Research Byte

Published in the RSAA Lunations
Vol1 Issue30 1–31 July 2022

A stellar embrace ending 10,000 years ago

Half of all stars are born into couples, called binaries. But unfortunately, we cannot compute their lives. Single stars, such as the Sun, we understand just fine: they are round and simple, and long timesteps in computer calculations are enough even in the late stages in their lives.

Binary stars are an issue. They are born wide as the clouds they collapse from need space to start with. And while they perform their distant orbital dance around each other during the main part of their lives, they are pretty much single stars, just two of them. But then one of them becomes a red giant and grows to engulf the other. Enter the “common-envelope phase”, postulated in 1976. Two stars orbit inside an opaque envelope of odd and changing shape. This phase has gained notoriety. We have never seen it. We cannot compute it. Time steps need to be on scales of minutes instead of millennia, the physics is too complex, and calculations rapidly lose their connection with reality.

The common-envelope (CE) phase was postulated to explain the abundance of implausibly tight binaries in the Galaxy. Inside the CE, friction slows down the two stars and their orbit spirals in. Friction produces heat, and eventually the envelope overheats and is blown away: the “CE ejection”. Revealed from within is a much tighter binary than nature would make without CEs. Here is the newborn seed population for stellar mergers, neutron-star mergers, black-hole mergers, gravitational-wave events, tracers of cosmic expansion, supernovae, kilonovae, other novae, contact binaries, cataclysmic variables, and a list of transient categories yet to be discovered when LSST starts collecting bytes, or Petabytes ideally.

Reconstructing anything from all those tight binaries is a problem. We understand single stars and we can put a clock on their evolutionary path. But as we do not know what exactly the CE does to a binary orbit, all the tight binaries we see are at “some” time after the CE ejection, whether that’s 10 million years or 10 billion years ago, and they have evolved since then. So, not only was the clock on the lives of the star pair reset, when they were reborn tight from a CE phase, but we even lost the clock. If we knew just how they get reborn, we could connect that to their present appearance and work out where they are along their way. But we can’t.

This is even a problem for interpreting galaxy SEDs, which are affected by tight binaries and their evolution, yet the population synthesis for the binary part should be taken with a solar mass of salt.

Now, this may slowly start to change, thanks to a serendipitous discovery.

Following the discovery of Blue Large Amplitude Pulsators (BLAPs) by the OGLE team, an exceedingly rare type of star that can change luminosity by a stunning 40% in 10 minutes, I decided to query the SkyMapper DR2 catalogue when it came out in 2019. After all, the SkyMapper Main Survey takes three pairs of u/v filter exposures in the space of 20 minutes as part of its ‘colour sequence’ and does that across the entire Southern hemisphere. Perhaps this might allow us to discover half of all BLAPs in the Galaxy that are not hidden by dust?

In June 2019, Chris Onken monitored the four resulting candidates with SkyMapper and Mike Bessell observed and reduced spectra with WiFeS at the ANU 2.3m telescope. While we found a BLAP, that wasn’t the scoop. Mike Bessell pointed out that one star had way too much Ca absorption given the Schlegel reddening (see figure), which doesn’t make sense either with how the ISM works or with how a stellar atmosphere works at 50,000 Kelvin. “Circum-stellar material then”, I said (“CSM”). A smile from a bit of sightseeing, and we move on.

Chris Onken found a light curve shape reminiscent of an eclipsing contact binary in a 3.5-hour orbit. The penny dropped. A very tight binary. With CSM. “Is that an ejected common envelope?”, I almost shouted (people have patience with my enthusiasm, at times). The more common, and indeed observed for real, things are nova shells around binaries. Mike confirmed the Ca line did not move over time, while the Balmer lines followed the orbit.

I asked Zhanwen Han from Yunnan to help out. He is both a theorist and observer working on tight binaries. We go back 20 years. His student, Jiangdan Li, did all the binary modelling from light curve, radial velocity curve and SED, and she prepared the paper. It turned into a journey with a couple of hiccups that will become personal anecdotes. In 2022 a manuscript seemed ready for submission. The object turned out to be a Roche-lobe filling subdwarf O star with a white dwarf companion and an accretion disk, and nearly a solar mass between them. In the meantime, two similar objects had been published by others. Astonishingly, our manuscript did not mention any Ca lines. “What happened to the common envelope?”, I asked. “What common envelope?” I heard. “The reason we are working with you,” I answered.

A period of email silence was followed by frantic activity to understand the Ca line. Its blue-shift (170 km/s relative to systemic) and column density allow putting the picture together. Luckily, the Kepler satellite and the Zwicky Transient Facility had observed the object as well. As a result, we see the orbital period shrinking by 0.1 sec over six years, which requires friction on residual material. The best model now is that a common envelope of over a solar mass was probably ejected about 10,000 years ago. And we are all best friends again :-).

The referee said, in a nutshell, that this paper was a sad case, as it seemed so significant and well-done, yet we claimed a radius of 0.2 solar radii for the white dwarf, which is bigger non-sense than most recent news on the air and invalidates the entire analysis. Of course, we had all proofread the paper diligently. Of course, we were all mortally embarrassed. We may have even scolded a colleague or three in years past for saying things like “a WD with 0.2 solar radii”, which meant someone spoke with the brain on standby. In any case that was the size of the secondary as derived from the light curve. We soon realised that it was the size of the accretion disk, which dominates over the small white dwarf in the light curve. After fixing this point, all was well.

The work is now published as: A Roche Lobe - filling hot Subdwarf and White Dwarf Binary: Possible detection of an ejected common envelope?  J. Li, C. A. Onken, Ch. Wolf, P. Nemeth, M. Bessell, et al., MNRAS, 2022, in press.

See also this ANU press release on Thursday 7 July

Christian Wolf