Abstract
Dissolved oxygen was measured during low-level oxygenation treatments. As previously reported in the literature, continual oxygen additions augmented the dissolved oxygen levels of wines. Results here showed concentrations up to 2.4 mg oxygen/L. Conversely, nonoxygenated wines had concentrations as low as 4 μg oxygen/L in the center of the tanks. The analysis of dissolved oxygen in a winery setting requires an appropriate sampling procedure and a sensitive measuring device.
Microoxygenation is a popular commercial process in which small amounts of oxygen are introduced in the wine matrix with the aim of improving its sensory profile (Parish et al. 2000, Rieger 2000). Although many trade articles have been written on microoxygenation, only a few well-controlled experiments have been published in major peer-reviewed scientific journals covering the effects of this practice on wine chemistry (Atanasova et al. 2002, Llaudy et al. 2006, Cano-López et al. 2006, Pérez-Magariño et al. 2007). Clearly, more research is required to understand and better control the outcome of microoxygenation.
To become reactive, oxygen in its natural triplet state, a biradical with two unpaired electrons, needs to be activated to its singlet radical forms (Singleton 1987), most likely with the involvement of transition metals catalysts (Danilewicz 2003). For activation to occur, oxygen must be dissolved in the wine in a process in which concentration gradients, partial pressure of gases, and wine temperature are fundamental. Compared with room temperature, oxygen solubility increases approximately 10% at temperatures of ~5°C. At atmospheric pressure and room temperature, oxygen saturation, with air, is reached at ~6 mL/L. Under continued agitation, that saturation level can be reached in approximately 30 sec (Ribéreau-Gayon 1933). After oxygen depletion, several new uptakes can occur, resulting in a total consumption of ~60 to over 600 mg/L O2 for white and red wines, respectively (Singleton 1987). If wine were instead exposed to high doses of pure oxygen, then saturation levels of ~30 mL/L would be expected (Singleton 2000). Nonetheless, with the typical oxygen flow rates used in microoxygenation, the dissolved oxygen (DO) concentration of wine is not expected to approach such levels (Waterhouse and Laurie 2006). Dissolved oxygen readings between ~10 and 120 μg/L have been reported for wines undergoing microoxygenation (Castellari et al. 2004).
The sensors most typically used for DO determinations are electrolytic cell based methods (Ramamoorthy et al. 2003, Cheynier et al. 2002). These sensors give access to partial pressures and can be calibrated in percentage saturation (Moutounet and Mazauric 2001, Cheynier et al. 2002). In these systems, the sensor consists of several electrodes submerged in an electrolyte solution and separated from the exterior by an oxygen-permeable membrane. When an electrical potential is applied between the two electrodes, the oxygen is reduced, thus generating an electrical response that is proportional to the oxygen content (Ramamoorthy et al. 2003).
The purpose of this research was to describe the levels of dissolved oxygen in industrial wine tanks during microoxygenation treatments. This leads to recommendations for dissolved oxygen measurement techniques and a discussion of the effectiveness of such measurements in a typical winery setting.
Materials and Methods
Wines.
Two different experiments, conducted over two consecutive vintages, 2002–2003 (experiment I) and 2003–2004 (experiment II), were performed at Wente Winery, Livermore, California. In experiment I, Cabernet Sauvignon grapes from Livermore (23.5 Brix), were destemmed, crushed, fermented, pressed, and transferred into a single 132,489-L tank where malolactic fermentation (MLF) took place. No sulfur dioxide was added at this stage. After MLF completion, the wine was settled for four days, and the clean supernatant (no turbidity measurements were performed) racked to four tanks of 11,553 L each to proceed with oxygenation treatments. At this point wine composition was ethanol concentration, 14.1%; pH, 3.9; titratable acidity, 6.6 g tartaric acid/L; total polyphenolics, 2393 mg gallic acid equivalents/L; residual sugars, 1.6 g/L; free SO2, 8 mg/L; and total SO2, 12 mg/L. In experiment II, to reduce the possibility of unwanted microbial growth, a similar type of wine was vinified as in experiment I and was treated using high-temperature short-time (~70°C for 15 sec followed by cooling to 15°C). Wine composition was ethanol concentration, 14.8%; pH, 3.75; titratable acidity, 5.7 g tartaric acid/L; total polyphenolics, 2734 mg gallic acid equivalents/L; residual sugars, 1.9 g/L; free SO2, 10 mg/L; and total SO2, 12 mg/L.
In both experiments, the wines were racked using nitrogen blanketing in the source tank and dry ice in the reception tanks to maintain the concentrations of DO as low as possible. Concentrations of DO in all tanks before the start of oxygenation treatments were between 15 and 25 μg/L.
Oxygen delivery system.
Oxygen gas, 99% purity (Airgas, Woodland, CA) was delivered into the wine tanks using Nylotube tubing (6.35 mm diam) (NewAge Industries, Southampton, PA) from a 7778-L (275 cubic feet) oxygen cylinder (Figure 1⇓). During experiment I, the oxygen flow applied was regulated with a custom-built flow controller and checked periodically with a calibrated standard gas bubble meter. The oxygen bubbles were created using a 1-micron porous stainless-steel cylindric sparger (6.35 mm diam x 12.7 mm length) located ~30 cm from the bottom of the tank. During experiment II, oxygen flow was regulated by commercial microoxygenation systems (Stavin, Sausalito, CA; Oenodev, Villeneuve-lés-Maguelones, France), while oxygen bubbles were generated using a 1-micron porous stainless-steel cylinder (6.35 x 12.7 mm) in two of the tanks and a commercial oxygen diffuser (Oenodev) in the third tank. The control tank was not oxygenated.
Oxygenation regimes.
For experiment I, Cabernet Sauvignon, with and without oak staves bundled to tank inner walls, was subject to four treatments: (1) no oxygen–no oak: no oxygen addition and no oak staves presence; (2) oxygen–no oak: oxygen addition at a fixed rate of 5 mL oxygen/L of wine per month for 6 months and no oak staves; (3) no oxygen–oak: no oxygen addition, and oak staves (80.3 cm2 French oak staves, medium plus toast [Stavin] per L of wine, bundled to the tank inside walls with stainless-steel wire); and (4) oxygen–oak: oxygen addition (as in treatment 2) and oak staves (as in treatment 3). For experiment II, no oak staves were included, and each wine treatment (labeled oxygen 1, 2, 3, and no oxygen) was treated with different oxygen rates (see Table 1⇓). These doses were suggested by a provider of the microoxygenation technique in California, after periodic winetasting. In this case, oxygenation treatment started before MLF, ceased during this process, and continued after MLF completion, which was induced by direct inoculation of lactic acid bacteria.
Dissolved oxygen in tanks.
To allow for wine sampling and DO determinations while keeping the tanks closed at all times, a sample circulation circuit constructed in 316 stainless-steel tubing (6.35 mm diam) (Webco Industries, Kansas City, MO) with a gas purge on top of the tanks was used (Figure 1⇑). The oxygen–no oak treatment was set up with four sampling points at different locations in the tank: (1) top front above the sparger, (2) top rear, (3) bottom rear, and (4) center. All other tanks were sampled from the center.
Dissolved oxygen concentrations were measured in situ using an Orbisphere 2713 series oxygen analyzer, model 26060 indicating instrument, and model 2110 sensor (Orbisphere, Geneva, Switzerland). The analyzing principle of the sensor is based on that of an electrolytic cell. Measurements were read on a weekly or monthly basis, depending on the stage of the treatment. Before every reading, the DO meter was calibrated using air-saturated water as a standard according to manufacturer instructions. The sensors were periodically purged with nitrogen to check the zero calibration. The sensors were reconditioned and the membranes replaced if the zero calibration was above 2 μg/L.
To test the reproducibility of the measurements at different DO concentrations (μg to mg/L), four calibrated gas samples from 0 to 4% oxygen (Air Liquide, Houston, TX) were sparged into a 5-L container with Cabernet Sauvignon wine until saturation was reached. Each treatment was sampled at three different dates, using a circulation circuit as previously described, obtaining percent standard deviation values <3% (Table 2⇓). Repeatability measures were performed at three different dates during experiment I, giving values below 4% (Table 3⇓).
Other analyses.
Wine temperature was measured using the Orbisphere equipment, and microbial populations were monitored using counting chambers, Petroff–Hausser for bacteria and Levy–Hausser for yeast according to manufacturer protocols (Hausser Scientific, Horsham, PA). While temperature has a marked effect on oxygen dissolution (Singleton 2000), microbes might consume part of the oxygen added and eventually spoil the wine being treated (Drysdale and Fleet 1988, Fornairon-Bonnefond et al. 2003). Wine composition analyses (ethanol content by ebuillometry, pH, titratable acidity, total polyphenolics by Folin-Ciocalteu, and free and total sulfur dioxide by aeration-oxidation) were carried out according to standard procedures (Ough and Amerine 1988).
Results and Discussion
Dissolved oxygen in tanks.
Experiment I.
After the first month of oxygenation, the oxygenated treatments, with or without oak staves, had increasing DO concentrations ranging from ~50 to 270 μg/L, while the treatments without oxygen addition showed lower and fairly consistent concentrations at ~25 to 30 μg/L. These concentrations are close to the median reported (24 mg/L) for 175 red wines in stainless-steel tanks (Castellari et al. 2004) and those reported elsewhere (Cheynier at al. 2002). After 6 month of oxygenation, the oxygen-treated wines had ~10-fold more DO than the untreated tanks (Figure 2⇓). In addition, the presence of oak staves did not seem to have a significant effect on DO, either at the beginning of the experiment or later. A similar effect was reported when oxygen was added at doses of 5 mL O2/L wine per month (Castellari et al. 2004).
Regular monitoring of wine temperature throughout the experiment showed minor diurnal variation (<2°C), and virtually no temperature disparity between tanks (<0.3°C), with a seasonal deviation from 13.5 to 17.5°C. The fluctuation in DO observed for the oxygenated treatment between May and July was consistent with an increase of 2 to 2.5°C in temperature observed in May, while no unusual microbial growth was detected at this or any stage throughout the experiment (data not presented). Although such temperature raises might result in a reduction of the oxygen solubility, we were unable to explain the observed DO drop.
Oxygen concentration measurements in the vented gas from these tanks suggested that more than 80% of the oxygen added was being consumed. The oxygen injection rates were ~1 mL/min, with vent gas rates varying between 3 and 8 mL/min over the 6 months of experimental period. Additional measurements to evaluate the DO at the four sampling points in the oxygen–no oak treatment (Figure 1⇑) showed very little variation (<4%), indicating an even DO distribution in this tank over 6 months of treatment.
Typical wine composition values were observed at the end of experiment I (Table 4⇓). The variation in phenolic concentration could be explained by the extraction of phenolic species from the oak staves.
Experiment II.
Similar to experiment I, the DO concentration in the oxygenated tanks showed consistently higher values than the control, especially before MLF (Figure 3A⇓), when higher oxygen flow rates were applied (Table 1⇑). Dissolved oxygen in the control tank was between ~4 and 14 μg/L, while DO in oxygenated tanks was between ~220 and 2400 μg/L. After malolactic fermentation, when the oxygen doses were reduced to between 4 and 0.5 mL oxygen/L wine per month, the differences between the oxygen-treated tanks and the control become less noticeable (Figure 3B⇓). DO values for the control were between 7 and 23 μg/L, while values for the oxygen-treated tanks were between ~8 and 60 μg/L.
Once again, temperature measurements were consistent across tanks, and diurnal variations were less than 2.5°C. The seasonal deviation in wine temperature registered was between 13.0 and 18.5°C. Wine composition data at the end of experiment II are presented in Table 4⇑. Considering the limitations to the experimental design of these industrial-scale experiments, these data can only be used as an indication of the possible effects of oxygen exposure on wine chemistry. The apparent lower levels of total phenolics observed in the oxygen-treated wines could have been the result of precipitation due to higher phenolic polymerization.
The above confirms previous observations indicating that discrimination between oxygen levels when using low rates of oxygen treatment demands the use of very sensitive equipment and appropriate sampling procedures that exclude oxygen before and during measurements. It is paramount that winemakers pay special attention to practices in which low temperatures are used, as higher dissolved oxygen rates will be expected (Castellari et al. 2004, Waterhouse and Laurie 2006).
Conclusions
Dissolved oxygen increased with low-level oxygenation treatments, suggesting that the rate of oxygen consumption is slower than that of oxygen dissolution. The low levels of oxygen demand specialized oxygen meters for accurate and precise measurements. In this case, these increases were noticeable with doses as low as 1 mL/L of wine per month, regardless of which oxygen delivery system was used.
Footnotes
Acknowledgments: The authors thank the American Vineyard Foundation, California Competitive Grant, and Wente Vineyards. V.F.L. also thanks the Fulbright, Laspau, Wine Spectator, Rhone Rangers, and Jastro Shields Scholarships, and the assistance of Annegret Rust. Oenodev microoxygenation system was provided by Vinovation, Sebastopol, CA.
- Received June 2006.
- Revision received September 2007.
- Revision received November 2007.
- Copyright © 2008 by the American Society for Enology and Viticulture