In a review paper of Whispering-gallery waves from A.N. Oraevsky, he writes the source-free monochromatic Maxwell's Equations as

$\nabla\times E = ikH$

$\nabla\times H = -ikE$

and he defines $k = (\omega/c)\sqrt{\mu\varepsilon}$ where $\mu$ and $\varepsilon$ are the magnetic permeability and electric permitivity respectively of the medium, $\omega$ is the frequency (rad/s) and $c$ is the speed of light in vacuum.

I am puzzled with the form Maxwell's Equations are written, since they should be

$\nabla\times E = i\omega\mu H$

$\nabla\times H = -i\omega\varepsilon E$

for a time dependence $e^{-i\omega t}$. His definition does not agree with cgs units neither. What is this different form of writing Maxwell Equations, and how can they be related to those written by me above?.

  • $\begingroup$ I've never seen that notation before, and it's definitely incorrect because both $k=\omega \mu $ and $k = \omega \epsilon$ have to be true at the same time. I suspect that the author has made a mistake. $\endgroup$
    – Judge
    Commented Jan 14, 2016 at 21:11
  • 1
    $\begingroup$ Maxwell's equations for the curls are $\nabla\times E = -\partial_t B$ and $\nabla\times B = \mu_0 (J + \epsilon_0\partial_t E)$. I neither understand your nor Oraesky's form of them, there is definitely some context lacking here. $\endgroup$
    – ACuriousMind
    Commented Jan 15, 2016 at 0:35
  • $\begingroup$ @ACuriousMind your equations are the same as mine. Since there is not sources $J=0$, $\partial_t \rightarrow -i\omega$ (harmonic monochromatic time-dependence), and of course $B=\mu H$. (Yours are in vacuum $\mu_0$, $\varepsilon_0$, mine are in a general medium $\mu$, $\varepsilon$) $\endgroup$ Commented Jan 15, 2016 at 11:59

1 Answer 1


The only explanation I found to this is the following:

From Maxwell's Equations in Lorentz-Heaviside units:

$\nabla\times E = (i\omega\mu/c)H$

$\nabla\times H = (-i\omega\varepsilon/c)E$

where now $\mu$ and $\varepsilon$ refer to the common relative permeability and permitivity respectively. Defining a sort of new $H' = \sqrt{\mu/\varepsilon} H$ then,

$\nabla\times E = (i\omega\sqrt{\mu\varepsilon}/c)H'$

$\nabla\times H' = (-i\omega\sqrt{\mu\varepsilon}/c)E$

which agrees with Oraevsky's equations. However, this requires to assume that $\nabla\times ( \sqrt{\varepsilon/\mu} H') = \sqrt{\varepsilon/\mu}\nabla\times H$ which is only valid for homogeneous media, and it does not seem to be the case for the context of the paper.


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