Posted by
James Pawley on
URL: http://confocal-microscopy-list.275.s1.nabble.com/Pulse-compression-and-in-vivo-imaging-tp6557894p6576792.html
*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy*****
Following on from the messages Jim and I posted
on the list on 'chirping' we had a lengthy
off-list exchange, and we thought that it might
be of interest to others, so here is the
discussion. Jim is in italics, I'm in normal
type.
Guy
Firstly, shorter pulses will not make the psf
larger. The psf depends on the wavelength,
nothing else (and is, in principle, infinite ...)
The only thing that will make it appear to spread
is if you are saturating the fluorescence at the
centre of the psf, and this will apply
irrespective of whether you are doing 2P or
regular confocal. There are all sorts of reasons
why saturation is a bad thing which we need not
go into here.
I think you misunderstand me. I agree that the
"shape of the converging rays" does not depend on
the pulse length. However, these converging rays
represent a light intensity at each z plane.
Normally, we assume that this intensity only
reaches the instantaneous level needed to elicit
(much) fluorescence in the "blurry blob" near
crossover. Admittedly, this blob has fussy edges
but the does have a "z-resolution," in the form
of some Gaussian-ish response function.
This has nothing to do with
multiphoton or pulse length - there is no true
zero to the confocal psf in Z so very bright
objects will intrude into the planes where they
might not (naively) be expected. I've seen this
effect any number of times in ordinary confocal
imaging. You will find it mentioned in my book!
The only time the psf is actually affected,
though, is if we saturate in the centre - and
this applies whatever illumination mode we are
using. I think we really agree about this.
Yes. And saturation at the peak is mentioned in my book.
However, it seems obvious to me that as the peak
power goes up, it will become high enough to
produce significant signal at some (short)
distance above or below the focus plane. I agree
that the in-focus plane will still produce more
more signal but as you mention, saturation is
possible (see duty cycle below) and I think that
the flattening caused in this way, coupled with
the effect just mentioned above, could reduce
z-res (depending on the actual peak intensity
levels).
Please recall, that my comment was addressed as a
possibility and asked is anyone had noticed this.
If you shorten the pulse while keeping the power
the same you will increase all 2 and 3 photon
processes, both damaging and not. Since 2P
follows a square law and 3P a cube law you will
change the relative proportions, as Jim says. If
you shorten the pulses and keep the peak
intensity the same you will reduce overall power
which at least will reduce heating of the sample
so there should be some benefit, without
affecting multi-photon processes. I suspect that
what most people do is something in between.
Surely, if you shorten pulses without increasing
peak power, you will reduce average excitation
(time x intensity squared) and have less signal.
I would guess that people seldom do this (except
to the extent that, when making the adjustments
for shorter pulses, you may affect the average
power inadvertently). They try to shorten pulses
while keeping AVERAGE power the same.
OK, sloppy writing on my part. I
should have said keeping average EXCITATION the
same. But since excitation is intensity squared,
this is NOT keeping the average power the same.
I think we are in total agreement about what
happens if they keep the average power the same.
Why is 2P more damaging? Arguably it's not -
Vadim Dedov and I were able to measure
mitochondrial membrane potential in nerve cells
with JC1 using 2-photon excitation while
equivalent single-photon excitation killed the
cells and we couldn't measure anything.
But I didn't say that "2-p is more damaging". I
said very specifically that 2-photon was more
damaging per excitation. But as the excitations
are confined to near the focus plane, then, on
thickish specimens, there are far fewer total
excitations and therefore less damage overall.
And mitochondria may be an anomalous test object
as they are set up to withstand singlet O2.
There are plenty of
non-mitochondrial examples. Remember the famous
Cornell sea-urchin egg division series. José
Feijó has shown this on many different botanical
examples. How much damage is due to out of plane
excitation is hard to know, though I suppose one
could use a very thin sample (film of bacteria,
for example) to test this. But UV is so
obviously and immediately damaging to many living
cells that I find your thesis hard to sustain.
Once you get past 20µm in (assuming some stain
through out), I think that out of plane damage is
by far the predominant mechanism, especially on
the embryoes you mention. (Jayne Sparrow and the
mouse egg that hatched?)
There are lots of other examples in the
literature. What is true is that 2P can cause
different sorts of damage. The most extreme is
breakdown caused by the electric field, which
appears as bright flashes as you scan and
'craters' thereafter. If you increase the peak
electric field you will naturally increase this
damage.
Another point is more subtle. Chemical
selection rules state that in a symmetrical
molecule, 2P excitation must occur to a different
state than 1P. This means you will not excite
the S1 state, and hence you have an enhanced rate
of inter-system crossing into a triplet state.
This is a very noticeable with fluorescein, since
it is symmetrical. There are lots of published
spectra out there now - if a fluorochrome shows
very different 1P and 2P spectra you'd do best to
avoid it.
Finally, when we compress pulses we may not get
what we think we are getting. Chirping gives a
pulse a strange shape, which we hope will even
out to a normal pulse after passing through our
optics. If in fact we excite with a chirped
pulse then the peak intensity may be much high
higher than we'd calculate from the nominal pulse
length and average power.
One more factor. As 2p is pulsed, the duty cycle
is usually less than 10%. This means that people
often work nearer to singlet-state saturation
when using 2photon (to get an image in the same
scan time). This means that a lot more excited
molecules are present in the very high excitation
field near the centre of the focus, and increases
the likelihood of "one-plus-one" (or maybe 2 plus
one) overexcitation. Many smart, 2-photon folks
blame this for much of the increased
bleaching/excitation noted.
Well, they may be right but I'd
like to see evidence. In all the cases I've
looked at where people reported rapid bleaching
in 2P it's been the selection rule issue that was
obviously to blame. People think that because
FITC is such a super-bright and relatively stable
dye it must be the label of choice. Then they
excite at about 750nm (where it does indeed have
a peak). And it bleaches very fast indeed. I
pointed this out at an FOM meeting more than 10
years ago, in discussion on a paper which had
measured FITC bleach rates in 2P. Actually if
you excite FITC at 900nm it's much more stable.
I accept the point. And Piston did note extreme
differences in bleaching rates of different dyes.
The other curious feature I've
seen in postings on this issue is that most of
these people seem to be using 'experimental'
systems with nothing (other than adjusting the
laser) to control the power reaching the sample.
So their problems are hardly surprising! Clearly
you need to use an EOM or whatever to control the
power. You cannot adjust a Ti-S at all
satisfactorily by adjusting the input excitation
- I know, I've tried it.
Again, we agree.
--
Jim Pawley (Summer address) c/o Postmaster,
Egmont, BC, Canada, V0N-1N0 604-883-2095,
[hidden email]