Analysis of 236U and plutonium isotopes, 239,240Pu, on the 1 MV AMS system at the Centro Nacional de Aceleradores, as a potential tool in oceanography

Analysis of 236 U and plutonium isotopes, 239,240 Pu, on the 1 MV AMS system at the Centro Nacional de Aceleradores, as a potential tool in oceanography Elena Chamizo a, ⇑ , Mercedes López-Lora a , María Villa b,c , Núria Casacuberta d , José María López-Gutiérrez a,e , Mai Khanh Pham f a Centro Nacional de Aceleradores (Universidad de Sevilla, Consejo Superior de Investigaciones Científicas, Junta de Andalucía), Thomas Alva Edison 7, 41092 Seville, Spain b Departamento de Física Aplicada II, Universidad de Sevilla, Av. Reina Mercedes 4A, 41012 Seville, Spain c Servicio de Radioisótopos, Centro de Investigación, Tecnología e Innovación, Universidad de Sevilla, Av. Reina Mercedes 4B, 41012 Seville, Spain d Laboratory of Ion Beam Physics, ETH Zürich, Otto-Stern-Weg 5, CH-8093 Zürich, Switzerland e Departamento de Física Aplicada I, Escuela Universitaria Politécnica, Universidad de Sevilla, Virgen de África 7, 41011 Seville, Spain f IAEA-Environment Laboratories, Monte Carlo 98000, Monaco article info Article history: Received 26 November 2014 Received in revised form 18 February 2015 Accepted 20 February 2015 Available online 11 March 2015 Keywords: Plutonium isotopes U236 Seawater Sediments Oceanography abstract The performance of the 1 MV AMS system at the CNA (Centro Nacional de Aceleradores, Seville, Spain) for 236 U and 239,240 Pu measurements has been extensively investigated. A very promising 236 U/ 238 U abundance sensitivity of about 3 10 11 has been recently achieved, and background figures for 239 Pu of about 10 6 atoms were reported in the past. These promising results lead to the use of conventional low energy AMS systems for the analysis of 236 U and 239 Pu and its further application in environmental studies. First 236 U results obtained on our AMS system for marine samples (sediments and water) are presented here. Results of two new IAEA reference materials (IAEA-410 and IAEA-412, marine sediments from Pacific Ocean) are reported. The obtained 236 U/ 239 Pu atom ratios, of 0.12 and 0.022, respectively, show a dependency with the contamination source (i.e. local fallout from the US tests performed at the Bikini Atoll and general fallout). The results obtained for a third IAEA reference material (IAEA- 381, seawater from the Irish Sea), are also presented. In the following, the uranium and plutonium isotopic compositions obtained on a set of 5 intercomparison seawater samples from the Arctic Ocean provided by the ETH Zürich are discussed. By comparing them with the obtained results on the 600 kV AMS facility Tandy at the ETH Zürich, we demonstrate the solidity of the CNA technique for 236 U/ 238 U determinations at, at least, 7 10 10 level. Finally, these results are discussed in their environmental context. Ó2015 Elsevier B.V. All rights reserved. 1. Introduction 236 U(T 1/2 = 23.4 My) and 239 Pu (T 1/2 = 24,110 y), alpha emitters, are present in the environment due to their natural and anthropogenic production. Their natural way of production is the thermal neutron activation of the two major uranium isotopes, 235 U (T 1/2 = 703.8 My, 0.72% isotopic abundance) and 238 U(T 1/2 = 4468 My, 99.27% isotopic abundance), respectively. The neutron associated to the cosmic radiation and the so-called nucleogenic component (i.e. neutrons produced by the spontaneous fission of 238 Uorin the ( a ,n) reactions on light elements) are responsible for these processes. The nucleogenic production is especially enhanced in uranium rich minerals. Indeed, the maximum natural 236 U/ 238 U and 239 Pu/ 238 U atom ratios have been measured in these samples, ranging from 10 12 to 3 10 10 and from 2 10 13 to 6 10 12 , respectively [1]. The minimum 236 U/ 238 U atom ratio, of 10 14 ,is anticipated for deeper layers of rocks. The estimated natural inventory of 236 U on Earth is 35 kg [2]. Since the onset of the nuclear energy in the 1940s, the anthropogenic production has overwhelmed those natural inventories. Due to its fissile nature, 239 Pu was produced in huge amounts for nuclear weapons and later on for nuclear reactors, in this last case usually combined with the naturally fissile radionuclide 235 U. Among the different associated by-products, 236 U and 240 Pu (T 1/2 = 6564 y, alpha emitter) deserve special attention. 240 Pu is mainly produced by neutron activation of 239 Pu; for 236 U, the nuclear reaction 238 U(n,3n) is, in addition to the neutron activation of 235 U, an important way of production when fast http://dx.doi.org/10.1016/j.nimb.2015.02.066 0168-583X/Ó2015 Elsevier B.V. All rights reserved. ⇑ Corresponding author. Nuclear Instruments and Methods in Physics Research B 361 (2015) 535–540 Contents lists available at ScienceDirect Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb neutrons are involved (i.e. thermonuclear atomic weapons). Therefore, the 240 Pu/ 239 Pu and the 236 U/ 238 U are in close relationship with the nature of the neutron component in a specific nuclear device. Whereas in nuclear detonations they inform about the original design of the gadget and the yield of the detonation; in nucle- ar reactors, these ratios report about the nature of the fuel (e.g. type of reactor) and the burning times. A comprehensive set of 240 Pu/ 239 Pu results has been reported which shows this dependency. For instance, the local fallout at the Trinity Test Site (New Mexico, USA), where the first American nuclear gadgets were tested, is characterized by a 240 Pu/ 239 Pu atom ratio of 1.3 10 2 , one of the lowest ones reported in the literature [3]. At sites influenced by the Chernobyl accident (1986 in the former Soviet Union, now Ukraine), 240 Pu/ 239 Pu atom ratios as high as 0.57 have been reported [4]. However, there is not information about the source dependency of the 236 U/ 238 U ratio, due to the dilution effect of the ubiquitous 238 U. Therefore, environmental studies of 236 U require the use of inventories to identify its source. The baseline level is associated to the 236 U that was released worldwide due to the atmospheric nuclear tests, with estimated global inventories ranging from 900 to 1400 kg [5,6]. On the other hand, the Sellafield Nuclear Reprocessing Plant (NRP), in United Kingdom, and La Hague, in Northern France, have released approximately between 115 and 250 kg of 236 U to the Northern Seas since 1960, and about 190 kg of 239 Pu [7]. The 236 U released to the environment represents only about 0.1% of the total budget that is estimated to be stored in nuclear power plants worldwide [6]. In the last years, 236 U has been established as a new isotopic tracer in oceanography due to its conservative nature in seawater. It is worth to mention the studies focused on the Pacific Ocean after the Fukushima accident [8] and on the Japan Sea [9], and the 236 U studies performed by the ETH Zürich group on the North Atlantic Ocean [6] and the North Sea [10]. Moreover, 236 U results in marine sediments have been reported in [9,11]. The development of this new application field has been possible thanks to the low 236 U/ 238 U abundance sensitivity provided by Accelerator Mass Spectrometry (AMS), orders of magnitude better than the provided ones by conventional MS techniques. The key is the removal of the uranium molecular isobars (mainly 235 U 1 H) in the so-called stripping process, which occurs in the terminal of a tandem accelerator. With an appropriate ion-optics design of the kinematic filters, the background caused by the neighboring 235,238 U molecular fragments can be minimized. Finally, the use of nuclear detectors in a final stage allows the discrimination of the ions of interest. The best 236 U/ 238 U background values, of about 10 13 , have been achieved on the 14 MV UD AMS facility at the Australian National University (ANU, Canberra, Australia) and on the 3 MV AMS facility at the Vienna Environmental Laboratories (VERA, Vienna, Austria) [1,12]. The 600 kV compact AMS facility at the ETH Zürich features very similar abundance sensitivity, despite the energy gap and its compact design [13]. Recently, it has been demonstrated that an abundance sensitivity of about 310 11 can be achieved on the 1 MV AMS facility at the Centro Nacional de Aceleradores (CNA, Seville, Spain) [14]. In this context, a potential complementary approach would be the study of the 236 U/ 239 Pu ratio. A very reliable range for the global fallout 236 U/ 239 Pu ratio of 0.19–0.23 has been obtained from the study of soil cores from different areas of the Northern Hemisphere, far from local actinides sources other than global fallout [5,15,16]. These ratios were deduced using the integrated inventories of 239 Pu and 236 U along the cores, due to the matrix-dependent profiles of both radionuclides. It is important to note the different physical–chemical behavior of uranium and plutonium, which might be particularly important for non steady state systems, such as seawater masses. Indeed, in surface seawater, the 236 U presence is enhanced by at least a factor of 2 over 239 Pu, due to the conservative nature of uranium and to the particle-reactive behavior of plutonium, which can be further strengthen in areas with a remarkable biological productivity [17]. For marine sediments, it can be deduced, from published 236 U and 239,240 Pu data, a few results for reference materials provided by the International Atomic Energy Agency (IAEA) and affected by different local sources. For instance, it can be estimated a 236 U/ 239 Pu atom ratio of about 0.09 for sample IAEA-135 [18,19]. This sample was collected at the Irish Sea and is clearly influenced by the discharges from the Sellafield Nuclear Reprocessing Plant (NRP). In this work, we explore the possibility that offers the 1 MV AMS system at the CNA for 236 U and 239 Pu studies in Oceanography. For this purpose, two groups of samples have been studied: (i) three reference materials provided by the IAEA: two new marine sediments from the Pacific Ocean (IAEA-410 and IAEA-412) and one seawater sample from the Irish Sea (IAEA- 381), and (ii) five seawater samples from the Arctic Ocean provided by the ETH Zürich group. In the case of the seawater samples, our results will be directly compared to the obtained ones by the ETH Zürich group as an intercomparison exercise. The obtained results for the two new IAEA reference materials will add unique information in specific environmental compartments about the rarely studied source-dependent 236 U/ 239 Pu atom ratio. The obtained results will be discussed in their corresponding environmental Table 1 Obtained 236 U and 239,240 Pu results for the different aliquots of the IAEA-410 and IAEA-412 samples studied on the 1 MV AMS facility at the CNA. The ones leached with HF have been highlighted ( ⁄ ). The 236 U/ 238 U atom ratios have been estimated considering the 236 U AMS results (leachable fraction) and the 238 U data (total fraction) given in the IAEA data evaluation reports [20,21], as explained in the text. The 239,240 Pu results have been also included in these IAEA reports. Sample Description 236 U(10 6 atoms/g) 236 U/ 238 U(10 9 atom ratio) 239+240 Pu (mBq/g) 240 Pu/ 239 Pu (atom ratio) 236 U/ 239 Pu (atom ratio) 410–1 IAEA-410 sediment from Bikini Atoll (11°26 0 N, 164°52 0 E) 266 ± 17 141 ± 5 6.32 ± 0.14 0.273 ± 0.011 0.112 ± 0.008 410–2 194 ± 22 103 ± 8 4.74 ± 0.11 0.284 ± 0.012 0.111 ± 0.013 410–5 223 ± 14 118 ± 4 5.33 ± 0.12 0.244 ± 0.011 0.105 ± 0.007 410–8 ⁄ 216 ± 14 114 ± 4 3.76 ± 0.08 0.226 ± 0.009 0.139 ± 0.009 Average 225 ± 8 120 ± 4 5.04 ± 0.28 0.257 ± 0.005 0.112 ± 0.005 Std. deviation 26 16 1.07 0.023 0.013 412–2 IAEA-412 sediment from Pacific Ocean (22°23 0 N, 152°40 0 E) 6.7 ± 1.9 0.99 ± 0.21 0.578 ± 0.019 0.183 ± 0.012 0.025 ± 0.007 412–3 7.2 ± 1.4 1.06 ± 0.13 0.597 ± 0.018 0.157 ± 0.010 0.025 ± 0.005 412–4 ⁄ 7.6 ± 0.9 1.12 ± 0.10 0.568 ± 0.017 0.193 ± 0.011 0.030 ± 0.004 412–6 ⁄ 3.4 ± 0.7 0.50 ± 0.06 0.589 ± 0.015 0.195 ± 0.010 0.013 ± 0.003 412–10 ⁄ 3.8 ± 1.1 0.56 ± 0.09 0.579 ± 0.018 0.182 ± 0.011 0.014 ± 0.004 Average 5.7 ± 0.6 0.85 ± 0.11 0.583 ± 0.005 0.182 ± 0.005 0.0218 ± 0.002 Std. deviation 2 0.29 0.011 0.015 0.008 536 E. Chamizo et al./Nuclear Instruments and Methods in Physics Research B 361 (2015) 535–540 contexts. Furthermore, the samples characterization, the radiochemical methods, the measurement technique, and the obtained results will be presented and discussed. 2. Materials and methods 2.1. Samples The two studied marine sediment samples were provided by the IAEA. The one coded IAEA-410 was sampled at 11°26 0 N–164°52 0 Ein the Bikini Atoll in 1997, by box coring at 0–24 depth mixture. The sample IAEA-412 was collected at 22°22 0 N–152°40 0 E in the Pacific Ocean also in 1997, by box coring at 0–35 m depth. The sediments were first dried in open air and subsequently in heating cabinets at 85 °C. They were then ground into powder, sieved through a 250 um, homogenized and stored in polyethylene flasks [20,21]. In addition, the so-coded IAEA-381 reference seawater sample was also studied. This sample was collected by the IAEA in 1993 during cruising between two shallow stations located at about 54°24 0 N–3°33 0 W at 5 m water depth at the Irish Sea, close to the Sellafield liquid discharge area. For this sample, the reported 236 U/ 238 U atom ratio is at the 10 6 level [18,22]. The studied seawater samples were collected during the ARK XVI/3 cruise in the Arctic Ocean (2011). The samples, without a further pretreatment, were directly stored in clean plastic containers and distributed among the laboratories. Two sampling stations were selected for the ETH-CNA intercomparison exercise. Different sampling depths were chosen in each case in order to have a broader range of 236 U/ 238 U atom ratio (Table 2). 2.2. Radiochemical procedure Both seawater samples and sediment samples were processed for the sequential separation of the plutonium and the uranium fractions. For sediment samples, 0.5 g aliquots were selected from the sample IAEA-410 and 5 g from the sample IAEA-412. Different masses where chosen according to their origin (i.e. lower actinide concentration was expected in the second case, solely influenced by the general fallout). In the case of the intercomparison seawater samples, 5 l aliquots were used except for the deepest one at station 218, where 10 l were processed. For that sample, very lowlevels 236 U concentrations were expected. Samples were spiked with approximately 2 pg of 242 Pu and 3 pg of 233 U. The sediment samples were then leached with 8 M HNO 3 and H 2 O 2 on a hot plate to put in solution the fallout material from the surface of the sediment grains, as discussed in [5]. In some cases, the leaching was enhanced by adding concentrated HF (Table 1). Afterwards, actinides were coprecipitated with Fe(OH) 3 after the addition of 50 mg of Fe 3+ from a standard solution provided by High Purity Standards (HPS, England). For the seawater samples, the actinides were directly coprecipitated with Fe(OH) 3 after the addition of 200 mg of Fe 3+ from the same standard solution. In both cases, plutonium and uranium were sequentially purified from the iron hydroxides using TEVA Ò and UTEVA Ò cartridges, respectively, in tandem in a vacuum box, following the guidelines given in [23,24]. Finally, for the AMS cathode preparation, 1 mg of Fe 3+ from the HPS standard solution was added to the corresponding uranium and plutonium fractions to coprecipitate them with Fe(OH) 3 . These precipitates were transferred to quartz crucibles, dried, baked at 600 °C for 1 h to convert the uranium or the plutonium to the oxide form, and finally mixed with about 3 mg of Nb powder and pressed into appropriate aluminum cathodes. In every step, the 238 U contamination was minimized using ultra-pure reagents. Glassware was avoided when possible. It is verified that 236 U and 238 U concentration is negligible in the HPS Fe 3+ standard solution used in the different Fe(OH) 3 preconcentration steps. An instrumental blank was prepared and measured on the Tandy AMS facility at the ETH Zürich, providing negligible figures. The possible 239,240 Pu and 236 U contamination from the 242 Pu and 233 U spikes was also studied. The 242 Pu standard solution supplied by National Physical Laboratory (NPL, England) was proved to contain insignificant amounts of any of the radionuclides of interest (i.e. 239,240 Pu/ 242 Pu and 236 U/ 242 Pu atoms ratios at the 10 6 level). However, it was found that the 233 U main solution, provided by the IAEA Monaco (unknown supplier) contains 236 U at the 10 3 level (i.e. 236 U/ 233 U atom ratio). Therefore, the deepest seawater sample from station 218 was not spiked with 233 U, to avoid a potential 236 U contamination in the sample. Given that very different environmental samples are routinely processed in our laboratory, a minimum of one chemical blank was included in every batch of samples to control the 236 U, 239,240 Pu and 238 U background contributions. 2.3. AMS measurement Plutonium isotopes, 239 Pu and 240 Pu, and uranium isotopes, 236 U and 238 U, were measured on the 1 MV AMS facility at the CNA. Details about the AMS measuring techniques can be found in [14,25]. Shortly, uranium or plutonium isotopes are extracted from the Cs-sputter ion source as oxide anions (i.e. 239,240,242 Pu 16 O  and 233,236,238 U 16 O  ); analyzed by a 90°sector magnet; injected into the tandem accelerator working at about 670 kV where they are stripped to 3+ (i.e. x Pu 3+ and x U 3+ ) in Ar gas; and analyzed by a 90°sector magnet and by a 120°electrostatic deflector. The overall transmission through the accelerator is 11%. The less abundant isotopes (i.e. 239,240,242 Pu and 233,236 U) are counted from the total energy signal provided by a gas ionization chamber. 238 U 3+ is measured as a beam-current in the off-axis Faraday cup (FC3) placed at the exit of the 90°high-energy magnet. The electronics of this FC3 has been recently updated with a new integration range that Table 2 Uranium and plutonium results for the five intercomparison seawater samples obtained on the 1 MV and on the 600 kV compact AMS systems at the CNA and ETH, respectively. The first number indicates the sampling station; the second one stems for the depth in the water column, in meters. The obtained results for the IAEA-381 reference material are also displayed, and compared with the 95% confidence intervals given in [22] (*). For the sample 204_250, the 240 Pu was below the detection limit (LOD), so the 240 Pu/ 239 Pu atom ratio is not reported. NM stands for ‘‘non-measured’’. Sample 236 U/ 238 U(10 9 atom ratio) 236 U(10 6 atoms/kg) 238 U( l g/kg) 240 Pu/ 239 Pu (atom ratio) CNA ETH CNA ETH CNA ETH CNA ETH IAEA-381 2120 ± 60 2031–2786 (*) 17,300 ± 322 19,726–21,423 (*) 3.22 ± 0.06 3.04–3.84 (*) 0.248 ± 0.007 0.18–0.25 (*) 218_5 2.74 ± 0.16 2.81 ± 0.07 21 ± 1.3 21.9 ± 0.9 3.13 ± 0.08 3.03 ± 0.05 0.35 ± 0.25 0.27 ± 0.06 218_750 2.08 ± 0.09 2.22 ± 0.06 18.6 ± 0.9 20.1 ± 0.8 3.60 ± 0.09 3.55 ± 0.05 0.21 ± 0.04 0.21 ± 0.01 218_1800 0.77 ± 0.05 0.81 ± 0.04 NM 7.1 ± 0.6 NM 3.44 ± 0.08 0.19 ± 0.07 0.20 ± 0.03 204_80 3.1 ± 0.18 2.92 ± 0.09 24.8 ± 1.7 23.9 ± 0.9 3.16 ± 0.12 3.21 ± 0.05 0.15 ± 0.08 0.16 ± 0.02 204_250 2.25 ± 0.11 2.40 ± 0.11 19.7 ± 1.1 19.6 ± 1.1 3.51 ± 0.10 3.21 ± 0.08 ( 240 Pu < LOD) 0.19 ± 0.02 E. Chamizo et al./Nuclear Instruments and Methods in Physics Research B 361 (2015) 535–540 537 supports the measurement of pA currents, which are the ones usually involved in natural samples [26]. During a routine measurement, the cycling between the different masses is performed using the Slow Sequential Injection System, a pulsing system designed by HVEE that allows the sequential change, in second pulses, of different parameters of the accelerator. In our case, these parameters are the voltage of the low-energy magnet (bouncer) and the terminal voltage. In the case of the rare isotopes, the 120°electrostatic deflector is also adjusted. Every sample is usually measured 13 runs together with standards, instrumental and chemical blanks. In every run, the masses of interest are measured 5 times. The instrumental precision of the measurement is 2%. The total efficiency of the measurement is a factor of 2 higher for plutonium compared to uranium isotopes, probably due to differences in the ionization yields in the Cs-sputtering process. Typical count rates are about 60 cps per pg of every plutonium isotope in a sample, against 30 cps for uranium. This means a total efficiency of about 3 10 5 for a normal counting time of 25 min per plutonium isotope (i.e. 2.5% counting statistics error for a sample containing 20 fg, or 0.04 mBq, of 239 Pu). The 236 U and 239,240 Pu ETH results for the five intercomparison seawater samples were obtained on the 600 kV AMS facility Tandy. Details about the setup and the measurement technique can be found in [13,27] and references therein. 3. Results and discussion 3.1. IAEA reference materials The uranium results measured for the two new IAEA reference marine sediments, coded 412 and 410, are presented in Table 1. In particular, the 236 U atomic concentrations and the 236 U/ 238 U and 236 U/ 239 Pu atom ratios are shown. The 236 U/ 238 U atom ratios were deduced considering the 238 U concentrations presented in the IAEA data evaluation reports that were obtained following a total decomposition of the samples [20,21]. This is because the 238 U fractions measured by AMS are not representative of the samples, as leaching procedures were applied to extract the fraction of interest (i.e. anthropogenic 236 U) as discussed in [5]. The reported 239,240 Pu results are also displayed. It can be concluded from Table 1 that the four analyzed 0.5 g aliquots from IAEA-410 sample show consistent results. The average 236 U atom concentration is (225 ± 8) 10 6 atoms/kg, and the average 236 U/ 238 U atom ratio is (120 ± 4) 10 9 , where the uncertainties come from error propagation. Each set of results shows a one standard deviation of about 10%, which is a factor of 2 lower than the one associated to the 239+240 Pu activity concentration (with a mean of 5.04 ± 0.28 mBq/g). It is probably that the plutonium present in the sample is more refractory than 236 U. Indeed, the nonhomogenous Pu results could be due to the leaching techniques, as similar Pu concentrations were reported to the IAEA by those participating laboratories using total decomposition procedures (with a 95% confidence interval ranging from 4.52 to 4.92 mBq/g, [20]). This reference material was taken in the Bikini Atoll where atmospheric nuclear detonations were conducted by United States between 1948 and 1958, with the first megaton test in 1952 [28]. Therefore, it is very likely that it contains plutonium and uranium in refractory forms. The different solubility of the plutonium and 236 U fractions will be further studied in a future work. Much lower uranium and plutonium concentrations have been obtained in the case of the sample IAEA-412 (Table 1). This sample might be solely affected by the general fallout, as it was collected in the Pacific Ocean far from any local source. Therefore, a much more homogenous set of results would have been expected. Indeed, this is the case of the plutonium results, as the 239+240 Pu activity concentrations and the 240 Pu/ 239 Pu atom ratios show standard deviations of 2% and 8%, respectively. However, the 236 U results are not that consistent: two out of the five processed aliquots (samples 412–6 and 412–10) provide different results. Specifically, they are about a factor of 2 lower (about 3.5 10 6 atoms/g) than the other three ones (samples IAEA-2, 3 and 4, with about 7 10 6 atoms/g). On average, the 236 U concentration is (5.7 ± 0.6) 10 6 atoms/g and the 236 U/ 238 U atom ratio is (0.85 ± 0.11) 10 9 . These figures have a standard deviation of about 35% including 412–6 and 412–10. It is interesting to mention that these two samples together with 412–4 were leached with HF. Therefore, it is difficult to explain the lower 236 U concentrations obtained. The hypothesis of an error in the added 233 U spike is unlikely, given that these samples provide very dissimilar results from the rest, but comparable within each other. This unexpected result, together with the one for sample 410, might be related to the geochemical nature of 236 U in the marine environment, which is fairly unknown. Thus, further tests are necessary to find an answer for those data. In the case of the IAEA-381 reference material (seawater), the CNA 238 U concentration is within the IAEA 95% confidence interval (Table 2). However, the 236 U concentration approaches that interval from below, 12% apart from the lowest side, and the 236 U/ 238 U atom ratio is close to the lowest value that can be estimated considering the reported 236 U and 238 U information [22]. The most likely explanation for this tendency towards low 236 U results is the loss of detection efficiency in the gas ionization chamber due to the high count rates involved in this case, in the kHz range, since dead-time losses are not considered in our routine protocol for low level environmental data evaluation. However, it is important to keep in mind that only information values have been reported for 236 U and 238 U by the IAEA. 3.2. Intercomparison seawater samples In Table 2 we present the obtained results for the ETH–CNA intercomparison exercise performed with the five different seawater samples from the Arctic Ocean. The CNA ones are corrected by the instrumental backgrounds (i.e. iron oxide mixed with Nb) that represent the cross contamination in the ion source and, besides, in the case of uranium results, by (i) the 236 U contribution of the 233 U spike and, (ii) by the 236 U background produced by the 235 U molecular fragments using the ‘‘ 239 Pu’’/ 238 U ratio as a proxy, as discussed in [14]. For the ETH results, none of these corrections were necessary. In sample 218_1800 only the uranium isotopic composition is displayed in Table 2 in the case of the CNA, as no 233 U spike was added to this sample as explained before. There is a very good agreement between the ETH and CNA results. This is also true for the sample with the lowest 236 U/ 238 U isotopic ratio (218_1800), where the CNA reports (0.77 ± 0.05) 10 9 and the ETH (0.81 ± 0.04) 10 9 . Reaching that isotopic abundance with that accuracy highlights the potential of the CNA facility in this exercise, as the value is a factor of 30 over the abundance sensitivity (i.e. 236 U/ 238 U310 11 [14]). This means that the applied background corrections are reliable for atomic ratios in this range. This limiting background is at least two orders of magnitude lower in the case of the ETH facility, thanks to the introduction of a second sector magnet on the high-energy side of the spectrometer [13]. The 236 U and 238 U atomic concentrations in both facilities are also in agreement within errors, which indicate the reliability of the calibration of the 233 U CNA spike solution and of the radiochemical procedure used to process the samples. The biggest discrepancy is found in the precisions of the reported 240 Pu/ 239 Pu atom ratios, which in the case of the CNA are at least a factor of 3 above the values obtained at ETH. It is very unlikely 538 E. Chamizo et al./Nuclear Instruments and Methods in Physics Research B 361 (2015) 535–540 that the problem is the chemical yield, as previous tests performed by alpha-spectrometry indicated similar recoveries for plutonium and uranium, ranging from 70% to 90%. The most likely explanation is the insufficient measuring time dedicated to the plutonium analysis, about 40 min per sample (i.e. 20 min for 240 Pu and 15 min for 239 Pu). Another reason might be the poor performance of Cs-sput- ter ion source, which was running at a much lower Cs output during plutonium than during uranium analysis to prevent instabilities. In any case, for similar measurement times and source outputs, a better statistical precision is always expected for the 600 kV AMS system at the ETH. It features a transmission through the accelerator for the 3+ charge state at least a factor of 3 higher than the CNA one, thanks to the use of He as stripper gas [29]. Finally, Table 3 displays the obtained 239+240 Pu activity concentrations for the seawater samples measured at the CNA. For the aliquots processed at the ETH no 242 Pu spike was added, so the Pu activity concentrations are not reported. Despite the low efficiency obtained during the plutonium analysis explained above, the uncertainties obtained for the plutonium activity concentrations are low enough to outline a depth profile. A clear 239+240 Pu activity concentration maximum is obtained for the 750 m depth sample due to the well-known remineralization process [30]. 3.3. 236 U and Pu isotopes as environmental tracers Discussion of the results in an environmental context shows a contribution of anthropogenic 236 U to all the studied samples. The measured 236 U/ 238 U atomic ratios are above 10 10 , at least 2 orders of magnitude higher than the expected one for natural sources (from 10 13 to 10 14 [2]). The studied marine sediment samples show ratios consistent with the involved actinides sources in each case. The sample IAEA-410, from Bikini Atoll, shows 236 U/ 238 U atom ratios of about 10 7 , two orders of magnitude higher than the IAEA-412, due to the influence of the local fallout from the thermonuclear tests conducted at that site. In the case of the intercomparison seawater samples, the most superficial ones show ratios between 2 and 3 times 10 9 , approximately in the same range than the ratios associated to general fallout. On the other hand, the 236 U/ 238 U ratio measured in the IAEA-381 sample is three orders of magnitude higher than the fallout ratios, as the 236 U signal from the Sellafield NRP is obviously clear at the Irish Sea [22]. In Fig. 1, 236 U/ 239 Pu average ratios are plotted against the corresponding 240 Pu/ 239 Pu ratios for the two studied marine sediments from Pacific Ocean, IAEA-412 and IAEA-410, and other results for the same kind of sample from different areas (Irish Sea [18,19]; North Atlantic Ocean [31]; and Mururoa and Fangataufa atolls [14]). Also, the reported isotopic composition for soils affected by the general fallout from different zones (Australia [32]; Japan [5]; Austria [14]) is displayed. It is observed that both ratios show a clear source dependency. In the case of marine sediments, the lowest 236 U/ 239 Pu atom ratio, of about 3.5 10 5 , has been reported for the IAEA-384 sample [14]. As mentioned before, this sample is influenced by the local fallout from the atmospheric tests performed in the French Polynesia (Mururoa and Fangataufa atolls) from 1966 to 1971. This influence can be also noticed in the 240 Pu/ 239 Pu atom ratio, of about 0.055, which is within the range of the so-called weapon-grade plutonium (from 0.02 to 0.07). Both ratios reflect the nature of these tests (i.e. low-yield nuclear tests and safetytests), whose debris usually keep the isotopic composition of the primary fuel [33,34]. Some 236 U/ 239 Pu ratios are associated to the general fallout in the Northern Hemisphere and detected using the 240 Pu/ 239 Pu range (from 0.16 to 0.19, [35]). Two different results have been considered: (i) IAEA-412 sample from the Pacific Ocean reported in this work ( 236 U/ 239 Pu = 0.0218 ± 0.02) and, (ii) the one obtained for a sediment core collected in the North Atlantic Ocean at the socalled Porcupine Abyssal Plain,with an integrated 236 U/ 239 Pu ratio of 0.032 [31]. These ratios are about one order of magnitude lower than the reported ones for soils (Fig. 1). This effect is caused by the solubility of uranium in seawater. In fact, in contrast to its presence in soils, only about 1% of the anthropogenic 236 U released to the Oceans is expected to be found in abyssal marine sediments [9], against a 20–35% range for the particle-reactive 239 Pu [36]. The IAEA-135 sample must be influenced by the Sellafield NRP since it was collected in the Irish Sea. As discussed before, its 239 Pu/ 240 Pu ratio is not dramatically different from global fallout group. However, it differs from the group in the 236 U/ 239 Pu isotopic composition, with a ratio of 0.09 [18,19]. Finally, the IAEA-410 sample from the Bikini Atoll shows remarkably higher isotopic ratios ( 236 U/ 239 Pu = 0.112 ± 0.005 and 240 Pu/ 239 Pu = 0.257 ± 0.005). These ratios might define the very characteristic uranium and plutonium composition of the local fallout released in the thermonuclear tests (yields up to 15 Mt) that were conducted in the Marshall Islands between 1952 and 1958 by United States [28]. As for the seawater samples, it can be observed in Table 3 that the 236 U/ 239 Pu ratios, ranging from 1 to 6, are in agreement with Table 3 239+240 Pu activity concentrations and 236 U/ 239 Pu atom ratios obtained on the 1 MV AMS system at the CNA for the six studied seawater samples (Table 2). For the sample 218_1800, the 236 U results obtained at the ETH have been used for the estimation of the 236 U/ 239 Pu atom ratio, as no 233 U spike was used for the samples processed at the CNA (see Section 3in the text) ( ⁄ ). The 239+240 Pu activity concentration for sample 204_250 refers solely to 239 Pu, as 240 Pu was below the detection limit ( ⁄⁄ ). Sample 239+240 Pu ( l Bq/kg) 236 U/ 239 Pu (atom ratio) 218 _5 8 ± 3 5.8 ± 2.4 218_750 16.4 ± 1.6 1.9 ± 0.2 218_1800 9.40 ± 1.6 1.2 ± 0.2 ( ⁄ ) 204_80 5.70 ± 1.3 6.2 ± 1.3 204_250 1.3 ± 0.8 ( ⁄⁄ )14±8 IAEA-381 13.82 ± 0.19 mBq/kg (13.1–14.5 mBq/ kg) 2.06 ± 0.04 Fig. 1. Correlation between the 240 Pu/ 239 Pu and 236 U/ 239 Pu atom ratios for the (i) IAEA-410 and IAEA-412 samples, (ii) a sediment core from the Porcupine Abyssal Plain site (PAP) from the North Atlantic Ocean [31] and, (iii) different soils and marine sediments published elsewhere (empty symbols). The results for the IAEA- 384 sample were reported in [14], the ones for IAEA-135 in [18,19], the one for soils from Australia in [32], from Japan in [5], and from Austria (IAEA-Soil-6) in [14]. For the sediment samples, a line has been drawn to guide the eye. The 240 Pu/ 239 Pu atom ratios for the weapon-grade plutonium and for the general fallout in the Northern Hemisphere are also indicated. E. Chamizo et al./Nuclear Instruments and Methods in Physics Research B 361 (2015) 535–540 539 the measured ones in [17] in samples mostly affected by the general fallout. It is noteworthy that the IAEA-381 sample, with a clear influence of Sellafield discharges, has a 236 U/ 239 Pu ratio in the same range. The Arctic samples are also influenced by the Sellafield and La Hague NRP [6]. However, the IAEA-135 marine sediment from the Irish Sea (Fig. 1) shows a noticeable different 236 U/ 239 Pu ratio from the global fallout one, as discussed before. Due to the very different biogeochemical characteristics of both radionuclides in the water, particulate and sediment fractions, this ratio is not expected to be constant in ocean waters in time and space. Besides, 236 U/ 239 Pu might not be an absolute ratio that can be used to discriminate the origin of the samples. In any case, the data available are scarce, so further sampling and measurements must be undertaken in the future to predict the behavior of 236 U in sea water. 4. Summary and conclusions Our results demonstrate the potential of the 1 MV AMS facility at the CNA for 236 U and 239 Pu determinations in oceanography studies. The reliability of the CNA 236 U measuring technique (sample preparation and AMS determination) has been demonstrated thanks to an intercomparison exercise performed with the ETH Zürich on a set of 5 seawater samples, with 236 U/ 238 U atom ratios ranging from 7 10 10 to 3 10 9 . The obtained 236 U and 239 Pu results on different aliquots of two IAEA reference samples (IAEA-410 and 412) provide new valuable information about the source dependency of the 236 U/ 239 Pu atom ratio in marine sediments. This information, together with other published results, has been compared with the one provided by the extensively studied 240 Pu/ 239 Pu atom ratio. Despite the different physical–chemical behavior of both elements, complementary and valuable information can be obtained through the study of both ratios. A much more extensive research is necessary to understand the behavior of anthropogenic 236 U in marine samples. Acknowledgments This work has been financed through the projects FIS2012- 31853 provided by the Spanish Ministry of Economy and Competitiveness, and P-10-FQM-5956, provided by the Andalusian Government (Junta de Andalucia, Spain). References [1] K.M. Wilcken, L.K. Fifield, T.T. Barrows, S.G. Tims, L.G. Gladkis, Nucl. Instr. Meth. B 266 (2008) 3614. [2] P. Steier, M. Bichler, L.K. Fifield, R. Golser, W. Kutschera, A. Priller, F. Quinto, S. Richter, M. Srncik, P. Terrasi, et al., Nucl. Instr. Meth. B 266 (2008) 2246. [3] P.P. Parekh, T.M. Semkow, M.A. Torres, D.K. Haines, J.M. Cooper, P.M. Rosenberg, M.E. Kitto, J. Environ. Radioact. 85 (2006) 103. [4] T. Bisinger, S. Hippler, R. Michel, L. Wacker, H.-A. Synal, Nucl. Instr. Meth. B 268 (2010) 1269. [5] A. Sakaguchi, K. Kawai, P. Steier, F. Quinto, K. Mino, J. Tomita, M. Hoshi, N. Whitehead, M. Yamamoto, Sci. Total Environ. 407 (2009) 4238. [6] N. Casacuberta, M. Christl, J. Lachner, M. Rutgers van der Loeff, P. Masqué, H.-A. Synal, Geochim. Cosmochim. Acta 133 (2014) 34. [7] P.J. Kershaw, D. McCubbin, K.S. Leonard, Sci. Total Environ. 237–238 (1999) 119. [8] A. Sakaguchi, A. Kadokura, P. Steier, K. Tanaka, Y. Takashi, H. Chiga, A. Matsushima, S. Nakashima, Y. Onda, Geochem. J. 46 (2012) 355. [9] A. Sakaguchi, A. Kadokura, P. Steier, Y. Takahashi, K. Shizuma, M. Hoshi, T. Nakakuki, M. Yamamoto, Earth Planet. Sci. Lett. 333–334 (2012) 165. [10] M. Christl, N. Casacuberta, J. Lachner, S. Maxeiner, C. Vockenhuber, H.-A. Synal, I. Goroncy, J. Herrmann, A.B. Daraoui, C. Walther, R. Michel, Nucl. Instr. Meth. B (2015). [11] O.J. Marsden, F.R. Livens, J.P. Day, L.K. Fifield, P.S. Goodall, Analyst 126 (2001) 633. [12] C. Vockenhuber, I. Ahmad, R. Golser, W. Kutschera, V. Liechtenstein, A. Priller, P. Steier, S. Winkler, Int. J. Mass Spectrom. 223–224 (2003) 713. [13] C. Vockenhuber, M. Christl, C. Hofmann, J. Lachner, A.M. Müller, H.-A. Synal, Nucl. Instr. Meth. B 269 (2011) 3199. [14] E. Chamizo, M. Christl, L.K. Fifield, Nucl. Instr. Meth. B (2015) (submitted for publication). [15] M.E. Ketterer, A.D. Groves, B.J. Strick, C.S. Asplund, V.J. Jones, J. Environ. Radioact. 118 (2013) 143. [16] M. Srncik, P. Steier, G. Wallner, J. Environ. Radioact. 102 (2011) 614. [17] R. Eigl, M. Srncik, P. Steier, G. Wallner, J. Environ. Radioact. 116 (2013) 54. [18] S.H. Lee, P.P. Povinec, E. Wyse, M.A.C. Hotchkis, App. Radiat. Isot. 66 (2008) 823. [19] Y. Muramatsu, S. Uchida, K. Tagami, S. Yoshida, T. Fujikawa, J. Anal. At. Spectrom. 14 (1999) 859. [20] Worldwide laboratory comparison on the determination of radionuclides in IAEA-410 Bikini Atoll sediment, IAEA/AQ/xxx, 2014, p. 78 (in preparation). [21] Worldwide laboratory comparison on the determination of radionuclides in IAEA-412 Bikini Atoll sediment, IAEA/AQ/xxx, 2014, p. 78 (in preparation). [22] P.P. Povinec, C. Baide, A. Baeza, G. Barci-Funel, T.D. Bergan, B. Bojanowski, W. Burnett, J. Eikenberg, L.K. Fifield, V. Serradell, et al., J. Rad. Nucl. Chem. 251 (2002) 369. [23] E.P. Horwitz, M.L. Dietz, R. Chiarizia, H. Diamond, S.L. Maxwell III, M. Nelson, Anal. Chim. Acta 310 (1995) 63. [24] E.P. Horwitz, M.L. Dietz, R. Chiarizia, H. Diamond, A.M. Essling, D. Graczyk 266 (1992) 25. [25] E. Chamizo, S.M. Enamorado, M. García-León, M. Suter, L. Wacker, Nucl. Instr. Meth. B 266 (2008) 4948. [26] E. Chamizo, F.J. Santos, J.M. López-Gutiérrez, S. Padilla, M. García León, J. Heinemeier, C. Schnabel, Nucl. Instr. Meth. B 361 (2015) 13. [27] M. Christl, C. Vockenhuber, P.W. Kubik, L. Wacker, J. Lachner, V. Alfimov, H.-A. Synal, Nucl. Instr. Meth. B 294 (2013) 29. [28] M.W. Carter, A.A. Moghissi, Health Physics 33 (1977) 55. [29] C. Vockenhuber, V. Alfimov, M. Christl, J. Lachner, T. Schulze-König, N. Suter, H.-A. Synal, Nucl. Instr. Meth. B 294 (2013) 382. [30] K. Hirose, M. Aoyama, C.S. Kim, C.K. Kim, P.P. Povinec, Radioact. Environ. 8 (2006) 67. [31] E. Chamizo, M. Villa, S. Hurtado, J.M. López-Gutiérrez, F.J. Santos, I. Gómez- Martínez, Sci. Total Environ. (to be submitted). [32] M. Srncik, S.G. Tims, M. De Cesare, L.K. Fifield, J. Environ. Radioact. 132 (2014) 108. [33] H.G. Hicks, D.W. Barr, Lawrence Livermore National Laboratory, UCRL-53499/1 (1984). [34] E. Chamizo, M. García-León, J.I. Peruchena, F. Cereceda, V. Vidal, E. Pinilla, C. Miró, Nucl. Instr. Meth. B 269 (2011) 3163. [35] J.M. Kelley, L.A. Bond, T.M. Beasley, Sci. Total Environ. 237 (238) (1999) 483. [36] G. Buffoni, R. Delfanti, C. Papucci, Mar. Geol. 109 (1992) 159. 540 E. Chamizo et al./Nuclear Instruments and Methods in Physics Research B 361 (2015) 535–540