Physics 522.

Quantum Mechanics II
Problem Set #10 – Time Dependent Perturbations
Due Thursday, May 5 2011

Problem 1: Photon absorption cross-section and photon scattering (10 points)

Given a laser beam at frequency ωL incident of an atom, we can define the absorption cross-
section in the usual way,
P
σ abs (ω L ) = abs ,
I inc

where Iinc is the incident intensity and Pabs is the absorbed power Assume a “weak” incident
(polarized) monochromatic (narrow band) plane wave near resonance to a two-level transition
g → e , for a time T>>1/Γ, where Γ is the atomic linewidth.

(a) Use Fermi’s Golden rule to show that

4π 2 ˆ  2
σ abs (ω L ) = e d ⋅ ε L g ω L g(ω L ) ,where g(ω L ) is the normalized atomic lineshape,
c
ˆ 
2
σ0 8πω L e d ⋅ ε L g
and for the case of lifetime broadening, σ abs (ω L ) = , where σ 0 =
1 + 4Δ2 / Γ 2 c Γ
is the absorption cross-section on resonance (Δ=0).

(b) Using the expression for the spontaneous emission rate (Einstein-A coefficient),
4 3 2
Γ= k L e d g , show that the resonant absorption cross section is,
3
2
• σ 0 = 6π 2 J e mg + q 1q J g mg , when the field is polarized along with eq and the atom is
“polarized”, i.e. prepared in a well defined m-state.
• σ 0 = 2π 2 if the field and/or the atom are unpolarized.

The bottom line here is that the resonant cross section is of the order of the square of the exciting
wave length, much larger that the physical cross section of the atom itself for optical resonance.

(c) If a photon is absorbed by an atom is will eventually be emitted. This is not necessarily true
of other matter, e.g. a molecule or condensed mater (liquid or solid) where the excited matter can
relax with other degrees of freedom such as vibrational motion (phonons) etc. The absorption of
a photon followed by spontaneous emission is a random direction is photon scattering.

Argue that for weak fields, the scattering cross-section is equal to the absorption cross-section,
σ scatt (ω ) = σ abs(ω ) , and show that the scattering-rate is,

I s Ω2 / 2
γ scatt = σ scatt = Γ , where s = 2 is the saturation parameter.
ω 2 Δ + Γ2 / 4
Problem 2: Two-photon transitions (20 points)

Suppose we want to excite a hydrogen atom to a state with a very large principle quantum
number n=100 (a so-called “Rydberg state”). To get there directly would require an x-ray
photon, for which lasers do not exist. If we would like to coherently excite the transition, we
must thus resort to a two-photon excitation via a virtual transition. We can enhance the
excitation rate by tuning one of the lasers frequencies relatively close to a dipole-allowed strong
transition, and then a second photon to connect to the Rydberg state.
For example, suppose we shine two laser beams on the gas with electric fields E1 (t) = Re(E1e−iω t ) 1

and E2 (t) = Re(E2 e−iω t ) . We can tune one laser near the 1s → 2 p transitions with a detuning Δ ,
2

and a second laser so that ω 1 + ω 2 is close to the energy difference between 1s and 100s , with
a two-photon detuning of δ , as shown.

δ 100s

ω2
Δ 2p

ω1

1s

(a) Use second order perturbation theory to argue that the transition probability P100 s←1s after a
time t is approximately,

⎛Ω Ω ⎞
2

P100 s←1s = ⎜ 1 2 ⎟ sin 2 (δ t / 2) ,

⎝ 2Δδ ⎠

where Ω1 = 2 p d̂ 1s ⋅ E1 and Ω2 = 100s d̂ 2 p ⋅ E2 define the Rabi frequencies of the two
dipole-allowed transitions.

Now let us solve the problem more exactly according to the time-dependent Schrödinger
equation. Restricting the dynamics to these three levels, Hamiltonian in the rotating-wave
approximation can be written Ĥ = Ĥ A + Ĥ int (t) , where

Ĥ A = E1s 1s 1s + E2 p 2 p 2 p + E100 s 100s 100s

Ω1 Ω2
Ĥ int (t) = −
2
(
2 p 1s e−iω1t + 1s 2 p e+iω1t −
2
) (
100s 2 p e−iω 2t + 2 p 100s e+iω 2t )
To simplify notation, let: a = 1s , b = 2 p , c = 100s .
(b) In the interaction picture, use the time-dependent Schrödinger equation show that the
probability amplitudes in the three states evolve according to

Ω1 +iΔt
ca = i e cb
2
Ω Ω
cb = i 1 e−iΔt ca + i 2 ei(δ −Δ )t cc
2 2
Ω
cc = i 2 e−i(δ −Δ )t cb
2

(c) In the limit that Δ >> Ω1 ,Ω2 , δ , the population in the 2p state is never large and can be
eliminated from the dynamics as a “virtual excitation”. To do so, we formally integrate

Ω Ω
t t

cb (t) = i 1 ∫ dt ′e−iΔt ′ ca (t ′ ) + i 1 ∫ dt ′ei(δ −Δ )t ′ cc (t ′ ) .

2 0 2 0

Plugging this back in and neglecting the rapidly oscillating terms (which contribute to order
Ω 3t / Δ 2 and Ω2δ t / Δ 2 ), show that we get the following approximate equations of motion the
couple the probability amplitude in 1s and 100s in the two-photon transition,

Va Ωeff
ca = −i ca − i cc eiδ t
 2
Vc Ωeff
cc = −i cc − i ca e−iδ t
 2

Ω12 Ω22 ΩΩ
where Va ≡ , Vc ≡ , Ωeff = 1 2 .
4Δ 4Δ 2Δ

(d) With ca (0) = 1, show that up to an overall irrelevant phase, the solution is,

⎛ 
Ωt δ eff  ⎞
Ωt Ωeff 
Ωt
ca (t) = e−iδ t /2 ⎜ cos +i sin ⎟ , cc (t) = ieiδ t /2 sin
⎝ 2 Ω 2⎠ 
Ω 2

 = Ω2 + δ 2
where δ eff = δ + (Vc − Va ) /  and Ω eff eff

This solution is known as two-photon Rabi flopping.

 (Hint: Compare to the Rabi solution of a

(e) Please interpret the parameters Ωeff ,Va ,Vc , δ eff , Ω
two-level resonance).

(f) Show that to lowest order in Ω1 ,Ω2 , you recover the solution from perturbation theory in (a)