I remember to have seen a setup where the light emitted by the laser was passed through a (pretty long) fiber before being fed into the imaging optics. The fiber was loosely wound up into a coil-like bundle (with sufficiently large bending radius such as to avoid losing total reflectance conditions) and the bundle was somehow shaken mechanically (can't remember the details, but must have been motorized). The ameliorization in imaging was striking.

Steven,

The article below lists a variety of approaches to improve your S/N if that is any use, approx micron resolution susceptibility maps are shown

Magneto-optical Kerr effect analysis of magnetic nanostructures

Allwood, DA (Allwood, DA); Xiong, G (Xiong, G); Cooke, MD (Cooke, MD); Cowburn, RP (Cowburn, RP)

JOURNAL OF PHYSICS D-APPLIED PHYSICS Volume: 36 Issue: 18 Pages: 2175-2182 Published: SEP 21 2003

Yes, Kai is right. We also did it like this and attached the long fiber to the membrane of a loudspeaker, which was then driven by an AC signal. It worked very well.

There is not enough information in your question to understand what are you trying to do in the first place. Please have a look at the following info to help you out.

http://www.study-on-line.co.uk/whoami/thesis/chap2.html

I commonly use MOKE sytems for both purposes. Asumming you are using Kerr Effect not Faraday, then I suggest to use image processing to remove unwanted optical "noise" - aka speckle. A few keywords would provide endless publication in this field. Good luck...........

In the case of imaging using monochromatic (or short spectra) laser source you have problem with the interference effect. It is typical. Good solution is to use simple matte glass or with matte glass mounted on a rotator. Some microscopes (e.g. Zeis, Olympus, Leica) has such kind of glasses.

I agree with the suggestion of Kai. We have tried also an automated "despeckle" unit which is a rotating disk with several optical windows. At certain rpm, you get the same effect. The trick with the loudspeaker is also very good to achieve want you want to have.

I have no experience in this, but you should contact Prof. A. Chyzyk, UPV/EHU.

as an expert in MOKE systems.

Besides the good solutions already mentioned, you can also apply optical (not digital) image processing: You have to mount a space filter into the optical illumination path. A space filter consists of a microscope objective with high numerical aperture that focusses the beam on a pinhole. The pinhole has to be precisesly adjustable. Then a second lens forms a parallel beam again. If the pinhole is small enough the speckles will be gone. The disadvantage of this method is a severe loss of intensity. The advantage is the ideal parallel beam behind the space filter.

Is the laser essential? If you are imaging perhaps a LED source would be better?

Buy a diode for 1$ and you are fine.

Usually I have used oscillating with low sound frequency (40 – 60 cycle per second) polyethylene film.

Thank you all for your suggestions. In regards to a few of the questions that were asked:

1. We use a laser because the Kerr rotation is tiny and the Kerr amplitude through the analyzer is consequently very small. The high power laser ensures that we can measure a signal--we are also very concerned about output power stability.

2. We are imaging using a reflected longitudinal Kerr mode, not Faraday.

We'll try out a rotating diffuser, since we already have an optical chopper head that could be used. We'll definitely play around with image processing as well.

1. We use a laser because the Kerr rotation is tiny and the Kerr amplitude through the analyzer is consequently very small. The high power laser ensures that we can measure a signal--we are also very concerned about output power stability.

Answer: Regarding stabilization of an output laser power - you can use an optical feedback. For this purpose usually use input polarizer - electrooptical modulator - output polarizer, i.e. analyzer. The half-transmitting mirror is sending a part of working beam to the photodiod, and after amplification this electrical signal from photodiod through additional amplifier, this amplified (antiphased) signal is returning to the electrooptical modulator.

I believe that you can also use your PC for the stabilization of an output laser power.

I can well believe this signal would be tiny. So are you scanning the laser to form this image pixel by pixel? From what little I've read about this it doesn't appear that the bandwidth of the light source is really the driving factor. As I recall HeNe lasers are neither high power nor particularly fast as lasers go. Diode sources in the watts both CW and pulsed are cheap. Anyway, just thoughts of a non-expert and probably not a valid one.

I believe, that high power laser diode would be the best solution. Cograts for nice idea of your experimental system! Greetings.

Yes, making the laser beam passing through a spinning plastic or glass disk semi-tranparent (a so called frosted glass) destroys the laser coeherence and, thereby, the speckle. Images are not as nice as using a standard incoherent source, but good enough to "see" the magnetic contrast due to the domains.

Paolo Vavassori's answer is correct, in interferometry experiments where optical fibres can be used speckle is removed by continously mechanically modulating the optical fibre (a mode scrambler).

Laser+scrambler will definitely works but this makes the source incoherent. I wonder if using a high power red LED might not be better because there is no need for scrambling and there are definitely LEDs that woul fit your power requirements, especially if you use a lens in a parabolic reflector, you can make a narrow spot over a short distance (< 1 m) that should be efficient enough. Particularly, i'm thinking that a 700 mW red LED (Thorlabs M625D2, 50$) would be a good solution if the scrambled laser is not incoherent enough.

To remove speckles, you have to deminish time coherence. One solution, apply phase modulation (piezomirror eg).

We have good experience with Optotune's speckle reducer. Details at

http://www.optotune.com/images/products/Optotune%20LSR-3000%20Series.pdf

Prodcut also sold by Edmund Optics.

If you want to keep the laser, you can get optical diffuser from throlabs or newport and attached to the spinning moter. That will remove speckles. But there are moving parts and it will generate some trace as well.

I recommend to use a non-coherent source which has no coherence (Arc lamp or high power LED). Those have high power that typical He-Ne laser.

I agree, the speckle reducer will be better or e.g. the second easy solution - rotating the opal glass (matte).

It is necessary use an oscillating with low frequency the plate of thick transparent polyethylene installed along the light beam.

I have been facing the same problem with my MOKE microscope set-up and after changing the laser by an LED the speckles were removed.

The cheapest solution is to replace the laser by a diode!! You remove the phase coherency and you have a robust and cheap system, and you can use the laser for something else. 

Greetings

I intend to see the domain images of Fe bulk, NiFe thin films. I built a moke microscope in which there is a perpendicular incidence of light on sample in longitudnal configuration.  How one can be sure that the images, one is getting, are actually the domain images? if my question seems vague, i can share some images i have taken with my setup.

Regards