Natural Clinopyroxene Reference Materials for in situ Sr Isotopic Analysis via LA-MC-ICP-MS

Clinopyroxene is a major host mineral for lithophile elements in the mantle lithosphere, and therefore, its origin is vital for constraints on mantle evolution and melt generation. In situ Sr isotopic measurement of clinopyroxene has been available since the recent development of laser ablation multicollector inductively coupled plasma mass spectrometry (LA-MC-ICP-MS) in the 2000s. Therefore, there is an increasing demand for natural clinopyroxene reference materials for Sr isotope microanalysis. In this contribution, we present six natural clinopyroxene reference materials from South Africa (JJG1424) and China (YY09-47, YY09-04, YY09-24, YY12-01, and YY12-02) for Sr isotope microanalysis. The Sr content of these clinopyroxenes ranges from 50 to 340 μg g−1, which covers most natural clinopyroxene compositions. Homogeneity of these potential reference materials were investigated and evaluated in detail over a 2-year period using 193-nm nanosecond and 257-nm femtosecond laser systems coupled to either a Neptune or Neptune Plus MC-ICP-MS. Additionally, the major and trace element of these clinopyroxenes were examined by electron probe microanalyzer (EPMA) as well as solution and laser ICP-MS. The in situ 87Sr/86Sr values obtained for the six natural clinopyroxene reference materials agree well with data obtained using the thermal ionization mass spectrometer (TIMS) method. The Sr isotopic stability and homogeneity of these clinopyroxenes make them potential reference materials for in situ Sr microanalysis to correct instrumental fractionation or as quality control materials for analytical sessions. The new Sr isotope data provided here might be beneficial for microbeam analysis in the geochemical community.

In situ Sr isotope measurement via ablation multicollector inductively coupled plasma mass spectrometry (LA-MC-ICP-MS) for clinopyroxene has been growing rapidly. For example, the first study demonstrated isotopic heterogeneity at the scale of individual grains in peridotite xenoliths and multiple measurements of the same grain; however, it indicated intragrain Sr isotopic disequilibrium (Schmidberger et al., 2003). Subsequently, Jackson and Hart (2006) reported in situ Sr isotopes in melt inclusions hosted by olivine phenocrysts and observed melts from high 3 He/ 4 He and EM II-type mantle end members, respectively. Sun et al. (2012) performed an in situ Sr isotopes for clinopyroxene in mantle xenoliths from Hebi, central North China Craton (NCC), and concluded that the clinopyroxene were crystallized from metasomatic melts. Xu et al. (2013) conducted in situ Sr isotopic composition of peridotite xenoliths from Kuandian and investigated Pacific slab subduction-related mantle modification of clinopyroxene beneath the eastern NCC.
According to the literature data from Georoc (Figure 1), the Sr content of natural clinopyroxene samples mainly distributed between 50 and 350 µg g −1 . Although a synthesized clinopyroxene glass with added Sr (CPX05G, ∼518 µg g −1 Sr) was developed as an in-house reference material, the content of Sr is higher than that of most natural clinopyroxene, and a limitation is its unavailability for other users (Tong et al., 2016). There is still lack of accessible clinopyroxene reference material for Sr isotope microanalysis. Despite a few published papers about in situ Sr analysis of clinopyroxene, the shortage of reference materials hinders the development of in situ Sr isotope measurements for clinopyroxene (Waight et al., 2002;Bizzarro et al., 2003;Schmidberger et al., 2003;Hart et al., 2005;Sun et al., 2012;Xu et al., 2013;Su et al., 2015;Tong et al., 2016;Deng et al., 2017;Tang et al., 2019).
Herein, we investigated whether the homogeneity of Sr isotopes for six natural clinopyroxene from South Africa (JJG1424) and China (YY09-47, YY09-04, YY09-24, YY12-01, and YY12-02), which covers a range of 50-350 µg g −1 , corresponds well to the natural distribution, at the micrometer scale, using MC-ICP-MS coupled with nano-and femtosecond laser over a 2-year period. Precise Sr isotope compositions of these samples were also determined using classic TIMS or solution MC-ICP-MS methods. Meanwhile, the major and trace element of these clinopyroxenes were examined by EPMA, ICP-MS based on solution, and laser sampling. Our work indicates that these six natural minerals might be employed as potential reference materials for in situ Sr isotope analysis.

ANALYTICAL METHODS
All six samples were prepared and mounted in epoxy resin blocks and polished to expose the interior of the crystals prior to analysis at the State Key Laboratory of Lithospheric Evolution (SKLLE), Institute of Geology and Geophysics, Chinese Academy of Sciences (IGGCAS). For comparison, one analytical session for in situ Sr isotope was performed at the State Key Laboratory of Geological Process and Mineral Resources (GPMR), China University of Geosciences (Wuhan), China.

Sample Description
Five of the samples were separated from lherzolite xenoliths entrained in the Yangyuan, in the Central Zone of the NCC (Zhao et al., 2015(Zhao et al., , 2017, and the other is from South Africa (Class and le Roex, 2011). According to the mineralogical and geochemical features of the clinopyroxene, they are Cr-diopside.

Major Element by EPMA
Major element composition of samples and backscattered electron (BSE) images were obtained from polished thin sections FIGURE 1 | Sr contents (µg g −1 ) distribution range over 10,000 published clinopyroxenes (based on data from Georoc).
using a JEOL-JXA8100 electron probe microanalyzer (EPMA) at IGGCAS. The operating conditions were as follows: 15 kV accelerating voltage, 12 nA beam current, 5 µm beam spot, and 10-30 s counting time on peak. Natural clinopyroxene and synthetic oxides were used for data correction, and the precision of all analyzed elements is better than 1.5%.

Trace Element Compositions Using Solution or Laser Ablation ICP-MS Analysis
Solution trace element contents in clinopyroxene were determined using a sector field (SF) ICP-MS (Finnigan MAT Element I) after digestion of about 40 mg of sample using a mixture of ultrapure 1 ml HF and 0.8 ml HNO 3 in Teflon bombs. After dissolution, the solution in the bomb is transferred into a polyethylene terephthalate (PET) bottle, which is weighed accurately to 50 g by addition of a 2% HNO 3 solution with 10 ng g −1 in internal standard addition. The carrier and makeup gas flows were optimized daily to obtain a sensitivity of 89 Y over 20 Mcps/µg g −1 while holding the ThO + /Th + ratio below 0.5%. Indium was used as an internal standard to correct for matrix effects and instrumental drift. A Chinese GSR-3 silicate reference material was measured to monitor the accuracy of the analytical procedure, and the results are in consistence with recommended values. The results are adopted as reference values for LA-ICP-MS analyses. According to the result of GSR-3, individual elemental precision is generally better than 5%.
An Agilent 7500a ICP-MS coupled with a 193-nm ArF excimer laser ablation system was employed to measure trace element. Helium was used as the carrier gas through the ablation cell and mixed with argon downstream of the ablation cell ( Table 1). Prior to measurement, the pulse/analog (P/A) factor of the detector was calibrated using a tuning solution. The carrier and makeup gas flows were daily optimized to obtain maximum signal intensity for 238 U + while keeping the ThO + /Th + ratio below 0.5%. All LA-ICP-MS determinations were conducted using time-resolved analysis in fast, peak jumping mode. Each spot analysis consisted of an ∼20-s background and 60-s sample data acquisition. The dwell time for each isotope was set at 6 ms for Rb, Sr, Ba, Nb, Ta, Zr, Hf, Pb, and rare earth element (REE) and 10 ms for 232 Th and 238 U. Trace element concentrations were calibrated against the NIST SRM 612 standard glass reference material with 43 Ca as the internal standard element and using USGS BCR-2G glass as a quality monitor. Data reduction, including concentration determinations, method detection limits, and internal uncertainties were obtained using the GLITTER laser ablation software (Achterbergh et al., 2001).

In situ Sr Isotopic Analysis by Laser Ablation MC-ICP-MS
Neptune or Neptune Plus MC-ICP-MS with either a 193-nm ArF excimer or 257-nm femtosecond laser was employed to measure Sr isotopes at IGGCAS, Beijing and GPMR, Wuhan (Yang et al., 2014(Yang et al., , 2019Zhang et al., 2018Zhang et al., , 2020. A spot size of 60-160 µm was used with a 6-8 Hz repetition rate and an energy density of ∼8 J cm −2 , depending on the Sr content of sample ( Table 1). The Sr isotopic data were collected by static multicollection  (Christensen et al., 1995;Bizzarro et al., 2003;Woodhead et al., 2005) Data processing package used For trace elements, Glitter software was used for isotopic and elemental fractionation, instrumental mass bias calibration and uncertainty propagation. For Sr isotope, an in house Microsoft Excel macro written in VBA (Visual Basic for Applications) was used for Sr isotope mass fraction correction, interference correction and uncertainty propagation.
mode, using X skimmer and Jet sample cone. Prior to laser measurement, the MC-ICP-MS was optimized using a standard solution to obtain maximum sensitivity. The integral process of data acquisition has one block of 200 cycles, and the integration time is 0.524 s per cycle. A typical data acquisition cycle consisted of a 30-s measurement of the Kr gas blank with the laser off, followed by 60 s of measurement with the laser on. In this work, YY09-47, YY09-24, and YY12-01 clinopyroxene samples were measured after every 10 unknown samples for external calibration (Yang et al., 2014(Yang et al., , 2019Lin et al., 2016;Zhang et al., 2018Zhang et al., , 2020. Data reduction was conducted offline and the potential isobaric interferences were accounted for in the following order: Kr + and Rb + . First, the interference of 84 Kr and 86 Kr on 84 Sr and 86 Sr, respectively, were removed using the 30-s Kr gas baseline measurement. The isobaric interference correction of 84 Kr and 86 Kr on 84 Sr and 86 Sr was carried out using the natural Kr isotopic ratios ( 83 Kr/ 84 Kr = 0.20175, 83 Kr/ 86 Kr = 0.66474; Christensen et al., 1995;Bizzarro et al., 2003). Second, the natural ratio of 85 Rb/ 87 Rb (2.5926) was used to correct for isobaric interference of 87 Rb on 87 Sr by the exponential law, assuming that Rb has the same mass discrimination behavior as Sr (Woodhead et al., 2005). It is observed that the obtained 87 Rb/ 87 Sr ratio is typically <0.001 during in situ clinopyroxene Sr analysis, indicating that the radiogenic 87 Sr contribution is negligible (Yang et al., 2011). Additionally, our previous work demonstrated that Ca argides and dimers had an insignificant influence on Sr isotope analysis using a Neptune MC-ICP-MS (Yang et al., 2011); this observation is also strongly supported by other studies (Ramos et al., 2004;Vroon et al., 2008). Therefore, interferences from Ca argides or dimers are not considered further in this work. Meanwhile, we also monitored the 167 Er 2+ , 171 Yb 2+ , and 173 Yb 2+ at masses 83.5, 85.5, and 86.5, indicating negligible interference of double-charged ion. Finally, the 87 Sr/ 86 Sr ratios were calculated and normalized from the interference-corrected 86 Sr/ 88 Sr ratio using the exponential law. The whole data reduction procedure was performed using an in-house Excel Visual Basic for Applications (VBA) macro program (Horstwood et al., 2008(Horstwood et al., , 2016Zhang et al., 2018;Zhang et al., 2020).

Solution Sr Isotope Measurement by Isotope Dilution Thermal Ionization Mass Spectrometer
Rb and Sr concentration and isotopic compositions of the clinopyroxenes were measured using a Thermo Scientific Triton Plus TIMS in IGGCAS. Clinopyroxenes from a number of samples were handpicked for radiogenic isotope analysis using the isotope dilution method detailed elsewhere (Li et al., 2012. About 50 mg of each sample was weighted into a 7ml round bottom Savillex TM Teflon screw-top capsule and 3.0 ml of a mixed acid added, composed of 29 M HF + 0.3 ml 14 M HNO 3 + 0.3 ml 11.8 M HClO 4 with the addition of a 87 Rb-84 Sr tracer. The samples were dissolved on a hotplate at 180 • C for 7 days. Each capsule was opened and evaporated to fume HClO 4 after cooling. The dissolved sample solution was then evaporated to dryness at ca. 120 • C. After that, the samples were redissolved once more in 1.0 ml of 6 M HCl and reheated to 180 • C for several hours to eliminate fluoride complexes. Finally, the vials were opened, and the resulting sample solution was evaporated to dryness and redissolved with 1.0 ml of 2.5 M HCl on a hot plate at 120 • C. Next, elements were separated on AG50W-X12 cation resin columns (Li et al., 2012. Rock reference materials BCR-2 and BHVO-2 from USGS were measured to monitor the accuracy of the analytical procedure, and our results [BCR-2: 46.2 µg g −1 Rb and 338 µg g −1 Sr, 0.705004 ± 0.000008 (2µ) of 87 Sr/ 86 Sr ratio; BHVO-2: 8.9 µg g −1 Rb and 378 µg g −1 Sr, 0.703419 ± 0.0000013 (2µ) of 87 Sr/ 86 Sr ratio] are almost identical to recommended values (Raczek et al., 2003;Yang et al., 2010Yang et al., , 2012.
In situ Sr Isotopic Measurement of Low Sr Content Clinopyroxene As mentioned above, it is still challenging when it comes to reliably determining Sr isotope compositions of samples with low Sr content of <100 µg g −1 (Sun et al., 2012;Tong et al., 2016). While performing Sr isotopic measurements, sensitivities vary due to instrumental conditions and parameters. Figure 5 shows variations in 87 Sr/ 86 Sr results for analyses, under identical instrumental condition, of NBS 987 solutions at variable 88 Sr signal intensity between 0.1 and 10 V. This suggests that, although the 87 Sr/ 86 Sr results deviate from the reference value as the signal drops, the longer integration time (1.049 s) can compensate and improve the precision. This improvement is especially pronounced when the intensity is lower than 0.5 V.
As summarized in Table 4, when the signal of the sample is lower than 1 V, particularly for samples with low Sr content (e.g., about 100 µg g −1 Sr). We utilized line scanning instead of single spot sampling in session 1 (120 µm spot size, nanosecond 193nm laser) and session 5 (60 µm spot size, femtosecond 257-nm laser). Comparing with results of constant single point ablation, the 88 Sr signal intensity lifted to 1 V, and the accuracy improved from 0.00051 (2SD, n = 20) to 0.00022 (2SD,n = 20). This reveals that we can get higher precision Sr isotopic data through nanosecond and femtosecond ablation by line scanning mode with moderate spot size even if the Sr content is <100 µg g −1 . Therefore, for low Sr content samples, raster rather than crater laser mode, combined with more integration time, is preferable and can promote the precision and accuracy of actual low clinopyroxene sample. Moreover, the ideal 88 Sr signal intensity is more than ca. 1 V for reliable 87 Sr/ 86 Sr data (Waight et al., 2002;Vroon et al., 2008;Jochum et al., 2009;Yang et al., 2014Yang et al., , 2019Tong et al., 2016;Zhang et al., 2018).
The Potential of Clinopyroxene for in situ Sr Isotopic Analysis As mentioned above, Figure 1 illustrates that over 65% (10,010/15,306) published clinopyroxenes (n = 15,306, Georoc) have more strontium than YY12-02 sample with 50 µg g −1 Sr content, more than 40% (6,077/15,306) clinopyroxene contain over 100 µg g −1 strontium, and over 90% (13,863/15,306) are <350 µg g −1 . Thus, clinopyroxene is a mineral that has low Sr content, and the suite of clinopyroxene reference materials have wide applicability and can cover the Sr content of the majority of natural clinopyroxenes. Low element contents and isobaric interferences precluded Sr isotopic analysis on some clinopyroxene samples via LA-MC-ICP-MS. Our previous study indicated that ∼500 µg g −1 Sr is enough to obtain an absolute precision of ± 0.0001 on the 87 Sr/ 86 Sr ratio when using a large laser spot size (Yang et al., 2009b(Yang et al., , 2014(Yang et al., , 2019. The extremely low Rb contents (and hence very low Rb/Sr ratios) of clinopyroxene mean that isobaric interference of 87 Rb on 87 Sr is usually negligible and can be easily accounted for (Yang et al., 2011). At more moderate to low Sr content of clinopyroxene (200 µg g −1 ), there is usually enough Sr for reasonable Sr isotopic analysis (± 0.0002). In our experiences, the Sr content of clinopyroxene determines the signal insensitivity and is the major factor for reliable Sr isotopic measurement by LA-MC-ICP-MS of this mineral phase Yang et al., 2014Yang et al., , 2019Zhang et al., 2018).

CONCLUSIONS
Considering the few, or unavailable, natural clinopyroxene reference materials for Sr microanalysis, we investigated thoroughly the assessment by both laser and solution measurements of the Sr isotopic ratios of six potential natural clinopyroxene reference materials from South Africa and China. The Sr isotopic compositions obtained for these samples FIGURE 5 | The relationship between the variation of 87 Sr/ 86 Sr ratios of NBS 987 and 88 Sr signal intensity (volt) with different integration times, which simulate for low Sr content clinopyroxene sample under the similar laser mode and investigate the effect on accuracy and precision of low signal intensity using a Faraday cup. The gray rectangle means the accepted value of 0.7103 ± 0.0002 (2SD). The error bar is smaller than the symbol labels. As indicated, crater laser mode time can yield reasonable Sr isotope with of ± 0.0002 on the 87 Sr/ 86 Sr ratio, for the moderate Sr contents of clinopyroxene (more than 200 ppm Sr). However, a raster laser mode together with more integration time is preferable and desirable over the crater mode because of the relatively low Sr content samples (ca. 100 ppm Sr). Moreover, the ideal 88 Sr signal intensity is more than ca. 1 V for reliable 87 Sr/ 86 Sr data. are all consistent with values obtained by solution methods [both MC-ICP-MS and isotope dilution TIMS (ID-TIMS)]. Moreover, the major and trace elements of these clinopyroxenes were also examined by EPMA as well as solution and laser ICP-MS. Due to the abundant supply of these natural samples and their homogeneous Sr isotopic compositions, these clinopyroxene samples (JJG1424, YY09-47, YY09-04, YY09-24, and YY12-01) might be potential reference materials for in situ LA-MC-ICP-MS Sr isotopic measurements, and YY12-02 is a potential material to monitor analysis quality for low-Sr samples. Our results demonstrate that these samples can be employed as reference material for in situ determination of Sr concentration and isotope composition using laser sampling. Based on our data, laser ablation can yield reasonable Sr isotope with 2σ precision of ± 0.0002 on the 87 Sr/ 86 Sr ratio for clinopyroxene with moderate Sr contents (more than 200 µg g −1 Sr). Moreover, our diverse investigation indicates that the raster laser mode is preferable over the crater mode when analyzing samples with relatively low Sr content (ca. 100 µg g −1 Sr). These reference materials are of sufficient amount and are available to the scientific community via contacting the corresponding author.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
HZ is a research scholar working on this study as part of his doctoral thesis, executed main experiments (major, trace, and Sr isotopes), analyzed and compiled the findings, and written this original manuscript preparation. X-MZ provided the five YangYuan samples and conducted trace element using solution ICP-MS and Sr isotope data by ID-TIMS. PL provided the JJG1424 sample and totally went through the draft. WZ is in charge of fs-LA-MC-ICP-MS at GPMR, Wuhan, conducted instrumental tuning, and supervised the fs-LA-MC-ICP MS work of HZ. L-WX, CH, and S-TW helped with the initial screening of CPX samples potentially suitable for LA-ICP-MS and ID-TIMS. HW, J-HY, and F-YW reviewed the draft and gave insightful comments. Y-HY supervised the whole study from planning to execution and result analysis. All authors contributed to the article and approved the submitted version.

FUNDING
This study was financially supported by the Natural Science Foundation of China (Grants 41525012 and 41973015).

ACKNOWLEDGMENTS
We thank Mao Q. and Zhang D. for their assistance with EPMA analysis. We are also grateful to the reviewers for their comments and suggestions to clarify our arguments.