P12
?
TI
12
;
EXISTEIXE
AYTINA
IN
17 E
.
T AS
L"
PRET
WEBESSTILSTAND
TV
WALL
3.
UNCLASSIFIED
ORNL
4
.
Mai
T IWA 2.
19A
, 1www
.7
.
PA
2
2
.
PERS
*
*
_*
1092
2
.
-
.--:-
:
N
Orph-P-1092
Griet
Cong -654-54
MAR 23 1065
MEASUREMENTS AND CALCULATIONS OF ENERGY AND ANGULAR DETAILS
OF FAST NEUTRON FLUX IN WATER FROM A POOL-TYPE REACTOR*
V. V. Verbinski and M. S. Bokharit
Oak Ridge National Laboratory
Oak Ridge, Tennessee
out responsaboratory
· MASTER
ABSTRACT
printly owned rhota; or
racy, cotapletosout, or w
with the couraglan, or de oraploymutat vith such contractor.
dienalmate, or pronoun now to, any taforation pursued to be explotant or contract
mech amployee or contractor at the Commisslon, or employee al such coatractor preparme,
ployu or contractor of the Countedion, or employee of wch contractor, to the aleat that
As word la the abo", "paruou acting a bowall of the Cowaluobo" includes may
wo of way lafortation, appsrat, wethod, or proceu discloud
3. Acousas day los futhat with repect to the wes of, or for damascus rotters from the
of may taformation, appunto, method, or procura dincloned to this report may sot latring
tulongs of the fasortation a tined to als reports or that the one
A. Makes any warranty or represetation, pred or Itapled, with respect to the accu-
Site, vor the Comuulon, nor way peru acting on bold the Coualaston:
Thla report wo progured on account of Government ponsored work. Neither the Vallad
LEGAL NOTICE --
tulu port,
Measurements have been made of the spectral, spatial,
and angular details of the fast-neutron flux at one position
within the Bulk Shielding Reactor I (BSR-I) and at several
positions in the water medium surrounding the reactor. A
shielded-diode spectrometer, developed especially for these
measurements, was used to obtain the angular flux. The
measurements were made in a gamma-ray flux that was as high
as 10° r/hr. The spectra for the oº (forward) angular flux
initially harden with increasing neutron-penetration depth in
water up to about 30 cm, after which the spectral shape re-
mains nearly constant. The spectra at 40.5° and 51° are
much softer and are nearly constant at depths in water where
the spectrometer does not view. the edge of the reactor. The
reactor power was calibrated against thin uranium foils, and
the spectral data were compared with predictions of two
neutron transport codes, NIOBE and DTK, all normalized to a
reactor power of 2.4 x 10-11 watts. The experimental and
calculational results were everywhere within about a factor-
of-two agreement with the measurements over a range of

PATENT CLEARANCE OBTAINED. RELEASE TO
THE PUBLIC IS APPROVED. PROFEDURES
ARE ON EILE IN THE RECEIVING SECOM.
attenuation of about 104 for angles of 0° - 52° and energies of
1.5 - 12 Mev.
*Research sponsored by the U. S. Atomic Energy Commission
under contract with the Union Carbide Corporation.
Pakistan Atomic Energy Commission, Lahore, Pakistan.
-3-
Introduction
A knowledge of the spatial, spectral, and angular distribution of
neutrons in the water shield of a standard pool-type reactor is partic-
ularly useful for testing neutron-transport calculations, and is generally
of interest for a wide variety of experiments and instrument calibrations
that can best be performed in a neutron flux of known intensity and
spectral shape. This paper reports spectral measurements of the fast-
neutron angular flux at various positions and angles within the water
shield of the Bulk Shielding Reactor I (BSR-I).?
The geometry of the reactor core and its water shield is simple in
that water acts as both reflector and shield. It should be relatively
easy to reproduce this geometry for a variety of shielding calculations
which one night wish to compare with these measurements.
The experiments were carried out with a shielded-diode spectrometer
developed especially for this work. Although the spectrometer efficiency
is low, it is particularly suitable for spectral measurements of reactor
neutrons, where both the neutron and garma-ray intensities are high,
because the shielding arrangement makes the spectrometer quite in-
sensitive to the high gamma-ray flux from the reactor. The response of
the spectrometer to monoenergetic neutrons is a simple Gaussian, and the
average pulse height varies linearly with energy. With this response,
the energy spectrum is obtained directly from the pulse-height spectrum
by a simple correction for the spectrometer efficiency whose energy by
dependence has been calculated. The efficiency, resolution, and garma-
ray tolerance are sensitive functions of instrument design and cost,
and all three could conceivably be improved by a large factor over the
present design, which should be considered as a pilot model. These
factors will be discussed below in appropriate parts of this communication.
Experimental Arrangement
The general arrangement of the reactor and the spectrometer in the
Bulk Shielding Facility pool is shown in Fig. 1. In the reactor loading
used for these measurements (Fig. 1), the control rods and regulating
rods have been moved back from their normal, more central location to
leave the area near the front face of the reactor homogeneous and un-
cluttered. The reactor is a rectangular 38 x 38 x 60 cm configuration,
except for the two additional elements in the rear that were required to
attain criticality and to restore some of the symmetry lost by moving
back the control and regulating rods. The shielded-diode spectrometer
was positioned in the horizontal midplane of the reactor and the data
were obtained as a function of e, the horizontal angle between the pro-
longed center line of the reactor and the center line of the spectrometer,
and of r, the distance from the reactor face to the end of the air-filled
spectrometer collimator. Sulphur pellets positioned 20- and 40-cm from
the plane of the reactor's face, 38-cm left of the reactor's vertical
midplane, and 35-cm above the horizontal midplane, were used to monitor
the relative power of the reactor for various runs. The relative power
level from one run to another is known to an accuracy of 2%.
This rela-
tive power was converted to absolute power with a reactor power cali-
bration which was performed for the core configuration shown in Fig. 1.
This calibration is discussed below, near the end of this paper.
Shielded-Diode Spectrometer
As shown in Fig. I, the shielded-diode spectrometer essentially
consists of a 580-ug/cm2-thick 'LiF layer on 70-ug/cma-thick formvar,

UNCLASSIFIED
2-01-058-844R4
BSR LOADING 96A
:
-
.
:-
.
ALUMINUM
AIR
-20 cm -
20 cm -
-50
SULFUR FOILSL
LEAD
1111111111
!
TUNGSTEN LINING
CONTROL RODS EQUAL TO
12 FUEL ELEMENT
A
SILICON DIODES
SILICON DIODES
F LAYER
'liF LAYER
A
> VACUUM
INCHES
LEAD -
Fig. 1. Experimental Arrangement. The relative positions of the
reactor, spectrometer, and sulphur pills are shown. The latter were
placed 38-cm to the left of the reactor's midplane center line and
35-cm above the horizontal plane shown. The dotted circle represents
one of the spherical source regions used with the calculations.
which is supported in vacuum between widely spaced silicon-gold surface.
barrier diodes. To shield the diodes against intense gauma-ray fields
and very fast neutrons, a lead shield and a tungsten-lined 18-in.-long
lead collimator is used. By further shielding against slow neutrons
with 'Li on all sides, the count rate from very low-energy neutrons is
drastically reduced and the spectrometer is found to work satisfactorily
for the fast-neutron angular-fJux measurements above 1-1.5 MeV.
A neutron, after passing through the lead collimator, impinges on
the 'Lif layer and produces an alpha particle and a triton emitted in
opposite directions in the center-of-mass system. These charged parti-
cles share between them the energy of the incoming neutron plus 4.78 MeV,
the Q value of the Li(n,a)T reaction. The two diodes placed on opposite
sides of the 'LiF, but in the shadow of the collimator, must each detect
a charged particle simultaneously before an event is accepted as a true
neutron count. This scheme greatly reduces background counts from gamma-
rays and from fast-neutron reactions produci the body of the diodes.
The remaining background is determined by substituting a 'Lif layer for
the 'LiF layer, and proper background subtractions are thereby made
possible.
Electronics
The block diagram of Fig. 2 shows the pulse routing from the two
diodes.
The output of diode A is sent to the charge-sensitive preampli.
fier A, whose output is split and sent to the side-channel amplifier A
and to a swmming circuit feeding amplifier C. Pulses from diode B feed
the other leg of the summing circuit. The circuitry is completely
symmetric about diodes A and B. Each of the side-channel amplifiers A
UNCLASSIFIED
2-04-058-781RI
0.001

PRE-AMP
DD-2
AMPLIFIER
DISCRIMINATOR
A
VARIABLE
DELAY
31 Meg
TEST PULSE
GENERATOR
AND CUR-
RENT SIMULA-
TION CIRCUIT
MIXER
CIRCUIT
DELAY
220 K
DD-2
AMPLIFIER
SIGNAL
000
PULSE-HEIGHT
ANALYZER
0.22
GATE
COINCIDENCE
CIRCUIT
100-V
POWER SUPPLY
A Meg
VARIABLE
DELAY
PRE - AMP
11 mo
DD-2
AMPLIFIER
DISCRIMINATOR
OscRimarok –
0.004
Fig. 2. Block Diagram of Shielded Diode Spectrometer Electronics.
Ti
..
.
-
.
and B has double-delay-line differentiation and a crossover pickoff
discriminator feeding the fast-coincidence circuit of 75-nsec resolving
time. This resolving time was chosen conservatively to avoid timing
drift problems during long runs. This long resolving time was poorer
than that of the DD-2 amplifier and discriminator, which, when properly
adjusted, have a walk of about 15 nsec for the factor-of-ten range in
pulse height in these measurements. Therefore, about a factor of 3 to
4 improvement in time resolution is easily attainable with any good
drift-free coincidence circuit and a properly adjusted commercial ampli-
fler with about a 10-nsec walk. The expected improvement in performance,
if coupled with improved shielding of the diodes, would significantly
extend the capabilities of the detector. A delay line is placed be-
tween each discriminator and the coincidence circuit to adjust for
fixed timing differences. An event that fires both discriminators
causes the coincidence circuit to activate the analyzer which then re-
ceives a pulse from the summing amplifier C. The delay line between C
and the analyzer ensures that the coincidence pulse and the pulse from
C arrive in the time sequence required by the particular analyzer in
use.
The side channels must, of course, be biased below the smallest
alpha pulses; otherwise, the efficiency of the spectrometer will be very
unstable and difficult to calculate for all neutron energies. Figure 3
shows a thermal-neutron pulse-height distribution from only one diode.
The dotted line represents the true alpha-particle pulse-height distri-
bution. It is obtained by feeding the output of only one diode to the
analog input of the analyzer and triggering the analyzer with a pulse
UNCLASSIFIED
2-01-058-785

-
WITHOUT DISCRIMINATION
Doo ALPHA PULSES OBTAINED BY
GATING THE ANALYZER WITH
TRITONS FROM THE OTHER DIODE
TRITONS
600
600
-
......
NUMBER OF COUNTS
-----
-
-----....................
.......------
-
-
H
ALPHA
PARTICLES
Ochase
0 20 40 60 80 100
PULSE HEIGHT
Fig. 3. Response of a single diode to thermal neutrons.
O
from the opposite diode whose side-channel amplifier has its discriminator
set just below the triton peak, i.e., above the alpha pulses. This scheme
allows the experimenter to set both side-channel discriminators optimally,
e.g., just below the alpha peak and above most of the smaller noise
pulses.
Spectrometer Response
The response of the shielded-diode spectrometer to three groups of
monoenergetic neutrons is shown in Fig. 4. It 18 linear, and as can be
seen in Figs. 5, 6, and 7, displays a Gaussian-like peak whose full
width at half height, a measure of the resolution, is a function of the
energy
The spectrometer energy calibration was obtained by removing the
vacuum-tight housing from the lead and ºli shielding and placing the
diode-end of the housing near neutron sources of thermal, 3 MeV, and
14.7 Mev, since these sources CORNL Graphite-reactor thermal column, and
0 to 300 keV accelerator using ?(, n)>He and 3H(d, n) *He reactions] were
too weak to be used with the long collimator and thermal shield. Figure
5 shows the thermal peak, Fig. 6 the 3-MeV peak with a contamination of
thermal neutrons, and Fig. 7 the 14-MeV peak. The contaminating pulses
associated with the 14-MeV neutrons become important below channel 130.
They arise mostly from (np) reactions in one diode, the protons then
proceeding to the opposite diode to give a coincidence count. These
pulses reached an intensity of about 2-1/2 orders of magnitude greater
than the 14-MeV peak, but were almost identically produced in the back-
ground run L1 replaced by 7.1).
Although the present scheme of background subtraction is very
accurate, because the same diodes are employed for both foreground and
.rar
UNCLASSIFIED
2-01-058-782 R1

tow
e chiamatinis
. - - * -
rec
.
. .
NEUTRON ENERGY (MeV)
-11-
20
40
60
80 100 120
PULSE HEIGHT
140
160
180
200
Fig. 4. Neutron energy versus pulse height when both alpha and
triton are detected by the diode pair.
-12
.
(x103,
UNCLASSIFIED
2-04-058-783

NUMBER OF COUNTS
caccccccccco
20
I bound
40 60
PULSE HEIGHT
80
100
Fig. 5. Pulse height spectrum from thermal neytrons with the two
diodes operating in coincidence, and utilizing the Li
31 reaction.
-13-
UNCLASSIFIED
2-04-058 - 784

3-MeV NEUTRONS
NUMBER OF COUNTS
..
.
SLOW NEUTRON PEAK
100
20 40 60 80
PULSE HEIGHT
Fig. 6. Pulse height spectrum from 3-MeV neutrons, spectrometer
shielding removed.
-14-
UNCLASSIFIED
2-01-058-822 R1

.
.
NUMBER OF COUNTS
_
C
160
Leccecooldoo
170 180 190
PULSE HEIGHT
200
210
Fig. 7. Pulse height spectrum from 14.7-MeV neutrons after back-
ground subtraction ('LiF foil replacing LiF foil). The spectrometer
was unshielded because the low source intensity made it necessary to
place the Li foil (and diodes) very near the source.
--15-
background measurements, shielding of the diodes is necessary in measur-
ing reactor spectra because the backgrounds are simply too high to be
subtracted with good statistical certainty, and also because the diode
characteristics change under the heavy gamma and neutron irradiation.
Subjecting the instrument to intense gamma radiation resulted in both an
increase in the width of the peak for monoenergetic neutrons and a channel
shift that is a constant for all pulse heights (1.e., a "baseline shift").
Both the width increase and the channel shift were measured by sending
pulses from a pulse generator through the spectrometer preamplifiers in
the actual operating environment. It was found that both were constant
(channel shift = 0) at low reactor power levels but began to rise linearly
when the rector power was raised past a critical level P... This behavior
is illustrated in Fig. 8. The value of P. increased with distance of the
spectrometer from the reactor face. The increase followed the gamma-ray
attenuation much more closely than the neutron attenuation and is therefore,
W9
mi?!
"
attributable to gamma-ray background. For all measurements reported here.
the reactor power was held below Pa. It should be pointed out that po
was <10% of the reactor's capability of 1 MW for • = 50 cm and 8 = 0°
(Fig. 1), whereas it was >1 MW for 2 = 41° and 52°, where the collimator
did not directly view the reactor. Better shielding of the diodes could
conceivably raise P. to 1 MW, for 0 = 0, so that the oº spectra could be
obtained at a 70-to 80-cm water depth with good reliability. The
simplest improvement would be that of increasing the diode separation
and thus improving the shadow shielding. One should also move the ºli
disk out of the throat of the collimator, farther from the diodes (Fig. 1)
to minimize the effects of gamma-ray and neutron scattering into the
"
.,,,
1:15:;:-..:.:*
:*
"..
.
.
.
.
..
..
...
.
.
UNCLASSIFIED
2-01-058-845 A
PULSE HEIGHT
(EPITHERMAL PEAK)

(a)

A, WIDTH AT HALF-MAXIMUM OF
EPITHERMAL PEAK (PHS CHANNELS)
0)
1
5
6 Pc 7 8
REACTOR POWER (kw)
9
10
Fig. 8. Response of shielded diode spectrometer at r = 10-cm and
= 0o. ) position of the epithermal neutron peak vs reactor power
and (b) width at half-maximum of the epithermal neutron peak vs reactor
power. The value of Pe is the reactor power below which the data are
taken at r = 10-cm and = 0°. P. increases with both r and 0.
-17-
diodes. A thin wafer of unclad - B,C could probably be used to better
advantage.
Spectrometer Efficiency
For the sake of simplicity, the overall spectrometer efficiency is
divided into two parts: 6, the collimator-independent part, and ec
that part due to the spectrometer proper. €, is numerically equal to
the solid angle of the collimator subtended at the center of the ºli
foil, and gives the flux at the center of the detector produced by an
angular flux of 1 neutron/cma-sec-steradian at the end of the collimator.
6. = 1.07 x 10-2 steradians for the collimator geometry of this
experiment.
€ (E) is the coincidence count rate for a flux of 1 neutron/om-sec
on the Oli foil. € (E) = No (90°) lw for thermal neutrons, where N 18 the
number of ºli nuclei lying within the projected area of the two diodes,
0(90°) is the Oli(n,a)r cross section at 90°, and tw is the average
solid angle of the diodes at the 'Li surface. Due to the coincidence
requirement, the effective solid angle for a point on the 'Li surface
increases from zero, at the rim of the diodes, to a maximum at a point
along the common axis of the diodes, for the simple case of thermal
neutrons. In Fig. 9 is shown the construction for integrating over the
ºli surface to obtain € (E) for circular diodes:
6
60 - $*()2trds Sacs) 068, 0u(x) = No(Eng),
where N(r) is the number of “Li nuclei per cma at r, o is the angle
between the incoming neutron and the element of solid angle dw subtended
-18-
UNCLASSIFIED
ORNL DWG 64 - 10796

AA12
I
DIODE
DIODE
GLIF
DIODE
AA/2-
R = 0.9 cm, 2D = 1.79 cm,
LIF THICKNESS = 580 kg/cm
Fig. 9. Geometric construction used in describing spectrometer
efficiency calculation. R = 0.9-cm, 2D = 1.79-cm, Lif thickness =
580ugr/cm2.

- 192
by a small increment of diode area da, and AA/2 is that part of the
area, lying on one of the diodes, that 18 defined by the overlap of the
dotted circle and the diode area. The dotted circle is the boundary of
the upper diode projected through point r. R is the effective radius of
the diodes. The dotted circle, shown in the lower part of Fig. 9,
describes the area into which a triton may fall for an alpha striking
any part of the upper diode. Because of the coincidence requirement,
only that area AA/2 of either diode can contribute to the solid angle
subtended at r.
For thermal neutrons, ole) = Comotor/451 = constant, sw = 1.35 ster-
adians, in which case a numerical integration yielded € = 3.2 x 10?.
Calibration of the instrument with neutrons from a thermal column
yielded €. = 3.5 x 10–3 (+10%). The instrument was calibrated against
a gold foil whose diameter and thickness to neutrons was the same as
the 'Lif layer so that the flux depression cancelled. The gold foil
activity was obtained for various sizes of discs punched out from a
large disc that replaced the material of the 'LiF film. The two sep-
arate exposures (Au activation and Lif counting) were each monitored
with gold foils exterior to the detector chamber.
The detector efficiency € (E) for neutrons in the Mev region of
energy was similarly calculated and the results are given in Fig. 10.
The average differential cross section for the 'Li(n,a)T reaction, where
known, was taken as o(1/2 - A) in the laboratory system. Here A is the
angle of inclination of both alpha and triton, where both are emitted
at equal angles to the incoming neutron. The value of A is zero for
thermal neutrons, about 50 for 1-MeV neutrons and 10° for 14-MeV neutrons.
UNCLASSIFIED
2-01-058-854

III
****
counts
-20-
I
FISSION SPECTRUM =
(CRANBERG et al.)
FISSION SPECTRUM
OBTAINED WITH THE
SHIELDED DIODE +
SPECTROMETER
40
50
60
70
80 90
PULSE HEIGHT
100
110
120
130
Fig. 10. Shielded-Diode Spectrometer Efficiency vs Neutron Energy.
This is the part called Es in the text, which is the probability of
-21-
It has no effect on the calculation of tw, to first order. The effec-
tive area of 'LIF 18 the same, except that this contributing circular
disc moves to the left in Fig. 9 by D x A. For this reason, the LIF
disc was made larger than the diode. The differential cross section
for the Ol1(n,a) reaction is known at 14 MeV and below 2.5 mev. In
these regions, ola/2 - A) is nearly equal to Totalin,C) This value
13.3
was taken for all energies with some degree of assurance because Omotay
displays no noticeable resonances between 2.15 and 14 MeV.
In calibrating the efficiency of the spectrometer, no correction
was made for double-charged-particle emission in ºli that would supple-
ment the ºu(n,a)r reaction, or for double-charged particle emission in
"Li that would overestimate the background when the 'Lif foil replaces
the 'LiF foil during a background run. There are two almost indis-
tinquishable side reactions in ºLi which are leading candidates for
study because they have the least negative Q value: “Li(n, an ")*He and
Li(n, a)>He-*He + n(10-21-sec lifetime). Similarly, with 7-1 we have
?Li(n, Tnº)*He and PL1(n,)He-*He + n. These reactions have been studied
in some detail by R. Batchelor and J. W. Towle. They all yield an
effective three-body breakup with the neutron taking off about 1-MeV
energy, on the average, for an incident neutron energy of 5-MeV. Assum-
ing this holds true at higher energies, an 8.3-MeV neutron producing
the contaminating 'Li(n, dn")*He reaction and its counterpart will, on
the average, give as much energy to the d and “He as a 1-MeV neutron
will give to the *He and T in the principal reaction of the spectrometer,
the Oli(n,Q)T reaction. With ?Li, a 9.3-MeV neutron can deposit the
same energy in the two diodes, on the average, as a 1-MeV neutron in
-22-
the 'Li(n, a)t reaction. The combined cross section for the two reactions
in 'Li is more than half of the combined cross section for the interfering
reactions in ºli, so that there should be a reasonable amount of cancel-
lation of effects from these side reactions when 'Li is used in a back-
ground measurement.
The effects of the side reactions can be summarized as follows:
(1) In general, the spectrometer is useful in the Mev region only for a
spectrum that has few neutrons above about 8 MeV; (2) However, if the
biases on the side channel amplifiers are set to just accept the Q + T
in the 'Li(n,a)t reaction, the competing reactions in 'Li and 'Li are
biased out to a considerable extent because a) the 3-body breakup will
reduce the probability that the two charged particles will strike both
diodes since they are not emitted in opposite directions in the center-
of-mass system, b) the 8- or 9-MeV incident neutron in the side reactions
brings in much more momentum than the l-MeV neutron in the 'Li(n, a)T
reaction, so that the charged particles in the side reaction are much
more forward peaked. They will therefore be less likely to strike the
diodes and they also will lose much more energy in the 580-48/cm2 °LIF
plus 50-4g/cm? formvar; (3) In the case of a reactor spectrum, the flux
at 1-MeV and the side reactions are unimportant for any bias setting.
However, the spectrum is quite flat for the special case of the for-
ward-directed flux at 50-cm water penetration, i.e., blu = 1, 50-cm, E)
is roughly independent of E from 1- to 8-MeV. One would expect the
measured spectrum at 50-cm to be too high in the l- to 3-MeV region
because of the side reactions since 'Li has a more favorable Q-value and
cross section for their production than has "Li. But a comparison of
-23-
the measured spectra with the NIOBE calculation (Fig. 16) shows about the
same degree of disagreement of the two for ølu = 1, 10-cm, E) as for
Ølu = 1, 50-cm, E). This strongly suggests that the side reactions in
ºli and 'Li are biased out by the spectrometer.
Several independent checks of both differential and integral effi-
ciency have been made. Figure 11 shows the results of two independent
measurements of the fission spectrum of 2350; a measurement with the L1
spectrometer, and a time-of-flight measurement made by Cranberg et al."
There is reasonably good agreement within the statistical limits of the
present experiment. This spectrum was obtained by placing a 1/32" 2350
foil in front of the cylindrical spectrometer vacuum-tight housing, all
covered with ca, all located at the end of the graphite thermal column
of the ORNL graphite reactor. The count rate was very low, requiring
about 32 hours to obtain the data from this low flux source. Therefore,
the statistical uncertainty remains large as shown. In Fig. 12 the
spectral intensity at the BSR-I core-shield interface (r = 0 and 0 = 0°)
has been compared with that measured at the same position with a proton-
recoil telescope by Cochran and Henry." The agreement in shape and
magnitude is excellent.
Data Treatment
A typical pulse-height distribution for the spectrum at the core-
shield interface is shown in Fig. 13. In this plot, background has been
subtracted. At the end of each foreground run, the background was measured
by replacing the Lif film with a 'Lif film of almost identical size and
mass. By filling the spectrometer's collimator (Fig. 1) with water, it
was established that the background from the neutrons leaking through the
sides was negligible with the collimator geometry employed.
UNCLASSIFIED
2-01-058-885 R1
· 5X

')
.W
. steradian
. sec. MeV
-24.
El (neutrons.cm
SHIELDED DIODE OLI SPECTROMETER DATA
+ PROTON RECOIL SPECTROMETER DATA: FROM R. G. COCHRAM
AND K. HENRY, ORNL-CF-53-5-105, (1953)
8
10
12
NEUTRON ENERGY (MeV)
Fig. ll. Uranium-235 Fission Spectrum (from thermal neutrons)
obtained with Shielded-Diode Spectrometer compared with Spectrum
Measured by Cranberg et al.
-25-
UNCLASSIFIED
2-04-058-846R

EPITHERMAL NEUTRON PEAK ;

:
F
p=0
0 = 0°
6.-
1
-
-
counts
-
-
-
........
..
2--
.
..
100 !
-
0 20 40 60 80 100 120 140
PULSE HEIGHT
2 4 6 8 10
NEUTRON ENERGY (MeV)
Fig. 12. Comparison of the BSR-I Fast Neutron Spectrum Measured
with the Shielded Diode Spectrometer and with a Recoil-Proton Telescope,
both normalized to a reactor power level of 1 watt, both given in terms
of neutrons/ster-see-MeV-watt.
:*::*
*
V
****
**
*
*
hon w
, 1:
.
..'
- 26.
As can be seen in Fig. 13, the peak due to epithermal neutrons
extends to about 1 MeV. The subtraction of this peak (solid line in
Fig. 13) resulted in a large error in data below 1 MeV while it had negli.
gible effect on the data above 1.5-Mev. Therefore, only the data above
1-MeV have been report here.
The net pulse-height distributions were converted to absolute spec-
tral intensities by correcting for the efficiency shown in Fig. 10, and
for the solid-angle factor of the spectrometer (, = 1.07 x 10-2; see
above). All spectra were normalized against sulphur-monitor activation,
and the sulphur monitoring was used to tie in the spectral measurements
with the absolute reactor-power calibration that is discussed below.
Results and Discussion
A. The NIOBE Transport Calculation
The NIOBE (Numerical Integration of the Boltzmann Transport Equation)
code was developed by S. Preiser, et al. several years ago and was
designed to handle multilayered shells of shielding material. It solves
a series of integro-differential equations by an iterative method, and
is semi-analytical. The code utilizes differential cross sections to the
P-8 approximation and can be used only with spherical geometry. Certain
precautions must be observed in applying the NIOBE code.' The source
radius must be large enough to intercept the first angular mesh point
(about 80 from a radial line for the 16-angle mesh size used in the present
calculations, and the radial mesh must be large enough so that at a shield
penetration of one radial increment, the perpendicular to the first engular
ray does not pass through the inner shell if the relaxation length in the
inner shell is very different from that of the adjacent shell. The
UNCLASSIFIED
2-01-058-853
10-5
i E, DETECTOR EFFICIENCY

10-6
-27-
10->
02
6 8 -
NEUTRON ENERGY (Mev)
10
12
Fig. 13. Pulse-Height Distribution of Neutrons from the BSR-I at
r = 0-cm and 0 = 0°.
-28-
running time for a single problem on an IBM 7090 computer was about 45
minutes for a reactor-source region of 26-cm radius and 80-cm thickness.
We used 60 radial mesh points, 16 angular points, and 36 energy points
for this problem.
B. The DTK Calculation
The DTK code is a version of the DSN code having more rigorous
conversion criteria. It has very recently been run by G. E. Whitesides
(Central Data Processing Facility, Oak Ridge Gaseous Diffusion Plant)
and employed 21 energy groups above 1 MeV. He used the GAM-II code to
generate cross sections tº (1t has a cross-section library of 99 groups),
interpreting the GAM-II cross sections in terms of cross sections accept-
able for the DTK code.
The P-O and P-l scattering cross sections that were used were ad-
justed by the P-2 cross sections in a manner that "preserved the second
spatial moment of neutrons moderated from a point source in an infinite
medium and preserves all energy loss moments."
The problem discussed below (26-cm source radius, 80-cu-thick water
shield) took 11 minutes to run for an energy grid of 26 groups, an
angular grid of 16 angles and a radial grid of 60 points.
C. Comparison of NIOBE Predictions with Experiment
A NIOBE calculation that was performed to compare with experimental
..
results had assumed a spherical fission source, 26-cm radius.
The radial
..
.
*
source distribution was the same as that measured along the BSR-I mid-
-
*
.
plane center line (Fig. 14). This calculational source matched the power
of the 38-cm x 38-cm x 60-cm high reactor as determined by the power cali-
bration discussed below. The two are normalized at 2.4 x 10-4 watts.
R
.
.
-
2
RA
UNCLASSIFIED
ORNL-DWG 64-10797
@

MEASURED SOURCE DISTRIBUTION
s(r) (fission neutrons /cmº)
-29-
..
..
..-
..
Oo
8 12 16 20 24
RADIAL DISTANCE ALONG THE CENTER LINE (cm)
Fig. 14. Calculational Source Distributions vs Distance from
Center of source. This represents the distribution measured along the
reactor's mid-plane center line.
-30-
Figure 15 shows a comparison of the measured oº flux, Ølu = 1.0,r,E)
and the NIOBE values of Olu = 0.989,r,E), with r measured from the surface
of the source regions in each case. The overall agreement is quite good
everywhere above 1.5 MeV, where the spectrometer is most reliable. At
r = 50-cm, the reactor power was 70 KW and the gamma flux was 7 * 10* r/hr.
In Figs. 16 and 17 are shown $10.755,r,E) and Ø(0.617,r,E) respect-
ively for both NIOBE and measurement. The overall agreement is very
good, except at smaller values of r where, in the experiment, the corner
of the reactor may be an important additional contributor. For the
6(0.618,40-cm, E) measurement, the reactor power was 5 x 10% watts and
the gamma flux at 50-cm was about 10° r/hr.
Comparison of DTK Results with Experiment
In Figs. 18, 19, and 20 are shown the predictions of the DTK code
for
26-am-roddu
-radius source whose radial distribution is identical to that
measured along the BSR-I's mid-plane center line.
In Fig. 18 the values of d(u = 0.978,r,E) are everywhere within a
factor of 2 of the measured oº flux, 01.0,r,E), for energies above
1.5 MeV where the spectrometer should be quite reliable. The measured
angular. flux at 41° and at 52° is compared with DTK results for $10.745,r,E)
and $(0.650,r,E) shown in Figs. 19 and 20 respectively. Overall agree-
ment with the experiment is very good at 30- and 40-cm from the source
region, where the corners of the reactor are far from the spectrometer's
"line of sight," and at energies above 1.5 MeV, where the spectrometer
is most reliable.
Neutron Spectrum Within the Reactor
A measurement was made of neutrons leaking from a region 7.8-cm
inside the reactor by removing the front, central fuel element in Fig. 1.
- 31.
UNCLASSIFIED
ORNL DWG 64-2936 R38

oglo
r = 0 cm
odpoqu002609co
10 cm
boo
EXPERIMENTAL
7
... NIOBE CALCULATION
20 cm
000 Kool
orlog. 0010
OD
30 cm
. .
.
.
.
ofio
.
•14=1,1, E )(neutron.cm-2. sec. Mev.steradian' per 2.4 x 10"W)
On one on one on one or
10 cm
1:
.
.,
.
-
i
.
.
..
2
4
1012
6
NEUTRON ENERGY (MeV)
Fig. 15. The NIOBE Calculated Values of Alu = 0.989,r, E) compared
with measured values of (1,r,E), both normalized to a power of 2.4 x
10-11 watts.
-32.
UNCLASSIFIED
ORNL DWG 64-108888
10-6

.
..
..
..
.
::,
-
-20 cm
i
.
..
L
do
1
-EXPERIMENTAL
.sec-. Mev.steradian' per 2.4 x 10-"W)
"
'
.....
.
NIOBE CALCULATION
----......
......
L
S
.
.
.....
.....
.
......
.--.
40 cm
....
...
.
..
..
.
...
,
..
.
.
..
.
l..
.
.
$14=0.755, 7, E ) (neutron.cm
..
.
.
-
m.
...ochen
10-12
0
2
4
6 8
NEUTRON ENERGY (MeV)
10
12
Fig. 16. Measured and NIOBE Calculated Values of Cu = 0.755,r,E)
(410), 26-cm Radius Calculational Source.
-33.
UNCLASSIFIED
ORNL-DWG 64 - 108878

.
"
...
poo dog
pooooooooooo
Ti
- 10 cm
20 cm
EXPERIMENTAL
-
Ó
.............
.:-NIOBE CALCULATION
.
. .
. .
. . .
. .
9. 30 cm
..
.
....
-
Q .--.
20000 2000 2003
40 cm
(p=0.617,1, E ) (neutron.cm?. sec'. Mev S. steradian' per 2.4 x 10-"w)
1
.
..
.
.
.
-.
.
-
.
.
..
-
-
-
-
-
-
-
.
10-12
2
1 8
NEUTRON ENERGY (MeV)
10
12
Fig. 17. Measured and NIOBE Calculated Values of ølu = 0.617,1,E)
(520), 26-cm Radius Calculational Source.
-34-
UNCLASSIFIED
ORNL OWG 64 - 2936 R3A

1
Lil Oecomposite r = 0 cm
1.10 cm
.
S
EXPERIMENTAL
b
$ 20 cm
. So
DTK CALCULATION
osobe 9.900
. steradian' per 2.4 x 10"W)
*
"
...
of... -
30 cm
non or an owo wo
:I
I.
.
.
cm
-
..
.
ooo
. sec. Mev
.
.
:.
.
.
..........
............
..
$18=1,r, E) (neutron.cm
........ www.
-
O**
.
.
.!-
-
- ......
..
..
2
4
8
6
B
NEUTRON ENERGY (MeV)
10
12
Fig. 18. DTK Calculated Values of Olu = 0.978,r,E) compared to
measured values of 0(1,1,E) for a 26-cm Radius Calculational Source.
- 35-
UNCLASSIFIED
ORNL DWG 64 - 10888 A

.
.
.....
.
..
..
1
boogpunt r = 20 cm
.
.
.-
.-
-
teradian 'per 2.4 x 10
-
-
1
00%,0 o
- EXPERIMENTAL
--.
-...
o
----
-----
OLDTK CALCULATION
40 cm
Sec.
.
.....
.
.
...
...-
..
-
1
El (neutron.cm
..
O
.
.
-
-
...
.
.
..
oo... .
..
to...--
-..
.
..
... -
-
-
..
10-12
2
4
8
1012
6
NEUTRON ENERGY (MeV)
Fig. 19. DIK Values of ølu = 0.745,r,E) and experimental values
of Olu = 0.755,r, E). A 26-cm radius DIK source was used, and the
calculational source density duplicated that measured along the reactor's
mid-plane center line.
- 36-
UNCLASSIFIED
ORNL-DWG 64 - 10887A

9000
0000 soorten
DawodooDoo
p= 10 cm
EXPERIMENTAL
lobo o
. 20 cm
0!. DTK CALCULATION
Ti 30 cm
190000 4000
1.
Oooo
bong
pool
$18=0.617,1,E) (neutron-cm2. sec. Mevr.steradian' per 2.4 x 10-"W).
-
-
-
-
-
10-12
O
.
2
2
4
6
- 8
NEUTRON ENERGY (MeV)
10
12
Fig. 20. DTK calculation of 0(4 = 0.650,r,E) compared to measured
values of Olu = 0.617,r, E).
-37-
This can be denoted by Ø11, r = -7.8 cm,E). The results, shown in Fig. 21,
are compared with a f1ssion spectrum and with $11, r = 0,E) of Figs. 15
and 18. As expected, the shape of spectrum inside the reactor is inter-
mediate to the other two shown. The measured reactor spectrum is a reason-
ably accurate representation of both the scalar flux spectrum and the
spectrum of angular flux. Both became identical over a region where the
flux gradient is zero, and from Fig. 14 it is seen that the flux gradient
is rather small at -7.8-cm from the reactor's edge. Since the reactor
.
contiguration was different for this measurement, the reactor power cali-
bration described below does not apply. The spectra in Fig. 21 are
therefore arbitrarily normalized.
.
.
Reactor Power Calibration
The power density was mapped throughout the reactor volume by copper-
wire-activation techniques combined with an absolute calibration of the
copper activation by means of very thin layers of uranium. The folls,
which were of the same enrichment as the reactor fuel elements, were
exposed both in the reactor and in a standard pile. All power-density
measurements were related to the sulphur pellet monitors used in the
spectral measurements so that the measured spectra could all be normal-
ized to a single, absolute value of reactor power. Further details of
the power calibration are given in an ORNL-TM in preparation, and in the
ORNL Neutron Physics Division Annual Report.
In both the NIOBE and DTK calculations, the spherical source region
.
1
.
*
.
:
F
that mocked up the reactor was given a power density identical to that
along the reactor's mid-plane center line. With this constraint, the
radius of the calculational source region was increased until the power
UNCLASSIFIED
2-01-058-851

--- AT r=0; 8 = 0°
AT r=-7.5 cm; 0 = 0°
- FISSION SPECTRUM
$ (v=1, E) (arbitrary units)
-38-
C
12
6
8
10 -
NEUTRON ENERGY (Mev)
Fig. 21. Comparison of Uranium-235 Fission Spectrum with BSR-I
Spectrum at r = 0(0 = 00) and at 7.5-cm within the Reactor (r = -7.5,
0 = 0°, along horizontal mid-plane center line ... see Fig. 1).
.
-39.
of the calculational source was equal to a reactor power of 2.39 x 10 db
watts; 1.e., comparisons of NIOBE and DTK results with experiment are
normalized to an absolute reacter power of 2.39 x 10- watts. This value
is an arbitrary value that has only historic meaning of no importance to
this presentation.
Conclusions
The shielded-diode spectrometer developed for these measurements can
successfully cope with the intense gamma-ray flux (up to 3 x 10°R/hr) and
low-energy-neutron flux in measuring high-energy-neutron spectra near a
reactor. The side reactions in L1 and "Li, which would place 8. to ?
9-MeV neutrons in the 1-MeV region of the spectrometer, can be completely
ignored in all cases except at deep penetrations where the spectrometer
18 aimed at the reactor core. If the side-channel biases are set suffi-
ciently high, as in these measurements, the comparison of calculation
and experiment indicates that the side reactions are biased out.
This spectrometer should be considered as only a pilot model.. The
resolution can be improved by decreasing the LF and formvar thickness
and the sensitivity increased by increasing the detector area. A mosaic
of detectors could be used on each side of the LIF, each element of the
mosaic going to a separate preamplifier and amplifier with circuitry to
eliminate low gamma pulses (a "rug sweeper") before sending the pulses
to an adder circuit.
The 1-Mw reactor could not be run above 70 kW for the oº spectrum
at 50-cm penetration because of excessive gamma noise in the diodes. At
angles of 41° and 52°, a 1-MW power level was used. Therefore, improve-
ments in the spectrometer collimator alone should increase the capawility
-40...
11
of measuring the oº spectra at greater depths by an intensity factor of
14 for the present reactor.
Measurements of directed flux that were made over a range of in-
tensity of 104 were reproduced by two calculations within a factor of
2 of the measured intensity.
Acknowledgements
The authors are indebted to Dr. F. C. Matenschein for helpful dis-
cussions, continued Interest and encouragement, to T. A. Love and R. E.
Zedler for their part in spectrometer design, and to H. A. Todd for
assembling the circuitry. S. K. Penny and D. K. Trubey helped clarify
the shielding problems. G. E. Whitesides adapted the DTK code to give
a detailed spectrum in the MeV region, ran the codes, and helped remove
errors in source strength. Henna Francis ran the NIOBE code. G. T.
Chapman and T. V. Blosser have contributed generously to the success of
of the power calibration. One of us (M.S.B..) is grateful for the support
received from the v. S. State Department and the Pakistan Atomic Energy
Commission.
..!41-
BIBLIOGRAPHICAL REFERENCES
1.
Goldstein, H., Fundamental Aspects of Reactor shielding, Addison-
Wesley Publishing Co., Inc., Reading, Mass. (1959). Chap. 4 con-
tains a description of the ORNL (alumimm clad) Bulk Shielding
Reactor.
Goldberg, M. D., et al., "Angular Distributions in Neutron-Induced
Reactions, Vol. 1, BNL 400, pp. 3,6,9.
Batchelor, R., and Towle, J. H., Muclear Physics 47, 385 (1963).
Cranberg, L., et al., Phys. Rev. 103, 662 (1956).
Cochran, R. G., and Henry, K. M., Fast-Neutron Spectra of the BSF
Reactor, ORNL CH-53-5-105 (May 29, 1953).
Preiser, S., et al., A Program for the Memberical Integration of the
Boltzmann Transport E uation, A.R.L. Technical Report 60-314, u. s.
Department of Commerce, Washington, D..C. (December 1960).
Penny, s. K., ORNL Shielding Irformation Center, private comunica-
tion.
The DTK code was obtained from J. W. Warlton of Los Alamos Scien-
tific Laboratory, Los Alamos, New Mexico.
Carlson, B., Lee, C. E., and Warlton, J. W., "The DSN and DC
Neutron Transport Codes," LAMS 2346, Feb. 12, 1960.
Joanou, G. D., arid Dudek, J. S., "GAM-II, a B 3 Code for the Cal-
culation of Fast Neutron Spectra and Associated Multigroup Constants,"
GA 4265, Sept. 16, 1963.
11. Lee, C. E., "Discrete S. Approximation to Transport Theory," LA 2595
(1962).
12. Klema, E. D., et al., "Recalibration of the X-10 Standard Graphite
Pile," ORNL-1398 Toct. 1952).

LE
-
.
.
***
17
DATE FILMED
4/28 165





--
--
-
-
-
- -
-
:
-
-
T
4.
*
*
*
-
*
:
A
-.
.
-
.
..
-
-
-
-
- LEGAL NOTICE -
This report was prepared as an account of Government sponsored work. Neither the Unitesi
States, nor the Commission, nor any person acting on behalf of the Commission:
A. Makes any warranty or representation, expressed or implied, with respect to the accu-
racy, completeness, or usefulness of the information contained in this report, or that the use
of any information, apparatus, method, or process disclosed in this report may not infringe
privately owned rights; or
B. Assumes any liabilities with respect to the use ol, or for damages resulting from the
use of any information, apparatus, method, or proce88 disclosed in this report.
As used in the above, "person acting on behalf of the Commission” includes any em-
ployee or contractor of the Comuission, or employee of such contractor, to the extent that
Buch employee or contractor of the Commission, or employee of such contractor prepares,
disseminates, or provides access to, any information pursuant to his employment or contract
with the Commission, or his employment with such contractor.

.-
'
" H
11'
i
n
1
IL
END



ht
.
I
TER