Largely a summary of Sleep et al. 2001, with some updating (see references).

A paradox about planets is that they must have formed in zones of the solar system where mass tended to concentrate: that is, where many bodies interacted and collided. As protoplanets formed, collisions between them became much more energetic. In particular, each collision generated a lot of heat. The last stages of planetary formation thus must have produced hot planets, probably with molten surfaces.

Certainly Earth (and the Moon) seem to have formed when the proto-Earth was struck by a body about the size of Mars. The Earth formed with a tilt to its axis, and with a Moon orbiting close to it. The chemical make-up of the Moon suggests that it formed largely from debris that was blasted off the outer layers of the Earth in this collision. That collision would have been powerful enough to vaporize much of the Earth.

So Earth was born hot, and then cooled some time later. It would be important to estimate how long that took. When did Earth cool to the point where life might form on it, or arrive on it?

Here is one scenario: others are possible, of course, because we have little data and must depend a lot on computer models and chemical and physical calculations, and both of these methods require us to make assumptions with varying degrees of confidence.

At first, the Earth was hot, and the early Sun seems to have been rather cool, maybe producing 70% of the radiation it puts out today. Earth's atmosphere was controlled by the heat coming from Earth's surface.

That early atmosphere would have been rock vapor soon after the giant collision, but that vapor would have condensed quickly into mineral droplets and crystals and fallen back into the seething molten planetary surface. This might have taken a few thousand years (quickly on a geological time scale!). However, powerful convection currents in the molten Earth would have carried not just molten rock but heat to the surface, where much of it would be lost by radiation out into space. As this powerful convection churned the Earth, the denser elements and compounds, particularly solid and liquid iron, would have sunk to begin to form a core. As the core formed, the convective zone would gradually have narrowed to define the Earth's mantle (there was no crust yet).

At the point when the "rock" compounds had largely rained out, Earth's surface and atmosphere would have been perhaps 1800C, and the atmospheric gases would probably have been dominated by carbon dioxide and water vapor. As the "magma ocean" of the planetary surface cooled, it began to form a thin, hot crust. Calculations suggest that only a few hundred years would be enough for the entire surface to become "solid". "Solid" surface means that it the new crust now sealed off the molten interior, except where many volcanic vents and new asteroid craters allowed lava to reach the surface. Even so, the convection that still churned the great mass of the molten Earth would no longer be delivering heat directly to the surface, but only to the underside of the thin crust. Solar radiation and the atmosphere itself now began to play a much greater role in the evolution of Earth's surface.

The dense early atmosphere, dominated by the "greenhouse gases" carbon dioxide and water vapor, trapped the radiation of the early Sun, while the solid crust radiated volcanic heat from the interior. The crustal heat was convected upwards through the thick atmosphere and radiated to space, so that Earth continued to cool relatively quickly at first, then more slowly as the crust cooled and thickened. Paradoxically, the more the greenhouse gases heated the atmosphere, the faster heat would have radiated to space.

The dense atmosphere might have been 25 times denser than today's, and at a temperature over 200C. However, at temperatures and pressures like these, it is possible to form liquid water (water at these temperatures exists today, deep in the oceans at the mid-ocean ridges). So in the end, perhaps after a few million years, the interior had lost enough heat that its contribution to the atmosphere was much reduced. As atmospheric temperatures dropped to perhaps 250, water vapor began to rain out, carrying carbon dioxide dissolved in it, and hot liquid water filled up low places on the crust to make shallow, slightly acid seas.

With massive amounts of volcanic action on the early sea floors, and large asteroid impacts, there would have been dramatic chemical reactions between hot rock and water. Sodium and chlorine would have condensed fairly early out of the atmosphere, and the early ocean would have been salty, perhaps super-salty (a hot brine) increasing the scope of rock/water reactions. We know from the submarine vents on today's mid-ocean ridges how impressive those chemical reactions can be, with mineral-rich deposits of all sorts forming vigorously. High levels of carbon dioxide in the water and atmosphere would also have aided the formation of carbonate minerals.

Earth could have spent many tens of millions of years or even a few hundred million years in this state, with a dense greenhouse atmosphere and hot briny oceans (too hot for life). Even so, with a surface over 200C, it was inevitable that Earth would continue to cool. As it did so, another geological process came into play: subduction, the process by which crustal rocks sink into the mantle.

As Earth's surface cooled, the crust that had interacted with water now became dense enough to sink back into the interior, carrying with it its new set of hydrothermal minerals for recycling. Subtly different suites of minerals then began to erupt as the chemical sources for making new magma became more heterogeneous, each suite of minerals providing new permutations for weathering and new compounds as it weathered again on the surface.

Perhaps more important, the sunken crustal rocks would have been carrying away with them the carbon dioxide (and water) that they had accumulated by weathering reactions at the surface. This process, which stores these "gases" in the mantle for millions of years at a time, still continues, of course, but by now it is in a "steady state", with as much water vapor and carbon dioxide erupting out of volcanoes as sinks with sea-floor crust. On the early Earth, however, it was this process that took away the dense early atmosphere and its greenhouse effect. On Venus, which never cooled enough to have liquid water, that never happened; and on Mars and the Moon, the gases were lost to space from these smaller bodies before they could be stored in the mantle.

The process continued until Earth's surface temperature was comparable with today's Equator, round about 30C. Earth's interior was hotter than today's, and would remain so for a long time, but the insulation of the crust (and mantle) quickly brought heat flow through the crust to comparatively low levels.

If all this is true, then Earth would have become hospitable for life quickly in geological terms, with large, perhaps shallow oceans, temperatures well in the range we experience today, and with weathering processes acting between atmosphere, water, and rocks to release many different chemical compounds. Results published in 2005 suggested that water was present on the surface, implying rain, weathering, erosion, and sediment deposition as early as 4350 Ma.

The question is how quickly Earth reached a temperature at which life might have been able to exist. Calculations suggest that once crust started to cycle the greenhouse gases into the interior, a runaway cooling set in, and Earth cooled very quickly from over 200C down to about 30, the temperature of today's equator. The time Earth spent at around 100C would have been at most a few million years, probably around one million years, and possibly much less than that.

There is an important implication here: Earth did not spend much time in a temperature regime at which "thermophilic" bacteria thrive today. Thermophilic bacteria seem to be very "primitive" on many charts of bacterial evolution, and there are suggestions and scenarios for the origin of life on Earth under hot conditions, perhaps even at deep-sea vents. But the geological constraints discussed here seem to point to an early Earth that reached what we consider to be "normal" conditions almost as soon as it reached habitability.

This does not mean that the life we see today evolved before 4 Ga, because the great impacts around that time may well have sterilized the surface of anything that had evolved before that. The life that was ancestral to us likely evolved after 4000 Ma.

NASA news story, January 2006

Written 2001, updated March 2006.

[Return to Chapter 1]