Article pubs.acs.org/est
Rapid Multisample Analysis for Simultaneous Determination of Anthropogenic Radionuclides in Marine Environment Jixin Qiao,*,† Keliang Shi,‡ Xiaolin Hou,†,§ Sven Nielsen,† and Per Roos† †
Center for Nuclear Technologies, Technical University of Denmark, DTU Risø Campus, DK-4000 Roskilde, Denmark Radiochemistry Lab, School of Nuclear Science and Technology, Lanzhou University, 730000, Lanzhou, China § Xi’an AMS Center and SKLLQG, Institute of Earth Environment, Chinese Academy of Science, Xi’an 710075, China ‡
S Supporting Information *
ABSTRACT: An automated multisample processing flow injection (FI) system was developed for simultaneous determination of technetium, neptunium, plutonium, and uranium in large volume (200 L) seawater. Ferrous hydroxide coprecipitation was used for the preliminary sample treatment providing the merit of simultaneous preconcentration of all target radionuclides. Technetium was separated from the actinides via valence control of technetium (as Tc(VII)) in a ferric hydroxide coprecipitation. A novel preseparation protocol between uranium and neptunium/plutonium fractions was developed based on the observation of nearly quantitative dissolution of uranium in 6 mol/L sodium hydroxide solution. Automated extraction (TEVA for technetium and UTEVA for uranium) and anion exchange (AGMP-1 M for plutonium and neptunium) chromatographic separations were performed for further purification of each analyte within the FI system where four samples were processed in parallel. Analytical results indicate that the proposed method is robust and straightforward, providing chemical yields of 50−70% and improved sample throughput (3−4 d/sample). Detection limits were 8 mBq/m3 (0.013 pg/L), 0.26 μBq/m3 (0.010 fg/L), 23 μBq/m3 (0.010 fg/L), 84 μBq/m3 (0.010 fg/L) and 0.6 mBq/m3 (0.048 ng/L) for 99Tc, 237Np, 239Pu, 240Pu and 238U for 200 L seawater, respectively. The unique feature of multiradionuclide and multisample simultaneous processing vitalizes the developed method as a powerful tool in obtaining reliable data with reduced analytical cost in both radioecology studies and nuclear emergency preparedness.
■
plasma mass spectrometry (ICP-MS)) quantifications.1 A number of bulky samples have to be transported to the laboratory for single radionuclide determination, which is not only costly but also physically cumbersome to handle. Even though with the improvement in detection limits via using modern accelerator mass spectrometry (AMS), sample volumes required for 236U, 239Pu, 240Pu, and 237Np determination can be reduced to 2−50 L, the AMS measurement for these radionuclides is still limited to a few laboratories especially for 99Tc quantification so far. As 99Tc has been the main interest in European (and the Arctic Ocean) waters as well as the different physical-chemical behavior of Tc relative to actinides (e.g., Np, Pu, and U), analytical procedure developed so far for Tc determination were in most cases performed separately from the ones for Np, Pu, and U.6−8 Although Ballestra et al.9 have proposed the applicability of iron hydroxide coprecipitation for simultaneous isolation of Tc and actinides (Np, Pu, Am) from environmental water samples, their work was mainly focused on Tc determination whereas protocols for actinides purification
INTRODUCTION Technium-99, as a fission product and pure beta emitter, is the dominating isotope of technetium. 237Np, 239Pu, and 240Pu are the most important isotopes of neptunium and plutonium and decay with alpha emissions. Large amount of 99Tc, 237Np, 239Pu, 240 Pu were released and distributed in the environment via anthropogenic nuclear-related activities1,2 Due to the long halflives of these radionuclides (99Tc: 2.12 × 105 y; 237Np: 2.14 × 105 y; 239Pu: 2.41 × 104 y; 240Pu: 6.54 × 103 y), once entering into the environment, they will have long-term persistence. On the other hand, 99Tc and U isotopes (e.g., 236U (t1/2= 2.34 × 107 y)) have relatively high mobility and bioavailability in oxic waters, making them rather valuable tracers in oceanographic studies.3−5 For the purpose of environmental risk monitoring and assessment, nuclear decommissioning, nuclear waste management, emergency preparedness and their tracer applications, the accurate determination of 99Tc, 237Np, Pu and U isotopes is imperative. A number of analytical methods have been developed during the past few decades for determination of 99Tc, 237Np, Pu and U isotopes in environmental samples. In seawater analysis, due to the extremely low concentrations of these radionuclides in the marine environment, large volume (e.g., 200 L) samples are usually needed to meet the required detection limits in radiometric or mass spectrometric (e.g., inductively coupled © 2014 American Chemical Society
Received: Revised: Accepted: Published: 3935
October 13, 2013 February 5, 2014 February 28, 2014 February 28, 2014 dx.doi.org/10.1021/es404584b | Environ. Sci. Technol. 2014, 48, 3935−3942
Environmental Science & Technology
Article
were not given. Furthermore, most of the published analytical procedures for Tc, U, Pu, and Np seawater analysis are operated in a manual fashion thus making them relatively timeconsuming and labor insensitive.9−12 Although the application of flow-based automated methods constitutes a great advantage in radiochemical assays, at present their development is still limited to a few laboratories, and gaps between method development and the requirements in routine analysis of large volume environmental samples still exist.13−19 Moreover, to the best of our knowledge, flow-based manifolds are not commercially available for multiple-sample simultaneous processing.13,14,18 This might be a consequence of the mechanical complexity and the proneness to instrumental failure when assembling various liquid drivers and rotary valves in the flow system. In this work, we aimed to develop an analytical protocol in which Tc, Np, Pu, and U can be simultaneously determined from large volume seawater. A flow injection system was exploited for automated sample processing wherein four samples can be processed in parallel. Ion exchange and extraction chromatography was used for chemical purification and both radiometric (beta counting) and mass spectrometric (ICP-MS) methods were used for quantification of target radionuclides.
Figure 1. Schematic illustration of the multisample processing FI system for simultaneous determination of Tc, Np Pu, and U in large volume seawater.
SV1−8 were connected with PTFE tubing of smaller diameter (0.8 mm i.d./1.6 mm o.d). Four Econo-Columns (Bio-Rad Laboratories Inc., Hercules, CA) packed with TEVA, UTEVA, or AGMP-1 M resin were integrated in the flow system through PEEK ferrules and fittings, whereupon the chemical purification of analytes were controlled automatically via the aid of the userfriendly FIAlab software (FIAlab Instruments, Bellevue, WA). Preliminary Investigation. To preliminarily investigate the effect of sample pH and concentration of Fe on the coprecipitation efficiency of iron hydroxide, as well as to select suitable reducing and oxidizing reagents for Tc, Pu, and Np valence adjustment, 200 mL of seawater spiked with certain amounts of 242Pu (ca. 10 mBq), 237Np (ca. 1 mBq), and 99mTc (1−100 Bq) was used. In some cases, 55Fe (1−100 Bq) was also spiked to monitor the precipitation efficiency of iron hydroxide. To each sample, 0.2−2 mL of concentrated HCl was first added to acidify the seawater (pH from 0.9 to 1.9). Then, 0.03−0.2 mL of 100 g/L Fe (in the form of FeCl3) and certain amount of reducing reagent (K2S2O5, NH2OH·HCl, N2H4·HCl or ascorbic acid) were added, respectively. After stirring the sample for 20 min via magnetic stirrer or N2 bubbling, 10 mol/ L NaOH was added to adjust the pH to 11. After discarding supernatant, the residue was dissolved with 5−10 mL of 2 mol/ L HCl and counted using a NaI gamma detector for 99mTc. A 1 mL aliquot was taken from the sample and diluted for 238U, 237 Np, and 242Pu measurement by ICP-MS. To the remaining solution, after adding reducing reagent (see details in SI Table S-2), 10% NH3·H2O was added to adjust pH to 9. The residue obtained after centrifugation was dissolved with 5−10 mL of 2 mol/L HCl and 99mTc, 238U, 237Np, and 242Pu were measured again to calculate their chemical yields. To select optimal oxidizing reagents for separating Tc from U, Np, and Pu, different oxidizing reagents (H2O2, NaClO, NaNO2, K2S2O8 or air) were applied. After oxidization of Tc (Pu and Np would also be oxidized), 10% NH3·H2O was added to pH 9. Both supernatant and residue were collected for 99mTc, 238 U, 237Np and 242Pu measurement as mentioned above. Uranium Dissolution Test in NaOH Solutions. To investigate the chemical behavior of U−Fe(OH)3 coprecipitate when treated with different NaOH solutions, a 200 mL acidified
■
EXPERIMENTAL SECTION Sample and Reagents. 99mTc was obtained from 2 to 4 GBq commercial 99 Mo-99mTc generators (Amersham, UK) and purified using alumina cartridges according to the method reported by Hou et al.20 Standard solution of 242Pu (0.1037 Bq/g in 2 mol/L HNO3) diluted from NBL-CRM 130 was purchased from New Brunswick Laboratory (Argonne, IL) and used as a chemical yield tracer for both Np and Pu. A 237Np solution of 0.01175 Bq/g in 2 mol/L HNO3 was diluted from NIST-SRM-4341 (National Institute of Standard and Technology (NIST), Gaitherburg, MD). 99Tc (in the form of NH4TcO4 with an activity of 4.17 Bq/g) and 239Pu (0.100 Bq/g in 2 mol/L HNO3) standard solutions were supplied by Center for Nuclear Technologies, Technical University of Denmark (DTU Nutech). A standard solution of 238U (1.000 g/L in 2 mol/L HNO3) was purchased from NIST (Gaithersburg, MD). TEVA and UTEVA extraction chromatographic resin (particle size was 100−150 μm for both) was purchased from TRISKEM International (Bruz, France). Macro porous anion exchange resin AGMP-1 M (particle size 100−200 mesh (corresponding to 75−150 μm)) was purchased from BioRad Laboriatories Inc. (Hercules, CA). All other chemicals used in this work were analytical grade and all solutions were prepared with high purity water (18 MΩ·cm). Seawater samples collected from Roskilde Fjord, Denmark (55°41′N, 12°5′E) in 2012 were used throughout the work for method development. For method application and evaluation, seawater samples collected from Greenland (see sample list in Supporting Information (SI) Table S-1) coast were used. Instrumentation. A multisample processing flow injection (FI) system was developed in this work (see Figure 1), wherein four seawater samples after preconcentration can be handled simultaneously. The system consists of a four-channel peristaltic pump (Watson-Marlow Inc. Wilmington, MA), a 10-port multiposition selection valve (MSV, Valco Instruments, Houston, TX) and 8 solenoid valves (SV1−8). All outlets of MSV were connected through PEEK ferrules and fittings with rigid PTFE tubing of 2.4 mm i.d./3.2 mm o.d., while all outlets 3936
dx.doi.org/10.1021/es404584b | Environ. Sci. Technol. 2014, 48, 3935−3942
Environmental Science & Technology
Article
Figure 2. Schematic flowchart of analytical procedure for simultaneous determination of Tc, Np, Pu, and U in large volume seawater (values in round brackets are chemical yields of 238U, uncertainties for all values are ≤15%).
Tc, U, Np, and Pu in 200 L seawater is summarized in the SI. The schematic flowchart of the overall procedure is also illustrated in Figure 2. In general, Tc (as Tc(IV)), U, Np, and Pu, were coprecipitated with Fe(OH)2 and preconcentrated from the large volume of seawater. After oxidizing Tc(IV) to Tc(VII) with NaClO, Tc was separated from U, Np, and Pu fraction for the further purification with TEVA column. Afterward, U was dissolved in 6 mol/L NaOH from the Fe(OH)3 coprecipitate and thus separated from Np and Pu fraction for further purification with UTEVA column. The remaining Np and Pu fraction was further purified with AG MP-1 M column.
seawater (pH 2) or 100 mL artificial solution of 0.2 mol/L HCl spiked with 1 μg of U was used as a sample. To each sample, 1 mL of 0.1 g/mL Fe was added and then the U−Fe(OH)3 coprecipitate was obtained after adjusting the sample pH to 9 with 10% NH3·H2O and centrifugation. Thirty mL of freshly prepared 0.0001−6 mol/L NaOH, old 6 mol/L NaOH stored for several months or mixture of 6 mol/L NaOH and 0.2 mol/ L Na2CO3 solution was added to the coprecipitate. After stirring for 10 min, the sample was centrifuged again, the supernatant and residue (dissolved with 5 mL of concentrated HCl) were saved for 238U measurement by ICP-MS. Analytical Procedure for 200 L Seawater Analysis. The detailed analytical procedure for simultaneous determination of 3937
dx.doi.org/10.1021/es404584b | Environ. Sci. Technol. 2014, 48, 3935−3942
Environmental Science & Technology
Article
Detection of Radionuclides. The detection of 55Fe in the preliminary investigation was carried out with ultra low level liquid scintillation spectrometer (1220 Quantulus, PerkinElmer life science, Shelton, CT). A NaI well gamma detector was used for detection of 99mTc and an anticoincidence shielded low-level GM gas-flow beta counter (DTU Nutech, Denmark) was exploited for 99Tc quantification. The detection of 238U, 237Np, 239Pu, 240Pu, and 242Pu was performed with ICP-MS instrument (X SeriesII, Thermo Fisher Scientific, Waltham, MA) equipped with an Xt-skimmer cone and an ultrasonic nebulizer (U5000AT+, CETAC, USA) under hot plasma conditions. All Pu/Np eluates were evaporated to dryness and dissolved with 5 mL of 0.5 mol/L HNO3. 238U concentrations in both raw seawater and U eluate were determined after appropriate dilution with 0.5 mol/L HNO3. 115 In (as InCl3) was exploited as an internal standard and prepared to a concentration of 1 μg/L in each sample for calculating the chemical yields of Np/Pu and U. A 0.5 mol/L HNO3 solution was used as washing solution among consecutive assays. Prior to each measurement, the ICP-MS instrument was tuned to maximum transmission of target analytes. Typical sensitivities of U, Np and Pu ranged from 1 × 106 to 5 × 106 cps per μg/L.
■
Figure 4. Effect of ion concentration on the coprecipitation efficiency of Tc, Np, Pu, and U.
RESULTS AND DISCUSSION Optimization for Iron Hydroxide Coprecipitation. In oxic seawater, substantial amounts of U(VI) exist as carbonate
Figure 5. Analytical performance of different oxidizing reagents for the separation of Tc from Np, Pu, and U.
Figure 3. Effect of sample pH on chemical yields of Tc and U after the second coprecipitation.
Ca(OH)2, and Fe(OH)2 (Figure 4(a)). While in the second coprecipitation, where 10% NH 3 ·H 2 O was used as a precipitating reagent to adjust the pH to 9, the chemical yields of Tc, Np, Pu, and U demonstrate positive correlation with the concentration of Fe (Figure 4(b)). This also confirms that in the second coprecipitation, the precipitate is mainly composed of iron(II) hydroxides. To obtain sufficiently high coprecipitation efficiency for all the analytes, the optimal addition of Fe was set to 50 mg/L in the present work. It is noted that 114.8 ± 8.6% chemical yields were obtained for Pu, which might be a consequence of matrix effect in the ICP-MS measurement. In the second coprecipitation, to prompt the settlement of the precipitate, solid KCl (or NaCl) was added. The experiment indicated that by adding solid KCl (or NaCl) after the formation of Fe(OH)2 precipitate and vigorous stirring, complete settlement of precipitate could be achieved within 1 h (≥3 h were normally needed without the addition of KCl or NaCl). This might be attributed to the increase of ionic strength and kinetic energy by extra addition of salts and vigious stirring, which provide more precipitation
complexes, therefore, release of U prior to the iron hydroxide coprecipitation is imperative to ensure the quantitative scavenging of U. At lower pH, carbonates can be converted to carbon dioxide, and thus U can be released as free ions into the seawater. Analytical results about the effect of pH on the coprecipitation of U (Figure 3) show that, chemical yields of U under investigated conditions are all around 80−85% after the second coprecipitation using 10% NH3·H2O, revealing pH of about 2 is sufficient to release the majority of U into free ions. Approximately 10−15% of U was lost during the coprecipitation which might be caused by insufficient addition of Fe (15 mg/L), since the chemical yields of 55Fe (Figure 3) were also not very high (70−85%). To pretreat 200 L seawater, a two-step coprecipitation procedure was exploited after valence adjustment of Tc(IV), Np(IV), Pu(III), and Fe(II), which is based on the control of sample pH with the use of different precipitating reagents. In the first coprecipitation, NaOH (10 mol/L) was used to adjust the pH to 11, under this condition Tc(IV), Np(IV), Pu(III), and U(IV) can be quantitatively coprecipitated with Mg(OH)2, 3938
dx.doi.org/10.1021/es404584b | Environ. Sci. Technol. 2014, 48, 3935−3942
Environmental Science & Technology
3939
0.084 ± 0.040 (0.010 ± 0.005) 0.023 ± 0.012 (0.010 ± 0.005) 0.0003 ± 0.0001 (0.010 ± 0.004) 0.008 ± 0.002 (0.013 ± 0.003)
Uncertainties for all chemical yields are less than 10%. bValues in round brackets are the corresponding concentrations in mass per volume (pg/L for 99Tc and fg/L for 237Np, 239Pu and 240Pu). cLimit of detection (LOD) for each individual radionuclide was calculated as the 3 times of the standard deviation of the procedure blank. a
0.151 ± 0.023 (11.500 ± 2.269) 1.396 ± 0.209 (0.166 ± 0.025) 0.308 ± 0.060 (11.846 ± 2.308) 55.5 57.5
0.157 ± 0.016 (0.248 ± 0.025)
2.520 ± 0.378 (1.096 ± 0.164)
0.172 ± 0.026 (11.500 ± 2.269) 0.241 ± 0.036 (11.500 ± 2.269) 0.221 ± 0.033 (11.500 ± 2.269) 1.280 ± 0.192 (0.144 ± 0.023) 1.306 ± 0.196 (0.155 ± 0.023) 1.047 ± 0.157 (0.125 ± 0.019) 2.026 ± 0.303 (0.881 ± 0.132) 1.476 ± 0.221 (0.642 ± 0.096) 1.290 ± 0.194 (0.561 ± 0.084) 0.299 ± 0.059 (11.500 ± 2.269) 0.385 ± 0.077 (14.808 ± 2.962) 0.448 ± 0.089 (17.231 ± 3.423) 0.094 ± 0.009 (0.148 ± 0.014) 0.087 ± 0.009 (0.137 ± 0.014) 0.188 ± 0.020 (0.297 ± 0.032) 58.9 50.3 56.3 61.6 61.0 62.7
Greenland-1C Greenland-1D GreenlandDANA1 GreenlandDANA2 LODc
239
Tc, Bq/m3 (pg/L) 99
Pu 242
Tc 99
sample ID
chemical yielda,%
Table 1. Selected Results for Method Application and Evaluation
237
seeds and molecular collision probability, thus better aggregation or flocculation. Selection of Reagents for Valence Adjustment. Selection of Reducing Reagent. The principle in Tc preconcentration using iron hydroxide coprecipitation is based on reducing the valence of Tc to Tc(IV) which will be coprecipitated with Fe(OH)2 under alkaline condition. In the first coprecipitation, the reduction of Pu and Np is also critical to obtain an isotopic equilibrium between intrinsic Np/Pu in the sample and the spiked 242Pu tracer. Therefore, it is important to select a suitable reducing reagent to ensure high chemical yields of Tc and analytical reliability. In this work, several reducing reagents have been tested and the results are summarized in SI Table S-2. Initially, salt-free reducing reagents including NH2OH·HCl, N2H4·HCl and ascorbic acid, were preferably tested to avoid the induction of extra matrix interferences for the following chemical separation. However, none of them provided sufficient reducing capacity to reduce Tc to Tc(IV) under the experimental conditions (SI Table S-2). Only 45.4 ± 4.5% of Tc (Group 1, SI Table S-2) can be recovered when using NH2OH·HCl (1.5 g/L), which might be due to an insufficient addition of NH2OH·HCl or slow reducing dynamic. N2H4·HCl demonstrated even lower reducing efficiency for Tc (chemical yields
