Dielectrics and Electric Polarization

Although the charges inside a dielectric (insulator) are bound to their individual atoms or molecules and cannot travel through the material, an external electric field can still act on each of them locally, producing a bulk effect called polarization -- and this effect has a real, measurable influence on the field within the dielectric.

Nonpolar molecules. Molecules such as $\text{N}_2$, $\text{O}_2$, $\text{CO}_2$, or $\text{CH}_4$ are, in the absence of any external field, symmetric enough that their centre of positive charge (the nuclei) and centre of negative charge (the electron cloud) coincide exactly at one point -- such a molecule has zero net dipole moment on its own. When an external field $\vec{E}_0$ is applied, the field pulls the positive and negative charge centres apart very slightly in opposite directions, creating a small induced dipole moment in each molecule, aligned along $\vec{E}_0$.

Polar molecules. Molecules such as $\text{H}_2\text{O}$ or $\text{HCl}$, by contrast, are already asymmetric enough (due to unequal sharing of bonding electrons, and a non-symmetric molecular geometry) to possess a permanent dipole moment even with no external field present. In the absence of a field, thermal agitation keeps these permanent molecular dipoles oriented randomly in every direction, so their contributions cancel on average and the material shows no net polarization. When an external field $\vec{E}_0$ is applied, each permanent dipole experiences the orientation-dependent torque and potential energy of Section 2.9, tending to rotate it toward alignment with $\vec{E}_0$; thermal motion continually disturbs this alignment, but a net partial alignment survives, increasing with field strength and decreasing with temperature.

The net macroscopic effect, either way, is the same in kind. Whether the individual dipoles are induced (nonpolar case) or partially aligned (polar case), the interior of the dielectric still contains, at every internal point, closely spaced pairs of opposite bound charge that continue to cancel each other's effect almost completely -- EXCEPT at the two outer faces of the dielectric that lie perpendicular to the field, where uncancelled bound surface charge appears: negative bound charge on the face nearer the external field's positive source, and positive bound charge on the face nearer its negative source (see the accompanying figure). This layer of bound surface charge produces its own induced field $\vec{E}_p$, directed OPPOSITE to $\vec{E}_0$, so the net field actually present inside the dielectric is the (smaller) difference $\vec{E} = \vec{E}_0 - \vec{E}_p$. …