Ouput Polarization of Corner Cube

The polarization is changed due to the action of the beam splitting coating. If you model the coating, I would expect even Zemax will show this change.

Please visit my web site at

Which coating? A corner cube is TIR, only AR coating on the surface.

--- sam | Sci.Electronics.Repair FAQ:

Repair | Main Table of Contents:

+Lasers | Sam's Laser FAQ: | Mirror Sites: Important: Anything sent to the email address in the message header above is ignored unless my full name AND either lasers or electronics is included in the subject line. Or, you can contact me via the Feedback Form in the FAQs.

I forgot to mention that the corner cube modeled in ZEMAX is a three perpendicular reflective (mirrors) surfaces object, there are no coating on them.

Sam Goldwasser wrote:

"Farsang" wrote >

Yes

Most likely... but when?

It's a recurring problem which has been extensively treated in the literature.

The corner cube will "pizza slice" the output beam.

My approach if I were to do it from scratch would be to use the Jones matrix calculus and the Fresnel reflection coefficients.

This assumes that you know the values of the angles of incidence on each face. I leave to you as an exercise. ;-)

I believe that somewhere Jim Wyant (College of Optics, formerly OSC, U of A) has treated the problem in its generality and in the context of the interferometric testing of corner cubes.

Following the above recipe it's very hard to do.

No, there's a significant change of polarization on TIR even with a clean glass/air interface. The s polarized light sticks out further into the low-index medium, so its dE/dz is different at the surface, leading to a different value of phase shift between the incident and reflected beams.

I sometimes use this effect as a phase vernier in corner-cube interferometers--you just rotate one cube around its 3-fold symmetry axis. Works up to the point where one of the cube edges starts to cross the beam.

Cheers,

Phil Hobbs

A quick look shows that your uncoated alum reflector has about a 5.5% difference in S&P reflectivity at a 45 degree AOI (I know its compound angle) with the S being higher, this is for one reflection, so it would be about 16% difference on the output of S & P with S being higher.

The phase difference was about 13 degrees with S being higher again on one reflection.

I agree with Richard it should be fine in Zemax wth polarization modeling.

Phil, a phase vernier with tilt angle, very cool - thanks for the mind candy!

Michael

A cube corner (erroneously called a corner cube) will change the polarization of outgoing light from that of the ingoing light. That even is true for a metal corner.

Those corners using total internal reflection modify the polarization because the p and s components totally reflect with differing phase shifts. As you look into the cube, each pair of circular segments provide a different pair of what I will call pseudo polarization states that are orthogonal. These are elliptical polarizations. They are described in an old JOSA paper.

A metallized corner will also display a change in polarization. For a perfect metal, it will be the same for all the sector pairs. The polarization change is "geometrical." It has to do with how vibration directions are defined rather than with the physics of reflection.

A simple way to see that is with a metallized roof prism. Take one of them and place a sheet of polarizing material such as Polaroid over the hypotenuse face. As you rotate the polarizer, you will see lightening and darkening through the polarizer. This effect is geometrical.

Bill

-- Ferme le Bush

"Berry's phase" or "Pancharatnam's phase". It happens when you make the k vector describe a nonplanar figure. The phase shift is equal to the included solid angle of the path, which is a very non-obvious fact. The corresponding effect in quantum mechanics is responsible for some spooky-looking effects, but as you say, it's just geometry.

Cheers,

Phil Hobbs

I must admit that I never heard of Berry's or Pancharatnam's phase. The geometrical phase I describe wrt metallized mirrors is probably much simpler. It does not require complex numbers as would be necessary for total internal reflection. It has more to do with sign convention for multiple reflections for p and s waves.

Bill

-- Ferme le Bush

"Charles Manoras" wrote

snip

For s and p polarization on each face.

Depending the polarization state of the input beam (i.e. partially polarized) the Mueller calculus may be more appropriate.

Mueller matrices may be easily derived from the Jones matrices.