We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Near Infrared Spectrometry of Clinically Significant Fatty Acids Using Multicomponent Regression.
- Authors
Kalinin, A.; Krasheninnikov, V.; Sviridov, A.; Titov, V.
- Abstract
We have developed methods for determining the content of clinically important fatty acids (FAs), primarily saturated palmitic acid, monounsaturated oleic acid, and the sum of polyenoic fatty acids (eicosapentaenoic + docosahexaenoic), in oily media (food products and supplements, fish oils) using different types of near infrared (NIR) spectrometers: Fourier-transform, linear photodiode array, and Raman. Based on a calibration method (regression) by means of projections to latent structures, using standard samples of oil and fat mixtures, we have confirmed the feasibility of reliable and selective quantitative analysis of the above-indicated fatty acids. As a result of comparing the calibration models for Fourier-transform spectrometers in different parts of the NIR range (based on different overtones and combinations of fatty acid absorption), we have provided a basis for selection of the spectral range for a portable linear InGaAs-photodiode array spectrometer. In testing the calibrations of a linear InGaAs-photodiode array spectrometer which is a prototype for a portable instrument, for palmitic and oleic acids and also the sum of the polyenoic fatty acids we have achieved a multiple correlation coefficient of 0.89, 0.85, and 0.96 and a standard error of 0.53%, 1.43%, and 0.39% respectively. We have confirmed the feasibility of using Raman spectra to determine the content of the above-indicated fatty acids in media where water is present.
- Subjects
CARBON compounds spectra; FATTY acids; NEAR infrared spectroscopy; INDIUM gallium arsenide; PALMITIC acid; PHOTODIODES; REGRESSION analysis
- Publication
Journal of Applied Spectroscopy, 2016, Vol 83, Issue 5, p811
- ISSN
0021-9037
- Publication type
Article
- DOI
10.1007/s10812-016-0368-0