From magnetized Coulombic quantum dynamics to magnetized fluids

Abstract: We extend the quantum modulated energy developed in [17] in order to de- rive the magnetized pressureless Euler-Poisson equation as a semiclassical and mean field semiclassical limit from the magnetized Schrödinger-Poisson and von-Neumann equations, respectively. Local well-posedness of the underlying monokinetic PDE is also addressed. In both limits, the magnetic field is external and may be spatially non-uniform. Our results fall in the broader scope of semiclassical and quantum mean field limits for magnetized quantum dynamics.

• The paper proves that magnetized Schrödinger-Poisson and many-body von Neumann dynamics converge to the magnetized pressureless Euler-Poisson system, with rates of order O(ℏ²) and O(ℏ² + N^−β), respectively.
• The analysis combines quantum modulated energy, magnetic commutator cancellations, and Coulomb-flow estimates to control singular interactions and establish convergence up to the first fluid blow-up time.
• The results apply to external magnetic potentials satisfying smoothness and decay assumptions, while global convergence, weaker velocity regularity, general non-decaying fields, and self-consistent magnetic fields remain open problems.

Overview

The paper under review establishes rigorous semiclassical and mean-field semiclassical limits for magnetized Coulombic quantum dynamics, deriving the magnetized pressureless Euler-Poisson system in three dimensions. Two limit regimes are treated: the semiclassical limit $\hbar \to 0$ from the magnetized Schrödinger-Poisson equation (SPA), and the joint limit $\hbar + 1/N \to 0$ from the $N$-body magnetized von Neumann equation (v-NA). In both cases the target is the monokinetic system

$$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$

where $V(x) = (4\pi|x|)^{-1}$ is the repulsive Coulomb kernel and $\mathbf{J} = D_x A - D_x^T A$ is the antisymmetrized Jacobian of the magnetic vector potential $A$. The method is a magnetized adaptation of the quantum modulated energy technique introduced by Golse–Paul, combined with the commutator estimates of Serfaty for Coulomb flows. To the author's knowledge, the $N$-body result is the first mean-field limit for quantum many-body dynamics that are simultaneously singular (Coulomb) and magnetized.

Main results

Both theorems are conditional on a technical assumption (A1–A2) on $A$: smoothness, together with polynomial decay of all derivatives of $\operatorname{curl}_x A$ at rate $\hbar + 1/N \to 0$0 and boundedness of all first derivatives of $\hbar + 1/N \to 0$1. This assumption is borrowed from Lührmann's work on mean-field quantum dynamics with magnetic fields and accommodates spatially non-uniform magnetic fields, including linear vector potentials corresponding to constant magnetic fields. The convergence is weak, in $\hbar + 1/N \to 0$2 for the density and $\hbar + 1/N \to 0$3 for the current, uniformly on $\hbar + 1/N \to 0$4 for any $\hbar + 1/N \to 0$5 preceding the first blow-up time of the Euler-Poisson solution, and requires the initial modulated energy to vanish as $\hbar + 1/N \to 0$6 (resp. $\hbar + 1/N \to 0$7).

The first theorem treats the one-body problem. Given $\hbar + 1/N \to 0$8 with $\hbar + 1/N \to 0$9 and classical data $N$0, the magnetized quantum density $N$1 and the magnetized current

$N$2

converge to $N$3 and $N$4 respectively, in the sense that $N$5 provided $N$6. The second theorem is the $N$7-body analogue: for symmetric density operators $N$8 with finite $N$9, the one-body marginal density $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$0 and current $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$1 converge to $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$2 and $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$3 as $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$4, again uniformly on $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$5.

Well-posedness of the underlying equations

The paper assembles the analytic prerequisites. Global well-posedness of (SPA) in the magnetic Sobolev space $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$6 is imported from Lührmann, with the observation that the $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$7-dependent problem follows from the $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$8 case by the scaling $$\partial_t \rho + \operatorname{div}_x(\rho u) = 0, \qquad \partial_t u + u D_x u + u\mathbf{J} + \nabla_x V * \rho = 0,$$9. The magnetic Sobolev spaces are necessary because $V(x) = (4\pi|x|)^{-1}$0 is not assumed bounded, which is precisely what allows constant magnetic fields; $V(x) = (4\pi|x|)^{-1}$1 does not coincide with the ordinary Sobolev space without boundedness of $V(x) = (4\pi|x|)^{-1}$2.

For the $V(x) = (4\pi|x|)^{-1}$3-body von Neumann equation, self-adjointness of $V(x) = (4\pi|x|)^{-1}$4 is proved via Kato's perturbation theorem. The key estimate controls the pair interaction $V(x) = (4\pi|x|)^{-1}$5 using the 3D Hardy inequality combined with the diamagnetic inequality $V(x) = (4\pi|x|)^{-1}$6, yielding $V(x) = (4\pi|x|)^{-1}$7 with $V(x) = (4\pi|x|)^{-1}$8 after choosing $V(x) = (4\pi|x|)^{-1}$9 small. Stone's theorem then produces the unitary propagator, and a trace estimate shows the modulated energy is well defined for all times.

The magnetized Euler-Poisson system is shown to be locally well posed in $\mathbf{J} = D_x A - D_x^T A$0 by a contraction argument in a Banach space of velocity fields controlled in $\mathbf{J} = D_x A - D_x^T A$1. The proof requires a stability estimate for the transport equation in the $\mathbf{J} = D_x A - D_x^T A$2 norm and an $\mathbf{J} = D_x A - D_x^T A$3 propagation estimate for transported densities. The choice of the bi-Laplacian control on $\mathbf{J} = D_x A - D_x^T A$4 is dictated by the later needs of the modulated energy argument, which requires $\mathbf{J} = D_x A - D_x^T A$5. Local, rather than global, well-posedness is natural since Euler-Poisson solutions may blow up in finite time for general data.

Formal derivation of the magnetized monokinetic system

A proposition of independent interest shows that the zeroth and first velocity moments of the magnetized Vlasov-Poisson equation,

$\mathbf{J} = D_x A - D_x^T A$6

satisfy a moment system with the Lorentz-type term $\mathbf{J} = D_x A - D_x^T A$7. Inserting the monokinetic ansatz $\mathbf{J} = D_x A - D_x^T A$8 and rewriting in terms of $\mathbf{J} = D_x A - D_x^T A$9 yields the target Euler-Poisson system. The computation exploits $A$0 and the antisymmetry of $A$1, the latter ensuring $A$2. This derivation clarifies why $A$3, rather than the full curl, appears in the limiting fluid equation.

The semiclassical limit

The core of the argument is the computation of the time derivative of the magnetized quantum modulated energy

$A$4

Working with the magnetic Hartree equation satisfied by the rank-one projector $A$5, the author derives

$A$6

where $A$7. The decisive new ingredient is a cancellation lemma: for antisymmetric $A$8, the magnetic contributions arising from $A$9 cancel identically against the $N$0 terms from the Euler equation. The identity rests on the fact that $N$1 and $N$2 are commuting multiplication operators. This cancellation is the magnetic counterpart of the structural alignment between the quantum commutator structure and the classical Lorentz force; without it, the Grönwall argument would fail.

Both residual terms are then bounded: the trace term by $N$3, and the interaction term by $N$4, using the identity $N$5. Grönwall's lemma yields $N$6, so the convergence rate in $N$7 is at least $N$8 modulo the initial data. The convergence of $N$9 to $A$0 then follows by standard arguments unaffected by the magnetic field.

The mean-field semiclassical limit

The $A$1-body argument uses a renormalized modulated energy

$A$2

where $A$3 and the additive $A$4 correction (from Duerinckx's lemma) guarantees non-negativity despite the Coulomb singularity. The time-derivative computation parallels the one-body case, with the same magnetic cancellation lemma applied particle by particle, and the interaction part reorganizes into

$A$5

Bounding this term requires Serfaty's commutator estimate for Coulomb flows, giving $A$6 for some $A$7. The resulting Grönwall inequality,

$A$8

is the main quantitative output: the joint convergence holds with rate $A$9.

Admissible initial data

The paper constructs wave functions realizing $\operatorname{curl}_x A$0 for initial data $\operatorname{curl}_x A$1 with $\operatorname{curl}_x A$2, $\operatorname{curl}_x A$3, and $\operatorname{curl}_x A$4 with $\operatorname{curl}_x A$5. The construction partitions the support of $\operatorname{curl}_x A$6 into cubes of width $\operatorname{curl}_x A$7 and builds a WKB-type superposition $\operatorname{curl}_x A$8 with $\operatorname{curl}_x A$9, subject to the scaling condition $\hbar + 1/N \to 0$00. The kinetic part is controlled through three terms: an $\hbar + 1/N \to 0$01 term from $\hbar + 1/N \to 0$02, a cutoff-gradient term $\hbar + 1/N \to 0$03, and a localization error $\hbar + 1/N \to 0$04 from the Lipschitz variation of $\hbar + 1/N \to 0$05. The tensor product ansatz $\hbar + 1/N \to 0$06 then reduces the $\hbar + 1/N \to 0$07-body initial energy to the one-body one, plus a $\hbar + 1/N \to 0$08 self-interaction remainder that vanishes since $\hbar + 1/N \to 0$09.

Notably, the author points out that the admissible-data construction of Ben-Porat–Chen–Yuan for the quantum quasi-neutral limit does not transfer here, since it perturbs the Laplacian by a gradient, which would force the magnetic field to vanish.

Limitations and open questions

Several restrictions are acknowledged. The assumption A1–A2 excludes general magnetic fields with non-decaying curl beyond the linear case; in particular, the treatment of arbitrary uniform fields via linear potentials is covered, but the decay condition on $\hbar + 1/N \to 0$10 is a genuine constraint for non-uniform fields. The convergence is weak ($\hbar + 1/N \to 0$11 and $\hbar + 1/N \to 0$12) and holds only up to the first blow-up time of the Euler-Poisson solution; no global-in-time result is claimed, and blow-up can occur in finite time for repulsive Euler-Poisson. The regularity demanded of $\hbar + 1/N \to 0$13 — Lipschitz with bounded Laplacian — is stronger than what the modulated energy argument strictly needs, and the author explicitly raises whether it can be relaxed. The obstacle is identified precisely: for $\hbar + 1/N \to 0$14 with $\hbar + 1/N \to 0$15, controlling $\hbar + 1/N \to 0$16 would require uniform-in-$\hbar + 1/N \to 0$17 $\hbar + 1/N \to 0$18 bounds on $\hbar + 1/N \to 0$19, which in the unmagnetized case follow from propagation of quantum moments but whose magnetized extension appears non-trivial for general Lipschitz $\hbar + 1/N \to 0$20. The author further notes that such a propagation of quantum moments result would likely be a key ingredient for deriving the magnetized Vlasov-Poisson equation as a semiclassical limit from (SPA), which remains open. Finally, the derivation targets external magnetic fields only; the self-consistent case, i.e. monokinetic PDEs with self-consistent magnetic fields from quantum many-body dynamics, is left as the principal open problem.

Conclusion

The paper extends the quantum modulated energy method to magnetized Coulombic quantum dynamics and obtains, with explicit rates $\hbar + 1/N \to 0$21, the magnetized pressureless Euler-Poisson system as both a semiclassical and a mean-field semiclassical limit. The mathematical substance lies in identifying the exact cancellation structure between the magnetic commutator $\hbar + 1/N \to 0$22 and the Lorentz term $\hbar + 1/N \to 0$23 in the limiting fluid equation, a cancellation that persists at the $\hbar + 1/N \to 0$24-body level and combines with Serfaty's Coulomb commutator estimates to close the Grönwall argument. The results are conditional on vanishing initial modulated energy, for which admissible WKB-type data are constructed, and on the decay assumptions A1–A2 on the vector potential.