Journal of Radioanalytical and Nuclear Chemistry

, Volume 314, Issue 2, pp 545–553 | Cite as

Primary activity standardization of 134Cs

  • Karsten Kossert
  • Justyna Marganiec-Gałązka
  • Ole J. Nähle


Cesium-134 has been measured by means of a new coincidence counting system which is equipped with a liquid scintillation detector to count β particles and a NaI crystal to detect γ-rays which are emitted simultaneously. In order to validate this new technique, additional measurements were carried out with an established 4πβγ-coincidence counting system with a proportional counter in the β channel. The coincidence counting experiments were complemented with measurements in further liquid scintillation counters with two and three photomultiplier tubes. The counting efficiencies for these systems were determined using CIEMAT/NIST efficiency tracing and the triple-to-double coincidence ratio (TDCR) method, respectively.


134Cs Activity standardization Coincidence counting Triple-to-double coincidence ratio method CIEMAT/NIST efficiency tracing 


The 4πβγ-coincidence counting technique is one of the most important methods for accurate activity determination in radionuclide metrology. At the Physikalisch-Technische Bundesanstalt (PTB), the method has been in use for a long time using proportional counters (PC) in the β channel and NaI crystals to detect γ-rays [1]. This method requires solid samples of the nuclide under study with low self-absorption. The sources must have a conductive surface and are then placed inside a proportional counter. Since the method makes high demands on the quality of the sources, the source preparation is laborious and time consuming.

In this work, we report on new coincidence counting experiments using a liquid scintillation (LS) counter in the β channel. To this end, a new counter was custom-built at PTB that incorporates a gamma detector in conjunction with a traditional 3-photmultiplier tube TDCR system. A liquid scintillation source is placed inside an optical chamber of the system and the ejected γ-rays are measured with a NaI(Tl) detector below the LS counter.

The three photomultiplier tubes (PMTs) in the new LS-γ coincidence system make simultaneous TDCR measurements possible. The LS samples were also measured in another TDCR system, which has been designed at PTB [2], as well as in a commercial LS counter to apply the CIEMAT/NIST method [3, 4]. The rather complex computation of the respective counting efficiencies was made with the aid of a stochastic model and will be described in detail.

The method was used to standardize 134Cs, which is an important radionuclide for calibration procedures in γ-ray spectrometry. Often, 134Cs is also used as a tracer for 137Cs activity determinations. Cesium-134 (T 1/2 = 2.0644 (14) a) decays mainly by β transitions which are accompanied by one or more coincident γ transitions [5]. A simplified decay scheme is shown in Fig. 1.
Fig. 1

Simplified decay scheme of 134Cs (see also Table 2)

At PTB, 4πβ(LS)-γ-coincidence counting has never been used before for this isotope. Thus, additional validation measurements were made with the conventional coincidence counting technique using a PC in the β channel.

All uncertainties stated in this work are standard uncertainties (coverage factor k = 1).


The measurements were performed using an aqueous 134CsCl solution which contains 50 μg inactive CsCl/g of solution and 0.1 mol/L HCl. A preliminary activity concentration of about 87.2 kBq/g on the reference date (1st June 2015, 0 h CET) was calculated from a gravimetrically determined dilution factor and from the known activity concentration of the master solution. A flame-sealed PTB-type glass ampoule with about 2 g of the master solution was measured in a calibrated ionization chamber. The master solution was also measured by means of γ-ray spectrometers and no radioactive impurity was found.

Sample preparation

Five LS samples were prepared with a 15 mL Ultima Gold™ scintillator in 20 mL low-potassium borosilicate glass vials. About 0.97 mL of distilled water and weighed portions of about 30 mg of the diluted 134Cs solution were added to each sample. Nitromethane (CH3NO2) was used to vary the counting efficiencies. In addition, one sample with 15 mL Ultima Gold™ scintillator and 1 mL of distilled water was prepared to measure the background counting rate, which was then subtracted.

Four sources were prepared using VYNS (polyvinylchloride–polyvinylacetate co-polymer) foils (15 μg/cm2 plus 15 μg/cm2 Au–Pd alloy on each side) mounted on stainless steel rings. The foils had been pretreated by electrospraying them with a colloidal SiO2 suspension on a circular area of 6 mm in diameter. The masses of the 134Cs solution deposited on a single source was determined by difference weighing of a pycnometer and were in the range from about 9.5 mg to 21 mg. The background was measured with a corresponding blank sample.

The masses of all samples as well as the dilution factor were determined gravimetrically using two Mettler balances traceable to the German national mass standard.

β(LS)-γ coincidence measurements

The 4πβ(LS)-γ coincidence system comprises a TDCR counter with an automated sample changer and a NaI scintillation detector (102 mm × 102 mm), which is located 32 mm below the LS vial in the optical chamber. The amplified and discriminated anode signals from the three Hamamatsu R331-05 PMTs are fed into an FPGA-based coincidence module which was designed at PTB [6]. This module—referred to as the 4KAM module—can also process an input signal from the NaI detector which is used as the γ channel. The 3 PMTs are mounted in the bore holes of an optical chamber which is made of the diffuse reflecting material OP.DI.MA (ODM98) produced by Gigahertz Optik GmbH, with a reflectivity of more than 98% over a wide wavelength range. The LS vial to be measured is placed in the centre of this chamber. Massive lead shielding around the optical chamber and around the NaI detector helps to reduce background counting rates.

The anode signals of the PMTs are amplified by two CAEN N978 fast amplifiers and are discriminated by an Ortec 935 constant-fraction discriminator. The amplification could also be realized with only one module, but two devices were used in order to eliminate potential miscounting due to electrical cross-talk between adjacent channels. The discrimination threshold was adjusted to just below the single electron peak, analysing the pulse height spectra of each PMT with an analogue to digital converter.

The anode signals of the PMT coupled to the NaI are amplified using an ORTEC 575A module and then fed into a single channel analyser (SCA) of type ORTEC 551. The SCA was adjusted to cover an energy range that includes the full-energy peaks of the two most intense γ-rays with energies of 604.72 keV (P γ = 97.63%) and 795.86 keV (P γ = 85.47%) [5], respectively. The signals are then fed into the 4KAM coincidence module. In this way, the TDCR system can be used for simultaneous 4πβγ coincidence measurements. The common extendable dead time of the 4KAM module was adjusted to 30 μs. The coincidence resolving times were defined to be 77.5 and 76.5 ns or the β channel (3 PMTs of the LS system) and for the βγ coincidences, respectively. Due to the slower signal processing of the γ-channel a delay for the β-channel of 2.482 μs had to be inserted in the coincidence logic of the 4KAM.

The efficiency range covered when using chemical quenching was too low for a sophisticated efficiency extrapolation for the coincidence counting method. Thus, the same samples were measured many times using thin neutral density (ND) filters (LEE Filters Worldwide, Hampshire, UK) with light transmission from 69.3% to 6.6% (ND0.15 to ND1.2), respectively. The filters reduce the light intensity almost equally at all wavelengths in the visible light spectrum. These grey filter foils were bent around the cylindrical vials. The bottom of the vial was also covered with a circular grey foil disc. The efficiency range was considerably extended when using these grey filters. A benefit of efficiency variation by means of grey filters is that this method is reversible, i.e. the initial (higher) counting efficiency can be obtained again when removing the grey filter foils.

β(PC)-γ coincidence measurements

The VYNS sources were measured in a 4πβγ coincidence system equipped with a pillbox-type PC flooded with pure (99.95%) methane under atmospheric pressure. The photons were counted by means of a 75 mm × 75 mm NaI(Tl) detector above and a 100 mm × 100 mm Na(Tl) below the PC. Thus, two measurements with different settings in the γ-channel can be carried out simultaneously. The analogue signal processing was accomplished by means of a preamplifier, an amplifier and an SCA, after which a non-extendable dead time of about 8 μs was introduced by means of high-precision dead time units [7]. A delay unit was used to minimise the delay between the γ- and β-channels. The digital outputs of the β- and γ-channels were fed into coincidence units with a resolving time of about 1 μs. The counting rates were corrected for background and nuclear decay and analysed on the basis of the expression given by Smith [8] for the coincidence counting rate (see also [9]).

A discriminator was used to adjust the threshold above noise level in the β channel. In the γ channels two different settings were used. In the first setting the whole γ spectrum was accepted and in the second setting the energy range was defined to include the full-energy peaks of the dominant peaks at 604.72 keV and 795.86 keV as used for 4πβ(LS)-γ coincidence (see previous section). The β counting efficiency was varied by successive addition of VYNS absorption foils.

The measurements of 134Cs samples and a background sample were repeated several times with the aid of an automated sample changer.

CIEMAT/NIST measurements

The LS samples were measured in a Wallac 1414 Guardian™ liquid scintillation spectrometer. Figure 2 shows a measured liquid scintillation spectrum obtained with this apparatus. The calibration curve, i.e. the counting efficiency of 3H as a function of the quenching indicator SQP(E), was measured with the aid of a PTB standard solution of 3H. The LS samples containing 3H have the same sample composition and geometry as the 134Cs LS samples.
Fig. 2

Measured LS spectrum of 134Cs. The measurement was carried out in a Wallac 1414 counter with logarithmic amplification. A background spectrum has been subtracted

TDCR measurements

The LS samples were measured in two custom-built TDCR systems of PTB. The first system makes use of the MAC3 coincidence module [10] with 40 ns coincidence resolving time and was described by Nähle et al. [2]. The second TDCR system is part of the new LS-γ counter with an automated sample changer as described above. For the (pure) TDCR measurements, the coincidence resolving time in this system was about 76.5 ns. Both TDCR systems are equipped with an optical chamber holding three Hamamatsu R331-05 photomultiplier tubes (PMTs) surrounding a liquid scintillation sample in its centre. Both systems are shielded with lead to reduce the background counting rate.

The anode signals of the PMTs are amplified by a CAEN N978 fast amplifier and discriminated by an Ortec 935 Constant-Fraction Discriminator. In both TDCR systems, the discrimination threshold was adjusted to just below the single electron peak.

An advantage of the TDCR-gamma configuration is that experimental TDCR and coincidence counting data can be obtained with the same instrument. In principle, the corresponding methods can even be applied simultaneously. However, the methods often require different adjustments (e.g. for the coincidence resolving time or discriminator thresholds) and, hence, separate measurements are needed.

Analysis of data from 4πβ(LS)-γ and 4πβ(PC)-γ coincidence counting

For both coincidence counting methods, the counting rates N β, N γ and N c were used to plot the ratio N β N γ/N c as a function of x = (1−N c/N γ) / (N c/N γ). First and second order polynomial functions f(x) were fitted to the data by means of the least squares method and the function was finally extrapolated to N c/N γ = 1 (Figs. 3, 4). Polynomials are often used for the fit when applying coincidence counting techniques for 134Cs [11]. However, the shape of the extrapolation curve depends on the detectors used as well as on the window settings [12] and also linear fits have been applied (see, e.g. [13]).
Fig. 3

N β N γ /N c measured with a 4πβ(LS)-γ coincidence system as a function of x = (1−N c/N γ ) / (N c/N γ ); function f(x) is a 2nd degree polynomial which was fitted to the experimental data, and the bars represent the statistical uncertainty of the individual data. In the left figure N β corresponds to the counting rate of double coincidences N D, whereas triple coincidences were used in the right figure (N β = N T)

Fig. 4

N β N γ/N c vs. f(x) as in Fig. 3 but measured with a 4πβ(PC)-γ coincidence system. For the right figure the γ-energy range was limited and includes the full-energy peaks of dominant γ emissions. For the left figure the whole γ spectrum was accepted

For 4πβ(LS)-γ coincidence counting the β counting rate was defined either as the counting rate for the logical sum of double coincidences N D (left side of Fig. 3) or as the counting rate for triple coincidences N T (right side of Fig. 3). Since the triple counting efficiency is lower than the double counting efficiency, the extrapolation range is somewhat larger when using triple coincidences. Also the shape of the fitted function f(x) changes when using N T instead of N D (Fig. 3). The range of data which were taken into account for the fitting procedure was varied in order to evaluate its influence on the final result. An alternative fit procedure was applied, taking into account the statistical uncertainty of individual data points as weighting factors. Eventually, 24 individual results were obtained from the fitting procedures. The arithmetic mean of these values was adopted as the final result. Half of the relative deviation between the highest and the lowest results amounts to 0.147% and was taken into account as an uncertainty component referred to as “fitting uncertainty”. The full uncertainty budget is shown in Table 1.
Table 1

Uncertainty budgets for the activity concentration of the 134Cs solution measured by four methods


u(a)/a in  %



β(PC)-γ CC

β(LS)-γ CC

Counting statistics










Dead time










Resolving time



Counting time










Decay correction





Extrapolation of efficiency curve



Impurities (no radioactive impurity detected)





3H tracer activity and interpolation of efficiency curve


TDCR value and interpolation of efficiency curve


Model and decay data





Ionization quenching and kB value



PMT asymmetry



Sample (in)stability



Fitting uncertainty (see text)



Square root of the sum of quadratic components





* Correlation coefficients are taken into account

Figure 3 contains data from 4 LS samples. The samples have different degrees of chemical quenching and measurements were carried out with and without various grey filters. The extrapolation was done with the whole data set shown in Fig. 3, i.e. using results from the 4 LS samples.

In an alternative approach, the results from the LS samples were analysed separately. The mean value of the four individual results obtained from the extrapolation method was almost identical to the above-stated result. Hence, the combination of both efficiency variation techniques seems to be a valid approach in the case of 134Cs, which yields a rather high β counting efficiency.

The data analysis for 4πβ(PC)-γ coincidence counting was done in a similar manner using first and second order polynomials as well as different ranges of data which were taken into account. Here, the β channel is defined as the PC, but two γ channels were used simultaneously. For the γ channels two window settings were used as described above. Thus, 48 individual results were obtained. Half of the relative deviation between the highest and the lowest results amounts to 0.186% and was taken into account as an uncertainty component (Table 1).

The uncertainty components related to dead time and coincidence resolving time were conservatively estimated in the case of 4πβ(LS)-γ coincidence counting. The 4KAM module uses different techniques to determine the coincidences for the β-channels (D and T) and βγ coincidences. The former are realized by a delay line with increments of 2 ns while the latter are bound to the FPGA clock, resulting in a jitter and increments of about 5.882 ns. Further investigation may help to reduce these uncertainty components in the future.

Computation of the LS counting efficiencies and analysis of data for CIEMAT/NIST and TDCR

The required decay data for the efficiency computation were taken from [5]. Cesium-134 decays by different β transitions which all lead to excited states of 134Ba, i.e. the β transitions are accompanied by at least one and up to four simultaneous γ transitions. An electron-capture branch with a low transition probability of 0.0003% was neglected in this work. Thus, the computations were made for four β transitions and eleven γ transitions as shown in Fig. 1 and Table 2. Fourteen different cascades (pathways from ground state to ground state of the daughter nuclide) were taken into account. The three dominant β transitions are of an allowed nature, while the β transition to the first excited level of 134Ba is of a non-unique second forbidden nature. Its maximum β energy is 1454.26 (33) keV and, thus, a high LS counting efficiency is achieved. Consequently, also the shape factor function of this transition is of minor importance. For all β transitions, the shape factor function was assumed to be C(W) = 1. The overall β emission probability was normalized to 100%.
Table 2

Fourteen cascades were taken into account for the efficiency calculations for CIEMAT/NIST efficiency tracing and the TDCR method

Cascade no.

Probability in %

Components with their energies in parentheses



β 3 (658.39 keV), γ 8 (795.8677 keV), γ 11 (604.7232 keV)



β 2 (415.64 keV), γ 7 (1038.6137 keV), γ 11 (604.7232 keV)



β 4 (1454.26 keV), γ 11 (604.7232 keV)



β 2 (415.64 keV), γ 6 (475.368 keV), γ 9 (563.2457 keV), γ 11 (604.7232 keV)



β 1 (89.06 keV), γ 3 (801.953 keV), γ 10 (1167.968 keV)



β 1 (89.06 keV), γ 4 (1365.1987 keV), γ 11 (604.7232 keV)



β 1 (89.06 keV), γ 1 (326.585 keV), γ 7 (1038.6137 keV), γ 11 (604.7232 keV)



β 2 (415.64 keV), γ 5 (242.746 keV), γ 8 (795.8677 keV), γ 11 (604.7232 keV)



β 1 (89.06 keV), γ 3 (801.953 keV), γ 9 (563.2457 keV), γ 11 (604.7232 keV)



β 2 (415.64 keV), γ 6 (475.368 keV), γ 10 (1167.968 keV)



β 1 (89.06 keV), γ 2 (569.2457 keV), γ 8 (795.8677 keV), γ 11 (604.7232 keV)



β 1 (89.06 keV), γ 1 (326.585 keV), γ 6 (475.368 keV), γ 9 (563.2457 keV), γ 11 (604.7232 keV)



β 1 (89.06 keV), γ 1 (326.585 keV), γ 6 (475.368 keV), γ 10 (1167.968 keV)



β 1 (89.06 keV), γ 1 (326.585 keV), γ 5 (242.746 keV), γ 8 (795.8677 keV), γ 11 (604.7232 keV)

Each cascade comprises one β transition and up to four γ/IC transitions (see also Fig. 2). The overall β emission probability was normalized to 100%

Theoretical coefficients for internal conversion (IC) were calculated with the conversion coefficient calculator BrIcc (v2.3S) using the “frozen orbital” approximation [14].

The CIEMAT/NIST efficiency computations were carried out with a locally developed program which has been used in several previous works (see, e.g., [15, 16]). The program applies procedures to calculate the counting efficiencies of β transitions accompanied by several γ transitions [17]. The ionization quenching function Q(E) was calculated by means of the procedures described in a previous article [15] and an ionization quenching parameter kB = 0.0075 cm/MeV was used. For a 3H efficiency of 50%, the counting efficiency of 134Cs is calculated to be 95.66%. The result of one sample was about 0.15% higher than those from other samples. This was also observed when using the TDCR method and the LS-based coincidence counting method and, thus, it can be excluded that this small bias is due to the CIEMAT/NIST model. This sample was excluded in the further analysis of the data in all presented methods. The individual results of all other samples were in good agreement. The 134Cs counting efficiency ranged from about 90.9–93.7%. This corresponds to a 3H efficiency range from about 20.7% to 33.9%.

In an alternative approach, the efficiency was computed by means of a stochastic model [18] which also proved its worth for other βγ isotopes [19, 20]. The model was even extended to make computations of β branches in coincidence with up to 7 γ/IC transitions possible [21]. The results were found to be in excellent agreement with the outcome from other institutes when measuring 166mHo with a very complex decay scheme [22].

The stochastic model can also be used for efficiency computations needed to analyse the TDCR data. This is a significant advantage since a sophisticated analytical approach to the calculation of counting efficiencies in a TDCR system for radionuclides with such a complex decay scheme is not yet available.

The counting efficiencies are computed in the following way. The number of electrons M i and their energies E il is computed and stored for all simulated decay events (here: N = 4 × 105 per cascade). The triple counting efficiency ε T as a function of the free parameter λ is then given by
$$\varepsilon_{T} (\lambda ) = \sum\nolimits_{i = 1}^{N} {\left\{ {1 - \exp \left[ {\frac{{ - \sum\nolimits_{l = 1}^{{M_{i} }} {E_{il} Q\left( {E_{il} } \right)} }}{3\lambda }} \right]} \right\}}^{3} \left/\vphantom{{\frac{{ - \sum\nolimits_{l = 1}^{{M_{i} }} {E_{il} Q\left( {E_{il} } \right)} }}{3\lambda }}}N \right.$$

The counting efficiency for double coincidences can be computed in a similar manner.

The results of the coincidence counting efficiencies of interest versus the free parameter were stored in files and later combined using the decay probabilities of each cascade (see Table 2) as weighting factors. The procedure is time consuming and must be repeated to study the influence of input parameters like the ionization quenching parameter kB.

The ionization quenching function was computed with the same method as recently described in [23] using an ionization quenching constant kB = 0.0075 cm/MeV.

The results for the activity concentration (Table 3) are identical for the TDCR method and for CIEMAT/NIST efficiency tracing. The uncertainty budgets were determined following the guidance presented in [24] and are shown in Table 1. The dominant contribution to the overall uncertainty is assigned to decay data and the model. Here, we take into account that some simplifications were made to compute the β spectra, since corrections for the atomic exchange effect or screening were neglected. The influence of these effects on the liquid scintillation results was demonstrated for the low-energy β emitter 63Ni [25]. However, in the case of 134Cs, the mean β energy is much higher, which yields a higher counting efficiency and, consequently, lower sensitivity to changes in the adopted β spectrum.
Table 3

Results for the determined activity concentration of the 134Cs solution under study (reference date: 1st June 2015, 0:00 CET)


a in kBq/g

Relative deviation from final result

TDCR (mean of results from 2 counters: 87.14 and 87.03 kBq/g)

87.08 (24)



87.08 (24)


β(PC)-γ coincidence counting

87.22 (22)


β(LS)-γ coincidence counting

87.24 (31)


Final result

87.15 (22)

The final result corresponds to the weighted mean of the four individual results. A relative uncertainty of 0.25% was estimated for the final result. This uncertainty is larger than the internal (0.14%) and the external (0.05%) relative uncertainty of the weighted mean

The influence of the ionization quenching function was investigated by using different parameterization for the electron stopping powers and by changing the value of the kB parameter. The influence of the electron stopping power was found to be low which can again be explained by the rather high counting efficiency. When the kB value is changed from 0.0075 cm/MeV to 0.0110 cm/MeV, the determined activity concentration will only decrease by less than 0.11% for the CIEMAT/NIST method, whereas it will increase by less than 0.2% for the TDCR method. Hence, the mean value of both methods is rather robust against changes in the kB parameter. A similar anti-correlation was also found for other β emitters [19, 26].

Table 3 also contains the individual results from the two TDCR counters. The relative deviation between these results is about 0.13%.

For the CIEMAT/NIST method the efficiency could be calculated either with the analytical calculation or by means of the stochastic model. The relative deviation between the results of both approaches is less than 0.01% when using the same ionization quenching function for both methods.

Prior to this standardization, a stability check indicated that the count rates, corrected for decay, were constant to within ±0.015% for about one week. However, the LS sample series prepared for this work was remeasured after about 6 months in the Wallac LS spectrometer. The resulting activity concentration was found to be 0.3% lower than the initial result and, thus, an uncertainty component of 0.05% was added to the uncertainty budget in Table 1 in order to allow for potential sample instability within the first 2 weeks after sample preparation when the relevant data were taken.

García-Toraño et al. [13] reported on stable LS samples with 134Cs over a period of 25 days when using the LS cocktail HiSafe III which was not tested in our work.

Combined result and submission to the international reference system

The results from the four methods are shown in Table 3. The weighted mean a = (87.15 ± 0.22) kBq/g is taken as PTB’s final result. The uncertainty corresponds to the uncertainty from 4πβ(PC)-γ coincidence counting which is more conservative than usage of the internal or external uncertainty of the weighted mean. It should be noted that the results are correlated since the same balance was used when preparing samples.

The uncertainty of the final result holds for the activity concentration of the diluted solution under study. When regarding the activity concentration of the master solution an additional uncertainty must be taken into account. Another aliquot of the master solution was diluted to prepare a solution which was then sent to the Bureau International des Poids et Mesures (BIPM) for purpose of comparison. Hence, the solution under study and the solution sent to the BIPM are linked via two dilution steps and, consequently, the corresponding relative uncertainty of both dilutions must be taken into account. In addition, the reference date for the BIPM submission differs and hence the uncertainty of the decay correction differs, too. However, the overall relative uncertainty of the activity concentration of the solution sent to the BIPM was found to be 0.25%, which is still rather low.

At this moment, the results of the comparison from the BIPM are still undisclosed.


The good agreement between the results supports usage of any of the four methods used in this work. However, a closer look reveals that the lowest uncertainty is obtained when using the classical 4πβ(PC)-γ coincidence counting. The β efficiency for 4πβ(LS)-γ coincidence counting could be improved by using PE vials instead of glass vials and by reducing the amount of water that is used. In this way, it might be possible to reduce the uncertainty due to the fitting procedures and related extrapolation. This could be investigated in future studies (also for other isotopes). Further improvements could be achieved by a more thorough analysis of the jitter effect in the 4KAM module.

The shutdown of 4πβ(PC)-γ coincidence counting at PTB was justified with the assumption that 4πβ(LS)-γ coincidence counting is a method that requires less workload and, consequently, less manpower. Indeed, the preparation of VYNS samples is rather time consuming whereas LS samples can be prepared in a shorter time. However, the major portion of the overall amount of work is required for the measurements and the analysis of data. The efficiency variation techniques (grey filters vs. absorber foils) and the data analysis require more or less the same efforts. Finally, the overall savings are rather low when replacing 4πβ(PC)-γ with 4πβ(LS)-γ coincidence counting and the price to pay is a somewhat larger uncertainty. Similar studies are ongoing to evaluate the change in uncertainties for other isotopes.

The uncertainties are also somewhat larger when determining the 134Cs activity by means of the TDCR method or CIEMAT/NIST efficiency tracing method. Here, one should keep in mind that these methods have some model dependence, e.g. due to imperfect knowledge about the ionization quenching [24] or the shape of β spectra [25]. This model dependence and, hence, the corresponding uncertainties are rather low in the case of 134Cs and are certainly acceptable for many standardizations. However, for other radionuclides (e.g. isotopes with electron capture) the corresponding uncertainties can be much higher and do not always meet the demands in radionuclide metrology.

It should also be noted that usage of various methods is often an inevitable means to identify unforeseeable problems of measurement techniques or model dependence [27].



We wish to thank S. Hennig for sample preparation.


  1. 1.
    Schönfeld E, Janssen H, Klein R, Hardy JC, Iacob V, Sanchez-Vega M, Griffin HC, Ludington MA (2002) Production of Co-60 sources for high-accuracy efficiency calibrations of gamma-ray spectrometers. Appl Radiat Isot 56:215–221CrossRefGoogle Scholar
  2. 2.
    Nähle O, Kossert K, Cassette P (2010) Activity standardization of 3H with the new TDCR system at PTB. Appl Radiat Isot 68:1534–1536CrossRefGoogle Scholar
  3. 3.
    Grau Malonda A (1999) Free parameter models in liquid scintillation counting. Colección Documentos CIEMAT, CIEMAT. ISBN 84-7834-350-4Google Scholar
  4. 4.
    Broda R, Cassette P, Kossert K (2007) Radionuclide metrology using liquid scintillation counting. Metrologia 44:S36–S52CrossRefGoogle Scholar
  5. 5.
    Bé MM, Chisté V, Dulieu C, Mougeot X, Chechev VP, Kondev FG, Nichols AL, Huang X, Wang B (2013) Table of radionuclides (Vol. 7A = 14 to 245). Monographie BIPM-5 Vol. 7, Bureau International des Poids et Mesures, Sèvres, ISBN-13 978-92-822-2248-5Google Scholar
  6. 6.
    Nähle O, Zhao Q, Wanke C, Weierganz M, Kossert K (2014) A portable TDCR system. Appl Radiat Isot 87:249–253CrossRefGoogle Scholar
  7. 7.
    Schönfeld E, Janssen H (1994) Precise measurement of dead time. Nucl Instrum Methods A339:137–143CrossRefGoogle Scholar
  8. 8.
    Smith D (1978) Improved correction formulae for coincidence counting. Nucl Instrum Methods 152:505–519CrossRefGoogle Scholar
  9. 9.
    ICRU Report 52 (1994) Particle counting in radioactivity measurements. Bethesda, Md., ISBN 0-913394-51-3Google Scholar
  10. 10.
    Bouchard J, Cassette P (2000) MAC3: an electronic module for the processing of pulses de-livered by a three photomultiplier liquid scintillation counting system. Appl Radiat Isot 52:669–672CrossRefGoogle Scholar
  11. 11.
    Rytz A (1982) Activity measurement of a solution of 134Cs: report on an international comparison. Nucl Instrum Methods 192:427–431CrossRefGoogle Scholar
  12. 12.
    Yunoki A, Kawada Y, Hino Y (2016) Improvements of the standardization of 134Cs by the critical window setting for 605 keV photopeak. Appl Radiat Isot 109:374–377CrossRefGoogle Scholar
  13. 13.
    García-Toraño E, Rodríguez Barquero L, Roteta M (2002) Standardization of 134Cs by three methods. Appl Radiat Isot 56:211–214CrossRefGoogle Scholar
  14. 14.
    Kibédi T, Burrows TW, Trzhaskovskaya MB, Davidson PM, Nestor CW Jr. (2008) Evaluation of theoretical conversion coefficients using BrIcc. Nucl Instrum Methods.589, 202–229. And: Accessed June 2015
  15. 15.
    Kossert K, Schrader H (2004) Activity standardization by liquid scintillation counting and half-life measurements of 90Y. Appl Radiat Isot 60:741–749CrossRefGoogle Scholar
  16. 16.
    Nähle O, Kossert K, Klein R (2008) Activity standardization of 22Na. Appl Radiat Isot 66:865–871CrossRefGoogle Scholar
  17. 17.
    Verdecia OP, Kossert K (2009) Activity standardization of 131I at CENTIS-DMR and PTB within the scope of a bilateral comparison. Appl Radiat Isot 67:1099–1103CrossRefGoogle Scholar
  18. 18.
    Carles AG (2007) MICELLE, the micelle size effect on the LS counting efficiency. Comput Phys Commun 176:305–317CrossRefGoogle Scholar
  19. 19.
    Kossert K, Nähle OJ, Ott O, Dersch R (2012) Activity determination and nuclear decay data of 177Lu. Appl Radiat Isot 70:2215–2221CrossRefGoogle Scholar
  20. 20.
    Kossert K, Nähle OJ (2014) Activity determination of 59Fe. Appl Radiat Isot 93:33–37CrossRefGoogle Scholar
  21. 21.
    Kossert K, Cassette P, Carles AG, Jörg G, Gostomski CL, Nähle O, Wolf C (2014) Extension of the TDCR model to compute counting efficiencies for radionuclides with complex decay schemes. Appl Radiat Isot 87:242–248CrossRefGoogle Scholar
  22. 22.
    Kossert K, Altzitzoglou T, Auerbach P, Bé MM, Bobin Ch, Cassette P, García-Toraño E, Grigaut-Desbrosses H, Isnard H, Lourenço V, Nähle O, Paepen J, Peyrés V, Pommé S, Rozkov A, Sanchez-Cabezudo AI, Sochorová J, Thiam C, Van Ammel R (2014) Results of the EURAMET.RI(II)-K2.Ho-166 m activity comparison. Metrologia 51:06022CrossRefGoogle Scholar
  23. 23.
    Kossert K, Grau Carles A (2010) Improved method for the calculation of the counting efficiency of electron-capture nuclides in liquid scintillation samples. Appl Radiat Isot 68:1482–1488CrossRefGoogle Scholar
  24. 24.
    Kossert K, Broda R, Cassette P, Ratel G, Zimmerman B (2015) Uncertainty determination for activity measurements by means of the TDCR method and the CIEMAT/NIST efficiency tracing technique. Metrologia 52:S172–S190CrossRefGoogle Scholar
  25. 25.
    Kossert K, Mougeot X (2015) The importance of the beta spectrum calculation for accurate activity determination of 63Ni by means of liquid scintillation counting. Appl Radiat Isot 101:40–43CrossRefGoogle Scholar
  26. 26.
    Nähle O, Kossert K (2011) Comparison of the TDCR method and the CIEMAT/NIST method for the activity determination of beta emitting nuclides. LSC2010, Advances in LS spectrometry: proceedings of the 2010 international conference on LS spectrometry, Paris 6–10 September 2010, Cassette P (ed), Radiocarbon, The University of Arizona, Tucson. ISBN 978-0-9638314-7-7, 313-320Google Scholar
  27. 27.
    Pommé S (2016) When the model doesn’t cover reality: examples from radionuclide metrology. Metrologia 53:S55–S64CrossRefGoogle Scholar

Copyright information

© Akadémiai Kiadó, Budapest, Hungary 2017

Authors and Affiliations

  1. 1.Physikalisch-Technische Bundesanstalt (PTB)BraunschweigGermany

Personalised recommendations