Publishing Co .

, Amsterdam
1 .B :3 .A Nuclear Physics 24 (1961) 400-411 ; C@ North-Holland
written permission from the publisher
Not to be reproduced by photoprint or microfilm without

MEASUREMENT OF THE ENERGY OF THE GAMMA RADIATION

FROM NEUTRON CAPTURE BY HYDROGEN
J. E. MONAHAN, S. RABOY and C. C. TRAIL
Argonne National Laboratory, Argonne, Illinois t

Received 16 January 1961

Abstract : A scintillation spectrometer with an anticoincidence annulus of NaI is used to measure

the energy of the gamma ray that follows the capture of a neutron by hydrogen . The measure-
ment is made simultaneously with the calibration of the spectrometer system in terms of six
reference gamma rays . The results indicate a value of 2219 ±2 keV for the binding energy of
the deuteron . A discussion of this result relative to other measurements of this and related
quantities is presented.

1. Introduction
The application of computational techniques 1 ) to the analysis of the spectra
of gamma, rays obtained with a scintillation spectrometer with an anticoinci-
dence annulus of sodium iodide 2) has led to convenient and precise determina-
tions of the energies of gamma rays 3) . It was decided to apply this technique
to the measurement of the energy of the gamma ray resulting from the capture
of neutrons by hydrogen. In a review article by Wapstra 4), the weighted
average of measured values of the binding energy of the deuteron is calculated
to be 2226.4±1.8 keV. Subsequently this has become the value commonly
accepted in the literature . It was the objective of this experiment to measure
with comparable accuracy a related quantity, namely the energy of the gamma
ray following neutron capture by hydrogen . Our results indicate a value of
2219±2 keV for the deuteron binding energy.
The discrepancy between these values is obviously greater than that conven-
tionally accepted as statistical . It is necessary, therefore, to devote a consider-
able part of this discussion to an analysis of possible systematic errors in our
own as well as in other measurements of this quantity.

2. Experimental Procedure
The technique of measurement is similar to that described in ref . s) . In order
to minimize drifts due to changes in amplifier gain as well as pulse-height shifts
due to different counting rates, the spectrum of the capture gamma ray is
obtained simultaneously with the spectra of various reference gamma rays.
t This work was performed under the auspices of
the U. S. Atomic Energy Commission .
400
 GAMMA RADIATION FROM NEUTRON CAPTURE BY HYDROGEN 401

The sources and the energies of the "standards" used in this experiment are
Nam (2753.3±1 .0 keV and 1368.6±0.3 keV) 5), Y88 (1836 .2±1 .7 keV and
898.7±0.8 keV) 3) and Bi" 7 (1063.9±0.4 keV s) and 570.8±0 .5 keV 3)) .

Fig. 1. Experimental arrangement of sources for the measurement of the energy of the gamma ray
resulting from capture of neutrons by hydrogen .
5
10 "
0.5708 MeV
0
CID
0.8987

4
10

N
1-
2
O
c

Q
N 1 ~3
V
Q

I 1 I I Î l Î I I
I +_~1 I ! I 1
l I ( i
10 1 1
0 50 100 150 200 250
CHANNEL NUMBER
Fig. 2. Spectrum of gamma rays obtained by simultaneous exposure of the scintillation spectro-
meter to radioactive sources (Na$°, Y88 and Bi207) and Pu-Be source in a barrel of water.

The experimental arrangement i.j shown in fig. 1 . A Pu-Be neutron source

immersed in a barrel of water (40 cm high and 40 cm in diameter) served as the
source of capture gamma rays . Th., radioactive sources (Na24, Yas and Bi207 )
402 J. E. MONAHAN et al.

were placed between the barrel of water and the spectrometer . The latter is
protected from background by 2.5 cm of boron-loaded paraffin and 5 em of lead .
A 0.64 cm. shield of Li's is placed between the spectrometer and the sources .
In order to reduce the number of events which normally would be scattered
through small angles into the crystal, a copper liner was inserted into the
collimator of the spectrometer .
Each run of data consists of the accumulation of counts over about two hours
with all sources in place. A typical spectrum obtained under such conditions is
shown in fig. 2. A total of ten such runs was made. In addition, in order to carry
out the analysis which is discussed below, it was necessary to measure the
spectra resulting from the following sources and combinations of sources:
(i) Na", (ii) Y88, Bi2°? and Pu-Be in H2 O, (iii) Bit °7 and Pu-Be in H2 O,
(iv) Y88 and Pu-Be in H2O, (v) Pu-Be in H2 O, (vi) Y88, Bi207 and Pu-Be in
D20 and (vii) no source (room background). In the remainder of this discussion
these spectra will be called "background runs" to distinguish them from the
"data runs" obtained with all sources in place. The data are recorded on punch-
ed paper tape in a form suitable for direct use on the digital computer
GEORGE at the Argonne National Laboratory.
The analysis of the data consists in describing the response of the spectro-
meter in the neighbourhood of a full-energy peak by a function of the form

C(x) ~ al G(x; a2, a3)+a4+asG(x+d ; a2, a3) .* (1 )

where C (x) is the measured response at pulse height x and a1 , . . ., a$ are para-
meters to be determined from the data by a modified chi-squared minimum
calculation; the quantities G are defined by

G (J ; oc2 , a3) = exp [-- (Y--12)2/a32], (')

In the analysis of a particular gamma ray, an attempt is made to remove the
influences of the other gamma rays in the spectrum. The full-energy peak of the
2.7533-MeV gamma ray is considered first . The data are corrected by subtract-
ing the normalized spectrum obtained in background run (ii) above, and a fit is
made to the corrected data. The parameters ocl , . . ., ab are also determined for
the full-energy peak of the 2 .7533-MeV gamma ray in background run (i) .
In this latter determination the spectrum of the normalized room
background
(vii) is subtracted before the parameters are evaluated . In
each of these cases
the normalization is-proportional to the ratio of the appropriate
life times of the
analyzer. The correction for background in the neighbourhood
of the full-
energy peak of the capture gamma ray is considered next
. The effect of the
neutron background, as determined in background run (vi),
is subtracted from
this peak in the data runs after normalization of the
background run to the
life time of the analyzer as before . The influence of tbc 24
N,~ sourer is removed by
subtracting the spectrum of this isolated source.,
background run (i), with a
 GAMMA RADIATION FROM NEUTRON CAPTURE BY HYDROGEN 403

normalization determined by the ratio of the areas of the full-energy peaks of

the 2.7533-MeV gamma ray of the data run and background run (i). This
procedure is repeated for each of the full-energy peaks in the data runs .
TABLE 1
Values ofparameters from least-squares fit of al exp[-- (0-120212321+24 to the full-energy peak of gamma
rays in a typical set (run 123) of data in the measurement of the energy ofthe hydrogen capture gamma
ray.

Energy Expected
12IL 22 a$ 12a
(MeV) (Counts) (Counts) X2 value of
(Channels) (Channels) x2

2.7533 11650+ 40 198 .020+0 .013 5.74 +0.02 387± 18 11 .6 19

2.2 (unknown) 6360± 90 159 .951±0 .028 5.13 ±0.09 350±100 8.5 11
1 .8362 7760±160 133 .077:L0 .031 4.45 ±0.12 300±180 11.2 9
1 .3686 27150± 80 100.407±0.007 3.66 ±0.01 1200± 40 6.3 12
1 .0639 15150±140 79.016±0 .022 3.28 ±0.04 140± 70 17.6 11
0.8987 17070±170 67.612±0 .017 2.82 ±0.04 690±160 4.3 ; 4
0.5708 37040±220 44.730±0.010 2.253±0.018 190±120 5.2 i 4

The result of the analysis of a typical set of data is shown in table 1 . The
analysis consists of a modified x2-minimum evaluation of the parameters
ai , . . ., as for the corrected data runs in the neighbourhood of each full-energy
peak. From eqs. (1) and (2) it is seen that al and 122 represent, respectively, the
amplitude and the mean pulse height of the full-energy peak and 1. is a measure
of its width. For those sources that emit gamma rays of different energies, it is
not possible to remove the influence of the higher energy gamma ray in the
vicinity of the lower energy peak . To take account of this the parameter 124 ,
which represents a constant level of counts superposed on the Gaussian, is
introduced in eq. (1) . As shown in table 1, this parameter is relatively large for
the lower energy peaks associated with each of the sources Na", Y138 and Bi207.
Finally a fifth parameter at, is introduced in eq. (1) to describe the response due
to back-scattered events within the crystal. These are caused by large-angle
Compton scattering in which the degraded gamma ray escapes the detector
through the entrance aperture, therefore, fails to trigger an anticoincidence
pulse. This type of event leads to a small peak not more than 250 keV lower than
the mean of the full-energy peak. Since the width and mean position, (12 2 --d) in
eq. (1), of this secondary peak can be calculated with reasonable accuracy, it is
necessary only to determine its amplitude a s in order to include these events in
the response function . For all full-energy peaks, with the exception of that at
2.7533 MeV, the contribution to the full-energypeakby125 is essentially zero . Even
in the case of the 2.7533-MeV peak, the correlation between a2 and 125 is suffi-
ciently small that any reasonable value of ocb produces a negligible shift in 0C .,-
The calculated value of x2 provides a measure of the "goodness of fit" of the
response function eq. (1) to the observed spectrum . The calculated and expec-
404 J. E. MONAHAN of at.

v
O=- cl GO
ti cq " Crl CM --4 N CO O O~
-+ ai
CD O
d+
O OO tiO OO MO ~O OO
â-4 id C; 0
-H I -H

r. Cà clZ
C~ O
ecq
G0 0
M .~-4
~O O
N
r- O
d+C4
c1 0
GÔ~
O'~ O
1-ice
0 0?
d~ e~
CD O
ce 004
.4
~-+ O W
MÔ mÔ MÔ MÔ
1 -H
II e,-~ MÔ d~r4 OO OÔ d*
-H " -H °' -H c' -H `° -H ' -l'I ° ' -H I -H
W
O
O

GV O d~ O CO O O o O 0 11~
O CC CC d~ O rr O eN

`° -H `" GÎ -H
°°
-H M -H °' -H ~ -H ' -H I -H

M = GO M cl cl 0 M CO c1 M ;g d* L-
CD M Oc7 "cl eMe CD -+
â
M~
ca O CM o -4 O cq O ea o L? oo .+ O -+ o m

CD ce 06 <:~ C6 C; m .6 id Q ~ -H I -H
P4 e-4
°' -H -H `° -H a -H -H
M -H -H

Û 0
cd Zn
6V O OM M d~ r" 00 ~ O
tU N d1 CD ."+ d~ "r t!O GV u9 ^ c-+ O r
G3, â~ r" O c1O d!O GOO M O A Q O~O MO OO 'e
~ c~ 14 iii .4~ ~ 0 m
w a ~ ~ ~ ~ 0 -0 = 0 ;; c
c
M -H M °H ° -H ' -H `° -H o I -H -H I -H
Ô y Ô
w w
71 N dé œ r.4 O N ti 00 t- O CD cl GO M 1p
- ce P. 0 ta r" 4 r" M t- 0 eo - = .-q CD O
en~
CO M
oo
O -4
~-+o 0
00 ~o ~o "-~o Gco ~o
a -W
â .1 -H -H Î ii ' -H I -H
-H '* -H ' -H
é
E,, ao
r.
R cc
4 ~, O M r- -4 n r" Co m cq r" 0 0 Go cq
0 cq r -4 r" M O O .-4 CV -4 -4 M r-4 GO CM O -1
91 M
= C4 O O O O d! O O O CO O r" O ti d! eN O OO CM
iz >% P4 ffl (Je Ô MÔ CD Ô 0Ô t-
m CD I* CD eMr-1 * Ô CD Ô eM
â ~° °: -H -H ° -H ' -H -H -H Î -H ' -H I -H
b~
a0)

M CO ~-+ - CD Ifi CO r" r" CO GO c1 M le 0

o GO -q 10 cq Go o = --, m ...0 .a o 1-1 O 1-4
w L-J O cl 0 CM O GO O CM O eN 0 -4 cq * 0 . 0 0
ed e --i Ci O *c; .40 ci O cd c; k6O "»*4 4 -4 C; OO M
ce

cß
m
00 CO * C- O O 0 CC cq r-1 d1
e~d 0 d+ d~ "-+ ut - GO e.V O O u' cl CM M 0 .-#
w eM 0 " -~ O GO O C~ O CO tr 0 0 F-+ O cl
0 94 -4 r:ô Md cc 0 ci r;ô ~c c ..+ dic cô ci
-11 I -H

u eD
a C= G~l d~ Ô ~ G0 "~+-~
O CC .-+ ~ .~ ~ Ô C~ M M M r -+
M O "i O Id! O d; O - O le O 0 le 0 O --i O t-
â,~., 10Q -ô Gôô r, ô m cc; Mô ~ - .i "0 ôô M
-H -H ' -H -H `° -H -H '-' -H I -H

M w Ca o r- GO
C45 q= 00 m OD CD
ti M
GO COO, ô GO
O, t~f,1 N .O C K
te cd 0 ci 14 M4 1-i 0 G .-
Ô
aT
t".
W m
(AalAi) ft-1 inums ;O ÀS.Iau 2[ sia3auimEd
 GAMMA RADIATION FROM NEUTRON CAPTURE BY HYDROGEN 405

tation value for this quantity are listed for each full-energy peak in the last two
columns of table 1.
The mean positions a2 of the full-energy peaks of all the gamma rays for all
the runs are presented in table 2. For each run, the energy E of the gamma ray
is related to the mean position by the quadratic function

E = u+ba2+ca22. (3)

The six reference energies thus provide an over-determined set of equations of

the form of eq. (3) for the calibration of an energy scale . The considerations
which led to a preference of the quadratic relation (3) over a linear relation are
discussed in refs. L $) . The resulting determinations of the energy of the hy-
drogen capture gamma ray are listed in table 3. The weighted average of this
set of data is 2217 .2±1 .5 keV for the hydrogen capture gamma ray. This gives
the value 2218.5±1.5 keV for the binding energy of the deuteron .
TABLE 3
Results of 10 determinations of the neutron-capture gamma ray in hydrogen, with assumption of
a quadratic relationship between energy and mean position

Position of mean
Run of capture Energy
gamma ray (keV)
(channel)

116 157 .840±0.031 2220.1 + 1 .5

117 160 .059±0 .033 2219.2±1 .5
122 161 .247±0 .039 2215 .3±1 .2
123 159 .951±0 .028 2215.6±1 .3
125 158 .089±0 .033 2215.8±1 .5
126 161 .371±0 .027 2215.0±2.4
147 153.839±0.032 2217.6±1 .6
148 150-734±0.031 2218.6±1.5
149 148-762±0.027 2217.1±1.5
150 146.356±0 .032 2217.8±1.5

Weighted average 2217.2±1.5 keV (error increased to allow for systematic deviation from
quadratic fit). Binding energy of the deuteron : 2219±2 keV. (This result contains an adjustment
for systematic errors, see text.)

In order to determine the effect, if any, that small-angle scattering of the

gamma rays in H2O would have on these measurements, the following experi-
ment was performed. A source of Cei" which decays through Pr'" to Nd144 was
placed in the centre of a barrel of water. The technique used for the measure-
ment of the hydrogen capture gamma ray was used to measure the energy of the
2.18-MeV gamma ray emitted by Ndi". We obtained the value 2184.1 ± 1.9 keV
for this measurement, which is to be compared. with the value 2186.0±2 .2
keV, the result of our previous direct measurement of this transition for a
400 J. E. MONAHAN 8: Q1.

include
source out of water 3) . The errors associated with these measurements
value for the effect
a measure of the systematic error in the method. To obtain a
and 0.4 keV for
of the water we consider the random errors which are 0.7 keV
the source out of water and in the water respectively . The effect of the water
is to give an apparently lower result by 1 .9±0.8 keV. We add this number to
our measured value of the energy of the hydrogen-capture gamma ray.
A measure of the systematic error in the method is reported 3) in the measure-
ment of the gamma rays from Ce144. Gamma rays with energies of 0 .696 MeV
.d 1.487 MeV are in cascade and the sum of the determination of their energies
a,-
is compared with the measurement of the 2.186-MeV crossover mentioned
above. This sum is measured to be 2183 .7 keV with a 0.5 keV random error.
The difference between these measurements is 2.3±0.8 keV. Since the x2 test of
the fit to the energy calibration curves indicate a systematic error in both
measurements, we conclude that the measurement of the energy of the cross-
over gamma ray is too high by 1 .2 keV with a systematic uncertainty of 1.1 keV.
When we combine the corrections for the effect of the water and the syste-
matic differences of the method, and when we combine the statistical error of
1.0 keV with the systematic error of 1 .1 keV, we obtain for our best value for
the binding energy of the deuteron 2219±2 keV.

3. Discussion of Results
As mentioned previously, the result of this measurement is lower than
the value usually quoted in the literature . In fact the discrepancy is sufficient
to indicate that one, or possibly both, of these values includes some unsuspected
systematic error. For this reason this section will include a discussion not
only of the present experiment but also of some independent measurements of
the binding energy of the deuteron and of related quantities.
3.1 . THE PRESENT MEASUREMENT

The technique discussed above is, of course, not an absolute measurement but
depends on a calibration of the pulse-height scale in terms of the independently
measured energies of a set of gamma transitions. In addition, the accuracy of
this method depends on the validity of using eq. (3) to represent the calibration
over the energy interval of this set of gamma ray energies.
It is possible that some of the energies of the standards used in this experiment
are in error, and, although there is some theoretical justification 1) for our choice
of eq. (3), it is also possible that this representation is not entirely adequate for
the energy calibration. However, since the use of six standard gamma-ray
energies provides an over-determined set for our assumed calibration function,
it is possible to obtain "best" (in the sense of least-squares) values for the para-
meters a, b, and c in eq. (3) and also to test the adequacy of eq. (3) by an
approximate calculation of x2. In the ten data runs the latter calculation re-
 GAMMA RADIATION FROM NEUTRON CAPTURE BY HYDROGEN 40 7

sulted in values of x2 which are from 1.1 to 3.2 times the expected value. Thus
eq. (3) seems to give no statistically significant deviation from the "correct"
calibration function .
Moreover, any random deviation introduced by the use of eq. (3) can be
included, at least approximately, in our final results by renormalizing our error
estimates so that xz takes on its expectation value. This has been done for each
data run ; and in combining the ten runs the errors have been treated as entirely
systematic and averaged in the same manner as the mean energy. The error
estimates, therefore, are considered to be relatively conservative.
There exists one possible flaw in the above argument . Because of the dis-
proportionate weight with which the 2.7533 MeV standard enters into the
calculations, particularly in the determination of the value of the parameter c
in eq. (3), an error of magntiude e in this standard energy will be reflected as an
error of about 2e in the measured value of the binding energy of the deuteron.
Thus the difference between the present result and that quoted by Wapstra 4)
would require that e be about 15 keV if this difference were to be attributed
entirely to this standard. This corresponds to slightly more than a full channel in
our pulse-height response . It is difficult to believe that any combination of
errors in the determination of the energy of this standard and in our measure-
ment of the corresponding value of a2 could account for this difference .
If subsequent measurements should indicate that an incorrect value of any
standard energy was used in this experiment, the results presented in tables 2
and 3, are complete enough to enable any interested party to recalculate the
value of the binding energy of the deuteron .

3 .2 . OTHER MEASUREMENTS

Measurements of the H3(p, n)He 3 threshold are among the experiments from
which precise evaluations of the binding energy of the deuteron have been
made . The Q value of this reaction, when combined with the end point of the
beta decay of tritium, gives the difference in mass between the neutron and the
hydrogen atom . This difference, when combined with the mass difference be-
tween H2 and D, gives the binding energy of the deuteron .
The threshold for neutron production by the reaction H3(p, n)He3 was
measured to be 1019± 1 .0 keV by Taschek et al . 7) . This threshold energy was
a
determined by calibrating the electrostatic analyzer against the Ale' (p, y)Si2
resonance at 993. 3 keV as measured by Herb et al. 8) . The calibration of
the energy scale was checked against the 873.5-keV resonance in F19 (p, ay)Ols
and the Li7(p, n)Be7 threshold") .
Bonner and Butler 9) have reported a value of 1020.3±1 .5 keV for the
magnetic analyzer
H3(p, n)He 3 threshold. These authors calibrated their
against the 873.5-keV resonance in F19 (p, ay)Ols and the 993 .3-keV resonance")
in A127 (p, y)Si28.
408 J. E. MONAHAN ßt al .

.A group at the Naval Research Laboratory 1 °) has measured the H3 (p, n)He$
threshold to be 1019.7±0 .5 keV. This group made an "absolute" measurement,
using an electrostatic analyzer to determine the proton ..energy.
If the Q values reported in these threshold measurements are combined with
the weighted average of 18.1±0.2 keV for the end-point energy of the beta-
decay of tritium as reported by King 11) and with a value 1441 .3±0 .6 keV for
the H2-D doublet difference, which is the average of two recent determina-
tions 12,13) the respective values obtained for the binding energy of the
deuteron are 2222 .3±1 .3 keV, 2223.3±1 .8 keV, and 2222 .7±0 .8 keV . These
results would be increased by 0.5 keV if the value of the beta-decay energy of
tritium (18.6±0.1 keV) obtained by Porter 14) is used.
This agreement is somewhat illusory. The Rice 9) and Los Alamos 7) groups
used the A127 (p, y)Si28 resonance at 993.3 keV to calibrate their accelerators.
In addition to the value of 993.3 keV for the energy of this resonance as
measured by Herb et al. 8), the following values are to be found in the literature :
990.8±0 .2 keV by Bumiller et al . 15 ), 992.4±0 .5 keV by Bondelid and Kenne-

dy
16)
and 994.0± 1 .0 keV by S. E. Hunt et al. 17) . An earlier bit of work in
Copenhagen 18) reports on a series of Al(p, y) resonances . Resonances at 986± 2
keV and 994+2 keV are reported. The 986-keV level of Brostrdm et al. is
indentified as the resonance at 991 keV by the authors of a survey article 19), but
these same reviewers report the 994 keV without adjustment of the energy
scale.
In our work on the F19 (p, ay)016 resonances 2°) in which we used the scintilla-
tion spectrometer with an anticoincidence annulus, we found that the relative
intensities of the gamma rays of energy 6.1 MeV, 6.9 MeV and 7.1 MeV change
as the proton energyis varied over the resonance . Also, at any particular proton
energy, the angular distribution differs measurably from one of these gamma
rays to another. Upon shifting to a plain NaI crystal and a pulse-height discri-
minator, it was possible to shift the position of the resonance at 1.385 MeV by
large amounts by varying the bias condition on the discriminator. It is suggest-
ed, therefore, that each of the measurements mentioned in the previous para-
graph may have been concerned with different quantities and that the spread in
the results represent a spread in the experimental conditions . In fact, it is
possible that the various experimental groups may have been looking at
different resonances. This calibration energy is, of course, quite critical in the
measurement of the H3(p, n)He3 threshold, and consequently in these deter-
minations of the binding energy of the deuteron. For example, the value
990.8 keV, if used for the A127(p, y) resonance, would lead to a value of about
2220 keV for the deuteron binding energy.
The report of the experiment at the Naval Research Laboratory 10) indicates
that their absolute measurement of the threshold of H3 (p, n)He3 would be in
disagreement with the earlier measurements 7,9) if their own value of the
 GAMMA RADIATION FROM NEUTRON CAPTURE BY HYDROGEN 40 9

A127 (p, y) resonance is) had been used. Their claim of agreement with the
earlier measurements, therefore, cannot be maintained.
Each of these measurements of the H3 (p, n) threshold used some system of
BF3 counters combined with a paraffin moderator as a neutron detector. In
order to determine the threshold energy it is necessary to extrapolate from
values of the measured neutron yield above threshold . Such an extrapolation
presupposes knowledge about the shape of the yield curve which depends not
only on the reaction itself but also on the response of the detector system (as a
function of energy). It is difficult to make a proper assessment of the uncer-
tainty in the results introduced by this procedure .
A direct measurement of the energy of the gamma rays emitted after capture
of neutrons by hydrogen was made by Bell and Elliott 21) . Their method in-
volves the use of a beta-ray spectrometer to analyze the photoelectrons ejected
from a uranium radiator by the capture gamma ray. A 13 cm thickness of paraf-
fin irradiated in the thermal column of a reactor was the source of the capture
rays. The value of the deuteron binding energy obtained by these authors is
2230± 7 keV. In the two runs usedfor this determination, the signal-to-noise ratio
is about 1 .08 and 1 .06, respectively. An accurate determination of the back-
ground is necessary for a precise measurement of the position of the peak since
this background is changing rapidly with energy in the region of the peak. The
background measurement is obtained by the substitution of lucite for the
uranium radiator. The contribution to the background by the continuum of
gamma rays producing photoelectrons from the uranium radiator will not be
detected by the lucite measurement. Similarly, the calibration standard, the
2.615-MeV gamma ray in ThC", is measured with the reactor off. Its peak
position is superposed upon a qualitatively different spectrum. Our own work
indicates that serious systematic errors may be introduced by poor signal-to-
background ratios and by inadequate methods for determining the background.
Direct measurements of the photodisintegration energy of the deuteron have
been made in at least two laboratories 22, 23) . Both these measurements are
threshold measurements and use a bremsstrahlungbeam produced by monoener-
getic electrons. The neutron yield is determined as the energy of the electrons is
varied and an extrapolation procedure is devised to determine the threshold of
the disintegration process. The values of the deuteron binding energy obtained
by this method are 2226±3 keV (Mobleyand Laubenstein 22 )) and 2227±3 keV
(Noyes et al . 23 ~
''
.

The extrapol ation procedure of Mobley and Laubenstein depends on the

assumption that the yield of photoneutrons is directly proportional to a power
of the energy excess above threshold. This was justified by a plausibility argu-
ment which is of questionable validity for the range of energies measured . The
combination of the behaviour of the photodisintegration cross section 24) and
the intensity distribution of the bremsstrahlung beam gives an energy depend-
410 J . E. NONAHAN et al.

ence for the photoneutron yield which differs quite markedly from a direct
power law for energies only a few keV above threshold. The energy of the elec-
tron beam is determined by reference to the Li7(p, n) Be? threshold with the
proton beam of the electrostatic accelerator.
As was noted by Marion 2s), the threshold measurement for Li7 (p, n)Be7 can
serve as an energy calibration point for charged particle accelerators but may
be somewhat unreliable as a measurement of the Q values of the reaction . It
must be emphasized, however, that the same reasoning demands that the
experimental conditions of the absolute measurement must be reproduced in
order to use these measurements for energy standards. The limitations associated
with the neutron counter, i .e. long counter 2s), and with threshold measurements
may cause systematic errors . Mobley and Laubenstein do not describe the coun-
ter, its shape, or position used in the calibration process.
The experiment by Noyes et al. consists of a similar measurement of the
photodisintegration of the deuteron by a bremsstrahlung beam. The energy
of the electron beam, however, is measured absolutely with an electrostatic
analyzer. The extrapolation technique is based on a different power law from
that of Mobley and Laubenstein, and again no quantitative estimate is made of
the error introduced by approximating the photoneutron yield by a pure power
law. There is a serious question about the detection of neutrons in this experi-
ment. A BF3 counter, at about 120° to 140° to the beam and a D20 target
is embedded in about 30 kg of paraffin . The quantity of paraffin is sufficiently
large to compete with the BF, counter, particularly for neutrons near threshold .
This competition probably introduces a prejudice against the data points near
the actual threshold.
All the results which have been discussed in this section seem to indicate a
rather close grouping of the values of the deuteron binding energy for those
measurements which are essentially similar. This would suggest the possibility
of systematic differences due to the method of measurement. In such a case it is
rather unrealistic to average all these measurements in the same manner as with
quantities exhibiting random differences.
A group at Livermore 27) has reported a direct measurement of the energy of
the capture gamma ray with a bent crystal spectrometer . Their mean value of
the binding energy of the deuteron is given as 2225 keV as computed from the
following set of measurements : 2224± 6 keV, 2224±16 keV, 22284- 3 keV and
2222-? keV. The data thus exhibit a large spread with the "best" values at the
extremes of the range. It is difficult to define a reasonable statistical
estimate of
the error to be assigned to the average of such a set . The Bragg angle
for this
energy is very small, so that the reflected beam is very close to
the main beam
direction. The background is known to be appreciable and
is changing very
rapidly. The report does not indicate how this background
is measured or what
corrections have been made for the background .
 GAMMA RADIATION FROM NEUTRON CAPTURE BY HYDROGEN

We gratefully acknowledge the Help received in setting up this experiment

from J. Peregrin of the Argonne National Laboratory. We are indebted to
R. Julke of the Applied Mathematics Division of Argonne for the actual pro-
gram on the digital computer. The patience and cooperation of L. Nelson and
J. Nelson and their staffs at the computer are appreciated. Finally we must
express our appreciation for the illuminating and encouraging discussions with
G. R. Ringo.

References
1) J. E. Monahan, S. Raboy and C. C. Trail, to be published
2) C. C. Trail and Sol Raboy, Rev. Sci. Instr. 30 (1959) 425
3) J. E. Monahan, S. Raboy and C. C. Trail, to be published
4) A. H. Wapstra, Physica 21 (1955) 367;
Encyclopedia of Physics (Springer-Verlag, Berlin, 1958), vol. 18, Part 1, p. 1
5) A. Hedgran and D. Lind, Ark. Fys. 5 (1952) 177 and private communication
6) D. E. Alburger, Phys. Rev. 92 (1953) 1257
7) R. F. Taschek, H. V. Argo, A. Hemmindinger and C - `. . Jarvis, Phys. Rev. 76 (1949) 325
8) R. G. Herb, S. C. Snowden and O. Sala, Phys. Re 6 75 (1949) 246
9) T. W. Bonner and J. W. Butler, Phys. Rev . 83 (, j51) 1091
10) R. O. Bondelid, J. W. Butler, C. A. Kennedy and A. del Callar, Phys. Rev. 120 (1960) 887
11) R. W. King, Revs. Mod. Phys . 26 (1954) 327
12) J . Mattauch and R. Bieri, Z. Naturforsch. 9A (1954) 303;
J. Mattauch and H. Hintenberger, Report on the activity of the Swiss study commission for
atomic energy from 1946 to 1958, Experimental Supplement VIII p. 50 (1960)
13) R. A. Dermirkhanov, T. I. Gutkin, V. V. Dosokhov and A. D. Rudenko, J . Nuclear Energy 3
(1956) 251
14) F. T. Porter, Phys. Rev. 115 (1959) 450
15) F. Bumiller, H. H. Staub and H. E. 'Weaver, Helv . Phys. Acta 29 (1956) 83
16) R. O. Bondelid and C. A. Kennedy, Phys. Rev. 1 :13 (1959) 1601
17) S. E. Hunt, R. A. Pope, D. V. Freck and W. W. Evans, Phys. Rev. 120 (1960) 1740
18) K. J. Broström, T. Huus and R. Tangen, Phys. Rev. 71 (1947) 661
19) P. M. Endt and C . M. Braams, Revs. Mod. Phys . 29 (1957) 683
20) J. E. Monahan, S. Raboy and C. C. Trail, to be published
21) R. E. Bell and L. G. Elliott, Phys . Rev. 74 (1948) 1552 : Phys. Rev. 79 (1950) 283
22) R. C. Mobley and R. A. Laubenstein, Phys. Rev. 80 (1950) 309
23) J. C . Noyes, J . E. Van Hoomissen, W., C . Miller and B. Waldman, Phys . Rev. 95 (1954) 396
24) H. A. Bethe, Elementary Nuclear Theory (Wiley, New York, 1947) p. 58
25) J. B. Marion, University of Maryland, Physics Department, Technical Report 184 (July- 1960)
(unpublished)
26) A. O. Hanson and J. L . McKibben, l'hys. Rev. 72 (1947) 673
27) E. L. Chupp, R. W. Jewell, Jr., and W. John, Phys. Rev. 121 (1961) 234