Study of Sequential Accelerated Solvent Extraction of Different Depths of Oak Tank Staves, Affected by Three Different Heat Sources, Analyzed by Gas Chromatography-Mass Spectrometry and Correlations to Sensory Descriptive Analysis of Their Model Wine Extr

A comparison of Soxhlet and a 4 solvent cycle of accelerated solvent extraction (ASE) using dichloromethane as solvent was performed on the same American oak tank staves toasted using three different heat sources, convection, fire and infrared, to demonstrate the efficiency of ASE for the extraction of oak samples. ASE recovered up to 93.5% of deuterated phenol-d5 used as a surrogate and the recovery by Soxhlet extraction was only 78.7%. A 4 solvent cycle was chosen after the comparison of up to 8 independent solvent cycles of the same sample. The comparison of 95% ethanol, dichloromethane and 1:1 dichloromethane/acetone solvents for a recovery test of spiked samples determined that blending solvents can have a diminishing effect due to solvent to solvent interactions. A sequential accelerated solvent extraction method with 2 cycles of dichloromethane followed by 2 cycles of 95% ethanol was suggested for future analysis of oak wood to better represent the targeted compounds formed during oak toasting, since 95% ethanol extracted better vanillin and syringealdehyde and dichloromethane did better with furan compounds and oak lactones.;The gas chromatography mass spectroscopy analysis (GC-MS) of three 2 mm layers of convection, infrared and fire toasted tank staves of French and American oak after sequential accelerated solvent extraction (ASE) with dichloromethane and ethanol, determined a general trend of less differentiation across the layers in convection toasting followed by fire toasting and the greatest gradient of extractives was found in infrared toasted tank staves.;A descriptive analysis of different depths of American and French oak tank staves toasted by infrared, fire and convection in a model wine, showed significant differences for 7 attributes, toasted oak, smoke, vanilla, solvent, ethanol, cardboard aromas, and bitter taste. And the comparison with the chemical results of the GC-MS analysis of the same samples extracted by sequential accelerated solvent extraction, showed that the main difference in the wood were caused by toasted oak and smoke aroma that were positively correlated to furan compounds. The lack of extractives resulted in higher perceptions of solvent aroma. French oak fire toasted and the deeper layers of French oak infrared toasted wood had the highest ratings for solvent aroma and were lower in toasted oak and smoke. American oak samples in general showed more intensity of aromas derived from toasting and lower impact of ethanol and solvent.