Too much already - I'll try to be brief.  References are
at the end.

        This problem is like the problem of focally photolyzing
caged compounds, which are UV-sensitive (300-400 nm).
My explanation is based on that, since it's what I do.

        UV fibers for focusing light onto a specimen certainly exist.
Two ways to aim the fiber are: (1) put it directly over the specimen
(you get a big spot, ~mm) or (2) put it behind the epifluorescence
port or another lens.  In epifluorescence the focused spot size
is (fiber optic diameter) divided by (objective power).
A typical value is 100 um / 40 = 2.5 um (Wang and Augustine).
If you put a dichroic in then you can merge with the confocal beam.

        If you don't mind illuminating a big field just use the lamp
directly, no fiber, through the epifluorescence port.  In this case you
could stop down the aperture, and make a spot tens of microns
wide (Khodakhah and Ogden).

        I am talking about multimode fiber optics (the common kind),
in which the light bounces around inside.  This means that even
if it's a laser coming in, the light comes out not collimated any more
and fills the 'acceptance cone' of the fiber (kind of like NA).  You lose
some light because some of it goes out in undesired directions.
To let UV through the fiber must be made of fused silica or quartz.
These are cheap and come from Ceram Optec, 3M, Polymicron,
and Spectrans.  Their Web sites have specs.

        Single-mode fibers retain collimation of a laser beam put in.
They can be bought but only last a few hundred hours.  If your
exposures are brief maybe that is okay.  They are called for only
if you have a laser, though in this case I suggest using mirrors and a
lens instead (Katz and Dalva). Personally, for my work (brain slices)
the tissue scatters light so such an approach is unwarranted.

        To summarize performance and costs:

Approach                    Spot size                   Cost

laser + mirrors          < 1 um (diffraction)   >=$10k (depends on laser)

laser + fiber               1-10 um                      similar but less
trouble

Xenon lamp + fiber  10 um - 1 mm            <$2000

        I hope this helps.  A lot of people are unsubscribing (and why
do we get those?) so sorry to go on.

       By the way, wouldn't big hair be *less* dense? ;-)

Sam Wang


Katz LC and Dalva MB (1994) J Neurosci Methods 54:205-218
Khodakhah K and Ogden D (1995) J Physiol 487.2:343-358
Wang SSH and Augustine GJ (1995) Neuron 15:755-760