ARTICLE
Determination of whole-rock trace-element compositions of siliceous rocks using MgO-diluted fused glass and LA–ICP–MS
2022 年 56 巻 6 号 p. 231-239
詳細
2022 年 56 巻 6 号 p. 231-239
The whole-rock trace-element compositions of igneous rocks provide primary geochemical information about their petrogenesis. Such compositions can be determined by LA–ICP–MS analysis of fused-glass samples as well as the conventional solution ICP–MS method. However, in contrast to basalt and andesite, the fused-glass method is not suitable for Si-rich rocks (granite and rhyolite) due to the difficulty of making homogeneous glasses. To extend the method to Si-rich rocks, we adopted an MgO dilution process to prepare fused glasses and evaluated the technique by analysis of granite and rhyolite reference materials. Dilution of 30 mg of powdered samples with 10 mg MgO facilitated the preparation of homogeneous fused-glass samples of felsic rocks. LA–ICP–MS analyses indicate that the fused glass is homogeneous in the contents of most incompatible elements considered in igneous petrology. The glass enables the analysis of these elements in felsic rocks within 20% deviation from reference data. The sample data are consistent with reference values for rhyolite although Zr and Hf data deviate significantly (by ~40%) from reference values for granitic rocks.
Knowledge of whole-rock compositions is essential for describing the geochemical features of igneous rocks. Trace-element data are useful in elucidating genetic processes, and major-element compositions provide primary geochemical data for the understanding of igneous rock affinities and magma differentiation. Trace-element compositions are conventionally determined by inductively coupled plasma–mass spectroscopy (ICP–MS) analysis of solutions prepared by acid digestion of rocks. Fused glasses have been prepared before digestion to avoid solution residues and to simplify the digestion procedure (Awaji et al., 2006; Shimizu et al., 2011; Senda et al., 2014; Akizawa et al., 2020).
Laser ablation (LA)–ICP–MS is widely applied in spot analyses of solid samples (e.g., Sylvester, 2001a, 2008) and can determine whole-rock compositions by analysis of pressed powder pellets, alkali-flux fusion glass, and direct fusion glass (e.g., Fedorowich et al., 1993; Sylvester, 2001b; Eggins, 2003; Orihashi and Hirata, 2003; Kurosawa et al., 2006; Nehring et al., 2008; Stoll et al., 2008; Ito et al., 2009; Zhu et al., 2013; Garbe-Schönberg and Müller, 2014; Kon and Hirata, 2015; Tamura et al., 2015; Wu et al., 2018; Rospabé et al., 2022). In our laboratory, analysis of direct-fusion glass by LA–ICP–MS and of conventional alkali-flux fused glass by X-ray fluorescence spectroscopy are applied in determining whole-rock trace-element and major-element compositions, respectively (Kusano et al., 2014; Tamura et al., 2015; Arai et al., 2018; Morishita et al., 2020). Although some trace-element data (e.g., for Pb and U) have high uncertainty in fused-glass LA–ICP–MS analyses because of volatilization during fusion, contents of many trace elements useful to igneous petrology can be determined for basalt and andesite samples (Tamura et al., 2015). However, it is widely known that direct-fusion glasses prepared from high-Si rocks (rhyolitic and granitic rocks; SiO2 ≈ 70 wt%) are inhomogeneous due to their high melt viscosity, and it is therefore difficult to obtain reliable whole-rock data (e.g., Nehring et al., 2008).
Pressed pellets of sub-nano-sized powder have proved reliable target materials for LA–ICP–MS in the analysis of granitic rocks (Garbe-Schönberg and Müller, 2014; Wu et al., 2018). Alkali-flux fusion with lithium tetraborate is a more popular method for the preparation of fused glasses for analysis by LA–ICP–MS and can be used for both basaltic and andesitic rocks, and also for siliceous rocks (e.g., Eggins, 2003; Kon and Hirata, 2015). The low fusion temperature (1000°C–1200°C) is advantageous in reducing element volatilization. In contrast, higher temperatures (1600°C–1800°C) and long fusion times (60 s to 2 h) are required for the production of homogeneous glasses of Si-rich rocks by direct fusion (Nicholls, 1974; Nehring et al., 2008; Wu et al., 2018). To obtain homogeneous fused glasses from such siliceous samples, Nehring et al. (2008) attempted an MgO-dilution method involving degassing and milling of sample–MgO mixtures. High temperatures (>1200°C) are still necessary with short fusion times (30 s), so volatilization cannot be avoided. However, unlike the alkali-flux fused glass, direct-fusion glass prepared with or without MgO dilution allows the determination of Li and many key trace elements for petrological study (e.g., Nehring et al., 2008; Tamura et al., 2015; Wu et al., 2018).
In contrast to several new techniques requiring extra equipment to improve data quality, the fused-glass LA–ICP–MS method is rapid and simple, and MgO-diluted fused glass has the potential to rapidly provide reliable trace-element data for siliceous rock samples that are of sufficient quality for petrological analyses (Nehring et al., 2008). In this study, whole rock trace-element compositions of high-Si reference materials were determined using the direct-fusion method and LA–ICP–MS. A simplified procedure for the preparation of MgO-diluted fused glass was developed for use with conventional fused-glass LA–ICP–MS methods. Improvements and limitations of the technique were evaluated.
Whole-rock fused-glass samples were prepared by a direct-fusion method using an iridium-strip heater (Nicholls, 1974; Nehring et al., 2008; Stoll et al., 2008; Tamura et al., 2015). We used heater comprising house-made copper electrodes connected to a direct-current power supply (PK6-130 Matsusada Precision, Japan) and an iridium strip (30–50 mm × 5 mm × 0.05 mm; 99.85% Ir foil; Nilaco Corporation, Japan) (Tamura et al., 2015). Because of the strip size between electrodes and the hotspot heating area, powdered rock samples of ~20 mg were fused for 30–60 s.
To test the method, we used high-Si rock samples (SiO2 >67 wt%), including Geological Survey of Japan (GSJ) powdered reference materials JG-1a and JG-3 (granodiorite), JG-2 (granite), and JR-1 (rhyolite) (Imai et al., 1995). MgO powder (MgO >98%; 122-00281 Wako Chemicals, Japan) was used for dilution. Siliceous rock powder was mixed with MgO in a 3:1 ratio (30:10 mg). The two powders were thoroughly mixed in parafine paper before fusion. Fused-glass beads of 2–3 mm diameter were placed on glass slides and polished to thick sections (≈1 mm). Examples of fused glasses with and without MgO dilution are shown in Fig. 1. The major-element compositions of the siliceous rock samples (Imai et al., 1995) and calculated MgO-diluted compositions are listed in Table S1. Fused glasses of basalt and andesite (JA-1, JA-2, JB-2 and BHVO-2 provided by GSJ and USGS) were also prepared without MgO dilution.
Examples of polished fused glasses in thick section. Images are in transmitted light, except for (c), which is shown with reflected light. (a, b) Fused glasses prepared from JG-2 and JG-3 without MgO dilution (FG), showing opaque dendritic inclusions (Iridium metal). Note the different bubble sizes in (a) and the heterogeneity in color in (b). (c, d) MgO-diluted fused glasses (MFGs) prepared from JG-2 and JG-3. All MgO-diluted fused glasses have similar features of color and shape. Note the homogeneous appearance of the polished surface in (c). Small bubbles (black dots) were infrequently observed (d). Red points indicate laser ablation pits.
LA–ICP–MS analyses involved a NewWave Research UP-213 Nd:YAG deep UV (213 nm) LA system (New Wave Research, Australia) coupled to an Agilent 7850 (S-lens option) quadrupole ICP–MS or Agilent 7500s ICP–MS (Agilent Technologies, Japan) at Kanazawa University (Kanazawa, Japan) (Table 1). As the Agilent 7850 was newly installed in our laboratory, replacing Agilent 7500s systems, this was used to analyze the newly prepared fused-glass samples together with glasses analyzed previously using an Agilent 7500s.
| ICP-MS* | |
| Model | Agilent 7850 (S-lens option) and Agilent 7500s (Agilent Technologies) |
| Forward power | 1200 W |
| Plasma gas flow | 15 L min–1 |
| Carrier gas flow | 1.10 L min–1 (Ar), 0.3 L min–1 (He) |
| Interface | Ni sampler/Ni skimmer |
| Laser | |
| Model | UP-213 (New Wave Research) |
| Wave length | 213 nm (Nd-YAG) |
| Spot size | 100 μm |
| Repetition rate | 5 Hz |
| Energy density at target | 7 J cm–2 (Attenuater: 50–60%) |
| Warming up | 10 sec |
* Agilent 7500s was replaced by Agilent 7850 (see text).
LA operating conditions included a spot size of 100 μm diameter, ablation frequency of 5 Hz, and an energy fluence of 7 J cm–2 (attenuator ~55%); with laser warm-up for 10 s, ablation for 50 s, and background acquisition for 50 s. The carrier gas comprised Ar (1.1 L min–1) and He (0.3 mL min–1). ICP–MS sensitivity and operating conditions were optimized by monitoring 7Li, 89Y, and 232Th during LA of a US National Institute of Standards and Technology (NIST) 612 glass. The oxide production rate was maintained at <0.5% by monitoring the 248ThO/232Th ratio. A total of 36 elements (39 isotopes) were analyzed for petrological use including 7Li, 11B, 29Si, 42Ca, 43Ca, 45Sc, 49Ti, 51V, 52Cr, 53Cr, 59Co, 60Ni, 62Ni, 85Rb, 88Sr, 89Y, 90Zr, 93Nb, 133Cs, 137Ba, 139La, 140Ce, 141Pr, 143Nd, 147Sm, 151Eu, 157Gd, 159Tb 163Dy, 165Ho, 166Er, 169Tm, 172Yb, 175Lu, 178Hf, 181Ta, 208Pb, 232Th, and 238U. The time-resolved mode was used with a dwell time of 20 ms except for rare-earth elements (REEs; 40 ms) and 29Si and 42Ca (10 ms). The sweep time was 1.069 s. Each analysis required a total of 125 s.
The US Geological Survey BCR-2G glass reference material was used as a calibration standard, with its reference value obtained from the GeoReM database (Jochum and Nohl, 2008). 29Si was used as internal standard with reported reference values (Longerich et al., 1996). Quality control of the LA–ICP–MS system involved analysis of BHVO-2G, BIR-1G (USGS), TIT-200 (Ødegård et al., 2005), and NIST SRM glasses 612 and 614, as reported in Table S2 and Fig. S1.
The non-diluted fused-glass siliceous rock samples varied in color and included dendritic opaque material (Fig. 1a, b). Bubbles of various sizes often formed in JG-1a and JG-2 fused glasses (Fig. 1a), whereas MgO-diluted fused glasses appeared homogenous with no visible opaque inclusions, although small bubbles (<50 μm diameter) appeared infrequently (Fig. 1c, d). The polished surface conditions of MgO-diluted fused glass (Fig. 1c) were similar to those of fused glasses prepared from basalt and andesite.
The fused-glass method was evaluated by assessment of sample homogeneity based on the precision of multi-spot analyses and comparison with reference values (Table 2; GSJ, Imai et al., 1995; GeoReM database, Jochum and Nohl, 2008). Results indicate that non-diluted fused glasses of Si-rich rock samples (JG-2 and JG-3) were considerably heterogeneous in composition (Figs. 2a and S1). Such heterogeneity causes uncertainties in whole-rock spot analyses, although averages of many spots may reflect approximate whole-rock compositions (Fig. 2b). The relative standard deviation (RSD) of ~10% for 6–12 analytical spots indicates that MgO-diluted fused glasses were compositionally homogeneous (Figs. 2a and 3a). Many elements of importance to igneous petrology, such as incompatible elements, were analyzed with <10% deviation from the reference values (Figs. 2b and 3b). The Li contents of Si-rich samples (20–80 μg g–1) were determined for MgO-diluted fused glasses with a precision of <5% RSD, matching that of Li analyses of non-diluted fused-glass basaltic and andesitic samples (4–10 μg g–1). Tamura et al. (2015) reported that Pb and U may be depleted in fused-glass samples owing to volatilization during fusion, although no U depletion was observed in the andesitic and basaltic fused glasses prepared in this study (Fig. 3b). Comparison with non-diluted fused-glass samples—even samples with heterogeneity—indicates that MgO-diluted fused glasses are severely depleted in Pb and U (Figs. 2b and 3b). Depletion in Rb and Cs was also noticeable, with MgO dilution possibly increasing their volatilization from granitic and rhyolitic rocks.
| Sample Fused glass points (in μg/g) | JG-1a (Granodiorite) | JG-2 (Granite) | JG-3 (Granodiorite) | JR-1 (Rhyolite) | isotope | Detection limit
μg/g |
Sensitivity
cps/(μg/g) |
||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ref.V | MFG n = 6 |
Ref.V | MFG1 n = 6 |
MFG2 n = 6 |
FG1 n = 6 |
FG2 n = 6 |
Ref.V | MFG
n = 12 |
FG
n = 3 |
Ref.V | MFG1
n = 9 |
MFG2
n = 6 |
|||||||||||||||||||||||
| AVE | RSD | RD | AVE | RSD | RD | AVE | RSD | RD | AVE | RSD | RD | AVE | RSD | RD | AVE | RSD | RD | AVE | RSD | RD | AVE | RSD | RD | AVE | RSD | RD | MFG* | MFG* | BCR2G | ||||||
| Li | 79.5 | 70.61 | 2% | 0.89 | 42.2 | 39.40 | 3% | 0.93 | 38.42 | 2% | 0.91 | 40.25 | 15% | 0.95 | 37.23 | 17% | 0.88 | 20.9 | 20.89 | 9% | 1.00 | 21.15 | 7% | 1.01 | 61.4 | 49.4 | 4% | 0.80 | 56.75 | 5% | 0.92 | Li7 | 0.2 | 4987 | 7407 |
| B | 3.95 | 2.86 | 58% | 0.72 | 1.78* | 2.63 | 61% | 1.48 | 3.88 | 37% | 2.18 | 15.64 | 37% | 8.79 | 17.97 | 37% | 10.09 | 2.15* | 1.58 | 27% | 0.74 | 2.13 | 14% | 0.99 | 117 | 3.1 | 40% | 0.03 | 6.63 | 59% | 0.06 | B11 | 1.1 | 998 | 1547 |
| Sc | 6.21 | 8.08 | 11% | 1.30 | 2.42 | 7.18 | 6% | 2.97 | 5.28 | 21% | 2.18 | 5.20 | 26% | 2.15 | 4.79 | 10% | 1.98 | 8.76 | 10.84 | 5% | 1.24 | 13.40 | 25% | 1.53 | 5.07 | 7.1 | 14% | 1.41 | 7.97 | 9% | 1.57 | Sc45 | 0.1 | 2713 | 4071 |
| Ti | 1500 | 1419.48 | 2% | 1.39 | 264 | 302.32 | 10% | 1.71 | 340.36 | 3% | 1.93 | 268.14 | 26% | 1.52 | 261.41 | 15% | 1.48 | 2880 | 2739.43 | 5% | 0.95 | 3234.41 | 9% | 1.12 | 660 | 641.4 | 4% | 2.36 | 879.29 | 6% | 3.23 | Ti49 | 0.5 | 195 | 292 |
| V | 22.7 | 16.87 | 4% | 0.74 | 3.78 | 4.05 | 29% | 1.07 | 2.78 | 14% | 0.74 | 0.63 | 13% | 0.17 | 0.62 | 18% | 0.16 | 70.1 | 69.57 | 6% | 0.99 | 77.00 | 15% | 1.10 | 7 | 2.6 | 11% | 0.38 | 7.45 | 6% | 1.06 | V51 | 0.05 | 4728 | 7088 |
| Cr | 17.6 | 3.70 | 38% | 0.21 | 6.37 | 1.33 | 27% | 0.21 | 1.42 | 25% | 0.22 | 1.57 | 74% | 0.25 | 1.57 | 58% | 0.25 | 22.4 | 39.01 | 16% | 1.74 | 23.17 | 23% | 1.03 | 2.83 | 0.5 | 30% | 0.17 | 0.64 | 27% | 0.23 | Cr53 | 1.0 | 3739 | 5620 |
| Co | 5.9 | 5.20 | 4% | 0.88 | 3.62 | 3.41 | 19% | 0.94 | 2.80 | 3% | 0.77 | 2.87 | 19% | 0.79 | 3.22 | 22% | 0.89 | 11.7 | 11.79 | 8% | 1.01 | 13.84 | 9% | 1.18 | 0.83 | 0.5 | 9% | 0.66 | 1.20 | 11% | 1.44 | Co59 | 0.03 | 5402 | 8113 |
| Ni | 6.91 | 5.41 | 9% | 0.78 | 4.35* | 6.90 | 33% | 1.59 | 2.38 | 14% | 0.55 | 1.74 | 29% | 0.40 | 2.06 | 43% | 0.47 | 14.3 | 37.80 | 10% | 2.64 | 18.69 | 10% | 1.31 | 1.67* | 1.0 | 8% | 0.60 | 3.95 | 17% | 2.36 | Ni62 | 0.2 | 1053 | 1588 |
| Rb | 178 | 133.49 | 7% | 0.75 | 301 | 213.02 | 3% | 0.71 | 258.58 | 4% | 0.86 | 331.34 | 6% | 1.10 | 315.78 | 6% | 1.05 | 67.3 | 57.39 | 5% | 0.85 | 67.92 | 3% | 1.01 | 257 | 100.6 | 13% | 0.39 | 148.68 | 2% | 0.58 | Rb85 | 0.08 | 6852 | 10271 |
| Sr | 187 | 162.81 | 2% | 0.87 | 17.9 | 15.70 | 9% | 0.88 | 16.35 | 2% | 0.91 | 13.70 | 22% | 0.77 | 14.31 | 16% | 0.80 | 379 | 328.38 | 5% | 0.87 | 400.51 | 11% | 1.06 | 29.1 | 24.9 | 5% | 0.86 | 30.66 | 9% | 1.05 | Sr88 | 0.01 | 7617 | 11437 |
| Y | 32.1 | 25.43 | 3% | 0.79 | 86.5 | 68.67 | 4% | 0.79 | 77.88 | 3% | 0.90 | 97.13 | 39% | 1.12 | 67.29 | 27% | 0.78 | 17.3 | 16.30 | 4% | 0.94 | 17.46 | 16% | 1.01 | 45.1 | 40.3 | 6% | 0.89 | 44.35 | 6% | 0.98 | Y89 | 0.01 | 5847 | 8779 |
| Zr | 118 | 76.22 | 4% | 0.65 | 97.6 | 60.97 | 3% | 0.62 | 150.64 | 6% | 1.54 | 58.86 | 65% | 0.60 | 37.06 | 42% | 0.38 | 144 | 236.32 | 10% | 1.64 | 98.97 | 53% | 0.69 | 99.9 | 107.4 | 6% | 1.08 | 115.73 | 6% | 1.16 | Zr90 | 0.01 | 2850 | 4280 |
| Nb | 11.4 | 10.70 | 3% | 0.94 | 14.7 | 12.42 | 2% | 0.85 | 13.60 | 2% | 0.93 | 14.21 | 28% | 0.97 | 17.13 | 38% | 1.17 | 5.88 | 5.57 | 5% | 0.95 | 6.86 | 13% | 1.17 | 15.2 | 14.6 | 3% | 0.96 | 15.37 | 4% | 1.01 | Nb93 | 0.006 | 6284 | 9443 |
| Cs | 10.6 | 5.57 | 18% | 0.53 | 6.79 | 3.70 | 12% | 0.55 | 5.32 | 8% | 0.78 | 7.90 | 9% | 1.16 | 7.32 | 8% | 1.08 | 1.78 | 1.28 | 11% | 0.72 | 2.04 | 7% | 1.15 | 20.8 | 4.0 | 24% | 0.19 | 6.55 | 4% | 0.31 | Cs133 | 0.04 | 9239 | 13841 |
| Ba | 470 | 451.15 | 3% | 0.96 | 81 | 60.58 | 3% | 0.75 | 57.07 | 4% | 0.70 | 58.43 | 21% | 0.72 | 57.10 | 15% | 0.70 | 466 | 442.69 | 6% | 0.95 | 487.71 | 7% | 1.05 | 50.3 | 43.2 | 4% | 0.86 | 46.55 | 4% | 0.93 | Ba137 | 0.03 | 1020 | 1538 |
| La | 21.3 | 20.69 | 5% | 0.97 | 19.9 | 14.94 | 7% | 0.75 | 21.02 | 4% | 1.06 | 12.54 | 28% | 0.63 | 12.37 | 23% | 0.62 | 20.6 | 18.60 | 5% | 0.90 | 32.15 | 49% | 1.56 | 19.7 | 16.8 | 5% | 0.85 | 18.89 | 7% | 0.96 | La139 | 0.005 | 6984 | 10519 |
| Ce | 45 | 49.03 | 5% | 1.09 | 48.3 | 42.74 | 7% | 0.88 | 61.19 | 3% | 1.27 | 33.56 | 25% | 0.69 | 35.31 | 26% | 0.73 | 40.3 | 43.51 | 6% | 1.08 | 66.32 | 50% | 1.65 | 47.2 | 46.0 | 3% | 0.98 | 49.79 | 4% | 1.05 | Ce140 | 0.003 | 8895 | 13394 |
| Pr | 5.63 | 4.91 | 5% | 0.87 | 6.2 | 5.09 | 9% | 0.82 | 6.99 | 4% | 1.13 | 4.06 | 30% | 0.65 | 4.06 | 25% | 0.66 | 4.7 | 4.31 | 5% | 0.92 | 6.70 | 45% | 1.43 | 5.58 | 5.2 | 4% | 0.94 | 5.89 | 5% | 1.06 | Pr141 | 0.003 | 9984 | 15042 |
| Nd | 20.4 | 18.37 | 4% | 0.90 | 26.4 | 20.98 | 9% | 0.79 | 27.82 | 4% | 1.05 | 17.83 | 29% | 0.68 | 17.70 | 27% | 0.67 | 17.2 | 15.73 | 5% | 0.91 | 23.32 | 38% | 1.36 | 23.3 | 20.8 | 5% | 0.89 | 23.53 | 6% | 1.01 | Nd146 | 0.019 | 1667 | 2507 |
| Sm | 4.53 | 3.95 | 3% | 0.87 | 7.78 | 6.64 | 10% | 0.85 | 7.94 | 5% | 1.02 | 6.54 | 29% | 0.84 | 6.13 | 37% | 0.79 | 3.39 | 3.01 | 5% | 0.89 | 3.86 | 24% | 1.14 | 6.03 | 5.1 | 4% | 0.85 | 7.77 | 10% | 1.29 | Sm147 | 0.025 | 1417 | 2131 |
| Eu | 0.7 | 0.62 | 4% | 0.89 | 0.1 | 0.09 | 13% | 0.87 | 0.09 | 6% | 0.91 | 0.08 | 15% | 0.78 | 0.08 | 10% | 0.78 | 0.9 | 0.77 | 5% | 0.85 | 0.89 | 8% | 0.99 | 0.3 | 0.2 | 6% | 0.78 | 0.29 | 10% | 0.95 | Eu151 | 0.005 | 6326 | 9547 |
| Gd | 4.08 | 4.04 | 2% | 0.99 | 8.01 | 7.71 | 7% | 0.96 | 9.00 | 3% | 1.12 | 9.40 | 30% | 1.17 | 7.89 | 44% | 0.98 | 2.92 | 2.86 | 5% | 0.98 | 3.33 | 18% | 1.14 | 5.06 | 5.3 | 6% | 1.05 | 6.02 | 6% | 1.19 | Gd158 | 0.025 | 1294 | 1948 |
| Tb | 0.81 | 0.66 | 3% | 0.82 | 1.62 | 1.42 | 5% | 0.88 | 1.64 | 4% | 1.01 | 1.86 | 34% | 1.15 | 1.49 | 50% | 0.92 | 0.46 | 0.42 | 5% | 0.92 | 0.46 | 14% | 1.01 | 1.01 | 0.9 | 6% | 0.89 | 1.01 | 4% | 1.00 | Tb159 | 0.004 | 9116 | 13710 |
| Dy | 4.44 | 4.27 | 3% | 0.96 | 10.5 | 9.84 | 5% | 0.94 | 11.48 | 4% | 1.09 | 13.43 | 37% | 1.28 | 10.37 | 48% | 0.99 | 2.59 | 2.60 | 4% | 1.00 | 2.79 | 13% | 1.08 | 5.69 | 6.0 | 6% | 1.06 | 6.70 | 6% | 1.18 | Dy163 | 0.012 | 2151 | 3233 |
| Ho | 0.82 | 0.85 | 4% | 1.04 | 1.67 | 2.09 | 4% | 1.25 | 2.45 | 3% | 1.47 | 2.92 | 39% | 1.75 | 2.17 | 44% | 1.30 | 0.38 | 0.54 | 4% | 1.41 | 0.55 | 13% | 1.44 | 1.11 | 1.3 | 5% | 1.14 | 1.39 | 5% | 1.25 | Ho165 | 0.005 | 8402 | 12665 |
| Er | 2.57 | 2.56 | 4% | 1.00 | 6.04 | 6.57 | 5% | 1.09 | 7.82 | 3% | 1.30 | 9.31 | 40% | 1.54 | 6.71 | 41% | 1.11 | 1.52 | 1.68 | 5% | 1.11 | 1.65 | 14% | 1.09 | 3.61 | 4.1 | 4% | 1.13 | 4.45 | 6% | 1.23 | Er166 | 0.009 | 2690 | 4037 |
| Tm | 0.38 | 0.37 | 3% | 0.99 | 1.16 | 1.00 | 4% | 0.86 | 1.20 | 3% | 1.04 | 1.42 | 40% | 1.22 | 0.96 | 39% | 0.83 | 0.24 | 0.25 | 4% | 1.06 | 0.24 | 15% | 1.02 | 0.67 | 0.6 | 4% | 0.94 | 0.69 | 7% | 1.03 | Tm169 | 0.004 | 8355 | 12653 |
| Yb | 2.7 | 2.52 | 4% | 0.93 | 6.85 | 6.52 | 5% | 0.95 | 8.11 | 3% | 1.18 | 9.83 | 40% | 1.43 | 6.47 | 36% | 0.94 | 1.77 | 1.87 | 5% | 1.06 | 1.66 | 16% | 0.94 | 4.55 | 4.5 | 5% | 0.98 | 4.90 | 7% | 1.08 | Yb172 | 0.017 | 1884 | 2827 |
| Lu | 0.44 | 0.39 | 5% | 0.88 | 1.22 | 0.98 | 5% | 0.81 | 1.28 | 3% | 1.05 | 1.53 | 40% | 1.25 | 0.96 | 30% | 0.78 | 0.26 | 0.31 | 9% | 1.21 | 0.25 | 18% | 0.98 | 0.71 | 0.7 | 5% | 1.04 | 0.79 | 5% | 1.11 | Lu175 | 0.006 | 7277 | 11017 |
| Hf | 3.59 | 2.51 | 5% | 0.70 | 4.73 | 2.72 | 2% | 0.58 | 7.30 | 9% | 1.54 | 4.00 | 128% | 0.85 | 1.46 | 43% | 0.31 | 4.29 | 6.52 | 11% | 1.52 | 3.08 | 67% | 0.72 | 4.51 | 4.9 | 5% | 1.08 | 5.28 | 6% | 1.17 | Hf178 | 0.045 | 2047 | 3083 |
| Ta | 1.9 | 1.67 | 3% | 0.88 | 2.76 | 2.02 | 1% | 0.73 | 2.31 | 3% | 0.84 | 1.81 | 24% | 0.65 | 2.50 | 55% | 0.91 | 0.7 | 0.58 | 5% | 0.83 | 0.65 | 18% | 0.92 | 1.86 | 1.8 | 4% | 0.99 | 1.81 | 5% | 0.97 | Ta181 | 0.009 | 6726 | 10012 |
| Pb | 26.4 | 0.89 | 90% | 0.03 | 31.5 | 1.00 | 100% | 0.03 | 3.06 | 65% | 0.10 | 29.60 | 25% | 0.94 | 19.96 | 32% | 0.63 | 11.7 | 1.61 | 59% | 0.14 | 7.27 | 22% | 0.62 | 19.3 | 0.1 | 52% | 0.00 | 0.10 | 61% | 0.01 | Pb208 | 0.052 | 5238 | 7862 |
| Th | 12.8 | 11.73 | 3% | 0.92 | 31.6 | 27.81 | 29% | 0.88 | 25.50 | 4% | 0.81 | 29.53 | 33% | 0.93 | 25.55 | 73% | 0.81 | 8.28 | 6.72 | 6% | 0.81 | 9.98 | 88% | 1.21 | 26.7 | 24.7 | 4% | 0.92 | 26.80 | 6% | 1.00 | Th232 | 0.008 | 6811 | 10264 |
| U | 4.69 | 1.66 | 22% | 0.35 | 11.3 | 4.52 | 37% | 0.40 | 6.02 | 22% | 0.53 | 13.30 | 18% | 1.18 | 10.26 | 51% | 0.91 | 2.21 | 1.90 | 16% | 0.86 | 1.69 | 21% | 0.76 | 8.88 | 0.9 | 30% | 0.11 | 2.86 | 12% | 0.32 | U238 | 0.006 | 11957 | 18126 |
Fused glass: MFG: MgO-diluted fused glass.
RSD: relative standard deviation, RD: relative deviation to reference value (Ref.V) from Imai et al. (1995) and GeoReM database (Jochum and Nohl, 2008).
See Table S3 for all spot data.
MFG*: calculated detection limit and sensitivity for MgO-diluted fused glass measurement.
Accuracies of analyses of granitic rock fused glasses (JG-1a, JG-2, and JG-3). (a) Relative standard deviation (RSD) of spot analyses. (b) Relative deviation (RD) of average values of spot analyses from the reference values. Note that the Ni reference value of JG-2 is highly uncertain and the error bar indicates uncertainty. FG, fused glass; MFG, MgO-diluted fused glass; N, number of analytical spots.
Accuracies of analyses of rhyolite (JR-1) and andesitic/basaltic rocks (JA-2, JB-2, and BHVO-2). (a) Relative standard deviation (RSD) of spot analyses. (b) Relative deviation (RD) of average values of spot data from reference values. Note that the Ni reference values of JR-1 and JA-1 are highly uncertain. Error bars indicate uncertainties in the reference values of Co and Ni in JR-1. Abbreviations as for Fig. 1.
The Zr, Hf, and REEs in MgO-diluted fused glasses were also homogeneous with 3%–9% RSD and improved over non-diluted fused glasses (>40% RSD), although there was slight heterogeneity in JG-3 glass for Zr and Hf (~11% RSD; Fig. 2a). The Zr and Hf contents for granodiorite (JG-1a and JG-3) and granite (JG-2) were >20% higher or lower than reference values in spite of the homogeneity (Fig. 2), and those of rhyolite (JR-1) were within 5% (Fig. 3b). In contrast to the discrepancy in Zr and Hf contents (~40%), Zr/Hf ratios were consistent with reference values (within 10%) in MgO-diluted fused glasses of granitic rocks (Fig. 2b). Some accessory minerals such as zircon contribute Zr and Hf to granitic rocks, so the discrepancy may be due to heterogeneity of the powder aliquot used in making the fused glass. Our MgO-diluted fused glass was generated from ~15 mg of sample in 20 mg total mass. Preparation of multiple fused glasses from a single sample powder aliquot may reduce the effects of heterogeneity of the powder. For zircon-rich granitic rocks, averaged values for multiple fused glasses would provide more reliable data for Zr and Hf (e.g., JG-2; Fig. 2b).
For compatible elements, the Co content is useful in discriminating arc volcanic rocks in the Co–Th diagram (Hastie et al., 2007). Cobalt contents were determined with an accuracy of better than ±20% for granitic rocks (JG-1a, 5.9 μg g–1; JG-2, 3.6 μg g–1; JG-3, 11.7 μg g–1; Fig. 2b). The measured Co content of JR-1 is consistent with the reference value within uncertainty (0.83 ± 0.59 μg g–1), although with a 40% deviation (Fig. 3b). Reliable Co data can thus be obtained using MgO-diluted fused glass. Chromium data are considered reliable using the fused-glass method for contents of >150 μg g–1 (Tamura et al., 2015), so the Cr content of Si-rich samples (<22 μg g–1) was too low for accurate measurement, whereas the results for high-Cr samples such as JA-2 (436 μg g–1) and BHVO-2 (293 μg g–1) were consistent with reference values (Figs. 3 and 4). The low accuracy achieved for Ni in JG-2 and JR-1 is due to the 70% uncertainty in the reference values (Figs. 2b and 3b). It is unclear what causes the discrepancy in JG-3 but, considering its precision of 10% RSD, we speculate that it may be due to heterogeneity of the powder aliquot.
Comparison of chondrite-normalized incompatible trace-element patterns of MgO-diluted fused glasses and reference values (green line). Chondrite values are from Sun and McDonough (1989). All analytical spot data are shown in each diagram (N, number of spots). Red and gray patterns indicate data determined using Agilent 7850 and Agilent 7500s, respectively. See Fig. S3 for results for fusion without MgO dilution.
Incompatible trace-element patterns are compared with those of the reference data in Fig. 4, demonstrating that data obtained with MgO-diluted fused glass successfully reproduced geochemical characteristics of high-Si rocks. This confirms that the LA–ICP–MS data are as reliable and useful for petrological and geochemical studies as those involving conventional fused glasses prepared from basalt and andesite (Fig. S2).
We determined whole-rock trace-element compositions of Si-rich igneous rocks (granitic and rhyolitic rocks) using LA–ICP–MS analysis with fused glasses. The MgO-dilution method conducted by simple mixing procedure successfully produced homogeneous fused glasses from Si-rich rocks. The method is useful in preparing fused glasses for LA–ICP–MS analysis of Si-rich rocks, enabling determine whole-rock trace-element compositions.
LA–ICP–MS analysis of fused glass with or without MgO dilution provides reliable data for many elements that are useful in petrological and geochemical studies. However, the results for some volatile elements such as Pb are highly uncertain due to volatilization during fusion, and the method is not suitable for use where such data are critical. However, where Li data are required, the method has advantages in terms of being a rapid and simple procedure relative to other methods where ultra-fine pressed powder pellets and acid digestion are required.
We would like to thank N. Hasebe and K. Fukushi for their help in maintaining the laboratory environment. Analytical and technical assistance was provided by S. Ishimaru and S. Umino. This wok was partially supported by Kanazawa University (grant no. SAKIGAKE 2020) and the Ministry of Education Culture, Sports, Science, and Technology of Japan (Grants-in-Aid nos. 18H01303 (KO), 19H01990 (TM), and 19KK0092 (TM)). This manuscript was improved by constructive comments from two anonymous reviewers and editorial comments.
