I think there should be two ways of writing the equation for the impulsive spherical wave from an impulsive point source at the origin, say $\delta(t) \delta(r)$:

$$(4\pi ct)^{-1} \delta(r-ct) \tag{1}$$

$$(4\pi ct)^{-1} \delta(t-r/c) \tag{2}$$

But these are not the same. In (1) the Delta Function has dimension 1/[L] and in (2) the Delta Function has dimension 1/[T].


$$(4\pi ct)^{-1} \delta(r-ct) = (4\pi ct)^{-1} \delta(c(r/c-t)) =(4\pi ct)^{-1} (1/c)\delta((t-r/c))$$

so the two equations differ by (1/c).

My question is which is right and why?

Why would one be preferred over the other?


The wave from an impulsive point source is the Green's function $G(\vec{r},t)$ for the inhomogenous wave equation. In other words, if we wish to solve the equation $$ - \frac{1}{c^2} \frac{\partial^2 \psi}{\partial t^2} + \nabla^2 \psi = \rho(\vec{r},t), $$ we can do so by solving the related equation $$ - \frac{1}{c^2} \frac{\partial^2 G}{\partial t^2} + \nabla^2 G = \delta^3(\vec{r}) \delta(t). \tag{A} $$ in a distributional sense. By convention, the Green's function is usually taken to be the solution to this equation (in a distributional sense) with a delta function of unit weight on the right-hand side. This allows us to then find the solution of a more general inhomogeneous equation via integration: $$ \psi(\vec{r},t) = \iiiint d^3\vec{r}' dt' G(\vec{r} - \vec{r}',t-t') \rho(\vec{r}',t'). \tag{B} $$ (This can be proven by simply applying the wave operator to the right-hand side and bringing the derivatives inside the integral.) If we had used a different weight for the delta functions on the right-hand side, then this equation would require extra pre-factors to compensate.

So which one is right? Dimensionally, the right-hand side of (A) has dimensions of $[\text{length}]^{-3} [\text{time}]^{-1}$, and so by this logic the dimensions of $G$ should be $[\text{length}]^{-1} [\text{time}]^{-1}$. In your question, version (2) has the right units, and so it would be the correct Green's function for this equation.

Of course, if you defined your inhomogeneous wave equation differently---for example, something like $$ - \frac{\partial^2 \psi}{\partial t^2} + c^2 \nabla^2 \psi = \rho(\vec{r},t) $$ then the appropriate Green's function would be the solution to this equation with the right-hand side replaced by a unit-weight delta-function. Via the same dimensional arguments as above, the correct Green's function (so that the solution is still given by eq. (B)) would be $$ G(\vec{r},t) = \frac{1}{4\pi c^3 t} \delta(t-r/c). $$

| cite | improve this answer | |
  • $\begingroup$ Super answer--thanks. I am surprised that the Green's fn depends on the algebraic arrangement of the wave equation. $\endgroup$ – user45664 May 29 at 23:39
  • $\begingroup$ Very roughly speaking, the Green’s function acts as the “inverse” of a differential operator. From that perspective, it’s not entirely surprising that multiplying the operator by a constant changes the Green’s function, just as multiplying a matrix by a constant changes that inverse. $\endgroup$ – Michael Seifert May 30 at 2:25
  • $\begingroup$ I see--like in $TT^{-1}=\delta$ $\endgroup$ – user45664 May 30 at 18:17

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.