RT Journal Article SR Electronic T1 Discrimination and Control of Toasting Intensity and Quality of Oak Wood Barrels JF American Journal of Enology and Viticulture JO Am J Enol Vitic. FD American Society for Enology and Viticulture SP 479 OP 494 DO 10.5344/ajev.1999.50.4.479 VO 50 IS 4 A1 Pascal Chatonnet YR 1999 UL http://www.ajevonline.org/content/50/4/479.abstract AB The volatile compounds and odors formed by toasting oak wood can greatly influence the composition and quality of wines and spirits matured in casks. The choice and mastering of the intensity of toasting can have a considerable influence on the quality of the products produced in the wood. Therefore, it is desirable to understand the barrel toasting process in the cooperage and to better test it. For a given cooperage, it is possible to differentiate barrel toasting levels by analyzing a certain number of volatile and semi-volatile compounds resulting from the thermal degradation of oak. We show that headspace analysis of toasted wood using a micro-extractive method with a stationary polymethyl-siloxane-type phase (Head Space Solid Phase Micro Extraction, HSSPME) is a promising technique for analyzing a sample taken from a toasted barrel. It is easier to use than the traditional maceration and extraction method. This method requires neither preparation nor preliminary extraction of toasted wood samples. It can, however, only be carried out in laboratories experienced in chromatographic analysis. The use of an array of metal oxide based odor sensors (MOS) has been also studied on laboratory and industrial scales. It is possible to differentiate between the intensities of toasting of barrels by analyzing the volatile compounds released into the gaseous phase using an ad hoc combination of electronic sensors ("an electronic nose"). Prior selection on a laboratory scale showed that six sensors should provide sufficient discrimination of the classic various toasting intensities. The toasting of the barrels can be tested immediately after pressing as long as they are rapidly sealed by a suitable method. Under these conditions similar profiles can be measured for at least 15 minutes. The injection of the still hot headspace of the barrels takes several tens of seconds. With the system designed it is possible to test, on average, 160 casks one after another with the apparatus working six hours per day. This system can be easily improved and completely automated. We have produced a reference database to allow differentiation between different toasting intensities in casks toasted in the normal way by coopers. This database is valid for a given type of wood in a given workshop. Generally speaking, the coopers find it difficult to produce regular medium toasting; light, medium+ or high toasting form much more uniform groups. At this stage of our research, the technical feasibility of our system to control the toasting level of oak wood barrels using an array of MOS has been proved.