Justin Hayles, Laurence Y Yeung, Martin Homann, Asmita Banerjee , Hehe Jiang, Bing Shen, Cin-Ty A. Lee

The 18O/16O ratios of ancient marine minerals show a puzzling increase over geologic time. Long-term changes in temperature, seawater 18O/16O ratios, and post-depositional overprinting can all explain this trend, but few tracers can distinguish between these scenarios. Here, we report high-precision 18O/16O and 17O/16O ratios of cherts through 3.4 Ga of Earth′s history. We find that Phanerozoic cherts are consistent with having formed in porewaters that are isotopically indistinguishable from modern (ice-free) seawater. In contrast, Precambrian cherts require either formation in waters isotopically distinct from Phanerozoic seawater, or a different mode of formation. If the early diagenetic formation pathway of Precambrian cherts resembles that of Phanerozoic cherts, and the Precambrian cherts are unaltered, then the results would imply that the oxygen-isotope composition of seawater has evolved on billion-year timescales before reaching its present composition by the Ordovician. Under this interpretation it is estimated that seawater had δ′18O < -11‰ at 3.41 Ga, with surface temperatures < 34 C. Although this scenario provides the simplest explanation for the observed 18O/16O trend of marine minerals, other scenarios which do not require a secular change in seawater 18O/16O cannot be ruled out.