Abstract
Iron plays a key role in wine oxidation. Polyphenols that contain catechol systems are the main reductants, and it has been proposed that the oxidation of these substances is mediated by the redox cycling of the Fe(III)/Fe(II) couple. At any time, the Fe(II):Fe(III) concentration ratio should depend on the rate of Fe(II) oxidation by oxygen relative to that of Fe(III) reduction by polyphenols. Fe(III) oxidation of polyphenols, although facilitated by sulfite, is somewhat slower than the reaction of Fe(II) with oxygen, which is strongly accelerated by Cu. Alongside this process, Fe(III) inhibits is own formation. Therefore, the Fe(II):Fe(III) concentration ratio is determined by the interplay of a number of competing reactions. However, because of the relative speed of Fe(II) oxidation, oxygen should be a major determinant of this ratio. A simple spectroscopic method involving ferrozine is used to measure Fe(II) concentration in wines collected under nitrogen with minimal disturbance so as to determine Fe(II) levels in the original wine container. However, Fe(III), which becomes a strong oxidant in the presence of ferrozine, oxidizes catechols in wine conditions. Therefore, Fe(II) concentration, which increases as a result of catechol oxidation, was monitored over time and extrapolated back to the moment of ferrozine addition. Total Fe concentration was determined by adding ascorbic acid to reduce the Fe(III). As expected, the Fe(II):Fe(III) ratio was higher in wines bottled with screw caps than in those bottled with natural cork or filled in boxes. Exposure of wines to oxygen lowered the ratio, which reached equilibrium after some days of aerial saturation. However, the ratio attained differed in the different wines, and this difference likely depends on wine constituents that alter the relative rate of Fe(II) oxidation to that of Fe(III) reduction.
Iron (Fe) plays a key role in wine oxidation (Waterhouse and Laurie 2006). Removal of Fe from white wine prevents its oxidation, and Fe addition accelerates it (Danilewicz and Wallbridge 2010). Polyphenols that contain a catechol or galloyl moiety are the principal reductants, and it has been proposed that the oxidation of these substances is mediated by the redox cycling of the Fe(III)/Fe(II) couple (Scheme 1) (du Toit et al. 2006, Waterhouse and Laurie 2006).
The concentration of oxidizable polyphenols, expressed as caffeic acid equivalents (CafE), in a short series of white wines was estimated to range from 25 to 105 mg/L (mean 56 mg/L) in a study using cyclic voltammetry and from 53 to 85 mg/L (mean 78 mg/L) using a ferric ion reducing antioxidant power (FRAP) assay (Danilewicz 2015). Taking a value of 80 mg/L, their concentration should be ~4.5 × 10−4 mol/L (CafE). On the other hand, white wine should contain 5 to 6 g/L of tartaric acid (TA), together with approximately the same amount of malic acid (MA), which co-ordinates similarly to Fe(III). Therefore, the combined concentration of these two acids should be ~7 × 10−2 mol/L, in vast molar excess to polyphenols, allowing Fe co-ordination with these acids to dominate. Also, TA co-ordinates strongly to Fe(III), forming mainly a dimeric complex, Fe2(L2−)2(−3H+), at wine pH, where LH2 represents TA (Timberlake 1964, Yokoi et al. 1994). This complex displays an absorption maximum at 335 nm, which is unaffected by the addition of 4-methylcatechol (4-MeC) (3.8 × 10−4 mol/L) to model wine (Danilewicz 2014). MA behaves similarly, and it appears, therefore, that the catechol cannot displace these acids significantly in white wine conditions.
In contrast, acetic acid co-ordinates to Fe(III) more weakly and therefore cannot prevent catechol co-ordination with Fe(III). Consequently, addition of 4-MeC (3.8 × 10−4 mol/L) to acetate buffer with pH 3.2 results in the immediate formation of the catecholate complex. This complex displays a broad absorption maximum at ~400 nm, which fades over ~12 hr as the catechol is oxidized (Danilewicz 2014). Although it appears that TA and MA are the dominant Fe(III) ligands in wine, it is proposed that catechol–Fe(III) complexes must form to some small extent to allow oxidation to occur because the aforementioned study indicates that electron transfer occurs within the complex.
Reduction potentials have been considered indicators of redox status in wines. However, such potentials, by definition, can be generated only by systems that are in thermodynamic equilibrium, which is not the case for wine. The two main reactions, O2 reduction and polyphenol oxidation, each involving Fe, are irreversible. The observed so-called reducation potentials depend on the O2 concentration and are now understood to be caused by ethanol oxidation (Danilewicz 2012). Consequently, although reduction potentials cannot be used to define the redox status of reacting redox couples in wine conditions, they can be used to determine the theoretical feasibility of such redox interactions. The reduced component of a couple will reduce the oxidized component of a second couple that has a higher reduction potential, where ΔE = −ΔG/Fν; ΔE is the difference in reduction potential, F is the Faraday constant, ν the number of electrons transferred, and ΔG is the change in Gibbs free energy. A positive ΔE will give a negative ΔG, indicating a favorable transformation, which will occur if a mechanistic pathway is available (Atkins 1999).
TA co-ordinates more strongly with Fe(III) than with Fe(II) (Timberlake 1964), and by displacing the Fe(III):Fe(II) redox equilibrium towards Fe(III), TA lowers the reduction potential of the Fe(III)/Fe(II) couple to ~345 mV (versus the standard hydrogen electrode) in model wine with a pH of 3.6 (Danilewicz 2013, 2014). This potential is, therefore, lowered below that of the O2/H2O2 couple (E3.6 = 570 mV), and Fe(II) is readily oxidized in wine conditions, a process which is accelerated by Cu (Ribéreau-Gayon et al. 2000, Danilewicz 2013).
More detailed studies have shown that Fe(II) oxidation slows markedly when a Fe(III):Fe(II) ratio of between ~2:1 and 3:2 is reached, depending on the concentrations of Fe and Cu. It has also been observed that Fe(III) inhibits Fe(II) oxidation, and a mechanism in which Fe(III) competes with Fe(II) for an intermediate Fe(III)-superoxo-complex has been proposed (Scheme 2). In effect, Fe(III) inhibits its own production, which would therefore slow as Fe(III) concentration increases. Net Fe(II) oxidation, which proceeds to completion in the absence of polyphenols, results in the irreversible production of H2O2, which is generated without the intermediate formation of hydroperoxyl radicals by the hydrolysis of a di-iron(III),(III)-dioxygen complex (Danilewicz 2013). H2O2 is rapidly reduced by Fe(II) to produce hydroxyl radicals in the Fenton reaction (Elias and Waterhouse 2010).
However, although TA facilitates Fe(II) oxidation, it also lowers the reduction potential of the Fe(III)/Fe(II) couple below that of a quinone/catechol couple (E3.6 ~575 mV), and the Fe(III) oxidation of a catechol is thermodynamically disfavored. In model wine, catechols are not oxidized alone or at least are oxidized extremely slowly (Danilewicz 2007, 2011). The coupling of the oxidation step with a thermodynamically more favorable process is necessary to drive the reaction forward. In wine, this process is provided by the reaction of sulfite with quinones, which are reduced back to catechols, a reaction that is essentially quantitative with (+)-catechin (Danilewicz and Wallbridge 2010). Studies with 4-MeC have shown that Michael addition of sulfite to the quinone can also occur to produce ~38% of the sulfonic acid (Danilewicz et al. 2008). In line with these findings, Fe(III) does not oxidize 4-MeC in model wine but does so when sulfite is added (Danilewicz 2014). Pyrogallol, which is a more powerful reductant than catechols, is oxidized alone, but its rate of reaction is still accelerated by sulfite (Danilewicz 2011). Other nucleophiles, such as thiols, may also compete with sulfite for addition to quinones (Nikolantonaki and Waterhouse 2012). In white wines, MA co-ordinates with Fe in a manner similar to TA and therefore affects the reactivity of the Fe couple similarly (Danilewicz 2014).
Given the evidence for Fe redox cycling, the Fe(II):Fe(III) concentration ratio should depend on the relative reaction rate of Fe(II) with oxygen and that of Fe(III) with polyphenols, which should depend on O2 concentration, the reactivity and concentration of polyphenols, and on the concentrations of Fe, Cu, and sulfite. Because of the speed of Fe(II) oxidation, O2 concentration should have a major effect on this ratio, which should largely depend on how much O2 a wine is exposed to and for how long, as it may take some time for the system to reach equilibrium.
Wine Fe(III) and total Fe concentration have been previously measured with adsorptive stripping voltammetry (Wang and Mannino 1989), with a flow-through fluorescent sensor (Pulido-Tofiño et al. 2000), and by extraction of the Fe(III) thiocyanate complex with methyl isobutyl ketone and sequential injection analysis with atomic absorption spectrometry (AAS) (Costa and Araújo 2001). The Fe(II):Fe(III) ratio has been determined by ion-exchange chromatography coupled with AAS (Ajlec and Štupar 1989) and also by methyl isobutyl ketone extraction of the Fe(II)-1,10-phenanthroline and Fe(III)-isocyanate complexes (Tašev et al. 2006). Methods have also been described for measuring free Fe and what has been described as bound Fe (Paleologos et al. 2002, Pohl and Prusisz 2009). However, as discussed above, Fe is likely to be co-ordinated principally to TA or MA in white wines.
The Fe(II):Fe(III) concentration ratio has been determined more simply by spectrophotometry at pH 5.5. Fe(II) concentration was measured with 2-(5-bromo-2-pyridylazo)-5-diethylaminophenol, which forms a colored complex (λmax 748 nm). Total Fe concentration was then determined by first adding ascorbic acid to reduce the Fe(III) to the ferrous state (Ferreira et al. 2007). A similar method, with the pH adjusted to 4.9 and using 2,2-dipyridylketone picolinoylhydrazone, which forms a blue-green complex (λmax 700 nm) with Fe(II), has also been recently described (López-López et al. 2015).
However, in these methods, samples are diluted and exposed to air for varying times and their pH is altered, which could disturb existing Fe(II):Fe(III) ratios. Also, the ratios, which should be variable, are reported as single values for each wine irrespective of the method of wine storage and how they were then handled. The aim of this investigation was to develop a simple method to determine the Fe(II):Fe(III) concentration ratio in white wines with minimal sample disturbance and to determine how the ratio may change on wine exposure to oxygen. In effect, to see if the ratio can give an indication of redox status.
Materials and Methods
Materials
Water (Emsure, Fe ≤ 1 μg/L, Cu ≤ 0.4 μg/L, E. Merck), Cu(II) sulfate pentahydrate, sodium hydroxide, l-(+)-tartaric acid (BDH AnalaR grade), and ethanol (96% GPR grade) were obtained from VWR International. Ascorbic acid, Fe(II) sulfate heptahydrate (99+% ACS reagent), 2,6-dichlorophenol-indophenol, 3-(2-pyridyl)-5,6-diphenyl-1,2,4-triazine-p,p′-disulfonic acid monosodium salt (ferrozine), and 2,4,6-tripyridyl-s-triazine (TPTZ) were obtained from Sigma-Aldrich, and potassium metabisulfite (Kadifit) was purchased from Erbslöh Geisenheim AG.
Wines
Sauvignon blanc (Western Cape, South Africa) and Pinot Grigio (Hungary) wines were obtained from Sainsbury’s Supermarket Ltd. in London EC1N 2HT. Soave (Italy) and Pinot Grigio (Las Moras, Argentina, 2013) wines were purchased from Co-operative Group Ltd. Manchester M60 0AG. The South African Sauvignon blanc, Hungarian Pinot Grigio, and the Italian Soave were all nonvintages from wine boxes. A Pinot Grigio (Blossom Hill, California, 2013) wine was from Blossom Hill Co. London NW10 7HQ. A Pinot Grigio (Lancinni Vino, 2013) wine was from Cita Paderno, Italy, and a Chardonnay wine (Terre & Vigne, Pays D’Oc, 2014) was from Saint Félix de Lodez, France. The pH, total Fe concentration, and initial O2 and sulfite concentrations of the wines are given in Table 1.
Preparation of model wine solutions
(+)-TA (8.0 g) was dissolved in ~750 mL water in a 1-L volumetric flask. Ethanol was added to give a 12% (v/v) final concentration. The pH was increased to 2.75 and 3.20 with 2.5 N sodium hydroxide, and water was added progressively to the mark as the required pH was approached. The TA concentration selected simulated the action of all the wine acids, including MA, and produced a final titratable acidity of ~4 g/L. Both MA and TA co-ordinate with Fe and should affect the reactivity of the Fe(III)/Fe(II) redox couple similarly (Danilewicz 2014).
Measuring the Fe(II):Fe(III) concentration ratio in wine
A calibration curve was obtained by adding FeSO4.7H2O to the model wine containing a nine-fold excess of ferrozine and 2 mole equivalents of ascorbic acid to ensure that no trace of Fe(III) was present, as Fe(II) is rapidly oxidized in model wine. The relationship: absorbance = 0.4952 × C + 0.0156 (R2 = 0.9999), where C equals the Fe(II) concentration in mg/L, was obtained for Fe(II) concentrations ranging from 0.5 to 4.0 mg/L, which gave a molar absorptivity of 27,645 L/mol.cm at pH 3.2. At pH 2.75, the relationship: absorbance = 0.5066 × C + 0.0075 (R2 = 0.9998) was obtained, with a molar absorptivity of 28,536 L/mol.cm (Lit. εmax 27,900 L/mol.cm at pH 5.5) (Stookey 1970). To check the calibration curve and overall procedure, the total Fe concentration in four wines was shown to be close to values obtained by AAS (Table 1).
Wine bottle closures were removed, and a gentle nitrogen stream was quickly introduced into the headspace. A setup for taking wine samples consisted of a cork fitted with a thin plastic tube that reached the bottom of the bottle and extended some 40 cm from the top and of a short tube that extended only into the headspace. The tubes were flushed with nitrogen, and the assembly was inserted into the bottle. Applying a gentle nitrogen pressure into the headspace helped expel wine (~120 mL) into a nitrogen-filled 250-mL flask; the initial ~20 mL of wine was discarded. Wine was also dispensed into nitrogen-filled ~68-mL brown bottles to measure the starting O2 concentration as described below. When the wine was taken from wine boxes, a small amount was first flushed through the tap, and the sample was dispensed directly into the nitrogen-filled 250-mL flasks.
Wine samples (5.0 mL) were transferred to reaction tubes, and 25 μL ferrozine solution (containing 175 mg/5.0 mL H2O) was immediately added. Absorbance at 562 nm was measured in triplicate after 1 min and at 1-min intervals for 10 min. Least-square regression lines were drawn as shown in Figures 1 and 3 with coefficients of determination (R2). These lines were extrapolated to zero time to obtain the absorbance at the time of mixing, from which Fe(II) concentrations were calculated. The errors of these values could not be calculated precisely but were anticipated to be small because they were obtained from individual triplicate readings displaying very small standard deviations and from least-square regression lines with R2 values very close to 1. Furthermore, it would be anticipated that the extrapolated absorbance values should have standard deviations of less than ±1 to 2%, similar to those obtained over the first 1 to 3 min. Error bars denoting standard deviations are not shown because they were smaller than the data point symbol dimensions. Ten-millimeter quartz cuvettes were used and readings were taken against a wine blank. The spectroscopy sample was returned to the reaction tube, and 25 μL ascorbic acid solution (containing 126 mg/10.0 mL model wine) was added. Absorbance, which generally stabilized after 10 to 30 min, gave the total Fe concentration. After the initial measurements, the nitrogen was air-flushed, and the flask was shaken to saturate the wine with air and then stored in the dark. The Fe(II):Fe(III) ratio was measured as described above over several days in triplicate as the wine oxidized.
Measuring Fe(II) concentration with TPTZ
A calibration curve was obtained by adding FeSO4.7H2O to model wine containing excess TPTZ and by measuring absorbance at 593 nm. The relationship: absorbance = 0.3752 × C − 0.0107 (R2 = 1), where C = Fe(II) concentration in mg/L, was obtained for Fe(II) concentrations ranging from 1.0 to 6.0 mg/L, which gave a molar absorptivity of 21,513 L/mol.cm at pH 3.2 (lit. 22,600 L/mol.cm) (Stookey 1970). Wine Fe(II) concentration was measured as described for the ferrozine method by adding 75 μL TPTZ solution (containing 26.8 mg/L in 1.5 mL 0.04 M HCl) to 5.0 mL of wine.
O2 measurements
Wine was dispensed into nitrogen-filled ~68-mL brown glass bottles until they were approximately three-quarters full, and the bottles were then closed with screw caps. Oxygen concentration was measured with an HI-9146 dissolved O2 meter fitted with an HI-76408 Clark-type electrode (Hanna Instruments Ltd.); the manufacturer specified the resolution and limit of detection of the meter as 0.01 mg/L O2. For the O2 measurements, the bottle caps were quickly removed, and a small stirrer bar was inserted, followed by the electrode so that its tip was ~5 mm above the briskly stirring magnetic bar. Readings stabilized within 1 min and then remained stable for more than 5 min, showing that, even though the system was not sealed during measurements, no measurable amount of external O2 reached the measurement area during that time.
Measuring Fe(II):Fe(III) concentration ratio in model wine
Caffeic acid (44 mg/L) and (+)-catechin (78 mg/L) were dissolved together in 500 mL of model wine pH 3.2. The catechols were first dissolved in a small amount of ethanol to facilitate the dissolution. Potassium metabisulfite was then added to give 52.0 mg/L free SO2 (modified Ripper method). Three portions (150 mL) of this solution were then transferred to 250-mL flasks. To one was added Fe(II) (3.0 mg/L) as FeSO4, to a second, Fe(II) (3.0 mg/L) plus Cu(II) (0.2 mg/L) as CuSO4, and to the third, the same amounts of Fe and Cu with additional (+)-catechin to give a total concentration of 2.0 g/L. The flasks were shaken to aerial saturation, and the Fe(II) concentration was followed over time in triplicate with the ferrozine method as described for wine.
SO2 measurement
Wine free and bound SO2 concentrations were measured with the aeration-oxidation method and also with the modified Ripper method for model wines using KIO3-KI/starch (Ough and Amerine 1988a).
Ascorbic acid measurement
Ascorbic acid concentration was determined with the 2,6-dichlorophenolindophenol method (Ough and Amerine 1988b).
Fe and Cu measurements
Fe and Cu analyses were done by Campden BRI, using a Perkin Elmer 3110 atomic absorption spectrometer (Perkin Elmer Corp).
Measurements were taken in triplicate and mean values (±SD) were calculated, and figures were drawn with Excel software (Microsoft), which was also used to obtain least-square regression lines and coefficients of determination (R2). Experiments were conducted at ambient temperature (19 to 24°C).
Results and Discussion
Ferrozine has been used to determine Fe(II) and total Fe concentrations in ground and sea water after Fe(III) reduction with hydroxylamine (Viollier et al. 2000). This bidentate ligand has high affinity for Fe(II), with which it forms a magenta complex with an intense absorption band at 562 nm. An Fe(II) concentration that ranges from 0.5 to 4 mg/L can be measured directly with this method, which is ideal for wine. Ferrozine is also readily available and relatively inexpensive, and importantly, the ligand and its Fe(II) complex are both freely soluble in water. Manganese, which may be present at the same concentrations as Fe (Ough and Amerine 1988b), does not form a colored complex with ferrozine, and, of the metals present in wine, Cu is the only one reported as being capable of causing some interference in this assay (Stookey 1970). To assess the extent of this possible complication, Cu (0.2 mg/L) was added to a Chardonnay wine containing 2.6 mg Fe and <0.1 mg/L Cu, as measured by AAS. The total Fe concentration measured with ferrozine after ascorbic acid reduction was 2.52 ± 0.0 mg/L, a result that did not change after the Cu addition. Therefore, at the low concentrations found in wine, Cu should not interfere with the Fe(II) determinations with the ferrozine assay.
As mentioned above, Fe(III) does not readily oxidize catechols in model wine. However, ferrozine, because of its preferential co-ordination to Fe(II), raises the reduction potential of the Fe(III)/Fe(II) couple to ~1 V, making Fe(III) a strong oxidant (Danilewicz 2013). Consequently, 4-MeC is rapidly oxidized when ferrozine is added to model wine containing Fe(III) (Elias and Waterhouse 2010). It was expected, therefore, that absorbance would need to be measured quickly after ferrozine addition to wine, since the Fe(II) concentration should increase as Fe(III) is reduced by polyphenols. The immediate recording of absorbance was also advised for measuring Fe(II) concentration with ferrozine in water that contains dissolved organic material (Voelker and Sulzberger 1996).
The change in Fe(II) concentration during measurements was illustrated with a French Chardonnay wine, taken under nitrogen from a bottle closed with natural cork. The initial O2 concentration in this wine was surprisingly high at 0.63 ± 0.1 mg/L. Absorbance was measured 1 min after ferrozine addition to 5 mL of the wine and then at 1-min intervals for 10 min (Figure 1, curve T = 0). Extrapolation of the trend-line to the moment of mixing (t0), where x = 0 in the trendline equation shown below the curves, gave an initial Fe(II) concentration of 0.84 mg/L. After 1 min (t1), this had increased to 0.86 mg/L and after 10 min (t10), to 0.95 mg/L. This increase illustrates that Fe(III) was being reduced, and it was important to determine Fe(II) concentration at the moment of mixing when ferrozine is added to determine the Fe(II) concentration actually present in the sample. The wine was then saturated with air, and the Fe(II) concentration was measured at the time shown in Figure 1. After 68.5 hr, Fe(II) concentrations at t0, t1, and t10 had declined to 0.46, 0.48, and 0.58 mg/L, respectively (curve T = 68.5 hr). The increase in absorbance was relatively small at 1 min but became quite significant by 10 min. The total Fe concentration was 1.28 mg/L, determined by ascorbic acid addition, and the percentage of Fe present as Fe(II) was plotted over time (Figure 2). In the bottle, 66% of the total Fe was Fe(II), and this percentage declined to and stabilized at 36%. Evidently, Fe(II) does not become fully oxidized but attains an equilibrium concentration after some days, when the wine is being oxidized at a maximum rate when saturated with air. The relatively lower initial Fe(II) content than in wine from bottles fitted with screw caps (95%, see below) was consistent with the unexpectedly high initial O2 content.
The oxygen ingress rates are lower for bottles closed with screw caps than in those with traditional natural corks (Lopes et al. 2006), and the effect of a screw cap on the Fe(II):Fe(III) ratio is illustrated by the results obtained with an Argentinian Pinot Grigio (Figure 3). The initial O2 concentration in this wine was nil and the total Fe concentration was 2.07 mg/L. The Fe(II) concentrations at t0 and t10 at the first measurement (T = 0) were 1.97 and 1.99 mg/L, respectively. The differences were small, as Fe(III) concentrations were very low and little polyphenol oxidation could occur in this wine. The differences were greater when more Fe(III) was available after 70.5 hr, and Fe(II) concentrations at t0, t1, and t10 were 1.18, 1.20, and 1.33 mg/L, respectively. Fe was 95.3% as Fe(II) in the bottle, and the Fe(II) levels stabilized at 57% after 95.75 hr (Figure 2). The curve for the last measurement point (T = 95.75 hr) is not shown because it overlapped with that for 70.5 hr (Figure 3). In an American Pinot Grigio wine, which was also closed with a screw cap and hence assumed to contain no O2, the Fe was found similarly to be 95% as Fe(II) in the bottle, but this percentage decreased to <39% by 87.8 hr after oxygen saturation, much lower than in the Argentinian wine (Figure 2).
The molar absorptivity of the Fe(II)-ferrozine complex increased slightly as the pH was decreased from 3.2 to 2.75, and a small correction can be made for more acidic wines. For example, with the pH 2.75 calibration curve for the Argentinian Pinot Grigio (pH 2.78), the total Fe concentration was 2.04 mg/L. The Fe(II) concentration at T = 0 hr was 1.94 mg/L and 1.17 mg/L at T = 95.75 hr, giving 95.3% and 57.3% as Fe(II), respectively. With the pH 3.2 calibration curve, the total Fe concentration was found to be 2.07 mg/L. The Fe(II) content at T = 0 hr was unchanged at 95.3%, but that at T = 95.75 hr was observed to be very slightly reduced to 57.0%.
A South African Sauvignon blanc wine from a wine box was also examined. The initial O2 concentration in this wine was 0.29 ± 0.08 mg/L, with 64.7% of the Fe being present as Fe(II). On air saturation, the Fe(II) concentration dropped to 26.8% after 47.5 hr. (Figure 2). When the box was then left untouched for 3 mos after the withdrawal of the initial wine sample, the O2 concentration had increased to 0.74 ± 0.08 mg/L, and the Fe(II) content decreased to 47.3%, which, on oxygen saturation reduced further to 28.1% after 54 hr (Figure 2). An Italian Soave wine also from a wine box that had been similarly opened and then had been left standing for 2.5 mos was found to contain 1.24 ± 0.12 mg/L O2 with 53.7% Fe(II). After saturation with air, Fe(II) percentage dropped to 39.6% after 76 hr (Figure 2).
Of the 10 white wines investigated in this study, four contained ascorbic acid, which was immediately obvious when the wines were treated with ferrozine, with the Fe(II) content rising rapidly as Fe(III) was reduced. This complication was first encountered with a Hungarian Pinot Grigio wine from a wine box, which contained 0.2 ± 0.07 mg/L O2. Addition of ferrozine gave an initial Fe(II) concentration of 1.27 mg/L at t1, which increased rapidly to 1.47 ± 0.03 mg/L at t10 (derived from absorbance in Figure 4), close to the value of 1.5 mg/L for the total Fe concentration found by AAS. Clearly, the wine contained a highly reductive substance, and titration with 2,6-dichlorophenolindophenol showed that the wine contained 54 mg/L ascorbic acid. A precise extrapolation to t0 was not possible, but it was apparent that Fe(II) had undergone substantial oxidation at these low O2 concentrations; even in the presence of ascorbic acid, the Fe(II) content in the wine box had decreased to at least 70%. With an Italian Pinot Grigio wine from a screw-cap bottle containing 124 mg/L ascorbic acid, ferrozine addition resulted in the total reduction of Fe(III) within 1 min.
TPTZ is a Fe(II)-selective tridentate ligand, which, like ferrozine, increases the oxidant power of Fe(III) and is used in the FRAP assay to measure the antioxidant capacity of polyphenols (Prior et al. 2005). Interestingly, the Fe(III)/TPTZ system appeared to be a stronger oxidant than the corresponding ferrozine system because all the Fe(III) was reduced in the Hungarian Pinot Grigio wine within 1 min (Figure 4). The absorbance values at t0 and t10 corresponded to 1.5 ± 0.0 mg/L Fe(II), similar to the value obtained with ferrozine for total Fe content attained after ~6 min. The absorbance was lower than that obtained with ferrozine because of the lower molar absorptivity of the Fe(II)-TPTZ complex.
The effect of Cu and catechols on the rate of Fe(II) oxidation was also briefly examined in the model wine (Figure 5). An air-saturated model wine containing caffeic acid (44 mg/L) and (+)-catechin (78 mg/L), concentrations that simulated a white wine, and containing 52 mg/L free SO2, oxidized slowly in the presence of Fe alone. Cu addition accelerated oxidation and quickly established a lower Fe(II):Fe(III) ratio in the model wine. Increasing the concentration of (+)-catechin to 2.0 g/L to simulate the polyphenol content of red wine accelerated Fe(III) reduction and prevented the Fe(II) concentration from dropping below ~55%. This result illustrated the effect of altering the relative rates of Fe(II) oxidation and Fe(III) reduction.
Conclusion
This study supports the proposal that the Fe(II)/Fe(III) couple redox-cycles as wine is oxidized. The Fe(II):Fe(III) concentration ratio was observed to depend on how a wine was stored, and the ratio changed when wine was exposed to oxygen. The Fe(II) concentration fell as wines were oxidized at air saturation, but the level at which the Fe(II):Fe(III) ratio stabilized differed for the different wines. Under these conditions, wine should be oxidizing at a maximum rate, and this finding implies that the relative rates of Fe(II) oxidation and Fe(III) reduction differ among different wines. This difference should be due to wine compositional differences, which could be well worth investigating, as it may provide a deeper understanding of the oxidative process in wines. The simple procedure reported here may be useful for determining the redox status of wines and, for instance, provide an additional way to examine the efficiencies of wine bottle closures and of different storage systems.
- Received September 2015.
- Revision received November 2015.
- Accepted November 2015.
- Published online December 1969
- ©2016 by the American Society for Enology and Viticulture