TechArchive

This article was originally written in the period 1995-2000

Christer Helenelund, Product Manager of the Sensor Systems Division of Vaisala, explains the thinking behind the new Carbocap optical carbon dioxide sensor.

Numerous technologies for measuring gas concentration have been developed over the years. The list includes chemical sensors, ceramic sensors, pelletisers and various types of electrochemical sensors. Each one, however, is only suitable for a limited number of applications. The environment must be sufficiently non- aggressive and frequent recalibration of the instruments is required.

Non-flammable gases like carbon dioxide have proven particularly puzzling. Chemical sensors are difficult to develop and they often have a short service life. Indirect measurement methods tend to suffer from poor sensitivity, cross sensitivity to other gases, low accuracy or drift.

Frequent recalibration, however, is not a realistic option for most users, especially for high volume or low-end applications. For this reason, there is a need for gas sensors that are stable enough to function accurately and reliably for long periods of time.

The problem with optical devices, on the other hand, has been the rice. With higher production volumes, however, prices could be decreased by integration and miniaturisation. Non-dispersive infra-red (NDIR) is one subset of a range of optical measurement techniques.

Gases absorb light. The type of gas determines the wavelength of light that it absorbs. Carbon dioxide, for instance, absorbs infra-red light (wavelength 4.26æm). In optical gas detection, gas is introduced to a measurement cell with a light source at one end and an interference filter and a detector at the other. The filter cuts out all other wavelengths, and the detector measures the amount of light that has passed through the cell. The amount of light reaching the detector depends on the amount of gas.

There are three commonly used NDIR methods:

• Single-beam, single-wavelength
• Dual-beam, dual wavelength
• Single-beam, dual wavelength

Vaisala’s Carbocap is a completely new NDIR sensor for the measurement of CO₂. Though based on the non-dispersive intra-red (NDIR), single-beam, dual- wavelength operating principle, it combines the performance of the single-beam, dual wavelength with the compactness, simplicity and reliability of the single-beam, single wavelength method. In particular, its drift characteristics as a function of time and temperature are exceptional.

The difference is in the filter – a micromachined, electrically tuneable Fabry-Perot interferometer (FPI). The FPI ensures the high accuracy and stability of this dual- wavelength instrument, without the problems of mis-matched filters and detectors, or the wear and tear of a rotating filter. The FPI is the only moving part, so the system is virtually solid-state.

The infra-red source at the end of the measurement chamber is electrically modulated (ie turned on and off continuously). Light enters the gas chamber, where the gas absorbs a fraction of the photons at a certain wavelength. The FPI is tuned so that its pass band coincides with the absorption wavelength of the gas. The detector measures the strength of the signal that gets through.

The pass band of the FPI is then shifted to either side of the absorption band – this band must not have any interfering absorption lines. Because no absorption occurs at this point, only the light transmission in the system is measured. This constitutes the reference signal. The ratio of these two signals indicates the degree of light absorption and thus the gas concentration. The measurement system can be truly miniaturised, since the mechanically rotating filters in the traditional method have been replaced by a small silicon component, the FPI.

The Carbocap prototype was designed and built using an optical path length of 50mm. The temperature dependence is extremely low and its stability is excellent. at 0 ppm, CO₂instability is negligible and at 2000 ppm it is well within 100 ppm. The susceptibility to variations in the optical transmission shows that the ratio between the measurement and reference channel varies by only 0.6% with a 50% reduction in the intensity of the lamp. So the new sensor is very insensitive to any deterioration of the optical transmission in the path.

Traditional NDIR optical gas sensors

Single-beam, single wavelength devices, based on one optical channel and one wavelength, offer the poorest performance. Stability in particular is easily affected by ageing of the lamp, contamination or changes in the reflecting properties of the optical path. Many of these devices are fairly unstable. Temperature changes also have an adverse effect on short-term stability. But the structure of these devices is simple and they are mechanically reliable and inexpensive.

Dual-beam, dual-wavelength devices have two optical channels, two detectors and two interference filters. Although these devices are more accurate and more stable, they are also more expensive. To achieve good performance over a larger temperature range, the detectors must form a perfectly matched pair. Another problem is non-symmetrical contamination of the two channels.

The single-beam, dual-wavelength structure is the most expensive – and the most accurate. This device has only one open channel, but the wavelength can be changed by switching the interference paths in the optical path. This is usually done with a rotating filter wheel. Since only one light source and one detector are used, temperature dependence and non-symmetrical contamination are not a problem. The filter wheel is bulky and difficult to protect from dust, however, and the filter wheel must be changed from time to time.

• Vaisala
• Tel: 01638 674400
• Fax: 01638 674411
• Contact: Helen Carson (Marketing Manager)