Now that we have that under our belts we can address the multipolar
expansion of the vector potential intelligently. To begin with, we
will write the general solution for the vector potential in
terms of the multipolar expansion for the outgoing wave Green's
function defined above:
() & = & ik &sum#sum;_L [ J_L()
&int#int;_r^&infin#infty;&mu#mu;_0 (') H_L(')^(&ast#ast;) d^3r' .
& & + . H^+_L() &int#int;_0^r &mu#mu;_0 (')
J_L(')^(&ast#ast;) d^3r' ]
where, by convention,
means that the
is
conjugated but the bessel/neumann/hankel function is not.
This is because the only point of the conjugation is to construct
from the
-sum for each
via the addition
theorem for spherical harmonics. We certainly don't want to
change
into
, which changes the time dependent behavior of
the solution13.9. Note that the integral over all space is broken
up in such a way that the Green's function expansions above always
converge. This solution is exact everywhere in space including
inside the source itself!
We can therefore simplify our notation by defining certain functions
of the radial variable:
() = &sum#sum;_L i k[ C_L(r) J_() +
S_L(r) H^+_L() ].
In this equation,
C_L(r) & = & &int#int;_r^&infin#infty;&mu#mu;_0 (') H_L(')^(&ast#ast;) d^3r'
S_L(r) & = & &int#int;_0^r &mu#mu;_0 (') J_L(')^(&ast#ast;) d^3r'.
Clearly and for , . At the origin the solution is completely regular and stationary. Outside the bounding sphere of the source distribution the solution behaves like a linear combination of outgoing spherical multipolar waves. From now on we will concentrate on the latter case, since it is the one relevant to the zones.