WorksPhysics and Astronomy
Energy partition in resonant instabilities in multi-component plasmas
Abstract
Space and astrophysical plasmas are often collisionless and exhibit non-equilibrium velocity distributions. When these departures from equilibrium are sufficiently large, the plasma becomes unstable to various kinetic micro-instabilities, which provide a mechanism for transferring energy from the particles to the electromagnetic (EM) fields. In multi-component plasmas, different species interact with the generated EM waves in distinct ways, acting as sources or sinks of wave energy through resonant wave–particle interactions. Consequently, the nonlinear evolution of kinetic instabilities is governed not only by the species driving the instability but also by the resonant response of other plasma populations. In this work, considering proton--particle plasmas with parameters representative of the solar wind, we investigate the generation, saturation, and energy partition of kinetic micro-instabilities, such as Alfvén/ion-cyclotron and fast-magnetosonic/whistler wave instabilities. We combine linear Vlasov–Maxwell theory and hybrid particle-in-cell (PIC) simulations to analyze these instabilities and to quantify the exchange of energy and momentum between ion populations during the instability evolution. Our results demonstrate that minor ions play a fundamental role in regulating the nonlinear saturation of kinetic instabilities and the transfer of energy among plasma populations.
Introduction
Space and astrophysical plasmas are often in a collisionless state, and many natural plasma systems exhibit non-equilibrium particle velocity distribution functions (VDFs; E. Marsch [34]; D. Verscharen et al. [58]). For example, space plasmas like the solar wind often exhibit temperature anisotropies with respect to the local magnetic field and field-aligned relative drifts between plasma species (A. J. Hundhausen et al. [24]; J. R. Asbridge et al. [2]; E. Marsch et al. [37]). Additional deviations from the Maxwell–Boltzmann distribution persist, for instance, in the form of suprathermal tails or more localized deviations in velocity space (C. C. Goodrich & A. J. Lazarus [19]; S. P. Christon et al. 1988; C. M. Espinoza et al. [11]; A. V. Eylade et al. 2025).
If these deviations from equilibrium are large, the plasma becomes unstable and excites kinetic micro-instabilities (S. P. Gary [16]). During the linear phase of these instabilities, the amplitudes of the unstable normal modes of the plasma grow exponentially in time at the expense of the free energy contained in the driving non-equilibrium features in the particle VDFs. In the case of resonant instabilities, particle scattering on the growing fluctuations in Email: [email protected] the electromagnetic field of the unstable wave modifies the particle energy in such a way that the total energy of the system remains conserved [26, 27]. Micro-instabilities are thus a mechanism for transferring energy from plasma particles to the electromagnetic field.
In the case of the solar wind, the dynamically most important plasma species are the protons, the electrons, and the -particles. All of these species exhibit deviations from Maxwellian equilibrium distributions. For instance, the protons often show a core () and a beam () population, each of which presents different bulk velocities and temperature anisotropies with , where and are the temperatures of species in the directions perpendicular and parallel with respect to the magnetic field [13, 10, 8]. Moreover, the -particles often have a bulk speed [36, 49]. Ultimately, all of these deviations from equilibrium are potential drivers of kinetic micro-instabilities [41, 35, 15, 9, 17, 18, 23, 40, 3, 47, 38, 4, 55, 25, 62, 28, 43, 54].
The response of different plasma species to the unstable normal modes depends on the individual VDFs as well as the charge and mass of the species. In such a multi-species plasma, some species may release energy while others absorb energy through resonant wave–particle interactions. An overall micro-instability with net energy transfer from the plasma particles to the electromagnetic field is possible only if at least one species releases more energy than the combined energy absorption of the other species.
While some species emit and others absorb energy, their VDFs change through pitch-angle scattering, modifying the velocity moments of the VDF according to the kinetic theory of resonant wave–particle interactions [50, 51] (C.-Y. Tu & E. Marsch 2002). Micro-instability can thus couple all involved particle species and cause free energy to flow from one driving species into other species via electromagnetic fluctuations. In the following, we quantify the energetics of this inter-species coupling via kinetic micro-instabilities in multi-species plasmas.
Here, using linear Vlasov–Maxwell theory and kinetic plasma simulations, we demonstrate inter-species coupling via kinetic micro-instabilities in plasma configurations comparable to those observed in the solar wind. This article is organized as follows. In § 2, we develop quantitative arguments for the energetics of inter-species coupling through instabilities. In § 3, we present the theory and methods we use to analyze wave–particle interactions and plasma energetics through kinetic theory and simulations, considering four Cases selected to excite kinetic instabilities. Then, Sections 4 and § 5 show the results considering the Alfvén/ion-cyclotron (A/IC) and fast-magnetosonic/whistler (FM/W) instabilities. Finally, in § 6, we discuss our results and summarize our main conclusions.
Energetics of the inter-species coupling through instabilities
In a plasma consisting of particles of species , the total energy is conserved 1:
where
is the kinetic energy of the particles of species . We define the electromagnetic energy , where
is the energy in the electric field 2, and
is the energy in the magnetic field . The integration over is to be taken over the entire velocity space, and the integration over over the entire volume of the plasma.
With this notation, we characterize instabilities by a net increase in at the expense of a reduction in over time, while the overall remains constant. A net increase in can be achieved even when one or more species gain energy, i.e., their individual , as long as the sum over all is negative.
The arguments above based on the total energy balance according to Equation (1) are sufficient for the description of the energetics of instabilities and, by extension, to cases with a net decrease in , of wave damping. However, for a closer inspection of the energy transfer involved, it is useful to separate into contributions from the bulk kinetic energy and the thermal energy of species . We assume that is gyrotropic. With the definitions of the number density
and the bulk velocity
of species , we then rewrite as
where
is the bulk kinetic energy contribution for species ,
the thermal energy contribution of species , and is the Boltzmann constant. In this derivation, we use the definition of the pressure tensor
and define the kinetic temperatures of species perpendicular and parallel to the magnetic field as
and
where is the unit vector of the magnetic field.
Equation (7) separates the particle kinetic energy of species into the contributions from the bulk kinetic energy and the thermal energy. This separation allows us to determine whether, for a given instability, the bulk motion or the thermal pressure is the main source or sink of particle energy for a given species . If there is a meaningful distinction between multiple populations of particles of the same species, e.g., proton core and proton beam populations, the framework also applies and allows for them to be separated as different .
Methodology
To investigate the redistribution of energy among ion populations by instability-driven fluctuations, we consider a set of drifting multi-species plasma configurations representative of conditions in the solar wind. We perform linear predictions and non-linear simulations for each configuration. The selected parameters are chosen to excite either Alfvén/ion-cyclotron (A/IC) or fast-magnetosonic/whistler (FM/W) instabilities, while simultaneously allowing substantial resonant interactions with non-driving ion species and thus energy transfer between species through the excited electromagnetic fluctuations. These configurations are designed such that is positive for one species but negative for another, while is positive.
According to quasilinear theory, as illustrated in Figure 1, wave–particle interactions with the emitted waves cause cyclotron-resonant particles (represented by the vertical green line) to pitch-angle scatter and subsequently diffuse locally in the direction of the tangents (red arrows) of circles given by , where and are the velocity components perpendicular and parallel to the local background field. These circles represent shells of constant energy in the frame comoving with the field-parallel component of the wave’s phase velocity [63, 58], where is the real part of the wave frequency and is the wavevector.
Figure 1. Sketch of quasilinear diffusion due to resonant interactions in space. Solid black and dashed red curves represent constant VDF contours and the constant energy shells in the wave reference frame, respectively. Dot-dashed dark red and green vertical lines correspond to the field-parallel component of the phase velocity of the wave and the field-parallel velocity of cyclotron-resonant particles. Red arrows indicate the directions of the diffusive flux of resonant particles through velocity space.
We model the plasma as a superposition of bi-Maxwellian background VDFs:
where and are the squares of the perpendicular and parallel thermal speeds of the species .
We study four Cases, detailed in Table 1. In all Cases, densities are normalized to the proton (core) density , the ratio of parallel thermal to magnetic pressure for each is defined as , and the drift speed is normalized to the proton (core) Alfvén speed , where is the background magnetic field, and the proton mass.
| Case | Species | ||||
| 1 | Protons (c) | 1.0 | 3.0 | 1.0 | 0.0 |
| -particles () | 0.05 | 0.6 | 4.0 | 0.9 | |
| Electrons (e) | 1.1 | 3.3 | 1.0 | 0.082 | |
| 2 | Proton Core (c) | 1.0 | 1.0 | 1.0 | 0.0 |
| Proton Beam (b) | 0.2 | 0.2 | 1.0 | 1.8 | |
| -particles () | 0.05 | 0.3 | 1.0 | 0.8 | |
| Electrons (e) | 1.3 | 1.3 | 1.0 | 0.338 | |
| 3 | Proton Core (c) | 1.0 | 1.0 | 1.0 | 0.0 |
| Proton Beam (b) | 0.2 | 0.2 | 1.0 | 1.8 | |
| -particles () | 0.01 | 0.06 | 1.0 | 0.8 | |
| Electrons (e) | 1.22 | 1.22 | 1.0 | 0.308 | |
| 4 | Proton Core (c) | 1.0 | 1.0 | 1.0 | 0.0 |
| Proton Beam (b) | 0.2 | 0.2 | 1.0 | 1.8 | |
| Electrons (e) | 1.2 | 1.2 | 1.0 | 0.3 |
Table 1. Plasma parameters for our four Cases under investigation. Note—Case 1 contains a drifting anisotropic -particle population driving the A/IC instability, while Cases 2–4 include a proton-beam component triggering the FM/W instability. Case 4 does not include -particles. For the linear theory, we have set the Alfvén speed to .
Case 1 is composed of three populations: protons (c), drifting -particles (), and electrons (e). This Case is designed to drive the A/IC instability through the anisotropic -particle population with and a relative drift similar to the proton Alfvén speed with respect to the isotropic proton population. This system has and .
In Case 2, we include a proton beam, leaving the system composed of four species: a proton core (c), proton beam (b), drifting -particles (), and electrons (e). In this Case, we study the proton-beam-driven FM/W instability, keeping all species isotropic and introducing a large beam drift (in units of the proton core Alfvén speed). The -particles, while also drifting, do not contribute to the driving of the instability but instead receive energy through wave–particle interactions with the proton-beam-driven unstable waves. This system has , and .
Finally, Cases 3 and 4 serve as control cases for Case 2. Here, we systematically reduce the -particle population concentration (to 1% in Case 3 and 0% in Case 4) to analyze the energy distribution between species when the concentration of a non-driving ion species decreases. These systems have and , while in Cases 3 and 4.
In all four Cases, the electron density and drift velocity with respect to the protons have been selected to guarantee quasi-neutrality and a zero parallel current condition.
Linear Theory
We perform a linear analysis of Vlasov–Maxwell theory for each Case shown in Table 1 using the PLUME (Plasma in a Linear Uniform Magnetized Environment; K. G. Klein & G. G. Howes [29]; K. G. Klein et al. [30]) dispersion-relation solver. PLUME solves the hot-plasma dispersion relation detailed in Chapter 10 of T. H. Stix [52] for a plasma system consisting of an arbitrary number of species with relatively drifting bi-Maxwellian VDFs. The solutions require specifying a reference ratio for the thermal to magnetic pressures and a reference ratio for the thermal speed to the speed of light for reference species . For our calculations, we select the proton core as the reference species (). Each component is further specified by a set of parameters for each species : the temperature anisotropy , the temperature disequilibrium , the relative density ratio , the relative charge ratio , the relative mass ratio , and the drift parallel to the mean magnetic field normalized to the reference Alfvén velocity .
For a given background specified by , PLUME determines the complex normal-mode solutions for a specified wavevector . Here, is the real part of the frequency of the normal mode oscillation and is its growth () or damping () rate. PLUME can follow solutions along user-defined paths in and/or . For this work, we focus on variations in with fixed and fixed for a given Case. If , the wave amplitude grows with time, meaning that . Likewise, if , the wave amplitude decreases and .
In addition to determining the normal-mode solutions, PLUME calculates the eigenfluctuations of the density, velocity, and electromagnetic fields [30], as well as of the fluctuations of the VDF itself [6]. The code also determines the collisionless power absorption or emission from each component:
where according to T. H. Stix [52], is the anti-Hermitian susceptibility tensor for component with evaluated at . From , PLUME calculates the contributions of species to the overall , such that if . If , species makes a positive contribution to the overall , meaning that . Likewise, if , species makes a negative contribution to the overall and .
The code further decomposes this component emission or absorption into contributions from the different resonance orders , which are defined in the resonance condition
The contributions for different fulfill if .
The full expressions are given by K. G. Klein et al. [30]. For this work, we focus on the orders (Landau and transit-time resonances) and (cyclotron resonances). An example of this decomposition is shown in rows 3 and 4 of Figure 2.
Figure 2. Linear dispersion analysis for Case 1, corresponding to an -particle-driven A/IC instability. From top to bottom, the panels show the normalized real frequency , growth/damping rate , proton contribution to the growth rate , and -particle contribution , as functions of parallel wavenumber . In the first row, solid and dashed curves denote and , respectively. From the second to the fourth row, solid and dashed curves represent and , respectively. Solid and dashed curves denote and , respectively. We identify four parallel-propagating modes: two FM/W (not shown) and two A/IC modes. The FM/W modes are stable over the investigated wavevector range. The second row shows the unstable behavior of the forward- and backward-propagating left-handed modes, A/IC A and A/IC B, respectively, consistent with the properties of the A/IC instability. This finding is supported by the wavevector-dependence of the real part of the frequency (first row) and the transverse polarization (not shown). The third and fourth rows give us information about the power absorption/emission of each species, showing that the unstable modes are driven by the -particles that cyclotron-resonate with , while they are subject to significant proton-cyclotron damping via the same resonances. This combined behavior shows a direct pathway for energy transfer from -particles to protons, mediated by A/IC waves.
Simulations
For our nonlinear study, we employ one-dimensional, non-relativistic () hybrid particle-in-cell (PIC) simulations for each Case, in which ions are modeled as kinetic particles and electrons as a charge-neutralizing fluid. We use the code Hybrid-VPIC [5] (A. Le et al. 2021; A. Le et al. 2023), a high-performance open-source hybrid PIC code3
The one-dimensional setup is built by cells in a box of length oriented along the -direction, and , where is the proton (core) inertial length. We initialize the system with particles per cell per component (core and beam protons, and -particles). We set the time-step to be , with the simulation running until . We use a number of particles per cell that ensures energy conservation (see Appendix § 9). The electron fluid is isothermal with scalar pressure. We don’t include explicit collisions or dissipation terms, and we set resistivity and hyper-resistivity to zero.
We scale the numerical particle weights for beam protons and -particles to their density ratios and . We use a quadratic-sum particle shape to deposit the current and charge density on the Cartesian mesh (A. Le et al. 2023, Appendix B); Also, we apply three passes of a “1-2-1” binomial filter to the current and charge on the mesh, and also to the electromagnetic fields when interpolating to particles.
Alfvén Ion-cyclotron instability
We begin our analysis with Case 1 in Table 1, representing a plasma in which the free energy to drive the instability is carried by the -particles with a temperature anisotropy and drift . The resulting shape of their VDF makes the -particles susceptible to the resonant driving of A/IC instability. Next, we perform a linear analysis, evaluate the dispersion relation to identify unstable modes, and determine each ion species’ contribution to the overall growth rate. We then present results from our PIC simulations and compare the linear predictions with the plasma’s nonlinear evolution.
Linear Prediction
Figure 2 shows the linear analysis for Case 1 for the A/IC instability. The panels show, from top to bottom, the normalized real frequency , growth/damping rate , proton contribution to the growth rate , and -particle
Simulation Results
We run a one-dimensional hybrid simulation for the parameters of Case 1 from Table 1. In this Case, the A/IC is driven unstable during the simulation at frequencies and wavenumbers consistent with those predicted in § 4.1 (see Appendix § 10). Figure 3 (left) shows the evolution of the total energy variation
Figure 3. Energy partition from the simulation for Case 1. (left) Evolution of the total-energy variation (black), magnetic-energy variation (red), total kinetic-energy variation (purple), proton kinetic-energy variation (blue), and -particle kinetic-energy variation (green). (right) Changes in bulk () and thermal energy () for protons and -particles.
the magnetic-energy variation (red), the total kinetic-energy variation (purple), the proton kinetic-energy variation (blue), and the -particle kinetic-energy variation (green). All energies are in code units, normalized to .
The magnetic energy increases at the expense of the system’s total kinetic energy, keeping the total energy practically constant (black line). There is a redistribution of particle kinetic energy: -particles transfer energy to the protons while driving the wave unstable.
In Figure 3 (right), we show the bulk-energy variation (solid), and the thermal-energy variation , for each species (blue for protons and green for -particles). Overall, protons gain bulk kinetic and thermal energy as the simulation evolves, while -particles lose energy in a non-monotonic way. In the initial stage of the instability, the bulk kinetic energy of the -particles increases, while their thermal energy decreases, resulting in a net reduction in total kinetic energy. At later stages of the evolution, the bulk kinetic energy decreases while the thermal energy slowly increases, but it remains below the initial value.
The growth of magnetic energy is accompanied by a relaxation of the temperature anisotropy of the -particles, as shown in the left panel of Figure 4, and by a decrease in their overall kinetic energy (left panel of Figure 3). At the same time, the protons, initially isotropic and stationary in our reference frame, undergo perpendicular heating and net acceleration along the magnetic field, increasing the proton temperature anisotropy and the bulk and thermal energies and .
Figure 4. Evolution of the moments for each species: protons (blue) and -particles (green). The left panel shows the temperature anisotropy , and the middle panel the parallel drift speed . The right panel shows parallel cuts of the -particle VDF as at four different times. We also superimpose vertical lines indicating the mean velocities at each time.
The drift-velocity evolution for the -particles shows the same initial increase as in Figure 3 (left panel). We also show field-parallel cuts of the -particle VDF at four different times to demonstrate the evolution of the -particle drift. We superimpose the mean velocity from these cuts as vertical lines. We observe that at , the red line shows a more asymmetric distribution, with a tail that is skewed towards the right. This skewness increases the overall bulk velocity of the -particles at that time, showing that this increase in drift speed does not correspond to uniform acceleration of the -particle population. Moments later, at (blue curve), the VDF becomes more symmetric, with a noticeable deformation developing on the left-hand side corresponding to a net reduction in the mean velocity. By the end of the simulation (green curve), the mean velocity has decreased further, resulting in a much more symmetric VDF. The simultaneous decrease in -particle kinetic energy and increase in proton kinetic energy provide direct evidence that the instability mediates the transfer of energy and momentum from the -particle population to the protons through resonant wave–particle interactions, while the wave grows (i.e., ).
Figure 5 shows the evolution of the proton and -particle VDFs between the initial and final stages of the simulation, together with the resonant velocities corresponding to the (Landau; red), (blue), and (green) resonances according to Eq. (15) for the two unstable modes shown in Figure 2. The -particle VDF, initially strongly anisotropic, undergoes substantial relaxation during the instability evolution, becoming broader in the parallel direction and more isotropic at later times, consistent with Figure 4. Simultaneously, the proton VDF develops a slightly enhanced anisotropy with , consistent with the proton heating and anisotropy growth observed in the evolution of the moments. The strongest deformation of the VDFs occurs near the that fulfills the cyclotron resonance, indicating that the nonlinear evolution is governed primarily by resonant wave–particle interactions and associated quasilinear diffusion in velocity space, as shown in the sketch in Figure 1 and supported by the analysis of in Figure 2.
Fast-Magnetosonic/Whistler instability
In this section, we analyze the linear and nonlinear behavior of an unstable plasma in which the resonant instability arises from the relative drift between plasma species. As shown in Table 1 for Case 2, we include a proton beam so that the system now consists of four species: a proton core (), proton beam (), 5% of drifting -particles (), and electrons (). We now study the proton-beam-driven FM/W instability, keeping all species isotropic and introducing a large beam drift (in units of the proton-core Alfvén speed).
In this Case, the -particles, while also drifting, do not contribute to the driving of the instability but may receive energy through wave–particle interactions with the unstable waves. Cases 3 and 4 serve as control cases for Case 2, in which we systematically reduce the -particle population concentration (to 1% in Case 3 and 0% in Case 4) to analyze the energy distribution between species when the concentration of the energy-absorbing species decreases. As in the previous section, we begin with linear predictions of the unstable modes and the contributions of each ion population to the growth rates, then present the PIC simulation results and compare them with the linear predictions.
Linear Predictions
For the FM/W instability, we perform a linear dispersion-relation analysis similar to that presented in § 4.1. Figure 6 shows the parallel-propagating FM/W solutions calculated using PLUME for the parameters of Case 2 in Table 1. We see in the second row that FM/W B, the backward-propagating solution, is damped (i.e., and thus ) for the entire range of investigated wavevectors. The resonances of the proton-core and -particle populations make a significant contribution to the overall , while the proton beam population does not, leading to enhanced damping near . For FM/W A, the forward-propagating solution, the proton beam emits sufficient power to overcome the damping by the proton-core and -particle populations to drive a strong instability near . At these wavenumbers, even though and .
Figure 6. Linear dispersion-relation analysis for Case 2, corresponding to a proton-beam-driven FM/W instability. Format follows Figure 2, but for the FM/W modes and including an additional, fourth, row for the proton-beam emission and absorption rates.
The relative damping or emission contributions from the three ion populations for FM/W A for Cases 2, 3, and 4 are shown in Figure 7. In all three Cases, the proton-beam emission is quantitatively similar in wavevector dependence and amplitude. As we reduce the -particle density from Case 2 to Cases 3 and 4, we see a reduction in the magnitude of the -particle contribution and an increase in the mode’s total growth rate .
Figure 7. Summary of the growth/damping rate for the proton-beam-driven FM/W instability in Cases 2, 3, and 4. The thin curves show the total , while the thicker curves correspond to the proton core, proton beam, and -particle contributions, respectively. Solid curves denote positive contributions to the wave growth (), whereas dashed curves indicate contributions to damping ().
Simulation Results
We run a nonlinear simulation for the parameters of Cases 2, 3, and 4 in Table 1. In all these Cases, the FM/W instability is driven at frequencies and wavenumbers consistent with those predicted in § 5.1 (see Appendix § 10). Similar to Case 1 of the A/IC instability, we expect the electromagnetic energy to increase at the expense of the total kinetic energy of the particles , while the energy of each species may increase or decrease in time due to wave–particle interactions.
We summarize the energy evolution obtained from the simulations for the FM/W instability in Figure 8. Cases 2, 3, and 4 are shown in the left, middle, and right columns. The top panels show the evolution of the variations of the total (black), magnetic (red), kinetic (purple), proton kinetic (blue), beam kinetic (orange), and -particle kinetic As the free energy for this instability is provided by the proton beam, we observe a reduction in the beam kinetic energy , along with increases in the magnetic energy and the core and -particle kinetic energies and . This increase in magnetic energy is translated into a decrease in the total kinetic energy .
Figure 8. Evolution of the energy partition from the simulation for Cases 2, 3, and 4, in the left, middle, and right panels, respectively. energies (green), while the lower panels show the separation of kinetic energy into bulk kinetic and thermal energies (with -particles shown only for Cases 2 and 3).
The comparison of the time evolution of the magnetic energy (red line in the first row) shows that, as the -particle density decreases, the instability becomes more effective, redirecting more of the initial free energy to the fields. Similarly, the higher the -particle density, the larger the fraction of the kinetic energy transferred to them, primarily from the beam. The beam primarily provides energy as bulk kinetic energy , as shown in the lower panels, while the thermal energy and bulk kinetic energy of the other species increase.
Figure 9 shows the evolution in the temperature anisotropies of the ion species (top panels) and the change in their drift speeds (lower panels). In all three Cases, the proton beam develops a modest perpendicular anisotropy while losing drift speed as the instability extracts energy from the beam’s bulk kinetic energy. The proton core remains nearly isotropic in all Cases, although it develops a small positive drift, indicating that a small part of the beam momentum is transferred to the core protons. The -particles develop a slight anisotropy with , while the drift speed increases. These changes depend on the -particle concentration. These results indicate that the -particle population modifies the instability’s nonlinear saturation by providing an additional resonant channel for the redistribution of energy.
Figure 9. Evolution of the anisotropy and drift variation from the simulation for Cases 2, 3, and 4.
Discussion and Conclusions
In this work, we analyze the generation and relaxation of kinetic micro-instabilities in collisionless multi-species plasmas, and describe inter-species coupling through wave–particle interactions. We perform a linear Vlasov–Maxwell mode analysis and PIC simulations of the A/IC and FM/W instabilities in plasmas composed of protons and -particles with plasma parameters similar to those observed in the solar wind.
Our linear and nonlinear analyses reveal a complex interplay between the particle species and the unstable waves in resonant wave–particle interactions. Depending on the VDFs of the involved particle species at the relevant resonance velocities, the diffusive flux of particles in velocity space modifies the shape of the VDFs according to quasilinear theory. As long as the resonant wave–particle interactions lead to a net , the electromagnetic field experiences , corresponding to the unstable growth of the waves. In plasmas with multiple species and even under overall unstable conditions, some species may have while others have . In these cases, the wave–particle interactions effectively transfer energy between the involved particle species while still retaining sufficient energy for the growth of the unstable modes.
In our example Cases, bi-Maxwellian initial VDFs create unstable configurations through a combination of relative drifts and temperature anisotropies. We illustrate with examples in which -particles or a proton beam provide the main free energy to drive instabilities. Our simulations confirm the energetic arguments proposed in § 2, in which the overall energy budget allows for transfer between particle species during the growth of an instability. While our work presents four selected example Cases to showcase inter-species coupling, similar couplings are likely relevant to many other configurations, including minor ion species (e.g., [64]). Indeed, P. S. Moya et al. [42] conclude based on a similar analysis including -particles and ions that the presence of anisotropic and drifting ion populations triggers an energy cascade from low to high frequencies, peaking near their corresponding resonance frequencies, with subsequent heating of ions as the charge-to-mass ratio increases.
The evolution of all ultimately leads to a marginally stable wave–particle equilibrium state for which the overall growth/damping rate vanishes. In this quasilinear saturation state, the of all involved particle species have – locally in velocity space as a result of quasilinear diffusion – adjusted to a configuration in which continuing wave–particle interactions keep for all and thus [60]. These wave–particle equilibria of all species with significant distortion at the relevant resonance speeds are an observable prediction of our study [61, 48]. The analysis of observed with the help of the ALPS code [57, 31], which evaluates the linear Vlasov–Maxwell dispersion relation without the assumption of bi-Maxwellian VDFs, has the potential to reveal the presence of such wave–particle equilibria.
Our analysis makes limiting assumptions on the plasma configurations under consideration. For instance, we assume that the background plasma is homogeneous and in a steady state. The solar wind, however, is an expanding and turbulent medium [39, 7], suggesting that the steady-state and homogeneity assumptions require careful consideration and justification based on arguments of scale separation [1, 22, 56, 21, 65]. Observations of the solar wind reveal, however, that predictions for instabilities based on linear Vlasov–Maxwell theory are often appropriate even in such a nonlinear medium [28, 44, 45]. Moreover, our analysis so far accounts for electrons either as an isotropic Maxwellian (in our linear calculations) or as a charge-neutralizing, massless fluid (in our hybrid simulations). In reality, though, electrons also exhibit a wide variety of non-equilibrium features [14, 46, 20] that can drive kinetic micro-instabilities [59]. It would be worthwhile to extend our study to examine the effects of non-Maxwellian electron background VDFs in both linear theory and full particle-in-cell simulations.
Acknowledgments
This research was supported by the International Space Science Institute (ISSI) in Bern, through ISSI International Team project “Excitation and dissipation of kinetic-scale fluctuations in space plasmas” (ISSI Team project #24-612). We acknowledge the support of ANID Chile through FONDECyT grant No. 1251712 (R.A.L.) and 1240281 (P.S.M.). K.G.K. was partially supported by NASA grant 80NSSC24K10171. A.T. was partly supported by U.S. National Science Foundation PHY-2010189 and the U.S. Department of Energy (DOE) Fusion Energy Sciences Postdoctoral Research Program, administered by the Oak Ridge Institute for Science and Education and Oak Ridge Associated Universities under DOE contract DE-SC0014664. D.V. is supported by STFC Consolidated Grant ST/W001004/1. High-performance computing resources were provided by Anvil at Purdue University’s Rosen Center for Advanced Computing via allocation PHY250393 from the U.S. National Science Foundation (NSF) Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program. Powered@NLHPC: This research was partially supported by the supercomputing infrastructure of the NLHPC (CCSS210001). The ALPS project received support from UCL’s Advanced Research Computing Centre through the Open Source Software Sustainability Funding scheme.
Appendix
Energy conservation
Since we are interested in the energy partition among plasma species, we must ensure that the total energy is conserved in our simulations. We take Case 1 from Table 1 and run three simulations with varying numbers of particles per cell per species. We use (dotted lines), 50,000 (dashed), and 100,000 (solid). Figure 10 shows the evolution of the variations in the total energy (black), magnetic energy (red), and total kinetic energy (purple) depending on .
Figure 10. Energy conservation test for Case 1. Evolution of the variations of total energy (black), magnetic energy (red), and total kinetic energy (purple) for three different numbers of particles per cell per species: Solid lines for , dashed lines for , and dotted lines for .
Increasing ensures better energy conservation. The main differences are observed in the total kinetic energy, while the magnetic energy is almost independent of . The overall trend in energy evolution is maintained, even when energy conservation is poor.
Nonlinear Spectra and Dispersion Relation
In Figure 11, we compare the dispersion relation from the simulation with the predictions by linear theory. We perform the Fast Fourier Transform (FFT) of the complete temporal series for the magnetic field fluctuations of our simulation data as . The first and second quadrants correspond to left-hand circularly polarized modes in forward and backward propagation, respectively, whereas the third and fourth quadrants correspond to right-hand circularly polarized modes in forward and backward propagation, respectively. We normalize Case 1 to the maximum of its power spectrum. For Cases 2, 3, and 4, we normalize all spectra to their common maximum, which corresponds to the maximum of Case 4. This ordering is consistent with the linear theory prediction in Figure 7.
Figure 11. Dispersion relation obtained from applied to the simulation data for all four Cases. The first and second quadrants correspond to left-hand circularly polarized modes in forward and backward propagation, respectively, whereas the third and fourth quadrants correspond to right-hand circularly polarized modes in forward and backward propagation, respectively.
Over the FFT results from our simulations, we superimpose the linear dispersion-relation solutions as black dashed lines on each spectrum, using the initial-condition parameters for each Case. The spectra agree well with the predicted linear normal modes of the plasma, as well as the unstable modes that develop throughout the simulation. For Case 1, the unstable modes lie in the first and second quadrants, consistent with the A/IC instability. Cases 2, 3, and 4 show the unstable branch in the third quadrant, consistent with the FM/W instability.
The linear prediction for Case 1 differs from the simulation spectrum slightly in that the real frequency is below the simulation spectrum for the A/IC branch in the first quadrant. This difference arises because the linear-theory prediction is a snapshot based on the initial condition, whereas the simulation spectra capture the entire simulation time window, and, as seen in Figure 5, the VDFs evolve significantly during the simulation.
Figure 5. Contours of the proton (left panels) and -particle (right panels) VDFs in the plane at the initial time (), black dashed contours) and final simulation time (), red contours) for Case 1. The vertical dashed lines indicate the resonant velocities associated with the (Landau), , and resonances for the two unstable modes: A/IC A (left panels) and A/IC B (right panels).
To corroborate this claim, we use the ALPS solver [57] to compute the linear dispersion relation using a Chebyshev polynomial representation of the actual proton and -particle VDFs obtained from the simulation at a representative time of ; details of the Chebyshev representation are given by [31]. By this simulation time, the instability has mostly saturated, and the system is close to a steady wave–particle equilibrium state, as we show in Figure 3. We add this solution as a white dashed line to the Case 1 spectrum, which we plot up to the wavenumber where to exclude strongly damped modes from our visualization. The observed late-time simulation spectrum for Case 1 is consistent with the dispersion relation obtained from ALPS. The agreement between the two is most notable at low wavenumbers until the wavenumber at which the wave power begins to decrease again. Although the dispersion relations deviate from the simulation spectrum at larger wavenumbers, the difference in the wavenumber ranges that satisfy the conditions for forward- and backward-propagating A/IC waves can also explain the reflectional asymmetry of the observed spectra.
References
- [1]Alexandrova, O., Chen, C. H. K., Sorriso-Valvo, L., Horbury, T. S., & Bale, S. D. 2013, SSRv, 178, 101, doi: 10.1007/s11214-013-0004-8DOI
- [2]Asbridge, J. R., Bame, S. J., Feldman, W. C., & Montgomery, M. D. 1976, J. Geophys. Res., 81, 2719, doi: 10.1029/JA081i016p02719DOI
- [3]Bale, S. D., Kasper, J. C., Howes, G. G., et al. 2009, PhRvL, 103, 211101, doi: 10.1103/PhysRevLett.103.211101DOI
- [4]Bourouaine, S., Verscharen, D., Chandran, B. D. G., Maruca, B. A., & Kasper, J. C. 2013, ApJL, 777, L3, doi: 10.1088/2041-8205/777/1/L3DOI
- [5]Bowers, K. J., Albright, B. J., Yin, L., Bergen, B., & Kwan, T. J. T. 2008, Physics of Plasmas, 15, 055703, doi: 10.1063/1.2840133DOI
- [6]Brown, C. R., Howes, G. G., Klein, K. G., & TenBarge, J. M. 2026, Physics of Plasmas, 33, 072101, doi: 10.1063/5.0335826DOI
- [7]Bruno, R., & Carbone, V. 2013, Living Reviews in Solar Physics, 10, 2, doi: 10.12942/lrsp-2013-2DOI
- [8]Bruno, R., De Marco, R., D’Amicis, R., et al. 2024, ApJ, 969, 106, doi: 10.3847/1538-4357/ad47b3DOI
- [9]Daughton, W., & Gary, S. P. 1998, J. Geophys. Res., 103, 20613, doi: 10.1029/98JA01385DOI
- [10]De Marco, R., Bruno, R., Jagarlamudi, V. K., et al. 2023, A&A, 669, A108, doi: 10.1051/0004-6361/202243719
- [11]Espinoza, C. M., Stepanova, M., Moya, P. S., Antonova, E. E., & Valdivia, J. A. 2018, Geophysical Research Letters, 45, 6362, doi: https://doi.org/10.1029/2018GL078631
- [13]Feldman, W. C., Asbridge, J. R., Bame, S. J., & Montgomery, M. D. 1973, J. Geophys. Res., 78, 2017, doi: 10.1029/JA078i013p02017DOI
- [14]Feldman, W. C., Asbridge, J. R., Bame, S. J., Montgomery, M. D., & Gary, S. P. 1975, J. Geophys. Res., 80, 4181, doi: 10.1029/JA080i031p04181DOI
- [15]Gary, S. P. 1991, SSRv, 56, 373, doi: 10.1007/BF00196632DOI
- [16]Gary, S. P. 1993, Theory of Space Plasma MicroinstabilitiesDOI
- [17]Gary, S. P., Yin, L., Winske, D., & Reisenfeld, D. B. 2000, Geophys. Res. Lett., 27, 1355, doi: 10.1029/2000GL000019DOI
- [18]Goldstein, B. E., Neugebauer, M., Zhang, L. D., & Gary, S. P. 2000, Geophys. Res. Lett., 27, 53, doi: 10.1029/1999GL003637DOI
- [19]Goodrich, C. C., & Lazarus, A. J. 1976, J. Geophys. Res., 81, 2750, doi: 10.1029/JA081i016p02750DOI
- [20]Halekas, J. S., Whittlesey, P., Larson, D. E., et al. 2020, ApJS, 246, 22, doi: 10.3847/1538-4365/ab4cecDOI
- [21]Hellinger, P., Matteini, L., Landi, S., et al. 2019, ApJ, 883, 178, doi: 10.3847/1538-4357/ab3e01DOI
- [22]Hellinger, P., Matteini, L., Landi, S., et al. 2015, ApJL, 811, L32, doi: 10.1088/2041-8205/811/2/L32DOI
- [23]Hellinger, P., Trávníček, P., Kasper, J. C., & Lazarus, A. J. 2006, Geophys. Res. Lett., 33, L09101, doi: 10.1029/2006GL025925DOI
- [24]Hundhausen, A. J., Bame, S. J., & Ness, N. F. 1967, J. Geophys. Res., 72, 5265, doi: 10.1029/JZ072i021p05265DOI
- [25]Jian, L. K., Moya, P. S., Viñas, A. F., & Stevens, M. 2016, AIP Conference Proceedings, 1720, 040007, doi: 10.1063/1.4943818DOI
- [26]Kennel, C. F., & Engelmann, F. 1966, Physics of Fluids, 9, 2377, doi: 10.1063/1.1761629DOI
- [27]Kennel, C. F., & Wong, H. V. 1967, Journal of Plasma Physics, 1, 75, doi: 10.1017/S0022377800000310X
- [28]Klein, K. G., Alterman, B. L., Stevens, M. L., Vech, D., & Kasper, J. C. 2018, PhRvL, 120, 205102, doi: 10.1103/PhysRevLett.120.205102DOI
- [29]Klein, K. G., & Howes, G. G. 2025, kgklein/PLUME: Zenodo release, v1.0.1 Zenodo, doi: 10.5281/zenodo.15215514
- [30]Klein, K. G., Howes, G. G., & Brown, C. R. 2025, Research Notes of the AAS, 9, 102, doi: 10.3847/2515-5172/add1c2DOI
- [31]Klein, K. G., & Verscharen, D. 2025, Physics of Plasmas, 32, 092104, doi: 10.1063/5.0286477DOI
- [34]Marsch, E. 2006, Living Reviews in Solar Physics, 3, 1, doi: 10.12942/lrsp-2006-1DOI
- [35]Marsch, E., & Livi, S. 1987, J. Geophys. Res., 92, 7263, doi: 10.1029/JA092iA07p07263DOI
- [36]Marsch, E., Rosenbauer, H., Schwenn, R., Muehlhaeuser, K.-H., & Neubauer, F. M. 1982a, J. Geophys. Res., 87, 35, doi: 10.1029/JA087iA01p00035DOI
- [37]Marsch, E., Schwenn, R., Rosenbauer, H., et al. 1982b, J. Geophys. Res., 87, 52, doi: 10.1029/JA087iA01p00052DOI
- [38]Maruca, B. A., Kasper, J. C., & Gary, S. P. 2012, ApJ, 748, 137, doi: 10.1088/0004-637X/748/2/137DOI
- [39]Matteini, L., Hellinger, P., Landi, S., Trávníček, P. M., & Velli, M. 2012, SSRv, 172, 373, doi: 10.1007/s11214-011-9774-zDOI
- [40]Matteini, L., Landi, S., Hellinger, P., et al. 2007, Geophys. Res. Lett., 34, L20105, doi: 10.1029/2007GL030920DOI
- [41]Montgomery, M. D., Gary, S. P., Feldman, W. C., & Forslund, D. W. 1976, J. Geophys. Res., 81, 2743, doi: 10.1029/JA081i016p02743DOI
- [44]Opie, S., Verscharen, D., Chen, C. H. K., Owen, C. J., & Isenberg, P. A. 2022, ApJ, 941, 176, doi: 10.3847/1538-4357/ac982f
- [45]Opie, S., Verscharen, D., Chen, C. H. K., et al. 2024, Journal of Plasma Physics, 90, 905900602, doi: 10.1017/S0022377824001375DOI
- [46]Pilipp, W. G., Miggenrieder, H., Montgomery, M. D., et al. 1987, J. Geophys. Res., 92, 1075, doi: 10.1029/JA092iA02p01075DOI
- [47]Podesta, J. J., & Gary, S. P. 2011, ApJ, 742, 41, doi: 10.1088/0004-637X/742/1/41DOI
- [48]Ran, H., Verscharen, D., Coburn, J., et al. 2026, ApJ, 1003, 124, doi: 10.3847/1538-4357/ae60fbDOI
- [49]Reisenfeld, D. B., Gary, S. P., Gosling, J. T., et al. 2001, J. Geophys. Res., 106, 5693, doi: 10.1029/2000JA000317DOI
- [50]Rowlands, J., Shapiro, V. D., & Shevchenko, V. I. 1966, Soviet Journal of Experimental and Theoretical Physics, 23, 651
- [51]Shapiro, V. D., & Shevchenko, V. I. 1966, in Plasma Physics and Controlled Thermonuclear Fusion, ed. K. D. Sinel’Nikov, 149
- [52]Stix, T. H. 1992, Waves in plasmas (American Institute of Physics)
- [54]Verniero, J. L., Larson, D. E., Livi, R., et al. 2020, ApJS, 248, 5, doi: 10.3847/1538-4365/ab86afDOI
- [55]Verscharen, D., Bourouaine, S., & Chandran, B. D. G. 2013, ApJ, 773, 163, doi: 10.1088/0004-637X/773/2/163DOI
- [56]Verscharen, D., Chandran, B. D. G., Klein, K. G., & Quataert, E. 2016, ApJ, 831, 128, doi: 10.3847/0004-637X/831/2/128DOI
- [57]Verscharen, D., Klein, K. G., Chandran, B. D. G., et al. 2018, Journal of Plasma Physics, 84, 905840403, doi: 10.1017/S0022377818000739DOI
- [58]Verscharen, D., Klein, K. G., & Maruca, B. A. 2019, Living Reviews in Solar Physics, 16, 5, doi: 10.1007/s41116-019-0021-0DOI
- [59]Verscharen, D., Chandran, B. D. G., Boella, E., et al. 2022, Frontiers in Astronomy and Space Sciences, 9, 951628, doi: 10.3389/fspas.2022.951628DOI
- [60]Villarroel-Sepúlveda, N., Verscharen, D., Moya, P. S., López, R. A., & Klein, K. G. 2026, ApJ, 1002, 49, doi: 10.3847/1538-4357/ae5639
- [61]Walters, J., Klein, K. G., Lichko, E., et al. 2023, ApJ, 955, 97, doi: 10.3847/1538-4357/acf1faDOI
- [62]Wicks, R. T., Alexander, R. L., Stevens, M., et al. 2016, The Astrophysical Journal, 819, 6, doi: 10.3847/0004-637X/819/1/6DOI
- [63]Yoon, P. H. 2017, Reviews of Modern Plasma Physics, 1, doi: 10.1007/s41614-017-0006-1DOI
- [64]Zhang, M. F., Kunz, M. W., Squire, J., & Klein, K. G. 2025, ApJ, 979, 121, doi: 10.3847/1538-4357/ad95fcDOI
- [65]Zhu, X., Verscharen, D., He, J., Maruca, B. A., & Owen, C. J. 2023, ApJ, 956, 66, doi: 10.3847/1538-4357/aced03DOI










