Measuring white wine colour without opening the bottle

Measuring white wine colour without opening the bottle Excessive brown colour development is undesirable in white wines and generally indicates that the wine is oxidised. The commonly accepted industry method for assessing oxidised colour in white wines is to measure wine absorbance at 420 nm (A 420 ) (Singleton et al. 1976, AWRI publication #283, Iland et al. 1988, Zoecklein et al. 1995). In fact, A 420 is really a measurement of yellow colour. In red wines, A 420 coupled with A 520 can give an indication of hue and colour density (AWRI publication #151). In 70 previous studies it has been shown that for a Chardonnay wine, an increase in the A 420 value was strongly correlated with decreasing concentrations of total sulfur dioxide (AWRI publication #528) and for a Semillon wine, an increase in the A 420 value was strongly correlated with decreasing concentrations of both total and free sulfur dioxide levels (AWRI publication #666). Thus, A 420 can be considered as a marker of white wine oxidation. It is important to note that the A 420 values of two non-identical wines should not be compared because each wine displays unique spectral characteristics with differing A 420 values, and overall wine colour is dependent on the combination of all absorbances in the visible region. The spectrum of light absorbance by a white wine is a continuous curve in the visible region of the UV-Vis spectrum (Figure 1). It is clear from Figure 1 that there is no peak or absorbance maximum at 420 nm or at any other wavelength. However, absorbance is higher at shorter wavelengths, and for this reason has been easier to measure than absorbance at longer wavelengths. Typically, a young white wine upon ageing will have a colour change from virtually colourless, green tinged, pale straw, straw, yellow, gold, amber and brown. This colour change is related to increasing absorbances throughout the visible spectrum. It is worth noting that our perception of colour is determined by a combination of all these absorbances. Technical Review No.147 December 2003

Absorbance in Flint bottle (a.u.) Figure 1. Absorbance from 380 to 700 nm, determined in the spectrophotometer using the modified sample port, of blends of an oxidised (100/0) and a non-oxidised (0/100) wine (v/v) in Flint-P 15587 Claret bottles (ACI). Reproduced from Skouroumounis, G.K.; Kwiatkowski, M.J.; Sefton, M.A.; Gawel, R.; Waters, E.J. In situ measurement of white wine absorbance in clear and coloured bottles using a modified laboratory spectrophotometer. Australian Journal of Grape and Wine Research 9: 138-148; 2003 (with permission from the Australian Society of Viticulture and Oenology). 71 Colour measurements are currently made on the wine after opening the bottle and cannot, therefore, be used to select bottled stock in which excessive browning has occurred. Currently, wineries and retailers need to visually assess bottled stock in order to do this. This demanding task is made more difficult, and is barely possible, when the wine is in coloured bottles. Being able to measure colour non-intrusively (i.e. without opening the bottle) would be of obvious value. The need for individual bottle measurements relates to the variability of closures and storage conditions. Variation in browning can occur during ageing due to different oxygen permeations exhibited by the closures and/or closure/glass interfaces (AWRI publication #550). Using A 420 to measure white wine browning without opening the bottle, however, presents a number of challenges, including the suitability of the wine bottle, with its curved surface of coloured glass and variable thickness. Absorbance measurements are generally made in non-coloured, transparent glass or in plastic cuvettes that have a flat surface that is held perpendicular to the light beam, and which have walls of controlled thickness; bottles made by the industry come in different sizes, shapes and colours. Despite these challenges, we have shown (AWRI publication #731) that it is possible to use A 420 to measure white wine browning without opening the bottle. We achieved December 2003 Technical Review No.147

this using a Shimadzu 1201 UV-VIS spectrophotometer with the following modifications: a modified sample port comprising a sample bay, lid component and a capping device. The modified sample port allows the measurement of colour absorbance in the visible region of the bottle, whereby maximum and minimum absorbances of different coloured bottles can be recorded and noted (Figure 2). In fact, bottles of the same colour manufactured at different sites can be scanned and compared, ensuring quality standards are in place for bottle manufacture and that bottle colour is reproducible. 72 Figure 2. Absorbance spectra of empty (a) Claret type bottles of Classic Green, Emerald Green, Georgia Green and Flint, and (b) Burgundy type bottles of French Green, Antique Green, Amber and a Claret bottle of Cobalt Blue from 300 to 800 nm, determined in the spectrophotometer using the modified sample port. All bottles sourced from ACI Reproduced from Skouroumounis, G.K.; Kwiatkowski, M.J.; Sefton, M.A.; Gawel, R.; Waters, E.J. In situ measurement of white wine absorbance in clear and coloured bottles using a modified laboratory spectrophotometer. Australian Journal of Grape and Wine Research 9: 138-148; 2003 (with permission from the Australian Society of Viticulture and Oenology). Technical Review No.147 December 2003

A linear correlation can be obtained between A 420 measured in a cuvette (10 mm pathlength) and A 420 of that same wine measured in a 750 ml Flint bottle (colourless transparent glass). This method can also be applied to bottles of different colours such as Emerald Green, Classic Green, French Green, Georgia Green, and Cobalt Blue. Light at 420 nm was almost totally absorbed by Antique Green and Amber glass (Figure 2 b]) and thus A 420 of white wines could not be measured directly in these bottles with the UV-Vis spectrophotometer used in this study. The A 540 value of a wine, however, could readily be determined in these bottles and a good linear relationship between A 540 in the bottle and A 420 in cuvettes was obtained for blends of oxidised and non-oxidised wines. A good linear relationship between A 600 in Amber bottles and A 420 in cuvettes was also obtained (note: we used these alternative wavelengths as this resulted in minimal interference from the coloured glass, see Figure 2). Indeed, because of the relatively low absorbance by white wine above 500 nm, absorbances could be measured more accurately in these bottles than in cuvettes because of the longer pathlength of the former. A linear relationship between A 540 (bottle) and A 420 (cuvette) that has been derived from one wine should not be used to measure A 420 indirectly in another wine, as different wines might have different shaped absorbance curves. When measuring A 540 in the bottle to determine the A 420 variation of wines bottled in Antique Green or Amber glass, it is necessary to open a subset of those wines first in order to demonstrate that the relationship between A 540 (bottle) and A 420 (cuvette) is linear and that an adequate calibration curve can be obtained for those bottles and for those wines. 73 A case study of wine oxidation was undertaken using the method described here. Winemakers at a large Australian winery had noted, upon tasting, a high incidence of excessive oxidation in a six-year old bottled Chardonnay wine. This wine was bottled in Antique Green Burgundy bottles and thus it was not possible to visually assess the wine colour. The bottles had front and back labels and some additional stickers, but there was a gap of approximately 27 mm between the labels through which absorbance measurements could be determined. Because the wines were bottled in Antique Green glass, in situ determination of A 420 of the wine inside the bottles was not possible. However, full spectral measurements of various wine blends indicated that determination of absorbance at longer wavelengths might be used to discriminate between oxidised and non-oxidised wine (Figure 1). Accordingly, absorbance at 540 nm was measured for these wines as the glass of the Antique Green bottles had an absorbance minimum near this point (Figure 2 b]). A 540 December 2003 Technical Review No.147

values ranging from 0 to 0.885 a.u. (Figure 3) with a mean value of 0.229 a.u. and standard error of 0.008 a.u. were established from 597 bottles. 74 Figure 3. The range of A 540 values shown in April 2002 by 597 bottles of a Chardonnay wine. Number of wines in 14 categories of A 540 spanning the complete range of values measured are indicated. Reproduced from Skouroumounis, G.K.; Kwiatkowski, M.J.; Sefton, M.A.; Gawel, R.; Waters, E.J. In situ measurement of white wine absorbance in clear and coloured bottles using a modified laboratory spectrophotometer. Australian Journal of Grape and Wine Research 9: 138-148; 2003 (with permission from the Australian Society of Viticulture and Oenology). Of these bottles, 16 were opened and tasted by the company s winemakers and the levels of free and total SO 2 and volatile acidity in these wines were also determined. The A 540 (bottle) values correlated with decreasing free and total SO 2 levels (r 2 of 0.712 and 0.779 respectively), and for this particular batch, wines with A 540 (bottle) values equal to or above 0.247 a.u. were considered by the winemakers as being commercially unacceptably oxidised. Collecting these data enabled the wine company to set a specification for absorbance in these bottles and to screen them, non-intrusively, for excessive oxidation. The non-intrusive nature of this method allows bottled stock to be screened at any time. Even without calibration curves, it is a simple matter to distinguish high absorbance wines from those with low absorbance in any given batch. Providing absorbance specifications have been set by wineries, incidents of sporadic or accelerated ageing can be dealt with before or after market release because wineries and retailers can identify and remove wines with an unacceptably high absorbance. This has the potential to result in an increase in overall quality of white wines available in the market. Technical Review No.147 December 2003

The spectrophotometer used in this study was a basic model and exhibited linearity up to 2.3 a.u. in the bottle (0.305 a.u. cuvette equivalents) at A 420. Improved technology should enable the automation of white wine monitoring. New instruments claim a linearity of up to 7 absorbance units, triple the capacity observed here for a basic model. This would allow colour changes in bigger bottles, such as Champagne Burgundy bottles, to be monitored. George Skouroumounis Research Chemist References Godden, P.; Francis, L.; Field, J.; Gishen, M.; Coulter, A.; Valente, P.; Høj, P.; Robinson, E. (2001) Wine bottle closures: physical characteristics and effect on composition and sensory properties of a Semillon Wine 1. Performance up to 20 months post-bottling. Aust. J. Grape Wine Res. 7, 62 105 (AWRI publication #666). Iland, P.G.; Ewart, A.J.W.; Bruer, D.R.G.; Hartnell, C. (1988) Effects of skin contact, ph and SO 2 on young red wine composition. Smart, R.E.; Thornton, R.J.; Rodriguez, S.B.; Young, J.E., eds. Proceedings of the second international symposium for cool climate viticulture and oenology, Auckland, New Zealand, 11 15 January 1988. Auckland: New Zealand Society for Viticulture and Oenology; 1988: 225 226. 75 Singleton, V.L.; Kramling, T.E. (1976) Browning of white wine and an accelerated test for browning capacity. Am. J. Enol. Vitic. 27, 157 160. Skouroumounis, G.K.; Kwiatkowski, M.J.; Sefton, M.A.; Gawel, R.; Waters, E.J. (2003) In situ measurement of white wine absorbance in clear and coloured bottles using a modified laboratory spectrophotometer. Aust. J. Grape Wine Res. 9, 138 148 (AWRI publication #731). Somers, T.C.; Ziemelis, G. (1985) Spectral evaluation of total phenolic components in Vitis Vinifera: grapes and wines. J. Sci. Food Agric. 36: 1275 1284 (AWRI publication #283). Somers, T.C.; Evans, M.E. (1977). Spectral evaluation of young red wines: anthocyanin equilibria, total phenolics, free and molecular SO 2 chemical age. J. Sci. Food Agric. 28, 279 287 (AWRI publication #151). Waters, E.J.; Peng, Z.; Pocock, K.F.; Williams, P.J. (1996) The role of corks in oxidative spoilage of white wines. Aust. J. Grape Wine Res. 2, 191 197 (AWRI publication #528). Waters, E.J.; Williams, P.J. (1997) The role of corks in the random oxidation of bottled wines Aust. N.Z. Wine Ind. J. 12 (2): 189 193 (AWRI publication #550). Zoecklein, B.W.; Fugelsang, K.C.; Gump, B.H.; Nury F.S. (1995) Wine analysis and production. Chapman & Hall, New York: 146 151. December 2003 Technical Review No.147