Page 153 - Numerical methods for chemical engineering
P. 153

Eigenvalue problems in quantum mechanics                            139



                  We next integrate each of these equations over [0, 2P] to obtain a set of algebraic
                  equations:
                                           N              N

                                               b m h nm = E  b m s nm               (3.209)
                                         m=−N           m=−N
                  The coefficients s nm are
                                                 '  2P
                                                      ∗
                                           s nm =    χ (x)χ m (x)dx                 (3.210)
                                                      n
                                                  0
                  and the coefficients h nm are
                                                 2P
                                               '
                                                         ˆ
                                                     ∗
                                         h nm =    χ (x)[Hχ m (x)]dx                (3.211)
                                                     n
                                                0
                  where the Hamiltonian operator is
                                                h ¯ 2  d 2
                                        ˆ
                                        Hχ =−         χ(x) + V (x)χ(x)              (3.212)
                                               2m e dx 2
                  We compute these integrals shortly, but first let us consider the structure of (3.209), which
                  we write as
                             N             N

                                h nm b m = E  s nm b m  n = 0, ±1, ±2,..., ±N       (3.213)
                           m=−N           m=−N
                  Let us define the new indices
                       p = m + N + 1  m = 0, ±1, ±2,..., ±N  p = 1, 2,..., 2N + 1
                                                                                    (3.214)
                       q = n + N + 1  n = 0, ±1, ±2,..., ±N  q = 1, 2,..., 2N + 1
                  so that, with B = 2N + 1, we define a B × B overlap matrix S,a B × B Hamiltonian matrix
                                             B
                  H, and a coefficient vector c ∈ C , such that
                                                                                    (3.215)
                                        H pq = h mn  S pq = s mn  c p = b m
                  The system of equations (3.213) then becomes
                                     B            B

                                       H qp c p = E  S qp c p  q = 1, 2,..., B      (3.216)
                                    p=1          p=1
                  This is the qth row of the generalized matrix eigenvalue problem

                                                 Hc = ESc                           (3.217)
                  H and S are both Hermitian and S is also positive-definite. Let the energy eigenvalues be
                                     [k]
                  E k with Hc [k]  = E k Sc . These E k are the allowable energy states of the system, whose
                  corresponding normalized wavefunctions are
                                                           B
                                       [k]
                                      ϕ (x)
                           [k]                    ϕ (x) =    c χ m=p−N−1 (x)        (3.218)
                                                              [k]
                                                   [k]
                                   2P         2               p
                         ψ (x) = -
                                        [k]
                                      ϕ (x) dx
                                   0                      p=1
                  with the corresponding electron probability densities
                                               [k]
                                                        [k]

                                              ρ (x) = ψ (x)     2                   (3.219)
                                               e
   148   149   150   151   152   153   154   155   156   157   158