Homework 2.

6
Quantum Mechanics II

Pavel Elian Morales Medina

May 28th , 2021

Problem 11.3
Solve Equation 11.17 for the case of a time-independent perturbation, assuming that ca (0) =
1 and cb (0) = 0. Check that |ca (t)|2 + |cb (t)|2 = 1. Comment: Ostensibly, this system
oscillates between “pure ψa ” and “some ψb ”. Doesn‘t this contradict my general assertion
that no transitions occur for time-independent perturbations? No, but the reason is rahter
subtle: In this case ψa and ψb are not, and never were, eigenstates of the Hamiltonian–a
measurement of the energy never yields Ea or Eb . In time-dependent perturbation theory
we typically contemplate turning on the perturbation for a while, and then turning it off
again, in order to examine the system. At the beginning, and at the end, ψa and ψb are
eigenstates of the exact Hamiltonian, and only in this context does it make sense to say that
the system underwent a transition from one to the other. For the present problem, then,
assume that the perturbation was turned on at time t = 0, and off again at time T – this
doesn´t affect the calculations, but it allows for a more sensible interpretation of the result.
Answer:
We start by recalling Equation 11.17 to solve
i 0 −iω0 t i 0 iωo t
c˙a = − Hab e cb , c˙b = − Hba e ca (1)
~ ~
Now, we will use substitution, to get rid of ca , obtain a single differential equation for cb . To
do this, first we solve for ca from the second equation, remembering that the perturbation
is independent of time,
i~ −iω0 t
ca = 0
e c˙b (2)
Hba
i~
⇒ c˙a = 0 [c¨b − iω0 c˙b ] e−iω0 t
Hba

1
Therefore, equalling with the first equation
i 0 −iω i~
− Hab e−iω
e  0 t cb = 0 [c¨b − iω0 c˙b ]  0t
  
~ Hba
1 0 ~
Hab cb = − 0 [c¨b − iω0 c˙b ]
~ Hba
~2
cb = − 0 2 [c¨b − iω0 c˙b ]
|Hab |
~2
cb = − 0 2 [c¨b − iω0 c˙b ]
|Hab |
" #
0 2
~2 |Hab |
0 = − 0 2 c¨b − iω0 c˙b + cb
|Hab | ~2
0 2
|Hab |
0 = c¨b − iω0 c˙b + 2
cb
~
2
|Hab
0
|
Setting g12 = ~2
, equation above becomes

c¨b − iω0 c˙b + g12 cb = 0

To solve the differential equation, se use the following ansätz

cb = eqt
⇒ c˙b = qeqt
⇒ c¨b = q 2 eqt (3)

Replacing this expressions into the differential equation we get

q2qt
e − iω0 q qt
e + g12 qt
e =0
q 2 − iω0 q + g12 = 0

Where we have obtained a second degree polynomial in q, which roots can be obtained using
the quadratic form
q
iω0 ± (iω0 )2 − 4g12
p
iω0 ± i ω02 + 4g12
q± = =
2 2
Hence, the full solution to the homogeneous differential equation, is the linear combination
of our ansätz, using q+ and q− , as we would expect, since it was a second degree problem.

cb (t) = Aeq+ t + Beq− t

√ 2 2
! √2 +4g 2
!
iω0 +i ω0 +4g1 iω0 −i ω0 1
2
t 2
t
= Ae + Be
" √ 2 +4g 2
√ 2 +4g 2
#
iω0 + ω0 1 − ω0 1
t it it
=e 2 Ae 2 + Be 2

2
By rearranging the constants, we know that the solution above, can be rewritten as
" p ! p !#
2 2 2 2
iω0 ω0 + 4g1 ω0 + 4g1
cb (t) = e 2 t A0 sin t + B 0 cos t
2 2

Using the boundary condition given,

cb (0) = 0
1 h
:0 :1
i
0
e + A0 
sin(0)
 + B0
cos
(0)
 =0
⇒ B0 = 0

Consequently
p !
2 2
iω0 ω0 + 4g1
cb (t) = e 2
t
A0 sin t (4)
2

Next, to compute, ca (t) we just simply substitute the expression above, into, previously
obtained, equation 2,
" p !#
2 2
i~ −iω0 t ∂ iω0 ω + 4g
ca (t) = 0 e e 2 t A0 sin 0 1
t
Hba ∂t 2
" p ! p p !#
i~ −iω0 t 0 iω0 ω02 + 4g12 ω02 + 4g12 ω02 + 4g12 iω0
= 0 e A sin t + cos t e 2 t
Hba 2 2 2 2
" p ! p p !#
i~ − iω0 t 0 iω0 ω02 + 4g12 ω02 + 4g12 ω02 + 4g12
= 0 e 2 A sin t + cos t
Hba 2 2 2 2
" p ! p !#
~A0 − iω0 t ω02 + 4g12 ω02 + 4g12
q
= 0
e 2 −ω 0 sin t + i ω02 + 4g12 cos t
2Hba 2 2

Then, using the boundary condition given, as well,

ca (0) = 1
~A0  1 1

0
q
0
e0 −ω0   :
 2 2
sin(0) + i ω0 + 4g1 cos
(0)
 :

=1
2Hba
~A0
q 
2 2
0
i ω0 + 4g1 = 1
2Hba
0
0 −2iHba
A = p
~ ω02 + 4g12

3
Substituting A0 into ca (t) above
" p ! p !#
0 2 2 2 2
−2iHba ω + 4g ω + 4g
q
~ iω0
ca (t) = 0
p e− 2 t −ω0 sin 0 1
t + i ω02 + 4g12 cos 0 1
t
2Hba ~ ω0 + 4g1
2 2 2 2
" p ! p !#
2 2 2 2
−i ω + 4g ω + 4g
iω0
q
=p 2 e− 2 t −ω0 sin 0 1
t + i ω02 + 4g12 cos 0 1
t
ω0 + 4g12 2 2
" p ! p !#
2 2 2 2
iω0 iω0 ω + 4g ω + 4g
= e− 2 t p 2 sin 0 1
t + cos 0 1
t
ω0 + 4g1 2 2 2
Doing the same for cb (t) from equation 4 we get
p !
−2iH 0 iω0 ω02 + 4g12
cb (t) = p 2 ba 2 e 2 t sin t
~ ω0 + 4g1 2
Collecting both results we have found
" p ! p !#
iω0 iω0 ω02 + 4g12 ω02 + 4g12
ca (t) = e− 2 t p 2 sin t + cos t (5)
ω0 + 4g12 2 2
!
0 2
p
0 2 2
−2iHba iω0
t ω0 + 4g1 |Hab |
cb (t) = p e 2 sin t where g12 = 2
(6)
2 2
~ ω0 + 4g1 2 ~
Finally with these expressions we proceed to verify
 p ! p !2
2 2
 iω0 ω02 + 4g12 ω02 + 4g12 
|ca (t)| + |cb (t)| =  p 2 sin t + cos t 

 ω0 + 4g12 2 2 
 p ! 2
 −2iH 0 ω02 + 4g12 
ba
+ p 2 sin t 

 ~ ω0 + 4g12 2 
p ! p !
ω02 ω0
2
+ 4g1
2
ω 2
0 + 4g 2
1
= 2 sin2 t + cos2 t
ω0 + 4g12 2 2
0 2
4 |Hab |
+ 2 2
~ (ω0 + 4g12 )
Notice, that in the last term we can substitute the definition of g1 we simplify
p ! p !
2 2 2 2 2 2
 
ω 4g ω + 4g ω + 4g
|ca (t)|2 + |cb (t)|2 = 2
0
2
+ 2 1 2 sin2 0 1
t + cos2 0 1
t
ω0 + 4g1 ω0 + 4g1 2 2
*1
 2 p ! p !
12 2 2 2 2

ω + 4g ω + 4g

ω0 +4g 0 1 0 1
= 2 2
sin2 t + cos2 t
ω

 0 + 4g1 2 2
p ! p !
2 2 2 2
ω 0 + 4g1 ω 0 + 4g1
= sin2 t + cos2 t
2 2

4
 ⇒ |ca (t)|2 + |cb (t)|2 = 1 (7)
As desired to verify.

Problem 11.9
The first term in Equation 11.32 comes from thee eiωt /2 part of the cos (ωt), and the second
from e−iωt /2. Thus dropping the first term is formally equivalent to writing Ĥ 0 = (V /2) e−iωt ,
which is to say,
0 Vba −iωt 0 Vab iωt
Hba = e , Hab = e
2 2
(The latter is required to make the Hamiltonian matrix hermitian–or, if you prefer, to pick
out the dominant term in the formula analogous to Equation 11.32 for ca (t).) Rabi noticed
that if you make this so-called rotating wave approximation at the beginning of the
calculation, Equation 11.17 can be solved exactly, with no need for perturbation theory, and
no assumption about the strength of the field.

(a) Solve Equation 11.17 in the rotating wave approximation (Equation 11.36), for the usual
initial conditions: ca (0) = 1, cb (0) = 0. Express your results (ca (t) and cb (t)) in terms of
the Rabi flopping frequency,
1
q
ωr = (w − ω0 )2 + (|Vab | /~)2 .
2
Answer: Writing down Equation 11.32 for the given perturbation,
i Vab iωt −iω0 t i Vba −iωt iωo t
c˙a = − e e cb , c˙b = − e e ca (8)
~ 2 ~ 2
Defining ∆ω = ω − ω0 , the system of differential equations becomes
iVab i∆ωt iVba −i∆ωt
c˙a = − e cb , c˙b = − e ca
2~ 2~
Then, to solve it, we proceed analogously as in the previous problem, first we solve for ca ,
from the equation on the right, differentiate it, and combine it with the expression on the
left, to obtain a single second order differential equation for cb . The process is shown below
2i~ i∆ωt
ca = e c˙b (9)
Vba
2i~
⇒ c˙a = [c¨b + i∆ω c˙b ] ei∆ωt
Vba
Equalling both equations for c˙a we get
iVab i∆ωt 2i~
− ei∆ωt
 
e  cb =
 [c¨b + i∆ω c˙b ] 
2~ Vba
4~2
cb = − [c¨b + i∆ω c˙b ]
|Vab |2
|Vab |2
⇒ 0 = c¨b + i∆ω c˙b + cb
4~2

5
 |Vab |2
Setting g22 = ~2
and using the ansätz cb (t) = eqt as in equation 3, we get the second degree
polynomial

2 g22
q + i∆ωq + =0
4

r  2
g2
−i∆ω ± (i∆ω)2 − 4 4 i∆ω i
q
⇒ q± = =− ± ∆ω 2 + g22
2 2 2
Therefore, as in the previous problem, the solution corresponds to a linear combination of
the ansätz using both q+ and q− , so we can write

cb (t) = Ceq+ t + Deq− t

 √   √ 
− i∆ω + 2i ∆ω 2 +g22 t − i∆ω − 2i ∆ω 2 +g22 t
⇒ = Ce 2
+ De 2

i∆ω
h i
√ 2 2 i
√ 2 2i
= e− 2 t Ce+ 2 ∆ω +g2 t + De− 2 ∆ω +g2 t

As we know, we can arrange the constants such that, the expression above becomes a linear
combination of sine and cosine.
  q   q 
− i∆ω t 0 1 2 2 0 1 2 2
⇒ cb (t) = e 2 C sin ∆ω + g2 t + D cos ∆ω + g2 t
2 2

Observe that using the definition of g22 and ∆ω into the argument of the trigonometric
functions, we can express it in terms of the Rabi flopping frequency
s
1 1 |Vab |2
q
∆ω 2 + g22 t = (ω − ω0)2 + t = ωr t
2 2 ~2
| {z }
ωr

So, that the equation above can be rewritten as

i∆ω
cb (t) = e− 2
t
[C 0 sin (ωr t) + D0 cos (ωr t)]

Next, imposing the boundary conditions

cb (0) = 0
⇒ e0 [C 0 sin (0) + D0 cos (0)] = 0
D0 = 0

Subsequently
i∆ω
cb (t) = e− 2
t
C 0 sin (ωr t) (10)

6
Using this result into equation 9 we can solve for ca (t) as follows
2i~ i∆ωt
ca (t) = e c˙b
Vba
2i~ i∆ωt ∂ h − i∆ω t 0 i
= e e 2 C sin (ωr t)
Vba ∂t
2i~C 0 i∆ωt
 
i∆ω i∆ω
= e − sin (ωr t) + ωr cos (ωr t) e− 2 t
Vba 2
0
 
2i~C i∆ω t i∆ω
= e 2 − sin (ωr t) + ωr cos (ωr t)
Vba 2
Now, we use the boundary conditions, to determine C’, recall

ca (0) = 1
0
 
2i~C 0 i∆ω
⇒ e − sin (0) + ωr cos (0) = 1
Vba 2
2i~ωr 0
C =1
Vba
Vba
C0 =
2i~ωr
Substituting into ca (t) we obtain
 
2i~ Vba i∆ω t i∆ω
ca (t) = e 2 − sin (ωr t) + ωr cos (ωr t)
Vba 2i~ωr 2
 
1 i∆ω t i∆ω
= e 2 − sin (ωr t) + ωr cos (ωr t)
ωr 2
 
i∆ω
t i∆ω
=e 2 − sin (ωr t) + cos (ωr t)
2ωr
Substituting C’ for cb (t) from expression 10 we get
i∆ω Vba
cb (t) = e− 2
t
sin (ωr t)
2i~ωr
−iVba − i∆ω t
= e 2 sin (ωr t)
2~ωr
Collecting our results
 
i∆ω
t i∆ω
ca (t) = e 2 − sin (ωr t) + cos (ωr t) (11)
2ωr

−iVba − i∆ω t
cb (t) = e 2 sin (ωr t) (12)
2~ωr
q
where ∆ω = ω − ω0 and ωr = 1
2
(w − ω0 )2 + (|Vab | /~)2 .

7
(b) Determine the transition probability, Pa→b (t), and show that it never exceeds 1. Con-
firm that |ca (t)|2 + |cb (t)|2 = 1.
Answer: Knowing that the transition probability from state a to b is given by the modulus
squared of cb (t) we compute

|Vab |2
2
Pa→b (t) = |cb (t)| = 2 2 sin2 (ωr t) (13)
4~ ωr

On the other hand, notice that

|Vab |2 < ∆ω 2 ~ + |Vab |2

⇒ |Vab |2 < ∆ω 2 ~2 + |Vab |2
!
~2 |Vab |2
|Vab |2 < 22 2 ∆ω 2 +
2 ~2
!
1 |Vab |2
|Vab |2 < 4~2 2 ∆ω 2 +
2 ~2

Substituting ∆ω, it becomes

" #
2
1 |Vab |
|Vab |2 < 4~2 2 (ω − ω0 )2 +
2 ~2
| {z }
ωr2
2 2
|Vab | < 4~ ωr2
|Vab |2
<1
4~2 ωr2

Thus, analysing from expression 13 for the transition probability that sin (ωr t) oscillates
|Vab |2
between 1 and 0, and that the amplitude 4~2 ω 2 < 1 means that
r

|Vab |2
Pa→b (t) ≤ <1 (14)
4~2 ωr2

8
Finally we proceed to verify
2
|Vab |2

2 2
 i∆ω 
|ca (t)| + |cb (t)| = −
 sin (ωr t) + cos (ωr t) + 2 2 sin2 (ωr t)

2ωr 4~ ωr
∆ω 2 2 2 |Vab |2
= sin (ω r t) + cos (ω r t) + sin2 (ωr t)
4ωr2 4~2 ωr2
" #
∆ω 2 |Vab |2
= 2
+ 2 2 sin2 (ωr t) + cos2 (ωr t)
4ωr 4~ ωr
" #
2
1 1 |Vab |
= 2 ∆ω 2 + sin2 (ωr t) + cos2 (ωr t)
ωr 4 ~2
| {z }
ωr2

ωr2
= sin2 (ωr t) + cos2 (ωr t) = sin2 (ωr t) + cos2 (ωr t)
ωr2

⇒ |ca (t)|2 + |cb (t)|2 = 1 (15)

As requested to confirm.

(c) Check that Pa→b (t) reduces to the small perturbation theory result (Equation 11.35)
when the perturbation is “small”, and state what small means in this context, as a constraint
on V .
Answer: Let us consider the case

|Vab | < ∆ω~

 
|Vab |
⇒ < ∆ω
~
 
|Vab |
~
⇒ <1
∆ω
 2
|Vab |
~
⇒ << 1 (16)
∆ω 2
Saving this result for later, now let’s focus on the Rabi flopping frequency,
1
q
wr = (ω − ω0 )2 + (|Vab | /~)2
2q
1
= ∆ω 2 + (|Vab | /~)2
2 s
1 (|Vab | /~)2
= ∆ω 1 +
2 ∆ω 2
1
∼ ∆ω
2

9
Where, in the last line, we have use the approximation subject to equation 16. Under this
conditions, we can also analyse Pa→b (t), from (b) we know that
|Vab |2
Pa→b (t) = |cb (t)|2 = sin2 (ωr t)
4~2 ωr2
|Vab |2
∼ sin2 (∆ωt/2)
4~2 (∆ω/2)2
|Vab |2
= 2 sin (∆ωt/2) sin (∆ωt/2)
~ ∆ω 2
Taking advantage of the parity of the sine function (odd function) we can write
|Vab |2
Pa→b (t) ∼ 2 [− sin (−∆ωt/2)] [− sin (−∆ωt/2)]
~ ∆ω 2
|Vab |2
= 2 sin2 (−∆ωt/2)
2
~ (−1) ∆ω 2

|Vab |2 2
= 2 sin (−∆ωt/2)
2
~ (−∆ω)
Substituting ∆ω = ω − ω0 we end up with
|Vab |2
 
2 ω0 − ω
Pa→b (t) ∼ sin t (17)
~2 (ω0 − ω)2 2
Recovering precisely the same result as in Equation 11.35. And as approximation 16 suggest,
this comes from the case where the perturbation Ĥ 0 = (V /2) e−iωt includes a potential smaller
than the energy difference (∆ω~) at which the system is oscillating.

(d) At what time does the system first return to its initial state?
Answer: Observe that both, ca (t) and cb (t), include trigonometric functions of argument
ωr t. Given that sine and cosine have a period of 2π, we just need to find the time t = T at
which this value is reached.
⇒ ωr T = 2π
2π
T = (18)
ωr
Hence at this time the system first returns to its initial state.

Problem 11.11
You could derive the spontaneous emission rate (Equation 11.63) without the detour through
Einstein’s A and B coefficients if you knew the ground state energy density of the electro-
magnetic field, rho0 (ω), for then it would simply be a case of stimulated emission (Equation
11.54). To do this honestly would require quantum electrodynamics, but if you are pre-
pared to believe that the ground state consist of one proton in each classical mode, then the
derivation is fairly simple:

10
(a) To obtain the classical modes, consider an empty cubical box, of side l, with one corner
at the origin. Electromagnetic fields (in vacuum) satisfy the classical wave equation18

1 ∂2
 
2
− ∇ f (x, y, z, t) = 0,
c2 ∂t2

where f stands for any component of E or of B. Show that separation of variables, and
the imposition of the boundary condition f = 0 on all six surfaces yields the sanding wave
patterns n π  n π  n π 
x y z
fnx ,ny ,nz = A cos (ωt) sin x sin y sin z ,
l l l
with
πc q 2
ω= nx + n2y + n2z .
l
There are two modes for each triplet of positive integers (nx , ny , nz = 1, 2, 3, ...), correspond-
ing to the two polarization states.
Answer: As instructions suggest, we propose that the solution to the equation above, is a
function f which is the product of 4 single variable functions, one per different variable.

f (x, y, z, t) = f1 (x)f2 (y)f3 (z)f4 (t)

Plugging this proposal into the classical wave equation (Explicit notation is dropped to
simplify expression),

1 ∂2
 
2
− ∇ f (x, y, z, t) = 0
c2 ∂t2
1 ∂2
 
2
− ∇ [f1 (x)f2 (y)f3 (z)f4 (t)] = 0
c2 ∂t2
1 ∂2
 2
∂2 ∂2
  
∂
[f1 (x)f2 (y)f3 (z)f4 (t)] − + + [f1 (x)f2 (y)f3 (z)f4 (t)] = 0
c2 ∂t2 ∂x2 ∂y 2 ∂z 2
1 ∂ 2 f4 ∂ 2 f1 ∂ 2 f2 ∂ 2 f3
f f f
1 2 3 − f f f
2 3 4 − f f f
1 3 4 − f f f
1 2 4 =0
c2 ∂t2 ∂x2 ∂y 2 ∂z 2

Dividing by f = f1 f2 f3 f4 both sides equation above becomes

1 ∂ 2 f4 ∂ 2 f1 ∂ 2 f2 ∂ 2 f3
c2 ∂t2 ∂x2 ∂y 2 ∂z 2
− − − =0
f4 f1 f2 f3
Observing that the expression obtained includes the sum of terms of a single variable, which
equal a constant. Subsequently each of this terms, have to be equal to a constant, such that

11
the sum of constants equals zero. Let’s propose
1 ∂ 2 f4
c2 ∂t2 ∂ 2 f4
= kt2 ⇒ = c2 kt2 f4 (19)
f4 ∂t2
∂ 2 f1
∂x2 ∂ 2 f1
− = kx2 ⇒ = −kx2 f1 (20)
f1 ∂x2
∂ 2 f2
∂y 2 ∂ 2 f2
− = ky2 ⇒ = −ky2 f2 (21)
f2 ∂y 2
∂ 2 f3
∂z 2 ∂ 2 f3
− = kz2 ⇒ = −kz2 f3 (22)
f3 ∂z 2
where kt2 + kx2 + ky2 + kz2 = 0. Using this last expression, we can actually, replace kt2 in terms
of the spacial k constants, so that equation 19 becomes
kt2 = − kx2 + ky2 + kz2

∂ 2 f4 2 2 2 2


⇒ = −c kx + ky + k z f4 (23)
∂t2
Notice that all the differential equations to solve 23, 20, 21, and 22, are of the Harmonic
Oscillator Differential Equation form, which solution can be written as a linear combination
of sine and cosine, or as a constant times cosine or sine with a phase; with frequencies
according to each equation. Then, for the spacial solutions, we choose the first representation
while for time we choose the second one.
f1 (x) = A1 sin (kx x) + B1 cos (kx x)
f2 (y) = A2 sin (ky y) + B2 cos (ky y)
f3 (z) = A3 sin (kz z) + B3 cos (kz z)
 q 
f4 (t) = A4 cos c kx2 + ky2 + kz2 t + B4

For the spacial functions, we have the boundary condition that f = 0 at any of the faces of
the cube of side l. For the faces with corner at the origin we have
f (0, y, z, t) = 0
⇒ [A1 sin (0) + B1 cos (0)] f2 (y)f3 (z)f4 (t) = 0
∴ B1 = 0

f (x, 0, z, t) = 0
⇒ [A2 sin (0) + B2 cos (0)] f1 (x)f3 (z)f4 (t) = 0
∴ B2 = 0

f (x, y, 0, t) = 0
⇒ [A3 sin (0) + B3 cos (0)] f1 (x)f2 (y)f4 (t) = 0
∴ B3 = 0

12
Hence the solutions become
f1 (x) = A1 sin (kx x)
f2 (y) = A2 sin (ky y)
f3 (z) = A3 sin (kz z)
 q 
f4 (t) = A4 cos c kx2 + ky2 + kz2 t + B4
In the same way, we use the boundary condition for the faces at x = l, y = l and z = l,
where we find
f (l, y, z, t) = 0
⇒ [A1 sin (kx l)] f2 (y)f3 (z)f4 (t) = 0
⇒ sin (kx l) = 0
πnx
∴ kx =
l
f (x, l, z, t) = 0
⇒ [A2 sin (ky l)] f1 (x)f3 (z)f4 (t) = 0
⇒ sin (ky l) = 0
πny
∴ ky =
l
f (x, y, l, t) = 0
⇒ [A3 sin (kz l)] f1 (x)f2 (y)f4 (t) = 0
⇒ sin (kz l) = 0
πnz
∴ kz =
l
Where ni = 1, 2, 3, ... and i = x, y, z. Additionally, notice that the time reference frame can
always be chosen by convenience, therefore, we can get rid of the phase B4 in f4 (t), obtaining
 q   cπ q 
2 2 2
f4 (t) = A4 cos c kx + ky + kz t = A4 cos 2 2 2
nx + ny + nz t
l
So that the solutions become
 πn 
x
f1 (x) = A1 sin x
 πn l 
y
f2 (y) = A2 sin y
l
 πn 
z
f3 (z) = A3 sin z
 cπl q 
f4 (t) = A4 cos n2x + n2y + n2z t
l
Therefore, constructing the full solution f = f1 (x)f2 (y)f3 (z)f4 (t) we get
 πn   πn   πn 
x y z
fnx ,ny ,nz (x, y, z, t) = A cos (ωt) sin x sin y sin z (24)
l l l
where A = A1 A2 A3 A4 and ω defined by the problem is given by ω = cπl n2x + n2y + n2z .
p

13
(b) The energy of the photon is E = ~ν = ~ω (Equation 4.92), so the energy in the mode
(nx , ny , nz ) is
π~c q 2
Enx ,ny ,nz = 2 nx + n2y + n2z .
l
What then, is the total energy per unit volume in the frequency range dω, if each mode gets
one photon? Express your answer in the form
1
dE = ρ0 (ω) dω
l3
and read off ρ0 (ω). Hint: refer to Figure 5.3.
Answer: Defining the wavevector

k = (kx , ky , kz ) ,

which comes from the spatial part of the solution. While using the same analysis from Figure
5.3, we have that in a shell of thickness dk k-space, contains a number of photons equal to
V 2
k dk (25)
π2
Where V is the volume of the box, so V = l3 . It is important to notice that this expression
does hold since now we have 2 modes per each triplet of integers, when for the electron case,
we had 2 electrons per triplet of integers due to the 2 possible spin configurations.
On the other hand, we can express the frequency ω in terms of the wavevector k, precisely
its magnitude.
r
cπ q 2 π 2 n2x π 2 n2y π 2 n2z q
ω= nx + n2y + n2z = c + + = c kx2 + ky2 + kz2 = ck
l l2 l2 l2
⇒ ω = ck and dω = cdk (26)

Finally, with this information we know that the energy of per mode, is given by
π~c q 2
Enx ,ny ,nz = nx + n2y + n2z
l
= ~ck

Notice that the factor of 2 from the original formula is not consider, since we are computing
the energy per mode. The duality of modes was already consider in equation 25 as previously
described.
V 2
dE = (~ck) k dk
π2

14
This expression comes from the energy of a single photon, multiplied by the amount of
photons in a shell of thickness dk. Changing variables as equation 26 states

l3 3
dE = ~ k dω
π2
l 3  ω 3
dE = ~ 2 dω
π c
dE ~ω 3
⇒ 3 = 2 3 dω
l |π{zc }
ρ0 (ω)

~ω 3
ρ0 (ω) = (27)
π 2 c3

(c) Use your result, together with Equation 11.54, to obtain the spontaneous emission rate.
Compare Equation 11.63.
Answer: From Equation 11.54 we have that the transition rate for stimulated emission from
state b to a, under the influence of incoherent, unpolarized light incident from all directions
is
π
Rb→a = 2
|℘|2 ρ0 (ω)
30 ~
Substituting ρ0 (ω0 ) from the previous exercise
3
π 2 ~ω0
⇒ Rb→a = |℘|
30 ~2 π 2 c3
ω03
= |℘|2 = A
30 ~πc3

⇒ Rb→a = A (28)
Obtaining the spontaneous emission rate (A from Equation 11.63) already calculated, but
computed from the ground state of energy density of the electromagnetic field.

15