Small-scale clumping at recombination and the Hubble tension

Despite the success of the standard $\Lambda$CDM model of cosmology, recent data improvements have made tensions emerge between low- and high-redshift observables, most importantly in determinations of the Hubble constant $H_0$ and the (rescaled) clustering amplitude $S_8$. The high-redshift data, from the cosmic microwave background (CMB), crucially relies on recombination physics for its interpretation. Here we study how small-scale baryon inhomogeneities (i.e., clumping) can affect recombination and consider whether they can relieve both the $H_0$ and $S_8$ tensions. Such small-scale clumping, which may be caused by primordial magnetic fields or baryon isocurvature below kpc scales, enhances the recombination rate even when averaged over larger scales, shifting recombination to earlier times. We introduce a flexible clumping model, parametrized via three spatial zones with free densities and volume fractions, and use it to study the impact of clumping on CMB observables. We find that increasing $H_0$ decreases both $\Omega_m$ and $S_8$, which alleviates the $S_8$ tension. On the other hand, the shift in $\Omega_m$ is disfavored by the low-$z$ baryon-acoustic-oscillations measurements. We find that the clumping parameters that can change the CMB sound horizon enough to explain the $H_0$ tension also alter the damping tail, so they are disfavored by current Planck 2018 data. We test how the CMB damping-tail information rules out changes to recombination by first removing $\ell>1000$ multipoles in Planck data, where we find that clumping could resolve the $H_0$ tension. Furthermore, we make predictions for future CMB experiments, as their improved damping-tail precision can better constrain departures from standard recombination. Both the Simons Observatory and CMB-S4 will provide decisive evidence for or against clumping as a resolution to the $H_0$ tension.