Article pubs.acs.org/ac
Five-Column Chromatography Separation for Simultaneous Determination of Hard-to-Detect Radionuclides in Water and Swipe Samples Xiongxin Dai* and Sheila Kramer-Tremblay Chalk River Laboratories, Atomic Energy of Canada Limited, Chalk River, Ontario K0J 1J0, Canada S Supporting Information *
ABSTRACT: There is a growing demand for the rapid determination of hard-to-detect radionuclides in environmental and biological samples for environmental monitoring, radiological protection, and nuclear forensic reasons. A new method using five-column chromatography separation has been developed for the simultaneous determination of Pu, Np, Th, U, Am, Cm, Pm, Y, and Sr isotopes, as well as iron-55, by inductively coupled mass spectrometry (ICPMS), α spectrometry, Č erenkov and liquid scintillation (LS) counting. Spiked swipe and water samples as well as proficient testing water standards were analyzed to validate the separation procedure, and the results are in good agreement with the expected values. The method provides quick sample turnaround time and high analysis throughput with low analysis cost. The flexibility of the method also allows for its easy adaptation to various emergency and routine radioassays.
I
techniques,18,19 would facilitate high sample analysis efficiency and cost-effectiveness. Recently, we have also developed a rapid method for the simultaneous determination of Pu, U, Am, Cm, and 90Sr isotopes in swipe samples using four stacked anion and extraction chromatographic columns.20 In this work, a simultaneous separation method based on stacked five-column anion/extraction chromatography was examined to cover additional radionuclides, including 237Np, 147Pm, and 55Fe. Spike and blank samples were prepared for method verification, and hard-to-detect radionuclides in water and swipe were measured by α spectrometry, ICPMS, and LS counting.
n the past several decades, the development of nuclear fission for nuclear weapon testing and nuclear power generation have led to global spread of both artificial and natural radionuclides in the environment at ultratrace levels. The release of hazardous radionuclides to the environment as the consequence of nuclear incidents (e.g., Chernobyl and Fukushima) further intensified the health and environmental concerns about the use of nuclear power and the risks associated with nuclear proliferation. In response to a nuclear emergency, there is an increasing need to develop simpler, faster, and high analysis throughput radioanalytical methods, particularly for hard-to-detect radionuclides (with no γ emission) that are impossible or difficult to identify with field instrumentation. Actinide isotopes, radiostrontium, 147Pm, and 55 Fe are among the most important hard-to-detect radionuclides, due to their high radiotoxicities and abundant activation/fission productions during the operation of nuclear power reactors.1,2 Therefore, rapid and accurate determination of these radionuclides in various sample matrixes, including water, swipe, food, soil, and urine samples, etc., is often required for the purposes of environmental monitoring, workplace radiohazard characterization, and dosimetry assessment, as well as for nuclear forensic reasons.3−10 Although a large number of procedures for actinide,3−6,11 strontium,7,8,12,13 promethium,9,14 and iron10,15−17 radioisotopes in environmental and biological samples have been developed, many of them are very time-consuming/tedious to operate and cover only few radionuclides of interest. In cases that multiple radionuclides in the same sample are required for analysis, simultaneous or sequential analytical procedures, using state-of-art stacked column chromatographic separation Published XXXX by the American Chemical Society
■
EXPERIMENTAL SECTION Reagents and Standard Solutions. All the chemicals and resins used in this study were analytical grade or better. The anion-exchange AGMP-1 M (Cl− form, 100−200 mesh) resin was purchased from Bio-Rad Laboratories Canada Ltd. (Mississauga, ON, Canada), and the resin was slurry-packed in a 2 mL cartridge before use. The extraction chromatography resins, including TEVA, UTEVA (50−100 μm), DGA-N (normal, 50−100 μm), Sr Spec, and TRU resin prepacked in 2 mL cartridges, were obtained from Eichrom Technologies, Inc. (Lisle, IL, U.S.A.). Nitric acid, hydrochloric acid, hydrofluoric acid, ammonium hydroxide, hydrogen peroxide, sodium nitride, titanium oxychloride solution, cerous nitrate, titanium trichloride solution, and ethanol were supplied by Sigma-Aldrich Canada (Oakville, ON, Canada). A liquid Received: February 7, 2014 Accepted: May 6, 2014
A
dx.doi.org/10.1021/ac500572g | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
were prepared and the corresponding calibration curves were established to enable efficiency correction using TDCR. Details of TDCR Č erenkov counting for 90Y has been described in recent publication.23 Procedure. A flow diagram showing the radioanalytical procedure used for the simultaneous determination of difficultto-detect radionuclides is shown in Figure 1.
scintillation (LS) cocktail Ultima Gold AB was purchased from PerkinElmer Canada (Woodbridge, ON, Canada). The ultrapure water (UPW) was produced using a Millipore Direct-Q5 ultrapure water system. Radiochemical isotope standard solutions of 238Pu, 239Pu, 241 Pu, 242Pu, 230Th, 243Am, 241Am, and 244Cm were supplied by the National Institute of Standards and Technology (NIST, Gaithersburg, MD, U.S.A.), while the 237Np, 228Th, 232U, 233U, 236 U, 90Sr, 89Sr, 147Pm, and 55Fe standards were obtained from the Eckert & Ziegler Isotope Products Laboratories (Valencia, CA, U.S.A.). In-house highly enriched and low-enriched uranium standards were also used, and thermal ionization mass spectrometry (TIMS) was used to accurately calibrate the uranium isotopic ratios in these U standards. The TraceCERT Sr and Fe standards for ICP, used as the stable tracers, were purchased from Sigma-Aldrich Canada. Samples. To examine the performance of the procedures, water and swipe samples spiked with different radionuclides (including actinides, 89/90Sr, 147Pm, and 55Fe), along with the procedural blanks, were analyzed. For water analysis, seven UPW spike samples labeled as “WS#”, four processed sewage spike samples (the processed discharge effluents were collected from an outfall to Ottawa River at Chalk River Laboratories site) labeled as “PSS#”, and 13 procedural blank samples were prepared. In addition, two water standards of the proficiency testing from the ERA water company (Colorado, U.S.A.) were also analyzed to validate the water procedure. For swipe analysis, two swipe filter spike samples labeled as “SFS#” and five filter blanks were prepared by dripping the spike and tracer solutions onto the filter. Four swipe procedural spike samples labeled as “SRS#” and six reagent blanks were also prepared with no filter used. The samples were processed following through the entire swipe procedure. Apparatus. An Octete Plus α spectroscopy workstation (AMETEK/ORTEC Inc., Oak Ridge, TN, U.S.A.), with eight 450 mm2 ULTRA-AS ion-implanted silicon detectors, was used for the measurements of 230Th, 228Th, 232Th, 242Pu, 238Pu, 239/240 Pu, 243 Am, 241 Am, 242 Cm, and 243/244 Cm by α spectrometry. The default counting time by α spectrometry was 2 days, except otherwise specified. The Th, Pu, Np, and U isotopes, including 230Th, 232Th, 242 Pu, 239Pu, 240Pu, 241Pu, 237Np, 233U, 234U, 235U, 236U, and 238 U, were measured using a high-resolution sector field ICPMS (Element XR, Thermo Scientific, Bremen, Germany) coupled with a high-sensitivity desolvation sample introduction unit (Apex-Q, Elemental Scientific Inc., Omaha, NE, U.S.A.) and an autosampler. Detailed instrument setup and working conditions for Th, Pu, Np, and U by ICPMS have been previously described.21,22 Stable strontium and iron were also determined by ICPMS. A Hidex 300SL TDCR (triple to double coincidence ratio) liquid scintillation counter (Hidex Personal Life Science, Finland) was used for 89/90Sr, 90Y, 147Pm, and 55Fe analysis. The TDCR values measured during liquid scintillation and Č erenkov counting were used to evaluate the counting efficiencies. An excellent match between the TDCR and counting efficiency has been observed for pure β emitters (i.e., 89/90 Sr, 90Y, and 147Pm) by LS counting. However, discrepancies were found in the TDCR value and counting efficiency for high-energy β emitters (89Sr and 90Y) by Č erenkov counting and for electron capture emitter (55Fe) by LS counting. Thus, a series of external standards at different quenching conditions
Figure 1. Flow diagram of rapid analytical method for difficult-todetect radionuclides in water and swipe samples.
Prior to sample processing, known amounts of radioactive (230Th, 242Pu, 233U, and 243Am) and stable (Sr and Fe) yield tracers were either added to 20 mL of water sample in a 50 mL plastic centrifuge tube or dripped onto the swipe filter (47 mm Whatman grade 41 filter paper) folded in a precleaned porcelain crucible. The water samples were acidified to a concentration of 8 M HNO3 by adding equal volumes of concentrated HNO3. For swipe analysis, the filter was gently dried in a muffle furnace. It was then ashed by successively raising and holding the temperature at 200 and 400 °C for 1 h each, and finally combusted at 600 °C until no black carbon was observed. After cooling to room temperature, the ash was transferred into a Teflon beaker with minimal amount of concentrated HNO3, heated, and dissolved in concentrated HNO3 and/or HCl on a hot plate, with the addition of H2O2 if necessary. Once the ash was completely dissolved, the solution was evaporated to just dryness. The sample was redissolved in 10 mL of 8 M HNO3 for further processing. The total preparation time for swipe ashing and digestion is usually between 6 and 8 h, depending upon the type of the swipe filter used. To ensure consistent valences of actinides for chromatography separation, 0.2 mL of 3 M NaNO2 was added to the sample for valence adjustment. After sitting for 10 min, the sample was loaded onto the columns for simultaneous radionuclide separation. As illustrated in Figure 2, five anion-exchange and extraction chromatography cartridges, stacked from top to bottom in a sequence of AGMP-1M, UTEVA, DGA, Sr Spec, and TRU resins, were used for the simultaneous extraction of actinides, promethium, yttrium, strontium, and iron. The columns were preconditioned with 15 mL of 8 M HNO3 before extraction. The sample was poured into a 60 mL syringe holder connected on top of the columns and passed through the resins at a flow rate of ∼1 mL/min. The columns were then rinsed with 15 mL of 8 M HNO3 and split for elution. B
dx.doi.org/10.1021/ac500572g | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 2. Simultaneous column chromatography separation and analysis procedure for difficult-to-detect radionuclides.
Thorium was first striped off the AGMP-1 M resin using 8 mL of concentrated HCl, followed by the elution of Pu/Np with 15 mL of 0.2 M HNO3 + 0.05 M HF. The Th eluate was then split: an aliquot of 1 mL eluate was evaporated to almost dryness and dissolved in 5 mL HNO3 for analysis of long-lived 232 Th by ICPMS; the remainder was coprecipitated with hydrous titanium oxide (HTiO), centrifuged, dissolved in 15 mL of 1 M HCl, and eventually microprecipitated with CeF3 onto a Resolve filter for preparation of a thin-layer counting source to count 228/232Th by α spectrometry. Detailed CeF3 microprecipitation has been previously reported.24 Uranium was eluted off the UTEVA resin using 2 mL of 4 M HCl followed with 13 mL of 0.01 M HCl, and the 234,235,236,238U isotopes were analyzed by ICPMS. Trivalent Am, Cm, Pm (and other lanthanides), and Y isotopes were extracted by the DGA resin and eluted with 15 mL of 0.05 M HCl. The eluate was then equally split: one-half was used for the counting of Am and Cm isotopes by α spectrometry after CeF3 microprecipitation; the other half was transferred into a 20 mL plastic scintillation counting vial for immediate Č erenkov counting of short-lived 90Y followed with LS counting of both β-emitting lanthanides (147Pm, 144Ce, 151 Sm, and 154/155Eu, if present in the sample) and 90Y in their β counting windows. The strontium extracted onto the Sr Spec resin was eluted with 8 mL of UPW. After elution, Č erenkov counting was first performed to measure 89Sra high-energy β emitter which produces sufficient Č erenkov light for detection. Then, a small aliquot (∼0.2 mL) of the Sr eluate was diluted to 50 mL in 0.1 M HNO3 for yield determination of stable Sr by ICPMS. The remainder of the eluate was mixed with 12 mL of Ultima Gold AB LS cocktail for the counting of 89/90Sr isotopes. The LS counting was also repeated ∼10 days later to check for the 90Y in-growth. Strontium-85, if available and not present in the sample, could also be used as the tracer for measuring 89/90Sr by LS counting to eliminate the need of additional yield determination of stable Sr by ICPMS.
Iron was adsorbed onto the TRU resin and striped into 8 mL of 0.02 M HNO3. A small fraction of the eluate was used for yield analysis of stable Fe by ICPMS, and the remainder was mixed with 12 mL of LS cocktail and counted for 55Fe by LS counting.
■
RESULTS AND DISCUSSION Simultaneous Column Separation. Stacked anion and extraction chromatographic cartridges coupled with a vacuum box system were employed for the simultaneous extraction and separation of actinide, lanthanide/yttrium, strontium, and iron in a single loading pass. Thus, division of the sample is not needed when analyses of multiple radionuclides are required. This significantly shortens the sample processing time and minimizes the waste production. The utilization of a vacuum box system streamlines the column separation and maximizes the sample analysis throughput. The simultaneous column separation procedure is simple, fast, and easy to operate for emergency preparedness. Using a couple of 12-hole vacuum boxes, the entire radiochemical separation procedure can be completed in 6−8 h for a batch of 12 water samples. After simultaneous extraction, the columns can be split for elution in parallel, provided that enough equipment and laboratory space are available. The separation scheme can also be adapted to fit specific analytical needs. For instance, UTEVA, Sr Spec, and TRU cartridges can be removed if the analysis of U, Sr, or Fe is not required. The AGMP-1 M cartridge can also be replaced by the TEVA cartridge for the separation of Pu, Np, and Th. When the TEVA resin is chosen, it is recommended to use 0.1 M HCl + 0.01 M HF for elution of Pu and Np, and other experimental conditions remain unchanged. The DGA resin showed good affinity for trivalent Am, Cm, lanthanides, and Y, and no notable chemical differentiation between these elements was observed in the DGA eluate under the experimental conditions used in this work. Therefore, utilization of nonisotopic tracer 243Am to correct for the Cm, Pm, and Y recovery was justified. Due to continuous β energy C
dx.doi.org/10.1021/ac500572g | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Table 1. Results of Average Chemical Recoveries for Water and Swipe Samples av chemical recovery water (%) Am/Pm Pu/Np Th U Sr Fe
101.0 96.6 103.4 100.6 86.2 98.5
± ± ± ± ± ±
swipe without filter (%)a
swipe with filter (%)b
95.1 ± 3.4 94.0 ± 10.9
96.0 ± 8.4 89.4 ± 10.3
83.7 ± 3.0 95.3 ± 5.1
80.8 ± 3.8 90.1 ± 1.8
6.9 4.3 3.8 3.0c 2.2 1.7
Samples processed following the entire swipe analysis procedure with no swipe filter used, to examine any potential loss in chemical recovery due to the interferences of the filter ash on the performance of column separation. bSamples processed following the entire swipe analysis procedure with swipe filter used. cAverage chemical recovery for U was determined using replicate samples with 232U tracer by α spectrometry. a
Table 2. Results of Average Procedural Blanks and Minimum Detectable Activity (MDA) for 20 mL Water Samples procedural blank (mBq) isotopes
α spectrometry
241
0.26 ± 0.24 0.03 ± 0.04 0.22 ± 0.17
Am Cm 243/244 Cm 237 Np 238 Pu 239 Pu 240 Pu 241 Pu 228 Th 232 Th 234 U 235 U 236 U 238 U 89/90 Sr 147 Pm 55 Fe 242
0.08 ± 0.04 0.08 ± 0.04d
0.22 ± 0.17 0.04 ± 0.02
a
ICPMS
MDA (mBq) LSC/Č erenkovb
α spectrometry
a
ICPMS
LSC Č erenkovb
MAC (Bq/L)c
140 250 210
0.67 12 1.1 1.3 0.6 0.55 0.55 29 1.9 0.59 2.8 2.9 2.9 3.1 54/5 525 388
0.33 0.15 0.35 1.1 ± 1.5 × 10−3
0.005 0.27 0.26d
0.096 ± 0.014 0.063 ± 0.022 9.6 ± 6.1 1.9 0.013 4.0 5 1.0
± ± ± ± ±
0.41 0.21
0.2 × 10−4 0.003 0.8 × 10−5 5 × 10−5 0.2 × 10−3
