The question of why galaxies stop forming stars and change their shape has been at the center of modern astrophysics for several decades. Observations show a clear bimodality: blue disks and red spheroids are separated by mass, size, and activity. The classical explanation assigns the leading role to galaxy mergers, which destroy disks and trigger black hole activity. However, mergers increase a galaxy's size, whereas quiescent galaxies are smaller than star-forming ones at the same mass. This contradiction points to an additional mechanism linking structure and quenching. The work fits into the line of research that uses cosmological simulations to trace galaxy evolution and search for the physical causes of observed correlations. Nearby lie studies of mergers as a source of black hole spin and of the influence of supermassive black holes on stellar clusters, but this study stands out for its detailed separation of channels and its test of the role of morphology.
The authors used the flagship EAGLE cosmological hydrodynamic simulation, which reproduces the evolution of dark matter and baryons in a cube 100 megaparsecs on a side. They selected about 2200 central galaxies with stellar masses of at least 10^10 solar masses and traced their main merger branches back in time to a redshift of z≈4. Morphology was described by three parameters: size (half-mass radius), the disk-to-total mass ratio (D/T), and triaxiality (T), which distinguishes oblate from prolate shapes. Merger history was assessed from the fraction of stars born outside the main branch (the ex-situ fraction). They then compared the evolutionary trajectories of star-forming and quiescent galaxies, controlling for mass and merger history, and computed correlations between morphology and black hole mass.
Down to a redshift of z≈1–2, the D/T and triaxiality trajectories of future quiescent and star-forming galaxies are indistinguishable, and then they diverge at the same time as quenching begins. The difference in size and black hole mass arises earlier, at z≈2. Four channels were identified. First: mergers simultaneously increase size, suppress rotation, raise triaxiality, and grow the black hole, which then quenches star formation. Second: at a fixed merger history, compact galaxies with low D/T grow their black holes faster and quench more often. The correlation between D/T and black hole mass is negative (R≈-0.5) in all samples. Third: at equal stellar and black hole masses, future quiescent galaxies are more compact by 0.1–0.2 dex, which points to a role for size in feedback efficiency. Fourth: disk instability turns compact oblate disks into prolate systems with increasing size and suppressed rotation, without any growth in stellar mass. This secular channel produces about half of the prolate galaxies at a mass of about 10^10.6 solar masses.
The work shows that morphology plays an active role in black hole growth and quenching rather than being a passive consequence. This sets it apart from neighboring studies that focused on individual aspects, such as mergers or the influence of black holes on clusters. Here, for the first time in a single simulation, four physical channels are separated and the contribution of disk instability is quantified. The result matters for semi-analytic models: reproducing the observed size difference between quiescent and star-forming galaxies is possible only if black hole accretion depends explicitly on galaxy structure.
The topic could develop toward a more detailed study of the physics of disk instability and its connection to gas accretion onto the black hole. It would be interesting to test how the identified channels depend on the simulation resolution and on the parameters of the feedback model. Comparison with other simulations, such as IllustrisTNG and SIMBA, will help reveal which results are universal and which depend on the implementation of the physics.
The results will influence the development of semi-analytic galaxy models and the interpretation of observational data on galaxy sizes and shapes.
The next step is to test the simulation's predictions against observations, for example data on compact massive galaxies at z≈2 and on prolate galaxies in the local Universe.
The work is directly connected to the unsolved problem of the origin of galaxy morphological bimodality and the role of supermassive black holes in halting star formation. It also addresses the question of why quiescent galaxies are more compact than star-forming ones and proposes a concrete mechanism through the modulation of black hole growth.
In the EAGLE simulation, galaxies with a mass of about 10^10.6 solar masses that never experienced major mergers can turn from a flat disk into a prolate "cigar" through internal instability, while barely gaining any stellar mass.