LAUSR

Chemical analysis using laser-induced breakdown spectroscopy (LIBS) is well suited for field applications and was applied here for shipboard characterization of a large sample set during the RR1310 rock dredging expedition to the Tuvalu Seamounts. Although recently the most common data treatment for LIBS has consisted of a partial least squares approach to define sample groupings, we show that quantitative data of useful quality can be obtained with a univariate approach. Here, our analysis goal was a quantitative comparison of the total alkali (Na2O + K2O) versus silica (SiO2) contents of 586 representative dredge samples with known ranges in common rock types. Out of those samples, >400 form a single large group of alkalic basalts with minor basanites/tephrites (SiO2: 43–48 wt%, Na2O + K2O: 3–5 wt%), similar to known shield-stage compositions of the Rurutu and Samoa hotspots in the sampling area. In contrast, several dredge hauls contain samples with compositions that do not overlap with the majority of samples. This includes three dredges performed on the northern boundary of the Lau Basin that contain similar SiO2 compositions, but slightly higher total alkali (Na2O+K2O) content. Despite this difference, they classify as basanite/tephrite, similar to a subset of the main group. More importantly, similar compositions were previously reported from the same tectonic boundary, ascribed to hotspot mantle source material mixed into the Lau Basin back-arc. Although the quality of the compositional data suffices to enable sample selection for time-intensive analyses, higher precision is required for more in-depth petrogenetic interpretation. Error analysis based on repeat standard measurements suggests averaging 100 spectra per sample is optimal here, while use of a higher resolution spectrometer, together with better laser control, would improve results and interpretations.