---
title: Electron Parallel Nonlinearity (EPN)
url: https://www.emergentmind.com/topics/electron-parallel-nonlinearity-epn
type: topic
---

# Electron Parallel Nonlinearity (EPN)

Electron Parallel Nonlinearity (EPN) most explicitly denotes the nonlinear fluctuation-induced advection of the electron distribution in parallel phase space, represented in gyrokinetics by terms such as $\dot v_{\parallel 1}\,\partial \delta f/\partial v_\parallel$ or $\dot p_{\parallel 1}\,\partial \delta f_{p_\parallel}/\partial p_\parallel$ [2509.21680]. Several adjacent literatures do not use the label explicitly, yet treat closely related electron-mediated phenomena: trapped-electron modification of Langmuir-wave response [1208.3864], heating and distribution reshaping by field-aligned electric fields in reconnection [2412.10188], nonlocal parallel response operators for heat flow, friction, and viscosity [1906.09201, 1906.09225], and hot-electron two-beam coupling through interference-induced temperature gratings [2008.12824]. This suggests that EPN is best understood as a family of nonlinear or nonlocal electron responses organized by motion, acceleration, or transport along a distinguished parallel direction.

## 1. Formal gyrokinetic definition

In the most direct usage, EPN is a specific nonlinear term in the electron drift-kinetic or gyrokinetic equation. In the parallel-velocity formulation, the perturbed distribution satisfies
\[
\frac{\partial \delta f}{\partial t} + \mathbf{v}_G \cdot \nabla \delta f + \dot v_\parallel \frac{\partial \delta f}{\partial v_\parallel}
= - \mathbf{v}_{G1}\cdot \nabla f_M - \dot{\mathcal E}_k \frac{\partial f_M}{\partial \mathcal E_k},
\]
and the EPN is the nonlinear contribution
\[
\dot v_{\parallel 1}\,\frac{\partial \delta f}{\partial v_\parallel}.
\]
In GEM, which uses the parallel canonical momentum
\[
p_\parallel = v_\parallel + \frac{q}{m}A_\parallel,
\]
the corresponding term is
\[
\dot p_{\parallel 1}\,\frac{\partial \delta f_{p_\parallel}}{\partial p_\parallel}.
\]
Physically, this is nonlinear coupling between fluctuation-induced parallel acceleration and fine structure in electron parallel velocity space [2509.21680].

This term is usually dropped under the Frieman–Chen ordering because it is formally $O(\delta^3)$, whereas the standard retained nonlinearities are $O(\delta^2)$. The 2025 GEM study is important because it shows that formal ordering smallness does not guarantee dynamical insignificance in electromagnetic turbulence. The same paper also makes clear that implementation is not limited to modifying the particle equation of motion: in a split-weight PIC formulation, EPN-consistent higher-order terms must also be retained in the electron weight equation and in the vorticity equation. A partially linearized treatment neglects precisely these higher-order contributions, including the distinction between $p_\parallel$ and $v_\parallel$ in the nonlinear terms [2509.21680].

A common misconception is that EPN is simply another name for parallel conductivity or parallel streaming. The gyrokinetic definition is narrower: it is a nonlinear phase-space advection term, not a constitutive coefficient. That distinction becomes important when comparing EPN to nonlocal closure models, which encode parallel electron response but do not evolve the full nonlinear electron distribution.

## 2. Electromagnetic turbulence and kinetic ballooning saturation

The clearest demonstration of EPN as a turbulence-saturation mechanism appears in the GEM study of the Cyclone Base Case above the kinetic ballooning mode threshold. For the global electromagnetic setup, the chosen primary mode is $n=20$, corresponding to
\[
k_y \rho_i = 0.31,
\]
and a linear $\beta$ scan shows transition from ITG to KBM at about
\[
\beta \gtrsim 0.03.
\]
At $\beta=0.04$, the linear $n=20$ mode has
\[
\omega_r a/c_s = 0.94, \qquad \gamma_{\rm lin} a/c_s = 0.34.
\]
The key nonlinear result is that a simulation without full EPN undergoes runaway early in the nonlinear phase, whereas the simulation including EPN saturates at a low initial amplitude and exhibits a strong reduction of heat transport relative to the runaway case [2509.21680].

The onset of nonlinear regulation is unusually early. With EPN, deviation from linear growth occurs at
\[
t c_s/a \approx 5, \qquad e\phi/T_0 = 0.0013,
\]
whereas in the electrostatic ITG comparison the corresponding deviation occurs only near
\[
t c_s/a \approx 30, \qquad e\phi/T_0 = 0.019.
\]
The same study states that the EPN-inclusive KBM case saturates at an initial amplitude two orders of magnitude below that of the electrostatic case. By contrast, in the electrostatic regime at $\beta=4\times 10^{-4}$, including EPN does not substantially alter the usual zonal-flow-mediated saturation picture. The paper therefore argues that EPN is not generically strong in all turbulence, but becomes decisive in electromagnetic KBM regimes [2509.21680].

The proposed mechanism depends on magnetic flutter. For the $n=20$ KBM, the linear simulation gives
\[
\frac{v_u A_\parallel}{\phi} = 0.14,
\]
which might seem modest, but for electrons
\[
\frac{v_{Te} A_\parallel}{\phi} = 5.9.
\]
Thus electron radial displacement is dominated by magnetic flutter rather than by $E\times B$ drift. The maximum field-line width near the mode location is reported as
\[
\Delta r/\rho_i \approx 0.2,
\]
while the characteristic radial fine-structure scale of the linear eigenmode is estimated as
\[
\delta_{\rm lin}\approx 0.49 \rho_i.
\]
Because the flutter-induced radial displacement depends on $v_\parallel$, radial fine structure is converted into velocity-space fine structure. The paper expresses this by differentiating the phase-space constancy relation
\[
f({\bf Z},t)=f({\bf Z}-\Delta {\bf Z}, t-\tau)
\]
with respect to $v_\parallel$, producing a term proportional to
\[
-\frac{\partial \Delta x}{\partial v_\parallel}\frac{\partial f}{\partial x}.
\]
The resulting estimate
\[
\frac{\partial \delta f}{\partial v_\parallel} \sim \tau \frac{\delta B_r}{B_0}\,\frac{\delta f}{\delta_{\rm lin}}
\]
shows how a large $\partial/\partial v_\parallel$ can be generated from pre-existing radial fine structure, thereby amplifying the EPN term [2509.21680].

The same work gives a plausibility-level nonlinear damping estimate,
\[
\gamma_{\rm nl} \sim \frac{e}{m_e}\omega_r A_\parallel \, \tau \frac{\delta B_r}{B_0} \, \frac{1}{\delta_{\rm lin}},
\]
with
\[
\frac{\gamma_{\rm nl}}{\gamma_{\rm lin}} = 0.018
\]
at $t c_s/a=5$, increasing to roughly
\[
\frac{\gamma_{\rm nl}}{\gamma_{\rm lin}} \sim 0.16
\]
by $t c_s/a=10$. A mass-ratio scan over
\[
m_i/m_e = 900,\ 1800,\ 3600,\ 7200,\ 14400
\]
finds an approximate scaling
\[
A_\parallel \propto m_e^{0.75},
\]
which the authors interpret as strong evidence for an electron-streaming-based phase-space mechanism. The paper is careful, however, to note that a quantitative confirmation would require direct diagnostics of the evolving velocity-space structure, and its model neglects $\delta B_\parallel$ [2509.21680].

## 3. Trapping nonlinearity in Langmuir waves

In laser-plasma interaction, the phenomenon most closely corresponding to EPN is electron trapping nonlinearity in Langmuir waves. Here the wave acts on electrons with parallel velocity near the phase velocity
\[
v_p = \frac{\omega}{k},
\]
and trapped electrons execute bounce orbits inside the electrostatic potential troughs. The resulting flattening of the velocity distribution near $v_p$ reduces Landau damping and contributes to nonlinear frequency shift. The paper emphasizes a crucial distinction: it is not enough for electrons to lie instantaneously inside the separatrix; the distribution must have time to flatten, which takes of order a bounce period [1208.3864].

For a Langmuir wave with density perturbation amplitude
\[
\delta N \equiv \frac{\delta n}{n_e},
\]
the bounce frequency and bounce period are
\[
\omega_B = \omega_{pe}\,\delta N^{1/2}, \qquad
\tau_B = \frac{2\pi}{\omega_B} = 2\pi \omega_{pe}^{-1}\delta N^{-1/2}.
\]
The paper introduces the bounce number
\[
N_B \equiv \frac{1}{\nu_d \tau_B},
\]
where $\nu_d$ is the detrapping rate. The practical onset criterion is
\[
N_B \sim 1.
\]
For a time-dependent wave, the dynamic bounce number is
\[
N_B^{\mathrm{dyn}}(t) = \int_0^t \frac{dt'}{\tau_B(t')},
\]
and the simulations show that significant trapping response begins near
\[
N_B^{\mathrm{dyn}} \approx 0.5.
\]
The authors explicitly stress that this is not a hard threshold; trapping nonlinearity increases continuously with $N_B$ [1208.3864].

The paper treats two classes of detrapping process. Convective loss arises because a finite wavepacket has finite longitudinal and transverse extent. For transverse side loss,
\[
\nu_{d,\mathrm{sl}} = K_{\mathrm{sl}} \frac{v_{Te}}{L_\perp},
\]
with
\[
K_{\mathrm{sl}}=
\begin{cases}
1.02, & \text{2D},\\
2.08, & \text{3D},
\end{cases}
\]
and the corresponding threshold amplitude is
\[
\delta N_{\mathrm{sl}} = \left[ 2\pi K_{\mathrm{sl}} \frac{\lambda_{De}}{L_\perp} \right]^2.
\]
Collisional detrapping is treated with a Fokker–Planck operator that includes electron-ion pitch-angle scattering, electron-electron pitch-angle scattering, and electron-electron parallel drag and diffusion in one framework. The collisional bounce number scales as
\[
N_{B,c} = \left(\frac{\delta N}{\delta N_c}\right)^{3/2},
\]
so collisions become less effective at detrapping as the wave amplitude increases [1208.3864].

This threshold theory is supported by 2D-2V Vlasov simulations with the code *loki*. In the driven-wave simulations, trapping nonlinearity increases continuously with $N_B$ and is significant for
\[
N_B \sim 1.
\]
For a single NIF-like $f/8$ speckle, the paper concludes that convective transverse loss is usually the most potent detrapping process, and under typical inner-cone conditions local reflectivities around $3\%$ are estimated to produce significant trapping effects. This makes trapped-electron EPN a quantitatively accessible effect in stimulated Raman backscatter, rather than a merely asymptotic kinetic correction [1208.3864].

## 4. Field-aligned electric fields, beam formation, and reconnection heating

A distinct but closely related usage arises in collisionless magnetic reconnection, where nonlinear electron dynamics are driven by magnetic-field-aligned electric fields. In a statistical study of 140 MMS magnetotail reconnection outflows, the acceleration potential
\[
e\Phi_\parallel(x) = -\int_\infty^x E_\parallel dl
\]
is inferred from the morphology of electron velocity distribution functions rather than from direct field-line integration. The paper reports that heating by $E_\parallel$ in the reconnection outflow can reach up to ten times the inflow electron temperature, and that the potential scales approximately as
\[
e\Phi_\parallel \propto T_{e\infty}^{1/2}V_{Ae\infty},
\qquad
\frac{e\Phi_\parallel}{T_{e\infty}} \propto \beta_{e\infty}^{-1/2}.
\]
The authors interpret this as evidence that $E_\parallel$ is set by the need to maintain quasi-neutrality and confine or trap electrons in the ion diffusion region [2412.10188].

The inferred nonlinear pathway is beam acceleration followed by wave-driven thermalization. The paper describes eVDFs that consist of a cold beam accelerated toward the X-line and a thermalized hot population. These non-equilibrium distributions can drive Buneman and electron two-stream instabilities, whose electrostatic waves scatter the beam and reshape the eVDF into a flat-top distribution. This links EPN, in the reconnection context, to field-aligned beam formation, trapping, pressure anisotropy, and phase-space flattening rather than to a single fluid coefficient [2412.10188].

The inference procedure is technically specific. The field-aligned part of the eVDF is fit with a generalized $(r,q)$ distribution,
\[
f_{(r, q)} = f_0\left [1 + \left (\frac{v_\parallel^2}{\xi v_{te,\parallel}^2} \right )^{r+1}\right ]^{-q},
\]
using a Levenberg–Marquardt algorithm minimizing the reduced weighted $\tilde{\chi}^2$. Only energies satisfying
\[
E_e\leq 5 T_{e\parallel}
\]
are included, and pitch-angle selection is restricted to within $9^\circ$ of $0^\circ$ and $180^\circ$ in the spacecraft frame. The “knee” of the plateau is defined using the threshold
\[
\varepsilon=1/e,
\]
which yields
\[
v_\Phi = \left ( \varepsilon^{-1 / q} - 1\right )^{1/2(r+1)}\xi^{1/2}v_{te,\parallel},
\qquad
e\Phi_\parallel = \frac{m_e v_\Phi^2}{2e}.
\]
This is therefore an inferred potential from nonlinear eVDF morphology, not a direct electrostatic potential measurement [2412.10188].

The paper also uses electron force balance along the field,
\[
-eE_\parallel = \nabla_\parallel T_{e\parallel} + T_{e\parallel}\nabla_\parallel \ln n_e + \left ( T_{e\bot} - T_{e\parallel} \right )\nabla_\parallel \ln B,
\]
to interpret the spatial structure of the potential. In the case study, the observed drop from current-sheet center to edge is
\[
e\Delta \Phi_\parallel^{obs}\approx 220~\mathrm{eV},
\]
decomposed approximately into
\[
e\Delta \Phi_{\parallel T}\approx 10~\mathrm{eV},\quad
e\Delta \Phi_{\parallel B}\approx 50~\mathrm{eV},\quad
e\Delta \Phi_{\parallel n}\approx 160~\mathrm{eV},
\]
so the density term dominates. The observations favor an ambipolar, quasi-neutrality-enforcing interpretation, although the paper explicitly does not separate electrostatic and inductive contributions to $E_\parallel$ directly. It further argues that as $\beta_{e\infty}$ increases, $E_\parallel$ becomes increasingly important to the ion-to-electron energy partition in reconnection [2412.10188].

## 5. Nonlocal closure formulations for parallel electron response

A different branch of the literature addresses EPN at the level of fluid closure rather than full nonlinear phase-space evolution. The papers on electron parallel closures formulate the parallel heat flux, friction force, and viscosity as kernel-weighted integrals of thermodynamic drives along the magnetic field line. The central claim is that when collisionality decreases, electrons free-stream along the field, so the response at one point depends on remote locations measured in units of the electron mean free path rather than on purely local gradients. The closed quantities are the parallel heat flux density $h_\parallel$, the parallel friction force density $R_\parallel$, and the parallel viscosity $\pi_\parallel$, driven by the parallel temperature gradient, the relative electron-ion parallel flow, and the parallel rate-of-strain tensor [1906.09201].

In this formulation, the response is spatially nonlocal in a normalized parallel coordinate. The fitted kernels provide a compact surrogate for a very large moment system and are written in the universal form
\[
K_{AB}(\eta) = -\bigl[d + a e^{-b\eta^{c}}\bigr] \ln\!\left[1-\alpha e^{-\beta \eta^\gamma}\right].
\]
For $Z=1$, the fits were constructed to reproduce the 6400-moment solution in the converged regime together with the collisionless asymptotic behavior. The same work reports a concrete failure mode of local closure: Braginskii’s local closure for the temperature-gradient-driven heat flow has about $16\%$ error at
\[
k\sim 0.1,\qquad \epsilon \equiv \frac{\lambda_{\mathrm{mfp}}}{\lambda}\sim 0.02,
\]
and about $100\%$ error at
\[
k\sim 0.4,\qquad \epsilon\sim 0.06.
\]
For reduced models that require an EPN-aware representation of electron parallel transport, this is precisely the regime where local conductivity becomes quantitatively unreliable [1906.09201].

The extension to arbitrary effective charge is provided by the companion study for various ion charge numbers. It keeps the same analytic kernel form and tabulates parameters for
\[
1\le Z\le 10,
\]
with smooth $Z$ dependence suitable for interpolation to noninteger $Z_{\rm eff}$. The recommended interpolation is linear in the parameters $a,b,c,d,\beta,\gamma$, together with asymptotic constraints for selected kernels so that the correct collisionless limit is preserved. Benchmark errors of the fitted-kernel closures relative to the 6400-moment closures are reported as typically a few percent or less over the convergent regime, with worst cases under $5\%$ [1906.09225].

These closures are directly relevant to EPN-oriented fluid or extended-fluid modeling, but they should not be confused with a complete nonlinear kinetic theory. The arbitrary-collisionality closure paper explicitly states that it is a linearized, moment-based, nonlocal closure model. It captures nonlocal parallel electron transport, finite-collisionality corrections, and cross-coupling between thermal gradients, relative flows, and strain, but it does not provide a fully nonlinear electron kinetic closure, nonlinear phase-space trapping, nonlinear Landau damping, or self-consistent nonlinear evolution of the electron distribution function. A plausible implication is that these kernels are best viewed as closure ingredients for EPN-aware reduced models, rather than as a full theory of EPN in the strict nonlinear kinetic sense [1906.09201, 1906.09225].

## 6. Hot-electron beam coupling and conceptual extensions

Outside plasma transport and reconnection, a related electron-mediated nonlinearity appears in epsilon-near-zero ITO. The two-beam coupling study does not use the term EPN explicitly, but it is highly relevant if EPN is understood as an electron-mediated nonlinear interaction between two beams in a parallel or crossing-beam geometry. The physical mechanism is hot-electron nonlinearity: optical absorption heats the conduction electrons, the electron temperature $T_e$ rises during the pulse, and the optical properties depend on $T_e$ through temperature-dependent Drude parameters. The temperature dynamics are described with a two-temperature model,
\[
C_e \frac{dT_e}{dt} = -g_{el}(T_e-T_l)+S(t), \qquad
C_l \frac{dT_l}{dt} = g_{el}(T_e-T_l),
\]
and the dielectric function is written as
\[
\varepsilon(\omega,T_e) = \varepsilon_\infty - \frac{Ne^2}{\varepsilon_0[m^*(T_e)\omega^2 +ie\omega/\mu(T_e)]}.
\]
The mechanism is therefore a temperature-driven Drude-parameter renormalization involving both $m^*(T_e)$ and $\mu(T_e)$, not an instantaneous bound-electronic Kerr response [2008.12824].

The central two-beam result is that pump and probe interference creates a spatially periodic heating term and hence a temperature grating,
\[
T_e(r,t) = T_e^s(t) + \left[T_e^g(t)e^{ik\cdot r}+c.c.\right].
\]
Because $\varepsilon$ depends on $T_e$, the medium develops a dielectric grating that diffracts pump light into the probe direction. The output field in the probe channel contains both the modified probe and the diffracted pump contribution:
\[
A^{\mathrm{trans}}_p(t) = t^s(t) A_p(t-t_d) + t^g(t) A_e(t)e^{i\delta t},
\]
\[
A^{\mathrm{refl}}_p(t) = r^s(t) A_p(t-t_d) + r^g(t) A_e(t)e^{i\delta t}.
\]
This is a beam-coupled nonlinear response proportional to a field product rather than merely to pump intensity alone, and it produces polarization-sensitive transients that can mimic an instantaneous $\chi^{(3)}$ unless the grating mechanism is taken into account [2008.12824].

The experiment uses a $316\ \mathrm{nm}$ ITO film on fused silica, near the ENZ wavelength of $1240\ \mathrm{nm}$, with $67\ \mathrm{fs}$ pulses at $1\ \mathrm{MHz}$ and a pump–probe angle of $10^\circ$. The observed smooth hot-electron response decays with a characteristic time around $82$–$85\ \mathrm{fs}$, while the coherent two-beam-coupling signal appears only for co-polarized beams during temporal overlap. The authors fit effective nonlinear coefficients
\[
\beta = -350\ \mathrm{cm/GW}, \qquad n_2 = 0.013\ \mathrm{cm^2/GW},
\]
and show that in the continuous-wave nondegenerate limit the delayed response renormalizes them as
\[
n_2^\prime = n_2 \left(1+\frac{1}{1+\delta^2 \tau^2} \right)- \beta \frac{c}{\omega} \frac{ \delta\tau}{1+\delta^2\tau^2},
\]
\[
\beta^\prime =  \beta \left(1+\frac{1}{1+\delta^2 \tau^2} \right)+n_2 \frac{\omega}{c} \frac{\delta \tau}{1+\delta^2\tau^2}.
\]
At degeneracy, each nonlinear coefficient is enhanced by a factor of $2$; away from degeneracy, refractive and absorptive nonlinearities mix. The paper also shows that chirp values of $-2600\ \mathrm{fs^2}$, $-700\ \mathrm{fs^2}$, $960\ \mathrm{fs^2}$, and $3600\ \mathrm{fs^2}$ modify the polarization-anisotropy transient because the coupling depends on $\delta\tau$ [2008.12824].

This optical literature defines neither a gyrokinetic EPN operator nor a Langmuir-wave trapping threshold. Nevertheless, it provides a concrete example of electron-mediated nonlinear coupling in which the internal state variable is electron temperature, the response is delayed rather than instantaneous, and beam interaction is mediated by an interference-induced grating. A plausible implication is that the broader EPN label can encompass not only parallel kinetic transport and reconnection physics, but also hot-electron beam coupling whenever the dominant nonlinearity is carried by electron dynamics along a privileged propagation geometry [2008.12824].

Source: https://www.emergentmind.com/topics/electron-parallel-nonlinearity-epn