This is a brief summary of the work that was involved in the development of the technique for quantitative XPS from analysis of the background of inelastically scattered electrons. About 30 years ago it became evident that these electrons must carry valuable information about the depth where the XPS electrons are excited. Theoretical modeling started and algorithms were developed. It was necessary to have an accurate description of the electron energy loss processes which at that time was not available. Theoretical calculations of inelastic cross sections based on a dielectric response description were done and a new experimental method to determine this from analysis of reflected electron energy loss spectra (REELS) was also developed.

To make these procedures for quantitative XPS analysis work in practice it is however not possible to use calculations valid only for specific sample compositions. Therefore an effort was made early on to find cross sections which can be used as an approximation for wide classes of materials and compositions. This resulted in the Universal cross sections which are now widely used and without which practical use of the formalism would have been very limited. The resulting XPS analysis technique was summarized in [1].

In the following years, much effort was then centered on applications to increasingly finer details of the morphology of nanostructures. This requires a careful data analysis since otherwise the uncertainty on the determined morphology may be large. Sometimes the detailed morphology is however not the most important issue for technological applications. Things like speed of analysis, robustness, and automation is often more important in industrial environments. It was therefore decided to develop a simpler algorithm which does not give as detailed information but which is very robust and therefore faster to use and less dependent on a meticulous analysis procedure. The resulting algorithm [2] has been shown to be very robust and therefore suitable for automation. It proved also effective in generating 3D images of nano-structures where automation is mandatory since thousands of spectra (one per pixel) must be analyzed.

Throughout, efforts were always exerted to test the validity of each step in the development of algorithms and procedures by designing and performing critical experiments. This is of utmost importance to ensure progress which does not lead to dead ends.

In this talk I will give an overview of the development of the technique and discuss some technological applications.

1. S. Tougaard, J. Vac. Sci. Technol. A14, 1415 (1996)

2. S. Tougaard, J. Vac. Sci. Technol. A23, 741 (2005)

1. http://www.nist.gov/srd/nist100.cfm.

2. W. Smekal, W. S. M. Werner, and C. J. Powell, Surf. Interface Anal. 37, 1059 (2005).

For each array, the intensity of the In, P, N and Si core levels, normalised with respect to that measured on the mask and InP open spaces, are compared to the corresponding ratio calculated from the geometry of the array and the analysis arrangement or configuration. Modelling takes into account the contribution of the various surfaces (mask, InP) in the line of sight of the analyser, and the rate of irradiation according to the geometry of the array and the nature of the materials (InP ridge, mask) the x-rays pass through. This comparison points out the relation between the intensity emitted from the bottom and the aspect ratio of the array. A good agreement is obtained when including the analyser acceptance angle to the model. Concerning the sidewall the discrepancy between experiment and simulation corroborates the presence of a passivation layer. The presentation will discuss in detail the influence of the plasma chemistry on the quantitative composition of the sidewall and the bottom.

Valence band X-ray photoelectron spectroscopy (XPS)¹ gives spectral features (peak positions and peak intensities) that arise from different physical principles than the core spectral region. This difference leads to the valence band region providing complimentary information to that of the core region. In many cases the valence band region can be used to detect subtle chemical differences that cannot be determined in core XPS studies. The value of using valence band XPS interpreted by calculation models will be demonstrated for various systems, and the use of core and valence band XPS for the study of buried interfaces will be discussed. Examples discussed will include the formation and study of thin (less than 100A) oxide-free phosphate films, polymer films, composite surfaces, and the identification of different oxide films (including aluminum oxides) with similar chemical composition. Studies of shallow buried interfaces will be discussed. Recent work involving the preparation of hydroxyapatite films formed on metals which were coated with a thin oxide free film of metal etidronate will be reported. The metals studied were stainless steel and titanium. The key to adhesion of the hydroxyapatite films is the initial formation of a thin, oxide free, etidronate film on the metal. It was not found possible to prepare the hydroxyapatite films directly on the metal surfaces. Since hydroxyapatite is a key component of bone and teeth, it is likely that the coated metals will have desirable biocompatible properties, and that these treated metals may find applications in the production of medical implants.

¹P..M.A. Sherwood, “XPS Valence Bands”, chapter in “Surface Analysis by Auger and X-ray Photoelectron

Spectroscopy” edited by D Briggs and J T Grant, SurfaceSpectra Ltd and IM Publications, Chapter 19, 531-555,

2003.

Optoelectronic devices, used for inter-conversion of light and electricity (e.g. photovoltaics and displayes), depend upon careful optimisation of chemical, electronic and structural properties for efficient operation and useful operating lifetime.

Characterisation of such a device will typically identify the chemical bonding states, valence band positions, band gap and work function for each component. Lateral and depth resolution may be required to evaluate cell/pixel and multilayer structures.

Electron spectroscopic surface analysis techniques are ideal for the detailed analysis of the electronic structures of optoelectronic devices. Such techniques allow full quantitative characterisation of materials with chemical state and structural information. Surface specificity of spectroscopic information ensures that thin films can be analysed without interference from deeper parts of the sample. Multilayer structures may be studied with depth profiling techniques, and imaging functionality may be used to study cell or pixel structures.

The Thermo Scientific Escalab250Xi offers several such spectroscopic techniques, which have been employed in this study. X-ray Photoelectron Spectroscopy (XPS) offers surface-sensitive, quantified chemical state analysis and imaging capabilities. Ultraviolet Photoelectron Spectroscopy (UPS) allows measurement of valence band positions. Reflected Electron Energy Loss Spectroscopy (REELS) yields information on the electronic band gap and hydrogen content of a material. These techniques are combined for a thorough characterisation of the electronic structure of optoelectronic devices.

A noncontact chemical and electrical measurement technique of XPS is performed to investigate a CdS based Photoresistor and a Si-Diode during their operation. The main aim of the technique is to trace chemical and location specified surface potential variations as shifts of the XPS peak positions under operating conditions. For the Photoresistor Cd3d and for the Diode (p-n junction) Si2p peaks positions have been recorded, respectively. The variations in the Cd3d peak without and under photoillumination with 4 different lasers is extracted to yield the location dependent resistance values, which are represented; (i) two dimensionally for line scans, and (ii) three dimensionally for area measurements. In both cases one of the dimensions is the binding energy. For the Si p-n junction the variations in the Si2p peak position under normal and reverse bias are recorded to differentiate and identify the nature of the doping (p- or n-). The main advantage of the technique is its ability to assess element-specific surface electrical potentials of devices under operation based on the energy deviation of core level peaks in surface domains/structures. Detection of the variations in electrical potentials and especially their responses to the energy of the illuminating source in operando, is also shown to be capable of detecting, locating, and identifying the chemical nature of structural and other types of defects.