Peter G. Kenny

Scatter diagrams and hotelling transforms: application to surface analytical microscopy

Abstract The collection of electrons from several Auger peaks together with other electrons scattered from a sample in a scanning Auger electron microscope provides a means of analysing many features of an inhomogeneous surface. The use of multi-image sets in the analysis of samples of Al on Si, W on Si, LaNi compounds and superconducting wires are given as examples. The use of scatter diagrams as a means of classifying the number and coverage of different surface phases is illustrated. The Hotelling transform is described and then applied both to data from a real W/Si sample and to simulated images which are chosen to illustrate the ways in which this transform can be useful. The design of a new 23 channel Auger microscope intended to take advantage of the multi-spectral approach to data acquisition is outlined.

Three-dimensional scatter diagrams : application to surface analytical microscopy

Abstract Multivariate statistical techniques, such as principal component analysis and two-dimensional scatter diagrams, are increasingly being applied to multispectral images in analytical electron and X-ray microscopy. Combining data from different sources can often reveal more information than would be obtained if each image was processed separately. Three-dimensional data manipulation and visualisation tools are gaining popularity in many scientific fields. However, three-dimensional scatter diagrams are not yet being exploited much in analytical microscopy. This paper demonstrates the advantages of three-dimensional scatter diagrams and interactive correlation partitioning. An introduction to the methodology is included. Examples using multi-element Auger and topographic images from the York multispectral analytical electron microscope (MULSAM) are presented. Implementation issues are discussed briefly to help other workers who may wish to investigate the techniques. Some areas of future work are suggested.

Correction of topographical artefacts in Auger images using four backscattered electron detectors

Abstract The York multi-spectral scanning Auger microscope has been used to examine the correlations between Auger and backscattered electron images collected simultaneously from an electron spectrometer and the quadrants of an Si p-n junction backscattered electron detector. Digital signal processing of the four backscattered electron images allows the complete characterisation of the surface topography and also gives the effective atomic number at each pixel of the image. The topographical images thus calculated can then be used to correct the topographical artefacts in the Auger images. This method, which allows quantitative surface analysis of samples with complicated surface topographies with no prior knowledge of the sample, is tested using a carbon-coated anisotropically etched Si sample.

Multi-imaging and multivariate statistics used for 3D characterization at surfaces

Several microanalytical imaging techniques—energy dispersive X-ray detection, parallel electron energy loss spectroscopy, secondary ion mass spectroscopy, photoelectron spectroscopy and XPS) and scanning Auger microscopy—have reached the stage where they are capable of producing images of a surface with a section of a spectrum in each pixel. The resulting image-spectrum is a complex data structure which requires the use of special methodologies if the data are to be interpreted effectively. Appropriate methods have been developed for Earth satellite image processing and are directly applicable to surface microanalysis. The use of scatter diagrams, interactive correlation partitioning, factor and target factor analysis and principal component analysis are outlined in this paper and their application to semiconducting, catalytic and magnetic structures is illustrated. This field of endeavour can be thought of as being the beginning of an area of study which may be called surface chemometrics.

Automatic and interactive correlation partitioning compared : Application to TiN/Ti/SiO2

A new method has been developed for the partitioning of sets of images with the objective of automatically identifying the number and locations of different regions in a material. The method is called automatic correlation partitioning and it involves the identification of clusters in the n-dimensional intensity histogram of a set of n images that are spatially registered. The method uses the peaks located in the simple intensity histograms of each image in the set to produce a list of all possible clusters in the entire data set. This list is then searched in order to find the actual clusters. The method is tested using data from a multi-imaging Auger electron microscope, which yields sets of Auger images characteristic of the spatial distributions of selected kinds of atoms in the surface of a solid. The first tests involve the use of a model sample consisting of a W overlay pattern on a Si substrate. The second tests are done on a TiN/Ti/SiO2 planar layer structure that has been ion beam bevelled to reveal a cross-section of the composition depth profile. The first set contains two images and the second set contains five images. The results of the new automatic method are compared with those obtained by the analyst working interactively with the data set to identify the clusters subjectively. Cluster analysis of the second sample reveals details of the interfacial layer chemistry not revealed by the interactive method and is consistent with published XPS depth profiling experiments reporting a titanium silicide layer at the Ti/SiO2 interface.

Quantitative surface microanalysis of samples with extreme topography utilising image interpretation by scatter diagrams and principal component analysis

Abstract Quantitative surface analysis (by Auger imaging, SIMS, EPMA, etc.) of many samples of technological interest is not possible due to the image artefacts arising from surface topography or other features of the sample. A methodology is presented for the identification and removal of regions within an image where artefacts dominate the contrast. This enables meaningful quantification of the regions not dominated by artefacts. The method employs multi-variate statistical techniques including 3D scatter diagrams and principal component analysis (PCA). PCA proves to be a powerful method for measuring the extent of any remnant artefact within image sets. The methodology is applied to the characterisation of a Pt Rh catalyst using the multi-spectral scanning Auger microscope at York.

Data processing for spectrum-images : extracting information from the data mountain

Several analytical methods of electron, X-ray and ion microscopy are now capable of generating images containing an entire spectrum at each pixel. These spectrum-images require new software tools for visualisation, classification and analysis. This paper outlines a new method of visualisation for spectrum-images and two image classification methods that can be used to identify the phases present in the sample.

Visualization of Smoothed Particle Hydrodynamics for Astrophysics

Scientific visualization still presents a number of challenges. Effective visua lization straddles several problem domains - the data structures needed to support visualization of large data sets, rendering techniques for fast and interactive display of this data, and enough understanding of the data involv ed to construct visualizations that provide real insight into the problem. Data from Smoothed Particle Hydrody namics simulations is of particular interest, due to its time-dependent, point-based nature and its prevalence insimulation in astrophysics in areas such as star formation and evolution. This paper looks at some of the issue s associated with building a useful, usable visualization tool for SPH data from astrophysics, and describes a p rototype of such a system. This paper describes work in progress.

Exploratory simulation for astrophysics

Exploratory simulation involves the combination of computational steering and visualization at interactive speeds. This presents a number of challenges for large scientific data sets, such as those from astrophysics. A computational model is required such that steering the simulation while in progress is both physically valid and scientifically useful. Effective and appropriate visualization and feedback methods are needed to facilitate the discovery process. Smoothed Particle Hydrodynamics (SPH) techniques are of interest in the area of Computational Fluid Dynamics (CFD), notably for the simulation of astrophysical phenomena in areas such as star formation and evolution. This paper discusses the issues involved with creating an exploratory simulation environment for SPH. We introduce the concepts of painting and simulation trails as a novel solution to the competing concerns of interactivity and accuracy, and present a prototype of a system that implements these new ideas. This paper describes work in progress.

Simulation trails improve accuracy and efficiency in astrophysical simulations

Harnessing expert user insight into simulations gives a promising technique to reduce computational time and improve efficiency in astrophysical simulations.