A complex frequency $\omega$ is related to dissipation. Is the idea that if your wave is described by $\mathbf A =\mathbf A_0 \exp(\mathrm i\omega t)$ as a function of time $t$, where $\mathbf A_0$ is an amplitude (vector if needed), then you can decompose it such that
$$\mathbf A=\mathbf A_0'(t) \exp[\mathrm i\Re(\omega) )t],$$
where $\Re$ is the real part, and
$$\mathbf A_0'(t)=\mathbf A_0\exp[-\mathrm \Im(\omega) t] ,$$
where $\Im$ is the imaginary part. The whole idea idea is that the wave now is described by a new amplitude $\mathbf A_0'(t)$ that decays exponentially with time (note the lack of imaginary $\mathrm i$).
Note: for this to work, you have to choose a convention, I choose here that the imaginary part of $\omega$ is always positive. The opposite convention exists too but you have to change the sign of the whole exponent.