Skip to content

A Dielectric Skin Screens the Conductor

Irodov 3.75DifficultyDifficulty 2 of 5UniquenessUniqueness 2 of 5

Problem Irodov 3.75

A conductor carrying surface charge density σ\sigma is in contact with a dielectric of permittivity ε\varepsilon. Show that the bound charge on the dielectric’s surface where it touches the conductor has the surface density σ′=−σ(ε−1)/ε\sigma' = -\sigma(\varepsilon - 1)/\varepsilon.

A small piece of the interface: the conductor below carries the surface charge σ, the dielectric of permittivity ε lies above it, and n is the normal out of the conductor (drawn for σ > 0).

A small piece of the interface: the conductor below carries the surface charge σ\sigma, the dielectric of permittivity ε\varepsilon lies above it, and n\vect n is the normal out of the conductor (drawn for σ>0\sigma > 0).

Given
  • surface charge density σ\sigma on the conductor
  • permittivity ε\varepsilon of the dielectric in contact
Find
show that the bound charge density on the interface is σ′=−σ(ε−1)/ε\sigma' = -\sigma(\varepsilon - 1)/\varepsilon

Solution

Step 1The displacement just outside the conductor

Inside the conductor E=0\vect E = 0, so D=0\vect D = 0 there too. The tangential component of E\vect E is continuous across any boundary and vanishes inside, so just outside the field, and with it D\vect D, is normal to the surface. Gauss’s theorem for D\vect D over a flat pillbox of face area ΔS\Delta S straddling the surface counts only the free charge σ ΔS\sigma\,\Delta S it encloses, and only the outer face carries flux:

Dn ΔS=σ ΔSD_n\,\Delta S = \sigma\,\Delta S

so with n\vect n the normal out of the conductor,

D=σ n(1) \vect D = \sigma\,\vect n \tag{1}
Step 2The polarization of the dielectric there

In a linear isotropic dielectric D=ε0εE\vect D = \varepsilon_0\varepsilon\vect E, so the polarization is the part of D\vect D beyond the vacuum term:

P=D−ε0E=D−Dε=ε−1ε D\vect P = \vect D - \varepsilon_0\vect E = \vect D - \frac{\vect D}{\varepsilon} = \frac{\varepsilon - 1}{\varepsilon}\,\vect D

Substituting (1):

P=ε−1ε σ n(2) \vect P = \frac{\varepsilon - 1}{\varepsilon}\,\sigma\,\vect n \tag{2}
Step 3The bound charge on the dielectric’s face

The bound surface charge of a dielectric is the normal component of P\vect P along the dielectric’s own outward normal. At the face that touches the conductor, the dielectric’s outward normal points into the conductor, that is along −n-\vect n. From (2):

σ′=P⋅(−n)\sigma' = \vect P\cdot(-\vect n)
σ′=−ε−1ε σ\sigma' = -\frac{\varepsilon - 1}{\varepsilon}\,\sigma
(3)

The bound charge is opposite in sign to the conductor’s charge and smaller in magnitude.

The polarization P points away from the conductor, and the bound charge σ' just inside the dielectric is opposite to σ and sparser (drawn for ε = 3, where |σ'| = 2σ/3).

The polarization P\vect P points away from the conductor, and the bound charge σ′\sigma' just inside the dielectric is opposite to σ\sigma and sparser (drawn for ε=3\varepsilon = 3, where ∣σ′∣=2σ/3|\sigma'| = 2\sigma/3).

Check

Gauss’s theorem for E\vect E, which counts free and bound charge together, gives the same result independently. The field just outside is D/(ε0ε)D/(\varepsilon_0\varepsilon), so the pillbox sees the total charge

σ+σ′=ε0En=σε\sigma + \sigma' = \varepsilon_0 E_n = \frac{\sigma}{\varepsilon}

and the boxed σ′\sigma' satisfies it. For ε=1\varepsilon = 1 there is no bound charge, as there is no medium.

Remark

The two layers together carry σ/ε\sigma/\varepsilon: the dielectric screens all but a fraction 1/ε1/\varepsilon of the conductor’s charge, which is why the field next to the conductor drops by the factor ε\varepsilon and why filling a capacitor raises its capacitance by the same factor. As ε\varepsilon grows without bound, σ′\sigma' tends to −σ-\sigma and the screening becomes complete, as it would be for a second conductor.