Integrated Textural and Mineralogical Data In Soil Forensic Analysis Based on Automated Mineralogy and Particle Classification.

Extensive and detailed casework and experimental experience has shown that the mineralogy of soils is typically very variable on a variety of scales (metres to kilometres).  This variability is not only a result of differences in the proportion of minerals but also the textural characteristics of the soils, sediments and dusts. For instance, minerals might be present as discrete mono-minerallic grains, as small poly-minerallic lithic grains (rock fragments) or as man-made materials such as concrete.  Routine mineralogical analysis using standard techniques such as XRD does not provide textural information, only the relative abundance of the minerals identified. A combination of techniques may be employed to derive textural and mineralogical information (such as optical analysis etc) but in many real-world cases, there is simply too little sample available for the full suite of analysis to be undertaken and a significant line of evidence is therefore potentially lost to the investigation.

We have been using automated mineralogy using QEMSCAN^® technology in criminal investigations (mostly murder cases) in the UK since 2003, but increasingly, rather than just utilising the modal mineralogy, we are using the ability to examine and quantify particles individually as "lithotypes".  A "lithotype" is effectively a compositional grouping based on the combination of minerals and their association within a particle; effectively what a geologist does when identifying a rock, but using an automated mineral analysis system.

In this talk we will demonstrate how soil mineralogy based on automated SEM-EDS analysis can either be reported as modal mineralogy or as "lithotypes".  This is based on an experimental study carried out in Cornwall, UK but will also be applied to casework scenarios.