1 Introduction to scattering theory

(lecture 2)
1.1 Electromagnetic formulation of the problem
1.2 Amplitude scattering matrix
1.3 Stokes parameters and scattering matrix
1.4 Extinction, scattering and absorption
Let us assume that medium surrounding the scattering particle is non-absorbing. The total
or extinction cross section is then the sum of the absorption and scattering cross sections:

σe = σs + σa , (1)

where
Z
1
σe = − dAS e · er ,
Ii A
Z
1
σs = dAS s · er , (2)
Ii A
when A is a spherical envelope of radius r containing the scattering particle.
Let the original field be of ex -polarized form E 0 = Eex . In the radiation zone,
exp[ik(r − z)]
Es ∝ XE, er · X = 0,
−ikr
k
Hs ∝ er × E s , (3)
ωµ
where the vector scattering amplitude X is related to the amplitude scattering matrix as
follows:

X = (S4 cos φ + S1 sin φ)es⊥ + (S2 cos φ + S3 sin φ)esk . (4)

By making use of the asymptotic forms of the scattered field shown above and ex -polarized
original field, the so-called optical theorem can be derived: extinction depends only on scat-
tering in the exact forward direction,
4π
σe = Re[(X · ex )θ=0 ]. (5)
k2
1
In addition,
Z
dσs
σs = dΩ , (6)
4π dΩ
where the differential scattering cross section is
dσs |X|2
= . (7)
dΩ k2
The extinction, scattering, and absorption efficiencies are defined as the ratios of the
corresponding cross sections to the geometric cross section of the particle A⊥ as projected in
the propagation direction of the original field:
σe
qe = ,
A⊥
σs
qs = ,
A⊥
σa
qa = . (8)
A⊥
For an unpolarized original field, the cross sections are
1 (1)
σe = (σe + σe(2) ),
2
1 (1)
σs = (σs + σs(2) ), (9)
2
where the indices 1 and 2 refer to two polarization states of the original field perpendicular
to one another.

2 Plane waves
The electromagnetic plane wave

E = E 0 eik·x−iωt
H = H eik·x−iωt
0 (10)
can, under certain conditions, fulfil Maxwell’s equations. The physical fields correspond to the
real parts of the complex-valued fields. The vectors E 0 and H 0 above are constant vectors
and can be complex-valued. Similarly, the wave vector k can be complex-valued:
k = k0 + ik00 , k0 , k00 ∈ Rn (11)

2
Inserting (11) into equation (10), we obtain
00 0
E = E 0 e−k ·x eik ·x−iωt
00 0
H = H e−k ·x eik ·x−iωt
0 (12)
00 00
In Eq. (12), E 0 e−k ·x and H 0 e−k ·x are amplitudes and k0 · x − ωt = φ is the phase of the
wave.
An equation of the form k · x =constant defines, in the case of a real-valued vector k, a planar
surface, whose normal is just the vector k. Thus, k0 is perpendicular to the planes of cons-
tant phase and k00 is perpendicular to the planes of constant amplitude. If k0 k k00 , the planes
coincide and the wave is homogeneous. If k0 ∦ k00 , the wave is inhomogeneous. A plane wave
propagating in vacuum is homogeneous.

In the case of plane waves, Maxwell’s equaitons can be written as

k · E0 = 0
k · H0 = 0
k × E0 = ωµH 0
k × H0 = −ωE 0 (13)

The two upmost equations are conditions for the transverse nature of the waves: k is
perpendicular to both E 0 and H 0 . The two lowermost equations show that E 0 and H 0
are perpendicular to each other. Since k, E 0 , and H 0 are complex-valued, the geometric
interpretaion is not simple unless the waves are homogeneous.
It follows from Maxwell’s equations (13) that, on one hand,

k × (k × E 0 ) = ωµk × H 0 = −ω 2 µE 0 (14)

and, on the other hand,

k × (k × E 0 ) = k(k · E 0 ) − E 0 (k · k) = −E 0 (k · k), (15)

so that

k · k = ω 2 µ. (16)
Plane waves solutions are in agreement with Maxwell’s equations if

k · E0 = k · H 0 = E0 · H 0 = 0 (17)

3
and if

k 02 − k 002 + 2ik0 · k00 = ω 2 µ. (18)

Note that  and µ are properties of the medium, whereas k0 and k00 are properties of the
wave. Thus,  and µ do not unambiguously determine the details of wave propagation.
In the case of a homogeneous plane wave (k0 kk00 ),

k = (k 0 + ik 00 )ê, (19)
where k 0 and k 00 are non-negative and ê is an arbitrary real-valued unit vector.

According to Eq. (16),

ω 2 m2
(k 0 + ik 00 )2 = ω 2 µ =
, (20)
c2
√
where c = 1/ 0 µ0 is the speed of light in vacuum and m is the complex-valued refractive
index
r
µ
m= = mr + imi , mr , mi ≥ 0. (21)
0 µ0
In vacuum, the wave number is ω/c = 2π/λ, where λ is the wavelength. The general homo-
geneous plane wave takes the form
2πmi s 2πmr s
E = E 0 e− λ ei λ
−iωt
(22)

where s = e · x. The imaginary and real parts of the refractive index determine the attenua-
tion and phase velocity v = c/mr of the wave, respectively.

3 Poynting vector
Let us study the electromagnetic field E, H that is time harmonic. For the physical fields
(the real parts of the complex-valued fields), the Poynting vector

S =E×H (23)

describes the direction and amount of energy transfer everywhere in the space.

4
 Let n be the unit normal vector of the planar surface element A. Electromagnetic energy
is transferred through the planar surface with power S · n A, where S is assumed constant
on the surface. For an arbitrary surface and S depending on location, the power is
Z
W = − S · ndA, (24)
A
where n is the outward unit normal vector and the sign has been chosen so that positive W
corresponds to absorption in the case of a closed surface.
The time-averaged Poynting vector

1 t+τ
Z
hSi = S(t0 )dt0 τ >> 1/ω (25)
τ t

is more important than the momentary Poynting vector (cf. measurements).

The time-averaged Poynting vector for time-harmonic fields is

1
hSi = Re{E × H ∗ } (26)
2
and, in what follows, this is the Poynting vector meant even though the averaging is not
always shown explicitly.
For a plane wave field, the Poynting vector is

1 n E × (k∗ × E ∗ ) o
S= Re{E × H ∗ } = Re , (27)
2 2ωµ∗
where

E × (k∗ × E ∗ ) = k∗ (E · E ∗ ) − E ∗ (k∗ · E). (28)

For a homogeneous plane wave,

k · E = k∗ · E = 0 (29)
and
√
1 n µ o 4πIm(m)z
S = Re ∗
|E 0 |2 e− λ êz . (30)
2 µ

5
4 Stokes parameters
Consider the following experiment for an arbitrary monochromatic light source (see Bohren
& Huffman p. 46). In the experiment, we make use of a measuring apparatus and polarizers
with ideal performance: the measuring apparatus detects energy flux density independently
of the state of polarization and the polarizers do not change the amplitude of the transmitted
wave.
Denote
E = E 0 eikz−iωt , E 0 = E⊥ ê⊥ + Ek êk
E⊥ = a⊥ e−iδ⊥
Ek = ak e−iδk a⊥ , ak ≥ 0, δ⊥ , δk ∈ R (31)

Experiment I
No polarizer: the flux density is proportional to
|E 0 |2 = Ek Ek∗ + E⊥ E⊥∗ (32)

Experiment II
Linear polarizers k and ⊥:
1) k: the amplitude of the transmitted wave is Ek and the flux density is Ek Ek∗
2) ⊥: the amplitude of the transmitted wave is E⊥ and the flux density is E⊥ E⊥∗

The difference of the two measurements is Ik − I⊥ = Ek Ek∗ − E⊥ E⊥∗ .

Experiment III
Linear polarizers +45◦ ja −45◦ : The new basis vectors are
ê+ = √12 (êk + ê⊥ )


ê− = √1 (êk − ê⊥ )

2

and

E 0 = E+ ê+ + E− ê−
1
E+ = √ (Ek + E⊥ )
2
1
E− = √ (Ek − E⊥ ).
2

6
1) +45◦ : the amplitude of the transmitted wave is E+ and the flux density is
E+ E+∗ = 21 (Ek Ek∗ + Ek E⊥∗ + E⊥ Ek∗ + E⊥ E⊥∗ )
2) −45◦ : the amplitude of the transmitted wave is E− and the flux density is
E− E−∗ = 21 (Ek Ek∗ − Ek E⊥∗ − E⊥ Ek∗ + E⊥ E⊥∗ )

The difference os the measurements is I+ − I− = Ek E⊥∗ + E⊥ Ek∗ .

Experiment IV
Circular polarizers R and L:

1
êR = √ (êk + iê⊥ ) êR · ê∗R = 1
2
1
êL = √ (êk − iê⊥ ) êL · ê∗L = 1 êR · ê∗L = 0
2
and

E 0 = ER êR + EL êL
1
ER = √ (Ek − iE⊥ )
2
1
EL = √ (Ek + iE⊥ ).
2

1) R: the amplitude of the transmitted wave is ER and the flux density is ER ER∗ =
1
2
(Ek Ek∗ − iEk∗ E⊥ + iE⊥∗ Ek + E⊥ E⊥∗ )
2) L: the amplitude of the transmitted wave is EL and the flux density is EL EL∗ =
1
2
(Ek Ek + iEk∗ E⊥ − iE⊥∗ Ek + E⊥ E⊥∗ )
∗

The difference of the measurements is IR − IL = i(E⊥∗ Ek − Ek∗ E⊥ ).

With the help of Experiments I-IV, we have determined the Stokes parameters I, Q, U ,
and V :

I = Ek Ek∗ + E⊥ E⊥∗ = a2k + a2⊥

Q = Ek Ek∗ − E⊥ E⊥∗ = a2k − a2⊥
U = Ek E⊥∗ + E⊥ Ek∗ = 2ak a⊥ cos δ
V = i(Ek E⊥∗ − E⊥ Ek∗ ) = 2ak a⊥ sin δ δ = δk − δ⊥ (33)