|DHAKAL, SAGAR - Us Forest Service (FS)|
|Chao, Kuanglin - Kevin Chao|
|Qin, Jianwei - Tony Qin|
Submitted to: Applied Sciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/20/2017
Publication Date: 1/25/2017
Citation: Dhakal, S., Chao, K., Qin, J., Kim, M.S., Chan, D.E. 2017. Identification and evaluation of composition in food powder using point-scan Raman spectral imaging. Applied Sciences. 7(1):1.
Interpretive Summary: This study reports a method based on Raman spectral imaging for identification and quantification of multiple components present in a complex food powder mixture, demonstrated using vanillin, melamine, and sugar mixed together into powdered non-dairy creamer at ten sample concentrations between 1% and 10% (w/w). Spectral data was extracted from Raman spectral images of the sample mixtures. Analysis by self-modeling mixture analysis (SMA) and spectral information divergence values was able to precisely extract identifiable pure spectra of vanillin, melamine, and sugar from the mixed-component sample spectra, and to create detection images of the sample mixtures in which detected pixels of the three components were visually color-coded to show particle distribution. The results showed effective detection of the vanillin, melamine, and sugar components, and accurate quantification of their concentrations in the mixture samples. This Raman-based method has great potential for application to food safety problems requiring authentication of powdered food ingredients, evaluation of component concentrations, or detection of contaminants, for a variety of products ranging from processed foods, health supplements, and animal feed. Given the widespread distribution of many powdered ingredients through food processing supply lines nationally and worldwide, this research will benefit food processors and food safety regulators seeking to ensure safety and quality of ingredients ultimately consumed by the public.
Technical Abstract: This study used Raman spectral imaging coupled with self-modeling mixture analysis (SMA) for identification of three components mixed into a complex food powder mixture. Vanillin, melamine, and sugar were mixed together at 10 different concentration levels (spanning 1% to 10%, w/w) into powdered non-dairy creamer. SMA was used to decompose the complex multi-component spectra and extract the pure component spectra and corresponding contribution images. Spectral information divergence (SID) values were calculated between the extracted pure component spectra and reference component spectra to identify the components corresponding to the extracted spectra. The contribution images obtained via SMA were used to create Raman chemical images of the mixture samples, to which threshold values were applied to obtain binary detection images of the components at all concentration levels. The detected numbers of pixels of each component in the binary images was found to be strongly correlated with the actual sample concentrations (correlation coefficient of 0.99 for all components). The results show that this method can be used for simultaneous identification of multiple individual components and estimation of their concentrations for authentication or quantitative inspection purposes.