Advanced Search

Indexed by SCI、CA、РЖ、PA、CSA、ZR、etc .

Volume 31 Issue 2
Apr.  2020
Turn off MathJax
Article Contents

Lüyun Zhu, Ganglan Zhang, Yongsheng Liu, Jie Lin, Xirun Tong, Shao-Yong Jiang. Improved in-situ Determination of Sr Isotope Ratio in Silicate Samples Using LA-MC-ICP-MS and Its Wider Application for Fused Rock Powder. Journal of Earth Science, 2020, 31(2): 262-270. doi: 10.1007/s12583-019-1214-0
Citation: Lüyun Zhu, Ganglan Zhang, Yongsheng Liu, Jie Lin, Xirun Tong, Shao-Yong Jiang. Improved in-situ Determination of Sr Isotope Ratio in Silicate Samples Using LA-MC-ICP-MS and Its Wider Application for Fused Rock Powder. Journal of Earth Science, 2020, 31(2): 262-270. doi: 10.1007/s12583-019-1214-0

Improved in-situ Determination of Sr Isotope Ratio in Silicate Samples Using LA-MC-ICP-MS and Its Wider Application for Fused Rock Powder

doi: 10.1007/s12583-019-1214-0
More Information
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Figures(6)  / Tables(3)

Article Metrics

Article views(130) PDF downloads(14) Cited by()

Related
Proportional views

Improved in-situ Determination of Sr Isotope Ratio in Silicate Samples Using LA-MC-ICP-MS and Its Wider Application for Fused Rock Powder

doi: 10.1007/s12583-019-1214-0
    Corresponding author: Lüyun Zhu

Abstract: The in-situ Sr isotope determination of solid samples has the advantages of high spatial resolution and convenient and rapid sample preparation. However, due to the lack of column chemistry separation of matrix elements from Sr, the interference correction (especially the isobaric interference from 87Rb on 87Sr) becomes a big issue for high precise in-situ Sr isotope analytical results. In this study, we evaluated the influence of isobaric interference systematically and adopted an improved strategy for isobaric interference correction of Rb. Specifically, as an external standard for data calibration, an in-house silicate standard (Rb-std) was fused from pure oxide powders. The new standard (Rb-std) shows very low contents of Sr (< 1 ppm), Yb (< 0.09 ppm) and Er (< 0.05 ppm), which directly avoid the trouble of interference of doubly charged ions and isobar, when we determinate the mass fractionation factors of Rb (βRb). Our improved method can provide more accurate data than previous researches, in which the mass fractionation were calculated from an external reference material StHs6/80-G with high Rb (30.07 ppm), Sr (482 ppm), Yb (1.13 ppm) and Er (1.18 ppm). Then, these βRb from Rb-std were applied by linear interpolation to the unknown samples for correction of 87Rb on 87Sr. The analytical results of serial international reference materials (BIR-1G, BCR-2G, BHVO-2G, T1-G and ATHO-G) demonstrate that our improved method can raise the upper limit of Rb/Sr for in-situ Sr isotope determination and reduce the relative standard deviation of results. This improvement of the upper limit of Rb/Sr (< 0.33) will expand not only range of microanalysis for silicate minerals, but also bulk Sr isotope determination of volcanic rock combining with rapid fusion technique.

Lüyun Zhu, Ganglan Zhang, Yongsheng Liu, Jie Lin, Xirun Tong, Shao-Yong Jiang. Improved in-situ Determination of Sr Isotope Ratio in Silicate Samples Using LA-MC-ICP-MS and Its Wider Application for Fused Rock Powder. Journal of Earth Science, 2020, 31(2): 262-270. doi: 10.1007/s12583-019-1214-0
Citation: Lüyun Zhu, Ganglan Zhang, Yongsheng Liu, Jie Lin, Xirun Tong, Shao-Yong Jiang. Improved in-situ Determination of Sr Isotope Ratio in Silicate Samples Using LA-MC-ICP-MS and Its Wider Application for Fused Rock Powder. Journal of Earth Science, 2020, 31(2): 262-270. doi: 10.1007/s12583-019-1214-0
  • The combination of laser ablation system (LA) and multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) instruments promote the development of in-situ determination of isotope ratios (Woodhead et al., 2005; Christensen et al., 1995). Due to the advantage of high precision and spatial resolution, this technique received wide attention from geological researchers quickly. Because the Rb/Sr ratio in the crust and mantle samples are significantly different, as the radiogenic daughter of 87Rb, the 87Sr in these samples would be different obviously. Therefore, the 87Sr/86Sr ratio has been an effective tool to trace origin of magma (Chen Y W et al., 2018; Liu et al., 2018; Chen H et al., 2017; Ju et al., 2017). Recently, the in-situ analysis technique of Rb-Sr isochron age has been used as an effective tool to date the magmatic event. For instance, Zack and Hogmalm (2016) have obtained in-situ Rb-Sr isochron ages of syenodiorite and pegmatite by LA-ICP-MS combining with a technique of online chemical separation of Rb and Sr in an oxygen-filled reaction cell. Generally, the technique of in-situ Sr isotope ratio determination has been applied to research different kinds of natural samples (e.g., mineral, rock and biological debris), in which Sr content is enough high for accurate analysis (Li et al., 2018; Xu et al., 2013; Vroon et al., 2008; Prohaska et al., 2002). As a pioneer, Christensen et al. (1995) first tried to determinate Sr isotope ratio in modern marine gastropods and plagioclase using LA-MC-ICP-MS. Then, Bizzarro et al. (2003) tried to measure the Sr isotopic homogeneity among igneous minerals (apatite and carbonate) which contain high Sr (> 3 000 ppm) and low Rb (< 1 ppm) contents. In general, these analyses are capable of producing measured 87Sr/86Sr values with enough high analytical precision based on enough high Sr intensity and low interferences on Sr masses. If we are to expend application of in-situ Sr isotope ratio analysis technique, there are two kinds of possible schemes for improvement: (1) increasing sensitivity of analysis for samples with lower Sr contents and (2) correcting interferences more effectively for samples with higher Rb/Sr ratios.

    In previous researches, following feasible ways have been adopted to improve sensitivity. When the signal intensities are close to ~20 μV, the Johnson noise from black-body radiation of the high-ohmic resistor (1 011 ohms) coupled to the Faraday collector will be the dominant source of signal (Konter and Storm, 2014). Therefore, some researchers tried to adopt a higher-ohmic resistor (1 012 or 1 013 ohms) to improve the signal stability of Faraday cup and decrease detection limit successfully for measurement of Sr, Nd, Ta isotope system (Pfeifer et al., 2017; Koornneef et al., 2014, 2013). The disadvantage brought by utilization of high-ohmic resistance is that the time-consume of analysis will increase. Moreover, another alternative approach is adopting ion counter to collect low intensity specially (Xie et al., 2017; Jochum et al., 2009), because it can multiply lower intensity signal by electron multiplication. For instance, Souders and Sylvester (2008) tried to collect all lead and uranium isotopes by ion counters, which could improve the spatial resolution and quantification limits of total Pb concentration in silicate glasses from > 80 ppm to < 15 ppm for measurement of LA-MC-ICP-MS. In a word, this kind of improvement are depended on upgrade and regeneration of scientific instruments, which require fund support.

    If we try to correct the isotopic interferences accurately during in-situ Sr isotope determination, it is necessary to distinguish the origin of these interferences firstly. These interferences consist of isobaric interferences by 87Rb, 86Kr and 84Kr on 87Sr, 86Sr, 84Sr, as well as doubly charged Er and Yb (168Er; 168Yb on 84Sr; 170Er; 170Yb on 85Rb needed for 87Rb correction; 172Yb on 86Sr; 174Yb on 87Sr; 176Yb on 88Sr). In addition, it is also possible that high temperature ICP will produce some dimers (44Ca43Ca and 44Ca40Ar) overlapping on 87Sr and 84Sr (Woodhead et al., 2005; Ramos et al., 2004; Waight et al., 2002). Indeed, isobaric interference from Kr is not an ignorable issue for MC-ICP-MS analyses, because Kr is usually present in small amount in the Ar gas, which is used for plasma generation and sample introduction. In special, the Kr contained in the gas background can be subtracted by baseline measurement for LA-MC-ICP-MS analysis (Yang et al., 2012; Woodhead et al., 2005; Ehrlich et al., 2001b). Particularly, the effects of doubly charged Er and Yb (168Er; 168Yb on 84Sr; 170Er; 170Yb on 85Rb; 172Yb on 86Sr; 174Yb on 87Sr; 176Yb on 88Sr) were estimated from real time intensity of 167Er, 173Yb (Waight et al., 2002). For 44Ca43Ca and 44Ca40Ar, the previous research has demonstrated that floated guard electrode is beneficial for suppressing these polyatomic ions (e.g., interference of Ca-related polyatomic ions were < 0.001% for samples with Ca/Sr≤500 and < 0.003% for samples with Ca/Sr=1 000) (Tong et al., 2016). Normally, Rb only occur two isotopes 87Rb and 85Rb, whose abundance are 27.84% and 72.17%, respectively. Taking the abundance of 87Sr (~7%) into account, the 87Rb isobaric interference will be an especially important issue for in-situ Sr isotope determination. For solution analysis, it usually reduces the Rb interference by careful column chemistry (Yuan et al., 2018). However, due to the lack of column chemistry separation, the in-situ Sr isotope determination only can be applicated for some minerals with high Sr but low Rb content (extremely low Rb/Sr ratio), which has been a major limit for wider application of this technique. Therefore, an effective correction for Rb interference is a key factor for this in-situ Sr isotope determination.

    At initial stage, it is common to correct Rb interference based on measured 85Rb and a fixed 87Rb/85Rb ratio (usually adopting natural constant 87Rb/85Rb ratio) (Ehrlich et al., 2001a). However, this approach ignored the mass fractionation (β is fractionation factor for calibration) during analysis. Then, some studies assumed that Rb has the same mass discrimination behavior as Sr(βRb)=βSr to correct interference of Rb (Ramos et al., 2004; Schmidberger et al., 2003). However, it recently found that elements with similar mass (e.g., Tl vs. Pb, Ir vs. Os) have different β values but constant β ratios (e.g., βTl/βPb and βIr/βOs)(Baker et al., 2004; Pearson et al., 2002). Hence, some studies tried to use synthetic glass or correlation between Sr and Rb constructed by solution experiments to obtain fractionation factor of Rb (Tong et al., 2016; Kimura et al., 2013; Davidson et al., 2001). For instance, (1) Tong et al. (2016) calculated a virtual 87Rb/85Rb ratio from an external reference material StHs6/80-G which composes high Rb (30.07 ppm), Sr (482 ppm), Yb (1.13 ppm), and Er (1.18 ppm). By this calculation, the shortage of different mass discrimination behavior between Rb and Sr would be covered. Nevertheless, the interference of doubly charged ions (Yb and Er) and isobar (87Rb and 87Sr) during analysis of this reference material will hinder accurate calculation of fractionation factor and increase resultant error. (2) Adopting solution experiment to obtain fractionation factor is a grueling process, which requires construction of a reasonable relationship between Rb and Sr at the beginning of each batch of sample analysis (Kimura et al., 2013). (3) According to LA-MC-ICP-MS results of solid reference materials (a modern coral and three USGS basaltic glasses: NKT-1G, TB-1G and BCR-2G with different Rb/Sr ratio), Zhang et al. (2018) successfully construct a reasonable relationship between raw 87Rb/86Sr ratio and the deviation of 87Sr/86Sr for Rb interference correction. Based on this relationship, precise and accurate in-situ Sr isotope ratio can be obtained if the basaltic glasses have Rb/Sr ratio less than 0.14. In brief, construction of correction relationship is a time-consuming process no matter which method is adopted. Moreover, the available time of this correction relationship is also unpredictable, which may be depended on state of instrument.

    Here, we developed another more convenient approach to raise the Rb/Sr upper limit for in-situ Sr analysis, which is also beneficial to extend application for more types of samples. At first, we specially developed an in-house silicate standard Rb-std that was fused from pure oxide powders without Sr, Yb and Er. Then, isobaric interference of 87Rb on 87Sr was corrected by applying standard-sample bracketing, whereby the βRb calculated from in-house standard Rb-std immediately before and after are applied by linear interpolation to the unknown samples. In general, free-Sr, Yb and Er in Rb-std avoid the trouble of interference of doubly charged ions and isobar, which is beneficial for more accurate correction. And the analytical results of serial international reference materials (BIR-1G, BCR-2G, BHVO-2G, T1-G and ATHO-G) demonstrate that our data strategy indeed raises the upper limit of Rb/Sr for in-situ Sr isotope determination and monitor the time drift of instrument more conveniently. If combining with the rapid fusion technique, our Improved analytical technique is also beneficial for determination of bulk Sr isotope ratio in volcanic rocks.

  • A ThermoFisher Neptune Plus MC-ICP-MS and the same 193 nm laser ablation system (Geolas 2005, Germany) were used for Sr isotope determination at the State Key Laboratory of Geological Processes and Mineral Resources (GPMR), China University of Geosciences (Wuhan). The Faraday collector configuration of our determination is composed an array from L4 to H3 to monitor Kr, Rb, Er, Yb and Sr mass respectively, which is suggested by Tong et al. (2016). All measurements were performed in low resolution and static mode. Helium was used as a sample gas due to its signal enhancement effect in LA-MC-ICP-MS analysis (Hu et al., 2008). Moreover, other details of the LA-MC-ICP-MS operational conditions are listed in Table 1. Before carrying routine analytical session, the electronic baseline and gain of Faraday collectors and amplifiers should be determined and calibrated. Each analysis composed a 30 s background of gas blank and 50 s data acquisition of sample signal. More details of LA-MC-ICP-MS operational conditions are listed in Table 1.

    MC-ICP-MS Neptune plus (Sr isotope ratio)
    Cup-configuration L4 L3 L2 L1 C H1 H2 H3
    83Kr 167Er2+ 84Sr 85Rb 86Sr 173Yb2+ 87Sr+87Rb 88Sr
    RF 1 100 W
    Acceleration valtage 10 kV
    Cool gas flow rate 16 L/min
    Auxiliary gas flow rate 0.75 L/min
    Carrier gas 0.7 L/min
    Cone Jet cone+X cone
    Integration time 0.524 s
    ICP-MS Agilent 7500a (chemical composition)
    RF 1 350 W
    Plasma gas flow 14 L/min
    Auxiliary gas flow 0.9 L/min
    Make-up gas flow 0.8 L/min
    Laser ablation system GeoLas 2005
    Wavelength 193 nm
    Energy density 5 J·cm-2
    Spot size and Repetition rate 120 μm and 10 Hz (LA-MC-ICP-MS);
    44 μm and 5 Hz (LA-ICP-MS)

    Table 1.  Routine operating condition of the MC-ICP-MS, ICP-MS and laser ablation system

    An Agilent 7500a ICP-MS and excimer 193 nm laser ablation system (Geolas 2005, Germany) were used for chemical composition determinations at the GPMR. And the analytical method has been reported by Liu et al. (2008). All measurements were performed in time-resolved analysis mode. More details of LA-ICP-MS operational conditions are listed in Table 1.

  • In order to reduce the instrumental background and memory effect (cross contamination caused by high Rb or Sr aerosols precipitation) after analysis of high Rb or Sr sample, a wash sample (Washer-1) was fused from pure oxide powders.

    According the composition of the natural clinopyroxene megacryst (HNB-8) (Tong et al., 2016), SiO2, MgO, Al2O3, CaO, FeO, TiO2 and Na2O were mixed in a proportion of 51.1%, 14.9%, 4.1%, 23.3%, 4.7%, 1.5%, 1.5% and 0.5%, respectively. After grinding, the mixture was placed in a Pt crucible and fused in a high-temperature furnace (Nabertherm, Germany) at 1 350 ℃ for 30 min. Then, the Pt crucible and melt was quenched together by cold ultrapure water. Because the cold shrinkage, the melt solidified quickly and disrupted into lots of glass grains, of which diameter were range from ~1 to ~10 mm. Then, several larger grains (~ > 8 mm) were chosen and mounted in resin. Following the same proportion, these oxide powders were mixed again, then spiked with Rb standard solution (1 000 ppm, GSBG62030, Central Iron & Steel Research Institute, China). After dried on an electric hot plate, the mixtures were fused along the same line to produce in-house standard Rb-std (high Rb and extremely low Sr) which is beneficial for determination of βRb. Furthermore, the following well-characterized reference materials were used in this research to verify the applicability of our analytical method: USGS glasses (basalt: BIR-1G, BHVO-2G and BCR-2G) and MPI-DING glassed (rhyolite: ATHO-G and quartz-diorite: T1-G). These reference materials have well defined Sr isotopic compositions and different Sr concentrations (from 94.1 ppm to 482 ppm) and Rb/Sr ratios (from 0.002 to 0.694). Furthermore, a natural clinopyroxene megacryst (HNB-8) was used for evaluation of the cross contamination during Sr isotopic determination. More details of these reference materials and samples are listed in Table 2.

    Samples Type Rb Sr Er Yb CaO TiO2 Rb/Sr 87Sr/86Sr
    (ppm) (wt.%)
    BIR-1G Basalt 0.2 109.0 1.7 1.6 13.3 1.0 0.002 0.703 105
    BHVO-2G Basalt 9.2 396.0 2.6 2.0 11.4 2.8 0.025 0.703 469
    BCR-2G Basalt 47.0 342.0 3.7 3.4 7.1 2.3 0.139 0.705 003
    ATHO-G Rhyolite 65.3 94.1 10.3 10.5 1.7 0.3 0.694 0.710 093
    T1-G Quartz-
    diorite
    79.7 284.0 2.5 2.4 7.1 0.8 0.281 0.703 271
    HNB-8 Natural Cpx
    megacrysts
    0.01 89.2 0.7 0.3 16.6 1.3 0.000 1 0.703 765
    Note: The reference values of USGS and MPI-DING glasses are from the GeoReM database (http://georem.mpch-mainz.gwdg.de). The values of HNB-8 from He et al. (2015).

    Table 2.  Details of the reference values of USGS, MPI-DING glasses and clinopyroxene megacrysts

  • Generally, there are different kinds of interference should be corrected during in-situ analysis of Sr isotopes. Firstly, the interference of Kr in working gas can be corrected by analytical background subtraction. Secondly, the interference of double charge ions (170Er2+ and 170Yb2+ on 85Rb, 172Yb2+ on 86Sr, 174Yb2+ on 87Sr) can be corrected based on the signal intensities of 167Er2+ and 173Yb2+ as following Eq. (1). Thirdly, the effect of Ca-related interferences (46Ca40Ar2+ and 46Ca42Ca2+) is limited for LA-MC-ICP-MS analysis in GE floated mode. By optimization of operation condition to limit ratio of oxidation products, the interference of Ga, Zn, Ca and Fe oxidation also can be controlled. Because lack of column chemistry separation, interference of 87Rb on 87Sr is an unneglectable problem for in-situ Sr isotope determination. Usually, the intensity of 87Rb was calculated from that of 85Rb. Previous researches demonstrate that the mass fractionation factor of Rb (βRb) is the key for this calculation. In this study, the factors (βRb) were measured by analyzing an in-house standard (Rb-std) before and after analyzing samples and then applied to samples via linear interpolation as following Eq. (2)

    where 85Total, 86Total, 87Total are measured signal intensity of Mass-85, Mass-86, Mass-87. 85Rb, 86Sr and 87Sr are the signal intensity of respective isotope after interference correction. R(167Er), R(170Er), R(170Yb), R(172Yb), R(173Yb) and R(174Yb) are the abundance of 167Er, 170Er, 170Yb, 172Yb, 173Yb and 174Yb, respectively (Wieser, 2006).

    where M is the mass of the isotope, βRb-sam and βRb-std are the mass fractionation coefficients of Rb for the sample and standard, respectively. 87Rb/85Rbref is the natural value of 87Rb/85Rb (0.385 756 from IUPAC)(Wieser, 2006), 87Rb/85Rbmea-std is the measured value of 87Rb/85Rb for Rb-std, 85Rbmea-sam and 87Totalmea-sam are the measured signal intensity of 85Rb and Mass-87, respectively, for the sample, and 87Srcor-sam is the signal intensity of 87Sr after interference correction for the sample.

  • It is widely accepted that the exponential law is suitable for the correction of Sr mass fractionation during LA-MC-ICP-MS analysis (Kimura et al., 2013). Considering the isotopic abundances, isobaric interferences of Sr and the collector configurations, the natural value of 87Sr/86Sr was used as the normalizing ratio to correct the mass fractionation of 87Sr/86Sr as following Eq. (3)

    where M is the mass of the isotope, 87Sr/86Srmea and 88Sr/86Srref are the measured ratio of 87Sr/86Sr and the natural value of 88Sr/86Sr (8.375 254 from IUPAC) (Wieser, 2006), respectively. Moreover, βSr is the mass fractionation factor of Sr during analysis. 88Sr/86Srcor is 87Sr/86Sr ratio after correction of mass fractionation.

  • In theory, the high Rb and Sr-free glass is especially suitable for obtaining the mass fractionation of Rb without considering interference of 87Sr on 87Rb. Then, Sr and Rb content of in-house reference glass Rb-std were measured by LA-ICP-MS (Fig. 1a). The analytical results demonstrated that the Rb-std glass contain ~186 ppm Rb and 0.94 ppm Sr. In the other words, the 87Sr/87Rb in this reference glass is ~0.001 32 which means the abundance of 87Sr in all Mass 87 signal is only ~1.32‰ (Fig. 1b). If considering the uncertainty of the measurement (average signal variation of Mass 87= ~4.81‰), this Mass 87 signal is approximated with 87Rb signal. Therefore, this fused Rb-std glass is suitable for obtaining real-time Rb mass fractionation factor.

    Figure 1.  (a) Variation of Rb and Sr content obtained for Rb-std by LA-ICP-MS, (b) uncertainty of 87Rb signal during analyzing glass Rb-std by LA-ICP-MS and interference of 87Sr on 87Rb which calculated from Rb and Sr content obtained for Rb-std by LA-ICP-MS.

    However, the high Rb sample or reference brings the risk of cross contamination between sample and reference. Previous researches have presented that mean particle diameters of nanosecond laser-induced aerosols are ~13 and ~16 nm under He atmosphere and fluence of 2.5 and 15 J·cm-2, respectively (Kuhn et al., 2004). During transportation of these aerosols, some of them precipitate in the ablation cell and inlet lines. Then, the precipitate aerosols contaminate the analysis of next sample. For instance, when directly analyzing HNB-8 after ablated Rb-std, these precipitated aerosols of last sample (Rb-std) may be carried by gas again which lead to the signal peak of Mass 87 at the beginning stage of laser ablation (Fig. 2a). If the precipitate aerosols were clean by ablating a washing sample (a fused glass with low Rb and Sr) before analyzing HNB-8, the signal of Mass 87 display an obviously lower peak at initial stage of laser ablation which means cross contamination generated by precipitated aerosols is weaken. Moreover, pre-ablation is also an effective method to reduce cross contamination between standard and sample (memory effect). However, the pre-ablation will increase the sample consumption, which will generate deeper ablation pit. The increase of ablation pit depth will aggravate isotopic fractionation. Therefore, we suggest that ablating a washing sample (about 40 s) between different samples or reference materials is beneficial for obtaining accurate analysis results.

    Figure 2.  Signal intensity variations of Rb and Sr isotopes obtain for Cpx crystal HNB-8 without and with washing precipitated aerosol.

  • For in-situ Sr isotope ratio determination, the appropriate calibration strategy plays a key role to obtain precise and accurate results. Therefore, a series of international reference glasses (BIR-1G, BHVO-2G, BCR-2G, T1-G and ATHO-G) were analyzed using LA-MC-ICP-MS to compare the effectiveness of our calibration strategy and that of previous researches (Tong et al., 2016) (Fig. 3, Table 3). Especially, the Rb/Sr ratios of these references range from 0.002 to 0.694, which are beneficial for evaluating different degrees of interference correction upon final results. As the Fig. 3 showed, no matter which data strategy was adopted for calibration, the 87Sr/86Sr ratio results of both BIR-1G and BHVO-2G agreed well with the reference values within an uncertainty of 5‰. Although the Rb/Sr ratio of BIR-1G is higher than that of BHVO-2G, single 87Sr/86Sr ratio analytical result of BIR-1G (σ= ~0.001 9) does not display lower error than that of BHVO-2G (σ= ~0.000 8). In theory, both the intensity of Sr and degree of interference correction of Rb on Sr will affect accuracy and precision of result. Therefore, the higher Sr content in BHVO-2G (Sr=396 ppm) than BIR-1G (Sr=109 ppm) is the main factor that causes the variation of data error. Therefore, for the sample with low Rb/Sr (< 0.025), the accuracy and precision of results were dominated by intensity of Sr during analysis. If the Rb/Sr ratio increases continually, the influence of interference correction on results will magnify. The 87Sr/86Sr ratio results of BCR-2G display similar level of single data error with that of BHVO-2G, but more deviation from reference value than that of BHVO-2G. This phenomenon can be explained by Sr content and Rb/Sr ratio in BCR-2G and BHVO-2G. The basaltic glass BCR-2G (Sr=342 ppm, Rb/Sr=0.139) has similar level of Sr content with BHVO-2G (Sr=396 ppm, Rb/Sr=0.025), but different level of Rb/Sr ratios. For results of BCR-2G and BHVO-2G, uncertainties from intensity of Sr are similar, but those from interference correction of Rb on Sr are different. In general, the more accurate and precise results of BCR-2G can be obtained when new data strategy is adopted for calibration, compared with the results of previous method suggested by Tong et al. (2016). As the Rb/Sr ratio (0.281 to 0.694) further increase, the accuracy and precision of 87Sr/86Sr ratio of T1-G and ATHO-G deteriorate quickly no matter which data strategy was adopted for calibration. A mounts of previous researches have demonstrated that the difference of matrix between reference material and analysis sample will lead the deviation of isotope ratio during LA-MC-ICP-MS (Zhang et al., 2013; Danyushevsky et al., 2011; Sylvester, 2008; Kroslakova and Günther, 2007). As the raise of Rb/Sr ratio, the accuracy of fractionation factor of Rb would become more important for accurate interference correction. Therefore, the matrix effect would be a key factor for deviation of results during analyzing T1-G and ATHO-G.

    Figure 3.  The measured 87Sr/86Sr ratio of international silicate reference materials (a) BIR-1G, (b) BCR-2G, (c) BHVO-2G, (d) T1-G and (e) ATHO-G.

  • As Fig. 4 shown, the analytical uncertainties display two trends of negative correlation over intensity of Sr, which demonstrated the influence of Sr intensity (Line A) and interference correction (Line B) on results. Generally, the interference correction may dominate the uncertainty when intensity is high enough (intensity of Sr > 0.2 V). The main difference between our data strategy and previous one is the way of obtaining Rb fractionation factor for interference correction. Therefore, all the original data of international reference materials are calibrated by these two strategies of interference correction for comparison (Fig. 5). As the increase of Rb/Sr, the resultant relative standard deviation by our strategy is obviously new Rb fractionation factor calculated from Rb-std (Sr=0.963 ppm, Rb/Sr=193.1) is closer to true value than that from extremal standard StHs6/80-G (Sr=30.7 ppm, Rb/Sr=0.063 7). Two factors may contribute to this improvement. Firstly, the Rb-std is a manufactured silicate glass from high-purity oxides, which is lower than that byprevious methods, which demonstrate that almost free of other interference elements (Er and Yb) on Sr and Rb isotope. As the fusion from volcanic rock powder (Sant Hellen Mountain), StHs6/80-G contain some Er and Yb unavoidably, which is not beneficial for obtaining pure Rb and Sr isotope signal. Secondly, because of high Rb and ultralow Sr content in Rb-std, it is likely to obtain more accurate and precise 87Rb/85Rb ratio without trouble of interference of 87Sr on 87Rb. In the other words, the Rb fractionation factor can be calculated from the in-house standard Rb-std just by normalizing measured 87Rb/85Rb ratio to natural ratio. Due to non-negligible interference of 87Sr on 87Rb, the calibration against StHs6/80-G is more complex. Before calculation of Rb fractionation factor, it is necessary to correct interference correction of 87Sr on 87Rb. Therefore, the error of additional process will be propagated to final result. Furthermore, no matter what method (solution or solid reference materials) was used to construct a relationship for Rb correction (Zhang et al., 2018; Kimura et al., 2013), it always was a time-consuming and grueling process. Because the lasting time of the constructed relationship is uncertain, it is probably best to constructed different and special relationship for time drift correction at the beginning of each batch of sample analysis. Comparing with these previous methods, Rb fractionation factor was obtain from more real-time analysis of Rb-std during each batch of sample measurement, which means more monitors for time-drift of instrument. According the polynomial equations of fitting curves in Fig. 5, our improved method can provide result with uncertainty less than 1‰, when the Rb/Sr ratio is less than 0.33. For the fused rock sample, the appropriate repetitive analysis of same sample will further raise accuracy and precision of results. In special, taking the time of repeat determination into account, our new method (e.g., BCR-2G: 87Sr/86Sr=0.705 01±0.000 14 (2SD, n=15)) also can provide accurate Sr isotope ratio of volcano rock by LA-MC-ICP-MS, which is even slightly better than results (e.g., BCR-2G: 87Sr/86Sr=0.705 08±0.000 10 (2SD, n=34)) from Zhang et al. (2018).

    Figure 4.  Influence of intensity of 88Sr on the resultant uncertainty of different international reference materials.

    Figure 5.  Rb/Sr ratio vs. relative error (RE) of in-situ Sr isotope ratio determination (RE=|average value–reference value|/reference value×100%).

    Results of improved method in this study Results of previous method suggested by Tong et al. (2016)
    Spot BIR-1G BHVO-2G BCR-2G T1-G ATHO-G BIR-1G BHVO-2G BCR-2G T1-G ATHO-G
    87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ* 87Sr/86Sr σ*
    1 0.703 03 20 0.703 57 8 0.705 03 15 0.709 53 24 0.698 72 330 0.702 98 21 0.703 52 8 0.703 86 16 0.707 52 27 0.691 56 337
    2 0.703 17 24 0.703 66 9 0.704 91 14 0.709 81 31 0.701 38 339 0.703 10 25 0.703 56 9 0.703 88 15 0.707 78 33 0.687 81 353
    3 0.703 21 24 0.703 72 8 0.705 02 14 0.709 49 28 0.699 52 326 0.703 17 24 0.703 64 8 0.704 48 15 0.707 84 29 0.688 82 336
    4 0.703 21 22 0.703 50 9 0.705 08 14 0.709 58 28 0.700 08 321 0.703 18 23 0.703 42 10 0.704 04 15 0.707 85 30 0.689 30 337
    5 0.703 14 18 0.703 64 9 0.705 04 15 0.709 44 28 0.699 61 336 0.703 09 19 0.703 56 9 0.704 05 16 0.707 84 30 0.688 99 339
    6 0.703 13 21 0.703 43 8 0.705 14 15 0.709 00 29 0.700 81 320 0.703 08 23 0.703 31 8 0.704 36 16 0.707 10 32 0.694 08 336
    7 0.703 26 21 0.703 59 7 0.705 00 15 0.709 32 36 0.699 71 340 0.703 25 23 0.703 52 7 0.704 02 15 0.708 05 40 0.691 18 374
    8 0.703 57 20 0.703 56 7 0.704 98 15 0.709 84 28 0.698 18 293 0.703 53 21 0.703 45 7 0.704 03 16 0.708 14 30 0.681 27 296
    9 0.703 01 18 0.703 64 8 0.704 87 13 0.709 31 29 0.698 90 305 0.702 94 18 0.703 53 8 0.703 97 14 0.707 66 29 0.681 02 336
    10 0.703 41 22 0.703 71 7 0.705 05 12 0.709 39 28 0.701 08 195 0.703 40 23 0.703 59 8 0.704 07 13 0.707 49 29 0.689 57 209
    11 0.702 97 24 0.703 68 8 0.705 10 16 0.709 22 32 0.700 14 221 0.702 94 24 0.703 51 8 0.704 20 18 0.707 07 35 0.687 85 237
    12 0.703 22 22 0.703 66 8 0.705 03 14 0.709 32 29 0.701 13 222 0.703 15 22 0.703 53 8 0.704 20 15 0.707 08 31 0.688 04 236
    13 0.703 10 21 0.703 62 9 0.705 02 13 0.709 74 23 0.701 01 212 0.703 04 22 0.703 52 9 0.704 10 13 0.707 55 24 0.693 12 231
    14 0.703 40 19 0.703 57 9 0.704 99 17 0.709 59 28 0.701 62 203 0.703 36 21 0.703 42 9 0.703 88 17 0.708 12 28 0.692 83 209
    15 0.702 82 20 0.703 62 9 0.704 94 17 0.709 77 27 0.702 12 135 0.702 78 21 0.703 52 10 0.703 83 18 0.707 93 28 0.691 63 142
    Ave. 0.703 18 19 0.703 61 8 0.705 01 7 0.709 49 24 0.700 27 116 0.703 13 19 0.703 51 8 0.704 07 19 0.707 67 36 0.689 14 381
    Ref. value (TIMS) - - - - - - - - - - 0.703 11 6 0.703 47 7 0.705 00 4 0.710 14 4 0.703 27 15
    The reference values of USGS and MPI-DING glasses are from the GeoReM database (http://georem.mpch-mainz.gwdg.de), -. no value, *. 103 times of their true values.

    Table 3.  Strontium isotope ratios of the reference glasses determined by different methods by LA-MC-ICP-MS

  • With development of rapid fusion technology of bulk silicate rock analysis, LA-MC-ICP-MS analysis can not only distinguish the variation of micro-region in silicate minerals, but also determinate chemical composition of the silicate rock powder (He et al., 2015; Zhu et al., 2013; Nehring et al., 2008; Stoll et al., 2008). For instance, some powder reference materials of basalt (BCR-2 and BHVO-2), andesite (AGV-2) and rhyolite (RGM-1 and RGM-2) were fused in a hermetic vessel by a special oven for bulk analysis (Zhu et al., 2013). Benefited from protection of the hermetic vessel, the violate elements (Zn, Pb) also can be kept in fusion glass for analysis. Then, this preparation was used to measure Pb isotope ratio of volcanic rock powder by LA-MC-ICP-MS (Bao et al., 2016). Because most of Sr and Rb can be kept during rapid fusion (Zhu et al., 2013), LA-MC-ICP-MS would be an effective tool to measure Sr isotope ratio in whole rock. Furthermore, the homogeneity of fused silicate glass is another important premise of accurate and precise result. Therefore, this rapid fusion technique is suitable for volcanic rock with enough low viscosity, which was mostly generate under the tectonic setting of Arc Island and Ocean Island. As Fig. 6 shown, there are a huge number of volcano meeting requirement of Rb/Sr (Rb/Sr < 0.33) for in-situ Sr isotope analysis. Due to raise of the upper limit of Rb/Sr ratio by our improve method, there are more samples can be measured in-situ Sr ratio directly by LA-MC-ICP-MS. Obviously, the analytical precision of this method in the whole rock Sr isotope analysis using LA-MC-ICP-MS is lower than traditional solution Sr isotope analysis using MC-ICP-MS and TIMS (Deng et al., 2017; Müller and Anczkiewicz, 2015; Montgomery et al., 2010; Harlou et al., 2009; Waight et al., 2002). However, the advantages of this method are less time consumption and simpler sample preparation, which is beneficial for quick identification and statistics of Sr isotope ratio in volcanic rock (e.g., huge numbers of basalt in island arc). Because the accuracy and precision required by different geological problems are different, this method should be more suitable for distinguish geological event with obvious variation of Sr isotope, such as the process of mantle and crust reaction. The Rb/Sr ratio in crust material is usually higher than that in mantle material. Therefore, raise of the upper limit of Rb/Sr ratio is also beneficial for determinate the different Sr isotope ratio in core and rim of mineral which is formed during mantle and crust reaction.

    Figure 6.  Variation of Rb/Sr ratio and Sr content in different geochemical reservoir, Ocean Islands and Arc Islands (used data of Ocean Islands basalt sample and Arc Islands lava are from geological database GEOROC: http://georoc.mpch-mainz.gwdg.de/georoc and RidgePetDB: www.earthchem.org/petdb).

  • Our work demonstrated that the interference correction plays a key role on in-situ determination of Sr isotopes by LA-MC-ICP-MS. In this study, an in-house silicate standard Rb-std was fused from pure oxide powders for interference correction specially. In general, isobaric interference of 87Rb on 87Sr was corrected by applying the standard-sample bracketing technique, whereby the βRb calculated from in-house standard Rb-std immediately before and after are applied by linear interpolation to the unknown sample. The lack of Sr, Yb, and Er in Rb-std avoid the trouble of interference of doubly charged ions and isobar. In addition, βRb obtained by a real-time way is beneficial for monitor the time-drift of instrument during analysis. Then, the analytical results of serial international reference materials (BIR-1G, BCR-2G, BHVO-2G, T1-G and ATHO-G) demonstrate that our data strategy indeed raise the upper limit of Rb/Sr for in-situ Sr isotope determination and reduce the relative standard deviation of result. Especially, the improvement of the upper limit of Rb/Sr will expand application of not only microanalysis in silicate minerals, but also bulk Sr isotope determination of volcano rock combining rapid fusion technique.

  • Two anonymous reviewers are thanked for the detailed comments that helped us to improve the manuscript. We highly appreciate the helpful suggestions of Dr. Wei Zhang regarding an earlier version of this manuscript. This research is co-supported by the National Natural Science Foundation of China (Nos. 41530211, 41125013, 41803012), the China Postdoctoral Science Foundation (No. 2017M622546) and the Most Special Funds of the State Key Laboratory of Geological Processes and Mineral Resources (No. MSFGPMR01). The final publication is available at Springer via https://doi.org/10.1007/s12583-019-1214-0.

Reference (49)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return