@Isbjoern

I’ve been doing some Jones calculations with OS lately, but I’m not an expert at all. What’s your matrix definition for the retarder (10° shift)?

I’m using this formula, from this article:

To avoid calculation errors, I recommend letting OS compute the matrix product, i.e. define one Jones Matrix surface per matrix you have (1 quarter-wave plate, 1 retarder, 1 polarizer or analyzer). Also, I prefer to just have a dummy surface where I can change the parameter of interest, such as φ in Eq. 19 above, and let OS calculate the matrix using ZPL and pickup solves. Here is how I would setup the retrarder:

And I have the following ZPL solves on the A real/imag columns:

CONS_PI = 2 * ACOS(0)

phi_deg = THIC( SURC("PHASE SHIFT IN DEG ON THICKNESS") )
phi_rad = phi_deg / 180 * CONS_PI

SOLVERETURN COSI( phi_rad / 2 )

CONS_PI = 2 * ACOS(0)

phi_deg = THIC( SURC("PHASE SHIFT IN DEG ON THICKNESS") )
phi_rad = phi_deg / 180 * CONS_PI

SOLVERETURN SINE( phi_rad / 2 )

It is not the most optimal in terms of computation (you could do the deg to rad conversion once elsewhere), but I hope you get the point.

It seems to work for me. When the Thickness of Surface 1 is 0 (or 0°):

When the Thickness of Surface 1 is 10 (or 10°):

The retardance measured with CODA in the Merit Function also gives 10°:

And the Jones Matrix looks like so:

Last confirmation was a retardance of 90° since we expect circular polarization:

I hope this helps.

Take care,

David

@David.Nguyen 

Dear David,

indeed I used the same formula. The approach to separate both retardances in Zemax is doing the trick.

Interestingly, when I matrix multiply the Jone matricies of lambda/4 and 10° mathematically and put the result into Zemax I get a wrong result. The stokes parameter “polarization degree” (S2) is bigger than one.
When I put both retardances seperately into Zemax it works. I guess Zemax does some normalization in the background that it finally results in a polarization degree (S2) below 1. This is unexpected it should, from mathematical point of view, work both ways.
Anyhow, thank you for pointing out the solution!

@Isbjoern

No worries, its good to know that the product doesn’t work when done outside OS though. I’ve made a sort of polarimeter recently, although not with a quarter-wave plate and in the context of polarization-resolved SHG. DM me if you’re interested to share about your experience making such a device.

Take care,

David

Sorry, I marked my post as answer. It should have been your post. Do you know how to change it? You should get the credit for it.

@Isbjoern 

Don’t worry about it. As far as I know, only Ansys personnel can do it. Your answer also has meaningful information in it

Have a great weekend,

David