Oxygen
Page 7
Although this is, perhaps, the most widely accepted model of banded iron formation, there are still some difficulties with it. The oldest iron formations, 3.8 billion years old, were surely formed before oxygen began to accumulate. Furthermore, most of the ironstones around the world do not consist of simple iron oxides such as haematite, as might be expected if oxygen levels were genuinely high and the reactions were no more than bucket chemistry. There are other biological mechanisms that can oxidize iron without any requirement for free oxygen. One was described in 1993
by Friedrich Widdel and his colleagues at the Max Planck Institute for Marine Microbiology in Bremen. They isolated a strain of purple bacteria from lakeside sediments which could use the energy from sunlight to produce iron ores without requiring free oxygen. The main product of the bacterial reaction is a brownish rust-like deposit, ferric hydroxide, which is commonly found in banded iron formations. Widdel argued that the same seasonal upwellings that brought nutrients and iron to the sunny surface waters could have stimulated great bursts of iron-ore formation by purple bacteria. Thus, while the presence of cyanobacteria and rusting iron in banded iron formations suggests that free oxygen may have played a role in their genesis, Widdel and his colleagues have shown that some iron formations could have been formed by purple bacteria in the absence of oxygen. Despite their promise, then, banded iron formations cannot give us a quantitative estimate of oxygen levels in the air during this period.
One possible solution to the problem of exactly when oxygen levels rose has been put forward by Donald Canfield of the University of Southern Denmark, a leading authority on Precambrian oxygen levels, in a series of papers published in Science and Nature. Canfield turned, rather elliptically at first sight, to the oxygen-hating sulphate-reducing bacteria that produce hydrogen sulphide under stagnant conditions, to estimate the timing of the increase in atmospheric oxygen. His rationale was founded on two observations.
42 • SILENCE OF THE AEONS
First, sulphate-reducing bacteria gain their energy from a reaction in which hydrogen reduces sulphate to produce hydrogen sulphide.
Although sulphate (SO 2–
4
) is found at high levels in modern sea water (at
about 2.5 grams per litre) it should not have been plentiful in the early Precambrian period, as its formation requires the presence of oxygen. This premise is supported by the absence of sulphate evaporites, such as gypsum, from the early Earth. If sulphate can only form in the presence of oxygen, then the sulphate-reducing bacteria could not have established themselves until there was some oxygen in the atmosphere. We can go further: because low sulphate is a rate-limiting factor for sulphate-reducing bacteria, virtually precluding their growth in freshwater lakes, their activity depends on the concentration of sulphate. This in turn depends on the concentration of oxygen. Put another way, even though sulphate-reducing bacteria are strictly anaerobic — they are actually killed by oxygen — they cannot exist in a world without oxygen, and their activity is ultimately governed by oxygen availability.
The second observation applied by Canfield relates to sulphur isotopes. Just as photosynthesis leaves a carbon signature in the rocks, the sulphate-reducing bacteria similarly discriminate between the two stable isotopes of sulphur, sulphur-32 and sulphur-34. As with carbon isotopes, the lighter sulphur-32 atoms have a slightly greater vibrational energy, and so their reactions are catalyzed more quickly by the action of enzymes. Sulphate-reducing bacteria therefore produce hydrogen sulphide gas enriched in sulphur-32, leaving more sulphur-34 behind in the oceans. In some conditions, both the hydrogen sulphide and sulphate can precipitate from the oceans to form rocks. Sulphur signatures can be read in these rocks. In particular, and perhaps surprisingly for those who still do not associate minerals with life, hydrogen sulphide reacts with dissolved iron to form iron pyrites, which then sinks to the bottom sediments. Iron pyrites can be formed by either volcanoes or bacteria. Against the consistent, unadulterated ratio of sulphur isotopes from volcanoes, the hand of biology signs off with a clear signature — in other words, a distortion in the natural balance of isotopes.
Canfield examined the sedimentary iron pyrites deposited during the Precambrian period for sulphur signatures, and found them. The first signs of a skewing in the sulphur-isotope ratios date to about 2.7 billion years ago, implying there was a build-up of oxygen at this time. Interestingly, this is very close to the date given to the first eukaryotic cells in the Hamersley shales by Jochen Brocks and his colleagues. After this, little
Three Billion Years of Microbial Evolution • 43
changed for half a billion years. Then, around 2.2 billion years ago, there was an abrupt rise in the sulphur-32 content of iron pyrites, suggesting that the amount of sulphate in the oceans must have risen to the point where they could support a much larger population of sulphate-reducing bacteria. This, in turn, indicates that much more oxygen must have been available to produce the sulphate. Thus Canfield’s work implies that there was a small rise in oxygen levels 2.7 billion years ago, followed by a much larger rise about 2.2 billion years ago.
Unequivocal evidence of free oxygen in the air and oceans requires proof of oxidation on land, as changes wrought by the thin air cannot be obscured or confounded by the rich biology and chemistry of the oceans.
More than a billion years before the invasion of the land by plants and animals, the terrestrial populations of microbes could not have compared in abundance or diversity with their marine cousins. The widespread rusting of iron minerals on land is therefore the most tangible evidence we have for oxygen in the atmosphere. These rusting iron minerals are found in fossil soils (palaeosols), and in the so-called continental red-beds.
In a classic series of measurements, the geochemists Rob Rye and Heinrich Holland from Harvard University examined the iron content of ancient fossil soils, and used these measurements to estimate the period when oxygen built up in the air. Their reasoning was as follows. Because iron dissolves in the absence of oxygen, but is insoluble in the presence of oxygen, iron could leach out of very ancient soils (when there was no oxygen in the air) but became trapped in more recent soils (when oxygen was present in the air). By measuring the iron content of fossil soils, Holland and Rye estimated that a large rise in atmospheric oxygen took place between 2.2 and 2 billion years ago. From the amount of iron left in the fossil soils, as well as its rustiness — its oxidation state — they estimated that the concentration of atmospheric oxygen at this time probably reached 5 to 18 per cent of present atmospheric levels.
In terms of timing, these findings are corroborated by the appearance of continental red-beds between 2.2 and 1.8 billion years ago. These sandstone rock formations were probably formed by free oxygen reacting with iron in the rocks during the erosion of mountain ranges. Rivers must have run red as they flowed over the barren surface of the Earth, a scene that conjures up images of nuclear winter. Rather than being washed out to sea, some eroded minerals deposited in valleys and alluvial
44 • SILENCE OF THE AEONS
plains, ultimately forming the beds of red sandstone. Because the red-beds were formed from eroded minerals, however, we cannot use them to estimate the concentration of oxygen in the air, only the timing.3 The timeline from the first carbon signatures in Greenland rocks to the formation of the red-beds is shown in Figure 2.
A bizarre microbial relic also attests to a rise in free oxygen around 2 billion years ago: the natural nuclear reactors at Oklo, in Gabon, West Africa. The solubility of uranium, like iron, depends on oxygen. But unlike iron, uranium becomes more soluble, rather than less, in the presence of oxygen. The chief uranium mineral found in rocks older than about 2 billion years is uraninite, but this ore is very rarely found in younger rocks. The sudden transition is associated with the rise in oxygen. What seems to have happened is that, as the oxygen levels increased, oxidized uranium salts leached out of uraninite ores in the rocks and washed away in streams. Their concentratio
n cannot have been higher than a few parts per million.
In Gabon, 2 billion years ago, several streams converged on shallow lakes encrusted with bacterial mats, similar to the mats that still exist today in the geyser pools at Yellowstone National Park in the United States and elsewhere. Some of the bacteria that lived in these mats had a penchant for soluble uranium salts as an energy source. They converted the soluble uranium back into insoluble salts, which precipitated out in the shallow water beneath them. Over the next 200 million years or so, the bacterial mats deposited thousands of tons of black uranium ore in their lakes.
There are two main isotopes of uranium, both radioactive, as most of the Cold War generation knows. Uranium-238 has a long half-life of 4.51 billion years. Half the uranium-238 that was present when the Earth condensed from its cloud of radioactive dust is still out there somewhere.
Its sister isotope, uranium-235, decays much faster, with a half-life of about 750 million years. Most uranium-235 has therefore already decayed into its daughter elements, by emitting neutrons. If one of these neutrons hits a nearby uranium-235 nucleus, however, the effect is to split the nucleus into one or more additional neutrons, plus large fragments of roughly equal mass, with a liberation of energy equal to the total loss of 3 The red colour of the continental red-beds shows that the iron was completely oxidized, as would be expected for deposits of eroded debris that had been exposed to the air for an indeterminate, but probably lengthy, period. Because there is no spectrum of oxidation, we cannot estimate the atmospheric oxygen levels from the red-beds.
Three Billion Years of Microbial Evolution • 45
Million
years ago
1800
Uranium reactors
First multicellular algae
0
Present day
1900
Rise in O2 to
Red-beds
about 5–18%
2000
of present
First eukaryotes with mitochondria
500
atmospheric
Cambrian
2100
Oxidized fossil soils
level deduced
explosion from oxidized
First eukaryotic fossils
2200
minerals and
Global tectonic activity
sulphate-reducing
1000
bacteria
2300
Snowball Earth
2400
1500
2500
Banded-iron formations
Rise in O2
to about 1%
2600
of present
2000
atmospheric
2700
Biomarkers of cyanobacteria
level deduced
Biomarkers of eukaryotes
from sulphate-
2800
reducing bacteria
2500
2900
3000
3000
3100
3200
3500
3300
3400
4000
3500
Fossil stromatolites
Fossil cyanobacteria?
3600
4500
3700
3800
Carbon signatures
3900
in Greenland rocks
Figure 2: Geological timeline expanding the mid-Precambrian period (Archaean and early Proterozoic). Note the burst of evolutionary activity in the period 2.3 to 2 billion years ago, as oxygen levels rose to about 5–18% of present atmospheric levels.
46 • SILENCE OF THE AEONS
mass. (Energy is related to mass according to Einstein’s famous equation E = mc 2.) If the uranium-235 atoms are closely packed together, there is a good chance that the newly emitted neutrons will hit more uranium-235
nuclei. In these circumstances, a chain reaction — nuclear fission — can take place, potentially causing a nuclear explosion.
For nuclear fission to take place, uranium-235 must be enriched to at least 3 per cent of the total mass of uranium. Today, uranium-235
accounts for only 0.72 per cent of uranium by mass, so we must enrich it ourselves if we wish to build a nuclear power station or an old-fashioned uranium atom bomb. Two billion years ago, however, less uranium-235
had already decayed. Its content in uranium ores would have been higher
— in fact about 3 per cent. The uranium-loving bacteria in Gabon therefore stockpiled enough ore enriched in uranium-235 to start a nuclear fission chain reaction. This, at any rate, was the conclusion of the French secret service in 1972. There had been something of a panic when uranium ores mined along the Oklo River, near the border with the Congo Republic, turned out to be depleted in uranium-235. Some consignments had less than half the expected 0.72 per cent uranium-235. In an Africa still emerging from colonial rule and beset by civil unrest, the implication that some tribal group had stolen enough uranium to make a nuclear bomb did not bear thinking about. The French threw everything at the problem, and it was not long before a large team of scientists from the French Atomic Energy Commission solved the case.
Samples of the Oklo ores showed clear relics of spent radioactive fission, even when they were extracted from undisturbed seams. Instead of decaying naturally, tons of uranium-235 had fissioned away in half a dozen separate locations, producing a million times the power of natural decay. The natural reactors in Gabon had apparently been sustained for millions of years by a steady flow of water from the streams that fed into the ancient uranium lakes. Water slows the speed of neutrons, reflecting them back into the core of the reactor, so instead of quelling the inner fires, water actually promotes nuclear fission. The streams did more than this, however — they also acted as safety valves against nuclear explosion.
Whenever the chain reactions approached danger levels, water boiled off, allowing neutrons to escape. This scuttled the chain reactions and shut down the reactors until flow was re-established. There is no evidence of a nuclear explosion. The entire system was finally buried beneath sediments where it remained undisturbed until the arrival of the French, a testament to the ingenuity of bacteria 1.8 billion years before Enrico
Three Billion Years of Microbial Evolution • 47
Fermi and his Chicago team applied their genius to making the first man-made atomic bomb; and indeed a testament to the potential long-term safety of burying nuclear waste.
What of the catastrophic mass extinction, the oxygen holocaust described by Lynn Margulis (see Chapter 2, page 19)? There is no trace of a holocaust in the rocks. Far from being a profound and debilitating challenge, the appearance of oxygen seems to have driven the evolution of new forms of metabolism, and new branches in the tree of life, as argued by Preston Cloud in the 1960s (see Chapter 2). But why did it take so long for oxygen to accumulate, despite more than a billion years of continuous production by cyanobacteria? To put it into context, this interlude is twice as long as the entire modern era of plants and animals (the Phanerozoic), or for that matter 15 times as long as the period since the demise of the dinosaurs. Is this long gestation perhaps hidden evidence of a difficult adaptation, concealing the throes of life as it struggled to cope with a poisonous gas? It seems unlikely. A number of speculations can explain the delay; for example, iron-loving bacteria may have dominated the ecosystem until the iron ran out, or the cyanobacteria may have been restricted to shallow-water stromatolite communities that absorbed as much oxygen as they produced, because of the presence of non-photosynthetic oxygen-respiring bacteria. The most likely explanation is simply that there was no change for a billion years because a stable equilibrium persisted for that time.
The long stasis was finally shattered by an apocalyptic climate change about 2.2 to 2.3 billion years ago. The Earth plummeted into the first ever ice age. This was no trivial ice age, to be compared with the recent Pleistocene cold sn
ap, but a global freeze that may have covered the tropics in glaciers a kilometre [3280 feet] thick — in Joseph Kirschvink’s memorable phrase, a ‘snowball Earth’. What made the pleasant Precambrian climate collapse so violently is not known. One theory, argued by the sometime NASA geochemist James Kasting, is that the appearance of free oxygen itself brought about the freeze. As it built up in the air, oxygen would have reacted with methane (produced in large amounts by bacteria), and so removed this important greenhouse gas from the early atmosphere. As the greenhouse effect was undermined, temperatures plummeted and the Earth succumbed to the grip of an ice
48 • SILENCE OF THE AEONS
age. Kasting’s theory has been advocated by James Lovelock among others, who claims an important role for methane-producing bacteria in his books on Gaia, but at present the theory suffers from lack of strong supporting evidence.
Whatever the reason, there is no doubt that the Earth plunged into a serious ice age about 2.3 billion years ago. It was to last for 35 million years. Hard on the heels of this ice age the planet was racked by a period of heightened tectonic activity, leading to major continental rifting and the uplift of mountain belts on a scale comparable with the Andes.
Joseph Kirschvink, a specialist in palaeomagnetism at Caltech (the California Institute of Technology) is a leading advocate of the snowball Earth theory, and one of its most thoughtful commentators. He argues that, after the glaciers finally melted, the stones and mineral dust scoured out by glacial erosion would have filled the oceans with minerals and nutrients, stimulating a cyanobacterial bloom and a rise in oxygen. As evidence for this claim, Kirschvink and his co-workers cite a huge deposit of manganese ore in the Kalahari desert in southern Africa, dated to right after the end of the snowball Earth. The Kalahari manganese field contains some 13.5 billion tons of manganese ores, or about 4 billion tons of manganese, making it by far the world’s largest economic reserve of this element.