A new preprint has tightened the constraints on planetary-mass primordial black holes (PBHs) as candidates for dark matter, showing that a recent claim of their discovery toward the Andromeda Galaxy was mistaken. The original preprint by Sugiyama et al. reported 12 short-timescale microlensing events (lasting less than one day) in high-cadence Subaru Hyper Suprime-Cam images of M31, suggesting a large population of PBHs with masses around that of Earth or smaller could explain the dark matter halos of both the Milky Way and Andromeda.

This independent reanalysis, also a preprint (arXiv:2604.00111), used a different difference-image-analysis photometric pipeline on the exact same dataset. Researchers examined the light curves of all 12 candidates and found they showed asymmetric shapes and variability across multiple nights—properties inconsistent with true microlensing, which should produce symmetric, single-event brightening. Their conclusion: all 12 objects are variable stars (10 RR Lyrae stars, one eclipsing binary, and one unclassified variable), not gravitational microlensing by black holes.

The study methodology involved careful vetting for stellar variability using an independent pipeline, directly addressing the original work's apparent shortcomings in variable-star rejection. The sample consisted of the full set of 12 reported candidates from the Subaru observations; no new data were collected. Limitations include reliance on reprocessing existing data from one target galaxy and the inherent challenge of distinguishing rare microlensing from common stellar variability in crowded fields.

This result aligns with—and was motivated by—prior null searches from the OGLE survey toward the Magellanic Clouds, which also found no evidence for such short-timescale events. Synthesizing these with a comprehensive 2021 review by Carr, Kohri, Sendouda, and Yokoyama (arXiv:2002.12778), which identified a possible window for planetary-mass PBHs to constitute all dark matter, the new null result further closes that window. Earlier microlensing campaigns (MACHO, EROS) had already ruled out larger compact objects, while this work targets the asteroid-to-planet mass range.

What the original Sugiyama coverage missed or downplayed was the inconsistency in both the temporal and spatial distribution of their candidates with expected microlensing signals, plus the well-known difficulty of variable-star contamination in high-cadence surveys. The pattern is familiar: initial excitement about massive compact halo objects (MACHOs) in the 1990s was later tempered by improved analysis showing they could not account for all dark matter.

This fits into the high-stakes, multi-decade effort to identify dark matter. With WIMP searches yielding null results at the LHC and direct-detection experiments, and axion haloscopes scaling up, each constrained PBH mass range narrows the viable parameter space. While PBHs remain interesting for other reasons (such as early-universe physics), the evidence increasingly suggests they are not the primary constituent of dark matter.