Q on intensifying screens --correct-correction of fallacies!

[David F. Spencer]

In article <7opiGDf98QgB-pn2-
hwr2Wdn2gWtM at dawntreader.bio.ukans.edu>, 
PGegen at UKans.nolospamare.edu says...

> Some of this dialogue contains serious errors, which might dissuade people 
> from trying two screens even when it might help them, and will certain confuse
> anyone who is trying to separate fact from fantasy!

Well Peter, I'm afraid I took the novel approach of actually 
doing the bloody experiment before I posted anything. Admittedly 
unusual considering what sometimes passes for "facts" in this 
group. 

I took a Geiger counter, an old Hi-Plus intensifying screen and 
some 32P in an Eppendorf tube.  I also used a piece of plexiglass 
to attempt to account for Bremsstrahlung in some cases. Some of 
what I found I must admit surprised me but what I described in my 
post is what I observed plus obvious deductions of the 
implications of what I observed. The Hi-Plus screen does not have 
the overall efficiency of Lightning Plus for 32P but I didn't 
want to risk burning out a good screen (if beta can even do such 
a thing). 

> The original question was simple:  
> 
> Does one need one or two intensifying screens, and if two are used, how should
> they be placed? 
> 
> The answer is equally simple: one screen works fine for most purposes. The 
> placement is sample:film:screen (active side against film). 

That is indeed the logical way to set things up but you are being 
very presumptious to assume that most or all readers realize 
that. 
 
> Greater sensitivity can be obtained with two screens, placed on either side of
> the film with the active side of each screen in direct contact with the film. 
> This will work for film with emulsion of one or two sides, although 
> double-sided film will be somewhat more sensitive. This is the only placement 
> of two screens that has ever been suggested, and its benefits were clearly 
> documented in the original paper describing the use of intensifying screens 
> for 32P samples. The sensitivity isn't twice that of a single screen, for 
> obvious reasons, but it is clearly higher (at least as reported). 

The original paper on use of intensifying screens in molecular 
biology is the Anal. Biochem paper from the late 70's (Swanson 
and Schenk ?) which I don't happen to have here at home. As I 
recall, their recommendation for two screens was for use with one 
of the iodine isotopes.

Incidentally it's easy to think of another [sample:film:screen] 
arrangement, namely, blot/sample in the middle, one piece of film 
and one screen on each side (a super sandwich, a double-single-
double), and after the exposure and film processing, clamp the 
two pieces of film over one another in register. That has to give 
you about double the sensitivity of a single screen. Someone must 
have tried that, if not well you read it first here. (that's 
Spencer with a 'c'). Of course someone will probably want to do 
this with 4 screens and two pieces of film.

I have never used single-sided X-ray film but I know that double-
sided film (and normal black-and-white photographic film) has 2 
pigments in the emulsions (which dissolve in the developer and 
fix respectively) and so light produced on one side of film will 
produce at best a haze on the emulsion on the second side. If 
single-sided film has similar pigments then such film is 
inappropriate for the "sandwich" method.


> Two screens are expected to yield a less-sharp image, for the obvious reason 
> that the beta particles will spread out from their source as they travel 
> through the first screen to the film. Those which travel through the film will
> be spread more, and the light they produce from the second screen will spread 
> out, too. However, we are not talking Ansel Adams here! If you're looking for 
> a few bands or spots, and your choice is between exposures of 1 week vs 2 
> weeks, or 1 month vs 3-4 months (because of reciprocity failure), you aren't 
> in a position to complain about image quality. 

If you read what I wrote (and indeed even what you admit here), 
you can't hope to achieve a factor of close to two.
The highest improvement in sensitivity (even with the unavoidable 
image degradation) for one screen with the best screen-type film 
is about 5- to maybe 10-fold. I don't contest that. What I'm 
challenging is the notion that a second screen improves that 
sensitivity anywhere nearly as much as one might at first 
intuitively suspect. The image degradation is more than 
an aesthetic concern (although if you're doing direct RNA 
sequencing with small amounts of and/or poorly labelled RNA the 
image clarity is far from irrelevant).
Even assuming only a minimal 5-fold enhancement that a single 
screen setup should give, a 1 month autoradiography exposure 
with a screen would be the equivalent of more than 5 months 
without a screen. What on earth would you be trying to detect? 
Good lord.
 
> I am troubled because one poster disparaged the use of two screens, and in 
> doing so made several erroneous statements which were not based in physical 
> reality. Edited portions follow (... represent my deletions): 
>  
> > This surprisingly common notion of using two intensifying in routine
> > autoradiography baffles me. 
>  ... 
> > There are certainly a large number of other variations (some bizarre) on
> > how you could use two screens. One would be to place the film between two
> > screens with phosphors facing in, then set that onto the blot or gel. 
> 
>  -- This is the correct way to use two intensifying screens. There is nothing 
> bizarre or outre' about it at all.

You clearly misapplied my modifier "bizarre". And I have known 
people to use extremely "innovative" arrangements of screen and 
film. I do take issue with the term "correct way".
 
> 
> > That might seem to be a way of improving sensitivity, but first consider how
> > thick an intensifying screen is, 
>  ... 
> 
> >... there will be some small attenuation of the 32P passing through the 
> plastic.
>  -- True, the attenuation will be small. 
> 
> > The biggest problem, however, would be ... the size of the 32P image" by the time
> > it hit the nearest screen would obviously be degraded, and because it is
> > now spread over a larger area, the intended increase in sensitivity will be
> > less than predicted. 
> 
>  -- There will be a spread of the beta radiation away from the original 
> source, but not as much as the above statement suggests. 

In the standard (single screen) setup the active part of the 
screen phosphor is immediately adjacent to the film emulsion, and 
the beta only passes through the film to reach the phosphor, the 
thickness of film being less than half the thickness of the 
backing of the screen, which in the two screen sandwich is the 
shortest distance to the nearest phosphor. Inverse square turns 
that into more than 4, not too difficult to calculate. Spread an 
image over 4 times the surface area and clearly the final 
intensity of image has got to be less than in the "logical" one 
screen setup. We are talking about situations where a screen 
makes sense, for samples without enough radioactivity to produce 
a decent image in a week or two (or longer) directly, with 
reciprocity being one factor. Note that the beta intensities here 
would not be high enough to overload the screen in the single 
screen scenario. Therefore the area over which the light emitted 
from the screen is dispersed is relevant (there is also the 
consideration that even allowing for reciprocity it may take more 
than one photon to convert a silver halide crystal). I'm actually 
ignoring the fact that when the beta strikes the "back" of the 
first phosphor layer (which is non-zero in thickness), the light 
produced to some extent must pass out through, and be diffused 
by, the remaining thickness of phosphor to hit the film.

> > Another factor in this 2 screen "sandwich" is that the 
> > screen nearest the 32P actually blocks (either by absorption or reflection)
> > close to half of the 32P beta hitting it (from behind, technically). 
> 
>  -- This statement has no basis in fact. The screen does _NOT_ "block" beta 
> particles: they either interact with the screen to produce light emission, or 
> they pass through it! There are no other physical possibilities. "Converting 
> an electron to a photon" is not the same as "blocking an electron." 

Oh nonsense. When you put a Hi-Plus screen between 32P and a 
Geiger tube about half of the counts "disappear"; that's BLOCKING 
if you're just measuring the output. Even if every beta captured 
produces a photon (a remarkable efficiency if true) that says 
that Hi-Plus screens are about 50% efficient at best; Lightning 
Plus screens will be slightly better (although there are clearly 
two ways it can achieve this, either by more efficient capturing 
of beta, or by producing a "better" wavelength of light). It also 
means that in the sandwich setup only about half the initial beta 
can reach the second screen, and half of what hits that screen 
goes through it. The beta that hits the second phosphor has gone 
through the entire first screen (about 25 mils on my micrometer), 
and through the entire X-ray film (about 7-8 mils), with a small 
percentage of the beta being captured (or at least having 
interaction with halide) in the X-ray film emulsion on the way 
through, which is of course unrelated to the presence of a 
screen. You do the math for the spread of beta at the second 
screen's phosphor. 

 
>  -- The notion that electrons could be "reflected" is simply absurd. The 
> closest phenomenon is Compton scattering, which is not "reflection" -- it more
> resembles refraction. 
> 
> ...
> > The beta that does make it to the second screen can result in light emission, or

>  -- As before, it is physically meaningless to speak of electrons as "reflecting". 

I'm no physicist (I don't know about you) but you can test this 
yourself. Hold a microcentrifuge tube with a bit of 32P in the 
bottom above a standard old style, cylindrical Geiger tube, just 
at the end. If you position it properly you will get few if any 
counts. Bring an intensifying screen towards the end of the tube. 
You will find that as the screen gets nearer the end of the 
Geiger tube you get counts, quite a few. The same effect occurs 
even with metal or plexiglass but not as efficiently. Obviously 
the Geiger tube can't detect light so I first assumed that this 
was probably Bremsstrahlung, because intensifying screens rely on 
various heavy metal-containing phosphors, such as calcium 
tungstate or lanthanides, etc. which presumably can generate 
these high energy photons when hit with high energy electrons.  
However, plexiglass blocked these "reflected" counts so it must 
be electrons (beta) being detected by the counter (and besides 
Geiger tubes are very poor at detecting high energy photons). My 
best guess here is that beta particles are going through multiple 
"bendings" (which do generate "braking radiation" but the Geiger 
counter's detection efficiency is too poor to pick them up) and 
some beta eventually can do the equivalent of reflecting 90 (or 
more) degrees. The 32P beta packs a pretty good starting clout 
and can lose a moderate amount of energy and still be reasonably 
efficiently detected by standard lab counters.

I don't get the Compton scattering reference. In the Compton 
effect electrons are "ejected" when high energy gamma (or X) 
irradiates low atomic number atoms. In other cases, with larger 
atomic number atoms in the target, you get the photoelectric 
effect. Beta cannot exhibit Compton scattering.

Whatever the explanation for my above observation, this is 
experimentally verifiable and means that, whether you like it or 
not, intensifying screens do not only either a) capture a beta 
and release light or b) do nothing at all and let the beta pass 
through.  Some of this incident beta (from my crude experiment it 
could be 5-10%) goes in a direction other than straight through 
the screen which means that it is reflected (technically 
repeatedly refracted I suppose). The Bremmstrahlung produced may 
or may not result in light formation from the phosphors because 
even 50-100 kev X-rays don't stop easily. I have no idea how this 
further lowers the sensitivity of the "classic two screen 
sandwich" but if my hypothesis is correct it means that cassettes 
with 32P and intensifying screens are throwing out nasty X-ray 
class photons. Surely that can't be true because I've never heard 
any warnings?!

I repeat my original point, use of one screen makes sense when 
necessary, using two rarely if ever does (in the conventional 
sandwich method that is, not of course in the 'Spencer' 
technique, patent pending).

-DFS
 
-----------------------
David F. Spencer, PhD
Dept. of Biochemistry
Dalhousie University
Halifax, Nova Scotia
Canada

dspencer at is.dal.ca
-----------------------