betaFeedback

WorksPhysics and Astronomy

Anisotropic momentum distributions due to radiation recoil in relativistic plasmas with electric and magnetic fields

Abstract

The interaction of strong electromagnetic fields with plasma generates radiation accompanied by a recoil force, which can significantly alter the plasma dynamics. In this work, we investigate the development of anisotropic momentum distributions induced by the combined action of electric and magnetic fields on a thermally relativistic plasma. We consider three distinct types of anisotropy. The first arises from a pure magnetic field acting on plasma with either isotropic or anisotropic initial momentum distributions, producing the characteristic ring-shaped momentum profile. The second is driven by a pure electric field, where radiation reaction generates a partial, anisotropic ring distribution in momentum space: significant modifications occur primarily in the 90∘^\circ–180∘^\circ and 270∘^\circ–360∘^\circ sectors of the pxp_x–pyp_y plane, while the remaining quadrants remain largely unaffected. The third case considers the combined effect of electric and magnetic fields. When the cyclotron frequency is very close to the upper hybrid frequency, the azimuthal symmetry of the ring-momentum distribution is broken. Conversely, in the regime where the cyclotron frequency is lower than the upper hybrid frequency, the rapid oscillations of the electric field dominate and preserve the symmetry of the ring-momentum distribution.

Introduction

The study of the interplay between quantum electrodynamics (QED) and collective plasma dynamics has gained considerable momentum over the past decade [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. This growing interest is largely driven by rapid advances in high-intensity laser technology [13, 14, 15, 16, 17, 18], which now enable experiments in regimes where both QED and plasma effects play a central role. In astrophysical settings, extreme environments near compact objects such as magnetars naturally give rise to the same physics, and similar conditions are expected to be reproduced in forthcoming next-generation strong-field laboratory experiments [7].

The interaction of electromagnetic fields with plasma drives particles into relativistic motion. Accelerating charged particles lead to photon emission, and depending on the photon energy, this radiation can substantially alter the dynamics of the emitting particle. The associated recoil force, known as radiation reaction, has been extensively studied in both classical and quantum frameworks. In the classical regime, radiation reaction is often described by the Abraham–Lorentz–Dirac (ALD) equation, though it is well known to exhibit unphysical runaway solutions [19]. When radiation reaction acts only as a small perturbation, this difficulty can be circumvented by adopting the Landau–Lifshitz (LL) equation [20]. Exact solutions of the LL equation have been obtained for single-particle motion in specific field configurations [21, 22], while in plasma physics, the LL equation has been widely implemented in Particle-In-Cell (PIC) simulations, both in its classical form [23] and in quantum-corrected generalizations [24, 25].

The incorporation of radiation reaction into kinetic theory was first established by Hakim et al. in the framework of many-particle systems [26]. Subsequent studies developed alternative derivations of radiation reaction–corrected Vlasov equations [27, 28, 29], with the primary focus on the formal construction of kinetic evolution equations rather than their plasma dynamical consequences. A kinetic analysis of Landau damping modified by radiation reaction was presented in Ref. [30]. In parallel, hydrodynamic models of relativistic plasmas that incorporate radiation reaction have also been proposed; see, for instance, Refs. [31, 32, 33, 34]. More recently, self-consistent treatments of radiation reaction in kinetic plasmas have been investigated: the electrostatic limit was examined in Ref. [35], while the case of a circularly polarized electric field was studied in Ref. [36]. For a plasma subjected to a strong external magnetic field—conditions typical of astrophysical plasmas—electrons can lose a significant portion of their kinetic energy. Previous studies have shown that radiation reaction in a constant and homogeneous magnetic field can drive a plasma initially in equilibrium to develop anisotropy, leading to the formation of a ring-shaped momentum distribution [4, 5]. However, the emergence of such ring-shaped anisotropy has not yet been explored in the case of radiation driven by an electric field, nor in situations where both electric and magnetic fields act simultaneously on the plasma.

In this work, we address this gap by investigating radiation reaction effects in plasmas subject to self-consistent electric fields in combination with external magnetic fields, with the aim of understanding the resulting anisotropy formation and momentum-space dynamics. We solve the relativistic Vlasov equation coupled with Ampère’s law, allowing the plasma current to self-consistently modify the electric field. Our objective is to study how radiation reaction influences the momentum distribution for both initially isotropic and anisotropic plasmas. The radiation reaction force is modeled using the Landau–Lifshitz (LL) equation. The magnetic field is treated as constant and homogeneous, generated by an external source; thus, only the plasma contribution to the electric field is considered, while its contribution to the magnetic field is neglected. This treatment of the magnetic field is consistent with previous works [4, 5], but our formalism additionally incorporates the effect of the self-consistent electric field as well as the magnetic force term v×B\mathbf{v} \times\mathbf{B}, where v\mathbf{v} is the velocity of the electrons in the plasma and B\mathbf{B} is the magnetic field. In the case of pure synchrotron radiation with an initially isotropic plasma, it is often convenient to neglect the magnetic force term, as it does not affect the plasma dynamics. However, for an initially anisotropic plasma subject to a strong magnetic field, or in the combined presence of electric and magnetic fields, the magnetic force term must be retained.

The radiation reaction force is a non-conservative force, i.e., it does not conserve phase-space volume, unlike the Lorentz force. Depending on the geometry of the electromagnetic field, the development of anisotropy in the plasma momentum distribution can take different forms. By considering the combined effects of electric and magnetic fields in plasma, we are able to study distinct types of anisotropic structures. In this work, we investigate three different cases. First, we study pure synchrotron radiation for both isotropic and anisotropic initial plasmas. We find that an initial anisotropy in the plasma momentum distribution reduces the overall energy loss due to radiation, leading to a smaller ring-momentum distribution compared to the isotropic case. In the second case, we examine how radiation induced by an electric field gives rise to a new form of anisotropy, distinct from that of synchrotron radiation. This anisotropy exhibits similarities to the magnetic-field case in certain regions of momentum space, while in other regions the initial momentum distribution remains largely unaffected. Interestingly, the characteristic size of the anisotropic structure in momentum space is found to be comparable in both cases, provided that the initial electric and magnetic fields have the same energy. In the third case, we examine the combined influence of electric and magnetic fields on radiation reaction and its impact on the plasma momentum distribution. When the cyclotron frequency is close to resonance with the upper hybrid frequency, the ring-momentum structure associated with pure synchrotron radiation can be strongly modified. In contrast, if the cyclotron frequency is much smaller than the upper hybrid frequency, the modification of the ring-momentum structure remains weak.

The organization of this paper is as follows. In Section II, we present the kinetic equation including the radiation reaction force considered in this work. In Section III, we present numerical simulation results for the three different cases of anisotropy development. Finally, in Section IV, we discuss the main findings of the study and provide concluding remarks.

Kinetic theory including a radiation reaction force

For sufficiently strong electromagnetic fields, the relativistic Vlasov equation describing the dynamics of an electron ensemble must be modified to incorporate quantum effects (see, e.g., Refs. [16, 17, 18, 37, 38]). When the field strength is well below the Schwinger critical value, electron–positron pair production via the Schwinger mechanism can be neglected. Nevertheless, photon emission by individual electrons through nonlinear Compton scattering may become significant whenever the parameter χA02\chi A_0^2 is not too small. Here, we define the quantum parameter

χ=γEcr∣E+vc×B∣2−(vc⋅E)2(1)\chi= \frac{\gamma}{E_{\mathrm{cr}}} \sqrt{\left|\mathbf{E} + \frac{\mathbf{v}}{c} \times\mathbf{B}\right|^2 - \left(\frac{\mathbf{v}}{c} \cdot\mathbf{E}\right)^2} \tag*{(1)}

Here, γ\gamma denotes the Lorentz factor, and Ecr=m2c3/(∣e∣ℏ)E_{\mathrm{cr}} = m^2 c^3/(|e|\hbar) is the critical field, where ee and mm are the electron charge and mass, respectively, ℏ\hbar is the reduced Planck constant, and cc is the speed of light in vacuum. We also define the dimensionless laser strength parameter as A0=eE/(mω)A_0 = eE/(m\omega), with ω\omega denoting the wave frequency. Nonlinear Compton scattering has been extensively investigated in the low-density plasma regime, where the electron density is sufficiently small such that the electromagnetic fields can be treated as vacuum solutions of Maxwell’s equations. In this limit, attention has largely been directed toward analyzing the characteristics of the emitted radiation spectra (see the recent reviews [16, 17, 18] for comprehensive references). At higher electron densities, however, plasma dynamics in strong fields becomes considerably more intricate, as self-consistent plasma currents must be included. Furthermore, if the radiated energy is significant, the electron dynamics—ordinarily governed by the Lorentz force—must be corrected to account for the radiation reaction force [23, 24, 25].

In this work, we model the plasma dynamics using the relativistic Vlasov equation, modified by an additional term representing radiation reaction. Before presenting the kinetic equation, we define the normalized notation

tn=ωcet, pn=pmc, ϵn=ϵmc2t_n = \omega_{ce}t,\ p_n = \frac{p}{mc},\ \epsilon_n = \frac{\epsilon}{mc^2}
fn=m3c3n0f, En=eEmcωce, Bn=eBmωce, n0n=c3ωce3n0(2)f_n = \frac{m^3c^3}{n_0}f,\ E_n = \frac{eE}{mc\omega_{ce}},\ B_n = \frac{eB}{m\omega_{ce}},\ n_{0n} = \frac{c^3}{\omega_{ce}^3}n_0 \tag*{(2)}

where ωce\omega_{ce} is the Compton frequency, n0n_0 is the plasma density, ϵ=1+p2\epsilon= \sqrt{1+p^2} and ff is the distribution of the plasma in the momentum space. Note that we will drop the subscript nn denoting the normalized quantities in what follows to simplify the notation. In this work, the radiation reaction force FR\mathbf{F}_{R} is added to the Lorentz Force FL=E+pϵ×B\mathbf{F}_{L} = \mathbf{E} + \frac{\mathbf{p}}{\epsilon} \times\mathbf{B} in the relativistic Vlasov equation as

[∂∂t+pϵ⋅∇r]f+∇p⋅[(FL+FR)f]=0(3)\left[\frac{\partial}{\partial t} + \frac{\mathbf{p}}{\epsilon} \cdot\nabla_r\right]f + \nabla_p \cdot\left[(\mathbf{F}_L + \mathbf{F}_R)f\right] = 0 \tag*{(3)}

where the radiation reaction force is given by

FR=2α3[ϵdEdt+(p⋅Eϵ)E+E×B+B×(B×pϵ)−ϵp((E+pϵ×B)2−(pϵ⋅E)2)](4)\mathbf{F}_{R}=\frac{2\alpha}{3}\left[\epsilon\frac{d\mathbf{E}}{dt}+\left(\frac{\mathbf{p}\cdot\mathbf{E}}{\epsilon}\right)\mathbf{E}+\mathbf{E}\times\mathbf{B}+\mathbf{B}\times\left(\mathbf{B}\times\frac{\mathbf{p}}{\epsilon}\right)-\epsilon\mathbf{p}\left(\left(\mathbf{E}+\frac{\mathbf{p}}{\epsilon}\times\mathbf{B}\right)^2-\left(\frac{\mathbf{p}}{\epsilon}\cdot\mathbf{E}\right)^2\right)\right] \tag*{(4)}

where α\alpha is the fine-structure constant. From this point onward, we focus on a field configuration where, in a suitable reference frame, the electromagnetic field is purely electric in nature. In addition, we assume a constant and homogeneous magnetic field. Such a configuration is representative of fields present in certain astrophysical environments, where the timescale governing the magnetic field variation is much longer than that of the electric field. The latter can be associated with coherent photon emission from plasma surrounding dense objects, such as magnetars [39]. Without loss of generality, we set B=B0ez\mathbf{B}=B_0\mathbf{e}_z. The contribution of the plasma current to the magnetic field is neglected, and the magnetic field is assumed to be generated by an external source. In this case, Faraday’s law can be disregarded, and the system is closed using Ampère’s law with a vanishing time-dependent magnetic field; that is, we use

∂tE=−4παn0∫d3ppϵf(5)\partial_t\mathbf{E}=-4\pi\alpha n_0\int d^3p\frac{\mathbf{p}}{\epsilon}f \tag*{(5)}

We consider electric field that is perpendicular to the magnetic field, i.e. E=Exex+Eyey\mathbf{E}=E_x\mathbf{e}_x+E_y\mathbf{e}_y. With these assumptions, the kinetic equation in Eq. (3) is reduced to

∂f∂t+Ex∂f∂px+Ey∂f∂py+B0ϵ[py∂f∂px−px∂f∂py]=−FR⋅∇pf−f∇p⋅FR(6)\frac{\partial f}{\partial t}+E_x\frac{\partial f}{\partial p_x}+E_y\frac{\partial f}{\partial p_y}+\frac{B_0}{\epsilon}\left[p_y\frac{\partial f}{\partial p_x}-p_x\frac{\partial f}{\partial p_y}\right]=-\mathbf{F}_R\cdot\nabla_p f-f\nabla_p\cdot\mathbf{F}_R \tag*{(6)}

where the radiation reaction force is

FR,x=2α3[ϵdExdt+Exϵ(pxEx+pyEy)+B0Ey−B02pxϵ−ϵpx((Ex−B0pyϵ)2+(Ey+B0pxϵ)2−1ϵ2(pxEx+pyEy)2)](7)F_{R,x}=\frac{2\alpha}{3}\left[\epsilon\frac{dE_x}{dt}+\frac{E_x}{\epsilon}(p_xE_x+p_yE_y)+B_0E_y-B_0^2\frac{p_x}{\epsilon}-\epsilon p_x\left(\left(E_x-B_0\frac{p_y}{\epsilon}\right)^2+\left(E_y+B_0\frac{p_x}{\epsilon}\right)^2-\frac{1}{\epsilon^2}(p_xE_x+p_yE_y)^2\right)\right] \tag*{(7)}
FR,y=2α3[ϵdEydt+Eyϵ(pxEx+pyEy)−B0Ex−B02pyϵ−ϵpy((Ex−B0pyϵ)2+(Ey+B0pxϵ)2−1ϵ2(pxEx+pyEy)2)](8)F_{R,y}=\frac{2\alpha}{3}\left[\epsilon\frac{dE_y}{dt}+\frac{E_y}{\epsilon}(p_xE_x+p_yE_y)-B_0E_x-B_0^2\frac{p_y}{\epsilon}-\epsilon p_y\left(\left(E_x-B_0\frac{p_y}{\epsilon}\right)^2+\left(E_y+B_0\frac{p_x}{\epsilon}\right)^2-\frac{1}{\epsilon^2}(p_xE_x+p_yE_y)^2\right)\right] \tag*{(8)}

Before studying the case of electric and magnetic field in plasma, we will look for an analytical solution of Eq. (6) in the limit of vanishing EE-field, in which the kinetic equation including radiation reaction Eq. (6) reduces to

∂f∂t=−FR⋅∇pf−f∇p⋅FR(9)\frac{\partial f}{\partial t}=-\mathbf{F}_R\cdot\nabla_p f-f\nabla_p\cdot\mathbf{F}_R \tag*{(9)}

where the radiation reaction term is FR=−2αB02p(1+p⊥2)/3ϵ\mathbf{F}_R=-2\alpha B_0^2\mathbf{p}(1+p_\perp^2)/3\epsilon, here p⊥p_\perp is the perpendicular momentum. Expressing ∇p\nabla_p in cylindrical momentum coordinates, and assuming that the initial plasma distribution is isotropic, i.e. the magnetic force term vanishes, the kinetic equation reduces to

32αB02∂f∂t=ηp⊥2ϵf+p⊥2ϵ∂f∂p⊥+p⊥2pzϵ∂f∂pz(10)\frac{3}{2\alpha B_0^2}\frac{\partial f}{\partial t}=\eta\frac{p_\perp^2}{\epsilon}f+\frac{p_\perp^2}{\epsilon}\frac{\partial f}{\partial p_\perp}+\frac{p_\perp^2p_z}{\epsilon}\frac{\partial f}{\partial p_z} \tag*{(10)}

where η=5−(p⊥2+pz2)/ϵ2\eta=5-(p_\perp^2+p_z^2)/\epsilon^2. In the ultra-relativistic limit, the factor η=4\eta=4. As shown in [4, 5], using the method of characteristic, the analytical solution to Eq. (10) is for any arbitrary initial distribution f0f_0 is

f(p⊥,pz,t)=f0(p⊥1−τp⊥,pz1−τp⊥)(1−τp⊥)4(11)f(p_\perp,p_z,t)=\frac{f_0\left(\frac{p_\perp}{1-\tau p_\perp},\frac{p_z}{1-\tau p_\perp}\right)}{(1-\tau p_\perp)^4} \tag*{(11)}

where τ=23αB02t\tau=\frac{2}{3}\alpha B_0^2t. This equation fully determines the time dependence of the distribution function undergoing synchrotron radiation. The solution is characterized by an upper bound p~⊥\tilde{p}_\perp, defined as p~⊥=τ−1\tilde{p}_\perp=\tau^{-1}, such that p⊥τ<1p_\perp\tau<1. In other words, the distribution function is progressively compressed, with the upper bound p~⊥\tilde{p}_\perp decreasing over time. As studied in [4, 5], an initially isotropic relativistic thermal plasma tends to develop anisotropy due to radiation reaction. The time evolution of the plasma distribution, presented in Eq. (11), exhibits an inverted Landau population; that is, there exist regions in the perpendicular momentum p⊥p_\perp where ∂f/∂p⊥>0\partial f/\partial p_\perp>0. This behavior leads to the formation of an anisotropy in momentum space, characterized by a ring-shaped momentum distribution. We define the radius of the ring, pRp_R, by the condition

∂f(p⊥,t)∂p⊥∣p⊥=pR=0,\left.\frac{\partial f(p_\perp,t)}{\partial p_\perp}\right|_{p_\perp=p_R}=0,

and, using Eq. (11), we obtain

pR=1+6pth2τ2−1+12pth2τ26pth2τ3(12)p_R=\frac{1+6p_{\rm th}^2\tau^2-\sqrt{1+12p_{\rm th}^2\tau^2}}{6p_{\rm th}^2\tau^3} \tag*{(12)}

This equation describes the time-dependence of the ring momentum radius developed due to pure synchrotron radiation. The ring radius is initially zero and increases over time as a result of the cooling mechanism. It reaches a maximum at the time determined by the condition ∂pR(t)/∂t∣t=tR=0\left.\partial p_R(t)/\partial t\right|_{t=t_R}=0, where

tR=34αB02pth(13)t_R=\frac{3}{4\alpha B_0^2p_{\rm th}} \tag*{(13)}

denotes the characteristic time required to attain the maximum ring radius in the pure synchrotron radiation. However, tRt_R can still be used to describe the characteristic timescale of the anisotropic development of the electric field, in which case the amplitude of B0B_0 is simply replaced by E0E_0. Later in this paper, we will employ tRt_R in the combined case of electric and magnetic fields, where only the magnetic field amplitude is considered. Beyond the maximum value of the ring radius, continued radiation causes the ring radius to decrease. Later in this work, we extend the conclusions from Eq. (11) to investigate the physics of the inverted Landau population in three distinct cases. The first case considers a constant external magnetic field with the plasma initialized in both isotropic and anisotropic configurations. The second case examines a relativistic thermal plasma subject to a time-varying electric field in the absence of a magnetic field. Finally, the third case addresses the situation where both electric and magnetic fields are present simultaneously.

Numerical simulations

Preliminaries

The kinetic system consisting of the Vlasov equation in Eq. (6) and Ampère’s law in Eq. (5) is solved numerically in two-dimensional momentum space with coordinates pxp_x and pyp_y. We assume that the parallel momentum component pzp_z is zero in order to simplify the calculation, without losing any relevant physics, since the system does not exhibit parallel acceleration. Spatial dependence is neglected as well, because there is no time-varying magnetic field—only a constant magnetic field B=B0ez\mathbf{B}=B_0\mathbf{e}_z and a time-varying electric field. The time evolution of the distribution function f(t,p)f(t,\mathbf{p}), and hence of the self-consistent electric field (i.e., the system Eqs. (5) and (6)), is computed in two steps. First, the Vlasov advection due to the Lorentz force is solved using the method of characteristics. The distribution function is interpolated from its previous value at each momentum grid point (px,py)(p_x,p_y) according to

f∗(px,py,t+Δt)=f(px−eFxΔt,py−eFyΔt,t),f^*(p_x,p_y,t+\Delta t)=f(p_x-eF_x\Delta t,p_y-eF_y\Delta t,t),

where FxF_x and FyF_y denote the Lorentz force components in the xx- and yy-directions. This procedure ensures phase-space density conservation along particle trajectories. In the second step, radiation reaction is included through the conservation equation

∂f∂t+∇p⋅(FRf)=0,\frac{\partial f}{\partial t}+\nabla_{\mathbf{p}}\cdot(\mathbf{F}_R f)=0,

where the interpolated distribution f∗f^* is used as the initial condition. This equation is solved numerically using a fourth-order Runge–Kutta (RK4) method, with flux divergences evaluated by finite differences on the momentum grid. Finally, a stabilization step is applied to remove small negative values of ff (arising from interpolation) and to truncate the high-momentum tail in order to prevent numerical instabilities. The resulting distribution ff therefore incorporates both Vlasov advection and radiation reaction effects at each timestep.

Through the paper, we consider a plasma that is initially in thermal equilibrium and follows a Maxwell–Boltzmann distribution.

f0=cne−px2/pth,x2e−py2/pth,y2(14)f_0=c_n e^{-p_x^2/p_{\mathrm{th},x}^2}e^{-p_y^2/p_{\mathrm{th},y}^2} \tag*{(14)}

where pth,xp_{\mathrm{th},x} and pth,yp_{\mathrm{th},y} denote the thermal spread in the xx- and yy-directions, respectively, and cc is the normalization constant, given by cn=1/∫dpxdpye−px2/pth,x2e−py2/pth,y2c_n=1/\int dp_xdp_y e^{-p_x^2/p_{\mathrm{th},x}^2}e^{-p_y^2/p_{\mathrm{th},y}^2}.

Pure synchrotron radiation

We consider here the case of pure synchrotron radiation; that is, we assume the absence of any electric field and the presence of only a constant magnetic field. Our aim is twofold: first, to benchmark our results against the analytical solution in Eq. (11) and the numerical results reported in [4, 5]; and second, to investigate how an initially anisotropic plasma evolves under radiation and whether a similar inverted Landau population emerges. The radiated power scales as P∝γ4(p×a)2P\propto\gamma^4(p\times a)^2, where aa is the acceleration of the electron. Since the system contains no electric field, the plasma must be relativistically hot, i.e., pth≫1p_{\mathrm{th}}\gg1, in order to produce significant radiation. The relevant quantum parameter in this configuration is given by χ=B0p⊥\chi=B_0p_\perp.

Considering an isotropic plasma with pth,x=pth,y=20p_{\mathrm{th},x}=p_{\mathrm{th},y}=20 and B0=0.05B_0=0.05, the quantum parameter is χ≲1\chi\lesssim1 for most particles in the plasma. Although the Landau–Lifshitz (LL) equation is derived under the classical assumption χ≪1\chi\ll1, it has been shown in [4, 5] that, for synchrotron radiation, the LL model agrees well with the full quantum radiation model even at χ∼1\chi\sim1 as the correction due to quantum radiation is small compared to the classical one. While some discrepancies start to arise when entering the fully quantum regime, the energy loss due to radiation reduces the thermal momentum of the particles reducing the quantum parameter χ\chi and the radiation can be described by the classical model. Furthermore, as demonstrated in [36], the agreement between Landau-Lifshitz and quantum radiation improves for higher initial plasma temperatures, that is for relativistic hot plasma such as the ones considered in this work, the Landau-Lifshitz model should be valid at χ∼1\chi\sim1.

In the first row of Figure 1, we present the momentum distribution of the isotropic plasma. The left panel shows the initial distribution, while the right panel corresponds to the distribution at t=tRt=t_R, when the radius of the ring momentum reaches its maximum. Comparing the two panels, we observe the formation of a ring structure in the right panel. The radius of the ring is approximately 13.3, which agrees well with the corresponding value in the analytical result in Eq. (12). Radiative cooling is clearly visible in the right panel, as the distribution becomes concentrated at lower momenta with a higher population density (as indicated by the colorbar maximum), compared to the initial distribution. The inverted plasma distribution arises because electrons with higher perpendicular momentum lose energy more rapidly, as the radiation reaction force is stronger at higher p⊥p_\perp. As a result, these high-p⊥p_\perp electrons decelerate and begin to bunch with lower-p⊥p_\perp electrons, leading to an accumulation of particles at lower perpendicular momenta. As the radiation process continues, successive layers of higher-energy electrons slow down and bunch in the same region, overlapping with the earlier bunched particles. This progressive bunching into a narrower zone causes the peak of the distribution to shift toward higher p⊥p_\perp, not because the electrons are gaining energy, but because the location of maximum particle density moves outward in momentum space due to the collective convergence. Throughout this process, the total energy of the particles continues to decrease due to radiation losses.

Momentum distribution of the plasma

Figure 1. Momentum distribution of the plasma. In the upper row, we used pth,x=pth,y=20p_{th,x}=p_{th,y}=20, where in the left figure we display the plasma at t=0t=0 and in the right show the distribution at t=tRt=t_R. In the lower row, we used pth,x=20p_{th,x}=20 and pth,y=10p_{th,y}=10. For the left figure, we set the radiation reaction force FR=0\mathbf{F}_R=0 to display the magnetic force effect, while in the right figure, we included the magnetic and radiation force. For all figures, we used B0=0.05B_0=0.05.

In the lower row of Figure 1, we show the momentum distribution of a non-isotropic plasma with thermal momenta pth,x=20p_{th,x}=20 and pth,y=10p_{th,y}=10. Both the left and right panels correspond to the distribution at t=tRt=t_R. In the left panel, the radiation force FR\mathbf{F}_R is set to zero in order to only display the effect of the magnetic force arising from the anisotropic thermal distribution. As seen, the plasma particles undergo gyromotion, resulting in a spiral-like structure in momentum space. The non-zero v×B\mathbf{v}\times\mathbf{B} force causes particles to drift from the pxp_x to the pyp_y direction. In the right panel, we present the momentum distribution including the effects of radiative cooling. The spiral structure persists, but the distribution becomes more concentrated at lower momenta. In particular, the spirals at high momenta become narrower due to the decreased particle occupation at high momentum caused by radiation losses. At lower momenta, we also observe the formation of a ring-like structure, similar to the one seen in the isotropic case but with a lower momentum radius pR=8.8p_R=8.8. This value is lower than for the case of isotropic plasma with pth=20p_{th}=20 but higher than the momentum radius of isotropic plasma pth=10p_{th}=10. One of the reasons why the ring radius is smaller in the non-isotropic case is the reduced energy loss due to radiation, compared to the isotropic case. Another contributing factor is the magnetic force, which induces momentum flux from the pxp_x to the pyp_y direction. This redistribution affects the electron bunching structure by modifying how particles accumulate in momentum space under the influence of radiation.

To further investigate how anisotropy affects the plasma distribution under radiation, we display in Figure 2 the dependence of different plasma distributions on the perpendicular momentum. In the left column of Figure 2, we show an initially isotropic plasma: both the initial distribution (solid curve) and the one formed at t=tRt=t_R (jagged curve). As expected, particles with initially high energies are cooled to lower energies due to radiation. However, the peak of the distribution does not shift to the lowest energies, as is typical for radiative cooling in plasmas. Instead, due to the development of a population inversion, higher energy states become more populated than the lowest ones. In the right column of Figure 2, we present the initial and evolved plasma distributions for three anisotropic cases. In the upper panel, we set pth,x=20p_{\mathrm{th},x}=20, pth,y=15p_{\mathrm{th},y}=15, in the middle panel pth,x=20p_{\mathrm{th},x}=20, pth,y=10p_{\mathrm{th},y}=10 and in the lower panel, pth,x=20p_{\mathrm{th},x}=20, pth,y=5p_{\mathrm{th},y}=5. Due to the temperature imbalance between the xx- and yy-directions, the magnetic force induces a drift of particles in momentum space from the higher-temperature region to the lower-temperature one. This redistribution reduces the overall energy loss due to radiation, as evidenced by the higher particle occupancy at larger momentum values. Comparing the high-momentum particle densities between the left and right columns, we observe a greater population at high momentum in the anisotropic cases. Furthermore, the distributions are less symmetric due to uneven energy loss, with certain regions of momentum space experiencing greater radiative damping than others, as a result of the gyromotion. These effects reduce the radius of the ring-shaped momentum distribution in the anisotropic case compared to the isotropic one.

Perpendicular momentum distribution of the plasma

Figure 2. Perpendicular momentum distribution of the plasma, initial plasma (solid curve) and the distribution formed at the peak of the ring momentum formation (jagged curve). In the left column, we have isotropic initial distribution, while in the right column we have anisotropic distribution.

To understand how the drift of particles from high to low temperature affect the radiation, we study the average values of the change of pxp_x and pyp_y using

⟨∣px∣⟩=∫dpxdpy∣px∣f(t)(15)\langle|p_x| \rangle= \int dp_x dp_y |p_x| f(t) \tag*{(15)}
⟨∣py∣⟩=∫dpxdpy∣px∣f(t)(16)\langle|p_y| \rangle= \int dp_x dp_y |p_x| f(t) \tag*{(16)}

For the perpendicular momentum, we use the definition ⟨p⊥⟩=⟨px2⟩+⟨py2⟩\langle p_\perp\rangle=\sqrt{\langle p_x^2\rangle+\langle p_y^2\rangle}. In Figure 3, we plot all three quantities using two different initial distributions. In the upper panel, we use an isotropic distribution, pth,x=pth,y=20p_{\mathrm{th},x}=p_{\mathrm{th},y}=20. As observed, the average values of pxp_x and pyp_y, and consequently the average perpendicular momentum p⊥p_\perp, decrease over time. This dynamical behavior is a direct consequence of radiation reaction, which continuously reduces the particle momenta in both the xx and yy directions. As a result, the perpendicular momentum exhibits an almost constant rate of reduction. In the lower panel of Figure 3, we consider an anisotropic initial plasma distribution with pth,x=20p_{\mathrm{th},x}=20 and pth,y=10p_{\mathrm{th},y}=10. Due to the unequal thermal spread in the xx and yy directions, the particle flow from xx to yy increases the average value of

Average momentum components and perpendicular momentum as functions of time

Figure 3. The average values of the momentum components pxp_x (solid curve), pyp_y (dashed curve), and the perpendicular momentum p⊥p_\perp (curve with circles) are plotted as functions of time for two cases. The upper panel shows the isotropic case with pth=20p_{\mathrm{th}}=20, while the lower panel presents the anisotropic case with pth,x=20p_{\mathrm{th},x}=20 and pth,y=10p_{\mathrm{th},y}=10.

pxp_x to approximately 1.5 times its initial value. This redistribution reduces the rate of energy loss, as evidenced by the evolution of the average perpendicular momentum, which exhibits a smaller slope compared to the isotropic case.

Anisotropic formation due to pure electric field

In this subsection, we consider only an electric field propagating in plasma and assume that the magnetic field is zero. We use the same assumption of the Maxwell–Boltzmann distribution Eq. (14) as in the previous subsection. We let the plasma interact self-consistently with an electric field that initially has peak values in the xx- and yy-directions. As the electric field induces a current in the xx- and yy-directions, the contribution from the current updates the values of the electric field through Ampère’s law Eq. (5).

The radiation reaction force in the vanishing magnetic field-case is

FR,i=2α3[dEidt+Eiϵ(pxEx+pyEy)−ϵpi(E2−1ϵ2(pxEx+pyEy)2)](17)F_{R,i}=\frac{2\alpha}{3}\left[\frac{dE_i}{dt}+\frac{E_i}{\epsilon}\left(p_xE_x+p_yE_y\right)-\epsilon p_i\left(E^2-\frac{1}{\epsilon^2}\left(p_xE_x+p_yE_y\right)^2\right)\right] \tag*{(17)}

where i=x,yi=x,y. Starting with an electric field, it is possible to accelerate the plasma to ultra-relativistic energies and obtain significant radiation without starting with a relativistically hot plasma. However, to investigate whether a similar ring-formation physics exists in the case of an electric field in plasma, we have to start from a relativistically hot plasma. Furthermore, the Lorentz factor due to acceleration by the electric field must be much smaller than the thermal Lorentz factor, to ensure that the same condition applied to synchrotron radiation is now applied to radiation due to the electric field.

We perform a numerical simulation with E0=0.05E_0=0.05, the same magnitude as in the synchrotron radiation case, and start with an isotropic relativistically hot plasma with pth=20p_{\mathrm{th}}=20. To restrict the Lorentz factor due to the electric field, we had to run the simulation with a dense plasma. Starting with the plasma densities n0=7×1028 cm−3n_0=7\times10^{28}\ \mathrm{cm}^{-3} we obtained a Lorentz factor (not accounting for the thermal Lorentz factor) of the order of unity. In Figure 4, we show the results from the simulation of the plasma distribution after radiation due to the electric field, together with the results from the synchrotron radiation simulation. In the upper row of Figure 4, the plots correspond to the synchrotron simulation, while the lower row represents the plots from the electric field simulation. In the first column, we plot the magnitude of the radiation force multiplied by the initial plasma distribution, ∣FR∣f0|F_R|f_0. For the synchrotron radiation, it is clear that the radiation is azimuthally symmetric as we know that FR=−2αB02pϵ/3F_R=-2\alpha B_0^2p\epsilon/3, with a peak at higher momenta of the plasma distribution, since the thermal Lorentz factor is larger there. In the lower row of the first column, the radiation force arising from the electric field is displayed. The radiation force is not azimuthally symmetric and exhibits pronounced peaks at large momenta, p=px2+py2p=\sqrt{p_x^2+p_y^2}. This behavior arises because the radiation force is dominated by the third term of Eq. (17), which is higher order in pp compared to the other two terms. We can rewrite the third term of Eq. (17) as

Evolution of the plasma distribution after radiation

Figure 4. Demonstration of the evolution of the plasma distribution after radiation. In the first row, the figures are for the case of only magnetic field, while the lower panel is for only electric field. In the first column we display the magnitude of the radiation force times the initial plasma distribution ∣FR∣f0|F_R|f_0. In the second column we plot the plasma distribution at t=0.4tRt=0.4t_R and in the third column we display the plasma distribution at t=4tRt=4t_R

F=−2αϵp3[E2−(p⋅E)2ϵ2](18)\mathbf{F}=-\frac{2\alpha\epsilon p}{3}\left[E^2-\frac{(\mathbf{p}\cdot\mathbf{E})^2}{\epsilon^2}\right] \tag*{(18)}

Compared to the synchrotron case, when E0=B0E_0=B_0, the ratio of the third term in Eq. (17) to the force term in the pure magnetic field case becomes

FEFB=1−(pxEx+pyEy)2E02ϵ2(19)\frac{F_E}{F_B}=1-\frac{(p_xE_x+p_yE_y)^2}{E_0^2\epsilon^2} \tag*{(19)}

The second term breaks the azimuthal symmetry observed in synchrotron radiation, since there are regions where FE/FBF_E/F_B can vanish. We assume Ex(t=0)=Ey(t=0)=E0E_x(t=0)=E_y(t=0)=E_0, which results in identical evolution for ExE_x and EyE_y. Because the plasma is relativistically thermal, we can approximate p∼ϵp\sim\epsilon. For particles in the distribution located at high (px,py)(p_x,p_y) and (−px,−py)(-p_x,-p_y), the second term in Eq. (19) cancels. In the remaining regions of momentum space, the second term in Eq. (19) is negligible, and the radiation reaction force behaves similarly to the pure magnetic field case. This leads to the momentum distribution that we can see in the lower row of the first column of Figure 4

In the second column of Figure 4, we show how the plasma distribution evolves over time, at t=0.4tRt=0.4t_R. In the upper row, the ring momentum distribution is clearly visible, although its radius has not yet reached its maximum. The radiation force, which is strongest at higher momentum, pushes the particles toward lower momentum. As the particles bunch toward lower momentum, increased radiation leads to the overlapping of different particle layers, causing the peak momentum to shift toward higher energy. In the lower row, a similar process occurs, but with a different anisotropy. The azimuthally asymmetric radiation force pushes particles with higher perpendicular momentum toward lower momentum, where the radiation force is weaker. Over time, as more particles are pushed toward lower momentum, the peak momentum shifts to higher values, as in the synchrotron case, but with an asymmetric shape. In particular, particles located at high (px,py)(p_x,p_y) and high (−px,−py)(-p_x,-p_y) remain almost unaffected by radiation, compared to the synchrotron case where the radiation is azimuthally symmetric. This is because, at high (±px,±py)(\pm p_x,\pm p_y), the second term of Eq. (18) nearly cancels the first term, resulting in minimal net radiation.

In the third column of Figure 4, we display the plasma distribution at t=4tRt = 4t_R, where the plasma exhibits a new shape due to radiation effects. In the upper row, the radius of the ring momentum has already reached its maximum, and the higher-momentum states remain unoccupied. In the lower row, the plasma distribution resembles that of the synchrotron case when considering the azimuthal angles 90∘90^\circ–180∘180^\circ and 270∘270^\circ–360∘360^\circ. At these angles, the momentum component pxp_x has an opposite sign to pyp_y, which leads to a cancellation in the second term of Eq. (18), resulting in a strong radiation reaction force. Distinct peaks appear particularly at 135∘135^\circ and 315∘315^\circ, whereas at 45∘45^\circ and 225∘225^\circ, the plasma remains almost unchanged from its initial state, since the radiation force is minimal at these angles.

In the case of synchrotron radiation, the symmetry of the ring-shaped momentum distribution allows us to follow the time evolution of the anisotropic distribution by monitoring the evolution of its radius. For the case of an external electric field, a similar scaling of the anisotropy can be observed, but only when focusing on specific angles in the pxp_x–pyp_y plane. As shown in the second and third panels of the lower row of Figure 4, the distance separating the two maxima increases with time, indicating that dynamics similar to those in the synchrotron case can also be studied here. In the synchrotron case, we defined the radius of the ring-momentum as the value of the perpendicular momentum for which ∂f/∂p⊥=0\partial f/\partial p_\perp= 0. In the case of electric field, this definition cannot be applied uniformly to all angles in the pxp_x–pyp_y plane. However, by fixing the angle to 135∘135^\circ, we can define a radius-like distance in the same way as we did for the synchrotron case.

In Figure 5, we show the time evolution of the ring-momentum anisotropy radius dd (denoted as dd rather than rr, since in the electric-field case there is no symmetric structure) for the magnetic field case, together with the corresponding behavior in the electric-field case. In the upper panel, the initial plasma distribution follows a Maxwell–Boltzmann distribution with pth=10p_{\rm th}=10. The radius in the magnetic-field case reaches a peak that is almost the same as in the electric-field case. However, the time required to reach the peak is longer in the electric-field case. In the lower panel of Figure 5, we show the same information as in the upper panel, but for pth=20p_{\rm th}=20. We observe nearly the same peak radius in both the electric- and magnetic-field cases, although the characteristic time

Time evolution of the ring momentum anisotropy radius for electric and magnetic fields

Figure 5. The distance dd for the ring momentum distribution for both cases of pure electric and magnetic field. In the first panel we used pth=10p_{\rm th}=10 while in the second one we had pth=20p_{\rm th}=20.

scales differ. The reason it takes longer for the electric field to reach a comparable peak is that the radiation process is slower. In the synchrotron case, the acceleration is constant in time, leading to a faster energy loss due to continuous radiation. In contrast, in the electric-field case the acceleration is maximized only when the field reaches its peak. Since the field oscillates in time, there are intervals when the field strength is below its maximum and the radiation is weaker. This results in a slower overall energy loss, and it takes longer for the bunching process to develop and produce the anisotropy.

Before we end our analysis of the anisotropy formation in a pure electric field, we examine the damping of the field amplitude. Due to radiation, the energy stored in the electric field is damped over time. In Figure 6, we show the time evolution of the electric field normalized to its initial value, using pth=20p_{\rm th}=20, E0=0.05E_0=0.05, and n0=7×1028 cm−3n_0=7 \times10^{28}\ \mathrm{cm}^{-3}. As seen in the figure, the peak field decreases with time. In particular, at t=4000t=4000—the time at which the radius of the ring momentum induced by radiation reaches its maximum—the field amplitude has decreased to 60% of its initial value.

Time evolution of the normalized electric field

Figure 6. Time evolution of the normalized electric field, Enor=E(tn)/E0E_{\mathrm{nor}}=E(t_n)/E_0, as a function of time using the plasma density n0=7×1028 cm−3n_0=7\times10^{28}\ \mathrm{cm}^{-3}.

Radiation due to electric and magnetic fields

In this subsection, we investigate the radiation effects on a plasma in the presence of both electric and magnetic fields. The electric field is treated self-consistently, as in the previous subsection, while the magnetic field is considered as an external and constant field, as discussed earlier in the manuscript. The initial plasma distribution is relativistic hot and is given by Eq. (14). To analyze how the combination of the electric and magnetic fields affects radiation in a relativistic plasma, we start by examining the radiation force in Eqs. (7) and (8) in more detail. We consider a relativistic temperature with pth≫1p_{\mathrm{th}}\gg1, and a dense plasma such that the gamma-factor γE\gamma_E associated with the motion induced by the electric field satisfies γE<γth\gamma_E<\gamma_{\mathrm{th}}, where γth=1+pth2\gamma_{\mathrm{th}}=\sqrt{1+p_{\mathrm{th}}^2}. Thus, the dominant contribution in Eqs. (7) and (8) comes from the term of higher order in pp (or equivalently, in ϵ\epsilon). We consider a plasmas with pth=20p_{\mathrm{th}}=20, and can approximate ϵ≈p\epsilon\approx p. The fourth term in Eqs. (7) and (8) scales as second order in pp, while the second and third terms are zeroth order in pp. The first term is proportional to dE/dtdE/dt, which is known to vanish in the relativistic regime compared with the fourth term. Keeping only the fourth term, we can express the radiation force as

∣FR∣=23αϵpB02[p2ϵ2+E02B02−2B0ϵ(Expy−Eypx)−1ϵ2B02(pxEx+pyEy)2](20)|\mathbf{F}_R|=\frac{2}{3}\alpha\epsilon p B_0^2\left[\frac{p^2}{\epsilon^2}+\frac{E_0^2}{B_0^2}-\frac{2}{B_0\epsilon}(E_xp_y-E_yp_x)-\frac{1}{\epsilon^2B_0^2}(p_xE_x+p_yE_y)^2\right] \tag*{(20)}

The first term is familiar from the synchrotron case, the second and fourth terms originate from pure electric-field radiation, and the third arises from the combined action of electric and magnetic fields. Note that both the first and second terms yield an azimuthally symmetric radiation force, whereas the third and fourth terms can break this symmetry depending on the values of ExE_x and EyE_y. In the limit of E0≪B0E_0\ll B_0, we can keep up to first order in E0/B0E_0/B_0 and get the radiation force ∣FR∣|\mathbf{F}_R| divided with the pure magnetic field radiation force ∣FR,E=0∣|\mathbf{F}_{R,E=0}|

∣FR∣∣FR,E=0∣=1−2ϵp2B0(Expy−Eypx)(21)\frac{|\mathbf{F}_R|}{|\mathbf{F}_{R,E=0}|}=1-\frac{2\epsilon}{p^2B_0}(E_xp_y-E_yp_x) \tag*{(21)}

This equation shows that the electric field can enhance and decrease the total radiation force depending on the values of the two components of the electric field. To investigate how the electric field influences radiation, we begin our analysis of electric-field propagation in a magnetized plasma by considering the limit E0≪B0E_0\ll B_0. The aim is to smoothly introduce the effect of radiation reaction due to the electric field on the azimuthally symmetric magnetic-field case studied in Section III B, and to examine how the ring momentum anisotropy is modified. In Figure 7, we present simulation results of radiation reaction in the combined presence of electric and magnetic fields. In particular, we illustrate the behavior of the radiation force ∣FRf0∣|\mathbf{F}_{R f0}| within the initial plasma distribution. We adopt pth=20p_{\mathrm{th}} = 20 to ensure a relativistic temperature. The plasma density is chosen as n0=7×1025 cm−3n_0 = 7 \times10^{25}\ \mathrm{cm}^{-3}, which corresponds to a rest-frame plasma frequency of ωp=0.0096\omega_p = 0.0096 (note that in the lab frame this frequency is reduced due to the ωp/γ\omega_p/\gamma scaling). For the magnetic field, we set B0=0.05B_0 = 0.05, giving a cyclotron frequency of ωc=0.05\omega_c = 0.05. Thus, we study the regime where ωc≫ωp\omega_c \gg\omega_p, implying that ωc∼ωuh\omega_c \sim\omega_{\mathrm{uh}}, with ωuh\omega_{\mathrm{uh}} denoting the upper hybrid frequency.

Momentum distribution of the radiation force in the plasma

Figure 7. The momentum distribution of the radiation force in the plasma, given by ∣FR∣f0|\mathbf{F}_R|f_0, is shown in panels (a)–(e). In (a), we consider a pure magnetic field (i.e., no electric field). In (b), we use E0=0.005E_0=0.005, B0=0.05B_0=0.05, and t=0t=0. In (c), the parameters are E0=0.01E_0=0.01, B0=0.05B_0=0.05, and t=0t=0. In (d), we take E0=0.005E_0=0.005, B0=0.05B_0=0.05, and t=0.6tRt=0.6t_R. Finally, in (e), we use E0=0.01E_0=0.01, B0=0.05B_0=0.05, and t=0.6tRt=0.6t_R.

In Figure 7(a), we show the radiation reaction force in the case of a pure magnetic field, which serves primarily as a reference for comparison with the subsequent subfigures where the electric field modifies the force structure. In Figure 7(b), we include an electric field with E0=0.005E_0 = 0.005 and display the radiation reaction force at t=0t = 0, when Ex=Ey=E0E_x = E_y = E_0. Comparing this result with Figure 7(a), two regions exhibit notable differences. The first is at large pxp_x and negative pyp_y, where the second term in Eq. (21) becomes positive, leading to an enhanced radiation force. The second is at large positive pyp_y and negative pxp_x, where the same term becomes negative, producing a weakened force, as seen in Figure 7(b). In the remaining regions of the pxp_x–pyp_y plane, the behavior closely resembles the pure magnetic-field case. In particular, for combinations of large positive pxp_x and pyp_y or large negative pxp_x and pyp_y, the contributions from the two parts of the second term in Eq. (21) effectively cancel.

In Figure 7(c), we consider E0=0.01E_0 = 0.01 and display the radiation reaction force at t=0t = 0. The same qualitative effects as in Figure 7(b) are observed, but the modifications induced by the electric field are more pronounced since the field amplitude is twice as large, resulting in a doubled contribution from the second term in Eq. (21). In Figure 7(d–Figure 7e), we show the radiation force at t=0.6tRt = 0.6t_R for E0=0.005E_0 = 0.005 and E0=0.01E_0 = 0.01, respectively. In contrast to Figure 7(b–Figure 7c), the electric field components are no longer equal, but instead ExE_x is negative while EyE_y is positive, with differing magnitudes. This asymmetry shifts the regions in the pxp_x–pyp_y plane where the second term of Eq. (21) contributes positively or negatively to the radiation force.

To understand how the new momentum-shape of the total radiation force presented in Figure 7 would affect the initial plasma distribution, we plot the plasma distribution after radiation at t=0.6tRt = 0.6t_R in Figure 8. In Figure 8(a), we plot the plasma distribution for the pure synchrotron case, which serves as a reference for the other subplots. In Figure 8(b), we display the plasma distribution in the presence of both electric and magnetic fields, with B0=0.05B_0 = 0.05, E0=0.01E_0 = 0.01, and plasma density n0=7×1025 cm−3n_0 = 7 \times10^{25}\ \mathrm{cm}^{-3}, the same parameters used in Figure 7(c,Figure 7e), i.e. we have ωc≫ωp\omega_c \gg\omega_p. As shown previously in Figure 7(c,Figure 7e), two distinct regions emerge where the net radiation force is stronger or weaker compared to the pure synchrotron case. In the right half of the pxp_x–pyp_y plane, the radiation force is enhanced due to the positive contribution of the electric field, causing some electrons to be pushed toward the weaker-radiation region in the left half of the plane. This redistribution can be confirmed by comparing the color map in Figure 8(b) with that in Figure 8(a): in the left half of the pxp_x–pyp_y plane, the number of electrons increases due to the weaker radiation force, consistent with the behavior shown in Figure 7(c,Figure 7e).

Plasma distributions in momentum space

Figure 8. The plasma distribution at t=0.6tRt = 0.6t_{R} is shown as follows: (a) pure magnetic field in plasma, (b) with an electric field E0=0.01E_{0} = 0.01 and plasma density n0=7×1025 cm−3n_{0} = 7 \times10^{25}\ \mathrm{cm}^{-3}, and (c) with E0=0.01E_{0} = 0.01 and a plasma density of n0=7×1027 cm−3n_{0} = 7 \times10^{27}\ \mathrm{cm}^{-3}.

Increasing the plasma density raises the oscillation frequency of the electric field and we enter the regime where ωp≫ωc\omega_p \gg\omega_c. Consequently, during the radiation time 0.6tR0.6t_R, the components of the electric field, ExE_x and EyE_y, undergo many oscillation periods. This leads to the possibility that the second term of Eq. (21) contributes both positively and negatively to the net radiation force within the same regions of the pxp_x–pyp_y plane. This effect is illustrated in Figure 8(c), where we use E0=0.01E_0 = 0.01, B0=0.05B_0 = 0.05, and n0=7×1027 cm−3n_0 = 7 \times10^{27}\ \mathrm{cm}^{-3}. The resulting plasma distribution is very similar to the pure synchrotron case in Figure 8(a). The high frequency of the self-consistent electric field causes the positive and negative contributions of the electric field to the net radiation force to shift rapidly across the pxp_x–pyp_y plane. As a result, different regions of momentum space experience alternating contributions at different times, which produces an almost symmetric plasma distribution. Moreover, the duration over which any given region of momentum space is influenced by these positive or negative contributions is very short, since the higher oscillation frequency causes the contributions to move quickly from one region to another.

Studying the time evolution of the ring radius in Figure 8(c), we find that it closely resembles the case of a pure magnetic field. One might suspect that the moderate electric field amplitude of E0=0.01E_0 = 0.01, compared to B0=0.05B_0 = 0.05, is the reason why the time evolution of the ring radius appears almost unaffected by the presence of the electric field. However, this is not the case, as the rapid oscillations of the electric field suppress the influence of the azimuthally asymmetric term on the total radiation-reaction force. Increasing the electric field amplitude to E0=0.05E_0 = 0.05, while keeping the same magnetic field and plasma density, the time evolution of the ring radius remains essentially unchanged. This behavior can be explained using Eq. (20). By taking E0=B0E_0 = B_0 and p∼ϵp \sim\epsilon, which is a valid approximation for a relativistic thermal plasma (as radiation primarily occurs at large momentum), we obtain

∣FR∣=23αϵpB02[1+1−2B0ϵ(Expy−Eypx)−1ϵ2B02(pxEx+pyEy)2]|\mathbf{F}_R| = \frac{2}{3}\alpha\epsilon p B_0^2 \left[1 + 1 - \frac{2}{B_0\epsilon}(E_xp_y - E_yp_x) - \frac{1}{\epsilon^2B_0^2}(p_xE_x + p_yE_y)^2\right]

By examining the time evolution of ExE_x and EyE_y, we find that their behaviors closely coincide. Thus, we may approximate Ey≈ExE_y \approx E_x. This leads to

1ϵ2B02(pxEx+pyEy)2=Ex2B02\frac{1}{\epsilon^2B_0^2}(p_xE_x + p_yE_y)^2 = \frac{E_x^2}{B_0^2}

asymmetric term to the total radiation force shifts quickly in momentum space. This stands in contrast to the first term of Eq. (22), which remains constant over time. Consequently, the plasma distribution is only weakly affected, and the dynamics of the ring-momentum distribution remain largely unchanged.

The maximum value of ExE_{x} is E0E_{0}, which is equal to B0B_{0}. At the time when the electric field reaches its peak, we obtain

∣FR∣=23αεpB02[1+1−2B0ε(Expy−Eypx)−1]=23αεpB02[1−2B0ε(Expy−Eypx)](22)\begin{aligned} \lvert F_{R} \rvert= \frac{2}{3}\alpha\varepsilon p B_{0}^{2}\left[1+1-\frac{2}{B_{0}\varepsilon}(E_{x}p_{y}-E_{y}p_{x})-1\right] \\ &= \frac{2}{3}\alpha\varepsilon p B_{0}^{2}\left[1-\frac{2}{B_{0}\varepsilon}(E_{x}p_{y}-E_{y}p_{x})\right] \tag*{(22)} \end{aligned}

The second and fourth terms cancel each other when ExE_{x} and EyE_{y} attain their peak values. At the time when the electric field components have the value 0, the results in Eq. (22) still hold as the second term that is E02/B02=(Ex2+Ey2)/B02E_{0}^{2}/B_{0}^{2}=(E_{x}^{2}+E_{y}^{2})/B_{0}^{2} would vanish together with the fourth term. At intermediate times, the first term in Eq. (22) takes values between one and two. After applying these assumptions, only the third term of Eq. (20) remains, whose effect is illustrated in Figure 7(b–Figure 7e) (note that we have used p∼εp \sim\varepsilon). Thus, the conclusions derived for E0≪B0E_{0} \ll B_{0} remain valid also in the case E0=B0E_{0}=B_{0}. Even for E0/B0=1E_{0}/B_{0}=1, the azimuthally asymmetric part of the radiation-reaction force does not alter the anisotropy formation, provided that the cyclotron frequency is much smaller than the upper hybrid frequency. Due to the rapid oscillations of the electric field driven by the high upper hybrid frequency, the contribution of the azimuthally asymmetric term to the total radiation force shifts quickly in momentum space.

Discussion

The main objective of this work is to investigate how the radiation reaction affect the momentum distribution of a plasma that is initial in equilibrium. We have used the Landau-Lifshitz force to model the recoil force of the radiation on the plasma. As the recoil force is non-conservative, the cooling of the plasma due to radiation becomes anisotropic in momentum-space, with stronger cooling on the region with stronger magnitude of the radiation reaction force. For the case of pure magnetic field it is shown that the azimuthal-free radiation reaction force lead to ring-momentum distribution. The time evolution characterizing the ring-momentum distribution is shown to agree well with the corresponding analytical solution. Studying radiation due to strong magnetic field in anisotropic plasma, it is shown that introduction of initial anisotropy in momentum-space reduces the energy loss rate and as consequence the radius of the ring-momentum distribution become smaller. Due to the temperature imbalance between the xx- and yy-directions, the magnetic force induces a drift of particles in momentum space from the higher-temperature region to the lower-temperature one. This redistribution reduces the overall energy loss due to radiation.

Studying the radiation due to electric field, we found that the radiation reaction force is not azimuthal-free, i.e. the force is zero in certain regions in px−pyp_{x}-p_{y}-plane, while in other regions it is non-zero. In particular, the force is non-zero at the 90°–180° and 270°–360° sectors of the px−pyp_{x}-p_{y}-plane, while the remaining quadrants it is zero. The region where the force is non-zero experience a similar bunching of the electron due to the cooling effect of radiation. This leads to the creation of sector-dependent ring-momentum distribution with a similar time-dependent dynamics of the ring-radius as in the magnetic field-case. For the rest of the sectors in the px−pyp_{x}-p_{y}-plane, the plasma remains unaffected.

When both electric and magnetic fields are present in the plasma, the radiation reaction force can be decomposed into two contributions. The first is an azimuthally independent force that remains constant in time and drives the formation of a ring-momentum distribution. The second depends on the azimuthal angle, enhancing or reducing the total radiation force depending on the region of the px−pyp_{x}-p_{y} plane. Since the second contribution is determined by the two components of the self-consistent electric field, which vary with time, the regions in momentum space where the total force is stronger or weaker continuously shift. The resulting anisotropy depends on whether the cyclotron frequency is near the upper hybrid frequency or not. When the cyclotron frequency is nearly resonant with the upper hybrid frequency, the electric-field components vary only slowly over the plasma cooling time. In this regime, the azimuthal-dependent force remains localized in the same region of momentum space long enough to significantly distort the plasma distribution, thereby breaking the azimuthal symmetry of the ring-momentum structure. In contrast, when the upper hybrid frequency is much larger than the cyclotron frequency, the rapid oscillations of the electric field components cause the positive and negative contributions of the azimuthal-dependent force to shift quickly across momentum space. As a result, the azimuthal symmetry of the ring-momentum anisotropy is preserved.

The physical parameters considered in this work, namely B0=0.05B_0 = 0.05, E0=0.01E_0 = 0.01, and the plasma density n0=7×1025n_0 = 7 \times10^{25}–1027 cm−310^{27}\ \mathrm{cm}^{-3}, occur naturally in the magnetospheres of compact astrophysical objects [10]. In a laboratory context, it is anticipated that next-generation strong-field experiments will be capable of creating analogous astrophysical conditions [7].

The Vlasov system with a linearly polarized electric field and a constant magnetic field, including Ampère’s law and the radiation reaction force, has been solved using a two-step numerical scheme. The numerical results show good agreement with the analytical solution in the limit of a vanishing electric field. Our focus was to investigate how the propagation of the electric field in a magnetized plasma affects the development of the ring-momentum distribution. Thus, the physics related to the onset of instabilities (which require much longer timescales) arising after the development of anisotropies is beyond the scope of this work. For future studies, it would be of interest to investigate such instabilities, for example the electron cyclotron maser instability (see [4, 40, 41, 42, 43]). Furthermore, studying the development of anisotropies in momentum space driven by the electromagnetic field geometry would provide deeper insights into the nonlinear dynamics of magnetized plasmas and their associated radiation processes.

ACKNOWLEDGMENT

The author acknowledges support from the Knut and Alice Wallenberg Foundation. A special thank to Gert Brodin and David Reis for fruitful discussions.

References

  1. [1]K. Qu, S. Meuren, and N. J. Fisch, Signature of collective plasma effects in beam-driven qed cascades, Physical review letters 127, 095001 (2021).DOI
  2. [2]A. Griffith, K. Qu, and N. J. Fisch, Radiation reaction kinetics and collective qed signatures, Physics of Plasmas 31 (2024).arxiv.org/abs/2312.04700
  3. [3]K. Qu, S. Meuren, and N. J. Fisch, Collective plasma effects of electron–positron pairs in beam-driven qed cascades, Physics of Plasmas 29 (2022).DOI
  4. [4]P. J. Bilbao and L. O. Silva, Radiation reaction cooling as a source of anisotropic momentum distributions with inverted populations, Physical Review Letters 130, 165101 (2023).DOI
  5. [5]P. J. Bilbao, R. J. Ewart, F. Assuncao, T. Silva, and L. O. Silva, Ring momentum distributions as a general feature of vlasov dynamics in the synchrotron dominated regime, Physics of Plasmas 31 (2024).DOI
  6. [6]S. Bulanov, G. Grittani, R. Shaisultanov, T. Z. Esirkepov, C. Ridgers, S. Bulanov, B. Russell, and A. Thomas, On the energy spectrum evolution of electrons undergoing radiation cooling, Fundamental Plasma Physics 9, 100036 (2024).arxiv.org/abs/2310.15623
  7. [7]H. Chen and F. Fiuza, Perspectives on relativistic electron–positron pair plasma experiments of astrophysical relevance using high-power lasers, Physics of Plasmas 30 (2023).DOI
  8. [8]Z. Gong, F. Mackenroth, X. Yan, and A. Arefiev, Radiation reaction as an energy enhancement mechanism for laser-irradiated electrons in a strong plasma magnetic field, Scientific reports 9, 17181 (2019).
  9. [9]L. Comisso and L. Sironi, Pitch-angle anisotropy controls particle acceleration and cooling in radiative relativistic plasma turbulence, Physical Review Letters 127, 255102 (2021).
  10. [10]V. Zhdankin, M. W. Kunz, and D. A. Uzdensky, Synchrotron firehose instability, The Astrophysical Journal 944, 24 (2023).DOI
  11. [11]T. Liseykina, S. Popruzhenko, and A. Macchi, Inverse faraday effect driven by radiation friction, New Journal of Physics 18, 072001 (2016).DOI
  12. [12]P. Zhang, S. Bulanov, D. Seipt, A. Arefiev, and A. Thomas, Relativistic plasma physics in supercritical fields, Physics of Plasmas 27 (2020).DOI
  13. [13]Z. Chen, S. Meuren, E. Gerstmayr, V. Yakimenko, P. H. Bucksbaum, and D. A. Reis, Preparation of strong-field qed experiments at facet-ii, in High Intensity Lasers and High Field Phenomena (Optica Publishing Group, 2022) pp. HF4B–6.DOI
  14. [14]D. L. Burke, R. C. Field, G. Horton-Smith, J. E. Spencer, D. Walz, S. C. Berridge, W. M. Bugg, K. Shmakov, A. W. Weidemann, C. Bula, K. T. McDonald, E. J. Prebys, C. Bamber, S. J. Boege, T. Koffas, T. Kotseroglou, A. C. Melissinos, D. D. Meyerhofer, D. A. Reis, and W. Ragg, Positron production in multiphoton light-by-light scattering, Phys. Rev. Lett. 79, 1626 (1997).
  15. [15]J. M. Cole, K. T. Behm, E. Gerstmayr, T. G. Blackburn, J. C. Wood, C. D. Baird, M. J. Duff, C. Harvey, A. Ilderton, A. S. Joglekar, K. Krushelnick, S. Kuschel, M. Marklund, P. McKenna, C. D. Murphy, K. Poder, C. P. Ridgers, G. M. Samarin, G. Sarri, D. R. Symes, A. G. R. Thomas, J. Warwick, M. Zepf, Z. Najmudin, and S. P. D. Mangles, Experimental evidence of radiation reaction in the collision of a high-intensity laser pulse with a laser-wakefield accelerated electron beam, Phys. Rev. X 8, 01020 (2018).
  16. [16]A. Fedotov, A. Ilderton, F. Karbstein, B. King, D. Seipt, H. Taya, and G. Torgrimsson, Advances in qed with intense background fields, Physics Reports 1010, 1 (2023).DOI
  17. [17]H. Abramowicz, U. Acosta, M. Altarelli, R. Assmann, Z. Bai, T. Behnke, Y. Benhammou, T. Blackburn, S. Boogert, O. Borysov, et al., Conceptual design report for the luxe experiment, The European Physical Journal Special Topics 230, 2445 (2021).DOI
  18. [18]A. Gonoskov, T. G. Blackburn, M. Marklund, and S. S. Bulanov, Charged particle motion and radiation in strong electromagnetic fields, Rev. Mod. Phys. 94, 045001 (2022).DOI
  19. [19]G. Ford and R. O’connell, Radiation reaction in electrodynamics and the elimination of runaway solutions, Physics Letters A 157, 217 (1991).DOI
  20. [20]L. D. Landau, The classical theory of fields, Vol. 2 (Elsevier, 2013).
  21. [21]S. V. Bulanov, T. Z. Esirkepov, M. Kando, J. K. Koga, and S. S. Bulanov, Lorentz-abraham-dirac versus landau-lifshitz radiation friction force in the ultrarelativistic electron interaction with electromagnetic wave (exact solutions), Phys. Rev. E 84, 056605 (2011).pubmed.ncbi.nlm.nih.gov/22181534
  22. [22]A. D. Piazza, Exact solution of the landau-lifshitz equation in a plane wave, Letters in Mathematical Physics 83, 305 (2008).arxiv.org/abs/0801.1751
  23. [23]M. Vranic, J. L. Martins, R. A. Fonseca, and L. O. Silva, Classical radiation reaction in particle-in-cell simulations, Computer Physics Communications 204, 141 (2016).DOI
  24. [24]E. Wallin, A. Gonoskov, C. Harvey, O. Lundh, and M. Marklund, Ultra-intense laser pulses in near-critical underdense plasmas–radiation reaction and energy partitioning, Journal of Plasma Physics 83 (2017).DOI
  25. [25]M. Vranic, T. Grismayer, R. A. Fonseca, and L. O. Silva, Quantum radiation reaction in head-on laser-electron beam interaction, New Journal of Physics 18, 073035 (2016).DOI
  26. [26]R. Hakim and A. Mangeney, Relativistic kinetic equations including radiation effects. i. vlasov approximation, Journal of Mathematical Physics 9, 116 (1968).DOI
  27. [27]D. A. Burton and A. Noble, Aspects of electromagnetic radiation reaction in strong fields, Contemporary Physics 55, 110 (2014).DOI
  28. [28]M. Kunze and A. D. Rendall, The vlasov-poisson system with radiation damping, in Annales Henri Poincaré, Vol. 2 (Springer, 2001) pp. 857–886.DOI
  29. [29]Y. Elskens and M.-H. Kiessling, Microscopic foundations of kinetic plasma theory: The relativistic vlasov–maxwell equations and their radiation-reaction-corrected generalization, Journal of Statistical Physics 180, 749 (2020).DOI
  30. [30]A. Noble, J. Gratus, D. Burton, B. Ersfeld, M. R. Islam, et al., Kinetic treatment of radiation reaction effects, in Proc. of SPIE Vol, Vol. 8079, pp. 80790L–1.
  31. [31]V. I. Berezhiani, R. D. Hazeltine, and S. M. Mahajan, Radiation reaction and relativistic hydrodynamics, Phys. Rev. E 69, 056406 (2004).pubmed.ncbi.nlm.nih.gov/15244950
  32. [32]R. D. Hazeltine and S. M. Mahajan, Closed fluid description of relativistic, magnetized plasma interacting with radiation field, Phys. Rev. E 70, 036404 (2004).pubmed.ncbi.nlm.nih.gov/15524643
  33. [33]V. I. Berezhiani, S. M. Mahajan, and Z. Yoshida, Plasma acceleration and cooling by strong laser field due to the action of radiation reaction force, Phys. Rev. E 78, 066403 (2008).pubmed.ncbi.nlm.nih.gov/19256958
  34. [34]G. T. Dalakishvili, A. D. Rogava, and V. I. Berezhiani, Role of radiation reaction forces in the dynamics of centrifugally accelerated particles, Phys. Rev. D 76, 045003 (2007).DOI
  35. [35]H. Al-Naseri and G. Brodin, Radiation reaction effects in relativistic plasmas: The electrostatic limit, Physical Review E 107, 035203 (2023).DOI
  36. [36]H. Al-Naseri and G. Brodin, Probing the transition from classical to quantum radiation reaction in relativistic plasma, arXiv preprint arXiv:2506.22577 (2025).arxiv.org/abs/2506.22577
  37. [37]G. Brodin, H. Al-Naseri, J. Zamanian, G. Torgrimsson, and B. Eliasson, Plasma dynamics at the schwinger limit and beyond, Physical Review E 107, 035204 (2023).DOI
  38. [38]H. Al-Naseri, J. Zamanian, and G. Brodin, Plasma dynamics and vacuum pair creation using the dirac-heisenberg-wigner formalism, Physical Review E 104, 015207 (2021).DOI
  39. [39]D. A. Uzdensky and S. Rightley, Plasma physics of extreme astrophysical environments, Reports on Progress in Physics 77, 036902 (2014).DOI
  40. [40]D. Melrose and M. Wheatland, Is cyclotron maser emission in solar flares driven by a horseshoe distribution?, Solar Physics 291, 3637 (2016).DOI
  41. [41]R. A. Cairns, D. Speirs, K. Ronald, I. Vorgul, B. Kellett, A. Phelps, and R. Bingham, A cyclotron maser instability with application to space and laboratory plasmas, Physica Scripta 2005, 23 (2005).
  42. [42]R. Bingham and R. A. Cairns, Generation of auroral kilometric radiation by electron horseshoe distributions, Physics of Plasmas 7, 3089 (2000).DOI
  43. [43]D. Le Quéau, R. Pellat, and A. Roux, Direct generation of the auroral kilometric radiation by the maser synchrotron instability: Physical mechanism and parametric study, Journal of Geophysical Research: Space Physics 89, 2831 (1984).DOI

Paper details

vlnc:W_bpgbj4s3
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents