High losses on 445nm wave plate/PBS setup

[Solarfire]

Is anyone else getting high losses on the 445nm wave plates?
With my diode being driven @ 1000mW I’m getting:
1192mW after the collimator
1104mW after the PBS = 7% loss, this is ok
875mW after the wave plate/PBS combo = 27% loss, 20% loss on the WP appears a little much.

[andythemechanic]

Hi,

Try to measure transmission trough the WP alone, to see if it is a coating problem. If Transmission is ok, it is likely that the plate is not l/2 but something a bit off from that. Waveplates are hard to align, and the wavelength of the diode also has to be within ~1nm of the design wavelength. Especially the tilt angle of the plate is very important to really get l/2. The tilt tolerance is inversely proportional to the order of the plate so if you have some high order plate this could cause trouble. To really get optimum alignment you need fine adjustment on rotation and tilt in both axes.

Andreas

[Solarfire]

Hi Andreas,

I measured the loss of the WP alone and it's <2.1%. This is supposedly a zero order 1/2 WP for 445nm, which raises the question of how well can a WP function with these 445nm multimode diodes if they are off ±2-3nm to the design wavelength of the WP. It’s hard to find any documentation in this direction concerning tolerances of wavelength, tilt and rotation. Tilt and rotation may not be spot on at the moment but not significantly off to justify a 20% loss. At the moment I’m tending more in the direction of the WP not being up to the task or the design wavelength being off. For example, would a WP for 457nm work to an extent with 445nm diodes?

Frank

[andythemechanic]

Hi Frank,

concerning wavelength tolerances I just did a quick calculation and this is the result:
The relative retardance error of the plate D_phi/pi=(2m+1)*D_lambda/lambda
where D_phi is the retardance eroor D_lambda the wavelength deviation, lambda the design wavelength and m the order of the plate. So for a zero order plate the relative phase error is proportional to the relative wavelength deviation, which shouldn't be so bad.
Are you sure that it is really a zero order plate? A zero order plate for 450nm is supposed to be only ~22 micron thick.
What also could be is that your crystal axis is not where you think it is. Did you rotate the plate at least 180 degrees to dtermine the best rotation angle?

Andreas

[absolom7691]

Andreas, you have officially fried my brain! This is good stuff to know, since I just got a 635~640nm WP. I am glad to know that there is more to it than shoving it in front of the laser. I have never messed with WPs before. Frank, I hope you get it sorted out

[Solarfire]

Here’s the data of these wave plates..

Material: Quartz-Crystal
Dia.: 12.50mm
Thickness: 0.5mm
Retardation：1/2 Lambda or Type: 1/2
Surface quality: 20/10
Parallelism: 1 sec
Retardation-tolerance lambda/300
Transmitted wavefront disortion: Lambda/8
BBAR: 430nm-670nm R<0.5%, 445nm, 532nm

I didn’t find any info’s explicitly saying that these are zero order WP, but it was explicitly said to not be multi order which leads me to believe that these are zero order WPs. Probably not true zero order WP.

[andythemechanic]

Ok, at this thickness it cannot be a zero order plate. Possibly it is a glued zero order plate on top of a thicker substrate? Should be svisible at the edge if there is something glued on top. Otherwise it is of higher order which would lead to a high wavelength sensitivity.

[planters]

This is similar to the problem I was having with my 445 module some months back. There were a few threads, I can't remember them all, that danced around the wave plate vs. PBS. Some were trying non-orthogonal orientations of the cube. What I found, was the wave plate was not the culprit. I tried the plates from OSLS and Lasershowparts and even a very expensive one from Tower optical and the differences were small, but the PBSs that are widely available are BROADBAND and often optimized for 405nm. Once I received a narrow band ( low energy product group) PBS optimized for 448 nm from JML optical approx $100 for the 12.5mm cube; night and day. Total loss for both polarization combined -2%.

[Solarfire]

This is similar to the problem I was having with my 445 module some months back. There were a few threads, I can't remember them all, that danced around the wave plate vs. PBS. Some were trying non-orthogonal orientations of the cube. What I found, was the wave plate was not the culprit. I tried the plates from OSLS and Lasershowparts and even a very expensive one from Tower optical and the differences were small, but the PBSs that are widely available are BROADBAND and often optimized for 405nm. Once I received a narrow band ( low energy product group) PBS optimized for 448 nm from JML optical approx $100 for the 12.5mm cube; night and day. Total loss for both polarization combined -2%.

Hmm.. I partially followed some of those threads but I can’t recall a wave plate in the discussions. I don’t quite understand how that could be, due to the PBS working perfectly on its own in both instances of transmission and reflection with no more than 7% loss. I don’t doubt this, but it would be nice to understand it though..