Reionization's two clocks disagree. The CMB clock, Planck's polarization-derived optical depth, prefers a relatively late, compact reionization centered near z of 7.7. The quasar clock, absorption spectra and Lyman-alpha emitter statistics tracing the actual state of the intergalactic medium, tells a messier story: islands of neutral hydrogen persisting to z near 5.3 to 5.5, large sightline-to-sightline scatter in opacity, and hints of ionizing activity beginning much earlier (Planck Collaboration 2020; Bosman et al. 2022).
The mismatch is one of duration and shape rather than mere midpoint. Mapping the CMB's single integrated number onto an ionization history requires a model of how reionization proceeded, and the simple histories ΛCDM prefers, a smooth sigmoid driven by the gradually building galaxy population, struggle to accommodate both clocks at once: ending late enough for the quasar data while staying compact enough for the optical depth, and patchy enough for the observed opacity scatter while smooth enough for the source population that drives it. The repairs adjust escape fractions, faint-source budgets, and clustering prescriptions per dataset, the familiar signature of a process whose underlying timeline the model is forcing into the wrong shape.
The standing sharpens quickly: JWST is measuring the actual ionizing output of early galaxies, 21-cm experiments are closing in on the neutral-hydrogen history directly, and the duration question, one epoch or many, is becoming empirical.