Computer-generated synthetic single-beam spectra and interferograms are used to study signal processing strategies for passive Fourier transform IR (FTIR) remote sensing. Four-component mixtures of ethanol, methanol, methyl ethyl ketone, and acetone vapors in two passive FTIR remote sensing scenarios are studied. Interferogram and spectral processing strategies are compared based on the ability to generate accurate pattern recognition and multivariate calibration models. Chemical detection and quantitative analysis is further assessed based on analyte band characteristics and spectral overlap. Temperature differential between the analyte cloud and the IR source is found to be the critical factor determining classification performance. Calibration model performance is further dependent upon the ability to reliably assess analyte plume temperature. Calibration model performance is shown to degrade significantly as the uncertainty in plume temperature increases or as the temperature difference decreases. These results suggest that in order to quantify passive FTIR spectra remotely either there must be either a large temperature difference or an accurate assessment of plume temperature.

Hyperspectral imaging is the latest advent in imagin technology, providing the potential to extract information about the objects in a scene that is unavailable to panchromatic imagers. This increased utility, however, comes at the cost of tremendously increased data. The ultimate utility of hyperspectral imagery is in the information that can be gleaned from the spectral dimensions, rather than in the hyperspectral imagery itself. To have the broadest range of applications, extraction of this information must occur in real-time. Attempting to produce and exploit compete cubes of hyperspectral imagery at video rates, however, presents unique problems for both the imager and the processor, since data rates are scaled by the number of spectral planes in the cube. MIDIS allows both real-time collection and processing of hyperspectral imagery over the range of 0.4 micrometers to 12 micrometers .

Passive standoff FTIR spectrometry relies on the radiance differential between a background scene and a target vapor analyte. Unlike traditional FTIR approaches controlling radiance levels within a narrow range, the passive configuration often encounters a large variance in radiance levels. This places higher demands on the passive FTIR configuration for maintaining linearity. The present study assesses the radiometric linearity of a passive FTIR configuration using controlled blackbody radiance in conjunction with target samples such as a polystyrene film and ethanol vapors.

Passive FTIR remote sensing is based on the spectral analysis of ambient IR radiation. If there are chemical agents in the field of five of the IR sensor, and if there is a temperature difference between the background and the chemical agent cloud, significant excursion will appear in the IR spectra. By calculation of the radiance and the brightness temperature the spectral analysis can be performed without prior background measurements. Numerous chemical compounds can be identified by their spectral fingerprints with an automated software. From laboratory measurements detection limits are determined in terms of concentration X path length X temperature difference. Field measurements of released ammonia are presented. The ammonia cloud was identified on-line from several observation points at 600 m, 1.5 km and 2.4 km. The analysis results clearly demonstrate that reliable IR remote sensing of hazardous chemical agents in the atmosphere is possible if a highly sensitive instrument, an optimized measurement technique and a sophisticated evaluation software are used.

We are investigating the use of small, transportable, Raman systems for remote Raman measurements at intermediate ranges. Previous work focused on the use of an imaging spectrograph and a fiber-optic coupled probe for making single point measurements. More recently we have considered the use of tunable filters for remote Raman imaging. For this work, acousto-optical and liquid crystal tunable filters are being used both with, an in place of dispersive spectrometers and fixed filtering devices. In addition, we have improved the system by the use of a modified holographic fiber-optic Raman probe for light and image collection. In these experiments, a 100 micron collection optical fiber is replaced with a small diameter image guide for light collection and imaging. The feasibility of tunable filter technology for remote Raman imagin will be discussed along with the merits of image transfer devices using small- diameter image guides.

A novel gas-sensing system based on a dynamic nonlinear response is reported. A sinusoidal temperature change was applied to a semiconductor gas sensor, and the resulting conductance of the sensor was analyzed by fast Fourier transformation (FFT). The output signals of the sensor changed characteristically depending on the composition of the sensed alcohol and were quantitatively characterized by the higher harmonics in FFT. The characteristic responses of the gas sensor are discussed in relation to the kinetics of alcohol at the sensor surface.

An array of four acoustic wave chemical sensors has been developed and tested for recognition and quantitation of six closely related chlorinated organic vapors at low part-per- million concentrations. These vapors include PCE, TCE, VC, and DCE, cis-1,2-DCE, trans-1,2-DCE, and 1,1-DCE. The developed sensor array can detect as little as 0.2 ppm of PCE, 0.8 ppm of TCE, 1.4 ppm of cis-1,2-DCE, 1.4 ppm of trans-1,2-DCE, 3 ppm of 1,1-DCE, and 3.5 ppm of VC in air. Quantitation of TCE and cis-1,2-DCE vapor mixtures was achieved using multivariate calibration techniques. The locally weighted regression analysis provides more accurate results over the partial least squares regression.

An array of semiconducting metal oxide (SMO) chemiresistive sensor can quantitatively measure the freshness of Atlantic salmon. A variety of SMO films were tested, including films containing oxides of tungsten and tin. Experiments were performed with these films to determine if they were sensitive to trimethylamine and/or dimethylamine, known Atlantic salmon degradation products. The electrical resistance of the films before and during exposures to concentrations of the amines was monitored. The sensor temperatures were varied to determine the operating conditions for sensitivity and selectivity to one or both gases. The most promising films were then used to monitor the degradation of Atlantic salmon is conjunction with a sensory evaluation panel and tests using gas chromatography mass spectrometry and total volatile base nitrogen.

Gas sensor development traditionally focuses on a single sensing platform that is optimized for a specific task. An ideal sensor is one that is completely sensitive as well as selective to the target gas of interest while being non- sensitive to interferent gases. Unfortunately, ideal sensors do not exist. Sensor arrays expand on the single sensor concept by incorporating a number of sensor utilizing the same sensing technology to provide a fingerprint of the particular measurand. Multiple sensor response are then analyzed with pattern recognition techniques such as neural networks. The limitation with single technology design methods is that single sensor limitations propagate. A sensor technology effective in detecting specific measurands may not be sensitive to related measurands. A novel sensor approach is underway that incorporates two sensing techniques, surface acoustic wave (SAW) and SMO, into an array of arrays. This integrated sensor array can provide marked improvements over either array alone in the increased bandwidth of measurands, the capability of cross verifying result with complimentary sensor technology responses, and the performance in the presence of interferents. Difficulties arise in the management, coordination and signal processing of the two methods. This is due to the fact that SAW_based sensor responded via a change in frequency, while SMO-based sensor response via a change in resistance. Each of these changes vary in magnitude proportional to the amount of analyte present. This paper will focus on aspects of: sensor selection, sensor data collection, manipulation, management, and processing for an integrated SAW and SMO sensor array.

The design of sensor arrays using polymer-coated QCRs for the practical detection of organic compounds in aqueous environments requires the appropriate selection of sensitive coatings. This selection of the coatings is performed through a study of the sorption processing terms of the partition coefficients of the analytes in the coatings. However, the determination of the partition coefficients in only accurate if all contributions to the total measured frequency shifts, (Delta) f_s can be accounted for. While mass loading is often assumed to be the dominant factor used in determining partition coefficients, viscoelastic effect may also contribute to changes in (Delta) f_s, especially in liquid environments. Both the effect of viscoelastic properties and the effect of the mass loading on the sensor responses are investigated by using a network analyzer and oscillator circuits. Different types of coatings including rubbery and glassy polymers are investigated, and the targeted analytes include classes of polar compounds, nonpolar compounds and chlorinated hydrocarbons. It is shown that changes in viscoelastic properties during and after analyte sorption can be significant enough to place the sensor in the non-gravimetric regime. However, for most applications invovling relatively low concentrations of organic compounds, the subsequent change in viscoelastic properties are still such that the sensors operate in the linear regime. In that case, the measured (Delta) f_s, hence the calculated partition coefficients can still be used for an approximate classification and selection of the coatings.

LRRS has been shown to have the potential to make Raman spectroscopy as practical and as widely used as IR spectroscopy. The advantages LRRS brings to Raman spectrometry are its order of magnitude lower price, size, and weight from that of laboratory grade Raman spectrometers. These allow the implementation of the Raman approach with all of its advantages in a small, lightweight, portable, and affordable instrument. The disadvantages of LRRS is a small degradation in performance from that of laboratory grade Raman spectrometers. It is shown in this paper that for one class of applications, measuring the concentrations of analytes in a solvent, the degradation in performance is insignificant. This class of application includes monitoring monomers and polymer concentrations during polymerization and the chemical constituents during crystallization processes.

A major source of ionic copper (Cu(I) and (II)) trace metal contamination in the marine environment is the copper containing antifoulant coatings on ship hulls. Tracing this contamination is difficult. Difficult because the ionic copper released from the coatings complexes rapidly with organic and inorganic ligands in the water column. This research examines the potential for using the organic dye 2,9-dimethyl-4,7-diphenyl-1, 10-phenanthroine (BCP) embedded in the ionomeric polymer, Nafion 117, as a membrane probe sensor for quickly measuring ionic copper in seawater. Results of this study show the Nafion 117 containing BCP measures Cu(I) within 17 percent of measurements made with the Standard Method bathocuproine di-sulfonic acid (BCS) procedure. To measure total ionic copper with the membrane probe, the reductant, hydroxylamine-hydrochloride, is added to the sample. The use of the polymer with BCP allows for a rapid optical measurement. Optical measurements are possible in two ways, as a comparator and as a modification to a fiber optic probe. Response times for these devices depend on the size of the membrane used. Presently, the optical comparator using a 2 X 3 membrane takes 20 minutes to develop color. The fiber optic probe uses a 0.5 cm diameter membrane that takes 1 minute. A comparator has no electronic parts and requires only a comparison of color depth to determine the level of ionic copper in the sample. This makes it useful for monitoring. The fiber optic probe is more suitable to buoy installation and long term monitoring where data logging and archiving are needed. Therefore, the Nafion 117 polymer impregnated with BCP is a rapid method useful for detecting ionic copper released in to the marine environment.

An instrument for in-situ monitoring of volatile organics in groundwater at pat-per-billion levels has been developed and field tested. The device is an acoustic wave thickness-shear mode sensor based on a 10-MHa AT-cut quartz resonator coated with a non-polar polymer film. The sensor demonstrates a detection limit of 8 and 12 parts per billion in water for trichloroethylene and toluene, respectively, and a rapid reversible response. This low detection limit is achieved by carefully minimizing the noise level in the electronic detection system, by optimizing the thickness of the sensing polymer film, and by performing measurements in the headspace. Preliminary field test demonstrated good correlation of sensor response with conventional laboratory purge-and-trap/GC analysis.

As the technology of gas turbines push the firing temperatures and pressures higher, the sensitivity of the turbine's performance on fuel quality is expected to increase. This, combined with the gas industry's practice of mixing several sources of natural gas in a given pipeline, may make the need for on-line fuel quality sensors more critical in the years ahead. This report describes the development of a new sensor system to determine the dewpoint of natural gas, an important indicator of fuel quality.

Microhotplate gas sensors can be operated in rapid temperature programmed sensing (TPS) modes due to their small size and mass. The temperature sequences can be optimized in order to enhance a sensor's sensitivity and to discriminate among similar volatile organic compounds. In the paper, we demonstrate that quantification of a single component gas in air with a microhotplate gas sensor operated in a TPS mode is possible by using partial least squares techniques. The calculations of the PLS models are based on gas sensitivity which is the ratio between the dynamic responses to the test gas and to air. We show that the gas concentration can be predicted accurately for different test gases, different temperature sequences, and different metal-dosed microhotplate gas sensors. THus the qualitative and the quantitative analysis of a single component gas in air can be achieved with a single microhotplate gas sensor operated in TPS mode.

This paper reviews several levels of interface electronics for chemical sensing system based on physical sensors. Many physical sensor, because of their interaction and vulnerability to the environment, have been eliminated from inclusion in sensing systems that require high precision and accuracy. With sufficient modeling, array design, and signal processing, however, these physical sensor are essential for providing a barrier to false alarms that is inherently not available in other sensors. We present our discussion under the assumption that physical sensor provide a useful dimension to solving chemical sensing problems. The responsibility of the interface electronics is to improve signal-to-noise ratio for a wider range and resolution of concentration detection and to extract features that enhance the robustness of chemical discrimination. Examples are presented of circuits that perform both of these functions for arrays of chemiresistors; these examples, aggregation of similar sensor signals and rank extraction, are a small subset of possible techniques that can be used to enhance the performance of physical sensors.

The feasibility of analyzing mixtures of spectrally similar analytes with low resolution mass spectrometers coupled with membrane introduction was studied. The performance of the multivariate calibration of an isometric ethyl benzene and p-xylene mixture remained essentially unchanged as the mass resolution degraded. The calibration performance also improved slightly as the data used for calibration decreased from the full mass spectra to only 12 or 18 fragment ions judiciously chosen by variable selection. The multivariate calibration and prediction for a more complex mixture of benzene and toluene only degraded slightly as the resolution decreased, while the result for the two isomers ethyl benzene and p-xylene got progressively worse. Depending on the variable selection algorithm and the number of fragment ions used, using only a select few ions for multivariate calibration and prediction gave results that were either similar to or slightly worse than results using the entire mass spectra. This paper demonstrates that mixture analysis performed with membrane introduction coupled to future miniature, low resolution mass spectrometers is possible.

Neural networks and fuzzy logic have merged as useful tools for the calibration of arrays of thin film gas senor. The choice of network parameters is essential for acceptable network performance. Often, choosing network-operating parameters involves a search of many possible candidate networks. When the neural network code is incorporated with other code, such as for data capture, presentation, and control, it is often the case that interdependencies are formed between these code segments. Frequently, enhancements, modifications, and fixes to the code lead to an extensive and time-consuming rewrite of many parts of the software. Thus, the need arise for neural network software modules that can be easily incorporated in application software but whose interface is well defined and whose implementation is entirely separate from the functionality it provides. By providing debugged and proven software modules encapsulating neural network functionality that can be simply inserted into any application, the entire software system can be modularized. These modules can be reused easily, and changing the neural network operating parameters no longer involves a complete software rewrite or even a recompile.

The outputs from a physics-based modeler of magnetometry data have been successfully used with a probabilistic neural network (PNN) to discriminate UXO from scrap. Model outputs from one location at a site were used to train a PNN model, which could correctly discriminate UXO from scrap at a different location of the same site. Data from one site location, the Badlands Bombing Range, Bull's Eye 2 (BBR 2), was used to predict targets detected at a different location at the site, Badlands Bombing Range, Bull's Eye 1 (BBR 1) containing different types of items. The UXO detection rate obtained for this analysis was 93 percent with a false alarm rate of only 28 percent. The possibility of discriminant individual UXO types within the context of a coarser two- class problem was demonstrated. The utility of weighting the sum of squared errors in cross-validation optimization of the (sigma) parameter has been demonstrated as a method of improving the classification of UXO versus scrap.

Over the past several years our group has been ex;poring the potential of sol-gel-derived glasses as platforms for advanced sensors and biosensors. In this paper we will outline the challenges that are associated with using these xerogels. Toward this end, we discuss recent results from our laboratories on the performance of polyclonal anti- dansyl antibodies sequestered within a series of xerogels. We assess antibody performance by determining the hapten/antibody association constant, the static excitation and emission for the dansyl hapten bound to the antibody combing site, and the excited-state fluorescence anisotropy and intensity decay kinetics for the dansyl/anti-dansyl system within a series of xerogels.

Conjugated polymers containing terpyridine receptors have been prepared and characterized which contain unusually high fluorescence sensitivity toward transition metal ions including Fe²⁺, Co²⁺, and Ni²⁺. This unique molecular structure combines inherently fluorescent conjugated polymers with a low lying electronic state formed upon binding between transition metal ions and terpyridine receptors on the polymer. The result is a dramatic fluorescence quenching response of the conjugated polymer. Fluorescence quenching measurements demonstrate quenching rate constants of 10^-9 M^-1s^-1 and as much as 5 percent quenching of the emission at concentrations of 4 X 10^-9 M in organic solution. The enhanced sensitivity of these structures is explained based on facile energy migration through the conjugated polymer system to any of the pendant excitation trapping sites. Thin films of these materials demonstrate further enhanced quenching efficiencies as well as moderate reversibility.

A new polymer material for pH sensing is described in this work. The optically sensitive membranes consists of lightly crosslinked poly(3-nitro-4-hydroxystyrene) microspheres embedded in a hydrogel matrix of polyvinyl alcohol (PVA). The microspheres are designed to swell and shrink as a function of hydrogen ion concentration accompanying the deprotonation and protonation of the hydroxyl group on the polymer backbone. The optical properties of the membranes are affected by both the change in size of the microspheres and the change in refractive index that accompanies swelling. These properties can be measured as a change in membrane turbidity at the interface between the polymer particle and the hydrogel matrix by a spectrophotometer or fiber optic sensor device. Microspheres are prepared by seeded emulsion polymerization of 4-acetoxystyrene followed by hydrolysis and nitration to introduce the nitro-phenol functional groups. The magnitude of the change in turbidity with pH changes were studies. Other parameters such as loading of the particles and ionic strength effects were also studied.

The response features of optical oxygen sensors based on luminescence quenching are complex and typically include downward curving Stern-Volmer plots and multi-exponential luminescence decays in the absence and presence of oxygen. The principle features and failings of some of the established models used to describe these response features are considered. A new model is introduced: 'a log-Gaussian distribution in (tau) ₀ and k_q,i ' model, or 'log- Gaussian' model for short. The key basic equations of this model are given and appear able to simulate the diverse response characteristics of many optical oxygen sensor. The model is used to fit the observed Stern Volmer plots and luminescence decays for two real sensors. The model appears to provide a reasonable physical basis for the very diverse response characteristics observed for optical oxygen sensors.

The sol-gel process is an exciting new technology that enables the production of gel glasses and ceramic materials at room temperature. Sol-gel technology offers simple methods for manipulation of the structure, configuration, composition and chemical characteristics of organic matrices. A novel longwave luminescent dye based on the europium luminescence initiated by a covalently bonded antenna fluorophore was designed, synthesized and characterized. The dye was successfully entrapped into various sol-gel and ormosil matrices and consequent optical- , leaching- and light fastness-properties were tested. Finally, sensor layer based on TMOS doped with Eu³⁺- complex and bromothymol blue was found to be most appropriate for purposes of sensing pH over the range 5-10.

Research in our group has shown that a multiple channel laser-induced fluorescence (LIF) system is an effective tool for in situ monitoring of polyaromatic hydrocarbons (PAHs). The system delivers ten laser beams to a samples and simultaneously detects the fluorescence signals from the individual channels, generating an excitation-emission matrix (EEM) of the sample. Speciation of chemical present in a mixture EEM is conducted by target factor analysis. To determine the capability of the technique for speciation, various aromatic compounds were measured with the LIF system. The pure-component EEMs were analyzed against each other as targets to separate the compounds into classes with the target factor analysis method. Then EEMs of mixtures containing 2-5 components were analyzed using the pure- component EEMs as standards. In this preliminary study of eleven target aromatic compounds, nine were successfully identified individually, while two spectroscopically very similar compounds, naphthalene and 2-methylnaphthalene, could be differentiated from the other nine but not from each other. In such cases, the technique can correctly identify the presence of the class, e.g., naphthalenes, instead of an individual components, and the quantitative results for the class must be interpreted accordingly.

Excitation-emission matrix (EEM) fluorescence spectroscopy with multi-way spectral deconvolution has proven successful when applied to the detection and quantification of aqueous pesticides. However, resolution becomes difficult when the EEM spectrum of one or more pesticides overlap the Raman scattering. The diagonal pattern across the spectrum is inefficiently modeled by trilinear multi-way spectral deconvolution and calibration methods. Weighted parallel factor analysis (W-PARAFAC) may be employed to reduce this effect. There are tow principal classes of weighting strategies, positive and negative. Positive weighting enhances each component's signal, while negative weighting focuses on eliminating any nonlinearities present in the background. These two classed may be further divided into the subgroups of hard and soft weighting. Hard weighting is a binary weighing strategy, while soft weighting applies a continuous distribution of weights. The optimal weighting strategy for a model is dependent on the analytes; signal intensity and overlap with the background scattering. When Raman scattering is present in the EEM spectra, a W-PARAFAC model will always exceed the performance of a model lacking background correction.

The validity of using Beer's Law to describe liquid waveguide capillary cells (LWCC) as absorption cells with increased optical pathlength was investigated. Experimental and theoretical results for two types of LWCC are presented. 'Type I' LWCCs are constructed with solid TEFLON AF tubing. 'Type II' LWCCs consist of quartz tubing with an outer coating of TEFLON AF. UV/Vis absorbance spectra versus chromophore concentration were found to be linear for both LWCC types within the wavelength range and absorbance accuracy of the spectrophotometer used. The ratio between 'effective' and 'physical' pathlength, EPLR was determined experimentally for both LWCC types. Type I cells had an effective optical pathlength that was statistically indistinguishable from the physical pathlength on a 95 percent probability basis. Type II cells had an effective optical pathlength that was slightly shorter than the physical pathlength, dependent on the cell's inner diameter and wall thickness. A theoretical model explaining Type I LWCC result is presented. Our results indicate that Beer- Lambert's Law can be applied to both types of LWCCs for UV/Vis absorption spectroscopy.

This paper presents integrated fault detection capability of the Spreeta biosensor technology. A specific feature is discussed based on multi-point image characterization. Multi-point image characterization provides a means to monitor biosensor surface damage, as well as sample anomalies such as macroparticulates are bubbles.

The integrity of most currently used fluorescence data analysis algorithms requires that a fluorophore present in a sample display the same spectral signature, apart from a single concentration-dependent scaling factor, regardless of its concentration or the concentrations of other components in the sample. Inner filter effects impede the utility of fluorescence spectroscopy by perturbing this spectral signature, rendering identification and quantification of fluorescent analytes difficult. The photophysical background theory of inner filtering is presented along with a comparison of existing correction schemes. The dual- pathlength method, a novel and simple experimental procedure for inner filter effect compensation, is introduced. The role of inner filtering in JP-4 a jet fuel contaminant detected by our laser-induced fluorescence/cone penetrometer instrument, is evaluated.

In the petroleum industry, the fluid produced from underground reservoirs can be a mixture of crude oil, water and gas. Knowing their volume fractions and flow rates at each point of entry along the well are essential for optimizing the production. An optical probe system has been developed to achieve these goals. It uses a laser- or LED- beam directed to the tip of a tiny optical probe situated in the flow stream. By measuring the reflection from the probe tip, gas can be distinguished from liquid due to their large difference in refractive index. In addition, fluorescent characteristics of crude oils are used to distinguish oil from non-oil. By integrating the time fraction of individual events in reflection and fluorescence respectively, the volume fraction of the three phases can be determined. In addition, the leading edge of the fluorescence signal is used to determine the speed of the oil drops, which, combined with the volume fraction of oil, gives the oil flow rate. The system requires no calibration for the volume fraction measurement; however, a simple calibration is needed for determining the oil drop speed. This article shows the apparatus of the probe system, response of different probe designs and data from lab experiments.

Novel, field portable, inexpensive, easy to fabricate, optical fiber waveguide Surface Plasmon Resonance (SPR) sensors have been developed. Such sensors show exceptional promise in process and environmental monitoring. These sensors generate SPR features that are broader than prism- based SPR features. While these wide features are troublesome to calibrate using the univariate method traditional to SPR, multivariate methods of calibration have been applied to these new SPR sensors. Multivariate calibration has shown a fifty-percent improvement in predictive accuracy compared to univariate methods. In select case, the multivariate calibrations have been an order of magnitude better than the univariate methods. Traditional and new calibration methods for SPR are discussed, and recommendations are made regarding choice of regression modes, data density, and calibration set makeup for calibration of SPR using novel optical fiber waveguide SPR.

The functional design of the smart electronic nose using polymer-film coated quartz resonator gas sensors, based on the solubility parameter of sensing membrane and gases, is carried out in order to develop the sensor with excellent selectivity and high sensitivity for harmful gases such as toluene, acetaldehyde and ammonia gases. The polymer-films such as propylene-butyl, poly-carbonate and acrylic-resin of which the solubility parameter almost coincide with that of toluene, acetaldehyde and ammonia gases, respectively, are chosen as a sensing membrane material coated on the quartz resonator. It is found that propylene-butyl-film coated quartz resonator gas senor exhibits a high sensitivity and an excellent selectivity for toluene and p-xylene gas, as it is expected from the functional design based on the solubility parameter. It is also found that poly-carbonate- film coated sensor and acrylic-resin-film coated sensor exhibit high sensitivity and excellent selectivity for acetaldehyde and ammonia, respectively, as it is also expected. The results strongly suggest that the solubility parameter is effective to functional design of the sensing membrane of quartz resonator gas sensors. The identification of gas kind is successfully possible by the principal component pattern recognition analysis of the transient response of the each sensor for gases.