The design and optical performance of a small-footprint, low-power, turnkey, Point-And-Stare hyperspectral analyzer,
capable of fully automated field deployment in remote and harsh environments, is described. The unit is packaged for
outdoor operation in an IP56 protected air-conditioned enclosure and includes a mechanically ruggedized fully reflective,
aberration-corrected hyperspectral VNIR (400-1000 nm) spectrometer with a board-level detector optimized for point
and stare operation, an on-board computer capable of full system data-acquisition and control, and a fully functioning
internal hyperspectral calibration system for in-situ system spectral calibration and verification. Performance data on the
unit under extremes of real-time survey operation and high spatial and high spectral resolution will be discussed.
Hyperspectral acquisition including full parameter tracking is achieved by the addition of a fiber-optic based
downwelling spectral channel for solar illumination tracking during hyperspectral acquisition and the use of other
sensors for spatial and directional tracking to pinpoint view location. The system is mounted on a Pan-And-Tilt device,
automatically controlled from the analyzer's on-board computer, making the Hyperspec<sup>TM</sup> particularly adaptable for base
security, border protection and remote deployments. A hyperspectral macro library has been developed to control
hyperspectral image acquisition, system calibration and scene location control. The software allows the system to be
operated in a fully automatic mode or under direct operator control through a GigE interface.
The technical challenges of applying hyperspectral imaging techniques to on-line real-time food monitoring is discussed.
System optimization must be applied to the design of the hyperspectral imaging spectrograph, the choice and operation
of the imaging detector, the design of the illumination system and finally the development of software algorithms to
correctly quantify the hyperspectral images. The signal to noise limitation of hyperspectral detection is discussed with
particular emphasis on the detection of moving objects at high measurement bandwidths. An example is given of the
development of a simple but accurate algorithm for the detection and discrimination of rust particles on leaves.
The use of hyperspectral technology in the NIR for food quality monitoring is discussed. An example of the use of
hyperspectral diffuse reflectance scanning and post-processing with a chemometric model shows discrimination between
four pharmaceutical samples comprising Aspirin, Acetaminophen, Vitamin C and Vitamin D.
On-line real-time monitoring of the gas-phase concentrations of sulfur containing molecules such as sulfur dioxide (SO2), hydrogen sulfide (H2S), carbon disulfide (CS2) and nitrogen containing molecules such as nitric oxide (NO), nitrogen dioxide (NO2), nitrous oxide (N2O) and ammonia (NH3) is of major importance in pollution monitoring and reduction and in the optimization of many gas-phase industrial processes. A UV optimized non-solarizing fiber-optic based diode-array analyzer system utilizing a 10,000 hour MTBF Xenon pulsed source has been developed and proven on-line. An on-board chemometric prediction engine allows for the simultaneous multi-component analysis of measured spectra of sample gases in real-time. Fiber-optic coupling of the analyzer to the gas flow cell housed within the sampling system allows intrinsically safe measurement to be carried out on samples gases at temperatures up to 310 C and pressures to 60 barg. Detection sensitivity down to ppm levels have been realized including such measurement applications as NH3, NO and H2S monitoring.
A diffuse reflectance spectroscopy system is described which can operate in a contact and non- contact mode on powders, slurries and other diffusely scattering materials. Diffuse reflectance spectra are presented for a number of samples including common household materials. A comparison is made of the probe with a Bio-Rad diffuse reflectance accessory. Second derivative spectra are shown of a calibration mixture of polymer additives. The use of the diffuse reflectance system for non-destructive tablet hardness measurements is discussed. Sensor multiplexing for diffuse reflectance spectroscopy is reviewed.
The quantification of chemical processes in real time using infrared fiber-remote absorption spectroscopy requires the accurate sampling of the chemical species at a well defined and reproducible absorption pathlength. Sensors and sampling methods are described which are tailored to the type of process and process conditions being monitored. The optical fiber parameters which effect process sampling and measurement accuracy are reviewed. Sampling techniques are discussed for homogeneous liquid phase and gas phase measurements. The use of flow through cells for side stream monitoring and immersible probes for in-line monitoring of liquid phase streams are covered in detail. Sensor pathlength calibration of a variable path immersible transmission probe is discussed.
Optical fiber-based Fourier transform infrared (FTIR) is investigated as an in-situ monitoring tool for two applications in organometallic chemical vapor deposition (OMCVD): measurement of the concentration of organometallic precursors fed to the system, and detection of gas-phase reactions relevant to the process. The feasibility of the first application is demonstrated using fluoride fibers in case studies with trimethylindium and trimethylgallium. With a short single pass gas-cell, a minimum detection limit of 0.5% is achieved for a one minute scan time. Further reduction in this limit may be realized by improved cell design and longer scan times. The second application of the FTIR technique, in situ monitoring of gas-phase reactions, is also demonstrated. The results are in excellent agreement with previous reported data for the pyrolysis of the two precursors.
Remote spectroscopic sensing in the infrared places tight constraints on measurement system stability and analog signal integrity. The stability of each component of the system needs to be considered, including detectors, electronics, fiber cables, spectrometer, optical components and light sources. A full discussion of the signal-to-noise limit of a fiber remote systems is given. The parameters affecting the ability of the system to be used for long term process sensing are reviewed. Environmental data is presented on the optical throughput stability of infrared fiber optics, fiber-optic cables and sensors with changing temperature. The effect of water and vibration on bare and protected infrared fibers is discussed. The measurement stability of each component of a FT-IR remote fiber-optic system is related to the final measurement stability of the complete system. It is shown that, within certain environmental limits, the signal-to-noise limit of the measurement may be realized with careful system configuration and calibration.
The evanescent-wave spectrum of a sample surrounding the core of an optical fiber is a complex function of the optical constants of the media involved as well as the geometry of the sensing fiber. We develop a simple theory for evanescent-wave absorption in the weak absorption limit where we show that the absorbance of a length of sensor fiber may be related linearly to the bulk sample absorption coefficient. We present experimental data that verifies the observed scaling between the evanescent-wave absorbance and the bulk absorption coefficient for an isopropanol sample. The application of evanescent-wave spectroscopy with different sensor fiber materials is discussed, along with experimental and theoretical data for the enhancement of evanescent-wave spectroscopy using tapered fibers. Finally we discuss the results of a numerical series of calculations based on the exact ray paths of radiation within the fiber and the fundamental theory of ATR absorption at an interface assuming a plane wave approximation. In the more complex theory the evanescent-wave absorption coefficient is a decreasing function of the bulk absorption coefficient.
A series of FT-IR spectrometer based remote sensing systems have been developed taking advantage of the
new technology of IR transmitting optical fibers. The systems may be used to monitor the chemical composition
of solid, liquid and gas phase samples. An array of remote sensors may be interfaced to a single FT-IR
spectrometer through a multi-fiber launch module. An optical channel selector (OCS) allows the sensors to be
addressed with a single opto-mechanically multiplexed detector system. Remote collimated beam sensors have
been developed for web monitoring and liquid and gas phase sensing. An optimized multi-detector web
monitoring system has been developed for moving web sensing on optically dttuse webs. Quantitative data will
be presented for a number of remote spectroscopic measurements.
Infrared transmitting heavy metal fluoride optical fiber has been used
to separate an FTIR analyzer from a remote measurement point. Several
types of remote sensors have been developed for species concentration
measurements. Remote transmission cells connected to fiber cables have
been used for the measurement of spectra of liquids and gases.
Evanescent wave probes have been developed to obtain spectra in highly
absorbing and highly scattering media. Remote spectra taken with an FTIR
fiber-optic analyzer in the 8000 - 2000 cm1 spectral region are
presented. A calculation of detectability limits for these species
based on the measured data will be presented. A discussion of sensor
multiplexing applied to remote fiber optic FTIR spectroscopy will be