OLEUM analytical solutions for olive oil authenticity
Dr. Alessandra Bendini, in the framework of the Food Fraud Session – Part I Authenticity and Integrity, organized during the EFFOST Congress 2018, held a speech on the advancements of the OLEUM project addressed to possible analytical solutions for olive oil authenticity.
Taking into consideration the need for fighting the fraudulent commercial practices that distress the olive oil sector reducing the confidence of consumers and the positive image of olive products, the OLEUM project plays an important role. Specifically, OLEUM aims to develop and validate reliable analytical solutions to be proposed as:
i) revisions of existing EU official methods suggesting replacement of specific steps of analytical protocols to reduce time analysis and solvent consumption
ii) new methods for detecting specific markers for specific frauds by target approaches
iii) new screening methods by non-target approaches.
In this framework, some examples of new analytical strategies to detect the three most common cases of olive oil fraud have been presented.
In particular, the first case concerns the quality non consistent with the declared grade, for example when a virgin or, in the worst-case scenario a lampante olive oil, is marketed as extra virgin olive oil (EVOO) that is the top quality grade: a target method to support the IOC Panel test (official sensory method) based on identification and quantitation of single volatile molecules as markers of sensory defects has been developed and its validation is in progress (partners involved: CSIC, ITERG, NESTEC, UB, UNIBO, UNIPG, UNIUD). This SPME-GC-FID/MS approach has to be able to discriminate samples according to their quality grades and has to be simple enough to be applied by the official control labs. A similar method does not officially exist yet. Moreover, taking into account the need to significantly decrease the daily work of sensory panels of olive oil companies, four different non-target rapid screening methods are in course of evaluation (partners involved: EUROFINS, UB, UNIBO). These methods are based on a fingerprinting approach that means a chemometric elaboration of the entire profiles in volatile molecules without steps of identification and quantification. The scope is to build robust classification models able to discriminate samples according to their quality grades. Now, classification models have been performed with 180 oils collected during the first year sampling (partners involved: EUROFINS, ITERG, IPTPO, UNIBO, UZZK, ZRS). The preliminary results show percentages higher than 85% of correct classification of EVOOs and non-EVOOs. When considering non-EVOOs similar percentages were found when distinguishing between virgin and lampante olive oils. These classification models will get stronger thanks to the addition of the other 154 samples collected during the second year sampling (partners involved: EUROFINS, ITERG, IPTPO, UNIBO, UZZK, ZRS).
According to the EU Regulation, the virgin olive oils have to be obtained from the olives solely by mechanical or physical treatments under conditions that do not lead to alteration in the oil. For this reason, the soft deodorization treatment is an illegal practice when realized with the aim to decrease the volatile compounds responsible for negative attributes and, in this way, to remove sensory defects of low quality olive oils. This soft deodorized oil is then mixed with a percentage of EVOO and the blend is fraudulently marketed as EVOO. However, the soft deodorization, that is realized at low temperature, does not remove Fatty Acid Ethyl Esters (FAEEs) that remain in the oil. FAEEs are formed in stored olives by the reaction between free fatty acids produced by enzymatic hydrolysis of triacylglycerols (TGs) and ethanol produced by the sugar microbial fermentation. For this reason, FAEEs are markers of low quality of oils and can be considered, also, indirect markers of soft deodorization treatment. Based on this, an official method to establish the concentration of FAEEs was approved in 2011 and the legal limit for EVOO was later modified. However, the official method is particularly laborious, time consuming and requires large volumes of solvents. In view of these drawbacks, in the OLEUM projects two alternative procedures to speed and mechanize the collection of the FAEEs have been realized and in-house validated with the aim to propose improvements to the official protocol (by CSIC and UNIBO). In the OLEUM project, a soft deodorization treatment by short path distillation has been applied to different kinds of defective virgin olive oils (by ITERG). Following this, different blends between each soft deodorized oil and EVOOs have been prepared in different ratios (by FERA) and analyzed. Based on this kind of sample set, OLEUM partners are working to set up new analytical screening strategies for verifying the conformity of EVOOs with FAEEs legal limit (partners involved: CSIC, ITERG, UNIBO) but, also, to detect possible new markers of this illegal treatment; specifically, new legal parameters based on free fatty acids and diacylglycerols (DGs) have been preliminarily proposed and the effectiveness of them is in course of evaluation (by CSIC).
The second case of common fraud concerns the illegal and legal blending between extra virgin olive oil or olive oil and other cheaper vegetable oils. According to the EU Reg. 29/2012, it is possible to produce and commercialize in EU countries (with some national exceptions as for Spain and Italy) legal blends between olive oil (OO) and vegetable oils. In this case, the presence of olive oil can be evidenced on the label only if it accounts for more than 50% of the blend. However, an official method to check the correctness of this declaration does not exist yet. To fight this possible fraud, in the OLEUM project researchers are setting up a target method to be applied by official control labs (by CSIC). This new method is based on the elaboration of a decision tree that fixes minimum contents of specific typical major (TGs) and minor compounds (linear aliphatic hydrocarbons, tocopherols, free sterols) of OO and EVOO.
To detect extraneous oils in extra virgin olive oil and olive oil the official method calculates a specific parameter named ECN 42. This legal limit is based on a mathematical calculation and the analytical protocol is laborious (long time of analysis, use of large quantity of solvents) and presents some limitations (low resolution of specific TGs). An alternative analytical approach, specifically a rapid screening method, is under study in the OLEUM project (by UB). This analytical approach presents many advantages compared to the official one. In fact, the instrument, based on HESI-UHRMS, works with high resolution power and accuracy and the analysis by direct infusion is very fast (only 2 minutes are requested for each analysis) as well as the sample preparation is simple. Namely, different kinds of possible adulterants of EVOO or OO have been considered and divided in vegetable oils characterized by i) high content in oleic acid as hazelnut, avocado and high oleic sunflower oil both virgin and refined typologies ii) high linoleic adulterant as refined conventional sunflower oil and iii) high palmitic adulterant as refined palm oil. For each one of this adulterant, three different percentages (2, 5 and 10%) were prepared in EVOO and OO. For this approach the chemometric data treatment, in which the TGs profile is used as a data matrix, is essential. Results obtained for both the high palmitic and the high linoleic models, in which the adulterants were the refined palm oil and the refined sunflower oil, respectively mixed at 2, 5 and 10 % in EVOO and OO, were satisfactory; in fact, all the adulterated samples, as well as those not adulterated, were correctly classified. Considering the most difficult case related to the high oleic model, in which the adulterants were the refined and virgin hazelnut oils, the refined and virgin avocado oils and the refined and desterolized high oleic sunflower oils, the prediction model resulted satisfactory only when the adulterant was mixed in EVOO or OO at 5 and 10 % excluding the 2% level.
The third case of common fraud concerns the not compliance with the designation of geographical origin. In general, the behaviour of consumers when purchasing food is oriented towards a greater preference for those products whose geographical origin is declared since this information clearly increases their confidence. According to the EU Reg. 29/2012, the EVOO and VOO must bear the geographical designation of origin on the label, following the specific cases related to the EU or non-EU origin and originating from more than one Member State or third country. Despite this, an official method (shared and validated) is not available to verify the compliance with the geographical origin that, up to now, can be verified only by documented traceability. The need to have availability of a reliable and possibly fast method able to verify the geographical origin of virgin olive oil is an urgent requirement. In the OLEUM project several analytical approaches based on spectroscopic and chromatographic analysis are under study (partners involved: CONICET-EHU, CNRIFFI, EAF, EUROFINS, FERA, ITERG, SMART ASSAYS, UB, UNIBO, UNIPG, UNIUD). One of the most promising method evaluates the semi-volatile sesquiterpene hydrocarbon fraction as markers of geographical origin and determines them by a SPME-GC-MS instrument with both a target and non-target approach (by UB). The non-target approach, that elaborates as a fingerprint the entire semi-volatile profile, allows much better discrimination than the target one. Moreover, the method can be proposed for routine analysis because this kind of instrument is generally available in control labs. Another innovative analytical approach for the analysis of the head space volatile compounds of virgin olive oils has been also developed and it is based on the Flash Gas Chromatography analysis (by UNIBO). The main advantage of this technique is the short analysis time, in fact, only 3 minutes are needed to perform the profile of volatile compounds. A chemometric elaboration applied with a non-target approach, to the chromatographic traces, has been realized. Up to now, the number of samples analysed with this method were 174 characterized by different geographical origin: from single EU countries, single non-EU countries and blends of EU and non-EU countries. The prediction results obtained from the classification model were satisfactory, in fact the full cross validation and the external validation allowed a correct classification of 98.5 and 92.5 % for non-EU samples, respectively, and 5 out of 6 of Extra EU blends were correctly classified. Considering the wider set of EU oils, the full cross validation and the external validation permitted to rightly classify the 86 and 80% of samples, respectively, and 9 out of 12 EU blends were also correctly classified. Other samples have to be analysed to verify the satisfactory results obtained, in particular it will be necessary to add non-EU samples in order to better balance the two sets of samples.
Alessandra Bendini (Università di Bologna, UNIBO), Stefania Vichi and Alba Tres (Universitat de Barcelona, UB), Wenceslao Moreda and Diego Luis García-González (Instituto de la Grasa, CSIC)