Rapid and Accurate Determination of 226Ra from Drinking Water Using Macropa-Mediated Cloud Point Extraction Coupled with ICP-MS/MS Authors: Jegan Govindaraj, Myriame Proulx, Guillaume Blanchet-Chouinard, Steeve Roberge, Denis Giguère, Dominic Larivière Published: 2026 ABSTRACT A novel protocol has been developed for the preconcentration, separation, and accurate quantification of Radium-226 (226Ra) from drinking mineral water samples by cloud point extraction (CPE) coupled with inductively coupled plasma tandem mass spectrometry (ICP-MS/MS). To the best of our knowledge, this is the first report on using CPE to preconcentrate Ra. This method was extended to determine 226Ra concentration in mineral water samples collected from tap and bottled water. To preconcentrate 226Ra, CPE was performed using Macropa as a selective chelating agent and Triton X-114 as the nonionic surfactant. Factors influencing the extraction efficiency (namely pH, ligand concentration, surfactant concentration, and hydrophobizing and salting-out agents) were studied and optimized for quantitative extraction (>95%). Adding hydrophobizing agents, such as octylamine, cetylpyridinium chloride, and benzyltrimethylammonium chloride, significantly enhanced the capacity to extract Ra (>95%) via CPE. The optimized conditions were applied to preconcentrate 226Ra from 250 mL of water samples (tap and bottled), which were subsequently quantified by ICP-MS/MS. The limit of detection (LOD) of the method was 2 mBq L–1 (55 fg L–1), and the limit of quantification (LOQ) was 8 mBq L–1 (219 fg L–1), significantly lower than national and international drinking water guidelines. Introduction Radium-226 (226Ra) is the principal isotope of radium and a long-lived naturally occurring radioactive material (NORM) with t 1/2 = 1622 years, belonging to the 238U-decay series (4n + 2). The radioactive equilibrium in the 238U–206Pb decay series nuclides leads to environmental fractionation of the intermediate nuclide 226Ra, which, due to its high specific activity of 3.7 × 1010 Bq g–1, is recognized as one of the most toxic natural radiopollutants. − Due to its chemical similarity to other alkaline earth metals, particularly Ca2+, 226Ra follows the same pathway in biological organisms and is strongly adsorbed by bones, cells, and tissues. Sustained accumulation of 226Ra, combined with its high radiotoxicity, can cause life-threatening diseases such as bone cancer through ingestion, lung cancer via inhalation of the decay products of 222Rn, and skin cancer from prolonged external exposure to contaminated soil or building materials. , Hence, the tracking, monitoring, and measurement of 226Ra in environmental samples such as water, soil, and food are vital for public health protection. Among environmental media, water is considered one of the most significant sources of 226Ra contamination due to its extensive daily use in large volumes. Therefore, the separation and quantification of 226Ra are essential to safeguard public health by ensuring compliance with standard regulations and guidelines. In Canada, the per-capita water withdrawal footprint is 960 m3/person/year, of which 13.43% (129 m3/capita/year) is used by municipalities for drinking-water supply. , In addition, bottled and canned water are also consumed in large quantities. The United States Environmental Protection Agency (USEPA) has established regulatory standards, including USEPA standards 901.1, 903.0, 903.1, and 903.2, that limit 226Ra in drinking water to 0.183 Bq L–1 (5,000 fg L–1). Health Canada guidelines set the maximum acceptable concentration (MAC) to 0.5 Bq L–1 (13,700 fg L–1), a limit also followed by the province of Québec. − Quantification of 226Ra in drinking water remains scarce due to its ultratrace level and the need for a rapid, simple, and effective preconcentration method before its analysis. − Methods such as precipitation/coprecipitation, ion exchange, and solvent extraction have the following typical disadvantages: they are labor-intensive and time-consuming, show inconsistent elemental selectivity, generate substantial chemical waste, are not cost-effective, and remain difficult to scale to large volumes. Therefore, they are not eco-friendly practices. − Conventional coupled detection techniques such as liquid scintillation counting (LSC), radon emanation, gamma spectrometry, and alpha spectrometry suffer from limited sensitivity and often require long acquisition times. Therefore, it is important to develop new preconcentration and quantification methods to overcome these limitations and to ease the determination of 226Ra in various water samples. Cloud point extraction (CPE) is an efficient preconcentration technique established for extracting organic molecules and trace elements from complex matrices. This method employs surfactants that form micellar aggregates with the analytes of interest at a specific temperature (known as the cloud point temperature), leading to phase separation, which is typically achieved by centrifugation. − CPE has never been used to preconcentrate 226Ra from water or any other matrices. Given its ubiquitous versatility, we decided to explore CPE as an effective preconcentration strategy. We selected Triton X-114 (Figure ) as a surfactant due to its low cloud point, which facilitates phase separation at 23 °C. − Choosing a ligand that selectively complexes 226Ra is more challenging because of its chemical similarities to other alkaline earth metals such as Mg2+, Ca2+, Sr2+, and Ba2. Macropa, a diaza crown ether with two picolinic acid side chains, originally known as H2bp18c6 (Figure ), , has recently been developed and used as a selective chelator for Ra in cancer therapy. Chemical structure of the ligand Macropa and the surfactant Triton X-114. Macropa exhibits a strong binding affinity for Ra, with radiolabeling efficiencies exceeding 95%, − and was therefore selected as a complexing agent for this study. Macropa was synthesized as previously reported, with a purity exceeding 98%. , We systematically studied and optimized CPE parameters such as pH, concentration of surfactant and add-on surfactant, ligand concentration, temperature, time of extraction, and the presence of a salting-out agent to achieve the quantitative recovery of 226Ra. ICP-MS/MS is a highly sensitive and precise analytical technique widely used to detect trace levels of 226Ra in water samples. This method offers detection limits down to the picogram per liter (pg L–1). − Its rapid analysis, high precision, and excellent sensitivity make it ideal for quantifying 226Ra in drinking water. However, ICP-MS/MS analysis faces challenges due to spectral and nonspectral interferences that are inherent to the sample matrix. Isobaric, polyatomic, and molecular ion interferences can often be mitigated using a collision/reaction cell or during the CPE procedure, though not always with complete effectiveness. The ionizing efficiency of plasma may also be reduced by matrix elements such as Mg2+ and Ca2+, but these effects can be mitigated during the preconcentration step. Therefore, we chose ICP-MS/MS to quantify 226Ra. The Macropa-assisted CPE method was initially demonstrated on simulated water samples, and we applied it to real mineral water samples to quantify 226Ra. The results indicate that Macropa is an effective chelating agent to preconcentrate Ra, and ICP-MS/MS is a robust technique capable of detecting 226Ra at the fg to pg L–1 scale. Experimental Section Materials Pyridine dicarboxylic acid (99%, TCI America, Portland, USA), sodium carbonate (99%, Merck, Oakville, ON, Canada), and hexane (99%, BDH Chemicals, Mississauga, ON, Canada) were used as received. Sodium borohydride (99%), dichloromethane (99.8%), dimethylformamide (99.8%), sodium dodecyl sulfate (99%), octylamine (99%), benzyltrimethylammonium chloride (99%), and thionyl chloride (99%) were purchased from Sigma-Aldrich, Oakville, ON, Canada, and used without further purification. Methanol (99.8%), ethyl acetate (99.9%), hydrochloric acid (37–39%, trace metal grade), and nitric acid (67–69%, trace metal grade) supplied by Fisher Scientific, Ottawa, ON, Canada, were used as received. Cetylpyridinium chloride (99.96%) supplied by AmBeed, Arlington Heights, USA, was used as received. Triton X-114 surfactant supplied by VWR, Avantor, Radnor, PA, USA, was used as received. Macropa, a complexing ligand, was synthesized through a multistep reaction as previously reported. The detailed synthetic protocol and corresponding spectra, including 1H Nuclear Magnetic Resonance (1H NMR), 13C Nuclear Magnetic Resonance (13C NMR), and High-Resolution Mass Spectrometry (HR-MS), are provided in the Supporting Information (pages S4 to S12). ,, The Plasma Cal standard solution containing 1000 mg L–1 of 205Tl, purchased from SCP Sciences (Baie D’Urfe, QC, Canada), was used to prepare calibration solutions, internal standards, and spiking solutions. 226Ra obtained from the National Institute of Standards and Technology (NIST; Gaithersburg, MD, USA), was used as a calibration standard and to assess separation recovery. All other aqueous stock solutions and reagents were prepared by dissolving accurately weighed quantities of respective substances in high-purity deionized water (18 MΩ cm) obtained from a Milli-Q purification system of EDM Millipore (Etobicoke, ON, Canada). Additional details on the materials and solution preparation are provided on page S14 of the Supporting Information. The pH of the solution was adjusted with 0.1 M NaOH or HNO3. Instruments 1H and 13C NMR spectra were recorded by an Agilent DD2–500 MHz spectrometer, and 1H chemical shifts were referenced to the TMS proton resonance (all at 299 K, CDCl3). An Agilent 6210 LC-MS TOF instrument was used to confirm the molecular weight of the samples. 226Ra was analyzed by inductively coupled plasma mass spectrometry (ICP-MS/MS) using an Agilent 8900 instrument (Agilent Technologies, Santa Clara, CA, USA). Instrument specifications for the ICP-MS/MS are provided in the Supporting Information (page S13). We used a temperature-controlled centrifuge (Sorvall Legend XTR, Thermo Fisher Scientific, Bremen, Germany) for phase separation during the CPE. General Procedure for Cloud Point Extraction To a 15 mL centrifuge tube containing 10 mL of an aqueous solution containing 226Ra at pH 7 (0.453 Bq L–1), 50 μL of 1% aqueous Macropa solution was added and vortexed for 2 min. Subsequently, 0.25 mL of 10% Triton X-114 was added, mixed, and heated at 60 °C for 15 min to reach opalescence. The relatively higher density of the resulting micellar phase caused it to settle to the bottom of the tube under gravity. For complete phase separation and equilibrium, the mixture was immediately centrifuged at 13,000 rpm and 23 °C for 10 min using a temperature-controlled centrifuge. To increase the viscosity of the micellar phase, the tube was placed in an ice bath for 10 min after centrifugation. The schematic representation of the cloud point extraction is shown in Figure . During the optimization step, recovery was first calculated by mass balance in the aqueous phase (before and after three successive CPE) and later by comparing the initial aqueous phase and the micellar phase after performing the back-extraction. Back-extraction was performed by dissolving the micellar phase in 2 mL of 6 M HNO3. The CPE procedure was then repeated for three times, and after the phase separation, the aqueous phase was evaporated and redissolved in 3 mL of 4% (v/v) HNO3 prior to analysis. A 4% (v/v) HNO3 solution was used as the medium for ICP-MS/MS analysis with the required amount of internal standard 205Tl. The recovery yield of 226Ra was quantified using a mass balance approach, comparing radionuclide activities in the samples prior to and following CPE. For unknown samples, the yield was monitored by spiking a known amount of 226Ra into the samples themselves. As ICP-MS/MS was employed for direct quantification of 226Ra, no external tracer was added to monitor recovery yield. The extraction efficiency for water samples was calculated using eq . During each optimization step, only the parameters under investigation varied, while all other experimental conditions were kept constant. The detailed optimization conditions are presented in Table . 1 EE(%)=[Ra]f[Ra]i Schematic representation of cloud point extraction of 226Ra by Macropa. 1Volume of Solutions Added to the Various Conditions of OptimizationsOptimizations Volume of 10% TX-114 (mL)Volume of 1% Macropa (μL)pH[226Ra] (Bq L–1)[HA] (M)[Salts] (M)Effect of Triton X-1140.1 to 1.010.07.000.45--Effect of Macropa1.010.0 to 200.07.000.45--Effect of solution pH1.010.01.00 to 10.000.45--Effect of initial 226Ra1.010.07.000.10 to 50.00--Effect of additional HA1.010.07.000.450.01-Effect of various salts (I) 1.010.07.000.45-0.1aAbove solutions are added to 10 mL of 0.453 Bq L–1 of 226Ra at pH 7.bHydrophobizing agents = octylamine/benzyltrimethylammonium chloride.c I = ionic strength of the salt solutions. Volume of Solutions Added to the Various Conditions of Optimizations Above solutions are added to 10 mL of 0.453 Bq L–1 of 226Ra at pH 7. Hydrophobizing agents = octylamine/benzyltrimethylammonium chloride. I = ionic strength of the salt solutions. EE = Extraction Efficiency, f = final, and i = initial. Degree of Complexation in Cloud Point Extraction (α CPE ) Degree of complexation in the CPE procedure refers to the ratio between the amount of metal ions transferred to the surfactant-rich phase as metal–ligand complexes and the total amount of metal ions initially present in the sample. The α CPE value ranges from 0 ≤ α CPE ≤ 1. For quantitative recovery of metal ions, the value is generally expected to be greater than 0.95. 2 ∝CPE=[Metal−LigandComplex]Micellarphase[TotalMetal]Micellar+Aqueousphase Results and Discussion CPE is a simple, rapid, and effective preconcentration method for water samples containing metal ions at ultratrace levels. Various experimental parameters influence the efficiency of the CPE, and their optimization is essential before applying the method to environmental monitoring. The primary considerations include the concentrations of the surfactant, ligand, salting-out agents, and hydrophobizing agents. To facilitate complexation between 226Ra and the added ligand, suitable pH, temperature, and time of complexation are also optimized. Effect of the Solution pH pH is a crucial parameter influencing metal ion complexation, as it regulates the availability of ligand coordination sites and significantly affects the degree of complexation (α), defined as the ratio of metal ion concentrations in the solution before and after ligand addition, as shown in eq . At pH < 2, the ligand displays carboxylic acid and pyridinium functional groups, resulting in relatively lower α values. In contrast, at pH > 8, the ligand exhibits negatively charged carboxylates and pyridine rings, which show more favorable metal binding, forming stable complexes with slightly higher α values. However, working at pH > 8 can pose challenges due to the potential hydrolysis and precipitation of metal ions. Therefore, the effect of pH on the recovery of 226Ra was investigated at pH 2, 4, 6, 8, and 10, and the results are shown in Figure . The recovery was marginally higher (35 to 40%) in the pH range of 7 ≤ pH ≤ 10, which we attributed to the increased abundance of deprotonated ligands. In contrast, recoveries were minimal (<25%) in the range of 2 ≤ pH ≤ 6. Based on these findings, we chose pH = 7 for the CPE procedure. In real water samples, especially in bottled mineral water, working in the alkaline range (8 ≤ pH ≤ 10) may lead to precipitation/coprecipitation of metal ions due to hydrolysis. Recovery of 226Ra by cloud point extraction as a function of the pH of the solution. Effect of Triton X-114 Concentrations The concentration of surfactants is a critical variable to control, influencing the CPE extraction efficiency. An optimal concentration is essential to facilitate phase separation between the aqueous and micellar phases in the analyte solution. While higher surfactant concentrations can enhance extraction efficiency, exceeding the optimum value may reduce mass transfer due to increased solution viscosity. For effective phase separation and mass transfer, a typical viscosity range of 5–10 mPa·s is desirable. A viscosity of 10–100 mPa·s is considered moderate but may slow separation, whereas >200 mPa·s complicates the process. Conversely, at low surfactant concentrations, the system remains below the critical micellar concentration, resulting in poor extraction due to insufficient micelle formation and no phase separation. Kojro et al., , established that the optimal concentration range lies between 1%≤ C ≤ 10%. In the present study, we tested a solution of 10% (w/v) Triton X-114 in varying volumes (0.1 to 1 mL), and the results are presented in Figure . To maintain an optimal viscosity, volumes beyond 1 mL were not investigated. As the volume of surfactant increased, the extraction efficiency improved, reaching a maximum of 53% at 1 mL of 10% Triton X-114 (w/v). This is attributed to the greater volume of micellar phase formed at higher surfactant concentrations, which can entrap more Macropa-complexed 226Ra. Therefore, we chose 1 mL of 10% (w/v) Triton X-114 for further optimization. Recovery of 226Ra by cloud point extraction as a function of Triton X-114 concentration. Effect of Macropa Concentration The concentration of the ligand is a key factor in achieving quantitative extraction, as it significantly impacts both recovery and selectivity. At higher ligand concentrations, it may also complex with other alkaline-earth metal ions such as Ba2+, Ca2+, and Mg2+, thereby reducing selectivity. Conversely, low ligand concentration results in incomplete complexation, negatively affecting recovery. Therefore, optimizing the ligand concentration in CPE is a critical but challenging task to ensure high recovery and selectivity. To determine the optimum concentration for quantitative and selective extraction of 226Ra, we conducted experiments with various volumes (10 to 200 μL) of 1% (18.78 × 10–3 M) Macropa. The results, shown in Figure , indicate that in the presence of an excess of ligand, the extraction efficiency was reduced (35%). An experiment conducted without the ligand showed no extraction, confirming its essential role in complex formation. The efficiency increased inversely with ligand volume and reached a maximum at 10 μL of 1% Macropa. 3 M2++L↔ML+2H+ Recovery of 226Ra by cloud point extraction as a function of Macropa concentration. The equilibrium for the complexation reaction is shown in eq . The increase in extraction efficiency is attributed to a shift in equilibrium toward the products, according to Le Chatelier’s principle. However, in this case, an excess of Macropa led to a marginal decrease in efficiency due to a drop in pH, shifting the solution to acidic conditions where complex formation is significantly reduced. A maximum recovery of approximately 47% was achieved at 10 μL of 1% Macropa, which was selected as the fixed concentration for subsequent optimization. Effect of Salting-Out Agents We investigated the effects of various common salting-out agents on the CPE of 226Ra. Salting-out agents significantly influence phase separation and cloud point temperature. Maintaining an optimal ionic strength helps the phase splitting by enhancing the hydrophilicity of the aqueous phase and promoting the transfer of the metal–ligand complex into the micellar phase. This results in a higher distribution ratio and faster phase separation. Salting-out agents also lower the cloud point temperature. All reagents were tested at an ionic strength of I = 1 mol L–1. The results are shown in Figure . For comparison, a control experiment was performed under identical conditions without any salting-out agents. As expected, the addition of a salting-out agent slightly improves the recovery, though not quantitatively. The highest recovery of 60% was achieved with I = 1 mol L–1 of (NH4)2SO4. The observed trend is attributed to the progressively stronger salting-out effect of the added salts, in the order: No salt < Na3C6H5O7 < NaCl ≈ KNO3 < Na2SO4 < K3PO4 < (NH4)2SO4. The sample run with added (NH4)2SO4 showed a marginally (10 to 15%) higher recovery than the others. Recovery of 226Ra by cloud point extraction in the presence of different salts. Effect Of 226Ra Concentration The initial aqueous phase concentration of 226Ra has a significant impact on extraction efficiency. We varied it from 0.1 to 50 Bq L–1 to determine the concentration range for achieving maximum extraction for a fixed concentration of ligand, while keeping all other conditions constant. As shown in Figure , the extraction efficiency decreases beyond 1 Bq L–1, most likely due to insufficient ligand availability for Ra complexation. This reduction arises from competitive extraction between Ra and Ba present in the solution, consistent with the Ba content reported in NIST SRM 4967A (80 μg BaCl2 per gram of the original solution). Therefore, under the given conditions, we found that the maximum extraction (35–50%) was achieved within the concentration range of 0.1 to 10 Bq L–1. Recovery of 226Ra by cloud point extraction as a function of 226Ra concentration. Effect of Ionic Surfactants and Hydrophobic Agents We investigated the effect of adding an ionic surfactant and hydrophobizing agent on the extraction of 226Ra, and the results are shown in Figure . The primary role of auxiliary surfactants is to enhance the interaction between micelles and metal–ligand complexes through hydrophilic–lipophilic interaction. We studied the effect of adding hydrophobizing agents (anionic surfactant, cationic surfactant, and neutral long-chain organic molecule) on the extraction of 226Ra. We conducted the experiments in four ways: with sodium dodecyl sulfate (SDS, anionic), cetyltrimethylammonium bromide (CTAB, cationic), cetylpyridinium chloride (CPC, cationic), and octanoic acid (OA, neutral). These compounds act as hydrotropic additives by balancing the hydrophilic–lipophilic properties of the system and increasing the hydrophobicity of the metal complex through the formation of a ternary complex, thereby enhancing its micellar solubility. We found that the anionic surfactant SDS (emulsion formation) and the cationic surfactant CTAB (Figure ) did not have any positive impact on 226Ra extraction, while the other cationic surfactant, CPC (phase formation only at 0.1 mL), showed some enhancement in extraction efficiency. Among them, CPC, containing an aromatic ring, had a more favorable effect, likely due to the effective interaction between Triton X-114, CPC, and the Ra-Macropa complex. This behavior aligns with previously reported studies involving secondary surfactants such as CPC and OA and their synergistic effects. Recovery of 226Ra by cloud point extraction as a function of CTAB concentration. We found that using the phase transfer catalyst benzyltrimethylammonium chloride (BTACl) at a concentration of 0.1% enhanced the hydrophobicity of the complex by associating with the metal–ligand complex, thereby balancing the hydrophilic–lipophilic properties of the system. Therefore, we tested it for the extraction of 226Ra. The concentration was fixed at 0.1% for CPC and 10 mM for OA, based on the conditions reported in the literature, while all other parameters remained constant. We recovered more than 99% of 226Ra using 0.1% CPC. In order to obtain quantitative recovery, the CPE procedure was repeated 4 times for each of our analyte solutions. A similar behavior was observed with the addition of hydrophobizing organic molecules such as octylamine (10 mM) and the phase transfer catalyst benzyltrimethylammonium chloride (0.1% BTACl) in 0.27 mM of CTAB; results are summarized in Table . Therefore, these conditions were adopted for the treatment of the bottled mineral water samples. However, using the cationic surfactant and octylamine together did not enhance the extraction efficiency beyond that achieved by each agent alone. 2Effects on Added on Surfactants in 226Ra RecoveryReagentsConcentration in mM Volume of 10% TX-114 (mL) Volume of 1% Macropa (μL) pH% RecoveryCPC2.95 (0.1) 1.00507.0099 ± 4%BTACl-CTAB5.39 (0.1) -0.27 (0.01)1.00507.0099 ± 5%OA10 (0.129) 1.00507.0099 ± 5%aAbove solutions are added to 10 mL of 0.453 Bq L–1 of 226Ra at pH 7.bValues in parentheses represent concentrations expressed as % w/v. Effects on Added on Surfactants in 226Ra Recovery Above solutions are added to 10 mL of 0.453 Bq L–1 of 226Ra at pH 7. Values in parentheses represent concentrations expressed as % w/v. Selectivity and Recovery of Ra over Other Interfering Ions Macropa was originally developed for the complexation of Ba2+ and was subsequently applied to Ra2+ because of the close chemical similarity between these alkaline-earth metals. In this context, the present section examines Ra2+ recovery under CPE conditions in the presence of Ba2+ and other potentially interfering ions. NIST Standard Reference Material SRM 4967A was used in this study and contains 80 μg of BaCl2 per gram of solution, corresponding to 53.68 mg L‑1 Ba. The original stock solution (24,874 Bq L‑1) was diluted to prepare working solutions ranging from 0.5 to 50 Bq L‑1. At the highest calibration level (50 Bq L‑1), after mixing 10 mL of the solution with 0.1 mL of 1% (18.78 mM) Macropa, the final molar concentrations were as follows: [Ra] = 6.09 × 10–12 M, [Ba] = 7.86 × 10–7 M, [Ra + Ba] ≈ 7.86 × 10–7 M (Ra concentration is negligible), and [Macropa] = 1.88 × 10–4 M. Macropa is 3.1 × 107 times (31 million times) excess to Ra and 239 times excess to Ba. Thus, Macropa was present in large molar excess relative to the total divalent metal ions (Ra2+ and Ba2+). This confirms that the system was operated well below its complexation capacity, thereby preventing competitive saturation effects. As a consequence of its original design for Ba2+ complexation and its subsequent application to Ra2+ based on their close chemical similarity, inherent selectivity between Ra2+ and Ba2+ cannot be achieved through ligand discrimination alone in the CPE system. Such selectivity could possibly be achieved by fine-tuning the molecular structure of Macropa through the introduction of various substituents that induce steric strain or by altering the cavity diameter (Table ). In this regard, our research laboratory is currently exploring the binding capacity and selectivity of Macropa derivatives with different alkyl substitutions. 3Ionic Diameters and Cavity Size of Crown Ethers and Their Interactions IonsIonic diameter (Å)Competition with Ra2+ Crown with cavity (Å)Ra2+ 2.96 Å Target ion12-crown-4 (1.2–1.5)Ba2+ 2.84 ÅStrong competitor15-crown-5 (1.7–2.2)Sr2+ 2.52 ÅModerate competitor 18-crown-6 (1.6–3.2) Pb2+ 2.38 ÅWeak–moderate competitor12-crown-4 (3.4–4.3)La3+ 2.32 ÅWeak competitor Ce3+ 2.28 ÅWeak competitor Bi3+ 2.06 ÅVery weak competitor Ca2+ 2.24 ÅWeak competitor Mg2+ 1.78 ÅWeak competitor Na+ 3.02 ÅWeak competitor (low ionic potential) K+ 2.36 ÅWeak competitor (low ionic potential) aIonic potential = Z(charge)/r(radius). Ionic Diameters and Cavity Size of Crown Ethers and Their Interactions Ionic potential = Z(charge)/r(radius). Experiments were also performed by varying the Ra/Ba concentration ratio while maintaining Macropa in solution under optimized CPE conditions. We performed independent cloud point extraction (CPE) experiments by externally adding Ba to a known amount of standardized pure 226Ra solution (0.516 Bq/L, supplied by Center d’expertise en analyze environnementale du Québec, Quebec, Canada). The Ba was added at ratios of 1000:1 (Level 1), 1:1 (Level 2), and 1:1000 (Level 3), while maintaining a fixed Macropa concentration. Each level was analyzed in triplicate. Unfortunately, the results were not reproducible upon repeated attempts. For instance, when level 1 was run three times, the first result was inconsistent with the second and third runs. This issue will be further investigated using radiation-based counting techniques, such as alpha spectrometry or liquid scintillation counting (LSC) via 222Rn ingrowth, rather than ion counting by ICP-MS. This approach is intended to minimize potential polyatomic interference from Ba during ICP-MS measurements. The use of ICP-MS equipped with a collision/reaction cell, utilizing gases such as He, O2, or NH3, may help reduce these interferences and improve the analytical performance. However, at present, we were unable to achieve significant Ra/Ba separation using CPE, unlike resin-based ion-exchange chromatography (IEC), although the complete extraction of 226Ra was consistently achieved. Even when 226Ra is not radiochemically pure, its progeny do not significantly influence extraction performance because of their limited affinity for Macropa. The 226Ra decay chain (Section 6 of the ESI) includes 222Rn, which is gaseous and is readily lost upon opening the sample container. Other daughter nuclides, such as Bi3+ and Po4+, exhibit a strong tendency toward hydrolysis under the experimental conditions employed and, therefore, remain in the aqueous phase rather than participating in the competitive or coextraction process. In addition, Pb2+, when present, is not expected to interact strongly with Macropa. The smaller ionic diameter of Pb2+ (2.38 Å) compared to that of Ra2+ (2.96 Å) makes Pb2+ less favorable for coordination within the 18-crown-6-like cavity of Macropa (Table ). Overall, these observations demonstrate that while Macropa does not provide inherent selectivity for Ra2+ over Ba2+ under CPE conditions, quantitative recovery of Ra2+ can be reliably achieved by operating the system with a sufficiently large molar excess of ligand. Competitive effects from Ba2+ and other potential interfering ions, including radionuclide progeny and Pb2+, were shown to be negligible under the experimental conditions employed. The findings highlight that, in the absence of ligand-based selectivity, Ra2+ recovery in CPE is governed primarily by ligand capacity rather than molecular discrimination, and that effective Ra/Ba separation would require structural modification of the ligand or alternative separation strategies beyond single-ligand CPE. Influence of Interfering Metal Ions in CPE Polyatomic species such as 208Pb18O+, 209Bi17O+, 88Sr138Ba+, 78Sr139La+, 68Sr140Ce+, and 88Sr138Ce+ can significantly elevate background signals and potentially interfere with the detection of 226Ra, particularly at ultratrace levels. Therefore, evaluating the ability of the CPE procedure to remove these interfering elements is critical for ensuring the method’s robustness. To address this, we performed additional experiments to evaluate the separation efficiency of the CPE system toward representative interfering elements (Pb, Bi, Sr, Ce, and La). The decontamination factor (DF) was calculated as in eq : 4 DF=MSMSRP=CSVSCSRPVSRP=CSVSCIVI−CSVS where MS and M SRP represent the mass of the analyte in the supernatant and surfactant-rich phase after CPE, respectively. The mass of the analyte was calculated based on the measured concentration (CS ) and volume (VS ) of the supernatant phase. Then, the DF is calculated by subtracting MS from the initial MI (CI initial concentration and VI initial volume). It is important to note that Ba cannot be separated from Ra using the CPE procedure due to their similar chemical properties and coordination behavior in solution, as stated previously. Consequently, most Ra purification strategies rely on BaSO4 coprecipitation, and Ba is inherently present even in NIST standard reference materials. Macropa does not effectively distinguish between Ra and Ba in mixed systems and, therefore, cannot separate them via CPE. To compensate for this limitation, Macropa is added in excess relative to Ra to minimize the influence of Ba. The results demonstrate that under optimized pH and ligand concentrations, the Macropa-based CPE system exhibits strong selectivity toward Ra and Ba over other heavy metal ions (Table ). This selectivity can be attributed to: 1) the higher thermodynamic stability of the Ra– and Ba–Macropa complex compared to other metal–ligand systems; 2) the minimal affinity of Pb and Bi toward the ligand under the working pH conditions; and 3) the preferential partitioning of the Ra complex into the surfactant-rich phase, while interfering elements remain predominantly in the aqueous phase. Consequently, the contribution of polyatomic interferences is significantly reduced by the CPE prior to ICP-MS analysis. 4Decontamination Factor (DF) for Interfering Metal Ions (n = 3)Element C I (mg/mL) V I (mL) C S (mg/mL) V S (mL) M S M SRP DFSr1.007 × 10–7 2509.298 × 10–8 247.52.301 × 10–5 2.162 × 10–6 11La8.648 × 10–8 2508.045 × 10–8 247.51.991 × 10–5 1.708 × 10–6 12Ce8.651 × 10–8 2508.065 × 10–8 247.51.996 × 10–5 1.666 × 10–6 12Bi1.192 × 10–7 2501.146 × 10–7 247.52.836 × 10–5 1.436 × 10–6 20Pb9.889 × 10–8 2501.018 × 10–7 247.52.519 × 10–5 4.730 × 10–7 53 Decontamination Factor (DF) for Interfering Metal Ions (n = 3) Validation of Developed CPE Method The recovery of the developed CPE–ICP-MS method was directly assessed using a tap water sample (TW) with a certified 226Ra concentration of 516 mBq L–1, provided by the Center d’Expertises en Analyze Environnementales du Québec (CEAEQ, Québec, QC, Canada). The measured concentrations showed good agreement with the certified value, confirming the accuracy of the method (Table ). 5Validation of CPE of 226Ra by Certified Reference MaterialsMatrices[226Ra] by CEAEQ (Bq L–1)[226Ra] in current work (Bq L–1)Tap Water0.516 ± 0.010.501 ± 0.01 Validation of CPE of 226Ra by Certified Reference Materials Recommended Conditions for Scale-Up and Treating Real Drinking Water Samples On the basis of the parameters we had studied, we selected the optimal conditions from each experiment and scaled the trials up to 100 mL to evaluate their reproducibility. We observed that scaling up significantly impacted recovery efficiency. No cloud point formation was observed with the 0.1% CPC system, while the 10 mM OA system led to emulsion formation. The only system that maintained consistent performance was the 0.1% BTACl system in 0.01% CTAB, which demonstrated good reproducibility. The results are summarized in Table . We, therefore, chose BTACl as our auxiliary surfactant for processing real water samples. 6Impact on Scale Up in 226Ra RecoveryReagentsConcentration (mM)[226Ra] i (Bq L–1) at 100 mL[226Ra] f (Bq L–1) at 100 mL% Recovery at 10 mL% Recovery at 100 mLCPC2.95 (0.1%)0.45 ± 0.01-99 ± 4%No-CPEBTACl-CTAB5.39 (0.1%)-0.27 (0.01)0.45 ± 0.010.45 ± 0.0299 ± 5%99 ± 5OA10 (0.129%)0.45 ± 0.01-99 ± 5%Emulsion Impact on Scale Up in 226Ra Recovery Back-Extraction and Quantification The micellar phase collected after extraction presents challenges for direct metal ion analysis, as the presence of organic compounds can reduce plasma ionization efficiency in ICP-MS/MS. Additionally, carbon deposits may damage instrument components and compromise subsequent analysis. Therefore, a suitable back-extraction technique is essential to remove surfactant from the analyte matrix using an analyte-compatible reagent. Initially, complexing agents such as citric acid and EDTA were tested but showed no 226Ra recovery. Subsequently, ligand protonation in the metal–ligand complex was explored as a strategy for back extraction. As illustrated in eq , increasing the H+ concentration shifts the complexation equilibrium toward the dissociated metal–ligand, releasing the metal ion from micelles. To test the hypothesis, back extraction was performed at 2, 4, and 6 M HNO3. The dissolved micellar phase was incubated to induce coacervate formation and the separated (repeated three times to ensure complete recovery) aqueous phase combined was evaporated to dryness and analyzed after redissolving in 4% HNO3. As expected, complete recovery (>99%) of 226Ra was achieved at 1 mL of 6 M HNO3 using a control sample containing 0.1 Bq L–1 of 226Ra. The recoveries obtained at 1 mL volumes of the 2 and 4 M solutions were not quantitative. Therefore, we chose 6 M HNO3 for the treatment of drinking water samples. The surfactant-rich phase was treated as mentioned in Section . We measured the recovered 226Ra using ICP-MS/MS; key analytical parameters are discussed below. Extraction and Quantification of 226Ra from Tap Water Using CPE Municipal tap water samples (250 mL) collected from the Laval University campus were immediately degassed for 1 h using a sonic bath at 25 °C. To this solution, 2.5 g of Triton X-114, 50 mg of Macropa, 250 mg of BTACl, and 25 mg of CTAB were added and thoroughly mixed for 5 min. The mixture was then incubated at 45 °C for 30 min to reach the cloud point. After phase formation, the solution was centrifuged at 4,500 rpm using a centrifuge rotor. To increase the viscosity of the micellar phase, the sample was placed on ice for 5 min before decantation. The CPE procedure was repeated three times to ensure complete recovery. The collected combined micellar phase was dissolved in a minimum amount (1–2 mL) of 6 M HNO3, and mixed for 5 min, then CPE was performed (three times) and collected combined aqueous phase was evaporated to dryness and redissolved in 4% HNO3. The resulting aqueous phase was analyzed for 226Ra by ICP-MS/MS using 205Tl as an internal standard. The main objective of the study was to estimate the Ra in drinking water, and hence identical experiments were also performed with carbonated mineral drinking bottled water samples. The specifications of the water samples and results of the investigation are presented in Table . 7Specifications of the Mineral Water Samples with Quantified 226Ra Concentration Concentration of mineral in mg/L(ppm) Sample[Ca2+][Na+][Mg2+][K+][Cl–][HCO3 –][NO3 –][SO4 2‑][F–][TDS] 226Ra (mBq L–1)MTW-Laval University151520.822N.A.0.320<0.1N.A.1.000 ± 0.002MBW- GerolsteinerCanada3481181081140N.A.N.A.38N.A.247913.000 ± 0.007aTDS = Total Dissolved Solids. Specifications of the Mineral Water Samples with Quantified 226Ra Concentration TDS = Total Dissolved Solids. Analytical Figure of Merit The calibration curve for the analysis is provided on page S15 of Supporting Information. The slope (S) obtained from the curve was 2.39 × 10–5 CPS/Bq/L, and the standard deviation σ ratio, blank of 1.9565 × 10–8 CPS was calculated from the three replicates of the blank. The Limit of Detection (LOD) and Limit of Quantification (LOQ) for the proposed method were estimated using eqs and , respectively. The calculated LOD of the method for 250 mL of water treated by CPE was 2 mBq L–1 (55 fg L–1) and the LOQ was 8 mBq L–1 (219 fg L–1)­ 5 LOD=3.3×σratio,blanks 6 LOQ=10×σratio, blanks The repeatability (r) and coefficient of variation (CV) for the method were calculated according to eqs and , respectively. Here, SD represents the standard deviation of the measurements. The turnaround time of the method was 3 h, and the sample preparation time was 1 h. 7 Repeatability(r)=t(0.975;n−1)×SD√n 8 CV(%)=r×100x̅ The preconcentration factor (PF) of the method was calculated based on eq , where V i is the volume of the initial water sample treated, and V SRP is the volume of the surfactant-rich phase with concentrated 226Ra. 9 PF=ViVSRP The current figure of merit of the protocol was compared with five reported methods for 226Ra separation and quantification from table and bottled mineral water samples, as shown in Table . It demonstrated that the proposed method has the benefits of a shorter turnaround time, marginally higher preconcentration factors, minimal chemical waste generation, and higher recovery compared to previous methods. Therefore, the CPE method can be deployed and adopted for separation and quantification of 226Ra from various drinking mineral water samples, as shown in Table . 8Comparison of Reported Five Protocols for 226Ra Determination in Drinking Water SamplesProtocol ReferenceDalencourt et al. Lagace et al. Verlinde et al. Yang et al. CEAEQ Current WorkMethodsIEC IECIECIECCoprecipitation & IECCPE Resins/LigandsAG50W-X8AG50W-X8 and Sr-specAnaLig Ra-01AG50W-X8AG50W-X8MacropaProcess TypeContinuousContinuousContinuousContinuousContinuousBatchRecovery84–8596–10587–9689–9594–9597–99Repeatability7–97–811–151–31–163–5Preconcentration factor50333316733383MDL;MQL (fg L–1)7;244;156;202;72;655;219MDL;MQL (mBq L–1)0.3;0.90.2;0.50.2;0.70.1;0.30.1;0.22;8Turnaround Time (h)61121118.53Sample volume (mL)5001001005001000240Cost (USD)9876407013250Waste (mL)908023133431243aIEC = ion exchange chromatography.bCPE = cloud point extraction. Comparison of Reported Five Protocols for 226Ra Determination in Drinking Water Samples IEC = ion exchange chromatography. CPE = cloud point extraction. The comparison between solid-phase extraction (SPE) and cloud point extraction (CPE), particularly with respect to batch processing and automation, remains a widely discussed topic among extraction chemists and analytical scientists. Currently, CPE is less convenient for high-throughput automation because it requires controlled heating and a phase-separation step. In contrast, SPE offers clear advantages in automated and cartridge-based workflows. However, CPE provides several distinct benefits, including lower organic solvent consumption, high preconcentration factors, strong tolerance to high-salinity matrices, lower consumable costs, rapid turnaround time, and selective complexation of desirable metal ions. Although SPE can be readily adapted for automation or online sampling, high-salinity matrices often adversely affect its selectivity and require additional scrubbing or washing steps, thereby increasing the processing time. Achieving selectivity in SPE typically involves multiple steps (loading, scrubbing, washing, and elution), making the process time- and labor-intensive and generating large volumes of waste, which is undesirable for radiochemical applications. In contrast, CPE relies on chelating agents specifically designed or synthesized to selectively bind target metal ions, enabling efficient separation from matrix ions through micellar extraction. High-salinity matrices are more easily handled by CPE, as increased ionic strength reduces the hydrophilicity of the metal–ligand complex, thereby promoting its transfer into the surfactant-rich phase. Consequently, CPE is particularly suitable for the analysis of saline samples such as seawater, river water, lake water, and mining effluents when the current ligand structure is fine-tuned by an alkyl substituent with various chain lengths. The main limitations of CPE include its temperature dependence, the requirement for centrifugation during phase separation, challenges in scalability, reproducibility concerns, and potential surfactant interference with metal analysis. However, CPE is not inherently incompatible with automation. Recent developments incorporating controlled heating and phase separation modules have enabled automated and flow-based implementations. Furthermore, conventional CPE has evolved into several advanced formats, including micro-CPE (M-CPE), displacement CPE (D-CPE), rapid synergistic CPE (RS-CPE), dual CPE (d-CPE), and acid-induced CPE (A-CPE). These approaches can be integrated into temperature-dependent aqueous micellar two-phase systems (AMTPSs), which allow automated phase separation and direct sample introduction into online ICP-MS systems. Accordingly, CPE should not be regarded as inherently incompatible with online or automated analytical workflows. While the present study focuses exclusively on method development and analytical validation, the extension of the approach to automated platforms constitutes a logical direction for future work. Conclusions We developed a precise, sensitive, and accurate method for the quantification of 226Ra in water samples using Macropa-based cloud point extraction. The method was applied to municipal tap water and bottled mineral water samples. The limit of detection and limit of quantification were found to be 2 mBq L–1 (55 fg L–1) and 8 mBq L–1 (219 fg L–1), respectively. The high sensitivity and low error associated with CPE-ICP-MS/MS demonstrate its versatility for various aqueous samples with low 226Ra activity. It compares favorably with other ICP-MS/MS-based methodologies for radium analysis by the key figures of merit. Since the method does not require loading samples onto a resin, as in most published protocols, and does not rely on nonselective coprecipitation, the associated loading and waiting steps for complete precipitation are eliminated. It is also considered a green method since no conventional or environmentally toxic organic solvents are used. The simple and rapid procedure enables the effective determination of 226Ra from water samples. This opens the door to measuring long-term naturally occurring 226Ra in environmental, domestic, and industrial effluent water samples. Supplementary Material