Page 336 - PRINCIPLES OF QUANTUM MECHANICS as Applied to Chemistry and Chemical Physics
P. 336

Series solutions of differential equations        327
                               1
                              X                                            k‡s‡1
                                 fa k‡1 [(k ‡ 1)(k ‡ 2l ‡ 2)] ‡ a k (ë ÿ l ÿ 1 ÿ k)gr  ˆ 0  (G:47)
                              kˆ0
                        Since the coef®cient of each power of r must vanish, we have for the recursion
                        formula
                                                        k ‡ l ‡ 1 ÿ ë
                                               a k‡1 ˆ                a k                 (G:48)
                                                     (k ‡ 1)(k ‡ 2l ‡ 2)
                        Thus, we obtain the following set of expansion constants

                                    a 0
                                         l ‡ 1 ÿ ë
                                    a 1 ˆ        a 0
                                          2l ‡ 2
                                         l ‡ 2 ÿ ë   (l ‡ 2 ÿ ë)(l ‡ 1 ÿ ë)
                                    a 2 ˆ        a 1 ˆ                  a 0
                                        2(2l ‡ 3)      2(2l ‡ 3)(2l ‡ 2)
                                         l ‡ 3 ÿ ë   (l ‡ 3 ÿ ë)(l ‡ 2 ÿ ë)(l ‡ 1 ÿ ë)
                                    a 3 ˆ        a 2 ˆ                           a 0
                                        3(2l ‡ 4)       3!(2l ‡ 4)(2l ‡ 3)(2l ‡ 2)
                                      .
                                      .
                                      .
                        so that the solution of (G.43) is
                                                                                   !
                                           1
                                          X  (l ‡ k ÿ ë)(l ‡ k ÿ 1 ÿ ë)     (l ‡ 1 ÿ ë)
                             F ˆ a 0 r l  1 ‡                                    r k      (G:49)
                                                k!(2l ‡ k ‡ 1)(2l ‡ k)     (2l ‡ 2)
                                          kˆ1
                        We have already discarded the second solution, equation (G.46).
                          The ratio of consecutive terms in the power series expansion F is given by equation
                        (G.48) as
                                            a k‡1 r k‡l‡1  k ‡ l ‡ 1 ÿ ë
                                                      ˆ                  r
                                              a k r k‡l  (k ‡ 1)(k ‡ 2l ‡ 2)
                        In the limit as k !1, this ratio becomes r=k, which approaches zero for ®nite r.
                        Thus, the series converges for all ®nite values of r. To test the behavior of the power
                        series as r !1, we consider the Taylor series expansion of e r
                                                            1  r k
                                                       r   X
                                                      e ˆ
                                                               k!
                                                           kˆ0
                        and note that the ratio of consecutive terms is also r=k. Since the behavior of F as
                        r !1 is determined by the expansion terms with large values of k (k !1), we see
                                         r
                        that F behaves like e as r !1. This behavior is not acceptable because S(r)in
                        equation (G.42) would take the form
                                                          l r=2
                                               l r ÿr=2
                                       S(r) ! r e e   ˆ r e   !1       as r !1
                        and could not be normalized.
                          The only way to avoid this convergence problem is to terminate the in®nite series
                        (equation (G.49)) after a ®nite number of terms. If we let ë take on the successive
                        values l ‡ 1, l ‡ 2, ... , then we obtain a series of acceptable solutions of the
                        differential equation (G.43)
   331   332   333   334   335   336   337   338   339   340   341