Globular Clusters: A Palimpsest of Ancient Gas Clouds

‹›

Imagine an ancient manuscript, where a new story is written over an old text. Over time, the upper layer wears thin, and lines of the past show through. That’s exactly how globular clusters—the oldest stellar systems in the Universe—are structured, like a palimpsest. Since the early 20th century, when Edwin Hubble used them as beacons for measuring distances, these objects have remained key to understanding galaxy evolution. Their chemical composition, measured for decades via spectroscopy, seemed contradictory: stars in a single cluster show strong anti-correlations between nitrogen and oxygen, as if they were born under completely different conditions. Traditional self-pollution models assumed that the first generation of massive stars polluted the gas for the second generation, but this picture required tens of times more 'ink'—stellar material—than could possibly exist in the cluster. A new study offers an elegant way out.

Using THESAN-ZOOM simulations, akin to a time machine, researchers peered into an era when the Universe was less than a billion years old. In these calculations, which account for the constraints of the speed of light and radiation transfer, they discovered giant molecular clouds with a chemical profile remarkably similar to globular clusters. The key turned out to be not local pollution, but galactic memory. A burst of star formation 20–50 million years earlier enriches the galaxy with oxygen from supernova explosions; then, as star formation fades, old stars on the asymptotic giant branch saturate the medium with nitrogen. Later, the cooled oxygen-rich cloud falls back into the galaxy, colliding with nitrogen-rich gas—just like an accretion stream onto a black hole, only spanning hundreds of light-years. This collision creates a compression zone where future cluster stars inherit ready-made anti-correlations, as if reading a record left by previous episodes of the galaxy's life.

This scenario untangles the mass deficit knot: no giant first generation is needed. All stars form simultaneously from already mixed gas, which is fully consistent with observations from Hubble, which spent decades collecting spectra of hundreds of clusters. Moreover, there’s no need to invoke mechanisms of selective evaporation of stars with normal composition—another headache for old models. The chemical palimpsest works naturally, blurring the line between stellar generations and shifting the focus to the gas cloud's prehistory.

This work builds a bridge from stellar chemistry to the baryon cycle of galaxies and even to the distribution of dark matter, the importance of which was first pointed out by Vera Rubin through galaxy rotation curves. In the coming years, simulations will be able to directly track cluster evolution on cosmological scales, including the production of sodium—a key element for testing the Na–O anti-correlation. Meanwhile, next-generation gravitational wave detectors will catch ripples from mergers of compact objects in cluster cores, providing independent insight into their internal dynamics. We are moving from self-pollution models to galactic archaeology, where every globular cluster becomes a palimpsest inscribed with the history of its parent system. And in this text, perhaps, lie the keys to unraveling the formation of the first galaxies and the nature of dark matter.