Research Article Received: 15 April 2014
Revised: 28 June 2014
Accepted article published: 21 July 2014
Published online in Wiley Online Library: 12 August 2014
(wileyonlinelibrary.com) DOI 10.1002/jsfa.6835
Effect of ethanol, dry extract and reducing sugars on density and viscosity of Brazilian red wines Flávia SPP Neto,a Maurício BM de Castilhos,a* Vânia RN Telisb and Javier Telis-Romerob Abstract BACKGROUND: Density and viscosity are properties that exert great influence on the body of wines. The present work aimed to evaluate the influence of the alcoholic content, dry extract, and reducing sugar content on density and viscosity of commercial dry red wines at different temperatures. The rheological assays were carried out on a controlled stress rheometer, using concentric cylinder geometry at seven temperatures (2, 8, 14, 16, 18, 20 and 26 ∘ C). RESULTS: Wine viscosity decreased with increasing temperature and density was directly related to the wine alcohol content, whereas viscosity was closely linked to the dry extract. Reducing sugars did not influence viscosity or density. Wines produced from Italian grapes were presented as full-bodied with higher values for density and viscosity, which was linked to the higher alcohol content and dry extract, respectively. CONCLUSION: The results highlighted the major effects of certain physicochemical properties on the physical properties of wines, which in turn is important for guiding sensory assessments. © 2014 Society of Chemical Industry Keywords: red dry wine; rheology; physicochemical properties; physical properties; principal component analysis
INTRODUCTION
J Sci Food Agric 2015; 95: 1421–1427
According to this definition, wines can be classified as full-bodied or light-bodied. Castilhos et al.6 found that full-bodied wines have a strong sensation of body because they present high alcohol content, phenolic compounds and expressive values of dry extract which enhance the sensation of filling the mouth. Light-bodied wines present the opposite. The correlation between a wine’s body and viscosity was perceived by Yanniotis et al.,9 as it affects the feeling of thickness in the mouth. Full-bodied wines generally have a high amount of sugars and persist longer on the palate. In the case of light-bodied wines, the time of persistence is low; the wine spreads easily through the mouth and is swallowed quickly. Although there are studies indicating that the body of a wine is linked to its alcohol content, sugar content or glycerol content,10 studies that relate physicochemical properties to this sensory attribute are scarce. In addition, knowledge of thermophysical and chemical properties of wines, especially data on density and viscosity, are essential for design and development of industrial
∗
Correspondence to: Maurício BM de Castilhos, Engineering and Food Technology Department, São Paulo State University, Cristóvão Colombo Street, 2265, São José do Rio Preto, São Paulo, Brazil. E-mail: [email protected]
a Engineering and Food Science PhD Program, São Paulo State University, São José do Rio Preto, São Paulo, Brazil b Engineering and Food Technology Department, São Paulo State University, São José do Rio Preto, São Paulo, Brazil
www.soci.org
© 2014 Society of Chemical Industry
1421
Processed and natural foods present different microstructures that result from distinct structural arrangements and multiple phases, which can be evaluated by macroscopic or microscopic analysis. The presence of these structures provides further functionality to food and beverages, such as nutritional value, texture control or shelf life. Rheological studies are employed as essential tools in the area of food engineering, since rheology is linked to food processing and stability, as well as to sensory perceptions, including texture and other oral sensations.1 Texture plays an important role in the acceptability of beverages and, before consumers have their first contact through mouthfeel, it is difficult to predict certain sensory perceptions by means of rheological properties or through application of measuring technics such as texture profile analysis (TPA).2,3 Wine is one of the beverages that causes a number of different oral sensations, and it is possible to describe some tactile ones, such as astringency, oiliness and pungency. Although these mouthfeel sensations are extremely important to wine quality, the literature concerning this subject is scarce and most research studies are related to volatile wine compounds responsible for aroma,4,5 or to the oenological practices that enhance and improve sensory characteristics of this beverage.6,7 The most important and studied feature in wine is the body mouthfeel. Jackson8 defines the body of wine as the tactile sensation induced by the presence of alcohol, but clearly influenced by sugars, glycerol (at high concentrations) and phenolic compounds.
www.soci.org equipment to be used in wine production lines. This information is needed for a variety of scientific research studies and applications in the field of engineering, particularly regarding the behavior of these properties as affected by concentration and temperature.11 Based on the above considerations, the present study aimed (i) to evaluate density and viscosity of commercial dry red wines under different temperatures and (ii) to obtain correlations with dry extract, alcohol and reducing sugar content in order to predict the influence of these physicochemical properties on rheological parameters.
MATERIAL AND METHODS Nine types of Brazilian commercial dry red wines from Serra Gaúcha, southern region of Brazil, were used. The wines were produced from different cultivars listed as follows: Nebbiolo (Wine W1), Cabernet Sauvignon (W2), Malbec (W3), Tannat (W4), Teroldego (W5), Cabernet Sauvignon (W6), Barbera (W7), Pinotage (W8) and Merlot (W9). Both W2 and W6 were Cabernet Sauvignon; however, they were produced by different Brazilian wineries. All determinations were repeated three times and each sample was collected in triplicate. The data were analyzed using a completely randomized design in order to avoid carryover effects. Determination of physical properties Steady shear rheological tests were carried out in a controlled stress rheometer, model AR-2000ex (TA Instruments, New Castle, DE, USA), using a concentric cylinder geometry with conical end (rotor radius 14 mm, cup radius 15 mm). Measurements were carried out following a logarithmic increasing shear rate ramp in the range of 1–250 s−1 , at different temperatures (2, 8, 14, 16, 18, 20 and 26 ∘ C), which were controlled by a Peltier system coupled to the outer cylinder. These temperatures were chosen based on the temperature range found in the winemaking process. The results were analyzed using Rheology Advantage software (TA Instruments) and all samples showed Newtonian behavior. Newton’s law of viscosity (Eqn (1))12 was fitted to the experimental shear stress versus shear rate data using linear regression, thus allowing determination of the viscosity of each system, which could be correlated with temperature. 𝜏 = 𝜇 𝛾̇ (1) In Eqn (1) 𝜏 is the shear stress (mPa), 𝛾̇ is the shear rate (s−1 ) and 𝜇 (mPa s) is the dynamic viscosity of the fluid. In general, the effect of temperature on the viscosity of Newtonian fluids can be represented by the Arrhenius equation (Eqn (2)): ( ) Ea (2) 𝜇 = B exp RT
FSPP Neto et al.
In this expression B is an empirical constant, R is the universal gas constant (8.314 J mol K−1 ), T is the absolute temperature (K) and E a (J mol−1 ) is the activation energy required for the flow.13 The correlation analysis between the physical and physicochemical properties was performed by density and viscosity measurements only at 16 and 18 ∘ C, because these temperatures are usually used in the sensory service of wine samples and in the red wine stabilization applied by wineries.8 Testing the accuracy of the rheometer The accuracy of the rheometer used in the rheological measurements can be observed through the results presented in Table 1, which compares the experimental viscosity of chlorobenzene with standard data presented by Lide.14 Validation of the rheometer was checked by the application of Student’s t-test at P < 0.05. Table 1 shows that the rheometer resulted in validated data, i.e. the results regarding experimental viscosity in different temperatures showed no significant differences (P-value > 0.050) compared to viscosity data cited by Lide.14 Density was determined using a digital electronic densimeter (model DMA 4500-M, Anton Paar, Graz, Austria) at different temperatures (2, 8, 14, 16, 18, 20 and 26 ∘ C), which were adjusted directly in the equipment. Wine samples of 50 mL were used for density measurements at each temperature. The densimeter performance was checked using an aqueous solution of acetic acid (50:50, v/v) with well-known densities described by Perry and Chilton.15 Eleven replicates were used at 0, 15 and 30 ∘ C. Validation of the densimeter was checked by applying Student’s t-test at P < 0.05 (Table 2). Table 2 shows that the density data were validated by the applied methodology, since results concerning the experimental density at the evaluated temperatures did not show significant differences (P-value > 0.050) when compared to values cited by Perry and Chilton.15 Determination of physicochemical properties Determination of the alcohol content was based on the official AOAC 920.57 method,16 in which a wine sample was distilled in a micro-distillator until 50 mL distillate was obtained. The distillate obtained was inserted into an electronic densimeter (model DMA-4500, Anton Paar) and, after temperature stabilization, the alcoholic degree was measured at 20 ∘ C. The densimeter used does not require the use of a density conversion table to percentage alcohol v/v, since readings are automatically converted to alcohol content by the instrument’s built-in tables. Tests were performed in triplicate. The content of dry extract was determined by the AOAC official method 920.62.16 Initially, 50 mL of each sample was weighed in a beaker and placed in a thermostatic bath at 100 ∘ C until the
Table 1. Validation of viscosity measurements using chlorobenzene as standard substance Standard substance Chlorobenzene (C6 H5 Cl)
a
Temperature (∘ C) 0 25 50
Viscositya (mPa s) 1.05 0.75 0.57
Mean ± SDb
CIc (95%)
1.05 ± 0.010 0.75 ± 0.007 0.57 ± 0.006
(1.05; 1.06) (0.74; 0.75) (0.57; 0.58)
P-value 0.596 0.377 0.087
Data from Lide (2004).
1422
b SD, standard deviation from 11 experimental values. c CI, confidence interval (95%).
wileyonlinelibrary.com/jsfa
© 2014 Society of Chemical Industry
J Sci Food Agric 2015; 95: 1421–1427
Effect of ethanol, dry extract and reducing sugars on Brazilian red wines
www.soci.org
Table 2. Validation of density measurements using aqueous solution of acetic acid (50:50, v/v) as standard substance Standard substance Aqueous solution of acetic acid
Temperature (∘ C)
Densitya (kg m−3 )
0 15 30
1072.9 1061.3 1049.2
Mean ± SDb 1073.3 ± 11.8 1061.0 ± 5.6 1041.2 ± 13.1
CIc (95%) (1065.4;1081.3) (1057.2;1064.8) (1032.4;1050.0)
P-value 0.899 0.876 0.073
a Data from Perry and Chilton (1986). b SD, standard deviation for 11 experimental values. c
CI, confidence interval (95%).
consistency of syrup was obtained. The samples were then placed in an oven at 100 ∘ C, up to constant weight. This step ranged from 3 to 5 h, depending on each type of wine. The final dry extract was obtained by the gravimetric method. Tests were performed in triplicate using a sample from each bottle for each analysis. Determination of reducing sugars was performed by the Lane and Eynon17 method, consisting of a sugar solution that is necessary to completely reduce a known volume of Fehling solution. This solution is actually blue and the final titration point is characterized by a reddish tone with a small amount of residue which presents the same color.16 These determinations were also performed in triplicate. Data analysis The statistical significance of density and viscosity were evaluated by analysis of variance (ANOVA), followed by Tukey’s multiple comparison post hoc test with P < 0.05. The influence of the physicochemical properties on physical parameters was subjected
to the principal component analysis (PCA) multivariate approach. Univariate and multivariate analysis was performed using Minitab 15 software (Minitab Inc., Coventry, UK) and Statistic 10 (StatSoft, Tulsa, OK, USA), respectively.
RESULTS AND DISCUSSION All the studied wines behaved as Newtonian fluids. Newton’s law of viscosity (Eqn (1)) was fitted to the experimental flow curves (Fig. 1), allowing the estimation of viscosity, which was subsequently described by a function of temperature. The fitting procedure resulted in good adjustment to experimental data (R2 > 0.99) in the evaluated temperatures. The Arrhenius model (Eqn (2)) was applied in order to describe the viscosity dependency with temperature. This approach enabled calculating the activation energy, E a , of each wine by linearization of viscosity and temperature data, leading to values in the range 20.14–24.80 kJ mol−1 (Table 3). These results are very
1423
Figure 1. Newtonian behavior of the wine samples.
J Sci Food Agric 2015; 95: 1421–1427
© 2014 Society of Chemical Industry
wileyonlinelibrary.com/jsfa
www.soci.org
Table 3. Calculated values of activation energy for the tested wines Sample
E a (kJ mol−1 )
W1 W2 W3 W4 W5 W6 W7 W8 W9
21.30 20.27 22.27 21.95 20.14 22.26 24.80 21.51 23.24
R2 0.985 0.997 0.989 0.999 0.994 0.990 0.995 0.994 0.988
Figure 2. Mean values of viscosity for the wine samples at different temperatures.
1424
similar to those reported by Yanniotis et al.9 for Greek wines. On the other hand, non-Newtonian pseudoplastic fluids usually exhibit higher activation energies for their consistency indexes, e.g. 43.65 kJ mol−1 for egg yolk, which indicates a high dependency of the corresponding rheological properties on temperature.13 Furthermore, viscosity decreased significantly with increasing temperature for all samples (Fig. 2), ranging from 2.90 mPa s at 2 ∘ C (W1) to 1.16 mPa s at 26 ∘ C (W7). In all wines, the viscosity differs significantly with respect to temperature (P < 0.001). On the other hand, when comparing different wines at the same temperature, the behavior of viscosity did not obey a regular pattern that could be detected; this irregular behavior could be explained by the fact that each of the assayed wines presents unique characteristics, since the samples include varietal and assemblage wines (Fig. 3). Regarding wine density, there were no significant differences either when comparing different temperatures or between different wine samples. The density values were nearly equal to water density, i.e. around 1000 kg m−3 . Density values around this value were also found by Etaio et al.18 for red wines produced from two different winemaking processes and by Yanniotis et al.9 for two different red dry wines at 16 ∘ C. The alcohol content varied between 11.2% and 14.7% (∘ GL), the dry extract ranged from 24.2 to 28.4 g L−1 and the reducing sugars were between 2.6 and 3.8 g L−1 , with significant differences among the samples (Table 4). The alcohol content and dry extract varied in ranges similar to those found by Yanniotis et al.9 and García-Muñoz et al.19
wileyonlinelibrary.com/jsfa
FSPP Neto et al.
The univariate analysis of density and viscosity data corresponding to each wine, from the standpoint of composition effect on the physicochemical properties, showed that viscosity and density were affected by different aspects of wine composition. The viscosity was favorably correlated only with dry extract, showing that the higher the wine dry extract the higher was its viscosity. Furthermore, density, at both temperatures, was significantly correlated with the alcohol content and dry extract, both being positive correlations (Table 5). Other authors have found that wine density decreased as the alcohol content increased;9 on the other hand, the dry extract is a physicochemical property that has been shown also to significantly affect wine density.6 Dry extract comprises the fixed wine compounds, i.e. mineral compounds, organic acids (malic, tartaric, lactic and others), phenolic substances (anthocyanin, flavonols, tannins, and hydroxybenzoic and hydroxycinnamic acids) and reducing sugars (arabinose, rhamnose and xylose) that are responsible to determine the body mouthfeel.20 Thus the highest density values observed could be related to the higher content of solids and not necessarily linked to lower alcohol levels. Density did not present any significant correlation with the wine reducing sugar content, which constituted only a small fraction ( 0.05).
Table 5. Pearson correlation coefficients (P-values) of viscosity and density in relation to alcohol content, dry extract and reducing sugar content Physicochemical properties
Rheological property Viscosity 16 ∘ C Viscosity 18 ∘ C Density 16 ∘ C Density 18 ∘ C
Alcohol content (∘ GL) 0.556 (0.112) 0.468 (0.204) 0.737 (0.023) 0.748 (0.021)
Dry extract (g L−1 ) 0.804 (0.009) 0.789 (0.012) 0.894 (0.001) 0.903 (0.001)
J Sci Food Agric 2015; 95: 1421–1427
0.260 (0.189) 0.437 (0.022) 0.003 (0.987) 0.051 (0.802)
© 2014 Society of Chemical Industry
wileyonlinelibrary.com/jsfa
1425
Significant correlations are highlighted in bold.
Reducing sugar content (g L−1 )
order to bring out their best sensory qualities.21,22 These features are closely linked to full-bodied wines like Teroldego wines, which present higher values of parameters responsible for wine color and body.23 In this context, it is possible to assume that these Italian cultivars produce full-bodied red wines, and this feature is closely linked to their high dry extract and alcohol potential. At the same time, wines W6 (Cabernet Sauvignon) and W8 (Pinotage) were located on the opposite side of component 1 and this result indicates the weak influence of dry extract and alcohol content on density and viscosity of these wines, leading to the conclusion that there are other features and properties that influence these physical properties. Sauvignon wine is blended with Merlot, a typical blend of the Bordeaux region of France, in order to provide to the Sauvignon wine a moderation of tannin content, making it smoother to the palate.21 Sauvignon wines are
www.soci.org
FSPP Neto et al.
(A)
(B)
Figure 4. Two-dimensional PCA plots for (A) physicochemical and physical properties and (B) wine samples.
1426
known as full-bodied wines because of the high amount of phenolic compounds and tannins, which provides strong astringency to the beverage.24 Pinotage wine is very popular in South Africa and its peculiar aroma gives a coffee,25 fruity and sweet flavor to the wine,26 which can influence flavor assessment. Based on these features, both the mentioned wines – W6 and W8 – could be classified as full-bodied owing the phenolic content or aroma composition,6,10 and not necessarily with a basis on the alcohol content or dry extract. These results demonstrate the relevance of multivariate
wileyonlinelibrary.com/jsfa
analysis. While the univariate approach presents a general trend of dependence between the analyzed variables, application of the multivariate technique allows identifying, in a punctual form, which variable is related to each wine sample. Reducing sugars did not present any relationship to density and viscosity. Sample W9 (Merlot) presented a higher reducing sugar content and sample W4 (Tannat) a lower value. This result is closely linked to the typicality of the cultivars and the winemaking procedure applied by the winery. Furthermore, additional factors such as soil type, sanitary conditions of the grapes, climate and
© 2014 Society of Chemical Industry
J Sci Food Agric 2015; 95: 1421–1427
Effect of ethanol, dry extract and reducing sugars on Brazilian red wines weather conditions, as well as management of the vine, influence the variation of reducing sugar content and other physicochemical properties.27 Results from other studies showed the correlation between reducing sugar content and body enhancement of Brazilian red wines produced from American grapes (Vitis labrusca);6,10 however, the results from the multivariate approach showed that reducing sugars did not influence the density or viscosity of the assessed wines. This can be explained by considering that the wine dry extract consists of a great number of components other than reducing sugars, such as organic acids, phenolic compounds, proteins, pectins and gums. These compounds are probably responsible for the higher densities and viscosities observed, independent of the reducing sugar content.21 Three samples did not present any correlation either for physical or for physicochemical properties: W2 (Cabernet Sauvignon), W3 (Malbec) and W7 (Barbera). The physical properties of these wine samples were not influenced by any of the physicochemical properties – dry extract, alcohol content or reducing sugar content. Therefore, it should be useful to study other physicochemical properties, such as phenolic compounds and tannins, since these substances exert a strong influence on the body of some wines.26 Based on these results, body evaluation of wines produced from Italian grapes could be predicted by physical (density and viscosity) and physicochemical properties (dry extract and alcohol content). For the other wines, other physicochemical properties such as phenolic compounds, tannins and aroma substances need to be studied in order to predict some influence on the physical properties and, consequently, on the body of these wine samples.
CONCLUSIONS The evaluation of commercial wines showed that density slightly decreased at higher temperatures; on the other hand, increasing temperature caused significant viscosity reduction. Density and viscosity, regardless of temperature, could be correlated with the alcohol content and dry extract, whereas reducing sugar content did not present any correlation with the physical properties. The multivariate approach showed that wine viscosity was mainly affected by dry extract, whereas wine density was mainly influenced by the alcohol content. Density and viscosity of wines produced by Italian grapes, such as Nebbiolo and Teroldego, were influenced by the alcohol content and dry extract, respectively; however, Sauvignon and Pinotage wines did not present any influence of dry extract or alcohol content on their physical properties. These results help in predicting the rheological behavior of wines during the winemaking process and contribute to guiding improvements in the sensory quality of wines, facilitating studies of descriptive analysis and sensory acceptance, through knowledge of density, viscosity and physicochemical properties by the tasters.
REFERENCES 1 Stokes JR, Boehm MW and Baier SK, Oral processing, texture and mouthfeel: from rheology to tribology and beyond. Curr Opin Colloid Interface Sci 18:349–359 (2013). 2 Bourne MC, Moyer JC and Hand DB, Measurement of food texture by a universal testing machine. Food Technol 20:522–526 (1966). 3 Stokes JR, Oral rheology, in Food Oral Processing: Fundamentals of Eating and Sensory Perception. Wiley-Blackwell, Hoboken, NJ, pp. 227–263 (2012).
www.soci.org
4 Jones PR, Gawel R, Francis IL and Waters EJ, The influence of interactions between major white wine components on the aroma, flavor and texture of model white wine. Food Qual Prefer 19:596–607 (2008). 5 Rigou P, Triay A and Razungles A, Influence of volatile thiols in the development of blackcurrant aroma in red wine. Food Chem 142:242–248 (2013). 6 Castilhos MBM, Cattelan MG, Conti-Silva AC and Del Bianchi VL, Influence of two different vinification procedures on the physicochemical and sensory properties of Brazilian non-Vitis vinifera red wines. LWT – Food Sci Technol 54:360–366 (2013). 7 Arbulu M, Sampedro MC, Sánchez-Ortega A, Gómez-Caballero A, Unceta N, Goicolea MA et al., Characterisation of the flavour profile from Graciano Vitis vinifera wine variety by a novel dual stir bar sorptive extraction methodology coupled to thermal desorption and gas chromatography–mass spectrometry. Anal Chim Acta 777:41–48 (2013). 8 Jackson RS, Wine Tasting: A Professional Handbook. Academic Press, San Jose, CA (2009). 9 Yanniotis S, Kotseridis G, Orfanidou A and Petraki A, Effect of ethanol, dry extract and glycerol on the viscosity of wine. J Food Eng 81:399–403 (2007). 10 Castilhos MBM, Conti-Silva AC and Del Bianchi VL, Effect of grape pre-drying and static pomace contact on physicochemical properties and sensory acceptance of Brazilian (Bordô and Isabel) red wines. Eur Food Res Technol 235:345–354 (2012). 11 Kosmerl T, Abramovic H and Klofutar C, The rheological properties of Slovenian wines. J Food Eng 46:165–171 (2000). 12 Rao MA, Rheology of Fluid and Semisolid Foods: Principles and Applications. Springer, New York (2007). 13 Telis-Romero J, Thomaz CEP, Bernardi M, Telis VRN and Gabas AL, Rheological properties and fluid dynamics of egg yolk. J Food Eng 74:191–197 (2006). 14 Lide DR, CRC Handbook of Chemistry and Physics. CRC Press, Miami, FL (2004). 15 Perry RH and Chilton CH, Manual de engenharia química. Guanabara Dois, Rio de Janeiro (1986). 16 Association of Official Agricultural Chemists, Official Methods of Analysis of AOAC International. AOC, Washington, DC (2005). 17 Lane H and Eynon L, Determination of reducing sugar by means of Fehling’s solution with methylene blue as internal indicator. J Soc Chem Ind 42:32–37 (1923). 18 Etaio I, Elortondo FJP, Albisu M, Gaston E, Ojeda M and Schlich P, Effect of winemaking process and addition of white grapes on the sensory and physicochemical characteristics of young red wines. Aust J Grape Wine Res 14:211–222 (2008). 19 García-Muñoz S, Muñoz-Organero G, Fernández-Fernández E and Cabello F, Sensory characterization and factors influencing quality of wines made from 18 minor varieties (Vitis vinifera L.). Food Qual Prefer 32:241–252 (2014). 20 Ribéreau-Gayon P, Glories Y, Maujean A and Dubourdieu D, Dry extract and minerals, in Handbook of Enology: The Chemistry of Wine, Stabilization and Treatments. Wiley, Chichester (2000). 21 Jackson RS, Wine Science: Principles and Applications. Academic Press, San Diego, CA (2008). 22 Armanino C, Forina M, Castino M, Piracci A and Ubigli M, Chemometrical investigation on four red wines from a single cultivar grown in the Piedmont region. Analyst 115:907–910 (1990). 23 Miele A, Rizzon LA and Zanus MC, Discrimination of Brazilian red wines according to the viticultural region, varietal and winery origin. Food Sci Technol 30:268–275 (2010). 24 Chira K, Pacella N, Jourdes M and Teissedre P, Chemical and sensory evaluation of Bordeaux wines (Cabernet Sauvignon and Merlot) and correlation with wine age. Food Chem 126:1971–1977 (2011). 25 Naudé Y and Rohwer ER, Investigating the coffee flavor in South African Pinotage wine using novel offline olfactometry and comprehensive gas chromatography with time of flight mass spectrometry. J Chromatogr A 1271:176–180 (2013). 26 Marais J, Effect of different wine-making techniques on the composition and quality of Pinotage wine. I. Low-temperature skin contact prior to fermentation. S Afr J Enol Vitic 24:70–75 (2003). 27 Lee SJ, Lee JE, Kim HW, Kim SS and Koh KH, Development of Korean red wines using Vitis labrusca varieties: instrumental and sensory characterization. Food Chem 94:385–393 (2006).
1427
J Sci Food Agric 2015; 95: 1421–1427
© 2014 Society of Chemical Industry
wileyonlinelibrary.com/jsfa
Revised: 28 June 2014
Accepted article published: 21 July 2014
Published online in Wiley Online Library: 12 August 2014
(wileyonlinelibrary.com) DOI 10.1002/jsfa.6835
Effect of ethanol, dry extract and reducing sugars on density and viscosity of Brazilian red wines Flávia SPP Neto,a Maurício BM de Castilhos,a* Vânia RN Telisb and Javier Telis-Romerob Abstract BACKGROUND: Density and viscosity are properties that exert great influence on the body of wines. The present work aimed to evaluate the influence of the alcoholic content, dry extract, and reducing sugar content on density and viscosity of commercial dry red wines at different temperatures. The rheological assays were carried out on a controlled stress rheometer, using concentric cylinder geometry at seven temperatures (2, 8, 14, 16, 18, 20 and 26 ∘ C). RESULTS: Wine viscosity decreased with increasing temperature and density was directly related to the wine alcohol content, whereas viscosity was closely linked to the dry extract. Reducing sugars did not influence viscosity or density. Wines produced from Italian grapes were presented as full-bodied with higher values for density and viscosity, which was linked to the higher alcohol content and dry extract, respectively. CONCLUSION: The results highlighted the major effects of certain physicochemical properties on the physical properties of wines, which in turn is important for guiding sensory assessments. © 2014 Society of Chemical Industry Keywords: red dry wine; rheology; physicochemical properties; physical properties; principal component analysis
INTRODUCTION
J Sci Food Agric 2015; 95: 1421–1427
According to this definition, wines can be classified as full-bodied or light-bodied. Castilhos et al.6 found that full-bodied wines have a strong sensation of body because they present high alcohol content, phenolic compounds and expressive values of dry extract which enhance the sensation of filling the mouth. Light-bodied wines present the opposite. The correlation between a wine’s body and viscosity was perceived by Yanniotis et al.,9 as it affects the feeling of thickness in the mouth. Full-bodied wines generally have a high amount of sugars and persist longer on the palate. In the case of light-bodied wines, the time of persistence is low; the wine spreads easily through the mouth and is swallowed quickly. Although there are studies indicating that the body of a wine is linked to its alcohol content, sugar content or glycerol content,10 studies that relate physicochemical properties to this sensory attribute are scarce. In addition, knowledge of thermophysical and chemical properties of wines, especially data on density and viscosity, are essential for design and development of industrial
∗
Correspondence to: Maurício BM de Castilhos, Engineering and Food Technology Department, São Paulo State University, Cristóvão Colombo Street, 2265, São José do Rio Preto, São Paulo, Brazil. E-mail: [email protected]
a Engineering and Food Science PhD Program, São Paulo State University, São José do Rio Preto, São Paulo, Brazil b Engineering and Food Technology Department, São Paulo State University, São José do Rio Preto, São Paulo, Brazil
www.soci.org
© 2014 Society of Chemical Industry
1421
Processed and natural foods present different microstructures that result from distinct structural arrangements and multiple phases, which can be evaluated by macroscopic or microscopic analysis. The presence of these structures provides further functionality to food and beverages, such as nutritional value, texture control or shelf life. Rheological studies are employed as essential tools in the area of food engineering, since rheology is linked to food processing and stability, as well as to sensory perceptions, including texture and other oral sensations.1 Texture plays an important role in the acceptability of beverages and, before consumers have their first contact through mouthfeel, it is difficult to predict certain sensory perceptions by means of rheological properties or through application of measuring technics such as texture profile analysis (TPA).2,3 Wine is one of the beverages that causes a number of different oral sensations, and it is possible to describe some tactile ones, such as astringency, oiliness and pungency. Although these mouthfeel sensations are extremely important to wine quality, the literature concerning this subject is scarce and most research studies are related to volatile wine compounds responsible for aroma,4,5 or to the oenological practices that enhance and improve sensory characteristics of this beverage.6,7 The most important and studied feature in wine is the body mouthfeel. Jackson8 defines the body of wine as the tactile sensation induced by the presence of alcohol, but clearly influenced by sugars, glycerol (at high concentrations) and phenolic compounds.
www.soci.org equipment to be used in wine production lines. This information is needed for a variety of scientific research studies and applications in the field of engineering, particularly regarding the behavior of these properties as affected by concentration and temperature.11 Based on the above considerations, the present study aimed (i) to evaluate density and viscosity of commercial dry red wines under different temperatures and (ii) to obtain correlations with dry extract, alcohol and reducing sugar content in order to predict the influence of these physicochemical properties on rheological parameters.
MATERIAL AND METHODS Nine types of Brazilian commercial dry red wines from Serra Gaúcha, southern region of Brazil, were used. The wines were produced from different cultivars listed as follows: Nebbiolo (Wine W1), Cabernet Sauvignon (W2), Malbec (W3), Tannat (W4), Teroldego (W5), Cabernet Sauvignon (W6), Barbera (W7), Pinotage (W8) and Merlot (W9). Both W2 and W6 were Cabernet Sauvignon; however, they were produced by different Brazilian wineries. All determinations were repeated three times and each sample was collected in triplicate. The data were analyzed using a completely randomized design in order to avoid carryover effects. Determination of physical properties Steady shear rheological tests were carried out in a controlled stress rheometer, model AR-2000ex (TA Instruments, New Castle, DE, USA), using a concentric cylinder geometry with conical end (rotor radius 14 mm, cup radius 15 mm). Measurements were carried out following a logarithmic increasing shear rate ramp in the range of 1–250 s−1 , at different temperatures (2, 8, 14, 16, 18, 20 and 26 ∘ C), which were controlled by a Peltier system coupled to the outer cylinder. These temperatures were chosen based on the temperature range found in the winemaking process. The results were analyzed using Rheology Advantage software (TA Instruments) and all samples showed Newtonian behavior. Newton’s law of viscosity (Eqn (1))12 was fitted to the experimental shear stress versus shear rate data using linear regression, thus allowing determination of the viscosity of each system, which could be correlated with temperature. 𝜏 = 𝜇 𝛾̇ (1) In Eqn (1) 𝜏 is the shear stress (mPa), 𝛾̇ is the shear rate (s−1 ) and 𝜇 (mPa s) is the dynamic viscosity of the fluid. In general, the effect of temperature on the viscosity of Newtonian fluids can be represented by the Arrhenius equation (Eqn (2)): ( ) Ea (2) 𝜇 = B exp RT
FSPP Neto et al.
In this expression B is an empirical constant, R is the universal gas constant (8.314 J mol K−1 ), T is the absolute temperature (K) and E a (J mol−1 ) is the activation energy required for the flow.13 The correlation analysis between the physical and physicochemical properties was performed by density and viscosity measurements only at 16 and 18 ∘ C, because these temperatures are usually used in the sensory service of wine samples and in the red wine stabilization applied by wineries.8 Testing the accuracy of the rheometer The accuracy of the rheometer used in the rheological measurements can be observed through the results presented in Table 1, which compares the experimental viscosity of chlorobenzene with standard data presented by Lide.14 Validation of the rheometer was checked by the application of Student’s t-test at P < 0.05. Table 1 shows that the rheometer resulted in validated data, i.e. the results regarding experimental viscosity in different temperatures showed no significant differences (P-value > 0.050) compared to viscosity data cited by Lide.14 Density was determined using a digital electronic densimeter (model DMA 4500-M, Anton Paar, Graz, Austria) at different temperatures (2, 8, 14, 16, 18, 20 and 26 ∘ C), which were adjusted directly in the equipment. Wine samples of 50 mL were used for density measurements at each temperature. The densimeter performance was checked using an aqueous solution of acetic acid (50:50, v/v) with well-known densities described by Perry and Chilton.15 Eleven replicates were used at 0, 15 and 30 ∘ C. Validation of the densimeter was checked by applying Student’s t-test at P < 0.05 (Table 2). Table 2 shows that the density data were validated by the applied methodology, since results concerning the experimental density at the evaluated temperatures did not show significant differences (P-value > 0.050) when compared to values cited by Perry and Chilton.15 Determination of physicochemical properties Determination of the alcohol content was based on the official AOAC 920.57 method,16 in which a wine sample was distilled in a micro-distillator until 50 mL distillate was obtained. The distillate obtained was inserted into an electronic densimeter (model DMA-4500, Anton Paar) and, after temperature stabilization, the alcoholic degree was measured at 20 ∘ C. The densimeter used does not require the use of a density conversion table to percentage alcohol v/v, since readings are automatically converted to alcohol content by the instrument’s built-in tables. Tests were performed in triplicate. The content of dry extract was determined by the AOAC official method 920.62.16 Initially, 50 mL of each sample was weighed in a beaker and placed in a thermostatic bath at 100 ∘ C until the
Table 1. Validation of viscosity measurements using chlorobenzene as standard substance Standard substance Chlorobenzene (C6 H5 Cl)
a
Temperature (∘ C) 0 25 50
Viscositya (mPa s) 1.05 0.75 0.57
Mean ± SDb
CIc (95%)
1.05 ± 0.010 0.75 ± 0.007 0.57 ± 0.006
(1.05; 1.06) (0.74; 0.75) (0.57; 0.58)
P-value 0.596 0.377 0.087
Data from Lide (2004).
1422
b SD, standard deviation from 11 experimental values. c CI, confidence interval (95%).
wileyonlinelibrary.com/jsfa
© 2014 Society of Chemical Industry
J Sci Food Agric 2015; 95: 1421–1427
Effect of ethanol, dry extract and reducing sugars on Brazilian red wines
www.soci.org
Table 2. Validation of density measurements using aqueous solution of acetic acid (50:50, v/v) as standard substance Standard substance Aqueous solution of acetic acid
Temperature (∘ C)
Densitya (kg m−3 )
0 15 30
1072.9 1061.3 1049.2
Mean ± SDb 1073.3 ± 11.8 1061.0 ± 5.6 1041.2 ± 13.1
CIc (95%) (1065.4;1081.3) (1057.2;1064.8) (1032.4;1050.0)
P-value 0.899 0.876 0.073
a Data from Perry and Chilton (1986). b SD, standard deviation for 11 experimental values. c
CI, confidence interval (95%).
consistency of syrup was obtained. The samples were then placed in an oven at 100 ∘ C, up to constant weight. This step ranged from 3 to 5 h, depending on each type of wine. The final dry extract was obtained by the gravimetric method. Tests were performed in triplicate using a sample from each bottle for each analysis. Determination of reducing sugars was performed by the Lane and Eynon17 method, consisting of a sugar solution that is necessary to completely reduce a known volume of Fehling solution. This solution is actually blue and the final titration point is characterized by a reddish tone with a small amount of residue which presents the same color.16 These determinations were also performed in triplicate. Data analysis The statistical significance of density and viscosity were evaluated by analysis of variance (ANOVA), followed by Tukey’s multiple comparison post hoc test with P < 0.05. The influence of the physicochemical properties on physical parameters was subjected
to the principal component analysis (PCA) multivariate approach. Univariate and multivariate analysis was performed using Minitab 15 software (Minitab Inc., Coventry, UK) and Statistic 10 (StatSoft, Tulsa, OK, USA), respectively.
RESULTS AND DISCUSSION All the studied wines behaved as Newtonian fluids. Newton’s law of viscosity (Eqn (1)) was fitted to the experimental flow curves (Fig. 1), allowing the estimation of viscosity, which was subsequently described by a function of temperature. The fitting procedure resulted in good adjustment to experimental data (R2 > 0.99) in the evaluated temperatures. The Arrhenius model (Eqn (2)) was applied in order to describe the viscosity dependency with temperature. This approach enabled calculating the activation energy, E a , of each wine by linearization of viscosity and temperature data, leading to values in the range 20.14–24.80 kJ mol−1 (Table 3). These results are very
1423
Figure 1. Newtonian behavior of the wine samples.
J Sci Food Agric 2015; 95: 1421–1427
© 2014 Society of Chemical Industry
wileyonlinelibrary.com/jsfa
www.soci.org
Table 3. Calculated values of activation energy for the tested wines Sample
E a (kJ mol−1 )
W1 W2 W3 W4 W5 W6 W7 W8 W9
21.30 20.27 22.27 21.95 20.14 22.26 24.80 21.51 23.24
R2 0.985 0.997 0.989 0.999 0.994 0.990 0.995 0.994 0.988
Figure 2. Mean values of viscosity for the wine samples at different temperatures.
1424
similar to those reported by Yanniotis et al.9 for Greek wines. On the other hand, non-Newtonian pseudoplastic fluids usually exhibit higher activation energies for their consistency indexes, e.g. 43.65 kJ mol−1 for egg yolk, which indicates a high dependency of the corresponding rheological properties on temperature.13 Furthermore, viscosity decreased significantly with increasing temperature for all samples (Fig. 2), ranging from 2.90 mPa s at 2 ∘ C (W1) to 1.16 mPa s at 26 ∘ C (W7). In all wines, the viscosity differs significantly with respect to temperature (P < 0.001). On the other hand, when comparing different wines at the same temperature, the behavior of viscosity did not obey a regular pattern that could be detected; this irregular behavior could be explained by the fact that each of the assayed wines presents unique characteristics, since the samples include varietal and assemblage wines (Fig. 3). Regarding wine density, there were no significant differences either when comparing different temperatures or between different wine samples. The density values were nearly equal to water density, i.e. around 1000 kg m−3 . Density values around this value were also found by Etaio et al.18 for red wines produced from two different winemaking processes and by Yanniotis et al.9 for two different red dry wines at 16 ∘ C. The alcohol content varied between 11.2% and 14.7% (∘ GL), the dry extract ranged from 24.2 to 28.4 g L−1 and the reducing sugars were between 2.6 and 3.8 g L−1 , with significant differences among the samples (Table 4). The alcohol content and dry extract varied in ranges similar to those found by Yanniotis et al.9 and García-Muñoz et al.19
wileyonlinelibrary.com/jsfa
FSPP Neto et al.
The univariate analysis of density and viscosity data corresponding to each wine, from the standpoint of composition effect on the physicochemical properties, showed that viscosity and density were affected by different aspects of wine composition. The viscosity was favorably correlated only with dry extract, showing that the higher the wine dry extract the higher was its viscosity. Furthermore, density, at both temperatures, was significantly correlated with the alcohol content and dry extract, both being positive correlations (Table 5). Other authors have found that wine density decreased as the alcohol content increased;9 on the other hand, the dry extract is a physicochemical property that has been shown also to significantly affect wine density.6 Dry extract comprises the fixed wine compounds, i.e. mineral compounds, organic acids (malic, tartaric, lactic and others), phenolic substances (anthocyanin, flavonols, tannins, and hydroxybenzoic and hydroxycinnamic acids) and reducing sugars (arabinose, rhamnose and xylose) that are responsible to determine the body mouthfeel.20 Thus the highest density values observed could be related to the higher content of solids and not necessarily linked to lower alcohol levels. Density did not present any significant correlation with the wine reducing sugar content, which constituted only a small fraction ( 0.05).
Table 5. Pearson correlation coefficients (P-values) of viscosity and density in relation to alcohol content, dry extract and reducing sugar content Physicochemical properties
Rheological property Viscosity 16 ∘ C Viscosity 18 ∘ C Density 16 ∘ C Density 18 ∘ C
Alcohol content (∘ GL) 0.556 (0.112) 0.468 (0.204) 0.737 (0.023) 0.748 (0.021)
Dry extract (g L−1 ) 0.804 (0.009) 0.789 (0.012) 0.894 (0.001) 0.903 (0.001)
J Sci Food Agric 2015; 95: 1421–1427
0.260 (0.189) 0.437 (0.022) 0.003 (0.987) 0.051 (0.802)
© 2014 Society of Chemical Industry
wileyonlinelibrary.com/jsfa
1425
Significant correlations are highlighted in bold.
Reducing sugar content (g L−1 )
order to bring out their best sensory qualities.21,22 These features are closely linked to full-bodied wines like Teroldego wines, which present higher values of parameters responsible for wine color and body.23 In this context, it is possible to assume that these Italian cultivars produce full-bodied red wines, and this feature is closely linked to their high dry extract and alcohol potential. At the same time, wines W6 (Cabernet Sauvignon) and W8 (Pinotage) were located on the opposite side of component 1 and this result indicates the weak influence of dry extract and alcohol content on density and viscosity of these wines, leading to the conclusion that there are other features and properties that influence these physical properties. Sauvignon wine is blended with Merlot, a typical blend of the Bordeaux region of France, in order to provide to the Sauvignon wine a moderation of tannin content, making it smoother to the palate.21 Sauvignon wines are
www.soci.org
FSPP Neto et al.
(A)
(B)
Figure 4. Two-dimensional PCA plots for (A) physicochemical and physical properties and (B) wine samples.
1426
known as full-bodied wines because of the high amount of phenolic compounds and tannins, which provides strong astringency to the beverage.24 Pinotage wine is very popular in South Africa and its peculiar aroma gives a coffee,25 fruity and sweet flavor to the wine,26 which can influence flavor assessment. Based on these features, both the mentioned wines – W6 and W8 – could be classified as full-bodied owing the phenolic content or aroma composition,6,10 and not necessarily with a basis on the alcohol content or dry extract. These results demonstrate the relevance of multivariate
wileyonlinelibrary.com/jsfa
analysis. While the univariate approach presents a general trend of dependence between the analyzed variables, application of the multivariate technique allows identifying, in a punctual form, which variable is related to each wine sample. Reducing sugars did not present any relationship to density and viscosity. Sample W9 (Merlot) presented a higher reducing sugar content and sample W4 (Tannat) a lower value. This result is closely linked to the typicality of the cultivars and the winemaking procedure applied by the winery. Furthermore, additional factors such as soil type, sanitary conditions of the grapes, climate and
© 2014 Society of Chemical Industry
J Sci Food Agric 2015; 95: 1421–1427
Effect of ethanol, dry extract and reducing sugars on Brazilian red wines weather conditions, as well as management of the vine, influence the variation of reducing sugar content and other physicochemical properties.27 Results from other studies showed the correlation between reducing sugar content and body enhancement of Brazilian red wines produced from American grapes (Vitis labrusca);6,10 however, the results from the multivariate approach showed that reducing sugars did not influence the density or viscosity of the assessed wines. This can be explained by considering that the wine dry extract consists of a great number of components other than reducing sugars, such as organic acids, phenolic compounds, proteins, pectins and gums. These compounds are probably responsible for the higher densities and viscosities observed, independent of the reducing sugar content.21 Three samples did not present any correlation either for physical or for physicochemical properties: W2 (Cabernet Sauvignon), W3 (Malbec) and W7 (Barbera). The physical properties of these wine samples were not influenced by any of the physicochemical properties – dry extract, alcohol content or reducing sugar content. Therefore, it should be useful to study other physicochemical properties, such as phenolic compounds and tannins, since these substances exert a strong influence on the body of some wines.26 Based on these results, body evaluation of wines produced from Italian grapes could be predicted by physical (density and viscosity) and physicochemical properties (dry extract and alcohol content). For the other wines, other physicochemical properties such as phenolic compounds, tannins and aroma substances need to be studied in order to predict some influence on the physical properties and, consequently, on the body of these wine samples.
CONCLUSIONS The evaluation of commercial wines showed that density slightly decreased at higher temperatures; on the other hand, increasing temperature caused significant viscosity reduction. Density and viscosity, regardless of temperature, could be correlated with the alcohol content and dry extract, whereas reducing sugar content did not present any correlation with the physical properties. The multivariate approach showed that wine viscosity was mainly affected by dry extract, whereas wine density was mainly influenced by the alcohol content. Density and viscosity of wines produced by Italian grapes, such as Nebbiolo and Teroldego, were influenced by the alcohol content and dry extract, respectively; however, Sauvignon and Pinotage wines did not present any influence of dry extract or alcohol content on their physical properties. These results help in predicting the rheological behavior of wines during the winemaking process and contribute to guiding improvements in the sensory quality of wines, facilitating studies of descriptive analysis and sensory acceptance, through knowledge of density, viscosity and physicochemical properties by the tasters.
REFERENCES 1 Stokes JR, Boehm MW and Baier SK, Oral processing, texture and mouthfeel: from rheology to tribology and beyond. Curr Opin Colloid Interface Sci 18:349–359 (2013). 2 Bourne MC, Moyer JC and Hand DB, Measurement of food texture by a universal testing machine. Food Technol 20:522–526 (1966). 3 Stokes JR, Oral rheology, in Food Oral Processing: Fundamentals of Eating and Sensory Perception. Wiley-Blackwell, Hoboken, NJ, pp. 227–263 (2012).
www.soci.org
4 Jones PR, Gawel R, Francis IL and Waters EJ, The influence of interactions between major white wine components on the aroma, flavor and texture of model white wine. Food Qual Prefer 19:596–607 (2008). 5 Rigou P, Triay A and Razungles A, Influence of volatile thiols in the development of blackcurrant aroma in red wine. Food Chem 142:242–248 (2013). 6 Castilhos MBM, Cattelan MG, Conti-Silva AC and Del Bianchi VL, Influence of two different vinification procedures on the physicochemical and sensory properties of Brazilian non-Vitis vinifera red wines. LWT – Food Sci Technol 54:360–366 (2013). 7 Arbulu M, Sampedro MC, Sánchez-Ortega A, Gómez-Caballero A, Unceta N, Goicolea MA et al., Characterisation of the flavour profile from Graciano Vitis vinifera wine variety by a novel dual stir bar sorptive extraction methodology coupled to thermal desorption and gas chromatography–mass spectrometry. Anal Chim Acta 777:41–48 (2013). 8 Jackson RS, Wine Tasting: A Professional Handbook. Academic Press, San Jose, CA (2009). 9 Yanniotis S, Kotseridis G, Orfanidou A and Petraki A, Effect of ethanol, dry extract and glycerol on the viscosity of wine. J Food Eng 81:399–403 (2007). 10 Castilhos MBM, Conti-Silva AC and Del Bianchi VL, Effect of grape pre-drying and static pomace contact on physicochemical properties and sensory acceptance of Brazilian (Bordô and Isabel) red wines. Eur Food Res Technol 235:345–354 (2012). 11 Kosmerl T, Abramovic H and Klofutar C, The rheological properties of Slovenian wines. J Food Eng 46:165–171 (2000). 12 Rao MA, Rheology of Fluid and Semisolid Foods: Principles and Applications. Springer, New York (2007). 13 Telis-Romero J, Thomaz CEP, Bernardi M, Telis VRN and Gabas AL, Rheological properties and fluid dynamics of egg yolk. J Food Eng 74:191–197 (2006). 14 Lide DR, CRC Handbook of Chemistry and Physics. CRC Press, Miami, FL (2004). 15 Perry RH and Chilton CH, Manual de engenharia química. Guanabara Dois, Rio de Janeiro (1986). 16 Association of Official Agricultural Chemists, Official Methods of Analysis of AOAC International. AOC, Washington, DC (2005). 17 Lane H and Eynon L, Determination of reducing sugar by means of Fehling’s solution with methylene blue as internal indicator. J Soc Chem Ind 42:32–37 (1923). 18 Etaio I, Elortondo FJP, Albisu M, Gaston E, Ojeda M and Schlich P, Effect of winemaking process and addition of white grapes on the sensory and physicochemical characteristics of young red wines. Aust J Grape Wine Res 14:211–222 (2008). 19 García-Muñoz S, Muñoz-Organero G, Fernández-Fernández E and Cabello F, Sensory characterization and factors influencing quality of wines made from 18 minor varieties (Vitis vinifera L.). Food Qual Prefer 32:241–252 (2014). 20 Ribéreau-Gayon P, Glories Y, Maujean A and Dubourdieu D, Dry extract and minerals, in Handbook of Enology: The Chemistry of Wine, Stabilization and Treatments. Wiley, Chichester (2000). 21 Jackson RS, Wine Science: Principles and Applications. Academic Press, San Diego, CA (2008). 22 Armanino C, Forina M, Castino M, Piracci A and Ubigli M, Chemometrical investigation on four red wines from a single cultivar grown in the Piedmont region. Analyst 115:907–910 (1990). 23 Miele A, Rizzon LA and Zanus MC, Discrimination of Brazilian red wines according to the viticultural region, varietal and winery origin. Food Sci Technol 30:268–275 (2010). 24 Chira K, Pacella N, Jourdes M and Teissedre P, Chemical and sensory evaluation of Bordeaux wines (Cabernet Sauvignon and Merlot) and correlation with wine age. Food Chem 126:1971–1977 (2011). 25 Naudé Y and Rohwer ER, Investigating the coffee flavor in South African Pinotage wine using novel offline olfactometry and comprehensive gas chromatography with time of flight mass spectrometry. J Chromatogr A 1271:176–180 (2013). 26 Marais J, Effect of different wine-making techniques on the composition and quality of Pinotage wine. I. Low-temperature skin contact prior to fermentation. S Afr J Enol Vitic 24:70–75 (2003). 27 Lee SJ, Lee JE, Kim HW, Kim SS and Koh KH, Development of Korean red wines using Vitis labrusca varieties: instrumental and sensory characterization. Food Chem 94:385–393 (2006).
1427
J Sci Food Agric 2015; 95: 1421–1427
© 2014 Society of Chemical Industry
wileyonlinelibrary.com/jsfa
