식품의 역할 ≫ 기호성 ≫ 관능검사

와인의 향기성분

와인
- 와인의 품질결정요소,  와인 풍미요인
- 와인의 제조공정, 와인의 분류

- 맥주의 향
- 위스키 : 위스키의 향
- 와인 향기성분,  오크통의 토스팅

향기성분의 분류

 

향기성분의 유래

 
Volatile aroma compounds and wine sensory attributes
V. Ferreira, University of Zaragoza, Spain
Abstract: This chapter presents a revision of our knowledge and understanding of the role played by the different aroma chemicals in the positive aroma attributes of wine. In Section 1.1, some basic concepts concerning the characteristics of aroma chemicals, such as thresholds, odour activity values (OAVs) and the relationship between the intensity of odour and the concentration are presented. After this, a systematic approach to classifying the different aroma chemicals of wine is presented. One basic idea is that all wines share a common basic aromatic structure formed by ethanol and 27 different aroma compounds, most of them by-products of fermentation. The mixture of those products has the typical vinous aroma and exerts an aroma-buffering effect with the ability to suppress the effect of many odorants added to it, particularly those with fruity characteristics. The ability of the different odour chemicals to break such a buffer, and hence transmit a different aroma nuance to the wine, and the relationship between the transmitted aroma nuance and the aroma of the chemical are used to define the different roles played by aroma compounds on wine aroma. These roles can be as impact compounds, major contributors, net contributors, subtle aroma compounds, aroma enhancers and aroma depressors. The subjects can be individual aroma chemicals or well-defined mixtures of molecules sharing chemical and odour properties (aroma families). Different examples of the aroma chemistry behind some of the most relevant wine aroma nuances from simple or complex wines are also presented and discussed.

1.1 Introduction: basic properties of aroma chemicals
All of us have experience with odours and with some odour chemicals and, for those involved with wine and other high-quality foods, the description of the odour and flavour attributes of a product is something familiar or even trivial. The words we use to describe the odour attributes of a product come from our personal experience and take the name of other products with which more or less everybody is familiar and whose aroma is clear, explicit and supposedly unambiguous.
However, most of us have received nearly no education about the way in which the chemical aroma molecules work to create an aroma. On the contrary, it is likely that we will have received some contradictory or even biased information about the role of chemicals in aroma. On the one hand, we know that some chemicals have clear and easily recognizable aromas that match nearly perfectly the aroma attributes of a given product (e.g. vanilla and vanillin, or corked wine and trichloroanisole – TCA), which suggests that single aroma molecules can drive the aroma attributes of a product. However, this happens only in a very limited number of cases, and we should be ready to accept the idea that, most often, the aroma nuances are the result of a complex equilibrium between different aroma chemicals. On the other hand, some complex products such as wine, coffee or cacao have in their composition more than 1000 volatile molecules (Maarse and Vischer, 1989), and therefore, one can be tempted to think that the final aroma perception is the result of the interactions of several hundred volatiles, which means that we have almost no chance to understand, model or predict the aroma of these products.
In this case, fortunately, things are not that complicated and, even in those complex products, the numbers of chemicals really contributing to the different aroma nuances are limited (let’s say some tens). It is also true that we are lacking well established tools to deal with the aroma of complex products and that we have to rely mainly on empirical evidence, but since the late 1990s there has been some progress that makes it possible to propose a first approximation of the roles played by the different aroma chemicals of wine in the perception of its different sensory attributes. This chapter deals with this question and will try to present more or less systematically a number of ideas, concepts and facts that will help in understanding the complex relationships between wine odour chemicals and wine odour nuances. Before dealing with those ideas, we will present some basic concepts about aroma chemicals.

1.1.1 Volatile compounds and aroma chemicals
An odour chemical is a chemical compound able to interact with our olfactory epithelium and elicit an odorous sensory response. As our olfactory epithelium is located in the nostrils in the upper interior part of our nose, only volatile molecules can reach this point. So far, all aroma compounds known at present are volatile compounds (Leffingwell, 2002). The opposite is also true, and nearly all volatile molecules (excluding permanent gases, water and a limited number of simple molecules) have some aroma. However, there is an extraordinary variability in the range of concentrations at which a volatile compound is really able to impact the pituitary and elicit the sensory response. While some simple molecules, such as ethane, chlorotrifluoromethane or ethanol, will only be smelled when the concentration in the air is as high as 1 g/m3 (van Gemert, 2003), there are some others, such as bis(2-furfuryl) disulphide or 1-methoxy-3-methyl-3-butanethiol (Guth and Grosch, 1991), which can be perceived at concentrations in the air as little as 0.1 ng/m3 : i.e. differences in sensitivities are up to 10 orders of magnitude. When a volatile compound is at concentrations well below (let’s say 10-fold) its threshold, then its odour contribution can be considered null and the molecule can no longer be, strictly speaking, considered an aroma compound.

1.1.2 Thresholds, odour activity values (OAVs) and I/logC
Thresholds in air are related to the real ability of the odorant to impact the pituitary, but are not related to the potential relevance of the aroma compound in a given product. This is because some odorants are so strongly retained or dissolved in the matrix of the product that they are released to the headspace in contact with the product to a very limited extent. This is the reason why thresholds in water are more useful for predicting the potential importance of a given aroma compound in an aqueous product. In water, some hydrophilic compounds, such as vanillin, are so well dissolved and hence are poorly released to the air that, even if we are very sensitive to them (the threshold in air is very low), a large concentration is required to get a clear odour signal (the threshold in water is very high) (van Gemert, 2003).
Thresholds are simple values and collecting them is relatively easy. This is the reason why, together with the odour description of the chemical, these values are almost the only parameters defining the behaviour of a given odour chemical which are used and tabulated. Many researchers rely then on the quotient between the concentration of a compound in a given product and the threshold of that compound in a matrix similar to that of the product. This parameter is known odour activity value (OAV) or aroma value (Grosch, 1993). However, the threshold, and hence the OAV, is just a first and rough approximation to the potential sensory importance of a given aroma compound in a product. As a first approximation, we can say that being at a concentration above threshold (OAV > 1) is necessary, but it is not enough, particularly in complex products. We can also say that all aroma chemicals present in a product at a concentration one order of magnitude below the threshold (OAV < 0.1) are irrelevant to the aroma of that product.
The relationship between the concentration of an odorant (usually in log scale) and its odour intensity is a very important plot, known as psychophysical curve. Figure 1.1 shows this representation for ethyl 2-methylbutyrate in hydroalcoholic solution. The function is a typical sigmoid curve reaching a saturation point (Chastrette et al., 1998).
As the figure shows, the concentration range can be divided roughly into three areas: the sub-threshold area, the supra-threshold area and the saturation area. The boundaries between these areas are not clearly delineated due to the individual differences in sensitivity; rather, these areas have quite blurred boundaries, particularly in the area around the threshold which is known as the peri-threshold area. Compounds can differ not only in thresholds but also in the dynamic ranges of response, which have a significant effect on the final sensory effect of the compounds. This can be clearly seen in Fig. 1.2, which shows the I/logC plot of four different compounds (Ferreira et al., 2003a). The thresholds of the compounds in the figure are widely distributed, and range from 0.1 ppt to 70 ppm, i.e., more than eight orders of magnitude. However, the plot also shows that the compounds have very different I/logC relationships. While the range between the 0 and the saturation area in ‘normal’ compounds, such as 4-methyl-4-mercaptopentanone or ethyl 2-methylbutyrate, is covered in four to five orders of magnitude of concentration, methyl benzoate reaches saturation just 1.5 orders to magnitude above threshold while β-damascenone requires 11 orders of magnitude to reach saturation. The practical impact of these facts is enormous and shows the serious limitations of the OAV concept for predicting the importance of a given compound in the aroma. For instance, it can be seen that the plots of β-damascenone and 4-methyl-4-mercaptopentanone intersect at a concentration of around 0.1 µg/L. This means that at this concentration both compounds have the same odour intensity; however, the OAV for β-damascenone is more than 1000 (OAV over-estimates its importance) while for 4-methyl-4-mercaptopentanone this parameter is just 20 (OAV under-estimates its importance).
There is an additional reason why OAVs can fail to predict the real relevance of a given aroma molecule in a complex mixture, and this reason is related to the ‘miscibility’ or ‘differentiability’ of the aroma under study in relationship with the aroma of the mixture (Livermore and Laing, 1998). Some odours can be very easily identified in a given mixture and, hence, a low OAV can have great impact (Escudero et al., 2004). On the contrary, some other aromas meld nearly perfectly with the aroma of the mixture so that, even at high OAVs, the aroma is barely perceived.

1.1.3 Distribution of thresholds
A further reason why thresholds and OAVs are merely rough approximations to the importance of a given aroma in a product is that thresholds vary among individuals (Punter, 1983). Usually, the distribution of the thresholds for a single aroma chemical among the population follows a log-normal law and, in the most frequent case, the difference in sensitivity between the least sensitive individuals and the most sensitive is one order of magnitude. Again, there is a relatively large variability in this respect too. In extreme cases, the population is segregated in groups of sensitive and insensitive (or anosmic) individuals and, in this case, differences in sensitivity can be higher than four or five orders of magnitude (Amoore, 1977). This has been found to happen to some wine off-odours, which means that if the quality control in a company was carried out by a single individual and he/she happened to be anosmic to the compound, then the off-flavoured wine could be bottled and dispatched unnoticed (Coulter et al., 2007). This is a good reason to rely on a sensory panel!

1.1.4 Odour, flavour and taste
Aroma compounds can reach the pituitary via the nose during normal olfaction (via orthonasal) or via the pharynx when the food is taken in the mouth and swallowed (via retronasal). In the first case, we smell the odour of the product; in the second case, we perceive the flavour of the product. There is no doubt that the smell of the product is almost entirely due to the signals elicited by the aroma molecules in the olfactory receptors (pure odour), although there are also trigeminal nerve endings in the olfactory mucosa able to interact with some aroma chemicals to produce trigeminal stimulation. In fact, some aroma chemicals produce a sensory perception that is a mixture of odour and trigeminal stimulation, such as acetic acid or menthol (Prescott, 1999a). The flavour is a complex perception that integrates information from three different sensory systems: odour, taste and the chemosensory receptors, responsible for ‘hot’, ‘cool’, ‘dry’, ‘irritant’ or ‘pungent’ attributes (Prescott, 1999b). In the case of aroma molecules, the amount and proportion of molecules reaching the olfactory receptors when the food is taken into the mouth and during swallowing is not exactly the same as that when the product is smelled, because the conditions for the release of the aroma molecules from the product are fairly different in both cases (e.g. temperature, dilution with saliva, aggregation) (Linforth et al., 2002). That complexity can make us think that flavour is totally different, while the truth is that most qualitative attributes of flavour are also produced by aroma molecules via the activation of the olfactory receptors.

1.2 Wine aroma ‘organization’
1.2.1 The base of wine aroma
A normal table wine contains several hundreds of volatile compounds, but most of them are at concentrations well below the threshold, which means that they are not really relevant in the perception of the sensory attributes of the wine. The number of odour molecules really active in a normal wine lies between 20 and 40, and the total number of odour molecules that can be really active in the different kinds (without odour problems) of wines is around 70. What we need now is to find a series of rules to put some order into these relatively high numbers. A first key to such issues consists of realizing that some aroma compounds are present in all wines, independent of their origin and kind. These groups of compounds are of course these volatile aroma compounds produced by fermentation at relatively well-defined proportions. These 27 compounds can be seen in Table 1.1. All these compounds are present at concentrations well above threshold in nearly all wines and they form a particular aroma mixture with the aroma we often define as vinous. It is slightly sweet, pungent, alcoholic and a little bit fruity (Escudero et al., 2004). Another compound, β-damascenone, can also be included in this list because although it is not formed by yeast during fermentation, it can also be found in nearly all wines at concentrations above threshold which, as we saw in Fig. 1.2, is very low.
Of course not all wines have exactly the same composition in this base. The concentration of sugar in the must, the strain of yeast and the degree of anaerobiosis are well-known factors influencing its composition. For instance, this last factor means that whites and rosés are richer in fatty acids and their ethyl esters while they contain less alcohols and isoacids than reds (Ferreira et al., 1996). Another less well-known factor is that the concentrations of fusel alcohols, fusel alcohol acetates, isoacids and their ethyl esters, all of them related to the yeast amino acid metabolism, are related to the varietal origin of the must (Ferreira et al., 2000; Hernandez-Orte et al., 2002). This is the reason why some wines, such as those made with Tempranillo, have a ‘soft’ and delicate aroma because of the low amount of fusel alcohols and isoacids they naturally contain. Finally, ageing is also an important factor and introduces important changes related to the different acid + alcohol = ester equilibria. As the wine concentrations in acetic acid are low, fusel alcohol acetates are hydrolyzed and can be found at appreciable concentrations only in young wines. In contrast, as the wine concentrations in ethanol are relatively high, the concentrations of fatty acid ethyl esters are relatively constant and the concentrations of the ethyl esters of isoacids steadily increase during ageing, which causes a softening in the aroma of some red wines. These compositional differences have some remarkable sensory consequences but yet, it is the combined effect of all these compounds which exerts the most remarkable influence on wine aroma and follows a pattern common to all wines.

1.2.2 The aroma-buffer effect
The mixture formed by the aforementioned compounds forms what we call an aroma buffer. We call it a buffer because in some sense it resembles the buffer systems we usually use to fix pH. Those buffer systems have the ability to counteract the effect of small additions of acid or of alkali and, as such, the aroma buffer has both the ability to counteract the effect of the omission from the mixture of one of its components and the ability to counteract the addition to the mixture of many single odorants. Both effects can be seen in Tables 1.2 and 1.3. Table 1.2 (Ferreira et al., 2002) shows the effects of omission from the mixture of one of the odorants. As can be seen, in most cases the omission from the mixture had no effect, or a just noticeable effect that the judges were not able to define. Only in the cases of isoamyl acetate and β-damascenone were there slight effects on the fruitiness of the mixture. The effect of the addition of different aroma compounds to a neutral wine is presented in Table 1.3 (Escudero et al., 2004). Results are again very surprising. It can be seen that the addition of huge amounts of some odorants has almost no effect, or even that the effect is not the perception in the mixture of the added odorant, but a decrease in some of the basic attributes of the mixture (except for isoamyl acetate). This buffering effect is something challenging for neurophysiologists and has such a strong influence on the way we should understand the hierarchical relationships between wine odorants that it can be used as a useful criterion to classify wine odorants.

1.2.3 Breaking the buffer
Fortunately, the aroma of many wines is very rich in aroma nuances that are quite different to the basic ‘vinous’ aroma of the aroma buffer. This clearly means that some aroma molecules succeed in some wines in breaking the buffer and produce a different sensory perception. As was mentioned earlier, OAVs are not a useful criterion to predict whether one aroma molecule is going to be clearly perceived in wine aroma, particularly in light of the previously explained fact that some odorants at very low OAVs cause large sensory impacts while others at such OAVs can be hardly detected. By observation, we have identified four different ways to break the buffer:
1. A single molecule at a concentration large enough, such as, for instance, isoamyl acetate in banana-smelling wines.
2. A group of molecules with close similarity in chemical and aromatic properties, such as, for instance, aliphatic γ-lactones in some red wines. Some of these associations can be so important, particularly if the group of chemicals is produced along the same chemical or biochemical pathway, that individual compounds have no significance and it is much better to work directly with the group of chemicals. These groups of chemicals are considered families of chemical aroma compounds.
3. A large group of molecules with some similarity in a generic (non-specific) aroma descriptor (for instance sweet), such as for instance linalool, γ-lactones and ethyl cinnamates in some white wines. Those large groups of molecules are in most cases combinations of different families of chemical aroma compounds (Loscos et al., 2007).
4. The association between an aroma enhancer and one or several aromatic molecules unable to break the buffer themselves. In this case, the aroma nuance can be that of the chemical enhanced or a new aromatic nuance mixture of those of the enhancer and the other aroma molecules.

1.2.4 Roles played by aroma chemicals
The exact role played by an aroma chemical in a complex aroma mixture can be assessed empirically by means of reconstitution, addition and omission sensory experiments (Grosch, 1993). In reconstitution experiments, a synthetic aroma mixture containing the aroma compounds present in the wine at the same concentrations is first prepared. Then, new mixtures similar in composition to the first one but in which one aroma chemical or a group of aromas have been omitted or added are further prepared and their sensory properties are assessed. These tedious procedures make it possible to observe the sensory effect linked to the presence of a compound in a chemical environment matching closely that of the wine. How such sensory impact relates to the odour properties of the aroma chemical or group of aroma chemicals under study gives us the potential to define the different roles played by the different aroma chemicals or groups of them. For instance, in some wines in which isoamyl acetate is found at relatively high concentrations, it is possible to detect the typical banana nuance of his compound; i.e., in those wines this compound is transmitting to the wine its specific odour nuances. However, in some other wines in which this aroma compound is present at more modest concentrations, isoamyl acetate just adds a fruity or even a sweet–fruity (unspecific) odour nuance. In these latter cases, it is a generic descriptor (primary or secondary) of the acetate that is transmitted to the wine. These ideas can be easily understood with the help of Figs 1.3 and 1.4. Figure 1.3 shows in schematic form the ability of isoamyl acetate to transmit its odour nuances to the wine as a function of concentration. As can be seen, below 200 µg/L this compound is just one of the many sweet–fruity compounds that wine contains, and removing it from the wine will have a null or very weak sensory effect. We can say that this compound at this concentration is a secondary or subtle contributor to the sweet–fruity note. Between 200 and 1400 µg/L, the importance of the compound in the wine grows to the point that it becomes a quite important contributor to the fruity note of the wine. If this compound is removed from the wine, then the intensity of the fruity note of the wine will decrease, but most surely we will not notice any qualitative change. At this point, this aroma molecule plays the role of net contributor to the fruity note. Between 1400 and 2200 µg/L we come into the transition phase. At this point, the people more sensitive towards this chemical will perceive its banana note in the wine, while the less sensitive will not be able to identify the product. However, if the compound was removed from the wine, then an important sensory change implying quantitative and qualitative changes would be noted. At this point, this compound plays the role of major contributor. Above 2000 µg/L, nearly everybody will perceive the typical banana note of the compound in the aroma of the mixture and, if the sensory impact becomes too dominant, it will even become a defect. At this last concentration, removing the compound from the wine has a dramatic effect on its aroma properties with notable qualitative changes, which means that, at these concentrations, the compound acts as a genuine impact compound. The second example is that of ethyl 2-methylbutyrate. This compound, as the plot in Fig. 1.4 shows, never reaches the concentration in wine at which it is used as an additive by the industry to produce a net impact. This means that, taken individually, this compound can play only the role of subtle or net contributor, but not that of impact compound. However, and as we will see later, this compound may act concurrently with the other members of its family, and the family in its entirety, can then play a role as net or major contributors to the fruit or red fruit aroma nuance of the wine.
A third different example is that of geraniol. Geraniol is a quite powerful aroma compound, but in many instances the sensory effect linked to its presence cannot be directly related to its specific or even generic sensory properties, but to the specific sensory properties of a different (but similar) molecule, such as linalool or cis-rose oxide, i.e., the addition of geraniol to a wine containing linalool will not bring about the increment of an odour nuance related to geraniol, but to linalool. This is a clear case of aroma enhancement, and the role played by geraniol is that of aroma enhancer. A completely different situation is most often observed with some chemicals of relatively bad aroma, although some relatively good-smelling molecules can also follow a similar pattern. A typical example of this behaviour is that of 4-ethyl phenol, responsible for the undesirable sweat, animal, leather odour nuances of some red wines (Chatonnet et al., 1992). The relationship between the sensory effect caused in wine by this chemical and its concentration can be observed in Fig. 1.5. At a concentration high enough (well above 1000 µg/L), the sensory effect of this molecule is the straightforward apparition of the off-odour note but, at smaller concentrations, certainly below the recognition thresholds, what the addition of this molecule to a fruity wine causes is the decrement of the fruity note (Aznar et al., 2003). At those low concentrations, the molecule plays the role of aroma depressor.
All these different roles can be systematized in a list which will make it possible to classify the aroma compounds attending to the role they can play in a given wine. The classification is a new proposal based on observation, but it is based on well established concepts of flavour chemistry (Belitz and Grosch, 1999). 1. Genuine impact compound. This role is played by individual compounds which, in a given wine, are at concentration high enough to transmit to that wine their specific aroma nuances, i.e., the aroma of the compound can be recognized in the wine.

2. Major contributor.
This role is played by individual compounds or by families of aroma compounds that are present in the wine at a concentration high enough to transmit to the wine a primary generic descriptor (red fruit, citric, minty) of its aroma, but not the specific descriptor of the compound (i.e., the compound cannot be clearly recognized in the wine). The transmitted descriptor in the wine is nearly entirely due to the compound so that, if the compound or family of compounds are removed, then the sensory effect will be very intense qualitatively and quantitatively.
3. Net contributors.
This role is played by individual compounds or by families of aroma compounds that are present in the wine at a concentration high enough to transmit to the wine a generic descriptor. Such a descriptor is also contributed by other compounds or families of compounds so that if the compound or family of compounds are removed, then a significant decrease in the intensity of the odour nuance will be noted, but a major change in the qualitative aroma profile will not be observed.
4. Secondary or subtle contributors. This role is played by those individual compounds or families of aroma compounds that are present in the wine at a concentration below that required to transmit individually to the wine one of its generic descriptors. However, such aromatic descriptor (usually very general, such as sweet or fruity) is noted because of the concerted action of many aroma molecules or families. Accordingly, if the compound or family of compounds is removed from the wine, the sensory effect will be very weak or even null.
5. Aroma enhancer. This role is played by those individual aroma molecules or families of aroma compounds which fail to transmit to the wine their specific or generic descriptors but nonetheless enhance the specific aroma of some other molecule or group of molecules present in the wine. In some cases, the enhancement brings about a new aroma quality as a consequence of the mixture of the odours of aroma and enhancer, while in some others the effect of the enhancement is merely the increase of the aroma intensity. In any case, if the enhancer is removed, then a decrease in the intensity of an aroma nuance, not directly related to the aroma of the enhancer, will be noted.
6. Aroma depressor. This role is played by those individual aroma molecules or families of aroma compounds whose presence in the wine causes a decrease in the intensity of an odour note. If the depressor is removed from the wine, then an increase in the intensity of the depressed odour nuance is noted. With these roles defined, it is now possible to present the aroma chemicals and families of aroma chemicals that are able to play some of the most significant roles.

1.3 Wine aroma molecules classified by their role
1.3.1 Impact compounds
According to our own research and from literature data, the following 16 compounds can act as impact compounds of some particular wines. Varietal impact aroma compounds Linalool. This was the first identified aroma component able to exert an impact on Muscat wines (Cordonnier and Bayonove, 1974; Ribéreau-Gayon et al., 1975). Its contribution to the characteristic aroma of several wines made with grape cultivars from Galicia has been clearly demonstrated (Versini et al., 1994; Campo et al., 2005; Vilanova and Sieiro, 2006). Similarly, it also contributes to the flowery or even citrus notes of some other white cultivars (Arrhenius et al., 1996; Lee and Noble, 2003; Campo et al., 2005; Palomo et al., 2006), always in combination with the other terpenols. cis-Rose oxide. This terpene of pleasant flowery character was first identified as a characteristic impact aroma compound of wines made with Gewürztraminer (Guth, 1997). Later, it was also found to be a key odorant in wines made with the varietal Devin (Petka et al., 2006), and was also detected in the hydrolyzed fractions from precursors obtained from different neutral grape varieties (Ibarz et al., 2006). As with linalool, it requires the presence of the other terpenols to be clearly perceived.
β-Damascenone. This compound is found in nearly all wines at concentrations around 1–4 µg/L. At these concentrations, the compound acts mainly as aroma enhancer, promoting the fruity aroma of wine esters (Escudero et al., 2007; Pineau et al., 2007). However, this compound can be present at much higher concentrations in wines made from sun dried grapes (Campo et al., 2008) or overripe grapes (Pons et al., 2008). It is a key odour compound in Pedro Ximénez wines (Campo et al., 2008).
4-Mercapto-4-methylpentan-2-one. This compound has a characteristic scent of box tree (Buxus spp.), which can be perceived in some wines made with Sauvignon blanc (Darriet et al., 1991, 1993, 1995) or Scheurebe (Guth, 1997). At lower concentrations, the compound is not strictly speaking an impact compound, but a major contributor to the fresh fruity notes (Escudero et al., 2004). 3-Mercaptohexan-1-ol. This compound has a smell reminiscent of green mango and box tree with some rubbery notes. Its odour is very complex and changes with concentration and with the aroma environment in which it is found. It was first identified in wines from Sauvignon blanc, Cabernet-Sauvignon and Merlot (Bouchilloux et al., 1998) but afterwards it was found in many others (Tominaga et al., 2000). It is an impact compound of some rosé wines (Murat et al., 2001; Ferreira et al., 2002), of white wines made with Petit Arvine (Fretz et al., 2005) and Sauternes wines (Bailly et al., 2006; Sarrazin et al., 2007; Campo et al., 2008).
3-Mercaptohexyl acetate. This compound was first found in wines from Sauvignon blanc (Tominaga et al., 1996), but it can also be found in many other wine types (Tominaga et al., 2000; Lopez et al., 2003; Culleré et al., 2004; GomezMiguez et al., 2007). It has been recently shown that it is the impact aroma compound of the wines made with the Spanish variety Verdejo, imparting the characteristic tropical fruit aroma nuance to the wine (Campo et al., 2005).
Rotundone. This compound is a sesquiterpene responsible for the spicy notes of Shiraz wines and also of black and white pepper, and was recently reported in Australian wines (Wood et al., 2008). It elutes out of the chromatographic phases very late, which precluded its identification when the aroma composition of pepper was first addressed. Its odour threshold in water was found to be around 25 ng/L, although 25% of tasters were insensitive to this compound. Concentrations in Shiraz grapes are highly variable, ranging from less than 10 to more than 600 ng/L (Wood et al., 2008).

Fermentative impact compounds
Diacetyl. This compound is another odorant with a complex role in wine aroma. It was one of the first identified wine aroma molecules (Fornachon and Lloyd, 1965), and it has often been blamed as the cause of a defect when it is present at high concentrations (Clarke and Bakker, 2004). Its sensory effect is extremely dependent on the type of wine (Martineau et al., 1995a; Bartowsky et al., 2002), and its concentration is also time dependent and related to the concentration of sulphur dioxide in the wine (Nielsen and Richelieu, 1999). Diacetyl is responsible for the buttery note appreciated in some Chardonnay wines (Martineau et al., 1995b; Bartowsky et al., 2002), and its role in the sweet notes of some Port wines has also been suggested (Rogerson et al., 2001). Several authors agree on its ambiguous character (Lonvaud-Funel, 1999; Bartowsky et al., 2004).
Isoamyl acetate. This is the only ester capable of imparting its characteristic aroma nuance to wines, sometimes too overtly. In wines made with Pinotage or Tempranillo varieties it is a characteristic aroma compound (van Wyck et al., 1979; Ferreira et al., 2000).

Age-related impact aroma compounds
(E)-Whiskylactone. This is an impact compound in wines aged in oak wood (Boidron et al., 1988). Above a given concentration it can produce an excessive and unpleasant woody characteristic (Pollnitz et al., 2000).
Sotolon (3-hydroxy-4,5-dimethyl-2(5H)-furanone). This is also an impact compound in wines made with botrytized grapes (Masuda et al., 1984), or wines from biological ageing (Martin et al., 1990, 1992; Moreno et al., 2005), natural sweet wines (Cutzach et al., 1998, 1999), Pedro Ximénez (Campo et al., 2008), Oporto (Ferreira et al., 2003b) or Madeira (Camara et al., 2004). Its concentration, in general, increases with oxidation (Escudero et al., 2000a).
Furfurylthiol (FFT, or 2-furanmethanethiol). This strong coffee-smelling compound is formed by reaction between furfural from the oak cask and sulphydric acid formed during the fermentation (Blanchard et al., 2001), and is able to transmit its aroma to some types of wine. There is not a lot of analytical data on the occurrence of FFT because of difficulties in its determination, but it has been found at relatively high concentrations in aged wines from Champagne (Tominaga et al., 2003a) and in some other wines (Tominaga et al., 2006; Mateo-Vivaracho, 2009).
Benzylmercaptan (or benzenemethanethiol). This is a compound with a powerful toasty aroma and, together with FFT, it can impart smoky and empyreumatic nuances to some aged wines, such as Champagne or Chardonnay sur lie (Tominaga et al., 2003a,b) but also to normal aged dry wines (MateoVivaracho, 2009).
Dimethyl sulphide (DMS). This compound was identified some time ago in aged wines (Marais, 1979) and apparently plays an ambiguous role in wine aroma. Quite often it is related to a defect (sulphury odour) (Park et al., 1994; Ferreira et al., 2003c), but some other authors have demonstrated that it exerts a powerful enhancing effect on the fruity note of some highly appreciated red wines (Segurel et al., 2004; Escudero et al., 2007).
Methional (3-(methylthio)propanal). This compound also plays an ambiguous role. In young white wines it causes unpleasant odours (Escudero et al., 2000b), but in complex wines, such as some Chardonnays or some great red wines, is a net contributor to some appreciated odour nuances (Ferreira et al., 2005).
Phenylacetaldehyde. This is also a compound with an ambiguous role. Its smell of honey is very pleasant but gives to the wine oxidation notes that are considered to be defective and that depress fruitiness (Aznar et al., 2003; Ferreira et al., 2003d). However, this compound can act as impact compound in Sauternes or Pedro Ximénez wines, in which it is found at very high concentrations (Sarrazin et al., 2007; Campo et al., 2008).
All the aforementioned aromas, at lesser concentrations do not play a role of impact compound, but that of major, net or even subtle contributor to some aroma nuance related to one of its more or less general or specific aroma descriptors.

1.3.2 Homogeneous aroma families
A particular case of additive (or eventually synergistic) action is that of all of the groups of compounds which share aromatic characteristics and also share common formation pathways (Jarauta et al., 2006), a case relatively frequent in fermented natural products. In this case, it is possible to define families of odorants. The exact role of these families has been less studied because of the difficulties in defining thresholds and sensory properties for groups of compounds. However, the concept has been latent in the aroma groupings made by some authors (Moyano et al., 2002; Aznar et al., 2003) and has been the subject of some research in our laboratory (Jarauta, 2004; Culleré, 2005; Jarauta et al., 2006; Culleré et al., 2007; Loscos et al., 2007). In this group different families can be identified.
1. Ethyl esters of fatty acids, responsible for fruity notes (apple-like, ester-like) of some white wines (Ferreira et al., 1995).
2. Aliphatic γ-lactones which contribute to the peachy aroma of some reds (Ferreira et al., 2004; Jarauta, 2004) but can also be contributors to the sweet nuances of many other wines (Loscos et al., 2007).
3. Volatile phenols such as guaiacol, eugenol, 2,6-dimethoxyphenol, iso-eugenol and allyl-2,6-dimethoxyphenol, responsible for phenolic and toasted notes of wines (Escudero et al., 2007).
4. Vanillas (vanillin, methyl vanillate, ethyl vanillate and acetovanillone) that can contribute to sweet flowery notes in many wines (Loscos et al., 2007).
5. Burnt-sugar compounds (furaneol, homofuraneol, maltol) that can contribute to the general fruitiness of red wines (Jarauta, 2004; Ferreira et al., 2005).
6. Fusel alcohol acetates that can contribute to the flowery and/or fruity notes of white wines (Campo et al., 2005).
7. Aliphatic aldehydes with 8, 9 and 10 carbon atoms, that can contribute to the citric notes of some wines (Culleré, 2005).
8. Branched aldehydes 2-methylpropanal, 2-methylbutanal and 3-methylbutanal, that contribute to the characteristic odours of aged red wines (Culleré, 2005).
9. Ethyl esters of branched or cyclic fatty acids (ethyl 2-, 3- and 4-methylpentanoates and ethyl cyclohexanoate) (Campo et al., 2006a,b), some of which have been recently identified. The aroma of these compounds could act additively with that of the other wine ethyl esters of branched acids (ethyl isobutyrate, ethyl isovalerate and ethyl 3-methylbutyrate) and contribute to the fruity notes of red wines, as has been recently suggested (Ferreira et al., 2009).
10. Ethyl cinnamate and ethyl dihydrocinnamate that can contribute to the sweet and floral notes of some wines, particularly Chardonnays (Loscos et al., 2007, 2009).
The previous classifications involve 16 potential impact compounds and 31 other aroma compounds grouped in 10 different aroma families. This makes a palette of 26 different odours which, by combination with the base of the aroma, can explain most wine aroma nuances. Not all the combinations have been completely described and studied up to this date, but it is possible to interpret some relevant wine aroma nuances.

1.4 Interpretation of some wine aroma nuances
The key message coming out of this chapter about the aroma of wine is how the basic aroma buffer produced by ethanol and the other major fermentation volatiles is broken. The way in which this happens (through more or less impact compounds, or families, or through large numbers of subtle compounds) determines the complexity and aroma characteristics of wines.
1.4.1 Wines whose aromatic perception is driven mainly by a single
odour chemical In general, wines whose aromatic perception is driven by a single odour chemical are simple and have a clear, simple and distinctive aroma nuance caused by a chemical acting as a genuine impact aroma compound. Its degree of complexity, of course, will depend on the concentration of such a chemical, and on the presence of other aroma compounds which can modify or add some more aroma nuances. The most typical and well-known example of these kinds of wines is Muscat. Other examples are some rosé wines whose aroma characteristics are due to the presence of high concentrations of 3-mercaptohexan-1-ol (Murat et al., 2001; Ferreira et al., 2002). Others are Sauvignon blanc wines, whose aroma characteristics are due mainly to 4-mercapto-4-methylpentan-2-one (Darriet et al., 1995), or wines made from Verdejo, whose aroma characteristics are due to 3-mercaptohexyl acetate (Campo et al., 2005). Other particular examples of this type of wine are some Cabernet Sauvignon or Cabernet Franc red wines made in some parts of New Zealand or France which show a very intense cassis aroma, nearly entirely due to the presence of high concentrations of 3-mercaptohexyl acetate. Of course, if the wines are rich in some other compounds, such as ethyl esters of fatty acids, linalool or isoamyl acetate, the final perception will be more complex, and surely more appreciated. In the case of Sauvignon blanc wines, some producers find value in the presence of methoxypyrazines, which no doubt, adds some complexity (even if this is controversial). Another case of simple wines that, nowadays, are not very appreciated is that of white wines with large amounts of isoamyl acetate, displaying a strong banana aroma. Finally, there are some white wines with a simple fruity aroma which is mainly due to the presence of high concentrations of ethyl esters of fatty acids.

1.4.2 Wines not having any genuine impact aroma compound
Wines that do not have any genuine impact aroma compounds include some very interesting wines showing complex aromas which cannot be attributed to a single chemical identity. In the case of whites made with Maccabeo or Chardonnay, for instance, their aroma nuances are related to the simultaneous presence of many relevant aroma families present at quite modest concentrations. For instance, the flowery notes of some of them can be related to the simultaneous presence of small amounts of linalool, γ-lactones, vanillins, ethyl cinnamates and norisoprenoids (Loscos et al., 2007). Their fruity notes are the result of a complex interaction between those compounds and ethyl esters of fatty acids, fusel alcohol acetates and small amounts of some cysteine-related mercaptans, such as 4-mercapto-4-methylpentan-2-one or 3-mercaptohexyl acetate and eventually also to some aliphatic aldehydes (Escudero et al., 2004; Loscos et al., 2007). In these cases, obviously, the quality vectors of wine are extremely complex and multivariate.

1.4.3 Complex wines containing several potential impact compounds
Typical examples of complex wines containing several potential impact compounds are some Chardonnays fermented in barrel or aged sur lies. In this case, the concentrations of some fermentation compounds are lower, and several powerful odorants appear. These are whiskylactones, of course, but also diacetyl, methional and furfurylthiol. The aroma is still complex, since it retains a large part of the compounds previously cited, but now the typical woody notes, together with the creamy–buttery nuance given by diacetyl and eventually a cauliflower undertone given by methional and a toasty-coffee-like note given by furfurylthiol, can be easily detected.
Other examples of these types of wines are Sherry-like or Sauterne-like wines. In Sherry, acetaldehyde, diacetyl and several isoaldehydes (isobutyraldehyde, isovaleraldehyde, 2-methylbutyraldehyde) act as a family of impact compounds (Culleré et al., 2007), but they also contain sotolon at high concentrations, which gives them their characteristic nutty flavour (Campo et al., 2008). In the case of Sauternes, wines contain relatively high concentrations of 4-mercapto-4-methylpentan-2-one, 3-mercaptohexan-1-ol, phenylacetaldehyde and sotolon (Campo et al., 2008).

1.4.4 Most complex wines: the case of big reds
Red wines are, by nature, much more complex since, among many other factors, they contain quite large amounts of volatile phenols which exert a suppression effect on fruity notes (Atanasova et al., 2004). This phenomenon is still more intense when the wines have been aged in oak casks, increasing the concentrations of volatile phenols and adding whiskylactones. In this chemical environment, the perception of the different notes, particularly fruity notes, is extremely complex. In addition, great red wines do not have explicit or specific odour nuances, but a large palette of many subtle odours. It is not surprising, therefore, that in red wines, leaving aside whiskylactones, most often we do not find genuine impact compounds, but relatively large groups of compounds which contribute to the different odour nuances. Up to this date, we have identified several major contributors to the fruity notes of red wines:
1. The concerted action of ethyl esters, including here several recently discovered branched ethyl esters, with norisoprenoids (β-damascenone and β-ionone) and with the enhancing effect of DMS, that can impart berry fruit notes to the wine (Escudero et al., 2007).
2. The concerted action of five γ-lactones (γ-octa-, nona-, deca-, undeca- and dodecalactones) that can be responsible for the peach notes of some reds, particularly from certain areas of Spain and Portugal (Jarauta et al., 2006).
3. The concerted action of furaneol, homofuraneol, maltol, sotolon, norisoprenoids and methional that can be responsible for some cherry and chocolate notes of some reds (Ferreira et al., 2005). It is very interesting to note the hierarchy of the sensory notes and hence of the aroma chemicals in wine. In a recent study relating the odorant composition, measured by gas chromatography-olfactometry, with wine quality, this parameter was found to be related to three major vectors formed by the summation of groups of odorants (Ferreira et al., 2009), as is shown in Fig. 1.6. The vector with major and more negative influence was formed by the summation of ethyl phenols, TCA and 3,5-dimethyl-2-methoxypyrazine, while most wines were of course not corked (no one was classified as such) nor too rich in ethylphenols. Therefore, this result not only means that a wine containing a high concentration of ethylphenols or TCA is a bad wine, but that the general quality of a commercial red wine is inversely related to the content it has in these compounds, i.e., all these compounds at concentrations far below the recognition threshold are exerting a strong depressing effect on the wine fruitiness as previously suggested (Aznar et al., 2003).

1.5 Conclusions and future trends
The complexity of wine aroma is in accordance with its chemical complexity. As happens in complex perfumes, and far from the artificially flavoured products, wine aroma is the result of complex interactions between many odour chemicals. Only in some particular and simple cases is it possible to find genuine impact compounds able to transmit to the product their primary sensory descriptors. In the most complex and most valuable products, however, the sensory notes are created by the concerted action of many molecules, many of which, surprisingly, are at concentrations near threshold. As the most important aroma compounds are known and as today there are analytical techniques available for their determination, it is only a matter of time before the different aroma compounds of the most valuable products will be unscrambled and fully understood and their vectors of quality defined. This potential is going to open a number of opportunities in the study of the precursors of the different wine aroma nuances. It is expected that in a first step, the chemical aroma quality vectors for different types of wine or for different aroma sensory attributes will be defined, and that further research will make it possible to determine which ones are the precursors and what is the potential of a given grape or of a recently fermented must to produce a wine of the desired quality. Particular attention will be paid to those aroma chemicals or families of aroma chemicals really able to create a sensory impact. Similarly, the discovery that many ‘bad’ aroma chemicals have negative consequences for wine quality at concentrations well below those usually considered risky will promote in the industry the need for more efficient and better directed quality control tools. An exciting world of knowledge with much better wines as recompense is ahead for those daring to deal with aroma chemicals in a systematic and comprehensive way.



오크통의 마법
[중앙일보] 박방주 과학전문기자  입력 2009.03.20 00:37

참나무 그을릴 때 생기는 타닌·락톤 같은 화합물이 술의 맛과 향 좋게 만들어
  
같은 레드 와인이라고 해도 묵직한 맛이 나는 ‘쉬라즈’는 미국산 참나무통(오크통)에서, 부드러운 ‘보르도’는 프랑스산 참나무통에서 주로 숙성시킨다. 미국산 참나무에 와인을 넣어 보관하면 참나무 속 화학성분이 진하게 우러나오지만 프랑스산 참나무는 그렇지 않기 때문이다. 레드 와인이 각양각색이듯이 참나무도 여러 가지다. 와인이 급속히 대중화되면서 와인의 숙성비법도 애호가들 모임에서 화제에 자주 오르곤 한다. ‘나무 박사’인 전남대 김윤수(수목학) 총장과 함께 오크통의 비밀을 과학적으로 풀어본다.

◆100년 넘은 고목이 좋아
대나무에 술을 담아 놓으면 한 달도 못 돼 술이 증발해 버리기 십상이다. 대나무에 수분과 자양분이 오가는 물관이 있는데 이것이 뚫린 채로 남아있기 때문이다. 참나무에도 그런 물관이 남아있다면 와인이나 양주 숙성용으로 애초 쓸 수 없었을 것이다. 다행히 참나무, 그것도 미국산 화이트 참나무는 고목이 될수록 물관이 타일로시스라는 알갱이 같은 세포로 꽉 막혀 버린다. 미국산 붉은 참나무는 그렇지 않다. 한국에도 상수리나무·굴참나무 같은 참나무류가 있지만 물관이 제대로 막히지 않는다. 오크통 재료로 아시아산 참나무가 거의 쓰이지 않는 연유다. 프랑스산 참나무는 미국산보다는 타일로시스가 적어 물관과 같은 방향으로 나무를 제재하지 않으면 와인이 샌다. 고목이 될수록 물관이 막히는 것은 술통으로서의 기본인 ‘술이 새지 않을 것’ ‘술맛을 좋게 할 것’ 중 하나를 충족한다고 할 수 있다. 이 때문에 술통으로 애용하는 참나무는 미국산 화이트 참나무와 프랑스 등 유럽산이다.

◆어떻게 향과 맛을 좋게 하나
오크통이 술이 새지 않는다는 이유만으로 술통으로 사랑받는 건 아니다. 물관이 타일로시스로 막혀 있는 것으로만 치자면 아카시아나무도 화이트 참나무에 결코 뒤지지 않지만 술통으로 쓰진 않는다. 술의 맛을 개선하는 효과가 적기 때문이다. 참나무는 와인이나 양주에 색상과 향기를 더하고 맛을 더 고급스럽게 변화시키는 ‘마력’을 지녔다. 오크통의 결정적 위력이다. 알코올 성분은 색이 없고, 특유의 독한 냄새가 난다. 이를 오크통에 넣어두면 술맛을 부드럽게 하고 매혹적인 색을 띠게 만든다. 알코올 농도 20여 도의 소주보다 40여 도의 양주의 맛이 부드럽게 느껴지는 건 오크통 덕분이다. 참나무에는 떫은맛을 내는 타닌, 쓴 아몬드향의 푸르푸랄, 캐러멜향의 말리톨 등 많은 종류의 화합물질이 들어 있다.

◆완성 직전 불에 그을려
오크통의 막판 완성 단계에서는 꼭 거치는 과정이 있다. 오크통 안쪽을 불에 그을린다. 이를 토스팅이라 한다. 이 과정 또한 숙성하는 와인 맛에 큰 영향을 준다. 코코넛향을 내는 락톤 화합물의 향은 이때 더 진해진다. 탄수화물 분해산물인 당분 역시 토스팅 때 많이 만들어진다. 토스팅은 불의 강도에 따라 약·중·강으로 나눈다. 약한 불에 잠깐 그을리면 과일향이 진하고, 떫은맛이 약간 나는 와인을 만드는 오크통이 된다. 약 15분을 그을리면 내부가 갈색으로 되며, 바닐라향이나 커피향을 내는 와인을 만들어낸다. 불에 태우다시피 강하게 그을리면 볶은 원두거피향, 구운 빵, 구운 고기 맛의 와인을 만드는 오크통으로 변신한다. 강하게 그을린 오크통이 약하게 한 것보다 타닌 성분이 적다. 강하게 그을리면 나무 속의 화합물이 우러나오는 양도 적다. 오크통 속에서는 효모가 참나무 성분을 전혀 다른 성분으로 바꿔 버리기도 한다. 쓴맛을 내는 푸르푸랄이 효모를 통해 훈제 고기 또는 가죽 냄새 등으로 좀 더 부드럽게 바꾼다. 화이트 와인은 거의 오크통에 숙성을 하지 않는다.

1.1. Volatile thiols in wine
Sulfur aromas can be present in wine as result of different causes. They can come from grapes
as non-volatile precursors, from microbial fermentation or from chemical reaction taking place
during storage. The extraction of thiol compounds from wood is another cause of sulpur related
aroma in wines (Landaud et al., 2000).
Some thiol aromas can be generated starting from sulfur-containing amino acids, fermentation
and metabolism products from the sulfur-containing pesticides. Thermal and chemical reaction
of sulfur compounds during winemaking and storage are responsible foe thiol aromas in wines
(Mestres et al., 2000).
Many volatile sulpur compounds, such as carbon sulfide, ethanethiol, methanethiol, and
hydrogen sulfide, which are mainly produced at high levels during alcoholic fermentation are
responsible for olfactory defect (Bartowsky & Pretorius, 2009). Those compounds are
responsible for notes as cabbage, onion, rotten egg, garlic and rubber (Vermeulen et al., 2005).
Among negative sulfur compounds, hydrogen sulfide is characterized by high volatility and
high reactivity. Dimethyl sulfide (DMS) is another sulfur-containing volatile which was found
in wines, with sub-part per billion to sub-part per million levels (Acinobar Beloqui et al., 1996).
S-methylmethionine has been reported as possible precursor (Segurel et al., 2005). These
compounds are responsible for reduced notes in wines.
On the opposite, some long-chain volatile sulphur compounds supply the typical pleasant
aromatic profile of certain wines. In particular, 3-mercaptohexan-1-ol (3-MH),
3mercaptohexylacetate (3-MHA) and 4-mercapto-4-methylpentan-2one (4-MMP) are regarded
as the most important pleasant volatile thiols in wines. 3-MH and 3-MHA are responsible for
passion fruit, grapefruit notes and their perception threshold is 4 ng L-1 and 60 ng L-1
,
respectively for 3-MHA and 3-MH. The 4-MMP aroma is described as box tree, black currant,
or cat urine at high concentration and its perception threshold in wine is 0.8 ng L-1 (Tominaga et
al., 1998a). 3-MH and 3-MHA can impart sweaty aromas at excessive concentrations.
4-mercapto-4-methylpentan-2-ol (4-MMPOH) is reminiscent of citrus zest and grapefruit. The
perception threshold for this compound is 55 ng L-1 in acqueous alcoholic solution (Tominaga
et al., 1998b).
The long-chain sulfur compounds mentioned above characterize the typical varietal aroma of
Sauvignon blanc wine (Tominaga T et al., 1998a). 3-MH and 3-MHA have been identified in
certain red wine varieties, such as Merlot and Cabernet Sauvignon (Bouchilloux et al., 1998).
These volatile thiols, together with 4-MMP, contribute to the aromas of white wine made from
different Vitis vinifera grape varieties, such as Gewurtztraminer, Muscat, Riesling, Sylvaner,
Pinot gris, Pinot blanc, Colombard, Petit Maseng, botrytized Semillon and Grenache
(Tominaga et al., 2000a, Ferreira et al., 2002).
3-MH and 3-MHA are present in wines as two different stereoisomers. The R and S forms of
these compounds are equivalent in the case of 3-MH (Tominaga et al., 2006). The perception
thresholds for the R and S enantiomers are similar for 3-MH (50 and 60 ng L-1, respectively for
S and R forms), but they are responsible for different aromas: the R form is described as
grapefruit, while the S form evokes passion fruit. The ratio between the two enantiomeric forms
is close to 1 in dry white wine; on the other hand, for sweet wines made from botryzized grapes
the proportion of R to S forms is measured as 30:70 (Tominaga et al., 2006).
3-MHA S enantiomer is more abundant than R form. Moreover, the two enantiomers of 3-MHA
show different aromas and perception thresholds (Tominaga et al., 2006). The less odoriferous
R form, which threshold is 9 ng L-1, is characterized by passion fruit descriptor. The S form is
reminiscent of box tree and its perception threshold is much lower (2.5 ng L-1). In dry
Sauvignon blanc and Semillon wines the R to S enantiomeric ratios have been reported to be
30:70, thus the most powerful isomer is even the most abundant (Tominaga et al., 2006).
Benzenemethanethiol, 2-furanmethanethiol and 3-mercaptopropionate represent another group
of volatile thiols responsible for pleasant notes in aged wines (Tominaga et al., 2003a). 2-
furanmethanethiol is a particularly strong-smelling compound reminiscent of roasted coffe and
its perception threshold in a hydroalcoholic solution is extremely low (0.4 ng L-1). This
compound has been identified in sweet white wines made from Petit maseng grape, and in
certain red Bordeaux wines made from Merlot, Cabernet franc and Cabernet sauvignon grape
varieties (Tominaga et al., 2000b). Benzenmethanethiol is a volatile thiol with a strong
empyreumatic aroma reminiscent smoke, identified in boxwood (Buxus sempervirens L.)
(Tominaga & Dubourdieu, 1997) and in both in red and white Vitis vinifera wines (i.e.
Sauvignon blanc, Semillon, Chardonnay) which contain several dozen nanograms per liter,
which represent 100 folds higher than its perception threshold (0.3 ng L-1 in model
hydroalcoholic solution) (Tominaga et al., 2003b). Moreover, both this compound, and 2-
furanmethanethiol and 3-mercaptopropionate, have been identified in aged Champagne wines
(Tominaga et al, 2003a). Nonetheless, 3-methyl-3-mercaptobutanal and 2-methylfuran-3-thiol,
together with 3-mercaptopropyl acetate, 3-MH and 3-mercaptoheptanal, play a key role in
Sautern wine (Bailly et al., 2006. Bailly et al., 2009). In a similar way, wines made from
Botrytis-infected grapes are characterized by the presence of thiol related aromas as 3-
mercaptopentan-1-ol, 3-mercaptoheptan-1-ol and 2-methyl-3-mercaptobutan-1-ol. The first two
have citrus and grapefruit aromas whereas the third compound is reminiscent of raw onion. The
concentration of such aromas in commercial botrytized wines ranges from tens to thousands ng
L
-1. Despite their perception threshold is similar to the measured quantity in wines, their
olfactory impact on the overall aroma of botrytized wines is confirmed (Sarrazin et al., 2007).
As described above, several odoriferous thiols have been identified in Vitis vinifera white and
red wines (figure 1.1). Nonetheless, the most important Sauvignon blanc varietal thiols are 4-
MMP, 3-MHA and 3-MH.