O 1s and C 1s binding energies decrease with increasing metal coordination, largely un-affected by co-adsorbates, with shifts of the order of 1 eV in the C 1s and 2 eV in the O 1s binding energies.

1.3 X-ray absorption

The excitation of a photoelectron is always accompanied by the annihilation of a photon. X-ray absorption spectroscopy is, therefore, in many ways complementary to X-ray photoelectron spectroscopy. In absorbance-vs-photon-energy spectra characteristic absorption edges are observed when the photon energy is equal to the excitation energy of a core level, because then a new absorption channel is opened. Extra features around these edges are due to excitations of electrons into unoccupied bound sates below the vacuum level or quasi-bound resonance states just above. These features are known as near-edge X-ray absorption fine structure (NEXAFS) or X-ray absorption near-edge spectroscopy (XANES). NEXAFS features often show strong polarisation dependence because the absorption cross section, σ, depends on the alignment between polarisation vector and the symmetry elements of the two orbitals involved (core, Ψcore, and unoccupied state, Ψunocc) according to Fermi’s Golden Rule:

[MATHEMATICAL EXPRESSION OMITTED] (2)

For adsorbed molecules the symmetries of Ψcore (usually a totally symmetric s orbital) and Ψunocc (an unoccupied molecular orbital) are usually known, hence, the polarisation dependence of σ can be used to determine the orientation of the molecules.

1.4 X-ray sources

1.4.1 Laboratory sources. Gas-discharge lamps and X-ray anode sources are used to produce monochromatic radiation for UPS and XPS in the laboratory. He discharge or plasma lamps are most commonly used for low photon energies. The natural widths of the He I (hv = 21.2 eV) and He II (hv = 40.8 eV) lines are less than 0.002 eV, therefore monochromators are only used to separate them but not to reduce their line width. Common X-ray sources are based on Mg Kα1,2 (hv =1253.6 eV) and Al Kα1,2 (hv = 1486.7/1486.3 eV) emission from stationary or rotating anodes. Their natural line widths are 0.6 and 0.8 eV, respectively, which is sufficient for standard elemental analysis, however most modern instruments use crystal monochromators to reduce the line width, typically to about 0.2–0.3 eV, and suppress satellite lines that would otherwise lead to “ghost peaks” in the photoelectron spectrum. For higher photon energies also Cr Kβ1,3 (hv =5946.7 eV) and Cu Kβ1,3 (hv =8905.3 eV) are in use but there are no laboratory sources readily available for photons in the energy range between 40.8 eV (He II) and 1253.6 eV (Mg Kα).

1.4.2 Synchrotron radiation. When charged particles at relativistic speed are forced onto a curved trajectory synchrotron radiation is emitted. Initially regarded as an unwelcome energy loss in particle accelerators, synchrotron radiation has been used as X-ray source for materials science studies since the mid 1960’s. Since the 1990’s so-called “third generation” synchrotron sources have become widely available. These are purpose-built high-brilliance light sources to suit the needs of chemical and structural characterisation and have become a standard tool in catalysis research in the last decade. To date, about 50 synchrotron radiation facilities worldwide offer their services to a wide user community. The “X-ray data booklet”, which is distributed by the Advanced Light Source (ALS) in Berkeley, provides a brief introduction into the history of synchrotron radiation facilities and a valuable compilation of relevant data. The advantages of synchrotron radiation over laboratory sources are listed below:

Tunability: Unlike anode line sources synchrotron bending magnets provide a continuous spectrum of intense radiation from the infrared to the hard X-ray regime (W100 keV). The signature of third generation synchrotron facilities are “insertion devices”, such as wigglers and undulators, which are arrays of small bending magnets. These create structured photon spectra with narrow lines of high intensity that can be moved within the spectrum by changing the distance between the magnets. In the standard beamline layout the radiation from a bending magnet or insertion device passes through a grating or crystal monochromator (below and above 2000 eV, respectively) with a resolving power (E/ΔE) of typically 5000–10000. This makes it possible to access fixed photon energies in the soft X-ray region below 1000 eV with high photoemission cross sections for elements that are particularly important for chemical research, such as C, N, O, and enables absorption spectroscopy by scanning the photon energy, which is impossible with laboratory sources.

High brilliance or spectral brightness: Synchrotron radiation is much more intense than conventional laboratory sources and highly collimated. Typical numbers for state-of-the art undulator sources are 1018–1019 photons/ (s·mm2·mrad·0.1% bandwidth), which corresponds to a flux of about 1014 photons per second on a typical 100 µm × 100 µm spot or up to 103 photons/s per surface site. This is 8 to 10 orders of magnitude higher than conventional laboratory sources.

Polarisation: The polarisation of synchrotron radiation is well-defined. Radiation emitted from bending magnets is linearly polarised within the plane defined by the electron beam; above and below the plane the radiation is elliptically polarised with opposite chiralities. The latest generation of APPLE II-type undulators allow arbitrary linear or elliptical/circular polarisation.

(Continues…)Excerpted from Spectroscopic Properties of Inorganic and Organometallic Compounds Techniques, Materials and Applications Volume 42 by J. Yarwood, R. Douthwaite, S. B. Duckett. Copyright © 2012 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.