Stampede2 supercomputer simulations help capture helium gas clouds escaping distant planet.
A planet located approximately 950 light years from Earth could be the Looney Tunes’ Yosemite Sam equivalent of planets, blowing its atmospheric ‘top’ in spectacular fashion.
Called HAT-P-32b, the planet is losing so much of its atmospheric helium that, according to observations by astronomers, the trailing gas tails are among the largest structures yet known of an exoplanet, a planet outside our solar system.
Three-dimensional (3D) simulations on the Stampede2 supercomputer of the Texas Advanced Computing Center (TACC) helped model the flow of the planet’s atmosphere, based on data from the Hobby-Eberly Telescope of The University of Texas at Austin’s McDonald Observatory. The scientists hope to widen their planet-observing net and survey 20 additional star systems to find more planets losing their atmosphere and learn about their evolution.
Discovering the Helium Tail
“We have monitored this planet and the host star with long time series spectroscopy, observations made of the star and planet over a couple of nights. And what we found is there’s a gigantic helium gas tail that is associated with the planet. The tail is large — about 53 times the planet’s radius — formed by gas that’s escaping from the planet,” said Zhoujian Zhang, a postdoctoral fellow in the Department of Astronomy & Astrophysics, University of California Santa Cruz.
Zhang is the lead author in a study on the helium tail detected from HAT-P 32b that was published in Science Advances in June 2023. The science team used data from the Habitable Planet Finder spectrograph, an instrument on the Hobby-Eberly telescope, which provides high spectral resolution of light in near infrared wavelengths.
The planet HAT-P-32b was discovered in 2011 using spectroscopic data from the Hungarian-made Automated Telescope Network. It’s known as a ‘hot Jupiter,’ a gas giant similar to our neighboring planet Jupiter, but with a radius twice as large. This hot Jupiter hugs closely in orbit to its host star, about three percent the distance from the Earth to the Sun. Its orbital period — what we consider a year here on Earth — is only 2.15 days, and this proximity to the star scorches it with both long and short-wave radiation.
Delving Into the Neptunian Desert
The main motivation for the scientists’ interest in studying hot Jupiters is their pursuit of the mystery of the Neptunian desert, the inexplicable relative scarcity on average of intermediate-mass planets, or sub-Jupiters, with short orbital periods.
“One of the potential explanations is that maybe the planets are losing their mass,” Zhang offered. “If we can capture planets in the process of losing their atmosphere, then we can study how fast the planet is losing their mass and what are the mechanisms that cause their atmosphere to escape from the planet. It’s good to have some examples to see like the HAT-P-32b process in action.”
The light analyzed in the study comes from the star HAT-P-32 A. It’s slightly hotter and similar in size to our own sun. The analyzed light is not just straight starlight. As the planet passes in front of the star, for just a couple of hours the starlight gets filtered the most by the planet’s gassy atmosphere. This filtering, called absorption, reveals features of the transiting planet, in this case huge outflows of helium when the spectra were analyzed.
Zhang and colleagues used a technique called transmission spectroscopy to separate the starlight into its component frequencies, like a prism separates sunlight into a rainbow spectrum. Gaps in the spectrum indicate light being absorbed by elements in the gaseous atmosphere of HAT-P-32b.
“What we see in our data is that when the planet is transiting the star, we see there’s deeper helium absorption lines. The helium absorption is stronger than what we expect from the stellar atmosphere. This excess helium absorption should be caused by the planet’s atmosphere. When the planet is transiting, its atmosphere is so huge that it blocks part of the atmosphere that absorbs the helium line, and that causes this excess absorption. That’s how we discovered the HAT-P-32b to be an interesting planet,” Zhang said.
3D Simulations and Atmospheric Dynamics
It got more interesting as they developed 3D hydrodynamical simulations of the HAT-P-32b and host star, led by Antonija Oklopčić, Anton Pannekoek Institute for Astronomy, University of Amsterdam; and Morgan MacLeod, Institute for Theory and Computation, Harvard-Smithsonian Center for Astrophysics, Harvard University.
The models examined the interactions between the planetary outflow and stellar winds in the tidal gravitational field of the extrasolar system. The models showed columnar tails of planetary outflow both leading and trailing the planet along its orbital path with excess helium absorption even far from the transit points that matched observations. What is more, the models suggest complete loss of the atmosphere in about 4 x 10e10 Earth years.
“We made use of TACC’s Stampede2 system’s Intel Skylake nodes for our calculations,” MacLeod said. “This computation involves tracking flow as it accelerates from a slow-moving subsonic ‘atmosphere’ near the planet to a supersonic wind as it moves further away. The HAT-P-32b system was identified to have a large-scale outflow similar in size to the planet’s orbit around the star. Taken together, these requirements suggest the need for a stable, high-accuracy algorithm for solving three-dimensional gas dynamics.”
The modelers utilized the Athena++ hydrodynamic software and a custom problem setup to do their calculation on Stampede2. With it they solve the equations of gas dynamics in a rotating frame of reference that matches the planet’s orbital motion. Athena++ is a Eulerian code — the flow is discretized with volume elements — and they used nested layers of mesh refinement to capture the large-scale star-planet system along with the much smaller scale of the atmosphere near the planet’s surface.
“Using the TACC HPC systems is a joy,” MacLeod said. “A few things go into this — the first, and most important is the level of support. Whenever I have a problem, I can call the support line, get help, and get back to doing the science that I am best at. Secondly, the vast majority of my time goes into developing and validating model results, rather than running a single, full-scale calculation. The TACC systems are incredibly well set up for this reality, and it hugely speeds up the pace of development. Being able to run test calculations through the development queues or submit larger calculations of a range of sizes in the lead up to an eventual final model is crucial and effective in these environments.”
The Future of Exoplanetary Research
Looking ahead, the scientists hope to continue to develop sophisticated 3D models that capture effects such as atmospheric mixing of gases and even winds within the atmosphere on more distant worlds hundreds and even thousands of light years away.
“Now is the time to have supercomputers with the computational power to make this happen,” Zhang said. “We need the computers to make real predictions based on recent advances in the theory and to explain the data. Supercomputers bridge the model and the data.”
“The best thing we can do is watch the night sky and try to recreate what we see through computer modeling,” MacLeod concluded. “Our universe is complicated. This means we need to have access to the absolute best supercomputing systems.”
Reference: “Giant tidal tails of helium escaping the hot Jupiter HAT-P-32 b” by Zhoujian Zhang, Caroline V. Morley, Michael Gully-Santiago, Morgan MacLeod, Antonija Oklopčić, Jessica Luna, Quang H. Tran, Joe P. Ninan, Suvrath Mahadevan, Daniel M. Krolikowski, William D. Cochran, Brendan P. Bowler, Michael Endl, Gudmundur Stefánsson, Benjamin M. Tofflemire, Andrew Vanderburg and Gregory R. Zeimann, 7 June 2023, Science Advances.
TThe study authors are Zhoujian Zhang of UCSC; Caroline V. Morley, Michael Gully-Santiago, Jessica Luna, Quang H. Tran, Daniel M. Krolikowski, William D. Cochran, Brendan P. Bowler, Michael Endl, Gudmundur Stefánsson, Benjamin M. Tofflemire, Gregory R. Zeimann of UT Austin; Morgan MacLeod of Harvard University; Antonija Oklopčić of University of Amsterdam; Joe P. Ninan of Tata Institute of Fundamental Research; Suvrath Mahadevan of The Pennsylvania State University; Andrew Vanderburg of MIT. Funding came from the NASA Exoplanets Research Program grant number 80NSSC20K0257; National Science Foundation grant 2108801; NASA Hubble Fellowship grants HST-HF2- 51522.001-A. Support also came from NSF grants AST-1006676, AST-1126413, AST-1310875, AST-1310885, AST 2009889, AST 2009982, ATI 2009955, and AAG 2108512 and the Heising-Simons Foundation via grant 2017- 0494.