Thumbnail Image


Color is one of the most important quality attributes that affect consumers' selection of food. The increasing demand of consumers for natural colorants over artificial ones has placed challenges in both product development and regulatory practices. However, current analytical solutions for food colorants are mostly limited to a sophisticated laboratory setting with tedious sample preparation procedures. Herein, this research focuses on the analytical developments toward cost-effective determination of colorant adulteration and stability analysis. The main technique explored is Raman spectroscopy, which measures the inelastic light scattering and allows one to obtain unique molecular fingerprints for specific molecules. Compared with chromatographic methods, e.g., high-performance liquid chromatography (HPLC) or gas chromatography (GC), Raman spectroscopic methods show great potential in the analysis of food colorants due to their fingerprinting capability non-destructive nature, good portability, and easy preparation protocols. In addition, the integration of gold or silver metallic nano substrate to Raman spectroscopy, known as surface-enhanced Raman spectroscopy (SERS), improves detection sensitivity tremendously. The first study was aimed to develop a fast screening and quantification method combining thin-layer chromatography (TLC) and Raman spectroscopy for saffron powder analysis, which is one of the most expensive agricultural products in the world. Sample solutions of pure saffron and spiked saffron were prepared and dropped on TLC chips for the identification of pure saffron quality as well as discrimination of spiked saffron. The result indicated that Raman spectroscopy was capable of quantifying the major coloring compound, crocin, on dried TLC sample patterns. The method achieved excellent prediction capability for crocin concentration (expressed via absorbance value at 440nm, according to ISO3632) in the range of 0 to 400 with an RMECV of 13 and an R2 of 0.99 using partial least squares (PLS) regression model. The TLC patterns between adulterants (safflower, turmeric, red 40, and yellow 5) were distinct under bright and UV lights. Their variation was demonstrated using principal components analysis (PCA). The lowest adulteration degrees that can be detected were 2.59% for red 40, 4.15% for yellow 5, 31.01% for safflower, and 41.98% for turmeric respectively, using the PLS analysis. The next study was aimed to utilize data fusion strategies to improve the classification and quantification accuracies of saffron adulteration. In the data processing protocol, the imaging data and featured Raman data were concatenated into one data matrix. The model performance of the fused data was compared against each data block. The result indicated that the classification accuracy for saffron adulterants were greatly improved with a classification accuracy of 99% among 4 different adulterants ranging from 2% to 100% (w/w) using the fused data block as compared to the imaging data (84% accuracy) or Raman data (86% accuracy) alone. The quantification performances of the developed PLS model were improved slightly using the fused data block, achieving higher R2 values with lower errors. To improve the detection capability of TLC-Raman for adulterant analysis, a mirror “stamping” TLC-SERS approach was developed. The aim of this study was to build a simple approach for sample separation (TLC) and signal enhancement (SERS) using easily available materials. Homemade silver nanoparticles were fabricated, formed to a mirror, and stamped on developed TLC patterns for Raman signal amplification. Pure saffron (1000 ppm) and spiked saffron (200 ppm red 40) solutions were used as a model system. The result indicated that both saffron colorant(crocin) and adulterated colorant (red 40) exhibited a strong characteristic peak using the mirror stamp approach, whereas there was no or little signal without the mirror application. An arising challenge was reported when food manufacturers were trying to develop products that contain natural colorants with iron fortification. In detail, the natural colorant, anthocyanins, developed unwanted blue color shifts due to iron-anthocyanin co-pigmentation. Thus, the last study was aimed to establish a quantitative model for the prediction of anthocyanin color stability in the presence of dissolved iron. Nine anthocyanin extracts obtained from various sources were purified and measured using SERS. The PCA model successfully discriminated anthocyanins from different plant sources. The stabilization index of each anthocyanin extract was measured using UV-vis spectroscopy under pH 3 and 6 with and without iron fortification (ferric sulfate) and used as input for PLS model. The PLS model demonstrated high accuracy of predicting the stability index with an RMSECV of 2.16 nm (bathochromic shift) and an R2 of 0.98 for external validations. The results from these studies demonstrated the capability of Raman or SERS conjoined with TLC techniques and multivariant analyses for natural colorant adulteration and stability analysis. The Raman/SERS spectral data obtained in the present studies provide useful references for the food colorant research. The established methods could provide useful methodologies for industry applications in regard to fast raw ingredients screening for food companies as well as quality control for natural colorant manufacturers.
Research Projects
Organizational Units
Journal Issue
Publisher Version
Embedded videos