The planet Mercury is oddly iron rich. So are some exoplanets. Scientists trying to work out how this happened have come up with a crazy theory. Starlight itself might be sorting out the solar system through a technique called electrophoresis.
When scientists started looking at the solar system, they noticed something odd. Mercury, the innermost planet to the sun, had more iron than it was supposed to have. At first, they thought that some kind of impact had blown away the outer layers of the planet, which would have mostly been silicates, leaving the iron-rich core. When they looked for all the other elements that would have been blown away with the silicates, though, they were right there on the surface of Mercury. And they were right there on the surface of iron-rich exoplanets. Whatever formed Mercury also formed other planets, and it formed them to have an unusual proportion of iron to silicates.
What formed Mercury, scientists think, might have been starlight itself, through a process called electrophoresis. Electrophoresis is already used in labs as a way of trapping or moving particles. When light hits a particle, it doesn't impart a lot of momentum, but it does heat up one side of the particle. Heat is the motion of molecules. If a molecule of gas were to hit the heated side, the side with molecules that are also hopping around, it would come away from the particle with more momentum than it would if it were to hit the particle on the cooler side. Because the gas molecule comes away with more momentum, it also imparts more momentum (in the opposite direction) to the hot side of the particle. So whichever side of the particle is heated gets harder pushes, and the particle moves.
It may seem like the particle will only be driven away from the light source, but electrophoresis can move the particle in either direction. If the particle is shaped just right, or has the right properties for transmitting the light, it can act as a lens, focusing the heat of the light on its far side. The side of the particle far from the light heats up, and the particle is driven, by gas molecules, toward the light source.
Given the huge distances in a solar system, and the relative scarcity of gas molecules, it would take a long time to drive the silicates away from the sun, and perhaps the iron towards the sun. But we are dealing with cosmic amounts of time. It's possible that starlight can organize a solar system.