Conductors, capacitance, and electrostatic energy

Intuition Beginner

A conductor is a material in which charges move freely. Metals such as copper, aluminium, and silver behave this way because their outermost electrons are not bound to individual atoms but float through the lattice. Push a charge into one end of a wire and the imbalance equalises within nanoseconds: electrons shuffle until no net force pushes them around.

The endpoint of that shuffling is electrostatic equilibrium. In equilibrium the electric field inside the conductor must be zero — if any field remained, the free charges would still feel a push and would still be moving, contradicting equilibrium. So the inside is field-free. The surface is then an equipotential: every point on it sits at the same voltage, because moving along the surface costs no work.

Any extra charge you deposit on a conductor migrates to its outer surface. The interior stays neutral; the skin carries it all. This is why a hollow metal sphere shields its interior from outside fields and why standing inside a metal cage during a lightning strike is safer than standing in the open.

A capacitor is two conductors holding equal and opposite charges. The voltage between them is proportional to the charge: doubling the charge doubles the voltage. The proportionality constant is the capacitance $C=Q/V$, measured in farads. Capacitance depends only on geometry (and on what is between the plates), not on how much charge you put on. Storing charge in a capacitor also stores energy; a charged capacitor can release that energy in a flash, as in a camera strobe.

Visual Beginner

Picture a flat metal plate placed in a uniform horizontal electric field pointing rightward. The field would push positive charge rightward and negative charge leftward, so the free electrons inside the plate immediately drift leftward, leaving a positive shadow on the right edge and a negative pile on the left edge. The induced surface charges create their own field, pointing leftward inside the plate, exactly cancelling the external field. Inside the plate the total field is zero. Outside, the original field is bent so that it meets the plate at right angles — the surface is equipotential, so the field has no tangential component there.

The picture captures three facts that all conductors obey in equilibrium: the field inside is zero, the surface is equipotential, and the field just outside is perpendicular to the surface with magnitude $\sigma /{ϵ}_0$, where $\sigma$ is the local surface charge per unit area.

Worked example Beginner

A simple parallel-plate capacitor has two flat metal squares of side $1\mathrm{cm}$ separated by a gap of $1\mathrm{mm}$ of air. Charge it to a voltage of $100\mathrm{V}$. How much charge sits on each plate, and how much energy is stored?

Step 1. The plate area is $A=(0.01{)}^2=1.0\times 10^{-4}{\mathrm{m}}^2$ and the gap is $d=1.0\times 10^{-3}\mathrm{m}$.

Step 2. The capacitance of a parallel-plate capacitor with vacuum (or air) between the plates is $C={ϵ}_{0}A/d$. With ${ϵ}_0=8.85\times 10^{-12}F/m$:

$$C=\frac{(8.85\times 10^{-12})(1.0\times 10^{-4})}{1.0\times 10^{-3}}=8.85\times 10^{-13}\mathrm{F}\approx 0.88\mathrm{pF}.$$

Step 3. The charge on each plate at $V=100\mathrm{V}$ is $Q=CV=(8.85\times 10^{-13})(100)=8.85\times 10^{-11}\mathrm{C}$, or about $89\mathrm{pC}$.

Step 4. The energy stored is $U=\frac{1}{2}C{V}^2=\frac{1}{2}(8.85\times 10^{-13})(100{)}^2=4.4\times 10^{-9}\mathrm{J}$, or about $4.4\mathrm{nJ}$.

What this tells us: tiny capacitors store only nanojoules at modest voltage, which is why a camera flash uses a capacitor a million times larger ($\sim 1\mathrm{mF}$) charged to a few hundred volts to deliver millijoules in a single discharge.

Check your understanding Beginner

Formal definition Intermediate+

A conductor is a region $\Omega \subset {\mathbb{R}}^3$ in which mobile charges respond to any electric field without resistance. In electrostatic equilibrium, the field inside vanishes, and the boundary of $\Omega$ becomes an equipotential surface.

Boundary conditions at a conductor surface. Let $S$ be the surface of a conductor with outward unit normal $\hat{\mathbf{n}}$, and let $\sigma$ denote the surface charge per unit area. Inside the conductor, $\mathbf{E}=0$. Just outside, the field decomposes into a tangential part ${\mathbf{E}}_{\parallel }$ along $S$ and a normal part $E_n\hat{\mathbf{n}}$. The equipotential condition forces ${\mathbf{E}}_{\parallel }=0$. Applying Gauss's law to a pillbox straddling $S$ gives the normal component:

$$E_n=\frac{\sigma }{{ϵ}_0},{\mathbf{E}}_{\parallel }=0.$$

The potential on $S$ is constant: $V{|}_S=V_S$.

Capacitance. Let ${\Omega }_1,\dots ,{\Omega }_N$ be $N$ disjoint conductors with surfaces $S_i$, held at potentials $V_i$. Each carries total charge $Q_i={\oint }_{S_i}\sigma dA$. Because Laplace's equation ${\nabla }^{2}V=0$ in the exterior region is linear, the charges depend linearly on the potentials:

$$Q_i=\sum _{j=1}^N{C}_{ij}{V}_j.$$

The matrix $C_{ij}$ is the capacitance matrix of the conductor system. For a pair of conductors carrying $\pm Q$, the capacitance is $C=Q/(V_1-V_2)$.

Definition (parallel-plate capacitor). Two parallel conducting plates of area $A$ separated by distance $d\ll \sqrt{A}$, with vacuum between them, have capacitance $C={ϵ}_{0}A/d$.

Definition (electrostatic energy). The energy stored in a static electric field on a region $\Omega$ is

$$U=\frac{{ϵ}_0}{2}{\int }_{\Omega }|\mathbf{E}{|}^2{d}^{3}r.$$

For a system of conductors at potentials $V_i$ carrying charges $Q_i$, this is equal to $U=\frac{1}{2}{\sum }_i{Q}_i{V}_i$. For a single capacitor, $U=\frac{1}{2}C{V}^2=Q^2/(2C)=\frac{1}{2}QV$.

Counterexamples to common slips

• The interior of a conductor is field-free, but the interior of a cavity inside a conductor need not be. If a charge sits in a cavity carved out of a metal block, the field inside the cavity is non-zero; only the bulk metal is field-free. The induced charge on the cavity wall, however, has total magnitude equal to the cavity charge with opposite sign — Gauss's law applied to a surface in the bulk metal forces this.
• Capacitance is a geometric quantity, not a dynamical one. It does not depend on $Q$ or $V$; it depends only on the shape and arrangement of the conductors and on the permittivity of the medium between them. Doubling $Q$ doubles $V$ leaving $C=Q/V$ unchanged.
• The energy $U=\frac{1}{2}C{V}^2$ and the work $W=QV$ done by an external battery to charge the capacitor differ by a factor of two. The battery does work $W=QV$ pushing charge through potential difference $V$; only half ends up as field energy, the other half is dissipated as heat in the connecting wires (or radiated, if the charging is done abruptly). The factor of $\frac{1}{2}$ comes from the fact that during charging, the voltage rises linearly from $0$ to $V$ so the average voltage seen by an arriving charge is $V/2$.

Key theorem with proof Intermediate+

Theorem (uniqueness of the electrostatic field with conductor boundary conditions). Let $\Omega \subset {\mathbb{R}}^3$ be a region whose boundary consists of conductor surfaces $S_1,\dots ,S_N$ plus a far boundary at infinity. Suppose either (a) the potentials $V_i$ on each $S_i$ are prescribed (Dirichlet data), or (b) the total charges $Q_i$ on each $S_i$ are prescribed (mixed data). Then the electrostatic potential $V$ in $\Omega$ is determined uniquely up to an additive constant in case (b), and uniquely in case (a) once the far-field decay is specified.

Proof. Suppose $V_1$ and $V_2$ both satisfy Laplace's equation ${\nabla }^{2}V=0$ in $\Omega$ together with the boundary conditions. Let $W=V_1-V_2$. Then ${\nabla }^{2}W=0$ in $\Omega$.

Case (a): Dirichlet. $W$ vanishes on each $S_i$ and decays at infinity. By the maximum principle for harmonic functions (a consequence of the mean-value property, unit 02.05.02), $W$ attains its extrema on the boundary. The boundary value is zero, so $W\equiv 0$ throughout $\Omega$.

Case (b): mixed. On each $S_i$, $W$ is some constant $W_i$ (the difference of two equipotential values). The total charge ${\oint }_{S_i}\sigma dA=-{ϵ}_0{\oint }_{S_i}{\partial }_{n}VdA$ is the same for $V_1$ and $V_2$, so ${\oint }_{S_i}{\partial }_{n}WdA=0$ for each $i$.

Apply Green's first identity to $W$ on the region $\Omega$:

$${\int }_{\Omega }|\nabla W{|}^2{d}^{3}r={\oint }_{\partial \Omega }W{\partial }_{n}WdA.$$

The far-field integral vanishes provided $W\to 0$ at infinity (in case (a)) or $W\to$ constant fast enough (in case (b)). On each $S_i$, $W=W_i$ is constant, so

$${\oint }_{S_i}W{\partial }_{n}WdA=W_i{\oint }_{S_i}{\partial }_{n}WdA=0$$

by the charge constraint. Thus the right-hand side of Green's identity is zero, which forces

$${\int }_{\Omega }|\nabla W{|}^2{d}^{3}r=0.$$

Since $|\nabla W{|}^2\ge 0$ and is continuous, $\nabla W\equiv 0$, so $W$ is constant on $\Omega$. In case (a) the boundary constraint pins the constant to zero; in case (b) the potential is unique up to an overall additive constant. $■$

Bridge. Uniqueness is the foundational reason that the capacitance matrix $C_{ij}$ is well-defined: any other potential satisfying the same boundary data would give the same charges, so the linear map $V_j\mapsto Q_i$ is single-valued. This is exactly the structural input that builds toward 10.01.04 dielectric capacitors, where the same Green's-identity argument runs through the modified permittivity $ϵ(\mathbf{r})$ inside the dielectric. The bridge to 24.01.03 weak variational formulations is direct: the Dirichlet integral $\int |\nabla W{|}^2{d}^{3}r$ in the proof is exactly the energy functional whose minimisation is the variational version of Laplace's equation, and the boundary-condition handling generalises to mixed boundary-value problems on Sobolev spaces. Putting these together, uniqueness underwrites the entire boundary-element-method industry: given prescribed conductor potentials, the field is determined, so numerical solvers can compute capacitance matrices for arbitrary geometry without ambiguity.

Exercises Intermediate+

Advanced results Master

Theorem 1 (Green's reciprocation theorem, Maxwell 1873). Let two systems of charges $\{Q_i\}$ and $\{Q_i^{\prime }\}$ on the same set of conductors produce potentials $\{V_i\}$ and $\{V_i^{\prime }\}$, respectively. Then

$$\sum _i{Q}_i{V}_i^{\prime }=\sum _i{Q}_i^{\prime }{V}_i.$$

Proof. Multiply Poisson's equation for the first system ${\nabla }^{2}V=-\rho /{ϵ}_0$ by $V^{\prime }$, integrate over all space, and use Green's second identity:

$$\int (V^{\prime }{\nabla }^{2}V-V{\nabla }^2{V}^{\prime })d^{3}r={\oint }_{\partial }(V^{\prime }{\partial }_{n}V-V{\partial }_n{V}^{\prime })dA=0,$$

with the surface integral at infinity vanishing for a localised system. The volume integral gives $-\int (V^{\prime }\rho -V{\rho }^{\prime })/{ϵ}_0{d}^{3}r=0$, i.e., $\int V^{\prime }\rho d^{3}r=\int V{\rho }^{\prime }{d}^{3}r$. For point or surface charges concentrated on conductors at fixed potential, $\int V^{\prime }\rho d^{3}r={\sum }_i{V}_i^{\prime }{Q}_i$ and similarly on the other side. $■$

Theorem 2 (Symmetry and positive-definiteness of the capacitance matrix). The capacitance matrix $C_{ij}$ defined by $Q_i={\sum }_j{C}_{ij}{V}_j$ satisfies $C_{ij}=C_{ji}$ and is positive-definite.

Symmetry is a corollary of Theorem 1: set $V_k={\delta }_{kj}$ in one system and $V_k^{\prime }={\delta }_{ki}$ in the other to obtain $C_{ij}=C_{ji}$. Positive-definiteness follows from the energy identity $U=\frac{1}{2}{\sum }_{i,j}{C}_{ij}{V}_i{V}_j=\frac{{ϵ}_0}{2}\int |\mathbf{E}{|}^2{d}^{3}r\ge 0$, with equality only when $\mathbf{E}\equiv 0$, which requires all $V_i$ equal (and zero at infinity), hence all $V_i=0$.

Theorem 3 (Thomson's minimum-energy theorem, Kelvin 1845). Among all conceivable distributions of a given total charge $Q$ on a conductor, the one that minimises the total electrostatic energy is the equilibrium distribution — the one in which charges lie on the outer surface and produce zero field inside.

Proof sketch. Consider any distribution $\rho$ on the conductor and write the energy $U[\rho ]=(1/2)\int \rho V{d}^{3}r$ subject to the constraint $\int \rho d^{3}r=Q$. Introduce a Lagrange multiplier $\mu$ and vary $\rho \to \rho +\delta \rho$ with $\int \delta \rho d^{3}r=0$. The variation of the energy is $\delta U=\int V\delta \rho d^{3}r+(1/2)\int \rho \delta V{d}^{3}r$, but by reciprocity (Theorem 1 applied to $\delta \rho$ and $\rho$) the second term equals the first, so $\delta U=\int V\delta \rho d^{3}r$. Stationarity with the constraint forces $V$ to be constant on the conductor — the equipotential condition. The second variation ${\delta }^{2}U=({ϵ}_0/2)\int |\delta \mathbf{E}{|}^2{d}^{3}r>0$ shows the stationary point is a minimum. $■$

Theorem 4 (Dirichlet principle for the Laplace equation, Riemann 1851, Hilbert 1904). Let $\Omega$ be a bounded domain in ${\mathbb{R}}^3$ with smooth boundary $\partial \Omega$, and let $g$ be a continuous function on $\partial \Omega$. The unique solution to the Dirichlet problem ${\nabla }^{2}V=0$ in $\Omega$, $V{|}_{\partial \Omega }=g$, minimises the Dirichlet integral

$$\mathcal{D}[V]={\int }_{\Omega }|\nabla V{|}^2{d}^{3}r$$

over all sufficiently smooth functions agreeing with $g$ on $\partial \Omega$.

The Dirichlet principle was used informally by Riemann in his 1851 thesis on Abelian functions; Weierstrass's 1870 critique pointed out that the existence of a minimiser must be proved separately. Hilbert restored the principle in 1904 by establishing the necessary lower-semicontinuity and existence theorems in his work on the Direct Method in the Calculus of Variations. The modern statement uses Sobolev spaces $H^1(\Omega )$ as the natural function class. In this language, capacitance is a quadratic functional on $H^1$, the equilibrium charge distribution is its minimiser, and the Dirichlet-to-Neumann map $V{|}_{\partial \Omega }\to {\partial }_{n}V{|}_{\partial \Omega }$ is a continuous self-adjoint operator on a trace space.

Theorem 5 (Capacitance as boundary-to-Neumann map). Let ${\Omega }^c$ be the exterior region of a system of conductors $\{S_i\}$ in ${\mathbb{R}}^3$. The map $\Lambda :V{|}_{S_i}\to {\partial }_{n}V{|}_{S_i}$ (the Dirichlet-to-Neumann map of the Laplacian on ${\Omega }^c$ with decay at infinity) is symmetric and positive-definite, and its quadratic form integrated over each $S_i$ gives the capacitance matrix entries.

The boundary-element method (BEM) for capacitance extraction in VLSI design implements this directly: parameterise $V$ on each $S_i$ as a piecewise constant, assemble the Dirichlet-to-Neumann map as a matrix, and read off $C_{ij}$ from the boundary integrals. The numerical cost is governed by the conditioning of the boundary integral operator, which Stratton-Chu and Costabel-Stephan analysed in the 1970s and 80s.

Theorem 6 (Polya-Szegő isoperimetric inequality for capacity, Polya-Szegő 1951). Among all compact bodies $K\subset {\mathbb{R}}^3$ of fixed volume $V$, the ball minimises the self-capacitance $C(K)$. Quantitatively,

$$C(K)\ge 4\pi {ϵ}_0{\left(\frac{3V}{4\pi }\right)}^{1/3},$$

with equality iff $K$ is a ball.

The Polya-Szegő symmetrisation argument runs as follows: capacity is monotone-decreasing under Schwarz spherical rearrangement of the conductor, because rearranging concentrates the indicator function while preserving volume and decreasing the Dirichlet energy of the equilibrium potential. The same rearrangement technique establishes related inequalities for the principal eigenvalue of the Laplacian (Faber-Krahn 1923), for torsional rigidity (Saint-Venant 1856), and for the heat content of a body. Polya and Szegő's 1951 monograph remains the canonical reference; modern presentations in geometric measure theory streamline the rearrangement proofs but do not change the conclusion.

Theorem 7 (Conformal invariance of capacity in two dimensions). In two dimensions the logarithmic capacity $\mathrm{cap}(K)$ of a compact set $K\subset \mathbb{C}$ is invariant under conformal maps that fix infinity, and equals $\exp (-V_K)$ where $V_K$ is the Robin constant of $K$ — the value at infinity of the Green's function with pole at infinity for the complement of $K$.

Two-dimensional capacity is the conformal-geometric notion underlying potential theory on Riemann surfaces. It connects via the Riemann mapping theorem to the conformal radius and via Wiener's criterion to regularity of boundary points for the Dirichlet problem. Hilbert's 1909 paper "Zur Theorie der konformen Abbildung" established the modern conformal framework; Tsuji's 1959 Potential Theory in Modern Function Theory is the standard monograph.

Synthesis. The capacitance matrix is the foundational reason that an arbitrary system of conductors admits a finite-dimensional electrostatic description: $N$ conductors give $N^2$ matrix entries, of which $N(N+1)/2$ are independent by symmetry (Theorem 2), and the symmetric positive-definite quadratic form $U(V)=\frac{1}{2}\sum C_{ij}{V}_i{V}_j$ is the entire energy content of the system. The central insight is that $C_{ij}$ is exactly the Dirichlet-to-Neumann map of the exterior Laplace operator (Theorem 5), so capacitance generalises through every linear elliptic boundary-value problem with conductor-style boundary conditions. Putting these together with the Dirichlet principle (Theorem 4), the bridge is between an analytic boundary-value problem and a finite-dimensional quadratic minimisation — the same identification that lets BEM solvers compute capacitances by minimising a discrete energy.

The pattern recurs across mathematical physics: Thomson's theorem (Theorem 3) appears again in 05.01.01 as Hamilton's principle for the action functional and in 24.01.03 as the variational formulation of weak elliptic problems; Polya-Szegő symmetrisation (Theorem 6) is dual to the Brunn-Minkowski inequality in convex geometry and identifies the ball as the extremal body for many isoperimetric quantities; conformal invariance in 2D (Theorem 7) builds toward potential theory on Riemann surfaces and the Bergman kernel. The capacitance matrix sits at the intersection of analysis, geometry, and numerics, and its symmetry is exactly the self-adjointness of the boundary-Laplace operator.

Full proof set Master

Proposition 1 (Spherical capacitor). Two concentric conducting spheres of radii $R_a<R_b$ with vacuum between, held at $V_a$ and $V_b$ respectively, have mutual capacitance $C=4\pi {ϵ}_0{R}_a{R}_b/(R_b-R_a)$.

Proof. Let the inner sphere carry charge $+Q$ and the outer sphere $-Q$. By spherical symmetry and Gauss's law, the field in $R_a<r<R_b$ is $E(r)=Q/(4\pi {ϵ}_0{r}^2)$. The potential difference is

$$V_a-V_b={\int }_{R_a}^{R_b}E(r)dr=\frac{Q}{4\pi {ϵ}_0}\left(\frac{1}{R_a}-\frac{1}{R_b}\right)=\frac{Q(R_b-R_a)}{4\pi {ϵ}_0{R}_a{R}_b}.$$

So $C=Q/(V_a-V_b)=4\pi {ϵ}_0{R}_a{R}_b/(R_b-R_a)$. $■$

Proposition 2 (Energy stored equals $\frac{1}{2}\sum Q_i{V}_i$). For a system of $N$ conductors at potentials $V_i$ carrying charges $Q_i$,

$$\frac{{ϵ}_0}{2}{\int }_{{\Omega }^c}|\mathbf{E}{|}^2{d}^{3}r=\frac{1}{2}\sum _{i=1}^N{Q}_i{V}_i.$$

Proof. Using $\mathbf{E}=-\nabla V$:

$${\int }_{{\Omega }^c}|\mathbf{E}{|}^2{d}^{3}r={\int }_{{\Omega }^c}\nabla V\cdot \nabla V{d}^{3}r.$$

Apply Green's first identity to the function $V$ on the exterior region ${\Omega }^c$:

$${\int }_{{\Omega }^c}\nabla V\cdot \nabla V{d}^{3}r+{\int }_{{\Omega }^c}V{\nabla }^{2}V{d}^{3}r=-\sum _{i=1}^N{\oint }_{S_i}V{\partial }_{n}VdA,$$

where the normal ${\partial }_n$ on each $S_i$ points out of the exterior region, i.e., into the conductor; the minus sign on the right comes from this orientation. Since ${\nabla }^{2}V=0$ in ${\Omega }^c$, the volume integral on the left reduces to $\int |\nabla V{|}^2{d}^{3}r$. On each $S_i$, $V=V_i$ is constant, and the inward normal derivative is $-E_n=-\sigma /{ϵ}_0$. So the right-hand side becomes

$$-\sum _i{V}_i{\oint }_{S_i}(-\sigma /{ϵ}_0)dA=\sum _i{V}_i{Q}_i/{ϵ}_0.$$

Multiplying by ${ϵ}_0/2$ gives the energy identity. $■$

Proposition 3 (Electrostatic pressure on a conductor surface). On a conductor surface with local surface charge density $\sigma$ and field $E=\sigma /{ϵ}_0$ just outside, the outward force per unit area is $f={\sigma }^2/(2{ϵ}_0)=\frac{1}{2}{ϵ}_0{E}^2$.

Proof. Consider a small surface element $dA$. The total electric field at the surface is $E=\sigma /{ϵ}_0$ outside, $0$ inside. Decompose this total field as the sum of the field $E_{\text{patch}}$ produced by the element $dA$ itself and the field $E_{\text{rest}}$ produced by everything else. By symmetry, $E_{\text{patch}}=\pm \sigma /(2{ϵ}_0)$ on the two sides of $dA$ (it behaves as an infinite charged plane on the scale of the element). For the total field to be $E$ outside and $0$ inside, we need $E_{\text{rest}}=+\sigma /(2{ϵ}_0)$ pointing outward on both sides.

The element $dA$ cannot exert net force on itself; it experiences only $E_{\text{rest}}$. The force on $\sigma dA$ is

$$dF=(\sigma dA)\cdot E_{\text{rest}}=\sigma \cdot \frac{\sigma }{2{ϵ}_0}dA=\frac{{\sigma }^2}{2{ϵ}_0}dA,$$

directed outward. So $f=dF/dA={\sigma }^2/(2{ϵ}_0)$. Substituting $E=\sigma /{ϵ}_0$ gives the equivalent form $f=\frac{1}{2}{ϵ}_0{E}^2$. $■$

Connections Master

• Laplace BVP 10.01.02. Capacitance is the canonical Dirichlet-to-Neumann data of the exterior Laplace problem; the uniqueness theorem proved here is the same one that underwrites separation-of-variables and method-of-images solutions in 10.01.02, and the capacitance matrix is the finite-dimensional summary of the Dirichlet-to-Neumann map.

• Dielectric polarisation 10.01.04. When the vacuum between two conductors is replaced by a linear dielectric of relative permittivity ${ϵ}_r$, the parallel-plate capacitance is multiplied by ${ϵ}_r$ and the field is reduced inside the dielectric. The Green's-identity argument for energy and uniqueness goes through unchanged with ${ϵ}_0\to ϵ(\mathbf{r})$, and the capacitance matrix becomes the Dirichlet-to-Neumann map of a divergence-form elliptic operator $\nabla \cdot (ϵ\nabla V)=0$.

• EM energy and Poynting 10.03.03. The energy density $u=\frac{1}{2}{ϵ}_0|\mathbf{E}{|}^2$ derived here is the static limit of the full electromagnetic energy density $u=\frac{1}{2}{ϵ}_0|\mathbf{E}{|}^2+|\mathbf{B}{|}^2/(2{\mu }_0)$; in the time-dependent theory it acquires a flux partner, the Poynting vector $\mathbf{S}=\mathbf{E}\times \mathbf{B}/{\mu }_0$, and the energy balance becomes Poynting's theorem.

• Weak variational formulation of elliptic PDE 24.01.03. The Dirichlet principle (Theorem 4) and the Green's-identity argument for energy and uniqueness are exactly the variational underpinning developed in 24.01.03: the electrostatic problem is the prototype linear elliptic boundary-value problem, and capacitance extraction by boundary element methods is its prototype numerical realisation.

• Method of images 10.01.06 pending (pending). Image charges are the explicit solution technique for capacitor problems involving planar or spherical conductor boundaries; they exploit the uniqueness theorem proved here to construct the field outside the conductor by replacing the conductor with one or more fictitious charges that enforce the equipotential boundary condition.

Historical & philosophical context Master

The capacitance concept emerged from the practical electrostatics of the 18th and early 19th centuries — Leyden jars, Volta's electroscope, condensers in early telegraphy — well before its mathematical formalisation. Cavendish in unpublished work circa 1771 (rediscovered by Maxwell in 1879 and edited as The Electrical Researches of the Honourable Henry Cavendish) measured capacitances of various conductors and obtained the inverse-square law independently of Coulomb. Faraday's experimental work in the 1830s introduced the dielectric concept and the term "capacity"; Maxwell systematised the theory in his Treatise on Electricity and Magnetism (1873) ^{[Maxwell1873]}, where Volume 1 Part I Chapter III gives the modern treatment of the capacitance matrix as a symmetric quadratic form.

Thomson (Lord Kelvin) introduced the method of images and the variational viewpoint in his 1845 paper ^{[Thomson1845]} in the Cambridge and Dublin Mathematical Journal. His minimum-energy theorem for the equilibrium distribution of charge on a conductor — that the actual charge configuration minimises the electrostatic energy at fixed total charge — anticipated the Dirichlet principle and supplied the first variational characterisation of equilibrium in field theory. Kirchhoff's 1845 work on current networks and Helmholtz's 1853 paper on the energy theorem for conductors completed the variational framework, which Hilbert in 1904 placed on rigorous existence-theoretic foundations through what he called the Direct Method of the calculus of variations.

The Polya-Szegő monograph Isoperimetric Inequalities in Mathematical Physics (1951) ^{[PolyaSzego1951]} in the Annals of Math Studies series brought together capacity, principal eigenvalues, torsional rigidity, and heat content under a single symmetrisation framework, establishing the ball as the extremal body for a wide class of physical-geometric functionals. The same techniques flow through Federer's geometric measure theory of the 1960s, Talenti's 1976 sharp constants for Sobolev inequalities, and modern symmetrisation in calculus of variations. The modern computational realisation — capacitance extraction in VLSI via boundary element methods — descends through Stratton-Chu's 1939 boundary-integral formulation, Costabel-Stephan's 1985 Math. Z. analysis of boundary integral operators, and Nabors-White's 1991 fast multipole accelerator.

Bibliography Master

@book{Maxwell1873,
  author    = {Maxwell, James Clerk},
  title     = {A Treatise on Electricity and Magnetism},
  volume    = {1},
  publisher = {Clarendon Press, Oxford},
  year      = {1873},
}

@article{Thomson1845,
  author    = {Thomson, William},
  title     = {Theorems with reference to the solution of certain partial differential equations},
  journal   = {Cambridge and Dublin Mathematical Journal},
  volume    = {1},
  year      = {1845},
  pages     = {75--95},
}

@book{PolyaSzego1951,
  author    = {Polya, George and Szeg{\H{o}}, Gabor},
  title     = {Isoperimetric Inequalities in Mathematical Physics},
  series    = {Annals of Mathematics Studies},
  number    = {27},
  publisher = {Princeton University Press},
  year      = {1951},
}

@book{Jackson1999,
  author    = {Jackson, John David},
  title     = {Classical Electrodynamics},
  edition   = {3},
  publisher = {Wiley},
  year      = {1999},
}

@book{Griffiths2013,
  author    = {Griffiths, David J.},
  title     = {Introduction to Electrodynamics},
  edition   = {4},
  publisher = {Cambridge University Press},
  year      = {2013},
}

@book{LandauLifshitzVol8,
  author    = {Landau, Lev Davidovich and Lifshitz, Evgeny Mikhailovich},
  title     = {Electrodynamics of Continuous Media},
  series    = {Course of Theoretical Physics},
  volume    = {8},
  edition   = {2},
  publisher = {Pergamon},
  year      = {1984},
}

@book{PurcellMorin2013,
  author    = {Purcell, Edward M. and Morin, David J.},
  title     = {Electricity and Magnetism},
  edition   = {3},
  publisher = {Cambridge University Press},
  year      = {2013},
}

@article{Hilbert1904,
  author    = {Hilbert, David},
  title     = {{\"U}ber das Dirichletsche Prinzip},
  journal   = {Mathematische Annalen},
  volume    = {59},
  year      = {1904},
  pages     = {161--186},
}