The MEGATRON project has released its first major findings, using supercomputing simulations to link James Webb Space Telescope observations of infant galaxies to chemical relics found in the Milky Way. The results, published in the Open Journal of Astrophysics, model the evolution of the early universe.
The earliest stars—known as Population III or Pop III stars—formed from pristine hydrogen and helium left over from the Big Bang, forging heavier elements like carbon, oxygen, and iron that were later spread across the cosmos via supernova explosions. Cosmologists have long struggled to explain how the universe transitioned from this darkness into the complex structures seen today.
A collaborative team from the University of Bath, the University of Chicago, and the Institut d’Astrophysique de Paris has resolved that disconnect by connecting direct observations of distant infant galaxies with the chemical composition of ancient stars orbiting in our own galactic neighborhood.
MEGATRON Uses Massive Computing Power
The MEGATRON project has already been awarded 40 million processor hours on the United Kingdom’s national supercomputers—an allocation equivalent to running five million laptops simultaneously for a full year. Initiated in 2023 and scheduled to run through 2030, the project requires this scale to model parsec-scale structures across an entire cosmological volume.
The simulation suite tracks the movement of dark matter particles with a mass of roughly 2.5 × 10⁴ solar masses, tracing a proto-Milky-Way environment from the early universe down to a redshift of approximately z ≈ 8. Spatial resolution shifts from 2.5 parsecs to about 5 parsecs over the course of the run.
Simulations incorporate non-equilibrium chemistry involving more than 80 primordial species, molecules, and metal ions. Starlight alters the thermal and chemical state of surrounding gas, which dictates cooling rates, gas collapse, and subsequent star formation.
Simulations Explain Iron Levels in Dwarf Galaxies
Iron concentrations appear constant regardless of mass in extremely faint systems, breaking a general Milky Way trend where smaller, fainter galaxies contain less iron. Previous computer models failed to replicate this behavior.
The MEGATRON simulations suggest explosions from exotic Population III stars cause this. Larger galaxies possessed enough gravitational pull to retain the metals produced by these primordial stars—which produced higher proportions of iron than standard supernovae—while smaller systems lost them to surrounding space.
The Chicago research team predicts galaxies devoid of active stars that continue to shine, either because their stars collapsed directly into black holes or because the systems contained only gas.
Models Link Cosmic Dawn to Stellar Fossils
Four initial studies published in the Open Journal of Astrophysics establish a unified physical framework. Researchers can now test competing theories of early star formation by comparing simulation outputs against two complementary datasets.
The Bath-Chicago-Paris collaboration demonstrated that simpler models often underestimate how much stellar radiation influences gas surrounding evolving galaxies by accounting for the interplay between starlight, gas dynamics, and chemical dispersal.
The research group is developing next-generation models using massive national supercomputing grants to refine spatial resolution as large-scale stellar surveys map out ancient stars in our cosmic backyard. Additional papers are expected as the project continues toward its 2030 completion date.
“The JWST gives us a direct glimpse of the infant cosmos, while stellar archaeology allows us to study the relics of those earliest times in our own Galactic neighbourhood, MEGATRON provides a physical bridge between the two.”
Dr Martin Rey, Department of Physics at the University of Bath