Outline:
Radio-continuum observations are nearly insensitive to gas obscuration, hence the radio luminosity (Lradio) of star-forming galaxies has long been used as a star formation rate (SFR) tracer (Condon 1992). However, our knowledge of the radio-SFR relation is limited only to massive galaxies (stellar mass M*>1010 Msun) and up to z~4 (Delvecchio et al. 2021, A&A, 647, 123), showing a significant decrease of the ratio [SFR/Lradio] with increasing M* (left Figure), which is not yet understood. In the upcoming years, the SKA1 (MID) will routinely detect ultra-faint (~10 nJy) radio sources in low-M* galaxies and at z>4, but their star-forming content totally relies upon uncertain extrapolations (by 2 dex at 108.5 Msun) of the radio-SFR relation not yet backed up by data. This proposed PhD project capitalizes on state-of-the-art JWST surveys to get M*-complete samples down to 108.5 Msun at z~8 (right Figure, Shuntov et al. 2025), using the largest JWST/NIRCam survey to-date: “COSMOS-Web” (which the supervisors are members of). From JWST-selected galaxies, the PhD candidate will perform image stacking in deep infrared and radio bands - reaching stacked flux levels comparable to those of future SKA detections - run SED-fitting of stacked photometry and radio-AGN removal, in order to infer accurate SFR and Lradio estimates in every bin of (M*, z). The newly calibrated SFR-radio relation will become a benchmark in the SKA era, enabling accurate data-model comparisons, better radio-quiet AGN identification, and an independent assessment of the SFR density in the early Universe. The (co-)supervisors have long expertise and all core-data already in-hand to ensure the full accomplishment of the project.