Origin of nitrogen in Earth's mantle constrained by models for partitioning and cycling

Abstract: We model nitrogen (N) partitioning in the magma ocean stage and cycling between the surface and mantle through Earth's history, and suggest that N in the present-day mantle may be set by subduction before the development of the modern N cycle.

Introduction: On present-day Earth, N cycling between the surface and mantle is largely controlled by biological N fixation and aerobic biological processing. Biological N fixation brings the majority of inorganic N into the modern N cycle. In the oceans, dissolved nitrate is the main form of nitrogen available for life, and dissimilatory denitrification leads to residual nitrate being kinetically enriched in ¹⁵N by ~6‰. The isotopically enriched nitrate is then reduced to ammonium and finally trapped in sediments (e.g., Stüeken al. 2016). 

Though secular subduction of ¹⁵N-rich sediments should cause ¹⁵N enrichment in the mantle, the mantle N sampled from mid-ocean ridge basalt (MORB) is rather depleted in  ¹⁵N (~-5‰), which is known as the N isotopic disequilibrium. Previous studies hypothesized that N in the mantle is a primordial component (Cartingy & Marty, 2013; Labidi et al. 2020). In the primordial origin scenario, the isotopic disequilibrium is attributed to atmospheric escape, which enriched the atmosphere with ¹⁵N. Another study proposed a recycling origin scenario in which the N isotopic composition of sediments has changed over time (Marty & Dauphas, 2003). Neither of these scenarios has been modeled quantitatively. Here we test the different scenarios by using numerical models coupled with N isotopes.

Methods: We developed two sets of models for the origin of observed mantle N isotopic composition: i) the primordial origin and ii) the recycling origin. The results are either accepted or rejected by the comparison to the amounts and isotopic compositions of N in contemporary surface reservoirs and mantle (Johnson & Goldblatt, 2015; Labidi et al. 2020).

For the primordial origin scenario, we calculate N partitioning between the atmosphere and mantle upon magma ocean solidification by using a melt-trapping model (Hier-Majumder & Hirschmann, 2017) and the partitioning coefficients between minerals, silicate melt, and the atmosphere (Li et al. 2013; Dalou et al. 2017). We consider the range of oxygen fugacity relevant to Earth's formation. We also estimate the ¹⁵N-enrichment effect due to EUV-driven escape (Watson et al. 1981) and solar-wind pick-up (Lichtenegger et al. 2010) to see how much atmospheric N should be removed to reproduce ~+5‰ difference between the atmosphere and mantle.

For the recycling origin scenario, we calculate secular N exchange between the surface reservoirs and mantle. Our model is based on that of Labidi et al. (2020). The isotopic fractionation between the atmosphere and subducting sediments is taken to be ~-9‰ and ~+6‰ before and after the Great Oxidation Event at 2.4 Ga, respectively, considering the change from abiotic fixation and anaerobic processing to biological fixation and aerobic processing. We fix the subduction and degassing fluxes on present-day Earth, and their power-law indices as a function of time as parameters. Bulk N partitioning and the isotopic difference between the reservoirs in the initial state are also treated as parameters.

Figure 1: Nitrogen partitioning between the atmosphere and mantle at the time of magma ocean solidification (PAN = present-day atmospheric nitrogen). The range of the modeled mantle N content reflects the uncertainty in the oxygen fugacity of the magma ocean.

Figure 2: Evolution of masses (left) and ¹⁵N/¹⁴N ratios (right) in the surface reservoirs (blue) and mantle (red). Curves show accepted models having different initial conditions and fluxes. Gray areas are the estimates for present-day surface reservoirs and mantle.

Results: Because N is relatively insoluble in silicate melts, it is mostly partitioned into the atmosphere even when trapping in interstitial melts is considered (Figure 1). Partitioning N into the mantle as much as present-day leads to 100 times excess in PAL N. We found that the excess amount of N can be removed neither by EUV- nor solar-wind-induced loss without excessive ¹⁵N enrichment in the atmosphere. Impact erosion by the late veneer bombardment removes atmospheric N without isotopic fractionation up to ~10 bar (Sakuraba et al. 2019), but it may not be sufficient to remove all excess N in the atmosphere.

Since the results of our partitioning model suggest that the primordial origin is unlikely, next we tested the recycling scenario in our N cycling model (Figure 2). In our successful runs, the mantle is initially depleted in N, and N in the present-day mantle is a result of higher net subduction flux on early Earth where sedimentary N is depleted in ¹⁵N due to abiotic N fixation and anaerobic N processing. The change to modern N cycle is visible in the kink in δ¹⁵N evolution, which may provide a way to test our model with evidence from geologic record.

Discussion: We suggest that N partitioning between the surface and deep Earth may be set by subduction driven by plate tectonics and partially by biology. This also suggests that the difference of atmospheric N contents between Venus and Earth, the former of which has three times more N in the atmosphere, is caused by their long-term evolution rather than early formation and differentiation processes.

Conclusions: We conclude that N in the present-day mantle may be set by subduction before the development of the modern N cycle. Further results for parameter survey and discussion on other geological and geochemical constraints will be shown in our presentation.