Purpose: The cost-efficient methods of analysis, such as rapid short-wave infrared (SWIR) spectral analysis, have been applied for the efficient exploration of critical raw materials (CRM), including mineral components and rare earth elements (REE) from the deep-ocean sediments. Methods: Gravity cored sediment samples were collected during an oceanographic mission to the Trans-Atlantic Geotraverse (TAG) hydrothermal field of the Mid-Atlantic Ridge (MAR, 26° N). SWIR reflectance spectra (dependent variable) of samples were mathematically tested against referent geochemical data (independent variable), obtained by conventional analysis (ICP/OES, ICP/MS), after applied full cross-validation multivariate partial least square regression (CVPLSR). Value of parameter-residual predictive deviation (RPD) was used for evaluation of CVPLSR modeling: RPD > 2.5 (satisfactory calibration model for the screening purposes) and RPD > 5.0 (model adequate for the quality control of the studied elements). Results: The CVPLSR modeling provided significant results for the determination of several mineral components: major elements (Fe and Si) had the values of RPD equal to 3.65 and 2.84, respectively, which indicated a viable potential for their routine analysis, whereas RPD for Ca was equal to 5.51, thus assuring its quality control by SWIR analysis, in sediment samples of the studied location. Among the REE, Ce (RPD = 2.55) and Er (RPD = 2.59) yielded the most satisfactory results. Conclusions: The findings highlight the benefit of rapidly obtained empirical SWIR-reflectance data, which can be used for near-real-time exploration of geochemical deposits hosted in deep-ocean sediments.
- Cross-validation partial least square regression (CVPLSR)
- Deep-ocean sediments
- Rare earth element (REE) geochemistry
- Residual predictive deviation (RPD)
- Short-wave infrared (SWIR) spectra
- TAG hydrothermal field