While testing the through focus mechanism on my new (under development) CNC microscope stage I found that I get this peculiar rippling effect while moving in Z with this silicon chip specimen - it looks like the sample is under water in a swimming pool and you are looking at it from above with the ripples on the water distorting the image (but the distortions here are more regular and periodic than water ripples). See this video clip of what I mean (ignore the initial vibrations, that’s just me bashing the keyboard to get the motor started):
I conducted a series of experiments to determine the cause and my conclusion is that the ripple effect is essentially due to an aliasing / moire effect of the regular fine straight edges of the silicon chip specimen on the Bayer grid of the camera (the camera here is a single chip colour CMOS with an overlying Bayer and the microscope is using coaxial epillumination via a beam splitter - it is the standard PUMA setup for metallurgical analysis with a polariser in the illumination path but no analyser, using a x40 dry objective). It also appears to have a component related to the de-Bayering algorithm. For example, the effect is reduced or not noticeable in all of these circumstances:
- the camera is used with colour saturation set to zero in the camera firmware (grey scale output)
- a monochrome CCD is used with the same silicon chip specimen.
- the original colour camera is used with an irregular specimen (tissue histology) in transillumination
So my questions are: Have any of you also noticed such an effect? Can you confirm that this is an aliasing / moire effect (which is my conclusion) and are there any published works on this you can point me to (I would be surprised if I was the first to notice this. I normally image tissue histology irregular structures - not silicon chips - so this is new to me)?
Thanks.
PJT