3D pollen - structured illumination microscopy

[Tardigrade37]

At some point, I will elaborate on my new setup, but I was excited to post my first images using the Zeiss Apotome structured illumination system I have acquired. This is a standard mixed pollen slide prepared by Carolina Scientific, so I don't know the origin of the pollen, but these slides are commonly used as test slides for fluorescence microscopy.

Excitation is via an EXFO Excite-120 metal halide and a Rhodamine filter cube. The objective lens is a Planapo 63x/1.4 and the camera used is a Hamamatsu ORCA-AG.

The Apotome system is essentially a grid that is placed in the illumination path that reduces out of focus fluorescence - in other words, a combination of acquisition settings and an algorithm to make sense of the images allow for an optical section of a fluorescence sample to be acquired. The result is very similar to (and some may argue, better than) images acquired with a confocal laser scanning microscope; although, it is considerably less expensive.

The image below is a volume rendering of 72 images. If you click on the image, you will be brought to a movie of the pollen spinning around the (semi)-central axis; the movie is 17MB.

See here for more info on the Apotome system.

Definitely more to come!

[rjlittlefield]

Two thumbs up, or maybe that should be four or more, 'cuz I'm very impressed by this result and everybody who contributed to it.

The still image is good, but that movie is wonderful!

--Rik

[Ken Ramos]

That is really wild! I would like to have something like that but dang if I'd know how to use it.

[Charles Krebs]

Amazing stuff!

About how long did it take to record a single image? (From the little I read at the link you provided this would be three exposures with the "grid" shifting for each.. is that correct?)

[Tardigrade37]

Thanks all. Yes Charles, each "single" Apotome image is a processed set of three images. I also used a 3 image average to reduce noise, so all told, the data set was 648 images. According to Nyquist, I undersampled this at every 500nm in Z, so the data set should have been twice as big for optimal resolution in Z (but I was impatient and didn't even know if this would work).

As far as time, everything on my scope is automated, so the upper and lower limits of focus are set, and the scope does the rest. The exposure time for each image was 51ms, so total acquisition time for a single optical section is around 150ms, then averaged 3 times = 450ms. The processing algorithm works on the fly, so there is a fair bit of overhead involved. For this data set, it took roughly 7 minutes from start to finish.

I'll post some images later that include "grid images", complete optical sections, and a comparison to normal fluorescence.