2022ApJ...924...81P

We present a novel approach to the riddle of star cluster multiple populations. Stars form from molecular cores. But not all cores form stars. Following their initial compression, such "failed" cores re-expand, rather than collapsing. We propose that their formation and subsequent dispersal regulate the gas density of cluster-forming clumps and, therefore, their core and star formation rates. Clumps for which failed cores are the dominant core type experience star formation histories with peaks and troughs (i.e., discrete star formation episodes). In contrast, too few failed cores results in smoothly decreasing star formation rates. We identify three main parameters shaping the star formation history of a clump: the star and core formation efficiencies per free-fall time, and the timescale on which failed cores return to the clump gas. The clump mass acts as a scaling factor. We use our model to constrain the density and mass of the Orion Nebula Cluster progenitor clump, and to caution that the star formation histories of starburst clusters may contain close-by peaks concealed by stellar age uncertainties. Our model generates a great variety of star formation histories. Intriguingly, the chromosome maps and O-Na anticorrelations of old globular clusters also present diverse morphologies. This prompts us to discuss our model in the context of globular cluster multiple stellar populations. More massive globular clusters exhibit stronger multiple stellar population patterns, which our model can explain if the formation of the polluting stars requires a given stellar mass threshold.