High-Performance Gas Sensing of CO ... - ACS Publications

May 8, 1999 - The two different sensor types investigated here were also compared to the well-known commercial Figaro TGS 822 sensor at similar ...
2 downloads 0 Views 85KB Size
Anal. Chem. 1999, 71, 2512-2517

High-Performance Gas Sensing of CO: Comparative Tests for Semiconducting (SnO2-Based) and for Amperometric Gas Sensors N. Baˆrsan,† J. R. Stetter,*,‡ M. Findlay, Jr.,§ and W. Go 1 pel†

Institute of Physical and Theoretical Chemistry, University of Tu¨bingen, Auf der Morgenstelle 8, D-72076 Tu¨bingen, Germany, BCPS Department, Illinois Institute of Technology, Chicago, Illinois 60616, and TSI Incorporated, P.O. Box 64394, St. Paul, Minnesota 64394

A comparison of the stability and sensitivity for two different sensor types (semiconductor SnO2 devices, amperometric electrochemical sensors) has been performed. Sensitivities and drifts in the signal and in the background for various concentrations of CO have been studied for thick-film SnO2 sensors (Pt and Pd doped) for a period in excess of 8 months. Similar performance data have been recorded for commercial amperometric sensors for a period in excess of 4 years. The two different sensor types investigated here were also compared to the well-known commercial Figaro TGS 822 sensor at similar concentrations. An objective approach for comparing different types of sensors has been developed using the “analytical sensitivity”. Two commonly used types of sensors for the detection of CO, NO2, and other toxic gases are the heated semiconductor SnO2 devices and the room-temperature amperometric electrochemical sensors.1-6 These gas sensors are used in many applications such as fire detection, hazard warning, process control, medical and environmental uses, and more recently consumer products such as the home CO alarm.7 Their sensitivity and stability are important performance parameters in trace detection. The stability of these sensors is crucial in consumer products that require no maintenance and long lifetime as well as in sensor arrays wherein multidimensional drift must be kept to a minimum. Further, it is important that the characteristics of the individual sensors are well understood and improved as much as possible before using the sensors in the most demanding applications or for extremely complex analytical tasks. To this end, a comparison of the stability and sensitivity of these different sensor types has been performed. †

University of Tu ¨ bingen. Illinois Institute of Technology. § TSI Inc. (1) Go ¨pel, W.; Schierbaum, K. D. Sens. Actuators B 1995, 1, 26-27. (2) Watson, J.; Ihokura, K.; Coles, G. S. V. Meas. Sci. Technol. 1993, 4, 711. (3) Ihokura, K.; Watson, J. The Stannic Oxide Gas Sensor, 1st ed.; CRC Press: Boca Raton, 1994. (4) Chang, S. C.; Setter, J. R.; Cha, C. S. Talanta 1993, 40, 461-467. (5) Cao, Z.; Buttner, W. J.; Stetter, J. R. Electroanalysis 1992, 4, 253-266. (6) Chemical and Biochemical Sensors, Part I; Go¨pel, W., Jones, T. A., Kleitz, M., Lundsto ¨m J., Seiyama, T., Eds.; VCH: Weinheim, 1991. (7) Stetter, J. R.; Pan, L. Amperometric CO Sensor for Residential Alarms. U.S. Patent 5,3312,310, July 19, 1994. ‡

2512 Analytical Chemistry, Vol. 71, No. 13, July 1, 1999

In order to interpret the data presented herein and perform a comparison of analytical capability, a short description of CO detection mechanism for both sensor types follows. SnO2 Sensors. Proposed Model of Response. In air, at temperatures between 100 and 450 °C the conduction of SnO2, an n-type semiconductor, is influenced by the chemisorption of oxygen and water vapor.8,9 Chemisorption takes place on the SnO2 surface and results in the trapping of free electrons, building a negatively charged layer at the surface. In the polycrystalline material studied here, there is a negative barrier at the surface of each grain which, when grains are placed together, forms the barrier to conduction at each grain boundary that each conducting electron experiences. During sensing of CO, COg (the upper case letter designations refer to the phase of the species) reacts with chemisorbed oxygen, decreasing its surface concentration and removing all or part of this barrier leading to increases in conductance. We have supporting evidence that the other reactions, which result from water vapor interactions, can also contribute to observed changes in conductance.10 Since the adsorption of CO may be described by a Freundlichtype isotherm over this temperature range, the removal of the barrier can be considered to occur exponentially. Low concentrations of CO create a large change in the initial conductance. Further contributing to the nonlinear change in conductance is the fact that the barrier height is not removed linearly with the addition of each CO molecule but the initial CO will remove the highest energy sites. These reasons, at least in part, explain both the high sensitivity of this sensor to low levels of CO and the decreasing sensitivity of this sensor with increasing CO concentration. The performance of SnO2-based sensors for detecting CO depends on many parameters, some of them “internal” the others related to the conditions in which the detection takes place. These parameters include the temperature and relative humidity of the sample atmosphere, the presence of interfering gases, the operat(8) Moseley, P. T.; Williams, D. E. Oxygen Surface Species on Semiconducting Oxides. In Techniques and Mechanisms in Gas Sensing; Moseley, P. T., Norris, J., Williams, D. E., Eds.; Adam Hilger: Bristol, 1991; Vol 2, p 46. (9) Kohl, D. Oxidic Semiconductor Gas Sensors. In Gas Sensors; Sberveglieri, G., Ed.; Kluwer: Dordrecht, 1992; Chapter 2, p 43. (10) Barsan, N.; Schweizer-Berberich, M.; Go¨pel, W. Fresenius J. Anal. Chem. submitted. 10.1021/ac981246d CCC: $18.00

© 1999 American Chemical Society Published on Web 05/08/1999

Table 1. Summary of Cross-Sensitivity (Major Interference) for M Series Electrochemical Sensors17a interfering gas sensorb

CO

H2S

SO2

NO2

NO

Cl

hydrogen

ethanol

NH3

CO-MNS CO-MFS H2S-MNS NO2-MNL SO2-MNS SO2-MFS NO-MNS NO-MFS

100 100 0 0 0 0 0 0

1000 0 100 -500 500 10 500 10

150 0 10 -100 100 100 50 0

-50 0 0 100 -100 -50 10 0

100 n/a 20