ABSTRACT The physicochemical and functional properties of wheat can be modified by exposing the whole grains to thermal pretreatment. Conventional analytical methods used to investigate such modifications are expensive, labour‐intensive and may be inaccurate due to interfering compounds. An economical alternative is to use Fourier transform near‐infrared (FT‐NIR) spectroscopy in combination with multivariate data analysis techniques. Analysis of variance simultaneous component analysis (ASCA) is an exploratory data analysis technique used to characterise the effects of experimental design factors on the chemical composition captured in the FT‐NIR spectral data. In this study, two hard wheat cultivars were exposed to 10 different temperatures by means of forced convection continuous tumble roasting. ASCA was applied to the standard normal variate preprocessed spectral data to evaluate the effects of cultivar, roasting temperature and their interaction. All three factors, cultivar, roasting temperature and their interaction, had a significant effect ( p 0.05) on the spectral data. Differences between the roasted wheat cultivars were associated with moisture, starch and aromatic compounds. The association with aromatic structures was supported by the differences in the phenolic contents of the two cultivars. Low roasting temperatures (108°C–150°C) were associated with starch and moisture changes particularly at approximately 1450, 1410 and 1940 nm. Water evaporated from the kernels, and the degree of starch polymerisation decreased. High roasting temperatures (170°C–232°C) were associated with starch and amino acids (ca. 2100 and 2294 nm), which likely underwent structural changes and participated in nonenzymatic browning reactions.
Niekerk et al. (Thu,) studied this question.