# Feasibility of the Diffusion Equation to Model Electric Charges in a Conductor

Background: Given that equilibrium solutions of charges in a spherical shell formed from a conductor satisfy the Poisson problem $$\nabla^2 \Phi = - \frac{\rho}{\epsilon_0}$$ as a corollary of the differential form of Gauß's Law $$\nabla \cdot E = \frac{\rho}{\epsilon_0}$$, would it be natural to model the events proceeding the equilibrium with the diffusion equation?

To make my question more rigorous, we prescribe the system with initial charge density $$\rho(r, \theta, \varphi)$$ that is not uniform. Without using a Galerkin method, does there exist an analytical solution for $$\Phi(r, \theta, \varphi, t)$$?

Pardon me if my intuition is grossly over simplified; however, could the transition to an equilibrium state be viewed as a progression to positive eigenvalues? Taking the diffusion pde in $$\mathbb{R}^3$$ ($$\partial_t u = k \nabla^2 u$$) to be parabolic with degenerate eigenvalues, would not the equilibrium in the Poisson equation represent an elliptic PDE ($$\nabla^2 \Phi = - \frac{\rho}{\epsilon_0}$$)? Additionally, the steady state implies that $$\partial_t u = 0$$, which enforces the requirement of being harmonic on the solution.

Question: Would the following initial value problem suffice for modeling the redistribution of the charge density and voltage to reach equilibrium. My system is underconstrained, which leads me to believe that I am neglecting some physical relationship.

$$\partial_t \Phi = \frac{-k \rho(r, \theta, \varphi, t)}{\epsilon_0}$$

$$\rho(r, \theta, \varphi, 0) = \rho_0(r, \theta, \varphi) : \mathbb{R} \times [0, 2 \pi) \times [0, 2 \pi) \to \mathbb{R}$$

I am struggling to find the evolution of $$\rho(r, \theta, \varphi, t)$$ in time. I understand that areas to which the $$\vec E$$-field lines converge will have increasing charge density, and the opposite will hold for the case of diverging lines. Could I take the divergence of the following expression for the $$\vec E$$-field? Would some clever application of Gauß's law suffice instead of evaluating the following integral?

$$\vec E(r, \theta, \varphi) = \int_0^{\pi} \int_0^{2 \pi} \int_{r_0}^{r_1} \frac{dq}{4 \pi \epsilon_0 d((r^*, \theta^*, \varphi^*), (r, \theta, \varphi))^{\frac{3}{2}}} \vec d((r^*, \theta^*, \varphi^*), (r, \theta, \varphi)) dr^* d \theta^* d\varphi^*$$

$$d$$ is the chordal distance and $$\vec d$$ is the chordal displacement. They will correspond to the euclidean distance between the two points on the sphere written in Cartesian coordinates.

Meanwhile: I have found a solution for Laplace's PDE on the spherical shell. I am now attempting to use a Green's function to reconstruct a solution to the particular equation.