Precise Pb isotope analysis of igneous rocks using fully-automated double spike thermal ionization mass spectrometry (FA -DS- TIMS)

Transcription

1 JAMSTEC-R IFREE Special Issue, November 2009 Precise Pb isotope analysis of igneous rocks using fully-automated double spike thermal ionization mass spectrometry (FA -DS- TIMS) Takashi Miyazaki 1*, Nobuyuki Kanazawa 2, Toshiro Takahashi 1, Yuka Hirahara 1, Bogdan Stefanov Vaglarov 1, Qing Chang 1 1, Yoshiyuki Tatsumi with a spike enriched The isotopic ratios of the of = ± , Keywords Takashi Miyazaki tmiyazaki@jamstec.go.jp 73

2 1. Introduction in understanding geodynamic processes such as material recycling in the Earth s interior, or in surface environments, thermal ionization mass spectrometry (TIMS) and multiple collector inductively coupled plasma mass spectrometry spectrometers has been a major obstacle preventing the 204 approaches have been developed to correct for this mass fractionation. In the case of TIMS, 1) a conventional fractionation correction using a mass fractionation factor determined by analyzing a standard, which is then applied to all unknowns, 2) a zero-time correction (Kuritani and Nakamura, 2002), and 3) double, triple or two-double spike Thirlwall, 2000; Kuritani and Nakamura, 2003) have been correction using Tl (e.g. Thirlwall, 2000; Tanimizu and Ishikawa, 2006), 2) a combination between double spike applied. These procedures have resulted in a dramatic particularly for samples containing relatively large amounts instrumental developments in TIMS have enabled reliable magazine to be performed. This cuts labor costs, although the ). However, the method is not applicable to analyze small amounts of sample (< 25 ng) (Miyazaki et al., 2003a). samples, we have developed a fully-automated analytical procedure of this method and report its results for Pb- 204 Pb DS-TIMS spike enriched in two isotopes (in this case in the mass spectrometer; the natural run refers to the isotope ratios are measured with a common denominator (i.e., either 206,207, ,207, ,206, ,206, isotope ratios of the spike, derive the mass discrimination ratios of the natural and the spiked runs. Several methods have been discussed in the literature to derive these values, Thirlwall, 2000; Kuritani and Nakamura, 2003). The latter was applied in this study. ( is illustrated in Fig. 1. The true isotopic compositions U and M, respectively, while points u and m correspond to the measured isotopic compositions of the unknown S corresponds to the isotope fu Measured (u) 206 Pb/ 204 Pb 208 Pb/ 204 Pb Unknown sample (U) fm Doped mixture (M) Measured (m) 207 Pb 204 Pb spike (S) 207 Pb/ 204 Pb Fig.1. Three dimensional graphical representation of the double spike, 74

3 T. Miyazaki et al., fractionation law, the true isotope ratios of the unknown (U i ) M i U i = u i (1 + i ) (1) M i = m i (1 + M i ) (2) where u i and m i are the measured isotope ratios of the and M are the fractionation U and M i is the mass difference between the numerator and denominator isotopes (subscript i between the U i and S i M i M i (1 + q) = U i + S i q (3) where q is the ratio between the reference isotope in the spike and the unknown ( 204 spike / 204 unknown (m i - S i ) q + m i i (1 + q) M - u i i = u i - m i (4) (5) where r = (1 + q A and vector B (6) (7), follows by A. The propagated error for the corrected isotope ratios of the unknown is calculated using a model from Hamelin et al. (1985). 3. Experimental 3.1. Reagents and isotope standard materials 3, 20 % HBr and 98 % H 3 4 (Tama Chemicals Co., Ltd) -1 by a Milli-Q system (Millipore silicic acid colloidal solution (Nissan Chemical Industries, and diluted with 0.5 M HNO 3 to give a concentration of DS calibration (1999). The with 0.5 M HNO 3 to 7 ppm, and calibrated using NIST SRM NIST SRM 982, used for this calibration and determined following the conventional method (Miyazaki et al. 2003a), were , and when normalized to composition of the , Sample preparation in a class 100 clean room. The powdered silicate rock ) with a 1 3 and 20 M HF on a hot plate at after which a further 1 ml of 8 M HBr was added. Once more achieved before 1 ml of 0.5 M HBr was added to digest the dried sample and it was centrifuged for 20 minutes at

4 The remaining precipitate was rinsed with 0.5 ml 0.5 M HBr and centrifuged, and the supernatant was then added to the mesh), which was previously washed and conditioned with 1 ml of 1 M HNO 3, 1 ml of H 2 O and 0.6 ml of 0.5 M HBr. 2 O after the composed of 0.25 M HBr and 0.5 M HNO 3. One drop of M H 3 4 which was previously washed and conditioned with 1 ml of 1 M HNO 3, 1 ml of H 2 fraction was collected with 2.5 ml of H 2 O after the elution of concomitant elements using 4 ml of 0.5 M HBr. The amount case of NIST SRM 981, a 10 ppm solution was split 10 ng was added to the 3 4 to the 7.5 Thermo Finnigan TRITON TI thermal ionization mass collectors at the Institute for Research on Earth Evolution with a amplifier) was assigned to the Low2 Faraday collector for the smallest 204 to each test and EL test). The advantage of using the it reduces the statistical error caused by a low ion beam intensity on a minor isotope, such as 204 in the signal to noise ratio by a factor of 3 (Schwieters et al., 2006). Fig. 2 compares the raw isotope ratios of and measured using a about 0.05V). For a given 204 ratios measured using the order to decompose organic materials. The total procedural routine analysis Mass spectrometry activator silica gel. The amount of ionization activator 15 ng). The loads were concentrated on the top of trapezoid- filament used in this study was supplied by Thermo Fisher rapidly decreased to zero. Fig.2. Comparison between raw isotope ratios of

5 T. Miyazaki et al., correlations than the those using the automatically by passing a constant electric current through them. The gain factor between the determined through repetitive measurements of NIST SRM 981 following conventional methods (Miyazaki et al. 2003a) Program for FA- DS- TIMS On the TRITON TI, operation conditions, such as gain calibration, baseline measurement, filament heating, focusing of the ion beam, and sample measurement, are performed during sample measurement runs. However, when there are only small amounts of sample, wasting sample result, we adopted a single long-term baseline measurement performed prior to the sample measurement runs. Table 1 shows the combination of four method files including the long baseline measurement. The baseline was measured for 15 minutes. The baseline data is stored and used for each successive sample measurement run. However, 60 minutes can elapse between the baseline measurement and the drift may occur. To monitor this possibility, one phantom blank sample measurement was performed at the very end of an ion source filament. This raw intensity data without the ion beam is used to perform the baseline correction for each sample by interpolating the baseline drift, if necessary. This Twenty-one filaments can be installed on one magazine set in the mass spectrometer. This allows ten samples, in addition to the standard, to be automatically Table 1. Combination of method files for fully-automated DS measurement (for one sample on filament). Method File Operation Acquired data Time 1 Gain calibration Gain data (stored in operating computer) 15 min 2 Baseline measurement Baseline data (stored in operating computer) 15 min 3-1 Filament heating ( C) and Focusing of the ion beam - < 60 min 3-2 Sample measurement Raw intensity data of sample (used for off-line calculation) 15 min 4 Blank measurement for baseline drift collection Raw intensity data without ion beam (used for off-line calculation) 15 min Table 2. Pb isotope ratios measured in NIST SRM ng #1 #2 #3 #4 #5 #6 #7 206 Pb / 204 Pb 2SE 207 Pb / 204 Pb 2SE 208 Pb / 204 Pb 2SE average ± 2SD RSD (%) ng #1 #2 #3 #4 # average ± 2SD RSD (%) ng and 7.5 ng average ± 2SD RSD (%) Previous studies (using DS or TS) 2SD 2SD 2SD Woodhead et al. (1995) TIMS 207 Pb Pb DS 100 ng Pb (n = 109) Todt et al. (1996) TIMS 202 Pb Pb DS 80 ng Pb (n = 11) Galer & Abouchami (1998) TIMS 207 Pb Pb Pb TS 10 ng Pb (n = 60) Thirlwall (2000) TIMS 207 Pb Pb DS 20 ng Pb (n = 41) Thirlwall (2002) MC-ICP-MS 207 Pb Pb DS 7 and 44 ng Pb (n = 36) Kuritani & Nakamura (2003) TIMS 207 Pb Pb DS 15 ng (n = 16) Baker et al. (2004) MC-ICP-MS 207 Pb Pb DS ng Pb (n = 95) Kuritani et al. (2006) TIMS 207 Pb Pb, ( 205 Pb Pb) Pb-ID-DS 40 ng Pb (n = 5) Makishima et al. (2007) MC-ICP-MS 207 Pb Pb DS 4 ng Pb SD = 2 standard deviation, RSD = relative standard deviation (1SD), 2SE = 2 standard error propagated through the double-spike mass bias correction Uncertainties shown in italics are the values re-calculated here from the RSD (%) in the previous studies. 77

6 Pb/ Pb Pb/ Pb Previous studies This study Fractionation line Pb/ Fig.3. Comparison between average values for NIST SRM 981 obtained and sources of previous studies are listed in Table 2. Reproducibility (%) Pb ( ng ) ); corr. TIMS (Kuritani and Nakamura, 2002). Pb Conventional TIMS (Miyazaki et al., 2007) DS TIMS (This study) ratios ( and spiked runs were used for offline calculations on an spreadsheet, a copy of which is available from the corresponding author. 4. Results and discussion 4.1. Measurements of NIST SRM 981 standard 15 ng) measured following this procedure at the IFREE correction using the ng of NIST SRM 981 are the average values for the 7.5 ng and 15 ng runs are within 1 standard deviation. Our results are compared with those correction TIMS Measurements of JB ± , (0.004 %, %, and %, respectively), they are since 2000 are compared with our data in Table 3 and Fig. 5. The in this study are consistent with those of the other studies, 78

7 T. Miyazaki et al., Table 3. Pb isotope ratios measured in JB ng #1 #2 #3 #4 #5 q 206 Pb / 204 Pb 2SE 207 Pb / 204 Pb 2SE 208 Pb / 204 Pb 2SE average ± 2SD RSD (%) Previous studies 2SD 2SD 2SD Ishizuka et al. (2003) (n = 5) Baker et al. (2004) (n = 14) Tanimizu and Ishikawa (2006) * (n = 7) * Corrected against NIST SRM 981 of Baker et al. (2004), 2SD = 2 standard deviation, RSD = relative standard deviation (1SD), 2SE = 2 standard error propagated through the double-spike mass bias correction. q is the ratio between the reference isotope in the spike and the unknown ( 204 Pb spike / 204 Pb unknown ) in the doped mixture. Pb/ Pb Previous studies This study within error. The analytical accuracy and reproducibility Acknowledgement Suzuki for constructive reviews and comments. Pb/ Pb previous studies are listed in Table 3. Pb/ Pb References isotopic analysis of standards and samples using a Chem. Geol., 211, standard, J. Res. Nat. Bur. Stand., 72A, internal standards for precise isotopic analysis by the order algebraic solutions, J. Sci. Inst., 40, for high precision lead isotopic measurement, Chem. Geol., 157, application of lead triple spiking for correction of instrumental mass discrimination, Mineral. Mag., 62A, Geochim. 79

8 Cosmochim. Acta, 49, Earth Planet Sci. Lett., 10, (2003), Fluid-mantle interaction in an intra-oceanic Earth Planet. Sci. Lett., 211, samples without use of double spikes, Chem. Geol., 186, Kuritani, T. and E. Nakamura (2003), Highly precise and using J. Anal. At. Spectrom., 18, Geostand. Geoanal. Res., 30, addition, J. Anal. At. Spectrom., 22, Miyazaki, T., T. Shibata and M. Yoshikawa (2003a), New synthesis method of silica-gel for lead isotope analysis, Proc. Japan Academy Ser. B, 79, Miyazaki T., M. Yoshikawa, T. Shibata and Y. Tatsumi to obtain high reproducibility in lead isotope measurements by thermal ionization mass spectrometer (TIMS), Frontier Research on Earth Evolution, 1, , Miyazaki, T., T. Shibata, M. Yoshikawa, T. Sakamoto, K. Iijima and Y. Tatsumi (2005), Evaluation of silicagel activator in order to find the optimal silica-gel activator for lead isotope measurement by thermal ionization mass spectrometer (TIMS), Frontier Research on Earth Evolution, 2, , 1-5. Miyazaki T., T. Takahashi, Y. Hirahara, Q. Chang, B.S. Vaglarov, K. Suzuki and Y. Tatsumi (2007), Low Sr analysis of minerals in volcanic rocks, Frontier Research on Earth Evolution, 3, , Improvements in high precision multicollector isotope ratio measurements for small sample sizes using Ohm current amplifiers and Faraday cup detectors, Geochim. Cosmochim. Acta, 70, A567. reference samples, Geochem. J., 40, Thirlwall, M. F. (2000), Inter-laboratory and other errors in double spike, Chem., Geol., 163, demonstrates up to 400 ppm/amu systematic errors in Tl-normalization, Chem. Geol., 184, Evaluation of a precision lead isotope analysis, In: A. Basu and S. Hart (Editors), Earth Processes: Reading the Isotopic Code, Geophys. Monogr., 95, Am. Geophys. Union, Washington, DC, Routine lead isotope determinations using a lead- analytical precision and accuracy, Analyst, 120,