Skip to main content
ARS Home » Research » Publications at this Location » Publication #123470
Title: EFFECTS OF INSTRUMENT VARIATIONS, AS REFLECTED BY THE FREQUENCY OF OBTAINING BACKGROUND SPECTRA, AND INSTRUMENT FUNCTIONS SUCH AS TYPE OF APODIZATION ON MID-INFRARED CALIBRATIONS FOR FEED COMPOSITION

Author
item REEVES III, JAMES
item REEVES, VALERIE - FDA

Submitted to: Journal of Applied Spectroscopy
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/2/2002
Publication Date: 9/5/2002
Citation: N/A

Interpretive Summary: Spectroscopy uses the interaction of light with materials to determine their composition, i.e., mid-infrared spectroscopy uses light beyond the range of human site. Past research demonstrated that mid-infrared spectroscopy can be used to accurately and rapidly determine the composition of materials ranging from forages to soils. However, there are many parameters which are adjustable when performing such determinations which have not been studied. These include how often a background or blank spectrum is obtained for comparing to spectra of samples, and many functions related to the mathematical processing of the data. Studies were conducted using 173 chemically treated forage samples and a DigiLab FTS-60 mid-infrared spectrometer. Results demonstrated that variations in the mathematical functions (zero filling and apodization) used to generate the spectrum (graphical representation of how the mid-infrared light interacts with a sample)had little or no effect on determining sample composition, although Kubelka-Munk transformed data did not perform well as absorbance data. Likewise, the frequency of taking a background spectrum did not seem to have any great effect on calibrations, although results were better with hourly or daily acquisitions.

Technical Abstract: Research has demonstrated that mid-IR spectroscopy can, like NIRS, be used to quantitatively determine the composition of ground samples of forages and soils without the need for KBr dilution. However, the influence of factors such as apodization function has not been determined. Results based on the spectra of 173 treated forage samples obtained using a DigiLab FTS-60 spectrometer have demonstrated that many parameters associated with mid-IR spectra have little or no effect on PLS calibrations. Additional zero filling of spectra had little effect other than to increase the derivative gaps found to produce optimal calibrations, but calibrations developed using Kubelka-Munk transformed data were not as accurate as absorbance based calibrations. Choice of apodization function also had little effect, although slightly better results were found using triangular or weak Norton-Beers. Likewise, the frequency of taking a background spectrum did not seem to have any great effect on calibrations, although results were better with hourly or daily acquisitions as opposed to one for each sample as is done in the NIR.