Search

ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured. ASA is captured.

Supernova remnants (SNRs), as the major contributors to the galactic cosmic rays (CRs), are believed to maintain an average CR spectrum by diffusive shock acceleration regardless of the way they release CRs into the interstellar medium (ISM). However, the interaction of the CRs with nearby gas clouds crucially depends on the release mechanism. We call into question two aspects of a popular paradigm of the CR injection into the ISM, according to which they passively and isotropically diffuse in the prescribed magnetic fluctuations as test particles. First, we treat the escaping CR and the Alfvén waves excited by them on an equal footing. Second, we adopt field-aligned CR escape outside the source, where the waves become weak. An exact analytic self-similar solution for a CR "cloud" released by a dimmed accelerator strongly deviates from the test-particle result. The normalized CR partial pressure may be approximated as P(p, z, t ) = 2[∣z∣5/3 + z 5/3 dif (p, t )]-3/5 exp[-z2/4DISM(p)t ], where p is the momentum of CR particle, and z is directed along the field. The core of the cloud expands as zdif α √DNL (p) t and decays in time as P α 2z-1 dif (t ). The diffusion coefficient DNL is strongly suppressed compared to its background ISM value DISM: DNL ~ DISM exp (-Π) ⪡ DISM for sufficiently high field-line-integrated CR partial pressure, Π. When Π ⪢ 1, the CRs drive Alfvén waves efficiently enough to build a transport barrier (P ≈ 2/ ∣z∣-"pedestal") that strongly reduces the leakage. The solution has a spectral break at p = pbr, where pbr satisfies the equation DNL (pbr) ≃z²/t. [ABSTRACT FROM AUTHOR]