Inhibitors of calcium tartrate in wines

Inhibitors of calcium tartrate in wines Maria de Fátima Gonçalves Marques maria.marques@tecnico.ulisboa.pt Instituto Superior Técnico, Lisbon, Portugal December 2014 Abstract The influence of different compounds, in its majority those which are part of the wine composition, on the precipitation of the calcium tartrate (CaT) was evaluated in a hydroalcoholic standard solution, at 15 C. Organic and phenolic s and amino s were among the compounds tested, as well as other compounds (the polysaccharide fractions) isolated from the red wine by ultrafiltration. The equipment used to conduct the ultrafiltration experiments was the LabUnit M20 (from DSS). These assays were performed with the FS40PP and FS61PP membranes, under the concentration mode, at a pressure of 3 bar. Polysaccharides of higher molecular weight were obtained in the concentrates, at different volumetric concentration factors, with the FS40PP membrane (100 kda). When the FS61PP membrane (20 kda) was used, polysaccharides of lower molecular weight were obtained in permeate. The induction periods associated with the calcium tartrate precipitation were either evaluated through the visual appearance of the solution and also with the use of a calcium ion-selective electrode. The citric and malic s have demonstrated to be responsible for a higher delay in the induction time. When added at the same molar concentration (0.02 M), these s have shown the following inhibition strength: malic > citric. Gluconic has demonstrated to be an efficient inhibitor of the calcium tartrate precipitation. Metatartaric has also demonstrated to prevent the occurrence of precipitation, such as it would be expected. It was found that the polysaccharides isolated from the fractions obtained by ultrafiltration did not significantly influence the induction times. The most important influence has occurred when the polysaccharides precipitates with volumetric concentration factors (VCF) of 4 and 4.5 were added. Key Words: Calcium Tartrate; Precipitation; Red Wine; Ultrafiltration; Inhibitors. 1. Introduction Tartaric (H 2T) is an important organic of the wine composition. Its concentrations vary between 1.5 to 4 g/l (Curvelo-Garcia, 1988). Tartaric is a relatively strong, responsible for wine ph values between 3-3.5. When present in a solution it coexists in equilibrium in three different forms, which are: H 2T, HT - and T 2- (Lasanta e Gómez, 2012). The proportion of each tartaric ion depends on the ph. At wine ph values, for instance for a ph between 3.5 and 4, HT - is present at a maximum proportion (Zoecklein et al., 1995). In this range of ph the percentage of this ion is 70%, as shown in figure 1. One of the most common causes of instability in bottled wines is related with the precipitation of potassium bitartrate (KHT), and in a less extent with the calcium tartrate (CaT). Those two salts result from tartaric dissociation. Figure 1. Tartrate Solubility Curve (Smith, V. 2012). Although tartaric precipitations occur naturally in wine, if they take place in a bottled wine they could be misinterpreted as a sign of lack of quality of that wine. That represents a problem for wine industry. For that reason, wines are stabilized before being placed on the market. Tartaric precipitations can be avoided by using different methods (Claus, H. et al. 2014), such as: 1) Induction of precipitation before bottling by cold stabilization; 1

2) Selective removal of ions in excess (potassium and/or calcium); 3) Use of crystallization inhibitors. The precipitation of CaT, usually, only occurs in wines after numerous years of aging. This is due to the fact that the time necessary for spontaneous nucleation of CaT is much longer in comparison with KHT (Ribéreau-Gayon, 2006). The nucleation is the limitation step of the CaT precipitation. This is the principle of minicontact test, one of the methods used to predict the stability of CaT, which is based on the addition of "seed crystals" in order to promote crystal growth. This is related with the fact that homogeneous nucleation is more rapidly induced than the primary or spontaneous nucleation (McKinnon et al. 1995). The CaT content in a wine depends on the concentrations of calcium and tartrate ions. As mentioned above the concentration of T 2- depends on the dissociation of H 2T. The concentration of calcium present in wine depends not only on the concentration of this ion in grapes (endogenous origin), but also on the quantity that is added by some technological operations (exogenous origin). Currently, the phenomenon of CaT precipitation in wine is better known due to the significant scientific research that has been conducted on the subject. Studies carried out by McKinnon et al. 1994 have shown that the first step in CaT precipitation is associated with the formation of a soluble CaT species which by aggregation will result in a critical nucleus before the beginning of precipitation and crystal growth. Additives such as metatartaric and carboxymethyl cellulose (CMC) currently marketed are compounds added to the wine with the main purpose of inhibiting the crystallization of the tartaric salts, namely the CaT. The metatartaric has been the most widely used inhibitor in the wine industry. Its addition to the wine is a method of tartaric stabilization which is considered to be both an economic and of short-term method. This is a short-term method because the use of this additive, in a long-term, can lead to a slow hydrolyzation of itinto tartaric, which can then be responsible for increasing even more the risk of precipitation (Jackson, 2008; Gonçalves, F., 2002). When added to the wine, this additive interferes with the growth of CaT crystals (Jackson, 2008). Carboxymethyl cellulose (CMC) is more stable than the metatartaric, and it is very effective in the inhibition of tartrate precipitation. The effect of CMC on the inhibition of tartaric salts precipitation can be explained through the interaction of the CMC negative charge with the crystals positive charge (potassium and calcium) (Crachereau et al. 2001). Nevertheless, it has been argued that CMC also interacts with other wine compounds besides the interaction with the tartrate. The presence of malic was shown to reduce the velocity of CaT crystallization (Clark et al. 1988; McKinnon, 1995). Later, the influence of other organic s, including the malic, on the precipitation of the CaT, was studied in a hydroalcoholic model solution. These studies were carried out by McKinnon et al. (1995) and the following order of inhibition for each of the additives tested was found: citric > malic > lactic > succinic. These s were added to the model solution at the same mass concentration. These results have demonstrated that the inhibitors containing carboxyl groups or carboxyl and hydroxyl groups were the most effective ones in the inhibition of CaT. Pellerin et al. (2013) have shown the influence that the wine polysaccharides have on the CaT precipitation. From the study carried out also in a wine hydroalcoholic solution, it was observed that the RG-ɪ had shown to be the most effective inhibitor as a result of having the ability to participate in the calcium ions sequestration, while the RG-ɪɪ, although it is structurally more complex than the RG-ɪ, it has not demonstrated such a strong influence on the induction time associated with the CaT precipitation. The AGP (arabinogalactan proteins) and MP (mannoproteins) have shown to have a reduced or even no effect in the induction time of CaT. 2. Materials and Methods 2.1. Red Wine The red wine used in this experimental work was produced in INIAV (Instituto Nacional de Investigação Agrária e Veterinária) (Dois Portos, Portugal). Its physical and chemical characterization was also performed in these facilities and it is shown in Table 1. 2.2. Ultrafiltration experiments The experiments were performed in a plate-andframe Lab-Unit M20 (from DDS) (see Figure 2). The ultrafiltration tests were performed at 3 bar and at a temperature of about 20 C. 2

Table 1. Physical and Chemical characterization of the red wine. Alcoholic strength (% v/v) 12.1 Density (g/ml at 20 C) 0.9929 Total Dry Extract (g/l) 27.53 Chromatic Intensity 8.947 characteristics (420, Tone 520 e 620 nm) ( nuance ) 0.636 Total polyphenols index 50.25 Free sulfur dioxide (mg de SO 2/L) 5 Total sulfur dioxide (mg de SO 2/L) 100 Total ity (g de ácido tartárico/l) 6.0 Volatile ity (g de ácido acético/l) 0.47 ph 3.28 Reducing sugars (g/l) 1.9 Malolactic fermentation Positive/Negative Figure 2. Scheme of a plate-and-frame Lab-Unit M20 (DSS). 1. Feed Tank; 2. Valve; 3. Safety filter; 4. Highpressure pump; 5. By-pass valve (or needle valve); 6. heat exchanger; 7. Manometer; 8. plateand frame module; 9. permeate collector; 10. Pressure-control valve (Cui et al. 2010). 2.3. Membranes In the UF experiments two DSS (Denmark) organic fluor polymer membranes were used. The UF membrane, FS40PP, with 100 kda MWCO and the UF membrane, FS61PP, with 20 kda MWCO. The membranes surface areas used in these experiments were 0.072 m 2 (for FS40PP membrane) and 0.036 m 2 (for FS61PP membrane). The compaction of membranes consisted on permeation experiments of deionized water in full recycle mode for a period of 2 to 3 hours at a pressure 20% higher than the maximum operating pressure, so as to minimize any effect on the structure of the membranes during subsequent experiments Both membranes were characterized in terms of hydraulic permeability, and this corresponds to the slope of straight line given by the plot of pure water flux versus pressure. Thus, the hydraulic permeability determined for membranes FS40PP and FS61PP were respectively 108 and 56 Lh -1 m -2 bar -1. The molecular weight cut off (MWCO) was determined, for both membranes, through the reference solute rejections. The solutions of reference solutes considered were dextrans (molecular masses of 10, 40 and 70 kda) with a concentration of 500 mg / L. Rejections of organic solutes were determined in terms of total organic carbon (TOC). UF experiments with reference solutions were performed at a pressure of 1 bar, at a temperature of 20 C and at the maximum recirculation flowrate (9 L / min). The values of 112 and 155 kda were obtained, respectively for membranes FS40PP and FS61PP, for 91% rejection. The permeation experiments of red wine in concentration mode were performed at a pressure of 3 bar. With FS40PP membrane 5 trials were carried out corresponding to the volumetric concentration factors (VCF) of 1.5, 2, 3, 4 and 4.5. 2.4. Analytical Methods Total polysaccharide content was determined following the procedure described by Segarra et al. (1995). According to this author the precipitate can be obtained by using ethanol in a proportion of 1:5, i.e, to 4 ml of wine add 20 ml of ethanol. The precipitate obtained was dried under vacuum for a period of about 24 hours. Part of that precipitate was later used in the precipitations experiments. The other part was dissolved in water and was used to determine the total polysaccharide content by the phenolsulphuric method (Dubois et al. (1956)). The calibration curve used for this determination was obtained using glucose and the results were expressed in mg/l of this sugar. 2.5. Precipitations experiments The experiments conducted with the model solution with and without additives aim to determine the induction time associated with the precipitation of the CaT. The induction time was considered as the time when there was a decrease in the potential read through the calcium ion-selective electrode. The lower value corresponds to a lower concentration of ionized calcium and therefore it can be inferred that the induction time corresponds to the time where the concentration of ionized calcium begins decreasing. In addition, the induction time were also visually evaluated through the beginning of the turbidity 3

J p (L.h -1.m -2 ) of the solution. The use of both methods is consistent since the onset of turbidity, i.e. when initiating precipitation of the CaT resulted in an immediate change of the value of the potential read by the electrode. The model solution results of the mixture oftwo hydroalcoholic solutions. One of tartaric and other of calcium chloride dihydrate, in order to obtain a solution with a final concentration of 2.2 g / L (0.0143 M) in tartaric and 140 mg / L (0.0035 M) in calcium. The ratio between the concentrations of calcium and tartaric was in agreement with that presented in McKinnon et al. 1995. Equal volumes of the two solutions were added to obtain a total volume solution of 140 ml and also the desired concentrations. The adjustment of ionic strength and ph were made at this stage, and before mixing the two solutions. The ionic strength of the model solution was adjusted to 0.038 M, a value that is within the range of ionic strengths (0.016 and 0.056) considered by Berg and Keefer (1958, 1959) for wines. The adjustment of ionic strength was done by the addition of 4 M sodium chloride (Normapur Analar) to the calcium hydroalcoholic solution. This solution was contained in a reactor, ~250 ml, under stirring (850 rpm), and at a temperature of 15 C. The ph of the two hydroalcoholic solutions was adjusted at 3.5 by addition of 0.1 M HCl (37%, Panreac) or 5M sodium hydroxide (98%, Panreac). During ph adjustments, the hydroalcoholic solutions were under stirring, to ensure a better homogenization of the solution, which leads to a faster stabilization of the ph. In the precipitation experiments with additives, these were added to the hydroalcoholic tartaric solution, in order to avoid the possibility of calcium binding occur before the two solutions were mixed. All solutions used in precipitation experiments were prepared with Millipore deionized water. Precipitation curves were obtained by the plot of potential versus time. The readings of potential and ph were initiated, as soon as the two working solutions were mixed. The displayed graphical representations consider only the values read after 8 minutes. Only at this time can be ensure a proper mixing between the two solutions and the stabilization of temperature at 15 C. Temperature, ph, and ionic strength influence the potential values given by the calcium ion-selective electrode. 3. Results and Discussion 3.1. UF experiments with red wine Before starting the concentration mode experiments with the FS40PP membrane (100 kda), it was studied how the fluxes varies with the pressure increase. Figure 3 displays the results obtained. 45 40 35 30 25 20 15 10 5 0 0 0,5 1 1,5 2 2,5 3 P (bar) Figure 3. Variation of permeation fluxes of red wine with pressure. Membrane FS40PP; Operating conditions: T=20 C, at maximum circulation flowrate (9L/min); Membrane surface area: 0,072 m 2. It was found that this variation is linear up to pressure values of 1 bar. For pressures between 1.5 bar and 2 bar, the fluxes increase occurs more gradually. The linear behavior seems to return for pressure values of 2.5 to 3 bar). The fluxes obtained from UF of red wine are much smaller than those obtained with white wine (Pinho, M.N, 2010). Gonçalves et al. (2001), using the same membrane considered in this study (FS40PP membrane), and varying the pressure between about 0.8 and 2.5 bar, at the maximum circulating flowrate, obtained permeation fluxes for white wine that were between 125 and 275 L.h -1.m -2, respectively. Although the red wine permeation fluxes obtained, using the same membrane at the same circulation flowrate, shown to be lower (Goncalves et al., 2002). Thus, varying the pressure between about 0.8 and 2.6 bar, this author obtained permeation fluxes for red wine between about 25 and 50 L.h -1 m -2. The fluxes obtained for red wine in this study (Figure 3), and using the same membrane, were similar to those obtained by this author. For pressures between 0.5 and 3 bar the fluxes obtained were equal to12 and 42 L.h -1.m -2, respectively. The fluxes values may vary depending on the type of red wine that was used. Thus, the difference between the values obtained by the author and those obtained in the present study may be due to composition differences resulting, for example, from different winemaking techniques. 4

J p (L.h -1.m -2 ) 50 40 30 20 10 0 1,5 2 2,5 3 3,5 4 4,5 FCV Figure 4. Variation of permeation fluxes with VCF. The different VCF was used with the intent of obtain different polysaccharides concentrates. Membrane: FS40PP (100 kda). Operating conditions: T=20 C, P=3 bar at maximum circulation flowrate (9 L/min); Membrane surface area: 0,072 m 2. Afterwards, the concentration mode experiments were carried out at a pressure of 3 bar, since in the previous experiment it had been shown that higher permeation flux were achieved at this pressure. In addition, up to this pressure, there was no flux stabilization trend. Thus, it can be concluded that at this conditions membrane fouling did not occur and so it did not represent a risk to its normal operation. Concentration mode assays with FS40PP membrane to obtain different polysaccharide precipitates In order to test the influence of different VCF fractions on the inhibition of CaT precipitation, different experiments were carried out at concentration mode. In concentration mode experiments, permeate was continuously collected leading thus to a change in feed composition. Thus, the higher the FCV the higher would be the concentration of the rejected compounds in the feed, i.e., those having a molecular weight higher than 100 kda. The AGP and the MP are the polysaccharides of the wine with molecular weights higher than the cut off of the membrane considered, so it is expected that they will be part of concentrates composition. Lower molecular weight polysaccharides, such as RG's (ɪ and ɪɪ) are permeated. Figure 4 shows the variation of permeate fluxes with VCF. As can be seen from figure 4, the increasing of VCF to 4.5, led to a reduction on permeation fluxes. The flux decrease with the increasing of VCF was maybe due to a polarization concentration phenomenon and/or to membrane fouling. The next step was to use the lower cut-off membrane, FS61PP (20 kda), in order to obtain in permeate the polysaccharides of lower molecular weight, such as RG-ɪɪ. To achieve this, a concentration mode, experiment was carried out using the permeate resulting from the previous tests with FS40PP membrane. The permeate considered was that with a VCF of 4.5, since, as already mentioned above, the higher the VCF, the higher concentration of polysaccharides of greater molecular weight feed. Consequently, as intended, a permeate richer in lower molecular weight polysaccharides is obtained. Precipitation tests to evaluate the influence of different additives on the precipitation of CaT The induction time allows evaluating the importance that a given additive has on the nucleation stage. Influence of the addition of different commercial additives in CaT precipitation The inhibitor effect of a given additive was evaluated by comparing the induction time obtained for the model solution with and without the additive. In Table 2 are shown some of the additives tested by classes of compounds. The additive addition may in some cases lead to a change in ionic strength value calculated for the initial model solution. As to the calcium ion selective-electrode is sensitive to variations in ionic strength it was important to determine an approximate value for ionic strength in the situations where different additives were used. As an approximation, in calculating ionic strength it was been considered the corresponding dissociation constants for dilute solutions at 25 C. Those values are also present in table 2. 5

Table 2. Effect of different additives, mainly compounds that are part of wine composition, on the induction time of calcium tartrate. Model solution 11% (v/v) in ethanol with: [Ca 2+ ]=140 mg/l e [Tartaric ]=2,2 g/l. Operating conditions: T=15 C; Stirring speed=850 rpm. Additive class Organic s present in wine Organic not existing in wine composition Phenolic s present in wine Model Solution Without additive With lactic With malic With citric With gluconic With adipic With hydroxybenzoi c With salicylic Additive structure Additive concentration g/l (M) Inductio n time (min) Ionic Strenght (M) 0 g/l 13 0,038 2,3 g/l (0,026 M) 3,1 g/l (0,035 M) 1,4 g/l (0,01 M) 3 g/l (0,02 M) 0,3 g/l (0,0013 M) 1,4 g/l (0,0067 M) 3,8 g/l (0,02 M) 1,1 g/l (0,005 M) 2 g/l (0,009 M) 0,979 g/l (0,0067 M) 0,03 g/l (0,00022 M) 0,03 g/l (0,00022 M) 15 0,046 18 0,049 26 0,044 93 0,050 15 0,039 27 0,044 67 0,054 >107 0,040 >162 0,041 13 0,038 9 0,038 9 0,038 Neutral Sugars present in wine With arabinose With ramnose 0,4 g/l (0,003 M) 0,36 g/l (0,002 M) 13 0,039 13 0,038 Sugar Acid present in wine With galacturonic 0,35 g/l (0,0016 M) 1 g/l (0,0047 M) 3,88 g/l (0,02 M) 19 0,039 24 0,041 97 0,049 Amino s present in wine With glutamic With threonine 1,47 g/l (0,01 M) 1,91 g/l (0,02 M) 13 0,039 13 0,053 Organic s: The malic and citric s have demonstrated to have a higher effect than lactic on the induction time associated with CaT precipitation. In both cases, the higher the concentration of each of these s added to the model solution the more remarkable seems to be the delay in induction time (Table 2). The delay is observed through the baseline extension that occurs before the decrease in the potential, that is, 6

Potential (mv) Potential (mv) which occur before the onset of the precipitation (Figure 5). Malic when added at the same molar concentration that citric (0.02 M) revealed to have a higher inhibitor effect (Figure 5). 320 315 310 305 Model solution with 3,8 g/l (0,02 M) of citric Model solution Model solution wth addiction of 3,8 g/l of citric 8 36 64 92 120 148 176 Time (min) Model solution with 3 g/l (0,02 M) of malic 320 315 310 305 300 Model solution 8 36 64 92 120 148 176 Time (min) Figure 5. Potential as function of time for the assays with addiction of the same molar concentration of citric and malic s. Model solution 11% (v/v) in ethanol with: [Ca 2+ ]=140 mg/l e [Tartaric ]=2,2 g/l. Operating conditions: T=15 C; Stirring speed=850 rpm. In studies carried out by McKinnon et al. (1995), those organic s (citric and malic s) have demonstrated the capacity to bind with calcium at wine ph values. This knowledge helps to better understand the mechanism associated with inhibition of CaT precipitation by these organic s. Gluconic is other organic that is part of the wine composition. This is known to have a higher sequestrating power and to be a good chelator at alkaline ph (Ramachandran et al., 2006). It is a compound capable of capturing metallic ions, such as calcium, resulting in chelates. For the reasons stated above, the influence of this on the CaT induction time, at the wine ph values, was studies. Initially, this was added to the model solution at a concentration of 2 g/l. This concentration corresponding to the limit concentration in which it may be present in wine (Curvelo-Garcia, 1988). In these conditions, and as shown in table 2, precipitation did not occur during the duration of the assay. Later, it was evaluated if a decrease to half of this initial concentration results in a lower inhibition effect. This was not observed. Among all the organic s that have been tested, gluconic was the one with a higher effect on the induction time associated with CaT precipitation. The spontaneous inhibition of CaT precipitation caused by this might be explained through: 1) a binding established between the calcium ion and the gluconic, at the ph value of the study, resulting in the formation of calcium gluconate; 2) the fact that this might difficult the aggregation process of the CaT aqueous soluble specie, which consequently slows the nucleation and the occurrence of precipitation. The inhibition of CaT precipitation by gluconic seems to point out to an association between the carboxylic and hydroxyls groups on the inhibition of CaT precipitation. Although adipic is not part of the wine s composition, its addition was studied in order to understand if there is any structural characteristics for a certain additive to act as a CaT inhibitor. Adipic was added to the model solution at the same molar concentration that has been considered in previous assays with citric (0.0067 M). The results showed that this, as opposed to citric, had no effect on the induction time of CaT precipitation (see table 2). A potential structural factor that could explain the effect observed with citric on the CaT nucleation inhibition could be the extra carboxylic group when comparing with the adipic. However, when comparing gluconic (one carboxylic group) with adipic (two carboxylic groups), this structural factor is not observed. For molar concentrations similar to those tested with adipic, gluconic showed a much higher effect on CaT nucleation (table 2). Similar findings were also obtained by McKinnon et al. (1995). These authors tested at the same mass concentrations (2 g/l) the following organic s: lactic (one carboxylic group) and succinic (two carboxylic groups), both of them presented in wine. Succinic appeared to have a minor influence on the induction time than lactic. This was related with the fact that, for the ph 7

Potential (mv) under study (ph=3.5), a binding between calcium and succinic does not occur, unlike to what is observed with lactic. Phenolic s: The compounds tested have demonstrated to have no effect on the delay of CaT induction time. In certain cases, the addiction of these compounds resulted in lower induction times when comparing with the model solution (see table 2). The following hypothesis can be pointed out: 1) These compounds can induce precipitation. 2) The precipitate obtained could be other compound rather than CaT (possibly a calcium salt of these s). 3) Low solubility of the compound in the study solution might have facilitated the occurrence of precipitation. Hydroxybenzoic and salycilic s have a similar chemical composition, but have different structures. These two compounds, when added at the same mass concentration (0.03 g/l), have demonstrated to have the same effect on the induction time (table 2). Neutral Sugars: The neutral sugars (arabinose and rhamnose) that were tested have shown no influence on the CaT inhibition, nor either at the wine concentrations (arabinose) nor even at higher concentrations than those found in wine (rhamnose) (table 2). Sugar Acid: Within this subclass, galacturonic was studied. This is the most abundant uronic in the wine, and is also part of the main structure of polysaccharides, for example, the RGs. It was intended to study if this sugar has any influence on the CaT nucleation. First, this compound was added to the model solution at a concentration within the limits found in the wine. Sponholz and Dittrich (1984) mentioned that this is found at the german white wines, at concentrations between 0.15 and 1 g/l. At the tested concentrations of 0.35 and 1 g/l, it was intended to assess whether this compound had any influence on the CaT nucleation at the concentrations at which it might be presented in wine. Then, the effect of concentration on the CaT induction time was evaluated. A concentration of 3.88 g/l was considered. The effect of galacturonic in the induction time has been shown to be related with its concentration at solution. The higher the concentrations, the higher were the induction times obtained (see table 2). When added at a concentration of 0.35 g/l, this additive has shown to delay the induction time and the potential decay over time (see Figure 6). McKinnon et al. (1996) found that the addition of 0.5 g/l of this to a model solution (ph = 3.5) had no significant effect on the induction time, but has led to a decrease in crystallization rate. Similar results were obtained in this work. 320 318 316 314 312 310 308 306 304 Model solution with 0,35 g/l of galacturonic 8 18 28 38 48 58 68 Time (min) Amino s: Some of the amino s that are present in the wine were tested. These compounds contain hydroxyl or carboxyl groups in the side chain. As it was observed with the organic s, these compounds might have the appropriate structural characteristics to act as potential calcium binding agents. However, the amino s tested showed no influence on the CaT induction time (table 2). Precipitation assays with polysaccharides obtained by UF experiments with red wine With FS40PP Membrane Model solution with addiction of 0,35 g/l of galacturonic Model solution Figure 6. Potential as function of time for the model solution and the model containing 0,35 g/l of galacturonic. Model solution 11% (v/v) in ethanol with: [Ca 2+ ]=140 mg/l e [Tartaric ]=2,2 g/l. Operating conditions: T=15 C; Stirring speed=850 rpm. In these assays, low concentrations of wine s precipitated polysaccharides (300 mg/l) were added. The results showed that the addition of the polysaccharides corresponding to the concentrated with higher VCF have a higher influence in the induction times obtained, although in some cases this is not significant. The most notable effect on the induction time was registered for the precipitated polysaccharides corresponding to VCF of 4 and 8

Total polysaccharides content (mg/l) Potential (mv) 4.5, as demonstrated in Figure 7. These fractions were associated with higher concentrations of total polysaccharides (Figure 8). The potential decay over time, indicating the precipitation rate, has demonstrated to be similar for these two situations. 330 325 320 315 310 Concentrates fractions obtained by FS40PP membrane (100 kda) 8 58 108 Time (min) With FS61PP Membrane VCF=1,5 VCF=2 VCF=3 VCF=4 VCF=4,5 Figure 7. Potential as function of time for the assay with addiction of the concentrates fractions obtained with FS40PP (100 kda) membrane to the model solution. Model solution 11% (v/v) in ethanol with: [Ca 2+ ]=140 mg/l e [Tartaric ]=2,2 g/l. Operating conditions: T=15 C; Stirring speed=850 rpm. 440 430 420 410 400 390 380 370 360 350 1 1,5 2 2,5 3 3,5 4 4,5 VCF Figure 8. Total polysaccharides as function of VCF. The polysaccharides present either in the concentrate (MP and RG-I: according to the literature) and in the permeate (RG-II) do not show any influence on the induction time associated with the CaT precipitation. The results obtained are shown in Table 3. Table 3. Effect of polysaccharides obtained from UF assays with the FS61PP (20 kda) membrane in induction time. Model solution 11% (v/v) in ethanol with: [Ca 2+ ]=140 mg/l e [Tartaric ]=2,2 g/l. Operating conditions: T=15 C; Stirring speed=850 rpm. VCF Additive concentration (mg/l) Induction time (min) Model solution 0 13 Model solution with addiction of the concentrate 2,5 200 17 polysaccharides fraction Model solution with addiction of the permeate polysaccharide fraction 2,5 200 13 Conclusions Among the various classes of additive compounds tested, the organic s showed to be the class having the greater delay in the induction time. The inhibition order observed when these additives were added, at similar molar concentrations, was the following: Gluconic >Malic >Citric >Lactic. Gluconic was tested at concentrations that were within the range in which it can be found in wine. At these concentrations, this has evidenced to be a good inhibitor of CaT precipitation, since during the duration of the assays (about 2 hours) no precipitation occurred. This is known to be a good chelator of calcium at alkaline ph. A large delay in the time of induction upon addition of this to the model solution was achieved. This could be an indicator of a possible binding between calcium and this compound, at the ph value tested (ranging between 3.5). These are typical ph values of the wine. Malic has a higher inhibition power than lactic, as mentioned earlier, which might influence the CaT precipitation in wine. The occurrence of malolactic fermentation, which occurs naturally in wine, might increase the instability of the CaT precipitation in wine, by decreasing the concentration of a more efficient inhibitor (malic ) and increasing the concentration of a less efficient one (lactic ). At a structural level, it has been found that the presence of a higher number of carboxylic groups did not necessarily mean a higher inhibition power. Gluconic with one carboxyl group has demonstrated, when tested at similar molar concentrations, a greater impact on 9

induction time than citric (two carboxyl groups). Galacturonic (subclass of sugar s) showed to have an influence in the induction time. This effect demonstrated to be much more significant with the increasing of the concentration. All the compounds discussed above, which are shown to have influence on CaT precipitation, have carboxylic and hydroxyl groups in their structure. Other compounds from other classes (amino s, neutral sugars and phenolic s) showed to have no effect on the onset of precipitation. Amino s and phenolic s, although having carboxylic groups in its structure, did not affect the induction time. A possible explanation for this situation might be due to calcium inability to bind with these compounds, at the studied ph. Polysaccharides isolated from the fractions obtained by ultrafiltration showed no significant influence in the induction times. The most notable influence occurred when the precipitated polysaccharides of higher molecular weight obtained from FCV= 4 and 4.5 were added to the model solution. References Berg, H.W. e Keefer, R.M., Analytical determination of tartrate stability in wine.i. Potassium bitatrate. American Journal Of Enology and Viticulture, 9 (4), pp 127-134 (1958) (Citado por Lasanta et al. 2012 e Mira, H., 2004). Berg, H.W. e Keefer, R.M., Analytical determination of tartrate stability in wine.ii. Calcium tatrate. American Journal Of Enology and Viticulture, 10, pp 105-109 (1959) (Citado por Lasanta et al. 2012 e Mira, H., 2004). Clark, J.; Fugelsang, K.; Gump, B., Factors Affecting Induced Calcium Tartrate Precipitation from Wine. Am. J. Enol. Vitic., 39, 155-161 (1988). Claus, H., Tenzer, S., Sobe, M., Sclander, M., König, H., Fröhlich, J. Effect of carboxymethyl celulose on tartrate salt, protein and colour stability of red wine. Australian Journal of Grape and Wine Research (2014). Crachereau J.C., Gabas N., Blouin J., Hebrard S. and Maujean A. Bull. OIV, 841 842, 151. (2001) Cui, Z. F.; Muralidhara, H.S. Membrane Technology. A pratical Guide for Membrane Technology and Applications in Food and Bioprocessing. 1st Edition, BH, United States of America (2010). Curvelo-Garcia, A.S., Controlo de qualidade dos vinhos. Química Enológica. Métodos Analíticos. Ed. Instituto da Vinha e do Vinho (1988). De Pinho, M.N.. Membrane Processes in Must and Wine Industries. Membranes for Food Applications, Peinemann K.-V, Pereira Nunes S., Giorno L. Ed. Wiley-VCH (2010). Dubois, M.; Gilles, K.A.; Hamilton, J.K.; Rebers, P.A; Smith, F. Colorimetric Method for Determination of Sugars and Related Substances. Analytical Chemistry, 28 (3), (1956). Gonçalves, F., Fernandes, C., De Pinho, M.N., White wine clarification by micro/ultrafiltration: effect of removed colloids in tartaric stability. Separation and Purification Technology, 22-23, 423 429 (2001) Gonçalves, F., Optimização da clarificação e Estabilização Tartárica de Vinhos por Processos de Membranas. Influência das macromoléculas e da composição iónica. Tese de Doutoramento. Instituto Superior Técnico (2002). Jackson, R.S., Wine Science. Principles and applications. Elsevier (2008). Lasanta, C.; Gómez, J. Tartrate stabilization of wines. Food Science & Technology, 28, 52-59 (2012). McKinnon, A.J., Scollary, G.R.; Solomon, D.H., Williams, P.J., The influence of wine components on the spontaneous precipitation of calcium L(+)-Tartrate in a model wine solution. American Journal of Enology and Viticulture, 46, 509-517 (1995). Pellerin, P., Doco, T., Scollary, G. Original Article. The Influence of wine polymers on the spontaneous precipitation of calcium tartrate in a model solution. International Journal of Food Science and Technology, 48, 2676-2682 (2013). Ramachandran, S.; Fontanille, P.; Pandey, A.; Larroche, C. Gluconic Acid: Properties, Applications and Microbial Production (a review). Food Technol.Biotechnol. 44 (2), 185-195 (2006). Ribéreau-Gayon, P., Glories, Y., Maujean, A., Dubourdieu, D. Handbook of enology. Volume 2: The chemistry of wine. Stabilization and treatments. 2nd edition. Jonh Wiley & Sons, Ltd. England (2006). Segarra, I.; Lao, C.; López-Tamames, E; de la Torre-Boronat, M.C., Spectrophotometric methods for the analysis of polysaccharide levels in winemaking products. Am. J. Enol. Vitic., 46, 564-570 (1995). Smith, V. Assessment of Cold Stabilization for tartaric in wine. Penn State Food Science Undergraduate (2012). 10

Vidal, S., Williams, P., Doco, T., Moutnounet, M., Pellerin, P. The polysaccharides of red wine: total fractionation and characterization. Carbohydrate Polymers, 54, 439-447 (2003). Zoecklein, B.W., Fugelsang, K.C., Gump, B.H., Nury. F.S. (1995). Wine analysis and production. The Chapman & Hall Enology Library. International Thompson Publishing. 11