Copyright © 2006 Penton Media, Inc., All rights reserved.
Printing of this document is for personal use only.
Reprints

To identify compounds or investigate sample composition, engineers often turn to infrared (IR) spectroscopy. Correlation tables can be found in various resources. One is available online at en.wikipedia.org/wiki/Infrared_spectroscopy_correlation_table.

IR spectrometry works because molecules can absorb energy at specific frequencies determined by the shape of the molecular potential energy surfaces, the masses of the atoms, and the associated vibrational coupling. Diatomic molecules with their single covalent bond can only elongate and contract, producing a single vibrational mode, with harmonics. More complex molecules, with more bonds, have more complex signatures. Families of related organic molecules can be identified by their similar signature absorption spectra (Fig. 1).

It’s common to perform IR spectroscopy on gases by using a single sensor that monitors a beam of infrared light that passes through a sample of the gas. From this, a transmittance or absorbance spectrum can be produced and analyzed. Solids can be analyzed reflectively or by looking at their combustion products.

A single sensor is adequate if the application simply involves identifying the presence or absence of certain chemicals. For example, it could be water in a stream of gas used in some manufacturing process or a contaminant in a stream of anesthetic in a hospital operating theater. Obviously, a 2D or 3D array of sensors offers opportunities for actual imaging.

SINGLE SENSORS
Cal Sensors, a company deeply steeped in single-sensor lore, recently introduced a pair of lead-salt IR arrays. The company specializes in thin-film sensors made from lead salts, specifically lead sulphide (PbS) and lead selenide (PbSe). Lead-salt sensors have less sensitivity (and cost less) than indium-gallium-arsenide (InGaAs) sensors, but are responsive to wider spectra: roughly 1 to 3 µm for PbS and 1 to 5 µm for PbSe.

According to Brian Elias, director of engineering, much of the market for his company’s single sensors is in checking for the presence of dangerous but odorless gases, such as carbon monoxide and dioxide, or oxides of nitrogen, or the presence of water vapor. He added that a strong new application comes from the recycling industry. Many recycled plastics look alike but require different handling at the recycling center. IR makes it possible to separate them readily.

A great deal of information about lead-salt detectors is available online. David A. Kondas notes in “Technical Report ARFSDTR- 92024: Introduction to Lead Salt Infrared Sensors” (www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA260781&Location=U2&doc=GetTRDoc.pdf) that, by the end of World War I, research was being conducted on various lead-salt materials for IR detection and communications applications. In the 1930s, there was considerable research in Germany on lead-salt infrared detectors for military applications. Eventually, the United States military began to study these detectors.

Infrared devices may be either “thermal” or “photon” (quantum) detectors. The most common type of thermal detector is the bolometer. Two types of photon detectors exist: photoconductors and photovoltaics (see the table).

PbS and PbSe detectors are photoconductive. Fabricated using thin-film techniques, they undergo a change in conductivity when exposed to radiation. When incident photons with energy levels in the infrared region bombard the surface of the thin-film semiconductor, they collide with electrons that have energy levels within the valence band of the detector material. This kicks the electrons across the material's energy bandgap and generates electron-hole pairs that can produce a current in the presence of an external electrical field.

If there’s an external voltage across the detector, changes in current can be detected. In simple terms, the more photons, the more current, but only if the photons are at the wavelength to which the material is sensitive. Incidentally, the photoconductor materials in the table are intrinsic semiconductors. They don’t need doping.

More concretely, there’s a cutoff wavelength for photons below which valence electrons can no longer acquire sufficient energy to promote the production of electron-hole pairs. That wavelength is given by ?_cutoff = hc/E_G, where h is Planck’s constant, c is the speed of light, and E_G is the bandgap energy of the photoconductor. For a PbS detector with an energy bandgap of 0.42 eV at 295 K, the cutoff wavelength would be 2.9 µm. Similarly, for a PbSe detector with an energy bandgap of 0.23 eV at 195 K, the cutoff wavelength is 5.4 µm, as stated by Elias.

According to Kondas, some IR detector systems directed at a fixed target may include a center-spun reticule or chopper wheel somewhere between the path of incoming radiation and the detector. There may also be a detector cooler to increase the responsivity, dark resistance, and detectivity at longer wavelengths.

Detectivity, or D*, is the reciprocal of noise equivalent power (NEP), normalized to unit area and unit bandwidth. It’s expressed in units of cm × vHz/W. NEP is the signal power that gives a signalto- noise ratio (SNR) of 1 for an integration time of half a second.

Most lead-salt detectors are designed to operate at three temperatures: ambient (295 K), intermediate (193 K), and low (77 K). The latter two correspond to the boiling point of Freon 13 and the boiling point of liquid nitrogen, respectively.

Generally, Elias said, resistance decreases consistently with temperature. For PbSe, the rate is 2.7%/°C, from –30°C to –80°C. For PbS, the rate is 3.4%/°C, from 23°C to –40°C (Fig. 2).

As photoconductive devices, the sensors require an external bias voltage. Typically, supplying it is straightforward. The sensor is the lower resistor in a voltage divider, and the signal at the top of the voltage divider is applied to a unity-gain buffer amplifier. In general, Elias said, the sensor signal is linearly related to bias voltage (up to approximately the detector maximum bias voltage).

Continue to page 2

At low bias values, PbS/PbSe detector noise shows relatively little dependence on voltage. However, after a given voltage value is reached, noise is linearly related to voltage. Larger detector areas require higher bias voltages to reach this range. Figure 3 shows how signal, noise, and SNR of PbSe detectors vary as functions of bias.

If chopping is used, detector noise is a function of the inverse of the chopping frequency. Therefore, at decreasing chopping frequencies, further diminished voltage bias may yield acceptable SNRs. Cal Sensors’ standard test rules are 50 V dc/ mm between electrodes for cooled and uncooled PbS, 35 V dc/mm between detector electrodes for uncooled PbSe, and 25 V dc/mm between detector electrodes for cooled PbSe. Thus, an uncooled PbSe 2- by 2-mm detector, which has a 2-mm active area between electrodes, would be tested with a 70-V dc bias.

PB-SALT ARRAYS
Cal Sensors’ new array products are the LIRA^5S Square Pixel PbSe Thermal Imaging Array and the MIRA4, a four-color sensor. Co-planar multicolor detectors typically consist of two or more detectors mounted side by side on a cooler cold plate or package header. Spectral filters are mounted above each detector in holders designed and built to minimize optical crosstalk between channels. The LIRA^5S is a 256-element multiplex thermal-imaging array for wavelengths from 1.0 to 5.5 µm. Its integrated electronics provide a 4-MHz data-readout speed. It can be programmed for high-speed hot-spot detection applications in manufacturing and assembly process lines, conveyor belts, buildings, and railway systems in which its 256 elements allow simultaneous measurement of that many discrete thermal points. Using PbSe provides for measurements at longer (cooler) wavelengths than alternative detector materials such as InGaAs. Previously, PbSe arrays with this type of sensitivity were custom units with long lead times.

Array elements are 40-µm square on 50-µm centers. They come in standard 28-pin packages. Product support includes a Windows-based evaluation/demonstration system with a USB interface board, USB cable, copper mounting block, heatsink with integrated fan, system and cooler power-supply modules, and array controller software.

The four-color MIRA4 PbSe sensor combines similar sensitivity with the ability to detect up to four distinct gases. Compared to the use of four individual detectors, it can reduce sensor cost in a system by as much as 60%. Applications include industrial and medical gas analysis, auto and aviation emissions analysis, underground applications (such as mining tunnels and walkways), and general industrial environmental monitoring, such as smokestacks and assembly lines.

BOLOMETERS
Older RF engineers are familiar with wattmeters based on Samuel Langley’s bolometer of 1878. In general terms, a bolometer consists of an “absorber” connected to a heatsink (area of constant temperature) through an insulating link. Radiation raises the temperature of the absorber relative to the heatsink, which may be exposed to the ambient environment or artificially cooled. In astronomy, they may be cooled down to a few hundreds of a Kelvin above absolute zero.

IR measurements use microbolometers, made of grids of vanadium oxide or amorphous silicon heat sensors on top of a silicon grid (Fig. 4). Honeywell developed microbolometers under a U.S. government contract in the mid-1980s. The government declassified the technology in 1992, and several companies have licensed the intellectual property (IP). Commercial microbolometers come in grid array sizes from 160-by-120 to 1024-by-768.

One microbolometer application is the Thermal Emission Imaging System (THEMIS) instrument on the Mars Odyssey Mission (Fig. 5). The orbiter began mapping Mars in February 2002 (see “Mapping Mars In Infrared”). Other Odyssey gear included the Gamma Ray Spectrometer (GRS) and the Mars Radiation Environment Experiment (MARIE).

The IR part of THEMIS, which also has a visible-light imager, covers 10 spectral bands: two at 6.78 µm and others at 7.93, 8.56, 9.35,10.21, 11.04, 11.79, 12.57, and 14.88 µm. THEMIS also boasts 100-m/pixel resolution. As initially conceived, THEMIS was designed to allow NASA to create global maps that showed the distribution of minerals across the Martian landscape. Carbonates, hydrothermal silica, sulfates, phosphates, hydroxides, and silicates have fundamental infrared absorption bands that THEMIS was tuned to.

Design-wise, THEMIS was engineered so the IR bolometer focal plane arrays (FPAs) and visible-light imagers could share the same optics and housing, while power supplies and I/O provisions were separate. The microbolometer FPAs were designed to operate at “ambient” temperatures. (Those can be mighty cold when you’re in orbit around Mars. Still, a thermal electric cooler stabilizes the IR FPA temperature to ±1 mK.)

Each array consists of 240 elements along 320 tracks with 50-µm dimensions in each direction. In fabrication, microbolometer arrays were grown directly on the surface of readout integrated circuits (ROICs), which allow image digital signal processing to be achieved at the focal plane.

Raytheon Vision Systems’ THEMIS design allows for a 100-m geometric instantaneous field-of-view (GIFOV) across a track approximately 32 km long. The FPAs were derived from a Raytheon handheld imager originally developed for rugged military use, which significantly reduced development cost compared to a custom design. They were produced by Raytheon Vision Systems under license from Honeywell.

OTHER SOURCES
Photoconductive sensors and bolometers do not comprise the entire world of IR sensor types, although they do make up a large part of it. Avalanche photodiodes (APDs) also enter into the picture, although APDs are used more often for fiber-optic communications than spectography.

With a long history (as Marconi and English Electric Valve) in sensors, the British company e2v offers Pellistor thermal-conductivity gas-detection sensors. These devices measure the change in heat loss (and hence temperature/resistance) of the detecting element in the presence of the target gas.

The company also has a range of IR sensors for specific gases, notably, but not exclusively, carbon dioxide. They include both source and detectors inside a small gas cavity/optical cell. Integral IR bandpass filters tune the sensors to the specific gases to be sensed.

At the far end of the cost/sophistication spectrum, Vigo System in Poland offers a range of mercury-cadmium-telluride (HgCdTe) and mercury-cadmium-zinc-telluride (HgCdZnTe) devices.