The fast TeV variability of the blazars Mrk 501 and PKS 2155--304 implies a compact emitting region that moves with a bulk Lorentz factor of Gamma_{em}~100 toward the observer. The Lorentz factor is clearly in excess of the jet Lorentz factors Gamma_j\simless 10 measured on sub-pc scales in these sources. We propose that the TeV emission originates from compact emitting regions that move relativistically {\it within} a jet of bulk Gamma_j~10. This can be physically realized in a Poynting flux-dominated jet. We show that if a large fraction of the luminosity of the jet is prone to magnetic dissipation through reconnection, then material outflowing from the reconnection regions can efficiently power the observed TeV flares through synchrotron-self-Compton emission. The model predicts simultaneous far UV/soft X-ray flares.Comment: Moderate changes to match the published version, MNRAS, 395, L29 (2009

A conceptual model of resistive magnetic reconnection via a stochastic plasmoid chain is proposed. The global reconnection rate is shown to be independent of the Lundquist number. The distribution of fluxes in the plasmoids is shown to be an inverse square law. It is argued that there is a finite probability of emergence of abnormally large plasmoids, which can disrupt the chain (and may be responsible for observable large abrupt events in solar flares and sawtooth crashes). A criterion for the transition from magnetohydrodynamic to collisionless regime is provided.PACS numbers: 52.35. Vd, 94.30.cp, 96.60.Iv, 52.35.Py Introduction. Magnetic reconnection is the process of topological rearrangement of magnetic field, resulting in a conversion of magnetic energy into various forms of plasma energy [1]. It is believed to cause solar flares and has been studied in tokamaks [2], dedicated laboratory experiments [3] and measured in situ in the Earth's magnetosphere [4]. The basic conceptual underpinnings of the modern understanding of resistive reconnection can be summarised in three points: (i) generic X-point configurations are unstable and collapse into current layers [5,6]; (ii) the structure of resistive current layers is well described by the Sweet-Parker (SP) model [7]: if B 0 is the upstream magnetic field, V A = B 0 / √ 4πρ is the Alfvén speed (ρ the plasma density), L the length of the layer, η the magnetic diffusivity, and S ≡ V A L/η the Lundquist number, then the layer thickness is δ ∼ L/ √ S, the outflow velocity is V A , and the reconnection rate is cE ∼ V A B 0 / √ S -"slow" because it depends on S, which is very large in most natural systems; (iii) when S exceeds a critical value S c ∼ 10 4 , the SP layers are linearly unstable [8] and break up into secondary islands, or plasmoids [9]. This fact has emerged as a defining feature of numerical simulations of reconnection as they have broken through the S c barrier [6,[9][10][11][12][13][14][15][16]. It seems that high-S reconnection generically occurs via a chain of plasmoids, born, growing, coalescing, and being ejected in a stochastic fashion [17,18]. Importantly, recent numerical evidence [11,[13][14][15][16] suggests that plasmoid reconnection is "fast", i.e., independent of S.

Using two-dimensional particle-in-cell simulations, we characterize the energy spectra of particles accelerated by relativistic magnetic reconnection (without guide field) in collisionless electron-positron plasmas, for a wide range of upstream magnetizations σ and system sizes L. The particle spectra are well-represented by a power law γ −α , with a combination of exponential and super-exponential high-energy cutoffs, proportional to σ and L, respectively. For large L and σ, the power-law index α approaches about 1.2.

It is generally accepted that astrophysical sources cannot emit synchrotron radiation above 160 MeV in their rest frame. This limit is given by the balance between the accelerating electric force and the radiation reaction force acting on the electrons. The discovery of synchrotron gamma-ray flares in the Crab Nebula, well above this limit, challenges this classical picture of particle acceleration. To overcome this limit, particles must accelerate in a region of high electric field and low magnetic field. This is possible only with a non-ideal magnetohydrodynamic process, like magnetic reconnection. We present the first numerical evidence of particle acceleration beyond the synchrotron burnoff limit, using a set of 2D particle-in-cell simulations of ultra-relativistic pair plasma reconnection. We use a new code, Zeltron, that includes self-consistently the radiation reaction force in the equation of motion of the particles. We demonstrate that the most energetic particles move back and forth across the reconnection layer, following relativistic Speiser orbits. These particles then radiate > 160 MeV synchrotron radiation rapidly, within a fraction of a full gyration, after they exit the layer. Our analysis shows that the high-energy synchrotron flux is highly variable in time because of the strong anisotropy and inhomogeneity of the energetic particles. We discover a robust positive correlation between the flux and the cut-off energy of the emitted radiation, mimicking the effect of relativistic Doppler amplification. A strong guide field quenches the emission of > 160 MeV synchrotron radiation. Our results are consistent with the observed properties of the Crab flares, supporting the reconnection scenario.