Unequivocal identification and quantification of PAHs content in ternary synthetic mixtures and in smoked tuna by means of excitation-emission fluorescence spectroscopy coupled with PARAFAC

Journal Pre-proof Unequivocal identification and quantification of PAHs content in ternary synthetic mixtures and in smoked tuna by means of excitation-emission fluorescence spectroscopy coupled with PARAFAC S. Catena InvestigationWriting - Original DraftSupervision , S. Sanllorente InvestigationSupervisionWriting - Review & Editing , L.A. Sarabia Formal analysisSupervisionWriting - Review & Editing , R. Boggia Writing - Original Draft , F. Turrini Writing - Original Draft , M.C. Ortiz ConceptualizatioSupervisionWriting - Review & Editing PII: S0026-265X(19)33369-7 DOI: https://doi.org/10.1016/j.microc.2019.104561 Reference: MICROC 104561 To appear in: Microchemical Journal Received date: 28 November 2019 Revised date: 19 December 2019 Accepted date: 19 December 2019 Please cite this article as: S. Catena InvestigationWriting - Original DraftSupervision , S. Sanllorente InvestigationSupervisionWriting - Review & Editing , L.A. Sarabia Formal analysisSupervisionWriting - Review & Editing , R. Boggia Writing - Original Draft , F. Turrini Writing - Original Draft , M.C. Ortiz ConceptualizatioSupervisionWriting - Review & Editing , Unequivocal identification and quantification of PAHs content in ternary synthetic mixtures and in smoked tuna by means of excitation-emission fluorescence spectroscopy coupled with PARAFAC, Microchemical Journal (2019), doi: https://doi.org/10.1016/j.microc.2019.104561 This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ©2019 Published by Elsevier B.V. 1 HIGHLIGHTS  PARAFAC a usefull tool coupled with excitation-emission fluorescence spectroscopy  EEM-PARAFAC, a cheap, green and usefull tool to determine PAHS in smoked tuna  EEM-PARAFAC succeeds in identifying unequivocally and quantifying three PAHs  Using a reduced orthogonal design to obtain mixtures for calibration standards 2 Unequivocal identification and quantification of PAHs content in ternary synthetic mixtures and in smoked tuna by means of excitation-emission fluorescence spectroscopy coupled with PARAFAC S. Catena (1), S. Sanllorente (2), L.A. Sarabia (3), R. Boggia (1), F. Turrini (1), M.C. Ortiz (2)* (1) Department of Pharmacy, University of Genova, Viale Cembrano 4, 16148 Genova Italy. (2) Department of Chemistry, Faculty of Sciences, University of Burgos, Plaza Misael Bañuelos s/n, 09001 Burgos, Spain (3) Department of Mathematics and Computation, Faculty of Sciences, University of Burgos, Plaza Misael Bañuelos s/n, 09001 Burgos, Spain. *Corresponding author. E-mail address: [email protected] 3 Abstract It is well known that several polycyclic aromatic hydrocarbons (PAHs), products of incomplete pyrolysis of organic material, have proved to be extremely toxic to humans. Food can be contaminated by these compounds in many different ways and diet represents nowadays the major source of exposure to PAHs for non-smokers population. In the present study, three of the most important carcinogenic PAHs in foods, according to the legislation currently in force, i.e. benzo[a]pyrene (BaP), benzo[a]anthracene (BaA) and chrysene (Chry), were firstly arranged in ternary mixtures, following an experimental design. Then, an organic extraction from a commercial smoked tuna, potentially affected by PAHs, was performed to investigate the presence of the three compounds under study. A spectrofluorimetric method based on the second order calibration of excitation-emission fluorescence matrices (EEMs) and parallel factor analysis (PARAFAC) decomposition was proposed in this work as analytical approach for PAHs detection. Both in the ternary mixtures and in the food matrix (smoked tuna), PAHs were unequivocally identified and quantified with decision limit (CCα) and capability of detection (CCβ) equal to 0.11 µg L-1 and 0.21 µg L-1 for BaP, 0.27 µg L-1 and 0.53 µg L-1 for BaA and 0.18 µg L-1 and 0.35 µg L1 for Chry, respectively, when the probabilities of false positive (α) and false negative (β) were fixed at 0.05. In the investigated smoked tuna, detectable levels of BaP were found, whereas BaA and Chry were absent. Keywords Polycyclic aromatic hydrocarbons; excitation-emission matrix fluorescence spectroscopy; parallel factor analysis; smoked tuna; unequivocal identification. Abbreviations 4 PAHs, Polycyclic aromatic hydrocarbons; BaP, benzo[a]pyrene; BaA, benzo[a]anthracene; Chry, chrysene; PARAFAC, parallel factor analysis; EEM, excitation– emission fluorescence matrix; CCα, decision limit; CCβ, capability of detection; DMSO, dimethylsulfoxide ; DMF, N, N-dimethylformamide; EPA, Environmental Protection Agency; IARC, International Agency for Research on Cancer; SCF, Scientific Committee on Food; EFSA, European Food Safety Authority; AOAC, Association of Official Analytical Chemists; SPE, solid phase extraction; SFE, supercritical fluid extraction; ASE, accelerated solvent extraction; MAE, microwave assisted extraction; SPME, solid-phase microextraction; GPC, gel permeation chromatography; HPLC, high performance liquid chromatography; GC-MS, Gas Chromatography–Mass Spectrometry; HRGC-MS, high resolution GC-MS. 1. Introduction Polycyclic aromatic hydrocarbons (PAHs) are a group of organic compounds composed by multiple aromatics rings which are widespread in the environment as products of incomplete combustion or pyrolysis of organic material, such as coal, petroleum, wood and natural gas [1,2]. 16 PAHs have been classified for decades as priority environmental pollutants by the European Union (EU) and the US Environmental Protection Agency (EPA) [3]. Food can be contaminated by PAHs that are present in water, air, soil or packaging materials, as well as those that are formed during food processing or certain home cooking practices, e.g. smoking, barbecuing, grilling, roasting, toasting, heating, drying, baking, frying and ohmic-infrared cooking [4]. Thus, the major source of PAHs exposure, in non-cigarette smokers and non-occupationally exposed workers, is diet, responsible for more than 90% of the total exposure to PAHs of population in many countries of the world [5]. 5 A substantial number of studies showed that PAHs are extremely toxic to lung, breast, stomach, esophagus and pancreas [6,7]. Several of these compounds, in fact, have been classified by the International Agency for Research on Cancer (IARC) as probably carcinogenic to humans (Group 2A) or possibly carcinogenic to humans (Group 2B) and benzo[a]pyrene (BaP) directly as carcinogenic to humans (Group 1) [8]. In 2002, the Scientific Committee on Food (SCF) [9] stated that benzo[a]pyrene (BaP) might be used as a marker of the occurrence of the carcinogenic PAHs in food products [10], but in 2008 a report of the European Food Safety Authority (EFSA), showed that BaP is not always detectable, providing a negative response in about 30% of all the testing samples, even if others PAHs, above all chrysene, were found [11]. Thus, the EFSA „Panel on Contaminants in the Food Chain‟ highlighted that benzo[a]pyrene (BaP), chrysene (Chry), benzo[a]anthracene (BaA) and benzo[b]fluoranthene (PAH4) plus four additional PAHs, i.e. benzo[k]fluoranthene, dibenzo[a,h]anthracene, benzo[g,h,i]perylene, indeno[1,2,3-cd]pyrene (PAH8), are currently the only possible indicators for the carcinogenic potency of PAHs in food. This led to the Regulation UE 835/2011, which sets the maximum level of BaP in smoked meat and smoked fish products to 5 µg kg-1 together with the additional limit of the sum of PAH4 (30 µg kg-1) [12]. Smoked foods present significant variations in PAH amounts, even within the same kind of food. This is due to the variability of smoke composition, which is greatly affected by process environment oxidizing power, combustion temperature, type of generator, smoking time, fuel used, presence of smoking flavour agents and fat content in the food [4]. Traditional smoking techniques, in which the smoke produced by the incomplete combustion of the wood comes into direct contact with the product, can lead to a high contamination by PAHs; for this reason, commercial alternatives are used nowadays, like immerging food items in smoke extracted solution (the so called "liquid smoke") [13]. 6 The determination of PAHs in foods requires a preliminary extraction followed by multistep clean-up procedures [14], to isolate these compounds, occurring at the level of µg kg-1 or ng kg-1, from all the interferers present in a complex fat and protein containing food matrix. Standard methods for PAHs detection in food have been published by Association of Official Analytical Chemists (AOAC) [15]. Grimmer and Böhnke‟s procedure [16], involving saponification, extraction with cyclohexane, liquid-liquid partition of cyclohexane extract with N, N-dimethylformamide (DMF), followed by silica gel column chromatography is still widely used. Its most common modifications concern change of extraction and partition solvents (e.g. dimethylsulfoxide, specific solvent for PAHs which allows to separated them from triglycerides, instead of DMF [17] and use of pre-packed cartridges for solid phase extraction (SPE) [18], which guarantee time and solvent savings as well as better reproducibility, compared to chromatographic columns. In addition to liquid–liquid extraction, column chromatography and SPE, other techniques like supercritical fluid extraction (SFE) [19], microwave assisted extraction (MAE) [20], solid-phase microextraction (SPME) [21]. Furthermore, concerning the analytical methods, Gas Chromatography–Mass Spectrometry (GC–MS) [22], high resolution GC-MS (HRGC–MS) [23] and high performance liquid chromatography (HPLC) fluorescence detection (FLD) or Diode Array Detector (DAD) [24,25] are nowadays the most applied techniques for qualitative–quantitative analysis of PAHs in food. In the present work, molecular fluorescence spectroscopy has been used for the determination of BaP, BaA and Chry in a commercial smoked tuna. A search with keywords “EEM” and (PAHs OR “polycyclic aromatic hydrocarbons”) in Scopus gave 276 papers since 1980, none of them dealt with the determination of PAHs by using excitationemission fluorescence in smoked food products. This technique was selected as analytical technique because highly sensitive, easy to use, fast, non-destructive and lowcost. The interpretation of fluorescence spectral data of the three compounds is complex 7 due to overlapping signals of other molecules or ions, since fluorescence is a nonseparative technique; for this reason, the excitation-emission matrices (EEM) need to be analyzed by means of three ways techniques in order to extract useful information from a huge amount of data. PARAFAC decomposition of trilinear data provides unique profiles estimations of the different fluorophores, when an appropriate number of factors is chosen to fit the model. This means that it is possible to identify unequivocally each factor with each analyte, thus separating the signals of each fluorophore. Moreover, thanks to the second order property of PARAFAC, it is also possible to quantify in presence of uncalibrated interferers. In this context, ternary mixtures of BaP, BaA and Chry were prepared and analysed with fluorescence spectroscopy coupled with PARAFAC as chemometric tool. Finally, the procedure was used to determination and identification of these PAHs in smoked tuna. 2. Materials and methods 2.1. Chemical compounds and reagents Benzo[a]pyrene (CAS no. 50-32-8; 96% purity) was purchased from Alfa Aesar Gmbh (Karlsruhe, Germany), benzo[a]anthracene (CAS no. 56-55-3; 99% purity) from Acros Organics (Geel, Belgium) and chrysene (CAS no. 218-01-9; 95% purity) from Merck KGaA (Darmstadt, Germany), n-Hexane (CAS no. 110-54-3; for liquid chromatography LiChrosolv®) was supplied by Merck KGaA (Darmstadt, Germany) and dimethyl sulfoxide (CAS no. 67-68-5; 99% purity) by VWR International S.A.S. (Fontenay-sous-Bois, France). (Chemical compounds in table S1 in supplementary material). Deionised water was obtained by using the Milli-Q gradient A10 water purification system from Millipore (Bedford, MA, USA). Sep-Pak Silica Plus Long cartridge (690 mg Sorbent per Cartridge, 55-105 µm Particle Size) and Sep-Pak Plus Short C-18 cartridges (360 mg Sorbent per Cartridge, 55-105 µm Particle Size) were supplied by Waters (Milford, MA, USA). 2.2. Standard solutions of selected PAHs and food matrix 8 Stock solutions of each polycyclic aromatic hydrocarbon were individually prepared in nhexane at a concentration of 100 mg L−1. Intermediate solutions for each analyte at 100 µg L−1 were prepared by dilution with n-hexane. Then, final solutions for the analysis were prepared daily by further dilution with n-hexane from the intermediate solutions. All these solutions were stored at low temperature (4°C) in amber flasks. Smoked tuna was obtained from a local supermarket in the area of Burgos, Spain. 2.3. Instruments For the lyophilization of smoked tuna, a freeze-drier FreeZone 12 L Console Freeze-Dry System with drying chamber, Labconco was used. Ultrasound-assisted extraction (UAE) of lyophilized smoked tuna was carried out in an ultrasonic bath (BandelinSonorex RK52, Berlin, Germany) with 35 kHz frequency, maximum power of 240 watts, and internal dimension of 150×140×100 mm. Sample pre-concentration was performed in a rotary evaporator at pressure of 335 mbar (ILMVAC, Ilmenau, Germany). The excitation-emission fluorescence measurements were performed at room temperature on a PerkinElmer LS 50B Luminescence Spectrometer (Waltham, MA, USA) equipped with a xenon discharge lamp. A standard cell holder and a 10 mm quartz SUPRASIL® cell with cell volume of 3.5 mL by PerkinElmer (Waltham, MA, USA) were used. EEMs were recorded in the following ranges: emission wavelengths from 340 to 500.5 nm (each 0.5 nm) and excitation ones from 230 to 305 nm (each 5 nm). Excitation and emission monochromator slit widths were both set to 10 nm and the scan speed was 1500 nm min−1. 2.4. Software 15 In the present work, the identification of each analyte was performed by means of the correlation between its emission and excitation reference spectra and the emission and excitation loadings estimated from the corresponding PARAFAC mode. The correlation coefficients for the emission and excitation profiles were 0.992 and 0.995 for BaP, 0.980 and 0.963 for BaA and 0.996 and 0.991 for Chry, respectively. A least square regression between the sample loadings and the true concentration of each compound in the mixtures was performed for the three PAHs. Then, the regressions „predicted concentration‟ versus „true concentration‟ [39]. Table 3 shows the parameters of this regression line. The procedure is unbiased, for all the analytes because the p-values are greater than 0.05 in both cases. 4.3. Validation and prediction of ternary mixtures’ model To validate the model built with the five-levels experimental design, the 31 analysis were performed again, but, in this case, they were preceded by the measurements of the single compounds in different concentrations to carry out three new calibration lines, in the range of the five levels. The concentration for each PAHs are shown in table 4. A new three-way array X5 (52x322x16) was built with the composition of 31 samples in table 1 plus the first 21 samples in table 4. Then, a PARAFAC decomposition with four factors (with non-negativity constraints) was applied to X5, obtaining a CORCONDIA index of 99%, with explained variance equal to 99.84%. Internal validation was carried out with the sample loadings of the PARAFAC model. Least square regressions between the sample loadings and the true concentration were performed using the first 37 loadings of the model that constitute the training set. The last 15 sample loadings (corresponding to the second half of the experimental design from table 1) were used as test set. This accuracy lines in prediction was validated as in the previous case, being the mean of 16 relative errors in absolute value in prediction for BaP, BaA and Chry 6.8%, 3.6% and 2.5%. 4.4 Smoked tuna extraction As already described in section 2.7, two portions of 5 g lyophilized smoked tuna were extracted as such (named extract A and B) and others two (extract C and D) were added with 300 µL of BaP, BaA and Chry intermediate solutions. After the extraction procedure, the four extracts were dissolved in different amount of nhexane to be analysed, based on the different intensity of each fluorescence signals: extracts B and C were dissolved in 30 mL of n-hexane, extract A in 20 mL and D in 10 mL. Then, all of them were diluted three times, thus analysing each extract in four different dilutions. As in the case of the validation of ternary mixtures (section 4.3), new calibration lines of BaP, BaA and Chry were built and analysed in the same day of the extracts measurements. The data for this study was built by combining the fluorescence spectra of the extracts A, B, C and D (16 samples) with those of the new calibration models again according to table 4 (26 samples) and of the 31 experiments that made up the ternary mixtures design. So, the three-way data tensor X6 (73x317x15). It can be seen from the dimension of the tensor X6 that the emission and excitation ranges have been reduced by some wavelength, to avoid signal saturation problems. The PARAFAC decomposition of this resulting tensor needed five factors and showed a CORCONDIA index equal to 88% and explained variance equal to 99.82%. Fig. 3 shows sample (3a), emission (3b) and excitation (3c) profiles. Again, for the identification of the PAHs, the correlation between the PARAFAC spectral loadings and the reference spectra have been used. Fig. 4 shows the comparison 17 between the emission reference spectra (in red) and the emission loadings of the PARAFAC obtained after smoked tuna extraction (in blue). It must be highlighted that the values have been normalized in this figure to compare them. The correlation coefficients for the emission and excitation profiles were 0.975 and 0.982 for BaP, 0.993 and 0.996 for BaA and 0.997 and 0.993 for Chry, respectively. Once the decomposition has been done by using the loadings corresponding to the sample profile, it is possible to conclude that only BaP is present in the smoked tuna investigated (see samples from number 58 to number 65, blue diamonds in Fig. 3 a). The average concentration of BaP in extracts A and B, bearing in mind the dilution factors of each measurement, was 5.42 µg kg-1, calculated for kg of wet smoked tuna (before lyophilization), since the wet form, not the dry one, is what you eat. This BaP concentration is just in accordance with the maximum allowed by law; however, it must be specified that the amount of BaP calculated by PARAFAC, before the modifications related to the dilution factors and to the difference between wet weight and dry weight, was found to be 0.1 µg L-1, which is just at the limit of CCα (see table 2). The concentrations of BaA and Chry, in both extracts A and B, were not significantly different from zero. Meanwhile, analysing the two extracts C and D obtained with the standard addition method, the average amount (n = 8) of each compound in both extracts resulted to be 2.23 µg L-1, 1.21 µg L-1, 1.94 µg L-1 for BaP, BaA and Chry, respectively. Their respective 95% confidence intervals, expressed in µg L-1 are [1.84, 2.62], [0.71, 1.72], and [1.68, 2.19]. The three PAHs were recovered with a percentage of 72.3 (BaP), 40.5 (BaA) and 64.6 (Chry). 5. Conclusions An analytical method based on EEM fluorescence spectroscopy associated with PARAFAC to detect polycyclic aromatic hydrocarbons (PAHs) has been set up and tuned by a five-level experimental design. This procedure has been performed for the first time 18 to detect three cancerogenic PAHs in a food matrix (smoked tuna fish). Ternary mixtures of BaP, BaA and Chry were prepared and analysed by the EEM/PARAFAC method for evaluating its feasibility in detecting and quantifying the three PAHs. Thanks to the „second order property‟ of PARAFAC, when mixtures were analized, BaP, Chry and BaA were unequivocally identified by means of the correlation between the pure spectra and the PARAFAC excitation and emission spectral loadings, being these correlations coefficients: 0.992 and 0.995 for BaP, 0.980 and 0.963 for BaA and 0.996 and 0.991 for Chry. Moreover, the accurate procedure for the three PAHs has been assured, being their capability of detection 0.21, 0.53 and 0.35 µg L-1 for BaP, BaA and Chry respectively in the case of individual calibration and 0.39, 0.25 and 0.29 µg L-1 for BaP, BaA and Chry in the ternary mixtures (in all case the probabilities of false positive and false negative were fixed at 0.05). In smoked tuna, only BaP was identified and quantified in a concentration just within the limit allowed by legislation; BaA and Chry were not detected. Sample CRediT author statement: S. Catena: Investigation , Writing - Original Draft, Supervision S. Sanllorente: Investigation ,Supervision, Writing - Review & Editing L.A. Sarabia Formal analysis, Supervision, Writing - Review & Editing R. Boggia: Writing - Original Draft F. Turrini: Writing - Original Draft M.C. Ortiz: Conceptualizatio, Supervision, Writing - Review & Editing Funding: The authors thank the financial support from Spanish MINECO (AEI/FEDER, UE) and Junta de Castilla y León through projects CTQ2017-88894-R and BU012P17 respectively (all co-financed with FEDER funds). Declaration of competing interest: There are no conflicts of interest to declare. 19 References [1] B. Zhao, S. Zhang, Y. Zhou, D. He, J. Fang, Characterization and quantification of PAH atmospheric pollution from a large petrochemical complex in Guangzhou: GC– MS/MS analysis, Microchem. J. 119 (2015) 140-144. https://doi.org/10.1016/j.microc.2014.12.004 [2] C.Y. dos S. Siqueira, M.V.P. Lemos, B.C. da C. Araujo, R. da R.P.E. de Oliveira, R.A. da S. San Gil, F.R. de Aquino Neto, Atmospheric distribution of organic compounds from urban areas near Olympic games sites in Rio de Janeiro, Brazil, Microchem. J. 133 (2017) 638-644 https://doi.org/10.1016/j.microc.2017.04.027 [3] Z. Zelinkova and T. Wenzl, The Occurrence of 16 EPA PAHs in Food – A Review. Polycycl. Aromat. Compd., 35 (2015) 248–284. https://doi.org/10.1080/10406638.2014.918550 [4] L. Singh, J.G. Varshney, T. Agarwal, Polycyclic aromatic hydrocarbons‟ formation and occurrence in processed food, Food Chem. 199 (2016) 768-781. https://doi.org/10.1016/j.foodchem.2015.12.074 [5] World Health Organization (WHO), (1998). Selected non-heterocyclic polycyclic aromatic hydrocarbons. Environment Health Criteria No. 202. Geneva, Switzerland: WHO. https://apps.who.int/iris/handle/10665/41958 [6] R. Goldman, L. Enewold, E. Pellizzari, J.B. Beach, E.D. Bowman, S.S. Krishnan, P.G. Shields, Smoking increases carcinogenic polycyclic aromatic hydrocarbons in human lung tissue, Cancer Res. 61 (2001) 6367–6371. [7] K.L. Harris, L.D. Banks, J.A. Mantey, A.C. Huderson, A. Ramesh, Bioaccessibility of polycyclic aromatic hydrocarbons: relevance to toxicity and carcinogenesis. Expert. Opin. Drug Metab. Toxicol. 9 (2013) 1465–1480. https://doi.org/10.1517/17425255.2013.823157 20 [8] IARC Monographs, (2010). Some non-heterocyclic polycyclic aromatic hydrocarbons and some related exposures, Vol. 92 (2010) Lyon, https://publications.iarc.fr/110 [9] European Commission Scientific Committee on Foods (2002). Opinion of the Scientific Committee on Food on the Risks to Human Health of Polycyclic Aromatic Hydrocarbons in Food. Brussels: Scientific Committee on Food. https://ec.europa.eu/food/sites/food/files/safety/docs/sci-com_scf_out153_en.pdf [10] Commission Regulation (EU) No. 1881/2006 of 19 December 2006 (2006). Setting maximum levels for certain contaminants in foodstuffs. This Regulation repels the Commission Regulation (EU) No. 208/2005 of 4 February 2005. [11] European Food Safety Authority (EFSA). (2008). Scientific opinion of the panel on contaminants in the food chain on a request from the European commission on polycyclic aromatic hydrocarbons in food. The EFSA Journal, 724, 1e114. [12] European Commission (2011). COMMISSION REGULATION (EU) No 835/2011of 19 August 2011 amending Regulation (EC) No 1881/2006 as regards maximum levels for polycyclic aromatic hydrocarbons in foodstuffs. Off. J. Eur. Union, 215/4. [13] J.M. Lingbeck, P. Cordero, C.A. O'Bryan, M.G. Johnson, S.C. Ricke, P.G. Crandall, Functionality of liquid smoke as an all-natural antimicrobial in food preservation, Meat Sci. 97 (2014) 197–206. https://doi.org/10.1016/j.meatsci.2014.02.003 [14] R. Fuoco, S. Giannarelli, M. Onor, A. Ceccarini, V. Carli, Optimized cleanup methods of organic extracts for the determination of organic pollutants in biological samples, Microchem. J. 79 (2005) 69-76. https://doi.org/10.1016/j.microc.2004.06.009 [15] AOAC Official Method 973.30. Polycyclic Aromatic Hydrocarbons and Benzo[a]pyrene in Food–Spectrophotometric method, in: W. Harwitz (Ed.), Official 21 methods of analysis, Association of Official Analytical Chemists, Inc., Arlington (Virginia), 2000, pp. 1-3. [16] G. Grimmer, H. Bohnke, Polycyclic aromatic hydrocarbon profile analysis of highprotein food & oils, and fats by gas chromatography, J. Assoc. Off. Anal. Chem. 58 (1975) 725-733. [17] E. Corradetti, C. Abbondanza, L. Mazzanti, G. Poli, Determinazione gascromatografica e spettrofluorimetrica degli Idrocarburi Policiclici Aromatici (IPA) nell‟olio extravergine di oliva prodotto da olive contaminate da condensa di pece di origine industriale. Considerazioni sulle possibili vie di contaminazione, Boll. Chim. Igien. 39 (1988) 297-317. [18] M.N. Kayali-Sayadi, S. Rubio-Barroso, M.P. Cuesta-Jimenez, L.M. Polo-Díez, Rapid determination of polycyclic aromatic hydrocarbons in tea infusion samples by highperformance liquid chromatography and fluorimetric detection based on solid-phase extraction, Analyst 123 (1998) 2145-2148. DOI: 10.1039/A803967D [19] E. Jävenpää, R. Huopalahti, P. Tapanainen, Use of supercritical fluid extraction-high performance liquid chromatography in the determination of polynuclear aromatic hydrocarbons from smoked and broiled fish, J. Liq. Chromatogr. Relat. Technol. 19 (1996) 1473-1482. https://doi.org/10.1080/10826079608007196 [20] G. Purcaro, S. Moret, L.S. Conte, Optimisation of microwave assisted extraction (MAE) for polycyclic aromatic hydrocarbon (PAH) determination in smoked meat, Meat Sci. 81 (2009) 275–280. https://doi.org/10.1016/j.meatsci.2008.08.002 [21] S. M. S. Jillani, M. Sajid, K. Alhooshani, Evaluation of carbon foam as an adsorbent in stir-bar supported micro-solid-phase extraction coupled with gas chromatography–mass spectrometry for the determination of polyaromatic hydrocarbons in wastewater samples, Microchem. J. 144 (2019) 361-368. https://doi.org/10.1016/j.microc.2018.09.027 22 [22] H.C. Menezes, M. J. N. Paiva, R. R. Santos, L. P. Sousa, Z. L. Cardeal, A sensitive GC/MS method using cold fiber SPME to determine polycyclic aromatic hydrocarbons in spring water, Microchem. J. 110 (2013) 209-214. https://doi.org/10.1016/j.microc.2013.03.010 [23] S. Wretling, A. Eriksson, G.A. Eskhult, B. Larsson, Polycyclic aromatic hydrocarbons (PAHs) in Swedish smoked meat and fish, J. Food Compos. Anal. 23 (2010) 264– 272. https://doi.org/10.1016/j.jfca.2009.10.003 [24] D. Thompson, D. Jolley, W. Maher, Determination of Polycyclic Aromatic Hydrocarbons in Oyster Tissues by High-Performance Liquid Chromatography with Ultraviolet and Fluorescence Detection, Microchem. J. 47(1993) 351-362. https://doi.org/10.1006/mchj.1993.1054 [25] N. Akvan, G. Azimi, H. Parastar, Chemometric assisted determination of 16 PAHs in water samples by ultrasonic assisted emulsification microextraction followed by fast high-performance liquid chromatography with diode array detector, Microchem. J. 150 (2019) 104056. https://doi.org/10.1016/j.microc.2019.104056 [26] B.M. Wise, N.B. Gallagher, R. Bro, J.M. Shaver, PLS Toolbox 6.0.1, Eigenvector Research Inc., Wenatchee, WA, USA (2013). [27] STATGRAPHICS Centurion 18 Version 18.1.11 (64 bit), Statpoint Technologies, Inc., Warrenton, VA, USA, 2018. [28] L. Sarabia, M.C. Ortiz, DETARCHI: A program for detection limits with specified assurance probabilities and characteristic curves of detection, Trac. Trends Anal. Chem. 13 (1994) 1–6. https://doi.org/10.1016/0165-9936(94)85052-6 [29] R.G. Brereton, Multilevel multifactor designs for multivariate calibration, Analyst 122 (1997) 1521-1529. DOI: 10.1039/A703654J [30] M.S. Garcıa-Falcon, M.J. Lopez de Alda Villaizan, S. Gonzalez Amigo, J. Simal Lozano, M.A. Lage Yusty, Enrichment of benzo[a]pyrene in smoked food products 23 and determination by HPLCFL, J. Chromatogr. 753 (1996) 207–215. https://doi.org/10.1016/S0021-9673(96)00522-5 [31] M.D. Guillen, Polycyclic aromatic compounds: extraction and determination in food, Food Addit. Contam. 11 (1994) 669-684. DOI:10.1080/02652039409374268 [32] C.D. Geddes, J.R. Lakowicz, Metal-Enhanced Fluorescence, J. Fluoresc. 12 (2002) 121-129. DOI:10.1023/A:1016875709579. [33] N. Rodríguez, B. D. Real, M. C. Ortiz, L. A. Sarabia, A. Herrero. Usefulness of parallel factor analysis to handle the matrix effect in the fluorescence determination of tetracycline in whey milk, Anal. Chim. Acta 632 (2009) 42–51. https://doi.org/10.1016/j.aca.2008.10.051. [34] R. Bro, H.A.L. Kiers, A new efficient method for determining the number of components in PARAFAC models, J. Chemom. 17 (2003) 274–286. https://onlinelibrary.wiley.com/doi/abs/10.1002/cem.801 [35] M.C. Ortiz, L.A. Sarabia, M.S. Sánchez, A. Herrero, S. Sanllorente, C. Reguera, Usefulness of PARAFAC for the quantification, identification, and description of analytical data, in: A. Muñoz de la Peña, H.C. Goicoechea, G.M. Escandar, A.C. Olivieri (Eds.), Data handling in science and technology: fundamentals and analytical applications of multiway calibration. Elsevier, Amsterdam, 2015, pp. 3781. [36] A. C. Olivieri, G. M. Escandar, Parallel Factor Analysis: Trilinear Data, in: A. C. Olivieri, G. M. Escandar (Eds), Practical Three-Way Calibration. Elsevier, Amsterdam, 2014, pp. 65-92. [37] J.A. Arancibia, C.E. Boschetti, A.C. Olivieri, G.M. Escandar, Screening of oil samples on the basis of excitation–emission room-temperature phosphorescence data and multiway chemometric techniques. Introducing the second-order advantage in a classification study, Anal. Chem. 80 (2008) 2789–2798. DOI:10.1021/ac702364n 24 [38] R. Rajkó, N. Omidikia, H. Abdollahi, M. Kompany-Zareh, On uniqueness of the nonnegative decomposition of two-and three-component three-way data arrays, Chemom. Intell. Lab. Systems 170 (2017) 91-98. https://dx.doi.org/10.1016/j.chemolab.2016.12.001. [39] International Organization for Standardization, ISO 5725–2, Accuracy (Trueness and Precision) of Measurement Methods and Results, Part 2: Basic Method for the Determination of Repeatability and Reproducibility of a Standard Measurement Method, Genève, Switzerland (1994). 31 Table 4 Distribution of concentrations for the three PAHs under study used to perform the calibration Sample BaA (μg L−1) BaP (μg L−1) Chry (μg L−1) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 0 0.5 1 1 3 5 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 6 3 0 0 0 0 0 0 0 0 0 0.5 1 1 3 5 7 0 0 0 0 0 0 0 0 3 6 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 1 1 3 5 7 0 6 3 6 0 Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.