Confronting interacting radiation models for the Hubble tension with Lyman-<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mi>α</mml:mi></mml:math> data

Models of interacting dark radiation have been shown to alleviate the Hubble tension. Extensions incorporating a coupling between dark matter and dark radiation (DM-DR) have been proposed as combined solutions to both the Hubble and ${S}_{8}$ tensions. A key feature of these extended models is a break in the matter power spectrum (MPS), suppressing power for modes that enter the horizon before the DM-DR interactions turn off. In scenarios with a massless mediator, modes that enter before matter-radiation equality get suppressed, whereas for massive mediators, the break is determined by the mediator mass, a free parameter. In this work, we test these models against probes of LSS: weak lensing, CMB lensing, full-shape galaxy clustering, and eBOSS measurements of the 1D $\mathrm{Ly}\ensuremath{\alpha}$ forest flux power spectrum. The latter are the most constraining since they probe small scales where many models predict the largest deviations. In fact, already within $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$, the eBOSS $\mathrm{Ly}\ensuremath{\alpha}$ data are in significant tension with Planck CMB data, with the $\mathrm{Ly}\ensuremath{\alpha}$ data preferring a steeper slope of the MPS at $k\ensuremath{\sim}h{\mathrm{Mpc}}^{\ensuremath{-}1}$. We find that the simplest dark radiation models, which improve the Hubble tension, worsen the fit to the $\mathrm{Ly}\ensuremath{\alpha}$ data. However, models with DM-DR interactions can simultaneously address both tensions.