1

Quantum Mechanics Note 1 (modified 3/9/20)

(A) Axiom #1 : (The state Axiom)

The state of a system is specified by a wave function  . For a system consisting of a


single point particle,  is a complex function of the position coordinate r of the particle

and the time t:    (r, t ) specifies the state of the system at time t. The temporal
development of the wave function is governed by the Schrödinger equation:

 (r , t )   2 2   
i      V (r , t )   (r , t ) .
t  2m 


(B) Statistical interpretation of the wave function  (r, t ) : (Max Born)

For a particle in the state  (r , t ) ,
   
P(r, t ) d3r  | (r, t ) |2 d3r (1)
 
gives the probability of finding the particle in a volume element d 3r about point r , at
time t. The definition of probability at once implies the following normalization condition
for  :
 
 | (r, t ) | d3r  1.
2
(2)

The integration is over the whole configuration space of the system. A wave function
satisfying the condition (2) is said to be normalized or normed. According to Eq. (1) the
position of a particle (at a given time t) is not uniquely determined but only given by a
probability distribution even when the particle is in a definite state, i.e. in a state fully
specified according to quantum mechanics.

(C) Axiom #2 : (Linear superposition principle)

 
If  1   1 (r, t ) and  2   2 (r, t ) are two possible states of a system, so is
  c1 1  c2  2 ,
where c1 and c2 are (arbitrary) complex numbers.
 2

(D) Definition : (Observable)

Quantities such as the position, momentum or energy of a particle, which can be

measured experimentally, are called observables.

(D1) Axiom #3 : (Observable-operator Axiom )

To each observable there corresponds a linear operator. The operators corresponding to

 
the position coordinate r and the momentum p of a particle are, respectively, given by
   
rˆ  r , p̂  i .
   
More generally, the operator corresponding to a function of r and p , F (r , p) , is given
by
   
Fˆ  F (rˆ , pˆ )  F (r,  i) .

(E) Problem :
The one-dimensional motion of a particle, restricted to the region 0  x  L , is described
by the wave function,

 ( x)  A sin x e  i E t /  ,
L
where A is a normalization constant and E is the energy of the system in this state.
Obtain the probability that the particle is (i) in the interval 12 L  x  L , (ii) in the interval
1
4
L  x  34 L .

Solution. Normalization constant A :

L L

1   |  |2 dx  | A |2  sin 2 x dx
0 0
L
1  2 
L
 | A |2 
2 0
1  cos x  dx
L 
L
1 L 2 
 | A| x
2
sin x
2 2 L  0

L
 | A |2
2
 A 2 L
 3

L L
L  2  2 1 L 2  1 L
P  ,
2
L  
 L L 2 
sin 2 x dx 
L
x
L 2 2
sin x
L  L2
  L    0.5
L 2
3L 4
  L 3L   2  2
3L 4
2 1 L  1  3L L L L 
P   ,     sin x dx  x  x        0.82
2
sin
 4 4  L L4
L L 2 2 L  L4
L  4 2 4 2 

(F) Axiom #4 : (Expectation-value Axiom)


For a system in the normed state  (r, t ) , the expectation value at time t of the observable
A, represented by the operator  , is given by

  
 A t    * (r, t ) Aˆ  (r, t ) d3r   | A  . (1)


(F1) Physical meaning of Axiom 4 :

Equation (1) is a statistical statement. It says nothing about the result of a single
observation but only about the expectation value, i.e. the mean value, obtained from
many repeated measurements. Before each of these measurements we must prepare the

system to be in the state  (r, t ) . Averaging over the results of these measurements, we
obtain a mean value. The axiom claims that this experimentally measured mean value is
the same as that given by the integral in Eq. (1).

(F2) Lemma :  A  B   A   B.

Proof.  A  B   | ( Aˆ  Bˆ )    | Aˆ     | Bˆ     A   B .

(G) Definition: (Inner or scalar product of  and  )



 |   :   * d 3r.


(G1) Theorem :

(i).  |     |  .
(ii).  |   > 0,  |    0    0.
(iii). (a).  | a  b   a  |    b b |   ; (b). a  b |   a  |   b   |  .
 4

Proof.
*
  
  
(i)  |  *     * d 3r    ( *) * * d 3r   *  d 3r   |   .
    

  
(ii) Let  (r )  f (r )  ig (r ) , for some real-valued functions f and g. Then

 2 f 2  g 2 > 0;  2 0  f  g  0    0 .
Therefore

 |      2
d3r > 0,  |    0    0.


(iii) (a).  | a  b     * (a  b ) d 3 r
 
 a   * d 3r  b   *  d 3r
 a  |    b  |   .

(b). Taking complex conjugate of both sides (b) follows.

(H) Definition and importance of Hermitian operator :

Observables such as momentum or energy are real quantities. Hence the expectation
value  A of an observable must be real for any state  :

 A   A * ,
  
3 3 3
or,  * Aˆ  d r   ( Aˆ  ) * d r   ( Aˆ  ) *  d r ,
  

or, in inner-product notation,

 | Aˆ     Aˆ  |   . (2)

An operator which satisfies condition (2) for all states  is called Hermitian (or self-
adjoint). We therefore conclude that observables must be represented by Hermitian
operators.
 5

(H1) Theorem : (Equivalence of two Hermiticity conditions )

[  | Aˆ     Aˆ  |   for any state  ] (3)


[  1 | Aˆ  2    Aˆ  1 | 2  for any two states  1 and  2 ] (4)

Proof. (: ) Suppose condition (3) holds for operator  . Substituting the state
2
  c1  1  c2  2   cn n ,
n 1

where c1 and c2 are arbitrary complex numbers, into (3) gives

 2m1 cm m | Aˆ ( 2n 1 cn n )   Aˆ ( 2m1 cm m ) | ( 2n 1 cn n ) ,

2 2
or,  *
cm cn  m | Aˆ  n    cm cn  Aˆ  m | n  ,
m, n1 m,n1

2
or,  cm* cn [ m | Aˆ  n    Aˆ  m | n ]  0 . (5)
m, n1

In deriving this result, we assumed that the operator  is linear :

Aˆ (c1 1  c2  2 )  c1 ( Aˆ  1 )  c2 ( Aˆ  2 ) .

Since Eq. (5) holds for arbitrary c1 and c2 ,

 m | Aˆ  n    Aˆ  m | n , (m, n  1, 2).

With m  1 and n  2 , condition (4) follows for  .

(: ) Conversely, let condition (4) holds for  . Setting  1   2   into (4),
condition (3) follows for  . QED.
 6

(H2) Theorem :
(i) If operators  and B̂ are Hermition, then so is Aˆ  Bˆ .
(ii) Position, momentum and Hamiltonian operators are Hermitian.
Proof.

(i)  | ( Aˆ  Bˆ )     | ( Aˆ   Bˆ  )   | Aˆ     | Bˆ  
  Aˆ  |     Bˆ  |     Aˆ   Bˆ  |    ( Aˆ  Bˆ ) |  .

(ii) Hermiticity of xˆ  x (or any real number) :

 | x     x dx   x  dx   x  dx ( x ) dx   x |   ( x  x) .

d
Hermiticity of px  :
i dx

If f  x  , g  x   L2 , then


dg
ˆ
f | pg  f
*
dx
 i dx

  df 
*


i
*
f g |

+   g dx
  i dx 

 pˆ f | g ,

where we have used integration by parts and throw away the boundary terms, as

f     g     0

2
d2
Hermiticity of Hˆ   V  x :
2m dx 2


  2
d2 
f | Hˆ g   f *   2
 V  g dx
  2m dx 

2
 2
* d g

  f dx   f *Vg dx
2m  dx 2 
 7

Integrating by parts twice:

  
 f * d g dx  f * dg 
2
df * dg
  dx
 dx 2 dx   dx dx
 
dg df *   d 2 f *
 f *
 g  | g dx
dx  dx   dx 2


d2 f *
 g dx .
 dx 2

Therefore


*
ˆ     
2
d2 f
f | Hg 2
 Vf  g dx  Hˆ f | g
  2m dx 

(I) Definition: (Eigenfunction and eigenvalue of an operator)

An equation of the type

Aˆ   a ()

tells us that the operator  acting on the state ψ simply multiplies ψ by the number a. A
state ψ for which equation () holds is called an eigenstate (or eigenfunction) of Â, and
a is the corresponding eigenvalue. Equation () is known as an eigenvalue equation. It
may happen that the eigenvalue a of the Hermitian operator  possesses more than one
eigenfunction. The eigenvalue a is called s-fold degenerate if there exist exactly s
linearly independent eigenfunctions with this eigenvalue. The collection of all the
eigenvalues of an operator is called its spectrum.

(I1) Lemma: Aˆ   a   A  a .

Proof.  A   | Aˆ     | a   a  |    a .
 8

(I2) Theorem :

For a system in an eigenstate  of the Hermitian operator Â, measuring the observable A

necessarily gives the corresponding eigenvalue for the result of the measurement.
Conversely, if for a system in the state  measuring A gives the unique result a, then 
is an eigenstate of  and a is the corresponding eigenvalue.

Proof. For a system in the normed state  , repeated measurements of an observable A

lead to a spread of results with the mean value  A . As a measure of this spread we take
the standard deviation A , defined by

(A) 2   ( A   A ) 2 
  | ( Aˆ   A ) 2 
  | ( Aˆ   A ) ( Aˆ   A )  
 ( Aˆ   A ) | ( Aˆ   A ) 
2 
or, (A) 2   ( Aˆ   A d 3r , (6)

where we have used the Hermiticity of  and  A , and hence also of Aˆ   A . If every
measurement gives the unique value a, their average is also a, i.e.  A  a , and A must
vanish. It then follows from Eq. (6) that

(A  0)  ( Aˆ   a for some number a, and  A  a) .

That is, for a system in an eigenstate  of the Hermitian operator Â, a measurement of

the observable A necessarily gives the corresponding eigenvalue for the result of the
measurement, i.e. A  0 . Conversely, if for a system in the state  measuring A gives the
unique result a (i.e. A  0 for this state), then  is an eigenstate of  and a is the
corresponding eigenvalue.

(I3) Theorem :
In case of discrete spectra show that
(i) The eigenvalues of Hermitian operators are real, and
(ii) The eigenfunctions of Hermitian operators belonging to different
eigenvalues are orthogonal.
 9

Proof.

(i) If  is Hermitian,  A is real for any state; in particular for an eigenstate with the
eigenvalue a, for which  A  a . Hence a is real.

(ii) If  i and  j are two eigenstates of the Hermitian operator  corresponding to

different (real) eigenvalues ai and a j , then

Aˆ  i  ai  i , Aˆ  j  ai  j , (ai  a j ) . (7a,b)

Hence,

(ai  a j )  i |  j   ai  i |  j   a j  i |  j 
  ai*  i |  j    i | a j j 
  ai  i |  j    i | a j j 
  Aˆ  i |  j    i | Aˆ  j 
 0, (8)

where we have used the Hermiticity of  and the reality of the eigenvalues of Â. Since
ai  a j , it follows from Eq. (8) that

 i | j   0 , (i  j ),

so that eigenfunctions belonging to different eigenvalues are orthogonal.

(J) Axiom #5 : (Expansion Postulate)

For a Hermitian operator Â, the totality of its linearly independent eigenfunctions

 1 ,  2 , ... ,  n , ... (which will be taken as orthonormal) form a complete set. By this
one means that if  is any state of a system and A is an observable of the system, then
 can be expanded in terms of the eigenfunctions of the corresponding Hermitian
operator  :

   cn  n . (1)
n
 10

Taking the inner product of Eq. (1) with  m and using the orthonormality relation
 m |  n    m n , one obtains the expansion coefficients:

 m |     m |  cn  n    cn   m |  n    cn  m n  cm ,
n n n

or, cn   n |   . (2)

(J1) Theorem: A measurement of an observable A yields only the eigenvalues of Â.

Proof. For a system in the normed state  , the expectation value of the observable A,
represented by the Hermitian operator Â, is given by
 A    | Aˆ   (3)
Expanding  in the complete set of orthonormal eigenfunctions  n  of Â,

   cn  n , (4)
n
with
cn   n   ,
we obtain

 A    m
cm  m | Aˆ ( n cn  n )    cm cn  m | Aˆ  n  .
m n
(5)

Using the eigenvalue equation

Aˆ  n  an  n

and the orthonormality relation

 m  n    m n , (6)
we can write

 m | Aˆ  n    m | an n   an  m |  n   an m n .

Hence Eq. (5) reduces to

 A    | cn | 2 a n . (7)
n
 11

Let us assume that all the eigenvalues a n are non-degenerate. We know that for a system
in the state  n a measurement of A necessarily gives the value a n . Equation (7) for the
mean value  A then leads us to interpret

P (an )  | cn | 2  |  n |   | 2 (8)

as the probability that a measurement of A on a system in the state  yields the value a n .
Furthermore, we can write the normalization condition

 |    1 ,

using the expansion (4) and the orthonormality relation (6), in the form :

 cn 2  1 .
n

On account of Eq. (8), the last equation can be written

 P ( a n ) 1 ,
n

from which we must conclude that the only values which can be obtained when
measuring A are the eigenvalues a1 , a2 , ... .

(Note: Other values, e.g. somewhere in between a1 and a2, can never be observed.)

(K) Momentum eigenstates and eigenvalues :

The momentum eigenstates are the solutions of the eigenvalue problem

     
pˆ  p (r )  i p (r )  p p (r ) .

One sees at once that for any momentum vector p , the plane wave
  
i pr /
 p (r )  Ae (1)
 
is a momentum eigenstate with p as eigenvalue. Thus the momentum operator p̂ has a
continuous eigenvalue spectrum :    px , py , px  .
 12

Unfortunately, the normalization integral


 
 | p (r) |
2
d 3r


of the eigenstate (1) is infinite. And yet we do recover a kind of ersatz orthonormality for
these eigenstates as follows:

    
     
 e

i (p  p )  r /  3
 p*  (r )  p (r ) d 3r  | A |2 d r  | A |2 (2 )3  (3) (p  p) .
 

If we choose A  1/(2 )3 / 2 , so that

  
1 i p r / 
 p (r )  e , (2)
(2 )3 / 2
then
 
 p  | p    (3) (p  p) . (3)

This is the desired orthonormality relation for continuum eigenfunctions. Instead of the
Kronecker delta, it contains the Dirac delta function.


Most important, the momentum eigenstates are complete. If  (r ) is a
normalizable state, its representation as a superposition of plane waves (to form a wave
packet) is given by the Fourier integral:

 
    1  i p  r /  3 
 (r)    (p)  p (r) d p  (2 )3 / 2 

3
(p )e d p. (4)



The expansion coefficient  (p) is obtained as follows:
 
      
 |     (p)  p  | p  d3 p 

p   (p) 
( 3)
(p  p) d3 p   (p) .
 


Alternatively, we can get the same expression for  (p) from Plancherel’s theorem, for
the expansion (4) is nothing but a Fourier transform.
 13

(L) Conservation of probability:

To begin with,

  
d       *   3 
dt  |  (r, t ) |2 d 3r   t
( * ) d 3r    t
  * dr
t 
(1)
  

Now the Schrödinger equation says that

 1 ˆ
 H , (2a)
t i
so that
 * 1
  ( Hˆ  ) * . (2b)
t i

  *
Substituting for and from Eq. (2) in Eq. (1), we obtain
t t
 
d 2 3 1 
dt  | | d r  
i  [( Hˆ  ) *  * ( Hˆ  )] d 3r .
 

But the Hamiltonian operator Ĥ must satisfy the Hermiticity condition:

 
 
 
( Hˆ  ) * d 3r   * ( Hˆ  ) d 3r .
 

Hence,

d 

dt  
|  | 2 d3r  0 .


In other words, once  (r , t ) is normalized at a given time t, it must remain so at all
future times.
 14

(M) Consider a system for which the Hamiltonian Ĥ has no explicit time dependence.
   i E t / 
Find conditions on  (r ) and E for which  (r, t )   (r) e is indeed a solution of
the Schrödinger equation.

Answer. Substituting the expression

  i E t / 
 (r, t )   (r) e (3)

in the Schrödinger equation


Hˆ   i (4)
t
we obtain

 i E t /   i E t /   i E t / 
Hˆ  (r) e  i  (r) e (iE / )  E  (r) e ,

or,
 
Hˆ  (r, t )  E (r, t ) , (5)

or, equivalently,
 
Hˆ  (r )  E (r ) , (6)

Equation (6) tells us that an expression of the form (3) is indeed a solution of the

Schrödinger equation (4), provided  (r ) is an eigenfunction of the Hamiltonian Ĥ with
the eigenvalue E: in the state (3) the system has the definite energy E by virtue of
eigenvalue equation (5). Equation (6) is called the time-independent Schrödinger
equation, in contrast to the time-dependent Schrödinger equation (4).

(M1) Definition : (Stationary state)

States for which the physical properties of a system are constant in time are called
stationary states.
 15

(M2) Theorem: Energy eigenstates are stationary states.

Proof. For the energy eigenstates

  i E t / 
 n (r, t )  n (r) e n
the probability distribution
   
P (r, t ) d3r  | n (r, t ) |2 d3r

reduces to the time-independent probability distribution

   
P (r, t ) d3r  |n (r) |2 d3r .

More generally, for a system in the state  n , the expectation value of an observable A at
the time t is given by

  
 A t   n* (r , t ) Aˆ  n (r , t ) d 3r


  
   n* (r ) e i Ent /  Aˆ  n (r ) e i Ent /  d 3r


  
   n* (r ) Aˆ  n (r ) d 3r .


If the operator  has no explicit time dependence, then  A t is constant in time.

(M3) Problem:

For a system, consider the superposition of two states given by

  
 (r, t )  c1 1 (r) ei E1t /   c2 2 (r) ei E2t /  (7)

(With 1 and 2 orthonormal and |c1|2 + |c2|2 = 1,  is normalized.) Calculate the

expectation value of an observable A in this state and discuss the result.

Solution. Measuring the energy of the system in this state gives either the value E1 or E2:
we can interpret |E1 – E2| as the energy uncertainty of the system in the state  in (7). For
this state,  A t is given by
 16

 
 A t   (r, t ) | A (r, t )
   
 c1  1 (r ) ei E1t /   c2 2 (r ) ei E2t /  | Aˆ (c1  1(r ) ei E1t /   c2 2 (r ) ei E2t /  )
 | c1 |2 1 | Aˆ 1   | c2 |2 2 | A2   c1c2 1 | A2  ei  t  c2 c1 2 | A1  ei  t

 | c1 |2 A11  | c2 |2 A22  c1c2 A12 ei  t  c2 c1 A21 ei  t , (8)

where
  ( E1  E2 ) /  ,
and
Amn  m | Aˆ n  .
*
Using the Hermiticity condition we have A21  A12 . This allows us to simplify Eq. (8) to

 A t  | c1 |2 A11  | c2 |2 A22  2 Re [c1c2 A12 ei t ]

This expectation value oscillates between the extreme values

| c1 |2 A11  | c2 |2 A22  2| c1c2 A12 |

with the period
2 2 
T  .
|  | | E1  E2 |

This oscillatory behaviour is typical of non-stationary states. The period T characterizes

how quickly the system changes with time. The larger the energy uncertainty |E1 – E2|,
the more rapidly do the properties of the system in the state  in (7) change.

(N) Axioms of Quantum Mechanics:

Axiom 1. (The state axiom)

The state of a system is specified by a wave function  . For a system consisting of a


single particle,  is a complex function of the position coordinate r of the particle and

the time t :    (r , t ) specifies the state of the system at time t. The temporal
development of the wave function is governed by the Schrödinger equation:

 (r, t )   2 2   
i      V (r, t )   (r, t ) .
t  2m 
 17

Axiom 2. (Linear superposition principle)

 
If  1   1 (r, t ) and  2   2 (r, t ) are two possible states of a system, so is

  c1 1  c2  2 ,

where c1 and c2 (arbitrary) are complex numbers.

Axiom 3. (Observable-operator axiom)

To each observable there corresponds a linear operator. The operators corresponding to

 
the position coordinate r and the momentum p of a particle are, respectively, given by
   
rˆ  r , p̂  i .
   
More generally, the operator corresponding to a function of r and p , F (r , p) ,
is given by
   
Fˆ  F (rˆ , pˆ )  F (r,  i) .

Axiom 4. (Expectation-value Axiom)


For a system in the normed state  (r, t ) , the expectation value at time t of the observable
A, represented by the operator  , is given by

  

 A t   * (r, t ) Aˆ  (r, t ) d 3r   | Aˆ  


Axiom 5. (Expansion Postulate)

For a Hermitian operator Â, the totality of its linearly independent eigenfunctions

 1, 2 ,..., n ,... (which will be taken as orthonormal) form a complete set. By this one
means that if  is any state of a system and A is an observable of the system, then  can
be expanded in terms of the eigenfunctions of the corresponding Hermitian operator Â:

   cn  n .
n