Any problem in gauge theory can of course in principle be solved in any gauge (or, in the case of classical gauge theory, without using gauge potentials at all), but some gauges are much more useful than others. In gauge quantum field theory, there are a variety of gauges that are particularly useful in various situations: e.g. Coulomb gauge in the canonical formalism of QED, and various $R_\xi$ gauges (Landau, Feynman-'t Hooft, Yennie, etc.) in nonabelian gauge theory.

But in classical EM, I don't know of any situation in which any gauge other than Lorenz gauge is more practically useful than Lorenz gauge for calculations. By far the second-most widely used gauge after Lorenz gauge is Coloumb gauge. I've seen it used in two situations in classical EM, but in neither do I think it's more useful than Lorenz gauge:

  1. In electrostatic and/or magnetostatic situations. But in these situations Coulomb and Lorenz gauges are actually equivalent, because you don't need to worry about the retardation effects that distinguish them in general. So while it's conceptually simpler to use the Coulomb gauge-fixing condition, I don't really consider it to be "better" than Lorenz gauge since it results in the exact same gauge fields.

  2. I've seen an exercise in Jackson that uses Coulomb gauge in an electrodynamic situation in order to explicitly demonstrate that the apparently acausal behavior of Coulomb gauge actually leads to completely causal behavior in all the gauge-invariant quantities. While this is certainly a very useful pedagogical exercise, I remember the calculation being a bit messy, and I'm not sure if it would ever be the most efficient way to actually get a solution.

(I know that Lorenz gauge in incomplete and that further gauge-fixing can sometimes be useful, but I'm not counting that as a separate gauge.)

Are than any situations in classical E&M in which a gauge other than Lorenz gauge is strictly more convenient than Lorenz gauge?

  • 1
    $\begingroup$ Yes, Coulomb gauge is very useful in classical EM, it is easier to do certain calcs in that gauge. Also, I think S. Weinberg quantizes EM in the coulomb gauge in his text just to illustrate the differences in technique and interpretation. $\endgroup$
    – user196418
    Dec 15 '18 at 23:02
  • 1
    $\begingroup$ @ggcg David Tong also does this in his excellent lecture notes, for those interested in the two different cases and their subtleties. $\endgroup$
    – J. Murray
    Dec 16 '18 at 6:56

The Coulomb gauge is still very useful is situations where you have individual moving charges and radiation in the same region. The reason is that the Coulomb gauge splits the electric field ${\bf E}$ up into two pieces, corresponding to the pole and vortex sources. The electric field directly due to the charges is entirely given by through scalar potential as $-\nabla\Phi$. The ${\bf E}$ generated by a changing ${\bf B}$ is entirely given by $-\partial{\bf A}/\partial t$.

In other gauges, such as Lorenz gauge, the electrostatic part of ${\bf E}$ comes from a combination of $\Phi$ and the longitudinal part of ${\bf A}$. This separation is still simple in Fourier space, but it is fairly awkward in position space.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.