The origin of nitrogen in Earth's mantle: Constraints from basalts 15N/14N and N2/3He ratios

Plate tectonics is thought to be a major driver of volatile redistribution on Earth. The budget of nitrogen in Earth's mantle has been suggested to be almost entirely surface-derived. Recycling would contribute nitrogen with relatively heavy N-15/N-14 isotope ratios to Earth's mantle. This could explain why the Earth's mantle N-15/N-14 isotope ratio is substantially higher than both solar gases and chondritic parent bodies akin to enstatite chondrites. Here, published nitrogen isotope data of mid-ocean ridge and ocean island basalts are compiled and used to evaluate the nitrogen subduction hypothesis. Nitrogen isotope ratios are used in conjunction with published N-2/He-3 and K2O/TiO2 ratios on the same basalts. Assuming that He-3 is not recycled, N-2/He-3 ratios are argued to trace nitrogen addition to mantle sources via subduction. Various mantle source enrichments for basalts are tracked with K2O/TiO2 ratios: elevated K2O/TiO2 ratios are assumed to primarily reflect the contributions of recycled components in the basalts mantle sources. The main result of our data compilation is that for most basalts, delta N-15 and N-2/He-3 remain constant across a vast range of K2O/TiO2 ratios. Mid-ocean ridge basalts have delta N-15 signatures that are lower than air by similar to 4 parts per thousand and an average N-2/He-3 ratio of 3.7 (+/- 1.2) x10(6) (95% confidence, n = 30). Published delta N-15 and N-2/He-3 are invariant across K2O/TiO2 ratios that vary over a factor of similar to 20. Using estimates of slab K2O/TiO2 and [TiO2], the observed invariant delta N-15 and N-2/He-3 may be fit with slabs containing similar to 0.1 ppm N. A mass balance shows that adding similar to 10% recycled slabs to the convective mantle only raises the N-2/He-3 by <5%. Lavas from Iceland, Galapagos and Hawaii have high N-2/He-3 and N-15/N-14 ratios relative to the convective mantle. Only seven samples show nitrogen isotopic signatures that are unaffected by air contamination, although those samples are poorly characterized for N-2/He-3. The seven basalts show delta N-15 between -2 and 0 parts per thousand that do not vary systematically with K2O/TiO2 ratios that vary over a factor of similar to 5. The N-2/He-3 ratios of these seven basalts is unknown, but the high N-2/He-3 mantle may be estimated by combining published N-2/Ar-36 to He-3/Ar-36 ratios. This yields a N-2/He-3 of 2.3 (+/- 1.2) x 10(6) (1 sigma uncertainty). This is indistinguishable from the MORB estimate of 3.7 (+/- 1.2) x 10(6). Invariant delta N-15 across variable degrees of mantle enrichments and MORB-like N-2/He-3 for the high N-2/He-3 mantle are not consistent with nitrogen addition to plume sources with elevated N-2/He-3 ratios. Lavas from the Society plume with low He-3/He-4 ratios show an enriched mantle source, and they have elevated delta N-15 >= +0.5 parts per thousand and N-2/He-3 > 10(7). For those, the addition of slabs with concentrations of similar to 0.1 ppm N to a mantle source can account for the integrated dataset. To summarize, the published data suggest that nitrogen subduction may explain a sub-set of published N isotope data on basalts, but that N recycling has an overall more limited impact on mantle nitrogen than previously thought.