Dark Matter-Electron Interactions Alter the Luminosity and Spectral Index of M87

Abstract

We investigate the possibility that cosmic-ray electron cooling through dark matter-electron scatterings contributes to the low radiative efficiency observed in radio-loud galaxies such as M87. Light dark matter can scatter efficiently off electrons in M87, lowering the observed bolometric luminosity compared to astrophysical expectations. This consideration allows us to probe previously unexplored regions of the parameter space of dark matter-electron interactions. We further model the cosmic-ray electron distribution by numerically solving a diffusion equation along the jet and find that efficient dark matter-electron interactions can induce a flattening of the spectral index at different distances from the central supermassive black hole, in better alignment with radio observations from M87.

Figures (5)

• Figure 1: Left panel: Cosmic ray electron timescales in M87. The red region corresponds to the synchrotron cooling timescale ($\tau_{\rm sync}$) given $B_{\text{jet}}$ in Eq. \ref{['eq:B_jet']} and the DM-electron scattering cooling timescales ($\tau_{\mathrm{\chi}e}$) are shown in blue, for different values of the DM mass, with $m_{A^{\prime}}=5$ MeV. Right panel: Variety of constraints on the dark photon kinetic mixing $\epsilon$ versus dark photon mass from Caputo_2021. The red region encloses all constraints that are based on photon to dark photon transitions with laboratory based searches (e.g. light-shining through wall experiments). In green, we show bounds arising from stellar cooling arguments, also relying on photon to dark photon transitions. In blue we show the bounds that rely on dark photons comprising the DM of the Universe. In gray, we show the direct detection bounds from devoted dark photon experiments, namely, SHUKET, WISPDMX, and Dark E-field Radio. The horizontal lines correspond to the constraints obtained in this work from M87 cooling, for various DM masses (in different colors), and fixed DM coupling $g_\chi = 0.1$. The dashed lines are derived using the magnetic field value $B_{\rm MAGIC}$, while the solid line uses $\langle B_{\rm VLBI}\rangle$ (see Appendix \ref{['sec:appendix_B']}). These bounds are evaluated at cosmic ray electron energy of $E_{e} = 10 \ \rm{GeV}$, and for an DM density $\rho_{\mathrm{DM}}(10R_s) = 10^{12} \rm{GeV/cm^3}$.
• Figure 2: Left panel: Upper limits on the non-relativistic Dirac DM-electron scattering cross section in terms of the DM mass. The blue band shows bounds from cosmic-ray electron cooling at M87 derived in this work, where the band accounts for two possible values of the magnetic field in the vicinity of the M87 BH inferred in the literature ($B=0.003-7$ G, see the Appendix \ref{['sec:appendix_B']} for more details). The DM density is taken at $10 R_S$, with value $\rho_{\rm DM}=10^{12}$GeV cm$^{-3}$; different values would rescale the bounds linearly (see the Appendix \ref{['sec:appendix_DM']} for a discussion on the DM density profile uncertainties). For comparison, we show as a vertical gray bound the lower limit on the DM mass obtained from measurements of the number of relativistic effective species ($N_{\rm eff}$) at Big Bang Nucleosynthesis Depta:2019lbeGiovanetti:2021izcKrnjaic:2019dzc. The red region corresponds to bounds from DM-electron and DM-nucleus scatterings at current direct detection experiments. For masses below $m_{\chi} \lesssim 20$ MeV, the best bound comes from DAMIC-M DAMIC-M:2025luv. For $20 \lesssim m_{\chi} \lesssim 200$ MeV, the leading bound comes from electron recoils in PANDAX4T Cheek:2025nulZhang:2025ajc. For masses above $m_{\chi} \gtrsim 200$ MeV, the leading bounds arise from nuclear recoils at CRESST-III CRESST:2019jnq. The dotted red line is obtained from the combination of projected sensitivities in various future DM direct detection experiments. In particular, we take sensitivity projections from QROCODILE QROCODILE, OSCURA aguilararevalo2022oscuraexperiment, and XLZD XLZD:2024nsu. The solid black band corresponds to the thermal relic abundance expectation in various models, see e.gKrnjaic:2025nojHochberg:2014draKuflik:2015isiGraesser:2011wiLin:2011gjSlatyer:2015jla. Right panel: Same plot as in the left panel, but for Majorana DM instead of Dirac DM. The direct detection bounds are relaxed compared to the Dirac case due to the velocity suppression of the scattering cross section $\sigma_{\rm \chi e} \propto v^2$.
• Figure 3: Left panel: Cosmic ray electron distribution as a function of energy, obtained by solving Eq. \ref{['eq:transfer_eq']}. The solid lines represent the cosmic ray-electron distribution obtained from adiabatic and synchroton cooling, in the absence of DM-electron interactions. The dotted lines corresponds to the evolution of the cosmic ray electron distribution when including DM-electron interactions. We set the product of couplings $g_\chi \epsilon =0.1$. The lines $q =0$ and $q =10$ are in agreement with Ref. Ro_2023. Depending on the configuration, DM can either leave the distribution unaffected or flatten it. This flattening occurs in the regions where DM cooling is more efficient than SM cooling. The vertical dashed lines represent the turn over energies where SM cooling becomes efficient in depleting the cosmic ray electron flux. Right panel: Spectral index $\alpha$ versus de-projected distance from the SMBH in units of $R_S$, with the legend as in the left panel. We show in solid lines the simulated spectral index obtained in absence of DM-electron interactions for $q=0$ and $q=10$, in agreement with Ro_2023. The dotted lines correspond to the modified spectral index in presence of DM-electron scatterings for different choices of parameters. For efficient DM-electron induced cooling, the spectral index flattens more than expected from conventional astrophysical mechanisms. For comparison, the shaded gray region shows the range of spectral index observations inferred from a combination of KaVA and VLBA data Niinuma:2014lga2011Natur.477..185HRo_2023, together with the $1\sigma$ error band.
• Figure 4: Possible dark matter density profiles in the inner parsecs of M87 under different astrophysical scenarios. The solid blue line corresponds to a canonical dark matter spike (see Eq. \ref{['eq:spike']}) formed under adiabatic growth of the SMBH and a pre-existing NFW profile with index $\gamma=1$ (also shown as a black solid line). The dashed blue line shows the depleted dark matter spike due to dark matter self-annihilation for a representative cross section value. The solid red line corresponds to a relaxed dark matter spike by stellar heating processes. We note that this effect is not expected to be efficient in M87 (see Appendix \ref{['sec:appendix_DM']} for details). The vertical dashed purple line indicates the edge of the M87 jet, taken at $r_{\rm jet}=2500R_S.$
• Figure 5: Magnetic field strength measurements and constraints of the M87 jet from VLBI cores at different frequencies (colored circle and diamond points) Kino:2014duaKino:2015dhaHada_2012Hada_2016Acciari_2010Kim:2018hul2014Natur.510..126ZEventHorizonTelescope:2019dseJiang_2021, Spectral Energy Distribution (SED) fits at MAGIC (dashed gray line) MAGIC:2020gbb, and EHT measurements EventHorizonTelescope:2021dvx. The black point illustrates the estimated magnetic field strength from Ro_2023, and the gray region shows the extrapolated magnetic field from that study at different radii.