I would like to understand the plane wave solution for the (3+1-dimensional) d'Alembert operator $$ \square = \nabla^2 - \frac 1{c^2}\frac{\partial^2}{\partial t^2}\tag{1} $$ in terms of its fundamental solutions. I know that the “function” (distribution) $$ G(\vec x, \, t) = \frac{c}{4\pi r}\delta(ct - r),\tag{2} $$ where $r = \lvert\vec x\rvert$, is a fundamental solution, that is, it satisfies $$ \square G(\vec x, \, t) = -\delta^3(\vec x)\delta(t)\tag{3}. $$ This fundamental solution can be seen as a mathematical statement of Huygens principle, since it states that the response to an instantaneous impulse at $t = 0$ and $\vec x = 0$ is a spherical propagation of that impulse with an attenuation given by the $1/r$ which ensures conservation of energy. So I wanted to see this principle in action in the simple example of a plane wave, but I'm having a hard time.
A plane wave $f(\vec x, \, t)$ propagating satisfies $\square f = 0$, but the fundamental solution method only allows me to find particular solutions of an inhomogeneous equation $\square f = g$, by convolution with the fundamental solution, i.e. $f = G \circledast g$. I figured that I could choose some singular $g$, i.e. $g \propto \delta(t)$, in order to have $g = 0$ most of the time. I thought that $g$ would then have the meaning of an “initial condition” or “initial impulse”. In particular, for a plane wave traveling in the $z$ direction, $g$ should also have translational symmetry with respect to $x$ and $y$, i.e. $$g(\vec x, \, t) = g(0, \, 0, \, z, \, t) \equiv g(z, \, t) = h(z)\delta(t).\tag{4}$$
But this doesn't seem to work since I'm not getting solutions of the form $g(ct - z)$.
Question: is this method correct in principle? i.e. is choosing a source proportional to $\delta(t)$ akin to choosing initial conditions? If not, is there a different way of doing that?