Pelagic barite precipitation at micromolar ambient sulfate

Geochemical analyses of sedimentary barites (barium sulfates) in the geological record have yielded fundamental insights into the chemistry of the Archean environment and evolutionary origin of microbial metabolisms. However, the question of how barites were able to precipitate from a contemporary ocean that contained only trace amounts of sulfate remains controversial. Here we report dissolved and particulate multi-element and barium-isotopic data from Lake Superior that evidence pelagic barite precipitation at micromolar ambient sulfate. These pelagic barites likely precipitate within particle-associated microenvironments supplied with additional barium and sulfate ions derived from heterotrophic remineralization of organic matter. If active during the Archean, pelagic precipitation and subsequent sedimentation may account for the genesis of enigmatic barite deposits. Indeed, barium-isotopic analyses of barites from the Paleoarchean Dresser Formation are consistent with a pelagic mechanism of precipitation, which altogether offers a new paradigm for interpreting the temporal occurrence of barites in the geological record.

). Sulphur could not be analyzed in these samples owing to the type of filter used. That Ba and Fe are not co-located supports the notion that the bright grains are not of terrigenous minerals lithogenic origin but rather barites. Note that the topography of the filtered particulate matter and the oblique orientation of the EDS detector may lead to a slight mismatch in co-location between BSE micrographs and EDS maps.

Supplementary Note 1
As noted in the main text, the particulate Ba excess, pBa XS , was calculated as the difference between the 'supported' Ba (lithogenic-and organic matter-associated Ba) and measured (in situ) Ba concentration: The following subsections deal with the considerations regarding the most robust proxy element for lithogenic material, then with the derivation of appropriate normalizing ratios for Ba : Fe and Ba : P (Eq. 1). In the discussion of ratios, we make two further assumptions that could affect the magnitude of pBa XS . Firstly, we assume that the Ba do not account for multiple particle types nor for processes that could alter Ba : Fe or Ba : P after a particle has formed, such as different lithogenic sources, chemical scavenging, or selective remineralization (e.g., 4,5 ). However, these assumptions appear to be of only secondary importance and serve only to alter the magnitude, but not the depth range or very existence of, the pBa XS in Lake Superior.
Based on the full dataset of lithogenic-type tracers ( Supplementary Fig. 2), potential caveats, and analytical considerations, we suggest that Fe is the best elemental proxy for lithogenic material in Lake Superior. . That the Ba : Fe of particulates in Lake Superior is slightly lower than the crustal average is not surprising since the watershed surrounding Lake Superior contains numerous banded iron formations (e.g., 8,9 ), which possess significant proportions of Fe and thus may contribute to lower-than-average Ba : Fe ratios in fluvially-derived particles entering Lake Superior. The lowest observed Ba : Fe ratio during our sampling was 3.79 ± 0.22 mM : M (±2 SD), measured within the core of the benthic nepheloid layer ≈ 180 m at St. WM (Supplementary Fig. 1). Though this benthic bottom layer is predominantly composed of resuspended terrigenous material (e.g., 10 ), 13.1 nM of pP were also measured in this sample ( Supplementary Fig. 3), which likely contribute some organic matter-associated Ba to the in situ total pBa. Thus, the 'true' Ba : Fe ratio of lithogenic material must be < 3. and for cultured plankton (≈ 0.4 mM : M; 11 ). However, it is worth noting caveats specific to each of these estimates of Ba : P. Firstly, in situ bulk particles likely overestimate organismal Ba : P as these samples invariably contain lithogenic material and/or barite, which both contribute significant quantities of Ba (but generally not P; 2 ). Indeed, there does not appear to be any indication of a significant biologically-associated component to particulate Ba cycling along the NAZT. 4 Secondly, more recent culturing experiments employing chelating agents to buffer free metal concentrations to marine-relevant levels yielded significantly lower organic-associated Ba : P of of 0.5 µM : M. 12 Thus, it highly likely that true organismal Ba : P is 1 mM : M. In the absence of culture data for freshwater organisms we assigned organismal Ba : P based on our own observations from Lake Superior. Using analogous reasoning as for assigning Ba : Fe of lithogenic material, the 'true' value for Ba : P must be less than or equal to the lowest observed value.
The lowest particulate Ba : P observed here was 0.44 mM : M., which was obtained for a zooplankton sample col-

Supplementary Note 2
Here we present representative BSE (backscatter electron) micrographs and EDS (energy-dispersive X-ray spectrography) maps of filters that evidence the presence of barite in Lake Superior particulate matter. Filters from BH15-11 were necessarily dissolved for the purposes of obtaining multi-element geochemical and Ba-isotopic data. Thus, the filters imaged in Supplementary Figs Fig. 2). It thus follows that Fe and Ti should co-occur in certain grains of lithogenic origin, whereas authigenic barites should exhibit Ba but not Fe (or Ti). Though this test is less assured than if measuring S directlyand relies on Fe and Ti co-occurring in the same minerals-we nonetheless identified numerous grains that were either Ba (or Ti)-rich or Fe-rich at CD1 (Supplementary Fig. 6). When considered in the context of the µm-size range and 'bright' characteristics of the grains when imaged using BSE, these additional tests are entirely consistent with the presence of barite in these samples.
Lastly, we note that these images are of secondary significance to the other geochemical data that evidences barite precipitation, since the identification of barite in Lake Superior particulate matter does not address the manner of precipitation. That is, the distinction between in situ (i.e., pelagically precipitated) versus ex situ barites (e.g., eroded from the surrounding watershed and carried into the lake via fluvial processes) cannot be made by BSE micrographs or EDS maps: the geochemical tests prescribed in the main text are the primary means by which the origin of the pBa XS can be discerned.