Schwinger's 1951 gauge-invariant proper-time formulation of vacuum polarization yielded the exact one-loop effective action of QED in a constant electromagnetic background, whose imaginary part gives the celebrated rate for electron-positron pair production in a constant electric field (the Schwinger effect). For general strong time-dependent backgrounds — the physically relevant case for intense laser foci and astrophysical fields — no exact nonperturbative result exists: only special backgrounds (constant fields, single plane waves, and sech-squared profiles in time or in space) are exactly solvable, general backgrounds are treated by approximations of bounded validity (worldline instantons, the locally constant field approximation, derivative expansions, quantum-kinetic numerics, dynamically assisted mechanisms), and the backreaction of the produced pairs on the field is handled only in mean-field, semiclassical, lattice, or reduced-dimension settings. The problem posed here is to determine, nonperturbatively and with controlled error, the pair-production probability for charged spin-1/2 fermions in an arbitrarily time-dependent strong electromagnetic field, and to solve the self-consistent evolution of the field including the induced current of the produced pairs, with exact reduction to the known constant-field, plane-wave, sech-squared-pulse, and perturbative limits.

Background

Quantum electrodynamics (QED) predicts that the quantum vacuum behaves as a nonlinear polarizable medium: virtual electron-positron pairs screen and modify any applied electromagnetic field, an effect called vacuum polarization. The standard way to encode this is the one-loop effective action — the action of the electromagnetic field obtained after integrating out the electron field, order by order in electron loops; its imaginary part measures the probability that the field promotes virtual pairs into real ones. Heisenberg and Euler (1936) computed this quantity for slowly varying fields as a power series; Schwinger (1951) then introduced the proper-time method (a representation of the electron propagator as an integral over an auxiliary evolution parameter) and used it to evaluate the exact one-loop effective Lagrangian for an arbitrary constant background field. Its imaginary part yields the rate at which a constant electric field EE produces real electron-positron pairs — the Schwinger effect, a tunneling process with rate per unit volume and time 2 Im L(1)=(eE)24π3∑n=1∞1n2exp⁡(−nπm2/eE)2\,\mathrm{Im}\,\mathcal{L}^{(1)}=\frac{(eE)^2}{4\pi^3}\sum_{n=1}^{\infty}\frac{1}{n^2}\exp(-n\pi m^2/eE) (natural units ℏ=c=1\hbar=c=1, with ee the charge and mm the electron mass), exponentially suppressed until fields approach the critical scale Ec=m2/e≈1.3×1018 V/mE_c=m^2/e\approx 1.3\times 10^{18}\,\mathrm{V/m}. Schwinger's paper also proved that a single plane wave (a field depending only on one light-like combination of coordinates, of arbitrary strength and spectral composition) produces no nonlinear vacuum phenomena and no pairs, computed the pair-production probability perturbatively (as an expansion in powers of the field) for general backgrounds, and treated the decay of a neutral meson into photons via vacuum polarization in a weak-field approximation. Throughout, the background field is prescribed: the created pairs are assumed not to act back on the field, and the paper poses no explicit open problems.

The exact nonperturbative result is thus confined to idealized backgrounds. Besides constant fields and single plane waves, the exactly tractable inhomogeneous backgrounds form the sech-squared family: spatially uniform electric fields with the hyperbolic-secant-squared time profile E(t)=E0 sech2(ωt)E(t)=E_0\,\mathrm{sech}^2(\omega t) — the Sauter pulse, named for Sauter's pioneering 1931 treatment of relativistic tunneling in a homogeneous electric field, and the only exactly solvable time-dependent case — together with static analogues depending on a single spatial coordinate; in each of these the Dirac and Klein-Gordon mode equations reduce to hypergeometric form. The alternating-field pair-production computation of Brezin and Itzykson (1970) and the equivalent scattering-amplitude expressions are summarized in Dunne's 2004 review, which states that these exhaust the analytically solvable cases with physically realistic inhomogeneous backgrounds. All realistic situations — pulsed and focused laser beams, neutron-star magnetospheres, heavy-ion collisions — involve strong fields varying in both space and time. Later literature therefore developed systematic approximations with bounded domains of validity: the derivative expansion (slowly varying fields); worldline instantons (classical closed loops in Euclidean spacetime that dominate the semiclassical pair-production path integral in inhomogeneous fields; Dunne and Schubert 2005); the locally constant field approximation, or LCFA (replacing the rate at each spacetime point by the constant-field rate of the local field value; its numerical domain of validity was mapped by Sevostyanov et al. 2021); quantum-kinetic methods (evolving the Bogoliubov coefficients — the amplitudes mixing positive- and negative-frequency solutions that quantify pair creation in a time-dependent background — as time-nonlocal rate equations; Hebenstreit, Alkofer and Gies 2008); and the dynamically assisted Schwinger mechanism, in which a weak rapidly varying field superimposed on a strong slow one dramatically enhances the tunneling rate (Schützhold, Gies and Dunne 2008), with higher-order perturbative treatments in the weak component by Torgrimsson (2019).

The second limitation is backreaction: the produced pairs accelerate in the field, carry away its energy, and induce an electric current that modifies the very field that created them, producing plasma oscillations and charge screening. In three spatial dimensions with spatially inhomogeneous fields no self-consistent treatment exists; what is available are mean-field approximations, real-time lattice simulations (Liu 2023), self-consistent semiclassical Maxwell–Schrödinger dynamics restricted to spatially homogeneous fields (Kaushal and Singh 2024), and exactly solvable two-dimensional analogues such as the 1+1-dimensional massive Schwinger model — quantum electrodynamics in one spatial dimension with a massive fermion, rewritten by bosonization as a bosonic theory (Gralla and Mizuno 2025). The reviews of Dunne (2004) and Gelis and Tanji (2016) both frame the field's central unfinished business exactly this way: exact results exist only for special backgrounds, general backgrounds are covered by approximations, and the self-consistent backreaction problem is treated numerically or in restricted settings. Direct experimental observation of the Schwinger effect itself is still outstanding.

Problem Statement

Determine, nonperturbatively and with controlled error, the rate at which a strong, arbitrarily time-dependent classical electromagnetic background produces electron-positron pairs from the vacuum, and determine the self-consistent evolution of the field including the backreaction of the produced pairs. In parts:

(i) For quantized spinor QED (a charged Dirac field of mass mm) in a prescribed classical background Fμν(x)F_{\mu\nu}(x) of general space-time dependence and field strengths of order the critical scale m2/em^2/e or larger, compute the probability of pair production from the vacuum — defined through the imaginary part of the one-loop effective action, with spacetime density 2 Im L(1)(x)2\,\mathrm{Im}\,\mathcal{L}^{(1)}(x) and 2 Im W(1)=∫d4x 2 Im L(1)(x)2\,\mathrm{Im}\,W^{(1)}=\int d^4x\,2\,\mathrm{Im}\,\mathcal{L}^{(1)}(x); in a constant field this density equals the rate per unit volume and time 2 Im L(1)2\,\mathrm{Im}\,\mathcal{L}^{(1)} — without expanding in powers of the field and without restricting to constant fields, single plane waves, or sech2\mathrm{sech}^2-type time profiles. For time-dependent backgrounds the corresponding particle-number observable must be fixed by an explicit definition (through asymptotic in/out Bogoliubov coefficients, instantaneous projections, or detector-based definitions), together with the limiting sense in which it reproduces 2 Im W(1)2\,\mathrm{Im}\,W^{(1)}. Determine the rate in closed form or as a systematically improvable convergent computation with certified error bounds, valid for a class of backgrounds broad enough to include arbitrary time dependence and spatial inhomogeneity, and not restricted to the exactly solvable profiles.

(ii) Solve the coupled self-consistent problem in which the produced pairs' induced current sources the field: the semiclassical Maxwell–Dirac system with pair-production backreaction, in three spatial dimensions and for spatially inhomogeneous fields, conserving total energy-momentum. Solve this system exactly or by a certified convergent numerical scheme with quantified errors.

The target result must reduce exactly, in the appropriate limits, to (a) the constant-field Schwinger rate 2 Im L(1)=(eE)24π3∑n=1∞1n2exp⁡(−nπm2/eE)2\,\mathrm{Im}\,\mathcal{L}^{(1)}=\frac{(eE)^2}{4\pi^3}\sum_{n=1}^{\infty}\frac{1}{n^2}\exp(-n\pi m^2/eE); (b) the vanishing of the effective action and of pair production for a single plane wave; (c) the exact hypergeometric-mode results for spatially uniform sech2(ωt)\mathrm{sech}^2(\omega t) pulses; and (d) the perturbative expansion of the pair-production probability in powers of the field for weak general backgrounds, including its leading multiphoton terms for weak, rapidly varying fields; and, in part (ii), the self-consistent solution must satisfy Maxwell's equations with the induced current and conserve total energy-momentum. Throughout, 'certified' (equivalently, controlled or quantified) error means the error of a claimed result is explicitly quantified and independently checkable — a stated bound on the truncation and regulator error where one exists, or, for a numerical scheme, a demonstrated convergence with a reproducible error estimate — rather than a qualitative claim that the error is small.

References

  1. J. Schwinger, 'On Gauge Invariance and Vacuum Polarization', Physical Review 82, 664–679 (1951). DOI: 10.1103/PhysRev.82.664, https://doi.org/10.1103/PhysRev.82.664
  2. J. Schwinger, 'On Gauge Invariance and Vacuum Polarization', in A Quantum Legacy (World Scientific Series in 20th Century Physics), World Scientific, pp. 116–131 (2000). DOI: 10.1142/9789812795694_0006, https://doi.org/10.1142/9789812795694_0006
  3. F. Sauter, 'Über das Verhalten eines Elektrons im homogenen elektrischen Feld nach der relativistischen Theorie Diracs', Zeitschrift für Physik 69, 742–764 (1931). DOI: 10.1007/BF01339461, https://doi.org/10.1007/BF01339461
  4. W. Heisenberg and H. Euler, 'Consequences of Dirac Theory of the Positron' (English translation by W. Korolevski and H. Kleinert of Z. Phys. 98, 714–732 (1936)), arXiv:physics/0605038 (2006), https://arxiv.org/abs/physics/0605038
  5. E. Brezin and C. Itzykson, 'Pair Production in Vacuum by an Alternating Field', Physical Review D 2, 1191–1199 (1970). DOI: 10.1103/PhysRevD.2.1191, https://doi.org/10.1103/PhysRevD.2.1191
  6. G. V. Dunne, 'Heisenberg-Euler Effective Lagrangians: Basics and Extensions', arXiv:hep-th/0406216 (2004); also published as a chapter of From Fields to Strings: Circumnavigating Theoretical Physics (World Scientific), pp. 445–522 (2005). DOI: 10.1142/9789812775344_0014, https://arxiv.org/abs/hep-th/0406216
  7. G. V. Dunne and C. Schubert, 'Worldline instantons and pair production in inhomogenous fields', Physical Review D 72, 105004 (2005). DOI: 10.1103/PhysRevD.72.105004, arXiv:hep-th/0507174, https://arxiv.org/abs/hep-th/0507174
  8. R. Schützhold, H. Gies and G. Dunne, 'Dynamically Assisted Schwinger Mechanism', Physical Review Letters 101, 130404 (2008). DOI: 10.1103/PhysRevLett.101.130404, arXiv:0807.0754, https://arxiv.org/abs/0807.0754
  9. F. Hebenstreit, R. Alkofer and H. Gies, 'Pair production beyond the Schwinger formula in time-dependent electric fields', Physical Review D 78, 061701 (2008). DOI: 10.1103/PhysRevD.78.061701, arXiv:0807.2785, https://arxiv.org/abs/0807.2785
  10. F. Gelis and N. Tanji, 'Schwinger mechanism revisited', Progress in Particle and Nuclear Physics 87, 1–49 (2016). DOI: 10.1016/j.ppnp.2015.11.001, arXiv:1510.05451, https://arxiv.org/abs/1510.05451
  11. G. Torgrimsson, 'Perturbative methods for assisted nonperturbative pair production', Physical Review D 99, 096002 (2019). DOI: 10.1103/PhysRevD.99.096002, arXiv:1812.04607, https://arxiv.org/abs/1812.04607
  12. D. G. Sevostyanov, I. A. Aleksandrov, G. Plunien and V. M. Shabaev, 'Total yield of electron-positron pairs produced from vacuum in strong electromagnetic fields: Validity of the locally constant field approximation', Physical Review D 104, 076014 (2021). DOI: 10.1103/PhysRevD.104.076014, arXiv:2012.10751, https://arxiv.org/abs/2012.10751
  13. W. Liu, 'Schwinger pair creation with the backreaction in 3+1 dimensions', Physical Review D 107, 054502 (2023). DOI: 10.1103/PhysRevD.107.054502, arXiv:2302.05143, https://arxiv.org/abs/2302.05143
  14. S. Kaushal and S. Singh, 'Backreaction inclusive Schwinger effect in flat and de Sitter spacetimes via a self consistent Maxwell Schrodinger semiclassical dynamics', arXiv:2412.09436 (2024; v2 September 2025), https://arxiv.org/abs/2412.09436
  15. S. E. Gralla and M. Mizuno, 'Schwinger effect with backreaction in 1+1D massive QED with a strong external field', Physical Review D 113, 085007 (2026). DOI: 10.1103/5jhz-cbf7, arXiv:2511.23464, https://arxiv.org/abs/2511.23464