Food Research International 54 (2013) 2025–2038

Contents lists available at ScienceDirect

Food Research International

j ourna l homepage: www.e lsev ie r .com/ locate / foodres

Authenticity of olive oil: Mapping and comparing official methods andpromising alternatives

Ramón Aparicio a,⁎, María T. Morales b, Ramón Aparicio-Ruiz b, Noelia Tena a, Diego L. García-González a

a Instituto de la Grasa (CSIC), 4 Padre Garcia Tejero, 41012 Sevilla, Spainb Department of Analytical Chemistry, Universidad de Sevilla, 2 Prof. García-González, 41012 Sevilla, Spain

⁎ Corresponding author. Tel.: +34 954 611550.E-mail address: aparicio@cica.es (R. Aparicio).

0963-9969/$ – see front matter © 2013 Elsevier Ltd. All rihttp://dx.doi.org/10.1016/j.foodres.2013.07.039

a b s t r a c t

a r t i c l e i n f o

Article history:Received 30 April 2013Received in revised form 3 July 2013Accepted 9 July 2013

Keywords:Olive oilAuthenticityChromatographySpectroscopyIsotopic ratioTrade standards

Trade standards are continuously updated to give plausible solutions to situations created by fraudsters whoapply the most sophisticated procedures to their objectives of olive oil adulteration. Clustered inside targetedand profiling approaches, methods based on spectroscopic, isotopic and chromatographic techniques arereviewed. Chromatographic methods, most of them being official methods, compete with newer methodsbased on spectroscopic, isotopic and trace element techniques for ensuring that the pace of research in thedetection of malpractices is rapid enough.The speed of the analyses, the need of statistical interpretation of the results, the quality parameters of themethods, limit of detection of the adulterants, and the applicability range among others are on the basis forthe absolute and comparative analyses of the most known methods, which results are unpacked in the paper.The new frontiers of research in the field of olive oil authenticity are also dissected together with the challengesfor the near future.The extensive and deep analysis of themethods for quantifying the chemical compounds responsible for olive oilauthenticity will contribute to a better comprehension of the complex analytical world of olive oil for the analystworking with this food product for the first time, as well as for experienced professionals.

© 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The high price of olive oil and its reputation as a healthy and delec-table oil makes it a preferred target for fraudsters. Thus, adulterationmay take place not only by accidental contamination during the stagesof oil processing but even more often by deliberate mislabeling of lessexpensive olive oil categories or by the addition of less expensive edibleoils to virgin olive oil for the purpose of financial gain.

Numerous adulterants have been found in virgin olive oil, and theyvary from refined olive oil, deodorized virgin olive oils, raw olive-pomace oil and synthetic olive oil–glycerol mixtures to almost all seedoils (e.g. maize, cottonseed, hazelnut, rapeseed and sunflower). In fact,the admixtures of expensive olive oils with less expensive and lower-grade oils have been traditionally more than a potential problem incountries that manufacture seed oils and import olive oils. Thisprocedure is harmful for new consumerswho buy olive oil for its healthbenefits and strict purity control and are surprised receiving oil thatdoes not fulfill their expectations (García-González, Aparicio-Ruiz, &Aparicio, 2009).

Several international institutions (e.g. International Olive Council –IOC – and AntifraudUnit of the EuropeanUnion –OLAF – amongothers)are actively involved in anti-fraud regulations, which are focused on

ghts reserved.

tighter control of producing and importing countries, clear definitionsfor olive oil products, uniform labeling regulations, and rapid, easy andaccurate instrumental techniques and analytical methodologies. Thefinal objective is to avoid any image of a hypothetical uncontrolled dis-tribution of adulterated olive oil into themarket and to ensure fair tradeas well as the safety and consumer protection.

Advances in knowledge and technology, which have been needed inthe detection of malpractices by fraudsters, have required a consider-able investment although no rapid and universal method has been offi-cially recognized for all the authenticity issues yet; e.g. adulteration,mislabeling, and misleading among others.

At this point, the most accepted definition for the genuineness of afood product is: “A product is authentic as long as it is firstly describedaccurately by the label and secondly complies with the current legisla-tion in force in the country where it is marketed or sold” (Lees, 1998).An authentic food is, in consequence, one which is truly derived from aspecified source where the term source must be clearly defined (e.g. aparticular category of olive oil). The ample number of olive oil categories,which are clearly defined by current regulations (EC, 2013; IOC, 2011),and the numerous edible oils that can be used in adulterations requireof a plethora of analytical techniques and methods for carrying out astrict olive oil authenticity control.

This work is not a systematic revision of methods and instrumenta-tion used in the authenticity of olive oil. There are several interestingworks already published for this purpose (Aparicio & Aparicio-Ruiz,

2026 R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

2000; Aparicio, Aparicio-Ruiz, & García-González, 2007; Aparicio, Conte,& Fiebig, 2013; Aparicio et al., 1998; Ben-Ayed, Kamoun-Grati, & Rebai,2013; Dais & Hatzakis, 2013; Frankel, 2010; García-González, Baeten,Fernández-Piernas, & Tena, 2013). In contrast to these reviews, thiswork presents an extensive and deep analysis of themost relevant com-pounds used as targeted analytes for virgin olive oil characterizationand authentication, the analytical problems derived from their determi-nation by targeted and profiling methods, and a critical approach oftheir utility today, and future trends.

2. Discussion

2.1. Official standards and promising alternatives: the state-of-the-art

Three are the different ordinances and legal sources that have tradi-tionally ruled olive oil production and international trading althoughthere are also regulations in force in some producer states (e.g. Australia,California). IOC trade standards have traditionally helped to design andimprove the other ordinances with which it shares much more than90%. Particular regulations exist within the European Union (EU) tocontrol its olive oil market which represents 72% of the world produc-tion. As olive oils are subjected to worldwide trade, a further level ofregulation is needed, and this is provided by the Codex AlimentariusCommission. EU regulations are in force for EU countries while IOCtrade standards and Codex standards are agreements that signatorycountries voluntarily have accepted to comply. They establish the limitsfor each quality and purity criterion, including the precision values ofthe applied methods, although IOC trade standards are more specificthan the Codex Alimentarius standards in, for instance, the labelingaspects (IOC, 2011).

Today, most of the methods refer to ISO (International Standardiza-tion Organization)methods although some still refer to AOCS (AmericanOils Chemists' Society) methods. In fact, IOC is an official Liaison-Member of ISO/TC 34/SC 11 – Animal and vegetable fats and oils,which is a Subcommittee of ISO/TC-34 – Food products. Based on thisrelationship, IOC methods, which were developed for olive oil tradestandards, have been able to be compared with already checked ISOmethods that were developed for a wide number of fats and oils; some-times, however, the peculiar characteristics of olive oils have allowedsome IOC particular methods to be accepted as ISO norms. The conse-quence is that the IOCmethods are appliedworldwide for the analyticalcontrol of olive oils though alternative methods fostered by other asso-ciations should be taken into account as well; i.e. AOAC International,Federation of Oil Seeds and Fats Association (FOSFA) and InternationalUnion of Pure and Applied Chemistry (IUPAC).

In those days when the adulteration was very simple, the detectionof adulterants was relatively easier than nowadays. The fraudulentpractices, however, becamemore complex and ingenious as technologyadvanced and it was widely accessible to everybody. In the objective ofdetermining chemical compounds that can be markers for olive oilauthenticity, many modern techniques have been proposed byresearchers and technologists. Thus, they have proposed methodsbased ongas, liquid, gas–liquid, quiral, silver-ion,mass, and supercriticalfluid chromatographies, stable carbon isotope ratio analysis (SCIRA),excitation–emission fluorescence (EEFS) and total synchronous fluores-cence (TSyF), pyrolysis–mass spectrometry (PyM), nuclear magneticresonance spectroscopy (NMR), and infrared and Raman spectroscopyamong others. However, any of those methods needs to be approvedor recommended by international associations to become an officialstandard. In fact, most of those numerous proposed methods can onlydetect adulterations greater than 10%, which scarcely represent anyadvantage over current tests and official methods.

Conceptually the methods can be naturally divided into “targetedanalyses” – based on definite information obtained from the frac-tionation of olive oil components – and “profiling or non-targetedanalyses”, which relies on the simultaneous contribution from

many known or unknown analytes belonging to a predefined meta-bolic pathway (Baeten et al., 2005; Rezzi et al., 2005). The formeranalyses, which quantify and identify series of chemical compounds,analyte by analyte, search for compounds that do not appear, or onlyat trace levels, in genuine olive oils but do appear in adulteratedones. As these analyses reveal under what circumstances theseanalytes appear in the adulterated oils, the information can also beused to remove or diminish the amount of those selected markersin an improved adulteration process; e.g. adding desterolized seedoils that cannot be detected with methods based on the detectionof sterols. This approach requires not only considerable investmentin perfecting the classical methods, or in developing new methods,but also in ensuring that the pace of research in the detection of mal-practices is rapid enough.

The profiling approach, which typically does not differentiatebetween analytes and sometimes neither quantify them, aims to rapidlydetermine the genuineness of olive oils based on information frommulti-target screening methods, which are gaining popularity as alter-native to targeted approaches based on gas liquid chromatography(GLC) or liquid chromatography (HPLC). In the case of profiling tech-niques, the fraudsters have no information since there is not a particularmarker but the analysts may have problems interpreting the informa-tion because multivariate statistical procedures are needed to arrive atcorrect conclusions, in addition obviously of plausible chemical or bio-chemical explanations, if analysts want to avoid that the authenticityis not based on random parameters or noise.

The current limits for the physicochemical parameters involved ineach purity or quality criterion (Tables 1a–1b) are results achieved,however, from the chromatographic techniques. It is so because oflower cost, rapid implementation and development, more versatilityfor quantifying diverse analytes, and superior reproducibility of thechromatography in comparison with other proposed techniques.

2.2. Targeted approaches

A standard method needs several years to be endorsed as officialmethod from its submission to the regulatory institutions. The reasonsfor that period of time can be found, among others, in that the proposalsare, in general, hyper-optimists due to a lax application of the statisticalprocedures, an inadequate selection of the validation samples or a casu-al relationship between the adulteration and the selected chemicalmarkers.

The chemical compounds, whose contents allow determining thedifference between genuine and adulterated olive oils with regard totheir designations, are shown in Tables 1a–1b. The chemical composi-tion of olive oil has been traditionally clustered into major and minorcompounds; the former are, in large part, responsible for the olive oilmain characteristics while the latter are markers for their peculiarities.This section, which has been structured around the series of chemicalcompounds that are currently used in olive oil authentication, analyzesthe series from three viewpoints: a) the main reasons for analyzingthem; b) the current standards, with practical comments and sugges-tions if possible, for quantifying them; and c) potential alternatives toofficial methods for determining them.

2.2.1. Fatty acidsFatty acids are, with a few exceptions, the major components of any

oil or fat. In small amounts they are present as free fatty acids butusually form esters, most often with glycerol, to produce glycerides(mono-, di- and tri-acylglycerols) and phospholipids but they can alsoform esters with aliphatic alcohols of linear structure (waxes) orterpenic structure (terpene and sterol esters).

2.2.1.1. Reasons for analyzing these compounds. The knowledge of thefatty acid composition has widely been used for characterizing edibleoils since 1960s when seed oils with a modified fatty acid composition

Table 1aOlive oils purity and quality characteristics according to the International Olive Council (IOC, 2011).

Designations (1) (2) (3) (4) (5) (6a) (7) (8) (9b)

Extra-virgin olive oil ≤0.05 ≤0.05 ≥1000 ≤4.5 ≤250 ≤0.10 ≤ 0.2 B ≤2.50c

Virgin olive oil ≤0.05 ≤0.05 ≥1000 ≤4.5 ≤250 ≤0.10 ≤ 0.2 B ≤2.60c

Ordinary virgin olive oil ≤0.05 ≤0.05 ≥1000 ≤4.5 ≤250 ≤0.10 ≤ 0.2 B –

Lampante virgin olive oil ≤0.10 ≤0.10 ≥1000 ≤4.5c ≤300c ≤0.50 ≤ 0.3 C –

Refined olive oil ≤0.20 ≤0.30 ≥1000 ≤4.5 ≤350 – ≤ 0.3 C –

Olive oil ≤0.20 ≤0.30 ≥1000 ≤4.5 ≤350 – ≤ 0.3 B –

Crude olive–pomace oil ≤0.20 ≤0.10 ≥2500 ≤4.5d N350d – ≤ 0.6 ≤1.4% –

Refined olive–pomace oil ≤0.40 ≤0.35 ≥1800 ≤4.5 N350 – ≤ 0.5 ≤1.4% –

Olive–pomace oil ≤0.40 ≤0.35 ≥1600 ≤4.5 N350 – ≤ 0.5 ≤1.2% –

Designations (10b,e) (11b) (12) (13) (14) (15) (16) (17) (18) (19)

Extra-virgin olive oil ≤0.22 ≤0.01 ≤0.8 ≤20 ≤0.5 ≤0.5 ≤0.1 ≤4.0 bCamp ≥93.0Virgin olive oil ≤0.25 ≤0.01 ≤2.0 ≤20 ≤0.5 ≤0.5 ≤0.1 ≤4.0 bCamp ≥93.0Ordinary virgin olive oil ≤0.30 ≤0.01 ≤3.3 ≤20 ≤0.5 ≤0.5 ≤0.1 ≤04.0 bCamp ≥93.0Lampante virgin olive oil – – N3.3 – ≤0.5 ≤0.5 ≤0.1 ≤4.0 – ≥93.0Refined olive oil ≤1.10 ≤0.16 ≤0.3 ≤5 ≤0.5 ≤0.5 ≤0.1 ≤4.0 bCamp ≥93.0Olive oil ≤0.90 ≤0.15 ≤1.0 ≤15 ≤0.5 ≤0.5 ≤0.1 ≤4.0 bCamp ≥93.0Crude olive–pomace oil – – – – ≤0.5 ≤0.5 ≤0.2f ≤4.0 – ≥93.0Refined olive–pomace oil ≤2.00 ≤0.20 ≤0.3 ≤5 ≤0.5 ≤0.5 ≤0.2f ≤4.0 bCamp ≥93.0Olive–pomace oil ≤1.70 ≤0.18 ≤1.0 ≤15 ≤0.5 ≤0.5 ≤0.2f ≤4.0 bCamp ≥93.0

Note:

(1) Trans-oleic fatty acid (%).(2) Sum of trans-linoleic & linolenic fatty acids (%).(3) Total sterol content (mg/kg).(4) Erythrodiol and uvaol content (% total sterols).(5) Wax content: C40 + C42 + C44 + C46 (mg/kg).(6) Stigmastadiene contents (mg/kg).(7) Difference between the actual and theoretical ECN42 triacylglycerol content.(8) Content of 2-glyceryl monopalmitate; B, ≤0.9 if total C16:0 ≤ 14.0% or ≤1.0 if C16:0 N 14:0%; C, ≤0.9 if total C16:0 ≤ 14.0% or ≤1.1 if C16:0 N 14:0%.(9) Absorbency in ultra-violet at K232.

(10) Absorbency in ultra-violet at K270.(11) Absorbency in ultra-violet (ΔK).(12) Free acidity (%m/m expressed in oleic acid).(13) Peroxide value (in milleq. peroxide oxygen per kg/oil).(14) Δ7-Stigmastenol (%).(15) Cholesterol (%).(16) Brassicasterol (%).(17) Campesterol (%).(18) Stigmasterol (%).(19) The value of β-Sitosterol is calculated as: Δ5,23-Stigmastadienol + Clerosterol + β-Sitosterol + Sitostanol + Δ5-Avenasterol + Δ5,24-Stigmastadienol.a Total isomers which could (or could not) be separated by capillary column.b Quality characteristics.c When the oil has a wax content of between 300 mg/kg and 350 mg/kg it is considered a lampante olive oil if the total aliphatic alcohol content is ≤350 mg/kg or if theerythrodiol + uvaol content is≤3.5%.

d When the oil has a wax content of between 300 mg/kg and 350 mg/kg it is considered a crude olive-pomace oil if the total aliphatic alcohol content is N350 mg/kg and if theerythrodiol + uvaol content is N3.5%.

e Maximumwavelengths of 268 if iso-octane is used and of 270 nm if cyclohexane is used.f Limit raised to ≤0.2 for olive-pomace oils.

2027R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

similar to olive oil appeared; e.g. high oleic safflower and sunflower oils,and high oleic rapeseed and soybean oils more recently. Furthermore, ahigh percentage of myristic acid in labeled olive oils may indicate thepresence of seed oils, mainly fractionated palm oil, while high percent-ages of linolenic or eicosanoic or eicosenoic or behenic acids are associ-ated to the presence of soybean or low erucic rapeseed oils, and a highvalue of lignoceric to peanut oil.

On the other hand, the presence of trans isomers of oleic, linoleic andlinolenic acids in percentages above the approved levels (Table 1a)is related to the presence of refined oils obtained from, for example, es-terified olive oils, (partially-)hydrogenated seed oils or seed oilsdesterolized at high temperatures, among others.

2.2.1.2. Official methods: comments and suggestions. There is broadagreement on the methodology for the quantification of fatty acids(Tables 1a–1b) although special attention should be paid to the methyl-ation step, which can be carried out by acid or alkaline catalysis. Theanalyst should bear in mind that methylation should be acid in case ofa medium or a higher concentration of free fatty acids since alkalinecatalysis is effective in trans methylation only.

2.2.1.3. Alternatives to officialmethods.While GC is the techniqueused forfatty acid quantification (Table 2a), spectroscopy is the alternative to theofficial method for quantifying trans fatty acids (Table 2a) in addition tothe determination of iodine value, saponification number and humiditycontent. Reasons can been found in its rapid sample screening. Ramanbands near 1656 cm−1 and 1670 cm−1 are the consequence of cis andtrans isomer contents while Mid-IR bands around 860–990 cm−1 and1650–1655 cm−1 respectively correspond to the contents of trans andcis isomers in olive oils (Ismail, Cocciardi, Alvarez, & Sedman, 2006;Ismail, Nicodemo, Sedman, van de Voort, & Holzbaur, 1999).

In the field of NMR (nuclear magnetic resonance), 1H NMR is thebest for the quantification of fatty acids although mathematical algo-rithms need to be used for the quantification of the sums of saturated,monounsaturated, and polyunsaturated fatty acids from appropriatesignal intensities. As the signals overlap, the individual determinationof fatty acids cannot be carried out like GC with the exception oflinolenic whose methyl protons are detected at δ 0.96. The isomers cisand trans have also been determined by using the allylic methyleneprotons adjacent to cis and trans double bonds (Sedman, Gao,García-González, Ehsan, & van de Voort, 2010). 13C NMR, on the

Table 1bOlive oils purity and quality characteristics according to the International Olive Council (IOC, 2011).

Designations (20a) (21a) (22a) (23) (24) (25) (26) (27) (28) (29) (30)

Extra-virgin olive oil Mf N 0 Md = 0 (b) ≤15 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.2 ≤0.2 (c)Virgin olive oil Mf N 0 0 b Md ≤ 3.5 – ≤15 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.2 ≤0.2 (c)Ordinary virgin olive oil – 3.5 b Md b 6.0d ≤15 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.2 ≤0.2 (c)Lampante virgin olive oil – Md N 6.0 – ≤15 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.2 ≤0.2 (c)Refined olive oil – – – ≤15 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.2 ≤0.2 (c)Olive oil – – – ≤15 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.2 ≤0.2 (c)Crude olive-pomace oil – – – ≤30 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.3 ≤0.2 (c)Refined olive-pomace oil – – – ≤30 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.3 ≤0.2 (c)Olive-pomace oil – – – ≤30 ≤0.05 ≤1.0 ≤0.6 ≤0.4 ≤0.3 ≤0.2 (c)

Note:

(20) Organoleptic assessment: median of fruity attribute (Mf).(21) Organoleptic assessment: median of defect (Md).(22) Fatty acid methylesters (FAMEs) and fatty acid ethylesters (FAEEs).(23) Unsaponifiable matter (g/kg).(24) Myristic acid (% m/m methylesters).(25) Linolenic acid (% m/m methylesters).(26) Arachidic acid (% m/mmethylesters).(27) Eicosenoic acid (% m/mmethylesters).(28) Behenic acid (% m/mmethylesters).(29) Lignoceric acid (% m/mmethylesters).(30) Other fatty acids (% m/m methylesters).a Quality characteristics;b (FAME + FAEE) ≤ 75 mg/kg or [75 mg/kg b ∑(FAME + FAEE) ≤ 150 mg/kg and (FAEE/FAME) ≤ 1.5]c Palmitic: 7.5–20.0; palmitoleic: 0.3–3.5; heptadecanoic: ≤0.3; heptadecenoic: ≤0.3; stearic: 0.5–5.0; oleic: 55.0–83.0; linoleic: 3.5–21.0.d Or when the median of the defect is less than or equal to 3.5 and the median of the fruity attribute is equal to 0.

2028 R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

other hand, is used when information about trans fatty acids is re-quired because the distance (5 ppm) between signals of the allyliccarbons of cis and trans double bonds (Gao et al., 2009). NMR instru-ments, however, require a large initial outlay and their methodsneed more time for the analysis than GC methods.

2.2.2. TriacylglycerolsA large extent of olive oil main characteristics is the responsibility of

triacylglycerols (trihydric alcohols esterified with three fatty acids)(TAG). The molecular structure of each individual TAG species hasthree basic characteristics: (i) the total carbon number (CN), which isthe sum of the alkyl chain lengths of each one of the three fatty acids(FAs); (ii) the degree of unsaturation in each FA; and (iii) the positionand configuration of the double bonds in each FA (Buchgraber,Ulberth, Emons, & Anklam, 2004). Thus, the large number of possibleFA combinations on the glycerol backbone of TAGs makes the TAGsvery revealing although, for this same reason, their analysis is a verychallenging task.

2.2.2.1. Reasons for analyzing these compounds.At the beginning of 1980s,Gegiou and Georgouli (1983) investigated the presence of re-esterifiedoils in olive oils by the ratios of TAGs (i.e. 1-oleo-2,3-dipalmitin to 1,3-dipalmito-2-olein, 1,3-dioleo-2-palmitin to 1-palmito-2,3-diolein, and1,3-dioleo-2-stearin to 1-stearo-2,3-diolein). Later, Salivaras andMcCurdy (1992) explored the information of TAGs in the detection ofadulterations of virgin olive oil with canola oil from 7.5% whileEl-Hamdy and El-Fizga (1995) detected adulterations with about 1% oflinoleic-rich vegetable oils (soybean, sunflower and corn oils) thoughthe detection of olive oil admixed with non-rich linoleic acid seed oilsis much more difficult.

In this decade, the interest of researchers was focused on the differ-ences between the experimental composition of TAGs and theirtheoretical composition determined from fatty acids according to their1,3-random 2-random distribution (Cortesi, Rovellini, & Fedeli, 1990).From a theoretical viewpoint, the difference between both values shouldbe zero for an authentic olive oil sample but the analytical error – for ex-ample, only some theoretical TAGs showgoodmathematical significance(García-Pulido & Aparicio, 1993) – directed the attention to the differ-encebetween the real and theoretical ECN42 (equivalent carbon number

42) triacylglyceride content (ΔECN42). The IOC trade standard (IOC,2011) and European regulation (EC, 2013) include a limit of ≤|0.2| forΔECN42 determined in edible virgin olive oils (≤|0.3| for lampante virginolive oil), which makes possible detecting a fraudulent addition of seedoils to olive oil, even at low percentages.

A different approach to TAG analysis was focused on the discrimina-tion of natural TAGs from chemically synthesized ones that wereobtained from free fatty acids esterified with glycerol; synthesizedTAGs were sold as olive oil in the early 1970s. The biosynthesis of TAGin vegetable oils, however, does not allow that high concentration ofsaturated fatty acids as well as fatty acids with a carbon chain longerthan 18 carbon atoms to be present at 2-position of the molecule; 1%of such acids esterified at the 2-position of glycerol is enough to con-clude that the oil is not genuine. Thus, the analytical proposals focusedon the direct analysis of the reaction of lipolysis by capillary gas chroma-tography that Lercker, Moschetta, Caboni, and Frega (1985) improvedby removing the step of separation by TLC, which made the analyticalprocess easier and faster. The method was validated, and is includedin the Italian Regulation (NGD, 2002).

2.2.2.2. Official methods: comments and suggestions. TAGs can be separatedbyHPLC, according to the number of carbon atoms and their unsaturation,using diverse sample preparations, stationary and mobile phases,columns and detectors. The current discussion inside IOC is, however,focused on the mobile phase: acetone/acetonitrile (1:1 v/v) adversuspropionitrile (Table 2a). A mobile phase of acetone/acetonitrile, inisocratic elution, allows that critical pairs are effectively separated butsaturated long-chain TAGs are not sufficiently dissolved with thisphase. The isocratic elution with propionitrile (Fiebig, 1985; Moreda,Pérez-Camino, & Cert, 2003) results in an improved separation of thegroups of peaks clustered as ECN42.

The selected particle size for reverse phase (RP)-18 column was4 μmbecause it was observed an increase of the olive oil TAG resolutionas the particle size diminished; in fact, the use of columns with a smallparticle size minimizes the problems of integration and quantificationof HPLC peaks. The refractive index (RI) detector is also themost appro-priate despite its drawbacks of low sensitivity, and intensity of the re-sponse depending on the saturation level of TAGs. Besides RI detectorsshould not be placed near sun-lit windows and the analyst should be

Table 2aThe standard methods for quantifying acyl lipids and fatty acids.

Compounds Technique Sample preparation Chromatographic characteristics

Triacylglycerols HPLC-RI 0.12 g oil in 0.5 mL hexane is charged into SPE-cartridge Mobile phase flow-rate (0.6 to 1.0 mL/min)Oven temperature: 20 °C

(1 g of Si) and solution pulled through and, then, elutedwith 10 mL hexane-diethylether (87:13 v/v).

Mobile phase: propionitrileColumn: RP-18 (4 μm)Detector: RI

HPLC-RI 0.5 g oil in 10 mL acetone or acetone/chloroform (1:1 v/v). Mobile phase flow-rate (0.6 to 1.0 mL/min)Oven temperature: 25 °CMobile phase: acetone/acetonitrile (1:1 v/v)Column: RP-18 (4 μm)Detector: RI

2-Glyceryl monopalmitate (%) GC-FID Hydrolysis with pancreatic lipase.Separation by LC or SPE.Require silanization

Column: capillary 12 m × 0.32 mm × 0.10–0.30 μmPhase: methylpolysiloxane or 5% phenylmethylpolysiloxane.Carrier gas: hydrogenOperation conditions: temperature gradientInjection mode: on-column

Fatty acids GC-FID Total fatty acids:Methylation with cold methanolic solution of KOH or doublemethylation in a methanolic medium with alkaline and acid catalysis.

Column: capillary 25–100 m × 0.2–0.8 mm × 0.1–0.2 μmStationary phase: polyglycol, polyester or cyanopropylsiliconeCarrier gas: hydrogen

GC-FID Trans fatty acids:Methylation with cold methanolic solution of KOH

Operation conditions: temperature gradientInjection mode: split

GC-FID Fatty acid in the 2-position:Hydrolysis with pancreatic lipase previously and methylation ina methanolic mediumwith alkaline and acid catalysis.

Waxes GC-FID Isolation on LC Si-column Column: capillary 12–15 m × 0.25–0.32 mm × 0.1–0.3 μmStationary phase: 5% phenylmethylpolysiloxaneCarrier gas: hydrogenOperation conditions: temperature gradientInjection mode: split or on-column

Note: GC, gas chromatography; FID, flame ionization detector; HPLC, high performance liquid chromatography; RI, refractive index detector; SPE, solid phase extraction.

2029R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

careful about possible bubbles of gas in the solvents, leaks in the systemand back pressure HPLC pump since they affect to chromatographicbaseline (Christie, 1992). The use of a thermostated systemand isocraticelution is recommended, and also to pass thepre-solved acetone samplethrough a 0.2 μm pore size filter, when determining raw olive-pomaceoils, to remove possible precipitates that can shorten the effective lifeof the chromatographic column.

The initialmethod for the determination of the percentage of palmiticacid at the 2-position of the triacylglycerols was substituted by the con-tent of 2-glycerylmonopalmitate (IOC, 2006a). Themethod and themax-imum percentages of the compound for each olive oil category wereincluded in the EURegulation702/2007 (EC, 2007) after analyzing the re-sults of collaborative trials whose aimwas to determine the content of 2-glyceryl monopalmitate in genuine extra virgin olive oils with high con-tent of palmitic acid. The result was to assign the value (%) for this com-pound according to the percentage of palmitic acid (Table 1a).

2.2.2.3. Alternatives to official methods. GLC is also attractive for theanalysis of acylglycerols because of the linear response and selectivityalthough GLC is not exempt of problems such as, for example, thehigh temperature needed for a suitable separation of TAGs. Thus, thecapillary columns (30 m × 0.25 mm × 0.1–0.15 μm) coated with 65%phenylsiloxane + 35% methylsiloxane have been suggested becausethey endure high temperatures for a long time. All the existing studiesof GLC analysis of olive oil TAGs suggest to use a cold on-column injec-tion although the analysis takes longer. As the time that TAGs are insidethe column increases, the results are both resolution loss and peakwid-ening; for example, trilinolein (LLL) results in a poor peak shape and aless reliable measurement than using RP-HPLC.

Analysis of TAGs of oil samples by liquid chromatography coupledwith mass spectrometry (HPLC–MS) is a widespread method becauseMS supplies identification of non- or partially resolved HPLC peaks. Infact, MS provides detailed information about the FAs composition ofTAG molecules though only a few ionization techniques are suitablefor coupling with HPLC. Thus, electrospray ionization (ESI) is an

excellent technique because it enables the identification of the individ-ual acyls without the need for authentic reference standards while at-mospheric pressure chemical ionization (APCI) coupled to HPLC isvery effective because it allows identifying the positional isomers(Holčapek, Jandera, Zderadička, & Hrubá, 2003); other signals at lowermasses, related to diacylglycerols (DAG) fragments, have been studiedby Jakab, Héberger, and Fornács (2002).

Recently, matrix-assisted laser desorption ionization time-of-flightmass spectrometry (MALDI-TOF-MS) is being used for a rapid differen-tiation of edible oils based on the detection of TAG molecules as theirsodium adducts and, sometimes, as their potassiated adducts, whichrender low intensities. The method possesses a good resolution in TAGmass range which contributes to distinguish vegetable oils whose TAGmolecules differ only by their degrees of unsaturation (Cozzolino & DeGiulio, 2011). An important advantage of this method is the rapid sam-ple preparation because of the absence of analyte purification, chemicalmodification or derivatization. Concerning applications in specific adul-terations, an approach based on the laser desorption ionization (LDI)coupled TOF-MS has been proposed for the detection of adulterationof olive oil with sunflower oil (Calvano, Palmisano, & Zambonin, 2005).

2.2.3. WaxesThe termwaxhas been applied to an ample variety of plant products

containing several kinds of fatty materials that are synthesized in theepidermal cells of the olives. The pathways for the biosynthesis of waxcomponents comprise an acyl reduction, which yields primary alcoholsand wax esters, and a decarbonylation pathway that synthesizessecondary alcohols, aldehydes, alkanes and ketones (Christie, 2012).Wax esters are, however, the most common form of the waxes, andthey consist of fatty acids esterified to long-chain alcohols with similarchain-lengths (Christie, 2011). The main waxes found in olive oil areesters of even numbered of carbon atoms from C36 to C46.

2.2.3.1. Reasons to analyze these compounds. Because waxes are synthe-sized from very-long-chain saturated fatty acids, their presence in

2030 R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

olive oils is of interest because their concentrations vary among the oliveoil categories, and the information about its presence can be used as indi-cator of purity (EC, 2013). In fact, the contents of waxes C36 and C38 areusually higher than thewaxes C40, C42, C44 andC46 inVOOs, on the con-trary of the concentrations in refined and olive-pomace oils (Table 1a).However, differences in the content of waxes cannot be used as a uniquecriterion to detect the presence of olive-pomace oil in olive oil but togeth-er with the amount of the erythrodiol and uvaol (IOC, 2011).

2.2.3.2. Official methods: comments and suggestions. All the methods forthe determination of waxes consist of a separation from the otherlipid constituents using silica gel chromatography and the subsequentquantification using GC (Table 2a). Although there is a specific methodfor the determination of waxes in olive oil (IOC, 2007), a new methodthat allows the simultaneous determination of waxes, fatty acid methylesters and fatty acid ethyl esters by capillary gas-chromatography (IOC,2010) is widely applied in laboratories because it saves time when theobjective is to determine the olive oil genuineness. Themethod is recom-mended for distinguishing between olive oil and olive-pomace oil, forthe detection of the presence of lower quality oils (ordinary, lampante)in extra-virgin olive oils, and for the detection of the fraudulent additionof some deodorized oils to extra virgin olive oils.

Although semipolar columns SE-54 or SE52 are used, better resolu-tion is obtained with phenyl-methyl-silicone columns (Table 2a),which can endure higher temperatures. Anyway, it is advisable to con-dition the column when used for the first time. It can be done bymeans of a gradual warm-up until it reaches 350 °C, after approximate-ly 4 h, and maintaining this temperature for at least 2 h. Once theinstrument has been regulated with the operating conditions (IOC,2007), the signal has to be adjusted in sensitivity at least twice higherthan that required for the analysis. Furthermore, the baseline shouldbe linear, with no peaks of any kind, and must not have any drift.

2.2.3.3. Alternative to official methods. Furthermore the replacing of silicagel columns by solid phase extraction (SPE) cartridges (1 g of silica gel)for the sample purification because they need smaller amount of sampleand a reduced volume of elution solvent, it has been proposed a newmethod based on the isolation of alkyl esters and waxes by SPE andtheir analysis by capillary column (DB5; 5–12 m long) gas chromatog-raphy with on-column injector and FID detector (Cert et al., 2011).The method seems to be more rapid than current standard with lowerconsume of organic solvents although there is no information aboutits precision, limit of quantification and the other analytical parameters.

Another analytical alternative is based on the use of TOTAD (ThroughOven Transfer Adsorption Desorption) interface (Aragón, Toledano,Cortes, Villén, & Vázquez, 2011). The oil, with an internal standard (C32wax ester) diluted in n-heptane, is injected directly with no sample pre-treatment step other than filtration. Normal-phase liquid chromatogra-phy (NPLC) separates the wax ester fraction from the triglycerides andthe TOTAD interface transfers it to the GC to be analyzed. Biedermann,Haase-Aschoff, and Grob (2008) also proposed that samples were pre-pared by NPLC but, in this proposal, the pre-column is attached to theinlet of the column of the GC × GC instrument by means of a press-fitconnector. The first dimension is performed with a PS-255 column(20 m × 0.25 mm i.d) and the second-dimension with a SOP-50 (50%phenyl polysiloxane) (1.5 m × 0.15 mm i.d).

2.2.4. SterolsSterols make up an extensive series of compounds with analo-

gous molecular structure, more than 200 have been reported inplants, that are grouped into three classes (4,4-dimethylsterols,4-monomethylsterols, and 4-desmethyl sterols) according to the num-ber ofmethyl groups at the C-4 position. They aremade up of a tetracycliccyclopenta[a] phenanthrene ring and a longflexible side chain at the C-17carbon atom (Piironen, Lindsay, Miettinen, Toivo, & Lampi, 2000). 4,4-dimethylsterols and 4-monomethylsterols are metabolic intermediates

in the biosynthetic pathway leading to end-product, 4-desmethylsterols.Thus, 4-desmethylsterols, the most abundant class in olive oil, are com-monly called phytosterolswhile 4,4-dimethylsterols are called triterpenicalcohols and 4-monomethylsterols are named methylsterols.

2.2.4.1. Reasons to analyze these compounds. Sterols, which comprise amajor portion of the unsaponifiable matter, are the most important ofthe olive oil minor compounds in authentication purposes. Because ofthe ranges of concentration of 4-desmethylsterols are characteristic ofthe genuineness of several edible oils, they are used widely and reliablyto detect fraudulent mixtures of olive oil with other oils. For example,rapeseed oils contain significant levels of brassicasterol while safflowerand sunflower seed oils contain high levels of Δ7-stigmastenol, and, inthe case of olive oil, sterols have a profile of high levels of β-sitosteroland Δ5-avenasterol, low levels of campesterol and stigmasterol.

The content of free brassicasterol would allow detection of theaddition of 2% rapeseed oil to olive oil. The addition of 5% sunflower oil in-creases the stigmasterol, campesterol andΔ7-stigmasterol concentrationssignificantly above those found in non-adulterated olive oils. An additionof 10% soybean oil to olive oil is easily detected by the concentration offree campesterol and stigmasterol since their concentrations will double,while an olive oil spiked with 10% rapeseed oil increases the concentra-tion of free campesterol by five, and there are evident increments in theconcentration of γ-tocopherol and campesteryl-C18-ester that are al-most negligible in the olive oils. The addition of 10% grape seed oilwould also increase the concentrations of free campesterol and stig-masterol. The profile of sterols is known as the “fingerprint” of theolive oil because also it varies among the cultivars (García-Gonzálezet al., 2009).

Some of the sterols, however, are not detected in VOOs but in refinedolive oils (ROO) due to the refining process (e.g., Δ5,23-stigmastadienol)what allowsdistinguishingROOs fromVOOs. Thus,Δ5,24-stigmastadienolincreases its concentration during the refining processwhile the concen-tration of Δ5-avenasterol decreases (Amelotti, 1985). More recently, thequantification of free and esterified sterols was accepted as one of themost successful proposals for the detection of the adulteration of refinedolive oil with refined hazelnut oil (Bowadt & Aparicio, 2003). A doublemathematical model based on three free and esterified sterols(campesterol, Δ7-stigmastenol and Δ7-avenasterol) was able to detectthe presence of hazelnut oil in olive oil at percentages in the range6–8% (Mariani, Bellan, Lestini, & Aparicio, 2006). Another model system,although based on the sum of campesterol and stigmasterol, has beenrecently proposed for determining the presence of other vegetable oils(i.e. corn, soybean, sunflower, and cotton seed) in olive oil (Al-Ismail,Alsaed, & Ahmad, 2010).

When oils with low concentration of sterols (desterolized) are usedto adulterate olive oil, which makes difficult their quantification, theconfirmation can be obtained through the analysis of the dehydrationby-products of sterols, which increase spectacularly with that fraudu-lent practice.

2.2.4.2. Official methods: comments and suggestions. Sterols are deter-mined as the overall contribution of both the free and esterified formsas their trimethylsilyl (TMS) ethers or sterol acetates (Table 2b), whichimproves their peak shape, volatility and response factor (Cercaci,Passalacqua, Poerio, Rodriguez-Estrada, & Lercker, 2007). For their de-termination, it is necessary to isolate previously the unsaponifiable frac-tion, preferably by means of diethyl ether, since it guarantees the totalextraction of all the sterols (IOC, 2001a). On the other hand, the possi-ble mobile phases for TLC, benzene/acetone (95:5, v/v) and hexane/diethyl ether (65:35, v/v) yield excellent results, with no need formultiple developments.

The individual separation of 4-desmethylsterols, however, is notexempt from difficulties, which can arise in determination of thecomponents present at low concentration. Such difficulties, which areassigned to the incomplete separation of the sterol fraction in thin-

Table 2bThe standard methods for determining minor compounds.

Chemical series Technique Sample preparation Chromatographic characteristics

Sterols GC-FID Unsaponifiable-matter isolation TLC or HPLC.Requires silylation

Column: capillary (25–30 m × 0.25–0.32 mm × 0.15–0.30 μm)Stationary phase: 5% phenylmethylpolysiloxaneCarrier gas: hydrogenOperation conditions: isothermalInjection mode: split

Erythrodiol + uvaol GC-FID Unsaponifiable-matter isolation TLC or HPLC.Requires silylation

Column: capillary (25–30 m × 0.25–0.32 mm × 0.15–0.30 μm)Stationary phase: 5% phenylmethylpolysiloxaneCarrier gas: hydrogenOperation conditions: isothermalInjection mode: split

Aliphatic alcohols GC-FID Unsaponifiable-matter isolation TLC or HPLC.Requires silylation

Column: capillary (25–30 m × 0.25–0.32 mm × 0.15–0.30 μm)Stationary phase: 5% phenylmethylpolysiloxaneCarrier gas: hydrogenOperation conditions: temperature gradientInjection mode: split

Aliphatic hydrocarbons and sterenes GC-FID Unsaponifiable-matter isolation on LC Si-column Column: capillary (25–30 m × 0.25–0.32 mm × 0.15–0.30 μm)Stationary phase: 5% phenylmethylpolysiloxaneCarrier gas: hydrogenOperation conditions: temperature gradientInjection mode: split

Note: GC, gas chromatography; FID, flame ionization detector; HPLC, high performance liquid chromatography; TLC, thin layer chromatography.

2031R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

layer chromatography on silica gel, also produce high values of the coef-ficients of variation in repeatability and reproducibility. The problemmight be attributed to various causes resulting from an excess of theunsaponifiable deposited on the thin layer, the inadequate developmentof the plate, and defective scraping of the sterol band. Empirical resultssuggest that the optimumamount to deposit on the thin layer should bearound 200 μL if the content of 4,4-dimethylsterols is high. The objec-tive is to prevent that small amounts of cycloarthenol can cause errorsas it overlaps with Δ7-stigmastenol. The presence of 24-methylene-cycloarthanol, which yields a peak at tr = 1.27 in relation to β-sitosterol, warns the analyst that the percentage of Δ7-stigmasterol canbe caused in part by the presence of cycloarthenol.

Although the sterols are thought to be the fingerprint of the edibleoils, Grob, Giuffré, Biedermann, and Bronz (1994) found that numerousedible oils were adulterated with cheaper oils whose sterols had beenpreviously removed in order to make these admixtures undetectable.This fraudulent practice should be detected by the amount of transunsaturated C18-fatty acids, which are produced from normal cis acidsduring the deodorizing process, but some of the illegal processes ofremoving sterols are carried out without forming fatty acid trans-isomers. To avoid the adulteration, IOC proposed that the total valueof sterols should be expressed in mg/kg instead of percentages (IOC,2011). Thus, a genuine VOO cannot have less than 1000 mg/kg of ste-rols; below this value, it is reasonable to think that the olive oil mighthave been spiked with “desterolized” seed oils. However, the investiga-tion on this kind of adulteration also focused on the analysis of theproducts resulting from the degradation of the olefinic sterols, whichis discussed in the section of hydrocarbons.

The analytical determination of triterpene dialcohols is carried outtogether with sterols, by scraping their thin layer chromatography(TLC) band with the band of sterols (Table 2b). They are suitablemarkers detecting the presence of olive-pomace oil in olive oil althoughthere are genuine varietal olive oils that do not fit IOC trade standards(e.g. Spanish var. Verdial de Huevar). Results are calculated, in percent-ages, as the ratio between the sum of the areas of erythrodiol + uvaoland the sum of sterols + erythrodiol + uvaol. A VOO must not have avalue exceeding 4.5% (IOC, 2011). This information, however, has tobe assessed together with the wax or aliphatic alcohol content to ascer-tain the presence of olive-pomace oil.

2.2.4.3. Alternative to official methods. Because of the potential problems inusing TLC to isolate the sterol fraction from the rest of the unsaponifiable

material, as well as the high elution times needed, Grob, Lanfranchi, andMariani (1990) simplified the analysis of this and other fractions of theunsaponifiable fraction by the in-tandem technique LC-GLC. The tech-nique, which removes almost all the manual work, is widely applied byanalysts at present because it allows the quantification of free and esteri-fied sterols in short time. By adding pivalic acid anhydride, LC allowscollecting a fraction composed of freemonoalcohols and the fatty acid es-ters of these alcohols, which are transferred to GC on-line. It also allowscollecting erythrodiol and uvaol. The whole analytical procedure is iden-tical to traditional except for the substitution of TLC purification forHPLC, which involves considerable saving in both time and solvent.Later, Dionisi, Prodolliet, and Tagliaferri (1995) set up RP-HPLC with am-perometric detection to detect the addition of palm and grape seed oils toany tocotrienol-free vegetable oil (e.g., olive oil) at 1–2%.

In his permanent fight against the olive oil adulteration, Marianiet al. (2006) developed a method for an independent and adequatequantification of the nonpolar fraction (containing esterified sterols)and the polar fraction (containing free sterols) of great success in thedetection of the adulteration with hazelnut oil. The proposed methodcomprises three steps: (i) the preparation of the silica gel chromatogra-phy column, (ii) the separation of the free and esterified sterols, and(iii) the independent quantification of free and esterified sterols byGLC according to IOC standard method (IOC, 2001a).

2.2.5. HydrocarbonsSterol compositionwas an important obstacle to the availability of ge-

neticallymodified seeds–with fatty acid composition similar to olive oil –to be used in adulteration. However, Lanzón, Albi, Cert, and Gracián(1994) was pioneer in discovering the presence of hydrocarbons – as n-alkanes, n-hexacosadiene, stigmasta-3,5-diene, isomerization productsof squalene, isoprenoidal polyolefins produced from hydroxy deriva-tives of squalene and steroidal hydrocarbons from 24-methylenecycloarthanol – in refined oils as a consequence of the high temperaturein the refining process, compounds that were are not detected in anyvirgin olive oil after analyzing thousands of samples (Aparicio, Alonso,& Morales, 1994).

2.2.5.1. Reasons to analyze these compounds. The sterol dehydrationproducts, just above cited, are good markers for refined edible oils,either deodorized or desterolized. The quantification of campestatrienein virgin olive oils can indicate the presence of desterolized rapeseedoil while the ratio between degradation products of sistosterol

Table 3Main characteristics of the techniques proposed for the authentication of olive oils.

Characteristics Techniques

Separation GC, RPLC–GC, HPLC.Structural & pattern recognition NMR, MS, NIR, FTIR, FT-Raman, DSC, TG, SF.Stable isotope analysis IRMS.Trace element analysis ICP-AES, AAS, FAAS, ETA-AAS.In-tandem GC–MS, HPLC–MS, ICP-MS, CG × GC, LC × LC,

SFC, δ2H-EA–Py–IRMS, δ2H-GC–Py–IRMS.

Note: Gas-chromatography (GC); reversed phase liquid chromatography (RPLC); liquidchromatography (HPLC); nuclear mganic resonance (NMR); near infrared spectroscopy(NIR), Fourier transform infrared spectroscopy (FTIR) and Fourier transform Ramanspectroscopy (FT-Raman); isotope ratio mass spectrometry (IRMS); inductive coupledplasma-atomic emission spectroscopy (ICP-AES); atomic absorption spectroscopy (AAS);flame absorption spectroscopy (FAAS); electrothermal atomization-AAS; massspectrometry (MS), GC–MS, LC–MS and ICP-MS; elemental analyser–pyrolysis–isotope ratio mass spectrometry (δ2H-EA–Py–IRMS) and δ2H-GC–Py–IRMS; bidimesionalchromatography (GC × GC, LC × LC); supercritical fluid chromatography (SFC);synchronous fluorescence (SF); differential scanning calorimetry (DSC) andsimultaneous thermogravimetry (TG).

2032 R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

(stigmastadienes) and campesterol (campestadienes) is a good markerfor desterolized grapeseed or palm or soybean or sunflower oils.The quantification of 3,5-cholestadiene, resulting from cholesteroldehydroxylation, indicates the presence of deodorized lards. Further-more, the desterolized process involves isomerization so convertingΔ7-sterol into Δ8(14)-and Δ14-sterols, which were not quantified untilMariani, Venturini, and Grob (1995) set up a technique that avoidedtheir overlaps with Δ5-sterols and hence allowing the detection oflow percentages of a desterolized high oleic sunflower oil in olive oil.Δ7-sterols are themselves useful for detecting desterolized sunfloweroil in olive oil, and the presence of ergosterol has been associated tothe addition of low quality olive oils obtained from olives infested bymolds and yeasts (Biedermann, Grob, & Mariani, 1995; Biedermann,Grob, Mariani, & Schmidt, 1996; Mariani, 1998).

2.2.5.2. Official methods: comments and suggestions. The initial proposalbased on the quantification of stigmasta-3,5-diene (derived fromβ-sitosterol) (Lanzón & Albi, 1989) was almost immediately includedinside IOC trade standards (Table 1a). Although 3,5-stigmastadiene isthe major compound, the dehydration reaction is not selective for β-sitosterol. In fact, the yield of dehydration reaction is not the same forevery sterol, and because of this, the presence of sterenes with three dou-ble bonds depends on the presence of hydroxysterols originated by steroloxidation (Bortolomeazzi, De Zan, Pizzale, & Conte, 2000) (Table 2b).

The EU regulation (EC, 2013) and IOC trade standards (IOC, 2011) in-clude a footnote (Table 1a) pointing out that the limits (i.e. ≤0.10 fornon-lampante virgin olive oils) is for the sum of all the isomers thatcan (or cannot) be separated by CG capillary columns (Table 2b).

2.2.5.3. Alternative to official methods. The official procedure involves aprevious step of saponification followed by liquid chromatography onsilica gel column, in order to remove interfering compounds such assqualene isomers and n-alkanes. However, despite the elegant initialmethod and the high separative power of capillary columns, some over-laps occur, whichmay produce false results. In the attempt of perfectingthe official method, approaches by using off-line or on-line pre-separation by HPLC have been proposed, any of them being validatedwith acceptable success (Gallina-Toschi, Bendini, & Lercker, 1996;Verleyen et al., 2002).

2.3. Profiling approaches

Because there is not a method qualified as official, rapid and globalfor all the authenticity issues of olive oil but particular solutions, numer-ous proposals have been designed to opt to that qualification or to com-pete with the current official methods (Aparicio et al., 2013; Frankel,2010). On the contrary to the target analysis, which is constrained toquantitative analysis of a particular analyte or analytes, the profilinganalysis represents a well established approach that can be consideredthe precursor for metabolomics (Ryan & Robards, 2006). This approachis not however easy if it has to accomplish the following four basicrequirements: (i) theproposedmethodology has to improve the currentstandard in any aspect (e.g. analysis time, reproducibility, limit of detec-tion, analysis cost); (ii) the samples for training and testing the methodhave to represent thewhole universe (i.e. olive oil cultivars and produc-er countries); (iii) the new methodology validation step has to becarried out with blind samples; and (iv) an inter-comparison studyhas to be carried out as well. In addition, the proposed method shouldattempt to select those variables that allow an extendable application.

Researchers have investigated on the application of other techniquesdifferent from the chromatographic ones, which were analyzed in theprevious section, with the aim of implementing rapid methods or withthe objective of identifying and quantifying until today undetectableanalytes. Table 3 shows the proposed techniques, which are describedin depth in recent revision works (Aparicio et al., 2013; Frankel,2010). The table includes targeted and profiling approaches now

clustered inside five groups, defined by their most remarkable charac-teristics, which pushing back the frontiers of research in the field ofolive oil authenticity. They vary from classical to very sophisticatedtechniques, which are capable of discriminatingminor differences asso-ciated with authenticity issues.

2.3.1. Vibrational spectroscopyInfrared absorption and Raman scattering give information about

molecular vibrations, yielding a vibrational fingerprint of themolecules,which have been used for the identification of oils with the help ofmultivariate statistical algorithms. Successful discriminations betweenauthentic and adulterated olive oils are, however, mostly centered onthe identification of trans double bonds in samples labeled virgin oliveoils as their unsaturated fatty acids have to be in their cis form. The in-formation from the fingerprint region, which has not been decipheredwith mathematical algorithms yet, have allowed reaching relativelylow percentages of detection but almost always circumscribed to akind of adulteration, the addition of refined edible oils to virgin oliveoils (García-González, Baeten, Fernández-Piernas, & Tena, 2013;García-González, Infante-Domínguez, & Aparicio, 2013). Its main disad-vantage is that its lowest percentage of detection is higher than thevalues reached with some official methods (e.g. stigmastadienes)although the technique is much more rapid. The fingerprint, however,still hides information that could contribute to the characterization oflabeled monovarietal virgin olive oils in the future.

2.3.2. NMR spectroscopyNMR instrumentation has been applied with undoubted success in

the authentication of many water-based foods. The detection by NMRof possible anomalies in the chemical composition of adulterated oliveoils is, however, less interesting when results are compared withthose fromofficial standards; for instance, the resonance in the carbonylregion ascribed to saturated fatty acids at the sn-2 position of glycerol.

13C NMR is the preferred technique to obtain information about thepositional distribution of the saturated, oleyl, linoleyl, and linolenylchains on the glycerol moiety. 1H NMR technique is adequate for thequantification of fatty acids, despite it is unable to determine individualfatty acids (unlike gas chromatography), and suitable for the detection ofminor compounds, such as free fatty acids, diacylglycerol isomers, totalsterols, squalene, cycloarthenol, etc. 31P NMR is used for the detectionand quantification of phospholipids like phosphatidic acid, lyso-phosphatidic acid and phosphatidylinositol. The resonances of thephosphitylated hydroxyl groups of phenolic compounds are used fordetecting simple phenols, lignans, and flavonols. Finally, the applicationof site-specific natural isotope fractionation-nuclearmagnetic resonance(SNIF-NMR) is focusedmainly on investigating the intramolecular distri-bution of deuterium in fatty acids and triacylglycerols (Dais, 2013).

2033R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

Empirical results show, however, that slight differences in NMR spec-tra are less reliable when this technique tackles a real adulteration prob-lem. It is so because the information reported by fatty acids/TAGs is not,in general relevant, in sophisticated adulterations (Table 4) and the infor-mation from identified minor compounds is related with VOO qualityrather than, for example, the detection of the presence of refined non-olive oils in olive oils or deodorized virgin olive oils in virgin olive oils.

The combination of multivariate statistical algorithms and databasesof NMR spectra has been used in determining the geographical prove-nance of virgin olive oils. Some of the publishing results are promisingin the steps of training and testing but, unfortunately, less valuablewhen model is validated by external laboratories (the reproducibilitystep). Reasons are in the enormous amount of external and internalparameters affecting to virgin olive oil chemical composition in compari-sonwith the small amount of useful information reportedbyNMRspectrathat is, besides, circumscribed to a few chemical series; SEXIA™ expertsystem, for example, uses 64 different chemical compounds of 9 chemicalseries, and hundreds of decision rules, only for determining the geograph-ical provenance of Spanish virgin olive oils (García-González, Tena, &Aparicio, 2012). The use of supervised statistical procedures without astrict procedure for selecting the variables of the mathematical modelsis on the basis of good testing but poor validation results; the bibliographyis plenty of those hyper-optimistic proposals that besides are usuallybased on banks of samples that do not represent the entire universe tobe studied (e.g., cultivar, geographical origin, ripeness) or even on chem-ical compounds that have not been identified but tentatively.

2.3.3. Stable isotope analysisPyrolysis–isotope ratio mass spectrometry (Py–IRMS) is a finger-

printing technique that, coupledwith suitablemultivariate data analysisprocedures, has been proposed for the detection of frauds in olive oil(Guillou et al., 1999); its main advantage is to be rapid and no need ofsample cleanup.

Information collected by elemental analysis–pyrolysis–IRMS (2H-EA–Py–IRMS) has allowed detecting refined olive oil samples adulteratedwith refined hazelnut oils at percentages higher than 10%. Its maindrawbacks, however, are the lack of D/H (deuterium/hydrogen) certifiedreference material and the geographical origin dependence. Other stud-ies, based on δ18O isotopic abundance in olive oils, have been focused onolive variety and geographical origin. Although compounds with lowoxygen content generate more problems than faced when chromato-graphic techniques are applied, authors have found correlation betweenenrichment in heavy isotopes and latitude (Aramendía et al., 2007)though the model has not been validated yet.

δ13C isotope ratio measurements have been scarcely used for oliveoil authentication (Angerosa, Camera, Cumitini, Gleixner, & Reniero,1997) but for other foods (e.g., wine and juices) because the values ofthis ratio by themselves did not allow fraud can be ascertained althoughthey could perhaps provide valuable complementary information

Table 4Main authenticity issues and sub-issues, and their current examples.

Issue Sub-issue Paradigm

Adulteration Addition of cheaper oilto olive oils

Detection of refined hazelnutoils in ROOs

Addition of refined oilsto VOOs

Detection of seed oils in VOOs

Addition of low to highoil categories

Detection of deodorized VOOsin VOOs

Geographical origin Inexact label Detection of VOO from severalorigins

Traceability Characterisation of PDOsAgricultural system Organic vs. conventional Addition of conventional to

organic VOOsCultivar Varietal VOOs Authentication of monovarietal

VOOs

Note: PDO, protected denomination of origin; ROO, refined olive oil; VOO, virgin olive oil.

(Kelly, Parker, Sharman, Dennis, & Goodall, 1997). Other attempts,based on the ratios δ13C16:0 adversus δ13C18:1, have been proposed todetect slight variations in the composition of fatty acids as result ofthe presence of edible oils with similar profiles to olive oil; i.e. higholeic sunflower, hazelnut, olive-pomace (Ogrinc, Košir, Spangenberg,& Kidrič, 2003). The variability in δ13C values of fatty acids, however, isaffected by a combination of geographical provenance (mainly environ-mental conditions) and genetic factors (Woodbury, Evershed, & Rossell,1998) which depreciates the proposal because of the percentages offalse positives. Bibliography does not describe that these methods hadbeen validated according to ISO 5725 (ISO, 1994) and hence they cannotbe described as alternatives to official standards. Therefore, it seemsimportant to establish a database that provides isotopic informationfor olive oils as a previous step to tackling olive oil authenticity withthese kinds of procedures. The protocol should be supplemented witha calibration step referred to a primary standard method.

2.3.4. CalorimetryAlthough differences in the solid–liquid phase transitions of olive oil

and seed oils are known by researchers from many years ago, there islittle information in the literature because of the reproducibility of thethermograms. The methods based on simultaneous thermogravimetry(TG) and differential scanning calorimetry (DSC), which seem to be intheir preliminary stages of application for determining olive oil authen-ticity and quality, are currently focused on identifying deconvolutedpeaks related to olive oil chemical composition (i.e. triacylglycerols,fatty acids and phenols) (Vecchio, Cerratini, Bendini, & Chiavaro,2009). In terms of authenticity, the main problem is, once again, thatthe information collected by the technique (i.e. TAGs and FAs) is notenough as to determine if an olive oil is adulterated due to the large pan-oply of possible adulterants and, as the same time, to be more efficientand precise than the current official methods.

2.3.5. Multidimensional chromatographyAlthough chromatography offers enough resolution to yield quantita-

tive information for chemical components, the probability of separationof every component ranges, depending on sample complexity,19%–37%. The detection of current sophisticated adulterations, however,requires techniqueswith increased resolution power that is not providedby single dimension chromatography. Solution canbe in themultidimen-sional chromatography, a separation technology that utilizes twoorthogonal separation mechanisms in order to increase the resolutionpower and peak capacity of an experiment by increasing the selectivityof the experiment (Cortes,Winniford, Luong, & Pursch, 2009). For exam-ple, GC × GC has been especially helpful in identifying low level odd-numbered fatty acids and species where the results of mass spectra canbe ambiguous (Tranchida et al., 2009) while GC × GC-FID has allowedstudying phytol esters, geranylgeraniol esters and the straight-chain es-ters of palmitic acids and the unsaturated C18 acids (Biedermann,Bongartz, Mariani, & Grob, 2008). Bidimensional chromatography is notthought for its application in olive oil authenticity, because of the instru-ment cost and need of analysts with particular skills among others, butfor helping in unsolved problems; for example, GC × GC–TOFMS hasbeen used for the characterization of fresh adversus stale olive oil at qual-itative level (de Koning, Kaal, Janssen, van Platerink, & Brinkman, 2009)whereas a GC × GC–TOF-MS was checked to determine polycyclic aro-matic hydrocarbons (Purcaro, Morrison, Moret, Conte, &Marriott, 2007).

2.4. Rapid methods for olive oil authenticity

The efficiency of the official food control is dependent on the possi-bility to proceed to rapid tests for particular analytes. It is more remark-able in the current globalmarket inwhich the international regulations,which are stricter day by day, have shot up the number of samples andthe analytes to be analyzed. The immediate consequence has been anincreasing demand of rapid methods that allow a fast overall view of

Table 5Basic characteristics of the methods proposed for the current analytical challenges of the olive oil authenticity issues.

Issue Addition of cheaper oils to olive oils

Objective Detection of the presence of crude or refined hazelnut oil in virgin or refined olive oil.Analyte/signal Free and sterified sterols.Technique Gas chromatography.Level of applicability Particular due to the design of the mathematical algorithm.Official method? No, but the method has been validated with blind trials by IOC.Time of analysisa Sample pre-treatment: 120 min; measurement: 45 min; data analysis: 15 min.Limit of detectionb b8%Advantages Easy to carry out. Low cost analysis. Fine repeatability.Disadvantages Time-consuming. Poor reproducibility because of the need of extensively well-trained analysts.References García-González, Viera, Tena, and Aparicio (2007); Mariani et al, (2006), Bowadt & Aparicio (2003).

Objective Detection of the presence of any edible oil (crude or refined) in virgin or refined olive oilAnalyte/signal Selected 13C & 1H NMR bands of the spectrum.Technique 13C NMR and 1H NMR spectroscopies.Level of applicability Universal although has been checked with only a few adulterants.Official method? No, but the adulteration with hazelnut oils have been validated with blind trials.Time of analysisa Pre-treatment: No; measurements: 4 h for 1H NMR and 1.45 h for 13C NMR; data analysis: 20 min applying procedures

of Artificial Neural Networks (ANN).Limit of detectionb N10% using bands from 13C NMR and 1H NMR for adulterations with hazelnut oils.

~15% using bands from 13C NMR or from 1H NMR for dulterations with hazelnut oils.Advantages Good repeatability.Disadvantages Time-consuming. Poor reproducibility. False positives. Hyper-optimist models.References Dais and Hatzakis (2013), Mannina, D'Imperio, Capitani, Rezzi, et al. (2009), García-González, Mannina, D'Imperio, Segre, and Aparicio (2004).

Objective Detection of the presence of any edible oil (crude or refined) in virgin or refined olive oil.Analyte/signal Infrared or Raman bands.Technique FTIR or FT-Raman.Level of applicability Universal although has been checked with only a few adulterants.Official method? No, some kinds of adulteration have been validated with blind samples.Time of analysisa FTIR: Pre-treatment: 5minc; measurement: 5 min; data analysis: 5 min applying ANN.

FT-Raman: Pre-treatment: nilc; measurement: 10 min; data analysis: 5 min applying ANN.Limit of detectionb N10%Advantages Rapid and easily implementable method.Disadvantages Full checked with hazelnut oils only. A large set of spectra is required. Unstable mathematical equations.References García-González, Baeten, Fernández-Piernas, and Tena (2013), García-González, Infante-Domínguez, and Aparicio (2013),

Baeten et al. (2005), López-Díez, Bianchi, and Goodacre (2003), Tay, Singh, Krishnan, and Gore (2002).

Objective Detection of the presence of any edible oil (crude or refined) in virgin or refined olive oil.Analyte/signal Triglycerides and fatty acids.Technique HPLC for quantifying triacylglycerides and GC for quantifying fatty acids. A free program for determining differences between

empirical and theoretical TAGsLevel of applicability Universal. With almost all the edible oils.Official method? No, but it is being validated with blind trials by IOC.Time of analysisa Pre-treatment: 60 min; measurement: 90 min; data analysis: 5 min.Limit of detectionb ~10% but it depends on the kind of edible oil.Advantages Global and rapid method with individual information from FAMEs & TAGs.Disadvantages Time-consuming. False positives. Unstable mathematical model.References García-González, Viera-Macías, Aparicio-Ruiz, Morales, and Aparicio (2007), IOC (2006b)

Issue Addition of refined oils to virgin olive oils

Objective Detection of the presence of any refined edible oil in virgin olive oils.Analyte/signal Hydrocarbons: stigmastadienes.Technique Gas chromatography.Level of applicability Universal.Official method? Yes (Table 1a)Time of analysisa Pre-treatment: 90 min; measurement: 20 min; data analysis: No.Limit of detectionb 2%Advantages Global method. Excellent LOQ (0.1 ppm) & LOD (2%).Disadvantages May fail detecting refined oils obtained under soft conditions (b175 °C).References IOC (2001b), Lanzón, Cert, and Gracian (1994).

Objective Detection of the presence of any refined edible oil in virgin olive oils.Analyte/signal cis/trans FTIR or FT-Raman bands.Technique Spectroscopy by FTIR or FT-Raman.Level of applicability Universal.Official method? No, but the method has been validated with blind samples.Time of analysisa Pre-treatment: Nil; measurement: 10 min; data analysis: 10 min.Limit of detectionb N8%Advantages Rapid method.Disadvantages Limit of detection. The method does not work properly with less unsaturated oils.References Baeten et al. (2005), Hourant, Baeten, Morales, Meurens, and Aparicio (2000), Baeten, Meurens, Morales, and Aparicio (1996).

Issue Addition of low to high VOO categories

Objective Detection of the presence of VOOs deodorized at low temperature in Extra-VOOsAnalyte/signal Alkyl and ethyl esters

2034 R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

Table 5 (continued)

Issue Addition of cheaper oils to olive oils

Technique Gas-chromatographyLevel of applicability Universal, but perhaps it is circumscribed to VOOs coming from olives that have undergone any fermentative process.Official method? Yes (Table 1a)Time of analysisa Pre-treatment: 60 min; measurement: 27 min; data analysis: 5 minCorrect classificationb (%) Not reported by IOCAdvantages Detect the presence of some deodorized VOOs in VOOsDisadvantages Numerous false negatives. No detection of deodorized VOOs with rancid/wood/frozen and other non-fermentative defects.References IOC (2010), Pérez-Camino, Cert, Romero-Segura, Cert-Trujillo, and Moreda (2008).

Issue Geographical traceability of VOOs

Objective Determination of the geographical provenance (country, region, county, PDO, PGI) of VOOsAnalyte/signal Several: fatty acids, alcohols, sterols, hydrocarbons etc.Technique Gas chromatography for chemical analysis and expert system (SEXIA®) for data analysise.Level of applicability Whole Spain and partially the other EU producer countries.Official method? No, but SEXIA® has been validated with hundreds of samples for years.Time of analysisa Pre-treatment: 180 min; measurement: 300 min; data analysis: 10 min using expert systemCorrect classificationb (%) Average certainty factors (CF): 92% for Andalusian PDOs, 95% for Spanish regions, and 96% for the identification of major

EU producing countries/varieties among othersAdvantages Results are associated to high CFs. It based on the largest VOO database.Disadvantages Time-consuming. Several different chemical analyses. It constantly needs to be updated.References García-González, Baeten, Fernández-Piernas, and Tena (2013), García-González, Infante-Domínguez, and Aparicio (2013);

Aparicio and Luna (2002); Aparicio (2000), Aparicio, Alonso, and Morales (1994).

Objective Determination of the geographical provenance of VOOsAnalyte/signal δ2H, δ13C or δ18OTechnique EA-IRMS or NMRLevel of applicability UniversalOfficial method? No.Time of analysisa Pre-treatment: nil; measurement: few minutes; data analysis: 5 minCorrect classificationb(%) Not reported by authorsAdvantages Rapid methodDisadvantages Reproducibility. Need of a previous large database. Harmonization calibration procedureReferences Camin, Larcher, Perini, Bontempo, et al., 2010; Chiavaro et al., 2011; Alonso-Salces et al., 2010.

Objective Determination of the geographical provenance of VOOsAnalyte/signal Multi-elementsTechnique ICP-MS or ICP-AESLevel of applicability UniversalOfficial method? NoTime of analysisa Pre-treatment: 75–90 min with digestions in microwave; measurement: 3–5 min; data analysis: 15 min using ANN.Correct classificationb(%) Not reported by authorsAdvantages Causal relationship between soil and oil. A large number of variables (elements). Repeatability.Disadvantages Low concentration of elements in the oils. Need of information of soils for training the model. Interference of fertilizers

and fungicidesd.References Llorent-Martínez, Ortega-Barrales, Fernández-de Córdova, Domínguez-Vidal, and Ruiz-Medina (2011), Benincasa, Lewis,

Perri, Sindona, and Antonio Tagarelli (2007), Zeiner, Steffan, and Cindric (2005).

a Checked by the authors at their labs and in the course of collaborative analyses of European funded projects.b The best figure reached in the course of collaborative analyses with blind samples.c The measurement is carried with the entire oil but if the measurement is of the unsaponifiable matter, 60 min has to be added to the total analytical procedure.d Foliar fertilizers can contain K, Fe, Mg,Mn, P and Zn in different proportions, together with other elements (i.e. B, Ca), which can be presented complexedwith amino acids such in the

cases of Ca, Fe, Mg, Mn and Zn. Fungicides can contain Cu among other elements.e Other authors have proposed the study of particular geographical production zones by diverse series of compounds, and data are analyzed by an umpteen different number of statis-tical procedures, either unsupervised (e.g. PCA, MDS) or supervised (e.g. LDA, PLS).

Addition of low to high VOO categories

2035R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

the olive oil traceability and safety; a demand widely claimed byproducers, sellers and consumers. However, there is not a uniformagreement on a valid definition for the term “rapid method” due tothe lack of conceptual clarity of the adjective “rapid”. The adjective isnot conditioned by a fixed time limit even though rapid methods arecharacterized obviously by time. In fact, the concept of being a “rapidmethod” is not only defined in the time domain, but it also expressesa general sense of convenience in lab work, with a automation compli-ance in all the analytical steps: sample pre-treatment or sample prepa-ration, the chemical or physical analysis, the data analysis, and theevaluation of the results.

As described previously, regulators and associations of consumershave suggested numerous criteria for defining the authenticity of oliveoil. Adulteration, geographical origin, production system and varietyare the main authenticity issues associated with the olive oils althoughtheir relevance depends on the current trends of the global market.Table 4 shows the issues, their most relevant sub-issues and their cur-rent paradigms, which are circumscribed to olive oil adulterations that

have traditionally produced high benefits to fraudsters. Solutions, whenpossible, come from classical techniques to very sophisticated techniquesthat are capable of picking up minor differences associated with theauthenticity issues. Some of them are responsible of irrefutable results(i.e. stigmastadienes in the addition of refined to crude oils) but un-fortunately only a fewcanbedefinedas rapid (Aparicio andAlonso 1994).

Table 5 describes the most common and successful methods pro-posed for the authenticity sub-issues of Table 4; they are comparedwith standard methods when possible. All the methods have beendissected according to the time needed for the whole analysis (samplepreparation, measurement and data analysis), the limits of detectionand quantification, when described by the authors, and their mainadvantages and disadvantages.

2.5. Unsolved problems with the current methods

The large collection of analytical methods and techniques availableat present and the fact that the olive oil is the most regulated edible

2036 R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

oil in theworld have not avoided the existence of someolive oil adulter-ations that are still difficult to be detected. Reasons can be found in theample information about the chemical composition of any edible oil thatis at the disposal of everybody through internet. Thus, the fraudsters caneasily prepare fraudulent recipes mixing edible oils at percentages cal-culated by computer software. For example, an adulterationcomposed of 65% refined olive oil, 15% refined hazelnut, 15%desterolized sunflower and 5% palm olein without free sterols isundetectable although the economical benefits of this fraudulentpractice might not really yield a profit. Furthermore, the industrialtechnology allows the deodorization of virgin olive oils with sensorydefects at moderate temperature (Biedermann, Bongartz, Mariani, &Grob, 2008), and the resulting oils might not be detected even withthe proposed methodologies based on chlorophylls and alkyl esters(García-González & Aparicio, 2010).

Short-sighted planning for the elaboration of regulations, whichmostly are not based on a large number of olive oil samples of almostall the varieties and producing regions, is the main forthcoming prob-lem that has abruptly increasedwith thenumerous olive orchards locat-ed beyond the latitude of the Mediterranean basin (e.g. New Zealand,Argentina). The consequence is an ample group of olive oil varietiesand olive tree orchards that produce genuine (virgin) olive oils withchemical compositions that are, in one or more parameters, outsidethe limits of the current trade standards. Those exceptions are notonly of olive tree orchards placed in latitudes that do not correspondto those of theMediterranean basin (e.g. some virgin olive oils producedin Argentina and Australia) (Ceci & Carelli, 2007; Mailer, Ayton, &Graham, 2010) as they also include Mediterranean varietal olive oils(e.g. cv. Verdial de Huevar of Spain) (García-González, Infante-Domínguez, & Aparicio, 2013). In this context, the purpose of olive oilauthenticity should not be exclusively focused on the labeling controlbut also on protecting the genuineness of olive oils with regard totheir geographical origin and botanical variety.

The growing global market together with the increasingly stricterinternational regulations has also increased the number of samplesthat have to be analyzed by means of numerous analytical methods.Therefore, there is a growing demand for a global procedure, whichcould be able to reduce to one or two all the current official methods,or for rapid methods, since the current methods are going to collapserecognized analytical laboratories (IOC, 2012). The former alternativewould be widely supported for all the olive oil actors although itconfronts with the wide variety of edible oils and their different chem-ical compositions. The current proposals are based on mathematicalalgorithms designed to detect low concentrations of extraneous oils inolive oil, which results unstable systems that oscillate between thepercentages of false positives and false negatives. Furthermore, eachnew datum added implies rebuilding the mathematical decision rulesin an endless process that in most of cases makes things worse. It isthe so-called Jackknife Paradigm.

An alternative to global and rapid methods is based on the determi-nation of DNA in the oils (Bracci, Busconi, Fogher, & Sebastiani, 2011),which could achieve indubitable results in the authenticity of olive oilalthough its current results are at the same order of magnitude of thechromatographic techniques (Consolandi et al., 2008; Kumar, Kalon, &Chaudhary, 2011), in addition to the difficulties in extracting good-quality DNA without contamination from other sources (Woolfe &Primrose, 2004). In the meantime, the solution may come from data-bases with chemical information, obtained by chromatographic tech-niques, which could be used to perfect certain limits of the officialparameters and also to aid in the calibration of those spectroscopic tech-niques which are less time-consuming.

3. Challenges for the near future

The olive oil authenticity is continuously evolving to situationsexplained by the consumer demands of a global market. IOC trade

standards, in addition to international and national regulations, are pe-riodically upgraded in the light of new challenges created by fraudstersand the advances in analytical instrumentation.

Sometimes, however, the problem lies in the semantic definition ofauthenticity, which is not obvious. If the fatty acid composition deter-mines the physical and chemical properties of the oil and itspotential applications (food industry is offering tailored oils for specificpurposes), mutants or genetically modified seed oils, for example,represent a challenge in oil authenticity not only for detecting contam-ination or fraud, but also for the global meaning of authenticity (García-González & Aparicio, 2010). Furthermore, forthcoming problemsmightnot be only focused on the authenticity of the olive oil added to cannedfish or inside bottles labeled as spiced VOOs, or even in the detection ofdeodorized oils in VOO, but in the authentication of VOO geographicalprovenance and the current overlap between the concepts of qualityand authenticity.

The highnumber of protected designations of origins (PDO)has raisedeven more the concern of producers and consumers about the particularcharacteristics of their VOOs. The European Community regulation (EC,2006), as result of the consumer demands, established a controlled label-ing for PDOs with the objective of assuring the consumers' expectationsand a better protection of VOOs, which have proved to have particularsensory and chemical characteristics, against falsification or mislabelingbecause their higher market price. Regulations, however, do not suggestany analytical procedure to verify the information provided on the labelbut exclusively administrative controls. As a consequence, the geograph-ical declaration of virgin olive oil, which is never controlled by physical–chemical parameters, is vulnerable to frauds.

The revolution in agricultural techniques, on the other hand, hasallowed to cultivate olive trees from varieties autochthonous fromMediterranean countries, there where never were thought they couldbe cultivated. When planted inside the Mediterranean Basin, the chem-ical composition of the varieties is within the limits of trade standards.The latitude, altitude and climate of the new producing areas haveplayed a dirty trick on farmers because those physical parameters affectbiochemical pathways of some cultivars more than would be expected,and the chemical composition of resulting oils does not fit with currenttrade standards. The solution can come from the development of deci-sion trees that, using other chemical compounds, can act as safeguardfor the genuineness of those olive oils without modifying the currentlimits although analysts should keep in mind that decision trees needof strict statistical algorithms that avoid an unacceptable number offalse positives.

Therefore, the futurework on geographical traceability should be fo-cused on building an Olive Oil Mapwhere themost productive cultivarsand all the approved PDOs are characterized by chromatographic(Aparicio et al., 1994; García-González et al., 2012) spectroscopic (Dais& Hatzakis, 2013; García-González, Baeten, Fernández-Piernas, & Tena,2013) and isotopic (Chiavaro et al., 2011) methods. The resulting data-bases, in conjunction with new procedures of classification and visuali-zation techniques, would allow a better authenticity of the geographicalprovenance of virgin olive oils.

The increasing importance of mathematical algorithms in thesuccessful control of the olive oil adulteration have encouraged someresearchers to believe that a strongmathematical relationship betweentwo factors can prevail over the lack of any scientific evidence thatexplains that relationship is causal. Inappropriate application of statisti-cal procedures, most of the time without the required validation tests,allows reaching over-optimist conclusions that usually present ump-teen exceptions in the model validation step; for example, modelsbased on the statistical procedures LDA (linear discriminant analysis)without a selection of the variables by means of a stepwise algorithm.Furthermore, that hypothetical success fighting against adulterationshas driven analysts to think that the same chemical compounds alsocan be used in determining virgin olive oil quality, in a mixture ofconcepts that is difficult to be understood.

2037R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

Thus, new trade standards should be analyzed under a magnifyingglass to prevent that a casual relationship between chemical com-pounds and authenticity/quality can be seen as causal. One of paradigmsis the use of chemical compounds, which are consequence of the degra-dation of the initial virgin olive oil color (i.e. pyropheophytins — PPP),for explaining VOO freshness. The casual relationship between the in-crease of PPP with shelf-life and the causal relationship between shelf-life and sensory quality has been used to relate PPP with theoverall grading of virgin olive oil quality (color is not among the qualityparameters) when there is not any scientific evidence that relates theconcentration of PPP with virgin olive oil flavor (Aparicio, Morales, &García-González, 2012). Conclusionswithout scientific basis – for exam-ple, NIR spectra explain intensities of sensory descriptors and concen-trations of volatile compounds of virgin olive oils – are unfortunately afact very common today.

References

Al-Ismail, K. M., Alsaed, A. K., & Ahmad, R. (2010). Detection of olive oil adulteration withsome plant oils by GLC analysis of sterols using polar column. Food Chemistry, 121,1255–1259.

Alonso-Salces, R. M., Moreno-Rojas, J. M., Holland, M. V., Reniero, F., Guillou, C., &Heberger, K. (2010). Virgin olive oil authentication by multivariate analyses of1H-NMRfingerprints and δ13C and δ2H data. Journal of Agricultural and Food Chemistry,58, 5586–5596.

Amelotti, G. (1985). Problematiche inerenti la valutazione della frazione sterolica. RivistaItaliana delle Sostanze Grasse, 62, 337–344.

Angerosa, F., Camera, L., Cumitini, S., Gleixner, G., & Reniero, F. (1997). Carbon stableisotopes and olive oil adulteration with pomace oil. Journal of Agricultural and FoodChemistry, 45, 3044–3048.

Aparicio, R. (2000). Characterization: mathematical procedures for chemical analysis.In R. Aparicio, & J. Harwood (Eds.), Handbook of Olive Oil: Analysis and properties(1st ed.). Gaithersburg: Aspen.

Aparicio, R., & Alonso, V. (1994). Characterization of virgin olive oils by SEXIA expert sys-tem. Progress in Lipid Research, 33, 29–38.

Aparicio, R., Alonso, V., & Morales, M. T. (1994). Detailed and exhaustive study of theauthentication of European virgin olive oils by SEXIA expert system. Grasas y Aceites,45, 241–252.

Aparicio, R., & Aparicio-Ruiz, R. (2000). Authentication of vegetable oils by chromato-graphic techniques. Journal of Chromatography. A, 881, 93–104.

Aparicio, R., Aparicio-Ruiz, R., & García-González, D. L. (2007). Rapid methods for testingauthenticity: The case of olive oil. In A. van Amerongen, D.M. Barug, & M. Lauwars(Eds.), Rapid methods for food and feed quality determination (pp. 163–188).Wageningen (Netherlands): Wageningen Academic Publishers.

Aparicio, R., Conte, L. S., & Fiebig, H. J. (2013). Olive oil authentication. In R. Aparicio, & J.Harwood (Eds.), Handbook of olive oil: Analysis and properties (2nd ed.). New York:Springer Science.

Aparicio, R., & Luna, G. (2002). Characterisation of monovarietal virgin olive oils. EuropeanJournal of Lipid Science and Technology, 104, 614–627.

Aparicio, R., McIntyre, P., Aursand, M., Eveleigh, L., Marighetto, N., et al. (1998). Oils andfats. In M. Lees (Ed.), Food authenticity issues and methodologies (pp. 211–269).Nantes (France): Eurofins Scientific.

Aparicio, R., Morales, M. T., & García-González, D. L. (2012). Towards new analyses ofaroma and volatiles to understand sensory perception of olive oil. European Journalof Lipid Science and Technology, 114, 1114–1125.

Aragón, A., Toledano, R. M., Cortes, J. M., Villén, J., & Vázquez, A. (2011). Wax estercomposition of monovarietal olive oils from Designation of Origin (DO) “Campos deHellin”. Food Chemistry, 129, 71–76.

Aramendía, M.A., Marinas, A., Moreno, J. M., Moalem, M., Rallo, L., & Urbano, F. J. (2007).18O measurement of Andalusian olive oils by continuous flow pyrolysis/isotoperatio mass spectrometry. Rapid Communications in Mass Spectrometry, 21, 487–496.

Baeten, V., Fernández-Piernas, J. A., Dardenne, P., Meurens, M., García-González, D. L., &Aparicio-Ruiz, R. (2005). Detection of the presence of hazelnut oil in olive oil byFT-Raman and FT-MIR spectroscopy. Journal of Agricultural and Food Chemistry, 53,6201–6206.

Baeten, V., Meurens, M., Morales, M. T., & Aparicio, R. (1996). Detection of virgin olive oiladulteration by Fourier transform raman spectroscopy. Journal of Agricultural andFood Chemistry, 44, 2225–2230.

Ben-Ayed, R., Kamoun-Grati, N., & Rebai, A. (2013). An overview of the authentication ofolive tree and oil. Comprehensive Review on Food Science and Food Safety, 12, 218–227.

Benincasa, C., Lewis, J., Perri, E., Sindona, G., & Antonio Tagarelli, A. (2007). Determi-nation of trace element in Italian virgin olive oils and their characterizationaccording to geographical origin by statistical analysis. Analytica Chimica Acta,585, 366–370.

Biedermann, M., Bongartz, A., Mariani, C., & Grob, K. (2008). Fatty acid methyl and ethylesters as well as wax esters for evaluating the quality of olive oils. European FoodResearch and Technology, 228, 65–74.

Biedermann, M., Grob, K., & Mariani, C. (1995). On-line LC-UV-GC-FID for the determina-tion of Δ7- and Δ8(14)-sterols and its application for the detection of adulterated oliveoils. Rivista Italiana delle Sostanze Grasse, 72, 339–344.

Biedermann, M., Grob, K., Mariani, C., & Schmidt, J. P. (1996). Detection of desterolizedsunflower oil in olive oil through isomerized Δ7-sterols. Zeitschrift für Lebensmittel-Untersuchung und -Forschung A, 202, 199–204.

Biedermann, M., Haase-Aschoff, P., & Grob, K. (2008). Wax ester fraction of edible oils:Analysis by on-line LC–GC–MS and GC × GC-FID. European Journal of Lipid Scienceand Technology, 110, 1084–1094.

Bortolomeazzi, R., De Zan, M., Pizzale, L., & Conte, L. S. (2000). Identification of newsteroidal hydrocarbons in refined oils and the role of hydroxy sterols as possibleprecursors. Journal of Agricultural and Food Chemistry, 48, 1101–1105.

Bowadt, S., & Aparicio, R. (2003). The detection of the adulteration of olive oil with hazel-nut oil: A challenge for the chemists. Inform, 14, 342–344.

Bracci, T., Busconi, M., Fogher, C., & Sebastiani, L. (2011). Molecular studies in olive (Oleaeuropaea L.): Overview on DNAmarkers applications and recent advances in genomeanalysis. Plant Cell Reports, 30, 449–462.

Buchgraber, M., Ulberth, F., Emons, H., & Anklam, E. (2004). Triacylglycerol profiling byusing chromatographic techniques. European Journal of Lipid Science and Technology,106, 621–648.

Calvano, C. D., Palmisano, F., & Zambonin, C. G. (2005). Laser desorption/ionizationtime-of-flight mass spectrometry of triacylglycerols in oils. Rapid Communications inMass Spectrometry, 19, 1315.

Camin, F., Larcher, R., Perini, M., Bontempo, L., et al. (2010). Characterization of authenticItalian extra-virgin olive oils by stable isotope ratios of C, O and H and mineralcomposition. Food Chemistry, 118, 901–909.

Ceci, L. N., & Carelli, A. A. (2007). Characterization of monovarietal Argentinian olive oilsfrom new productive zones. Journal of the American Oil Chemists' Society, 84,1125–1136.

Cercaci, L., Passalacqua, G., Poerio, A., Rodriguez-Estrada, M. T., & Lercker, G. (2007).Composition of total sterols (4-desmethyl-sterols) in extra virgin olive oils obtainedwith different extraction technologies and their influence on the oil oxidativestability. Food Chemistry, 102, 66–76.

Cert, A., Di Giacinto, L., Faberi, A., Mariani, C., Lacoste, F., et al. (2011). New methodsproposed to IOC. Internal communication.

Chiavaro, E. C., Cerratini, L., Di Mateo, A., Barnaba, C., Bendini, A., & Iacumin, P. (2011).Application of a multidisciplinary approach for the evaluation of traceability of extravirgin olive oil. European Journal of Lipid Science and Technology, 113, 1509–1519.

Christie, W. W. (1992). Detectors for high-performance liquid chromatography of lipidswith special reference to evaporative light-scattering detection. In W. W. Christie(Ed.), Advances in Lipid Methodology — One (pp. 239–271). Ayr (UK): Oily Press.

Christie, W. W. (2011). What is a lipid? In W. W. Christie (Ed.), The AOCS Lipid Library. :AOCS Press (Available from: http://lipidlibrary.aocs.org/lipids/whatlip/index.htm).

Christie, W. W. (2012). Waxes: Structure, composition, occurrence and analysis. InW. W.Christie (Ed.), The AOCS Lipid Library. : AOCS Press (Available from: http://lipidlibrary.aocs.org/lipids/waxes/index.htm).

Consolandi, C., Palmieri, L., Severgnini, M., Maestri, E., Marmiroli, N., Agrimonti, C., et al.(2008). A procedure for olive oil traceability and authenticity: DNA extraction,multiplex PCR and LDR — Universal array analysis. European Food Research and Tech-nology, 227, 1429–1438.

Cortes, H. J., Winniford, B., Luong, J., & Pursch, M. (2009). Comprehensive two dimensionalgas chromatography review. Journal of Separation Science, 32, 883–904.

Cortesi, N., Rovellini, P., & Fedeli, E. (1990). I trigliceridi degli oli naturali. Nota I. RivistaItaliana delle Sostanze Grasse, 67, 69–73.

Cozzolino, R., & De Giulio, B. (2011). Application of ESI andMALDI-TOF-MS for triacylglyc-erols analysis in edible oils. European Journal of Lipid Science and Technology, 113,160–167.

Dais, P. (2013). Nuclear magnetic resonance: Methodologies and applications. In R.Aparicio, & J. Harwood (Eds.), Handbook of olive oil: Analysis and properties (2nded.). New York: Springer Science.

Dais, P., & Hatzakis, E. (2013). Quality assessment and authentication of virgin olive oil byNMR spectroscopy: A critical review. Analytical Chimica Acta, 765, 1–27.

de Koning, S., Kaal, E., Janssen, H. G., van Platerink, C., & Brinkman, U. A. T. (2008).Characterization of olive oil volatiles by multi-step direct thermal desorption-comprehensive gas chromatography–time-of-flight mass spectrometry using aprogrammed temperature vaporizing injector. Journal of Chromatography. A, 1186,228–235.

Dionisi, F., Prodolliet, J., & Tagliaferri, E. (1995). Assessment of olive oil adulterationby reversed phase high performance liquid chromatography/amperometric de-tection of tocopherols and tocotrienols. Journal of the American Oil Chemists' Society,72, 1505–1511.

El-Hamdy, A. H., & El-Fizga, N. K. (1995). Detection of olive oil adulteration bymeasuring itsauthenticity factor using reversed-phase high-performance liquid chromatography.Journal of Chromatography. A, 708, 351–355.

European Communities (EC) (2006). Official Journal of the European Union (Regulation nº510/2006 L93, 12–25, 20 March 2006).

European Communities (EC) (2007). Official Journal of the European Union (Regulation nº702/2007 L161, 11–27, 22 June 2007).

European Communities (EC) (2013). Official Journal of the European Union (Regulation nº299/2013 L90, 52–70, 28 March 2013).

Fiebig, H. J. (1985). HPLC-Trennung von Triglyceriden. Fette, Seifen, Anstrichm, 87, 53–57.Frankel, E. N. (2010). Chemistry of extra virgin olive oil: Adulteration, oxidative stability,

and antioxidants. Journal of Agricultural and Food Chemistry, 58, 5991–6006.Gallina-Toschi, T., Bendini, A., & Lercker, G. (1996). Evaluation of 3,5-stigmastadiene con-

tent of edible oils: Comparison between the traditional capillary gas chromatographicmethod and the on-line high performance liquid chromatography-capillary gaschromatographic analysis. Chromatographia, 43, 195–199.

Gao, L., Sedman, J., García-González, D. L., Ehsan, S., Sprules, T., et al. (2009). 13C NMR as aprimary method for determining saturates, cis and trans-monounsaturates and

2038 R. Aparicio et al. / Food Research International 54 (2013) 2025–2038

polyunsaturates in fats and oils for nutritional labeling purposes. European Journal ofLipid Science and Technology, 11, 612–622.

García-González, D. L., & Aparicio, R. (2010). Research in olive oil: Challenges for the nearfuture. Journal of Agricultural and Food Chemistry, 58, 12569–12577.

García-González, D. L., Aparicio-Ruiz, R., & Aparicio, R. (2009). Olive oil. In A. Kamal-Eldin,& R. Moreau (Eds.), Gourmet and health-promoting specialty oils (pp. 33–72). UrbanaIL: AOCS Press.

García-González, D. L., Baeten, V., Fernández-Piernas, J. A., & Tena, N. (2013). Infrared,Raman and fluorescence spectroscopies: Methodologies and applications. In R.Aparicio, & J. Harwood (Eds.), Handbook of olive oil: Analysis and properties (2nded.). New York: Springer Science.

García-González, D. L., Infante-Domínguez, C., & Aparicio, R. (2013). Tables of olive oilchemical data. In R. Aparicio, & J. Harwood (Eds.), Handbook of olive oil: Analysisand properties (2nd ed.). New York: Springer Science.

García-González, D. L., Mannina, L., D’Imperio, M., Segre, A. L., & Aparicio, R. (2004). Using1H- and 13C-NMR techniques and artificial neural networks to detect the adulterationof olive oil with hazelnut oil. European Food Research and Technology, 219, 545–548.

García-González, D. L., Tena, N., & Aparicio, R. (2012). Describing chemical singularity ofSpanish protected designations of origin of virgin olive oil. Grasas y Aceites, 63, 26–34.

García-González, D. L., Viera, M., Tena, N., & Aparicio, R. (2007). Evaluation of the methodsbased on triglycerides and sterols for the detection of hazelnut oil in olive oil. Grasas yAceites, 58, 344–350.

García-González, D. L., Viera-Macías, M., Aparicio-Ruiz, R., Morales, M. T., & Aparicio, R.(2007). Validation of a method based on triglycerides for the detection of lowpercentages of hazelnut oil in olive oil by column liquid chromatography. Journal ofAOAC International, 90, 1346–1353.

García-Pulido, J., & Aparicio, R. (1993). Triacylglycerol determination based on fatty acidcomposition using chemometrics. Analytical Chimica Acta, 271, 293–298.

Gegiou, D., & Georgouli, M. (1983). A rapid argentation TLC method for detection ofreesterified oils in olive oil and olive-pomace oil. Journal of the American Oil Chemists'Society, 60, 833–835.

Grob, K., Giuffré, M.A., Biedermann, M., & Bronz, M. (1994). The detection of adulterationwith desterolized oils. Fat Science Technology, 9, 341–345.

Grob, K., Lanfranchi, M. L., & Mariani, C. (1990). Evaluation of olive oils through the fattyalcohols, the sterols and their esters by coupled LC–GC. Journal of the American OilChemists' Society, 67, 626–634.

Guillou, C., Lipp, M., Radovic, B., Reniero, F., Schmidt, M., & Anklam, E. (1999). Use ofpyrolysis–mass spectrometry in food analysis: Applications in the food analysis labo-ratory of the European Commissions' Joint Research Centre. Journal of Analytical andApplied Pyrolysis, 49, 329–335.

Holčapek, M., Jandera, P., Zderadička, P., & Hrubá, L. (2003). Characterization of triacyl-glycerol and diacylglycerol composition of plant oils using high-performance liquidchromatography–atmospheric pressure chemical ionization mass spectrometry.Journal of Chromatography. A, 1010, 195–215.

Hourant, P., Baeten, V., Morales, M. T., Meurens, M., & Aparicio, R. (2000). Oil and fatclassification by selected bands of near-infrared spectroscopy. Applied Spectroscopy,54, 1168–1174.

International Olive Council (IOC) (2001a). COI/T.20/Doc. No 10/Rev. 1. Determination of thecomposition and content of sterols by capillary-column gas chromatography. Madrid,Spain.

International Olive Council (IOC) (2001b). COI/T.20/Doc. no. 11/Rev. 2. Determination ofstigmastadienes in vegetable oils. Madrid, Spain.

International Olive Council (IOC) (2006a). COI/T.20/doc. No. 23. Method for the determina-tion of the percentage of 2-glyceryl monopalmitate. Madrid, Spain.

International Olive Council (IOC) (2006b). COI/T.20/Doc. No. 25 Global method for thedetection of extraneous oils in olive oils. Madrid, Spain.

International Olive Council (IOC) (2007). COI/T.20/doc. No. 18/Rev.2. Determination of waxcontent by capillary column gas chromatography. Madrid, Spain.

International Olive Council (IOC) (2010). COI/T.20/Doc. No 28/Rev. 1. Determination of thecontent of waxes, fatty acid methyl esters and fatty acid ethyl esters by capillary gaschromatography. Madrid, Spain.

International Olive Council (IOC) (2011). COI/T.15/NC No 3/Rev. 6. Trade standards applyingto olive oils and olive-pomace oils. Madrid, Spain.

International Olive Council (IOC) (2012). T.21/Doc. nº 13/Rev. 15. Liste des laboratoiresd'analyses chimiques agréés par le Conseil Oléicole International pour la période du1.12.2012 au 30.11.2013. Madrid, Spain.

International Organization of Standardization (ISO) (1994). ISO 5725:1994, Accuracy(trueness and precision) of measurement methods and results. Geneva, Switzerland.

Ismail, A. A., Cocciardi, R. A., Alvarez, P., & Sedman, J. (2006). Infrared and Raman spec-troscopy in food science. In Y. H. Hui (Ed.), Handbook of food science, technology andengineering (pp. 44.1–44.19). Boca Raton (FL): CRC Press.

Ismail, A. A., Nicodemo, A., Sedman, J., van de Voort, F., & Holzbaur, I. E. (1999). Infraredspectroscopy of lipids: Principles and applications. In R. J. Hamilton, & J. Cast (Eds.),Spectral properties of lipids (pp. 235–269). Boca Raton (FL): CRC Press.

Jakab, A., Héberger, K., & Fornács, E. (2002). Comparative analysis of different plant oils byhigh-performance liquid chromatography–atmospheric pressure chemical ionizationmass spectrometry. Journal of Chromatography. A, 976, 255–263.

Kelly, S., Parker, I., Sharman, M., Dennis, J., & Goodall, I. (1997). Assessing the authenticityof single seed vegetable oils using fatty acid stable carbon isotope ratios (13C/12C).Food Chemistry, 59, 181–186.

Kumar, S., Kalon, T., & Chaudhary, S. (2011). A rapid screening for adulterants in olive oilusing DNA barcodes. Food Chemistry, 127, 1335–1341.

Lanzón, A., & Albi, T. (1989). Detección de la presencia de aceite de oliva refinado en elaceite de oliva virgen. Grasas y Aceites, 40, 385–388.

Lanzón, A., Albi, T., Cert, A., & Gracián, J. (1994). The hydrocarbon fraction of virgin oliveoil and changes resulting from refining. Journal of the American Oil Chemists' Society,71, 285–291.

Lees, M. (1998). Food authenticity: Issues and methodologies. Nantes (France): EurofinsScientific.

Lercker, G., Moschetta, M., Caboni, M. F., & Frega, N. (1985). Determinazione degli oliesterificati negli oli provenienti dalla lavorazione delle olive. Rivista Italiana delleSostanze Grasse, 62, 15–18.

Llorent-Martínez, E. J., Ortega-Barrales, P., Fernández-de Córdova, M. L., Domínguez-Vidal,A., & Ruiz-Medina, A. (2011). Investigation by ICP-MS of trace elements in vegetableedible oils produced in Spain. Food Chemistry, 1127, 257–1262.

López-Díez, E. C., Bianchi, G., & Goodacre, R. (2003). Rapid quantitative assessment of theadulteration of virgin olive oils with hazelnut oils using Raman spectroscopy andchemometrics. Journal of Agricultural and Food Chemistry, 51, 6145–6150.

Mailer, R. J., Ayton, J., & Graham, K. (2010). The influence of growing region, cultivar andharvest timing on the diversity of Australian olive oil. Journal of the American OilChemists' Society, 87, 877–884.

Mannina, L., D'Imperio, M., Capitani, D., Rezzi, S., et al. (2009). 1H NMR-based protocol forthe detection of adulterations of refined olive oil with refined hazelnut oil. Journal ofAgricultural and Food Chemistry, 57, 11550–11556.

Mariani, C. (1998). Sulla presenza di ergosterolo negli oli di oliva. Rivista Italiana delleSostanze Grasse, 75, 3–10.

Mariani, C., Bellan, G., Lestini, E., & Aparicio, R. (2006). The detection of the presence ofhazelnut oil in olive oil by free and esterified sterols. European Food Research andTechnology, 223, 655–661.

Mariani, C., Venturini, S., & Grob, K. (1995). Individuazione dell'olio di girasole alto oleicodesterolato nell'olio d'oliva. Rivista Italiana delle Sostanze Grasse, 72, 473–482.

Moreda, W., Pérez-Camino, M. C., & Cert, A. (2003). Improved method for the determina-tion of triacylglycerols in olive oils by high performance liquid chromatography.Grasas y Aceites, 54, 175–179.

Norme Grassi e Derivati (NGD) (2002). NGD C 86-03. Determinazione della percentuale diacido palmitico in posizione 2 dei trigliceridi mediante dosaggio del 2 glicerilmonopalmitato. Stazione Sperimentale per le Industrie degli Oli e dei Grassi (SSOG),Milano, Italy.

Ogrinc, N., Košir, I. J., Spangenberg, J. E., & Kidrič, J. (2003). The application of NMR andMSmethods for detection of adulteration of wine, fruit juices, and olive oil. A review. An-alytical and Bioanalytical Chemistry, 376, 424–430.

Pérez-Camino, M. C., Cert, A., Romero-Segura, A., Cert-Trujillo, R., & Moreda, W. (2008).Alkyl esters of fatty acids a useful tool to detect soft deodorized olive oils. Journal ofAgricultural and Food Chemistry, 56, 6740–6744.

Piironen, V., Lindsay, D.G., Miettinen, T. A., Toivo, J., & Lampi, A.M. (2000). Plant sterols:Biosynthesis, biological function and their importance to human nutrition. Journalof Food Science and Agriculture, 80, 939–966.

Purcaro, G., Morrison, P., Moret, S., Conte, L. S., &Marriott, P. J. (2007). Determination of poly-cyclic aromatic hydrocarbons in vegetable oils using solid-phase microcomprenhensivetwo-dimensional gas chromatography coupled with time-of-flight mass spectrometry.Journal of Chromatography. A, 1161, 284–291.

Rezzi, S., Axelson, D. E., Héberger, K., Reniero, F., Mariani, C., & Guillou, C. (2005). Classifi-cation of olive oils using high throughput flow 1H-NMR fingerprinting with principalcomponent analysis, linear discriminant analysis and probabilistic neural networks.Analytical Chimica Acta, 552, 13–24.

Ryan, D., & Robards, K. (2006). Metabolomics: The greatest omics of them all? AnalyticalChemistry, 78, 7954–7958.

Salivaras, M., & McCurdy, A.R. (1992). Detection of olive oil adulteration with canola oilfrom triacylglycerol analysis by reversed-phase high-performance liquid chromatog-raphy. Journal of the American Oil Chemists' Society, 69, 935–939.

Sedman, J., Gao, L., García-González, D. L., Ehsan, S., & van de Voort, F. R. (2010). Deter-mining nutritional labeling data for fats and oils by 1H NMR. European Journal ofLipid Science and Technology, 112, 439–451.

Tay, A., Singh, R. K., Krishnan, S. S., & Gore, J. P. (2002). Authentication of olive oiladulterated with vegetable oils using Fourier transform infrared spectroscopy.Lebensmittel-Wissenschaft und Technologie, 35, 99–103.

Tranchida, P. Q., Purcaro, G., Conte, L. S., Dugo, O., Dugo, G., & Mondello, L. (2009).Enhanced resolution comprehensive two-dimensional gas chromatography appliedto the analysis of roasted coffee volatiles. Journal of Chromatography. A, 1216,7301–7306.

Vecchio, S., Cerratini, L., Bendini, A., & Chiavaro, E. (2009). Thermal decomposition studyof monovarietal extra virgin olive oil by simultaneous thermogravimetry/differentialscanning calorimetry: relation with chemical composition. Journal of Agricultural andFood Chemistry, 57, 4793–4800.

Verleyen, T., Szulczewska, A., Verhe, R., Dewettinck, K., Huyghebaer, A., et al. (2002). Com-parison of steradiene analysis between GC and HPLC. Food Chemistry, 78, 267–272.

Woodbury, S. E., Evershed, R. P., & Rossell, J. B. (1998). Purity assessments of major vege-table oils based on δ13C values of individual fatty acids. Journal of the American OilChemists' Society, 75, 371–379.

Woolfe, M., & Primrose, S. (2004). Food forensics: Using DNA technology to combatmisdescription and fraud. Trends in Biotechnology, 22, 222–226.

Zeiner, M., Steffan, I., & Cindric, I. J. (2005). Determination of trace elements in olive oil byICP-AES and ETA-AAS: A pilot study on the geographical characterization.Microchemical Journal, 81, 171–176.