Maxwell equation in absence of charges and currents are

$$\nabla \cdot \bf{E} = 0 \\ \nabla\cdot B=0 \\ \nabla \times E=-\frac{\partial B}{\partial t} \\\nabla \times B=\mu \epsilon \frac{\partial E}{\partial t}$$

Wave equation is $$\nabla ^2 \bf{E}=\mu \epsilon \frac{\partial^2 E}{\partial t^2}\tag{1}$$

How can I prove that, given a solution $\bf{E}$ that satisfies both Maxwell equations and $(1)$, then the vector $\bf{E+E'}$ where $\bf{}{}E'$ is any vector such that $\nabla \times \bf{}{}E'=0$ (and not necessarily a solution of Maxwell equations) is a solution of $(1)$?

  • $\begingroup$ Your question doesn't quite make sense. In reciprocal space your condition $\nabla \times \vec{E}'=0$ can be written as $\vec{k} \times \vec{E}'=0$, which means that $\vec{E}'$ is parallel to the direction of the wave propagation, or zero. In both cases $\vec{E}'$ adds zero information to the radiation field. Also, if you consider conventional ED/QED, $\vec{E}'$ and $\vec{B}'$ in free space are equivalent, hence your new solution IS the solution of the Maxwell equations if it satisfies Helmholtz's/wave equation. $\endgroup$ – MsTais Apr 19 '17 at 14:59
  • $\begingroup$ The wave equation is found by algebraic manipulation of Maxwell's equations, so I'm not sure how you expect to find a solution consistent with one but not the other. $\endgroup$ – The Photon Apr 19 '17 at 16:04

You cannot prove it, since it isn't true.

The curl of your E-field would be unchanged $$\nabla \times (\vec{E} + \vec{E}^{\prime}) = \nabla \times \vec{E} + \nabla \times \vec{E}^{\prime} = \nabla \times \vec{E}$$

The left hand side of the wave equation is the Laplacian of the E-field $$\nabla^2 (\vec{E} + \vec{E}^{\prime}) = \nabla (\nabla \cdot (\vec{E} + \vec{E}^{\prime})) - \nabla \times (\nabla \times (\vec{E} + \vec{E}^{\prime})) $$ $$\nabla^2 (\vec{E} + \vec{E}^{\prime}) = \nabla(\nabla \cdot \vec{E}^{\prime})-\nabla \times (\nabla \times \vec{E}) $$ $$ \nabla^2 (\vec{E} + \vec{E}^{\prime}) = \nabla(\nabla \cdot \vec{E}^{\prime}) + \nabla^2 \vec{E} - \nabla (\nabla \cdot \vec{E}) = \nabla^2 \vec{E} + \nabla(\nabla \cdot \vec{E}^{\prime})$$

The right hand side of the wave equation becomes $$ \mu \epsilon \frac{\partial^2 (\vec{E}+\vec{E}^{\prime})}{\partial t^2} = \mu \epsilon \frac{\partial^2 \vec{E}}{\partial t^2} + \mu \epsilon \frac{\partial^2 \vec{E}^{\prime}}{\partial t^2} $$

Thus the only way that the vacuum wave equation is still satisfied is if

$$\nabla (\nabla \cdot \vec{E}^{\prime} )= \mu \epsilon \frac{\partial^2 \vec{E}^{\prime}}{\partial t^2} $$

This is not a general identity! For example, $\vec{E}^{\prime}= \vec{E}_0 \sin \omega t$ has a zero curl, a zero divergence, but a non-zero second time derivative. However, if the additional field is a solution to Maxwell's vacuum equations with zero curl, then this condition is automatically satisfied.

  • $\begingroup$ Thanks! If I may ask: writing $$\mu \epsilon \frac{\partial^2 (\vec{E}+\vec{E}^{\prime})}{\partial t^2} = -\nabla \times \frac{\partial \vec{B}}{\partial t}\tag{2}$$ you are assuming that $E+E'$ satisfies the last Maxwell equation, while it should not be something to assume, is that right? Secondly, non-zero divergence solutions of wave equations are the ones I was looking for (i.e. solutions of wave equation that do not obey the zero divergence Maxwell equations). So is it possible to have non zero divergence solutions $E+E'$ ($\nabla \cdot E' \neq 0$) that satisfy $(1)$ but not $(2)$? $\endgroup$ – Sørën Apr 20 '17 at 0:05
  • $\begingroup$ @Sørën You are right, I need to look again. $\endgroup$ – Rob Jeffries Apr 20 '17 at 6:59

I write a complement to the answer by Rob Jeffries. A possibility to add such a vector $\vec{E}'$ to $\vec{E}$ means that outside charges the total field is not obligatory transversal, but may contain (and contains) a longitudinal field. Normally it happens next to the radiating system of charges (a "near" field). Only far-far away the field becomes "radiated" and transversal.

From mathematical point of view, the equations are not sufficient for determining the field. One needs boundary conditions too.


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.