SEPA
United States
Environmental Protection
Agency
Offica of
Radiation Programs
Washington DC 20460
EPA 520/5-82-012
1982
Radiation
The Eastern Environmental
Radiation Facility's Participation
in Interlaboratory and Intralaboratory
Comparisons of Environmental
Sample Analyses: 1979 and 1980
-------
EPA 520/5-82-012
The Eastern Environmental Radiation
Facility's Participation in Interlaboratory
and Intralaboratory Comparisons of Environmental
Sample Analyses: 1979 and 1980
__
R. Blanchard, J. Broadway, and J. Moore
January 1982
Office of Radiation Programs
Eastern Environmental Radiation Facility
U.S. Environmental Protection Agency
P.O. Box 3009
Montgomery, Alabama 36193
-------
TABLE OF CONTENTS
LIST OF ILLUSTRATIONS iv
PREFACE v
ACKNOWLEDGEMENTS . v1
I. INTRODUCTION 1
II. INTERLABORATORY COMPARISON PROGRAMS .... 3
A. EMSL-LV Intercomparison Program 3
B. WHO and IAEA Intercomparison Programs 8
III. INTRALABORATORY COMPARISON PROGRAMS 10
A. Replicate Analyses 10
B. Blind Analyses 14
IV. SUMMARY 19
REFERENCES 24
APPENDIXES 25
A: EMSL-LV Water Cross-Check Samples 25
B: EMSL-LV Milk Cross-Check Samples 35
C: EMSL-LV Food Cross-Check Samples 39
D: WHO Intercomparison Sample Results 42
E: IAEA Intercomparison Sample Results 45
n
-------
ILLUSTRATIONS
FIGURES
1. Cumulative Probability vs. the Percent Sample
Measurement Error 20
2. The Cumulative Probability vs. the Sample Coefficient
of variation 23
TABLES
1. Summary of EMSL-LV Cross-Check Samples 7
2. Minimum Detectable Concentrations for Routine
Cross-Check Sample Analyses 9
3. Values Used for Calculating the Mean Range
Control Limits 12
4. Analytical Precision for Various Analyses 12
5. Summary of Replicate Analyses Results 13
6. Results of Intralaboratory Blind Analyses
of Water Samples 15
-------
PREFACE
The Eastern Environmental Radiation Facility (EERF) helps solve
problems defined by the Office of Radiation Programs. The facility
provides analytical capability for evaluating and assessing radiation
sources through environmental studies and surveillance and analysis.
The EERF provides special analytical support for Environmental
Protection Agency Regional Offices and other federal government
agencies as requested as well as technical assistance to the
radiological health programs of state and local health departments.
Readers of this report are encouraged to comment freely. Comments
may be directed to the EERF directly or to the Office of Radiation
Programs in Washington, D.C.
)/
Director
Eastern Environmental Radiation Facility
-------
ACKNOWLEDGEMENTS
The authors would like to offer their special appreciation to the
staffs of the Monitoring and Analytical Services Branch and the
Counting Section of the Technical Support Branch of the Eastern
Environmental Radiation Facility. The analytical work performed by
these staff members is fundamental to producing a document of this
nature.
-------
I. INTRODUCTION
Since 1964 the Eastern Environmental Radiation Facility (EERF),
Office of Radiation Programs (ORP), U.S. Environmental Protection
Agency, has compared its results of analyses of radionuclides in
environmental and biological samples with those of other agencies.
Such intercomparisons are sponsored by several agencies and the
results are routinely published by the quality assurance reference
center of the respective sponsoring agency. All samples in these
intercomparisons are treated anonymously, identified by a code known
only by the originating laboratory and appropriate reference center.
The EERF is committed to making its results of inter!aboratory
comparison studies a matter of public record. This provides a basis
for judging the validity of routinely reported results. The
Environmental Measurements Laboratory (EML), formerly the Health and
Safety Laboratory (HASL), was first to publish the results of their
participation in an intercomparison program (We77). In 1979, the EERF
followed by publishing all results of intercomparison programs prior
to 1979 that were sponsored by the EPA National Quality Assurance
Program at the Environmental Monitoring and Support Laboratory - Las
Vegas (EMSL-LV), the World Health Organization (WHO), and the
International Atomic Energy Agency (IAEA) (B179).
This report presents our results for 1979 and 1980 in
intercomparison studies sponsored by EMSL-LV, WHO, and the IAEA as
-------
well as our intralaboratory analyses results. The latter results
include replicate, blind, and spiked sample analyses. We plan to
publish similar reports on a biannual schedule.
-------
II. INTERLABORATORY COMPARISON PROGRAMS
A. EMSL-LV Intercomparison Program
The most comprehensive intercompari son program, relative to
numbers of samples, sample types and radionuclides, has been conducted
with the EMSL-LV Reference Center. Routine sample types and the
radionuclides that were included in each are shown below:
Sample Type
Radionuclides Included in Sample
Water
51Cr(6), 60Co(10), 65Zn(7), 89Sr(10),
90Sr(10), 106Ru(8),
134
Milk
Food
Cs(10),
137Cs(10), 226Ra(13), 228Ra(13), 234>238U(5)
00 Q
"yPu(6), Gross Alpha (15), Gross Beta (15)
137
89Sr(8), 90Sr(8),
Cs(8),
140
Ba(8), K(8)
Sr(5), 90Sr(5),
90C ,c, 131TICi 137r
•1(5), i0'Cs(5),
140
Ba(5), K(5)
Note.--The number of measurements of each type analysis is given
in parentheses.
Each analysis was made in triplicate. The results of these
analyses with the known concentrations are presented in Appendixes A
through C. The values reported by EERF were compared in two different
ways to the known values supplied by the reference center. For the
first comparison, the average of the triplicate analyses (column 2)
-------
was divided by the known concentration (column 3) resulting in a ratio
(R) shown in the fourth column. The magnitude by which R differs from
unity is a measure of the agreement between measured and known
concentrations. However, for low concentration measurements, large
values of R are often acceptable, since the uncertainty of measurement
is also frequently large at low concentrations.
A second method of comparison is a qualitative statistic known as
the coefficient of variation (CV) (Mo51). The coefficient of
variation is particularly useful in comparing dispersion of two or
more sets of positive variates measured in the same or different
units. For a sample with a true mean (u), a sample mean(X), and a
sample standard deviation (S ), the coefficient of variation (CV) is
rt
usually defined as
sx
CV= — . 100%, (1)
where
s_, .., J211/2 (2)
T /Y. \211/2
=
-------
The degrees of agreement between the EERF results and the
concentration supplied by the reference center, as indicated by the
ratio R, are summarized in Table 1. The radionuclide is given in the
first column with the number of samples shown in parentheses. Results
of the analyses of the different sample media—water, milk, and
food—are combined. The ranges in the values of R are listed in the
second column followed by the percent of the analyses differing by +_
10%, _+ 20%, and by more than 20% of the known reference center value.
Agreement within 10% is considered very good; those within 20% of the
known value are considered satisfactory.
Approximately 80% of all EMSL-LY cross-check results were within
20% of the known value. However, 29 (64%) of the 39 EMSL-LV
cross-check samples on which our results differed from the known value
by more than 20% are judged satisfactory because of the low sample
concentrations, the uncertainty in the measurements, and the small
absolute difference between the results. These results have asterisks
by them in Appendixes A through C. Thus, about 95% of the EMSL-LV
cross-check analyses are acceptable by this test.
Similarly, using the coefficient of variation as a measure of
consistency with the EMSL-LV cross-check samples (see values in the
fifth column of Appendixes A-C), 42% had coefficients of variation
within 10% of the known value; 74% were within 20%; and only 2% had
coefficients of variation greater than 50%. A more detailed
examination of those samples having high coefficients of variation
showed that serious analytical difficulty was not indicated in most
cases. For example, the Co water cross-check of April 1980 (see
-------
Appendix A) contained 6 pCi/1 but was measured to contain 15 pCi/1,
resulting in a coefficient of variation of 150%. Since the error in
the analytical determination of this sample is about 8 to 10 pCi/1 at
the 95% confidence level, the measured value of 15 pCi/1 is not
significantly different from the known value of 6 pCi/1. In contrast,
the results of the IAEA water sample listed in Appendix E reveal the
144
existence of a problem in measuring Ce. The mean concentration
determined by the EERF was 112 pCi/1 and the known value was only
5.6 pCi/1, resulting in a coefficient of variation of 1900%. In
summary, when using the coefficient of variation to determine whether a
specific analytical problem exists, one must consider both the
magnitude of the measurement and the associated analytical errors.
The mean coefficients of variation for each nuclide analyzed in
the EMSL-LV water, milk, and food samples are listed in the last column
of Table 1. This provides a quick method for comparing the relative
accuracy of the analytical procedures for different radionuclides. For
example, the mean coefficient of variation of the H analyses, 5.8%,
indicates a relatively accurate procedure compared to that for Cs,
for which the mean coefficient of variation was 40%, indicating a
relatively inaccurate procedure. However, this comparison does not
consider the concentrations involved.
?26
The mean coefficient of variation for Ra of 9.4% compares
favorably with the 12% value previously determined by Williams (Wi81).
Also, the mean coefficients of variation of 16.8% and 14.4% for gross
alpha and beta measurements, respectively, compare favorably with the
20-40% range reported by Jarvis (Ja76).
-------
TABLE 1
Summary of EMSL-LV Cross-Check Samples
Nuclide
H-3U1)
Cr-51(3)
Co-60(7)
Zn-65(4)
Sr-89(18)
Sr-90(18)
Ru-106(3)
1-131(17)
Cs-134(8)
Cs-137(19)
Ra-226(13)
Ra-228(ll)
U-238/234(5)
Pu-239(6)
Alpha (15)
Beta (15)
K (12)
Range in R
0.90-1.13
0.98-1.21
0.95-2.50
0.90-1.29
0.60-1.17
0.93-2.78
0.90-1.05
0.69-1.48
0.85-3.08
0.81-1.31
0.98-1.18
0.57-1.50
1.00-1.15
0.67-0.93
0.58-1.53
0.51-1.10
0.82-1.01
Percent of values differing
from R =1.00 by: Coefficient of variation, CV (%)
<10%
64
67
57
75
28
33
100
53
38
79
69
9
60
33
40
53
83
<20%
100
67
71
75
67
78
100
71
88
95
100
27
100
67
73
73
100
>20%
0
33
29
25
33
22
0
29
12
5
0
73
0
33
27
27
0
Range
1.5-12.7
12.0-35.5
5.1-150.6
4.4-31.6
7.3-40.0
3.7-178
10.2-25.6
1.2-47.5
4.7-208
1.7-31.6
3.5-17.6
3.2-49.8
0 -15.4
7 -33.4
0 -54.2
0 -49
1.0-18.2
Mean
5.8
22.9
35.9
16.7
18.7
24.5
18.7
14.9
40.3
12.2
9.4
30.3
7.9
16.8
16.8
14.4
7.9
-------
Table 2 lists the minimum detectable concentrations for the
radionuclides commonly measured by the procedures used at the EERF
(1182). These concentrations correspond roughly to one half the
detection limit as defined by Currie (Cu68), in that they are a priori
concentrations which should lead to detection at a confidence level of
a=0.05 and 8=0.05. Although results presented in this report are not
sufficient to corroborate the tabulated minimum detectable
concentrations, general consistency with those values is demonstrated.
Of the 41 analyses of EMSL-LV cross-check samples which were reported
by EERF as being below their respective detection limits, 40 were
correctly reported. The analysis of only one food sample containing
89
8 pCi/kg of Sr was incorrectly reported as having less than
5 pCi/kg.
B. WHO and IAEA Intercomparison Programs
While participating in the EMSL-LV intercomparison program, the
EERF simutaneously participated in similar programs conducted by the
World Health Organization (WHO) and the International Atomic Energy
Agency (IAEA), although on a more limited basis. The intercomparison
results for samples provided by these two reference centers are given
in Appendixes D and E. The two programs consisted of 11 samples on
which a total of 41 specific radionuclide analyses were made.
Agreement between our results and the concentrations supplied by
the two international reference centers was similar to that attained in
the EMSL-LV program. Seventy-one percent of the analyses agreed within
10% of the reference center value, while 84% of the analyses agreed
-------
TABLE 2
Minimum Detectable Concentrations for Routine
Cross-Check Sample Analyses
Water Samples (pCi/1)
3H -
51
60
65
Cr -
Co -
In -
89
Sr -
200
30
10
20
5
90
106
134
137
226
Sr - 1
Ru - 30
Cs - 10
Cs - 10
228]
239r
Ra - 1
7Pu - 0.015
Gross alpha - 2
Gross beta - 1
Ra - 0.1
Milk (pCi/1) and Food (pCi/kg) Samples
89
90
Sr - 5
Sr - 1
I - 10
137
140
Cs - 10
Ba - 10
Soil Samples (pCi/g)
60
106
Co - 0.010
Ru - 0.030
134
Cs - 0.010
137
239
Cs - 0.010
Pu - 0.015
-------
within 20%. Of the 6 analyses that differed by more than 20% of the
known concentrations, 3 are judged acceptable because of the small
90
concentrations present (K in WHO freshwater fish, Sr in WHO marine
90
sediment, and Sr in IAEA water). The concentration we reported for
106Ru/106Rh -n the WHQ seawater sanlpie was for 106Rh only and
should have been multiplied by two to include Ru, resulting in a
total concentration of 783 pCi/1, which is 92% of the IRC value. Thus,
overall agreement is quite good.
Using the coefficient of variation as a measure of consistency with
the WHO and IAEA cross-check samples (see values in the fifth column of
Appendixes D and E), 68% were within 10% of the known value, 84% were
within 20%, and only 4 analyses had coefficients of variation greater
than 50%. Discarding the coefficient of variation for the Ce
analyses of the IAEA water sample as an outlier (we are presently
investigating the problem associated with the spectral analysis of
144
Ce), the mean coefficient of variation for all 38 analyses is 18%.
III. INTRALABORATORY COMPARISON PROGRAMS
A. Replicate Analyses
Replicate analyses are performed on every tenth sample analyzed at
the EERF and on each inter!aboratory cross-check sample. Usually,
there were two or three replicate analyses on each cross-check sample,
but some samples were analyzed as many as six times. To analyze the
10
-------
precision of these analyses, we calculated the mean range (R) between
duplicate analyses from the standard deviation of the analyses. The
mean range (Ro64; Ka77) is defined by the equation
I = d2o, (5)
where d2 is a function of the number of replicates involved (see
Table 3) and a is the standard deviation (see Table 4). The control
limits are computed as follows:
R" + 3o = D^ = D4d2a (6)
where aR is the standard deviation of the range and D^ is a
function of the number of replicates involved (see Table 3). Therefore,
a = "R~(D4-l)/3 . (7)
The range limits were computed for each type analysis by the
procedure described above. The observed ranges between replicates were
classified as £("R + OR), <_{1T+ 2aR), (£ + 3aR).
The number of replicate analyses that fall into each category is a
measure of the laboratory's performance. Table 5 summarizes the
results of the replicate analyses performed at the EERF. The
theoretical distribution is listed in the last line of the table.
Overall, the distribution of the precision attained approaches the
expected theoretical distribution. Analyses showing less precision
were for 144Ce and 65Zn.
To determine the precision of the analyses for naturally occurring
radionuclides in soil, 10 identical samples of harbor silt were submitted
11
-------
TABLE 3
Values Used for Calculating the Mean Range Control Limits
No.
of
Observations
2
3
4
5
6
Central line
factor (d2)
1.128
1.693
2.059
2.326
2.534
Control limit
factor (04)
3.267
2.575
2.282
2.115
2.004
Source: Rosenstein and Goldin, 1964.
TABLE 4
Analytical Precision for Various Analyses
Nuclide
89Sr> 131I§ 137CSj
9°Sr
K
3H
226Ra
239Pu
Gross Alpha
Gross Beta
Concentration
(pCi/1 or kg)
140Ba 5-100
>100
2-30
>30
>0.1*
<4000
>4000
iO.lpCi/1
>0.1**
<20
>20
<100
>100
Standard Deviation, a
(single determination)
5 pCi/1
5%
1.5 pCi/1
5%
5%
5%
10%
15%
10%
5 pCi/1
25%
5 pCi/1
5%
* Units are g/1 or Kg.
** Units are pCi/l> g, or sample.
12
-------
"blind" to the analytical laboratory. The results of the 10 replicate
analyses are listed below (units are pCi/g):
228Th = 1.10+0.13 226Ra = 1.32+0.14
230Th = 1.42+0.27 234U = 1.59+0.27
232Th = 1.14+0.10 238U = 1.44+0.21
Each uncertainty listed is the standard deviation of an individual
measurement. That is, if another analysis were to be made, the result
would fall within that range 66.7% of the time. The mean standard
deviation for these analyses is 13.6%, which we believe to be
acceptable.
TABLE 5
Summary of Replicate Analyses Results
Range
Analytical Program <_(
EMSL-LV QA
Sample analyses
(n=134)
Routine analyses
(n=590)
All replicate analyses
made at EERF (n=724)
Theoretical distribution
TT+(T?+3aR)
8%
(11)
1%
(6)
2%
(17)
0
Note.--Numbers of analyses are given in parentheses.
13
-------
B. Blind Analyses
A routine program of submitting water samples of known
concentration to the analytical laboratory was instituted in 1980.
these samples were submitted "blind" to the analytical staff with a
request to perform specific analyses. The results of these analyses
are reported in Table 6. The known value given in the third column of
the table is either the concentration determined from previous analyses
or the concentration made by "spiking" the sample. A total of 52
"blind" analyses were performed for 16 radioisotopes.
Two samples containing less than the detectable concentrations
were correctly identified. Of the 50 remaining analyses, 64% of the
results were within 10%, 96% were within 25%, and only 4% (2 analyses)
differed by more than 30% of the known concentration. The latter two
939
analyses were for Pu and totaled 15% of the plutonium analyses.
14
-------
TABLE 6
Results of Intralaboratory Blind Analyses of Water Samples
Date
1/22/81
4/16/81
4/13/81
1/30/81
1/30/81
8/29/80
8/29/80
7/10/81
EERF Known R Date EERF Known
Values Value Values Value
3H
4230+_300 4/28/81 1300+^00 1060+^50
4280+300 4080+200 1.04
4200+300 4/21/80 400+200 500+200
8257^500
8241+500 8510+400 0.97
823U500
40,
1048+100 1150+60 0.91
Co
195M 202+3 0.97 1/30/81 11+4 13+_2
92+4 99+2 0.93
89Sr
<5 5.0
90-
Sr
6+^2 7+1 0.86
106Ru
209MO 219+30 1.01 1/30/81 66+15 61+2
232+40
R
1.23
0.80
0.85
1.08
-------
TABLE 6--Continued
Date
10/2/80
10/15/80
8/29/80
4/13/81
EERF Known
Values Value
9+7
l+_7 6+Jl
10+_7
6.9+0.6
6.3+0.9 6.8+1.4
7.4+1.0
11+8 <10
30+9 32+5
30+^9
R Date EERF
Values
2/18/81 123_+5
1.11 121+6
121+6
4/13/81 69+1
1.01 68+2
137Cs
1/30/81 384+9
0.94 1/30/81 105+8
1/30/81 20+3
Known
Value
138+^10
75+5
390HO
100+2
20+5
R
0.88
0.91
0.98
1.05
1.00
1/30/81 206+4 203+5 1.01
134Cs
1/30/81 157+5 151+5 1.04 1/30/81 81+3 80+2
1.01
12/11/80 18+4 21+2
210
Pb
0.86
-------
TABLE 6--Continued
Date EERF Known
Values Value
4/15/81 12.8+0.4 12+1
4/15/81 30.5+0.9 30+2
4/15/81 55+2 60+5
R Date EERF Known
Values Value
222n
Rn
1.07 5/18/81 24+1 24+2
1.02 5/18/81 6.0+0.3 6+1
0.92 5/18/81 12.1+0.4 12+2
R
1.00
1.00
1.01
8/15/80
2/8/81
8/7/80
10/9/80
17.8+0.2 16+1
226Ra
1.11 9/12/80
230,
28.1+0.2 26+2
Th
54+4
2.5+0.4
5.6+0.7
65+5
2.8+0.3
5.6+0.3
0.83
238,
0.89
1.00
2/5/81
6/1/81
28+3
37+5
28+2
45+3
1.08
1.00
0.82
11/7/80 28+3
28+3
1.00
-------
TABLE 6--Continued
Date
9/5/80
9/5/80
3/4/81
3/4/81
3/4/81
3/24/81
3/24/81
Note. --The
EERF
Values
15+2
12+2
4.2+0.2
3.8+0.4
4.0+0.6
16+2
8.3+0.8
known value
Known
Value
20_+2
15+3
4.0+0.6
4.0+0.6
4.0+0.6
20+3
20+3
is either the
R Date
238n
Pu
0.75 10/21/80
239D
Pu
0.80 3/24/81
1.05 6/1/81
0.95 7/10/81
1.00 7/10/81
0.80 7/10/81
0.42
concentration determined
EERF Known
Values Value
33+1 30+2
19+2 20+3
11+1 14+2
4.4+0.4 5.7+0.6
4.7+0.6 5.7+0.6
3.9+0.3 5.7+0.6
from previous analyses
R
1.10
0.95
0.79
0.77
0.82
0.68
or the
concentration made by spiking the sample.
-------
IV. SUMMARY
This report compiles the results of the EERF's inter!aboratory and
intralaboratory quality assurance programs during 1979 and 1980. The
inter!aboratory program consisted of participating in cross-check
analyses with the EMSL-LV, IAEA, and WHO Reference Centers. The
intralaboratory program consisted of blank, blind, spiked, and
replicate sample analyses. In general, the results were similar to
those observed in previous years, and we believe them to be
satisfactory-
A summary of the results from the EMSL-LV cross-check sample
program is presented in Table 1. These results are also presented
graphically in Fig. 1, where the cumulative probability is plotted with
the observed percent error in increments of 5%. The probability that
the result of an analysis will be within a selected error of the true
concentration during the reporting period can be easily ascertained
from the graph. For example, the probability that an analysis will be
within 25% of the correct concentration is seen to be 86%. One percent
of the analyses were in error by more than 50%. However, this graph
does not consider those results judged acceptable because of low sample
concentrations and the larger uncertainty in the results of those
analyses. With these additional samples included, about 95% of the
cross-check analyses were judged satisfactory.
In Fig. 2, the percent coefficient of variation is plotted with
the cumulative percent of samples. This figure gives an effective
19
-------
ro
O
1.0
0.9
0.8
>,
1 0.7
2
Q.
0)
3 0.6
0
0.5
0.4
0.3
s
•:•::;;;:::;::;: i:
!"!!"!
;;;;;
Hi!!
:| •:::::
•: :!!!!!
I ililil
! iiiii!
• ::::•;
" i:::::
i|| :::::::::
:::::::::::::
: :::::::::::::::»::
: !:::::::::!::::!:::
:::::::::::::::::::
:jj|:::::::j::::::{
I::::::::::::::::::
:|:|:|::|::|::|::::
:|::j:::::::|::::||
:;||{:j|j|:::::::::
: •:::•::::::::::::•:
fimzzA
!!!i!!!!!!!!!!!i!!!!!!Hi!!!!!!!!!!!!!!!:!!!!!!!!!!i!!!!!i!!!i!!!in!!!!!!!!i!i!!i!iHiil!
i!!!!!:!!!!!!!!!ii!!!!!!!H!i!!i!!i!!i!!!i!!!!!:i!!!:!!!!!i!J!!!!:!!!ii!!!!!!!!!!i!ii!!!!!
[[[j:::::::
::::::|:::::::::::::::::::::::::::::::::::::i:::::::::::::::::::::::::::::::::::::::::::::
!:!!!!!!:ir:i|!!!!i!!!!:!!i!:!!!!!!!i!i:::!i!!!!!!ii!:ii!!!i!!!!!!!i:!i:!yi!ii!!!!i::!!!::!
:||::i|:::|::H^||::i:::H:||:n::HH:Hi:|::H::;H|H:H:HHHHHH:H:H:H^
i!i!:!i::!:!:!!:!i!ii!i!!!!!!!!!!:iiii!!!:!!i!iiii:i!!!!!i!!!!::!::!r:!jii!!!:i!!ii!!!i!i:
[[[
!i!!l!!!!!l!li!!!iHl!!!!!!!l!iiii!!!lll!l!i!!H
5 10 15 20 25 30 35 40 45 50
-------
standard deviation for all sample types as calculated from the
normalized variances measured with respect to each known value.
Similar to the results shown in Fig. 1, this figure shows coefficients
of variation less than 25% for 80% of the samples. Approximately 2% of
the sample results had coefficients of variation greater than 50%.
In general, the results of the intralaboratory quality assurance
program were also satisfactory. The distribution of the precision
attained for all replicate analyses performed during 1979 and 1980
approached the expected theoretical distribution (see Table 5). Also,
the results of 96% of the blind or spiked sample analyses were within
25% of the known concentrations.
The most important function of the quality assurance program
described in this report is to identify problem areas in our analytical
laboratory. If a problem exists, immediate remedial action is
initiated. As is apparent in some of our reported results, errors can
occur for many reasons—improperly following a tested procedure,
arithmetical errors in the calculations, permitting contamination to
enter the sample during analysis, undetected fluctuations in counting
efficiencies and backgrounds, and using incorrect weights, absorption
factors, and abundances. It requires continual alertness and expedient
action to recognize and correct these problems when they arise.
The EERF is continually trying to improve the quality of its
analytical chemistry program. For example, a major problem that has
plagued the quality of some of our analytical results has been the
stripping of complex gamma-ray spectra. To remedy this problem, we are
now converting the matrix method to a more general least-squares method
of spectral stripping.
21
-------
In accord with our belief that laboratory quality results should
be public record, the EERF plans to issue in the future brief biannual
reports similar to this which will present the results of our quality
assurance programs.
22
-------
ro
CO
to
CL
E
c
0)
o
o3
0.
Q)
>
to
D
E
3
O
10
15 20 25 30 35
% Coefficient of variation
40
45
50
Fig. 2. The cumulative probability vs. the sample coefficient of variation
-------
REFERENCES
B179 Blanchard, R. L., Strong, A. B., Lieberman, R. and Porter, C. R.,
1979, "The Eastern Environmental Radiation Facility's
Participation In Inter!aboratory Comparisons of Environmental
Sample Analyses," Office of Radiation Programs, EPA, Technical
Note, ORP/EERF-79-1.
Cu68 Currie, L. A., 1968, "Limits for Qualitative Detection and
Quantitative Determination", Anal. Chem. 40, 586-593.
Ja76 Jarvis, A. N., Smiecinsky, R. F. and Easterly, D. G., 1976, "The
Status and Quality of Radiation Measurements in Water," U. S.
Environmental Protection Agency Rept., EPA-600/4-76-017.
Ka77 Kanipe, L. G., 1977, "Handbook for Analytical Quality Control in
Radioanalytical Laboratories," U. S. Environmental Protection
Agency Report, EPA-600/7-77-088.
Li82 Lieberman, R., "Radiochemical Analytical Procedures Manual-EERF,"
to be published.
Mo51 Mode, E. B., 1951, "Elements of Statistics," 2nd Edition,
Prentice-Hall, New York.
Ro64 Rosenstein, M., and Go!din, A. S., 1964, "Statistical Technique
for Quality Control of Environmental Radioassay," AQCS Report
Stat - 1, U. S. Public Health Service, Winchester, Mass.
We77 Welford, G. A. and Harley, J. H., 1977, "HASL Participation in
IAEA Intercomparisons," Energy Research and Development
Administration Report, HASL-322.
Wi81 Williams, A. R., 1981, "An International Comparison of 226Ra
Analysis by the Emanation Method," Health Phys. 41, 179-183.
24
-------
Appendix A—EMSL-LV Water Cross-Check Samples (pCi/1)
Date
4-79
6-79
8-79
10-79
12-79
2-80
2-79
6-79
10-79
EERF
Values
2370
2310
2370
1650
1780
1690
1530
1540
1500
1730
1720
1740
1960
1950
1950
1620
1560
1530
<50
<50
<50
<30
<30
<30
117
120
135
Known R Coeff. of Date EERF Known R Coeff. of
Value Var. (%) Values Value Var. (%)
4-80 3430
2270 1.04 3.7 3480 3400 1.01 1.5
3420
6-80 2160
1538 1.11 11.5 2180 2000 1.08 7.9
2130
8-80 1340
1480 1.03 3.2 1340 1210 1.11 11.0
1350
10-80 3270
1560 1.11 10.9 3210 3200 1.01 1.5
3240
12-80 2550
2040 0.96 4.2 2510 2240 1.13 12.7
2510
1750 0.90 10.5
51 Cr
2-80 101
0 — — 103 101 1.21* 35.5
163
6-80 <20
0 — — <20 13
<20
10-80 105
113 1.10 12.0 61 86 0.98 21.1
87
-------
Appendix A—Continued
to
cr\
Date
2-79
4-79
6-79
10-79
10-79(A)
2-79
6-79
10-79
2-80
EERF
Values
<10
<10
<10
14
15
19
44
45
45
<10
<10
<10
39
41
41
17
19
21
<20
<20
<20
<20
<20
<20
33
33
31
Known R Coeff. of Date EERF Known
Value Yar. (%) Values Value
2-80 11
9 — — 15 11
8
4-80 14
15 1.07 15.9 15 6
16
6-80 <10
47 0.95 5.1 <10 5
<10
10-80 20
6 — — 20 16
15
10-80(A) 13
33 1.22* 22.4 12 12
14
65Zn
6-80 22
21 0.90 12.3 22 23
22
10-80 20
0 — — 34 25
16
10-80(A) <10
0 — — <10 0
<10
25 1.29* 29.6
R Coeff. of
Var.(%)
1.03 26.2
2.50* 150.6
— —
1.15 20.7
1.08 10.8
0.96 4.4
0.93 31.6
—
-------
Appendix A—Continued
Date EERF
Values
1-79
4-79
5-79
9-79
10-79(A)
1-79
4-79
5-79
16
15
17
8
6
6
20
22
24
<5
<5
<5
11
10
11
7
7
6
9
10
9
33
33
27
Known R Coeff. of Date EERF
Value Var.(X) Values
1-80 6
14 1.14 15.4 6
6
4-80 3
9 0.74* 28.0 3
3
5-80 <5 '
23 0.96 8.3 <5
<5
9-80 26
3 — — 26
26
10-80(A) 7
12 0.89 11.8 7
7
90Sr
1-80 23
6 1.11 13.6 26
21
4-80 <1
8 1.17 17.7 <1
<1
5-80 15
30 1.03 10.0 14
16
Known R Coeff. of
Value Var.(X)
10 0.60* 40.0
4 0.75* 25.0
5
24 1.08 8.3
6 1.17 16.7
20 1.17 19.6
0
12 1.25* 25.9
-------
Appendix A—Continued
NJ
OO
Date
9-79
10-79(A)
2-79
6-79
10-79
2-80
4-79
EERF
Values
32
32
31
8
8
8
<30
<30
<30
<30
<30
<30
<30
<30
<30
56
37
68
59
59
59
Known R Coeff. of Date EERF Known
Value Var. (%) . Values Value
9-80 17
28 1.13 13.2 18 15
19
10-80(A) <1
7 1.14 14.3 <1 0
<1
106Ru
4-80 ND
0 ND 37
ND
6-80 34
0 — — 34 37
32
10-80 42
0 — ~ 35 46
57
10-80{A) <10
51 1.05 25.6 <10 0
<10
4-80 32
40 1.48 47.5 25 44
34
R Coeff. of
Var. (%)
1.20 20.7
__
—
0.90 10.2
0.97 20.2
__
0.69 32.3
-------
Appendix A—Continued
NJ
Date
8-79
12-79
2-79
4-79
6-79
10-79
10-79(A)
2-79
4-79
EERF Known R Coeff. of
Values Value Var. (%)
26
29 26 1.04 6.7
26
47
50 53 0.92 8.0
50
134r,
<10 6
<10
21
23 19 1.19 20.4
24
67
70 71 0.96 4.7
67
<10
<10 7
<10
52
55 56 0.99 5.3
59
137p,
13
11 12 1.00 6.8
12
<10
<10 0
<10
Date EERF Known
Values Value
8-80 29
26 36
29
12-80 20
20 22
20
2-80 9
9 10
12
4-80 24
25 8
25
6-80 14
7 11
8
10-80 18
16 20
17
10-80(A) 10
16 12
14
2-80 30
35 30
25
4-80 23
24 18
24
R Coeff. of
Var. (%)
0.78* 22.6
0.91 9.1
1.00 14.1
3.08 208.4
0.88 30.6
0.85 15.6
1.11 23.6
1.00 13.6
1.31* 31.6
-------
Appendix A—Continued
Date EERF
Values
6-79
10-79
10-79(A)
3-79
4-79
6-79
9-79
10-79
<10
<10
<10
17
11
11
<10
<10
<10
9.4
9.4
9.5
6.8
7.2
6.5
7.9
8.4
8.5
11.2
11.4
11.2
11.2
11.4
11.1
Known R Coeff. of Date EERF
Value Var. (%) Values
6-80 18
0 — — 21
17
10-80 12
11 1.18' 31.5 10
13
10-80(A) 18
0 — — 20
19
226Ra
3-80 16.2
8.2 1.15 15.1 17.0
16.3
4-80 17.4
5.9 1.16 16.5 17.6
17.4
6-80 2.0
8.4 0.98 3.5 2.0
2.0
9-80 14.2
10.2 1.10 10.2 13.7
14.3
10.2 1.10 10.2
Known R Coeff. of
Value Var. (%)
17 1.10 14.0
12 0.97 10.8
20 0.95 6.4
16.0 1.03 3.8
16.0 1.09 9.2
1.7 1.18 17.6
13.1 1.07 7.6
10-79{A) 11.2 10-80(A) 14.6
11.8 11.0 1.05 5.1 14.7. 12.8 1.14 14.1
11.5 14.5
-------
Appendix A—Continued
Date EERF
Values
12-79
3-79
4-79
6-79
9-79
10-79
10-79(A)
12-79
2-79
10.8
11.0
11.0
7.0
9.2
9.7
4.0
4.3
5.1
9.1
9.7
9.5
8.8
8.2
9.7
8.8
8.2
9.7
0.6
0.6
0.5
0.4
0.5
0.8
40
38
41
Known R Coeff. of Date EERF
Value Var. (%) Values
12-80 12.4
11.0 0.99 1.0 13.9
14.7
228Ra
3-80 ND
13.6 0.63* 37.5 ND
ND
4-80 31.6
6.2 0.72* 28.9 31.3
ND
6-80 1.0
14.4 0.66* 34.5 1.0
2.0
9-80 14.4
12.7 0.70* 30.3 13.7
14.0
12.7 0.70* 30.3
10-80(A) 11.9
0 — -- 9.6
10.4
12-80 12.6
1.0 0.57* 46.5 12.4
12.4
238/234u
8-80 28
35 1.13 13.8 28
28
Known R Coeff. of
Value Var. (35)
13.2 1.04 8.0
2.6
21 1.50 49.8
1.7 0.78* 35.1
14.4 0.97 3.2
9.2 1.16 18.7
10.5 1.19 18.8
27 1.04 3.7
-------
Appendix A—Continued
Date
8-79
2-80
1-79
7-79
1-80
1-79
3-79
EERF
Values
38
37
39
34
35
33
4.6
4.4
4.4
2.1
2.1
2.1
3.2
3.2
3.1
6
6
6
11
11
10
Known R Coeff. of
Value Var.(%)
33 1.15 15.4
32 1.06 6.8
239Pu
4.8 0.93 7.2
2.4 0.88 12.5
3.4 0.93 7.0
Gross Alpha
6 1.00 0
10 1.07 8.2
Date EERF Known R Coeff. of
Values Value Var. (%)
10-80(A) 5
5 5 1.00 0
5-80 6
7 8 0.79* 21.7
6
7-80 2.8
3.2 4.5 0.67* 33.4
3.0
11-80 5.4
5.9 7.2 0.81 ' 19.2
6.3
3-80 12
11 13 0.87 13.3
11
-------
Appendix A—Continued
U)
00
Date EERF
Values
4-79
5-79
9-79
10-79(A)
11-79
1-80
1-79
24
23
23
19
20
18
8
8
7
17
18
18
10
10
9
21
20
21
12
12
11
Known R Coeff. of Date EERF
Value Var. (%) Values
4-80 57
22 1.06 6.4 57
57
5-80 23
18 1.06 7.2 23
23
7-80 31
5 1.53* 54.2 31
31
9-80 25
21 0.84 16.0 24
24
10-80(A) 43
12 0.81 19.8 43
44
11-80 16
30 0.69 31.2 16
17
Gross Beta
16 0.73* 27.2
Known R Coeff. of
Value Var. (%)
98 0.58 41.8
23 1.00 0
36 0.86 13.9
32 0.76* 24.0
39 1.11 11.2
16 1.02 3.6
-------
Appendix A—Continued
U)
Date EERF Known R Coeff. of
Values Value Var. (%)
3-79
4-79
5-79
9-79
10-79(A)
11-79
1-80
16
17 16 1.02 3.6
16
31
33 44 0.73* 27.3
32
23
21 22 1.05 8.7
25
41
38 40 0.99 3.2
40
44
47 49 0.92 8.7
44
24
21 27 0.85 15.7
24
41
40 45 0.90 10.4
40
Date EERF
Values
3-80 19
20
20
4-80 51
51
51
5-80 16
15
15
7-80 34
31
31
9-80 20
20
20
10-80(A) 47
47
48
11-80 13
13
13
Known R Coeff. of
Value Var. (%)
22 0.89 10.8
100 0.51 49.0
14 1.10 10.1
38 0.84 16.2
21 0.95 4.8
60 0.79* 21.1
13 1.00 0
Notes: (1) R - The ratio of the average EERF value divided by the known value.
(2) Coeff. of var. (%) - The percent of the coefficient of variation.
(3) Date(A)-Blind performance evaluation study for radionuclides in water.
(4) ND - Not determined.
(5) AsterisM*) indicates results which differed from the EMSL-LV value by more than 20% but were judged
acceptable because of low sample concentration, the uncertainty in the measurement, and the
small absolute difference.
-------
Appendix B--EMSL-LV Milk Cross-Check Samples (pCi/1)
U)
ui
Date
1-79
4-79
7-79
11-79
1-79
4-79
7-79
11-79
EERF
Values
27
29
30
25
25
26
<5
<5
<5
16
18
16
22
19
21
56
56
56
13
13
12
21
21
21
Known
Value
33
42
5
25
19
54
11
17
R Coeff. of Date
Var.(%)
1-80
0.87 13.7
4-80
0.60 39.7
7-80
— —
10-80
0.67* 33.5
90$r
1-80
1.09 11.0
4-80
1.04 3.7
7-80
1.15 15.7
10-80
1.24* 23.5
EERF Known R Coeff. of
Values Value Var. (%)
ND
ND 10
ND
10
11 10 1.00 8.2
9
51
51 55 0.93 7.3
51
23
23 23 1.00 0
23
ND
ND 25
ND
16
15 15 1.07 8.6
17
19
18 17 1.04 8.3
16
<1
<1 0
<1
-------
Appendix B--Cont1nued
CO
cr>
Date
1-79
4-79
7-79
11-79
1-79
4-79
7-79
EERF Known
Values Value
108
101 105
106
95
95 96
94
19
19 17
20
628
639 637
628
52
50 49
51
156
158 154
154
13
15 12
11
R Coeff. of Date EERF Known
Var. (%) Values Value
131T
i
1-80 <10
1.00 2.8 <10 0
<10
4-80 38
0.99 1.5 42 33
36
7-80 <10
1.14 14.0 <10 0
<10
10-80 12
0.99 1.2 13 18
12
1-80 ND
1.04 4.4 ND 40
ND
4-80 27
1.01 1.7 28 28
26
7-80 27
1.08 16.0 28 35
30
R Coeff. of
Var.(%)
__
1.17 18.8
—
0.69* 31.6
__
0.96 4.6
0.81 19.4
-------
Appendix B--Cont1nued
Date
11-79
1-79
4-79
7-79
11-79
1-79
EERF
Values
46
48
50
<10
<10
<10
<10
<10
<10
<10
<10
1420
1440
1480
Known
Value
49
0
0
0
0
1560
R Coeff. of Date EERF
Var. (%) Values
10-80 23
0.98 3.9 25
22
1-80 ND
ND
ND
4-80 <10
7-80 <10
10-80 <10
K(mg/l)
1-80 ND
0.93 7.4 ND
ND
Known R Coeff. of
Value Var. (%)
21 1.11 12.6
0
0
0
0
1600
-------
Appendix B—Continued
Date EERF Known R Coeff. of
Values Value Var. (%)
4-79 1350
1450 1560 0.90 10.8
1390
7-79 1470
1550 1630 0.91 9.5
1430
11-79 1340
1380 1470 0.94 6.5
1420
Date EERF Known
Values Value
4-80 1160
1150 1190
1160
7-80 1460
1460 1550
1430
10-80 1340
1430 1700
1410
R Coeff. of
Yar.(%)
0.97 2.8
0.94 6.5
0.82 18.2
U)
(2) Coeff. of Var. (%) - The percent of the coefficient of variation.
(3) ND - Not determined.
(4) Asterisk(*) indicates results which differed from the EMSL-LV value by more than 20% but were
judged acceptable because of low sample concentration, the uncertainty in the measurement, and
the small absolute difference.
-------
Appendix C--EMSL-LV Food Cross-Check Samples (pCi/kg)
CO
VD
Date
3-79
7-79
11-79
3-79
7-79
11-79
3-79
7-79
11-79
EERF
Values
37
31
31
<5
<5
<5
63
57
61
27
29
28
9
8
8
33
30
31
90
88
88
20
22
22
124
119
134
Known R Coeff. of Date EERF Known
Value Var.(%) Values Value
8V
48 0.69 31.8
7-80 75
8 — — 79 94
77
11-80 7
73 0.83 17.7 7 8
7
9°Sr
22 1.27* 27.5
7-80 18
3 2.78* 178 14 18
18
11-80 <1
27 1.16 16.7 <1 0
-------
Appendix C--Continued
o
Date
3-79
7-79
11-79
3-79
7-79
11-79
3-79
7-79
EERF
Values
70
73
75
35
35
32
22
24
15
<10
<10
38
<10
2722
2741
2694
2590
2570
2640
Known R Coeff. of Date EERF Known
Value Var. (%) Values Value
74 0.98 3.3
7-80 27
33 1.03 5.2 30 27
28
11-80 12
22 0.92 19.1 13 12
8
0 — —
7-80 <10
0 — — <10 0
11-80 <10
0 — — <10 0
K(mg/kg)
2700 1.01 1.0
7-80 2550
2650 0.98 2.2 2520 2660
2450
R Coeff. of
Var.(%)
1.05 6.8
0.92 19.8
—
—
0.94 6.0
-------
Appendix C—Continued
Date
11-79
Notes
EERF
Values
1240
1300
1310
: (1)
(2)
(3)
(4)
Known R Coeff. of
Value Var.(X)
1511 0.85 15.2
Date
11-80
EERF
Values
2340
2290
2310
Known
Value
2520
R- The ratio of the average EERF value divided by the known value.
Coeff. of Var. (%) - The percent of the coefficient of variation.
ND- Not determined.
Asterisk(*) Indicates results which differed from the EMSL-LV value by
judged acceptable because of low sample concentration, the uncertainty
small absolute difference.
R Coeff.,
Var. (%)
0.92 8.2
more than 20% but were
in the measurement, and
of
the
-------
Appendix D--WHO Intercomparison Sample Results
Nuclide
EERF
Values
IRC R
Value
Coeff. of Nuclide
Var. (%)
Freshwater Fish (No.
Sr-90(pCi/kg)
Cs-137(pCi/kg)
Ra-226(pCi/kg)
6130+251
6020+235
5860+252
1500+750
1500+750
1600+800
3620+36
3820+38
3860+39
63400+300 0.95
1470+90 1.04
3500+700 1.08
5.3
5.4
8.2
EERF IRC R Coeff. of
Values Value Var. (% )
F290) November 1979
U(ng/kg)
Ca(g/kg)
K(g/kg)
fc! Marine Sediment (No. E114) December 1977
Sr-90
Ru-106/Rh-106
Sr-90(pCi/l)
H-3
0.14+0.10
0.15+0.10
0.16+0.10
6.6+0.7
6.4+0.7
6.5+0.7
15+4
15+3
4920+450
4470+450
4860+450
0.21+0.02 0.71
6.4+0.3 1.02
Liquid
16.0+0.8 0.92
Rainwater
4-640+40 1.02
29
2.0
Milk (No.
8.8
(No. F856)
4.9
Cs-137
Ce-144/Pr-144
E414) March 1978
Cs-137(pCi/l)
1770+248
1815+254 1620+370 1.02 12.6
1368+287
58+5
58+6 62+2 0.94 6.4
58+6
10+1
12+1 9+1 1.22 2.4
(pCi/g)
0.17+0.3
0.17+0.3 0.18+0.01 0.92 8.1
0.17+0.3
4.6+0.5
4.4+0.5 4.3+0.2 1.03 4.2
4.3T0.5
34+10
32+^10 31+1 1.02 7.0
October 1980 (pCi/1)
-------
Appendix D--Continued
Nuclide EERF
Values
Sr-90(pCi/kg) 20+5
20+5
20+5
Cs-137(pCi/kg) <5
<5
<5
Ra-226(pCi/kg) 17.0+0.9
16.6+0.8
12.4+0.7
i.
J
Sr-90 46+7
45+7
46+8
Ru-106/Rh-106* 405+61
388+54
382+57
Ca(g/l) 1.2+0.1
1.2+0.1
1.1+0.1
IRC R Coeff. of Nuclide
Value Var. (% )
Cereals (No. F553) March 1980
Ca(g/kg)
18+1 1.11 11.1
Mg/kg)
5.9+0.4
U(ug/kg)
4.2_+l 3.65 270
Seawater (No. F712) June 1980 (pCi/1
- Sb-125
53+5 0.86 13.9
Cs-137
850_+70 0.46 54
K(g/l)
1.13+0.03 1.03 5.3
EERF IRC R Coeff. of
Values Value Var. (%)
0.36+0.04
0.36+0.04 0.40+0.02 0.90 10.0
0.36+0.04
5.0+0.5
4.6+0.5 4.4+0.2 1.06 8.3
4.4+0.4
1.9+0.7
1.3+0.6 <0.4
1.3+_0.6
47+12
53+JL5 42+8 1.17 18.5
<5
<5 1.3+0.1
<5
1.6+0.2
1.5+0.2 1.48+0.04 1.04 4.8
1.5+0.2
Pond Water Samples (Nos. E888,889) October 1978(pCi/l)
H-3 5500+400
5400+400
5600+400
H-3
5760+260 0.95 4.7
19,100+600
19,200+600 19,100+900 1.00 0.96
18,800+600
* Failed to multiply the Rh-106 concentration by two to include Ru-106.
-------
Appendix D—Continued
Nuclide
Sr-90(pCi/l)
Cs-137(pCi/l)
H-3UCi/l)
Mu-54(pCi/l)
EERF
Values
47+7
47+7
46+7
292+25
283+25
281+25
2.4+0.2
2.3+0.2
2.3+0.2
723+110
726+110
742THO
IRC
Value
42+2
250+7
Low-Level
2.5+_0.1
780+J70
R
Liquid
1.11
1.14
Coeff. of
Var. (%)
Milk (No.
11.2
14.3
Radioactive Liquid
0.93
0.94
6.9
6.5
Nuclide
F140) June 1979
Ca(g/l)
K(g/l)
Waste (No. FIDO)
Co-58(pCi/l)
Co-60(pCi/l)
EERF IRC R Coeff. of
Values Value Var. (*)
1.2+0.2
1.2+0.2 1.23+0.03 0.98 2.4
1.2+0.2
1.4+0.2
1.5+0.2 1.55+0.04 0.95 6.2
1.5+0.2
March 1979
47,400+510
47,700+510 51,800+1,700 0.91 8.6
47,000+510
2,300+200
2,400+200 2,310+120 1.02 3.2
2,400+~200
-------
Appendix E—IAEA Intercomparison Sample Results
Nuclide EERF
Values
H-3 2200+200
2400+200
2500+200
Sr-90 7.4+1.6
7.6+1.7
7.2^1.6
Sr-89 1.59
>£>
Ul
True R Coeff. of Nuclide
Value Var. (%)
Water (No. W-l/2) pCi/1
Ce-144
2,330 1.02 5.6
4.15 1.78 78
Liquid Milk (No. A-9/1) 1979-1980
Cs-137
1.52 1.05 (4.6)
EERF
Values
103+52
115+58
118+59
(pCi/1)
1004+20
1009+20
1003+20
1006T20
998+20
True R Coeff. of
Value Var. (%)
5.6 20 1900
1040 0.97 3.5
Sr-90
0.86
0.87
0.99 (1.2)
------- |