Adjustments to the Rock-Eval® thermal analysis for soil organic and inorganic carbon quantification

Abstract. Quantifying both soil organic carbon (SOC) and soil inorganic carbon (SIC) is essential to understand carbon (C) dynamics and to assess the atmospheric C sequestration potential in calcareous soils. The procedures usually used to quantify SOC and SIC involve pretreatments (decarbonation, carbonate removal) and calculations of the difference between C contents estimated by elemental analysis on raw and pretreated aliquots. These procedures lead to analytical bias associated with pretreatments, measurement deviations associated with sample heterogeneity, and cumulative errors associated with calculations. The Rock-Eval® analysis is a ramped thermal analysis that has been used in soil sciences since the 2000s, consisting of pyrolysis of the sample followed by oxidation of the residue. A single Rock-Eval® analysis on non-pretreated aliquots provides two parameters estimating the organic (TOC) and inorganic (MinC) C contents of the samples. Nevertheless, the Rock-Eval® protocol was standardised in the 1970s by IFP Energies Nouvelles for studying oil-bearing rocks and is thus not perfectly suited for soil study. Previous studies have suggested statistical corrections of the standard parameters to improve their estimations of C contents assessed by elemental analysis, but only a few of them have focused on the estimation of inorganic C content using the MinC parameter. Moreover, none of them have suggested adjustments to the standard Rock-Eval® protocol. This study proposes to adapt this protocol to optimise SOC and SIC quantifications in soil samples. Comparisons between SOC and SIC quantifications by elemental analysis and by Rock-Eval®, with and without statistical corrections of the standard TOC and MinC parameters, were carried out on 30 agricultural topsoils with a wide range of SOC and SIC contents. The results show that the standard Rock-Eval® protocol can properly estimate SOC contents once the TOC parameter is corrected. However, it cannot achieve a complete thermal breakdown of SIC amounts > 4 mg, leading to an underestimation of high SIC contents by the MinC parameter, even after correcting for this. Thus, the final oxidation isotherm is extended to 7 min to complete the thermal breakdown of SIC before the end of the analysis. This work is a methodological step to measure SOC and SIC contents in a single analytical run on a non-pretreated aliquot. More work is needed (i) on a wider range of soil samples with differing land use and other forms of carbonate mineral and sampling depths and (ii) to avoid the use of statistical corrections of the TOC and MinC parameters.