SIMS microanalysis of the Strelley Pool Formation cherts and the implications for the secular-temporal oxygen-isotope trend of cherts

Abstract

The significance of oxygen isotope ratios in Archean chert has long been debated. Cherts from the c. 3.4 Ga Strelley Pool Formation (SPF) (Pilbara Craton, Western Australia) host some of the oldest stromatolite and microfossil evidence for life, but the genesis and timing of silica cements has been unclear. Field relations, petrography and a combination of laser fluorination and in-situ SIMS measurements of δ18O in quartz show that bedded cherts of the SPF were originally precipitated as carbonates and were later widely replaced by quartz. Three localities were studied and analyzed for δ18O(Qz) in chert: 1) Camel Creek: foliated, metamorphosed, bedded cherts and meter-scale black chert veins; 2) Unconformity Ridge and ABDP8 drill core: stromatolitic and bedded chert overlying basal detrital quartz sandstone; and 3) the Trendall locality: “bedded” stromatolitic chert replacing original dolomite, low temperature hydrothermal quartz, and mm- to decimeter-scale chert-quartz veins. Laser fluorination (mm-scale) values of δ18O(Qz) range from: 14.2 to 18.2‰ VSMOW at Camel Creek; 9.3 to 18.9‰ at Unconformity Ridge; and 13.7 to 25.7‰ at Trendall. Values of δ18O(Qz) in cm to decimeter-scale hydrothermal chert veins cutting bedded carbonates at Trendall range from ca. 15 to 16‰, whereas “bedded cherts” are 17 to 26‰. These laser data include the highest δ18O values reported for cherts in Paleoarchean sediments and are up to 4‰ higher than the upper limit of ∼22‰ reported in other studies, in apparent contrast to the long-standing secular-temporal trend which shows such high δ18O only in younger chert. However, analysis by laser fluorination at the 1-mm scale cannot resolve microtextures seen petrographically. In contrast, in-situ SIMS analyses can resolve petrographic microtextures and show δ18O(Qz) at 10-μm scale have an even greater range of 7–31‰ in “bedded” cherts at the Trendall locality, up to 9‰ above the secular-temporal trend.Textures observed optically at the Trendall locality were classified as: microquartz, mesoquartz, chalcedony, megaquartz veins, and cavity megaquartz. SEM-CL imaging shows two generations of meso- and megaquartz; bright CL with well-developed growth zoning, and dark CL with massive or mottled texture. Microquartz is the earliest textural generation of quartz and has a maximum δ18O(Qz) of ∼22‰ by SIMS. Dark-CL mesoquartz has similar δ18O to microquartz and is interpreted to also be early. Bright-zoned-CL mesoquartz, which formed post-Archean, has even higher δ18O, up to 29‰. Vein megaquartz crosscuts most quartz generations and has a restricted range of δ18O, mostly from 16 to 19‰. Chalcedony pseudomorphs rhombic cavities and fractures, lines the edges of veins, and has similar δ18O to veins (16–19‰). Late cavity megaquartz is bright and zoned by CL, grows into late open cavities, and has the highest δ18O(Qz) values reported from the Pilbara, up to 31.3‰. Thus, the highest-δ18O quartz cements at the Trendall locality are the youngest and may be related to weathering. Early silicification and the formation of microquartz, chalcedony and low δ18O mesoquartz occurred during low temperature hydrothermal activity in the Archean.None of the SPF quartz examined is interpreted to have formed as a direct precipitate from Paleoarchean seawater. Thus, values of δ18O(Qz) do not record either water chemistry or temperature of Archean oceans. In-situ SIMS analysis shows that high-δ18O(Qz) values above 22‰ are only found in late-forming cavity megaquartz and bright-zoned-CL mesoquartz at the Trendall locality.The SPF results from our sample suite demonstrate the ability to resolve complex history using detailed petrography and SIMS analysis. Similar studies may show equal complexity of δ18O(Qz) data for other localities that are interpreted to show secular-temporal trends for chert. The apparent increase of δ18O(Qz) through time may reflect differences in diagenesis, and/or an inherent and previously unrecognized sampling bias that compares fundamentally different populations of quartz, such as Archean hydrothermal chert from volcanic greenstone belts, with unrelated Phanerozoic biogenic quartz.