Documents

A Simple Method for the Precise Determination of -40 Trace Elements

Description
determinacion por ICP-MS ETR, en matriz acuosa
Categories
Published
of 16
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  n Zi- ELSEVIER Chemical Geology 134 (1997) 311-326 CHEMIC L GEOLOGY IN LUDING ISOTOPE GEOSCIENCE A simple method for the precise determination of > 40 trace elements in geological samples by ICPMS using enriched isotope internal standardisation S.M. Eggins *'1, J.D. Woodhead 2, L.P.J. Kinsley, G.E. Mortimer, P. Sylvester, M.T. McCulloch, J.M. Hergt 3, M.R. Handler Research School of Earth Sciences Australian National University Canberra A.C.T. 0200 Australia Received 9 August 1995; accepted 22 June 1996 Abstract The combination of enriched isotopes and conventional elemental internal standards permits the precise determination of > 40 trace elements by ICPMS in a broad spectrum of geological matrixes. Enriched isotopes expand the suite of available reference isotopes spaced through the mass spectrum, so that the complex mass-dependent variations in sensitivity encountered during ICPMS analysis can be monitored and deconvolved. The method we have developed is straightforward, entailing simple sample preparation, instrument calibration, and data reduction procedures, as well as providing extended element coverage, improved precision, and both time and cost benefits compared to alternative analytical strategies. Analytical precision near or better than 1% RSD (relative standard deviation) is achieved for most dements with mass > 80 amu and between 1% and 4% (RSD) for elements with mass < 80 amu, while maintaining low detection limits (< 1 to < 10 ng g- l for elements with mass > 80 amu and < 10 ng g- i to 1 Ixg g- i for elements with mass < 80 amu). The subtle geochemical differences which can be resolved using this method are demonstrated by analyses of Nb, Ta, Zr, and Hf in magmas from ocean islands and subduction zones. These data reveal significant departures from chondritic Zr/Hf and Nb/Ta values, and systematic trends which are consistent with greater incompatibility of Zr relative to Hf and also of Nb relative to Ta during melting of the upper mantle. The occurrence of significantly subchondritic Zr/Hf and Nb/Ta ratios in Nb-poor subduction zone magmas, supports the notion that the depletion of high-field strength elements in subduction magmas is due to their removal from the mantle wedge by prior melting events. Keywords: Inductively coupled plasma methods; Mass spectrometry; Trace element analysis; Subduction zones; Partial melting I. Introduction * Corresponding author. t Present address: DepaJament of Geology, Australian National University, Canberra, A.C.T. 0200, Australia. 2 Present address: School of Earth Sciences, University of Melbourne, Parkeville, Vic. 3052, Australia. 3 Present address: DepaJ:tment of Geology, Australian National University, Canberra, A.C.T. 0200, Australia. ICPMS (inductively coupled plasma mass spec- trometry) has rapidly become established as a pre- ferred method for the analysis of trace elements in geological samples, offering rapid analysis capabili- ties, an ability to measure most elements, relatively straightforward sample preparation, very low detec- 0009-2541/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved. PH S0009-2541 (96)00100-3  312 S.M. Eggins et al. / Chemical Geology 134 1997) 311-326 tion limits, a large dynamic range, and largely inter- ference-free spectra. However, it is increasingly ap- parent that fully quantitative measurement by ICPMS can be compromised by variation in instrument sen- sitivity during the course of an analytical run and with the introduction of different sample matrixes. Accurate and precise data may only be obtained if these effects can be minimised or, monitored and corrected for. The adverse effects of sensitivity variations dur- ing ICPMS analysis can be avoided or minimised by employing either standard addition or isotope dilu- tion (e.g., Jenner et al., 1990; Xie et al., 1994). However, these methods have drawbacks which compromise the inherent capabilities of ICPMS. Iso- tope dilution requires accurately prepared and main- tained enriched isotope spikes, and optimal spik- ing of individual samples. Furthermore, it is limited to multi-isotope elements and, when applied to a large number of element determinations, is expensive and time consuming. Standard addition likewise re- quires preparation of accurate elemental standard solutions, employment of optimal spiking strategies on a sample by sample basis, and the aliquoting of sample solutions in order to conduct the multiple analyses necessary for analyte determination. Faced with the rigors of standard addition and isotope dilution, many ICPMS laboratories have re- sorted to external calibration (using either synthetic or natural reference materials) with internal and/or external standardisation. These alternate strategies offer considerable time, sample preparation, and cost benefits but until now have not delivered good ana- lytical reproducability due to the complex sensitivity variations encountered during analysis. The strate- gies employed with most success to date include: (1) the use of multiple elemental internal standards (Thompson and Houk, 1987; Doherty, 1989); (2) relatively short analytical procedure duration; and (3) external standardisation based on multiple repeats of the same solution run at intervals during an analyti- cal procedure (e.g., Cheatham et al., 1993), or a combination of the above (e.g., Schonberg, 1993). It should be noted that most existing multi-ele- ment ICPMS studies on geological samples fall well short of the analytical precision capabilities of mod- em ICPMS instruments. The ability to measure iso- tope ratios to a precision of ~ 0.2-0.3% RSD (rela- tive standard deviation), should translate into an ability to reproduce measurements to + 1% or better, given cumulative sample preparation, analytical methodology, and counting statistic errors. This can be compared to reported precision usually in the range 3-5% RSD (Jenner et al., 1990; Cheatham et al., 1993; Schonberg, 1993; Xie et al., 1994), and sometimes > 10% or 20% despite favourable con- centration levels. Herein we describe a straightforward method em- ploying external calibration along with a combina- tion of external standardisation and multiple internal standardisation, using both elemental and enriched isotope internal standards. Improved precision and extended element coverage (> 40 elements) is at- tained compared to other ICPMS analytical strategies reported to date. Moreover, because mass-dependent variations in instrument sensitivity are monitored and deconvolved on a sample by sample basis, the method is matrix tolerant and does not necessitate strict matrix-matching protocols. It has been applied suc- cessfully to peridotites, a spectrum of magma types ranging from komatiite through basalt to rhyolite, granitic samples, banded iron formations, and min- eral separates including olivine, pyroxene, spinel, feldspar, various oxide phases, and apatite. 2. Mass-dependent sensitivity variation during ICPMS analysis Initial experience with our instrument (Fisons PQ2 + ) employing a single internal standard (ele- mental In) revealed an inability to reproduce deter- minations on the same sample within analytical er- ror, even during a single analytical run. The relative differences between repeated analyses were found to be a systematic but often complex function of mass. This behaviour is symptomatic of variations occur- ring in instrument sensitivity as a function of mass with time (Fig. 1), and has been described in detail by Cheatham et al. (1993) for the restricted mass region spanned by Ba and the rare-earth elements (REE). We have found that the magnitude and com- plexity of these sensitivity variations can be severe where the analysed isotopes span the entire mass spectrum, particularly in the low-mass region (< 80 amu). Interestingly, it appears that this problem is  S.M. Eggins et al. / Chemical Geology 134 1997) 311-326 313 ~ 1.0 -~ ~ 0.95 ---~ 0.90 O~ Q. O0 0.85 0.80 ._> ¢n 0.75 ,- CO 0.70 ._> -,,-- 0 65 r'¢ 0.60 aU IntscnalStsndards ~ Sr Rh In 47E;m Tm Re ~ ~ SU I Ot~ ~ ~aga~m o -t~ ........................ .............................. .......................... .............................. ............................... =.~ ................. -~.-~' ............. ............................ .--~,~v-i ........................ i I 5 li :_._;.._._._1._ ....~.~ ............... ............................ ....... # 12 I..: ~71 ~i ~: I # 191 : -,',, Ui-,,, ,'i .... i ........ 50 100 150 200 250 atomic mass units Fig. 1. Relative sensitivity changes for internal standard and analyte isotopes during the course of an analytical run, based on repeat analyses of the same solution at intervals during the run. #n denotes the analysis number of each repeat measurement within the analysis sequence. The continuous line is a linear interpolation between internal standard isotopes for analysis # 19. This compares to the dashed curve, which is a closer approxima- tion to the actual sensitivity changes, particularly in the low-mass region between 6Li and 84Sr. the principal limitation to quantitative analysis of many low-mass eleme.nts rather than molecular inter- ferences. Given the complex nature of the sensitivity varia- tions illustrated in Fig. 1, it can be readily appreci- ated that the conventional use of a single or even several elemental internal standards will be inade- quate to monitor and deconvolve these variations during the course of an analytical run. If internal standardisation is to be effective in such cases then multiple internal standards, spaced at appropriate intervals through the entire mass spectrum (i.e. ~ 20 mass units apart), are required. This presents an insurmountable difficulty for conventional internal standardisation using modern high-sensitivity instru- ments, as elemental internal standards are largely restricted to ultra-trace elements (e.g., Rh, In, Re, Bi) whose natural abundances do not contribute sig- nificantly (ideally < 0.1-0.2%) to any added inter- nal standard. Enriched isotopes not only expand the pool of useable internal standards but enable the abundance of their parent elements to be determined. The use of particular enriched isotopes is subject to their posi- tion and spacing within the mass range, stability in dilute nitric acid solutions, cost, and the limitations which they can place on the dynamic range and detection limits of parent element analyte isotopes. The latter is a product of the uncertainities arising from subtracting the natural and enriched isotope contributions from the internal standard and analyte isotopes, respectively. Multiple internal standards also contribute to im- proved analytical precision by offsetting the effects of higher frequency noise in ICPMS. For example, with a typical mass spectrometer sweep time of 1 s to traverse the mass spectrum, ten uniformly spaced internal standards will account for noise of < 10 Hz whereas a single internal standard will be limited to <1 Hz. 3. Analytical technique 3 1 Sample preparation and introduction Samples and geochemical reference materials are dissolved by conventional digestion methods, using concentrated HF-HNO 3 mixtures (10:1) in 25-ml Savillex screw-top Teflon beakers. Several times during the digestion, the beakers are placed in an ultrasonic bath in order to disaggregate granular material and render it more susceptible to acid at- tack. In the case of samples containing resistant phases, such as zircon, microwave or bomb di- gestions are employed to ensure complete dissolu- tion. Following digestion the samples are evaporated to incipient dryness, refluxed in 6 N HNO 3, taken again to incipient dryness, and the sample cake then dissolved in 2 ml of concentrated HNO 3. Final sam- ple preparation is undertaken, following transfer to 125-ml polypropylene bottles and addition of a known weight of internal standard solution, by dilu- tion with ultrapure H20 to a sample/solution weight ratio of 1 : ~ 1000-1250 (i.e. ~ 80-100-rag sample in 100 g of final solution). This dilution factor is a compromise between sample-size related heterogene- ity, ease of digestion, availability of large volumes of clean reagents, required detection limits, and analyte suppression effects which can be severe where total dissolved solid contents exceed 0.2%. All reagents  314 S.M. Eggins et al. / Chemical Geology 134 1997) 311-326 Table 1 Internal standard and analyte isotopes, interferences, and internal standard concentrations Analyte Isotope Significant interferences a6 (ng/g) Int. Std. 6Li 6 Li natural 7 10 Li 7 Li Int. Std. Be 9 Sc 45 COOH? Ti 49 V 51 Cr 53 Co 59 Ni 60 44 CAO49 Cu 65 TiO, 4s TiOHSO Zn 66 rio, 49 TiOHI42 Ga 71 Ce z+, 142Nd2+ Int. Std. 84Sr 8484 Sr natural, s4Kr 16 Rb 85 Sr 86 Sr Int. Std., ~Kr Y 89 Zr 91 Nb 93 Mo 98 Int. Std. Rh 103 n4 5 Cd 114 Sn Int. Std. In 115 10 Sn 120 Sb 121 Cs 133 Ba 137 La 139 Ce 140 Pr 141 Nd 146 147 Int. Std. 147Sm 147 Sm natural 149 10 Sm 149 Sm Int. Std. 13s Eu 151 BaO and 134BaOH143 Tb 159 NdO 144 Gd 160 SmO, 144NdO, 16097145 Dy 161 NdO 149 Ho 165 SmO Er 167 t51 Eul6OI69 Int. Std. Tm 169 Tm natural 10 Yb 173 157GdO Lu 175 J59TbO Hf 178 162DyO Ta 181 ~65HoO Int. Std. Re 187 5 T1 205 Pb 208 Int. Std. Bi 209 10 Th 232 235 Int. Std. Z35U 235 U natural 23a 2.5 U 238 U spike Table 1 (continued) Analyte Isotope Significant interferences a (ng/g) Other possible internal standards: 6 J Ni 61 44 CaOH t34Ba 134 135Ba 135 2°apb 204 a Bold typeface indicates corrected interferences. are prepared using double-distilled concentrated acids and ultra pure water (> 18 MII). Sample introduction to the ICPMS is automated via a Gilson Autosampler at a rate of ~ 1 ml/min- 1. The uptake time is set to facilitate stable analyte signals prior to analysis (typically 90-120 s). A washout procedure incorporating a dilute surfactant (e.g., ~ 0.5% Triton X-100 or DECON) and suffi- cient wash time is critical to achieving adequate instrument blank levels. A procedure appropriate for most applications involves a 60-90 s surfactant wash and a 180 s wash with 2% HNO 3. The total time for analysis of a single sample solution is ~ 10-12 min, resulting in typical run times of between 3 and 7 hr for the analysis of 18-36 solutions, of which 12-28 are unknowns, 2 are calibration solutions, 2-4 proce- dural blanks, and an instrument sensitivity monitor- ing solution which is repeated every 5-7 solutions. To assure low instrument blank levels, particu- larly for the analysis of low-level concentrations of elements which have poor washout characteristics (e.g., Nb and Ta), prior analysis of samples with high concentration levels of such elements is avoided. The spray chamber and nebuliser are also cleaned and replaced regularly along with the sample intro- duction tubing. Externally lubricated silicone tubing is employed on the peristaltic pump in preference to Tygon tubing, which develops internal cracking after a short period of use. It should be noted that mea- sured reagent blank levels for many elements with poor-washout characteristics are often not a true reflection of instrument blank levels during the intro- duction of analyte solutions, as these elements can be more efficiently scavenged from the sample intro- duction system than by blank 2% HNO 3 acid solutions (see also Xie et al., 1994).
Search
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks