Mass fractionation of Rb and Sr isotopes during laser ablation-multicollector-ICPMS: in situ observation and correction

One of the most critical issues concerning in situ mass spectrometry lies in accounting for elements and molecules that overlap target isotopes of analytical interest in a sample. This study traced the instrumental mass fractionation of Rb and Sr isotopes during laser ablation-multicollector-inductively coupled plasma mass spectrometry (LA-MC-ICPMS) to obtain reliable 87Sr/86Sr ratios for high-Rb/Sr samples. In the LA-MC-ICPMS analysis, Kr interferences were corrected using Ar and He gas blanks measured without ablating material. Contributions from doubly charged Er and Yb ions were corrected using the intensities of half masses and isotopic compositions reported in the literature. After Kr correction, the calculated 166Er2+ intensity of NIST SRM 610 approached the measured intensity at mass 83, and the 173Yb2+/171Yb2+ ratio agreed with the recommended value within error ranges. Kr- and REE2+-stripped peak intensities were further corrected for Rb interference. Use of the Sr mass bias factor for the calculation of measured 87Rb/85Rb yielded 87Sr/86Sr ratios consistent with the recommended and expected values for low-Rb/Sr materials, such as NIST SRM 616, modern shark teeth, and plagioclase collected from Jeju Island, but failed to account for the 87Rb interference from high-Rb/Sr materials including NIST SRM 610 and SRM 612. We calculated in situ mass bias factor of Rb from the known 87Sr/86Sr ratios of the standards and observed a correlation between Rb and Sr mass fractionation, which allowed inference of the Rb bias from the standard run. Reliable 87Sr/86Sr and 85Rb/86Sr ratios were obtained for SRM 610 and SRM 612 using the inferred mass bias factor of Rb calculated by the standard bracketing method. This study revealed that Rb and Sr isotopes behave differently during LA-MC-ICPMS and suggests the potential usefulness of the standard bracketing method for measuring the Rb–Sr isotopic compositions of high-Rb/Sr materials.


Introduction
The natural radioactivity of rubidium was demonstrated in 1906 by Campbell and Wood and came into use in the field of isotope geology when mass spectrometry analysis became available in the 1950s. Rubidium-87, which presently makes up 27.83% of Rb, decays to stable 87 Sr by emitting a negative beta particle with a decay constant of 1.42 × 10 −11 year −1 (Steiger and Jäger 1977) that corresponds to a half-life of 48.8 × 10 9 years. Strontium consists of four naturally occurring stable isotopes ( 84 Sr, 86 Sr, 87 Sr, and 88 Sr). Its isotopic composition, generally represented by the 87 Sr/ 86 Sr ratio, varies widely in the geosphere and hydrosphere resulting from contrasting geochemical behavior of Rb and Sr. The Sr isotope signatures of local ecosystems transferred through the food chain have been used extensively to address archeological and forensic questions since the 1980s (Ericson 1985;Aggarwal et al. 2008).
The Rb-Sr isotope data have long been obtained using isotope dilution thermal ionization mass spectrometry (TIMS) or multicollector-inductively coupled plasma mass spectrometry (MC-ICPMS) combined with the separation of Rb and Sr by cation exchange chromatography. Despite the constraint of time-consuming sample preparation, these wet techniques for bulk samples continue to be used widely, particularly when a high level of precision and accuracy is demanded. However, microsampling or in situ analysis is inevitably required to extract the time-series information stored in biogenic materials or to retrieve petrogenetic information from a mineral composed of geochemically heterogeneous micro-domains.
The potential of laser ablation (LA)-MC-ICPMS for in situ Sr isotope analysis was realized in the 1990s (Christensen et al. 1995). Although this technique has become a standard methodology for measuring the Sr isotopic composition of Sr-rich minerals (e.g., feldspar, apatite, and clinopyroxene), carbonate and phosphate rocks, and biogenic materials (e.g., teeth and otoliths), considerable debate remains over isobaric interferences and instrumental mass fractionation (IMF) of Rb and Sr isotopes (Ramos et al. 2004;Vroon et al. 2008;Müller and Anczkiewicz 2016;Willmes et al. 2016).
This study traced the IMF of Rb and Sr isotopes during laser ablation-multicollector-inductively coupled plasma mass spectrometry (LA-MC-ICPMS) and proposed an adequate procedure for correcting Rb interference on 87 Sr using synthesized standards and natural samples. The in situ data obtained in this study indicate the feasibility of the LA-MC-ICPMS technique for measuring the Rb-Sr isotopic composition of high-Rb/Sr materials.

Materials
This study used glass standard reference materials (SRMs) provided by the National Institute of Standards and Technology of the USA (NIST), modern shark teeth, and plagioclase crystals separated from a Quaternary basalt on Jeju Island, Korea.
The teeth of two shark species, the bull shark (Carcharhinus leucas) and spot-tail shark (Carcharhinus sorrah), were purchased during the Denver Gem and Mineral Show in 2018. The enameloid (external tooth layer) of the shark teeth consists typically of fluoroapatite (Ca 5 (PO 4 ) 3 F) (Kemp and Park 1974)

Methods
Solution and laser ablation-assisted Rb-Sr isotopic analyses, including sample preparation and mass spectrometry, were performed at the Korea Basic Science Institute Ochang Center. The instrument operating parameters of (LA)-MC-ICPMS are summarized in Table 1.

Solution ICPMS
Solution ICPMS was carried out for the shark teeth and Jeju plagioclase. Fragments (~40 mg) of tooth enameloid were digested overnight in 8 mL of 8 M HNO 3 at 160°C, then dried and stored in 10 mL of 2 M HNO 3 . The sample solutions were diluted 20,000 times with 2% HNO 3 . The concentrations of Rb and Sr in the teeth were measured using an X-series quadrupole ICPMS (Thermo Fisher Scientific). Plagioclase crystals (~10 mg) were digested overnight in 3 mL of mixed acid (HF:HNO 3 : HClO 4 = 10:3:1) at 160°C and then dried and redissolved in 2.5 N HCl. Chemical separation of Sr was performed twice using ion-exchange chromatography with Eichrom Sr-spec resin (100-150 μm) for the shark teeth and Bio-Rad AG50W-X8 resin (200-400 mesh) for the Jeju plagioclase. Sample digestion and Sr separation were conducted in a clean room equipped with laminar flow benches. The total Sr in the procedural blank was less than 200 pg.
Strontium isotopic compositions of the shark teeth and Jeju plagioclase were measured with a Neptune MC-ICPMS and a Nu Plasma II MC-ICPMS instrument, respectively. Strontium isotope data were acquired in static mode (n = 20). The IMF was exponentially normalized to 86 Sr/ 88 Sr = 0.1194. Isobaric interferences from Kr and Rb were negligible. Replicate analyses of NBS 987 with the Neptune and Nu Plasma II MC-ICPMS instruments yielded average 87 Sr/ 86 Sr ratios of 0.710283 ± 0.000007 [2 standard errors (SE), n = 6] and 0.710259 ± 0.000012 (2 SE, n = 18), respectively. Errors of isotopic ratios in this study denote 2 SE unless otherwise stated. 87 Sr/ 86 Sr data were corrected for interlaboratory bias by adjusting them to the recommended NBS987 value (0.710248; Thirlwall 1991).

Laser ablation MC-ICPMS
Fragments of the shark teeth were sonicated in distilled water for 15 min and then dried in an oven at 60°C. The NIST SRMs were washed with ethanol. The dried teeth and SRMs were mounted in epoxy and ground to expose the sample surface. Plagioclase phenocrysts in polished thin sections (0.1 mm thickness) were used for LA-MC-ICPMS analysis.
The Rb-Sr isotopic measurements were performed with a Nu Plasma II MC-ICPMS connected to an NWR193-nm ArF excimer laser ablation system. The LA-MC-ICPMS measurements consisted of five analytical sessions. SRM 610 was analyzed in all sessions but SRM 612 and SRM 616 were analyzed in session 1 only. The shark teeth were analyzed in sessions 1-4, and the Jeju plagioclase was measured in session 4. The ThO + / Th + (248/232) ratio was measured using path ablation mode before the beginning of each analytical session.
Three Faraday detectors (L1, H3, and H5) were placed at half mass positions to correct the isobaric interferences from doubly charged rare earth elements (REEs). The laser beam was focused to a spot diameter of~150 μm (sessions 1, 3, 4, 5) or~100 μm (session 2), and the intensities of mass 88 for SRM 610 ranged from ca. 5 to 1 V. Raw data (ca. 80-40% in the middle of the total 150 cycles) were processed using an Excel spreadsheet or Iolite 2.5 within Igor Pro 6.3.5.5 software (Paton et al. 2011).

Laser ablation MC-ICPMS
Isobaric interferences on Sr isotopes may arise from Ca dimers and Ca argides, Fe dioxides, Ga and Zn oxides, doubly charged REEs and Hf, and singly charged Kr and Rb (Ramos et al. 2004;Vroon et al. 2008). Among them, critical interferences from Kr, doubly charged Er and Yb ions, and Rb are discussed below. The other molecular species have only a limited impact on Sr isotope ratios (Ramos et al. 2004).

Gas blank and REE 2+ interference
Kr is usually included in Ar and He gas. The signals of Kr isotopes were monitored for 300 s without ablating material before the beginning of each analytical session. The observed 84 Kr intensities were 0.26 mV (session 1), 1.60 mV (session 2), 0.08 mV (session 3), 0.004 mV (session 4), and 5.4 mV (session 5  84, 85, 86, 87, and 88. Peak intensities at masses 83, 85.5, and 86.5 measured for SRM 610 were in the range of 10-3 mV when the 88 peak intensity was around 5 V (i.e., session 5). We confirmed that the Krstripped 166 Er 2+ intensity at mass 83 corresponded to the value calculated from the intensity at mass 83.5 ( 167 Er 2+ ) and the Er isotopic composition reported in the literature ( 166 Er, 33.5027%; 167 Er, 22.8691%; Chang et al. 1998). We also confirmed that the 173 Yb 2+ / 171 Yb 2+ ratio measured at masses 86.5 and 85.5 in each analytical session agreed with the recommended value in the literature (1.132338 ± 0.000050; Vervoort et al. 2004) within errors (for example, session 1 173 Yb 2+ / 171 Yb 2+ = 1.13 ± 0.02). These correspondences indicate that the Kr and REE 2+ intensities were measured correctly.
The contributions of Er and Yb to Rb and Sr isotopes were calculated from the measured intensities at masses 83.5 ( 167 Er 2+ ) and 86.5 ( 173 Yb 2+ ) (Eq. (1), "I" denotes a peak intensity) and the isotopic compositions and ratios of Er and Yb (see Table 2). The IMF was not considered in this calculation but was certainly insignificant at the level of REE 2+ in the samples and standards analyzed in this study.

Rb interference
Rubidium limits the types of samples that can be analyzed for Sr isotopic composition without chemical purification. Rb can be removed by "on-filament chemistry" in TIMS, which is not available in ICPMS. The mass resolution required for the separation of 87 Rb and 87 Sr masses (M/ΔM = ca. 296,000) cannot be achieved by a normal TIMS or ICPMS (M/ΔM = ca. 500).
The main problem of Rb interference correction arises from the IMF; the lighter and heavier isotopes of an element behave differently during the course of mass spectrometric measurement. Each element has its own mass bias factor. For Sr, the assumed ratio of 86 Sr/ 88 Sr (= 0.1194) is used for the correction of the IMF, but this internal normalization cannot be applied for Rb because this element consists of only two isotopes.
During ICPMS measurement, isotopic fractionation of two elements of similar mass should be similar, albeit not identical (White et al. 2000). Therefore, for measurement of the 87 Sr/ 86 Sr ratio for low-Rb/Sr materials such as biogenic apatite and plagioclase, it has been assumed that Rb and Sr isotopes behave in the same way during LA-ICPMS analysis (Ramos et al. 2004;Vroon et al. 2008;Willmes et al. 2016). After Kr and REE 2+ interference correction, the measured 87 Rb/ 85 Rb ratios of NIST SRM 616, Jeju plagioclase, and shark teeth were calculated using the IUPAC 87 Rb/ 85 Rb ratio (= 0.3856; Meija et al. 2016) and mass bias factor (β) of Sr calculated exponentially from measured 86 Sr/ 88 Sr and 0.1194. Then, the 87 Sr peak intensity was calculated by subtracting the 87 Rb contribution from the intensity of mass 87. As shown in Fig. 1 (Horstwood et al. 2008;Lewis et al. 2014;Müller and Anczkiewicz 2016;Willmes et al. 2016). The ThO + /Th + ratio is typically used as an indicator of oxide formation, because ThO + is the strongest oxide product within the plasma (Kent and Ungerer 2005). The ThO + /Th + ratio measured in each session varied from 0.77 to 0.16%. The enameloid of the shark teeth yielded average 87 Sr/ 86 Sr ratios of 0.70929 ± 0.00001 (session 1, n = 11), 0.70922 ± 0.00002 (session 2, n = 19), 0.70921 ± 0.00002 (session 3, n = 16), and 0.70922 ± 0.00002 (session 4, n = 21) (Fig. 2). These ratios approached the modern seawater value (0.70918; Faure and Mensing 2005) and solution MC-ICPMS results (0.709205 ± 0.000035, bull shark; 0.709170 ± 0.000025, spot-tail shark). It is noted that the LA-MC-ICPMS 87 Sr/ 86 Sr ratio was highest in session 1, in which ThO + /Th + was also highest (0.77%). However, small 87 Sr/ 86 Sr differences require additional studies to verify the reproducibility of this effect. In fact, it has been reported that polyatomic interference is negligible for samples with Sr contents higher than 1900 μg/g (Simonetti et al. 2008). The shark teeth analyzed in this study are sufficiently high in Sr (1 600 μg/g). The 85 Rb/ 86 Sr ratios of glass SRMs measured during laser ablation varied inconsistently (Fig. 3), manifesting that the IMF of Rb and Sr isotopes occurred differently. The employment of β (Sr) as β (Rb) yielded widely variable 87 Sr/ 86 Sr ratios for SRM 610 and 612. The averages for SRM 610 (0.7138 ± 0.0011, n = 81) and SRM 612 (0.7116 ± 0.0004, n = 15) deviated substantially from the recommended values.
We calculated β (Rb) using known 87 Sr/ 86 Sr ratios of standard materials as described in Eq. (2), where "m" and "c" denote the measured and corrected values, respectively, and "M" indicates atomic weight. The 87 Rb/ 85 Rb ratio of 0.3856 was taken from Meija et al. (2016).
In summary, Rb and Sr isotopes behaved differently during LA-MC-ICPMS but there was a correlation (albeit not strong) between the mass bias factors of these elements.

Bracketing calibration
The correlation between the mass bias factors of Rb and Sr led us to apply the standard bracketing method (Albarède and Beard 2004)

Conclusions
(1) After the Kr and REE 2+ correction, the 87 Sr/ 86 Sr ratios of low-Rb/Sr materials (NIST SRM 616, Jeju plagioclase, and shark teeth) were measured accurately by assuming that Rb and Sr isotopes behave in the same way during LA-MC-ICPMS. During measurements of the shark teeth, polyatomic isobaric interference on mass 87 was seemingly insignificant.
(2) Use of the Sr mass bias factor for the correction of Rb interference yielded highly variable and inaccurate 87 Sr/ 86 Sr ratios for high-Rb/Sr materials such as NIST SRM 610 and SRM 612. The mass bias factor of Rb was calculated from the recommended 87 Sr/ 86 Sr ratios of these SRMs. There were weak to moderate correlations between the mass bias factors of Rb and Sr. Reliable 87 Sr/ 86 Sr and 85 Rb/ 86 Sr results were obtained using the interpolated mass bias factor of Rb calculated by the standard bracketing method.
(3) This study showed that the Rb-Sr isotopic composition of high-Rb/Sr materials can be measured accurately using LA-MC-ICPMS by monitoring the IMF of Rb and Sr isotopes and applying the standard bracketing method.