Aims. In this paper, we carry out a phenomenological investigation of the viability of fuzzy dark matter, which is composed of coherent waves of non-interacting ultralight axions with a mass of ma ≈ 10−22 eV. We did so by confronting the predictions of the model, in particular, the formation of a solitonic core at the center of dark matter halos, with a homogeneous and robust sample of high-resolution rotation curves from the LITTLE THINGS in 3D catalog. This comprises a collection of isolated, dark matter-dominated dwarf-irregular galaxies that provides an optimal benchmark for cosmological studies. Our aim is to find evidence of fuzzy dark matter in the observations; alternatively, we seek to set exclusion ranges for its mass.
Methods. We used a statistical framework based on a χ2 analysis of the rotation curves of the LITTLE THINGS in 3D catalog using a fuzzy dark matter profile as the theoretical model. This allows us to extract relevant parameters such as the axion mass and mass of the solitonic core, as well as the mass of the dark matter halo and its concentration parameter. We fit the data using current Markov chain Monte Carlo techniques with a rather loose set of priors, except for the implementation of a core-halo relation predicted by simulations. The results of the fits were then used to perform various diagnostics on the predictions of the model.
Results. Fuzzy dark matter provides an excellent fit to the rotation curves of the LITTLE THINGS in 3D catalog, with axion masses determined from different galaxies clustering around ma ≈ 2 × 10−23 eV. However, we find two major problems from our analysis. First, the data follow scaling relations of the properties of the core, which are not consistent with the predictions of the soliton. This problem is particularly acute in the core radius-mass relation with a tension that (at face value) has a significance of ≳5σ. The second problem is related to the strong suppression of the linear power spectrum that is predicted by fuzzy dark matter for the axion mass preferred by the data. This can be constrained very conservatively by the galaxy counts in our sample, which leads to a tension that exceeds 5σ. We estimate the effects of baryons in our analysis and discuss whether they could alleviate the tensions of the model with observations.