Armand H Delsemme. American Scientist. Volume 89, Issue 5. Sep/Oct 2001.
The first 600 million years of our planet’s history have been erased from its surface. Between the time it was formed about 4.6 billion years ago and the formation of the oldest known sedimentary rocks, which are about 4 billion years old, the Earth changed from a hot, dry little rock to a world with an ocean and an atmosphere-a planet that was primed for the origin of life. Those missing years hold the key to everything we hold dear on the surface of our planet. Remarkably, we are still not sure exactly what took place during that time.
The subsequent evolution appears to have been unique in our solar system. None of the other planets is covered with an ocean of liquid water, and none of their atmospheres is especially receptive for the evolution of life. Most of the water on Mars is locked in its polar ice caps, and only recently has any water at all been found on the Moon– frozen on its poles. Mercury and Venus are both much too hot, and the outer planets and moons are much too cold. There is reason to believe that at least one of Jupiter’s icy moons, Europa, may harbor an ocean of water, but it must be at some depth below the surface where the pressure and temperature could accommodate the liquid state. All told, the Earth appears to have been in just the right spot. But in the right spot for what, exactly?
The question bears on the larger issue of whether life might exist on planets in other stellar systems as well. Dozens of extrasolar planets have been discovered so far, and it’s becoming clear that their formation around other stars may be quite common. But whether the formation of earthlike planets is common, is another question. Would an Earth-sized planet, formed at just the right distance from a Sunlike star, always develop an ocean and an atmosphere friendly to life? What else might be necessary?
Planetologists have been struggling with such questions for decades, and I now believe we have found some answers. There is little doubt in my mind that our oceans and our atmosphere were delivered on the backs of comets that bombarded the newly formed Earth in its first few hundred million years. What is more, the comets also appear to have brought prebiotic molecules-organic building blocks that could be used to get life started. These ideas have a fairly long history but have been resisted for various reasons over the decades. I have been studying the chemistry of comets for more than 50 years, and I admit that early in my career I too was reluctant to accept the possibility that comets had played such a crucial role in our planet’s history. But the evidence has continued to accumulate over the decades, and it now seems irrefutable. Here I provide an overview of the reasoning behind this extraordinary idea.
A World Without Water
How do we know that the infant Earth did not come fully formed with its continents, oceans and atmosphere? Surprisingly, the answer lies within small rocks that fall from the sky. Certain meteorites, the variety known as chondrites, are relics from the era when the solar system was first formed. Examine a chondrite under a microscope and you’ll see that it consists of very fine dust grains of various compositions, gently compressed together to form a “sedimentary” rock. The sedimentation of chondrites did not take place in a river or an ocean, however. Rather, these rocks were formed in the accretion disk that surrounded the nascent Sun, or protosun.
The general course of events in the formation of an accretion disk in the early solar nebula is now reasonably well understood (see Protostars, July-August, and How Were the Comets Made? May-June). Although the gas and dust that formed the protosun and its circumstellar disk were originally quite cold, the process of accretion heats the materials as they fall into the gravitational well of the newly forming solar system. This heating was modeled in 1978 by Alastair Cameron of the Smithsonian Astrophysical Observatory, who showed that the temperature of the disk in the zones where the planets are formed reaches a maximum just before the dust sediments into the midplane. This is because mass infall induces both heating and turbulence. When mass infall stops, gas turbulence subsides, and dust, which is no longer carried by gas eddies, then sediments onto the midplane.
After this, the dust grains collect into ever-larger pieces (planetesimals), which eventually coalesce to form the planetary bodies. The chondrites are remnants of the circumstellar disk that were not incorporated into a planet, and so escaped the forces (heat and gravity) that would have destroyed their “pristine” state. Because of this, the composition of a chondrite tells scientists much about the processes that must have taken place in the protosolar accretion disk.
When we take a closer look at the chondritic meteorites, we notice three general varieties: enstatite chondrites, ordinary chondrites and carbonaceous chondrites. The enstatite chondrites are not relevant here. Ordinary chondrites and carbonaceous chondrites are identical except for the depletion of volatile elements (such as carbon, lead and bismuth) in the ordinary chondrites. The difference indicates that the ordinary chondrites experienced some heating that cooked away the more volatile components. The identity of the depleted elements also serves as a temperature gauge. Edward Anders of the University of Chicago and his colleagues have shown that these volatile elements are vaporized at temperatures above 450 kelvins. This temperature separates the ordinary chondrites from their cooler, carbonaceous cousins.
So where in the disk can we place this 450-kelvin boundary line between the two classes of chondrites? Fortunately, we have a good idea where the chondrites came from before they landed on the Earth. There’s little doubt that they are pieces of asteroid families that lie in the region between Mars and Jupiter. Scientists have been able to take this a step further and deduce that most of the carbonaceous chondrites come from the dark, C-type, asteroids, whereas the ordinary chondrites are pieces of the brighter, S-type, asteroids. Using the orbital statistics of C and S asteroids, I was able to show that the two classes of chondrites were originally separated at a distance of 2.6 astronomical units (AU) from the protosun. (One AU equals the average distance between the Earth and the Sun.) This tell us that the protosolar disk had a temperature of 450 kelvins about 2.6 AU from the protosun just as the dust sedimented into the midplane.
Using this information, we can extrapolate a temperature gradient for the accretion disk at various distances from the protosun. It’s a negatively sloping curve, based on the thermal equilibrium of the solar nebula, with the temperatures increasing as one gets closer to the protosun and decreasing as one moves farther away. In the region where the primordial Earth formed (between 0.8 and 1.3 AU from the protosun), the accreting grains of dust reached temperatures between 900 and 1,400 kelvins. Such searing temperatures would have thoroughly degassed these materials, leaving dust grains almost solely consisting of silicates and reduced iron. (Not surprisingly, iron and silicate-bearing rocks are the primary components of the deep Earth.) The volatiles in the zone of the protoearth’s formation were effectively separated from the solid components. All of the carbon was in carbon monoxide, all nitrogen in gaseous N2 and all of the water in hot steam. No organics could have been present in the dust.
Could these volatiles and organics have been incorporated into the protoearth during the planet-building process? This seems unlikely. In the earliest stages, as the dust grains came together to form larger pieces (about a kilometer in diameter), the sheer size of the planetesimals would have prevented any gas absorption. At an intermediate stage of accretion when the protoearth was about 10,000 kilometers across-about 83 percent of its final mass-growing gravitational interactions between ever larger planetesimals would have begun to stir up the contents of the protoplanetary disk. Materials from as close in as 0.4 AU and as far away as 2.6 AU could have been mixed into the forming planet. Yet even this radial mixing of components could not have contributed volatiles to the protoearth. As we’ve seen, materials from a distance of 2.6 AU, cooked to 450 kelvins, were devoid of water and gas, much as are ordinary chondrites.
We must, however, account for the final 17 percent of the Earth’s mass, which roughly corresponds to the upper mantle. This material represents the last stages of the planet’s accretion. Theoretical considerations suggest that a few relatively large, asteroid-like objects may have collided with the protoearth during this time. Some of these large projectiles could have come from beyond 2.6 AU and would have contained volatile-rich carbonaceous materials. Could such objects have brought the Earth its water?
The composition of the upper mantle can help us answer this question. In 1983, Heinrich Wanke, of the Max Planck Institut fur Chemie in Mainz, and his colleagues analyzed the composition of the mantle based on the material spewed up in a volcanic eruption. They found that volcanic ejecta are typically enriched in heat-resistant elements by a factor of 1.3, but depleted of moderately volatile elements by factors of 0.1 to 0.2, and depleted of very volatile elements by factors of 0.01 to 0.0001! This is effectively the opposite of what one would expect if a large, carbonaceous chondritic asteroid had hit the protoearth gently enough not to produce any ejecta. Such a depletion of volatiles can only be explained by the intense heating produced by the collision that formed the Moon.
Even if an asteroid had made a contribution to the Earth’s upper mantle, it could not have brought the world its oceans. In 1991, I calculated that a giant asteroid-about 1,000-kilometers across (about twice as massive as Ceres, the largest known asteroid)-could not have brought more water than a few percent of the oceans’ mass. All of these studies suggest that the oceans and the atmosphere must have come after the primordial Earth had acquired most of its mass.
A Heavy, Heavy Rain
Anyone who has looked at the Moon through a telescope cannot help but be impressed by its pockmarked appearance. Its surface is covered with craters caused by the hammer-like blows of impacting objects. Photographs from the orbiting Apollo missions reveal that the far side of the Moon is even more heavily cratered. The Moon is not unique: Spacecraft missions to Mars and Mercury have shown that these planets are also heavily cratered. Moreover, the craters on all three bodies are remarkably similar with respect to their sizes and densities, suggesting that all experienced the same heavy bombardment at some point in their histories. Despite the relative absence of impact craters on the Earth, its proximity to these other planets leaves little doubt that it too must have faced such a period of bombardment. The question is, when did these impacts take place, and what could have caused them?
Fortunately, we needn’t rely solely on theory to assess the age of these impacts. Lunar rocks, brought back by the Apollo and Luna missions, have been dated by their radioactive isotopes-a measure of when the rocks solidified. The ages of the rocks, taken from various regions of the Moon, correspond to the time when these areas were covered with molten lava, and so had obliterated any trace of previous craters. By counting the density of the craters in these regions, we can assess the intensity of the bombardment and its age. These types of data suggest that the lunar impact rate was most intense during the first 600 million years of the Moon’s history. Since the Moon was formed merely 50 million years after the Earth’s formation (4.56 billion years ago), this 600-million-year period also corresponds to the missing era in the primordial Earth’s history.
The intensity of the bombardment during this period is simply too large to be explained by planetesimals from the inner solar system-most of these would have been already incorporated into the planets themselves. The long duration of the impact flux also implies that the bombarding projectiles had to come from farther out, and a clue to their identity comes from the work of the Russian astronomer Viktor Safronov.
In the late 1960s, Safronov demonstrated that in the final stages of their growth, the giant planets, especially Jupiter, tugged on the planetesimals in the outer parts of the solar system, sending them on hyperbolic trajectories into interstellar space. Since the planetesimals were ejected in random directions, a fair number of them must have passed through the inner solar system. In doing so, some of them inevitably would have collided with the inner planets, including the primordial Earth.
The nature of these planetesimals can be deduced by their location in the accretion disk. At Jupiter’s orbital distance (and beyond), the temperature was so low that the dust grains never lost their frosty cover of water ice and organic volatiles, originally acquired in interstellar space. When they accreted into larger planetesimals they formed the icy bodies that we now call comets. Indeed, a remnant population of ejected comets still exists in the form of the Oort Cloud, an extended sphere of comets that encloses our solar system at a distance as far as 1,000 times the orbit of Pluto. All of the Oort Cloud comets, still gravitationallly bound by the Sun, originally came from the zone of the accreting giant planets. Their existence is a testimony to that era.
So comets must have peppered the Earth early in its history, but could they have brought enough water to fill the world’s oceans? Safronov’s model predicts the mass of the scattered comets to be about 6 or 7 times the masses of the giant planets’ solid cores. This equates to nearly 400 Earth masses’ worth of comets in the region between Jupiter and Neptune. Given that roughly half the mass of a typical comet is water, there must have been thousands of frozen oceans in orbit beyond Jupiter. One estimate holds that it would have taken a million comet impacts to account for all the seawater on our planet. Could there have been that many? More, actually. My calculations, based on Safronov’s model, suggest that the comets brought 5.8 times the amount of seawater now in our oceans and 680 times the gas in our atmosphere!
From a dearth of volatiles to an overabundance. How do we account for the “missing” water and gases? The mystery is solved if one considers the velocity of the impacting objects. Hurled by the gravitational energy of the giant planets, and accelerated by their fall toward the sun, the comets would have hit the Earth with an average speed of more than 42 kilometers per second. This greatly exceeds the escape velocity of our planet-about 11.2 kilometers per second-so that the whole process must have thrown several oceans and hundreds of atmospheres back into space.
If comets did bring the Earth its oceans and its atmosphere, then we might expect some chemical evidence of their contribution on the surface of our planet. Our seawater, the air we breathe and even the ground we walk on may well have some trace of the massive bombardment-even though it occurred more than 4 billion years ago. The trick lies in determining what it is that we should find. It turns out to be less straightforward than we might have hoped.
Consider the first piece of evidence, which involves the simple hydrogen (H) atom. Garden-variety hydrogen merely contains a single proton in its nucleus. But its more exotic cousin, deuterium (D), holds one proton and one neutron-forming an unstable isotope that burns into helium at stellar temperatures. The deuterium atom is interesting because every natural bit of it was created in the Big Bang– about 20 to 30 parts per million (ppm) hydrogen atoms. It’s not easy to make deuterium any more, and its abundance has been diminishing over the eons to the point where the local interstellar clouds only have about 5 to 15 ppm deuterium. When the solar system was formed 4.56 billion years ago, the deuterium abundance was still about 20 ppm, as demonstrated by its presence in the atmospheres of Jupiter and Saturn, which were captured from the gas of the solar nebula before it dissipated. Earthly seawater contains 160 ppm of deuterium, an enrichment of eight times relative to the deuterium in the solar nebula. So we would expect the comets to be similarly enriched, no?
No. As it happens, the water in the three bright comets-Halley, Hyakutake and Hale-Bopp-that recently swung past the Earth averaged about 320 ppm deuterium, an enrichment of 16 times over the deuterium in the solar nebula, and twice that found in seawater! How do we account for these differences?
We must again return to Safronov’s model of how the newly forming giant planets scattered the comets. Because of their differing locations and masses, the planets flung the comets in their vicinity in different directions and speeds. My calculations show that the largest planet, Jupiter, accelerated the nearby comets to more than 35 kilometers per second, effectively sending most of them out to interstellar space, beyond the grasp of the Sun’s gravity. The comets accelerated by the other giant planets-Saturn, Neptune and Uranus-did not attain sufficient velocities to escape the Sun and so became trapped in the Oort Cloud. In fact, we can calculate that the three outer giants contributed about 96 percent of the Oort Cloud’s cometary mass, whereas comets from the vicinity of Jupiter only account for the remaining 4 percent. The same model predicts that 80 percent of the comets that hit the Earth (after the Moon was formed) came from Jupiter’s orbit, whereas only 20 percent came from the orbits of the three outer planets.
These differences help to explain the paradoxical values of deuterium in seawater and the three bright comets. Simply put, the comets that contributed the bulk of the Earth’s seawater are different. They were formed at a higher temperature, in the zone of Jupiter’s orbit, where they acquired more water but less deuterium than the comets formed at much lower temperatures in the zones of the outer giant planets. The Jupiter-zone comets are enriched in water because here the icy grains create a “cold front” that condenses any steam vaporized in the hotter zones of the inner solar system. This steam carries much less deuterium than the interstellar frost and so dilutes the relative proportion of this isotope in the Jupiter-zone comets. The extra steam also adds an extra layer of ice to the dust grains before they accrete onto the comets, and so increases their total content of water.
There are several ways in which water can be enriched in deteurium relative to interstellar hydrogen. One reaction involves the exchange of neutral isotopes:
This reaction is extremely sluggish at very cold temperatures. However, there are many other ways to produce deuterated water, which involve ion– molecular reactions. Charge exchanges between ions and molecules easily cross the thermal-energy barriers, and so easily exchange deuterium with hydrogen, as for instance in:
Reactions of this type are fast even at temperatures close to absolute zero, hence they prevail in interstellar space and help to explain why the frost that covers interstellar dust grains is so enriched in deuterium. Although the degree of enrichment depends on the local interstellar temperature, it is generally about 10 to 20 times the deuterium levels of the solar nebula. This is consistent with the sixteenfold enrichment seen in the three bright comets, and it suggests that they must have formed at very low temperatures. It also suggests that the outer zones never experienced temperatures that would have altered the relative enrichment of the interstellar grains.
Near the zone of Jupiter’s formation, however, the temperatures are considerably warmer, about 225 kelvins. This zone corresponds to the frost line of the accretion disk, where steam condenses to form snow. Laboratory experiments at this temperature show that deuterated snow approaches a sixfold enrichment in the presence of hydrogen. The comets formed in Jupiter’s zone would have had less than half the deuterium of the outermost comets.
Now, if we consider the proportions of comets that came from Jupiter’s zone (80 percent, with a sixfold enrichment) and from the outermost planets (20 percent, with a sixteenfold enrichment), the eightfold enrichment of deuterium in our seawater is entirely consistent with a cometary origin. The paradox is not only solved, but theory actually predicts it!
Among the comets we observe nowadays, only four percent of the long-period comets and none of the short-period comets originated in Jupiter’s zone. However, a Jupiter– zone comet may have passed by the Earth recently. Mike Mumma of the NASA Goddard Space Flight Center and his colleagues have shown that Comet C/1999 S4 LINEAR contains four to five times more water (in proportion to cometary volatiles) than most other comets, and the authors argue that its unusual composition may indicate its formation in Jupiter’s zone. Unfortunately, C/LINEAR was very fragile, breaking into fragments that vaporized in just a few days-before its deuterium to hydrogen ratio could be measured.
The comet-bombardment scenario also helps to explain the abundances of two other chemical markers-krypton and xenon-in our atmosphere. Because these inert gases don’t combine with the other elements, and their large atomic masses prevent them from escaping the Earth’s gravity, their atmospheric abundances reflect a primordial composition. Compared to solar abundances, however, their isotopic proportions are anomalous, suggesting a fractionation process in which some isotopes are enriched relative to others. The xenon in our atmosphere is enriched in the heavier isotopes, but the krypton isotopes are not similarly enriched. This decoupled fractionation cannot be explained by any evolutionary history on the Earth.
There is, however, a cosmic explanation. Laboratory studies by Akiva Barn-Nun and his coworkers, of Tel Aviv University, show that krypton and xenon can be trapped within crystalline clathrate hydrates or, at very low temperatures (with practically the same results), within amorphous ice. These studies suggest that the krypton and xenon in the Earth’s atmosphere were trapped within amorphous ice formed at about 30 to 75 kelvins-temperatures found in the comet-forming zones of Uranus and Neptune. On the other hand, the temperatures near the comet-forming zones of Jupiter (225 kelvins) and Saturn (130 kelvins) were not cold enough to trap xenon and krypton. The relative paucity of comets from Uranus and Neptune thus helps to explain the very low abundance of xenon and krypton in our atmosphere. Accurate measures of the isotopic gas fractions in comets of different origins should eventually resolve any lasting doubts.
The last piece of chemical evidence I’ll consider here lies in the Earth’s crust and upper mantle. Much as oil and vinegar separate according to their densities in a bottle of salad dressing (with the lighter oil on top), the core and mantle of the Earth are layered according to their densities. The separation of these layers dates to the earliest period of the Earth’s formation, when it was still accumulating mass by the accretion of planetesimals. The energy of the accretionary impacts was transformed into a heat so intense that Earth’s surface was covered with a thick layer of molten lava, perhaps to very great depths. This liquid state provided a medium for differentiation: The heaviest elements-notably iron–fell towards the nascent planet’s core.
As the iron sank into the planet’s depths, it would have carried siderophile (“iron-loving”) metals, such as nickel, cobalt and molybdenum, with it. The crust and the upper mantle should be fairly depleted of these elements, but they are not. In fact, the siderophile metals are present in nearly solar proportions, suggesting that they must have been brought to the planet after its core had formed.
The Primeval Biosphere
The primordial surface of the comet-barraged Earth would surely have been unrecognizable to us. It would have looked rather like the lunar highlands, where the density of impact craters is so great that they overlap one another other. This helps to explain why there are almost no geological traces of sediments dating from this period: They were obliterated. In fact, as two studies reported earlier this year, all that may remain of the oldest known continental crust are grains of zircon-existing as detritus in much younger rocks!
Simon Wilde, of Curtin University of Technology in Australia, and his colleagues have dated grains of zircon from Western Australia to 4.4 billion years. And Stephen Mojzsis, now at the University of Colorado, and his coworkers dated zircon grains from the same region to 4.3 billion years ago. Both studies suggest that a continental crust and an ocean existed within 200 to 300 million years after the Earth was formed.
The comet flux would have been very heavy when these early continents arose, but the rate decreased exponentially over the course of the bombardment. After 400 million years, the comet flux fell to one percent of its initial value, and then to merely 0.1 percent after 600 million years.
All told, about 1018 tons of cometary material crashed onto the Earth’s surface during this era. This is equivalent to one billion small cometary nuclei (about one kilometer across) or one million large nuclei (about 10 kilometers across). The average rate of bombardment equates to a few dozen medium-sized comets hitting the Earth every century.
We can tally their material contribution to the young planet based on the average composition of a comet– which is about 50 percent water, 13 percent volatile organics (carbon monoxide, carbon dioxide and various volatile organics bearing nitrogen), 15 percent refractory organic molecules (relatively resistant to destruction) and 22 percent silicate rocks. If much of the material hadn’t been ejected back into space, it would have coated the Earth with a uniform layer of cometary matter several kilometers deep.
A picture of the young Earth near the end of the bombardment period would show a cloudy atmosphere, dozens of times thicker than our own. Such an atmosphere would protect the ground and prevent it from cooling rapidly in an era when the young Sun was about 30 percent less bright than it is now. Even so, most of the cometary water would have turned into steam. As the ground temperature fell below the boiling point (403 kelvins at a pressure of 30 atmospheres), the steam would have condensed to form a hot ocean. In the residual atmosphere, the solar ultraviolet radiation and the high temperatures would have induced both photodissociations and polymerizations of some cometary molecules. Formaldehyde (CH2O) would have polymerized to form a fog of solid polyformaldehyde particles in the stratosphere. Ammonia (NH3) would have been photodissociated into molecular nitrogen (N2) and hydrogen (H2), and the hydrocarbons would oxidize and lose their hydrogen. All of the hydrogen would have escaped from the top of the atmosphere and been quickly lost to space.
At this point the atmosphere would consist of 80 percent carbon dioxide (CI2), 10 percent methane (CH4), 5 percent carbon monoxide (CO) and 5 percent nitrogen (N2). The atmosphere has yet to reach a steady state, however, because it is still being fed by a (diminishing) cometary bombardment. It is turbulent in the extreme. Incessant torrential rains of water acidified with CO2 attack the silicate rocks on the ground. The ground temperature is soon low enough (much below 373 kelvins) that the carbonates made by this reaction become stable, forming solid sediments that bury the CO2 in the ground, and so cause the atmospheric pressure to fall drastically, to 10 and then perhaps to 5 atmospheres and less. Indeed, major sediments of limestone (calcium carbonate, CaCO3) and dolomite (a carbonate of calcium and magnesium) still exist in what is now Greenland and are 3.8 billion years old.
At the same time, some of the organic molecules delivered by the comets may have had a few interesting chemical interactions of their own-actually giving a “jump start” to the first life on our planet. Although some have questioned whether organics could survive the heat of an impact, the issue now seems to be resolved. The survival of 74 different amino acids (most of which are not known on the Earth) on carbonaceous chondrites, such as the Murchison meteorite, suggest that organics could at least survive a minor impact. And recent studies by Elisabetta Pierazzo, of the University of Arizona, and Christopher Chyba of the SETI Institute in Mountain View, California, suggest that some amino acids could even survive the shock heating of kilometer-sized cometary impacts. In any case, Anders and I have, independently, argued that an extremely large flux of interplanetary dust particles (derived from the tails of comets that missed the Earth during its first 600 million years) could have salted the young Earth with enormous quantities of prebiotic molecules. Indeed, in 1985 Don Brownlee of the University of Washington, Seattle, showed that cometary dust grains, captured in the upper atmosphere, contain undamaged organic molecules.
About 3.5 billion years ago large cometary impacts would have become increasingly rare, but when they did occur, they produced enormous cataclysms. The oceans would have boiled near the impact site, causing hurricanes and gigantic waterspouts with fantastic ejections of gas and water into space. Under these chaotic and seemingly inhospitable conditions, a phenomenon occurs that is going to have astonishing consequences: Bacteria begin to multiply in the hot waters of the first oceans.
There is now considerable evidence to support the idea that we owe the existence of our biosphere to a heavy bombardment of comets in the very early history of our planet. Indeed the delivery of water and prebiotic molecules explains why life emerged so soon after the conditions ceased to be utterly hostile. The oldest fossil imprints of bacteria date to about 3.456 billion years ago (in Australian rocks), and there is indirect evidence that life was present 3.8 billion years ago in the ancient sediments of Greenland.
This process could have occurred many times on rocky planets in other stellar systems. And, if our solar system is any indication, the scattering of comets by cold giant planets might be a necessary condition for the emergence of life in the universe.