SIXTH IAEA INTERCOMPARISON OF LOW-LEVEL

TRITIUM MEASUREMENTS IN WATER (TRIC2000)

REPORT

prepared by

M. Gröning1, C.B. Taylor2, G. Winckler1,3, R. Auer1, H. Tatzber1

1
Isotope Hydrology Laboratory, IAEA, Vienna
2
28 Wyndrum Avenue, Lower Hutt 6009, New Zealand
3
now at Lamont-Doherty Earth Observatory, Columbia University, New York

Isotope Hydrology Laboratory

Agency’s Laboratories Seibersdorf and Vienna
International Atomic Energy Agency

Vienna, October 2001

0
 1

1. INTRODUCTION

It is now 36 years since the isotope intercalibration programmes of the International Atomic
Energy Agency (IAEA) were initiated. Regular exercises are conducted for stable isotopes of
light elements and for tritium, the radioactive isotope of hydrogen. The Agency’s
environmental isotope intercalibration programme is now much larger than originally
conceived, being a component of the Analytical Quality Control Services (AQCS). The
purpose of AQCS is to assist laboratories engaged in various fields of research to check and
maintain measurement quality.
Five previous tritium intercomparisons were conducted [1-6](Cameron, 1967; Florkowski,
Payne and Sauzay, 1970; Taylor, 1978; Hut, 1986a, 1986b: Östlund, Scott and Taylor, 1995).
Basic details of these and of the recent one are summarized below:

Table 1: Details of conducted IAEA tritium interlaboratory comparison exercises.

Year Number of participants, Water samples distributed

(approached) and tritium concentration [TU]
1965 12 (29) Sample A: < 100 TU
Sample B: 200 - 800 TU
Sample C: 1000 – 6000 TU

1970 35 (38) A: ~ 10 TU
B: ~ 50 TU
C: ~ 250 TU

1975 41 (79) A: < 0.2 TU

B: 8.37 ± 0.04 TU
C: 33.1 ± 0.1 TU
D: 678.5 ± 2.2 TU

1985 57 (85) A: < 0.2 TU

B: 5.94 ± 0.03 TU
C: 25.66 ± 0.13 TU
D: 100.2 ± 0.5 TU

1994 57 (91) A: ~ 0 TU
B: ~ 1 TU
C: ~ 5 TU
D: ~ 19 TU

2000 86 (102) T3: 0 TU

T2: 2.152 TU
T4: 5.252 TU
T1: 10.112 TU
T5: 26.118 TU
T6: 505.83 TU

The quoted precise values for 1975, 1985 and 2000 refer to the chosen reference dates for
these intercomparisons. Bracketed 2nd column no. are the numbers of laboratories approached.
 2

In each case, a set of water samples was distributed, covering the concentration range then
prevailing in environmental waters. In 1965, 1970 and 1995, the tritium values were known
only approximately. In 1975 and 1985, samples were prepared by gravimetric dilution of
tritiated standard water NBS4926 (now referred to as SRM4926) with water of near-zero
tritium concentration. In the present exercise of 2000, samples were prepared by gravimetric
dilution of tritated standard water SRM4927F with virtually tritium-free water.

Laboratories were advised of the approximate concentrations in advance, and were requested
to provide detailed information on techniques and procedures. Results of the first two
exercises were reported with identification of the laboratories, but anonymity has applied
since 1975.

Participation in IAEA intercomparison exercises is voluntary. However, leading laboratories

of recognized experience and quality are usually requested to participate, and some may also
be involved in the data evaluation. Results are often discussed at meetings of invited experts,
from which decisions and recommendations are passed to the scientific community.

Most laboratories applying environmental isotopes in geochemistry, hydrology, oceanography,

agriculture, plant physiology, life sciences and other environmental studies have participated
in the Agency’s intercomparison programmes. These are valuable opportunities to check
measurements against those of other laboratories, to discover the occurrence of errors and
systematic deviations, to stimulate improvements, and to promote confidence in data from
international research projects encompassing the efforts of several laboratories.

Table 1 reveals a continuing increase in the number of laboratories measuring tritium at

environmental levels. A further increase of participants occurred in this 6th intercomparison
(code-named TRIC2000). The increase has occurred despite the decline of tritium
concentrations in rainfall to very low levels, and the elapse of 3 half-lives since maximum
levels were recorded in the mid-1960s following atmospheric H-bomb tests. The present low
tritium environment has persisted now for about 15 years, and levels in the ocean and
groundwater will continue to decline. Levels in the southern hemisphere are only marginally
above pre-bomb levels. These markedly lower concentrations impose much greater technical
difficulty in measurement. It is therefore important to judge from the present exercise how
well laboratories have improved their sensitivity and accuracy.

In 1982, IAEA recommended a fixed tritium scale based on reference standard water
NBS4926C (now SRM4926C, US National Institute of Standards and Technology, (NIST))
and half-life 4540 days, (12.43 years) [7]. With SRM4926C now entirely used up, a new set of
NIST tritiated standard waters has been prepared and issued, and massic activity ratios have
been measured for all their previous and present standards [8,9]. It is very opportune that
NIST has simultaneously conducted a comprehensive review and critical evaluation of the
tritium half-life [10]. This has demonstrated that the previously adopted half-life is too long,
and evaluated a new value of 4500 ± 8 days. The coincidence of these two projects allows all
the scale parameters to be readjusted simultaneously. This is believed to be the best possible
approach to end the continuing diversity of standardization applied in different laboratories.
A recommendation to this effect is being prepared for publication. Some details are necessary
for the present report; these are considered in Appendix B.

102 laboratories declared their willingness to participate in this 6th interlaboratory comparison
and have received the set of six water samples. Results on their tritium concentration have
 3

been received in 89 independent data sets from 86 laboratories at the time of writing this
report. Six samples of different tritium concentrations were prepared at the Agency’s
laboratory. One was a virtually tritium-free water, collected from the same deep artesian
aquifer at Grafendorf (near Graz, Austria), which was used in earlier IAEA tritium
interlaboratory comparisons. The other samples were prepared by accurate gravimetric
dilution of the new NIST tritiated water standard SRM4927F, as described below.

2. SAMPLE PREPARATION

As was the case in the earlier intercomparison exercises, the submitted results indicate that
various tritiated standard waters and half-lives are still adopted by the laboratories to calibrate
their measurements, despite the 1982 recommendation to adopt a fixed tritium scale. In light
of the new robust tritium half-life assessment [10], and the associated re-evaluation of all
NIST tritiated water standards [9], we have decided on this occasion to calculate the prepared
values and their uncertainties based on the newly determined half-life 4500 ± 8 days and the
NIST tritiated water standard SRM4927F. To achieve this, we have incorporated four
uncertainty sources: (a) the calibration uncertainty attached to the certified NIST standard
tritiated water SRM4927F (calibration date 3 Sep 1998) used for sample preparations; at a
stated standard uncertainty of 0.36%, this is the dominant contributor; (b) the small
uncertainty in correcting the values to our reference date (1 July 2000) using the half-life 4500
± 8 days; (c) the smaller errors of our dilution procedures; (d) the calculated uncertainties have
been broadened by including an upper limit for a possible tritium concentration for sample T3,
which was used in preparing the other five samples. As detailed below, this upper limit
resulted from statistical analysis of the submitted results for that sample.

Throughout this report, the results on the sample tritium concentrations are expressed in terms
of their massic tritium activity [Bq/kg] and as their tritium/hydrogen isotope ratio in terms of
TU (tritium unit). Strictly speaking, this is not a “unit” but a ratio and may be referred to more
rigorously as tritium ratio. The tritium ratio of 1 TU corresponds to a tritium/hydrogen ratio of
T/H=10-18. The massic tritium activity is related to the tritium ratio via the following
relationship:
1 Bq/kg corresponds to 8.390 ± 0.015 TU [11], using the new value for the tritium half-life
[10] and the CODATA 1998 value for the Avogadro constant [12].

Adoption of half-life 4500 days means that results from laboratories using other half-lives for
their calibrations will be biased relative to the “true” tritium values and ranges of the samples
as calculated for this exercise. For example, those laboratories calibrating their measurements
by the former NIST water standard SRM4926C (calibration date 9 Sep 1982) and 4540 day
(12.43 year) half-life, should bear in mind that their results will be positively biased by 1.55%
relative to the prepared values we provide here, based on SRM4927F (calibration date 3 Sep
1998). Some results may therefore appear to be marginally high outliers, but would be within
the correct range if adjusted to the new half-life. Similarly, some results, which are within the
ranges given here, may actually be marginal outliers on the negative side. We recommend that
laboratories make their own final assessments of the quality of their results. We have added
Appendix B, which provides help to make the adjustments according to published procedures
[7]. A different approach for this adjustment is suggested in [13].
 4

2.1 Massic Activity of SRM4927F at Reference Date

NIST’s calibration certificate gives the massic activity of SRM4927F as 634.7 kBq.g-1 on 3
September 1998, with a combined standard uncertainty of 0.36%; this uncertainty is the
square root of the sum of squares of all error components in the calibration. We have applied
the tritium half-life value 4500 ± 8 days evaluated by NIST [10] to calculate the massic
activity at our intercomparison reference date (1 July 2000). The standard uncertainty of the
half-life introduces an additional error component, which increases with time elapsed since 3
September 1998. Expressed as a percentage uncertainty, to be added in quadrature with the
combined standard uncertainty (0.36%), this error is 2.378×10-5×t, where t is the elapsed time
in days; this is only 0.016% to our reference date 1 July 2000 (667 days). The massic activity
of SRM 4927F at that date is 572.7 ± 2.1 kBq.g-1.

The website http://srmccatalog.nist.gov/srmcatalog/certificates/srm-4927f.htm provides full

details of the SRM4927F calibration.

2.2 Sample Dilution Procedure

SRM4927F was diluted in 3 stages to obtain a daughter water of appropriate tritium activity to
serve for common laboratory purposes (see Table 4 in Appendix A for details). The diluting
water was a de-ionised sample from an artesian well in Grafendorf, Austria, which was used
as “tritium-free” water in earlier interlaboratory comparisons and is used in the IAEA Isotope
Hydrology Laboratory for the daily routine work. The third dilution of SRM48927F served as
a common mother water for a final stage of dilution to obtain samples T1, T2, T4, T5, T6 (see
Table 5 in Appendix A). Sample T3 taken in June 2000 from the same artesian well in
Grafendorf as mentioned above was used as the diluting water for this final stage. Full details
of the dilution procedures are given in Appendix A. All dilutions were performed using
calibrated balances, checking the zero reading and a test weight before and after each
weighing process. The process consisted of the weighing of (a) the empty containers, (b) the
filled containers and, in case of the tritium standard transfer, (c) the empty transfer container
(syringe). The combined uncertainty of the weighing procedure was calculated according to
the law of uncertainty propagation, taking into account balance precision, maximum drift and
display uncertainty. No buoyancy correction was applied.

2.3 Tritium Concentration of Diluting Water (T3)

Sample T3 was used as diluting water to prepare the five other samples. This water was
collected in 150 litre stainless steel barrels and was not de-ionised. Its tritium concentration
was assessed using results submitted by the participating laboratories. Only the 35 results with
stated uncertainties less than 0.25 TU were considered in the evaluation as “high precision”
laboratories. Obvious outliers were removed before evaluating a first arithmetic mean value
with its standard deviation. A range of acceptability to judge individual results was based on
the mean value ± twice its standard deviation. Results differing by more than twice the stated
laboratory measurement uncertainty from those upper or lower limits were discarded, and a
second mean value evaluated. The process was repeated until further repetition produced no
more changes to the outlier population. The 26 remaining accepted results provided a mean
value of 0.028 ± 0.022 TU (standard error of the mean) (see Table 2). This value is
indistinguishable from zero, which is certainly its value at the source. Therefore it was
 5

decided to set the tritium ratio for T3 to 0 TU with an expanded upper uncertainty limit.
Under this condition its upper limit is 2×0.022 + 0.028 = 0.072 TU at the two-σ level. This
value for the tritium ratio of the diluting water T3 has been applied in calculating the tritium
ratios of the other five samples, whose stated ranges thereby include all components of
uncertainty at the two-σ level.

The same evaluation procedure for the tritium ratio of sample T3 can be used to compare
results based on the three individual measurement methods, liquid scintillation counting
(LSC), gas proportional counting (GAS) and 3He mass spectrometry (3HE) for the same sub-
set of laboratories. In addition, in the third column of Table 2 similar results are calculated
including all participating laboratories.

Table 2: Evaluated mean tritium ratios with standard errors of the mean for sample T3, obtained from
the sub-set of “high precision” laboratories (left column) and by all laboratories (right
column). In brackets, the number of accepted/reported results is given. Note: For the further
evaluation a value of 0 TU ± 0.072 TU at 2-σ level is used (see text).

Tritium ratio and standard error of the Tritium ratio and standard error of
mean for “high precision” laboratories the mean for all participating
[TU] laboratories [TU]
All data 0.028 ± 0.022 (26/35) -0.004 ± 0.093 (63/80)
LSC 0.006 ± 0.029 (16/26) 0.067 ± 0.120 (52/66)
GAS 0.063 ± 0.074 (3/3) -0.033 ± 0.086 (7/7)
3HE 0.020 ± 0.014 (5/6) 0.020 ± 0.014 (5/6)

The results of all data sets in Table 2 are consistent with each other within their evaluated
standard errors of the mean. All these mean values are indistinguishable from zero (tritium
free water) at the 2-σ level.

2.4 Reference Values for the Prepared Samples T1 to T6

The reference tritium ratios and combined standard uncertainties for the prepared six samples
T1 to T6 can now be calculated from the combination of all dilution factors for the original
standard solution SRM4927F. The uncertainties are not symmetrical around the reference
values, but larger at the upper side. By considering all possible sources of uncertainty we have
ensured that the true tritium ratios lie with 95% probability within the calculated upper and
lower uncertainty limits.

The calculated reference values and their ranges are given in Table 3. They are believed to be
the best possible approximation of the true tritium values and are consequently used in the
further data evaluation, in the graphical display and for the determination of outliers in the
data sets.
 6

Table 3: Reference values and uncertainty ranges for the tritium ratio of samples T1 to T6 (see also
Table 4 in Appendix A).

Sample Name Reference Value Lower Uncertainty Upper Uncertainty

[TU] Limit (at 2-σ level) Limit (at 2-σ level)
[TU] [TU]
T1 10.112 10.038 10.258
T2 2.152 2.136 2.240
T3 0 0 0.072
T4 5.252 5.214 5.362
T5 26.118 25.928 26.380
T6 505.83 502.15 509.58

All data evaluation and especially the outlier determination was performed with the PASCAL-
based software ICE (InterComparisonEvaluation, version 4.5), specially developed at IAEA
for the purpose of providing a standardized evaluation tool for the interlaboratory comparisons
and proficiency tests performed in the field of stable isotopes and tritium. The program was
tested and results were verified by simulated data sets and comparison with the output of the
same algorithms used with commercial spreadsheet programs. The main advantage was found
to be the easy application of outlier detection procedures and robustness of ICE for different
data sets.

3. PRESENTATION OF RESULTS

The submitted results are listed in Table 6 in Appendix C, with columns listing laboratory
identification number, reported tritium ratio and reported measurement uncertainty in TU for
the six samples, the used measurement method, and remarks on any necessary data
modification. Results are ordered according to laboratory code number. Used measurement
methods include Liquid Scintillation Counting, Gas Proportional Counting and 3He Mass
Spectrometry. Empty cells indicate data not submitted. Some results could not be assessed due
to failure to submit measurement uncertainties. Laboratories which submitted 2 sets of results
by applying different methods have suffixes a and b attached after their lab code numbers
(except for 93a and 93b, where two laboratories measured the same set of samples).

Additionally, results for each sample are depicted graphically, in increasing order of tritium
ratio, using S-plots (Figs.1-6); these show the relation to the reference value and its
uncertainty range, represented by the lower and upper uncertainty limits from Table 3. Error
bars are added at 2 times the submitted uncertainties, enabling easy visual identification of
outliers (open markers instead of filled ones). The laboratory code numbers (Lab ID) appear
along the x-axes. In all figures, the suffix “-D” added to Lab ID’s signalises direct counting
without prior tritium enrichment for samples T1 to T5.

All the S-plots are asymmetric around the reference values and their uncertainty limits, with
more results on the high side. A part of this asymmetry is likely to be due to the use of older
standard waters and higher half-lives (e.g. 12.43 years = 4540 days) by many laboratories for
their calibrations, whereas the reference values for the prepared samples are based on the new
SRM4927F standard water and half-life 4500 ± 8 days (see section 2). However, this cannot
completely account for the overall biases of the population means from the reference tritium
 7

ratios for the five tritium bearing samples. Those biases decrease with increasing tritium ratio
from +20.3% (T2), +7.6% (T4), +3.3% (T1) and +2.1% (T5) to +0.9% (T6). The underlying
reasons for the systematic biases towards higher tritium values cannot be interpreted
conclusively from the available information.

Because of the asymmetry of the S-plots and the above-mentioned underlying reasons,
average results for the samples calculated from the whole population of laboratories are of
limited usefulness, and are therefore not reported here.
 8

S-Plot for Sam ple T1

16

15
Accepted data, sam ple T1
R ejected data, sam ple T1
14

13
Tritium ratio [TU]

12

11

10

6
5
98
65
68
55

59
82
3
7

111
28

47
51
69
79
108
54
12
23
77

66
35
27

95

10b
21
31b
33
9
103
57

91
94
101
31a
46
10a
16
32
15
39

19
92
14
17
22
88
1
36
38
8
6
29
43
18
71
76
25
41
89
67
45
24
48

86

44
42
61
83
37

72
106
93b-D
85-D

75-D

73-D

26-D

84-D

56-D

4b-D

49-D

4a-D

96-D

58-D
Laboratory ID

Fig.1: S-plot for sample T1. Laboratory results at the utmost left and right sides are outside the chosen scale. Rejected values are marked with open symbols.
 9

S-Plot for Sample T2

8

7 Accepted data, sample T2

Rejected data, sample T2

6
Tritium ratio [TU]

0
98

65
25
5
3
22
95
61
45
28

76
59
17

82
79
111
12
32
43
69
94
57
33
39
108
15

10b
6
9
16
23
48
77
89
92
8
14
18
10a
101
106
31b
55
7
47
68
88
38
103
41
1
67
36
31a
71
29
42
21
27
35
51
93a
54

24
19

46
86
37

44
91

72
83
85-D
73-D

56-D

4a-D

58-D

4b-D

96-D

84-D

26-D

75-D

93b-D
Laboratory ID

Fig.2: S-plot for sample T2. Laboratory results at the utmost left and right sides are outside the chosen scale. Rejected values are marked with open symbols.
 10

S-Plot for Sam ple T3

5

4
Accepted data, sample T3
Rejected data, sample T3
3

2
Tritium ratio [TU]

-1

-2

-3

-4
61
25
51
65
77
101
18
68
27
103
29
28
3

16
32

1
10a
12
9
22
35
59
66
82
92

94
15
6
5
69
55
7

67
48
71
95
98
21
38
76
8
31b
57
39
41
10b
108
46
23
31a

45
14
93a
88
24
91
19
36
17
47
89
42
54
86
72
44
33
37

83
106
85-D
73-D
56-D

4a-D

96-D
26-D

93b-D

4b-D

58-D

84-D
Laboratory ID

Fig.3: S-plot for sample T3. Laboratory results at the utmost left and right sides are outside the chosen scale. Rejected values are marked with open symbols.
 11

S-Plot for Sample T4

10

9
Accepted data, sample T4
8 Rejected data, sample T4

7
Tritium ratio [TU]

0
91

98
5
111
65
27
45
22
51
79
29
86
43

82
88
103
93a
3
69

57
21

59
18
23
95
15
66
9
17
31a
41
76
77
10b
35
6
8
12
32
47
94
101
10a
16
48
54
71
92
108
7
28
55
89
14
38
1
31b
68
39

67
46
25
24
36
19

83

44

61
42
33
37
72
106
96-D
85-D

56-D

4a-D

4b-D

84-D

58-D

26-D

73-D

93b-D
Laboratory ID

Fig.4: S-plot for sample T4. Laboratory results at the utmost left and right sides are outside the chosen scale. Rejected values are marked with open symbols.
 12

35 S-Plot for Sample T5

Accepted data, sample T5
Rejected data, sample T5

30
Tritium ratio [TU]

25

20

15
93a
5
65
98
54

68
59
43
88
3
111
71
79

12
45
44
39
76
57
27
31a
33
91
10b
66

23
32
47
55
82
19
69
8
31b
94
95
103
29
35
6

10a
15
9
92
46
101
21
77
25
36
41
17
14
89
28
18
16

51
24
1
7

86
38
72
48
67
83
108
42
22
37
61
106
49-D
85-D

96-D

4b-D

4a-D

84-D

58-D

73-D
56-D

93b-D
26-D

75-D

81-D
Laboratory ID

Fig.5: S-plot for sample T5. Laboratory results at the utmost left and right sides are outside the chosen scale. Rejected values are marked with open symbols.
 13

S-Plot for Sample T6

600

580
Accepted data, sample T6
Rejected data, sample T6
560

540
Tritium ratio [TU]

520

500

480

460

440

420

400
106
98
65
54
44
3
18
43
93a
69
95
57

12
68
19
7
79
22
103
55
8
45
15
72
94
39
14
48

28
76
86
92
62
29
47
51
35

10a
59
33
21
36

32
42

66
71
37
46
88
10b
17
1
101
24
9

83
41

82
27
16
91

89
108
6
67
38
31a
77
31b
25
109-D

93b-D
49-D

96-D

4b-D

73-D

4a-D

85-D
84-D

58-D

26-D
99-D

75-D

81-D
56-D
Laboratory ID

Fig.6: S-plot for sample T6. Laboratory results at the utmost left and right sides are outside the chosen scale. Rejected values are marked with open symbols.
 14
3.1 Implications of Results for Sample T3

Due to the very low range of environmental tritium concentrations now prevailing, sample T3
has special importance in this exercise. Results for the equivalent sample in the 5th
intercomparison revealed that slight contamination was prevalent at several laboratories. The
overall mean average of the results, after removal of outliers, was slightly higher by about 0.1
TU, suggesting a small degree of contamination on average; 7 out of 51 results (14%) were
obvious outliers.

In a similar assessment for T3 in the present exercise, 80 laboratories submitted values, but 5
could not be used in the evaluation, due to absence of the requested measurement uncertainty.
Of the remaining 75 results, 15 (20%) were outliers on the high side. 2 negative results were
obviously incorrect, being negative even if the reported measurement errors were taken into
account. The results for sample T3 for all accepted laboratories yielded a mean value of zero
(-0.004 ± 0.093 TU), see Table 2. However, this mean value is biased by one laboratory
(no.85) reporting a quite negative value and shifting the mean value considerably towards
zero.

In nearly all cases, high outliers must be due to contamination of the sample during storage
and/or processing. 14 of the 15 high outlier laboratories applied sample pre-distillation and
electrolytic enrichment before counting. Handling in a contaminated atmosphere is the most
likely cause of such results. To achieve satisfactory results at such low levels, the tritium
concentrations of atmospheric vapour in processing laboratories or sample storage should not
noticeably exceed that of outside uncontaminated air. At all laboratories, exposure of the
samples to atmosphere during storage and processing should be rigidly kept to the absolute
minimum. However, some laboratories are known to be unsuitably sited in contaminated
areas; this now severely limits their effectiveness for hydrological research. A very common
cause of laboratory contamination is the insidious influence of tritiated vapour escaping from
luminous watches. To avoid this contamination, it is essential to exclude the use of such
watches by laboratory personnel, and to prevent emitted vapour from entering the laboratory
from other areas of buildings, particularly via air-conditioning. Policing of entry, and regular
and thorough monitoring of vapour samples within and outside laboratories is the only way to
combat this problem. Ideally, a tritium laboratory should be housed in its own building,
allowing better control. Prolonged contamination within a building takes a long period to
decay, as vapour is absorbed by many materials, and only very slowly released after the
sources are identified and removed.

Other possible causes of high results for near-zero samples include contamination of
electrolytes (usually hygroscopic Na2O2 or NaOH), sample storage in insecure bottles which
allow evaporation, or memory effects in enrichment cells and gas preparation lines.

A conforming result for sample T3 is not necessarily evidence of lack of contamination if the
measurement uncertainty is high. This may not be of too much concern to some laboratories,
but regular measurements of a tritium-free water as control sample would soon show up any
problem.

A likely cause of negative outliers (counting methods) is the use of tritium-free background
water which has become contaminated.
 15
3.2 Results for Sample T6

This sample can be measured accurately by direct gas or liquid scintillation counting, and was
presumably treated this way by the participants. Although the spread of the submitted results
can be slightly greater due to the various used calibrating standard waters and half-lifes, it
should have been easily possible to achieve conformity with the prepared value. However, as
in previous intercomparisons, a disappointingly high number of laboratories submitted results
which were gross outliers. 82 laboratories provided results with errors attached, with 16 (20%)
identified as outliers. 14 of these applied liquid scintillation counting, and 2 used the 3He
method. 5 results were on the low side (range 341 – 459 TU) and 9 were too high (530 – 631
TU). 2 further outliers are close to the reference value, but indicate an overestimation of
measurement precision by the laboratories concerned.

Several calibrating “standard” waters were reported for these 16 outliers, including NIST
standards in 6 cases. High results can be due to storage of standard waters over time in
insecure bottles (shift to lower concentrations due to net evaporation into an atmosphere of
lower concentration). Low results can be due to a variety of handling or calculation errors, or
perhaps reliance on standard waters obtained at second or third hand without performing
adequate cross-calibration checks.

3.3 Comparison of Reported Uncertainties

To enable laboratories to judge how their reported precision compares with that of other
participating laboratories, we have compiled two distributions of the reported uncertainties,
one for samples T1 to T5 (Fig.7) and one for sample T6 (Fig.8).

Several parameters, which are highly variable among laboratories, have to be considered in
judging these two diagrams:
(a) it is by no means certain that all the laboratories have correctly assessed their
measurement uncertainties by combining all contributing components according to the
Guide on the Uncertainty of Measurements (GUM) [14]; although several reported
values were considered much too low to be acceptable (laboratory 14 at the sensitive end
of all the diagrams is such a case), all have been included in the diagrams;
(b) longer counting times (accumulation times for 3He), lower and more stable background
count rates and better counting sensitivity significantly improve the uncertainty;
(c) several laboratories counted directly the samples without prior tritium enrichment.

The laboratories which counted without prior enrichment for samples T1 – T5, have been
identified in the figures by “-D” attached to the laboratory ID number; this has not been done
for T6, because the sample was presumably counted directly by nearly all participants. In Fig.
7, the cumulative reported uncertainties for five samples are stated. In cases where less than
five samples were measured, the number of samples is indicated in brackets at the laboratory
ID number. As could be expected, most of the direct counting laboratories are near the upper
end of the uncertainty distributions. However, they are grouped among laboratories using
enrichment before counting. This may give cause for thought to the latter, which are applying
much more effort and expense for no obvious gain in precision.
 16

Cumulative uncertainties for tritium ratios of

14
sam ples T1 to T5 as reported by laboratories
Reported tritium ratio uncertainties [TU]

12 T1 uncertainty
T2 uncertainty
T3 uncertainty
10
T4 uncertainty
T5 uncertainty
8

93b-D(4)
49-D(2)

75-D(3)
111(4)

93a(4)
79(4)

71(4)
98(4)

43(4)
4b-D

84-D

4a-D

26-D
58-D

96-D
85-D

73-D
56-D
103

10b
108
10a

31b

31a

101
14

67
15
59
92
25
12
69
66
65
29
88
36
89
94
45
91

76

41

38
95
16
23
21
18
32
42
46
57
77
28
17
24
47
55

68
19
61
51
39
35
48
72
54
22
82

37

33
44
86
5
6

3
7
8
9

Laboratory ID

Fig.7: Cumulative uncertainties for samples T1 to T5 as reported by laboratories in increasing order. Laboratories using direct measurements without enrichment
are marked with the suffix “-D” at the Lab ID. For laboratories with less than 5 analysed samples (biasing the ranking), the respective number is added in brackets
to the Lab ID.
 Stated uncertainty for T6 [TU]

0
5
10
15
20
25
30
14
19
59
7
92
67
15
46
36
22
72
73-D
91
6
25
42
56-D
79
103
8
41
71
88
12
76
94
16
18
29
55
66
26-D
9
17
31a
31b
10a
4b-D
89
93a
95
45

Fig.8: Reported uncertainty for sample T6 as stated by laboratories.

10b
81-D
35
4a-D
65
96-D

Laboratory ID
68
32
39
49-D
57
Reported Uncertainty for Sample T6

24
62
48
108
33
44
101
98
51
3
69
58-D
82
109-D
1
21
47
37
85-D
86
84-D
75-D
43
28
38
99-D
77
93b-D
54
17
 18

3.4 Laboratory Performance Indicators – Accuracy and Precision

An attempt was made to provide a rough estimate of laboratory performance by judging their
results versus the well-determined reference values of the samples as prepared by the IAEA.

3.3.1 Accuracy:

The most obvious performance ranking can be performed by calculating the deviation of the
tritium ratio reported by the laboratory from the reference value in TU. The sum of deviations
for samples T1 to T5 for each laboratory is shown in Fig. 9 (sorted according to absolute sum
of deviations). This provides a good estimate for the accuracy of values reported for an
individual laboratory within the span of covered tritium ratios in TRIC2000.

Several features can be derived from Fig. 9. The sum of deviations from the reference value
ranges from 0.5 TU (lab 94) to 64 TU (lab 93b), not considering lab 106 with a deviation of
more than 300 TU. The majority of laboratories show distinct deviations and bias in positive
or in negative direction for all samples. A balanced distribution of deviations around the
reference values (zero in the plots), – as to be expected for random errors following a normal
distribution – occurs only for a minority of about 30 laboratories. This indicates problems
with the calibration of measurements. Obviously, the ranking of laboratories in Fig. 9 is only
correct for laboratories having measured all five samples T1 to T5. Those labs with less than
five results are consequently marked at the Lab ID with the respective number of analysed
samples indicated in brackets.

Due to the different activity level of sample T6 and the commonly used direct measurement
approach without enrichment and with much higher uncertainties, the deviations from the
reference tritium ratio for sample T6 are presented separately in Fig.10.
 19

Sum of Deviations from Reference Tritium Ratios for T1 to T5

20

T1: (Lab - Reference)

15 T2: (Lab - Reference)
T3: (Lab - Reference)
Deviations from reference values [TU]

T4: (Lab - Reference)

10
T5: (Lab - Reference)

-5

-10

-15

-20
49-D(2)

75-D(3)
111(4)

93a(4)

93b-D
66(4)

79(4)

43(4)
4b-D
4a-D

56-D

58-D
84-D

26-D
73-D

96-D

85-D
10b

103

31b
10a

31a

101

108

106
94
32

95
69
57
23
15

82
35
92
12
76
77
39
21
55
29
47
16
27
41
18
17
28
14
71
59
88
46
19

45
25
89
51
38
36

68
48
67
24
22

33
86
54
91
44
42
65
98
61

37
83
72
9

7
1

5
Laboratory ID

Fig.9: Sum of deviations from the reference tritium ratios for samples T1 to T5 in increasing absolute order in TU.
 Deviation from reference tritium ratio [TU]

-25
-20
-15
-10
-5
0
5
10
15
20
25
109-D
14
48
28
76
86
39
92
62
94
29
72
47
51
15
35
73-D
10a
45
59
33
8
55
21
103
36
22
79
7
4a-D
32
19
4b-D
42
85-D
84-D
66
58-D
68
71
12
37
46
88
10b
57
96-D
17
1
101

Laboratory ID
24
95
9
69
26-D
83
99-D
41

Fig.10: Deviation from the reference tritium ratio for sample T6 in increasing order
93a
75-D
82
43
27
49-D
16
91
81-D
89
56-D
108
18
6
3
44
67
93b-D
38
54
65
Deviation from Reference Tritium Ratio for Sam ple T6

31a
77
31b
25
98
106
20
 21

3.3.2 Precision (Reliability of Precision Statement):

An additional aspect of visualising laboratory performance is the actual precision stated for
performed measurements. This approach focuses on the reported uncertainty associated with
each measurement and on the assessment of its reliability. The absolute deviation for each
sample is again calculated, and then divided by the reported uncertainty. Therefore, the
deviation for each sample is expressed in a sigma scale, directly comparing the reported 1σ-
uncertainty for each sample of a laboratory with the actual deviation from the reference value.
Each laboratory’s “sigma-deviations” for the six samples T1 to T6 are presented in Fig. 11,
ordered by increasing sum of absolute “sigma-deviation” values.

Assuming the absence of any systematic bias for each individual laboratory, all results should
be normal distributed. Under this assumption the respective references value should lie with
95% probability within 2-σ uncertainty ranges around the reported values. As this holds for all
six samples following in each case a normal distribution, the average standard deviation at a
1σ-level (67% probability) for a set of six measurements should be close to one; or expressed
differently, the sum of all absolute standard deviations for six samples should be close to six
sigma in Fig. 11.

As a subjective quality criterion, for six samples reported, the sum of deviations should not
exceed 8σ and it should not fall below 4σ. As it can be seen from Fig. 11, about 50% of all
laboratories exceed significantly the value of 8σ, which clearly is an indicator for an
underestimation of uncertainty in those laboratories (including the possibility of systematic
biases). On the other side, about 25% of all laboratories provided data deviating in sum less
then 4σ from the reference values of samples T1 to T5, which signalises their quite
conservative approach in stating significantly too high uncertainties.

It has to be stressed, however, that the results of Fig. 11 do not tell anything about the actual
suitability of reported data for a certain purpose e.g. for hydrology. This is illustrated in Fig.
12, where a similar plot as in Fig. 11 is presented for samples T1 to T5. In addition to the sum
of absolute “sigma-deviations” for the five samples, the cumulated reported uncertainty for
the five samples is plotted in the lower part of the figure. Obviously, no correlation exists
between the magnitude of uncertainty stated by the laboratories and its reliability as
approximated by the sum of absolute “sigma-deviations”.

Comparing the 12 laboratories with lowest reported cumulative uncertainties for samples T1
to T5 from Fig. 7 with their position in Fig. 12 reveals that 11 out of these 12 laboratories do
considerably overestimate their precision. This statement even holds true when considering a
possible +1.55% positive bias of laboratories in case of using a higher tritium halflife and old
standards. Seven labs show average offsets of more than 5% from the reference tritium ratios.
Five of these 12 laboratories are among the dozen of laboratories with the largest sigma-
deviations from the reference values (lab 5 and 14 deviating both by more than 100σ), all with
average offsets from the reference tritium ratios between 5 to 35%.
This indicates the strong need to further improve the evaluation and uncertainty statements for
tritium measurements. An IAEA advisory group meeting was conducted in 1998 to discuss
uncertainties of nuclear analytical measurements. A contribution discussed in detail the
uncertainty evaluation for tritium at environmental low-level activities using electrolytic
enrichment and liquid scintillation counting [10], providing numerical examples to facilitate
the application of the described principles.
 22

Sum of Deviations (in multiples of reported uncertainty)

20 from Reference Tritium Values for Samples T1-T6
Deviations from reference values (in sigma-units)

T1
15 T2
T3
T4
10 T5
T6
5

-5

-10

-15

-20

93b-D(5)
75-D(4)
49-D(3)

81-D(2)

111(4)

93a(5)
66(5)

23(5)

43(5)

71(5)

79(5)

61(5)

98(5)
4a-D

4b-D

58-D

73-D

96-D

56-D

26-D

85-D

84-D

5(5)
10a

101

103
10b

108

31a

31b
82
35
32
39
94
86
57

95

48
55

76
21
28
51
33
77

17
22
16

69
68
54

47
41
44
15
29
12
46
18
24
92

19
88
38
89
37

42
36
59
45
72

91
25
65
67
14
9

3
7

6
Laboratory ID

Fig.11: Sum of deviations from the true values for samples T1 to T6 in increasing absolute order, here reported as sigma-deviation by dividing the calculated
deviation for each sample by the stated uncertainty. For details see text.
 23

Criterion for Reliability of stated Precision

A) Sum of Deviations from Ref.Value for T1-T5 (mult.of uncert.)
25 B) Stated Uncertainty for Samples T1-T5
20
T1
Deviations from ref.values [sigma-scale]

T2
15
T3 A)
T4
10 T5

-5 Cumulative Uncertainty 20
for T1 to T5

Cumulative Uncertainty [TU]

-10 15
B)
-15 10

-20 5

-25 0

93b-D(4)
49-D(2)

75-D(3)

111(4)
66(4)

43(4)

23(4)

71(4)

79(4)

98(4)
4a-D

56-D

4b-D

58-D

73-D

96-D

26-D

85-D

84-D
101

31b
10a

31a

103

10b

108
82

32
57

35

94

39

95
86

77
55

16

48

76

18
22
21
17

44
28
51

33
41

69
68
46
54
12

19

15
47

29
24
92

88
89
38
37

42
61
36
59

45
72
25

91
65
67
14
9

1
3

5
Laboratory ID

Fig.12: Sum of deviations from the reference value for samples T1 to T5 in increasing absolute order, reported as sigma-deviation by dividing the calculated
deviation for each sample through the stated uncertainty.
 24

3.5 Contrast of Reported Uncertainties in 5th and 6th Tritium Intercomparisons

It is worthwhile to assess whether there has been any indication of a general improvement in
sensitivity and accuracy since the 5th intercomparison in 1994. This is attempted here by
examining the error distributions for the zero samples (A in 1994 and T3 in 2000, Fig.13) and
samples of moderate concentration (D in 1994 and T5 in 2000, Fig.14).

100

90
5th intercomparison
80

70

60

50

40

30

20 Comparison of uncertainty distributions for zero TU

sample in 5th and 6th intercomparisons
10

0
0 1 2 3 4 5

Reported measurement uncertainty (TU)

Fig.13: Uncertainty distributions for tritium-free samples in the 5th and 6th tritium intercomparisons.
The upper line denoting the recent exercise indicates a slightly better performance (as
percentage of all participating laboratories).

100

90 6th intercomp. sample T5

80
5th intercomp. sample D
70

60

50

40

30

20
Comparison of uncertainties between
10 5th and 6th intercomparisons

0
0 5 10 15 20 25

Reported % uncertainty

Fig.14: Uncertainty distributions for samples with approx. 25 TU in the 5th and 6th tritium
intercomparisons. The upper line denoting the recent exercise indicates a slightly better
performance (as percentage of all participating laboratories).
 25

For the zero samples there is little difference between the distributions at uncertainty values
less than 0.25 TU; because more laboratories are involved in 2000, this indicates that more
laboratories are reporting low detection limits. At greater uncertainties, the sensitivities have
improved slightly overall. However, it should be realised that the choice of limit 0.25 TU for
this comparison was arbitrary. It is possible to achieve standard measurement uncertainty
significantly lower than that. Very low detection limits have become necessary for studies
involving very low concentrations, as for example ocean measurements.

Samples D and T5 had slightly different tritium ratios (sample D: 19.2 TU; T5: 26.1 TU). The
comparison (Fig. 14) was therefore made using the percentage uncertainties. The distributions
again reveal some overall improvement in TRIC2000.

This comparison definitely includes a number of incorrectly calculated or reported

uncertainties. In both intercomparisons, a few laboratories grossly underestimated their
measurement uncertainty, while some others did not include all uncertainty components.
Therefore, the indications are of a very general nature.

3.6 Reporting of Results

Methods of reporting results are often neither consistent nor appropriate. Many results in
TRIC2000 were reported with different numbers of decimal points for result and error. Others
had too few decimal points, and some too many. Details of standardisation were unclear in
many cases, which has caused problems in looking for factors influencing the data
individually and overall. The following ideas may help to improve this situation for the future.

Results should always be expressed in an appropriate number of decimal places, with the
same number applying to result and error. The appropriate number can be judged by the
relative magnitudes of the reported value and measurement error, and the range of tritium
concentrations pertaining to samples handled by the laboratory. As an example, consider a
laboratory operating normally now within a range lower than a tritium ratio of 20 TU, and
able to achieve 3% standard measurement uncertainty for all but near-zero results. A result of
10 TU would bear an uncertainty of 0.30 TU; one of 5 TU would have an uncertainty of 0.15
TU. In this case, it is important to report results and uncertainties to 2 decimal places.
Moreover, for results less than 1 TU, where the % error increases sharply, reporting to 3
decimal places becomes appropriate. If samples of order 100 TU are measured reporting to 2
decimal places is no longer of any advantage, 1 being sufficient. (Note: Due to the very small
uncertainties associated to the reference values in TRIC2000, those data are reported with
more decimal places than recommended for routine samples).
Similar considerations should be applied in determining the appropriate number of decimal
places for results reported as massic activities (e.g. Bq.kg-1).

The sample date to which a result refers, should always be reported.

All laboratories should be able to clarify the basis of their calibrations, and even include this
as a footnote to results in published papers or reports to clients. Many laboratories reported
4500 days as their applied half-life. This value was stipulated for decay correction of the
TRIC2000 results to 1 July 2000, but it is unlikely that that very recently determined and
published half-life was the basis of any calibration procedures. Only few laboratories stated to
 26

use already the “new generation” of NIST standards with calibration date 3 Sep 1998. The
necessary fundamental information comprises the identity and certified calibration (with date)
of the tritiated water standard which is the cornerstone of the calibration procedures, and the
half-life applied to estimate its decay over the years. A number of laboratories reported to use
“Packard” or “Wallac” tritium standards; these were probably sealed control cocktails
monitoring counter stability during routine liquid scintillation counting, which are not relevant
for calibration purposes. Tritium calibration involves counting cocktails with daughter water
standards of appropriate count rates and background waters, both having identical
composition to the unknown sample cocktails.

3.7 Suggestions for Improved Sensitivity

Given the need for optimum sensitivity, it is useful to examine whether any of the results and
technique details provided by the laboratories indicate that they are individually getting the
best value from equipment available to them. For liquid scintillation counting, the background
count rate and its stability are very important factors. A more stable background count rate
will reduce the uncertainty for the same counting time, or allow samples to be counted for
shorter times, enabling greater sample throughput. Background is not only counter-dependent,
but it also depends on the material of the sample vials and the cocktail constituents. From the
available data provided by the participants, the only two counter groups giving assessable
background information were those laboratories using Wallac type Quantulus and Packard
type 2770/3170TR counters. These counters are well protected against cosmic ray variations
(including altitude dependence) and against other factors influencing many other counters. 26
respectively 12 laboratories reported using these counter types mentioned above, with
backgrounds ranging from 0.44 to 1.38 cpm (Quantulus) and 0.89 to 1.90 cpm
(2770/3170TR). The lower values give a huge advantage, while the upper values seem to be
perhaps unnecessarily high, indicating not-optimised settings. Slightly higher backgrounds can
be expected for other instrument types, or those without all the electronic options aiding
background reduction, but spectral analysis capability allows all counters of these types to be
tuned carefully to give the best compromise between detection sensitivity of tritium decay
events and background count rate. Spectral analysis capability also enables the selection of
counting vials and cocktails with lowest background from available commercial options, and
determination of the most advantageous water/cocktail proportions. Low-level tritium samples
should now be counted in polyethylene or teflon vials (i.e. not glass, which has a higher
background), using a cocktail type which does not diffuse through the vial wall; teflon vials
are expensive, and there are several types of cheaper polyethylene vial which are equally
effective.

AUTHORS

Manfred Gröning, Randolf Auer and Herbert Tatzber are staff members of the IAEA Isotope
Hydrology Laboratory. Claude Taylor is now a groundwater and laboratory consultant
(address details: Dr. C. B. Taylor, 28 Wyndrum Avenue, Lower Hutt 6009, New Zealand, tel.
64 4 566 3571, fax. 64 4 586 2124, E-mail: sandbagger5@xtra.co.nz). Gisela Winckler is a
former staff member of the Isotope Hydrology Laboratory, now at LDEO, Columbia
University, New York.
 27

REFERENCES

[1] Cameron, J.F., Survey of systems for concentration and low background counting of
tritium in water. In: Radioactive Dating and Methods of Low-level Counting. IAEA,
Vienna, 1967, 543-573.

[2] Florkowski, T., Payne, B.R., Sauzay, G., Interlaboratory comparison of analysis of tritium
in natural waters. Int. J. Appl. Rad. Isot., 1970, 21, 453-458.

[3] Taylor, C.B., Interlaboratory comparison of low-level tritium measurements in water. Int.
J. Appl. Rad. Isot., 1978, 29: 39-48.

[4] Hut, G., Low-level tritum counting facilities: a survey. Nucl. Inst. Meth., 1986, B17: 490-
492.

[5] Hut, G., Intercomparison of low-level tritium measurements in Water. IAEA Internal
Report, 1987, Vienna.

[6] Östlund, G., Scott, E.M., Taylor, C.B., Fifth IAEA Intercomparison of low-level tritium
measurements in water, IAEA Internal Report, Vienna, 1995.

[7] Taylor, C.B., Roether, W., A uniform scale for reporting low-level tritium measurements
in water. Int. J. Appl. Rad. Isot., 1982, 33: 377-382.

[8] Unterweger, M.P., Lucas, L.L., Calibration of the National Institute of Standards and
Technology tritiated-water standards. Appl. Radiat. Isot., 52: (2000) 527-531.

[9] Lucas, L.L., Massic activity ratios of the NBS/NIST tritiated-water standards issued
between 1954 and 1999. J. Res. Nat. Inst. Stand. Technol. 105: (2000) 535-539.

[10] Lucas, L.L., Unterweger, M.P., Comprehensive review and critical evaluation of the half-
life of tritium. Ibid. 105: (2000) 541-549.

[11] Rozanski, K., Gröning, M., Quantifying uncertainties of tritium assay in water samples
using electrolytic enrichment and liquid scintillation spectrometry. In: Quantifying
Uncertainty in Nuclear Analytical Measurements. Proceedings of an IAEA Advisory Group
Meeting, 11-14 May 1998, IAEA (IAEA-TECDOC in preparation)

[12] Mohr, P., Taylor, B.N., CODATA recommended values of the fundamental physical
constants: 1998, Review of Modern Physics, Vol.72 No.2 (2000), 351-495. See also in:
http://www.codat.org

[13] Mann, W.B., Unterweger, M.P., Coursey, B.M., Comments on the NBS Tritiated-
Water Standards and Their Use. Int. J. Appl. Rad. Isot., 1982, 33: 383-386.

[14] ISO Guide to the Expression of Uncertainty in Measurement, International

Organisation for Standardisation, Geneva, Switzerland, First edition, corrected and reprinted
1995.
 28

APPENDIX A. DETAILS OF PROCEDURES TO PREPARE SAMPLES

Because SRM4927F is an extremely active tritiated water, the first two dilution stages were
not performed in the low-level Isotope Hydrology Laboratory, but at the IAEA Laboratories at
Seibersdorf. A de-ionised “tritium-free background water” was used for these 3 stages. This
water dates from 1992 and was used for the previous tritium interlaboratory comparison; it is
from the same Grafendorf well which was re-sampled to provide the water for the final
dilution stage (sample T3). Repeat analyses of the de-ionised water indicated that its tritium
ratio is less than 0.1 TU. Its massic activity is less than 5×10-7 that of the resulting daughter
water after 3 dilution stages, and its contribution to the prepared values can therefore be
neglected.

Table 4: Dilution procedure for the tritiated water standard SRM4927F

Dilution stage 1 2 3
Weight of tritiated water (in 4.9063 (0.0003) 20.9142 (0.0003) 83.585 (0.003)
brackets uncertainty) [g]
Weight of added T3 dilution 994.38 (0.07) 999.89 (0.07) 8453.95 (0.05)
water (uncertainty) [g]
Dilution factor (relative 203.674 48.8091 (0.007%) 102.142 (0.004%)
uncertainty) (0.009%)
Combined dilution factor 203.674 9941.12 (0.012%) 1015409 (0.012%)
(relative uncertainty) (0.009%)

A sufficient supply of diluting water for the final dilution stage was sampled from the
Grafendorf well. This is sample T3. The possible tritium concentration range of this water was
judged to be between 0 - 0.072 TU, using the submitted results.

The diluting water T3 was not distilled or de-ionised, in order to ensure that the samples
would be put through the routine pre-measurement distillation step by the participating
laboratories. In the final dilution stage, the measured weights of diluting water were corrected
for the dissolved solid content; this correction is 218 mg/kg water, i.e. –0.0218%.

The dilution stages of the daughter water and the calculated TU values are summarised in the
accompanying table. Numbers in brackets are 1 standard uncertainty. The factor 8.390 was
applied to convert Bq/kg to TU.
 29

Table 5: Dilution procedure for the six samples T1 to T6 used in this interlaboratory comparison
exercise, using as tritiated standard the ‘Dilution Stage 3’ of Table 4. Values in brackets are
uncertainties at the 1-σ level.

Step T1 T2 T4 T5 T6

Weight of SRM4927F 166.52 33.451 82.400 412.19 1806.37

daughter [g] (0.03) (0.003) (0.003) (0.03) (0.03)
Weight of T3 water [g] 77782.9 73541.7 74173.3 74286.9 15096.3
(0.2) (0.2) (0.2) (0.2) (0.2)
Correction of above T3 77765.9 73525.7 74157.1 74270.7 15093.0
weight for TDS [g] (0.2) (0.2) (0.2) (0.2) (0.2)
Dilution factor 468.006 2199.01 900.965 181.186 9.35543
(0.018%) (0.009%) (0.004%) (0.007%) (0.004%)
Total dilution factor [×10- 4.75218 22.3289 9.14848 1.83978 0.0949959
8
] relative to SRM4927F (0.022%) (0.015%) (0.013%) (0.014%) (0.013%)
Massic tritium activity at 1 1.20519 0.25650 0.62604 3.11303 60.2899
July 2000 [Bq/kg] (0.363%) (0.363%) (0.363%) (0.363%) (0.363%)
Tritium ratio at 1 July 2000 10.112 2.152 5.252 26.118 505.83
[TU] (0.037) (0.008) (0.019) (0.095) (1.84)
Possible range of added (0 - 0.072) (0 - 0.072) (0 - 0.072) (0 - 0.072) (0 - 0.072)
tritium [TU] due to diluting
water T3 (additional
uncertainty component)
Combined uncertainty 10.038 - 2.136 - 5.214 - 25.928 - 502.15 -
range of tritium ratio at 1 10.258 2.240 5.362 26.380 509.58
July 2000 [TU] at 2-σ level
 30

APPENDIX B: ADJUSTMENT OF RESULTS TO THE PREPARED VALUE RANGE

The previously recommended tritium scale [7] was based on two primary parameters. The first
was the certified massic activity (Bq/kg) of the primary tritiated water standard on 3
September 1978. This standard water was suitable for dilution by environmental laboratories
to provide daughter standard waters for calibration of routine measurements (for enrichment,
counting and 3He mass spectrometry). The second adopted parameter was the former tritium
half-life 4540 days. These were fixed parameters, i.e. defined to be error-free. Their actual
uncertainties (for example for massic activity as quoted by NIST on accompanying standard
certificates) were not relevant for intercomparability with data obtained using other standards
and half-life values. In fact, it was not possible till now to judge the uncertainty of any half-
life value, making the fixed parameter scale the only option. The new NIST half-life (4500 ± 8
days) is the first to provide such a confident uncertainty estimate. Absolute measurements (not
based on fixed parameter scale) must take into account reliably determined uncertainties in
standard calibration and half-life value.

The two primary parameters defined the old scale for massic activity measurement, but a
conversion factor had to be applied to yield TU. This factor was deduced from the molar
weight assumed for water, the Avogadro constant (determining the number of hydrogen atoms
in 1 kg of water) and the assumed half-life, yielding 1 Bq/kg = TU = 8.464 (old scale!).

The new conversion factor 8.390 is based on the newly determined half-life (4500 days), the
present value of the Avogadro constant and the molecular weight of water.

Laboratories can adjust their intercomparison results to the prepared values using the
following conversion equation.

TR(adjusted) n1s as2(t02) exp(-λ2(tc - t02))

---------------- = ---- × --------- × -------------------
TR(original) n2s as1(t01) exp(-λ 1(tc - t01))

Here the quantities as1(t01), as2(t02) are respectively the massic activities of the laboratory’s
standard water and SRM4927F at their calibration dates t01, t02.. λ 1, λ 2 are the old, new
radioactive decay constants, and tc is the intercomparison reference date. n1s is the count rate
of the daughter standard used to calibrate the original measurement. n2s is the count rate which
would have been recorded by an equally-diluted daughter of the SRM4927F. It must be
emphasised that data adjustment to the prepared values must involve an accurately
measured ratio n1s/n2s. Luckily ratios are available for all old and newer NIST standards [9].
It would be extremely useful if other used tritiated water standards could be compared against
NIST tritiated standard waters.

Very few laboratories would have used any of the very recent NIST standards or the new half-
life to calibrate their measurements for TRIC2000 (only seven laboratories indicated having
used a NIST standard of the new series: LabIDs 1,12,29,31a/b,57,75,103). Many reported
values using the SRM4926C/half-life 4540 days fixed parameter scale. Adjustment to the
values calculated for the intercomparison samples involves incorporation of the NIST-
determined massic activity ratio [9]
 31

SRM4927F/SRM4926C = 576 758 ± 0.066%

into the above equation. Converted values should have this uncertainty component added in
quadrature to the original % measurement error. In nearly all cases, this will not change the
already higher uncertainty, but corrections from standard waters not originated from NIST
may involve larger uncertainties in the ratio n1s/n2s. Laboratories which used other standard
waters and/or half-lives can evaluate and apply the appropriate factors in the above conversion
equation.

APPENDIX C: LISTING OF RESULTS FOR THE SIX SAMPLES

In Table 6, all submitted results from the 86 laboratories participating in TRIC2000 are listed
and sorted according to the assigned laboratory ID number. The results are presented with as
many digits as originally reported (only in one case, a fourth digit was omitted). Necessary
alterations to the raw data are marked under the column ‘Remarks’. The gray-shaded fields are
data which were regarded as outliers. Three criteria for outlier-detection were applied:
1. all data without stated uncertainty could not be evaluated and were regarded as outliers
irrespective of the reported value (see the part requested information in the TRIC2000
invitation); 2. all data strongly deviating from the remaining data set were regarded as outliers
(interquartile criterion); 3. all data deviating from the stated limits around the reference value
by more than twice their reported uncertainty.
In Table 7 a list of correction data is reported which were received after the reporting deadline
and after an indication of possible problems in the sample analysis, was sent to 26
laboratories. For this reason, those data were not included in the original data set.
 32

Table 6: All sample results as reported by the individual participating laboratories. The uncertainties u(sample) were requested to be stated at the 1-σ level; only
obvious changes were applied (see remarks). Blank fields indicate not-submitted data. The three measurement methods are indicated: LSC – Liquid Scintillation
Counting; Gas – Gas Proportional Counting; 3He – Mass Spectrometric 3He analysis. All outlier values in the tables are gray-shaded, for details on the outlier
determination procedure see text. The TRIC2000 reference values are stated in the header lines.

Lab T1 u(T1) T2 u(T2) T3 u(T3) T4 u(T4) T5 u(T5) T6 u(T6) Meas. Remarks

Method
Ref. 10.112 2.152 0 5.252 26.118 505.83 Reference value for TRIC2000
All tritium values and uncertainties are stated as tritium ratio in [TU]

1 10.78 0.52 2.62 0.22 -0.07 0.15 5.88 0.33 28.35 1.21 521 24 LSC
3 9.1 0.5 1.6 0.2 -0.1 0.2 5.1 0.3 24.4 0.9 457.7 18.1 LSC See also revised data (T6)
4a-D 10.5 0.8 1.90 0.68 -0.10 0.70 5.12 0.69 25.8 0.9 514.8 12.6 LSC-D
4b-D 10.03 0.42 2.28 0.32 0.07 0.32 5.18 0.37 25.3 0.7 495.8 10.8 Gas-D
5 6.367 0.045 1.370 0.019 0.010 0.004 3.301 0.028 16.352 0.094 3He Values+uncertainties adjusted for ref.date
6 10.931 0.113 2.293 0.027 0.004 0.010 5.424 0.055 26.654 0.270 553.4 5.5 3He See also revised data (T6)
7 9.17 0.34 2.46 0.25 0.04 0.19 5.71 0.32 28.39 1.01 496.88 3.06 LSC
8 10.9 0.5 2.31 0.16 0.15 0.12 5.43 0.27 26.3 1.1 498 7.0 LSC
9 10.1 0.4 2.3 0.3 0.0 0.3 5.3 0.3 26.8 0.9 522 10 LSC
10a 10.4 0.43 2.38 0.26 -0.02 0.22 5.58 0.33 26.75 0.8 510.8 10.6 LSC
10b 10.05 0.42 2.28 0.11 0.21 0.08 5.34 0.23 25.71 1.03 519.04 11.59 Gas
12 9.76 0.20 2.07 0.09 -0.02 0.06 5.48 0.14 25.0 0.40 493 8 Gas
14 10.70 0.028 2.34 0.024 0.56 0.024 5.81 0.027 27.60 0.033 505.40 0.145 LSC Probably relative uncertainty stated?
15 10.46 0.12 2.23 0.04 0.003 0.018 5.29 0.07 26.75 0.27 501.4 4.4 LSC
16 10.4 0.6 2.3 0.3 -0.1 0.2 5.6 0.4 27.8 1.3 528.8 8.3 LSC
17 10.7 0.7 2.0 0.5 1.1 0.5 5.3 0.6 27.5 1.2 520 10 LSC
18 11.07 0.54 2.36 0.42 -0.20 0.42 5.20 0.60 27.78 1.08 459.00 8.44 LSC
19 10.6 1.0 3.4 1.0 0.9 0.2 6.5 0.8 26.2 1.3 496 2 LSC
21 10.10 0.57 2.97 0.27 0.11 0.30 5.16 0.73 27.16 1.07 514.01 24.87 LSC
22 10.7 1.4 1.6 1.4 0.0 1.4 4.7 1.5 32.0 1.8 497.1 5 LSC See also revised data (T5)
23 9.8 0.7 2.3 0.4 0.25 0.1 5.2 0.3 25.8 1.3 3He
24 11.84 0.75 3.37 0.56 0.75 0.52 6.15 0.67 28.24 1.06 521.6 15.3 LSC
25 11.13 0.19 1.35 0.13 -0.54 0.11 6.09 0.15 27.24 0.24 631.23 5.88 LSC
26-D 9.46 1.38 4.88 1.35 -0.09 1.15 7.17 1.18 28.71 1.32 522.88 9.99 Gas-D
27 10 3.0 -0.17 4.2 25.6 528 LSC
28 9.38 0.69 1.84 0.24 -0.12 0.21 5.76 0.43 27.72 1.91 506.76 34.75 LSC
 33

Lab T1 u(T1) T2 u(T2) T3 u(T3) T4 u(T4) T5 u(T5) T6 u(T6) Meas. Remarks

Method
Ref. 10.112 2.152 0 5.252 26.118 505.83 Reference value for TRIC2000
All tritium values and uncertainties are stated as tritium ratio in [TU]

29 11.00 0.32 2.95 0.24 -0.14 0.22 4.93 0.24 26.36 0.44 508.4 8.6 LSC
31a 10.3 0.4 2.8 0.4 0.3 0.4 5.3 0.4 25.6 0.9 613 10 LSC
31b 10.1 0.5 2.45 0.3 0.15 0.3 5.9 0.4 26.3 0.7 628 10 LSC
32 10.4 0.6 2.1 0.4 -0.1 0.4 5.5 0.5 26.0 1.2 515 15 LSC
33 10.1 3.1 2.2 0.7 3.0 0.8 9.9 4.4 25.6 2.8 512 17 LSC See also revised data (T4)
35 9.9 1.2 3.0 0.8 0.0 0.5 5.4 1.0 26.6 1.7 510.4 12.3 LSC
36 10.85 0.32 2.73 0.15 1.07 0.15 6.45 0.23 27.35 0.67 514.39 4.85 LSC
37 16.5 2.5 4.85 0.75 3.1 0.6 10.1 1.5 33 5 519 26 LSC See also revised data set (T1-T6)
38 10.88 0.59 2.54 0.18 0.11 0.02 5.85 0.34 28.91 1.46 575.9 36.1 LSC See also revised data set (T5,T6)
39 10.5 1 2.2 1 0.2 1 6.0 1 25.5 1 504 15 LSC
41 11.2 0.5 2.6 0.4 0.2 0.4 5.3 0.4 27.4 0.7 524 7 LSC
42 12.99 0.64 2.95 0.53 1.39 0.51 7.9 0.59 30.8 0.83 516.4 6 LSC
43 11 1 2.1 0.8 5 1 24 1 484 30 LSC Reported: T3<0.8
44 12.7 5.8 5.5 1.8 2.9 2.5 7.4 2.0 25.1 3.0 445 17 LSC See also revised data set (T1-T6)
45 11.4 0.5 1.8 0.1 0.5 0.04 4.4 0.2 25.0 1.0 500.7 11.2 LSC
46 10.3 0.72 3.95 0.51 0.235 0.352 6.05 0.59 27.0 1.1 519 4.4 LSC Reported relative uncertainty converted
47 9.5 0.7 2.5 0.3 1.1 0.2 5.5 0.5 26.0 2.0 510 25 LSC
48 12.1 1.4 2.3 0.5 0.1 0.4 5.6 0.7 29.9 2.9 505.4 15.9 LSC
49-D 10.2 1.3 21.6 2.3 483 15 LSC-D
51 9.5 0.8 3.0 0.6 -0.5 0.5 4.7 0.7 28.2 2.0 510 18 Gas
54 9.6 1.5 3.2 0.9 1.6 0.5 5.6 0.9 20 3.5 430 85 LSC
55 8.85 0.75 2.45 0.65 0.03 0.60 5.80 0.70 26.00 1.00 497.90 9.25 LSC
56-D 10 2 1 2 -1 3 5 2 28 3 536 6 LSC-D See also revised data (T6)
57 10.19 0.7 2.15 0.4 0.16 0.2 5.15 0.5 25.54 1.5 492 15 LSC
58-D 12.4 1.4 2.0 1.3 0.3 1.3 6.5 1.3 27.2 1.9 518 21 LSC-D
59 9.00 0.14 1.99 0.10 0.00 0.08 5.19 0.11 23.83 0.20 511.52 2.13 Gas
61 13.5 0.8 1.72 0.94 -0.80 0.80 7.67 0.82 37.0 1.0 LSC
62 507.9 15.6 LSC
65 7.82 0.16 1.15 0.12 -0.40 0.15 3.82 0.26 17.01 0.49 400.26 13.33 LSC See also revised data set (T1,T5,T6)
66 9.85 0.24 0.00 0.05 5.29 0.19 25.73 0.55 517.6 9.4 3He
67 11.34 0.14 2.66 0.05 0.075 0.032 6.04 0.06 29.93 0.21 571.2 4.3 3He See also revised data set (T5,T6)
 34

Lab T1 u(T1) T2 u(T2) T3 u(T3) T4 u(T4) T5 u(T5) T6 u(T6) Meas. Remarks

Method
Ref. 10.112 2.152 0 5.252 26.118 505.83 Reference value for TRIC2000
All tritium values and uncertainties are stated as tritium ratio in [TU]

68 8.2 0.8 2.5 0.8 -0.2 0.7 5.9 0.8 23.7 0.9 493.5 14.8 LSC
69 9.5 0.2 2.1 0.1 0.01 0.01 5.1 0.1 26.2 0.5 489 19 3He
71 11.1 0.7 2.9 0.3 0.1 5.6 0.4 24.8 1.0 518.2 7.0 Gas
72 17.75 1.34 7.26 0.93 2.06 0.89 17.92 1.16 29.44 2.38 501.75 5.13 LSC
73-D 9.3 2.2 0.6 2.1 -1.6 2.1 7.63 2.1 27.9 2.2 510.43 5.2 LSC-D
75-D 8.9 3.1 6.1 2.8 29 3.5 526 29 LSC-D See also revised data (T2)
76 11.1 0.5 1.95 0.32 0.13 0.20 5.3 0.45 25.5 0.7 507 8.0 LSC
77 9.8 0.7 2.3 0.6 -0.4 0.5 5.3 0.7 27.2 0.9 615 40 LSC
79 9.54 0.16 2.04 0.11 4.92 0.08 24.80 0.54 496.96 6.51 3He Reported : T3<0.4
81-D 37.8 2.5 530.4 12.0 LSC-D
82 9 2 2 1 0 1 5 2 26 3 526 21 LSC
83 14.8 11.9 5.0 6.8 30.0 523 LSC
84-D 9.8 0.78 3.4 0.31 3.2 0.29 6.0 0.45 26.7 1.60 517 28.4 LSC-D Reported relative uncertainty converted
85-D 4.70 1.85 -1.93 1.51 -3.86 2.35 1.93 1.68 22.6 2.1 516.8 26.5 LSC-D Reported values in Bq/kg converted to TU
86 12.391 4.044 4.038 3.858 1.612 3.989 4.960 4.014 28.728 4.11 507.150 27.33 LSC
88 10.7 0.3 2.5 0.2 0.7 0.2 5.0 0.3 24.2 0.5 519.0 7.8 LSC See also revised data set (T1,T4)
89 11.3 0.4 2.3 0.3 1.1 0.3 5.8 0.3 27.6 0.5 535 11 LSC
91 10.2 0.5 5.9 0.3 0.8 0.3 0.1 0.3 25.7 0.5 530.0 5.2 LSC
92 10.6 0.2 2.3 0.1 0.0 0.1 5.6 0.1 26.9 0.3 507.8 3.6 LSC
93a 3.02 0.61 0.57 0.58 5.09 0.61 7.86 0.72 486 11 LSC
93b-D 16.9 1.7 28.0 2.7 0 34.7 3.2 28.4 2.8 573.9 44.2 LSC-D Reported relative uncertainty converted
94 10.2 0.4 2.1 0.2 0.0 0.1 5.5 0.2 26.3 0.9 503.7 8.2 LSC Reported negative uncertainty (T3) corrected
95 10 0.6 1.7 0.4 0.1 0.2 5.2 0.5 26.3 1 490 11 LSC
96-D 12.2 1.7 3.2 1.8 -0.1 1.8 0.6 1.8 24.9 2.1 492 14 LSC-D
98 6.4 0.7 0.8 0.3 0.1 0.0 3.1 0.5 19.2 0.9 341 17.2 LSC
99-D 523.30 38.20 LSC-D Reported relative uncertainty converted
101 10.3 0.64 2.44 0.39 -0.32 0.35 5.54 0.45 27.0 1.3 521 17 ?? Average value of two data sets calculated
103 10.18 0.31 2.57 0.24 -0.15 0.15 5.02 0.25 26.32 0.44 497.4 6.8 LSC
106 110.69 2.442 10.581 23.60 229.11 244.98 LSC
108 9.6 0.37 2.2 0.24 0.21 0.20 5.6 0.28 30 0.9 550 16 LSC
109-D 505.91 23.27 LSC-D Reported relative uncertainty converted
 35

Lab T1 u(T1) T2 u(T2) T3 u(T3) T4 u(T4) T5 u(T5) T6 u(T6) Meas. Remarks

Method
Ref. 10.112 2.152 0 5.252 26.118 505.83 Reference value for TRIC2000
All tritium values and uncertainties are stated as tritium ratio in [TU]

111 9.35 0.20 2.06 0.07 3.81 0.10 24.68 0.62 3He Reported: T3<0.05; relative uncertainties
converted

Table 7: Corrections to the original submitted data of Table 6. The data were not used for any compilation, as they were received after the deadline passed and
nearly all only after a letter was sent to the laboratories to alert them on probable problems with their results. The TRIC2000 reference values are stated in the
header lines.

Lab T1 u(T1) T2 u(T2) T3 u(T3) T4 u(T4) T5 u(T5) T6 u(T6) Meas. Remarks

Method
Ref. 10.112 2.152 0 5.252 26.118 505.83 Reference value for TRIC2000
All tritium values and uncertainties are stated as tritium ratio in [TU]

3 501.3 18.5 LSC Correction after indication of possible problem

6 507.0 5.0 3He Correction after indication of possible problem
22 29.0 1.8 LSC Correction after indication of possible problem
33 7.3 LSC Correction after indication of possible problem
37 12.4 2.6 0.7 1.1 -1 1 6.0 1.7 28.5 5 519 26 LSC Correction after indication of possible problem
38 21.611.09 537.98 21.52 LSC Correction after indication of possible problem
44 3.59 3.4 9.52 2.3 6.80 2.5 13.2 3.8 25.1 3.0 328 30 LSC Correction after indication of possible problem
56-D 513 5 LSC-D Correction after indication of possible problem
65 8.536 0.188 17.667 2.044484.81 9.644 LSC Correction after indication of possible problem
67 25.84 0.58 538.7 9 3He Correction after indication of possible problem
75 <6.7 LSC-D Correction after indication of possible problem
88 10.3 0.3 24.8 0.9 LSC Re-analysis confirming former result
 36

APPENDIX D: LIST OF PARTICIPANTS

TRIC2000 - Participants

Mr. Héctor O. PANARELLO Mr. David HILL

Instituto de Geocronología y Geología Environment Division, ANSTO
Isotópica (INGEIS) PMB 1, Menai NSW
Pabellon INGEIS - Ciudad Universitaria AUSTRALIA 2234
1428 Buenos Aires Tel.: +61-2-9717-3539
ARGENTINA Fax: +61-2-9717-9286
Tel: +54-11-47833022/23 E-mail: David.Hill@ansto.gov.au
Fax: +54-11-47833024
E-mail: hector@ingeis.uba.ar
hector@mail.retina.ar

Ms. E. HOFFMANN Mr. Viktor KARG

Radiochemistry, Building 34 Labor für Strahlenschutz
ANSTO Wien des BMLFUW
Private Mail Bag 1 Berggasse 11
Menai, 2234, NSW A-1090 Wien
AUSTRALIA AUSTRIA
E-mail: emh@ansto.gov.au Tel.: +43.1.317 54 18;
Fax: +43.1.317 03 99
E-mail: vkarg@attglobal.net

Ms. Ulrike PRUEFERT Mr. Roland TESCH

Magistratsabteilung 15-Gesundheitswesen Mr. Vincent RAJNER
Institut für Umweltmedizin Arsenal Research
Fachbereich Umweltphysik Cluster Umwelt und Leben
Feldgasse 9 Geschäftsfeld Umweltforschung, Objekt 214,
1080 Vienna Faradaygasse 3
AUSTRIA A-1030 Wien
Tel.: +43-1-40413-87846 AUSTRIA
Fax: +43-1-40413-99-87800 Tel.: +43/1/797 47 - 516
E-mail: pru@M15.magwien.gv.at Fax: +43/1/797 47 - 587
E-mail: Tesch.R@arsenal.ac.at
E-mail: Rajner.V@arsenal.ac.at
 37

Mr. Herbert TATZBER Mr. Zildete ROCHA

Ms. Daniela BRUMMER Mr. Jefferson Vianna BANDEIRA
Isotope Hydrology Laboratory Brasilian Nuclear Energy Commission (CNEN)
International Atomic Energy Agency Nuclear Technology Development Centre
P.O.Box 100 (CDTN)
A-1210 Vienna Cidade Universitária – Pampulha
AUSTRIA P.O. Box 941
Tel: +43-1-2600-21761 30123-970 Belo Horizonte, MG
Fax: +43-1-2600-21763 BRAZIL
Email: H.Tatzber@iaea.org Tel: +55 31 499-3120
D.Brummer@iaea.org Fax: +55 31 499-3127
E-mail: jvb@urano.cdtn.br
rochaz@urano.cdtn.br

Mr. Robert J. DRIMMIE Mr. Wayne WORKMAN

Environmental Isotope Laboratory AECL
C/O Chemistry Stores Chalk River Laboratories
University of Waterloo Chalk River, Ontario, K0J 1J0
200 University Avenue CANADA
Waterloo, ON N2L 3G1 Fax: +(613) 584-9530
CANADA E-mail: workmanw@aecl.ca
Tel.: +519 8884567 x.2580
Fax: +519 7460183
E-mail: rdrimmie@uwaterloo.ca

Mr. Jiashen CHEN Mr. Yang HUAIYUAN

Shenzhen Health and Antiepidemics Station Low background radioactivity lab
Quandong China Institute for Radiation Protection
CHINA Box 120,Taiyuan, Shanxi
CHINA
E-mail: yanghaiyuan@163.net

Ms. Nada HORVATINCIC Mr. Sven P. NIELSEN

Radiocarbon and Tritium Laboratory Radioecology Programme
Rudjer Boskovic Institute Nuclear Safety Research Department
P.O. Box 180 Risø National Laboratory
10002 Zagreb DK-4000 Roskilde
CROATIA DENMARK
Tel.: +385-1-4560943 Tel.: +45 4677 4172
Fax: +385-1-4680239 Fax: +45 4677 4193
E-mail: nada.horvatincic@irb.hr
 38

Mr. Aly Islam METWALLY ALY Mr. Martial DRAY

National Centre for Nuclear Safety and Centre de Recherches Géodynamiques
Radiation Control Université Pierre et Marie Curie
Atomic Energy Authority 74200 Thonon-les-Bains
3, Ahmed Elzomor Street FRANCE
Nasr City 11762, Elzoohor District Tel.: +00 33 4 50 81 99 64
P.O. Box 7551 Fax: +00 33 4 50 26 68 34
Cairo E-mail: dray@biogeodis.jussieu.fr
EGYPT
E-mail: alyi45@yahoo.com

Ms. Christine FLEHOC Mr. Marc FOURNIER

BRGM Institut de Protection et de Sureté Nucléaire
ANA / ISO Laboratoire de Mesure de la Radioactivité dans
3, avenue Claude Guillemein l'Environnement
BP 6009 Bat.501, Bois des Rames
45060 Orleans cedex 02 F-91400 Orsay cedex
FRANCE FRANCE
Tel.: +02 38 64 34 13 Fax : +(33)169855837
Fax: +02 38 64 37 11 E-mail : marc.fournier@ipsn.fr
E-mail: c.flehoc@brgm.fr

Elise FOURRE, P. JEAN-BAPTISTE Mr. Fabien POINTURIER

Laboratoire des Sciences du Climat et de CEA/DAM/DIF/DASE/SRCE
l'Environnement PO Box 12
Orme des Merisiers bât 709 91680 Bruyeres-Le-Chatel
91 191 Gif-sur-Yvette Cedex FRANCE
FRANCE E-mail: pointuri@dase.bruyeres.cea.fr
Tel.: +33 1 69 08 41 72 Baglan@dase.bruyeres.cea.fr
Fax: +33 1 69 08 77 16
E-mail: elise.fourre@cea.fr

Mr. L. EICHINGER Mr. D. HEBERT

Hydroisotop GmbH TU Bergakademie Freiberg
Woelkestr. 9 Institut für Angewandte Physik
D-85301 Schweitenkirchen D - 09596 Freiberg
GERMANY Bernhard-von-Cotta-Strasse 4
Tel.: +49 8444 92890 GERMANY
Fax: +49 8444 928 929 Tel.: +0049-3731-39-2731
E-mail: Hydroisotop@t-online.de Fax: +0049-3731-39-4004
E-mail: hebert@tu-freiberg.de
 39

Mr. Walter KRAUSE Olaf KLATT, Monika RHEIN

Federal Institute of Hydrology Universität Bremen, FB1
(Bundesanstalt für Gewässerkunde) Institut für Umweltphysik
Referat G4 Abt: Ozeanographie
Kaiserin-Augusta-Anlagen 15-17 Kufsteinerstr.
D-56068 Koblenz Gebäude NW1; Raum S0400
GERMANY 28335 Bremen
E-mail: walter.krause@bafg.de GERMANY
Tel.: +49-421-2184562
Fax: +49-421-2187018
E-mail: oklatt@physik.uni-bremen.de
mrhein@physik.uni-bremen.de

Mr. Willi STICHLER Mr. Axel SUCKOW

Institut fuer Hydrologie Institut für Geowissenschaftliche
GSF-Forschungszentrum München Gemeinschaftsaufgaben (GGA)
Ingolstädter Landstrasse 1 Joint Geoscientific Research Institute (GGA)
85764 Neuherberg Geochronology and Isotope Hydrology (S3),
GERMANY Stilleweg 2
E-mail: stichler@gsf.de 30655 Hannover
GERMANY
Tel.: +49 511 6432527
Fax: +49 511 6433665
E-mail: Axel.Suckow@GGA-Hannover.de

Mr. Rolf TRETTIN Mr. Ioannis LEONTIADIS

Umweltforschungszentrum Leipzig-Halle National Center for Scientific Research
GmbH “Demokritos”
Sektion Hydrogeologie Institute of Physical Chemistry
Theodor-Lieser-Str. 4 15310 Agh Paraskevi
D-06120 Halle P.O. Box 60228
GERMANY GREECE
Tel.: +49/0 345 558 5207 Tel.: +30-1-6536873
Fax: +49/0 345 558 5559 Fax: +30-1-6511766
E-mail: trettin@hdg.ufz.de E-mail : leontiad@mail.demokritos.gr

Mr. Miklos SUVEGES Mr. Zsuzsa SZANTO

Water Resources Research Centre Plc. Institiute of Nuclear Research of the HAS
(VITUKI Rt.) P.O. Box 51
Institute for Pollution Control H-4001 Debrecen,
Isotope Hydrology Laboratory HUNGARY
H-1095 Budapest, Kvassay J. ut 1. Tel.: +36-52-417266/1407
HUNGARY Fax: +36-52-416181
Fax: +36-1-2168140 E-mail: aszanto@moon.atomki.hu
E-mail: suveges@vituki.hu
 40

Mr. Peter VANCSURA Mr. Bhishm KUMAR

ADUKÖFE Isotope Hydrology Laboratory
Baja National Institute of Hydrology
Péter-Pál u. 8-12. Roorkee University Campus
6500 Roorkee- 247667
HUNGARY Hardwar, U.P.
Tel.: +36-79-421-010 INDIA
Fax: +36-79-423-816 Tel.: +91 1332 72906 to 72909
E-mail: vapety@freemail.hu ext. 224(O), 324 (R)
adukvfmk@ktm.x400gw.itb.hu +91 1332 70846 (R)
Fax: +0091 1332 72123 & 73976
E-mail: bk@cc.nih.ernet.in

Ms. A.R. NAIR Mr. K.K. NARAYAN

Isotope Hydrology Section Low-Level Tritium Laboratory
Isotope Applications Division Radiation Safety Systems Division
B.A.R.C., Trombay B.A.R.C. Hospital
Mumbai 400 085 Anushakti Nagar
INDIA Mumbai- 400 094
E-mail: svnavada@apsara.barc.ernet.in INDIA
E-mail: nkk@magnum.barc.ernet.in

Mr. Zainal ABIDIN Ms. Elisabetta BOARETTO

Hydrology Isotope Laboratory Radiocarbon Dating Lab.
Pusat Penelitian dan Pengembangan Kimmel Center for Archaeological Science
teknologi ESER dept.
Isotop dan Radiasi - BATAN The Weizmann Institute of Science
Jl. Cinere Ps.Jumat P.O Box 7002 JKSKL 76100 Rehovot
Jakarta, 12070 ISRAEL
INDONESIA Tel.: +972 8 934 3213
Tel: +062 21 7659376 Fax: + 972 8 934 4124
Fax: +062 21 7513270 & 062 21 7691607 E-mail: elisa@wis.weizmann.ac.il
E-mail: HIS45@bit.net.id

Mr. Davide CAPRA Mr. Onelio FLORA

CESI SpA - B.U. Environment University of Trieste
39, via Reggio Emilia Department of Geological, Environmental and
20090 Segrate Marine Sciences
Milan Via E. Weiss, 6
ITALY 34100 Trieste
Tel.: +39 02 21258154 ITALY
Fax: +39 02 21258691 Tel.: +39-040-6762153
E-mail: capra.davide@cesi.it Fax: +39-040-6762152
E-mail: flora@univ.trieste.it
 41

Mr. Mario MUSSI Mr. G. Sergio TAZIOLI

Laboratorio chimico-isotopico Dept. of Scienze di Materiali e della Terra
Istituto Internazionale per le Ricerche Faculty of Engineering
Geotermiche Via Brecce Bianche
Consiglio Nazionale Delle Ricerche 60131 Ancona
56010 Ghezzano, Pisa ITALY
Via V. Alfieri, 1 Tel.: +39-071-2804234
ITALY Fax: +39-071-2204714
Tel.: +39 050 315 2319 E-mail: tazioli@popcsi.unian.it
Fax: +050 3152323/+050 3152322
E-mail: mussi@iirg.pi.cnr.it

Mr. Ichiro GOTO Mr. Keisuke ISOGAI

Institute of Research and Innovation Research & Development Section 1
Kashiwa Laboratory Japan Chemical Analysis Center
Nuclear Chemistry and Chemical 295-3, Sanno-cho, Inage-ku, Chiba-shi,
Engineering Center Chiba, 263-0002
1201, Takada, Kashiwa JAPAN
Chiba 277-0861 Tel.: +81-43-423-5325
JAPAN Fax: +81-43-424-8660
Tel.: +81-471-44-9098 E-mail: k-isogai@jcac.or.jp
Fax: +81-471-44-7602
E-mail: goto@iri.or.jp

Mr. Kiriko MIYAMOTO Mr. Maki TSUJIMURA

National Institute of Radiological Sciences Hydrology Laboratory
4-9-1 Anagawa Institute of Geoscience
Inage-ku, Chiba-shi 263-8555 University of Tsukuba
JAPAN Tsukuba 305-8571
Tel.: +81-43-206-3156 JAPAN
Fax: +81-43-251-4853 Tel.: +81-298-53-2568
E-mail: kiriko@nirs.go.jp Fax: +81-298-51-9764
E-mail: mktsuji@atm.geo.tsukuba.ac.jp

Ms. Susan KILANI Mr. Yongkwon KOH

Laboratories and Water Quality Dept. Radioactive Waste Disposal Team
Environmental Isotope Section Korea Atomic Energy Research Institute
1 Subhi Ameen Amr ST. P.O. Box 105, Yusung
8th Circle Taejon 305-600
P.O. Box 2412 KOREA
Amman Tel: +82-42-868-8159
JORDAN Fax: +82-42-868-2063
Tel.: +00962 6 5864361/2 E-mail: nykkoh@kaeri.re.kr
Fax: +00962 6 5825275
E-mail: isolab@go.com.jo
 42

Mr. Mohd Shahid AYUB Mr Hamid MARAH

Tracer Application Group 65, Rue Tansift Agdal
MalaysiaDivision of Technology Industry Rabat
n Institute for Nuclear Tech. Research MAROCCO
Bangi, 43000 Kajang Tel :(00) + 212 7 71 20 23
MALAYSIA Fax:(00 ) +212 7 71 19 40
Tel.: +6-03-8925-0510 E-mail: marah@cnesten.org.ma
Fax: +6-03-8925-2577
E-mail: shahid@mint.gov.my

Mr. Pavel POVINEC Mr. J. (Hans) VAN DER PLICHT

Marine Environment Laboratory Centre for Isotope Research
IAEA Radiocarbon Laboratory
4, Quai Antoine 1er University of Groningen
BP 800, MC98012 Nijenborgh 4, 9747 AG Groningen
MONACO THE NETHERLANDS
E-mail: P.Povinec@iaea.org Tel.: +31-50-3634730
Secretary: +3634760
Res.: +5419390
Fax: +31-50-3634738
E-mail: plicht@phys.rug.nl

Mr. Uwe MORGENSTERN Dag Øistein ERIKSEN

Institute of Geological and Nuclear Sciences Dept. for Environmental Technology
Ltd. Institute for Energy Technology
PO Box 30-368 P.O. Box 40
Lower Hutt N-2027 Kjeller
NEW ZEALAND NORWAY
E-mail: U.Morgenstern@gns.cri.nz Tel.: +47 6380 6312
Fax: +47 6381 5553
E-mail: dag.eriksen@ife.no

Mr. Manzoor Ahmad CHOUDHRY Mr. Tadeusz KUC

Radioisotope Hydrology Group University of Mining and Metallurgy
Radiation & Isotope Application Division Faculty of Physics and Nuclear Techniques
Pakistan Institute of Nuclear Science and Dept. of Environmental Physics
Technology Al. Mickiewicza 30
P.O. Nilore, Islamabad 30-059 Krakow
PAKISTAN POLAND
Fax: +92 51 9290275 Fax: + 48 12 6340010
E-mail: manzoor.pins@dgcc.org.pk E-mail: kuc@novell.ftj.agh.edu.pl
 43

Mr. Wojciech SOLTYK Ms. Maria de Fátima ARAÚJO

Institute of Nuclear Chemistry and Tech. Departamento de Química
Shipment address: Instituto Tecnológico e Nuclear
Laboratorium Izotopow Srodowiskowych Estrada Nacional 10
Instytut Chemii I Techniki Jadrowej 2686-953 Sacavém
Zaklad V1 PORTUGAL
Dorodna 16, 03 Tel.: +351 21 994 6204
195 Warszawa Fax: +351 21 994 1455
POLAND E-mail: faraujo@itn1.itn.pt
Fax: (+4822) 811 15 32, 811 19 17
E-mail: sekdyrn@orange.ichtj.waw.pl

Ms. Niculina PAUNESCO Mr. Augustin TENU

IFIN-HH National Institute of Meteorology and
Department of Applied Physics Hydrology (NIMH)
P.O. Box MG-6 Soseaua Bucuresti-Ploiesti, 97, Sector 1
Bucharest-Magurele 71552, Bucharest
ROMANIA ROMANIA
Tel.: +40-1-2215756
Fax: +40-1-2303143
E-mail: isotopes@fx.ro

Ms. Carmen VARLAM I. FERRONSKY, V. POLYAKOV

National Research Institute for Isotopic All-Russian Scientific Research Institute of
Technologies Hydrogeology & Engineering Geology
1000 Rm.Valcea, P.O.Box10 Nuclear Geophysics and Isotope Method
ROMANIA Laboratory
Tel.: +40-50-733890 142452, Zeleny Village, Noginsk District,
Fax: +40-50-732746 Moscow Oblast
E-mail : mvarlam@ns-icsi.icsi.ro RUSSIAN FEDERATION
Fax: +(095) 913-51-26
E-mail: polyakow@online.ru

Mr. Alexander NIKITIN Mr. Janja VAUPOTIC

Typhoon Laboratory Dept. of Physical and Organic Chemistry
Lenin Street 82 J. Stefan Institute
249038 Obninsk Jamova 39
Kaluga region POB 3000
RUSSIAN FEDERATION SI - 1001 Ljubljana
Tel.: +(08439) 7 13 05 SLOVENIA
Fax: +(08439) 7 11 47 Tel.: +386-1-4773213
E-mail: nikitin@iem.obninsk.ru Fax: +386-1-4773811
E-mail: janja.vaupotic@ijs.si
 44

Mr. Balt VERHAGEN Ms. M. Antonia SIMON

Mr. Mike BUTLER Laboratorio de Isótopos Ambientales Ed.3
Schonland Research Centre University of CIEMAT, Avda Complutense 22
Witwatersrand E-28040 Madrid
P.O. Wits SPAIN
2050 Johannesburg Fax: +34-91-3466005
SOUTH AFRICA E-mail: simon.arauzo@ciemat.es
E-mail: Verhagen@schonlan.src.wits.ac.za
E-mail: butler@schonlan.src.wits.ac.za

Mr. Luis ARAGUAS Mr. M. K. AHMED

Díaz TEIJEIRO Local Coordinator
Laboratorio de Hidrología Isotópica Groundwater and Wadis Administration
CETA/CEDEX Kilo 10, c/o UNDP Office
Centro de Estudios y Experimentación de The Res. Rep. of the U.N.
Obras Públicas Development Programme in the Sudan
C/ Alfonso XII, 3 y 5 P.O. Box 913
28014 Madrid Khartoum
SPAIN SUDAN
Tel.: +34 91 335 72 18/ 64 E-mail: gww_infocent@sudanmail.net
Fax: +34 91 335 72 67
E-mail: Luis. Araguas@cedex.es
M.Teijeiro@cedex.es

Mr. R. KIPFER Mr. Ulrich SCHOTTERER

Environmental Physics Physikalisches Institut
Swiss Federal Institute of Environmental Abteilung Klima- und Umweltphysik
Science and Technology (EAWAG) Universitaet Bern
Swiss Federal Institute of Technology(ETH) Siderstrasse 5
Ueberlandstr. 133 CH-3012 Bern
8600 Duebendorf SWITZERLAND
SWITZERLAND E-mail: schotterer@climate.unibe.ch
Tel.: +41-1-823 55 30/31/32
Fax: +41-1-823 52 10
E-mail: kipfer@eawag.ch

Mr. A. SHAKHASHIRO Mr. Hassani J. MJENGERA

Head, Quality Assurance Office Water Laboratories Unit
Syrian Atomic Energy Commission Ministry of Water
P.O.Box 6091 Maji Ubungo
Damascus P.O. Box 35066
SYRIA Dar Es Salaam
Fax: +963-11-6112289 THE UNITED REPUBLIC OF TANZANIA
E-mail: atomic@net.sy Tel.: +255-022-2450047
Fax: +255-022-2451447
 45

Mr. Pisit SUNTARAPAI Mr. O. UZUN

Health Physics Division DSI Teknik Arastirma ve Kalite Kontrol
Office of Atomic Energy for Peace (OAEP) Dairesi Baskanligi
16 Vibhavadee-Rangsit Rd., Ladyao 06100 Ankara
Chatuchak, Bangkok 10900 TURKEY
THAILAND Fax: +90 - 312 3992795
Fax: +662-5613013 E-mail: takk@dsi.gov.tr
E-mail: pisit@oaep.go.th

Mr. Michael BUZINNY Mr. W. J. JENKINS

LSC and Methodology group School of Ocean and Earth Sciences
Radioecology Laboratory Southampton Oceanography Centre
Research Center for Radiation Medicine Southampton, SO14 3ZH
Academy of Medical Sciences of Ukraine UNITED KINGDOM
04050, Melnikovastr.53 Phone: +44-2380-596573
Kiev Fax: +44-2380-593052
UKRAINE E-mail: W.Jenkins@soc.soton.ac.uk
Tel.: +380 44 2445874
Fax: +380 44 2137202
Res.: +380 44 2661937
E-mail: Buzinny@Bigfoot.com

R.I. OTLET/Jil WALKER CHAO-LI/ Jack LIU

RCD-RadioCarbon Dating Isotope Geochemistry Laboratory
The Old Stables Illinois Geological Survey (ISGS)
East lockinge Champaign
Wantage Illinois 61820
Oxfordshire OX12 8QY USA
UNITED KINGDOM Tel.: +217-244-2192
Tel./Fax: +44 1235 833667 Fax: +217-333-4722
E-mail: rcd.lockinge@virgin.net E-mail: jliu@isgs.uiuc.edu

Mr. Steven J. FRITZ Ms. Charlene GRALL

Dept. of Earth and Atmospheric Sciences Mr. James D. HAPPELL
Purdue University Mr. Gote OSTLUND
West Lafayette, Indiana 47907 Tritium Laboratory
USA University of Miami/RSMAS
Tel.: +(765) 494-3276 4600 Rickenbacker Cswy.
Fax: +(765) 496-1210 Miami, Florida 33149
E-mail: sfrito@eas.purdue.edu USA
Phone: +(305)361-4119
Fax: +(305)361-4112
E-mail: cgrall@rsmas.miami.edu
E-mail: jhappell@rsmas.miami.edu
E-mail: gostlund@rsmas.miami.edu
 46

Mr. Austin LONG Mr. Robert L. MICHEL

Departments of Geosciences and Hydrology US Geological Survey
and Water Resources 345 Middlefield Rd
University of Arizona MS 434
Tucson 85721-0077 Menlo Park CA 94025
USA USA
E-mail: along@geo.arizona.edu Phone: +650-329-4547
Fax: +650-329-5590
E-mail: RLMICHEL@usgs.gov

Mr. Michael P. NEARY Mr. Steve PELPHREY

Low-Level Tritium Laboratory Isotech Laboratories
Center for Applied Isotope Studies 1308 Parkland Court
University of Georgia Champaign IL. 61821
120 Riverbend Road USA
Athens, Georgia 30602 Tel.: +217-398-3490
USA Fax: +217-398-3493
Tel: +(706) 542-6115 E-mail: steve@isotechlabs.com
Fax: +(706) 542-6106

Mr. Robert POREDA Mr. Kip SOLOMON

Dept. of Earth and Environmental Sciences University of Utah
227 Hutchison Hall Dept. of Geology and Geophysics
University of Rochester 135 South, 1460 East
Rochester, NY 14627 Room 719
USA Salt Lake City,
Tel.: +716-275-0051 Utah 84112-0111
Fax: +716-244-5689 USA
E-mail: Poreda@earth.rochester.edu Tel.: +801-581-7231
Fax: +801-581-7065
E-mail: ksolomon@mines.utah.edu

Mr. Zafer TOP Ms. Nada MILJEVIC

Research Professor Vinca Institute of Nuclear Sciences
University of Miami P.O. Box 522
4600 Rickenbacker Causeway 11001 Belgrade
Miami, FL 33149 YUGOSLAVIA
USA Tel.: +381 11 453 967
Tel: +(305) 361 4110 Fax: +381 11 444 72 07
Fax : +(305) 3614911 E-mail: emiljevi@rt270.vin.bg.ac.yu
E-mail: ztop@rsmas.miami.edu

APPENDIX E: LISTING OF ANNOUNCEMENT, DATA REPORTING

INSTRUCTIONS, REPORTING FORM
 47

E1: Invitation letter with requirements:

ANNOUNCEMENT

Sixth IAEA Intercomparison Exercise of Low-Level Tritium Measurements in

Water - TRIC2000

The Isotope Hydrology Laboratory of the International Atomic Energy Agency is organizing
the sixth interlaboratory comparison exercise of tritium analysis of water at environmental
activity levels (TRIC2000). All laboratories and institutes having analytical capabilities for
low-level tritium measurements (gas proportional counting, liquid scintillation counting, He3
ingrowth method) are invited to take part in the exercise. Purpose of the intercomparison
exercise is to give the participating laboratories the opportunity to check the agreement of
their measurements with those of other laboratories, to discover the occurrence of errors and
systematic deviations and, eventually, to improve the quality and reliability of their analytical
procedures.

A set of six water samples has been prepared for distribution to the interested laboratories:

Codes: TRIC2000-T1 to TRIC2000-T5, 5 samples, 500ml each with tritium concentrations

between 0 and about 25 TU (0-3 Bq/kg)
and
TRIC2000-T6, 1 sample, in a marked 100ml bottle with a tritium concentration of about 500
TU (120 Bq/kg)

In addition to the range from 0-25 TU (T1-T5) covered by previous intercomparison

exercises we will provide one sample (T6) with a tritium concentration of approximately 500
TU. This sample should be analyzed by direct counting without enrichment and thus should
allow to check the proper calibration of the used laboratory standards as well as the
precision of laboratories engaged in the analysis of higher tritium concentration levels.
The samples will be prepared by spiking a tritium-free water from an Artesian well
(Steiermark/Austria) with well specified dilutions of a NIST tritiated water standard (NIST
4927F). The samples have not been distilled and will be shipped in polyethylene bottles.

The laboratories will be requested to report for each sample the results of the tritium
measurement in TU and its assigned standard uncertainty at the 1 sigma level. It would be
useful to specify in addition the individual uncertainty components as assessed in the
individual laboratories (standardization, enrichment, counting, etc). Mandatory information
include the used equipment and analytical procedures as well as the calibration method
(specifying the used tritium standards and dilutions). Any other information which is
considered relevant and useful will be appreciated. The respective forms including detailed
information and instructions for preparation of the report will be sent out together with the
samples. The results will be evaluated in the IAEA and each participating laboratory will
receive a report with the individual data, the statistical evaluation of all results and their
graphical display. TRIC2000 is open to all low-level tritium laboratories with full confidentiality
with respect to link between identity of participants and results.

We particularly request mass-spectrometric laboratories using the He3 ingrowth method to

participate in the intercomparison exercise. In these cases we will take into account that the
analytical procedure might take longer than for the beta-counting laboratories. Please
specify in your response if you do not want to receive the higher-TU sample.

Participation in the exercise will be limited to approximately 120 laboratories. If you are
interested in taking part in the intercomparison, please respond directly BY EMAIL at your
 48

earliest convenience, but not later than August, 15th, to the TRIC2000 account
(isotope.hydrology.lab@iaea.org) specifying the subject TRIC2000. Please indicate also in
your response your full address for shipment of the samples as well as your email and fax
number for further correspondence.

In case email is not available, please send your information to the following address:

Isotope Hydrology Laboratory

"TRIC2000"
International Atomic Energy Agency
P.O. Box 100
Wagramerstrasse 5
A-1400 Vienna
Austria
Fax +43-1-26007
Tel: +43-1-2600-21766
Email: isotope.hydrology.lab@iaea.org

Time schedule

Deadline for registration of your participation: August, 15th

Shipment of samples: August/September 2000
Deadline for submission of the results to IAEA: December, 31th.

We are looking forward to your active participation in the forthcoming exercise.

Best regards,

Manfred Groening / Gisela Winckler

For more information please refer to the home page of the IAEA Isotope Hydrology
Laboratory
http://www.iaea.org/programmes/rial/pci/isotopehydrology/
 49

E2: Instructions for Data Reporting (attached to shipped samples):

ISOTOPE HYDROLOGY LABORATORY

WAGRAMER STRASSE 5, P.O. BOX 100, A-1400 VIENNA, AUSTRIA
TELEPHONE: (+43 1) 260021766, FACSIMILE: (+43 1) 26007 21763, E-MAIL: isotope.hydrology.lab@iaea.org INTERNET: http://www.iaea.org

September 1st, 2000

Dear Colleague:

Thank you very much for your interest in the 6th IAEA Tritium Interlaboratory Comparison
Exercise TRIC2000.
Enclosed please find the set of six samples which we prepared for this exercise. The samples
T1-T5 cover the range between 0 and 25 TU (0-3 Bq/kg). Note that the consecutive
numbering of the samples T1-T5 does not imply increasing tritium concentrations, i.e. the
samples are in random order. Sample T6 has a higher tritium activity of ca. 500 TU.

An Laboratory Identification Number (Lab ID) is assigned to each laboratory and is given
on the bottle labels. This identification number ensures the confidentiality and will be used
throughout the exercise, i.e. it will allow you to identify your data in the final report of the
exercise.

Reporting your results

For analysing the samples, we ask you to follow the routine procedure as practised in your
laboratory. You are kindly requested to provide detailed information on the measurement
technique applied as well as on the calibration methods used (measurement scale,
standardisation procedure, half-life value, etc.). Please also note the Annex providing
information on the newly determined tritium half life and its implications for future tritium
work.

We kindly ask you to carefully complete the Reporting Sheet. We would ask you to submit
the results by e-mail in order to avoid typing errors and to simplify the evaluation procedure.
We will send the electronic version of the Reporting Sheet (tric2000.doc, MicrosoftWord 6.0
for Microsoft Windows) as an e-mail attachment. In case you do not use Microsoft Windows
or/and Microsoft Word the Reporting Sheet will also be sent to you as plain text by e-mail
 50

(ASCII). In addition, both electronical versions of the Reporting Sheet are available for
download on our Internet site.
You are kindly requested to fill out the electronic version and send it back to us by e-mail
(Word6.0-file attachment preferred). For your convenience, a printed copy of the Reporting
Sheet is included in the sample shipment. The deadline for sending back the results will be
December, 31th.

Statement of uncertainty

Please give a careful statement of the uncertainty of your data. The uncertainty should not only
include the standard uncertainty of the individual measurement (counting). It rather means an
assessment of all uncertainty components of the complete analysis procedure including all
steps. For details please refer to the draft paper by Rozanski & Gröning which will be
shortly sent to you by email and will be available to download from our Internet site.

Evaluation of the results/Quantitative Assessment

The exercise is designed as a proficiency test. All samples were prepared by dilution of a
NIST water standard (NIST 4927F) by tritium-free water. The assigned value of each sample
(the best estimate of the “true” activity) is the tritium activity as calculated from the dilution
procedure.
For evaluation of the results we will use the following procedure:
(i) The submitted data and uncertainties will be compared to the assigned value allowing a
direct assessment of the performance of each laboratory.
(ii) In a second approach, a two-stage statistical treatment which has been adopted in previous
IAEA interlaboratory exercises will be applied to the population of the submitted results
allowing the identification of obvious outliers (stage 1) and examination of the deviation of
the individual results from the mean value of all laboratories.

If you have any questions concerning the exercise or the completion of the report sheet, please
do not hesitate to directly contact us. Information related to the exercise as well as documents
for download will be provided on our Internet site at
http://www.iaea.org/programmes/rial/pci/isotopehydrology/ic.htm.

We understand that participating in the exercise means a lot of work for you and, therefore all
the more are grateful for your kind co-operation;

Yours sincerely,

G. Winckler M. Gröning

Annex: Implication of the newly determined tritium half-life for future tritium work
 51

ANNEX

Implication of the newly determined tritium half-life for future tritium work

It is highly desirable that all laboratories participating in the exercise should aim to produce
data which are consistent with that of other laboratories, i.e. to express their results on the
same measurement scale.
In the past, the main reference has been a scale based on the NIST standard water
NBS 4926 C (calibration date 3 September 1978) which was proposed by C.B. Taylor and W.
Roether in 1982 [1]. As the stock of this NIST standard generation is now depleted, the US
National Institute of Science and Technology has prepared a new set of standard waters
(SRM4926E, SRM4927F from September 1998, [2]). Furthermore, NIST has re-evaluated all
tritium half—life determinations and has provided a new half-life value for tritium of 4500 ± 8
days [3].
On the basis of these new developments, the IAEA Isotope Hydrology Laboratory
recommends for future work to adopt the new NIST standard generation as the new scale for
low-level tritium work and to use the new value for the tritium half-life. In this scale, 1TU is
0.11919 ± 0.00021 Bq/kg H2O.

Compatibility of older data (based on an old scale) with the new standard scale
Calibration of the data on the new scale leads to slightly but significantly different values.
However, we would like to point out that adopting the new scale means a complete re-
calibration of all standard solutions used in the laboratory [2]. For consistency reasons, only
laboratories that completely re-calibrate their standard procedures should use the new
conversion equation given above. For details concerning the adjustment to the new scale
please contact us directly. A short note on this issue is under preparation by C.B. Taylor and
M. Gröning.

References

[1] C.B. Taylor and W. Roether (1982). A uniform scale for reporting low-level tritium
measurements in water. Applied Radiation and Isotopes, Vol. 33, 377-382.
[2] L.L. Lucas (2000). Massic activity ratios of the NBS/NIST tritiated-water standards issued
between 1954 and 1999. Journal of Research of the National Institute of Standards and
Technology, Vol. 105.
[3] L. L. Lucas and M.P. Unterweger (2000). Comprehensive review and critical evaluation of
the half-life of tritium. Journal of Research of the National Institute of Standards and
Technology, Vol. 105.
 52

E3: Data Reporting Sheet (attached to shipped samples):

Sixth IAEA Intercomparison Exercise of Low-Level Tritium Measurements in Water -

TRIC2000
Reporting Sheet
Participant Information
Participating Laboratory Lab ID (identical
to the number on
the sample
bottles)

Method applied
Liquid scintillation counting
gas proportional counting
He3/T ingrowth (no need to follow the query below)
Standardization/Calibration
Which standard do you use ?
(e.g. from NIST, PTB, IRMM,
other;
Please specify name, activity and
certification date)
Which lab standard do you use?

How was your lab standard

prepared?

Please specify the dilution

procedure

Which tritium half-life value do

you use for calculations of
tritium activities

General Lab information

Elevation of the laboratory [in m
above sealevel]
Counter location [ground
floor/sub floor/etc.]
Enrichment procedure used?
[Y/N]
Number of analyses per year
Range of typ. tritium values
measured
 53

Description of applied procedures

First destillation
Predistillation of samples
[Yes/No]
If yes, distillation at atmospheric
pressure or under vacuum
Conductivity after pre-
destillation [microS/cm]
Remarks:

Enrichment
Which kind of anode do you use
(e.g. steel)
Which kind of kathode do you
use
(e.g. Nickel)
Number of cells
Volume of the electrolytic cell
[ml]
Initial volume of sample water
[ml]
Final volume of sample water
[ml]
Temperature during enrichment
[C]
Coolant medium during
enrichment
Which type of electrolyte do you
use ?
(e.g. NaOH, Na2O2)
Which amount of electrolyte [g]
?
how long do you run the
electrolysis?
[hours]
Which current do you apply [A]
if different sequences, please
note
Total charge used
[Amperehours]
Number of enrichment runs [per
year]
Average enrichment factor
How do you determine the
enrichment factor?
Remarks:
 54

Second destillation and Neutralization

neutralization after enrichment
[Yes/No]
Type of chemical added [e.g.
PbCl2]
Amount of chemical added [g]
Final distillation [Yes/No]
If yes, distillation at atmospheric
pressure or under vacuum
Remarks:

Counting
For Laboratories using proportional counters
How do you prepare the
hydrogen gas
(e.g. reduction on Mg turnings)
temperature during reduction
process
which counting gas do you use?
Geometric volume of prop
counter [l]
Effective Volume of counter [l]
Voltage applied [V]
pressure [atm]
Temperature during counting [C]
Counting time [min]
Counting efficiency [%]
Calibration factor [TU/cpm]
Background countrate [cpm]
Used background water (origin
etc.)
Detection limit [TU]
For laboratories using liquid scintilation counters
Type/Model of counter
Type of vials (e.g. glass,
polyethylene)
Volume of the counting vials
[ml]
Volume of water in the vial [ml]
Type of scintillation cocktail
Amount of scintillation cocktail
Quench indicating parameter
Counting efficiency for an
 55

unquenchend standard [%]

Counting efficiency for a
quenched standard [%]
Calibration factor [TU/cpm]
Temperature during counting [C]
Counting time [min]
Sequential/cycle counting or in
block?
If sequential, specify sequences
Composition of each enrichment
run, please specify how many
samples, blanks, spikes and
standards per run
Background countrate [cpm]
Used background water (origin
etc.)
Detection limit [TU]
Uncertainty Assessment
How do you determine the uncertainty of
your tritium results?
Which uncertainty components contribute
to your over all error?
Uncertainty of enrichment factor [%]
Uncertainty of background countrate [%]
Uncertainty of standardization [%]
Uncertainty of other components
Measurement

Please read this section carefully before you fill out the results table.
1. Please calculate the tritium values for the date of sample collection (July, 1st of 2000).
2. Please express your results in TU. If you use other units (e.g., Bq/kg), please specify it!
3. Please present your numerical TU values and errors even if they are negative and/or below
the usually stated limit of detection. Negative numbers can occur because the net tritium
count rate is in principle the difference between the count rate of the sample and that of a
tritium-free sample (background count or blank sample). The negative values should be
reported as such and will allow us an unbiased statistical treatment of sets of the data.
Please do NOT give any numbers using the < symbol (e.g., < 1 TU) as these data cannot be
used in the statistical analysis!
4. Please give the uncertainty of your results as standard uncertainty at 1 sigma-level
5. If you perform multiple analyses of the samples, please report in Table 1 the mean value
and specify the number of analyses in the respective column in Table 1.
6. Please supply the results of the multiple measurements in the extra table below.
 56

Final Results

Cod Tritium activity Standard date of Numbe

e [TU] uncertainty measurement r
(one sigma) of
Analyse
s
T1

T2

T3

T4

T5

T6

Used Half-Life:
If you used any conversion equation (e.g., from Bq/kg to TU), please specify:

In case you performed multiple measurement of the samples please add the results of the
individual measurements here
Individual Results of multiple analyses
T1 T2 T3 T4 T5 T6
1.
2.
3.

Laboratory Manager
Date
Please send the Reporting Sheet (electronic version strongly preferred) to:

email: isotope.hydrology.lab@iaea.org

Isotope Hydrology Laboratory

TRIC2000
International Atomic Energy Agency
P.O. Box 100
A-1400 Vienna
Austria
Tel: +43-1-2600 21766
Fax: +43-1-2600721763
 57

E4:Problem-indicating email to 26 laboratories

After the deadline had passed the submitted data were sreeened and obvious outlying
measurements identified. Emails similar as stated below were sent to altogether 26
laboratories. 20 replies and 12 corrected datasets were received

Evaluating your TRIC2000 report we found

the T6 value out of range
Please check your calibration.
Best regards,
Isotope Hydrology Laboratory
International Atomic Energy Agency
P.O. Box 100
A-1400 VIENNA
Austria

Evaluating your TRIC2000 report we found

the T1,T2,T3 and T5 value out of range
Please check for offset.
Best regards,
Isotope Hydrology Laboratory

Evaluating your TRIC2000 report we found

the T1,T4,T5 and T6 values out of range
Please check your calibration.
Best regards,
Isotope Hydrology Laboratory