Fresnel formul

Baptiste Augui
1st July 2009
1 Single interface
Let us consider the situation depicted in gure 1, where we dene the Fresnel coecients as the
1
z
x
y
H
i
E
i E
r
E
t
H
r
H
t

i
2

t
(a) TE polarisation (b) TM polarisation
1
E
i
H
i
H
r
H
t
E
r
E
t

i
2

t
Figure 1: Schematic of reection and transmission by a single interface. Note that the orientation
of the vectors has been chosen so as to have a consistent picture at normal incidence.
ratio of the complex amplitude of the electric elds, r = E
r
/E
i
for the reection coecient, and
t =E
t
/E
i
for the transmission coecient. e Fresnel coecients can be obtained by considering
the continuity of the tangential component of the E and H elds at the interface. e continuity
of the normal components of D and B does not yield supplementary conditions and will not be
considered. It is worth noting that the continuity of H
y
is only valid for materials that do not
sustain a surface current resulting from the addition of external charges.
1.1 TE-polarised light
For TE-polarised light, the continuity of E
y
reads,
E
i
+E
r
=E
t
which yields,
1+r =t . (1)
e continuity of H
y
can be written as,
H
i
cos
i
H
r
cos
i
=H
t
cos
t
1
It is convenient to express the angles in terms of the normal component of the k-vectors,
k
z
=nk
0
cos. (2)
for the incident, reected, and refracted elds, so that cos
t
/cos
i
=
k
z2
n
1
k
z1
n
2
. e complex mag-
nitude of the magnetic and electric elds are linked in each medium by the optical impedance. Its
expression can be obtained from the induction law,
E =
t
B
expressed in Fourier space as,
kE=i B
e ratio E/H is the impedance, written as,
E
H
=c
0
/n =

0
=Z Z
0
. (3)
We can dene the admittance Y as the inverse impedance, Y =
1
Z Z
0
Using 2 and 3, the continuity
of H
y
can be expressed as,
Y
1
(1r ) =Y
2
n
1
k
z2
n
2
k
z1
t .
Aer rearranging, we obtain,
1r =

1
k
z2

2
k
z1
t . (4)
We can summarize the two continuity relations in the following system,

1+r =t
1r =

1
k
z2

2
k
z1
t
Solving for r and t yields the result,
t
s
12
=
2
2
k
z1

2
k
z1
+
1
k
z2
, r
s
12
=

2
k
z1

1
k
z2

2
k
z1
+
1
k
z2
. (5)
1.2 TM-polarised light
For TM-polarised light, the continuity of H
y
reads,
Y
1
(1r ) =Y
2
t
rewritten as,
1r =

1
t . (6)
Note that in the case of normal incidence, we can combine equation 1 and equation 6,
t
12
=
2

1
+

2
, r
12
=

2
+

1
.
assuming here a homogeneous plane wave
2
e continuity of E
y
can be written,
E
i
cos
i
+E
r
cos
i
=E
t
cos
t
and can be expressed in terms of the normal component of the wave-vectors as,
1+r =
n
1
k
z2
n
2
k
z1
t .
We can summarize the two continuity relations in the following system,

1r =

1
t
1+r =
n
1
k
z2
n
2
k
z1
t
Solving for r and t yields the result,
t
p
12
=
2
2
k
z1

2
k
z1
+
1
k
z2

2
, r
p
12
=

2
k
z1

1
k
z2

2
k
z1
+
1
k
z2
. (7)
To summarize, for a single interface from 1 to 2 with normal along the z direction, the Fresnel
coecients read,
r
p
12
=

2
k
z1

1
k
z2

2
k
z1
+
1
k
z2
, r
s
12
=

2
k
z1

1
k
z2

2
k
z1
+
1
k
z2
t
p
12
=
2
2
k
z1

2
k
z1
+
1
k
z2

2
, t
s
12
=
2
2
k
z1

2
k
z1
+
1
k
z2
.
(8)
Note that,
r
i j
=r
j i
. (9)
To verify the conservation of energy, one must consider the irradiance dened in terms of the elec-
tric eld as I =Y |E|
2
, yielding,
R =
I
r
I
i
=|r |
2
, T =
I
t
I
i
=
Y
1
Y
2
|t |
2
. (10)
and we can verify that, indeed, R +T =1 for a single interface.
2 Reectivity of a layer
From the viewpoint of ray optics, a thin layer will support an innite number of internal reections
(absorption and irregularities will however reduce the intensity in a physical situation). e innite
series of reected orders can be expressed in the form a geometric sum, leading to a closed form
formula as shown below. An incident plane wave with amplitude A impinges on the rst interface.
It can be reected, B = r
01
A, or transmitted. e total response of the slab can be obtained by
following each order of reection inside the slab (C, D, ).
Upon transmission at the rst interface, the wave amplitude is t
01
A. Application of Fermats
principle yields a phase change =k
z1
d when the wave hits the second interface. e reection
coecient at this interface is r
12
. e partial wave reected from this path, noted C, is therefore
C =t
10
t
01
r
12
exp(2i k
z1
d)A.
Similarly, in D,
D =t
10
t
01
r
10
r
2
12
exp(4i k
z1
d)A
3
0
1
2
3
N
N+1
A B C D

Figure 2: Schematic of reection and transmission by a multilayer stack. A few reected orders are
noted A, B, C and D for the rst lm.
And, for the j
th
partial wave,
t
10
t
01
r
j
12
r
j 1
10
exp(2i j k
z1
d)A
e wave reected by the slab is the sum of these contributions,
r
slab
A = B +C +D + =

r
01
+t
10
t
01
r
12

j =0
r
j
12
r
j
10
exp(2j i k
z1
d)

A
For clarity, I write =r
12
r
10
exp(2i k
z1
d). e summation of all partial waves is thereby expressed
as a geometrical sum,
r
slab
=r
01
+

t
10
t
01
r
12
exp(2i k
z1
d)

j =0

j
4
Recalling that the sum of a geometric series of common ratio q is
1
1q
, we can write,
r
slab
=r
01
+
t
10
t
01
r
12
exp(2i k
z1
d)
1+r
12
r
10
exp(2i k
z1
d)
We note from equations 1 and 6 and the relation r
i j
=r
j i
that for either polarisation we have
the following identity,
t
s
i j
t
s
j i
=(1r
s
i j
)(1+r
s
i j
) =1(r
s
i j
)
2
, t
p
i j
t
p
j i
=
Y
i
Y
j
(1r
p
i j
)
Y
j
Y
i
(1+r
p
i j
) =1(r
p
i j
)
2
(11)
Using equation 11 and the substitution r
10
=r
01
we nally obtain,
r
slab
=
r
01
+r
12
exp(2i k
z1
d)
1+r
01
r
12
exp(2i k
z1
d)
. (12)
For the transmission, one obtains,
t
slab
=
t
01
t
12
exp(i k
z1
d)
1+r
01
r
12
exp(2i k
z1
d)
. (13)
When N layers are stacked together, the reection coecient of the structure can be found by
applying recursively the preceding formula for a single layer. is amounts to considering one of
the reection coecients to be the eective reection accounting for all the layers behind.
5