SN Refsdal Time-Delay Analysis Delivers 3% Hubble Constant at 65 km/s/Mpc
A flat-ΛCDM analysis of SN Refsdal time delays plus extended host-galaxy surface brightness yields the most precise single-lens H0 to date at 3% uncertainty. The study isolates the information gain from 77,000 HST pixels while keeping the mass model fixed, showing that lens-mass and time-delay errors are now comparable. Continued progress hinges on sub-percent delay measurements and relaxed mass-model assumptions.
The arXiv preprint reports the first cosmological inference from a strongly lensed supernova that exploits full surface-brightness reconstruction rather than point images. Four mass-model realizations were tested while holding the cluster total-mass parametrization fixed to an earlier point-like model; the added pixels tightened the H0 posterior by a factor of two, producing joint constraints H0 = 65.0 +2.0/-1.8 km/s/Mpc and Ωm = 0.30 +0.08/-0.09 in flat ΛCDM. Time-delay and lens-mass uncertainties now contribute comparably to the error budget.
This result sits between Planck CMB and local Cepheid-SN ladders, reinforcing the existing 4–6σ tension without resolving it. Because the lens model was deliberately frozen, the quoted 3% figure isolates the gain from extended-source constraints and does not yet include systematic shifts that would arise from refitting the mass parametrization. Future sub-percent time-delay measurements at cluster scales could push single-lens precision below 2%.
Next steps require independent mass models with free parameters and new cluster-scale lenses discovered by Euclid and Roman. Only then can the method move from proof-of-concept to a primary anchor in the Hubble-constant debate.
Schuldt et al.: Within three years, a second cluster-scale lensed SN with 1% time-delay precision will deliver an independent H0 constraint below 2.5% when the mass model is allowed to vary.
Sources (2)
- [1]Primary Source(https://arxiv.org/abs/2610.12595)
- [2]Supporting Source(https://arxiv.org/abs/1907.01644)