Quantification of carbon derived from lignite in soils using mid-infrared spectroscopy and partial least squares

The total organic carbon (TOC) of many recultivated mine soils is composed of a fraction that is lignite-derived as well as a fraction that is derived From recent plant litter. In these soils, precise quantification of the li gnite contribution to the TOC content can only be achieved with expensive a nd time consuming methods. In the present study, we tested diffuse reflecta nce infrared Fourier transform (DRIFT) spectroscopy in combination with mul tivariate data analysis partial least squares (PLS)I as a rapid and inexpen sive means of quantifying the lignite contribution to the TOC content of so il samples. The conceptual approach included analysis of samples with diffe rent lignite content (bulk soil and particle size fractions) by DRIFT-spect roscopy and C-14 activity measurements. Afterwards, with both data sets a c alibration curve was established by PLS and the lignite content predicted f rom the DRIFT spectra. A good fit was obtained between this approach and th e radiocarbon analysis. Loading factors showed that this prediction was bas ed on structural differences between the two organic matter types. We concl ude that DRIFT spectroscopy can be used in combination with multivariate da ta analysis for the differentiation of carbon derived from lignite and carb on derived from recent organic matter in soils. (C) 2001 Elsevier Science L td. All rights reserved.