For an explanation of some of the terms in this list, see our Mini-Glossary.
There is no best way to categorize the whole range of gas sensors. Our list is arranged arbitrarily as (1) reactive, (2) physical property, and (3) sorptive sensors, depending on how they transduce the presence of a chemical into an electrical signal.
Reactive sensors generate a signal by measuring some aspect of a chemical reaction of the analyte. The analyte may be destroyed in the process. The smallest and cheapest sensors tend to be reactive types.
Electrochemical Sensors - A more precise term for this type is the "porous-electrode amperometric gas sensor". It is also called a fuel cell sensor. It will respond to gases that can be electrolytically reduced or oxidized on a metallic catalyst such as platinum or gold. Typical gases measured are O2, CO, NO2, NO, and H2S, and organic vapors such as alcohols, aldehydes, or ketones. Typical sensitivities are in the 3-30 PPM range, but some proprietary sensors are capable of detecting as little as 2 PPB of gases such as ozone, NO2, H2S or arsine. Selectivity is generally modest unless auxiliary methods are used.
Solid State Semiconductor Sensors - This sensor typically consists of a bead of tin oxide formed around two fine coils of platinum wire. When the bead is heated using one of the coils, the analyte gas will oxidize on the bead surface, changing the electrical conductivity as measured between the heated and unheated coils. Nearly all oxidizable gases can be detected on the SSS sensor. Selectivity is poor, but it can be modified slightly by changing the operating temperature and by doping the tin oxide with various elements. SSS sensors are inexpensive, but the sensitivity is not good (20-100 PPM). There is also a serious problem with baseline drift.
Combustible Gas Sensor - The electrical resistance of most metals will increase with temperature. The combustible gas sensor consists of little more than a coil of platinum wire which is electrically heated. When gases combust on the surface, some of the heat of combustion is transferred to the wire coil. The increase in coil temperature is reflected as an increase in electrical resistance.
Flame Ionization Detector (FID) - The FID works by burning the analyte gas in a hydrogen flame. In this environment, organic compounds produce positive ions, which are collected at a cylindrical electrode above the flame. A very small current will be generated between the collector and the metal flame jet. The FID is very sensitive and linear over many orders of magnitude. Because of the needs for hydrogen and a mechanically stable environment for the flame, the resulting instruments are complex, but some companies such as Foxboro have succeeded in making reliable FID instruments. FIDs are nearly nonselective among organic compounds, but they do limit their responses to organic compounds only.
Chemiluminescence - Certain chemical reactions generate light, which can be measured with great sensitivity. The most common application of chemiluminescence in gas detection is the measurement of nitric oxide by reaction with ozone; measurements of NO can be made down to the parts per trillion (10-12 level. NO2 is usually measured in the same instrument, by reducing it to NO before reaction with ozone. An older, and now obsolete, monitor for ozone used the chemiluminescent reaction with ethylene gas.
Physical Property Sensors
Physical property sensors generally leave the analyte gas undisturbed, and measure some property such as absorption of light or thermal conductivity. These sensor types tend to be more complex and expensive, but also more selective.
Nondispersive Infrared (NDIR) - These are the simplest of the spectroscopic sensors. The key components are an infrared source, a light tube, an interference (wavelength) filter, and an infrared detector. The gas is pumped or diffuses into the light tube, and the electronics measures the absorption of the characteristic wavelength of light. NDIR sensors are most often used for measuring carbon dioxide. The best of these have sensitivities of 25-50 PPM. Typical NDIR sensors are still in the $200- $1000 range. Most are used for carbon dioxide, because no other sensing method works reliably for this gas.
Spectroscopic Sensors - These use conventional means to generate monochromatic light in the ultraviolet or infrared and to measure its absorption by a gas. An ultraviolet spectrometer, for example, is the 'gold standard' method for measuring ozone. Specific organic compounds can sometimes be individually measured by measuring absorption of infrared light at one or more wavelengths.
Photoacoustic Sensors - If a short pulse of infrared light is passed through an absorbing gas, the absorbed light energy becomes heat. The sudden expansion of the gas generates a pressure, or acoustic, wave, which can be measured with a microphone. These are a variation of infrared spectroscopic sensors, with an important twist: The PA sensor measures the light absorbed by the sample. This is in contrast to conventional spectroscopy, which measures the light not absorbed. Since photometric error is eliminated, very sensitive detection is possible. These sensors only became practical in recent years, when digital signal processing (DSP) chips became available to distinguish signal from background noise.
Many types of sensors depend on the physical or chemical sorption of the analyte into a coating on the sensing surface. Depending on the device, the sorption phenomenon may be detected by measurement of mass, refractive index, color change, electrical resistance, etc.
Fiber-Optic - A thin glass or plastic fiber is coated with a thin layer of a compound that will absorb the analyte. When light is passed through the fiber and reflects from its inside surface, some of the light energy extends beyond the surface of the fiber. This effect is known as the evanescent wave, and its influence is usually no more than a few nanometers. A simple surface coating may absorb organic gases, changing its refractive index. The amount of light reflected inside the fiber is changed; this is detected by a receiver at the other end of the fiber from the light source. Other surface coatings may react with the analyte gas and change color, which will affect the spectrum of the reflected light. Fibres coated with pH indicator dyes, for example, have been used to measure pH in hard-to-reach places.
Microbalances - The simplest form of this sensor uses a quartz crystal which is electronically made to vibrate at its natural frequency. The crystal is coated with a material that absorbs the analyte gas. The mass of the coating increases and slows down the natural rate of vibration of the crystal. The resulting frequency shifts can be measured electronically with great sensitivity. The most commercially successful device of this type uses the amalgamation reaction between mercury vapor and a gold coating on the crystal to measure traces of Hg. The basic microbalance has been elaborated into more sophisticated devices such as SAW (surface acoustic wave) and resonator devices, which are more sensitive than the simple bulk crystal. This class of sensors is sometimes referred to as gravimetric.
Conductive Polymer - Certain polymers, such as polyanilines and polythiophenes, are electrically conductive. The conductivity changes when certain gases are absorbed by the polymers. The polymers can also be "tuned" to certain compounds by carrying out the polymerization in the presence of the analyte. Although there are many technical problems, like temperature- and humidity-sensitivity, the potential low cost and the "tuning" feature are attracting a great deal of attention from developers.
Elastomer Chemiresistor Sensors - These measure the very slight physical expansion of a film of an elastomeric material that occurs when it absorbs a gas. The elastomer, silicone rubber, for example, contains electically conductive particles such as carbon. The concentration of particles is adjusted so that there are relatively few conducting paths through the elastomer. Slight expansion of the elastomer causes some of these paths to be broken, and the electrical resistance rapidly increases. Following the same mathematics that describes the "percolation" of water through a bed of gravel, the resistance changes very rapidly as the polymer swells.
Reactive-Gate Semiconductor Devices - Most active semiconductor devices, such as MOSFET transistors, use voltages directly applied to the gate to control the flow of charge carriers. Chemically-sensitive devices, however, use a chemical interaction to change the transconductance. For example, the gate of a MOSFET may be plated with palladium, which selectively absorbs hydrogen gas.