Food Chemistry 158 (2014) 534–545

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Improvements in the malaxation process to enhance the aroma quality of extra virgin olive oils P. Reboredo-Rodríguez, C. González-Barreiro, B. Cancho-Grande, J. Simal-Gándara ⇑ Nutrition and Bromatology Group, Analytical and Food Chemistry Department, Faculty of Food Science and Technology, University of Vigo, Ourense Campus, E-32004 Ourense, Spain

a r t i c l e

i n f o

Article history: Received 11 January 2014 Received in revised form 21 February 2014 Accepted 24 February 2014 Available online 6 March 2014 Keywords: Extra-virgin olive oil Malaxation Volatile compounds Odour activity value (OAV) Odorant series

a b s t r a c t The influence of olive paste preparation conditions on the standard quality parameters, as well as volatile profiles of extra virgin olive oils (EVOOs) from Morisca and Manzanilla de Sevilla cultivars produced in an emerging olive growing area in north-western Spain and processed in an oil mill plant were investigated. For this purpose, two malaxation temperatures (20/30 °C), and two malaxation times (30/90 min) selected in accordance with the customs of the area producers were tested. The volatile profile of the oils underwent a substantial change in terms of odorant series when different malaxation parameters were applied. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Virgin olive oil (VOO) is the principal source of fat above all in the Mediterranean diet (Aparicio & Harwood, 2003). Spain is a traditionally olive-growing country of the Mediterranean which produces and exports high quality VOO from a wide variety of cultivars. The Spanish olive grove is present in 34 of the 50 Spanish provinces and occupies an area of 2,509,677 ha. The areas of olive production in Spain – in descending order – are Andalusia (60.4%), Castilla-La Mancha (15.8%), Extremadura (10.2%), Catalonia (4.6%), Valencia (3.7%), Aragon (2.3%) and others (3.1%) including Galicia (AAO-Agencia para el aceite de oliva, 2013). Galicia (Northwestern Spain) was centuries ago, in the Middle Ages, a great producer of oil. Nowadays, there is a resurgence of this culture, from a family and artisan production to half-scale production. The oils produced in this area are thought to possess a characteristic aroma profile, but – to our knowledge – there are scarce data on their composition (Reboredo-Rodríguez, González-Barreiro, Cancho-Grande, & Simal-Gándara, 2012; Reboredo-Rodríguez, González-Barreiro, Cancho-Grande, & Simal-Gándara, 2013a, 2013b). VOO oil is highly appreciated by consumers for its healthy and sensory properties. The olfactory attributes of VOO arise mainly from the occurrence of C5 and C6 saturated and unsaturated aldehydes, alcohols and esters responsible for some typical sensory notes (such as ‘cut grass’, ‘fruity’ and ‘floral’). ⇑ Corresponding author. Tel.: +34 988 387000; fax: +34 988 387001. E-mail address: [email protected] (J. Simal-Gándara). http://dx.doi.org/10.1016/j.foodchem.2014.02.140 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.

The processing and characteristics of different quality olive oils are controlled and defined by the European Commission Implementing Regulation No 29/2012 (European Union Commission, 2012). The production of a high quality VOO is not only strongly dependent on the nature of the cultivar and the use of healthy olive fruit with an appropriate degree of maturity, but also it is influenced by other several factors like edaphoclimatic conditions, agricultural practices, extraction methods, processing techniques and/or storage conditions (Boselli, Di Lecce, Strabbioli, Pieralisi, & Frega, 2009; Inarejos-García, Gómez-Rico, Salvador, & Fregapane, 2009). The extraction process of VOO, consisting only of physical methods, includes olive crushing, malaxation of the pastes and separation of the oil phase. Malaxation of the olive paste, obtained previously from the olive crushing, is a crucial step of the process where the olive paste is subjected to a slow continuous kneading, aimed at breaking off the emulsions formed during the crushing process and facilitating adequate coalescence (Angerosa, Mostallino, Basti, & Vito, 2001; Stefanoudaki, Koutsaftakis, & Harwood, 2011). According to Clodoveo (2012) malaxation of olive paste must be considered much more than a simple physical separation, because a complex bioprocess takes place that is very relevant to the quality and composition of the final product. During malaxation considerable changes in the oil’s chemical composition occur because of the partition phenomena between oil and water and vice versa and the catalytic activity of fruit enzymes. Different enzymatic reactions of oxidoreductases naturally present in olive

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

pulp (such as polyphenoloxidases-PPO, lipoxygenases-LOX and peroxidases-POD) involving in the transformation of volatile and phenolic compounds take place (Boselli et al., 2009; Taticchi et al., 2013). The rate and the extent of these reactions are greatly affected by malaxation time and temperature, two technological parameters that can markedly modify not only the oil yield but also the composition and quality of the final VOO produced (Inarejos-García et al., 2009). In the present work, we have examined the effect of different malaxation operating conditions commonly used by Galician producers on the standard quality parameters, as well as the volatile profile of EVOOs obtained from two olive fruits, Manzanilla de Sevilla and Morisca. The aim of this work was to find the right combination of malaxation time and temperature in order to ensure the best quality of the resulting oils with a characteristic aroma.

2. Material and methods 2.1. Oil samples For this study, olives from two olive orchards with a different variety each were collected in November 2011 in the Southeast of Galicia (NW Spain). Each variety presented one maturation index (MI) according to the method proposed by the International Olive Oil Council (IOOC-International Olive Oil Council, 1984), based on the evaluation of the olive skin and pulp colours of the fruit. Two different oils were done: a mixture of Morisca and Verdial de Badajoz cultivars (90:10%) (MI = 3.4) and a mixture of Manzanilla de Sevilla and ‘unknown’ cultivars (95:5%) (MI = 2.1), with one in a higher proportion than the other. It should be noted the huge difficulty to obtain monovarietal oils in the olive oil mills of this area; a little percentage of a different olive variety is usually found in the olive batches processed. The oils were elaborated under identical conditions at a semiindustrial scale. Thus, all oil samples were processed in an oil mill plant (Almazara Profy, Industrias Céspedes e Hijos, S.L.) with a production capacity of 200 kg/h equipped with an olive washing machine, a hammer crusher, a horizontal kneader with a nonhermetic closure and a two-phase horizontal decanter. Leaves and dirt were removed by washing under cold running water before extraction. The olive paste corresponding to a mixture of Morisca and Verdial de Badajoz (90:10%) cultivars was kneaded at 20 ± 3 °C during 30 and 90 min, as well as 30 ± 2 °C during 30 and 90 min. On the other hand, the olive paste corresponding to a mixture of Manzanilla de Sevilla and ‘unknown’ (95:5%) cultivars were kneaded only at 30 ± 2 °C during 30 and 90 min. The temperature at three different points (left, centre and right of the malaxer) was checked by using an infrared thermometer at 10 min intervals. The conditions used in terms of temperature and time of malaxation were selected according to producers customs. Four oil samples were obtained for each set of conditions and were stored in dark-brown glass bottles without headspace at 10 °C in the dark. The samples were allowed to settle and racked for about 4 months. This is the procedure typically used by local producers before marketing their oil (Reboredo-Rodríguez et al., 2013b). Genetic and morphological determinations of representative olive samples were performed by the Pomology Group of the Department of Agronomy at the University of Cordoba (Spain), using fingerprinting based on Simple Sequence Repeat (SSR) markers. The results were used to characterise the studied cultivars. Accurately identified varieties included in the database of the World Olive Germplasm Bank of Cordoba (WOGB), which is the main repository of olive genotypes in Spain, were used as reference samples.

535

2.2. Analytical methods 2.2.1. Quality indices, fatty acids, sterols and erythrodiol+uvaol assays Standard quality indices (viz., free acidity; peroxide value; UV absorption characteristic, K270, K232; waxes, trilinolein and DECN), fatty acids, sterols and triterpenic dialcohols composition were carried out, following the analytical methods described by European Commission’s Regulation EEC/2568/91 and subsequent amendments (European Union Commission, 1991, 2003, 2007). The values of these parameters in different olive oils can be limited by regulations established by the European Union. Fatty acid assays were determined according to EEC/2568/91 and subsequent amendments (European Union Commission, 1991, 2003). Sterols and erythrodiol+uvaol were determined by following the procedures set out in Annexes V and VI of Regulation EEC/2568/91 and subsequent amendments (European Union Commission, 1991, 2003). 2.2.2. Extraction and identification of volatiles Volatile compounds were extracted from the EVOO samples by Dynamic Headspace (DHS) with an automatic sampler device, the Master DHS (DANI Instruments S.p.A., Cologno Monzese, Milan, Italy), following our previous work (Reboredo-Rodríguez et al., 2012). In short, the samples (volume: 9.0 mL of EVOO, fast shaking) were directly placed into the DHS sampler in standard 20 mL vials that can be heated at a chosen temperature (40 °C). An inert gas flow (He, flow: 150 mL/min) was used to purge the sample (time: 90 min) in order to carry out the volatile compounds; then the purged gas passed through a cooled trap (0 °C) where the compounds were concentrated. The trap was quickly heated in backflush to a high preset temperature (250 °C) transferring the compounds to the chromatographic column in a narrow band and a reduced volume of gas. The Master DHS has a device especially designed to remove humidity from the desorbed gas before entering into the GC–MS system. It was kept at low temperature during the injection phase and was heated during the baking step to eliminate retained water. Afterwards, volatile compounds were separated and identified on a Trace GC gas chromatograph with a PolarisQ ion trap mass selective detector (ITMS) interfaced to a PC computer running the software Xcalibur 1.4, from Thermo Finnigan (Rodano, Italy). Chromatographic separations were done with a ZB-WAX fusedsilica capillary column (60 m  0.32 mm ID, 0.50 lm film thickness, Phenomenex, Torrance, CA, USA). The carrier gas, helium, was circulated at 1 ml/min in the constant flow mode. A split/splitless injector in the split mode was used (split ratio, 1:10). The injector temperature was 200 °C. The oven temperature programme was as follows: 40 °C for 5 min; 2 °C/min ramp to 125 °C; 10 °C/min ramp to 250 °C and holding for 5 min. The transfer line temperature was 250 °C, and the ion trap manifold temperature 200 °C. The ion energy for electron impact (EI) was set constantly at 70 eV. Identification of the volatile compounds was achieved by comparing the GC retention times and mass spectra over the mass range 35–300 amu for the samples with those for pure standards analysed under the same conditions. Mass detection was performed in the selected ion recording (SIR) mode for quantification. The concentration of volatile compounds in EVOO samples was calculated taking into account the method recoveries (Reboredo-Rodríguez et al., 2012). 2.3. Calculation of the odorant series values An odorant series is defined as a group of volatile compounds with similar aroma descriptors (Peinado, Mauricio, & Moreno, 2006). The total intensities for every odorant series were calculated as sum of the odour activity value (OAV) (defined as concentration

536

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

Table 1 Standard quality indices (free acidity; peroxide value; UV absorption characteristic, K270, K232, DK; waxes, trilinolein and DECN), fatty acids, sterols and triterpenic dialcohols composition of the studied EVOOs. Parameter

20 °C

A. Morisca+Verdial de Badajoz (90:10%) Acidity (%) Peroxides (meq. O2/kg oil)* K232 K270 DK Waxes (mg/kg) Trilinolein (%) DECN

30 min 0.45 ± 0.01a 6.11–9.11 1.68 ± 0.19a 0.16 ± 0.03a 0.0021 ± 0.0000a 44.1 ± 1.6a 0.33 ± 0.12a 0.17 ± 0.02a

90 min 0.37 ± 0.01b 5.90–9.38 1.69 ± 0.31a 0.19 ± 0.01a 0.0054 ± 0.0047ab 49.6 ± 4.9a 0.40 ± 0.04a 0.14 ± 0.01a

30 °C 30 min 0.42 ± 0.00ab 6.66–8.79 1.71 ± 0.22a 0.15 ± 0.01a 0.0022 ± 0.0001a 50.7 ± 6.2a 0.40 ± 0.06a 0.16 ± 0.05a

90 min 0.52 ± 0.02c 7.48–14.70 1.86 ± 0.39a 0.20 ± 0.01a 0.0130 ± 0.0000b 54.0 ± 0.0a 0.33 ± 0.07a 0.10 ± 0.13a

Fatty acid composition by GC (% m/m methyl esters) Myristic C14:0 Palmitic C16:0 Palmitoleic C16:1 Margaric C17:0 Margaroleic C17:1 Stearic C18:0 Oleic C18:1 Linoleic C18:2 Linolenic C18:3 Arachidic C20:0 Eicosenoic C20:1 Behenic C22:0 Lignoceric C24:0 trans oleics trans L+Ln

0.013 ± 0.002a 12.88 ± 0.00a 0.92 ± 0.02ab 0.061 ± 0.001a 0.098 ± 0.003a 3.04 ± 0.13a 67.99 ± 0.31a 13.10 ± 0.07a 0.91 ± 0.04a 0.48 ± 0.03a 0.29 ± 0.01a 0.17 ± 0.02a 0.063 ± 0.010a 0.015 ± 0.007a 0.025 ± 0.007a

0.009 ± 0.001a 12.76 ± 0.19a 0.88 ± 0.01a 0.056 ± 0.006a 0.088 ± 0.004a 2.97 ± 0.04a 68.14 ± 0.25a 13.24 ± 0.02ab 0.92 ± 0.01a 0.48 ± 0.00a 0.28 ± 0.01a 0.15 ± 0.00a 0.053 ± 0.010a Traces 0.020 ± 0.000a

0.010 ± 0.001a 12.88 ± 0.23a 0.94 ± 0.01ab 0.067 ± 0.004a 0.097 ± 0.005a 2.99 ± 0.04a 67.74 ± 0.28a 13.41 ± 0.01b 0.91 ± 0.01a 0.49 ± 0.01a 0.28 ± 0.00a 0.15 ± 0.00a 0.068 ± 0.004a 0.020 ± 0.000a 0.027 ± 0.004a

0.011 ± 0.001a 12.98 ± 0.08a 0.96 ± 0.01b 0.064 ± 0.005a 0.101 ± 0.001a 2.98 ± 0.00a 67.92 ± 0.04a 13.16 ± 0.06a 0.89 ± 0.01a 0.46 ± 0.00a 0.28 ± 0.00a 0.15 ± 0.01a 0.059 ± 0.001a 0.011 ± 0.000a 0.020 ± 0.000a

Sterols by GC relative amount (%) Cholesterol Brassicasterol Campesterol Stigmasterol Apparent b-Sitosterol D7-Stigmastenol Total sterols (lg/g) Erythrodiol+Uvaol

0.053 ± 0.010a Traces 2.33 ± 0.01a 1.14 ± 0.01a 95.60 ± 0.31a 0.12 ± 0.05a 1600.0 ± 34.0a 3.29 ± 0.08a

0.061 ± 0.027a Traces 2.39 ± 0.02b 1.02 ± 0.01b 95.64 ± 0.20a 0.15 ± 0.01b 1342.6 ± 146.6a 3.13 ± 0.01a

0.069 ± 0.001a Traces 2.35 ± 0.01ab 1.12 ± 0.01a 95.46 ± 0.24a 0.18 ± 0.08a 1616.1 ± 55.1a 3.48 ± 0.16a

0.062 ± 0.002a Traces 2.32 ± 0.01a 1.33 ± 0.03c 95.46 ± 0.23a 0.10 ± 0.00c 1596.5 ± 40.3a 3.63 ± 0.19a

Parameter

30 °C

B. Manzanilla de Sevilla+Unknown (95:5%) Acidity (%) Peroxides (meq. O2/kg oil)* K232 K270 DK Waxes (mg/kg) Trilinolein (%) DECN

30 min 0.25 ± 0.01a 2.24–5.23 1.45 ± 0.11a 0.12 ± 0.03a 0.0036 ± 0.0002a 34.5 ± 3.5a 0.07 ± 0.01a 0.06 ± 0.03a

90 min 0.28 ± 0.05a 1.94–4.35 1.38 ± 0.13a 0.13 ± 0.01a 0.0034 ± 0.0001a 42.9 ± 4.5a 0.11 ± 0.01a 0.13 ± 0.06a

Fatty acid composition by GC (% m/m methyl esters) Myristic C14:0 Palmitic C16:0 Palmitoleic C16:1 Margaric C17:0 Margaroleic C17:1 Stearic C18:0 Oleic C18:1 Linoleic C18:2 Linolenic C18:3 Arachidic C20:0 Eicosenoic C20:1 Behenic C22:0 Lignoceric C24:0 trans oleics trans L + Ln

0.014 ± 0.002a 10.80 ± 0.15a 0.86 ± 0.01a 0.180 ± 0.057a 0.280 ± 0.141a 3.23 ± 0.08a 78.80 ± 0.00a 4.11 ± 0.08a 0.80 ± 0.06a 0.43 ± 0.15a 0.30 ± 0.01a 0.14 ± 0.04a 0.072 ± 0.005a 0.023 ± 0.003a Traces

0.011 ± 0.001a 11.15 ± 0.05a 0.87 ± 0.01a 0.210 ± 0.014a 0.285 ± 0.134a 3.05 ± 0.08a 77.66 ± 0.03b 4.96 ± 0.06b 0.81 ± 0.01a 0.43 ± 0.12a 0.33 ± 0.01a 0.14 ± 0.01a 0.116 ± 0.062a 0.022 ± 0.013a Traces

Sterols by GC relative amount (%) Cholesterol Brassicasterol Campesterol Stigmasterol Apparent b-Sitosterol D7-Stigmastenol Total sterols (lg/g) Erythrodiol+Uvaol

0.050 ± 0.013a Traces 2.30 ± 0.01a 0.54 ± 0.02a 96.52 ± 0.08a 0.12 ± 0.01a 1443.3 ± 54.2a 2.02 ± 0.02a

0.068 ± 0.025a Traces 2.61 ± 0.01b 0.62 ± 0.00b 96.06 ± 0.09b 0.15 ± 0.01a 1503.4 ± 133.5a 2.00 ± 0.00a

Values are mean ± standard deviation (n = 2). Different letters within rows indicate statistical differences as per ANOVA (p < 0.05) and Tukey’s HSD test. Peroxide values were determined three months apart.

*

537

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545 Table 2 Concentrations of volatile compounds of the studied EVOOs. A. Morisca:Verdial de Badajoz (90:10%) Volatile compound

Odour descriptor*

Odorant series (Odour threshold; ng/g)

Concentration (ng/g) 20 °C/ 30 min

20 °C/ 90 min

30 °C/ 30 min

30 °C/ 90 min

C6 compounds trans-2-Hexenal

Lawn; green-apple like; bitter almonds

Grass (1125); apple (424); bitter-like (420)

Apple; leaf-like Green, bitter Green grass, leaves; green, grassy, sweet n.f.

Apple (6000); leaf (1100) Bitter-like (1500) Grass (8000); leaf (5000)

Olive fruity; green notes; banana

Grass (750); olive fruit (750); banana (200)

C6/LnA-Aldehydes cis-3-Hexen-1-ol trans-3-Hexen-1-ol trans-2-Hexen-1-ol cis-2-Hexen-1-ol C6/LnA-Alcohols cis-3-Hexenyl acetate trans-2-Hexenyl acetate C6/LnA-Esters

2037 ± 207a 1867 ± 217a 2204 ± 193a 1096 ± 87b 2037

n.f.

1867

2204

1096

1003 ± 74ab 1104 ± 160a 1011 ± 43ab 897 ± 58b 12.2 ± 1.1a 13.3 ± 1.6a 11.8 ± 0.8a 13.7 ± 1.0a 2162 ± 147a 1787 ± 188b 531 ± 29c 2586 ± 105d 7.7 ± 0.8a 3185

11.0 ± 1.5b 2915

4.7 ± 0.3c 1559

14.1 ± 1.2d 3511

667 ± 43a

764 ± 94a

764 ± 80a

210 ± 29b

4.9 ± 0.5ab

5.9 ± 0.9bc

4.7 ± 0.4a

6.0 ± 0.4c

672

770

769

216

Hexanal C6/LA-Aldehydes

Grass; green apple; green-sweet

Grass (300); apple (80); sweet-like (75)

415 ± 38a 415

412 ± 48a 412

455 ± 56a 455

284 ± 28b 284

1-Hexanol C6/LA-Alcohols

Fruit; banana, soft

Olive fruit (400); banana (400)

1574 ± 120a 2360 ± 229b 986 ± 34c 1574 2360 986

1840 ± 137a 1840

Hexyl acetate

Green; fruity; sweet

Grass (1040); olive fruit (1040); sweet-like (1040)

113 ± 9a

134 ± 9b

118 ± 7a

57.4 ± 3.4c

113

134

118

57.4

C6/LA-Esters C5 compounds trans-2-Pentenal

Green, apple; green, bitter almond

Apple (300); bitter-like (300)

C5/LnA-Aldehydes 1-Penten-3-ol

Lawn; olives; leaf; pungent

cis-2-Penten-1-ol

Banana; sweet; green fruity; fresh olive fruits n.f.

trans-2-Penten-1ol

Grass (400); olive fruit (400); leaf (400); pungent-like (400) Banana (250); sweet-like (250); olive fruit (250)

C5/LnA-Alcohols 1-Penten-3-one

Leaf; bitter; pungent

Leaf (50); bitter-like (50); Pungent-like (50)

C5/LnA-Ketones

17.7 ± 1.5a

15.5 ± 1.7a

24.4 ± 1.3b

12.1 ± 0.7c

17.7

15.5

24.4

12.1

650 ± 84a

624 ± 64a

765 ± 88a

634 ± 40a

160 ± 7a

168 ± 16a

156 ± 27a

167 ± 19a

44.3 ± 1.1ac

40.9 ± 2.8a

52.1 ± 2.3b

49.7 ± 5.3bc

854

833

973

851

205 ± 6a

173 ± 14a

589 ± 49b

24.2 ± 3.7c

205

173

589

24.2

Pentanal C5/LA-Aldehydes

Woody; bitter; oily

Wood (240); Bitter-like (240)

102 ± 8ac 102

121 ± 9a 121

176 ± 16b 176

92.9 ± 6.0c 92.9

1-Pentanol C5/LA-Alcohols

Fruity; strong, sticky, balsamic

Olive fruit (470); pungent-like (3000)

55.7 ± 5.9a 55.7

36.9 ± 4.7b 36.9

30.1 ± 2.7b 30.1

89.1 ± 6.7c 89.1

3-Pentanone

Fruity; green; sweet

Grass (70,000); olive fruit (70,000); sweet-like (70,000)

517 ± 37ac

448 ± 43a

340 ± 17b

565 ± 64c

C5/LA-Ketones

517

448

340

565

R Total compounds

9747

10,085

8223

8639

B. Manzanilla de Sevilla:Unknown (95:5%) Volatile compound Odorant series (Odour threshold; ng/g)

Concentration (ng/g) 30 °C/ 30 min

30 °C/ 90 min

C6 compounds trans-2-Hexenal C6/LnA-Aldehydes

Grass (1125); apple (424); bitter-like (420)

820 ± 73a 820

1733 ± 28b 1733

cis-3-Hexen-1-ol trans-3-Hexen-1-ol trans-2-Hexen-1-ol cis-2-Hexen-1-ol C6/LnA-Alcohols

Apple (6000); leaf (1100) Bitter-like (1500) Grass (8000); leaf (5000)

2564 ± 103a 112 ± 4a 2713 ± 132a 21.0 ± 2.3a 5410

2743 ± 132a 142 ± 9b 3985 ± 189b 33.3 ± 3.3b 6903

cis-3-Hexenyl acetate trans-2-Hexenyl acetate C6/LnA-Esters

Grass (750); olive fruit (750); banana (200)

2237 ± 63a 12.3 ± 1.3a 2250

121 ± 11b 5.4 ± 1.0b 127

Hexanal C6/LA-Aldehydes

Grass (300); apple (80); sweet-like (75)

537 ± 11a 537

751 ± 32b 751

1-Hexanol

Olive fruit (400); banana (400)

5781 ± 172a

5857 ± 112a (continued on next page)

538

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

Table 2 (continued) B. Manzanilla de Sevilla:Unknown (95:5%) Volatile compound Odorant series (Odour threshold; ng/g)

Concentration (ng/g) 30 °C/ 30 min

30 °C/ 90 min

5781

5857

Grass (1040); olive fruit (1040); sweet-like (1040)

216 ± 8a 216

31.3 ± 0.8b 31

Apple (300); bitter (300)

8.1 ± 0.4a 8

15.7 ± 0.7b 16

1-Penten-3-ol cis-2-Penten-1-ol trans-2-Penten-1-ol C5/LnA-Alcohols

Grass (400); olive fruit (400); leaf (400); pungent-like (400) Banana (250); sweet-like (250); olive fruit (250)

885 ± 125a 262 ± 36a 54.5 ± 7.4a 1201

907 ± 143a 358 ± 59b 83.9 ± 12.3b 1348

1-Penten-3-one C5/LnA-Ketones

Leaf (50); bitter-like (50); pungent-like (50)

41.0 ± 4.2a 41

53.6 ± 2.5b 54

Pentanal C5/LA-Aldehydes

Wood (240); bitter-like (240)

156 ± 3a 156

74.3 ± 4.7b 74

1-Pentanol C5/LA-Alcohols

Olive fruit (470); pungent-like (3000)

115 ± 12a 115

130 ± 6a 130

3-Pentanone C5/LA-Ketones

Grass (70,000); olive fruit (70,000); sweet-like (70,000)

996 ± 83a 996

942 ± 74a 942

17,531

17,996

C6/LA-Alcohols Hexyl acetate C6/LA-Esters C5 compounds trans-2-Pentenal C5/LnA-Aldehydes

R Total compounds

Values are mean ± standard deviation (n = 4). Different letters within rows indicate statistical differences as per ANOVA (p < 0.05) and Tukey’s HSD test. n.f.: not found.

of volatile compound in oil sample/odour threshold) of each compound assigned to this series. According to its aroma descriptors, a volatile compound can be included in one or several odorant series. The odour descriptors of C6 and C5 compounds that contribute to the sensory properties of EVOOs were taken from literature (Angerosa, Mostallino, Basti, & Vito, 2000; Kalua et al., 2007; Tura, Failla, Bassi, Pedò, & Serraiocco, 2008); and grouped in 9 different odorant series: grass, leaf, wood, bitter-like, sweet-like, pungentlike (or rasping), olive fruit, apple and banana. Some compounds were included in two or more odorant series. 2.4. Statistical analysis Analysis of variance (ANOVA) was carried out using the statistical package Statgraphics Centurion XV for windows Version 15.2.06 (Statistical Graphics Corp., Herndon, Va, USA). Tukey’s HSD test was used as a single-step multiple comparison method in conjunction with ANOVA to identify significantly different means. Partial least squares (PLS) regression was implemented by using the statistical package Unscrambler v. 9.1 for Windows (CAMO Software, Oslo, Norway). 3. Results and discussion 3.1. Quality indices, fatty acid, sterol and triterpene dialcohols composition of EVOOs extracted at different malaxation conditions It is of great importance to deepen in the effect of the malaxation conditions on the quality of the VOO produced in order to strike an appropriate balance between the economic turnover and the quality of the commercial product (Inarejos-García et al., 2009). The analysed quality indices of the olive oil samples in the different malaxation conditions studied met the standards of the European Community for the classification as ‘extra virgin’ category (European Union Commission, 1991, 2011). All the oils from Morisca and Manzanilla de Sevilla analysed showed values below the upper limits established by EU Regulations for the extra virgin olive oil

category (acidity 6 0.8°; peroxide index 6 20 meq O2/kg; K270 6 0.22; K232 6 2.5; DK 6 0.01; Waxes 6 250 mg/kg; Trilinolein < 0.5%; DECN 6 0.2). There was an increase, in some cases significant, when either the malaxation temperature and/or time was risen (Table 1a, b) such it was observed in the case of Cornicabra variety (InarejosGarcía et al., 2009). Ranalli, Contento, Schiavone, and Simone (2001) attributed this behaviour to the increase in activity of the lipase enzymes (responsible for the increase of free acidity) and to an intensification of the primary oxidation processes (responsible for the increase of the K232 and peroxide index values) and the secondary oxidation processes (responsible for the increase of the K270 and carbonyl index values) when malaxation temperature rises. The fatty acid levels, as well as being in accordance with the limits mentioned in the EU Regulation for ‘extra virgin’ category (European Union Commission, 1991, 2003), were not significantly influenced by both malaxation parameters (time and temperature) for both oil varieties (Morisca and Manzanilla de Sevilla) (Table 1a, b). Concerning sterols and the sum of erythrodiol and uvaol, expressed as sum of percentage of total sterols, it should be noted that for both cultivars, Morisca and Manzanilla de Sevilla, no significant differences in total sterol composition and erythrodiol+uvaol of oil samples were detected (Table 1a, b). Nevertheless, a trend is observed by evaluating the effect of malaxation temperature in Morisca oils, higher contents of total sterols and triterpene dialcohols were observed when olive paste was malaxed at the highest temperature (30 vs. 20 °C) (Table 1a). According to Allouche et al. (2010), this rise could be explained by the fact that higher temperatures decrease the oil viscosity, thus the extraction of these compounds from the olive paste is favoured.

3.2. Effect of malaxation process on the volatile profile of EVOOs Three branches of volatile C6 metabolites are generated from linolenic acid (LnA) and linoleic acid (LA) through the lipoxygenase (LOX) pathway. LOX transforms LnA and LA into their corresponding 9- and 13-hydroperoxides, in a ratio ranging between 65:35

539

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

(a)

20 ºC

C6 compounds

9000 8000

30 ºC

7000

20 ºC

ng/g oil

6000

30 ºC

5000 4000

30 ºC

20 ºC

3000 2000 1000 0 30 min

90 min

30 min

90 min

30 min

90 min

LnA compounds

30 min

Aldehydes

2500

90 min

30 min

90 min

LA compounds

(b)

Alcohols

30 min

90 min

Total compounds Esters

C5 compounds

30 ºC

20 ºC

2000 30 ºC

ng/g oil

1500 20 ºC 20 ºC

1000

30 ºC

500

0 30 min

90 min

30 min

90 min

30 min

90 min

LnA compounds

30 min

LA compounds Aldehydes

Alcohols

90 min

30 min

90 min

30 min

90 min

Total compounds

Ketones

Fig. 1. (a) Sum of C6 volatile compounds in Morisca:Verdial de Badajoz (90:10%) oils at different malaxation process conditions: Temperature: 20/30 °C; Time: 30/90 min. (b) Sum of C5 volatile compounds in Morisca:Verdial de Badajoz (90:10%) oils at different malaxation process conditions: Temperature: 20/30 °C; Time: 30/90 min. (c) Sum of C6 volatile compounds in Manzanilla de Sevilla:Unknown (95:5%) oils at different malaxation process conditions: Temperature: 30 °C; Time: 30/90 min. (d) Sum of C5 volatile compounds in Manzanilla de Sevilla:Unknown (95:5%) oils at different malaxation process conditions: Temperature: 30 °C; Time: 30/90 min. Volatile content of C6/LnA aldehydes is trans-2-hexen-1-al. Volatile contents of C6/LnA alcohols are the sum of: cis-2-hexen-1-ol, trans-2-hexen-1-ol, cis-3-hexen-1-ol and trans-3-hexen-1-ol. Volatile contents of C6/LnA esters are the sum of: cis-3-hexenyl acetate and trans-2-hexenyl acetate. Volatile content of C6/LA aldehydes is hexanal. Volatile content of C6/LA alcohols is hexanol. Volatile content of C6/LA esters is hexyl acetate. Volatile content of C5/LnA aldehydes is trans-2-pentenal. Volatile contents of C5/LnA alcohols are the sum of: 1penten-3-ol, cis-2-penten-1-ol and trans-2-penten-1-ol. A volatile content of C5/LnA ketones is 1-penten-3-one. Volatile content of C5/LA aldehydes is pentanal. Volatile content of C5/LA alcohols is 1-pentanol. Volatile content of C5/LA ketones is 3-pentanone.

and 55:45, respectively. Only the 13-hydroperoxides, from both LnA (13-HPOT) and LA (13-HPOD), are cleaved by hydroperoxide lyase (FAHL) into C12 oxo-acids, cis-3-hexenal and hexanal, as the enzyme has a high substrate specificity. Enzymatic transformation of the two aldehydes mediated by isomerases (IR), alcohol dehydrogenases (ADH) and alcohol acetyl transferases (AAT) yields the corresponding C6 esters and C6 alcohols. An additional branch of short-chain green volatiles, including oxygenated C5 compounds, is biosynthesised through another LOX pathway. In this case, 13-HPOT undergoes a b-scission yielding pentene dimers and pentenols through the alkoxyl radical. The subsequent oxidation of pentenols catalysed by an alcohol dehydrogenase yields C5 carbonyl compounds (Ranalli, Pollastri, Contento, Iannucci, & Lucera, 2003).

3.2.1. Effect of malaxation time In our oils, an opposite behaviour for the most abundant C6 aldehydes, trans-2-hexenal and hexanal, is observed taking account the effect of time: in Morisca oils C6 aldehydes decreased in concentration with increasing times of olive paste kneading, such decreases were statistically significant (p < 0.05) at 30 °C (Table 2a, Fig. 1a). In the case of Manzanilla de Sevilla oils, the trend differed completely (Table 2b, Fig. 1c) and was in good agreement with other varieties such as the Italian Coratina and Frantoio (Angerosa et al., 2001), Leccino, Dritta and Caroleo (Ranalli et al., 2003), the Spanish Cornicabra (Gómez-Rico, Inarejos-García, Salvador, & Fregapane, 2009) and the Tunisian Chemlali and Chetoui (Youssef et al., 2013). With these results, it is clear that the behaviour of enzymes pattern during malaxation is linked to the variety.

540

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

C6 compounds 18000

(c)

Total compounds

16000 14000

ng/ g oil

12000

LnA compounds

LA compounds

10000 8000 6000 4000 2000 0 30 min

90 min

30 min

90 min

Aldehydes

Alcohols

30 min

90 min

Esters

C5 compounds 3000

(d) Total compounds

2500

2000

ng/ g oil

LnA compounds

LA compounds

1500

1000

500

0 30 min

90 min

30 min

Aldehydes

90 min

Alcohols

30 min

90 min

Ketones

Fig. 1 (continued)

The content of C6 alcohols also experienced an increase with lengthy times for both varieties except trans-2-hexen-1-ol at low temperatures (20 °C) for Morisca oils (Table 2a, b, Fig. 1a, c). A considerable decrease was detected for C6 esters at 30 °C in both cultivars, especially for cis-3-hexenyl acetate (Table 2a, b, Fig. 1a, c). Ranalli et al. (2003) attributed this behaviour to a progressive inactivation of AAT. The concentrations of individually C5 compounds were influenced differently by time adopted during the malaxation for both varieties: for Morisca oils, ketones were the compounds that experienced the most significant changes, only at 30 °C. Thus, the concentrations of 1-penten-3-one diminished drastically while the amount of 3-pentanone increased (Table 2a). Nevertheless, the concentration of ketones in Manzanilla de Sevilla oils rarely varied; on the other hand, the amount of LnA-alcohols in these oils,

especially cis-2-penten-1-ol, was higher at prolonged malaxation times (Table 2b). In general terms, the total concentration of C6 volatiles increased with extended times of olive paste kneading for both studied oils regardless of temperature (Fig. 1a, c). For C5 volatiles, this behaviour was maintained in the case of Manzanilla de Sevilla oils (Fig. 1d) but not in the case of Morisca oils (Fig. 1b): C5 compounds tended to decrease at high times. 3.2.2. Effect of malaxation temperature Changes in malaxation temperature (at 20 and 30 °C) were evaluated only with one cultivar (Morisca). According to Salas and Sánchez (1999) the malaxation temperature generally causes a decrease of levels of volatile compounds from LOX pathways, as a consequence of proved inactivation of hydroperoxide lyases. This

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

541

Fig. 2. (a) One-dimension PLS-2: scores plot for Morisca:Verdial de Badajoz (90:10%) oils extracted at 20/30 °C for 30/90 min (x), together with loadings plot of X-variables for the 18 volatile descriptors (y) and loadings plot of Y-variables for the 9 aroma descriptors (z). One principal component accounted for 86% and 78% of the variance in X and Y data, respectively. (b) One-dimension PLS-2: scores plot for Manzanilla de Sevilla:Unknown (95:5%) oils extracted at 30 °C for 30/90 min (x), together with loadings plot of Xvariables for the 18 volatile descriptors (y) and loadings plot of Y-variables for the 9 aroma descriptors (z). One principal component accounted for 97% of the variance in both X and Y data, respectively.

Fig. 2 (continued)

trend was evidenced by other authors as well (Angerosa et al., 2001; Gómez-Rico et al., 2009; Kalua, Bedgood, Bishop, & Prenzler, 2006; Ranalli et al., 2001). With prolonged times (90 min) most C6-compounds underwent statistically significant variations when temperature is increased. In this sense, C6-aldehydes, alcohols and esters coming

from LnA (trans-2-hexenal, cis-3-hexen-1-ol, cis-3-hexenyl acetate, respectively) and LA (hexanal, 1-hexanol and hexyl acetate, respectively) decreased. Only an increase in C6/LnA-alcohols was observed, the amount of trans-2-hexen-1-ol was higher at tested extreme temperatures and times: 30 °C and 90 min (Table 2a, Fig. 1a).

542

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

Fig. 2 (continued)

Fig. 2 (continued)

Nevertheless, interactions between malaxation time and temperature were observed for C5-compounds: the behaviour of the majority compounds at short times (30 min) was opposite to high times (90 min) when temperature was changed (Table 2a). In conclusion, the total concentration of C6 volatiles decreased with high temperatures regardless of paste kneading time (Fig. 1a). For C5 volatiles, when malaxation time was set to 30 min an increase in the total amount of C5 compounds was

observed at 30 °C, when time was set to 90 min C5/LnA compounds diminished and C5/LA compounds increased slightly (Fig. 1b). Angerosa et al. (2001) found the same trend for Coratina oils, with the exception of 1-penten-3-one, although these authors assessed only C5/LnA compounds. In a previous paper, the same authors (Morales, Angerosa, & Aparicio, 1999) highlighted that the production of 1-penten-3-one increased with temperature regardless malaxation time, suggesting this volatile is produced through an oxidation process.

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

543

Fig. 2 (continued)

Fig. 2 (continued)

3.3. One-dimension partial least squares regression (PLS2) between volatile compounds and odorant series To estimate quantitatively the overall oil aroma, a method based in the addition of OAVs was employed. This procedure makes it possible to relate quantitative information obtained by chemical analysis to sensory perception, providing a tentative aroma profile (Sánchez-Palomo E & Alonso-Villegas R, 2010).

Two PLS2 modelling (one for Morisca and another for Manzanilla de Sevilla oils) between two data matrices (both volatile compounds and odorant series) were performed providing in both cases one-factor or 1D model. The models were evaluated by cross validation via the root mean square error for predictions (RMSEP), which was calculated to be lower than 10 in both cases. In general, there were found significant positive and negative correlations (r = |0.7–1.0|) amongst many of the data of both matrices.

544

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545

Connecting 18 volatile compounds with 9 aromatic series in Morisca oils extracted at different malaxation conditions (20 or 30 °C for 30 or 90 min) resulted in two different groups of oils. The other two groups of oils were remaining together in the middle of both extreme groups (Fig. 2a(x)). The findings obtained can be summarised as follows: 1. According to the scores figure (Fig. 2a(x)), the upper half only included the Morisca oils extracted at 30 °C for 30 min, while the lower half mainly included the Morisca oils extracted at 30 °C for longer times (90 min). However, along the middle baseline of 0 were placed the non-differentiated groups of Morisca oils extracted at 20 °C for both 30 and 90 min. 2. According to the X- and Y-loadings figures (Fig. 2a(y), Fig. 2a(z)), for the highest values of PC1, Morisca/30 °C/30 min oils were mainly associated to the following volatile compounds: 1-penten-3-one, trans-2-hexenal and cis-3-hexenyl acetate, and to the odorant series: bitter-like, pungent-like, leaf, apple and grass. For the lowest values of PC1, Morisca/30 °C/90 min oils were mainly associated to the volatiles trans-2-hexen-1ol, 1-hexanol and 3-pentanone, and to the odorant series olive and banana fruits. Morisca/20 °C/30 or 90 min oils were more similar to Morisca/30 °C/90 min oils. A clear separation of two groups of Manzanilla de Sevilla oils was obtained (Fig. 2b) by linking 18 volatile compounds with 9 aromatic series: 1. According to scores figure (Fig. 2b(x)), the upper half only included Manzanilla de Sevilla oils extracted at 30 °C for 30 min, while the lower half mainly included Manzanilla de Sevilla oils extracted at the same temperature for longer times (90 min). 2. According to the X- and Y-loadings figures (Fig. 2b(y), Fig. 2b(z)) for the highest values of PC1, Manzanilla de Sevilla/30 °C/30 min oils were mainly associated to the volatiles cis-3-hexenyl acetate, hexyl acetate and pentanal, and to the odorant series banana, olive, grass and wood. For the lowest values of PC1, Manzanilla de Sevilla/30 °C/90 min oils were mainly associated to the volatiles trans-2-hexen-1-ol, trans-2-hexenal and hexanal, and to the aromatic series apple, sweet-like, bitter-like and leaf. 4. Conclusions The results obtained in this work shows that it is possible to modulate the volatile fraction of an oil in accordance with malaxation temperature ranging from 20 to 30 °C and above all malaxation time ranging from 30 min to 90 min. The total concentration of C6 volatiles increased with extended times of olive paste kneading for both oils regardless of temperature. For C5 volatiles, this behaviour was maintained in the case of Manzanilla de Sevilla oils but not in the case of Morisca oils. Concerning temperature conditions, the total concentration of Morisca C6 volatiles decreased with 30 °C regardless of paste kneading time. Nevertheless, for C5 volatiles, when malaxation time was set to 30 min an increase in the total amount of C5 compounds was observed, and when time was set to 90 min C5/LnA compounds diminished and C5/LA compounds increased slightly. A clear difference was found between oils processed at 30 °C/ 30 min and 30 °C/90 min for both varieties in terms of odorant series:  The oils from Morisca obtained at 30 °C and 30 min were characterised mainly by bitter-like, pungent-like, leaf, apple and grass odorant series, while for Manzanilla de Sevilla oils, fruity (banana, olive), grass and wood notes predominated.

 The sensory profile of oils from Morisca obtained at 30 °C and 90 min changed with respect to shorter times, thus olive and banana fruits were the characteristic aromatic series instead. For Manzanilla de Sevilla the sensory profile changed to apple, sweet-like, bitter-like and leaf notes.

Acknowledgements This work was funded by EU FEDER, and also under contracts 09TAL045E (Xunta de Galicia and Aceites Abril S.L.) and 2009/ 060 (Xunta de Galicia). The authors are grateful to Aceites Abril S. L., and especially indebted to J. M. Pérez-Canal and F. OsorioRodríguez. References AAO-Agencia para el aceite de oliva (2013). Available at: http://aplicaciones. magrama.es/pwAgenciaAO/OlivarEspanol.aao?opcion_seleccionada=2100&control _acceso=S&idioma=ESP. Accessed 11 June 2013 Allouche, Y., Jiménez, A., Uceda, M., Paz Aguilera, M., Gaforio, J. J., & Beltrán, G. (2010). Influence of olive paste preparation conditions on virgin olive oil triterpenic compounds at laboratory-scale. Food Chemistry, 119(2), 765–769. Angerosa, F., Mostallino, R., Basti, C., & Vito, R. (2000). Virgin olive oil odour notes: Their relationships with volatile compounds from the lipoxygenase pathway and secoiridoid compounds. Food Chemistry, 68(3), 283–287. Angerosa, F., Mostallino, R., Basti, C., & Vito, R. (2001). Influence of malaxation temperature and time on the quality of virgin olive oils. Food Chemistry, 72(1), 19–28. Aparicio, R., & Harwood, J. (2003). Manual del aceite de oliva (1st ed.). Madrid: AMV Ediciones, Mundi-prensa. Boselli, E., Di Lecce, G., Strabbioli, R., Pieralisi, G., & Frega, N. G. (2009). Are virgin olive oils obtained below 27 °C better than those produced at higher temperatures? LWT – Food Science and Technology, 42(3), 748–757. Clodoveo, M. L. (2012). Malaxation: Influence on virgin olive oil quality. Past, present and future – An overview. Trends in Food Science and Technology, 25(1), 13–23. European Union Commission (1991). The characteristics of olive oil and oliveresidue oil and on the relevant methods of analysis. Official Journal of the European Communities. Regulation EEC/2568/91. § L 248. European Union Commission (2003). The characteristics of olive oil and oliveresidue oil and on the relevant methods of analysis. Official Journal of the European Communities. Regulation EC/1989/2003. § L 295. European Union Commission (2007). The characteristics of olive oil and oliveresidue oil and on the relevant methods of analysis. Official Journal of the European Communities. Regulation EC/702/2007. § L 161. European Union Commission (2011). The characteristics of olive oil and oliveresidue oil and on the relevant methods of analysis. Official Journal of the European Communities. Regulation EC/61/2011. § L 23. European Union Commission (2012). Marketing standards for olive oil. Official Journal of the European Union. Implementing Regulation (EU) N° 29/2012. § L 12. Gómez-Rico, A., Inarejos-García, A. M., Salvador, M. D., & Fregapane, G. (2009). Effect of malaxation conditions on phenol and volatile profiles in olive paste and the corresponding virgin olive oils (Olea europaea L. Cv. Cornicabra). Journal of Agricultural and Food Chemistry, 57(9), 3587–3595. Inarejos-García, A. M., Gómez-Rico, A., Salvador, M. D., & Fregapane, G. (2009). Influence of malaxation conditions on virgin olive oil yield, overall quality and composition. European Food Research and Technology, 228(4), 671–677. IOOC-International Olive Oil Council (1984). Document No 6. Kalua, C. M., Allen, M. S., Bedgood, D. R., Jr., Bishop, A. G., Prenzler, P. D., & Robards, K. (2007). Olive oil volatile compounds, flavour development and quality: A critical review. Food Chemistry, 100(1), 273–286. Kalua, C. M., Bedgood, D. R., Jr., Bishop, A. G., & Prenzler, P. D. (2006). Changes in volatile and phenolic compounds with malaxation time and temperature during virgin olive oil production. Journal of Agricultural and Food Chemistry, 54(20), 7641–7651. Morales, M. T., Angerosa, F., & Aparicio, R. (1999). Effect of the extraction conditions of virgin olive oil on the lipoxygenase cascade: Chemical and sensory implications. Grasas y Aceites, 50(2), 114–121. Peinado, R. A., Mauricio, J. C., & Moreno, J. (2006). Aromatic series in sherry wines with gluconic acid subjected to different biological aging conditions by Saccharomyces cerevisiae var. capensis. Food Chemistry, 94(2), 232–239. Ranalli, A., Contento, S., Schiavone, C., & Simone, N. (2001). Malaxing temperature affects volatile and phenol composition as well as other analytical features of virgin olive oil. European Journal of Lipid Science and Technology, 103(4), 228–238. Ranalli, A., Pollastri, L., Contento, S., Iannucci, E., & Lucera, L. (2003). Effect of olive paste kneading process time on the overall quality of virgin olive oil. European Journal of Lipid Science and Technology, 105(2), 57–67.

P. Reboredo-Rodríguez et al. / Food Chemistry 158 (2014) 534–545 Reboredo-Rodríguez, P., González-Barreiro, C., Cancho-Grande, B., & Simal-Gándara, J. (2012). Dynamic headspace/GC-MS to control the aroma fingerprint of extravirgin olive oil from the same and different olive varieties. Food Control, 25(2), 684–695. Reboredo-Rodríguez, P., González-Barreiro, C., Cancho-Grande, B., & Simal-Gándara, J. (2013a). Concentrations of aroma compounds and odor activity values of odorant series in different olive cultivars and their oils. Journal of Agricultural and Food Chemistry, 61(22), 5252–5259. Reboredo-Rodríguez, P., González-Barreiro, C., Cancho-Grande, B., & Simal-Gándara, J. (2013b). Effects of sedimentation plus racking process in the extra virgin olive oil aroma fingerprint obtained by DHS-TD/GC-MS. Food and Bioprocess Technology, 6(5), 1290–1301. Salas, J. J., & Sánchez, J. (1999). The decrease of virgin olive oil flavor produced by high malaxation temperature is due to inactivation of hydroperoxide. Journal of Agricultural and Food Chemistry, 47(3), 809–812. Sánchez-Palomo, E., Gómez García-Carpintero, E., Alonso-Villegas, R., & GonzálezViñas, M. A. (2010). Characterization of aroma compounds of Verdejo white

545

wines from the La Mancha region by odour activity values. Flavour and Fragrance Journal, 25(6), 456–462. Stefanoudaki, E., Koutsaftakis, A., & Harwood, J. L. (2011). Influence of malaxation conditions on characteristic qualities of olive oil. Food Chemistry, 127(4), 1481–1486. Taticchi, A., Esposto, S., Veneziani, G., Urbani, S., Selvaggini, R., & Servili, M. (2013). The influence of the malaxation temperature on the activity of polyphenoloxidase and peroxidase and on the phenolic composition of virgin olive oil. Food Chemistry, 136(2), 975–983. Tura, D., Failla, O., Bassi, D., Pedò, S., & Serraiocco, A. (2008). Cultivar influence on virgin olive (Olea europea L.) oil flavor based on aromatic compounds and sensorial profile. Scientia Horticulturae, 118(2), 39–148. Youssef, O., Mokhtar, G., Abdelly, C., Mohamed, S. N., Mokhtar, Z., & Guido, F. (2013). Changes in volatile compounds and oil quality with malaxation time of Tunisian cultivars of Olea europea. International Journal of Food Science and Technology, 48(1), 74–81.