Re: TIRF and p- s-polarized incident linght
Posted by
Shigeo Watanabe on
URL: http://confocal-microscopy-list.275.s1.nabble.com/paper-help-tp4115966p4136539.html
John,
Thank you so much for explaining in
a way for me to understand.
Now I am getting to understand better.
I just could not believe the cartwheeling polarized light still, though.
Sincerely,
Shigeo Watanabe
I just remembered another physical analogy that can help
explain the reasons why certain polarizations of light reflect or transmit
through a refractive index interface. Imagine the polarization of light
is represented by a long wooden stick. Now imagine throwing this stick
into a large pool of water at an oblique angle such that its long axis
is oriented paralell to the surface of the water. The stick will
enter the water easily in this case. Now imagine throwing the stick towards
the water at an oblique angle with the long axis perpendicular to the water
surface. I think you can imagine in your mind that as the lower part of
the stick strikes the surface first and begins to move slower in the water,
the top part of the stick still outside the water will begin to rotate
and "cartwheel" towards the surface because of its momentum.
It's a very crude mechanical analogy and there are several things wrong
with it, but I think you get the general similarity with the situation
encountered with light waves. Perhaps someone else out there knows of a
better analogy. I think I remember reading a version of this story with
bouncing sticks on an interface in Richard Feynman's books or in Eugene
Hecht's textbook on optics.
John Oreopoulos
On 2009-12-07, at 11:07 PM, John Oreopoulos <john.oreopoulos@...>
wrote:
Shigeo,
Since the core of my PhD project involves exploiting the
polarization of light in the TIRF mode of illumination, I guess I can try
to explain this. As you know, light can normally be linearly polarized
in any direction that is transverse (perpendicular) to the direction of
travel of the light beam using polarization optics. Polarization here refers
to the direction of oscillation associated with the electric field vector
of the light beam. In the epi-illumination mode for microscopy, the beam
of light is directed straight through the sample along the optic axis of
the microscope objective - call this direction the z-direction. In this
case, this means that the light can be polarized along any direction in
the xy plane (the sample/image plane) - the plane that gets projected onto
your CCD detector. Now consider what happens when you adjust the beam for
TIRF illumination - you set up the beam to impinge on the substrate surface
at an oblique angle instead of going straight through. From Snell's law
of refraction at an interface, you know that the critical angle for a glass/water
interface is ~60 degrees. If you set the beam to impinge the surface
at at the critical angle, the refracted beam that emerges from the interface
travels horizontally (x-direction) along the surface. You can see this
with a bright laser on a TIRF microscope setup and solution of fluorscent
dye. Since the laser beam travels along the x-direction in this situation,
the same polarization rule applies - now the beam can be polarized in the
yz plane depending on the input polarization of the beam before it emerged
from the interface. Here is the tricky part: if you increase the angle
of incidence beyond the critical angle of the interface, total internal
reflection occurs and an evanescent wave/field is created in the lower
index medium (water). As a consequence of Maxwell's electromagnetic field
equations solved at an interface, the polarization characteristics of the
evanescent wave are a bit unusual. These equations have certain boundry
conditions that demand that the transition of light (more specifically
the light oscillation wavefronts) from one medium to another be continuous
for all angles of incidence, and as a consequence, the elecric field vector
of the evanescent wave in both mediums (glass and water) must adjust to
satisfy this condition. This is a general property that is observed in
many physics situations - this idea of continuaty at an interface is the
reason that quantum electron tunneling occurs as well. In fact, the evanescent
wave is the photon analog of electron tunneling. I would suggest reading
Griffeth's textbook Indroductory Electrodynamics for the full details of
this dervation about light interactions at an interface. Anyways, it turns
out that past the critical angle, light that was originally polarized along
the y-axis remains the same ("s-polarization", the s stands for
the first letter of the German word for perpendicular to the plane of incidence)
but light polarized along the z-direction ("p-polarization, the p
stands for the "paralell" to the plane of incidence) for critical
angle illumination becomes split into a linear combination of light polarized
along z and the x-direction (which is the direction of travel of the evanescent
surface wave!). The relative amount of z and x polarization will depend
in the refractive indcies and the angle of incidence. This is one of the
few examples in nature where light can be setup to partially oscilate along
its direction of travel (a longitudinal wave) and this is why the website
you mentioned said that the "p-polarized" evenescent electric
field vector "cartwheels" along the direction of travel.
After reading what I just wrote I see that it still might
be unclear. Working with polarization can be tricky because it is the feature
of light we as humans are most unfamiliar with since our eyes cannot sense/detect
polarization. It's hard to visualize in your mind imaginary field vectors
oscillating and traveling in space. If it's still unclear, try reading
Dan Axelrod's review papers on TIRF microscopy where he discusses the applications
of polarized TIRF illumination. I think there is even a useful Java applet
on the Molecular Expressions website that does a bettter job of depecting
this. Polarized illumination (epifluorescence or TIRF mode) can be useful
for studying the orientations of fluorescent molecules embedded in a sample
(linear dichroism, fluorescence anisotropy, etc.)
John Oreopoulos
On 2009-12-07, at 5:44 PM, Shigeo Watanabe <[hidden email][hidden email]>
wrote:
Dear all,
I have questions about TIRF and polarized light.
Incident light can be devided into p- and s- polarized light.
Then what happen to these two light when evanescent light is induced by
these light.
Olympus webpage explains these, but I could not understand these phrase.
| "A non-zero longitudinal
component and phase lag manifests the p-polarized incident light, which
has an evanescent electric field vector direction that remains in the plane
of incidence. The longitudinal component induces the p-polarized light
electric field vector to "cartwheel" along the interface and
produce elliptical polarization of the evanescent field in the plane of
propagation." |
Olympus web page is here.
http://www.olympusmicro.com/primer/java/tirf/evaintensity/index.html
I appreciate if there are somebody who can explain these to me.
Sincerely
Shigeo Watanabe
Hamamatsu Photonics KK.