Rapid quantification of oil content in castor seeds using FT-NIR spectroscopy

Abstract

Fourier Transform Near-Infrared Reflectance Spectroscopy (FT-NIR) explored as a technique to predict oil content in castor seeds (Ricinus communis L.). The castor seeds were taken from Agricultural Research Station, Agriculture University Jodhpur, India during 2023-2024 with the objective to determine the oil content by rapid method. Castor seeds analysed using both reference methods, NMR and FT-NIR spectroscope with an integrating sphere in reflectance mode, covering the wavelength range of 10,000-4,000 cm⁻¹ (1000-2500 nm). Partial least squares (PLS) regression used to develop calibration models for the examined components, and further models validated through full cross-validation. The coefficients of determination (r²) for intact seeds of castor found 0.9972 with standard error of performance (SEP) 0.0848. Pairwise concentration of actual versus calculated oil % in total samples used in calibration and validation plotted and results showed that the correlation coefficient of both the content was noted 1.00. The results of present study indicates that FT-NIRS calibrations are effective for estimating seed quality parameters, making it a highly desirable technique for quick and non-destructive analysis of oil content in castor breeding programs and further may be utilized in other oil seed crops.