NIR spectroscopy application for determination caffeine content of Arabica green bean coffee

Abstract Caffeine is considered as an important quality indicators of coffee. Caffeine content in green coffee bean influence the flavour and price of coffee. Commonly, percentage of caffeine content is determined by chemical method, which is known time consuming, destructive and expensive, so it is not suitable for a real time coffee content prediction system. The objective of this study was to develop a PLS model based on NIR-Spectroscopy to predict caffeine content of Indonesia coffee beans nondestructively. The wavelength used in this study was 1000-2500 nm and the caffeine content of samples were determined by LCMS method. Several data pretreatments such as multiple scatter correction (MSC), first derivative (dg1), combination of dg1+MSC, with PLS factors variation were applied to obtain the best prediction of caffeine content by NIR. The best prediction was obtained using MSC and 6 PLS factors indicated by a high coefficient correlation value which was >0.9. NIR method can be used for predicting the caffeine content of green bean coffee nondestructively.