Whole grapes have been commonly used to measure the two main smoke taint compounds, guaiacol and 4-methylguaiacol. Whole berries (skin, pulp and seeds) were ground up and measurement presented on a unit weight basis. Where grapes were sent from Phylloxera Infested Zones the grapes were frozen for quarantine purposes and this procedure could alter the extraction of compounds from the berries compared to non-frozen samples, as has been found when measuring organic acids and potassium.
Free-run juice may include a small contribution of smoke taint compounds from the skin, as the free-run juice produces relatively low concentrations compared to whole berries. The concentrations in free-run juice are potentially influenced by the degree of maceration of the skins, pressing fraction and the period of time skins spend in contact with the juice. In 2003, the concentration of smoke taint compounds in juice from macerated grapes was shown to increase in a linear fashion up to about four days on contact with skins.
The concentration of smoke taint compounds observed in wine will be influenced by many factors such as processing temperature, press fraction, period on skins, and fining agents used. These aspects need to be considered when comparing wine results.
Solubility
No previous research has investigated the solubility of smoke taint compounds in juice or wine. In related work, the solubility of smoke related compounds was greater in a model wine solution than in water (Barrera- Garcia et al. 2006). Barrera-Garcia et al. (2003) attribute this to the effect of the alcohol increasing the solubility. This difference may have some impact when comparing measurements between grapes and juice (water solution) and wine (alcohol and water solution) given that the high solubility of many aroma compounds in alcohol reduces their volatility.
Of the compounds of interest in smoke affected grapes, guaiacol has the highest solubility (17.9 g/L in model wine at 25ºC), then, in decreasing order, 4-methylguaiacol (10.4 g/L), 4-ethylphenol (5.9 g/L), 4-ethylguaiacol (4.2 g/L) and eugenol (2.1 g/L) (Barrera-Garcia et al. 2006). Thus, the measurement of guaiacol and 4-methylguaiacol captures the two compounds that are most likely to be absorbed into grapes and wine.
Measurement
The extraction method used by the main provider of analytical services, the Australian Wine Research Institute (AWRI), only collects free guaiacol and 4- methylguaiacol. There may well be further amounts of the compounds of interest bound to glycosides that may be released as free conjugates over time or under particular conditions.
Pollnitz et al. (2004) examined the extraction and measurement of guaiacol, 4-methylguaiacol and other volatile oak compounds and concluded that artefacts in measurement may occur resulting in the observation of up to 10 times the level of guaiacol. These artefacts were particularly obvious with diethyl ether extraction and gas chromatography injector block temperatures above 225°C. Data from previously published determinations for guaiacol should be examined carefully with respect to sample preparation and analysis.
Testing of the same 16 wine samples was conducted at two laboratories, the AWRI and a commercial independent laboratory in Victoria. In the range between 10 and 550 µg/L, the results showed a close relationship between the data from both laboratories (r2 = 0.997) but it was not a 1:1 relationship.
The ratio of the commercial laboratory readings to those of the AWRI was 1.47:1. With 4-methylguaiacol, the relationship was also very close (r2 = 0.995) but again the ratio was 1.4:1 with the Victorian independent laboratory showing relatively higher readings. Similar techniques were used, i.e. liquid-liquid extraction with pentane and pentane:diethyl ether (2:1), and measurement by gas chromatography. Both techniques take into account potential issues with inaccurate guaiacol extraction and measurement identified by Pollnitz et al. (2004).
Further investigation is required to determine the reasons for the different results observed between the two laboratories and to validate whether the relationship holds with samples of less than 10 µg/L. Grape samples should also be compared between the laboratories particularly at the lower end of the concentration spectrum, i.e. less than 10 µg/kg guaiacol, since this is the critical range in which grape samples may be rejected by wineries.
Detection limits
Most of the analyses reported in this report were conducted by the AWRI. The level of detection in wine was 1.0 µg/L with an uncertainty of ±1.0 µg/L or ±10% (whichever was greater). The commercial laboratory in Victoria also had a minimum detection limit of 1.0 µg/L with an uncertainty of ±1.0 µg/L or ±10% (whichever was greater).
