xvEPA
United States
Environmental Protection
Agency
Development of Rapid
Radiochemical Method for Gross
Alpha and Gross Beta Activity
Concentration in Flowback and
Produced Waters from Hydraulic
Fracturing Operations
This report was prepared for the National Analytical Radiation Environmental Laboratory of the Office of
Radiation and Indoor Air and the National Exposure Research Center of the Office of Research and
Development, United States Environmental Protection Agency. It was prepared by Environmental
Management Support, Inc., of Silver Spring, Maryland, under contract EP-W-07-037, work assignment 2-
43, and EP-W-13-016, task order 014, both managed by Dan Askren. This document has been reviewed in
accordance with U.S. Environmental Protection Agency (EPA) policy and approved for publication. Note
that approval does not signify that the contents necessarily reflect the views of the Agency. Mention of trade
names, products, or services does not convey EPA approval, endorsement, or recommendation.
RESEARCH AND DEVELOPMENT
-------
-------
EPA/600/R-14/107
July 2014
Development of Rapid Radiochemical
Method for Gross Alpha and Gross Beta
Activity Concentration in Flowback and
Produced Waters from Hydraulic
Fracturing Operations
Brian Schumacher
U.S. Environmental Protection Agency
Office of Research and Development
National Exposure Research Center
Environmental Sciences Division
Las Vegas, NV 89119
John Griggs and Dan Askren
Office of Air and Radiation
Office of Radiation and Indoor Air
National Analytical Radiation Environmental Laboratory
Montgomery, AL 36115
Bob Litman and Bob Shannon
Environmental Management Systems
Silver Spring, MD 20910
Marinea Mehrhoff, Andrew Nelson, and Michael K. Schultz
University of Iowa
Iowa City, Iowa 52242
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
CONTENTS
Acronyms, Abbreviations, Units, and Symbols vii
Radiometric and General Unit Conversions ix
1.0 Introduction 1
2.0 Characterization of the Sample Matrix 5
3.0 Initial Radioanalytical Methods 9
4.0 Method Development Process Summary 11
5.0 Participating Laboratories 13
6.0 Measurement Quality Objectives 13
7.0 Method Development Plan 17
7.1 Method Uncertainty 17
7.2 Detection Capability 19
7.3 Method Bias 19
7.4 Analyte Concentration Range 21
7.5 Method Specificity 21
7.6 Method Ruggedness 21
8.0 Techniques Used to Evaluate the Measurement Quality Objectives for Methods
Development Project 23
8.1 Required Method Uncertainty 23
8.2 Required Minimum Detectable Concentration 23
9.0 Evaluation of Experimental Results 25
9.1 Summary of the Method 25
9.2 Required Method Uncertainty 25
9.3 Required Minimum Detectable Concentration 31
9.4 Evaluation of the Absolute and Relative Bias 36
9.5 Method Ruggedness 38
10.0 Timeline to Complete a Batch of Samples 39
11.0 Reported Modifications During Development and Recommendations for Future
Work 41
12.0 Summary and Conclusions 47
13.0 References 49
Attachment I: Method Development Trials 51
Attachment II: Time Lapse 59
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Attachment III: Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity
in Flowback and Produced Waters from Hydraulic Fracturing Operations
(FPWHFO) 61
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Figures
Figure 1 - Po-209 Alpha Spectrum 52
Figure 2 - Alpha Spectrum of Separated 230Th 53
Figure 3 - Liquid scintillation Spectrum of 209Po 55
Figure 4 - Flow Chart for Method Development Using Tracers 56
Figure 5 - Elution Profile for Traced solution with 0.1 M Ammonium Bioxalate
in4MHCl 56
Figure 6 - Total Recovery of Tracers (%) with (0.1 M Ammonium Bioxalate in
4MHC1) 57
July-2014 iii
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 iv
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Tables
Table 1 - Non-Radiological Analysis of FPWHFO and Surrogate Sample 6
Table 2 - MQOs Targeted for Gross Alpha and Gross Beta in FPWHFO 15
Table 3 - Method Development As-Spiked Concentrations, MQOs and Mean
Measured Results 16
Table 4 - Proposed Sample Processing and Counting Combinations 17
TableS - Actual Sample Concentrations Used for MD Process 18
Table 6A - Th, U. Po Gross Alpha by LSC - Analytical Results for Required Method
Uncertainty Evaluation 27
Table 6B - 226Ra Alpha by Gamma Spectrometry - Analytical Results for Required
Method Uncertainty Evaluation 28
Table 6C - 228Ra Beta by Gamma Spectrometry - Analytical Results for Required
Method Uncertainty Evaluation 29
Table 7A - Th, U, Po Gross Alpha by LSC - Experimental Standard Deviation of the
Seven PT Samples by Test Level (230Th) 30
Table 7B - 226Ra Alpha by Gamma Spectrometry - Experimental Standard Deviation
of the Seven PT Samples by Test Level 30
Table 7C - 228Ra Beta by Gamma Spectrometry - Experimental Standard Deviation
of the Seven PT Samples by Test Level 31
Table 8A - Th, U, Po Gross Alpha by LSC - Blank Water Samples 31
Table 8B - Th, U, Po Gross Alpha by LSC - Surrogate Water Samples 32
Table 8C - 226Ra Alpha by Gamma Spectrometry in Demineralized and Surrogate
Water Samples 32
Table 8D - 228Ra Gross Beta by Gamma Spectrometry in Demineralized and Surrogate
Water Samples 33
Table 9A - Reported Results for Samples Containing Th, U, Po Gross Alpha by LSC
at the As-Tested MDC Value (12.45 pCi/L) 34
Table 9B - Reported Results for 226Ra Alpha by Gamma Spectrometry at the As-Tested
MDC Value (55.2 pCi/L) 35
Table 9C - Reported Results for 228Ra Beta by Gamma Spectrometry at the As-Tested
MDC Value (30.4 pCi/L) 36
Table 10 - Absolute and Relative Bias for Gross Alpha and Gross Beta by the Three
Measurements 37
Table 11 - Estimated Elapsed Times for Gross Alpha Analysis 59
Table 12- Estimated Elapsed Times for Gross Beta Analysis 59
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 vi
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Acronyms, Abbreviations, Units, and Symbols
AAL analytical action level
ACS American Chemical Society
APS analytical protocol specification
Bq becquerel
CZ/NC critical net concentration
CSU combined standard uncertainty
Ci curie
d day
EPA U.S. Environmental Protection Agency
ESD Environmental Sciences Division
FPWHFO Flowback and Produced Waters from Hydraulic Fracturing Operations
g gram
GPC gas-flow proportional counter
h hour
ID [identifier] [identification number]
L liter
LCS laboratory control sample
LSC liquid scintillation [count] [cocktail]
m meter
M [molar]
MARLAP Multi-Agency Radiological Laboratory Analytical Protocols Manual
MD method development
MDA minimum detectable activity
MDC minimum detectable concentration
MDRM method development reference material
min minute
MQO measurement quality obj ective
MVG Method Validation Guide for Qualifying Methods Used by Radiological
Laboratories Participating in Incident Response Activities
NAREL EPA's National Analytical Radiation Environmental Laboratory,
Montgomery, AL
July-2014 vii
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
NERC National Exposure Research Center
NHSRC EPA's National Homeland Security Research Center, Cincinnati, OH
NIST National Institute of Standards and Technology
ORD U.S. EPA Office of Research and Development
ORIA U.S. EPA Office of Indoor Air and Radiation
^MR required relative method uncertainty
PT proficiency testing
QA quality assurance
QAPP quality assurance proj ect plan
QC quality control
STS sample test source
MMR required method uncertainty
y year
July-2014 viii
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Radiometric and General Unit Conversions
To Convert
years (y)
disintegrations per second (dps)
Bq
Bq/kg
Bq/m3
Bq/m3
microcuries per milliliter
(nCi/mL)
disintegrations per minute (dpm)
cubic feet (ft3)
gallons (gal)
gray (Gy)
roentgen
equivalent man (rem)
To
seconds (s)
minutes (min)
hours (h)
days (d)
becquerels (Bq)
picocuries (pCi)
pCi/g
pCi/L
Bq/L
pCi/L
HCi
pCi
cubic meters (m3)
liters (L)
Rad
sievert (Sv)
Multiply by
3.16xl07
5.26xl05
8.77xl03
3.65xl02
1
27.0
2.70xl(T2
2.70xlO~2
io-3
109
4.50xlO~7
4.50x10"'
2.83xlO~2
3.78
IO2
io-2
To Convert
s
min
h
d
Bq
pCi
pCi/g
pCi/L
Bq/L
pCi/L
pCi
m3
L
rad
Sv
To
Y
dps
Bq
Bq/kg
Bq/m3
Bq/m3
HCi/mL
dpm
ft3
gal
Gy
rem
Multiply by
3.17x10^
1.90XKT6
1.14x10^
2.74xlO~3
1
3.70xlO~2
37.0
37.0
IO3
io-9
2.22
35.3
0.264
io-2
IO2
NOTE: Traditional units are used throughout this document instead of the International System of Units (SI). Conversion to SI
units will be aided by the unit conversions in this table.
July-2014
IX
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Acknowledgments
This method development report was developed by the National Analytical Radiation
Environmental Laboratory (NAREL) of the Environmental Protection Agency's (EPA) Office of
Radiation and Indoor Air (ORIA) in cooperation with the Environmental Sciences Division of
the National Exposure Research Center (ESD NERL). Dr. John Griggs was the project lead for
this document. Several individuals provided valuable support and input to this document
throughout its development. Special acknowledgment and appreciation are extended to Dr. Brian
Schumacher of ESD NERL. We also wish to acknowledge Ms. Marinea Mehrhoff (State
Hygienic Laboratory at the University of Iowa) and Dr. Michael Schultz and Mr. Andrew Nelson
(University of Iowa), who conducted the method development studies. Numerous other individuals,
both inside and outside of EPA, provided comments and criticisms of this method, and their
suggestions contributed greatly to the quality, consistency, and usefulness of the final method.
Technical support was provided by Environmental Management Support, Inc.
July-2014 xi
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 xii
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
1.0 Introduction
This report summarizes the development and testing of an improved method for the
Determination of Gross Alpha and Gross Beta Activity in Flowback and Produced Waters from
Hydraulic Fracturing Operations (FPWHFO). Flowback and produced waters are characterized
by high concentrations and complex mixtures of inorganic salts, organic compounds and other
materials. They may commonly contain concentrations of naturally-occurring radionuclides from
the uranium and thorium decay chains as high as 103-104 times the level of activity routinely
observed in environmental water samples. The complex nature and high concentration of matrix
constituents in these water samples present significant technical challenges especially for gross
alpha and gross beta determinations in such samples.
It is critical when working with gross alpha and gross beta measurements that one
understand what the screening results represent. These measurements are not nearly as simple as
many people may believe. There is a relatively a short list of natural chain radionuclides.
Uranium and thorium chain radioactivity (i.e., 238U and 232Th, and progeny), and 40K, usually
account for >95% of the naturally occurring alpha and beta activity. A relatively small amount of
actinium (Ac) chain activity may also be present (less than 5% of total activity is associated with
the Ac chain, i.e., 235U and its decay progeny). Thus, five radioelements in the sample either
directly comprise or indirectly support the naturally occurring activity of the sample:
• uranium (238U; 235U; 234U);
• thorium (232Th; 230Th; 228Th; 234Th and its short-lived decay progeny 234mPa)
• radium
o 228Ra and its short-lived decay progeny 228Ac;
o 226Ra and its short-lived decay progeny 222Rn, 218Po, 214Pb, 214Bi, 214Po;
o 224Ra and its short-lived decay progeny 220Rn, 216Po, 212Pb, 212Bi, 212Po, 208T1;
• lead (210Pb and its short-lived decay progeny 210Bi; and
• polonium (210Po).
Method 900.0 is the benchmark method for gross alpha and gross beta determinations.
Method 900.0 involves evaporating the sample to a thin layer of solid residue in a stainless-steel
planchet and analyzing alpha and beta particles emitted from the source using gas-flow
proportional counting. The benefit of this method is its rapidity and relatively low-cost in
comparison to performing radionuclide specific testing. This method focuses heavily on
identifying 226Ra, with a secondary concern on natural uranium and thorium chain activity. This
method will identify some non-volatile man-made radionuclides that are present as
environmental contaminants.
Method 900.0 has several notable limitations that impact the reliability and
intercomparability of gross alpha and beta measurements, especially for naturally occurring
radioactivity. The method under-responds to low-energy beta emitters such as 210Pb and 228Ra.
Lax control of the timing of the count relative to preparation and collection of the sample may
cause the method to dramatically over-respond, or to completely fail to detect radionuclides
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
present at the point of collection. For example, the gross alpha activity of samples containing
226Ra will increase following the preparation with the count stabilizing at 400% of the actual
226Ra activity 3-4 weeks after preparation.1 In contrast, delaying the analysis of the sample for
more than two weeks after collection will lead to a failure to detect 224Ra or its decay progeny
that were present at the time of collection. Finally, the method is not applicable for the
determination of analytes that are volatile under the conditions of analysis. This would
potentially impact 210Po.
Finally, Method 900.0 is designed for low-solids samples such as drinking water. This
is because alpha and beta particles are severely attenuated by matter they encounter as they travel
to the detector (including "self-absorption" by the solids in the sample test source). The amount
of solids in a sample limits the size of sample that can be processed and thereby the sensitivity of
the measurement. Method 900.0 restricts residue thickness for gross alpha measurements to a
maximum of 5 mg/cm2. In a 50-mm diameter stainless-steel planchet, this is equivalent to 100
milligrams of solid residue. While this method is applicable to drinking water samples which
generally have solids content under 500 mg/L, FPWHFO samples may have solids content in the
hundreds of thousands of mg/L and sample sizes would be restricted to a small fraction of a
milliliter. Since the sensitivity of the measurement is inversely proportional to the size of sample
processed, the capability of the method to detect activity could be decreased by a factor of a one-
thousand and the ability of the evaporation approach to detect radioactivity in FPWHFO
becomes questionable at best.
One approach to addressing this limitation would involve using a measurement
technique that is not sensitive to the solids content of the sample being analyzed. Gamma
spectrometry, for example, is capable of determining isotope specific gamma-ray emissions from
samples of 3 - 5 kg or larger. Unfortunately, this technique is not as sensitive as techniques such
gas-flow proportional counting (GPC) or liquid scintillation counting (LSC). Gamma
spectrometry cannot detect pure alpha and beta emitters, of which there are several of concern
for FPWHFO samples. Liquid scintillation spectrometry is a technique that is capable of
sensitive measurements of alpha and beta particle emissions and can tolerate somewhat larger
amounts of solids (up to about /^ gram in the sample test source) than can GPC.
A second approach can be taken to improve the sensitivity of gross alpha and gross beta
measurements. This involves developing chemical separation methods to remove non-radioactive
constituents from the sample thereby allowing the radioactive constituents to be concentrated
into a source that can be measured using a method sensitive to alpha and beta particle emissions
(e.g., gas-flow-proportional counting or liquid scintillation spectrometry). Method 900.0 relies
on evaporation to accomplish this. Water, which is not in itself radioactive, is removed from the
sample by evaporation. For this reason, Method 900.0 is limited to determining non-volatile
constituents in samples.2
1 This is due to ingrowth of short-lived alpha-emitting decay progeny 222Rn, 218Po, and 214Po. Similar effects are noted with beta-
emitting decay progeny of 226Ra, 224Ra, 238U, and 210Pb with time frames ranging from days to months.
2 This raises questions about whether Po can be reliably determined using method 900.0.
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
As a result of these considerations, method development efforts in Phase 1 of this
project focused on exploring:
• Detection methods that are less subject to self-absorption (e.g., gamma spectrometry or
liquid scintillation), and if possible, less sensitive to differences in the energy of the
decay particle;
• The viability of using chemical separation methods that selectively isolate longer-lived
radionuclides of concern from non-radioactive constituents that may be present in high
enough concentrations to limit sample size (i.e., U, Th, Ra, Pb, andPo need to be
separated from Ca, Sr, Ba, Mg, Na, K, Fe)\ and
• Preparation methods that are less subject to losses due to analyte volatility.
• Developing and validating a method that could rapidly achieve detection limits in the
range of those recognized for gross alpha and gross beta in drinking water.
Since each FPWHF sample matrix is different, it is important to initially characterize
whether or not U/Th isotopes may be present prior to validating any method. This will ensure
that necessary chemistry is addressed for each radionuclide.
Another important aspect of gross alpha and gross beta determinations needs to be
considered. As mentioned above, the time elapsed between collection of the sample and the
sample preparation, and the time between the preparation of the sample and the count,
significantly impact the final result. While it is not practicable to simulate this during the development
process, the theoretical impact of decay and ingrowth is well established and should
be addressed in the scope and applicability and interferences statements when a laboratory
adopts this method and validates it in their method format. Failure to delineate these concerns
may result in a failure of users to recognize that results may vary by a factor of two or more, and
that delay in analysis may result in complete failure of the method to detect short-lived
radionuclides such as 224Ra, present in significant quantities at the time of sample collection.
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
2.0 Characterization of the Sample Matrix
Per instruction by the Environmental Protection Agency (EPA), the method
development and testing targeted a single FPWHFO sample from the Marcellus shale region
of Pennsylvania. A 55-gallon drum of the water was received at the National Analytical and
Radiation Environmental Laboratory (NAREL) in Montgomery, Alabama in late April, 2013. At
the beginning of May 2013, it was then transferred to the State Hygienic Laboratory at the
University of Iowa for method development and testing studies.
The sample was described by the Iowa laboratory as a clear, light amber liquid, of high
specific gravity (-1.25 g/cc) with less than -0.5% by volume (by visual inspection) of a light,
suspended reddish-brown precipitate. The sample was not preserved prior to receipt at the
laboratory. After the sample arrived at the laboratory, the drum was mixed to ensure
homogeneity and two carboys of the water withdrawn. No sample filtering or preservation was
performed.
Preliminary analytical work was performed to characterize the major non-radiological
(i.e., chemical) composition of the water. These analyses provided perspective on the matrix so
that one could assess how best to configure a method as well as to help understand difficulties
encountered while developing the method. The data were also used to create a surrogate sample
for the method development process whose composition was very similar to that of the sample. The
chemical analyses of the field sample and the surrogate solutions are shown in Table 1.
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 1 - Non-Radiological Analysis of FPWHFO and Surrogate Sample
Analyte
Chloride (CO
Strontium (Sr)
Sodium (Na)
Calcium (Ca)
Barium (Ba)
Magnesium (Mg)
Potassium (K)
Manganese (Mn)
Iron (Fe)
Lead (Pb)
Aluminum (Al)
Fluoride (F~)
Sulfate (SO42~)
Nitrate nitrogen as
N (NO3~)
Nitrite nitrogen as
N (MV)
Ortho-phosphate
as P (PO43~)
Alkalinity
Bicarbonate
alkalinity (HCO3~)
Carbonate
alkalinity (CO321
Silicon (Si)
Total Solids (TS)
Total Suspended
Solids (TSS)
Total Extractable
Hydrocarbons
(TEH)[2]
Concentration
(mg/L)
Field Sample
147,000±6,000
36,000±2,000
29,000±1,000
13,000±1,000
9,000±400
850±40
160±10
3.4±0.2
43±2
1.0±0.1
LOD
LOD
LOD
LOD
LOD
8±1
10
10
LOD
3.7±0.3
278,000±6,000
780
2.3±0.1
LOD [1]
(mg/L)
1
200
5,000
10,000
500
120
2.5
1
5
0.5
5
0.1
40
10
5
0.02
1
1
1
2.5
1
1
0.1
Concentration
in Surrogate
(mg/L)
100,000
34,000
30,000
12,000
8,500
850
160
—
23
—
—
—
—
—
—
—
—
—
—
—
—
—
—
LOD[1]
(mg/L)
1
0.10
2.5
5.0
0.25
2.5
5.0
—
0.10
—
—
—
—
—
—
—
—
—
—
—
—
—
—
Method
Chloride, EPA 300.0
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Metals, EPA 200.7
Fluoride, SM 4500-F C 18th
Sulfate, EPA 300.0
Anions, EPA 300.0
Anions, EPA 300.0
ortho-Phosphate as P,
LAC10-115-01-1A
Alkalinity as CaCO3,
SM 2320 B 18th
Alkalinity as CaCO3,
SM 2320 B 18th
Alkalinity as CaCO3, SM
2320 B 18th
Silicon, EPA 200.7
Total Solids (Dried at
103°C), SM2540B 18th
Total Suspended Solids
(Dried at 103 °C) USGS
1-3765-85
Total Extractable
Hydrocarbons, Iowa OA-2
[1] LOD -Limit of Detection at 99% confidence determined per 40 CFR Part 136,
of the Method Detection Limit (MDL).
[2] Chromatographic patterns do not conform to fuel standards for this range.
Appendix B Definition and Procedure for the Determination
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
A radiological characterization of the water was also conducted during the preliminary
phase to determine the radionuclides present in the sample. Gamma spectrometric analysis of
raw FPW showed high levels of activity for the 226Ra sub-chain of 238U with a 226Ra
concentration of 17,600 ± 1,400 pCi/L (*=1). Significantly elevated activity of 228Ra (1,950 ±
110 pCi/L) was also noted with no indication of significant levels of 224Ra which may be present
in freshly collected samples containing 228Ra. This indicates that activity concentrations of 228Th
rjrj A. 0 1 0
were likely low and that the sample had probably aged enough that Ra and Pb had decayed
prior to receipt at the laboratory (longer than one month). The activity concentrations of 238U
(234Th, and 234mPa), and 235U were not detectable using gamma spectrometry.3 Thus, it is possible
to make theoretical conclusions about the state of equilibrium at the time of sampling. Since the
gross radioactivity of naturally-occurring decay chains changes with time, control of time
elapsed between sampling and preparation of the sample, and the preparation and analysis will
be factors that will significantly impact gross alpha and gross beta measurements and which will
need to be closely controlled.
Based on the elevated levels of radioactivity in the sample received, it became apparent
that a method development reference material (MDRM) would need to be created that closely
matched the composition of the FPWHFO but did not contain significant concentrations of
naturally occurring radionuclides. In previous method development studies of matrices such as air
filters and soil, low concentrations of naturally occurring radionuclides were present that could
easily be accounted for when creating test samples of low activity. However with the high
concentrations of the naturally occurring radionuclides in this sample, an effective measure of
uncertainty at low activities, or of a minimum detectable concentration (MDC), would not be
possible.
Reproducing the predominant non-radioactive constituents in the sample, Cl, Ca, Ba,
Sr, Mg, Na, K, and total solids in the test samples was critical both for the method development
and testing process. Based on the non-radiological analysis of the sample, a "surrogate"
method development reference material (MDRM) solution composition was formulated. It was
evident from some of the first attempts at gross radioactivity analysis that minimum detectable
concentrations would be much higher than that for drinking water.
The elevated concentrations of solids in the sample meant ordering very large quantities
of the soluble salts needed to make the surrogate used to prepare development process blank, MDC
and spiked samples.
The extremely high levels most particularly of Ca, Sr, and Ba, all Group II chemical
analogues for Ra, placed into question the initially proposed analytical approach and required
significant modification of the techniques used. While the procurement of salts was in process,
laboratory staff attempted several gross separation and analysis processes (noted below) on the
real sample to see how they would fare as preparation for the method development process.
3 Informal alpha spectrometry results obtained from personnel working with NAREL staff in May 2013 indicated that no 238U or
234U was detected in the sample at the 0.1 pCi/L level.
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Once created, the surrogate matrix was spiked with 226Ra, 230Th and 228Ra based on the
magnitude of blank analyses to create MDRMs at three levels for testing of the required
method uncertainty and one MDRM for verification of the required minimum detectable
concentration. The radionuclides for spiking differed from the outline in the method development
plan.
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
3.0 Initial Radioanalytical Methods
A series of different approaches were explored for the gross alpha and gross beta
analysis to identify an approach optimized to accommodate the matrix of the FPWHFO sample
received from the EPA. While the testing matrix used was quite challenging, these results
may or may not be applicable to matrices that differ from this sample matrix. Approaches
explored ranged from tests of EPA Method 900.0 through attempts to explore different detection
techniques, to approaches to separate/concentrate radioactive constituents of the sample. Several
examples include:
• The Method 900.0 approach failed to produce a usable test source with an extremely
small 0.1 mL aliquant.
• Direct addition of the sample to liquid scintillation cocktail resulted in precipitation of
matrix components and failure to produce a useable test source.
• Gamma spectrometry analysis provided results for a number of the natural emitters in the
sample. This approach to quantitation is not based on measurement of alpha emissions,
but the results are radionuclide specific and can be used to calculate the total gross alpha
activity associated with a radionuclide such as 226Ra in a sample when it is in secular
equilibrium4 with its progeny (that are gamma emitters).
Attachment I provides additional details on these and other attempts to processing the
FPWHFO sample, as well as testing if the relative robustness of several techniques for chemical
separation of key radionuclides (Th, U, and Po).
The goal of developing a rapid, and economically viable approach for isolating and
analyzing alpha emitting radionuclides (Ra, Th, U and Po) in a single test source and
determining gross alpha activity using a single sensitive measurement of alpha emissions,
however, has yet to be realized. This is due to the complex matrix composition, and the levels of
Group II elements (Mg, Ca, Sr, and Ba) in the FPWHFO matrix. Group II constituents were
present in overwhelming amounts (gram/liter concentrations) in this sample. The chemical
behavior of these elements closely mimics that of radium, which foiled all attempts to achieve a
group separation of Ra, Th, U, and Po.
As a result, the most viable approach for gross alpha analysis involved the use of two
physical determinations of alpha activity. The first of these determinations involves a group
separation of Th, U, and Po using a rapid solid-phase extraction chromatographic separation
technique5 followed by liquid scintillation analysis of alpha emissions. This technique is
described in greater detail below. The other contributor to gross alpha, 226Ra (+progeny) is
4 If secular equilibrium is to be assumed steps need to be taken in a laboratory's method that indicates how this will be
controlled. If secular equilibrium cannot be assured then the measured values should be qualified noting that these activity
concentrations maybe larger.
5 This involves the use of a commercial solid phase extraction resin such as Eichrom™. This material has a very high selectivity
coefficient for U, Th and Po (and other high atomic number radionuclides) over first second and third row transition series
elements.
July-2014
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
determined by high purity germanium (HPGe) gamma spectrometric measurement of gamma
emissions from the sample.
Since the gamma spectrometric determination of 226Ra has very different levels of
measurement uncertainty than does the liquid scintillation approach for thorium, uranium, and
polonium, it was necessary to establish separate measurement quality objectives (MQOs) for
required method uncertainty for each of the two analyses. The results of the analyses may be
used separately as the individual results are quite informative about radioactivity in the sample,
or the two results and their respective uncertainties mathematically combined to yield a single
combined activity and standard uncertainty for gross alpha activity in the sample.
Attempts to perform gross beta in a method along the lines of Method 900.0 also
proved fruitless as extremely elevated solids in the sample produced unacceptably high-levels of
self-attenuation or resulted in sample aliquants that were far too small to provide meaningful
detection levels. Preparation of a reproducible test source proved to be unsuccessful due to non-
uniform drying and spattering of sample solids. While a chemical separation of naturally
occurring beta emitting members and progeny of the uranium and thorium decay chains may
theoretically be possible, there would still be concern about significant differences in decay
energy of the beta emitters that could negatively impact quantitation. Using gamma emissions
from the radionuclides, however, permits use of a non-destructive method that avoids lengthy
separations, and minimizes time performing the separations. The trade-off, however, is that the
counting times for samples can be significant (6 hours or more) depending upon the gamma
detector efficiency.
While each gamma ray detected can be ascribed to a specific radionuclide and the
activities of these radionuclides and their progeny are easily summed, an assessment of the total
uncertainty can only account for the uncertainty of those beta-gamma emitters identified. It will
not account for beta-gamma emitters present at low, sub-detectable levels. This is a more
significant issue for low-activity samples where no gamma rays are detected. Based on the high
activity of radium in this type of sample matrix, this seems to be a more remote possibility.
July-2014 10
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
4.0 Method Development Process Summary
The method development (MD) plan for gross alpha and gross beta analysis of FPWHFO
follows the guidance provided in Method Validation Guide for Qualifying Methods Used by
Radiological Laboratories Participating in Incident Response Activities [MVG] (EPA 2009),
Validation and Peer Review of U.S. Environmental Protection Agency Radiochemical Methods
of Analysis (EPA 2006), and Chapter 6 of MARLAP (2004). This MD process was conducted
under the generic Quality Assurance Project Plan Validation of Rapid Radiochemical Methods
for Radionuclides Listed in EPA 's Standardized Analytical Methods (SAM) for Use During
Homeland Security Events (EPA 2012). This method is considered a "first time approach."
Therefore, the method needed to be evaluated according to MARLAP method validation "Level
D" or "Level E."
The proposed method, for both gross alpha and gross beta, in the MD plan was
validated against acceptance criteria for the required method uncertainty about a specified
analytical action level (AAL) concentration and at the required MDC. After several
modifications to the MD plan, and due to significant challenges related to the matrix of the
sample, the technique used for gross beta analysis was changed from the traditional screening
using beta particle emission counting (gas-flow proportional counting or liquid scintillation
counting) to an inferential gamma spectrometric analysis method. Since the gamma spectrometry
allows the determination of several beta-gamma emitters, the overall relative required method
uncertainty for beta-gamma emitters was expanded to 50% to accommodate the increased
uncertainties of the multiple measurements needed.
In addition to determination of the required method uncertainty and verification of the
required minimum detectable concentration, analytical results were evaluated for statistical bias,
absolute bias for blank samples and relative bias at each of the three test level radionuclide
activities. The radiochemical yield of the method was also evaluated as a characteristic of
method ruggedness.
The method development process was divided into four phases as follows:
1. Phase 1
a. Laboratory familiarization with the method concept.
b. Set-up of the laboratory and acquisition of reagents, standards and preparation of
in-house performance test (PT) samples.
c. Perform preliminary tests of the methods on an actual sample.
d. Make changes to improve the method based on the preliminary tests.
2. Phase 2
a. Conduct method testing for required method uncertainty using internal PT
samples.
3. PhaseS
a. Conduct verification of the required MDC using internal PT samples.
July-2014 11
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
4. Phase 4
a. Report results.
b. Written report and narratives by laboratory describing the process/method.
c. Feedback comments on the method.
d. Method development report written by EMS and reviewed by laboratory.
The objective of Phase 1 was to have the laboratory become familiar with the draft
method. While the laboratory obtained hands-on experience, areas of the method that required
optimization were identified. Based on the information and experience gained during Phase 1
runs, the analysis method was optimized and modified prior to initiating the formal testing
process in Phases 2 and 3.
The draft method was run (see Attachment I for description of Radionuclide: 226Ra and
Radionuclides: U, Th, and Po) with the field sample obtained from the EPA using the innate
226Ra as a measure of the success of each purported separation step. The laboratory also prepared
blank and spiked surrogate solutions (as described in the introduction above) to assess the
feasibility of the draft method with elements representing the longer-lived U and Th decay chain
radionuclides (i.e., 230Th, natural uranium, 209Po) since these radionuclides and their short-lived
progeny comprise the alpha radioactivity of the sample. Sample test sources were prepared after
the separation of the respective alpha emitters for alpha spectrometric analysis. The radionuclide
specific testing also allowed assessment of the robustness of the method to recover the added
radionuclides from the matrix. Attachment 1 presents data on the recovery of these
radionuclides/radioelements. Due to the challenging matrix, quantitative recovery was neither
expected nor was it achieved, but the final method was effective and obtained recoveries that
approached 80-90% for the targeted analytes. As a gross screening method, this is deemed to be
acceptable.
The initial objective for detection levels was aimed at drinking water MFC
concentrations. However, the difficulties encountered in this initial evaluation of the proposed
methods with the matrix evoked a reevaluation of what could be the attainable action levels and
detectability concentrations. The values that are found in Tables 2 and 3 were based on this
reevaluation.
During Phases 2 and 3 of the MD process, the laboratory analyzed in-house PT samples
consisting of method development reference materials [MDRM] prepared from the surrogate
solution (described in the introduction above) spiked with known amounts of 230Th, 226Ra and
228Ra. The laboratory was instructed to spike the PT samples with concentrations consistent with
test levels for the required method uncertainty and the required MDC. The targeted test levels
and applicable MQOs for Phases 2 and 3 are listed in Table 2.
Following completion of the method developkment studies, comments from the laboratories
were evaluated per guidance in the MVG. The final method was finalized to conform to the
documented "as-tested" conditions in Phases 2 and 3. The report was also edited to ensure clarity
based on comments received from the laboratories. Thus, the test data presented in this report
reflect the final method included in the attachments to this document.
July-2014 12
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
5.0 Participating Laboratories
The State Hygienic Laboratory and at the Radiology and Radiation Oncology Free
Radical and Radiation Biology Program Laboratory, both at the University of Iowa, performed
all of the analyses (radiological as well as non-radiological).
6.0 Measurement Quality Objectives
The gross alpha and gross beta analysis method was developed to meet MQOs
established for the project. The MQOs selected include the radionuclide concentration range, the
required method uncertainty at a specified radionuclide concentration (i.e., AAL), and the
required MDC. Since the EPA did not provide regulatory action levels, MQOs were defined in
the method development plan based on the presumed sensitivity of the techniques to be used. Once
the sample was received and initial testing proved how challenging nature of the matrix would
be, new technical approaches were needed and MQOs were updated to maintain consistency with
the capabilities of the techniques being used. The AAL for 226Ra by gamma spectrometry is
higher than initially envisioned while that for gross alpha by liquid scintillation is somewhat
lower. Gross beta MQOs are based on gamma spectrometric analysis (see Attachment 1
descriptions for not using traditional gross beta analysis) with separate MQOs for 228Ra (for
gross beta analysis) and 226Ra (contribution to gross alpha and gross beta activity), based on the
relative sensitivity for these radionuclides using gamma spectrometry. The targeted test levels
and applicable MQOs for Phases 2 and 3 are listed in Table 2.
The required method uncertainty for gross alpha (liquid scintillation count [LSC]
contribution) assumes tolerable error rates of 10% (zi_a = ZI_P= 1.282) for Type I and Type II
errors. It is calculated for gross alpha using the following equation:
,, ., , pCi (AAL-DL) (30-5)
UMR , Soss AlphaLSC>—= T——— -
( 1.282 + 1.282)
25 = 9.8 pCi/L
2.564
Where AAL is the analytical action level and DL is the discrimination level for gross
alpha analysis.
And the value for the relative required method uncertainty is:
(pMR, Gross AlphaLSc, = 100 x MMR/ (AAL) = (9.8 / 30) = 33%
The required method uncertainty for gross alpha (gamma spectrometry contribution) is:
(PMR> t/rOSS AipHdgarnrrla spectrometry ^-MR 33/o
as determined by the gamma spectrometry software algorithm.
July-2014 13
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
The potential contributors to the gross beta activity are 234Th, 234mPa, 214Pb, 214Bi, 212Pb,
212Bi, 208T1, 228Ra, 228Ac, 210Pb, and 210Bi.
Several of these are linked in shorter decay chains so contributions to the gross beta at
full ingrowth can be calculated by measuring one of the radionuclides and assuming secular
equilibrium with the others in the respective decay chain. These would be summarized as
follows:
234Th = 234mPa (full secular equilibrium within 10 minutes)
214Pb = 214Bi (full secular equilibrium within 3 hours, Rn-222 is in secular equilibrium
with Ra-226)
212Pb = 212Bi = 208T1 / 0.36 (full secular equilibrium 72 hours after separation)
228Ra = 228Ac (full secular equilibrium 36 hours)
210Pb = 210Bi (full secular equilibrium 15-25 days after collection)
Since only the non-bolded radionuclides are measured, the bolded radionuclides are
determined inferentially. The activity and uncertainty of the measured values will be used as the
values for the inferred radionuclides 234Th/ 234mPa, 214Pb/ 214Bi, 212Pb/ 212Bi/ 208T1/, 228Ra/ 228Ac,
and 2
The gross beta activity of samples is calculated based on the gamma spectrometric
measurement of beta-gamma emitting radionuclides from the uranium and thorium chains that
are present in detectable levels in the sample. The concentration of beta emitters that are detected
may be calculated by inference when their gamma-emitting parents or progeny where these are
present at detectable levels. Gamma emitting radionuclides are considered to be detectable when
the relative combined standard uncertainty of the radionuclide is determined by the gamma
spectrometry software to be less than or equal to 50%. The overall beta uncertainty is calculated
as the quadratic sum of the uncertainties of the detected radionuclides contributing to the total
gross beta activity.6
Ra-228, based on the gamma emission of its short-lived progeny 228Ac, was chosen in
lieu of other naturally occurring beta-gamma emitters to determine action levels and MDC for
gross beta analysis. This choice is based on three factors:
• The abundance of the 228Ac gamma ray at 91 1 keV is 25.8%. The abundance of the 214Bi
at 609 keV peak is 45.9% and that of 212Pb at 238 keV is 43.6%.
• The efficiency of the 911 keV peak is lower than the other two gamma ray peaks.
• The abundance of 228Ra in many geologic formations used for oil and gas production is
much lower than that of Ra-226 or their respective progeny.
This provides assurance that if the 228Ac action levels and MDC can be confirmed that those
for the other principal naturally occurring beta-gamma chains can also be achieved.
6 The exception for this will be for 226Ra with a WMR = 9.8 pCi/L and a cpMR= 33%.
July-2014 14
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
The targeted required method uncertainty for each of the beta-gamma emitters that will be
measured by gamma spectrometry is 50% at the AAL of 60 pCi/L.
Gross Betagammasectrometry, = 50%
Although the evaluations for the other beta/gamma emitters were not performed pending
development of the gamma spectrometry measurement, further work needs to be performed with
the 224Ra and 226Ra progeny fully ingrown to assess the achievable MDC and AALs. Table 2 also
presents targeted values that will be evaluated once a gamma spectrometry method for a broader
list of analytes has been fully developed.
Table 2 - MQOs Targeted for Gross Alpha and Gross Beta in FPWHFO
Measurement
Gross Alpha
230Th (liquid scintillation)
226Ra (based on the 186 keV
gamma emission)
Gross Beta[4]
228Ra (228Ac)
Beta-emitting progeny of 226Ra
(214Pb/214Bi) and other gamma
emitters in sample[5]
Beta-emitting progeny of 224Ra[5]
(inferred based on gamma
emissions of 212Pb)
Th-234[5] (and inferred 234mPa)
Matrix [1]
FPWHFO
FPWHFO
FPWHFO
FPWHFO
FPWHFO
FPWHFO
AAL
pCi/L
30
150
60
~
~
~
Required
MDC/MDA[2]
pCi/L
5
55
30
~
~
~
u [3]
"MR
pCi/L
9.8
50
30
~
~
~
(PMK™
(%)
33%
33%
50%
50%
50%
50%
[1] A single sample offlowback and produced water supplied by EPA was used as the prototype for a surrogate matrix for method testing.
The method development reference material (MDRM) consisted ofaliquants of the surrogate matrix spiked with radionuclide solutions
traceable to the National Institute of Standards and Technology (NIST).
[2] The MDC represents the value for the individual measurement. A combined MDC for gross alpha or beta will not be determined due to
significant differences in the sensitivity of the respective measurement techniques.
[3] Required method uncertainty at and below the AAL.
[4] Relative required method uncertainty above the AAL.
[5] Preliminary development of the method for gross beta is based on measurement of228Ac (gamma emitting decay progeny of228Ra). Final testing
will be conducted pending further refinement of the gamma spectrometry measurement technique.
A synopsis of the final method testing results for all three test levels and the MDC is
found in Table 3. The 230Th values represent the results by chemical separation and liquid
scintillation counting. The results for 226Ra and 228Ra are from direct sample analysis by gamma
ray spectrometry.
July-2014
15
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 3 - Method Development As-Spiked Concentrations, MQOs and Mean Measured Results
Concentration
Relative to
Action Level
MDC
MDC
MDC
!/2 x AAL
!/2 x AAL
!/2 x AAL
AAL
AAL
AAL
3 x AAL
3 xAAL
3 x AAL
Isotope
230Th
226Ra
228Ra
23oTh
226Ra
228Ra
23oTh
226Ra
228Ra
23oTh
226Ra
228Ra
Targeted
PT
Activity
(pCi/L)
15
60
30
20
90
30
40
180
60
120
540
90
Known
Value111 ±
csu[2]
(pCi/L)
12.45±0.79
55.2±1.5
30.4±0.2
22.5 ±1.2
75.0 ±2.0
41.4 ±0.65
37.5 ±2.1
164.9 ±4.5
62.47 ±
0.99
104.9 ±6.3
449 ± 12
180.1 ±2.8
Measured
Mean ±
SD[3]
(pCi/L)
10.2 ±1.9
85±28
39±12
14.9 ±2.2
94 ±32
43.2 ±9.8
31.4 ±4.2
138 ±18
74 ± 12
73 ± 12
431 ±52
199.3 ±6.8
Required
Method
Uncertainty'41
"MR (pCi/L)
5
55
30
12 pCi/L
54pCi/L
22 pCi/L
12 pCi/L
54 pCi/L
31pCi/L
n/a
n/a
n/a
Required
Relative
Method
Uncertainty
9>MR
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
33%
33%
50%
Measured
Relative
Standard
Deviation
(k=l)
19%
33%
31%
15%
34%
23%
13%
13%
16%
16%
12%
3.4%
[1] The known values were calculated by dilution of the tracers added to the surrogate matrix.
[2] CSU — combined standard uncertainty (k=l).
[3] SD — Standard deviation of the observed measurements
[4] UMR =
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
7.0 Method Development Plan
The methods for gross alpha and gross beta activity were evaluated for four important
performance characteristics for radioanalytical methods specified in Quality Assurance Project
Plan Validation of Rapid Radiochemical Methods for Radionuclides Listed in EPA 's
Standardized Analytical Methods (SAM) for Use During Homeland Security Events (EPA 2012).
These include method uncertainty, detection capability, bias, and analyte activity range. A
summary of the manner in which these performance characteristics were evaluated is presented
below. Chemical yield of the method, an important characteristic for method ruggedness, is not
specifically evaluated on an ongoing basis as this method is not radionuclide specific. Tracers
were used in method development to assess recovery, however, to provide information on
method ruggedness. See Attachment I for further discussion.
Three counting methods were evaluated during the method development to determine
the counting method to be used for method testing. Gas proportional counting was eliminated
as a method of gross alpha or beta analysis because the high concentration of dissolved solids
prevented the achievement of activity concentration detection limits that were reasonable due to
self-attenuation and due to challenges encountered in preparing the test source. The final method
utilizes liquid scintillation counting and high purity germanium gamma spectrometry. The
proposed testing combinations are shown in Table 4.
Table 4 - Proposed Sample Processing and Counting Combinations
Method
Gross Alpha
Gross Beta
Processing Method
Chemical Separation
Non-destructive Count
Gamma
Spectrometry
X
Liquid Scintillation
Counting
X
Gas Proportional
Counting [11
X
[1] Eliminated from final testing.
7.1 Method Uncertainty
The method uncertainty for gross alpha and gross beta activity was initially estimated
based on the gamma analysis of the untreated field sample and spiked demineralized water and
surrogate MDRM sample spikes. Based on those trials, sample target concentrations were
derived and finalized as noted in Table 5.
July-2014
17
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 5 - Actual Sample Concentrations Used for MD Process
Sample Type
Demineralized
Water Blanks
Surrogate MDRM
Blanks
Surrogate MDRM
Spikes
Surrogate MDRM
Spikes
Surrogate MDRM
Spikes
Surrogate MDRM
Spikes
Minimum
Number
Analyzed
7
7
10
7
7
7
Spike
Concentration
Blank
Blank
MDC
!/2 x AAL
AAL
3 x AAL
Gross Alpha
by LSC (as
fraction of
the AAL)
AAL = 37.5
pCi/L
0
0
0.33
0.60
1.0
2.8
226Raby
Gamma
Spectrometry
(as fraction of
the AAL)
AAL = 164.9
pCi/L
0
0
0.33
0.45
1.0
2.7
228Raby
Gamma
Spectrometry
(as fraction of
the AAL)
AAL = 62.47
pCi/L
0
0
0.49
0.65
1.0
2.9
Internal MDRM proficiency test samples were prepared:
• The performance of gross alpha as evaluated using two difference radioelements as
follows:
o Thorium, uranium, and polonium: the performance of the chemical separation / LSC
approach was evaluated based on recovery of 230Th added to the MDRMs since
characterization of the method in Phase 1 demonstrated that Th, U, and Po behave
similarly using this method.
o 226Ra: The alpha activity as determined by gamma Spectrometry was used to validate
the method based on recovery of 226Ra. This MDRM contained the alpha-emitting
radionuclide 226Ra.
• Gross beta as determined by gamma Spectrometry reflects the activity of beta-emitting
• 226T
228-
228T
decay progeny of Ra and Ra and the progeny of the Ra.
The respective concentrations for the test samples are noted in Table 5. The method
was evaluated against the required method uncertainty values identified in Table 2.
July-2014
18
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
7.2 Detection Capability
The detection capability for gross alpha by LSC (Th/U/Po) was estimated at 10 pCi/L.
For 226Ra by gamma spectrometry, it was estimated at 60 pCi/L, and for gross beta based on
228Ra7 for all significant beta emitters was estimated at 30 pCi/L. In accordance with the
guidance provided in the MVG, the critical net concentration (CLNc) of the method was to be
estimated based on the results of at least seven blank samples. Results from ten replicate samples
having an "as tested" concentration at the required MDC being tested were compared to the
critical net concentration to determine method detection capability. It should be noted that the
verification of the required MDC for gross beta failed to meet criteria when tested at the 15
pCi/L level. The test was repeated using test samples spiked at twice the initial activity indicating
that the method is capable of meeting a required minimum detectable concentration value at the
30 pCi/L level.
7.3 Method Bias
Two types of method bias were evaluated, absolute and relative.
Absolute Bias
Absolute bias was determined as a method performance parameter. The results from the
seven blank samples for the required MDC evaluation were evaluated for absolute bias according
to the protocol and equation presented in the Method Validation Guide for Qualifying Methods
Used by Radiological Laboratories Participating in Incident Response Activities (EPA 2009).
Absolute bias was to be determined as a method performance parameter; however, there was no
acceptance limit for bias established for the method in this method development process.
The following protocol was used to test the gross alpha water method for absolute bias:
1. Calculate the mean (X) and estimated standard deviation (sx) for "N" (at least seven) blank
sample net results (in the case of gamma spectrometry measurements 13 were used as the
demineralized water blank and the blank surrogate were indistinguishable from each other)
2. Use the equation below to calculate the |T| value:
X
(1)
T
3. An absolute bias in the measurement process is indicated if:
(2)
where:
T represents the absolute value of the statistical t-value determined from analysis of the
samples analyzed, and
The analysis for 228Ra was performed using the gamma ray emissions of its first progeny, 228Ac, which emits a gamma ray at 911 keV. See
discussion at the end of Section 6.1.
July-2014 19
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
ti-o/2 (N-l) represents the (1 - a/2)-quantile of the t-distribution with N-l degrees of
freedom.
For seven blanks (or in the case of gamma spec 13 blanks), an absolute bias is identified
at a significance level of 0.05, when |T| > t1_a/2.
Relative Bias
The results from the replicate samples for each of the three test levels and at the MDC
level were evaluated for relative bias according to the protocol and equation presented in the
Method Validation Guide for Qualifying Methods Used by Radioanalytical Laboratories
Participating in Incident Response Activities (EPA 2009). No acceptable relative bias limit was
specified for this MD process.
The following protocol was used to test the gross alpha method for relative bias:
1. Calculate the mean (X) and estimated standard deviation (sx) of the replicate results for
each method test level.
2. Use the equation below to calculate the |T| value:
X-K
T= ls2/N + u2 K <3)
where:
17] is the average measured value
sx is the experimental standard deviation of the measured values
N is the number of replicates
K is the reference value
u(K) is the standard uncertainty of the reference value
A relative bias in the measurement process is indicated if:
T>t1 /2(v ) (3a)
The number of effective degrees of freedom for the t-statistic is calculated as follows:
Where veff as calculated by the equation generally is not an integer so veff should be
truncated (rounded down) to an integer. Then, given the significance level, 0.05, the critical
value for T is defined to be ti-a/2(veff), the (1 - a/2)-quantile of the t-distribution with veff degrees
of freedom (see MARLAP Appendix G, Table G.2).
July -20 14 20
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
7.4 Analyte Concentration Range
The gross alpha method was evaluated for the required method uncertainty at three test
level activities. The seven replicate PT samples from each test level concentration were
analyzed. The three "targeted" and "known" test level activities are presented in Table 3. Note
that the final test concentration values for the PT samples varied from the proposed test levels
but that these values were within the sample preparation specifications provided to the
laboratory.
The gross beta method was evaluated for required method uncertainty using the gamma
spectrometry software criterion of < 50% and the combined standard uncertainty as calculated by
the gamma spectrometry software.
7.5 Method Specificity
The purpose of this method is to perform a gross activity measurement. The specificity
relates only to gross alpha or gross beta analysis. The separation steps employed were intended
to isolate groups of radionuclides that were either alpha or beta emitters. In the development of
the final method, certain tracers were used to follow the separation of the surrogate radionuclides
used for the analysis.
7.6 Method Ruggedness
Method ruggedness was evaluated primarily in Phase 1 of the MD process for the effects that the
extreme combination of chemical compounds and high dissolved solids content would have on
the method. In this phase, the method was modified to use a general chemical separation process
for isotopes of uranium, thorium and polonium, and gamma spectrometry was selected for
analysis of beta emitting radionuclides and 226Ra. The ruggedness of the method was testing
during Phase 1 using the EPA provided sample and with different radionuclides spiked into the
surrogate matrix. See Attachment 1 for details on this testing.
July-2014 21
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 22
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
8.0 Techniques Used to Evaluate the Measurement Quality
Objectives for Methods Development Project
A general description of the specifications and techniques used to evaluate the required
method uncertainty, required MDC and bias was presented in Section 6. The detailed method
evaluation process for each MQO and bias is presented in this section.
8.1 Required Method Uncertainty
The gross alpha and gross beta activity method was evaluated following the guidance
presented for "Level E Method Validation: Adapted, Newly Developed Methods, Including
Rapid Methods" in Method Validation Guide for Qualifying Methods Used by Radiological
Laboratories Participating in Incident Response Activities (EPA 2009) and Chapter 6 of Multi-
Agency Radiological Laboratory Analytical Protocols Manual (MARLAP 2004).
MARLAP "Level E" method validation requires the laboratory to conduct a method
development study wherein seven (or more) replicate samples from each of the three concentration
levels are analyzed according to the method. The concentration test levels analyzed are listed in
Table 3. For validation "Level E," internally prepared PT samples consisting of NIST-traceable
230Th, 228Ra and 226Ra spiked into the surrogate matrix to create a method development reference
material (MDRM). In order to determine if the proposed method met the project MQO
requirements for the required method uncertainty, each PT sample result was compared with the
method uncertainly acceptance criteria listed in Tables 6A, 6B, and 6C. "Level E" validation
stipulates that for each test sample analyzed, the measured value must be within ±3.0 wMR8 for
test level activities at or less than the AAL, or ±3.0 cpMR9 for test level activities above the AAL.
8.2 Required Minimum Detectable Concentration
The analytical results reported for the PT samples having a gross alpha and gross beta
activity at the tested MDCs (Th/U/Po by LSC alpha at 12 pCi/L; 226Ra alpha by gamma
spectrometry at 55 pCi/L; 228Ra beta at 30 pCi/L) were evaluated according to Sections 5.5.1 and
5.5.2 of Testing for the Required MDC mMethod Validation Guide for Qualifying Methods
Used by Radiological Laboratories Participating in Incident Response Activities (EPA 2009).
For this method development process, the terms "MDC" and "Critical Net Concentration" are used.
The State Hygienic Laboratory at the University of Iowa analyzed the prepared PT samples in
accordance with the final proposed method.
Critical Net Concentration
In order to evaluate whether the combined method can meet the required MDC
(Th/U/Po by LSC alpha at 12 pCi/L; 226Ra alpha by gamma spectrometry at 55 pCi/L; 228Ra beta
at 30 pCi/L), the critical net concentration, as determined from the results of analytical blanks,
must be calculated. The critical net concentration (CL>jc) with a Type I error probability of a =
is the required method uncertainty
9
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
0.05, was calculated using the following equation (consistent with MARLAP, Chapter 20,
Equation 20. 35):
i) = tl_a(n-\)xSBlanks (7)
Where:
-^Blanks is the standard deviation of the n blank-sample net results (corrected for
instrument background) in radionuclide concentration units of pCi/L, and
ti-a(n-i) is the (1 - a)-quantile of the t-distribution with n-1 degrees of freedom (see
MARLAP Table G.2 in Appendix G). For this method development study, a Type I error
rate of 0.05 was chosen.
For example, for seven (minimum) blank results (six degrees of freedom) and a Type I error
probability of 0.05, the previous equation reduces to:
(8)
The use of the above equations assumes that the method being evaluated has no bias.
RequiredMDC
Each of the analytical results reported for the PT samples having concentration at the
required MDC (Th/U/Po by LSC alpha at 12 pCi/L; 226Ra alpha by gamma spectrometry at 55
pCi/L; 228Ra beta at 30 pCi/L) were compared to the estimated critical net concentration for the
method. The following protocol was used to verify a method's capability to meet the required
method MDC for a radionuclide-matrix combination:
I. Analyze a minimum of seven matrix blank samples for the radionuclide.
II. From the blank sample net results, calculate the estimated Critical Net Concentration,
U/jNC-
III. Analyze ten replicate samples spiked at the required MDC.
IV. From the results of the ten replicate samples spiked at the required MDC, determine the
number (Y) of sample results at or below the estimated
V. If Y < 2, the method evaluated at the required MDC passes the test for the required MDC
specification.
VI. If Y > 2, the method evaluated at the required MDC fails the test for the required MDC
specification.
July-2014 24
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
9.0 Evaluation of Experimental Results
Only the experimental results for Phases 2 and 3 of the MD process are reported and
evaluated in this study. Information presented in this section will include results for Sections 6
and 7. The gross alpha and gross beta activity concentration results were evaluated for the
required method uncertainty, required MDC, and bias.
9.1 Summary of the Method
The full method for the analysis is located in Attachment III. A brief description of
each part of the method is described here.
Th, U, Po Gross Alpha by LSC
A 300 mL sample is treated with permanganate in ammonaical solution to coprecipitate
all radionuclides in the naturally occurring series. The mixture is settled and the supernatant
solution discarded. The residue is dissolved in hydrochloric acid and passed through a TRU
Resin™ column. 10 Thorium, uranium and polonium are eluted from the resins using ammonium
oxalate. The eluent is transferred to a liquid scintillation vial as the final sample test source and
counted in the alpha emission region of the liquid scintillation spectrum. The final sample test
source is counted for 30 minutes.
226Ra Alpha by Gamma Spectrometry
A sample test source of 3 L is counted directly by gamma ray spectrometry. The
activity concentration of 226Ra is determined by gamma ray analysis using the 186 keV gamma
ray region (assumes no significant amount of 235U compared to the amount of 226Ra in the
sample). The sample is counted for 12 hours or long enough to achieve the required method
uncertainty.
228Ra Beta by Gamma Spectrometry
The sample test source of 3 L is counted directly by gamma ray spectrometry. In
addition to 228Ra being determined indirectly by analysis of its first progeny 228Ac (911 or 338
keV) several other radionuclides are determined directly by their gamma ray emission and others
may be determined indirectly through analysis of a parent or progeny radionuclide that is a
gamma ray emitter. The sample is counted for 12 hours or long enough to achieve the required
method uncertainty.
9.2 Required Method Uncertainty
Tables 6 A through 6C show the measured results for each of the samples analyzed and
the testing requirements for each level. The "known" values and combined standard
uncertainty of the known values ("CSU known") are based on the amount of NIST-traceable
standards added to the samples. The allowable range of results is calculated by taking the known
' Eichrom Technologies, LLC, Lisle, IL.
July-2014 25
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
value and adding (to get the upper value of the range) and subtracting (to get the lower value of
the range)
three times the required method uncertainty (3.0 x UMR). For example, in Table 6A, the
allowable range for gross alpha for Test Level 2 is 37.5 ± (3.0 x 12) pCi/L or 1.5 to 74 pCi/L
(rounded to two significant figures).
July-2014 26
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 6A- Th, U. Po Gross Alpha by LSC—Analytical Results for Required Method Uncertainty Evaluation
Nuclide: Gross Alpha (Th-230) Matrix: FWHFPO AAL Tested: 37.5 pCi/L
Proposed Method: U, Th, Po Gross Alpha by LSC
Required Method Validation Level: MARLAP "E"
Required Method Uncertainty, MMB: 12 pCi/L at and below AAL; 33% above AAL
Acceptance Criteria:
Test Levels 1 and 2: 3.0 x «MR = known value of sample in test level ± 36 pCi/L
Test Level 3: 3.0 x g>MK = quoted known value of sample ± 99% of known value of sample (pCi/L)
Test Level 1
Test Value = 22.5 ± 1.2 pCi/L
Sample #
1
2
3
4
5
6
7
pCi/L
Known
22.5
csu*
Known
1.2
pCi/L
Measured
13.68
18.1
14.93
11.74
15.7
13.45
16.9
csu**
Measured
0.96
1.1
0.99
0.92
1.0
0.96
1.0
Allowable Range
(pCi/L)
-14 to 59
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Test Level 2
Test Value = 37.5 ± 2.1 pCi/L
Sample #
1
2
3
4
5
6
7
8
pCi/L
Known
37.5
CSU*
Known
2.1
pCi/L
Measured
31.0
38.8
32.8
32.0
24.8
33.0
27.2
31.5
CSU**
Measured
1.3
1.5
1.3
1.3
1.2
1.4
1.2
1.3
Allowable Range
(pCi/L)
1.5 to 74
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Y
Test Level 3
Test Value = 104.9 ± 6.3 pCi/L
Sample #
1
2
3
4
5
6
7
pCi/L
Known
104.9
CSU*
Known
6.3
pCi/L
Measured
72.6
60.7
67.5
60.2
76.5
86.5
89.3
CSU**
Measured
2.0
1.8
1.9
1.8
2.1
2.2
2.3
Allowable Range
(pCi/L)
1.0 to 210
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Quoted uncertainty (one sigma)
Combined standard uncertainty, k = 1
July-2014
27
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 6B - Ra Alpha by Gamma Spectrometry - Analytical Results for Required Method Uncertainty Evaluation
Nuclide: Gross Alpha (226Ra) Matrix: FWHFPO AAL Tested: 164.9pCi/L
Proposed Method: 226Ra Alpha by Gamma Spectrometry
Required Method Validation Level: MARLAP "E"
Required Method Uncertainty, «MB : 54 pCi/L at and below AAL; 33% above AAL
Acceptance Criteria:
Test Levels 1 and 2: 3.0 x UMR = known value of sample in test level ±162 (pCi/L)
Test Level 3: 3.0 x (3MR = quoted known value of sample ± 99% of known value (pCi/L)
Test Level 1
Test Value = 75.0 ± 2.0 pCi /L
Sample #
1
2
3
4
5
6
7
pCi/L
Known
75.0
CSU*
Known
2.0
pCi/L
Measured
58.55
93.7
107.3
101
153
62.5
84.1
csu**
Measured
18.1
41.8
36.45
41.8
42.8
19.5
33.2
Allowable Range
(pCi/L)
-87 to 240
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Test Level 2
Test Value = 164.9 ± 4.5 pCi/L
Sample #
1
2
3
4
5
6
7
8
pCi/L
Known
164.9
CSU*
Known
4.5
pCi/L
Measured
125.8
140
144
110
167
131
150
125.8
CSU**
Measured
22
21.6
22.2
22.6
43.6
22.9
22.7
22
Allowable Range
(pCi/L)
2. 9 to 330
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Y
Test Level 2
Test Value = 37.5 ± 2.1 pCi/L
Sample #
1
2
3
4
5
6
7
8
pCi/L
Known
37.5
CSU*
Known
2.1
pCi/L
Measured
31.0
38.8
32.8
32.0
24.8
33.0
27.2
364
CSU**
Measured
1.3
1.5
1.3
1.3
1.2
1.4
1.2
32.2
Allowable Range
(pCi/L)
1.5 to 74
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Y
Quoted uncertainty (one sigma)
Combined standard uncertainty, k=l.
July-2014
28
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 6C - Ra Beta by Gamma Spectrometry—Analytical Results for Required Method Uncertainty Evaluation
Nuclide: Gross Beta (228Ra) Matrix: FWHFPO AAL Tested: 62.47 pCi/L
Proposed Method: 228Ra Beta by Gamma Spectrometry
Required Method Validation Level: MARLAP "E"
Required Method Uncertainty, UUP'. 3 1 pCi/L at and below AAL; 50% above AAL
Acceptance Criteria:
Test Levels 1 and 2: 3.0 x UMR = known value of sample in test level ± 93 pCi/L
Test Level 3 : 3 .0 x /pMR = quoted known value of sample ± 1 50% of known value (pCi/L)
Test Level 1
Test Value = 41.4 ± 0.65 pCi/L
Sample #
1
2
3
4
5
6
7
pCi/L
Known
41.40
CSU*
Known
0.65
pCi/L
Measured
44.9
50.2
54
39.2
52.2
30.6
31.1
CSU**
Measured
6.03
7.13
6.46
4.61
5.08
5.07
5.5
Allowable Range
pCi/L
-52 to 130
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Test Level 2
Test Value = 62.47± 0.99 pCi/L
Sample #
1
2
3
4
5
6
7
8
pCi/L
Known
62.47
csu*
0.99
pCi/L
Measured
77.2
80.2
53.2
82.4
85.3
62.3
76.7
77.2
CSU**
6.67
8.44
6.13
10.5
10.5
6.52
9.17
6.67
Allowable Range
pCi/L
-31 to 160
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Y
Test Level 2
Test Value = 37.5 ± 2.1 pCi/L
Sample #
1
2
3
4
5
6
7
pCi/L
Known
37.5
csu*
Known
2.1
pCi/L
Measured
31.0
38.8
32.8
32.0
24.8
33.0
27.2
CSU**
Measured
1.3
1.5
1.3
1.3
1.2
1.4
1.2
Allowable Range
(pCi/L)
1.5 to 74
Acceptable
Y/N
Y
Y
Y
Y
Y
Y
Y
Quoted uncertainty (one sigma)
Combined standard uncertainty, k=l.
July-2014
29
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Based on the results in Tables 6A to 6C it may be concluded that:
• A required method uncertainty of 12 pCi/L can be achieved at an action level of 38 pCi/L
for gross alpha analysis using LSC
• A required method uncertainty of 54 pCi/L can be achieved at an action level of 165
pCi/L for 226Ra analysis using gamma spectrometry
• A required method uncertainty of 31 pCi/L can be achieved at an action level of 62 pCi/L
for 228Ra analysis using gamma spectrometry.
These results meet the measurement quality objectives from the method development plan.
Tables 7 A to 7C provide summary information regarding the expected variability of the
results for an individual analysis and test level. In each case, the experimental standard deviation
was considerably less than the required method uncertainty indicated that the method is capable
of meeting the method uncertainty MQO.
Table 7A- Th, U, Po Gross Alpha by LSC—Experimental Standard Deviation of the Seven PT Samples by Test Level (^h)
Test Level
1
2 (AAL)
3
Mean
Concentration Measured
(pCi/L)
14.9
31.4
73
Experimental Standard
Deviation'11
(pCi/L)
2.2
4.2
12
Required Method
Uncertainty
(pCi/L)
12
12
35[2]
[1] Standard deviation of the seven measurements.
[2] Calculated by multiplying the known value of Test Level 3 by the required relative method uncertainty.
Table 7B - 226Ra Alpha by Gamma Spectrometry—Experimental Standard Deviation of the Seven PT Samples by Test Level
Test Level
1
2 (AAL)
3
Mean
Concentration Measured
(pCi/L)
94
138
431
Experimental Standard
Deviation'11
(pCi/L)
32
18
52
Required Method
Uncertainty
(pCi/L)
54
54
150[2]
[1] Standard deviation of the seven measurements.
[2] Calculated by multiplying the known value of Test Level 3 by the required relative method uncertainty.
July-2014
30
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 7C - 228Ra Beta by Gamma Spectrometry—Experimental Standard Deviation of the Seven PT Samples by Test Level
Test Level
1
2(AAL)
3
Mean
Concentration Measured
(pCi/L)
43.2
74
199.3
Experimental Standard
Deviation'11
(pCi/L)
9.8
12
6.8
Required Method
Uncertainty
(pCi/L)
31
31
90[2]
[1] Standard deviation of the seven measurements.
[2] Calculated by multiplying the known value of Test Level 3 by the required relative method uncertainty.
9.3 Required Minimum Detectable Concentration
Critical Net Concentration
The method for Th, U, Po gross alpha by LSC in an FWHFO matrix was evaluated for
a required MDC using surrogate water blanks that were taken through the analytical processes of
gross alpha chemical separation and LSC. Ra-226 and Ra-228 were evaluated for the MDC by
direct analysis of the MDC verification samples by gamma counting.
Demineralized water blanks for the gross alpha determination by LSC yielded an
expected blank value of zero since the average of these blanks was used to determine the
background subtraction value, while surrogate samples yielded values greater than zero. The
elevated result for surrogates is likely due to the contribution of low levels of naturally occurring
radionuclides in the reagents used to make the surrogate. Thus, the surrogate water blanks were
used to determine a matrix appropriate critical net concentration for the verification of the MDC.
The results of the demineralized water and surrogate blanks for gross alpha are shown in Tables
SAandSB.
Table 8A - Th, U, Po Gross Alpha by LSC—Blank Water Samples
Sample ID
DI Water Blank 1
DI Water Blank 2
DI Water Blank 3
DI Water Blank 4
DI Water Blank 5
DI Water Blank 6
DI Water Blank 7
Mean** (pd/Liter)
Standard Deviation ** (pCi/Liter)
Critical Net Concentration (pCi/Liter)
Concentration (pCi/L)
0.07
-0.12
0.54
-0.12
-0.59
0.07
0.15
0. 00
0.34
0.66
CSU* (pCi/L)
0.69
0.68
0.71
0.68
0.66
0.69
0.69
* Combined standard uncertainty, k=l or coverage factor of 1.
* * Mean and standard deviation were calculated before rounding.
July-2014
31
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 8B - Th, U, Po Gross Alpha by LSC—Surrogate Water Samples
Sample ID
Surrogate Blank 1
Surrogate Blank 2
Surrogate Blank 3
Surrogate Blank 4
Surrogate Blank 5
Surrogate Blank 6
Surrogate Blank 7
Mean** (pCi/Liter)
Standard Deviation ** (pCi/Liter)
Critical Net Concentration (pCi/Liter)
Concentration (pCi/L)
8.86
6.95
7.92
6.53
7.81
7.03
7.03
7.45
0.79
1.5
CSU* (pCi/L)
0.85
0.85
0.83
0.79
0.82
0.80
0.80
Combined standard uncertainty, k=l or coverage factor ofl.
Mean and standard deviation were calculated before rounding.
The results for the demineralized water blanks and surrogate blanks measured by
gamma ray spectrometry were indistinguishable from each other (i.e., statistically the same).
Therefore these data were pooled and used to calculate the critical level net concentration for the
226Ra and 228Ra measurements. The pooled results are shown in tables 8C and 8D.
226
Table 8C - Ra Alpha by Gamma Spectrometry in Demineralized and Surrogate Water Samples
Sample ID
Surrogate 1
Surrogate 2
Surrogate 4
Surrogate 5
Surrogate 6
Surrogate 7
Surrogate 8
DM Water Blank 1
DM Water Blank 2
DM Water Blank 3
DM Water Blank 5
DM Water Blank 6
DM Water Blank 7
Mean** (pCi/Liter)
Standard Deviation ** (pCi/Liter)
Critical Net Concentration (pCi/Liter)
Concentration (pCi/L)
-12.6
-17.9
-15.78
-8.99
14.3
-12.1
-6.73
53.5
-6.4
-1.94
-45
-18.3
-9.94
-7
22
40
CSU* (pCi/L)
17.3
18.3
34
19.5
24.5
17.4
22.9
26
17.8
28.6
30.7
24.8
21.3
Combined standard uncertainty, k=l or coverage factor of 1.
' Mean and standard deviation were calculated before rounding.
July-2014
32
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 8D - Ra Gross Beta by Gamma Spectrometry in Demineralized and Surrogate Water Samples
Sample ID
Surrogate 1
Surrogate 2
Surrogate 4
Surrogate 5
Surrogate 6
Surrogate 7
Surrogate 8
DM Water Blank 1
DM Water Blank 2
DM Water Blank 3
DM Water Blank 5
DM Water Blank 6
DM Water Blank 7
Mean** (pCi/Liter)
Standard Deviation ** (pCi/Liter)
Critical Net Concentration (pCi/Liter)
Concentration (pCi/L)
11.8
13.3
14.9
-9.54
-16.2
5.65
-11.8
-10.7
-1.87
-17.1
-9.42
-3.08
-11.7
-4
11
20
CSU* (pCi/L)
5.24
6.62
5.7
11.1
23.8
4.24
8.19
13.4
1.2
34.1
10.1
4.34
13.2
* Combined standard uncertainty, k=l or coverage factor of 1.
* * Mean and standard deviation were calculated before rounding.
RequiredMDC
The method development plan based the estimated MDC values for gross alpha and gross
beta on demineralized water blanks. The initial estimates of the MDCs targeted for 226Ra and
OOQ
Ra proved to be too low and were adjusted to a higher concentration during the method development
process. The most likely reason for this was increased background counts from low
activity concentration of natural radioactivity in the surrogate matrix.
The MDC data for each of the three evaluations are shown in Tables 9A through 9C.
The determination of gross alpha using 226Ra as the radionuclide and gamma spectrometry as the
measurement technique is not as sensitive as using liquid scintillation as the measurement
technique.11 For this reason, the two different techniques for the gross alpha measurement have
different MDC values. This is highlighted in the MDC results shown in Tables 9A and 9B.
11 The abundance of the 186 keV gamma ray for 226Ra is only 3.3 % while the abundance of the alpha particle emission is 100 %.
July-2014
33
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 9A - Reported Results for Samples Containing Th, U, Po Gross Alpha by LSC[1] at the As-Tested MDC
Value (12.45 pCi/L)
Sample ID
MDC1
MDC 2
MDC 3
MDC 4
MDC 5
MDC 6
MDC 7
MDC 8
MDC 9
MDC 10
Concentration (pCi/L)
12.08
7.73
13.72
9.05
11.89
8.97
8.62
10.10
10.57
9.21
CSU[2] (pCi/L)
0.93
0.82
0.96
0.85
0.92
0.85
0.84
0.88
0.89
0.86
Mean[3](pCi/L)
Standard Deviation of Results [3] (pCi/L)
CXNC[4](pCi/L)
Acceptable maximum values < CLNC
Number of results > CLNC
Number of results < CLNC
Evaluation
Test Result
< CLNC (Y/N)
N
N
N
N
N
N
N
N
N
N
10.2
1.9
1.5
2
10
0
PASS
[1] Samples spiked with 230Th.
[2] Combined standard uncertainty, coverage factor k=1.
[3] Mean and standard deviation were calculated before rounding.
[4] Critical net concentration. CLNC based on the water blanks taken through the entire method.
July-2014
34
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 9B - Reported Results for 226Ra Alpha'11 by Gamma Spectrometry at the As-Tested MDC Value (55.2 pCi/L)
Sample ID
MDC1
MDC 2
MDC 3
MDC 4
MDC 5
MDC 6
MDC 7
MDC 8
MDC 9
MDC 10
Concentration (pCi/L)
148
99.6
79.2
51.7
102
74.7
72.8
68
58.9
98.1
CSU[2] (pCi/L)
42.1
42.4
34.13
17
44.9
31.8
36.6
33.2
25.4
34
Mean[3](pCi/L)
Standard Deviation of Results [3] (pCi/L)
CLNC[4](pCi/L)
Acceptable maximum values < CLNC
Number of results > CZNC
Number of results < CLNC
Evaluation
Test Result
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 9C - Reported Results for 228Ra Beta'11 by Gamma Spectrometry at the As-Tested MDC Value (30.4 pCi/L)
Sample ID
MDC1
MDC 2
MDC 3
MDC 4
MDC 5
MDC 6
MDC 7
MDC 8
MDC 9
MDC 10
Concentration (pCi/L)
35.8
44.2
57.9
35.5
58.3
40.5
26.2
34
30.8
26.2
CSU [2] (pCi/L)
6.03
5.45
5.55
5.59
6.6
4.53
5.06
3.52
4.02
5.31
Mean[3](pCi/L)
Standard Deviation of Results [3] (pCi/L)
CXNC[4](pCi/L)
Acceptable maximum values < CLNC
Number of results > CLNC
Number of results < CLNC
Evaluation
Test Result
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Table 10 - Absolute and Relative Bias for Gross Alpha and Gross Beta by the Three Measurements
Type of
Bias
Absolute
MDC
Level 1
Level 2
Level 3
Gross
Measurement
(Radionuclide
Measured)
Gross Alpha
(230Th, LSC)
Gross Alpha
(226Ra, GS)
Gross Beta
(228Ra/228Ac, GS)
Gross Alpha
(230Th, LSC)
Gross Alpha
(226Ra, GS)
Gross Beta
(228Ra/228Ac, GS)
Gross Alpha
(230Th, LSC)
Gross Alpha
(226Ra, GS)
Gross Beta
(228Ra/228Ac, GS)
Gross Alpha
(230Th, LSC)
Gross Alpha
(226Ra, GS)
Gross Beta
(228Ra/228Ac, GS)
Gross Alpha
(230Th, LSC)
Gross Alpha
(226Ra, GS)
Gross Beta
(228Ra/228Ac, GS)
Known
Value ±
CSU(A=1),
pCi/L
-
-
-
12.45±0.79
55.2±1.5
30.4±0.2
22.5±1.2
75.0±2.0
41.40±0.65
37.5±2.1
164.9±4.5
62.47±0.99
104.9±6.3
449±12
180.1±2.8
Mean of
Measurement
±SD(A=1),
pCi/L
7.45 ± 0.79
-7 ±22
-4 ±11
10.2 ±1.9
85±28
39±12
14.9±2.2
94±32
43.2±9.8
31.4±4.2
138±18
74±12
73±12
431±52
199.3±6.8
Difference
from
Known,
pCi/L(%)
7.45
7
4
2.25 (-18)
27.2 (54)
8.6 (28)
7.6 (-34)
19 (25)
1.8(4.3)
6.1 (-16)
26.9 (-16)
11.5(18)
3 1.9 (-30)
18 (-4.0)
19.2(11)
Number of
Measurements/
(Effective)
Degrees of
Freedom111
7/6
13/12
13/12
10/70
10/9
10/9
7/64
7/6
7/6
8/61
8/12
8/6
7/55
8/11
8/28
T [i]
2.447
2.179
1.12
2.28
3.36
2.32
5.05
1.59
0.47
2.39
3.23
2.52
4.14
0.76
5.06
t W
tdr
24.8
1.09
2.17
9
1.99
2.26
2.26
2.00
2.45
2.45
2.00
2.18
2.45
2.00
2.20
2.05
Biaspl
(Y/N)?
Y
N
N
Y
Y
Y
Y
N
N
Y
Y
Y
Y
N
Y
[1] Assessment of relative bias was performed at MDC and Levels 1, 2, and 3.
[2] | T | is the absolute value of the f-statistic, and fjf is the effective degrees of freedom of the f-statistic. They are defined in Section 7.3 and
represent the statistical factors for the actual measurement data and for data from a normal distribution.
July-2014
37
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
9.5 Method Ruggedness
Ruggedness represents the ability of a method to be unaffected by changes in the
components of the sample or the measurement system. The matrix selected for this project was
based on that of a single sample of FPWHFO and will not necessarily represent method
performance in samples of significantly different composition such as FPWHFO collected at
different times in the life of a well or from different wells or regions. Nonetheless, this matrix is
considered to be a particularly challenging one due to its extremely high dissolved solids content
and its complexity (very elevated levels of divalent cations). This affected the bias and
detectability of each of the three measurements associated with the method.
Based on testing performed at the end of the method development process, gross alpha
activity determined by liquid scintillation counting of the separated U, Th, and Po fraction of the
sample was measured with recoveries ranging from 84-96 % of the expected value (see Section
11 and Attachment II, "Flow Chart for the Alpha Beta Methods Attempted " and Figure 6). The
low bias was confirmed during formal development testing with an average gross alpha recovery for
the 14 test measurements at test levels 2 and 3 of 77±11% (k=\) of the known value and
results ranging from 57-104 %.13 Although these results easily meet testing acceptance
criteria, the systematic low bias and the observation that quantitative (or near-quantitative)
recovery of 230Th is not achieved raises concerns about the ruggedness of the chemical separation
method. Final method testing was performed using 230Th, which had the lowest recovery of the
three radioelements evaluated in final preliminary tests, but not uranium or polonium. Since
uranium and polonium were not quantitatively recovered in preliminary testing, it is reasonable
to expect similar low bias for these elements with real test samples.
Gamma spectrometry, by its nature as a non-destructive method, is more rugged than
methods that rely on chemical separations since there can be no losses prior to the analytical
measurement and variability of the matrix will have less potential to impact analytical results.
The average recovery calculated from values presented in Table 6B for the 14 test-level-2 and 3
analytical measurements of 226Ra was 90±13% (k=l) with individual values ranging from 67-
108%. The average recovery calculated from values presented in Table 6C for the 14 analytical
measurements of 228Ra spikes at test levels 2 and 3 was 114±14% (£=1) with individual values
ranging from 85-137%. While it likely would be possible to improve the process used for the
gamma spectrometry measurement, better control of the sample (i.e., no introduction of added
variability from chemical separations) makes gamma spectrometry a very rugged technique.
Thus, there is little concern that measured values will adequately reflect the performance of the
gamma spectrometry technique on real-world samples.
' Note that the level of uncertainty for samples for the verification of the MDC and test level 1 due to the low concentrations in these samples.
July-2014 38
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
10.0 Timeline to Complete a Batch of Samples
It is important for the laboratory to know how long an analysis will take so that they
can provide timely results back to their clients. Historically a batch of gross alpha and gross beta
analyses will take on the order of 3-4 days (based on the SDWA methods). There are two
separate timelines for the analysis based on the two different analytical methods that are used for
measuring the gross alpha and gross beta activity concentrations. Gross Th, U, and Po using
gross alpha analysis by LSC, from sample preparation to final results reporting, can take place in
3 hours for a single sample, or approximately one day for a batch of 20 samples (assuming a
single LSC is available). The gamma spectrometry count for 228Ra beta and 226Ra alpha can
require about 12 hours for a single sample and can process about 2 samples per day per
spectrometer. Assuming a single gamma ray spectrometer is available, a batch of 20 samples
would require 10 days for analysis. Laboratories will need to judge their processing times for a
batch of samples based on the number of instruments they have available.
See Attachment II for a breakdown of the times for the individual steps shown in the
flowchart in Section 17.4.
July-2014 39
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 40
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
11.0 Reported Modifications During Development and
Recommendations for Future Work
The development of a straight-forward, inexpensive, single measurement method for
gross alpha and gross beta screening of FPWHFO samples was hampered by the complexity of
the matrix itself. Significant preliminary effort was spent in Phase 1 modifying the approach
initially conceived to accommodate a FPWHFO matrix which turned out to be particularly
challenging. The development process will be briefly summarized here to provide perspective.
Attachment I is a more complete presentation of various trials and modification that had to be
made during the Phase 1 method development process that preceded the final method development
process.
Several classic water methods for gross activity testing were explored in hopes that a
simple non-destructive analysis (i.e., an analysis without chemical separation) could be
performed. Unfortunately solids content in the gram per liter range is 200-1000 times above the
range that Method 900.0 evaporation or direct counting by liquid scintillation (e.g., ASTM
D7282) can accommodate. Digestion and fusion techniques were considered but in view of the
high non-volatile solids content of the sample matrix, would not have made the matrix more
amenable to processing. All attempts to use these classic approaches failed to yield acceptable
results.
A variety of classic chemical separation techniques were explored for chemically
isolating radium, thorium, uranium, or polonium. Among the approaches explored were barium
sulfate precipitation techniques (e.g., EPA Method 900.1 or 903.0, SM 71 IOC), barium chromate
precipitations (e.g., EPA Method 905.0) and solid phase extraction with chelating agents, such as
Empore™ Radium RadDisks. Again, the gram per liter levels of barium, strontium, and calcium
interfered with every attempt to chemically isolate radium (also a divalent cation) together with
thorium, uranium, and polonium.
A radon deemanation method (Rad7™ electronic radon detector) proved to be
reasonably successful but this approach had to be abandoned due to concerns about foaming of
the sample and extended turn-around times as long as 7-21 days (required for ingrowth of
222Rn). Other radon emanation techniques, such as mineral oil extraction of radon with liquid
scintillation counting or radon deemanation similar to the classic approach used in EPA Method
903.1 could be considered as possible avenues for future exploration for more sensitive
T7£\
radionuclide-specific determinations of Ra.
One of the methods used to initially characterize the sample, gamma spectrometry,
proved to be one of the most viable, attractive, and rugged alternatives for the analysis of
FPWHFO. The primary advantage of gamma spectrometry is its ability for simultaneous
measurement of all gamma emitters present in the sample together with its nominal insensitivity
to solids in the sample. In contrast to alpha and beta emissions, where sample solids above 100-
200 mg interfere with measurements, the highly penetrating nature of gamma rays allows
measurement of solid samples in the 3-5 kg range without chemical separations. The response of
gamma spectrometers is also limited relative to that of liquid scintillation counters. Gamma
July-2014 41
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
detection efficiencies are generally less than 10% of the efficiency for either alpha or beta
emissions by liquid scintillation. Similarly, the gamma emission rate associated with natural
chain alpha and beta emitters (3% to 45%) is low relative to the corresponding alpha and beta
particle emission rates (nearly 100%). By increasing the size of the aliquant and the count time,
however, it is possible to compensate, as least in part, for low response in the gamma detector.
As a result, detection capabilities may be obtained for gamma spectrometry that are generally
within a factor of 10-50 of those possible by direct measurement of alpha and beta emissions.
Gamma spectrometry was ultimately the technique chosen to perform determinations of 226Ra
and gross beta.
Efforts were focused on finding a group chemical separation for thorium, uranium and
polonium. Iron hydroxide coprecipitation was initially explored as a preconcentration technique
but due to the possibility of high iron content in samples that would interfere with chemical
separations and subsequent liquid scintillation measurements, the decision was made to use a
manganese dioxide coprecipitation in reducing environment (i.e., to remove iron). A number of
extraction chromatographic techniques were tested including a stacked TEVA-TRU followed by
Sr Resin approach. Sr Resin is effective only for polonium and would be overwhelmed by
barium resulting in low yields. Testing of the TRU Resin alone, however, showed excellent
uptake of thorium, uranium, and polonium from a 4 M HC1 load solution (containing ascorbic
acid to reduce iron) and effective retention of the analytes during 4 M HC1 rinse steps. Elution
profiles (see Attachment 1, Figure 5) indicated that the three analytes could effectively be
stripped from the column with 5 mL of ammonium bioxalate. Analysis of subsequent data
indicated that a more complete recovery of analytes, especially uranium, was obtained by
modifying the single 5-mL elution to five sequential 1-mL additions of bioxalate.
Source preparation techniques were also investigated. A number of coprecipitation
techniques were explored but obtaining quantitative coprecipitation for all three radioelements in
a single step proved to be elusive. In the end, a simple technique was chosen - direct addition of
the eluent to liquid scintillation cocktail.
Pulse-shape discrimination liquid scintillation analysis is an attractive option because it
was less sensitive to solids in the sample than gas flow proportional counting, has very high
detection efficiencies due to the 4rc counting geometry (-90%), and reasonably low backgrounds.
The result was reasonably short count times (e.g., 30 minutes) and good detection capabilities.
The final method for gross alpha (Th, U, and Po) by liquid scintillation was thus
assembled based on the most successful approaches identified: manganese dioxide
coprecipitation followed by group separation of thorium, uranium and polonium on TRU Resin,
stripping with ammonium bioxalate, and pulse-shape discrimination liquid scintillation analysis.
oin on ono
The draft method was tested using Th, U, and Po tracers and recoveries of-89% and
above were obtained for each of the three elements with 230Th, the test isotope selected for testing,
showing the lowest recoveries (See Attachment 1, Figure 6).
Gamma spectrometry was selected for the 226Ra alpha measurement and the gross beta
measurement. Beyond ensuring that solids remained suspended using an agar gelling technique,
the classic technique for gamma spectrometry was used without significant modification.
July-2014 42
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
No technical changes were made to the final measurement techniques following the
initiation of development testing. Since the technical content of the method was finalized prior to
Phase 2 and 3 testing, the data obtained in Phases 2 and 3 alone should be used reflect the
performance of the final method as presented in Attachment III.
The method validated for gross alpha by LSC measures the activity of isotopes of
thorium, uranium and polonium. Alpha-emitting radium isotopes, which are likely to be more
prevalent in these samples (in terms of activity concentration), could not be easily or
economically incorporated into a single-step gross alpha method at this time. The uncertainty for
alpha-emitting radium isotopes determined by gamma spectrometry is much larger than that for
the liquid scintillation analysis for the other alpha-emitting isotopes.14 This result complicates the
reporting process, and the determination of uncertainty, and prevents calculation of a single
meaningful value for gross alpha detection capability.
While radionuclide-specific determinations of the naturally occurring progeny of
uranium and thorium decay chain members are possible by gamma spectrometry, the same
differences in efficiency and abundance will result in counting times that are significantly longer
than that for the alpha analyses. As mentioned in Section 1, gross alpha and beta activity is
temporally dependent due to the inherent complexity of radioactive ingrowth and decay in the
uranium and thorium decay chains. As a result of this, the physical activity of some of the
radionuclides being measured will change over time such that the results at the beginning of an
analysis sequence may not be comparable to those at the end of the run. Although the activity of
long-lived radionuclides such as 226Ra would not be impacted, delaying count times for
measurements of shorter-lived beta-emitters such as 212Pb, 214Pb, and 214Bi will have an impact
on the activity measured.
Overall the detection limits achieved, the MDC target and the action levels that were
used were based on the results of the preliminary testing. These values were not as low as had
originally been hoped for and can be improved upon with further analysis and development studies.
Thus, two sets of recommendations follow for use of this method and possible further
development. The first set of recommendations falls within the scope of the current method and
possible improvements (recommendation 8) on it:
1. Sample results using this method should be reported separately as:
a. Gross alpha (U, Th, Po) - LSC
b. Gross alpha (226Ra) - Gamma Spectrometry15
The magnitude of differences in uncertainty and counting time is attributable to the relative detection efficiency of liquid scintillation and
gamma spectrometry instruments and differences in the abundance of decay particle emission rates (i.e., abundance) for the radionuclides. The
detection efficiency of gamma spectrometry instruments is generally less than 10% of that for liquid scintillation counters. The abundance for
gamma rays is often significantly less than 100%, but for radionuclides that decay by alpha emission the abundance is close to 100%.
15 Combined gross alpha results provide no information about the radionuclides present. For example, it is not uncommon to find high levels of
radium in FPWHFO samples even though there may be next to no U, Th, or Po present. By separately screening samples for Ra-226, and for
alpha activity associated with U, Th, or Po, it may be possible to decide there is no need for radionuclide-specific testing for U, Th, or Po. This
strategy can be used to streamline the characterization of samples.
July-2014 43
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
c. Gross beta - Gamma Spectrometry
2. In principal, a screening test should be designed to minimize the risk of not detecting analyte
present at the analytical action level to a rate that is acceptable in the decisionmaking
process. Low bias in a screening measurement may be tolerable as long as the criteria used to
decide that the analytical action level has been exceeded take into account uncertainty and
the bias associated with the measurement. Thus, additional work is needed to estimate the
bias and uncertainty associated with the Th, U, Po gross alpha measurement, and to propose
decision rules that ensure that measurements will reliably identify samples with activity that
exceeds the specified analytical action level.16
3. Although testing criteria were met, an observed low bias raises questions about the
ruggedness of the gross alpha measurement, especially with regard to analysis of FPWHFO
of different compositions, from different regions, or at different times in the hydraulic
fracturing life cycle. Future work should be done to explore the chemical separation scheme
to improve the individual recoveries for Th, U, and Po, and to test the approach using
FPWHFO of varying compositions such as will be encountered in samples resulting from
routine hydraulic fracturing operations.
4. Additional work is needed to validate the gamma spectrometry measurement for each of the
longer-lived beta-emitting members of the natural decay chains used to quantify gross beta
activity by gamma spectrometry (in addition to 228Ra).
5. Additional work is needed to optimize the data inputs (e.g., libraries), analysis parameters
(i.e., spectrum analysis parameters) and calculations used for the gamma spectrometry
analysis of 226Ra and radionuclides of interest to the gross beta analysis with a goal of
minimizing levels of bias and uncertainty in the measurement.
6. Additional work is needed to explore options for optimizing the preparation of the gamma
spectrometry sample so that the radioelements can be concentrated, thereby improving
measurement sensitivity and reducing the sample counting time.
7. Additional work would be needed if all four of the long-lived radioelements (Ra, U, Th, and
Po) associated with natural decay chain alpha-emissions are to be included in a single
separation and analysis.
8. Build on experience gained in this project in the area of chemical separations in FPWHFO
samples to develop sensitive radionuclide-specific te
210Pb, 228Th, 230Th, 232Th, 238U, 235U, 234U, and 210Po.
samples to develop sensitive radionuclide-specific testing methods including: 226Ra, 228Ra,
The second set of recommendations has to do with temporally dynamic processes of
radioactive decay and decay-progeny relationships is common to all gross alpha and beta
16 See Appendix VI of Radiological Laboratory Sample Analysis Guide for Incidents of National Significance — Radionuclides in Water (EPA
2008) for a discussion of the directed planning process (DQOs/MQOs) and guidance on controlling decision errors with analytical decision
levels (ADLs). In this case, there is an additional concern about low bias in the measurement that must be worked out.
July-2014 44
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
measurements of natural decay chain radioactivity. As discussed above in Section 1, the activity
of samples containing natural decay chain activity, especially 226Ra and 224Ra, changes over time
due to radioactive ingrowth and decay processes. Specifically, the activity physically present in
the sample test source and the interpretation of gross alpha and gross beta results, depend on the
timing of processes preceding the sample collection, the elapsed time between sample collection
and preparation, and the elapsed time between sample preparation and the count. Parameters that
either must be controlled, or at least taken into consideration, to ensure reliable interpretation of
results include:
1. How long did FPWHFO fluids recirculate in the geologic formation prior to discharge to
a settling pond, basin or waterway?
2. Does the activity present in the FPWHFO at the time of the sample measurement reflect
that at key points in the life cycle of the FPWHFO (e.g., time of discharge, point of full
ingrowth, point of transport or release of fluids)?
3. How does the time elapsed between collection, preparation and counting of the samples
impact results. Can the timing of event be managed to minimize variability of results and
provide useful and intercomparable measurements for gross alpha and beta.
4. Does the use of 230Th or 210Po for LSC provide the best method isotope for gross alpha
activity?
5. The analysis of treated FPWHFO may prove to be a completely different matrix where
some radionuclides have been removed (e.g., Ra). Thus future work should also examine
these types of treated matrices.
Addressing these questions as part of the entire process, or identifying individual times for each
of the separate steps until the laboratory receives the samples will ensure that the analytical
results support defensible decisionmaking regarding the handling and disposition of the fluids.
July-2014 45
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 46
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
12.0 Summary and Conclusions
Three parts of The Gross Alpha and Gross Beta Method in FPWHFO were tested using
a matrix based on the composition of a FPWHFO sample received from the EPA to determine
whether they would satisfy method development guidelines outlined in the Method Validation
Guide for Qualifying Methods Used by Radiological Laboratories Participating in Incident
Response Activities (EPA, 2009). Two of the three parts comprise measurements of alpha
emitters in the sample while the third is designed to measure beta emitters.
The MQOs for each of the three parts differed based on the matrix complexity, the
instruments used for analyses, and the nuclear constants associated with the principal
radionuclides used for the development process, and variation associated with preparation of the test
samples. The as-tested MQOs and measured results are presented in Table 3. The final method
with flow diagram used in this method development study is presented in Attachment III.
Each of the three parts of the method validated met all of the acceptance criteria for
method uncertainty as shown in Tables 6A, 6B, and 6C. A summary of the observed levels of
uncertainty at each of three activity levels is summarized in Tables 7A, 7B, and 7C. Detectable
levels of bias were observed across the activity levels for each of the three measurements as
summarized in Table 10. The levels of bias, however, were so large that they compromised the
determinatoin of method uncertainty. The detection capability for each of the three parts was
successfully verified as summarized in 9A, 9B, and 9C.
Although all testing criteria were met as described in this report, the complexity of
the matrix prevented development of a single-measurement method for gross alpha and beta in
FPWFIFO samples that will be simple, economical, and sufficiently rugged in matrices beyond
the one used for the testing. Performing this analysis required a level of effort that was much
different from previous analytical methods in other water matrices for alpha or beta emitters.
Several unique approaches were attempted in order to identify an analytical approach that would
accommodate this particularly challenging matrix. Section 11 provides a brief synopsis of
development activities and Attachment 1 provides additional detail supporting the method
development activities preliminary to final testing.
The final approach for gross alpha requires two measurements. The first measurement
involves gross alpha by liquid scintillation counting following chemical separation to isolate
thorium, uranium and polonium from the matrix. Method testing in the surrogate matrix indicates that
a measureable bias is associated with the technique. Average recovery were 74±11% (k=\) of the
known concentration of 230Th. Recoveries ranged from 57-104%. Although all of the testing
criteria were met, the observed low bias raises possible questions about the ruggedness of the
technique, especially with regard to use of the method for analyzing of FPWHFO of different
compositions, from different regions or different times in the hydraulic fracturing life cycle.
Possible future work should be done to improve the ruggedness of the method and to develop
July-2014 47
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
estimates of uncertainty and decision criteria that would protect against decision errors using this
17
screening technique. See Section 11 for recommendations for possible future work in this area.
The second measurement for alpha activity associated with 226Ra is performed by
gamma spectrometry. The gamma spectrometry measurement is used to simultaneously
determine the activity of longer-lived members of the thorium and uranium decay chains for
calculation of gross beta activity. Although the development process detected bias in the gamma
spectrometry measurements at some levels, the magnitude of the bias is lower than that observed
for the alpha and there is no need for concern about the ruggedness of the non-destructive
measurement technique since there are no variables such chemical separations that will introduce
variable levels of bias into the method. Section 11 suggests the possibility of future work to
improve the sensitivity of the gamma spectrometry measurement.
Due to the physics of the measurement technologies, radionuclide determinations
performed by gamma spectrometry are generally less sensitive and have higher uncertainty that
those performed by the liquid scintillation counting. This complicates the reporting process, the
determination of uncertainty, and prevents calculation of a single meaningful value for gross
alpha detection capability. Section 11 recommends that measurements of gross alpha by LSC and
of 226Ra be reported and interpreted separately and suggests the possibility of future work that
would improve the sensitivity of the gamma spectrometry measurement thereby minimize the
disparity in the sensitivity of the two techniques.
Finally, as mentioned in the introduction in Section 1, all gross alpha and beta measurements are
limited by the complexities of radioactive decay and ingrowth in the uranium and thorium decay
chains which causes the alpha and beta activity physically present in the sample to change over
time. Thus gross alpha and beta measurements are often not (inter-) comparable from
measurement to measurement or laboratory to laboratory. This significantly complicated the
interpretation of gross alpha and beta results. Section 11 recommends that future work explore
the impact of timing on the performance of the method and the interpretation of results, a project
that would benefit gross alpha and beta measurements of natural products in all water matrices.
A screening test should generally be designed to minimize the risk of not detecting analyte when it is present at some action level to a rate
tolerable to the data end user. For this reason, screening tests tend to be structured to deliver results that systematically bias high. Low bias in a
screening measurement may be tolerable as long as the criteria used to decide whether the action level has been exceeded take into account
uncertainty and bias associated with the measurement.
July-2014 48
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
13. References
ASTMD7282 "Standard Practice for Set-up, Calibration, and Quality Control of Instruments
Used for Radioactivity Measurements, " ASTM Book of Standards 11.02, current version,
ASTM International, West Conshohocken, PA.
ASTM D7283 "Standard Test Method for Alpha and Beta Activity in Water By Liquid
Scintillation Counting, " ASTM Book of Standards 11.02, current version, ASTM
International, West Conshohocken, PA.
Multi-Agency Radiological Laboratory Analytical Protocols Manual (MARLAP). 2004. EPA
402-B-1304 04-001A, July. Volume I, Chapters 6, 7, 20, Glossary; Volume II and Volume
III, Appendix G. Available at www.epa.gov/radiation/marlap/.
SM 711OC. "Coprecipitation Method for Gross Alpha Radioactivity in Drinking Water, "
Standard Methods for the Examination of Water and Wastewater, current version, American
Public Health Association, American Water Works Association, Water Environment
Federation, APHA, Washington, DC.
Taylor, JohnK. 1987. Quality Assurance of Chemical Measurements, Lewis Publishers, Chelsea,
MI.
U.S. Environmental Protection Agency (EPA), 1980. Prescribed Procedures for Measurement of
Radioactivity in Drinking Water. Office of Research and Development, Environmental
Monitoring and Support Laboratory, Cincinnati, OH. EPA 600/4-80-032. August.
U.S. Environmental Protection Agency (EPA). 2006. Validation and Peer Review of U.S.
Environmental Protection Agency Radiochemical Methods of Analysis. FEM Document
Number 2006-01, November 8.
U.S. Environmental Protection Agency (EPA). 2008. Radiological Laboratory Sample Analysis
Guide for Incidents of National Significance - Radionuclides in Water. EPA 402-R-07-007,
January. Available at: www.epa.gov/narel.
U.S. Environmental Protection Agency (EPA). 2009. Method Validation Guide for Qualifying
Methods Used by Radiological Laboratories Participating in Incident Response Activities.
Revision 0. Office of Air and Radiation, Washington, DC. EPA 402-R-09-006, June.
Available at: www.epa.gov/narel.
U.S. Environmental Protection Agency (EPA). 2012. Quality Assurance Project Plan Validation
of Rapid Radiochemical Methods For Radionuclides Listed in EPA 's Standardized Analytical
Methods (SAM) For Use During Homeland Security Events. Revision 4. Office of Air and
Radiation, National Analytical Radiation Environmental Laboratory.
July-2014 49
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 50
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Attachment I:
Method Development Trials
Classic Approach to Gross Alpha Beta: Evaporation and Gas Proportional Counting
One ml of the field sample (no dilution or treatment) was evaporated directly onto a
planchet. The dry weight was 260.3 mg and 276.4 mg for replicate samples. These values were
well outside the efficiency calibration curves for either gross beta or gross alpha. Spattering of
the material was significant which would lead to an underestimate of the count rate. Based on
these two significant issues no further analysis by this method was deemed to be fruitful.
Radionuclide: 226Ra
Precipitation of BaSC>4 to coprecipitate radium was tried once. It was evident that too
much precipitate formed (other solids besides the barium/radium precipitate salt out and provide
too heavy a matrix for gross analysis by GPC). The resulting precipitate could not be processed
any further, even when small sample volumes (< 100 mL) were used.
A more selective precipitation using barium chromate was attempted. This process
yielded similar problems to using BaSC>4 with a large mass of other salts precipitating, forming a
black/green 'dusty' precipitate that could not be loaded onto a planchet for gas flow proportional
counting. Using this material for liquid scintillation analysis yielded severe color quench18.
Radium recovery was only about 1%.19
Replicate aliquants of the sample (6*250 mL and 6x40 mL) were analyzed for 222Rn
r)r)^ TM
following a period for ingrowth to assess Ra concentration using a Durridge Rad7 electric
radon detector. This method provides high quality and consistent results, when the foaming of
samples, relative humidity is controlled. Some issues where encountered with foaming of the
FPWHFO samples. Methods could be developed for unsupported 222Rn/ 220Rn as well. The
success of this approach indicates that a radon de-emanation approach to the analysis of 226Ra
(direct analysis of 50 mL of sample without prior chemical concentration) would be capable of a
reasonably sensitive determination of 226Ra with MDCs of-10 pCi/L or less. The method would
OT?
require waiting several days allowing the Rn to ingrow into the sample. This approach was not
pursued further, however, because it would require significant additional development and would
not provide the longer list of beta-emitting analytes possible for gamma spectrometry analysis.
Radionuclides: U, Th, and Po
A classical approach was attempted to obtain a gross separation of uranium and thorium
by Fe(OH)3 precipitation. The iron hydroxide precipitate was dissolved in hydrochloric acid and
loaded onto an Eichrom TRU/TEVA column for separation of the U/Th, and then eluted through
a SR Resin™ for polonium.
8 Quench refers to any process that reduces the production of, or transmission of light from the sample to the photomultiplier detector. If a
correction cannot be applied, quench may result in questionable results.
9 The recovery of 226Ra in this attempt was based on the analysis of the original sample for 226Ra by gamma spectrometry.
July-2014 51
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
In order to assess the success of the gross alpha attempts, the sample test sources were
analyzed by alpha spectrometry. This not only helped to quantify yields but also to determine the
effectiveness of separation of the alpha emitters.
The recovery for the radionuclides was assessed using alpha spectrometry and was as follows:
228n
• Th recovery indeterminate due to Th ingrowth
• Po recovery -29.6 ± 1.7%
U recovery -26-49 ±11%
Po can be selectively electroplated onto a silver disk. However, it must be isolated first
with SR Resin™. Attempts to electrodeposit without a SR Resin clean-up resulted in no recovery
with significant Th/Th-daughter contamination (as determined by alpha spectrometry).
Figure 1 shows the alpha spectrum for the separated polonium (most abundant alpha in the
spectrum at channel number ~ 950 is 4.88 MeV.
Figure 1 - Po-209 Alpha Spectrum
160
Channel
July-2014
52
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Group Separation Using MnCh/FefOH)^ Precipitation
This is a non-specific precipitation technique for radionuclide separation. The
MnO2/Fe(OH)3 precipitate is dissolved in hydrochloric acid followed by TRU/TEVA group
separation of actinides and SR-Resin for Po (U/Th). Run in quadruplicate.
Adding MnO2 substantially increases recovery of U. Note that 228Th cannot be used as a
tracer since it may be present in the sample as an ingrowth product from 228Ra; this provides an
interesting possibility for dating of flowback waters. Natural 230Th and its parents are generally
not detectable in samples. As such, 230Th may be a suitable candidate as a Th tracer for yield
calculations. The U recovery for this mixed precipitate was 71.7±4.8%.
228T
Thorium recovery from the field sample was indeterminate due to 228Th ingrowth20
from ZZ5Ra Analysis for either radium isotope was not attempted. When the surrogate solution
was used with added 230Th tracer, the alpha spectrum in Figure 2 was obtained (the most
abundant alpha at channel number ~ 970 is 4.688 MeV).
230-,
Figure 2 - Alpha Spectrum of Separated Th
4500
3500
3000
<3 2000
1500
1000
A
20 Initially the concentration of 228Th may have been very low. However the age of the sample (at least 3 months) and the high concentration of
226Ra (determined by gamma spectrometry) have allowed significant ingrowth of the 228Th.
July-2014
53
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Empore™ RAD Disks
This method of radionuclide concentration from water is used for several different
radionuclides. The disadvantage is that disk material shows significant uptake of barium.
Additionally, high concentrations of non-radionuclides, such as Ca and Sr, can outcompete Ra
and significantly decrease retention on the disk. The recovery for radium was low, -13.0 ± 1.1%,
even with volumes as low as 50 mL of field sample diluted ten-fold to reduce the concentration
of the non-radionuclide ions. There was too much competition for the exchange sites from non-
radiological elements, and in particular for this sample, from barium. Therefore the use of the
Empore disks was abandoned because of the significant non-radiological interfering ions.
Direct Addition to LS Cocktail
Addition of 0.10 to 10 mL aliquants of the field sample to liquid scintillation cocktail
were used to determine the feasibility of performing direct gross alpha analysis. Even less than
0.1 mL led to salting out in the cocktail yielding a two-phase mixture with a high quench.
Although liquid scintillation measurement remained attractive, this experience indicated that
analysis of the sample would only be possible after clean-up from the FPWHFO matrix (i.e.,
chemical separations).
Gamma Spectrometry
Gamma spectrometry was one of the first analyses to be performed to assess the level
of beta-gamma-emitting radionuclides native to the sample. A 3-L Marinelli beaker was counted
in three different manners: for 17 hours as a straight sample, for 6 hours as a straight sample, and
for 17 hours with agar added to ensure that settling of suspended solids did not occur.
226Ra and 228Ra (or their progeny) were detectable; however, no peaks were observed
for 238U (234Th, 234mPa) or 235U or 228Th (224Ra). The low concentration of 235U in the sample,
allowed for straightforward analysis of 226Ra using the 186 keV peak.
Acid Digests Prior to Separations
Gross alpha analysis was attempted using 10-100 mL of sample following digestion in
nitric acid. However, the amount of residue formed from this digest clogged the TRU Resin
preventing completion of preparation for analysis by alpha spectrometry. In the case of LSC, the
final digestate resulted in unacceptably high amounts of color quench.
Platinum Crucible
Initially acid digestion or fusion techniques were considered as an initial step in the
process. The quantities of dissolved solids in the field sample limit the size of the aliquant
making the fusion method using platinum crucibles unfeasible.
Polonium by Sulfide Precipitation
A method of isolating polonium from the bulk solution used CuS precipitation under
acidic conditions was attempted as PoS will coprecipitate with CuS. Figure 3 shows the liquid
scintillation spectrum from that gross separation. However, this separation technique was not
July-2014 54
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
used in the final method as it caused an unacceptable level of quench in the liquid scintillation
spectrum which would degrade the measurement sensitivity and reliability of the method.
209 r
Figure 3 - Liquid scintillation Spectrum of Po
Channels
While this separation technique was fairly specific for polonium, the low yield and the
quantity of salts (high quench) made this separation technique undesirable.
Thiacalix Resins
Thiacalix resins initially appeared to be a possibility for 226Ra separation since they are
the most selective of Ra resins. However at a Ba: Ra mass ratio of 104, recovery drops to 20%.
The field sample obtained had a mass ratio for Ba: Ra of 109. Furthermore, these resins
experience interference from other divalent cations such as Mg, Ca, Sr, which are present in
similar concentrations to Ba in our sample. Therefore this method of radium separation was not
used.
Flow Chart for the Alpha Beta Methods Attempted
Figure 4 shows the flow chart for the methods attempted for the analysis of gross alpha
and gross beta. In this instance, tracers were used with the surrogate solution to assess the
viability of these separations. The combined iron hydroxide manganese dioxide precipitate,
combined with the elution of the dissolved precipitate through a TRU column provided a good
separation of the alpha emitters uranium, polonium and thorium. Figure 5 shows that with a few
milliliters of bioxalate solution (0.1 molar), the major alpha contributors are eluted together.
July-2014
55
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Figure 6 shows the recovery of three radioelements (U, Th, Po) using the optimized method prior
to method testing.
Figure 4 - Flow Chart for Method Development Using Tracers
Add29mgKMn04,boil,+ 6N
NH4OH,topH7.0
Filtrate
Gross Alpha Analysis
300 ml original sample, filtered and
acidified, 209Po,232U,230Th
Gross Beta Analysis
3- 4 L of original sample,
filtered and acidified
DirectGamma
Spectrometry Measurement
DilutetoSOOmL, boil-60min
Mn02 :Th + U +Po precipitates
Dissolve precipitate in 10 ml 6
N HCI + 2 ml 1.0 N ascorbic
acid.
Discard supernatant and
wash solutions
Discard residue.
Add tracers
226Raand228Ra
\
214Pb,212Pb
210
' n
214Bi,212Bi, „„
Inferential
Determination
Load solution.
Discard waste.
Load HCI solution 1 mL/min,
onto TRU resin
Rinse with 3 5-mL portions
4NHCI.
Discardwaste.
| Alpha Spectrometry [<
Elutewith10-mL0.1 M
(NH4)2Ox
Th,U,
Ce(OH)4(orCeF4)
microprecipitation
Figure 5 - Elution Profile for Traced solution with 0.1 M Ammonium Bioxalate in 4 M HCI
mL of Eluent
Polonium
Thorium
•Uranium
July-2014
56
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Figure 6 - Total Recovery of Tracers (%) with (0.1 M Ammonium Bioxalate in 4 M HCI)
Note: These elutions were performed with 5-1 mL (blue) volumes or 1- 5 mL volume (red).
100
15x1 mL
11x5 mL
Polonium
Uranium
Thorium
These final results of the tracer trials provided evidence that the proposed methods of
chemical separation for the gross alpha (for Th, U, and Po) would be successful.
July-2014
57
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 58
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Attachment II:
Time Lapse
Table 11 - Estimated Elapsed Times for Gross Alpha Analysis
Gross Alpha Analysis Step
Measure aliquant, take "initial" volume
of sample
Precipitation and removal of
supernatant
Redissolution
Loading and elution of sample through
TRU Resin
Perform eluate transfer to LSC vial
Count samples (maximum count time)
Data analysis and review
Total Time
Time for 1 Sample (min.)
2
90
5
30
2
30
20
179
Time for Batch of 20 (min.)
20
200
50
300
20
600*
300
1490
*Assumes only one LSC unit is used
Table 12 - Estimated Elapsed Times for Gross Beta Analysis
Gross Beta Analysis Step
Measure aliquant, take "initial" volume
of sample
Count samples (maximum count time)
Data analysis and review
Totals
Time for 1 sample (min.)
4
720
20
744
Time for Batch of 20 (min.)
40
14400*
300
14740
*Assumes only one gamma-ray spectrometer is used.
July-2014
59
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 60
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Attachment III:
Rapid Radiochemical Method for Gross Alpha and Gross
Beta Activity in Flowback and Produced Waters from
Hydraulic Fracturing Operations (FPWHFO)
1. Scope and Applicability
1.1. This method was developed for flowback and produced water from hydraulic
fracturing (FPWHF) operations. This type of sample will contain several
radionuclides from any of the three natural decay chains. However, based on the
variability in of chemical compounds in the fracturing fluid, the geologic formation
that is being developed and other environmental factors, no assumptions about
radioactive equilibrium should be made when performing the analyses.
1.2. The following alpha and beta activity is determined in this method:
1.2.1. All alpha radioactivity in a hydraulic fracturing sample is associated with
naturally occurring uranium, thorium and radium, short-lived progeny of
these radioelements, and 210Po. The gross alpha screen is accomplished with
two measurements.
1.2.1.1. The first involves a group separation of U, Th, and Po followed by
liquid scintillation counting for gross alpha. This approach
quantifies gross alpha activity associated with uranium isotopes
(238U, 235U , 234U), thorium isotopes, (232Th 230Th , 228Th), short-
lived progeny from 228Th as may have ingrown following the
separations, and 210Po.
1.2.1.2. A non-destructive gamma spectrometry screen is used to determine
alpha-emitting radium activity, primarily associated with 226Ra.
1.2.1.3. The results of the two measurements may be used, either
individually, or the measurements may be mathematically summed
and their uncertainties combined to yield an estimate the combined
gross alpha activity of the sample.
1.2.2. Gross beta screening is accomplished by gamma spectrometry measurement
of radium isotopes (226Ra, 228Ra, 224Ra) and other gamma-emitting natural
chain radionuclides. The beta activity of Ra and its short-lived progeny
(assuming equilibrium ingrowth) are summed and uncertainties combined to
yield a result for the gross beta screen.
1.2.3. The gamma spectrometry measurement also achieves a definitive
determination of the isotopic activity of gamma-emitting tracers used to
assess the efficacy of the hydraulic fracturing process that have not decayed
between collection and the gamma spectrometric measurement.
1.3. The gross alpha and beta screening results are not corrected for decay.
July-2014 61
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
1.4. The method is not intended to demonstrate compliance with the requirements of the
Safe Drinking Water Act (SOWA).
1.5. FPWHFO matrices may contain different chemical mixtures. Method validation may
need to be performed for mixtures that significantly differ in composition from a fluid
matrix that has been previously validated.
1.6. The method is capable of achieving the following MQOs:
1.6.1. Gross alpha screen consisting of the following two components:
1.6.1.1. Gross Alpha by liquid scintillation counting for Th, U, and Po and
decay progeny. A 300 mL sample can achieve a required method
uncertainty of 9.0 pCi/L at less than or equal to an Analytical
Action Level (AAL) of 40 pCi/L. This result is referenced to the
radionuclide used for calibration of the detection system, 230Th, or
210Po.
1.6.1.2. 226Ra alpha by gamma spectrometry: A 3.0 L sample can achieve a
required method uncertainty of 33% (relative), at less than or equal
to an AAL of 165 pCi/L.
1.6.2. 228Ra beta by gamma spectrometry: A 3-L sample can achieve a relative
required method uncertainty of 50%, at less than or equal to an AAL of 60
pCi/L.
1.6.3. Application of the method must be validated by the laboratory using the
protocols provided in Reference 16.1. The sample turnaround time and
throughput may vary based on additional project MQOs, the time for
analysis of the final counting form and initial sample volume.
2. Summary of Method
2.1. See Section 17.4 for a flow chart overview of the process.
2.2. Gross Alpha
2.2.1. The sample is treated with potassium permanganate and ammonia to
coprecipitate radionuclides. The alpha emitting radionuclides are separated
using a TRU resin. The activity of the eluate from the resin is determined
using a liquid scintillation counter.
2.2.2. Ra-226 is determined by direct counting of the 186 keV photopeak using
gamma-ray spectrometry. It may also be used to determine the activity of its
beta-gamma emitting progeny.
2.3. Gross Beta
2.3.1. A 3L aliquant of the sample is counted directly using gamma ray
spectrometry. Gross beta activity is determined by summing the gamma
emitters activity plus those beta only-emitters that are progeny or parents of
gamma emitters detected with less than 50% relative CSU (see Step 2.2.2
and 2.3.2).
July-2014 62
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
2.3.2. Ra-228 is determined by direct counting of the 911 or 338 keV photopeak
using gamma ray spectrometry.
3. Definitions, Abbreviations and acronyms
3.1. Analytical Protocol Specifications (APS). The output of a directed planning process
that contains the project's analytical data needs and requirements in an organized,
concise form.
3.2. Analytical Action Level (AAL). The term "analytical action level" is used to denote
the value of a quantity that will cause the decision maker to choose one of the
alternative actions.
3.3. Flowback and Produced Water from Hydraulic Fracturing (FPWHF) Operations. A
fluid used when gas or oil is extracted from shale rock formations. The produced
water from fracking operations contains up to 200 different chemicals that comprise
the solution, plus any minerals that are extracted from the shale formation during the
process. Each type of FPWHF may be unique in its chemical and physical properties.
3.4. Multi-Agency Radiological Analytical Laboratory Protocols Manual (MARLAP) (see
Reference 16.2).
3.5. Measurement Quality Objective (MQO). The analytical data requirements of the data
quality objectives that are project- or program-specific and can be quantitative or
qualitative. These analytical data requirements serve as measurement performance
criteria or objectives of the analytical process.
3.6. Required Method Uncertainty (WMR). The required method uncertainty is a target value
for the individual measurement uncertainties and is an estimate of uncertainty (of
measurement) before the sample is actually measured. The required method
uncertainty as an absolute value is applicable at or below an AAL.
3.7. Relative Required Method Uncertainty (^MR). The relative required method
uncertainty is the WMR divided by the AAL and is typically expressed as a percentage.
It is applicable above the action level.
3.8. Sample Test Source (STS). This is the final form of the sample that is used for
nuclear counting. This form is usually specific for the nuclear counting technique in
the method, such as a solid deposited on a filter for alpha spectrometry analysis.
4. Interferences
4.1. Radiological:
4.1.1. The elapsed time between sampling and arrival at the laboratory and the
start of the analytical process to the counting time should be recorded for
each sample. These time frames can have significant effects on the final
activity concentrations determined. This is due to the various radiochemical
equilibria that exist for the naturally occurring radionuclides. It will be
important for the laboratory to know at which point in time (e.g., sample
time, arrival at laboratory, time of discharge) the client desires to have the
activity concentrations corrected to so that they can meet the needs of the
project.
July-2014 63
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
4.2. Non-radiological:
4.2.1. FPWHFO samples will be high in dissolved solids content as well as having
a high concentration of other ionic substances not usually encountered in
surface or drinking water samples (e.g., barium, strontium, and silica).
4.2.2. Although the method has been designed to accommodate high levels of
solids, including Group II elements, concentrations of non-radioactive
barium in the grams/L range may require a decreased sample size (i.e.,
smaller) to be selected.
4.2.3. Similarly, high concentrations of non-radioactive calcium, magnesium or
strontium in the sample may require the use of a decreased sample size.
5. Safety
5.1. General
5.1.1. Refer to your safety manual for concerns of contamination control, personal
exposure monitoring and radiation dose monitoring.
5.1.2. Refer to the laboratory chemical hygiene plan for general chemical safety
rules.
5.2. Procedure-Specific Non-Radiological Hazards:
5.2.1. Solutions of potassium permanganate can rapidly oxidize organic materials
and generate significant heat. Do not mix large quantities of permanganate
solution with solutions of organic solvents as the potential for conflagration
exists.
6. Equipment and Supplies
6.1. Toploader balance with a 0.1-g readability.
6.2. Beakers, Pyrex®: 250, 400 mL.
6.3. Centrifuge capable of holding 300 mL vessels (optional).
6.4. Hot plate, or other suitable device for heating ammoniated sample volume.
6.5. Glass stirring rods.
6.6. Graduated cylinders: 500, 1000, 4000 mL capacities.
6.7. Scintillation vials: 22 mL glass.
6.8. pH paper; range 5.0-9.0 pH units.
7. Reagents and Standards
NOTES: All reagents are American Chemical Society (ACS) reagent grade or equivalent unless otherwise
specified.
Unless otherwise indicated, all references to water should be understood to mean Type I Reagent water (ASTM
D1193, Reference 16.3).
7.1. Potassium Permanganate, (KMnO4)
7.2. Ammonium Hydroxide (NHtOH): Concentrated (15 M)
7.2.1. Ammonium Hydroxide (NELiOH): (6 M); dilute 40 mL of concentrated
ammonium hydroxide to 100 mL with water.
July-2014 64
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
7.3. Hydrochloric Acid (HC1): Concentrated (12 M)
7.3.1. Hydrochloric Acid (HC1): (6 M); dilute 50 mL of concentrated hydrochloric
to 100 mL with water.
7.3.2. Hydrochloric Acid (HC1): (4 M); dilute 33 mL of concentrated hydrochloric
to 100 mL with water.
7.4. Ascorbic Acid, crystals
7.4.1. Ascorbic Acid (1 M); dissolve 176 g of dried crystals in 100 mL of water
7.5. Ammonium Bioxalate, crystals
7.5.1. Ammonium Bioxalate (0.1 M); dissolve 12.4 g of dried crystals in 100 mL
of demineralized water
7.6. UltimaGold AB™ scintillation cocktail, (available from PerkinElmer Inc.) or
equivalent.
7.7. TRU Resin™ (Eichrom Technologies, LLC.); 100-150 micron
7.8. Calibration sources, Traceable to a national standards body such as the National
Institute of Standards and Technology (NIST) in the U.S.
7.8.1. For liquid scintillation: 230Th or 210Po in a counting configuration that
matches that of the sample test source as closely as possible. If sample
quench varies enough that the efficiency varies by more than 10%, prepare
standards with varying levels of quench that span the range of efficiencies
encountered during the analysis of samples.
7.8.2. For gamma spectrometry: Either 228Ra and 226Ra or a mixed gamma source
with energies spanning the entire range of energies to be used during
analysis of samples. The geometry of the calibration standard shall match
that of the sample test source as closely as possible (e.g., Marinelli beaker
3000-mL volume, density, average "z", etc.). See Section 11.1.3 for a
description of preparation of the sample test source.
8. Sample Collection, Preservation and Storage
8.1. None recommended
9. Quality Control
9.1. Batch quality control results shall be evaluated and meet applicable Analytical Project
Specifications (APS) prior to release of unqualified data. In the absence of project-
defined APS or a project-specific quality assurance project plan (QAPP), the quality
control sample acceptance criteria defined in the laboratory quality manual and
procedures shall be used to determine acceptable performance for this method.
9.2. A laboratory control sample (LCS) shall be run with each batch of samples. The
concentration of the LCS should be at or near the action level or a level of interest for
the project.
9.3. One method blank shall be run with each batch of samples. The laboratory blank
should consist of demineralized water.
July-2014 65
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
9.3.1. If possible a matrix blank that includes a surrogate solution made from
reagent grade chemicals should be analyzed.
9.4. One laboratory duplicate shall be run with each batch of samples. The laboratory
duplicate is prepared by removing an aliquant from the original sample container after
mixing/stirring.
9.5. A matrix spike sample may be included as a batch quality control sample if there is
concern that matrix interferences, such as the presence of elemental barium in the
sample, may compromise chemical yield measurements, or overall data quality.
9.5.1. Matrix spike added activity may be difficult to estimate unless there is prior
historical data to identify an existing concentration of some of the
radionuclides.
10. Calibration and Standardization
10.1. The liquid scintillation counter is set-up, calibrated, verified, quality controls
performed according to manufacturer's specifications for alpha^eta discrimination,
and as specified in ASTM D7282 Sections 9, 12, 13, 19, and 25.
10.1.1. Set up the scintillation counter to discriminate between alpha and beta
pulses according to manufacturer's specifications. Samples are counted in
an energy window that encompasses the alpha peak observed in the middle
section of the alpha spectrum.
10.1.2. Calibrate the LSC for alpha efficiency in the same energy window of the
alpha spectrum used to count samples.
10.1.3. Use standard 230Th or 210Po incorporated in the same counting configuration
used for samples (i.e., vial type, cocktail, chemical make-up of the aqueous
matrix, sample to cocktail ratio, and quench levels should match that of the
samples as closely as possible). See Sections 11.1.2.9 and 11.1.2.10 for a
description of preparation of the sample test source.
10.1.4. The subtraction background is determined in the same energy window of the
alpha spectrum used to count samples using a background sample that
matches that of samples being analyzed.
10.2. The gamma spectrometer is set-up, calibrated, verified, quality controls performed
according to manufacturer's specifications for alpha/beta discrimination, and as
specified in ASTM D7282 Sections 9, 12, 13, 17, and 23.
10.2.1. Calibrate the gamma spectrometer across the entire energy range to be used
during analysis of samples.
10.2.2. Count the calibration standard (Section 7.1.2) in the same juxtaposition to
the detector as the samples.
11. Procedure
11.1. Water Sample Preparation
11.1.1. A well-mixed portion of the sample is taken for each of the separate analysis
flow paths:
NOTE: The sample is not filtered.
July-2014 66
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
11.1.1.1. 300 mL for gross alpha analysis
11.1.1.2. 3.0 L for gamma spectrometry analysis
11.1.2. Gross Alpha Analysis
11.1.2.1. Add 29 mg of KMnC>4 and bring the solution to a boil.
11.1.2.2. Add enough concentrated ammonium hydroxide to bring the
solution into a pH range of 5.0 to 9.0.
11.1.2.3. Dilute the solution to 500 mL with demineralized water and bring
to a boil for approximately 60 minutes.
11.1.2.4. When the solution has cooled, and settled, decant off and discard
the supernatant solution
11.1.2.5. To the residue after decantation add approximately 10 mL of 6 M
HC1 and 2 mL of 1.0 M ascorbic acid.
NOTE: this is the load solution for the solid phase extraction column and
should be kept to a minimum. Fresh ascorbic acid should be used to ensure
all the Fe+3 have been reduced to Fe+2.
11.1.2.6. Load the solution in 11.1.2.5 onto a TRU Resin column at the
rate of 1 mL/min
11.1.2.7. Di scard the eluate from the 1 oad soluti on
11.1.2.8. Rinse the TRU Resin column with three 5-mL portions of 4M
HC1 at a flow rate of 1 mL/min. Discard the wash solutions.
11.1.2.9. Elute the radionuclides using 10 mL of 0.1 M ammonium
bioxalate solution
11.1.2.10. Add the entire eluate to 15 mL of UltimaGold AB cocktail in a
22-mL scintillation vial. This is the sample test source (STS) for
the gross alpha measurement.
11.1.2.10.1. Liquid Scintillation Counting
After dark adaptation, the sample is counted for
approximately 30 minutes or long enough to meet
the MQOs noted in Step 1.6.
11.1.3. Determination of Gross Beta and Radium 226 activity by gamma
spectrometry
11.1.3.1. If necessary to prevent settlement of solids during the counting
time either:
11.1.3.1.1. Acidify the solution with 16 M nitric acid until
solids dissolve, or
11.1.3.1.2. Add sufficient agar to suspend any solid materials.
11.1.3.2. Count the sample on a gamma ray spectrometer for
approximately 6-12 hours or until the MQO's noted in Step 1.6
can be achieved.
NOTE: The exact time to count the sample will depend upon the exact sample size taken and the
efficiency of the detectors used for the analysis. In either case the counting time should be
adjusted so that the MQOs noted in Step 1.6 can be achieved.
11.1.3.3. Using the information in Figure 17.1:
July-2014 67
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
11.1.3.3.1. Determine which radionuclides that are to be
inferentially determined will be in secular
equilibrium based on the time between sampling
and analysis.
11.1.3.3.2. Sum the activities of the radionuclides that have a
CSU of less than 50% and those inferred assuming
secular equilibrium.
11.1.3.3.3. Review the gamma-ray spectrum report for any
gamma rays that are unidentified and ascribe
identities to them.
Determine if these activities need to be added to
the sum in Step 11.1.3.3.2.
11.1.3.3.4. The gross beta activity is estimated by the sum of
Steps 11.1.3.3.2 and 11.1.3.3.3 (See Section 12)
12. Data Analysis and Calculations
12.1. Results are reported as gross alpha and gross beta activity as follows:
12.1.1. Gross Alpha Activity
Gross Alpha ActivityLSC, ^ = (^ffi (D
Where:
cpms is the sample gross count rate, in counts per minute
cpnib is the instrument background count rate in counts per minute
e is the detector efficiency based on the instrument quench curve and a
230 Th or 210Po calibration source
V is the sample volume used in liters
Gross Alpha Activitygamma spec, ^ = Zi (/a-y)n (2)
Where:
Ia-Y are the individual radium alpha emitter activities inferred from
gamma-emitting counterparts (usually only 226Ra)
12.1.2. Gross Beta Activity
Gross beta activitygammaspec,^- = l{(My)y + Zi(7y)fc (3)
Where:
My is the activity of each measured gamma-ray emitting radionuclides
Iy are the activities of the non-gamma, beta- emitting radionuclides that
are inferred.
12.1.3. Measurement uncertainty
NOTE: The uncertainty for the indirectly measured radionuclides should be made equal to the
uncertainty from their gamma-emitting parent or progeny (for either the alpha-only or beta
only emitters). Since this is a gross analysis this approximation should be satisfactory.
12.1.3.1. Gross alpha
July-2014 68
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
(cpms j
Gross Alpha Uncertainty LSC , Is = 022x^x10
Gross Alpha Uncertainty 'gamma spec, Is = /£ J ufn_y (4)
Where:
Ujn_ is the uncertainty of an individual gamma emitter that is used to determine the
alpha activity.
12.1.3.2. Gross Beta
Gross Beta Uncertaintygammaspec, Is = ^{k ]u2Mvj + u2/y J (5)
Where:
UM = 1 a from y — spectrum report for directly measured radionuclides (6)
Yj
and
u, = 1 a for indirectly measured radionuclides (7)
13. Method Performance
13.1. Results of method validation performance are to be archived and available for
reporting purposes.
13.2. Expected turnaround times are:
13.12.1 Gross alpha ~3 hours for an individual sample and -25 hours per batch
(see Section 17.2).
13.12.2 Gross beta -13 hours for an individual sample and -10 days per batch (see
Section 17.3).
14. Pollution Prevention
14.1. The use of potassium permanganate to produce MnO2 and a TRU® resin reduces the
amount of solvents that would otherwise be needed to co-precipitate and purify the
final sample test source.
14.2. Ultima Gold AB is a non-hazardous waste after it is used.
15. Waste Management
15.1. Nitric acid and hydrochloric acid wastes should be neutralized before disposal and
then disposed of in accordance with local ordinances.
16. References
16.1. U. S. Environmental Protection Agency (EPA). 2009. Method Validation Guide for
Radiological Laboratories Participating in Incident Response Activities. Revision 0.
July-2014 69
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
Office of Air and Radiation, Washington, DC. EPA 402-R-09-006, June. Available
at: www.epa.gov/narel/incident guides.html.
16.2. Multi-Agency Radiological Laboratory Analytical Protocols Manual (MARL AP).
2004. EPA 402-B-1304 04-001A, July. Volume I, Chapters 6, 7, 20, Glossary;
Volume II and Volume III, Appendix G. Available at: www.epa.gov/radiation/
marlap/index.html.
16.3. ASTM Dl 193, "Standard Specification for Reagent Water," ASTM Book of
Standards 11.02, current version, ASTM International, West Conshohocken, PA.
16.4. NNDC, Brookhaven National Laboratory, sonzogni@bnl.gov.
16.5. Data Source: National Nuclear Data Center, Brookhaven National Laboratory, based
on ENSDF and the Nuclear Wallet Cards (2014); http://www.nndc.bnl.gov/chart/.
16.6. ASTM D7282 "Standard Practice for Set-up, Calibration, and Quality Control of
Instruments Used for Radioactivity Measurements," ASTM Book of Standards 11.02,
current version, ASTM International, West Conshohocken, PA.
17. Tables, Diagrams and Flow Charts
17.1. Table of Direct and Inferential Determinations for Gamma Ray Emitters
Radionuclide
208Tj[l]
210pb
212Pb
212Bi
214Fb
214Bi
226Ra[2]
228Ac
234Th
Gamma Ray
Energy, keV
583
46.5
238
727
295
352
609
1120
186
338
911
93
63
Abundance,%
85.0
4.25
43.6
6.67
18.4
35.6
45.5
14.9
3.64
11.3
25.8
4.2
3.7
Half-Life,
Days
2.1X10'3
8.1xlO+3
4.4XKT1
4.2xKT2
1.9xlO'2
1.9xlO'2
1.4xlO'2
1.4X10'2
5.8xlO+5
2.6XKT1
2.6X10"1
2.4xlO+1
2.4xlO+1
Time from
Sampling to
Counting,
Days
<0.1
365
30
2
2
18
18
18
18
—
1.5
1.5
<0.1
<0.1
Inferred
Activity
212Bi
210Po
2iaBi
224Ra
224Ra
226Ra
226Ra
226Ra
226Ra
—
228Ra
228Ra
234mPa
234mpa
[1] Tl-208 also has a gamma at 2615 keV that is 99% abundant.
[2] Note that this gamma ray is interfered with by the gamma ray from 235U at 185. The 226Ra can be determined directly if no 235U is present or
if the 235U peak issued to subtract out the interference at 186 keV. See Reference 16.5.
July-2014
70
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
17.2. Estimated Elapsed Times for Gross Alpha Analysis
Gross Alpha Analysis Step
Measure aliquant, take "initial"
volume of sample
Precipitation and removal of
supernatant
Redissolution
Loading and Elution of sample
through TRU resin columns
Perform eluate transfer to LSC vial
Count samples (maximum count
time)
Data analysis and review
Total Time
Time for 1 Sample, Minutes
2
90
5
30
2
30
20
179
Time for Batch of 20, Minutes
20
200
50
300
20
600*
300
1,490
* Assumes only one LSC unit is used.
17.3. Estimated Elapsed Times for Gross Beta Analysis
Gross Beta Analysis Step
Measure aliquant, take "initial"
volume of sample
Count samples (maximum count
time)
Data analysis and review
Totals
Time for 1 Sample, minutes
4
720
20
744
Time for Batch of 20, minutes
40
14,400*
300
14,740
* Assumes only one gamma ray spectrometer is used.
July-2014
71
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
17.4. Sample Processing for FPWHFO
Filtrate
r
Add 29 mg KMn04, boil + 6 M
NH4OH,topH7.0
Dilute to 500 ml, boil - 60 min
Mn02 :Th + U +Po precipitates
Dissolve precipitate in 10 ml
6 MHCI + 2mL1.0 M
ascorbic acid.
Gross Alpha Analysis
300 ml original Sample
Gross Beta Analysis
3 - 4 L original sample
Direct Gamma
spectrometry Measurement
A.
Discard supernatant and
wash solutions
Inferential
Determination
210 228
235
'34 226
U)
^
imu
f
Ratal
™mPa, 228Ra, ™E
( 210Po, 226Ra - f
i, 212Bi,
Ipha)
\r
^m«* ilnh
Rinse with 3 5-mL portions
4MHCI,
Discard waste.
Th, U, Po'2, Po"1 |
Add combined eluate to 15 ml
Ultima Gold AB.
July-2014
72
-------
Development of Rapid Radiochemical Method for Gross Alpha and Gross Beta Activity in FPWHFO
July-2014 73
-------
&EPA
United States
Environmental Protection
Agency
Office of Research
and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-14/007
July 2014
www.epa.gov
Please make all necessary changes on the below label, detach
or copy and return to the address in the upper left hand corner.
If you do not wish to receive these reports CHECK HERE Q;
detach, or copy this cover, and return to the address in the
upper left hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA PERMIT No. G-35
Recy cl ed /Recy cl able
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |