Screening for dioxin contamination in ®sh oil by PARAFAC and N-PLSR analysis of ¯uorescence landscapes ² Dorthe Kjœr Pedersen, Lars Munck and Søren Balling Engelsen* Food Technology, Department of Dairy and Food Science, Royal Veterinary and Agricultural University, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark Received 7 September 2001; Revised 28 February 2002; Accepted 4 April 2002 A preliminary investigation of fish oils demonstrates that fluorescence excitation±emission landscapes evaluated by three-way chemometric methods may be a candidate for an inexpensive screening method to indicate the level of contamination by dioxins and PCBs, which are normally analysed by expensive and time-consuming physicochemical separation techniques such as GC-MS. Fluorescence landscapes of 88 fish oils have been investigated and showed great variation due to species, season and treatment, depicting a variation in natural fluorescent components. The fluorescence landscapes were analysed by PARAFAC. Samples with similar fluorescence fingerprints were selected from a PARAFAC score plot, and local significant prediction models with PARAFAC/MLR, N-PLSR and PLSR were established with correlation coefficients in the range from r = 0.69 (n = 10) to r = 0.97 (n = 75) for dioxin and of r = 0.92 (n = 12) for PCB. Application of PARAFAC/MLR and N-PLSR to fluorescence landscapes of fish oils resulted in local regression models for dioxin determination with prediction errors below 1 ng kg 1 , which is comparable to the reference method. In the PARAFAC model, two of the modes give the excitation and emission spectra of the pure underlying fluorophores, while the third mode gives their individual concentrations. Excitation and emission optima for three or four PARAFAC components in each data set were identified, representing both positive and negative (quenching) correlation components. It is hypothesized that the quenching correlation may be effected by the joint contribution of chlorinated organic compounds in the fish oil, including dioxins and PCBs. Other explanations for the results are discussed. Copyright # 2002 John Wiley & Sons, Ltd. KEYWORDS: dioxin; PCB; fluorescence; multi-way; PARAFAC; N-PLSR; fish oil 1. INTRODUCTION To an increasing extent, food and feed contain residues of environmental contaminants. Monitoring programmes are required to analyse food and feed for the presence of trace amounts of toxic substances such as heavy metals, pesti- cides, polyaromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs) and dioxins, flame retardants such as polybrominated diphenyl ethers (PBDEs), and oestrogenic compounds such as nonyl phenols and phthalates. Dioxins are widely encountered toxic substances. Dioxins is a short expression for polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs). Both are tricyclic, chlorine-substituted, aromatic organic compounds. Polychlorinated biphenyls (PCBs) are another class of environmental contaminants. In practice, dioxins appear as mixtures of various congeners with different concentrations and with extreme variations in toxicity. Dioxins are persistent and lipophilic compounds which bioaccumulate and bioconcentrate in the food chain. They are very toxic, acutely as well as chronically, and some dioxins might be carcinogens, immunotoxics, endocrine disruptors or terato- gens. The number of chlorine substituents may range from one to eight, which means 75 possible PCDD congeners and 135 possible PCDF congeners [1]. Congeners with chlorine substitution in the 2,3,7,8-positions (TCDDs) (Figure 1) are very persistent, bioaccumulative and toxic [1]. Dioxin is formed involuntarily as a by-product of many industrial processes involving chlorine; for example, in producing inflammable transformer oil. It is also produced when chlorinated substances are burned in the presence of carbon and oxygen and thus released into the environment through air pollution, chemical waste and soil contamination *Correspondence to: S. B. Engelsen, Food Technology, Department of Dairy and Food Science, Royal Veterinary and Agricultural University, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark. E-mail: se@kvl.dk ² Paper presented at the 7th Scandinavian Symposium on Chemometrics, Copenhagen, Denmark, 19±23 August 2001. Contract/grant sponsor: Ministry of Food, Agriculture and Fisheries. Copyright # 2002 John Wiley & Sons, Ltd. JOURNAL OF CHEMOMETRICS J. Chemometrics 2002; 16: 451±460 Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/cem.735