ABSTRACT Quantification of oxalate salts in soil clay minerals is necessary to study oxalate biogeochemistry, but existing analytical techniques are expensive and time-consuming. We aim to develop an efficient attenuated total reflectance mid-infrared (MIR) spectroscopic technique to quantify oxalate salts in a clay mineral matrix. We calibrated MIR models for analysis of oxalate anion concentrations in standard solutions (0–0.01 M) by using a partial least-squares regression algorithm. MIR models were also developed for analysis of sodium oxalate (NaOx) and calcium oxalate (CaOx) content in clay mineral mixtures with composition like soils of a semi-arid region and with contrasting concentrations (0–1.0 g g−1 for both oxalate salts) to test sensitivity of analyses. Validation plots (true vs predicted values) showed excellent model fit (R2 ≥ 0.96) and accuracy (normalized root mean squared error of prediction ≤ 0.06) for CaOx, NaOx and oxalic acid components. Once predictive models are stored in analytical software, MIR spectroscopic analyses of samples are much more efficient than chemical techniques. Our MIR spectral-based models are suitable for direct quantification of oxalate salts in clay mineral mixtures for samples like those used for model calibration.
Read full abstract