🚀 go-pugleaf

Steve, I have read and reread the Czirr-Jensen paper and examined each
of the figures to which you refer.  Your comments, and those of your
collaborators, have clarified some points, but if you will bear with
me there are still some things I am not clear on.  First off it is the
2.9 MeV calibrations spectrum, rather than the 5.2 MeV one that would
seem to have the most bearing on the issue.  If, as you indicate, this
spectrum was produced by deuteron bombardment of a deuterated titanium
target, I have trouble accepting your explanation of the recorded spectrum
as consisting of a peak due to neutrons from the d + d reaction, and
a tail due to deuteron breakup on titanium, copper, or whatever.  The
reason for my skepticism is simple reaction kinematics.  If we note
that zero incident kinetic energy is to result in 2.45 MeV neutrons, 
what deuteron bombardind energy was used to produce the 2.9 MeV
neutrons?  It was a few 100 keV was it not?  What is the threshold
for deuteron break-up?  It is above 2 MeV is it not?  I think the only
way you can non-monoenergetic neutrons is from deuteron stripping
reactions, and here the coloumb barrier does play a significant
role.  In short the bombarding energy makes a big difference so you
have to put together a picture which includes a background run at
the same bombarding energy before you can explain the lower portion
of the 2.9 MeV calibration run as being due to something else.

It would seem that you have put alot of faith in the calculated
response of the detector.  We all know that without the coincidence
requirement for the neutron capture in the glass, the pulseheight
response for the liquid alone would be a continuum extending down
to zero pulse height.  In fact that spectrum would rise a low
energies.  I can see that the coincidence requirement would reduce
that low energy side of the spectrum, but the Monte Carlo calculation
looks a bit wierd.  Note, the two lowest calculated points indicate
a very steep rise near zero pulse height.  I am tempted to
speculate that the 450 counts in those two lowest bins should
really have been distributed over channels 0-30.

Your assertion that the gamma efficiency for the glass is low does
not square totally with the literature on the subject.  It depends
on the thickness of the glass plates, a bit of data not contained
in the paper so I have to guess that it is 2 or 3 mm.  We then
need to recall that although the pulse heights for neutron capture
in the glass is independent of energy, as a gamma detector the glass
acts more like a normal scintillation detector showing a pulse height
roughly linear as a function of electron energy loss.  Hence an
evaluation of how much "leakage" there is of the gamma response
into the neutron peak is dependent on the gamma energy spectrum.
The reponse deduced for a radium source is not universal.  While
it is true that the cosmic ray spectrum seems to show a gamma
peak and a neutron peak, determining how those two peaks merge
and what events belong to which peak is a bit of a guess isn't it?
I guess I would like some way of knowing, independent of measurements
made with the Czirr-Jensen detector itself, what the background
neutron rate really is relative to the gammas.

I had hoped that my question concerning the normalization of the
subtracted background could have been resolved by now, but you haven't
made a completely clear statement as I read your latest message.
Was the normalization of the background subtracted based solely on
the ratio of times spent recording the two spectra?

Dick Blue
NSCL @ MSU