Results from the Zurich sea water (ZSW) intercomparison study for U- and Pu-isotopes

Results from the Zurich sea water (ZSW) intercomparison study for Uand Pu-isotopes ☆ Marcus Christl * , Habacuc Perez-Tribouillier , Philip Gautschi Laboratory of Ion Beam Physics, ETH Zurich 8093 Zurich, Switzerland ABSTRACT Approximately 350L of sea water samples were collected to prepare a large volume (Zurich Sea Water, ZSW) intercomparison sample for 236 U and other actinides. Here we report the results of the intercomparison study for Uand Pu-isotopes. ZSW was processed in six different chemistry labs and measured on five AMS systems and one MC-ICP-MS system, respectively. The analysis of the reported U-isotopic data shows that the scatter, indicated by the one sigma uncertainty for a single measurement, of the 236 U/ 238 U ratios and the 236, 238 U-concentrations is 5 % and 3 %, respectively. A Chi-squared analysis shows that the external error (lab-scatter) is 1–2 % and 3 % for the 236, 238 U-concentrations and the 236 U/ 238 U ratios, respectively. The fact that ZSW has been characterized by many different (A)MS labs makes it a valuable internal standard for quality control for the analysis of anthropogenic Uand Pu-isotopes in seawater. 1. Introduction The long-lived anthropogenic U-isotope 236 U (T ½ =23.4 Myr) and other actinide isotopes (e.g. Pu-isotopes and 233 U) are increasingly used in oceanographic studies to trace ocean currents and determine mixing ratios of water masses, or to identify the source of an anthropogenic contamination [1–6]. This success is supported by substantial developments in accelerator mass spectrometry (AMS) that allow the fast, precise, and almost background free determination of Puand U-isotopes in environmental samples [7–10]. The success of compact, low energy AMS systems in actinide research is not limited to U and Pu isotopes alone, it also triggered several studies relying on the highly sensitive and selective detection of the minor actinides (e.g. Np-, Am-, and Cm-isotopes) in environmental and biological samples [11–17]. Worldwide, the number of AMS systems capable of performing actinide measurements at environmental levels is increasing [18–20] and only recently, the ability to perform competitive analyses of 236 U and other actinides with the compact 300 kV AMS system MILEA has been reported [21]. With the number of AMS systems increasing the question arises of how accurate those measurements are, particularly when performed on different systems, applying different radiochemical methods, and using different internal standards for normalization. Here we report on the results of an intercomparison experiment for 236 U and Pu-isotopes carried out with a large volume sea water sample. 2. Material and methods 2.1. Sample treatment and preparation (pre shipment) The sea water used for this experiment was collected in 2011 during the Fenice 2011 cruise [22] which took place from 25 October to 11 November 2011 in the Tyrrhenian Sea (Fig. 1). A total of about 350L of sea water was collected in 35 ten-liter plastic canisters, intended for 236 U/ 238 U analysis at ETH Zurich. Unfortunately, the sample labels were not readable any more when they arrived at ETH Zurich. Instead of discarding the samples we decided to combine them and use the sea water as a large volume intercomparison sample for 236 U and the actinides. The sea water was pumped through 0.45 μ m membrane filters using a peristaltic pump. Filters were exchanged after every second sample and the filtered sea water was collected in a 500L plastic tank that has been previously rinsed with deionized water. By adding concentrated nitric acid (p.a.) the pH value was adjusted to 3.0 ±0.5 using pH indicator strips. After some months of equilibration pH was checked again. Then the seawater was filled into 35 ten-liter plastic cubitainers before shipment to the participating labs. The sample was named Zurich Sea Water intercomparison sample (ZSW). ☆ This article is part of a special issue entitled: ‘Special Issue of AMS16’ published in Nuclear Inst. and Methods in Physics Research, B. * Corresponding author. E-mail address: [email protected] (M. Christl). Contents lists available at ScienceDirect Nuclear Inst. and Methods in Physics Research, B journal homepage: www.elsevier.com/locate/nimb https://doi.org/10.1016/j.nimb.2025.165750 Received 11 March 2025; Received in revised form 30 April 2025; Accepted 10 May 2025 Nuclear Instruments and Methods in Physics Research B 565 (2025) 165750 Available online 19 May 2025 0168-583X/© 2025 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). 2.2. Sample processing and analysis (post shipment) The primary goal of the intercomparison study was to collect information on the 236 U/ 238 U ratio and the 236 U and 238 U concentrations of ZSW. Nevertheless, the labs were asked to provide additional information on concentration of other actinides, if possible. Depending on request, one or two ZSW-cubitainers were sent out to each participating laboratory. Each participant was asked to apply their standard sample processing procedures for uranium and other actinides. Similarly, the AMS labs were asked to use their standard instrumental setups, measurement routines and to employ their own in-house reference materials. 2.3. Data evaluation and analysis ZSW has been processed in six different chemistry labs and actinide results were reported from five different AMS systems and from one Multi Collector Inductively Coupled Plasma Mass Spectrometry system (MC-ICP-MS). Results from the different participants and instruments are presented in an anonymized way. In all figures, different labs are indicated by different colors and different (A)MS systems are identified by different symbols. Several combinations of colors and symbols are possible as one laboratory can prepare samples that are measured on different AMS systems or one AMS lab can measure samples prepared in different labs. For those labs that agreed, the identities of the prep labs and (A)MS facilities are revealed in Table S2 in the supplementary information. In total, 26 data reports were returned, while one data report may include information on several isotopes or isotopic ratios e.g., on the 236 U/ 238 U ratio, the concentrations of 236 U and 238 U, etc. Throughout the paper, the term isotopic ratio always refers to an atom ratio. In all figures of this paper, each data point represents one reported result. Reported data points were not combined or averaged. For each set of isotope concentrations or isotopic ratios reported by the different labs, the error weighted mean and the standard deviation of the weighted mean was calculated using Eq (1) and Eq (2): Y= ∑iyi•1 σ 2 i ∑i1 σ 2 i (1) ΔY= ∑i(yi−Y)2•1 σ 2 i ∑i1 σ 2 i (2) Accordingly, the reduced Chi-Squared value for each data set is given by: X2 red = ∑i(yi−Y)2•1 σ 2 i (n−1)(3) where n is the total number of reported data points (y i ) for each isotope or isotopic ratio and i indicates one single reported result including its one sigma uncertainty. The upper and lower limits for X2 red were calculated for each data set using the X2-distribution with (n-1) degrees of freedom using a 95 % confidence interval. In case the scatter of the data set (and therefore the X2 red value) exceeded the range calculated from the X2-distribution, it was assumed that the dataset contains some additional uncertainty e.g., introduced by the different labs, due to, for example, different chemical procedures, different AMS-techniques or standards, etc. In this case, an external scatter was added to each data point according to EQ (4): Fig. 1. Cruise track of the Fenice 2011 cruise (. Source: Google Maps) M. Christl et al. Nuclear Inst. and Methods in Physics Research, B 565 (2025) 165750 2 σ 2 i,tot = σ 2 i+ (yi• σ ext%)2(4) The relative external error ( σ ext%, Table S1) was varied until the X2 red value (EQ (3) equaled the upper value of the according X2-distribution. We note that this implies the assumption that the external scatter is represented by an additional (normal distributed) statistical fluctuation and not by a systematic error (e.g., due to different standards used for normalization). 3. Results and Discussion 3.1. Results for the 236 U/ 238 U (and 233 U/ 236 U) ratios All participating labs reported results on the 236 U/ 238 U ratio of ZSW. In total, 26 data points from 6 different labs, 5 AMS systems, and 1 MCICP-MS were reported. One AMS lab, however, indicated that their result is indicative only as they were still optimizing their laboratory protocols for 236 U and sample preparation suffered from low yield. These results (magenta triangle) for 236 U/ 238 U (and for the concentrations of 236 U and 238 U) are shown for completeness only, and are not included in the calculation of the average value nor the Chi-squared distribution (Fig. 2 and Table S1). Also, the group providing 236 U/ 238 U results measured with MC-ICP-MS considered their setup experimental at that stage and reported that the results might be systematically biased because they strongly rely on the performance and calibration of the secondary electron multiplier detector. We therefore decided not to include those results in the calculations as well (open purple hexagons, Fig. 2). We, nevertheless want to mention that the ICP-MS lab only used 1 L of sea water to perform their measurement and reported their data with a remarkable small uncertainty. This result demonstrates that MC-ICP-MS in general has the capability of analyzing 236 U/ 238 U ratios in sea water at a level of 10 -9 [23]. The remaining 22 data points from 5 labs measured on 5 different AMS systems were all included in the following calculations (Fig. 2 and Table S1). Based on these data, the average value for the 236 U/ 238 U ratio is (1689 ±77) x10 -12 . The X2 red value of 2.86 for this result indicates that the reported uncertainties do not fully explain the scatter of the data set based on a 95 % confidence level. The statistical analysis indicates that an external scatter of 2.0 % is needed to explain the distribution of the data. Taking into account that the typical uncertainty of a 236 U/ 238 U measurement in a surface sea water sample with AMS is 3–5 % this is an acceptable result. It indicates that the contributing 5 labs and 5 AMS systems are able to provide consistent results on 236 U/ 238 U in sea water that agree within 5 % or less. Three labs provided additional information on the 233 U/ 236 U ratio of ZSW (Fig. 3, Table S1). The sample was analyzed on two different AMS systems only, therefore statistics on the 233 U/ 236 U results is limited. Nevertheless, the data from Lab 1 shows a clear discrepancy from the results reported by Lab 2 and Lab 3. In this lab, the intercomparison sample was processed and measured together with other sea water samples that were spiked with about 1 pg of 233 U, each. The fact that the two results are significantly higher than the ones reported by Lab2 and Lab 3 indicates that the samples processed in Lab 1 were probably contaminated by the 233 U spike either during chemical processing in the lab or suffered from cross-contamination during the AMS measurement. An alternative explanation would involve surviving 232 ThH 3+ molecules due to insufficiently high stripper pressure [24]. The results on 233 U/ 236 U reported from Lab 1 were therefore excluded from the calculation of the average. For the remaining six data points from two different labs, measured on one AMS system an average value of 233 U/ 236 U =(0.67 ±0.07) x10 -2 was calculated. The statistical analysis of the results indicates that the reported uncertainties are in accordance with the scatter of the data points. Therefore, no additional external uncertainty is needed to explain the scatter of the data points. It has to be noted that the determination of 233 U in sea water is almost always limited by counting statistics and uncertainties of about 10 % are common for AMS analyses of 233 U/ 236 U. Fig. 2. Reported results for 236 U/ 238 U of ZSW. Datapoints marked with a star (*) were not included in the average. M. Christl et al. Nuclear Inst. and Methods in Physics Research, B 565 (2025) 165750 3 Fig. 3. Reported results for 233 U/ 238 U of ZSW. Data points marked with a star (*) were not included in the average (see text for details). Fig. 4. Reported results for 236 U-concentration of ZSW. M. Christl et al. Nuclear Inst. and Methods in Physics Research, B 565 (2025) 165750 4 Since the results were produced by only one AMS lab no statement can be made about inter-comparability of 233 U/ 236 U between different AMS systems. The average 233 U/ 236 U ratio reported in this study, nevertheless, may serve as indicative value for future AMS analyses and intercomparisons. Based on the average 236 U concentration of ZSW presented in the next section, an average 233 U concentration of (1.0 ± 0.1) x10 5 at/kg in ZSW is estimated. 3.2. Results for the 236 Uand 238 U-concentrations Eighteen results were reported for the concentrations of 236 U and 238 U, respectively (Fig. 4, Fig. 5). The number of reported data points is smaller compared to 236 U/ 238 U because not all labs (i) used a spike (typically 233 U) that allows determining the concentration of these isotopes by isotopic dilution or (ii) determined the concentration of 238 U separately in an aliquot of the sample (typically by ICP-MS). In total, three labs reported 236 U and 238 U concentration results that were measured on four different AMS systems. The reported data on the 236 U concentration scatters by 3.5 % around its average value of 15.5 x10 6 at/kg. The statistical analysis (Table S1, Fig. 4) shows that an external uncertainty of 1.8 % is needed to explain the scatter of the data points compared to their reported uncertainties. This result is very similar to the 236 U/ 238 U ratios where the external uncertainty was estimated to 2 %. Also in this case, the results indicate that the participating AMS labs are able to produce consistent data for 236 U concentrations typical for surface sea water within an uncertainty of 4 % or better. The reported data on the 238 U concentration of ZSW scatters by 2.7 % around the average value of 3.67 μ g/kg and statistical analysis indicates an external error of 1 % (Fig. 5, Table S1). For the determination of the 238 U concentration by the different participating laboratories two independent methods were applied. While Lab 1 and Lab 4 used a 233 U spike to determine both, 236 U and 238 U concentrations by isotopic dilution, Lab 2 analyzed 238 U by ICP-MS in an aliquot of the sample. The fact that the reported 238 U concentrations for ZSW determined by two independent methods agree well, indicates that the reported results are not only precise but also accurate within an uncertainty of 3 % or less. 3.3. Results for Pu-isotopes In addition to the information of U-isotopes three labs reported results on 239 Pu-, and 240 Pu-concentrations and on the 240 Pu/ 239 Pu ratio (Fig. 6, Table S1. The results were produced on three different AMS systems and average values of (1.19 ±0.06) x10 6 at/kg and (6.17 ±0.19) x10 6 at/ kg are calculated for 240 Pu and 239 Pu, respectively. All labs added a 242 Pu spike to the sample in order to determine the concentration of 239 Pu and 240 Pu by isotopic dilution. The statistical analysis shows that no or only negligible (<1%) external uncertainty has to be added to the results. One lab also analyzed the sample without adding a spike. Therefore, for this preparation only information on the isotopic ratio was reported. This preparation, however, suffered from a low chemical yield and low counting rates during the measurement. Therefore, the reported uncertainty is relatively large. In a similar way as for Pu concentrations, for the ratio measurements of 240 Pu/ 239 Pu no external uncertainty is needed to explain the scatter of the data points around their average value of (0.19 ±0.01) at/at. In summary, the results for Puisotopes of the intercomparison sample reported by three labs and measured on three AMS systems indicate that all participating labs are able to produce consistent results for Pu-concentrations and isotopic ratios in sea water with 3–5 % uncertainty. 4. Summary and Conclusion The main goal of this study was to perform an intercomparison study for the analysis of 236 U/ 238 U in sea water. For this purpose, a large volume Zurich Sea Water (ZSW) intercomparison sample was produced and sent out to all participating labs. Results on 236 U/ 238 U for the ZSW sample have been returned from 6 different labs and measurements were performed on 5 different AMS systems and on one MC-ICP-MS system, Fig. 5. Reported results for 238 U-concentration of ZSW. Data points marked with a star (*) were not included in the average (see text for details). M. Christl et al. Nuclear Inst. and Methods in Physics Research, B 565 (2025) 165750 5 respectively. Additional information was reported by the participating labs regarding the concentrations of 236 U, 238 U, 239 Pu, and 240 Pu and on their isotopic ratios ( 233 U/ 236 U and 239 Pu/ 240 Pu). The results collected from all participating labs for 236 U-, 238 U-, 239 Pu-, 240 Pu-concentrations and 239 Pu/ 240 Puand 236 U/ 238 U-isotopic ratios agree well within 3–5 % uncertainty. A Chi-squared analysis of the scatter of the data shows that the external scatter for all concentrations and isotopic ratios is equal or below 2 %. This shows that all participating labs are able to perform precise 233, 236, 238 Uand 239, 240 Pu-isotope analyses (including their isotopic ratios) in ocean surface water samples with an uncertainty of 3–5 %. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The ETH Zurich Laboratory of Ion Beam Physics (LIP) is partially funded by its consortium partners PSI, EMPA, EAWAG, and WSL. HPT and MC received partial funding from the European Partnership on Metrology, co-financed from the European Union’s Horizon Europe Research and Innovation Program and by the Participating States is partially funded by the European Partnership on Metrology: Funder ID: 10.13039/100019599, Grant number: 21GRD09 MetroPOEM. We further want to thank Maxi Castrillejo, Nuria Casacuberta, (ETH Zurich, Switzerland), Elena Chamizo, Mercedes L´ opez-Lora (CNA Spain), Karin Hain (VERA, Austria), Jixin Qiao, Mu Lin (DTU, Denmark), Huei-Ting (Tina) Lin (National Taiwan University) for processing and analyzing samples in their labs. We additionally want to thank Joze Kotnik (Jozef Stefan Institute, Slovenia), Arne Bratkiˇ c (University of Li` ege, Belgium), and Laura Fantozzi (CNR, Italy) for sampling and sample handling during the Fenice cruise. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.nimb.2025.165750. References [1] K. Hain, et al., 233 , Nat. Commun. 11 (2020) 1275. [2] J. Qiao, et al., An unknown source of reactor radionuclides in the Baltic Sea revealed by multi-isotope fingerprints, Nat. Commun. 12 (2021) 823. [3] M. Christl, et al., A depth profile of uranium-236 in the Atlantic Ocean, Geochim. Cosmochim. Acta 77 (2012) 98. [4] A. Sakaguchi, et al., Uranium-236 as a new oceanic tracer: A first depth profile in the Japan Sea and comparison with caesium-137, Earth Planet. Sci. Lett. 333–334 (2012) 165. [5] N. Casacuberta, et al., A first transect of 236 U in the North Atlantic Ocean, Geochim. Cosmochim. Acta 133 (2014) 34. [6] M. Villa-Alfageme, et al., Distribution of 236 U in the U.S. GEOTRACES Eastern Pacific Zonal Transect and its use as a water mass tracer, Chem. Geol. 517 (2019) 44. [7] M. Christl, et al., Status of 236 U analyses at ETH Zurich and the distribution of 236 U and 129 I in the North Sea in 2009, NIM B 361 (2015) 510. [8] C. Vockenhuber, et al., The potential of He stripping in heavy ion AMS, Nucl. Instrum. Methods Phys. Res., Sect. B 294 (2013) 382. [9] C. Vockenhuber, et al., Accelerator mass spectrometry of 236 U at low energies, Nucl. Instrum. Methods Phys. Res., Sect. B 269 (2011) 3199. [10] P. Steier, et al., Natural and anthropogenic 236 U in environmental samples, Nucl. Instrum. Methods Phys. Res., Sect. B 266 (2008) 2246. [11] A. Wallner, et al., 60 Fe and 244 Pu deposited on Earth constrain the r-process yields of recent nearby supernovae, Science 372 (2021) 742. [12] C. Tighe, et al., Local and global trace plutonium contributions in fast breeder legacy soils, Nat. Commun. 12 (2021) 1381. [13] M. L´ opez-Lora, et al., 236 , Sci. Total Environ. 765 (2021) 142741. [14] Y. Wu, et al., Ultrasensitive Analytical Method for Direct Search of Primordial 244 Pu in Bastnaesite, ACS Earth Space Chem. 5 (2021) 1316. [15] M. Christl, et al., A novel chronometry technique for dating irradiated uranium fuels using Cm isotopic ratios, J. Radioanal. Nucl. Chem. 322 (2019) 1611. [16] K. Gückel, et al., Scavenged 239 Pu, 240 Pu, and 241 Am from snowfalls in the atmosphere settling on Mt. Zugspitze in 2014, 2015 and 2016, Scientific Reports 7 (2017) 11848. Fig. 6. Reported results for 239 Pu and 240 Pu concentrations and the 240 Pu/ 239 Pu ratio of ZSW. M. Christl et al. Nuclear Inst. and Methods in Physics Research, B 565 (2025) 165750 6 [17] X. Dai, et al., Determination of Attoto Femtogram Levels of Americium and Curium Isotopes in Large-Volume Urine Samples by Compact Accelerator Mass Spectrometry, Anal. Chem. 88 (2016) 2832. [18] M.A.C. Hotchkis, et al., Actinides AMS on the VEGA accelerator, Nucl. Instrum. Methods Phys. Res., Sect. B 438 (2019) 70. [19] E. Chamizo, et al., Measurement of 236 U on the 1 MV AMS system at the Centro Nacional de Aceleradores (CNA), Nucl. Instrum. Methods Phys. Res., Sect. B 358 (2015) 45. [20] S.R. Winkler, et al., He stripping for AMS of 236 U and other actinides using a 3MV tandem accelerator, Nucl. Instrum. Methods Phys. Res., Sect. B 361 (2015) 458. [21] M. Christl, et al., 236 , NIM B 534 (2023) 61. [22] M.G. Nerentorp Mastromonaco, et al., Seasonal and spatial evasion of mercury from the western Mediterranean Sea, Mar. Chem. 193 (2017) 34. [23] H.T. Lin, et al., 236 U/ 238 U Analysis of Femtograms of 236 U by MC-ICPMS, Anal. Chem. 93 (2021) 8442. [24] J. Lachner, et al., Existence of triply charged actinide-hydride molecules, Phys. Rev. A 85 (2012) 022717. M. Christl et al. Nuclear Inst. and Methods in Physics Research, B 565 (2025) 165750 7