B920341a 1033.1038

www.rsc.org/jaas | Journal of Analytical Atomic Spectrometry Lithium isotope analysis of natural and synthetic glass by laser ablationMC-ICP-MS† Received 30th September 2009, Accepted 27th April 2010First published as an Advance Article on the web 21st May 2010DOI: 10.1039/b920341a This contribution presents the first laser ablation multicollector ICP-MS (LA-MC-ICP-MS)technique for the measurement of Li isotopes. A Nu Instruments NuPlasma HR operating at 6 kV andfitted with a non-standard ‘‘low-mass, high abundance’’ skimmer cone is used, coupled toa NewWave UP213 213 nm laser ablation unit operating under a He atmosphere. Asymmetric tuning ofthe quad lenses on the NuPlasma allows the simultaneous measurement of 7Li and 6Li on a Faraday-ioncounter configuration. Use of the standard-bracketing technique for analysis of unknowns obviatesany cross calibration between Faraday cup and ion counter. Use of a natural glass as bracketingstandard (e.g. BCR-2G) yields accurate Li isotope data with external, measured 2s d7Li precision of<1& at Li concentrations of 3–35 ppm.
mineral phases, surface contamination etc. Glass samplesprepared from rock powders, such as the MPI-DING glass The stable isotope pair of 7Li and 6Li are significantly fraction- standards used here,9 can also be successfully analysed for Li ated, at least 60&, between different globally important reser- isotope compositions using the proposed technique. Such voirs. This makes Li isotope compositions and systematics a fusion process is probably not a viable routine sample prepa- potentially useful when attempting to constrain a variety of ration route for Li isotope analysis.
issues in igneous geochemistry ranging from subduction and fateof crustal material, the origin and composition of so-calledmantle end-member compositions, the nature of compositional variability of the mantle beneath mid-ocean ridges and the rate of magma crystallization and cooling, to name but a few e.g.1,2–4A comprehensive review of Li isotope geochemistry is presented All measurements reported were performed using a NewWave UP213 213 nm Nd-YAG laser ablation unit with a standard Nevertheless, geochemical studies of large sample sets using Li sample cell coupled to a Nu Instruments NuPlasma HR housed stable isotopes are limited due to the difficult and time in the AEON EarthLAB, Department of Geological Sciences, consuming chemical preparation required for TIMS and solution University of Cape Town. The NuPlasma is fitted with the larger MC-ICP-MS analysis, or the cost and time required for SIMS Edwards E2 M80 rotary pump evacuating the sampler/skimmer interface for increased instrumental sensitivity. The fixed A rapid, but still precise, analytical technique capable of collector array on this NuPlasma contains 12 Faraday cups fitted producing Li isotope data for a significant number of samples at with 1011 ohm resistors, 3 ETP electron multiplier ion counters reasonable costs would therefore be a useful investigative tool.
and a single channeltron ion counter. In this study the NuPlasma Precise (<1& 2smeasured), B isotope data have been obtained on was operated at 6 kV for increased sensitivity. Table 1 lists the natural and synthetic glasses (<1 ppm B) by laser ablation general NuPlasma and laser ablation unit operating conditions coupled to a MC-ICP-MS.8 Results are presented here for a similar approach to obtain Li isotope (<1& 2s measured) Two sets of sampler and skimmer cones were used, the first compositions of glass samples with Li concentration of 3– being the general so-called ‘‘dry plasma’’ NuPlasma set. The 35 ppm. This technique offers the opportunity of greatly alternative set is comprised of the standard so-called ‘‘wet increasing the assessibility of Li isotope data.
plasma’’ sampler cone and an experimental ‘‘low-mass, high Although limited to analysis of glass samples the in situ nature abundance’’ skimmer cone. This experimental ‘‘low-mass, high of the presented technique can provide information currently abundance’’ skimmer cone was provided by Nu Instruments for only offered by SIMS analysis, such as detecting compositional evaluation purposes. At present this skimmer cone is not effects associated with cracks, fractures, presence of small generally available, and Nu Instruments should be contacted forfurther details. The increase in sensitivity with this second set of AEON EarthLAB/Department of Geological Sciences, University of Cape cones is marked (see Fig. 1). To maintain the measured Li signal Town, Rondebosch, South Africa. E-mail: [email protected]; Fax: at optimal levels (see discussion below) the laser parameters were retuned for this second set of cones. This led to the use of less † This article is part of a themed issue highlighting some of the most intense laser ablation parameters (see Table 1), resulting in more recent and significant developments in the area of Sector FieldICP-Mass Spectrometry.
This journal is ª The Royal Society of Chemistry 2010 J. Anal. At. Spectrom., 2010, 25, 1033–1038 | 1033 Typical LA-MC-ICP-MS operating settings for Li isotope Nickel‘‘wet’’ plasma sampler‘‘light-element high-abundance’’ Total Li signal in volts obtained during analysis of the labelled standard glasses using the standard NuPlasma ‘‘dry’’ plasma cone set and the experimental ‘‘low-mass, high-abundance’’ cone set, versus Li concentrations (ppm) of these standard glasses.
compositions of these glasses have also recently been deter- The Li concentrations and isotopic compositions of a range of standard glasses6,7 with natural major element compositions have recently been determined.9 The subset used in this study are presented in Table 2. In the final procedure detailed below, the BCR-2G glass standard was used as bracketing and therefore reference standard, and the remaining standards in Table 2 were used as ‘‘unknowns’’ to test the success of the technique.
Low-mass, high-abundance cone setCleaningSpot size Specific instrument tuning for Li isotope analysis The large relative mass difference ($16%) between 7Li and 6Li makes tuning the NuPlasma HR for Li isotope analysis a chal- lenge. Due to the large relative difference in abundance between the two Li isotopes (7Li 92.5% and 6Li 7.5%), the larger 7Li signalwas placed on a Faraday detector (H6) and the much smaller 6Lisignal on an ion counter (IC2) (see Table 3).
Ablation was conducted in a He atmosphere, set to a constant Although less beam manipulation is required to place the 7Li 0.9 L minÀ1. The He sweep gas stream was mixed immediately on signal on the channeltron detector (see Table 3), the 7Li signals exiting the sample chamber with a constant stream of 0.9 L minÀ1 encountered here (>2 000 000 cps, see discussion below) were too Ar make-up gas via a y-connector, before delivering the ablated high for measurement on the channeltron.
material into the plasma. All gas streams by-passed the standard Achieving alignment of the 7Li and 6Li peaks, while also valve system on the laser ablation unit, resulting in increased manipulation of the ion beams through the respective quadru-pole lenses and ion counter deflector voltage settings. Thequadrupole lenses are comprised of an array of plates above and below the ion beam path situated between the magnet and the The certified NIST 610, 612 and 614 international glass reference collector array. By applying a series of quadratically increasing standards10 were initially used for setup, tuning and as bracketing voltages to the two quadrupole lenses, the beam dispersion at the standards. These standards are synthesized Si–Na–Ca–Al-oxide collector array can be manipulated. These voltages are usually glasses produced by the U.S. National Institute of Standards and symmetric to provide uniform mass spacing between for example Technology (NIST10), with 484.6 Æ 27 ppm, 41.54 Æ 2.87 ppm,11 Sr isotope signals. However due to the significant mass difference and 1.74 ppm12 Li concentrations respectively. The Li isotope between Li isotopes, asymmetric quadrupole lens voltages are 1034 | J. Anal. At. Spectrom., 2010, 25, 1033–1038 This journal is ª The Royal Society of Chemistry 2010 Published Li concentrations and isotope compositions of standards used in this study,6,7 along with Li isotope values obtained in this study a 7Li/6Li corrected for instrumental drift, mass fractionation and referenced through BCR-2G to LSVEC (NIST 8545, 7Li/6Li 12.1716). b Analysis doneusing standard NuPlasma ‘‘dry plasma’’ cone set.
required to obtain aligned, flat-topped Li isotope signals in the The standard NuInstruments Time-Resolved Analysis (TRA) H6-IC2 configuration (see typical settings in Table 1).
software was used for all data collection. Once satisfactory 7Li-6Li Initial setup also included a Faraday-Faraday (H6-L5) peak shape and alignment were obtained, the magnet position was configuration for high Li concentration samples. Quad lens set to the mid-point of the flat-topped peaks. No automatic peak voltage settings to achieve similar aligned, flat-topped Li isotope centring was performed during subsequent data acquisition. Each peaks on these Faraday detectors are also asymmetric analysis started with a 30 s on-peak zero (OPZ) measurement of (not shown in Table 1). Due to the L5 Faraday detector being the gas background Li isotope levels without the laser firing, but positioned on the ‘‘outside’’ of the IC2 detector, slightly less all instrumental settings were the same as during ablation.
manipulation of the ion beam is required when using the H6-L5 Subsequent data acquisition during ablation yielded just more configuration. All the analytical results presented here were than 2.75 min of measured 7Li and 6Li signals. The software allows obtained using the H6-IC2 configuration, as the 3–35 ppm Li for the start and end of the range of TRA data used for reduction concentration range of the glass standards analysed did not to be selected. The starting point was usually selected to avoid any deliver a 6Li signal sufficient for measurement on a Faraday initial instability in the 7Li/6Li ratio (typically eliminating the first 5–10 s) and the end of the range placed just prior to the end of These quad lens settings were found to be stable once defined, ablation to clearly avoid instability once the signal starts to with only slight adjustments from session to session or day to decline. The typical TRA result presented in this study was derived from a 2.3–2.5 min measured Li isotope signal.
As mentioned the gas tubing from the laser ablation sample cell bypassed all valves on a direct route to the plasma. Any sample change therefore affected the stability of and balancebetween the He sweep gas and Ar makeup gas. Therefore, Laser ablation was done along a line, to obtain a stable signal following a sample change the instrument tuning was quickly re- lasting longer than 2 min. All laser ablation analyses in this study optimised. This slight retuning changed the instrumental mass were performed along 500mm lines. After rapid surface cleaning fractionation from session to session, which is shown in Fig. 2 by with a 175 mm laser spot at 25 mm sÀ1, the line was sampled at 3 the change in measured 7Li/6Li ratios of the BCR-2G bracketing mm sÀ1 using a 150 mm laser spot. Therefore an area of 500 mm by 150 mm is required for an analysis. During sampling the laser wasdefocussed $250 mm below the surface, following tests at defocusdepths up to 1000 mm, achieving improved ablation signal stability. When the laser is focused below the sampling surface Instrument sensitivity and ion counter usage limits more mass is ablated and, due to more vertical particle distri-bution, particle deposition surrounding the sampling crater is The typical OPZ total Li background levels were <0.001 V, significantly reduced13 (see Table 1 for laser details).
corresponding to <30 000 cps 7Li and <2500 cps 6Li. Total Collector layout of the NuPlasma HR (NP051) at the AEON EarthLAB, and collector assignment for Li isotope analysis as presented here a Channeltron ion counter fitted as non-standard equipment to NP051. b Faradays labelled ‘‘Hx’’ are on high-mass side of axial (Ax) Faraday, andFaradays labelled ‘‘Lx’’ are on low-mass side of the axial Faraday.
This journal is ª The Royal Society of Chemistry 2010 J. Anal. At. Spectrom., 2010, 25, 1033–1038 | 1035 Top panel. Measured 7Li/6Li ratios of bracketing BCR-2G standard analyses and analyses of other standards as unknowns. Bottom panel. d7Li (&) of labelled standards as measured in this study; corrected for instrumental drift, mass fractionation and referenced through BCR-2G to LSVEC(NIST 8545, 7Li/6Li 12.1716). Filled symbols individual analyses, black filled fields average Æ2s of each bracketed session, hatched fields publishedMC-ICP-MS and gray filled fields published SIMS values for each standard (see Table 2). Vertical-lined symbols data collected using standard cones, allother data collected using ‘‘low-mass, high-abundance’’ cone set. Date label indicates day of analysis.
measured Li signals during data acquisition ranged from 0.125 V once a day, if not more frequently. In the H6-IC2 configuration ($7500 000 cps 7Li; $520 000 cps 6Li) for BCR-2G (9 ppm Li) to used in this study, no separate calibration of the IC2 gain is 0.03 V ($1800 000 cps 7Li; $120 000 cps 6Li) for BIR-1G required due to the standard bracketing technique employed. By referencing the measured Li isotope composition of the brack- During the development of this technique, reliable results were eted samples against that of the standard, both instrumental only obtained with the 6Li signal on IC2 significantly below mass fractionation correction and IC gain calibration are 1000 000 cps. An upper limit of 600 000 cps for the 6Li signal on IC2 was preferred to ensure consistently reliable results. Above For the analysis of the higher Li concentration MPI T1-G and this limit, the 6Li signal is variably over-counted when analysing MPI ATHO-G standard glasses the 6Li signals were significantly a high Li concentration sample, resulting in too low to measure >600 000 cps. In these instances the total Li signal during the 7Li/6Li values, compared to a concomitantly analysed low Li analysis of these glasses was reduced by partially closing the concentration sample. The rational of the standard-sample- source slit. For the analysis of the bracketing standard, see standard bracketing technique is therefore compromised. This below, the source slit was opened to the default maximum observation agrees with previously reported non-linearity of position. Therefore, although the Li concentration of the stan- these specific discrete-dynode ion counters at high signal dards in Table 2 range from 3.6 ppm to 28.6 ppm, the corre- sponding total Li voltage measured only ranged from $0.03 V to The standard approach when employing a mixed Faraday-ion counter collector array in MC-ICP-MS analysis would be tocalibrate the ion counter response, or gain, against a Faraday detector. This would entail either measuring the same ion signalon both ion counter and Faraday, or measuring a known isotope Initial development of this technique employed the NIST 610, ratio in a Faraday-IC configuration. In both cases, the Faraday 612 and 614 standard glasses for instrument tuning, as well as for measurement would be considered the constant, or ‘‘true’’, the bracketing standard when analyzing samples with unknown response and the ion counter response can thus be calibrated.
Li isotope compositions. When analysing the NIST glasses This ion counter gain calibration would be performed at least a significant deposition of material around the tips of the cones, 1036 | J. Anal. At. Spectrom., 2010, 25, 1033–1038 This journal is ª The Royal Society of Chemistry 2010 especially the skimmer cone, developed rapidly. This necessitated difference between results from analysis of the same standards frequent cleaning, after 1–2 sessions, to avoid serious degrada- using either the standard cone set or the experimental low-mass, high-abundance skimmer cone set (see data for analysis of The Li isotope signals from NIST 610 and 612 were too intense BCR-2G, BIR-1G and MPI GOR128-G in Table 2 and Fig. 2).
for the IC measurement as discussed above. Only NIST 614 was The Li isotope compositions obtained with the new LA-MC- evaluated against the USGS and MPI standard glasses as ICP-MS technique also compare variably with previously a bracketing standard. Within an analytical session significantly determined secondary ion mass spectrometry (SIMS) results,6,7 different slopes of drift in measured Li isotope values between as plotted in Fig. 2. The isotopic variability observed in MPI repeat analyses were observed for NIST 614 and a USGS or KL2-G, with the resulting large 2s error, agrees with previous MPI-DING standard glass. It was therefore inappropriate to use heterogeneous results obtained by SIMS on this specific stan- the measured Li isotope values of NIST 614 to correct the dard.7 The choice between SIMS and the new LA-MC-ICP-MS measured Li isotope values of the USGS or MPI standard glasses technique is a trade-off between spatial resolution and precision, NIST 614 is significantly more transparent than the dark and All precisions listed in Table 2 and plotted in Fig. 2 are opaque USGS and MPI standard glasses, leading to significantly measured, external 2s errors based on 2–4 repeat analyses. With different coupling with the laser. Particle populations ablated the exception of MPI KL2-G, routinely achieved 2s precisions from the more transparent material (NIST 614) are biased were <1& over the 3–35 ppm Li concentration range of the to larger sizes compared with the more opaque material (e.g. BCR-2G).17 The incomplete vaporisation, atomisation and The flat-topped peak shapes of the Li isotope signals were ionisation of larger particles delivered to the plasma from laser found to be significantly sensitive to instrument tuning. Any ablation analyses has been shown to affect elemental fraction- slight slope on the tops of the peaks would result in compromised ation,18 which might also be the case here where the progressive precision. When measured external 2s errors were observed to closing of the skimmer cone aperture by the observed deposit have deteriorated to >1&, slight retuning to again achieve flat buildup bias successive analyses progressively towards the Li and level Li isotope peak top surfaces proved to rectify the isotope composition of smaller sized particles which might differ from that of the complete particle population.19 The use ofa matrix-matched bracketing standard from which the ablated particles are of small size and the population size distribution issimilar to that of the samples should eliminate or significantly This contribution demonstrates the feasibility of Li isotope analysis by LA-MC-ICP-MS following a simple and straight- For the above reasons the USGS and MPI glass standards in forward approach. Li isotope compositions are successfully Table 2 were evaluated as potential bracketing standards.
reproduced with true external measured 2s d7Li precision of BCR-2G was selected as it has moderate Li concentration and <1& at Li concentrations of 3–35 ppm for a suite of USGS and isotope composition (9 ppm, 5 Æ 0.8& d7Li7), and in this instance MPI standard glasses. A standard glass with natural composi- was available in relatively large size. The ablation characteristics tions (BCR-2G) was found to be a more successful bracketing of BCR-2G was found to be significantly more stable compared and reference standard compared to the NIST glasses. BCR-2G to the NIST glasses, and the rate of material buildup on the cones is widely available and its use should not limit other facilities was found to be drastically reduced. All the analytical results presented in this study used BCR-2G as bracketing standard to Further key issues in the development of this technique was: correct for any instrumental drift, instrumental mass fraction- the asymmetric quad lens voltage settings on the NuPlasma HR ation (typical a factor $1.20) and referencing to LSVEC (NIST to achieve stable 7Li and 6Li flat-topped peak alignment; and, the 8545, 7Li/6Li 12.1716) (see Fig. 2).
use of the experimental low-mass, high-abundance skimmer conewhich yielded significantly increased sensitivity.
As can be seen in Table 2 and Fig. 2, the Li isotope compositionsof the USGS and MPI glass standards obtained in this study with The assistance, especially by Ian Bowen, of Nu Instruments and the new LA-MC-ICP-MS technique reproduced the published the provision of the low-mass, high-abundance cone are greatly appreciated. This work benefited from discussions with Anton le The LA-MC-ICP-MS data in Fig. 2 fall mostly on the low Roex and Chris Harris, as well as insightful and constructive side of the solution MC-ICP-MS literature data, with the reviews. This is AEON contribution 73.
exception of data from MPI ATHO-G. This does not appear tobe due to variable measured Li signal intensities, as the measured voltages listed in Table 2 show. By using the source slit to 1 T. Zack, P. B. Tomascak, R. L. Rudnick, W. F. McDonough and moderate the Li signal as discussed earlier, analysis of the high Li e, Earth Planet. Sci. Lett., 2003, 208, 279–290.
concentration standards (e.g. $0.13 V for 28.6 ppm Li MPI 2 C. Bouman, T. Elliot and P. Z. Vroon, Chem. Geol., 2004, 212, 59–79.
ATHO-G) did not involve a significantly higher Li signal 3 T. Elliot, A. B. Jeffcoate and C. Bouman, Earth Plan. Sci. Lett., 2004, compared to lower Li concentration standards (e.g. $0.1–0.12 V 4 T. Elliot, A. Thomas, A. B. Jeffcoate and Y. Niu, Nature, 2006, 443, for 5.1 ppm Li MPI KL-2G). Furthermore, there is no systematic This journal is ª The Royal Society of Chemistry 2010 J. Anal. At. Spectrom., 2010, 25, 1033–1038 | 1037 5 P. B. Tomascak, in Rev. Mineral., ed. C. M. Johnson, B. L. Beard and 10 J. S. Kane, Geostand. Geoanal. Res., 1998, 22, 7–13.
ede, Mineralogical Society of America, Geochemical Society, 11 N. J. G. Pearce, W. T. Perkins, J. A. Westgate, M. P. Gorton, Washington, 2004, vol. 55, pp. 153–195.
S. E. Jackson, C. R. Neal and S. P. Chenery, Geostand. Geoanal.
6 S. A. Kasemann, A. B. Jeffcoate and T. Elliot, Anal. Chem., 2005, 77, 12 I. Horn, R. W. Hinton, S. E. Jackson and H. P. Longerich, Geostand.
7 K. P. Jochum, B. Stoll, K. Herwig, M. Willbold, A. W. Hofmann, M. Amini, S. Aarburg, W. Abouchami, E. Hellebrand, B. Mocek, 13 C. C. Garcia, H. Lindner and K. Niemax, J. Anal. At. Spectrom., I. Raczek, A. Stracke, O. Alard, C. Bouman, S. Becker, M. Ducking, H. Bratz, R. Klemd, D. de Bruin, D. Canil, D. Cornell, C.-J. de Hoo, 14 D. L. Hoffmann, D. A. Richards, T. R. Elliot, P. L. Smart, C. Dalpe, A. Eisenhauer, Y. Gao, J. E. Snow, N. Groschopf, C. D. Coath and C. J. Hawkesworth, Int. J. Mass Spectrom., 2005, d. Gunther, C. Latkoczy, M. Guillong, E. H. Hauri, H. Hofer, Y. Lahaye, K. Horz, D. Jacob, S. Kasemann, A. J. R. Kent, 15 S. Richter, S. A. Goldberg, P. B. Mason, A. J. Traina and T. Ludwig, T. Zack, P. Mason, A. Meixner, M. Rosner, K. Misawa, J. B. Schwieters, Int. J. Mass Spectrom., 2001, 206, 105– B. Nash, J. Pfander, W. Premo, W. Sun, M. Tiepolo, R. Vannucci, T. Vennemann, D. Wayne and J. D. Woodhead, Geochem., 16 H. P. Qi, P. D. P. Taylor, M. Berglund and P. De Bi Geophys., Geosyst., 2006, 7, Q02008.
Spectrom. Ion Processes, 1997, 171, 263–268.
8 P. J. le Roux, S. B. Shirey, L. Benton, E. H. Hauri and T. D. Mock, 17 M. Guillong and D. Gunther, J. Anal. At. Spectrom., 2002, 17, 831– 9 K. P. Jochum, D. B. Dingwell, A. Rocholl, B. Stoll, A. W. Hofmann, 18 D. Gunther and J. Koch, in Laser ablation ICP–MS in the earth S. Becker, A. Besmehn, D. Bessette, H.-J. Dietze, P. Dulski, sciences: Current practices and outstanding issues, (ed. P. Sylvester), J. Erzinger, E. Hellebrand, P. Hoppe, I. Horn, K. Janssens, Mineralogical Association of Canada, Vancouver, 2008, vol. 40, pp.
G. A. Jenner, M. Klein, W. F. McDonough, M. Maetz, K. Mezger, C. Munker, I. K. Nikogosian, C. Pickhardt, I. Raczek, D. Rhede, 19 H.-R. Kuhn and D. Gunther, J. Anal. At. Spectrom., 2004, 19, 1158– H. M. Seufert, S. G. Simakin, A. V. Sobolev, B. Spettel, S. Straub, L. Vincze, A. Wallianos, G. Weckwerth, S. Weyer, D. Wolf and 20 M. Guillong, K. Hametner, E. Reusser, S. A. Wilson and D. Gunther, M. Zimmer, Geostand. Geoanal. Res., 2000, 24, 87–133.
Geostand. Geoanal. Res., 2005, 29, 315–331.
1038 | J. Anal. At. Spectrom., 2010, 25, 1033–1038 This journal is ª The Royal Society of Chemistry 2010

Source: http://aeon.org.za/content/pdf/publications/073_2010_lithium%20isotope%20analysis%20of%20natural%20and%20syntehtic%20glass.pdf