ir
,
.
."
.
.-
:
-
:-
.
473
ORNL
UNCLASSIFIED
NY
M
UM HIMOYA
::
..
!

-**..
.
*
ORNG-P-403
.
DIES
MASTER
CONF-641001-603
OCT disug
LEGAL NOTICI --

mic, d
,
d an
www
to me
naudingas
been
dhe
...
h
ow much I would not
2
EXPORDENAL STUDY OF ARTIFACTS AN ERRORS
COUNTERED. IN GAMMAPAUS ROG
FOR RADIOCHEMISTRY AND ACTIVATION ANALYSIS*
W. S. Lyon
J. S. Eldridge
P. Crowther**
Analytical Chemistry Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee
Research sponsored by the U. S. Atomic Energy Commission under contract
with the Union Carbide Corporation.
Present address: South African Atomic Energy Board, -Pelindaba, Pretoria,
South Africa.
The
o
ry
:
"7

1. Introduction
The use of l'al timma-ray spectrometry has enabled many radiochemical
determinations to be made nondestructively or with a minimum of chemical
A paration. Quantitative measuriment of gama-ray photopeaks coupled with
knowledge of the branching ratio, peak efficiency, and geometry enables the
absolute disintegration rate of a radionuclide to be obtained. Phenomena
that remove counts from photopeaks, such as cascade suming, obviously must
be taken into account in such measurements. Similarly, events or a combina-
tion of events, thet produce artifacts, spurious photopeaks, or errors in
measurements will affect interpretation or analysis of a radionuclide or a
radionuclide mixture. A number of such effects have been idertified and
studied, and results of these studies are given below.
2. Coincident or Cascade Garna-Ray Summing
This effect, which has been treated in an excellent article by Lazar [1],
occurs when two coincident gimna rays are detected simultaneously by the crys-
tal. It can be shown that the true disintegration rate of a radionuclide,
such as cool that decays by emission of two cascade gamma rays, 18 given by
PG)
where
bo Ex172)7(1-(72) AT
P(9.) • area under photopeak of 720
E (72) - peak efficiency of 77,
Enly) - total efficiency of yz,
of solid angle, and
Io - d/s of radionuclide.
At 10 cm. this correction 18 usually sma!.1, being about 2% for cobo.
It becomes increasingly larger as the source-detector distance becomes less,
3. 180° Compton-Annihilation Gamma-Ray Summing
When a radionuclide decays by positron emission, it is possible that
one of the 510-keV annihilation gamma rays will be counted in coincidence
with the 180° Compton back-scattered gamma ray from either the coincidence
-
..
510-k@V annihilation canna ray or another coincident gamma ray. This sum-
minc phenomena results in the production of an apparent photopeak of ~ 720
kev, whose magnitude varies inversely with the square of the solid angle
subtended by the source to the detector.
We have obtained gamma-ray spectra by use of a source of Cu°e under
several different experimental conditions that illustrate this effect.
Measurements have been made at high and low geometries both shielded and
encapsulated. These data illustrato especially the enhancemeat of the same
ming effect at high gocmetry with a large amount of scattering material
present.
In 78. 1, there is only a slight indication of the ~ 700-keV photo-
peak; here the positrons are annihilated not at the source, but either in
the detector or in the walls of the shield. When the source is encapoulated
in wax, the result 1.8 annihilation and scattering close to the source or the
positrons emitted in all directions. Thus, the conditions are maximized for
the production of the sum peak as shown in Fig. 2. At the greater distance
of 8.6 cm, there is a small amount of sum peak production even with no encap
sulation (F18. 3). This is explained by the greater probability of scato
tered radiation reaching the detector from the sides and walls of the 30-cm
diameter load shield. The effect here is small, however. But when the
source 18 encapsulated (Fig. 4), the sum poak again appears in greater inten-
sity as explained above.
The distribution of Compton electrons from scattering results in a
shary maximum corresponding to 180° scattering. The 180° Scattering peak
18 clearly seen at w channel 19 in Mo. 3 and 4. In the high geometry
experiments shown in Figs. 1 and 2, the analyzer discriminator was set to
cut oft about channel 16 so as to increase the live time; hence, this peak
is not as obvious us in Figs. 3 and 4. For the energy region 400 keV to
2 MeV, the energy corresponding to this 180° scattering slowly rises from
about 160 keV to about 200 keV as Illustrated graphically in reference [27.
Thus, one vould expect to find the sum peal from the 510-keV annihilation
gamma ray and 180° scattered photon at 680 kev. The higher value reported
1s duo either to a slight error in energy-pulse height calibration or more
..
.
WWW.WI
TT? -- VIPi
.. -
*** ***
*** ....... en Por
de tenir un predvsem
probably to the wel.l. known nonproportionality of Nal detectors for sum peaks.
The difference between the expected enerxy of 180 keV and the observed onervy
1. about 30 keV. This 18 the same discrepancy t'ound by Peelle and Love 37
in their study of sum garona ruyu.
Probably the safest, most reliable, and most accurate method of coun-
tinc a sample is to position the source at a fairly low geometry, 1.6., at
leust 3 cm and preferably 10 cm from source to detector. Whore sample and
standard are · be compared, the amount of absorber close to the sources
~ 700 keV does not prevent accurate measurement of pure samplee, its presence
maliy obscure an impurity or cause uncertainty in the mind of the experimenter.
For absolute measurements, 1.0., without the use of compas'ator standards,
all sumaning effects should be minimized and must be corrected for in the
calculat-ons. The more familiar suming phenomena, y-y cascade suming and
8*-summing, (as discussed above in section 3) are reduced at 10 cm to a
maximm of only a few percent of the dinintegration rate.
4. Gamma-Ray Spectrometry_of8* Emitters
In the gamma-ray spectrometry of positron emitters, the position of
annihilation relative to the detector is of great importance. The height
(and area) of the 0.51-MeV gamma-ray photopeak obtained when a source has the
absorber close to it varies cons..derably from that obtained when the absorber
18 far away. As is well know, the positron annihilation results in the pro-
duction of two 0.51-MeV gamma-ray quanta per positron, and these two quanta
are emitted in directions 180° from each other. It is thus necessary to
divide the total integrated area of a positron annihilation gamma-ray photo-
peak by two to obtain the actual number of positrons triat it represents.
Using sources of Na“, we have studied the effect of absorber position on
the number of annihilation quanta detected in the crystal. The disintegra-
tion rate of the Na22 was obtained from the integrated area of the 1.28-MeV
gamma-ray photopes.k by absolute gamma ray spectrometry. A source of Na<
was dried on polystyrene and measured on a 3" x 3" NaI(Tl) ganma-ray spectro-
Ti
-
MeV gamma-ray photopeak wus massured and the apparent disintegration rate"
-
calculated:
where
? . fraction of decay by position omission,
Mo - solid angle, and
En intrinsic peak efficiency.
Table I lists the results obtained under a variety of absorber conditions.
From Table I a number of Interesting observations can be made. Frst, 1t
1s obvious that the correct value for the disintegrotion rate is obtained
only when the gamma rays are anninilated close to the source, as evidenced
by the first entry. This arrangement was a beryllium sandwich. With only
one absorber piaced diractly below the sample, the observed disintegration
rate is about 20% lower than the true value.
With no absorber (annihilation in aluminum cover of detector and in
detector) or with absorber on the crystal, the observed disintegration rate
18 seen to be 15-18% too bighiPlacing absorber behind the sample, either
with or without additional absorber on the crystal, results in a disintegra-
tion rate ~' 40% above the true value and ~ 20% above the value obtained
with no absorber directly below the crystal. This agrees with the 20% lä-
crease obtained when the source placed on the absorber (#2 in Table I) was
backed by beryll1um (#1 in Table I). These effects are essentially indepe.:d-
ent of the source-detector distance.
5. Production of Ann:lhilation Radiation by Interaction
of High Ener Carma Rays With Walls
There are many instances in which a positron end tter must be measured
in the presence of a large amount of a radionuclide with one or more high
energy gamma rays. If this situation arises, it is almost always more advan-
tageous to measure an accompanying gamma ray (17 one is present in the decay),
rather than the 0.51-MeV annihilation radiation peak. This is because the
high energy gamma rays can interact with the sides and top of the shield to
produce amihilation radiation, which results in the appearance of a photo-
peak at 0.51 MeV ia the observed gamma-ray spectrum.- such a photopeuk
Amma
1
causes indirect interference in the determination of a true positron emitter,
and in many instances presents so large an error that the desired determina-
tion cannot be made. For example, the determination by activation analysis
of trace quantities of copper in the presence of large amounts of sodium 18
seriously affected by this effect. The production of annihilation radiation
by gamma rays increases with increasing gamma-ray energy, and the 15-bour Nac
produced by activation has a 2.8-MeV gamma ray in 100% abundance. The 12.9-
hour cu º produced by neutron activation decays only 19% by positron emission
and so 18 particularly susceptible in annihilation gamma-ray interference.
We have made some studies in our laboratories to establ.ish the magni-
tude of this errect in our standard shields at various geometries for gamma
rays of several energies. Our pickle-barrel shield 18 a right circular
cylinder having inside demensions 50-cm uigh, 30-cm diameter. The top of.
the photowltiplier-crystal assembly 18 22 cm from the top of the shield.
Data were taken at heights 5.4 cm, 9.3 cm, 20 cm (top open), 20 cm (top
closed), and 36 cm (sample above the top of the shielå) from the detector.
Gamma-ray spectra from sources of four different radionuclides were obtained
at these distances, and the total integrated area beneath the 0.51-MeV gamma-
rey photopeak and the high-energy gama-ray photopeak corrected for peak
efficiency and geometry /17. The four radionuclides used and their gamma-
ray energies are given in Table II.
Figure 5 shows the gamma-ray spectrum obtained from g3 at 36 cm. The
two small photopeaks at ~ 1.3 MeV and ~ 1.8 MeV arise from small amounts of
1.8-hour Art and 2.27-minute A14%, respectively, which were present as
slight impurities in the sample. Note the large contribution from the anni-
hilation garwa ray. Figurs 6 compares the 0.51-MeV and 3.1-MeV gamma-ray
photopaak regions from experiments made at two different sources to detector
distances: 36 cm and 5.1 cm. Note the decrease in the ratio 0.51-Mev/3.1.
MeV photopeak as the source is brought closer. Because the geometry for
annihilation gama production and detection decreases as the source 18
brought closer. to the detector, it is apparent that high geometries tend to
minimize this effect. Ia 718. 7 18 plotted the results of these measure-
ments. As expected, the higher the gamma-ray energy and the greater the
distance from source to detector, the larger the effect. Note also the
increase in the ratio 0.51/principle gamma ray at 20cmwhen the lead
1
2
3
vir
10 PA
MALL
Ho
cover is placed behind the source. Some additional experiments were made
with sources of Na24 in a large (81 cm x 81 cm x 81 cm) shield at different
hesabte. Table III gives a comparison of data obtained in the small and
large shield.
The advantage of using a large shield when large source-detector dis-
tances are required is obvious. In addition, M8, 7 and Table III Indicate
the difficulty of determining stall amounts of cuºe in the presence of NaC.
8ince Cu°* decays only 19% by 8* emission, there will be only 38% 0.51-MeV
Earma rays aval able for photopoak production por disintegration. This means
that usch gran of scdium exposed to irradiation and present at analysis time
will produce 0.51-MeV gamma photopeak intensity equivalent to 12 mg of cuo
when the sample io counted at 10 cm.
t
6. Resolution and Shape of Low Energy Cama-Ray Photopeak
Iodine-125 is an electron-capturing nuclide which has a highly con-
verted 0.035-MeV ganilia ray. Thus, the gamma-ray spectrum consists essen-
tially of only TeK X-rays. Because the X X-rays from capture are in coinci-
dence with the K X-rays from the converted gamma rays at high geometries,
a cascade sum peak 18 observed at twice the X-ray energy. A method has been
devised to assay rks by this sum coincidence method 247; in this method the
areas of the singles photopeak and the sum photopeak are used. Figure 8
shows an excellent gamma-ray spectrum of phas obtained with the source on
the detector face. Figure 9 compares the spectra obtained by use of two
different 3" x 3" Nal detectors. The non-symetrical shape of the K X-ray
peak in the lett spectruim was a cause of some concern, since both detectors
were apparently identical. If one were to use the left detector for a' meas-
urement of 1", bo vould not be able to determine the disintegration rate
by the sum coincidence method. from the appearance of this spectrum, we
might conclude that the X X-ray peak was conteminated with a X-ray of same.
what lover wergy.
We examined resources on a variety of detector assemblies, include
ing both well and solid counters, and observed excellent resolution of the
K X-ray peak in only a few. A surprising finding was that the shape of the
X X-ray photopeak was 1886 symetrical when neajured with detector assemblies
having good resolution for CS gamma rays than when measured with what we
:........................
...........
..
usually describe as a poor roboliation detector. A NaI crystal with poor 125
measurement characteristics but sood C8-31 resolution (7.8%) was succes-
sively polished with grit papers. Pigure 10 shows the X X-ray spectra of
I obtained after each treatment, together with the Css resolution.
The crystal surface was first polisbod with 180-grit emery paper; this
Improved the Cotsresolution tut worsened the X-ray response. Polishing
with 300-grit paper resulted in bitter X-ray resolution, but poor cssr
response. Additional polishings with 600 grit, and finally 2000 steel wool,“
gave very good low energy resolution and brought the C8-37 value vo 7.90%,
close to the original value of 7.8%. It 18 apparent that crystal prepara-
tion by hand polishing can be a prime cause of poor K X-ray resolution for
125.
From these studies we conclude that spectral distortiou8 for low-energy
responses in many detectors may be attributed to optical effects caused by
light scattering at the surface where most low-energy interactions occur.
These effects could possibly be confused with electronic noise distortions.
Witia proper care, however, suitable low-energy spectral shapes as well as
high resolution for C8+5gamma rays may be achieved.
7. Conclusions
We bave described a number of possible errors or artifacts encountered
in gamaa-ray spectrometry. The radiochemist must be ever alert to the possi-
bility of misinterpretation or error. Simultaneous measurement of standards
is always helpful but not always possible. Where mixtures of radionuclides
are to be assayed, especially positron emitters, extreme care must be taken
in the measurement and interpretation of gamma-ray spectral data.
8. Bibliographical References
LI] LAZAR, N. 8., I.R.E. Trans. Nucl. Sci. NS-5, #3 (1958) 138/146.
27 JOHNSON, N. R., BICHLER, E. and O'KELIXY, G. D., Nuclear Chemistry.
Techniques of Inorganic Chemistry, vol. II, Interscience, New York
(1963) 41.
[3] PEELE, R. W. and LOVE, T. A., Rev. Sci. "Instr. 31 (1960) 31.
17 HARPER, P. V., Y AL., J. Nuclear Med. 4 (1963) 277.
..
.
PT
LE
*
::.1
WT.
11
TABLE I
Na?? MEASUREMEXIT BY 8+,COUNTING
Calculated Na22
Absorber
Calculated/Trus
1.0
2. 1:23 Be Dehired uple.
1.23 Be below spla.
4.36 x 204
0.80
1.18
2. 1.23 Be below sple.
3. None
4. 1.23 mg/cm2 Be on xital
5. 1.23 mg/cm2 Be behind
Be on %'tal
1.23 mc/cm2 Be behind sple.
3.49 x 204
5.16' x 104
5.04 x 204
6.19 x 204
6.07 : 204
1.15
1.42
1.39
TABLE II
RADIONUCLIDES MEASURED IN ANNIHILATION GAMMA-RAY PRODUCTION
Nuclide
I, 12
Game-Ray Hinergy
(Mey)
**
o
12.4
37.3 m
15.02
5.1 m
1.5
2.15
2.75
3.1
TABLE III
RATTO 0.51-MeV 7/2.8-MeV in POR Na24 IN TWO SHIELDS
Source-Detector
Distance (cm)
Cylindrical Shield
(n = 50 cm; d. 30 cm)
Cubic Shield
(81 cm x 81 cm x 81 cm)
.5.3
0.012
20 (top open)
20 (top closed)
0.012
0.029
0.069
0.10
0.11
0.027
0.051
0.065
0.077
-10
ORNL-DWG. 63-3636
Fig. 1. Gamma-ray spectrum of cuºt higio geometry, no absorber on source.
UNCLASSIFIED
ORNL-DWG. 63-3636
100,000

Cu64 at 1.74 cm
3 in. x 3 in. NaI (TI)
No absorber
Energy scale = 10 KEV/PHU7
10,000
LLULU
COUNT RATE
1000
100 €
TTT
10
Liteit
. PULSE HEIGHT, arbitrary units
50
100
150
***;4...vernow.in...
.
is.
ORNL-DWG. 63-3637
Fig. 2. Gamma-ray spectrum of a
surrounding source.
high geometry, absorber
UNCLASSIFIED
ORNL-DWG. 63-3637
100,000

Cu 64 at 1.74 cm
3 in. x 3 in. NOI (TI)
0.5 in. wax absorber
Energy scale i 10 KeV/PHU
:10,000
NT RATE
-
-
-
-
-
50
100
PULSE HEIGHT, arbitrary units
ORNL-DWG. 63-3635
Fig. 3. Gamma-ray spectrum of cuº* low geometry, no absorber.
.

oor.......................
UNCLASSIFIED
ORNL-DWG. 63-3635

Cu64 at 8.6 cm
3 in. x 3 in. NOI (TI)
No absorber
Energy scale 2 10 KEV/PHUT
1000
100
.
150
50
100...
PULSE HEIGHT, arbitrary units
.-.-
.-.
-.-
.-.-
.
-.
-
.-
-.
-.
..
...
...
..-
-.-
ORAL-DWG. 63-3634
Fig. 4. Gamma-ray spectrum of Cuº* low geometry, absorber surrounding
source.
T
.
i
..
::
.
2
UNCLASSIFIED
'.
.
*
100,000

il "L
.
4.Y.
1:
4
Y
VY
Cu64 at 8.6 cm
3 in. x 3 in. NaI (TI)
0.7 in. lucite absorber
Energy scale i 10 KeV/PHU
1.
;
:
:
YAK
ART
XL-
IX
10,000
.
.
COUNT RATE
1000
NII
MT
10
.
OGb.
50
. 100.-
PULSE HEIGHT, arbitrary units
ORNL-DWG. 64-7799
Fig. 5. Ganma-ray spectrum of 837 at 36 cm.
UNCLASSIFIED
ORNL-DWG. 64-7799

:
: :
10,000
3.1 MeV
.
.
.. ..
1000
TRATE
COUNT
الدليل البايتالايام
0
20
40 60 80 100 120 140 160 180 200 .
PULSE HEIGHT (~20 keV / PHU)
ORNL-DWG. 64-7797
Fig. 6. Comparison of 0.51-MeV and 3.1-MeV gamma-ray photopeak
regions obtained at 5.4 cm and 36 cm.
UNCLASSIFIED
ORNL-DWG. 64-7797


537
5.4 cm
337
36 cm
10,000
-
-
-
-
COUNT RATE
TT
1000
2.-
aber
rari**'--'an Sendi
1001 I dhu III
10 20 30 40 50 140 150 160 170 180 10 20 30 40 50
PULSE HEIGHT (~20 keV/ PHU)
140 150 160 170
..
.
.
-
ORNL-DWG. 64-7798
Ms. 7. Photon ratio of high-energy produced gamma rays to
producing gamma ray m hunction of energy.

1
LAR
U
..
11.
14
*
W
WW
?
UNCLASSIFIED
ORNL-DWG. 64-7798
0.14

.
o 36 cm
A 20 cm closed
o 20 cm open
09.3 cm
5.4 cm
0.10+
NAO.08+
0.064
: 0.51 y
PRIN. Y
0.04
0.025
1.5 2.0 2.5 3.0
ENERGY OF PRINCIPAL Y-RAY (MeV)
IMAM HUJILL
WAY.
.4.
ve
l
.
L,'.
.
.
ORNL-DWO. 643021
Fig. 8. Gamma-ray spectrum of 60-day Is, showing area contribution
to singles peak (4) and coincidence peak (A).
.
-
--
-
-
-
-

i
10%
ORNL-EWEES

60 day 1128
3x3 in. Noi (TI)
Source on Detector Face
Energy Scale, 1 kov/ch1
[Az+ 2A, 12
No.
4 A
.
COUNTS / CHANNEL
:
::::::
:
..:.
.Biri...
40.
60
CHANNEL NUMBER
.

.
/
1
inicios
ORNI-DWO. 64-3017
Mg. 9. Gamma-ray spectrum of 60-day 1" with (a) detector showing
lov-energy distortion due to optical effects and (b) detector
sboring correct response.
il
14
.
.
UNCLASSIFIED
ORNL-OWG, 64-3017


.
.
60 day 1125
3x3 in. NaI(TI)
Source on Detector Face
Energy Scale, 1 kev/ch
.
.
WUNTSX CHANNEL
.:
:.:
.:
.
60 80 0 20
CHANNEL NUMBER
1
-
-
.
ili
1
.
i.
-
-
ORNL-DWG. 64-3019
Mg. 20. Gamma-ray spectrum of 60-day I for different treatment
of the incident crystal face,
(a) Bad face polish with 180 grdt
(6) Ind face polish with 300 grit
(c) Bad face polish with 600 grat
(a) Dad face polish with 0000 steel wool

,
?
.
7
-
*
*
.
11
VIS
RRU
AMAT!
IM
UNCLASSIFIED
ORNL-OWO. 64.3019
60 Day 1126
· 3x3 in. NoI(TI).
Energy Scale, 1 kev/ch

.:
FEnd Face Polish,
180 grit
| End Face Polish,
300 grit
mmmm
TTIR
.662 MeV
Cs 137
.662 MeV Cs 137
Resolution, 7.3%
Resolution, 7.6%
:.
..
COUNTS/CHANNEL
FEnd Face Polish,
600 grit
End Face Polish,
0000 steel woola
sa...
...
:
.662 MeV
.662 MeV Cs 137
Cs137
Resolution, 7.8% Resolution, 7.95%
20., 40
20 40
CHANNEL NUMBER:
:
::..::,:'.
"
..
.
DATE FILMED





2/4 165




LEGAL NOTICE
This roport was prepared as an accovat ol Qovorament sponsored work. Neither the United
Stator, nor the Commission, nor any person soting on behall of the Commission:
A. Makos any warranty or roprosentation, expressed or implied, with respect to the acou-
rioy, completeness, or wohinose of the information contained in this report, or that the use
of any inlormation, apparatus, mothode or process disoloned in this roport may not infringe
privately owned rigato; or
B. Assumor any liabilities with rospoot to the use of, or lor damages resulting from the
un of way Information, apparatus, method, or procesu disclosed in this report.
As used in the above, "person uting on behall of the Commission" inoludes any om-
ployee or contractor of the Commission, or omploys of such contraotor, to the extent that
such omploys or contractor of the Commission, or employs of such contractor preparos,
disuminatos, or provides access to any information pursuant to his employment or contract
with the Commission, or his employmout with such contractor.

END