Preview only show first 10 pages with watermark. For full document please download

Determination Of Multiple Trace Element Compositions In Thin (> 30 ?m) Layers Of Nist Srm 614 And 616 Using Laser Ablation-inductively Coupled Plasma-mass Spectrometry (la-icp-ms

To understand and/or avoid small-scale chemical heterogeneities within geological materials prepared as normal thin sections, in situ multiple trace element determination coupled with the simultaneous microscopic observation of the sample during

   EMBED


Share

Transcript

   The chemical compositions of rock-forming mineralshave been investigated to understand the chemicaland physical processes related to rock formationand alteration. In situ  trace element and isotopic Determination of Multiple Trace Element Compositionsin Thin (< 30 µm) Layers of NIST SRM 614 and 616Using Laser Ablation-Inductively CoupledPlasma-Mass Spectrometry (LA-ICP-MS) Vol. 29 — N°1 p.107-122  To understand and/or avoid small-scale chemicalheterogeneities within geological materials prepared as normal thin sections, in situ  multipletrace element determination coupled with the simultaneous microscopic observation of the sampleduring analysis is preferable. We have examinedfifty trace elements in thin (< 30 µm) layers of theNIST SRM 614 and 616 glass reference materials by LA-ICP-MS using different pit diameters and internalstandard elements (Ca and Si). Compositional heterogeneities of Tl, Bi, As and Cd were found in NIST SRM 614 and 616 at the spatial resolutionof ca  . 100 µm. Except for these elements, the RSDsof six determinations for most elements were better than 10% in NIST SRM 614 when ablation diameters were > 50 µm. The measured concentrations for most elements in NIST SRM 614 and 616 agree  with previous values in the literature at the 95%confidence level with the exception of W and Bi.New LA-ICP-MS data for K, As and Cd are alsoreported. The results support the view that the latest LA-ICP-MS is a powerful and flexible analyticaltechnique for the determination of multiple ultra-trace element compositions in geologicalmaterials prepared as normal thin sections of thetype that has been used for polarising opticalmicroscopic observations since the end of the 19th century. Keywords: LA-ICP-MS, thin sections, trace elements,NIST SRM 614 and 616. Pour comprendre et/ou éviter les problèmes d’hétérogénéités à petite échelle existant dans les matériaux géologiques qui se présentent en lames minces, l'analyse multi-élémentaire des éléments entrace couplée à l'examen microscopique simultanée est une excellente solution. Nous avons analysé par LA-ICP-MS cinquante éléments dans des tranches fines (< 30 µm) des verres de référence NIST SRM614 et 616, en utilisant des diamètres d'analyse variables et Ca et Si comme standards internes. Des hétérogénéités dans la répartition de Tl, Bi, As et Cd ont été observées dans NIST SRM 614 et 616 à une résolution spatiale de 100 µm. Mis à part ces éléments, les RSD obtenues sur six déterminations de tous ces éléments étaient meilleures que 10%dans NIST SRM 614 pour un diamètre d'impact > 50 µm. Les concentrations mesurées pour la plupart de ces éléments dans NIST SRM 614 et 616 concordent avec les valeurs publiées, avec un degré de confiance de 95%, à l'exception de W et Bi. De nouveaux résultats de K, As et Cd obtenus par LA-ICP-MS sont aussi présentés. Les résultats confirment que les dernières LA-ICP-MS sont des outils analytiques puissants et adaptés à l'étude de compositions multi-élémentaires (traces et ultra- traces) dans des matériaux géologiques préparés sous forme de lames minces de l'épaisseur classiquement utilisée en microscope polarisant depuis la fin du 19ème siècle. Mots-clés : LA-ICP-MS, lame mince, éléments en trace,NIST SRM 614 et 616. 107 0305  Tomoaki Morishita (1) *, Yoshito Ishida (2)  , Shoji  Arai (2) and Miki Shirasaka (2) (1) Graduate School of Natural Science and Technology, Kanazawa University, Kakuma, Kanazawa 920-1192, Japan(2) Department of Earth Sciences, Kanazawa University, Kakuma, Kanazawa 920-1192, Japan*Corresponding author. e-mail: [email protected] Received 02 Mar 04 — Accepted 16 Sep 04 GEOSTANDARDS and RESEARCH  GEOANALYTICAL   compositions of minerals can provide fundamentalinformation to solve diverse geological problems. Oneof the best developed techniques for in situ  trace ele-ment analysis is secondary ionisation mass spectrome-try (SIMS). In SIMS analysis, however, the analyticaltime is relatively long, and complex matrix-relatedinterference corrections are often required to obtainaccurate results. In the last decade, laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) has emerged as a powerful and flexible analyticaltechnique for direct elemental and isotopic determina-tion of solid materials. Advantages of the LA-ICP-MStechnique include its rapidity, spectral simplicity andrelatively straightforward data reduction. LA-ICP-MS isthus becoming widely used for geological samples to yield large amounts of precise geochemical information. A significant difference between SIMS and LA-ICP-MS is the volume of material sampled during an anal- ysis, with laser ablation consuming significantly morematerial than SIMS. This is particularly significant in thecontext of analysing geological material prepared inthe form of thin sections, which are only 30 µm thick.Optical microscope observation of thin sections permitsidentification of the minerals present, investigation of their textural relations and heterogeneity within grainssuch as zoning, exsolution and the presence of inclu-sions. When combined with capabilities for in situ  micro-analysis this represents a powerful approach tounderstanding geological processes. Studies of theoptimisation of analytical conditions for the LA-ICP-MSof geological materials prepared as normal thin sec-tions have been limited (e.g.,  Jackson et al. 1992),because of the ease of penetrating such thin samplesby repetitive sampling with high powered lasers. LA-ICP-MS analyses have usually been applied to relati- vely thick sections or minerals mounted in an epoxy disk so that they can be ablated for the longer timesrequired to improve precision. Trace elements in the NIST SRM glass referencematerials 614 and 616, which are synthetic silicateglasses with a nominal concentration of 1 µg g -1 and20 ng g -1 for sixty-one elements, respectively, are use-ful as reference materials for microbeam determina-tions of ultra-trace elements in geological materials as well as for demonstrating the accuracy and precisionof determinations. However, there is little publisheddata for these materials, particularly NIST SRM 616.High laser pulse rates (e.g., 10 Hz) coupled with longablation times have usually been used for LA-ICP-MSanalysis of these materials (e.g., Horn et al. 1997,Kurosawa et al. 2002, Gao et al. 2002). The latest generation of quadrupole ICP-MS instruments coupled with a 193 nm excimer laser system offer lower back-grounds and higher sensitivity, and raise the possibility of accomplishing quantitative analysis at ultra-tracelevels from small amounts of material.In this study, we measured multiple trace elementscompositions covering the mass range from 7 Li to 238 Uin thin (< 30 µm thick) layers of the NIST SRM 614 and616 by LA-ICP-MS using different pit diameters. Theresults allowed us to examine the presence of localdepletion and/or enrichment of elements in NIST SRM614 and 616. These concentrations were compared with the previous values from the literature to evaluatethe ability to analyse trace element compositions fromsmall amounts of samples. New LA-ICP-MS data for Cd, As and K in NIST SRM 614 are included in thedata presented here, noting that these are important elements, especially in the environmental sciences. Experimental Instrumentation and operating conditions Quartered NIST SRM glass wafers (614 and 616) were embedded in 25 mm diameter epoxy discs and were polished using 1 µm diamond paste to producea flat, scratch-free surface that was suitable for anal- ysis. Determinations were performed at the IncubationBusiness Laboratory Centre of Kanazawa University,using a quadrupole ICP-MS (Agilent 7500s) equipped with a commercially available laser-ablation system(MicroLas: GeoLas Q-Plus) (Ishida et al. 2004). The GeoLas Q-Plus used an argon fluoride gasmixture to produce 193 nm laser light and was equip-ped with a beam homogeniser and aperture-imagingoptical system. The diameter of the ablation spot couldbe varied from 4 to 160 µm across, depending on thesize of the aperture used. In this study we used four different pit sizes (30, 50, 70 and 100 µm) for anal- ysis. Due to the homogeneous illumination of the aper-ture, the energy density on the sample surface wasconstant at all beam sizes, leading to the projection of a flat top beam onto the sample surface. The laser fluence at the sample surface could be changed by  varying the discharge voltage of the laser and by using a beam splitter, with accessible energy on thesample ranging up to 35 J cm -2 . Craters on the surfaceof a normal slide glass used for thin section making were cut vertically through the centre after 500 and 108 GEOSTANDARDS and RESEARCH  GEOANALYTICAL   1000 pulses at an energy of 10 J cm -2 on the samplesurface and were then observed using an opticalmicroscope (Ishida et al. 2004). As a result, the rate of material removal was estimated to be approximately < 0.2 µm/pulse. Günther et al. (1997) suggested that the ablation rate with a 193 nm excimer laser systemis relatively matrix-independent when comparing theNIST SRM 612 and natural minerals. Therefore, 150pulses were selected for all the analyses in this study;i.e., 30 seconds were available for data acquisition at a laser pulse repetition rate of 5 Hz (Figure 1). Thedepth of the holes we drilled was simply confirmed by optical microscope and was less than 30 µm. The ablation cell used allowed a normal thinsection and two external reference glasses (e.g., NIST SRM 610 and 612) to be loaded and analysed in asingle uninterrupted session. In order to increase theefficiency of aerosol transport into the ICP-MS system(Bleiner and Günther 2001), the volume of the abla-tion cell was reduced (from ca. 38,000 mm 3 to 5,600mm 3 ) by an in-house constructed sample folder madeof acrylic resin and designed to fit the ablation cell. A high-quality CCD camera and LCD monitors wereincorporated into the laser system and allowed bothtransmitted- and reflected-light viewing when sear-ching for points to be analysed. All data were obtained by ablating in a He atmos-phere ( ca. 0.2 l min -1 ) prior to combining with thedominant Ar carrier flow ( ca  . 1.2 l min -1 ). Helium wasused to minimise post-ablation surface condensationand to maximise sample transport efficiency into theICP (Eggins et al. 1998a , Günther and Heinrich 1999,Ishida et al. 2004). Platinum sampler and skimmer cones were used for all of the analyses. The operatingparameters of the instrumentation used in this study are summarised in Table. 1. We usually optimised the ICP-MS using a signalobtained from the ablation of NIST SRM 612 glass witha laser raster procedure to facilitate invariant signalintensities over the relatively long time used for tuningof instrumental conditions. Taking into account the massablation rate, instrument sensitivity was tuned to giveapproximately 4000 cps/µg g -1  , 10000 cps/µg g -1 and 5000 cps/µg g -1 for 7 Li, 89  Y and 209 Bi, respectively, when ablating a 70 µm circular spot at a laser pulserepetition rate of 5 Hz. 248  ThO/ 232  Th was maintainedbelow 0.5%, and other potentially interfering oxides 109 GEOSTANDARDS and RESEARCH  GEOANALYTICAL  101001000100000 20 40 60 80 100 120Y89Pb208Th232U238Laser-on Laser-off Figure 1. Typical LA-ICP-MS calibration spectra of 89  Y, 208 Pb, 232  Th and 238 U showing intensity (counts per second) vs time onNIST SRM 614 glass using a 50µm pit diameter. Backgrounds onthe dry plasma were collected for 50-60 seconds prior to ablatingfor 30 seconds. Ablation wasstopped at about 90 seconds,and the signal returned to background levels after about 10 seconds. Time (seconds)    C  o  u  n   t  s  p  e  r  s  e  c  o  n   d    Table 1. Typical operating conditions for the LA-ICP-MS method in this study  ICP-MS Model7500s (Agilent)Forward power1200 W Reflected power1 W Carrier gas flow1.16 l min -1 (Ar)0.2 l min -1 (He) Auxiliary gas flow1.0 l min -1 Plasma gas flow15 l min -1 ConesPt sample conePt skimmer cone Laser  ModelGeoLas Q+ (MicroLas) Wavelength193 nm, ArFRepetition rate5 HzEnergy density at target10 J cm -2  110 GEOSTANDARDS and RESEARCH  GEOANALYTICAL   Table 2. Analyte elements, isotopes, dwell time per element, observed sensitivities, background countsrates and calculated detection limits (Ca-normalised) for LA-ICP-MS using different pit diameters ElementIsotopeD.T.BG cpsSensitivity (cps/µg g -1  )Detection limit (µg g -1  )(ms)(100)(70)(50)(30)(100)(70)(50)(30) Li 7 25 300 8219 4238 2206 954 0.011 0.019 0.039 0.12Be 9 50 20 1095 583 312 138 0.014 0.025 0.056 0.12Mg 24 25 250 5804 3003 1566 683 0.013 0.027 0.42 0.14Si 29 10 300000 169 87 46 21 51 83 167 480K 39 25 150000 13752 6951 3628 1608 0.17 0.18 0.31 0.31Ca 42 10 20000 114 59 31 14 11 20 50 127Sc 45 25 2000 13023 6804 3612 1646 0.018 0.034 0.067 0.17 Ti 47 25 150 898 460 245 111 0.043 0.090 0.18 0.43 V 51 25 700 15180 7757 4080 1852 0.008 0.017 0.031 0.085Cr 53 50 5000 1263 645 343 159 0.21 0.37 0.77 2.1Mn 55 25 25000 15800 8073 4328 1952 0.061 0.13 0.20 0.47Co 59 25 1300 12195 6129 3326 1502 0.017 0.033 0.049 0.14Ni 60 25 150 2531 1286 707 321 0.027 0.050 0.084 0.21Zn 66 25 200-1300 1307 691 401 187 0.070 0.12 0.35 1.1Ga 69 25 200 11146 5684 3044 1374 0.007 0.013 0.025 0.062Ge 72 25 2000 4208 2137 1157 523 0.053 0.10 0.18 0.51 As 75 50 350 1451 729 413 191 0.051 0.075 0.14 0.42Rb 85 25 800 16248 8153 4363 1992 0.009 0.018 0.034 0.086Sr 88 25 250 19958 10203 5404 2488 0.005 0.008 0.015 0.04 Y 89 25 20 18241 9430 5084 2412 0.001 0.002 0.004 0.011Zr 90 25 20 9540 4960 2682 1270 0.002 0.004 0.006 0.020Nb 93 25 20 17784 9086 4918 2293 0.001 0.002 0.004 0.010Mo 95 25 25 2968 1487 815 370 0.008 0.018 0.031 0.074Cd 111 50 100 956 512 295 146 0.037 0.068 0.12 0.28In 115 25 25 18827 9444 5228 2333 0.001 0.003 0.005 0.014Sn 118 25 100 5383 2795 1458 543 0.010 0.021 0.038 0.091Sb 121 25 125 4686 2529 1344 506 0.013 0.025 0.046 0.11Cs 133 25 100 22516 11632 6170 2314 0.005 0.009 0.014 0.038Ba 137 50 25 2605 1357 720 266 0.013 0.012 0.024 0.19La 139 25 20 20092 10610 5720 2155 0.001 0.002 0.004 0.010Ce 140 25 20 21643 11267 5972 2237 0.001 0.002 0.003 0.010Pr 141 25 20 24307 12798 6749 2558 0.001 0.002 0.003 0.008Nd 146 50 20 3760 2003 1056 407 0.004 0.008 0.015 0.039Sm 147 50 20 3084 1650 864 333 0.006 0.011 0.017 0.041Eu 153 25 20 11836 6334 3329 1255 0.002 0.004 0.007 0.017Gd 157 50 20 2798 1524 802 313 0.006 0.011 0.019 0.048 Tb 159 25 25 18818 10220 5389 2077 0.001 0.002 0.004 0.010Dy 163 50 25 4216 2298 1206 472 0.005 0.008 0.015 0.037Ho 165 25 25 16701 9107 4820 1853 0.002 0.003 0.006 0.014Er 166 50 25 5609 3059 1623 623 0.003 0.006 0.011 0.027 Tm 169 25 25 16609 9117 4853 1829 0.002 0.003 0.005 0.011 Yb 172 50 25 3527 1937 1024 395 0.005 0.008 0.016 0.050Lu 175 25 25 14681 8077 4300 1666 0.002 0.003 0.006 0.013Hf 178 25 25 4423 2418 1300 494 0.005 0.010 0.019 0.045 Ta 181 25 20 12798 6966 3692 1424 0.002 0.004 0.007 0.017 W 182 25 25 3294 1756 934 347 0.007 0.012 0.027 0.062 Tl 205 50 45 9090 5026 2593 1018 0.003 0.006 0.010 0.063Pb 208 25 90 6276 3431 1829 689 0.007 0.016 0.028 0.068Bi 209 25 40 11710 6437 3461 1294 0.003 0.005 0.010 0.025 Th 232 25 40 9769 5414 2947 1128 0.003 0.005 0.009 0.14U 238 25 30 11490 6305 3395 1277 0.003 0.004 0.008 0.021 D.T. dwell time. BG background count. Number in parenthesis pit diameter (µm).   were assumed to be negligible, based on the relativeease of Th oxide production (Leichte et al. 1987).  Analytical elements and data reduction Dwell time and the number of elements to be deter-mined are important parameters in optimising dataacquisition procedures (Günther et al. 1999). Data werecollected by peak hopping in time-resolved mode tomonitor possible compositional heterogeneities. We chosedwell times of 10 ms for internal standard elements (i.e., 29 Si and 42 Ca) and 25 or 50 ms for other elementsdepending on sensitivity and relative isotope abundance(settling time is approximately 1.8 ms; Table 2). In practice,about thirty isotopes covering light to heavy masses were flexibly selected for determination depending onthe analytical objectives. In order to resolve chemicalheterogeneity effects in samples, the shortest sweep timeis required. In order to reduce sweep time (< 1000 ms),the isotopes were divided into two groups based onmass number ( 7 Li to 115 In and 118 Sn to 238 U). The sameanalytical conditions were used for these two sets. All signal intensities were corrected for the back-ground signal obtained from measurement of a gasblank for 50-60 seconds prior to initiating ablation(Figure 1). Typical sensitivity and gas blank for eachelement are given in Table 2. The reason is not clear for the wide range of background count rates for Zn(200-1300 cps). After the end of the ablation period,the signal usually returned to the background levelafter 10 seconds (Figure 1). Total analysis time wasthus 120 seconds per spot including backgrounds and washout of the sample prior to the next analysis.Data reduction followed a protocol essentially identical to that outlined by Longerich et al. (1996). The use of internal standards produces a much morerobust calibration method by allowing a correction tobe applied for differing ablation yields betweensample and reference material. Previous studies haveindicated that the ablation rate with excimer systems isrelatively matrix-insensitive when comparing NIST SRM612 and minerals (hornblende, augite and garnet;Günther et al. 1997). Compositions of minerals, cali-brated against NIST SRM 612 or 610 using a major element as an internal standard, agreed well withindependent data obtained by other methods despitethe very considerable difference in matrix between theNIST SRM glasses and the samples ( Jackson et al. 1992, Fedorowich et al. 1995 , Ludden et al. 1995,Norman et al. 1996, 1998, Günther et al. 1997, Eggins et al. 1998b). Both 42 Ca and 29 Si were used as inter-nal standards to correct for variations in the absoluteamount of material that was ablated and transportedduring analysis. Major element compositions of NIST SRM 614 and 616 were determined using a JEOL JXA-8800 Superprobe at the Centre for CooperativeResearch of Kanazawa University (Table 3). The externalcalibration sample was the NIST SRM 612 referencematerial. Many investigations have concluded that theNIST SRM glasses are essentially homogeneous inmajor and trace elements (Si, Al, Ca, Na, Rb, Cs, Sr,Ba, Sc, Y, REE, Ti, Zr, Hf, Nb, Ta, Pb, Th and U), but com-positional heterogeneity problems have been reportedfor some elements (Hinton et al. 1995 , Norman et al. 1996 , Rocholl et al. 1997, Kane 1998, Hinton 1999 , Sylvester and Eggins 1997, Eggins and Shelley 2002).In this study, the elemental concentrations of NIST SRM612 used for calibration were selected from the prefer-red values of Pearce et al. (1997), because these datahave been widely used in the external calibration of many LA-ICP-MS analyses of NIST SRM glasses as wellas geological materials (e.g., Horn et al. 1997, Mason 111 GEOSTANDARDS and RESEARCH  GEOANALYTICAL   Table 3.Major element compositions (% m/m) of NIST SRM 614 and 616 determined by EPMA  NIST SRM 614n = 22NISTSRM 616n = 20averagesdaveragesd SiO 2  71.83 0.44 72.35 0.23 Al 2 O 3  1.95 0.03 1.95 0.03CaO 11.71 0.08 11.75 0.09Na 2 O 13.78 0.08 13.84 0.15total 99.27 0.48 99.89 0.35 Major element compositions were determiend using a JEOL JXA-8800 microprobe at the Cooperative Centre of Kanazawa University. The analyses were performed under an accelerating volatage of 15 kV and a beam current of 15 nA using 30 µm diameter beam. JEOL software using ZAF corrections was employed. No loss of Na was observed during the analysis under these analytical conditions.n number of analyses. sd one standard deviation.