Accepted Manuscript Geographical Traceability of Virgin Olive Oils from South-Western Spain by Their Multi-Elemental Composition María Beltrán, María Sánchez-Astudillo, Ramón Aparicio, Diego L. GarcíaGonzález PII: DOI: Reference:
S0308-8146(14)01152-2 http://dx.doi.org/10.1016/j.foodchem.2014.07.104 FOCH 16170
To appear in:
Food Chemistry
Received Date: Revised Date: Accepted Date:
5 September 2013 23 May 2014 22 July 2014
Please cite this article as: Beltrán, M., Sánchez-Astudillo, M., Aparicio, R., García-González, D.L., Geographical Traceability of Virgin Olive Oils from South-Western Spain by Their Multi-Elemental Composition, Food Chemistry (2014), doi: http://dx.doi.org/10.1016/j.foodchem.2014.07.104
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Geographical Traceability of Virgin Olive Oils from South-Western Spain by Their Multi-Elemental Composition María Beltrán1, María Sánchez-Astudillo1, Ramón Aparicio2*, Diego L. García-González2
1
Department of Chemistry and Science of Materials, Faculty of Experimental Sciences, University of Huelva, Avda. Tres de Marzo S/N. 21071, Huelva, Spain 2
Instituto de la Grasa (CSIC), Padre García Tejero 4, 41012, Sevilla, Spain
*Author to whom correspondence should be sent. E-mail: [email protected]; Tel: +34 954 61 15 50; Fax: +34 954 61 67 90
1
ABSTRACT The geographical traceability of virgin olive oil can be controlled by chemical species that are linked to the production area. Trace elements are among these species. The hypothesis is that the transfer of elements from the soil to the oil is subjected to minor variations and therefore this chemical information can be used for geographical traceability. In order to confirm this hypothesis, the trace elements of virgin olive oils from south-western Spain were analyzed, and the same elements were determined in the corresponding olive-pomaces and soils. The differences in the concentration were studied according to cultivars and locations. Results show some coincidences in the selection of elements in soils (W, Fe, Na), olivepomace (W, Fe, Na, Mg, Mn, Ca, Ba, Li) and olive oils (W, Fe, Mg, Mn, Ca, Ba, Li, Bi), which supports their utility in traceability. In the case of olive oils, 93% of the samples were correctly classified in their geographical origins (96% for Beas, 77% for Gibraleón, 91% for Niebla, and 100% for Sanlúcar de Guadiana).
Keywords: Virgin olive oil, Geographical traceability, Trace elements, Inductively coupled plasma-mass spectrometry
2
1. Introduction The quantitative determination of trace elements in foodstuffs has always been a challenge in analytical chemistry since they have evident nutritional (Boufleur, Dos Santos, Debastiani, Yoneama, Amaral & Dias, 2013), safety (Zand, Chowdhry, Wray, Pullen & Snowden, 2012) and quality implications (Benedet & Shibamoto, 2008). Although there is extensive literature concerning the analysis of other materials, such as some lubricants and fuels (Maryutina & Soin, 2009), the information concerning food analysis is relatively scarce. In regards to edible oils, elemental analysis has received little attention so far and, on the contrary, other major (e.g. fatty acids) or minor components (e.g. sterols) has been the targets of the analytical effort in the last decades for solving authenticity/quality issues. However, it is a well-known fact that certain metallic contaminants (e.g. Cu and Fe) speed up the oxidation processes of edible oils thereby having a negative effect on their sensory quality (Benedet & Shibamoto, 2008). The unimpeachable importance of metals in olive oil stability explains that the International Olive Council (IOC) established concentration limits of Cu and Fe, and also contaminants such as Pb and As, in olive and olive-pomace oils (IOC, 2013). Other elements (e.g. Ca, Mg, Mn) are also present in a wide concentration range in olive oils (Benincasa, Lewis, Perri, Sindona & Tagarelli, 2007). Other elements can be found and they might be transfer into the oil from the metallic surfaces of the processing equipment or storage material. They might be also incorporated into de oil from the soil although their concentrations are modulated by biochemical pathways of each cultivar (Chatzistathis, Therios & Alifragis, 2009). As olive trees are closely linked to land, the importance of the chemical species coming into virgin olive oils (VOOs) from the soil is taking on special relevance day by day, as it is the case of elements. The importance of the elements lies in their potential use in geographical traceability, in particular in the characterization of protected designations of origin (PDOs) or protected geographical indications (PGIs) (EU, 2012), and they can also 3
contribute to determine VOO geographical provenance of non PDO oils. Thus, a complete element characterization may back up a hypothetical warfare against illicit practices as consequence of the financial benefits associated with these prestigious labels. Thus, a nonPDO product may be labelled as a PDO one and also it may be adulterated with olive oils that do not fulfil the PDO/PGI requirements. The use of elemental analysis for detecting these frauds can be an alternative to other approaches to tackle VOO geographical traceability based on chromatographic, spectroscopic, isotopic, and in-tandem analytical techniques, (Alonso-Salces et al. 2010; Aparicio, 1988; Benincasa, Lewis, Perri, Sindona & Tagarelli 2007; Camin et al., 2010a; García-González, Tena & Aparicio, 2011; Woodcock, Downey, O’Donnell, 2008). The needs of characterizing the trace elements of olive oil, and other edible oils, has led researcher to optimize methodologies with the most sensitive techniques. Thus, edible oils have been analysed for different elements using potentiometry (Dugo, La Pera, La Torre & Giuffrida, 2004), inductively coupled plasma atomic emission spectrometry (ICP-AES) (Zeiner, Steffan & Cindric, 2005), electrothermal vaporization inductively coupled plasma mass spectrometry (Huang & Jiang, 2001) and mostly atomic absorption spectrometry (AAS) (Mendil, Uluözlü, Tüzen & Soylak, 2009), which is included in some official methods (IOC 2011; Codex Alimentarius, 2009). Since some elements are present at very low concentration, the inductively coupled plasma-mass spectrometry (ICP-MS) is the most suitable tool because of its low detection limits, multi-elemental capacity and wide linear range that result of combining the remarkable characteristics of ICP for atomising and ionising samples with the sensitivity and selectivity of mass spectrometry (Castillo et al., 1999). Thus, the number of papers dealing with the analysis of organic samples by ICP-MS has increased in recent years, particularly in the analysis of olive oil (Benincasa et al., 2007; Jiménez, Velarte, Gomez & Castillo, 2004; Llorent-Martínez, Ortega-Barrales, Fernández-de Córdova, Domínguez-Vidal
4
& Ruiz-Medina, 2011a; Llorent-Martínez, Ortega-Barrales, Fernández-de Córdova & RuizMedina, 2011b). In this paper, we analyse the availability of elements, determined by ICP-MS, for implementing a reliable method of geographical traceability of olive oils. As elements can come from other sources than soil, samples of soils, wet olive-pomaces (“alperujo”) (WOP) and virgin olive oils from diverse geographical places have been analyzed. The presence and concentration of some elements in VOOs and WOPs in comparison with the results of analysing the soils of the orchards will help to understand the usefulness of this technique for explaining the olive oil geographic provenance, and hence for protecting virgin olive oils from PDOs and PGIs against false copies. A geographical zone of Southern Spain was selected for its diversity of cultivars that are cultivated in modern irrigated orchards with different characteristics of soils.
2. Materials and Methods 2.1. Samples Table 1 summarizes the number of samples of virgin olive oils (VOOs) and olivepomaces according to cultivars (var. Arbequina, Picual and Verdial de Huévar) and their geographical provenances in terms of the municipalities of the Huelva province, and the samples of the orchards of those municipalities where olive trees are cultivated. Seventeen orchards of olive trees located in four municipalities of the Southern Spanish province of Huelva - Beas (3), Gibraleón (2), Niebla (8), and Sanlúcar de Guadiana (4) - were selected because cultivars are harvested in orchards with diverse soil characteristics (REDIAM, 2013). Soil samples (40) were collected in each orchard at two depths, 30 cm and 60 cm, because the depth of the roots can vary among the olive trees of cultivars and also in order to study the possible variability of the element concentrations with the depth. The number of samples per orchard is two (one per depth) with the exception of the orchards
5
located at Beas as their size are larger and their soil compositions are very diverse in comparison with the orchards of the other municipalities (REDIM, 2013). The orchards can have more than one cultivar depending on the orchard size and its geographical location. The samples of VOOs and olive-pomaces were 82, which can be clustered in terms of cultivars (40 for Arbequina, 29 for Picual and 13 for Verdial de Huévar) or geographical provenance (28 from Beas, 14 from Gibraleón, 21 from Niebla and 19 from Sanlúcar de Guadiana). Olives were harvested by mechanical means at the same step of ripeness, according to the classification of Hermoso, Uceda, García, Morales, Frías and Fernández (1991). All the olive oils were processed by two-phase centrifugation systems in cooperative societies of farmers under similar conditions of milling and malaxation. The resulting products of this extraction system were VOO and olive-pomace (“alperujo”). The latter is a by-product that consists of vegetation water and solids (stone and pulp of the olives) and a small percentage of olive oil. 2.2. Sample preparation and digestion Virgin olive oils and olive-pomaces (“alperujo”) resulting of processing olives from the different orchards as well as samples from their soils were prepared for digestion. The number of elements quantified in the samples were 34 although some of them were not detected (e.g., Strontium) or at trace level in the samples of VOOs and olive-pomaces. The samples were digested in an Anton Paar (multiwave 3000 SOLV) oven with programmable power control (10 W increments, maximum power 1000 W) with segmented rotor XQ80 (35 bar of maximum operating pressure and 260 ºC of maximum operating temperature).
2.2.1. Olive oil samples Microwave-assisted acid decomposition was performed to dissolve the oil sample for elemental analysis. The digestion was carried out with 0.5 g aliquot of sample, weighed
6
directly into the digestion vessel, to which were added 5 mL of nitric acid at 65% v/v, 3 mL of hydrogen peroxide at 30% v/v, and 1 mL of hydrochloric acid (Sigma-Aldrich, Madrid, Spain) (Acar, 2012). The microwave operation parameters of the Anton Paar oven were firstly a ramp of 15 minutes to reach 280 ºC and 80 bar that was maintained for 20 minutes with minimum level of ventilation, and later, the samples were vented for 15 minutes. After digestion, samples were stored at 25 ºC for 12 h and finally, all the digestion liquors were diluted to 25 mL with ultrapure water (Llorent-Martínez et al., 2011a). Samples were thoroughly shaken prior to analysis by ICP-MS.
2.2.2. Olive-pomace samples The olive-pomace (alperujo) samples were frozen and lyophilized in a laboratory freeze-dryer Cryodos 80 (Telstar, Tarrasa, Spain) at -80 ºC. Two subsequent digestions were performed to dissolve the freeze-dried alperujo for elemental analysis. The first digestion was carried out as follows: 5 mL of nitric acid at 65%, 1 mL of hydrogen peroxide at 30%, 1 mL of hydrofluoric acid at 40% and 1 mL of hydrochloric acid at 30% were added to 0.20 g of lyophilized alperujo. An Anton Paar Microwave-Assisted oven was used under the condition described for olive oil with a ramp of 12 min to reach 210 ºC and 40 bar, and this conditions were maintained for 20 min. The second digestion was carried out as follows: 6 mL of boric acid were added to the residue of the previous digestion. The program of microwave operation parameters were a ramp of 5 minutes to reach 210 ºC and 40 bar and these conditions were maintained during 15 minutes with a minimum level of ventilation, and then the sample was vented for 15 minutes. After digestion, samples were stored at 25 ºC for 12 h and finally, the residues were diluted in 25 mL with ultrapure water.
7
2.2.3. Soil samples The method described by De la Rosa et al. was applied for the digestion of soil samples (De la Rosa, Chacón, Sánchez de la Campa, Carrasco & Nieto, 2001). The soil samples, previously frozen, were grounded many cycles in a special vibrating mill using titanium rods under cryogenic conditions. An aliquot of 0.1 g of each sample was placed into a 60 mL PTFE/PFA bomb (Savillex, Eden Prairie, MN) where 8 mL of hydrofluoric acid (HF) and 3 mL of nitric acid were added. The mixture was heated at 90 ºC in the closed bomb for 24 h. Soil samples were homogenized in the titanium mill and a complete digestion with HF would release more titanium and zirconium in to the digest which could affect the determination of copper and cadmium by ICP-MS through the formation of oxide ions. Bomb was opened for the evaporation of acids, and the mixture was heated at 130º C. Then, 3 mL of nitric acid were added, the bomb was closed and the mixture was heated at 90º C for 12 h. Then the acids were re-evaporated and 3 mL of hydrochloric acid were added, the bomb was closed and the mixture was heated at 90º C for 12 h. The residue was recovered with nitric acid at 2 % into a 100 mL volumetric flask once total dryness was achieved. In all samples (soils, olive-pomaces, virgin olive oils) Rh (103) was added as internal standard in a 5 µg/L concentration.
2.3. ICP-MS Analyses The concentration of 34 elements (Table 2) were determined by a quadrupole inductively coupled plasma mass spectrometer (ICP-MS) (Agilent 7700X Model G3281A, Agilent Technologies, CA, USA) working under the following operating conditions: RF power, 1.5kW; plasma Ar flow rate, 15 L/min; auxiliary Ar flow rate, 0.9 L/min; carrier Ar flow rate, 1.1 L/min; sample depth, 9.0 mm; spray interface temp. 2 ºC; sample flow rate, 400 µL/min. The sampler and skimmer cones were of nickel. The instrument was run under a linear multipoint calibration 1-200 µg/L. Analyses were carried in duplicate.
8
Glassware was not used to avoid metal releases and all the plastic containers, like PFA Teflon digestion vessels, were checked for contamination. Vessels were cleaned using the same microwave operating program for digestion but adding 7 mL HNO3 to each digestion vessel after each analytical batch. Later, all the vessels were thoroughly rinsed with Milli-Q water. Ultrapure deionised water was obtained from Milli-Q system (Millipore, Bedford, MA).
2.4 Calibration procedure The calibration standard solution was prepared from a multi-element standard solution (SCP Science, Paris, France) by dilution with HNO3 in ultrapure water (Llorent-Martínez et al., 2010b). For the quantitative analysis of oils calibration curves were built at five different concentrations (Benincasa et al., 2007). The concentration range was 0.2-60 ng/mL for all the elements excepting Ba, Ca, Sr, V, Zr, which were calibrated in a wider concentration range (10-200 ng/mL). The recoveries were determined by analysing spike solutions (Boqué, Maroto, Riu & Rius, 2002), and the values were within the range 78-218%, being 50% of them in the range 90-110%. These recoveries were inside the range described by other authors (Arunachalam, Mohl, Ostapczuk & Emons, 1995; Benincasa et al., 2007; Llorent et al., 2011a-b).
2.5 Data analysis The data matrix (elements × samples) was analyzed by uni and multi-variate mathematical procedures. Brown-Forsythe test (Brown & Forsythe, 1974) was used to determine homogeneity of the variances and to select variables (elements) with univariate discriminate ability. Stepwise linear discriminant analysis (SLDA), a supervised statistical procedure, was applied under the strictest conditions (F-to-Enter ≥ 4.0). Principal Component Analysis (PCA), an unsupervised statistical procedure, was used as it allows reducing the
9
dimensionality of the original data by means of equations (principal components) that are linear combination of the elements and encapsulate their variability. All statistical data treatments were performed by Statistica 6.0 (StatSoft, Tulsa, OK).
3. Results and discussion The trace elements that are incorporated to the olive tree from the soil are partially transferred to the olives (Bakircioglu, Kurtulus & Yurtsever, 2013). In consequence, trace elements are determined in olive oil and/or olive-pomace samples, the only two materials resulting of processing the olives. The province of Huelva (Southern Spain) is an excellent geographical zone to study the importance of the soil composition in geographical traceability because it provides different cultivars planted in the same land. The orchards of var. Verdial de Huévar, which meant more than 90% of the whole production only a few years ago, have gradually been substituted by var. Arbequina and Picual cultivated in irrigated intensive plantations. Thus, in this area it is possible to analyze the concentration of elements in olive oil and olive-pomace of several cultivars with respect to their availability in the soils. As the root depth in olive trees varies in accordance with the soil characteristics and the olive tree age (Fernández, Moreno, Cabrera, Arrue & Martín-Aranda, 1991), the soil samples were collected at 30 cm and 60 cm depth because the plantations are in its juvenile step –with the exception of var. Verdial de Huévar – and the tree roots were estimated to be around 40-50 cm depth. Table 2 shows the mean and standard deviation of the concentrations of 34 elements at those two depths. The half of them was quantified at concentrations higher than 10 mg/kg. In regards to elements quantified at low concentrations, it is remarkable the low concentrations of Mn that were lower than 1 mg/kg in all cases. The concentrations of eighteen elements are higher in the samples of soils collected at 60 cm depth (Table 2) (p
S0308-8146(14)01152-2 http://dx.doi.org/10.1016/j.foodchem.2014.07.104 FOCH 16170
To appear in:
Food Chemistry
Received Date: Revised Date: Accepted Date:
5 September 2013 23 May 2014 22 July 2014
Please cite this article as: Beltrán, M., Sánchez-Astudillo, M., Aparicio, R., García-González, D.L., Geographical Traceability of Virgin Olive Oils from South-Western Spain by Their Multi-Elemental Composition, Food Chemistry (2014), doi: http://dx.doi.org/10.1016/j.foodchem.2014.07.104
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Geographical Traceability of Virgin Olive Oils from South-Western Spain by Their Multi-Elemental Composition María Beltrán1, María Sánchez-Astudillo1, Ramón Aparicio2*, Diego L. García-González2
1
Department of Chemistry and Science of Materials, Faculty of Experimental Sciences, University of Huelva, Avda. Tres de Marzo S/N. 21071, Huelva, Spain 2
Instituto de la Grasa (CSIC), Padre García Tejero 4, 41012, Sevilla, Spain
*Author to whom correspondence should be sent. E-mail: [email protected]; Tel: +34 954 61 15 50; Fax: +34 954 61 67 90
1
ABSTRACT The geographical traceability of virgin olive oil can be controlled by chemical species that are linked to the production area. Trace elements are among these species. The hypothesis is that the transfer of elements from the soil to the oil is subjected to minor variations and therefore this chemical information can be used for geographical traceability. In order to confirm this hypothesis, the trace elements of virgin olive oils from south-western Spain were analyzed, and the same elements were determined in the corresponding olive-pomaces and soils. The differences in the concentration were studied according to cultivars and locations. Results show some coincidences in the selection of elements in soils (W, Fe, Na), olivepomace (W, Fe, Na, Mg, Mn, Ca, Ba, Li) and olive oils (W, Fe, Mg, Mn, Ca, Ba, Li, Bi), which supports their utility in traceability. In the case of olive oils, 93% of the samples were correctly classified in their geographical origins (96% for Beas, 77% for Gibraleón, 91% for Niebla, and 100% for Sanlúcar de Guadiana).
Keywords: Virgin olive oil, Geographical traceability, Trace elements, Inductively coupled plasma-mass spectrometry
2
1. Introduction The quantitative determination of trace elements in foodstuffs has always been a challenge in analytical chemistry since they have evident nutritional (Boufleur, Dos Santos, Debastiani, Yoneama, Amaral & Dias, 2013), safety (Zand, Chowdhry, Wray, Pullen & Snowden, 2012) and quality implications (Benedet & Shibamoto, 2008). Although there is extensive literature concerning the analysis of other materials, such as some lubricants and fuels (Maryutina & Soin, 2009), the information concerning food analysis is relatively scarce. In regards to edible oils, elemental analysis has received little attention so far and, on the contrary, other major (e.g. fatty acids) or minor components (e.g. sterols) has been the targets of the analytical effort in the last decades for solving authenticity/quality issues. However, it is a well-known fact that certain metallic contaminants (e.g. Cu and Fe) speed up the oxidation processes of edible oils thereby having a negative effect on their sensory quality (Benedet & Shibamoto, 2008). The unimpeachable importance of metals in olive oil stability explains that the International Olive Council (IOC) established concentration limits of Cu and Fe, and also contaminants such as Pb and As, in olive and olive-pomace oils (IOC, 2013). Other elements (e.g. Ca, Mg, Mn) are also present in a wide concentration range in olive oils (Benincasa, Lewis, Perri, Sindona & Tagarelli, 2007). Other elements can be found and they might be transfer into the oil from the metallic surfaces of the processing equipment or storage material. They might be also incorporated into de oil from the soil although their concentrations are modulated by biochemical pathways of each cultivar (Chatzistathis, Therios & Alifragis, 2009). As olive trees are closely linked to land, the importance of the chemical species coming into virgin olive oils (VOOs) from the soil is taking on special relevance day by day, as it is the case of elements. The importance of the elements lies in their potential use in geographical traceability, in particular in the characterization of protected designations of origin (PDOs) or protected geographical indications (PGIs) (EU, 2012), and they can also 3
contribute to determine VOO geographical provenance of non PDO oils. Thus, a complete element characterization may back up a hypothetical warfare against illicit practices as consequence of the financial benefits associated with these prestigious labels. Thus, a nonPDO product may be labelled as a PDO one and also it may be adulterated with olive oils that do not fulfil the PDO/PGI requirements. The use of elemental analysis for detecting these frauds can be an alternative to other approaches to tackle VOO geographical traceability based on chromatographic, spectroscopic, isotopic, and in-tandem analytical techniques, (Alonso-Salces et al. 2010; Aparicio, 1988; Benincasa, Lewis, Perri, Sindona & Tagarelli 2007; Camin et al., 2010a; García-González, Tena & Aparicio, 2011; Woodcock, Downey, O’Donnell, 2008). The needs of characterizing the trace elements of olive oil, and other edible oils, has led researcher to optimize methodologies with the most sensitive techniques. Thus, edible oils have been analysed for different elements using potentiometry (Dugo, La Pera, La Torre & Giuffrida, 2004), inductively coupled plasma atomic emission spectrometry (ICP-AES) (Zeiner, Steffan & Cindric, 2005), electrothermal vaporization inductively coupled plasma mass spectrometry (Huang & Jiang, 2001) and mostly atomic absorption spectrometry (AAS) (Mendil, Uluözlü, Tüzen & Soylak, 2009), which is included in some official methods (IOC 2011; Codex Alimentarius, 2009). Since some elements are present at very low concentration, the inductively coupled plasma-mass spectrometry (ICP-MS) is the most suitable tool because of its low detection limits, multi-elemental capacity and wide linear range that result of combining the remarkable characteristics of ICP for atomising and ionising samples with the sensitivity and selectivity of mass spectrometry (Castillo et al., 1999). Thus, the number of papers dealing with the analysis of organic samples by ICP-MS has increased in recent years, particularly in the analysis of olive oil (Benincasa et al., 2007; Jiménez, Velarte, Gomez & Castillo, 2004; Llorent-Martínez, Ortega-Barrales, Fernández-de Córdova, Domínguez-Vidal
4
& Ruiz-Medina, 2011a; Llorent-Martínez, Ortega-Barrales, Fernández-de Córdova & RuizMedina, 2011b). In this paper, we analyse the availability of elements, determined by ICP-MS, for implementing a reliable method of geographical traceability of olive oils. As elements can come from other sources than soil, samples of soils, wet olive-pomaces (“alperujo”) (WOP) and virgin olive oils from diverse geographical places have been analyzed. The presence and concentration of some elements in VOOs and WOPs in comparison with the results of analysing the soils of the orchards will help to understand the usefulness of this technique for explaining the olive oil geographic provenance, and hence for protecting virgin olive oils from PDOs and PGIs against false copies. A geographical zone of Southern Spain was selected for its diversity of cultivars that are cultivated in modern irrigated orchards with different characteristics of soils.
2. Materials and Methods 2.1. Samples Table 1 summarizes the number of samples of virgin olive oils (VOOs) and olivepomaces according to cultivars (var. Arbequina, Picual and Verdial de Huévar) and their geographical provenances in terms of the municipalities of the Huelva province, and the samples of the orchards of those municipalities where olive trees are cultivated. Seventeen orchards of olive trees located in four municipalities of the Southern Spanish province of Huelva - Beas (3), Gibraleón (2), Niebla (8), and Sanlúcar de Guadiana (4) - were selected because cultivars are harvested in orchards with diverse soil characteristics (REDIAM, 2013). Soil samples (40) were collected in each orchard at two depths, 30 cm and 60 cm, because the depth of the roots can vary among the olive trees of cultivars and also in order to study the possible variability of the element concentrations with the depth. The number of samples per orchard is two (one per depth) with the exception of the orchards
5
located at Beas as their size are larger and their soil compositions are very diverse in comparison with the orchards of the other municipalities (REDIM, 2013). The orchards can have more than one cultivar depending on the orchard size and its geographical location. The samples of VOOs and olive-pomaces were 82, which can be clustered in terms of cultivars (40 for Arbequina, 29 for Picual and 13 for Verdial de Huévar) or geographical provenance (28 from Beas, 14 from Gibraleón, 21 from Niebla and 19 from Sanlúcar de Guadiana). Olives were harvested by mechanical means at the same step of ripeness, according to the classification of Hermoso, Uceda, García, Morales, Frías and Fernández (1991). All the olive oils were processed by two-phase centrifugation systems in cooperative societies of farmers under similar conditions of milling and malaxation. The resulting products of this extraction system were VOO and olive-pomace (“alperujo”). The latter is a by-product that consists of vegetation water and solids (stone and pulp of the olives) and a small percentage of olive oil. 2.2. Sample preparation and digestion Virgin olive oils and olive-pomaces (“alperujo”) resulting of processing olives from the different orchards as well as samples from their soils were prepared for digestion. The number of elements quantified in the samples were 34 although some of them were not detected (e.g., Strontium) or at trace level in the samples of VOOs and olive-pomaces. The samples were digested in an Anton Paar (multiwave 3000 SOLV) oven with programmable power control (10 W increments, maximum power 1000 W) with segmented rotor XQ80 (35 bar of maximum operating pressure and 260 ºC of maximum operating temperature).
2.2.1. Olive oil samples Microwave-assisted acid decomposition was performed to dissolve the oil sample for elemental analysis. The digestion was carried out with 0.5 g aliquot of sample, weighed
6
directly into the digestion vessel, to which were added 5 mL of nitric acid at 65% v/v, 3 mL of hydrogen peroxide at 30% v/v, and 1 mL of hydrochloric acid (Sigma-Aldrich, Madrid, Spain) (Acar, 2012). The microwave operation parameters of the Anton Paar oven were firstly a ramp of 15 minutes to reach 280 ºC and 80 bar that was maintained for 20 minutes with minimum level of ventilation, and later, the samples were vented for 15 minutes. After digestion, samples were stored at 25 ºC for 12 h and finally, all the digestion liquors were diluted to 25 mL with ultrapure water (Llorent-Martínez et al., 2011a). Samples were thoroughly shaken prior to analysis by ICP-MS.
2.2.2. Olive-pomace samples The olive-pomace (alperujo) samples were frozen and lyophilized in a laboratory freeze-dryer Cryodos 80 (Telstar, Tarrasa, Spain) at -80 ºC. Two subsequent digestions were performed to dissolve the freeze-dried alperujo for elemental analysis. The first digestion was carried out as follows: 5 mL of nitric acid at 65%, 1 mL of hydrogen peroxide at 30%, 1 mL of hydrofluoric acid at 40% and 1 mL of hydrochloric acid at 30% were added to 0.20 g of lyophilized alperujo. An Anton Paar Microwave-Assisted oven was used under the condition described for olive oil with a ramp of 12 min to reach 210 ºC and 40 bar, and this conditions were maintained for 20 min. The second digestion was carried out as follows: 6 mL of boric acid were added to the residue of the previous digestion. The program of microwave operation parameters were a ramp of 5 minutes to reach 210 ºC and 40 bar and these conditions were maintained during 15 minutes with a minimum level of ventilation, and then the sample was vented for 15 minutes. After digestion, samples were stored at 25 ºC for 12 h and finally, the residues were diluted in 25 mL with ultrapure water.
7
2.2.3. Soil samples The method described by De la Rosa et al. was applied for the digestion of soil samples (De la Rosa, Chacón, Sánchez de la Campa, Carrasco & Nieto, 2001). The soil samples, previously frozen, were grounded many cycles in a special vibrating mill using titanium rods under cryogenic conditions. An aliquot of 0.1 g of each sample was placed into a 60 mL PTFE/PFA bomb (Savillex, Eden Prairie, MN) where 8 mL of hydrofluoric acid (HF) and 3 mL of nitric acid were added. The mixture was heated at 90 ºC in the closed bomb for 24 h. Soil samples were homogenized in the titanium mill and a complete digestion with HF would release more titanium and zirconium in to the digest which could affect the determination of copper and cadmium by ICP-MS through the formation of oxide ions. Bomb was opened for the evaporation of acids, and the mixture was heated at 130º C. Then, 3 mL of nitric acid were added, the bomb was closed and the mixture was heated at 90º C for 12 h. Then the acids were re-evaporated and 3 mL of hydrochloric acid were added, the bomb was closed and the mixture was heated at 90º C for 12 h. The residue was recovered with nitric acid at 2 % into a 100 mL volumetric flask once total dryness was achieved. In all samples (soils, olive-pomaces, virgin olive oils) Rh (103) was added as internal standard in a 5 µg/L concentration.
2.3. ICP-MS Analyses The concentration of 34 elements (Table 2) were determined by a quadrupole inductively coupled plasma mass spectrometer (ICP-MS) (Agilent 7700X Model G3281A, Agilent Technologies, CA, USA) working under the following operating conditions: RF power, 1.5kW; plasma Ar flow rate, 15 L/min; auxiliary Ar flow rate, 0.9 L/min; carrier Ar flow rate, 1.1 L/min; sample depth, 9.0 mm; spray interface temp. 2 ºC; sample flow rate, 400 µL/min. The sampler and skimmer cones were of nickel. The instrument was run under a linear multipoint calibration 1-200 µg/L. Analyses were carried in duplicate.
8
Glassware was not used to avoid metal releases and all the plastic containers, like PFA Teflon digestion vessels, were checked for contamination. Vessels were cleaned using the same microwave operating program for digestion but adding 7 mL HNO3 to each digestion vessel after each analytical batch. Later, all the vessels were thoroughly rinsed with Milli-Q water. Ultrapure deionised water was obtained from Milli-Q system (Millipore, Bedford, MA).
2.4 Calibration procedure The calibration standard solution was prepared from a multi-element standard solution (SCP Science, Paris, France) by dilution with HNO3 in ultrapure water (Llorent-Martínez et al., 2010b). For the quantitative analysis of oils calibration curves were built at five different concentrations (Benincasa et al., 2007). The concentration range was 0.2-60 ng/mL for all the elements excepting Ba, Ca, Sr, V, Zr, which were calibrated in a wider concentration range (10-200 ng/mL). The recoveries were determined by analysing spike solutions (Boqué, Maroto, Riu & Rius, 2002), and the values were within the range 78-218%, being 50% of them in the range 90-110%. These recoveries were inside the range described by other authors (Arunachalam, Mohl, Ostapczuk & Emons, 1995; Benincasa et al., 2007; Llorent et al., 2011a-b).
2.5 Data analysis The data matrix (elements × samples) was analyzed by uni and multi-variate mathematical procedures. Brown-Forsythe test (Brown & Forsythe, 1974) was used to determine homogeneity of the variances and to select variables (elements) with univariate discriminate ability. Stepwise linear discriminant analysis (SLDA), a supervised statistical procedure, was applied under the strictest conditions (F-to-Enter ≥ 4.0). Principal Component Analysis (PCA), an unsupervised statistical procedure, was used as it allows reducing the
9
dimensionality of the original data by means of equations (principal components) that are linear combination of the elements and encapsulate their variability. All statistical data treatments were performed by Statistica 6.0 (StatSoft, Tulsa, OK).
3. Results and discussion The trace elements that are incorporated to the olive tree from the soil are partially transferred to the olives (Bakircioglu, Kurtulus & Yurtsever, 2013). In consequence, trace elements are determined in olive oil and/or olive-pomace samples, the only two materials resulting of processing the olives. The province of Huelva (Southern Spain) is an excellent geographical zone to study the importance of the soil composition in geographical traceability because it provides different cultivars planted in the same land. The orchards of var. Verdial de Huévar, which meant more than 90% of the whole production only a few years ago, have gradually been substituted by var. Arbequina and Picual cultivated in irrigated intensive plantations. Thus, in this area it is possible to analyze the concentration of elements in olive oil and olive-pomace of several cultivars with respect to their availability in the soils. As the root depth in olive trees varies in accordance with the soil characteristics and the olive tree age (Fernández, Moreno, Cabrera, Arrue & Martín-Aranda, 1991), the soil samples were collected at 30 cm and 60 cm depth because the plantations are in its juvenile step –with the exception of var. Verdial de Huévar – and the tree roots were estimated to be around 40-50 cm depth. Table 2 shows the mean and standard deviation of the concentrations of 34 elements at those two depths. The half of them was quantified at concentrations higher than 10 mg/kg. In regards to elements quantified at low concentrations, it is remarkable the low concentrations of Mn that were lower than 1 mg/kg in all cases. The concentrations of eighteen elements are higher in the samples of soils collected at 60 cm depth (Table 2) (p
