Kilonova Light-Curve Interpolation with Neural Networks
ADS link 
arXiv link
Peng, Y.; Ristić, M.; Kedia, A.; O'Shaughnessy, R.; Fontes, C. J.; Fryer, C. L.; Korobkin, O.; Mumpower, M. R.; Villar, V. A.; Wollaeger, R. T.
Kilonovae are the electromagnetic transients created by the radioactive decay of freshly synthesized elements in the environment surrounding a neutron star merger. To study the fundamental physics in these complex environments, kilonova modeling requires, in part, the use of radiative transfer simulations. The microphysics involved in these simulations results in high computational cost, prompting the use of emulators for parameter inference applications. Utilizing a training set of 22248 high-fidelity simulations, we use a neural network to efficiently train on existing radiative transfer simulations and predict light curves for new parameters in a fast and computationally efficient manner. Our neural network can generate millions of new light curves in under a minute. We discuss our emulator's degree of off-sample reliability and parameter inference of the AT2017gfo observational data. Finally, we discuss tension introduced by multi-band inference in the parameter inference results, particularly with regard to the neural network's recovery of viewing angle.
Interpolated kilonova spectra models: Examining the effects of a phenomenological, blue component in the fitting of AT2017gfo spectra
ADS link arXiv link
Ristić, M.; O'Shaughnessy, R.; Villar, V. A.; Wollaeger, R. T.; Korobkin, O.; Fryer, C. L.; Fontes, C. J.; Kedia, A.
In this paper, we present a simple interpolation methodology for spectroscopic time series based on conventional interpolation techniques (random forests) implemented in widely available libraries. We demonstrate that our existing library of simulations is sufficient for training, producing interpolated spectra that respond sensitively to varied ejecta parameter, postmerger time, and viewing angle inputs. We compare our interpolated spectra to the AT2017gfo spectral data and find parameters similar to our previous inferences using broadband light curves. However, the spectral observations have significant systematic short-wavelength residuals relative to our models, which we cannot explain within our existing framework. In line with previous studies, we consider the contribution of a third component as a radioactive heating source characterized by light, slow-moving, lanthanide-free ejecta with Mth=0.003 M⊙ , vth=0.05 c , and κth=1 cm2/g . When included as part of our radiative transfer simulations, our choice of third component reprocesses blue photons into lower energies, having the opposite effect and further accentuating the blue-underluminosity disparity in our simulations. As such, we are unable to overcome short-wavelength deficits at later times using an additional radioactive heating component, indicating the need for a more sophisticated modeling treatment.
Heavy element nucleosynthesis associated with a gamma-ray burst
ADS link arXiv link
Gillanders, J. H.; Troja, E.; Fryer, C. L.; Ristić, M.; O'Connor, B.; Fontes, C. J.; Yang, Y.-H.; Domoto, N.; Rahmouni, S.; Tanaka, M.; Fox, O. D.; Dichiara, S.
Kilonovae are a novel class of astrophysical transients, and the only observationally-confirmed site of rapid neutron capture nucleosynthesis (the r-process) in the Universe. To date, only a handful of kilonovae have been detected, with just a single spectroscopically-observed event (AT 2017gfo). Spectra of AT 2017gfo provided evidence for the formation of elements heavier than iron; however, these spectra were collected during the first ~ 10 days, when emission from light r-process elements dominates the observations. Heavier elements, if synthesised, are expected to shape the late-time evolution of the kilonova, beyond the phases for which we have spectral observations. Here we present spectroscopic observations of a rapidly-reddening thermal transient, following the gamma-ray burst, GRB 230307A. Early (2.4 day) optical spectroscopy identifies the presence of a hot (T ~ 6700 K) thermal continuum. By 29 days, this component has expanded and cooled significantly (T ~ 640 K), yet it remains optically thick, indicating the presence of high-opacity ejecta. We show that these properties can only be explained by the merger of compact objects, and further, leads us to infer the production of the heavy lanthanide elements. We identify several spectral features (in both absorption and emission), whose cause can be explained by newly-synthesised heavy elements. This event marks only the second recorded spectroscopic evidence for the synthesis of r-process elements, and the first to be observed at such late times.
Constraining inputs to realistic kilonova simulations through comparison to observed r-process abundances
ADS link arXiv link
Ristić, M.; Holmbeck, E.; Wollaeger, R.; Korobkin, O.; Champion, E.; O'Shaughnessy, R.; Fryer, C.; Fontes, C.; Mumpower, M.; Sprouse, T.
Kilonovae, one source of electromagnetic emission associated with neutron star mergers, are powered by the decay of radioactive isotopes in the neutron-rich merger ejecta. Models for kilonova
emission consistent with available modeling and the electromagnetic counterpart to GW170817 also
predict characteristic abundance patterns, determined by the relative balance of different types of
material in the outflow. Assuming the observed source is prototypical, this inferred abundance pattern in turn must match r -process abundances deduced by other means, such as what is observed in
the solar system. We report on analysis comparing the input mass-weighted elemental compositions
adopted in our radiative transfer simulations to the mass fractions of elements in the Sun. We characterise the extent to which our parameter inference results depend on our assumed composition
for the dynamical and wind ejecta and examine how the new results compare to previous work. We
find that a mass ratio of Mw/Md = 2.81 reproduces the observed AT2017gfo kilonova light curves
while also producing the abundance of neutron-capture elements in the solar system.
Surrogate light curve models for kilonovae with comprehensive wind ejecta outflows and parameter estimation for AT2017gfo
ADS link arXiv link
Kedia, A.; Ristić, M.; O'Shaughnessy, R.; Yelikar, A. B.; Wollaeger, R. T.; Korobkin, O.; Chase, E. A.; Fryer, C. L.; Fontes, C. J.
The electromagnetic emission resulting from neutron star mergers have been shown to encode properties of the ejected material in their light curves. The ejecta properties inferred from the kilonova emission has been in tension with those calculated based on the gravitational wave signal and numerical relativity models. Motivated by this tension, we construct a broad set of surrogate light curve models derived for kilonova ejecta. The four-parameter family of two-dimensional anisotropic simulations and its associated surrogate explore different assumptions about the wind outflow morphology and outflow composition, keeping the dynamical ejecta component consistent. We present the capabilities of these surrogate models in interpolating kilonova light curves across various ejecta parameters and perform parameter estimation for AT2017gfo both without any assumptions on the outflow and under the assumption that the outflow must be representative of solar \emph{r}-process abundance patterns. Our parameter estimation for AT2017gfo shows these new surrogate models help alleviate the ejecta property discrepancy while also illustrating the impact of systematic modeling uncertainties on these properties, urging further investigation.
Interpolating detailed simulations of kilonovae: Adaptive learning and parameter inference applications
ADS link arXiv link
Ristić, M.; Champion, E.; O'Shaughnessy, R.; Wollaeger, R.;
Korobkin, O.; Chase, E. A.; Fryer, C. L.; Hungerford, A. L.; Fontes, C. J.
Detailed radiative transfer simulations of kilonovae are difficult to apply directly to observations; they only sparsely cover simulation parameters, such as the mass, velocity, morphology, and composition of the ejecta. On the other hand, semianalytic models for kilonovae can be evaluated continuously over model parameters, but neglect important physical details which are not incorporated in the simulations, thus introducing systematic bias. Starting with a grid of two-dimensional anisotropic simulations of kilonova light curves covering a wide range of ejecta properties, we apply adaptive learning techniques to iteratively choose new simulations and produce high-fidelity surrogate models for those simulations. These surrogate models allow for continuous evaluation across model parameters while retaining the microphysical details about the ejecta. Using a code for multimessenger inference developed by our group, we demonstrate how to use our interpolated models to infer kilonova parameters. Comparing to inferences using simplified analytic models, we recover different ejecta properties. We discuss the implications of this analysis which is qualitatively consistent with similar previous work using detailed ejecta opacity calculations and which illustrates systematic challenges for kilonova modeling. An associated data and code release provides our interpolated light-curve models, interpolation implementation which can be applied to reproduce our work or extend to new models, and our multimessenger parameter inference engine.