Vous êtes sur la page 1sur 16

ANGULAR MOMENTUM_1

Classical angular momentum


For a classical particle, the angular momentum is
defined by L
L  r p
p
 Lx i  Ly j  Lz k
r
In components
Lx  ypz  zp y F

Ly  zpx  xpz
Same origin for r and F
Lz  xp y  ypx

Angular momentum is very important in problems dL d dr dp


 (r  p)   p  r ×
involving a central force (one that is always dt dt dt dt
directed towards or away from a central point) p
because in that case it is conserved  (  p)  (r  F)  0.
m
Hermitian operators for quantum
angular momentum
In quantum mechanics we get linear
Hermitian angular momentum
r  rˆ  r, p  pˆ  i 
operators from the classical expressions
L  r p  L ˆ  i r 
using the postulates

ˆ   
Lx  i  y  z 
Lx  ypz  zp y Lˆx  ypˆ z  zpˆ y  z y 
  
Ly  zpx  xpz Lˆ y  zpˆ x  xpˆ z ˆ
L y  i  z  x 
 x z 
Lz  xp y  ypx Lˆz  xpˆ y  ypˆ x   
ˆ
Lz  i  x  y 
 y x 
Commutation relations
 Lˆx , Lˆ y   i Lˆz
The different components of angular momentum
do not commute with one another, e.g.
 
Proof:  Lˆx , Lˆ y   Lˆx Lˆ y  Lˆ y Lˆx
 
Lˆx  ypˆ z  zpˆ y
Lˆ y  zpˆ x  xpˆ z
Lˆz  xpˆ y  ypˆ x
[ x, p x ]  i
[ y, p y ]  i
[ z , pz ]  i

Similar arguments give  Lˆ y , Lˆz   i Lˆx  Lˆz , Lˆx   i Lˆ y


the cyclic permutations    

Summarize these as  Lˆi , Lˆ j   i Lˆk where i, j, k obey a


  cyclic (x, y, z) relation
Commutation relations (2)
The different components of L do not commute with
each another, but they do commute with the squared Lˆ2  Lˆ2x  Lˆ2y  Lˆ2z
magnitude of the angular momentum vector:

Proof:

Similar proofs for the other components

 Lˆx , Lˆ2    Lˆ y , Lˆ2    Lˆz , Lˆ2   0


     
Commutation relations (3)
The different components of angular momentum do not commute
• Lx, Ly and Lz are not compatible observables
• They do not have simultaneous eigenfunctions (except when L = 0)
• We can not have perfect knowledge of any pair at the same time

BUT, the different components all commute with L2


• L2 and each component are compatible observables
• We can find simultaneous eigenfunctions of L2 and one component

CONCLUSION
We can find simultaneous eigenfunctions of one Lˆz   
component of angular momentum and L2 .
Conventionally we chose the z component. Next step is Lˆ2   2
to find these eigenfunctions and study their properties.

What determines the direction of the z-axis?


In an experiment we usually have one or more privileged directions (e.g. the
direction of an external electric or magnetic field) which gives a natural z axis.
If not, this direction is purely arbitrary and no physical consequences depend
on what choice we make.
Angular momentum in
spherical polar coordinates
z
Spherical polar coordinates are the natural coordinate system
in which to describe angular momentum: x, y, z  r , , θ
r y
The angular momentum operators only depend on the angles
θ and φ and not on the radial coordinate r. φ
  cos   
  
Lˆx  i  y  z  Lˆx  i  sin    x
 z y    tan   
     sin    x  r sin  cos 
Lˆ y  i  z  x  Lˆ y  i   cos   
 x z 
  tan   
  
y  r sin  sin 
Lˆz  i  x  y  
 y x  Lˆz  i z  r cos 


 1     1 2 
Lˆ  Lˆ  Lˆ  Lˆ  
2 2 2 2 2
 sin    sin   2 2
  sin   
x y z
 
Note: The angular momentum operators commute with any operator which only depends on r.
L2 is closely related to the angular part of the Laplacian 1   2   L2
(see 2B72 and Section 6).   2 r
2
 2 2
r r  r  r
Lz in spherical polars
x  r sin  cos 
y  r sin  sin 
Proof that 
Lz  i z  r cos 

  
Lˆz  i  x  y 
 x  y  z   y x 
  
  x  y  z

x
x  r sin  cos    r sin  sin    y

y
y  r sin  sin    r sin  cos   x

z
z  r cos   0


   i ˆ
  y  x  Lz
 x y
Eigenfunctions of Lz
Look for simultaneous eigenfunctions of L2 and Lz

Lˆz         
First find the eigenvalues and
eigenfunctions of Lz. Can only Lz  i

depend on the angle φ 
i           

     A exp  i 
2

 d     
2
Normalize 1
solution 0

    A* exp  i  A exp  i   A


2 2

2
1
 d  2 A  1  A
2 2
A
0 2
1
     exp  i 
2
Eigenfunctions of Lz (2)
Boundary condition: wave-function must be single-valued 1
     exp  i 
 (  2 )   ( ) 2

 exp  i   exp  i  exp  2 i 


 exp  2 i   1
   m  0, 1, 2, 3

The angular momentum about the z-axis is Lz  m


quantized in units of hbar (compare Bohr model).
The possible results of a measurement of Lz are m  integer

So the eigenvalue equation and eigenfunction solution for Lz are

1
Lˆz m    m m    m    exp  im 
2
Orthonormality and
completeness
   2     
Lz is a Hermitian operator. Its eigenfunctions are
orthonormal and complete for all functions of the angle
φ that are periodic when φ increases by 2π.

Orthonormality

2 2

 exp  i  n  m  d  
1
0  m   n  d  2
*
mn
0

Completeness  1
 m    exp  im 
    a
m 
m  m   2
2
am   m    d
 *

0
Eigenfunctions of L2
Now look for eigenfunctions of L2 Lˆ2 f  ,     2
f  ,  

ˆ 2  1     1 2 
L  
2
 sin   2 2
 sin      sin   
Try a separated solution of the form
1
f ( ,  )  ( )( )  exp(im )( )
2
(this ensures the solutions remain eigenfunctions of Lz)

Eigenvalue equation is

Lˆ2 exp(im )( )   2


exp(im )( )

We get the equation for Θβm(θ) which depends on both β and m

 1     1 
      m    0
2
 sin    sin
 

m
 
2
 sin 
Eigenfunctions of L2 (2)
Make the substitution  1     1 
      m    0
2
 sin    sin
 

m
 
2
 d    sin 
  cos      sin 
 d  

This gives the Legendre equation, solved in 2B72 by the Frobenius method.

d  d  m      m2 
1          m     0
2

     1    
2

We need solutions that are finite at μ = ±1 (i.e. at θ = 0 and θ = π since μ =


cosθ). This is only possible if β satisfies

  l (l  1) where l  0,1, 2, and l  m

This is like the SHO where we found restrictions on the energy


eigenvalue in order to produce normalizable solutions.
Eigenfunctions of L2 (3)
Label solutions to the Legendre equation by the values of l and m

 m     lm    d  2 d lm    
 m2 
1      l  l  1   lm     0
  l (l  1)     
 1   2
 
For m = 0 the finite solutions are the Legendre polynomials

P0     1
l ,m0     Pl     Pl  cos 
P1     

P2    
1
2
 3 2  1

For non-zero m the solutions are the associated Legendre polynomials


m
lm     Pl m
    Pl  cos 
m  d 
Pl m (  )  (1   2 )  Pl ( )
m /2

 d 
Note that these only depend on the size of m not on its sign
Eigenvalues of L2
So the eigenvalues of L2 for physically allowed solutions are

 2
 l (l  1) 2 , where l  0,1, 2, and l  m

The possible results for the measurement of the squared magnitude of the
angular momentum are L2   2  l (l  1) 2
The possible results for a measurement of the magnitude of the angular
momentum are
L  l (l  1)

From l  m we get  l  m  l
For each l there are 2l+1 possible integer values of m

The restriction on the possible values of m is reasonable. The z-component


of angular momentum can not be greater than the total!
In fact, unless l = 0, the z-component is always less than the total and can
never be equal to it. Why?
Summary
f ( ,  )  Pl m  cos  exp  im 
The simultaneous eigenfunctions
of Lz and L2 are

Eigenvalues of Lˆ2 are l (l  1) 2 , with l  0,1, 2,

The integer l is known as the principal angular momentum quantum


number. It determines the magnitude of the angular momentum

Eigenvalues of Lˆz are m , with  l  m  l


(i.e. m  l , l  1 , 1, 0,1, l  1, l )

The integer m is known as the magnetic quantum number. It


determines the z-component of angular momentum. For each
value of l there are 2l+1 possible values of m.

The simultaneous eigenfunctions of L2 and Lz do not correspond to definite values


of Lx and Ly, because these operators do not commute with Lz. We can show,
however, that the expectation value of Lx and Ly is zero for the functions f(θ,φ).

Vous aimerez peut-être aussi