Dr Daniel Williams

Research Associate (Physics & Astronomy)

telephone: 01413307168
email: Daniel.Williams@glasgow.ac.uk
pronouns: He/him/his

Biography

I'm an astronomer working on the analysis of gravitational wave data produced by advanced ground-based observatories. I completed my PhD during the first two observing runs of the Advanced LIGO detectors, and moved into post-doctoral research during the third observing campaign. As we move into regime of routine observation in the fourth observing run I am responsible for managing the flag-ship analysis for LIGO-Virgo-KAGRA (LVK) collaboration data.

Unlocking the mysteries of black holes from their death-spirals

The Universe is full of the remnants of stars which have reached the end of their lives, collapsed under their own gravity, and become ultra-dense remnants: black holes and neutron stars. It is common for stars to form in systems where a pair are orbiting each other: a binary star system. These arrangements are also common with black holes and neutron stars, which can also form binaries.

In 1915 Albert Einstein's General Theory of Relativity changed our understanding of how the force of gravity worked througout the Universe, and predicted that orbits decay slowly over time. As the orbit decays energy is radiated away in the form or gravitational waves. In the last decade new frontiers in astronomy have been opened-up by our ability to observe gravitational waves.

The special circumstances which exist in a binary containing black holes or neutron stars make observing the decay of these orbits possible with our current technology, and we are currently able to observe the very end of these orbits, just before the two objects collide with each other, as well as the collisions, which are the most energetic explosions in the Universe. We can work backwards from the signal we receive in the gravitational wave to better understand the system which generated it.

This is parameter estimation, and requires computational and statistical techniques working together to unpick the information which is contained in signals which are often extremely weak by the time we can detect them.

Scaling up: from a handful of detections a year to dozens per day

I have pioneered techniques for performing these analyses, and performing these at scale. Every year technological developments make gravitational wave detectors more sensitive, allowing us to make more detections, and within the decade ground-based detectors will make many detections every day.

My work allows us to perform these analyses faster, manage them more easily, and allow us to get as much information as we can from our data as efficiently as possible, allowing the astronomy community to focus on the results of our analyses as quickly as possible. I have developed a software tool called asimov which is used extensively within the gravitational wave community to orchestrate these large and complex analyses.

Understanding gravitational waveforms

Accurate parameter estimation, and inference of the astrophysical systems producing gravitational waves relies upon a solid understanding of the underlying physics which produces the gravitational waves.
While this is ultimately derived from Einstein's General Relativity, it is normally very difficult to move from this theory to predictions about gravitational waves, and their precise properties.

The best solution to this is to use supercomputing facilities to perform simulations using Einstein's theory, however these require a long time to run, and require very powerful computing facilities. In order to investigate signals from systems which don't match the ones which have been simulated this way the gravitational astrophysics community has created a number of approximants; sets of equations and computer programs which are able to approximate the expensive simulations quickly, but at the expense of losing some accuracy.

Agreement between these approximants, and between the approximants and the simulations is normally very good, however as the sensitivity of gravitational wave detectors increase, and as the number of detections increase it is increasingly likely that we will observe systems for which this is not the case, either because the configuration of the system is very different to anything which has been simulated before (and thus used to inform the approximant).