Introduction

Radiostrontium is a major concern in environmental contamination and nuclear processes and is considered one of the more important radionuclides in the view of the radiation protection, radioecology, radioactive waste management and environmental monitoring [1]. This is due to (1) 90Sr having a significant accumulative thermal fission product yield of 5.88 and 2.1% for 235U and 239Pu, respectively, (2) 90Sr has a long effective half-life (~18 years) in the human body [2], and (3) 90Sr can have a high or a moderate mobility in soils and sediments where its environmental behavior is governed mainly by its chemical similarity to calcium [3].

The concentration of radiostrontium in the environment is generally low, so a good radiological monitoring system requires highly selective, sensitive and stable techniques especially for pure β-emitters such as 90Sr that has high probability of interference from other radionuclides. Strontium-90 may be detected by several detection techniques include gas-flow proportional, liquid scintillation counting (LSC), Cerenkov counting, and ICP-MS but almost all of them require multi-step radiochemical procedures to separate strontium from other beta emitting nuclides. Significant spectral interferences can be problematic for the nuclear counting techniques. A variety of methods, including solvent extraction, ion exchange, precipitation, etc., have been developed for the separation of strontium [4]. Solvent extraction has been very successfully used in 90Sr separation. Also, solid-phase extraction (SPE) RAD disks [5] and extraction chromatographic materials that combine the power and selectivity of solvent extraction and the simplicity of chromatography were applied effectively in isolating 90Sr from complex matrices [6]. In all cases, there is a selective molecule or a functional group must be used to extract Sr2+ from other ions (Ca2+, Pb2+) and especially 133Ba, 140Ba, 226Ra, 228Ra radionuclides that have a similar chemical behavior. Among the studied selective functional groups, macrocyclic polyethers (crown ethers) that first synthesized by Pedersen in 1967 were found to form strong complexes with alkali or alkaline earth metals. Two of the most effective for strontium extraction are dicyclohexano-18-crown-6 (DCH18C6) and ditert-butylcyclohexano-18-crown-6 (DtBuCH18C6) [7]. Due to their high extraction ability and good selectivity, these two crown ethers have been considered as promising extractants for strontium. Although, the methods of using selective crown ethers for Sr separation proved to be simpler and faster than the traditional precipitation methods, they have limitations for use in conjunction with in situ monitoring systems. This is because they form strong complexes with Sr2+ are typically conducted at high nitric acid concentration (3–8 M) that requires special handling for the aqueous sample before concentration. Also, damage to analytical equipment and the bleeding of the extractant/diluent layer can occur, especially in long-term unattended sampling locations.

Traditional methods of extracting groundwater samples and performing laboratory analyses are expensive, time-consuming and induce significant disposal challenges. The ideal in situ monitoring system should have very limited handling and should measure the radioactivity in aqueous samples under environmental conditions without using harsh chemicals. To achieve this goal, Runkle et al. presented a prototype compact counter capable of measuring 90Sr groundwater concentrations in situ at or below the drinking water limit of 0.33 Bq/L (8 pCi/L) [8]. The measurement is based on detection of Cerenkov light created by the high-energy electrons produced by 90Y, the daughter of 90Sr. Although, the counter design is unique, the authors assumed that the 90Sr and 90Y were in a secular equilibrium in groundwater that is difficult to be proved or be controlled due to the clear difference in the chemistry of both isotopes. This weakness was solved through extracting 90Sr using a SuperLig® 620 column (strontium selective solid-phase extraction material) from groundwater to create a pure 90Sr source from that subsequent 90Y ingrowth can be measured [9]. This method required 64.25 h at which 90Y ingrowth is 50.1% of 90Sr activity on the column. The 90Y was fluidically transferred from the column to a 5 mL volume Cherenkov detection flow cell. Following concentrating 90Sr out of a 350 mL groundwater sample; they reported a detection limit of 0.057 Bq/L. In addition, SuperLig® 620 was shown be highly selective for Sr2+ under a wide range of chemical conditions, where Sr2+ ions can be extracted between 0.1 and 8 M nitric acid [10]. Egorov et al. mixed BC-400 plastic scintillator beads with solid-phase extraction material SuperLig® 620 for the development of 90Sr sensors using a 1:1 weight ratio. The 90Sr sensor exhibited excellent analyte retention characteristics and good detection efficiency of 63 ± 3% [11].

The main objective is to benefit from the unique extraction behavior of SuperLig® 620 material for radiostrontium in developing a simple and easy to handle in situ monitoring system through combining solid-phase extraction and scintillation detection. SuperLig® 620 will be modified using two different approaches, (1) mixing the SuperLig®620 extractant with inorganic scintillating beads, (2) incorporation of the SuperLig®620 extractant within plastic scintillating beads. The developed extractive scintillating technique gives a near real-time indication of activity levels.

Experimental

Materials and chemicals

All chemicals were used as received except monomers, which were disinhibited by passing through a column of basic alumina before use. Ethanol and methanol were purchased from BDH (UK). 4-methylstyrene monomer, divinylbenzene (DVB), azobisisobutyronitrile (AIBN), toluene, poly(vinyl alcohol) (PVA, average MW 65,000–124,000 Da, degree of hydrolysis 87–89%) and 4-chloromethylstyrene (CMS) were purchased from Sigma-Aldrich. Reagent grade NaCl, diethyl ether, dimethylformamide (DMF), and chloroform were purchased from Fisher Scientific. Hydroxypropyl methylcellulose (HPMC) came from Dow Chemical Co. (USA). The inorganic scintillator GS20 was purchased from Saint-Gobain Crystals (USA), calcium fluoride, CaF2:Eu and yttrium silicate, Y2SiO5:Ce were from Rexon Components Inc. (USA). The solid phase extractant SuperLig® 620 was purchased from IBC Technologies (USA).

Preparation of extractive scintillating sensors

The main materials, preparation techniques and corresponding codes applied to make extractive scintillating materials using the two different approaches are summarized in Table 1.

Table 1 The two different approaches proposed for preparing extractive scintillating sensors for quantification of radiostrontium
Full size table

Combination of extractant/scintillator beads

The heterogeneous flow-cell was prepared through mixing SuperLig® 620 in varying ratios with an inorganic scintillator such as europium-activated calcium fluoride, CaF2:Eu (SL-CaF), cerium-activated yttrium silicate, Y2SiO5:Ce (SL-YSO) and cerium-activated lithium silicate glass, GS20 (SL-GS20). The heterogeneous beads approach aimed to evaluate the performance of the SuperLig® 620 in its original form comparing to the incorporated one. Because 90Sr has medium beta energy (546 keV), the heterogeneous sensor is expected to overcome the energy transfer limitation from the beads containing the extractant to the scintillating beads resulting in a good detection efficiency. This setup also has high stability due to the inorganic nature of the extractant and the scintillating beads. The SuperLig® 620 was mixed well with each of three inorganic scintillators and dry packed into a translucent FEP Teflon tube for final application.

SuperLig®620/polymer composite by suspension polymerization

The second extractive scintillating sensor material was prepared by incorporating SuperLig® 620 in plastic scintillating beads using suspension polymerization (SL-PVT). The SuperLig® 620 solid phase particles were mixed as received with 4-methylstyrene monomer, divinylbenzene crosslinker, benzoyl peroxide initiator, 2-(1-naphthyl)-4-vinyl-5-phenyloxazole (vNPO) (the vinyl form of αNPO scintillator) organic fluor monomer [12] and a toluene porogen to form a dispersed oil phase. The components of the oil phase were mixed at room temperature for 15–60 min before adding to the aqueous phase. An emulsion was prepared by dispersing the oil phase into the aqueous phase in a four-neck glass reactor. The aqueous phase contained poly(vinyl alcohol) (PVA) as a polymer surfactant and emulsion stabilizer, NaCl to adjust the electrolyte concentration and hydroxypropyl methylcellulose (HPMC) as a seeding (nucleation) agent for dispersing of the organic phase in water. The mixture of the aqueous and the organic phases were dispersed by stirring for 30 min at room temperature and 500 revolutions per minute (rpm) to form the beads size required before polymerization. Polymerization reaction was performed for 12 h at 70 °C temperature and 500 rpm in a thermostatic water bath. The resin beads settled in the reactor (bottom) were filtered, washed three times with methanol and air dried in a fumehood. Also, the suspension polymerization aqueous media was dried and residual fine powder (top) was washed and dried.

Characterization and application of extractive scintillating materials

The physical and chemical behavior of materials resulting from the above four formulations of the two approaches were subjected to extensive characterization. The extractive scintillating resin from the second approach was evaluated using thermogravimetric analysis (TGA). The extraction efficiency was tested using batch experiment and 85Sr in 0.1 M HNO3 solution. The filtered solution was measured with NaI:Tl detector and compared with the initial concentration. The capacity of the modified scintillating SuperLig® 620 particles were tested using batch experiment at different stable strontium concentrations (1–1000 ppm) after spiking with an 85Sr tracer in 0.1 M HNO3, the maximum resin capacity (q m) was modeled using Langmuir equation.

The luminosity and absolute detection efficiency was measured for each extractive scintillating bead formulation. For the luminosity measurement, about 25 mg of material was placed in a 7 mL LSC vial while a 37,000 Bq 241Am point source was positioned at 0.5 cm above the bead surface. The light output from α particles deposition into the scintillating beads was measured using a Hidex Triathler liquid scintillation counter (LSC). The absolute detection efficiency for the developed sensor materials was measured after packing about 50 mg of the functionalized resin beads to a bed length of 4.0 ± 0.1 cm in a FEP Teflon tubing (1/8″ external, 1/16″ internal diameters). After conditioning the column with 0.1 M HNO3, 90Sr in 0.1 M HNO3 was passed through where the selective group catches it. The minicolumn was centered in the middle of LSC vial and counted offline with a Quantulus LSC (Perkin-Elmer). In all column experiments, known volume of influent and effluent solutions was combined with 10 mL of Ultima Gold AB scintillation cocktail to be counted offline with LSC. This to determine the initial and the remaining radioactivity to calculate bound efficiency.

The final application of extractive scintillating materials were evaluated using online detection mode. The beads were dry packed into a FEP Teflon column with a small amount of glass wool packed at each end to prevent resin washout from the tubing. The FEP Teflon tube with an inner diameter of 0.16 cm (1/16″) was filled with ~50 mg of resin beads (bed length of 4.0 ± 0.1 cm). After packing the columns, they were placed as a U-shape flow-cell in the flow scintillation analyzer (FSA) IN/US β-Ram model 5 (LabLogic Systems, Inc). The instrument digital output was connected to a laptop using the Laura software package (LabLogic Systems, Inc., USA) and the data were acquired as a multichannel scaling spectrum with a 30-s dwell time. The calculated uncertainties for all measurements dependent mainly on statistical counting as major elements of errors and one sigma as a confidence interval. In several places where more than one measurement is conducted the mean and the associated one sigma standard deviation is reported.

The minimum detectable activity concentration in water (MDC) for the radiometric preconcentrating flow-cell can be calculated according to Eq. 1 [13].

$${\text{MDC}}_{\text{Bq}} = \frac{{2.71 + 4.65\sqrt {C_{\text{b}} t} }}{{tE_{\text{s}} E_{\text{d}} V}}$$
(1)

t is counting time, C b is the background count rate, V is sample volume, E d is the absolute detection efficiency (ratio of the counting rate detected to the decay rate in the loaded sample) and E s is bound efficiency (the ratio of activity loaded on the sensor to the injected sample activity).

Results and discussion

Detection of 90Sr with heterogeneous flow-cell

Sensors characterization and application

The three inorganic scintillators that were investigated have different physical properties that can affect their final applications. The density is 2.4, 4.45, and 3.18 g/cm3 for GS20, Y2SiO5:Ce and CaF2:Eu, respectively. The emission wavelength (λ em) from three inorganic scintillators ranges from 395 nm for GS20 to 435 nm for CaF2:Eu. Both Y2SiO5:Ce and CaF2:Eu have light yield of approximately 24,000 photons/MeVγ, while GS20 has significant lower light yield of 3800 photons/MeVγ [14]. The three inorganic scintillators were chosen for different reasons. GS20 was chosen because it has similar density for SuperLig® 620 that can result in more uniform mixing. Y2SiO5:Ce and CaF2:Eu were used based on their high luminosity comparing to the organic scintillator BC-400 applied before in a similar application [11]. The radioluminosity of the three inorganic materials and BC-400 were measured by weighing about 25 mg of the scintillating crystals in 7 mL LSC vial followed by irradiation from a 37,000 Bq 241Am (1 μCi) point source positioned approximately 0.5 cm above the bead surface. The light output from α particles irradiation was measured and presented in Fig. 1. Y2SiO5:Ce and CaF2:Eu gave a peak height around channel 815 and 850, while the peak height for the GS20 and BC-400 is channel 540 and 560, respectively. This result shows that replacing the GS20 with Y2SiO5:Ce or CaF2:Eu in the heterogeneous flow-cell is expected to improve the detection efficiency under the same conditions.

Fig. 1
figure 1

Pulse height spectra were collected for the inorganic scintillators GS20, CaF2:Eu and Y2SiO5:Ce used to setup heterogeneous flow-cells; BC-400 was added for comparison. (The channel numbers are on the logarithmic scale)

Full size image

Different sensors were prepared by mixing SuperLig®620 with each of the three inorganic scintillator GS20 (SL-GS20), Y2SiO5:Ce (SL-YSO) and CaF2:Eu (SL-CaF). The effect of the type of inorganic scintillator on the detection efficiency was investigated using two weight ratios of 1:1 and 1:2 SuperLig®620 to inorganic scintillator. The high accessible capacity of the SuperLig®620 (0.225 mmol/g) allows for a higher scintillator ratio that can help to detect most of the beta particles. Due to the similarity of Y2SiO5:Ce and CaF2:Eu in their pulse heights and scintillation efficiency, only Y2SiO5:Ce was compared to GS20 beads using 1:1 weight ratio. Following that, the Y2SiO5:Ce and CaF2:Eu were compared for the 1:2 weight ratio. To study the effect of scintillator type and extractant/scintillator ratio on the detection efficiency of 90Sr measurement, about 250 mg of the SuperLig®620 and 250 mg of GS20 or Y2SiO5:Ce were mixed to give 1:1 ratio, then dry packed into a U-shape translucent FEP Teflon tube.

Based on previous research it was determined that Pb, K, Tl and Sr in 2 M HNO3 are retained by the SuperLig®620 and are eluted with 0.49 M ammonium citrate while other common ions, Ca, Bi, V, were demonstrated not to be retained on the column [15]. The most critical interference comes from 90Y (The direct daughter of 90Sr) which has high energy β−. Acidifying the solution to 0.1 M HNO3 is essential to avoid the co-extraction of 90Sr and 90Y. Before loading sample solutions, the flow-cell was stabilized in the β-ram flow scintillation analyzer and left in dark for 15 min. About 10–15 mL 0.1 M HNO3 was pumped through the flow cell at 0.5 mL/min flow rate followed by loading ~108 Bq of 90Sr in 0.1 M HNO3. After loading the activity, about 25 mL of 0.1 M HNO3 solution was pumped through the sensor. The activity accumulated in the flow-cell is measured continuously by FSA and a 5 mL of each effluent was counted for 30 min by LSC to measure E s. Knowing the radioactivity loaded onto the flow-cell and the net count rate, E d was calculated for each extractive scintillating sensor.

E s was >99.95% and E d for the 1:1 extractant/scintillator ratio was 30.6 ± 1 and 34.1 ± 1% for the SL-GS20 and SL-YSO sensor, respectively. The loading profile and corresponding loading and detection data are shown in Fig. 2. The results reveal that the Y2SiO5:Ce has a higher detection efficiency than the GS20 inorganic scintillator; therefore, the same conditions were used to compare the performance of CaF2:Eu relative to Y2SiO5:Ce but at 1:2 extractant/scintillator ratio, Fig. 2. The online profiles show that increasing the extractant/scintillator ratio of SL-YSO sensor from 1:1 to 1:2 increased the detection efficiency from 34.1 ± 1.0 to 46.2 ± 1.2%. This improvement is due to increasing the probability of the beta particles depositing energy in the Y2SiO5:Ce particles that surround the SuperLig®620 extractant. The 1:2 ratio was selected for further study to balance the sorption capacity and detection efficiency of the sensor.

Fig. 2
figure 2

Effect of inorganic scintillator type and SuperLig®620/scintillator weight ratio on 90Sr detection efficiency

Full size image

Comparing the performance of CaF2:Eu scintillator to Y2SiO5:Ce scintillator, for the 1:2 ratio, shows that the SL-CaF sensor has the highest detection efficiency among the three inorganic scintillators with a mean value of 58.7 ± 1.4%. The developed extractive scintillating sensor SL-CaF proved to combine the two important properties required to have a high performance radiation detection system for 90Sr quantification.

Sensor stability and limit of detection

To evaluate the regeneration capability of the heterogeneous extractive scintillating sensors, multiple loading and elution trials were conducted using SL-GS20 sensor. In these trials no degradation in the loading efficiency or the detection efficiency was observed. The uptake efficiency was almost >99.95%, while the average detection efficiency for the five trials was 32.8 ± 0.84%. The regeneration process was simple and can completely elute 90Sr using about 5 mL of 0.5 M sodium citrate at pH 6 as eluent (~50 column volume). This good stability and performance make this design promising to develop a similar heterogeneous flow-cell using a high luminosity inorganic scintillator to improve the detection efficiency.

The effect of flow rate used to load the aqueous sample was investigated using the SL-CaF sensor. The peristaltic pump of the β-Ram FSA was utilized at three different flow rates: 0.44, 0.95 and 1.67 mL/min. The effluent after each load was collected and counted by LSC to measure E s. The online analysis and corresponding profiles are presented in Fig. 3. The results show that the SL-CaF sensor was not affected by increasing the flow rate from 0.44 to 1.67 mL/min and gave almost the same count rate with a mean value of 49.50 ± 1.6 cps. This means that the analysis time can be reduced by 3.8 times without noticeable effect on the sensor performance.

Fig. 3
figure 3

The effect of flow rate on the uptake and the detection efficiency of 90Sr on SL-CaF sensor

Full size image

To measure the limit of detection of the developed radiation measurement system, the online quantification of 90Sr in 0.1 M HNO3 using the β-Ram FSA and SL-CaF sensor was tested and illustrated in Fig. 4. The MDC of 90Sr (Bq/L) in water for the radiometric preconcentrating flow-cell was calculated and evaluated. About 10 mL 0.1 M HNO3 at a flow rate of ~0.5 mL/min was pumped first to equilibrate the flow cell, followed by 90Sr (4 Bq/L) in 0.1 M HNO3. After pumping about 400 mL, that contained only 1.6 Bq of 90Sr, the SL-CaF sensor gave good signal response relative the background count rate. To achieve the U.S. MCL of 0.33 Bq/L [16], for a detection efficiency of 58.7%, a background count rate of 0.653 cps, a sample volume of 350 mL, extraction yield of >99.95% and a count time of 1 h is required, (Eq. 1).

Fig. 4
figure 4

Analysis of 4 Bq/L 90Sr using heterogeneous SL-CaF sensor

Full size image

Detection of 90Sr using SL-PVT prepared by suspension polymerization

Sensors characterization and application

The SuperLig® 620 was incorporated within scintillating polymer beads prepared using mainly polyvinyl toluene (PVT) with 2-(1-naphthyl)-4-vinyl-5-phenyloxazole (vNPO) fluor. Two different samples were prepared by using the same recipe but changing the type of porogen. In the first sample, a mixture of methyl ethyl ketone (MEK) and toluene (MEK/toluene) were used as a porogen (SL-PVTM) while the second sample utilized Span80 porogen (SL-PVTS). The optical images in the same magnification of the two samples were presented in Fig. 5. The images of the synthesized samples showed that the SL-PVTM prepared using the MEK/toluene porogen has spherical beads structure. The beads contain small amount of the SuperLig® 620 extractant attached to the beads surface.

Fig. 5
figure 5

The optical images of the SL-PVT extractive scintillating materials prepared using suspension polymerization in MEK/toluene porogen, SL-PVTM (a), and in Span80 porogen, SL-PVTS (b)

Full size image

To determine the amount of SuperLig® 620 incorporated in each fraction, TGA was conducted for four samples: silica, SuperLig® 620, SL-PVTM and SL-PVTS. The TGA data reveal that the four measured samples showed a first mass loss of about 2.7% around 100 °C attributed to water. Between 200 and 800 °C the mass loss was 14.0% from the SuperLig® 620 extractant, mainly attributed to the organic selective material attached to silica surface. The TGA of SL-PVTM sample shows that the polymer decomposition starts around 250 and 300 °C and gave a final polymer mass loss of 70.2%, leaving a pure SuperLig® 620 content around 15.3% (0.153 g/g polymer composite). So, the final total calculated capacity is 0.034 mmol/g. This capacity is not fully accessible, as a significant fraction of the SuperLig® 620 was encapsulated in the plastic scintillating material (Fig. 5). The result for SL-PVTS shows that the final SuperLig® 620 residue in composite is 0.222 g/g. The final mean extractant capacity was calculated to be 0.048 mmol/g. Although, SuperLig® 620 incorporation in polymer during suspension polymerization is better with the Span80 porogen, the MEK/toluene porogen gives good spherical transparent particles that are better for final application.

The capacity of SL-PVTM composite and the SuperLig® 620 extractant was measured at constant temperature and different concentrations of stable Sr2+ ions spiked with 85Sr. The maximum accessible capacity (q m) estimated from Langmuir isotherm is 0.28 mmol/g for the SuperLig® 620, that is close to the maximum capacity reported by the producer (0.225 mmol/g). On the other hand, the maximum accessible capacity for SL-PVTM composite is only 0.005 mmol/g that is approximately 45 times lower than the capacity of the raw material. Also, the material showed about 7 times lower capacity than the calculated value of 0.034 mmol/g from the TGA analysis where, small amount of SuperLig® 620 were incorporated in the plastic beads. This means that only 14.3% of the total sites are accessible to strontium ions.

The online detection response as a function of successive loading of different 90Sr activity without regeneration was investigated using the FSA. This evaluation refers to the possibility of using the same SuperLig®620/polymer sensor to quantify the 90Sr radioactivity level in aqueous samples that have a clear variation in strontium concentrations that can represent a contamination plume. The successive loading profile is presented in Fig. 6a and the corresponding calibration curve for the total activity levels of the loaded 90Sr was plotted in Fig. 6b. The correlation coefficient (R 2) of the calibration curve is 0.9998 with E d = 51.2% as determined by the slope of the fit line. Table 2 shows that the detection efficiency decreases slightly with each subsequent loading with the average of the four loaded activities of 53.6 ± 2.4%.

Fig. 6
figure 6

Online successive analysis of 90Sr using SL-PVTM extractive scintillating resin: a the accumulated count rate for water samples with different radioactivity levels; b the measured net count rate versus the total bound activity

Full size image
Table 2 The loaded activity, uptake percentage and detection efficiency of 90Sr in aqueous media using successive loading and online detection mode for the SL-PVTM extractive scintillating resin
Full size table

In addition to the decrease in E d, Table 2 also summaries the decrease in E s. About 7.8% decrease in the uptake percent was reported for the second loading step relative to the first one. Also, about 3.0% decrease in the uptake efficiency was measured in the fourth step compared to the third step. This decrease opened a concern about the sensor stability and behavior for online detection and application that need more deep evaluation.

Evaluation of SL-PVTM sensor stability

To investigate the cause of the decrease in the uptake efficiency during the use of successive loading-elution cycles, a new flow-cell was packed with the SL-PVTM resin and loaded with ~108 Bq of 90Sr according to the black profile presented in Fig. 7. A 0.1 M HNO3 solution was pumped through the flow cell for 15 min at 0.5 mL/min to collect the background count rate. After pumping ~108 Bq of 90Sr, the pump was turned off while counting for about 20 min. The average net count rate was 53.1 ± 1.3 cps. Following this step, about 42 mL of 0.1 M HNO3 was pumped through the sensor at the same flow rate used to load the 90Sr sample. As shown in Fig. 7, the activity of the flow-cell started to decrease slowly. To determine the cause of this behavior, the effluent was collected and counted by LSC giving a total activity of 43.7 Bq in the 42 mL effluent, that equals about ∼55.0% of the bound 90Sr activity. This unexpected behavior may be due to loss of affinity of SuperLig®620 particles for 90Sr after incorporating in the plastic scintillator or to the low sorption capacity of 0.005 mmol/g that was measured for SL-PVTM composite.

Fig. 7
figure 7

Online loading profiles of 90Sr using SL-PVTM and SL-GS20 extractive scintillating sensors

Full size image

The SL-PVTM sensor was compared with the heterogeneous SL-GS20 sensor by testing the latter under the same experimental conditions, Fig. 7. Comparing the results of both systems revealed that (1) the uptake efficiency is >99.95 and 87.6% for SL-GS20 and SL-PVTM composite, respectively; (2) the detection efficiency is 30.6 and 51.0% for SL-GS20 and SL-PVTM, respectively; (3) the SL-GS20 retains almost all the loaded activity after pumping >400 column volume, while the SL-PVTM sensor loses approximately 55% of its loaded activity under the same conditions.

The large difference between SL-PVTM and SL-GS20 performance in the column flow tests can be explained by the difference in 90Sr affinity for these two sensors in 0.1 M HNO3. Twenty-five mg each of SL-PVTM and Superlig® 620 was added to 5 mL of 0.1 M HNO3 solution, spiked with 41.5 Bq of 90Sr, then tumbled end-over-end for 5 h. The suspensions were filtered through 0.45 μm nylon syringe filters and the 90Sr concentration in the filtrate was determined using LSC. The K d value was calculated using Eq. 2,

$$K_{\text{d}} = \frac{{\left( {C_{\text{o}} - C_{\text{e}} } \right)}}{{C_{\text{e}} }} \times \frac{V}{m}\left( {\frac{\text{mL}}{\text{g}}} \right)$$
(2)

where C o and C e are the initial and equilibrium aqueous activity of 90Sr (cpm), V is the solution volume (mL), and m is the weight of the extractive scintillating material (g).

The calculated K d value for the SuperLig®620 and SL-PVTM materials are 3.78 × 105 and 1.03 × 103 mL/g. The smaller K d value reported for the SL-PVTM composite will cause greater Sr desorption during flow experiments where the influent solution does not contain Sr. This sensitivity to water chemistry makes SL-PVTM unstable in case of using the online monitoring application. Because the main objective of the current research is developing stable extractive scintillating sensors to be used for online monitoring systems and measuring 90Sr in situ, a new approach should be applied to maintain the high capacity of SuperLig®620, while preserving scintillation functionality. This can be achieved through growing scintillating polymer chains on the surface of SuperLig®620 silica particles and this is a subject of ongoing research.

Conclusion

This paper describes extractive scintillating sensors for the detection of current and changing conditions of radiostrontium contamination in the subsurface groundwater. Sensors capable of in-situ quantification of 90Sr are not commercially available. These sensors also facilitate real-time measurement, and decrease the risk to health and the cost of long term monitoring. A simple and rapid procedure was developed and characterized using extractive scintillating sensors for online environmental radiation monitoring of 90Sr in aqueous samples under field conditions. Two different approaches were applied to prepare highly selective sensors for strontium extraction. In the first approach, the SuperLig® 620 was mixed with inorganic scintillating beads. Three different granulated scintillators were investigated:Y2SiO5:Ce, GS20 and CaF2:Eu. Although, GS20 formed a good homogeneous mixture with SuperLig®620, it showed low detection efficiency of 30.6%. The best extractive scintillating performance was reported for the SuperLig®620/CaF2:Eu with average detection efficiency of 58.7 ± 1.4%. The second approach involved the incorporation of SuperLig®620 extractant into porous scintillating polyvinyltoluene (PVT) beads formulated with the organic fluor monomer 2-(1-naphthyl)-4-vinyl-5-phenyloxazole (vNPO). The new extractive scintillating composite showed good selectivity and detection efficiency (51–55%) for 90Sr after loading radioactivity. However, 90Sr desorption occurred when radionuclide free solutions were passed through the column. This behavior is proposed to be due to the relatively low K d of SL-PVTM compared with the unmodified SuperLig®620 extractant. Both approaches have sufficient sensitivity for 90Sr in drinking water down to US maximum contaminant level of 0.33 Bq/L for 90Sr in drinking water.