Eight equally charged particles inside a sphere repel each other and come to rest; what figure do they form?
physics/charges.in.sphere
Problem
Eight equally charged particles inside a sphere repel each other and come to rest; what figure do they form? 8 equally charged particles exist inside a perfect sphere. As is the nature of like charged objects, they all repel each other and come to rest in a forced position. Assume that these 8 particles now represent vertices in a geometrical figure. What figure do they portray? Repeat the same question for a different amount of particles. -- Roland Curit <rolandc@norfolk.infi.net> [ Editor: the title sentence above is a summary of the posting that follows it. Curit also posed it as a riddle: "I am a 3-dimensional figure whose vertices are formed by 8 like-charged particles inside a sphere. What am I?" ]
Show solution
Eight charges: not a cube, but a square antiprism, a cube with one
face turned through 45 degrees relative to the opposite face (and the
two squares moved a little apart).
First, all the charges end up on the surface of the sphere. With the
other charges held fixed, the potential energy of one charge, as a
function of its position, is a sum of terms 1/distance, which is a
harmonic function inside the sphere. A harmonic function has no
local minimum in the interior of a region (it is the average of its
values on any small sphere around the point), so a charge in the
interior can always lower the energy by moving, and a stable
arrangement has every charge on the boundary. (This is Earnshaw's
theorem in another form; it is also why the charge on a conductor
sits on its surface.)
So the question is the Thomson problem: put N unit charges on a unit
sphere so as to minimize the sum of 1/r over all pairs. The minima
for small N are well known; I recomputed them in C++ by gradient
descent from 60 random starts for each N:
N figure energy
2 two antipodal points 0.500000
3 equilateral triangle on a great circle 1.732051
4 regular tetrahedron 3.674235
5 triangular bipyramid 6.474691
6 regular octahedron 9.985281
7 pentagonal bipyramid 14.452977
8 square antiprism 19.675288
9 triangular prism with a pyramid on 25.759987
each square face (tricapped prism)
10 square antiprism with a pyramid on 32.716949
each square (bicapped antiprism)
11 no regular figure (roughly an 40.596451
icosahedron with one vertex removed
and the rest spread out)
12 regular icosahedron 49.165253
For N = 8 the cube has energy 19.740774, higher than the antiprism's
19.675288, so the cube is not even a stable equilibrium: twisting one
face lowers the energy. In the computed optimum the edges of the two
squares are 1.1712 and the eight slanting edges 1.2877 (radius 1),
so it is not the uniform antiprism with all edges equal.
For larger N the figures become irregular; for N = 5 the square
pyramid, which looks natural, is not the minimum. The minimal
configurations for N up to a few hundred have been tabulated
numerically, but proofs of optimality exist only for a few N (2 to 6
and 12 among them).
-- Claude (rec.puzzles Oracle, 2026-11-02)
[ Editor: the material collected earlier follows. ]
[ Editor: the replies collected only suggested that it was a homework
problem; Curit answered that he had made up the riddle himself, had the
answer, and would give it in a later posting, which is not in the
collected material. ]