Читать книгу Handbook of Enology, Volume 2 - Pascal Ribéreau-Gayon - Страница 25
1.4.3 The “Buffer” Capacity of Musts and Wines
ОглавлениеThe acid–base buffer capacity of wines is largely responsible for their physicochemical and microbiological stability, as well as their flavor balance. For example, the length of time a wine leaves a fresh impression on the palate is directly related to the formation of acid salts by basic proteins in saliva, i.e. the expression of the buffer phenomenon and its capacity. In contrast a wine that tastes “flat” has a low buffer capacity, but this does not necessarily mean that it has a low acidity level. At a given total acidity level, buffer capacity varies according to the composition and type of acids present. This point will be developed later in this chapter.
In a particular year, a must's total acidity and acid composition depend mainly on geography, soil conditions, and climate, including soil humidity and permeability, as well as rainfall patterns, and, above all, temperature. Temperature determines the respiration rate, i.e. the combustion of tartaric and, especially, malic acids in grape flesh cells. The predominance of malic acid in must from cool‐climate vineyards is directly related to temperature, while malic acid is eliminated from grapes in hotter regions by combustion and is thus found in much lower amounts than tartaric acid.
Independently of climate, grape growers and winemakers have some control over total acidity and even the acid composition of the grape juice during ripening. Leaf thinning and shoot trimming restrict acid biosynthesis and, above all, combustion by reducing the greenhouse effect of the leaf canopy. Another way of controlling total acidity levels is by choosing the harvesting date. Grapes intended for sparkling wines must be picked at the correct level of technological ripeness to produce must with a total acidity of 9–10 g/l as H2SO4. This acidity level is necessary to maintain the wines' freshness and, especially, to minimize color leaching from the red grape varieties, Pinot Noir and Pinot Meunier, used in Champagne. At this stage in the ripening process, the grape skins are much less fragile than they are when completely ripe. The last method for controlling the total acidity of must is by taking great care in pressing the grapes and keeping the juice from each pressing separate (Volume 1, Section 14.3.2). In the Champagne region, the cuvée corresponds to juice from the mid‐part of the flesh (furthest from the skin and seeds), where it has the highest sugar and acidity levels.
Once the grapes have been pressed, winemakers have other means of raising or lowering the acidity of a must or wine. It may be necessary to acidify “flat” white wines by adding tartaric acid after malolactic fermentation in years when the grapes have a high malic acid content. This is mainly the case in cool‐climate vineyards, where the malic acid is not consumed during ripening. The disadvantage is that it causes an imbalance in the remaining total acidity, which then consists exclusively of a diacid, tartaric acid, and its monopotassium salt.
One method that is little‐known, or at least rarely used to avoid this total acidity imbalance, consists in partially or completely eliminating the malic acid by chemical means using a mixture of calcium tartrate and calcium carbonate. This method precipitates the double salt, calcium tartromalate (Section 1.4.4, Figure 1.9), and is a very flexible process. When the malic acid is partially eliminated, the wine has a buffer capacity based on those of both tartaric and malic acids, and not just on that of the former. Tartrate buffer capacity is less stable over time, as it decreases due to the precipitation of monopotassium and calcium salts during aging, whereas the malic acid salts are much more soluble.
Another advantage of partial elimination of malic acid over malolactic fermentation, followed by the addition of tartrate, is that, due to the low acidification rate, it does not produce wines with too low a pH. Low pH can be responsible for difficult or stuck secondary fermentation in the bottle during sparkling winemaking via the traditional method (méthode champenoise), leaving residual sugar in the wine.
Traditional acidification and deacidification methods are aimed solely at changing total acidity levels, with no concern for the impact on pH and even less for the buffer capacity of the wine and with all the unfortunate consequences this may have on flavor and aging potential.
This is certainly due to the lack of awareness of the importance of the acid–base buffer capacity in winemaking. Changes in the acid–base characteristics of a wine require knowledge of not only its total acidity and pH but also of its buffer capacity. These three parameters may be measured using a pH meter. Few articles in the literature deal with the buffer capacity of wine (Genevois and Ribéreau‐Gayon, 1935; Vergnes, 1940; Hochli, 1997; Dartiguenave et al., 2000a). This lack of knowledge is probably related to the fact that buffer capacity cannot be measured directly but rather requires readings of four or five points on a neutralization curve (Figure 1.3), and this is not one of the regular analyses carried out by winemakers.
It is now possible to automate the plotting of a neutralization curve, based on the wine's initial pH and total acidity, and thus measuring buffer capacity at the main stages in winemaking should become a routine.
Mathematically and geometrically, buffer capacity, β, and buffer range are deduced from the Henderson–Hasselbalch equation (Section 1.4.2, Equation (1.2)). Buffer capacity is defined by Equation (1.3):
where ΔB is the number of strong base equivalents that cause an increase in pH equal to ΔpH. Buffer range is a way of assessing buffer capacity. For an organic acid alone, with its salt in solution, it may be defined as the pH interval in which the buffer effect is optimum (Equation (1.4)):
Buffer capacity is normally defined in relation to a strong base, but it could clearly be defined in the same way in relation to a strong acid. In this case, the pH = f (strong acid) function decreases, and its β differential is negative, i.e.:
Strictly speaking, buffer capacity is obtained from the differential of the Henderson–Hasselbalch expression, i.e. from the following derived formula:
as only the Napierian logarithm is geometrically significant and provides access to the slope of the titration curve around its pKa (Figure 1.4).
The differential of the equation is as follows:
FIGURE 1.4 Determining the buffer capacity β from the titration curves of two model buffer solutions.
Making the assumption that the quantity of strong base added, d[B], generates the same variation in acidity in salt form, d[A−], and leads to an equal decrease in free acidity d[HA], per unit, i.e.:
the differential equation for pH is then
or,
Dividing both sides of the equation by d[B] gives the inverse of Equation (1.3), defining the buffer capacity. Theoretically, variations ΔB and ΔpH must be infinitely small, as the value of the ΔB/ΔpH ratio at a fixed pH corresponds geometrically to the tangent on each point on the titration curve (Figure 1.4). More practically, buffer capacity can be defined as the number of strong base equivalents required to cause an increase in pH of 1 unit per liter of must or wine. It is even more practical to calculate smaller pH variations in much smaller samples (e.g. 30 ml). Figure 1.4 clearly shows the difference in buffer capacity of a model solution between pH 3 and 4, as well as between pH 4 and 5.
This raises the issue of the pH and pKa at which buffer range should be assessed. Champagnol (1986) suggested that pH should be taken as the mean of the pKa of the organic acids in the must or wine, i.e. the mean pKa of tartaric and malic acids in must and tartaric and lactic acids in wine that has completed malolactic fermentation.
This convention is justified by its convenience, provided that (Section 1.4.2) the neutralization curves of the must or wine have no inflection points representative of the pKa of the organic acids present, since their buffer ranges overlap, at least partially. In addition to these somewhat theoretical considerations, there are also some more practical issues. An aqueous solution of sodium hydroxide is used to determine the titration curve of a must or wine in order to measure total acidity and buffer capacity. Sodium, rather than potassium, hydroxide used as the sodium salt of tartaric acid is soluble, while potassium bitartrate would be likely to precipitate out during titration. It is, however, questionable to use the same aqueous sodium hydroxide solution for both must and for a dilute alcohol solution like wine.
Strictly speaking, a sodium hydroxide solution in dilute alcohol should be used for wine to avoid modifying the alcohol content and, consequently, the dielectric constant and, thus, the dissociation of the acids in the solution during the assay. It has been demonstrated (Dartiguenave et al., 2000b) that the buffer capacities of organic acids, singly (Tables 1.4 and 1.5) or in binary (Table 1.6) and tertiary (Table 1.7) combinations, are different in water and in 11% dilute alcohol solution. However, if the solvent containing the organic acids and the sodium hydroxide is the same, there is a close linear correlation between the buffer capacity and the acid concentrations (Table 1.4).
TABLE 1.4 Equations for Calculating Buffer Capacity (mEq/l) Depending on the Concentration (mM) of the Organic Acid in Water or Dilute Alcohol Solution (11% vol.) Between 0 and 40 mM (Dartiguenave et al., 2000b)
| Solvent | Water | Dilute alcohol solution |
|---|---|---|
| Tartaric acid | Y = 0.71x + 0.29; R2 = 1 | Y = 0.60x + 1.33; R2 = 1 |
| Malic acid | Y = 0.56x + 0.43; R = 0.998 | Y = 0.47x + 0.33; R2 = 0.987 |
| Succinic acid | Y = 0.56x − 1.38 × 10−2; R2 = 0.993 | Y = 0.53x + 0.52; R2 = 0.995 |
| Citric acid | Y = 0.57x + 0.73; R2 = 1 | Y = 0.51x + 0.62; R2 = 1 |
Table 1.5 shows the values (mEq/l) calculated from the regression lines of the buffer capacities for acid concentrations varying from 1 to 6 g/l in water and in 11% dilute alcohol solution. The buffer capacity of each acid alone in dilute alcohol solution is lower than in water. Furthermore, the buffer capacity of a four‐carbon organic acid varies more as the number of alcohol functions increases (Table 1.8). Thus, the variation in buffer capacity of malic acid, a diacid with one alcohol function, in a dilute alcohol medium, is 1.4 mEq/l higher than that of succinic acid. When the hydroxy acid has two alcohol functions, the increase is as high as 5.3 mEq/l (17.7%) between tartaric and malic acids, even if the buffer capacities of the three acids are lower than in water.
However, the fact that the buffer capacities of binary (Table 1.6) or ternary (Table 1.7) combinations of acids in a dilute alcohol medium are higher than those measured in water is certainly unexpected. This effect is particularly marked when citric acid is included, and it reaches spectacular proportions in a ternary TMC blend (Table 1.7), where the buffer capacity in dilute alcohol solution is 2.3 times higher than that in water.
TABLE 1.5 Buffer Capacity (mEq/l) Depending on the Concentration (g/l) of Organic Acid in Water and in Dilute Alcohol Solution (Dartiguenave et al., 2000b)
| Acid concentration and type of medium | Tartaric acid | Malic acid | Succinic acid | Citric acid | |
|---|---|---|---|---|---|
| 1 g/l | Water | 5.0 | 4.6 | 4.7 | 3.7 |
| Dilute alcohol | 5.3 | 3.8 | 4.0 | 3.5 | |
| 2 g/l | Water | 9.7 | 8.8 | 9.5 | 6.7 |
| Dilute alcohol | 9.3 | 7.3 | 9.4 | 5.9 | |
| 4 g/l | Water | 16.4 | 17.1 | 19.0 | 12.6 |
| Dilute alcohol | 14.9 | 14.3 | 17.5 | 11.3 | |
| 6 g/l | Water | 28.7 | 25.5 | 28.4 | 18.5 |
| Dilute alcohol | 25.3 | 21.3 | 26.4 | 16.6 |
TABLE 1.6 Demonstration of Interactions Between Organic Acids and the Effect of Alcohol on the Buffer Capacity of Binary Combinations (Dartiguenave et al., 2000b)
| Medium | Buffer capacity (mEq/l) | Composition of equimolar mixes of two acids Total acid concentration (40 mM) | ||
|---|---|---|---|---|
| Tartaric acid Malic acid | Tartaric acid Succinic acid | Tartaric acid Citric acid | ||
| Water | Experimental value | 21 | 20 | 23.5 |
| Calculated value | 25.7 | 25.7 | 26.3 | |
| Difference (Calc. − Exp.) | 4.7 | 5.7 | 2.8 | |
| EtOH (11% vol.) | Experimental value | 18.3 | 20.1 | 29 |
| Calculated value | 24 | 23.3 | 24 | |
| Difference (Calc. − Exp.) | 5.7 | 3.2 | −5 | |
| Effect of ethanol | (EtOH − H2O) Exp. | −2.7 | 0.1 | 5.5 |
These findings indicate that the acids interact among themselves and with alcohol, compensating for the decrease in buffer capacity of each individual acid when must (an aqueous solution) is converted into wine (a dilute alcohol solution). From a purely practical standpoint, the use of citric acid to acidify dosage liqueur for bottle‐fermented sparkling wines has the doubly positive effect of enhancing the wine's aging potential while maintaining its freshness on the palate.
Table 1.9 shows the changes in buffer capacity in successive pressings of a single batch of Chardonnay grapes from the 1995 to 1996 vintages, at the main stages in the winemaking process.
The demonstration of the effect of alcohol and interactions among organic acids (Tables 1.6–1.8) led researchers to investigate the precise contribution of each of the three main acids to a wine's buffer capacity in order to determine whether other compounds were involved. The method consisted in completely deacidifying a wine, containing all of its malic acid, by crystallizing and precipitating the double calcium tartromalate salt. After this deacidification, the Champagne base wine had a residual total acidity of only approximately 0.5 g/l as H2SO4, whereas the sample's buffer capacity was still 30% of the original value. This shows that organic acids are not the only compounds involved in buffer capacity, although they represent 90% of total acidity.
TABLE 1.7 Demonstration of Interactions Between Organic Acids and the Effect of Alcohol on the Buffer Capacity of Ternary Combinations (Dartiguenave et al., 2000b)
| Composition of equimolar mixes of three acids (13.3 mM) Total acid concentration (40 mM) | |||
|---|---|---|---|
| Medium | Buffer capacity (mEq/l) | Tartaric acid Malic acid Succinic acid | Tartaric acid Malic acid Citric acid |
| Water | Experimental value | 9.4 | 11.6 |
| Calculated value | 25.4 | 25.5 | |
| Difference (Calc. − Exp.) | 16.0 | 13.9 | |
| EtOH (11% vol.) | Experimental value | 21.7 | 26.4 |
| Calculated value | 22.8 | 23.2 | |
| Difference (Calc. − Exp.) | 1.1 | −3.2 | |
| Effect of ethanol | (EtOH − H2O) Exp. | 12.3 | 14.8 |
TABLE 1.8 Effect of Hydroxyl Groups in the Structure of the Four Carbon Diacids on Buffer Capacity (mEq/l) (Dartiguenave et al., 2000b)
| Medium | 1 Hydroxyl group | 2 Hydroxyl groups | ||||
|---|---|---|---|---|---|---|
| Malic acid | Succinic acid | Δ (Mal. − Suc.) | Tartaric acid | Malic acid | Δ (Tart. − Mal.) | |
| Water | 23.8 | 23.4 | 0.4 | 29 | 23.8 | 5.2 |
| 11% vol. dilute alcohol solution | 22.0 | 20.6 | 1.4 | 25.9 | 22 | 3.9 |
TABLE 1.9 Changes in the Buffer Capacity of Must from Different Pressings of Chardonnay Grapes at Various Stages in the Winemaking Process (Buffer Capacity Is Expressed in mEq/l) (Dartiguenave, 1998)
| Cuvée | Second pressing | |||
|---|---|---|---|---|
| 1995 | 1996 | 1995 | 1996 | |
| Initial value of must | 77.9 | 72.6 | 71.2 | 65.9 |
| After alcoholic fermentation | 60.7 | 63.6 | 57.5 | ND |
| After malolactic fermentation | 51.1 | 60.1 | 48.4 | ND |
| After cold stabilization | 48.1 | 50.3 | ND | 42.4 |
Among the many other compounds in must and wine, amino acids have been singled out for two reasons: (1) in Champagne musts and wines, their total concentration is always over 1 g/l and may even exceed 2 g/l, and (2) their (at least bifunctional) character gives them a double‐buffer effect. They form salts with carboxylic acids via their ammonium group and can become associated with a non‐dissociated acid function of an organic acid via their carboxyl function, which is mostly dissociated at wine pH, thus creating two buffer pairs (Figure 1.5).
An in‐depth study of the interactions between amino acids and tartaric and malic acids focused on alanine, arginine, and proline, present in the highest concentrations in wine, as well as on amino acids with alcohol functions, i.e. serine and threonine (Dartiguenave et al., 2000a).
The findings are presented in Figures 1.6 and 1.7. Hydrophobic amino acids, such as alanine, were found to have only a minor effect, while amino acids with alcohol functions had a significant impact on the buffer capacity of an aqueous tartaric acid solution (40 mM). An increase of 0.6 mEq/l was obtained by adding 6.7 mM alanine, while addition of as little as 1.9 mM of serine produced an increase of 0.7 mEq/l, and addition of 4.1 mM of threonine resulted in a rise of 2.3 mEq/l.
FIGURE 1.5 Diagram of interactions between amino acids and organic acids that result in the buffer effect.
The impact of amino acids with alcohol functions was even more spectacular in dilute alcohol solutions (11% by volume). With only 200 mg/l serine, there was a 1.8 mEq/l increase in buffer capacity compared with only 0.8 mEq/l in water. It was also observed that adding 400 mg/l of each of the five amino acids led to a 10.4 mEq/l (36.8%) increase in the buffer capacity of a dilute alcohol solution containing 40 mM tartaric acid.
It is surprising to note that amino acids have no significant effect on the buffer capacity of a 40 mM malic acid solution (Figure 1.7).
All these observations highlight the role of the alcohol function, both in the solvent and in the amino acids, in interactions with organic acids, particularly tartaric acid with its two alcohol functions.
The lack of interaction between amino acids and malic acid, both in water and dilute alcohol solution, can be interpreted as being due to the fact that malic acid has one alcohol function as compared with the two alcohol functions of tartaric acid. This factor is important for stabilizing interactions between organic acids and amino acids via hydrogen bonds (Figure 1.8).
