SN 2023ixf, a Type II supernova in M101, showed infrared excess beginning around 60 days after explosion. The natural interpretation: dust. But Singh and colleagues, tracking the supernova's nebular phase evolution from 200 to 680 days, identify two completely different dust populations responsible for the infrared signal at different epochs.
The early infrared excess — appearing before day 200 — comes from pre-existing circumstellar dust. This dust was already there, shed by the progenitor star in its final centuries, now heated by the supernova's flash. It's an infrared echo: the explosion's light arrives at distant dust shells, heats them, and they re-radiate thermally. The dust didn't form from the explosion. It witnessed it.
Note (2026-11-14; read against the paper's full text): the essay's timeline differs from the paper's. The paper reports an early infrared excess already by 1.8 days (consistent with a radiative flash echo from pre-existing circumstellar dust over 1.8 to 33.6 days), a secondary near-infrared rebrightening over about 89 to 175 days that it attributes to a more extended echo from structured wind material, progressive red-wing attenuation of H-alpha from about 132 days, and carbon monoxide emission by about 217 days. It identifies the onset of H-alpha asymmetry, from about 132 days, as the first direct signature of internal dust formation and models an internal dust mass of roughly 1.5 times 10 to the minus 6 up to 6 times 10 to the minus 5 solar masses from the line profile over 140 to 418 days; the essay's statements that the infrared excess begins around 60 days and that the ejecta dust appears only after day 400 do not match this. JWST spectral modelling shows a multi-component continuum in which a cold graphite component traces lingering echo emission. The essay's closing about the mechanism switching underneath the same observable is its own.
The later signal is different. After day 400, asymmetric blueshifting of the hydrogen-alpha emission line indicates that the receding side of the ejecta is being selectively obscured. New dust is forming within the ejecta itself — condensing from the metal-rich gas as it cools below the condensation temperature. This dust is a product of the explosion, born in the debris. The infrared excess and the line asymmetry are both signatures, but of different dust in different locations made by different mechanisms.
The two dust populations overlap temporally. Between days 200 and 400, both contribute to the infrared signal. Disentangling them requires combining photometric evolution — the echo fades as the light front passes — with spectroscopic diagnostics — the line asymmetry grows as ejecta dust accumulates. Neither measurement alone separates the two sources.
The same observable — infrared excess from a supernova — has two completely different physical origins depending on when you look. The measurement stays the same. The mechanism switches underneath it.