In the derivation, we assume the eigenfunctions of $H$ to have the form $$\psi (x)=e^{i \phi(x)/\hbar}$$ where $\phi (x)$ is allowed to take any complex value.
But then suddenly we assume this expansion of $\phi (x)$ in terms of $\hbar$:
$$\phi (x)=\phi _0 (x)+ \phi _1 (x)\hbar + \phi _2 (x) \hbar^2+....$$
We assume the higher powers of $\hbar$ are negligible (in our approximation) and then feed this into the Schrodinger equation.
My question is about the expansion. What on earth is this expansion? It's not a Taylor series or anything (where we can neglect higher order terms). How is this expansion of $\phi (x)$ even unique? What determines $\phi _0 (x)$, $\phi_1 (x)$, etc in this expansion? Can't we make this expansion whatever we want?