Dwarf-galaxy simulations find supernova energy matters more than failure rates

Core-collapse supernovae (CCSNe) are more than spectacular stellar deaths. They heat and stir a galaxy’s gas, can help drive gas out into space, and return newly made elements to the material from which future stars form. But not every massive star explodes successfully, and successful explosions do not all have the same energy. How much might those differences change the evolution of a small galaxy?

Natalia Lahén and colleagues incorporated a new prescription for supernova outcomes into SPHGal, a computer code for galaxy evolution. The model uses a star’s carbon–oxygen core mass and metallicity—a measure of how much material heavier than hydrogen and helium it contains—to estimate whether it explodes and, if it does, its explosion energy. It also allows for black-hole formation when some supernova ejecta falls back onto the remnant.

The team ran five versions of the same isolated, low-metallicity dwarf galaxy for one billion years. The versions varied how supernovae were modelled, including a conventional case that assigns successful explosions a fixed energy of one “Bethe” (10⁵¹ ergs), a model with progenitor-dependent energies, and a case with lower fixed energy. Another version switched off photoionizing radiation from stars, allowing the researchers to explore the effects of that feedback channel too. This controlled numerical experiment shows what changes within this particular setup when the input assumptions change.

Weaker average explosions reshape the simulated galaxy

The new model produced about 2.7 times as many black holes as the conventional model. Yet the lower number of successful supernovae had little effect on the galaxy’s star formation by itself, the authors report. In their simulations, failed explosions were too small a fraction of events to substantially change that outcome.

The larger difference came from explosion energy. In the progenitor-dependent model, successful supernovae had an average energy of about 0.33 Bethe, roughly a third of the one-Bethe value used in the conventional version. With less energy injected on average, the galaxy formed stars at a higher rate and with less variation over time. Its outflows were weaker, and the mass-loading factor—the outflow rate relative to the star-formation rate—was lower. The authors also found greater star-cluster formation efficiency and more retention and recycling of supernova-produced metals in star-forming regions.

Changing the assumed energy affected more than individual blasts. It altered how effectively gas was removed from the galaxy and how much enriched gas remained available to make stars. Black-hole formation contributed to removing matter from the galaxy’s baryon cycle, but outflows remained the larger route of mass removal in these models. Most compact-remnant mass, meanwhile, was stored in white dwarfs rather than black holes or neutron stars.

What the comparison can—and cannot—tell us

The results complicate a common modelling shortcut. The authors note that many lower-resolution galaxy simulations use one-Bethe explosions and can match observed galaxy scaling relations with that choice. Their dwarf-galaxy simulations instead produce lower outflow rates with the new, lower-energy prescription. This difference between modelling approaches does not prove that one is generally wrong: the paper does not survey cosmological galaxies or show that its prescription will reproduce every observed galaxy property.

Several limits matter. The study follows one idealized galaxy in isolation, at an initial metallicity of about one-tenth the Sun’s, rather than a population of galaxies growing in a cosmological environment. The authors identify uncertainties in their implementation. For example, the treatment of partial fallback after some successful explosions is an adopted modelling choice that may need revision. The supernova yields used were not calibrated to the variable explosion energies, so the paper does not provide a detailed element-by-element abundance analysis. The simulations also omit pristine gas flowing into the galaxy, which the authors say means their enrichment estimates may be upper limits. Finally, the model does not include binary-star evolution, which can change the structure and fate of massive stars.

This work is an arXiv preprint; the manuscript lists journal publication details as pending. Its findings should be read as results from a specific simulation study, not settled consensus. The central takeaway is a testable modelling point: in this dwarf galaxy, average supernova energy had a stronger influence on star formation and outflows than the added explosion failures did. Further simulations, improved stellar-evolution inputs, and comparisons with observations will be needed to determine how broadly that pattern applies.