Erratum: Can magnetized turbulence set the mass scale of stars? (Monthly Notices of the Royal Astronomical Society (2020) (496) (5072-5088) DOI: 10.1093/mnras/staa1883)

The paper Can magnetized turbulence set the mass scale of stars? was published in MNRAS, 496, 5072 5088 (2020). In the original paper we found a large number of very low-mass sink particles (representing individual protostars) near the mass resolution limit (see fig. 10 of the original paper). After publication of the paper a detailed code review was carried out that found an uninitialized variable in the sink particle algorithm that could occasionally lead to erroneous behaviour. After re-running the simulation with the more thoroughly-developed sink particle methods used in Grudíc et al. (2020) and Guszejnov et al. (2020), we found that this population of low-mass sink particles was drastically reduced (see Fig. 1), suggesting that a sub-population of these was unphysical in origin (strengthening our conclusions about the necessity of additional physics to prevent an overly top-heavy IMF). Note that these low-mass objects represented a minor fraction of the total stellar mass. Since the main subject of our analysis was the mass-weighted median mass M50, the main conclusions of the original paper are not strongly affected by this issue, as shown by Fig. 2. However we also conjectured that non-isothermal gas physics (e.g.The opacity limit for fragmentation) may be necessary to prevent an unphysically-large number of brown dwarfs from forming, as has been argued in many other works (Bate 2009; Offner et al. 2009; Lee & Hennebelle 2018; Colman & Teyssier 2019). Because a significant number of the brown dwarfs predicted by the simulation were unphysical in origin, the actual factor by which the brown dwarf population must be suppressed was overstated, and potentially our assessment of the importance of additional physics in turn. (Figure Presented).