A Galaxy Rotation Law May Also Probe Dark Matter and the Cosmos

A spiral galaxy's rotation curve is one of astronomy's most reliable clues to its mass. The faster it spins at its outskirts, the more mass—stars, gas, and dark matter—must be pulling it inward. For decades, the baryonic Tully-Fisher relation (BTFR) has compressed that idea into a tight empirical rule: a rotating galaxy's total ordinary matter is tied to its maximum circular velocity. Astronomers use the relation as a distance indicator. A new preprint by Francesco Sinigaglia asks whether it might also be a direct probe of cosmology and dark matter.

The BTFR's Day Job: Distances and H0

The BTFR connects baryonic mass, the combined mass of stars and gas, to a galaxy's maximum circular velocity. That speed reflects the gravitational pull of the entire galaxy, including its dark matter halo. It has estimated distances and the local expansion rate, but rarely cosmological parameters beyond that.

Can a Rotation Curve Remember Ωm and σ8?

The puzzle is whether the BTFR's slope, normalization, and scatter carry information about cosmology, dark matter, and astrophysical feedback at once. Supernovae and active galactic nuclei blow gas out of galaxies, changing how much baryonic mass remains. Warm dark matter, made of particles lighter than standard cold dark matter, would suppress the formation of small halos. These effects could blur one another.

Normalizing Flows Learn the BTFR's Parameter Dependencies

The study uses the DREAMS suite, 1,024 cosmological simulations that follow dark matter and gas in boxes tens of millions of parsecs across. The runs vary six parameters: the matter density Ωm, the amplitude of matter clustering σ8, two numbers controlling supernova feedback, one controlling AGN feedback, and the warm dark matter particle mass M_wdm, sampled from 1.8 to 16 keV.

For each simulated universe, the author identifies rotation-supported galaxies, measures their baryonic mass and maximum circular velocity, and bins the galaxies by velocity. The BTFR is extracted at three cosmic epochs, corresponding to redshifts 0, 0.54, and 1.05. Combining redshifts improved the inference.

A deep neural network called a normalizing flow is trained on almost all the simulations to learn the mapping from a BTFR measurement back to the six parameters. It outputs posterior distributions: not single answers, but ranges of plausible values with probabilities attached. The final test uses 50 simulations held out from training.

Ωm and σ8 Recovered With Subpercent Deviations

The network recovered Ωm and σ8 with high accuracy in the test simulations. The paper reports subpercent average deviations and median precision of about 2.6 percent for Ωm and 3.9 percent for σ8. Ωm describes how much matter the universe contains; σ8 measures how clumpy matter is on large scales. The two parameters were anti-correlated, with a correlation coefficient of about -0.74.

That matters because it suggests a low-redshift galaxy scaling relation could provide cosmological information independent of the cosmic microwave background, galaxy clustering, weak lensing, and other traditional probes. Independence is valuable when different probes disagree, as they do in current discussions of the σ8 tension—a mismatch between some measurements of cosmic clumpiness.

Warm Dark Matter's 30–35 Percent Blur

The simulations also allowed the network to infer the warm dark matter particle mass, but less sharply. The reported precision is about 30 to 35 percent.

The physical reason is subtle. Warmer, lighter dark matter suppresses the formation of small halos. Fewer low-mass halos form, which affects the BTFR. The analysis found that σ8 and M_wdm anti-correlate, with a correlation coefficient of about -0.40. If dark matter is warmer, a higher σ8 can help restore small halos in the fit. That trade-off makes the warm dark matter mass harder to pin down, and the paper flags it as a possible systematic uncertainty.

SNe Feedback Pinned; AGNs Lost to Shot Noise

The framework also constrained the two supernova feedback parameters, with median precision around 10 to 13 percent. It could not constrain AGN feedback. The reason is practical: the simulation boxes are small, and AGNs are sparse, so shot noise—random scatter from too few objects—is likely larger than the signal.

This limitation is instructive. The BTFR is not a clean cosmological ruler that ignores galaxy physics. It is a joint record of cosmology, dark matter, and baryonic processes. The hope is that one measurement can constrain all of them together.

A Proof of Concept, Not a Data Measurement

This is a proof of concept, not a measurement of the real universe. The BTFR data come from simulations, and the inferred parameters are compared with the parameters that generated those simulations.

The paper notes two major hurdles. First, different galaxy formation models do not fully agree, so cosmological inference from the BTFR must be tested against model choices. Second, real observations bring systematics: noisy rotation curves, uncertain galaxy inclinations, selection effects, incompleteness at low masses, and cosmic variance. Those effects must be modeled before the method can be trusted on data.

The next step is the Square Kilometer Array, which the paper says will observe the BTFR directly out to redshift near 1, the range the paper simulates. Combining BTFR constraints with other dark matter probes, such as the Lyman-alpha forest, could sharpen the picture of what dark matter is.

The BTFR earned its place in astronomy by turning spinning galaxies into distance markers. The new work suggests that the same relation may also preserve a faint record of the cosmos that shaped those galaxies—and that reading it will require both future SKA observations and a careful accounting of the messy astrophysics in between.