Abstract
A new rapid, safe, and cost-effective Sr separation method to determine 89Sr and 90Sr in water samples has been developed. First, Ba and Ra were precipitated with chromate at pH 6–7 and Sr was precipitated with sulfate. Then, the potential interferences in the SrSO4 salt were chelated and dissolved with EDTA. The dry SrSO4 precipitate was weighed to determine the chemical recovery and then dissolved. Strontium-89 and strontium-90 were measured by liquid scintillation counting (LSC). The chemical recovery was optimized (∼75%), the decontamination factors for potential interferences were determined, the method was validated using spiked samples, and then the method was applied to environmental water samples.
Introduction
Strontium-89 and strontium-90 are anthropogenic isotopes, mainly produced during the fission of 235U in nuclear reactors (4.690 ± 0.057% and 5.73 ± 0.13% fission yield, respectively). These isotopes are both pure beta emitters and have radioactive half-lives (t 1/2) of 50.57 ± 0.03 d and 28.80 ± 0.07 y, respectively. Also, 90Sr decays to 90Y (t 1/2 = 2.6684 ± 0.0013 d) and is in complete equilibrium with its progeny in about 3 weeks. The highest activity concentrations of 89Sr and 90Sr are mainly found at nuclear sites, but these isotopes have been dispersed to the environment due to nuclear installations, atmospheric nuclear detonations, and nuclear accidents such as the ones that occurred at Chornobyl and Fukushima. − These Sr isotopes emit high-energy beta particles. As they are health concerns, it is important to have effective and accurate analytical methods to determine them in bioassay and environmental samples.
Strontium-89 and strontium-90 are often measured by either liquid scintillation counting (LSC) or proportional counting. , These measurement methods are relatively sensitive and accessible, but most of the time cannot discriminate the beta emitters measured. Therefore, Sr needs to be well separated from other elements to prevent potential interferences. The main methods used to separate Sr from potential interferences are precipitation, extraction chromatography (EXC), solid phase extraction (SPE), and ion exchange chromatography (IEC). Solvent extraction methods are occasionally used. , A detailed summary of Sr separation methods has been published by Vajda and Kim. Note that90Sr can also be determined with mass spectrometry methods, such as ICP-MS, if 89Sr is absent. In addition, 90Sr can be indirectly determined using 90Y after ingrowth and chemical separation.
The selective precipitation of Sr with Sr(NO3)2 is the classical method to separate Sr. − Sr(II) is first precipitated with fuming HNO3 as Sr(NO3)2, leaving most of the Ca(II) in solution. The Sr(NO3)2 precipitation is repeated until enough Ca(II) is removed. Then, Sr(NO3)2 is dissolved in water and Ba(II) and Ra(II) are precipitated with chromate at pH 4–6. The metals still present in the solution and the Y(III) are coprecipitated with Fe(OH)3 at pH 8–9. This method is simple, but a major drawback is the use of fuming HNO3, which is hazardous.
Strontium can be separated using EXC resins and SPE discs, which contain a crown ether (CE) selective to Sr. For the EXC method, the CE is dissolved in a solvent adsorbed onto a resin, and for the SPE method, the CE is covalently bonded to a filtering membrane. − The acidified water sample is passed through the support (EXC resin or SPE disc). The support is rinsed to remove impurities and the Sr is eluted. These methods are straightforward and safe but very costly, because CEs are expensive.
The IEC methods are usually less selective than the ones previously mentioned and employed less often, but recently Xu et al. developed an improved IEC separation method for Sr. Strontium was preconcentrated with carbonate and dissolved with HNO3. The solution was evaporated to dryness and the nitrate salt dissolved with a diethyldithiocarbamic acid (DTCA) solution at pH 6. The solution was passed through the IEC resin and Sr was eluted with 6 M HNO3. Under these conditions, Ca was not retained by the resin. This method is safe and robust and can be very useful for complex samples with a high Ca content. However, it is time-consuming and large volumes of acidic and basic solutions are needed.
The recovery tracers usually used to determine 89Sr and 90Sr are stable Sr and 85Sr. Both recovery tracers have to be measured with an instrument other than LSC, which increases the analysis cost and the time needed to obtain results. When performing the Sr(NO3)2 precipitation method, it is possible to determine the chemical recovery by precipitating Sr as SrCO3 before the LSC measurement and by weighing the dry salt obtained. When carrying out EXC, SPE, and IEC methods, the amount of Sr that can be added is limited by the number of active sites. For this reason, less Sr can be added to the sample and Sr is often measured by inductively coupled plasma mass spectrometry (ICP-MS). If 85Sr is used as a recovery tracer, Sr is determined by gamma spectrometry.
The methods available to separate Sr are limited and are either hazardous or costly. A potential solution could be to precipitate Sr in the presence of a chelate, which would complex the potential interferences and prevent them from precipitating. This method, often named the masking agent method, is sometimes used to determine 226Ra. In that case, Ra is coprecipitated with BaSO4 in the presence of a chelate. The chelate has a stronger affinity for the potential interferences than for Ba and Ra, preventing them from precipitating with the sulfate. The separation of 89Sr and 90Sr with the chelating separation method would be safe, rapid, and cost-effective.
In this work, a new and simple method to separate Sr based on the chelation of potential interferences is presented. First, Ba and Ra were removed by precipitating Ba(Ra)CrO4. Then, SrSO4 was precipitated and the remaining impurities in the SrSO4 salt were dissolved using EDTA. The SrSO4 salt was weighed to determine the chemical recovery and then dissolved to measured 89Sr/90Sr by LSC. The chemical recovery was optimized, the decontamination factor and figures of merit were determined, and the method was validated and tested.
Experimental Section
Reagents
All solutions used for this work were prepared using ultrapure water from a Millipore Direct-Q5 water purification system (Billerica, MA, USA). Trace-metal-grade acids (sulfuric acid (H2SO4) and nitric acid (HNO3)), trace-metal-grade salts (tetrasodium ethylenediaminetetraacetate (Na4EDTA) and sodium hydroxide (NaOH)), and reagent-grade salts magnesium chloride ((MgCl2), calcium chloride (CaCl2), strontium chloride (SrCl2), barium chloride (BaCl2), potassium chromate (K2CrO4)), silver nitrate (AgNO3), aluminum nitrate (Al(NO3)3), cobalt chloride (CoCl2), potassium chloride (KCl), manganese chloride (MnCl2), lead nitrate (Pb(NO3)2), zinc chloride (ZnCl2), zirconium oxychloride (ZrOCl2)), and bis-tris methane were purchased from Fisher Scientific (Fair Lawn, NJ, USA). TiCl3 in 12 M HCl was purchased from Sigma-Aldrich (Oakville, ON, Canada). Hionic-Fluor cocktail was obtained from PerkinElmer (Waltham, MA, USA). Certified solutions of 3H, 14C, 55Fe, 63Ni, 90Sr, 99Tc 129I, 209Po, 226Ra, 231Pa, 237Np, 239Pu, and 243Am were obtained from Eckert and Ziegler (Valencia, CA, USA). Standard solutions (1,000 mg L–1) of As, Au, Bi, Cd, Ce, Cs, Cr, Fe, Ga, Gd, Ge, Hf, Hg, In, Ir, Nd, Ni, Nb, Os, Pb, Pd, Pt, Re, Ru, Rh, Sb, Sc, Ta, Th, Tl, Sn, U, V, W, Y, Zn, and Zr were obtained from SCP Science (Baie-d’Urfé, QC, Canada).
Instruments
Strontium-90 was measured using a Hidex 300 SL liquid scintillation counter (Hidex Oy, Finland). The decontamination factors for potential interferences were determined using an Octete Plus alpha spectrometer with eight 450 mm2 ULTRA-AS ion-implanted silicon detectors (AMETEK/ORTEC, Oak Ridge, TN, USA), an Agilent 8800 ICP-MS (Santa Clara, CA, USA), and a Fisher UV–Visible absorption spectrometer (Fair Lawn, NJ, USA). All pH measurements were done with a pH meter (Fisher Accumet AE150, Fair Lawn, NJ, USA).
Procedure
A schematic representation of the method is shown in Figure .
1.
Schematic representation of the method.
1) Preparation and Preconcentration
A drying oven was preheated to 60 °C (Figure , step 1.1). A series of 50 mL centrifuge tubes, 15 mL centrifuge tubes, 20 mL polypropylene LSC vials, and filter funnels were labeled (step 1.2). Resolve filters (0.1 μm pore size) (Eichrom, Lisle, IL) were preweighed with their corresponding empty LSC vials (step 1.3). The filters were assembled and conditioned with about 2 mL of ethanol and 2 mL of water (step 1.4).
The water samples were filtered using 0.45 μm filters and preserved at pH ≤ 2 with HNO3. About 10 mL of sample were weighed in a 50 mL centrifuge tube (step 1.5) and 0.5 mL of a 1 M SrCl2 carrier solution was added (step 1.6). The amount of Sr carrier added (∼44 mg) is much higher than the typical amount of stable Sr found in most freshwater and seawater samples for a 10 mL water sample (∼0.011 mg and ∼0.08 mg, respectively) and can usually be considered negligible. , In case of doubt, the Sr concentration in the sample can be measured by either ICP-MS or ICP-OES (inductively coupled plasma optical emission spectroscopy).
The method description is for a 10 mL water sample, but it is possible to preconcentrate Sr from a larger sample volume (maximum 1 L) if a lower minimum detectable activity (MDA) is needed. In this case, 2 mL of 1 M Na2CO3 were added to the sample. Note that it is better to double the amount of Sr tracer to obtain higher recoveries when starting with a 1 L sample. The pH was adjusted to 10 using a 10 M NaOH solution (step 1.7).
The solution was mixed for 10 min using a magnetic stir bar. The carbonate salts were left to settle, the supernate was decanted. The carbonate salts were centrifuged (step 1.8). The carbonate salts were dissolved using 15.7 M HNO3 (the quantity added depended on the amount of precipitate obtained) and the solution was completed to 10 mL with water (step 1.8).
2) Barium Chromate
A volume of 0.2 mL of 0.8 M BaCl2 was added (step 2.1). Barium was precipitated by adding 0.2 mL of a 1 M K2CrO4 solution (step 2.2). The pH was approximately adjusted to 6–7 using 10 M NaOH and 15.7 M HNO3 as needed (step 2.3). The pH adjustment could be completed faster by adding 0.2 mL of a 1 M bis-tris methane buffer solution. The solution was left to react for 10 min and was then filtered (step 2.4). The eluate was collected in a 15 mL centrifuge tube. It was necessary to first remove Ba and Ra using chromate, because they were less strongly bound to the EDTA than Sr.
3) Strontium Sulfate
A volume of 0.05 mL of 18 M H2SO4 was added to the sample (step 3.1). The sample was mixed and left to react for 30 min (step 3.2). The SrSO4 precipitate was centrifuged (step 3.3), rinsed with 1 mL of water and centrifuged again (step 3.4).
4) EDTA
The SrSO4 precipitate was redispersed with 10 mL of water (step 4.1) in the same 15 mL centrifuge tube. Then, 0.4 mL of 15.7 M HNO3 was added (step 4.2) to obtain a pH of 4 after addition of the EDTA solution. A volume of 4 mL of 1 M EDTA was added to the sample to complex potential impurities. The tube contents were mixed for 30 s and SrSO4 was filtered on a preweighed filter (step 4.3). The filter was rinsed with 5 mL of water and then twice with 2 mL of ethanol (step 4.4).
5) Recovery and LSC Measurement
The filter was placed in its corresponding LSC vial and dried for 30 min at 60 °C in a drying oven (step 5.1). The vial and filter were weighed together to calculate the mass of SrSO4 obtained (step 5.1). Note that the SrSO4 mass was stable after drying as the salt did not adsorb any significant amount of humidity (result not shown). Strontium sulfate was dissolved by adding 1 mL of 1 M EDTA, 0.3 mL of 10 M NaOH, and 4 mL of water (steps 5.2–5.4). The sample was shaken for 4 to 5 min and the filter carefully removed (step 5.5). The sample was first counted to measure the Cherenkov emission (step 5.6). Then, 15 mL of Hionic-Fluor LSC cocktail were added (step 5.7). The LSC vials were left in the dark to reduce chemiluminescence (1 h for rapid measurement and 24 h for low-level measurement) and counted (e.g., 5–60 min) (step 5.8).
Calculations
Strontium-89 and strontium-90 activity concentrations were calculated based on Tayeb et al. calculation method. The efficiency was determined by preparing a graph of the efficiency as a function of the quenching index (triple-to-double coincidence ratio (TDCR) for this LSC). Atrazine was used as a quenching agent.
Method Development
The Sr chemical recovery was optimized at the SrSO4 precipitation step (step 3) and the EDTA step (step 4). For each test, the Sr recovery was determined by measuring the mass of SrSO4 obtained on a preweighed filter. In all cases, the filter was rinsed, dried, and weighed as described in the procedure (steps 4.4 and 5.1).
The optimal amount of 18 M H2SO4 to precipitate SrSO4 was first determined (step 3.1). Solutions of 10 mL of 0.05 M SrCl2 were prepared. Various known amounts of 18 M H2SO4 were added and SrSO4 was recovered on the filter after 5 min.
The time needed to maximize the SrSO4 precipitation yield was determined with and without Ca (step 3.1). For these tests,
Ten mL solutions of 0.05 M SrCl2 were prepared, various amounts of CaCl2 were added, and Sr was left to precipitate for a defined time after addition of 0.05 mL of 18 M H2SO4. When Ca was present, the CaSO4 formed was selectively dissolved by adding 15.7 M HNO3 and 4 mL of 1 M EDTA (the amount of 15.7 M HNO3 needed to obtain a pH of 4 after the addition of the EDTA was predetermined for each concentration of Ca). The solution was mixed for 30 s and filtered as described in the procedure (step 4). The same test was done with various concentrations of MgCl2 for a fixed reaction time of 30 min.
EDTA was used to dissolve the potential interferences in the SrSO4 salt, but it also slowly dissolved the SrSO4 salt. Therefore, SrSO4 recovery at that step (step 4.3) was optimized by studying the effect of time, pH, and EDTA concentration. Solutions of 10 mL of 0.05 M SrCl2 were prepared, 0.05 mL of 18 M H2SO4 were added, and SrSO4 was left to precipitate for 30 min. Then, 15.7 M HNO3 was added to adjust the pH (the amount to obtain the desired pH after addition of EDTA was predetermined using a surrogate test tube). When testing for Sr recovery as a function of time at various pH values, 4 mL of 1 M EDTA were added. When testing for the effect of EDTA concentration, various amounts of 1 M EDTA were added. The solutions were mixed for 30 s and filtered as described in the procedure (step 4).
Decontamination Factor
The decontamination factor (DF) for potential interferences was determined by spiking 10 mL water samples with a known activity or amount of a potential interference: 50 Bq for the beta emitters measured by LSC (3H,14C, 55Fe, 63Ni, 99Tc, 129I,), 0.1 Bq for the alpha emitters measured by alpha spectrometry (226Ra, 237Np, 239Pu, 243Am,) except for 209Po, 232Th, and 238U, for which 1 Bq was added, 1 mg for the elements measured by UV–visible (Cr, Ru, Pd, Os), 1 g for the elements determined by gravimetry (Mg, Al, K, Ca, Ti, Mn, Co, Zn, Zr, Ag, Ba, Pb), and 0.1 mg for the other elements measured by ICP-MS (Sc, V, Cu, Ga, Ge, As, Rb, Y, Rh, In, Sn, Sb, Cs, Ce, Gd, Hf, Ta, W, Re, Ir, Pt, Au, Tl, Bi). Then, the separation method was applied and the remaining potential interferences measured.
The potential interferences had to be measured using different instrumental techniques to obtain the highest possible sensitivity. The sample preparation for each type of measurement differed slightly as follows:
- LSC: the SrSO4 salt was dissolved as described in the procedure and the radionuclide counted for 1 h.
- Alpha spectrometry: the potential interferences were dissolved from the SrSO4 salt using 10 mL of 0.1 M HCl. Strontium sulfate was insoluble and removed by filtration. Then, a thin-layer source of the different isotopes was prepared by microprecipitation (CeF3 for the actinides, BaSO4 for 226Ra, and CuS for 209Po). The alpha sources were counted for 48 h by alpha spectrometry.
- UV–visible: the potential interferences were dissolved from the SrSO4 salt using 10 mL of 2 M HCl. The SrSO4 precipitate was removed by centrifugation and the supernate transferred to a 20 mL glass vial. The solutions were measured using the maximum absorbance wavelength.
- ICP-MS: the SrSO4 salt was dissolved in 1 L of 0.16 M HNO3. Then, the solution was diluted by a factor of 4,000 in 0.16 M HNO3.
- Gravimetry: A soluble salt of the potential interference was added. The SrCl2 tracer was not added. The method was applied and the mass of salt obtained on the filter was determined (step 5.1). If the potential interference was precipitated by sulfate or as an oxide/hydroxide under these conditions, it would be retained on the filter. The DF factor was calculated based on the most likely salt to be obtained (sulfate, oxide, or hydroxide).
The DF factor was calculated as the amount of interference added divided by the amount measured. The result presented is the DF average of two replicates and their standard deviation.
Figures of Merit
The minimum detectable activity (MDA) was determined by preparing 10 blank samples with 10 mL, 100 mL, and 1,000 mL of ultrapure water. The procedure was applied. For the 100 mL and 1,000 mL samples, the carbonate preconcentration steps were performed (steps 1.7 and 1.8). The MDA was calculated based on ISO 11929. Also, spiked water samples of 10 mL were prepared with various amounts of90Sr to validate the method. Then, the procedure was applied. The MDA, mean relative bias, and precision were calculated. The counting efficiency for 89Sr was estimated to 50% to calculate the MDA based on the counting efficiency of 90Y using Cherenkov measurement because no 89Sr standard solution was available. The 89Sr counting efficiency value of 50% is comparable to that of other work.
Application
The developed method was tested using drinking water (DW) samples from the town of Deep River (ON, Canada), groundwater (GW) samples collected at the Chalk River Laboratories site (Chalk River, ON, Canada), and seawater (SW) samples collected from the St. Lawrence River near Baie-Comeau (QC, Canada). The drinking water and seawater samples were spiked with a known activity of 90Sr as they were not expected to contain any measurable radiostrontium activity concentration.
Results and Discussion
Method Development
The Sr recovery as a function of the volume of 18 M H2SO4 is presented in Table . The Sr recovery was similar for the H2SO4 volumes tested. It is preferable to minimize the amount of H2SO4 added to minimize the amount of CaSO4 formed and to facilitate its dissolution with EDTA. It was decided to add 0.05 mL of 18 M H2SO4 for each 0.5 mL of 1 M SrCl2 tracer added (molar excess of 1.8) (Figure , step 3.1).
1. Sr Recovery as a Function of 18 M H2SO4 Volume Added and Corresponding Molar Excess (n = 2, ± SD).
| H2SO4 mL | Molar excess | Sr recovery (%) |
|---|---|---|
| 0.05 | 1.8 | 90 ± 1 |
| 0.10 | 3.6 | 92 ± 1 |
| 0.20 | 7.2 | 86 ± 4 |
| 0.30 | 11 | 86 ± 7 |
| 0.40 | 14 | 82 ± 2 |
| 0.50 | 18 | 85 ± 2 |
The Sr recovery as a function of time for solutions containing various concentrations of Ca is presented in Figure . The optimal recovery was obtained after 30 min for Ca concentrations ≤0.25 M (∼70%). For the 0.25 M Ca solution, waiting 30 min significantly improved the Sr recovery (Figure ). It was decided to fix the reaction time to 30 min for that step (step 3.1) as the Ca concentration of a water sample is usually unknown.
2.
Sr recovery as a function of time for solutions of various concentrations of Ca (n = 2, ± SD).
For the test samples with a higher Ca concentration (≥0.25 M), no Sr was recovered regardless of the time. Only CaSO4 precipitate was observed. Under these conditions, it is most likely that Ca reacted faster than Sr with SO4 2–, preventing SrSO4 from forming even if SrSO4 is less soluble than CaSO4 (0.135 g L–1 (25 °C) and 2.02 g L–1 (20 °C), respectively).
The Sr recovery as a function of time for solutions containing various concentrations of Mg after 30 min of reaction is presented in Figure . Magnesium sulfate was soluble under these conditions, but an Mg concentration ≥0.1 M reduced the Sr recovery. High concentrations of Ca and Mg significantly reduced or prevented the precipitation of SrSO4. In the majority of environmental waters, Ca and Mg concentrations are below 0.25 M. , However, if a preconcentration step is needed for some samples concentrated in Mg and Ca (for example seawater), the developed method could not work without a prior reduction of the amount of Mg and Ca, for example by using a nitrate precipitation or hydroxide precipitation.
3.
Sr recovery as a function of Mg concentration (n = 2, ± SD).
After precipitation of Sr as SrSO4, the impurities were removed using EDTA (step 4). The Sr recovery as a function of time after adding the EDTA solution is presented in Figure for different pH values. The Sr recovery decreased as a function of time regardless of the pH due to the partial dissolution of SrSO4 with EDTA. The dissolution was slower at pH 4, and for this reason it was decided to perform the addition of EDTA at that pH. Strontium dissolves more slowly at lower pH because of the competition between the [SrEDTA]2– complex and [H2EDTA]2– complex. However, it is not advisible to perform the reaction at a lower pH than 4, because H4EDTA will precipitate. It was also decided to mix each sample for 30 s to ensure a full dissolution of CaSO4 and to minimize SrSO4 dissolution. Note that improper pH adjustment at that step (step 4), can significantly affect the chemical recovery.
4.
Sr recovery as a function of time at different pHs after adding 4 mL of 1 M EDTA (n = 2, ± SD).
The Sr recovery as a function of EDTA concentration is shown in Figure . The Sr recovery was constant (∼75%) for all concentrations of EDTA tested. Therefore, the SrSO4 dissolution rate did not depend on EDTA concentration.
5.
Sr recovery as a function of the EDTA concentration after mixing for 30 s at pH 4 before filtration (n = 2, ± SD).
Decontamination Factor
The DFs for the elements tested are presented in Table . The DF was ≥1,000 for about two-thirds of the elements tested. These values are estimations of the DF and not absolute DF values. The uncertainty in some DF values was high, because the net number of counts measured was extremely low (for example 1–3 counts for alpha spectrometry), which therefore led to significant standard deviations. Note that a high uncertainty due to a low number of counts in this situation, is a statistical effect and does affect the order of magnitude of the DF. Also, the DF values should be interpreted as the lowest measurable DF. The direct measurement of some elements in the SrSO4 salt was challenging and different measurement methods had to be used. It led to an underestimation of the real DF value for some elements. For example, the measured DF for Cs was six times lower than for K, when these values are expected to be similar. In addition, some elements were not measured, but were expected to behave similarly and have similar DF values. For example, trivalent lanthanides are expected to behave like Ce(III) and Gd(III).
2. Decontamination Factors (n = 2, ± SD).
| Ag ≥ 2,830 ± 40 | Fe ≥ 1,670 ± 50 | Os ≥ 1,000 ± 30 | Ta ≥700 ± 300 | |||
| Al ≥ 1,700 ± 300 | Ga ≥ 600 ± 100 | Pb ≥ 1,500 ± 50 | Tc ≥ 1,730 ± 10 | |||
| Am ≥ 983 ± 3 | Gd ≥ 480 ± 70 | Pd ≥ 360 ± 20 | Th ≥ 40,000 ± 18,000 | |||
| As ≥ 140 ± 10 | Ge ≥ 230 ± 30 | Po ≥ 30,000 ± 20,000 | Ti ≥ 1,630 ± 10 | |||
| Au ≥ 400 ± 100 | H ≥ 1,250 ± 10 | Pt ≥ 300 ± 100 | Tl ≥ 1,000 ± 100 | |||
| Ba ≥ 2,200 ± 10 | Hf ≥ 570 ± 50 | Pu ≥ 3,000 ± 1,000 | U ≥ 24,970 ± 80 | |||
| Bi ≥ 1,200 ± 100 | I ≥ 1,660 ± 10 | Ra ≥ 3,000 ± 2,000 | V ≥ 1,100 ± 100 | |||
| C ≥ 1,690 ± 10 | In ≥ 1,500 ± 200 | Re ≥ 640 ± 70 | W ≥ 80 ± 40 | |||
| Ca ≥ 3,460 ± 40 | Ir ≥ 1,000 ± 100 | Rb ≥ 600 ± 100 | Y ≥ 1,200 ± 600 | |||
| Ce ≥ 1,600 ± 500 | K ≥ 2,900 ± 300 | Rh ≥ 200 ± 100 | Zn ≥ 1,560 ± 50 | |||
| Co ≥ 1,700 ± 100 | Mg ≥ 3,800 ± 300 | Ru ≥ 810 ± 30 | Zr ≥ 1,300 ± 900 | |||
| Cr ≥ 1,000 ± 30 | Mn ≥ 1,580 ± 10 | Sb ≥ 200 ± 20 | ||||
| Cs ≥ 450 ± 50 | Ni ≥ 1,500 ± 20 | Sc ≥ 2,100 ± 200 | ||||
| Cu ≥ 350 ± 40 | Np ≥ 5,370 ± 10 | Sn ≥ 120 ± 10 |
Gravimetry.
Alpha spectrometry.
ICP-MS.
LSC.
UV–visible.
When measuring 90Sr in environmental samples, the main potential radioactive interferences are the isotopes from the 232Th, 235U, and 238U decay chains and other naturally occurring isotopes such as 3H, 14C, and 40K. At nuclear sites and near nuclear weapon testing and nuclear accident sites, other potential interferences could also be present such as transuranic elements (e.g., Np, Pu, Am, and Cm), neutron activation products (e.g., 60Co, 55Fe, and 63Ni), and fission products (such as 133Cs, 134Cs, 135Cs, 137Cs, 135I, 131I, 129I, 99 Mo, 99Tc, 93Zr, 147Pm, 149Sm, 106Ru, and 107Pd). All the potential interferences were well removed by the developed method. Also, 90Y was well removed. Note that if a laboratory is expecting a potential interferent to be present at a higher concentration than what was tested in this work, it is recommended to verify that the DF is sufficient before using the developed method.
The developed method allowed for a rapid removal of potential interferences due to a combination of steps. The BaCrO4 precipitation step (steps 2.1 to 2.4) removed several elements as chromate, oxides and hydroxides including lanthanides, actinides, Ra, Ba, and most metals and metalloids. The precipitation of SrSO4 (steps 3.1 to 3.3) enabled the removal of soluble potential interferences such as Cs and K. Finally, EDTA dissolved Ca and the remaining potential interferences still present (step 4.3).
Figures of Merit
The MDA for water samples was determined for different counting times and volumes, and the results are shown in Table . A longer counting time and a larger sample volume led to a lower MDA. The MDAs obtained are typical of LSC counting analyses of 89Sr and 90Sr and comparable to those of other methods. , The MDA was determined for samples counted soon after separation. Lower MDAs could be obtained for 90Sr by waiting for the ingrowth of 90Y, which would double the number of counts for the sample. The counting time and sample volume will need to be adjusted based on the laboratory analysis objectives.
3. MDA for Water Samples of Various Sample Volumes and Counting Times in Bq L–1 (n = 2, ± SD).
| Counting time (min) | 5 | 10 | 60 | 5 | 10 | 60 |
|---|---|---|---|---|---|---|
| Volume (mL) | 90Sr | 89Sr | ||||
| 10 | 37 ± 4 | 26 ± 4 | - | 44 ± 6 | 31 ± 6 | - |
| 100 | 3.1 ± 0.3 | 2.1 ± 0.2 | - | 3.7 ± 0.5 | 2.6 ± 0.5 | - |
| 1,000 | 0.4 ± 0.1 | 0.27 ± 0.05 | 0.11 ± 0.02 | 0.5 ± 0.1 | 0.33 ± 0.05 | 0.12 ± 0.02 |
The developed method had MDA values that could suit most environmental water analysis needs. For example, the guideline levels for safe consumption of drinking water containing 90Sr in a routine situation are 5 Bq L–1, 0.3 Bq L–1, and 10 Bq L–1 according to Health Canada, the Environmental Protection Agency (EPA), and the World Health Organization, respectively. − The operational intervention levels for 90Sr in the event of a nuclear emergency is 30 Bq L–1 according to Health Canada. However, the method recovery might not be sufficient for some applications where low MDAs are required, especially when the concentrations of Mg and Ca are high. The laboratory will have to determine if the method can be used.
Spiked solutions of blank water samples with known activities of 90Sr were analyzed using the developed method, and the results are shown in Figure . The activity measured corresponded to the activity added. A mean relative bias and relative precision of −6.44% and 14.7%, respectively were obtained for 10 min counts by LSC for the samples above the MDA (26 ± 4 Bq L–1). The spiked solution test confirmed that stable Sr was a good recovery tracer for this method and enabled accurate determination of this method’s recovery.
6.
Spiked samples (n = 2, ± SD).
Application of the Method
The method was successfully used to determine 90Sr in some environmental water samples, and the results are presented in Table . The 90Sr activity added corresponded to the 90Sr activity measured for the spiked environmental water samples. The developed method could successfully determine 0.3 Bq L–1 of 90Sr in drinking water, which is the EPA guideline value. A lower recovery was obtained for the seawater samples as expected due to the high Ca and Mg concentrations. Some groundwater samples also had low recoveries. It is believed that these samples had higher concentrations of Ca and Mg, which is relatively common for groundwater samples (note that Ca and Mg concentrations were not determined). Seawater and groundwater samples were chosen to demonstrate that the developed method can be used to determine radiostrontium in very challenging samples; thus, it can be used for a wide variety of water samples with equally or less complex matrices Table .
4. Application of Method to Determine 90Sr Activity Concentration in Some Environmental Water Samples (n = 2, ± Sd) .
| Sample | Counting time | Volume | Activity concentration added | Activity concentration measured | Recovery | Bias |
|---|---|---|---|---|---|---|
| ID | (min) | (L) | (Bq L–1) | (Bq L–1) | (%) | (%) |
| DW1 | 60 | 1 | 0.34 ± 0.01 | 0.30 ± 0.04 | 62 ± 1 | –11 |
| DW2 | 60 | 1 | 0.34 ± 0.01 | 0.34 ± 0.03 | 72 ± 1 | 1 |
| GW1 | 10 | 0.01 | - | 570 ± 10 | 62 ± 1 | - |
| GW2 | 10 | 0.01 | - | 180 ± 10 | 45 ± 1 | - |
| GW3 | 10 | 0.01 | - | 1,250 ± 30 | 35 ± 1 | - |
| GW4 | 10 | 0.01 | - | 60 ± 5 | 62 ± 1 | - |
| GW5 | 10 | 0.01 | - | 37 ± 5 | 58 ± 1 | - |
| GW6 | 10 | 0.01 | - | ND | 60 ± 1 | - |
| GW7 | 10 | 0.01 | - | ND | 60 ± 1 | - |
| GW8 | 10 | 0.01 | - | 970 ± 20 | 61 ± 1 | - |
| GW9 | 10 | 0.01 | - | 6,360 ± 60 | 62 ± 1 | - |
| GW10 | 10 | 0.01 | - | 2,350 ± 30 | 67 ± 1 | - |
| SW1 | 10 | 0.01 | 840 ± 30 | 890 ± 40 | 25 ± 1 | 6 |
| SW2 | 10 | 0.01 | 800 ± 30 | 780 ± 30 | 35 ± 1 | –2 |
DW: drinking water, GW: groundwater, SW: seawater, ND: not detected.
5. Comparison between the Main Sr Separation Method Types and the Developed Method .
| Reagent and material cost | Separation time | Safety concerns | Waste concerns | Matrix limitations | |
|---|---|---|---|---|---|
| EDTA | low | rapid | - | chromate | Maximum Ca concentration of 0.25 M and reduce recovery if Mg concentration ≥0.1 M. |
| Nitrate precipitation | low | long | fuming HNO3 | chromate | Requires several nitrate precipitations depending on Ca concentration. |
| EXC resins and SPE discs using a CE | high | rapid | - | - | Reduce recovery if Ca ≥0.5 M on EXC resins. Strontium amount is limited to the number of active sites. |
| IEC | low | moderate | - | - | Can tolerate high Ca concentrations. Strontium amount is limited to the number of active sites. |
| Solvent extraction | high | rapid | solvents | solvents | Depends on the amount of solvent and/or extracting agent used. |
Developed method.
Extraction chromatography.
Solid phase extraction.
Crown ether.
Ion exchange chromatography.
Note that a comparison for the MDA is not mentioned as it depends on too many factors such as the sample volume, the instrument used, the counting time, and the chemical recovery. Similarly, the chemical recovery was not compared because it depends on the separation method itself, but also other factors such as the sample volume and the sample matrix, which makes the comparison difficult unless it is performed in the exact same conditions.
Reagent and material cost and separation time descriptors were interpreted based on ref .
Method Performance
It was possible to comfortably process 10 samples in half a workday for one person using this method (∼4 h). The separation steps took about 2 h (from the BaCrO4 step (step 2.1) to the isolation of the SrSO4 precipitate (step 5.1)). In comparison, it would take about 1.5 h to process 10 samples using an EXC method. The methods’ analytical throughputs were similar, but the developed method was about 5 times less expensive (see Supporting Information for a detailed cost analysis). The recovery was quickly determined with the mass of SrSO4 and thus no ICP-MS nor ICP-OES measurements were needed, which also saved time and reduced expenses. The developed method was very safe in comparison with the fuming nitric acid method. The reagents used were easily accessed by the laboratories and did not use any proprietary technology. The method could be useful in the event of a nuclear emergency due to its simplicity and rapidity.
The method has three main drawbacks, which can be partially mitigated. First, chromate solutions are toxic for the environment. Chromate salts and solutions produced during the method, can be kept apart and then easily reduced with ammonium iron(II) sulfate ((NH4)2SO4·Fe(SO4)2) to Cr(III) in a large beaker. Chromium(III) has a low toxicity (small amounts are essential for living beings) and Cr(III) salts and solutions can be safely disposed according to national and local regulations. Second, adjusting the pH at the chromate step can be time-consuming. Because environmental water samples are normally acidified with a known amount of HNO3, the amount of base needed to adjust the pH should be relatively similar for all samples. The use of a buffer would compensate for the small pH variations. Third, the method was limited by the Mg and Ca concentrations. In theory, it would be possible to reduce the concentration of Mg and Ca in the sample prior to the separation method by first precipitating Sr and Ca as carbonate and Mg as hydroxide. Then, the carbonate would be decomposed to CO2 by adding HNO3 and Ca and Mg would be selectively precipitated from Sr as hydroxide by adjusting the NaOH concentration to 0.2 M. This method was proposed and described by Chen et al.
Conclusion
A new method to separate and determine 90Sr in water samples using a chelate (EDTA) has been developed. This method enabled the effective removal of potential interferences using simple precipitations (BaCrO4 and SrSO4) and EDTA. The chemical recovery has been optimized. The method can be applied to most environmental water samples, but the chemical recovery will be lower for samples containing high concentrations of Mg and Ca. The method was validated and successfully applied to environmental water samples. This new separation method is safe, rapid, and cost–effective, and is an interesting alternative to current methods. It could be a useful method for routine monitoring and emergency situations.
Supplementary Material
Acknowledgments
This work was supported by Atomic Energy of Canada Limited under the Canadian Federal Science and Technology Program (project no. 51200.65.21.01.08).
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c10844.
Detailed cost analysis of the EXC resin method and the developed method (PDF)
All authors contributed to the writing of the manuscript, and all have given approval to the final version.
The authors declare no competing financial interest.
References
- Nichols, A. L. ; Aldama, D. L. ; Verpelli, M. . Handbook of Nuclear Data for Safeguards. In Database Extensions Cumulative; International Atomic Energy Agency: Vienna, Austria, 2008. [Google Scholar]
- Laboratoire National Henri Becquerel, online reference, last consulted, 2023. http://www.lnhb.fr/nuclear-data/nuclear-data-table.
- Maderich V., Jung K. T., Bezhenar R., de With G., Qiao N., Casacuberta N., Masque P., Kim Y. H.. Dispersion and fate of 90Sr in the Northwestern Pacific and adjacent seas: Global fallout and the Fukushima Dai-ichi accident. Sci. Total Environ. 2014;494:261–271. doi: 10.1016/j.scitotenv.2014.06.136. [DOI] [PubMed] [Google Scholar]
- Masson O., Romanenko O., Saunier O., Kirieiev S., Protsak V., Laptev G., Voitsekhovych O., Durand V., Coppin F., Steinhauser G.. et al. Europe-Wide Atmospheric Radionuclide Dispersion by Unprecedented Wildfires in the Chernobyl Exclusion Zone, April 2020. Anal. Chem. 2021;55(20):13834–13848. doi: 10.1021/acs.est.1c03314. [DOI] [PubMed] [Google Scholar]
- Volchok H. L.. Strontium 90: Estimation of Worldwide Deposition. Science. 1964;145(3639):1451–1452. doi: 10.1126/science.145.3639.1451. [DOI] [PubMed] [Google Scholar]
- Tazoe H., Yamagata T., Tsujita K., Nagai H., Obata H., Tsumune D., Kanda J., Yamada M.. Observation of Dispersion in the Japanese Coastal Area of Released 90Sr, 134Cs, and 137Cs from the Fukushima Daiichi Nuclear Power Plant to the Sea in 2013. Int. J. Environ. Res. Public Health. 2019;16(21):4094. doi: 10.3390/ijerph16214094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shao Y., Yang G., Tazoe H., Ma L., Yamada M., Xu D.. A review of measurement methodologies and their applications to environmental 90Sr. J. Environ. Radioact. 2018;192:321–333. doi: 10.1016/j.jenvrad.2018.07.013. [DOI] [PubMed] [Google Scholar]
- Letho, J. ; Hou, X. . Chemistry and Analysis of Radionuclides. Laboratory Techniques and Methodology; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2011. [Google Scholar]
- Vajda N., Kim C. K.. Determination of radiostrontium isotopes: A review of analytical methodology. Appl. Radiat. Isot. 2010;68:2306–2326. doi: 10.1016/j.apradiso.2010.05.013. [DOI] [PubMed] [Google Scholar]
- ISO. ISO 4721:2024 Water quality – Strontium 90 – Test method using ICP-MS; ISO, 2024. [Google Scholar]
- Xu P., Ding C., Yu G., Chen Z.. Determination of 90Sr in different matrices via ion-exchange chromatography and LSC. J. Radioanal. Nucl. Chem. 2022;331:3269–3274. doi: 10.1007/s10967-022-08401-2. [DOI] [Google Scholar]
- Thakur P., Ward A. L., González-Delgado A. M.. Optimal methods for preparation, separation, and determination of radium isotopes in environmental and biological samples. Environ. Radioact. 2021;228:106522. doi: 10.1016/j.jenvrad.2020.106522. [DOI] [PubMed] [Google Scholar]
- Health Canada Ottawa. Strontium in drinking water, Guideline technical document for public consultation; Health Canada Ottawa: ON, Canada, 2018. [Google Scholar]
- Holden, N. E. CRC Handbook Of Chemistry And Physics; CRC Press: Boca Raton, 2004. [Google Scholar]
- Tayeb M., Dai X., Sdraulig S.. Rapid and simultaneous determination of Strontium-89 and Strontium-90 in seawater. J. Environ. Radioact. 2016;153:214–221. doi: 10.1016/j.jenvrad.2016.01.003. [DOI] [PubMed] [Google Scholar]
- ISO. ISO 11929–1:2019 Determination of the characteristic limits (decision threshold, detection limit and limits of the coverage interval) for measurements of ionizing radiation Fundamentals and application Part 1: elementary applications; ISO, 2019. [Google Scholar]
- Health Canada. Guidelines for Canadian Drinking Water Quality: guideline Technical Document – Calcium; Health Canada, 1987. [Google Scholar]
- Health Canada. Guidelines for Canadian drinking water quality: guideline Technical Document – Radiological Parameters; Health Canada, 2010. [Google Scholar]
- U.S. Environmental Protection Agency. EPA Method 905.0: radioactive Strontium in Drinking Water, Prescribed Procedures for Measurement of Radioactivity in Drinking Water; U.S. Environmental Protection Agency, 1980. [Google Scholar]
- World Health Organization Global. Guidelines for drinking-water quality; World Health Organization, 4th ed., 2017. [Google Scholar]
- Health Canada. Generic Criteria And Operational Intervention Levels For Nuclear Emergency Planning And Response; Health Canada, 2018. [Google Scholar]
- Eichrom, Analytical method, strontium-89/90 in water, 2024. https://www.eichrom.com/wp-content/uploads/2018/02/srw01-15_sr-water-vbs.pdf.
- Chen O., Hou X., Yu Y., Dahlgaard H., Nielsen S. P.. Separation of Sr from Ca, Ba and Ra by means of Ca(OH)2 and Ba(Ra)Cl2 or Ba(Ra)SO4 for the determination of radiostrontium. Anal. Chim. Acta. 2002;466(1):109–116. doi: 10.1016/S0003-2670(02)00571-8. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.