Astrophysics: The Planet Search

Image

Tim Morton

Vera Gluscevic

Princeton astrophysicist Tim Morton has confirmed the existence of more than a thousand new planets outside our solar system.

Since it was launched in 2009, NASA’s Kepler spacecraft has identified thousands of “planet candidates” outside our solar system, leaving scientists mostly to verify their identity one by one. Thanks to a technique developed by Morton, an associate research scholar, scientists this spring confirmed in one swoop that almost 1,300 of the objects in NASA’s latest catalog were indeed planets — all without looking into a telescope. The finding nearly doubled the number of planets found by Kepler that might be able to support life.

Morton developed a new computer model, called Vespa, to weed through and analyze the Kepler data, allowing scientists from Princeton and NASA to determine which of the signals recorded by the telescope were actually planets and which only looked like them. Their results were published May 10 in The Astrophysical Journal.

Morton’s findings include nine new “Goldilocks” planets, so-called by scientists because they are neither too hot nor too cold to support life

“False positives” — objects that appear to be planets but are not — have bedeviled astronomers since they started tracking the known number of planets in the universe. To find planets, astronomers look for very faint dimming of stars that occurs in regular periods, indicating a planet’s orbit around the star. Ground-based telescopes, such as those used in Princeton’s HATNet Exploratory Survey, find that 60 to 70 percent of the signals they observe turn out to be something other than a planet — usually another star in a binary star system.

As a grad student at Caltech in 2011, Morton was the first researcher to show that the space-based Kepler telescope was much more accurate in its predictions. Partly due to lack of interference from the sun and the atmosphere, fewer than 10 percent of its signals were likely to be false positives. The question, however, was: Which 10 percent? “Even if you know that 90 percent of them are true planets, you’d like to know which ones of them they are,” says Morton. 

His new formula works in two parts. The first looks at the part of space in which a planet is found, comparing it to known data about binary stars and other planetary phenomena to determine the probability a planet is real. The second part of the formula examines the characteristics of the signal, including the depth and duration of the dimming and its shape on a graph over time. “A typical planet signal goes from full brightness to partial pretty quickly, making a very boxy-looking signal,” says Morton. For most binary stars, he says, “it’s a more gradual dimming, making it look more like a V.”

When he applied his formula to a database of some 4,000 unconfirmed signals from the Kepler mission, Morton was able to validate 1,283 of them as planets with at least a 99 percent degree of certainty — the bar that NASA sets to qualify a new planet. That discovery more than doubles the number of planets Kepler has found, to 2,325, giving astronomers studying planets a much richer set of data for their work. Just as important, Morton’s formula assigns a planet-probability figure for each signal, so researchers can expand and analyze their data. 

Perhaps most excitingly, Morton’s findings include nine new “Goldilocks” planets, so-called by scientists because they are neither too hot nor too cold to support life, bringing the number of potentially habitable planets found by Kepler to 21. It’s a particular goal of the Kepler mission to find such planets, which usually are extremely difficult to verify from the ground. “To say how common these inhabitable planets are, you have to use these false-positive probability results,” says Morton. “Now we can put a number on those candidates to say whether one is in or out.”