Quantitative estimation of polysaccharides in molasses using Near Infra Red spectroscopy

Non-invasive near infrared spectroscopy (NIR) in transmittance mode between 1100–2500 nm was employed for the quantitative estimation of dextran, starch, gum and pectin in final molasses one at a time. Partial least square regression analyses was used to develop a calibration model with 40 samples each in 0 th , 1 st and 2 nd derivatives which gave low SEC values. This method was applied for the estimation of unknown concentrations of the four polysaccharides in molasses, which gave low prediction error in the 0 th derivative explaining the accuracy of the results with correlation coefficient of 0.98. Stepwise multi linear regression has also been applied and wavelengths have been identified specific to the polysaccharides and a correlation has been tried on the basis of the structure of the polysaccharides. The above results lead to the most promising future use of NIR spectroscopy as on-line processing tool in sugar industry.