ARTICLE pubs.acs.org/ac
Determinations of Rare Earth Element Abundance and U-Pb Age of Zircons Using Multispot Laser Ablation-Inductively Coupled Plasma Mass Spectrometry Takaomi D. Yokoyama,† Toshihiro Suzuki,‡ Yoshiaki Kon,§ and Takafumi Hirata*,† †
Laboratory for Planetary Sciences, Kyoto University, Kitashirakawa Oiwakecho, Kyoto 606-8502, Japan Institute for Research on Earth Evolution, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokosuka, Kanagawa 237-0061, Japan § Geological Survey of Japan, Tsukuba, Ibaraki 305-8551, Japan ‡
ABSTRACT: We have developed a new calibration technique for multielement determination and U-Pb dating of zircon samples using laser ablation-inductively coupled plasma mass spectrometry (ICPMS) coupled with galvanometric optics. With the galvanometric optics, laser ablation of two or more sample materials could be achieved in very short time intervals (∼10 ms). The resulting sample aerosols released from different ablation pits or different solid samples were mixed and homogenized within the sample cell and then transported into the ICP ion source. Multiple spot laser ablation enables spiking of analytes or internal standard elements directly into the solid samples, and therefore the standard addition calibration method can be applied for the determination of trace elements in solid samples. In this study, we have measured the rare earth element (REE) abundances of two zircon samples (Nancy 91500 and Presovice) based on the standard addition technique, using a direct spiking of analytes through a multispot laser ablation of the glass standard material (NIST SRM612). The resulting REE abundance data show good agreement with previously reported values within analytical uncertainties achieved in this study (10% for most elements). Our experiments demonstrated that nonspectroscopic interferences on 14 REEs could be significantly reduced by the standard addition technique employed here. Another advantage of galvanometric devices is the accumulation of sample aerosol released from multiple spots. In this study we have measured the U-Pb age of a zircon sample (LMR) using an accumulation of sample aerosols released from 10 separate ablation pits of low diameters (∼8 μm). The resulting 238U-206Pb age data for the LMR zircons was 369 ( 64 Ma, which is in good agreement with previously reported age data (367.6 ( 1.5 Ma).1 The data obtained here clearly demonstrate that the multiple spot laser ablation-ICPMS technique can become a powerful approach for elemental and isotopic ratio measurements in solid materials.
I
t is widely recognized that the combination of inductively coupled plasma mass spectrometry (ICPMS) and laser ablation (LA) sample introduction led to one of the most sensitive and versatile analytical techniques for the elemental analyses of solid samples.2 4 Recently we have developed a new analytical technique using a femtosecond (fs)-LA-ICPMS technique, which enabled us to determine the trace-element abundances directly from metallic or various opaque minerals, including sulphides or elemental deposits, with a spatial resolution of better than 10 μm.5,6 In femtosecond-laser ablation there is no significant difference between the depths of ablation pits in glass and zircon materials ablated under identical experimental conditions, while in conventional nanosecond (nanosecond) laser ablation, the resulting crater depth on the zircon crystals was almost half the level of that obtained for glass material. Both, the thermally induced and particle size-related elemental fractionations, which have been thought to be the main sources of analytical error in LA-ICPMS analysis, could be reduced in the femtosecond-laser ablation.5,7 9 Despite the obvious success in obtaining less elemental fractionation during the laser ablation or ionization processes, there still r 2011 American Chemical Society
remains systematic and nonsystematic error in the results. Nonspectroscopic interferences (i.e., matrix effect) are the major source of analytical error.10,11 Although the nonspectrometric interferences can be successively reduced by the internal-standardization protocol, the choice of the internal standard elements would be severely restricted. Unlikely with the elemental analysis using solutions, spiking of the internal standard elements is also very difficult for most solid samples. Faced with these problems, we have employed the in-cell mixing technique using the galvanometric optics in the laser ablation sample introduction technique. Fernandez et al. first demonstrated the analytical capability of the incell mixing technique using the galvanometric device.12 They reported the abundance values for several trace-elements obtained with isotope dilution mass spectrometry (IDMS) using an in-cell mixing technique with analytes and isotopically enriched isotopes ablated from solid samples (in-cell femtosecond-LA-ICP-IDMS).12 Received: May 16, 2011 Accepted: October 17, 2011 Published: October 17, 2011 8892
dx.doi.org/10.1021/ac2012448 | Anal. Chem. 2011, 83, 8892–8899
Analytical Chemistry With the use of galvanometric optics, a laser rastering speed of higher than 200 mm s 1 could be achieved and the resulting sample aerosols released from two solid materials (unknown sample and standard material spiked with enriched isotopes) were mixed within the cell and were transported into the ICP. With the solid-spiking LA-ICP-IDMS technique, a relative deviation of better than 10% from certified values could be achieved for Cu, Zn, Sn, and Pb.12,13 One great advantage of using the ID technique is that the measured isotope ratio data should be less sensitive to matrix components of the solid sample, leading to an improvement of precision or trueness of measurement results. Another advantage of using galvanometric devices is the possibility to ablate comparably larger sample areas, covering 100 μm 1 mm. With laser ablation at high repetition rates (e.g., 10 kHz) and fast scanning speeds, the potential contribution of sample heterogeneity effects, as common in single spot ablation, could be minimized. Fernandez et al. reported that the relative standard deviation of the results could be significantly improved at repetition rates higher than 7.5 kHz.12 Despite the success in obtaining reliable elemental concentration data, practical application using the solid-spiking LA-ICPIDMS technique has been slowed down due to a lack of commercially available solid reference materials spiked with enriched isotopes. Moreover, mixing ratios between the sample and spike must be carefully optimized to minimize the error propagation factor in IDMS analysis,14 and this sometimes proved to be very difficult when multiple elements were measured simultaneously using the suggested IDMS technique.14 Nevertheless, the in-cell mixing technique can provide wider flexibility for elemental and isotopic analyses of trace-elements. To extend the analytical capability of the in-cell mixing technique, we have developed a new femtosecond-laser ablation system utilizing a galvanometric optics for elemental and isotopic analyses for natural zircon samples. Zircon crystals occur in the most of felsic igneous and sedimentary rocks as well as their metamorphic equivalents. Because of the ubiquitous presence of zircon in many rock samples and also due to the robustness of the mineral, the elemental and isotopic signature obtained from the zircon grains can provide various geochemical information such as chemical composition of source magma or provenance studies of single detrital zircon grains from sedimentary rocks. More importantly, the U Pb zircon geochronology is one of the most versatile and precise dating methods in determining ages of zircon crystallization or metamorphic overgrowth events, and therefore, it is widely recognized that the zircon is one of the most important mineral to decode the formation sequence of the rocks in nature. In this study, two natural zircon samples (Nancy 9150015,16 and Plesovice17,18 were subjected to the determination of the rare earth element (REE) abundances and the series of zircon grains from Lake Mountain Rhyodacite (LMR) were used for the U-Pb ages.
’ EXPERIMENTAL SECTION Ti:S Laser Ablation with Galvano Devices. The titaniumsapphire femtosecond laser used in this study was a Cyber IFRIT (Cyber Laser Inc., Tokyo, Japan) equipped with a third harmogenic convertor (Table 1). Wavelength and energy used were 260 nm and 54 μJ with a maximum repetition rate of 1 kHz.5 Although both the UV wavelength (260 nm) and fundamental near-infrared (NIR) wavelength (780 nm) can be applied for the laser ablation of solid samples, the UV wavelength was used
ARTICLE
Table 1. Instrumentation and Operational Conditions 1. ICP Mass Spectrometer instrument
AttoM (Nu Instruments, Wrexham, U.K.)
scan mode
electrostatic scan (E-scan)
monitored isotope REE deteminations
La (139), Ce (140), Pr (141), Nd (146), Sm (147), Eu (153), Gd (157) Tb (159), Dy (163), Ho (165), Er (166), Tm (169), Yb (172), Lu (175)
U Pb dating
Pb (204, 206, 207, 208), Th (232), U (238)
dwell time
500 μs for 204Pb and 1000 μs for others
detector
pulse counting
total analysis time
10 s/run
2. Laser Ablation System instrument
Cyber Probe (Cyber Inc., Tokyo, Japan)
laser pulse duration
IFRIT Type-C Ti:S femtosecond laser 230 fs
wavelength
260 nm (THG)
galvano device
Cannon GM-1000GC-201 2
interface positioning time dwelltime objective lens covered area raster speed repetition rate fluence ablation pit size dwell time no. of shots total ablation time
(for x and y-axes) Scan Lab RTC4PC 0.4 1 ms 10 ms f-θ lens (f = 100 mm) 20 mm 20 mm 200 mm/s 100 Hz 9.57 J/cm2 for REE determinations 5.77 J/cm2 for U Pb dating 15 μm for REE determinations 8 μm for U Pb dating 100 ms/spot 100 shots/spot 10 s/spot
3. Standardization REE deteminations U-Pb dating normalization
NIST SRM 612 (glass standard) Nancy 91500 (natural zircon) 206 Pb/238U = 0.179 17 (Nancy 91500) 238 U/235U = 137.88
throughout this study. This is because the particle size distribution can be modified toward smaller sizes using a shorter wavelength laser irradiation, and this results in smaller elemental fractionation during sample ablation and ionization.19 22 An inhouse He-flushing ablation cell (T200 cell) was employed in our experimental setup to minimize elemental and isotopic fractionation during the laser ablation stage.10 Galvanometric optics are widely used for laser microfabrication applications. With the angles of the two mirrors changed using two galvanometric motors, the position of the ablation pit can be moved in short time intervals without moving the sample itself as common in conventional LA setups (Figure 1a). The system used allows changing the ablation position on the sample within 0.4 1 ms. The number of laser shots at the targeted points can be controlled by the dwell time, which is defined as the sitting time for the angle of the galvanometric mirrors. Hence, the dwell time of 10 ms was employed throughout this study. With the dwell 8893
dx.doi.org/10.1021/ac2012448 |Anal. Chem. 2011, 83, 8892–8899
Analytical Chemistry
ARTICLE
shorter than 100 μs). Note that the mass scan rate achieved by the E-scan is comparable or even faster than those achieved by conventional quadrupole-based ICP mass spectrometers. The data acquisition using the E-scan mode was used for both the determination of 14 REEs (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu) and the U-Pb isotopic ratio (206Pb/238U and 207Pb/235U) measurements. The analysis sequence begins with the analysis of the carrier gas without any laser ablation (gas blank), and this is followed by three separate laser ablation sequences for NIST SRM612, series of three Hamilton Glass (DLH6, blank; DLH7, 75 μg/g; DLH8, 150 μg/g) and the natural zircon sample (Nancy 91500 and Plesovice) (Figure 1b,c). Details of the instrumentation and the operational settings for the ICPMS and the laser ablation system were summarized in Table 1. For REE determinations, no internal standardization was made throughout the analysis. Abundances for the 14 REEs were calibrated based on the standard addition method using solid-spiking with the NIST SRM612 glass standard. Another application of the present femtosecond-laser ablation technique using the galvanometric optics are the U-Pb age determination on zircon samples based on the 206Pb/238U and 207Pb/235U isotope ratio measurements. Hence, pseudosimultaneous laser ablation was used for the accumulation of the sample aerosols released from 10 separate zircon grains. Because of the large contribution of the counting statistic errors, mainly due to very low isotopic abundance of 235 U (0.7%) in the natural samples, the signal intensity of 235U was not measured throughout the U Pb isotopic analysis. Instead, the signal intensities of 235U for natural samples were calculated from the signal intensity of 238U based on a 238U/235U isotopic ratio of 137.88.23 The mass bias effect for the 206Pb/238U ratio was corrected by normalization to the 206Pb/238U ratio for Nancy 91500 zircon of 0.179 17.15 Figure 1. Schematic diagram of the laser ablation system equipped with galvanometric optics (a) and laser ablation protocols for elemental analysis (b,c).
time of 10 ms and laser repetition rate of 100 Hz, the number of laser shots per point was unity per single scan sequence. In this study, the number of laser scans of 100 was adopted, and this meant that the number of laser shots was 100 per point per analysis. The overall covered area achieved by our galvanometric system was 20 mm 20 mm, and the highest rastering speed was about 200 mm s 1. Although faster rastering can be achieved on the present galvanometric driver motors, precision of the ablation positioning could be deteriorated when further high rotation speed was adopted. This is mainly due to an oversteering of the galvanometric mirrors by the inertial force under the high rotation speed of the Galvanometrix mirrors. With multiple laser ablation events at different spots within very short time intervals, laser ablation of two or more different sample materials is feasible. Triggering of the laser firing was controlled by the galvanometric controllers, and therefore the laser ablation of precisely defined sample spots can be made. With the present femtosecond-laser and the galvanometric optics, pseudosimultaneous ablation of two or more samples can be applied for elemental and isotopic analyses. Analytical Protocol. The ICPMS used in this study was a Nu Instruments AttoM (Nu Instruments, Wrexham, U.K.) high resolution-magentic sector field-ICP mass spectrometer. Mass scanning using an electrostatic reflector (E-scan) provides fast successive monitoring of multiple isotopes (i.e., dwell time of
’ RESULTS AND DISCUSSION Standard Addition Capability. The sample aerosols released from multiple ablation points within a short time interval (∼10 ms) were well mixed and homogenized in the ablation cell, since the typical washout time for the sample aerosols (typically 1 1.5 s) was significantly longer than the time intervals between the laser firing. The resulting mixture of the sample aerosols was further homogenized by a signal smoothing device after the ablation cell.24 To evaluate the level of aerosol mixing through the solidspiking, we have measured the abundances of 14 REEs by means of the standard addition calibration method using three standard glass materials (DLH6, DLH7, and DLH8) (Figure 1b). For this experiment, the NIST SRM612 glass standard was adopted as the unknown sample. Signal intensities of 14 REEs were obtained from three different material combinations: (a) NIST SRM612 + DLH6 (blank), (b) NIST SRM612 + DLH7 (REE 75 μg/g), and (c) NIST SRM612 + DLH8 (REE 150 μg/g). The sample materials were set into the sample cell (Figure 1a). To test the reproducibility of the analysis, the experiment was repeated three times. In Figure 2, the resulting signal intensity data were plotted against the concentrations of the REEs in the three REE standards (DLH6, 7, and 8). The clear linear correlation with the signal intensity and the REE concentrations revealed that the sample aerosols released from two solid samples were well mixed and homogenized within the sample cell or through sample transport stages. The x-axis intercept of the regression line should reflect the concentrations of REE in the NIST SRM612. The resulting REE abundance values for the NIST 8894
dx.doi.org/10.1021/ac2012448 |Anal. Chem. 2011, 83, 8892–8899
Analytical Chemistry
ARTICLE
Figure 2. Correlation of signal intensities of REE and concentrations obtained from solid-spiking of DLH glass standards (DLH6, 7, and 8) into NIST SRM612 as an unknown sample.
Table 2. Rare Earth Element Abundances for NIST SRM612 and Two Zircons Determined by Standard Addition Calibration Technique, Together with Abundances Data for Calibration Standard (Hamilton Glass) elements (isotope)
NIST SRM612a
Hamilton Glassa
Nancy 91500 Zircona
this study literatureb DLH6 DLH7 DLH8 Leedey Chondritec
this study
I&Hd
Plesovice Zircona
Wiedene Sano f
this study
36.3 ( 2.3
35.77
0
75
150
0.378
0.079 ( 0.980
0.031
0.016
0.006 1.99 ( 17.57
Ce(140) 42.9 ( 3.6
38.35
0
75
150
0.976
2.83 ( 1.34
2.77
2.75
2.56
Pr(141) 36.0 ( 1.8 Nd(146) 36.7 ( 3.4
37.16 35.24
0 0
75 75
150 150
0.136 0.716
0.031 ( 0.035 0.264 ( 0.278
0.022 0.324
0.036 0.32
0.024 1.07 ( 0.47 0.24 12.9 ( 4.5
La(139)
5.27 ( 1.58
meang ming maxg 0.32
0.18
2.7
0.93
9.8
0.34 3
0.13 0.97
1.7 11.9 11.7
Sm(147) 38.0 ( 4.1
36.72
0
75
150
0.230
0.262 ( 0.223
0.624
0.304
0.5
8.85 ( 1.38
4.5
2.1
37.9 ( 3.8
34.44
0
75
150
0.087
0.220 ( 0.193
0.31
0.212
0.24
2.97 ( 0.98
1.2
0.26
Gd(157) 37.8 ( 5.7
36.95
0
75
150
0.311
1.41 ( 0.91
2.72
1.6
2.21
28.3 ( 5.1
14.7
Eu(153)
Tb(159) 38.0 ( 5.8
35.92
0
75
150
0.059
1.54 ( 0.74
Dy(163) 36.9 ( 7.1
35.97
0
75
150
0.390
13.55 ( 4.74
1.11 14.2
0.8 12.9
0.86 11.8
14.4 ( 2.0 84.0 ( 8.1
5.7 61
7 2.6 28
1.5
3.1 32 11.2 121
Ho(165) 38.0 ( 4.4
37.78
0
75
150
0.089
4.72 ( 1.41
5.83
5.5
4.84
28.8 ( 3.7
17.1
8
35
Er(166) 37.7 ( 5.8 Tm(169) 38.2 ( 6.4
37.43 37.55
0 0
75 75
150 150
0.255 0.039
28.7 ( 8.0 8.56 ( 1.28
28.2 8.44
29.2 7.6
24.6 6.89
89.6 ( 10.2 17.0 ( 2.2
59 10.6
27 5
123 24
81
38
185
Yb(172)
36.9 ( 5.7
39.95
0
75
150
0.249
102.9 ( 6.6
Lu(175)
37.5 ( 6.0
37.71
0
75
150
0.039
16.6 ( 1.8
101 13.9
84
73.9
178 ( 11
18.3
13.1
16.5 ( 1.6
9.8
4.6
23
Abundance values given in μg/g. b Reference 25. c References 27 and 36. d Iizuka and Hirata, Geochem. J. 2004, 38, 241 229. e Reference 16. f Reference 37. g Reference 18. a
SRM612 are listed in Table 2. For comparison, reported REE abundance data by Pearce et al.25 was also given in this table. The obtained REE abundance values for NIST SRM612 are in good agreement with literature data within the analytical uncertainties estimated based on statistical errors of slope and intercept through least-squares regression calculations.26 Relative deviations from the literature data are typically smaller than 6%, except for Ce (12%) and Eu (10%). Despite the slightly larger deviation
in the resulting Ce and Eu abundance data, the abundance data obtained here demonstrated that the trace element abundances could be well calibrated by means of the standard addition technique with reproducibilities ranging from 5 to 15%. In the next experiment we have applied the standard addition technique to estimate of REE abundances in two natural zircon (ZrSiO4) samples galvanometric (Nancy 91500 and Plesovice zircons). 8895
dx.doi.org/10.1021/ac2012448 |Anal. Chem. 2011, 83, 8892–8899
Analytical Chemistry
ARTICLE
Figure 3. Correlation of signal intensities of REE and concentrations obtained from multispot laser ablation of NIST SRM612, DLH6 (blank) and zircon samples.
Determination of REE in Zircon Using the Standard Addition Technique. The Nancy 91500 zircon was collected from
Kuehl Lake, Renfrew Country, Canada, and originally consisted of a single crystal with a mass of 238 g.19,16 The Plesovice zircon comes from a potassic granulite from the Plesovice quarry situated in the eastern part of the “Blansky les granulite” body in the southern Bohemian Massif, Czech, Republic.17 These two zircons have been widely used as calibration standards for the U-Pb dating method or trace element determinations in zircons. Among the trace elements, abundances of REEs have been given attention because the REE abundance pattern can provide key information about the growth sequence of zircon crystals (e.g., refs 10, 11, and 27), the coexisting paragenesis (e.g., refs 28 33), or about the chemical composition of the source magma.34,35 In this study, the 14 REE abundances of two zircon samples were determined by the standard addition technique using NIST SRM 612 as an external standard. Two zircon samples were mounted in a resin (Resine Mecaprex KM-U, PRESI). The surfaces of zircons and standard material were polished with a diamond paste (8000 mesh), and the resulting sample pieces were washed with ethanol for 3 min using an ultrasonic bath to remove possible surface contamination of trace elements due to sample preparation. The zircon samples, calibration standard (NIST SRM 612) and blank standard (DLH6) were set into the sample cell (Figure 1a). Laser ablation of the blank standard (DLH6, blank glass) was very important to keep the time proportion of the laser ablation for the unknown sample constant (Figure 1c). Hence the time proportion of the laser ablation of unknown sample was kept at 16.7% (100 shots per total 600 shot in each analysis sequence) throughout the REE analysis. This is analogous to the standard addition calibration technique using liquid samples, in which the final volume of the analysis solution would
be constant despite the different amounts of standard solution added. Contribution of the NIST SRM612 (i.e., amount of NIST SRM612 spiked to the sample) was controlled by changing the time proportion for laser ablation (i.e., adjusted by changing the number of laser shots from 100 to 500 shots) (Figure 1c). The resulting signal intensities of 14 REEs for aerosol mixtures from Nancy 91500 zircon and NIST SRM612 standard were plotted against the concentration of REE in the mixture of NIST SRM612 and blank (Figure 3). Clear linear correlations with the REE concentration values and the signal intensities were found for almost all REEs except for Ce. For the Ce data, measured deviation from the regression line was about 20% (Figure 3). This level of discrepancy was significantly larger than the typical analytical uncertainties achieved here (10 15%). The deviations of the data points from the regression lines for Ce can be attributed to the heterogeneous distribution of Ce within the zircon crystal, possibly induced through the crystal growth or secondary alteration processes. Despite the slightly poor linear correlation of the data points, the measured signal intensities for Ce were closely correlated with the Ce abundances which was approximated using a straight regression line (R2 = 0.989). The concentrations of REEs in Nancy 91500 zircons could be calculated based on the x-axis intercepts of the regression lines (x = 0). The calculated concentrations of 14 REEs are listed in Table 2 and are plotted as REE abundance or chondritic pattern after normalization to REE abundances in the Leedey chondrite, which is considered a representative material of the solar composition (Figure 4a).27,36 For comparison, previously reported REE abundance data of the investigated samples10,25,37 are also plotted in Figure 4. The estimated REE abundances for Plesovice zircon with the same technique are also given in Table 2 and Figure 4b. For Plesovice zircon, although the measured abundance values for all REEs were systematically higher than the reported values,17 the present 8896
dx.doi.org/10.1021/ac2012448 |Anal. Chem. 2011, 83, 8892–8899
Analytical Chemistry
ARTICLE
Figure 4. Chondrite normalized REE patterns for two zircon samples. Previously reported REE abundance patterns were also plotted for comparison.
REE abundance values were almost within the variation range. The systematical deviation found in the present REE values may reflect the grain-by-grain heterogeneity of the Plesovice zircons. For the Nancy 91500 zircon, the resulting REE abundances for almost all REEs are in good agreement with reported values by Wiedenbeck et al.25 The only exception is the abundance data for La. The measured abundance for La is almost 1 order of magnitude higher than the reported value. The most plausible explanation for the present discrepancy found in La abundance was the large contribution of counting statistics of La signal. The measured signal intensity of 138La was 50 cps, which was almost the same level of the background intensity. Because of such low signal intensities, the contribution of the counting statistics could be expected to be >50%, and this could result in the large deviation of the measured La abundance values. Another possible explanation for the large discrepancy found in the La value was the contribution of small (
