Scattered colorimetry is an innovative spectroscopic technique that provides integrated and global measurement of the color and turbidity of liquids. The appearance of a liquid is determined objectively by means of a measuring instrument in such a way that it does away with subjectivity.
Once appearance is measured, similar liquids may be mapped, i.e., they may be classified according to like features, and compared.
This technique applies to any other liquid moderately colored and moderately turbid.
Scattered colorimetry provides multi-wavelength and multi-angle absorption spectroscopy in the visible spectral range. The spectral data are processed by means of multivariate analysis to create a 2D or 3D map. The map is populated with points, each of which represents a liquid with its individual and global characteristics of color and turbidity. Similar liquids are mapped as clusters so that samples can be correctly assigned according to class.
Probes for scattered colorimetry: with (right) and without (left) optical fibers
SCATTERED COLORIMETRY AND MULTIVARIATE DATA PROCESSING: HOW THEY WORK
The instrumentation for scattered colorimetry is an optoelectronic device for measuring the absorption spectra of samples at different angles. It consists of four white-light LEDs spanning the 450-630 nm spectral range, with a miniaturized optical fiber spectrometer serving as a detector. The sources, which could either be fitted to the probe or guided by optical fibers, are located at 0°, 30°, 60°, and 90° in relation to the detector.
The LEDs were switched on sequentially to measure the transmitted and scattered spectra: The transmitted spectrum yields information regarding color which is also affected by turbidity; the scattered spectra yields information on turbidity which is also affected by color.
Given the spectrometer’s spectral resolution, the sample is characterized by means of 184 spectral values coming from the four absorption spectra, each of which consisted of 46 wavelengths. Multivariate data processing is used to achieve a reduction in data dimensionality to better extract the significant information for sample identification.
The spectral data are processed by principal component analysis (PCA) or linear discriminant analysis (LDA) that provides the coordinates for identifying the samples on 2D or 3D maps.
Several successful applications of scattered colorimetry and multivariate data analysis have been demonstrated, the most relevant of which are:
· mapping of extra virgin olive oils according to geographic area of origin;
· mapping of frying oils according to degree of degradation and substance being fried;
· mapping of surfactants according to type and mixture;
· Mapping of beers according to type;
· Mapping of lubricant oils according to types and the degree of degradation.
Collections of extra virgin olive oil (EVOO)
samples and their classification according to
the geographic area of origin:
Right: crop 2003/2004
Bottom: crop 2004/2005
Belgian beers and their
classification according to type
· Sens. Act. B, vol. 90(1-3), 2003, pp. 157-162 (olive oil)
· Sens. Act. B, vol. 111-112, 2005, pp. 363-369 (olive oil)
· SPIE vol. 6189, 2006, pp. 61892E-1/E-4 (beer)
· OSA ISBN 1-55752-B16-0, paper WB3 (lubricant oil)
Multi-angle and multi-wavelength absorption spectroscopy in the VIS