Dynamical Friction Constraints on the Dark Matter Hypothesis Across Astronomical Scales

TL;DR

This work applies Chandrasekhar dynamical friction as a self-consistency test for the existence of particle dark matter halos across scales from individual stars to galaxy groups. Using analytical, semi-analytical, and N-body approaches, it shows that in many systems the required dark halos would induce orbital decay on timescales shorter than the systems’ ages, in tension with observations (e.g., ultrafaint dwarfs, wide binaries in Reticulum II, Fornax globular clusters, fast galactic bars, MW satellites, the MW/LMC/SMC triple, and the M81 group). The results collectively challenge the standard dark-matter interpretation of these kinematics, while discussing possible escapes via core formation or exotic DM models, which in turn introduce their own conflicts with structure formation. Overall, the paper argues that dynamical-friction-based constraints significantly undermine the particle dark matter hypothesis unless new physics (e.g., very large de Broglie wavelengths) is invoked, and casts doubt on the universality of Newtonian gravity supplemented by cold dark matter for galaxy-scale dynamics.

Abstract

Dynamical friction implies a consistency check on any system where dark matter particles are hypothesised to explain orbital dynamics requiring more mass under Newtonian gravity than is directly detectable. Introducing the assumption of a dominant dark matter halo will also imply a decay timescale for the orbits in question. A self-consistency constraint hence arises, such that the resulting orbital decay timescales must be longer than the lifetimes of the systems in question. While such constraints are often trivially passed, the combined dependencies of dynamical friction timescales on the mass and orbital radius of the orbital tracer and on the density and velocity dispersion of the assumed dark matter particles leads to the existence of a number of astronomical systems where such a consistency test is failed. Here, we review cases from stars in ultrafaint dwarf galaxies, galactic bars, satellite galaxies, and, particularly, the multi-period mutual orbits of the Magellanic Clouds, as recently inferred from the star formation histories of these two galaxies, as well as the nearby M81 group of galaxies, where introducing enough dark matter to explain observed kinematics leads to dynamical friction orbital decay timescales shorter than the lifetimes of the systems in question. Taken together, these observations exclude dark matter halos made of particles as plausible explanations for the observed kinematics of these systems.

Figures (5)

• Figure S1: The red curves show loci of constant $\tau_{{\rm DF}\sigma}$ at the values shown in the labels at the top left. Inferred positions in the ($r_{1/2},\sigma$) parameter space shown are given by the points with error bars for a small sample of 4 Galactic UFD satellite galaxies. In particular, Ursa Major III has dynamical friction internal decay timescales of <2 Gyr, which precludes the presence of a dominant cuspy dark matter halo (an NFW profile, NFW1997) for this system, of the type shown by Errani2024 to be required to explain the system's survival against galactic tides, making it an impossible object under a Newtonian/dark matter scenario.
• Figure S2: Dynamical friction internal binary orbital decay timescales as a function of the internal separation distances, $\rm s_{ob}$, for the wide binaries in Reticulum II observed by Safarz2022, shown by the solid curve. The thin curves show the confidence intervals for the timescales given, and the horizontal line the age of the stellar population of this galaxy at $13.5$ Gyr. Decay timescales become shorter than the age of the stellar population for $\rm s_{ob}>0.56$ pc, indicating substantial decay is expected to have occured for wide binaries in this range.
• Figure S3: The baryonic mass ($\approx M_*$) vs. dark matter halo mass, $M_{200} \approx M_{\rm DM}$, from the APOSTLE $\Lambda$CDM simulation (using the smooth-particle-hydrodynamics code P-GADGET3, Sales+2017), in which the average particle gas mass is $1.0\times 10^4\,M_\odot$ and the average dark matter particle mass is $5.0\times 10^4\,M_\odot$, with a maximal softening of 94 pc. Dark, filled circles indicate the results of individual AP-L1 galaxies. Crosses indicate galaxies in halos considered not converged numerically, with the convergence limit being $6.0\times 10^9\,M_\odot$. The thick blue line indicates the median trend for the AP-L2 simulation set. This figure highlights that the $\Lambda$CDM model of hierarchical structure formation theory cannot produce galaxies with baryonic masses $<\,{\rm few}\,10^6\,M_\odot$ that are in dark matter halo masses $>10^{10}\,M_\odot$. The Chandrasekhar dynamical friction test on the four dSph satellite galaxies excludes the shaded region, with the shading being stronger towards lower baryonic masses based on fig. 1 of Sales+2017.
• Figure S4: The three pair-wise distances vs. time in the MW/LMC/SMC triple system, assuming each component has a dark matter halo corresponding to $\Lambda$CDM theory. The best solution obtained by OehmKroupa2024 for the orbit of the MW/LMC/SMC triple system is documented. Given the observational constraints on the proper motions of the LMC and SMC, the solution shown has a probability of $10^{-9}$, which is not a viable one, but it is the best here, fulfilling the condition that the LMC and SMC needed to have an encounter 1--4 Gyr ago with an encounter distance of 20 kpc or smaller in order to have launched the gas that evolved into the currently observed Magellanic Stream. The LMC--SMC orbit is shown as a red solid line, while the MW--LMC and MW--SMC distances are shown with lines defined in the inset key. The four horizontal black lines indicate the pericenter passages that the SMC must have had with the LMC in order to account for their synchronised SFH, as documented by Massana+2022. It is impossible to produce a SMC/LMC orbital period that is short, as indicated by the SFH constraints, as the binary merges within an orbital time scale due to Chandrasekhar dynamical friction. This figure was produced by Wolfgang Oehm based on the results of OehmKroupa2024.
• Figure S5: A composite radio-optical image of the M81 group. Shown in the photometric V-band is M81 (centre), M82 (left, $\approx$33 kpc from M81 in projection), NGC 3077 (lower right, $\approx$42 kpc distant from M81 in projection), and NGC 2976 (upper right, not used in the dynamical analysis here). Hydrogen gas is shown in red, with additional hydrogen gas detected by the Very Large Array depicted in green. The circles show five clouds of hydrogen gas discovered using the National Science Foundation's Robert C. Byrd Green Bank Telescope (GBT). Credit: Chynoweth et al., NRAO/AUI/NSF, Digital Sky Survey, based on Chynoweth+2008, shown here under creative Commons Attribution 4.0 Unported license.