Imager locates toxic liquids at stand-off distances
The remote location and identification of potential liquid hazards is desirable in a variety of civilian and military applications, such as toxic industrial spills or the deployment of chemical warfare agents (CWAs). When these substances exhibit low volatility, the inherent lack of vapor emanating from the sample challenges traditional chemical sensors. This limitation could lead to military users, first responders, or civilians inadvertently encountering involatile toxic chemicals, resulting in casualties and requiring laborious decontamination procedures.
Direct detection of persistent liquid contamination is one route to overcome this issue. Ideal configurations include isolating the user from the hazard with either an unmanned system intimately sampling the scene (remote detection) or a detection system co-located with the user at a distance from the scene (stand-off detection). The latter format clearly circumvents the need for subsequent decontamination. Such systems should also be able to differentiate between threatening and benign materials and survey a wide area. Hyperspectral imaging techniques combine these aspects as they exploit characteristic spectral signatures to classify or, ideally, identify target chemicals. Furthermore, imagery delineates the contamination extent.
Previous incarnations of stand-off liquid detection systems primarily focused on detecting chemicals at a single point using techniques such as laser-induced breakdown spectroscopy or Raman spectroscopy.1, 2 Those techniques provide chemical-specific information, but lack the spatial information required to conduct rapid area surveys. Passive hyperspectral imagery has been primarily used for stand-off detection of gases using the midwave IR (MWIR, 3–5μm) and longwave IR (LWIR, 8–12μm) bands to exploit the temperature difference of gases and vapors relative to the background environment.3 Similar systems have also been used to detect deposited chemicals and explosives,4 but the sample and background are in thermal equilibrium and sufficient ambient lighting is necessary, factors that can constrain detector sensitivity. To overcome these limitations, we have developed an active IR hyperspectral imaging system called the Negative Contrast Imager (NCI).
The NCI is a compact (33.5cm × 25cm × 13cm) detection system based on IR absorption/reflection spectroscopy (see Figure 1). If the illuminating radiation wavelength is coincident with the IR absorption band of a molecular species in the scene, then the incident radiation is attenuated, resulting in the detection of a reduced signal, and hence a negative contrast in the acquired image. The system works by acquiring several images at laser wavelengths that are commensurate with either absorption features or a background absorption level to create a hyperspectral data cube. Each image gives spatial information on the presence of the species of interest. The system is currently specified to cover a 1m2 area at a distance of 5m, although this can be modified using onboard zoom functions.
Two principal subsystems make up the NCI: a broadly tunable intracavity optical parametric oscillator (ICOPO) that serves as the active illumination source and a galvanometer-based raster scanning system for image acquisition.6–9 The intracavity approach confers two key advantages. First, the circulating power inside the cavity is much higher than outside, thereby reducing the pump-laser power requirements and those of the overall system. Second, this configuration is inherently compact, which enables a portable, stand-off liquid detection system. The ICOPO is based upon magnesium-oxide-doped periodically poled lithium niobate, which enables a narrow band of shortwave IR (SWIR) and MWIR photons to be produced with tuning ranges of 1.5–1.8μm and 2.6–3.8μm, respectively. This output light is scanned across the scene using an x-y galvanometer-actuated mirror pair. The returned signal is collected by the same mirrors and focused onto indium gallium arsenide and zinc-doped mercury cadmium telluride point detectors.
We have assessed the performance of the NCI system as a function of its angle of incidence. This informs us about the system's performance in a configuration representative of a realistic situation. We applied 10μl of thiodiglycol (TDG, a CWA simulant) to swatches of plasterboard, tarpaulin, and painted metal and imaged at angles to the surface of 90°, 70°, and 30°. These angles correspond to distances between the NCI and the surfaces of 1.36m, 1.39m, and 1.66m, respectively. Figure 2 shows the images acquired at 3.13μm, resonant with a spectral band in the TDG absorption spectrum. The TDG droplet on the plasterboard is clearer at angles of 70° and 30° and is not discernible at 90°. We acquired more images at key wavelengths to form a hyperspectral data cube. From this data, the TDG spectral signature can be extracted from the TDG droplet (see Figure 3), enabling location and detection of this liquid. Further results with deposited liquid CWA can be found in our SPIE article.5
Using the NCI, we have been able to detect and locate low-volatility liquid chemicals deposited on realistic surfaces at a variety of angles, thereby showing its potential utility as a portable stand-off CWA detection system. Our next step is to move from the current SWIR/MWIR optical parametric oscillator to develop a system that can access the LWIR band, also known as the spectral fingerprint region, where feature-rich spectra could enable greater differentiation between chemical substances.
We would like to thank the UK Ministry of Defence and the US Department of Defense for funding this work.
© British Crown Copyright 2013
DSTL-published with the permission of the Controller of Her Majesty's Stationery Office
Chris Howle is the technical lead for chemical and biological spectroscopy in DSTL's detection department, where his research interests include stand-off chemical detection. He obtained both an MSci and PhD in chemistry from the University of Birmingham, followed by research at the University of Bristol and Lawrence Berkeley National Laboratory. In addition to being a SPIE member, he is part of the SPIE Scholarship Committee, which makes recommendations to the Board of Directors on the scholarship and grant awards made by the society.
Rhea Clewes is part of DSTL's emerging chemical sensors team, specializing in novel spectroscopic research. She received an MChem degree (with honors) from Cardiff University in 2011 while undertaking research projects in electron paramagnetic resonance and surface-enhanced Raman spectroscopy. She is an early career member of SPIE.
Jason Guicheteau, research chemist with the US Army ECBC, focuses his research on proximal detection of chemical, biological, and energetic materials using Raman spectroscopy and surface-enhanced Raman spectroscopy. He is also on the committee for the chemical, biological, radiological, nuclear, and explosives sensing program within the SPIE Defense, Security, and Sensing conference.
Keith Ruxton received an MEng degree (with merit) in electronic and electrical engineering with business studies, followed by a PhD researching advances in tunable diode laser spectroscopy using residual amplitude modulation techniques, both awarded by the University of Strathclyde. He joined M Squared Lasers Ltd. as an applications engineer in 2010. His current research includes hyperspectral imaging and spectroscopy techniques using mid-IR sources.
Graeme Malcolm achieved a BSc in laser physics and optoelectronics at Strathclyde University, followed by a PhD incorporating research into compact solid-state laser sources. In 1992 he co-founded Microlase Optical Systems Ltd., which was acquired by Coherent Inc. in 1999, where he led the development of a new manufacturing facility. In October 2005 he co-founded M Squared Lasers Ltd., which specializes in novel laser solutions for a wide range of sectors, including defense and security, oil and gas, and food and drink.
