Experimental constraints on reconstruction of Archean seawater Ni isotopic composition from banded iron formations

The Ni isotopic systematics in banded iron formations (BIFs) potentially recorded the Ni isotopic composition of ancient seawater over Precambrian geological history. However, the utility of BIFs as proxies requires quantitative knowledge of how Ni isotopes fractionated as dissolved Ni was initially incorporated into iron-rich sediments and how diagenesis may have affected the Ni isotopic systematics. Here we report results of synthesis experiments to investigate the behavior of Ni isotopes during Ni coprecipitation with ferrihydrite and then transformation of ferrihydrite to hematite. Ferrihydrite coprecipitation experiments at neutral pH demonstrated that the dissolved Ni was variably heavier than coprecipitated Ni (likely a mixture of surface-adsorbed and structurally incorporated Ni), with the isotope fractionation becoming larger as the fraction of Ni associated with solid increased (Δ60/58Nisolution-solid = +0.08 to +0.50‰). Further experiments at lower pH (3.7-6.7), in which structurally incorporated Ni likely dominated in solids, documented a decrease in Δ60/58Nisolution-solid from +0.44‰ to −0.18‰ as the pH decreased. The negative value for Δ60/58Nisolution-solid at low pH indicates the enrichment of heavier isotopes in incorporated Ni relative to dissolved and adsorbed Ni, possibly as a result of the presence of a small amount of tetrahedral Ni2+ in addition to octahedral Ni2+ in the ferrihydrite structure. The results of the ferrihydrite experiments thus reflect equilibrium isotope fractionation between three pools of Ni, with δ60/58Ni values in the order of incorporated > dissolved > adsorbed. Hematite was synthesized by transformation of Ni-bearing ferrihydrite in aqueous solution at ∼100 °C. A significant amount of Ni (up to 60%) was released (desorbed) from solids into solutions as pH dropped from ∼7 to 4.5-5.5 upon phase transformation. Rinsing of the synthesized hematite in 2 M acetic acid released only very small amounts of Ni (<4% of total Ni, presumably surface-adsorbed) that were isotopically heavier (δ60/58Ni = +0.11 ± 0.06‰) than the residues (presumably dominated by incorporated Ni), which had δ60/58Ni of −0.26 ± 0.07‰. The preference of lighter isotopes for the incorporated Ni relative to the surface-adsorbed Ni after phase transformation (most had been released into solution) is probably due to distortion of NiO octahedra in the hematite structure, with weaker NiO bond strengths on average. Hence, the more variable Δ60/58Nisolution-solid values (−0.04 to +0.77‰) observed in hematite experiments most likely reflect thermodynamically driven Rayleigh fractionation, with incorporated Ni unavailable to exchange with dissolved Ni due to continuous reduction in size of the highly reactive surface pool of Ni, through which all solid-solution exchange must occur. Overall, the synthesized hematite was isotopically lighter than the ferrihydrite by ∼0.08‰ in δ60/58Ni, which is however within the current analytical uncertainties (±0.09‰). This implies that earliest diagenesis of BIFs results in very limited change in the isotopic composition of solid-associated Ni. Our experimental results, although conducted in a very simple system that differs from Archean seawater, represent an important step toward reconstruction of the Ni isotopic composition of ancient seawater from Ni isotopic signatures in BIFs.