The Quantification of Adulteration in Arabica Coffee using UV-Visible Spectroscopy in Combination with Two Different PLS Regressions

Diding Suhandy, Meinilwita Yulia

Abstract


Arabica coffee is being considered to be of better quality than robusta because of their superior taste and aroma. Adulteration of arabica with other cheaper coffee like robusta coffee has become a great problem related to authentication of food products. Ground coffee samples are most challenging to be discriminated each other and visual inspection by our eyes or even machine vision method becomes very difficult. For this reason, in this research we propose a new analytical method based on UV-visible spectroscopy for quantification of adulteration in arabica-robusta coffee blend. The proposed method is easy to use, low cost with affordable spectrometer and safe for environment with free-chemical analysis. A number of 100 samples was used as samples with different degree of adulteration (10-60% of robusta concentration in arabica-robusta coffee blend). Spectral data of aqueous samples was obtained using a UV-visible spectroscopy in the range 200-400 nm. The result shows that calibration model using selected subintervals (iPLS model) can improve the prediction performance of calibration model using full spectrum (FS-PLS). It is noted that iPLS model removed wavelengths that is not related to the quantification of adulteration in arabica-robusta coffee blend. This study has shown a potential application of using UV-visible spectroscopy for simple and low cost tool to detect the authentication of arabica coffees


Keywords


UV-visible spectroscopy, Chemometrics, Authentication, Adulteration, PLS regression

Full Text:

PDF


DOI: https://doi.org/10.13170/aijst.6.2.8457

Refbacks

  • There are currently no refbacks.


______________________________________________________________________________________________________________

This work  is licensed under a Creative Commons Attribution-Non Commercial 4.0 International License (CC BY-NC 4.0).