For materials with large Kerr effect, MOKE microscopy images can be clearly seen in polar geometry. However, for materials with small effect, it is difficult for see the domains even with the division or subtraction of a reference image (usually at saturation).
Did anyone try lock-in amplifier or other techniques to improve the images contrast ?
MOKE microscopy is suitable for imaging the dynamics of magnetic domains. That's why we are interested in this technique. MFM for example has a good resolution but no chance to see domain movement.
From my experience it is very important to use good polarizers with high extinction ratio. You need a stable illumination source in order to keep the same light distribution for all images. Further, you have to use a least 2 image processing steps:
1. Subtraction of a reference image (saturated state)
2. Normalization of the difference image by division by the reference image (this step eliminates the remaining inhomogeneous light distribution).
Finally the contrast of the normalized image can be increased. In general, it is important to optimize the angular position of the analyzer. Therefore, I recommend to repeat this procedure with different angular positions of the analyzer until you get the maximum contrast.
Thanks Stephan. We substracted the images from a reference one but I think we need to work more on the angle between the polarizer and the analyzer.
There is another option. You can significantly increase the contrast by an antireflective coating. Obviously this can only be used if this coating doesn’t influence the magnetic layer.
I think it is too hard: garnet film has thickness about 10 micron, Lock-in can improve of a signal up one order (for laser beam), dielectric coating about 50%. It was experiment there MFM is combined with Kerr microscope - it possible to see movement of domain wall under field of tip (Aist microscope as I remember).
May be help You www.evico-magnetics.de
Dear Rachid,
Instead of subtracting saturated state from the domain image, one can introduce a lambda quarter plate just before the analyzer. Domain contrast can be flipped rotating this plate. Then try subtracting one state from the other.
I have built a moke microscope but i am having trouble with image processing. I am trying to see domain of Bulk Iron (Fe) in longitudnal configuration. When i subtracted (Using ImageJ) the image(B=0) from the image (B saturated), i simply gets a dark image. No contrast. If anyone have a suggestion regarding this or moke microscope, it would be highly appreciated.
Dear Ali,
Download and install 'Calculator Plus' plugin into the ImageJ program. Then upload both saturated and domain images. Use 'Divide' function while i1 is background and i2 is domain. Let k1 to be mean value of background.
Please let me know if you manage.
Good luck.
Thank you Umut.
I used calculator plus but still nothing. Could it because of the angle of incidence is very small in my case? If AOI is zero then no longitudnal moke occurs. How one can overcome this problem without major changes in orientation of the setup? In my scheme, light after beam splitter, goes through objective to the sample perpendicular. Please see the rough diagram of my scheme as attachment. Any suggestion why i am unable to get domain images of Bulk iron.
Sometimes adding a lambda quarter plate before analyzer helps. That would compensate any ellyptical polarization effects. Besides, you may replace your objective with a higher numerical aperture objective.
Adding to Umut Parlak:
* High quality polarizer and analyzer can definitely improve the MOKE/Imaging signal
* Quarter wave plate (before analyzer) is necessary in most cases, since you are using other optical elements like objective lenses or prism.
* The quality of camera can also be a bottle neck. Camera with high bit resolution is necessary (12 bit preferred, a cooled camera). The camera should have very low noise floor or dark current (should give you close to zero intensity under no light condition).
* Before checking for images, try to see if you can optimize the setup (angles between polarizer -QW plate - analyzer ) for obtaining a regular MOKE signal (average intensity of images versus magnetic field sweep). If you get a hysteresis loops as expected for the sample under investigation, then you can try image processing tools. (Try hysteresis measurements with background image subtraction process.)
In longitudinal MOKE, the angle between incident and reflextion beam should be lager than 0. We can do this by shifting the beam splitter a litle and use small magnefication lens.
After getting images from MOKE micoscopy, is there any efficient way to calculate the area of circular ( expanded)domain waals? I tried by using edge (Canny/ Prewitt) function in MATLAB before finding area? But it gave no distinctive edges!!
Dear Babu, I would employ an image processing software like ImageJ. If the image quality allows, you can basically binarize it and then calculate the area using build-in commands.
Is there any precise way/tool for the calculation of ( very small distance) the expansion of the elliptic / asymmetric domain wall images obtained from MOKE micoscopy?
- Badges
- Science topic
- Similar topics
- Chemistry
- Nanotechnology
- Nanostructures
- Thin Films