Spectacular simulations of circling stars

The colourful images on YouTube are almost hypnotising. In spectacular animated videos, neutron stars rotate around each other at increasingly speeds until they begin to merge and cause waves in the surrounding space while flinging energy and matter. The outcome is an exciting jumble of swirling matter, plasma and radiation.

When supercompact stars collide, the ensuing space-time tremors that occur can be observed as far away as on Earth. In the case of neutron stars, the collision can also be seen as a flash of light. That phenomenon was first observed with telescopes on Earth in August 2017.

The German Tim Dietrich (31) previously worked at the Max Planck Institute in Potsdam and came to Nikhef at the start of last year with a Marie Curie Grant. He has received the prize thanks in part to the animated videos that he produces of such stellar collisions.

Are the videos intended for the wider public?
‘Certainly. Outreach is important, but that is not the sole purpose. For researchers, it can also be very enlightening to actually see what happens during such a neutron star merger. It allows them to visualise what the numbers and graphs represent.’

Surely, the videos are not the main objective?
‘The videos are not a standard outcome of the simulations. What is most important for us is the exact form of the measurable gravitational waves that a collision will produce and how that depends on the properties of the system. There is also a practical reason why we do not make animations for everything, namely the data storage. A simulation of 100 milliseconds around a collision easily consumes terabytes of data in our case. Of course, that is slowed down in the animation, as otherwise you would not see anything.’

What is the aim of the simulations?
‘In the case of colliding neutron stars, we need to solve field equations from Einstein’s general theory of relativity. There are analytical methods for that as long as a distance between the stars is large, but those do not work anymore once the stars get closer and the speeds and gravitation become many times larger. At that point, we have to work with computer simulations. These approximate the outcomes that we are actually seeking.’

Is that complex work?
‘Numerical relativistic simulations certainly entail extremely intensive calculations. Even on a supercomputer, a simulation can easily take weeks or months to run. The longest calculation that I have ever performed took nine months. The prize is also in part for the methods I have come up with to nevertheless limit the time needed.’

Are the results of that work also used in the case of real observations by the LIGO and Virgo detectors in the United States and Italy?
‘Our calculations lie behind the templates that are used to analyse the wave patterns measured. From the exact waveforms, it is possible to derive the masses of the colliding objects and how much light was produced during the collision.’

So far, it seems that neutron star collisions are scarce.
‘I had a bet with colleagues that we wouldn’t see the first neutron star merger with LIGO and Virgo until 2021. I was dumbfounded when it was already observed in 2017. And since then, the detectors have observed several serious candidates for such events. This is a fantastic time to be working on gravitational waves.’

Did you lose a lot on that bet?
‘The damage was limited. Just a few beers. And we all won a fantastic first observation.’

Newsletter Inside NWO-I, December 2019