Skip to main content
ARS Home » Research » Publications at this Location » Publication #114728
Title: MID-INFRARED DIFFUSE REFLECTANCE SPECTROSCOPY FOR THE QUANTITATIVE ANALYSISOF AGRICULTURAL SOILS.

Author
item Reeves Iii, James
item McCarty, Gregory
item REEVES, V - CVM, FDA

Submitted to: Journal of Agricultural and Food Chemistry
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/4/2000
Publication Date: 2/25/2001
Citation: N/A

Interpretive Summary: The interaction of light with various materials (spectroscopy) has come to be widely used to determine the composition of a variety of materials ranging from animal feeds to foods. Infrared light is beyond the range of human sight and is what we commonly perceive as heat. Soil samples were analyzed conventionally and by mid-infrared spectroscopy for total carbon, total nitrogen, pH and measures of biological activity. Ground, but otherwise untreated samples (n = 180), from experimental farm plots (two locations, three replicate plots, under plow and no- till practices, three rates of nitrogen fertilizer and from five depths), were scanned from using a custom made linear sample transport (50 by 2 mm sample area scanned). Results showed that accurate determination of a number of compositional parameters: total carbon, total nitrogen, pH and many measures of biological activity could be achieved. In general, the results achieved using mid-infrared spectra were at least as accurate as those found previously using near-infrared spectra (another type of light), and were sometimes significantly better, i.e., pH.

Technical Abstract: Soil samples were analyzed conventionally and by mid-infrared diffuse reflectance spectroscopy for total C, total N, pH and measures of biological activity. Ground, non-KBr diluted, samples (n = 180) from experimental plots (two locations, three replicate plots, under plow and no-till practices, three rates of N fertilizer and from five depths), were scanned from 4000 to 400 cm-1 ( 4-cm-1 resolution, 64 co-added scans) on a DigiLab FTS-60 Fourier transform spectrometer using a custom made linear sample transport (50 by 2 mm sample area scanned). Results using partial least squares regression showed that accurate calibrations can be developed for the determination of a number of compositional parameters: total C, total N, pH and many measures of biological activity. In general, the results achieved using mid-infrared spectra were at least as accurate as those found previously using near-infrared spectra, and were sometimes significantly better, i.e., pH.