The source of the cosmic-ray excess in the Centaurus region -- constraints on possible candidates, mass composition and cosmic magnetic fields

TL;DR

The paper tackles the origin of the Centaurus region overdensity observed in UHECRs with $E\gtrsim 40$ EeV by testing multiple source scenarios under realistic Galactic and extragalactic magnetic-field models. It employs a Fisher-distributed EGMF deflection framework, a suite of GMF models, and scans over signal fraction $f$, charge $Z$, and $\beta_{\rm EGMF}$ to reproduce the LiMa significance, angular scale, and directional stability of the observed excess, including its energy evolution. The main findings constrain subdominant-source scenarios to $Z\lesssim6$ (with $1\lesssim\beta_{\rm EGMF}\lesssim100$) for near sources, or $Z\sim6$ with $1\lesssim\beta_{\rm EGMF}\lesssim20$ for a more distant source like Sombrero, while a single dominant source above the ankle is viable for Cen A or M83 with $\beta_{\rm EGMF}\sim20$–$30$ and a mixed composition; however, below $\sim30$ EeV an additional component is likely required. These results integrate GMF/EGMF uncertainties to bound the emitted composition and magnetic-field strengths in the local Universe, with future composition measurements and improved GMF modeling expected to further sharpen the constraints.

Abstract

The most significant excess in the arrival directions of ultra-high-energy cosmic rays with energies $\gtrsim40\,\mathrm{EeV}$ is found in the direction of several interesting source candidates, most prominently the nearby radio galaxy Centaurus A. Naturally, Cen A has been suspected to create the anisotropy - but very different scenarios have been proposed. This includes a subdominant source contribution in combination with isotropic background sources, as well as a scenario where Cen A supplies the whole cosmic-ray flux above the ankle. Recently, it was suggested that the overdensity could instead consist of strongly deflected events from the Sombrero galaxy. Thanks to the recent development of several models of the Galactic magnetic field, it is now possible to test these proposed scenarios explicitly. Leveraging the measured overdensity direction, significance, angular scale, and energy evolution, we place limits on the allowed signal fraction, the possible ejected charge number and the strength of the extragalactic magnetic field between the respective source and Earth. We find that the scenario of a subdominant source in the overdensity region requires the charge number to be $Z\lesssim6$ and the extragalactic magnetic field quantity $B/\mathrm{nG} \sqrt{L_c/\mathrm{Mpc}}$ to be between $~1$ and $~100$. For the Sombrero galaxy to be the source, the dominant charge number has to be around $Z=6$ with $1\lesssim B/\mathrm{nG} \sqrt{L_c/\mathrm{Mpc}}\lesssim20$. We find that a scenario where all the flux above $30\,\mathrm{EeV}$ is supplied by Cen A or M83 is possible for $20\lesssim B/\mathrm{nG} \sqrt{L_c/\mathrm{Mpc}}\lesssim30$ and a mixed composition - explaining both the Centaurus region excess and the distribution of the highest-energy events - however, another contributing source is possibly required in the energy range $<30\,\mathrm{EeV}$.

Figures (23)

• Figure 1: Local LiMa significance in Galactic coordinates from a scan of the arrival direction data of the Pierre Auger Observatory g_golup_on_behalf_of_the_pierre_auger_collaboration_update_2024. The direction of several proposed source candidates of the most significant overdensity in the Centaurus region are indicated by the colored star markers.
• Figure 2: Simulation examples, energy spectrum in the upper row and corresponding mass composition (represented by the mean over the natural logarithm of the mass number $A$) in the lower row. The solid lines in the spectra denote the total simulated contribution of a mass group, the dashed lines that of the signal events only. The background always follows a mixed composition while the source composition and signal fraction $f$ can vary, see figure captions. The error bars correspond to Auger data from Aab_measurement_2020 for the spectrum and Yushkov:2019J8 and Auger_Mass_DNN_2025 for the mass. The gray area denotes energies below $10^{19.6}\,\mathrm{eV}\simeq40\,\mathrm{EeV}$ corresponding to the energy threshold where the Centaurus excess becomes maximally significant.
• Figure 3: Constraints on the parameter space of signal fraction $f$ and EGMF parameter $\beta_\mathrm{EGMF}$ for Cen A as the source of the observed excess, for different charge numbers as indicated in the figures. Parameter combinations in the orange horizontally hatched area lead to too large maximum LiMa significances $\sigma_\mathrm{40\,EeV}\gg5.1\sigma$ not in agreement with the data, while in the pink horizontally hatched area the significance is too small. In the green vertically hatched area, the angular scale $\psi_\mathrm{40\,EeV}\ll27^\circ$ is too small, in the blue vertically hatched area it is too large. The red shaded region leads to a mass composition not in agreement with Auger mass composition measurements Mayotte_ICRC_2025. The color bar (background gray scale) denotes the minimum (over the 10 simulations per GMF model) of the average (over the scanned energy thresholds the_pierre_auger_collaboration_flux_2024, see also Fig. \ref{['fig:cena_AD']}) of the angular difference between the observed excess direction and the simulated ones. Only very few simulations can reproduce the excess direction well over all energy thresholds, so that $\langle\theta\rangle$ is often a lot larger than the observed value of $3.3^\circ$, especially for larger charge numbers. The white contour denotes a value of $\langle\theta\rangle\simeq10^\circ$. The black cross marker (for $Z=2$) indicates a parameter combination that is discussed more thoroughly below. For more details see text.
• Figure 4: Evolution of the excess direction with the energy for example simulations with Cen A (star marker) as the source, $f=0.05$, $\beta_\mathrm{EGMF}=15$, using and different charges (see figure title) and four different GMF models. For each GMF, four randomly drawn simulations are shown. Each simulation's excess directions are indicated by 6 connected markers of decreasing size denoting the energy thresholds $>20\,\mathrm{EeV}$, $>25\,\mathrm{EeV}$, $>32\,\mathrm{EeV}$, $>40\,\mathrm{EeV}$, $>50\,\mathrm{EeV}$, and $>63\,\mathrm{EeV}$ used in the_pierre_auger_collaboration_flux_2024. The red crosses mark the same for the data. The red circle indicates the best angular scale of $\psi_\mathrm{40\,EeV}=27^\circ$the_pierre_auger_collaboration_flux_2024 for $>40\,\mathrm{EeV}$. Note that the random seeds of the four simulations are the same for each GMF and charge number, so that the background events are the same. This is why the excess directions are similar for some GMFs, especially for $>63\,\mathrm{EeV}$ where statistics can dominate.
• Figure 5: Four example simulations with Cen A as the source, with $Z=2$, $f=0.05$, $\beta_\mathrm{EGMF}=15$ (black cross in Fig. \ref{['fig:cena_constraints']}). The upper two rows are two examples with the UF23-base-Pl GMF model, the lower two rows with the KST24-Pl model. The left two figures are for $E_\mathrm{min}=20\,\mathrm{EeV}$, the right two for $E_\mathrm{min}=40\,\mathrm{EeV}$. All skymaps are in Galactic coordinates. The background events are half-transparent. Note that for $E_\mathrm{min}=20\,\mathrm{EeV}$, the angular scale of the LiMa significance scan is not optimized and instead fixed to $27^\circ$ as in the_pierre_auger_collaboration_flux_2024.