Kyle Kremer

Globular clusters are host to significant populations of dynamically active stellar remnants that connect to a variety of astrophysical sources. Using simulations performed with the Cluster Monte Carlo dynamics code, we study the formation of ultracompact binaries in which a stellar-mass black hole accretes material from a white dwarf companion in an orbit of less than an hour. These binary systems are prime multimessenger targets, as they can be observed both as luminous X-ray sources and as millihertz gravitational-wave sources detectable by the Laser Interferometer Space Antenna (LISA). We find that black hole+giant collisions are the primary mechanism through which such systems form. We model the outcomes of these “common envelope”-like events using the smoothed particle hydrodynamics code StarSmasher , and verify that these collisions yield black hole+white dwarf binaries that enter Roche contact on timescales ≲1 Gyr via gravitational-wave inspiral. We construct a mock catalog of local ultracompact X-ray sources and compare to candidate sources observed in globular clusters in the Milky Way (e.g., 47 Tuc X9) and external galaxies (e.g., RZ 2109 in NGC 4472). Finally, we compute the gravitational-wave strain for these sources, and show that of order one source may be resolvable in the Milky Way by LISA, representing a potentially powerful tool for observing new black holes in globular clusters.

Observations and cosmological simulations indicate that the early Universe hosted numerous compact, high-density stellar systems, where close encounters and physical collisions between stars were likely common. We develop a bottom-up framework for stellar dynamics in such environments, spanning systems with and without intermediate-mass and supermassive black holes, and covering regimes where stellar collisions may or may not dominate the evolution. This radially resolved analytic model connects dense star clusters in their cosmological context to observable outcomes mediated by stellar collisions. Initial conditions and environmental properties are drawn from high-resolution cosmological simulations, enabling exploration across a broad region of parameter space. The analytic predictions are validated against Monte Carlo simulations, demonstrating good agreement across key regimes. We find that stellar collisions are ubiquitous in many high-redshift environments, with runaway sequences naturally leading to the formation of very massive stars at early times. Finally, we show that high rates of destructive collisions can rapidly build up extremely dense gaseous environments around massive black holes, potentially providing an analog to the observed population of Little Red Dots.

Globular clusters (GCs) display much lower binary fractions than found among main-sequence (MS) stars in the solar neighborhood. The physical cause of this difference is debatable: does it reflect different star formation outcomes at low metallicity and/or high density, the dynamical processing of primordial binaries over cluster lifetimes, or a combination of the two? Starting from the assumption that the initial binary distribution in GCs is the same as the binary distribution observed in the solar neighborhood, we show with a straightforward analytical calculation that the dynamical dissolution of “soft” primordial binaries can largely explain the MS binary fractions in present-day GCs. We validate our estimates against a set of detailed N -body simulations with the Cluster Monte Carlo code. We find broad agreement between the binary populations and dynamical characteristics of the theoretical models and a sample of well-studied GCs in the Milky Way. The process of “black hole burning” provides a major energy source for soft binary dissolution on gigayear timescales; in the densest clusters (initial virial radii ≲ 0.5 pc), direct collisions between stars rapidly reduce the binary fraction at early times. Our results reinforce the need to include realistic initial conditions in the numerical modeling of GCs.

In the coming decade, the millihertz gravitational-wave observatory the Laser Interferometer Space Antenna (LISA) will provide the best constraints yet on the tens of thousands of close white dwarf binaries in the Milky Way, yielding unprecedented insights into the most abundant class of compact object binaries. Following inspiral via gravitational wave emission, interacting white dwarf binary pairs can lead to a multitude of outcomes, including AM Canum Venaticorum binaries, R Coronae Borealis stars, young, rapidly spinning single white dwarfs, (millisecond) magnetars, and a variety of explosive transients, most notably Type Ia supernovae. Current and future electromagnetic observations of these various outcomes coupled with the forthcoming flood of data from LISA place us on the precipice of a significant advance in our understanding of the long-term fate of white dwarf binaries. In this paper, we present a suite of mock catalogs of the Milky Way’s white dwarf merger history, created using the population synthesis code COSMIC combined with a metallicity-dependent star formation history from FIRE-2 galaxy simulations. We summarize the various merger outcomes expected (based upon varying white dwarf masses and chemical compositions) and explore ways the rates of these outcomes may vary with model uncertainties pertaining to binary evolution. We publicly release these merger catalogs as a tool for facilitating connections between gravitational-wave science and white dwarf binary astrophysics.

Stellar collisions can occur frequently in dense cluster environments and play a crucial role in producing exotic phenomena, from blue stragglers in globular clusters to high-energy transients in galactic nuclei. Successive collisions and mergers of massive stars could also lead to the formation of massive black holes, serving as seeds for supermassive black holes in the early Universe. While analytic fitting formulae exist for predicting collision outcomes, they do not generalize across different energy scales or stellar evolutionary phases. Smoothed particle hydrodynamics (SPH) simulations are often used to compute the outcomes of stellar collisions, but even at low resolution, their computational cost makes running on-the-fly calculations during an N -body simulation quite challenging. Here, we present a new grid of 27,720 SPH calculations of main-sequence star collisions, spanning a wide range of masses, ages, relative velocities, and impact parameters. Using this grid, we train machine learning models to predict both collision outcomes (merger versus disruption or flyby) and final remnant masses. We compare the performance of nearest neighbors, support vector machines, and neural networks, achieving classification balanced accuracy of 98.4% and regression relative errors as low as 0.11% and 0.15% for the final stars 1 and 2, respectively. We make our trained models publicly available as part of the package collAIder , enabling rapid predictions of stellar collision outcomes in N -body models of dense star cluster dynamics.

Binaries in the cores of globular clusters are known to prevent the gravitational collapse of the cluster, and simulations predict that the core of NGC 6397 contains a high number of white dwarfs (WDs), of which many are expected to be part of a binary system. In this work, we report the discovery of a compact binary system consisting of two WDs in the centre of the Galactic globular cluster NGC 6397. The system, known in the literature as NF1, was observed as part of a MUSE radial-velocity survey aiming at characterizing the binary population in the centre of NGC 6397. The spectral analysis of NF1 provides an effective temperature of 16 000 K and a surface gravity (log g ) of 5.72 (cgs), which is consistent with the characteristics of an extremely low-mass He-core WD. This is further supported by the mass of 0.23 ± 0.03 M _⊙ obtained from fitting the star’s spectral energy distribution using its HST magnitude in various filters. The system has a circular orbit with a period of 0.54 days. The radial velocities show a large semi-amplitude of 200 km/s, implying a minimum mass of 0.78 M _⊙ for the invisible companion, which is likely another WD, or a neutron star if the inclination of the system is smaller than about 50°. Some significant residuals in radial velocity remain with our best orbital solution, and we tested whether a model with a third body can explain these deviations. While this possibility seems promising, additional measurements are needed to confirm whether the star is actually part of a triple system.

The LIGO/Virgo/Kagra (LVK) Collaboration has detected numerous binary black hole (BBH) mergers with properties that challenge standard binary evolution scenarios, such as component masses above the pair-instability gap and high spin magnitudes. Dense stellar environments such as globular clusters provide a natural channel for producing such systems through hierarchical mergers, where black hole (BH) remnants formed in earlier mergers are retained in the cluster and undergo successive mergers. However, gravitational-wave recoil kicks often eject merger remnants from typical globular clusters, which limits hierarchical growth. Massive clusters with deeper potential wells, such as those found in giant elliptical galaxies like M87, may overcome this barrier, but direct simulations of such massive globular clusters remain computationally challenging. In this study, we present a 10-million-body cluster simulation performed with the Cluster Monte Carlo ( CMC ) code, referred to as colossus , which serves as a proxy for the most massive low-metallicity globular clusters observed in the local Universe. This simulation demonstrates that extended chains of hierarchical mergers can occur in massive globular clusters, producing BHs up to the fifth generation with masses approaching 250 M _⊙ , comparable to the most massive LVK events observed to date (e.g., GW231123). Combining the colossus simulation with the previous CMC Cluster Catalog , we develop a framework to extrapolate BBH merger predictions for the thousands of globular clusters seen in the Virgo Supercluster.