Diffuse reflection collection optics for analysing chicory content in coffee

Coffee is one of the most traded agricultural commodities in the world. Coffee consumption increases every year in India, along with coffee powder adulteration. In some cultures, chicory is added to enhance taste, as per some customers’ requirements. However, even in those cases, the percentage of chicory added is often falsely reported as lower than the actual value. Diffuse reflection near-infraredd spectroscopy gives information about the physical and chemical properties of the sample. In order to do this, the weak diffuse reflection from a powder sample must be efficiently collected using an optical system. Such a system for collecting diffuse reflection was designed in Zemax with off-the-shelf components. The system was then assembled and coupled to the fiber input of a NIR spectrometer (1350-2200 nm). Coffee samples from a well-known Indian brand were used for testing by mixing them with various quantities of chicory. Diffuse reflection spectra recorded from the designed system were preprocessed using Multiplicative scatter correction to eliminate noise. Multivariate calibration methods like Principal Component Analysis were employed to differentiate the coffee and coffee-chicory mixtures. Partial Least Squares Regression was used to develop a model to quantify the chicory content. After applying multiplicative Scatter Correction, the calibrated model achieved an R² value of 0.96. Results showed that the current model can predict the amount of chicory with an accuracy of 5% in the range of up to 20% chicory or above 35% chicory.