Titanium nitride films were deposited using reactive magnetron deposition and studied in-situ using Xray Photoelectron Spectroscopy (XPS). The films were then analysed ex-situ using Extended X-ray Absorption Fine Structure (EXAFS) and X-ray diffraction (XRD). The films had a larger degree of disorder at high nitrogen content and this correlated with a more complex structure in the N1s peak as seen by XPS. No evidence for the postulated CaF2 structure for highly nitrided TiN films was found.

A new type of electrostatic electron energy analyzer is described that can acquire an electron energy spectrum in 'one shot.' It uses a hyperbolic field to focus electrons emitted from a solid in the energy range 50 eV to 2500 eV into a dispersive plane of about 50 mm length. An expression for the energy resolution is given and the effect of side and base plates on the behavior of the device is discussed. The main intended area of application for this type of analyzer is parallel data acquisition in Auger electron spectroscopy (AES) and X-ray photoelectron spectroscopy (XPS). The potential to acquire a spectrum is just a few seconds is possible with this device.

Synchrotron-excited ultraviolet photoemission spectroscopy (UPS) was used to acquire spectra at photon energies between 85 and 100 eV from titanium nitride. The TiN was sputter deposited at several compositions, and manoeuvred into the analysis position in situ. By ratioing the area of the unhybridised Ti 3d peak near the Fermi edge with the area of the Ti 3p core peak, an estimate of the 3d electron density can be made. Electron energy loss spectra acquired from the same samples show a low energy plasmon loss peak, the energy of which is dependent on sample composition. Three methods for determining the electron effective mass from the plasmon loss energy, the d electron band density and the d electron band depth as measured using UPS, are described. These methods show that the plasmon is screened and suggest that the effective electron mass increases with titanium content.

Samples of titanium nitride were magnetron sputter deposited and transferred in situ to be studied by X-ray photoelectron spectroscopy using high-energy CrKβ radiation and high-resolution AlKα radiation. The determination of the Auger parameter of titanium in titanium nitride using the Ti KLL Auger peak has been carried out at a series of compositions using the CrKβ radiation. The Auger parameter changes by a relatively small amount with composition showing that the extra-atomic relaxation in TiNx is similar to that of the metal, and suggests that charge transfer is largely responsible for the titanium core level peak shifts. The spectral structure of the N1s photoemission peak at high-energy resolution revealed the development of multiple nitrogen environments at high nitrogen concentration, suggesting vacancies on the titanium sublattice or a highly defective structure. The various titanium photoelectron and Auger peaks reveal satellites which can be explained either by two or more local environments for the titanium atoms in fcc titanium nitride, or by a screening mechanism giving two different peaks for equivalent titanium environments.

Examples of the evolution of the topography of selected materials (synthetic graphite, molybdenum disulphide and muscovite mica) as a function of reactive (in oxygen) and nonreactive (in argon) etching in a 13.56 MHz plasma are considered. The dependence of the observed topography and its regular and semi-regular characteristic details on the power levels, gas pressures and treatment times is demonstrated, particularly at sub-micron scales. The topographical images, derived roughness and other data (obtained variously using scanning tunnelling microscopy and atomic force microscopy) are discussed with respect to the layer lattice structure of the materials processed.

The nature of the electron state band in the amorphous metal alloys CuHf and CuZr has been studied using synchrotron radiation as a source in photoemission experiments. In the photon energy range - 45 eV the valence band undergoes resonant photoemission which emphasizes Hf 5d-like states in CuHf and Zr 4d-like states in CuZr, while at higher photon energies contributions from these states are strongly de-emphasized because of Cooper minima in their cross-sections. The ratio of the intensities of the valence band as compared with other peaks in the spectrum characteristic of Cu or Hf (Zr) shows that the near-Fermi-level region is not wholly Hf (Zr) d-like in character, but that the degree of hybridization of these states is less than in alloys such as FeZr where both elements have partially filled shells.

Samples of titanium nitride were grown in an Ar/N atmosphere with various Ar/N gas ratios using reactive DC magnetron deposition onto stainless steel and silicon substrates. The samples were transferred to the analysis position in-situ and were studied there using AES, EELS and synchrotron-radiation excited UPS.Using AES, the composition of the samples was found to vary with the Ar/N gas ratio, as expected. Changes in the hybridised Ti 3d/N 2p band seen via UPS agreed with the theoretical prediction that the band narrows as the nitrogen concentration is reduced. A large resonance in the Ti 3d band was observed at similar to 73 eV, which is probably associated with the Ti 3s level. The UPS and EELS data is consistent with vacancies being present on the titanium sublattice.The EELS data displayed low-energy loss peaks which shifted to higher loss energies as the nitrogen concentration in the film was lowered. These loss peaks can be explained by a screened Drude plasmon loss rather than by intraband transitions. The intensity of the loss peaks was much weaker than previously reported, with attendant implications for explanations of the shape of the Ti 2p photoelectron peak observed in XPS. (C) 1997 Elsevier Science B.V.

We have used electron spectroscopy to study amorphous CaAℓ alloys across the entire composition range. Synchrotron radiation photoemission was used to examine the region of the Aℓ 2p core level, with the energy of plasmon loss features giving the electron density around the Aℓ atoms, while energy loss spectroscopy excited by low energy electrons probed the average electron density in the alloys. The plasmon energy is found to vary continuously with composition. For alloys with low Ca concentrations (below ∼20 at%) there is reasonable agreement with expectations for free-electron-like plasmons with Ca and Aℓ having their elemental valencies. At higher Ca concentrations the Aℓ plasmon energy is larger than expected. Although this would be consistent with an increased electron density in the vicinity of the Aℓ atom, the plasmon energies are affected because part of the oscillator strength of the electron gas is associated with one-electron excitations around the Ca atoms. We argue that there is little charge transfer in the alloys, a result which agrees with our studies of the Auger parameter.

The first spin-resolved photoemission experiment on an iron - boron amorphous alloy using a synchrotron source is presented. The experimental spin polarization of the d band of has been compared with three theoretical predictions and found to be in best agreement with self-consistent spin-polarized calculations based on a supercell LMTO approach. The observed average spin polarization of the valence band is approximately twice that of 10 eV secondary electrons. Hysteresis loops for determined from the 1 eV and 20 eV secondary-electron asymmetry are similar to those determined using the magnetooptic Kerr effect, but show a lower coercivity. The differences are attributed to a combination of different sampling depths of the two techniques and to the surface inhomogeneity of the sample.

The compositional dependence of the binding energies and Auger parameters on Nb and Si in NbxSi1-x has been measured at high resolution using a Scienta spectrometer. Alloying increases the binding energy of Nb and decreases that of Si by up to 0.6 eV. The Auger parameter of Si increases on alloying, indicating enhanced screening in the amorphous silicide, but that of Nb is exceptionally sensitive to the presence of oxide at the surface.

Analysis of Auger parameter shifts in amorphous CuHf alloys of varying composition give insight into charge transfer and core hole screening, which is useful in understanding the anomalous electronic properties of these materials. The copper Auger parameter, which can be determined very precisely, increases upon addition of hafnium, but shifts in the hafnium Auger parameter are difficult to isolate due to peak overlap. However, Hf4f and Cu2p shift in opposite direction with composition variation, which, in conjunction with the Auger parameter data, suggests electron transfer from hafnium to copper.

Photoemission from a synchrotron source along with in situ resistivity measurement is used to study conduction hand states in amorphous CaxAl1−x and CaxMg1−y alloys. In the mid-composition range CaAl shows an anomalously broad Fermi edge, while CaMg behaves normally at all compositions. The 3p → 3d giant resonance in Ca leads to resonant photoemission from the conduction band of Ca metal but it greatly decreases in strength with increasing Al concentration.

Synchrotron radiation has been used to investigate resonant photo- emission in amorphous metal systems CuxZr100−x, CuxTi100−x and FexZr100−x. Constant initial state spectra reveal resonant enhancement of Zr 4d-like conduction band states near the Fermi level in Cu45Zr55 and 3d-like Ti states in Cu33Ti67 at np → nd phonon eexcitation energies; this is in contrast to the absence of resonance effects in a lower lying “filled” Cu 3d-like band. Observation of a Cooper minimum in the photoemission cross section in Cu45Zr55 gives further support to the importance of Zr 4d states near the Fermi edge. In contrast both conduction band features in Fe90Zr10 show Fe 3p → 3d resonant enhancement with some evidence of Zr 4p → 4d resonance. This is consistent with Cu45Zr55 and Cu33Ti67 exhibiting split band behaviour, while the d electrons in Fe90Zr10 appear to form a common band.

Films of amorphous NbSi prepared by magnetron sputtering were found to have a higher concentration of silicon at the surface than in the bulk material. In this paper we report on measurements of both ultraviolet photoemission spectroscopy and Auger electron spectroscopy carried out on samples prepared in situ at the UK Synchrotron Radiation Source, Daresbury. Spectra taken immediately after sputter deposition show that the surface is rich in silicon, while following argon ion bombardment the bulk composition is slowly revealed with a much greater concentration of niobium. This is observed by a comparison of the relative heights of peaks in the Auger spectra and by resonance effects and the sharpening of the Fermi edge in the UP spectra. Of particular interest are changes around the metal-insulator transition that occur at approximately 11 at.% niobium.

Ultraviolet photoemission experiments using radiation from a synchrotron source have been used to study details of the valence bands of amorhous NbxSi1-x on either side of the metal-insulator transition. The alloys were prepared in situ by RF magnetron sputtering with surface and bulk characterization obtained by Auger electron spectroscopy and EXAFS respectively. The variation of the intensities of several features in the UV spectra with photon energy are due to the presence of resonant photoemission and Cooper minima in photoionization, and these effects can be used to identify the atomic origins of the partial densities of states in different parts of the valence bands. In the metallic phase such measurements show 'pure' Nb 4d states close to the Fermi edge, whereas for the insulating material there is no Fermi edge and little evidence for the Nb resonance.

Synchrotron radiation has been used to investigate resonant photoemission in the amorphous metal systems CuTi, CuZr and CuHf. Constant initial state spectra reveal resonant enhancement at np --> nd photon excitation energies of conduction band states near the Fermi level; these are Ti3d-like states in CuTi, Zr4d-like states in CuZr, and Hf5d-like states in CuHf. In contrast there is an absence of such resonance effects in a lower lying "filled" Cu 3d-like band. The results are consistent with CuTi, CuZr and CuHf all exhibiting split band behaviour, with little hybridisation between Cu 3d and transition metal nd states.

The design and construction of an ultrahigh vacuum multi-imaging scanning electron microscope is described. The microscope is designed to contain two field electron emission columns and can acquire simultaneous digital images from a 16-channel electron spectrometer, a four-quadrant back-scattered electron (BSE) detector, an Si(Li) x-ray detector, a SEM detector and the current flowing to ground through the sample. Because there is exact spatial registration between corresponding pixels in each of the images, it is possible to use the image set to make quantitative interpretations of the surface and subsurface chemistry. This is done using mathematical manipulations of the image set, together with models for the SEM, BSE and Auger signals. Techniques are described for setting up the alignment and characterizing the field of view and transmission function of the microscope and its spectrometer. Examples of multi-imaging from simple samples are given. The close coupling between the microscope and its control and interpretation computers provides considerable power for the analysis of inhomogeneous surfaces.

A UHV compatible backscattered electron detector has been constructed using a four quadrant Si detector. Particular attention has been paid to the vapour pressures of the materials used, the residual magnetic properties after fabrication and the screening of any insulators. The result is a reliable and practical device for UHV operation.

AbstractA method for quantifying Auger spectra and including matrix effects is described. This method corrects iteratively for the effects of electron back‐scattering, the inelastic mean free path and atomic density in the sample using established methods. New algorithms have been developed to assist the user interactively with the analysis of spectra collected from surfaces covered by thin films of unknown composition or thickness, or by coverage of another material. Error analysis has been included for both homogeneous and thin film cases. A ‘figure of merit’, used to inform the user if a thin film model of the surface is accurate, is introduced. The algorithms can be implemented on a small microcomputer. Their application to quantitative Auger analysis by multi‐spectral imaging is also described. Each point in a set of images has been corrected for the factors that influence the Auger current to arrive at surface compositions at each pixel for each element. Some of the advantages and problems of quantifying Auger images will be outlined. The use of multi‐spectral Auger images is seen as a powerful means of identifying surface phases of materials and is given the acronym MULSAM—MULti‐Spectral‐Auger Mapping.