Solving 'cosmological puzzles' (Credit: Bingqing Sun)

Ultra-diffuse galaxies

Ultra-diffuse galaxies (UDGs) are a recently discovered class of galaxies with sizes comparable to the Milky Way but luminosities as faint as dwarf galaxies. Their origin remains under active debate.

I lead a project using the Auriga simulations, a suite of state-of-the-art zoom-in magnetohydrodynamic simulations, to investigate the formation of these puzzling systems. In Auriga, field UDGs tend to form in dark matter haloes with higher spin compared to normal field dwarfs. Satellite UDGs, on the other hand, have two distinct origins: about half were already UDGs before infall, while the other half were originally normal dwarfs that transformed into UDGs due to tidal interactions after becoming satellites.

Dwarf galaxies in filaments

In the classical picture of galaxy formation, primordial gas collapses into the potential wells of dark matter haloes. Similarly, massive dark matter filaments also act as large-scale potential wells that can trap gas. Understanding these processes in detail will be helpful in understanding the observed filamentary dependences of galaxy properties.

Using zoom-in hydrodynamic simulations, we show that cold, dense gas preprocessed by filaments is funneled into embedded low-mass haloes along the filament direction, leading to highly anisotropic gas accretion. These filament haloes exhibit higher baryon and stellar mass fractions compared to their field counterparts. Our results suggest that filaments promote gas cooling and enhance star formation in the haloes they host.

Galaxy angular momentum

Angular momentum is a fundamental property of galaxies, and understanding its origin and evolution is key to studying galaxy formation.

In Liao et al. (2017b), we tested the standard assumption that baryons and dark matter are initially well-mixed and thus acquire identical specific angular momentum through tidal torques before radiative cooling. Using non-radiative hydrodynamical simulations, we found that this assumption breaks down: gas and dark matter can become segregated due to their different physical behaviors during halo assembly. As a result, they often originate from distinct Lagrangian regions, experience different torques, and develop misaligned angular momentum vectors. This challenges the accuracy of semi-analytical models that infer gas properties solely from dark matter merger trees.

(Image: M101. Credit: Subaru, HST, Robert Gendler)

SMBH binaries in galaxy formation simulations

Current galaxy formation simulations cannot resolve the evolution of SMBH binaries below separations of ~1 kpc, due to softened gravity and limitations in SMBH subgrid modeling. During my postdoctoral research in Helsinki, I joined the KETJU team to develop the KETJU code, which can simultaneously follow galaxy (hydro-)dynamics and small-scale SMBH dynamics with post-Newtonian corrections.

To better capture SMBH behavior during gas-rich galaxy mergers, we developed a new SMBH binary accretion and feedback subgrid model. This model extends the classic Bondi–Hoyle–Lyttleton framework into the binary phase and incorporates preferential accretion onto the secondary SMBH, motivated by small-scale circumbinary disc simulations. We tested the model using idealized gas-rich disc galaxy mergers, incorporating either pure thermal or pure kinetic AGN feedback. Our model yields more physically realistic SMBH mass ratios -- an important parameter for predicting GW recoil velocities -- highlighting its relevance for future studies of SMBH coalescence and GW signals.

RABBITS series study

With the newly developed KETJU code, we launched the RABBITS (Resolving supermAssive Black hole Binaries In galacTic hydrodynamical Simulations) series to systematically study SMBH merger time-scales and improve predictions of GW event rates and the stochastic GW background.

In the first two papers, we investigated how key galaxy formation processes -- including gas cooling, star formation, SMBH accretion, and feedback -- influence SMBH coalescence. AGN feedback plays a crucial role in regulating central galaxy properties and influencing the SMBH binary evolution during the dynamical friction phase, with kinetic feedback proving especially effective at clearing central gas and shaping the stellar environment. Nuclear star formation -- triggered by gas inflows -- accelerates SMBH binary hardening through enhanced three-body interactions, reducing merger timescales. Together, these studies underscore the importance of baryonic physics in predicting realistic SMBH merger timescales for upcoming GW observations.

SMBH mergers in the early Universe

The upcoming space-based GW observatory, LISA, is expected to detect GW signals from SMBH mergers occurring at high redshifts. However, understanding the origin and growth of SMBHs in the early Universe remains an open problem in astrophysics.

In this work, we utilize the First Light And Reionization Epoch Simulations (FLARES), a suite of cosmological hydrodynamical zoom-in simulations, to study SMBH mergers at 5≲z≲10 across a wide range of environments. Most mergers in FLARES involve secondary SMBHs near the seed mass (m_seed≈1.5×10^5 M⊙) while primary SMBHs span up to 10^9 M⊙, resulting in mass ratios from q∼10^−4 to 1. The number of mergers increases rapidly towards lower redshifts, and the comoving total number density scales with overdensity as n_merger=10^−3.80 (1+δ)^4.56. Within the FLARES redshift range, LISA is expected to detect mergers with a rate of 0.030 yr^−1 for events with signal-to-noise ratio SNR≥10. Our study demonstrates the sensitivity of GW predictions at high redshifts to SMBH seed models and merger time delays, highlighting the need for improved modeling in future cosmological simulations to maximize LISA's scientific return.

(Image credit: Lovell et al. 2021)

Simulation initial conditions

Grid and glass configurations are the two most commonly used pre-initial conditions for cosmological N-body simulations.

In Liao (2018), I introduced an alternative approach: the Capacity Constrained Voronoi Tessellation (CCVT), originally developed in computer graphics. As a geometrical equilibrium state, CCVT is uniform and isotropic, follows the minimal power spectrum P(k) ~ k^4 precisely, and remains stable under gravitational evolution. We showed that CCVT performs comparably to grid and glass configurations in cosmological simulations and offers a valuable tool for investigating the numerical convergence of pre-initial conditions.

In Zhang et al. (2021), we carried out a detailed numerical convergence study, quantifying the impact of different pre-initial conditions on the final properties of dark matter haloes.