Journal of Food Engineering 81 (2007) 200208 www.elsevier.

com/locate/jfoodeng

Eects of extraction time, temperature and solvent on concentration and antioxidant activity of grape marc phenolics
Giorgia Spigno *, Lorenza Tramelli, Dante Marco De Faveri
` Cattolica Sacro Cuore, Via Emilia Parmense, 84, 29100 Piacenza, Italy Institute of Oenology and Food Engineering, Universita Received 13 July 2006; received in revised form 13 October 2006; accepted 25 October 2006 Available online 12 December 2006

Abstract This study was aimed to optimize the extraction of phenolic compounds from grape marc investigating extraction kinetics (from 1 to 24 h) at 45 and 60 C, and the eect of solvent (ethanol with dierent water content) on phenols yield and quality of extracts (phenols concentration and antioxidant power). Extraction was a slow process, with higher yields at 60 C than at 45 C, and with apparent thermal degradation of constituents beyond 20 h. Phenols yield increased for water content of ethanol from 10% to 30% and remained constant for water content from 30% to 60%, while phenols concentration of extracts decreased for water content above 50%. Antioxidant power (ABTS test) highly correlated to total phenols concentration, and was not inuenced by water content of ethanol, suggesting that this variable inuenced only the amount but not the nature of the extracted compounds. Freeze-drying did not alter composition and antioxidant property of extracts. 2006 Elsevier Ltd. All rights reserved.
Keywords: Antioxidant activity; Freeze-drying; Phenolic compounds; Solvent extraction; Wine-making wastes

1. Introduction Importance of natural antioxidants for medical and food application has been underlined by numerous works as reported by Spigno and De Faveri (2007). The same work reported also about the increasing processing of agricultural wastes (such as wine-making wastes) as a low-cost source of antioxidants (phenolic compounds). Recovery of these components is commonly performed through a solvent-extraction procedure but, at present, unambiguous data on the methods and conditions for extraction are available and sometimes contradictory, particularly if different raw materials are compared. Moreover, results are
Abbreviations: AOP, antioxidant power; CAE, caeic acid equivalents; GAE, gallic acid equivalents evaluated by the FolinCiocalteu method; GAE-280, gallic acid equivalents evaluated by direct reading of absorbance at 280 nm; QE, quercetin equivalents; TAN, tannins; TEAC, Trolox equivalent antioxidant capacity. * Corresponding author. Tel.: +39 0523599181; fax: +39 0523599232. E-mail address: giorgia.spigno@unicatt.it (G. Spigno). 0260-8774/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2006.10.021

dicult to compare even because phenols are measured in dierent ways and sometimes only the phenolics content of the nal extracts is reported, but not the total yield. The aim of an extraction process should be, of course, to provide for the maximum yield of substances and of the highest quality (concentration of target compounds and antioxidant power of the extracts). Focusing on those authors who applied a single-stage solvent extraction, only a few literature works on antioxidant recovery from grape by-products were aimed to the optimisation of some process parameters (Table 1). The variables investigated up to now have been: pre-treatment of the sample (degreasing and size reduction), solvent/sample ratio, type of solvent, time and temperature of extraction. As concerns the inuence of pre-treatment of the sample, it was shown that degreasing reduced total phenols yield, increasing only slightly the phenolic content of extracts (Spigno & De Faveri, 2007). Reduction of particle size should increase the supercial area available for mass transfer and, then, increase extraction yield. However, size

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208 Table 1 Comparison of literature works on optimisation of solvent extraction of phenolic constituents from grape and grape by-products Sample Entire dried seeds from red grape marc after vinication Variables Water content in ethanol: 54060 80% Extraction time (without stirring): 3 612 h Water content in ethylacetate: 3.35 101520% Extraction time 11224 h Crushing of marc Extraction time: 530 min Solvent: acetone, ethylacetate, methanol, ethylacetate with 1015 20% water Solvent: ethanol, methanol, water Solvent/solid ratio: 135 (v/w) Extraction time: 306090 min Extraction temperature: 2537.5 50 C In a continuous process: Solvent ow rate: 22.53 ml/min Sample amount: 2.557.5 g Particle size: 0.535 mm Solvent: water, 70% ethanol, 70% methanol Extraction time: 11224 h Water content of ethanol, methanol, and acetone Degreasing pre-treatment Solvent: 90% ethylacetateethanol Extraction time: 524 h Extraction temperature: 2860 C Inuence on Yield of catechins and proanthocyanidins Reference

201

Alonso et al. (1991)

Entire dried seeds from white grape marc after pressing Red grape marc after vinication Dried powdered defatted seeds from fresh red grape Dried red and white grape marc

Yield of proanthocyanidins

et al. Pekic (1998) Bonilla et al. (1999) Jayaprakasha et al. (2001) Pinelo et al. (2005a)

Yield of total phenols, concentration and reducing power of extract Yield of total avanols, concentration and antioxidant activity of extract Yield of total phenols, concentration and antioxidant activity of extract

Powdered white grape marc after distillation

Yield of total phenols, antioxidant activity of extract

Pinelo et al. (2005b)

Red grape pressed marc after vinication

Yield of total phenols and anthocyanins, concentration and antioxidant activity of extract Yield of total phenols and antioxidant activity of extract Yield of total phenols, concentration and antioxidant activity of extract

Lapornik et al. (2005) Yilmaz and Toledo (2006) Spigno and De Faveri (2007)

Seeds and skins from pressed marc before(white grape) and after vinication (red grape) Dried powdered red grape marc (after vinication) and stems

obtained after powdering of the sample was not always specied. Bonilla, Mayen, Merida, and Medina (1999) reported a higher extraction of phenolic compounds by acting on crushed than on uncrushed marc, but crushing , Kovac , Alonso, and Revilla details were omitted. Pekic (1998) wrote that grinding of grape seeds could shorten the extraction time but did not increase the yield of proanthocyanidins, and, furthermore, caused a signicant increase in the extraction of undesired concomitant components, so they used entire grape seeds, such as Alonso, Bourzeix, and Revilla (1991). Pinelo, Del Fabbro, Marzocco, Nunez, and Vicoli (2005a) investigated the eect of dierent particle size in a continuous phenol extraction, and they concluded that a higher amount of total polyphenols was obtained with the lower ow rate, sample amount and particle size. Actually, the authors expressed the yield as an index of mg phenols l1 h (the area under the polyphenol concentration curve as a function of time), but if these values are transformed into the yield of polyphenols (mg phenols/mg of sample), that is to say taking into account the solvent ow rate, very similar results for particle size of 0.5 and 5 mm are obtained. In the present paper, the particle size of 2 mm already used by Spigno and De Faveri (2007) was chosen and kept constant.

The eect of the solvent/sample ratio has been investigated by Pinelo et al. (2005a) and by other authors for dif kerget, & Knez, ferent raw materials (Herodez , Hadolin, S 2003; Cacace & Mazza, 2003): the higher the ratio, the higher the total amount of solids obtained, despite the solvent used, according to mass transfer principles. However, on the basis of the results reported in the cited works and of the extraction procedures summarised in Spigno and De Faveri (2007) and relative to Table 1, we retained to maintain constant the ratio of 4/1 (v/w) already used in our previous research. Type of solvent has been the most investigated factor. Ethylacetate was reported as one of the best solvent for extraction of polyphenols from grape seeds (it is capable of selectively extracting proanthocyanidins) and water addition up to a certain level (10%) increased proanthocyanidins yield because of increased permeability of grape seeds, but beyond this level signicant amount of et al., concomitant substances were extracted (Pekic 1998). On the other hand, alcoholic solvents have been commonly employed to extract phenolics from natural sources: they give quite high yield of total extract even though they are not highly selective for phenols. Particularly, mixtures of alcohols and water have

202

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208

revealed to be more ecient in extracting phenolic constituents than the corresponding mono-component solvent system (Yilmaz & Toledo, 2006; Pinelo et al., 2005a). Regarding grape and wine-making wastes, inuence of water content of alcohols has been studied for extraction from only seeds (Alonso et al., 1991; Yilmaz & Toledo, 2006) and, since the use of ethanol (a dietary alcohol) may be preferable than methanol in view of a food application of the extracts, in the present paper the eect of dierent aqueous ethanol mixture on extraction from grape marc was investigated. In fact, ethanol, a polar solvent, eectively extracts avonoids and their glycosides, catecols and tannins from raw plant materials (Bazykina, Nikolaevskii, Filippenko, & Kaloerova, 2002), but solubility of these compounds can be enhanced using a mixed solvent over a limited compositional range (Cacace & Mazza, 2003). Time and temperature of extraction are important parameter to be optimised even in order to minimise energy cost of the process. Many authors agree in the fact that an increase in the working temperature favours extraction enhancing both the solubility of solute and the diusion coecient, but also that beyond a certain value phenolic compounds can be denatured (Spigno & De Faveri, 2007; Pinelo, Rubilar, Jerez, Sineiro, & Nunez, 2005b; Yilmaz & Toledo, 2006). Spigno and De Faveri (2007) concluded that phenols yields at 60 C were higher than at 28 C, but an intermediate temperature of 45 C was selected for the next trials in order to verify if it could be possible to obtain the same result as for 60 C (or even better in case a certain degree of thermal degradation occurred at 60 C) with reduction of energy costs. More contradictory are the data available for extraction length: some authors chose quite short extraction times (Bonilla et al., 1999; Pinelo et al., 2005b; Yilmaz & Toledo, et al., 1998; Pinelo 2006); other quite long times (Pekic et al., 2005a; Lapornik, Prosek, & Wondra, 2005; Jayaprakasha, Singh, & Sakariah, 2001; Spigno & De Faveri, 2007). Furthermore, previous works reported opposing trend of phenols yield from grape marc: none signicant dierence between 5 and 24 h, at both 28 and 60 C (Spigno & De Faveri, 2007), signicant increase from 12 to 24 h at room temperature (Lapornik et al., 2005). That is why we decided to study extraction kinetics at both 60 C and 45 C. Finally, it was investigated whether the freeze-drying of the extracts might be used without leading to any phenols loss and/or degradation. 2. Materials and methods 2.1. Materials Pressed marcs (by Barbera red grape) were kindly provided from a wine-making factory in Piacenza (northern Italy) in 2004. They were collected after devatting, oven dried at 60 C up to moisture content of about

24% (determined by dry weight in oven at 105 C until constant weight) and milled through a 2 mm sieve (nal powder size 62 mm). All chemicals were of analytical grade. FolinCiocalteu reagent was purchased from Merck, quercetin and ABTS from SigmaAldrich Chemie GmbH, gallic acid, caeic acid and Trolox from Fluka. 2.2. Extraction of phenolic constituents Dried and milled marc was extracted with the solvent in a thermostatic rotary shaker (Infors AG, CH-4103 Bottmingen/Switzerland) with a 4/1 (v/w) ratio solvent/sample (wet weight of oven dried marc). The liquid extract was separated from solids by centrifugation (5350 g for 5 min, ALC 4237R centrifuge). In the rst phase of experimentation two temperatures (45 C and 60 C) and dierent maceration times (135 79152024 h) with absolute ethanol as solvent were tested. Liquid extracts were analysed for content of total phenols (FolinCiocalteu method), tannins and anthocyanins. In the second phase, extractions at 60 C for 5 h were repeated using absolute ethanol containing dierent volumes of water (102030405060%), and the liquid extracts were freeze-dried. Both liquid and freeze dried extracts were analysed for content of total phenols (according both to FolinCiocalteu method and direct reading at 280 nm), tannins, cinnamic acids and avonols. For each phenolic class, yield and extract concentration were calculated as: Yield (%): gphenols/100 gdried marc (dry weight) Phenols content (%): gphenols/100 gfreeze-dried extract Total extract yield (%) was calculated as gfreeze-dried extract/100 gdried marc (dry weight) Extracts were also characterized for their antioxidant power (ABTS assay). All the extraction trials were carried out in triplicate.

2.3. Chemical analyses 2.3.1. Total phenols Total phenols were determined according with two different methods: reau-Gayon, Glories, Mau(1) FolinCiocalteu (Ribe jean, & Dubourdieu, 2000). This method was used since it is the one adopted in almost all the published works about natural antioxidant recovery, being considered the best method for total phenolics (including tannins) determination (Engelhardt, 2001). (2) Direct reading of the absorbance of the sample at reau-Gayon et al., 2000). This a faster 280 nm (Ribe procedure based on the absorbance of the aromatic ring. Most catechins have a maximum absorption at

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208

203

around 280 nm, (+)-catechin is reported to be the major catechin monomer in all grape skins and total phenols and tannins contents are reported to be highly correlated with absorbance at 280 nm (Yilmaz & Toledo, 2006). In both cases total phenols were expressed as gallic acid equivalents (GAE) by means of calibration curves with standard gallic acid. 2.3.2. Tannins The applied analytical method (acid butanol assay) is based on the ability of monomer and condensed 34, avandiols to oxidise in acid and alcoholic medium at high temperature to give coloured procyanidins. Tannins were calculated by comparison with a standardised olig reau-Gayon et al., omeric procyanidin solution (Ribe 2000). 2.3.3. Anthocyanins Free anthocyanins and anthocyanins combined with tannins were measured through a chemical method based on their specic properties of bleaching by SO2, and calculated by comparison with a standardised anthocyanin solu reau-Gayon et al., 2000). tion (Ribe 2.3.4. Cinnamic acids and avonols Cinnamic acids were determined by reading absorbance of the sample at 320 nm and expressed as caeic acid equivalents (CAE) thorugh a calibration curve (Di Stefano & Cravero, 2001). Flavonols were determined by reading absorbance of the sample at 370 nm and expressed as quercetin equivalents (QE) through a calibration curve (Di Stefano & Cravero, 2001). 2.3.5. Antioxidant power Antioxidant power of the freeze dried extracts was assessed according to the ABTS assay (Re et al., 1999), which is based on the ability of antioxidants to interact with the radical ABTS decreasing its absorbance at 734 nm. Antioxidant power (AOP) was calculated as percentage inhibition:   ABlankt6 AExtractt6 %Inhibition 100 AABTSt0 where ABlank was the value of absorbance for the blank (ethanol), AExtract was the absorbance of the extract (dissolved in ethanol), t indicates the time (in min) at which absorbance was read. AOP was also converted into Trolox equivalents antioxidant activity (TEAC) by a calibration curve obtained with standard Trolox (115 lM nal concentration in the cuvette). TEAC is the ratio of mM Trolox to mM phenols in the extract (as GAE).

2.4. Statistical analysis The results reported in this paper are the averages of three replicates. Signicant variables were calculated, subjecting results to a linear regression, using SPSS statistical program version 11.5 at a condence level superior to 95% (P < 0.05). Dierence between means was evaluated by Tukeys post-hoc test. For statistical analysis all the percent data were transformed into arcsin values. 3. Results 3.1. Extraction kinetics Since Spigno and De Faveri (2007) had concluded that phenols yield from grape marc was higher at 60 C than at 28 C, and that it did not signicantly change from 5 h to 24 h, but, as explained in the Section 1, contradictory data are available from the literature, the aim of the rst part of experimentation was:  to verify whether an intermediate temperature between 28 C and 60 C could give the same recovery yield as 60 C (or even higher in case a certain degree of thermal degradation occurred at 60 C) in order to reduce the energy cost of the process;  to demonstrate if the equivalence of yield at 5 and 24 h was due either to the achieving of a plateau at 5 h, or to any antioxidants degradation taking place after an intermediate maximum. Statistical analysis indicated that both time and temperature highly inuenced antioxidants yields for all the measured compounds, with higher yields at 60 C (Fig. 1). Trend of anthocyanins extraction was not shown because yields were always lower than 0.012%, due to the levels of anthocyanins in grapes and to the fact that for red wines, they are mainly extracted during fermentation and vinication processes. Tukeys post-hoc test conrmed that yield increased with length of maceration but at 60 C, and, especially

Yield %

0 0 5 GAE 45C TAN 45C 10 15 20 25 GAE 60C Time (h) TAN 60C

Fig. 1. Yield of antioxidants for trials at dierent temperatures and times (error bars indicate SD).

204

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208

for anthocyanins and tannins, there seemed to be a reduction beyond 20 h due to thermal degradation, or polymerization which may inuence analytical quantication (Pinelo, Rubilar, Sineiro, & Nunez, 2005c). Eect of temperature cannot be generalised since it strongly depends on typology of compounds. For example, Cacace and Mazza (2003) found a maximum of 3035 C for extraction of anthocyanins from ribes with ethanol 85%; while Herodez et al. (2003) indicated 20 C, 0 C and 0 C for the highest yields of ethanolic extraction of carnosic acid, ursolic acid, and oleanolic acid, respectively, from Balm leaves. On the other hand, it was shown (Larrauri, Ruperez, & Saura-Calixto, 1997) that drying red grape pomace peels at 60 C did not signicantly aect the stability of polyphenols and antioxidant activity, and, indeed, it was reported an increase in the antioxidant capacity of grape extract by means of a simple thermal treatment at 60 C, due to phenols polymerization (Pinelo et al., 2005c; Kim et al., 2006). Considering the time eect, time was a signicant variable, particularly at 45 C and for total phenols, while at 60 C yields after 5 and 24 h were dierent but means fell into two adjacent homogenous groups (discrimination by means through Tukeys test, Table 2). Intermediate and higher values were reached in agreement with other litera et al., 1998) and many timetemperature ture works (Pekic combinations gave actually the same results. From a recovery point of view it would be more convenient to work at lower temperature for longer time. However, considering extraction rates of Fig. 1 and bearing in mind an industrial application of the process, it could be more convenient and energetically less expensive to work at higher temperature for shorter times (possibly <8 h). That is why extraction at 60 C for 5 h was selected, even though leading to lower yields. An accurate economical evaluation of the incidence of energy cost of the extraction stage on the overall produc-

tion cost per unit mass of nal extract will allow conrmation of this choice. The kinetic plots of antioxidants extraction from Balm leaves have been explained by the presence of two extraction stages (Herodez et al., 2003): an initial fast step corresponding to recovery of solutes from the supercial sites of the raw material, and a second lower step corresponding to molecular diusion of solutes from the internal sites through the porous medium. For both the steps, application of the steady-state model leads to the rst-order rate equation:   c1 ln k obs t c1 c where c is the concentration of the extracted constituent in the solution at time t, c1 is its concentration at equilibrium (t = 1), and kobs is the overall rate constant (s1). The three rate governing steps of the process are: surface-controlled infusion, diusion of the soluble constituents through the solid with a diusion coecient, and diusion of the constituents through the Nerst layer with another diusion coecient. Considering the second step alone is rate determining, the kobs takes into account the diusion coecient, the partition coecient of the extracted constituents between the solvent and the solid, the total surface area, the volume of solvent and the size and geometry of solid particles. The model, a quite simple one, was applied to phenols and tannins recovery, considering the equilibrium concentrations as those reached before the corresponding apparent reductions of Fig. 1. Obviously, plots in Fig. 2 refer to the second low step previously described (the tting equations do not origin from zero), and the model appears to be valid in describing experimental data, with an higher accuracy (smaller standard deviations) for the short extraction time range previously selected. Further kinetics inves-

Table 2 Means discrimination (phenols and tannins yields) by Tukeys test (a = 0.05, means with the same letter were not statistically dierent) h C GAE yield a 1 3 1 5 7 3 9 5 7 15 9 24 20 15 20 24 45 45 60 45 45 60 45 60 60 45 60 60 45 60 60 45 0.97 1.41 1.46 1.53 1.62 b c d e f g h i 1 24 3 1 3 5 7 5 24 7 9 9 15 15 20 20 45 45 45 60 60 45 45 60 60 60 45 60 45 60 45 60 h C Tannins yield a 0.67 0.81 1.06 1.08 1.17 1.25 1.30 1.31 1.08 1.17 1.25 1.30 1.31 1.32 b c d e f g h

1.62 1.82

1.82 1.92 2.02 2.03

2.02 2.03 2.19 2.21

2.03 2.19 2.21 2.28

2.19 2.21 2.28 2.32 2.35

1.17 1.25 1.30 1.31 1.32 1.42

2.21 2.28 2.32 2.35 2.47

1.25 1.30 1.31 1.32 1.42 1.47 1.47

1.32 1.42 1.47 1.47 1.59

2.47 2.65

1.42 1.47 1.47 1.59 1.63 1.65

1.59 1.63 1.65 1.81

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208

205

y = 5E-05x + 0.7863

1n[c/(cc)]

1n[c/(cc)]

R = 0.9766
2 1

y = 6E-05x + 0.3496 2 R = 0.988

y = 2E-05x + 0.4272 R = 0.9864

2

1
y = 3E-05x + 0.778 2 R = 0.9964

0 0 60 C GA E 20000 40000 45 C GA E 60000

0 0 60 C TA N 15000 30000 45 C TA N 45000

Time (sec)

Time (sec)

Fig. 2. First order plot for the slow stage of phenols and tannins extraction at 45 and 60 C (error bars indicate SD).

tigation relative also to the initial fast step should, then, provide a useful tool for scaling-up and process design. In order to further increase the recovery, we moved onto investigating the inuence of addition of water to ethanol, since, as summarized in the introduction, many literature references report the use of this kind of mixture. 3.2. Improvement of the process The following trials were carried out extracting at 60 C for 5 h and using ethanol containing dierent volumes of water (102030405060%), because mixtures of alcohols and water have revealed to be more ecient than the corresponding mono-component solvent system in extracting phenolic constituents from grape seeds (Alonso et al., 1991; Yilmaz & Toledo, 2006; Pinelo et al., 2005a). Extracts were analysed for total phenols, tannins, cinnamic acids and avonols content, before and after freeze-drying. Anthocyanins were not determined, considering the extremely low yields obtained in the previous phase. Increase of water content of ethanol was statistically inuent in improving extraction yield for GAE, GAE280, tannins, and total extract, but not for CAE and QE (Fig. 3).

5 12 4 9

3 2 1 0 0 GAE CAE 20 40 TAN QE 6 3

0 60 % H2O GAE 280 Extract

Fig. 3. Inuence of water content of ethanol on yields of phenolic compounds and total extract (error bars indicate SD).

The same yields were recalculated on the basis of the phenols content of freeze-dried samples. It must be pointed out that extracts had to be dissolved in the same mixture ethanol/water used for their extraction, in order to get a complete solubilization without heating assistance. Solubilization problems should be taken in consideration for nal employment of the extracts in food systems. For example, avonoids modication by lipase-catalysed esterication has been reported to improve their solubility in lipidbase media (Gayot, Santarelli, & Coulon, 2003; Kontogianni, Skouridou, Sereti, Stamatis, & Kolisis, 2003), and in natural antioxidants commercially exploited a proper medium, such as maltodextrine and propylene glycol, is incorporated to the same purpose. Statistical analysis (Students t-test for paired samples; a = 0.01) conrmed freeze-drying did not lead to reduction in phenols content, except for the cinnamic acids which seemed to get lost through freezedrying. Actually, they are very sensitive to oxidation processes both due to enzymes (such as tyrosinase which is easily co-extracted from marc) and oxygen. Liquid extracts were stored in closed asks but not under nitrogen, under refrigeration (oxygen solubility in water is increased at low temperature) and for a variable period of time before freeze-drying, and this may have brought to their rapid degradation. Fig. 3 shows also clearly that yields of total phenols based on GAE-280 were lower but highly correlated to those based on GAE (Pearsons correlation coecient +0.967, Sig. 0.000). The dierent results are due to the different principle of the two analytical methods: reaction with an oxidising reagent in the FolinCiocalteu analysis, absorption of the aromatic ring in direct reading at 280 nm. Tukeys post-hoc test (Table 3) conrmed that phenols yield was improved increasing the water percentage of ethanol from 10% to 30%, and, then, it did not signicantly changed for water contents between 30 and 60%, while total extract yield kept on increasing with water content. Similar trends were reported by other authors. Yilmaz and Toledo (2006) found out that phenol content (as GAE) of ethanol extracts from grape seed powder increased increasing water in the mixture from 0% to 30%, kept constant for 304050%, and decreased for higher percentage. Cacace and Mazza (2003) obtained that

Yield% Extract

Yield%

206

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208

Table 3 Means discrimination (phenols and tannins yields at dierent water content of ethanol) by Tukeys test (a = 0.05, means with the same letter were not statistically dierent) % H2O 10 20 30 40 50 60 Total extract a 6.04 6.4 b 6.4 7.38 9.1 9.35 12 c d 10 20 40 30 60 50 % H2O GAE a 2.45 2.94 4.05 4.13 4.15 4.25 b 10 20 30 50 40 60 % H2O GAE 280 a 1.4 1.53 b 1.53 1.91 c 10 20 60 30 50 40 % H2O Tannins a 1.71 2.62 3.8 3.88 4.11 4.15 b c

1.91 2.06 2.08 2.16

extraction of anthocyanins from black currants using aqueous ethanol increased with ethanol concentration up to a maximum at about 60% and then decreased with further increase in solvent concentration. In the same work it is suggested a dierent optimum ethanol content for the extraction of each group of phenols. According to the fact the total extract yield kept on increasing with water content, concentration of phenolic constituents in the extracts increased for water content from 10% to 30%, and decreased for water content above 50% (Fig. 4, Table 4). From comparison of literature works on phenols recovery from grape (Spigno & De Faveri, 2007), phenols content of our extracts were very high and comparable to those obtained from only grape seeds. Phenols concentra60 12

Phenols content %

6 20 3 0 0

Yield %

40

tion is always around 50%, this is because alcoholic and aqueous extracts (solvents not selective for phenols) from fruits processing by-products contain inevitably sugars and polysaccharides. From a commercial point of view, a higher purity would give a higher value to the extract. However, purication by C18 resins or other columns to eliminate sugars, non volatile acids and amino acids (Siriwoharn & Wrolstad, 2004) is an expensive step, inevitably separates free from bound polyphenolics, or dierent phenols classes (mixing of which may bring to antioxidant synergistic eect, Saint-Cricq de Gaulejac, Provost, & Vivas, 1999), while, depending on the extracts application (dietary supplies, antioxidants into food systems) sugars removal might be not necessary. Furthermore there are still many uncertainties about polyphenols eective bioavailability and metabolism, i.e., in vitro positive eects have been shown for concentrations that are not reached in vivo, and even though much has been learned about possible mechanisms of actions, it is largely unknown whether they can reach their multiple intended sites of action and whether they are better absorbed as aglycones or glycosides (Walle, 2004). 3.3. Antioxidant power of extracts Antioxidant power of freeze-dried extracts changed with the water content of ethanol exactly as the phenols content of extract (Fig. 5): this was probably due to the fact that ABTS assay was carried out initially with the same concentration of total freeze-dried extract, and extracts had dierent phenols content. To verify the actual trend of antioxidant power with the phenols concentration, AOP

0 GAE CAE

20

40 TAN QE

60 % H2O GAE 280 Extract

Fig. 4. Inuence of water content of ethanol on phenolic compounds content of freeze-dried extracts (error bars indicate SD).

Table 4 Means discrimination (phenols and tannins content of extracts at dierent water content of ethanol) by Tukeys test (a = 0.05, means with the same letter were not statistically dierent) % H2O 60 10 20 50 30 40 GAE a 33.8 38.8 b 38.8 39.3 45.1 47.1 47.4 c 60 10 20 50 30 40 % H2O 280 a 16.7 19.7 21.3 21.3 22.7 23.4 23.4 b c 60 10 20 30 50 40 % H2O Tannins a 32.3 38.2 38.6 b 38.2 38.6 44.8 45.1 c 60 50 40 20 10 30 % H2O QE a 0.9 1.29 1.36 1.37 1.41 1.47 b 60 10 50 20 40 30 % H2O CAE a 1.5 1.94 2.04 2.08 2.24 2.04 2.08 2.24 2.34 b c

44.8 45.1 48.2

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208

80 55 50

207

100 80

Phenols content %

60

AOP (%)

AOP %

45 40 40 20 35 30

60 40 Liquid extract 20 0 0 1 2 3 4 5 6 7 8 9 10 Freeze-dried extract

0
GAE

20
Tannins

40

60 % H2O GAE 280 AOP

GAE mg/L
Fig. 7. Antioxidant power of extracts before (liquid) and after freezedrying (error bars indicate SD).

Fig. 5. Antioxidant power as a function of extraction solvent and phenols content of extracts (error bars indicate SD).

was measured at dierent extracts dilutions. In Fig. 6a, the nal concentration of phenols in the cuvette was reported (extract in the cuvette resulted 100 folds diluted). Since there was no dierence between the various extracts, water content seemed only to inuence the amount of phenols recovered but not their nature. To compare our results with other similar literature works, % inhibition was transformed into TEAC values (Fig. 6b). TEAC value was concentration-dependent as observed in other works (Fukumoto & Mazza, 2000; Cao, Soc, & Prior, 1997; Saint-Cricq de Gaulejac et al., 1999; van den Berg, Haenen, van den Berg, & Bast, 1999), because in the considered range antioxidant activity linearly increased with concentration up to a certain value, above which the increase was lower or absent, while a linear Trolox calibration curve was used for TEAC calculation. Values of TEAC were comparable to those of many other phenolic compounds and vegetable extracts (Re et al., 1999; van den Berg et al., 1999; Yu, Ahmedna, & Goktepe, 2005; Arts, Dallinga, Voss, Haenen, & Bast, lez-Parama s, Esteban-Ruano, Santos2003). Only Gonza Buelga, de Pascual-Teresa, and Rivas-Gonzalo (2004) reported very high TEAC data (from 10 to 140) for grape marc extract. A screening trial assured that freeze-drying did not reduce antioxidant activity of extracts (Fig. 7).
100 80

4. Conclusions The aim of the present research was to get a better insight into optimisation of solvent extraction of antioxidants from grape marc, investigating some variables which were selected on the basis of the available literature about the same subject. Analysis of results brought to the following conclusions and aspects that should be further investigated.  For optimisation purposes direct reading of absorbance at 280 nm may be preferable than FolinCiocalteu method for total phenols evaluation.  In order to maximise recovery yield it could be better to work at 45 C rather than at 60 C for longer times, but from an economical point of view it would be advisable to work at 60 C for shorter times (preferably <8 h). This needs to be veried by evaluation of the energy cost of the extraction step on the overall production cost.  Beyond 20 h of extraction there was an apparent reduction in the amount of extracted phenols, but it is under investigation whether this was due to a real degradation or to polymerization reactions bringing new compounds with a dierent response to analytical measurements.  Trend of phenols yield could be well described by a rst order kinetics, particularly up to the suggested 5 h time. A detailed study on extraction rate (even for shorter
1.5

a
10% 30% 50% Trolox 20% 40% 60%

b
TEAC
1.0

AOP %

60 40 20 0 0 1 2 3 4

0.5

10% 40%
0.0 6 7 0 1 2

20% 50%
3

30% 60%
4 5 6

GAE mg/L

GAE mg/L

Fig. 6. Antioxidant power (A) and TEAC values (B) of extracts as a function of total phenols concentration.

208

G. Spigno et al. / Journal of Food Engineering 81 (2007) 200208 Kim, S.-Y., Jeong, S.-M., Park, W.-P., Nam, K. C., Ahn, D. U., & Lee, S.C. (2006). Eect of heating conditions of grape seeds on the antioxidant activity of grape seed extracts. Food Chemistry, 97, 472479. Kontogianni, A., Skouridou, V., Sereti, V., Stamatis, H., & Kolisis, F. N. (2003). Lipase-catalysed esterication of rutin and naringin with fatty acids of medium carbon chain. Journal of Molecular Catalysis B: Enzymatic, 21, 5962. Lapornik, B., Prosek, M., & Wondra, A. G. (2005). Comparison of extracts prepared from plant by-products using dierent solvents and extraction time. Journal of Food Engineering, 71, 214222. Larrauri, J. A., Ruperez, P., & Saura-Calixto, F. (1997). Eect of drying temperature on the stability of polyphenols and antioxidant activity of red grape pomace peels. Journal of Agricultural and Food Chemistry, 45, 13901393. , B., Kovac , V., Alonso, E., & Revilla, E. (1998). Study of the Pekic extraction of proanthocyanidins from grape seeds. Food Chemistry, 61(1/2), 201206. Pinelo, M., Del Fabbro, P., Marzocco, L., Nunez, M. J., & Vicoli, M. C. (2005a). Optimization of continuous phenol extraction from Vitis vinifera byproducts. Food Chemistry, 92, 109117. Pinelo, M., Rubilar, M., Jerez, M., Sineiro, J., & Nunez, M. J. (2005b). Eect of solvent, temperature, and solvent-to-solid ratio on the total phenolic content and antiradical activity of extracts from dierent components of grape pomace. Journal of Agricultural and Food Chemistry, 53, 21112117. Pinelo, M., Rubilar, M., Sineiro, J., & Nunez, M. J. (2005c). A thermal treatment to increase the antioxidant capacity of natural phenols: catechin, resveratrol and grape extract case. European Food Research and Technology, 221, 284290. Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., & RiceEvans, C. (1999). Antioxidant activity applying an improved abts radical cation decolorization assay. Free Radical Biology and Medicine, 26(9/10), 12311237. reau-Gayon, P., Glories, Y., Maujean, A., & Dubourdieu, D. (2000). Ribe Handbook of Enology. The chemistry of wine Stabilization and Treatments (vol. 2). West Sussex, England: John Wiley and Sons Ltd, pp. 157162. Saint-Cricq de Gaulejac, N., Provost, C., & Vivas, N. (1999). Comparative study of polyphenol scavenging activities assessed by dierent methods. Journal of Agricultural and Food Chemistry, 47, 425431. Siriwoharn, T., & Wrolstad, R. E. (2004). Polyphenolic composition of marion and evergreen blackberries. Journal of Food Science, 69(4), 233241. Spigno, G., & De Faveri, D. M. (2007). Antioxidants from grape stalks and marc: inuence of extraction procedure on yield, purity and antioxidant power of the extracts. Journal of Food Engineering, 78, 793801. van den Berg, R., Haenen, G. R. M. M., van den Berg, H., & Bast, A. (1999). Applicability of an improved Trolox equivalent antioxidant capacity (TEAC) assay for evaluation of antioxidant capacity measurements of mixtures. Food Chemistry, 66, 511517. Walle, T. (2004). Absorption and metabolism of avonoids. Free Radical Biology and Medicine, 36(7), 829837. Yilmaz, Y., & Toledo, R. T. (2006). Oxygen radical absorbance capacities of grape/wine industry byproducts and eect of solvent type on extraction of grape seed polyphenols. Journal of Food Composition and Analysis, 19, 4144. Yu, J., Ahmedna, M., & Goktepe, I. (2005). Eects of processing methods and extraction solvents on concentration and antioxidant activity of peanut skin phenolics. Food Chemistry, 90, 199206.

times) and its relationship with temperature and antioxidant power of extracts will allow scaling-up and design of the process.  Addition of water to ethanol improved extraction rate, but too high water content brought an increased concomitant extraction of other compounds, and, then to lower phenols concentrations in the extracts. Phenols extracted with dierent water content revealed the same antioxidant activity, suggesting that only dierent amounts but not dierent compounds were recovered.  Freeze-drying did not reduce antioxidant power of extracts, but further studies are necessary to assess AOP maintenance during extracts storage and to dene extracts solubility in dierent media.

References
Alonso, E., Bourzeix, M., & Revilla, E. (1991). Suitability of water/ ethanol mixtures for the extraction of catechins and proanthocyanidins from Vitis vinifera seeds contained in a winery by-product. Seed Science and Technology, 19, 542552. Arts, M. J. T. J., Dallinga, J. S., Voss, H.-P., Haenen, G. R. M. M., & Bast, A. (2003). A critical appraisal of the use of the antioxidant capacity (TEAC) in dening optimal antioxidant structures. Food Chemistry, 80, 409414. Bazykina, N. I., Nikolaevskii, A. N., Filippenko, T. A., & Kaloerova, V. G. (2002). Optimization of conditions for the extraction of natural antioxidants from raw plant materials. Pharmaceutical Chemistry Journal, 36(2), 100103. Bonilla, F., Mayen, M., Merida, J., & Medina, M. (1999). Extraction of phenolic compounds from red grape marc for use as food lipid antioxidants. Food Chemistry, 66, 209215. Cacace, J. E., & Mazza, G. (2003). Optimization of extraction of anthocyanins from black currants with aqueous ethanol. Journal of Food Science, 68(1), 240248. Cao, G., Soc, E., & Prior, R. L. (1997). Antioxidant and prooxidant behaviour of avonoids: structureactivity relationships. Free Radical Biology and Medicine, 22(5), 749760. Di Stefano, R., & Cravero, M. C. (2001). Metodi per lo studio dei polifenoli delluva. Rivista di Viticoltura ed Enologia, 2, 3745. Engelhardt, U. (2001). Flavonoids-analysis. Critical Review in Food Science and Nutrition, 41, 398399. Fukumoto, L. R., & Mazza, G. (2000). Assessing antioxidant and prooxidant activities of phenolic compounds. Journal of Agricultural and Food Chemistry, 48, 35973604. Gayot, S., Santarelli, X., & Coulon, D. (2003). Modication of avonoid using lipase in non-conventional media: eect of the water content. Journal of Biotechnology, 101, 2936. lez-Parama s, A. M., Esteban-Ruano, S., Santos-Buelga, C., de Gonza Pascual-Teresa, S., & Rivas-Gonzalo, J. C. (2004). Flavanol content and antioxidant activity in winery byproducts. Journal of Agricultural and Food Chemistry, 52, 234238. kerget, M., & Knez, Z. (2003). Solvent Herodez , S. S., Hadolin, M., S extraction study of antioxidants from Balm (Melissa ocinalis L.) leaves. Food Chemistry, 80(2), 275282. Jayaprakasha, G. K., Singh, R. P., & Sakariah, K. K. (2001). Antioxidant activity of grape seed (Vitis vinifera) extracts on peroxidation models in vitro. Food Chemistry, 73, 285290.