The Fifth Element in Three Dimensions: Boron Systematics Along and Across the Aleutian Volcanic Arc

This study uses abundances and isotope ratios of the fluid-mobile element boron (B) to better understand the role of fluids in volcanism at the Aleutian arc in Alaska. In volcanic rocks from the Aleutian volcanic front, patterns in B abundance and isotope ratios (δ11B) point to three endmembers: a sediment melt, a slab interior melt, and a slab-top melt. Most lavas from the eastern Aleutian Islands have δ11B of -2 to +5‰, similar to Aleutian sediment. Hence, sediment is likely to be the dominant B source in the Aleutian Island samples. In lavas from the western Aleutian seafloor, high δ11B (0 to +13.4‰) is negatively correlated with B concentrations (2-9 ppm). These samples have geochemical characteristics of partial melts of the subducting slab in eclogite facies. Such heavy B is likely to be derived from a fluid released by serpentinite in the mantle section of the subducting slab, which triggers melting of minimally altered rock in the slab interior. The existence of the third endmember (slab-top melt) is inferred from a subset of Aleutian Island lavas with relatively low concentrations of isotopically light B (< 11 ppm, δ11B< -2‰). These samples may point to partial melts of altered oceanic crust near the surface of the slab, which has lost primarily 11B through dehydration reactions as the slab heats up during subduction. Because this slab-top melt endmember is poorly defined by data from Aleutian island lavas, samples dredged from the Aleutian rear-arc were analyzed to better constrain this source component. In the rear-arc, the slab is deeper and hotter, making the slab surface more likely to melt. In rear-arc lavas, B enrichments (e.g. B/Ta) decrease with increasing slab depth, but B abundances and isotope ratios are not clearly linked to slab depth and are generally indistinguishable from typical Aleutian island lavas. Sr-Nd-Pb isotopes in rear-arc basalts with K2O > 1 wt% indicate a reduced role for sediment compared to typical volcanic front lavas. This implies an increased relative contribution from slab-top melts, which is not evident in B/δ11B data for rear-arc lavas. However, addition of only 0.5% of bulk sediment (10-20% of the total subducted component) to the melt source is sufficient to maintain sediment-like δ11B. More work is necessary to better define the low-light B endmember. Finally, the model proposed to explain Aleutian data patterns is compared to a global dataset for arc lavas. Global data patterns do not show a clear link between slab temperatures and δ11B-B systematics, though these links have been observed at individual arcs. However, two isotopic trends emerge in the global dataset which appear to be dominated by either sediment-derived or serpentinite-derived boron. The serpentinite-dominated group shows a positive correlation between δ11B and B abundance, in contrast to the negative correlation observed in seafloor lavas from the western Aleutians. Areas where lavas may have similar B-δ11B characteristics to western Aleutian seafloor lavas are suggested as sampling locations for future B work.