Skip to main content

A Trillion Worlds

Big-data statistics have revealed, among other things, that our own solar system is kind of an oddball

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


Times sure change. A little over two decades ago, the discovery of a new extrasolar planet was invariably treated as front-page news. In recent years, a newly discovered world elicits barely a mention, and simply adds another data point to the thousands of planetary confirmations that are accumulating in NASA’s Exoplanet Archive. Like many fields of endeavor, exoplanetology is transitioning to the realm of big data.

While this situation may be frustrating for planet hunters facing ever-diminishing accolades in exchange for their worlds, it is thrilling to watch the fine-grained statistical unveiling of the galactic planetary census. It is hardly an exaggeration to state that nearly everything found to date has come as a surprise.

In retrospect, the first exoplanetary system was discovered by Galileo in 1610 when he turned his telescope to Jupiter and noticed the movements of its four largest moons. Jupiter’s Galilean Satellites (Io, Europa, Ganymede and Callisto) form a compact, almost perfectly co-planar family of worlds on near-circular orbits that together add up to about two ten-thousandths of Jupiter’s mass, and have orbital periods ranging from a scant two days (Io) to more than two weeks (for Callisto).


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


This basic architecture, as defined by the range of orbital periods and the ratio of the satellites’ masses to that of the planet, is pleasingly replicated for the systems of moons orbiting Saturn and Uranus. But it doesn’t resemble the architecture of the solar system, where the largest planet has a thousandth of the sun’s mass, and where the orbital periods of the planets are much longer than those of the large planets’ satellites.

To astronomers’ lasting surprise, it is now clear that the average planetary system in the galaxy looks a lot more like the Jovian moons than like the sun’s planets. Of order half of nearby stars have at least one (and often several) worlds with masses substantially greater than Earth and orbital periods ranging from mere days to weeks; in our system, the space interior to tiny Mercury’s 88-day orbit is entirely empty.

The first hints of the ubiquity of compact, short-period planetary systems came 10 years ago, with the results from the HARPS spectrograph located in La Silla, Chile. HARPS employed the Doppler velocity technique to detect a profusion of low-mass planets. The results were so unexpected that they were met with skepticism. Doubts, however, were eliminated when NASA’s Kepler mission found hundreds of systems with multiple, transiting super-Earth–sized planets.

The transit technique and the Doppler velocity technique are complementary. The former measures a planet’s size, and the latter senses its mass. For many of the small planets found with Kepler, however, it’s difficult to get accurate masses with the Doppler method. The signals are very small, and the parent stars often generate confounding noise in the spectra.

Fortunately, however, transiting planets on tightly spaced orbits often show deviations from clockwork-like regularity. The ebb and flow of gravitational attraction between the planets causes the transits times to vary, and by analyzing the deviations, the masses and orbital details of the planets can often be worked out. If one has both a mass and a size measurement for a planet, then one knows the density, which gives a key insight into the planet’s structure and what it is made of.

Density estimates now exist for well over a hundred planets with masses below 30 Earth masses, and the aggregate of results is bewildering. In a nutshell, planetary densities are all over the map. If you know the radius of a planet (but have no other information), you have virtually no basis for guessing its mass. Among planets that fall in the size range between Earth and Neptune (which has four times Earth’s radius), the range of masses at a given radius can vary by over a factor of 10. This huge variation points to a startling range in compositions that must run the gamut from dense worlds made of iron to puffy spheres that include ample supplies of hydrogen.

Very recently, however the confusion induced by the exoplanet density diagram has started to lift. Clarity is emerging from the extensive use of large ground-based telescopes to better measure the properties of the transiting planets’ parent stars, particularly their radii. Accurate knowledge of the star’s size leads directly to an improved handle on the size of its planets.

It is now clear that the distribution of planet sizes contains a distinct double-peaked shape. Planets with radii of about 1.75 times that of Earth are substantially less common either than planets that are slightly bigger than Earth or than those with 2.5 or more times Earth size. This planetary “radius gap” is telling us something important. Most likely, it delineates a fairly sharp transition between roughly Earth-sized worlds that are fundamentally terrestrial in character and planets with deep hydrogen-helium atmospheres that are hotter, smaller cousins to Uranus and Neptune.

A separate, and even more intriguing insight was elucidated by a team led by Lauren Weiss, a postdoctoral researcher at the University of Montreal, who established the existence of a curious organizing principle for the sprawl of planetary masses and sizes. The planets within a particular system tend to have very similar sizes (and masses) and tend to adhere to a geometric uniformity of spacing, much like the arrangement of peas in a pod.

While there are hints of this phenomenon among the sun’s planets (Earth and Venus present a uniform pair, as do Uranus and Neptune) it is much more pronounced among the systems discovered by the Kepler mission. The origins of this curious intra-system uniformity are as-yet completely unclear, and they provide an interesting constraint on the planet formation process. That is, a mechanism must exist to coordinate the growth and spacing of planets within a given system, even as these attributes vary wildly from star to star.

Astronomers are on the cusp of adding very substantially to the planetary data trove. NASA’s TESS Mission, the successor to Kepler, was successfully launched in April of this year and is currently monitoring stars for transiting planets. During its two-year prime mission, TESS will survey the entire sky, sifting over a thousand new planets from a sample of roughly 200,000 bright nearby stars.

Simultaneously, a slew of new high-precision spectrographs for making Doppler velocity measurements are coming online. A recent conference featured presentations from 23 separate teams, many of whom described instruments (such as EXPRES in the Northern Hemisphere and ESPRESSO in the Southern) which are already taking data, and which will reach the velocity precisions of order centimeters per second that are required to sense Earth-like planets in Earth-like orbits.

The Milky Way contains over a trillion planets. While we are nowhere remotely close to a full accounting of this vast population, the outlines of the distribution, and how it came to be, are rapidly falling within our grasp.