Optimization of collision/reaction gases for ...

Spectrochimica Acta Part B 135 (2017) 82–90

Contents lists available at ScienceDirect

Spectrochimica Acta Part B journal homepage: www.elsevier.com/locate/sab

Optimization of collision/reaction gases for determination of 90Sr in atmospheric particulate matter by inductively coupled plasma tandem mass spectrometry after direct introduction of air via a gas-exchange device Yoshinari Suzuki a,b,⁎, Ryota Ohara a, Kirara Matsunaga a a b

Faculty of Life and Environmental Science, Shimane University, 1060 Nishikawatsu-cho, Matsue-shi, Shimane 690-8504, Japan Estuary Research Center, Shimane University, 1060 Nishikawatsu-cho, Matsue-shi, Shimane 690-8504, Japan

a r t i c l e

i n f o

Article history: Received 19 April 2017 Received in revised form 18 July 2017 Accepted 24 July 2017 Available online 27 July 2017 Keywords: 90 Sr ICP-MS/MS Atmospheric particulate matter Real-time analysis Collision/reaction cell

a b s t r a c t Nuclear power plant accidents release radioactive strontium 90 (90Sr) into the environment. Monitoring of 90Sr, although important, is difficult and time consuming because it emits only beta radiation. We have developed a new analytical system that enables real-time analysis of 90Sr in atmospheric particulate matter with an analytical run time of only 10 min. Briefly, after passage of an air sample through an impactor, a small fraction of the sample is introduced into a gas-exchange device, where the air is replaced by Ar. Then the sample is directly introduced into an inductively coupled plasma tandem mass spectrometry (ICP-MS/MS) system equipped with a collision/ reaction cell to eliminate isobaric interferences on 90Sr from 90Zr+, 89Y1H+, and 90Y+. Experiments with various reaction gas conditions revealed that these interferences could be minimized under the following optimized conditions: 1.0 mL min−1 O2, 10.0 mL min−1 H2, and 1.0 mL min−1 NH3. The estimated background equivalent concentration and estimated detection limit of the system were 9.7 × 10−4 and 3.6 × 10−4 ng m−3, respectively, which are equivalent to 4.9 × 10−6 and 1.8 × 10−6 Bq cm−3. Recoveries of Sr in PM2.5 measured by real-time analysis compared to those obtained by simultaneously collection on filter was 53 ± 23%, and using this recovery, the detection limit as PM2.5 was estimated to be 3.4 ± 1.5 × 10−6 Bq cm−3. That is, this system enabled detection of 90Sr at concentrations b 5 × 10−6 Bq cm−3 even considering the insufficient fusion/vaporization/ionization efficiency of Sr in PM2.5. © 2017 Elsevier B.V. All rights reserved.

1. Introduction

are also problems. Radionuclide-specific analysis and quantification of Sr require alternative approaches based on radiochemical separation. The currently available detectors are liquid scintillation counters and gas ionization detectors, but sample pretreatment and measurement take 2 weeks to 1 month. The nuclear disaster at the Fukushima Daiichi Nuclear Power Plant released radionuclides into the environment, and confidence in the safety of nuclear power generation subsequently decreased [2]. We may soon enter an era in which many nuclear reactors will be decommissioned. Safe decommissioning of nuclear reactors will require the establishment of new methods for real-time monitoring of 90Sr because the currently available methods are time consuming. Radioisotopes can be measured by means of various mass spectrometry techniques, such as accelerator mass spectrometry [3], resonance ionization mass spectrometry [4], and thermal ionization mass spectrometry [5]. Inductively coupled plasma mass spectrometry (ICP-MS) is a particularly powerful tool for analyzing long-lived radioisotopes owing to its high sensitivity for alkaline earth metals, high analytical speed, being widely installed in many laboratories, low sample 90

Nuclear power plant accidents can release radioactive materials into the environment; the major radionuclides released by such accidents are radioactive strontium (89Sr and 90Sr), iodine (129I and 131I), and cesium (134Cs and 137Cs). Exposure to 90Sr, which has a relatively long half-life (29 years), results in long-term delivery of beta radiation to the skeleton and surrounding tissues because its chemical properties are similar to those of Ca. Therefore, monitoring of 90Sr after a nuclear power plant accident is important. However, unlike 129I, 131I, 134Cs, and 137Cs, which are gamma emitters and thus can be detected by means of a Ge semiconductor detector or a NaI(Tl) scintillation detector, 90 Sr emits only beta radiation, which is difficult to detect because its energy can take any value up to the maximum value for a given isotope [1]. In addition, interferences from 90Y, which is produced by β decay of 90Sr, ⁎ Corresponding author at: Faculty of Life and Environmental Science, Shimane University, 1060 Nishikawatsu-cho, Matsue-shi, Shimane 690-8504, Japan. E-mail address: [email protected] (Y. Suzuki).

http://dx.doi.org/10.1016/j.sab.2017.07.007 0584-8547/© 2017 Elsevier B.V. All rights reserved.

Y. Suzuki et al. / Spectrochimica Acta Part B 135 (2017) 82–90

consumption, ease of sample preparation, and non-necessity for radioisotope qualification [6,7]. Zoriy et al. [6] determined 90Sr in groundwater samples from the Semipalatinsk Test Site area in Kazakhstan by double-focusing sector field ICP-MS operated at medium mass resolution under cold plasma conditions. Takagai et al. [7] quantified 90Sr by means of single quadrupole ICP-MS (ICP-QMS) in combination with online chelate column separation and an oxygen reaction; these investigators reported a detection limit (DL) of 2.3 Bq L−1 (0.46 ng L−1 as 90Sr) with an analytical run time (from injection to mass spectrometric detection) of 14.6 min. Wu et al. [8] analyzed Pu isotopes by using sector field ICP-MS after two-stage ion-exchange chromatography separation. More recently, radioisotope analysis by means of tandem quadrupole ICP-MS (ICP-MS/MS) has been reported [9–11]. Radioisotopes have traditionally been analyzed after separation and purification, but ICP-MS/MS allows for chemical and physical separation in the instrument by the use of two quadrupole mass filters (Q1 and Q2) and a collision/reaction cell (CRC) located in-between [12,13]. Ohno et al. [9] developed an ICPMS/MS method for the determination of 129I in soil samples by using O2 as a reaction gas to reduce spectral interferences from 127IH2, 127ID, and 129 Xe. Ohno and Muramatsu [10] used nitrous oxide as the reaction gas to reduce isobaric interferences from 134Ba, 135Ba, 137Ba, and 134Xe and thus were able to determine 134Cs/137Cs and 135Cs/137Cs ratios in rainwater samples by means of ICP-MS/MS. Zheng et al. [11] analyzed radiocesium isotopes in soil and litter samples by using ICP-MS/MS after ammonium molybdophosphate-selective adsorption of Cs and subsequent two-stage ion-exchange chromatographic separation. Isotopic analysis of Sr has been carried out via ICP-MS/MS using different reaction gases (e.g., O2 and CH3F) [14–16]. For analysis of 90Sr, the use of reactions of Sr and Zr with O2 by means of ICP-MS with a collision/reaction cell (CRC) have been reported [17–20]. In addition, Agilent described a method for analysis of 90Sr by means of ICP-MS/ MS; [21] specifically, reaction with O2 and H2 effectively reduces isobaric interference from Zr and results in low estimated background equivalent concentration (BEC) and DL values (0.08 ng L−1 and 0.23 ng L−1, respectively) for 90Sr. However, the effects of isobaric interference from 90 + Y or 89Y1H+ and the removal of these species were not described. Methods for analysis of 90Sr released into the atmosphere have not yet been established. Previously, we reported a system for real-time determination of metal concentrations in atmospheric particulate matter (APM) by means of the combination of a gas-exchange device (GED) and ICPMS. [22,23] In the case of the Fukushima nuclear accident, spherical microparticles, which diameter is b 2.5 μm, containing radioactive Cs and U were found in aerosol samples collected on March 14th and 15th, 2011, in Tsukuba, 172 km southwest of the Fukushima Daiichi Nuclear Power Plant [24,25]. PM2.5 was set as the target APM in this study. The concentration of Y in PM2.5 is lower than 1.0 ng m−3 [26], and the effect of spectral interference is expected to be small. However, when particles that temporarily contain a large amount of Y enter the analysis system, spike-shaped signals may be observed in real-time analysis. Therefore, it is necessary to evaluate chemical species such as 89Y1H+. In this study, we modified this previously reported system and examined the capability about real-time analysis of 90Sr in PM2.5 by means of GED-ICP-MS/MS. The Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT) regulates the lowest concentrations of 90Sr as strontium titanate in atmosphere to 3 × 10−4 Bq cm−3 [27]. And aim of this study is to develop a real-time monitoring method with limit of detection closed to the regulated value in Japan.

83

Tokyo, Japan) were used for preparation of the final sample solution. Single-element standard solutions (1000 mg L−1) of Ti, Mn, Fe, Ge, Sr, Y, and Zr were purchased from Kanto Chemical Co. Oxygen, helium, ammonia (10% in helium), and argon gases were purchased from Sanin Sanso Corp. (Shimane, Japan). Hydrogen was loaded into hydrogen gas storage purification equipment (HB-SC-0050-Q, HBank Technologies, New Taipei City, China) by IAS (Tokyo, Japan). 2.2. System for real-time analysis of 90Sr in PM2.5 A schematic diagram of the analytical system is shown in Fig. 1. Outdoor air samples were collected with a 10-m-long Tygon tube by means of a diaphragm pump equipped with a mass flow controller. In addition, a blank air sample was collected from a clean booth (class 100) equipped with HEPA filters. Each air sample was introduced to an impactor (Sioutus Personal Cascade Impactor, SKC Inc., Eighty Four, PA, USA) by means of accelerator plate A at a flow rate of 9.0 L min−1 to select PM2.5. A diaphragm pump equipped with a mass flow controller was used to remove 8.75 L min−1 of each air sample, and the remaining 0.25 L min− 1 was introduced into a GED (J-Science, Kyoto, Japan), which replaced the air with Ar. Then this portion of the sample was introduced to the ICP-MS/MS system (8800, Agilent Technologies, Tokyo, Japan) by means of a micro diaphragm gas-sampling pump (NMP05L, KNF Neuberger AG, Balterswil, Switzerland). The ICP-MS/MS consists of two quadrupole mass filters (Q1 and Q2) separated by an octopole CRC set between Q1 and Q2. Each mass filter can pass any selected mass-to charge (m/z) ratio with unit mass resolution in MS/MS mode. Agilent 8800 has collision/reaction cell with 4 cell gas lines. In the MassHunter software of the Agilent 8800, addition of 0.5 mL min−1 He is mandatory when using NH3. The ICP-MS/MS system can also be operated in the single quadrupole mode, in which case Q1 operates as an ion guide. In this study, we used the single quadrupole and MS/MS for non-gas mode and gas modes, respectively. During analysis of air or blank air, a high-efficiency sample introduction system (HESIS, Apex-Q, Elemental Scientific Inc., Omaha, NE, USA) was used to introduce ultrapure water into the ICP-MS/MS. During analysis of standard solution, element standard solutions were introduced into ICP-MS/MS via the HESIS. 2.3. Optimization of CRC gas conditions The operating conditions for the ICP-MS/MS system are summarized in Table 1. All other parameters were optimized for the O2 conditions by means of the autotune function in the MassHunter 4.1 software (version C.01.01, Agilent Technologies, Tokyo, Japan), and the optimized parameters were used even under the other gas conditions. A mixed solution containing 100 μg L−1 Sr, 10 mg L− 1 Y, and 10 mg L−1 Zr was used to optimize the reaction gas flow rate. In this measurement, CRC condition was summarized in Table S1 (Appendix). For analysis of mass spectra, a mixed solution of Sr, Y, and Zr, each at 100 μg L− 1, was used. 10% ethanol solution and a mixed solution containing 100 μg L− 1 Sr, 10 mg L−1 Y, Zr, Ti, Mn, Fe, and Ge, 1% HNO3, and 3% HCl were used to evaluate polyatomic ion of 38Ar40Ar12C+, 50 40 Ti Ar+, 55Mn35Cl+, 54Fe36Ar+, and 74Ge16O+ for analysis of mass spectra. For optimization, the mixed solution was introduced via the HESIS without the use of the GED. Finally, for application to real-time monitoring, we manually optimized CRC cell conditions under the optimized gas flow rate to obtain the highest intensity of 59Co.

2. Experimental 2.4. Calibration of the HESIS 2.1. Reagents Nitric acid (61%, electronic laboratory grade, Kanto Chemical Co., Tokyo, Japan), hydrochloric acid (36%, electronic laboratory grade, Kanto Chemical Co.) and ultrapure water (Direct-Q UV3, Millipore,

We determined the introduction efficiency for Sr and other elements via the HESIS by means of a procedure described by Suzuki et al. [23] with slight modification to trapping in tandem impingers. Using the introduction efficiency, we converted the element concentrations in the

84


Fig. 1. Schematic diagram of analytical system for real-time 90Sr analysis.

multi-element standard solutions, in nanograms per milliliter, to element mass flow rates in nanograms per minute. 2.5. Sr recoveries of real-time analysis by comparing with simultaneous filter-collection data To evaluate recoveries of Sr in PM2.5 in real-time analysis, PM2.5 were simultaneously collected on a cellulose nitrate filter while real-time analysis was performed. In this analysis, only natural isotopes were analyzed. Measurements were performed on March 19, April 14, May 3, June 20, and July 11 in 2016. Samples collected on filter were aciddigested as described in the literature with modification [28]. HNO3 (6 mL), HF (3 mL), and H2O2 (1 mL) were added to the samples, which were then digested with digestion vessel for high pressure (HSU-50, San-Ai Kagaku Co. Ltd., Aichi, Japan) at 180 °C for 12 h. After digestion, HF was evaporated by means of a hot plate (230 °C). When the solution was reduced to a single transparent droplet (ca. 0.05 mL), it was diluted to 10 mL with ultrapure water. Trace element concentrations in the solution were determined by ICP-MS/MS, and the concentration units of nanograms per milliliter were converted to nanograms per cubic meter by using an air flow of 8.75 L min−1. 2.6. Data analysis The acquisition parameters are summarized in Table 1. The single quadrupole and MS/MS modes were used for the no-gas conditions and collision/reaction gas conditions, respectively. All data, except for outdoor air measurements, were obtained in triplicate. Table 1 Operating conditions of inductively coupled plasma tandem mass spectrometry. Plasma conditions RF power RF matching Sampling depth Plasma gas (Ar) Auxiliary gas (Ar) Carrier gas (Ar) Acquisition parameters MS mode No. of peaks per mass

Integration time per mass

1550 W 1.8 V 8.5 mm 15 L min−1 0.95 L min−1 0.95 L min−1 without the gas-exchange device 0.70 L min−1 with the gas-exchange device Single quadrupole, no gas MS/MS, with reaction gas(es) 1 point/mass for optimization of the cell gas(es) 6 points/mass for product-ion scans 20 point/mass for neutral mass gain scan 0.1 s, all scans except neutral mass gain scan 0.2 s, neutral mass gain scan

To optimize the CRC gas conditions, we used 1 point/mass for peak quantitation. For product-ion scans, Q2 was set to scan the mass range from m/z 80 to 200, and Q1 was set at a fixed mass; these settings cover the product ions formed from Sr, Y, and Zr under O2 + H2 + NH3 conditions. We used 6 points/mass for peak quantitation to obtain peak shapes. For neutral mass gain scans both Q1 and Q2 were set to scan from m/z 80 to 100. And we used 20 points/peak for peak quantitation to evaluate the abundance sensitivity by MS/MS method. For real-time analysis of atmospheric 90Sr, time-resolved analysis was used with an integration time per mass of 0.1 s, 1 point/peak for peak quantitation, and an integration time of 600 s to shorten the measurement cycle. For analysis of atmospheric natural isotopes (88Sr), we used 1 point/peak for peak quantitation with an integration time per mass of 0.1 s and 3 times repetition.

3. Results and discussion 3.1. Reactions of Sr, Y, and Zr in the CRC The effects of the reaction of gas flow rates in the CRC on 88Sr+, 89Y+, and 90Zr+ signal intensities are shown in Fig. 2. Note that the reactivities of O2 with Y+ and Zr+ are known to be high, and the sensitivities of the respective + 16 reaction product ions (89Y16O+ and 90Zr16O+) under the O2 gas mode are 52% and 46% of the sensitivities of Y+ and Zr+ under the no-gas mode [29]. We found that at an O2 flow rate of 1.0 mL min−1, the intensities of the signals for 89Y+ (m/z = 89 at Q1 and Q2) and 90Zr+ (m/z = 90 at Q1 and Q2) were 1/100,000 and 1/ 10,000 the corresponding values in the absence of O2 (Fig. 2a). In contrast, the 88Sr signal intensity was reduced by a factor of only approximately 1/10. Shikamori and Nakano reported that introduction of O2 and H2 reduces isobaric interference on 90Sr from 90Zr, but these researchers did not investigate the utility of this method for reducing interference from 89Y [21]. We found that although the introduction of H2 and O2 (O2 flow rate 1.0 mL min−1) did reduce the 90Zr signal intensity by at least a factor of 1/100, these conditions increased the 89Y signal intensity by a factor of 2.5–10 (Fig. 2b). That is, the introduction of H2 was useful for eliminating interference from Zr but did not eliminate interference from Y. The reactivity of NH3 with Zr and Y is reported to be low: the yields of the +15 (NH) reaction product ions (89Y14NH+ and 90Zr14NH+) are reported to be only 15% and 17%, and 32% and 20% of 89Y+ and 90Zr+ were remained not to produce product ion [29]. In addition, the rate of production of molecular clusters containing two or more NH1–3 species is reported to be 1 order of magnitude lower than those containing


85

Fig. 2. Effect of collision/reaction cell gas flow rates on 88Sr, 89Y, and 90Zr signal intensities: (a) O2, (b) H2 in combination with 1.0 mL min−1 O2, (c) NH3 in combination with 1.0 mL min−1 O2, and (d) NH3 in combination with 1.0 mL min−1 O2 and 10 mL min−1 H2. Error bar indicate standard deviation from triplicate analysis.

one NH1–3 species. These results suggested to us that the utility of NH3 alone as a CRC gas for analysis of Y and Zr was likely to be low. Therefore, we evaluated the combination of O2 and NH3. We varied the NH3 flow rate from 0.0 to 8.0 mL min−1 in combination with O2 at a flow rate of 1.0 mL min−1. Although these conditions eliminated isobaric interference from Zr, the Sr signal intensities were 1/700 those with O2 alone at a flow rate of 1.0 mL min−1 (Fig. 2c). Note that we actually used a mixture containing 90% He and 10% NH3, and collision with He, which occurs at the same time as reaction with NH3, markedly decreased the signal intensity. This result suggested that although NH3 could effectively eliminate spectral interference on 90Sr, the flow rate of NH3 gas had to be as low as possible. Therefore, we conducted an experiment with the following flow rates: NH3, 0–1.0 mL min− 1; O2, 1.0 mL min− 1; H2, 10.0 mL min− 1; and He, 0.5 mL min− 1. Under these conditions, the Sr signal intensity was 1/10 of that at 1.0 mL min− 1 of O2 gas. The intensities of the Y signals decreased by 3 orders of magnitude but Zr signals did not change as the NH3 flow rate was varied from 0 to 1.0 mL min − 1 . (Fig. 2d). The ratio of the signal of the analyte ion (88 Sr +) to the interfering ions (sum of 89Y+ and 90Zr+) showed maximum value (7.1 × 10 3) under the condition of 1.0 mL min − 1 O2 , 10.0 mL min− 1 H2, and 1.0 mL min− 1 NH3. These results indicated that by using a mixture of gases in the CRC, we could minimize spectral interference. To elucidate the reaction of Zr+ with CRC gases, we carried out experiments in product-ion-scan mode to identify the molecular ions generated in the CRC. For the product-ion scan, Q1 was fixed at m/z = 90, and Q2 was scanned from 80 to 200 (Fig. 3). For the Zr+ and O2 reaction, signals at m/z = 90 for Q2 were not observed, indicating that Zr+ ions reacted readily to form oxides. The highest signal intensities were observed at m/z = 106 (ZrO) for Q2, and the intensities of the ZrO+ 1–6 signals decreased as the number of oxygen atoms increased (Fig. 3a). In contrast, in NH3 mode, the highest signal intensity was observed at m/z = 174, which corresponds to the ZrN5H+ 14 product ion (Fig. 3b). In addition, the signal intensity at m/z = 90, indicating the presence of unreacted Zr ion, was comparable to that at m/z = 174. In other words, although the reaction of NH3 and Zr formed relatively large

product ions, the reaction efficiency was not high enough to prevent unreacted Zr ion from remaining in the CRC. When the combination of O2 and H2 was used, the basic mass spectra were similar to those obtained under O2 conditions (Fig. 3c). Furthermore, the signal intensity was highest at m/z = 174 for Q2, and this signal was attributed to ZrO5H+ 4 ; the higher signal intensities were observed at the higher m/z as molecular ion bound to many oxygen atoms. When the combination of O2, H2, and NH3 was used in the CRC, the signal intensities at m/z = 90 for Q2 were 0 cps (Fig. 3d). We believe that molecular ions were efficiently produced under these conditions, but no signal was detected at m/z = 90 for both Q1 and Q2. The highest signal was observed at m/z = 175 for Q2; we attributed this peak to ZrOH(NH3)4. Under O2 and O2 + H2 conditions, the mass spectral patterns of the molecular ions ZrOmHn, where m = 3–5 and n = 0–7, were bimodal; whereas unimodal mass spectra were observed under O2 + H2 + NH3 conditions, similar to the case for NH3 conditions. These results suggest that the generated molecular ion under O2 + H2 + NH3 conditions was ZrOH(NHm)n, where m = 2, 3 and n = 3–5. Under O2 + H2 + NH3 conditions, molecular ions such as ZrO1–3 were not detected. Furthermore, when NH3 was used, (NH1–3)1–3 adducts with molecular weights of b 140 were produced. This result suggests that the reactivity of ZrO with NH3 was high and that Zr(NHm)n, where m = 2, 3 and n = 3–5, was efficiently produced. In addition, it was suggested that formation of a large molecular ion suppressed the fragmentation of molecular ions by collision with He. The reactions of Y with O2, O2 + H2, NH3, and O2 + H2 + NH3 were similar to those of Zr (Fig. 4). The intensity of the signal for Y at m/z = 89 for Q2 (in the form of 89Y, which we considered to be a potential source of isobaric interference on 90Y) was lowest when the gas flow rates were 1.0 mL min−1 for O2, 10.0 mL min−1 for H2, and 1.0 mL min−1 for NH3 (Fig. 4d). No 89Y1H+ ion, a polyatomic ion that is derived from naturally occurring isotopes and that interferes with 90Sr, formed in the CRC under O2 + H2 + NH3 conditions, although this ion was generated under NH3 conditions. Although we could not find it in our research, simultaneous measurement of 90Sr+, 90Y+, and 90Zr+ might be possible

86


Fig. 3. Mass spectra for product-ion scan for m/z = 90 at Q1 and scan from m/z = 80 to 200 at Q2 obtained under the following reaction gas conditions: (a) 1.0 mL min−1 O2; (b) 2.0 mL min−1 NH3; (c) 10.0 mL min−1 H2 and 1.0 mL min−1 O2; and (d) 1.0 mL min−1 O2, 10.0 mL min−1 H2, and 1.0 mL min−1 NH3.

by examining the use of other reaction gases, the optimal combination of reaction gases, and the generation of higher order reaction products in detail. Fig. 5 shows overlaid mass spectra of a mixed solution of Sr, Y, and Zr, each at 100 μg L−1 for natural-gain scan for m/z = 83–97 (Q1 = Q2) under no-gas conditions (gray area) and conditions for 1.0 mL min−1 O2, 10.0 mL min−1 H2, and 1.0 mL min−1 NH3 (black area). Although the intensities of 88Sr signals had decreased to 1/100 times, the 90Zr signals could be eliminated and the 89Y+ signals could be reduced to 1/ 100000. Under this condition, we could measure 90Sr with by minimizing the influence of 89Y1H+, 90Y+, and 90Zr+. Under the conditions of 1.0 mL min−1 O2, 10.0 mL min−1 H2, and 1.0 mL min−1 NH3, the spectral interferences of 38Ar40Ar12C, 50Ti40Ar, 55Mn35Cl, 54Fe36Ar, and 74 Ge16O were also eliminated (Fig. S1 and S2, Appendix).

3.2. Estimation of BEC and DL of 90Sr by real-time monitoring Using the optimized CRC gas conditions (1.0 mL min− 1 O2, 10.0 mL min−1 H2, and 1.0 mL min−1 NH3), we attempted to measure 90 Sr concentrations in PM2.5 by using the GED-ICP-MS/MS system. Fig. 6a and b shows chromatograms of 88Sr intensity for a blank solution and a 100 μg L−1 Sr solution obtained with continuous introduction of the blank or standard solution and clean air. As the autosampler moved, the signal intensity fluctuated for the first 50 s of the measurement. Therefore, after correcting with the signal intensity for In, which we used as an internal standard, we integrated the signal intensities during the period from 50 to 600 s and used the resulting data for the subsequent calculation. Fig. 6c shows the signal intensity of 90 Sr at m/z = 90 for both Q1 and Q2 for the blank solution. The average


87

Fig. 4. Mass spectra for product-ion scan at m/z = 89 for Q1 and scan from m/z = 80 to 200 for Q2 under the following reaction gas conditions: (a) 1.0 mL min−1 O2; (b) 2.0 mL min−1 NH3; (c) 10.0 mL min−1 H2 and 1.0 mL min−1 O2; and (d) 1.0 mL min−1 O2, 10.0 mL min−1 H2, and 1.0 mL min−1 NH3.

blank signal after internal standard correction and integration was 23.0 ± 2.8 counts. Using the method of Suzuki et al. [23], we calculated the introduction efficiency of Sr to be 8.8%. Then we converted the Sr concentration in the solution into the mass concentration of Sr in the atmosphere by using Eq. 1: C air ¼ C l F l EF=F air

ð1Þ

where Cair is the concentration (ng m−3) of elements (here in Sr, Y, and Zr); Cl is the concentration (ng mL−1) of elements in the multi-element external standard solution introduced into the HESIS; Fl is the uptake rate (mL min−1) of the multi-element external standard solution into the HESIS; EF is the sample introduction efficiency of HESIS; and Fair is the flow rate of air (m3 min−1) introduced into the ICP.

The slope of 90Sr was estimated from Eq. 2:

90

Slope of Sr ¼

88

Slope of Sr 88

Abundance of Sr ð0:8258Þ

88

Mass of Srð88Þ 90

Mass of Srð90Þ

ð2Þ

Using an estimated calibration curve for 90Sr (Fig. 6d), we estimated BEC and DL values for this system to evaluate its performance for analysis of 90Sr (Table 2). The estimated BEC and DL values for our analytical system were 9.7 × 10−4 and 3.6 × 10−4 ng m−3, which are equivalent to 4.9 × 10−6 and 1.8 × 10−6 Bq cm−3 when the specific radioactivity of 90Sr (5.09 × 1012 Bq g−1) is considered. Insufficient fusion or vaporization of particles directly introduced into plasma has been reported previously [30,31]. In addition to incomplete fusion and vaporization, signal reduction for high-melting metal oxides was reported for particles generated by laser ablation and direct

88


Fig. 5. Overlaid mass spectra for neutral mass gain scan for m/z = 83–97 (Q1 = Q2). Gray area, no-gas conditions, single quadrupole mode; black area: 1.0 mL min−1 O2, 10.0 mL min−1 H2, 1.0 mL min−1 NH3, MS/MS mode.

introduction of PM1.0 via GED [23,32–35]. We simultaneously collected PM2.5 on a filter during real-time analysis, and examined Sr recoveries of real-time analysis by comparing with filter-collection data. Recovery of Sr was 53 ± 23%, and the DL as PM2.5 was estimated to be 6.8 ± 3.0 × 10− 4 ng m− 3 and 3.4 ± 1.5 × 10− 6 Bq cm− 3 after divided DL by this recovery. It is considered that the aerodynamic diameter of the particulate matters containing Sr is not constant at all times. And ionization efficiencies are expected to depend on particle size. In that respect, it may be necessary to consider large uncertainty with quantitative results. For 90Sr in compounds other than strontium titanate, MEXT regulates that the lowest concentrations that can be determined in atmosphere and exhaust are 7 × 10−4 and 5 × 10−6 Bq cm−3, respectively, and the corresponding values for strontium titanate are 3 × 10−4 and

8 × 10−6 Bq cm−3, respectively [27]. With the analytical system reported herein, we could measure 90Sr in PM2.5 below these regulation values in only 10 min, even considering the fusion/vaporization/ionization efficiency of Sr in PM2.5. 3.3. Real-time monitoring of environmental 90Sr with our analysis system To evaluate the practical utility of our system, we measured 90Sr concentrations over the course of 1 h in PM2.5 outside our laboratory at Shimane University, which is located within a radius of 20 km from the Shimane Nuclear Power Plant. We observed some spikes in the concentration of the naturally occurring isotope, 88Sr, which was measured at m/z = 88 for both Q1 and Q2 as a positive control (Fig. 7a). In sizeclassified particulate matter collected at Shimane University, the

Fig. 6. Standard and blank signals for naturally occurring Sr and 90Sr obtained upon introduction of clean air: (a) 88Sr signal for blank solution; (b) 88Sr signal for 100 μg L−1 of Sr solution; (c) 90Sr signal for blank; and (d) estimated calibration curve for 90Sr after integration.

Y. Suzuki et al. / Spectrochimica Acta Part B 135 (2017) 82–90 Table 2 Performance of inductively coupled plasma tandem mass spectrometry system with gasexchange device for analysis of 90Sr in PM2.5 (analytical run time, 10 min). Blank signal at m/z = 90 Estimated slope for 90Sr Estimated background equivalent concentration for 90Sr Estimated detection limit for 90Sr Estimated detection limit for 90Sr as PM2.5

23.0 ± 2.8 counts 23,616 (cps m3 ng−1) 9.7 × 10−4 (ng m−3) 4.9 × 10−6 (Bq cm−3) 3.6 × 10−4 (ng m−3) 1.8 × 10−6 (Bq cm−3) 6.8 ± 3.0 × 10−4 (ng m−3) 3.4 ± 1.5 × 10−6 (Bq cm−3)

average Sr concentration in PM2.5 has been measured to be 0.12 ng m−3, which corresponds to about 1000 counts and is in good agreement with the results we obtained by real-time analysis with our system (Table S2, Appendix). The signal intensity for 90Sr in the outside air was 22.8 ± 6.7 counts for 10 min, which was less than the DL for our system; that is, 90 Sr was not detected in the PM2.5 that we collected at Shimane University (Fig. 7b). Although 90Sr was not be detected under general circumstances, the fact that 88Sr was observed with a reasonable value indicates that 90Sr would be detected in our system if 90Sr over the DL value exists. We obtained no signal at m/z = 90 for Q1 and Q2, and Zr in the atmosphere was detected as 90Zr(NH3)5 (m/z = 175 for Q2; Fig. 7c). These results demonstrate the capability to monitor 88Sr and 90Sr simultaneously while also eliminating Zr interference even in the real

89

environment. In addition, we confirmed that the 90Sr concentration in the atmosphere at Shimane University was below the regulation value. It is expected to be applied to the case in high urgency such as nuclear accident and monitoring during decommissioning of nuclear reactors.

4. Conclusion We have optimized CRC gases and their combination for determination of 90Sr in PM2.5 by ICP-MS/MS after direct introduction of air via a GED. Under the conditions of 1.0 mL min− 1 O2, 10.0 mL min− 1 H2, and 1.0 mL min−1 NH3, although the intensities of 88Sr+ signals had decreased to 1/100 times, spectral interference from 89Y (89Y1H+ and 90 + Y ) and 90Zr could be reduced. Under optimized condition, our analytical system enabled real-time analysis of 90Sr in PM2.5 over the course of only 10 min and showed the 90Sr DL of 3.4 ± 1.5 × 10−6 Bq cm−3 even considering the fusion/vaporization/ionization efficiency of Sr in PM2.5. Although the DL of the developed system was not enough to detect 90Sr in general environment, our analytical system would be useful in the high urgency situation such as nuclear accident and monitoring during decommissioning of nuclear reactors. By mixing a plurality of reaction gases, it is possible to form larger molecular ions. Therefore, if reactions in the cell can be grasped correctly, mixing of multiple reaction gases may be useful for solving the spectral interference which is difficult to remove by using single gas.

Fig. 7. Real-time analysis of (a) 88Sr, (b) 90Sr, and (c) Zr (as 90Zr(NH3)5) in PM2.5 matter. Sampling was conducted on 9 November 2015.

90


Acknowledgements We acknowledge Mr. Kohei Nishiguchi (J-Science Lab Co.) for his support and invaluable comments. This research was partially supported by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan (no. 15K21171). The authors thank the Faculty of Life and Environmental Science at Shimane University for financial assistance for the publication of this paper.

[15]

[16] [17]

[18]

Appendix A. Supplementary data [19]

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.sab.2017.07.007.

[20]

References [1] N. Vajda, C.K. Kim, Determination of radiostrontium isotopes: a review of analytical methodology, Appl. Radiat. Isot. 68 (2010) 2306–2326. [2] International Atomic Energy Agency (IAEA), Fukushima Nuclear Accident Information Sheet, http://www.iaea.org/About/japan-infosheet.html (accessed Feb 3, 2017). [3] F. Arslan, M. Behrendt, W. Ernst, E. Finckh, G. Greb, F. Gumbmann, M. Haller, S. Hofmann, R. Karschnick, M. Klein, W. Kretschmer, J. Mackiol, G. Morgenroth, C. Pagels, M. Schleicher, Trace Analysis of the Radionuclides Sr-90 and Sr-89 in Environmental-Samples. 2. Accelerator Mass-Spectrometry (Ams), Angewandte Chemie-International Edition in English, 34, 1995 183–186. [4] K. Wendt, G.K. Bhowmick, B.A. Bushaw, G. Herrmann, J.V. Kratz, J. Lantzsch, P. Muller, W. Nortershauser, E.W. Otten, R. Schwalbach, U.A. Seibert, N. Trautmann, A. Waldek, Rapid trace analysis of Sr-89,Sr-90 in environmental samples by collinear laser resonance ionization mass spectrometry, Radiochim. Acta 79 (1997) 183–190. [5] Y. Shibahara, T. Kubota, T. Fujii, S. Fukutani, T. Ohta, K. Takamiya, R. Okumura, S. Mizuno, H. Yamana, Analysis of cesium isotope compositions in environmental samples by thermal ionization mass spectrometry-1. A preliminary study for source analysis of radioactive contamination in Fukushima prefecture, J. Nucl. Sci. Technol. 51 (2014) 575–579. [6] M.V. Zoriy, P. Ostapczuk, L. Halicz, R. Hille, J.S. Becker, Determination of Sr-90 and Pu isotopes in contaminated groundwater samples by inductively coupled plasma mass spectrometry, Int. J. Mass Spectrom. 242 (2005) 203–209. [7] Y. Takagai, M. Furukawa, Y. Kameo, K. Suzuki, Sequential inductively coupled plasma quadrupole mass-spectrometric quantification of radioactive strontium-90 incorporating cascade separation steps for radioactive contamination rapid survey, Anal. Methods 6 (2014) 355–362. [8] J. Wu, J. Zheng, M. Dai, C.A. Huh, W. Chen, K. Tagami, S. Uchida, Isotopic composition and distribution of plutonium in northern South China Sea sediments revealed continuous release and transport of Pu from the Marshall Islands, Environ. Sci. Technol. 48 (2014) 3136–3144. [9] T. Ohno, Y. Muramatsu, Y. Shikamori, C. Toyama, N. Okabe, H. Matsuzaki, Determination of ultratrace I-129 in soil samples by Triple Quadrupole ICP-MS and its application to Fukushima soil samples, J. Anal. At. Spectrom. 28 (2013) 1283–1287. [10] T. Ohno, Y. Muramatsu, Determination of radioactive cesium isotope ratios by triple quadrupole ICP-MS and its application to rainwater following the Fukushima Daiichi Nuclear Power Plant accident, J. Anal. At. Spectrom. 29 (2014) 347–351. [11] J. Zheng, W.T. Bu, K. Tagami, Y. Shikamori, K. Nakano, S. Uchida, N. Ishii, Determination of Cs-135 and Cs-135/Cs-137 atomic ratio in environmental samples by combining ammonium molybdophosphate (AMP)-selective Cs adsorption and ionexchange chromatographic separation to triple-quadrupole inductively coupled plasma-mass spectrometry, Anal. Chem. 86 (2014) 7103–7110. [12] L. Balcaen, E. Bolea-Fernandez, M. Resano, F. Vanhaecke, Inductively coupled plasma - tandem mass spectrometry (ICP-MS/MS): a powerful and universal tool for the interference-free determination of (ultra) trace elements - a tutorial review, Anal. Chim. Acta 894 (2015) 7–19. [13] E. Bolea-Fernandez, L. Balcaen, M. Resano, F. Vanhaecke, Overcoming spectral overlap via inductively coupled plasma-tandem mass spectrometry (ICP-MS/MS). A tutorial review, J. Anal. At. Spectrom. (2017) http://dx.doi.org/10.1039/C7JA00010C. [14] E. Bolea-Fernandez, S.J.M. Van Malderen, L. Balcaen, M. Resano, F. Vanhaecke, Laser ablation-tandem ICP-mass spectrometry (LA-ICP- MS/MS) for direct Sr isotopic

[21] [22]

[23]

[24] [25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

analysis of solid samples with high Rb/Sr ratios, J. Anal. At. Spectrom. 31 (2016) 464–472. E. Bolea-Fernandez, L. Balcaen, M. Resano, F. Vanhaecke, Tandem ICP-mass spectrometry for Sr isotopic analysis without prior Rb/Sr separation, J. Anal. At. Spectrom. 31 (2016) 303–310. T. Zack, K.J. Hogmalm, Laser ablation Rb/Sr dating by online chemical separation of Rb and Sr in an oxygen-filled reaction cell, Chem. Geol. 437 (2016) 120–133. D.R. Bandura, V.I. Baranov, S.D. Tanner, Reaction chemistry and collisional processes in multipole devices for resolving isobaric interferences in ICP-MS, Fresenius J. Anal. Chem. 370 (2001) 454–470. A.P. Vonderheide, M.V. Zoriy, A.V. Izmer, C. Pickhardt, J.A. Caruso, P. Ostapczuk, R. Hille, J.S. Becker, Determination of 90Sr at ultratrace levels in urine by ICP-MS, J. Anal. At. Spectrom. 19 (2004) 675. G. Favre, R. Brennetot, F. Chartier, P. Vitorge, Understanding reactions with O-2 for Sr-90 measurements by ICP-MS with collision-reaction cell, Int. J. Mass Spectrom. 265 (2007) 15–22. J. Feuerstein, S.F. Boulyga, P. Galler, G. Stingeder, T. Prohaska, Determination of (90)Sr in soil samples using inductively coupled plasma mass spectrometry equipped with dynamic reaction cell (ICP-DRC-MS), J. Environ. Radioact. 99 (2008) 1764–1769. Y. Shikamori, K. Nakano, Agilent 8800 ICP-QQQ Application Handbook5991-2802EN https://www.agilent.com/cs/library/Brochures/5991-2802EN_electronic.pdf 2013. Y. Suzuki, H. Sato, S. Hikida, K. Nishiguchi, N. Furuta, Real-time monitoring and determination of Pb in a single airborne nanoparticle, J. Anal. At. Spectrom. 25 (2010) 947. Y. Suzuki, H. Sato, K. Hiyoshi, N. Furuta, Quantitative real-time monitoring of multielements in airborne particulates by direct introduction into an inductively coupled plasma mass spectrometer, Spectrochim. Acta B 76 (2012) 133–139. K. Adachi, M. Kajino, Y. Zaizen, Y. Igarashi, Emission of spherical cesium-bearing particles from an early stage of the Fukushima nuclear accident, Sci. Report. 3 (2013). Y. Abe, Y. Iizawa, Y. Terada, K. Adachi, Y. Igarashi, I. Nakai, Detection of uranium and chemical state analysis of individual radioactive microparticles emitted from the Fukushirna nuclear accident using multiple synchrotron radiation X-ray analyses, Anal. Chem. 86 (2014) 8521–8525. Q.L. Dai, L.W. Li, T.K. Li, X.H. Bi, Y.F. Zhang, J.H. Wu, B.S. Liu, J.X. Gao, W.C. Gu, L. Yao, Y.C. Feng, Atmospheric signature and potential sources of rare earth elements in size-resolved particulate matter in a megacity of China, Aerosol Air Qual. Res. 16 (2016) 2085–2095. The Ministry of Education, Culture, Sports, Science and Technology of Japan, Science and Technology Agency Notification No. 5 in Heisei 12(Quantity of isotopes emitting radiation) http://www.mext.go.jp/b_menu/hakusho/nc/k20001023001/ k20001023001.html (accessed Feb 3, 2017). N. Furuta, A. Iijima, A. Kambe, K. Sakai, K. Sato, Concentrations, enrichment and predominant sources of Sb and other trace elements in size classified airborne particulate matter collected in Tokyo from 1995 to 2004, J. Environ. Monit. 7 (2005) 1155–1161. N. Sugiyama, K. Nakano, Agilent Technical note, 2014, 5991-4585EN, http://www. agilent.com/cs/library/technicaloverviews/public/5991-4585EN_TechNote8800_ ICP-QQQ_reactiondata.pdf. R.S. Houk, R.K. Winge, X.S. Chen, High speed photographic study of wet droplets and solid particles in the inductively coupled plasma, J. Anal. At. Spectrom. 12 (1997) 1139–1148. D.B. Aeschliman, S.J. Bajic, D.P. Baldwin, R.S. Houk, High-speed digital photographic study of an inductively coupled plasma during laser ablation: comparison of dried solution aerosols from a microconcentric nebulizer and solid particles from laser ablation, J. Anal. At. Spectrom. 18 (2003) 1008–1014. M. Guillong, D. Günther, Effect of particle size distribution on ICP-induced elemental fractionation in laser ablation-inductively coupled plasma-mass spectrometry, J. Anal. At. Spectrom. 17 (2002) 831–837. M. Guillong, H.R. Kuhn, D. Gunther, Application of a particle separation device to reduce inductively coupled plasma-enhanced elemental fractionation in laser ablation-inductively coupled plasma-mass spectrometry, Spectrochim. Acta B 58 (2003) 211–220. H.R. Kuhn, M. Guillong, D. Gunther, Size-related vaporisation and ionisation of laserinduced glass particles in the inductively coupled plasma, Anal. Bioanal. Chem. 378 (2004) 1069–1074. H.-R. Kuhn, D. Günther, Laser ablation-ICP-MS: particle size dependent elemental composition studies on filter-collected and online measured aerosols from glass, J. Anal. At. Spectrom. 19 (2004) 1158–1164.