1 Introduction

During the past decade, a new class of relativistic jet sources has been discovered in the Galaxy: Galactic superluminal radio sources (or microquasars), exemplified by the prototypical sources GRS 1915+105 and GRO J1655-40. Radio outbursts from these transient X-ray binaries, containing accreting black holes or neutron stars, are associated with ejections of plasma at relativistic bulk velocities. The bulk Lorentz factors of these flows inferred from observations are of the order of $\Gamma_{\rm jet} \sim 3 {-} 10$, though exact observational determination of $\Gamma _{\rm jet}$ is difficult and in any case it is unlikely that a universal value of $\Gamma _{\rm jet}$ holds for all jets. For reasons of simplicity and for lack of better knowledge we will adopt a fiducial value of $\Gamma_{\rm jet} \sim 5$ for numerical examples, with the explicit understanding that Galactic jets most likely operate at a variety of Lorentz factors, which are currently not well determined. These ejections can contain energies in excess of $10^{44}~ \rm {ergs}$(Mirabel & Rodríguez 1994; Hjellming & Rupen 1995; Fender et al. 1999a), and they occur at a Galactic rate of $\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$ }}}\hbox{$>$ }}}$ $few\ {\rm yr^{-1}}$.

A large part of the kinetic energy transported by these jets is transferred into random, isotropic particle energy at the interface between the jet and the ambient medium, the working surface. Because the jets are relativistic, the particles leaving the working surface must a priori be relativistic without any need for diffusice acceleration. This mechanism of accelerating relativistic cosmic ray (CR) particles is fundamentally different from CR production in the non-relativistic shocks of supernova remnants (SNRs), which provides the bulk of the Galactic CRs.

While the momentum gain for particles crossing a non-relativistic shock is small (of order $\delta~p/p \sim v/c$), the large momentum gains encountered in relativistic shocks (of order $\delta ~p/p \sim \Gamma$, where $\Gamma$ is the shock Lorentz factor) should lead to the formation of distinct spectral features in the spectrum (see Sect. 3.3.1). Thus, unlike the CRs produced in SNRs, which follow a smooth powerlaw spectrum, the CRs produced in relativistic flows, like those encountered in microquasars, should show clearly distinguishable, and possibly narrow, spectral features.

If these particles can escape the working surface without suffering significant adiabatic energy losses, they will diffuse through interstellar space, and will thus contribute to the Galactic cosmic ray (CR) spectrum.

Figure 1: Cartoon of the proposed model of CR production in microquasars: The interface between the relativistic jet and the ISM is a natural site for the production and release of relativistic particles.

Based on the arguments presented in this paper, we conclude that an additional component of CRs generated by relativistic jets in microquasars should exist in the Galaxy. Initially, this component should consist of narrow peaks, with peak energies corresponding to $\Gamma_{\rm jet}~m_{\rm p}~ c^2$ from different jet sources.

There are many mechanisms which might broaden these features. However, any observational limits on their existence would give us additional information about the physics of microquasar jets and the physical conditions in relativistic shocks. Below, we will discuss the main mechanisms which could smooth out the component under discussion. In this paper, we content ourselves with presenting order of magnitude estimates only, since the goal of the paper is to point out to the CR community that, in addition to CR acceleration in supernova remnants (SNRs), there is another very effective mechanism to release relativistic particles in the Galaxy. Traces of these particles might be hidden in the observed CR spectra, in $\gamma$-rays with energies of a few 100 MeV, and possibly in electron-positron annihilation line emission from regions close to the location of microquasars in the Galactic plane.