Though hobbled by age, NASA’s Kepler planet-hunting telescope is proving to be an almost inexhaustible engine of discovery. The observatory found thousands of new worlds before an equipment malfunction in 2013 slowed its planetary torrent to a trickle, but clever researchers have managed to squeeze remarkable new findings out of its vast trove of archival data. The latest, announced yesterday, is a technical tour-de-force: using 88 million measurements drawn from four years of Kepler observations, a team of astronomers has for the first time weighed a planet smaller than Earth, finding the world Kepler-138b to be roughly two-thirds the mass of Mars. The results appear in the journal Nature (Scientific American is part of Nature Publishing Group).

“Even the visionaries who developed the Kepler mission never imagined that Kepler data itself would measure the mass of planets like these,” says study co-author Eric Ford, an astronomer at Pennsylvania State University. According to lead author Daniel Jontof-Hutter, also of Penn State, the feat is akin to measuring the dimensions of a golf ball 10 million kilometers away.


The animation above shows the masses and radii (sizes) of 127 exoplanets in comparison to the planets in our own solar system, slowly zooming in from the scale of Jupiter-size worlds to display the dimensions of the Kepler-138 planets relative to Earth, Venus, Mars and Mercury. The crosses around each exoplanet illustrate the range of uncertainty associated about that planet's mass and size.

Kepler-138b resides in a scorching 10-day orbit around a dim red dwarf star some 200 light-years from Earth, accompanied by Kepler-138c and Kepler-138d, two known planets slightly larger than Earth. All three planets are too close to their star to be considered habitable, falling well within what would be the orbit of Mercury around our Sun. Each world “transits” the star as seen from Earth, casting shadows toward us that Kepler registers as minuscule dips in starlight. The bigger those transitory dips are, the larger the planets casting the shadows must be.

Transits typically allow astronomers to learn little more than a world’s size, but the transiting planets of Kepler-138 offer something more. They are a tight-knit trio, in orbits so close that each world’s gravity significantly, cyclically influences the motions of the others. This gravitational ebb and flow alters the timing of each planet’s transit, making them arrive early or late, like off-schedule trains. By analyzing these “transit timing variations” for all three planets, the team learned how forcefully each world was being perturbed, generating estimates for each world’s mass.

To appreciate the difficulty of these measurements, consider this illustrative analogy, courtesy of study co-author Jason Rowe of the SETI Institute: Imagine the Empire State Building with all its windows lit up and all its window shades open. Simply detecting Kepler-138b, c, and d is equivalent to seeing the partial lowering of three window shades across the whole building from about 100 kilometers away. Now imagine that each lowering window shade slightly alters the timing of the others, and that the three window shades only lower for minutes or hours at most a few times a month. That, in short, is why measuring the masses of the Kepler-138 planets required four years of observations.

All that hard work yields a great reward. Find a planet’s size and its mass, and you can guess at its density and composition, gaining a sense of that world’s true substance. In recent years, Kepler and other projects have shown that most planet-bearing stars harbor strange worlds that weigh more than Earth but less than Neptune. No such planets exist in our own solar system—these worlds are alien to us in every sense of the word. Hopeful astronomers called these planets “super-Earths,” imagining them to be larger versions of our own terrestrial globe, but the few with measured densities proved to be more like “mini-Neptunes,” lighter and puffier than any rocky world could ever be. Why and how this unpredicted population of planets exists in such abundance has become a central question vexing planet-hunters, particularly those searching for other Earths.

Exoplanets “seem to present an unending ability to thwart expectations,” says Greg Laughlin, an astrophysicist at the University of California-Santa Cruz who wrote a commentary to accompany the Kepler-138 study. “It’s clearer all the time that super-Earths are neither super nor Earths, and I think that essentially the same confounding trends will hold for actual Earth-mass planets.”

Kepler-138b, at least, seems to be relatively familiar. Its estimated density is similar to that of Mars—about what would be expected for a small, rock-dominated world. The next planet out from the star, the approximately Earth-sized Kepler-138c, is heavier but also familiar: with a density similar to Earth’s, it may well have a mixed composition of metal and rock, just like our own planet. But the outermost known planet, Kepler-138d, is less than half as dense, suggesting an unearthly world puffed-up with large volumes of water and gas. It is only familiar in following the alien archetype of so many other so-called super-Earths. Laden with delicate, volatile elements that tend not to linger in the broiling temperatures close to stars, Kepler-138d may have formed further out in the cold before spiraling down to its present toasty position.

All this variety, Rowe says, means “we cannot infer a planet’s bulk physical properties based on size alone,” particularly when a planet is close to the size of Earth. What appears as an Earth-like shadowy blip in a planet-hunting telescope’s detectors could, upon further investigation, prove to instead be an Earth-sized orb of iron, or a world made mostly of ice, or gas. Only by measuring the densities of a large sample of Earth-sized and smaller planets can astronomers hope to learn how often rocky worlds like our own arise, and where exactly the transition between rocky and gaseous planets takes place.

The tools to do exactly this will soon be in-hand. Two Kepler successors—NASA’s TESS mission, and ESA’s PLATO mission—will launch in coming years. Each will search for transiting worlds around many of the Sun’s nearest, brightest neighboring stars, providing plentiful small, potentially rocky planets for other facilities like the James Webb Space Telescope to examine in more detail. Before long, astronomers will have the weights of many worlds on their minds—and maybe on their shoulders, too, if the planets once again fail to fit their models.