Chemometric Coffee

Skip to Navigation


  • Published: Sep 15, 2012
  • Author: David Bradley
  • Channels: Infrared Spectroscopy / Chemometrics & Informatics
thumbnail image: Chemometric Coffee
Coffee roaster 

An informatics approach to coffee that uses a least-squares analysis of near-infrared spectroscopic data could allow coffee roasters to treat their green beans perfectly every time.

João Rodrigo Santos and António Rangel of the Universidade Católica Portuguesa in Porto and Mafalda Sarraguça and João Lopes of the Universidade do Porto, Portugal, explain that one of the primary aims of coffee bean quality assessment is based on the relative numbers of defective beans among non-defective beans. If an automated process for characterising green beans before roasting can be used to reduce the ratio of bad to good beans, then the final roasted product can be improved and so attract a higher price for coffee roasters as well as improving their reputation for supplying a higher quality product.

The team has now investigated whether a methodology based on fast and low-cost near infrared spectroscopy (NIRS) might be suitable for such quality assessment of green coffee beans. NIRS requires no complex sample preparation with organic solvents and also offers short response times making it entirely suitable for online food analysis.

The applicability NIRS has now been tested on Arabica and Robusta varieties of unroasted coffee beans from various parts of the world (Robusta from Indonesia and Vietnam, Arabica from Colombia and Nicaragua).

NIR spectra were recorded with a Fourier-transform near infrared spectrometer with an InGaAs detector. Spectra were recorded in diffuse reflectance mode at 2 wavenumber resolution between 10000 and 4000 wavenumbers and averaged over 64 scans. The team then used partial least-squares regression to correlate the NIR spectra with the mass fraction of defective versus non-defective beans and obtained a relative error of just 5%. "This shows that NIRS can be a valuable analytical tool to be used by coffee roasters, enabling a simple and quantitative evaluation of green coffee quality in a fast way," the team says.

"Green coffee quality depends on the relative amount of defective beans and/or foreign bodies among non-defective beans," the researchers add. They point out that in this context, defects might include the presence of non-coffee bean debris, such as fragments of wood and stones, bean husks and hulls). It may also encompass diseased beans, black beans and sour beans, mouldy or insect-eaten beans. Given the variety of problems that can occur in any given batch of unroasted coffee beans, there is no standard way to test and sort a batch.

NIRS has previously been used in the food and drinks industry for various analyses in quality control. Indeed, earlier work has demonstrated its utility in obtaining a measure of water content in coffee beans, as well as being able to distinguish between different types of bean, and characterisation of beans before and after roasting, for roasting control itself. The Portuguese team has now extended the technique's repertoire using a statistical approach to NIR spectra that allows more effective quality control to be facilitated before roasting.

Flavour enhancer

There is no single standard for assessing beans. However, for export beans, the International Coffee Organization (ICO) asserts that there should be no more than 86 defects per 300 grams of Arabica beans or 150 defects/300 g for Robusta. The team concludes that chemometric modelling of NIR spectra of green coffee beans bodes well for developing a handheld system for objective coffee classification that is not as time consuming nor as subjective as a visual examination. The same approach might also be hooked to a conveyor sorting system to automate the removal of defective beans and others debris very efficiently prior to roasting.

The team has also now developed an application based on similar  method to ascertain the authenticity of coffee brands. They are currently implementing the methodology in collaboration with a coffee roaster and expect to deliver the application for commercial use in a couple of months involving a hand-held NIR instrument and specially designed software for control and data processing.

Related Links

Article by David Bradley

The views represented in this article are solely those of the author and do not necessarily represent those of John Wiley and Sons, Ltd.

Social Links

Share This Links

Bookmark and Share


Suppliers Selection
Societies Selection

Banner Ad

Click here to see
all job opportunities

Copyright Information

Interested in separation science? Visit our sister site

Copyright © 2018 John Wiley & Sons, Inc. All Rights Reserved