UC Irvine

Neutron stars harbor dense nuclear matter in density and temperature regimes inaccessible in terrestrial laboratory experiments. Understanding the interior of these stars and determining the equation of state of such matter has been a forefront area of nuclear and astrophysics research for decades, as the properties of neutron star matter are of key importance in comprehending the early Universe, laboratory and particle physics, and other astrophysical phenomena like supernovae and stellar mergers.

In this work, the determination of the equation of state of dense neutron star matter and neutron star structure is approached in several ways. The first is through theoretical modeling carried out in the framework of relativistic quantum field theory at both zero and finite temperatures. The calculated equation of state models are then used to compute the properties of neutron stars using Einstein’s theory of general relativity. Special emphasis is placed on the structure and stability of differentially rotating compact objects, which may exist on short timescales following extreme astrophysical events like binary neutron star mergers.

The second approach to determining the equation of state of dense neutron star matter is based on machine learning, utilizing X-ray spectra and observable properties such as mass and radius of neutron stars. A novel inference of the equation of state directly from simulated high-dimensional spectra of observed stars is compared with a calculation of the full likelihood of the equation of state parameters, accomplished by replacing intractable elements of the likelihood with machine learning models trained on samples of simulated stars.