Loading...

Development and Application of Pattern Recognition Methods Combined with FT-IR Spectroscopy and HPTLC Techniques to Detect the Type and Amount of Fraud in Iranian Saffron Sample

Amirvaresi, Arian | 2020

400 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 52730 (03)
  4. University: Sharif University of Technology
  5. Department: Chemistry
  6. Advisor(s): Parastar, Hadi; Daraei, Bahram; Amir Ahmadi, Maryam
  7. Abstract:
  8. Nowadays, with increasing concerns about health of food and its influence on human health,food authenticity has become a vital issue and a major challenge for consumers and regulatory agencies. It's especially remarkable about food with high nutritional and economical value. On the other hand, solving these challenges necessitates the use of fast and reliable techniques.Among the various foods, saffron is the main candidate for food adulteration due to its low production and high economic value and from this viewpoint it's in fourth place. Since the scope of fraud in this food is vast and also Iran is the major producer of saffron in the world,the present project has highlighted the importance of investigating this issue. Therefore, in order to verify the authenticity of this food, two domains including geographical origin and common adulteration has been investigated by near infrared, KBr pellet, attenuated total reflectance spectroscopy and high performance thin layer chromatography (HPTLC) based on pattern recognition techniques. The analysis of fingerprints pattern obtained by different instruments due to the complexity of saffron texture as well as the responses of different techniques requires the use of different data preprocessing and pattern recognition methods. In the first step of this study, a data exploration method, Principal Component Analysis (PCA)was used to investigate the geographical origin of 100 saffron samples from two provinces include southern Khorasan and Khorasan Razavi. The results showed that HPTLC technique with 90% accordance with geographical origin of the samples had the best results in this regard.Afterward, NIR technique with 80% and KBr spectroscopy with 70% are most compliance with the geographical origin of the samples. ATR spectroscopy technique is not suitable for use in this field Due to the lack of a regular pattern and consistent with the geographical origin of the samples. According to the results obtained from the geographical origin of the samples,four common adulterant materials in saffron, namely safflower, calendula, madder, and style were studied. The PLS-DA method was used to build the model of adulterated samples for all techniques and for evaluate the quality of the developed models, two sample categories including a number of adulteration market samples and a mixture of adulteration were modeled. Model accuracy range for different classes model of NIR, ATR and KBr techniques were 93-100%, 90-100% and 91-100%, respectively. Of all the techniques, only NIR was able to determine the fraud rate, with a mean R2 = 0.98 for all classes. In the study of ATR samples due to the nonlinear behavior of the data a new method based on special preprocessing and its integration with the support vector machine method was applied and the accuracy of the saffron class was increased from 50 of PLS-DA to 100% of QSVM model. Using the results from HPTLC images, it was also shown that this technique is very effective in detecting adulteration and can even detect fraud in some cases up to 1%. As a result, the developed models can be used as a standard method for determining the geographical origin and investigating the types of saffron adulteration
  9. Keywords:
  10. Saffron ; Spectroscopy ; Chromatography ; Pattern Recognition ; Chemometrics Method ; Supervised Analysis ; Unsupervised Analysis ; Skulduggery

 Digital Object List

 Bookmark

No TOC