The already tentative detection of phosphine is under heavy scrutiny from the scientific community, and a problem with the way ALMA’s software interpreted the original observation isn’t helping.
Jane Greaves (Cardiff University, UK) and colleagues published the possible discovery of phosphine gas in the cloud decks of Venus. The study garnered a lot of media attention because on Earth the molecule is a biosignature associated solely with life. At the very least, the sighting indicated that some unexpected chemistry was taking place. What’s more, the team couldn’t explain the presence of phosphine based on lightning, volcanoes, or meteorites.
A considerable amount of scientific scrutiny followed in the two months since the announcement. It’s too early for any definitive refutations, but four critical studies by prominent groups have already been put up on the arXiv preprint server. Therese Encrenaz (Paris Observatory), Ignas Snellen (Leiden Observatory), Mark Thompson (University of Hertfordshire), and Geronimo Villaneuva (NASA), each investigated the phosphine detection in different ways: by looking at old data, critiquing the statistical analysis, and suggesting that the phosphine signal was actually a misinterpretation of sulphur dioxide. All of these studies are still under review and Greaves’ team is currently preparing responses to them.
The phosphine line was seen first with the James Clerk Maxwell Telescope (JCMT) in 2016, and then again by the Atacama Large Millimeter/submillimeter Array (ALMA) in 2019. If the “methane on Mars” saga is illustrative, it could be years and many observations before phosphine on Venus is definitively verified or excluded. But the biggest recent upset to the story has come not from skeptics in the scientific community, but from ALMA itself.
Greaves’s team released their calculations, models, and methods to the public immediately upon publication, so that other astronomers could vet the find. Greaves’ willingness to expose the study to scrutiny brought a large amount of attention to the observatories involved. That in turn led ALMA staff to find a problem with the data, specifically with the way Jupiter’s moon Calisto is used for the Venus calibration. (See the official statement here.)
“The publicity drew the attention of people with experience on other interferometers and ALMA experts specializing in other fields who made unexpected connections,” explains Anita Richards, the team’s ALMA liaison. “All of this has ultimately contributed to a new reduction.”
The problem arose because Venus is not a “standard” target for ALMA. The array of radio dishes spends most of its time looking at things that very far away, and thus very faint.
“Observing solar system objects is quite a small part of what ALMA does,” Greaves says. “ALMA is learning from this situation as well, in terms of future observations. We're not the only project affected. They've had to contact a number of other principal investigators.”
Looking for something barely visible (the phosphine absorption line) inside a large, bright, nearby object (Venus) creates large uncertainties with the interpretation of the subsequent data. It’s possible to see things that aren’t there, in the same way that looking at a bright light causes afterimages when your eyes are closed.
“ALMA told us, ‘we've never tried to push the limits of getting a faint line in this way before,’” Greaves says. “So it was a risk. And obviously I'm not comfortable with finding out the drawbacks afterwards. But we did want the community to look at it and that's what they have done.”
This issue doesn’t necessarily mean that ALMA isn’t a good telescope for Venus. The team has now come up with a new observational strategy. They plan to take a longer exposure, pointing the telescope at slightly different places on the planet to make a kind of mosaic.
In the meantime, Greaves has performed a quick analysis on the “fixed” data from ALMA, undertaken in the past two weeks. The results have just been posted on the arXiv as part of the team’s response to Villaneuva’s critique. Although the phosphine feature has not disappeared, she says the signal isn’t as strong as it was before.
“When ALMA starts back up, which will be sometime next year, we hope to redo the observation,” she says. “Because now the feature just looks promising rather than really good." Whereas the initial announcement met the “five sigma gold standard,” in which the signal is five times stronger than the background noise, the fixed data no longer provide that certainty.
It has become evident that in this context, both ALMA and JCMT are working at the limit of their observatory capabilities. Astronomers must obtain new spectra yet it isn't clear that ground-based telescopes are up to the task. Even so, the team is moving forward with a new approach that will allow them to focus on distinct locations on the planet where they think phosphine is present, which is what they wanted to do originally.
“Because of the extra ‘wiggles’ in the signal, we were forced to average north and south detections across the planet,” Greaves says. “But now I'm excited to say we can go back and look at little bits of the planet, and hopefully say, ‘There it is, and there, and there.’”
Worldwide attention combined with the team’s total transparency has resulted in prompt and widespread engagement, which Greaves admits has sometimes been difficult. But it also enabled this quick correction, which will benefit future ALMA users observing solar system objects. And it has inspired retrospective analyses, like the study coming out soon by Rakesh Mogul (Cal Poly Pomona), which found signals in the data from NASA’s Pioneer Venus spacecraft that are consistent with the detection of phosphine.
Overall, Greaves is pleased with the level of interest her study has inspired: “This process has sometimes been painful, but it's also been good for getting the data right."