Wine Reduction Potentials: Are These Measured Values Really

Apr 5, 2019 - ... University of Southampton , Southampton SO17 1BJ , United Kingdom ... of Auckland , Private Bag 92019, Auckland 1142 , New Zealand...
2 downloads 0 Views 1MB Size
Download PDF
Review Cite This: J. Agric. Food Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/JAFC

Wine Reduction Potentials: Are These Measured Values Really Reduction Potentials? John C Danilewicz,*,† Peter Tunbridge,‡ and Paul A. Kilmartin§ †

Private laboratory, 44 Sandwich Road, Ash, Canterbury, Kent CT3 2AF, United Kingdom Faculty of Natural and Environmental Sciences, University of Southampton, Southampton SO17 1BJ, United Kingdom § School of Chemical Sciences, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand

Downloaded via BOSTON UNIV on April 8, 2019 at 03:35:22 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

ABSTRACT: During its production wine can react with substantial amounts of aerial oxygen. Some oxidation can be beneficial, especially in red wine, but if allowed to occur in excess it is highly detrimental, making oxygen management an important aspect of wine making. The use of reduction potentials at platinum electrodes to measure the redox state of wines extends back over 80 years. The premise is that reductants in wine produce oxidized derivatives and the balance between the two determines the reduction potential, as in classical electrochemistry. As the detailed mechanism of wine oxidation becomes better understood, it is apparent that redox couples in wine do not function in this way. It is proposed that the observed potentials are mixed potentials largely due to ethanol oxidation coupled with oxygen reduction. Under low oxygen conditions, further redox couples can contribute to the mixed potential, both directly and via adsorption effects at the platinum electrode. KEYWORDS: redox potential, wine, polyphenols, ethanol, oxygen, iron, thiols

■

INTRODUCTION Wine reacts readily with oxygen, red more so than white due to the greater concentration of polyphenols, which are the main initial reductants. Some oxygen exposure of red wine improves wine quality, such as by lowering bitterness and astringency and removing vegetative character, but overexposure can be highly damaging. On the other hand, white wines are best protected from aerial oxidation to preserve their fresh fruity character.1,2 However, wines are repeatedly exposed to aerial oxygen during production and to a varying extent during storage and postbottling.3 The control of oxygen exposure is therefore an extremely important aspect of wine production. The measurement of wine reduction potentials has a long history in enology, and careful attention has been given to the method of analysis and electrode preparation, given issues with obtaining a stable and steady value at platinum electrodes.4,5 The drift in potential values was ascribed to poisoning of the electrode surface during exposure to wine, meaning that a consistent chemical pretreatment was recommended prior to taking the measurement, and the measurement time could take a number of hours before the potential changed by less than 0.2 mV/min. In repeat measurements with separately prepared platinum electrodes, an average difference of 5.5 mV was obtained for 21 wines after 3−4 h of stabilization, while with platinized platinum (black) electrodes, the average difference was 2.2 mV; however the values were 150−200 mV lower than those obtained with unplatinized electrodes, as used in all other studies.6 Wine reduction potentials have been measured on the understanding that they are determined by the relative concentration of oxidized and reduced components of redox couples,7,8 which in wine would be determined by the extent of oxygen exposure. Values obtained have, therefore, been considered to be a measure of “redox state”, as in classical electrochemistry. The method has been refined, and there is now an official AOAC international method for the © XXXX American Chemical Society

determination of wine oxidation−reduction potentials (OIVMA-AS2-06).7,9 However, some concerns have been expressed that so-called wine “reduction potentials” are actually mixed potentials generated at the measuring platinum electrodes by redox processes that are quite different from those occurring in the wines themselves.10−13 In recent publications, interesting effects at platinum electrodes in wine have been observed for both oxygen saturation of fermenting musts14 and wines subject to anoxic storage.15 Reduction potentials have also been used to study the effect of oxygen on wine postbottling16 and on the barrel aging of port wine.17 The origin of the trends seen for the “wine redox potentials” observed in these studies are relevant to wine oxidation chemistry. The aim of this review is to examine how the more recent understanding of the reactions that occur during wine oxidation could relate to changes in measured potentials.

■

REDUCTION POTENTIALS Classically, the ability of a substance to accept or donate electrons is determined using reversible oxidation−reduction electrodes. For example the reduction potential of the Fe(III)/ Fe(II) couple may be measured by placing a platinum electrode in a solution of known concentrations of the two and measuring the potential generated against a reference standard hydrogen electrode (SHE), which is taken as having zero potential. The reduction potential of redox couples depends on the relative concentration of the oxidized and reduced components, as shown by the Nernst equation for Fe (eq 1, Figure 1). Received: January 8, 2019 Revised: March 20, 2019 Accepted: March 23, 2019

A

DOI: 10.1021/acs.jafc.9b00127 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Review

Journal of Agricultural and Food Chemistry

Figure 1. Reduction potential of the Fe(III)/Fe(II) redox couple.

EH is the measured potential against a SHE, or more conveniently against a Ag/AgCl or calomel electrode, with the necessary correction (+207 mV for Ag/AgCl in 3 M NaCl and +241 mV for a saturated calomel electrode at 25 °C). All potentials quoted in this review are with reference to SHE unless otherwise stated. E0 is the formal potential of the couple and results when [Fe(III)] = [Fe(II)] in the electrolyte solution under study at pH 0. At low concentration, when the activity coefficients approach unity, E0 approximates to E°, the standard reduction potential, when all components of the system are at unit activity (including pH = 0). R is the gas constant, T is the absolute temperature, F is the Faraday constant, and n is the number of electrons transferred. At 25 °C and when n = 1, the equation can be more conveniently written as eq 2 (Figure 1). EH increases as the relative concentration of Fe(III) increases and falls as it is lowered. EH is, therefore, determined by the redox state of the couple. The formal reduction potential (E0) is determined when the concentrations of the reduced and oxidized components of a couple are equal. Importantly, the system needs to be at equilibrium to be able to measure the potential of redox species in solution directly at an inert electrode. The comparison of reduction potentials allows the ability of a reduced component of a redox couple to transfer an electron to the oxidized component of another couple to be determined. The basic premise, therefore, is that the so-called reduction potential of wine is a measure of its redox state; that is the balance between reduced and oxidized components. The reduction potential of a mixture will be achieved when all the redox couples attain the same potential and are at equilibrium. If a couple is at a higher potential than others, and a pathway exists for electron transfer, a reaction will occur and the overall measured potential will change until equilibrium is achieved. Considerable progress has been made in recent years in understanding wine oxidation, and it is now possible to consider what the reduced and oxidized components might be and how their concentration may be affected by oxygen.

Figure 2. Quinhydrone redox couple.

pH (pH 3.5), the reduction potential (E3.5) of the couple would be ∼497 mV. The 1,4-quinone is stable enough to measure its electrochemical properties up to pH ≈ 8. The catechols of wine behave similarly, but the resulting 1,2quinones are much less stable. Being strong electrophiles, they react rapidly with nucleophiles such as sulfite (HSO3−) and thiols that would be invariably present in wine.19 The formal reduction potential of the 1,2-benzoquinone/benzene-1,2-diol couple has been determined as E0 = 793 mV by titrating the catechol against Ce(SO4)2 in acetate buffer. However, to allow for the instability of the quinone, the experiment was designed so as to obtain the reduction potential at the moment of reagent addition.20 The reduction potential at pH 3.5 (E3.5) would be ∼587 mV. The oxidation of polyphenols is catalyzed by iron, assisted by copper, where the Fe(III)/Fe(II) redox couple redox cycles as oxygen is reduced and the polyphenol oxidized (Figure 3).21

Figure 3. Redox cycling of Fe as oxygen oxidizes a catechol with involvement of sulfite.

The reduction potential of the quinone/catechol couple for (+)-catechin was found to be ∼577 mV at pH 3.6 by cyclic voltammetry,22 close to the above value obtained for catechol itself by potentiometric titration. However, the reduction potential of the Fe(III)/Fe(II) couple is only ∼350 mV under wine conditions (see below), lower than that of the catechol system. Consequently, the oxidation of catechols by Fe(III) is thermodynamically disfavored and should not occur. Indeed, when the aerial oxidation of (+)-catechin was attempted in model wine, the oxygen uptake increased as Fe(II) and Cu(II) concentrations were raised (Figure 4, curves a, b, and c).23 However, only enough O2 was taken up to oxidize ∼3/4 of the Fe(II). It is evident that under these conditions the Fe(III) produced is not reduced by the catechol, as the resulting Fe(II) would react with oxygen and oxygen uptake would continue. Accordingly, when Fe(III) was added directly, no reaction was observed (curve d). Evidently, some additional assistance is necessary to allow Fe to redox cycle and for (+)-catechin oxidation to proceed. In contrast, when sulfite was added, a steady oxygen uptake was obtained (Figure 5, curve a).23 When Fe(II) was added at higher concentration an initial faster rate was apparent, consistent with initial Fe(II) oxidation (curve b), but when the Fe couple

■

POLYPHENOL OXIDATION Polyphenols include the cinnamic acid derivatives, such as caftaric acid, and flavanols and derivatives, which include oligomers, the proanthocyanidins, and polymers, the condensed tannins, in red wine. These substances, which contain pyrogallol and catechol groups, are the most abundant initially oxidizable wine constituents; ethanol and added antioxidants are further oxidizable compounds present in wine. When the redox processes also involve the transfer of protons, reduction potentials will be pH-dependent. This involvement is illustrated by the quinhydrone electrode, in which 1,4benzoquinone (Q) is reduced to 1,4-benzenediol (QH2) (Figure 2). The reduction potential of the system will follow the Nernst equation (eq 3), which simplifies to eq 4, where E0 is 704 mV.18 For each unit increase in pH, the reduction potential is lowered by 59 mV. With the quinhydrone electrode, where [Q] = [QH2], the reduction potential is dependent on pH at a given temperature. Therefore, at wine B

DOI: 10.1021/acs.jafc.9b00127 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Review

Journal of Agricultural and Food Chemistry

Pyrogallol, which is a stronger reductant than catechols and which can serve as a model for gallocatechins, is oxidized alone in model wines, but its oxidation is also accelerated by sulfite.23 It would appear, therefore, that quinones are unstable in wine conditions and can only have transient existence. Their presence is seen over a period of seconds to minutes in voltammetric scans taken at inert electrodes, where polyphenols are first oxidized, and indicated by a return peak seen for some “reversible” polyphenols, for example, catechin.22 As far as reduction potentials are concerned these results indicate that the quinone component of the quinone/catechol couple is largely absent in wine and polyphenols cannot constitute a classical “reduction potential”. In support of this conclusion is the observation that there is little difference in reduction potentials when SO2 is added to wine,7 which would ensure the absence of any quinones. An increase in potential of just 19 mV was observed following a large 750 mg L−1 SO2 addition to sweet white wine.7 On addition of 50 mg L−1 of SO2 to a red wine, a difference of not more than 10 mV was observed, even as the O2 concentration was increased from near zero to 7 mg L−1 and the potential increased from 200 to 520 mV.27 In this same study, red and white wines to which 100 mg L−1 ascorbic acid had been added, showed potentials around 100 mV lower than control wines when subject to elevated O2 concentrations, an observation that will be discussed further below. Similar responses to changes in O2 concentration are observed in white and red wines and importantly in model wines, which do not contain polyphenols.7,11 It is also reported that reduction potentials only change slightly when polyphenols are removed with charcoal from both white and red wines.7 It is evident that polyphenols and their oxidation products are not present at equilibrium in wine and so cannot constitute their own reduction potentials. It can also be noted that as the detailed mechanism of wine oxidation is becoming better understood, the naming of polyphenols as antioxidants is giving a misleading impression of their actual activity. Polyphenols are the initial reductants, which by interacting with iron play a major role in activating oxygen. They therefore initiate the oxidation cascade, which would lead to wine spoilage if not inhibited by sulfite by removing H2O2 and quinones. They do not prevent oxidation as the term antioxidant could imply. Perhaps, with regard to these substances in wine, antioxidant assays should be renamed “reductant capacity assays”. Oxygen Reduction. In wine, oxygen is first reduced to H2O2, which is proposed to occur in two concerted single electron steps by two Fe(II) (Figure 6). This mechanism was first proposed to explain the kinetics of oxidation of the Fe(II)−ethylenediaminetetracetic acid (EDTA) complex.28 However, in wine the oxidation of Fe(II) was found to slow markedly before completion, and it was shown that it was

Figure 4. Oxidation of (+)-catechin (524 mg L−1) in model wine in the absence of sulfite. (a, b, c) Effect of increased Fe(II) and Cu(II) concentration; (d) attempted oxidation of (+)-catechin (524 mg L−1) when Fe(II) is replaced by Fe(III). Reproduced from ref 23. Copyright 2011 American Society for Enology and Viticulture.

Figure 5. Oxidation of (+)-catechin (524 mg L−1) in the presence of Fe, Cu, and sulfite. (a, b) Samples sealed immediately after metal addition without any air headspace to measure O2 uptake. (c) Samples sealed after the added Fe(II) was oxidized overnight in the presence of air and system was resaturated with air. Reproduced from ref 23. Copyright 2011 American Society for Enology and Viticulture.

was allowed to equilibrate first and the model wine was resaturated with air before measurements were taken, a smooth O2 uptake (curve c) was obtained, showing pseudo-first-order kinetics with respect to [O2]. Presumably, in curve a, the initial rate of Fe(II) oxidation was sufficiently slow that no overall change in rate was apparent as the Fe couple equilibrates at the onset. Sulfite cannot react with oxygen in wine conditions as originally thought.24 Its main function is to react with quinones and with H2O2 (Figure 3). In further studies, it was found that when (+)-catechin is oxidized in model wine in the presence of sulfite, essentially all the quinone is reduced back to the catechol and therefore its concentration does not change.25 It appears, therefore, that oxidation of the catechol is only possible when the reaction is coupled with a thermodynamically favorable process in which the quinone is removed. Other polyphenols may also undergo 1,4-conjugate addition as is observed with 4-methylcatechol, used as a model catechol.26 Other nucleophiles such as azide and benzenesulfinic acid similarly allow (+)-catechin to oxidize.23 In the latter case, the quinone is seen to react by forming the benzenesulfonic acid 1,4-adduct quantitatively.25

Figure 6. Reduction of oxygen by Fe(II) and inhibitory action of Fe(III). Reproduced from ref 29. Copyright 2013 American Society for Enology and Viticulture. C

DOI: 10.1021/acs.jafc.9b00127 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Review

Journal of Agricultural and Food Chemistry inhibited by Fe(III),29 which would, therefore, be increasingly inhibitory as the reaction progressed. It was proposed that Fe(III) oxidized the Fe(III)−superoxo complex (Figure 6), and it follows that the Fe(III)/Fe(II) ratio determines the rate of Fe(II) oxidation. The H2O2 may then be rapidly further reduced by Fe(II) in the Fenton reaction to produce hydroxyl radicals (HO•), which are powerful oxidants, which oxidize mainly ethanol to acetaldehyde, since it is the most abundant substrate. Alongside this chemical process, recent studies have drawn attention to the role of viable yeasts in the production of acetaldehyde during wine oxidation.30 In the presence of sulfite, H2O2 is rapidly reduced to water with the formation of sulfate, so preventing ethanol oxidation. The overall reduction of oxygen to water may, therefore, be represented by the following half reaction: O2 + 4e− + 4H+ = 2H2O. It can be noted that in wine the reaction is not reversible; oxygen and water are not at equilibrium, which is a fundamental requirement if the O2/H2O couple is to constitute a reduction potential. The intermediate reduction of O2 to H2O2 is also irreversible, as H2O2 is rapidly reduced to water by either Fe(II) or sulfite. It was found that the standard potential of the O2/H2O redox couple cannot be reliably measured with a platinum electrode, presumably due to low electroactivity of the species involved, the formation of metal-oxide films, and the contribution of intermediates such as H2O2,31 and so has to be calculated based upon other experimental results.32,33 The change in Gibb’s free energy (ΔG°) was obtained from the reaction enthalpy and entropy data for the reaction of O2 with H2. Since ΔG° = −nFE°, the standard reduction potential (E°) for the O2/H2O couple was then calculated to be 1.229 V at 25 °C, where n is the number of electrons transferred and F is the Faraday constant. This value was derived with the standard state of O2 taken as 105 Pa. The value increases to 1.27 V for 1 mol L−1.33 The reduction potential of an oxygen half-cell versus SHE can be obtained from eq 5, which simplifies to eq 6, taking water as having unit activity. E H = E° + (RT /nF )ln[O2 ][H+]4 /[H 2O]2 V

(5)

E H = 1270 − 59 pH + 14.8 log[O2 ] mV

(6)

E H = 1264 − 59.9 pH + 150.9 log[O2 ] mV

(7)

Figure 7. Typical effect of O2 concentration on wine reduction potential7 and comparison with the Nernst relationship.

presence of either ethanol or tartaric acid, the components added to make up a model wine. Oxygen reacts slowly in wine, one O2 saturation taking some days to react in red wine and some weeks in white wine. In wine, the so-called “reduction potential” is very dependent on O2 concentration as shown in Figure 7. When O2 is added to a wine, there is an immediate increase in the measured potential,14 but the redox status of the wine takes some time to change. The measured potential will, therefore, be dependent on O2 concentration at the moment of measurement and has no bearing on the existing redox status, which will then change over time as oxygen reacts. The Fe(III)/Fe(II) Couple. Jean Ribéreau-Gayon (1931) noted in his doctoral thesis that Fe exists as Fe(II) in wine when protected from air but when O2 is introduced a stronger oxidant was produced, which he proposed was Fe(III).34 The proposal, therefore, was that Fe redox cycled as shown in Figure 3. The Fe(III)/Fe(II) ratio should depend on the relative rates of Fe(II) oxidation and Fe(III) reduction and so be determined by “redox status”. It would be expected that the wine Fe(III)/Fe(II) ratio would be at a minimum in the absence of oxygen and at a maximum at O2 saturation after equilibration. The formal reduction potential of the Fe(III)/Fe(II) couple is often quoted as 770 mV, but this applies to pH 0, where the metals are as hexahydrates with the water fully protonated. The reduction potential will change depending on the relative selectivity of ligands for the two redox states and extent of ligand ionization. 2,2′-Bipyridyl coordinates preferentially with Fe(II) and so, by stabilizing Fe(II) relative to Fe(III), makes Fe(III) a strong oxidant, E3.5 = 1.03 V (Figure 8). In the presence of this ligand Fe(II) can no longer reduce O2. In contrast EDTA is Fe(III) selective and by stabilizing Fe(III) renders Fe(II) highly reducing and capable of rapidly reducing O2. Tartaric acid is also Fe(III) selective, but since it binds less strongly to Fe(III) than EDTA, it only lowers the reduction potential of the Fe couple down to 350 mV.35 However, this is still sufficient to allow Fe(II) to rapidly react with O2 under wine conditions.36 The reduction potential of the Fe-couple may, therefore, be determined in wine conditions by knowing the Fe(III)/Fe(II) concentration ratio and using the Nernst equation (Figure 8). A method has been developed using ferrozine to measure this ratio in white wine. The procedure involves minimal wine

Consequently, when O2 concentration is increased from 0.1 to 7 mg L−1 (approximately wine saturation), it would be expected that EH would increase by ∼27 mV (Figure 7). However, using an improved platinum electrode, which was less prone to poisoning and equilibrated more readily than conventional electrodes, it was shown that this was not the case in model wine or wine. A ∼350 mV increase in reduction potential was observed for the above increase in O 2 concentration (Figure 7).7 Equation 7 was proposed to explain the result. A 10-fold increase in the coefficient of the log[O2] term was proposed, but without any thermodynamic justification.7 In a study of the barrel aging of a Port wine a plot of EH against log[O2] gave a slope of 177.7 and 143.3 at ∼11 and 21 °C respectively.17 These two values are of the same order as that of 150.9 obtained by Vivas et al. shown in eq 7.7 Evidently, the O2/H2O couple is not involved as this would require a value of 14.8 as shown in eq 6. It is apparent, from this major divergence from the Nernst relationship, that the process occurring on the platinum electrode is not the same as in a simple aqueous solution and is due to the D

DOI: 10.1021/acs.jafc.9b00127 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Review

Journal of Agricultural and Food Chemistry

It has been recently reported that when 15 red wines were subjected to accelerated anoxic aging by being stored under strict anoxic conditions at 50 °C, the reduction potential dropped slowly from ∼222 mV to equilibrate at ∼197 mV after ∼12 days.15 It is possible that this small lowering in potential could be due to the release of thiols from oxidized precursors under anoxic conditions.38 The slow reduction in Fe(III), which is likely to be present at the onset, could be another cause. This latter possibility is suggested by the finding that a classic sigmoidal curve was obtained when EH was plotted against the [Fe(III)]/[Fe(II)] ratio in model wine under anoxic conditions. The potential increased from ∼310 to ∼450 mV as the ratio increased. The midpoint potential was 377 mV (E3.5) close to the value of 350 mV derived by cyclic voltammetry quoted above. Also, the tangent at the midpoint indicated a one-electron process. It could be that the Fe-system is producing real reduction potentials in anoxic conditions and that oxidation of ethanol, linked to proton reduction, has only a minor effect. However, it may also be due to a mixed potential, where Fe(III) reduction is coupled to ethanol oxidation. Thiols. There has been considerable uncertainty concerning the standard reduction potential of glutathione, cysteine, and other thiols, as measured by electrochemical methods, as these substances bind to metal electrodes causing drifting of potentials and poor reproducibility, known as poisoning. Other approaches have, therefore, been explored to obtain these potentials. Thermodynamic analysis has given a value of ΔG° = −1510 cal M−1 for the reduction of cystine to 2 cysteines (CSSC/2CSH), where ΔG° is the change in free energy under standard conditions. Since ΔG° = −nFE° under standard conditions (pH 0), as described above for the O2/ H2O couple, E° was calculated as +32 mV for the CSSC/ 2CSH couple. On the basis that the reduction potential is lowered by ∼59 mV for each unit increase in pH, since one H+ is added per electron transferred, as for quinhydrone, E3 was given as −157 mV, that is at pH 3, and E°′ = −390 mV, where E°′ is defined as the “biochemical standard reduction potential” and is determined at pH 7 versus SHE.39 However, a somewhat higher value for the reduction potential of the glutathione disulfide/glutathione couple (GSSG/2GSH) was derived by determination of the equilibrium constant (Kequ) for the glutathione reductase catalyzed reduction of the GSSG by NADPH at pH 6.8 (Figure 10).40 From the relationship nF log Kequ = ΔE, where ΔE is the difference in reduction potential between the two redox systems, and the E°′ for the NADP+/NADPH couple being established as −318 mV, E°′ for the glutathione couple was found to be −250 mV. A similar value (−230 mV) was obtained for the cysteine couple by equilibrating GSH with

Figure 8. Ligand dependence of Fe(III)/Fe(II) reduction potential.

disturbance and compensates for any change in ratio during measurement.21 In a Pinot grigio wine, from a bottle closed with a screwcap, the Fe was 95% as Fe(II). In a Chardonnay wine, from a bottle closed with a natural cork, Fe was 66% as Fe(II). In a Sauvignon blanc in a wine box Fe(II) content dropped to 47%. This reflected increased O2 content. No measurable amounts of O2 were found in the bottle closed with a screwcap; however, the Chardonnay and Sauvignon blanc wines contained 0.63 and 0.74 mg·L−1 O2 respectively. The three wines were saturated with air, and the Fe(II) content equilibrated over several days to 57%, 36%, and 27% for the Pinot grigio, Chardonnay, and Sauvignon wines, respectively (Figure 9). These values correspond to calculated reduction

Figure 9. Change in % Fe(II) concentration in air saturated white wines.

potentials of 340, 364, and 375 mV, respectively. These wines, being saturated with air, would have similar measured so-called “reduction potentials” with values over 500 mV (Figure 7), which does not match the values calculated from the actual Fe(III)/Fe(II) ratios. The above procedure has been modified to allow Fe(II)/ Fe(III) ratios to be determined in red wines.37 A Cabernet Sauvignon wine closed with a screwcap had 97.3% Fe(II). Shiraz wines from three wine boxes contained between 84.0 to 90.0% Fe(II) initially, which corresponds to reduction potentials of 302 to 287 mV. These wines contained no measurable amounts of O2 (