|— United States
Environmental Protection
1*1 M*Agency
Office of Radiation and Indoor Air
National Analytical Radiation
Environmental Laboratory
EPA 402-B-17-001
October 2019
High Resolution Gamma-Ray
Spectrometry Analyses For
Normal Operations and
Radiological Incident Response
lypothetical Cs Gamma Spectrum
Peak centroid:
661.64 keV
(calculated)
FWHM= 1,31
-------
-------
EPA 402-B-17-001
October 2019
Revision 0
High Resolution Gamma-Ray
Spectrometry Analyses for Normal
Operations and Radiological Incident
Response
National Analytical Radiation Environmental Laboratory
Office of Radiation and Indoor Air
U.S. Environmental Protection Agency
Montgomery, AL 36115
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
This report was prepared for the National Analytical Radiation Environmental Laboratory of the Office of
Radiation and Indoor Air, United States Environmental Protection Agency. It was prepared by
Environmental Management Support, Inc., of Silver Spring, Maryland, under contract EP-W-13-016, Task
Order 014, managed by Daniel Askren. Mention of trade names or specific applications does not imply
endorsement or acceptance by EPA.
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
PREFACE
This document is one of several initiatives by EPA's Office of Radiation and Indoor Air (ORIA)
designed to provide guidance to radioanalytical laboratories that will support EPA's response
and recovery actions following a radiological or nuclear incident. This guide examines the
analysis of samples by gamma-ray spectrometry during normal operations and following a
radiological incident. The guidance provided in this document for the screening and analysis of
samples should assist those federal, state, and commercial radioanalytical laboratories that will
be challenged with a large number of such samples when responding to a radiological or nuclear
incident. This document is applicable to different types of events; a radiological transportation
incident, a radiological dispersal device (RDD or "dirty bomb"), a release from an emergency
condition at a nuclear power plant, or the detonation of an improvised nuclear device (IND),
other potential radiological releases as well as normal laboratory operations. These samples will
be contaminated with varying levels of radionuclides and will represent matrices of varied
composition. Advance planning by national and regional response teams, as well as by
radiological laboratories, will be critical to ensure uninterrupted throughput of large numbers of
radioactive samples and the rapid turnaround and reporting of results that meet required data
quality objectives associated with the protection of human health and the environment. EPA's
responsibilities, as outlined in the National Response Framework, Nuclear/Radiological Incident
Annex, include response and recovery actions to detect and identify radioactive substances and to
coordinate federal radiological monitoring and assessment activities.
Detailed guidance on recommended radioanalytical practices can be found in the Multi-Agency
Radiological Laboratory Analytical Protocols Manual (MARLAP), which provides detailed
radioanalytical guidance for project planners, managers, and radioanalytical personnel based on
project-specific requirements (www.epa.gov/radiation/marlap/links.html). Familiarity with
Chapters 2, 3, 14, 15 and 18-20 of MARLAP will be of significant benefit to users of this guide.
This document is one in a series designed to present radioanalytical laboratory personnel,
Incident Commanders (and their designees), and other field response personnel with key
laboratory operational considerations and likely radioanalytical requirements, decision paths, and
default data quality and measurement quality objectives for analysis of samples taken after a
radiological or nuclear incident.
Documents currently completed include:
• Radiological Laboratory Sample Analysis Guide for Incidents of National Significance -
Radionuclides in Water (EPA 402-R-07-007, January 2008)
• Radiological Laboratory Sample Analysis Guide for Incidents of National Significance -
Radionuclides in Air (EPA 402-R-09-007, June 2009)
• Radiological Laboratory Sample Screening Analysis Guide for Incidents of National
Significance (EPA 402-R-09-008, June 2009)
• Method Validation Guide for Qualifying Methods Used by Radiological Laboratories
Participating in Incident Response Activities (EPA 402-R-09-006, June 2009)
• Guide for Laboratories - Identification, Preparation, and Implementation of Core
Operations for Radiological or Nuclear Incident Response (EPA 402-R-10-002, June 2010)
l
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
• A Performance-Based Approach to the Use of Swipe Samples in Response to a Radiological
or Nuclear Incident (EPA 600-R-l 1-122, October 2011)
• Guide for Radiological Laboratories for the Control of Radioactive Contamination and
Radiation Exposure (EPA 402-R-12-005, August 2012)
• Radiological Laboratory Sample Analysis Guide for Incident Response - Radionuclides in
Soil (EPA 402-R-12-006, September 2012)
• Uses of Field and Laboratory Measurements During a Radiological or Nuclear Incident
(EPA 402-R-12-007, August 2012)
Comments on this document, or suggestions for future editions, should be addressed to:
Dr. John Griggs
U.S. Environmental Protection Agency
Office of Radiation and Indoor Air
National Analytical Radiation Environmental Laboratory
540 South Morris Avenue
Montgomery, AL 36115-2601
(334) 270-3450
Griggs.John@epa.gov
li
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
ACKNOWLEDGMENTS
This guide was developed by the National Analytical Radiation Environmental Laboratory
(NAREL) of EPA's ORIA. Dr. John Griggs was the project lead for this document. Technical
support was provided by Dr. Robert Litman, Mr. Sherrod Maxwell, Dr. Daniel Montgomery, Dr.
David McCurdy, Mr. Stan Morton and Mr. Robert Shannon of Environmental Management
Support, Inc.
Numerous other individuals both inside and outside of EPA provided peer review of this
document, and their suggestions contributed greatly to the quality and consistency of the final
document.
Data for some examples were used with permission of the following:
• Martin Wright of Diablo Canyon Nuclear Power Plant
• James Westmoreland of GEL Laboratories, Inc.
• Tammy Just, Southern Company
• Virginia Kammerdiener, Texas Department of Safety and Health Services
Dedication
This guide is dedicated to the memory of our friend and colleague, Dr. David McCurdy. Dave
worked as a consultant for EMS for over 23 years, beginning with the Multi-Agency
Radiological Laboratory Analytical Protocols (MARLAP) manual in 1994. Dave served as
EMS's QAPP Primary Investigator, instructed courses on topics such as MARLAP and basic
radiochemistry, assisted in the method validation process, and was a key contributor to the
publication of incident response guidance. In his spare time, Dave enjoyed fly-fishing, tennis,
and traveling. He left many friends throughout EPA and the radioanalytical community, and he
will be greatly missed.
111
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
CONTENTS
Preface i
Acknowledgements iii
Contents iv
Acronyms and Abbreviations viii
Glossary x
Radiometric Unit Conversions xi
I. Background 1
A. Structure of the Document 2
B. Goals 4
C. Objectives 4
II. Processing Spectra for Identifying and Quantifying Gamma-ray Emitting Radionuclides 4
III. Radioactive Decay Modes 7
IV. Gamma Ray Identification 11
A. Gamma Ray Interaction 11
B. Potential Threat Radionuclides 13
C. Effects of High Activity Concentrations on Interferences and Identification 17
Low abundance gamma rays 17
Gamma-ray interferences not anticipated 19
Single-escape, double-escape and multiple sum peaks 20
D. Decay Correction and Count Time 27
Ensuring correct half-life 27
Correction for decay during counting 27
Adjustments to Decay Corrections Due to High Sample Activity (Correction for
Dead Time) 29
E. Decay Correction and Radioactive Equilibrium 30
Three types of radioactive equilibria exist: secular equilibrium, transient equilibrium,
and no equilibrium 30
Parent-Progeny Equilibrium not Established 38
Parent -Progeny Equilibrium Disturbed or Created by Chemical Effects 38
Incorrect Gamma Ray Used to Analyze a Radionuclide 40
F. Software Preset Functions 40
Peak search sensitivity 40
Peak uncertainty cutoff 41
Energy comparison 42
Half-life period exceeded 42
Key Line 43
Ensuring Correct Gamma-ray Abundance 43
Radionuclide libraries optimized for the type of sample or matrix 43
Abundance or fraction limit 44
Calculating individual gamma-ray activities 44
Uncertainty-based weighted mean 45
Abundance weighted mean 45
Detection capability 46
Compton background determination 46
Peak Background Subtraction 49
iv
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Analysis Range 49
Decay Correction 50
Decay Build up Options 52
Interference Correction (Peak Overlap) 52
TCS or True coincidence summing 55
Directed Fit 55
Suspect Library 56
Nuclide ID Confidence 56
G. Selecting detectors 57
H. Sample preparation 58
I. Selecting Geometries, Counting Containers, Detection Equations, Count Intervals
and Uncertainty Equations 59
Detector Geometries 59
Counting Interval and Detection Equations 61
Critical Level Concentration (Lc) 61
Minimum Detectable Concentration (MDC) 63
Minimum Detectable Activity (MDA) 64
Lower Limit of Detection (LLD) 64
Uncertainty Determination 66
J. Detector Calibration and Non-Routine Counting Geometries 67
Energy and Resolution Calibration 67
Efficiency Calibration 70
K. Correcting Efficiency for Matrix/Geometry Differences 72
Empirical Corrections for Attenuation 73
Mathematical, Modeled, and other Corrections for Efficiency and Attenuation 74
L. Establishing sample/matrix specific libraries in the software 74
M. Gamma spectrometry report review processes (verification) 76
N. Data validation and reporting protocols 78
0. References 80
Attachment I: Instructions To Laboratories Analysis Of Fresh Mixed Fission Product
Proficiency Test Samples Round III - December 2012 83
1. Introduction 83
II. PT Sample Description 83
III. Sample Analysis Instructions 84
IV. Gamma Spectrometry 84
V. Reporting Requirements 85
Attachment II: Examples of Radioanalytical Data Review And Reporting Problems 88
Example 1: Analysis of 140Ba/140La in Rainwater Samples- Incorrect Use of Decay
Correction and Equilibrium 88
Example 2: Results from the Irradiated Uranium PT 90
Example 3: Analysis of Miner's Lettuce Following the Fukushima Event-Incorrectly
Identifying Radionuclide Gamma Rays 92
Example 4: Unidentified Gamma Rays from Insufficient Library 94
Example 5: Incorrect Radionuclide Equilibrium Used to Analyze Rainwater Samples
from Fukushima 95
Example 6: Incorrect Preservation of Samples and Its Effect on Analysis 96
Example 7: Incorrect Abundance 97
v
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 8: Software Preset Functions and Radionuclide Misidentification and
Unidentification 99
Example 9: Effect of Differing Chemistry of Parent and Progeny 101
Example 10: Incomplete Preset Library to Detect Gamma Rays 102
Example 11: Sample-Source Geometry Mismatch 103
Example 12: Incorrect Placement of Samples in Detectors 104
Example 13: Incorrect Placement of Samples in Detectors and Sample Bias 105
Example 14: Incorrect Use of Half-Lives and Inappropriate Detector 107
Example 15: Incorrect Identification of Gamma Rays from Insufficient Libraries and
Incorrect Identification of Peaks 110
Example 16: Incorrectly Identified Gamma Rays Based on Energies 112
Example 17: How Progeny Activity Can Be Used To Calculate Parent Activity 113
Figures
Figure 1. Software Process Flow for Gamma-Ray Spectrum Analysis 7
Figure 2. Line of Stability (Black Squares) and Inset (Fission Product Region for A = 131 to
A 147) 9
Figure 3 A. Thorium, Uranium, Neptunium, and Plutonium -Some Nuclear Data 10
Figure 3B. The Distribution of Fission Products from 235U 10
Figure 4. Probability of Interaction as a Function of Energy for an HPGe 12
Figure 5. Comparison of a Software Library (from that used at a nuclear power plant) and
NNDC Database for 1321 18
Figure 6. Compton Edge Location as a Function of Gamma-ray Energy 20
Figure 7. The Pair Production Process 21
Figure 8. HPGe Spectrum of 60Co (the SE and DE peaks are from the 1332 keV gamma ray)... 22
Figure 9. Coincidence decay in a beta emitter 24
Figure 10. Correction Factor for Decay during Counting (DDC); Zero Dead Time 28
Figure 11. Correction Factor for Decay during Counting (DDC); Non-Zero Dead Time
(Assumed half-life of 900 s, live time is 600 s) 29
Figure 12. Secular Equilibrium Displayed by 137Cs-137mBa 31
Figure 13. Timeline from Sampling to Analysis 33
Figure 14. Transient Equilibrium for 132Te/132I and "Mo/"mTc 34
Figure 15. No-Equilibrium Case for 143Ce/143Pr 36
Figure 16. Method of Net Peak Area Determination 48
Figure 17. Decay Correction for 95Zr-95Nb 51
Figure 18. U-235 and 226Ra Overlapping Gamma Rays 53
Figure 19. A Spectrum of 60Co Showing Peak Areas (y-axis is log-scale) 55
Figure 20. Distribution of Results for a Blank sample and Sample with Detectable Activity 62
Figure 21. Screen Shots from Software Identifying Selectable Constants for MDA Equation.... 64
Figure 22. Calibration for Resolution 69
Figure 23. Example of a HPGe Efficiency Calibration Curve 70
Figure 24. Fourth Order Calibration Curve 71
Figure 25. Fifth Order Calibration Curve 72
Figure 26. HPGe Detector Efficiency Curves for Four Solids 73
Figure 27. Incorrect Equilibrium Assumed for 140Ba/140La 88
Figure 28. Parent Activity Inferred from Progeny; Progeny Half-Life Used for Its Decay
vi
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Correction 90
Figure 29. Miners Lettuce Following Fukushima Event 92
Figure 30. Unidentified Gamma Rays from Miners Lettuce 94
Figure 31. Rainwater Samples Following Fukushima Event 95
Figure 32. Dry Deposition Samples Following Fukushima Event 96
Figure 33. Waste Liquid Release Sample 97
Figure 34. Radioiodine Analysis of a Charcoal Cartridge 99
Figure 35. Proficiency Test Results for 95Zr 101
Figure 36. One Gamma Ray Used for two Radionuclides 102
Figure 37. Sample Containers of Similar but Different Size and Composition 103
Figure 38. Sample Container Geometry Holder 104
Figure 39. Incorrect Geometry Leads to Bias 105
Figure 40. Progeny Decay Correction with Incorrect Half-Life 107
Figure 41. Energy Lines from Gamma-ray Analysis of a Filter 108
Figure 42. Proficiency Test Sample Spiked with Fission Products 110
Figure 43. Analysis of a PT Fission Product Sample Using X-ray Region 112
Figure 44. Pm-147 Detected 20 Days Following an Event 113
Tables
Table 1. Possible Gamma Emitting Radionuclides Resulting from a Radiological Event 14
Table 2. High-Activity Radionuclides Resulting from a Fission Event 16
Table 3. Commonly Encountered Radionuclides with True Coincidence Sum Effects 25
Table 4. Time for Radionuclide Pairs to Achieve Maximum Progeny Activity 37
Table 5. Detection Equation Calculations 66
Table 6. Potential Radionuclides on Charcoal Cartridge 75
Table 7. Potential Radionuclides on Air Particulate Filter 75
Table 8. Checklist (partial) for Gamma Spectrometry Data Verification 78
Table 9. Example Data Validation Table 79
vii
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
ACRONYMS AND ABBREVIATIONS
(Excluding chemical symbols and formulas)
a alpha particle
a probability of Type I decision error
A mass number
ADC analog-to-digital converter
ASTM American Society for Testing and Materials
P beta particle
P probability of Type II decision error
P" negatron or beta-minus particle
P+ positron or beta-plus particle
Bq becquerel (1 dps)
BNL Brookhaven National Laboratory
CFR Code of Federal Regulations
Ci curie
cm centimeter
cpm counts per minute
d day
DDC decay during counting
DDEP Decay Data Evaluation Project
DOE United States Department of Energy
DP decay product(s)
dpm disintegration per minute
dps disintegration per second
DRL derived response levels
DRP discrete radioactive particle
8 electron capture
Epmax maximum energy of the beta-particle emission
ERLN Environmental Response Laboratory Network
EPA United States Environmental Protection Agency
FRMAC Federal Radiological Monitoring and Assessment Center
FWHM full width half maximum
y gamma radiation
g gram
Ge germanium semiconductor
Gy gray
h hour
HPGe high-purity germanium [detector]
IC incident commander
IC internal conversion
ICLN Integrated Consortium of Laboratory Networks
IND improvised nuclear device (i.e., a nuclear bomb)
in inch
IT isomeric transition
kg kilogram (103 gram)
keV kiloelectronvolts (103 electronvolts)
viii
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
L liter
LNHB Laboratoire Nationale de Henri Becquerel
MARLAP Multi-Agency Radiological Laboratory Analytical Protocols Manual
MCA multichannel Analyzer
mCi millicurie (10"3 Ci)
MeV megaelectronvolts (106 eV)
mg milligram (10"3 g)
mL milliliter (10"3 L)
mrem millirem (10"3 rem)
[j,Ci microcurie (1CT6 curie)
[j,g microgram (10"6 g)
MDA minimum detectable activity
MDC minimum detectable concentration
min minute
MQO measurement quality objective
NAREL EPA National Analytical Radiation Environmental Laboratory
NNDC National Nuclear Data Center, Brookhaven National Laboratory
NORM naturally-occurring radioactive materials
NUDAT NNDC NUDAT database
ORIA EPA Office of Radiation and Indoor Air
PAG protective action guide
pCi picocurie (10"12 Ci)
PT proficiency test (sample)
QA quality assurance
QC quality control
rad radiation absorbed dose
RCS reactor coolant system
RDD radiological dispersal device (i.e., a dirty bomb)
RFA responsible federal agency
rem roentgen equivalent: man
s second
SI International System of Units
SOP standard operating procedure
STS sample test source
Sv Sievert
t time frame
TCS true coincidence summing
TRU transuranic elements
WCS working calibration standard
y year
IX
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
GLOSSARY
Excellent glossaries of radiological and radiochemical terms including those for gamma
spectrometry can be found in ASTM D7902-14, "Standard Terminology for Radiochemical
Analyses", and also in the Multi-Agency Radiological Laboratory Analytical Protocols Manual
(MARLAP). The terms used here are in the context of those references.
x
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
RADIOMETRIC UNIT CONVERSIONS
To Convert
To
Multiply by
To Convert
To
Multiply by
years (y)
seconds (s)
minutes (min)
hours (h)
days (d)
3.16xl07
5.26* 105
8.77* 103
3.65xl02
s
min
h
d
y
3.17xl0-8
1.90x10^
1.14x10^
2.74xl0-3
disintegrations per
second (dps)
becquerels (Bq)
1
Bq
dps
1
Bq
Bq/kg
Bq/m3
Bq/m3
picocuries (pCi)
pCi/g
pCi/L
Bq/L
27.0
2.70x 1CT2
2.70x 1CT2
10 3
pCi
pCi/g
pCi/L
Bq/L
Bq
Bq/kg
Bq/m3
Bq/m3
3.70x 10~2
37.0
37.0
103
microcuries per
milliliter ((.iCi/niL)
pCi/L
109
pCi/L
(iCi/mL
10 "
disintegrations per
minute (dpm)
(iCi
pCi
4.50xl(T7
4.50x 1CT1
pCi
(iCi
dpm
2.22
2.22x 106
cubic feet (ft3)
cubic meters (m3)
2.83 xlO"2
m3
ft3
35.3
gallons (gal)
liters (L)
3.79
L
gal
0.264
gray (Gy)
Rad
102
rad
Gy
10 2
roentgen equivalent
man (rem)
sievert (Sv)
10 2
Sv
rem
102
Note: Traditional units are used throughout this document instead of International System of
Units (SI) units. PAGs and their derived concentrations appear in official documents in the
traditional units and are in common usage. Conversion to SI units will be aided by the unit
conversions in this table.
XI
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
I. BACKGROUND
Many of the country's radiological laboratories have not had the experience of analyzing
environmental samples contaminated with levels of radioactivity up to one thousand times
environmental levels and above. Even for those laboratories that have had experience with high
activity concentration samples, fresh fission products have rarely been seen in environmental
samples for the past 40 years1. The potential threat of an improvised nuclear device or a nuclear
facility accident where fresh fission products are released requires that our nation have both the
capability and capacity to analyze these specific types of radionuclides. Most of the fission
products that would result from either type of event are gamma-ray emitters; this makes gamma-
ray spectrometry a prime analytical tool for environmental analysis.
Historically, proficiency test (PT) samples have contained multiple radionuclides. With few
exceptions2, these have tended to be long-lived, many of which have not been fresh fission
products (i.e., short-lived). The same radionuclides were repeatedly used in the PT samples, and
most of the radionuclides used did not exhibit parent-progeny radiochemical equilibria. Thus,
experience identifying and quantifying short-lived fission products and parent-progeny
equilibrium radionuclides was lacking.
ORIA sponsored a contract to prepare four rounds of PT samples with fresh, short-lived fission
products to address this specific issue. The PT samples were prepared by irradiating natural
uranium, dissolving it and preparing aqueous samples containing elevated concentrations of
radionuclides with half-lives of ~2 days or longer. These samples were analyzed several times
by a reference laboratory and then sent to radiochemistry laboratories willing to participate in
that project.
The first two rounds of samples were analyzed only by gamma-ray spectrometry by 11 to 30
participating laboratories. The results were tabulated and reports provided individually to
laboratories on their performance as well as the group performance (with identities of other
laboratories being maintained as confidential). The results were surprising as there was an
overwhelming lack of experience with this type of sample and the use of the gamma-ray
spectrometry systems used to analyze such a sample. Additionally, laboratories had trouble
adhering to the analytical requirements and reporting criteria requested by EPA. Afterward, the
criteria and instructions for analysis were refined to include more specific directions on analysis
and reporting requirements. These can be found in Attachment I.
Two additional rounds of PT samples were prepared and distributed to 40 laboratories. There
was an improvement in the analytical results that were reported and the reporting protocols that
were followed in these next two events. However, there remained some concerns about how the
laboratories were validating the results of their analyses and whether or not there was sufficient
training for laboratory staff in the theory of gamma spectrometry and the operation of the gamma
spectrometry systems and software.
1 The obvious exception is for those samples taken during the Chernobyl and Fukushima events. However, in both
cases the types and activity concentrations of the fission products that were determined were far reduced due to the
long atmospheric transport times of the radionuclides.
2 For example,1311 in water and milk.
1
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Under a separate contract, site-specific training was provided to twelve state radiochemistry
laboratories. One of the topics for onsite training was gamma spectrometry theory and practice.
Some of the laboratories had also participated in the mixed fission product PT sample programs
described above. While some of them provided very good results on the PT program, it was
evident that their theoretical background in gamma spectrometry and software operation was
limited. An example of the limited theoretical background was that laboratories relied too
strongly on software analysis even though the reported results were improbable. Although
software can be adapted to analyze a spectrum for different types of samples and radionuclides,
many users did not understand the capabilities and limitations of the software, and how critical
their role was in the configuration of the software to obtain accurate and reliable results.
Rapid measurements of activities that are precise and of known and reliable quality (e.g.,
accurate identification of radionuclides) are essential in both normal and emergency response
situations. Radionuclides that are misidentified or activities that are inaccurate may lead to
inappropriate actions on the part of the client. Thus, training in data review, as well as
interpretation of spectral results, is essential. This guide was developed to be a tool to assist in
training current and new gamma-ray spectrometrists.
A. Structure of the Document
Gamma-ray spectrometry analysis has progressed significantly since single-channel analysis was
used in the 1950s and early 1960s. The energy resolution and efficiency of detection systems
have improved since that era. Sophisticated software has been developed that increases the speed
with which these analyses can be performed.
Prior to the advent of sophisticated software, practitioners gained more experience in assessing
the identities of radionuclides from the direct review of the gamma-ray spectrum, understanding
how radiochemical relationships affect activity calculations, and the determination of activity
through hand calculations. A further complicating factor for current practitioners is that they
likely have never seen samples that contain high activity (one thousand or more times
background detection limits) or spectra that contain fresh fission products or a diversity of
activation products (ones that would potentially be used in an RDD). This also makes it difficult
for today's practitioners to assess how the radioactive decay laws are applied in the logic,
algorithms, and calculations used in the internal software programs for the gamma-ray
spectrometry system. Finally, in spite of the improved capabilities and sophistication of modern
software, no standard commercially available gamma spectrometry analysis software has the
capabilities to properly account for all parent-decay product activity relationships.
2
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
The purpose of this document is to provide guidance to laboratories performing gamma-ray
spectrometric analyses of samples following a radiological or nuclear incident. Specific
examples of analyses where environmental and process samples containing fresh fission products
or other radionuclide combinations are addressed in this document in Attachment II. The
following topics are described in general theoretical terms and some of each topic can be found
in the examples:
• Processing gamma-ray spectra for identifying radionuclides.
• Radionuclide identification: gamma-ray energies and photons,
o Radioactive decay modes.
o Gamma-ray interactions with the detector,
o Potential threat radionuclides.
o Effects of high activity concentrations on interferences and identification.
> Low abundance gamma rays of high activity radionuclides which are not listed in the
libraries.
> Gamma-ray interferences not anticipated.
> Limited sensitivity due to high background.
> Single-escape, double-escape and multiple-sum peaks.
> Dead time corrections/adjustments due to high sample activity.
• Decay correction and count time,
o Ensuring correct half-life.
• Decay correction and radiochemical equilibrium,
o Parent-progeny equilibrium not established.
o Parent-progeny equilibrium disturbed by chemical effects.
o Proper use of software algorithms for decay and ingrowth correction.
o Activity correction for cases where equilibrium has not yet been established (manual).
• Software.
o Proper configuration for sample/matrix type, and geometry.
o Validation tests used during analysis (peak identification, radionuclide association with a
gamma ray, etc.).
o Ensuring correct abundance.
o Radionuclides are appropriate for the type of sample or matrix,
o Activity (quantification) and detection equations,
o Uncertainty.
• Selecting Detectors.
• Sample Preparation.
• Selecting Geometries, Counting Containers, Detection Equations, Count Intervals and
Uncertainty Equations.
• Detector calibration and non-routine counting geometries.
• Correcting efficiencies for matrix effects.
• Establishing Sample Matrix/Specific Libraries in the Software.
• Gamma-ray spectrometry report review processes.
• Data validation and reporting protocols.
• Examples of radioanalytical results.
3
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
The information provided herein is a starting point for those performing gamma-ray
spectrometric analyses. The document should be used in conjunction with analysis of actual
samples containing activity levels that range from low-level to those which are much higher than
normal, and where less commonly encountered short-lived radionuclides are present. Analyses
performed in this manner will allow analysts to experience the types of problems that will be
encountered during analysis of radiological event samples and resolved by using the guidance
provided in this document.
B. Goals
This guide presents a generic approach to gamma-ray spectrometry. It describes the effects that
gamma rays create in high-purity germanium (HPGe) detectors and in other materials and the
general set-up and maintenance of HPGe systems and software. It also addresses various aspects
of samples that may be encountered (radionuclide mixtures and matrix) and provides guidance
that allows the user to review the results of measurements to ensure that the results will satisfy
applicable measurement quality objectives (MQOs).
C. Objectives
The objectives of this document are as follows:
1. Describe the basic theoretical principles of gamma-ray spectrometry.
2. Show how the interactions of gamma rays with the HPGe detector can yield artifacts that
cannot be used to quantify radionuclides
3. Explain the radioactive equilibria and demonstrate how to calculate radionuclide
concentrations when these equilibria are present.
4. Provide examples of problems that can be encountered when analyzing specific matrices
5. Provide descriptions of the different software functions and how they are used in analyzing
the gamma-ray spectrum.
6. Provide examples of analyses that were incorrectly performed by software based on
preselected functions that were inappropriate for the type of sample analyzed, and how these
problems can be avoided.
7. Identify the different types of detection equations and how they differ in their determination
of detection.
II. PROCESSING SPECTRA FOR IDENTIFYING AND QUANTIFYING GAMMA-RAY
EMITTING RADIONUCLIDES
A gamma-ray spectrum is generated when gamma rays interact with a detector. Electronic pulses
are produced with amplitude proportional to the energy these gamma rays deposited in the
detector. The pulses are shaped, amplified, and stored by signal processing electronics that
normally includes an amplifier, an analog-to-digital converter (ADC), and a multi-channel
analyzer (MCA) that stores the pulses in digital memory according to energy. The resulting data
is displayed graphically as a histogram or spectrum showing the number of observations (counts)
as a function of gamma-ray energy. An example of a spectrum of 60Co is shown in Figure 8. The
characteristics of a gamma-ray spectrum are discussed below.
4
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Modern gamma-ray spectrometry systems include a high-resolution high purity germanium
(HPGe) detector plus electronics interfaced to a computer system with software to process
spectra and calculate the concentration of radionuclides in samples. The complexity of gamma-
ray spectra that may be encountered in emergency response scenarios precludes the processing of
spectra manually. Many software packages provide the tools for setup and calibration of the
systems for processing spectra. Two major suppliers in the United States3 provide sophisticated
software for processing spectra. The individual components of the software necessary for
analyzing samples will be discussed in Section F. The reader is also referred to an excellent
discussion of "Computer Analysis of Gamma-Ray Spectra" in Chapter 9, Reference 22.
Figure 1 (from Reference 21) shows a generic flow process for spectral analysis of a sample
applicable to most software capable of complete analyses of spectra. In practice, software
packages may implement unique approaches that optimize performance to address different
analytical challenges. Thus, one approach may be more appropriate for analyzing low-activity
samples with a limited set of radionuclides (e.g., naturally-occurring radionuclides), whereas a
different approach may be optimal for analyzing high-activity samples containing many
radionuclides, or for samples that may contain radionuclides that may be present in more
complex decay relationships.
The laboratory's Quality Manual for Radiochemistry should include any specific objectives of
the lab, but more importantly, it should identify the source documents from which all parameters
in the library originate. This kind of document is vital for succession planning, training of new
personnel, and the important standardization of parameters like energy lines, abundances, and
half-lives. For example, this document may state that all parameters of this sort are obtained
from the NUDAT database.
3 Canberra Industries, Inc. and AMETEK ORTEC
5
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
NOTE
A number of nuclear databases are commonly used by gamma spectrometrists. Some of
these, such as Kocher (.Radioactive Decay Data Tables, D.C. Kocher, DOE/TIC-11026,
1981) and several libraries supplied by instrument manufacturers, are outdated and do
not reflect the currently, best-available nuclear data. The Decay Data Evaluation Project
(DDEP) is an international collaboration of radiological laboratories. DDEP data files
can be accessed at Henri Becquerel Laboratories (LNHB,
http://www.nucleide.org/DDEP WG/DDEPdata.htm). The data in DDEP is quite
reliable as it has undergone extensive validation. It also contains extensive
details regarding nuclear constants.
This guide, however, recommends using the Decay Radiation Database at the National
Nuclear Data Center (NNDC) website, commonly referred to as NUDAT. Although this
online database is routinely updated and contains up-to-date nuclear data, it may not be
as carefully reviewed as DDEP data. However, it contains a much more comprehensive
collection of nuclear data that is much more easily accessible using a powerful query
interface. The data are also presented in a manner that is more compatible with the
preparation of gamma spectrometry libraries. The more comprehensive collection of
data is also invaluable when searching for unidentified photopeaks in gamma spectra.
While we recommend routinely using NNDC/NUDAT, we also recommend using the
more reliable DDEP data to validate the abundant library lines that drive the
quantification of gamma emitters (the top 3-4 for a nuclide). Similarly, we recommend
relying on DDEP data whenever there is any concern over the correctness of nuclear
data found elsewhere.
This effort will prevent a common failure related to lab performance degradation after key
people leave the lab through retirement, etc. and also improve general communication should
more than one person be involved in maintenance/review of the libraries.
6
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Computer
User Action Routine Data File
Edit List
Spectra
input
Control Info
Input
Peak Location
Edit List
Peak Region
Edit List
Previous
Calibration
Calibration
E(x),W(x)
Peak Shape
Parameters
Decay Data
Subset
Detector Data
Library
Analyze
Peaks
Secondary
Peaks to-be-fit
Secondary
Peak Analysis
Nuclide
Interferences
Activity
Calculation
Detector
Efficiency
Identify
Nuclides
Present
Nuclide
Identification
of Peaks
Detector
Efficiency
Parent
Daughter
Relations
Figure 1. Software Process Flow for Gamma-Ray Spectrum Analysis
The functionality of the various computer routines identified in the above table is discussed in
detail in Section F, Software Preset Functions.
Hi. RADIOACTIVE DECAY MODES
Two decay modes that can precede gamma-ray emission are alpha (a) and beta ((3) decay. Beta
decay is further subdivided into negatron (P"), positron (P ) and electron capture (£) decay.
7
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Decay only by the emission of a gamma ray (without beta or alpha particle emission) is referred
to as isomeric transition4 (IT).
Most radioactive decay by an alpha particle or beta particle emission does not result in the
release of sufficient energy for the decay product to be left at its lowest possible energy (ground)
state. The remaining energy to reduce the energy state of the decay product to the ground state is
normally released in the form of a gamma ray. Gamma rays are released from the excited state of
the decay products at discrete energies.
Radioactive decay may also include emission of characteristic X-rays. X-rays are the result of an
electron in a bound shell being expelled and the subsequent rearrangement that occurs when
higher energy electrons fall to vacant, lower energy states. The emitted characteristic X-ray will
have energy equal to the difference between the two energy states. X-ray emission will always
follow electron capture decay and may follow P", P+, alpha, internal conversion (IC) and IT decay
processes. Internal conversion is a result of a nucleus in an excited state transferring all its
energy directly to a bound electron (highest probability is a K-shell electron), and decreasing the
probability of gamma-ray emission.
Figure 2, from the Brookhaven National Laboratory NuDat database5, shows the entire range of
known radioactive and stable nuclides. A small section of the chart centered at about 140La is
expanded to show some of the detail from the NuDat website diagram. The black boxes represent
stable nuclides6. All other colors represent nuclides that are radioactive. A diagonal drawn
through the average trend of the black boxes is referred to as the "line of stability". All
radionuclides to the right of the line of stability have a neutron excess, while those to the left of
the line have a neutron deficiency. Those radionuclides to the right will decay mostly by P"
emission (neutron ejecting a negatron and becoming a proton), while those to the left will decay
mostly by P+ emission (proton ejecting a positron and becoming a neutron), or electron capture (a
K-shell electron captured by the nucleus decreasing the proton number by one and increasing the
neutron number by one). All three of these decay modes can collectively be called isobaric
transitions as the nuclide mass number remains constant while the ratio of protons to neutrons
changes. Once beyond the last stable nuclide, 209Bi alpha emission becomes a more common
form of radioactive decay from either side of the imaginary line drawn from 209Bi to 238U.
In general, the further away from the line of stability a nucleus lies the more unstable is the
nucleus and the shorter will be its half-life. The line of stability ends at 209Bi; beyond that nuclide
all nuclides are radioactive. Those elements beyond 209Bi also tend to be unstable with respect to
alpha emission, as well as other forms of nuclear decay. Figure 3 A shows the region of the Chart
4 While all gamma ray emissions are isomeric transitions, the specific decay mode of IT is given to those
radionuclides that have a singular emission to a ground state without cascading gamma ray emission, and have half-
lives that are significant in terms of being separate from any other decay mode. Typically, the IT mode is noted
when the decay half-life is more than about 1 microsecond (however there is no formalism identified for this
identification process).
5 See www.nndc.bnl. gov/chart
6 Also note that in the Chart of the Nuclides that naturally-occurring radionuclides will also have a black bar across
the top of their boxes. Thus, even though they are not stable, they are present in nature as very long-lived
radionuclides or their progeny.
8
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
of the Nuclides surrounding 2"U. In this region there are many radionuclides that decay by alpha
particle emission.
Most beta decay processes are followed by gamma-ray or X-ray emission. However, there are
several radionuclides that will decay by beta emission only. Examples of radionuclides that do
not emit gamma rays or X-rays and which may be expected to be found in samples after a
radiological event may include 3H, 89Sr, 90Sr, and "Tc; °Fe and 63Ni are also non-gamma-
emitting radionuclides but both emit X-rays.
N=50
l39Pa
140Pffl
141Pa
l*ZPm
!3«Nd
I WHcJ
H
142Nd
14 3Nd
146Nd
IJXPr
1J&Pr
IWPr
I40P*
|41Pr
14 m
144 Ft
I 36C<
!37Cc
l}tC(
140Cc
142Cc
¦
""
I3*U
I JTLi
I3*U
IJ»t«
Hit.
1421,1
I43L»
IMBi
135Bi
IJ*B»
1 37Bi
U*Bi
1J9B*
...
I418»
142Bi
133C«
13SC*
l)Xl
140C#
141C.
1 J2Xe
1 J4Xc
J«Xt
1 J7Xe
1 J*Xc
140Xc
""
1321
im
1341
4 351
1361
1371
1381
im
79
81
$3 85
m
Slisile
eo-s«
*
«
p
N
$F
Unknown
N, number Of neutron:
N=126
Figure 2. Line of Stability (Black Squares) and Inset (Fission Product Region for
A = 131 to A = 147)
9
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
ZtSPo
20.9 M
234Pu
8 8 hi
235Pu
25 .3 M
236PU
2.858 Y
237Po
45.64 D
238Pu
87.7 Y
239Pu
24110 Y
240Pu
6561 Y
241PU
14.325 Y
2
&99.8ttb
0tO.L2fe
I* 94009b
G* 6.00»
I 100005b
a 2.8&3
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
IV. GAMMA RAY IDENTIFICATION
A. Gamma Ray Interaction
The energy that is emitted in the form of a gamma ray is a direct result of a nuclear transition of
nucleons. The gamma-ray energies most often associated with radionuclides of interest in a
radiological event and which could be readily detected by gamma-ray spectrometry systems
commonly in use fall in the range of the frequently used calibration standards; 59.5 keV (241Am)
to 1836.1 keV (88Y). Operating outside this energy range (e.g., using an extended range detector)
requires additional standards to extend the detection range of the detector. This will be discussed
in more detail in Section J on calibration.
The three forms of gamma-ray interaction with the matter of significance in gamma-ray
spectrometry are the Photoelectric Effect, Compton Effect, and Pair Production Effect. The
Photoelectric Effect occurs when a gamma ray interacts with a bound electron and imparts all of
its energy to the electron. The Compton Effect occurs when there is a scattering of a gamma ray
after interaction with a bound electron such that the initial energy of the photon is shared
between the scattered gamma ray and the scattered electron. The Pair Production Effect occurs
when a photon with energy greater than 1.02 MeV interacts in the vicinity of a nucleus such that
the photon is converted into a positron and an electron. Any energy greater than 1.02 MeV is
distributed between the positron and electron as kinetic energy (this is discussed in more detail in
Section C).
Several features will be evident in gamma-ray spectra as a result of these different gamma-ray
interactions. A full energy peak is observed in the gamma-ray spectrum from multiple gamma-
ray interactions with the detector in which all of the gamma-ray energy is deposited in the
detector via the Photoelectric Effect (i.e., the full energy of the gamma ray is always absorbed).
The Compton Effect (sometimes called Compton Scattering) and Pair Production will only
produce the full energy peak if all of their subsequent interactions also occur within the detector
(these are much less probable events as they involve multiple photon interactions of low cross-
section in a small detector volume).
If the Compton-scattered gamma ray leaves the detector after any scattering interaction, no full
energy event is recorded and the resulting signal is effectively noise (i.e., contributes to the
background) at a lower energy. The collection of noise in the gamma-ray spectrum from
Compton scattering is called the Compton continuum.
Figure 4 shows the relative response of an HPGe detector to gamma rays for these three effects
over the energy range of 1 keV to 100 MeV.
11
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
1.0E+04
1.0E+03
• Compton
— — Photoelectric
1.0E+02
Pair Production
1.0E+01
ToteT
1.0E+00
c
o
1.0E-01
+j
re
3
c
01
4-<
<
1.0E-02
1.0E-03
1.0E-04
1.0E-05
1.0E-06
1
10
100
1000
10000
100000
Energy (keV)
Figure 4. Probability of Interaction7 as a Function of Energy for an HPGe
While we focus on the calibration energy range for quantitative purposes, it is important to
remember that photopeaks also may be identified by the software in the quantifiable regions of
the spectrum that are due to gamma-ray artifacts produced in the detector. Artifact photopeaks
that occur as the result of:
• Single-escape peaks;
• Double-escape peaks;
• Summation peaks (either random or coincidence);
• Peaks within -0.5 keV of the annihilation peak (511 keV);
• An apparent peak at a Compton edge8;
• X-rays, especially from high-Z elements9; and
• Germanium x-ray escape peak 11 keV below a photopeak (in particular a low energy
photopeak);
are not used to quantify radionuclide activities using only the algorithms contained in the
gamma-ray spectrometry software. Software used for identification of gamma-ray peaks may
have different methods of allowing the user to identify these peaks as artifacts and not use them
7 Diagram taken from, "Handbook of Radioactivity Analysis", 3rd Edition, Michael F. L'Annunziata, Academic
Press, 2012 (Reference 26)
8 The Compton Edge refers to the highest energy point of the continuum. The background counts drop off
precipitously above this energy (see Figure 8 for an example).
9 Z refers to the number of protons in an element.
12
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
in quantitative peak analysis. It is critical that these user-selectable functions be understood and
spectra for non-routine samples (especially those from a radiological emergency) be reviewed for
their applicability to those radionuclides and activity concentrations present.
Some artifacts also may be formed by several events yielding a detector response at a specific
energy in the gamma-ray spectrum, but the software determines that it has the shape and
characteristics of a gamma ray. The only interaction of a gamma ray with the detector that has
both quantitative and qualitative value is through the Photoelectric Effect. This interaction
provides a measure of the full energy of the original gamma ray. Two other interactions of note
that lead to routine features in the gamma-ray spectrum are the Compton Effect and the Pair
Production Effect. Review of Figure 4 shows the relative probabilities of interactions of gamma
rays with an HPGe detector as a function of energy. From low gamma-ray energy up to about
140 keV, the principal means of interaction is by the Photoelectric Effect. The Compton Effect is
dominant between 140 keV and about 8,000 keV. Above 8,000 keV, pair production becomes
dominant. Note that the threshold energy for pair production to occur is 1,022 keV. For several
naturally-occurring and man-made radionuclides, there are gamma rays with energies in the
range of 2,000 to 3,000 keV that will produce peaks (single- and double-escape peaks-see
description in next section) in the energy range used in gamma-ray analysis. These may appear
as unknown peaks in the gamma-ray report since these energies do not correspond to real gamma
rays. In addition, the gamma-ray source of these artifacts may lie beyond the calibration range
normally used. Thus, these gamma-ray energies are often not included in software libraries.
Some gamma-ray software uses a "suspect library" to assist in identifying unknown peaks such
as the single- and double-escape peaks and their sources. This software will assign potential
radionuclide identities based on a separate library that can be used after initial analysis of the
spectrum.
Calibration curves for efficiency, energy, and resolution will be discussed in more detail in
Section M.
B. Potential Threat Radionuclides
Table 1 lists some radionuclides10 that potentially could be used in a radiological dispersal device
(a "dirty bomb" or RDD). Table 2 identifies radionuclides that could result from a release of
fresh fission products (such as a nuclear power plant breach or an IND). These radionuclides
could subsequently contaminate environmental samples in the vicinity of ground zero and
downwind of the event.
Several of the radionuclides in these tables have progeny that also are radioactive. For example,
if 241Pu is identified, 241Am likely will be present. However, the extent to which progeny may be
present depends on the age of the material used in the RDD.
10 Radionuclides with half-lives less than about 12 hours have not been included in this list (unless they are short-
lived progeny of a long-lived progenitor) as most samples analyzed by laboratories supporting the event would be
days old.
13
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Table 1. Possible Gamma Emitting Radionuclides Resulting from a Radiological Event
Alpha Emitters[1]
Radionuclide
Gamma
Energy, keV
Gamma-ray
Abundance [6]
Half-Life
Radionuclide
Gamma
Energy, keV
Gamma-ray
Abundance
[6]
Half-Life
Am-241
59.5
0.359
432.7 y
Ra-226
186.2
0.0364
1.599xl03 y
Cm-242
44.1
0.000035
162.8 d
Th-228
84.4
0.0122
1.91 y
Cm-243
277.6, 228.2
0.14,0.106
29.1 y
Th-230
67.7
0.0038
7.56xl04 y
Cm-244
42.8
0.0026
18.1 y
Th-232
63.8
0.000263
1.4\10 \
Np-237
86.5
0.124
2.14xl06 y
U-234
53.2
0.000123
2.46xl05 y
Pu-238
43.5
0.000392
87.7 y
U-235
185.7
0.570
7.04x10s y
Pu-239
51.6
0.000272
2.41xl04 y
U-238
49.6
0.00064
4.47x10 \
Pu-240
45.2
0.000447
6.56xl03 y
U-Nat
185.7 (235U)
0.570
4.47x10 \
Beta Emitters
Radionuclide
Gamma
Energy,
keV
Gamma-ray
Abundance [6]
Half-Life
Radionuclide
Gamma
Energy,
keV
Gamma-ray
Abundance [6]
Half-Life
Ac-227 PI /
Th-227
236
0.129
21.7 y/18.7 d
Ba-140/La-
140
537/1596
0.2439,
0.9540
12.8 d/1.68 d
Bi-212 [4]
727
0.0667
60.6 min
Mo-99/Tc-
99m [3]
740, 141
0.1226,0.89
2.75 d/6.01 h
Bi-214 W
609
0.455
19.9 min
Pd-103
39.7
0.00683
17.0 d
Co-57
122, 136
0.856, 0.1068
271.8 d
Pb-210 [4]
46.5
0.0425
22.3 y
Co-60
1173, 1332
0.9985,0.9998
5.271 y
Pb-212 I 'l
239
0.436
10.6 h
Cs-137[2]
662
0.851
30.0 y
Pb-214 [4]
352
0.356
27 min
1-125
35.5
0.0668
59.4 d
Pu-241|2S|/
Am-241
59[5]
0.359
14.3 y
1-129
39.6
0.0751
1.57x107 y
Ra-228 [2]/
Ac-228
911 (Ac)
0.258
5.76 y/6.15 h
1-131
364
0.815
8.01 d
Ru-106 PI /
Rh-106
511.9,
622
0.204, 0.0993
1.02 y/299 s
Ir-192
317
0.8286
73.8 d
Se-75
265, 136
0.589, 0.585
119.8 d
Notes:
[1] The radionuclides in this table are alpha emitters generally with low abundance gamma rays. However, if any of
these were part of an event (e.g., from an RDD, or a nuclear power or reprocessing plant materials) as pure
radionuclides, detection using gamma-ray spectrometry could be very effective.
[2] Beta only emitter; progeny emits gamma ray.
[3] Parent is a low abundance or non-gamma emitter; progeny used for quantification by gamma spectrometry. See
discussion of decay equilibria in Section E.
[4] These radionuclides are found in many enviromnental samples as a result of being decay progeny of 226Ra or 224Ra.
Care should be taken in the assigmnent of their half-lives in gamma spectrometry libraries. See discussion of decay
equilibria in Section E.
[5] The production of the 239+24uPu isotopes results in the production of 241Pu as a result of multiple neutron captures.
Pu-241 has a 14-year half-life. If the nuclear material is "old", measurable activity of241 Am (a beta decay product of
241Pu) may be present, even if this radionuclide was not originally present in the source term for an event.
[6] Abundances, half-lives, and energies were taken from the National Nuclear Data Center, Brookhaven National
Laboratory, http://www.nndc.bnl.gov/chart/. Uncertainties are not included here but may be found at the referenced
site.
14
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
When the alpha emitters noted in Table 1 are present in samples at normal environmental
concentrations, they are best determined by means other than gamma-ray spectrometry.
However, if these radionuclides are the major source of the radioactive contamination and
present at high activities, gamma spectrometry (using the gamma rays noted) may be a good
method of analysis. U-235 and 241 Am fall into the range of gamma-ray emitters where their
analysis by gamma spectrometry is possible, but their abundances are relatively low compared
with other gamma emitters. Additionally, when analyzing for 241Am, with a gamma ray at 59
keV, special consideration should be given for matrix density (more attenuation of the gamma
ray because of the low energy) and background from Compton electrons (high background in
lower energy area when high energy gamma rays are present).
Several of the Table 1 radionuclides emit gamma rays (in addition to the ones shown) in the
range of 29 to -100 keV. This region of the electromagnetic spectrum also contains X-rays from
the decay of fission product radionuclides such as I, Xe, Cs, Ba, La, and Ce as well as from Pb
X-rays due to excitation of the shielding material from the sample gamma rays. If more than one
of these radionuclides is present there will be multiple X-rays from these radionuclides which
will overlap and significantly compromise the ability to make a quantitative assessment of the
activity due to a particular radionuclide.
Table 2 identifies those radionuclides that would be most prevalent in fission events. While a
number of these radionuclides are present as single radionuclides (i.e., no radioactive progeny)
several parent-progeny pairs are shown in Table 2 (see Section E for discussion of decay
equilibria). In a fission event, if the parent (listed first) is found in a sample, then the progeny
eventually will be present. Depending on its gamma-ray abundance and the half-lives of the
radionuclides, the progeny may be the radionuclide determined by gamma spectrometry. For
example, 140La (ti/2 =1.7 days) may be readily detected but its parent, 140Ba (ti/2 = 12.8 days),
may not be detected due to low abundance of the 140Ba gamma rays and its much longer half-life
(i.e., lower specific activity). Also, parent-progeny pairs may not be in radiochemical
equilibrium due to differences in the chemical reactivity in the environment, or the time elapsed
since the event. A specific example of such an issue arises with 95Zr/95Nb. While their chemical
similarities may keep them "together" in the environment, it takes about 180 days under stable
chemical conditions for these two radionuclides to approach transient equilibrium.
Radionuclides that are "beta-only" emitters, such as 90Sr/90Y, may have a distinct effect on the
low energy region of the gamma-ray spectrum. The effect is an increase in background counts
due to bremsstrahlung and any photons produced as a result of Coulombic interactions with
bound electrons. The visual result of these interactions can be what appears to be a Compton
edge in the spectrum that does not correspond to any specific gamma ray.
Unless extraordinary circumstances apply, we should not assume to have control over processes
in the environment (at least without stating such in the analytical report). Exceptions could
include, for example, refractory or extremely insoluble radionuclide pairs in particulates/soils
where it is reasonable to assume that there will be no fractionation between their formation in the
event and sample collection. Once the sample is collected, however, it is presumably preserved
and stored under controlled conditions to prevent fractionation from occurring.
15
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Even when it appears that appropriate measures have been taken to properly preserve samples,
the chemistry involved with mixtures of elements present in fresh fission products can be quite
complex. For example, noble gasses, like Xe and Kr will emanate from the sample regardless of
the type of container. Similarly, acidification will oxidize the iodide which can then be
volatilized from solution. An example that demonstrates the importance of sample preservation
is discussed under Figure 32 in Attachment 2.
Table 2. High-Activity Radionuclides Resulting from a Fission Event
Alpha Emitters
Beta/Gamma Emitters
Am-241[1]
Ba-140/La-140
Mo-99/Tc-99m [3]
Ru-103/Rh-103
U-235 [1]
Ce-141
Nd-147/Pm-147 [2]
Ru-106/Rh-106 [3]
U-238 [1]
Ce-143/Pr-143[2]
Eu-155
Sb-125
Pu-238
Ce-144/Pr-144
I-131/Xe-131
Te-132/1-132 [3]
Pu-239 [1]
Cs-134
1-133
Zr-95/Nb-95
Pu-240 [1]
Cs-137 [3]
Np-239
Zr-97/Nb-97
Pu-241
Eu-154
Pm-151/Sm-151[2]
Activation Products
Co-58
Ag-llOm
Cr-51
Mn-54
Np-239
Co-60
Fe-59
Na-24
Notes:
[1] Principally an alpha emitter with low abundance gamma rays
[2] These radionuclide pairs represent examples of "No Equilibrium". For instance, after approximately
70 days the 147Pm activity exceeds the 147Nd activity
[3] Parent is a low abundance or non-gamma emitter; progeny used for quantification by gamma
spectrometry.
The activation products shown in Table 2 are the ones most likely to be associated with fission
events. However, depending upon the specific source of the fission materials and location of the
event other activation products for the first or second row transition series are also possible.
An RDD event may not necessarily have one radionuclide; several different radionuclides and
their progeny may be present. In a fresh fission product event, activation products (such as those
listed in Table 2) also may be present depending upon the type of material involved in the
incident (includes both anthropogenic and natural materials).
An RDD or fission product event will produce environmental samples that have a mixture of
radionuclides that varies depending on how soon after the event the sample is counted and
environmental weathering or physical/chemical effects. These effects will also determine which
radionuclides will dominate the activity (and gamma-ray spectrum) from the sample. Thus, the
library used for analysis should be adjusted for these factors. For example, a sample taken 2 days
following an IND event would be expected to have unsupported, short-lived fission products like
133I (ti/2 = 20.8 hours), 105Rh (ti/2 = 35.4 hours), or 97Zr (ti/2 = 16.7 hours) present, as well as some
with much shorter half-lives than these radionuclides. However, two weeks following the event it
would be unlikely for these radionuclides to be seen based on their relatively short half-lives
compared to the decay time from the event.
16
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
C. Effects of High Activity Concentrations on Interferences and Identification
Samples containing high activity concentrations of radionuclides generally allow easier
identification and quantification (e.g., shorter count times, smaller sample size, more distinct
gamma-ray peaks above background) of those radionuclides. At the same time, however, they
can also create certain problems.
Low abundance gamma rays
Gamma rays with an abundance of less than about 1% are not always listed in the libraries
provided by the software manufacturer. Libraries provided by the manufacturer should never be
used as received. The software manufacturer does not know what the potential applications are
and cannot cover every possibility, but they supply libraries with a convenient framework for the
user so that additions, deletions or edits can be made appropriate to the user's application.
However, when a sample with a very high activity is counted, gamma rays that have relatively
low abundances may be detected. If these low-abundance gamma rays are not in the library, they
will be classified as 'unidentified'. If they are within the energy tolerance (see page 54) they may
also correspond to another radionuclide that is in the library and may be misattributed to that
radionuclide.
For example, Figure 5 compares the entire listing for 132I in the software library from a particular
gamma-ray system11 with part of the listing in the Brookhaven National Laboratory National
Nuclear Data Center (NNDC) listing. In the useful quantification range, the software contains 27
gamma rays (1921 keV and 2002 keV are excluded as they are beyond the usual calibration
energy of 1836 keV used by most laboratories), while the NNDC has 125 gamma rays. The
NNDC contains gamma rays for 132I with abundances as low as 7.9xl0"4%. While it may not be
reasonable or practicable to maintain all low abundance gamma rays in the analytical library
used for radiological response-event analyses, it is advisable to add some of the gamma rays that
were excluded from the analytical library to the suspect library so that the software can provide
accurate indications of the possible source of these gamma rays, when they are present.12
11 Software libraries supplied with gamma-ray detection systems may not be rigidly controlled and may not be based
on a recognized national or international standard reference for nuclear parameters. All libraries should be reviewed
routinely ensuring that they are consistent with NNDC compilations and that undocumented changes have not been
made during day-to-day use of the libraries.
12 Using the master library as the suspect library can lead to misleading suspect identifications. Instead, the suspect
library should only include radionuclides that could actually be present in samples at the time of analysis. It should
also supplement the library used for the analysis of the samples. For example, the main analysis library may only
need a handful of lines to properly identify a radionuclide and resolve interferences with other radionuclides present
in the library. Including a longer list of photopeaks for the radionuclides in the analysis library will ensure that they
are listed as suspects when higher activity samples are analyzed.
17
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Software Library
NNDC
Nuclide
Name
1-132
Half-Life
(Seconds >
8.280E+00 3
Erie rgy
(keV )
262
505
522
S47
621
630
650
667
€69
671
723
772
780
809
812
876
910
954
1034
1136
1143
1173
1290
1295
1372
1398
14 4 2
1921
2002
.700
. 900
.650
. 100
.200
.220
.600
.690
.800
.600
.500
.610
.200
.800
.200
. 800
. 300
.550
.700
.030
.400
.200
.700
.300
.070
.570
.560
.080
.200
Yield
<%)
1.4400
5.0300
16.1000
1.2500
1.5000
13.7000
2.6600
98.7000
4.9000
5.2000
1.1000
76.2000
1.2300
9000
6000
0000
9200
18,1000
4700
9600
3500
0900
1400
1. 9700
4700
1000
4200
1800
0900
Energy
(keV) lntensity(%)
620.9
0.39 %
621.2
1.58%
630.19
13.3%
650.5
2.57%
667.714
98.7%
669.8
4.6%
671.4
3.5 %
684.4
0.039 %
6S7.8
0.039 %
706.4
0.01971
727.0
2.2%
727.2
3.2%
728.4
1.6 %
771.7
0.020 %
772.60
75.6 %
780.0
1.18%
784.4
0.38 %
Figure 5. Comparison of a Software Library (from that used at a nuclear power plant)
and NNDC Database1-' for 132l
When some of these low-abundance gamma rays appear in a high activity sample, if they are not
in the library, they may either be misidentified as a different radionuclide or they go into the
unidentified gamma rays file. When this occurs, the radionuclides associated with these gamma
rays should either be manually identified the unidentified gamma rays compared to a well-
constructed 'suspect' library for the sample type under analysis.
For example, the NNDC database identifies 727.0 (2.2%) and 727.2 keV (3.2%) gamma rays but
the software library does not contain these gamma rays. These could be:
• Misidentified as 212Bi (6.67%);
• Misidentified as 128Sb (4.0%);
• Identified and the peak area used to determine all 2l2Bi, 128Sb and 132I;
• Placed into the unidentified gamma rays report; or
• Potentially identified in a "suspect library."
13 This figure also will be referred to in Section I where mismatch of software nuclear constants and the NNDC are
discussed. Note that not all the gamma-ray energies from the NNDC library have been listed in this figure.
18
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
The peak resolution would likely suffer a little as they are only 0.2 keV apart and it is not likely
that they would be resolved14. Thus, these peaks might be rejected by the software due to high
full width at half maximum (FWHM). It may be possible to resolve these peaks by adjusting the
peak search or Gaussian sensitivity factor for the spectrum. In addition to these two gamma rays
not appearing in the library, there is about a 30% difference in the abundance of the 728.5 keV
gamma ray between the software and NNDC database value.
A critically important aspect of the software analysis to consider is that interference corrections
rely on having a complete list of lines that will be present in a sample and that will interfere with
other radionuclides. This would also include naturally-occurring radionuclides. For example, the
620.9 and 621.2 keV gamma rays for 132I noted in Figure 5 should be included in the library as
they potentially interfere with the analysis of 106Ru (621.9 keV) because both are fission
products.
Gamma-ray interferences not anticipated
The Compton Effect produces an "edge"15 which is seen as a sharp decrease in the slope of the
Compton continuum. If the gamma spectrometry software is not properly tuned to the
performance of the detector, the software may see the Compton edge as a peak. The Compton
edge occurs when Compton scattering occurs within the active volume of the detector and the
scattered photon is undetected. The Compton edge is created by the maximum energy the
scattered electron (A'"'ax) can have based on particle scattering formulas involving conservation
of momentum, energy, and matter. The energy relationship between where the edge occurs and
the gamma ray that creates that edge is derived from these scattering formulas and can be
calculated using the following equations:
0.511 x E
^min y
7 ~ (2xEr +0.511) (1)
/7 _ /7max _ r _ z^min
Compton Edge e~ 7 7 (2)
If we plot the location of the Compton Edge as a function of gamma-ray energy the curve shown
in Figure 6 is obtained.
14 In those instances where there are two gamma rays that will be present and cannot be resolved, it may be
advantageous to use the average energy of the gamma rays and the sum of their abundances as a library entry.
15 See Figure 8 for examples of Compton Edges.
19
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
2.2
2
> L8
| 1.6
<£ 1.4
to
S 1.2
1 1
| 0.8
2 0.6
0.4
0.2
0
0
i l l
i i .
i . i
i i i
! 1 1
1 1 1
3
1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2
Gamma Ray Energy, MeV
Figure 6. Compton Edge Location as a Function of Gamma-ray Energy
This curve can be of value when attempting to assess whether or not a gamma ray identified by
the software is real, an artifact of the Compton edge, or interfered with by the Compton edge.
This makes qualification and quantification more challenging and requires advanced software
tools. For example, a peak at a Compton edge will have a significantly different background
before and after the peak (leading edge & trailing edge) and will affect peak definition as well as
quantification. Figure 8 shows the Compton edges associated with the two primary gamma rays
from 60Co. Note that although they have a significantly different shape than a gamma-ray peak,
the software may identify these as photopeaks.
Another type of interference in the analysis of gamma rays appearing in the energy range of
about 180 to 250 keV is the "backscatter peak". This feature in the gamma-ray spectrum is due to
a gamma ray striking an object other than the detector (usually the detector shielding) via the
Compton Effect and the scattered gamma ray impinges on the detector, interacting via the
photoelectric or Compton Effect. All gamma rays can interact with matter to produce backscatter
radiations, which are not monoergic. Each gamma ray has a minimum energy for this interaction,
but with multiple gamma rays of different energies, there is no exact minimum value for the
backscatter peak. However, the probability of interaction of these backscatter gamma rays within
the detector drops off below about 140 keV (due to the detector housing on a normal p-type
detector), giving the appearance of a 'reverse' Compton edge (See Figure 8).
Single-escape, double-escape and multiple sum peaks
The phenomenon referred to as 'pair production' occurs when a photon with energy greater than
1,022 keV interacts with the electromagnetic field of a nucleus resulting in the creation of an
electron-positron pair. The energy of the photon must be greater than 1,022 keV to preserve
equivalence with the rest masses of the two beta particles of 0.511 keV each. Any energy greater
than 1,022 keV is shared almost equally between the two beta particles as kinetic energy. Figure
7 shows the sequence of particle transformations in this event.
20
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
511 keV (as mass)
V
©
511 keV (as mass)
Pair
/ Production
Photon 1
{511 keV)
Photon 2
(511 keV)
any p*
511 keV (as mass)
Annihilation
Figure 7. The Pair Production Process
Annihilation will occur when the antiparticle (positron) encounters matter inside or outside of the
detector. If the encounter is made within the detector volume then it is a near certainty that the
annihilation also occurs within the detector volume. When this interaction occurs within the
active detector volume, two other effects can lead to features in the gamma-ray spectrum. To
adequately describe these effects, it is important to note that the interaction of the gamma ray
with the detector via the Photoelectric Effect produces an electron which is accelerated and
amplified in the potential bias field across the active detector volume. Any electron produced in
the active detector volume will yield a similar response. Also keep in mind that there is a finite
time constant, t, associated with the collection of charge by the detection system after a detector
interaction occurs (this time constant is very long compared to the interaction). When the initial
pair production occurs, 1,022 keV of the energy of the incident gamma ray is converted to
negatron and positron masses with any energy in excess of 1,022 keV being converted into
kinetic energy for those particles.
1. The negatron interacts with the detector during any normal Photoelectric Effect process; the
negatron interacts with the atoms in the detector creating electron-hole pairs, which under the
influence of electric field generated by the high voltage bias, are collected during the time
frame, t.
2. The positron undergoes a slightly different phenomenon. The positron also produces
electron-hole pairs as it interacts with the atoms of the detector, then as it slows down it
interacts with a single bound electron causing an annihilation event. The electron-hole pairs
created, under the influence of electric field generated by the high voltage bias, are collected
during the time frame, t.
a. At the point of annihilation, the kinetic energy of the positron is essentially zero and its
collision with the electron yields two 511 keV photons, commonly referred to as
annihilation radiation
21
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
b. One or both of the 511 keV photons can interact with the detector through the
Photoelectric Effect producing a photoelectron.
If processes 1, 2, and 2a occur and one of the two 511 keV photons escapes and one of the two
511 keV photons is absorbed by the detector within time frame x, a single energy response is
produced that is 511 keV less than the initial gamma-ray energy. Multiple observations of this
phenomenon are referred to as a "single-escape peak" because the other 511 keV photon has
escaped the active detector volume without photoelectric interactions.
If both 511 keV photons escape detection during time frame x, then a single peak is produced
with 1,022 keV less energy than the original gamma ray. Multiple observations of this
phenomenon are referred to as a "double-escape peak".
A specific example of these events can be seen in Figure 8. The single- and double-escape peaks
are barely visible in this spectrum, but it should be noted that the single escape peak is smaller
than the double-escape peak. This is generally true even in large volume detectors, as the
probability of the same initiating event causing three separate events (the initial pair production,
detection of the negatron produced and the annihilation yielding a Photoelectric Effect) then
creating a single response in time frame, x, is very small.
1.0E0S
DE
10000
SE
Coincidence
Sum Peak
100
Backscatter
Peak
10
Compton
Edges
3.00
1352.00
Energy (keV)
2026 00
Figure 8. HPGe Spectrum of 60Co (the SE and DE peaks are from the 1332 keV gamma ray)
Summing of pulses from gamma-ray interactions in the detector that arrive within the resolving
time of the electronics will occur and be registered as the sum of the energies of the gamma rays.
Summing is generally characterized as random summing or true coincidence (cascade) summing.
Random summing will occur for all photons that simultaneously interact within the detector
giving rise to photoelectrons that are detected and processed by the electronics. A random sum
peak results from two photons, most probably from twice the energy of the highest count rate
gamma ray or from the two highest count rate gamma rays (from different atoms). If both
22
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
photons of energy, Ei and Ei, produce electrons by the Photoelectric Effect in the detector and
the detection system views the total energy (as measured by the pulse size generated) as coming
from a single event, then a count will be stored at the Channel (i.e., energy) where a sum peak
will be recorded. The probability of summing is related to the rate of photon-detector interactions
in time period t and also to pile up pulses in the amplifier. The analog-to-digital conversion
(ADC) dead time can be corrected electronically but does not eliminate the random summing
problem. If the system amplifier has pulse-pile-up rejection capability this will help to reduce
random summing.
Any full-energy photon that is summed with another pulse is not recorded in the single full-
energy photon peak and represents a loss of counts or efficiency in that full-energy peak. This
loss is count-rate dependent and therefore geometry dependent. Moving the sample farther from
the detector reduces the count rate and therefore reduces the effect of random summing.
The net result of either type of summing is a decrease in the full energy count rate of each photon
registered by the analyzer and an apparent loss of efficiency of the detector. Random summing is
related to the number of pulses per unit time processed by the electronics. High activity samples
counted in close proximity to the detector with high counting efficiency may produce count rates
greater than 1000 cps and, subsequently, cause significant losses in the recorded counts in those
gamma rays that are involved in the random summing effect. These losses will be a function of
the energy of the gamma rays and are difficult to predict for complex spectra.
Random summing and dead time effects can be minimized by limiting the count rate to less than
1000 cps by counting smaller samples at greater distances from the detector. Older analog
electronics are more sensitive to random summing effects than the newer digital electronics
which may be able to process much higher count rates with minimal losses from pulse pileup.
Laboratories routinely establish acceptance limits for dead time at levels low enough to ensure
that results will not be adversely impacted (e.g., 5-10%). Although most gamma spectrometry
systems electronically correct for dead time, corrections for high count rates can be determined
empirically to account for losses from pulse pileup. One approach involves counting a mixture of
gamma-ray emitters in a fixed position with total counting rate of less than 1000 cps. The
spectrum is processed to determine the count rate at each gamma-ray energy. Without moving
this source, the count rate is increased by exposing the detector to a second gamma-ray source
containing different gamma-ray energies than the first. The count is repeated several times
adjusting the distance to the detector of the second source to yield higher count rates. The spectra
are analyzed and the count rate of each gamma ray in the mixture plotted as a function of the
overall count rate of the mixture to determine the effect of the increased count rate on each
component of the original mixture.
True coincidence summing or cascade summing (TCS) can occur when a single atom of a
radionuclide in one excited state decays from that state, and then directly to another and the
lifetimes of the two excited states are very short. If the gamma rays in that cascade interact with
the detector simultaneously (i.e., femto- to nanosecond time frame) compared to the resolving
23
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
time of the electronics (microseconds), they will be recorded as one event at the summed energy
of the two coincident gamma rays.16
A generic decay scheme is shown in Figure 9. The result of summing events is a lower count rate
for each of the gamma rays in the cascade since a certain fraction of the events are registered as
sum counts. Fewer counts are recorded for each of the cascade gamma rays (yi and 72 in Figure
9) compared to a gamma ray of the same energy that is not in coincidence with another gamma
ray. This process is referred to as "summing out". If there is a gamma transition (73) from the
same radionuclide that is equal to the energy of two gamma rays in coincidence then the sum
peak (71 + 72) is counted and adds to the total counts recorded for the independent transition. This
results in a higher count rate for the independent gamma ray. This process is termed "summing
in". In order to identify coincidence sum peaks, a decay scheme for each of the radionuclides
identified in the spectrum should be used.
Parent
Yi
N
f
V3
V2
Progeny
Figure 9. Coincidence decay in a beta emitter
In contrast to random summing, true coincidence or cascade summing is not a function of count
rate but is related to the probability that coincident full-energy peaks of two gamma rays will be
detected simultaneously by the detector. The probability of true coincidence summing is also
geometry dependent and can be calibrated for directly and normalized.
This probability also is inversely proportional to the product of the efficiencies of yi and 72. As
the efficiency of detection for each gamma ray increases, the number of TCS events increases
and the loss of counts in each photopeak, yi and 72 increases. For the crossover transition, 73. the
count rate increases with increasing apparent detector efficiency. This effect is highly geometry
dependent (especially with regard to the size and shape of the source, and the juxtaposition of
sample to detector). Any corrections made must be unique to the calibration geometry. Some
16 Gamma rays also sum with X-rays if the two radiations interact with the detector within the resolving time of the
detection system. This effect, called y-x summing, is pronounced in detectors that are sensitive to low-energy
radiations such as X-rays. These include N-type detectors with thin entrance windows. The effect is described by
Robert C. McFarland in Demonstration of Coincidence Summing Effects Obser\>ed with N-Type Germanium
Detectors in the 20- to 40-keVEnergy Range When Counting 1291, 1251, and 125Sb, in The Counting Room: Special
Edition, Radioactivity & Radiochemistry, Vol. 4, No. 2, 1993 (Reference 27).
24
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
radionuclides that are commonly encountered that display coincidence sum effects are shown in
Table 3.
Table 3. Commonly Encountered Radionuclides with True Coincidence Sum Effects
Radionuclide
Gamma 1, keV
Gamma 2, keV
Sum Peak, keV
60
Co
1173.2
1332.5
[i]
2505.7
134
Cs
604
796
1400
154
Eu
58.4
1274
1332.8
123.1
1047.2
1170.3
00
00
1836
898
[2]
2734
Notes:
[1] Non-coincidence abundance is 2.0xl0~6%
[2] Non-coincidence abundance is 7.10xl0_1%
Total coincidence summing can have a significant impact on the quality of gamma-ray
spectrometry measurements. Calculations of radionuclide activity should always use detection
efficiencies that reasonably reflect the response of the detector to gamma rays emitted from the
sample test source. The mixed gamma-ray standards that are used for routine calibrations of
gamma-ray detectors will generally incorporate radionuclides with no or minimal TCS summing
effects17. Relying on routine efficiency determinations for radionuclides with significant TCS
effects will not yield accurate results.
The effects of TCS on efficiencies can be minimized, or corrections determined in several ways
including the following:
• TCS can be minimized by counting samples in a calibrated geometry further from the
detector (e.g., the center of the sample is at least 8-10 cm from the detector center). The
increased distance between the sample and the detector may be the most practical way when
measuring an unknown complex sample, especially for air filters or samples of small mass or
volume. The disadvantage of this strategy is that lower efficiencies may require significantly
longer counting times to reach desired detection sensitivity.
• The most accurate process for determining corrections for TCS involves empirical calibration
using single-nuclide standards for each TCS radionuclide being measured. This approach is
accurate and effective since the summing effects that impact the sample are exactly
reproduced in a calibration standard of the same size and shape as the sample counted in the
17 One of the more commonly used mixtures is described in ANSI N42.14, Section B.6.2 (Reference 16). It contains
109Cd, 57Co, 139Ce, 203Hg, 113Sn, 85Sr, 137Cs, 88Y and 60Co. Often241 Am or 210Pb are used to extend the range of this
mixture. This mixture of radionuclides will result in a low bias for efficiency of monoergic gamma-emitting
radionuclides in the higher energy range (e.g., Zn-65, Na-22, and Na-24).
Section B.6.3 of ANSI N42.14 recommends substituting 65Zn and 54Mn for 88Y and 60Co when counting closer to the
detector than about 5 cm as is common practice for environmental samples. It may be noted that for close-in
environmental geometries, when using 60Co and 88Y the discrepancy due to summing is generally less than about 5%
(relative).
25
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
same juxtaposition to the detector. Unfortunately, this methodology may not be practical.
Spectra containing many radionuclides will require having traceable sources of each of the
radionuclides for which TCS is of concern.
• A third strategy is to utilize gamma-ray software that corrects for TCS effects. This option
can be effective but requires special efficiency calibrations to determine the total efficiency
(as opposed to photopeak efficiency) of detectors as a function of gamma-ray energy, and
may require the use of dedicated libraries that are not readily customizable. Software
developed in recent years may or may not require special calibrations. The user must
determine which version of the software is available on their system prior to making
assumptions about how coincidence summing is affected. Such corrections are also subject to
interferences in complex spectra that contain interfering lines such as is common with
complicated mixtures of radionuclides.
• Finally, manual calculations can be performed to determine TCS correction factors by
utilizing the methodology outlined by K. Debertin and R.G. Helmer18, or Menno Blaauw 19.
This methodology also requires the peak-to-total efficiency of the detector be determined for
the gamma-ray energies being evaluated. This technique may not be practical for complex
spectra with many radionuclides with TCS.
With random summing, an apparent gamma ray will be observed in the spectrum that cannot be
attributed to the decay of any radionuclide and also cannot be accounted for by TCS. The
random sum peak may be identified by doubling the energy of the largest peak area or adding the
energies for the two largest peak areas in the gamma-ray spectrum below the energy of the
suspected sum peak. It is possible that in samples with multiple high activity radionuclides that
random sum peaks may be possible from three or more radionuclide gamma-ray emissions.
While this is a less frequent event, it can still produce a small spurious peak in the spectrum.
Another feature of the random sum peak is that it cannot have a Compton edge associated with it
since the random sum peak energy originates within the detector as the result of a simultaneous
electron production from other photon events and thus there is no way to form a Compton Effect
peak.
In Figure 8, the 2505.7 keV gamma ray in the 60Co spectrum is an example of a coincidence sum
peak. Although the 2505.7 keV energy level goes directly to ground state its abundance is only
about 2.5x10"6%. Thus, it is likely that it contributes negligibly to the observed Compton edge
and that the major contribution comes from Compton interactions of the two other gamma rays
occurring within the detectors time constant.
Two other examples of sum peaks are:
• In reactor coolant spectra a peak will be seen at 1321 keV from the random sum of the 810
keV 58Co gamma ray and the 511 keV annihilation peak.
• In a fresh fission product spectrum, a peak will be seen at 1400 keV from the coincidence
sum of the 604 and 796 keV gamma rays of 134Cs.
18 See Reference 20
19 "The Use of Sources Emitting Gamma-rays for Determination of Absolute Efficiency Curves of Highly Efficient
Ge detectors", NIM A322, 1993, pp. 483-500 (Reference 28)
26
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
D. Decay Correction and Count Time
Ensuring correct half-life
Many references, including vendor-supplied software nuclide libraries, provide what they deem
to be accepted values for the half-life of radionuclides. Inspection of these references quickly
shows, however, that the half-lives vary from reference to reference as a result of the
measurements selected for the evaluation. It is important to verify that the half-lives (and other
nuclear data) recorded in the gamma spectrometry software library are in good agreement with
half-lives identified in nationally- or internationally-recognized standards20. If this process is
performed with the "master library" used by the software, other libraries created from the master
should have reliable data.
In cases where it may be assumed that equilibrium has been fully established since the point of
sample collection, the half-life of the progeny may be modified in the library to allow the
software to use the half-life of the parent to make appropriate decay corrections. In those cases
where it is indeterminate as to whether equilibrium is established, decay correction must be made
cautiously (sometimes requiring calculations outside the gamma spectrometry software
capabilities). Non-routine calculations should be disclosed in the laboratory's final report.
Decay corrections are complicated when radionuclides are members of a decay chain and are
present in radioactive equilibria with a progenitor (parent). Therefore, when setting up libraries,
half-lives used should reflect whether or not the radionuclide is part of a radioactive decay chain
and whether the software can be used to make a decay correction. The decisions on how to
establish these factors in the library should be part of the laboratory's Quality Manual for
radiochemistry. This topic will be discussed in more detail in Section L.
Correction for decay during counting
Long-lived radionuclides (that decay directly to stable nuclides) like 137Cs, 110mAg or 54Mn will
have a negligible activity change (i.e., negligible loss of atoms) over the course of a few hours
between the start and finish of the counting interval. However, for many radionuclides that may
be present following a radiological event, half-lives may be short, and the counting intervals used
will be short (usually reflecting the urgency of needed results).
When the counting interval is greater than 10% of the radionuclide half-life a correction factor
for decay during counting (DDC) should be applied21. If the half-life of the progeny is used
rather than the half-life of the progenitor, a significant error in activity concentration of the
progeny could occur, and the use of the progeny half-life to perform DDC should be noted in the
case narrative.
A general formula implemented in most gamma spectrometry software for this correction is:
20 See References 12 or 13.
21 For example, if a count time of 1,000 seconds is used, the DDC factor should be applied if the radionuclide half-
life is less than 9,000 seconds.
27
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
cf = <3>
Where:
Cf is the correction factor (DDC, a dimensionless quantity)
X is the decay constant for a particular radionuclide (s"1)
tc is the clock time of the analysis (s)
This equation corrects activity to the start of the counting interval. Figure 10 has plots of the
magnitude of the correction factor as a function of half-life in days for three counting intervals in
seconds (the plots assume zero dead time). Two different examples of how this correction factor
functions are noted here:
• If a sample containing 134I (ti/z = 52.6 min) was counted for 1 hour, the value for Cf would be
-1.45. Not applying this correction factor would result in the activity measured being -45%
low.
• On the other hand, if a sample that contained l32Te (ti/2 = 3.2 d) in equilibrium with its
progeny 132I (tin = 2.3 hours) was counted for 5400 seconds (1.5 hours), and the 1321 half-life
was used to calculate the 132I activity, the calculation would be biased high by 23% (a Cf
multiplicative factor on the activity of 1.23 - See inset in Figure 10). Since the halt-life of the
progenitor is 3.2 days, the correction factor for DDC is negligible for count time of 5400
seconds (1.5 hours) or less.
4.50
1.50
cf, 1000 s
Cf, 3600 s
Cf, 5400S
1.40
4.00
1.30
1.20
3.50
1.10
« 3.00
1.00
l.E-02
l.E-Ol
Half life In days
5 2.50
2.00
1.50
1.00
0.50
l.E-02 l.E-01 1.E+00 l.E+01 1.E+Q2 l.E+03 l.E+04 l.E+05
Half life in days
Figure 10. Correction Factor for Decay during Counting (DDC); Zero Dead Time
For gamma-ray spectrometry with no chemical separations performed, the software is able to
correct the activity concentration of the sample to the start time of the analysis te in Figure 13
(i.e., it accounts for the decay of any radionuclide detected during the counting interval; Figure
28
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
10). Thus, an activity concentration can be positively associated with the start time of the
counting interval22.
Adjustments to Decay Corrections Due to High Sample Activity (Correction for Dead
Time)
When a sample is counted and there is measurable dead time, two-time intervals are of concern.
The first is the "preset" or "live" time for the sample. This is the time period that is preselected
by the user and represents only that time that the detection system is actively recording counts
from the sample. Other preselect functions can be used in place of "live" time, for example
counting to a preselected uncertainty or counting to achieve a particular minimum detectable
activity (MDA). The second is the "real"23 time which is the live time plus the dead time. Since
sample decay occurs relative to real time, that time should be used to calculate the correction
factor for decays during the counting interval.
The software used in some gamma spectrometry systems uses the live time for Cf and not the
real time. Figure 11 shows the difference between using the live time (blue curve) and the real
time (red curve) when calculating this correction factor. The green line shows how Cf changes as
a function of dead time. These curves were generated for a radionuclide with a half-life of 15
minutes (900 seconds), a live time count interval that is 67% of the half-life and count dead times
up to 13.9%. At the highest dead time for this example, the activity bias will be negative 3%. The
negative bias will be most significant when the radionuclide half-life approaches the sample live
times and where there is a dead time larger than about 10%.
1.29
1.285
1.28
1.275
1.27
u
1.265
1.26
1.255
1.25
1.245
580 600 620 640 660 680 700
time, seconds
Figure 11. Correction Factor for Decay during Counting (DDC); Non-Zero Dead Time
(Assumed half-life of 900 s, live time is 600 s)
22 This function is the decay during counting (DDC) correction, and is an option that can be selected by the user.
23 "Real" time may also be referenced in other documents as "clock" time.
Cf, Live time
Cf, Clock time
%DT
29
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Use of the correct half-life for a progeny radionuclide when it is present at any time with its
parent can lead to significant bias in the activity concentration unless the proper formulations for
time decay are used for analysis. A separate issue is whether or not the library contains the
correct half-life of the radionuclides; these should be checked against a national or international
standards organization.
Therefore, half-lives used in the gamma spectrometry libraries should both be verified to be
correct versus an accepted standard (see Reference 19 or 20) while reflecting whether or not the
radionuclide is part of a radioactive decay chain. If part of a decay chain and equilibrium is
established a half-life of a progenitor may be more appropriate to use in calculations (See
Section L). These actions should be documented in the laboratory's Quality Manual and in the
laboratory report to the client since the assumption may impact the use of the data.
The second portion of the significance of correct half-life deals with decay corrections that may
need to be made back to a specific time or the time of the event to determine dose consequences
ex-post facto.
E. Decay Correction and Radioactive Equilibrium
Three types of radioactive equilibria exist: secular equilibrium, transient
equilibrium, and no equilibrium.
"Secular equilibrium" occurs when the parent half-life is much greater than that of its progeny.
In general, once secular equilibrium has been established the activity of the progeny will equal
that of the parent. Although there is no specific ratio of parent-to-progeny half-lives that defines
when this relationship begins to be a secular equilibrium if the ratio24 is greater than about 50 the
difference in parent and progeny activities at equilibrium will be less than a factor 1.05 (a
difference of less than 5%). Once secular equilibrium is achieved it is a useful approximation to
say that the activities are "equal". An example of a radionuclide and its progeny displaying
secular equilibrium is shown in Figure 1225.
24 Other references provide different values for this factor. There is not a firm definition of what factor constitutes
that an equilibrium is transient or secular.
25 Note that the activity of the progeny 137mBa does not equal the 137Cs activity due to branching to the 137Ba ground
state.
30
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
40
1.1
0.7.2
5-20
137CS
10
137mBa
Ratio 137mBa/137Cs
600
200
400
800
1000
1200
Seconds
Figure 12. Secular Equilibrium Displayed by 137Cs-137mBa
"Transient equilibrium" occurs when the parent half-life is between 1 and approximately 50
times that of the progeny (See Figure 14). In general, the activity of the progeny will exceed that
of the parent once equilibrium is achieved. In these two cases, once equilibrium is established the
progeny decay with the half-life of the parent as long as there are no physical or chemical effects
that would separate the two radionuclides. If the activities of parent and progeny are plotted,
their curves will be parallel once equilibrium is established.
In the case of "wo equilibrium", the half-life of the parent is shorter than that of the progeny and
there is no point at which the activity curves of the two radionuclides are parallel (See Figure
15).
Decay correction during counting, to a sample collection time or to event initiation can be
problematic if the radionuclides are part of a parent-progeny relationship. Performing decay
corrections using the "true" half-life (e.g., as listed in the NNDC database), assumes that the
radionuclide is unsupported.26 Using the parent half-life to decay correct progeny activity prior to
equilibrium being established, however, can lead to bias in the decay correction unless the proper
formulations for decay are used for analysis.
A second issue that needs to be addressed involves using the correct half-life to make decay
corrections to a specific time, such as sample collection or the time of the event, to determine
dose consequences ex-post facto.
26 Supported" refers to the dynamic situation when a parent radionuclide is present and produces new atoms of the
progeny radionuclide, thereby, "supporting" its activity. The activity of the parent of a decay chain is described as
being "unsupported" when no new atoms are produced and the radionuclide will decay with its "true" half-life as
listed in a nuclear reference such as the NNDC database.
31
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Depending on many factors, including the half-lives of parent and progeny, environmental
conditions, or chemical effects, radionuclides that are part of radioactive equilibria may appear to
decay with the half-lives of their progenitors (transient and secular equilibria). Long-lived
progeny of short-lived parent radionuclides may not appear in the analysis until much after the
event has occurred (No equilibrium). Since the value for activity concentration is reported at a
specific date and time and since activity changes as a function of time, corrections for decay
must use the half-life of the radionuclide based on sample-specific radionuclide interactions, and
the sample history. When radionuclides are present following a fresh fission event where very
short-lived radionuclides are present in parent-progeny relationships (like 132Te/132I or
99Mo/99mTc,), the correct half-life will likely be that of the parent and not the half-life of the
progeny.
However, even though the decay that occurs between sampling and analysis time (t3+t4+ts in
Figure 13) may be significant, a decay correction for that interval may be performed only if the
proper sample preservation occurred at the time of collection. Lacking proper sample
preservation, the physical and chemical stability of parent and/or progeny radionuclides may be
in question and assumptions underlying radiochemical equilibrium untenable. If the sample is
successfully preserved, it is possible to determine the concentration of a parent radionuclide back
to the point of sample collection using the straightforward decay correction equation (see
Equation 4) implemented by most gamma spectrometry software packages.
Progeny radionuclides that are in radiochemical equilibrium relationships will need to have
careful consideration of how long the decay interval (t3+t4+ts) is compared to the half-life of the
progeny27. If the interval exceeds the half-life of the progeny by a factor of about 5-10 (i.e.,
equilibrium is established), then the DDC should use the parent half-life. Figure 14 and Figure
15 are examples of radioactive equilibrium relationships. The point of equilibrium is shown as
"E" on Figure 14. For decay correction prior to the "E" point, Equation 5 should be used. All
assumptions made must be stated with results appropriately qualified and included in the
analytics report (e.g. the case narrative).
The timeline for sample radionuclide decay from the start of the event to gamma-ray analysis is
shown in Figure 13.
Performing decay corrections from the time of analysis backward in time may be necessary to
assess potential doses to the public during the exposure period between the beginning of
sampling (t2) 28 and the count (te). Assuming no sample preparation/separation is performed on
the sample and no chemical effects have occurred during the sampling/transport, the interval
times (t2 to te) in Figure 13 may be used to make decay corrections.
27 A special case exists for decay during air particulate filter sampling. It is normally assumed that a constant
concentration occurs during sampling.
28 Decay correction prior to the onset of sampling and the time of the event should not be attempted because
environmental conditions can unpredictably separate parent and progeny radionuclides, due to physical or chemical
effects (See Attachment II for examples)
32
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Start of
sampling
interval
Sample
analysis
starts
Transport of
sample to
laboratory
Decay
Reference
Date (time
of the
event)
End of
sampling
interval
Laboratory
handling
until
analysis
Sample
analysis
completed
Figure 13. Timeline from Sampling to Analysis
For the specific examples of 132Te/132I and "Mo/"mTc, the 'as-received' software libraries
usually identify the unsupported progeny half-life (132I = 2.3 hours, "mTc = 6.01 hours). A
sample that is analyzed 16 hours or 36 hours, respectively, after collection will have established
transient equilibrium for these two short-lived radionuclides. Once transient equilibrium is
established, the activity of the progeny will decay with a half-life equal to that of the parent. This
relationship is shown in Figure 14 for 132Te/132I and "Mo/"mTc. For the specific examples
shown here, decay correction from the start of analysis backward to a time after which transient
equilibrium is established simply use the decay formula
Ax = A0e"At (4)
Where:
X is the decay constant for the parent
Ao is the activity at time zero (e.g., sample collection date)
Ai is the decay corrected activity (after a time t has elapsed), and
t is the time difference between the start of analysis and the
decay correction point.
33
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
l.E+04
Te-132,dps
9.E+03
1-132, dps
2.5
8.E+03
l-132/Te-132
7.E+03
6.E+03
(/i
O.
13
5.E+03
1.5
4.E+03
3.E+03
2.E+03
l.E+03
0.5
0
10
20
30
40
50
60
Hours
l.E+04
5.E+00
5.E+00
l.E+04
Tc-99m, dps
4.E+00
— Ratio Tc-99m/Mo-99
4.E+00
8.E+03
3.E+00
a.
u
3.E+00 '€
6.E+03
2.E+00
4.E+03
2.E+00
1.E+00
2.E+03
5.E-01
l.E+02
0.E+00
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48
hours
Note: Even though after "E" transient equilibrium has been established, the progeny activity is less
than the parent activity due to split decay modes of the parent.
Figure 14. Transient Equilibrium for 132Te/132l and "Mo/99mTc
In order to decay correct before transient equilibrium was established (See point E in Figure 14),
a different methodology is used. Knowing the parent activity decay correction of the progeny
activity back to the time prior to equilibrium being established (e.g., the sample time) can be
calculated using Equation 5.
34
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
a2n2 = a2 = (XjNj)
{e-*lAt_e-A2At}
02-*l)
(5)
Where:
Xi is the decay constant of the parent
X2 is the decay constant of the progeny
>uNi° (Ai°) is the activity of parent at the initiating time sample
collection
X2N2 (A2) is the activity of progeny at the end of time interval t
N2 is the number of atoms of progeny
At is the time between the event and the time to which the
progeny activity concentration is being corrected
Decay correction for these equilibria prior to the sample collection time is a complicated task and
would require a great deal of information about environmental conditions, the time of the event
and the particular chemistry of the radionuclides with the environmental matrix.
Equation 4 and Equation 5 can be used for calculations involving transient and secular equilibria.
These equations should be used with caution especially when it is not firmly established if the
"E" point in the above graphs has been passed29.
For radionuclide pairs that form "no-equilibrium" conditions a different equation may be used as
shown in Equation 6:
An example of this condition is shown in Figure 15 for 143Ce/143Pr (33.04 h/13.57 d). Here the
activity ratio of progeny to parent as a function of time never reaches plateau.
29 Point "E" in these graphs (and any others with these types of equilibria) is usually between 5 and 10 progeny half-
lives after the onset of ingrowth of progeny. The position of "E" depends upon the half-life differences. The
approach to equilibrium is exponential so there is no specific definition that states when this should occur. Generally
speaking it is convenient to use the point when the ratio of the parent to progeny does not change by more than
about 5%.
(6)
Where:
A\ is the activity of progeny at the time to which the activity is
corrected
A\ is the activity of the parent at the time of sample collection
A\ is the activity of the progeny at the time of sample collection
35
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
l.E+09
l.E+08
l.E+07
l.E+06
l.E+05
l.E+04
l.E+03 o
l.E+02 5
l.E+01
1.E+Q0
l.E-01
l.E-02
l.E-03
l.E-04
0 10 20 days 30 40 50
Figure 15. No-Equilibrium Case for 143Ce/143Pr
Within 5 days of an event, the activity of the parent is less than that of its progeny (assuming no
chemi cal effects di sturb the chemical equilibrium). If an estimate of the activity concentration for
' l3Ce were needed 1 day after the incident using a sample that was procured 50 days after the
incident (when l43Ce would likely not be detected) Equation 6 could be used to make that
calculation (assuming no chemical effects occurred). Generally speaking, decay correction back
to the time of sampling is only possible if the sample has been properly preserved. In this
equilibrium as well, decay correction prior to sampling is very tenuous as the effects of
environmental weathering, etc., would need to be taken into account and these are not easily
predicted for most radionuclides.
There are several important radionuclide equilibrium pairs that immediately follow a fission
event (like an END or nuclear plant accident) that are worth noting. These radionuclides and
several of the naturally-occurring radionuclide equilibrium relationships (which occur in
environmental samples or could result from an RDD) are shown in Table 4 along with their
decay constant values and the time to achieve peak progeny activity. The equation used to
calculate the time to peak progeny activity is:
_ (JhA1-2?iA2) /-7X
tlTtiepgak progeny activity ~ (A1-A2)
The time calculated in Equation 7 assumes that there were no progeny atoms present initially
(i.e., radiochemically pure parent). This equation holds true for the case of No Equilibrium as
well, since there is a time when the ingrowth of the progeny reaches a maximum. It should also
be noted that the activity ratio of progeny/parent is calculated using Equation 5 and will be at a
time that is different than the time to peak progeny activity.
l.E+02
l.E+01
1.E+00
K l.E-01
sl l.E-02
l.E-03
l.E-04
l.E-05
Ratio Pr-143/Ce-143
36
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Table 4. Time for Radionuclide Pairs to Achieve Maximum Progeny Activity
Radionuclide
Pair
X Parent
(Days1)
X Progeny
(Days1)
Time to
Peak
Progeny
Activity^
(Days1)
Type of
Equilibrium
Activity Ratio
Progeny/Parent
Post Equilibrium!111
Fission Products
95'7 /9SKTU
Zr/ Nb
1.08x10 2
1.98x10 2
6.73xl0+1
Transient
2.2
99 99m ri_ 1
Mo/ Tc[b]
0.252
2.77x10°
9.52X101
Transient
0.96
140 140
Ba/ La
5.44x10 2
4.13x10 1
5.65x10°
Transient
1.15
106 106
Ru/ Rh
1.87x10 3
2.00xl0+3
6.94xlO"3
Secular
1
132 132
Te/ I
2.17xl0_1
7.30x10°
4.96X101
Transient
1.03
131 131m
1/ Xe
8.64x10 2
5.82x10 2
1.40xl0+1
No
N/A
137 137
Cs/ Ba
6.31x10 5
3.91xl0+2
4.0xl02
Secular
1
147 147
Nd/ Pm
6.31x10 2
7.23 xio"4
7.16xl0+1
No
N/A
143 143
Ce/ Pr
5.03x10 1
5.11x10 2
5.06x10°
No
N/A
Naturally-occurring Radionuclides
238U/234Th
4.25xl0"13
2.88xl0"2
8.66xl0+2
Secular
1
228Ra/228Ac
3.29xlO"4
2.58x10°
3.48x10°
Secular
1
228Ra/(228Ac)/228Th
3.29xl0"4
9.92xl0"4
1.66xl0+3
Transient
1.4
226Ra/222Rn
1.19x10~6
1.81X101
6.59xl0+1
Secular
1
214Pb/214Bi
3.70xl0+1
5.01xl0+1
2.31xl0"2
Transient
3.8
212Pb/212Bi
1.56x10°
1.66X101
1.57xl0_1
Transient
1.1
210Pb/210Bi
8.51xl0"5
1.38X101
5.36xl0+1
Secular
1
Notes:
[a] Assumes no physical or chemical effects that disturb the equilibrium.
99 99
[b] The branching factor for the decay of Mo direct to Tc, of 12%, is taken from Table of
Radionuclides, ISBN 2 7272 0201 6, Bureau National De Metrologie (France 1999). Calculations
were performed using the Bateman equation for the parent-progeny relationship between
99m
radiochemical^ pure parent and daughter as direct fissions to Tc would not contribute to the
measurements greater than 1 day after the irradiation.
[c] The time to peak progeny activity should be used with caution See Footnotes 25 and 26.
When radionuclide Secular or Transient equilibrium relationships are determined to be present
after counting a sample, decay corrections for the progeny can be made by the software to a
previous point in time, if that point in time is after the equilibrium has been established
(Equation 4 would be used). If that is done, care should be used to ensure that the parent half-life
is used for that correction (i.e., the library for the progeny half-life matches that of the parent).
Equation 5 can be used to calculate the activity of parent or progeny after the activity of a parent
or progeny has been determined, but it assumes that there was no progeny present when the
37
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
progeny ingrowth started. This equation may be selectable in some software packages (check the
specifics of the technical information in the software manual).
To decay correct for parent or progeny in No Equilibrium relationships, or anytime the decay
correction point in time precedes equilibrium for Secular and Transient Equilibrium, Equation 6
would be used: this is generally not available in the software (i.e., a calculation 'outside' the
software using the activities determined at the start of the count time would be input into
Equation 6 to decay correct).
Parent-Progeny Equilibrium not Established
It is important to know the time that an event occurred and to a certain extent the type of material
that was used when fission products are involved since there are several different parent-progeny
equilibria that can be present. Some are so rapid to reach equilibrium (as in the case of
106Ru/106Rh, a little more than 4 minutes) that by the time the samples reach the laboratory
equilibrium can be reasonably assured30. Others (like the 143Ce/143Pr in Figure 15), are "no-
equilibrium" conditions where the parent half-life is short and radionuclide progeny appear many
days after the event or during remediation, without parent activity ever being measured.
There is also the concern that parent-progeny equilibrium has not yet been established;
specifically earlier in time before point "E" on Figure 14. At these times on the ingrowth curve,
the activity of the progeny will first increase then decrease with the parent half-life. Therefore
when attempting to make decay corrections from the beginning of the analysis start time to any
other time, past or future, it is imperative that the event and sample history are well established.
Parent -Progeny Equilibrium Disturbed or Created by Chemical Effects
Gamma spectrometry has the unique, purported advantage in radiochemical analysis that
chemical separations typically are not needed when analyzing for gamma-ray emitters. Sample
preservation of aqueous samples beyond acidification is often not considered important since
many of the radionuclides of interest are best preserved at pH of less than 2.0.
However, some examples of radionuclides in specific matrices requiring unique preservation
techniques are noted here:
• Radioiodines in aqueous matrices. Acidification will cause oxidation of iodine to I2 which
can cause loss due to volatilization or container adsorption. Samples suspected of containing
radioiodines would need two samples: one should be preserved via acidification processes
and one should be preserved with a slightly basic solution containing a reducing agent such
as thiosulfate.
• Ra-226 in solids like concrete or soil. Many analyses rely on the radioactive equilibrium
yielding 214Bi (after an appropriate in-growth period) whose 609 keV gamma ray is more
abundant than that of the 186 keV line of 226Ra. However for the decay chain to attain
equilibrium between 226Ra and 214Bi, the loss of 222Rn during sampling, transport, and
30 However, this assumes proper preservation of the sample.
38
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
analysis must be prevented. Thus a technique which will uniquely (for each matrix) fix the
emanated radon in place is necessary. One such technique for soil is to immerse the soil in
sufficient mineral oil to just cover the top surface of the sample test source (STS) container.
Radon is soluble in the mineral oil (preventing loss through emanation/volatilization) and
will provide a more representative geometry for the progeny when equilibrium is established
after approximately 21 days.
• Charcoal and silver zeolite canisters. These collection matrices are used to collect
radioiodines or noble gases during air sampling. The canisters are usually preceded in the
flow path by a small pore size particulate filter (typically about 1 |im) to eliminate
particulates from interfering with the analysis of the iodines or noble gases when analyzed by
gamma spectrometry. Depending on the manufacturer and style of the canister the depth of
active absorbent is about 4-6 cm (Reference 22). A simple way to reduce the appearance of
inhomogeneity of the sample deposition in the canister is to spin the canister just above the
detector.
For the case of the charcoal absorber, it is usually assumed that the adsorption of iodine
occurs in a very thin layer (i.e., face-loaded) on the inlet side. On the surface, this is a valid
assumption since the number of atoms is very, very small compared to the active adsorption
sites available. However, there are other materials in the air sample that can also be adsorbed
on these surfaces or may affect adsorption. High humidity or presence of organic vapors can
reduce the adsorption or cause it to be more uniformly distributed (i.e., fully loaded) in the
canister volume. Thus, it is important that the sampling method ensures that the radionuclide
distribution on the charcoal will match the loading of the standard that is used for calibration.
Additionally, since both iodines and noble gases are volatile, storage and transport of the
canisters prior to analysis should ensure that they are not exposed to unnecessarily high heat
environments as this may cause loss of these radionuclides.
If the samples are collected over an extended time period (several days) or analysis is
delayed for several days, the ingrowth of several progenies (e.g., 140Ba, 137Cs, and 88Rb) may
occur and be observed on the filter and charcoal/zeolite canisters. It may be necessary to
consider adding some of these gamma emitters to the specific protocol library so that the
lines are easily identified by the software.
If 135mXe is present a short purge with an inert gas (like nitrogen or argon) can remove this
radionuclide and its 526 keV gamma ray which yields a Compton edge at -360 keV.
Removal of that radionuclide from the sample will allow for a better baseline for the 365 keV
gamma ray of131I.
• Aqueous samples with fission products and transuranic elements (TRUs). Analysis of an
aqueous sample containing radionuclides like Zr, Nb and some TRUs can prove challenging.
Most fission products are best stabilized by bringing the solution pH to less than 2.0. The
aqueous chemistry of zirconium is such that it is best stabilized with fluoride ion. However
fluoride ion can cause other elements to precipitate (e.g., both uranium and plutonium if they
are in the +4 (IV) oxidation state). An example of this problem is shown in Attachment II,
Figure 35. Fission products and TRUs will provide examples of preservation needs that can
be unique.
39
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Incorrect Gamma Ray Used to Analyze a Radionuclide
Many radionuclides have gamma-ray energies that are so close in energy to one another that the
software may misattribute the gamma-ray activity to one or more radionuclides in the library
even if only one gamma-ray emitter is present. Some software will even calculate an activity for
each of the radionuclides using the same peak area more than one time. An example of this
situation is seen in a sample of lettuce taken from an environmental sampling program following
the Fukushima event (see Figure 29) where a 177 keV with a peak area of 148 counts is used to
calculate an activity for both 136Cs and misidentified 251Cf).
Part of the concern here is that the software has not been properly configured to perform
interference corrections. This problem can be exacerbated when too many radionuclides are
included in the library that cannot be in the sample, or if interfering photopeaks from a
radionuclide are not included in the library.
F. Software Preset Functions
The software available to gamma spectrometry practitioners has many user selectable preset
functions. Each manufacturer uses its own terminology to describe these functions, and
functionality may differ from software package to software package. Additionally, the logic and
order for certain steps may vary from manufacturer to manufacturer, or even using options
available within a given manufacturer's software package. Thus only generic descriptions of
typical software functions are identified here. It is the individual user's responsibility to
understand how the particular algorithms implemented in their software function and perhaps
even more importantly, how they can, under circumstances, work at cross-purposes with one
another.
Peak search sensitivity
This is generally part of proprietary algorithms developed by software manufacturers to
differentiate between peaks and noise in a spectrum. The specific technique employed by the
software should be described in the software manual, but generally consists of a comparison
between gross observations (counts) ascribed to the peak area and counts ascribed to the
background continuum. The sensitivity specified for peak analysis is generally based on the
relative sample activity.
Once the software locates a peak and fits the spectral data or establishes a tentative region of
interest, the software may perform a test (if elected by the user) to determine whether the peak
should be considered to be present in the spectrum. One commonly used technique is the critical
level test:
Lc = kVB + AB2 (8)
Where:
k is the confidence level (1.645 if 95% confidence level is
used),
B is counts in the Compton continuum under the "peak", and
AB is the 1-sigma uncertainty in B.
40
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
1/2 1/2
(For Poisson statistics AB = B , and Lc = 2.33xB )
If the net peak area is greater than Lc then the peak is considered to be identified and will be used
in subsequent calculations. Note that there are many forms of peak acceptance or rejection
criteria employed by software packages, some of which include user-adjustable parameters for
the acceptance criteria; it is incumbent upon the software user to understand the implications of
these parameters.
The option to adjust peak search sensitivity or peak rejection criteria allows the sensitivity of the
analysis to be matched to the activity in the sample being analyzed. For high activity samples,
where peaks will be relatively easy to locate, a low sensitivity will minimize the false
identification of numerous smaller peaks that may not be of concern to the analysis. In contrast,
for low activity samples, a higher sensitivity is needed to reliably distinguish small peaks from
the background.
It is the job of the gamma spectrometrist to ensure that the peak search sensitivity is properly
selected to ensure that all significant peaks in the spectrum are identified. For example,
specifying too low a sensitivity for an environmental sample may allow peaks actually present in
the spectrum to go unidentified. This is problematic since when a peak is not identified, it may
not show up in the unidentified peaks report, and review of the data output will fail to show
evidence that the radionuclide is present in the sample. The best way to ensure that a spectrum
has been analyzed with sufficient sensitivity involves visually inspecting the spectrum to ensure
that all peaks have been identified.
The spectrometrist must also be careful that the peak search sensitivity is properly coordinated
with other peak identification tests to ensure that they are not working at cross-purposes with one
another. This is dependent on the specifics of the software used.
Peak uncertainty cutoff
This option allows users to specify which peaks should be used in the calculation of sample
results based on their relative uncertainty. If the peak uncertainty cutoff is specified to be 100%
(e.g., at two standard deviations), any peaks with a one-sigma uncertainty of less than 50% will
be included in for subsequent calculations, whereas any peaks with uncertainty greater than 50%
will be ignored.
While the peak uncertainty cutoff is a powerful test, if not used properly, it can work at cross-
purposes with other peak identification tests such as peak search sensitivity. For example, for
some applications, it may be acceptable to ignore peaks with high uncertainties (e.g., > 100%).
Assume that high peak search sensitivity was selected in the hope of identifying low activity
radionuclides in the sample. Simultaneously specifying a two-sigma peak uncertainty cutoff of
less than 30% would be more restrictive than the peak identification test described for peak
search sensitivity above (which is based on detectability) and could result in a failure to identify
detectable peaks and include them in subsequent analysis of radionuclides in the sample.
While some software packages protect the user against such errors, others do not. In any case, it
is always the gamma spectrometrist's responsibility to understand how the software works, to
41
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
ensure that it is configured properly, and that review steps will identify issues that could
compromise the proper identification and quantification of radionuclides in the sample.
Alternatively, the peak uncertainty cutoff test can be disabled and the user may instead use the
critical level function to address detection based only on the critical level. The user should ensure
that this is satisfactory for both the client and any regulations that are involved in the analysis.
Energy comparison
This function allows the user to select the delta energy range that will be used to conclude that a
found gamma-ray peak matches the energy listed for a radionuclide photopeak in the library. The
energy range is either specified as a fixed energy (keV) value or is based on the calibration of
FWHM versus energy (where the match width increases as energy increases). While specific
values will depend on the detection system in question, examples of each would be an energy
tolerance of ± 0.5 keV or an energy tolerance of 0.6 x FWHM, respectively. An overly wide
tolerance (for example ± 2 keV) will result in the inclusion of many incorrectly identified
isotopes but may compensate for a less-than-ideal energy calibration (not recommended). An
overly narrow tolerance (± 0.25 keV) may miss isotopes and lead to many 'unidentified lines'.
Ensure proper energy calibration periodicity and monitor energy calibration control charts to
assure that settings used are in accord with the observed stability of the instrument. Also be
aware of vendor-recommendations for settings. For example, ORTEC specifies that the same
setting be used for samples as was used for the calibration, and recommends a default setting of
± 0.5 keV with values ranging between ± 0.45 and 0.7 keV. In addition to having a good
understanding of default software requirements, users should maintain good environmental
controls in the lab because temperature variations (as small as 5-6 °F) and humidity changes (on
the order of 15%) can affect the electronics and detector response.
Half-life period exceeded
If the time period between the time of sampling and the start time of analysis exceeds a
predetermined number of half-lives (based on the specific radionuclide half-life) then this
radionuclide may be rejected as a candidate since its activity is likely too low to be determined.
For example, a sample is analyzed one week after a radionuclide with a half-life of 2 hours, is
suspected of being present. The radionuclide would have gone through:
1 weekx (168 hour/week)/(2 hours/half-life) = 89 half-lives.
Its original activity would have been decreased by a factor of 289, or 6.2xl026. Generally
speaking, most preset functions will default to a value of about 8 to 12 for half-lives passed,
representing a decrease in activity of 256 to 4,096.
This test is useful for unsupported radionuclides or progeny radionuclides in secular or transient
equilibrium when the progeny half-life is adjusted to that of the parent. Some software packages
do not allow this test to be disabled; in such a case it is mandatory that supported progeny be
given the half-life of the progenitor otherwise the radionuclides will not be properly identified.
42
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Note
The Key Line and Abundance (or Fraction) Limit are tests of radionuclide
presence that are redundant and should not be used together unless the user
understands how the simultaneous use of these tests will affect the results.
Key Line
Many radionuclides emit several gamma rays of different abundances. Usually at least one
gamma ray has a significant abundance and is interference free. Typically this gamma ray may
be designated the 'key line'. The software may automatically designate the first line listed in the
library as a key line, or it may have the capability of designating more than one gamma ray as a
key line. If the key line(s) is selected for a particular gamma ray in the library, then the software
will not identify the radionuclide as being present unless that key line(s) is found.
Ensuring Correct Gamma-ray Abundance
This nuclear constant has several terms that identify it, such as abundance, branching ratio, yield
or intensity. Two commonly used definitions are:
Abundance - the probability of emission of a given radiation during the decay of an atom
of a given radionuclide; see intensity
Intensity- the probability of emission of a given radiation during the decay of one atom of
a given radionuclide; sometimes called abundance.
Whichever definition is used, this is a value assigned to each gamma ray and may be changed by
the user (with extreme caution; independent checks should be performed prior to making
changes). The abundance, just like the half-life, is an experimentally derived value. The library
that accompanies the gamma spectrometry software will contain the abundance for those gamma
rays that are part of the library. However, the library may not contain all of the gamma rays
emitted by certain radionuclides. For example, the library may only include photopeaks with
abundances of greater than 1%31. (See Figure 5 for the example using 132I). When unidentified
gamma rays appear in the gamma spectrum (especially if they have small peak areas) it may be
necessary to use a library or to consult a reference that has all of the gamma rays and their
abundances listed to either identify or eliminate possibilities for the radionuclide(s) that may be
present.
Radionuclide libraries optimized for the type of sample or matrix
The peak search algorithm identifies the energy, FWHM and peak areas for the gamma rays it
finds. It then attempts to match the located energies to those listed in a library selected for
analysis of that sample. Libraries need to be selected so that those radionuclides which are most
likely to be found in a particular sample are easily identified, minimizing the chances of
31 An exception is when the radionuclide only has one gamma ray and it is less than 1% abundant. This is a rare
exception.
43
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
incorrectly identifying a radionuclide that cannot be present. As an example, a sample taken from
the spent fuel pool of a reactor that has been shut down for 18 months displays a gamma ray at
657 keV. Both 110mAg and 97Nb have gamma rays at that energy. However, the 97Nb only has a
72-minute half-life (and does not have a long-lived parent) while the 110mAg has a 249-day half-
life. Having short-lived 97Nb in the analysis library used to analyze an 18 month-old sample
would not be appropriate.
Abundance or fraction limit
Each gamma ray emitted by a radionuclide has an abundance (see Equation 6 above for other
terms used for gamma-ray abundance) associated with it. The abundance for a gamma ray is the
probability of emission of a given radiation during the decay of an atom of a given radionuclide.
The abundance limit entered by the user is compared to the ratio of the abundance of the gamma
rays found for a particular radionuclide to the sum of all gamma rays listed in the library for that
radionuclide. If the calculated ratio does not exceed the user entered preset 'abundance limit'
then the radionuclide is rejected and the lines moved to an unidentified or rejected lines report.
This test is often incompatible with the key-line test; they perform similar functions and can
work at cross-purposes and result in arbitrary rejection of radionuclides that are actually present
in the sample. It should not be used with libraries that contain more than one radionuclide since it
is not possible to determine a single threshold that applies to different radionuclides that contain
various specific lines. An exception is a software that allows the user to set different abundance
limits in the library for each radionuclide. If this option is used, however, the user must carefully
consider whether the limit will reject a radionuclide even with one or more photopeaks are
actually present in the library. Some software packages may not allow the user to disable this
function. Thus the user must know if it is or is not being used and carefully review the
radionuclides identified based on how the software functions. The process for how this is
accomplished should be documented in the laboratory's Quality Manual for Radiochemistry.
Calculating individual gamma-ray activities
The basic equation used for calculating radionuclide activity for each gamma-ray energy that is
to be used for activity calculation is as follows.
_ Rj(net)XCf
1 (EiXlj) 1 )
Where:
Ci is the activity associated with the ith gamma ray,
Ri is the net count rate associated with the ith gamma-ray full energy
peak,
Si is the detector efficiency at the energy of the ith gamma ray,
Nt
£,= A,
Cf is the factor for radioactive decay during the counting interval
Ni is the net area under the ith gamma-ray peak, counts
A is the certified activity of the calibration standard at the ith
gamma-ray peak, number of gamma rays per second
44
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Ii is the absolute abundance for the i111 gamma ray
This equation is for single radionuclides with no ingrowth contribution from a parent
radionuclide and no decay correction to time to. The term Cf is included in Equation 9 as a
reminder that if the count time is long compared to the half-life of the radionuclide (e.g.,
counting interval is greater than 10% of the radionuclide half-life), then Cf, must be included as
noted in Equation 3.
Uncertainty-based weighted mean
If multiple gamma rays are used for calculating activity then a weighted mean activity would be
reported when more than one of those gamma rays is detected.
One formula that is used for this calculation is as follows:
EP-iCCi/c^.)
Cavg = ;n i (10)
yn i
M=1 a2
°r-
Where:
Cavg is the decay corrected weighted mean activity
n is the number of gamma-ray energies identified and usable
Ci is the decay corrected activity for the ith gamma ray
Oci is the standard deviation of Ci
This equation puts the emphasis on the uncertainty of the individual peak activity; the lower the
uncertainty the more weight is ascribed to that gamma-ray peak. It is important that the software
has resolved interfering peaks and that those adjusted peak areas determine the activities of the
radionuclide whose weighted mean average is being determined.
Abundance weighted mean
Another formula for determining the weighted mean activity uses the following equation:
Y?-. CiXlr.
^i-i i 4 n
avg sF=1iCi 1 ;
Where:
Cavg = the decay corrected weighted mean activity
n is the number of gamma-ray energies identified and usable
Ci is the decay corrected activity for the ith gamma ray
Ici is the abundance of the gamma ray associated with Ci
This equation puts the weighting on the abundance of the individual gamma rays.
For both methods, a statistical test in the software may also be used to identify individual activity
that is an outlier from the mean. That activity would be eliminated and the mean recalculated
without it. This test should be used with caution making sure that the associated uncertainty of
45
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
each peak is also considered, along with resolution of interfering/interfered with peaks prior to
quantification.
Detection capability
There are many formulas that have been used to describe how low an activity can be measured
by the gamma-ray software. Some of these are:
• Critical Level Concentration.
• Minimum Detectable Concentration.
• Minimum Quantification Level.
• Lower Limit of Detection.
• Minimum Detectable Activity.
• Safe Drinking Water Act Detection Limit.
Each of the equations associated with these concepts, and other equations used to assess
detection make certain assumptions about tolerable error rate and background factors. The user
has the ability to select the detection equation to be used in a specific analysis based on the client
needs. Some software allows the user to select more than one equation to be used to calculate
detection and be included in the analysis report. Some software allows the user to change the
coverage or tolerance factor on the detection limit or critical level calculations. Thus it is
extremely important that the user review the actual equations and the influence of user-adjustable
parameters and not just the word titles, to ensure the correct detection equation is used (see
Section I for more details on these equations). Any of these bulleted concepts may be
implemented using many different formulations - even within a single software package.
MARLAP Chapter 20 goes into detail on the theory of detection. But it is still important to
underscore that
Compton background determination
The gamma peaks that will be used to quantify radionuclides in a sample effectively sit "on top"
of a background continuum produced by scattered Compton counts. The software must
determine the fraction of counts under the peak attributable to the Compton continuum so that
they can be subtracted from the total counts in that region of interest to determine the net number
of counts attributable to the gamma-ray full-energy peak.
The following is a typical method used to determine the Compton background in each channel of
a singlet gamma-ray peak.
The width of the peak is determined using the shape or resolution calibration of the detector.
This allows the software to identify a region of the spectrum over which the total counts will be
integrated to determine the gross counts under the peak. The region is bounded on the left by
channel L and on the right by channel R. The average number of counts from a specified number
of channels to the left side and to the right side of the peak is then used to determine the number
of counts at the left and right boundaries of the peak. After this, the background in each channel
under the peak, Bi, is calculated for each channel under the peak (using Equation 12 and summed
to yield the total number of background counts under the peak. The background is then
subtracted from the gross counts to yield the net counts attributable to the gamma ray. The
46
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
algorithms used to determine the background under multiple peaks can be much more
complicated.
Different software packages may allow the spectrometrist to specify the number of channels that
will be used for background determination or to allow the software to make the best
determination of the number of channels to use based on algorithms programmed in the software.
Most modern software packages also allow the spectrometrist to visually review the spectrum to
determine whether the background was adequately determined, and in some cases, to interact and
modify the integration and background determination32. This is an important step which should
always be performed to ensure that peak determinations are accurate and reliable.
Bi = Lavg + ^ (Ravg - Lavg) (12)
^j = L Ij
T _ Hh=L-N+l Yi
§ ~ N
rR+N-ly
R _ Lj = R Ij
av§ N
Where:
i is the channel at which the background is computed,
L is the channel that defines the left limit of the peak region,
R is the channel that defines the right limit of the peak region,
N is the number of channels used in the average background calculations (limited
choices but may be selected by the user),
Lavg is the average background to the left of the peak,
Ravg is the average background to the right of the peak,
Yj is the spectral contents of channel j, and
Bi is the computed background at channel i.
A good visual of how this looks for a normal peak in the gamma-ray spectrum is shown on the
next page (note that the parameter definitions in Equation 12 do not align with those in Figure
16).
32 See References 9 and 11
47
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
~~~~~~~
Figure 16. Method of Net Peak Area Determination
Figure 16 identifies the geometric means for determining net peak area. The variables shown in
the diagram are identified as:
N is the number of channels in the peak ROI,
n is the number of continuum channels on each side (currently the same on
both sides),
Bi is the sum of counts in the continuum region to the left of the peak, and
B2 is the sum of counts in the continuum region to the right of the peak.
S is the net peak area
G is the gross peak area and,
B is the peak background
This is an idealized picture for a gamma ray that has no interferences and does not sit on a
variable background (like a Compton Edge). The software algorithms that are used to calculate
the net peak area rely greatly on the calibration for resolution (FWHM), energy and low energy
tailing parameters. It is beyond the scope of this document to detail these algorithms. However, it
is always advantageous when complex backgrounds are present to visually review the spectrum
(and if needed performing a by-hand estimate of the net peak area) ensuring that the software is
adequately representing the net peak area.
48
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Peak Background Subtraction
Compton background results from scattered gamma rays not depositing their full energy in the
detector (see discussion above). In contrast, peak background refers to activity from
radionuclides in the detector, shielding, as well as ambient gamma radiation that are able to
penetrate the detector shielding to create a full-energy peak in the spectrum. It is quite common
to see relatively low-activity full-energy peaks in background spectra from naturally-occurring
radionuclides in shielding and detector materials, such as 40K, 234Th, 235U, 226Ra, 214Pb, 214Bi,
228Ra, 228Ac, 224Ra, 212Pb, 210Pb, to name a few. Other peaks may also be present if the detection
system has been contaminated with other radionuclides. The most common of these is the
annihilation radiation at 511 keV, which cannot be used for quantification of any radionuclides
since its appearance is not directly related to any individual radionuclide.
The purpose of peak background subtraction is to measure this background so that it can be
subtracted from sample results. There are several different approaches used to accomplish this.
A long background spectrum is measured in a configuration equivalent to that which will be used
to count samples. In cases where the sample shields the detector from radiation coming from
shielding materials (or beyond), such as a four-liter Marinelli beaker full of water or milk, the
background should be collected with a blank sample in place. In cases where shielding is
minimal, laboratories often use no sample container at all within the detector shield to obtain a
background. The duration of the background count should generally be longer than the samples
being counted - often by a factor of three or more. This ensures that very low-activity peaks are
well-quantified which minimizes the magnitude of bias and uncertainty associated with the
correction. The background spectrum is analyzed and the full-energy peak background activity
for each of the gamma rays detected is quantified. This activity is then subtracted from the peak
activity of the sample spectrum prior to comparison of gamma rays to the library as part of the
nuclide identification and quantification. Modern software should account for the uncertainty of
the background and propagate this into an estimate of total uncertainty to be reported with the
final result.
It should be noted that there may be no need to perform peak background subtraction if analysis
of the long background count shows no background activity that would interfere with any of the
radionuclides to be reported.
Analysis Range
One of the software settings allows the user to set the analysis range to be used when analyzing
spectra. The user should keep in mind that although energy calibrations may be reliable beyond
the range of the standards used for calibration of the detection system, efficiency calibrations are
not. Thus, although one might be able to defend making qualitative identification of
radionuclides outside the range of the calibration standards used, quantification of activity is not
reliable and should not be performed. Thus the user must make a decision as to how to use the
information obtained from gamma rays detected outside the calibration range.
49
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Decay Correction
Most gamma spectrometry software packages allow the user to perform corrections for decay
during counting, and to decay correct from the start of the count to a specified date in the past33.
These decay corrections assume unsupported decay as described by Equation 3 above and can be
applied a radionuclides gamma ray is detected without further consideration to radionuclides
present as unsupported species in the sample (i.e., they are not decay progeny of other
radionuclides in the sample). Equations are also available that describe the relationship of the
decay of supported radionuclides (i.e., decay progeny) as a function of the activity of their
respective parents (e.g., Equation. 4 and Equation 5).
If secular or transient equilibrium is established at the beginning of the decay period (e.g., the
date of sample collection or the start time of the count), the half-life of the parent can be applied
to members of the decay chain and decay correction for progeny can be made using Equation 4
as implemented in most gamma spectrometry software packages by simply substituting the half-
life of the parent for that of its decay progeny in the library. As was previously stated, all
assumptions made regarding equilibrium and sample preservation should be stated in the
analytical report.
If equilibrium has not yet been attained by the time of the count, the activity of the decay
progeny at a prior reference date cannot be calculated using most software packages, although
manual calculations of the progeny activity at the reference date may be possible using Equation
13 (using the variables as defined in Section E, Equation 6).
The following is an example of these calculations as applied to the activity of 95Nb present on the
decay correction date for a mixed fission PT sample. It also emphasizes the importance of data
review as the data produced were improbable. The activity of 95Zr and 95Nb at the time of the
count, the elapsed time between the decay correction date and the count, and the half-lives of the
two radionuclides need to be known to be able to make the final calculations.
33 Each software application is unique and it is, as always, the user's responsibility to understand the capabilities and
limitations if each software application, keeping in mind theoretical limitations of these decay corrections as pointed
out in this section.
(13)
50
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Acquisition information
Start time:
Live time:
Real time:
Dead time:
Corrections
Decay correct to date:
Decay during acquisition:
Decay during collection:
12/6/2012 14:28:49
60000
60077
0,13 %
Status
YES
NO
YES
Comments
11/23/2012 11:00:00
11/23/2012 11:00:00
11/23/2012 11:00:00
Nucli de
SUMMARY
Time of Count
ACti vi ty
pci/L
OF N U C L I
Time corrected
Activity
pci/L
D e s
I N
Uncertainty
Counting
SAM
1 Sigma
Total
P L E
MDA
pci/L
CO-58
FE-59
CO-60
ZN-65
NB-95
ZR-95
1-132
Te-132
Np-239
ru-103
#A
#A
#A
#A
-3.8309E+00
-6.5565E-01
6.4002E+00
4.9813E+00
9.526OE+02
1.4724E+03
1.4423E+02
1.4033E+02
1.8190E+03
2.3314E+03
-4.3568E+00
-8.0464E-01
6.4306E+00
5.1708E+00
1.2359E+03
1.6975E+03
2.4218E+03
2.3564E+03
8.6974E+04
2.9406E+03
6.4764E+02%
1.1151E+03X
1.0304E+02%
1.7241e+02%
9.5369E-01%
1.0129E+00%
3.2951E+00%
1. 5971E+01S6
2.6S60E+00%
5.4900E-01%
6,4767E+02%
1.115lE+03%
1.0318E+02%
1.7250E+029S
5.4469E+00%
5.4581E+00%
6.7720E+005S
1.7075E+01%
6.4979E+00% 5
5.4018e+00% 2
751E+01
997E+01
193E+01
973E+01
578E+01
446E+01
280E+02
140E+03
348E+03
148E+01
Decay correction Date and Time
11/23/12 1100
Count Date and Time
12/6/12 1429
Elapsed Time (days)
13.14
95Zr measured at time of count, Ai1
1472.4 pCi/L
9?Nb measured at time of count, A21
952.6 pCi/L
Figure 17. Decay Correction for 95Zr-95Nb
Using the data from Figure 17 and Equation 13:
/1472 4\
952.6 - (() 3fa74j (2.205)(0.8674 - 0.7708)
0.7708
= 766.8 pCi/L
These results indicate that 766.8 pCi/L of 95Nb was present at decay correction date (the day the
sample was preserved) based on the 95Zr and 95Nbactivity at the time of the count. The date of
irradiation of the sample was known to he 11/9/2012. Had 1698 pCi/L of 95Zr been present on
11/23/2012, we could calculate that there should have been 1972 pCi/L of 95Zr on 11/9/2012, and
only 438 pCi/L of 95Nb should have been produced by 12/6/2012 (the count date). It is also
worth noting that very little 95Nb is produced as a fission product (<0.1% relative to 95Zr). Where
did the extra 95Nb come from?
Investigations showed that due to differences in their chemistry and inadequate preservation of
the sample, some 95Zr had been lost from the PT sample while 95Nb had not (see further
discussion under Figure 35). This example demonstrates the importance of:
• Reviewing the data to make sure they make sense radiochemically;
• Ensuring the proper preservation of samples; and
• Making sure to be very careful about making assumptions about the sample prior to
collection and preservation.
When multiple fission events occur, a fission product and one or more of its decay products will
be produced. The parent/progeny decay and ingrowth corrections employed by gamma
51
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
spectrometry software are not capable of establishing the relative amounts of each decay product
at the time of fission; this calculation must be accomplished manually.
Finally, if the parent and progeny are already in equilibrium at the time of the count, and we
cannot know at what point they reached equilibrium it will not be possible to (unless the progeny
half-life is very short) calculate the activity of decay progeny at a point in the past, either with
software or manually without the assumption that the sample was in equilibrium. Such an
assumption may or not be defensible. Such decisions should always appear in the final report to
the client.
Some software packages provide support that allows the user to take into account more complex
radioactive equilibria, such as secular and transient equilibria. These applications are unique and
it is, as always, the user's responsibility to understand the capabilities and limitations if each
software application, keeping in mind theoretical limitations of these decay corrections as
pointed out in this section.
As discussed above, care must be taken to consider all factors although decay corrections are
performed based on the half-lives of the nuclides, and the time elapsed, other factors such as the
chemical and physical stability of the radionuclides in the matrix may be critical and need to be
known throughout the entire decay period in question.
Decay Build up Options
Many gamma spectrometry software packages offer to correct the results for the "build up" of
activity and some of these routines apply to the activity of build up during activation analysis
(which is beyond the scope of this discussion). The most common application to environmental
testing is the sampling of radionuclides in air samples, to reflect the average activity present
during the period of time during which the sample was collected. They generally require the date
and time of the start and the completion of sample collection (note that this assumes a constant
flow rate during the entire sample collection period which may not always be the case). The
formula used is similar to that used for decay during counting (see Equation 3). These
corrections should only be applied to unsupported radionuclide activity and not for that of decay
progeny (where the parent is known to be present or must be present for the progeny to be
present) unless it is known that the progeny were in equilibrium at the time they were captured in
the air sample (see Example 8).
Interference Correction (Peak Overlap)
Two or more radionuclides may emit gamma rays that are located very close to one another such
that their emissions would form a single peak that is a composite of the counts from each of the
two peaks. Thus, is it not uncommon for a library to have what may be termed energy
duplication. In other words, energy lines for two (or more) radionuclides may fall within the
energy comparison match width specified by the user to assign photopeaks to radionuclides in
the library. In the best case, the overlapping gamma rays will be minor emissions and will not be
needed for quantification of the radionuclides in question and the conflict can be avoided by
omitting the offending lines from the library (or better, by including them in the suspect library).
When any of the gamma rays are needed for quantification, however, it is important to include
52
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
all overlapping gamma rays in the library and to make sure that the interference correction option
in the software is properly configured and enabled.
Gamma spectrometry software packages take a number of approaches to address library energy
duplication conflicts. While it will not be possible to describe all of them, they tend to rely on a
common principle. Assume that there is a radionuclide that emits a gamma ray at an energy
where there may by conflicting gamma rays emitted by other radionuclides. If the other
radionuclides emit other reliable ("clean") gamma rays that are abundant enough to be accurately
measured in the sample, we can calculate the contribution of counts at the duplicated energy
since we know the abundance and the efficiency for gamma rays at both energies. We can then
subtract them from the interfered peak area to calculate the activity attributable to our first
gamma ray.
For example, given the data acquired for the spectrum in Figure 18 we see a classic example of
an energy conflict between 226Ra with 3.64% abundance at 186.2 keV and 235U with 57.0%
abundance at 185.7 keV.
1 63
Figure 18. U-235 and 226Ra Overlapping Gamma Rays
Since 235U and 226Ra both contribute to the peak at 185 keV (i.e., there is significant overlap of
the gamma-ray energies and the peak is not well resolved), a manual determination of the
individual peak areas can be made by using the secondary gamma ray of 235U. The software
determines that over the counting period of 2400 seconds, there are 8174 net counts in a peak at
186.0 keV. There are 1419 net counts in a peak at 143.8 keV (solely attributable to 235U). The
activity of 235U, A235U. in the sample is calculated based on the clean line at 143.8 keV as follows:
Rnetl43 0.5913 counts/second
A235tt = -— = 7 r- = 77.91 Bq
U e x f0,0690=ts) x (O.IIO . / t. )
V y / \ disintegration/
Where:
Act235uis the activity of 23 5U
Rnet 143 is the net count rate in the 143.8 keV peak (counts / second)
53
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
s is the detection efficiency (counts/gamma)
I is the gamma-ray abundance (gammas/disintegration)
Knowing the gamma-ray abundance for 235U and the efficiency at 186 keV, we can calculate the
contribution of counts from 235U in the interfered peak.
Rnet = A235u x e x I = 77.91 Bq x ^0.0716 ^ x ^0.570^. . ^^ = 3.179 counts/second
The contribution of 235U to the 186 keV count rate is subtracted from the 186 net peak count rate.
The remainder is attributable to 226Ra. Now we can calculate how much 226Ra is present.
Rnet 186 — R235u 3.406 — 3.179 counts/second
A i86p = = — = 87.1 Bq
eXl (o.0716—) X fo.0364,. . Y .—)
V y / \ disintegration/
Where:
Rnet 186 = net count rate in the 186 keV peak (counts/second)
Rnet 235u = count rate in 186 keV peak from 235U (counts/second)
Clearly, this is a simple example designed to show the underlying principles involved in
interference corrections. Modern gamma spectrometry software packages can accommodate
much more complex situations, including multiple interfering radionuclides and propagation of
uncertainty associated with the corrections. It is important to understand how specific
interference correction schemes work and to correctly configure the software. This includes
ensuring that any libraries used to calculate interference corrections include all interfering
radionuclides and peaks.
One weakness of such an approach is apparent from this example. The 143.8 keV gamma ray in
this example is significantly less abundant than the 185.7 keV gamma ray for 235U. Unless there
is enough activity in a sample to quantify 235U, no correction may take place (if 235U is not
detected). Similarly, the levels of uncertainty introduced into the corrected value depend on the
levels of uncertainty in the measured 235U activity. Unfortunately, the software will not generally
point this out to the user, so checking interference corrections during the review of the final
results is important. One method of determining if results for activity from different gamma rays
agree is using a statistical test such as the relative error ratio (RER) or a Z-score test34. These
have validity in that they take into account the uncertainty of the measurements, and not merely
based on an arbitrary percentage.
The second concern with peak overlap and how the software "views" the overlap is the function
of peak deconvolution. The software may actually perform the deconvolution, but it may
automatically assign the closest energy to both peaks as opposed to really resolving the
interference.
34 See Volume III f Reference 7.
54
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
TCS or True coincidence summing
The coincidence or cascade summing effect was described in Section IV.C. Most software does
not have a provision for doing this unless the system has been calibrated with a special source
and with a true coincidence calibration method. Some software attempts to correct for this using
some mathematical formulations. The best way to present this is by example using a radionuclide
that exhibits this effect: 60Co.
Figure 19 shows an example spectrum
1332 keV,
446550 cts
1173 keV,
486094 cts
Figure 19. A Spectrum of 60Co Showing Peak Areas (y-axis is log-scale)
One method of calculating coincidence correction is to make the assumption that all the counts in
the sum peak came equally from the two peaks that make the summed energy. One-half of the
counts from the summed peak is added to each of the lower energy peaks after they are corrected
for the separate efficiencies of the two energies. While this accounts for the majority of the loss
of counts, it creates an indeterminate bias.
However, coincidence correction is most effectively accommodated by specific calibration (See
Reference 27 for details).
Directed Fit
"Directed fit" refers to an option used for analysis and reporting of results that is most frequently
used to support environmental monitoring programs where a value is calculated for a
radionuclide at a library energy even though a peak is not detected by the software. These
calculated values should be reported together with their associated uncertainties even when the
55
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
activity for the measurement may yield results that are not detectable (i.e., the result may be
positive, negative or zero). By not censoring values (i.e., reporting "less than" or "ND" as a
result), a series of measurements close to background can be averaged over time which may
facilitate quantification of very low levels of activity, or detection of changes in very low levels
of activity before they become detectable in individual measurements.
The process used by the software establishes a region at the energy where the peak would
normally be present (e.g., three times the width of the calibrated FWHM centered on the peak
energy). A background is calculated across this region and then subtracted from the integrated
total counts for this region. The net count rate thus determined, together with the detection
efficiency, abundance, and any other corrections (e.g., decay correction, sample size, etc.) are
used to calculate an activity and an associated uncertainty.
The preset functions noted above may need to change with the type of sample, the type of
radiological event, and the radionuclides being analyzed. Some generic issues where the
functions have either misidentified or not identified a radionuclide are described in general here
(with specific examples are shown in Attachment II).
Suspect Library
The library selected for analysis of the gamma-ray spectrum, sometimes called the "analytical"
library, is usually a subset of the main software library that has been edited to contain
radionuclides of concern and information that is needed to support the analysis (e.g., lines
needed to perform interference corrections that also reflect the half-lives for those radionuclides
that are in radioactive equilibrium). Gamma rays that are not contained in the analytical library,
however, are often observed in spectra. Gamma-ray spectrum reports should always contain a
section that lists all gamma rays that do not correspond to those in the analytical library (i.e.,
"unidentified" gamma rays). Most software packages also have a separate library, sometimes
called the "suspect" library that assists the analyst in establishing the identity of unknown
gamma rays in the spectrum. This should be a library that has been optimized to the type of
sample being analyzed. It generally contains all possible photopeaks for any radionuclides listed
in the analytical library (overlap is not problematic), as well as information for non-target
radionuclides that may possibly be present in spectra, along with information that will assist in
identifying spectral anomalies (e.g., sum, escape or non-target background peaks). Since the
suspect library is not used for the analysis, it may be updated and maintained as needed without
risking the integrity of the analytical library or negatively impacting results.
Alternatively, the stored spectrum can be reanalyzed using a different or newly created library
that addresses the diversity of radionuclides contained in the samples analyzed during a specific
event.
Nuclide ID Confidence
This is a calculated parameter. The radionuclide listed in the library must have at least one
gamma-ray energy within the user-selected energy tolerance of a peak identified in the spectrum.
For each gamma-ray energy of a radionuclide in the library that is within the energy tolerance of
an observed photo-peak, the nuclide confidence value (which starts out as 1.0) is multiplied by a
56
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
penalty function that reduces the value to less than one. The penalty function includes the
following factors: all the gamma rays associated with this radionuclide in the library, the
abundance of each of these gamma rays, the deviation of the measured energy from the library
listed energy the energy tolerance used in the gamma spectrum analysis and the MDA in the
measured spectrum for each of the gamma rays for that radionuclide. This function should be
used with caution for several reasons:
• A low activity sample with a radionuclide that has multiple gamma rays will get a low
confidence ID even though it is present.
• Gamma rays from other radionuclides that may be close in energy can cause the ID
confidence of a different radionuclide to be significant when the second radionuclide is not
present.
• There is no exact value for the ID confidence that can be used for detection purposes.
• Depending upon how the library is configured falsely high ID confidence values may be
recorded.
Ultimately it is the responsibility of the reviewer to ensure that the gamma rays used for the ID
confidence factor are appropriate for that factor (see Examples 3 and 15 in Attachment II).
G. Selecting detectors
Bigger is not always better. For samples where we expect to find very low activity
concentrations of only a limited number of radionuclides, a larger volume detector would be of
benefit because it will have a greater efficiency of detection. A greater efficiency may lead to
either smaller sample or a shorter count time (or both) to achieve the required detection level.
However larger detectors generally will have a wider FWHM (i.e., poorer resolution) than a
smaller detector. Therefore it is necessary to ensure that the radionuclides to be determined will
have gamma rays of sufficiently different energies so that the poorer resolution will not affect
detection (although in recent years this has improved with digital filter technology). Also note
that larger detectors suffer from higher levels of true coincidence summing (depending upon the
radionuclides being analyzed), especially when samples are counted close to the detector such as
will likely be done to optimize sensitivity.
Larger detectors may also have higher Compton background at lower energies due to an
increased probability of interaction with scattered high-energy gamma rays, and gamma-ray
peaks will be slightly wider.
Conversely, for high activity samples with multiple radionuclides, it would more likely be
necessary that high resolution (i.e., small FWHM) is needed so that gamma rays close in energy
can be easily resolved.
A second decision that laboratories must make is between P-type coaxial detectors, with good
response in the mid-to high-energy range, or extended-range N-type coaxial detectors with thin
entrance windows (e.g., carbon-fiber or beryllium) and good response at low-energies. While it
may be tempting to conclude that having better response across all energies is better, this may
not result in selecting a detector that is optimal for the measurements a lab needs to make. Since
N-type detectors are sensitive to low-energy radiations, they suffer elevated rates of true-
coincidence summing, especially gamma plus X-ray summing. This will impact the accuracy of
57
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
measurements unless the laboratory takes the effort to perform single-nuclide calibrations, or
they implement true-coincidence summing corrections, both of which can be costly and time-
consuming. The thin entrance windows are also notoriously fragile and can break leading to
expensive and time-intensive detector repairs.35 For this reason, unless the laboratory needs to
use the detector to make sensitive measurements below 60-80 keV (e.g., 210Pb, 129I, 55Fe, etc.),
the extended range of the detector may not be needed or could even be deleterious. If the
laboratory is planning on performing low-energy measurements, it should perhaps look into
planar detectors (or other low-profile configurations) that are sensitive to low-energy photon
radiations but have much lower backgrounds than coaxial detectors and are thus the best-suited
detectors for low-energy measurements.
H. Sample preparation
The sampling process is not addressed in this manuscript as there are separate documents that
deal with these practices for many different matrices. Also not addressed in this document is a
methodology for screening samples from a radiological event. This particular issue is addressed
in "Radiological Laboratory Sample Screening Analysis Guide for Incidents of National
Significance" (see Reference 13).
However, once the sample is obtained, the process of preservation and preparation of the sample
for gamma spectrometric analysis needs to consider the different radionuclides that may be
present and their potential chemical and physical forms. Samples obtained following a
radiological event such as an IND or RDD will likely contain discrete radioactive particles
(DRP). The physical and chemical nature of radionuclides in a sample containing DRPs will be
significantly different from those radionuclides found in environmental sampling. The high
temperature from the devices will cause some radionuclides to fuse with other materials such as
sand, metals, even plastics, creating a small particle that contains a very high activity of one or
more radionuclides. If these DRPs exist as suspended particulate in water samples, they cannot
be dissolved simply by adding acid, and their distribution in the solution may be questionable
due to settling or surface attraction to the container walls. Thus, sample preparation may need to
consider filtration, fixation of the material in suspension (by addition of a gelling agent) or
something else.
Sampled solids may also suffer from lack of homogeneity. A method has been developed for
homogenizing soil samples (see Reference 12).
Air particulate filters will generally have their pore size specified in sampling documents. During
radiological events, it is possible that the particulate loading of the filters will be elevated above
normal. The mass of material captured on these filters and the average Z value of those
particulates will determine if correction factors for self-attenuation are required as compared to
the standard used to calibrate the detector. This type of information should be requested by the
laboratory of the incident command sampling group so that corrections can be made if
warranted.
35 Detectors with thin entrance windows can and should be protected, except when being actively used for low-
energy measurements. While plastic covers are often used to this ends, graded end caps composed of layers of
copper and tin have been used to both protect the fragile end cap while minimizing gamma-x summing by reducing
the number of low-energy radiations that can reach the detector.
58
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
It may be necessary to assess whether or not a silver zeolite or an iodine charcoal cartridge is
face- or fully loaded. Most calibrations assume these cartridges as face loaded in a thin film on
the absorber's surface (the side facing the air flow into the housing). Transferring the sorbent
material to a secondary container after homogenizing, or counting the canister on its side, and
flipping or rotating the sample half-way through the count have been used to minimize bias.
Laboratories should have specific instructions on how to perform these options.
Climactic conditions such as heat, humidity, and presence of other non-radioactive, absorbable
gasses may prevent the radiological kryptons, xenons, or radons from forming a thin film on the
inlet to the canister. In addition, the presence of high activity concentrations of the noble gases
may have a negative impact on the detection limit for 13 *1 as there may be increased background
around the principal peak at 364 keV. A slow short purge with nitrogen or argon can help to
remove the radioactive noble gases and reduce background in the gamma-ray spectrum.
Samples other than liquids or air particulate and charcoal cartridges may need to be prepared by
grinding, sifting, or gelling, and be analyzed as wet, dry, or as received. The event authority will
need to specify what the appropriate method of reporting these activity concentrations is. In
general sample preparation must be carefully considered for any matrix that potentially contains
DRPs.
I. Selecting Geometries, Counting Containers, Detection Equations, Count
Intervals and Uncertainty Equations
Detector Geometries
HPGe detectors have shielding that cannot be easily reconfigured. Thus the sample-to-detector
geometry needs to fit inside the shielding. Generally, a container as large as 4-liter Marinelli
beakers can fit within the shield. A very important aspect of sample placement geometry is that
the sample container should never touch the detector; the detector is not a sample stand. A
sample holder or sample geometry made of low Z material should always be used to offset the
sample from the detector surface. The best reasons for this offset for the sample will include at
least some of the following:
• Cascade summing effects are minimized
• Probability of detector contamination is minimized
• Sample mass will not 'push' the detector cold finger into the cryostat
• Sample does not act as a 'heat sink' for the liquid nitrogen (or the new "cryo-coolers") via
contact with the detector housing.
Using low Z material for fixed counting geometries minimizes attenuation by the Photoelectric
and Compton Effects so that more gamma radiation reaches the detector. Plastics such as high-
density polyethylene, Delrin™, Lexan ™, or polycarbonate are most frequently used. Plastics
like polyvinyl chloride (PVC), Teflon™, Tefzel™ or other polymers containing halogens should
be avoided as these elements raise the average Z value of the material.
1. Sample Counting Containers and Sample Size
a. Standards containers (i.e., working calibration sources, WCS) used to calibrate the
detector need to have the same volume, geometric configuration, and composition as
59
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
those containers that will be used to count samples. As with the counting geometries,
the sample containers should be of a material that has low Z value. Container sizes
should be selected so that the routine mass or volume of sample can be accommodated
and match closely to that of the WCS.
b. Sample size and counting interval (see next section) are important considerations when
evaluating the detection limit that is needed to be achieved. The basic equation for
activity concentration is:
A = (NT~NB)Cf (l4)
c £XVxtxIxe~XxAt
Where:
Ac is the activity concentration, Becquerels per mass or volume
Nt are the total counts for the gamma-ray peak
Nb is the Compton background counts for the gamma-ray peak
Cf is the correction factor for decay during counting (DDC)
s is the efficiency with which gamma rays are detected from the geometry in
question
V is the sample volume or mass
I is the gamma-ray abundance
t is the live time
X is the decay constant for the radionuclide, and
At is the time interval between sampling and start of counting
Optimizing a counting geometry to obtain the best limit of detection involves balancing several
factors. As the sample size increases, the limit of detection generally decreases (i.e., the signal-
to-noise and sensitivity improve). However, as the sample size increases, the geometric shape of
the STS increases and the average distance of the sample to the detector increases. As the
average distance increases (depending on the matrix and density and the energy of the gamma
ray, the efficiency may also decrease due to self-absorption) efficiency decreases with the
inverse square of the distance, and fewer net counts per unit sample mass or volume will be
recorded. The number of counts collected for a given gamma peak plateaus as the size of the
sample approaches "infinite thickness"36. At infinite thickness and beyond, increasing the sample
size results in no net increase in photopeak signal (counts). In fact, if the matrix being counted
contains other radionuclides (e.g., soil), increasing the sample size may result in a net loss of
sensitivity since, although the efficiency may plateau, the background continues to rise from the
Compton interactions of the other radionuclides (see discussion of uncertainty and MDC below).
36 Infinite thickness is the point at which gamma rays from the sample are absorbed within the sample and no longer
can reach the detector. This "thickness" depends also on the energy of the gamma ray and is generally larger for
high-energy gamma rays than for low-energy gamma rays.
60
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Counting Interval and Detection Equations
The measurement quality objectives (MQOs) required by both normal operations and incident
response will need to specify the detection limits to be achieved and most importantly the
required method uncertainty for the analyses. There are many different terms used for detection.
Four of these terms will be described here with equations often used to implement them, and a
comparison of how the values for each differ.
It is important to note that while we identify a detection capability, it is the numerical result of an
analysis obtained from measurements, along with its associated uncertainty, that should always
be reported. Terms such as "Not Detected (ND)", "
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Sample
0
MDC
Figure 20. Distribution of Results for a Blank sample and Sample with Detectable Activity
A commonly used formula for the critical level concentration38 is:
B is the background counts
t is the live time
£ is the detector efficiency at the gamma-ray energy
V is the sample volume or mass
A is the gamma-ray abundance
The final units for Lc from this calculation will be in Becquerels per unit mass or volume. The
critical level concentration is a sample specific determinant of activity concentration (i.e., an a
posteriori determination) based on a true blank having "zero" concentration and measuring the
sample-specific value for "B". In gamma spectrometry, the sample Compton background is used
to determine the Lc.
The critical level concentration (Lc) can also be defined as "the smallest concentration of
radioactive material in a sample that will yield a net count (above background) that will be
38 The critical level, as calculated here, assumes that count time used to determine the background activity is equal to
that of the sample. This is the case in a routine gamma spectrometry measurement where the background subtraction
is determined from the same spectrum as the sample. If peak background subtraction is performed, however, the
background should be determined from a separate count of duration at least equal to, but generally significantly
longer, than that of the sample.
Lc = 2.33 X
txsxAxV
(15)
Where:
62
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
detected with a 95% probability."39 This definition is more closely aligned with that of the MDC
(see next paragraphs).
Minimum Detectable Concentration (MDC)
A general formula40 for the minimum detectable concentration is shown in the next equation.
MDC = (16)
tcX£XAxVxkwxkxxFxUf v y
Where:
Ld is the twice critical level plus the square of the tolerable error rate,
,2
Ld - 2 X Lc + (z^a))
£' is the attenuation corrected efficiency
F is the sample mass/volume conversion factor
V is the volume or mass
A is the abundance factor for the gamma ray
Cf is the correction for decay during counting
kw is the decay correction from the start of the count time to the time the sample
was obtained.
tc is the live time
Uf is the unit conversion factor from Bq to the desired units (e.g., pCi)
kx is the correction for decay during sample accumulation time.
The final units for MDC from this calculation would be, for example, pCi per unit volume or
mass. A definition of MDC that is commensurate with the above equation is:
"The smallest (true) value of the net state variable that gives a specified probability that
the value of the response variable will exceed its critical value i.e. that the material
analyzed is not blank."
Figure 20 displays the location of a series of measurements that would represent the MDC
distribution and how that distribution relates to both the critical level and "true" zero. Another
way to describe the MDC (when the Type II error ((f) is set to 5%) is as follows:
The smallest a priori radionuclide concentration for which there is:
• 5% probability of obtaining a result below the critical level (i.e., 5% probability of falsely
concluding that the result represents a blank);
— and simultaneously —
• 95% probability of obtaining a result greater than the critical value (i.e., 95% probability of
concluding that there is activity in the sample and it is not a blank).
39 See Reference 11
40 See Reference 8.
63
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Minimum Detectable Activity (MDA)
A general formula40 for the minimum detectable activity is shown in Equation 17.
MDA = A+VB+Cxcounts
Live timex(efficiencyxyield)
The units for the MDA from this equation are in Becquerels/(unit sample). This is specifically
different than other detection equations in that it does not perform a calculation per unit mass or
volume, but per sample regardless of size. This detection formula is used in those instances
where only the total activity present is of concern as for a swipe or perhaps an air particulate
filter. In this formula, the constants A, B, and C are selected by the user from the available
software settings a snapshot of which is seen in Figure 21. The first two constants A and B
represent the user's ability to add background counts if the background about the region of
interest (ROI) is zero. This prevents calculating a zero uncertainty. The value of C, often selected
as 21.7 is statistically derived from the tolerable error rate of Zi-a2 from the critical concentration
formula at 5% tolerable error rate (see MARLAP for a detailed discussion of this factor).
Properties for DSPEC-50
Amplifier | Amplifier 2 ] Amplifier PRO | ADC | Stabilizer
High Voltage | About | Status | Presets MDA Preset | Nuclide Report
MDA Preset
Nucl Energy
150.0000
Co-60| 1173.24
150.0000
Co-60 1332.50
1
f J N er'/-' ' |
Update
Delete
MDA Preset
Correction
Nuclide:
150.0000
cA.
Co-60
Energy: 1332.50 ^ keV
Coefficients
A: 10.000000
B: 10.000000
C: [21^7
Dugg
dose
Figure 21. Screen Shots from Software Identifying Selectable Constants for MDA Equation
When the values for constants are set at A=0, B=0, and C=21.7, the formula will yield the same
results as that presented below for the LLD equation (with the exception that the units for the
sample are omitted).
Lower Limit of Detection (LLD)
While there are several definitions for LLD the one that is used by the Nuclear Regulatory
Commission and defined in both Regulatory Guide 4.1441 and NUREG-1301 (Reference 9) is
41 See Reference 10
64
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
LLD, [iCi/mL = 4-66xgb tt- (18)
^ ' (3.7xl04xExVxYxe"XAt) v '
Where:
ofe = standard deviation of the background, dps
E = detector efficiency at the specified gamma-ray energy
V = volume of sample in mL
Y = Chemical yield (=1 for gamma spec)
At = elapsed time between sampling and counting
This equation determines an a priori limit of detection based on optimal characteristics of the
detection system and the sample. The two most significant factors in this equation that are
subject to change from the a priori measurement to the actual sample analysis are the value of CTb
and At. The determination of background is made by counting a sample container filled with the
same material as a sample and with the same geometry42, but completely void of radioactive
materials. This measurement assumes that the background is constant regardless of the presence
of other radionuclides in a sample. However, this is not necessarily true as the presence of other
radionuclides can elevate a sample background in certain regions of the gamma-ray spectrum
due to Compton and other beta decay effects that produce photons. Thus the LLD can provide an
estimate of the detection limit under ideal conditions where the sample and blank background are
the same.
The time between sampling and analysis brings into play the factor that if the radionuclide has a
short half-life and you have counted the sample too long after sampling so that the radionuclide
is undetectable then decay correction for an LLD at the time of sampling is useless.
Table 5 shows the values calculated from the different detection equations noted above. The
parameters for determination of 137Cs concentration, sample volume (1 L was used), count time,
detector efficiency, and sample background are kept constant for each detection equation
calculation on each row of the table.
When selecting an equation and constants to use for detection, the analyst should always ensure
that the end user or data user (client) is in agreement with the equations and values used and that
they properly accommodate the data and measurement quality objectives of the project.
42 Required inNUREG 1301, "Offsite Dose Calculation Manual Guidance: Standard Radiological Effluent Controls
for Pressurized Water Reactors" and in ASTM D7282 (Reference 15)
65
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Table 5. Detection Equation Calculations
Activity at Beginning of Count
Interval, pCi/L
Bg, cps
Detector
Efficiency (at
661 keV)
Lc
MDA
LLD
MDC
21600 sec
0.01
0.01
5.0
8.6
10.1
10.5
(6 Hours)
0.05
0.01
11.3
19.2
22.5
22.9
0.1
0.01
15.9
27.2
31.8
32.2
1
0.11
4.6
7.8
9.2
9.2
10
0.21
7.6
12.9
15.2
15.2
100
0.31
16.2
27.7
32.5
32.4
3600 sec
0.01
0.01
12.3
21.0
24.7
27.0
(1 Hour)
0.05
0.01
27.6
47.0
55.2
57.5
0.1
0.01
39.0
66.5
78.0
80.3
1
0.11
11.2
19.1
22.4
22.6
10
0.21
18.6
31.7
37.1
37.2
100
0.31
39.8
67.9
79.6
79.5
Note that the value for the constants in the MDA equation used were A=0, B=0, and C=5.43.
This was done to allow the numerator to align with the LD value of twice the Lc.
Uncertainty Determination
The values reported for the analyses of samples by gamma-ray spectrometry are of little value
without the associated measurement uncertainties of the results. MQO's associated with the
analyses should specify the required method uncertainties for analytical results. Calculation of
uncertainties can be very complex and are covered in significant detail in MARLAP (Reference
7) and GUM (Guide to Uncertainty Measurements-Reference 17). Presented here is a basic
outline of the calculation process and the inputs that should be considered.
The basic process for evaluating and reporting the combined standard uncertainty (also referred
to as total propagated uncertainty) of a measurement requires:
1. Identifying the measurand, Y, and all input quantities, Xi, for the mathematical model used
to calculate the result. Include all quantities whose variability or uncertainty could have a
potentially significant effect on the result (uncertainties <0.1% are not significant).
2. Express the mathematical relationship, Y = f(xi, X2,...xn ) between the measurand and the
input quantities. In most cases for radionuclide analysis, the model will be based on the
equation used to calculate the activity concentration.
3. Determine an estimate of the value of each input quantity, xi
4. Evaluate the standard uncertainty, u(xi), for each input estimate, xi, using either a Type A
or Type B method of evaluation
66
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
5. Evaluate the covariances, u(xi, xj), for all pairs of input estimates with potentially
significant correlations (this is occasionally difficult to evaluate and is often ignored since
it is usually small compared to other uncertainties.
The combined uncertainty can then be calculated using guidance in MARLAP (Chapter 18).
Alternatives for calculating combined uncertainty include commercial computer programs and
free programs that follow GUM principles. These include GUMCALC available at
http://www.mccroan.com/gumcalc.html and others that can be obtained by searching the web.
These programs allow the user to develop a model for calculations and input variables to be
included in uncertainty calculations.
While the most convenient method for the uncertainty of activity concentrations is the software
that is used for analysis of the gamma spectrometry data, this only uses input parameters that are
specified in the equations in the software. Usually, the uncertainties of input parameters included
are:
• Peak background
• Peak area
• Gamma-ray abundance
• Efficiency at the specific gamma-ray energy
User input includes providing uncertainties for nuclear data including uncertainties for branching
ratios, uncertainties for half-lives or decay constants. The software also will calculate the
uncertainties associated with decay corrections to the start or midpoint of the count interval, and
from the start of the count interval to the sample collection time. There is usually an option for
user-defined uncertainties that have been evaluated such as uncertainties associated with sample
positioning. This option is usually available in more sophisticated software packages.
Other sources of uncertainty that may need to be considered, but may not be included in the
gamma spectrometry software, are:
• Use of an aliquant versus the entire sample
• Density difference between standard and sample
• Sample inhomogeneity
• Mass or volumetric measurements
It is recommended that the user carefully review the source code to validate the methodology
used to calculate the combined standard uncertainty.
J. Detector Calibration and Non-Routine Counting Geometries
Energy and Resolution Calibration
The process of instrument set-up includes assuring that proper electrical capacity and voltage
regulation/surge protection have been installed, the detector shielding is securely in place,
several calibration geometries have been selected, the detector has had sufficient time at liquid
nitrogen temperature to ensure that it is at equilibrium, and the manufacturer's voltage for that
detector has been applied. Once the manufacturer's method for this set up has been completed,
the next step is energy calibration. The user must decide the number of channels that will be used
67
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
for the energy range of gamma-ray emitters of interest. That number of channels usually spans
59 to 1836 keV (the energies of the routine low and high energy calibration sources). In most
cases the energy calibration used will be approximately one of the following three:
• 1 keV/channel
• 0.5 keV/channel
• 0.25 keV/channel
A radioactive source is obtained that contains radionuclides that emit gamma rays at energies
that span the energies expected in samples to be analyzed.
The source used should be of sufficient strength so that the uncertainty of the net counts in each
gamma-ray peak contributes negligibly to the combined standard uncertainty. Guidance provided
in ANSI 42.1443 discusses using 20,000 net counts and 50,000 net counts for the efficiency and
shape calibration functions, respectively. These net counts should be achieved in a reasonably
short period of time so that any variability in detector environmental conditions is also
minimized.
The source is counted and the channel numbers of each photopeaks' centroids are plotted against
the known energy of those peaks and the slope of the curve calculated. The gain on the amplifier
may be adjusted and the standard recounted. This process is repeated until the peaks are in the
channel desired to fit the energy calibration. The source is recounted so that sufficient counts in
each of the peak centroid channels are accumulated and a calibration for energy and resolution
can be calculated by the software. Energy calibration over the range of 59 to 1836 keV is quite
linear. Beyond those limits, the energy may stray from perfect linearity. Energy calibration (E) as
a function of channel number x is shown in Equation 19 (a commonly used formula for energy
calibration):
E =CQ +ClX+ C2x2 +C3x3 (19)
Where:
CQ, Cp C2, and C3 are coefficients determined by the calculation performed
by the software
E is the photopeak energy in units of keV
x is the channel number of the centroid
This equation is a typical approach to energy calibration and assumes a cubic fit of the data.
Other equations for energy calibration may be available in the software.
43 See Reference 16
68
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
CENTROID CENTROID
TAIL
FWHM
FWI-IM
TAIL
Figure 22. Calibration for Resolution
The term resolution refers to the full-width-at-half-maximum (FWHM) of a photopeak. The
FWHM is determined by taking the number of net counts in the peak centroid and dividing by
two. The channels that correspond to these counts on either side of the photopeak energy are
converted to energy and difference between these two energies is the FWHM.
One method for calibration of resolution uses peaks like those shown in Figure 22 which are
fitted to a modified Gaussian curve and assumes that the FWHM = 2.355xo, where o is the
standard deviation for the modified Gaussian curve44. The peak centroid is determined from the
data points and may not be one of the actual data points since the curve is fitted to the Gaussian
shape.
A FWHM calibration curve is generally constructed using data collected during the energy
calibration. The observed FWHM is plotted as a function of energy (or channel). A commonly
used formula for resolution calibration is
FWHM = Fo+FlV^ (20)
Cl
Where:
E is the gamma-ray energy in keV,
Cl is the "gain" term from Equation 19, and
F and FQ are the coefficients of the FWHM equation
The coefficients for equation parameters (e.g., Fi, Ci, etc.) are empirically determined for each
detector based on the response of the instrument.
44 See Reference 21
69
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Efficiency Calibration
It is of paramount importance that the physical geometry of the sample is identical to that of the
standard used to calibrate the detector. In addition, the juxtaposition of the sample container with
respect to the detector must also match the standard container position when the standard was
used to calibrate the detector. Figure 23 is an example of an HPGe efficiency calibration curve. It
displays the energy dependence of the efficiency for one particular sample container in a specific
sample container to detector configuration.
0 018
^ 0.014
§ 0.012
•g 0.01
0 0.008
W 0.006
0.004
0.002
0
0 200 400 600 300 1000 1200 1400 1B00
Energy (keV)
Figure 23. Example of an HPGe Efficiency Calibration Curve
The parameters that affect the ability of this calibration curve to accurately determine the activity
concentration in a sample are:
• The height to which the sample container is filled;
• The density and chemical composition of the sample matrix;
• The physical composition of the sample container; and
• The thickness of the sample container,
when these parameters are compared to the working calibration source (WCS).
Once the analysis of the working calibration source (WCS) has been completed, the peak areas
for each of the gamma rays to be used for determining efficiency are compared to the gamma
activity of the WCS, as stated on the certificate received from the WCS manufacturer, using the
following equation:
Where:
8 is the efficiency for the gamma-ray peak of interest
Ns are the net counts in the gamma-ray peak of interest
ts is the WCS count time
70
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Ad is the certified activity at the gamma-ray peak of interest,
gammas/second45
Once all the efficiencies for the gamma rays selected for use in the detector calibration are
calculated, a curve is fitted to the data using one of the model equations in the software that is
selected by the user. It is important to review the shape of the curve and to assess the fit of
measured points to the curve to ensure it will be suitable for use. The importance of this is shown
by the following two figures. In Figure 24 a fourth-degree polynomial was selected to create the
calibration curve. While the overall shape of the curve parallels the response of germanium for
gamma rays, note especially that two of the points fall rather far below the smoothed curve. This
is particularly bad since it is in the region of the "knee" where the curve reaches a maximum and
the slope changes dramatically from positive to negative. In Figure 25 the same measured data
was fitted to a fifth order polynomial. In this case, the model equation used for fitting the curve
produces a smooth curve that parallels the response of germanium for gamma rays and all of the
points lie very close to the calibration curve. It is recommended that documented criteria for
acceptance of calibrations be defined in advance to ensure that the fitted curves adequately
reflect measured efficiencies. One approach that has been used is to establish a limit to the
acceptable deviation between measured and fitted values for each point in a curve.46
Dual Efficiency Calibration Curve
La* E'WCV
P l.'ulSulU
i
U
1
¦
P
¦
B
a
* 1
*
0 500 1M0 liCo £0C0 £500
Energy IkeV)
Figure 24. Fourth Order Calibration Curve
45 Note the units for the activity of the WCS are gammas per second and not dps, curies or another derived unit.
Thus the numbers of gamma rays interacting via the Photoelectric Effect are compared directly to the number of
gamma rays emitted by the WCS.
46 ASTM D7282, Standard Practice for Set-up, Calibration, and Quality Control of Instruments Used for
Radioactivity Measurements, describes such an approach (Reference 15).
71
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Figure 25 demonstrates that even when an acceptable curve is obtained by fitting measured
points, it is important not to use the efficiency calibration curve beyond the bounds of measured
points. In this case, the curve takes a dramatic upswing in efficiency after the final calibration
point, which does not reflect the physical response observed in an HPGe detector.
Empirical Efficiency Calibration Curve
B Measured - Calculated
_
¦
i
I
/
/
I
h
/
.
¦V
h
"V.
/
/
I
1
J
I
10 100 1000 10000
Energy (keV)
Datasource: DET08_8K
In(Eff) =-4.217e+000 + 7.494e-00rx -1.859e-002*xA2 - 1.486e-001*xA3
+ 1.225e-00rxM - 3.619e-002*xA5
where: x = ln( 9.478e+002/E)
Figure 25. Fifth Order Calibration Curve
Once the user is assured that the data produced by counting the WCS have produced a curve that
is satisfactory, the efficiency curve is verified to be correct with an independent source of known
activity. When this also produces results for each gamma-ray emitter in the second source that is
deemed acceptable, the calibration is complete. The user should determine acceptability using a
statistical model based on the uncertainty of both measurements (e.g., the Z-score).
K. Correcting Efficiency for Matrix/Geometry Differences
The sample container, the size of the sample (the fill-volume of the sample in the container), the
sample density, and the sample composition all have an impact on interactions of gamma rays
that will affect the number of gamma rays that reach the detector. The size of the sample
container, the sample and its fill volume in the container may lead to differences in attenuation.
These factors can be controlled by the laboratory staff by ensuring that samples are only
prepared and counted in geometries which closely match that of the calibration standards used
for the analysis.
Figure 26 shows efficiency curves for four different food-related matrices analyzed on the same
detector in the same container, sample size and juxtaposition to the detector, but for which the
density and the composition of the sample varies. The differences in efficiency observed,
especially at lower energies, are greater than thirty percent. These are attributable to differences
72
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
in self-attenuation associated with the density and to a somewhat lesser extent, the composition,
of samples.
5.5%
5.0%
?r 4.5%
Corncob - 0.4-5 g/cc
4.0%
Coffee -0.60g/cc
3.5%
Water Eq. - 1.15 g/cc
3.0%
Honey - 1.45 g/cc
2.5%
1.5%
1.0%
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Gamma Ray Energy (keV)
Figure 26. HPGe Detector Efficiency Curves for Four Solids
Empirical Corrections for Attenuation
The best method for calibrating detectors in a radiochemistry laboratory involves the use of
physical standards containing radionuclides (traceable to a National Metrological Institute such
as NIST in the United States) in a geometry that exactly matches that of the samples. When
differences in the size, density, and composition of samples being measured result in a large and
significant difference in detector efficiency, the detector should be calibrated using a calibration
standard matrix that exactly matches the levels, geometry, and levels of attenuation in the
samples being analyzed.
The traceability and reliability of measurements are one of the more important reasons to rely on
fixed laboratories. In contrast to field measurements, where mathematical approaches to
efficiency calibration may represent the only viable approach for performing measurements, the
fixed laboratory can more easily control different aspects of the measurement situation. The
laboratory can homogenize samples, place them in the same container using the same fill-
volume, and count them in a very tightly controlled juxtaposition to the detector. It can identify
and isolate between samples of different density and matrices. Finally, and most importantly, it
can obtain traceable standards in various containers, matrices, and densities and calibrate in
configurations that match those most commonly used for samples. However, economic and
operational factors will limit the number of configurations that a lab can afford to maintain.
73
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Mathematical, Modeled, and other Corrections for Efficiency and Attenuation
While empirically-based calibrations allow laboratories to obtain results that are defensible and
relatively accurate, it may be possible to improve the quality of results by generating and
applying correction factors to account and correct for minor differences between sample
geometries and the calibration standards. Two factors generally are outside of the control of the
laboratory, the density and matrix composition (i.e., average Z) of the samples, are the most
likely of these.
A number of approaches, some of which are integrated into gamma spectrometry software
packages, are available to accomplish this. These range from Monte-Carlo simulations with well-
characterized detectors and geometries, to manual or spreadsheet-based calculations to correct
for differences in attenuation due to density. In general, the corrections being applied should be
relatively small and should be minimally related to the skill and experience of the analyst. It
would not be advisable to rely on software to extrapolate efficiencies for a Marinelli beaker
geometry starting with a point-source geometry for the characterization of the detector since this
requires substantial skill for determining the relationship between the two geometries (e.g., to
build a model of a Marinelli beaker).
No matter what approach is used, it is important to ensure that the method/approach has been
validated using traceable standards that span or bracket the range of the corrections being
applied. The method should include an approach for estimating the measurement uncertainty and
the approach should consistently improve the accuracy of measurement results (i.e., combined
bias and uncertainty). Finally, it is important that quality control samples be used to show that
the performance of the approach is reliable and consistent with the documented results of the
method validation under the conditions applied.
L. Establishing sample/matrix specific libraries in the software
The main library that is often supplied with the original software package has hundreds of
radionuclides and their nuclear constants listed. However, using such a large library to analyze
each sample can prove to be counterproductive to obtaining a correct analysis of the sample as
many of the radionuclides cannot possibly be in that sample. A specific example is a sample
taken which contains fresh fission products. There are many radionuclides from the left side of
the Line of Stability47 that could not possibly be in the sample, yet they are listed in the main
library. It should be emphasized here again that the library should reflect the nuclides that would
be expected in the type of sample being analyzed.
Laboratory staff should be aware of the source (origin) of the sample and which gamma emitting
radionuclides are likely to be present. The library used for the analysis should consider the
following data regarding the sample:
• The specific matrix.
47 Radionuclides on the left of the Line of Stability are neutron deficient; fission products are neutron rich. See
www. nndc .bnl. gov/chart
74
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
• The potential list of radionuclides that can be present based on the event.
• The half-lives of the potential radionuclides with respect to the age of the sample.
• Specific requirements of the software package (e.g., peaks needed to resolve interferences).
Once these parameters are identified, the potential list of radionuclides to be included in the
sample-specific library will be minimized. Additionally, when short-lived radionuclides may be
present, it may be necessary to create a different library for different sample ages. A specific
example provides some of the thought processes that would go into such a sample specific
library.
Example:
A batch of air particulate and charcoal cartridge samples from a nuclear power plant
that had an atmospheric release about 5 days ago has been received by your laboratory.
The plant was in full operational mode and a spill of several hundred gallons of reactor
coolant occurred when the ventilation filtration was out of service. The air particulate
filter precedes the charcoal cartridge in the air flow. About 4.1xl07 cm3 of air sample
was passed through the collection device. The time period from sampling to the
laboratory is one day. Sample analysis turn-around time is to be one day.
The sample matrices are air particulate filter and charcoal cartridge. Each matrix should have
separate libraries since the physical forms of radionuclides that can be collected on each will be
different.
Table 6. Potential Radionuclides on Charcoal Cartridge
Radionuclide
00
00
l-t
85mK_r
85Kr
f§
00
00
"True"
Half-life
2.84 h
4.48 h
10.8 y
17.8
min
Radionuclide
135Xe
133mXe
133Xe
131mXe
133 J
131I
129|
"True"
Half-life
9.14 h
2.20 d
5.25 d
11.8 d
20.8 h
8.03 d
1.57xl07 y
Table 7. Potential Radionuclides on Air Particulate Filter
Radionuclide
f§
00
00
86Rb
91 Sr
91 y
92Sr
"True"
Half-life
17.8
min
18.6 d
9.63 h
58.5 d
2.66 h
Radionuclide
92y
137Cs
140Ba
140La
141 La
141Ce
143Ce
144Ce
"True"
Half-life
3.54 h
30.1 y
12.8 d
1.68 d
3.92 h
32.5 d
33.0 d
285 d
The primary consideration for the list of radionuclides for this type of incident is that they would
be volatile fission or activation products. Tables 6 and 7 contain the most likely candidates for
the library based on volatility, parent-progeny relationship, and half-lives greater than two hours
(less than two hours it becomes extremely unlikely that they would be observed since the
radionuclides would have undergone 72, 2-hour half-lives). Kr-88 and its progeny 88Rb are
eliminated (grey boxes) since they too would have gone through significant decay during that
time period. 1-129 is eliminated (cross-hatched box) due to its extremely long half-life (an
insignificant number of counts would be realized in the short time available to analyze the
75
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
samples). Because there is a large volume of air collected it is likely that there will also be some
naturally-occurring radionuclides in this sample such as 212Pb, 212Bi, 214Pb, 214Bi, 7Be, and 208T1;
all of which could be seen in both samples.
The next step would be to decide how many gamma-ray energies for each radionuclide need to
be included in the library. The main library will usually have all gamma-ray energies down to
about 1% abundance included. Consideration should be made for potentially using energies that
are down to 0.1% for this specific library.
The half-lives of progeny should be evaluated based on the timing of the sample following the
event. Specifically, 140Ba/140La which may be in Transient Equilibrium should be carefully
examined when decay correcting to sample data/time from the beginning of the count. All other
radionuclides present in the tables are in the No Equilibrium situation and their "true" half-lives
can be used.
Once the samples have been counted it is imperative that a thorough review of any unidentified
lines be performed to ensure that the basic assumptions regarding the release are correct (e.g.,
volatile fission or activation products and their progeny).
Creating a library that is specific for an event or specific to a sample type (and not using the
master library) is an extremely important step in minimizing incorrect identifications and
quantifications.
M. Gamma spectrometry report review processes (verification)
The analysis of a sample by gamma spectrometry yields a report containing activity results based
on the input parameters that were selected by the analyst. The data review process should ensure
that not only are the usual editorial inputs correct (i.e., date/time of the sample, date/time of
count, correct detector, correct library, etc.) but that the results for all of the reported
radionuclides make sense. The review should also consider:
• Sample related data are correct,
o Correct sample size entered,
o Correct density used.
• Impact on the results of the following selected preset software configurations,
o Correct calibrations used.
o Correct peak background subtraction used (if any),
o Correct summing and attenuation corrections used,
o Half-life ratio function value.
o Allowable energy difference between identified peaks and library values,
o Abundance limit selected (if used at all),
o Weighted mean average calculation,
o Correct date for decay correction used,
o Decay correction to start of the count interval,
o Decay correction to the end or midpoint of the sampling interval.
• Peak Search and Identification.
o The FWHM of all peaks used in the analyses are reasonable for the energy range that the
peaks were found.
76
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
o Deconvoluted peaks do not negatively impact quantification,
o Peak flags investigated including visual inspection of the spectrum,
o More than one radionuclide activity concentration is not calculated using the same data
from the same gamma-ray peak also ascribed to another radionuclide.
• Library.
o Correct library used - if a different library was used, rationale for change documented,
o Library correctly accounts for potential parent-progeny relationships,
o If the library was modified, changes made are documented and validated by a cognizant
reviewer prior to final reporting of results.
• Nuclide Identification and Quantification.
o Presence of a radionuclide is detected but not one that is closely associated with it (e.g.,
95Nb is detected but not 95Zr).
o Calculated, individual gamma-ray activities are in the appropriate ratios based on the
abundance of those gamma rays and the efficiency of the detector at their energies,
o Presence of unexpected radionuclides is investigated.
• Measurement quality objectives meet the clients' specifications/needs (or laboratory defaults
if none from the client).
o Critical level concentration or minimum detectable concentration,
o Method uncertainty requirements were met for each radionuclide,
o Results for requested radionuclides have been reported.
• Data Reporting.
o Results and uncertainty and units reported for all required radionuclides (and any other
parameters required by the client or laboratory default such as critical level),
o Actual values reported (not results such as "zero", "
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Table 8. Checklist (partial) for Gamma Spectrometry Data Verification
Sample
Sample Matrix Date/time Sample ID
Geometry Detector Count date/time
Y N
• Are all of the above inputs identified correctly
in the report?
• Are all identified radionuclides included based
on half-life?
• Have appropriate members of decay chains
been identified?
• Are proper half-lives used for radionuclides in
parent-progeny relationships?
• Are all the FWHM used to calculate activity
concentrations at the approximate value of the
gamma-ray energy?
• Are all identified radionuclides expected or
probable?
• Was the correct library used?
•
• Any "N" requires a description and resolution
The exact nature of the questions for this form are to be decided by the individual laboratory
staff in cooperation with the client, but clearly when any of the noted issues does not meet
expectations a description of why this occurred and how it may need to be resolved should be
part of the review process.
N. Data validation and reporting protocols
Data should always be validated before they are deemed acceptable for decision making. The
validation process will go through the entire sample history including the sampling process, preservation, and
storage as well as the parameters used to analyze the samples. All of these are important in gamma spectrometry as
the assumptions made about parent-progeny relationships and decay correction will all be tied into these actions. An
example of a data validation guide is shown in Table 9.
78
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Table 9. Example Data Validation Table
Project: Client:
Project QA Document: Analytical Laboratory Used:
Are the following satisfactory:
Sample COC? Y N
Sample Preservation? Y N
Sample holding time? Y N
For any "N" provide explanation:
All verification report inputs satisfactory? Y N
If "N" provide explanation:
All QC analyses Satisfactory? Y N
For any "N" provide explanation:
Have all software preset functions been optimized
based on the client requirements and sample history
to identify the radionuclides present? Y N
Client Requirements Met? Y N
Sensitivity Factor:
Half-life ratio:
Energy Difference:
Abundance factor:
Key line:
Weighted Mean:
Have all unknown gamma-ray lines with a cps 1-
sigma uncertainty less than 50% been identified? Y N
List all unidentified gamma rays:
79
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Quality system requirements generally specify two levels of validation: one from a technician
and one from a manager. Most software packages provide this option. Each facility should
procedurally establish expected criteria for software and analytical validation (electronic or
otherwise). This is another laboratory protocol that should be stated in the radiochemistry
Quality Manual.
0. References
Documents
1. U.S. Environmental Protection Agency (EPA). 1992. Manual of Protective Action Guides
and Protective Actions for Nuclear Incidents. Washington, DC. EPA 400-R-92-001, May.
Available at: https://www.epa.gov/sites/production/files/2016-03/documents/pags.pdf.
2. U.S. Environmental Protection Agency (EPA). 2002. Final Implementation Guidance for
Radionuclides, EPA 816-F-00-002. 40 CFR 141.26(a)(2)(iii). Available at:
https://www.epa.gov/sites/production/files/2015-
09/documents/2009_04_16_radionuclides_guide_radionuclides_stateimplementation.pdf
3. U.S. Environmental Protection Agency (EPA). 2004. Multi-Agency Radiological Laboratory
Analytical Protocols Manual (MARLAP). 2004. EPA 402-B-04-001A, July. Volume I,
Chapters 3, 6, Volume II, Volume III. Available at: https://www.epa.gov/radiation/marlap-
manual-and-supporting-documents.
4. U.S. Environmental Protection Agency (EPA). 2004. Response Protocol Toolbox: Planning
for and Responding to Drinking Water Contamination Threats and Incidents. Interim Final B
December. Office of Water. EPA 817-D-03-001 through EPA 817-D-03-007. Available at:
https://www.epa.gov/sites/production/files/2015-
05/documents/drinking_water_response_protocol_toolbox.pdf
5. U.S. Department of Energy (DOE). 2010. The Federal Manual for Assessing Environmental
Data During a Radiological Emergency, SAND2010-1405P, FRMAC Assessment Manual,
Volume 1, Overview, and Methods. Available at: https://energy.gov/nepa/downloads/ea-
1405-final-environmental-assessment.
6. U.S. Food and Drug Administration (FDA). 1998. Accidental Radioactive Contamination of
Human Food and Animal Feeds: Recommendations for State and Local Agencies. 13 August.
Available at: https://www.fda.gov/downloads/MedicalDevices/.../UCM094513.pdf.
7. U.S. Department of Health, Education, and Welfare (HEW). 1970. Radiological Health
Handbook, p. 123. National Nuclear Data Center, Brookhaven National Laboratory. Available
at
https://www.orau.org/ptp/PTP%20Library/library/Subject/health%20physics%20general/RH
H.pdf.
8. Genie™ 2000 Customization Tools Manual, Version 3.3, Canberra Industries
9. U.S. Nuclear Regulatory Commission (NRC). 1978. Offsite Dose Calculation Manual
Guidance: Standard Radiological Effluent Controls for Pressurized Water Reactors.
NUREG-1301. Available at http://www.nrc.gov
10. U.S. Nuclear Regulatory Commission (NRC). 1980. Radiological Effluent and
Environmental Monitoring at Uranium Mills. Regulatory Guide 4.14. Available at
http://www.nrc.gov
11. ORTEC a subsidiary of AETEK. 2010. GammaVision®-32 Gamma-Ray Spectrum Analysis
and MCA Emulator. ORTEC Part No. 783620 0910, Manual Revision G
80
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
12. U.S. Environmental Protection Agency (EPA). 2012. Rapid Methodfor Fusion of Soil and
Soil-related Matrices prior to Americium, Plutonium, Strontium, and Uranium Analyses.
EPA-600-R-12-636, -600-R-12-637, or -600-R-12-638. Available at
https://www.epa.gov/sites/production/files/2015-
06/documents/soil_dissolution_by_fusion_for_am_pu_u_09-17-12_epa-600-r-12-
636637638.pdf
13. U.S. Environmental Protection Agency (EPA). 2009. Radiological Laboratory Sample
Screening Analysis Guide for Incidents of National Significance. EPA 402-R-09-008.
Available at: https://www.epa.gov/sites/production/files/2015-05/documents/402-r-09-
008_sample_screening_guide.pdf
Consensus Standards
Some recognized standards (i.e., published by a standards organization either nationally or
internationally) and some published in the refereed literature for the analysis of radionuclides in
soil, are included below. The American Society for Testing and Materials (ASTM) methods may
be purchased online from www.astm.org. However, in every case, the individual laboratory must
perform method validation in their own laboratory for a soils matrix to ensure that the results will
conform to the needs of the incident.
14. ASTM C1402-09. (2009). Standard Guide for High-Resolution Gamma-Ray Spectrometry of
Soil Samples.
15. ASTM D7282. (2006). Standard Practice for Set-up, Calibration, and Quality Control of
Instruments Usedfor Radioactivity Measurements
16. ANSI N42.14. (1999). Calibration and Use of Germanium Spectrometers for the
Measurement of Gamma-Ray Emission Rates of Radionuclides
17. JCGM 100:2008. Evaluation of measurement data — Guide to the expression of uncertainty
in measurement (GUM, Revised 2008)
18. ASTMD7902. (2014). Standard Terminology for Radiochemical Analyses.
Other References
19. Brookhaven National Laboratory National Nuclear Data Center website,
www.nndc.bnl.gov/chart/
20. Bureau National de Metrologie, Laboratoire National Henri Becquerel. (1999). Table of
Radionuclides. France. ISBN 2 7272 0201 6.
21. Debertin, K. & Helmer, R.G. (1988). Gamma-and X-ray Spectrometry with Semiconductor
Detectors. K. Debertin and R.G. Helmer
22. "Practical Gamma-Ray Spectrometry 2nd Edition", Gordon Gilmore. John Wiley & Sons,
Ltd., 2008.
23. "Radiation Detection and Measurement", Glenn F. Knoll, John Wiley and Sons, 1979, page
739.
24. "A Ba-133 Loaded Charcoal Cartridge as a Counting Standard for 1-131"D.G.Olson,
J.S.Morton, C.D. Willis, Int. Appl. Radiation. Vol.35, pp. 574-577, 1984.
25. "Gamma-ray Spectrum Catalogue, Ge and Si Detector Spectra", Fourth Edition, Idaho
National Engineering & Environmental Laboratory, y-Ray Spectrometry Center (March
1999).
26. "Handbook of Radioactivity Analysis", Third Edition, Michael F. L'Annunziata, Academic
Press (2012).
81
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
27. "Demonstration of Coincidence Summing Effects Observed with N-Type Germanium
Detectors in the 20- to 40-keV Energy Range When Counting 1291, 1251, and 125Sb", in The
Counting Room: Special Edition, Radioactivity & Radiochemistry, Vol. 4, No. 2, 1993.
28. "The Use of Sources Emitting Gamma-rays for Determination of Absolute Efficiency Curves
of Highly Efficient Ge detectors", NIM A322, 1993, pp. 483-500
82
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
ATTACHMENT I: INSTRUCTIONS TO LABORATORIES ANALYSIS OF FRESH
MIXED FISSION PRODUCT PROFICIENCY TEST SAMPLES ROUND III -
DECEMBER 2012
I. Introduction
EMS, as a contractor to the U.S. Environmental Protection Agency, is providing your laboratory
with the third in a series of irradiated uranium proficiency testing (PT) samples as a simulation of
the type of environmental water sample that might be encountered following a fission event (e.g.,
an improvised nuclear device (IND) or reactor accident event). The mission of this program is to
provide laboratories with an opportunity to analyze samples that contain a complex mixture of
radionuclides that are representative of what might occur following a fission event. This
objective of the study is for these PT samples to allow laboratories to identify areas of strength
and improvement in their analyses and reporting of such complex radioanalytical samples.
II. PT Sample Description
The sample you will receive directly from our supplier48 will contain fission and activation
products resulting from a one-hour neutron irradiation of uranium oxide (U3O8). The oxide will
have undergone irradiation in early November 2012. The PT provider will prepare a stock
solution by dissolving the oxide in nitric acid. The solution will be stabilized with an iodate
carrier. Within several weeks of the irradiation, you will receive a proficiency test (PT) sample
of diluted stock solution.
The reference date (i.e., sample collection date) for decay correction purposes, as well as the
irradiation date (i.e., date of the "event"), will be sent to participating laboratories by EMS
PRIOR to the laboratories receiving the samples.
The sample is assumed to be surface water from an estuary that has been affected by the event.
The reference date (i.e., sample collection date) will be at least two weeks after the irradiation of
the oxide.
The activity will be significantly greater than environmental sample matrices that are normally
encountered but will not exceed 0.1 microcuries (0.1 /ud) total activity. The PT provider will
verify prior to shipment that the level of radioactivity in the samples is consistent with inventory
restrictions established under each facility's radioactivity materials license (or the equivalent
inventory restriction at DOE facilities). The sample will also contain fission products that are not
gamma emitters such as 90Sr and 239+240pu (in sub-nanogram quantities), and sub-milligram
quantities of uranium isotopes. The concentrations of these non-gamma emitting radionuclides
that will be used as "accepted values" for general comparisons will be calculated from
mathematical models.
48 Eckert & Ziegler Analytics, 1380 Seaboard Industrial Boulevard, Atlanta, GA 30318. Samples will be shipped
overnight express on or about December 5, 2012; you will receive advance notice of the actual shipping date.
83
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
III. Sample Analysis Instructions
Gamma-ray spectrometry analysis is requested of all study participants. Additionally, the
determination of 90Sr, 89Sr, and 239 + 240Pu activity concentrations should be performed by
laboratories who maintain capabilities for these analyses. The laboratory should use the same
approved processes and procedures that would be used for receipt, handling, analysis, and
reporting for a similar sample during an incident response. Deviations from approved procedures
that are necessary to comply with the requirements of this SOW should be documented in the
case narrative.
Prior to receiving the sample, we recommend that the laboratory have a process by which
summing effects can be minimized (e.g., counting samples several inches from the detector
surface, or diluting the sample), and that process should be described in the case narrative.
The sample will be shipped in nitric acid solution so that the sample pH is less than 2.0. Upon
receipt, the laboratory's protocol for screening samples shall be used. Reporting of the screening
results (either as count rate, dose rate or activity) for the sample and its package upon receipt is a
required deliverable for this study. These screening measurements should be converted into a
pCi/L concentration for the sample. If the laboratory does not have a method for doing this, that
should be stated in the report.
Dilution of the sample may be performed if your laboratory protocol requires dilution, and your
available sample geometries provide adequate sensitivity to meet the combined standard
uncertainty requirement stated below. Nitric acid must be used for this dilution, maintaining a
pH of less than 2.0 during the dilution process. Note that radioiodine will be carried in the
solution by added stable iodate. It is strongly recommended that the laboratories provide
adequate means to ensure all radionuclides remain soluble during sub-sampling and throughout
the gamma analysis counting interval. Based on the laboratory's experience, chemical additives
may be an acceptable method of ensuring that all components remain soluble.
IV. Gamma Spectrometry
The gamma spectrometry sample results should be reported based on a single count.
NOTE: This is a change from the previous studies.
• The sample count date will be specified by e-mail prior to the arrival date (within 1-2
days) of the sample.
• For each laboratory, the sample should be counted only once and the results of that count
decay corrected to the "sample collection date" and reported; no averages from multiple
detectors or multiple sample counts are to be reported.
• The laboratory is to count the sample for gamma spectrometry analysis so that relative
combined standard uncertainty (1-sigma CSU, i.e., k= 1) for 147Nd is less than or equal to
13%. The count time may be adjusted and the count repeated to achieve this value. Please
note that it is not intended to limit the count time for the sample.
• Upon achieving the relative CSU specification for 147Nd, all radionuclides in the sample
which have less than 50% counting uncertainty (1-sigma, k=\\ and are considered
"detects" (no "U" qualifier) should be identified, and the measured activity and
associated combined standard uncertainty (1-sigma, k=1) shall be reported.
84
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
o It is understood that all peaks with > 50% counting uncertainty may not be used in the
quantification of the radionuclide of interest. For peaks that are not used in the
quantification of a radionuclide that is otherwise being reported, the raw data should
be annotated or the report or case narrative should correctly attribute the peak to the
source radionuclide(s) of interest.
• In cases where specific radionuclide concentrations are present due to parent or progeny
activity but may be indirectly determined from gamma spectrometry results, the indirect
radionuclide activity concentration and combined standard uncertainty should be
reported. Assumptions of equilibrium (e.g. secular, transient, etc.) should be noted in the
report or the case narrative.
• Non-routine activity calculations, such as those addressing non-equilibrium ingrowth
factors, which are not performed by the gamma spectrometry software, should be
included in a separate Excel spreadsheet or other document attached to the report.
The reference date (i.e., sample collection date) for this sample and the irradiation date
(i.e., the date of the "event") will be sent to participating laboratories by EMS prior to the
laboratories receiving the samples. NOTE: The method of decay correction to the sample date for
some radionuclides may require a method not commonly used by the laboratory's gamma
spectrometry software.
V. Reporting Requirements
Two documents are necessary to evaluate the laboratory's data and information. EMS requires
only emailed files: paper copies are neither required nor desired.
Document 1. The following information shall be included in an Excel spreadsheet (to be
provided by EMS upon confirmation of your participation in this study) for the gamma
spectrometry, radiostrontium, and plutonium analyses measurements.
Do not modify the format of the spreadsheet or add any other information besides:
• The Laboratory Name (as specified on the shipping papers for the source to the
laboratory), and sample identification.
• Final analytical results and combined standard uncertainty (1-sigma, k = 1) for each
radionuclide identified on the lines dedicated for each radionuclide in the top section of
the worksheet.
• The activity reference date (i.e., the "sample collection date") has been filled in. Please
do not alter this date. Fill in the date /time of the count for each of the different analyses
performed. For radiostrontium, analysis use the first count date/time if two are used.
• Any gamma spectrometry result that the laboratory would qualify as a "non-detect" (e.g.,
a "U" qualifier) should not be included on the spreadsheet, even if space is provided for
that analysis.
• The bottom section of the worksheet will be for those detected radionuclides that are not
listed on the top section. Fill in the radionuclide identity and, again please include only
the data for those radionuclides that would be considered a "detect".
85
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Results from screening data should not be included in the Excel worksheet.
Document 2. A more detailed document should be used to transmit the final analytical results
and the requirements identified in this document for each analysis, respectively (a single pdf file
that combines all the submitted information is preferred).
The results of the gamma spectrometry analysis should be e-mailed to EMS within one week of
the analysis date of the sample. EMS requires only an e-mailed pdf file: paper copies are neither
required nor desired.
The analytical values for the gamma spectrometry measurements shall be reported with the
following information:
1. A summary report of the final activity concentration and the associated combined
standard uncertainty (coverage factor k = \) for each radionuclide identified in the sample,
in pCi/L, using appropriate rounding methods, decay corrected to the date on the
documentation received from the PT sample supplier (the reference date).
2. A description of any sample pre-treatment performed, including additional preservation,
filtration, or steps taken in the laboratory to ensure quantitative transfers from the original
container or representative sub-samples for the individual analyses.
3. Volume of sample counted for each of the analyses. Dilution process if sample was
diluted. The reported combined uncertainty shall reflect uncertainty from volumetric/
gravimetric dilutions.
4. Gamma spectrometry raw data printout with:
a. Summary of all peaks located by the peak search algorithm - should include peak
energy, FWHM, background activity at that peak and any qualifiers that the software
uses to identify gamma rays that may not be acceptable for quantitation.
b. Activity and CSU for each individual gamma-ray peak used in the computation of the
final reported activity.
c. A list of all gamma-ray peaks located by the software routine but not identified during
analysis or review of the spectrum.
5. In addition to the information presented in the gamma spectrometry raw data printout:
a. User selectable analysis options should either be reported in the raw data instrument
printout or narrated. This should include: peak search sensitivity, half-life ratio limit,
peak matching energy tolerance, background, and modifications to peak-search
algorithms based on the sample activity.
b. If any modifications to routine parameters were performed, these should be addressed
in the narrative. This should include coincidence summing corrections, nuclide
interference calculations, and algorithms used for parent/progeny calculations.
6. Gross activity screening results (dose or cpm and the algorithm used to convert to activity
and the reference nuclide used to calibrate the screening instrument), uncertainty, units
(pCi/L), instrument(s) used, and a description of any safety precautions taken as a result
of the screening process.
The reports do not need to include the following information, as it is not used to analyze
laboratory results:
• Calibration spectra
86
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
• Calibration certificates
• Quality control sample results (specifically do not include results for LCS, instrument
blanks, or duplicates)
• Method blank sample results
• Instrument performance checks
• Certifications for various agencies
Please do not include any of this information with your reports.
87
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
ATTACHMENT II: EXAMPLES OF RADIOANALYTICAL DATA REVIEW AND
REPORTING PROBLEMS
The data presented in this attachment were taken from reports submitted for various types of
samples. Each example starts with a figure. Each figure is followed by a discussion of issues that
arose with the analysis which include at least some of the following:
• Half-life
• Equilibrium type
• Library assignment
• Preset analysis configuration
• Identification of energy associated with more than one radionuclide
• Sample geometry
• Detector geometry
• Preservation of the sample when it was analyzed
In all the examples presented here the issues identified have been appropriately addressed by the
contributing organizations.
Example 1: Analysis of 140Ba/140La in Rainwater Samples- Incorrect Use of Decay
Correction and Equilibrium
- Sigma 2 Sigma
Isotope Ruii Date Qualifier Activity Uncertainty MDC LLD TPU Units
"Ba-140 03728/11 U I43E-5T i".'86E+00 T2QE+00 T.'50EK)i Offi^OO pCi/L
La-140 03/28/11 U 2.43E-01 1.S6E+00 3.20E+00 1.50E+O1 1.S6E-00 pC'iL
Figure 27. Incorrect Equilibrium Assumed for 140Ba/140La
The above results are excerpted from an analysis report on rainwater samples taken on March 23,
2011, received by the contract laboratory on March 25, 2011, and analyzed on March 28-29,
2011 following the Fukushima event (which occurred on March 11, 2011). Sample activity was
to be corrected to the date of sampling. The "U" qualifier denotes that the MDC for each
radionuclide was calculated, and the measured activity is less than the MDC. Ba-140 and 140La
form a transient equilibrium pair and assuming that there were no 140La initially present and no
environmental fractionation, after approximately 5 days the transient equilibrium activity ratio of
140La/140Ba would be 1.15. In this analysis the same gamma ray, 1596 keV, was used for the
analysis of both radionuclides; it is the most abundant gamma ray for 140La. A half-life of 140Ba,
12.8 days, was used to decay correct the 140La activity concentration back to the date of
sampling. The activity concentration of the 140Ba was determined inferentially by assuming
secular equilibrium with the 140La (based on the 1596 keV peak area).
However, the report did not take into account:
1. That Ba-140 was not measured rather an assumption of secular equilibrium with 140La was
made.
2. That making a decay correction for 140La (decay progeny) prior to the date of collection
assumes that equilibrium had been achieved during the aerosol transport across the Pacific
88
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Ocean. Radioactive equilibrium, however, may not have been achieved at the date of
sampling.
3. That the activity concentration of 140La at transient equilibrium is higher than that of 140Ba.
Concerns listed in #1 and #2 notwithstanding, calculating parent activity based on
measurement of progeny in equilibrium, would require applying the correct factor to account
for the ratio of parent and progeny at equilibrium (1.15).
4. If any sample preservation was performed.
5. Using two different detection criteria (LLD and MDC) is very confusing to any reviewer.
Why would two be necessary?
6. Use of the critical level concentration as the detection criterion49.
How could this analysis and reporting have been improved?
The laboratory could have notified the customer that, although they could decay correct the
parent nuclides to the date and time of sample collection, it is not possible to properly decay
correct results for 140La to the date of collection unless they assume that the pair had been in
equilibrium at the time of collection. If the customer agreed with such an assumption, they
should decay correct the result for the progeny to the time of collection using the half-life of the
parent, clearly qualify the data, and note in the narrative the assumption that was made.
In any case, the activity, uncertainty, and critical level/MDC for 140Ba should have been
determined separately from 140La using the most abundant gamma ray for 140Ba at 537.3 keV
(potentially with confirming lines at 162.7, 304.8, and/or 423.7 keV).
Finally, the data review process should also include a check of the reasonability of
measured/reported results to verify that the measured data for the parent/progeny pairs conform
to the ratios expected at equilibrium.
49 The MDC is an a priori determination of activity concentration that can potentially be achieved. It does not take
into account gamma ray interference corrections, or different backgrounds than a blank matrix sample. The critical
level concentration should be used for 'detection' decisions (although in this case it did not matter).
89
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 2: Results from the Irradiated Uranium PT
Laboratory
Activity
Concentration,
pCi/L
Activity
Concentration, pCi/L
140La
140Ba
Measured
Ratio/Theoretical
(progeny/parent)
"Mo
Measured
Ratio/Theoretical
(progeny/parent)
1
Activity1 at
the start of
the counting
interval
1980
1879
1.05/ 1.13
—
—
—
Corrected
for decay
back to time
of collection
207,000
3457
59/1.00
l.OxlO26
5.03xl07
2x1022/0.96
2
Activity1 at
the start of
the counting
interval
—
—
—
—
—
—
Corrected
for decay
back to time
of collection
2.49xl06
8.97xl03
2.78xl02/1.00
4.17xl019
2.59xl03
1.6xl016/0.96
The results signified by "—" were reported only with decay corrected values so the activity concentrations at the
start of the counting intervals were not available.
Figure 28. Parent Activity Inferred from Progeny; Progeny Half-Life Used for Its Decay Correction
Problems with results were found in the Irradiated Uranium PT results presented above (see
Attachment 1 for details). These involved:
• making incorrect assumptions about equilibrium that resulted in the incorrect use of half-life
for decay correction, and
• incorrectly performing an inferential determination of activity - one radionuclide was
measured and the activity of the other radionuclide reported based on an assumed secular
equilibrium concentration (which was incorrect as they are both transient equilibrium cases).
These results are shown in Figure 28.
The laboratory was instructed to decay correct results to the date of collection (and preservation).
As noted in previous examples, 140Ba/140La and "Mo/"mTc form transient equilibrium pairs. If
stable chemical conditions exist for approximately 15 days 140Ba and140La will reach
equilibrium, while approximately 2 days are needed for "Mo and "mTc. Once equilibrium has
been reached, the ratio of parent to progeny will approach the value characteristic of transient
equilibrium for the pair. If the sample has reached equilibrium at the time of the count, it is no
longer possible to make conclusions about what the activity would have been prior to the time of
the count without making an assumption about when the sample reached equilibrium.
In this case, both pairs had sufficient time between the collection date and the count time to
effectively reach transient equilibrium (after ~11 days, assuming no 140La at day zero, the
140La:140Ba ratio would be 1.13).
90
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
The principal gamma rays used for analysis of 140Ba and 140La were 487.0 keV and 1596.2 keV
respectively (both lines are actually 140La). The "mTc line at 140.5 keV was used to measure
both "mTc and "Mo.
Although sufficient time had passed between sample collection and the count time for the
samples to reach equilibrium, there would be no way to know that physical or chemical
processes had not interfered with equilibrium prior to the preservation of the sample. Thus, the
laboratory would have to assume that the pairs were in equilibrium at the point of collection if
they wanted to decay correct the progeny activity back to the time of the event using the half-life
of the parent. If on the other hand, the laboratory decay corrected the progeny activity using the
half-life of the parent, they would have to assume that the sample had been in equilibrium at the
point of collection. In some cases, this assumption is reasonable (e.g., "Mo and short-lived
"mTc) while in other cases, the situation may be less clear (e.g., 140Ba and the longer-lived
140La).
Prior to making such an assumption, it would always be advisable to contact the customer and let
them know that decay correction would require an assumption. It would be imperative that, if the
decay correction were performed, that the laboratory qualify the results and clearly document
that they have made an assumption (e.g., in the case narrative).
In this case, however, the decay corrected progeny activity was calculated using the half-life of
the progeny even though the progeny activity was supported by a parent. The resulting decay
corrected values for 140La and "mTc were extremely high and clearly unrealistic. Had the half-
life of the parent been applied to the decay progeny, the ratios would have been within 10% of
the theoretical ratio. While it would never be acceptable to decay correct the progeny results
assuming they were unsupported, it might not be unreasonable to assume transient equilibrium
was in place at the time of collection.
A second problem resulted since the laboratory used gamma rays emitted by the progeny to
inferentially determine the activity of the parent. Even assuming that the pair was in equilibrium
at the time of collection (and count), the laboratory incorrectly assumed that the pair were in
secular equilibrium, and failed to adjust the activity of the parent to correctly reflect the expected
ratio of activity of progeny to parent if equilibrium were established.
In both cases, however, using gamma rays directly attributable to the parent would have avoided
the need to make an inferential determination based on the parent based on the activity of the
progeny. Thus, the results for 140Ba could have been determined using the most abundant gamma
ray for 140Ba with at 537.3 keV (potentially with confirming lines at 162.7, 304.8, 423.7 keV).
Since "Mo emits a potentially interfering gamma ray at 537.8 keV (0.0033%), it would be
important to enable interference corrections and to include the more abundant and reliable "Mo
line at 739.5 keV in the library used for interference correction (even though the levels of
interference of the low abundance "Mo gamma ray with the higher abundance gamma ray from
140Ba may be minimal).
Finally, the data review process should also include a check of the reasonability of
measured/reported results to verify that the measured data for the parent/progeny pairs conform
to the ratios expected at equilibrium.
91
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 3: Analysis of Miner's Lettuce Following the Fukushima Event-
Incorrectly Identifying Radionuclide Gamma Rays
Configuration
Sample dare
sample ID
Detector name
Elapsed live time
Energy tolerance
Abundance limit
Nuclide
3E-7
K-40
1-131
DKA100:[CANBERRA.GAMMA.ARCHIVE.GAMMA]G274875003.CNF;1
TE-132
CS-134
CS-136
3A-137M
CS-137
31-214
RA-226
TH-230
U-234
CF-251
27-MAR32OH 18:00:00 Acquisition dare
Sample quantity
Detector geometry
Elapsed real time
Analyst Initials
Sensitivity
G274875003
GAM11
0 08:20:00.00
1.50000 keV
75.00000
Nucli.de Line
31-MAR-2011 19:22:15
4 . 72110E+02 GRAM
500MLMB
0 08:20:04.94 0.0%
MXR1
5 .00000
energy
477.60
1460.82
80.19
284.31
364.49
636.99
49.72
111.76
116.30
228.16
563.25
569.33
604.72
795.86
801.95
1365.19
153.25
176.60
273.65
340.55
818.51
1048 .07
1235.36
661.66
661.66
609
1120
1764.49
609.32
1120
1764.49
609.32
1120.29
1764.49
609.32
1120
1764.49
177.52
227.38
285.41
Area
1662
2646
332
3965
230
178
270
447
2868
2124
225
84
148
74
125
214
2805
2805
69
11
69
11
69
.32
,29
Activity Rersort
Uncorrected Decay Corr
%Abn pCi/GRAM pCi/GRAM
10.44 9.633E-01 1.018E+00
10.66 3.894E+00 3.894E+00
2 .62 Line Not Found
6.12 2.254E-01 3.247E-01
81.50 2.405E-01 3.465E-01
7.16 2.455E-01 3.537E-01
15.00 Line Not Found
1.74 Line Not Found
1.96 Line Not Found
11
69
11
148
178
332
88.00 7
8.34 2
15.37 2
97.62 2
76.91 2
8.69 2
3.02 4
5.75
10 .00
11.10
42.20
99.70
80.00
20.00
89.90
85.10
45.49
14 .92
15! 30
45.49
306E-03
235E-01
022E-01
147E-01
553E-01
407E-01
120E-01
1.826E-02
2.244E-01
2.029E-01
2.155E-01
2.563E-01
2.416E-01
4.136E-01
Line Not Found
4.661E-02 5.825E-Q2
2.713E-02 3.391E-Q2
1.394E-02 1.742E-02
2.037E-02 2.545E-02
1.810E-02 2.261E-02
Line Not Found
2.458E-01 2.459E-01
2.597E-01 2.597E-01
1.110E-02 1.110E-02
Line Not Found
1.280E-02 1.280E-02
1.110E-02 1.110E-02
14.92 Line Not Found
15.30 1.280E-02 1.280E-02
45.49 1.110E-02 1.110E-02
14.92 Line Not Found
15.30 1.280E-02 1.280E-02
45.49*1.110E-02 1.110E-02
.14.92 Line K:r rcur.d
15.30 1.280E-02 1.280E-02
17.30*2.694E-02 2.694E-02
6.80 9.455E-02 9.455E-02
1.13 1.221E+00 1.221E+00
Figure 29. Miners Lettuce Following Fukushima Event
The following concerns were noted on the final "Nuclide Line Activity Report" for the analysis
of Miner's Lettuce following the Fukushima event (Figure 29):
251
1. The 177.52 keV gamma ray is used to identify Cf in the sample. It was not possible for
2MCf to be present in this sample. This is a library configuration error. Including nuclides in a
library that are impossible or extremely unlikely to increase the risk of false detection of
radionuclides.
92
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
2. The same peak area for a single gamma ray at 176.60 keV is used for both 136Cs and 251Cf. If
two lines with the same energy are in a library and the interference correction is not enabled
or is configured improperly, activity calculations will be incorrect. In this case, the activity is
double counted by being attributed to two different nuclides.
3. The % abundance of the gamma ray at 661.66 keV is different for 137mBa and 137Cs. The beta
decay branch to the 661.66 keV excited state is 94.7%, however, in order to correctly
calculate the activity of the 137mBa from the gamma emitted from that state, the reduced
abundance (89.9% of the 94.7%) due to internal conversion must be taken into account;
85.10%.
234 226
4. The same gamma rays are used for U, Ra, and 230Th and are not due to any of these
214
radionuclides but to Bi. Note that all these radionuclides have exactly the same activity, so
it appears that the laboratory was assuming that the uranium decay chain was present in
secular equilibrium. Since there is no basis for assuming secular equilibrium of the uranium
chain in this matrix, these values are not defensible.
5. The sample transport time was 4 days. The non-volatile radionuclide, Te-132, is correctly
identified. However in 4 days the progeny, 132I should have been in transient equilibrium
with the 132Te. No gamma rays for 132I were identified in the spectrum (one gamma ray from
132I at 667 keV was detected but is in the unidentified lines list). In fact, during the 8-hour
count interval the ratio of progeny to parent would have been close to 1.0; almost
equilibrium. In this instance, the problem is lack of proper sample preservation. While it is
impossible to know with certainty what happened prior to sampling, had the sample been
preserved, any 132I that was present would not have been able to volatilize as elemental
iodine vapor prior to or during the count and iodine may have been determined as being
present.
6. The abundance (or fraction) limit for this analysis was set at 75%. In other words, the
summed abundance of lines identified in the sample divided by the sum of the abundance of
all lines in the library must be 0.75 or greater before the software would consider the
radionuclide identification to be valid. Using the library shown here and a limit of 75%, the
software will conclude that 136Cs is not present in the sample until three lines are identified
including one that has only 42.2% abundance. In contrast, 131I or 132Te will require a single
line with 82% or 88% abundance, respectively, to be considered valid.
Cs-134
Lines
Abnd.
to ID
97.62
46%
76.91
83%
15.37
90%
8.69
95%
8.34
99%
3.02
100%
1-131
Lines
Abnd.
to ID
81.5
84%
7.16
91%
6.12
97%
2.62
100%
Te-132
Lines
Abnd.
to ID
88 82%
15
97%
1.96
98%
1.74
100%
Cs-136
Lines
Abnd.
to ID
99.7
37%
80
67%
42.2
83%
20
90%
11
94%
10
98%
5.75
1
Bi-214
Lines
Abnd.
to ID
45.49
60%
15.3
80%
14.92
100%
This test proved to be unreliable for determining the presence of Cs. It should probably
have either been avoided (i.e., turned off), and one of the following functions enabled: a
nuclide specific abundance limit, or key-lines to designate which lines are needed to
constitute a valid identification.
93
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
7. An additional problem with the library usage is noted in the 'Energy' and 'Area' columns of
Figure 29. Note that the following gamma ray and the same area have been used to calculate
the activity of two different radionuclides: 177 keV with a peak area of 148 calculates an
activity for both 136Cs and 251Cf. Cf-251 is not possible to be found in this sample and should
not have been included in the analysis library. Also, note that the 177 peak has a significantly
different activity and uncertainty for the 136Cs activity as compared with its more reliable and
abundant lines. Clearly, the 177 keV peak should not be used in the final analysis of 136Cs.
Example 4: Unidentified Gamma Rays from Insufficient Library
Unidentified Gamma Rays
Energy
Area
Bkgnci
FWHM
€3.42
18
941
0.94
£€7.43
114
169
1.33
722.85
42
164
1. 56
1400.34
106
42
1. 97
1745.53
17
14
3.40
Figure 30. Unidentified Gamma Rays from Miners Lettuce
132
1. The 667.43 line from I was not identified due to an incorrectly constructed library.
2. The library listed the key line for 132I as the 772.6 keV gamma ray which is a less abundant
gamma-ray energy. Although 132I was present, it was considered not detected since the 772.6
keV gamma ray was not detected. The laboratory staff should have identified this problem
when reviewing unidentified peaks, but their review process failed to properly attribute the
activity at 667.43 keV to 132I.
134
3. The 1400 keV photon is a coincidence sum peak from Cs. This should be included in the
library as a "sum peak" or at least should have been noted as such in the laboratory's raw
data.
4. The 1745 keV peak may be eliminated as a real gamma ray due to a wide FWHM (3.4 keV).
Given the low number of counts, this can be attributed to poor fitting of data for the
statistically weak peak.
5. In general, all unidentified lines should be resolved by the analyst or reviewer prior to
sending out a final report. Any gamma rays that cannot be attributed to a radionuclide, an
artifact of the other gamma rays in the spectrum, or determined to be of poor quality or
exceptionally high uncertainty should be entered into the laboratory's condition reporting
system to be researched and resolved.
94
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 5: Incorrect Radionuclide Equilibrium Used to Analyze Rainwater
Samples from Fukushima
Sample Date
Analysis Date
1-132
(iCi/mL
Te-132
(iCi/mL
Ratio
132I/132Te
3/19/2011 9:00
3/22/2011
(8.52 ± 3.25)xl0"9
(1.18 ± 0.40)xl0"8
7.22E-01
3/21/2011 11:00
3/24/2011
(1.92±0.49)xl0~8
(2.40±0.66)xl08
8.00E-01
3/21/2011 10:00
3/24/2011
(2.27±0.39)xl0~8
(2.83±0.59)xl0~8
8.02E-01
3/21/2011 8:00
3/23/2011
(2.04±0.48)xl0~8
(2.49±0.55)xl08
8.19E-01
3/24/2011 14:36
3/28/2011
(1.63±0.46)xl0~8
(2.49±0.73)xl08
6.55E-01
3/24/2011 14:30
3/28/2011
1.99±0.45)xl0~8
(2.56±0.66)xl08
7.77E-01
3/24/2011 8:00
3/28/2011
(8.35±6.09)xl09
(1.25±0.54)xl08
6.68E-01
3/25/2011 9:30
3/30/2011
ND
(1.96±0.72)xl08
0.00E+00
3/25/2011 9:00
3/30/2011
(2.32±0.70)xl0~8
(2.50±0.74)xl0~8
9.28E-01
3/25/2011 8:00
3/30/2011
(6.76±5.61)xl0~9
(1.04±0.52)xl08
6.50E-01
3/28/2011 9:13
3/31/2011
(7.19±3.64)xl09
(6.98±6.35)xl09
1.03E+00
3/28/2011 9:00
3/31/2011
ND
(8.08±4.30)xl09
0.00E+00
Figure 31. Rainwater Samples Following Fukushima Event
This table shows an example of radiochemical equilibrium disturbed by chemical effects in
rainwater samples collected following the Fukushima incident and analyzed by gamma
spectrometry within a few days of being taken. It may have been advantageous to count these
samples somewhat longer to reduce the uncertainties which range from about 35 to 90%.
The theoretical ratio for 132I/132Te is 1.03 when equilibrium has been established. The samples
were not preserved and sent in plastic containers. The time between sampling and analysis was
more than sufficient for transient equilibrium to have been established in the sample. At the time
it was assumed that there would be insignificant losses for such a short transport time and
subsequent analysis. The low values for the ratios are likely due to loss of iodine either due to
volatilization or deposition on the container walls as h.
Following the Fukushima event, several air samples were analyzed by gamma spectrometry. The
transient equilibrium pair 140Ba/140La was determined to be undetected. However, the gamma
spectrometry report50 in each case had the same exact values for calculated activity
concentration, uncertainty, and LLD (to three significant figures) for both of these radionuclides.
This is an extremely improbable situation. As can be seen in Table 4, the activity ratio should be
1.15 for 140La/140Ba when equilibrium has been established. The reason this occurred was that the
library contained interfering lines and the software had not been configured to perform
interference correction so that these two activity calculations used the same peak area(s) data to
calculate the activity. In addition, the secular equilibrium case was assumed (in the library)
instead of the transient equilibrium case.
50 See Attachment II Figure 27 for the excerpt from this report
95
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 6: Incorrect Preservation of Samples and Its Effect on Analysis
Isotope
Run Date Qualifier Activity
2 Sigma
Uncertainty
MDC
LLD
2 Sigma
tpi:
Units
Gamma Spec
Be-7
03/25/11
2.54E-H02
8.47E+01
6.03E+01
8.48E-01
pCi Filter
Te-132
03/25/11
2.31E—01
9.77E+O0
9.25E+00
9-78E—00
pCi/ Filter
1-131
03/25/11
5.28Et01
1.22E+01
7.29E+00
1.00E-01
1.23E-01
pCr Filter
1-132
03/25/11
1.32E-KH
1.04E+01
8.44E+00
1.04E-01
pCiFilter
Cs-134
03/25/11
9.20E—01
1.55E+01
7.34E+00
1.56E-01
pCiFilter
Cs-137
03/25/11
8.65E-H31
1.34E+01
7.11E-KH)
5.00E-01
1.35E-01
pCi.'Filter
Figure 32. Dry Deposition Samples Following Fukushima Event
A similar situation of 132I being 'lost' due to volatilization (as in Figure 32) occurred with
samples taken in the same area by "sticky" pad for dry deposition. The sticky pad sample sheet
was removed from the stack and folded into itself to prevent the deposited material from
sloughing off. It was then shipped in a Ziploc™ bag to the contract laboratory.
The time period between sampling and gamma spectrometry analysis was approximately 3 days.
The ratio of 132I/132Te calculated from the above data is 0.57. With a half-life of 2.3 hours, there
was ample time for the 132I to reach equilibrium prior to the analysis. The activity ratio at
equilibrium is 1.03. Clearly, the preservation technique for this type of sample needs to include a
means of preventing iodine loss due to volatilization.
96
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 7: Incorrect Abundance
Peak Energy alet Peak Net Area Continuum Tentative
No. (keV) Area Uncertainty Counts Nuclide
1
510.98 1.35E+03
91.50
5. 67E+02
1-133
ZN-71M
n
661.654.10E+02
54.2 8
2 .60E+02
CS-137
3
795.75 5.08E+01
29.54
1.38E+02
CS-134
4
810.69 3.33E+03
118.73
1.36E+02
CO- 5 8
5
834.71 3.47E+01
20.39
6 . 06E+-01
KR-8 8
MN-54
KR-89
6 1
173.30 3.81E+02
41 . 99
4.13E+01
CO-60
1-132
7 1332.46 3.17E+02
35.61
0.00E+00
CO-60
TE-133
8 1
460.69 1 . 4 0E+01
8 .51
4.OOE+OO
IDENTIFIED NUCLIDES
Nuclide
Id
Energy
Yield(%)
Activity
Activity
Name
Confidence
(keV)
(uCilmL)
Uncertainty
MN-54
0 . 99
834.84 *
99. 98
1.16E-
07
6.83E-08
CO-58
0 . 98
810.78 *
99.45
1 . 09E-
05
6.10E-07
863.96
0 .68
1674.73
0.52
CO-60
1 . 00
1173.24 *
99. 97
1.66E-
06
2.OOE-07
1332.50 *
99. 99
1.52E-
06
1.88E-07
CS-137
1 . 00
661.66 *
85 . 10
1.34E-
06
1.87E-07
Figure 33. Waste Liquid Release Sample
The peak search algorithm used for analysis of the sample identified in Figure 33, found a peak
at 795 keV, and tentatively identified it as 134Cs (note that the peak has a relatively high
uncertainty). However, the abundance limit function for this analysis was selected at 90%.
The method of calculation for this check is as follows:
I ?Ai
% Abundance Found = —
Where:
Ai is the abundance of the gamma rays identified by the software in the
sample
Al is the sum of all the abundances for gamma rays listed in the library.
The 795 keV peak is 85% abundant. The 604 keV peak has a 97% abundance but was not
detected. So even if the 795 and the 604 peaks were the only ones in the library, the peak would
97
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
not have been identified as 134Cs since, 85/(85+97) = 46.7%. Since no other peaks were found
the "Identified Nuclides" list does not include 134Cs. Inspection of the spectrum indicated that the
reason for that gamma ray not being detected is that it sits on top of the Compton Edge for the
810 keV gamma ray from 58Co. This raised the background sufficiently so that the peak was not
distinguishable. Had this peak been detected and correctly identified the abundance limit would
have been exceeded {(85% +97%)> 90%} and the software would have been determined that the
radionuclide was identified. The peak at 796 keV is interference free. The gamma ray at 604 keV
also suffers from interferences due to 125Sb (a fission product) at 601 keV and 124Sb (a direct
fission product and also an activation product) at 603 keV. For 134Cs it may be a better option to
either choose the 796 keV gamma ray as the keV (even though it is less abundant).
This emphasizes the importance of correlating the gamma spectrometry report for unidentified
peaks with what is visible on the spectrum in relation to potential interferences. It bears repeating
that the abundance or fraction test generally should not be used for libraries containing more than
a single radionuclide,
Also, note that a 1460 keV peak was detected, but not identified. This is because 40K was not
present in the library used for this analysis. While the 40K was an expected radionuclide in such a
sample (as a naturally-occurring radionuclide from water in contact with a concrete sump) the
library did not reflect this fact. Review of unidentified peaks should have but failed to identify
this problem. Including 40K in a suspect library would help alleviate such a situation.
98
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 8: Software Preset Functions and Radionuclide Misidentification and
Uoidentification
GAMMA SPECTRUM ANALYSIS
Sample Identification
Sample Description
Sample Type
Unit
Sample Point
Sample Begin Date/Time
Sample End Date/Time
Nuclide Library
Procedure
Sample Size
Live Time
Dead Time
Peak Significance Threshold
Peak Locate Range (in channels)
Efficiency Calibration Date
12Aug14-012
MAIN STACK CHAR
GRW
1
MAIN STACK
8/5/2014 9:40:00AM
B/12/2014 11:00:00AM
IODINE
GRW Iodine SO
T.saoE+oscc,^^^
2000.0 seconds
0.01 %
2.50
100 - 4096
8/25/2009
Facility
Acquisition Started
Operator
Detector Name
Geometry
Real Time
Identification Energy Tolerance
Energy Calibration Date
Apex Number
: 8/12/2014 1:50:25PM
DET2
LVCC0
2000.2 seconds
1.000 keV
8/11/2014
58605
PEAK WITH NIP REPORT
Peak Energy Net Peak Net Area Continuum Tentative
No. (keV) Area Uncertainty Counts Nuclide
1
80.15
9.51E+01
53.77
4.18E+02
1-131
2
165.80
2.05E+02
58.76
4.04E+02
• ¦ . . .
3
249.74
6.36E+01
37.14
1.87E+02
4
284.27
2.73E+02
53.62
2.36E+02
1-131
5
364 .49
2.87E+03
113.26
1.79E+02
1-131
6
462.80
7.12E+01
25.45
5.56E+01
7
510.63
7.19E+01
26.46
6.83E+01
1-133
8
529.90
8.73E+02
63.22
8.39E+01
1-133
9
537.08
1.83E+01
16.85
3.94E+01
. ¦ • . ¦
10
546.70
2.53E+01
17.90
4.34E+01
1-132
11
555.51
3.90E+01
17.83
3.00E+01
12
636.95
1.2 9E+02
25.23
2.28E+01
1-131
13
722.87
4.06E+01
15.36
1.48E+01
1-131
14
875.33
2.84E+01
17.06
3.11E+01
1-133
15
1009.52
4.10E+01
13.71
3.95E+00
16
1131.43
1.12E+01
12.10
1.96E+01
17
1260.32
2.83E+01
13.30
1.35E+01
18
1435.85
4.39E+01
15.95
1.62E+01
19
1460.46
1.35E+01
12.33
1.29E+01
Nuclide Nuclide Wt mean Wt mean
Name ld Activity Activity
Confidence (uCi/cc) Uncertainty
1-131 0.982 1.751E-12 1.324E-13
1-133 0.877 3.504E-12 3.473E-13
Total Activity: 5.255E-12
UNIDENTIFIED PEAKS
Peak No. Energy (keV) Peak Rate (CPS) Peak Rate (%) Uncertainty
2
165.80
1.02E-01
14.34
3
249.74
3.18E-02
29.21
6
462.80
3.56E-02
17.87
9
537.08
9.15E-03
46.05
10
546.70
1.27E-02
35.35
11
555.51
1.95E-02
22.85
15
1009.52
2.05E-02
16.71
16
1131.43
5.60E-03
54.05
17
1260.32
1.41E-02
23.53
18
1435.85
2.20E-02
18.16
19
1460.46
6.77E-03
45.54
Figure 34. Radioiodine Analysis of a Charcoal Cartridge
The excerpts above are from the analysis of a charcoal cartridge that was used to sample a
nuclear power plant vent for one week. Several different preset functions play into radionuclide
99
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
misidentification in this sample analysis. The results have been decay corrected to the mid-point
of the sampling interval (3.5 days prior to the analysis). Four of the unidentified peaks belong to
135I. Four of the unidentified peaks belong to 138Cs. The report header is missing information
regarding the half-life ratio function which was selected as the default value of 12. The library
selected was the "Iodine" library. Why were 135I and 138Cs not identified? The ratio of decay
from sample time to radionuclide half-life was in excess of 12; (3.5 days x24 hours per day/6.57
hours) = 12.8. Thus the software eliminated 135I as a possible radionuclide based on the logic that
its half-life is too short to be present after 12 half-lives. But this function does not take into
account that the sample was counted only two hours after it was removed from the active sample
line. 1-135 was being continuously released to the plant vent during the sampling interval. Thus
the half-life ratio preset function should not have been used. This error in the use of the preset
function caused a released radionuclide not to be reported properly. Decay correcting to the
midpoint of the sampling interval is recommended for this process but as in this case may not
always be appropriate.
The correct time to be used as a reference for decay correction should be reconsidered.
Four of the unidentified gamma rays, 1009, 1435, 547 and 462 keV, belong to 138Cs. While the
particulate filter that precedes the charcoal cartridge normally captures this as particulate, 138Xe,
which is also present in the plant vent effluent, is the precursor to 138Cs, and the xenon gas has an
affinity for the charcoal cartridge. The library chosen for analysis of this spectrum was "Iodine"
which did not contain either 138Xe or 138Cs as one of its entries. Furthermore, even if they had
been in the library, their short half-lives (14 and 32 minutes, respectively) would have prevented
the software from identifying them for the same reason the 135I was not identified. Review of the
unidentified peaks report in the spectrum should have caught the fact that these peaks had been
disqualified, and led to resolution of the issue (using an appropriate library and increasing or
disabling the half-life test).
100
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 9: Effect of Differing Chemistry of Parent and Progeny
1600
1400
1200
~ Laboratory Results
-Certified Result
1000
600
800 ~ . ~ A ^ A ~ ~ ~ . A +
i ~ ~ ~ ~
400
200
9 11 13 15 17 19 21 23 25 27 29
Figure 35. Proficiency Test Results for 95Zr
A fission product PT sample was prepared by dissolving irradiated uranium. Aliquants of the
final solution were sent to various laboratories for analysis (see instructions sent to laboratories
in Attachment 1). The method of sample dissolution and subsequent dilution for distribution did
not take into account the solubility characteristics of one of the radionuclides, 95Zr (the solution
should have been preserved with fluoride to complex the zirconium maintaining it in solution).
The laboratory that prepared the PT samples performed multiple analyses of replicates of the
undiluted solution so that the activity concentrations for 16 radionuclides could be used as a
reference for comparison with the participating laboratories. Their analyses were performed very
shortly after the undiluted solution was dissolved. The laboratories received the diluted samples
about a two weeks later. The results are shown in Figure 35 for 95Zr with all 30 participating
laboratories results included. The average value from the participating laboratories was biased
-34% low compared to the reference laboratory value.
The disparity of the reference value versus a single laboratory result was investigated by one of
the participating laboratories and by the contract laboratory. The participating laboratory had
performed a partial transfer of the sample to their own gamma spectrometry geometry without
any additional chemical means. They recounted the original shipping container after it had been
emptied and found residual 95Zr. It was discovered that a substantial portion of the 95Zr had
either settled or plated out in the shipping container. The contract laboratory had examined the
same possibility and obtained similar results. It was recognized that preservation of the sample
would need to include fluoride ion to ensure that zirconium did not precipitate out on the
container surfaces.
Analysis results like these emphasize the importance of preservation of samples even over short
periods of time.
101
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 10: Incomplete Preset Library to Detect Gamma Rays
COLLECTOR...,
ANALYST
SAMPLE TIME..
ANALYSIS TIME
DETECTOR
25-JAN-2Q1G 23:03:00.00
25-JAN-2010 23:06:23.78
HPGE2
NUCLIDE LIBRARY.
GEOMETRY
COUNT TIME.......
SAMPLE VOL/MASS.,
DEADTIME
GENLIQ
1LIQ1L
0 00:50:00.00
1.00000E+03 ML
0.0%
SAMPLE POINT.
REMARK. ......
CONFIG FILE..
SYS$SYSDEVICE:[CRU.SAMP]HPGE2_SAMP 6 961.CNF;1
Energy
Area Bkgnd
FWHM Nuclides
511.21
699 136
2
.41
CO-58(A)
513.61
50
75
1
.24
SR-85
KR-85
810.99
1898
52
1
.35
CO-58
835.94
19
23
0
.81
MN-54
864.27
22
22
1
. 11
CO-58
1099.48
74
12
1
. 53
FE-59
1173.65
181
8
1
.51
CO-60
1291.98
39
3
1
.81
FE-59
1333.12
174
3
1
.61
CO-60
1462.17
10
3
0
. 82
Nuclide
Activity
1-
Sigma
UCI/ML
Error
MN-54
1.193E-07
6.
291E-08
CO-58
1.180E-05
7.
545E-07
FE-59
9.3 04E-07
1.
156E-07
CO-60
1.575E-06
1.
096E-07
KR-85
4.749E-05
2.
195E-05
SR-85
2.077E-07
9.
586E-08
Activity :
6.213E-05
Figure 36. One Gamma Ray Used for two Radionuclides
This sample was a waste liquid release from a nuclear power plant. The library used was their
general liquid library which included all the main library radionuclides. The annihilation peak
was seemingly resolved from the 513 keV gamma ray by the software and tentatively identified
as either 85Kr or 85Sr. However, the FWHM of the annihilation peak had an effect on the
background in the 513 keV peak, likely causing its area to be underestimated. The 85Kr activity
was calculated but was likely underestimated due to the potential overlap from the annihilation
peak.
To start, this library should not have been used since 85Sr cannot possibly be present in the waste
liquid of a pressurized water reactor (it is neither an activation product of reactor materials nor is
it a fission product). Similarly, the identification and quantitation as 8-Sr should have been
rejected in the review process as incorrect and the radionuclide to which it truly belonged
identified. Since more than one radionuclide used the same gamma ray to calculate the activity, it
is unclear whether the software apportioned the peak between the two radionuclides (i.e., via
deconvolution of the peak or interference correction) or whether is used the same area counts for
both.
The 1462 keV peak is from 40K. This radionuclide and other naturally occurring radionuclides
should be included in the waste liquid library as many of the waste streams can come into
contact with concrete and leach these radionuclides from the concrete. If 40K is really not of
102
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
concern, including it in the suspect library would at least help facilitate identification of the
unidentified line. Some software packages have the ability to qualitatively identify a gamma ray
without including it for quantitation.
The only gamma ray for 58Co that should have been used in the calculation was the 810 keV
gamma ray. While the "(A)" notation indicates that the 511 peak is annihilation radiation, the
library contained a user added abundance factor, and thus the 511 peak was used in the activity
concentration calculation. The areas for improvement for this analysis are:
• Revision of library to:
o Better address gamma rays with gamma rays that have energies that are close to each
other
o Remove any improbable radionuclides from the individual sample stream library
o Use only true gamma ray lines for 38Co
• The review process also needs to be improved to include basic facts about radionuclides that
can be present in the system and those which cannot.
• The energy calibration, and possibly the shape (resolution) calibration could be improved.
This certainly led to the 1462 peak being off by 1.2 keV and may have led to the overlap at
511 and 514 keV.
Example 11: Sample-Source Geometry Mismatch
Figure 37. Sample Containers of Similar but Different Size and Composition
Figure 37 shows a number of different sample containers that a laboratory used to count samples
using one sample to detector geometry. Although the containers are close in size and shape, they
do not have exactly the same dimensions. Additionally, they are made of different plastics with
103
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
different wall thicknesses. The standard (not shown) used to calibrate the detector was prepared
in a container that matched the clear plastic ones seen in the figure. Also disconcerting was the
sample containers were filled to different levels.
This situation represents a sample-calibration source mismatch. Samples analyzed under this
condition will have an unknown, but definite bias.
Example 12: Incorrect Placement of Samples in Detectors
Shelf Holding Geometry
0.9"
3.5"
0.6"
3.5"
CTi
d
6
ZL
/
N
Figure 38. Sample Container Geometry Holder
The sketch of a sample container holder shown in Figure 38 is based on an assessment of a
laboratory's count room practices. The dimensions of the holder show that the circular groove
for fitting a charcoal cartridge or air particulate filter was not exactly centered on the square shelf
used to hold it. None of the edges were marked to show which edge should be placed in the same
position when counting a sample. Additionally, there were nine of these shelves for 3 different
detectors, with three different sample holding geometries also shown in Figure 37. Position zero
was fixed in the holding geometry; the shelves were unmarked so any shelf could be used in any
holding geometry on any detector. The holding geometries were not assigned to a specific
detector and showed signs of physical wear at each surface contact point. It is also important to
keep in mind that in HPGe detectors, the germanium crystal is quite frequently not perfectly
centered under the detector protective "can".
The result of these minor inconsistencies is a potential for sample-to-standard geometry
mismatch unless the same holding geometry is assigned to a specific detector and each shelf is
coded so that it has one position on a specific holding geometry. There was no such coding at
this laboratory and results from some PT samples are seen in Figure 39 A, B, and C.
104
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 13: Incorrect Placement of Samples in Detectors and Sample Bias
(A)
(B)
(C)
The graphs above show the % Bias or % Difference for aqueous PT samples containing several
radionuclides (only three are shown here as an example). The activity concentrations for all three
were in the range of 1x10' to 2xl08 pCi/L (lxl0~2 to 2xl0_1 jiCi/mL).
Cs-134 Bias
1.0E+01
5.0E+00
2ndQtr
2011
4th Qtr
2011
2nd Qtr
2012
T
O.OE+OO
A
<« -5.0E+00
5
^ -1.0E+01
-1.5E+01
-2.0E+01
A /
t
/ J
f
I
kr
J
""A Reference A
value ¦# 1 Bias
f
-2.5E+01
0
5 10
Semi-Annual result
15
Cr-51
Co-58
Figure 39. Incorrect Geometry Leads to Bias
105
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
The principal gamma rays for 134Cs suffer from the coincidence sum effect. This causes the
results of activity for this radionuclide to be biased negative. The closer the sample is to the
detector the larger will be the negative bias. This effect cannot be accounted for in any simple
manner. Figure 39 (A) shows results for PT samples for an air particulate filter geometry.
For each quarter-year period shown in Figure 39A, the distance from the sample to detector
increases from left to right. It can be seen that the bias for the 134Cs measurements decreases with
distance that the sample is counted from the detector. The bias in these results was exacerbated
by the sample holder used to 'center' these filters on the detector (see Figure 38). A
mispositioning of the sample holder likely caused the +5% bias of the 4th quarter 2011 sample on
the farthest shelf from the detector.
Figures 39 (B) and (C) show the results for 51Cr and 58Co in the same PT sample. The percent
difference in these analyses is very wide for samples that had a relatively high activity (i.e., low
uncertainty due to counting). Both radionuclides have gamma rays of good abundance and, in
these samples; there are no interfering gamma rays from other radionuclides. These results were
not trended, and the laboratory staff had accepted all these analyses as satisfactory.
106
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 14: Incorrect Use of Half-Lives and Inappropriate Detector
Configuration
Sample date
Sample ID
Detector name
Elapsed live time
Energy tolerance
Abundance limit
Batch ID
Matrix Spike DPM
DKAO:[CANBERRA.GAMMA.ARCHIVE.GAMMA]G174206007.CNF;1
8-APR-2006 08:25:00. Acquisition date : 18-OCT-2006 15:45:27
G174206007 Sample quantity : 1.00000E+00 FILTER
WELL Detector geometry: HPETRI
0 02:00:00.00 Elapsed real time: 0 02:09:07.18 7.1%
2.00000 KEV Analyst Initials :
75.00000 Sensitivity : 3.00000
579869 Detector SN# :
LCS DPM :
Energy
Area
Bkgnd
FWHM
67.63
1379
153177
1. 48
74.92'
1D376
187455
1.48
B4 .68
16035
211874
1.69
87 .17
5456
212158
1.28
122.01
95437
369559
1 .68
J 32.57
1243
170570
1.03
136.36
9700
229518
1 .56
I76.ia
4216
255963
1.72
137.94*
1535
25 '367
1 .56
510.59
1911
171913
1 .14
320.08
34 691
235339
1.81
365.37
1031
119657
1.11
391.76
15966
179627
1 .84
427.97
9537
143321
1 .91
4 63.52
287 4
129181
1 .55
4 ~9.89
799
130343
2.30
511.14*
129122C
413413
3.03
6'J1.72
5134
131333
3.21
614.38
2467
145545
2.81
635.94
3175
118114
1.20
724.22
101402
110407
2.02
756.76
121446
83601
2.22
765.84*
452958
83826
2.23
810.85
2898967
154325
2.08
831.93*
469261
112589
2 .10
864.01
17878
79273
2.15
684 .44
411
32398
0.92
930.32
153
48536
1.27
1099.39
10023
51061
2.12
1115.78
9192
54079
2.22
1173.45*
865763
53748
2.29
1291.81
74 32
12115
2 .29
1322.06
3063
10377
2.54
1332.71'
7554 59
7829
2.55
1447.72
130
2625
2.59
1576.21
199
1838
0.95
1621.10
995
2177
2.83
16 lb.13
7914
2325
2.57
1691.48
925
1498
2.54
1843.04
168
825
2.84
1852.80
55
544
2.29
1930.32
68
511
2.73
1356.65
133
602
5.88
1984.OS-
406
67 4
2.57
Uncorrected
Decay Corr
pCi/FILTER
Decay Corr
2-Sigm!
Nuclide
Hlife
Decay pCi/FILTER
2-Sigma Error
%Erro:
BE-7
53 . 44D
12.3
8.092E+03
9.936E+04
15.59E+04
156.90
CR-51
27.70D
126.
2.554E+05
3.224E+07
0.29 6E+07
9.17
MN-54
312.70D
1.54
8.111E+05
1.245E+06
0.07 5E+06
6.00
CO-57
270.90D
1 .64
4.401E+04
7.218E+04
0.486E+04
6. 73
CO-58
70.80D
6.64
4.910E+06
3.260E+07
0.200E+07
6.14
FE-59
44.63D
20.1
3.817E+04
7.689E+05
0.896E+05
11.65
CO-60
5. 2 7 y
1.07
1.892E+06
2.028E+06
0.115E+06
5.69
ZN-65
244 .40D
1.73
3.941E+04
6.819E+04
0.855E+04
12.54
NB-95
35.06D
45.7
7.288E+05
3.332E+07
0.220E+07
6.59
ZR-95
64 .02D
8.11
3.4 91E+05
2.832E+06
0.214E+06
7.55
CD-109
4 64 .00D
1.33
S.835E+04
7.789E+04
2.381E+04
30. 57
SN-113
115.10D
3.20
2.150E+04
6.887E+04
0.840E+04
12.20
SB-124
60.20D
9.27
5.7 7 4E+03
5.350E+04
1.113E+04
20.81
SB-125
2.77Y
1.14
3.088E+04
3.526E+04
0.558E+04
15.83
SN-126
1.0OE+05Y
1.00
5.977E+Q3
5.977E+03
1.827E+03
30.57
CE-144
284.30D
1 .60
4.638E+03
7.432E+03
7.8 61E+03
105.77
PM-147
2.62Y
1 . 15
9.408E+10
1.082E+11
0.073E+11
6.71
NP-237
2.14E+Q6Y
1.00
1.755E+04
1.755E+04
0.647E+04
36.88
ANH-511
1.OOE+09Y
1.00
1.448E+06
1.448E+06
0. 095E+06
6.56
Figure 40. Progeny Decay Correction with Incorrect Half-Life
This gamma-ray analysis is of a filter sample from a nuclear power plant. The filter was used to
remove radioactivity from the reactor coolant system (RCS). This means it removed fission as
well as activation products from the RCS. The filter was used to process liquid through an entire
18-month fuel cycle. The filter was allowed to decay in the storage container for at least two
107
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
months prior to the sample of the filter being removed on the 8ih of April 2006. The sample
resided at the power plant until the beginning of October 2006 and sent offsite to be analyzed.
Nuclide
Type:
Uncorrected
Decay Corr
2-Sigma
Nuclide
Energy
Area
SAbn
%Ef f
pCi/FILTER
pCi/FILTER
tError
BE-7
'177 .59
799
10.42*
3.556E-01
8 .092E + 03
9.936E+04
156.90
CR-51
320.08
34 691
9.83*
5.187E-01
2.554E+05
3.224E+07
9.17
MN-54
834.83
469261
99.83*
2 .175E-01
8 .111E + 05
1.245E+06
6.00
CO-57
122.06
95437
85.51*
9.520E-01
4.401E+04
7.218E+04
6.73
136.47
9700
10. 47
9.276E-01
3.749E+04
6.149E+04
18.62
CO-58
810. '8
2398967
99.45*
2.229E-01
4 .910E+06
3.260E+07
6.14
FE-59
192.35
3.11
7.836E-01
Lin
e Not Found
1059.25
10028
56.50*
1.745E-01
3.817E+04
7,689E+05
11.65
12 51.60
7432
43.20
1.537E-01
4 .201E + 04
8.463E+05
10.50
CQ-60
1 L <'3 .2 4
366768
99. 90
1.658E-01
1.964E+06
2.106E+06
6.14
13.>2. 50
7 5 5 4 5 9
99.98*
1,499E-QI
1 .892E+06
2.028E+06
5.69
2N-65
1115.55
9192
50.75*
1.725E-01
3.941E+04
6.819E+04
12 .54
N8-95
765 .79
452958
99.81*
2.337E-01
7.288E+05
3.332E+07
6.59
2R-95
724 . 18
101402
43.70
2.451E-01
3.554E+C5
2.S83E+06
7.55
7?6.74
121446
55.30*
2.361E-01
3.491E+05
2.832E+06
7.55
CD-109
3 8 . 0 3
5456
3.79*
9.261E-01
5.835E+04
7
30.57
SN-113
331.69
15 966
64.90*
4.296E-01
2.150E+04
6!887E+04
12.20
SB-124
602.73
5134
97.87
2.883E-01
6.829E+03
6.328E+04
30 .76
645.85
7 . 26
2.707E-01
Liri
e Not Found
713.78
2 . 38
2.482E-01
Line Not Found
722.79
101402
11 . 10
2.451E-01
1.399E+06
1.296E + 07
CO
1368.16
2.51
1.468E-01
Line Not Found
1690.98
925
49.00*
1.227E-01
5.774E+03
5.350E+04
20.81
SB-125
176.33
4216
6.89
8.272E-01
2.777E+04
3.170E+04
4 2.65
427.89
9537
29.33*
3.952E-01
3.088E+04
3.526E+04
15.83
463.38
287 4
10.35
3.666E-01
2.843E+04
3.246E+04
44 .08
600.5 6
5134
17 . 80
2.883E-01
3.75SE+04
4.287E+04
30.95
606.64
5.02
2.862E-01
Line Not Found
635-90
3175
11. 32
2.745E-01
3.836E+04
4.379E+04
39.92
SN-126
64 .28
9.60
7.708E-01
Line Not Found
86 .94
5456
8. 90
9.261E-01
2.485E+04
2.4 85E+04
50.70
8 ' .57
5456
37.00*
9.261E-01
5 .977E-t-03
5.977E+03
30.57
CE-144
133.51
1248
10.80*
9.351E-01
"4.638E+03
7.4 32E + 03
105.77
PM-14?
121.30
95 4 37
0.00*
9.520E-01
9.408E+10
1.082E+11
6.71
NP-237
86.48
5456
12.60*
9.261E-Q1
1.755E+04
1.755E+04
36.88
95.87
2.60
9.4 98E-01
Line Net Found
Figure 41. Energy Lines from Gamma-ray Analysis of a Filter
As can be seen in the header information, the sample was placed into a Petri dish and counted
using a well-type detector on October 18, 2006. The dead time was about 7%. The FWHM of the
peaks located by the software is wide, as expected, compared to a non-well-type detector (and it
is likely the high dead time also contributed to this factor). The use of a well-type detector for the
analysis of high activity samples such as this one also leads to larger and greater varieties of
random sum peaks (such as the ones at 1621 and 1322 keV, both from 58Co). In this instance, a
planar p-type HPGE would have been better suited for the analysis as there may be low
abundance gamma rays from other radionuclides that are not seen due to the wide FWHM of the
high activity gamma rays.
The half-lives used for the analysis are listed in the second column in Figure 40. Note that the
95Nb activity concentration was calculated using a 34-day half-life. The ratio of the 95Nb/9:,Zr
activity should be about 2.2, as they are well into their transient equilibrium when the sample
was taken. The activity ratio from the report excerpt above is 11.73. If the activity of 95Nb
activity is decay corrected using the 64.02-day half-life of 95Zr, the corrected activity is 5.76xl06
pCi/filter, and the 9\Nb/95Zr ratio is 2.035, in much better agreement with the theoretical value of
2.2.
108
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
In addition to the half-life issue, the same gamma-ray lines were used to calculate the activity of
several different radionuclides as seen in Figure 41. Specifically: 109Cd, 237Np, 126Sn were all
calculated using the same area under the 87 keV peak. Co-57 and 147Pm were both calculated
from the same area under the 121 keV peak. Also, note that the peaks under 126Sn at 86.94 and
87.57 are displayed with the same peak areas.
These issues can be avoided if:
• The library is constructed to resolve interferences, and interference correction features are
properly configured and enabled for that analysis protocol
• Alternate key lines are chosen for those radionuclides identified in the sample
• Radionuclides that cannot be found in the sample are eliminated from the sample-specific
protocol library
109
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 15: Incorrect Identification of Gamma Rays from Insufficient Libraries
and Incorrect Identification of Peaks
31-214
0.517
Nuclide
Name
BI-214
Id
Con f idence
0.517
7€.56
C.45
79.29'
0.76
2.95394E-004
S9.80
C - 34
273.70
0.17
387.0C
0.29
359.10
0.37
405.74
0.18
454.77
0.28
469.69
0.14
6C9.31*
44.80
3.94054E-006
665.45
1.29
719.86
0.42
768.36
4.80
736.10
0.30
306.17
1.12
934.C6
3.03
964.06
0.38
1051.96
0.34
IC59.96
0.28
1120.29*
14.SC
5.17414E-0C7
Energy
Yield
Activity
(keV)
(1)
{uCi/ml )
1133.66
1135.19
120"?. 68
1238.11
1280.96
1377.67
1385.31*
1401.50*
1407.98*
1509.23
1538.50
1543.32
1583.22
1594.73
1599.31
1661.28
1683.99
1729.59
1764.49*
1836.36
1847.42
1973.16
2118.55
2204.21
2293.36
2447.86
Activity
Uncertainty
0.28
1.64
0.49
5.86
1.44
3.92
0.89
1.55
2.80
2.12
0.51
0.33
0.70
C.31
0.38
1.14
0.25
2.88
15.36
0.40
2.04
0.25
1.14
4.86
0.30
1.50
1.002222-005 2.85794E-0Q6
1.46739E-005 3.65183S-006
2.77374E-036 9.83748E-007
3.18320E-007 1.O6641E-007
Figure 42. Proficiency Test Sample Spiked with Fission Products
This is an excerpt taken from a gamma-ray report for analysis of a proficiency test sample that
was spiked with some fission and activation products. The PT sample had the following
radionuclides at the reference activity concentrations listed here:
110
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
95Zr
95Nb
103Ru
106Ru
llGm^g
131I
134Cs
137Cs
139Ce
141Ce
152Eu
Accepted
Value,
pCi/L
38132
73800
29560
401
424
131400
3948
211
2014
68020
252
The laboratory staff had never dealt with anything but environmental samples with only NORM
radionuclides. The sample had been counted for a relatively short period of time compared to
their normal environmental samples, and this sample did not have any measurable NORM. The
final activity (4.53xl0"7 |iCi/mL) was determined by using the weighted mean of all the peaks
the software identified as belonging to 214Bi. The review of the sample did not identify that the
range of values varied from 2.95xl0"4 to 3.18xl0"7 |iCi/mL. As one of the first steps in the
review process, seeing if the data make sense may have started the unraveling of this data to
yield true gamma-ray identities.
The only gamma rays that are associated with the background 214Bi activity are the peaks at 1764
and 1120 keV. Both yield activities that are reasonably close in activity concentration.
The following peaks should be associated with other gamma-ray emitters:
• The 79.29 keV gamma ray is a 2.6% abundance peak from 131I (only the most abundant 131I
gamma ray at 364 was in the library)
• The 609.3 keV peak is from 103Ru (was not in the library)
• The 1120 keV peak is a random sum peak from 131I and 95Nb (364 +756 keV) and
contributed to the peak area causing that activity to be larger than the activity calculated for
the 1764 keV peak.
• The 1385.3 keV peak is from 110mAg (this radionuclide was not in the library and all the other
more prominent peaks from this radionuclide were sent to the unidentified lines report)
• The 1401 keV peak is a coincidence sum peak from 134Cs
• The 1407.9 keV peak is from 152Eu (not in library, more prominent gammas went to the
unidentified lines report)
The results of this report could have been improved if the library had been adjusted to recognize
the fission products and activation products that would likely have been in the sample. A more
thorough review should have been performed of the individual line activities (which would have
identified discrepancies in the activities) and also the potential for the random and coincidence
sum effects creating interferences in identification and quantification.
Ill
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 16: Incorrectly Identified Gamma Rays Based on Energies
Nuclide Peak Centroid Background
Channel Energy Counts
J-129
117.10
29.46
13792.
J-129
118.38
29.78
12739.
Ba-133
131.90
33.16
28434.
CE-139
131.37
33.03
18768.
CE-139
133.00
33.44
21828.
J-129
13 3.64
33.60
18507.
XE-131M
136.8*
34.40
45389.
J-129
136 JiA
34.40
18261.
CE-141
141 43
35.55fc
32984.
CE-141
143.34
36.03j\
26530.
J-131
1174.32
283.99 \
V 10890.
EU-152
13 rz. 28
343.52
^ 9191.
J-13 1
1> 54.80
364 . 16
\\8473 .
Net Area
Counts
1910 ,
3972 .
6797 .
11498.
19944.
8333 .
[99871H
intensity Uncert FWHMi
Cts/Sec 1 Sigma * keV
0. 099
15.90
0.8511
0.325
3.45
0.8511
0.418
5.48
0.855.
0.177
7.65
0.8551
0.296
6.38
0.855!
0.067
22.44
0.856:
0.368
7.76
0.856:
0.629
3 .06
0.8562
1.065
6.50
0.8581
1.847
1.04
0.858:
0 .772
2.92
1.263
0 .038
32.81
1.173
9.247
0. 38
1.283
X-rays: not specific to 129l
(131l has I.C./y = 1.2 at 80 keV and
0.05 at 283 keV
X-rays: not specific to 141 Ce
Figure 43. Analysis of a PT Fission Product Sample Using X-ray Region
In this specific example, the X-rays of iodine were used to determine 129I a very long-lived
isotope of radioiodine, and the X-rays of cerium were used for 141Ce analysis instead of using the
145 keV gamma ray (which was not identified). Since X-rays of I or Ce are representative of any
isotope of that element, they are not useful for identification or quantification of any specific
isotope.
Note the net area counts highlighted in yellow for the various iodine peaks that were identified.
The counts for the 29.46 and 33.6 keV peaks if only attributable to one radioiodine would have
been -1880 and 1030 counts, respectively. The gamma rays emitted by 131I at 80 and 283 keV
have internal conversion ratios of 0.05 and 1.2, respectively; thus both yield significant X-rays
which cannot be associated with 129I.
Both these analyses were significantly biased by the shorter-lived, higher activity radioisotopes
of the other isotopes, Xe and Pr, the progeny of multiple iodine and cerium fission products
present in the sample.
An additional issue with this analysis was that the detector was calibrated between 59 and 1837
keV. Thus quantitative assessment of activity concentration outside of that energy range is not
valid.
112
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Example 17: How Progeny Activity Can Be Used To Calculate Parent Activity
1.0E+05
IMd Activity
1.0E+04
Pm Activity
S 1.0E+03
1.0E+02
1.0E+01
0
50
100
Days
Figure 44. Pm-147 Detected 20 Days Following an Event
A sample taken 20 days after an IND event51 was recounted and found to contain 6.86xl06 Bq/g
of 147Pm. During the initial analysis, the 147Nd peak at 91.1 keV was interfered with by a large
set of 234Th X-rays at 92.2-92.8 keV. A better value for the 147Nd was requested by the incident
command to correctly assess the dose during the early phase of the incident.
This example is not from real data but presented here to show how the progeny of a "no
equilibrium" case can also be used to calculate the activity of a parent that has decayed away.
In this instance, the sample is counted 80 days following the initial event. The assumptions are:
• The 1-liter soil sample was collected in a Marinelli beaker
• It was collected approximately 2 hours following the incident and was analyzed immediately
(thus there was negligible ingrowth of the 147Pm at that time)
• The sample was subsequently preserved in a temperature controlled room for 80 days and
recounted
The equation that relates the activity of progeny to parent at any time, t, is (Equation 6):
A\ = J4^e~A2t + - e—
In this instance, however, the activity of the progeny is known at 80 days and it is necessary to
calculate the initial activity of parent.
51 An IND that contained both enriched and natural uranium was detonated. The natural uranium had fully ingrown
234Th.
113
-------
High Resolution Gamma Spectrometry Analyses For Normal Operations and Radiological Incident Response
Rearranging the above equation we can calculate the activity of 147Nd at time zero.
0 = (A12-A°2xe~^t)x(A2-A1)
1 A2 x (e-/llt — e-/l2t)
The term A2 is assumed to be zero at 2 hours post event since the direct (independent)
fission yield, A20, of this radionuclide is very small (< 2.5xl0"6%) compared to the production
from the decay of its short-lived progenitors. Substituting in the equation we find:
n (6.86xl06 -0xe-723xl°"4x8°)x(7.23xl0-4-6.31xl0-2)
A\ = ^; J-—- r ^ = 6.31 X 108 Bq
1 23xl0-4 X (e~6-31xl° x80 — S — 7.23x10-4x8(A n
114
-------
-------
------- |