Induced Charge and Electrostatic Shielding

Slip your phone into a metal cookie tin, close the lid, and call it. It rings straight to voicemail. No exotic material, no power supply — a few hundred grams of tin-plated steel has just erased, in the volume it encloses, an electromagnetic environment thick with radio towers, Wi-Fi, and static from every doorknob in the building. The same trick, scaled up, is why a car struck by lightning protects its passengers and why a hundred-million-dollar quantum computer lives inside nested metal cans. The previous article established what a conductor in equilibrium is: field-free inside, charge exiled to the surface, the whole body one equipotential. This article is about what a conductor does to the world around it. The moment an external field arrives, the conductor's free charges redistribute — negative charge piling up where field lines land, positive where they leave — and this induced charge builds a counter-field that cancels the intrusion exactly, everywhere inside the metal. Hollow the conductor out, and something stronger happens: the empty cavity inherits the calm. No matter how violent or lopsided the outside field, the inside of the box stays at exactly zero field. That is Faraday's cage, and proving it takes nothing but Gauss's law and the curl-free character of electrostatic fields. But the shield has a one-way valve built into it, and most people get its direction wrong. A cage hides the world from its contents — yet an ungrounded cage does not hide its contents from the world. Understanding why, via a single Gaussian surface drawn through the metal, is the heart of this article.

Key takeaways

An external field drives a conductor's free charges to its surfaces within $\sim10^{-19}\,\mathrm{s}$ ; the rearranged induced charge produces a counter-field that cancels the applied field exactly throughout the metal.
An empty cavity inside a conductor is perfectly shielded: $\mathbf{E}_{\text{cavity}}=0$ regardless of the strength or shape of the external field — the Faraday cage.
A charge $q$ placed in the cavity induces exactly $-q$ on the cavity wall (forced by a Gaussian surface in the metal) and pushes $+q$ to the outer surface of a neutral conductor.
The outer surface has amnesia: its charge arranges itself as if the cavity didn't exist, so a spherical conductor's exterior field is that of a point charge at its center, wherever $q$ sits inside.
Shielding is asymmetric: the cage always protects inside from outside, but only a grounded cage hides inside charge from outside observers.
KEY IDEA: induced surface charge is the conductor's automatic, exact answer to any static field — it cancels the field in the metal, empties every enclosed cavity of field, and telegraphs enclosed charge to the outer surface unless a ground wire carries it away.

The core insight

In electrostatic equilibrium, induced surface charges arrange themselves to keep $\mathbf{E}=0$ in conducting material — shielding any empty cavity completely, while a cavity charge $q$ pins $-q$ to the inner wall and (unless grounded) $+q$ to the outer surface.

Where this lives today

Every electronics lab you will ever work in is an exercise in this article. Oscilloscope probes and microphone cables are miniature Faraday cages — a grounded braid whose induced charges terminate external interference before it reaches the signal conductor. Electrophysiology rigs measuring microvolt nerve signals sit inside mesh cages; EMC test facilities certify cars and pacemakers inside room-sized shielded chambers; secure facilities extend the same physics to information security, wrapping entire rooms in copper so that keystroke emissions cannot leak (the attenuation requirements are set by standards, but the mechanism is Faraday's). The frontier version is colder. Superconducting quantum processors idle at the bottom of dilution refrigerators inside nested shields — a grounded RF can, then mu-metal for magnetic fields — because a single stray charge rearrangement near a qubit can decohere it. And the reason your phone dies in an elevator but limps along near a window is mesh size: a cage of holes small compared to the relevant wavelength still shields well, which is also why the perforated screen on a microwave-oven door counts as "solid" at $12\,\mathrm{cm}$ wavelength while staying transparent at $500\,\mathrm{nm}$ so you can watch your food.

Origins · 1 minMichael Faraday and how this was discovered

Michael Faraday

1791–1867 · Royal Institution, London

Faraday was a bookbinder's apprentice who read the encyclopedias he was binding, talked his way into Humphry Davy's lab as a bottle-washer, and became the finest experimentalist of his century — all without ever learning calculus. The cage came in 1836: he built a twelve-foot cube of wood and wire mesh in the Royal Institution's lecture theatre, charged it with an electrostatic generator until sparks flew from its outside, and sat inside it with his most sensitive electroscopes. "I went into the cube and lived in it," he wrote, "but though I used lighted candles, electrometers, and all other tests of electrical states, I could not find the least influence upon them." The 1843 ice-pail experiment completed the picture — lowering a charged ball into a metal pail, he showed the induced charge on the outside exactly equals the enclosed charge, wherever the ball hangs. His field-line way of thinking, dismissed by contemporaries as a craftsman's crutch, turned out to be the deeper language: Maxwell spent the 1860s translating it into mathematics.

History

1729

— Stephen Gray discovers electrical conduction, splitting materials into conductors and insulators

1755

— Benjamin Franklin notices a cork ball hanging inside a charged silver can feels no force — the first recorded cavity-shielding observation

1836

— Faraday builds and inhabits his cube, demonstrating that an enclosure's interior ignores external electrification

1843

— The ice-pail experiment quantifies induction: induced charge exactly equals enclosed charge

1873

— Maxwell's Treatise recasts Faraday's cage and ice-pail as exact theorems of potential theory

The physical picture

Think of the conduction electrons in a metal as a liquid that cannot tolerate a slope. An applied field is a tilt; the liquid flows downhill — electrons drift against the field — until the surface puddles of excess charge build a counter-tilt that levels everything again. The flow is self-terminating: charges move because the interior field is nonzero, and they stop only when it isn't. The induced charge is just the bookkeeping of where the liquid piled up: a negative film where external field lines land, a positive film where they emerge.

Figure

Induced charge — a conductor edits the field around it

Field lines respect the arrangement: an external line cannot pass through the metal, so it terminates on the induced negative film and a fresh line is re-issued from the positive film on the far side. From outside, the conductor appears to "swallow and re-emit" the field; inside, there is nothing at all. Now hollow the conductor out. Could the cavity — empty space, no metal, no charge — carry a field smuggled in from outside? Intuition says the walls "block" it, but the honest argument is more beautiful: a field line inside the empty cavity would have to start on the wall and end on the wall, running from higher to lower potential. Close that path through the metal, where $\mathbf{E}=0$ , and you have a loop with a nonzero $\oint \mathbf{E}\cdot\mathrm{d}\boldsymbol{\ell}$ — which the curl-free electrostatic field flatly forbids. The cavity is clean because the alternative would break the most basic law the field obeys. Derivation A makes this airtight. The story changes completely if you place a charge inside the cavity. Its field must be terminated before reaching the metal, so the inner wall dresses itself with exactly $-q$ ; conservation then strands $+q$ with nowhere to go but the outer surface. The outside world sees that $+q$ — the cage leaks outward unless you ground it and let the stranded charge drain to Earth.

Figure

Two cavities — the empty one is shielded, the charged one telegraphs

Notice the strangest detail in the right-hand panel: the $+q$ on the outer surface distributes itself uniformly (for a spherical conductor) even though the cavity charge sits off-center. The metal wall is a perfect information firewall — the inner-wall charge cancels the cavity charge's field so completely inside the conductor that the outer surface never learns where the charge is, only how much there is.

From equilibrium to induction

Everything descends from the equilibrium condition established in the previous article — inside conducting material,

\mathbf{E} = 0 \qquad (1)

where: $\mathbf{E}$ — the macroscopic electric field in the bulk of the conductor, the total field of all sources, external and induced. Equation (1) is not an approximation about good conductors; it is the definition of electrostatic equilibrium. If it failed anywhere, free charge would still be flowing and "static" would be false. Combined with Gauss's law in differential form, $\nabla\cdot\mathbf{E}=\rho/\varepsilon_0$ , it forces $\rho=0$ throughout the bulk: every scrap of net charge, whether native or induced, lives on surfaces. How fast is "instantaneous"? Model the metal with Ohmic conductivity $\sigma_c$ and let charge continuity do the accounting. Any interior charge density decays as

\rho(t) = \rho_0\, e^{-t/\tau}, \qquad \tau = \frac{\varepsilon_0}{\sigma_c} \qquad (2)

where: $\sigma_c$ — the material's conductivity in $\mathrm{S/m}$ (the subscript keeps it distinct from surface charge density $\sigma$ ); $\tau$ — the charge relaxation time. In a dielectric medium the $\varepsilon_0$ here becomes the material's permittivity $\varepsilon$ , which matters whenever that is not close to $\varepsilon_0$ . For copper, $\sigma_c = 5.96\times10^{7}\,\mathrm{S/m}$ gives $\tau \approx 1.5\times10^{-19}\,\mathrm{s}$ — shorter than the period of visible light. Even seawater, a miserable conductor by metallic standards ( $\sigma_c \approx 4\,\mathrm{S/m}$ ), relaxes in about $0.2\,\mathrm{ns}$ . On any electrostatic timescale, induction is done before you finish causing it. The induced charge announces itself through the boundary condition at the surface: just outside, the field is normal to the surface with magnitude $\sigma/\varepsilon_0$ . This article uses that fact numerically; deriving it — and the surprising force it implies on the conductor's own skin — is the entire business of the next article, Surface Charge and Field at a Conductor. The cavity results deserve honest proofs rather than plausibility, and they come in a matched pair.

Derivation AThe empty cavity is exactly field-free

We prove the Faraday-cage theorem: an empty cavity inside a conductor has $\mathbf{E}=0$ no matter what fields exist outside. The strategy is proof by contradiction, and it uses only two facts — Gauss's law and $\nabla\times\mathbf{E}=0$ — so the result is as solid as electrostatics itself. The subtle step is realizing that Gauss's law alone is not enough: zero enclosed charge kills the net flux, not the field. It takes the curl-free condition, wielded on a cleverly closed loop, to finish the job.

Step 1. Draw a Gaussian surface $S$ entirely within the conducting material, snugly surrounding the cavity. Since $\mathbf{E}=0$ everywhere on $S$ by (1), the flux through $S$ vanishes, and Gauss's law gives

Q_{\text{cavity wall}} = \varepsilon_0 \oint_S \mathbf{E}\cdot\mathrm{d}\mathbf{a} = 0.

Why this isn't the end: zero net wall charge still permits a patch of $+$ here and a patch of $-$ there, with field lines running across the cavity between them. Gauss's law cannot rule that out.

Step 2. Suppose, for contradiction, that $\mathbf{E}\neq 0$ somewhere in the cavity. Then some field line threads the cavity. The cavity is empty, so the line cannot begin or end mid-flight; it must run from a positive patch on the wall to a negative patch on the wall.

Step 3. Build a closed loop: follow the field line across the cavity, then return to the starting point along any path through the metal. Along the field line, $\mathbf{E}\cdot\mathrm{d}\boldsymbol{\ell} > 0$ at every step (you are walking with the field), so that segment contributes a strictly positive amount. Through the metal, $\mathbf{E}=0$ and the return segment contributes exactly nothing.

Step 4. The total is $\oint\mathbf{E}\cdot\mathrm{d}\boldsymbol{\ell} > 0$ . But the electrostatic field is curl-free, so every closed loop integral must vanish. The contradiction leaves one exit: no such field line exists, and

\mathbf{E}_{\text{cavity}} = 0 \quad\text{everywhere in the cavity.}

Why the wall helps: the same argument shows the induced charge on the outer surface plus the external sources conspire to cancel inside the metal — the cavity simply sits inside that protected zone, walled off by a surface at constant potential. Constant $V$ on the cavity wall matches the constant solution, and uniqueness guarantees that is the only solution.

Go deeper · Derivation B · 2 minCharge in the cavity — the bookkeeping and the outer surface's amnesia

Derivation BCharge in the cavity — the bookkeeping and the outer surface's amnesia

Now place a point charge $q$ inside the cavity of a neutral, isolated conductor. We want three results: the total charge induced on the cavity wall, the charge stranded on the outer surface, and — the punchline — the fact that the exterior field carries no information about where the cavity charge sits. The tools are the same Gaussian surface as before plus charge conservation, and the amnesia result previews the uniqueness reasoning that E2 makes rigorous.

Step 1. Draw the Gaussian surface $S$ through the metal, enclosing the cavity. $\mathbf{E}=0$ on $S$ , so the enclosed charge vanishes:

q + Q_{\text{inner}} = 0 \quad\Longrightarrow\quad Q_{\text{inner}} = -q.

Why exactly: not approximately $-q$ , and not dependent on the cavity's shape or the charge's position — the flux through $S$ is identically zero, so the books must balance to the last coulomb.

Step 2. The conductor started neutral and no charge crossed its surfaces, so

Q_{\text{inner}} + Q_{\text{outer}} = 0 \quad\Longrightarrow\quad Q_{\text{outer}} = +q.

Step 3. Where on the outer surface does $+q$ sit? The inner-wall charge $-q$ terminates every field line from $q$ before it reaches the metal; together, $q$ and its inner-wall screen produce zero field everywhere in the conductor and beyond. The outer surface therefore feels nothing from the cavity system — its $+q$ must solve its own problem: distribute on a conductor surface with no other fields present.

Step 4. For a spherical outer surface, that standalone problem is solved by the uniform distribution, so the exterior field is

\mathbf{E}_{\text{out}} = \frac{1}{4\pi\varepsilon_0}\,\frac{q}{r^2}\,\hat{\mathbf{r}} \qquad (r > R),

centered on the sphere's center — wherever $q$ wanders inside the cavity. Move the cavity charge; the inner-wall charge redistributes instantly to keep the screen perfect; the outer surface never hears about it. Why "amnesia" matters: the exterior field reveals the total enclosed charge (Gauss's law guarantees that much leaks out) but conceals all geometry. A detector outside can count coulombs, never draw maps. Why "distribute as if alone" works: the inner-wall charge $-q$ screens the cavity charge so completely that the outer surface experiences no field from the cavity system at all. It therefore solves its own standalone boundary-value problem — an application of the first uniqueness theorem, which guarantees that the solution with no cavity influence is the unique one.

Limiting cases

Case	Result	Meaning
Deep inside conducting material	$\mathbf{E}=0$ , $\rho=0$	Equilibrium definition; all charge on surfaces
Empty cavity, any external field	$\mathbf{E}_{\text{cavity}}=0$ exactly	Faraday cage; independent of external field strength or shape
Cavity charge $q$ , isolated neutral conductor	$Q_{\text{inner}}=-q$ , $Q_{\text{outer}}=+q$	Exterior sees $q$ but not its position
Cavity charge $q$ , grounded conductor	$Q_{\text{inner}}=-q$ , $Q_{\text{outer}}=0$	Ground wire drains the stranded charge; two-way isolation
Real metal, sudden field switch-on	$\rho\to 0$ in $\tau=\varepsilon_0/\sigma_c \sim 10^{-19}\,\mathrm{s}$ (copper)	"Instantaneous" is an understatement in statics
Rapidly oscillating fields	shielding degrades once frequency $\to 1/\tau$ scale; penetration governed by skin depth	Electrostatic results; AC shielding is a different (E8) story

Worked examples

1 · A charge in a spherical shell (confidence builder)

A point charge $q = +2.0\,\mu\mathrm{C}$ hangs off-center inside the irregular cavity of a neutral conducting sphere of outer radius $R = 10\,\mathrm{cm}$ . Find the total charge on the cavity wall and outer surface, the field at $r = 50\,\mathrm{cm}$ from the sphere's center, and the potential of the conductor. Then ground it.

Step 1. Gaussian surface through the metal: $Q_{\text{inner}} = -q = -2.0\,\mu\mathrm{C}$ . Neutrality: $Q_{\text{outer}} = +2.0\,\mu\mathrm{C}$ . The cavity's irregular shape changes the distribution on the inner wall, not these totals.

Step 2. By the amnesia theorem, the outer $+2.0\,\mu\mathrm{C}$ spreads uniformly over the spherical outer surface, so outside the conductor the field is that of a $2.0\,\mu\mathrm{C}$ point charge at the center:

E = \frac{1}{4\pi\varepsilon_0}\frac{q}{r^2} = \frac{(8.99\times10^{9})(2.0\times10^{-6})}{(0.50)^2}\,\mathrm{V/m} = 7.2\times10^{4}\,\mathrm{V/m}.

Step 3. The conductor's potential (relative to infinity) is set by its outer surface:

V = \frac{1}{4\pi\varepsilon_0}\frac{q}{R} = \frac{(8.99\times10^{9})(2.0\times10^{-6})}{0.10}\,\mathrm{V} \approx 1.8\times10^{5}\,\mathrm{V}.

Step 4. Now connect the sphere to ground. The stranded $+2.0\,\mu\mathrm{C}$ drains away, $Q_{\text{outer}} \to 0$ , the exterior field collapses to zero, and $V \to 0$ — while inside the cavity absolutely nothing changes: $q$ and its $-q$ screen were already a closed conversation. Exterior field before grounding: $7.2\times10^{4}\,\mathrm{V/m}$ ; after: zero.

2 · What shielding actually costs the metal (exam level)

A cubical steel instrument enclosure sits in a uniform field $E_0 = 100\,\mathrm{kV/m}$ (a healthy thunderstorm-scale field). Estimate the induced surface charge density where field lines land squarely on a face, the number of electrons per square micrometer this represents, the fraction of the metal's available conduction electrons involved, and the time the rearrangement takes.

Step 1. Where field lines meet the surface at normal incidence, the boundary condition gives the local induced density

\sigma = \varepsilon_0 E_0 = (8.854\times10^{-12})(1.0\times10^{5})\,\mathrm{C/m^2} = 8.9\times10^{-7}\,\mathrm{C/m^2} \approx 0.89\,\mu\mathrm{C/m^2}.

Step 2. In electrons per area:

\frac{\sigma}{e} = \frac{8.9\times10^{-7}}{1.602\times10^{-19}}\,\mathrm{m^{-2}} = 5.5\times10^{12}\,\mathrm{m^{-2}} = 5.5\ \text{electrons per}\ \mu\mathrm{m}^2.

A field fierce enough to make your hair crackle is canceled by five or six electrons per square micrometer.

Step 3. A single atomic layer of a typical metal (atom spacing $\sim 0.25\,\mathrm{nm}$ , one conduction electron each) offers about $1/(0.25\times10^{-9})^2 \approx 1.6\times10^{19}$ electrons per $\mathrm{m^2}$ . The shield uses

\frac{5.5\times10^{12}}{1.6\times10^{19}} \approx 3\times10^{-7}

of one atomic layer — a part-per-million-scale perturbation. The conductor is never remotely "used up"; ideal-conductor behavior survives any field short of one that rips atoms apart.

Step 4. The rearrangement time is the relaxation time of (2). For steel ( $\sigma_c \approx 1.0\times10^{7}\,\mathrm{S/m}$ ),

\tau = \frac{\varepsilon_0}{\sigma_c} = \frac{8.854\times10^{-12}}{1.0\times10^{7}}\,\mathrm{s} \approx 8.9\times10^{-19}\,\mathrm{s}.

The shield deploys $\sim$ 6 electrons per $\mu\mathrm{m}^2$ in under an attosecond — a $10^{-7}$ -of-a-monolayer correction that erases the field completely.

Going further

References. Griffiths, Introduction to Electrodynamics, 4th ed., §2.5 (Conductors) · Purcell & Morin, Electricity and Magnetism, 3rd ed., Ch. 3 · Jackson, Classical Electrodynamics, 3rd ed., §1.11 · Zangwill, Modern Electrodynamics, Ch. 5. Key results used throughout: surface boundary condition $\sigma=\varepsilon_0 E_\perp$ , cavity theorems, and surface pressure $\sigma^2/2\varepsilon_0$ . Constant $\varepsilon_0=8.854\times10^{-12}\,\mathrm{F/m}$ and metal charge-relaxation time $\sim10^{-18}\,\mathrm{s}$ . Note the factor-of-two in the surface pressure formula $\sigma^2/2\varepsilon_0$ , which arises from the pillbox boundary-condition derivation.

Builds on: Conductors in Electrostatic Equilibrium · Gauss's Law in Integral Form · Gauss's Law in Differential Form · The Curl of the Electrostatic Field · The Electrostatic Potential

Leads to: Surface Charge and Field at a Conductor · Uniqueness Theorems for the Potential · Method of Images

If the field inside the metal is zero, how does the outside "find out" about a charge in the cavity at all?

Through charge conservation, not field penetration. The cavity charge's field is fully terminated by the $-q$ induced on the inner wall — no field crosses the metal. But that $-q$ had to come from somewhere: it was separated out of the neutral metal, leaving $+q$ behind, and that orphaned charge has nowhere to live except the outer surface, where it generates a perfectly ordinary exterior field. The message is carried by bookkeeping, not by transmission.

Why does shielding work one way but not the other for an ungrounded box?

The asymmetry is in who has a boundary at infinity. External fields are canceled inside because the cavity is wrapped in an equipotential surface — the cage — and a charge-free region bounded by an equipotential has the constant potential as its unique solution. An internal charge, though, forces $+q$ onto the outer surface, and there is no outer conductor enclosing the universe to screen that. Grounding fixes it by giving the stranded charge an escape route: with $Q_{\text{outer}}=0$ , the exterior solution is zero too.

Does the induced charge on the inner wall move when I move the charge around the cavity?

Constantly and instantly (on the $10^{-19}\,\mathrm{s}$ relaxation timescale). Bring $q$ close to one side of the cavity and the induced $-q$ crowds toward it, always arranging so the total field in the metal vanishes. The distribution changes; the total stays exactly $-q$ . Meanwhile the outer-surface charge never budges — the amnesia theorem — because the cavity system's external field is identically zero.

Is my car really a Faraday cage during lightning? It has windows.

It's a good approximate one. A lightning stroke's dominant energy sits at wavelengths of kilometers, vastly larger than a car's window openings, so the metal body behaves like a continuous enclosure and the current flows over the outside skin. The protection is not the rubber tires (a strike that jumped a kilometer of air is not stopped by 10 cm of rubber). Convertibles, fiberglass bodies, and open cabs genuinely lose the protection.

Why can't gravity be shielded the same way?

Shielding works because conductors contain mobile charge of both signs, so they can build a counter-field by separating them. Gravity has only one sign of "charge" — mass attracts, always — so no material can rearrange itself to cancel a gravitational field in a region. The mathematical structure (Poisson's equation) is identical; the missing ingredient is negative mass.

Where does the energy go when induced charges rearrange?

Rearrangement is driven by the applied field doing work on the free charges; during the $\sim\tau$ transient, that work partly builds the new field configuration and is partly dissipated as Joule heat in the metal. Once equilibrium is reached, no further energy flows. Interestingly, the final configuration is always the one of minimum field energy consistent with the constraints — a theme that becomes a theorem (Thomson's) in the energy articles of this chapter.

Is E exactly zero inside a real metal, atom by atom?

No — between atoms the microscopic field is enormous and wildly varying. The $\mathbf{E}=0$ of this article is the macroscopic field, averaged over many lattice spacings. A related quantum refinement: an applied static field actually penetrates a metal to about a Thomas–Fermi screening length, $\sim 0.05$ – $0.1\,\mathrm{nm}$ — less than one atomic layer. The classical "all charge on the surface, zero field inside" is the $\hbar\to$ irrelevant limit of that screening, and for anything larger than nanometers it is exact for all practical purposes.