Since quite early in the history of the discovery of planets around other stars it's been apparent that the likelihood of certain types of planets around a star is related to the abundance of heavy elements in that system. Specifically, astronomers can study the spectrum of light from a star and deduce the mix of hydrogen, helium, and heavier elements present in the stellar plasma. The assumption is that this mix is a fair representation of the original mix in the interstellar gas and dust from which the star formed. By extension it would be the mix in the so-called 'proto-planetary disk' (or circumstellar disk) from which any planets formed.
So far so good. Past data has shown that planets like 'hot Jupiters' - gas giants orbiting close to their stars - tend to occur in systems with a higher proportion of heavier elements, stuff like iron, oxygen and so on. In astronomer parlance these are systems with higher 'metallicity' - meaning more of the elements heavier than hydrogen or helium (don't ask, astronomy can sometimes be remarkably imprecise).
Flash forward to a new study published last week in Nature by Buchhave and colleagues. In a careful investigation of more than 400 planet-hosting stars (with a total of 600 planets on relatively small orbits) they find good evidence for three regimes of planetary 'type' that correspond to distinct grades of heavy element abundance in the stars.
Specifically: There are terrestrial-type (rocky) planets less than about 1.7 Earth radii in size, there are gas-dwarf planets with rocky cores and hydrogen-helium envelopes (mini-Jupiter's if you will, although not really the same) between 1.7 and 3.9 Earth radii in size, and there are ice or gas giant planets larger than 3.9 Earth radii. Remarkably the transitions between these regimes are determined by statistically significant variations in the heavy element mix of the stars, and by extrapolation the original planet-forming material. The more heavy elements there are, the more likely you are to find larger planets.
We don't understand the details of planet formation enough to know exactly why this should be, although the presence of more heavy elements presumably means an enhancement of building material during the agglomeration of matter around a young star. Or at least that's the standard rule of thumb. In the Buchhave et al. study stars hosting smaller, terrestrial-type, planets have an average heavy element abundance that's about 5% lower than that of our Sun. By comparison, stars hosting the larger ice or gas giant worlds have an average heavy element abundance about 51% higher than that of the Sun. These are not enormous differences, and there's a lot of variation - low 'metallicity' stars can still sometimes host large planets and high metallicity stars can host all sizes of planet.
And of course these are all planets on quite small orbits, circling their stellar parents in less than about 100 days, and often far less than that. We don't have a good handle on what the full architecture is of these systems, what sibling planets lurk out on larger orbits. There is however a clue in this present work, the data suggest that the mass of solid material needed to accumulate a thick hydrogen and helium atmosphere increases with distance from the star. In other words the transition size between rocky and gaseous planets grows the further from the star - and there could be more giant worlds out where we've not yet detected them.
But change and chaos can rule in planet formation and planetary dynamics, complicating and blurring the picture we see. It's likely to be a long while before we know how to fully unravel the history of any given system - including our own.