Assume a plane wave is incident on a plane interface between two dielectric media. Let the interface lie in the z = 0 plane. Rotate the coordinate system so that the k vector lies in the y = constant planes.
At the boundary, the plane wave will be partially transmitted and partially reflected. Let
Ei(r,t) = Ei exp(i(ki×r-wt)) denote the incident wave,Er(r,t) = Er exp(i(kr×r-wt)) denote the reflected wave,
Et(r,t) = Et exp(i(kt×r-wt)) denote the transmitted wave.
We want to find Er, Et, kr, and kt as a function of Ei and ki. The frequency of the incident, reflected and transmitted wave will be the same.
![]()
For dielectric-dielectric interfaces Maxwell’s equations yield the following boundary conditions.
| The normal component of B is continuous. | |
| The normal component of D is continuous. | |
| The tangential component of H is continuous. | |
| The tangential component of E is continuous. |
Let us consider plane waves in an lih medium. Then Ei ^ ki. We may write Ei = Eip p + Eis s, where p is a unit vector perpendicular to ki in the plane of incidence and s is a unit vector perpendicular to ki and perpendicular to the plane of incidence. The plane of incidence is the plane containing ki and the vector normal to the interface. In our example it is a y = constant plane.
The behavior of a plane wave at a boundary depends on its polarization. We can treat p-polarization and s-polarization separately and then use the principle of superposition to treat arbitrary polarization.
![]()
A wave incident on an interface with p-polarization is called a transverse magnetic (TM) wave. Assume the wave is incident on an interface between medium 1 and medium 2 from medium 1. Then
E1 = Ei + Er, E2 = Et.

Let Ei, Er, and Et denote the magnitudes of the field vectors and assume that they point into the directions shown in the figure above.
_______________________________________
The tangential component of E is continuous at z = 0.
Let
. Then
E2×
= Et×(-
)
= Et exp(i(kt×r-wt))
cos(qt)
E1×
= Ei×(-
)
+ Er×(-
)
= Ei exp(i(ki×r-wt))
cos(qi) – Er exp(i(kr×r-wt))
cos(qr)
At z = 0 we have Etx = Eix + Erx, or
Et exp(i(kt×r)) cos(qt) = Ei exp(i(ki×r)) cos(qi) – Er exp(i(kr×r)) cos(qr),
which must hold for all r in the z = 0 plane.
We therefore need
exp(i(kt×r)) = exp(i(ki×r)) = exp(i(kr×r)), or kt×r = ki×r = kr×r
for all r in the z = 0 plane.
For r =
we then obtain ki sinqi
= kr sinqr = kt sinqt.
Since ki = w/vi = wn1/c,
kr = w/vr = wn1/c,
kt = w/vt = wn2/c,
we obtain
sinqi = sinqr, the law of reflection, and
n1 sinqi = n2 sinqt, the law of refraction or Snell’s law.
_______________________________________
The tangential component of H is continuous at z = 0. Hty = Hiy + Hry
From Maxwell’s equations we have
.
We therefore have
.
Let
=
-
. Then, at
z = 0,
or
.
This equation combined with Et cos(qt) = Ei cos(qi) – Er cos(qr) at z = 0 yields the relation between the amplitudes Et, Ei, and Er.
r12p is called the Fresnel reflection coefficient for p-polarization.
t12p is called the Fresnel transmission coefficient for p-polarization.
Note that the reflection coefficient can be negative. Ei and Er denote the magnitudes of the field vectors assuming that they point into the directions shown in the figure above. A negative reflection coefficient denotes that given the direction of Ei as shown in the figure, the direction of Er must be reversed.
For nonmagnetic materials m1 = m2 = m0.
Defining a = cosqt/cosqi and b = n2/n1 we may then write
r12p = (b-a)/(b+a) , t12p = 2/(a+b).
_______________________________________
For any plane in a medium we define the characteristic wave impedance Z for a plane waves as the magnitude of the ratio of the transverse electric field E|| to the transverse magnetic induction H||. For a TM wave and planes perpendicular to the z-axis, as shown above, the wave impedance at position z is Z(z) = Ex/Hy. (Here transverse refers to directions transverse to the normal and therefore parallel to the plane of the interface.)
Since the tangential components of E and H are continuous for a dielectric-dielectric interface, the wave impedance must be continuous.
For a single TM plane wave propagating in a medium with a wave vector k making an angle q with the z-axis, we have
Z(z) = Ex(z) /Hy(z) = (mc/n) cosq = (k/ew) cosq = (kz/ew) º Z0.
[mc/n = mv = mev/e = 1/(ve) = k/(ew), since v = (me)-1/2 = w/k.]
Z0 is independent of position. It depends only on the properties of the medium and the angel q.
At z = 0 in medium 2 we have, for the TM wave shown above,
Z(0)(2) = (m2c/n2) cosqt = (kt/e2w) cosqt = (ktz/e2w) = Z0(2).
In medium 1 an incident and a reflected wave propagate.
At z = 0 in medium 1 we have, for the TM waves shown above,
Z(0)(1) = (m1c/n1) cosqi(Hiy - Hry)/(Hiy + Hry) = Z0(1)(Hiy - Hry)/(Hiy + Hry).
Here H is the magnetic induction at z = 0.
Since the wave impedance is continuous we have
Z0(1)(Hiy - Hry)/(Hiy + Hry) = Z0(2).
The reflection coefficient r12p is equal to
r12p = Hry/Hiy = -(Z0(2) - Z0(1))/( Z0(2) + Z0(1))
This is an expression of the reflection coefficient r12p in terms of the wave impedance.
![]()
s-polarization
A wave incident on an interface with s-polarization is called a transverse magnetic (TE) wave.
We analyze its behavior at the interface in a similar way as the behavior of a wave with p-polarization.

The laws of reflection and refraction are unchanged. For the reflection and transmission coefficients we now find
.
r12s is called the Fresnel reflection coefficient for s-polarization.
.
t12s is called the Fresnel transmission coefficient for s-polarization.
If m1 = m2 = m0, then
r12s = (1 - ba)/(1 + ba) , t12s = 2/(1 + ab).
_______________________________________
For a TE wave and planes perpendicular to the z-axis, the wave impedance at position z is Z(z) = -Ey(z)/Hx(z).
Since the tangential components of E and H are continuous for a dielectric-dielectric interface, the wave impedance must be continuous.
For a single TE plane wave propagating in a medium with a wave vector k making an angle q with the z-axis, we have
Z(z) = -Ey(z)/Hx(z) = (wm/kz) º Z0.
[wE = -k ´
B . E = (cm/n) H(
).
-Ey/Hx
= cm/(n cosq) = vm/(cosq)
= wm/(k cosq).]
At z = 0 in medium 2 we have for the transmitted TE wave
Z(0)(2) = (wm2/ktz) = Z0(2).
At z = 0 in medium 1 we have for the incident and reflected TE wave
Z(0)(1) = Z0(1)(Eiy + Ery)/(Eiy - Ery)
Here E is the electric field at z = 0.
Since the wave impedance is continuous we have
Z0(1)(Eiy + Ery)/(Eiy - Ery) = Z0(2).
The reflection coefficient r12s is equal to
r12s = Ery/Eiy = (Z0(2) - Z0(1))/( Z0(2) + Z0(1))
This is an expression of the reflection coefficient r12s in terms of the wave impedance.
![]()
The ratio of the reflected intensity to the incident intensity is obtained by squaring the Fresnel reflection coefficient. The reflected and incident wave have the same speed and their wave vectors make the same angle with the normal to the interface.
R = |r12|2 = <S>r×
/<S>i×
º
reflectance
is a unit vector normal to the interface.
The ratio of the transmitted intensity to the incident intensity is
T = (cosqt n1)/(cosqi
n2) ´ |t12|2 = <S>t×
/<S>i×
º
transmittance
We compare the incident and transmitted energy per unit area perpendicular to the z-axis per unit time. Energy conservation requires that in lossless media
T + R = 1.
Link:
Excel Spreadsheet: Reflectance vs. qi (s and p polariztion)