The trade in counterfeit medicine and illegal recreational drugs remains prevalent along with the damaging consequences for consumers and society. Technology is sought which reliably detects illicit materials during security screening while causing minimal disruption to legal activity and without causing damage to traded goods. Energy dispersive x-ray diffraction (EDXRD) is a technology in an advanced stage of research and development which has shown potential for non-destructive, fast, and reliable screening. It has been demonstrated in experiments to identify illicit drugs concealed in courier parcels and to determine accurately the chemical composition of pharmaceutical formulations.The diffraction signals which result from EDXRD are fingerprints from which we deduce the chemical composition of a sample under investigation. However, detection of materials of interest is made difficult by signal contributions from other materials as well as from extraneous physical effects. Moreover, a high level of data noise is inherent in rapid screening systems, which further inhibits material detection. This research focuses on statistical methodology to identify the relevant materials from these complex signals and to develop an analytical framework for screening procedures which use EDXRD.
This research uses data from two contexts:
We model data from both EDXRD experiments within a multivariate analysis framework and generalise the methodology to be extended to other screening contexts. Through the combination of soft spectral unmixing algorithms and physical hard-modelling we recover the characteristic EDXRD profiles of the materials of interest as well as the quantity or concentrations in which they are present.Inspiration for the statistical methodology is taken from the fields of analytical chemistry and spectroscopy, referred to as chemometric methods. Screening for food, agriculture and medicine quality is common and technologies such as Near Infra-Red spectroscopy and Raman spectroscopy are often employed in these applications. We extend such analytical methods to EDXRD, incorporating relevant contextual and physical insights, in order to extract relevant information from our screening experiments to inform detection.
I. Drakos, P. S. Kenny, T. Fearn, R. D. Speller, Multivariate analysis of energy dispersive X-ray diffraction data for the detection of illicit drugs in border control, J. Crime Sci. 6 (2017)C. Crews, P. S. Kenny, D. O'Flynn, R. D. Speller, Multivariate calibration of energy-dispersive X-ray diffraction data for predicting the composition of pharmaceutical tablets in packaging, Journal of Pharmaceutical and Biomedical Analysis (2018)