Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Evidence for dark matter interactions in cosmological precision data?

Julien Lesgourgues, Gustavo Marques-Tavares, Martin Schmaltz

TL;DR

The paper proposes a two-parameter extension of ΛCDM with a self-interacting dark radiation component and a weak dark matter–dark radiation drag to damp structure growth during radiation domination. Through precision fits to Planck CMB, BAO, LSS, and $H_0$ data, the authors find a preference for a non-zero drag rate $Γ_0$ and a finite dark radiation density $ΔN_\mathrm{fluid}$, improving $\chi^2$ relative to ΛCDM by about 11–12 points. The model yields a smoother suppression of small-scale power, alleviating the $σ_8$ tension while remaining compatible with CMB observations, and allows higher $H_0$ values when combined with external priors. They discuss concrete particle-physics realizations (non-Abelian dark sectors and dark photons) that realize the required parameters and highlight potential signatures in Lyman-α and future large-scale structure surveys.

Abstract

We study a two-parameter extension of the cosmological standard model $Λ$CDM in which cold dark matter interacts with a new form of dark radiation. The two parameters correspond to the energy density in the dark radiation fluid $ΔN_\mathrm{fluid}$ and the interaction strength between dark matter and dark radiation. The interactions give rise to a very weak "dark matter drag" which damps the growth of matter density perturbations throughout radiation domination, allowing to reconcile the tension between predictions of large scale structure from the CMB and direct measurements of $σ_8$. We perform a precision fit to Planck CMB data, BAO, large scale structure, and direct measurements of the expansion rate of the universe today. Our model lowers the $χ$-squared relative to $Λ$CDM by about 12, corresponding to a preference for non-zero dark matter drag by more than $3 σ$. Particle physics models which naturally produce a dark matter drag of the required form include the recently proposed non-Abelian dark matter model in which the dark radiation corresponds to massless dark gluons.

Evidence for dark matter interactions in cosmological precision data?

TL;DR

The paper proposes a two-parameter extension of ΛCDM with a self-interacting dark radiation component and a weak dark matter–dark radiation drag to damp structure growth during radiation domination. Through precision fits to Planck CMB, BAO, LSS, and data, the authors find a preference for a non-zero drag rate and a finite dark radiation density , improving relative to ΛCDM by about 11–12 points. The model yields a smoother suppression of small-scale power, alleviating the tension while remaining compatible with CMB observations, and allows higher values when combined with external priors. They discuss concrete particle-physics realizations (non-Abelian dark sectors and dark photons) that realize the required parameters and highlight potential signatures in Lyman-α and future large-scale structure surveys.

Abstract

We study a two-parameter extension of the cosmological standard model CDM in which cold dark matter interacts with a new form of dark radiation. The two parameters correspond to the energy density in the dark radiation fluid and the interaction strength between dark matter and dark radiation. The interactions give rise to a very weak "dark matter drag" which damps the growth of matter density perturbations throughout radiation domination, allowing to reconcile the tension between predictions of large scale structure from the CMB and direct measurements of . We perform a precision fit to Planck CMB data, BAO, large scale structure, and direct measurements of the expansion rate of the universe today. Our model lowers the -squared relative to CDM by about 12, corresponding to a preference for non-zero dark matter drag by more than . Particle physics models which naturally produce a dark matter drag of the required form include the recently proposed non-Abelian dark matter model in which the dark radiation corresponds to massless dark gluons.

Paper Structure

This paper contains 10 sections, 9 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Generalizing $\Lambda$CDM with dark radiation and dark matter drag
  3. Modified cosmological perturbation equations
  4. Effects on the CMB and LSS spectrum
  5. Example models and calculation of $N_\mathrm{fluid}$ and $\Gamma_0$
  6. Data and methodology
  7. Reconciling cosmological data sets
  8. Physical interpretation of the results
  9. Conclusions
  10. Acknowledgments

Figures (6)

  • Figure 1: Ratio of the dark matter density perturbation $\delta_\mathrm{dm}$ for an interaction rate $\Gamma_0=2 \times 10^{-7}~\mathrm{Mpc}^{-1} \simeq 2 \times 10^{-21}~\mathrm{s}^{-1}$ over the same perturbation in the standard non-interacting limit, in the Newtonian gauge, as a function of conformal time, and for six representative wavenumbers. The interaction rate causes a suppression of $\delta_\mathrm{dm}$ inside the Hubble radius, efficient especially during radiation domination, and continuing during the beginning of matter domination (the vertical dashed line shows the time of equality between radiation and matter). Apart from $\Gamma_0$, the two cosmological models share the same parameters, including $\Delta N_\mathrm{fluid}=0.21$.
  • Figure 2: Residual of the temperature (left) and $E$-polarisation (right) power spectrum in several extended models compared to the minimal $\Lambda$CDM model. Two models have ordinary decoupled cold dark matter, but either free-streaming (blue) or self-interacting (green) extra relics with respectively $\Delta N_\mathrm{eff}=0.21$ or $\Delta N_\mathrm{fluid}=0.21$. The text explains which quantities have been kept fixed in these comparisons. The last model (red curves) shares the same parameters as the latter model (green curves), excepted that the DM-DR interaction is switched on, with $\Gamma_0 = 2 \times 10^{-7}~\mathrm{Mpc}^{-1} \simeq 2 \times 10^{-21}~\mathrm{s}^{-1}$. The boxes show the binned error bars of the Planck High Frequency Instrument 2015 data , which covers $\ell \geq 30$. All models are well within the error bars of the Low Frequency Instrument, which covers $\ell < 30$.
  • Figure 3: Residual of the matter power spectrum $P(k, z=0)$ in the same extended models as in the previous figure, compared to the minimal $\Lambda$CDM model (see the caption of figure \ref{['fig:CMB']} for details).
  • Figure 4: 68% and 95% CL contours for ($\sigma_8, H_0$) and ($\sigma_8, \Omega_m$): first, for the $\Lambda$CDM model and CMB+BAO data (green); next, for our model and CMB+BAO data (black), CMB+LSS data (blue), CMB+BAO+LSS data (red). This figure can be compared with Fig. 33 of Planck 2015 Planck:2015xua, to show a clear difference between our model and all the massive active/sterile neutrino models used in that figure: our model can explain a lower $\sigma_8$ without requiring at the same time a lower $H_0$ or a higher $\Omega_m$ (on the contrary, it is compatible with higher $H_0$ values).
  • Figure 5: Posterior probabilities for the eight parameters forming the basis of our model and for two derived parameters ($\Omega_m$, $\sigma_8$), for CMB data combined with BAOs (black), LSS (blue), BAO+LSS (red), BAO+LSS+$H_0$ (yellow). See the text for details on parameter definitions and priors, and for the precise content of each dataset.
  • ...and 1 more figures