- Research Notes -
Monitoring Selected Variables in Oak-aged Wines by Gas Chromatography - A Progress Report
by
Barry Gump, Steven Glossner, K.C. Fugelsang
and C.J. Muller
CATI Publication #960501
© copyright May 1996, all rights reserved
INTRODUCTION
Barrels have been used in the winemaking process for more than two thousand years. In this process, it was discovered that specific types of wood could impart positive flavor attributes to wines stored therein. White oak, whether from America or Europe, has been found to have the desirable combination of 1) structural integrity, 2) ease of coopering, and 3) beneficial sensory properties; thus it has become the most popular wood for barrel construction.
Compounds found in barrels can be classified by their bio-genetic or chemical origins. These include 1) products of lignin degradation (examples include vanillin, that has the aroma of vanilla; and eugenol, that smells like cloves); 2) compounds from the pyrolysis of cellulose and hemicellulose (examples include furfural and hydroxymethylfurfural. These compounds impart a caramel or butterscotch character to the wine); 3) various terpene constituents; and 4) components derived from hydrolyzable tannins (examples include ellagic and gallic acids. These compounds may impart bitterness and 20 astringency; however, their sensory impact at this time is only widely assumed, and further research is needed before their actual roles are fully understood).
In designing experiments to monitor some facet of the coopering process (such as the effect of grain tightness in air-dried oakwood barrel staves) one must be concerned with how to measure results. The traditional method of conducting sensory analyses with proper controls and training of panelists is costly in time and effort. There is an obvious desire to see if objective, "chemical" measurement(s) might provide the same answers. For this reason, researchers have measured several parameters such as total extracted solids, total phenols, and, lately, some specific components. The techniques of liquid and gas chromatography have allowed quantitative analyses of a number of compounds possessing flavor characteristics.
EXPERIMENTAL DESIGN
The wine utilized in these experiments was a commercially made Chardonnay from grapes harvested in the 1991 season. Each barrel type contained the same blend of grapes, and the wine was barrel-fermented and aged sur-lie. The wine was clarified and bottled after approximately nine months of barrel aging. The barrels used for this study were made from American Oak from the Ozarks; the control barrel was a three-year old French Oak barrel.
CHROMATOGRAPHIC PROCEDURES
A micro-extraction procedure utilizing Freon 113 was used for separation and concentration of volatile compounds in the wines. Fourteen compounds, known to exist in oak-aged wines, were used as gas chromatographic standards (see Table 1). Chromatography was performed on a 30-meter fused silica, 0.32 mm-diameter column coated with OV-1, on an instrument operated in splitless mode with flame ionization detection.
TABLE 1. Compounds used as GC standards (in order of retention time).1. Furfural 6. Ethyl Maltol 11.Vanillin 2. Furfuryl Alcohol 7. Hydroxymethyl furfural 12. b-Ionone 3. Phenol 8. cis-Oak Lactone 13. Syringaldehyde 4. Guaiacol 9. trans-Oak Lactone 14. Coniferaldehyde 5. Maltol 10. Eugenol
CHROMATOGRAPHIC RESULTS
Figure 1 shows a chromatogram from a wine sample spiked with the standards listed in Table 1. Figure 2 shows a series of overlaid chromatograms that include: A) extract of wine standards; B) extract of wine from the control barrel; C),D), and E) extracts of wine in contact with barrels made from 6-, 12-, and 18-month air-dried wood, respectively. The chromatograms are obviously strikingly similar, and thus it is difficult to make definite statements about the presence of some of the standards; however, small amounts of phenol, maltol, ethyl maltol, oak lactones, eugenol and vanilla are indeed discernible. There are no major distinguishing characteristics among the three air-dried oak samples, or between them and the control barrel sample. However, from the presence of compounds associated with "smoke," "toast," "clove/spice," "coconut" and "vanilla" notes, one can surmise that those notes affect the sensory characteristics of the wines. Yet, it would be difficult to appropriately describe sensory differences among the wines directly from the chromatograms.
Figure 3 shows a similar overlay of chromatograms as follows: A) extract of wine standards, and B), C), and D) extracts from wine held in barrels whose staves had been subjected to 25-, 45-, and 60-minute toasting time in order to reach "medium- toast." While the results are similar to those depicted above, there are gradual increases in some peaks areas as a function of toasting time. Peak areas for compounds associated with "smoke," "toast," "clove/spice," and "vanilla" have gradually increased. The implication here being that the longer toasting time, 60 minutes, would have the greater intensity of these sensory attributes.
Two notes of caution are worth pointing out however: (1) It is rather difficult in practice to uniformly extract and concentrate the extracts prior to GC analysis; often larger peaks simply indicate more concentrated samples. (2) What is being observed is the response of the flame ionization GC detector. As mentioned previously, this detector does respond to compounds as one's palate does. It is conceivable then that we might have missed some sensory significant compounds, and that, conversely, compounds shown might be beneath the human sensory threshold.
CONCLUSIONS
As it is evident from the gas chromatographic results, peaks corresponding to "impact" flavor compounds commonly associated with "smoke," "toast," "cloves/spice," and "vanilla" are indeed present in wines exposed to oak that had undergone several different treatments. However, there seems to be no discernible differences in the relative amounts of these compounds among air- dry treatment of staves. Length of barrel toasting time, on the other hand, shows a definite pattern of increasing concentration of these compounds with increasing toasting time. It can be surmised that such an increase might translate into increased sensory perception of these flavor characteristics.
In conclusion, while current laboratory methodology is capable of detecting and quantifying "impact" flavor compounds, correlation between analytical methodology and sensory perception should be carried out to properly evaluate significant changes in oak-wood treatments.
AUTHORS' NOTE
This publication contains preliminary results and has not undergone peer review.
ACKNOWLEDGEMENTS
We would like to thank the California Agricultural Institute (CATI) and Independent Stave, Lebanon, MO (U.S.A.) for their support.
{ page top }
{ CATI , VERC , VERC - Research Publication }
Copyright © 1996. All rights reserved.
CALIFORNIA AGRICULTURAL TECHNOLOGY INSTITUTE - CATI
College of Agricultural Sciences and Technology
California State University, Fresno
