Following the London transatlantic plot of 2006,1 in which hydrogen peroxide was to have been mixed during flight with a fuel such as acetone to create an explosive mixture, passengers have been required to remove liquids from their hand luggage prior to boarding. Several solutions have been proposed to address the challenge of liquid identification for security screening, including Raman scattering, nuclear quadrupole resonance, and ion mobility spectrometry, to mention but a few likely candidates. These approaches are unfortunately not generally applicable and can fail, for instance, when used on hermetically sealed vacuum flasks. Conventional computed tomography has the drawback of not being material-specific, and it also cannot distinguish between liquids and solids.
We have recently applied to liquids a technique known as x-ray diffraction (XRD) imaging that spatially resolves the XRD properties of materials within volume elements (voxels) of extended objects such as suitcases.2 The method is already used for the material-specific identification of solid-state explosives and narcotics in hold luggage. Our work aims to eliminate the inconvenience, uncertainty, and expense associated with monitoring liquids separately from hand luggage at checkpoints. The application of XRD imaging to security screening has so far been limited to organic explosives. However, the potential of this technique for identifying liquids still remains to be tapped.
Figure 1. Comparison of the molecular interference function for a 30% by volume aqueous H2O2 solution (solid line) with the Percus–Yevick (P-Y) plot (dashed line), defined by parameters R and φ, determined by equating curve height and peak position with that of a water curve. H2O2: Hydrogen peroxide.
The XRD profiles of liquids can be described as the product of three factors.3,4 These include the x-ray scattering from a single, unbound electron, called the Thomson cross-section, modification of the Thomson cross-section to account for the spatial structure of atomic orbitals (atomic form factor), and the spatial correlations between neighboring atoms originating in the molecular structure of the liquid. Our technique is based on determining this last factor, called the molecular interference function (MIF). MIF depends on momentum transfer, and is closely related to the radial distribution function (RDF), which describes the spatial distribution of electronic charge throughout a liquid. In principle, at least, the RDF and associated MIF can be considered the ‘fingerprints’ of the liquid that they can help to identify.
The Thomson cross-section is well-known, and we have previously shown how to extract the effective atomic form factor from the high-momentum region of the XRD profile of a liquid.5 After appropriate calibration procedures, the influence of these two factors can be removed from the measured XRD signal to reveal the MIF.
Figure 2. Molecular interference function for acetone (solid line) and best fit to the P-Y function (dashed line), normalized to the maximum experimental peak height and position.
For a hypothetical liquid consisting of hard spheres, whose only mode of interaction is to exclude one another from the space occupied by each, the MIF can be calculated using the so-called Percus–Yevick equation.4 This has only has two defining variables, the sphere diameter, R, and the packing fraction, Φ, expressing the fraction of the total volume occupied by the spheres. Synchrotron XRD measurements performed at the Hamburg HASYLAB on two liquid samples, an aqueous hydrogen peroxide solution and acetone, are shown in Figures 1 and 2, respectively. These liquids are representative of the majority of current oxidizer-fuel combinations. For comparison purposes, the best fit to the Percus–Yevick MIF is also shown.
The aim of this work is to establish XRD as a ‘one-stop’ screening technique, valid for the full range of explosives, including crystalline, amorphous, liquid, and home-made explosives, as well as special nuclear materials.6 A detection modality becomes fully characterized only when its receiver operating characteristic (ROC), which plots the probability of raising a false alarm against the detection probability, is known. The ROC analysis of XRD-based liquid screening now requires an extensive, long-term systematic study, which we are currently initiating.
GE Security Germany GmbH
Geoffrey Harding has pioneered x-ray diffraction imaging and its application to security screening
University of Hamburg
Johannes Delfs studied physics in Hamburg and Southampton before starting his PhD studies at the University of Hamburg. He is interested in the development of x-ray diffraction imaging applications in the fields of security screening and medical diagnosis.