An exploding cloud of aluminium-rich and silica-rich vapour could have been the result of a planetary body having suffered one of two processes: pressure release following shock compression in a collision, or a rapid rise in temperature occurring from within.

Within a multimillion-year timescale for solar system formation, the current approach is to imagine a sea of planetesimals slugging it out in repeated impacts until a few attain the rank of planet. And certainly impacts were a common occurrence. However, the oldest datable particles in the solar system – the condensates and melt droplets – suggest that bodies of planet rank existed prior to the era of destruction. Out of their disintegration arose much smaller bodies, which then spawned still smaller ones as they collided. In relation to the largest bodies, an impact scenario only commends itself by default, on the grounds that the heat generated by radioactivity, although much greater than now (both because there was much more radioactive matter and because some half-lives were much shorter), would not have been high enough to cause whole planets to burst.

As we shall see, however, radioactivity at the rates presumed was insufficient even to account for limited melting. If the rates had been higher, the radioactivity-based timescale would be shorter, in accordance with the meteorite evidence that the solar system did not emerge from a sea of accreting particles. Collisions would then be an unlikely cause of planetary explosions. The only bodies initially orbiting the Sun were the planets themselves, and they were too few and far between for collisions to occur, even if they could have somehow been disturbed from their orbits. In this situation, therefore, an internally generated catastrophic surge in temperature needs to be considered. Were higher rates of radioactivity perhaps themselves the cause? Shortly before the explosions, aluminium-26 in particular seems to have been abundant and could hardly not have been involved.

If heat deep within a terrestrial planet were to have increased more quickly than it could have been conducted to the surface and radiated away, the outer parts would have remained solid. Before long, thermal expansion of the interior would have ruptured the primeval surface, magma pouring up through the fractures and overwhelming the planet. This appears to be what actually happened. The oldest rocks of every terrestrial planet are igneous: Mercury, Venus, Mars and the Moon were once blanketed with magma to a depth of several kilometers. So was the Earth, only here the effects were mitigated by the simultaneous release of an ocean of subterranean water.

But the planets still in existence did not explode. Why not, since others did? Perhaps we should delve more fundamentally into the origin of the radioisotopes. We can rule out a supernatural origin for the short-lived isotopes, because they came into existence only shortly before the chondrules and CAIs in which they were trapped. But in that case, it is likely that the long-lived isotopes also had a natural origin, since whether a radioisotope is considered short-lived or long-lived is a matter of degree. As measured on the basis of present rates of decay, half-lives can have lengths of minutes, billions of years or anything inbetween.

The usual explanation for the presence of radioisotopes in the solar system is that they were all synthesised in supernovas. Long-lived isotopes such as uranium were generated by supernovas and injected into the nebula long before it collapsed into a star-forming disc; the most ephemeral isotopes came from a contemporaneous supernova. But as we have seen, that scenario has its problems, including the inability of supernovas to synthesise ¹⁰Be and ³⁶Cl. Moreover, if chondrules, CAIs and matrix all came from already existing planets, then this is an indication that unstable isotopes were synthesised within the planets themselves rather than anywhere outside the solar system. New elements would have formed by nuclear fusion and nuclear fission.

Thermonuclear fusion is the process that provides the energy for all stars, including the Sun. At their centre, temperatures are high enough to overcome the electrostatic forces that keep atoms apart and the pressure causes the atoms to fuse, thereby creating new elements. As a by-product excess binding energy is converted into heat energy. The Sun is a relatively small star, so the temperature at its core – around 10 million °C – is sufficient only to fuse hydrogen, the simplest of elements, into helium. More massive stars can go further and fuse helium into carbon, nitrogen and oxygen (in what is known as the C-N-O cycle). Typically the temperature required here is around 100 million °C.

Radioactivity rates are related to the speed of light. The higher the speed, the faster radioisotopes decay, because a faster speed of light results in a lower atomic mass and hence lower binding energy. Likewise the threshold above which nuclear fusion begins to take place is lowered. The natural disintegration of atoms and the natural synthesis of atoms are both controlled by the same factor. Thus, if the speed of light had once been much faster, the same temperatures that currently enable the Sun to fuse only hydrogen atoms would also once have facilitated the synthesis of carbon, nitrogen and oxygen.

Since we do not know exactly what elements lie at the centre of the Sun, the only evidence we have to go on is the abundance of elements in its atmosphere. Depending on the degree to which convection mixes the interior, these may or may not be representative of what lies within. We also need to be bear in mind that plunging debris from exploded planets (or planetesimals, if you prefer) would also be a source of heavier elements. Consequently, if only helium nucleosynthesis had ever gone on, we would expect atmospheric abundances for all elements other than hydrogen and helium to reflect chondritic abundances. But that is not the situation. There are three other elements that substantially exceed chondritic abundances: nitrogen (40.6 times more abundant), carbon (9.9 times) and oxygen (1.8 times) – the very elements that today are the product of nuclear fusion in much more massive stars. Relative to each other, they are also found in much the same proportions. Thus we have evidence that C-N-O nucleosynthesis did occur in the past, and that thresholds for thermonuclear fusion were once much lower than now. Heavier elements may also have been synthesised, as a result of processes requiring still higher energy levels, but because their density may put them beyond the power of convective currents to dredge them up, any such elements are likely to remain in the core, beyond detection. The possibility must therefore remain hypothetical.

With electrostatic forces being so much lower, nucleosynthesis may also have taken place in planets, but in terrestrial planets the paths taken would have been rather different, since, necessarily, these bodies already had a full complement of stable elements. A radioactive element might have been generated simply by capturing a neutron or knocking out a neutron. As with stars, the thresholds for nuclear reactions would have been proportional to size, with larger planets being more capable of nucleosynthesis than smaller ones. Consequently, the answer to the question why some planets exploded but not others is that it was the biggest ones that exploded – hence the large volume of rocky matter in, for example, the cores of the giant gas planets, far in excess of the volume comprised by Mercury, Venus, Earth and Mars.

Since the gaseous planets are much more massive than the terrestrial planets, fusion may also have gone on in their interiors. Jupiter’s composition is approximately 90% hydrogen and 10% helium. Being smaller, Saturn has a lower proportion of helium, around 6%. One of the greatest mysteries of solar system physics concerns the fact that both planets radiate twice as much energy as they receive from the Sun (Smith et al. 2007), one suggestion being that the excess heat arises from deuterium fusion. Given that the gaseous planets started as pure hydrogen bodies, however, the simplest explanation would be that the excess is residual heat from the fusion of hydrogen into helium.

Whatever the mechanism, the strength of the case for attributing chondrules to large-scale explosions is becoming widely acknowledged. ‘Most chondrules formed by impacts’ (Hutchison et al. 2005). ‘Chondrules and refractory inclusions are impact-melt objects’ (Sears 2005). ‘Chondrules and metal grains in the CB chondrites formed from a vapour-melt plume produced by a giant impact between planetary embryos’ (Krot et al. 2005). ‘Entire planetesimals were melted internally by the decay of ²⁶Al, and … collisions led to the splashing of their liquid contents. Cascades of molten droplets would thereby have been lofted into the nebula where they would have cooled to become chondrules.’ (Sanders & Taylor 2005) Such an interpretation of chondrules has become feasible as it has become clear that differentiated bodies – namely those that gave rise to iron meteorites and angrites – existed already before the formation of chondrules (Sokol et al., Burckhardt et al., Kleine et al. 2009). Chondrites do not represent the primordial material from which asteroids accreted and then differentiated.

Thus it is no longer controversial that chrondrules originated from mantle silicates that underwent catastrophic heating. Extreme pressure, whether generated internally or externally, caused the planet’s crust to burst. Unloading vaporised the exploding mantle, and as the gases dispersed they condensed into silicate-rich droplets. In outer space they coalesced with pulverised crustal material, shock-heated rock fragments (‘metamorphic clasts’) and water vapour to form a multitude of compact aggregates, the parent bodies of what we now call chondrites. Some of the fragments may have been 100s of kilometres in diameter. As they flew apart, some crashed into other planets, some went into orbit around them, and some formed a circumsolar belt in the region previously occupied by their progenitor. Radiogenic heating continued for a time within the new bodies.

There is thus an accumulating pressure of evidence to indicate that CAIs must also derive from planetary explosions. As yet, the idea is controversial, but it is now being discussed (Sears; Sanders 2009).