Dirac relativistic electron theory

Masatsugu Sei Suzuki

Department of Physics, SUNY at Binghamton
(Date: January 10, 2014)

I taught the advanced quantum mechanics for the graduate students about 10 years
ago. Recently I have an opportunity to read my lecture notes (hand written) on the Dirac
theory. I realized that my lecture notes are still useful to the students who study the
relativistic quantum mechanics. So I revised my old lecture notes. This time I use the
Mathematica for the calculation of the commutation relations of Dirac matrices,
eigenvalue problems, the calculation of exponent of the matrices, series expansion for the
relativistic hydrogen atom, and so on. The complicated mathematical calculations can be
replaced by simple Mathematica calculations.

_______________________________________________________________________
In particle physics, the Dirac equation is a relativistic wave equation derived by
British physicist Paul Dirac in 1928 and later seen to be an elaboration of the work of
Wolfgang Pauli. In its free form, or including electromagnetic interactions, it describes
all spin-½ particles, such as electrons and quarks, and is consistent with both the
principles of quantum mechanics and the theory of special relativity, and was the first
theory to account fully for special relativity in the context of quantum mechanics.
It accounted for the fine details of the hydrogen spectrum in a completely rigorous way.
The equation also implied the existence of a new form of matter, antimatter, hitherto
unsuspected and unobserved, and actually predated its experimental discovery. It also
provided a theoretical justification for the introduction of several-component wave
functions in Pauli's phenomenological theory of spin; the wave functions in the Dirac
theory are vectors of four complex numbers (known as bispinors), two of which resemble
the Pauli wavefunction in the non-relativistic limit, in contrast to the Schrödinger
equation which described wave functions of only one complex value. Moreover, in the
limit of zero mass, the Dirac equation reduces to the Weyl equation.
http://en.wikipedia.org/wiki/Dirac_equation

1
http://en.wikipedia.org/wiki/Paul_Dirac

We discuss the relativistic theory of electron which was derived by Dirac. Here we use
the notations

x1  x , x2  y , x3  z , x4  ict ( x0  ct )

           
p1   , p2   , p3  
i x1 i x i x2 i y i x3 i z

      i
p4     E
i x4 i  (ict ) c t c

xk  ( x, y, z ) , pk  ( p x , p y , p z )

x  ( x1 , x2 , x3 , x4 ) , p  ( p1 , p2 , p3 , p4 )

We also use the abbreviation for the summations,

3 4
x k pk   x k pk , x p   x p
k 1  1

1. One particle theory

2
We consider the simplest case, one particle with no forces. The energy of the particle is
given by

1 2
E p .
2m

Using the relation

 
E  i , p ,
t i

we get the Schrodinger equation for the wave function  ,

 2 2
i    .
t 2m

In order to give a physical meaning to  , we state that

   * .

is the probability of finding the particle at the point (x, y, z) at time t. The Probability
current density is

p  
J  Re( *  )  Re( *  )  ( *   * .
m mi 2mi

The probability is conserved because


 J  0,
t

What happens to the particle relativistically. The energy of particle is given by

E 2  m 2c 4  c 2 p 2

which leads to the Klein-Gordon equation,

1 2 m 2c 2
   2 
2

c 2 t 2 

In order to give a physical meaning to  , we must have a continuity equation. This


condition    J  0 can be satisfied only if we take J as before and
t

3
 i *   *
 (   )
2mc 2 t t

Note that

 i * 1  2 1  2 *
 ( 2 2   2 )
t 2m c t c t 2
i * 2 m 2c 2 m 2c 2
 [ (   2  )   ( 2 *  2  * )]
2m  

 ( * 2   2 )
2mi
   J

The probability density thus defined is not always positive. We have negative probability.

2 Hamiltonian
The time dependent Schrödinger equation for the particle is given by

 2 2
i   
t 2m

We find that the time t and the space position (x, y, z) are treated very non symmetrically.
We need to search for relativistic equation for the particle of first order in t, x, y, and z,
where the equation should be symmetrical in space and time coordinates. Thus H is
required to be linear in the momentum operator.


i   H
t

The form of H is introduced by Dirac as

H  cα  p   mc 2

By squaring H, we should get the relation from the relativity (the Einstein relation),

1
H 2  (cα  p   mc 2 ) 2  c 2 [ pi p j { i , j }  mcpi { i ,  }   2 m 2c 2 ]
2
m c c p  E
2 4 2 2 2

which leads to the relations

4
 { i ,  j }  2 ij I , { i ,  }  0 , 2  I

where the curly bracket denotes an anti-commutator,

{ i , j }   i j   j i

Here we do not show how to derive the form of matrices  and  (4 x 4). The matrices
are Hermitian matrices (4x4). Thus the Hamiltonian is also Hermitian.

3. The matrices  and 

The matrices  and  can be expressed in terms of the Pauli spin matrices,

0 1 0  i 1 0 
 1    ,  2   ,  3    ,
1 0 i 0   0  1

1 0
I 2   .
 0 1 

Using the Kronecker product, the matrices  x ,  y ,  z , and  are given by

0 0 0 1
 
0 0 1 0   0 1 
1   1   1    
0 1 0 0    1 0 
 
1 0 0 0 


0 0 0  i
 
0 0 i 0   0 2 
2   2   2    
0  i 0 0    2 0 
 
i 0 0 0 
 

0 0 1 0
 
0 0 0  1  0  3 
3   3   3    
1 0 0 0    3 0 
 
0 1 0 0 


1 0 0 0
 
0 1 0 0   I2 0 
   z  I2    
0 0  1 0   0  I 2 
 
0 0 0  1


5
 0 0 0  i
 
0 0 i 0 
 1  i 1   ,
0 i 0 0
 
i 0 0 0 


0 0 0  1
 
0 0 1 0
 2  i 2  
0 1 0 0
 
1 0 0 0 


 0 0  i 0
 
0 0 0 i 
 3  i 3  
i 0 0 0
 
 0  i 0 0
 

I 0 
  4  
0  I 

 0 k 
 k  i 4 k   
 k 0 

 4 k    k 4

The Hamiltonian H is described by

 mc 2 0 cpz c( px  ip y ) 
 
 0 mc 2
c( px  ip y )  cpz 
H  
 cpz c( px  ip y )  mc 2 0 
 c( p  ip )  cp 0  mc 2 
 x y z 

H 2 can be evaluated as

1 0 0 0
 
20 1 0 0
H E 
2
  E 2I4
0 0 1 0
 
0 0 0 1 


where

6
 E 2  m 2c 4  c 2 p 2

((Mathematica))

7
Clear"Global`"; x   ; y   ;
0 1 0 
1 0  0
z   ; I2  IdentityMatrix2;
1 0
0 1

x  KroneckerProductx, x;
x  MatrixForm
0 0 0 1
0 0 1 0
0 1 0 0
1 0 0 0

y  KroneckerProductx, y;
y  MatrixForm
0 0 0 
0 0  0
0  0 0
 0 0 0

z  KroneckerProductx, z;
z  MatrixForm
0 0 1 0
0 0 0 1
1 0 0 0
0 1 0 0

8
   KroneckerProductz, I2;   MatrixForm
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1

x   .x  Simplify; x  MatrixForm

0 0 0 
0 0  0
0  0 0
 0 0 0

y   .y  Simplify; y  MatrixForm

0 0 0 1
0 0 1 0
0 1 0 0
1 0 0 0

z   .z  Simplify; z  MatrixForm

0 0  0
0 0 0 
 0 0 0
0  0 0
f1  c px x  c py y  c pz z   m 2 c2
c2 m 2 , 0, c pz, c px   c py,
0, c2 m 2 , c px   c py,  c pz,
c pz, c px   c py,  c2 m 2 , 0,
c px   c py,  c pz, 0,  c2 m 2 

g1  f1.f1  FullSimplify
c2 c2 m 4  px2  py2  pz2 , 0, 0, 0,
0, c2 c2 m 4  px2  py2  pz2 , 0, 0,
0, 0, c2 c2 m 4  px2  py2  pz2 , 0,
0, 0, 0, c2 c2 m 4  px2  py2  pz2 

4. Dirac equation
9
 We now have the Dirac equation given by


i   H
t

with

H  cα  p  mc 2

Then we get

 
i   (α  p  mc)  (i 4γ    mc)
 (ct ) i

or

  mc
[ k  4  ]  0 ,
xk  (ix0 ) 

with x1=x, x2 = y, x3 = z, and x4 = ict., since

α  i 4 γ ,  4 2  1,   4

Simply we have

 mc
(   )  0
x 

where  = 1, 2, 3, and 4.

5. Alternative method (Sakurai)

E2
 p 2  m 2c 2
c2

E ( op ) E ( op )
(  σ  p)(  σ  p)  m 2 c 2
c c

where

 
E ( op )  i  ic
t x0

10
with x0 = ct

This enables us to write a second order equation

 
(i  iσ  )(i  iσ  )  m 2c 2
x0 x0

for a free electron.  is now a two component wave function

1 
 ( R)  (i  iσ  )
mc x0

and

 (L )  


(i  iσ  ) ( R )  mc ( L )
x0

or


( i   i  ) ( R )   mc ( L ) (1a)
x0

and


(i  i  ) ( L )  mc ( R ) (1b)
x0

((Dirac equation))

Taking the sum and the difference of Eq.(1).

 (R)
 iσ  ( ( R )   ( L ) )  i (   ( L ) )   mc( ( R )   ( L ) )
x0

 ( R)
iσ  ( ( R )   ( L ) )  i (   ( L ) )   mc( ( R )   ( L ) )
x0

We define

11
  A   (R )   (L )

and

 B   (R)   (L)

Then


 iσ   B  i  A  mc A
x0


iσ   A  i  B  mc B
x0

  
  i  iσ   
 x0  A   mc A 
   B   
 B
 iσ   i
x0  


 A 
  is the 4x1 column matrix.
 B 

Here note that

  
  i  i σ   
 x0      I 0    0  i 

    (ix0 )  0  I   i 0 
 iσ   i
x0  


µ: gamma matrices (or Dirac matrices)

1 0 0 0
 
 I 0  0 1 0 0
 4     
0  I  0 0 1 0 
0 
 0 0  1

and

12
 0 0 0  i
 
 0  i 1   0 0  i 0 
 1   
 i 1 0   0 i 0 0 
 
i 0 0 0 
 

0 0 0  1
 
 0  i 2   0 0 1 0
 2   
 i 2 0   0 1 0 0
 
 1 0 0 0 


0 0  i 0
 
 0  i 3   0 0 0 i 
 3   
 i 3 0   i 0 0 0 
 
0  i 0 0
 

with

0 1 0  i 1 0 
 1   ,  2   ,  3   
1 0 i 0   0  1

Dirac spinor

 
 1 
 A   2 
 
      
 B   3 
 
 4 

Using these definitions, we have Dirac equation

 mc
[γ     4  ]  0
 (ix0 ) 

with

{  ,   }           2 

     (Hermutian)

For example

13
 { 1 ,  2 }   1 2   2 1
 0  i 1  0  i 2   0  i 2  0  i 1 
     
 i 1 0  i 2 0   i 2 0  i 1 0 
 [ , ] 0 
  1 2  
 0 [ 1 , 2 ] 
0

where

[ 1 , 2 ]   1 2   2 1  0

6. Lorentz invariance of the Dirac equation

6.1. Lorentz transformation

1 0 0 0  1 0 0 0 
   
0 1 0 0  0 1 0 0 
a aT  
0 0   i  0 0  i 
   
0 0 i   0 0  i  
 

det(a) = 1

a 1  aT

a 
1
  
 aT   a

x  '  a  x , a  x  '  a  a  x   x  x

x ' x '  a a x x   x x  x x

a  a    

a ( 1 )a(  2 )  a( 12 )

with

14
 1   2
12  .
1  1 2

6.2. Relativistic co-variance

 mc
  ' ( x' )   ' ( x' )  0
x  ' 

where

  is not primed.

x '  a x
x  a  x  '

 '  S .

where S is a 4 x 4 matrix. We note that

 x  
   a 
x  ' x  ' x x

Thus we get

 mc
  a  S ( x )  S ( x )  0
x 

Multiplying S 1 from the left, we have

 mc
S 1  a  S ( x )   ( x)  0
x 

or

 mc
S 1  Sa   ( x)   ( x)  0 .
x 

If

S 1  Sa    

15
or

S 1  Sa a   a   

or

S 1  S  a    ,

the problem of demonstrating the relativistic covariance of the Dirac equation is

now reduced to that of finding the form of S.

    4   4    4 2   

  mc  mc
(  )      0 , (   4 )     4  0
x   x  

or

 mc
(  )     0
x 

6.3 Relativistic co-variance

 mc
(  ' )    '  0 .
x  ' 

From this we get

 mc
 a   '    '  0
x 

We assume that

 '   T 1

 mc
 a   T 1    T 1  0
x 

Multiplying T from the left

 mc
 a   T 1  T    0
x 

16
    a  T 1  T  T 1  Ta 

or

T 1  T  a   

Here we can choose T = S. Then we have

    4 ,   4  

 '  S ,  '   S 1

since

 '   '   4    4 S  4   S 1

or

 4 S  4  S 1

or

S  4   4 S 1

In summary we have the following relations,

S 1  S  a   

S  4   4 S 1

 '  S

 '   S 1

6.4 Infinitesimal Lorentz transformation

a      

with the addition

   

17
since

a a   

  is pure-imaginary when one of ,  is equal to 4.

((Note))

a a  (     )(   )   

or

             

Neglecting    , we get

          

or

         

or

   

6.5 The expression of S

We now consider the relation

S 1  S  a   

We assume that

S  1  T  O(T 2 )

Then we get

(1  T )  (1  T )  (     ) 

or

18
 (   T  )(1  T )        

or

   [  , T ]        

or

[  , T ]     

The solution of this commutation relation is seen to be

1
T       
4

Then we have

1
S  1       
4
1
 1    [  ,   ]
8
i
 1     
4

where

1
   [  ,   ]
2i

((Note-1))
Noting that

{  ,   }  2 

we get

19
 1
[  , T ]  [  ,      ]
4
1
  [  ,     ]
4
1
  (            )
4
1
  [(     2  )     (     2  )]
4
1
 (          )
2
1
 (        )     
2

((Note-2))

1
       (            )
2
1
   (         )
2
1
   [  ,   ]
2

6.6 The relation S  4   4 S 1

We show that S satisfies the relation S  4   4 S 1 .

1
S  1      
4

1 *   1 *
S   1         1       
4 4

Then we get

1 * 1 *
S  4  (1        ) 4   4         4
4 4

1 1
 4 S 1   4 (1        )   4     4   
4 4

So we need to show that

  *     4     4   

20
For   i ,   k (i, k = 1, 2, 3),

 ik * k  i 4   ik  4 i k   ik ( k  i   i k ) 4
  ik 2 ik  4
0

For   i ,   4 (i = 1, 2, 3),

 i 4* 4 i 4   i 4 4 i 4   i 4 i 4 2   i 4 i 4 2
0

6.7 Rotation matrices

Infinitesimal rotation

 cos  sin  0 0  1  0 0   0  0 0 
     
  sin  cos  0 0     1 0 0     0 0 0 
a  1 
0 0 0 0  0 0 0 0 0 0 0 0
     
 0 0 0 0   0 0 0 0   0 0 0 0 
  

 12   ,  21  

1 1 1 1
S rot  1         1  ( 12 1 2   21 2 1 )  1  12 [ 1 ,  2 ]  1  i 12
4 4 4 2

where

 12   3  i 1 2

Finite rotation

S rot ( )  lim [ S rot ( )] N

N 

1 
 lim (1  i  3 ) N
N  2 N
1
 exp(i  3 )
2
 
 1cos  i sin 3
2 2

21
where

  N

Note that

 i2 
e 0 0 0 
 
i 
S rot ( )   0 0 0 
2
e
 i

 0 0 e2 0 
i
  
 0 0 0 e 2

S rot ( ) 1  1S rot ( )   1 cos    2 sin 

S rot ( ) 1  2 S rot ( )   1 sin    2 cos 

Clear"Global`";
((Mathematica))

exp_  : exp . Complexre_, im_  Complexre,  im ;

x   ; y   ; z   ; I2  IdentityMatrix2;
0 1 0  1 0
 0 0 1
I4  IdentityMatrix4;
1 0

x  KroneckerProductx, x ; y  KroneckerProductx, y;

z  KroneckerProductx, z ;
  KroneckerProductz, I2;

1   .x  Simplify; 2   .y  Simplify;

3   .z  Simplify; 4  ; 1   2.3; 2   2.3;

3   1.2; S1  MatrixExp  3;



S1  MatrixForm
2


 2 0 0 0
 
0  2 0 0

0 0  2 0
 
0 0 0  2

22
 MatrixExp  3   Cos  I4  I Sin  3  Simplify
  

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

2 2 2

InverseS1.1.S1  1 Cos  2 Sin  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

InverseS1.2.S1  1 Sin  2 Cos  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

6.8 Lorentz transformation

The Lorentz transformation is expressed by

 cosh  0 0 i sinh  
 
 0 0 0 0 
a
0 0 0 0 
 
  i sinh  0 0 cosh  


where

  tanh  , cosh    , sinh   

The pure Lorentz transformation is nothing more than a rotation in the 1-4
plane by imaginary angle;

   
S rot ( )  1cos  i sin (i 1 2 )  1cos  sin  1 2
2 2 2 2

  i , 2 4

Then we have

  
 cosh 0 0  sinh 
 2 2
   
   0 cosh  sinh 0
S Lor  1 cosh  i 1 4 sinh  2 2 
2 2  0  sinh

cosh

0

 
 2 2
  sinh   
0 0 cosh 
 2 2 

23
where

   
cos  cosh , sin  i sinh
2 2 2 2

The Hermitian conjugate of S Lor is obtained by

    
S Lor  1cosh  i 4  1 sinh
2 2
 
 1cosh  i 4 1 sinh
2 2
 
 1cosh  i 1 4 sinh
2 2
 S Lor

and

  
 cosh 0 0 sinh 
 2 2
  
   0 cosh sinh 0 
1
S Lor  cosh  i 1 4 sinh  2 2 
2 2  0 sinh

cosh

0 


 2 2
 sinh   
0 0 cosh
 2 2

Thus S Lor is not unitary. It is important to note that for both pure rotation and
pure Lorentz transformation, we have

S 1   4 S  4 , S    4 S  1  4

((Note))

 '  '    S Lor  S Lor 

((Mathematica))

24
Clear"Global`";
exp_  : exp . Complexre_, im_  Complexre,  im ;
x   ; y   ; z   ; I2  IdentityMatrix2;
0 1 0  1 0
 0 0 1
I4  IdentityMatrix4;
1 0

x  KroneckerProductx, x ; y  KroneckerProductx, y;

z  KroneckerProductx, z ;
  KroneckerProductz, I2;

1   .x  Simplify; 2   .y  Simplify;

3   .z  Simplify; 4  ; 1   2.3;
S1  I4 Cosh    1 .4 Sinh ;
 
2 2

S1  MatrixForm
Cosh    Sinh 2 
Cosh    Sinh 2 
0 0
2

 Sinh 2  Cosh  
0 0
2

 Sinh 2  Cosh  
0 0
2
0 0
2

S1H  TransposeS1 ; S1H  S1  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

25
 InverseS1  Simplify;   MatrixForm
Cosh   Sinh  
Cosh  Sinh 
0 0
2 2
 

Sinh   Cosh  
0 0
2 2

Sinh   Cosh  
0 0
2 2
0 0
2 2

InverseS1  4 .S1H.4  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

InverseS1  I4 Cosh    1 .4 Sinh   Simplify

 

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

2 2

S1H  4 .InverseS1.4  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

Eigensystem S1  FullSimplify

2 , 2 ,  2 ,  2 ,
1, 0, 0, 1, 0, 1, 1, 0,  1, 0, 0, 1, 0,  1, 1, 0

7. Space Inversion (parity operator)

7.1 Definition
When r '   r and t '  t ,

1 0 0 0
 
 0 1 0 0
a
0 0 1 0
 
0 0 0 1 


det(a) = -1: improper Lorentz transformation.

Under the parity operation,

E' E , B'   B .

and

26
 1 A
E  A0  , B   A
c t

Then we get

Ak ' (r ' t ' )   Ak (r , t ) , A4 ' (r ' t ' )  A4 (r , t ) .

Dirac equation is given by

 ie mc
(  A ' )  '  '  0
x ' c 

or equivalently

 ie  ie mc
[ (  Ak ) k  (  A4 ) 4 ] '  '  0 .
xk c x4 c 

We try as before

 '  S p

 ie  ie mc
[ (  Ak ) k  (  A4 ) 4 ]S p  S p  0
xk c x4 c 

1
Multiplying S p from the left,

1  ie  ie mc
S p [ (  Ak ) k  (  A4 ) 4 ]S p    0.
xk c x4 c 

When this equation is compared with the original Dirac equation,

 ie mc
(  A )     0,
x c 

we have

1 1
S p  k S p   k , S p  4 S p   4

We assume that

S p   4

27
where  is some multiplicative constant. Here we use

1 0 0 0
 
0 1 0 0
Sp   4    
0 0 1 0 
 
0 0 0  1

for an electron wave function.

7.2 Definition of the parity operator P

We define the parity operator as

P  

where  operates on the Dirac space and  operates on the coordinate space. Note that

P 2     2 2  1

The commutation relation between H (for the free particle) and P is given by

[ P, H ]  [  , c k pk   mc 2 ]
 [  , c k pk ]
 c k pk  c k pk 
 c kpk  c k pk 
 c(  k   k  )pk
 c{ , k }pk
0

where pk  pk  0 . Then the eigenket of P is the same as that of H.

What happened to the wave function?

 '    ;

 ' (r )  r  '
 r  
 r
 r
  (  r )

28
For the change in r   r , we have

 ' (  r )   ( r )   4 ( r )

When the time t is taken into account, we have

 ' (r , t )   4 (r , t ) .

7.3 Two component wave functions  A and  B under the parity operation

Since

 
   A 
 B 

we get

 ' (r , t )   I 0  A (r , t ) 

 ' (r , t )   A    4 (r , t )    
 B ' (r , t )   0  I  B (r , t ) 

Then we have

 A ' (r , t )   A (r , t ) (the same as the non-relativistic case)

 B ' (r , t )   B (r , t )

The upper and lower components of the wave functions have different behaviors under a
parity transformation.
This is expected from the following discussion of Dirac equation.

H  E

  e
   A  , H  cα  ( p  A)  mc 2
 B  c

Eigenvalue problem:

 e 
 mc 2 cσ  ( p  A)  
H   c  A   E  A 
   
 cσ  ( p  e A)  mc 2  B   B
 c 

29
Or

e
cσ  ( p  A) A  mc 2 B  E B
c

Then we have

1 e
 B ( r, t )  cσ  ( p  A) A ( r , t )
E  mc 2
c

Under the parity,

p  p, A  A,

 A (r , t )   A ' ( r , t ) ,

then we get

1 e
 B ' (r, t )   cσ  (p  A) A ' (r, t )
E  mc 2
c
1 e
 cσ  (p  A) A (r, t )
E  mc 2
c
  B (r, t )

where

 A ' (r, t )   A (r, t )

7.4 Even or odd parity in  A and  B

Suppose that  A and  B has different parities (even or odd).

  A   A ,  B  B

or

r A  rA  rA

and

r  B   r B    r B

30
Then we get

 I2 0  A (  r , t )    A (  r , t )   ( r , t ) 
 (  r , t )           A    ( r , t )
0  I 2  B (  r , t )     B (  r , t )   B ( r , t ) 

or, formally

 4 (  r , t )   ( r , t ) .

since

 A (  r , t )   A ( r , t ) ,  B (  r , t )   B ( r , t ) .

We assume that  A and  B are the eigenstates of the orbital angular momentum.

 A (  r , t )  ( 1) l  A ( r , t )   A ( r , t )
A

 B (  r , t )  ( 1) l  B ( r , t )   B ( r , t )
B

where lA and lB are the orbital angular momenta of the two-component wave function
 A ( r , t ) and  B ( r , t ) , respectively. Thus we have

( 1) l A  ( 1) l B 1

This implies that if  A ( r , t ) is a two-component wave function with an even (odd) orbital
angular momentum, then  B ( r , t ) is a two-component wave function with an odd (even)
orbital angular momentum.

((Example))
We consider the case of a central force.

A  0, A0   , eA0  V ( r )

 
   A  , H  cα  p  mc 2  V ( r )
 B 

 mc 2  V ( r ) cσ  p  A   A 
H   

 
  E  
    B 
2
 c σ p mc V ( r )   B 

31
Then we get

c
 B ( r, t )  (σ  p) A ( r , t )
E  V ( r )  mc 2

Let us suppose that  A ( r , t ) is an 2S1/2 state wave function with spin up (l = 0, s = 1/2)

 1 
iEt
 A ( r , t )  R( r ) e 

 0

Then

    
 i 
 ic x3 x1 x2  R ( r )   
iEt
 B ( r, t )    e
E  V ( r )  mc 
2     0 
 x  i 
 1 x2 x3 

Note that

 r dR x3 dR
R( r )   ,
x3 x3 dr r dr

 x dR  x dR
R( r )  1 , R( r )  2
x1 r dr x 2 r dr

Then we get

 i c 1 dR  x3 x1  ix2  1   iEt
 B (r , t )    e
E  V (r )  mc r dr  x1  ix2
2
 x3  0 
 i c 1 dR  x3   
iEt
   e
E  V (r )  mc 2 r dr  x1  ix2 
 i c 1 dR  1   0   iEt
 [ x    ( x  ix2 
) 1 ]e
E  V (r )  mc 2 r dr  0 
3 1
 

Here we note that

1 3 1 3 x1  ix2
Y11  sin ei 
2 2 2 2 r

32
 1 3 3 x3
Y10  cos 
2  4 r

1 3 1 3 x1  ix2
Y11  sin e i 
2 2 2 2 r

where

x1  sin  cos  , x2  sin  sin  , x3  cos

Thus we have

 ic dR 4 0  1  8 1  0   
iEt
 B (r , t )  [ Y    Y   ]e
3  0  3  1 
1 1
E  V (r )  mc 2 dr

The first term of the parenthesis is l = 1 and s = 1/2

j = 1/2 (2P1/2)

The second term of the parenthesis is l = 1 and s = 1/2

j = 3/2 (2P3/2)
j = 1/2 (2P1/2)

____________________________________________________________________
((Note)) Parity operator  in non-relativistic quantum mechanics

 '  ˆ  with ˆ ˆ  1̂ and ˆ   ˆ

 ' xˆ  '    xˆ  , ˆ  xˆˆ   xˆ

ˆ x   x

x  '  x ˆ  , or  ' ( x)   ( x)

Even parity: x  '  x ˆ   x  ,  ' ( x)   ( x)   ( x)

Odd parity x ˆ    x  ,  ' ( x)   ( x)   ( x)

______________________________________________________________________
8. Eigenvalue problem (degenerate case)

33
We solve the eigenvalue problem using the Mathematica.

 mc 2  E 0 cpz c( px  ip y ) 
 
 0 mc  E
2
c( px  ip y )  cpz 
H  EI 4   
 cpz c( px  ip y )  mc 2  E 0 
 c( p  ip )  cpz 0  mc  E 
2
 x y

From the condition that det( H  I 4 )=0, we get

E   R   m 2c 4  c 2 p 2

where

R  m 2c 4  c 2 p 2 (>0)

Thus we see that there are four eigenvalues which are degenerate in pairs, i.e.

E = +R, +R, -R, and -R

For simplicity we assume that

px  p y  0

For E = +R

 mc 2  R 0 cpz 0  u1   0 
    
 0 mc 2  R 0  cpz  u2   0 

0  2
 0  u    0 
 cp z mc R  3   
 0
  cpz 0  mc  R  u4   0 
2

or

(mc 2  R)u1  cpz u3  0

(mc 2  R)u2  cpz u4  0
cpz u1  ( R  mc 2 )u3  0
 cpz u2  (mc 2  R)u4  0

It is clear from the above equations that at the zero momentum limit ( p z  0 ) the first
two equations do not give us any information on the unknowns. Thus we need to solve

34
the second two equations. The two independent solutions, corresponding to the
eigenvalue +R,

cpz
u1 = 1, u2 = 0, u3  , u4 = 0.
R  mc 2

cp z
u1 = 0, u2 = 1, u3  0 , u 4   .
R  mc 2

For E = -R

 mc 2  R 0 cpz 0  u1   0 
    
 0 mc  R
2
0  cpz  u2   0 
  u    0 
 cp z 0  mc 2
 R 0  3   
 0
  cpz 0  mc  R  u4   0 
2

or

(mc 2  R)u1  cpz u3  0

(mc 2  R)u2  cpz u4  0
cpz u1  ( R  mc 2 )u3  0
 cpz u2  (mc 2  R)u4  0

It is clear from the above equations that at the zero momentum limit ( p z  0 ) the second
two equations do not give us any information on the unknowns. Thus we need to solve
the first two equations. The two independent solutions, corresponding to the eigenvalue -
R,

cpz
u1   , u2 = 0, u3 = 1, , u4 = 0.
R  mc 2

cpz
u1 = 0, u2  , u3  0 , u4  1 .
R  mc 2

((Summary))

For E  R (positive energy)

35
  1   0 
   
R  mc 2  0  R  mc 2  1 
 cpz   0 ,
2R  2R 
R  mc 2  cpz 
 0   
   R  mc 2 

For E  R (negative energy)

  cpz   0 
 2   cpz 
2  R  mc 
R  mc 0 R  mc 2  2 
   R  mc 
2R  1  2R  0 
   1 
 0   

If pz = 0, we have

1  0
   
 0 1
 0 ,  0 ,
   
 0  0
   

The non-relativistic spin sttes. These are degenerate and have energy eigenvalue +R.

 0  0
   
 0  0
1 ,  0
   
 0 1
   

The nonrelativistic spin states. These are degenerate and have energy eigenvlaue -R.

9. The use of the Mathematica to derive the eigenkets of H

The eigenvalue problem can be solved Using the Eigensystem[H] of the
Mathematica,.

((Mathematica))

36
 Clear"Global`"; exp_  : exp . Complexre_, im_  Complexre,  im ;
x   ; y   ; z   ;
0 1 0  1 0
 0 0 1
I2  IdentityMatrix2; I4  IdentityMatrix4;
1 0

x  KroneckerProductx, x; y  KroneckerProductx, y;

z  KroneckerProductx, z;   KroneckerProductz, I2;
H  m c2   c x px  y py  z pz;

rule1   c2 m 2   px2  py2  pz2   ;

R
c

rule2    px2  py2  pz2    c2 m 2 ;

R2
2
c

rule3  R  c c2 m 2  px2  py2  pz2 ;

Eigenvalue problem of the Hamiltonian of the Dirac free particle

eq1  Eigensystem H . rule1  FullSimplify

 R,  R, R, R,
c px   py c px   py
 , 0, 1,  2 , 1, 0,
c pz c pz
, 2 ,

c px   py c px   py

2
c mR c mR c mR c2 m  R
 , 0, 1,  2 , 1, 0
c pz c pz
, 2 ,
 c2 m  R c mR c m  R  c2 m  R
Orthogonality

eq12, 1 .eq12, 2 . rule3  Simplify

0

eq12, 1 .eq12, 3 . rule3  Simplify

0

eq12, 1 .eq12, 4 . rule3  Simplify

0

eq12, 2 .eq12, 3 . rule3  Simplify

0

eq12, 2 .eq12, 4 . rule3  Simplify

0

37
 eq12, 3.eq12, 4 . rule3  Simplify
0

Normalization constant

A1  eq12, 1.eq12, 1  Simplify; A11  A1 . rule2  Simplify

2R
2
c mR

A2  eq12, 2.eq12, 2  Simplify; A21  A2 . rule2  Simplify

2R
2
c mR

A3  eq12, 3.eq12, 3  Simplify; A31  A3 . rule2  Simplify

2R
 2
c mR

A4  eq12, 4.eq12, 4  Simplify; A41  A4 . rule2  Simplify

2R

c2 m  R

_______________________________________________________________________
The results are as follows.

For E  R (positive energy)

 cpz   c( px  ip y ) 
 2   
 R  mc   R  mc 2 
R  mc 2  c( px  ip y )  R  mc 2   cpz 
,
2 R  R  mc  2 R  R  mc 
2 2
 1   1 
   
 0   0 

For E  R (negative energy)

  cpz    c( px  ip y ) 
   
 R  mc 2
  R  mc 2 
R  mc 2   c( px  ip y )  R  mc 2  cpz 
2 R  R  mc    
2 2
2R R mc
 1   0 
   
 0   1 

10. Conserved current

38
  
 1 
 2 
   
 3 
 
 4 

    *1  2 *  3 *  4 * 


   *1  2*   3*   4* 
   4
1 0 00
 

  *1  2
*
 3* 
* 0
4 
0
1 0 0
0 1 0 
 
0 0 0  1


In order to obtain the wave equation for  , we start from the Dirac equation

 mc
[γ     4  ]  0
 (ix0 ) 

or

  mc
( k 4  )  0 (k = 1, 2, 3).
xk x4 

or

 mc
(   )  0 ( = 1, 2, 3, 4).
x 

where x1 = x, x2 = y, x3 = z, x4 = ix0 = ict

We take the hermitian conjugate of the Dirac equation,

   mc
  k  *   4     0 (1)
xk x4 

Multiplying Eq.(1) by 4 from the right, we get

   mc
  k  4  *   4 2    4  0 .
xk x4 

39
Since  4 k    k 4 , we have

    2 mc 
   4 k   4   4  0
xk x4 

or

  mc
   4     4  0
x 

or

 mc
(  )     0
x 

Now we have

 mc
(   )  0 (2)
x 

and

 mc
(  )     0 (3)
x 

 x Eq.(2) leads to

 mc
 (   )  0 (2')
x 

Eq.(3) x  leads to

 mc
(  )    )  0 (3')
x 

The subtraction of Eq.(3') from Eq.(2') yields

 
   (  )   0 .
x x

or
40
 
(   )  0
x

Thus we see that

S   ic     ( c   k , ic   )


S  0
x

or


 S  ic   0
 (ict )

or

 
  S  0
t

The flux density S is defined by


Sk  ic k  c   k

The probability density is defined by

   

((Note))

ic k  ic  4 k  c  k  Sk

ic 4  ic  4 4  ic 

where

 k  i 4 k

11. Simple solutions: nonrelativistic approximation

In the presence of electromagnetic fields,

41
   e
p   p  A
i x c

or

 i e
 ( p  A )
x  c

Dirac equation

 mc i e mc
(   )  0  [  ( p  A )  ]  0
x   c 

or

e
[  (p  A )  imc]  0
c

or

 0  iσ  e I 0  e 
 ( p  A)   ( p4  A4 )  imc   0
 iσ 0  c 0  I  c 

or

 e   e 
 0  iσ  ( p  A)   p4  A4 0 
 c    c   imc
 iσ  ( p  A)
e
0   0  ( p4  A4 ) 
e
 c   c 

   
Noting that A4  iA0 and p 4  
i x4 c t

 e     ie 
 0  i  ( p  A)      A0 0    
 c     c t c
A
 A   imc A 
    ie
 i  ( p  e A) 0  B   0  A0 )  B   B 
 c   c t c 

or
42
 e i 
 iσ  ( p  A) B  (i  eA0 ) A  imc A
c c t

e i 
iσ  ( p  A) A  (i  eA0 ) B  imc B
c c t

or

e 1 
σ  ( p  A) B  (i  eA0  mc 2 ) A
c c t

e 1 
 σ  ( p  A) A  (i  eA0  mc 2 ) B
c c t

Assuming that
 iEt / 

   0 e

  iE
i  A  E A or A  A
t t 

  iE
i  B  E B or B   B
t t 

Then we have

e 1
σ  (p  A) B  ( E  eA0  mc 2 ) A
c c

e 1
 σ  (p  A) A   ( E  eA0  mc 2 ) B
c c

((Note)) Aµ is time-independent.

c e
B  2 [  (p  A)] A
E  eA0  mc c

Substitution of this eq. into the first equation

e c e 1
[ (p  A)] 2 [  (p  A)] A  (E  eA0  mc 2 ) A
c E  eA0  mc c c

43
We now assume that E ≈ mc2 and eA0  mc2 .

Defining the energy measured from mc2,

(NR) 2
E  E  mc

c 1  2mc 2 
  
E  eA0  mc 2
2m  2mc  E
2 ( NR )
 eA0 
1 1

2m E  eA0
( NR )
1
2mc 2
1 E ( NR )  eA0
 [1  ( )  ...]
2m 2mc 2
1

2m

1 e e
[σ  (p  A)][σ  (p  A)] A  ( E ( NR )  eA0 ) A
2m c c

which becomes

1 e e
[ ( p  A) 2  σ  B  eA0 ] A  E ( NR ) A
2m c 2mc

((Note))

e e e e e
[σ  ( p  A)][σ  ( p  A)]  ( p  A) 2  iσ  [( p  A)  ( p  A)]
c c c c c
e e
 ( p  A) 2  σ  B
c c

since

e e e
(p A)  ( p  A)   ( p  A  A  P )
c c c
ie
  A
c
ie
 B
c

((Comment))

44
To zeroth order in (v/c)2, A is nothing more than the Schrödinger-Pauli two component
2
wave function in nonrelativistic quantum mechanics, multiplied by exp(imc
 t /  ).

B is “smaller” than A by a factor of roughly p  eA / c / 2mc  v /( 2c) if E ≈ mc2 and

eA0  mc2 .

For this reason with mc2, A and B are known as the large and small components of the
Dirac wave function .

Since

e 1
 σ  ( p  A) A   ( E  eA0  mc 2 ) B  2mc B
c c

1 e
B  σ  ( p  A) A
2mc c

12. Approximate Hamiltonian for an electrostatic problem

e c2 e
[σ  ( p  A)] [σ  ( p  A)] A  ( E  eA0  mc 2 ) A
c E  eA0  mc 2
c

For simplicity A = 0.

1 E ( NR )  eA0
(σ  p)[1  ](σ  p) A  ( E ( NR )  eA0 ) A
2m 2mc 2

or

H(ANR) A  E ( NR) A

with

1 E ( NR )  eA0
H A( NR )  (σ  p)[1  ( )](σ  p)  eA0 (1)
2m 2mc 2

It might appear that Eq.(1) is the time-independent Schrödinger equation.

However, there are three difficulities with this interpretation.

(1) Normalization

45
  (

A  A   B  B ) d 3 x  1

(2) H(ANR) contains a non-Hermitian term ( iE

  p)

(3) Since H(ANR) contains E (NR) itself, Eq.(1) is not an eigenvalue equation.

Since

1
B  (σ  p) A
2mc

1
 B    A (σ  p)
2mc

Normalization:

1
 (
 
A A  2 2  B (  p ) A ) d x  1
3

4m c

to order (v / c) 2 .

This suggests that we should work wit a new-two component wave function  defined by

   A

or

 A   1 

where

p2
 1
8m 2 c2

With this choice

p2 2
  d x   ( A (1 

 3
2 2
)  Ad 3 x
8m c
p2
  ( A (1 

) A d 3 x
4m 2 c 2
1

46
 H(ANR) A  E ( NR) A

H(ANR) 1   E ( NR)  1 

or

 1 H(ANR) 1   E ( NR) 2 

p2 p2
 1H A( NR )  1  (1  ) H ( NR )
(1  )
8m 2c 2 8m 2c 2
A

p2 p2
 (H A 
( NR )
H A )(1  2 2 )
( NR )

8m 2c 2 8m c
2
p
 H A( NR )  { 2 2 , H A( NR ) }
8m c
p2 1 2
 H A( NR )  { 2 2 , p  eA0 }
8m c 2m

where

1 2 1 E ( NR )  eA0
H A( NR )  p  eA0  (σ  p)( )](σ  p)
2m 2m 2mc 2

p2
E ( NR)  2   E ( NR) (1  )
4m2 c 2

Thus we have

1 2 p2 1 2
[ p  eA0  { 2 2 , p  eA0 }
2m 8m c 2m
1 E ( NR )  eA0 p2
 (σ  p)( )](σ  p )]  E ( NR )
(1  )
2m 2mc 2 4m 2 c 2
Note

p2 1 2 p2 p4 1
{ 2 2
, p  eA0 }  E ( NR )
2 2
  2 2
 2 2 { p 2 , E ( NR )  eA0 }
8m c 2m 4m c 8m c 8m c

Then we have

1 2 p4 1
[ p  eA0  2 2  2 2 [{ p 2 , E ( NR )  eA0 }  2(σ  p)( E ( NR )  eA0 )](σ  p)]  E ( NR ) 
2m 8m c 8m c

Here we use the formula

47
 {A2 , B}  2 ABA  [ A,[ A, B]]

When
( NR )
Aσ p, BE  eA0

A2  (σ  p)(σ  p)  p 2

Thus we have

{ p 2 , E ( NR )  eA0 }  2(σ  p)( E ( NR )  eA0 )(σ  p)

 [(σ  p), [(σ  p), E ( NR )  eA0 ]]

or

{ p 2 , E ( NR )  eA0 }  2(σ  p)( E ( NR )  eA0 )(σ  p)  [(σ  p), [(σ  p), E ( NR )  eA0 ]]

Here

[(σ  p), E ( NR )  eA0 ]  [(σ  p),eA0 ]

 e{(σ  p) A0  A0 (σ  p)}
 eσ  [ p, A0 ]

 eσ  A0
i
 ieσ  E

Note that A0  A0

[(σ  p), [(σ  p), E ( NR )  eA0 ]]  [(σ  p),ie (σ  E )]

 ie[(σ  p), (σ  E )]
 ie[(  p)(  E )  (  E )(  p)]

Note that

(σ  p)(σ  E )  p  E  iσ  ( p  E )

and

(σ  E )(σ  p)  E  p  iσ  ( E  p)

Then we have
48
 
(σ  p)(σ  E )  (σ  E )(σ  p)    E  2iσ  ( E  p)
i

Finally we obtain

1 2 p4 1 
{ p  eA0  2 2  2 2
(ie )[   E  2iσ  ( E  p)]}  E ( NR ) 
2m 8m c 8m c i

or

1 2 p4 e e 2
[ p  eA0  2 2  σ  ( E  p )    E ]  E ( NR ) 
2m 8m c 4 m 2 c 2 8m 2 c 2

((Physical meaning))

Third term: relativistic correction

p2
m 2c 4  p 2c 2  mc 2  mc 2 1  2 2
 mc 2
mc
p2 1 p4
 mc 2 [1    ...]  mc 2
2m 2 c 2 8 m 4 c 4
p2 1 p4
   ...
2 m 8 m 3c 2

The fourth term (Thomas correction)

e
Thomas term =  σ  ( E  p)
4m 2 c 2

For a central potential

eA0 = V(r)

1 dV
E  A0   r
r dr

1 dV 1 dV
E p   ( r  p)   L
r dr r dr

where L is an orbital angular momentum. Then the Thomas term is rewritten as

49
 e e 1 dV e 1 dV
Thomas term =  2 2
σ  ( E  p)   2 2 (  )σ  L  SL
4m c 4m c r dr 2m 2c 2 r dr

(Spin-orbit interaction)

The spin angular momentum is defined by


S σ
2

which is an automatic consequence of the Dirac theory.

The last term is called the Darwin term.

For a hydrogen atom,

  E   ( 3) (r ) .

It gives rise to an energy shift

e 2  2 ( 3) ( Schrodinger ) 2 3 e 2 2 2
 8m 2 c 2  (r ) (r ) d x 
8m c2 2
 (r )( Schrodinger ) r  0

which is non-vanishing only for the s state.

13. Free particle at rest

Each component of the four-component wave function satisfies the Klein-Gordon
equation if the particle is free.

 mc
(   )  0 (1)
x 

 0 i k 
 k  
 
i k 0  

I 0 
 4   
0 I 


Multiplying Eq.(1) from the left by  
x

50
   mc 
 (  )    0
x x  x

  mc 
      0
x x  x

1   mc
(        )   ( ) 2  0
2 x x 

Since {  ,   }  2  , , we have

  mc
  ,   ( ) 2  0 (2)
x x 

Note that Eq.(2) is to be understood as four separate uncoupled equations for each
component of . Because of Eq.(2), the Dirac equation admits a free particle solution of
the type

i
  u(p) exp[ (p  r  Et)]
 

with

E   c 2 p2  m2 c 4

u(p) is a four-componenet spinor independent of r and t.

((Note)) The following relations are always valid.

 
i   E ,   p
t i

For a particle at rest (p = 0)

 mc
(   )  0
x 

or

51
     mc
( 1 2  3 4  )  0
x1 x2 x3 x4 

Since


  p  0 , E   mc
2

 mc
[ 4  ]  0
 (ict ) 

or

1 E mc
[ 4  ]  0
ic i 

or

E mc
 4  
c 

For E = mc2,

 u (p  0) 
u   A 
 u B (p  0) 

mc 2  I 0  u A ( p  0)  mc  u A ( p  0) 
    
c  0  I  u B ( p  0)    u B ( p  0) 

or

 u A ( p  0)   u A ( p  0) 
    
  u B ( p  0)   u B ( p  0) 

or

52
 u B ( p  0)  0

For E = -mc2,

 u ( p  0) 
u   A 
 u B ( p  0) 

mc 2  I 0  u A ( p  0)  mc  u A ( p  0) 
     
c  0  I  u B ( p  0)    u B ( p  0) 

or

uA (p  0) uA ( p  0)


 u (p  0) 
  
 

 B  uB (p  0)

uA (p  0)  0

So there are four independent solutions

Positive energy solution

1 0
   
0 mc2 t 1 mc2 t
 exp(i ),  exp(i ),
0  0 
0 0
 
  

spin up spin down

Negative energy solution

0 0
   
0 mc 2 t 0 mc 2 t
 exp(i ),  exp(i ),
1  0 
0 1
 
  


The existence of negative-energy solutions is intimately related to the fact that the Dirac
theory can accomodate a positron.

53
((Note))
Nonrelativistic limit E = mc2, the upper two component spinor A coincides with the
2
Schrödinger wave function apart from the factor eimc t /
.

Let us define

 1 2   2 1  3 0 
3   
 0  
2i  3 

((Note))

 0 i 1  0 i 2   1 2 0 

 1 2  
     
i  1 0  
i 2 0  
  0  1 2 


Similarly

  2 1 0 
 2 1   
 0  2 1 

  1 2   2 1 0 
 1 2   2 1   
 0  1 2   2 1 
  0 
 2 1 2 
 0  1 2 
 0
 2i 3 
 0 3 

1 1 1

     
0  3 0 0 0
 3    
     
0  0  3 
0 0
0 0 0
     

The eigenstate of  3 is interpreted as the spin component in the positive z-direction in

units of /2.

____________________________________________________________________
14. Plane wave solutions (p ≠ 0).

54
  A  uA (p) i
  
   
u ( p)
exp[ (p  r  Et)]
  B   B  

or

1 
(i  mc 2 ) A  (σ  p) B ,
c t

1 
(i  mc 2 ) B  (σ  p) A
c t

or

1 1
(σ  p)u B  ( E  mc 2 )u A , (σ  p)u A  ( E  mc 2 )u B
c c

or

c
u A ( p)  (σ  p)u B ( p)
E  mc 2

c
u B ( p)  (σ  p)u A ( p)
E  mc 2

Fro simplicity we use

R p 2c 2  m 2c 4

(i) For E  R  0 (positive energy state)

(1) 1 ( 2) 0

u A ( p)    , u A ( p)   
 0 1

55
 (1) c  p3 p1  ip2  1 
u B ( p)    
2 
R  mc  p1  ip2  p3  0 
 cp3 
 2 
  R  mc 
 c( p1  ip2 ) 
 R  mc 2 

( 2) c  p3 p1  ip2  0 
u B ( p)    
R  mc 2  p1  ip2  p3  1 
 c( p1  ip2 ) 
 2 
  R  mc 
  cp3 
 R  mc 2 

Then we have

 1   0 
   
2 
0  2 
1 
R  mc  cp3 , R  mc  c( p1  ip2 ) 
u (1) ( p)  u ( 2 ) ( p) 
2 R  R  mc 2  2 R  R  mc 2 
 c( p1  ip2 )   cp3 
   
 R  mc 2   R  mc 2 

We take into account of the normalization factor.

(ii) For E   R  0

(1) 1 ( 2) 0

u B (p)    , u B (p)   
 0 1

(1)  c  p3 p1  ip2  1 
u A ( p)    
R  mc 2  p1  ip2  p3  0 
 cp3 
  2 
  R  mc 
  c( p1  ip2 ) 
 R  mc 2 

56
  c( p1  ip2 ) 
 c  p3 p1  ip2  0    R  mc 2 
uA
( 2)
( p)       
R  mc 2  p1  ip2  p3  1   cp3 

 R  mc  2

Then we have

 cp3   c( p1  ip2 ) 
  2   
 R  mc   R  mc 2 
R  mc 2  c( p1  ip2 )  R  mc 2  cp3 
u ( 3) ( p)   , u ( 4 ) ( p) 
2 R  R  mc    
2 2
2R R mc
 1   0 
   
 0   1 

________________________________________________________________________
15. Formulation
Since

i
u ( r ) (p) exp[ {p  r  Et )]


satisfies the free-field Dirac equation

 mc
(   )  0 ,
x 

 i i i
{u ( r ) ( p) exp[ { p  r  Et )]}  u ( r ) ( p) exp[ { p  r  Et )] pk
xk   

 i i 1 i
{u ( r ) ( p) exp[ { p  r  Et )]}  u ( r ) ( p) exp[ { p  r  Et )]( )( E )
x4   ic 
i E
 u ( r ) ( p) exp[ { p  r  Et )]( )
 c

Since

E
p  (p, i )
c

57
we get

i E mc ( r )
( pk  k   4  )u ( p)  0
 c 

or

E
(ipk  k   4  mc)u ( r ) ( p)  0
c

or

(i  p  mc)u ( r ) ( p)  0

regardless of whether E>0 or E<0.

In summary we have

 1 
 
2 
0 
R  mc  cp3 ,
u (1) ( p)  (E>0)
2R  R  mc 2 

 c( p1  ip2 ) 
 
 R  mc 2 

 0 
 
2 
1 
R  mc  c( p1  ip2 ) 
u ( 2 ) ( p)  , (E>0)
2 R  R  mc 2 
 cp3 
 
 R  mc 2 

 cp3 
  2 
 R  mc 
R  mc 2  c( p1  ip2 ) 
u ( 3 ) ( p)   , (E<0)
2 R  R  mc 
2

 1 
 
 0 

58
  c( p1  ip 2 ) 
 
 R  mc 2 
R  mc 2  cp3 ,
u ( 4 ) ( p)   (E<0)
2R R  mc 2 
 0 
 
 1 

Suppose that p = 0. R  mc 2 .

1  0
   
 0 1
u (1) ( p  0)    , u ( 2 ) ( p  0)    .
0 0
   
 0  0
   

 0  0
   
 0  ( 4)  0
u ( p  0)    , u ( p  0)   
( 3)

1 0
   
 0 1
   

The first two solutions look like the spin state of the non-relativistic theory. They are
degenerate and have enery eigenvalue E  R . In the same limit, the last two solutions
also look like the non-relativistic spin states, but they belong to the energy eigenvalue
E  R
__________________________________________________________________
13. Dirac's hole theory
13.1. Overview on Dirac's hole theory
Dirac made the astounding suggestion that all the negative-energy states should be
already occupied. This ocean of occupied negative-energy states is now referred to as the
‘Dirac sea’. Thus, according to Dirac, the negative energy states are already full up; by
the Pauli principle, there is now no room for an electron to fall into such a state. But, as
Dirac further reasoned, occasionally there might be a few negative-energy states that are
unoccupied. Such a ‘hole’ in the Dirac sea of negative-energy states would appear just
like a positive-energy particle (and hence a positive-mass particle), whose electric charge
would be the opposite of the charge on the electron. Such an empty negative-energy state
could now be occupied by an ordinary electron; so the electron might ‘fall into’ that state
with the emission of energy (normally in the form of electromagnetic radiation, i.e.
photons). This would result in the ‘hole’ and the electron annihilating one another in the
manner that we now understand as a particle and its anti-particle undergoing mutual
annihilation.

59
Fig. 2mc2 = 1.02 MeV.

Conversely, if a hole were not present initially, but a sufficient amount of energy (say
in the form of photons) enters the system, then an electron can be kicked out of one of the
negative-energy states to leave a hole. Dirac’s ‘hole’ is indeed the electron’s antiparticle,
now referred to as the positron.
At first Dirac was cautious about making the claim that his theory actually predicted
the existence of antiparticles to electrons, initially thinking (in 1929) that the ‘holes’
could be protons, which were the only massive particles known at the time having a
positive charge. But it was not long before it became clear that the mass of each hole had
to be equal to the mass of the electron, rather than the mass of a proton, which is about
1836 times larger. In the year 1931, Dirac came to the conclusion that the holes must be
‘anti-electrons’—previously unknown particles that we now call positrons.

13.2 The discovery of positron by Carl D. Anderson

In the next year after Dirac’s theoretical prediction, Carl Anderson announced the
discovery of a particle which indeed had the properties that Dirac had predicted: the first
antiparticle had been found!
Carl David Anderson (September 3, 1905 – January 11, 1991) was an American
physicist. He is best known for his discovery of the positron in 1932, an achievement for
which he received the 1936 Nobel Prize in Physics, and of the muon in 1936. Anderson
was born in New York City, the son of Swedish immigrants. He studied physics and
engineering at Caltech (B.S., 1927; Ph.D., 1930). Under the supervision of Robert A.
Millikan, he began investigations into cosmic rays during the course of which he
encountered unexpected particle tracks in his (modern versions now commonly referred
to as an Anderson) cloud chamber photographs that he correctly interpreted as having
been created by a particle with the same mass as the electron, but with opposite electrical
charge. This discovery, announced in 1932 and later confirmed by others, validated Paul
Dirac's theoretical prediction of the existence of the positron. Anderson first detected the
particles in cosmic rays. He then produced more conclusive proof by shooting gamma
rays produced by the natural radioactive nuclide ThC'' (208Tl) into other materials,
resulting in the creation of positron-electron pairs. For this work, Anderson shared the
1936 Nobel Prize in Physics with Victor Hess.

60
http://en.wikipedia.org/wiki/Carl_David_Anderson

13.2. Dirac sea

The Dirac equation for the free particle leads to a negative energy solution as well as
a positive energy solution. The positive energy solutions and the negative solutions are
separated by a gap shown in Fig.1. Classically, no transition is expected between the
energy gap (= 2 mc2). So we can restrict the energy to be positive classically. On the
other hand, in quantum mechanics, it is expected that the transition can occur. Since
electrons are fermions, all the negative-energy levels are filled with electrons, in accord
with the Pauli exclusion principle. The vacuum state (so called Dirac sea) is one with all
negative-energy levels filled and all positive-energy level empty. We note that the Dirac
sea is a theoretical model of the vacuum as an infinite sea of particles with negative
energy (E<-mc2). The positron, the antimatter counterpart of the electron, was originally
conceived of as a hole in the Dirac sea, well before its experimental discovery in 1932
(C.D. Anderson).

61
 E = mc2

E=0

E =-mc2

Fig.1 The Dirac sea. Positrons as holes in the Dirac sea of negative-energy
electron states. Dirac proposed that almost all negative energy states of
the electron are filled. Pauli principle prevents an electron from falling
into such a filled state. The electron states with the energy above mc2
(denoted by blue open circles) are empty.

Charge:

Qhole  (Qvacuum  ( | e |)  Qvacuum | e |

where Qvacuum is infinite but we have seen such infinite renormalization before.

Momentum:

Phole  ( Pvacuum  p)  Pvacuum   p

where Pvacuum  0 since for each negative energy state with p there is another
with (-p)

62
Energy:

Ehole  [ Evacuum  ( R)]  Evacuum   R  m 2c 4  c 2 p 2

Spin:

    
Σ hole  ( Σ vacuum  Σ )  Σ vacuum   Σ
2 2 2 2 2

In summary
The positron is the antiparticle or the antimatter counterpart of the electron.
The positron has an electric charge of +|e|, a spin of ½, and has the same mass as
an electron.
Positive energy R  m 2c 4  c 2 p 2
Positive charge |e| (>0)
Momentum -p

Spin  Σ
2
Helicity Σ  pˆ

13.3 Pair production

If sufficient energy (more than 2mc2) is given to the system in the form of
radiation, one of the negative energy electrons is excited into an empty state with
a positive energy. Thus we observe an electron of charge -|e| and energy R, and
in addition a hole in the Dirac sea. This hole(anti-particle) has the same mass as
the electron but opposite charge. This hole is called the positron. The process is
called the pair production. The hole registers the absence of an electron of charge
-|e| and energy –R, and would be interpreted by an observer relative to the
vacuum as the presence a particle of charge |e| and energy R.

63
 E = mc2

Radiation
E=0

E =-mc2

Fig.2 The pair production. The supplying of sufficient energy to the Dirac sea
could produce an electron-positron pair:   e   e .

13.4. Annihilation of electron and positron

When a low-energy positron collides with a low-energy electron, annihilation occurs,
resulting in the production of gamma ray photons. An electron falling into a hole would
be interpreted as the annihilation of the electron and the positron, with the release of
energy in the form of radiation

64
 E = mc2

Radiation
E=0

E =-mc2

Fig.3 The annihilation of electron and positron; e− + e+ →γ + γ

___________________________________________________________________
14. Orbital angular momentum L
The Hamiltonian of the free particle is given by

H  cα  p   mc 2

The orbital angular momentum L is defined by

65
 L1  x2 p3  x3 p2 , L2  x3 p1  x1 p3 , L3  x1 p2  x2 p1

We now consider the commutation relation between these operators,

[ H , L1 ]  [c k pk  mc 2 , x2 p3  x3 p2 ]
 [c k pk , x2 p3  x3 p2 ]
 [c 2 p2 , x2 p3 ]  [c 3 p3 , x3 p2 ]
 c 2 [ p2 , x2 ] p3  c 3[ p3 , x3 ] p2
c
 ( 2 p3   3 p2 )
i
c
 (α  p)1
i

Then we get the Heisenberg's equation;

dL i
 [ H , L]  c(α  p) .
dt 


15. Spin angular momentum Σ
2
Here we note that

 0
 k   k   i  i  j (i, j, k; cyclic)
0  k 

or, simply,

σ 0
Σ   
0 σ

0 k 
 k     i 4 k   k  5   5 k
 k 0 

or, simply,

66
 0 σ
α   
σ 0

with

[ 5 ,  k ]  0 , [ , k ]  0 , [ 5 , k ]  0 , [ ,  5 ]  0

[i ,  j ]  2i k , i  j   j i  i k (i, j, and k; cyclic)

[ 5 k ,  j ]   5 k  j   j 5 k   5 [ k ,  j ]

((Note))

0 1 0 0 0  i 0 0 
   
1 0 0 0 i 0 0 0 
1   , 2  
0 0 0 1 0 0 0  i
   
0 0 1 0  0 0 i 0 
  

1 0 0 0
 
0 1 0 0
3   ,
0 0 1 0
 
0 0 0  1


 0 0 1 0 
 
 0 0 0  1
5  
1 0 0 0 
 
 0 1 0 0 
 

Then we get

H  cα  p  mc 2  c 5 k pk  mc 2

We consider the commutation relation,

67
 [ H ,  1 ]  [  c  5  k p k   mc 2 ,  1 ]
  cp k [  5  k ,  1 ]  mc 2 [  ,  1 ]
  cp 2 [  5  2 ,  1 ]  cp 3 [  5  3 ,  1 ]
 cp 2  5 [  1 ,  2 ]  cp 3 5 [  3 ,  1 ]
 2 icp 2  5  3  2 icp 3 5  2
 2 ic ( 2 p 3   3 p 2 )
 2 ic ( α  p )1

which leads to the Heisenberg's equation,

d i i 2c
Σ  [ H , Σ ]  2ic(α  p)   (α  p)
dt   

16. Total angular momentum J

The time derivative of the total angular momentum J is obtained

d d 
J  ( L  Σ )  c (α  p )  c (α  p )  0
dt dt 2


Although L and Σ are not constants of the motion, the total angular momentum J
2
should be identified with the total angular momentum and is a constant of the motion.

  
 L σ 0 
J  L Σ  2 .
2  0 
L  σ 
 2 

As is well known, the constancy of J is a consequence of invariance under rotation.

Hence J must be a constant of the motion even if a central (spherically symmetric)
potential V(r) is added to the free particle Hamiltonian.

((Note))

 ' Hˆ  '   Hˆ 

or

68
 Rˆ Hˆ Rˆ  Hˆ


with

 '  R̂ 

Since

i
Rˆ  exp( Jˆ )


[ Hˆ , Jˆ ]  0 .

17. Helicity Σ  pˆ
We define the helicity operator as

Σ  pˆ

p
where p̂ is the unit vector ( pˆ 
p

σ 0 
Σ    .
0 σ

The eigenstate of helicity with eigenvalue +1 and -1 are referred to, respectively, as the
right-handed state and the left-handed state.

p 1  σ1  p1 0  1  σ 2  p2 0  1  σ 3  p3 0 
Σ  pˆ  Σ        
p p 0 σ1  p1  p  0 σ 2  p2  p  0 σ 3  p3 

or

 p3 p1  ip2 0 0 
 
1  p1  ip2  p3 0 0 
Σ  pˆ  
p 0 0 p3 p1  ip2 
 
 0 0 p1  ip2  p3 

69
  cp3  1 

 p3 0 0  0
2  R  mc 2   
1 R  mc p  ip  p3 0 0  cp3 
( Σ  pˆ )u (1) ( p)   1 2
p 2R  0   R  mc 2 
0 p3 p1  ip2 
  c( p1  ip2 ) 
 0 0 p1  ip2  p 3  
 R  mc 2 

When p1  p2  0 (for simplicity, p3=1)

Σ  pˆ   3
 c 
1 0 0  1 

 2
R  mc 2  0 
R mc 0
1 0 0  
3u (1) ( p)    c 
2R  0 0 1 0  R  mc 2 
  
0 0 0  pˆ 3  0 
 1 
 
R  mc 2  0 
  cp3   u ( p)
(1)

2R 
R  mc 2 
 0 
 

where

R p 2c 2  m 2c 4

Similarly, we have

3u ( 2) ( p)  u ( 2 ) ( p) ,

3u ( 3) ( p)  u ( 3) ( p)

3u ( 4) ( p)  u ( 4 ) ( p)

18. Plane wave solution

Hamiltonian

H  cα  pˆ  mc 2

70
 0 k 
 k     i 4 k   k  5
 k 0 

1 0
   4   
 0  1

 5   1 2 3 4 ,

 k  i  i  j (i, j, k; cyclic).
1  i 2 3 ,
 2  i 3 1 ,
3  i 1 2

H 2  [cα  pˆ  mc 2 ][cα  pˆ  mc 2 ]

 c 2 (α  pˆ ) 2   2 m 2c 4  c(α  pˆ )( mc 2 )  c( mc 2 )(α  pˆ )

(α  pˆ )(α  pˆ )  ( i  pˆ i )( j  pˆ j )
  i j pˆ i pˆ j
1
 ( i j   j i ) pˆ i pˆ j
2
1
 2 ij pˆ i pˆ j
2
 pˆ 2

Note that

2 1

(α  pˆ )    (α  pˆ )   i pˆ i    i pˆ i  ( i    i ) pˆ i  0

Thus we have

H 2  c 2 p 2  m 2c 4 .

We now consider a plane wave given by

71
 i
  u (p) exp[ {p  r  Et )]


H  icα     mc 2


H  i
t

Left-hand side

i
(icα     mc 2 )  (icα     mc 2 )u ( p) exp[ { p  r  Et )]

 (cα  p  mc )
2

right-hand side

 i i
i  iu (p)( E ) exp[ {p  r  Et )]  E
t  

or

H  E

or

Hu (p)  Eu (p)

where

H  cα  p  mc 2

H 2u( p)  EHu ( p)  E 2u( p)

or

E 2  c2 p2  m2c 4

72
or

E  c p 2  m 2 c 2

We now discuss

Hu( p)  Eu( p)

Since

H  cα  p  mc 2  c(i 4 k ) pk   4 mc 2

[c(i 4 k ) pk   4 mc 2 ]u( p)  Eu( p)

Multiplying this equation by  4

E
(i k pk   4   mc)u (p)  0
c

Noting that

E 1
  4   p0 4   p4 4  ip4 4 , p0  E / c, p4  ip0
c i

we obtain

(ip  mc )u( p)  0

where p    p

19. Simultaneous eigenket of H and Σ  pˆ .

We show that H is commutable with Σ  pˆ .

Σ  pˆ   k p̂k

73
  0
 k   k   i  i  j (i, j, k; cyclic)
0  k 

0 k 
 k     i 4 k   k  5   5 k
 k 0 

where

[ 5 ,  k ]  0 , [ , k ]  0 , [ 5 , k ]  0 , [ ,  5 ]  0

Suppose that u( p) is the eigenket of H,

Hu ( p)  Eu( p)

with

H  cα  p  mc 2   c 5 k pk  mc 2

Here we note that

[ H ,  j p j ]  [ c 5 k pk  mc 2 ,  j p j ]
  cpk p j [ 5 k ,  j ]  mc 2 p j [  ,  j ]
  cpk p j [ 5 k ,  j ]

We note that

[ 5 k ,  j ]   5 k  j   j 5 k   5 [ k ,  j ]

[i ,  j ]  2i k , i  j   j i  i k (i, j, and k; cyclic)

Then

c c
[ H ,  j p j ]  cpk p j 5 [ k ,  j ]   p k p j  5 [ k ,  j ]  p j p k  5 [  j ,  k ]
2 2

or

74
 c
[H ,  j p j ]   pk p j { 5 [ k ,  j ]   5[ j ,  k ]}  0
2

So we can demonstrate that

[ H , Σ  pˆ ]  0 .

This implies that u(p) is a simultaneous eigenket of H and Σ  pˆ ,

Hu ( p)  Eu ( p) , ( Σ  pˆ )u( p)  hu( p)

with pˆ  p / p . Since

( Σ  p)( Σ  p)  (i pi )( j p j )

 pi p j (i  j )
1
 pi p j (i  j   j i )
2
 p2

or

( Σ  p) 2  1

( Σ  p ) 2 u ( p )  h 2u ( p )  u ( p )

h2 = 1, or h = ±1.

_______________________________________________________________________
20. Classification of the simultaneous eigenket of H and the helicity
In summary, we have

Hu ( p)  Eu ( p) , ( Σ  pˆ )u( p)  hu( p)

with E   c p 2  m 2c 2 and h = ±1.

75
 E  0  E0  E  0  E0
  ,   ,  ,  
 h 1  h  1  h 1  h  1

Eigenket of the helicity

 σ  pˆ 0  u 
Σ  pˆ   , u   A 
 0 σ  pˆ   uB 

 σ  pˆ 0  u A  u 
( Σ  pˆ )u      h  A 
 0 σ  pˆ  u B   uB 

or

 (σ  pˆ )u A  u 
   h  A 
 (σ  pˆ )u B   uB 

or

(σ  pˆ )u A  hu A , (σ  pˆ )uB  huB (1)

with h  1 .

Eigenvalue problem with the Hamiltonian

H  cα  p  mc 2
 0 σ  p 1 0 
 c   mc 2  
σ  p 0   0  1
 mc 2 c (σ  p) 
  
2 
 c(σ  p)  mc 

Eigenvalue problem:

Hu ( p)  Eu ( p)

or

76
  mc 2 c(σ  p)  u A  uA 
 
 c(σ  p)  mc 2  u   E  u 
  B   B

mc 2u A  c(σ  p)u B  Eu A

c(σ  p)u A  mc 2uB  Eu B

or

c ( σ  p )u B cp
uA   (σ  pˆ )u B (2)
E  mc 2
E  mc 2

c ( σ  p )u A cp
uB   (σ  pˆ )u A
E  mc 2
E  mc 2

(i) h = 1 and E>0

(σ  pˆ )u A  u A , (σ  pˆ )u B  u B

cp cp
uA  (σ  pˆ )u B  uB
R  mc 2
R  mc 2

cp cp
uB  (σ  pˆ )u A  uA
R  mc 2
R  mc 2

We choose

1  cp 
u A    , u B   R  mc 2 
 0  
 0 

or

 1 
 
 0 
u (1)   cp 
 R  mc 2 
 0 
 

77
(ii) h = -1 and E>0

(σ  pˆ )u A  u A , (σ  pˆ )uB  uB

cp cp
uA  (σ  pˆ )u B   uB
E  mc 2
R  mc 2

cp cp
uB  (σ  pˆ )u A   uA
E  mc 2
R  mc 2

We choose

 0  0 
u A    , uB   cp 
1  
 R  mc 2 

or

 0 
 
 1 
u ( 2)  cp 

 R  mc 2 
 0 
 

(iii) h = 1 and E<0

(σ  pˆ )u B  u B , (σ  pˆ )u A  u A

cp cp
uA   (σ  pˆ )uB   uB
R  mc 2
R  mc 2

1  cp 

u B    , u A   R  mc 2 
 0  
 0 

or

78
  cp 
 2 
 R  mc 
u ( 3)  0 
 1 
 
 0 

(iv) h = -1 and E<0

(σ  pˆ )u B  u B

cp cp
uA   (σ  pˆ )u B  uB
R  mc 2
R  mc 2

 0  0 
u B    , u A   cp 
1  
 R  mc 2 

or

 0 
 cp 
 
u ( 4)   R  mc 2 
 0 
 1 
 

21. Foldy-Wouthuysen (FW) transformation

Discussion

H an  En an H   En 

  an ,

Suppose that

U b n  an U'  

 '  bn

Then we have

79
 HU bn  EnU bn HU  '  EnU  '

or

U  HU bn  Ea bn U  HU  '  E n  '

Here we define

H '  U  HU

Then we get

H ' bn  En bn H '  '  En  '

Then  ' is the eigenket of H' with the same eigenvalue En.
________________________________________________________________________
Suppose that

U  e  iS

with S is the Hermitian operator. Then we have

H '  e iS He iS

We note that

 H   ' H'  ' (from the definition)

since

U'   ,   'U

Then we have

 H    ' U  HU  '   ' H '  '

or

H '  U  HU

________________________________________________________________________

80
 U  e  iS

 '  U    e iS    U  '  e  iS  '

H '  U  HU  e iS He  iS

We choose S of the form

i 1
S   (α  p) , eiS  exp[  (α  p) ]
p p

 is a real function to be determined. S is a Hermitian operator.

i  i i
S  ( α  p)   ( p)  ( α  p)  ( p)    ( α  p) ( p)
p p p

since { i ,  }  0 .

H '  eiS He  iS
 eiS (cα  p   mc 2 )e  iS
 eiS [  (c α  p)  mc 2 )e  iS
 eiS  (c α  p  mc 2 )e  iS
 eiS  e  iS eiS (c α  p  mc 2 )e  iS
 eiS  e  iS (cα  p  mc 2 )

since [ S , α  p]  0 . Furthermore

e iS  e iS 

So that

H '  e 2iS  (c α  p  mc 2 )  e 2iS (cα  p   mc 2 )

where  2  1

α  p
H '  (cos 2  sin 2 )(cα  p   mc 2 )
p
α p
  (mc 2 cos 2  cp sin 2 )  ( pc cos 2  mc 2 sin 2 )
p

81
where

α  p α  p   p 2
and

α  p
e 2iS  cos 2  sin 2 (see the Mathematica)
p

R c
p

2q

mc

If we choose (so that odd terms disappear)

p
tan 2  .
mc

p cp mc mc 2
sin 2   , cos 2  
p 2  m 2c 2 R p 2  m 2c 2 R

Then we have

82
 H '   (mc 2 cos 2  cp sin 2 )
mc p
  (mc 2  cp )
p 2  m 2c 2 p 2  m 2c 2
m 2c 2 p2
  c(  )
p 2  m 2c 2 p 2  m 2c 2
  c p 2  m 2c 2  R

where

R  c p 2  m2c 2

So that, H' is now diagonalized. The eigenstate of H’ is the same as that of .

 1  0  0  0
       
 0  1  0  0
 0  0  1  0
       
 0  0  0  1

  U  '  e  iS  '

where

 pz ( p x  ip y ) 
 cos 0  sin   sin  
 p p 
 ( p  ip y ) pz 
 0 cos  x sin  sin  
 iS
 p p 
e 
 pz ( px  ip y ) 
 sin  sin  cos 0 
p p
 
 ( px  ip y ) sin  p
 z sin  0 cos 
 
 p p 

Note that

1 R  mc 2
cos  (1  cos 2 ) 
2 2R

1 R  mc 2
sin   (1  cos 2 ) 
2 2R

83
Then the eigenstate of the original Hamiltonian H is given by

 cos   1 
   
 0   0 
 pz
sin   R  mc 2  cpz 
 p  2R  R  mc 2 
 ( p  ip )   c( px  ip y ) 
 x y
sin    
  R  mc 2 
 p 

 0   0 
   
 cos   1 
 ( px  ip y )  R  mc 2  c( px  ip y ) 
 sin   
p 2 R  R  mc 2 
  
  pz sin   cpz 
   
 p   R  mc 2 

 p   cpz 
  z sin    
 p   R  mc 2 
 ( px  ip y )  R  mc 2  c( px  ip y ) 
 sin   
2 R  R  mc 
2
p
   1 
 cos    
 0   0 
 

 ( px  ip y )   c( px  ip y ) 
 sin    
 p   R  mc 2 
 pz  R  mc 2  cpz 
 sin  
p 2R  R  mc 2 
   0 
 0   
 cos    1 
 

((Mathematica))

84
Clear"Global`"; x   ; y   ; z   ;
0 1 0  1 0
 0 0 1
I2  IdentityMatrix2; I4  IdentityMatrix4;
1 0

x  KroneckerProductx , x ; y  KroneckerProductx, y;

z  KroneckerProductx , z ;
  KroneckerProductz , I2;

H1  c px x  c py y  c pz z   m 2 c2  Simplify;

S    .x px  y py  z pz ;
1
p

K1 

MatrixExp 2  S .   px2  py2  pz2   p,

1   px2  py2  pz2   1  p2   ExpToTrig  Simplify;

K11  K1 .  px2  py2  pz2  p2   Simplify;

K2  Cos2  I4   .x px  y py  z pz Sin2   Simplify;

1

K2  MatrixForm
p

Cos2  0 pz Sin2  px py Sin2 

Cos2  px py Sin2  pz Sin2 

p p
0 
 pz Sinp2   px pyp Sin2  Cos2 
p p

 px pyp Sin2  pz Sin2 

0 Cos2 
p

K11  K2  Simplify
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

Eigensystem 
 1,  1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0

MatrixExp  S..MatrixExp  S    Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

85
 K3 

MatrixExp   S .   px2  py2  pz2   p,

1   px2  py2  pz2   1  p2   ExpToTrig  Simplify;

K31  K3 .  px2  py2  pz2  p2   Simplify;

K31  MatrixForm
Cos 0  pz Sin
 px pyp Sin
Cos  px pyp Sin pz Sin
p
0
pz Sin px py Sin
Cos
p
0
px py Sin pz Sin
Cos
p p
 p
0
p

22. Charge conjugate operator

We start with the Dirac equation for the free particle.
Dirac equation:

 mc
(   )  0
x 

The replacement of

  ie
  A
x x c

leads to the Dirac equation in the presence of A  ( A, iA0 )

 ie mc
(  A )   0
x c 

or

 ie mc
(  A )    0 (1)
x c 

Hermite conjugate of Eq.(1):

 ie  ie *  mc 
(  Ak )  k  ( *  A4 )  4   0
xk c x4 c 

86
or

 ie  ie mc 
(  Ak )  k  (  A4 )  4   0
xk c x4 c 

Multiplying Eq.(1) by  4 from the right, we get

 ie  ie mc 
Ak )  k  4  ( A4 )  4   4  0
2
(  
xk c x4 c 

Here we note that

  4   , { k ,  4 }  0 ,

Then we get

 ie  ie mc 
(  Ak )  4 k  (  A4 )  4 4   4  0
xk c x4 c 

or

 ie  ie mc
(  Ak )  k  (  A4 )  4   0
xk c x4 c 

or

 ie mc
(  A )     0 (2)
x c 

We consider the Dirac equation for the charge conjugate wave function  C ,

 ie mc C
(  A )  C   0 (3)
x c 

We assume that

 C  C T

Then we get

87
  ie mc
(  A )  C T  C T  0 (4)
x c 

Taking the transpose of Eq.(4), we get

 ie mc
 A ) C T     CT  0 .
T
( (5)
x c 

Multiplying Eq.(5) by (C T ) 1 from the right,

 ie mc
 A ) C T   (C T ) 1    0,
T
( (6)
x c 

Comparing Eq.(6) with Eq.(2), we havbe

C T   (C T ) 1   
T

or

C 1  C   
T

Note that

 1T   1 ,  2T   2 ,  3T   3 ,  4T   4

Then we have

[ 1 , C ]  0 , [ 3 , C ]  0

{ 2 , C}  0 , { 4 , C}  0

The above relations can be satisfied when

0 0 0 1
 
0 0 1 0
C   2 4   .
0 1 0 0
 
 1 0 0 0 


C T  ( 2 4 )T   4  2   4 2   2 4  C
T T

(see the Mathematica program below)

88
Then we have

 C  C T   2 4 (  4 )T   2 4 4T (  )T   2 4 4 *   2 *

where

0 0 0  1
 
0 0 1 0
 2  i 2  
0 1 0 0
 
 1 0 0 0 


((Example))

 1   c( p1  ip2 ) 
0 0 0  1   
  0   R  mc 2 
R  mc  0
2
0 1 0  cpz  R  mc 2  cp3 
 2u1* 
2R  0  2    
2
1 0 0 R  mc 2R R mc
  c( p  ip )   0 
1 
0 0 0 
 x y
  
 R  mc 2   1 

 0   cp3 
0 0 0  1   2 
  1   R  mc 
R  mc 2  0 0 1 0  c( px  ip y )  R  mc 2  c( p1  ip2 ) 
 2u 2 *  
2R  0 1 0 0  R  mc 2  2 R  R  mc 
2
    1 
1 0 0 0   cpz 
  
 R  mc  2
 0 

  cp3   0 
0 0 0  1   
 R  mc 
2
  1 
R  mc 2  0 0 1 0   c( p1  ip2 )  R  mc 2  c( p1  ip2 ) 
 2 u 3*   
2R  0 1 0 0  R  mc  2 R  R  mc 2 
2
  1   
 1 0 0 0  cp3
   
 0   R  mc  2

  c( p x  ip y )   1 
0 0 0  1   
 R  mc
2
   0 
R  mc 2  0 0 1 0  cpz  R  mc 2  cpz 
 2u 4* 
2R  0   R  mc  2 R  R  mc 2 
2
1 0 0
  0    c( p x  ip y ) 
 1 0 0 0 
   
 1   R  mc 2 

89
We note that u1, u2, u3, and u4 are obtained from the FW transformation

 1 
 
 0 
R  mc 2  cp z 
u1 
2R  R  mc 2 
 c( p x  ip y ) 
 
 R  mc 2 

 0 
 
 1 
R  mc 2  c( p x  ip y ) 
u2 
2 R  R  mc 2 
 cpz 
 
 R  mc 2 

 cpz 
 
 R  mc 2 
R  mc 2  c( px  ip y ) 
u3 
2 R  R  mc 
2
 1 
 
 0 

 c( px  ip y ) 
 
 R  mc 2 
R  mc 2  cpz 
u4   
2R R  mc 2
 0 
 
 1 

When p1  p2  p3  0

 0  0  0  
       
 0  0 1  0
 2u1    ,
*
 2u 2   
*
 2u3    ,
*
 2u4   
*

0 1 0 0
       
1  0  0  0
       

((Mathematica))

90
Clear"Global`";
exp_  : exp . Complexre_, im_  Complexre,  im ;
x   ; y   ; z   ;
0 1 0  1 0
 0 0 1
I2  IdentityMatrix2;
1 0

x  KroneckerProductx, x; y  KroneckerProductx, y;

z  KroneckerProductx, z;
  KroneckerProductz, I2;

1   .x  Simplify; 2   .y  Simplify;

3   .z  Simplify; 4  ;

Transpose1  1  Simplify
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

Transpose2  2  Simplify
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

Transpose3  3  Simplify
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

Transpose4  4  Simplify
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

91
Charge conjugate C

C1  2.4  Simplify
0, 0, 0, 1, 0, 0,  1, 0, 0, 1, 0, 0,  1, 0, 0, 0

C1  MatrixForm
0 0 0 1
0 0 1 0
0 1 0 0
1 0 0 0

1.C1  C1.1  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

2.C1  C1.2  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

3.C1  C1.3  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

4.C1  C1.4  Simplify

0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

TransposeC1   C1
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

TransposeC1  C1
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

InverseC1  C1
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0

________________________________________________________________________
23. Heisenberg's equation of motion (Heisenberg’s picture)

All operators are given by those in the Heisenberg picture. Here we omit the superscript
(H).

92
 k (H )  k ,  (H )  

 k ( H ) and  ( H ) must be regarded as dynamic variable.

The Hamiltonian

H  cα  π  eA0   mc 2  c k  k  eA0  mc 2

with

e
π  p A, A  ( A, iA0 )
c

1 A
B   A, E  A0
c t

Heisenberg equation for the operator in the Heisenberg picture

d i 
O  [ H , O]  O
dt  t

(a)

d i
Rk  [ H , Rk ]
dt 
i ie i
 c j [ j , Rk ]  [ A0 , Rk ]  [ mc 2 , Rk ]
  
i e
 c j [ p j  Aj , Rk ]
 c
i
 c j [ p j j , Rk ]

i  
 c j Rk
 i R j
 c j j , k
 c k

or

d
R  cα  v
dt

93
(b)

d i
pk  [ H , pk ]
dt 
i ie i
 c j [ j , pk ]  [ A0 , pk ]  [ mc 2 , pk ]
  
i e ie
 c j [ p j  A j , pk ]  [ A0 , pk ]
 c 
i e ie
 c j [ pk , A j ]  [ pk , A0 ]
 c 
i 
e  j ie  A0
A
 c j 
 c i xk  i xk
A j ie A0
 e j 
xk  xk
 ( j A j ) A0
e e
xk xk

or

d
p  e( α  A)  eA0
dt

(c)

d i 
Ak  [ H , Ak ]  Ak
dt  t
i ie i 
 c j [ j , Ak ]  [ A0 , Ak ]  [  mc 2 , Ak ]  Ak
   t
i e ie 
 c j [ p j  Aj , Ak ]  [ A0 , Ak ]  Ak
 c  t
i 
 c j [ p j , Ak ]  Ak
 t
i  Ak 
 c j  Ak
 i x j t
Ak 
 c j  Ak
x j t

or

94
 dA 
 c( α  ) A  A .
dt t

(d)

dπ d e
 ( p  A)
dt dt c
e 
 e( α  A)  eA0  [c( α  ) A  A]
c t
1
 e(  A  A0 )  eα  (  A)
c t
 e( E  α  B )
v
 e( E   B )
c

where

(α  A)  (α  ) A  α  (  A)

(e)
We note that

e
α ( H  eA0 )  ( H  eA0 )α  2c( p  A)  2cπ
c

Using this we obtain a quantum mechanical analogue of the Lorentz equation

d 1 H  eA0 H  eA0
[v ( 2
)( )v ]  π  e( E  α  B )
dt 2 c c2

where

dR
 cα  v ,
dt

d d e
π  ( p  A)  e( E  α  B) .
dt dt c

If H  eA0   mc 2 , depending on whether the state is made of positive or negative energy

solutions of the Dirac equation, we get the equation of motion under the Lorentz forc,

d 1
m v   e( E  v  B )
dt c
95
((Note))

e
{α, H  eA0 }  2c( p  A)
c

or

e
{ i , H  eA0 }  { i , c j ( p j  Aj )  mc 2 }
c
e
 c{ i , j }( p j  Aj )
c
e
 2c ij ( p j  A j )
c
e
 2c( pi  Ai )
c

where

{ i ,  }  0 , { i , j }  2 ij .

(f)

d i
( Σ  π )  [ H , Σ  π ]  eΣ  E
dt 

where

H  cα  π  eA0   mc 2  c( 5 ) Σ  π  eA0   mc 2

with

 k  i 4 k   5 k , and [ 5 ,  k ]  0 .

Then we get

96
 d i
( Σ  π)  [H , Σ  π]
dt 
i
 [c( 5 ) Σ  π  eA0   mc 2 , Σ  π ]

ie
 [ A0 , Σ  π ]

A
 e k 0
xk
 eΣ  E

where

[  , Σ  π ]  [  , 1 1   2 2  3 3 ]
 [  , 1 ] 1  [  ,  2 ] 2  [  , 3 ] 3
0

[ 5 Σ  π, Σ  π ]   5 ( Σ  π )( Σ  π )  ( Σ  π ) 5 ( Σ  π )
0

(g) Zitterbewegung ((Free particle))

A  0 .

dp
0 p = constant
dt

The Hamiltonian of free particle is given by

H  cα  p   mc 2

Heisenberg's equation for the operator ,

97
 dα 1 dv

dt c dt
i
 [H , α]

i
 ( Hα  αH )

i
  ( Hα  αH  2 Hα )

2i
  (cp  Hα )


where

Hα  αH  2cp

((Note))

{H , k }  cpi { i , k }  mc 2 { , i }  2cpi

Since H = const., this equation has a simple solution:

dα 2i
 H (α  H 1cp)
dt 

or

v (t ) 2i
α  H 1cp  exp( Ht )[α (0)  H 1cp]
c 

or

v (t )  H 1c 2 p  ce 2iHt /  [α (0)  H 1cp]

This equation can be integrated:

c 2iHt / 
r (t )  r (0)  H 1c 2 pt  (e  1)[α (0)  H 1cp]
2iH

The first two terms on the right-hand side describe simply the uniform motion of a free
particle. The last term is a feature of relativistic quantum mechanics and connotes a high-
frequency vibration (Zitterbewegung) of the particle with frequency mc 2 /  and
amplitude  /(mc) , the Compton wavelength of the particle.

98
(h) Free particles (continued)
For a free particle Hamiltonian,

{ , H }  { , cα  p  mc 2 }  2mc 2  2  cp k { , k }  2mc 2

and

{ 5 , H }  { 5 , cα  p  mc 2 }
 cpk { 5 ,  k }  mc 2 { 5 ,  }
 cpk { 5 ,  5 k }
 2cpk  k

Hence, in a state of energy E, the operator  has the expectation value

2
mc 2  cp 
     1  
E E

Similarly,

cp
  5    Σ  pˆ 
E

where p̂ is the unit vector of p . The operator  5 is called the chirality.

(i) Lorentz force

A0  0 , A  0 (vector potential)

The Hamiltonian H is given by

e
H  cα  ( p  A)  mc 2 .
c

d i
Σ  [H , Σ ]
dt 

 k   k  5   5 k , [ , k ]  0

 k  i i j (i, j, k: cyclic)

99
 [ 5 ,  k ]  0 , , [ 5 , k ]  0 , [ ,  5 ]  0

[i ,  j ]  2i k , i  j   j i  i k (i, j, and k; cyclic)

[ 5 k ,  j ]   5 k  j   j 5 k   5 [ k ,  j ] .

Using the Mathematica, we get

e
[ H , 1 ]  [c k ( pk  Ak )  mc 2 , 1 ]
c
e
 [c 5 k ( p k  Ak ), i ]
c
e e
 2ic[ 2 ( p3  A3 )   3 ( p2  A2 )]
c c
e
 2ic[α  ( p  A)]1
c

or

e
[ H , Σ ]  2ic[α  ( p  A)]
c

leading to the Heisenberg's equation,

dΣ i 2c e
 [ H , Σ ]   [α  ( p  A)]
dt   c

So we have

dΣ dΣ 2c 2 e e
H  H  [α  ( p  A)][α  ( p  A)]
dt dt  c c
e e
 α  ( p  A)][α  ( p  A)]
c c
 2ecΣ  B

In the relativistic approximation, H  mc 2 . Then we have

100
 dΣ 2ec e
 2
ΣB  ΣB
dt 2mc mc

Since

σ 0 
Σ   
0 σ

dσ e
 σB
dt mc

For the one electron spin operator


S σ
2

dS e
 S  B  μ B
dt mc

or

e e
μ S σ
mc 2mc

24. Central force problem: hydrogen atom

The Hamiltonian is given by

H  cα  p  mc 2  eA0

where

A  ( A, iA0 )

with

A=0 eA0  V (r ) (spherical symmetry)

(a) The commutation relation between J3 and H

[H , J3 ]  0

((Proof))

101
 [ H  eA0 , L3 ]  [cα  p   mc 2 , L3 ]
 c[ pk , L3 ] k  mc 2 [  , L3 ]
 c[ pk , x1 p2  x2 p1 ] k
 c[ p1 , x1 p2  x2 p1 ]1  c[ p2 , x1 p2  x2 p1 ] 2
 
 cp21  cp1 2
i i

 c( p21  p1 2 )
i
c
 (α  p ) 3
i

or

[ H , L3 ]  ic(α  p) 3  [eA0 , L3 ]  ic (α  p) 3

since

[eA0 , L3 ]  [eA0 , xp y  ypx ]

  x[ p y , eA0 ]  y[ px , eA0 ]
   
 x eA0  y eA0
i y i x
0

where

A0  A0 (r ) , r  x2  y2  z 2

  r r A xy xy A
x A0  y A0  ( x  y ) 0  (  ) 0  0
y x y x r r r r

Similarly

102
 [ H  eA0 , 3 ]  [c k pk   mc 2 , 3 ]
 [c k pk , 3 ]  [ mc 2 , 3 ]
 cpk  5[ k , 3 ]
 cp1 5 [1 , 3 ]  cp2 5 [ 2 , 3 ]
 2icp1 5 2  2icp2 51
 2ic(1 p2   2 p1 )
 2ic(α  p)3

or

[ H , 3 ]  2ic(α  p)3  [eA0 , 3 ]

 2ic(α  p)3

or


[ H , 3 ]  ic(α  p)3
2

where

 k   5 k , [ 5 ,  k ]  0 , [ , k ]  0 , { 5 ,  }  0

[ k , 1 ]  [ 5 k , 1 ]   5 [ k , 1 ]

[i ,  j ]  2i k (i, j, k; cyclic)

Thus we have


[ H , J 3 ]  [ H , L3  3 ]  ic(α  p)3  ic(α  p)3  0
2

Similarly, we have

[ H , J1 ]  0 , [H , J 2 ]  0 .

or

[H , J ]  0 .

103
(b) Definition of the operator K and the commutation relation [ K , H ]  0

First we show that


[ H , Σ  J ]  [ H ,  ]
2

where

[H , J ]  0 and [ , Σ ]  0

((Proof))

[ H ,  Σ  J ]  H ( Σ  J )   ( Σ  J ) H
 [ H ,  ]( Σ  J )   [ H , Σ ]  J
 2c (α  p)( Σ  J )  2ic (α  p)  J

Here we note that

[ H ,  ]  [cα  p,  ]  cα  p   cα  p  2cα  p

[ H , Σ ]  2ic(α  p)

( α  p)( Σ  J )   5 ( Σ  p)( Σ  J )
  5 [ p  J  iΣ  ( p  J )]
  5 p  J  iα  ( p  J )

Then we have

[ H , Σ  J ]  2c [  5 p  J  iα  ( p  J )]  2ic ( α  p)  J
 2c 5 p  J

 2c 5 p  ( L  Σ)
2
 c 5 p  Σ
  cα  p

 [H ,  ]
2

where

(α  p)  J  α  ( p  J )

104
Then we define the operator K as


K  Σ  J  
2

  (Σ  J  )
2
 
  (Σ  L  Σ 2  )
2 2
  ( Σ  L  )

where


J  L Σ
2

Then K commutes with H,

[K , H ]  0 .

This also implies that

[K 2 , H ]  0 .

25. Commutation relations (continued)

(i) [ , J ]  0

 
[  , J 1 ]  [  , L1  1 ]  [  , 1 ]  0
2 2

(ii) [ Σ  J , J ]  0

105
 
[ Σ  L, J1 ]  [  1 L1    2 L2    3 L3 , L1  1 ]
2

 [  1 L1    2 L2    3 L3 , L1 ]  [  1 L1    2 L2    3 L3 , 1 ]
2
 
    2 [ L1 , L2 ]    3[ L3 , L1 ]  [   2 L2 , 1 ]  [  3 L3 , 1 ]
2 2
 
 i  2 L3  i  3 L2  L2  [1 ,  2 ]  L3  [ 3 , 1 ]
2 2
 i  2 , L3  i  3 L2  iL2   3 iL3  2
0

since

[ , k ]  0

(iii) [ Σ  J , J ]  0


[ Σ  J , J 1 ]  [ Σ  ( L  ), J 1 ]
2

 [ Σ  L, J 1 ]  [   2 , J 1 ]
2

 [3 , J 1 ]
2
0

or

[ Σ  J , J ]  0

which leads to

[ K , J ]  [  ( Σ  L  ), J ]  0

26. K2 and J2

K 2   ( Σ  L  ) ( Σ  L  )
 ( Σ  L)( Σ  L)  2Σ  L   2
 L2  i  ( L  L)  2Σ  L   2
 L2  Σ  L   2

since

106
 [ , k ]  0 .

We note that

 
J 2  (L  Σ )  (L  Σ )
2 2

 L2  Σ 2  Σ  L
4
3 2
L 
2
 Σ  L
4

Thus we obtain

2
K2  J2 
4

Since [ K 2 , H ]  0 , we also have the commutation relation

[J 2 , H ]  0

27. Simultaneous eigenket

For an electron in a central potential, we can conduct a simultaneous eigenfunction of H,
K, J2 , and J3,

H  E , K   ,

J 2   2 j ( j  1) , J 3  j3

P  

since

[H , K ]  0 , [H , J 3 ]  0 , [H , J 2 ]  0 , [ P, H ]  0

We also note that

2
K J 
2 2

This implies that

107
 2 1
K 2   2 2  [ 2 j ( j  1)  ]   2 ( j  ) 2
4 2

or

1 1
 2  [ j ( j  1)  ]  2 ( j  ) 2
4 2

So we must have

1
  ( j  ) ,
2

Note that j is a half-integer and  is an integer (   1,2,... ). So  has integer

eigenvalues not zero.

28. Operator K
We now consider the matrix of K.

K   ( Σ  L  )
 1 0  σ  L   0 
   
 0  1 0 σ  L   
σ  L   0 
  
 0  (σ  L   ) 

The wave function  is a simultaneous function of K , J 2 , and J 3 ,

 
   A 
 B 

Then we have

K  

or

σ  L   0  A   
      A 
 0  (σ  L   )  B   B 

or

108
 (σ  L   ) A   A , (σ  L   ) B   B

or

(σ  L) A   (  1) A , (σ  L) B  (  1) B

29. Operators J 2

J 2   2 j ( j  1)

  
 L σ 0 
J  L Σ  2 
2  0 
L  σ 
 2 

    
L σ 0  L  σ 0
J2  2  2 
 0    
L  σ  0 L σ
 2  2 
 2

  L   σ  0 
 2  
  2
   
0  L  σ  
  2  

Then we get

J 2 A, B   2 j ( j  1) A, B

or

2 2
     
 L  σ   A   j ( j  1) A ,  L  σ   B   j ( j  1) B
2 2

 2   2 

30. Operator J z

  
  L3  3 0    
J 3  ( L3   3 )   2  A   j3 A 
L3   3  B   B 
2  
0
 2 

or

109
  
( L3   3 ) A  j3 A , ( L3   3 ) B  j3 A
2 2

31 The operator L2

Since [ H , L2 ]  0 ,  is not the eigenfunction.

3 2
L2  J 2  Σ  L 
4

we have

 2 3 2 
 J  σ  L  0  
L2   4  A 
2 
 3  B 
 0 J 2  σ  L  
 4 

Then we get

3 2 2
( J 2  σ  L  ) A  [ J 2   (σ  L  )  ) A
4 4
2
 [ 2 j ( j  1)   2  ] A
4
  l A (l A  1) A
2

where

1
j ( j  1)     l A (l A  1)
4

Similarly,

3 2 2
( J 2  σ  L  ) B  [ J 2  (σ  L  )  ] B
4 4
2
 [ 2 j ( j  1)   2  ] B
4
  lB (lB  1) B
2

with

110
 1
j ( j  1)     l B (l B  1)
4

Thus  A and  B are separately the eigenfunctions of L2 . These eigenvalues are denoted
by l A (l A  1) 2 and  2lB (lB  1) , respectively.

Using these two equations, we can determine lA and lB for the given eigenvalue .

((Nonrelativistic case))

n = 1; l = 0

D0  D1 / 2  D1 / 2 (j = 1/2)

n = 2; l = 0, 1

D0  D1 / 2  D1 / 2 (j = 1/2)
D1  D1 / 2  D3 / 2  D1 / 2 (j = 3/2, 1/2)

n = 3; l = 0, 1, 2

D0  D1 / 2  D1 / 2 (j = 1/2)
D1  D1 / 2  D3 / 2  D1 / 2 (j = 3/2, 1/2)
D2  D1 / 2  D5 / 2  D3 / 2 (j = 5/2, 3/2)

________________________________________________________________________

(a) For j = 1/2,

1
  ( j  )  1 .
2

(i)   1 , lA = 1, and lB = 0.
(ii)   1 , lA = 0 and lB = 1.

(b) For a half integer j,

1
  ( j  ) .
2

1 1 1
(i)   j , lA  j  , lB  j 
2 2 2

111
 1 1 1
(ii)   ( j  ) ,` l A  j  , lB  j 
2 2 2

________________________________________________________________________
32. Normalized spin angular function

Spin orbit coupling

Dl  D1 / 2  Dl 1 / 2  Dl 1 / 2

For j  l  1 / 2 ,

l  j3  1 / 2 j3 1 / 2  1  l  j3  1 / 2 j3 1 / 2  0 
y lj l 1 / 2, j3  Yl    Yl  
2l  1  0 2l  1  1
 1 
 l  j3  Yl j3 1 / 2 
1  2 

2l  1  1 j3 1 / 2 
 l  j3  Yl 
 2 

For j  l  1 / 2 ,

l  j3  1 / 2 j3 1 / 2  1 
y lj l 1 / 2, j3   Yl  
2l  1  0
l  j3  1 / 2 j3 1 / 2  0 
 Yl  
2l  1 1
 1 
  l  j3  Yl j3 1 / 2 
1  2 

2l  1  1 j3 1 / 2 
 l  j3  Yl 
 2 

_____________________________________________________________________
33. Radial wave functions

1 1 1
(a) For   j  , lA  j  , lB  j 
2 2 2

 g ( r ) y j , j3 1 
 lA  j 
  j , j3
2

 if ( r ) y 1 
lB  j 
 2 

112
 1 1 1
(b) For   ( j  ) , lA  j  , lB  j 
2 2 2

 f ( r ) y j , j3 1 
 lA  j 
  j , j3
2

 ig ( r ) y 1 
lB  j 
 2 

The parity of y j , j3 1 is given by (1)l A , while the parity of y j , j3 1 is given by (1)l B .

lA  j lB  j 
2 2

Theses parities are different, since l B  l A  1 . The radial functions f and g depend on .
The factor i multiplying f and g is inserted to make f and g real for bound-state.

((Note))

 A (  r , t )  ( 1) l  A ( r , t )   A ( r , t )
A

 B (  r , t )  ( 1) l  B ( r , t )   B ( r , t )
B

Thus we have

( 1) l A  ( 1) l B 1

Thus we get the relation

l A  l B  1 .

Table-1

________________________________________
 lA lB 
1 1 1
  ( j  ) lA  j  lB  j 
2 2 2
1 1 1
 (j ) lA  j  lB  j 
2 2 2
________________________________________

Table-2

1
j
2
  1, lA  1, lB  0
  1 , lA  0 , lB  1

113
 3
j
2
  2, lA  2 , lB  1
  2 , lA  1, lB  2
5
j
2
 3, lA  3, lB  2
  3 , lA  2 , lB  3

34. Expression of the two-component wave function

For a fixed  [= j+1/2, or -(j+1/2)] we assume that the wave function is given by

 A   f ( r ) y l A 3 
j, j

  
   
j , j3 

 B  ig ( r ) y lB 

This function satisfies the Dirac equation given by

c(σ  p) B  ( E  V (r )  mc 2 ) A ,

c(σ  p) A  ( E  V (r )  mc 2 ) B

We note that

1
σ p (σ  r )(σ  r )(σ  p )
r2
1
 2 (σ  r )[ r  p  σ  ( r  p)]
r
1
 2 (σ  r )( r  p  iσ  L)
r
1 
 2 (σ  r )( ir  iσ  L)
r r

where

(σ  r )(σ  r )  r 2  iσ  (r  r )  r 2

 
r p r
i r

114
    1  1 
p    (e r  e  e )
i i r r  r sin  

For two arbitrary vectors A and B,

(σ  A)(σ  B )  ( A  B)1̂  iσ  ( A  B)

Then we get

(σ  p) B  i (σ  p) f ( r ) yljB, j3
i 
 (σ  r )( ir  iσ  L) f ( r ) yljB, j3
r 2
r
df (1   )
 (σ  rˆ)[   f ] yljB, j3
dr r
df j , j3 (1   ) j , j3
  yl  fyl A
dr A r

where

(σ  L) B  (  1) B , (σ  rˆ) y ljB, j3  y ljA, j3 .

Similarly, we get

(σ  p) A  (σ  p) gyljA, j3
1 
 (σ  r )(ir  iσ  L) gyljA, j3
r 2
r
 dg
 2 (σ  r )[ir  i (  1) g ] yljA, j3
r dr
dg
 [i  i (  1)g ] ylBj , j3
dr
dg j , j3 (  1) j , j3
 i ylB  i gylB
dr r

where

(σ  L) A   (  1) A , (σ  rˆ) y ljA, j3   y ljB, j3

35. The operator (σ  rˆ)

(a) {P, Σ  rˆ}  0

115
with P  

((Proof))

{P, Σ  rˆ}  { , Σ  rˆ}

 Σ  rˆ  Σ  rˆ
 Σ  rˆ  Σ  rˆ
 Σ  rˆ  Σ  rˆ
 [  , Σ ]  rˆ
0

or

  (σ  rˆ )  (σ  rˆ ) 0 
 0
 0   (σ  rˆ)   (σ  rˆ) 

or

 (σ  rˆ)  (σ  rˆ)  0

where, P   and rˆ  rˆ  0 . Thus we have

σ  rˆ is odd under the parity.

(b) (σ  rˆ) 2  1

((Proof))

(σ  rˆ)(σ  rˆ)  rˆ  rˆ  iσ  ( rˆ  rˆ)  1

(c) [ J 3 , σ  rˆ]  0

((Proof))

J 3  L3  S 3  L3   3
2

 
L3 
i 

σ  rˆ   1 sin  cos    2 sin  sin    3 cos 

Then we can evaluate the commutation relation

116
 
[ J 3 , σ  rˆ]  [ L3 , σ  rˆ]  [ 3 , σ  rˆ]
2

[ L3 , σ  rˆ]  L3 [(σ  rˆ) ]  (σ  rˆ) L3

 
 [( 1 sin  cos    2 sin  sin    3 cos )
i 
 ( 1 sin  cos    2 sin  sin    3 cos ) L3
 
  ( 1 sin  cos    2 sin  sin    3 cos )
i 
 ( 1 sin  cos    2 sin  sin    3 cos ) L3
 ( 1 sin  cos    2 sin  sin    3 cos ) L3
 i (  1 sin  sin    2 sin  cos  )

or

[ L3 , σ  rˆ]  i 1 sin  sin   i 2 sin  cos 

We also have

 
[  3 , σ  rˆ]  [ 3 ,  1 sin  cos    2 sin  sin    3 cos ]
2 2
 
 [ 3 , 1 ] sin  cos   [ 2 , 3 ] sin  sin 
2 2
 i 2 sin  cos   i 1 sin  sin 

Thus we have

[ J 3 , σ  rˆ]  0

36. Evaluation of (σ  rˆ) yljA, j3 and (σ  rˆ) yljB, j3

(a)

[ J 3 , σ  rˆ] yljA, j3  0

or

[ J 3 (σ  rˆ) yljA, j3  (σ  rˆ) J 3 yljA, j3  j3 (σ  rˆ) yljA, j3

which means that (σ  rˆ) yljA, j3 is the eigenfunction of J3 with the eigenvalue j3.

117
(b)

 (σ  rˆ) ylj , j  (σ  rˆ )ylj , j  (1)l

A
3
A
3 A 1
(σ  rˆ) yljA, j3

 (σ  rˆ) ylj , j  (σ  rˆ )ylj , j  (1)l

B
3
B
3 B 1
(σ  rˆ) yljB, j3

leading to the relation

(σ  rˆ) yljA, j3  cyljB, j3 , (σ  rˆ) yljB, j3  cyljA, j3

where c is constants. We note that

(σ  rˆ) 2 yljA, j3  c(σ  rˆ) yljB, j3  c 2 yljA, j3  yljA, j3

since

(σ  rˆ) 2  1

Then we get c  1 . Here we choose c = -1.

(σ  rˆ ) yljA, j3   yljB, j3 , (σ  rˆ ) yljB, j3   yljA, j3

((Note)) In the non-relativistic quantum mechanics, it is well known that

ˆ l , m  ( 1) l l , m

37. Radial wave function in hydrogen atom

Now we solve the Dirac equation as

c(σ  p) B  ( E  V ( r )  mc 2 ) A ,

or

df j , j3 (1   )c j , j3
c(σ  p) B   c yl A  fyl A  ( E  V ( r )  mc 2 ) gy ljA, j3
dr r

or

df (1   )c
 c  f  ( E  V ( r )  mc 2 ) g
dr r

118
Similarly,

c(σ  p) A  ( E  V ( r )  mc 2 ) B

or

dg (1   )c
c  g  ( E  V ( r )  mc 2 ) f
dr r

Introducing

F ( r )  rf ( r ) , G ( r )  rg ( r )
then we have a radial equations,

dF 
c(  F )  ( E  V ( r )  mc 2 )G
dr r

dG 
c(  G )  ( E  V ( r )  mc 2 ) F
dr r

We assume that V(r) is given by a Coulomb potential

Ze 2
V (r )  
r

We put

mc 2  E mc 2  E
1  , 2 
c c

Ze 2 2
   Z ,   1 2 r , 
c 1

where  is the fine structure constant,

e2 1
 =7.29735257 x 10-3,  137.035999074( 44).
c 

Then we get the coupled equations we need to solve,

119
 d  
(  ) F  (   )G  0
d  

d  1 
(  )G  (  ) F  0
d   

The analysis of the radial equation proceeds as usual.

 ,

dF dG 1
 G ,  F
d d 2

d 2F  2 dG  2 1
  FF
d 2
 1 d 1  2

Similarly,

d 2G
G
d 2

F  e  , G  e 

We assume that

F  e    s  am  m
m 0

G  e    s  bm  m
m0

We solve the problem using a series expension method. These series forms are
substituted into the coupled differential equation. We use the Mathematica to determine
the value of s and the recursion relation. The results are as follows.

38 Indicial equation to determine the value of s

( s   )a0  b0  0
 a0  ( s   )b0  0

or

120
 s    a 0   0 
     
  s    b0   0 

Since a0 and b0 are not zero (non-trivial solution), the determinant of the matrix should be
equal to zero.

s   2  2

Note that

s 2   2   2  min( 2 )   2  1  Z 
2

So we get approximately s  1 , or s<-1. However, we must require that

  (r)
2
r 2 dr  

The requirement amounts to

2
F (r)
 f ( r ) r dr   r 2 dr   F ( r ) dr   F (  ) d  
2 2 2 2
2
r

2
G(r)
 g (r) r dr   r 2 dr   G ( r ) dr   G (  ) d  
2 2 2 2
2
r

Around the origin,

F  s, G  s

Then we have

 2 s 1
 F (  ) d    2 s d 
2

2s  1

 2 s 1
 G(  ) d    d 
2 2s

2s  1

So in order to get the finite value of the probability near the origin, it is required that

121
 1
s
2

So we need to take

1
s   2   2  ( j  ) 2  Z 2 2
2

39. Mathematica (series expansion method)

122
40. Recirsion relation

(i) The second recursion relations

( s  1   )a1  a0  b1  b0  0

1
( s  1   )b1  b0  a1  a0  0


(ii) The recursion relations (the general case)

( s  q   )a q  a q 1  bq  bq 1  0
1
( s  q   )bq  bq 1  a q  a q 1  0


The functions F and G would increase exponentially as    if the power series do not
terminate. Assuming that the two series terminates with the same power, there must be
exist nr with the property. For q = nr, we assume that

a nr 1  bnr 1  0 , anr  0 , bnr  0

123
Then we get

a nr   bnr (1)

From the recursion relation (in general)

( s  q   )a q  bq  a q 1  bq 1

[( s  q   )bq  a q ]  a q 1  bq 1

we get the relation

[( s  q   )bq  a q ]  ( s  q   )a q  bq

or

[  ( s  q   )   ]bq  ( s  q     )a q . (2)

or

aq bq
Cq  
 1
s  q   (s  q   )  
 

for q = nr, nr-1, ...,0.

41. Derivation of the energy eigenvalue

From Eqs.(1) and (2) with q  n r , we have

[  ( s  nr   )   ]bnr  ( s  nr     )a nr    ( s  n r     )bnr

or

[  ( s  n r   )   ]    ( s  n r     )

or

1 2 (   2 )
s  nr    1
2 2 1 2

or

124
 2 1 2 ( s  n r )   (1   2 )

Noting that

2E m2c 4  E 2
1   2  , 1 2 
c c

we have the energy eigenvalue as

m 2 c 4  E 2 ( s  n r )  E

or

mc 2 mc 2
E 
2 Z 2 2
1 1
(nr  s ) 2 1 1
(n  j   ( j  ) 2  Z 2 2 ) 2
2 2

This is famous fine structure formula for the hydrogen atom. The quantum numbers j and
nr assume the values

1 3 5
j  , , ,... . nr = 0, 1, 2, 3,…
2 2 2

The principal quantum number n of the nonrelativistic theory of the hydrogen atom is
related to nr and j by

1
n j  nr
2

n=1

1
nr = 0, j 
2
(l = 0, s = 1/2) j = 1/2 1 2S1/2   1

n=2

3
nr = 0, j 
2
(l = 1, s = 1/2) j = 3/2 2 2P3/2   2

125
 1
nr = 1, j 
2
(l = 0, s = 1/2) j = 1/2 2 2S1/2   1
(l = 1, s = 1/2) j = 1/2 2 2P1/2  1

n=3

5
nr = 0, j 
2
(l = 2, s = 1/2 j = 5/2 3 2D5/2)   3

3
nr = 1, j 
2
(l = 1, s = 1/2 j = 3/2 3 2P3/2)   2
(l = 2, s = 1/2 j = 3/2 3 2D3/2)  2

1
nr = 2, j 
2
(l = 0, s = 1/2 j = 1/2 3 2S1/2)   1
(l = 1, s = 1/2 j = 1/2 3 2P1/2)  1

___________________________________________________________________
Table-2

1
j
2
  1, l 1
  1 , l 0
3
j
2
  2, l2
  2 , l 1
5
j
2
 3, l 3
  3 , l2

_______________________________________________________________________
Table 3 Notation in the nonrelativistic case

n=1
l = 0, s = 1/2 j = 1/2 1 2S1/2
n=2

126
 l = 0, s = 1/2 j = 1/2 2 2S1/2
l = 1, s = 1/2 j = 3/2 2 2P3/2
l = 1, s = 1/2 j = 1/2 2 2P1/2
n=3
l = 0, s = 1/2 j = 1/2 3 2S1/2
l = 1, s = 1/2 j = 3/2 3 2P3/2
l = 1, s = 1/2 j = 1/2 3 2P1/2
l = 2, s = 1/2 j = 5/2 3 2D5/2
l = 2, s = 1/2 j = 3/2 3 2D3/2

42. Energy levels

The energy E

mc 2
E   mc 2
Z 
2 2
1
1 1
(n  j   ( j  ) 2  Z 2 2 ) 2
2 2

can be expanded by using a Taylor expansion in a power of Z 2 2 .

127
 E = mc2
En hydrogen atom

E=0

E =-mc2

Fig. The energy levels of electron in the hydrogen atom (in the relativistic quantum
mechanics)

Using the Mathematica, we have

128
 E  E  mc 2
mc 2 mc 2 (6 j  3  8n)
  2 ( Z ) 
2
( Z ) 4  ...
2n 8(1  2 j )n 4

mc 2 mc 2 1 3
  2 ( Z  )  3 ( Z ) 4 (
2
 )  ...
2n 2n 1 4n
j
2

The first tem is the non-relativistic limit

mc 2 13.6057 Z 2
 ( Z  ) 2
  [eV]
2n 2 n2

The second term is the relativistic correction to E.

The principal quantum number n are n = 1, 2, 3, 4, … and j  1 / 2  n . There is the
degeneracy between 2 2S1/2 ans 2 2P1/2.states (similarly 3 2S1/2 ans 3 2P1/2, 3 2P3/2 ans 3
2
D3/2) persists in the eaxct solution to the Dirac equation. This degeneracy is lifted by the
Lamb shift due to the coupling of electron to the zero-point fluctuation of the radiation
field.

129
Fig. Detail of an energy-level diagram for the hydrogen atom. The manifolds of the n
= 2 and n = 3 levels are shown, based on the Dirac theory, without radiative
corrections (Lamb shifts) or hyperfine splittings. The energy differences are given
in the units of cm-1. 1eV = 8065.56 cm-1. ((Merzbacher, Quantum Mechanics)

((Mathematica))
The energy is in the units of cm-1; E ( erg ) /(2c ) .
1 eV = 8065.56 cm-1

130
 Clear"Global`";
rule1  c  2.99792  1010 , —  1.054571628 1027 ,
me  9.10938215 1028 , eV  1.602176487  1012 ,
  7.2973525376  103 , Z  1;

me c2
E0   me c2 ;
Z2 2
1
j1  1   Z2 2
2
2
n1  j1  1 
2 2

. rule1
1 eV
2—c
8065.56

SeriesE0,  , 0, 4 
FullSimplify , j1  0, n1  0 &

c2 me Z2  2 c2 me 3  6 j1  8 n1 Z4 4
 O5
8 1  2 j1 n1
 
E1 n_, j_ : E0  2  — c . n1  n, j1  j . rule1
2 4
2 n1

E13, 5  2  E13, 3  2
0.0360719

E13, 3  2  E13, 1  2
0.108219

E12, 3  2  E12, 1  2
0.365241

E13, 1  2  E12, 1
 2
15 241.6

43. Wave function for the ground state

Suppose that nr = 0. Then we have

131
 a1  b1  0 , a0  0 , b0  0

From the recursion relation,

 a0  b0  0

or

a0   b0 (1)

From the indicial equation

 a0  ( s   )b0  0 (2)

Using these two equations, we have

a0 s  
    0
b0 

where

s  2  2  

Then we have

s   0 and 0s 

or

  s  0

The absence of the   0 state for nr = 0 corresponds to the familiar rule in relativistic
quantum mechanics.

Ground state:

1
n j  nr
2

with n = 1. nr = 0, j = 1/2.

132
 mc 2 mc 2
E   mc 2 1  Z 2 2 (ground state energy)
Z 2 2 1
1
1  Z 2 2 1  Z 2 2

a0 mc 2  E 1  1  Z 2 2 2
      
b0 mc  E
2
Z Z

1 2 r mc
  ( Z ) r .
c 

with

1 2  (mc 2  E )(mc 2  E )  m 2c 4  E 2  mc 2 ( Z )

1
  ( j  )  1 since j = 1/2.
2

1
s   2   2  1  Z 2 2  1  Z 2 2
2

Then the radial wave function of the ground state are given by

1 2   s
f (r )  a0 e 

 a0 1 2 e    s 1
1
 Z 2 2

 a0 mc ( Z )e 
2 2

1
 Z 2 2
 A0e   2

and

133
 1 2   s
g (r )  b0 e 

 b0 1 2 e    s 1
1
 Z 2 2

 b0 mc ( Z )e 
2 2

1
1  Z 2 2
  a0 mc 2 ( Z ) 2 e    2
2
1
1  Z 2 2
  ( Z ) A0e    2
2

with

Z
A0  a0 (mc 2 ) Z , b0   a0
2

The upper component f(r) is very similar to the non-relativistic wave function except for
1
 Z 2 2
an enhanced (singular) part at small  which goes like  2
. This singularity is very
weak, and the solution is still integrable near the origin. The lower component g(r) is
1
very much smaller (by a factor of Z ) than the upper component. Thus the relativistic
2
solution differs from the non-relativistic solution only to the order of Za, or at very short
distances.

((Note)) The radial function for the ground state in the non-relativistic theory

2Z 3 / 2  a
rZ
mc
R10  3/ 2
e  2( )3 / 2 ( Z )3 / 2 e  
a 

with

rZ rZ mc
  2 me 2  r ( Z )
a  

where

2
a (Bohr radius) ,
me 2

Z mc
 Z .
a 

134
44. Heisenberg's principle of uncertainty
In the Dirac theory,

Ze 2
H  cα  p  mc 2 
r

From the Heisenberg's equation of motion, we get the relations,

1
α v
c

H  H  2mc 2

When

mc 2 1 v2
2  H  2mc 2 ,    1 2
E  c

The Heisenberg's principle of uncertainty:

pr r  

((Special relativity))

p  mv , E  mc 2  c m 2c 2  p 2

with

1

v2
1
c2

So we get the Hamiltonian

1 Ze 2
H  v  p  mc 2 
 r
Ze 2 1
 mv  mc 
2 2

 r
Ze 2
c p m c 
2 2 2

135
Note that

1 1
v  p  mc 2  v  mv  mc 2
 
mv 2 v2
  mc 2 1 
v2 c2
1 2
c
2
mc

v2
1
c2
 m 2 c  cp

We now consider the Hamiltonian given by

Ze 2
H  c (pr ) 2  m 2c 2  pr


We take a derivative of H with respect to pr

 cpr Ze 2
H  0
 (pr ) (pr ) 2  m 2c 2 

Then we get

cpr Ze 2

(pr ) 2  m 2c 2 

or

pr 2  Z  
2

(pr )  m c
2 2 2

From this, pr  can be obtained as

2

pr 2  m2c 2 Z  2

2

1  Z  

or

136
 mcZ
pr 
1  Z 
2

From the relation pr r   we get

  
1  Z  
2
r  
pr mcZ mcZ

Then the local minimum of H is given by

H  mc 2 1  ( Z ) 2

which is exactly the same as the value of Eground in the relativistic theory.

45. Determination of Cq (aq, bq)

We derive the recursion relation for Cq fro the relation

( s  q   )a q  bq  a q 1  bq 1

aq bq
Cq  
 1
s  q   (s  q   )  
 

From these equations we get

1
 (   )  2( s  q  1)

Cq  Cq 1
q(2s  q )

where

For q = nr+1, Cq = 0. Then we have

1
 (   )  2( s  nr )  0


Finally we get the recursion formula,

137
 2( s  q  1)  2( s  nr )
Cq  Cq 1
q(q  2s)
2(q  1  nr )
 Cq 1
q(q  2s )

For q = 1,

2(nr )
C1  C0 ,
1(1  2s )

For q = 2,

2(1  nr )
C2  C1
2(2  2s )
2(1  nr ) 2( nr )
 C0
2(2  2s ) 1(1  2s )
22 (1) 2 nr (nr  1)
 C0
2!(1  2 s )(2  2 s )

For q = 3,

2(2  nr )
C3  C2
3(3  2s )
23 (1)3 nr (nr  1)(nr  2)
 C0
3!(1  2s )(2  2s )(3  2s )

In general

2k (1) k [nr !/(nr  k )!]

Ck  C0
k!(1  2s )(2  2 s )...(k  2 s )

where

aq bq
Cq  
 1
s  q   (s  q   )  
 

________________________________________________________________________
REFERENCES
E. Fermi, Notes on Quantum Mechanics (The University of Chicago Press, 1961).

138
J.D. Bjorken and S.D. Drell, Relativistic Quantum Mechanics (McGraw-Hill, New York,
1964)
J.J. Sakurai, Advanced Quantum Mechanics (Addison-Wesley, New York, 1967).
M.E. Rose, Relativistic Electron Theory (John Wiley & Sons, New York, 1961).
A. Das, Lectures on Quantum mechanics, second edition (World Scientific, 2012).
F.J, Dyson, Advanced Quantum Mechanics (World Scientific, 2007).
S.S. Schweber, An Introduction to Relativistic Quantum Field Theory (Row, Peterson,
1961).
F. Schwabl, Advanced Quantum Mechanics (Springer Verlag, Berlin, 2005).
F. Gross, Relativistic Quantum Mechanics and Field Theory (Wiley-VCH, 1993).
Eugen Merzbacher, Quantum Mechanics, third edition (John Wiley & Sons, New York,
1998).
Ramamurti Shankar, Principles of Quantum Mechanics, second edition (Springer, New
York, 1994).
Leonard Schiff, Quantum Mechanics (McGraw-Hill Book Company, Inc, New York,
1955).
B.R. Holstein, Topics in ADVANCED Quantum Mechanics (Addison-Wesley, 1992).
H. Bethe and E.E. Salpeter, Quantum Mechanics of one and two-electron atoms
________________________________________________________________________
APPENDIX-1
Commutation relations

0 1 0  i 1 0 
 1    ,  2   ,  3    ,
1 0 i0   0  1

1 0
I 2   .
0 1

Using the Kronecker product, the matrices 1 ,  2 ,  3 , and  are given by

0 0 0 1
 
0 0 1 0   0 1 
1   1   1    
0 1 0 0    1 0 
 
1 0 0 0 


0 0 0  i
 
0 0 i 0   0 2 
2   2   2    
0  i 0 0    2 0 
 
i 0 0 0 
 

139
 0 0 1 0
 
0 0 0  1  0  3 
3   3   3    
1 0 0 0    3 0 
 
0 1 0 0 


1 0 0 0
 
0 1 0 0   I2 0 
   z  I2    
0 0  1 0   0  I 2 
 
0 0 0  1


0 0 0  i
 
0 0 i 0 
 1  i 1   ,
0 i 0 0
 
i 0 0 0 


0 0 0  1
 
0 0 1 0
 2  i 2  
0 1 0 0
 
1 0 0 0 


 0 0  i 0
 
0 0 0 i 
 3  i 3  
i 0 0 0
 
 0  i 0 0
 

I 0 
  4  
0  I 

 0
 k   k 
 0  k 

0 1 0 0 0  i 0 0 
   
1 0 0 0 i 0 0 0 
1   , 2  
0 0 0 1 0 0 0  i
   
0 0 1 0  0 0 i 0 
 

140
 1 0 0 0
 
0 1 0 0
3   ,
0 0 1 0
 
0 0 0  1


 0 0 1 0 
 
 0 0 0  1
5  
1 0 0 0 
 
 0 1 0 0 
 

 0 k 
 k  i 4 k   
 k 0 

 4 k    k 4

{ i ,  j }  2 ij I ,

{ i ,  }  0 , 2  I

 k   k  5   5 k ,

[ , k ]  0

 4

 5   1 2 3 4 ,

 k  i  i  j (i, j, k: cyclic)

[ 5 ,  k ]  0 ,

[ 5 , k ]  0 ,

[ 4 ,  k ]  0

141
 [i ,  j ]  2i k , i  j   j i  i k (i, j, and k; cyclic)

[ 5 k ,  j ]   5 k  j   j 5 k   5 [ k ,  j ] .

{ 5 ,  }  0 .

 1T   1 ,  2T   2 ,  3T   3 ,  4T   4

The helicity operator

Σ  pˆ

_______________________________________________________________
For two arbitrary vectors A and B,

(σˆ  A)(σˆ  B )  ( A  B )1̂  iσˆ  ( A  B )

where

A  ( Ax , Ay , Az ) and B  ( Bx , B y , Bz )

  
 L σ 0 
J  L Σ  2 
2  0 
L  σ 
 2 
______________________________________________________________

((Mathematica))

142
143
144
APPENDIX II
Klein-Gordon equation
((Problem))
The relativistic wave equation for bosons of rest mass m may be obtained by the relation

145
 E 2  p2c 2  m2c 4

through the identifications

 
E  i , p 
t i

(a) Obtain the wave equation relevant to bosons of rest mass m. This equation is
called the Klein-Gordon equation.
(b) What form does this equation assume for photons?
(c) Suppose that the wavefunction is independent of time t. It depends only on r.
Using the spherical co-ordinates; {r , ,  } , find the differential equation for the
er / a
wavefunction  (r ) . Show that  (r ) has the form of  ( r )  A , where A and
r
a are constants. We assume that l = 0.
(d) Find the expression for the characteristic length a.
(e) Use this equation to show that there is a local conservation law of the form


 j  0
t

with


j ( *  * ) .
2mi

Determine the form of  ( r , t ) . From this form for  , give an argument for why
the Klein-Gordon equation is not a good candidate for a one-particle relativistic
wave equation in plane of the Schrodinger equation, for which    *

((Solution))
(a)
We start with

E 2  ( p 2 c 2  m 2 c 4 ) ,

with

 
E  i , p 
t i

Then we have

146
 2
 2    2 c 2  2  m 2 c 4
t 2

or

1 2 m2c2
   2 2  2 
2
(Klein-Gordon equation)
c t 

(b) For photon, the mass m is equal to zero. Then we have the wave equation as

1 2
   2 2 0
2

c t

(c) Suppose that the wavefunction is independent of time t. It depends only on r.

1   m2c2
( r ) ( r )  2 
r r r 

u
We assume that  .
r

d2 m2c2 1
2
u ( r )  2
u( r )  2 u(r ) .
dr  a

Then we have the

u  Ae  r / a

or

e r / a
A
r

(d)

a is the characteristic length and is defined by


a .
mc

(e) The current density is given by

147
 
j ( *  * )
2mi


 j  [  ( * )    ( * )
2mi

 ( *    * 2   *  2 * )
2mi

 ( * 2  2 * )
2mi

Using the equation of continuity, we have

 
   j   ( * 2  2 * )
t 2mi

We use the Klein-Gordon equation

1 2 m2c2 1  2 * m2c2 *
 2    ,  2 *    2 
c 2 t 2 2 c 2 t 2 

Then we get

  1 2 m2c 2 1 2 m2c2
 [ * ( 2 2   2  )  ( 2 2  *  2  * )]
t 2mi c t  c t 
  2
 2
 ( * 2   2  * )
2
2mc i t t
  *  
 (    * )
2mc i t
2
t t

Thus we have

i  
 ( *    * )
2mc 2
t t

Suppose that


*     i
t

where  and  are real. Then we have

148
  *
     i
t

Then we have

i i 
 2
[  i  (  i )]  2
2i   2
2mc 2mc mc

When   0 , the probability density could be negative, which is inconsistent with the
requirement that  should be positive. In this sense, the Klein-Gordon equation is not a
good candidate for a one-particle relativistic wave equation in plane of the Schrodinger
equation, for which    *

149