One hundred and thirty-eight oil samples have been analyzed by visible and near-infrared transflectance spectroscopy. These comprised 46 pure extra virgin olive oils and the same oils adulterated with 1% (w/w) and 5% (w/w) sunflower oil. A number of multivariate mathematical approaches were investigated to detect and quantify the sunflower oil adulterant. These included hierarchical cluster analysis, soft independent modeling of class analogy (SIMCA method), and partial least squares regression (PLS). A number of wavelength ranges and data pretreatments were explored. The accuracy of these mathematical models was compared, and the most successful models were identified. Complete classification accuracy was achieved using 1st derivative spectral data in the 400-2498 nm range. Prediction of adulterant content was possible with a standard error equal to 0.8% using 1st derivative data between 1100 and 2498 nm. Spectral features and chemical literature were studied to isolate the structural basis for these models.

Visible and near-infrared reflectance spectra have been examined for their ability to classify extra virgin olive oils from the eastern Mediterranean on the basis of their geographic origin. Classification strategies investigated were partial least-squares regression, factorial discriminant analysis, and k-nearest neighbors analysis. Discriminant models were developed and evaluated using spectral data in the visible (400-750 nm), near-infrared (1100-2498 nm), and combined (400-2498 nm) wavelength ranges. A variety of data pretreatments was applied. Best results were obtained using factorial discriminant analysis on raw spectral data over the combined wavelength range; a correct classification rate of 93.9% was obtained on a prediction sample set. Though the overall sample set was limited in numbers, these results demonstrate the potential of near-infrared spectroscopy to classify extra virgin olive oils on the basis of their geographic origin.

The limits of quantitative multivariate assays for the analysis of extra virgin olive oil samples from various Greek sites adulterated by sunflower oil have been evaluated based on their Fourier transform (FT) Raman spectra. Different strategies for wavelength selection were tested for calculating optimal partial least squares (PLS) models. Compared to the full spectrum methods previously applied, the optimum standard error of prediction (SEP) for the sunflower oil concentrations in spiked olive oil samples could be significantly reduced. One efficient approach (PMMS, pair-wise minima and maxima selection) used a special variable selection strategy based on a pair-wise consideration of significant respective minima and maxima of PLS regression vectors, calculated for broad spectral intervals and a low number of PLS factors. PMMS provided robust calibration models with a small number of variables. On the other hand, the Tabu search strategy recently published (search process guided by restrictions leading to Tabu list) achieved lower SEP values but at the cost of extensive computing time when searching for a global minimum and less robust calibration models. Robustness was tested by using packages of ten and twenty randomly selected samples within cross-validation for calculating independent prediction values. The best SEP values for a one year's harvest with a total number of 66 Cretian samples were obtained by such spectral variable optimized PLS calibration models using leave-20-out cross-validation (values between 0.5 and 0.7% by weight). For the more complex population of olive oil samples from all over Greece (total number of 92 samples), results were between 0.7 and 0.9% by weight with a cross-validation sample package size of 20. Notably, the calibration method with Tabu variable selection has been shown to be a valid chemometric approach by which a single model can be applied with a low SEP of 1.4% for olive oil samples across three different harvest years.

Infrared attenuated total reflectance spectroscopy has been assessed for the analysis of extra virgin olive oil samples from various Mediterranean sites and their adulteration by sunflower oil. In this study two different silver halide fiber-optic probes were separately tested for the mid-infrared spectroscopic measurement of pure olive oil samples and these same oils adulterated with sunflower oil. One fiber-optic probe contained an exchangeable U-shaped section of the silver halide fiber, whereas the second probe consisted of a fiber-coupled diamond crystal, which performed slightly less well than the whole fiber probe. The optimum standard error of prediction for the sunflower oil concentrations in spiked olive oil samples, obtained by optimized partial least-squares (PLS) calibration models and leave-one-out cross-validation, was 1.2% by weight with the use of a special variable selection strategy based on a pairwise consideration of significant respective minima and maxima of the optimum PLS regression vector, calculated for broad spectral intervals. Calibration robustness was proven by also using packages of 10 randomly selected samples within a further cross-validation for calculating independent prediction values. The implications for product monitoring are discussed.

Determ ination of the authenticity of extra virgin olive oils has become more important in recent years following some infamous adulteration and contamination scandals. There is signi cant economic advantage to be gained by falsely m arketing lower quality oils as top-quality products or by adulterating the end product with cheaper alternative oil. The analysis of olive oils is com plicated by their complex nature. The conventional analytical techniques involve complicated analyses of individual components in the oil in order to determine their authenticity; however, molecular spectroscop y combined with m odern chem ometric techniques offers a rapid analytical solution to this complex problem. This paper describ es the initial results from the experim ental application of Fourier transform Raman spectroscopy in this area.