Unexpected Findings Change the Picture of Sulfur on the Early Earth

Scientists believe that until about 2.4 billion years ago there was little oxygen in the atmosphere—an idea that has important implications for the evolution of life on Earth. Evidence in support of this hypothesis comes from studies of sulfur isotopes preserved in the rock record. But the sulfur isotope story has been uncertain because of the lack of key information that has now been provided by a new analytical technique developed by a team of Caltech geologists and geochemists. The story that new information reveals, however, is not what most scientists had expected.

"Our new technique is 1,000 times more sensitive for making sulfur isotope measurements," says Jess Adkins, professor of geochemistry and global environmental science at Caltech. "We used it to make measurements of sulfate groups dissolved in carbonate minerals deposited in the ocean more than 2.4 billion years ago, and those measurements show that we have been thinking about this part of the sulfur cycle and sulfur isotopes incorrectly."

The team describes their results in the November 7 issue of the journal Science. The lead author on the paper is Guillaume Paris, an assistant research scientist at Caltech.

Nearly 15 years ago, a team of geochemists led by researchers at UC San Diego discovered there was something peculiar about the sulfur isotope content of rocks from the Archean era, an interval that lasted from 3.8 billion to about 2.4 billion years ago. In those ancient rocks, the geologists were analyzing the abundances of stable isotopes of sulfur.

When sulfur is involved in a reaction—such as microbial sulfate reduction, a way for microbes to eat organic compounds in the absence of oxygen—its isotopes are usually fractionated, or separated, from one another in proportion to their differences in mass. That is, ³⁴S gets fractionated from ³²S about twice as much as ³³S gets fractionated from ³²S. This process is called mass-dependent fractionation, and, scientists have found that it dominates in virtually all sulfur processes operating on Earth's surface for the last 2.4 billion years.

However, in older rocks from the Archean era (i.e., older than 2.4 billion years), the relative abundances of sulfur isotopes do not follow the same mass-related pattern, but instead show relative enrichments or deficiencies of ³³S relative to ³⁴S. They are said to be the product of mass-independent fractionation (MIF).

The widely accepted explanation for the occurrence of MIF is as follows. Billions of years ago, volcanism was extremely active on Earth, and all those volcanoes spewed sulfur dioxide high into the atmosphere. At that time, oxygen existed at very low levels in the atmosphere, and therefore ozone, which is produced when ultraviolet radiation strikes oxygen, was also lacking. Today, ozone prevents ultraviolet light from reaching sulfur dioxide with the energy needed to fractionate sulfur, but on the early Earth, that was not the case, and MIF is the result. Researchers have been able to reproduce this effect in the lab by shining lasers onto sulfur dioxide and producing MIF.

Geologists have also measured the sulfur isotopic composition of sedimentary rocks dating to the Archean era, and found that sulfides—sulfur-bearing compounds such as pyrite (FeS₂)—include more ³³S than would be expected based on normal mass-dependent processes. But if those minerals are enriched in ³³S, other minerals must be correspondingly lacking in the isotope. According to the leading hypothesis, those ³³S-deficient minerals should be sulfates—oxidized sulfur-bearing compounds—that were deposited in the Archean ocean.

"That idea was put forward on the basis of experiment. To test the hypothesis, you'd need to check the isotope ratios in sulfate salts (minerals such as gypsum), but those don't really exist in the Archean rock record since there was very little oxygen around," explains Woody Fischer, professor of geobiology at Caltech and a coauthor on the new paper. "But there are trace amounts of sulfate that got trapped in carbonate minerals in seawater."

However, because those sulfates are present in such small amounts, no one has been able to measure well their isotopic composition. But using a device known as a multicollector inductively-coupled mass spectrometer to precisely measure multiple sulfur isotopes, Adkins and his colleague Alex Sessions, a professor of geobiology, developed a method that is sensitive enough to measure the isotopic composition of about 10 nanomoles of sulfate in just a few tens of milligrams of carbonate material.

The authors used the method to measure the sulfate content of carbonates from an ancient carbonate platform preserved in present-day South Africa, an ancient version of the depositional environments found in the Bahamas today. Analyzing the samples, which spanned 70 million years and a variety of marine environments, the researchers found exactly the opposite of what had been predicted: the sulfates were actually enriched by ³³S rather than lacking in it.

"Now, finally, we're looking at this sulfur cycle and the sulfur isotopes correctly," Adkins says.

What does this mean for the atmospheric conditions of the early Earth? "Our findings underscore that the oxygen concentrations in the early atmosphere could have been incredibly low," Fischer says.

Knowledge of sulfate isotopes changes how we understand the role of biology in the sulfur cycle, he adds. Indeed, the fact that the sulfates from this time period have the same isotopic composition as sulfide minerals suggests that the sulfides may be the product of microbial processes that reduced seawater sulfate to sulfide (which later precipitated in sediments in the form of pyrite). Previously, scientists thought that all of the isotope fractionation could be explained by inorganic processes alone.

In a second paper also in the November 7 issue of Science, Paris, Adkins, Sessions, and colleagues from a number of institutions around the world report on related work in which they measured the sulfates in Indonesia's Lake Matano, a low-sulfate analog of the Archean ocean.

At about 100 meters depth, the bacterial communities in Lake Matano begin consuming sulfate rather than oxygen, as do most microbial communities, yielding sulfide. The researchers measured the sulfur isotopes within the sulfates and sulfides in the lake water and sediments and found that despite the low concentrations of sulfate, a lot of mass-dependent fractionation was taking place. The researchers used the data to build a model of the lake's sulfur cycle that could produce the measured fractionation, and when they applied their model to constrain the range of concentrations of sulfate in the Archean ocean, they found that the concentration was likely less than 2.5 micromolar, 10,000 times lower than the modern ocean.

"At such low concentration, all the isotopic variability starts to fit," says Adkins. "With these two papers, we were able to come at the same problem in two ways—by measuring the rocks dating from the Archean and by looking at a model system today that doesn't have much sulfate—and they point toward the same answer: the sulfate concentration was very low in the Archean ocean."

Samuel M. Webb of the Stanford Synchrotron Radiation Lightsource is also an author on the paper, "Neoarchean carbonate-associated sulfate records positive Δ³³S anomalies." The work was supported by funding from the National Science Foundation's Division of Earth Sciences, the Henry and Camille Dreyfus Foundation's Postdoctoral Program in Environmental Chemistry, and the David and Lucile Packard Foundation.

Paris is also a co-lead author on the second paper, "Sulfate was a trace constituent of Archean seawater." Additional authors on that paper are Sean Crowe and CarriAyne Jones of the University of British Columbia and the University of Southern Denmark; Sergei Katsev of the University of Minnesota Duluth; Sang-Tae Kim of McMaster University; Aubrey Zerkle of the University of St. Andrews; Sulung Nomosatryo of the Indonesian Institute of Sciences; David Fowle of the University of Kansas; James Farquhar of the University of Maryland, College Park; and Donald Canfield of the University of Southern Denmark. Funding was provided by an Agouron Institute Geobiology Fellowship and a Natural Sciences and Engineering Research Council of Canada Postdoctoral Fellowship, as well as by the Danish National Research Foundation and the European Research Council.