.
IT !!
WITH..
.
P
.?
L
A
A 11
AWARD
www
A
W
UNCLASSIFIED
ORNL
2
27
.
FIY
548
17.
ORrup.d-48.
) - VAزہی
ا
x11/
BUITIODS OF FAST-NEITRON SPECTROSCOPY POR REACTOR SHIEI-DING *
F. C. Maienschein
Oak Ridge National Laboratory
OCT 30 1909
INTRODUCTION
At the session on Neutron Attenuation in Optically Thick Shieldo,
discrepancies were notod even between results from calculational methods
which were supposed to be "rigorous." The obvious suggestion was made
that experiments should be performed to determine the "correct answer."
However, experience has shown that critical examinations of the capabil.
ities of calculations on the basis of experiment have been few. "Real
experiments" usually involve geometries which are complex with respect to
those that the calculational methods can bandle as well as severe prob.:cm::
of determining source strengtho. In addition, the capabilities of the
detection Instruments are limited. When careful measurements are pos-
sible, the most meaningful tests of calculations and cross-section inputs
result from study of fast-neutron spectra.
Recent developments of spectrometers offer promise for future measure-
ments. However, further work is needed, and it should not be anticipated
that a simple, foolproof fast-neutron spectrometer requiring little effort
will become available. Any of the plausible methods currently requires
several man years to put into operation. Hopefully, calibrations for cor-
tain spectrometers can be provided which will reduce this effort. These
calibrations might take a form similar to that of the Heath catalog of
gamma-ray scintillation spectrometer functions plus a computer-code package
for treating experimental data which are produced. Comments on the use-
11'11
fulness of this approach would be appreciated.
*Research sponsored by the U. S. Atomic Energy Commission under contract
with the Union Carbide Corporation.
The most desirable features of fast-neutron spectrometers for the
measurement of continuous spectra in the fission energy recion include:
a kaim absolute efficiency of suficient magnitude that moasurements can
be completed in a few hours, gond gamma-ray discrimination, at least
moderate energy resolution (~10-20%), nondirectional response, and moderate
size, cost, and complexity.
Spectrometers with similar characteristics are required in related
fields, including studies of radiation damage and spectra in fast reactors.
However, for these applications, the energy region of most interest falls
below 1 MeV. For radiation-damage studies of most materials, the integrated
flux levels permit, or demand, the use of spectrometers of low sensitivity,
and high-radiation-damage resistance. For the measurement of fast-reactor
spectra, extraction of a representative neutron beam is required since the
spectrometer cannot be placed in the reactor core without introducing an un-
desirable perturbation. For shielding measurements, on the other hand, the
beam extraction problem does not arise and comparisons of the fast-neutron
leakage can be made with the calculational methods.
In the past, most of the possible concepts for fast-neutron spec-
troscopy have been tried, although new approaches, such as magnetic deflec-
tion of protons from neutron decay, are still being examined. The in- :
struments to be discussed in this talk are selected for comment because of
their demonstrated or anticipated use for reactor-shielding measurements.
The coverage is parochial not by choice but due to ignorance of other de-
velopments. In general, the useful neutron spectrometers depend upon
production of recoil prot ans, nuclear reactions in counters, activations
with cross sections of lowo enerwy dependence, or on time of flight.
PROTON-RECOIL METHODS
Useful shielding data have been produced using proton recoilo in
nuclear plates.,894 The rosults come blowly although improvement has
been achieved in this respect with uomi-automatic coordinate recording.
Gemma-ray backgrounds may be limiting and sensitivity 18 inadequate for
very large attenuations. So far as I loow, no development work is cur.
rently under way with nuclear plates.
Proton-recoil-proportional-counter telescopes exhibit good carama-ray
rejection, and, for moderate energy resolution, very low efficiency. Use of
a scintillator as the radiator and/or the proton-energy detector in the
telescope increases the efficiency but also the gamma-ray response, at-
though to my knowledge pulse-shape discrimination has not been tried. It
is also possible that use of a solid-state proton spectrometer ag an element
of a telescope should be investigated for shielding measurements."
Ilydrogenous scintillators give an integral pulse-height spectrum which
must be differentiated to produce the desired neutron spectrum. These de-
vices have efficiencies large with respect to those for the above instru-
ments and they are relatively nondirectional. However, the scintillator
suffers from a very large gamma-ray background and the differentiation re-
quired to obtain a spectrum is complicated by the finite pulse-height reso-
lution of the scintillator, nonlinearities in light output, and multiple
interactions. In spite of these problems, recent developments offer promise
of a useful shielding spectrometer of this type. The developments include
the use of pulse-shape discrimination against gamma rays and careful cali-
bration of the scintillator response with monoenergetic neutrons, together
with analysis of the pulse-height spectra obtained by a competent method.
Such a method will be discussed today.
For low-neutron energies (< 1 Mov) a hydrogenous proportional
counter can be used with pulse-shape gamma-ray discrimination in a manner
similar to that of the scintillator. This technique is of particular
interest for the determination of fast-reactor spectra.
NUCLEAR REACTION METHODS
A second major class of neutron spectrometers includes those which
depend upon the determination of the energy of nuclear reaction products
produced by neutrons. The only useful reactions are Sheln,p) and 21(n,x).
Other reactions lead to oxcited states and thus lose the one-to-one cor-
respondence between measured charged-particle energy minus the Q value
and the neutron energy. Even for these reactions complications arise.
especially in the form of elastic She recoils and competing reactions in
lithium such as Li(n,a) for high-neutron energies.
SHe can be used in a proportional counter to provide a spectr meter
of reasonable efficiency. The Columbia group has developed a method
based on pulse rise time which can be used to discriminate against the
She recoils.' This method will be discussed today. The value for the
Beln,p)r reaction is positive but small (+ 0.75 MeV) so that gamma-ray
background from the reactor may be troublesome. However, the pulse-shape
discrimination used against the recoils will also eliminate some of the
gamma-ray response and tests in a reactor spectrum are needed.
Since Lithium forms no gaseous compounds at or near room temperature,
considerable attention was directed toward incorporating lithium in a
crystalline scintillator. The resulting Z11(Eu) scintillator® gives
usefu efficiency and energy resolution when operated at low temperature.
However, the gamma-ray response, in spite of the + 4.785-MeV Q value, has
prevented its use for shielding measurements.
oso
A lator application of the 'Li(n,ayr reaction de in the form of a
sandırich of two solid-state spectrometers surrounding a thin layer of
ht. Coincident detection and energy determination of the triton and
alpha are required in the two diodes, thus reducing the gamma-ray response
and other backgrounds. The mi layer must be thin to permit oscape of
the charcod particles and the efficiency to low. Further, use of the
solid-state detectors (diodes), which are largely silicon, gives rise to
neutron-induced charcod-particle reactions in the silicon which may satisfy
the coincidence requirement by the charged particle emerging from the surface
of one diode and entering the other. This effect has been determined by sub-
stituting a layer of 'h in place of the Owl and subtracting the background
thus determined. A future approach will use the 18 otope 30si, to construct the
solid-state spectrometers, since the Q values for (n, charged-particle) reactions
In sºs1 are larger than for the other isotopes.
In spite of the coincidence requirement, the gamma-ray backgrourd is
probably the major limitation on the use of the mi dicde sandwich as a non-
directional neutron spectrometer in reactor shields. Shielding of the diodes
can be done by moving them away from the 'll layer and using a collimated
beam of neutrons (and gamma-ray background) incident only on the ºut. The
efficiency is thereby reduced further but data will be presented today for
water-shield thicknesses up to 50 cm using this technique.
The Sheln,p) reaction can also be used with a pair of diodes but the
lower Q value increases the problems of overcoming the gamma-ray and silicon
hackgrounds. Further, the marked anisotropy of the reaction products produces
problems of efficiency determination at high-neutron energies.
.
..-
.
.
.
MIRESHOLD-DETECTOR METHODS
A variety of nuclear reactions lead to radioactive nuclei which may
be detected by counting. Analysis of the results, taking into account the
dependence upon neutron energy of the reactions used, can give, in principle,
the incident differential neutron spectrum with very limited enorby reso-
lution. Much work of this type is available, and Vol. I of Neutron Dosimetry
is an especially rich source. This "threshold-detector" type of spec-
trometer 18 most applicable in studies of radiation damage a.lthough it has
been widely used, because of the lack of methods giving better energy reso-
lution, in shielding research. Results of this type will be discussed today
in two papers.
Variants of the above method which deserve mention are the use of
counting of the reaction products as in the case of fission chambers and the
use of absorbers to modify the response of a single detector. In the use of
fission-threshold reactions, the gamma-ray-induced-fission background must be
carefully considered, especially for measurements behind good neutron shields.
For the second approach, hydrogenous layers have been added to thermal-neutrori
detectors, 'LLI(Eu) scintillators and BF, counters, to extend their useful
response to beyond 14 MeV. The response must be calibrated and a powerfu
"unfolding technique" employed to product useful spectra with limited enervy
is large, and the energy range covered 16. very large. Boron absorbers can
be employed in a similar manner for lover energy neutrons.
-7-
TIME OF FLIGHT
It is true that the neutron time of flight cannot be usefully measured
for a reactor source (considering the MeV enorky region). However, with a
puised accelerator such as an electron linear accolerator, the cood enormy
ann spatial resolution and the large energy range make this a promising method
of neutron spectroscopy for basic shielding studies of moderate attenuations.
The photoneutron spectrum from the accelerator target 18 similar to a reactor
spectrum in enere' but not in angular distribution. The source spectrum can
be measured and used as the input for attenuation calculations which are to
be checked by experiment. Measurements of this type will be described today.
SUMMARY
What conclusions can be drawn from these comments? If I were asked to
recommend spectrometer types for shielding studies, I would give the following
answers. For the fortunate organization with an already established larce
group of nuclear-plate readers, the best approach is to simply use the plates,
with their concomitant disadvantages, where intensity and gamma background
permit. For other groups wanting to make the most detailed possible checks
of attenuation calculations, the time-of-flight technique should be considered
first, if an electron linear accelerator can be made available.
For shielding measurements at the reactor, the shielded 'Li diode spec-
trometer has demonstrated its userulness. If neutron intensities do not
permit use of this instrument, the proton-recoil scintillator offers con-
siderable promise at the present time. For radiation-damage experiments, the
potentialities of the threshold-foil technique have been demonstrated but the
possibilities for shield spectral determinations of using powerful unscrambling
techniques have not been exhausted. All who disagree with these recommenda-
tions should say so.
-8.
If there exists even partial agreement on spectrometer methods or
choice and if there is sufficient interest, it might be possible to share
the considerable burder. of calibration including efficiency and resolution
determinations for selected spectrometers. This cooperative approach would
require agreement on detector materials, sizes, and geometries to be used,
and the details would depend on the method under consideration.
If you have any comments on these possibilities, please see me at
this meeting or send me a lotter.
September 16, 1964
R. A. Karam and T. F. Parldnson, Neutron Decay Spectrometer,
Rev. Sci. Instr. 35, 600 (1964).
G. T. Western, Energy and Ancular Distribution Experiment,
MARP-62-16T, FZK-9-183-2, Sept. 30, 1963.
S. Blaize et al., Fast Neutron Spectrometry and Dosimetry, Proc,
Intern. Cont. Peaceful Uses Atomic Energy, 2nd, Geneva, 1958
Vol. 21, 230 (1959).
S. Passe, Mesures de Spectres de Neutrons Rapide 3 a l'Aide
d'Emulsions Nucléaires, Neutron Dosimetry, vol. I, p. 481,
IAEA, Vienna (1963).
A. Keith Furr and R. S. Runyon, A Fast Neutron Spectrometer for
Reactor Flux Measurements, Nuci. Instr. Methods 27, 292 (1964).
E. F. Bennett, Proportional Counter Proton-Pecoil Spectraveter
with Gamma Discrimination, Rev. Sci. Instr. 33, 1153 (1962).
A. Sayres and M. Coppola, He Neutron Spectrometer Using Pulse
Risetime Discrimination, Rev. Sci. Instr. 35, 431 (1964).
R. B. Murray, Use of °LII(EU) as a Scintillation Detector and
Spectrometer for Fast Neutrons, Nucl. Instr. Methods 2, 237 (1958).
T. A. Love and R. B. Murray, The Use of Surface-Barrier Diodes for
Tast-Neutron Spectroscopy, IRE Trans. Nucl. Sci. 8(1), 91 (1961).
Neutron Dosimetry, vol. I, IAEA, Vienna (1963).
R. L. Bramblett, R. I, Eiring, and T. W. Bonner, A New Type of
Neutron Spectrometer, Nucl. Instr. Methods 2, 1 (1960).
1. R. Burrus, Neutron Phys. Div. Ann. Progr. Rept. Sept. 1, 1962,
ORNL-3360, p. 19.
.
.
...
DATE FILMED
12/10/64







LEGAL NOTICE

The report we prepared as an account of Government sponsored work. Neither the United
state, nor the Commission, mor my pornos acting on behalf of the Commission:
A. Makes may warranty or representation, expressed or implied, with respect to the accu-
racy, completeness, or wefulness of the information contained in this report, or that the we
of any tifortation, apparatus, method, or procesu diecloud to the report may not Infringe
privately owned rights; or
8, Asom my Habilities will respect to the one of, or for damages rosulting from the
w of any information, apparatus, method, or procou diancloned in the repriko
Ao wond to the abova, "pornon acting on behalf of the Commission” includes may be
ployun or contractor of the Commission, or employee of much contractor, to the extent that
nich omployee or contructor of the Commission, or or.ployw of much contractor properes,
denominatoa, or provides mcon to, any information pur munat to do employmeat or contract
with the Communion, or Me employment with such contractor.
END



-