51
bem
I
.
.
ber
| OF T
ORNL P
1604.
:
1.1.4 1.6
MICROCOPY RESOLUTION TEST CHART
NATIONAL BUREAU OF STANDARDS -1963
I
!...INIL
.,PJI!!
:
-.
*--*/2*7
T
"*,*-
**.*.*.*..",+"
O
"
-
.11
.
7217.
.
...-
1
7
..
-"
"."
**
" *"."
--
- ---- *
/f-
- -* :-:- --

ORNU -P-1604
Conf. 650928-6
OCT6.14
Method for Obtaining Bursts of Polarized Neutrons
with Energies 10 - 700 kev*
auritars0 Lonca
REI:EASED FOR AUTOUNCEMENT
J. W. T. Dabbs and J. A. Harvey
Oak Ridge National Laboratory
Oak Ridge, Tennessee, USA
IN NUCLITAR SCIECE' ABSTRACTS
As first suggested by Schwinger\), fast neutrons can be polarized
by scattering from helium at w 90° (C.M.) and might be moderated without
loss of polarization.
The observation was made that the energy at which neutrons would
be produced in an electron linear accelerator with a heavy target (w1.2
MeV for 100 MeV electrons on w or U) agrees almost exactly with the peaks
in polarization and cross-section for scattering of neutrons by the well-
known p-wave resonance in "He, for an angle (Lab) of 90° $ 15º. The
calculated polarization of such a scattered beam 18 - 50 to 60 percent
with only slight dependence on angle, and the effective intensity of the
scattered beam is ~0.5 percent of that without the scattering (solid angle
factors included). The scattering requires al cm thickness of liquid
helium. For the proposed ORNL linac, the estimated heat input from the
scattering alone would be sufficient to evaporate about 0.5 liter/hr, and
in actuality, about twice this evaporation rate should be expected. The
median neutron energy after a single 90° scattering from *He is approxi-
mately 700 keV.
A survey of possible moderators which could be expected not to
depolarize the beam led to 12c as the lowest-mass spin-zero material with
small (p-wave and d-wave) scattering phase-shift differences for energies
below 1 MeV. Because of the high mass and small "slowing-down power" of
carbon, the number of collisions required 18 large, as is the escape
probability for usable sizes of moderator (in the present case, a graphite
:
A
-
....
.
.
...
.........
....
...
... ini
.....
.......asi
block of dimensions 10 x 10 x 8 cm was chosen). In spite of these diffi.
culties, a usable intensity appears to be available down to energies as
low as ~ 10 keV.
The estimated intensities for both graphite and H,0 (for comparison)
moderators are given in Table I, as obtained from a Monte Carlo calculation.
The N(E) are the numbers of neutrons which escape from the moderator in
the stated energy interval A E for each neutron incident on the moderator.
The values of I are the averages over 1000 neutrons of the time delay before
escape, converted to a length at the final velocity: 'that is,
(1)
where lg is the actual path lengtih associated with the ith trajectory
within the moderator. The standard deviation in I is denoted by AL. The
number of collisions in passing through the interval 4 E 18 also given.
A Monte Carlo calculation which also included the polarization of
the escaping neutrons was carried out, based on the Wolfenstein formu-
lation (2) with the parameters calculated from the phase shift formulas
given by Giamati and Thaler." The phase shift differences were pessi.
mistically chosen to be somewhat larger than the correct differences by
using a slower variation with energy than expected in extrapolating the
known values to the lower energies. The phase shift for s-wave and the
differences for p and d waves were taken to be so = -77 E,
A4, = - 1.5 E, and AS, - +1.5 respectively; here E 18 in MeV and
the values are in degrees. It should be noted that a coordinate trans-
formation was required both before and after each collision to utilize the
Wolfenstein formulation. For 1000 neutrons, the results are given in
Table II. P is the portion of the polarization remaining. The agreement
with the first calculation (Table I) is gratifying. The loss in polarization
18 seen to be negligible; indeed, no appreciable depolarization occurs at
any energy according to the calculation.
Since the inception of this approach, a method which is clearly
superior in the energy range 0 - 10 keV has been demonstrated by Shapiro,
Taran, and coworkers. (4) Their method, if applied to the linac, would
consist of passing water-moderated neutron burst through a dynamically
polarized proton target." Their method has the advantage of higher in-
tensity, since the values in the lower half of Table I would apply to the
intensity incident on the polarizer. The intensity loss associated with
the liquid helium single scattering (factor 200) would not apply; a loss
of a factor of 10 because of the small size of the polarizer'°) and a
factor of 5 for transmission loss in the polarizer gives an overall factor
of -4 in favor of their method in addition to the intensity factors between
the upper and lower halves of Table I. The polarizations expected from the
two methods are rather comparable. It would appear that the two methods
are complementary with respect to energy range and that together they will
provide means for obtaining bursts of polarized neutrons over the entire
range 0 - 700 keV with only slow variations of polarization with energy.
Questions of background, such as that from direct transmission to
the moderator and from scattering by the surrounding air have not been
studied intensively; it appears that an evacuated target room may be required.
The authors are especially indebted to J. G. Sullivan and A. M.
Craig, Jr. of the ORNL Mathematics Division, who carried out the Monte Carlo
calculations. We wish to thank A. Michaudon for an illuminating discussion
regarding polarization losses during moderation.
- .......
..
..........
..
.
.
.
....
..
...
.
in
....
..
.
.,'2'* .,..--:--.-..-',*
4
REFERENCES
* Research sponsored by the U. S. Atomic Energy Commission under
contract with the Union Carbide Corporation.
1. J. Schwinger, Plays. Rev. 69, 681 (1946).
2. L. Wolfenstein, Ann. Rev. Mucl. Sct. 6, 43 (1956).
3. C. C. Giamati and R. M. Thaler, Nucl. Phys. 59, 159 (1964).
4. P. Draghicescu et al., Phys. Lett. 12, 334 (1964); F. L. Shapiro,
International Conference on the Study of Nuclear Structure with
Neutrons, Antwerp, Belgium, July 19-23, 1965.
5. See, e.g., C. D. Jeffries, Dynamic Nuclear Fularization, Interscience,
New York, 1963.
6. Pointed out by A. Michaudon.
TABLE I
Neutron Moderation; Monte Carlo Calculations Based on 1000 Neutrons
AE
N(4 E)
ib(cm)
4L(cm)
No. Collisions
in AE (av
700-100 keV
100-10 keV
E = 700 keV; Graphite 10 x 10 x 8 cm
0.978
5.76
3.75
0.021 12.0
4.6
0.0009 14.0
0.00002 14.0
10-1 keV
4.9
1-0.1 keV
5.7
1000-100 keV
3,4
100-10 keV
1.78
10-1 keV
E = 1000 keV; Water 10 x 10 x 2 cm
0.822 2.44 0.74
0.087
1.51
0.044
1.43
1.46
0.021 1.37
1.56
0.012
1.60
0.008 0.95
2.2
1-0.1 keV
2.0
100-10 eV
1.27
2.0
cleV
1.06
3.2
AN(4 E) 18 Praction of neutrons escaping within A E.
BI 18 effective source distance behind exit face of moderator.
CAL 18 standard deviation in I.
.........
..........
....................
..
.
TABLE II
Moderated Neutron Polarization; Monte Carlo Calculations Based on
1000 Neutrons
Graphite, 10 x 10 x 8 cm
AE
No. Collisions
in AS
AB
700-100 keV
100-10 keV
10-1 keV
1-0.1 keV
N(AE)
0.977
0.022
0.0010
0.00005
I (cm)
5.7
12.4
24.6
15.2
AL
3.7
4.5
4.3
4.2
P
0.998 + 0.001
0.993 + 0.003
0.988 = 0.006
0.983 $ 0.008
No. Collisions
3
14
14
....
.
.
..-
.'.
.
.!"..h
:51'
71 '
"
WR
I TT
:..'

El.
m
en tot.
times the most wanted tow.
i
rar-
a.starTVIK
.
:
.'
IT
IZVOR:
11/8/65
DATE FILMED
END
is to
t
v
.
t
:
.
.
..
.'
L
E
::..
1
.
1.
*
.
S
....
S
TA