Metal-Poor Accretion as the Hidden Engine of Cosmic Noon Clumps

A new preprint (arXiv:2606.05284, v1, June 2026) uses near-infrared spectroscopy from LEGA-C, MOSDEF, and CANUCS to examine gas-phase metallicities in roughly 300 star-forming galaxies at 0.6 < z < 2.6. Clumpy systems sit systematically below the mass-metallicity relation while non-clumpy galaxies lie above it; clumps hosted by the most metal-poor galaxies are younger and show elevated star-formation rates, yet display no significant stellar-mass offset. Gini-M20 analysis of stellar-mass maps further indicates that most clumpy galaxies are not mergers. These findings point to cold, metal-poor accretion as the dominant trigger for kiloparsec-scale clump formation rather than the merger-driven scenario still dominant in many simulations. Earlier HST-based studies (e.g., Guo et al. 2018, ApJ) documented clump demographics but lacked the metallicity leverage now supplied here; complementary zoom-in simulations (e.g., Mandelker et al. 2020, MNRAS) already predicted that pristine inflows can both dilute metallicity and fragment disks into massive clumps, yet observational confirmation at the population level has been sparse. The present work bridges that gap, revealing that the same accretion events simultaneously explain both the metallicity offset and the clump properties, thereby challenging models that over-predict merger fractions at cosmic noon. Limitations remain: the sample is still modest for rare high-merger systems, metallicities are integrated rather than spatially resolved, and the redshift baseline does not yet reach the peak epoch of clump visibility. Future JWST/NIRSpec IFU maps of individual clumps will test whether the youngest, highest-SFR clumps indeed coincide with the lowest local metallicities predicted by accretion-driven fragmentation.