Map: Research / Early Earth / Sulfur cycling	< Back		Next >

3.3 Sulfur cycling on the early Earth (Farquhar, Fitz-Gibbon, Lyons, Marcus, McKeegan, Mojzsis, Rubin, Runnegar, Sander)

	Figure 3.3. The Archean sulfur cycle (Farquhar et al. 2002). The principal products of atmospheric chemistry were ultimately pyrite (positive 33S) and sulfate (negative 33S) as shown in Fig. 3.3.1. The atmospheric photochemistry (box) is one target of the research proposed here.

The recent discovery by Farquhar et al. (2000) and Bao et al. (2000) that mass-independent isotope effects (§3.3.1) may survive in terrestrial geological materials has invigorated sulfur and oxygen isotope geochemistry. As these effects are thought to be solely due to gas-phase reactions (Gao and Marcus 2001), their geological record provides an unprecedented source for atmospheric history. Most obviously, the anoxic atmosphere of the early Earth (older than ~2.2 Gyr) promoted UV-induced sulfur photochemistry whereas the subsequent oxygen-rich atmosphere was, and is, one in which ozone is a principal photochemical product of UV radiation (Lyons 2001).

The "Great Oxidation Event" (Holland, 2002), which took place about 2.2 Gyr ago, is believed to represent the transition between these two states and is recorded by the sudden disappearance of mass-independently fractionated (MIF) sulfur compounds (principally BaSO₄ and FeS₂) from the geological record. In contrast, the preservation of the MIF sulfur signal in rocks older than ~2.2 Gyr results from the fact that there were at least two main products of the atmospheric reactions, elemental sulfur particles and sulfate aerosols, each having a different MIF signature (Figure 3.3). As the elemental sulfur particles were insoluble in surface waters, these two components did not mix, and their positive and negative MIF signatures were transmitted to minerals formed from them in sedimentary environments.

3.3.1 Ion microprobe measurements of the four stable isotopes of sulfur

Figure 3.3.1. Sulfur isotopic compositions of 3.5 Gyr-old sulfides and sulfates, North Pole area, Western Australia. Pyrite in organic-rich cherts are enriched in 33S relative to normal terrestrial materials (33S > 0), indicating their orign from the reduction of elemental sulfur (Fig. 3.3). Pyrites that are intimately associated with bedded and cross-cutting sulfates (barites) are depleted in 33S (33S < 0), as are the sulfates. The 15‰ difference in 34S between the barite and the barite-associated pyrite has been regarded as evidence for bacterial sulfate reduction (Shen et al. 2001). However, in the expected setting for sulfate-reducing bacteria (black cherts and shales) the pyrite has a 33S signature which cannot have come from sulfate.

Our role in the development of this fruitful area of research has been the following:

(1) The demonstration that the effect is real, and not an artefact of sample preparation as had been argued by Ohmoto (2000), by making ³³S/³²S, ³⁴S/³²S, and ³⁶S/³²S measurements in a completely different way using the UCLA Cameca ims 1270 ion microprobe in monocollector and multicollector modes (Runnegar et al. 2002)

(2) The development of new sulfide (FeS, FeS₂, CuFeS₂) and sulfate mineral (BaSO₄) standards for ion microprobe analysis (Greenwood et al. 2000; Runnegar et al., in prep.)

(3) Evidence that early Archean sulfate minerals traditionally regarded as barite (BaSO₄) replacements of primary gypsum (CaSO₄.2H₂O) evaporites were, in fact, originally hydrothermal barite; this bears on the composition of the Archean ocean (Runnegar et al., 2001b, in preparation)

(4) The recognition of three different sulfur reservoirs in Archean sedimentary rocks based on their three and four sulfur isotope compositions (Figure 3.3.1), results that argue against widespread microbial sulfate reduction during the Archean (Runnegar et al. 2001a, in preparation) in contrast to work based on only the two most common isotopes (Shen et al. 2001)

(5) The use of MIF sulfur signatures as tracers of base metal sources and hydrothermal fluid circulation in Archean ore deposits (Runnegar et al. 2002)

(6) A survey of Precambrian and Phanerozoic sulfates and sulfides of many different ages and environments to provide a background for the interpretation of MIF sulfur signatures throughout Earth history (Runnegar et al., in preparation)

(7) The development of a quantitative understanding through numerical modeling, of the role that ozone plays in acquiring and transmitting a substantial MIF signature to oxygen-bearing compounds at all levels in the modern atmosphere (Lyons 2001); and (8) Application of the photochemical understanding gained in (7) to the important problem of self-shielding and the production of MIF oxygen in the solar nebula.

Given these discoveries, and the excitement of complementary studies by our NAI colleagues (Hiroshi Ohmoto, Pennsylvania State University; Douglas Rumble, Carnegie Institution of Washington; Andrey Bekker, Harvard University), we propose to proceed over the next five years on the following fronts:

(1) Explore sulfur cycling during and following the Archean by investigating key sedimentary environments in greater detail than has been done in our current survey that involved approximately 300 multiple sulfur isotope measurements on ~50 samples ranging in age from 3.85 Gyr to the Miocene. We are particularly interested in syngenetic pyrite in organic-rich shales and banded iron formation, in sulfates and sulfides associated with volcanic-hosted massive sulfide (VHMS) deposits and ancient "back smokers", and sulfate evaporites. Each of these environments, if tracked through time, has great potential for the preservation of sulfur isotope biosignatures.

(2) Investigate the disappearance of MIF sulfur from the sedimentary cycle at or after the Great Oxidation Event (GOE). Although this problem is being explored effectively by others in a primary stratigraphic context, there is the need to consider other avenues of investigation that involve the recycling of previously MIF sulfur compounds. These include ore deposits formed by meteoric hydrothermal actvity (Runnegar et al. 2002) and material recycled mechanically from older terrains. A combination of geochemical exploration and numerical modeling will be used to investigate the history of sulfur cycling on a global scale during the GOE transition.

(3) We have preliminary data, that needs to be confirmed and supports an earlier report (Rees and Thode 1977), of a small MIF effect in sulfide-bearing phase in the Allende carbonaceous chondrite. Rubin and McKeegan propose to explore this anomaly using the numerous Allende and other carbonaceous chondrite samples in the UCLA meteorite collection maintained by Wasson.

3.3.2 Laboratory experiments and theoretical analysis of the kinetics and photochemistry of mass-independent sulfur isotope effects

The chemical origin of the MIF in sulfur is still unclear. Farquhar et al. (2001) have shown in laboratory experiments that photodissociation of H₂S, SO₂, and SO₂/CO₂/H₂O mixtures all produce elemental sulfur with a wavelength-dependent MIF signature. However, the experimental results that best reproduce MIF in sulfur-bearing rocks older than ~2.2 Gyr involve photodissociation at single wavelengths < 200 nm. In a real atmosphere, MIF would not be produced at a single wavelength, since solar radiation is broadband. Instead, we believe that MIF in sulfur arises, at least in part, from the reaction S + S₂ S₃.

It has been clearly shown (Mauersberger et al. 1999) that MIF in O₃ is produced during the ozone formation reaction, O + O₂ O₃. Because O and S are isovalent atoms, the ozone formation reaction and the S₃ (thiozone) formation reactions are also isovalent. It is therefore likely that formation of S₃ isotopomers occurs in a mass-independent manner, just as is true for the isotopomers of O₃. We propose experiments to test this hypothesis and to determine the rate coefficients for formation of several isotopomers of S₃.

In the first experiment, we intend to pass an electrical discharge through flasks of H₂S and SO₂. Elemental sulfur has been observed previously in such experiments. We will collect elemental sulfur and SO₂, and measure 32S, 33S, 34S, and 36S relative to the initial gas isotope values. The measurements will be done in collaboration with Farquhar. Although very simple, the discharge experiments are an essential first step. If MIF is observed in the collected SO₂ and elemental sulfur residue, then we have demonstrated that an inherently broadband process (discharge) can produce MIF of sulfur. Conversely, if no MIF is observed, then it is likely that photodissociation alone (and not subsequent reactions of photoproducts) is the principal cause of the MIF.

Assuming that MIF is observed in the sulfur residue collected in the discharge experiments, then flow tube experiments will be undertaken to determine the isotope-specific rate coefficients for the formation of several S₃ isotopomers, for example:

³²S + ³⁴S³⁴S ³²S³⁴S³⁴S
³²S + ³²S³²S ³³S³²S³²S

We shall also determine the rate coefficient for sulfur atom exchange with S2,

³⁴S + ³²S³²S ³²S + 32S³⁴S

It may also be necessary to measure isotope-specific rate coefficients for the S4 formation reaction, S₂ + S₂ S₄, and for the SO dimer reaction, SO + SO (SO)₂, since these are important pathways to elemental sulfur and sulfate formation (Kasting et al. 1989).

A flow tube will be used because the variable-length reactant port allows for the determination of accurate rate coefficients. Additional ports are for reactant S₂ and the He flow gas. Atomic sulfur will be generated by microwave discharge of isotopically-pure carbonyl sulfide (OCS). Diatomic sulfur will be generated by heating isotopically-pure solid sulfur. The flow tube will be quartz (~1 meter in length), and will be heated to > 150 °C to prevent condensation of S₂ and S₄. Neutral products will be converted to positive ions by electron impact ionization and will be detected with an Extrel MAX 500 quadrupole mass spectrometer purchased for this purpose. Flow tube measurements will be done in collaboration with Sander.

Lyons will investigate the kinetic theory of the thiozone formation reaction in collaboration with Marcus. This will be an extension of the non-RRKM (i.e., non-statistical) unimolecular dissociation theory developed for the O + O₂ O₃ reaction by Marcus and his colleagues (Gao and Marcus 2001). The sulfur kinetic theory will provide an important link to the much better understood chemistry of gas-phase oxygen isotopes, and may be applicable to higher-order polysulfur compounds such as S₄ and S₈. With knowledge of the isotope-specific rate coefficients for thiozone isotopomers, full photochemical models of Archean sulfur chemistry can be developed.

3.3.3 Genomic approach to understanding of the evolution of sulfur metabolisms

Easily cultured organisms are estimated to constitute a small fraction of all microbial species and are rarely numerically dominant in the communities from which they were obtained (Hugenholtz 2002). Thus, much remains to be learned about the distribution and diversity of microorganisms and their geobiological activities. This can be overcome by purifying total DNA from environmental samples, constructing random DNA fragment libraries, and then using random sequencing or probing for genes of interest. Fitz-Gibbon and colleagues will use these exploratory methods to search for novel genes and novel genetic settings for sulfur metabolism.

Genes involved in dissimilatory sulfate reduction (ATP sulfurylase, APS reductase and sulfite reductase) are known from a wide range of Bacteria. Phylogenetic trees based on these genes tend to follow conventional taxonomic groupings (Wagner et al. 1998; Friedrich 2002). On the other hand, few of these genes have been identified from members of the Archaea, hampering efforts to determine whether the ability to reduce sulfate was a property of the last common ancestor of Bacteria and Archaea. Dissimilatory sulfate reduction is well studied in the euryarcheal genus Archaeoglobus but may be more widespread within the Archaea. For example, growth with sulfate as an electron acceptor has been reported for Caldivirgas (Itoh et al. 1999) and possibly also Thermocladium (Itoh et al. 1998), both members of the Crenarchaeota. Moreover, a set of sulfate reductase genes was found in the genome of the crenarchaeote Pyrobaculum aerophilum (Fitz-Gibbon et al. 2002), although frameshift mutations have made some of the genes non-functional.

The identification of a small number of novel archaeal sulfate reduction genes would be sufficient to determine whether they will produce phylogenetic trees follow the 16S rRNA pattern. If so, and if the genes were broadly distributed across the Archaea, it would be likely that sulfate reduction had been present in the common ancestor of Bacteria and Archaea. Conversely, the novel archaeal genes may show a clear pattern of more recent horizontal transfer from the Bacteria, and thus the possibility that sulfate reduction postdates the Archaea-Bacteria divergence. Sulfur metabolism genes tend to cluster in many genomes (Fitz-Gibbon et al. 2002). For example in Pyrobaculum a 26.5 Kb-long section of the genome carries approximately 35 genes, eleven of which are clearly involved in sulfur metabolism. Some of the other 24 genes appear to be in operons that also contain the sulfur metabolism genes, suggesting that they may also be involved in sulfur biochemistry. Thus, the identification of one sulfur metabolism gene by random sequencing can be followed by further sequencing along the clone (BAC clones are ~250 Kb in length) in the expectation of discovering additional genes for sulfur metabolism.

< Previous Section	Next Section >