You are on page 1of 42

Chapter 6

The Approximation Methods

When quantum mechanics is applied to realistic physical systems, we


encounter a potential energy that doesn’t have a simple form. Then, the
solution of the Schrodinger equation is usually a complicated mathematical
problem that requires the use of approximation methods.
In this chapter we will consider two main approximation methods:
1) The perturbation theory
2) The variational Theory
The Perturbation Theory
The Hamiltonian we shall consider is in the form

H  H o  gV (1)

with Ho is the unperturbed Hamiltonian with well-known eigenfunctions and


eigenvalues. gV is the perturbation with g is a small parameter. The main
mathematical ideas of the perturbation theory rests on two arguments:

(i) A small changes in the parameters of a linear differential equation produce


small changes in its solution. It is known that the Schrödinger equation is linear
differential equation.

(ii) Any vector (or any function) can be expanded in terms of a basis vectors (or
a basis functions). The eigenfunctions of a physical Hamiltonian can be
considered as basis functions since they are normalized functions.
Time Independent Non-degenrate Perturbation Theory:
Let H o no   no no (2)

and H n   n n (3)

Substituting for H from Eq. (1), Eq. (3) becomes

H o  gV  n   n n (4)

g 0
with n  no (5)

It will be assumed that

n  n0  g n1 g 2 n2   (6)

and  n   n0  g 1n  g 2 n2   (7 )

where ni and  ni represent the ith order correction to the eigenvectors and
eigenvalues, respectively. Now, substituting Eqs. (6) and (7) into Eq. (4) 

H o  gV  n0  g n1 g 2 n2    
 n0  g 1n  g 2 n2  n0  g n1 g 2 n2   (8)

Collecting terms of like power of g 

H o n0   n0 n0  g H o n1  V no  n0 n1   1n no 


g 2 H o n2  V n1  no n2  1n n1   n2 n1    0 (9)
Since Eq.(9) is to be valid for any g  every coefficient of g must equal to zero 

H o n0   n0 n0 (10)

H o n1  V no  n0 n1   1n no (11)

H o n2  V n1  n0 n2   1n n1   n2 no (12)


First Order Correction:
Using the 2nd argument we expand 1 in terms of the basis o, i.e.,

n1   a1nj  oj n  1,2, (13)


j 1

Substituting Eq. (13) into Eq. (11) we get

H o  a1nj  oj  V no  n0  a1nj  oj   1n no (14)


j 1 j 1

Now from Eq.(10), H o  0j   0j  0j then Eq. (14) reads

  0j   n0 a1nj  oj   1n  V  no (15)


j

Multiplying Eq. (15) by ko we get

  0j   n0 a1nj ko  oj  ko  1n  V  no 


j

  0j   n0 a1nj  kj   1n kn  ko V no (16)


j
Applying the properties of the Kronekel Delta we obtain

 k0   n0 a1nk   1n kn  ko V no (17)

To find the value of n1 we set k=n 

 1n  no V no  Vnn  V n


(18)

from Eq. (7) to first order we get

 n   n0  gVnn (19)

To find the value of ank we set k≠n in Eq.(17) 

 k0   n0 a1nk   ko V no  Vkn (20)


From Which we get
V
a1nk  0 kn 0 (21)
n k
From Eqs. (6) and (13) to first order we get

n  n0  g  a1nj  oj  n0  ga1nn n0  g  a1nj  oj (22)


j 1 j n

Now substituting for anj1 from Eq.(21) we obtain

n  
1  ga1nn  n0 g 
V jn
 o
j (23)

j n n
0
  0
j

Multiplying the above equation by its complex conjugate and dropping terms of
order g2 
1 2
n n  1  gann no no  g 2 () 

1 2
1 1  gann  1  2 ga1nn  Re a1nn  0 (24)

Now Eq.(23) becomes


V jn
 n  n0 g   o
j (25)

j n n
0
  0
j

The wave-mechanical analogue of Eq. (25) is

V jn
n  no g   o
j (26)
j n  n  j
0 0

with V jn    j V n dv (27)

Example: Find the first order correction for the eigenvalues and the
eigenfunctions of the Hamiltonian
p2 1
H  H o  gx 2 with H o   2  2 x 2
2

Solution: To find the 1st order correction in the energy we have, from Eq.(18)
we have
 1n  no V no  n x 2 n

But x 

2 
aa  †
 x 
2 
2
aa  aa †  a † a  a † a †  

Vnn 

2
 
n aa  aa †  a † a  a † a † n

Knowing that

n aa n  n n a n  1  nn  1 n n  2  0

and n a†a† n  n  1 n a† n  1  n  1n  2 n n  2  0

and n a†a n  n n a† n  1  n n n  n

and n aa† n  n  1 n a n  1  n  1 n n  n  1

 Vnn 

2n  1   n  12 
2 
From Eq.(19) we have

 n   n0  gVnn 

   
 n           
g 1  g   g 
  2    n 1   2 
0
n  12 n  12 n  12
   

p2 1
Returning to the Hamiltonian H  H o  gx   2 k  2 g  x 2
2
2
Which has an exact solution given by  n    n  1 ,
2
   
k  2g

1
k  2g  2  g 
or    1     1     
 k   k 

0  
 n    
 g  g
n  12 1       n 1   
 k   
2

Which is the same as the perturbed result for first order of g.
To find the correction in the eigen-function, we have from Eq.(25)

V jn
n  n0 g   o
j

j n n
0
  0
j

Knowing that V jn 

2

j aa  aa†  a†a  a†a† n 
but j a†a† n  n  1 j a† n  1  n  1n  2 j n  2  n  1n  2 jn2

and j aa n  n j a n  1  nn  1 j n  2  nn  1 jn  2

and j aa† n  n  1 j a n  1  n  1 j n  n  1 jn

and j a†a n  n j a† n  1  n j n  n jn

 V jn 

2
 nn  1 jn2  n  1 jn  n jn  n  1n  2 jn2 
V jn
  oZero except for j=n-2
j 
Zero except for j=n+2
j n  n   j
0 0

  nn  1 jn 2  2n  1 jn  n  1n  2 jn 2 


  oj
2  j  n  n0   0j
Zero for every j

  nn  1 n  1n  2 
 n  n0 g  0 0 n2  n2 
2   n   n2  n0   n0 2 


but  n0   n0 2   n  12  n  2  12  2 

and  n0   n0 2   n  12  n  2  12  2

 n  n0 
g
 nn  1 n  2  n  1n  2 n  2 
4 2
Second Order Correction:
To find the second order correction we have to solve Eq.(12) which is:

H o n2  V n1  n0 n2   1n n1   n2 no (12)

Make the expansion

n2   anj
2
 oj n  1,2, (28)
j 1

Substituting in Eq. (28) into Eq. (12) 

H o  anj
2
 oj  V n1  n0  anj
2
 oj   1n n1   n2 no (29)
j 1 j 1

Multiplying Eq. (29) by no 

 anj2 no H o  oj  no V n1  n0  anj2 no  oj   1n no n1   n2 no n0
j 1 j 1

From Eq. (10) we have H o  0j   0j  0j 


The last eq. becomes

 anj2  0j no  oj  no V n1  n0  anj2 no  oj   1n no n1   n2 no n0
j 1 j 1
Using n  j   nj (30) 

 anj2  0j  nj  no V n1  n0  anj2  nj   1n no n1   n2 nn 


j 1 j 1

2 0
ann  n  no V n1   n0 ann
2
  1n no n1   n2 

no V n1  1n no n1   n2 (31)


V jn
But from Eq.(25) n1 is given by n1    o
j

jn n
0
  0
j
Then Eq.(32) now becomes  nj
V jn V jn
 0 0 no V  oj  1n  n
o
 o
j   2
n (32)
j n  n   j j n  n   j
0 0
Using identity (30) the summation of the R.H.S vanish 
2
V jn V jn Vnj V jn
 n2   0 0 no V  oj   0 0  0 0 (33)
j n  n   j j n  n   j j n  n   j

And for the energy we get from Eq.(7)

and  n   n0  g 1n  g 2 n2

with  1n &  n2 are to be evaluated from Eqs. (18) and (33)


.
For the second order correction of the wave-function, the result is
straightforward but a very long and difficult identity given by

1 
 VknV jk VnnV jn 1 V jnVnj   o
n2   0 0 0 0  0  0 j

j n  n   j 
k  n  n   k  n   j  n   j 
0 2 0
Example: Find the eigenvalues of the Hamiltonian

H
p2 1
2

 2  2 x 2  2bx 
(a) By exact solution (b) By perturbation method
Solution: (a) Let us rewrite the Hamiltonian as

p2 1
H  2  2  x  b 2  12  2b 2  H o  12  2b 2
2
p2 1
with H o   2  2  x  b 2
2

Is the Hamiltonian of a harmonic oscillator with x   x  b 

 n0   n  12 
and n0 Is the same as that of a harmonic oscillator but with x  x  b

Now H n   n n 

H o  12  2b 2  n  n n 

H o n  12  2b 2 n   n n 

but H o n   n0 n   n  12 n   

 n0  12  2b 2  n  n n 

 
 n   n0  12  2b 2   n  12   12  2b 2
(b) Let us rewrite the Hamiltonian as

p2 1
H  2  2 x 2   2bx  H o  gx with g    2b
2
From Eq.(18) we have with V  x 

2
a  a†  
 1n  Vnn  no V no  n x n 
2 


n a  a† n 


 n n n 1  n  1 n n  1  0
2
From Eq.(18), for the 2nd order correction, we have
2
 
V jn 
 n2   0 0 with V jn  j x n  j a  a† n 
j n  n   j 2 

V jn 

 n j n 1  n  1 j n  1   
 n  j ,n1  n  1 j ,n1
2 2
Zero except for j=n-1 Zero except for j=n+1

 n jn 1  n  1 jn1
  n2    
2  j n  n   j
0 0 2  j  n  n0   0j

  n n  1 
  n2   0 0  0 0 
2   n   n 1  n   n 1 


but  n0   n01   n  12  n  1  12   

and  n0   n01   n  12  n  1  12   
  n n  1 1
  n2      
2      2  2
 2 4 2
0 1 2 2   0
 n   n  g n  g  n   n
  b 
2 
  n  12  12  2b 2  
 2  
Time-Independent Perturbation Theory for Degenerate Case:
Let us go back to Eq.(6) where we have assumed

n  n0  g n1 g 2 n2  

That is, we have assumed that the perturbed eigen-state n differ slightly
o
from the unperturbed eigen-state n

When the unperturbed energy  no is degenerate, there are several eigenstates


(say d-states) corresponding to this eigenvalues, nk
o
k  1,2,3,, d  , and
we don’t know to which state n will tend as g  0 .
o
For non-degenerate case we have expanded n1 in terms of n and obtain
V jn
n   0
1
 o
j

jn n   0
j

Now for degenerate case the denominator is zero for jd and so the 1st order
correction to the wave-function blows up to infinity.
To avoid these problems we have to construct a new basis vectors such that
the numerator is zero.
This new basis vectors (the selected states) is written as a linear combination of
the old unperturbed degenerate states corresponding to the same eigen value,
i.e.,

d
 no   bnk ko n  1,2,  d (34)
k

We will assume that the perturbation removes the degeneracy so that n must
be unique and will be approximately equal to this selected state  n , i.e.,
o

g 0
n   no (35)

Since the linear combination of the degenerate eigenstates is also a


solution of the unperturbed Hamiltonian with the same eigenvalue, the first
o
order shift in the energy will be given by, from Eq.(19) with n is replaced
by  no
 n   n0  g  no V  no (36)

We will consider the perturbation of an d-fold degeneracy and set  no   o

It is clear from Eq.(34) that the selected state can be one of the unperturbed or
a linear combination of more than one unperturbed state depending on the
values of bnk. Our task in the problem then is to find the values of bnk and also
find the 1st order correction to the eigenvalues.

Multiplying Eq.(34) by its complex conjugate 


d d

 no  no   bnj  bnk  oj ko  1 
j k

 bnj bnk jk 1 
j ,k
2
 bnk 1 (37)
k
 o
Now rewrite Eq. (11) for n we get
H o n1  V  no  n0 n1   1n  no (38)
d
Using (13), and (36), we have n1   a1nj  oj &  no   bnk
1
no 
j 1 k 1
d d
 a1nj H o  oj   bnk V ko  n0  a1nj  oj   n  bnk
1
ko (39)
j 1 k j 1 k

But from Eq.(10) we have H o n0   n0 n0 


d d
 a1nj  0j  oj   bnk V ko  n0  a1nj  oj   n  bnk
1 ko (39)
j 1 k j 1 k
d
 a1nj  0j  oj   anj j  j   bnkV ko
1 0 o
j d j d k
d
  n0  a1nj  oj  n0  anj  j   n  bnk ko
1 o 1
j d j d k

since  no   oj   o for n  d  Eq.(39) becomes


d d
 a1nj 0j  oj   bnkV ko  n0  a1nj  oj   1n  bnk ko (40)
j d k 1 j d k 1

Now, multiplying Eq. 40) by mo m  1,2,3,, d  


d d
 a1nj  0j mo  oj   bnk mo V ko  n0  a1nj mo  oj   1n  bnk mo ko (41)
j d k 1 j d k 1

 mj  mj  mk

d d
 a1nj  0j  mj   bnk mo V ko  n0  a1nj  mj   1n  bnk  mk (41)
j d k 1 j d k 1

Because of the Kronekkel delta the first sum of both sides in Eq.(41) will vanish
and we have
d
 bnk mo V ko   1nbnm
k 1

d
or  Vmk bnk   1n bnm (42)
k 1

or in matrix form

 V11  V1d  bn1   bn1 


   1 
        n    (43)
V   b  b 
 d1 Vdd  nd   nd 
Example: Find the eigenvalues of the H-atom with the application of a uniform
electric field directed along the z-axis (the Stark effect).
 
Solution: It is known that U  pE
  
with p  er and E  Ezˆ  U  eEz
the perturbation with g=eE and
H  H o  eEz  H o  eE r cos V  z  r cos 

For the ground state (n=1) we have no degeneracy (d=1)

V  V11  100 z 100   100 r cos 100 d 3 x

1 r 1  2 
3 2r a
but 100  e a  V  V11   sin  cos  d  d  r e dr  0
a 3
a 3
0 0 0
11  0
For the first excited state (n=2) we have 4-fold degeneracy (d=4), namely

200 , 210 , 211 , 211 


 200 z 200 200 z 210 200 z 211 200 z 211 
 
 
 210 z 200 210 z 210 210 z 211 210 z 211 
V  
 211 z 200 211 z 210 211 z 211 211 z 211 
 
 
 211 z 200 211 z 210 211 z 211 211 z 211 

1  r   r 2a r  r 2a
1
200  14 3 
2  e 210  14 e cos
2a a 2a 3a

1 r  r 2a i 1 r  r 2a i
211  18 e sin  e 211  8
1 e sin  e
a3 a a3 a
since z  r cos  is independent of the angle , and noting that

 Ylm 'Ylm d  N mm'  all terms with different m vanish.
2 2
i
Because you will encounter the integral e d   e  2i d  0 
0 0
 200 z 200 200 z 210 0 0 
 
 
 210 z 200 210 z 210 0 0 
V  
 0 0 211 z 211 0 
 
 
 211 z 211 
 0 0 0

1  r   r 2a 1 r  r 2a
200  14 3 
2  e 210  14 e cos
2a a 2a 3a

1 r  r
2a sin  ei
1 r  r 2a i
211  18 e 211  18 e sin  e
a3 a a3 a

1  2  3 2
 r  r r a
V11  200 z 200  16
1
 cos  sin  d  d    2   e dr  0
2a 0 0 a a
3
0

zero
1  2  2 3
 r  r r a
V22  210 z 210  16
1
 cos 3
 sin  d   d     e dr  0
2a 0 0 a a
3
0
zero
 2  2 3
1  r  r r a
V33  V44  211 z 211  16
1
3
cos  sin 3
 d 
   a  a e dr  0
d
a 0 0 0 
zero
 200 z 200 200 z 210 0 0 
 
 
 210 z 200 210 z 210 0 0 
V  
 0 0 211 z 211 0 
 
 
 211 z 211 
 0 0 0

0 200 z 210
 V
210 z 200 0
V12  V21  200 z 210  210 z 200
1  2
 r
 r 4
r a
 16
1
3
sin  cos  d2
 
d  2   e dr
2a 0 0 0 a a

2/3 2π

1 1
 2  4 r a 1  5 r a 
V12  V21  8 3a 3  r e dr  2  r e dr 
a 0 a 0 

4!a 5 5!a 6

n!
But  x n eax dx  n1 , a 0, n is  ve integer
0 a
0  3a
 V12  V21  3a  V
 3a 0
Now from Eq.(43) we obtain

 0  3a  b1   b1   0  3a    21 0 
     1      
   2        0
  3a 0  b2   b2    3a 0   0  21 

   12  3a 
 
   0  1 2
 
 2  9a 2   12  3a 
  3a   1 
 2
 2   20  g 12   20  3aeE

 0  3a  b1   b1 
for  21  3a      3a  
    
  3a 0  b2   b2 
  3ab2   3ab1   3ab2  3ab1   b2  b1 
   
  3ab1   3ab2 
1 1 1 1
b  N    b
2
 1  N 2 1  1   1  b 
  
  1
 
  1
  2   1

 0  3a  b1   b1 
for  21  3a      3a  
    
  3a 0  b2   b2 
  3ab2    3ab1   3ab2  3ab1  b2  b1 
   
  3ab1    3ab2 
 1 1  1 
b  N    
 1 2 1
 
d
Now from Eq.(36)  no   bnk ko n  1,2, d we obtain
k

 2 
1
 200  210  and  2 
1
 200  210 
2 2
VRIATIONAL METHOD:
The variational method is the other main approximate method used in quantum
mechanics. Compared to perturbation theory, the variational method can be
more efficient in situations when there is no small dimensionless parameter in
the problem, or the system is not exactly solvable when the small parameter is
sent to zero.
The basic idea of the variational method is to guess a ``trial'' wave-function for
the problem, which consists of some adjustable parameters called ``variational
parameters.'' These parameters are adjusted until the energy of the trial wave-
function is minimized.
Why would it make sense that the best approximate trial wave-function is the
one with the lowest energy? This results from the Variational Theorem, which
states that the energy E of any trial wave-function is always an upper bound to
the exact ground state energy. This can be proven easily by the following
Lemma.
Lemma: Let  be an arbitrary normalized ket and 1 be the ground state
energy of the Hamiltonian H. Then the following inequality always holds:

 H   1 (44)
Proof: Write  as a linear combination of the different eigenfunctions, i.e.,
   ai i with H i   i i (45)
i
Now since  is normalized then we have
    aj ai  j i 1 
i
 ai
2
1  ij (46)
i

Using Eq.(45), the L.H.S of Eq.(44) becomes

   
 H     i ai  H  a j  j    aia j i H  j


i   j  i, j

H i   i i
 H    aia j  j i  j 
i, j
 ij
 H    ai a j  j  ij   ai  i  a1 1   ai  i
2 2 2
i, j i i 1

But   i  1    i  1 
i 1 i 1

 H   a1 1  1  ai   1  ai
2 2 2
(47)
i 1 i 1

Using Eq.(46) we get

 H   1

So the Lemma is proved.


The method is applied by choosing a reasonable trial wave-function which
depends on a set of N-parameters 1, 2 ,   N, i.e.,

trial  trial 1, 2 ,N  

trial dv  I 1 , 2 ,   N 
 H
 H    trial (48)
Now from the last lemma, Eq.(44), we have
I 1 , 2 ,   N   1 

To find the ground state energy we have to minimize the energy function
I 1, 2 ,N  with respect to the parameters , i.e.,

I
0 

 var  I min (49)

The success of this method still depends on having a good trial wavefunction
with the good set of parameters. This is what is called the variational method:
you vary the parameters within the trial wavefunction and find the parameter set
that minimizes the energy expectation value. If the trial wavefunction is close
enough to the true ground state, so is the minimum energy expectation value. If
the trial wavefunction has the correct functional form, it leads to the exact
ground state.
Example: Let us solve the H-atom by a adopting a trial function of the form

1 r
trial  e 
3
32 r
1 1 
Solution: It is known that the exact wave function is 100    e ao
  ao 

  p 2 e 2 
I     H  
  trial Htrial dv   trial   trial dv

 2 r 

1  2  1     1 2
knowing that   2  r  2  sin   2
2
r r  r  r sin      r sin 2   2

then
p2
2
2 2
trial    trial 
2
1  2 d 2 r 
 2r dr
3 2
r e  
p2 1  2  2 r  1 r  
or trial   e  e 
2 3
2   r  
 4    2 2 e2  2r  2
 I     3      e r dr
   0  r 2 2
r 


  2  2r 
 
 2  2r  2
  

 4  2r  

 I     3   e r dr   e r dr  e 2
 e r dr 
  
  0 2  0
2
0 

 22 2 2 3  22



n!
using again  x n e  ax dx  n 1
, a 0, n is  ve integer
0 a

 4   2   2  e 2 2  2 e2
 I     3     

   4  8 
4  2  2 

I      2 e 2   2 e2 2
    3  2 0    ao 
 
  2  2   
d  e 2

2 e2  e4
  I min   2  2
    Which is the exact result.
2  2  e 2
2
 e 2
2
 r 2
Let us now try the trial wave-function. trial  N e

To find the normalization constant we have

   trial r sin dddr  1


 2 
 2 r 2
4   e
2 2
 N 2
r 2 dr  1 
00 0 0
 n  1
    
n  ax 2
using  x e dx   2 
n 1 4 2 3 2
0 2a  2
34
 2 
N  2 
2 32
1 N   
 
p2 2 2  2 1   2 trial 
then trial    trial   r  
2 2 2  r r 
2 r 

p2  2  d  3  r 2 
then trial  N 2  r e  
2 r dr  
p2  2   2  r 2 4  r 2 
trial  N 2  3r e  2 r e 
2 r  
 2
3 2
 2  
2 2
e 
 I    4 N    2
2
 r  e
2 2 r
r 2 dr
 r 
0  

2
3 2
 2  2 r 2 2 2  2  4  2  r 2 
 2 r 2  
 I    4N  r e dr   r e dr  e 2
 re dr 
 
 0  0 0 

 3  1
4 2 3 2 8 2 5 2 2 2 
 n  1
   
n  ax 2
using  x e dx   2 
n 1
0 2a  2
3 2
e2   2   e2 
32
 2   3 2  3 2    3 2  
 I    4        4     
  
 8  2  16  2  4  
    
16  2  4  

3 2  2
 I     2e 2
2 

I   3 2 2 2 e 2 2
  e2 0  
 2  3 2 

3 2  8 2 e 4  4 e 2
e 4
 8 16 
 I min   2 e 2
   

2   9  
4  3 2
2  3 3 
2

8  e4
  I min    0.85 1
3 2 2

which is a good answer within 15%.

You might also like