11

E. T. Contis et al. (Editors) Food Flavors: Formation, Analysis and Packaging Influences © 1998 Elsevier Science B.V. All rights reserved

111

A simulated mouth to study flavor release from alcoholic beverages S. J. Withers, J. M. Conner, J. R. Piggott and A. Paterson University of Strathclyde, Department of Bioscience and Biotechnology, 204 George Street, Glasgow G l IXW, United Kingdom

Abstract Static headspace techniques have contributed much to our understanding of the interactions of Scotch malt whisky solutions. Such methods, however, do not account for the numerous and changing conditions of the mouth. The Buccal Headspace Technique addresses such effects by sampling air directly from the mouth as the whisky is warmed and mixed with saliva. The complexities of Scotch malt whisky can be more fully imderstood by creating simple model systems in the form of whisky analogues. These analogues, which may not be suitable for human consumption, are analysed using a Simulated Mouth, the conditions of which were set using data from the Buccal Headspace technique. The Simulated Mouth may provide a useful tool in the understanding of Scotch malt whisky flavor.

1. INTRODUCTION

Sensory and chemical techniques have made important contributions to flavor research. For the past twenty years such methods have been used by this laboratory to study the flavor characteristics of Scotch whisky (1). Aided by statistical techniques such as Partial Least Squares Regression analysis (2) an overall impression of the characteristics of Scotch whisky has been established. However, flavor perception is a dynamic system which assesses aroma and taste simultaneously through a complex series of reactions. When food or drink are introduced to the buccal cavity the non-volatiles are detected by the receptors for the four basic taste qualities; these are located throughout the surface of the tongue. As the food is warmed in the mouth and mixed with saliva, the volatiles from the material are released and passed through the retronasal passage where they are detected. Therefore, to obtain an accurate assessment of flavor we need to take account of a number of factors such as the mouth warming of the food towards the physiological temperature of 37°C; the saliva interactions with the food throughout the period of consumption and frictional forces contributed by the tongue and the teeth. A method which accounts for all of these factors is Buccal Headspace Analysis

112

(3). This technique involves the measurement of volatiles in the headspace directly above a food or drink in the buccal cavity. Scotch whisky is chemically a very complex system with many different reactions. It is often simpler to break down the reactions into smaller component parts by creating model systems. However, panellists are unwilling to sample solutions of alcohols and pure chemicals. Therefore we found that it was neccessary to develop a simulated mouth system. Simulated mouths have been constructed by other investigators (4,5and 6) but these systems were more concerned with mastication. We wanted to create a simple system capable of sampling both real and model systems, with the following attributes: an artificial saliva, constant temperature (37°C), agitation and frictional forces working within the artificial buccal cavity.

2. METHOD A series of experiments were conducted to measure the effect of mouth warming on the volatiles of model whisky systems. The model solution consisted of ethyl decanoate dissolved in 23% v/v alcohol. The headspace volatiles of the model solution were compared at 25°C and 37°C. The same comparison was made with the addition of wood extract. To study the effect of temperature increase in a real system we decided to use Buccal Headspace Analysis (3) The apparatus for this technique, which is illustrated in Figure 1, consisted of teflon nosepieces which were inserted into the nostrils of the panellist. The nosepieces were attached via PTFE tubing to a Tenax trap. The air from the buccal cavity was drawn through this apparatus using a pump. The Tenax trap was thermally desorbed using a Purge and Trap Injector Control unit. The desorbed volatiles were analysed by gas chromatography mass spectrometry (GC-MS) with a Finnegan -MAT ITS-40. The apparatus for the Simulated Mouth apparatus is illustrated in Figure 2. It consisted of a glass flask containing 8.4mL of whisky(23% v/v), 3.3mL of artificial saliva and thirty-two glass beads, to contribute a frictional force to the system. The flask was contained in a shaking water bath heated to 37°C. Hydrated air was passed over the headspace of the flask. The headspace of the whisky and saliva mixture was sampled using a Tenax trap and sampled by the GC-MS as in the previous experiment.

3. RESULTS AND DISCUSSION Our initial experiment indicated that the activity coefficient of ethyl decanoate decreased in the model solution at 37°C (Figure 1 ) . So in effect the flavor release of ethyl decanoate from the model solution was reduced upon heating . The effect of wood extract addition to the solution is illustrated in Figure 2.

113

Teflon nose pieces

Figure 1. Apparatus for Buccal Headspace Analysis

Tenax Trap

Hydrated Air /

Shaker Water Bath r a t 37°C Whisky(23%v/v)+Artificial Saliva+ 32 Glass Beads Figure 2. Strathclyde's simulated mouth.

114

Again the activity coefficient of the ethyl decanoate in the headspace was reduced, but to an even greater extent. The release of volatile compounds from alcoholic beverages in the mouth appears to be limited by the formation of ethanol agglomerates. The presence of ethanol agglomerates was suggested from reductions in the activity coefficient of hydrophobic ethyl decanoate. In wood maturations, increasing concentrations of short and medium chain organic acids decreased the critical aggregation concentration of ethanol resulting in decreased activity coefficients from 5 to 40% (v/v) ethanol. On the basis of these results the Buccal Headspace Analysis was carried out. Unfortunately a number of problems were encountered with this methodology: reproducibility, as everyone has a unique breathing and eating pattern. Over a long period of time this technique can be uncomfortable and for reasons of safety, panellists are unable to participate in more than two whisky sessions per day. It was thought that our Simulated Mouth would solve the panel effect we found with Buccal Headspace Analysis. However, reproducibility was again a problem and measurements of air flow and pressure proved unreliable.

6 -r

10

15

20

25

Ethanol concentration (% v/v)

Figure 3. The effect of temperature on the activity coefficient of ethyl decanoate at different ethanol concentrations.

115 6 -r

5.5

+

5 + -4—25 -C -•— + wood ext 25 'C 4.5 + Hi— + wood ext 37 °C

10

15

20

25

30

35

40

Ethanol concentration (% v/v) Figure 4. The effect of changing ethanol concentration on the activity coefficient of ethyl decanoate in different model systems.

4. CONCLUSION

Our initial experiments showed a decrease in the flavor release of ethyl decanoate in an alcohol and water solution at 37°C, and a further decrease with the addition of wood extract. By using Buccal Headspace Analysis and our Simulated Mouth system we hoped to examine these effects in greater detail. However, our trapping and sampling method proved to be unreliable and for the moment our findings remain inconclusive.

116 Acknowledgements:

The UK Biotechnology and Biological Sciences Research Council (BBSRC) and The Chivas and Glenlivet Group provided financial support and technical assistance for this work.

References: 1 S.J. Withers, J.R. Piggott, J.M. Conner and A. Paterson, Journal of the Institute of Brewing, 1995, Vol 101, pp359-364. 2 M. Martens and H. Martens. In: Statistical Procedures in Food Research (J. R. Piggott, ed.), Elsevier Applied Science, London, 1989, p293. 3 C. M. Delahunty, J. R. Piggott, J. M. Conner and A. Paterson, Journal of the Food and Agriculture, 1996, Vol 71, No 3 pp273-281.

Science of

4 W.E. Lee, Journal of Food Science, 1986, Vol 51, No 1 pp249-250. 5 S.M. Van Ruth, J.P. Roozen and J.L. Cozijnsen, Chemical Senses, 1995, Vol 20 Nol ppl46-149. 6 D.D. Roberts and T.E. Acree, Chemical Senses, 1995, Vol. 20, No.6, pp246-249