First Predicted Cosmic Ray Spectra, Primary-to-Secondary Ratios, and Ionization Rates from MHD Galaxy Formation Simulations

TL;DR

This paper presents the first detailed simulations of cosmic ray spectra in galaxies, incorporating kinetic-MHD models and comparing results with observations, revealing key insights into cosmic ray transport, ionization, and source characteristics.

Contribution

It introduces novel numerical methods for simulating resolved cosmic ray spectra in galaxy formation models, including effects often neglected in previous studies.

Findings

Reproduces observed cosmic ray spectra and ratios with simple models.

Identifies the importance of a ~10 kpc CR scattering halo in galaxies.

Shows that cosmic ray ionization rates vary naturally across different galactic environments.

Abstract

We present the first simulations evolving resolved spectra of cosmic rays (CRs) from MeV-TeV energies (including electrons, positrons, (anti)protons, and heavier nuclei), in live kinetic-MHD galaxy simulations with star formation and feedback. We utilize new numerical methods including terms often neglected in historical models, comparing Milky Way analogues with phenomenological scattering coefficients $\nu$ to Solar-neighborhood (LISM) observations (spectra, B/C, $e^{+}/e^{-}$, $\bar{p}/p$, $^{10}$Be/$^{9}$Be, ionization, $\gamma$-rays). We show it is possible to reproduce observations with simple single-power-law injection and scattering coefficients (scaling with rigidity $R$), similar to previous (non-dynamical) calculations. We also find: (1) The circum-galactic medium in realistic galaxies necessarily imposes a $\sim10\,$kpc CR scattering halo, influencing the required $\nu(R)$. (2) Increasing the normalization of $\nu(R)$ re-normalizes CR secondary spectra but also changes primary spectral slopes, owing to source distribution and loss effects. (3) Diffusive/turbulent reacceleration is unimportant and generally sub-dominant to gyroresonant/streaming losses, which are sub-dominant to adiabatic/convective terms dominated by $\sim0.1-1\,$kpc turbulent/fountain motions. (4) CR spectra vary considerably across galaxies; certain features can arise from local structure rather than transport physics. (5) Systematic variation in CR ionization rates between LISM and molecular clouds (or Galactic position) arises naturally without invoking alternative sources. (6) Abundances of CNO nuclei require most CR acceleration occurs around when reverse shocks form in SNe, not in OB wind bubbles or later Sedov-Taylor stages of SNe remnants.