Could hidden CMB systematics in recombination physics explain tension without invoking new low-redshift physics?

This is a great question that gets at the heart of current cosmological tensions. The short answer: modified recombination physics shows genuine promise, but the archive suggests it alone probably can't explain the full Hubble tension without some contribution from new low-redshift physics.

What recombination modifications can and can't do

Modified recombination—changes to how ionized gas couples to radiation in the early universe—can shift the sound horizon (the characteristic scale imprinted on the CMB) without invoking late-time dark energy [AG-2025.04-307]. This is attractive because it potentially resolves the tension between DESI's BAO measurements and the CMB in a single early-universe fix. However, the tension you're asking about—between CMB-inferred $H_0 \approx 67$ km/s/Mpc and local measurements around 74 km/s/Mpc—is fundamentally geometric: recombination physics sets a ruler (the sound horizon), and comparing that ruler to nearby distance measurements tests the entire expansion history.

The problem is that recombination can't act in isolation. A diagnostic called the "matter-era distance excess" reveals this clearly [AG-2026.03-1548]. This quantity measures the geometric mismatch between late-time and early-time probes. The analysis shows that low-redshift physics modifications alone cannot explain this excess—it's a genuine high-redshift signal. Modified recombination can help, but only by altering what happened between recombination and now in specific ways.

The tension between solutions

Here's where systematics enter darkly: when researchers combine pre- and post-recombination new physics (for instance, a time-varying electron mass *plus* a sign-switching cosmological constant), they find the proposals work *against each other* [AG-2024.07-006]. The matter density $\Omega_m$ becomes a battleground—one mechanism needs it larger, the other smaller—and neither combination successfully resolves the tension. This suggests that either:

1. Hidden systematics exist in how we measure or model $\Omega_m$ across datasets, or

2. No single coherent physics operates at both early and late times; the tension reflects multiple independent problems.

Can early-universe modifications alone work?

Yes, but with constraints [AG-2026.04-1527]. A model-independent reconstruction shows that purely early-time solutions *geometrically* exist and are consistent with CMB data—but they require a specific signature: a ~15% expansion rate enhancement around matter-radiation equality. This is not a recombination modification per se; it's a broader change to the pre-recombination expansion history. The question is whether such an enhancement can be realized without exotic new physics.

Several concrete early-universe proposals do attempt this: self-interacting ultralight dark matter with eV-scale mass [AG-2025.02-194], scale-invariant vacuum dynamics [AG-2026.02-021], or nonstandard recombination coupled to new dark sector interactions [AG-2025.03-358]. Near-future high-resolution CMB data (measuring the damping tail) will test whether these leave distinctive signatures [AG-2025.03-358].

The systematics question

The CosmoVerse white paper flags that observational systematics remain inadequately characterized across different probes [AG-2025.04-033]. Could unidentified biases in recombination modeling, baryon density measurements, or the precise ionization history account for part of the tension? Possibly—but the geometric constraints suggest any such systematics would have to be large and coordinated across independent datasets (CMB, BAO, supernovae) to fully dissolve the problem. That seems unlikely, though not ruled out.

Bottom line

Hidden systematics in recombination physics could plausibly account for 20–30% of the Hubble tension, but the preprints suggest the full tension probably requires *some* new physics, whether early-universe (expansion rate enhancement, dark sector interactions) or late-time (or both). Pure recombination systematics alone appear insufficient as a complete explanation [AG-2026.03-1548, AG-2024.07-006].