Illuminating Spectra - Part 4: Protein
Although a bit difficult to spot at lower concentrations, protein manifests itself as two distinctive bands at 2050 and 2180nm. These happen to be on either side of a starch band at 2100nm, resembling shoulders. The proximity of the protein and starch bands can often cause problems in the interpretation of starch content i.e. changes in protein content may be confused with changes in starch content.
The above example shows a series of spectra with varying protein content - ranging from 10 to 60% in an increasing order from bottom to top.
Starch content is consistent in all of these spectra (~12%) although it appears to be changing! In this case, low starch content can be wrongly interpreted as high starch if protein content is also low, and high protein content can mask low starch content.
So protein and starch need to be considered simultaneously for an accurate interpretation. This is one example why univariate (i.e. looking at one band only) analysis of NIR spectra can be misleading.
Note: To convert the wavelength units from nm to cm-1 please check out this webpage.
(Just be aware that this is a guideline that we use for solid agricultural samples. From time to time you might encounter samples that don’t react in the same way depending on specific sample composition.)