Advanced XPS characterization: XPS-based multi-technique analyses for comprehensive understanding of functional materials

X-ray photoelectron spectroscopy (XPS) has achieved maturity as an analytical technique in that it is a ubiquitous tool in the materials community, however as made apparent by recent reviews highlighting it’s misuse as a means of chemical deduction, it is a practice which is greatly misunderstood even in its simplest form. Advanced XPS techniques, or a combination of XPS and a complementary surficial probe may elicit auxiliary information outside of the scope of the standard sphere of appreciation. This review aims to bring to the attention of the general materials audience a landscape of some atypical applications of lab-based XPS and combinatorial approaches of related surface analysis, such as ion scattering, ultraviolet photoelectron, electron energy loss and auger emission spectroscopies found on many lab-based instrument set-ups.


Introduction
Functional materials on the nano-to macro-scale provide solutions for some of the most critical scientific problems of the current age, from the use of heterogeneous catalysts to improve the environmental footprint and commercial viability of fine chemical synthesis (e.g.pharmaceuticals, food additives, cosmetics) 1 to solid state materials as energy storage materials, 2 fuel cells 3 or photovoltaics 4 to tackle humanity's dependence on dwindling resources such as fossil fuels.Such is the scope in the application of functional materials that examples may be found across the spectrum of the physical sciences; including drug-delivery systems, 5 molecular sensors, 6 tackling drug resistant pathogens, 7 aerospace components, 8 electronics 9 and many more.In order to design and develop the most effective materials, a thorough understanding of the material properties is often paramount -particularly at the surface of a material which forms the interactive boundary between the functional material and the materials or substances on which they will act. 10-ray photoelectron spectroscopy (XPS) has become a wellknown technique in the field of materials science, one which may be used to probe the elemental, electronic and chemical characteristics of a material surface.[11][12][13][14] While the interpretation of XPS data via peak deconstruction is commonplace, fundamental understanding of the technique and an appreciation for correct data handling are often missing.15 Recently, a series of invaluable guides have been prepared from a community effort among the leading practitioners in the field of XPS, with the aim of enabling researchers new to XPS to plan experiments and understand their data to a high level.This series is published under the title 'Practical Guides for X-Ray Photoelectron Spectroscopy', for example 'Practical Guides for X-ray Photoelectron Spectroscopy: First Steps in Planning, Conducting, and Reporting XPS Measurements' 16 and 'Practical guide for curve fitting in X-ray photoelectron spectroscopy.'17 Furthermore, there exist a number of advanced techniques with which many materials scientists will be unfamiliar.Furthermore, a current focus of several of the XPS manufacturers is high throughput examination of a single analysis spot using an assembly of surface probes and even non-surface specific techniques such as Raman spectroscopy.As the availability of such systems becomes more widespread, the a wider knowledge of the capabilities, strengths and weaknesses of multi-technique surface analysis is required.
This review aims to highlight the benefits of such a combinatorial approach to the analysis of materials utilising lab-based XPS and related surface techniques.Applications of those experimental probes which may be found most regularly on lab-based systems (ion scattering, ultraviolet photoelectron, auger emission and electron energy loss spectroscopies) are covered in detail here, though many other complementary

Roxy Lee
Roxy graduated with an MSci Chemistry from UCL in 2018.In her fourth year project she developed and modified nanostructured TiO 2 electrodes and used these to fabricate and analyse their performance in solar cells.She is currently enrolled on an EngD programme (2019-2023), sponsored by Thermo Fisher Scientific, and focuses on photoemission spectroscopy and valence band spectrum analysis.She also uses computational methods to simulate photoemission spectra, and uses a joint experimental and theoretical approach in her research.

David J. Morgan
David J. Morgan obtained his PhD in surface science and catalysis in 2002 from Cardiff University under the supervision of Prof. P. R. Davies, where he then undertook postdoctoral research under the supervision of Prof. Graham Hutchings FRS focusing on model catalytic studies using high pressure XPS and oversaw the ESPRC Access to Research Equipment initiative for X-Ray Photoelectron Spectroscopy.Since 2007 he has been the Surface Analysis manager at Cardiff University and the technical manager of HarwellXPS since 2017.He has interests in standards, experimental design, data analysis and multi-technique analysis.Roxy graduated with an MSci Chemistry from UCL in 2018.In her fourth year project she developed and modified nanostructured TiO 2 electrodes and used these to fabricate and analyse their performance in solar cells.She is currently enrolled on an EngD programme (2019-2023), sponsored by Thermo Fisher Scientific, and focuses on photoemission spectroscopy and valence band spectrum analysis.She also uses computational methods to simulate photoemission spectra, and uses a joint experimental and theoretical approach in her research.

Shaoliang Guan
Shaoliang Guan is currently a Research Associate and Co-Investigator at the EPSRC National Facility for Photoelectron Spectroscopy (HarwellXPS), UK.He received his BS degree in Applied Chemistry from Beijing University of Chemical Technology (China) and MSc in Chemical Enzymology from Queen's University Belfast (UK).After that, he received his PhD degree in Electrochemistry from Cardiff University (UK) under the supervision of Prof. Gary Attard in 2015.His research interests include investigating selectively catalytic hydrogenation using electrochemical methods and CO oxy-chlorination reactions with various surface characterization techniques such as X-ray photoelectron spectroscopy and Surface enhanced Raman spectroscopy.approaches may provide pathways through which to tackle research problems in materials analysis.9][20][21][22] Such is the potential of this approach, selected commercial instruments now afford the option of an 'off-the-shelf' Raman module upgrade -permitting users to perform tandem experiments with little to no experience of the modification of UHV chambers.
The sheer scope of combinatorial analysis renders a full discussion of all methods impractical, and hence only a selection (of those deemed most readily available) will be presented here, however for a look into other powerful conjunctive techniques the reader may be directed towards some reviews highlighting the use of XPS/Scanning Tunnelling Microscopy (STM), [23][24][25] XPS/Scanning Probe Microscopy (SPM), 25,26 XPS/Infrared Reflectance Absorption Spectroscopy (IRRAS), [27][28][29] XPS/Time-of-Flight Secondary Ion Mass Spectroscopy (TOF-SIMS) 25,26 and XPS/Atomic Force Microscopy (AFM) 25,27 though even this is itself by no means an exhaustive list.

Advanced XPS techniques
As well as providing a method by which to qualitatively and quantitatively describe the chemical nature of the surfaces of materials, XPS can also provide insights into the topology, morphology and homogeneity of a sample as a function of depth.Such techniques may facilitate the understanding of new interfaces (e.g. in semiconductors, electrodes), 30,31 solid synthesis pathways, 32 catalytic and surface processes [33][34][35] and electronically active materials, 36,37 among others.

Tougaard background analysis
Each individual photoemission peak in an XP spectrum is accompanied by a characteristic stepped background (Fig. 1) which is the result of photoelectrons which have undergone an inelastic scattering process. 38In order to correctly analyse XP data, this inelastic scattering process must be properly accounted for in the form of fitted model background. 39he most widely used model, known as the Shirley background after its creator D. A. Shirley, 40 uses an iterative process to calculate the number of inelastically scattered electrons based upon the integrated peak area either side of the energy E. It has proven to be a reliable and robust method for accounting for inelastic scatter.However, the physical origin of the Shirley background is rather poorly defined. 41Detailed theoretical analysis with comparison to experimental data has led to much improved models of electron inelastic scattering 42 and a new method for quantitatively describing the XPS background named after its creator Sven Tougaard has been developed. 43A discussion of the merits of the various background types for standard XPS analysis is beyond the scope of this review but we will highlight how Tougaard's approach provides a non-destructive technique by which depth profile and structural information, including the spatial arrangement of overlayers, islands and sandwich structures can be obtained from the background profiles to complex composite materials. 44ucleation and growth modes in systematic samples can be obtained with an even greater degree of accuracy and consistency than that provided by ellipsometry or Rutherford backscattering spectroscopy. 45,46Tougaard background analysis has yielded excellent agreement with atomic force microscopy (AFM) measurements in determining Stransky-Krastanov growth of NiO deposition onto highly ordered pyrolytic graphite (HOPG). 47The technique has advanced to the point where analysis may be performed quickly and easily using specialist software (QUASES). 48,49erlayer thickness analysis Nanocomposite materials in the form of thin films or coreshell particles often offer unique solutions to critical research problems in areas such as electronics, 50 medical devices 32 and catalysis. 51Control over the film or shell thickness can be vital to obtain the precise physical and electronic properties required of such composite materials and therefore a robust method is needed to determine said thickness.X-ray reflectivity (XRR) 52 and electron microscopy 53 are often used for this purpose, however they are not without drawbacks; XRR requires highly ordered samples with almost no surface roughness, whilst electron microscopy requires long analysis times to obtain a statistically significant dataset.XPS can provide an alternative approach to film thickness determination involving a large area analysis with few limitations regarding sample chemistry.For planar materials, such as semiconductor thin films deposited onto a substrate, there are a number of approaches to using XPS to determine film thickness.For simple metal oxide on metal systems such as Al 2 O 3 /Al for example, one may use the Strohmeier equation (eqn (1)) based on a single high resolution region spectrum. 54,55 Strohmeier equation; l = inelastic mean free path, N = volume density (density/mass), I = intensity and y = photoelectron take-off angle.This method assumes a uniform overlayer and that the photoelectrons from both metal and oxide are of similar energy (hence is largely restricted to metal oxide layers of the same element as the underlying metal).
Differences in the inelastic mean free path (IMFP) between oxide and metal permit a facile calculation of path length using only intensity ratios between the two species.An important note for users relatively unfamiliar with XPS, is that between IMFP and the information depth of the sample.IMFP (l) is defined as the distance between successive inelastic collisions of photoelectrons and the surrounding medium (i.e.atoms in the sample as the photoelectron travels to the surface).The information depth (ID), refers to the depth of analysis from which the majority (495%) of the signal originates. 57This is dependent on both the material IMFP and the energy of the incoming X-rays and is typically found to be approximate to 3l.
6][67][68] Application of ARXPS to the valence orbitals is also extensively used to map electronic band structures. 69ne method with which ultrathin-film thickness may be determined is by monitoring the intensity of the substrate before and after deposition of an overlayer and using known inelastic mean free path (IMFP) lengths to calculate the path length of the attenuating material (overlayer thickness). 70his method is not without it's restrictions, since it may not account for differences in surface contamination, and surface topography and it also assumes identical experimental conditions and a pure substrate sample -which may not always be possible.It is, however, an excellent probe of uniform flat surfaces such as those prepared via deposition. 71A technique developed by Cumpson known as the ''thickogram'' facilitates overlayer analysis from a single XPS spectrum via graphical means. 72he method uses minimal calculation, and has the advantages of being independent of contamination layer and instrument factors, and works well over a large range of thicknesses and energies.The ratio of kinetic energies and the ratio of integrated peak areas for a substrate and overlayer peak, are plotted onto a thickogram template.The intersection of a line joining these two points with the third axis on the thickogram provides a value for the coating thickness.Since sample morphology may still influence the analysis when using this method, a series of correction factors (known as topofactors) have been derived by Shard which may permit the determination of non-uniform sample surfaces and even complex structures such as core shell nanoarchitectures. 73he analysis of smaller particles (o10 nm) requires particular consideration since the substrate on which the particles are mounted may also contribute to the XP spectrum and in these cases considerable care must be taken with methods such as Tougaard's background analysis.Angle-resolved measurements are typically unsuitable for such systems, since at such small size ranges we have topological considerations rendering the technique inappropriate.Models have been developed which permit calculations to account for the morphological differences of non-planar samples at a wide range of sample dimensions. 74mploying a twin energy X-ray source (for example Mg and Al) is one method by which to overcome some limitations regarding experimental conditions, since different energy X-rays may analyse the same region of the same sample but result in photoelectrons with different mean free paths and thus provide more information on the overlayer structures. 32This method has been shown to be in good agreement with electron microscopy for Ag@SiO 2 core shells over a small size range. 32he overlayer analysis methods described so far have the major advantage of being non-destructive and the superb depth resolution (in the order of Angstroms) of such non-destructive methods makes them ideal for analysing small particles (o20 nm for energy sources below 5 keV) thin films, overlayers or inhomogeneous samples. 75However, all these techniques are limited by the maximum kinetic energy of the chosen X-ray source (typically around the order of 10 nm for a 1486.69eV Al ka X-ray source).To address this limitation, a range of high energy laboratory X-ray sources such as the liquid gallium jet (9.25 keV), 76 are being developed which complement the variable energy photons offered by synchrotrons (up to 18 keV) 75,77,78 and facilitate examination of more deeply lying interfaces.An additional limitation when dealing with HAXPES is the decreased sensitivity due to diminishing photoionisation cross-sections at higher excitation energies [79][80][81] (thus also requiring higher powered sources which has the potential to lead to greater X-ray induced sample damage).One result of this is the need for high energy analysers with a large angle of acceptance, creating a further barrier in the development of HAXPES systems. 82,83Nevertheless there will always be an upper limit to the depth that non-destructive XPS methods can probe.

Depth profiling
An alternative method for exploring composition gradients in the Z-dimension by XPS is to use a destructive bombardment technique.This typically involves an ion beam (most commonly Ar + ) in which accelerating voltage, current and raster size may be tuned in order to control the removal of surface atoms via collisions with the ions. 84,85Obtaining snapshot spectra at regular intervals during the etching process (at which point the ion beam is turned off) permits the construction of a depth profile in which elemental or chemical state may be charted as a function of total sputter time (Fig. 3); converting the data to a depth profile requires an estimate of the etching rate for the material under study.The major drawbacks in ion bombardment depth profiling are a reduced depth resolution due to a mixture of non-uniform sputtering (e.g.cratering), 86 preferential sputtering (e.g.oxygen in metal oxides), 87 chemical reduction during sputtering 88 and the need to calibrate etching time with depth. 84ntil recently another major issue with ion bombardment sputtering was that it was limited when applied to organic materials, since ion etching of polymers results in significant chemical degradation. 90However, the advent of gas cluster ion sources/beams (GCIS/GCIB) in which ionised multi-atomic clusters are used in place of single ions has led to more stable etch craters due to multiple secondary collision events 91 and exhibit minimal damage for organic molecules, enabling z-dimension profiling and 3D imaging for the vast majority of organic materials. 92Depth resolutions as low as 7 nm have been recorded and may be readily improved through the Fig. 3 XPS depth profile composition measurements of pristine and magnesiated V 2 O 5 cathodes indicating the relative contents of carbon, oxygen, vanadium, tin and magnesium as a function of overall ion bombardment time (in itself a function of depth from the surface of the sample down into the bulk). 89his journal is © The Royal Society of Chemistry and the Chinese Chemical Society 2021 utilisation of a rotating sample stage and a decrease in X-ray flux, since exposure to X-rays can also degrade the steadiness of the etching process. 93Further improvements to depth resolution may come from combining ion bombardment with a technique such as TOF-SIMS. 94vanced sample handling for lab-based XPS systems While not necessarily an individual technique, the development of technology designed to overcome the imitations of analysing samples within a UHV system has widened the reach of XPS as a tool for probing materials.The advent of near-ambient pressure (NAP)-XPS 95 is one pertinent example of the science adapting to accommodate samples previously difficult or even impossible to analyse -though this has historically required specialist instrumentation often only found at a synchrotron.Recent years have seen a marked rise in the number of lab-based NAP-XPS systems, 96 with manufacturers such as Omicron Scienta and SPECS offering bespoke and 'off-the-shelf' instruments capable of performing ambient pressure measurements.This review will, however, focus on techniques by which we can use traditional lab-based XPS systems to probe air-sensitive materials or perform in situ experiments.97][98][99][100] Studying electrochemical systems using XPS One discipline in which such an approach has had great impact is the field of electrochemistry which is currently of great research significance due to its role in the energy sector where electrocatalysis represents a viable solution to hydrogen production 101 and electrode materials to that of energy storage. 102It is generally accepted that an improved understanding of the underlying mechanisms that occur at the interface between the electrolyte and the electrode, and any restructuring of the electrode surface during polarisation, activation and deactivation is extremely important. 103Ex situ and in situ XPS studies of electrode materials, 104 ionic liquids and electrochemical processes including electrochemical double layer (EDL) studies have been employed to provide insights into such mechanisms.
The surface sensitivity of XPS and its UHV requirements make it quite complicated to apply to the electrode/electrolyte interface; liquid electrolytes are obviously not compatible with vacuum conditions, making true in situ measurements difficult.
Ex situ electrochemical XPS refers to investigations where the material studied is removed from the original electrochemical cell and transferred to the vacuum chamber, usually this would involve disassembly in a glovebox and transfer using a suitable inert transfer device (Fig. 4).Such transfer modules may involve either the use of a vacuum sealed system (Fig. 4a-c), in which samples may be mounted inside a glovebox and the transfer module evacuated within the antechamber, where a check valve permits evacuation inside the cell and holds the vacuum until loading inside the instrument.Alternatively, through the use of a sealed manipulator, gate valve and standard KF flange, samples may be held under vacuum or inert gas, transferred from glovebox to instrument and the KF flange directly mounted onto the vacuum chamber, permitting the sample to be moved into the instrument after the gate valve is opened.Such measurements always introduce the possibility that the surfaces may be different from their original states but valuable electrochemical information can still be obtained following such disassembly as demonstrated by XPS studies of the oxygen evolution reaction (OER) on electrodes. 1058][109] The latter, attempts to retain the original conditions of the samples from the electrochemical cell whilst in the XP spectrometer.Pioneering work by Kolb et al. 110 showed that it was possible to remove electrodes from the electrolyte during potential control and keep the electrical double layer (EDL) intact.Gold electrodes were immersed from various electrolytes prior to performing XPS analysis, which was used to determine the surface excesses and adsorption isotherms for a number of ions and metal deposits.Later, Hecht et al. 111 examined the EDL on Ag in alkaline NaCl after immersion with various potentials, and studied the concentrations of the cations (Na + ) and anions (Cl À and OH À ) adsorbed on the silver electrode via quantitative XPS analysis.Their analysis revealed that the anions tend to bind Na + into the double layer for charge compensation, and that roughly one monolayer of water was found for all potentials studied, in agreement with classical double layer studies.

Ex situ electrochemical measurements
The importance of the surface sensitivity and chemical state information available from XPS is demonstrated well by recent studies carried out by Andreu et al. 104 into lithium ion batteries.The research involved modification of the positive electrode material, a mixed transition metal oxide consisting of LiNi x Mn x Co 1À2x O 2 where (0.33 r x r 0.5) [112][113][114][115][116] to improve its electrochemical performance.8][119] Investigations into how the transition metal nature and Al 2 O 3 coating on LiCoO 2 and LiNi 1/3 Mn 1/3 Co 1/3 O 2 (NMC) surfaces affect performance were carried out by Andreu et al. 104 studying the adsorption of SO 2 on the surface with XPS.The resulting spectra showed only one adsorption mode present for the LiCoO 2 surface, signified by the sulfate species (Fig. 5a and a 0 ) whereas the NMC electrode material exhibited both sulfite and sulfate species (Fig. 5b and b 0 ).In the case of the LiCoO 2 electrode, the coating modifies the surface reactivity, indicated by the appearance of both sulfite and sulfate species, and the authors attribute this to Al/Co substitution shifting the adsorption mode from redox to acid-base.In the case of the NMC electrode, a decrease of sulfite species is seen at the surface, a phenomenon that can be related to the cation mixing effect in the NMC. 120PS depth profiling has also been important in research into alternative energy storage with pseudocapacitors, which are based on redox reactions of electroactive materials, have a longer life than batteries and a higher energy density than conventional capacitors. 121 study of the pseudocapacitive performance of an Mn oxide electrode in the ionic liquid (IL) EMI-SCN was carried out by Chang et al. 122 ILs are composed solely of ions and exhibit several desirable properties including high electrical conductivity (up to 100 mS cm À1 ), a wide electrochemical window (up to 5.8 V), and extended liquid-state temperature range (ca.173-523 K). 123 Furthermore, their negligible vapour pressure overcomes the issue of incompatible liquid electrolytes in UHV systems.In the case of EMI-SCN, the cation (EMI + ) and the anion (SCN À ) both contain nitrogen atoms and so can be conveniently compared using the N 1s region of the XP spectrum, (Fig. 6).Chang et al. were able to identify the working species in the pseudocapacitor as the SCN À anions by recording sequential N 1s spectra of the Mn oxide electrodes in EMI-SCN during argon ion etching; 10 spectra recorded with the electrode under a potential of À0.3 V are shown superimposed in Fig. 6.The peak at 398.7 eV (due to the EMI + cation) disappears after only one Ar + sputtering cycle, indicating that it is only adsorbed on the surface signal whereas the at 401.6 eV (due to the SCN À anion) only gradually decreases with depth, indicating that it has penetrated into the Mn oxide surface.

In situ electrochemical cells
Despite advances brought by ex situ characterisation, postanalysis can mean that some metastable species are not analysed since they are only present under reaction conditions, 124 therefore in situ, or in operando, analysis is of great value.An in situ XPS cell reported by Weingarth et al. 125 demonstrated that XPS analysis can be performed simultaneously whilst carrying out electrochemical measurements, to determine the direct electrochemical shift as a function of applied potential.The system studied was a three-electrode set up involving a counter electrode (CE) and reference electrode (RE) prepared of activated carbon, and a Pt working electrode (WE) in contact with the IL electrolyte [EMIM][BF 4 ].XPS analysis was carried out at the interphase boundary between the IL and Pt WE, with selected associated spectra in Fig. 7.The binding energies (BE) for the electrolyte species core level spectra (B 1s, C 1s, N 1s) exhibited an electrochemical shift of À1.0 eV V À1 under the applied voltage because these atoms are in the outer Helmholtz plane (OHP) or the bulk of the electrolyte, and due to a lack of direct contact with the electrode, subsequently experience the full potential drop.The WE was grounded by the XPS system and therefore the Pt 4f BE and peak shape did not change.N 1s spectra were also recorded at very negative potentials, where the appearance of an extra peak at a potential of À1.9 V indicated a degradation product of an electrochemical reaction.Further examples of in situ UHV XPS cells include those of Wibowo et al. 126 and Gokturk et al. 127,128 Developments in XPS techniques that use synchrotron-based radiation sources as well as lab-based differentially pumped analysers and advanced electrostatic lenses now mean that it is possible to make in situ electrochemical XPS measurements at near ambient pressure (NAP-XPS), where NAP refers to a working pressure of up to a few millibars. 1291][132][133][134] In situ electrochemical XPS studies offer huge potential in electrochemistry research and further developments are expected to be made as the field progresses.

Cryo-XPS
Cryo-XPS employs a similar approach to the cryo-TEM technique that won Dubochet, Frank and Henderson the 2017 Nobel prize  for chemistry, and enables the characterisation of solid-liquid interfaces and volatiles.Samples are prepared as suspensions which are supported on a holder, fast frozen and introduced into the vacuum chamber.Usually, the sample is warmed slightly to a stage where multilayers of frozen volatiles sublime and a virtually unaltered liquid monolayer interface can be studied.
Cryo-XPS provides several clear advantages over conventional preparation by avoiding steps such as freeze drying which can damage the surfaces of more sensitive samples while providing ostensibly in situ measurements. 136,137Furthermore, the presence of a hydrated layer also reduces levels of adventitious carbon because these carbon contaminants tend to be trapped in the first few layers, which are lost during sublimation. 138,139Key considerations are the condensation of volatile species from the laboratory atmosphere onto the surface of the cooled sample and potential issues with sample warming on transfer between cooling, transfer and analysis stages.Ideally then an XPS system suitable for cryogenic analysis would have a glove box attached directly to the load dock allowing for sample preparation in an inert atmosphere 140,141 and LN 2 cooling in both the load dock and analysis chamber, ensuring the sample is maintained at a low temperature during analysis.
Schukarev et al. have carried out several investigations probing water-mineral interfaces using the cryo-XPS approach, delineating the interaction between hydrated surfaces and the chemical speciation of absorbates. 142,143Pastes of metal oxides were prepared through the centrifugation of particles suspended in different pH environments.The wet pastes are then placed onto precooled transfer rods at 170 1C and 1 Â 10 À8 torr before introduction to the analysis chamber. 144The zwitterionic nature of N-(phosphonomethyl)glycine (PMG) absorbed on goethite and manganite was studied using this method and provided information that would not have been possible for traditional XPS analysis.Fig. 8a), shows the stark difference between the N 1s regions recorded at cryogenic and room temperatures, with the absence of the B402 eV peak due to a protonated nitrogen species at room temperature. 135,145Fig. 8b demonstrates the sensitivity of the approach to pH changes in the flash frozen solution with dramatic changes to the relative ratios of NH 2+ and NH at 402 eV and 399.8 eV respectively with varying pH.The 402 eV decreases in intensity as pH increases.The experiment clearly demonstrates the potential for direct monitoring of the changes to the chemical speciation as a function of pH, which could not be extracted from a dry surface.
Similar studies have been carried out on sulfur rich systems such as pyrites and sphalerite, which until recently could not be accurately characterised in a UHV system due to the loss of sulfur species in vacuum. 146,147Ejtemaei et al. 148 combined cryo-XPS and depth profiling to study the activation of metal sulfides through the presence of Cu.Polished mineral samples were submerged in a copper sulfate solution for 10 minutes before being precooled to below À160 1C and maintained at that temperature for analysis by XPS.The samples were then etched 1-10 nm by argon ion bombardment.Survey scans of copper activated sphalerite (ZnS) showed copper ions present up to 10 nm deep (Fig. 9).Furthermore, the Cu and S 2p spectra show a greater proportion of Cu + and S À compared to Cu 2+ and S 2À species present on the surface following copper oxidation.
These findings were interpretated as evidence for a 2 step mechanism in the copper activation of sphalerite involving: Step 1: Cu ion exchange, Step 2: oxidation of the surface forming a CuS layer.
The study of soft matter including biological samples such as bacteria, is an emerging field within cryo-XPS.Fast freezing of cell walls can reduce the alterations and degradation caused by dehydration and beam damage, allowing for a more natural environment for analysis. 149][152] Ramstedt et al. 149 have used Cryo-XPS to study bacterial cell walls treated in different environments, using a multivariate Fig. 8 (a) N 1s spectrum of PMG recorded on a MnOOH surface at room temperature versus fast frozen, evidencing an addition nitrogen species following preparation via cryo-transfer.(b) N 1s spectrum of PMG on MnOOH at pH 3.9, 4.9, 6.0, 7.9 and 8.6 (top to bottom spectral features at around 402 eV).Here we can see the pH dependence on the amount of protonated PMG species on the surface of MnOOH.Copyright r 2004 John Wiley & Sons, Ltd. 135 curve resolution analysis of carbon spectra to analyse the lipid, protein, and sugar content of the bacteria cell surface layers.The work demonstrates the capability to reveal changes in the cell wall chemistry depending on the solution pH, 150 with the chemical composition of Bacillus subtilis being radically different at low pH values compared to high pH, with much higher proportions of protein compared to lipids and carbohydrates.Bacillus subtilis tend to survive in environments between 6-9 pH and when exposed to more acidic environments the bacteria form a 'spore coat' which is higher in protein and low in sugars and lipids.Recently Shchukarev et al. 153 applied this method to successfully characterise the surface layers of both fungi and viruses, demonstrating that the method can be applied to a large range of biological systems, and illuminating mechanisms such as absorption and antimicrobial interactions.
It is interesting to consider the similarities and differences between cryo-XPS and near ambient pressure XPS, which both offer opportunities to examine surfaces in the presence of high concentrations of adsorbates, albeit in very different physical states.Kjaervik et al. 154 used the two techniques in a characterisation of the surface of Pseudomonas fluorescens.Fig. 10 shows that both methods give spectra with at least three clearly defined C 1s species at the surface of the bacteria and very little disparity can be immediately discerned between the two sets of data.However, the scattering of photoelectrons by gas phase components in NAP produces much lower signal intensities, and as a result additional S 2p and N 1s species were identified in the cryo-XPS experiments (not shown).Furthermore, there is a slight but significant increase in the 285.0 eV peak of the NAP spectra over that of the cryo-XPS.The authors propose that this difference arises from the presence of adventitious carbon on the surface of the NAP sample which was not present in its cryogenic counterpart because of the sublimation of the condensed water layers.

Ion scattering spectroscopy
Ion scattering spectroscopy (ISS), also known as low energy ion scattering (LEIS), is a surface technique which involves probing the outermost surface layers of a solid using elastic collisions of ions.While ISS is generally be used to provide surface identities and ratios, it may also be used in a semiquantitative manner through comparison with standard samples. 155][159] It was the discovery that ion-surface collisions could be described kinematically by a simple two body elastic collision model that allowed for facile determination of surface composition: The dependence of peak energy position on atomic mass in ISS.E 0 = primary ion E k , E 1 = primary ion E k following a scattering process at y1 to initial trajectory, m 1 = mass of primary ion, m 2 = mass of target atom.
][201][202]207,208 While ISS remains a valuable tool in materials analysis, care needs to be taken over sample choice and preparation. For exaple, 'rough' surfaces such as silicas commonly used in heterogeneous catalysis may exhibit broadened peaks in eventual ISS profiles. 209 To overcome this, compactin samples into pellets may improve surface suitability, however, care must be taken so as not to alter the material properties.210 Furthermore, ion scattering may have limitations depending on sample composition, since preferential sputtering can lead to erroneously low concentrations of easily sputtered components such as sodium.211 Fig. 10 A comparison of C 1s regions obtained by NAP and cryo-XPS evidencing the benefits cryo-XPS may bring to the analysis of materials in a traditional UHV atmosphere.These graphs show the application of multivariate curve resolution analysis to the spectra.154

Morphological analysis
Owing to its enhanced surface sensitivity in comparison to XPS, which has an information depth of as much as 9-10 nm (using a typical Al ka X-ray source), 212 ISS has found much use in extracting qualitative information on elemental distributions in the very uppermost atomic layers of a sample.The ability to probe only a single monolayer allows for a more sophisticated analysis of the atomic arrangement at the surface and near surface.
This information is of critical concern in fields such as heterogeneous catalysis, wherein performance and activity are directed by surface interactions of reactants onto active sites.Ceria based mixed oxides for example, possess unique surface properties which render them highly useful in a variety of applications.Surface segregation of either dopant or ceria, determined by ISS, has enabled a deeper understanding of how different thermal treatments or synthesis methods affect the resultant materials properties as catalysts for CO/alkane oxidations 164,180,181,213 and dehydrogenation/hydrogen transfer 214 as well as for oxygen ion conductors/oxygen storage. 181,215urthermore, since this is in effect an ion bombardment technique, the correct experimental parameters may induce sputtering of the surface of a material (even at low accelerating voltages), successive ISS spectra may therefore reveal an elemental gradient, affording further information which with knowledge of sputtering yields (or in comparison with alternative methods of probing the depth distribution of the outer surface) may even be used to prepare an elemental depth profile (Fig. 11). 215,216hin film development requires methodologies to determine surface coverage and monolayer build-up.ISS is uniquely suited to measuring a relatively large field area to probe full thin film coverage or detect remaining substrate signal. 177For example, TiCrN thin films for drill coatings, developed using DC magnetron sputtering exhibit different properties when manufactured at different pulse frequencies.Analysis by XPS and ISS enabled the link between surface construction (and in particular, surface segregation of different species) and material properties to be determined. 198Surface morphology probed by ISS may be advantageous, compared with purely XPS methodologies such as Tougaard's inelastic background modelling since it is free from any contributions from subsurface species (for example, underlying pores containing materials of competing photoemission).The caveat, however, is that multiple samples/measurements are often required for meaningful representation.

Structural changes following reaction
ISS may also be used to determine changes to the outermost surface following chemical or physical processing and has been used to study systems undergoing many common processes such as thermal treatments, [217][218][219] hydrogen reduction 161,220 or exposure to gases or vapours such as O 2 and H 2 O. 221 ISS has also been used to study sintering, an important consideration in the field of catalysis, both in terms of the stability and reusability of the catalytic materials but also in the preparation of the materials.Vapour deposition, 222 wet impregnation 223 and even cluster deposition 224 are often followed by a thermal treatment step to nucleate particles or remove organics, counterions or capping agents.Changes to surface coverage may give an insight into sintering processes which are difficult to observe by alternative means such as XRD or STM. 170This principle applies to monitoring synthesis procedures to optimise conditions.For example, determining the content of transition metal sites within a PtCo electrocatalyst/ polymer electrolyte membrane fuel cell.Marked performance changes are observed depending on the number of transition metal sites within the surface layer.Thermal processing and ISS/ XPS was used to optimise preparation methods and monitor surface properties in combination with catalytic studies. 217SS is also a viable method for monitoring the surface of a system during exposure to working conditions.For example, alloy oxidation is a key parameter to optimal material properties, ISS/XPS confirmed a Grimley-modified Cabrera-Mott model for the oxidation of a multitude of mixed metal oxides during plasma oxidation, as well as using ISS etching to profile the distribution of elements. 205Similarly, in catalysis, Ag foils Fig. 11 (a) Te : Cd ratio from a sequence of ARXPS measurements compared with (b) a profile of sequential 5 keV Ne + ISS measurements as a comparison probe of the two methods efficacy in the analysis of ultrathin films. 216xposed to NO/O 2 mixtures to model the catalytic process of NO x removal showed structural reorganisation, forming oxidic and nitric sandwich layers. 193Pt based automotive three-way catalysts were studied by ISS to monitor coking under different conditions the enhanced surface sensitivity of ISS in comparison to XPS enabling the analysis of catalysts with active component loadings below the XPS limit of B0.5 wt%. 167mperature dependant ISS (TD-ISS) ISS has found use in several sectors in conjunction with variable heating stages in order to utilise the unique information available to ISS as a probe for surface changes resulting from an increase in system temperature.
In one example of this approach, size selective Ir clusters deposited onto TiO 2 surfaces for model catalytic studies made use of a very simplistic form of TD-ISS in which model surfaces were analysed by XPS and ISS before and after a the temperature programmed desorption (TPD) of CO in order to assess surface restructuring. 160It was observed that following a TPD procedure, the surface Ir intensity (ISS) greatly diminished.
While XP data indicated that Ir desorption was not the likely cause of this attenuation, some attenuation suggested that perhaps the composite was restructuring either via sintering or encapsulation of the Ir centres.A second TPD was performed and revealed markedly differing behaviour, which combined with the ISS and XPS data suggested the Ir centres were undergoing encapsulation by the reducible support (a strongmetal support interaction which has been observed for a number of metal-titania systems). 166,225,226onitoring dynamic changes to ISS spectra as an in situ method of determining temperature-structure relationships, rather than an ex situ pre-and post-process snapshot, can offer even greater insights.For example, following exposure to CO, small (Pd 2 ) and large (Pd 20 ) clusters on TiO 2 were analysed by ISS whilst heating from 100 K to 500 K (Fig. 12).
Comparisons of these changing ISS ratios with CO TPDs led to the conclusion that low temperature CO desorption in the TPD was due to CO adsorbed onto Pd-TiO 2 peripheral sites, rather than directly on-top of Pd clusters.A similar experimental procedure has been used to study other Pd cluster/metal oxide systems 172,173 as well as the relationships between Pt cluster sizes and ethylene 227 and diborane 228 surface reactions.

Modelling growth modes
When considering nanoparticle or thin-film growth over a substrate (of non-matching identities), knowledge of nucleation and growth modes permits sophisticated and precise control over nanomaterial synthesis.ISS analysis combined with XPS affords a simple methodology for modelling the growth based on the calculated surface coverage or peak intensity relative ratios. 189It is important to note, that the related higher energy technique, RBS, provides beneficial complementary here in that it may probe film thicknesses up to B200 nm due to the deeper probing depth compared with ISS. 229The efficacy of RBS, however, diminishes with smaller overlayer thicknesses (o3 nm) however, and as such ISS is much more suited for applications such as modelling growth modes on the atomic scale.
2]230 ISS was also used to shed light on the nucleation and growth of iron oxides on Au (111), in which the rate of change of surface coverage with calculated monolayer coverage provided insights into a change in growth mode at 0.5 ML from layer-by-layer to island growth. 196STM was used alongside the XPS and ISS to confirm these observations.More complex systems, involving mixed oxide interfacial layers of multiple metal site systems may make use of the combined power of ISS analysis and subsequent modelling, with the chemical information gleaned from XPS (particularly angle resolved XPS) to construct sophisticated models of changing interfacial species and growth modes.When modelling the deposition of titanium onto Si(100) Palacio and Arranz measured the Ti and Si ISS signals at regular time intervals during the deposition process and used them to rule out Volmer-Weber growth during the initial stages. 195Combining ISS with XPS, they were also able to determine that the process involved a period of titanium silicate growth (up to 4 ML) followed by Stranski-Krastanov growth of metallic Ti.
Similarly, Liu et al. analysed platinum deposition on nickel using ISS and XPS (plus the application of overlayer models to the XPS data) to develop a model for the alloying and growth of the adsorbate layer. 231u/CeO 2 catalysts are widely used for oxidations, water-gas shift and methanol synthesis.Measuring the ISS spectra of a range of depositions of discrete amounts of Cu onto a Ce (111) surface and normalising the signal to that of bulk (440 ML) Cu/CeO 2 resulted in the profiles observed in Fig. 13. 169heoretical profiles for layer-by-layer (Frank-van der Merwe) and hemispherical cap growth (island growth of uniform hemispherical particles) 232 are superimposed onto the measured datapoints to illustrate the correct growth formation.Combination of the surface coverage from ISS with the XPS data and knowledge of the growth mode (and parameters such as IMFP) allows for high sophisticated calculations of particle thickness/radius. 232

Measuring surface potential
Ion-induced secondary electron (IISE) emission is an important material property to be considered in the production of coatings for scientific equipment, such as non-evaporable getters for distributed pumping at CERN. 233 ISS has been used as an effective probe for determining the surface potential of thin films under ion irradiation.Nagatomi et al. derived an experimental approach for the study of MgO films on silica (at the time an important material composite for plasma screen displays) in which irradiation of the surface with 950 eV He + ions with an electron flood gun was used to determine the resulting surface potential. 234,235

Ultraviolet photoelectron spectroscopy (UPS)
The process giving rise to UPS is in essence identical to that of XPS.The only difference is the energy of the incoming photons and thus the source of the photoelectrons that are emitted.Typically, UPS is performed using a He(I) or He(II) lamp emissions (21.22 and 40.8 eV respectively), relating to 1s 2p -1s 2 relaxations from He atoms (He(I)) and ions (He(II)). 236These lower energy sources limits this technique to the analysis of valence shell electrons and molecular orbitals, or density-ofstates (DOS) structure, and also results in a shallower sampling depth (due to the reduced kinetic energy of emitted photoelectrons).7][238] It is particularly powerful in the study of molecules in the gas phase where multiple emission states may arise as a result of vibrational lines on top of purely energetic transitions but UPS is also a powerful probe of solid and composite materials, occasionally as a simple fingerprint technique or through more detailed understanding of the emissions.UPS has found use in the analysis of polymers, 239,240 self-assembled monolayers (SAMs), 241 catalysts [242][243][244] and semiconductors/electronic materials. 245,246lence band structure A simple application of UPS is as a tool to fingerprint valence structure to obtain information either not available using X-rays or to glean information from closer to the surface.In the case of metal-semi/non-conductor systems (prevalent in the field of catalysis), emissions near the Fermi energy allow monitoring of the presence of metallic bonding.For example, Roberts et al. used the presence of photoemission intensity near E F to discern the amount of metallic bonding in the deposition of size selective clusters of Pd on TiO 2 and correlate high catalytic activity with higher Pd 4d occupancy (vs.Pd 5s). 247pectral intensity (or lack thereof) around the Fermi energy of MoO 3 based catalysts for alkane/ene cracking was used to monitor changes to the molybdenum centres following hydrogen reduction. 161Prior to reduction, no structure near the Fermi level could be observed, however following thermal treatment under H 2 at 673 K, p and s bands from MoO 2 could be observed in the DOS as well as a clear emission around the Fermi energy.While the XPS did indeed show a marked reduction from MoO 3 to a mixed MoO 3 -MoO 2 system, no metallic Mo could be observed.The UPS however, indicated that the system was largely MoO 2 with metallic sites present, suggesting these species may be found predominantly at the surface.UPS and XPS were further used to determine the degree of reduction following doping with Pt and K where it was revealed that K doping enables the full reduction of MoO 3 to MoO 2 .Such utilisation of UPS is often hindered by strong overlapping signals in mixed or complex systems, however small windows in the spectra can provide critical insights in many structures.

Electronic structure
Possibly the most widely used application of UPS is the analysis of electronic states (for example in semiconductors).This requires slightly more detailed analysis and understanding of the photoemission in a system than the simple fingerprint approach.
Since UPS concerns itself with the energy region associated with valence orbitals, it is primarily analysing the electrons concerned when discussing DOS.XPS may also be used to record these of course, but with the penalty of reduced resolution due to the greater line width of the exciting photons sources. 248hile not often a requirement for pure metals, the enhanced resolution allows for more detailed analysis of composite materials where determination of orbital characteristics in the DOS may be more complicated. 249here are a number of terms which are vital to understanding molecular electronics including, but not limited to: the vacuum level (E VAC -the energy level at which an electron is at rest a few nm outside of the atom), the Fermi energy (E F -in metals this is the boundary between occupied and unoccupied states in the continuum, in semiconductors this is a statistical state within the band gap) and work function (the energy required to take an electron from E F to E VAC ). 250Understanding the work function is key to improving our knowledge of semiconductors and developing improved materials.While it is possible to measure work function using XPS, UPS is traditionally favoured due to its superior line width and flux. 251ample work function may be determined using the spectral width (this is the energy difference between the secondary electron cut-off and the Fermi energy), 252 though while this process may be simple for conductors, it is less straightforward for semiconductors because E F lies within the band gap.Analysis of the high kinetic energy edge permits determination of the Fermi edge of a material, while the low kinetic energy edge provides the position of the secondary electron cut-off.Sample biases may also be applied in order to decouple emissions arising from secondary electrons associated with the spectrometer. 253It should also be noted, that while this review does not intend to dive deeply into (N)AP-XPS, powerful methodologies have been determined for the determination of material work functions through the analysis of gas phase molecules in the vicinity of the material surface. 254While this analysis requires the use of a (N)AP-XPS system -it serves as an additional probe and may be a powerful combinatorial approach with UPS, enabling the more facile determination of the valence band maximum (VBM) of semiconductors and assisting with the interpretation of the UP spectrum.
][257] TiO 2 represents a major class of semiconducting solids in which the understanding of the electronic structure of the pure material, as well as modified, doped and mixed oxide systems is important for the development of advanced functional materials in fields such as photocatalysis, 258 solar cells 259 and gas sensors. 260s a result it has been studied extensively using UPS with particular interest in the calculation of absolute values of the CBM and VBM when for example, Tauc-plots. 261generated from UV-Vis spectroscopy permits only relative values.Experimentally, recording UPS spectra from non-or semiconducting samples represents a challenge due to a more demanding need for sample homogeneity and electrical contact with the system.Maheu et al. compared two preparation methods which have shown some promise as means of acquiring UPS spectra from granular semiconducting materials: electrophoretic deposition (EPD) and drop casting a suspension of TiO 2 in alkanes. 262The objective was to study anatase, rutile and P-25 (a mixture of rutile and anatase often used as a benchmark material for photocatalytic applications due to it's beneficial electronic properties) in an effort to understand the nature of the unusually high photoactivity of P-25 when compared with other forms of titania.
While EPD prepared a sample in which electrical contact was sufficient to overcome any potential charging, drop casting required the determination of the absolute position of the CBM and VBM energies using the signal from the Ag foil substrate.Both methods were able to determine the E VB of anatase and rutile from which an energy level diagram and representation of DOS was developed (Fig. 14), correlating well with theoretical predictions. 263he inclusion of heating stages in many modern spectrometers raises the possibility of combining procedures such as those demonstrated by Maheu et al. with annealing processes capable of producing oxygen defects in TiO 2 with the ability to effectively neutralise charge build up and further augment the spectral acquisition process. 264Note this method makes use of a modified background developed by Li et al, designed and optimised for UPS spectra. 265hilst the ability to calculate the work function of a material is vital to semiconductors and electronics, 266 its also very important to chemical performance in many other areas including catalysis, 267,268 photochemistry 269,270 and chemical sensors. 271,272For example, work function changes were correlated with molecular dipole moment measurements of partially fluorinated alkyl and aromatic phosphonate SAMs on a Zn-terminated ZnO(0001) surface to create a tuneable ZnO work function that could be tailored to improve the efficiency of organic-inorganic electronic devices (Fig. 15). 273 some cases, work function measurements can also reveal insights into growth modes.Monitoring the work function of Ag deposited onto MgO for example, revealed a valley-like gradient which was ascribed to Ag atoms binding to surface defects and forming sites of nucleation for further growth. 274rganic semiconductors have gained popularity in the 21st century, wherein widely conjugated p-systems take the role of the solid system. 275AR-UPS has been able to evidence the electronic structure and charge carrier ability of such systems, though experimental methods are often not so straightforward, requiring organic single crystals which may need extra consideration for charge compensation and supplementary excitation sources. 276,277The result of the utilisation of AR-UPS, however, is that a picture of the electronic structure of the HOMO-derived valence band may be derived, which is fundamental to understanding the nature of positive charge carriers in organic semiconductors. 277sorbate analysis UPS analysis of molecular adsorbates has for a long time been the preserve mainly of the heterogeneous catalysis community, but the advent of hybrid organic-inorganic systems has resulted in a much wider interest in the interaction between substrate and adsorbate as materials designed for electronics, sensors and optics benefit from an increased understanding of the bonding and interfaces involved. 278O adsorption was an early model for this field, since metal-CO interactions are prevalent among a multitude of chemical catalytic transformations.CO adsorbed onto Pd(111) produced 2 of the three spectral emissions associated with the gas phase spectrum of CO (namely, the 1p and 5s features).279 The 5s is primarily situated on the carbon, the 1p on the oxygen and the 4s an oxygen lone pair.Due to final-state screening even in the absence of any bonding 280,281 adsorbate binding energies will decrease by 2-3 eV which led to the proposed model that the 1p and 5s orbitals are non-bonding and the 4s orbital increases in energy (due to the interaction with the metal) to the point where it is no longer observable on the UPS spectrum.This was observed to be true for a number of systems 282 and led to the explanation that the CO 5s orbital is responsible for the bonding of this system via the carbon atom, and donates electron density to the vacant d-orbitals of the metal and subsequent back-donation from filled d-orbitals is directed into unoccupied CO 2p* antibonding orbitals, a model which would come to be known as the Dewar-Chatt-Duncanson (DCD) model.
The 5s peak can actually be found underneath the 1p peak using angle resolved UPS, 283 while inverse photoemission spectroscopy was used to identify the partially filled 2p* antibonding orbital. 284PS was also used to study the effect of potassium and caesium doping of Pt(111) (plus O 2 preadsorption) on the adsorption of ethylene. 285Firstly, ethylene was adsorbed (p-bonded -mild disruption to double bond) on the surface of the crystal at 37 K so as to avoid any double-bond dissociation before heating to 57 and 300 K which resulted in the formation   273 of a di-s species (major disruption to the double bond followed by the formation of an ethylidyne molecule on the Pt surface.Multilayer formation at 37 K was observed in the presence of alkali-dopants, with minimal distortions to the adsorbed molecular orbitals, however the introduction of halides, resulted in a more weakly bound species with UPS revealing that with increasing alkali coverage there were fewer di-s species formed at higher temperatures, to the point where p-ethylene is the only adsorbate (desorbing at 130 K). 286,287 This analysis is not limited to observing changes in the adsorbate itself, however, as electronic changes induced by the process of adsorption may also produce simultaneous spectral features which provide insight into surface reaction mechanics.6][287][288][289][290] For a clean Pt (111) surface, ethylene adsorption reduces the work function (j) by 1.05 eV, consistent with the adsorbate UPS and vibration spectroscopy measurements in observing a large charge transfer from ethylene to Pt(111) upon forming the di-s species. 287dsorption of potassium onto Pt(111) reduces j by 4.8 eV, after which adsorption of ethylene induces very little change in j, again consistent with the observation from UPS/TPD that alkali doping of Pt( 111) inhibits s-bonding of ethylene to the metal surface.In contrast, CO adsorbed onto Pt (111) increases the work function by 1.5 eV, 291 consistent with metal to adsorbate charge transfer in the form of p-back bonding.[294][295][296][297] Monolayer orientation UPS has been used to reveal information regarding the orientation and organization of self-assembled alkyl chains adsorbed on to metal films, often in conjunction with XPS in order to probe monolayer thickness. 298Chain tilt angles for alkanethiols on metal facets (e.g.Ag(111) vs. Au(111)) differ significantly, and this difference was noted in the UPS spectra. 299,300It was observed that not only can UPS identify adsorbates in different orientations (e.g.standing vs. lying chains), but it can also identify them independently, unlike techniques such as Fourier transformed infrared (FT-IR) and near-edge X-ray absorption (NEXAFS) spectroscopies which report an average tilt-angle. 301Measurements recorded at differing take off angles (TOA) may also yield information regarding the degree of homogeneity in the orientation of the chains, 302 where spectral features which do not change as a function of TOA are indicative of randomly oriented chains while TOA dependence suggests a good molecular orientation. 303Furthermore, the intensity of the Fermi edge as a function of TOA may yield additional insights into the order of the adsorbates. 304This technique has found use with more complex surface molecules, such as 2-and 4-chlorobenzylmercaptan, adsorbed on gold films.Identification of spectral features from model gas phase UPS spectra permitted the identification of both the ortho-and para-positions of the benzene ring as well as the orientation of the molecules, although this technique is often performed using a related technique called metastable induced emission spectroscopy (MIES), due to its superior surface sensitivity. 305

Auger electron spectroscopy
Auger electron spectroscopy (AES) is a type of secondary electron emission spectroscopy, resulting from the bombardment of a surface by an electron beam.AES complements XPS and ISS through it's highly surface sensitive probing (B5 atomic layers depending on energy) and is capable of extremely specific elemental identification. 306Furthermore, unlike ISS, it is capable of determining chemical or valence states in certain cases. 307In the case of carbon, for example, analysis of the KLL Auger peak (involving valence electrons) may provide more chemical information than even XPS. 307uger emission is a widely known process and occurs as an inter-atomic response to the photoemission process caused from the irradiation by X-rays or electrons.Following emission of an electron, the atom is left in an ionised, energetically unfavourable state with a core-hole formed in a low-lying orbital.Relaxation for this state comes as an electron in a higher orbital may drop down to fill the core-hole, releasing excess energy which may then facilitate the emission of a secondary electron, the ''Auger'' electron, with a kinetic energy that is independent of the exciting radiation/electron.

Analysis of films and overlayers
The enhanced surface sensitivity of AES (when applied to labbased sources) results in a lower window for film thickness analysis than for XPS for example, with its suitability limited to ultrathin layers only (though it may be possible to analyse even up to 100 Å in some cases depending on the element). 308Much as with XPS, this surface sensitivity may be used to produce depth profiles in combination with a destructive ion bombardment technique -though with highly surface sensitive measurements, care must be taken to avoid topological changes which can affect the resulting profile. 309n example of Auger spectroscopy being used in this way is for the determination of graphene layer thickness.Multilayer graphene thickness is commonly determined by Raman spectroscopy, 310 however, there are major drawbacks with using Raman from complications arising from graphene on substrates that negatively affect the Raman signal (e.g.Ni 311 and GaAs 312 ).The use of AES avoids these issues.Importantly, it is also possible to determine the thickness or number of graphene layers from a single, non-destructive spectrumproviding an advantage over most XPS methods and may also determine crucial sample properties such as electronic coupling between graphene and substrate, observe structural defects and quantify dopants or chemical impurities. 313or adsorbates, monolayers and multilayers the analysis of AES peak maxima of substrate (I S ) and adsorbate (I a ), can readily probe adsorbate growth modes and determine monolayer coverage even if sticking probability changes with surface coverage. 314Such a technique enables the calibration of monolayer coverage and subsequent calculation of a coefficient of attenuation, which may be used to determine the surface coverage of an unknown sample (e.g.carbon monolayer coverage on a spent catalyst). 315Furthermore, a simple graphical model of I S /I a allows for the facile distinction between Volmer-Weber, Stranski-Krastanov and Frank-van der Merwe growth of films. 316

Surface mapping
Since AES utilises an exciting beam of electrons, rather than X-rays, there exists a much finer spatial resolution on the area of illumination than for traditional XPS.This enhancement permits much more detailed surface imaging than XPS imaging, with image resolutions capable of reaching 10 s of nm, rather than 1000 s. 317 Modern spectrometers often combine the traditional XPS configuration with a field emission gun and provide both AES map scanning and scanning emission microscopy into one unit. 318This combination facilitates the simultaneous acquisition of a morphological/topological image with that of a chemical surface map, which may be superimposed upon one another (Fig. 16).

X-ray generation of auger electrons
The scope of this review has so far focused on AES using an electron gun as the exciting beam, however X-rays are capable of producing augers too and the generation of X-ray induced auger electrons represents an invaluable tool for the chemical analysis of several key systems.Auger peaks often provide a fingerprint method by which to identify chemical compounds by which XPS may not be possible or straightforward (such as some first row transition metal compounds). 319The classic example of the importance of auger electrons in XPS is the speciation of copper.Copper metal and copper(I) are virtually indistinguishable upon analysis of the Cu 2p photoemission, however the Cu LMM auger produces a difference of around 2 eV, facilitating the chemical distinction of the sample. 320This may be taken further with the use of the modified auger parameter (a 0 ), 321,322 which is defined as the product of the kinetic energy of the auger emission and the binding energy of the primary photoelectron peak (e.g.Cu LMM E k + Cu 2p E B ).This parameter is advantageous, since it is independent of both exciting energy and any calibration factors applied to nonconductive material data.When working with the modified auger parameter, it is often beneficial to refer to, or produce, a Wagner plot 322,323 -a graphical means by which to visualise trends in a 0 across a series of compounds.
In the analysis of nanomaterials, a 0 may provide even more insight into the surface chemistry and inter-particulate interactions. 324,325Particle size effects and metal-support interactions between supported metal nanoparticles and oxidic supports have been observed to influence the resultant a 0 .Since a 0 depends on initial and final state effects, an estimate of the relaxation energy or degree of screening in the presence of core holes may be extracted from changes to a 0 . 327An increase to a 0 is associated with a higher relaxation energy/screening energy -which may be correlated with increasing particle sizes (and subsequently an increase in the average number of neighbouring atoms capable of screening the core-hole). 324Metal-support interactions may also influence this parameter, with more polarizable supports enhancing the resultant a 0 . 328

EELS & REELS
Electron energy loss spectroscopy (EELS) involves a beam of monoenergetic electrons, (typically with kinetic energies o2 keV), incident with a sample of interest.Inelastic scattering collisions occur from which information about the electronic state of the surface can be deduced.
In reflection electron energy loss spectroscopy (REELS), both elastically and inelastically scattered electrons are collected.The elastically scattered electrons mostly reflect directly from  326 the surface without losing any energy, forming the most intense part of the REELS spectrum which defines the 0 eV position on the energy loss scale.The low-energy loss part of the spectrum (o50 eV) can then be used to determine information about the band structure and dielectric properties of the sample.
The high-energy loss region (generally 450 eV) contains the ionisation edges due to inner shell ionisation, whose features are equivalent of an absorption edge in X-ray absorption spectroscopy (XAS).They are characteristic of the elements present in the sample surface (ca.15 nm) and their local bonding environment.Therefore, the information from EELS is complementary to XPS and other surface sensitive techniques.

Probing surface valence and conduction band electronic properties
EELS is particularly sensitive to the valence region of a solid surface promoting electrons from the occupied valence band to the unoccupied conduction band and as a result EELS is capable of identifying chemical states that may not be determined through other methods.
For instance, SnO and SnO 2 have similar binding energies in XP spectra.][334] Non-destructive depth profile information may also be obtained by the simple process of tuning the primary electron beam energy.Such information is crucial for understanding the mechanisms of various surface modification processes.In the example of EELS analysis of the oxidation of tin surfaces, it was revealed that the oxidation process underwent a threedimensional growth mechanism within the subsurface layers, while the more oxidic species were formed at the metal-oxide interface. 333,334 major drawback in the use of EELS is that, for many of the most electronically relevant solids (e.g.mixed oxides/perovskites), the EELS spectra are too complex to interpret.Typically, the intensities and the energy positions of EELS features from interband transitions are determined by the dipole selection rules for single-particle excitations.It also depends on the joint density of filled and unfilled states of the solid. 335However, with the presence of delocalised electrons in a polycrystalline sample, a combination of dipole and non-dipole active transitions in the EELS spectra may occur.For accurate interpretation, EELS then requires a detailed knowledge of the electronic structure within the solid, prior to unravelling the various components of the resultant spectra.
EELS can provide a useful insight into the interaction between probe molecules and a model surface and, like many surface science techniques has found much use, within the field of catalysis as a result.One system to have been investigated thoroughly is that of Ni and NiO which are strongly catalytically active (yet cheap and abundant) materials; EELS studies have been made of the adsorption of various gases, such as O 2 , 336-339 CO, 336,[340][341][342] H 2 , 340,341,343 H 2 O, 344 N 2 [345][346][347] and NO, 348

on clean
Ni and on Ni alloyed with, or following the deposition of, other metals, such as Al, 349,350 Cu, 351,352 Fe, 353,354 Mo, 355 and the alkali metals (Na, and Cs). 356,357In the case of NiO, both EELS studies of single crystalline 336,[358][359][360][361][362][363][364] and polycrystalline NiO [365][366][367] have been carried out and EELS has also been used to monitor changes to other material properties of NiO, such as temperature dependence 368,369 and optical behaviours. 370,371owever, it becomes apparent from a survey of the available data that the literature lacks consistency of peak assignments and interpretations for the Ni and NiO EELS spectra.Part of this could be due to the lack of detailed systematic investigation but can also be attributed to the need, as mentioned above, for an accurate knowledge of the electronic structure.Hoflund et al. have addressed that issue with an extensive study of the electronic structure of Ni and NiO both experimentally and theoretically. 372,373With the aid of this relevant information, Hoflund et al. were able to carry out the peak assignments of EELS spectra of Ni and NiO obtained from different depths of the samples with systematic alteration of the primary electron energy from 100 to 1150 eV. 335In this approach, the surface and bulk loss peaks can be more clearly distinguished, since an enhancement in the surface sensitivity is more noticeable than the surface loss features, which leads to the improvement of the EELS peak assignments.
Finally, of note is the use of EELS in the probing of plasmons, screened and unscreened final states -of great importance when confronted with the complex multiplet structures associated with transition metal photoemission spectra.As previously mentioned, EELS was used to study the electronic structure of Ni metal, however as well as purely evaluating EELS data of the metal, the combined use of EELS with XPS has resulted in the determination of peaks in the XPS spectrum too. 374Nickel metal, as well as producing a typically asymmetric photoemission, produces two additional broad peaks at +3.7 and +6.0 eV.While there had been some debate on the origin of the major emission at 6.0 eV, Grovesnor et al. used REELS to determine both surface and bulk plasmons as the primary origin of these emissions (though small shake-up features could not be discounted as contributing), and further observed a similar process within Co metal. 375Similar insights are available to more complex systems, such as metal oxides 376,377 and semiconductors. 378For example, REELS has been used in combination with XPS and theoretical calculations to probe OsO 2 where again, REELS was able to identify the presence and energy of a plasmon.This measurement contributed to the complete understanding of the electronic configuration through theoretical modelling of the density-of-states applied to the valence region of the photoelectron spectrum. 376

Analysis of inelastic scattering of electrons
As discussed at the beginning of this review, upon exiting a solid material, photoelectrons may undergo scattering processes which reduce their kinetic energies, contributing to asymmetric peaks, spectral backgrounds and satellite peaks. 38e have covered the importance of the analysis of overlayers and introduced several non-destructive techniques by which we can probe these such as the Tougaard background analysis method. 44One parameter which is of paramount importance in this analysis is an understanding of the inelastic background produced by the overlayer.While the QUASES software provides some elegant solutions to this, one opportunity provided by REELS is the ability to probe the inelastic scattering via electron beam.This may then be modelled to not only build a background model, but also provide additional insights into the surface chemistry.For example, TiO 2 overlayer formation on Ti metal was analysed using REELS to produce an inelastic background, 379 which was then applied to the Ti 2p region and the resultant background subtraction revealed weak suboxides which may otherwise have not been observed.

Band gap determination for semiconductor materials
Modern device construction often requires the controlled deposition of monolayers with desired properties one of which, that is often crucial for optimised device performance, is the band gap of the materials involved.However, the usual optical means (i.e. a Kubelka-Munk function on UV-Vis spectra) 261 often fail when a layer is thin or grown on a different substrate.For samples with a well-defined band gap, the REELS technique provides a convenient method of measurement in these cases since REELS' electrons must give up at least the energy equivalent to the band gap.Furthermore, while it is already a surface sensitive technique, decreasing the energy of the incoming electron beam can augment the surface sensitivity and hence, the band gap of thin films may be more accurately determined.
Vos et al. have used REELS under controlled conditions to extensively study a series of semiconductors with wellestablished band gap values. 380They discovered that, without surface preparation, it was possible to get a meaningful value of the band gap if a relatively high incoming energy of B5 keV was used.When the energy was below 1 keV however, the intensity of the REELS spectrum inside the gap is insignificant.This usually caused a fading of the onset, and resulted in an inability to record the band gap energy.
2][383][384] Transition metal oxides such as HfO 2 , ZrO 2 , and their compounds deposited on SiO 2 or Al 2 O 3 have been exploited as alternative high-k gate dielectrics [385][386][387] to avoid quantum tunnelling 388 during further scale reduction process of semiconductors.Kang and Tougaard and their colleagues obtained the band gap of HfZrO 4 gate oxide thin film and inelastic mean free path (IMFP) values through a quantitative analysis of REELS spectra. 383

Investigation of conjugation or aromaticity in organic materials
For materials with conjugated carbon bonding systems, REELS can provide complementary information to that obtained via XPS.An example is shown in Fig. 17, involving the organic light emitting diode (OLED) material poly(9,9-di-n-octylfluorenyl-2,7diyl), known as PFO, which was probed using XPS and REELS.The C 1s XPS spectrum shows two very weak peaks at a binding energy position slightly greater than the main peak.These are assigned as shake-up satellites promoting an electronic transition from HOMO (p) into LUMO (p*) of PFO.In the REELS spectrum the same two peaks are much more intense and can therefore be more easily identified.Quantification of sp 2 /sp 3 hybridisation of carbons may also be achieved by REELS, in cases where it is not always possible via other techniques such as XPS/AES. 390mparison of relative hydrogen content in polymer materials XPS has great difficulty detecting hydrogen (H 1s) because of its very low photoionisation cross section, its only really feasible with a synchrotron source.REELS however has the capability to detect and quantify hydrogen because of a process that resembles that of the ISS phenomenon.When the REELs electron strikes a hydrogen atom, the atom recoils because it is so much lighter than other elements and results in a discrete energy loss for the electron.This is demonstrated in Fig. 18    which REELS with a 1 keV beam energy was used to analyse a series of polymers.Strong features relating to conjugated carbon bonding are readily observed, however another set of peaks at around 1.8 eV from the primary elastic peak are also present and these are assigned to hydrogen, an assignment that is supported by the REELS spectrum for PTFE (which has no hydrogens), where this peak is not visible.
The relative intensity of the hydrogen peak to the primary elastic peak is proportional to the amount of hydrogen relative to other atoms (e.g.C, O and N).A simple peak deconvolution of the two peaks can therefore be used to quantify the hydrogen present.

Summary
In this review, we hope to have emphasised the great importance of utilising a multi-technique approach to materials analysis, in order to extract the maximum amount of information and enable a greater understanding of materials properties at play.Surface analysis of materials is of enormous importance in almost every aspect of science and technology and combinations of techniques can provide much more information than is often apparent through the utilisation of a single technique alone.Whilst XPS is arguably the most widely appreciated technique in this area, we have highlighted some of the many ways in which it may be used in conjunction with a broad range of complementary analyses to elucidate additional sample information, far more than simply the chemical nature of the material surface as is often the extent of the information gleaned in the general analysis of materials.Electronic, 2-dimensional, topological and structural information are readily obtained through the combination of multiple techniques associated with photoelectron spectroscopy.Manufacturers have developed multi-technique systems for many years now, largely focused on users in the surface science field and relying on knowledge and experience of working with complex UHV chambers and systems.The advent of current gen systems such as the Kratos Analytical Ltd AXIS Supra + and Thermo Scientific TM NEXSA TM G2 has signalled a pronounced shift towards the development of multifunctional XPS based instruments aimed at a more general user base.Particularly the introduction of automated sample handling, more robust software and data handling and easy switching between experimental modes represents a great opportunity for such techniques to really enter mainstream use across research laboratories of many fields.Whilst the relatively large investment will restrict the uptake somewhat, the reduced requirement for specialised skills and knowledge will certainly broaden the general userbase and increase the number of researchers looking to utilise XPS and multi-technique surface analysis.The field is even starting to see more niche general use systems, such as the SPECS GmbH EnviroESCA TM and Scienta Omicron GmbH HAXPES Lab -capable of user friendly NAP-XPS and HAXPES experiments respectively.While some manufacturers will remain focused on one or two applications, others seem to be moving in the direction of adding more and more functionality to individual instruments.As the demand for such systems grows, it is entirely possible we see an increase in the number of available add-ons to standard commercial instruments, for example Diffuse Reflectance UV-Vis (DRUVS), Attenuated Total Reflectance Infrared (ATR) and even traditionally synchrotron exclusive techniques such as Near-edge X-ray Absorption Fine Structure (NEXAFS) 392 as improvements to radiation sources and detectors are developed.The combination of these, in particular with the in operando capabilities of systems such as the lab-based NAP-XPS instruments -represents the potential for incredibly powerful analytical investigations to operate within the confines of even a small scale laboratory.

Fig. 1 X
Fig. 1 X-ray photoelectron survey spectrum of a clean silver surface.

Fig. 2
Fig. 2 (a) Angular dependence of the electron path length in ARXPS (vertical black arrows represent photoelectron pathway to the sample surface the escape depth of which, perpendicular to the sample surface, is dependent on IMFP and sample angle), (b) example of ARXPS applied to Hf oxide films on a silicon substrate in which dependence of silicon signal on angle of emission may be clearly observed and (c) relative depth profile calculated from the quantitatively processed spectra in (b).Reprinted (adapted) with permission from ref. 56.Copyright {2019} Kratos Analytical.

Fig. 4
Fig. 4 Typical inert sample handling devices.(a-c) Sample transfer under vacuum from Thermo Scientific TM .(a) Disassembled transfer unit, (b) assembled transfer unit and (c) check valve to hold internal vacuum -(d-f) sample transfer under vacuum or inert gas from Kratos Analytical.(d) Transfer arm attached to UHV chamber, (e) transfer arm removed from UHV chamber and (f) sample stub held on manipulator probe.

Fig. 5
Fig. 5 XPS spectra for the S 2p core peaks after SO 2 adsorption on (a) LiCoO 2 and (a 0 ) Al 2 O 3 coated LiCoO 2 and XPS spectra for the S 2p core peaks after SO 2 adsorption on (b) NMC and (b 0 ) Al 2 O 3 coated NMC.Reprinted (adapted) with permission from ref. 104.Copyright {2015} American Chemical Society.

Fig. 7
Fig. 7 In situ electrochemical XPS measurements as reported by Weingarth et al. 125 for their three electrode in situ XPS cell, performed using a VG ESCALAB 220iXL spectrometer equipped with a monochromatic Al Ka X-ray source.(a) The N 1s binding energy shift under control of the working electrode (polycrystalline platinum foil) potential (as indicated); (b) the dependence of the binding energies of the B 1s, C 1s and Pt 4f 7/2 core levels on the applied electrode potential.The system used a [EMIM][BF 4 ] ionic liquid electrolyte.

Fig. 13 (
Fig. 13 (a) Normalised ISS intensities for Cu on CeO 2(111) with model growth modes for comparison and (b) average Cu thickness versus Cu surface coverage.The dashed line represents the normalised ISS signal to be observed if Cu grew in a layer-by-layer fashion, whereas the solid line represents Cu growing as hemispherical caps with a fixed radius and a fixed particle density of 7.8 Â 10 12 particles per cm 2 .This model is only reasonable up to B35% surface coverage (above this, potential particulate overlaps render the model unreliable) and the dotted lines are purely for visualisation.Reprinted (adapted) with permission from ref.169.Copyright {2015} American Chemical Society169

Fig. 14
Fig.14He(I) UPS profiles, fit with an established empirical model and an overlayed electronic structure of the two TiO 2 polymorphs anatase and rutile relative to the vacuum level.E thr.corresponds to the deepest state that can be excited in a photocatalytic test (E photons r 4.3 eV).262

Fig. 16 (
Fig.16(a) Scanning electron microscopy (SEM) micrograph indicating region of analysis for corresponding AES spectrum (blue line), (b) resultant AES profiles providing chemical analysis along the aforementioned spatial dimension.326 in

Fig. 17
Fig. 17(a) Photoelectron spectrum of poly (9,9-dioctylfluorene) (PFO), including observation of weak shake-up peaks caused from inelastic collisions between emitted photoelectrons and p/p* HOMO/LUMO (inset) and (b) REELS profile from PFO in which p/p* shake-up satellites may be clearly observed and energies recorded.Polystyrene is included as a reference for position of phenyl p-p* transition.Reprinted (adapted) with permission from ref. 389.Copyright {2013} Thermo Scientific.
Fig. 17(a) Photoelectron spectrum of poly (9,9-dioctylfluorene) (PFO), including observation of weak shake-up peaks caused from inelastic collisions between emitted photoelectrons and p/p* HOMO/LUMO (inset) and (b) REELS profile from PFO in which p/p* shake-up satellites may be clearly observed and energies recorded.Polystyrene is included as a reference for position of phenyl p-p* transition.Reprinted (adapted) with permission from ref. 389.Copyright {2013} Thermo Scientific.