Oscillation-Center Quasilinear Theory

• Oscillation-center quasilinear theory is a Hamiltonian approach that decouples fast oscillations from slow resonant transport in wave–particle interactions in plasmas.
• The method uses a near-identity Lie-transform to derive a quasilinear diffusion equation with an identical diffusion tensor in both oscillation-center and particle spaces.
• It extends to broader Hamiltonian systems and magnetized plasmas, incorporating collision effects, ponderomotive corrections, and invariant-space transport.

Oscillation-center quasilinear theory is a Hamiltonian formulation of weak wave–particle interaction in which the rapid oscillatory motion driven by small-amplitude fields is removed by a near-identity transformation, leaving a slow evolution equation for an oscillation-center distribution. In the unmagnetized electrostatic setting, the formulation recovers Dewar’s 1973 result and yields a quasilinear diffusion equation in oscillation-center variables with the same diffusion tensor as in particle space (Brizard, 22 Sep 2025). In broader Hamiltonian treatments, the same oscillation-center viewpoint supports a dressed distribution, a Fokker–Planck operator, ponderomotive corrections, and, in more general settings, collision and wave-kinetic terms (Dodin, 2022).

1. Historical placement and conceptual content

Oscillation-center quasilinear theory emerged as a reformulation of quasilinear transport in terms of coordinates adapted to the wave-induced fast motion. Instead of evolving the particle variables directly, one transforms to oscillation-center variables so that nonresonant oscillations are removed order by order, while resonant interactions remain as secular transport. In the specific unmagnetized plasma problem rederived by Lie transform, the theory applies to a charged particle in an electrostatic wave of small amplitude and produces a slow-time diffusion equation for the background distribution $F_0$ (Brizard, 22 Sep 2025).

In a more general Hamiltonian setting, quasilinear theory can be stated in terms of a dressed or oscillation-center distribution

$$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$

which satisfies

$$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$

Here $H_s^{\rm OC}=H_{0s}+\Delta_s$ includes the ponderomotive energy $\Delta_s$, $\Theta_s^{\,ij}$ is a dressing tensor, $D_s^{ij}\ge 0$ is the quasilinear diffusion tensor, and $C_s$ is a Balescu–Lenard collision operator (Dodin, 2022). This places Dewar’s electrostatic oscillation-center construction within a wider Hamiltonian program rather than as an isolated derivation.

2. Extended-phase-space Hamiltonian and Lie-transform structure

The Lie-transform derivation for the unmagnetized electrostatic problem is formulated in extended canonical phase space

$z^{\alpha}=(\mathbf{x},\mathbf{p},w,t),$

with Poisson bracket

$$\{f,g\} =\bigl(\partial_{w}f\,\partial_{t}g-\partial_{t}f\,\partial_{w}g\bigr) +\nabla f\cdot\partial_{\mathbf p}g-\partial_{\mathbf p}f\cdot\nabla g.$$

For a particle of charge $$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$0 and mass $$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$1 interacting with a prescribed electrostatic wave $$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$2 of small amplitude $$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$3, the extended Hamiltonian is

$$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$4

The oscillation-center variables $$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$5 are introduced by a near-identity canonical transformation generated by scalar functions $$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$6 (Brizard, 22 Sep 2025).

The push-forward of the Hamiltonian is expanded as

$$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$7

with

$$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$8

$$F_s(x,p)=\bar f_s+\tfrac12\,\partial_{p_i}\bigl(\Theta_s^{\,ij}(x,p)\,\bar f_s\bigr),$$9

$$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$0

where

$$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$1

The distribution is transformed by pull-back: $$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$2 If $$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$3, this generates the usual order-by-order relations between particle-space and oscillation-center distributions. The central structural point is that the Lie transform is not introduced as a mere change of notation: it is the mechanism that isolates fast oscillatory dynamics from slow secular transport.

3. Fast-phase separation, resonance, and the first-order problem

The derivation separates fast and slow scales through an eikonal ansatz,

$$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$4

with fast phase

$$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$5

The slow variable $$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$6 is the quasilinear time scale, while the eikonal phase carries the rapid wave oscillation (Brizard, 22 Sep 2025).

At first order, $$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$7 is chosen so that the oscillatory part of $$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$8 vanishes except possibly in a narrow resonant window $$\partial_tF_s+\{H_{s}^{\rm OC},\,F_s\}_z =\partial_{p_i}\!\Bigl(D^{ij}_s(x,p)\,\partial_{p_j}F_s\Bigr)+C_s.$$9. Using the Bateman–Kruskal method, the formal solution is

$H_s^{\rm OC}=H_{0s}+\Delta_s$0

This removes the fast nonresonant ponderomotive oscillations. The residual resonant first-order Hamiltonian is

$H_s^{\rm OC}=H_{0s}+\Delta_s$1

The first-order oscillation-center Vlasov equation is

$H_s^{\rm OC}=H_{0s}+\Delta_s$2

which yields the resonant solution

$H_s^{\rm OC}=H_{0s}+\Delta_s$3

The formal role of $H_s^{\rm OC}=H_{0s}+\Delta_s$4 is therefore explicit: it retains the resonant sector of the first-order Hamiltonian while the remaining nonresonant part is absorbed into oscillation-center dynamics.

4. Second-order closure and the quasilinear diffusion equation

The slow evolution of the background distribution appears at second order. The second-order oscillation-center Vlasov equation is

$H_s^{\rm OC}=H_{0s}+\Delta_s$5

After averaging over the fast eikonal phase, the Lie-transform derivation gives

$H_s^{\rm OC}=H_{0s}+\Delta_s$6

so that

$H_s^{\rm OC}=H_{0s}+\Delta_s$7

The diffusion tensor is

$H_s^{\rm OC}=H_{0s}+\Delta_s$8

Moreover,

$H_s^{\rm OC}=H_{0s}+\Delta_s$9

and the final oscillation-center quasilinear diffusion equation becomes

$\Delta_s$0

In this derivation, the oscillation-center diffusion tensor is identical to the particle-space result (Brizard, 22 Sep 2025).

This identity is sometimes misread as implying that the oscillation-center reformulation adds no content. The Lie-transform result shows the opposite: the transformed formulation makes explicit which pieces are removed as nonresonant oscillatory dynamics, which pieces survive as resonant transport, and how the second-order closure is obtained. In particular, the explicit second-order oscillation-center Vlasov equation and the proof that $\Delta_s$1 are stated as gaps in Dewar’s original presentation that the Lie-transform derivation fills.

5. General Hamiltonian formulations and magnetized extensions

The unmagnetized electrostatic theory sits inside a broader Hamiltonian framework for inhomogeneous turbulence. In that setting, the particle Hamiltonian $\Delta_s$2 is kept arbitrarily general, including relativistic electromagnetic, Newtonian gravitational, and relativistic gravitational cases. The diffusion tensor can be written in Weyl-symbol form and, in canonical variables, reduces to

$\Delta_s$3

where $\Delta_s$4 is the positive-semidefinite Wigner function of the perturbation Hamiltonian. In the same theory, a Balescu–Lenard-type collision operator emerges, conserving particles, momentum, and energy and satisfying the H-theorem; off-shell waves are permitted; and for nonresonant waves the full local wave-kinetic equation conserves wave action when $\Delta_s$5 (Dodin, 2022). Dewar’s electrostatic oscillation-center quasilinear theory is identified there as a particular case, with

$\Delta_s$6

$\Delta_s$7

$\Delta_s$8

For magnetized plasmas, the oscillation-center construction is expressed in guiding-center or action-angle variables. In a uniform background field, one works with $\Delta_s$9; in a nonuniform axisymmetric field, with three actions $\Theta_s^{\,ij}$0, $\Theta_s^{\,ij}$1, and $\Theta_s^{\,ij}$2. The perturbation Hamiltonian is built from $\Theta_s^{\,ij}$3, expanded in eikonal and gyro-Fourier form, and the slow evolution of the background distribution becomes a Fokker–Planck operator in invariant space (Brizard et al., 2022). In a nonuniform magnetized plasma, the resulting $\Theta_s^{\,ij}$4 diffusion tensor

$\Theta_s^{\,ij}$5

naturally incorporates radial diffusion together with energy and pitch-angle diffusion. The off-diagonal components $\Theta_s^{\,ij}$6, $\Theta_s^{\,ij}$7, and $\Theta_s^{\,ij}$8 encode the fact that these channels are not independent.

6. Validity domain, comparison with Dewar, and extensions

The Lie-transform derivation in the unmagnetized case assumes an unmagnetized, collisionless Vlasov–Poisson plasma that is spatially uniform at leading order, weak turbulence with $\Theta_s^{\,ij}$9, and a single monochromatic or slowly evolving eikonal wave with wave–wave coupling neglected (Brizard, 22 Sep 2025). The resonant approximation retains only particles satisfying

$D_s^{ij}\ge 0$0

as contributors to secular diffusion, while nonresonant effects survive only through the ponderomotive Hamiltonian. The scale separation condition is

$D_s^{ij}\ge 0$1

and the background $D_s^{ij}\ge 0$2 must remain smooth.

The comparison with Dewar’s original formulation contains several technical clarifications. The Lie-transform derivation uses the sign convention of Brizard and Hahm, whereas Dewar’s equations use the opposite sign in the definition of the generating functions; after adjustment, the final $D_s^{ij}\ge 0$3 is identical. Dewar did not explicitly present the second-order oscillation-center Vlasov equation and did not derive $D_s^{ij}\ge 0$4; those steps are supplied in the later derivation. The narrow-window function $D_s^{ij}\ge 0$5 is identified as shorthand for retaining only the resonant part of $D_s^{ij}\ge 0$6, and outside resonance the standard ponderomotive removal is recovered. A further difference is normalization: the Lie-transform formulation uses the extended-phase-space variables $D_s^{ij}\ge 0$7 and $D_s^{ij}\ge 0$8, whereas Dewar worked strictly in $D_s^{ij}\ge 0$9 and enforced the dispersion relation by hand.

Possible extensions listed for the unmagnetized derivation include magnetized plasmas, multiple electromagnetic wave branches and polarization, relativistic particle dynamics, multi-species and inhomogeneous background profiles, and the inclusion of collisions on long time scales (Brizard, 22 Sep 2025). A plausible implication, supported by the broader Hamiltonian and magnetized formulations, is that oscillation-center quasilinear theory is best regarded not as a special electrostatic trick but as a general perturbative architecture for organizing resonant diffusion, ponderomotive structure, and invariant-space transport across a wide class of plasma Hamiltonians.