Report for Quantum Optics

Ye Guo
February 2025

1 From Cavity QED to DCE

During the lecture, the topic of the behavior of quantized electromagnetic waves in a static cavity has
been brought up. The questions below will provide a guide for more in depth examination of such a
configuration, extending it to the case of a non-static cavity where photons could be spontaneously
created with energy extracted from vacuum fluctuations.

2 Properties of a Static Cavity

2.1 Mode Structure and Quantization
Consider a one-dimensional electromagnetic field confined in a perfectly conducting cavity of length
L with mirrors at x = 0 and x = L. The field satisfies Dirichlet boundary conditions:

A(0, t) = A(L, t) = 0. (1)

We will show that the vacuum state of this set up has divergent energy and fluctuation despite its
simplicity.
1. Mode Quantization:
(a) Derive the allowed mode frequencies ωn of the standing wave solutions.
(b) Write the general form of the normal mode expansion for the vector potential operator
ân (x, t) in terms of annihilation and creation operators.

1
 Solution

(a) As in the lecture, the boundary condition restrains λ with L = n × λ2 , where n is a

natural number. Then, since we have the dispersion relation for light in a vacuum as
λf = c, we get that
nπc
ωn = 2πf = (2)
L
(b) From (a), we also have kn = 2π/λ = πn/L. Together with ωn , an arbitrary vector
potential permitted by our boundary conditions are superpositions of standing waves
of the form n An (t) sin nπx
P
L . This could be written as:
r
X h̄ nπx
αn e−iωn t + αn∗ eiωn t sin


A= (3)
n
2ωn ε0 L L
(4)
q
h̄ωn
Where αn = (ωn Ãn −iẼn )/2 2ϵ 0L
, Ã, Ẽ being the field components in k representation
(Same notation as ”Quantum Optics” by Grynberg, Aspect, Fabre). Promoting αn
and αn∗ in this self consistent expression to operators ân and â†n then provides the
corresponding operator with adequate normalization:
r
X h̄ nπx
ân e−iωn t + â†n eiωn t sin


 = (5)
n
2ω n ε0 L L

2. Energy and Photon States:

(a) Derive the Hamiltonian for the quantized field.
(b) Show that the energy spectrum consists of discrete photon number states.
(c) What is the zero-point energy of the cavity, and what does it tell us?

Solution

(a) As in the lecture, the Hamiltonian in second quantization language for a specific
ω is just h̄ω(↠â + 12 ). Since we are considering all normal modes, the Hamiltonian
becomes:
X 1
Ĥ = h̄ωn (â†n ân + ) (6)
n
2

(b) Since the â†n ân is just the number operator, this is trivially true.
P
(c) The zero-point energy is defined as ⟨0| Ĥ |0⟩, this is just n h̄ωn /2. Given that we
think of the eigenstate of the number operator with eigenvalue 0 as a vacuum, |0⟩ is
a
Pvacuum state, and this vacuum state has a none zero, and in fact infinite energy of
n h̄ωn /2. This commands renormalization such as by choosing a cut off frequency.

3. Field Commutation Relations and Uncertainty Principle:

2
(a) Compute the commutator [Â(x, t), Π̂(x′ , t)], where Π̂(x, t) = ϵ0 Ê is defined as the conju-
gate momentum.
(b) Compute the expectation value and variance of the electric field in the vacuum state,
show that the variance diverges.
(c) Use this to show that there are fundamental quantum fluctuations in the electromagnetic
field, consistent with Heisenberg’s uncertainty principle.
(d) Comment and conclude on your findings.

3
Solution
∂
(a) We know that Ê = − ∂t Â, this gives us:
r
X h̄ nπx
iωn â†n eiωn t − ân e−iωn t sin


Ê = (7)
n
2ωn ε0 L L

The cross terms do not contribute to the commutator between this and  since we have
[âm , â†n ] = δmn . therefore it remains only to compute the commutator for each n, this
is just:

h̄  nπx   nπx′  
ân e−iωn t + â†n eiωn t â†n eiωn t − ân e−iωn t



iωn sin sin
2ωn ε0 L L L
† iωn t −iωn t
ân e−iωn t + â†n eiωn t



− ân e − ân e
ih̄  nπx   nπx′ 
= sin sin 2[ân â†n − â†n ân ]
2ε0 L L L
ih̄  nπx   nπx′ 
= sin sin (8)
ε0 L L L

So we have overall:
X ih̄  nπx   nπx′ 
[Â(x, t), Π̂(x′ , t)] = sin sin = ih̄δ(x − x′ ) (9)
n
L L L

(b) We have that ⟨0| Ê |0⟩ = 0 since the creation and annihilation operators do not
preserve particle number. Regarding the Variance, since we are computing in the
vacuum state (a definite particle number state), only the terms which have equal number
of creation and annihilation operators of the same particle will survive the sandwich.
This gives:
X −h̄ωn nπx
⟨0| Ê 2 |0⟩ = â†n ân − ân â†n sin


(10)
n
2ε0 L L
X h̄ωn nπx
= sin (11)
n
2ε 0 L L

Which diverges.

(c) The results of (a) directly imply that

h̄
∆A∆Π ≥ (12)
2
h̄
=⇒ ∆A∆E ≥ (13)
2ϵ0
(d) Hence there are always fluctuations of electromagnetic fields independent of the
state of the system. In particular, the fluctuation in the vacuum state is maximized
as it is divergent. This may be unphysical and commands efforts to renormalize the
result.

4
3 The Dynamical Casimir Effect in a Moving Cavity
Now, consider the case where the right boundary of the cavity moves periodically:

L(t) = L0 + ϵ cos(Ωt), where ϵ ≪ L0 . (14)

1. Time-Dependent Mode Frequencies:

(a) Derive how the mode frequencies ωn (t) change as a function of time, provide a first order
approximation.
(b) Guess a parametric resonance condition by invoking upon the fact that each mode of
oscillation is equivalent to a harmonic oscillator.

Solution

(a) We proceed as in the static case, only with the introduction of time dependence on
L:
nπc
ωn (t) = 2πf (t) = (15)
L(t)

The first order approximation is:

 
ϵ cos(Ωt)
ωn (t) ≈ ωn0 1 − . (16)
L0
√
(b) In a classical pendulum, the frequency is proportional to 1/ l, where l is the
string length of the pendulum. If we allow the string length (hence ω) to be perturbed
sinusoidally at twice the natural frequency (pull string at lowest point, release string at
highest point), the system will be experience an injection of energy and the energy of
the pendulum would increase. Accordingly, we can expect this phenomenon to occur
for any system with the same original Hamiltonian (harmonic oscillator), if its new
frequency is dependent on a parameter of order −1/2 that is perturbed at periodicity
twice the natural frequency.
p
In other words, for any harmonic oscillator with frequency ω(t), if ω(t) ∝ 1/ l(t)
and l(t) = l0 + ϵP (Ωt), where ϵ is small and P is a function with period 2π, setting
Ω = 2ω(0) will cause parametric resonance. Then, we expect this property to hold for
a quantum harmonic oscillator by the correspondence principle

Since a sum of normal modes is nothing but a sum of quantum harmonic oscillators,
we expect the systempto have this behavior as well. The only complication is that
instead of a ω(t) ∝ 1/ L(t), we have ω(t) ∝
q1/L(t). This though is quickly reconciled
ϵP (t)
if we realize that the Taylor expansion of 1/ 1+ L0 at the first order is nothing but
ϵP (t)
1− 2L0 + O(ϵ ). Therefore the parametric resonance condition should be Ω = 2ωn0 .
2

2. Bogoliubov Transformation and Photon Creation: The normal mode expansion of the

5
vector potential operator is now modified due to time dependence of L(t) and thus ω(t):
s
X h̄  Rt ′ ′
Rt ′ ′
 nπx
 = ân (t)e−i ωn (t )dt + â†n (t)ei ωn (t )dt sin . (17)
n
2ωn (t)ε 0 L(t) L(t)

The general form of the Hamiltonian is:

Z L(t)  
1 1 2
Ĥ(t) = ϵ0 Ê 2 (x, t) + B̂ (x, t) dx (18)
2 0 µ0

Deriving Ê and B̂ from (17), using the results of part (3.1.a), and inserting it into (18) gives
the following Hamiltonian:
X
h̄ωn0 â†n ân + ih̄gn (t) â†2 2


Ĥ = n − ân (19)
n

ϵ 0
With gn (t) = 2L0 ωk cos(Ωt), and the time dependence of the operators suppressed.

(a) Compute the time evolution of the operators in this Heisenberg picture.
(b) Derive an expression for the number of photons created as a function of time.
(c) Show that the photon number grows exponentially when the parametric resonance con-
dition Ω = 2ωn0 is satisfied.
(d) Comment on this result in relation to the previous section.

6
Solution

(a) Applying the Heisenberg equation of motion,

d i
âk = [Ĥ, âk ] (20)
dt h̄
d † i
â = [Ĥ, â†k ] (21)
dt k h̄
Using the standard commutators,

[âk , â†k ] = 1, [âk , âk ] = 0, [â†k , â†k ] = 0, (22)

we know that only terms in the sum of the hamiltonian with n = k survive. This gives
us:

[H, âk ] = h̄ωk0 [â†k âk , âk ] + ih̄gk (t)[â†2 2

k − âk , âk ] (23)

Using:

[â†k âk , âk ] = −âk (24)

[â2k , âk ] = −2âk (25)

we get:

[H, âk ] = −h̄ωk0 âk − 2ih̄gk (t)â†k (26)

So
d
âk = −iωk0 âk − 2gk (t)â†k (27)
dt
Similarly,
d †
â = iωk0 â†k + 2gk (t)âk (28)
dt k
We therefore obtain the coupled system:

−iωn0 −ign (t) ân

    
d ân
= (29)
dt â†n ign (t) iωn0 â†n

(b) This system can be solved by considering the ansatz:

ân (t) = un (t)ân (0) + vn (t)â†n (0), (30)

where un (t) and vn (t) are the Bogoliubov coefficients, which obey:

−iωn0 −ign (t) un

    
d un
= (31)
dt vn ign (t) iωn0 vn
Using the initial conditions:

un (0) = 1, vn (0) = 0, (32)

the number of created photons is given 7by:

Nn (t) = ⟨0|â†n (t)ân (t)|0⟩ = |vn (t)|2 . (33)

 Solution

(c)When the resonance condition Ω = 2ωn0 is satisfied, the equations for un and vn
become the standard form of the Mathieu equation, which exhibits an exponentially
growing solution. In this case, the function gn (t) takes the form:
ϵ 0
gn (t) = ω cos(Ωt). (34)
2L0 n
For small ϵ, the solutions to the system yield:

vn (t) ≈ sinh(γt), (35)

where γ is the growth rate, given by:
ϵ 0
γ≈ ω . (36)
2L0 n
Thus, the photon number grows as:

Nn (t) = |vn (t)|2 ≈ e2γt . (37)

This confirms the exponential growth of photon number when the parametric resonance
condition is met.

(d) The big picture is, energy is extracted from the vacuum via the moving boundary
to create more and more particles over time. This is most significant when Ω = 2ωn0 .
Realistically we know that the vacuum could not provide an infinite amount of energy,
and this apparent contradiction is reconciled as we know that only differences in en-
ergy have physical energy. Heuristically, the casimir force ”generated by these energy
changes” could then be considered to be related to the generation process.

4 Thoughts on Videos I-III

It is somewhat unfortunate that we did not go into more of the EPR paradox, I have read the
original paper and am quite fascinated by just what motivated E.P.R. to come up with condition
”If, without in any way disturbing a system, we can predict with certainty (i.e., with probability
equal to unity) the value of a physical quantity, then there exists an element of physical reality
corresponding to this physical quantity.” Although I’ve come back to it several times, I feel like I
still do not understand it completely.
I have met Alain Aspect in his nobel prize speech at Ecole Polytechnique, and have also in fact
asked him a question during the conference, although I do regret I was not knowledgable enough
for it to be technical. Nevertheless, I would have been very eager to have written the report on his
experimental setup which ultimately proved Einstein wrong, but the condition that it should adhere
more to course content prompted me to focus more on the quantization formalism and cavity QED.

8
5 Difference Between Semiclassical and Classical Light Mat-
ter Interaction
Fundamentally speaking, in the semiclassical case light is treated as a perturbation in the Hamilto-
nian of the system of matter, whilst the fully quantum system involves a hilbert space of external
product states and their linear combinations. Immediately, the latter allows for more exotic quan-
tum phenomenon such as entanglement and Decoherence.

Other important differences include:

• The photoelectric effect, where the continuous energy transfer by light in the classical picture
would fail to explain the inability os stimulation under a certain threshold.
• The (spontaneous) emission of photons despite a lack of optical stimulation, which is only
explained by the commutation relation between the ladder operators in the quantum theory

• Vacuum energy or photon generation, as mentioned in this paper is also only derivative of a
quantum model. Although we see here the appearance of infinities also illustrate the limitations
of the quantum model.