Ultralight Scalar Decay and the Hubble Tension

Abstract: We examine whether the Hubble tension, the mismatch between early and late measurements of $H_0$, can be alleviated by ultralight scalar fields in the early universe, and we assess its plausibility within UV physics. Since their energy density needs to rapidly redshift away, we explore decays to massless fields around the era of matter-radiation equality. We highlight a concrete implementation of ultralight pseudo-scalars, axions, that decay to an abelian dark sector. This scenario circumvents major problems of other popular realizations of early universe scalar models in that it uses a regular scalar potential that is quadratic around the minimum, instead of the extreme fine-tuning of many existing models. The idea is that the scalar is initially frozen in its potential until $H\sim m$, then efficient energy transfer from the scalar to the massless field can occur shortly after the beginning of oscillations due to resonance. We introduce an effective fluid model which captures the transition from the frozen scalar phase to the radiation dark sector phase. We perform a fit to a combined Planck 2018, BAO, SH$0$ES and Pantheon supernovae dataset and find that the model gives $H_0=69.9{-0.86}^{+0.84}$ km/s/Mpc with $\Delta\chi^2 \approx -9$ compared to $\Lambda$CDM; while inclusions of other data sets may worsen the fit. Importantly, we find that large values of the coupling between fields is required for sufficiently rapid decay: For axion-gauge field models $\phi F\tilde{F}/\Lambda$ it requires $\Lambda\lesssim f/80$, where $2\pi f$ is the field range. We find related conclusions for scalar-scalar models $\sim\phi\,\chi^2$ and for models that utilize perturbative decays. We conclude that these sorts of ultralight scalar models that purport to alleviate the Hubble tension, while being reasonable effective field theories, require features that are difficult to embed within UV physics.