Near- infrared (NIR) spectroscopy for food quality applications
Abstract
Near-infrared (NIR) spectroscopy measures the wavelength and wavelength-dependent absorption of NIR light by a sample. The NIR region of the spectrum is defined as the wavelengths from 700 to 2500 nm (wavenumber 14300 to 4000 cm-1), between the visible light and the mid-infrared light region.1 The overall goal of this project was to evaluate the applicability of NIR spectroscopy for the prediction of such food quality characteristics as starch digestion rate, total moisture and total starch content in corn mutant lines, infant formula contamination with melamine and determination of total protein and total fructan content in green leaf tissue. It was concluded that NIR spectra of ground corn samples could be used only for rough screening of corn population for complex traits such as starch digestion rate, but for the determination of total starch and total moisture content in corn mutant samples the method is good for all analytical tasks. It was concluded that the NIR method enables rapid detection of 1 ppm melamine in infant formula powder. A partial least squares (PLS) model was established for correlating NIR spectra to melamine concentration in infant formula powder: R2>0.99, RMSECV≤0.616, and RPD ≥ 35. Factorization analysis of spectra was able to differentiate unadulterated infant formula powder from samples containing 1 ppm melamine with no misclassifications. NIR spectroscopy of leaf tissue was able to detect differences in fructan content and protein content with a high degree of probability. NIR detection of the ratio of fructan to protein was less precise. The robustness of the NIR models created in this research can be improved by using more samples over longer periods of time in the calibration equations.
Degree
M.S.
Advisors
Weil, Purdue University.
Subject Area
Food Science|Analytical chemistry|Biochemistry
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.