1
...
i
;
• I OF L
ORNLP
2432
r .LOI
.
I
,
:
.
.
!
1
.
IEETEEEE
IV
111.25 .1.4 1.1.6
.
.
MICROCOPY RESOLUTION TEST CHART
INATIONAL BUREAU OF STANDARDS -1963
i
$
.ORNL-P- 2432
Conf-66/012-1TIAC
SEP 2 2 1963

LEGAL NOTICE
RELEASED FOR ANNOUNCEMENT
IN NUCLEAR SCIENCE ABSTRACTS
TWI roport nus propered u uwcoounal al Goronumant spruworod wrt. Neither the Uniad
stat, nor the countesson, nor wey porsas acting na behalf of the countsator:
A. Makos vynarraty or represeаtation, exprosnad or implied, we respect to the accu-
moy, complacemeus, or ustalonss of the talormation coataimed la this report, or that the wo
o! way information, apparatus, matbod, or procu, disclosed in this ropar Buy not talriamo
primusly owned niets; or
B. Asmuss my lasines with respect to the use of, or for damages resulting brown the
: um d w labormation, apparatus, method, or proces disclosa in the report.
As wood to the above, porrow athag ao beball of the Comodostaa" tacirdas any on-
ployw or contracter of the Cor mis son, or employee of such contractor, to the extent that
med vaploymo or contractor of Blas Commission, or employee of mcb coatraclor proparus,
denomina in, or provides accous to, any lalormation pursuant to his omploymeat or coatract
with the Commission, or we rmploy.nant with such contactor.

SECONDARY STANDARDIZATIONS USING DIG:TAL COMPUTER TECHNIQUES
TO NORMALIZE MULTIPLE GAMMA-RAY SPECT:{A FOR DIRECT EFFICIENCY DETERMINATIONS*
1
CESTI PRICES
J. S. Eldridge
Analytical Chemistry Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee
HC. 17.001 and
1. INTRODUCTION
Gamma-ray scintillation spectrometer systems have been used for quanti-
tative determinations of radioactive source strengths for many years. HEATH (1,2)
and LAZAR 13) gave detailed instructions for the application or the spectro-
metric technique for the quantitative determination of photon intensities. Both
authors describe systems using 3 in. x 3ja. TaI(11) detectors and 9-10 cm
source-to-detector distances. Suitable efficiency factors are given in those
reports so that gamma-emission rate determinations may be made with accuracies
of about 3% for many nuclides. The success of the scintillation spectrometric
method is based on the fact that once the efficiency of a crystal has been
obtained, another crystal of the saine size will have exactly the same effi-
ciency for a source at the same distance and direction. The pulse heights may
differ between the two crystals and the detectors may give different resolution,
but the counts under the photopeak in a gamma-spectrum are independent of
pulse height or resolution.
To provide a collection of gamma-ray spectra suitable for direct quanti-
tative determinations in a multichannel analyzer system, multiple sources of
absolutely assayed nuclides were combined to give a "library" of spectra from
which direct, assays could be performed. Multiple sources were combined so as
to increase the overall accuracy of the collection. The library was collected
for use with a least-squares gamma-Spectral resolution program; an assay system
was developed based on the digital data in the form of punched paper tape.
2. COLLECTION OF LIBRARY DATA
'Absolute disintegration rates were determined for a master solution of each
library nuclide. Disintegration rates were determined by 41B-coincidence
counting (4), ionization chamber measurements (5), and absolute gamma, spectro-
metry [2,3); the primary method was 47B-y coincidence counting where applicable.
"Research sponsored by the U. S. Atomic Energy Commission under contract
with the Union Carbide Corporation.
Triplicate sample mounts were prepared from each master solution by evaporat:ing
a precise aliquot onto l in. diameter watch glasses. These watch glasses were
then mounted in the center of a 2-1/2 x 3-1/4 inch cardboard source holder.
Each of the three sample mounts was counted twice at a distance of 10 cm
from the face of a 3 in, x 3 in. NaI(T1) detecvor (Harshaw Integral Line Scin-
tillation Detector, Model 12512) housei in a idad detector shield having
internal dimensions of 14 x 14 x 22 inches cov:red with a graded' liner of .
cadmium-copper. The detector face was covered with an 0.73 g/cm2 polystyrene
diac which served as a bete part 1020 quárbar, the detector resolution was 7.6%.
Background determinations were made immediately before and after sample
counting. Therefore, the output from the multichannel analyzer (Radiation
Courter Laboratories Model 20624 equipped with a punched-paper tape readout
device) consisted of two background spectra and six sample spectra from the
master solution for each librazy nuclide. Random coincidence effects were
minimized by keeping the deadt ime of each measurement below 10%.
3. POOLING PROGRAM FOR LIBRARY STANDARDS
A program was developed for the IBM 7090 computer to combine the six
standard spectra, correct for background, decay, and gain shift and convert
the spectral data from counts per second per channel to counts per disintegra-
tion per channel using the disintegration rate of the standard sample as an
input parameter. The computer program períorms an analysis of variance for
each channel for replicate counts within replicate aliquots by comparing swas
of squares with appropriate chi-squared and F-test values. A list of channels
whether undue variance was present in any of the samples. The statistically
tested spectral data inake up the standard library which is stored on punched
cards or magnetic tape at the computer center.
A typical computer output sheet showing the library data for the 59Fe
standard is shown in Fig. 1. Lines 1 and 2 of the output sheet show sample
identification and input data. Response 13 indicates the library number;
analyzer number 1 with 255 channe is was used with a low channel limit (LOCN)
of 2 and an upper channel limit (HCN) of 1.60 to give a standard scale factor
(SIDSF) of 10.00 keV/channel. A low and high channel limit may be used in
those cases where there is no valid spectral data for the standard in some
regins of the multichannel analyzer output. This serves as a time-saving
fecure in the computer program. Line 2 gives :urther input data: the nuclide
is 59Fe, and the library was made from solution number 51. Also shown is the
standard error associated with the absolute assay of the standard. The decay
constant, 0.178 x 106, has units of reciprocal seconds corresponding to a
half-life of 45 days, while the symbolism ior the beta absorber indicates that
an 0.736 g/cm polystyrene absorber was used. The final number, 1290.0, is
the energy in keV of the prominent feature of the spectrum that is used for
gain-shift corrections.
In section 3 the numbers 98-103 identify each of the six spectra making
up the pool. The numbers 0.2780E 055 rezer to tlie disintegration rate of each
standard (0.278 x 105 dis/sec) while the number:; following the disintegration
rate, such as 10.036, 10.033 etc., indicate the measureå values in keV/channel
for the scale factor of each spectrum.
Section 4 is a channel-oy-chenre: readout oi the corrected data, comprising
the library for the 59Fe standard. Ise vaiäe listeà for channel 1, 4937-5,
rafers to 0.4937 x 1095 counts/dis. . searching routine in the program reduced
the upper channel to 157; that is, the spectrur showed no valid data above
channel 157.
To recall the library data in a form suitable for insertion into a multi-
channel analyzer memory, it is necessary to nornalize the data in such a way
as to have whole numbers in each channel instese of the small fractions that
are given in section 4 of Fig. 1. Inis is accc.nplished in the computer by
converting the integral counts from the spectral data (channels 2-157 for the
59Fe example) to a value of 100,000 counts. Surmation of channels 2-157
results in 0.1914 x 10°? counts/dis for the 5952 standard. The normalization
factor (F) is ther found by dividing 100,000 by this value - giving in the
59Fe case 5.224 x 106. Fig. 2 shows tre reculled spectrum of the 59Fe library
standard where each channel plotted is the corresponding value from section 4
in Fig. 1 multiplied by 5.224 x 106. The normalization factor is calculated
in the .recall routine of the library standard program.
Forty-two nuclides have been measured in the standard library collection
and have been recalled for use in graphical as well as punched-paper tape forms.
Table I gives a list of the nuclides collected in the program.
4. USE OF THE LIBRARY STANDARDS
The collection of response functions for the 3 in. x 3 in. NaI(T2) spectro-
meter has been found to proviàe a convenient means of unscrambling unknown
spectra for on-tne-spot identifications and determinations. Punched paper
tape spectra corresponding to the constituents of the unknown sample may be
successively "stripped" or "peeled" using a spectrum-stripping technique [6].
Comparison of the relative photopeak areas of the standards and unknown con-
stituents can provide absolute disintegration rates.
Because the standard library consists of channel-by-channel data in counts
per disintegration, the library comprises a chanzel-by-channel measure of the
detection efficiency of specific nuclides for the given spectrometer shield
and source-to-detector configuration. Thus by combining the channel-by-channel
efficiencies for some region of the garma-ray spectrum which is not affected
greatly by external processes (for example, a prominent photopeak) it is possible
to directly assign an efficiency factor for measuring each of the library
nuclides.
The absolute emission rate determinations by HEATH (2) and LAZAR [3] .
require an expression of the form:
dis/sec = (counts/sec)
(1)
ErP AG,
where the subscript P refers to the photopeak corresponding to a specified
gamma ray, G is the branching ratio (emitted photons per disintegration) for
that gamma ray, Et is the calculated total efficiency, P is measured peak-
to-total ratio, and A is a correction factor for attenuation of the gamma
ray in the beta absorber. Thus the use of those techniques requires a knowledge
of four tabulated quantities.
.
......
...
..
..
.
.
.
.
.
.
.
...
...
...
..
...
.
.
le
M
M
The library of absolutely assayed sources provides a convenient method of
determining the disintegration rate o: an unen sampia by providing an overall
efficiency for any portion of the gamma-ray cirum and permits the use of
combined photopeaks for increased statistics occüracy. The expression for
determining absolute disintegration rates with the standard library may be
given as:
dis/sec = (counts/se :)
(2)
where (counts/sec)p refers to the integrated counts for some region of the
unknown sample's spectrum (usually a protopeak or combination' of close-Lying
photopeaks), F is the normalizing factor for tre standard (see section 3),
and Cp is the integrated counts from the library spectrum for the same energy
region as chosen for, the unknown sample.
An example will help to illustrate the use of expression (2) for an abso-
lute disintegration-rate determination. In Fig. 2, showing the 59Fe library
spectrum with F = 5.224 x 106, the region from channel 102 to 140 was chosen
for the comparison because it encompasses two photopeaks, and the extra counts
in the region increase the statistical accuracy. The shaded area (102-140)
was integrated with a summation device on the ultichannel analyzer and found
to contain 3.478 x 10° counts. A 59Fe sampie containing 2.05 x 10* dis/sec
was analyzed with the spectrometer where the integrated counts for channels
102-140 were found to be 1.346 x 104 or a 100 sec counting period. Thus the
integrated counts per second were 1.346 x 102. Substitution of these quantities
into Eq. 2 yields:
dis/sec = (1.346x102) (5.224x70e
(7.346x102) (5.224x106)
(3.478x104))
=
2.02x104
(3)
This value is ilì excellent agreement with the value of 2.05 x 104 dis/sec
determined with an independent assay.
Four sources from set number 116 of IAEA calibrated gamma-ray standards
(January 1, 1965) were measured with the punched tape assay system on July 24,
1966. Results from the assay were corrected for decay to the assay date of
the IAEA calibration. A comparison of the measured values with the IAEA cali-
brations is given in Table II. Half-life values used for the corrections were
those given with the calibration data. In the case of the 6ºCo assay, the
energy range of 1.09 to 1.42 MeV was chosen for the comparison so as to include
both photopeaks. For the other sources, the area encompassing the single
photopeak was used.
A further test of the assay method was macle to check its validity in
other spectrometer systems. Five radionuclide mounts were deasured in four
separate spectrometers at Oak Ridge National Laboratory (ORNL). Results of
those tests are shown in Table III. Systems B, C, and D contained Harshaw
Model 12812 NaI(TI) detectors, while system A contained a 3 in. x 3a. NaI(T1)
detector whic: had been canned at ORNL. The results show an overall low bias
of ~3-5% for spectrometers F, C, and D. Becau::e the apparent reason for the
discrepancy is due to an inaccuracy in determining the exact source-to-detector
distance, the bias can be corrected by using a single radionuclide standard
in conjunction with the tape assay etiroa. use of a single 1870s source
with system B for normalizing the icteruiol : iciency resulted in values of
1.00, 0.986, and 1.005 for the ratio or a udio vivier to amourt found for
198Au, 59Fe, aná 5+Cr respectively. Similar i grovements were found for
spectrometers C and D.
Resolution of the detector is not a critical factor in using the tape
assay method as long as the area cliosen for the integration and comparison
encompasses the entire photopeak.
5. LEAST-SQUARES GAMMA-RAY SPECTRAL RESOLUTIO!!
ELDRIDGE and BROOKS (7) deve.loped a computer program to use the library
of standards in a gamma-ray spectral resolution, technique. The program is
based on the method of least-squares to determine radionuclide concentrations
in complex mixtures. The program was tested with a quaternary mixture made
from known quantities of 465c, 5tcr, 595e, aná 134Cs. Figure 3 shows the
gamma-ray spectrum of the quaternary mixture superimposed on the individual
nuclide spectra at the level determineri by the program. Satisfactory resolu-
tion of the mixture was achieved by the method as seen by the results presented
in Table IV. It should be emphasized that the standard library of nuclides
developed in this program was used as the basis of the least-squares fitting.
6. CONCLUSIONS
A library of statistically tested gamma-ray spectra has been collected
for absolutely assayed sources of 42 raciionuclides. Techniques for use of the
library data in a direct assay method well as a computer system of gamma-
ray spectral resolution has been developed. It has been shown that secondary
standardizations of radionuclides may be performed with this library with
accuracies of 1-5%. A single nuclide standard used in conjunction with the
library data permits a rapid determination of radioactive source strength
in many spectrometer systems.
RIFERENCES
(1) HEATH, R. L., Scintillation spectrometry, gamma-ray spectrum catalogue,
USAEC Rep. IDO-16408 (1957).
[2]
HEATH, R. L., Scintillation spectrometry, gamma-ray spectrum catalogus,
USAEC Rep. IDO-16880-1 (1964).
[3] LAZAR, N. H., I.R.E. Trans. ?:cl. Sci. NS-5, No. 3 (1958) 138-46.
[4] CAMPION, P. J., Int. J. Appl. Rad. Isotopes 4 (1959) 232.
PEELE, R. W., USAEC Report ORNL CT-61-4-32., unpubl.shed (Oak Ridge National •
Laboratory 1961).

[6] ELDRIDGE, J. S., "Scintillation counting techniques," Ch. 6, Guide to
Activation Analysis (LYON, W. S., Ed.), Van Nostrand, Princeton, N.J.
(1964).
ELDRIDGE, J. S., BROOKS, A. A., Nucleonics 24 4 (1966) 54.
TABLE I
RADIONUCLIDES CURRENTLY AVAILABLE IN THE STANDARD LIBRARY
Standard
Number -
Standard
Member
Nuclide
Тве
Nuclide
95ND
99M0 (a)
1067, (a)
Zara
nő ~5 w
110 A8
124gb
125 sb (a)
1304
520r
54Mn
1347
$6CC
Event 20.0
137C: (a)
56Mn
14032 (a)
59Fe
ü
60CO
141ce
144ce(a)
64cu
147Na
6520
7262
o o ü ş
1535m2
75se
31
:
:
76AS
82BT
8557
BERD
ño o X
192 IT
198 Au
(a) Indicates equilibrium with radioactive daughter.
TABLE II
ASSAY OF IAEA STANDARD SOURCES USING THE PUNCHED-PAPER TAPE SYSTEM
Source
Half-life (y
IABA Value (uci).
Assay Value")
% Difference
54Mn
+0.2
*22N8
0.830
2.62
10.51
20.46
10.53
10.41
10.55
6ºCo
137CS
5.263
29.
10.59
10.88
-0.5
- -0.4
-2.5
20.61
la sources were from IAEA set #116 calibrated gamma-ray sources of January 1, 1965.
"Measurements were made on July 24, 1966 and corrected to January 1, 1965.
.
..
'
-.
-
-
TABLE III
RESULTS OBTAINED FROM LIBRARY ASSAY SYSTEM WITH FOUR SPECTROMETERS AT ORNL
Ratio of amount adiled to amount found
Spectrometer System
)
12456
1.013.
198A
1.004
0.959
0.964
0.990
1 370s
0.992
0.957
0.943
0.969
59Fe_
1.004
0.944
0.9148
0.978
51cr
0.990
0.962
.
:
0.983
0.991
D
----
1) System A:
B:
C:
D:
14 x 14 x 22-in. shield
14 x 14 x 22-in. shield
14 x 14 x 32-in.shield
12-in. diameter x 20 in. high round shield
.
.
.
TABLE IV
RESOLUTION OF FOUR COMPONENT MIXTURE OF RADIONUCLIDES •
USING LIBRARY AND LEAST-SQUARES METHOD
Nuclide
Amount added (uci)
Amount found (uci)
46SC
52Cr
595e
0.389
0.865
0.592
0.627
0.392 $ 0.008
0.841 + 0.043
0.603 + 0.014
0.632 0.005
1340s
•
..
.
o..
.

FIGURE CAPTIONS
Fig. 1
Fig. 2
Computer output sheet showing library data for 59Fe.
Recalled 598e library standard. Shaded region is integrated area
discussed in section 4.
Composite spectrum of quaternary mixture of radionuclide's used
in computer resolution test. Dashed lines indicate individual
components.
Fig. 3
.
**are
**
...
not.

"
--
--
:Y
-
+
---
•
-
:--
.=+*
-
-
i
-
i
.
5
--
17
4.
.
:
Y
: Figure li imputen. Output Data Prom Pooling Program For "Fe Standard i .
1. POOLED GAMMA-RAY SPECTRA STANDARD RESPONSE 13 FOR ANALYZER NO. 1 NO. CHAN 255 LOCN 2 HICN 160 STDSF 10.00
2. ISOTOPE FE 59 STD. NO 51 STDERR 3.0 DECAY 0.1780E-06 BETA ABS 1 # 0.736 PKLOC 1290.0
3. · POOL OF SPECTRA NOS. (AMOUNIS SCALE FACTORS)
98 0.27802 05 20.036 99 0.2780E 05 10.033 100 0.2780E 05 10.050
101 0.2780E 05 10.041 102 0.2780E 05 10.046 103 0.2780E 05 10.047
17
.
CHIN
• '*
IK
*
.
I
12
21
31
41
51
CORRECTED COUNTS
4937-5 1629-3
1195-3 1290-3
3611-3 2707-3
1310-3 1910-3
10683 1064-3
1035-3 1067-3
1051-3 2046-3
1144-3 1165-3
1309-3 1301-3
7825.4 7373-4
8465-4 9042-4
4829-3 4718-3
3245-4 3800-4
2653-3 2166-3
4662-5 3532-5
5347-6 6883-6
3401-4
1346-3
2342-3
1332-3
1039-3
1041-3
1069-3
1170-3
1312-3
6988-4
61
72
81
91
101
111
121
132
141
151
1126-3
1684-3
2089-3
1369-3
1035-3
1032-3
1081-3
1205-3
1305-3
6812-4
1256-3
2161-3
7585-4
1115-3
2305-5
1069-7
1108-3 1065-3
2232-3 2081-3
1823-3 1577-3
1258-3 1212-3
1043-3 1051-3
1029-3 1036-3
1078-3 · 1090-3
1324-3 1251-3
1234-3 1167-3
6893-4 6887-4
1639-3 2212-3
1397-3 8566-4
1126-3 1613-3
7483-4 4496-';
1479-5 2049-6
2383-7 1322-8
1098-3
2715-3
1360-3
1115-3
1046-3
1043-3
1122-3
12843
9016-4
1030-3 1071-3
1557-3 1633-3
1493-3 1424-3
1153-3 1122-3
1016-3. 1032-3
1042-3 1048-3
1067-3 1089-3
1242-3 1253-3
2102-3 9989-4
6898-4 7209-4
3060-3 4032-3
5451-4 3819-4
2151-3 2611-3
2708-4 1749-4
2615-6 7316-6
1460-7 4731-7
1161-3
3846-3
1305-3
1089-3
1031-3
1060-3
1134-3
1288-3
8203-4
7914-4
5026-3
2842-4
2950-3
7427-5
4690-6
5394-7
17
E
5354-4
1609-3
2724-5
2468-6
4746-3
3029-
2899-3
1063-4
9456-6
2020-6
UPPER CHANNEL CHANGED TO 157
.
M
.
.
-
T'
-
'
--
'
*
+
A
-
Z
+
-
*
'
a
i
.-
*
.
-
.
Figure 2. Recalled 59Fe library standard. Shaded region is integrated area
discussed in section 4.
ORNL-DWG. 66-1758
1
-
COUNTS PER DISINTEGRATION ( X 5.224

59 Fe Standard 13
Source Distance, 10 cm
Absorber, 0.736 g/cm2
Polystyrene
Energy Scale, 10 keV/channel
mm
50
100
CHANNEL NUMBER
950.
-
CHAD
ORNL-DWG. 66-1759

Composite for y-Spect, Resolution
134 CS ----
-
---
-
46 SC.
0
59Fe ---
0
51 Cr .............
-
-
700 59Hihnemr etendard.
Source Distance, 10 cm
Absorber, 0.736 g/cm Polystyrene
Energy Scale, 10 keV/channel
.
.
-
-
.
(x100)
-
O
-
-
-
COUNTS PER MINUTE ;
-
.....
......go
.
....
..
MIT
Shaded region is integrated area
DS
TWITTTT
150
250
100
CHANNEL NUMBER
hr
Figure 3. Composite spectrum of quaternary mixture of radionuclides used
in computer resolution test. Dashed lines indicate individual
components.
--------
LL
W
IN
UT
Die
MY!
22
W
19
bmw
I.
by
SW White
FILMI

a way
NUTUPA
"
A
+ L
MINU
END
i
-
.
,
1
DATE FILMED
10/25 /66



.
un
-
.
TE
x
M',
,'
wiii
..'
W
V
I.
WIN
WKWA
.com
0.
01
11
M
.