Makarnalık Buğday ve Makarnada Ekmeklik Buğdayın Tespit Edilmesi için Bazı Spektroskopik Analiz Yöntemlerinin Kullanılması

Abstract

Although common wheat (Triticum aestivum) has poor pasta making qulity, some pasta producers blend common whea with durum wheat (Triticum durum) in pasta production because of economic concerns and lower price of common wheat. However, addition of common wheat to pasta without indicating on the label is considered adulteration. Hence, there is a need for developing fast and reliable methods to detect adulteration. Therefore, in this thesis, the performance of different spectroscopic methods was investigated for discrimination of common and durum wheat and the detection of adulteration in flour and pasta. Raman spectroscopy (RS), near infrared spectroscopy (NIRS), attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIRS), synchronous fluorescence spectroscopy (SFS) and laser induced breakdown spectroscopy (LIBS) were used to analyse samples with combination of some chemometric methods such as principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA) and partial least squares regression analysis (PLSR). In the first part of the thesis; RS, LIBS, ATR-FTIRS, NIRS, SFS and RP-HPLC were used to discriminate 120 common wheat and 119 durum wheat flour samples. The most effective method was SFS for discrimination of common and durum wheat flour. It was observed that common and durum wheat were clearly separated from each other by SFS with a sensitivity and specificity ratio of 1.000. Performance parameters of SFS; root mean square error of calibration (RMSEC), root mean squared error of cross-validation (RMSECV) and root mean square error of prediction (RMSEP) were 0.164, 0.171, and 0.194, respectively. The performance of NIR and ATR-FTIR was very similar. It has been found that the cross validation sensitivity and specificity values of ATR-FTIR and NIR were 0.990 while their calibration and prediction sensitivity and specificity values were 1.000. Common and durum wheat discrimination performance of LIBS was less sensitive than NIRS and ATR-FTIRS. RMSEC, RMSECV and RMSEP values of LIBS were 0.206, 0.185 and 0.189, and Hotelling's T2 value was 99.90. The discrimination ability of RS was less than those of other spectroscopic methods. The prediction sensitivity of the RS was 0.350, while the cross validation sensitivity and specificity of RS were 0.960 and 1.000. In addition to spectroscopic methods, reverse phase high performance liquid chromatography (RP-HPLC) was used for extensive characterization of the samples. The results of the chemometric analysis of RP-HPLC; true positive rate (TPR) and true negative rate (TNR) were 1.000 and false positive rate (FPR) and false negative rate (FNR) were 0.000. The RMSEP of RP-HPLC had the lowest value (0.144) in all methods. In the second part of the thesis, the aim was to determine the rate of adulteration in blended flour and pasta samples. The same spectroscopic methods were used to analyse 28 blended flour samples and 28 pasta samples. According to the result of blended flour samples, the most effective method was LIBS. LIBS had the lowest limit of detection (LOD, 0.523), limit of quantification (LOQ, 1.584) and RMSEP (5.297) values compared to other methods, as well as the highest coefficient of determination (R2, 0.961) for detection of common wheat flour in durum wheat flour samples. The calibration and cross validation R2 values of LIBS were 0.971 and 0.951. As in the first part of the thesis, the performance of ATR-FTIRS and NIRS was very similar. The Hotelling’s T2 (%) and Q-Residuals (%) values of NIRS and ATR-FTIRS were 100.00 and 0.00. RS had shown the worst performance in detecting the rate of adulteration in blended flour samples similar to its performance in discrimination of the wheat samples. The RMSEC, RMSECV and RMSEP values of RS were 10.893, 15.022 and 13.884. Blended common and durum wheat samples were also analyzed using RP-HPLC. The highest calibration R2 (0.994) and cross validation R2 (0.990) values were obtained from RP-HPLC. It was found that the performance parameters of RP-HPLC were 2.251 and 3.522 for RMSEC and RMSECV values, respectively. According the result of the pasta samples prepared from common and durum wheat blends, the lowest LOD (0.494) and LOQ (1.496) values were obtanied from ATR-FTIRS. ATR-FTIRS had also lower RMSEC (7.781), RMSECV (9.417) and RMSEP (8.908) values compared to the other spectroscopic methods. The calibration R2, cross validation R2 and prediction R2 values of ATR-FTIRS were 0.922, 0.885 and 0.903, respectively. NIRS was also quite able to determine the addition level of farina in pasta. The LOD and LOQ values of NIRS were 1.079 and 3.270. The other performance parameters of NIRS; RMSEC, RMSECV and RMSEP were 9.595, 12.926 and 8.747, respectively. Although RP-HPLC had the highest LOD (6.473) and LOQ (19.616) values, the performance of RP-HPLC for the detection of adulteration was better than other spectoscopic methods with calibration R2, cross validation R2 and prediction R2 values 0.997, 0.995 and 0.995, respectively. Hence, RP-HPLC seems to have a high potential to determine the adulteration rate in pasta. The results revealed that spectrocopic techniques combined with chemometrics showed generally good potential to discriminate common and durum wheat flour samples and to determinate the addition level of common wheat in durum wheat flour and pasta samples. It was found that the most effective method for discrimination of common and durum wheat flours was SFS. On the other hand, LIBS and ATR-FTIRS had the best performance in determination of the adulteration in blended flour and blended pasta samples, respectively.