Three ghostly galaxies appear to be missing their dark matter

About 65 million light-years from Earth lie three ultradiffuse galaxies, described as “pallid objects dimly illuminated by a smattering of stars”. They are unusually large for their luminosity and have very low surface brightness.

The galaxies, DF2, DF9 and DF4, appear to lack dark matter. Scientists speculate that they formed after an ancient collision, which may have left them without it.

The result matters because dark matter has long been used to explain how galaxies move. Approximately forty years ago, astronomers plotting star rotation speeds in other galaxies found that the stars’ speeds barely fell as their distance from the centre increased. The resulting flat rotation curves became one of the strongest arguments for dark matter.

Dark matter and dark energy together make up 95% of the universe. The three galaxies therefore offer a way to examine both what dark matter may be and whether gravity needs a different explanation on large scales.

The collision scenario remains speculative. In theory, a galaxy could lose dark matter when an invisible mass is “stolen” by a more massive neighbour during a past collision. Gas could also slow down and separate from a dark halo.

Another possibility is that the galaxies were tidally whacked out of equilibrium and are slowly expanding towards ultimate dissolution. It remains unknown whether that process will occur, or exactly how any collision removed their dark matter.

Galaxies with little or no dark matter have been identified before, including NGC 1052-DF4. The apparent absence of dark matter in DF2, DF9 and DF4 creates difficulties for alternative gravity theories. Quanta magazine has published a story about observations presented as evidence against MOND.

Separate work is also probing the properties of dark matter itself. Theories of self-interacting dark matter, known as SIDM, predict that dark-matter halos eventually undergo core collapse and become extremely efficient gravitational lenses.

An analysis of 29 lensed quasars provided the first population-level constraint on the timescale for SIDM core collapse. Its results favoured SIDM with core collapse over cold dark matter, or CDM, with Bayes factors of 8:1 to 24:1.