Good evening,
unfortunately, there is going to be some physics to answer this question by
Dr. Zucker. I beg your pardon.
>This is an Interesting observation with a question for the confocal user
>group
>
>I was trying to do Huygens deconvolution ( Bitplane software) to increase
>the resolution of some confocal images.
Just as a comment, it's not the main question which is going to be answered
below:
In fact, you do NOT increase resolution by means of deconvolution.
Resolution is not a matter of software, it is something which is inherent
to your optical system. What you do by deconvolution is to reduce the
contribution by out-of-focus signals to the image of a certain focussing.
This is why a properly adjusted confocal microscope with sufficiently small
pinhole does have a better resolution than a widefield deconvolution
system. Deconvolution does not ADD information to the raw data images, it
substracts un-desired information.
>In order to do this effectively you
>are supposed to use 0.17 micron beads ( balls) or 0.5 micron beads( balls)
>with high power objectives to get a reference particle for the software.
The size of the beads should be sub-resolution.
>While doing this test to acquire the necessary reference images from the
>0.17 u beads, they appeared as doublets( two beads) and the 0.5 micron
>beads appeared egg shaped.
This has to do with the fact, that you had your Nomarski prism installed,
as you describe later on in your letter.
> What will this type of distortion observed on
>beads do to the fluorescent images of cells and particles?
It appears there just as it does for the spheres.
> On closer inspection we notice the Wallistron prism, which is necessary
No. It is not a "Wallistron" prism. The man who build this prism was Mr.
"Wollaston". However, the prism which you use in DIC- or Nomarski
microscopy is a modified Wollaston prism. Since the modification was first
suggested by Dr. Nomarski, this special kind of Wollaston prism sometimes
is also called "Nomarski prism".
>for interference contrast, was in the light path. I interpret these results
>to mean that while doing interference contrast simultaneously with
>fluorescence there will be a degradation of the fluorescence signal.
Well. The fact that you probably do have a standard laser as your light
source means that you have polarized light. If your system - I do not know
which one it actually is - is properly adjusted, the laser beam will arrive
more or less linearly polarized in the microscope. Now, the Nomarski prism,
uncorrectly but perhaps more easily understandable expressed, "splits the
beam up into two beams" in a kind of birefringence (do NOT interpret this
too literally). These beams are laterally displaced by a distance which in
normal widefield DIC microscopy generally would be smaller than your
resolution limit. This is why the Nomarski system also is called
"Differential" Interference Contrast, contrary to other interference
microscopic systems (Jamin-Lebedeff a.s.o.).
Now, if you really do widefield Nomarksi microscopy in order to visualize
phase-objects, the rays which are laterally displaced differentially by the
main DIC-prism, i.e. the one in the condensorrevolver, are re-combined and
brought to interference by the secondary DIC prism, also known as objective
DIC-prism. (I am not talking now about epi-illumination DIC systems which
geologists, mineralogosts, metallurgists a.s.o. sometimes use). So, in
other words, the two DIC prisms correctly working in series, will not cause
the problem you have. BUT since you are only illuminating with polarized
light - your fluorescence light will generally be un-polarized - the
differential displacement of the two laserbeams will cause two images to be
produced, which are slightly displaced and not re-combined. In case of high
NA objectives, this will then show up as a kind of "double image". Contrary
to "ghost images" caused by other problems, this two double images are
generally equal in strength., unless the polarization orientation of your
laser beam is parallel to one of the optical axes of the Nomarski prism -
what actually might be a solution to your problem, albeit demanding some
practical work to realize.
> Does
>any one know by how much? Should we not use interference contrast when
>acquiring high resolution fluorescent confocal images?
You can try for yourself. Take a standard fluorescent preparation with some
structure (sharp edges), e.g. some plant preparation. Then scan images at
increasing zoom-factors. There will a limit when the "double" image starts
to show up. According to my personal experience, this mainly is a problem
in case of strong objectives when approaching the Nyquist limit of pixeling.
There is another "egg" phenomenon which should not be mixed up with what
Dr. Zucker has described. The "eggs" Dr. Zucker describes, as far as I
understand, are lateral eggs.
Micro"spheres" will generally show up as axial eggs, even without the DIC
problem described above. This then is a consequence of the
Point-Spread-Function of the objective.
I hope this answer will help. I beg your pardon for the physics, I do not
know any simpler way to explain the observation. Perhaps somebody else
does. Would be great.
Best wishes
Johannes Helm
>Although interference contrast is a useful signal, it should not degrade
>the fluorescence signal if ones wants optimum resolution from a specimen. I
>would like to have some input on this observation and any suggestions or
>recomendations on the simultaneous use of both signals. Thanks
>Bob
--
********************************************************
Paul Johannes Helm, M.Sc., Ph.D.
Mail Address:
Institute of Basic Medical Sciences
Department of Anatomy
P.O. Box 1105 - Blindern
NO-0317 Oslo
Norway
Visiting Address:
Institute of Basic Medical Sciences
Department of Anatomy
Sognsvannsveien 9
Rom 0346/0347
NO-0372 Oslo
Norway
Voice: +47 22851159
Fax: +47 22851278
Email: [log in to unmask]
WWW: http://www.uio.no/~jhelm
********************************************************
|
