Experimental Polarization Techniques

[reasonableman]

I am currently doing experiments with polarizing optics and reflection from metal surfaces. However it's not as simple as it sounds in textbooks, so I was wondering if anyone had experience in these type of experiments.

There seems to be a fundamental problem with identifying the transmission axis of a polarizer. I've tried using a reflection from a surface at the Brewster angle but it's not that accurate, I've also used a method of reflecting from a metal. This gives answers to a high accuracy (mainly because it has curve fitting analysis), however isn't supported by simpler experiments (eg; calibrate 2 polarizers with this method and cross them).

I'd currently feel comfortable saying that I know the axis to 0.5 degrees accuracy however in the literature much higher accuracy is quoted.

Also the polarizers I'm using are quoted by the manufacturer to have an extinction ratio of <10,000. This doesn't seem to be what I get...

Can anyone help with these problems?

 

[Andy Resnick]

It's not clear what you are doing: do you have a polarized source? What kind of polarizers are you using? What's the bandwidth of illumination?

 

[reasonableman]

Thanks for the reply!

The ultimate aim is to calibrate the scales on rotating mounts that polarisers have been mounted in.

The source is a fibre-diode laser with a collimator on the output PM fibre. Bandwidth is very small, laser is easily within +/- 1 nm of the 1550 nm. It is reasonably polarised when it is emitted from the collimator.

I have 1 Glan-thompson (GT) and 2 film polarisers.

It was thought it would be easy:

Pass the emitted laser light through the GT, then calibrate the film polarisers against the GT by finding the extinction caused by crossing the GT and film polariser.

However it is not that easy...noise on the laser power/power meter makes finding the minimum difficult, even plotting around the minimum and curve fitting does not seem particularly repeatable. Then if you try testing the result by trying to find maximum transmission it is not at 90 degrees!

 

[Andy Resnick]

Oh- ok. You almost have it.

Rather than try to find the maximum or minimum transmission orientations, it's best to find the 45-degree inclinations: the point at which small rotations cause equal and opposite changes in the transmission. Less noise, better signal, more accuracy.

Let me know if that works better.