419

The C anadian M ine rala gi s tVol. 33, pp. 419434 (1995)

APPLrcAilONS OF I-AM-ICP-MS ANALYSISTO MINERALS

JOHN N. LLIDDENI, RIN FENCI, GILLES GAUTHIER AND JOHN STD(

Ddpanement de g6obgie, Universitd de MontrAal, MofirAaL Case postale 6128, succursale Centre-ville,Montrdal, Quibec H3C 3J7

LANG SHI AND DONALD FRANCIS

Depanment of Earth and Pl.atatary Sciences, McGill University, 3450 university Steet, Montreal, Quebec H3A 2A7

NUNO MACHADO AND GUOPING WU

Ddpanement des Sciences d.e IaTerre, I|niyersitd du Qulbec d MontrbaL Mont6al, Qu.6bec H3C 3P8

ABSTRACT

Laser-ablation microprobe ICP-MS (LAM-ICP-MS) is rapidly emerging as an exciting new analytical tool in the earthsciences. Here we present the results of a series of isotopic and ftace-element studies of various minerals; we used acommercially available Fisons-VG laserprobe and PQ2+ ICP-MS capable of ablation at spot-sizes of 20 to 60 !tm. Results arepresented for Pb/Pb dating of zircon, for which an :rccuracy of. <IVo is attainable. For carbonate and mafic minerals of igneousorigin, detertion limits of 100 ppb and precisions of 2-5Vo are attained for many trace elements. In general, calibration ispossible usinij synthetic or natural standards; as elemental responses are controlled by the efftciency of the ionization of ablatedmaterial in the ICP-plasma matrix-mafched calibrations are not as imFortant as fot ion-probe analysis, owing to the efficiencyof ionization in the plasma torch. In the future, new high-sensitivity ICP - mass spectrometers, coupled with high-spatial-resolution UV laser microprobes, should provide detection limits at the ppt level and spatial resolution comparable with thatattained with an ion microprobe.

Keywords: laser ablation, inductively coupled plasma - mass specfometry, ICP-MS, trace elements, Pb isotopes, minerals.

Sotr/INaelRe

L'ablation au laser couplde i la technique de plasma i couplage inductif avec spectrom6trie de masse QCP-MS) est devenuerapidement un outil analytique performant dans les sciences de la terre. Nous pr6sentons ici les r6sultats de plusieurs dtudes surles compositions isotopiques et les teneus en 6l6ments traces de divers mindraux. Ces r6sultas oil 6t6 obtenus d partir d'unesonde laser Fisons-VG coupl6e I un appareil ICP-MS PQ2+; avec ce systbme nous pouvons produire des points d'ablafion parlaser de I'ordre de 20 i 60 pm. Les quelques donn6es gdochronologiques sur cristaux de zircon nous pennettent de calculer des6ges Pb/Pb d'une pr6cision infErieure d 17o. Pour les carbonates et les mindraux mafiques, les seuils de d6tection des 6l6mentstaces sont de l'ordre de I 00 ppb, et les pr6cisions , de 2 d 5Vo . Et g6ndral, le calibrage est possible h partir d'6talons synth6tiqueset naturels. Par rapport aux analyses obtenues par sonde ionique, les effets de matrice sur I'analyse sont moins importants. Ceciest d0 h I'efficacit6 d'ionisation dans la torche h plasma- A l'avenir, I'augmentation de la sensibilit6 des spectromBtres de masseet de la performance des sondes lasen i r6solution spatiale 6lev6e dewait permeffe d'atteindre des seuils de d6tection de l'ordredu ppt, et une rdsolution spatiale comparable I celle d'une microsonde ionique.

Mots-cl6s: ablation au laser, plasma d couplage inductif - spectrombtre de masse, ICP-MS, 6l6ments traces, isotopes de Pb,min6raur

I Present address: CRPG{NRS, 15, rue N.D. des Pauvres, BP 20, 54501, Vandoeuwe-lds-Nancy Cedex, France.2 hesent address: Institute of Sedimentary and Petroleum Geology, Geological Survey of Canada, 3303 33rd Street NW,Calgary, N&ra'nL2N.

420 TIIE CANADIAN MINERALOGIST

InrnonucrroN

Laser-ablation microprobe - inductively coupledplasma - mass spectrometry (LAM-ICP-MS) is ahighly sensitive technique capable of direct elementaland isotopic measurements in solid samples at the scaleof 100 pm or less (Gray 1985, Arrowsmith 1987,Hager 1989, Denoyer et al. l99L,Peuce et al. 1,992,Perktns et a\.1993, Longerich et al. L993). Focusing anincident Nd:YAG (1064 nm) laser results in qatersizes of 20-60 pm, depending on target material andablation time. Provided that sufftcient malerial can betransported to the plasma torch of the ICP-MS, quanti-tative analysis is possible for these small-volumecraters (i.e., Jackson et al. l992,Feng et al. 7993,Feng1994). Frequency quadrupling of the Nd:YAG laser toproduce a wavelength of 266 nm can result in craters of<10 pm (Longerich et al. L993), but rhe significantreduction in the volume of material carried to theplasma torch results in signals that are commonly atthe current limits of detection of the ICP-MS.Improvements in the sensitivity of the ICP-MS, theefficiency of ionization in the plasma torch, andthe means of tansport of the ablated materials arerequired in order to analyze such small volumes ofsample.

The processes of mechanical and chemical fraction-ation that occur during the interaction of the incidentlaser-beam and the sample are poorly unders0ood.Some of the processes have been summarized byReady (1971), Moenke-Blankenbvrg et al. (1990),Hager (1989) and Remond et al. (L990). Experimentsinvolving focused lasers of different energy, differentcarier gases *4 sample volumes for specificgroups of elements (i.e.,Feng et al. lgg3,Ludden et al.

1995) are required in order to provide quantitativeanalysis by LAM-ICP-MS. Nonetheless, the availabledata suggest that mass fractionation, matrix effects, andthe mechanical tansport of the ablated material areconholable, and LAM-ICP-MS can thus producequantitative trace-element and isotope data witlablation cralers 20-60 pm in size (Jackson et al.1992,Feng et al. 1993, Fryer et al. L993).

In this paper, we present a surnmary of results forminerals using a commercially available Fisons-VGLaserhobe and PQ2+ ICP-MS. Initial experimentscentered on (1) the analysis of zircon for U, Pb and Th,with potential applications in geochronology (Fenget al. 1993, Ludden et al. t995), (2) analysis ofcarbonate minerals and foraminiferal tests (Feng 1994,Wu & Hillaire-Marcel 1995), and (3) analysis of maficminerals in igneous rocks for trace elements and theestablishment of minerals as reference standards.

INsrrurrm:qrenoN

A Fisons-VG PQ2+ ICP mass spectrometer and aFisons-VG LaserProbe were used for this work. TheLaserProbe uses a Specton 300 mJ Nd:YAG laseroperating at l0& nm with a 5-nanosecond half-pulseduration for Q-swirched mode. The laser is attenuatedby inserting circular ceramic apertures into the beampath to reduce beam divergence, resulting in an outputenergy of 2-10 mJ per pulse, about l-5Vo of the uncol-limated laser output. However, power density at thetarget is maintained at the same level as (or higherthan) the uncollimated laser by reducing the area ofablation to <100 p"m in diameter in Q-swirched mode(e.g., Pearce et al. L992). The LAM-ICP-MS isoptimized for marimum count-rates by ablating NBS

Computer controlledbypass valve

Nd:YAG laser

t u i n>

IEEE toPlasmaQuad

Computer

Laser ManagementUnit

Ftc. 1. Schematic of theFisons-VG laser micro-probe system (adaptedfrom Fisons-VG bro-chure).

x,y Translatlon Stages

z Translaton(elevaton) Stage

Plasma Torch

APPLICATIONS OF I.AM-ICP-MS ANALYSF TO MINERALS 42L

Iass EodeI'l6hlaElr vollageIassr Flptl6o! EtaApprodmte lasot ouqutIeer Inwer al*dtjr

Saqrls plt stzsPls-ablalto! tlBeFru @ndttlon

Ertmadon lruColl*to! lsENsbullss flowC@lalt gaAudlerry gEBsotudm (an)U@udng porntsDwell d@

ICP. !dASS gPECTBOMETM

Q-srltobsd?@750 v4-!O Ez4-6 nllpulsez*toD tl grtolo wi oP20 - 80 lm aueEter2 0 sOa sEplg sutfre

0 . 8 - l . Yvar,{able1 .1 - 1 .3 L /d !1.4 L/d!1 - 1.2 Lldrappm. lamlEkwtdtiS ldnts wtl.h DAc stsfr of 12 0 - 3 0 6

TAAI.A 1. I,AM.ICP-W OPEBATING CONDITIONS (now NIST) silicate-glass reference material 6L2(NIST 612); tle usual operating conditions are listed inTable 1. d simFlified diagram of the instrument isshown in Figure 1. The instrument is optimized differ-ently depending on the isotopes of interest (e.g.,n8Pbfor zircon, 88Sr atrd l40ce for carbonateo and14Ce for 62fis minerals).

During ablation, the laser beam passes through thetop of a silica-glass sample cell and is focused ontothe sample, which is either an individual grain of amineral mountgd in resia. s1 l rhin section, The carriergas flows tbrougb the sample cell, fransporting theablated material through approximately 1.5 m of 3 mm(i.d.) polyurethane tube to the injector tube of thestandard ICP torch. The sample is held in an argonafinosphere, and as no vacuum is required changingthe sample is rapid. These operating conditions resultin sample pits of 20 to 60 trrm, depending on the

Frc. 2. Examples of laser-ablation craters in different materials: a) zircon, b) pyrope (two sizes of crater), c) foraminifera"and d) calcite.

422 THE CANADIAN MINERALOGIST

M.MT-NT BIIWK VAII'A LIMIT OT D&TECTION

TASLE 2. BI,ANK VAI,UES AND IJ!'TTS OT'DRTECTIONIN LAI[-ICP-I{I'

and Th, we used a 5-s acquisition-time repeated15-30 times for each laser-sampling spot @eng et al.1993). A statistical routine was used to eliminatespurious (outside 957o confidence level) blocks ofdataand average isotope and elemental ratios were takenover the same time-interval for different samples [seesection on zircon analy$is, and also Feng et al. (1993)1.For U, Th and Pb isotopes, a zircon crystal was used as astandard (CN-92-1); tlte change in the signal with timeafter a 20-s pre-ablation period is shown in Figure 3a.Several factors may cause variation in the amount ofsample ablated by the laser, with consequent variationin the observed signal. These include pulse-to-pulsevariation in laser energy, defocusing ofthe laser beamduring ablation, and differences in physical propertiesbetween samples (e.9., Arrowsmith 1987). To furtherevaluate these variations, Ludden et al. (L995) drd aseries of tests under different laser-ablation conditionson NIST 610 glass and zircon (CN-92-1). As shown inFigures 3a-c, the relative fractionation for differentisotopes of an individual element (ffiPb, zmPb, zoap6;is similar, and the ratios cau be measured at a precisionof <l7o (see section on zircon analysis). However,substantial fractionation occurs among differingelements (e.g-U, Pb and Th), ar shown by the mari-mum for the Th signal after 25-pm profiling, and for Uat 6f-170 prn. This results in element ratios that changein the ablation process. This relative fractionationduring prolonged ablation is contary to what would bepredicted from the vaporization energies of theelements: Pb > U > Th. However. these elements seemto fractionate according to their vaporization energiesin the 20-s pre-ablation stage of the analysis (Luddenet al. 1995). Once a high-temperature plasma isestablished above the crater, it appears that otherfactors (such as Pb diffrrsion iffo the crater) confrol therelative abundance of the elements in the plasma.Despite these problems, measurement on standardsshow that it is possible to measure 232Thl2MPb alldn8IJP06Pb with a precision of 2,-5Vo. Preliminarystudies show that elements of similar character li.e.,therare-earth elements (REE1, highly charged elements,alkaline earthsl do not exhibit this type of decouplingduring ablation.

Mass fractionation

For isotopes of the sarne elemento massdiscrimination may be caused by a bias inmeasurement due to Coulombic collision during iontransmission; generally, the heavier isotope isenhanced relative to the lighter isotope (Hieftje 1992).For an element with two or more stable isotopes, amass-fractionation correction can be done by internalnormalization, and it is usually less than l%o. Forelements where there are no stable pain of isotopes(i.e., Pb), mass-fractionation corrections must be doneby using an external standard. A similar method is used

Ll, 86 60 - 100 sIESc, V, C!, Co, Nl 600 - 30@Bb, Sr, Y 60 - 100Y,Zt 60 - 200Ba, BEE 18 - 20sg 5@-700P b 3 0 - 4 0u , T h 1 5 - 2 0

0.1 - 0.? ppn0.8 - 6 .00 . t - 0 . 20 . 1 - 0 , 80.01 - 0,021 . 0 - 2 . 00.02 - 0.040.01 - 0.02

material ablated and the duration of the ablation.Examples of sample pits in zircon, NIST glass, calciteand a foraminiferal test are shown in Figure 2. Ablationrates at a pulse energy of 5-6 mJ (750 V), for aninstrument operated at 44 Hzo are estimated at2040 p'mlmin. for zircon crystals and up to100 pm/nin. for glass samples. In most materialsanalyzed, these conditions are required to maintain anadequate signal in the ICP-MS. Lowering the laserenergy and repetition rate results in slower rates ofclriling and less material ablated, degrading sensitivityand detection limits. However, improvements inICP-MS sensitivity should allow ablation of smallervolumes using lower-energy UV and IR lasers in thefuture.

Sensitivities of the instrument operated in laser-abla-tion microprobe mode vary from 200 to 700 cps/ppmbased on results from NIST 612 glass. Detection limitsin ppm are estimated at three times the "gas blank"(analysis of argon carrier gas); for <50 F.m sample pits,limits of detection (l.o.d.) for NIST-612 glass varyfrom 0.01 ppm to 2-3 ppm (Table 2). It is possible rooperate the laser in a "bulk-sampling mode", producingcraters of 150 to 250 pra. In this mode, sensitivityincreases to 2000 - 10000 cps/ppm, with a resultant5- to lO-fold decrease i1 1tre limit of detection.Although usefirl for the analysis of glass and powdersamples (Fedorowich et al. l993,Pewce et al.1992),this mode is of limited use in studies of minerals owingto the poor spatial resolution of the laser.

Tnpammr oF DATA

Signal stability

Because of the transient nature of theLAM-ICP-MS signals, the ICP-MS is operated in apeak-jumping mode. The detection system scansindividual peaks for at least 40 ms (30 ms on peak andl0 ms on background) on low-abundance isotopesand 20 ms acquisition time for abundant or mono-isotopic species. In our early experiments for U, Pb

APPLICATIONS OF I-AM_ICP_MS ANALYSIS TO MINERAI,S 423

a t 4 6 ' 65 85 t06 125 145 166 185 A!5 25 26 26s 2E5 305

Time (seconds) afier ablation starts

20 N 60 80 100 120 llm 160

Laser drilling depth (pm)

Frc. 3. Variations in laser-ablation signal witl time andprofiling depth: a) relativevariation in the signal for U,Th and Pb isotopes in stan-dard zircon (CN-92-1) (fromFerg a al. 1993); b) relativeintensity variation in thesignal for standard zircon(CN-92-l), both fromLudden er al. (1995), afic) relative variation in thesignal for MST 610 glassrelative to profiling depth(calculated at a rate of0.05 pm per pulse at 750 V).

EO '67c

E(d

! ' 15 " / o'6goi -2tiTc'6

-E -gsyooE

4o/o

3Wo

2WoE

€ 1eh(U'H e/o

.e "loe/"anE -*o.S €0"/oo.E <et".EO -50"/oE,

ff/o

\J ,i.* i A . + .' . i i r A

41";;"'"'

.0.1

42

-0.3

Zirconstandard at 750 V and 4 HZ @

NBS610 at 750 V and 6 HZ

M4 TTIE CANADIAN MINERALOGIST

for correcting mass fractionation in ICP-MS analysisof solutions (Hinners et al. 1987). In laser ablation,if an external standard is used to calculate the mass-discrimination factor, we must assume that this factorremains constant for the standard and the nnknown.This may not be the case if the standard aud tneunknown have different elemental concentrations anddifferent absorption characteristics of IR and UV light.In LAM-ICP-MS Pb/Pb zircon geochronology, themass-fractionation factors applied are generally lessthan'17o.

Matrix-matched calibrarton and internal standard;

Unlike electron-microprobe analysis, for whichtheoretical calibration procedures are well established,techniques such as SIMS and LAM-ICP-MS analysisdepend upon empirical methods. These usually consistof extemal calibrations using well-established, matix-matched standards, coupled with the use of an intemalstandard to correct for variable yields during thesampling @erkins et al. 1992, 1993). The choice ofinternal standard is based on the followingconsiderations: (1) proximit5r in mass to the traceelements of interesq (2) proximity in concentration tothe unknown samples, (3) accurate abundances ofsufEcient concen[ation to produce high count-rates. Inour studies, low-abundauce isotopes of major elements(generally zesi, oACa or oaca, and 57Fe) are used asinternal standards, the major-element concentrationshaving been determined prior to laser ablation byelecfton-microprobe and X-ray fluorescence analyses.

For single standard calibrations, NIST-612 is usedfor a suite of elements, as this sandard has been dopedwith many geologically interesting trace elementsranglng in concentration from 15 to 78 ppm (Jacksonet al. L992, Fedorowich et al. 1993). This standard hasbeen used to analyze for trace elements in diversemafrices such 4^s glass, amahibole, and carbonate, gen-erally using 2esi and aCa as internal standards.MST-612 seems to produce more accurate resultswhen aCa is used as the internal standard for maficglasses and for minerals with calcium contents similarto that of NIST-6L2 @i9.4). In tlis case, one potentialproblem is that the standard and the unknowns are notmatched in terms of overall major-element contents(e.9., Si, A, Mg). This may cause problems in accu-racy under certain conditions. For felsic glassescalib'rated with NIST-612, it is generally necessary touse 2eSi as the internal standad, as silica contenti arematched, and other elements are eitler at lowconcentrations (Ii, Mg, Ca, Fe) or have high gas-blanks (Na" Al).

In minerals, intemal standardization can be done byobtaining an estimate of the proportion of an element inthe minslal and monitoring this element duringLAM-ICP-MS analysis. A correction for the RelativeSensitivity Function (RSD of an unknown element

relative to an element of known concenftation can bemade using the following formula:

R.rflt,*ll*p.x(Ct")*n.*(C.;r)sto.(Cr)*-1. =

[rqrf t-,t"1]"4* (CrJ.o

where (s) is an intemal standard of choice, I,.., is theintensity at mass x of the element being sought, andC(.,*) is the concentration of the element of mass x inthe standard.

The concenfration of the internal standard can bedetermined by (1) assuming a stoichiometric concen-tration in a mineral (e.9., Ca in calcite,Z or Si in zircon), or (2) by elecfron-microprobeanalysis of spots in the mineral to be analyzedby LAM-ICP-MS. The ascuracy of the analysisdepends on the accuracy of the determination of theinternal standard, and on accurately monitoring drift inthe RSF. The results presented in this paper have beenacquired using 2esi and a3Ca or oACa as the internalstandard.

The accuracy of the results presented here and inother studies (1e., Jackson et al. 1992, Fedorowichet al. 1993) suggests that maffix effects are relativelyunimportant in the laser-ablation process. Experimentsin progress with multiple glass standards of differingFe, Ca and Si composition indicate some dependenceon the matix, particularly for samples with high Fecontents (Stix el aI. L995). Both Y aad Zr displayincoherent behavior in glasses of different composition(see results for MST 672, Fig. 4), suggesting thatmatrix effects may play a signifi.cant role for someelements. This observation may partly explain thecommon diffrculty of obtaining accurate Y and 7rconcenfrations by laser-ablation analysis of maficglasses when using NIST-612 w a standard, sincethe iron content is significantly different between thestandard and the unknown samples.

Inte rfe renc e corrections

Oxide formation is a significant problem in deter-mining ̂ REE concenilations by solution-mode ICP-MS(e.9., Jarvis et al. 1992). Howevero in laser-ablationmode, the absence of a solvent entering the ICP torchresults in much less oxide formation. Jackson et al.(1992) observed that oxide formation is typicallyaround l.S%o, and, even for light-ftEE-enrichedminerals, the difference between the results by oxide-corrected and oxide-uncorrected calibrations are gener-ally less than l%o. Accordingly, no oxide correctionswere applied in the trace-element studies reported here.However, at levels of. l%o oxrde formation, inter-ferences may be a problem in isotope-ratiodeterminations. During the experiments on zirconcrystals, possible interferences due to REE dioxidewere evalualed by analyzing solutions of Yb, Lu and

APPLICATIONS OF I-AM_ICFMS ANALYSIS TO MINERAIJ

FIc. 4. Relative accuracy of MST 612 determinations by laser-microprobe analysis:a) using Si as an internal standard; b) using Ca as an internal standard.

425

:=E ''oooI E

EoL

I

t r0(l)

EQ - 5I

ooE -rott

Iff; no interferenceb were detected. Monoxide inter-ferences on U or Pb may come from platinum-groupelements, which are at low concentations in mineralsof interest in geocbronology.

Blank corrections are a more significant problem.High blanks will degrade detection limits, and it is inthe operator's interest to limit the gas blanks across fhemass spectrum. As is evident from Table 2, the gasbla* is highly variable. High blanks are caused by apoorly sealed instument, glow discharge in the mid-mnss range, and memory ftom eadier samples. Theblanks in the middle masses result from elow-

dischmge and complex ionization of gases, whereas tlehigh Hg flank (similar blanks may be present for lightmasses) results from contamination in the insrumentand can only be lowered by a complete slsaning of thesamFling area and the lens stack

Data acquisition and calibration

In our early work on zircon crystals, the data-collection philosophy closely followed that used forthermal-ionization mass specfromety. As individualblocks of data could be obtained over a 5 to l0-second

C a V C o R b Y

Eaoh repoat ls 4o.second acquisiton.Preablaton tme = 20 s€cond

Nb La Pr Sm Gd Dy Er Yb Hl Th

C a V C o R b Y Nb La Pr Sm Gd Dy Er Yb Hf Th

426 TIM CANADIAN MINERALOGIST

counting period with precisions of. L-3Vo (Feng et aI.1993), statistical compilation of 15-30 data blocksshould lead 1e imFroved precisions. However, as isevident from the signals displayed in Figure 3, signifi-cant changes in relative intensity occur that probablyresult from the sampling process. These variations canbe averaged by sampling over a longer period.However, increasing the sampling interval may lead toinclusion of spurious data caused by sample hetero-geneity. In Figure 5, the results for variation inwwP{KPb and 238UP6Pb are shown for fewer blockswith longer durations. The decrease in RSD and itsconvergence to the theoretical value indicate thatPoisson counting statistics dominate the sources oferror. The lack ofconvergence to the correct U/?b ratioconfirms the problematic behavior of U and Pb, dis-cussed elsewhere in the paper. As a general procedure,longer counting-times and higher-intensity signals willlead to higher precision. If a single spot can be ablatedfor several minutes, precisions in isotope ratios ofsingle elements of less than 0.5Vo and in interelementratios of less than LVo are attainable.

For a 30-s acquisition for 10 isotopes, themeasurement time for each isotope is about onesecond; the raw count-rate is therefore equal to theintegated rate of counting (cps). The 30-s acquisitionis repeated three times to give a total time of ablationof 90 seconds for each laser-sampling pit. The meanintensity measured from the three repeat analyses isextracted off-line and calculated using a spreadsheetprogran.

A typical analytical routine includes analyzing sixsampling pits of the reference standard (NIST 612;zircon crystal CN-92-1; ooin house'o mineral standard),twelve unknown samples and five gas-blanks inbetween the standard measurements. The mean inten-sity of the five gas-blank measurement$ is taken as thebackground level ofthe instrument and subtracted fromthe mean intensity of each measurement. The relativeinstrument response factors @SF) are established byusing the glass standard. The concentrations for NIST6 12 were takeu from Jackson et al. (L992) and our ownanalyses in solution mode. Ideally, 'oin house" mineralstandards are analyzed in solution mode by ICP-MSand in several laboratories by multiple techniques. TheRSF is used to correct for relative ablation-yield, trans-portation effrciency and maftix effects. As the RSFmay drift over time, it is monitored by repeat analysisof the standards. The effectiveness of the drift correc-tion was confirmed by analyzing the standard as anunknown. As shown in Figure 5, with few exceptions,the LAM-ICP-MS results for NIST 612 marched withsolution mode and reported values.

Pb GEocHRoNoLocy

Conventional U-Pb geochronology involvingmechanical separation, abrasion and dissolution ofU-

'....12 \t'.r....,..-u

, \o

''---"";"""........3..................... .......r.

0 1 0 2 a 3 0 4 0 5 0 6 0

Frc. 5. Changes in relative staadard deviation (RSD, 1 sigma)using different data-collection times and repeats. Takenfrom Ludden at al. (1995): a) exarnple for PblPb variation;b) example for UlPb variation; estimated error: RSD for(7a+7b+,,,n)/(6a+6b+...n); measured error RSDfor IQ / 6)

"+ Q | 6)6+....]/n.

enriched minerals and analysis by thermal-ionizationmass spectrometry (IIN4S: Krogh 1973, 1982) remainsthe technique of choice to obtain high-precision(<O.LVo) ages. Kober (1987) and Krtiner & Todt (1988)have developed 6 single-zircon method of analysis byTMS that involves mounting a zircon grain on a fila-ment to produce sequential evaporation of Pb to obtain2wWP6Pb radiogenic ages. This method is relativelyrapid but does not have the precision of conventionalTIMS analysis and does not provide U/Pb ratiosrequired to prove zircon concordance; slepwise evapo-ration moves the 2aPbP06Pb value toward that ofzircon on the concordia.

Grains of zircon and other minerals of interest inU/Pb geochronology may have a core and over-growths; analysis by conventional TIMS is possible onmineral fractions of abraded cores, etc., but in situaaalysis of grains is the only realistic means of

a6(L

Ibql so-5 4AI

; e goU)t r 2

-oo-

E _ gf,

@

N

s1oac E 1

25t'.

1

^ \ 1 2\ '.,

-#

*--.O"-.

M@rodercr

EstLnated ercr

Me@rcdercr

Estmat€d srcr

0

APPLICATIONS OF I-AM_ICP-MS ANALYSIS TO MINERAI.S 427

resolving these complicated growth-histories. TheSHRIMP (Sensitive High Resolution Ion Mcroprobe)technique has pioneered this area (Compston er a/.1986) by using a focused ion-beam to sputter craters ofabout 2G-30 Fm. This ablated material is transportedinto a high-resolution mass spectrometer that producesages of 1-2Vo precision and provides both Pb- andU-isotope data. Despite the potential of the SHRIMP,the high price and operating costs have hampered thewidespread use ofthis technique. As discussed above,with the LAM-ICP-MS, it is possible to measureisotope ratios of the same element with a precision of<LVo. T\e technique thus presents an exciting possi-bility for Pb/Pb geocbronology in zircon and otherminerals @eng et aI. L993, Fryer et al. 1993); at aprecision of l%o, ages of Precambrian zircon can bedetermined to t30-40 Ma. Obviously, fhe precision ofthe technique does not allow the high-resolutionstratigraphy required in unraveling the evolution ofPrecambrian terranes (e.g., Corfu et al. 1,989).However, the SHRIMP also does not provide the ageresolution to solve many stratigraphic problems.However, both laser-microprobe ICP-MS and theSHRIMP provide in situ ages that can be used todecipher the history of zhcon crystals, to analyzepopulations of detrital zircon, and to provide regionalgeological information in relativd unknown terranes.Both in sirz rcchniques should be used in concert withthe conventional method of U/Pb analysis.

Pb geochronology

To evaluate the Pb/Pb laser-ablation lechnique, weanalyzeA a series of zircon crystals by conventionalU/Pb techniques and by Pb/Pb LAM-ICP-MS. The207Pbl206Pb ratios were determined relative to astandard zircon. CN-92-1" from a skarn in theGrenville hovince. Canada. This zircon standard wasused to monitor mass fractionation, No optical zona-tion was observed in CN-92-1. However, examinationin different spots indicates variations in Pb, U, and Thabundances and m8PbP6Pb ratios, but homogeneous2wPbPooPbrxios. Four analyses of different fragmentsof CN-92-l bv the conventionat TIMS U-Pb methodgpve zwVa?wPb = 0.07784 t 6, and rhe average of2vpbp?6pb from ten individual sampling pits can bereproduced withn 0.5Va of the TIMS data lw.ll.h a 0.6Vo(1 sigma) RSD @eng et al. L993).

In cases of high-quality transparent zircon with arelatively simple history, ages are within 1% of theU/Pb (TMS) aCe (Fig. 6). From these results and forthe present configuration of the instrumenl we candefine the limits for accurate analysis as follows:approximately twice the sample-pit size (i.e., 80 1t.m, or200 mesh) and abundances exceeding 5 ppm 2ffiPb(Fre. 7).

Application of this technique to a population ofdeftital zircon crvstals from the Amazon hecambrian

shield in Brazil (Frg. 8) indicates a bimodal population,3.2 and 2.9 Ga. The eastern Amazon shield is thusunderlain by rocks much older than hitherto known,and distinct source-rocks were eroded to fill the basins.These data provide an invaluable background to single-grain conventional geochronology, as following laserablation, the grains can be removed ftom the mount,dissolved and analyzed.

The limitations of LAM-ICP-MS Pb geochro-nology can be reduced by using a higher-sensitivitymass spectrometer, smaller spot-size, and a UV laser.The destructive nature of laser ablation relative to theSHRIMP tecbnique is disadvantageous in terms of ourability to collect data for long times, although thehigher sensitivities (5-10 times those of the SHRIMPfor Pb) partly overcome this problem. However, profil-ing through a zircon grain does provide importantinformation that would otherwise be unobtainable oncore-rim history @eng et al. 1993).

CensoNA,TFs

We have examined three types of carbonatematerial calcite, dolomite, and benthic foraminiferaltests. In fhe studies of calcite and dolomite" the aim wasto establish a routine trace-element package forcarbonate analysis, and to evaluate the possibility ofusing NIST glasses as standads. In the foraminiferal-test study, we determined key element-ratios that aredependent on the temperature of formation of theforaminifers, with the aim of circumventing laboriouscleaning and separation of these foraminiferal tests forroutine solution-analysis.

Analysis of calcite and dolomite

Feng (1994) presented results for calcite anddolomite samples for which we had completed bothLAM-ICP-MS and solution ICP-MS determinationsusing standard addition (Jenner et al. 1990).LAM-ICP-MS analysis followed the general proce-dures outlined in this paper, with a3Ca as an internalstandard. There is excellent agreement betweensolution analyses and laser analyses for most samples,but significant disparities in some cases @ig. 9). Forsampling pits of <50 pm, the limits of detection areabout 15 ppb for the light REE, L0 ppb for the heavyKEE, and 200 ppb for Sr, Y and Ba. For homogeneousmineral samples, based on sensitivities of about500 cps/ppm, the counting-statistics uncertainty for anisotope of 1 ppm concentration is approximately 57oRSD; relative accuracy for the same elements in NISTglass 612 is generally 4Vo (Fig.4).

In some minerals. there are considerable differencesbetween results of the bulk analysis in solution andspot analysis by laser. In general, these differences areattributable to tace-element inhomogeneify withincrystals. Figure 9 shows results for igneous dolomite

428 TIIE CANADIAN MINERALOGIST

Difference between Laser ICP-MS ntp6lzoap6 and TIMS U/Pb age

*o

6

"1

%*oeooS4l"r,)'/',",ft , j*,f ",*o{u-{oef,y*}e;1.,"&*-Ftc. 6. Pb/Pb age results by laser-ablation microprobe ICP-MS: comparison of differences

relative to U/Pb TIMS determinations (sample descriptions nFeng et aL 1993).

from a carbonatite, which is very rich in light.REE and sampled was as much as 4QVo for the light REE andpoor in heavy REE. The relative variation for the tbree approximately lOVo for the heavy REE The mostdifferent samFling pits from a single grain was from reasonable explanation for this variation is primary2-lOVo, whereas the variation among all gtains differences in concenfations of the lnghr REE in these

Errors for the average of mean

:B: \g \

11\.

FAed [ne tor tho RSD of the average ol meany=0.03803-0.01215. log(x) F =0.2394

El RSD ("/.) foraverage of mean-F" RSD (7.) estimaled for 5 s.''A. RsD (7") *flmatedfor 10s.

E t':e- tr

o 25 50 75 100 1 150 17s n0 25 2@ 275 300

2o7en lppm;Flc. 7. Changes i1 analytical enor (1 sigma) relative to the 2mPb concentration of the

mineral.

1Wo

d rz.

o&, *o

APPLICATIONS OF I.AM_ICFMS ANALYS$ TO MINERAIJ

IDetricalZircon

Rio Fresco Group

ISsna d@ Caralas fuea (n=25)

ffiRlo MadaArea (n=24

ffi

I

I#ffiI ffi# ffi

am @0 3m0 u@ 3600 s800Age (Ma)

FIG. 8. Pb/Pb ages from dehital zircon from the Rio Fresco Group in the Amazon Shield.

429

-c=$e.aat,

f rec

minerals. Nonetheless, the excellent coherence of thesolution-IcP-MS results and the composite of all grainanalyses by LAM-ICP-MS is proof of the accuracy ofthe LAM-ICP-MS technique. TWo calcite sampleswith 4, ppm for all RZE (Fig. 9) show fractionation ofthe heavy REE and reasonably good correspondencebetween results of laser-ablation and solution analysesdespite the very low abundances.

The results for carbonates show that it is possible toeltain quantitative analyses of calcite and dolomileusing LAM-ICP-MS with non-matrix-matchedstandardization, even though calcite and dolomitehave poor absorption of the IR-laser energy used inthis study. As discussed by Feug (L994), the successof the non-matix-matched stendardization is relatedto a stable and reproducible laser-coupling process.In particular, the laser beam used in this study can befocused to about 10 irm, forming craters 30-50 pm.across; the power density of tle focused beam(i.e., L}LL Wcm2) is higher than an unfocused beamle.g., l}e Wlcd, forming a crater of >200 pm:Denoyer et aI. (L991)). As suggested by Abell (1992),under this high-power density, the dominant absorptionof laser light and the resulting processes of therrnalvaporization at lower power (Ready 1971) give way todirect ablation due to laser-induced plasma-plumeexpansion. The induced plasma ir d[s major source ofenergy for the ablation, not the absorption of laserlight. Thus, this secondary coupling process allowsablation even when the material is transparent to thelaser (e.9., calcite).

Analysis of benthic foraminiferal tests

The trace-element composition of fglaminiferal testshas received much attention from marine geochemistsas a means of i::ferring the physical and chemicalconditions of the seawater in which the foraminiferallests grew @ender et aL L975, Boyle 1981, Lea &Boyle 1991). Current lsshniques require severalcleaning and leaching steps before analysis in solutionby ICP-MS, graphite-fumace atomic absorption orother tecbniques. Commonly, several foraminiferaltests are required to provide sufftcient material foranalysis. The LAM-ICP-MS technique offersthe advantage of being able to analyze singleforaminifera tests for several elements simultaneously.In a pilot study, Wu & Hillaire-Marcel (1995) analyzedlarge (approximately 250 ttm.) benthic foraminifera forfour minor cations (Sr, Mg, Mn and Ba) that arerelatively concentrated in benthic samples (generally>10 ppm). The instrument was optimized atmaximum sensitivity for 88Sr, 43Ca was used asan internal standard, calibration was done usingNIST 612 glass, and the ratios Sr/Ca Mn/Ca, MglCaand Ba/Ca were determined (Fig. 10); ablation cratersare shown in Figure 2. In general, the ratios can bereproduced to <L0Vo. T\e in sin analysis shows thatSr is present primarily in the carbonate of tlebenthic foraminifera and not in the Fe-Mn coatings,Mg and Mn are concentated in the coatings, andBa is associated either with the coatings or with baritecrvstals.

430 TI{E CANADIAN MINERALOGIST

Frc. 9. Comparison of chondrite-normalizecl concen$ations ofrare-earth elements in caxbonates by LAM-ICP-MS andby solution-mode ICP-MS: a) igneous dolomits; b) veincalcite from an Archean metavolcanic rock and from aGrenvillian marble.

MaFrc MnrFRAr,s AND STANDARDS

As matrix matching is not as strict a requirement asin ion-probe analysis, the fact tlat the calibration forLAM-ICP-MS can be done using synthetic glasses isextremely encouraging. Whereas more tests arerequirc{ pafiiculady in comparing coupling betweentransition-metal-rich samples and transition-metal-poor samples (1.e., Stix et aI. L995), the use of syntheticglasses provides a signifis26 advantage, as mineralsthat are homogeneous within the detection limits of theLAM-ICP-MS are rare. At current limits of detection.repeat LAM-ICP-MS analyses of NIST glassesindicates homogeneity at least at the l-2Vo level ofprecision.

Minerals equilibrated at high temperatures and pres-sures commonly are homogeneous for major elements.However, these minerals are commonly poor in manytrace elements of interes! in terms of advantages foruse as standards, their homogeneity is therefore offsetby the low abundances of some elements. To test thefeasibility of analyzing mantle minerals and to createworking standards, we selected a kaersutite and a

diopside megacryst from Kakanui, New Zealand.Repeat electron-microprobe analyses for the majorelements define an RSD of 4Vo, rncludrngK (I.SVo)and Na (1.87o) for the amphibole and Na (l.S%o) forfteclinopyroxene; Ti has an RSD less than l%o, andthe RSD of Si is approximately 0.5Vo 1e1 lqth minerals.Given the detection limits of the electron microprobe,these minerals are therefore homogeneous. The resultsfrom a LAM-ICP-MS sampling taverse across ttreamphibole and the clinopyroxene arc gtven in Table 3;included are the limits of detection based on threetimes the gas blank (see earlier discussion). These dataillustrate the homogeneity of llsss minelalg, 6gconcentrations of elements such as Ba La and Sr arereproduced at better than2Vo in many cases.

C. WuellerstorflHU-91-M5-18- a a ,t t t t t ; t . ! . .

. . '

200

100 150 200 250

0 50 100 150 200 250Mn/Ca (pm/mole)

Frc. 10. Ratios of Sr/Ca. Mg/Ca and Ba/Ca in benthic tests ofC. wwlkrstotfi, showing the effects of Fe-Mn-coatingand Mn-carbonate overgrowth; data from Wu & Hillaire-Marcel (1995).

@

o

E2EE; 1R(t)

9rsE

Ero;e5o=

0

9goE

i 20;B10!o

0

I

250

@

1 " " 1 " " r " " r0 50 100 150

20

40@

Calcite:L2 calclte veln trom Abldbi grsnsione BaltL1 €lcitE from matbls

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

C. WuellerstorfrHU-91-U5-18

Mn-Calbonate

r r r 1 ' P l

Fe-Mn 7'coatln2/ .a a / .

C. WuellerstorfiHU-91-A5-18 .

BaSOamicrorysials

t t 'r l a

.' . '. '.:'.-0 l'. '

APPLICATIONS OF I..A,M_ICP_MS ANALYSS TO MINERALS 431

The data of Table 3 were calculated using aCa as aninlernal standard (Frg. 11). For the REE, significantdifferences exist in the absolute values depending onwhich major element is used for normalization; in thisexampleo both Ca and Si were used. These differencesseem constant, at least across the RZE specftum, withthe Ca-normalized results being higher and invariablycloser to the values determined for dissolved samples(Ftg. 1 l). For several elements, the laser reproduces thesolution data within error. The LAM-ICP-MS datapresented here show lower (by about 5Vo) absolutevalues relative to results of solution analyses from theUniversit6 de Monft6al and the Memorial Universitylaboratories. In most cases. the ratios of the elementsare constant between LAM-ICP-MS and solutionICP-MS analyses. The integrify of the trace-element profiles produced by LAM-ICP-MS, in par-ticular the relative fractionation of high-field-strengthelements and the heavy REE, can be used to provideimportant constraints on mngma genesis.

Trn Frrn:nr

The encouraging results summarized here for Pb/Pbanalysis of zircon crystals, carbonates and maficminerals, in addition to data presented elsewhere, inparticular by Jackson et al. (1992), Jenner et aI. (7993),and Fryer et al. (1,993), have established theLAM-ICP-MS technique as a powerfirl new analyticaltool. The technological advances in LAM-ICP-MSanalysis are evolving rapidly. ICP-MS insffuments arenow capable of producing in excess of 200 millioncps/ppm for solutions; the insffument used for the workdiscussed here has a sensitivity of about 49 millioncps/ppm. The results given here were collected using afinely tuned Nd:YAG laser at low power (4-5 Hz),which operates in the IR region at L0& nm and canproduce ablation pits of 20-40 trlm depending on thematerial analyzed. Quadrupling (fourth harmonic)the laser beam produces a UV laser of 2ff nm, capableof ablating pits one quarter the size of the IR Nd:YAG,and results in better coupling between the laser andtransparent minerals (Jenner et al, 7993, Fryer et al.1995). However, the amount of ablated materialejected from the pit is directly related to the volume ofthe crater. To maintain count rates equivalent to thosereported here for NIST glasses, tle sensitivity of theICP-MS instrument must be increased by an order ofmagnitude, through a combination of enhancedplasma-ionization efficiency (i.e., using mixed carriergases), modified geomefiry of the-.,ICP-MS - plasmainterface, and more efficient hansport of ions in themass analyzer. High-sensitivity analysis of 30-50 pmcraters should result in an order-of-magnitude increasein count ra!e, and the ability to attain detection limits oftens of ppb. However, the requirement to monitor aminor isotope of a major element in a mineral andattain these detection limits requires a linearity of the

TAILE 3. BS'I'I]TS OF REPTICATE LAM-IgP-ldfI ANAI,YSNS

Dlolrglde' K@Eutlto

N ' 1 2 B S D S N " 9 B S D S

stanf

SovBbBaThTauNbIaCsP!S!NdSmZtEfEllGdTbDyYHoE!TmYbLu

38 ppm 2804 70.10 >200.09 )200.06 >200.o7 >240.01 >200.34 181.81 78.02 16t ,s1 588.30 28 .16 I2.49 t427.80 61.46 40.80 s2 .87 110.34 72 .10 78 .50 60.88 80.83 I0.09 710.62 150.08 18

19.5 ppn 13 6 ppn3 8 9 I I14.8 10 0.4

2 1 0 . 30.08 >20 0.041.66 I 0.050.01 >20 0.0228.1 2 0.14 .7 4 0 .0610.8 4 0 .062.74 4 0.05&4 | 0,216.78 3 0.24 .18 6 0 .2u .4 4 0 .22 ,4 8 0 .11 ,4 6 0 .08s ,4 10 0 .30.48 72 0.032,35 5 0.047,8t 5 0.10.3i3 t2 0.040.61 10 0.10.05 >20 0.090.it3 >20 0.!0.01 >20 0.0

I Data ssF @loulatod uslng ca 6s e ,ntsmal standald.2 Bkok valus m basd oB s avoage fo! tlu two data-6ate.

detection system of at least 10e orders of magnitude.Most of the applications described here require

refinement, and are in the development stage. Inparticular, we emphasize the importance of studiesrelating mafrix aJfects, coupling processes and varia-tion in elemental yield. The accuracy of the resultsobtained for carbonate analysis is good, and the resultspresented reflect a bulk analysis of a pit of 30-40 lrmin diameter and depth. Most low-temperature mineralsare finely zoned, and variation in the results fromrepeated ablations reflects this. Higher spatialresolution and ablation taverses using time-resolvedcollection are required in these studies and will become6rr imFortant part of LAM-ICP-MS applications.Nonetheless, sampling at a scale that may averagesome of these heterogeneities, (i.e., in studies ofpartitioning of an element in mantle minerals) providesfundamental constraints on manfle melting and theevolution of magmatic systems.

As the resolution and sensitivity of theLAM-ICP-MS increase, standardization will becomeincreasingly important. The NIST standards now usedmay well be heterogeneous at lower spatial resolution.The technique is destructive and hence requires acontinuous supply of mineral or glass standards; thefuture supply of high-quality standards is an importantconsfraint on LAM-ICP-MS analysis.

Clinopyroxene and amphlbole laserablation ICP-MS analyses

TTIE CANADIAN MINERALOGIST

La Ce Pr Nd Sm Eu Gd Tb Dy Er Tm Yb Lu

Amphibole

432

@100

Eo.12(!FE

810o<tco-co

Ho

LuTm

FIc .11 . Resu l ts o fLAM- ICP_MSanalysis of amphi-bole and clinopy-roxene megacry$sfrom Kakanui,New Zealand: a)comparison of Ca-normalzeA verswS i - n o r m a l i z e ddata; note tlecrossover in heavyREE, which ismaintained in bothnormalizations; b)and c) illustrateresults for clino-pyroxene and arn-phibole megacrysn,respectively dis-played relative tosolution data oncleaned mineralfractions. All dataare chondrite-nor-malized usingvalues given inSm&McDonough(le8e).

@100

Eo.a(tlI

8too

Ego-c(,

1

/:\( c )

100:oo.9,ct=o

I toEEo(t

1

-.-- CPX Sl-norm.

-_-__-_.-_____- CPX Ca-norm,

-----tr- AMPH Sl-norm.

---+__ AMPH Ca-nonn.

+ LaserCa-norm.

----r- Soln. data (MUN)

----{- Soln. data (UdeM)

Clinopyroxene

----^- Laser Ga-norm.

-------{-- Soln. dab (MUN)

-----tr- Soln. data (UdeM)

APPLICATIONS OF LAM_ICP-MS ANALYSIS TO MINERAI.S 433

AcKr.IowlEDcEr,fihlrs

The Universit6 de Mont6al ICP-MS laboratorv isfunded by NSERC-Canada as a regional faciliry. Weacknowledge the contributions of Bdgitte DionneoAmira Kouhri and Jean-Pierre Bourque in variousaspects of sample preparation. Glenn Poirier providedhelp with the McGill electron microprobe. Kurt Kyser,Brian Fryer, Frank Hawthome and Robert Martin areacknowledged for discussion and helpfirl reviews ofthis paper.

RSFEREI,{CES

Anelq I. (1992): rmFrovements in spatial analysis byI-A-ICP-MS. ICP Infornation Newsletter 17, 116-1,19.

Annowsurrn, P. (1987): Laser ablation of solids for elementalanalysis by inductively coupled plasma mass spectro-mety. Annl. Chent 59,1437-1444.

Bworn, M.L., LoRs.rs, R.B. & Wn.l.nus, D.F. (1975):Sodium, magnesium and strontium in tests of planlronicforaminifera- Micropaleontobgy 21, 448459.

Boru, E.A. (1981): Cadmium, zinc, copper and barium inforaminiferal tests. Earth Platut. Sci. Izx. 53. 11-15.

Couruou, W., KNNY, P.D., Wn-uarras, I.S. & Fosrau J.J.(1986): The age and Pb loss behaviour ofzircons from theIsua supracrustal belt as aletermined by ion microprobe.Earth Planet. Sci, Lett. E0. 71-81.

Conru, F., Ifuocn, T.E., Kwo& Y.Y. & JB{SB,{, L.S. (1989):U-Pb zircon geocbronology in the southwestern Abitibigreenstone belg Superior hovince, Can J. Eanh Sci. ?.6,L747-t763.

Dm.loym, E.R., I(mwsur J.F. & HeosR J.W. (1991): Iasersolid sampling for inductively coupled plasma msssspectometry. Anal. Chem 63, M5 A457 A.

Fmonowrcl J.S., fucrunos, J.P, JAN, J.C., Kmrucr, R. &Fer, J. (1993): A rapid method for REE and trace-elementanalysis using laser sampling ICP-MS on direct fusionwhole-rock glasses. Chem. GeoI. 106, 229-249.

Fhvo, Rut (1994): In rit, trace-element determination ofcarbonates by LaserProbe inductively coupled massspectrometry using non matrix makhed standardization.Geochim- Cosmochim Acta 58. 161.5-1623.

MACHADo, N. & LnDDBN, J.N.(1993): Leadgeochronology of zircon by laserProbe - inductivelycoupled plasma mass spectrometry (-P-ICPMS).Geochim. Cosmochim. Acn 57, 3n9-3486,

Fnrm, B.J., Jacrsolr, S.E. & LoNcmtcs, H.P. (1993): Theapplication of laser ablation microprobe - inductivelycoupled plasma - mass strcctrometry (.AM-ICP-MS) toin-situ Q,r-l>b geochronology. Chem- GeoI.lll9, 1-8.

& - (1995): The design, opera-tion and role of the laser-ablation microprobe coupledwith an inductively coupled plasma - mass spectrometer(LAM-ICP-MS) in the earth sciences. Can, Mineral.33,303-312.

Gnav, A.L. (1985): Solid sample introduction by laser abla-tion for inductively coupled plasma mass spectomety.Analyst ll0,551-556.

Hecnn, J.W. (1989): Relative elemental responses for laserablation - inductively coupled plasma mass spechometry.Anal. Cheru 61, 1243-1,248.

Htu'ns, G.M. (1,992): Plasma diagnostic techniques forunderstanding and control. Spectrochbn. Acta 4lTB,3-25.

Hnwm.s, T.A., I{unnuan, E.M., Sprru-;m, T.M. & Iftnsnaw,J.M, (1987): Inductively coupled plasma mass spectro-metric determination of lead isotopes. hnl. Chzm 59,2658-2662.

JAcKsoN, S.E., LoNGRrcH, H.P., DuNNnqc, G.R. & Fnvsn,B.J. (1,992): The application of laser-ablation microprobe- inductively coupled plasma - mass spectrometry(LAM-ICP-MS) to in situ trace-element determimtionsin minerals. Can. Mineral.30. lM9-10@.

Ianvr, K.E., Gnev, A.L. & Hourg R.S. (1992): HandbookofInductively Coapled Pla^rma Mass Spectrom.etry. Blaskie& Son Ltd., Glasgow, U.K.

JSNNER, G.A., For,sv, S.F., Jacrson, S.E., GRm{, T.H.,FRyB, B.J. & LoNGRrcH, H.P. (1993): Determination ofpartition coefficients for trace-elements in highpressure-temperatue experimental run products by laserablation microprobe - inductively coupled plasma - massspectrometry (LAM-ICP-MS). Geochim" Cosmochim-Acta 57,5099-5103.

LoNGmrcH, H.P., JacrsoN, S.E. & FnvsR, B.J.(1990): ICP-MS - a powerfirl tool for high-precisiontrace-element analysis in eaxth sciences: evidence fromanalysis of selected U.S.G.S. reference samples. Chem.Geol. E3,133-148.

Kosm, B. (1987): Single-zircon evaporation combined withPb+ emitter bedding for wPbPwPb-age investigationsusing thermal ion mass spectrometry and implications forzirconology. Contrib. Mineral. Petrol. 96, 63-7 I.

KRocH, T.E. (1973): A low-contamination method forhydro0rcrmal decomposition of zircon and extraction ofU and Pb for isotopic age determinations. Geochirn.C o smochim Acta tf , 485 494.

(1982): Improved accuracy ofU-Pb zircon ages bythe creation of more concordant systems using an airabrasion lsshnique. Geochirn. Cosmachim, Acta 460637-649.

KR6NER, A. & Toor, W. (1988): Single zircon dating con-straining the maximum age of the Barberton GreenstoneBelf southem Africu J. Geoplrys. Res. 93, 1,5329-1,5337.

434 TIIE CANADIAN MINERALOGIST

IjEA, D.W. & Bovre, E.A. (1991): Barium content of benthicforarninifera controlled by botiom water composition.Naure 338,751-753.

LoNcmrcr, H.P., JAcKsoN, S.E., FnvsR, B.J. & Srnoxc, D.F.(1993): Machinations: the laser ablation microprobe -inductively coupled plasma - mass spectrometer(LAM-ICP-MS). Geosci. Can. 20, 2l-2i1.

Luonw, J.N., Fwc, Rur, MACHADo" N. & KERRTcTL R.(1995): Temporal behaviour of Pb, Th and U signals andsources of error in LAM-ICP-MS. Soectrochim. Acta(in press).

MoBttt<s-BI-ANrcNBunc, L., GAoc,q M., GilNrHrR, D. &KAtln/H,, J. (1990): hocesses oflaser ablation and vapourtransport to the ICP. In Plasma Source MassSpectrometry. Proc. Third Suney Conf. on Plasma SourceMass Spectrometry (I<,8, Jarvis, A.L. Gray, I. Jarvis & J.Williams, eds.). The Royal Society of Chemistry,Cambridge, England (l- 17).

PEARCE, N.J.G., Psru<nvs, W.T., AsHr, I., Duu.im, G.A.T. &Fucs, R. (1992): Mineral microanalysis by laser ablationinductively coupled plasma mass spectrometry. J. knl.Atom Spectrom 7, 53-57 -

PERKNS, W.T., PEARCE, N.J.G. & Fucr, R. (1992): Analysisof zircon by laser ablation and solution inductivelycoupled plasma mass spectometry. J. Anal. Atom.Spectrom. 7r 6Ll-6L6.

& Jwrrurs, T.E. (1993): Laser ablationinductively coupled plasma mass spectrome!ry: a newtechnique for the determination of trace and ultra-taceelements in silicates. Geochim. Cosmochim" Acta {lo475482.

READv, J.F. (1911): Effects ofHigh-Power Inser Radiation.Academic Press, New York, N.Y.

RH\IoND, G., Bers., A., RocQl'rs-Canues, C., WHsl, D.,ABHI, I. & SRoussI, G. (1990): Scanning mechanicalmicroscopy oflaser ablated volumes related to inductivelycoupled plasma - mass spectrometry. ScanningMicroscopy 4, 239-'247.

Snx, J., Gaurm" G. & Luooun, J.N. (1995): A critical lookat quantitative laser-ablation ICP-MS analysis of naturaland synthetic glasses. Can. Minzral.33,435 444.

Srn, S.-S. & McDoNoucH, W.F. (1989): Chemical andisotopic systematics of oceanic basalts: implications formande compositions and processes. /n Magmatism in theOcean Basins (A.D. Saunders & M.J. Norry, eds.). Geol.Soc. Inndou Spec. Publ.4,313-345.

Wu, cuom.rc & IItrIARE-MARcE-, C. (1995): Application ofLP-ICP-MS to benthic foraminifers. Geochirn.Cosmochim- Acra (in press).

Received. April 13, 1994, revised manuscript acceptedFebruary 22, 1995.