Dielectric Materials and Insulators

In 1837 Michael Faraday built two identical spherical capacitors, charged one, shared its charge with the other, and found something the theory of his day had no slot for: with a cake of shellac filling the gap, the apparatus held more charge at the same potential — reproducibly, by a fixed factor that depended only on the filling material. He called the factor specific inductive capacity; we call it the dielectric constant $\kappa$ , and the materials that produce it — coined for Faraday by William Whewell — dielectrics. Every capacitor in every device you own is built on that 1837 factor. This article opens the electric-fields-in-matter chapter, and its job is the map. A dielectric is an insulator that responds to a field not by conducting but by polarizing: its charges are bound to atoms and molecules, so instead of flowing they stretch and turn, each molecule becoming or reorienting a tiny dipole — the very object the multipole chapter just spent four articles anatomizing. The chapter ahead builds the full machinery: the two microscopic mechanisms (Induced Dipoles, Permanent Dipoles), the macroscopic field $\mathbf{P}$ (Polarization Density), the bound charges it implies (Bound Charges: Surface and Volume), the fields they make (Field of a Polarized Object, Uniformly Polarized Sphere), and the bookkeeping field $\mathbf{D}$ (The Electric Displacement). Here we earn the ideas that need no machinery: why insulators insulate and conductors don't, why a polarized slab partially screens a field where a conductor screens it completely, how Faraday's measured $\kappa$ pins down the bound surface charge before we can derive it, and why the response of ordinary matter to laboratory fields is so reliably linear — until, at a few megavolts per meter, it catastrophically isn't.

Key takeaways

A dielectric is an insulator whose charges are bound: an applied field cannot make them flow, but it can stretch atoms and align polar molecules, producing a distributed polarization instead of a conduction current.
The contrast with a conductor is total versus partial screening: free charges rearrange until $\mathbf{E} = 0$ inside a conductor; bound charges shift slightly and only reduce the interior field, by the factor $\kappa$ .
Two microscopic mechanisms feed the response — induced dipoles (charge clouds stretch) and orientational alignment of permanent dipoles (molecules rotate against thermal noise); their anatomies are the next two articles' business.
Faraday's measurable $\kappa = C/C_0$ already fixes the bound surface charge on a filled capacitor's faces: $\sigma_b = \sigma_f(1 - 1/\kappa)$ — 62% of the free charge for polystyrene, 100% in the conductor limit $\kappa \to \infty$ .
Linearity is the rule because lab fields ( $\sim 10^6\,\mathrm{V/m}$ ) are minuscule beside the atomic fields binding the charges ( $\sim 10^{11}\,\mathrm{V/m}$ ); above the material's dielectric strength the bound charges tear free and the insulator fails as a spark or avalanche.
KEY IDEA: a dielectric answers an applied field by polarizing rather than conducting — bound charge builds on its surfaces and (for non-uniform response) in its bulk, partially screening the field by the measurable factor $\kappa$ .

The core insight

A dielectric responds to an applied field by polarizing — bound charges shift to produce a distributed dipole moment — rather than by conducting; the resulting bound charge partially screens the field, raising capacitance by the measurable factor $\kappa$ , with complete screening (the conductor) as the $\kappa \to \infty$ limit.

Where this lives today

The most consequential dielectric decision of this century sits in your pocket. Shrinking transistors thinned the SiO $_2$ gate dielectric ( $\kappa = 3.9$ ) toward five atomic layers, where quantum tunneling leaked current straight through the "insulator." The industry's 2007 fix was Faraday's factor: hafnium-based high- $\kappa$ dielectrics ( $\kappa \approx 20$ – $25$ ) recover the same gate capacitance at a physically thicker, tunnel-proof film. The interconnect layers want the opposite — low- $\kappa$ insulation ( $\kappa \to 2$ , porous carbon-doped oxides) to cut parasitic capacitance — so a modern chip engineers dielectrics in both directions at once. The same trade-offs run through power engineering (SF $_6$ and vacuum interrupters standing off megavolt transients), 5G boards (low-loss laminates for tens of GHz, where dispersion turns $\kappa$ complex), and pulsed power, where polymer–ceramic nanocomposites chase the $u_{\max} = \tfrac12\varepsilon_0\kappa E_{\text{bd}}^2$ frontier. Faraday's observation — that the material between conductors is a degree of freedom, not packaging — has become an industry.

Origins · 1 minMichael Faraday and how this was discovered

Michael Faraday

1791–1867 · Royal Institution, London

Faraday came to dielectrics from a conviction the mathematicians of his day did not share: that electric action happens in the medium, not by charges leaping across empty distance. In 1837, with two matched spherical capacitors and a torsion balance, he showed that shellac, sulphur, or glass in the gap raised the charge held at a given potential by a substance-specific factor — the specific inductive capacity — and he read it as the medium itself straining along the lines of force, molecule by molecule (his term for the strained state, dielectric polarization, survives intact; the word "dielectric" was supplied at his request by William Whewell, the Cambridge polymath who also coined "scientist"). It was a bookbinder's apprentice's third revolution at the Royal Institution, after electromagnetic rotation (1821) and induction (1831) — and the one most directly aimed at the future: Maxwell said his own theory began as an attempt to translate Faraday's strained-medium pictures into mathematics, and the field concept modern physics runs on is the translation's residue. Faraday declined a knighthood and, twice, the presidency of the Royal Society; the farad, coined for the unit this article's $\kappa$ multiplies, is the accepted honor.

History

1745

— The Leyden jar stores charge in glass, the first working dielectric — used for a century before being understood

1837

— Faraday measures specific inductive capacity: the material between conductors sets a reproducible factor

1873

— Maxwell's Treatise embeds Faraday's polarized medium in field equations, with displacement as a central actor

1912

— Debye separates induced from orientational polarization by its temperature dependence, making molecular dipoles measurable

2007

— Hafnium-based high-kappa gate dielectrics enter mass-market processors, ending the silicon-dioxide era

The physical picture

Two slabs between charged capacitor plates, one copper and one glass, tell the whole story of this chapter. In the copper, conduction electrons are free: the instant the field appears they surge (on the $\tau = \varepsilon_0/\sigma_c \sim 10^{-19}\,\mathrm{s}$ timescale established in Conductors in Electrostatic Equilibrium) until their surface pile-up cancels the applied field exactly. Interior field: zero. In the glass, every electron is spoken for — locked in a bond or a filled shell, no vacant states nearby to move into (that is the band-picture sentence behind "insulator," and one sentence is all we spend on it). The field cannot make glass conduct. What it can do is distort: each electron cloud shifts a fraction of an ångström off its nucleus, each polar molecule leans slightly into the field. A hundred million million million small compliances, all aligned.

Figure

Conductor versus dielectric — total versus partial screening

Now look at the sum of those compliances the way we looked at the polarized medium in the multipole chapter. Deep inside the slab, the displaced positive end of each molecule sits beside the displaced negative end of its neighbor: cancellation. Only at the two faces does the chain terminate unpaired — a thin sheet of uncompensated positive bound charge on the face the field points toward, negative on the other. Those sheets produce a field of their own, opposing the applied one. But unlike the conductor's surface charge, the bound sheets cannot grow without limit; each molecule can stretch only so far for a given field. The screening stalls at a material-dependent fraction, the interior field survives at reduced strength, and the reduction factor is precisely Faraday's $\kappa$ . Complete anatomy of the sheets — including the volume charge $\rho_b = -\nabla\cdot\mathbf{P}$ that appears when the polarization is non-uniform — is the business of Bound Charges: Surface and Volume; here the slab picture is all we need. Where do the molecular dipoles come from? Two mechanisms, cleanly separated by what the field must do. If the molecule has no dipole of its own (N $_2$ , CH $_4$ , any atom), the field creates one by stretching the charge cloud: induced polarization, stiff and essentially temperature-blind. If the molecule carries a permanent dipole (H $_2$ O at $1.85\,\mathrm{D}$ , HCl), the field need only turn it — but thermal agitation randomizes orientations relentlessly, so alignment is statistical and fades as $1/T$ . The next two articles derive each mechanism's strength; the chapter opener's point is only that both funnel into the same macroscopic object, the dipole moment per unit volume, and from there into one number a bridge measures.

Figure

Two mechanisms, one response — stretch and rotate

Faraday's number, and what it already tells us

The dielectric constant is defined by the experiment Faraday actually did — fill a capacitor, compare:

\kappa \equiv \frac{C_{\text{filled}}}{C_{\text{vacuum}}} \qquad (1)

where: $C = Q_f/V$ — capacitance, the free charge stored per volt; $\kappa$ — the dielectric constant (also written $\varepsilon_r$ ), a pure number, a property of the material alone, independent of the capacitor's size or shape so long as the filling is complete. Selected static values at room temperature: vacuum $1$ exactly, air $1.00059$ , polystyrene $2.6$ , glass $\approx 5$ , water $80.1$ . Every $\kappa$ exceeds $1$ : polarization always opposes the applied field, never assists it — a stability statement we unpack in the FAQ. The slab picture converts this measured number directly into a statement about the bound charge, and it is worth doing before any of the chapter's formal machinery exists — physics first, formalism later:

E_{\text{inside}} = \frac{\sigma_f - \sigma_b}{\varepsilon_0}, \qquad \sigma_b = \sigma_f\left(1 - \frac{1}{\kappa}\right) \qquad (2)

where: $\sigma_f$ — the free charge per area on the plates (what the battery delivered); $\sigma_b$ — the bound charge per area on the adjacent dielectric faces; $E_{\text{inside}}$ — the macroscopic field in the filled gap. Read the second formula at its ends: for a feeble dielectric ( $\kappa \to 1$ ) the bound charge vanishes; for $\kappa = 2.6$ (polystyrene) the faces carry $62\%$ of the free charge; as $\kappa \to \infty$ , $\sigma_b \to \sigma_f$ — complete cancellation, and the dielectric has become a conductor. The conductor is not a rival theory; it is the top rung of the dielectric ladder.

Derivation AThe slab audit — from measured kappa to the bound surface charge

The goal is equation (2): the interior field of a filled parallel-plate capacitor written in terms of free and bound sheet charges, and the bound charge extracted from the measured $\kappa$ — no constitutive law, no $\mathbf{P}$ , no $\mathbf{D}$ , none of which we have built yet. The strategy is bare superposition of charged sheets, whose fields we know exactly from the symmetry catalog of Gauss's Law: Applications and Symmetry. The one physical input is the slab picture: a uniformly polarized slab presents two face sheets $\pm\sigma_b$ and no bulk charge (the full derivation of that claim belongs to Bound Charges: Surface and Volume; here it is the model). Watch that we never assume how $\sigma_b$ depends on the field — that is the point.

Step 1. Fields of the four sheets. The plates carry $\pm\sigma_f$ , the dielectric faces $\mp\sigma_b$ (negative bound charge faces the positive plate — the molecules' negative ends are pulled toward it). Each infinite sheet contributes $\sigma/2\varepsilon_0$ on each side, directed away from (toward) a positive (negative) sheet. Between the plates all four contributions align along the gap axis; outside they cancel pairwise:

E_{\text{inside}} = \frac{\sigma_f}{2\varepsilon_0} + \frac{\sigma_f}{2\varepsilon_0} - \frac{\sigma_b}{2\varepsilon_0} - \frac{\sigma_b}{2\varepsilon_0} = \frac{\sigma_f - \sigma_b}{\varepsilon_0}.

Why the bound sheets subtract: they are arranged by the field itself — negative bound charge nearest the positive plate — so their own field points against the applied one. Polarization is a compliance, and a compliance opposes its cause.

Step 2. Bring in the measurement. At fixed free charge, $V = E_{\text{inside}}\,d$ and $C = Q_f/V$ , so

\frac{C_{\text{filled}}}{C_{\text{vacuum}}} = \frac{E_{\text{vac}}}{E_{\text{inside}}} = \frac{\sigma_f/\varepsilon_0}{(\sigma_f - \sigma_b)/\varepsilon_0} = \frac{\sigma_f}{\sigma_f - \sigma_b} \equiv \kappa.

Faraday's factor is exactly the screening ratio: measure capacitances, learn charge bookkeeping at the atomic scale.

Step 3. Solve for the bound charge:

\sigma_b = \sigma_f\left(1 - \frac{1}{\kappa}\right).

Why this is remarkable: a bridge measurement made at the terminals — two capacitance readings — has determined the surface density of molecular charge ends inside the gap, decades before anyone could picture a molecule. It also hands the chapter its target: any microscopic theory of dielectrics must predict this $\sigma_b$ , which is precisely what the $\mathbf{P}$ -machinery of the next articles will do.

Step 4. Audit the limits. $\kappa = 1$ : $\sigma_b = 0$ , vacuum, no medium to respond. $\kappa \to \infty$ : $\sigma_b \to \sigma_f$ , the interior field $\to 0$ — the conductor result from Conductors in Electrostatic Equilibrium, recovered as a limit rather than assumed. For water ( $\kappa = 80$ ), $\sigma_b = 0.9875\,\sigma_f$ : water screens $98.75\%$ like a near-conductor, yet conducts (ideally) nothing. Note that the fixed-free-charge assumption is legitimate because the dielectric adds charge only in bound pairs — equal and opposite ends of the same molecule, so the free charge on the plates is unchanged.

One measured number, three equivalent working statements — worth deriving once because exam problems and datasheets use all three interchangeably:

\text{fixed } Q_f:\; E \to \frac{E_0}{\kappa},\; V \to \frac{V_0}{\kappa}; \qquad \text{fixed } V:\; Q_f \to \kappa Q_{f0}; \qquad \text{always}:\; C = \kappa C_0 \qquad (3)

where: subscript $0$ denotes the empty capacitor. The fixed- $Q$ statements follow from (2) directly; the fixed- $V$ statement follows because holding $V$ (a battery) forces $E = V/d$ back to its vacuum value, and the battery must ship in $\kappa$ times the free charge to overcome the screening. Which quantity is held fixed decides everything downstream (fields, energies, forces — the fixed- $Q$ /fixed- $V$ fork returns with teeth in Energy Stored in Dielectric Systems and Forces on Dielectrics), so build the reflex now: first question for any dielectric problem — battery connected, or isolated?

Go deeper · Derivation B · 2 minWhy the response is linear — and where linearity dies

Derivation BWhy the response is linear — and where linearity dies

The goal is the pair of scales that bracket every dielectric in this chapter: why $\sigma_b$ (equivalently the polarization) is so reliably proportional to the applied field at laboratory strengths, and why the proportionality fails catastrophically — breakdown — at a well-defined material threshold. The strategy is an honest comparison of two fields: the one we apply, and the one the atom applies to its own electrons. No model of the atom is fitted; only Coulomb's law at the ångström scale. (The quantitative response coefficient $\alpha$ is the next article's derivation; here we establish only smallness, which is what linearity needs.)

Step 1. The internal field scale. An electron a Bohr-ish distance $a \approx 0.5\times10^{-10}\,\mathrm{m}$ from a proton sits in a field

E_{\text{at}} = \frac{e}{4\pi\varepsilon_0 a^2} = \frac{(8.99\times10^9)(1.6\times10^{-19})}{(0.5\times10^{-10})^2} \approx 5.8\times10^{11}\,\mathrm{V/m}.

This is the stiffness of matter: the field an applied field must compete with to displace bound charge.

Step 2. The applied field scale. A harsh laboratory field — the edge of what air tolerates — is $E \approx 3\times10^6\,\mathrm{V/m}$ . The ratio:

\frac{E}{E_{\text{at}}} \approx \frac{3\times10^6}{5.8\times10^{11}} \approx 5\times10^{-6}.

The distortion of any atom is a few parts per million of its natural charge separation. A displacement that small samples only the linear zone of whatever restoring force binds the electron — every smooth force is linear close enough to equilibrium — so the induced response is proportional to the field, with material-dependent slope. That slope, packaged macroscopically, is $\chi_e = \kappa - 1$ , whose formal apparatus arrives in Linear Dielectrics, Susceptibility and Permittivity. Why this argument is trustworthy despite its crudity: it needs only orders of magnitude. Linearity would survive even a factor-100 error in $E_{\text{at}}$ , and the same reasoning predicts where nonlinearity must appear — modern femtosecond lasers reach $10^{11}$ – $10^{12}\,\mathrm{V/m}$ , comparable to $E_{\text{at}}$ , and there nonlinear optics is not a correction but the show.

Step 3. Where insulation dies. Breakdown is not the smooth end of linearity but an instability: a stray free electron (cosmic ray, field emission) accelerating in field $E$ gains energy $eE\lambda$ over a mean free path $\lambda$ ; if that exceeds an ionization energy, each collision breeds new free electrons, and the avalanche turns the insulator into a plasma channel — the spark. The threshold, the dielectric strength $E_{\text{bd}}$ , is set by collision physics, not by $\kappa$ : air $\approx 3\,\mathrm{MV/m}$ , polystyrene $\approx 20\,\mathrm{MV/m}$ , mica $\approx 100\,\mathrm{MV/m}$ . Note the scales: breakdown arrives around $10^{-5}\,E_{\text{at}}$ , long before the bound charges themselves would let go — insulation fails by avalanche of the few free stragglers, not by tearing the lattice.

Step 4. The design consequence. Since a capacitor's ceiling is a maximum field, the figure of merit for energy storage is the energy density at breakdown, $u_{\max} = \tfrac{1}{2}\varepsilon_0\kappa E_{\text{bd}}^2$ (the energy formula is developed properly in Energy Stored in Dielectric Systems; we borrow its form for a scale). Both $\kappa$ and $E_{\text{bd}}$ multiply into performance — worked example 2 shows the pair buying a factor of six thousand.

Limiting cases

Case	Result	Meaning
vacuum	$\kappa = 1$ , $\sigma_b = 0$	nothing to polarize; the reference all materials are measured against
feeble dielectric (air)	$\kappa = 1.00059$	correction at the $10^{-4}$ level — why vacuum formulas work in air
ordinary solids	$\kappa \approx 2$ – $10$	the workhorse regime: partial screening, linear response
water	$\kappa = 80$	permanent dipoles add orientational response — large, and $T$ -dependent
conductor limit	$\kappa \to \infty$ , $\sigma_b \to \sigma_f$ , $E_{\text{in}} \to 0$	complete screening; conductors are the ladder's top rung
optical frequencies	water $\kappa \to 1.77$	molecules can't rotate at $10^{15}\,\mathrm{Hz}$ : orientation drops out, only stretch survives
$E > E_{\text{bd}}$	avalanche, spark	insulator becomes plasma; the dielectric picture ends

Common trapA dielectric does not stop the field — and bound charge is not fictitious

Two opposite misreadings of the slab picture, both fatal. First: a dielectric between plates does not "block" or "absorb" the field the way a conductor does — the interior field is reduced by $\kappa$ , not eliminated, and it does real work on any charge passing through. Second, the compensating error: the bound charge sheets are not a bookkeeping device — they are actual molecular charge ends, they produce genuine fields inside and outside the material, and an electrometer touched to a polarized face reads them. What is bookkeeping is the later split of Gauss's law into free and bound ledgers via $\mathbf{D}$ — and the chapter's final article is devoted to not confusing that convenience with physics.

Worked examples

1 · A polystyrene-filled capacitor, fully audited (confidence builder)

A parallel-plate capacitor, plate area $A = 25\,\mathrm{cm}^2$ , gap $d = 1.0\,\mathrm{mm}$ , is filled with polystyrene ( $\kappa = 2.6$ ) and connected to a $12\,\mathrm{V}$ battery. Find $C$ , the free charge, the field, and the bound surface charge; compare with the empty capacitor on the same battery.

Step 1. Capacitances. Empty: $C_0 = \varepsilon_0 A/d = (8.854\times10^{-12})(25\times10^{-4})/(10^{-3}) = 22.1\,\mathrm{pF}$ . Filled: $C = \kappa C_0 = 2.6 \times 22.1 = 57.6\,\mathrm{pF}$ .

Step 2. Free charge and field at fixed $V$ . $Q_f = CV = 57.6\,\mathrm{pF}\times12\,\mathrm{V} = 0.69\,\mathrm{nC}$ (the empty capacitor would hold $0.27\,\mathrm{nC}$ — the battery shipped $2.6\times$ more). The field is battery-pinned: $E = V/d = 12\,\mathrm{kV/m}$ , same as empty.

Step 3. Charge densities. $\sigma_f = Q_f/A = 0.276\,\mu\mathrm{C/m^2}$ . Bound: $\sigma_b = \sigma_f(1 - 1/\kappa) = 0.276\times(1 - 1/2.6) = 0.17\,\mu\mathrm{C/m^2}$ — the polystyrene faces carry $62\%$ of the free charge, in molecular ends.

Step 4. Cross-check by the audit. $E = (\sigma_f - \sigma_b)/\varepsilon_0 = (0.276 - 0.17)\times10^{-6}/8.854\times10^{-12} = 12\,\mathrm{kV/m}$ ✓. The battery sees a better capacitor; the gap sees a screened field; the two statements are the same bookkeeping read from opposite ends.

2 · Designing for breakdown — why mica beats air by a factor of 6000 (exam level)

A pulse circuit needs $C = 1.0\,\mathrm{nF}$ holding $5.0\,\mathrm{kV}$ . Compare an air-gap design with a mica design ( $\kappa = 5.4$ , $E_{\text{bd}} \approx 100\,\mathrm{MV/m}$ ; air: $\kappa \approx 1$ , $E_{\text{bd}} \approx 3\,\mathrm{MV/m}$ ), running each at its breakdown-limited minimum thickness.

Step 1. Minimum gap is set by field, not voltage: $d_{\min} = V/E_{\text{bd}}$ . Air: $5000/3\times10^6 = 1.67\,\mathrm{mm}$ . Mica: $5000/10^8 = 50\,\mu\mathrm{m}$ — thirty-three times thinner.

Step 2. Plate area for $1.0\,\mathrm{nF}$ from $A = Cd/(\varepsilon_0\kappa)$ . Air: $A = (10^{-9})(1.67\times10^{-3})/(8.854\times10^{-12}\times1) = 0.19\,\mathrm{m}^2$ — a coffee-table of foil. Mica: $A = (10^{-9})(5.0\times10^{-5})/(8.854\times10^{-12}\times5.4) = 10.5\,\mathrm{cm}^2$ — a postage-stamp stack. Same charge stored ( $Q = CV = 5.0\,\mu\mathrm{C}$ ), same energy ( $\tfrac12CV^2 = 12.5\,\mathrm{mJ}$ ), volume smaller by $A_{\text{air}}d_{\text{air}}/A_{\text{mica}}d_{\text{mica}} \approx 6000$ .

Step 3. See the 6000 directly in the figure of merit. $u_{\max} = \tfrac12\varepsilon_0\kappa E_{\text{bd}}^2$ : air $= \tfrac12(8.854\times10^{-12})(1)(3\times10^6)^2 = 40\,\mathrm{J/m^3}$ ; mica $= \tfrac12(8.854\times10^{-12})(5.4)(10^8)^2 = 2.4\times10^5\,\mathrm{J/m^3}$ . Ratio $\approx 6000$ — of which only $5.4$ came from $\kappa$ ; the rest, a factor $(100/3)^2 \approx 1100$ , came from dielectric strength.

Step 4. Read the design lesson. In energy-dense capacitor engineering, breakdown field beats dielectric constant, because it enters squared. High- $\kappa$ ferroelectric ceramics ( $\kappa \sim 10^3$ ) win a different game — miniaturizing capacitance, not energy — and pay in nonlinearity and temperature drift, exactly the beyond-linear territory flagged for Ferroelectrics and Spontaneous Polarization.

Going further

References. Griffiths, Introduction to Electrodynamics, 4th ed., Ch. 4 · Jackson, Classical Electrodynamics, 3rd ed., Ch. 4 · Purcell & Morin, Electricity and Magnetism, 3rd ed., Ch. 10 · Zangwill, Modern Electrodynamics, Ch. 6. Key relations used throughout: $\mathbf{D}=\varepsilon_0\mathbf{E}+\mathbf{P}$ ; bound charges follow from the integration-by-parts derivation; $C=\varepsilon_r C_0$ from $D=\sigma_f$ . Numerical permittivities: water static $\approx80$ , optical $\approx1.77$ ; air $\approx1.0006$ ; air breakdown $\approx3\times10^6\,\mathrm{V/m}$ ; $\varepsilon_0=8.854\times10^{-12}\,\mathrm{F/m}$ . The Curie $1/T$ orientational scaling and $n=\sqrt{\varepsilon_r}$ are standard results. Worked numbers: $C_0 = 22.1$ pF, $C = 57.6$ pF, $Q_f = 0.69$ nC, $\sigma_f = 0.276\,\mu$ C/m², $\sigma_b = 0.17\,\mu$ C/m²; design example $d_{\min}$ 1.67 mm vs 50 $\mu$ m, areas 0.19 m² vs 10.5 cm², $u_{\max}$ 40 vs $2.4\times10^5$ J/m³ (ratio ≈ 6000), stored energy 12.5 mJ.

Builds on: Conductors in Electrostatic Equilibrium · Gauss's Law: Applications and Symmetry · Electric Dipole Moment · Multipole Expansion of the Potential

Leads to: Induced Dipoles and Atomic Polarizability · Permanent Dipoles and Orientational Polarization · Polarization Density P · Bound Charges: Surface and Volume · The Electric Displacement D · Linear Dielectrics, Susceptibility and Permittivity

Is there any difference between an "insulator" and a "dielectric"?

Same materials, different emphasis. "Insulator" names what the material refuses to do — conduct; "dielectric" names what it does — polarize. In circuit contexts you care that your cable jacket blocks current (insulator); in capacitor and optics contexts you care how strongly the material polarizes (dielectric, quantified by $\kappa$ ). Every dielectric is an insulator below its breakdown field; the perfect vacuum is an insulator with nothing to polarize, $\kappa = 1$ , and hence the one "material" that is an insulator but a trivial dielectric.

Why doesn't the bound charge just flow away or neutralize like charge on a conductor?

Because each bound charge is the end of a molecule that is still holding on to it. The surface sheet $\sigma_b$ is not a population of independent charges sitting on the surface — it is the exposed ends of dipole chains anchored throughout the bulk. Discharge would require either breaking the molecules (that is breakdown) or removing the field that stretches them (then the polarization, and the sheet, relax away on molecular timescales). This is also why you cannot "wipe off" bound charge with a grounding wire: touch a wire to a polarized face and almost nothing flows, because the charge is not free to leave.

Can the dielectric constant ever be less than 1, so the material assists the field?

Not in statics. A static $\kappa < 1$ would mean the material polarizes against its cause — energetically, a medium that lowers its energy by anti-aligning dipoles with the field would be unstable and would have already rearranged. So static screening always reduces: $\kappa \geq 1$ , with equality only for vacuum. At optical frequencies the story changes: dispersion can push the real part of the permittivity below 1 (even negative, in metals and plasmas below the plasma frequency) because the response is dynamic — charges oscillating out of phase with the drive, not settling into equilibrium. That physics belongs to the electromagnetic-waves chapter; the static inequality stands.

Water has kappa = 80 — why isn't it the standard capacitor filling?

Because $\kappa$ is one line of the datasheet. Real water carries ions, so it conducts — a water-filled capacitor is a resistor in disguise. Even ultrapure water self-ionizes, and its polar response is lossy at AC: reorientation lags the field and dissipates energy (your microwave oven's operating principle, catastrophic in a capacitor). Practical fillings optimize the whole bundle — $\kappa$ , breakdown strength, loss, leakage, stability — and polymers and ceramics win it. Water's 80 is a biophysics superpower (it is why salt dissolves), not an engineering material.

If lab fields distort atoms by only parts per million, how can the effect matter at all?

Because the per-atom effect is multiplied by Avogadro-scale numbers: a parts-per-million shift of each electron cloud, summed over $\sim 3\times10^{28}$ molecules per cubic meter, produces face densities — our worked example's $0.17\,\mu\mathrm{C/m^2}$ — that reshape fields by tens of percent. Negligible individual responses, macroscopic consequences by sheer census. It is also why the macroscopic average fields of the coming articles are the right variables: no capacitor-scale experiment resolves single-molecule wiggles, only their sum.

Does a dielectric increase or decrease the energy stored in a capacitor?

Both, depending on what is held fixed — the fork from equation (3). At fixed voltage, $W = \tfrac12CV^2$ rises by $\kappa$ : the battery pumped in extra charge. At fixed charge, $W = Q^2/2C$ falls by $\kappa$ : the energy-storing field has been screened, and the "missing" energy went into work drawing the dielectric into the gap — dielectrics are pulled into strong-field regions, a force computed honestly in Forces on Dielectrics. Quoting either result without stating the constraint is the classic exam trap.

What actually happens during dielectric breakdown — is the material destroyed?

A stray free electron accelerates, gains more than an ionization energy between collisions, and each impact liberates more electrons: an exponential avalanche turns a filament of insulator into conducting plasma within nanoseconds — the spark. Gases self-heal (the plasma recombines; air "repairs" after lightning). Solids often don't: the current's heat carbonizes a permanent conductive track — Lichtenberg figures in acrylic are fossilized avalanches — which is why solid insulation fails terminally while gas-insulated switchgear survives flashover. Engineering margins run a factor 3–10 below $E_{\text{bd}}$ .