Project 3.1.1

Identification and control of volatile compounds responsible for important sensory attributes

Project summary

The flavour of wine is determined to a large extent by volatile compounds that are perceived during consumption by the sense of smell. The overall flavour of a wine, as well as specific flavour notes, is provided by dozens of naturally occurring chemical compounds of widely varying potency and concentration.

Within this project, the formation of, and factors which influence aroma compounds recently found to be responsible for ‘stone fruit’ (apricot-peach), which is common in Chardonnay and other white varieties, will be investigated, with fermentation and vineyard studies and analyses of commercially produced wines.

The role of the capsicum-like methoxypyrazine compounds in Shiraz wines, previously considered not relevant to green flavour in this variety, will be explored through whole bunch fermentation studies, analysis of stalks and leaves, and determination of the genetic basis of the expression and regulation of this metabolite in grape bunches through collaborative work with CSIRO.

Jammy/raisin flavour will be studied in collaboration with Project 4.1.1 (Managing the impact of vintage advancement and compression). The identity of the compound(s) that give overripe jammy/raisin flavour of Shiraz, and more generally the effect of bunch exposure, have not been adequately established. The lack of knowledge of the compound(s) responsible for this flavour is a major gap in setting a measurable target in viticultural projects, especially given increases in growing season temperatures and vintage compression. The effect of bunch exposure on this and other flavour compounds, including TDN which contributes bottle-aged character to Riesling wines, will also be examined in this project.

The role of thiols and other sulfur compounds in red varieties will be determined. Foliar nitrogen and sulfur vineyard sprays have been shown to have the potential to positively affect thiol concentrations in wine, and will be investigated.

Working with the rotundone mapping Project 4.4.4, the compound causing ‘musk’ and (non-pepper) spice in Shiraz will be investigated. In addition, investigations of Shiraz wines sourced from the rotundone (4.4.4) and terroir projects (3.3.1, 4.4.1) will allow determination of new or less understood volatiles that are key to premium wine flavour.

The effect of blending alternative grape varieties with established varieties produced in the Riverina, Riverland and Murray Valley regions will be investigated, to provide enhanced flavour characteristics, acid, colour, and weight to these types of wines. The project will also have a component whereby off-flavours and taints will be identified and studied.

Current practice in sensory evaluation of research wines uses sensory descriptive analysis, and while considered the most powerful and sophisticated method available, requires several weeks of data generation and substantial time for data analysis. In recent years, alternative faster methods have been developed in food science applications. This project will evaluate these protocols to determine their utility in wine studies and for wider industry use. Advantages of using untrained consumers compared to trained panellists will also be assessed. The project will also investigate technology for simulating wine experiences in lifelike environments, to better capture consumer responses.

Latest information


Methoxypyrazines in Shiraz and Pinot Noir: the role of whole bunch fermentations
The inclusion of stalks in a Shiraz fermentation was previously shown to cause an elevated concentration of methoxypyrazines, compounds that give a ‘green’ herbaceous character to wines. Methoxypyrazines were not previously thought to be important to Shiraz wines. A winemaking experiment was conducted with Adelaide Hills Shiraz and Pinot Noir grapes from the 2018 vintage. Wines were made with an increasing proportion of whole bunches, from 100% crushed and destemmed fruit (0% whole bunches) to 25%, 50%, 75% and 100% whole bunches. The results were very clear for both varieties, with a linear increase in ‘capsicum’ aroma and flavour with an increasing proportion of whole bunches, and an equally strong linear increase in isobutyl methoxypyrazine concentration (Figure 7). These results provide confirmation that inclusion of grape stalks in wine fermentations can change flavour towards ‘green’ attributes.

Effects of whole bunch fermentation on Pinot Noir and Shiraz phenolics
As part of the same study, researchers in the AWRI’s texture team examined the effects of whole bunches on wine phenolics. A key question was whether tannins extracted from the stems would alter colour and astringency of the wines.
For Shiraz, it was found that a linear increase in tannin concentration occurred with the inclusion of whole bunches (600 to 1,400 mg/L) and this correlated strongly with the astringency of the wines. The type of tannin also changed with whole bunch inclusion, with tannins from stems found to be larger (polymer length) and more galloylated than those extracted from the skins and seeds of the grapes. Although the study could not distinguish different effects in terms of tannin concentration and structure on astringency, it will be of interest to further compare the mouth-feel properties of tannins derived from the different bunch components. Whole bunch inclusion in Shiraz also slightly increased wine colour (assessed both in the laboratory and visually by a sensory panel), but this did not change as whole bunch inclusion was raised over 25%.
In the Pinot Noir trial, very different results were seen. Tannin concentration was low (200 mg/L) in the control wine and was decreased with 25% whole bunch inclusion. As whole bunch addition was raised between 50% and 100%, tannin concentration varied, but did not exceed that of the control wine (200 mg/L). Wine colour density measured in the laboratory was reduced by all whole bunch addition treatments for Pinot Noir, but interestingly it was found that for the 75% and 100% whole bunch treatments, red colour intensity determined by a sensory panel was remarkably increased (Figure 8).

The drop in visual brown colour intensity rated by the sensory panel was found to be linear as whole bunch inclusion rate increased for Pinot Noir (Figure 9). Looking more closely at the analytical results for wine colour, it was surprising that the hue (i.e. the ratio of red and brown coloured compounds) of the wines remained constant. This was because both red and brown compounds in decreased with whole bunch inclusion. The results highlight that at low concentrations of coloured phenolics, such as in Pinot Noir wines, losses of brown material may result in enhancement of perceived red colour.

Further study of the relative ripeness of the stalk material would be of interest, as well the effect of whole berries, but this work provides winemakers with knowledge to consider the trade-off between improved colour and higher tannin against the possibility of enhanced ‘green’ flavour.

Understanding the role of varietal thiols in red varieties
Thiol compounds such as 3-mercaptohexanol (3-MH) and 3-mercaptophexyl acetate (3-MHA) are important flavour compounds in Sauvignon Blanc and other white varieties, where they contribute ‘tropical fruit’, ‘passionfruit’, ‘grapefruit’ and ‘box hedge’ characters. Much less is known about the sensory significance of thiol compounds in red wines. An aroma-only sensory evaluation was conducted to assess the sensory effects of the varietal thiols 3-MH and 3-MHA in red wine. The compounds were added at the highest concentrations found in a recent red wine compositional survey and in a Grenache winemaking project that assessed a wide range of yeasts. More ‘fruity’, ‘green’ and sometimes ‘tropical’ characters were described in the Cabernet Sauvignon, Pinot Noir and Grenache wines spiked with 3MH and 3MHA compared to the unspiked wines, but the effect was less obvious in the Shiraz wine. Further sensory studies are planned.

Foliar nutrient sprays and changes in varietal thiols

It is known that nitrogen fertilisation in the vineyard can influence thiol concentration in wines. The effect of foliar application of nitrogen and sulfur on Shiraz and Chardonnay grapevines in the Barossa Valley was assessed over two seasons, with two doses applied. The sprays resulted in increased concentration of amino acids and yeast assimilable nitrogen.

Sensory analysis of the 2018 Chardonnay and Shiraz wines showed only a small increase in ‘grapefruit’/’tropical fruit’ descriptors with the lower dose foliar spraying rate, but a surprisingly large sensory effect was found for the wines made with fruit from vines sprayed at the higher dose, for both varieties. The 3-MH concentration was strongly enhanced for both wine sets, with 3-MHA also increased in the Chardonnay foliar spray treatments. Importantly, low molecular weight sulfur compounds such as methanethiol (usually associated with off-flavours) showed no increase with the treatment. The results provide a straightforward means for producers to enhance the thiol-related ‘fruity’ characters in wines, including in warmer regions.

Riesling flavour: TDN and aged character

Riesling flavour characteristics, especially after some time in bottle, can be affected by the presence of TDN (1,1,6-trimethyl-1,2-dihydronaphthalene), sometimes described as ‘kerosene-like’. If TDN is present early in a Riesling wine’s life it can be considered detrimental. A collaboration with Hochschule Geisenheim University, Germany continued during the year, with the aim of assessing ways to reduce the propensity of Riesling wines to develop TDN and undesirable ‘kerosene’ characters. Aspects such as light quality and quantity at the bunch zone, temperature, row orientation and width have been investigated in vineyards in Germany and in the Eden and Barossa Valleys over several seasons. Accelerated ageing of wines from two vintages from these studies has been performed to predict TDN formation and sensory properties, as well as changes to other key Riesling compounds such as monoterpenes.

In vintage 2019 different shade cloth materials were used to manipulate bunch light interception and there was a clear decrease in total TDN (free TDN plus TDN from precursor forms) in juice at harvest as a result of the shade cloth treatments, with the decrease related to the amount of light transmitted to the bunches. Using HPLC-MS analysis it was demonstrated that the effects of light exposure on carotenoid precursors to TDN and other flavour compounds were quite pronounced. From the 2018 and 2019 vintage experiments, black and green shade cloth were found to reduce overall carotenoid degradation, while red shade cloth strongly affected the carotenoid profile. Work is underway to identify the specific carotenoids that are the precursors to TDN. Results from this work were presented at the Oeno 2019/In Vino Analytica Scientia conference in Bordeaux in June 2019 and Yevgeniya Grebneva was awarded the inaugural Denis Dubordieu award for outstanding presentation by a young scientist.

‘Stone fruit’ flavour in white wine
Continuing earlier work on ‘apricot’ flavour, studies were conducted to investigate the influence of grapevine clone and harvest time on the concentration of monoterpenes in Viognier wine, with wines made from fruit from two regions. In addition, a new analytical method was developed combining the quantification of flavour-active lactones into one straightforward analytical procedure. Red wine samples from 10 varieties were analysed to assess the contribution of these compounds to red wine flavour.

To determine the flavour compounds responsible for ‘peach’ flavour, particularly in Chardonnay wines, a sensory-directed analytical study of wines with high and low ‘peach’ character was completed. A ‘peach’ reconstitution sensory study was conducted, to find which group of compounds was most important. The results clearly showed the influence of several fermentation-derived esters on ‘peach’ aroma. This study provides a major advance in knowledge on this sensory attribute, which should allow winemakers to adjust the level of this character through simple fermentation practices such as oxygen management, decreased juice clarification, yeast selection and lees contact.

‘Raisin’/’jammy’ flavour in ripe Shiraz
The volatile compounds that cause ‘raisin’ or ‘cooked fruit’ aroma, especially in late-picked Shiraz, are not well understood. A vintage experiment was completed to produce ripe and overripe grape samples and wines which will be studied through sensory and chemical analyses. GC-olfactometry analysis on grape berries from the previous vintage (with individuals acting as detectors to smell compounds as they are emitted from a gas chromatograph instrument) showed that several compounds are implicated and work is underway to quantify the relevant aroma compounds. Model chemical reaction experiments with sugars and amino acids provided evidence of the importance of pH and other conditions in the formation of several of these compounds, and this will be studied further.

Improved understanding of the impact of region and winemaking techniques on Pinot Noir wine
Commercial Pinot Noir wines were selected from five regions around Australia and five regions from around the world. Wine colour, tannin and volatile compounds were analysed and collaborators at the University of Tasmania will compare these results to sensory information and winemaking details to assess differences in chemical composition associated with wine regionality.

Sparkling wine autolysis trials
As part of a collaborative project with the University of Tasmania, Chardonnay and Pinot Noir sparkling wines made with different treatments to induce ‘autolytic’ or ‘aged’ character were analysed. Treatments included ultrasound, microwave, enzyme addition, storage at 15°C or 25°C and spiking with aged wine or lees. Comprehensive analysis of volatiles and non-volatiles were completed. As a final step, results will be compared to sensory data.

Assessment of rapid, alternative sensory methods
The Pivot© Profile rapid sensory evaluation method was applied using several groups of assessors with several variables investigated. A set of wines was assessed by a relatively large group of USA wine experts as part of Wine Australia’s Australian wine showcase event held in Lake Tahoe, California, and broadly the same set of wines was tasted by a group of Masters of Wine at an event in Adelaide. The results have confirmed the usefulness of this method as a rapid technique for wine sensory assessment, and a computerised data acquisition system has been produced to allow simple data collection and analysis.

The collection of sensory information from consumers is fraught with challenges, with difficulties sometimes experienced in obtaining reliable data without biasing or misinterpreting results. A relatively new method which involves consumers tasting a set of samples and positioning them on a grid according to preference or choice has been developed by overseas researchers. This method was applied at the AWRI for the first time in red wine evaluation, with 56 consumers assessing a set of wines made from ‘alternative’ varieties, both fairly well-known varieties and more obscure ones. The wines were assessed under three conditions: knowing the label, vintage and variety, but with no tasting; blind tasting with no information; and informed tasting. A Shiraz wine was included as a benchmark. Varieties included Tempranillo, Lagrein, Saperavi, Sangiovese, Nero d’Avola, Graciano and Montepulciano.

The results showed that the method provides rich insight into consumer attitudes and preferences, including appropriate occasions and attitudes. The method was simple for consumers to apply, and appeared to be more discriminating than conventional hedonic scoring. The Lagrein wine included in the study showed potential as it was well liked both blind and with knowledge of variety, with the results indicating this variety, as well as Saperavi, was of interest to the consumers, and suited for consumption at restaurants and with friends. The Nero d’Avola was one of the least preferred wines under each testing condition. The results for informed and label-only conditions were quite similar for most wines, with the exception of Graciano (where taste helped consumers to accept the wine), and Sangiovese (where label knowledge increased acceptance of the wine). It should be noted that only one example was used of each variety, and these results should not be generalised.