EPA-660/2-74-045
June 1974
Environmental Protection Technology Serit.
Comparison of Germanium Detectors
For Neutron Activation Analysis For
Mercury
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and -non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-660/2-74-045
June 1974
COMPARISON OF GERMANIUM DETECTORS
FOR NEUTRON ACTIVATION ANALYSIS FOR MERCURY
by
Robert V. Moore
Oliver W. Propheter
Southeast Environmental Research Laboratory
College Station Road
Athens, Georgia 30601
Project #16ADN 42
Program Element 1BA027
National Environmental Research Center
Office of Research and Development
U. S. Environmental Protection Agency
CorvalUsu Oregon 97330
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price Sfi cents
-------
•^ABSTRACT
V
Two types of lithium-drifted', -solid-state, germanium
detectors were compared for their ability to detect and
measure mercury in matrices of different complexity. We
compared (1) a large, coaxial detector with relatively high
efficiency and a good peak-to-Compton ratio, and (2) a thin
wafer detector, called a low energy photon detector (LEPD),
which has a good resolution for low energy photons. In
samples with relatively few elements primarily of low atomic
number, the large detector is preferred because of its
greater counting efficiency. In complex samples containing
many elements that interfere with the mercury peak, e.g.,
samarium, thorium, barium, and tungsten, the detector of
choice is the LEPD because of its ability to resolve the
gamma photons. The choice of detector for intermediate
samples would depend on the quantity of interfering elements
present.
11
-------
CONTENTS
Abstract
Section
I Conclusions and Recommendations 1
II Introduction 2
III Experimental 3
IV Results and Discussion 6
V References 12
111
-------
SECTION I
CONCLUSIONS AND RECOMMENDATIONS
In instrumental neutron activation analysis (INAA) for mercury
in samples composed primarily of water or organics with rela-
tively few metallic elements, the large coaxial, lithium-
drifted, germanium, solid state detector is preferred because
of its greater efficiency. In complex samples containing
many elements, especially those containing samarium, thorium,
barium, and tungsten, the low energy photon detector (LEPD)
is preferred because of its significantly better ability to
resolve low energy photons.
The laboratory using instrumental neutron activation analysis
should be equipped with germanium detectors of both types.
The large detector will provide information over a wide
energy range with better accuracy and precision. In the
analysis of complex samples in the low energy range (below
150 keV), the LEPD is mandatory in order to resolve the multi-
tude of gamma and x-ray emissions.
When a sample containing a large quantity of phosphorus is
activated, it emits large amounts of bremsstrahlung that can
prevent instrumental analysis of all low energy emissions.
-------
SECTION II
INTRODUCTION
Mercury continues to be a problem in environmental pollution.
Mercury in sediments, in water, in fish, and in other environ-
mental samples is measured by cold vapor atomic absorption
and other analytical methods in which the mercury must be
separated from the sample prior to measurement. Loss of .
mercury during separation is a problem to the analyst. A
method that does not use pre- or post-treatment is desirable
to referee the usual analytical methods. Instrumental
neutron activation analysis (INAA) is such a method. INAA is
a nondestructive, multielement, qualitative and quantitative
method. It would probably be the preferred method for
determination of mercury at low concentrations if more nuclear
reactors were available.
When mercury is irradiated with thermal neutrons, two
radioisotopes are produced, Hg-197 and Hg-203. Hg-197, which
has a higher activity, is usually the preferred isotope for
measurement. Hg-203 is used for confirmation when necessary.
The primary gamma photon of Hg-197, with an energy of 77.3
keV, is in the range of maximum interference from Compton
scattering, natural background, x-rays, and bremsstrahlung.
In addition, the sample matrix may contain elements whose
decay produces photons that would interfere with the measure-
ment of the 77.3 keV gamma of Hg-197.
Two types of detectors are available for INAA, both of which
are of the Ge(Li) type. Sodium iodide detectors, Nal(Tl), are
not suitable for INAA because of their poor resolution. The
first Ge(Li) type is a relatively large coaxial or modified
coaxial detector designed to have a relatively large active
volume, the purpose of which is to increase the efficiency of
detection of gamma photons over a large range of energies.
The second type is a thin wafer Ge(Li) called a low energy
photon detector, LEPD. It was designed to allow most high-
energy photons to pass through undetected, and has its
greatest efficiency of detection for low-energy photons, i.e.,
below 150 keV. The more notable difference in these two
detector types is in their ability to resolve energies. The
larger types have resolutions of the order of 2 to 7 keV full-
width-half-maximum (FWHM) of the photopeak. The LEPD's have
resolutions in the 0.2 to 0.7 keV FWHM range.
This paper compares the two Ge(Li) detectors for INAA of
mercury in various matrices.
-------
SECTION III
EXPERIMENTAL
MATERIALS
Mercury standards were made from reagent grade mercury
dissolved in nitric acid and maintained in solution in 1.6 N
nitric acid to prevent adsorption onto the glass vessel.
Serial dilutions gave solutions ranging from 10~5 to 10~9 g
Hg/ml of solution. Dilutions were made in distilled water,
water from the Chattahoochee River taken just north of
Atlanta, Georgia, and ocean water taken near Charleston,
South Carolina.
Mercury analyses were performed on three types of organic
matter: orchard leaves, National Bureau of Standards standard
reference material 1571; bovine liver, National Bureau of
Standards standard reference material 1577; and three fish
samples used in the "Mercury-in-Fish Round Robin, 1972,"
Environmental Protection Agency, Surveillance and Analysis
Division, Chemical Services Branch, Athens, Georgia.
Twenty-one sediment samples were analyzed, eighteen sediments
from various sourcesl and three samples from the "Sediment
Round-Robin, Mercury—1972," Environmental Protection Agency,
Surveillance and Analysis Division, Chemical Services Branch,
Athens, Georgia.
EQUIPMENT
The large Ge(Li) detector was manufactured by Nuclear Diodes,
Incorporated. It is a cylindrical, modified coaxial detector
with a diameter of 38.5 mm and a depth of 35.5 mm. The well
is 6.9 mm in diameter. The window is composed of aluminum
0.5 mm thick. Relative to a 3 x 3 inch Nal(Tl) detector, its
efficiency for the 1.333 MeV gamma of cobalt-60 was 7.9% with
a peak-to-Compton ratio of 33/1. Its resolution was 1.40 keV
FWHM at 122 keV. , The preamplifier and linear amplifier were
manufactured by Nuclear Diodes, Inc.
The LEPD used was obtained from Ortec Incorporated. Its
dimensions are 16.0 mm active diameter by 4.76 mm active depth
with an 0.13 mm beryllium window. The resolution was 0.502
keV FWHM at 122 keV. "The preamplifier and linear amplifier
were also manufactured by Ortec Inc.
-------
The amplifier signals were processed by a Nuclear Data,
Incorporated ND 2200 multichannel analyzer system with a 4096
channel resolution and a 1024 channel memory. The ADC had a
100 MHz digitizing rate. The multichannel analyzer was
calibrated so that the gamma peak occupied 5 memory channels
when either detector was used.
PROCEDURES
All samples were encapsulated in quartz vials to prevent the
loss of mercury experienced with plastic vials.2 Eight one-
gram replicates of each liquid sample and four half-gram
replicates of each solid sample were analyzed. In the prepa-
ration of the encapsulated samples for irradiation, two vials
containing samples were packaged with two vials of standard in
a square configuration in alternate positions. One gram of
solution containing 1.0811 x 10~6g Kg/ml was placed in each
standard vial. This standard concentration was chosen to give
a good statistical count in moderate counting time. All
samples and standards were weighed, as it was impractical to
deliver a quantitative volume into a quartz vial. The density
of the standard was close enough to 1 that the concentration
differential was negligible.
Thermal neutron irradiations were made in the one-megawatt
experimental nuclear reactor at the Nuclear Research Center,
Georgia Institute of Technology, Atlanta, Georgia. Irradia-
tions were made in a vertical thimble inside the reflector
for periods of 5 to 9 hours at a flux of 8.6 x 1012 neutrons/
cm2/sec.
Irradiated samples were allowed to decay three days or longer,
depending on the sample matrix. A minimum of three days was
necessary to permit decay of the activated elements in the
quartz. Solutions of mercury in distilled water or river
water could be analyzed after three days of decay. Ocean
water and the sediments required at least 10 days of decay
before they could be counted. The organic material could be
counted after 6 days if absorbers were used to attenuate
moderate amounts of bremsstrahlung originating from the
phosphorus-32. If the phosphorus-32 concentration is too
large, bremsstrahlung obscures the mercury-197 peak.
During gamma analysis each replicate and standard was counted
with the quartz vial mounted as close to the detector face as
possible for the best available geometry that would permit no
more than 10% dead-time. (Dead-time is the time spent by the
multichannel analyzer to process gamma photons.) The
-------
replicates and their standards were counted in the same
position and were rotated during counting to minimize geometry
effects, especially for the solids. Each replicate and
standard was counted by both detectors.
-------
SECTION IV
RESULTS AND DISCUSSION
Table 1 presents the results of analysis of the solutions of
mercury made up in distilled water, river water, and ocean
water. The data show that the large detector is better in
both accuracy and precision. In ocean water an increase in
background diminishes the sensitivity of the method as
demonstrated by the failure to detect mercury in the 10~9g
Hg/g range.
Table 2 presents the data for the biological samples. The
nominal values for orchard leaves and bovine liver are taken
from the Certificate of Analysis of the National Bureau of
Standards. Those for the fish samples were the mean values
from the round-robin results. The determined values were
higher than nominal values except for bovine liver, in which
mercury could not be detected because of bremsstrahlung from
phosphorus-32. No specific reasons can be ascribed for the
higher values. Results for the two detectors are essentially
equivalent.
Table 3 presents data for the sediments. Nominal values for
the "Sediment Round-Robin, Mercury—1972" samples are the mean
values of all the participants' results. The other samples,
being natural sediments,,have no nominal values. All of these
sediments were complex in composition with measurable quanti-
ties of many rare elements. Consequently, the background was
high and direct interference was appreciable. Since many of
the elements had half-lives as long or longer than that of
mercury, the background and interference did not decay away.
This background made the INAA for mercury less sensitive in
sediments than in the solutions or biological samples tested.
The better results for the LEPD can be attributed to its
ability to resolve the many peaks. The LEPD was at least one
order of- magnitude more sensitive than the large detector.' ,
"Not detected" was noted rather than "less than" values for
those samples in which mercury was not detected, because the
limitation of measurements was not primarily a function of the
detector but was largely a matter of the composition of the
individual sample. For two samples, 37933 and 37934, mercury
is listed as "present". The 77.3 keV gamma did not produce a
peak with enough counts to justify a concentration computation,
but, along with the x-rays, was sufficient to indicate that
mercury was present in the sample.
6
-------
Table 1.
Standards
Sample Actual Cone.*
A
B
'c
D
E
A
B
C
D
E
A
B
C
D
E
*
**
_5
1.0811x10 ,
' 1.0811x10"°
1.0811x10"!
1.0811x10"°
1.0811x10"
-5
1.0811x10 ,
•• ri
1.0811x10 °
1.0811x10"'
1.0811x10"°
1.0811xlO"y
_5
1.0811x10 ,
— n
1.0811x10 °
1.0811x10"!
1.0811x10"°
1.0811x10"
Concentration g Hg/g sample
Relative oer cent Standard Devi;
LEPD
Detector
Results
Coaxial
Cone.* Rel.% S.D.**
Distilled Water
_5
1.06x10 ,
1.10x10"°
1.09x10"'
1.23x10"°
3.17x10"*
River
c
1.06x10"^
1.06x10"°
1.08x10"'
1.35x10"°
2.51x10"*
Ocean
_5
1.03x10 ,
— n
1.03x10 °
1.12x10"'
1.14x10
5.0
3.7
3.9
7.2
19.
Water
4.8
3.7
312
8.9
35.
Water
6.0
4.8
6.7
64.
Not Detected
ation
Cone.*
_5
1.07x10 ,
1.08x101 -j
1.05x10 R
1.12x10"°
1.82x10
_5
1.06x10 ,
~ f\
1.09x10 °
1.06x10"!
1.12x10"°
1.59x10"
c
1.05x10"^
1.07x10" 5
1.06x10"!
1.46x10"
Not
Rel.% S.D.**
3.0
1.9
2.3
6.9
31.
3.4
2.7
1.8
6.3
26.
4.3
3.2
6.2
40.
Detected
-------
Table 2.
Biological Samples
Detector
Sample
Orchard leaves
Bovine liver
Fish
72cl222
72cl223
72cl224
Nominal
Cone.
1. 55x10" 7
1.6 xio"8
2.06x10"?
5.75x10 "£
7.23x10"
Value
Rel. % S.D.
9.7
12.5
40.
28.
28.
LEPD
Cone. Rel.
2.08xlO"7
Not Detected
2.71x10"!?
7.63x10"°
9.21x10"
% S.
25.
13.
5.0
2.7
Results
Coaxial
D. Cone. Rel.
2.95xlO"7
Not Detected
2.84xlO"g
7*00x10"°
8.57x10
% S.D.
18.
8.1
2.9
5; 2
00
-------
Table 3.
Sample
Round- Robin
72c5643
72c5644
72c5645
Natural
37913
37914
37915
37916
37917
37918
37919
37920
37921
37922
37923
37933
37934
37935
39736
37947
37950
37961
Sediments
Nominal Value LEPD
Cone. Rel. % S.D. Cone. Rel. % S.D.
_
-4 -4
1.096x10 5 20 1.60x10 ^ 2.5
4.39 xio"^ 21 6U9xlO 4.2
2.2 xlO~ 170 Not Detected
6.56x10"^ 4.0
1. 11x10":: 3.1
1.9 xlO"' 53.
5.22x10"° 8.6
2.5 xlO "' 48.
3.8 xlO"' 42.
1. 17x10° 9.4
Not-Detected
1.41x10"; 40.
•• £L
3.36x10 * 4.8
Not^Detected
M /
8.7 xlO ' 31.
Present
Present
2.37x10", Hi
1.8 xlO"' 56.
Not Detected
Not Detected
-
Coaxial
Cone. Rel. % S.D.
-4
1.17x10 * 5.3
4.69xlO"D 1.3
Not Detected
S.SSxlO"4, 3.1
1. 25x10" J 3.9
NotfiDetected
3.36x10"° 14.
Not Detected
Not -Detected
1.23x10 9.8
Not Detected
Not. Detected
.»/!
2.55x10 * 3.8
Not Detected
Not Detected
Not Detected
Not -Detected
1.85x10"° 7.6
Not Detected
Not Detected
Not Detected
-------
Compton backscatter is the primary general source of
interference in INAA. When a gamma photon interacts with any
atom and transfers some of its energy to that atom, a degraded
photon and an active atom result. The degraded photon, known
as Compton backscatter, and the emission from the active atom,
if detected, will add to the spectrum background. When the
sample matrix is essentially hydrogen, oxygen, carbon, and
nitrogen, as in water and biologicals> the background is low
since these elements either are not activated or their half-
lives are very short. As the complexity of the sample
increases to include appreciable quantities of elements with
atomic numbers greater than ten, the background goes up and
the possibility of direct photon interference increases. For
these reasons, sediment background problems are great.
Two elements that cause trouble in mercury analyses are
samarium and thorium. Thorium-232 when activated, quickly
decays to protactinium-233, which interferes with mercury-197.
Other elements such as gadolinium and rhodium could also inter-
fere but have not been encountered in the sediments tested in
this laboratory. Samarium and thorium appear to be ubiquitous
in nature. They are amenable to activation analysis, and they
have half-lives that are appreciable with respect to that of
mercury. The gamma photon energy for mercury-197 is 77.3 keV;
for samarium-153 it is 75.4 keV, and for protactinium-233 it
is 75.3 keV. If the ratio of detected gamma photons from the
interfering elements is high with respect to number of
detected mercury gamma photons, the mercury photons could be
completely masked and thus mercury could not be distinguished
using the large detector. The LEPD, because of its better
resolution, can separate the interfering gamma photons from
mercury gamma photons resulting in a positive analysis for
mercury.
When a radioactive element decays to a new daughter eliement,
the daughter element is in a metastable state. In decaying to
the stable state, the daughter emits energy in the form of
x-rays, which can have energies up to 120 keV depending upon
the source. Mercury-197 decays to gold-197, which in turn
emits four x-rays at 67.0, 68.8, 78.0 and 80.2 keV. The
first, second, and fourth x-rays can make it difficult to
determine the baseline. The third x-ray, 78.0 keV, is included
in the count of the Hg-77.3 keV gamma; however, since the
standard also includes it, it causes no error. Mercury-203
decays to thallium-203, which emits x-rays at 70.8, 72.9, 82.6,
and 84.9 keV, adding to the background problem. The LEPD is
able to resolve these x-rays, permitting the counting of the
10
-------
77.3 keV gamma photons. Other possible sources of x-rays in
the 77 keV range are osmium, iridium, platinum, lead, bismuth,
polonium, astatine, and radon. Polonium, astatine, and radon
are daughter elements of the decay of natural thorium and
uranium and are so rare as to pose no problems. Osmium,
iridium, and platinum are very rare, and normally will not be
a problem. The weak polonium x-rays caused by bismuth, which
is rare, also do not interfere. The large amounts of lead
used in a detector shield absorb large quantities of energy
and will produce interfering x-rays. However, these x-rays
can be prevented from reaching the detector by proper cladding
of the shield with copper and cadmium.
A few elements emit gamma photons that, while not interfering
directly with the 77.3 keV gamma of mercury-197, can add to
the difficulty of determining the baseline, since they have
energies not far removed from 77.3 keV. For example, tungsten-
187 has a 72.3 keV gamma and barium-133 has an 81.0 keV gamma.
The LEPD resolves these energies, but the large detector does
not.
A special problem arises when a sample contains an appreciable
amount of phosphorus. Phosphorus-32 emits high energy beta
particles, which in turn create a continuum of x-rays called
bremsstrahlung when the beta particles are decelerated in the
Coulombic field of atomic nuclei. Bovine liver and, to a
lesser extent, fish had appreciable quantities of phosphorus.
The half-life of phosphorus is significantly longer (14.28
days) than that of mercury-197 (65 hours) precluding waiting
for it to decay to measure the mercury. The bremsstrahlung
prevented the analysis of mercury in bovine liver but was not
sufficient to prevent the analysis for mercury in the fish
samples.
11
-------
SECTION V
REFERENCES
1. Moore, R. V. and O. W. Propheter. "Neutron Activation
Analysis of Bottom Sediments." Environmental Protection
Technology Series. EPA-R2-73-009, March, 1973.
2. Bate, L. C. "Loss of Mercury from Containers in Neutron
Activation Analysis." Radiochem, Radional. Letters. J5
(3): 139-144, 1971.
12
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
1. Report No.
2.
w
4 Tift-
COMPARISON OF GERMANIUM DETECTORS FOR
NEUTRON ACTIVATION ANALYSIS FOR MERCURY
Authr> '.-,)
Robert V. Moore and Oliver W. Propheter
United States Environmental Protection Agency
Southeast Environmental Research Laboratory
Athens, Georgia
5. Report Date s
6.
8. P^-iotrm, ^ Organization
Report No.
16ADN 42
No
1. Type jf Repc,. and
Period Covered
''12. S; • nsorit Organisation
Environmental Protection Agency Report Number, EPA-660/2-74-045
June 197U
7'j. Abstract
Two types of lithium-drifted, solid-state, germanium detectors
were compared for their ability to detect and measure mercury in
matrices of different complexity. We compared (1) a large, coaxial
detector with relatively high efficiency and a good peak-to-Compton
ratio, and (2) a thin wafer detector, called a low energy photon
detector (LEPD), which has a good resolution for low energy photons.
In samples with relatively few elements primarily of low atomic number,
the large detector is preferred because of its greater counting
efficiency. In complex samples containing many elements that interfere
with the mercury peak, e.g., samarium, thorium, barium, and tungsten,
the detector of choice is the LEPD because of its ability to resolve
the gamma photons. The choice of detector for intermediate samples
would depend on the quantity of interfering elements present.
173. Descriptors
*Neutron activation analysis, *mercury, analytical techniques,
17b. Identifiers
*INAA, gamma ray detectors, germanium detectors
l-c. COWRR Field & Group Q5A
18. Availability ^iP'We^Fc'/als. '
i , (Reptiii)
30. Security Cl.ss.
(Page) _^
21. No. $ ''•''>
Pages '
J2. Price ' i
fa^-^8- — _ — _
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. ZO24O
Abstractor
R. V.
Institution
SERL. Athens. Georgia 30601
IO2 (REV JUNF 1S71-
------- |