Kepler Is Dead. Long Live Kepler
NASA’s planet-hunting Kepler spacecraft succumbed to mechanical failure, but there is more data—and perhaps some life yet—for the mission
Hi, I’m Stephen Cass for IEEE Spectrum’s “Techwise Conversations.” It’s pretty easy to see that we live in a galaxy composed of hundreds of billions of shining stars. But it’s not so easy to see if planets like our own Earth orbit any of those stars. To get a handle on answering that big “if,” in 2009, NASA launched a space telescope called Kepler. Kepler’s goal was to sample the exoplanet population in our galaxy and give us a sense of whether or not Earth-like planets—the kind of relatively temperate places that could cameo in an episode of “Star Trek”—are actually pretty common.
Kepler’s mission came to an end earlier this year, but new planetary discoveries are still rolling in, and the telescope might actually still have some life in it yet. To help sort it all out, I’m here with IEEE Spectrum Associate Editor Rachel Courtland. She covered Kepler’s technical trouble and has a soft spot for pretty much everything exoplanetary.
Rachel, welcome to the podcast.
Rachel Courtland: Thanks very much for having me, Stephen.
Stephen Cass: So bring us up to speed. What’s going on with Kepler?
Rachel Courtland: Well, at the moment not too much. Right now the telescope’s in what’s called a point rest state. That’s basically just a very stable, fuel-saving orientation, where the spacecraft nods back and forth in the solar wind and pulses its thrusters occasionally to keep it in line. And it’s been that way since August. That’s when NASA engineers made their last attempt to get the spacecraft going again after they ran into some technical issues earlier this year.
Stephen Cass: What kind of technical issues are we talking about?
Rachel Courtland: Issues with the spacecraft’s reaction wheels. These are electrically driven flywheels that are used for orientation. They’re mission-critical because Kepler was designed to look for planets by basically staring at one patch of sky— the same 150 000 or so stars—for pretty much four years straight. To find planets that way, the spacecraft has to be very stable, because it’s looking for brief, tiny drops in brightness that happen when a planet blocks out a bit of light as its orbit takes it between its host star and us.
Kepler was built with four of these reaction wheels, and it needed at least three to keep it stable. But in 2012, one of them started showing elevated levels of friction and was taken offline, and a second one failed in May 2013.
The team gave the wheels a rest, and in August, they made one last attempt to get a third wheel online. I actually caught up with Kepler mission manager Roger Hunter not too long after to get the play-by-play, so I think I’d just let him explain what happened in his own words.
Just to warn you, he’s going to throw out some reaction-wheel names. Kepler now has two good wheels, which are wheels number 1 and 3. And there are two bad ones, 2 and 4. Here’s Roger Hunter explaining what happened in August, when they tried to boot up the better of the two bad reaction wheels, which was number 2.
Roger Hunter: We could still tell then that reaction wheel 2 still had signs of elevated friction on it. And one of the first things we wanted to do after we got the three wheels turned up was get it into what is called Earth point, which meant that we had to maneuver the spacecraft to point the high-gain antenna directly at Earth so that we could download the data that had been on the solid state recorder on the spacecraft since science operations ended back on May 15th, when reaction wheel number 4 failed.
Rachel Courtland: So, this actually worked.
Roger Hunter: We were able to get all of the science data down, and it took, you know, a few hours to get this done.
Rachel Courtland: But then they ran into trouble. They were getting ready to point Kepler back at its target patch of stars, when reaction wheel No. 2 started giving them trouble again. They started to see some increase in the resistance in the wheel and then some erratic movement.
Roger Hunter: That’s when the spacecraft kicked itself into safe mode, in other words as a self-protective measure. By then we could tell that reaction wheel 2 was failing again…as a matter of fact, it had started spinning down to zero, which meant the friction had started increasing and we could no longer turn the wheel.
Rachel Courtland: And that was it for Kepler’s mission.
Stephen Cass: Do they know what happened to the reaction wheels?
Rachel Courtland: Hunter says it’s hard to say for certain. But it seems like the bearings inside the wheels essentially wore out and fractured. But Kepler is millions of miles away, and the telemetry can only tell you so much, so they can’t be sure.
Stephen Cass: So I guess that’s it then for Kepler?
Rachel Courtland: Actually, maybe not. Here’s Roger Hunter again:
Roger Hunter: A lot of people here are still optimistic, because we’ve been asked by NASA headquarters to not throw Kepler away but to see if we can repurpose it and still use it. We have a photometer on this telescope. It’s an amazing photometer; it’s one of the most precise instruments ever built. We have 96 million pixels up there that we want to put to work.
Rachel Courtland: In the months since the last attempt to get Kepler back up and running, the team has come up with a plan to continue using the spacecraft with just two reaction wheels. Kepler can’t point at its original patch of stars, because sunlight hits the solar panels differently at different points in orbit, and it pushes the spacecraft around. But Kepler could be fairly stable if it looks at stars just along the ecliptic, which is the plane that the solar system’s planets orbit in.
Stephen Cass: What will that do?
Rachel Courtland: The team thinks it should cut down on the instability caused by the sun and let the spacecraft stare at patches of stars for up to 80 days. That should turn up dozens if not hundreds more planets.
Stephen Cass: So when will Kepler start doing that?
Rachel Courtland: Well, it’s not clear whether it will, actually. The Kepler team’s proposal still needs to be reviewed by NASA, and the spacecraft will have to compete with other proposals for funding. NASA may decide that an extended Kepler mission is not a top priority. But even if the plan doesn’t get approved, this isn’t the last we’ll hear from Kepler.
Stephen Cass: Oh, no? Why is that?
Rachel Courtland: Well, when Kepler failed, the spacecraft had already finished collecting four years of data. But astronomers aren’t anywhere close to having analyzed all of it. In fact, they only just recently announced the new planetary candidates they were able to identify in the first three years of data.
In fact, one of the most exciting results out of Kepler came out just last month. A team of astronomers found that, by looking at nearly all four years of Kepler data, they could actually estimate how many Earth-sized planets in the Milky Way are in the habitable zone of their host stars.
Stephen Cass: Before we dive into those results, maybe you can remind us what astronomers mean by habitable?
Rachel Courtland: Sure. It doesn’t have anything to do with building codes or working plumbing. Right now, habitability comes down to temperature. The idea is that each star has a region around it that could support life as we know it. It’s often called the Goldilocks zone, because it’s the area around each star where the temperatures are not too hot and not too cold but are in just the right range for there to be liquid water on the surface.
Erik Petigura, the graduate student who led the analysis, put it very nicely at the press conference announcing the new results:
Erik Petigura: The habitable zone depicts the range of orbits where liquid water could plausibly exist on the surface. Inside, liquid water boils off into steam, and outside, oceans are frozen solid.
Rachel Courtland: Now, it’s probably worth mentioning that temperature alone doesn’t tell you everything you need to know about whether life can survive on a planet. It’s still uncertain how big a planet can be and still have a nice rocky surface to walk on. There’s also debate over the requirements for plate tectonics, which helps keep our atmosphere in balance. But temperature is a pretty good starting point, since Kepler can only really tell you two things about a planet: how big it is and how far away it is from its host star.
Stephen Cass: All right, so let’s talk numbers. How many habitable Earth-sun pairings are there?
Rachel Courtland: So, when it comes to systems like our own, ones with planets about the size of the Earth orbiting stars that are more or less like the sun, the magic number is 22 percent.
Stephen Cass: 22 percent.
Rachel Courtland: Right. Basically, one in every five stars like the sun has a small, potentially rocky planet in that Goldilocks zone. It’s tough to translate that into raw numbers, because no one knows how many stars there are in the Milky Way. But if you assume a common estimate, about 200 billion, there should be about 50 billion sunlike stars in the Milky Way. Eleven billion of those will have an Earth-sized planet that could potentially have oceans or lakes or fjords.
Stephen Cass: Wait a second, though. I do remember talking with you about some of the news stories that came out after the press conference, and you mentioned the number of Earth-sized planets they actually found in the habitable zone was pretty small.
Rachel Courtland: Yeah, that’s right. They only actually saw 10 Earth-sized planets in the Goldilocks zone.
Stephen Cass: Hmm. That seems like a pretty big jump to go from 10 detected planets to estimating that there are 11 billion of them out there. How do they know they’re making a reliable extrapolation?
Rachel Courtland: That’s a really good question. And part of the answer is that they didn’t look at just a handful of stars—they were looking at about 40 000 that are the same type as our own sun. But you’re totally right—you can’t just count up the planets that you detect and say that applies to the entire stellar population. When you do that, that 10 Earth-size planets in a pool of 40 000 stars actually translates to a really small percentage, just a quarter of a tenth of a percent.
Here’s Erik Petigura again, explaining the problem:
Erik Petigura: Taking a catalog of planets is one thing, but we want to understand what the underlying distribution of planets is. You can think of it like we’re doing— we’re taking a census of extrasolar planets, but not everybody is answering the door.
Rachel Courtland: So there are two ways that a planet doesn’t answer the door. One is a matter of orientation. Kepler can only see planets if their orbit crosses in between their host star and our own. That’s what Kepler’s looking for—small dips in the brightness from those transits.
But the thing is, there’s no special orientation. All around the Milky Way, planets are orbiting their stars at all sorts of angles, and Kepler can only see the orbits that block out starlight along our line of sight. This alignment issue was a known problem, and since it’s just a matter of geometry, it’s a pretty easy one to account for when you’re trying to estimate the overall abundance of planets.
But there is a second problem that is much harder to deal with. And it sort of took the Kepler team by surprise.
Stephen Cass: Okay, you’ve piqued my curiosity. What is it?
Rachel Courtland: Well, before I come out and tell you, let me start by playing this. Just hold on one second while I queue it up. Okay, here it goes:
[eerie star sounds]
Stephen Cass: Okay. What is that?
Rachel Courtland: That is one of the stars that Kepler looked at. What you’re hearing is basically what happens when you take brightness variations in the star and transform them into something we can hear.
There’s actually a lot of noise in stars: Their surfaces brighten and dim because of internal vibrations; there are sunspots and solar flares and coronal mass ejections. The Kepler team expected to see this kind of activity. But sunlike stars turned out to be, on average, noisier than expected. And that’s a problem because the noise takes the form of brightness variations, which is just what Kepler was looking for when it was hunting planets.
Stephen Cass: That does seem like a problem.
Rachel Courtland: So that’s where some handy analytical tricks come in. To figure out how many planets they might be missing because of this haze of stellar noise, the team inserted some fake planetary signals into the actual data. Then they used their normal analysis techniques to see how many of those fake planets they could actually find. Since they knew how many planets they fed into the system, they could measure their detection rate and say, for however many planets they might find, how many might be missing.
Stephen Cass: And now they have this fabulous number that tells us that all those “Star Trek” episodes showing Starfleet officers wandering around on alien planets might not be so far off. That seems like a pretty good high note for Kepler to end on.
Rachel Courtland: Oh, well, it’s not over yet. The Kepler team has only just started the process of analyzing all four years of data, and they’re working on ways to better pull real signals out of the noise. This could better the chances of finding planets on Earth-like orbits around sunlike stars, because they’re rare events: They’ll only block out the light from their hosts roughly once a year.
Stephen Cass: It sounds like whether the spacecraft gets a new mission or not, we haven’t heard the last from Kepler.
Rachel Courtland: That’s definitely true. Astronomers tell me Kepler will have quite a legacy; that it really transformed planet finding from a serendipitous kind of thing into a systematic, almost big data–like endeavor. Erik Petigura told me that so far, astronomers have only been able to skim the cream off of what Kepler has to offer and that we’ll be hearing exciting news from Kepler for years to come.
Stephen Cass: Well, thanks very much for joining us, Rachel.
Rachel Courtland: My pleasure, Stephen.
For IEEE Spectrum’s “Techwise Conversations,” I’m Stephen Cass.
This interview was recorded 12 December 2013.
Audio engineer: Francesco Ferorelli
Read more “Techwise Conversations,”follow us on Twitter,or find us in iTunes.
NOTE: Transcripts are created for the convenience of our readers and listeners and may not perfectly match their associated interviews and narratives. The authoritative record of IEEE Spectrum’s audio programming is the audio version.