A team of cosmologists has shown that as the universe recombined, neutral hydrogen and helium scattered cosmic‑microwave background (CMB) photons through Rayleigh scattering (RS). The effect adds a part of the CMB anisotropies that grows like the fourth power of frequency (ν⁴) at leading order and peaks between 200 GHz and 800 GHz.

To assess how useful this signal could be, the researchers performed a multi‑frequency Fisher‑matrix forecast that included realistic noise and foreground models. They simulated two next‑generation CMB experiments – the proposed PICO satellite and a combination of LiteBIRD with the existing Planck data – and measured how the additional RS contribution would tighten constraints on cosmological parameters. For the standard ΛCDM cosmology, adding RS improves the constraints on all parameters by at most 7 % for PICO and by less than 2 % for LiteBIRD + Planck.

However, when the analysis allows the electron mass (m_e) or the fine‑structure constant (α) to vary during recombination – scenarios motivated by attempts to resolve the present Hubble‑tension – the improvement can reach roughly 31 % for key parameters such as the Hubble constant, m_e, α, the baryon density Ω_b h², and the cold‑dark‑matter density Ω_c h².

The study also quantified the danger of ignoring RS. Neglecting the signal biases the inferred cosmological parameters by up to about one standard deviation for PICO and by less than one σ for LiteBIRD + Planck. This level of bias shows that future analyses of CMB data must include RS to avoid systematic errors.

Figure 2 in the paper depicts how changing m_e or α shifts the Rayleigh opacity as a function of redshift and frequency, confirming the ν⁴ scaling predicted for the RS signal. Table 1 lists the foreground components that were modeled in the forecast, illustrating the complexity of the covariance matrix and the importance of accurate foreground simulations for robust parameter estimation.