The Uniqueness Theorem
Chapter 3 — Conductors
When you solve for the field in a region containing conductors, you often find yourself guessing — assuming a symmetric charge distribution, placing an image charge, picking a potential that looks right. The uniqueness theorem is your license to do that: if your guess satisfies the equations and the boundary conditions, it's the answer.
Statement
In a region of space where charge density \(\rho(\vec r)\) is specified, the potential \(V(\vec r)\) is uniquely determined by any one of the following on the boundary:
- Dirichlet condition: the value of \(V\) everywhere on the boundary, or
- Neumann condition: the normal derivative \(\partial V/\partial n\) (equivalently \(E_\perp\), or equivalently the surface charge) everywhere on the boundary — this one is unique up to an additive constant.
In particular: if \(N\) conductors are held at fixed potentials \(V_1,\dots,V_N\) and the charge density outside them is given, then the field everywhere is determined.
Sketch of proof (Dirichlet version)
Suppose two candidate potentials \(\phi\) and \(\psi\) both solve Poisson's equation \(\nabla^2\phi = -\rho/\epsilon_0\) with the same boundary values. Let \(W = \phi - \psi\). Then:
- \(\nabla^2 W = \nabla^2\phi - \nabla^2\psi = 0\) — so \(W\) satisfies Laplace's equation.
- \(W = 0\) on every boundary (both candidates match the same BCs).
- Solutions of Laplace's equation have no interior maxima or minima (mean-value property). So \(W\)'s extremes occur on the boundary — where \(W = 0\). Therefore \(W \equiv 0\) everywhere.
Conclusion: \(\phi = \psi\). The solution is unique.
Why this is so useful
The theorem legitimizes the "guess and check" style of conductor problem-solving. In particular, it's the foundation of the method of images: replace a conductor with fictitious point charges that reproduce the same boundary condition, and by uniqueness you've solved the original problem in the physical region.
Reverse method of images
Going the other direction: pick any nice charge distribution, compute its equipotentials, then "replace" one of those equipotential surfaces with a conductor at that potential. Outside the conductor, the field is unchanged. This is a useful trick for designing electrode shapes.
A corollary: Earnshaw revisited
You can't stably trap a point charge using only other static charges and conductors. A local potential minimum would require \(\nabla^2 V > 0\), impossible in a charge-free region. The uniqueness + maximum-principle structure rules it out.
Practice Problems
Hint
Solution
Inside the cavity, Laplace's equation \(\nabla^2 V = 0\) holds, and the boundary (the inner wall of the conductor) is at some constant potential \(V_0\). The function \(V(\vec r) \equiv V_0\) satisfies both. By uniqueness it's the only solution, so \(\vec E = -\nabla V = 0\) throughout the cavity.
Answer: \(\vec E = 0\) inside, regardless of external field.
Hint
Solution
Between the shells, \(\nabla^2 V = 0\) and by symmetry \(V\) depends only on \(r\). The radial Laplacian gives \(\frac{1}{r^2}\frac{d}{dr}(r^2 V') = 0\), with general solution \(V(r) = A + B/r\). Applying \(V(R_1) = V_1\) and \(V(R_2) = V_2\) fixes \(A, B\). This satisfies both the PDE and the BCs, so by uniqueness it's the answer.
Answer: \(V(r) = \dfrac{V_1(R_1^{-1}-R_2^{-1})^{-1}(1/r - 1/R_2) + V_2\cdots}{\cdots}\) — any guess of the form \(A + B/r\) that matches the two endpoints works.
Hint
Solution
\(W = \phi_1-\phi_2\) satisfies \(\nabla^2 W = 0\) and \(\partial W/\partial n = 0\) on the boundary. The identity \(\int (W\nabla W)\cdot d\vec A = \int (|\nabla W|^2 + W\nabla^2 W)\,dV\) becomes \(0 = \int |\nabla W|^2\,dV\), forcing \(\nabla W = 0\). So \(W\) is constant.
Answer: \(\phi_1 - \phi_2 = \text{const}\).
Hint
Solution
Inside the cavity, Poisson's equation reads \(\nabla^2 V = -q\delta(\vec r)/\epsilon_0\) (just a point source). A function of the form \(V(r) = \frac{1}{4\pi\epsilon_0}\frac{q}{r} + C\) solves this. To hit \(V=0\) at \(r=R\) (cavity radius), pick \(C = -\frac{q}{4\pi\epsilon_0 R}\). Uniqueness guarantees no other function does the job.
Answer: \(V(r) = \frac{q}{4\pi\epsilon_0}\left(\frac{1}{r} - \frac{1}{R}\right)\) for \(r < R\).