Quantum Electrodynamics (QED) Overview

• Quantum Electrodynamics (QED) is a relativistic quantum field theory that describes the interactions of electrons, positrons, and photons through perturbation theory.
• It employs multiple calculational frameworks, including path-integral and worldline approaches, to handle bound states, renormalization, and asymptotic series in precision measurements.
• Recent advances in QED focus on infrared behavior, strong-field effects, and analog systems, highlighting its foundational role in modern theoretical and experimental physics.

Searching arXiv for recent and foundational QED papers to ground the article. Quantum electrodynamics (QED) is the relativistic quantum field theory of electrons, positrons and photons. Its basic ingredients are the Dirac field $\psi(x)$, the electromagnetic field $A^\mu(x)$, and the minimal coupling $e\,\bar\psi\gamma^\mu\psi\,A_\mu$; perturbation theory expands $S$-matrix elements in powers of the fine-structure constant $\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$ (Consa, 2021). In atomic and precision contexts, QED provides the only systematic framework in which the interaction of atomic electrons with the quantized electromagnetic field can be treated to the accuracy demanded by precision spectroscopy and nuclear-moment determinations (Wehrli et al., 2022).

1. Field content and basic dynamical structure

In the formulation summarized in the literature, the fermionic sector is described by

$$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$

the electromagnetic sector by

$$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$

and the interaction by minimal coupling $e\,\bar\psi\gamma^\mu\psi\,A_\mu$ (Consa, 2021). These ingredients define the standard relativistic description of charged spin-$\frac12$ matter interacting with radiation.

Several modern reformulations preserve this content while changing the calculational organization. In the worldline approach, matter and gauge fields are integrated out explicitly, resulting in a many-body Lorentz covariant theory of $0+1$ dimensional worldlines describing super-pairs of spinning charges interacting through Lorentz forces (Feal et al., 2022). In the path-integral treatment of bound systems, the Fock–Feynman–Schwinger representation yields a relativistic Hamiltonian with an auxiliary einbein variable $A^\mu(x)$0, from which both exact Dirac spectra and nonrelativistic expansions can be recovered (Simonov, 2014).

These formulations do not replace the standard field-theoretic definition; rather, they reorganize it. A plausible implication is that QED is unusual among quantum field theories in the range of mathematically distinct but physically equivalent calculational frameworks that remain directly useful across perturbative, bound-state, and infrared problems.

2. Perturbation theory, renormalization, and asymptotic character

Perturbative QED organizes observables as expansions in $A^\mu(x)$1. For the electron magnetic moment one writes

$A^\mu(x)$2

with the one-loop Schwinger result

$A^\mu(x)$3

and higher orders involving 72 diagrams at sixth order, 891 at eighth order, and more than 12 000 at tenth order (Consa, 2021). This perturbative structure underlies the standard statement that QED is the most accurate theory in the history of science, although one of the papers in the source set emphasizes that this precision is commonly anchored to the anomalous magnetic moment of the electron and criticizes aspects of its historical computation (Consa, 2021).

Renormalization enters through bare parameters and counterterms. In the notation given in the source material, one rewrites the Lagrangian using $A^\mu(x)$4, $A^\mu(x)$5, and renormalization constants $A^\mu(x)$6, $A^\mu(x)$7, together with

$A^\mu(x)$8

choosing $A^\mu(x)$9 order by order in $e\,\bar\psi\gamma^\mu\psi\,A_\mu$0 so as to absorb infinities into $e\,\bar\psi\gamma^\mu\psi\,A_\mu$1 and $e\,\bar\psi\gamma^\mu\psi\,A_\mu$2, leaving finite renormalized $e\,\bar\psi\gamma^\mu\psi\,A_\mu$3-point functions (Consa, 2021). In finite-temperature applications the same logic appears through $e\,\bar\psi\gamma^\mu\psi\,A_\mu$4 and $e\,\bar\psi\gamma^\mu\psi\,A_\mu$5, with a temperature-dependent renormalized coupling $e\,\bar\psi\gamma^\mu\psi\,A_\mu$6 (Masood, 2018).

The perturbative series is not expected to converge. Dyson’s argument indicates that QED expansions are asymptotic, and the Chandrasekhar-limit analogy gives an estimate for the optimal truncation order,

$e\,\bar\psi\gamma^\mu\psi\,A_\mu$7

so that for physical QED

$e\,\bar\psi\gamma^\mu\psi\,A_\mu$8

(Kolomeisky, 2014). The same analysis predicts much smaller optimal orders in condensed-matter implementations of QED: around $e\,\bar\psi\gamma^\mu\psi\,A_\mu$9 for narrow band-gap semiconductors and Weyl semimetals, and severe limitations for graphene (Kolomeisky, 2014). This suggests that asymptoticity is a conceptual constraint in particle-physics QED but can become a practical one in strongly coupled analogue systems.

3. Infrared structure, soft theorems, and dressed states

A major development in QED concerns its infrared sector. In abelian gauge theories, the soft-photon theorem is equivalent to the Ward identity of an infinite-dimensional large $S$0 symmetry acting nontrivially at null infinity (Kapec et al., 2015). In retarded radial gauge, the residual gauge parameter approaches an angle-dependent constant,

$S$1

and the $S$2-matrix obeys

$S$3

with the conserved charges split into soft and hard parts. The resulting Ward identity reproduces the Weinberg soft factor,

$S$4

(Kapec et al., 2015). For massive charged particles, the hard charge is carried by the Liénard–Wiechert fields at timelike infinity, with antipodal matching imposed at spatial infinity (Kapec et al., 2015).

A complementary development is the worldline reformulation, in which soft singularities are captured and removed by endpoint photon exchanges at infinity that are equivalent to the soft coherent dressings of the Dyson $S$5-matrix proposed by Faddeev and Kulish (Feal et al., 2022). In that language, generalized Wilson loops and lines arise naturally, self-energy and exchange subgraphs exponentiate in one formal expression, and soft theorems emerge from worldline contour averages (Feal et al., 2022).

A more recent proposal studies QED in the infrared regime using the adiabatic approximation and the functional Berry phase, arguing that the physical state space is exact, nonperturbatively dressed, and endowed with a topological structure (Gamboa et al., 15 Jul 2025). In that construction, electrons do not exist as bare particles but as topologically protected electron-photon clouds, with a binding scale estimated at $S$6 and quantized functional Berry flux labeling infrared sectors (Gamboa et al., 15 Jul 2025). This is a proposal rather than established consensus, but it illustrates the extent to which the infrared behavior of QED continues to motivate new nonperturbative descriptions.

4. One-loop effective actions in strong external fields

Strong-field QED is often formulated through one-loop effective actions in prescribed electromagnetic backgrounds. In the in-out formalism for constant magnetic and electric fields, the one-loop spinor QED action can be represented through the gamma function and then rewritten in terms of Hurwitz zeta functions (Kim et al., 2014).

For a constant magnetic field $S$7, the unrenormalized action is given as a sum over Landau-level modes,

$S$8

with $S$9 (Kim et al., 2014). The Ramanujan expansion of $\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$0 reproduces the weak-field Heisenberg–Euler asymptotic series, but this series diverges for supercritical $\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$1 (Kim et al., 2014). Using instead the Whittaker–Watson expansion yields a closed Hurwitz-zeta form,

$\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$2

which remains valid in the supercritical regime (Kim et al., 2014).

For a constant electric field $\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$3, analytic continuation leads to a complex effective action with $\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$4,

$\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$5

whose real part gives vacuum polarization and imaginary part gives vacuum persistence (Kim et al., 2014). The imaginary part reproduces the Schwinger pair-production rate,

$\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$6

(Kim et al., 2014). The formal duality

$\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$7

maps the magnetic and electric results into one another, with the physical distinction that the magnetic field purely polarizes the vacuum whereas the electric field both polarizes the vacuum and produces pairs (Kim et al., 2014).

5. Bound-state QED and precision observables

Bound-state QED in few-electron ions is organized in the Furry picture, in which the electron is embedded in the Coulomb field of the nucleus to all orders and radiative as well as interelectronic interactions are treated perturbatively (Indelicato, 2019). For hydrogenlike ions the one-body one-loop shift is written as

$\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$8

with self-energy and vacuum-polarization contributions; two-loop terms take the analogous form

$\alpha = e^2/(4\pi \hbar c) \approx 1/137.036$9

(Indelicato, 2019). In heliumlike and lithiumlike ions, two-body radiative and non-radiative corrections, screening effects, and many-body methods such as MCDF, RMBPT, configuration interaction, coupled cluster, and model-QED operators are combined to extend predictive reach (Indelicato, 2019). The resulting comparisons with experiment across $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$0 show agreement, for example, for H-like Ly$$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$1, He-like $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$2 lines, and Li-like $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$3–$$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$4 fine structure, while Penning-trap measurements of the bound-electron $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$5-factor in $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$6 and $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$7 agree with theory to better than $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$8 (Indelicato, 2019).

Nuclear magnetic shielding supplies a distinct precision application. In hydrogen- and helium-like systems the leading QED correction to the shielding constant first appears at order $$L_\psi = \bar\psi(i\gamma^\mu\partial_\mu - m)\psi,$$9 and is built from self-energy, vacuum polarization, and vertex contributions (Wehrli et al., 2022). For the hydrogenic $$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$0 state the source material gives

$$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$1

and

$$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$2

with $$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$3 and $$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$4 (Wehrli et al., 2022). The total shielding constants quoted for $$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$5, $$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$6, and neutral $$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$7 are

$$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$8

respectively (Wehrli et al., 2022).

A separate path-integral line of work derives relativistic Hamiltonians directly from the Fock–Feynman–Schwinger representation. In that approach, the calculated spectrum reproduces exactly that of the Dirac hydrogen atom, the Breit–Fermi nonrelativistic expansion is obtained via Foldy–Wouthuysen transformation, and the positronium spectrum, including spin-dependent terms, coincides with standard QED perturbation theory up to $$L_A = -\frac14 F^{\mu\nu}F_{\mu\nu}, \qquad F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu,$$9 (Simonov, 2014). Taken together, these results show that QED is both a high-energy scattering theory and a precision theory of bound levels, splittings, and magnetic response.

6. Finite-temperature, lattice, and analogue realizations

At temperatures around nucleosynthesis, the early universe behaved as a relativistic QED plasma while the temperature was below the neutrino decoupling temperature (Masood, 2018). In the real-time formalism of thermal field theory, renormalization constants become effective thermal parameters, and one obtains a temperature-dependent coupling

$e\,\bar\psi\gamma^\mu\psi\,A_\mu$0

with the high-temperature limit

$e\,\bar\psi\gamma^\mu\psi\,A_\mu$1

(Masood, 2018). Screening and collective behavior are encoded in the Debye scale and plasma frequency,

$e\,\bar\psi\gamma^\mu\psi\,A_\mu$2

with the transverse dispersion relation

$e\,\bar\psi\gamma^\mu\psi\,A_\mu$3

(Masood, 2018). The source explicitly states that light is slowed down and trapped due to bending in such a medium and that the frequency of electromagnetic radiation becomes a function of temperature (Masood, 2018).

QED interactions also enter nonperturbative hadronic calculations. In lattice QCD plus quenched QED, heavy-meson masses receive few-MeV QED corrections arising almost entirely from valence-quark electric charges (Hatton et al., 2020). By comparing physical and sign-flipped valence charges, the two-body interaction term

$e\,\bar\psi\gamma^\mu\psi\,A_\mu$4

is exactly isolated to $e\,\bar\psi\gamma^\mu\psi\,A_\mu$5 (Hatton et al., 2020). In the nonrelativistic limit $e\,\bar\psi\gamma^\mu\psi\,A_\mu$6, and the continuum-extrapolated values reported are $e\,\bar\psi\gamma^\mu\psi\,A_\mu$7 for $e\,\bar\psi\gamma^\mu\psi\,A_\mu$8, $e\,\bar\psi\gamma^\mu\psi\,A_\mu$9 for $\frac12$0, and $\frac12$1 for $\frac12$2, corresponding to effective sizes $\frac12$3, $\frac12$4, and $\frac12$5 (Hatton et al., 2020).

A broader usage of the acronym appears in circuit QED. For a flux qubit interacting with an LC resonator, the closed-system Hamiltonian is the quantum Rabi model,

$\frac12$6

with ultrastrong coupling defined by $\frac12$7 and deep-strong coupling by $\frac12$8 (Napoli et al., 2024). In this regime counter-rotating terms cannot be dropped, Bloch–Siegert shifts appear, Jaynes–Cummings selection rules break down, and spectral line intensities depend on whether the device is read out through a mutual or capacitive port (Napoli et al., 2024). This suggests that methods and concepts originating in QED have become templates for experimentally controlled light–matter systems whose spectra probe beyond-RWA physics.

7. Historical critiques, debates, and conceptual open questions

The historical and conceptual status of QED remains the subject of debate. One paper in the source set argues that the history of QED includes “errors, suspicious coincidences, mathematical inconsistencies and renormalized infinities swept under the rug,” focusing especially on the Karplus–Kroll two-loop calculation of the electron anomalous magnetic moment and later recalculations of higher-order terms (Consa, 2021). The same source emphasizes that QED’s formal consistency remains unproven, citing Dyson’s divergence argument and the Landau pole, and calls for transparent publication of intermediate steps and codes (Consa, 2021).

The same source also records the mainstream response: defenders of QED note that the theory’s predictive accuracy now reaches parts in $\frac12$9 for $0+1$0, and that independent cross checks, analytic and numerical, have converged on the same results in many cases (Consa, 2021). It concludes that most mainstream physicists would characterize the historical “scandals” as part of the normal process of error-checking at the frontier of multi-loop perturbative calculations rather than evidence of a deep flaw in QED itself (Consa, 2021).

A second conceptual issue is the relation between practical success and asymptotic perturbation theory. The Dyson–Chandrasekhar estimate of an optimal truncation order near $0+1$1 in QED implies that the observed stability of presently computed orders is not in tension with asymptoticity (Kolomeisky, 2014). A plausible implication is that QED occupies a distinctive position: it is simultaneously an empirically unmatched perturbative framework, a theory with nontrivial infrared and renormalization subtleties, and a source of ongoing foundational discussion about what counts as completeness, rigor, and observability in quantum field theory.