I have recently learned about the Israel junction conditions in GR (as explained in for example Gravitation by MTW). I then tried to generalize it when including Electromagnetism, i.e. matching two spacetimes over a junction but when there is some non-zero electromagnetic content in, but I ran into difficulties. Does anyone know where this has been studied in the literature?

Basically the problem I ran into was that the Einstein equations are schematically on the form


Here $T$ is a matter stress energy tensor for some delta function matter at the junction, $F$ is the electromagnetic field tensor and $G$ is the Einstein tensor. When deriving the Israel junction conditions we would integrate over a small slice over the junction and it turns out that the Einstein tensor has a delta function which comes from a discontinuity in the extrinsic curvature so we obtain relations between the jump of the extrinsic curvature and the stress energy tensor of the shell. However, no such delta function can appear in F^2 if we assume a discontinuity in e.g. the four potential. Differentiating A then gives a delta function in $E$ and/or $B$ but $F^2$ is then a delta function squared and contributes nothing.

If we argue that it is really $B$ and $E$ that are the physical quantities of interest and they can not possess a delta function, then $F^2$ would contribute zero to the junction integral since there are not derivatives wrt E or B. Thus the Israel junction conditions should be unchanged by adding electromagnetism, but I dont think this is the right answer?

  • $\begingroup$ Are you looking for a resource (given your first paragraph) or a resolution of your issue (given your last paragraph)? $\endgroup$
    – Kyle Kanos
    Commented Aug 26, 2014 at 13:49
  • $\begingroup$ Well, both or any of them would be great! But I was primarly looking for resources, but if I did some obvious mistake in my reasoning please point it out. $\endgroup$ Commented Aug 26, 2014 at 15:46

1 Answer 1


I believed I found the answer to my own question. My reasoning seems to be right, the junction conditions obtained by a small integration over the shell are not altered by adding electromagnetism. However, to find the dynamics, these equations must be appended by at least one more equation, for example the Hamiltonian constraint, and these are indeed changed when there is a background stress energy tensor for the matter content.


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.