🚀 go-pugleaf

In article <00966113.17594E40.9677@dancer.nscl.msu.edu>, 
blue@nscl01.nscl.msu.edu writes:
> Recent post by S. Jones relays information provided by Profs. Jensen and
> Palmer concerning the detector used in the first Jones experiment reporting
> low yield neutron production from cold fusion.
>  .....
> Clearly I was mistaken as to which mode of operation had been employed
> to obtain the spectrum reported in the Nature paper. 

I'm glad to see we have settled that point.

> There do, however,
> remain some questions as to what the reponse function of the detector
> is for neutrons and for gammas.  Both Jensen and Palmer in their replies
> refer to spectra obtained at 2.9 MeV and 5.2 MeV as reported in the
> Czir-Jensen paper, but Jones discounts those as calibration spectra in
> favor of some more recent calibrations with monoenergetic neutrons.
> Not only do we seem to be getting differing answers from different
> members of the team, we seem to have a situation in which information
> crucial to the intepretation of the spectra was subject to revision
> long after the experiment was completed.  If we are to be convinced
> that the spectrum reported in Nature shows a "peak", it would be nice
> to know whether the detector as operating at the time was in fact
> capable of producing such a peak in response to fusion neutrons.  One
> question that remains in my mind is what mechanism accounts for
> the roll-off of the proton recoil response on the low side of the
> peak.  Off hand I would expect the response to be close to what is
> shown in the Czirr-Jensen paper.

I will try to show figures from the Czirr-Jensen papers so readers can
understand Dr. Blue's questions, and to provide the answers. 
       Figure 1:  Neutron spectrometer response to 5.2 MeV neutrons
              (Fig. 5 of NIM A284 (1989):365, Czirr and Jensen, BYU)
|          -
|          -
|           \                  --
|            -               -   - -
|            -             -        -  -
|            -            -          -
|  COUNTS     \          -              -
|              -        -                -
|              -       -                 -
|               - - - -                   -
                                           -
                                            - -----------
 --------------------------------------------------------
                 PULSE-HEIGHT CHANNEL (neutron energy information)

                               /|\  Clearly the detector shows a peak here.
         /|\  But what is this low-energy tail?  This is a principal question
of Dr. Blue's; as he says in reply to BartB:
"one question I am raising about this experiment has to do with the nature
of the signal the detector would record even in the absence of background.
In the Czirr-jensen paper that describes the detector, the calibration spectrum
[given in Fig. 1 above]
shown is not in accord with the signal claimed in Jones paper [Nature 1989]"
which looks roughly like this:

    Fig. 2  Neutron spectrum from original Nature paper, background subtracted
|               (agrees with monte carlo prediction for monoenergetic neutrons)
|
|
|           - -
|          -   -
|        -      -
| COUNTS         -
|
|       -         -
|
|      -           -          -
|     -             - - -    - - - -       -  -
|   -                 - -    -   - - ---- -  -   - etc.
|-----------------------------------------------------------------
     -                    -
                   Pulse Height Channel
The peak is lower than for the 5.2 MeV neutrons used in the calibration curve
above.  (We also ran a calibration for 2.9 MeV neutrons; the above peak appears
just below this 2.9MeV peak, and an short extrapolation based on both calibra-
tions gave the energy of the above peak as approx. 2.5 MeV, consistent with
neutrons from deuteron-d fusion.  See Nature
papers, 27 April 1989 and 22 Feb 1990 - correspondance from us.  This responds
to BartB's questions in a recent post.)

But, Dr. Blue questions, where is the low-energy tail?
First, as explained in the Czirr-Jensen paper, the calculated response of the
spectrometer based on the Monte Carlo program MCBD provides a bump as given
in Fig. 2 -- WITHOUT A LOW ENERGY TAIL.  (Didn't you notice this in the paper
Dick?)  This was the subject of a BYU Master's Thesis, in part.  In fact, the
observed bump shape/width fit pretty neatly with the monte carlo prediction.

Second, in work since the Czirr-Jensen paper (rec'd 7 June 1989), the two
demonstrated that the low-energy tail arises from high-energy deuteron beam
interactions in the Ti-d target, due to deuteron break-up.  When the d beam
from the BYU Van de Graaf impinges on metal foils (not containing deuterons
for d-d fusions), they found the following response in the spectrometer:

     Fig. 3  Spectrometer response for 1.9 MeV deuteron beam on blank
       copper disk (instead of titanium deuteride target with Cu backing)
          (From Prof. Jensen; included in BYU pre-print 1992,"Single-Tube
                                 Neutron Spectrometer")
       -
|       -
|       -
|       -
|        -
|         -
|         -
| COUNTS   -
|           -
|           -
|           -
|            -    -   -          -
|             --   ---  - --- ----- ------ etc.
|----------------------------------------------------------------
                PULSE-HEIGHT CHANNEL
Aha!  So that's what the low-energy tail is -- nothing but neutrons from
deuteron break-up.  NOT the response of the detector per se.

So in the latest calibrations, Jensen et al. measure the background as in
Fig. 3 for d break-up, and subtract this from Fig.1 (since there both d-d
fusion and d break-up occur) to yield a peak as in Fig. 2 and 3, that is,
the corrected calibration spectrum agrees both with the monte carlo calcula-
tion and the original data spectrum!
Send me your address, Dick, and I'll send the write-up on this.
Please understand, there is no "discounting" of earlier calibrations, nor
differing answers from different team members.  I hope we're clear on these
points now and can move on.

> The next question that can perhaps be laid -to rest if I understand
> Jensens reply has to do with the way in which the background subtraction
> was made.  Jensen states: "The background HAD to be normalized [Approx
> 4 times more background than foreground hours], but the background
> was featureless and could not generate a peak."  I take that to
> mean that the background was scaled in strict ratio in accord with
> the different recording times before subtraction, and no other
> adjustments were made such as matching forground and background in
> a region of the spectrum away from the "peak".

There are different, legitimate ways of scaling background discussed in our
papers.  I refer you particularly to our response in Nature 22 Feb. 1990.
There we describe an analysis of runs 1 to 7; the scaling was done as you
state above "in strict ratio in accord with the different recording times
before subtraction."  Note that background runs were run more or less
alternately with foreground runs, and that background runs included metals
involved in the foreground runs.   Please read for more info. another BYU
paper in J. Fusion E. 9:199-208, Dec. 1990, where we state:
 "Background runs were made using operating cells containing standard
electrodes and electrolytes, except that H2O replaced the D2O:  numerous
light-water control runs were performed prior to submission of [Nature 27
April 1989 paper].  Other background runs were made using both new and 
previously used standard cells containing D2O plus the usual electrolyte
but with no electrical current.  In the case of used D2O cells, the
current had been off for many hours.  The individual background runs
followed the featureless pattern of the integrated background illustrated 
in Fig. 4."  (J. Fusion Energy 9:199, Dec 1990,"Anomalous Nuclear Reactions
in Condensed Matter:  Recent Results and Open Questions,"  S.E. Jones et al.)

> 
> We are still left with the issue of the gamma response of the
> detector and what fraction of the response was in fact due to
> neutrons.  From other experiments which employed liquid
> scintillation counters with pulse-shape discrimination to 
> separate neutron and gamma response on sees cosmic-ray-induced
> backgrounds showing gamma-to-neutron ratios of something like
> 10E3 or 10E4.  Clearly this ratio can be altered by effects
> specific to the surroundings of a given experiment, but I
> see a potential problem in making a determination of this
> ratio with a detector that may well respond with no better
> than a 100 to 1 rejection ratio for gammas.  To make that
> explicite let us assume that the true ratio of gammas to
> neutrons is 10E4.  Then the detector will respond to gammas
> at a rate 10E2 times the neutron rate.  Under those circumstances,
> or something approaching them, how do you tell what the ratio
> of neutrons to gammas really is?  Jensen asserts that about 1/4
> of the background is due to gammas, but how does one go about
> making a determination of that number?
> 
> Dick Blue
> NSCL @ MSU

Our COINCIDENCE spectrometer differs significantly from other detectors
which use liquid organic scintillator (which do indeed show a low-energy
tail as part of the response).  The key is to require a pulse from
neutron capture in 6Li-doped glass in delayed COINCIDENCE with the
pulse from the liquid scintillator.  This coincidence is a powerful
way to reject gammas since the glass response to gammas is of very low
efficiency and differs from the t+alpha pulse arising from n capture
on 6Li.  We do NOT rely only on pulse-shape discrimination.  

(This is important:  understand the nature of our coincidence
spectrometer.  We DO get a bump from mono-energetic neutrons, without
a low-energy tail.)
Perhaps quoting from the most recent Jensen-Czirr write-up will further
clarify this:  "A series of low-resolution neutron spectrometers has been
developed in our laboratory for spectral measurements in the MeV energy
range.  These detectors operate on the "coincidence calorimeter" principle
in which two signals from a single neutron are required to assure total
kinetic energy deposition within the detector.  The first signal is produced
by the multiple elastic collisions of an incident neutron with protons in
the organic scintillator comprising the bulk of the dector body.  The second
signal arises if the neutron remains within the detector and is captured by
a ^Li nucleus incorporated in thin lithium glass scintillators that are
dispersed among the several organic plastic scintillators.  The slowing
down of the incident neutron occupies approximately 50 ns and the mean capture
delay is 11 microseconds."

This standard mode of operation, used also in original BYU experiments, is to
be distinguished from the "capture-spectrum mode" which is described in detail
in the Czirr-Jensen 1989 paper (NIM A284:365-369).  Basically, the capture-
spectrum mode provides pulse-height spectra from the glass alone, rather than
from the plastic scintillator.  We cannot determine the neutron energy from
the 6Li-doped-glass scintillator, but we can see neutrons clearly, and the
distinct signal from gammas.  Background at BYU looks like this:
  
   Fig. 4:  Capture-spectrum mode distribution for background events.
            (Differs from standard mode in Figs 1-3 above!)
|        -         (Figure 8 in Czirr-Jensen paper in NIM A284(1989):365)
|       - -
|
|
|      -   -
| COUNTS
|
|     -     -
|                         --
|    -       -         /- -  - -
| ---         --   -----        - ----\
|---------------------------------------------------------------
                    Pulse-height channel

The low-energy peak is caused by gamma interactions in the glass, while
the higher-energy bump (smaller) is caused by neutrons, presumably originating
from cosmic-ray interactions.  In the standard mode, we select only light-
pulses from the glass scintillator which occur in the higher-energy bump,
thus discriminating against gammas.  The cross-section for gamma interaction
in the thin glass scintillator plates is very low compared to the cross-section
for thermal neutron capture in the plates.

Using the same mode with a radium source shows the gamma peak only, with a tail
that extends into the neutron-bump region.  By scaling this gamma peak to match
the gamma peak in the background, then counting events in the gamma tail which
extend into the neutron bump region, and comparing to background- 
neutron-generated events in the neutron-bump region leads to an estimate of 
the events in the neutron region arising from gamma-ray leakage. 

This is explained in the NIM paper by Czirr and
Jensen, but I can understand some confusion due to their term "capture-spectrum
mode"  which refers to the spectrum from the GLASS scintillator only, whereas
the standard mode involves capture also, but the spectrum is that yielded by
neutrons slowing in the PLASTIC scintillator, with subsequent neutron capture
required in the glass.
[Note:  I discussed these points with Prof. Jensen, co-inventor of the 
coincidence spectrometer, but neither he nor Prof. Czirr are now available to
proofread what I have written.  Copies of their papers are available on
request.]

In short, the BYU neutron spectrometer discriminates well against gammas
and produces a peak in the pulse-height spectrum given a monoenergetic neutron
source.  It is thus much better than using organic scintillator with pulse-
shape discrimination alone for low-level neutron studies.


Respectfully,
Steven E. Jones