Gas sensors arrays ('electronic noses'): a study about the speed/accuracy ratio

New Fingerprint Mass Spectra (FMS) systems claim for a higher precision, a reduced analysis time, and a lower drift, compared with Gas Sensors Arrays 'Electronic Noses'. We demonstrated in this study that metal oxide gas sensors may have highly improved performances fur both accuracy and run time, when used in optimum conditions. The different noise components were quantified referring to popular words (equivalent of ethanol concentration) using the transfer function of the sensor. It was pointed out that sample surrounding greatly affects the noise level, and that the signal-to-noise ratio may be dramatically increased, reducing the distortion generally assessed between sensor and system detection limits. The standard deviation (std) for 100 ppm of ethanol in water was < 4 ppm, the quantification noise < I ppm; and the short-term sensor drift < 2.5 ppm. It was also shown that the cycle time may be reduced to less than 10 min, keeping the high accuracy over a large concentration range (10 to 10,000 ppm), The sample throughput could not be improved more because of the headspace equilibration, which cannot be reduced. These performances are truly similar to those measured by using Mass Spectra systems. (C) 2000 Elsevier Science S.A. All rights reserved.