University of Oxford

Keble College
Hilary term
Tutor: Alex Lvovsky

CP4: Mathematical Methods 2

Additional homework problem

a)
e
  
e e e

e 
e

e

 e
e e



e


e
 
e e e e 
e


e

e

e
 
e e

e e

e
e
e


e



e

e
e

b) Re n  1

  

Im n

FWHM = 

  

Figure 1: Classical theory of dispersion. (a) Model; (b) Real and imaginary parts of the refractive
index.

Problem B.1. A light wave propagating along the z axis inside a linear medium (e.g., a gas of
atoms) is governed by the equation
 2
1 ∂2 1 ∂2

∂
2
− 2 2
E= P, (1)
∂z c ∂t 0 c2 ∂t2

where P is the polarization of the medium, i.e. the density, per unit volume, of the dipole moment
induced in the medium by the field. The solution to this equation is

E(z, t) = ReE0 eikz−iωt . (2)

The polarization can be calculated as shown in Fig. 1(a). We treat each atom as a damped harmonic
oscillator driven by the oscillating field, where ω0 is the atomic transition frequency and γ its width.

1
 a) Find the displacement of the electron x(z, t) for an “atom” located at z. Hint: don’t forget
that the electron’s charge is negative.

b) Write down the polarization P (z, t) given that the dipole moment of each atom is −ex(z, t)
and the number density (number per unit volume) of atoms is N . Simplify your answer under
the assumption γ, ∆  ω0 for ∆ = ω − ω0 . Hint: under this assumption, you can also replace
iωγ → iω0 γ in the damping term.
c) Find the wavenumber k(ω) as a function of the frequency ω such that Eq. (2) is a solution to
Eq. (1). Hint: k(ω) is significantly complex.
c ck
d) Find the refractive index n(ω) = vph = ω . Check that your result is consistent with Fig. 1(b).

The remainder of the problem should be solved under the assumptions that |n − 1|  1 for all
frequencies (i.e. the absolute effect of the atomic gas on the continuous electromagnetic wave is very
small).
e) Show that the field intensity (power per unit area) behaves as I(z) ∝ e−αz (Beer’s law). The
quantity α in this equation is called the absorption coefficient. Find it for ω = ω0 . Hint: the
intensity is proportional to the square of the amplitude’s absolute value.
f) Find the group velocity at ω = ω0 . Find the conditions under which the group velocity is (i)
superluminal and (ii) negative. Hint: it is more convenient to calculate v1gr = dRek
dω .

g) Suppose you wish to perform an experiment to observe negative group velocity in atomic
rubidium vapor. Atomic rubidium has a transition with the wavelength of λ = 795 nm and
the linewidth of γ/2π = 5.7 MHz. Find the number density required. Check that the condition
|n − 1|  1 is satisfied at this number density.
h) For the experiment, it is optimal to use a pulse with the duration τ ∼ γ −1 (so the group velocity
−1
is approximately constant over the pulse’s spectrum). Justify this choice. For vgr = 0, estimate
the length of the vapor cell such that the absolute value of the group delay, in comparison with
a freely propagating pulse, is equal to the pulse duration. Estimate the fractional loss of the
pulse’s energy after traversing the cell.

See L. J. Wang, A. Kuzmich and A. Dogariu, Gain-assisted superluminal light propagation,