The Universe's initial conditions, in particular baryon and cold dark matter (CDM) isocurvature perturbations, are poorly constrained on sub-Mpc scales. In this paper, we develop a new formalism to compute the effect of small-scale baryon perturbations on the mean free-electron abundance, thus on cosmic microwave background (CMB) anisotropies. Our framework can accommodate perturbations with arbitrary time and scale dependence. We apply this formalism to four different combinations of baryon and CDM isocurvature modes, and use Planck CMB-anisotropy data to probe their initial amplitude. We find that Planck data is consistent with no small-scale isocurvature perturbations, and that this additional ingredient does not help alleviate the Hubble tension. We set upper bounds to the dimensionless initial power spectrum $\Delta_{\mathcal{I}}^2(k)$ of these isocurvature modes at comoving wavenumbers $1~\textrm{Mpc}^{-1} \le k \le 10^3$ Mpc$^{-1}$, for several parameterizations. For a scale-invariant power spectrum, our 95% confidence-level limits on $\Delta_{\mathcal{I}}^2$ are 0.023 for pure baryon isocurvature, 0.099 for pure CDM isocurvature, 0.026 for compensated baryon-CDM perturbations, and 0.009 for joint baryon-CDM isocurvature perturbations. Using a Fisher analysis generalized to non-analytic parameter dependence, we forecast that a CMB Stage-4 experiment would be able to probe small-scale isocurvature perturbations with initial power 3 to 10 times smaller than Planck limits. The formalism introduced in this work is very general and can be used more widely to probe any physical processes or initial conditions sourcing small-scale baryon perturbations.
Read full abstract