Combination of high-resolution laser-induced breakdown spectroscopy and least square method for reducing soil carbon overestimation due to iron interference

Abstract

Laser-induced breakdown spectroscopy (LIBS) has been used for soil carbon (C) estimation by detecting the 247.8561 nm line of atomic carbon (C I). However, the Fe II line at 247.8572 nm interferes with the LIBS C measurement, resulting in C content overestimation. Due to the resolution limitation of common LIBS instruments, Fe II could not be identified and eliminated spectroscopically, and in natural soil the conventional method confirming constant emission intensity ratio of two Fe emission lines (Fe I line at 247.8572 nm and another Fe line which the upper energy level close or equal to the former) could not be used directly. Thus, in order to obtain the C content in soil, a high-resolution spectrometer (0.012 nm resolution) was employed to identify the C and Fe emission lines near 247.8561 nm in the standard reference soils 2709a (San Joaquin Soil) and 2711a (Montana II soil). For interference correction the emission line of Fe at 249.7819 nm was selected due to its constant ratio with delay time, and the least square method was employed to obtain the β coefficient of Fe 247.8572 nm emission line and the α coefficient of C 247.8561 nm emission line. The results demonstrated that suitable emission lines near 247.8561 nm could be identified using the high-resolution spectrometer, and the Fe interference could be reduced, even if the pure Fe 247.8572 nm intensity was unknown for the soil sample. The coefficient of α = 1600 and β = 3.02 were calculated from the case study and then the C content was corrected. With the Fe interference, the percentage of overestimated C was in the range of 15% to 58% in the used soil samples. This method minimizes C overestimation, and is less affected by soil heterogeneity and matrix effects, demonstrating its potential use in soil C quantification.