In September, a team of astronomers announced a breathtaking finding: They had detected a molecule called phosphine high in the clouds of Venus, possibly indicating evidence of life.

That discovery shook the scientific establishment. Once thought of as Earth’s twin, Venus — though nearby and rocky — is now known to have a hellish environment, with a thick atmosphere that traps solar radiation, cranking surface temperatures high enough to melt metal, and accompanied by surface pressure akin to that thousands of feet below Earth’s ocean surface.

But the detection, led by researchers from Cardiff University in Wales, the Massachusetts Institute of Technology, and the University of Manchester in England, was high in the atmosphere, where conditions are far more hospitable and the idea of microbial life more plausible. It was accomplished using spectroscopy, a method of determining the presence of different molecules in a planet’s atmosphere by analyzing how those molecules alter the light reflected from the planet. A key member of the team was fellow Clara Sousa-Silva, who had spent years studying the molecule’s spectroscopic signature and who believes that phosphine is a promising way to track the presence of extraterrestrial life.

Sousa-Silva shifted her fellowship from MIT to the Center for Astrophysics | Harvard & Smithsonian and will spend the next two years advancing her work on biosignatures and life on other planets. She spoke with the Gazette about the recent discovery and what the future of the search for life may hold.

Q&A

Clara Sousa-Silva

GAZETTE: You study biosignature gases, and your website says phosphine is your favorite. What is a biosignature gas and what’s so special about phosphine?

SOUSA-SILVA: A biosignature gas is any gas in the planetary atmosphere that is produced by life. That by itself is not particularly interesting because molecules that can be produced by life can often be produced by many other things. So another question is: What is a good biosignature? And the answer to that also explains why phosphine is my favorite.

A good biosignature isn’t just produced by life and released into an atmosphere. It is also able to survive in that atmosphere and be both detectable and distinguishable. So, if we’re looking at an atmosphere from far away, say from a different planet, and we detect an interesting molecule, that’s great. But maybe, because of low resolution in the instruments, lots of molecules look very similar to one another and the spectral signature also corresponds to a different molecule than one we thought we saw. So, you want a biosignature to be distinguishable.

A good biosignature has a final characteristic: It has limited or accountable false positives. That means if it is produced by life, if it survives in the atmosphere, and you can detect it unambiguously, you still need to know if it was in fact produced by life or if it was accidentally produced by some other nonbiological process like photochemistry or volcanism. So, a good biosignature is all of these things: It is produced by life in large quantities and survives; it’s unambiguously detectable; and is unambiguously assigned to life.

Famous biosignatures like oxygen and methane rank very well in the first few of these parameters. But methane, for example, looks an awful lot like every other hydrocarbon. And so knowing if you’re looking at methane versus a different molecule that also has carbons and hydrogens is quite hard. And even if you can unambiguously assign the thing you saw to methane, you don’t know if you can unambiguously assign it to life.

Phosphine has a unique spectral signature, because the spectrum for phosphine is composed of the behavior of the bonds between hydrogen and phosphorus, and that’s a very rare bond in gas molecules. So phosphine is quite easy to distinguish, meaning it’s easy-ish to detect, and it is also produced by life. But it’s not produced by life in large quantities, so that’s a negative point for phosphine. But then, it’s so hard to produce without the intervention of life on rocky planets that it’s very low on false positives. I think phosphine is a well-balanced biosignature: produced in detectable quantities by life, being distinguishable, and having low false positives. That’s why it’s my favorite.

Clara Sousa-Silva, a fellow who grabbed headlines in September because of new findings of a potential signature for life on Venus, discusses that research.

Kris Snibbe/Harvard Staff Photographer

GAZETTE: Your site also says that phosphine is toxic to life that uses oxygen metabolism. So why is it a likely sign of life on Venus?

SOUSA-SILVA: I don’t know if it’s likely. I wouldn’t dare put a probability on that. It is toxic to life on Earth that uses oxygen. And that is, obviously, us and everything we love. But lots of life on Earth does not rely on oxygen, and for the majority of time that life existed on Earth it also didn’t rely on oxygen. Granted, it wasn’t the most thrilling life. It wasn’t writing great works of literature, but it was nevertheless popular on Earth and seemingly very happy, thriving in forms that had no need for oxygen.

The reason why phosphine on Venus, if it’s there, may signify life is more that we cannot explain it in any other way. We have no good explanation for the presence of phosphine on Venus, and we do know it can be produced by life. That doesn’t mean that’s what’s happening on Venus. That’s just, as extraordinary as it might sound, the best guess we have at this point.

GAZETTE: Let’s talk specifically about the findings from September. What did you and your colleagues find on Venus?

SOUSA-SILVA: It was an analysis of two separate observations done about 18 months apart. One was done with the JCMT, the James Clerk Maxwell Telescope, which is on Mauna Kea [in Hawaii]. That observation has a tentative signal that could be assigned to phosphine. We then applied for time on ALMA [Atacama Large Millimeter/submillimeter Array in Chile], which is a much more powerful array of telescopes and which seemingly got a slightly stronger signal that also corresponded to phosphine. This is encouraging because the odds that a random signal will appear in the same place 18 months apart, using two different instruments, are very slim.

The analysis was figuring out: One, is the signal real, because both of these instruments were collecting data very much at the limits of their capabilities. Two, if the signal is real, is the most plausible candidate phosphine rather than a different molecule? And three, if indeed the signal is real and it is phosphine, who or what is making it? Those are the three steps of the main article. This was about two years of work on top of my many years of work investigating phosphine as a biosignature.

It took a long time and a large international team, including Anita [Richards, of the University of Manchester, U.K.] and Jane [Greaves, of Cardiff University]. Jane is the lead author of the paper that came out in September specifically extracting the signal from the data. Then lots of us were trying to figure out if the signal belongs to phosphine and if so, at what abundances. My contribution is that I know the pure spectroscopy of phosphine very well. My entire Ph.D. was dedicated to the spectroscopy of phosphine. So I was able to help figure out, if it was phosphine, what kind of abundances it was present in.

I was also able to provide a list of other candidate molecules that could mimic the signal. The most promising one is phosphine, but the second-most-promising one is SO² (sulphur dioxide), which would be a strange molecule to find in that location of Venus, but not anywhere near as strange as finding phosphine. So it was an important candidate to check. Then, if it is indeed phosphine and the signal is real, figuring out what is producing it was led by William Bains [at MIT]. It was also a large team, figuring out every process that might make phosphine and excluding a near-infinite list of negatives. It’s very, very hard to know if you’ve reached the end of that list.

GAZETTE: So they’re working through the ways you might make phosphine that likely didn’t occur on Venus?

SOUSA-SILVA: We’re trying to find an explanation, any explanation, and we did find a few methods that could produce small amounts of phosphine, but they were always quite trivial and always many orders of magnitude below what our estimates were for the signal detected in the clouds of Venus.

GAZETTE: Is this discovery a warmup for finding phosphine and detecting biosignatures on planets around other stars?

SOUSA-SILVA: I think it’s exactly a warmup for the search for life. It’s an excellent case study in the world of astrobiology.

The odds that we find life beyond Earth from a booming, unambiguous, intelligent signal from the heavens is very slim. It’s likely, if we ever find life, that it is going to be something with quite a lot of uncertainty, and it will be really hard to even estimate that uncertainty. We won’t be able to say, “Oh, we found life with 80 percent certainty.” Those numbers are not ones we can do right now.

What we can do is look at planets that have potentially habitable environments, look for molecules that can be associated with life, and then try to explain what’s going on there. We found a biomarker in a place that is potentially habitable. That’s a crucial first step, but it’s very far from the final step. We now need to figure out what other molecules would that biosphere produce? How will they interact with one another? How do we disentangle those behaviors from the spontaneous behaviors of a dead atmosphere?

So, it’ll take a lot of work. We are very lucky to have Venus right next door so that we can use it as a lab. We can test all these theories in a way that we won’t be able to when we find a biomarker on an exoplanet, where there’s no hope of actually going in and probing the atmosphere to check. So this is a really important step.

This has been reasonably controversial — and it should be — but we will have to do this many times. And every time we hope to be better prepared and have a better tool kit so that there’s less uncertainty. But it’ll take a long time before we can unambiguously confirm life elsewhere.

GAZETTE: Before this discovery, Venus had been largely dismissed as a place for life because of its surface conditions. Your discovery has highlighted that a biosphere can be in places that may not immediately come to mind: high in the clouds where conditions are different. Is there a lesson here for thinking unconventionally when we evaluate places for life, especially since even here on Earth we’ve found life to be tough and enduring and in surprising places?

SOUSA-SILVA: Life is very resilient and very resourceful on Earth and there’s no reason to think that’s some special characteristic of life on Earth rather than of life itself. We have ignored Venus because Venus is quite horrid to us. When we sent probes, they melted dramatically so we didn’t feel particularly welcome. It seems easier to imagine a place like Mars as habitable, even though actually there’s so little atmosphere and so little protection from the sun’s radiation that it’s really not an easily habitable surface.

Mars is mostly uninhabitable, like Venus, just in a much quieter way. Mars will kill you, but it doesn’t melt you, so it feels more habitable, though I have no loyalty to either planet as a place to find life. This is hopefully going to help us think of habitability in a less anthropocentric way — or at least a less terra-centric way — and to think of habitability not just as a rocky planet with liquid water on the surface, but to think of subterranean habitats, moons of gas giants — something people already consider — and envelopes of an atmosphere as potentially habitable places in an otherwise uninhabitable planet.

GAZETTE: What did you think when it became apparent that it might be life on Venus? Was that an exciting moment?

SOUSA-SILVA: It was kind of a strange reversal. I had for years been working on this completely hypothetical investigation: If we found phosphine on a terrestrial planet what would it mean? I had concluded that because it has so few false positives on terrestrial planets that it could only mean life. I submitted the paper with this conclusion, and it was not controversial. The reviewers were fine with the idea — they had issues with other parts of the paper, but this didn’t bother them at all. No one cared because it was hypothetical: I was imagining this exotic, distant planet.

When I was contacted by Jane, who had this tentative detection of phosphine on Venus, my not-so-controversial statement was now really extraordinary. And Venus is next door, so my hypothetical scenario became very concrete, very quickly. That was two years ago. We spent about a year and a half basically redoing and refining the analysis that we had done for my paper. This was, again, led by William Bains to try to figure out whether this is what happened on Venus. Venus is not your classic, potentially habitable exoplanet. It’s a pretty infernal place and maybe there phosphine could be made abiotically. So I never got to be as excited as I might at the first mention that phosphine had been found on a terrestrial planet. I expected this to happen hopefully before I die, but probably after I retire, not within months of submitting my hypothesis.

I also immediately felt like I could not be trusted because I’m so biased. I’ve been working on phosphine for so long. I am a junior scientist without a permanent job. It would be so valuable to me for it to be life that I can’t be trusted to assess this accurately. So I was very careful to not get too excited. I had a strong glass of whiskey that evening, but that was it. Then I went and did the same work that we always do, which was to check every possible mechanism that can make phosphine, every possible molecule that can mimic the signal, and look again at everything I’ve done before and check for mistakes. It was nerve-racking to explore this expression of my prediction so nearby, so quickly.

GAZETTE: Have you had a chance since the original paper was published?

SOUSA-SILVA: Well, we did a good thing and paid a cost. Unlike a lot of observations of this kind, we published all our data and all our code. Everything was ready for people to come and tear it apart. So people did, which meant I never did get a little time off to enjoy it. It was great because they found a calibration mistake, and ALMA was able to rectify that, which allowed our team to reanalyze the data — they’re still doing it now. There was just way too much press and then way too much criticism, and I still haven’t taken time off.

GAZETTE: About the scientific debate, how to you respond to the failure of other research groups to replicate the results?

SOUSA-SILVA: This is the part of the work where I’m only tangentially involved, since I’m not doing any data reduction [of readings from Venus’ atmosphere]. This debate is a consequence of working at the edge of instrument capabilities, and the data are always going to be very noisy and delicate until we have better telescopes. Any discoveries made from these data, from the edges of our ability, are always going to be up for discussion. It’ll be nice when there’s a gold standard method for reducing these data, but there isn’t, so people disagree on the best way of extracting a signal without introducing spurious signals.

The disagreement comes in a variety of forms, but the teams that didn’t replicate the results, don’t replicate the results in different ways. For example, the [Ignas] Snellen team [from Leiden University in the Netherlands] looked at the ALMA data before the calibration error had been corrected. I’m looking forward to seeing their revised analysis of the better data. The Villanueva team [led by Geronimo Villanueva at the NASA-Goddard Space Flight Center] that looked at both the ALMA data and the JCMT data, did find signals in the JCMT data, which, of course, begs the question of “Where does the signal go in the ALMA data?”

They do disagree on the source of the JCMT signal, though. SO² [sulfur dioxide], our second-most-plausible candidate, is their first-most-plausible candidate. And that is an even more complicated question of how you choose between two molecules that can simulate the same signal at these resolutions. Our team’s argument is that the SO² [spectra] is a little off — you would expect SO² to show up in different areas of the white bandpass. There also isn’t enough SO² to justify the signal, so phosphine would need to complement the size of the signal. It’s a difficult argument to make — and we’re at the edge of the statistical significance of the signal — but it’s a totally valid argument.

Then there’s the archival Pioneer data that was revisited and that they think could correspond to phosphine. It’s hard to bring all of this data to a place where they agree with one another, sadly, because people want to know the truth — I do, too. But the only real conclusion we have is that we don’t know Venus well enough, and we need more data. We need more observations that are not at the edge of instrument capabilities so that there’s no ambiguity in what we’re looking at.

GAZETTE: Let’s talk a little bit about what you’ll be doing here at Harvard. You’ve been a fellow at MIT. Is the fellowship split between there and here?

SOUSA-SILVA: No, I moved it. I am 100 percent Harvard — for the last two months, I think. It’s very new.

GAZETTE: Who will you be working with and what will you be doing?

SOUSA-SILVA: The 51 Pegasi b Fellowship is a wonderful three-year prize fellowship that is provided by the Heising-Simons Foundation. I did one year at MIT, and I’ve moved to Harvard for the last two years of the fellowship. My host is Dave Charbonneau — part of the reason I moved to Harvard is because of the expertise he has — and the team that surrounds him — on exoplanet atmospheres. There’s also the HITRAN [High resolution Transmission molecular absorption database] group, led by Iouli Gordon — and previously, Larry Rothman — who are world leaders in spectroscopic databases, which is the bread and butter of my work. So that combination of expertise made Harvard perfect.

GAZETTE: Are you doing most of your work out of your home now or are you able to commute to the CfA physically?

SOUSA-SILVA: No, I don’t even know where my office is yet. I would love to be commuting to the CfA, but because my work can be done remotely, it shall be done remotely.

GAZETTE: Are you continuing to work on phosphine and Venus or are you moving on to other topics?

SOUSA-SILVA: I’ll give it the same percentage of my time as I have in the past. Phosphine is very much my expert molecule, but 50 percent of my work is pushing against the notion of looking for single indicators of life. Because unless we get a radio signal in prime numbers or an unambiguous sign of CFCs [chlorofluorocarbons] or other really complex pollutants, we are going to need more than one molecule; we’re going to need a whole array of molecules that together paint the picture of a biosphere with all its complexity and interactions.

So most of my work is trying to provide a tool kit that can detect every molecule that could potentially be in a habitable atmosphere. I started the work at MIT. They had come up with a list of all the possible molecules that could form in the context of a biosphere: 16,367. I know that number because I’ve been working on it for so long.

Out of those thousands, we have spectra of some quality — and some of them are rough — for less than 4 percent of them. For the majority of molecules, we don’t even have even a crude ability to detect them. So most of my work is trying to simulate that spectra so we have at least some idea of what these molecules look like. That’s the connection to HITRAN. They have extremely high accuracy and extremely careful data on a handful of molecules, a little over 50. That is wonderful, but only a small dent in the list of 16,000-plus.

I created a small program called RASCALL, for Rapid Approximate Spectral Calculations for All. The idea is to make really rough, very quick spectra for all of these molecules, and then build on it. Without RASCALL, the way I did my phosphine spectra took me a bit over four years and many extremely expensive supercomputers. I can’t repeat that for the 16,000 molecules. I calculated that it would take me over 62,000 years. I’m trying to shorten that timescale into something that resembles my lifetime, and that’s where RASCALL comes in.

GAZETTE: Folks like you will be helping answer an interesting question in the decades to come: whether life is something rare or whether it’s not really that rare after all. It seems the thinking on that has been shifting in recent decades.

SOUSA-SILVA: I do like that the shift is happening and that people are thinking that life is more common. I’m hoping that shift will go so far as thinking that life is not that special. It’s just an inevitable occurrence in a variety of contexts. If it can appear in places as different as Earth and Venus, which are at first glance similar because of their size and location but otherwise very different, then it must be extremely common because it would be the height of hubris to think that only the solar system can have life, but it has arisen twice in totally different environments.

That seems really implausible. The sun is average, rocky planets are extremely common, the molecular cloud that formed the solar system was not special. Life on Earth came to be in a huge diversity of forms, and life changed Earth’s atmosphere many times. We only have one planet where we know life existed, but Earth has been many planets, which is something an astronomer colleague of mine, Sarah Rugheimer, likes to say. We have quite a lot of data points that basically show that life is pretty good at making itself happen in many ways throughout history.

Interview was edited for clarity and length.

Astrochemist brings search for extraterrestrial life to Harvard