利用 2009 到 2022 年、每 6 个月一次的中红外监测，该源自 2014 年起在 2 年内增亮约 50%，之后在 1 年内降到低于初始通量，并持续变暗到 2022 年。独立的光学与近红外检查加强了这一判断：2016 到 2019 年光学发射下降约 100 倍，地面光学资料在 2023 年无法探测到它，Hubble 在 2022 年未见光学对应体且只见微弱近红外发射，而 Keck 在 2023 年的光谱确认了微弱近红外源。前身星初始质量估计约 13 个太阳质量，并在强风驱动的质量流失后于死亡时约为 5 个太阳质量。
一个关键物理含义是：在这一质量附近，核心坍缩的结果并非决定性的；中微子驱动的激波可能会也可能不会抛射包层，而激波失败会允许回落并直接形成黑洞。另一个唯一相似候选体 N6946-BH1（reported in 2010）位于 NGC 6946，距离约 25 million light-years（about 7.7 megaparsecs, original: 25 million light-years; about 2.4 x 10^20 kilometers），约比此次 M31 事件远 10:1，因此其资料品质较差。从统计上看，这是迄今针对银河系与邻近星系族群最大规模的红外变源搜寻，但只得到 1 个明确新案例，意味著要么可探测性偏低，要么大质量恒星死亡被低估；值得注意的是，这个事件实际上超过 5 years 几乎未被注意，因而促使未来利用 10-year Vera Rubin LSST program 扩大样本。
The study reports a likely “failed supernova” in the Andromeda Galaxy (M31) where a supergiant, M31-2014-DS1, appears to have collapsed directly into a stellar-mass black hole instead of exploding. This channel of stellar death is predicted by theory but has had little direct observational support, even though models suggest it may be common. The team led by Kishalay De (Columbia University) identified the event in public NASA NEOWISE archives, highlighting that a major signal can remain unnoticed for years in large time-domain datasets.
Using 6-month cadence mid-infrared monitoring from 2009 to 2022, the source brightened by about 50% over 2 years beginning in 2014, then dropped below its initial flux within 1 year and kept fading through 2022. Independent optical and near-infrared checks strengthened the case: optical emission declined by a factor of about 100 from 2016 to 2019, ground-based optical data could not detect it in 2023, Hubble saw no optical counterpart in 2022 and only faint near-infrared emission, and Keck spectroscopy in 2023 confirmed a faint near-infrared source. The progenitor was estimated at about 13 solar masses initially and about 5 solar masses at death after strong wind-driven mass loss.
A key physical implication is that core-collapse outcomes at around this mass are not deterministic: neutrino-driven shocks may or may not eject the envelope, and shock failure allows fallback and direct black-hole formation. The only other similar candidate, N6946-BH1 (reported in 2010), lies in NGC 6946 at about 25 million light-years (about 7.7 megaparsecs, original: 25 million light-years; about 2.4 x 10^20 kilometers), roughly 10:1 farther than this M31 event, so its data quality is poorer. Statistically, this was the largest infrared-variable search across Milky Way and nearby-galaxy populations yet yielded only 1 clear new case, implying either low detectability or undercounting of massive-star deaths; notably, this event went effectively unnoticed for more than 5 years, motivating larger future samples from the 10-year Vera Rubin LSST program.