Quantum Mechanics I, Correction Sheet 7, Spring 2013
Responsible for this sheet: J. Guillod (julien.guillod@unige.ch), office 212, Sciences I
                           April 24, 2013 (Ecole de Physique, Auditoire Stückelberg)
                              Prof. D. van der Marel (dirk.vandermarel@unige.ch)
                 Tutorials: J. Guillod (julien.guillod@unige.ch), O. E. Peil (oleg.peil@unige.ch)
       I.   (*) FUNCTIONAL CALCULUS AND GENERALIZED COMMUTATORS
Consider an analytic function F : C → C so that
                                                           ∞
                                                           X
                                               F (x) =           fn xn .
                                                           n=0
The function F (Â) of an operator  is then defined as
                                                           ∞
                                                           X
                                              F (Â) =           fn Ân .
                                                           n=0
   1. Let |ψi be an eigenvector of  with eigenvalue a. Therefore, for all n ∈ N, we have,
                                                       Ân |ψi = an |ψi .
      so by definition of F (Â),
                                             ∞
                                             X                     ∞
                                                                   X
                               F (Â)|ψi =         fn Ân |ψi =          fn an |ψi = F (a)|ψi ,
                                             n=0                   n=0
      and |ψi is an eigenvector of F (Â) with eigenvalue F (a).
                 2. If B̂, Â , Â = 0, we prove by induction that
                                           B̂, Ân = n B̂, Â Ân−1 .
                                                                The base case is n = 1. For the inductive step, by assuming the relation true of n, we have
                     B̂, Ân+1 = B̂ Ân  − Ân ÂB̂ = B̂ Ân  − Ân B̂  + Ân B̂  − Ân ÂB̂
                             
                                = B̂, Ân  + Ân B̂,  = n B̂,  Ân + Ân B̂, Â
                                                                                                                     = (n + 1) B̂, Â Ân ,
                                                 
                 
      since B̂, Â commutes with Â. For a general function, we obtain
                          ∞               ∞                     ∞
                        X               X                    X
                                      n              n
                                                                  nfn Ân−1 = B̂, Â F 0 (Â) ,
                                                                              
             B̂, F (Â) =    B̂, fn  =    fn B̂,  = B̂, Â
                           n=0                 n=0                              n=0
      because by definition,
                                                             ∞
                                                             X
                                                   0
                                                F (x) =            nfn xn−1 .
                                                             n=0
                                                                                                         1/5
              
   3. If X̂, P̂ = i~, then we deduce
                                   X̂, T (P̂ ) = X̂, P̂ T 0 (P̂ ) = i~T 0 (P̂ ) ,
                                                           and
                                         P̂ , V (X̂) = P̂ , X̂ V 0 (X̂) = −i~V 0 (X̂) .
                                                                  These two relations will be used in the next exercise.
               II.   EHRENFEST THEOREM AND HAMILTON’S EQUATIONS
In this exercise we consider a particle in three dimensions in a potential V , and the aim is to link
and see the differences between classical and quantum mechanics. The classical Hamiltonian is
                                                        p2
                                                 H=        + V (q) ,
                                                        2m
where p = (p1 , p2 , p3 ) and q = (q1 , q2 , q3 ) are the generalized coordinates. The quantum Hamilto-
nian is
                                                        p̂2
                                                 Ĥ =       + V (q̂) ,
                                                        2m
where p̂ = (p̂1 , p̂2 , p̂3 ) and q̂ = (q̂1 , q̂2 , q̂3 ) are the momentum and position operators.
                                A.       Poisson brackets and commutators
The Poisson bracket of two classical observables is defined as
                                          3                    
                                         X    ∂A ∂B       ∂A ∂B
                              {A, B} =                 −          ,
                                              ∂qi ∂pi ∂pi ∂qi
                                                  i=1
and the commutator between two quantum observables by
                                                                           Â, B̂ = ÂB̂ − B̂ Â .
   1. By definition of the Poisson bracket, we obtain
                                       3                     X 3
                                      X    ∂qi ∂qj   ∂qi ∂qj
                         {qi , qj } =              −          =     (δik 0 − 0δjk ) = 0 ,
                                           ∂qk ∂pk   ∂pk ∂qk
                                      k=1                       k=1
                                       3                     X 3
                                      X    ∂pi ∂pj   ∂pi ∂pj
                         {pi , pj } =              −          =     (0δjk − δik 0) = 0 ,
                                           ∂qk ∂pk   ∂pk ∂qk
                                      k=1                       k=1
                                       3                     X 3
                                      X    ∂qi ∂pj   ∂qi ∂pj
                         {qi , pj } =              −          =     (δik δjk − 0) = δij .
                                           ∂qk ∂pk   ∂pk ∂qk
                                     k=1                                    k=1
   2. In the representation given by
                                                                                           ∂
                           q̂i ψ(q) = qi ψ(q) ,                          p̂i ψ(q) = −i~       ψ(q) ,
                                                                                          ∂qi
                                                                                                       2/5
   we have
                     
             q̂i , q̂j ψ(q) = q̂i q̂j ψ(q) − q̂j q̂i ψ(q) = qi qj ψ(q) − qj qi (q) = 0 ,
                                                                    ∂ ∂                 ∂ ∂
            p̂i , p̂j ψ(q) = p̂i p̂j ψ(q) − p̂j p̂i ψ(q) = −~2             ψ(q) + ~2
                     
                                                                                               ψ(q) = 0 ,
                                                                   ∂qi ∂qj            ∂qj ∂qi
                                                                   ∂              ∂
             q̂i , p̂j ψ(q) = q̂i p̂j ψ(q) − p̂j q̂i ψ(q) = −i~qi       ψ(q) + i~       (qi ψ(q))
                                                                    ∂qj            ∂qj
                                        ∂                  ∂              ∂qi
                            = −i~qi        ψ(q) + i~qi        ψ(q) + i~        = i~δij .
                                       ∂qj                ∂qj             ∂qj
   The commutation relations between position and momentum operators in quantum me-
   chanics are the analog of the Poisson brackets between position and momentum in classical
   mechanics.
                                              B.      Ehrenfest theorem
1. By using the chain rule, the evolution of a classical observable A = F (q, p, t) is given by
                         3                                    3                                
      dA   ∂A X                   ∂A        ∂A              ∂A X         ∂A ∂H ∂A ∂H                      ∂A
         =    +                       q̇i +     ṗi       =    +                 −                    =      + {A, H} ,
      dt   ∂t                     ∂qi       ∂pi             ∂t           ∂qi ∂pi   ∂pi ∂qi                ∂t
                        i=1                                    i=1
   where we used Hamilton equations,
                                            ∂H                                            ∂H
                                    q̇i =       ,                               ṗi = −       .
                                            ∂pi                                           ∂qi
2. The evolution of the expectation value of a quantum observable  evolving under the action
   of the Hamiltonian Ĥ, is
                                                                                             
      d        d                       ∂ Â            d                               d
         Â = hψ(t)|Â|ψ(t)i = hψ(t)|       |ψ(t)i +      hψ(t)| Â|ψ(t)i + hψ(t)|Â      |ψ(t)i
     dt       dt                       ∂t              dt                              dt
                          ∂ Â            1                 1
                 = hψ(t)|      |ψ(t)i − hψ(t)|Ĥ Â|ψ(t)i + hψ(t)|ÂĤ|ψ(t)i
                   *      +∂t            i~                i~
                     ∂ Â        1 D       E
                 =           +       Â, Ĥ ,
                     ∂t         i~
   where we used the Schrödinger equation and its adjoint,
                             d                                                 d
                        i~      |ψi = Ĥ|ψ(t)i ,                         −i~      hψ(t)| = hψ(t)|Ĥ .
                             dt                                                dt
                                            C.      Hamilton’s equations
1. By applying Ehrenfest theorem and the conclusions of the first exercise, we have
                     d         1 D       E   1               1
                                                     q̂, p̂2 =
                                                           
                       hq̂i =       q̂, Ĥ =                      hp̂i ,
                    dt        i~             2i~m              m
                     d         1 
                                 D        E  1            
                       hp̂i =       p̂, Ĥ =     p̂, V (q̂) = − h∇V (q̂)i .
                    dt        i~             i~
                                                                                                                      3/5
      For the classical system, the Hamilton’s equations are given by
                                d     1                                  d
                                   q = p,                                   p = −∇V (q) .
                                dt    m                                  dt
   2. For a quadratic potential
                                             mω 2 2
                                   V (q) =       q           ⇒        ∇V (q) = mω 2 q ,
                                              2
      and therefore the quantum-classical correspondence
                                    hq̂i ↔ q ,                                 hp̂i ↔ p ,
      provides an exact analogy. This correspondence is not true for a generic potential. For
      example for a quartic one,
                                                  λ 4
                                      V (q) =       q        ⇒       ∇V (q) = λq3 ,
                                                  4
      and consequently
                                        h∇V (q̂)i = λ q̂3 6= λ hq̂i3 ↔ λq3 .
                                     III.    EVOLUTION OPERATOR
The time-evolution of a quantum state |ψ(t)i ∈ E where E is an Hilbert space is given by the
Schrödinger equation
                                                 d
                                            i~      |ψ(t)i = Ĥ(t)|ψ(t)i ,
                                                 dt
where Ĥ(t) = Ĥ(t)† is the Hamiltonian of the system.
   1. The evolution operator Û (t) : E → E is defined as Û (t)|ψ(0)i = |ψ(t)i. This operator is
      linear because if
                         d                                                   d
                    i~      |a(t)i = Ĥ(t)|a(t)i ,                      i~      |b(t)i = Ĥ(t)|b(t)i ,
                         dt                                                  dt
      then
                        d
                   i~      |ψ(t)i = Ĥ(t)|ψ(t)i ,           with      |ψ(t)i = α|a(t)i + β|b(t)i ,
                        dt
      which prove that
               Û (t) (α|a(0)i + β|b(0)i) = Û (t)|ψ(0)i = |ψ(t)i
                                                  = α|a(t)i + β|b(t)i = αÛ (t)|a(0)i + β Û (t)|b(0)i .
   2. By using the Schrödinger equation, we have
                             d                   d
                        i~      Û (t)|ψ(0)i = i~ |ψ(t)i = Ĥ(t)|ψ(t)i = Ĥ(t)Û (t)|ψ(0)i ,
                             dt                  dt
                                                                                                           4/5
   and also by definition,
                                           Û (0)|ψ(0)i = |ψ(0)i .
   Since the last two relations are valid for all |ψ(0)i ∈ E, the evolution operator satisfies the
   following differential equation
                          d                                                          ˆ
                         i~ Û (t) = Ĥ(t)Û (t) ,                          Û (0) = I.
                         dt
   This equation also defines the evolution operator uniquely.
3. (*) By using the differential equation satisfied by the evolution operator and its conjugate,
   we have
                    d                  d                         
                                                                           d
                                                                                   
                 i~     Û (t)† Û (t) = i~ Û (t)† Û (t) + Û (t)† i~ Û (t)
                    dt                      dt                            dt
                                                     †
                                       = − Ĥ(t)Û (t) Û (t) + Û (t)† Ĥ(t)Û (t) = 0 ,
   and
                                           Û (0)† Û (0) = Iˆ† Iˆ = Iˆ ,
   which together prove that
                                               Û (t)† Û (t) = Iˆ .
                                                                                               ˆ In
   Strictly speaking, to prove that Û (t) is unitary, it remains to show that Û (t)Û (t)† = I.
   fact this is automatically true by the fact that Û (t) is surjective: for all |ψ(t)i ∈ E there
   exists |ψ(0)i ∈ E such that |ψ(t)i = Û (t)|ψ(0)i and therefore,
                   Û (t)Û (t)† |ψ(t)i = Û (t)Û (t)† Û (t)|ψ(0)i = Û (t)|ψ(0)i = |ψ(t)i .
4. If the Hamiltonian is time-independent, the evolution operator satisfies the linear differential
   equation
                            d
                          i~   Û (t) = Ĥ Û (t) ,                   Û (0) = Iˆ .
                            dt
   Since this differential equation is linear it is sufficient to check that
                                               Û (t) = e−iĤt/~ ,
   is a solution:
               d           d
             i~ Û (t) = i~ e−iĤt/~ = Ĥe−iĤt/~ = Ĥ Û (t) ,    Û (0) = e−iĤ0/~ = Iˆ
               dt          dt
                     
5. (*) If Ĥ(t), Ĥ(s) = 0, then the evolution operator is given as for an ordinary differential
   equation by
                                                Z t            
                                                 −i
                                  Û (t) = exp          Ĥ(s) ds .
                                                 ~ 0
   However, if the Hamiltonian does not commute at different times, then
            Z t          2       Z t            Z t                          Z t
         d
                  Ĥ(s) ds = Ĥ(t)     Ĥ(s) ds +     Ĥ(s) ds Ĥ(t) 6= 2Ĥ(t)     Ĥ(s) ds ,
        dt      0                   0              0                            0
   and the evolution operator does not satisfies the differential equation.
                                                                                                 5/5