Most massive stars are expected to end their lives in brilliant supernovae, but theory predicts that stars above roughly 15 solar masses can sometimes bypass the explosion entirely, collapsing directly into a black hole. A new preprint (arXiv:2604.05019, not yet peer-reviewed) by Raquel Forés-Toribio and collaborators re-examines JWST infrared imaging and photometry of the leading candidate for this process: N6946-BH1 in the spiral galaxy NGC 6946, located about 7.7 megaparsecs away.

The team's methodology involved high-resolution point-spread-function fitting on JWST MIRI and NIRCam data to disentangle the candidate from its surroundings, followed by spectral energy distribution (SED) modeling. They identified four unrelated near-infrared stellar neighbors within a few arcseconds; these stars do not contribute to the mid-infrared excess previously observed. The SED is best fit by a remnant source of approximately 50,000 solar luminosities shrouded in an optically thick silicate dust shell containing grains up to 3 micrometers across. This dust absorbs nearly all light below 2 micrometers and re-radiates it at longer wavelengths, producing the characteristic mid-IR signature.

This work builds on the seminal 2017 discovery paper by Adams et al. (Astrophysical Journal), which used Large Binocular Telescope monitoring over nearly a decade to document a slow, years-long fade in optical bands between 2009 and 2015 with no detectable supernova outburst — a sample of one object tracked at high cadence. That study noted the star's initial mass was likely 15–25 solar masses but could not rule out source confusion or alternative explanations such as an exotic stellar merger. The new preprint synthesizes those optical data with a third source — theoretical models of stellar mergers from Kochanek and others (e.g., 2014–2020 papers on luminous red novae) — to perform a direct luminosity-ratio comparison across 4 Galactic and 7 extragalactic merger remnants.

The contrast is striking and what much early coverage missed: merger remnants brighten by factors of 10–100 in the late infrared phase as they expand and cool, whereas N6946-BH1's remnant is roughly 10 times dimmer than its progenitor. The authors explicitly test whether asymmetric, disk-like dust geometries could mimic the observed dimming; they cannot. This rules out merger impostors more cleanly than previous analyses and sharpens the observational template for failed supernovae: prolonged optical disappearance, isolated mid-IR dust echo, and net dimming relative to the pre-collapse star.

Contextually, only one other credible failed-supernova candidate exists — M31-2014-DS1 in Andromeda — making the sample size tiny. Both objects suggest that the fraction of massive stars taking this route may be 10–30 percent, a crucial parameter for predicting black-hole birth rates that feed gravitational-wave observatories like LIGO/Virgo. The pathway illuminates a poorly understood regime of stellar evolution where neutrino losses sap the rebound shock before it can unbind the envelope.

Limitations must be stated clearly. The study relies on broadband photometry rather than spectroscopy, so dust chemistry and exact remnant temperature remain somewhat degenerate. The merger comparison sample is modest (11 events total), and dust models assume spherical symmetry as a baseline. As a preprint, the results have not yet survived formal peer review. Nonetheless, by isolating the true remnant emission from neighboring stars, this analysis supplies the clearest observational signature yet of a star that simply winked out — providing astronomers a search template for future wide-field infrared surveys and tightening constraints on the initial-mass function slice that ends in silence rather than spectacle.