Page 511 - Electromagnetics
P. 511

of order n. From (A.143), we see that 4
                                                                   2
                                                                n
                                                            1 d (η − 1) n
                                                   P n (η) =
                                                            n
                                                           2 n!   dη n
                        is a polynomial of degree n, and that
                                                                      d m
                                                m
                                                                 2 m/2
                                                          m
                                              P (η) = (−1) (1 − η )      P n (η).
                                                n
                                                                     dη m
                        Both the associated Legendre functions and the Legendre polynomials obey orthogonality
                        relations and many recursion formulas.
                          In problems where the z-axis is included, the product %(θ)#(φ) is sometimes defined
                        as the spherical harmonic
                                                     $
                                                       2n + 1 (n − m)!  m     jmθ
                                           Y nm (θ, φ) =             P (cos θ)e  .
                                                                      n
                                                         4π  (n + m)!
                        These functions, which are complete over the surface of a sphere, were treated earlier in
                        this section.
                                            2
                          Remembering that k = ν(ν + 1), the r-dependent equation (A.139) becomes
                                            r
                                          1 d     2 dR(r)        2  n(n + 1)
                                                r        + k +            R(r) = 0.           (A.144)
                                           2
                                          r dr      dr              r 2
                        When k = 0 we have
                                               2
                                              d R(r)  2 dR(r)   n(n + 1)
                                                    +         −        R(r) = 0
                                               dr 2   r  dr       r 2
                        so that
                                                             n
                                                    R(r) = A r r + B r r  −(n+1) .
                                                         √
                                                  ¯
                        When k  = 0, the substitution R(r) =  kr R(r) puts (A.144) into the form
                                                                           2
                                           2 ¯
                                                     ¯
                                          d R(r)   d R(r)    2 2       1
                                                                             ¯
                                         2
                                        r       + r      + k r − n +        R(r) = 0,
                                           dr 2      dr                2
                        which we recognize as Bessel’s equation of half-integer order. Thus
                                                          ¯
                                                          R(r)   Z  n+ 1 (kr)
                                                   R(r) = √   =    √ 2   .
                                                            kr      kr
                        For convenience we define the spherical Bessel functions
                                                 )
                                                   π
                                           j n (z) =  J  1 (z),
                                                   2z  n+  2
                                                 )                   )
                                                   π               n+1  π
                                                                              1 (z),
                                          n n (z) =  N   1 (z) = (−1)     J −(n+ )
                                                   2z  n+  2           2z     2
                                                 )
                                                   π
                                          (1)          (1)
                                         h (z) =     H   1 (z) = j n (z) + jn n (z),
                                          n
                                                   2z  n+  2
                                                 )
                                                   π
                                          (2)          (2)
                                         h (z) =     H   1 (z) = j n (z) − jn n (z).
                                          n
                                                   2z  n+  2
                        4 Care must be taken when consulting tables of Legendre functions and their properties. In particular,
                        one must be on the lookout for possible disparities regarding the factor (−1) m  (cf., [76, 1, 109, 8] vs.
                                                     m
                        [5, 187]). Similar care is needed with Q (x).
                                                     n
                        © 2001 by CRC Press LLC
   506   507   508   509   510   511   512   513   514   515   516