Electric dipole moment (source code)

= Electric dipole moment
{wiki}

The <electric dipole moment> measures the first spatial moment of signed <electric charge>. For localized <point charges> $q_i$ at positions $\mathbf r_i$, or an integrable <charge density> $\rho(\mathbf r)$ with a finite first moment, define
$$
\mathbf p=\sum_i q_i\mathbf r_i,
\qquad
\mathbf p=\int\mathbf r\rho(\mathbf r)\,d^3r,
\qquad
Q=\sum_iq_i=\int\rho(\mathbf r)\,d^3r.
$$
https://farside.ph.utexas.edu/teaching/jk1/lectures/node20.html[Richard Fitzpatrick, Electric Dipoles: continuous definition]{ref}.

The <electric dipole moment> is a <vector> with <SI> units $\mathrm{C\,m}$: <coulombs> times <length> in metres. Opposite <electric charges> can cancel in $Q$ while contributing in the same direction to $\mathbf p$. For a neutral pair with positive <electric charge> $q>0$ at $\mathbf r_+$ and negative <electric charge> $-q$ at $\mathbf r_-$,
$$
\mathbf d=\mathbf r_+-\mathbf r_-,\qquad \mathbf p=q\mathbf d.
$$
Thus the <electric dipole moment> points from negative to positive <electric charge>, not along the attraction between the two <point charges>. https://openstax.org/books/university-physics-volume-2/pages/5-7-electric-dipoles[Ling, Moebs and Sanny, University Physics 2, 5.7: charge-pair geometry and dipole moment]{ref}.

Changing the coordinate origin changes the position vectors, not the physical <electric charge> distribution. If the new origin is at $\mathbf a$ relative to the old origin, $\mathbf r_i'=\mathbf r_i-\mathbf a$, so direct substitution gives
$$
\mathbf p'=\sum_iq_i(\mathbf r_i-\mathbf a)=\mathbf p-Q\mathbf a.
$$
Consequently <charge neutrality> makes the <electric dipole moment> independent of the origin. For a system with nonzero total <electric charge>, always specify the origin: a single <point charge> at the old origin has $\mathbf p=0$ but $\mathbf p'=-Q\mathbf a$. This is coordinate bookkeeping, not a newly created <electric dipole>. The full <electric potential> stays the same at the same physical point; the terms of its <electric multipole expansion> and the observation coordinates change together.

As a stipulated numerical example, put $+2\,\mathrm{nC}$ at $(1,0,0)\,\mathrm{mm}$ and $-2\,\mathrm{nC}$ at $(-1,0,0)\,\mathrm{mm}$. Their total <electric charge> is zero and $\mathbf p=(4\times10^{-12},0,0)\,\mathrm{C\,m}$. Translating the origin by $(3,-2,0)\,\mathrm{mm}$ leaves this <electric dipole moment> unchanged. Adding $+1\,\mathrm{nC}$ at the old origin instead gives $Q=1\,\mathrm{nC}$ and $\mathbf p'=(1,2,0)\times10^{-12}\,\mathrm{C\,m}$. Distinguishing these two cases prevents the mistaken claim that every <electric dipole moment> is independent of the origin.

The usefulness of the <electric dipole moment> is compression: many internal <electric charges> determine one <vector> governing the leading distant <electric potential> of a neutral localized object when $\mathbf p\ne0$. https://www.feynmanlectures.caltech.edu/II_06.html#Ch6-S2[Feynman, Leighton and Sands, II.6.2: distant dipole fields]{ref}. A <polar molecule> need not literally contain two <point charges>; its <atomic nuclei> and <electron density> collectively give an <electric dipole moment>. For example, the <chemical polarity> of <water> reflects its <bent molecular geometry>, whereas the symmetric linear <molecular geometry> of <carbon dioxide> makes its permanent <electric dipole moment> vanish. https://openstax.org/books/chemistry-2e/pages/7-6-molecular-structure-and-polarity[Flowers, Theopold, Langley and Robinson, Chemistry 2e, 7.6: molecular geometry and polarity]{ref}. A collection of such <electric dipole moments> per unit volume gives <electric polarization>. https://farside.ph.utexas.edu/teaching/jk1/Electromagnetism/node40.html[Richard Fitzpatrick, Polarization: dipole moment per unit volume]{ref}.