Back

Legendre Polynomials and Multipoles

Far from any lump of charge, its potential settles into a tidy series, the multipole expansion, and the angular shape of each term is a Legendre polynomial $P_l(\cos\theta)$. The pieces are familiar by name: the monopole is the same in every direction, the dipole is $\cos\theta$ with a single belt of zero around the equator, the quadrupole adds another, and each higher multipole carves the sphere into more zones. The reason is in the polynomial: $P_l$ has exactly $l$ zeros between $-1$ and $1$, and each zero is a nodal cone, a direction where that multipole contributes nothing, separating $l+1$ lobes that alternate in sign. The scene draws those lobes as a polar diagram around the vertical axis, red where $P_l$ is positive and blue where it is negative, with the nodal cones marked; the whole figure is what you get spinning this cross-section about the axis, since the pure multipoles do not depend on the azimuth. A probe sweeps the polar angle and ties the radius of the lobe to the value of the polynomial in the panel below, where $P_l(x)$ with $x = \cos\theta$ is plotted with its $l$ roots, the cosines of the cone angles. Every $P_l$ starts at $1$ when $\theta = 0$ and ends at $\pm 1$ at the south pole, and any two different ones are orthogonal, which is exactly why this set is the natural alphabet for expanding a field on a sphere. Step the multipole and watch a lobe and a cone appear with each increment of $l$.

Figure 1. Legendre polynomials and multipoles. Top: the angular shape P_l(cos theta) as a polar lobe diagram about the vertical axis, red positive and blue negative, with the nodal cones (dashed) and a sweeping probe. Bottom: the polynomial P_l(x) with x = cos theta, its l roots marked (the cone angles), starting at 1 and ending at (-1)^l. Method: Bonnet recurrence and root-finding. Source: Jackson, Classical Electrodynamics, 3rd ed., Sec. 3.2-3.3.

WHAT TO TRY

  • Step the multipole: each increment of $l$ adds one nodal cone and one lobe, the sphere carved into more alternating zones.
  • Compare the dipole and quadrupole: $\cos\theta$ has one equatorial node, $P_2$ has two cones at about 55 and 125 degrees.
  • Watch the probe: the lobe radius is exactly $|P_l(\cos\theta)|$, the value read off the curve below at $x = \cos\theta$.
  • Note the endpoints: every $P_l$ is $1$ along the axis and $\pm 1$ at the back, the polynomials normalized at the poles.