Introduction "Electronic nose" systems have raised a lot of interest since they appeared on the market at the beginning of the nineties. Since then, many searchers and industrials have tested different systems, and the limits of these instruments were revealed. One of the most important limitations is the lack of sensibility of "electronic noses" to some volatile compounds. The use of a preconcentration technique instead of the non-preconcentrated static headspace commonly employed with such

more »

... ectronic nose" systems could cross this drawback. The Purge-and-Trap technique has already been used to improve the selectivity of the sensors. This method was successfully employed as a filter for ethanol [1, 2] , i.e. the ethanol contained in the samples was not adsorbed by the porous polymer material, and therefore, was not delivered to the sensors. Consequently, the sensors were not blinded by the ethanol content, and could response to other components. The same Purge-and-Trap technique was used by Aishima [3] as a preconcentration method for coffee aroma with his laboratory-made instrument based on MOS sensors. Marsili [4] has worked with milk samples using a Solid Phase MicroExtraction (SPME)-MS system as an "electronic nose", and has made a comparison to static and dynamic headspace methods. Abstract. Four Swiss Emmental cheeses from four different factories were analysed with an "electronic nose" based on a mass spectrometer detector, the SMart Nose™. Three sampling methods, i.e. non-preconcentrated static headspace, Purge-and-Trap and solid phase microextraction (SPME), were compared in order to discriminate the cheese ripening ages. The use of a preconcentration technique was found to be helpful for this application due to the possibility to extract volatile compounds with higher molecular masses. From the two systems tested, the SPME was considered from far the best method because of its better repeatability, its simplicity and its compatibility with an autosampler.