Speciation of Ca-bearing minerals and its relation to MAOM in semi-arid agricultural soils: a combined XRPD - XANES approach

Calcium (Ca) is increasingly recognized as a key factor in promoting carbon persistence in semi-arid soils, which are typically characterized by low soil organic carbon content, alkaline pH, and Ca abundance. However, direct evidence or a clear mechanistic understanding of Ca contribution to carbon persistence remains limited, due partly to the lack of established techniques for determining Ca species in soils, especially Ca-bearing minerals. We therefore aimed to investigate the relationship between Ca species and mineral-associated organic matter (MAOM) fraction (>1.7 g cm^-3, <53 µm) in semi-arid agricultural soils of India using two non-destructive methods.

Specifically, we used X-ray powder diffraction (XRPD) and Ca K-edge X-ray absorption near-edge structure (XANES) spectroscopy to characterize Ca species across six surface soils from agricultural fields in south India, exhibiting a wide range of soil Ca content (6.0 - 79.2 g kg^-1) and MAOM-C content (1.5 - 7.4 g kg^-1) with alkaline pH (8.0 – 8.9). Ca-bearing mineral concentrations, namely Ca-plagioclase, carbonates (calcite and dolomite), and clay minerals (smectite and vermiculite), were determined by XRPD. The contributions of Ca-plagioclase and carbonates to soil Ca were calculated by multiplying their concentrations by Ca contents estimated from the ideal mineral formula. Exchangeable Ca was operationally estimated by assuming that it is dominantly associated with the interlayer sites of smectite and vermiculite. Furthermore, the Ca species were determined by XANES using linear combination fitting with standards, namely Ca-plagioclase (anorthite), calcite, and exchangeable Ca (montmorillonite). The Ca content of each species was calculated by multiplying the Ca ratio from XANES fitting and the soil Ca content measured by X-ray fluorescence (XRF).

The sum of the estimated Ca content from XRPD was highly correlated with the soil Ca content by XRF, with high accuracy (y = 0.81x, p<0.001, r = 0.98). Each Ca ratio estimated from XRPD showed a similar trend to that obtained from XANES, with one exception. Given that the XRPD fitting can generally utilize a large library of standards with distinct diffraction patterns, the differences between XRPD and XANES may reflect the representativeness of the selected XANES standards for Ca species in the soils. These results suggest that the non-destructive XRPD approach is effective for quantifying Ca-bearing minerals in semi-arid soils and, when combined with XANES, it offers complementary information on Ca speciation. MAOM-C was positively correlated with exchangeable Ca (p<0.05, r = 0.90) and carbonates (p<0.05, r = 0.84) according to XRPD-based Ca estimates. On the other hand, XANES-based Ca analysis showed that MAOM-C was positively correlated with calcite (p<0.05, r = 0.82), and weakly, though not significantly, correlated with exchangeable Ca (p=0.12, r = 0.70), indicating that XRPD-based Ca showed stronger relationships with MAOM-C. There was no significant relationship between MAOM-C and Ca-plagioclase in both XRPD- and XANES- Ca estimates. Our results suggest that both exchangeable Ca and carbonates might be more strongly associated with MAOM-C accumulation compared with Ca-plagioclase in the semi-arid agricultural soils. The current approach of distinguishing soil Ca species would help to elucidate the mechanisms underlying Ca-MAOM associations in soils.