High-resolution line-scan analysis of resin-embedded sediments using laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS)

Laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) line-scanning is a promising technique for producing high-resolution (µm-scale) geochemical records on resin-embedded sediments. However, this approach has not yet been thoroughly tested on sediment samples of known elemental compositions. Here, we address this through the analysis of resin-embedded quartz, calcite, and clay (montmorillonite) sediments spiked with Al, V, Mo, and Ba across a range of concentrations. LA-ICP-MS spot analyses on these samples were compared to data independently obtained by conventional techniques: solution nebulization-ICP-optical emission spectroscopy (SN-ICP-OES), SN-ICP-MS, and X-Ray Fluorescence on fused glass beads. Data were reported as (log-)ratios of the analyte elements normalized to Ca, Si, and Al for calcite, quartz and clay respectively to correct for varying ablation yields. Our LA-ICP-MS data demonstrate close agreement to within 6% of the reference values determined by the conventional techniques, with high correlation coefficients (R2 ˜ 1.0) across the full range of concentrations, and high precision (< 3% relative standard deviations after three repeated analyses). Barium was hosted in both aluminosilicates and carbonates in the clay matrix, giving different yields (elemental fractionation), and leading to variable accuracy (mean deviation of 15%). A selection of the spiked sediments was used to produce artificial laminated sediments to further test the effects of line scanning on LA-ICP-MS accuracy and precision in resin-embedded samples. It appears that LA-ICP-MS line-scan analyses have good accuracy (deviations from the reference values generally < 5%) and high precision (relative standard deviations of repeated analyses on the same sample generally < 5%) for our target elements. Moreover, the LA-ICP-MS line-scanning closely records the alternating geochemical profiles of our artificial laminations. However, calcium showed a clear tailing signal at the transition from a calcite to a quartz layer, indicating that geochemical signals can be smeared at transitions to sediment layers devoid of analyte element(s). We also analyzed resin-embedded natural sediments for comparison of LA-ICP-MS line-scan data to parallel sub-samples measured by SN-ICP-OES and SN-ICP-MS. In general, the results from LA-ICP-MS line-scans of natural sediments correspond well with reference values and with high reproducibility, corroborating the results of the artificial laminations. The results for Ca, V, Mn, Sr, Mo, Ba, and U (in log-ratios to Al) suggest that these elements can be analyzed semi-quantitatively, showing correlation coefficients (R2) ranging from 0.59 to 0.85 compared to conventional analytical techniques. However, we show that using non-matrix matched calibration standards such as NIST SRM 610 may induce deviations from the LA-ICP-MS values to the reference values. Consequently, we recommend a simple protocol for further data correction in which binned-mean LA-ICP-MS values are calibrated to a parallel series of discrete samples analyzed by these conventional techniques