I encountered a problematic statement about electromagnetic radiation and I would be grateful if someone could shed some light on it.

The situation is the following:

Flat Minkowski space in 4D, with coordinates $(u,r,z,\bar z)$, where $(z,\bar z)$ are (complex) angular coordinates (whose exact form is not of interest here), $r$ is the usual radial distance and $u=t-r$ is the retarded coordinate. Future null infinity $\mathscr{J}^+$ is the surface at $r=+\infty$ with coordinates $(u,z,\bar z).$ The gauge is $A_r=0$ everywhere, and $A_u=0$ (on $\mathscr{J}^+$).

The problematic statement is the following:

radiation flux through $\mathscr{J}^+$ is proportional to $\int_{\mathscr{J}^+} {F_u}^z{F_u}_z $,

encountered in the article "New symmetries in massless QED" (here).

So my questions are:

  1. How is the integrand related to radiation? The closest thing I can imagine is the $T_{uu}$ component of the stress-energy tensor (without the term $g_{uu}(F)^2=g_{uu}(B^2-E^2)$ using the approximation $B^2\sim E^2$. But, again, I can't manage to link it to radiation.

  2. It looks to me that we're missing the volume element of the surface, which turns out to be $r^2 \gamma_{z,\bar z}$ with $\gamma$ a function of $(z,\bar z)$. Since this integral is later asked to be nonzero and finite to impose some asymptotic behaviour on potentials $A$, the $r^2$ should be important here. Am I right on this?

I would be very grateful for any help on this. This is my first question so forgive me if I've missed something. Suggestions are also welcome.

  • 2
    $\begingroup$ Good question. Welcome to the site! $\endgroup$ Aug 2, 2016 at 10:32

1 Answer 1


I think it's solved now. I'll write the reasoning as a reference.

Energy is the conserved quantity associated to invariance under time translation.
In that coordinates I would say that a time translation is represented by a vector parallel to $\partial_u$, since it is $u$ that has the temporal role here.
The conserved current associated to this translation (energy density current) is $T_{\mu \nu} X^{\nu}$ with $X$ parallel to $\partial_u$. Its flux through $\mathscr{J}^+$ gives the energy flux through that surface. The flux is:

$ \int_{\mathscr{J}^+} * ( T_{\mu\nu} X^{\nu} dx^{\mu})$

which, using $dr=0$ on that surface, and the form of the metric, is written as

$\int_{\mathscr{J}^+} (-T_{uu} + T_{ur} )r^2 \gamma \, du dzd\bar z $

As in
$-T_{uu} + T_{ur} = -F_{u a}F_{z}^a - F_{ua}F_r^a -\frac{1}{4}(g_{uu}-g_{ru})F^2 $
the term with $F_r^a$ is subleading and $g_{uu}=g_{ur}$, one obtains what originally asked for in the question, complete with the volume element that seemed missing from the quoted expression.

  • $\begingroup$ Hello. I have a follow up question related to this article, if anyone can reply. Why is the charge associated to the large gauge transformation defined on $\mathcal{I}^{+}_{-}$ only, instead of let's say being the zeroth component of the current density associated to the large gauge transformation inside a three dimensional space?? Thanks in advance $\endgroup$
    – schris38
    May 11, 2022 at 17:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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