Optical detection systems play a vital role in the early warning detection of airborne biological agents. Fluorescence and elastic scattering signals have been historically exploited in order to characterize and profile bioaerosols and yield information that can help suggest the occurrence of a biological attack. More recently, other optical methods, including Raman, infrared, and laser-induced breakdown spectroscopy, have shown promise as candidate bioaerosol detection systems. The selection of an optimal approach involves careful consideration of advantages and disadvantages among these various alternative optical methods. Key considerations are detection probability, false alarm rate, time to detect, and sensitivity. These four parameters are interrelated functions of the nature of the optical signal - characterized by absorption and/or emission cross-section, information content, and signal measurement system technology limitations. Evaluation of prototype systems that exploit optical signatures to detect and warn of the presence of biological aerosols involves a careful, deliberate process of developing a standardized aerosol challenge that mimics the properties of not only a biological agent release, but also the highly complex natural and anthropogenic aerosol background. The key to developing a test methodology involves 1) interpretation of the limited background aerosol data, 2) development of dynamic aerosol challenge capabilities, and 3) integration of experimental design principles in the development and execution of artificial challenge tests and in the reduction and interpretation of sensor system performance based on the test results. Evaluation of prototype systems that exploit optical signatures to detect and warn of the presence of biological aerosols involves a careful, deliberate process of developing a standardized aerosol challenge that mimics the properties of not only a biological agent release, but also the highly complex natural and anthropogenic aerosol background. The key to developing a test methodology involves 1) interpretation of the limited background aerosol data, 2) development of dynamic aerosol challenge capabilities, and 3) integration of experimental design principles in the development and execution of artificial challenge tests and in the reduction and interpretation of sensor system performance based on the test results.

Effective application of point detectors in the field to monitor the air for biological attack imposes a challenging set of requirements on threat detection algorithms. Raman spectra exhibit features that discriminate between threats and non-threats, and such spectra can be collected quickly, offering a potential solution given the appropriate algorithm. The algorithm must attempt to match to known threat signatures, while suppressing the background clutter in order to produce acceptable Receiver Operating Characteristic (ROC) curves. The radar space-time adaptive processing (STAP) community offers a set of tools appropriate to this problem, and these have recently crossed over into hyperspectral imaging (HSI) applications. The Adaptive Subspace Detector (ASD) is the Generalized Likelihood Ratio Test (GLRT) detector for structured backgrounds (which we expect for Raman background spectra) and mixed pixels, and supports the necessary adaptation to varying background environments. The structured background model reduces the training required for that adaptation, and the number of statistical assumptions required. We applied the ASD to large Raman spectral databases collected by ChemImage, developed spectral libraries of threat signatures and several backgrounds, and tested the algorithm against individual and mixture spectra, including in blind tests. The algorithm was successful in detecting threats, however, in order to maintain the desired false alarm rate, it was necessary to shift the decision threshold so as to give up some detection sensitivity. This was due to excess spread of the detector histograms, apparently related to variability in the signatures not captured by the subspaces, and evidenced by non-Gaussian residuals. We present here performance modeling, test data, algorithm and sensor performance results, and model validation conclusions.

Liquid crystal tunable filters (LCTF) have been used in systems developed for Raman Chemical Imaging spectroscopy of chemical, biological and explosives threat materials. However, an ongoing challenge in detecting trace levels of materials is the limited throughput provided by previous generation LCTFs. In this article, we describe a new class of birefringent LCTFs based on a Multi-Conjugate Filter design that provides high throughput over an extended wavelength range (440 nm-750 nm). The spectral resolution, tuning accuracy, out-of-band rejection efficiency have been evaluated and are demonstrated on a Raman chemical imaging microscope platform. Detection of trace threat particulate matter in the presence of complex background with improved overall detection performance is demonstrated.

In recent years a number of analytical devices have been proposed and marketed specifically to enable field-based material identification. Technologies reliant on mass, near- and mid-infrared, and Raman spectroscopies are available today, and other platforms are imminent. These systems tend to perform material recognition based on an on-board library of material signatures. While figures of merit for traditional quantitative analytical sensors are broadly established (e.g., SNR, selectivity, sensitivity, limit of detection/decision), measures of performance for material identification systems have not been systematically discussed. In this paper we present an approach to performance characterization similar in spirit to ROC curves, but including elements of precision-recall curves and specialized for the intended-use of material identification systems. Important experimental considerations are discussed, including study design, sources of bias, uncertainty estimation, and cross-validation and the approach as a whole is illustrated using a commercially available handheld Raman material identification system.

Pacific Northwest National Laboratory (PNNL) has recently recorded the infrared (IR) and far-infrared (FIR, sometimes also called the terahertz, THz) spectral signatures of four common explosives in the condensed phase. The signatures of RDX, PETN, TNT and Tetryl were recorded both in the infrared and the THz domains, using Fourier transform infrared (FTIR) spectroscopy. Samples consisted of thin films and were made by depositing and subsequent evaporation of an acetone-explosive mixture. The complete spectrum spanned the range from 4,000 to 8 cm-1 at 2.0 cm-1 spectral resolution. Preliminary results in the infrared agree with those of previous workers, while the THz signatures are one order of magnitude weaker than the strongest IR bands.

Surface enhanced Raman spectroscopy (SERS) has been widely demonstrated to be capable of single molecule detection. In addition to enhancement of Raman scattering, the substrates used for SERS also display other unique optical properties such as photoluminescence and blinking. In this work, the photoactivation of Ag thin metal films as it relates to the mechanism of SERS enhancement and the production of Ag cluster SERS active sites was explored. Specifically, the photodynamics of SERS-active thin Ag films were qualitatively studied using a combination of optical imaging and high and low resolution spectroscopy. A key hypothesis tested in this work addressed the role of oxygen in thin metal film photodynamics. Based on spectroscopic and kinetic differences observed from thin Ag films under both ambient and nitrogen atmospheres, a simple photochemical mechanism for blinking in optical phenomena was developed and tested. The proposed mechanism relies on the photoreduction of silver oxide to produce an active species, which was postulated to be silver clusters.

Wavelength Modulation Spectroscopy (WMS) has been extensively used as a tool for sensitive detection through precise measurements of the absorption lineshape function of gaseous species. In this paper pathlength saturation in wavelength modulation spectroscopy is studied. New effects are found when one takes advantage of demodulation at higher harmonics of the modulation frequency. We show here that modulation spectroscopy is a much more sensitive probe of these effects. In particular, when synchronous detection is performed at higher harmonics of the modulation frequency, even very small pathlength saturation effects become clearly visible. The method discussed allows one to probe lineshape profiles by observing how the signal profile varies with absorption pathlength. In particular, the signal around line center displays effects of saturation that are characteristic of the lineshape. This method is powerful because, ultimately, all the information about any measurement is contained in the lineshape profile. Since different lineshape profiles exhibit different saturation behavior, higher harmonic detection provides a new method to perform sensitive detection. We have shown effects of saturation on the central lobes of harmonic signals. We also show that there are definite relationships between the variation of the individual side lobes as well as their relative magnitudes that yield further information about the lineshape function.

Enhancing our understanding of atmospheric processes and transformations require a suite of ever more sensitive, selective, versatile, and fast instruments that can measure trace atmospheric constituents at and below mixing ratios of 100-parts-per-trillion on airborne platforms. Instruments that can carry out such measurements are very challenging, as airborne platforms vibrate, experience accelerations, and undergo large swings in cabin temperatures and pressures. These challenges notwithstanding, scientists and engineers at the National Center for Atmospheric Research (NCAR) in collaboration with Rice University have long been employing mid-infrared absorption spectroscopy to acquire atmospheric measurements of important trace gases like formaldehyde on a variety of airborne platforms. The present paper will discuss two very recent airborne formaldehyde instruments employing tunable diode laser and difference frequency generation mid-IR laser sources. Both instruments employ second-harmonic absorption spectroscopy utilizing astigmatic multipass Herriott cells. This paper will discuss the performance of both instruments during recent airborne campaigns, focusing on the many steps necessary for minimizing the various aircraft perturbations. Prospects for the detection of other trace gases will also be presented.

Frequency-stabilized cavity ring-down absorption spectrometry (FS-CRDS) with single-mode excitation using a tunable continuous-wave diode laser is being developed to help support the delivery of reference gas concentration standards. This paper describes initial efforts to compare FS-CRDS measurements with National Institute of Standards and Technology (NIST) methane-in-air standard reference materials to demonstrate the potential of this method to deliver standards-grade measurements with uncertainties of 1 % or lower. The current work demonstrates measurements with residual standard deviations of approximately 1 % for methane sample mole fractions of 50 μmol mol^-1 and above. The results for lower mole fraction samples are poorer due to the poor signal-to-noise ratios and the higher pressures required for the measurements. The current results are potentially limited by the Voigt line shape which was used to model the data.

Type I antimonide diode lasers operate in the 2000 to 2800 nm spectral region. Compared to the 1300 to 1650nm communications spectral band, the antimonide band can access stronger molecular transitions and thus potentially achieve higher sensitivity. Compared to quantum cascade or lead-salt lasers operating at longer infrared wavelengths,antimonide lasers have the advantage that both laser and detector technology support room temperature, cw operation. This paper describes experiments to measure ammonia and methane simultaneously, with high sensitivity and fast response, using a distributed feedback laser at 2200 nm. Our approach is based on scanning the laser over a small spectral regionthat encompasses several lines, either by varying the laser temperature or current, while simultaneously using wavelength modulation with harmonic detection to record the spectrum. Temperature scanning is slower but can cover a wider spectral interval. Digital signal processing methods, including classical least squares and singular value decomposition, extract the gas concentrations from the measured spectra. The accuracy and precision of these algorithms are compared in two limits: the limit when both gases are absent or present only at low levels, and the limit when the concentration of one gas is high.

The recent development of a dense pattern, multiple pass optical cell based on cylindrical mirrors makes possible a differential spectroscopic method that removes (nearly) all common mode features including laser noise, laser distortion, and unwanted optical interference fringes (etalons). The cylindrical mirror cell is similar to other astigmatic cells in that the beam enters through a hole drilled in the center of one mirror. Key differences, however, include the property that for most re-entrant beam trajectories have N passes (where N/2 is odd) through the cell, the N/2 spot is always located at the center of the far mirror. In the differential cell approach, a pellicle beamsplitter located just behind a hole in the far mirror transmits a portion of the beam and reflects the remainder to continue the second set of N/2 passes before exiting through the entrance hole. The two beams - one exiting at the far mirror after N/2 passes, the other exiting at the entrance mirror following N passes - are the reference and sample beams, respectively, applied to a noise canceler circuit. Proof-of-principle experiments reported here using near-infrared measurements of methane absorbance show the differential method does work. The optical system used, however, introduced excessive astigmatism in the beam reflected from the pellicle beam splitter because of the displace of the pellicle from the cell mirror surface. That astigmatism made it difficult to align the return beam for the complete set of N/2 passes and to focus the exit beam onto the photodiode detector. Design improvements are discussed.

A compact, diode laser-based sensor has been developed for meteorological balloons to measure atmospheric carbon dioxide profiles. The sensor achieves a precision of better than 1 ppmv using a novel pressure/temperature compensating reference cell. This device weighs less than 1 kg and uses less than 4 Watts of battery electrical power. Turnkey operation is achieved by a digital signal processor. A full description of the sensor and a discussion of its performance are provided.

Optical interference fringes due to unwanted etalons are often the limiting uncertainty in diode laser spectroscopic trace gas measurements. Temporal variations in the fringe spacings, phases, and amplitudes introduce systematic baseline changes that limit useful signal averaging times to ~1000 seconds, and constrain minimum detectable absorbances to between one and three orders of magnitude worse than the fundamental limiting noise sources (shot noise and/or detector thermal noise). We describe an adaptive numerical filtering method based on singular value decomposition (SVD) that shows, for one system studied, a five-fold reduction in baseline drift due to unwanted etalons over a one week measurement period. The adaptive algorithm is fast (< 1 msec per computation), robust, and uses linear methods. It is computationally equivalent to principal component analysis (PCA). The test systems were acetylene detected using a near-infrared telecommunications laser operating near 6542 cm^-1 and methane detected using a vertical cavity surface emitting laser (VCSEL) operating at 6057cm^-1. The acetylene detection limit was 20 ppb (1 σ) over a 1 week measurement.

A longwave-infrared (LWIR) passive-spectrometer performance was evaluated with a short-pathlength gas cell. This cell was accurately positioned between the sensor and a NIST-traceable blackbody radiance source. Cell contents were varied over the Beer's Law absorbance range from the limit of detection to saturation for the gas analytes of sulfur hexafluoride and hexafluoroethane. The spectral impact of saturation on infrared absorbance was demonstrated for the passive sensor configuration. The gas-cell contents for all concentration-pathlength products was monitored with an active traditional-laboratory Fourier Transform Infrared (FTIR) spectrometer and was verified by comparison with the established PNNL/DOE vapor-phase infrared (IR) spectral database. For the passive FTIR measurements, the blackbody source employed a range of background temperatures from 5^oC to 50^oC. The passive measurements without the presence of a gas cell permitted a determination of the noise equivalent spectral noise (NESR) for each set of passive gas-cell measurements. In addition, the no-cell condition allowed the evaluation of the effect of gas cell window materials of low density poly(ethylene), potassium chloride, potassium bromide, and zinc selenide. The components of gas cell, different window materials, temperature differentials, and absorbances of target-analyte gases supplied the means of evaluating the LWIR performance of a passive FTIR spectrometer. The various LWIR-passive measurements were found to simulate those often encountered in open-air scenarios important to both industrial and environmental monitoring applications.

Pollutants detection by tunable diode laser spectroscopy is conventionally achieved by scanning the emission frequency of the laser around an isolated absorption line of the species under investigation. Absolute quantification relies on the comparison of the measured absorption signal with the absorption signal of a calibrated sample at the same pressure, or with a calculated line profile when the spectroscopic parameters are available and accurate. We developed an alternative procedure : with the laser emission frequency actively stabilized on top of the absorption line, both the pressure inside the cell and the absorption signal are measured while the cell is progressively filled with the sample up to about 12 Torr. The slope at origin of the signal vs. pressure curve is proportional to the concentration in the sample and absolute concentration is obtained with a calibrated mixture injected into the cell at regular intervals. This procedure, which proves as efficient as the conventional one, has been applied together with a mobile spectrometer to the quantification of formaldehyde in outdoor and indoor (buildings and cars) environments.

The ability to distinguish endospores from each other, from vegetative cells, and from background particles has been demonstrated by PNNL and several other laboratories using various analytical techniques such as MALDI and SIMS. Recent studies at PNNL using Fourier transform Infrared (FTIR) spectroscopy combined with statistical analysis have shown the ability to characterize and discriminate bacterial spores and vegetative bacteria from each other, as well as from background interferents. In some cases it is even possible to determine the taxonomical identity of the species using FTIR. This effort has now grown to include multiple species of bacterial endospores, vegetative cells, and background materials. The present work reports on advances in being able to use FTIR, or IR in combination with other techniques, for rapid and reliable discrimination.

Since the distribution of anthrax causing spores through the U.S. Postal System in the autumn of 2001, numerous methods have been developed to detect spores with the goal of minimizing casualties. During and following an attack it is also important to detect spores on surfaces, to assess extent of an attack, to quantify risk of infection by contact, as well as to evaluate post-attack clean-up. To perform useful measurements, analyzers and/or methods must be capable of detecting as few as 10 spores/cm², in under 5-minutes, with little or no sample preparation or false-positive responses, using a portable device. In an effort to develop such a device, we have been investigating the ability of surfaceenhanced Raman spectroscopy (SERS) to detect dipicolinic acid (DPA) as a chemical signature of bacilli spores. In 2003 we employed SERS to measure DPA extracted from a 10,000 spores per μL sample using hot dodecylamine. Although the entire measurement was performed in 2 minutes, the need to heat the dodecylamine limits field portability of the method. Here we describe the use of a room temperature digesting agent in combination with SERS to detect 220 spores collected from a surface in a 1 μL sample within 3 minutes.

Reagentless water and surface sensors employing laser induced native fluorescence (LINF) and resonance Raman spectroscopy (RRS) in the deep UV are making significant progress in detecting chemical and biological targets and differentiating them against a wide range of background materials. Methods for optimizing sensor performance for specific target and backgrounds materials will be discussed in relationship to closed industrial environments and open natural environments. Limits of detection and chemical specificity will be discussed for high and low spectral resolution systems for a wide range of compounds and composite particles such as spores and cells. Detection and identification of single spores at working distance of several meters is illustrated. A range of sensors will be described along with their physical and performance specifications including sample, sipper and immersion sensors for water and fixed point and scanner systems for surfaces. In addition, the use of UV LINF and RRS for detection in capillary electrophoresis and liquid chromatography will be described with limits of detection in the range of a few nmol L^-1.

A novel handheld Fourier-transform infrared spectrometer (FTS) system, currently in development, is described. Estimated and measured performance data presented here are based on modeling and preliminary testing. The basic instrument will be useful for a variety of sensing applications, including chemical agent detection. One novel aspect is a refractively-scanned, field-widened interferometer, providing, in a miniature footprint, energy equal to a laboratory spectrometer. A second novel aspect is the use of solid-phase extraction to concentrate airborne chemicals for infrared detection. FTS instruments provide a powerful approach to identification of chemical and biological substances. The specificity is very high, while the sensitivity varies with sampling interfaces and detection methods. Cost, size, sensitivity and weight have impeded the widespread deployment of FTS systems. Cost can be reduced by a variety of means, including improved designs, mass production, and the on-going electronics and manufacturing revolutions. Size can be addressed by the use of field widening, which has been known for many years, though seldom used. Photoacoustic detection provides a very low-cost and relatively sensitive sampling interface. Modeling indicates that the sensitivity can reach part per billion to part per trillion levels.

We report advances made on the development of a fiber optic nerve agent sensor having its entire length as the sensing element. The optical fiber is multimode, and consists of a fused-silica core and a nerve agent sensitive cladding. Upon exposure to sarin gas, the cladding changes color, resulting in an alteration of the light intensity throughput. The fiber is mass produced using a conventional fiber optic draw tower. This technology could replace, or be used with, a collection of point-detectors to protect personnel, buildings and perimeters from dangerous chemical attacks.

An optical fiber, stripped partially out of its cladding is used to sense refractive index of a liquid to a precision to fifth place of decimal. The dependence of the light output of the sensor on the refractive index of the test liquid is nonlinear. The light output of the sensor depends on the thickness to which the cladding is stripped. It shows both positive and negative slope with increasing refractive index of the test liquid. The slope of the plot of sensor output against liquid refractive index shows a change of sign at around the fiber refractive index. The sensor is unaffected by the presence of absorption and is insensitive to the chemical nature of the solute. The sensor is sensitive in the whole of the tested range of refractive indices 1.33 to 1.52. Experiments that show the significance of cladding modes in sensing are described.

Oxytetracycline (OTC) is extensively used in aquaculture worldwide for preventive and therapeutic purposes. Most of the drug, however, is discharged into the marine environment due to leaching from medicated feed and poor gastrointestinal (GI) absorption. Without exposure to sun light OTC has a long lifetime in the marine environment, therefore it is important to monitor and study its occurrence, distribution, fate and impact on the ecosystem. A portable tetracycline (TC) analyzer was developed in this laboratory for this purpose based on europium-sensitized luminescence. In this study, an assay method is developed for OTC analysis in water using this instrument. Water samples are filtered with glass wool; and solid phase extraction (SPE) is performed using Oasis HLB cartridges for OTC extraction and cleanup. Following reagent application, the samples are excited by 385 nm LED pulses; and time-resolved luminescence (TRL) is measured at 610 nm by a photomultiplier tube. A 0-3 ppm linear dynamic range (r² = 0.9988) and a 0.021 ppb limit of detection were achieved with a typical <5% relative standard deviation.

Protection of military personnel and civilians from water supplies poisoned by chemical warfare agents requires an analyzer that has sufficient sensitivity (μg/L), selectivity (differentiate the warfare agents from its hydrolysis products), and speed (less than 10 minutes) to be of value. In an effort to meet these requirements, we have been investigating the ability of surface-enhanced Raman spectroscopy (SERS) to detect these chemicals in water. The expected success of SERS is based on reported detection of single molecules, the one-to-one relationship between a chemical and its Raman spectrum, and the minimal sample preparation requirements. It is equally important to detect and distinguish the hydrolysis products of these agents to eliminate false-positive responses and evaluate the extent of an attack. Previously, we reported the SER spectra of GA, GB, VX and most of their hydrolysis products, as well as a preliminary study of HD, and its principle hydrolysis product thiodiglycol. Here we expand this study to include half-mustard, its hydrolysis product, 2-hydroxyethyl ethylsulfide, and ethyl ethylsulfide to better characterize the observed SER spectra. We also report the measurement of 10 μg/L of thiodiglycol as we continue to improve sensitivity.

Molecularly imprinted polymers (MIPs) have the potential to provide a unique combination of high chemical selectivity and environmental stability and are, therefore, being widely studied in chemical sensor applications. Optical interrogation of the MIP-chemical interaction is very convenient for the detection of fluorescent compounds, but is problematic for the detection of non-fluorescent species. Doping MIPs with Eu3+ is one approach that can facilitate the optical detection of non-fluorescent species. Eu3+ has absorption in the near UV and the doped MIP can, therefore, be excited with a commercially available laser diode at 375nm. In the present paper MIPs doped with Eu3+ and imprinted to methyl salicylate (MES), a chemical warfare agent simulant, were prepared in the form of a thin film on a quartz substrate. Non-imprinted (Blank) polymer films were also prepared using the same imprinting procedure, but without introducing the MES template. Both polymers were tested to MES and the structurally similar compound methyl 3,5-dimethylbenzoate (DMB) in hexane. For MES, the fluorescence intensity of the MIP was significantly stronger than for the Blank, while for the methyl 3,5-dimethylbenzoate, the Blank polymer exhibited the stronger fluorescence signal. A portable chemical sensor employing differential fluorescence from MIP/Blank polymer pairs is under development and allows target discrimination without the need for spectroscopic analysis of the emission spectra.

A diode laser based natural gas leak detector has been developed that can measure methane concentrations over six orders of magnitude, from ambient (1.7 ppm) to pure gas levels. The detection method utilizes a small multipass cell and wavelength modulation absorption spectroscopy. At high methane concentration, various forms of unmodulated absorption spectroscopy are used. The instrument is a handheld unit that operates on less than 2 W of power and weighs 1.4 kg (including battery). A small pump on the unit pulls outside gas into the enclosed optical cell through an extendable probe. The response time of the instrument is approximately 1 - 2 sec.

Fluorescence resonance energy transfer (FRET) is a process in which energy is transferred nonradiatively from one fluorophore (the donor) in an excited electron state to another, the chromophore (the acceptor). FRET is distinctive in its ability to reveal the presence of specific recognition of select targets such as the nerve agent stimulant Methyl Salicylate (MES) upon spectroscopic excitation. We introduce a surface imprinted and non-imprinted thin film that underwent AC-Electrospray ionization for donor-acceptor pair(s) bound to InGaP quantum dots and mesoporous silicate nanoparticles. The donor-acceptor pair used in this investigation included MES (donor) and 6-(fluorescein-5-(and-6)- carboxamido) hexanoic acid, succinimidyl ester bound to InGaP quantum dots (acceptor). MES was then investigated as a donor to various acceptor fluorophore: InGaP: mesoporous silicate nanoparticle layers.

One of today's primary security challenges is the emerging biological threat due to the increased accessibility to biological warfare technology and the limited efficiency of detection against such menace. At the end of the 90s, Defence R&D Canada developed a standoff bioaerosol sensor, SINBAHD, based on intensified range-gated spectrometric detection of Laser Induced Fluorescence (LIF) with an excitation at 351 nm. This LIDAR system generates specific spectrally wide fluorescence signals originating from inelastic interactions with complex molecules forming the building blocks of most bioaerosols. This LIF signal is spectrally collected by a combination of a dispersive element and a range-gated ICCD that limits the spectral information within a selected atmospheric cell. The system can detect and classify bioaerosols in real-time, with the help of a data exploitation process based on a least-square fit of the acquired fluorescence signal by a linear combination of normalized spectral signatures. The detection and classification processes are hence directly dependant on the accuracy of these signatures to represent the intrinsic fluorescence of bioaerosols and their discrepancy. Comparisons of spectral signatures acquired at Suffield in 2001 and at Dugway in 2005 of bioaerosol simulants, Bacillius subtilis var globiggi (BG) and Erwinia herbicola (EH), having different origin, preparation protocol and/or dissemination modes, has been made and demonstrates the robustness of the obtained spectral signatures in these particular cases. Specific spectral signatures and their minimum detectable concentrations for different simulants/interferents obtained at the Joint Biological Standoff Detection System (JBSDS) increment II field demonstration trial, Dugway Proving Ground (DPG) in June 2005, are also presented.

We describe algorithm development for a trigger system for bio-aerosol detection using bulk collection of aerosols. Two key problems inherent to any system which collects or probes a volume of air are presented - the "mixture" problem and the "spike" problem. We describe a background suppression and detection algorithm and show why knowledge of background endmembers is important. We present an endmember selection algorithm and show examples. Integrating these two algorithms solves both the mixture and spike problems and has applications to both bio-aerosol point detectors which collect samples from a volume of air, and to bio-aerosol stand-off detectors which probe a column of air.

In this paper we describe a method of fabricating a Fabry-Perot filter array consisting of four distinct wavelengths using a stopping layer, which in turn is discriminately measured. Precise control of the oxide thickness is demonstrated by using reflectance to measure center wavelengths (CWL) between 645nm-822nm with full width half maximum (FWHM) values of 15 nm. These parameters are used to confirm good narrow band filter characteristics. The physical and chemical properties of an oxide layer converted from a silicon-carbon-nitride (SiCN) etch stop layer (ESL) is reported for both as-deposited and the resultant oxidized film. The filter array can be fabricated directly on top of silicon photo diodes, to form a complete multi-wavelength sensor system. Fabricating a multi-wavelength filter array using etch-stop layers can provide better thickness control and across wafer uniformity compared to a timed-etch approach.

Infrared spectrometry allows detection and identification of gases in the atmosphere as well as analysis of solids and liquids from long distances. An application of the method which has received increased attention over the last years is the detection of hazardous compounds. These may be present in the gas phase in the atmosphere but also as liquid droplets on surfaces. In this study, imaging Fourier transform spectrometry (IFTS) was applied to detect both liquids and gases. Measurements were performed with an imaging Fourier transform spectrometer developed at TUHH. The imaging spectrometer and first results of measurements are presented.

The CH stretching overtone transitions of the nerve agent sarin (O-isopropyl methylphosphonofluoridate) are of interest to the standoff detection of chemical warfare agents, as many of these transitions occur near regions where small, efficient, portable diode lasers (originally developed for use in the telecommunications industry) operate. However, the interpretation of experimental vibrational overtone spectra is often difficult, and the computational simulation of overtone transitions in a molecule is challenging. Presented herein are the simulated CH overtone stretching transitions in sarin. Spectral regions are simulated from overtone transition energies and intensities, both of which are calculated within the harmonically coupled anharmonic oscillator (HCAO) model. Data for HCAO calculations are obtained from ab initio calculations, without any recourse to experimental data.

It is shown that 2, 4, 6-Trinitrotoluene (TNT) displays strong and distinct structures in differential reflectograms, near 420 nm and 250 nm. These characteristic peaks are not observed from approximately two dozen organic and inorganic substances which we tested and which may be in or on a suitcase. This exclusivity infers an ideal technique for explosives detection in mass transit and similar locations. The described technique for detection of explosives is fast, inexpensive, reliable, portable, and is applicable from some distance, that is, it does not require contact with the surveyed substance. Moreover, we have developed a curve discrimination program for field applications of the technique. Other explosives such as 1, 3, 5-trinitro-1, 3, 5 triazacyclohexane (RDX), 1, 3, 5, 7-Tetranitro-1, 3, 5, 7- tetraazacyclooctane (HMX), 2, 4, 6, N-Tetranitro-N-methylaniline (Tetryl), Pentaerythritol tetranitrate (PETN), and nitroglycerin have also been investigated and demonstrate similar, but unique, characteristic spectra. The technique utilizes near-ultraviolet to visible light reflected from two spots on the same sample surface yielding a differential reflectogram corresponding to the absorption of the sample. The origin of the spectra is attributed to the highest occupied molecular orbital to lowest unoccupied molecular orbital (HOMO-LUMO) transitions of the respective explosive molecule. Experiments using transmission spectrophotometry have also been performed to compliment and confirm the specific transitions. The results are supported by computer modeling of the molecular orbitals that yield UV and visible transitions.

Passive millimeter-wave (mmW) systems have been used in the past to remotely map solid targets and to measure low-pressure spectral lines of stratospheric and interstellar gases; however, its application to pressure-broadened spectral detection of terrestrial gases is new. A radiative transfer model was developed to determine the detection feasibility and system requirements for passive mmW spectral detection. A Dicke-switched multispectral radiometer that operates at 146-154 GHz was designed and built for remote detection of stack gases. The radiometer was tested in the laboratory using a gas cell; the spectra of acetonitrile were detected passively against a cold background, which mimicked typical remote detection scenarios in the field. With Dicke-switched integration of radiometric signals, on-line calibration, and novel signal processing to minimize atmospheric fluctuation, spectral line detection of polar molecules is possible from chemical plumes a few kilometers away.

This paper presents a band selection technique for spectral feature characterization in hyperspectral data, referred to as band selection for hyperspectral signature feature characterization (BSHSFC). Since a hyperspectral signature is characterized by its spectral profile, the number of bands to be selected is totally determined by spectral features that uniquely characterize the signature. As a result, two hyperspectral signatures may require different sets of bands for spectral characterization. The proposed BSHSFC is a variable-size variable-band selection (VSVBS) where the number of selected bands varies with a hyperspectral signature to be processed. In order for BSHSFC to select an appropriate set of bands for a hyperspectral signature, a new band prioritization criterion, referred to as orthogonal subspace projectorbased band prioritization criterion (OSP-BPC) is derived for this purpose. It assigns a different priority score to each spectral band of a hyperspectral signature such that various features can be captured by the BSHSFC from the original set of bands. Accordingly, the BSHSFC can be interpreted as a spectral feature extraction technique for signature characterization. Finally, experiments using two sets of data are conducted to demonstrate that the BSHSFC-based BS can improve the performance of hyperspectral signature characterization.

An endmember is an idealized, pure signature for a class and provides crucial information for hyperspectral image analysis. Recently, endmember extraction has received considerable attention in hyperspectral imaging due to significantly improved spectral resolution where the likelihood of a hyperspectral image pixel uncovered by a hyperspectral image sensor as an endmember is substantially increased. Many algorithms have been proposed for this purpose. One great challenge in endmember extraction is the determination of number of endmembers, p required for an endmember extraction algorithm (EEA) to generate. Unfortunately, this issue has been overlooked and avoided by making an empirical assumption without justification. However, it has been shown that an appropriate selection of p is critical to success in extracting desired endmembers from image data. This paper explores methods available in the literature that can be used to estimate the value, p. These include the commonly used eigenvalue-based energy method, An Information criterion (AIC), Minimum Description Length (MDL), Gershgorin radii-based method, Signal Subspace Estimation (SSE) and Neyman-Pearson detection method in detection theory. In order to evaluate the effectiveness of these methods, two sets of experiments are conducted for performance analysis. The first set consists of synthetic imagebased simulations which allow us to evaluate their performance with a priori knowledge, while the second set comprising of real hyperspectral image experiments which demonstrate utility of these methods in real applications.

In this paper, we present a modified Fisher's linear discriminant analysis (FLDA) to hyperspectral remote sensing image dimension reduction and classification. The basic idea of FLDA is to design an optimal transform which can maximize the ratio of between-class scatter matrix to within-class scatter matrix. The practical difficulty of applying the FLDA to hyperspectral images includes the unavailability of enough samples for all the classes. So the original FLDA is modified to avoid the requirement of class samples. In the following data classification using the FLDA-transformed low-dimensional data, a more powerful classifier generally is required. Fortunately, we find this is not difficult to achieve. A simple distance based classifier, such as Spectral Angle Mapper (SAM), can provide satisfactory classification performance. This approach is particularly useful to the data sets with small classes.

Multispectral sensors are still widely used in satellite remote sensing. They usually have spectral bands less than ten channels. The problem for so few channels is that it can not directly solve linear mixture model by least square unmixing for subpixel target detection. In order for least square approach to be effective, the number of bands must be greater than or equal to that of signatures to be classified, i.e., the number of equations should be no less than the number of unknowns. This ensures that there are sufficient dimensions to accommodate orthogonal projections resulting from the individual signatures. It is known as band number constraint (BNC). Such constraint is not an issue for hyperspectral images since they generally have hundreds of bands, however, this may not be true for multispectral images where the number of signatures to be classified might be greater than the number of bands. In order to relax this constraint, we present two signature reduction methods to reduce the number of unknowns, based on signature selection and signature fusion. A SPOT image scene will be used for experiment to demonstrate the performance.

Our environment has been changed continuously by nature causes or human activities. In order to identify what has been changed during certain time period, we need to spend enormous resources to collect all kinds of data and analyze them. With remote sensing images, change detection has become one efficient and inexpensive technique. It has wide applications including disaster management, agriculture analysis, environmental monitoring and military reconnaissance. To detect the changes between two remote sensing images collected at different time, radiometric calibration is one of the most important processes. Under the different weather and atmosphere conditions, even the same material might be resulting distinct radiance spectrum in two images. In this case, they will be misclassified as changes and false alarm rate will also increase. To achieve absolute calibration, i.e., to convert the radiance to reflectance spectrum, the information about the atmosphere condition or ground reference materials with known reflectance spectrum is needed but rarely available. In this paper, we present relative radiometric calibration methods which transform image pair into similar atmospheric effect instead of remove it in absolutely calibration, so that the information of atmosphere condition is not required. A SPOT image pair will be used for experiment to demonstrate the performance.

For hyperspectral imagery, greedy modular eigenspaces (GME) has been developed by clustering highly correlated hyperspectral bands into a smaller subset of band modules based on greedy algorithm. Instead of greedy paradigm as adopted in GME approach, this paper introduces a simulated annealing band selection (SABS) approach for hyperspectral imagery. SABS selects sets of non-correlated hyperspectral bands for hyperspectral images based on simulated annealing (SA) algorithm while utilizing the inherent separability of different classes in hyperspectral images to reduce dimensionality and further to effectively generate a unique simulated annealing module eigenspace (SAME) feature. The proposed SABS features: (1) avoiding the bias problems of transforming the information into linear combinations of bands as does the traditional principal components analysis (PCA); (2) selecting each band by a simple logical operation, call SAME feature scale uniformity transformation (SAME/FSUT), to include different classes into the most common feature clustered subset of bands; (3) providing a fast procedure to simultaneously select the most significant features according to SA scheme. The experimental results show that our proposed SABS approach is effective and can be used as an alternative to the existing band selection algorithms.