Characterization of zircon reference materials via high

1 downloads 0 Views 3MB Size Report
dSENS – Advanced Mass Spectrometry, Abelardo Vergueiro Cesar, 555 – Vila. Alexandria .... fractionation. R(sample)true ¼ R(sample)measured  [R(ref. mat.) ...

JAAS

Published on 09 August 2017. Downloaded by Universidade Federal de Ouro Preto on 04/09/2017 17:47:50.

PAPER

View Article Online View Journal

Cite this: DOI: 10.1039/c7ja00167c

Characterization of zircon reference materials via high precision U–Pb LA-MC-ICP-MS† Cristiano Lana, *a Federico Farina,b Axel Gerdes,c Ana Alkmim,a Guilherme O. Gonçalves a and Antonio C. Jardimd Recent improvements in laser ablation and multicollector (MC)-ICP-MS technology offer higher sensitivity, robust stability, and larger dynamic ranges. Such advances are translated into improved accuracy and precision on U–Pb age determinations, reduced spot sizes for high spatial resolution and reliable corrections for any amount of common Pb (Pbc). Although a number of studies have focused on the high-spatial resolution of the LA-MC-ICP-MS method, comparatively fewer experiments have explored the precision and stability of multi-collector instruments for high-precision U–Pb geochronology. In this contribution, we describe the precision and accuracy of Pbc corrected U–Pb ages obtained by LA-MCICP-MS, and discuss the homogeneity of two batches of zircon megacrystals (the Rio do Peixe (RP) and Blue Berry (BB)) as potential reference materials. The enhanced sensitivity and stability of the LA-MCICP-MS system allowed the determination of apparent Pb/U and Pb/Pb ratios with uncertainties of 0.3 to 1% (2s) for nearly all reference materials available. The measurements are done in automated mode, with a spatial resolution of 20 mm-wide spots. The short- and long-term reproducibility of 207

235

Pb/

U and

207

206

Pb/238U,

206

Pb/

Pb ratios were assessed based on several reference materials, including GJ-1

(96 analyses), M125 (96 analyses), Temora (34 analyses), 91500 (29 analyses), FC-1 (17 analyses), BB (121 analyses) and Pleˇsovice (133 analyses). The results show that the accuracy and precision of the relevant ratios are often between 0.5 and 1.0% (RSD). U–Pb LA-MC-ICP-MS and TIMS dates of four new megacrystals of the BB zircons agreed within error at ca. 560  1 Ma, and demonstrate that they

Received 27th April 2017 Accepted 8th August 2017

constitute suitable reference materials for LA-ICP-MS analyses. U–Pb LA-MC-ICP-MS for the RP zircons DOI: 10.1039/c7ja00167c

revealed low concentrations of U and Pb and a scattering of ages from 580 to 600 Ma (mean of 593  5

rsc.li/jaas

Ma). The RP zircons may only be suitable as a quality control reference material.

1. Introduction U–Pb geochronology via laser ablation inductively coupled plasma-mass spectrometry (LA-ICP-MS) is now a wellestablished and highly accessible method that plays a fundamental role in the eld of Earth Sciences. It is also known as a rapid and comparably inexpensive procedure that has achieved great improvement since the rst LA-ICP-MS U–Pb ages published in the literature.1–6 Initial problems with mass bias, interference and fractionation have been largely minimized

a

Applied Isotope Research Group, Departamento de Geologia, Escola de Minas, Universidade Federal de Ouro Preto, 35400-000 Ouro Preto, MG, Brazil. E-mail: [email protected]

Department of Earth Sciences, University of Geneva, Rue des Maraˆıchers 13, 1205, Geneva, Switzerland

b

c Institut f¨ ur Geowissenschaen, Goethe-Universit¨ at Frankfurt, Altenh¨oferallee 1, D-60438, Frankfurt am Main, Germany d SENS – Advanced Mass Spectrometry, Abelardo Vergueiro Cesar, 555 – Vila Alexandria, S˜ ao Paulo, SP, 04635-080, Brazil

† Electronic supplementary 10.1039/c7ja00167c

information

(ESI)

available.

This journal is © The Royal Society of Chemistry 2017

See

DOI:

aer signicant developments in laser-ablation (wavelength/ sample cell) technology and in data-reduction protocols.7–22 For instance, available soware such as Iolite,20 Glitter,23,24 Lamtrace,25 UranOS,26 and UPb.age17 have largely improved our ability to efficiently process large quantities of data. On the other hand, the U–Pb LA-ICP-MS method has oen been regarded as an inferior to high-sensitivity in situ method (via Secondary Ion Microprobe System – SIMS) due to large uncertainties in ratios and dates (1.5–6%, at the 2 sigma level – 2s), low spatial resolution and inability to precisely and accurately measure 204Pb for common-Pb (Pbc) correction. Even for the highest precision U–Pb geochronology studies, quoted uncertainties on the nal calculated date should not be higher than twice the precision of the individual (single-spot) analyses. Consequently, the uncertainty of the nal calculated date (or age) is a function of the uncertainty of the individual analyses, no matter which statistical method of date calculation is used.27 Among the high precision (magnetic sector eld) instruments, the uncertainty of the measured U–Pb ratios is oen above 1.5% (2s) for sector-eld SF-ICP-MS data3,11,12,15 and around 1% (2s) for multi-collectors.10,22,28,29 The main contributions to

J. Anal. At. Spectrom.

View Article Online

Published on 09 August 2017. Downloaded by Universidade Federal de Ouro Preto on 04/09/2017 17:47:50.

JAAS

Paper

such high uncertainties come from matrix effects between reference materials and samples28,30 and from low sensitivity and instrumental dri, which translate into low counting statistics or poor external reproducibility. The results can however be improved by using the right combination of carrier gases Ar–He– N–H, an adequate pair of skimmer/sample cones, sensitive and stable ion detectors and, more importantly, matrix-matched reference materials. For that matter, the multicollector (MC)ICP-MS is the ideal instrument because it is designed for applications in which precision and stability are more important than elemental coverage. The main advantage of the MC-ICP-MS over other systems is that all masses are counted simultaneously, which improves counting statistics, within run stability and therefore better internal and external uncertainties.10 By measuring masses simultaneously, small dri effects by the plasma source or laser are minimized because all counters see the same dri and thus the mass ratios are less affected. Simultaneous detection also reduces multiplicative/icker noise produced in the plasma or by laser ablation. LA-ICP-MS geochronology relies on well-characterized reference materials for matrix-matched external calibration and for quality control purposes. It therefore demands continuous development of reference materials with a wider range of ages and compositions (e.g., low Th–U vs. high Th–U zircons). Moreover, recent studies observed systematic age offsets between reference materials during laser ablation tests,28,30 suggesting that further characterization of the available reference material is required. Surprisingly, only a limited number of studies have explored the robustness and stability of multi-collector instruments for

Fig. 1

high-precision characterization of reference materials,29,31 despite a number of recent developments focusing on the high spatial resolution (e.g., single-shot) of the method.21,29,32,33 In this paper, we report on the precision and accuracy of Pbc corrected U–Pb dates obtained by LA-MC-ICP-MS and discuss the homogeneity of the Rio do Peixe (RP) and the Blue Berry (BB) reference materials. Both reference materials are abundant in quantity for distribution to LA-ICP-MS laboratories, but individual stones (or mm-wide megacrysts) may require characterization via TIMS and homogeneity tests via in situ geochronology.31 The BB zircon31 comes from a placer deposit of the Ratnapura gemstone eld,34 located in the southwestern region of the Sri Lanka Highland Complex.35 It occurs as large 1–3 cm-wide, purple megacrysts that are relatively homogeneous and similar to other Sri Lankan zircons36,37 in terms of Hf and O isotopes and trace element chemistry.31 The RP zircon comes from the Rio do Peixe alkaline complex in central Brazil.38–40 It occurs as large 1–3 cm wide megacrysts that may be suitable as a reference material for LA-ICP-MS. We have selected eleven transparent, gem-like megacrysts of the BB and RP for homogeneity tests via high precision LA-MC-ICP-MS dating and Chemical Abrasion (CA)-ID-TIMS.

2.

Materials and methods

2.1. Reference materials The RP and BB zircons are generally subhedral to anhedral, ranging from a few millimetres to several centimeters in diameter. The RP zircon comes from an 400 gram batch,

Transmitted and cathodoluminescence images of the RP and BB zircons that were selected for homogeneity tests via LA-MC-ICP-MS.

J. Anal. At. Spectrom.

This journal is © The Royal Society of Chemistry 2017

View Article Online

Published on 09 August 2017. Downloaded by Universidade Federal de Ouro Preto on 04/09/2017 17:47:50.

Paper

containing more than 70 megacrystals. The ve RP crystals selected for this study are purple, mostly anhedral, and vary in size from 1 to 2 cm in diameter. The BB zircon comes from an 300 gram batch, initially studied by Santos et al., (2017).31 Detailed TIMS analyses showed however that the BB zircons have some degree of heterogeneity in the U–Pb system, which led Santos et al. (2017)31 to conclude that each BB crystal should be characterized via high-precision geochronology prior to distribution. We selected ve new BB megacrysts that are homogeneous and sufficiently large for distribution to LA-ICPMS laboratories worldwide. The selected eleven samples (BB38–BB39–BB40–BB41–BB42) and (RP1–RP3–RP5–RP6–RP7–RP9) have large homogeneous domains that are separated by cm-long fractures. Some grains may show small fractured domains marked by fractures and a number of pits and minor mineral inclusions. Fractured domains are common around the rims and, for the case of the BB zircon, the fractures are oen lled with recrystallized zircons. Four megacrysts (RP03, RP5, RP06 and RP07) were sectioned in half and polished to reveal internal features (e.g., Fig. 1). For most parts, such megacrysts are homogeneous with

JAAS

weak CL zoning, or show small domains or rims that are rather bright under CL imaging. 2.2. Instrumental setup The ThermoScientic Neptune Plus is equipped with miniaturized secondary electron multipliers (SEMs) and compactdiscrete-dynode-ion-counters (CCD-ICs), fully integrated into the Faraday cup array (Fig. 2). Five of the ion detectors (CCD and SEM, making the “Plus” module of the Neptune) are xed on the outermost Faraday cup of the low side (cup L4) while one CCD is xed on the central cup. IC4, IC5, and IC6 are CCDs whereas IC2, IC1C and IC1B are SEMs. IC1C and IC1B sit behind retarding potential quadrupoles (RPQs), which in turn act as secondary electrostatic lters to improve abundance sensitivity. The outer ve ICs are xed at dened distances referring to one mass unit between L4 and IC6, one mass unit between IC6 and IC1B and IC2, and two mass units between IC2 and IC5 and IC4. The signals collected on the ve ICs are 208Pb on IC6, 207Pb on IC1B, 206Pb on IC2, 204Hg + Pb on IC5, and 202Hg on IC4. 232Th and 238U are measured on Faraday cups H2 and H4 respectively. The cup intensity measured in volts is converted into cps by multiplication by a factor of 62.500. For U–Pb measurements, the magnet is settled on a virtual mass (223.2 – Fig. 1) in the center of the collector array whereas all relevant masses (202Hg, 204 (204Pb + 204Hg), 206Pb, 207 Pb, 208Pb, 232Th and 238U) are measured simultaneously. Daily baseline and gain calibration are performed before tuning using script available from Thermosher. Ion counter crosscalibrations are done on a monthly basis, and factors are derived by direct comparison of the detector response by peak jumping of the same 206Pb signal across all ion counters. The operation voltage for all counters ranges from 2000 to 2750 V (IC1) and dark noise ranges from 0.002–0.008 cps. The yields of the ion counters vary from 92% (IC6) to 95% (IC2). The stability of the counters is oen measured in solution mode (with a 50 ppm Th solution) aer the cross-calibration procedure. In this case, the yields of each IC were measured 10 times over a period of 5 hours. The relative standard deviations (RSDs) of the yield values varied over the period of 6 months from 1 to 2%. We assume that the ion counters were stable within 1–2% over the course of an analytical session. 2.3. Data acquisition

Fig. 2 Multicollector array of the ThermoScientific Neptune Plus used in this study. The system consists of five ICs that are fixed on the outermost Faraday cup of the low side (cup L4), and one IC behind the central cup. IC4, IC5, and IC6 are CCDs. IC2, IC1C and IC1B are conventional dynode secondary electron multipliers (SEMs).

This journal is © The Royal Society of Chemistry 2017

The raw data for each spot are in the form of 550 cycles (or sweeps of 0.1 s each) of all relevant masses and ratios (207Pb/206Pb, 208Pb/232Th, and 206Pb/238U). The 207Pb/235U ratio is calculated off-line, where 235U is calculated from 238U (235U ¼ 238 U/137.818).41 The analyses are obtained in automated mode via a trigger cable between the laser and the MC-ICP-MS. Aer the laser triggers the ICP-MS, a 15 second background is measured, followed by a signal + background as the laser beam is activated for 40 seconds. A typical analysis, therefore, takes about 55 seconds. We apply a standard-sample-standard bracketing technique, analyzing 5 points on primary and 5 points on secondary reference materials, followed by 10 points on unknowns and back to 5 points on each reference material.

J. Anal. At. Spectrom.

View Article Online

JAAS Table 1

Paper Operation conditions and instrument settings

Laboratory and sample preparation

Published on 09 August 2017. Downloaded by Universidade Federal de Ouro Preto on 04/09/2017 17:47:50.

Laboratory name

Sample type/mineral Sample preparation

Imaging

LA-MC-ICP-MS (UFOP) U–Pb Laborat´ orio de Geoqu´ımica Isot´ opica, Departamento de Geologia, Universidade Federal de Ouro Preto (UFOP) Zircon Conventional mineral separation, 2.5 cm resin mount, 1 mm polish to nish CL, SEM JEOL JSM-6510LV, 20 nA, 20 kV, 20 mm working distance

Laser ablation system Make, model and type Ablation cell and volume Laser wavelength (nm) Pulse width (ns) Fluence (J cm2) Repetition rate (Hz) Ablation duration (s) Ablation pit depth/ablation rate Spot diameter (mm) Sampling mode/pattern Carrier gas

Cell carrier gas ow (l min1)

Photon Machines G2 excimer laser Low volume HelEx 193 nm 4 ns 1–2 J cm2 6 Hz 40 s 5–10 mm pit depth 20 mm Static spot ablation 100% He in the cell, Ar and N2 make-up gas, combined using two Y-piece 50% along the sample transport line to the torch 0.1 l min1

ICP-MS instrument Make, model and type Sample introduction RF power (W) Make-up gas ow (l min1) Detection system Masses measured

Integration time per peak/dwell times (ms); quadrupole settling time between mass jumps Total integration time per output data point (s) IC dead time (ns)

ThermoFisher Scientic, Neptune Plus, MC-ICP-MS Ablation aerosol 1100 W 0.5 l min1 SEM/compact discrete dynode ion counters/Faraday cups Faraday 232Th and 238U and IC 202 Hg, 204Pb, 206Pb, 207Pb, 208 Pb #N.A.

0.131 s IC1–2: 20 ns; IC3–4–5: 70 ns and IC6–7–8: 5 ns

The data are processed using an in-house spreadsheet modied from Gerdes and Zeh (2006),15 and the main corrections include background, downhole fractionation, instrumental mass bias dri and common Pb (Table 1). Because detectors have different sensitivities for different masses, raw isotopic and elemental ratios are never in agreement with the true values. For ICP-MS, this problem has been

J. Anal. At. Spectrom.

detected since the earlier studies were published in the literature.1,2 Under optimum conditions, there is a 10% to 20% difference between the measured ratio and the accepted (true) value of the reference material (compare for instance Fig. 3a and b). The source for this is debatable but may include differential ion transmission and space-charge effects.42 This is however corrected for standards and unknowns by normalizing the data against the data from the primary reference material. Fig. 3b shows the data already normalized against GJ-1 and applied to standards Pleˇsovice, Temora and BB9. The normalization according to the eqn (1) corrects for mass bias and fractionation. R(sample)true ¼ R(sample)measured  [R(ref. mat.)true/ R(ref. mat.)measured]

(1)

Laser-induced fractionation is another important factor to consider during LA-MC-ICP-MS data reduction.30 Even at a low level of ablation uence (2 mJ cm2) and frequency (6 Hz), some degree (

Suggest Documents