Reading Starlight: In Conversation with Michelle Kunimoto - Part II
Interview with Michelle Kunimoto
The Blue Hour — CiTR 101.9 FM
Recorded live in studio at CiTR 101.9 FM at the AMS Nest, University of British Columbia.
Listen to the interview on CiTR
The Blue Hour airs live every Tuesday at 2 p.m. on CiTR 101.9 FM and citr.ca.
Transcript lightly edited for clarity while preserving the natural rhythm of the live conversation.
Read Part I — Reading Starlight.
Transcript
Farha Guerrero: And I can’t help myself with jumping right into the idea of rogue planets because that’s something very TESS — if I’m not mistaken?
Michelle Kunimoto: Well, so I believe my team are the only people in the world who have actually done a rogue planet search with TESS.
It’s a very new idea because TESS was never designed to search for rogue planets.
But there’s going to be a mission launching later this year, the Nancy Grace Roman Space Telescope. It’s the next flagship mission out of NASA, and searching for rogue planets is going to be a really big goal of that.
And so I had an idea with a few collaborators a few years ago to use TESS as a precursor to Roman to understand how to design rogue planet searches with space-based telescopes.
Farha Guerrero: Now you’ve got to define what rogue planets are because I was very excited when I was listening to you talk about rogue planets in one of your lectures,— and there’s an acronym too, FFP.
Michelle Kunimoto: So rogue planets are also called free-floating planets, and those are those planets that don’t orbit any star that I mentioned earlier.
So depending on your definition of exoplanet, they are not exoplanets because they don’t orbit another star. I personally consider them exoplanets, but not everybody agrees.
Farha Guerrero: Why do you consider them exoplanets?
Michelle Kunimoto: Because their origins are probably consistent with having been exoplanets at one point.
So there are kind of these two major origin scenarios for how rogue planets can exist.
And one of them is that they originated in a planetary system like other exoplanets, like our own system. But then in the early days of planet formation, it was much more chaotic. You had planets crashing into each other. You had them scattering. You had all sorts of things, and they might have been ejected from their system, so that they’re now no longer gravitationally bound to any star.
So because they were at some point an exoplanet, I don’t know, I feel like...
Farha Guerrero: You’re acknowledging what you believe is their origin.
Michelle Kunimoto: Yeah. But then there’s another class of these rogue planets — some of the bigger ones might have just been something that was starting to form into a star but never got quite massive enough to reach the phase of nuclear fusion.
So it was kind of like a failed star, so it remains this planetary mass object.
So in that respect, I kind of see not calling that an exoplanet. But as long as it’s a planetary mass object, I consider it an exoplanet.
Farha Guerrero: Okay, so go back to where you were. So they’re free-floating, they’re not orbiting any star, they’re quite unique, and this is like a byproduct. This is not what the TESS mission, as you said, was intended to find.
However, this is something that you have a vested interest in.
And so what do you envision and what do you imagine that you’re going to find as you keep investigating this more?
Michelle Kunimoto: Yes, I’m actually pursuing this with one of my students right now, so I’m really excited about it.
So the idea behind the search in the first place was TESS is not designed to find these types of objects. It’s designed to find transiting exoplanets, right, with the transit method.
But you can find rogue planets using the same data set, but a different technique.
So this is a technique called microlensing. It’s unfortunately one of the harder methods to explain because it involves General Relativity.
The analogy I can give is imagine you’re looking at a really faraway light bulb, and somebody holds a magnifying glass and passes it between you and the light bulb. And as soon as that magnifying glass lines up just right, the light from that light bulb in the distance is going to get brighter, right? Because you’ve got a magnifying glass in front of it.
And then the brightening effect will end as the magnifying glass continues on its way.
So in this sense, that background light bulb is a background star. The magnifying glass is the rogue planet passing in front of that distant background star, and the magnifying effect is caused by the planet’s gravity bending light around it.
And so we have what’s called a lensing effect, where the light from that background star will temporarily be increased multiple times.
So we can use the same data sets that are measuring brightness over time, but instead of looking for these periodic decreases of brightness, we’re looking for these one-time increases of brightness.
So the idea was, TESS is taking data every 200 seconds, and some of these events caused by a terrestrial-mass free-floating planet might only last for an hour or so, so we have enough data over a given hour to see these events happen.
And because TESS is looking at a million stars every month, these are extremely low-probability events to have such an alignment that this effect happens, and so you need to look at a lot of stars.
And so because TESS is doing this for hundreds of millions of stars over the whole TESS mission, surely this will have happened at least once.
So we did some predictions, and we found that there have only been three free-floating planets of terrestrial mass ever discovered across decades of searching, and so we really don’t know how common they are.
And so depending on how common they are, maybe we would expect one detection, maybe we would expect 20 detections, maybe we would expect to see nothing.
But at least if we try to search, we could maybe constrain that a little bit better.
So what happened is we picked just a random month of TESS data. It was a completely arbitrary choice about which subset to use, and this was about a million stars in this data set.
And we searched it, and we found something that looks like the lensing event from a free-floating terrestrial-mass planet in our very first sector.
So there’s a few explanations for this. Either rogue planets are significantly more common than anybody would have expected, either we were extremely lucky to have found this, or it’s not a rogue planet. It’s something else.
So there are, just like how the transit method has false positives, the microlensing method does too.
Asteroids crossing the field — light from the Sun will bounce off the asteroid into the telescope and cause a brightening event.
But what the microlensing community seemed to be most concerned about was that this is a flare.
So stars are magnetically active, and our Sun goes through these cycles, and sometimes it’s more active than not. We just got through a really big peak in magnetic activity from the Sun, which is why we had all those aurora borealis that you could see. That was basically caused by a flare from our Sun and all the solar wind hitting the Earth’s atmosphere and causing these auroras.
This is basically a one-time increase in the brightness of the star that might look like a microlensing event.
These flare events are usually very asymmetric. They’ll have a really quick increase followed by a very gradual decay, whereas microlensing events are very nice and symmetric.
So we initially didn’t think that this was a false positive.
But when we posted our paper online for other people to read, that was what the microlensing community was most concerned by. But what’s interesting is that before our paper, these had never been considered as a false positive.
And I talked to one of the lead mission scientists for the Roman mission’s microlensing search, and he said, “Yeah, the paper that you posted — the fact that you found this event — is causing everybody to rethink how to do microlensing searches.”
Farha Guerrero: That’s got to make you feel a sense of satisfaction. This is groundbreaking.
Michelle Kunimoto: I’d like to think so.
I think it’s just because it’s an event that has confused a lot of people. Like I cannot say that it is a rogue planet. I also can’t say it’s a false positive. We don’t know.
And so some people are totally adamant — no, it’s not a rogue planet at all — and we’ve had a lot of responses from both sides that have been both positive and negative.
Farha Guerrero: But the thing is that the idea exists, and that that is exciting. That is really novel. And that has to keep you coming back to the office because that is something you really want to get after.
Michelle Kunimoto: Yeah. So one of my undergrads is actually going to continue the search and do it for the full TESS data set.
So I mentioned that this was based off a search of a single month. We then searched another five months and didn’t find anything.
So maybe we really were just lucky, but now we’re going to do all roughly 90 months of data that we have.
So we have potentially a lot more of these events to find, and if it turns out that this still remains the only possible free-floating planet, then that actually makes it more likely that it’s a real rogue planet, because it matches our predictions for the number of detections over the full TESS data set.
Isa: And for rogue planets, do they just drift with the expansion of space, or do some of them — because you said how you think some of them are formed is from other planets hitting — and I know there’s nothing to slow anything down in space, would some of them still be moving, or do they just drift with the expansion of space?
Michelle Kunimoto: So they would all still have some velocity from their initial ejection, and some of them might even be captured eventually by other systems.
So how do we know when we find the transiting exoplanet, for example, there’s a chance that this never even originated from the system that we’re observing. It was captured at some point.
So we’re not going to be able to tell that answer from the planets we see, but what people have done is simulated how often do we think these objects are being ejected. If we model the early solar system, do we think that this was a common process? Was this a rare process? What are the typical size distributions of objects that are getting ejected?
And we actually predict that free-floating planets outpopulate our bound planets by about 10 to 1. So there’s rogue planets everywhere floating around.
Isa: And also, I know it’s not that uncommon for other things to go into another solar system. I think it’s Oumuamua — I don’t know how to say it — but I know that was the first thing we discovered that was from another solar system that reached ours.
Michelle Kunimoto: So ʻOumuamua was the first interstellar object that came through our solar system in 2017. There was 2I/Borisov — I might want to double-check the name later — and then the one that was most recent in the last year was 3I/ATLAS.
So we’ve had three interstellar visitors so far. They’re all probably comets that were ejected from their original systems that have come through now.
But of course, those are the only ones we’ve detected. For all we know, there are several more that have passed through the solar system and we just didn’t have the images at the time to be able to see them.
But we think that this ejection process is really common, especially for small objects, and that’s why we think these terrestrial-mass free-floating planets are probably dominantly caused by this process.
Farha Guerrero: That is just absolutely fascinating.
And this Roman Space Telescope is going to be launched very soon.
Michelle Kunimoto: On my birthday.
Farha Guerrero: Amazing.
Michelle Kunimoto: Yes. As always with these big missions, they might get delayed. They have a projected launch date of September 7th of this year, so that would be a great birthday present if it launches and it works perfectly and everything goes smoothly.
Obviously it might get delayed, but yeah, it will be very soon.
Farha Guerrero: That’s really exciting.
Now I want to talk about the idea of finding planets that are similar to ours, that are habitable, about Earth-like planets and that big question of “Are we alone?” which everyone wants to ask people like you.
But there are a few things that UBC does here — a UBC-led TESS faint star search that you are a big part of, and you’re hoping to attract more students and scientists like yourself.
And as well, something that’s really cool called LEO — Lazy Exoplanet Operations. This is a way to distinguish planets from some of this kind of noise out there and stellar variability, et cetera.
So if you could just quickly touch on that, and then maybe we can talk about the Habitable Worlds Observatory that NASA has projected to launch in the 2040s.
Michelle Kunimoto: Sure.
Yeah, so when I worked on the TESS mission, I was the lead of MIT’s planet search team. So this was called the Quick Look Pipeline, and the QLP’s job was basically to take all million stars that TESS observed every month, make light curves — those measurements of brightness over time — and then search for those transiting exoplanets using various algorithms to make it automated, because we can’t manually go through a million objects.
But the resources that are needed to do such a big search — to search a million stars in a given month — you have to be finished before the next month’s data becomes available.
Essentially, we didn’t have the algorithms that could do this completely automated, and so we had to restrict ourselves to only about 10% of that sample.
So we made light curves for everything, but we only searched about 100,000 stars.
So as part of my postdoc project, I was thinking, well, I just finished my PhD at UBC. I designed planet search pipelines. I designed practices to make this very automated. So what if I take the tools that I developed as a PhD student and make this a more automated process?
So that turned into what you just mentioned, the faint star search.
Essentially those 100,000 stars that TESS looked at — that QLP looked at — were all the brightest stuff, like the highest-priority stuff, but that left 90% of the stars not looked at for planets, so the fainter stuff.
So I basically designed a bunch of new code bases, ran my planet search on this, and out of the roughly 8,000 TESS planet candidates that have been found so far, that has found about half of them. I think we’re almost to 4,000.
And this is something that I continue to lead out of UBC. And I continue to collaborate a lot with the TESS team to make sure that my candidates are alerted as official TESS Objects of Interest so that other people around the world can start following them up and potentially confirming them.
But yeah, as compared to all the other search teams out of MIT and NASA, this is actually the most successful in terms of number of candidate detections.
Farha Guerrero: Faint star search — what a great description. That’s what you’re doing.
Michelle Kunimoto: Yeah, it’s pretty straightforward.
So as part of the faint star search, we have to try to automate the process of finding planets.
And one of the most important things in this process is once we find a potential transiting periodic signal, we have to distinguish planets from brown dwarfs, from stars, from stellar activity. That’s where LEOvetter comes in.
So LEO-Vetter was my answer to this whole problem of how to automate this process.
It essentially does what’s called vetting. So it takes the planet candidate signal that we see in the light curve, and then it runs a bunch of tests on it, and then it tells me if it’s most likely a planet or some false positive.
I came up with the name because Leo is the name of the sports mascot for my favourite sports team, the BC Lions — Leo the Lion.
And so I came up with what the L, the E, and the O stand for after I came up with the name. So: Lazy Exoplanet Operations.
Yeah, we love our acronyms in astronomy.
So LEO-Vetter is a publicly available code base. It’s on my GitHub. Anybody can download it and use it for their own planet searches.
And I published a paper on it last year, and I know that there are several teams around the world who are already using it.
So it’s been very exciting collaborating with them and hearing that it’s been useful for the TESS community.
Farha Guerrero: Amazing.
Isa: I just wanted to ask, if most of the exoplanets that we’ve discovered — can you see their stars or their systems with the naked eye?
Because I go stargazing a lot, and obviously I don’t find exoplanets like you. I just look at constellations. But I’m just wondering if any of them that have been discovered or are notable or similar to Earth — can you see them with the naked eye, or are they mostly only seen with a telescope?
Michelle Kunimoto: Most of the planets we’ve found would only be accessible with a telescope, but there is a lot that you can see with the naked eye.
In fact, a lot of the planets that have been found around stars observed with the radial velocity technique — those are going to be really bright nearby stars. You can definitely see those with your own eyes.
Obviously you wouldn’t be able to see the planet because it’s so tiny and faint. The stars that you see will have exoplanets around them.
Isa: And my other question is, I don’t know if you’re familiar with it — it’s called SpaceEngine. It’s a game, and it has the entire universe.
And I actually searched up one of the exoplanets you found, and they had it in there.
Michelle Kunimoto: Oh, that’s hilarious. I’ll have to look that up.
Isa: It was developed by an astronomer. It has theoretically the whole universe. Because it’s mostly procedurally generated, you can search for planets, and and I looked up the ones that you discovered, and one of them was there.
Michelle Kunimoto: Thank you, yeah. Thanks for telling me that.
I’m also a really big fan of the Mass Effect series from BioWare, and they made some remastered Legacy Edition more recently, and you can see this galaxy map of various famous planetary systems that have been discovered, and they added a lot of the ones from Kepler.
So there you go — clearly some very big science fans on the development team for Mass Effect.
Farha Guerrero: So finally, a little bit of an answer to this burning question: if we find an exoplanet that is habitable, that we can live on one day in our future. And this incredible observatory by NASA is in the works. Maybe in the 2040s potentially it will be launched?
Can you talk briefly about it?
Michelle Kunimoto: Right. So the Habitable Worlds Observatory, as you said, will hopefully be launching in the 2040s.
And so we as the astronomical community are currently in the mission design phase.
And one of the most important questions is to say: how big of a telescope do we actually need?
Because the purpose of this telescope — one of the main purposes — is going to be to directly image other Earths around other Suns.
So currently a lot of the methods I’ve talked about, like the transit method, radial velocity, microlensing — they’re all indirect methods of detection. We’re not seeing the planet itself.
But direct imaging is literally taking a picture of a star, trying to block out the light from that star as much as possible so that you can see the fainter objects, like planets, orbiting around it. Literally taking a picture of another exoplanet.
So we have been able to do this for really big planets that are really far from their stars, but we’ve never been able to do this for something that looks like our solar system — relatively close-in Earth-like planets.
So the Habitable Worlds Observatory will be the first telescope capable of doing this.
But in order to design that, we need to know how common other Earths are in our galaxy. Because if they’re all over the place, then we can pretty much pick any star that looks like our Sun, observe it, and we should be able to see an Earth-like planet.
But if they’re really rare, we have to think a lot more carefully about how many stars we want to follow up, which ones are the most likely to host a potentially habitable planet.
And so one of the research projects that I’ve worked on in collaboration with NASA is to answer: how common are other Earth-like planets? You know, the planets that fit the basic properties of another Earth. They’re similar size, similar orbital period, orbit around a similar type of star.
And our answer is: there’s a lot of uncertainty.
Kepler, because it ended early because of those instrumental failures, it wasn’t really able to give us the constraints that we would have hoped for.
But we estimate around half — so 50% — of all Sun-like stars probably have another Earth-like exoplanet.
So if there’s hundreds of billions of Sun-like stars in our galaxy, that could mean hundreds of billions of potentially habitable Earths.
But again, depending on what that number actually is, that’s going to be really important for defining the size of the telescope that you need.
Isa: Most of them that we find are like super-Earths or sub-Neptunes, right? Isn’t that the most common?
Michelle Kunimoto: That’s correct. So planets between the sizes of the Earth and Neptune are the most common planets we’ve found so far, and we don’t have one in our solar system.
So that’s an open question.
Farha Guerrero: I feel that this conversation has been very Galilean — I don’t know if that is the right word — but what it means to me, what makes me so inspired by people like you, Michelle, and I say this sincerely, is that what you and your team are doing is extraordinary.
This is, as I said earlier, is happening in real time. This is extraordinary work that is changing, evolving, and transforming the narrative as we speak.
And it’s a collaboration, but— there is also discourse. It’s got everything.
And I think that’s what the cosmos is. A place of chaos as well as a place of beauty and harmony.
But I’m going to end it there.
Thank you so much for coming and sharing your enthusiasm, and we will be following you! You’re going to have to come back because your research is so fast-paced. We’ll need you back, to give us an update.
Michelle Kunimoto: Sounds good.Thank you so much.
Farha Guerrero: Yeah, I hope you enjoyed this.
Michelle Kunimoto: Really did. Thank you.
Farha Guerrero: And thank you, Isa, for joining in.