P5.3 - Efficient Calibration and Recalibration of Metal Oxide Gas Sensors by a New Mathematical Procedure
- Event
- SENSOR+TEST Conferences 2009
2009-05-26 - 2009-05-28
Congress Center Nürnberg - Band
- Proceedings SENSOR 2009, Volume II
- Chapter
- P5 - Gas-Sensors
- Author(s)
- R. Seifert, H. Keller - Forschungszentrum Karlsruhe GmbH, Eggenstein-Leopoldshafen, Germany, K. Frank, H. Kohler - Hochschule Karlsruhe, Karlsruhe, Germany
- Pages
- 393 - 397
- DOI
- 10.5162/sensor09/v2/p5.3
- ISBN
- 978-3-9810993-5-5
- Price
- free
Abstract
There is a growing need of economic online and in-situ field analysis applications like discriminated monitoring of toxic gas leakages, online monitoring of volatile components in chemical and biochemical processes, quality monitoring in food processing, etc. In this context, the isothermally operated metal oxide gas sensors (MOGs) with tin oxide as base material are manifold introduced due to their high sensitivity, long-term stability and low price. Their sensitivity to specific gas components, however, cannot be cultivated with high discrimination to others. On the other hand, it could be shown that by periodic variation of the sensor temperature and simultaneous sampling of the so called Conductanceover-Time-Profiles (CTPs), by appropriate choice of sensor additives to the tin oxide material and by using an innovative mathematical calibration and evaluation procedure, valuable signal information can be extracted to be numerically analysed for substance identification and concentration determination. Although MOGs are very low-priced with respect to production and operation, they show, nevertheless, the disadvantage that for more accurate monitoring tasks every individual sensor element has to be calibrated. This is very time consuming and expensive. For a mono-component analysis, a single sensor element typically has to be measured for calibration at about five calibration points, i.e. dosed concentrations of the gas under consideration have to be adjusted, and the resulting signal patterns have to be sampled.
These extraordinary expenses for calibration are a consequence of unavoidable production inaccuracies of the sensor elements which lead to unreproducibilities of the gas analytic attributes. Therefore, although the signal patterns of the various sensor elements of a production batch are quite similar, each sensor element (SE) has to be costly calibrated, in order to yield high analytic performance. The same is true if a sensor chip has to be exchanged in case of a defect or simply has to be recalibrated.