PHYS 415: OPTICS Polarization (from Trebino’s lectures)
description
Transcript of PHYS 415: OPTICS Polarization (from Trebino’s lectures)
www.bilkent.edu.tr/~ilday
PHYS 415: OPTICS
Polarization (from Trebino’s lectures)
F. ÖMER ILDAY
Department of Physics,Bilkent University, Ankara, Turkey
www.bilkent.edu.tr/~ilday
45° Polarization
0
0
( , ) Re exp[ ( )]
( , ) Re exp[ ( )]
x
y
E z t E i kz t
E z t E i kz t
Here, the complex amplitude, E0 is the same for each component.
So the components are always in phase.
~
www.bilkent.edu.tr/~ilday
Arbitrary-Angle Linear Polarization
0
0
( , ) Re cos( )exp[ ( )]
( , ) Re sin( )exp[ ( )]
x
y
E z t E i kz t
E z t E i kz t
Here, the y-component is in phase
with the x-component,
but has different magnitude. x
y
E-field variation over time (and space)
www.bilkent.edu.tr/~ilday
Define the polarization state of a field as a 2D vector—
Jones vector—containing the two complex amplitudes:
For many purposes, we only care about the relative values:
(alternatively normalize this vector to unity magnitude)
Specifically:
0° linear (x) polarization: Ey /Ex = 0
90° linear (y) polarization: Ey /Ex =
45° linear polarization: Ey /Ex = 1
Arbitrary linear polarization:
The Mathematics of Polarization
sin( )tan( )
cos( )y
x
E
E
1
x
y
yx
x
EE
E
EE
EE
www.bilkent.edu.tr/~ilday
Circular (or Helical) Polarization
The resulting E-field rotatescounterclockwise around the k-vector (looking along k).
0
0
( , ) cos( )
( , ) sin( )x
y
E z t E kz t
E z t E kz t
Here, the complex amplitude of the y-component is -i times the complex amplitude of the x-component.
So the components are always 90° out of phase.
0
0
( , ) Re exp ( )
( , ) Re exp ( )
x
y
E z t E i kz t
E z t iE i kz t
Or, more generally,
www.bilkent.edu.tr/~ilday
Right vs. Left Circular (or Helical) Polarization
The resulting E-field rotatesclockwise around the k-vector (looking along k).
0
0
( , ) cos( )
( , ) sin( )x
y
E z t E kz t
E z t E kz t
Here, the complex amplitudeof the y-component is +i times the complex amplitude of the x-component.
So the components are always90° out of phase, but in the other direction.
x
yE-field variation over time (and space)
kz-t = 0°kz-t = 90°
0
0
( , ) Re exp ( )
( , ) Re exp ( )
x
y
E z t E i kz t
E z t iE i kz t
Or, more generally,
www.bilkent.edu.tr/~ilday
Unequal arbitrary-relative-phase components yield elliptical polarization
The resulting E-field can rotate clockwise or counter-clockwise around the k-vector (looking along k).
x
yE-field variation over time (and space)
0 0x yE E
where and are arbitrary
complex amplitudes.
0
0
( , ) Re exp ( )
( , ) Re exp ( )
x x
y y
E z t E i kz t
E z t E i kz t
0
0
( , ) cos( )
( , ) cos( )x x
y y
E z t E kz t
E z t E kz t
Or, more generally,
0 0 .x yE Ewhere
www.bilkent.edu.tr/~ilday
Circular polarization has an imaginary Jones vector y-component:
Right circular polarization:
Left circular polarization:
Elliptical polarization has both real and imaginary components:
We can calculate the eccentricity and tilt of the ellipse if we feel like it.
The mathematics of circular and elliptical polarization
1x
y
EE
E i
/y xE E i
/y xE E a bi
/y xE E i
www.bilkent.edu.tr/~ilday
When the phases of the x- and y-polarizations fluctuate, we say the light is unpolarized.
where x(t) and y(t) are functions that vary on a time scale slower than1/, but faster than you can measure.
The polarization state (Jones vector) will be:
As long as the time-varying relative phase, x(t)–y(t), fluctuates, the light will not remain in a
single polarization state and hence is unpolarized.
( , ) Re exp ( )
( , ) Re exp ( )
x x x
y y y
E z t A i kz t t
E z t A i kz t t
1
exp ( ) ( )yx y
x
Ai t t
A
In practice, the amplitudes vary, too!
www.bilkent.edu.tr/~ilday
Light with very complex polarizationvs. position is also unpolarized.
Light that has passed through “cruddy stuff” is often unpolarized for this reason. We’ll see how this happens later.
The polarization vs. position must be unresolvable, or else, weshould refer to this light as locally polarized.
www.bilkent.edu.tr/~ilday
Birefringence
The molecular "spring constant" can be different for different directions.
www.bilkent.edu.tr/~ilday
Birefringence
The x- and y-polarizations
can see different refractive
index curves.
www.bilkent.edu.tr/~ilday
Uniaxial crystals have an optic axis
Light with any other polarization must be broken down into its ordinary and extraordinary components, considered individually, and recombined afterward.
Uniaxial crystals have one refractive index for light polarized along the optic axis (ne) and another for light polarized in either of the two directions perpendicular to it (no).
Light polarized along the optic axis is called the extraordinary ray, and light polarized perpendicular to it is called the ordinary ray. These polarization directions are the crystal principal axes.
www.bilkent.edu.tr/~ilday
Birefringence can separate the twopolarizations into separate beams
Due to Snell's Law, light of different polarizations will bend by different amounts at an interface.
no
ne
o-ray
e-ray
www.bilkent.edu.tr/~ilday
Calcite
Calcite is one of the most birefringent materials known.
www.bilkent.edu.tr/~ilday
Polarizers take advantage of birefringence, Brewster's angle, and total internal reflection.
Here’s one approach:
Combine two prisms of calcite, rotated so that the ordinary polarizationin the first prism is extraordinary in the second (and vice versa).
The perpendicular polarization goes from
high index (no) to low (ne) and undergoes total internal reflection, while the parallelpolarization is transmitted near Brewster's angle.
www.bilkent.edu.tr/~ilday
Polarizers
Air-spaced polarizers
www.bilkent.edu.tr/~ilday
Wollaston Polarizing Beam Splitter
The ordinary and extraordinary rays have different refractive indicesand so diverge.
The Wollaston polarizing beam splitter uses two rotatedbirefringent prisms, but relies only on refraction.
www.bilkent.edu.tr/~ilday
Dielectric polarizers
A multi-layer coating (which uses interference; we’ll get to this later) can also act as a polarizer.
Ealing Optics catalog
Melles-Griot catalog
Glass
www.bilkent.edu.tr/~ilday
Wire Grid Polarizer
The light can excite electrons to move along the wires, which thenemit light that cancels the input light. This cannot happen perpen-dicular to the wires. Such polarizers work best in the IR.
Polaroid sheet polarizers use the same idea, but with long polymers.
Input light contains both polarizations
www.bilkent.edu.tr/~ilday
Wire grid polarizer in the visible
Using semiconductor fabrication techniques, a wire-grid polarizer was recently developed for the visible.
The spacing is less than 1 micron.
www.bilkent.edu.tr/~ilday
The Measure of a Polarizer
The ideal polarizer will pass 100% of the desired polarization and 0% of the undesired polarization.
It doesn’t exist.
The ratio of the transmitted irradiance through polarizers oriented parallel and then crossed is the Extinction ratio or Extinction coefficient. We’d like the extinction ratio to be infinity.
Type of polarizer Ext. Ratio Cost
Calcite: 106 $1000 - 2000
Dielectric: 103 $100 - 200
Polaroid sheet: 103 $1 - 2
0° Polarizer
0° Polarizer
0° Polarizer
90° Polarizer
www.bilkent.edu.tr/~ilday
Wave plates
When a beam propagates through a birefringent medium, one polarization sees more phase delay than the other.
This changes the relative phase of the x and y fields, and hence changing the polarization.
Polarization state:
0
0
( , ) Re exp
( , ) Re exp
x
y
E z t E i kz t
E z t E i kz t
1
1
exp( )
exp( )
1
2exp
o
e
o e
ikn d
ikn d
i n n d
Output:
}
}
0
0
( , ) Re exp
( , ) Re exp
o
y e
xE z t E i kz t
E
kn d
knz t E i dkz t
Wave plate
+45° Polarization
Optic axis
-45° Polarization
x
y
z
Input:
www.bilkent.edu.tr/~ilday
Wave plates (continued)
Wave plate output polarization state:
1
2exp o ei n n d
0 1 45 Linear/2 Left Circular
1 45 Linear3 /2 Right Circular2 1 45 Linear
Output2 2 exp
Polarization Stateo e o e
i
i
n n d i n n d
A quarter-wave plate creates circular polarization, and a half-wave plate rotates linear polarization by 90.
We can add an additional 2m without changing the polarization, so the polarization cycles through this evolution as d increases further.
Quarter-wave plate
Half-wave plate
(45-degree input polarization)
www.bilkent.edu.tr/~ilday
Half-wave plate
If the incident polarization is +45° to the principal axes, then the output polarization is rotated by 90° to -45°.
When a beam propagates through a half-wave plate, one polarizationexperiences half of a wavelength more phase delay than the other.
www.bilkent.edu.tr/~ilday
Wave plates and input polarization
Remember that our wave plate analysis assumes 45° input polarization relative to its principal axes. This means that either the input polarization is oriented at 45°, or the wave plate is.
Wave plate w/ axes at 0° or 90°
±45° Polarizer
Wave plate w/ axes at
±45°
0° or 90° Polarizer
If a HWP, this yields 45° polarization.
If a QWP, this yields circular polarization.
±
If a HWP, this yields 90° or 0° polarization.
If a QWP, this yields circular polarization.
www.bilkent.edu.tr/~ilday
How not to use a wave plate
If the input polarization is parallel to the wave plate principal axes, no polarization rotation occurs!
Wave plate w/ axes at 0° or 90°
Wave plate w/ axes at
±45°
0° or 90° Polarizer ±45° Polarizer
This arrangement can, however, be useful. In high-power lasers, we desire to keep the laser from lasing and then abruptly allow it to do so. In this case, we switch between this and the previous case.
www.bilkent.edu.tr/~ilday
Thickness of wave plates
When a wave plate has less than 2 relative phase delay, we say it’s a zero-order wave plate. Unfortunately, they tend to be very thin.
Solve for d to find the thickness of a zero-order quarter-wave plate:
2
2o en n d
4 o e
dn n
Using green light at 500 nm and quartz, whose refractive indices are ne – no = 1.5534 – 1.5443 = 0.0091, we find:
This is so thin that it is very fragile and very difficult to manufacture.
d
d = 13.7 m
www.bilkent.edu.tr/~ilday
Multi-order wave plates
A multi-order wave plate has more than 2 relative phase delay.
We can design a twentieth-order quarter-wave plate with 20¼ waves of relative phase delay, instead of just ¼:
240
2o en n d
4141
4 zero ordero e
d dn n
This is thicker, but it’s now 41 times more wavelength dependent!
It’s also temperature dependent due to n’s dependence on temperature.
d
d = 561 m
www.bilkent.edu.tr/~ilday
A thick zero-order wave plate
The first plate is cut with fast and slow axes opposite to those of the second one.
The Jones vector becomes:
1 2 1 2
1 1
2 2 2exp expo e e o o ei n n d i n n d i n n d d
First plate Second plate
Now, as long as d1 – d2 is equal to the thin zero-order wave plate, this optic behaves like the prohibitively thin one! This is ideal.
Inputbeam
Optic axesx
y
zOutputbeam
d1 d2
www.bilkent.edu.tr/~ilday
Polarization Mode Dispersion plagues broadband optical-fiber communications.
Imagine just a tiny bit of birefringence, n, but over a distance of 1000 km…
Newer fiber-optic systems detect only one polarization and so don’t see light whose polarization has been rotated to the other.
Worse, as the temperature changes, the birefringence changes, too.
1
2exp i n d
Distance
Polarization state at receiver
=
If = 1.5 m, then n ~ 10-12 can rotate the polarization by 90º!
www.bilkent.edu.tr/~ilday
Circular polarizers
A circular polarizer makes circularly polarized light by first linearly polarizing it and then rotating it to circular. This involves a linear polarizer followed by a quarter wave plate
Unpolarized input light
Circularly polarized light
QWP45°
Polarizer
45° polarized light
Quarter wave plate (QWP)
±45° Polarizer
Additional QWP and linear polarizer comprise a circular "analyzer."
45° PolarizerQWP
-45° polarized light
www.bilkent.edu.tr/~ilday
Polarization Spectroscopy
The 45°-polarized Pump pulse re-orients molecules, which induces some birefringence into the medium, which then acts like a wave plate for the Probe pulse until the molecules re-orient back to their initial random distribution.
0 polarizer
90 polarizer
Yellow filter (rejects red)