Accepted Manuscript Effects of thermal annealing and chemical abrasion on ca. 3.5Ga metamict zircon and evidence for natural reverse discordance: Insights for U-Pb LA-ICP-MS dating
Daniel Wiemer, Charlotte M. Allen, David T. Murphy, Irina Kinaev PII: DOI: Reference:
S0009-2541(17)30377-7 doi: 10.1016/j.chemgeo.2017.06.019 CHEMGE 18376
To appear in:
Chemical Geology
Received date: Revised date: Accepted date:
18 February 2017 9 June 2017 15 June 2017
Please cite this article as: Daniel Wiemer, Charlotte M. Allen, David T. Murphy, Irina Kinaev , Effects of thermal annealing and chemical abrasion on ca. 3.5Ga metamict zircon and evidence for natural reverse discordance: Insights for U-Pb LA-ICP-MS dating, Chemical Geology (2017), doi: 10.1016/j.chemgeo.2017.06.019
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Effects of thermal annealing and chemical abrasion on ca. 3.5 Ga metamict zircon and evidence for natural reverse discordance: insights for U-Pb LAICP-MS dating
SC RI PT
Daniel Wiemer1*, Charlotte M. Allen1,2, David T. Murphy1 and Irina Kinaev2
1
Queensland University of Technology, 2 George St, QLD 4000, Brisbane, Australia
2
Central Analytical Research Facility, 2 George St, QLD 4000, Brisbane, Australia
NU
*Corresponding author, e-mail:
[email protected]
MA
Abstract
ED
We present a microstructural and U-Pb systematics study comparing pristine, thermally annealed (TA) and chemically abraded (CA) ~3500 Ma zircon from a quartz-dioritic gneiss,
PT
with the aim to improve pre-analytical workflows for more accurate and precise LA-ICP-MS U-Pb dating of ancient zircon. Four zircon domains are identified: i) low- to medium-U
CE
concentric-oscillatory zoned cores, ii-iii) two porous high-U outer alteration domains, and iv)
AC
low-U narrow inward growing recrystallization rims. Raman spectroscopy on the pristine zircon reveals a positive correlation between increasing structural damage and U content. The porous high-U outer domains show a drastic increase in non-formula Ca above estimated amorphous fractions of ~0.8, which we ascribe to hydrothermal alteration of high-damage zircon characterized by percolating networks of amorphous areas. Upon treatment (TA, CA), hyper-spectral CL and Raman spectroscopy suggest structural recovery of point defects, but not full repair of high-damage amorphous areas. Calculated Raman radiation damage ages
1
ACCEPTED MANUSCRIPT suggest that natural annealing affected all domains at ~500 Ma, consistent with negligible laser ablation matrix effects comparing pristine and treated cores. We show that reverse discordance in pristine and TA cores is not an analytical artifact. HighU alteration domains are normal discordant and were partially reset at ca. 3410 Ma. Upon CA, U-Pb discordance and scatter are reduced in the cores, yielding intercepts of 3503 ± 14
SC RI PT
Ma and 560 ± 550 Ma (MSWD=1.0). The lower intercept matches the timing of Pb-loss recorded in the alteration domains and the Raman radiation damage age. We argue that short distance Pb redistribution within the cores, which led to reverse discordance, was controlled by U-Pb systematics recording both the crystallization age and the timing of Pb
NU
redistribution. Some excess Pb was redistributed from the partially re-set alteration domains into the cores, leading to mixed U-Pb systematics and further reverse discordance. Removal
MA
of the latter excess Pb upon CA suggests that radiogenic Pb from the alteration domains accumulated within distinct damage sites accessible for CA leaching.
ED
We conclude that standardization of pre-analytical treatment is not recommended; we propose to investigate the structural state of unknown specimen by means of Raman
PT
spectroscopy, in combination with commonly applied CL imaging. During LA-ICP-MS
CE
analyses, co-measurement of non-formula and selected trace elements is deemed highly
AC
valuable in detecting ablations for use in reliable U-Pb age determination.
Keywords: U-Pb LA-ICP-MS dating; metamict Archean zircon; thermal annealing; chemical abrasion; reverse discordance;
1. Introduction
2
ACCEPTED MANUSCRIPT Uranium-lead geochronology of zircon represents a reliable, and widely used method to determine the absolute timing of magmatic crystallization and metamorphic events recorded in rocks (e.g. Compston et al., 1992; Davis et al., 2003). Zircon provenance studies from sedimentary rocks, and the increasing global zircon record have greatly improved our understanding of secular continental crustal growth (e.g. Campbell and Allen, 2008; Condie
SC RI PT
and Aster, 2009; Belousova et al., 2010; Cawood et al., 2013; Condie, 2014), and as the oldest physically preserved material on Earth (Jack Hills; e.g. Froude et al., 1983; Compston and Pidgeon, 1986; Wilde et al., 2001), zircon has become the key to insights into early Earth crustal formation and evolution, and ambient planetary surface conditions (e.g. Wilde et al.,
NU
2001; Valley et al., 2002, 2014; Bell et al., 2011; Bizzarro et al., 2012).
However, accurate and precise age determination of zircon remains biased particularly by
MA
natural discordance of the 238U-206Pb and 235U-207Pb systems through partial isotopic resetting usually ascribed to loss of Pb, resulting in normal discordance (e.g. Mattinson et al.,
ED
1996; Metzger and Krogstad, 1997). The fundamental cause of elemental and isotopic redistribution in zircon is based on the physical modification of its crystal structure induced
PT
mainly by the alpha decay of 238U (e.g. Davis and Krogh, 2000; Marsellos and Garver, 2010).
CE
Physically damaged zircon is characterized by a reduced chemo-physical durability, subject to diffusion, leaching and recrystallization processes, even under low-temperature conditions
AC
(e.g. Mezger and Krogstad, 1997; Davis and Krogh, 2000; Geisler et al., 2001a; Geisler, 2002), far below the closure temperature for crystalline zircon (>900 °C; e.g. Cherniak and Watson, 2000). More rarely, excess in radiogenic Pb relative to parental U is observed. Such reverse discordance has been ascribed to either analytical artifacts, or natural intra-grain redistribution of Pb from high- to low-U zircon zones (Mattinson et al., 1996; Carson et al., 2002; Zhao et al., 2014). Recent analytical advances attest for radiation damage related Pb mobility on the nano-scale, forming isolated clusters of high 207Pb/206Pb ratios, while on the
3
ACCEPTED MANUSCRIPT µm-scale of commonly used U-Pb analytical spots closed-system behavior is observed (Valley et al., 2014; Kusiak et al., 2015; Peterman et al., 2016). The extent of intra-grain Pb redistribution in affecting common U-Pb analyses, and potentially resulting in reverse discordance, remains poorly understood. The cause of both normal and reverse U-Pb isotopic discordance is seemingly case-specific,
SC RI PT
though often not investigated in detail, and rather treated as a common phenomenon. Elimination of discordance through pre-analytical thermal annealing, and mechanical and chemical abrasion techniques to recover the crystal structure and to remove affected portions of zircon grains have been successfully applied to ID-TIMS (isotope dilution - thermal
NU
ionization mass spectrometry) analysis (e.g. Krogh and Davis, 1974; 1975; Krogh, 1982; 1994; Mattinson, 1997; 2005). For the faster and more cost-efficient SHRIMP (sensitive
MA
high-resolution ion microprobe) and LA-ICP-MS (laser ablation inductively coupled plasma mass spectrometry) techniques, in which only small volumes of the zircon are targeted for U-
ED
Pb, pre-analytical workflows are debated. Mattinson (2005) proposed that thermal annealing alone could sufficiently reduce discordance to improve precision and accuracy in LA-ICP-
PT
MS apparent ages. In fact, Allen and Campbell (2012) demonstrated that the structural
CE
recovery upon thermal annealing reduced a radiation-damage induced matrix effect during laser ablation. Chen et al. (2002) performed step-wise chemical abrasion on thermally
AC
annealed metamict zircon, and found no improvement, until affected discordant domains were completely removed. In contrast, Crowley et al. (2014) recently argued that CA should be applied. In the present contribution, we investigate the effects of TA and CA on the U-Pb systematics and the microstructure of metamict Archean zircon to improve LA-ICP-MS pre-analytical workflow. Pristine and experimental grains are structurally and chemically characterized and compared using CL imaging, Raman spectroscopy, and LA-ICP-MS chemistry and U-Pb
4
ACCEPTED MANUSCRIPT isotopic composition. The aim of the study is to constrain the effect of pre-analytical treatment, and its usefulness in LA-ICP-MS age determination of particularly ancient zircon. In addition, we aim to investigate the mode and cause of discordance.
SC RI PT
2. Analytical and experimental details
2.1 Sample background
In the present experiment, we use zircon from a quartz-dioritic gneiss (DW-370-1-14) from
NU
the southwestern margin of the multiphase Archean Muccan Granitic Complex, East Pilbara Terrane (Western Australia). The rock sample was collected ca. 200 m south of the Newman-
MA
Tabba-Tabba Road (138), east of 1 Mile Creek (20°52’31.98” S, 119°43’56.67” E; Supplementary Fig. S-1), and belongs to one of the ‘older’ gneiss components described in
ED
Wiemer et al. (2016). The quartz-dioritic gneiss comprises a mineral assemblage of hornblende, plagioclase, quartz, biotite, minor epidote, titanite and opaque phases, and
PT
accessory zircon and apatite. We selected this sample for the present experiment due to its
CE
morphologically homogeneous population of relatively large (~400 µm) zircon.
AC
2.2 Zircon separation, sample preparation, and experimental conditions
Zircon grains were extracted from bulk-rock at the University of Queensland (UQ) using standard techniques of mechanical, magnetic and density separation. From 240 selected zircon grains, about 100 grains were left untreated to characterize pristine zircon, and about 140 grains were selected for pre-analytical thermal annealing (TA).
5
ACCEPTED MANUSCRIPT TA was performed in ceramic crucibles with lid at 850 ºC for 60 hours in a Ar-atmosphere furnace to prevent oxidation. The samples were cooled within the oven over a few hours. Due to previously observed changes in laser ablation behavior and down-hole fractionation between untreated and treated zircon (Allen and Campbell, 2012; Crowley et al., 2014), fractions from the reference materials (Plešovice, 337.1 Ma, 465-3084 ppm U; Sláma et al.,
SC RI PT
2008, and Temora2, 417 Ma; 82-320 ppm U; Black et al., 2004) were thermally annealed under the same conditions to create a set of physically matching standards (Crowley et al., 2014). Archean zircon reference material OG1 (207Pb/206Pb age of 3465 ± 0.6 Ma for chemically abraded, almost concordant thermal ionization mass spectrometry analyses; Stern
NU
et al., 2009) was analyzed in a pristine state.
A total of 40 of the thermally annealed grains of unknown sample 370-1 were selected for
MA
chemical abrasion under two experimental conditions (CA-I and CA-II, ~20 grains each). In both experiments the grains were put into Savillex Teflon vials, and partial dissolution was
ED
achieved within a diluted hydrofluoric acid solution of 0.2 ml HF and 1.8 ml 1:1 HNO3 + H2O, using a Milestone bench-top Ultra-wave single reaction chamber microwave digestion
PT
system, with an external chamber microwave absorbing base solution of 25 ºC and 40 bar.
CE
The experiments were performed at 1500 W microwave power under following internal chamber conditions: CA-I, 2 steps (10 min) at 175 ºC and 60 ºC, respectively, both steps at
AC
120 bar; CA-II, 2 steps (15 min) at 200 ºC and 70 ºC, respectively, both steps at 110 bar. The samples were then cleaned through repeated i) adding of 6ml MilliQ-H2O, ii) centrifuging for 15min, and iii) pipetting off of ~7ml. After repeated cleaning, the residues were evaporated dry on a hot plate at 100 ºC. After thermal annealing (TA) and chemical abrasion (CA-I, CA-II), both pristine and treated grains were mounted in epoxy, aligned with their crystallographic c-axis parallel to horizontal. Prior to polishing, reflected and transmitted light images were acquired. For
6
ACCEPTED MANUSCRIPT electron beam excited imaging (scanning electron microprobe/SEM, electron microprobe/EMPA), polished mounts were carbon-coated. The C-coating was removed prior to Raman spectroscopic and LA-ICP-MS analyses.
SC RI PT
2.3 Instrumentation
For reflected and transmitted light imaging of the entire zircon mount, a ZEISS Axio Imager M2m microscope system was used. Cathodoluminescence (CL) imaging was performed with a ZEISS Sigma field emission scanning electron microscope (SEM) in variable pressure (VP)
NU
operation mode at 20 kV and 1.2 nA. Backscattered electron (BSE), hyper-spectral CL, and x-ray (EDS) mapping of selected elements were acquired with a JEOL JXA 8530F
MA
Hyperprobe field emission electron microprobe (EMPA), equipped with both a standard and high-resolution xCLent spectroscopic CL detectors. For the spectral CL imaging a dwell time
ED
of 40 ms were allowed and operating conditions held at 20 kV and 20 nA. The spectral CL data were evaluated with the xCLent V software.
PT
Inelastic Raman scattering was stimulated at room temperature using a Renishaw inVia
CE
Raman microscope fitted with a green 530 nm excitation laser, edge filter, orthogonal polarization, and 1800 l/mm spectrographic gratings, and focused beam size of 5 µm. Spectra
AC
were acquired perpendicular to the zircon c-axis, for Raman shift emissions from 600 to 1400 cm-1 with 16 seconds exposure and 4 accumulations. Raman data were treated with the GRAMS/AI Thermo Scientific spectroscopic data processing software, including background and multipoint baseline correction routine. Band position, height, and width were determined using an iterative Gaussian peak fitting above a calculated linear baseline with minimization of the reduced Chi2 value, giving a ‘goodness-of-fit’.
7
ACCEPTED MANUSCRIPT Isotopic concentrations were determined with LA-ICP-MS. Selected analytical 30 µmdiameter spots were ablated using a 193 nm excimer laser. Ablated material was transported in He carrier gas and analyzed with an Agilent 8800 triple quadrupole ICP-MS. Zircon standards Temora2 and Plešovice were used as primary and secondary reference material, respectively. The NIST 610 glass (Jochum et al., 2011) was used as a primary standard for
SC RI PT
trace element composition. LA-ICP-MS raw data counts were imported to the Igor Pro-based Iolite software (WaveMetrics) for data de-convolution and evaluation, including baseline correction, integrations for U-Pb primary reference standards, down-hole fractionation correction and propagated error calculation (Paton et al., 2010). The LA-ICP-MS data were
NU
reduced relative to two different reference materials, pristine and annealed Temora2 in order to investigate the effects of treatment on the reference materials as well as the Archean
MA
unknown. All other parameters in the data reduction routine (Iolite) were held constant. U-Pb
PT
3. Results and discussion
ED
Concordia diagrams were plotted in the Isoplot Excel add-in.
CE
In the following sections analytical results are presented and discussed to structurally and chemically characterize the zircon and quantify radiation-induced structural damage and
AC
recovery thereof upon experimental TA and CA. Isotopic U-Pb systematics of pristine and experimental zircon reference standards and unknowns are investigated to distinguish potential laser ablation analytical artifacts from radiation-induced isotopic disturbance that can be directly associated with the zircon evolution. Morphological and optical characteristics are provided as Supplementary Material S-2 in the online version.
8
ACCEPTED MANUSCRIPT 3.1 Identification and characterization of distinct CL zircon domains and effects on CL upon annealing
The CL imaging reveals grain interiors (core domain) characterized by mostly concentric zoning of alternating medium to light CL bands, parallel to the crystal faces, consistent with
SC RI PT
magmatic derivation. Rare sector zoning and convolute zoning are observed. Most grains display four distinct domains best observed in treated zircons, as within pristine grains the CL contrast is extremely faint using the scanning electron microscope (Fig. 1a-e): An inner core with moderate CL response,
A domain of high CL response (alt-1) that forms a concentric band around the core, and
NU
distribution of CL response,
MA
showing a mostly sharp outer boundary, and is characterized by a heterogeneous
A domain of low CL response (alt-2), marking the outermost concentric band of the
A thin outermost zircon shell (rim domain) of medium-CL that is similar in intensity to the core domain.
PT
ED
grains, and is characterized by the lowest CL intensity observed,
CE
The heterogeneity in CL intensity within domain alt-1 can be ascribed to two features: i) some regions are characterized by a spongy texture, where more or less homogeneously
AC
distributed dark CL (no CL response?) holes of up to 1 µm size are observed within overall high CL (Fig. 1c), and ii) regions, in which CL intensity gradually increases away from either boundary parallel, or perpendicular dark CL/no CL fractures, occasionally resulting in an overall convoluting texture formed by varying transitional CL intensities, apparently controlled by the fracture network (Fig. 1d). Many of the low-CL perpendicular fractures affecting the inward core domain, seem to terminate in domain alt-1 (Fig. 1c). Other, in- and outward extending fractures display the same high-CL intensity as domain alt-1. These
9
ACCEPTED MANUSCRIPT fractures extend into low-CL core domains and/or into fractures or zones parallel to the concentric zoning bands of the cores, connecting low-CL zones within the zircon (Fig. 1b, d). Similar to domain alt-1, domain alt-2 comprises regions of a spongy heterogeneous CL texture. Here, however, the latter texture is much more developed, showing tight
SC RI PT
accumulations of “CL holes” that are confined to very restricted regions and/or domainparallel bands and zones (Fig. 1e). Towards the outer extent of domain alt-2, some larger (~5-10 µm) irregular-shaped areas of absence of material, connected to outward-penetrating fractures, occasionally obscure the band and cannot be correlated with any primary crystal-
NU
specific features (Fig. 1e). The latter outer zircon domains alt-1 and alt-2 are mostly present in grains that overall show darker CL, where both domains reach width of up to 20 µm, while
MA
in light CL zircon grains the domains are much thinner (<5 µm), or absent. The rim domain defines the well-developed crystal faces at its outward limits, but displays a
ED
curved inner boundary that penetrates into the interior of the zircon (Fig. 1e). Overall, the rim domain shows relatively consistent widths of ~5-10 µm, but very rarely it reaches into the
PT
zircon’s interior where it replaces larger portions of affected grains. The rim domain shows
CE
intensive radial fracturing. Most of the fractures are restricted to the rim, or terminate in the adjacent alt-2 domain (Fig. 1).
AC
The back-scattering electron (BSE) response correlates negatively with CL intensity in most instances, with high BSE intensity in zones of low CL intensity, and vice versa (e.g. Nasdala et al., 2002; Fig. 2). However, this seems not to be true for the alt-2 domain within the pristine grains, where low-CL corresponds to low BSE, or absence thereof (Fig. 2, top row). Only upon annealing do the dark CL alt-2 domains display expected bright BSE intensity. The alt-1 domain shows lowest BSE, even in the pristine grains, where its high CL intensity relative to the cores, as visible in the TA grains, is not too conspicuous.
10
ACCEPTED MANUSCRIPT
3.1.1 Change in CL upon annealing
The drastic increase in the integral panchromatic CL intensity upon annealing (TA) was investigated further using hyper-spectral CL. Spectral intensities were acquired for a photon
SC RI PT
energy range of 1.7 to 3.5 eV. Representative spectra from all domains (core, alt-1, alt-2, rim) are compared in Figure 3 from one pristine and one TA grain, highlighting the overall increase in CL intensity of almost up to an order of magnitude within rims and alt-1, and ca. five times higher intensities in cores and alt-2 compare to pristine grains. The spectra are
NU
dominated by the presence of very broad asymmetric emission bands, covering almost the entire spectral range, and additional narrow emission lines are superimposed on the broad
MA
bands of the alt-1, alt-2, and core domains of the TA grains (Fig. 3). The luminescence centers responsible for the here observed asymmetric broad band emissions (Fig. 3) could not
ED
be resolved unambiguously through Gaussian curve fitting due to multiple solutions. However, in a simple, 2-component solution, two overlapping broad band centers, one in the
PT
yellow-green range (A in Fig. 3) for the pristine zircon, the other in the blue range (B in Fig.
CE
3) for the thermally annealed zircon, are identified. Following previous studies, we propose that the broad blue CL band (Fig. 3) is of intrinsic nature, and its increasing intensity upon
AC
TA is the result of point defect recovery associated with SiO4 groups and overall reestablishment of the crystal field (e.g. Kempe et al., 2000; Nasdala et al., 2003). The yellowgreen broad band that particularly dominates the spectral range of the pristine cores (Fig. 3) possibly represents CL emission from electron transitions associated with additional radiogenic point defects, which healed during the thermal annealing (Gaft, 1992; Kempe et al., 2000; Tsuchiya et al., 2013, abstr.).
11
ACCEPTED MANUSCRIPT A striking change upon TA is the appearance of narrow high intensity emission lines, particularly in the alt-1 and alt-2 domains relative to the broad band emissions. The narrow bands can be ascribed to the presence of various extrinsic trivalent rare earth element (REE) impurities, based on their specific spectral peak positions (e.g. Götze and Kempe, 2009). The following extrinsic REE3+ luminescence centers are identified based on their peak position in
SC RI PT
the alt-1 and alt-2 domains of the TA grains: Er3+ duplet at ~405 nm, Dy3+ triplet at ~475 nm, an undefined duplet at ~485 nm, Er3+ multiplet at ~525 nm, Tb3+ at ~557 nm, Dy3+ multiplet at ~580 nm, and Sm3+ at 617 nm (Fig. 3). The superposition of these peaks on the broad emissions is not observed in the rims and any of the pristine grain domains. The annealed
NU
cores show minor peaks of the 475 nm, 485 nm, 557 nm, and 580 nm REE3+ emission lines. In both the TA alt-1 and alt-2 domains the relative intensities between the REE3+ emissions
MA
are approximately constant, while the relative intensities between the REE3+ peaks show a slightly different pattern in the TA cores (Fig. 3), possibly due to varying concentration
ED
levels of REE.
Previous CL and photoluminescence studies suggested that a decrease in REE emissions
PT
could be the result of incorporation of non-formula elements, and/or increasing radiation
CE
damage (e.g. Lenz and Nasdala, 2015). In the present study, semi-quantitative X-ray maps of selected elements (Fig. 4a) and corresponding grain profiles (Fig. 4b) demonstrate the
AC
presence of elevated non-formula element content (Fe and Ca) within both alt-1 and alt-2 domains. The alteration domains show the most intensive REE CL emission peaks upon TA, compared to the cores and rims that do not show significant contents of non-formula elements (Fig. 4). Therefore, it is more likely that the suppression of the REE emissions in the pristine grains is dominated by radiation-induced structural damage (Nasdala et al., 2002).
3.2 Microstructural state and recovery revealed by Raman spectroscopy
12
ACCEPTED MANUSCRIPT
Raman vibrational spectra on each of the four identified domains from each experiment have been acquired, in order to determine the degree of structural damage and test the proposed recovery thereof, upon TA. Representative Raman spectra of each domain from both a pristine (pr04) and a TA grain (TA94) are displayed in Fig. 5a-c for comparison. A list of
SC RI PT
identified peaks for each spot is given in the Supplementary Table S-1.
The Raman spectra (Fig. 5a) clearly show an increase in the sharpness and intensity of two dominant peaks, situated at a spectral frequency of around 970 and 1000 cm-1, respectively, in the TA grain domains compared to the pristine zircon. The relative increase in intensity of
NU
the two peaks is different, with the higher frequency (~1000 cm-1) peak displaying drastic intensity increase, while the peak situated at ~970 cm-1 shows relatively minor increase.
MA
Within the pristine grains, the two peaks show similar intensities and are well developed within the rims, and cores, but show poor intensity and increased broadness within alt-1 and
ED
alt-2. The increasing sharpness and intensity from pristine to thermally annealed is most drastic in the alt-1 domain, where the peak at ~1000 cm-1 reaches the highest intensity
PT
observed. Another important observation is the systematic shift towards higher frequencies
CE
with increasing intensity and sharpness in the TA domains. Within the pristine grains, the two dominant peaks are situated at 962-972 cm-1 and 996-1004 cm-1, respectively, with the
AC
highest values in the rims and cores, shifting towards 973-979 cm-1 and 1004-1008 cm-1, respectively, with the highest frequencies in the rims, cores and alt-1 in both the TA and CAI/CA-II grains. We focus on the ~1000 cm-1 peak (Table 1) that has been well characterized in previous studies (internal stretching mode of Si-O bonds, v3 SiO4, e.g. Dawson et al., 1971). In crystalline synthetic zircon, this v3 SiO4 mode shows intensive Raman scattering of a phonon frequency at 1008 cm-1 characterized by a very narrow full-width at half-maximum
13
ACCEPTED MANUSCRIPT (FWHM~1) peak (e.g. Geisler et al., 2001b). In contrast, the v3 SiO4 peak of metamict zircon is less intense, broader, and situated at lower phonon frequencies (e.g. Geisler et al., 2001b). The destruction of the zircon crystal structure during metamictization is caused by simultaneous accumulation of both point defects, and formation of amorphous domains (e.g. Murakami et al., 1991; Weber et al., 1994; Nasdala et al., 1995; Geisler et al., 2001b;
SC RI PT
Trachenko et al., 2001; Devanathan et al., 2006). Geisler et al. (2001b) compared the Raman response of variously naturally damaged zircon with metamict zircon annealed at different temperatures and duration. The results are displayed as the radiation damage trend and the annealing trend in the frequency versus line-width (FWHM) plot of the v3 SiO4 internal
NU
stretching mode (Fig. 6). In combination with XRD powder diffraction, the authors demonstrated that the structural recovery upon annealing is a two-stage process that differs
MA
from the process of metamictization. This means that annealing radiation damage in zircon does not directly reverse the process of damage accumulation. During the first stage, at
ED
annealing temperatures up to ~1000 K, Geisler et al. (2001b) observed significant decrease in unit-cell volume (XRD) with minor decrease in FWHM, while during the second stage
PT
(>1000 K), growth of crystalline domains corresponded to drastic decrease in FWHM. Due to
CE
the fact that the volume swelling is caused by overlap of damaged regions characterized by point defects, and not a result of the structure of amorphous domains (e.g. Trachenko et al.,
AC
2001), the slightly decreasing FWHM and increasing frequency during stage 1 is interpreted as the recovery of point defects in crystalline domains, while during stage 2 amorphous domains recrystallize (e.g. Geisler and Pidgeon, 2002). The pristine domains follow the radiation damage trend (Fig. 6). However, zircons from the highly damaged alt-1 and alt-2 domains plot above the damage trend, and at correspondingly high FWHM values (Fig 6). This may indicate some degree of natural annealing of the finer structure point defects (e.g. Geisler et al., 2003a,b,c). All of the annealed domains plot at high
14
ACCEPTED MANUSCRIPT frequencies (>1004 cm-1; Table 1; Fig. 6), indicating significant recovery of short-range point defects, but still displaying a scatter in the FWHM. Hence, amorphous domains are likely still present. Because different grains are used for each experiment, and due to the fact that the initial damage state determines the annealing process (Zhang et al., 2000), it remains unclear, if amorphous domains recrystallized at all at the used annealing conditions (850 °C,
SC RI PT
60 h). For example, some of the annealed cores plot at higher FWHM than others. This hypothesis agrees with our CL images, in which some cores show darker CL in the TA grains, possibly reflecting higher initial damage. This means that some of the higher frequency, low FWHM annealed domains did not necessarily experience the second
NU
annealing stage, but started off from lower initial damage, somewhere on the damage trend line (Fig. 6). The alt-1 domains must have reached a critical level of amorphization, and the
MA
amorphous regions could not be healed under the experimental annealing conditions. The presented evidence for increased structural damage within the alt-1 and alt-2 domains
ED
can explain the development of some of the observed fractures (Fig. 1). The distribution and alignment of radial micro-cracks within the outermost rims, compared to rather concentric
PT
fractures along the inward boundaries of alt-1 and alt-2, fit numerical models by Lee and
CE
Tromp (1995) based on predicted internal stress distribution due to the macroscopic volume
AC
swelling of highly damaged domains (up to 18%; Weber, 1993).
3.3 Insights on structural damage from zircon chemistry
In this section zircon chemical data of selected elements are presented to further estimate the radiation damage based on actinide content and to chemically characterize the domains. Quantification of non-formula element incorporation is discussed in the light of chemical
15
ACCEPTED MANUSCRIPT disturbance and potential alteration mechanisms in correlation to the microstructural damage. LA-ICP-MS zircon chemical data is given in the Supplementary Table S-2.
3.3.1 Actinide content and alpha-dose
SC RI PT
Uranium concentrations range widely in the studied zircons, but can be correlated with the CL domains. For the pristine grains, average U concentrations are 141 ± 79 ppm for core, 437 ± 212 ppm for alt-1, 2,545 ± 1,211 ppm for alt-2, and 140 ± 77 ppm for rim domains. Thorium concentrations correlate positively with U content, but overall show minor
NU
variations. In Figure 7a, Th versus U concentrations are plotted for pristine grains only, but treated grains show similar trends. Cores plot along relatively consistent “magmatic” Th/U
MA
values (~0.37; e.g. Belousova et al., 2002), while the rims and the highly damaged alt-1 and alt-2 domains plot at lower Th/U values (~0.1). Secondary loss of Th or gain of U relative to
ED
each other is rather unlikely in this context, as both elements should behave relatively similar during alteration (e.g. Geisler et al., 2002). Rather, the low Th/U ratios within the alt-1 and
PT
alt-2 domains are probably features related to late-stage magma composition, after Th was
CE
exhausted, or incorporated preferentially, for example, in competing, co-existing apatite that is found abundant in the host rock and often shows intergrowth with zircon.
AC
In order to correlate the U and Th contents to the structural damage revealed by Raman spectroscopy, alpha-doses (D) were estimated from the equation given by Holland and Gottfried (1955):
𝐷 = 8 238𝑈[𝑒𝑥𝑝(λ238 ∗t) − 1] + 7 235𝑈[𝑒𝑥𝑝(λ235 ∗t) − 1] + 6 232𝑇ℎ[𝑒𝑥𝑝(λ232 ∗t) − 1] [1]
16
ACCEPTED MANUSCRIPT with λ, decay constants for 238U, 235U, and 232Th, respectively (Steiger and Jäger, 1977); 238U, 235
U, and 232Th, actinide content in atoms/mg. An age (t) of 3.50 Ga was used, based on
preliminary U-Pb age results (Section 3.5). The range in estimated alpha-doses for each domain are (average values in brackets): cores 0.7-13.3 1015α/mg (3.6), alt-1 3.1-34.4 1015α/mg (14.4), alt-2 4.5-93.4 1015α/mg (35.6), and rims 0.8-5.8 1015α/mg (3.2),
SC RI PT
respectively.
Previous X-ray diffraction studies have shown that the alpha-dose dependent progressive metamictization advances in three stages (Murakami et al., 1991; Pidgeon, 2014). In the first stage, up to D ~3*1015 a/mg, dominantly point defects accumulate, resulting in the unit cell
NU
expansion; in the second stage, up to D ~8*1015 a/mg (corresponding to the percolation point observed in the changing Raman response; Fig. 6), progressive distortion of crystalline
α /mg, zircon is fully amorphous.
MA
remnants and increase of amorphous regions occurs; and in the third stage, above D ~8*1015
ED
The amount of the amorphous fraction (p) can be estimated based on the received alpha-dose
[2]
CE
𝑝 = 1 − 𝑒𝑥𝑝−𝐵(𝐷−𝐷𝑖𝑛𝑐)
PT
by the equation (Geisler et al., 2003a):
AC
with B = 0.28*10-18 as the amorphous mass (g) produced per α-decay, D as the α-dose, and Dinc = 0.47*1018 (α/g) as the incubation dose accounting for natural annealing during the geological history (e.g. Geisler et al., 2003a). In Fig. 7b, the plot of the amorphous fraction p versus the received alpha-dose D for each domain of the pristine grains demonstrates that the identified domains overall represent the full range of possible radiation damage, from crystalline in the cores to almost completely amorphous within the alt-1 and alt-2 domains, if natural annealing has not occurred.
17
ACCEPTED MANUSCRIPT
3.3.2 Non-formula and rare earth element incorporation
In agreement with the X-ray maps (Fig. 4), LA-ICP-MS analyses reveal non-formula Ca concentrations of ~1,600 ppm in alt-1, and ~4,650 ppm with values >10,000 ppm, within alt-
SC RI PT
2. Such high concentrations of Ca must be of secondary origin, likely supplied by alteration fluids into the metamict zircon domains. This is consistent with the proposed natural annealing indicated in the Raman frequency versus line-width plot (Fig. 6). The observed drastic increase in Ca above an amorphous fraction of p~0.8 (Fig. 7c) corresponds to the
NU
critical alpha-dose value of D ~8*1015 α/mg that marks the transition between metamictization stage two and three (Murakami et al., 1991). Based on studies using
MA
percolation-type theory (e.g. Salje et al., 1999; Ríos et al., 2000), the critical dose marks the second percolation point characterized by a transition between a damage state in which
ED
amorphous material starts to form percolating networks, and a damage state in which crystalline material is not connected any more and amorphous material dominates. Such a
PT
structure of connected amorphous domains provides low-temperature diffusion pathways,
CE
promoting elemental redistribution (e.g. Meldrum et al., 1998; Geisler et al., 2001a; Balan et al., 2003; Utsunomiya et al., 2004). The drastic increase in Ca above an amorphous fraction
AC
of p~0.8 fits well with the existence of the second percolation threshold value, previously experimentally determined at p~0.7 (Geisler et al., 2001b; 2003a). The second percolation point at p~0.7 is associated with a dramatic increase in diffusion distance, interpreted as the appearance of high-diffusivity paths of reduced density in the damaged structure due to regions of depleted matter, as a result of alpha-radiation-induced atomic displacement cascades (Geisler et al., 2001a,b; 2003a; Trachenko etal., 2004).
18
ACCEPTED MANUSCRIPT As shown in Figure 7d, the increasing Ca content in alt-1 and alt-2 positively correlates with increasing La (LREE) incorporation. Although progressive substitution of REE with decreasing ionic radii is facilitated during magmatic growth (e.g. Belousova et al., 2002), the coupled incorporation of La (i.e. LREE) and non-formula Ca in the more damaged alt-1 and alt-2 domains likely indicates addition of LREE into the damaged structure during an
SC RI PT
alteration event. This is in agreement with our observation of dark/no CL holes (Fig. 1) indicating porous textures in the alteration domains (e.g. Hay and Dampster, 2009). Solidstate recrystallization of the chemically weakened metamict material through lowtemperature hydrothermal fluid dissolution-reprecipitation likely resulted in the formation of
NU
porous domains (e.g Hay and Dampster, 2009). Besides significant Ca and LREE content, the composition and origin of the alteration fluids is unclear, but given the presence of altered
MA
feldspar in the host rock, a proximal source for the elemental redistribution may be inferred. Such fluid alteration process could have also caused some degree of natural annealing, as
CE
PT
3.4 Radiation damage ages
ED
indicated in the Raman frequency versus line-width plot (Fig. 6).
As indicated in Figure 6, the pristine grains that plot above the un-annealed radiation damage
AC
trend likely experienced natural re-heating during their geological history, partially recovering the zircon structure. The quantification of the degree of structural damage in regards to known alpha flux determined by actinide content can provide constraints on the conditions and timing of natural annealing of zircon (e.g. Holland and Gottfried, 1955; Pidgeon et al., 1998; Pidgeon, 2014). The dependence of the v3 SiO4 line-width (i.e. FWHM) on the α-dose has been used to establish a calibration curve for un-annealed Sri Lankan zircon (Nasdala et al., 2001), where
19
ACCEPTED MANUSCRIPT an exponential relationship of damage accumulation determined by Raman spectroscopy is given as (Palenik et al., 2003):
𝐹𝑊𝐻𝑀 = 𝐴[1 − 𝑒𝑥𝑝(−𝐵𝐹𝑊𝐻𝑀 𝐷𝑒𝑑 ) ]
[3]
SC RI PT
with A, asymptotically approached FWHM value of 35.64; B, related to mass of material damaged per alpha decay event (computed at ~5.49 x 10-19/g); Ded, equivalent damage dose determined from U and Th contents and age of radiation damage (375 Ma) for calibration using Sri Lankan zircon. Solving equation [3] for the equivalent damage dose Ded allows the
NU
calculation of radiation damage ages of unknown specimen, using U and Th concentrations representative of the Raman point analyses (Pidgeon, 2014).
MA
The relationship between the v3 SiO4 FWHM and D for a selection of representative pristine zircon is shown in Figure 8, which confirms that at a given alpha-dose (D) the broadness of
ED
the v3 SiO4 band lies below the calibration curve (Ded), hence natural annealing occurred (e.g.
PT
Nasdala et al., 2001). Calculated radiation damage ages yield relatively consistent values for all pristine zircon domains (Table 2), with an average damage accumulation age of 504 ± 19
CE
Ma (n=9). Note that doses calculated from the Raman relationship yield slightly higher values than the estimated total alpha dose for low damage core analyses (spot 01-2, 04-3).
AC
This is ascribed to the fact that the calibration curve does not take into account a minimum FWHM (1.8 cm-1) at zero alpha-dose (Nasdala et al., 2001; Pidgeon, 2014). Even the highly damaged alteration domains alt-1 and alt-2 show some degree of natural annealing. It is expected that at natural low-temperature conditions, primarily point defects recovered along a frequency – line-width vector similar to the annealing trend in Figure 6 (Geisler et al., 2001b). Both the state of structural distortion prior to ~504 Ma, and the degree of recovery (i.e. vector length in Fig. 6), remain speculative. Renewed accumulation of
20
ACCEPTED MANUSCRIPT radiation damage after natural annealing would have shifted the frequency – line-width relationship to its present position, along a vector defined by the radiation damage trend. Because the degree of long-range order defect recovery, if occurring at all during initial (i.e. low-temperature) annealing, is poorly understood, the impact of natural annealing on the present amorphous fraction remains unclear. Our observations of the second percolation point
SC RI PT
for enhanced non-formula element diffusivity above an amorphous fraction of p~0.8, may approach the value (p~0.7) determined by Geisler et al. (2001b; 2003a), when accounting for possible long-range order recovery during natural annealing.
Radiation damage ages provide reliable last cooling ages that can be interpreted in a similar
NU
way to Ar-Ar or Rb-Sr cooling ages (Nasdala et al., 2001; Pidgeon, 2014). The radiation damage ages calculated here for the East Pilbara quartz-diorite gneiss from the Muccan
MA
Granitic Complex are in agreement with radiation damage ages of 420 ± 110 Ma calculated for zircon from the Darling Ranges, Yilgarn Craton, Western Australia (Pidgeon, 2014). The
ED
~420 Ma ages confirm the conditions of natural zircon annealing, evident in Rb-Sr age isopleths for biotite (400-500 Ma; Nemchin and Pidgeon, 1999; Pidgeon, 2014).
PT
It should be noted that natural annealing at ~500 Ma does not exclude the occurrence of prior
CE
annealing events. Similarly, the full extent of non-formula and LREE incorporation through hydrothermal fluids cannot be ascribed to the Phanerozoic event alone. As evident in the
AC
estimated total α-doses (D), it is expected that at least the high-U zircon domains were already highly damaged prior to ~500 Ma. Multiple alteration and annealing events may have contributed to the present chemical and structural state of the zircon.
3.5 U-Pb isotopic systematics
21
ACCEPTED MANUSCRIPT Below, we systematically investigate the effects of pre-analytical treatment on the U-Pb systematics measured by LA-ICP-MS. First, the effects on annealed and pristine standards Temora2, Plešovice and OG1 are examined relative to their accepted TIMS U-Pb values. Then, we will compare the effects of using annealed and pristine reference standard Temora2 as applied to our unknown pristine zircon cores. Pristine grains are used to distinguish
SC RI PT
between analytical artifacts (i.e. ablation matrix effects) and natural characteristics in regards to the nature and degree of U-Pb discordance. Finally, we will assess the U-Pb data of pristine and experimental grains, integrating the physico-chemical state of the studied unknown zircon domains, towards the reliability and accuracy of recorded apparent ages. U-
NU
Pb data are given in the Supplementary Table S-3.1 and S-3.2 using annealed and pristine
MA
Temora2, respectively, as the primary reference material.
ED
3.5.1 Effects on U-Pb systematics of reference material
known TIMS results
PT
a) Matching of matrices among reference zircons in LA-ICP-MS analyses compared to
CE
Table 3 demonstrates several features well understood about the precision of ICP-MS measurement of zircon ages using excimer lasers, especially U-Pb fractionation (Allen and
AC
Campbell, 2012; Marillo-Sialer et al., 2014; Marillo-Sialer et al., 2016). For Temora2 and Plešovice zircons for both pristine and annealed types, Table 3 gives the TIMS (“accepted”) 206
Pb/238U, 207Pb/235U and 207Pb/206Pb values, and fractionation factors, defined here as the
accepted divided by the average measured ratios. These zircon samples are both Paleozoic or “middle aged” reference materials that are old enough to have built up measurable radiogenic Pb, but young enough to not be overly damaged by radioactive decay products. Note that although Plešovice is younger (337.1 Ma; Sláma et al., 2008), it has a greater U+Th content
22
ACCEPTED MANUSCRIPT and therefore has a higher calculated radiation dose, which is about twice that of Temora2 (Black et al., 2004). Down-hole variation of Pb-U ratios is observed, as is instrument drift, but to emphasize the effects of relative standardization, what is reported here is the average ablation, then averaged as a group (n=12 to 15; Table 3). Once instrumental settings are fixed, most fractionation is caused at the ablation site and is controlled by laser-substrate
SC RI PT
interactions, which depend on structural integrity and trace element content (Marillo-Sialer et al., 2014; Marillo-Sialer et al., 2016). More damaged grains should appear older because they liberate their Pb more easily with every ablation pulse (sampled volume) compared to U. During ablation, energy from the laser radiation causes the breakdown of zircon to crystalline
NU
and amorphous ZrO2 and amorphous SiO2, which deposit around the ablation crater (Košler et al., 2005). Preferred incorporation of both U and Pb into the newly formed ZrO2 and SiO2
MA
deposits, results in elemental fractionation characterized by a decoupling of radiogenic Pb that escapes and is measured in higher Pb/U ratios, yielding older apparent ages (Košler et
ED
al., 2005; Allen and Campbell, 2012). Within highly damaged zircon, this effect is enhanced due to the physically and chemically weakened zircon being ablated more readily, as evident
PT
in deeper ablation pits (e.g. Crowley et al., 2014; Marillo-Sialer et al., 2016). As reviewed
CE
earlier, annealing improves crystallinity but does not fully repair it; radiation damage is a non-reversible process.
AC
What can be gleaned from Table 3 is that for this analytical session, in all cases, the measured Pb/U ratios are less than the accepted values and thus the fractionation factors to correct the measured ratio to known values are greater than 1. For annealed Temora2, the fractionation factors for both 206Pb/238U and 207Pb/235U are similar (~1.11) meaning that the correction factor for 207Pb/206Pb is unity, which is expected given that intra-element mass fractionation is not expected, except for mass bias associated with instrument tuning. If annealed Temora2 is used to treat the pristine Temora2, the resulting 206Pb/238U ages are
23
ACCEPTED MANUSCRIPT about 2% too old, and this stems from the fact that after annealing, the reference material is physically different and thus has a different fractionation factor. The annealed grains from Plešovice should be well matched to annealed Temora2 but the calculated 206Pb/238U age is about 3% too old. The most extreme physical mismatch is between annealed Temora2 and pristine Plešovice. Using the tabulated bulk ages, the calculated age for pristine Plešovice is a
SC RI PT
full 8% (28 Ma) too old. Note that the uncertainties on fractionation factors are 1 to 2%, which limits the ultimate quoted accuracy for an individual ablation age even when standards and samples are well matched. The effect of mismatch is shown in Figure 9, in which the individual ablations are plotted having used annealed Temora2 as the reference material. The
NU
TIMS values are shown (orange dots in Fig. 9). Annealed Temora2 plots concordantly where dictated given that it is the reference. Pristine Temora2 appears somewhat older than
MA
accepted and is more scattered. Annealed Plešovice is somewhat older than its accepted age and pristine Plešovice is significantly older than accepted, is scattered, and is reversely
ED
discordant. The bulk 207Pb/206Pb fractionation factor for pristine Plešovice is 1.025 ± 0.025 (Table 3) making the 207Pb/206Pb fractionation factor “1” within calculated uncertainty. This
PT
test demonstrates potential fractionation factor complications due to matrix effects for
CE
Paleozoic zircons, but what of the effects on the Archean zircons? To test this, we analyzed pristine grains of the 3465 ± 0.6 Ma reference zircon OG1, which is slightly normally
AC
discordant after chemical abrasion, according to TIMS results (Stern et al., 2009; orange dot, inlet in Fig. 9). The pristine OG1 zircons are relatively concordant, when reduced using physically matched pristine Temora2 (Fig. 9, inlet). We obtain a mean 207Pb/206Pb age of 3445 ± 42 Ma for pristine OG1 (5 analyses), or an upper Concordia intercept of 3458 ± 36 Ma (MSWD=0.09). Figure 9 (inlet) shows that using the physically mismatched annealed Temora2 (upper intercept of 3475 ± 92; MSWD=0.12), introduces reverse discordance for the pristine OG1 grains.
24
ACCEPTED MANUSCRIPT In synthesis, we confirm the notion of a radiation-damage induced LA-ICP-MS matrix effect with up to 8% difference in age compared to the known TIMS age for pristine Plešovice using the physically mismatched TA Temora2. This highlights the importance of preanalytical treatment of standards along with unknowns.
SC RI PT
b) The effects on U-Pb systematics of reference zircons and their physical state as applied to pristine Archean unknowns
The effects of switching from pristine to annealed reference Temora2 on the studied unknowns is examined in Figure 10a, showing a Concordia plot of the 21 pristine cores. The
NU
dots represent results when pristine Temora2 is applied as the reference (no uncertainty shown for clarity). The ellipses represent the application of annealed Temora2 as the
MA
reference to these pristine unknowns. Only one of the 21 points is normally discordant. The upper intercept age is the same within uncertainty no matter the reference (3507 ± 48 Ma;
ED
lower intercept: -564 ± 3100 Ma; MSWD=0.10 for pristine reference; 3498 ± 23 Ma; lower intercept: -226 ± 980 Ma; MSWD=0.82 for annealed reference; Fig. 10a). For the pristine
PT
unknowns, the matrix effect of using annealed Temora2 shows some shift towards reverse
CE
discordance, but is overall insignificant. Instead, the physical match between pristine standard and unknown clearly demonstrates that in this case, reverse discordance is not an
AC
artifact of mismatch between standards and unknowns; reverse discordance is a genuine characteristic of the studied cores. As demonstrated in Section 3.4, the studied zircons experienced natural annealing at ~500 Ma. The accepted TIMS age for Temora2 is ~417 Ma, and U concentrations are similar to the core domains in zircon from this study (Black et al., 2004). We suggest that the little effect observed between standard match and mismatch is grounded in a similar matrix effect during ablation between our (naturally annealed) unknowns and Temora2. Previous studies, in which zircons of different age and damage state
25
ACCEPTED MANUSCRIPT without reported natural annealing, have been compared, matrix effects due to greater mismatch were more significant (e.g. Crowley et al., 2014; Marillo-Sialer et al., 2016). This is consistent with our observation that effects of standard mismatch are greater for the more radiation damaged Plešovice than for our unknowns (Fig. 9, Fig. 10a).
SC RI PT
3.5.2 U-Pb systematics of unknowns and the effects of thermal annealing and chemical abrasion
Pre-analytical treatment workflows that are derived based on the study of high-quality
NU
international reference standard zircon should not be readily justified for application to unknowns (Schaltegger et al., 2015). In fact, we will demonstrate below that the U-Pb
MA
systematics of the here studied unknowns are tied to a unique, complex zircon history, which requires detailed examination. In order to assess the nature and intensity of pre-analytical
ED
treatment effects, it is necessary to integrate observations on the natural characteristics of observed U-Pb systematics and discuss processes responsible for isotopic disturbance and U-
CE
PT
Pb discordance.
a) Natural reverse discordance and Pb redistribution
AC
As demonstrated in Section 3.5.1b, reverse discordance of unknown core analyses is genuine, and not an analytical artifact. In contrast to normal discordance that is mostly attributed to Pb loss, reverse discordance implies the presence of excess, unsupported radiogenic Pb (or U loss). Concordia evaluation, on first hand, reveals co-linearity of U-Pb systematics within distinct structurally and chemically identified zircon domains: i) slightly normal to mostly reversely discordant cores (Fig. 10a-c), ii) slightly reversely discordant to medium normal discordant rims (Fig. 10d), and iii) rare reversely discordant/concordant to mostly highly
26
ACCEPTED MANUSCRIPT discordant alt-1 and alt-2 domains (Fig. 10e,f). As exemplified for our TA cores in Figure 10b, reversely to normal discordant clusters of co-linear Discordia-fit generally correlate with U content. Reversely discordant spots are low in U concentration, while normal discordant spots are characterized by elevated U concentrations. Following previous studies, we argue that radiogenic Pb was mobilized and redistributed from high- to low-U domains, as the
SC RI PT
result of either event-diffusion, or alpha-recoil implantation (e.g. Williams et al., 1984; Mattinson et al., 1996; Carson et al., 2002; Valley et al., 2014; Kusiak et al., 2015; Peterman et al., 2016). Due to the similar behavior of 238U and 235U during diffusion, and/or the similar total decay chain energies of both U-Pb systems, isotopic fractionation is thought to be
NU
negligible for both diffusion and (at least instantaneous) alpha-recoil implantation (Mattinson et al., 1996). Therefore, Discordia chords can be used as in normal discordant “one-event”
MA
systems, providing geological meaning for either event-related (i.e. diffusion), or zero (i.e. instantaneous alpha-recoil implantation) lower intercept ages. However, we observe negative
ED
lower intercept apparent ages for the pristine core domains (Fig. 10a). According to Mattinson et al. (1996), this indicates increasing alpha-recoil implantation affecting
PT
increasingly damaged zircon, over time. The existence of this process cannot be excluded,
CE
nor conclusively confirmed, in the present study. Below, we discuss an alternative process, in which event-related diffusion caused a spatially associated co-existence of different excess
AC
Pb components in the cores, sourced from different domains, resulting in steeper Discordia slopes and negative lower intercepts through analytical mixing.
b) Analytical mixing of U-Pb components and comments on their apparent ages For the TA cores, Figure 10b shows a Discordia yielding an upper intercept of 3472 ± 17 Ma, and a lower intercept of 786 ± 450 Ma (N=26; MSWD=2.7). Both the slightly wider data dispersion along Discordia and the younger upper intercept are somewhat conspicuous, in
27
ACCEPTED MANUSCRIPT comparison to the pristine cores (Fig. 10a), and cannot be readily explained by thermal annealing effects. Closer examination of the data reveals that the normally discordant “cluster” defined by four analyses is characterized by highest U concentrations (Fig. 10b), as expected, and also highest (>200 ppm) non-formula Ca concentrations, which are in the range of Ca concentrations in the alteration domains. The U-Pb ratios and ages of the normally
SC RI PT
discordant “cluster” spots are significantly lower/younger than the concordant to reversely discordant analyses of the cores. In our view, the TA core Discordia (Fig. 10b) represents an example of mixing of different U-Pb components. In fact, by discarding the four spots of the normal discordant cluster, a Discordia defined by the remaining 22 spots yields upper and
NU
lower intercepts of 3498 ± 13 Ma and -521 ± 1200 Ma (MSWD=1.6), respectively, which are indistinguishable from the pristine cores (Fig. 10a). The inclusion of the high-U cluster data
MA
lowered the Discordia slope and shifted the upper intercept towards too young, and the lower intercept towards negative (i.e. future) ages (Fig. 10b).
ED
Here we relate the high-U normally discordant cluster in the TA cores to the alteration domains (alt-1, alt-2) that are strongly normally discordant to concordant, and rarely
PT
reversely discordant. For convenience we plot U/Pb data for both alt-1 and alt-2 domains
CE
from each experiment together (Fig. 10e,f). For both domains, high dispersion of data spots along Discordia is observed, without obvious systematic difference between the experiments.
AC
Although poorly fit, both alteration domains give similar intercept ages of 3413 ± 62 Ma and 599 ± 180 Ma (MSDW=33; Fig. 10e) for alt-1, and 3409 ± 110 Ma and 614 ± 210 Ma (MSDW=140; Fig. 10f) for alt-2, respectively. The lower intercepts are roughly in agreement with the ~500 Ma radiation damage age, and the upper intercepts (~3410 Ma) make sense in that they are younger than the core ages and are defined by discordant data with U-Pb systematics similar to the high-U TA core cluster spots. We propose that the alteration domains and the high-U core components experienced partial U-Pb isotopic re-setting,
28
ACCEPTED MANUSCRIPT possibly close to ~3410 Ma. The latter resetting is likely related to a tectono-magmatic event that culminated in the study area around 3460-3420 Ma (Wiemer et al., 2016). Major Pb loss (lower intercepts) occurred contemporaneous to the ~500 Ma event recorded in the radiation damage ages. Revisiting the reversely discordant cores, we argue for a process, in which radiogenic Pb
SC RI PT
(slightly younger than the cores) was redistributed from the partially re-set high-U domains to explain the observed negative lower intercepts in pristine and TA cores. As indicated in Figure 10c, CA treated cores do not include the excess Pb component that results in negative lower intercepts. Instead, the lower intercept (560 ± 550 Ma) coincides with those of the
NU
alteration domains and with the Raman radiation damage age. Slight reversely discordant data persisting in the CA cores, is well matched (MSWD=1.0) along the Discordia that
MA
defines the upper (3503 ± 14 Ma) and lower intercepts.
Structural recovery during natural thermal annealing results in rejection of Pb, progressing
ED
from low- to higher damage portions (Kusiak et al., 2015). If higher damage sites become isolated and inaccessible for leaching, Pb will be trapped. Recent studies using high-
PT
resolution scanning transmission electron microscopy (Kusiak et al., 2015) and atom probe
CE
tomography (Valley et al., 2014; Peterman et al., 2016), have demonstrated the existence of randomly distributed nm-scale Pb clusters, particularly in ancient zircon, accumulated and
AC
trapped during metamorphism that lead to Pb loss. Outside these clusters new radiogenic Pb accumulated after annealing from low Pb zircon. This results in discordance characterized by discrete Pb reservoirs recording different isotopic compositions (Peterman et al., 2016). Our interpretation of co-existing Pb components is similar to these models, with the difference that, in the present case, Pb mobilization was related to a single event, but distinct reservoirs were sourced from distinct domains, one having experienced prior isotopic disturbance. One may argue that the younger excess radiogenic Pb component may have been
29
ACCEPTED MANUSCRIPT accumulated from recent self-irradiation, after natural annealing. This is inconsistent with our data, in which highest relative Pb excess is associated with low-U domains. Recent (leachable) radiogenic Pb accumulation should show higher relative excess in proportion to higher U content. We argue that both short distance Pb redistribution within the cores, and Pb diffusion from adjacent high-U domains into the cores occurred contemporaneously during
SC RI PT
major Pb loss at ~500 Ma. In agreement with expected longer diffusion distances, and the fact that only the radiogenic Pb that was sourced from the high-U alteration domains was leached upon CA, it is argued that the latter Pb accumulated within distinct damage sites. Accumulated damage controls dissolution and leaching behavior in zircon (Mattinson, 1994;
NU
Davis and Krogh, 2000; Romer, 2003). The observed extensive radial and fir-tree fracture network observed in CL (Fig. 1) likely formed as a result of volume swelling due to
MA
radiation-induced point defect accumulation (Trachenko et al., 2003) in the alteration domains, and provided enhanced diffusion pathways during hydrothermal alteration,
ED
connecting alteration and core domains (Fig. 1c). There is no evidence for any hightemperature metamorphic event in the studied host quartz-dioritic gneiss, or anywhere else
PT
within the East Pilbara Terrane, that can be correlated with the 500 Ma natural annealing.
CE
Therefore, significant volume diffusion of Pb particularly within the low-U, low-damage cores is unlikely (Cherniak and Watson, 2000; Valley et al., 2014). Instead, it is proposed
AC
that micro-fracture networks and possibly short distance alpha-recoil damage-related pathways allowed limited fluid-enhanced Pb diffusion redistribution. Longer distance diffusion required to redistribute Pb from the alteration domains into the cores must have occurred along distinct enhanced diffusion pathways. X-ray mapping shows that elevated Ca and Fe concentrations in cores mostly focus along fractures (Fig. 4a), and hence these are likely the sites of excess radiogenic Pb concentrations derived from the adjacent high-U alteration domains. In contrast, radiogenic Pb concentrations in fracture-absent portions
30
ACCEPTED MANUSCRIPT would have rather formed from Pb that accumulated within the cores themselves (see Kusiak et al., 2015). The fractures also provided accessibility for experimental CA leaching, while in-between the fractures, zircon cores were left unaffected by CA. In conclusion, we accept the upper intercept of 3503 ± 14 Ma of the CA cores (Fig. 10c) as the timing of magmatic core crystallization.
SC RI PT
Finally, there are seven discrete rim ablations, which are either pristine (n=4), TA (n=2) or CA (N=1), ranging from reversely to normally discordant, and give an upper intercept age of 3469 ± 68 Ma and a lower intercept of 445 ± 1200 Ma (MSWD=2.9; Fig. 10d).
NU
c) Constraints on the timing of alteration and the validity of the lower intercept apparent ages
MA
All high-U normal discordant spot analyses in the alteration and core domains are characterized by elevated non-formula Ca concentrations, and these domains experienced
ED
alteration fluid interaction after significant radiation damage occurred. We observe that increased normal discordance correlates with actinide contents in excess of ca. 400 ppm U
PT
and 100 ppm Th. Using these minimum concentrations to define high-U domains
CE
characterized by Pb loss, we can approximate minimum α-doses to constrain the time required to produce sufficient damage: in order to reach the first and second percolation point
AC
(3.0*1015 α-mg-1 and 8.0*1015 α-mg-1; e.g. Murakami et al., 1991), the ~3000 Myr before the natural annealing event (~500 Ma) are required, yielding a minimum alpha-dose of 5.99*1015 α-mg-1 for the high-U domains, compared to 0.72*1015 α-mg-1 for 500 Myr of recent damage accumulation (note: full dose calculates at 7.71*1015 α-mg-1 for 400 ppm U and 100 ppm Th, and 3500 Myr). Similarly, post-crystallization damage accumulation until partial re-setting of the alteration domains occurred (~3410 Ma) would not have allowed significant non-formula Ca incorporation. Re-calculation of the amorphous fractions for 3000 Myr-damage
31
ACCEPTED MANUSCRIPT accumulation, shifts the second percolation point, here defined as the drastic increase in nonformula Ca in pristine cores (Fig. 7c), to the expected value of p~0.7, experimentally determined by Geisler et al (2001b). This confirms that most damage accumulated in the ~3000 Myr prior to the ~500 Ma event, and that the most extensive alteration and associated Pb loss occurred during the ~500Ma event, as indicated by the lower intercept ages (Fig. 10c-
SC RI PT
f).
d) Synthesis of the effects of thermal annealing and chemical abrasion on the U-Pb systematics of unknown DW-370-1-14
NU
Previous studies have suggested that thermal annealing removes ablation artifacts in LA-ICPMS (Mattinson, 2005; Allen and Campbell, 2012). In our study little effects are observed
MA
between pristine and TA grains using matching pristine and annealed standard Temora2. Importantly, the studied zircons experienced natural annealing at ~500 Ma. We argue that the
ED
coincidence of the natural annealing age of the studied grains with the crystallization age of Temora2, in combination with similar U concentrations in both our cores and Temora2,
PT
resulted in a physical match between unknowns and standard (i.e. similar α-dose and damage
CE
state). The upper intercept ages of both pristine (3507 ± 48 Ma) and TA (3498 ± 13 Ma, excluding the high-U cluster) cores, are in good agreement. The only difference is observed
AC
in the smaller errors for TA cores compared to the pristine ones. A reduced U-Pb data dispersion has been reported upon experimental TA in previous studies (Solari et al., 2015). In general, two processes may cause the reduced dispersion: i) improved physical integrity at the ablation site (e.g. Allen and Campbell, 2012), and ii) rejection of Pb due to the structural recovery upon TA (e.g. Kusiak et al., 2015). As a compromise of our study in using different grains for each experiment, an arbitrary error reduction due to a sampling bias cannot be excluded. However, as demonstrated in the present study, structural recovery of point defects
32
ACCEPTED MANUSCRIPT that likely accumulated since the natural annealing, upon experimental TA, is relatively significant. This may imply that improved zircon density and associated smaller ablation volume in fact resulted in higher precision (reduced error). Single ablations show a slight reduction in internal 2-σ errors for 206Pb/207Pb ages between pristine (average ± 51 Ma) and TA (± 45 Ma) cores. The upper intercept age error of ± 48 Ma is reduced to ± 23 Ma upon
SC RI PT
introduction of annealed Temora2 for reduction of pristine cores, and further reduced to ± 13 Ma after TA of the unknowns. Both upper and lower intercept ages are statistically indistinguishable, and no difference in the range of discordant data are observed, excluding a process in which significant Pb was removed during experimental TA structural recovery. In
NU
synthesis, the full potential of thermal annealing alone could not be explored in regards to Archean zircon, due to the fact that natural annealing occurred. Thermal annealing of the
MA
more recent (~500 Ma to present) damage shows no effect on the general U-Pb systematics, but precision is somewhat improved, which we ascribe to the change in physical parameters
ED
influencing laser ablation.
The results of our CA treatment, on the other hand, not only show improved accuracy and
PT
precision of the U-Pb zircon age, but allow us to reconstruct a model of the case-specific U-
CE
Pb systematics history of the studied unknowns as discussed in Section 3.5.2b. Importantly, the favored Discordia (Fig. 10c) is most strongly supported by integrating the Raman
AC
spectroscopic and chemical zircon data, providing constraints and validation of the lower intercept (Section 3.5.2c). Previous HF dissolution experiments have demonstrated that U and other trace elements (radiogenic Pb) are preferentially and selectively removed from high-U domains, and/or that high-U domains affected by Pb loss are completely dissolved (Mattinson, 1994; Davis and Krogh, 2000; von Quadt et al., 2014; Crowley et al., 2014). Davis and Krogh (2000) found that threshold α-doses slightly in excess of 1.5*1015 α-mg-1 are required for experimental HF dissolution of U-rich zircon. In our experiment, distinct
33
ACCEPTED MANUSCRIPT dissolution textures are not observed upon experimental CA, and we argue that no significant physical volume was dissolved/removed, not even from the high-U alteration zones. Instead, porous dissolution textures are observed in all experiments (Fig. 1, Fig. 2) within the high-U alteration domains. As stated in Section 3.5.2c the minimum α-dose for recent damage accumulation in the high-U domains is estimated at 0.72*1015 α-mg-1, hence below the
SC RI PT
threshold for dissolution proposed by Davis and Krogh (2000). However, these domains clearly reached a damage state susceptible for partial hydrothermal fluid dissolution – reprecipitation prior to the ~500 Ma natural alteration-annealing event, evident in the incorporation of non-formula elements. Based on the observations from reflected light
NU
imaging (Supplementary Material S-1), the introduction of milky-white areas that we ascribe to residue from the CA experiment, focuses most significantly along high-damage
MA
fracture networks (Supplementary Fig. S-2a). We predict that the fracture network promoted redistribution of radiogenic Pb from the partially reset alteration domains into the
ED
core domains during overall Pb loss at ~500 Ma, leading to reverse discordance (in the cores). Similarly, experimental HF exposure would have allowed preferential removal of the
PT
excess radiogenic Pb from these sites (i.e. fractures), while in between the fractures
CE
substantial natural annealing resulted in chemical resistance to CA under the used conditions, particularly in the low-U cores.
AC
The physically matched pristine cores - pristine Temora2 show larger scatter in 206Pb/238U ages (713 Ma) compared to 207Pb/235U ages (244 Ma). Upon CA (using matched annealed Temora2 standard), the scatter in 207Pb/235U remains the same (reduced by <1 %, i.e., negligible), while the scatter in 206Pb/238U ages is reduced by >30 %. Furthermore, it is particularly the high 206Pb/238U end-member data spots that “disappear”, and overall the latter ages are observed to approach the corresponding 207Pb/235U values/ages, hence improving concordance. Using weighted mean ages of only the most concordant (± 1.5 % discordance)
34
ACCEPTED MANUSCRIPT analyses, yielding 3510 (± 9) and 3538 (± 21) Ma for 207Pb/235U and 206Pb/238U ages, respectively, for pristine (N=8), and 3511 (± 35) and 3522 (± 61) Ma, respectively, for CA cores (N=10), shows that the 206Pb/238U weighted ages approach the 207Pb/235U ages, which are basically identical for both experiments. The 207Pb/235U ages are more consistent with the accepted upper intercept of 3503 ± 14 Ma for CA cores. In our view, this demonstrates a
SC RI PT
preferential “removal” of particularly high excess 206Pb, where excess 206Pb was sourced from the adjacent partially reset alteration domains, redistributed during the ~500 Ma event.
NU
4. Conclusions
We present a microstructural and U-Pb systematics study comparing pristine, thermally
MA
annealed (TA) and chemically abraded (CA) early Archean zircon from a quartz-dioritic gneiss, in order to improve LA-ICP-MS U-Pb zircon dating workflows. Overall, we find that
ED
most reliable (i.e. accurate and precise) U-Pb ages are ultimately achieved through CA treatment, which includes annealing of radiation damage and chemical abrasion of high-U
PT
discordant portions. However, our study demonstrates following critical new insights into the
CE
use of pre-analytical treatment of zircon, which leads us to conclude that standardization of pre-analytical treatment is not recommended:
AC
Firstly, we show that the effects of pre-analytical treatment are case-specific and controlled by the initial state of the zircon or zircon domain. In general, this means that treatment conditions must be adjusted according to the initial damage state of unknowns, highlighting the importance of characterizing pristine grains in terms of microstructure and U-Pb systematics, first. In the present case, the core domains record part of the primary U-Pb signature, but they also comprise a secondary U-Pb component likely as the result of Pb redistribution from adjacent high-U alteration domains that experienced partial isotopic
35
ACCEPTED MANUSCRIPT resetting. Only upon CA, the latter U-Pb component was removed, indicating that radiogenic Pb moved into specific damage sites, accessible for CA leaching. Without the prior characterization of pristine grains, identification of the secondary excess Pb component would not have been possible. Here, most fundamental information comes from Raman spectroscopy, revealing that the zircons experienced natural annealing during a Pb loss event,
SC RI PT
which constrains the true slope of the Discordia and its lower intercept. In addition, our data supports the notion that radiation damage is irreversible, and more data from natural samples are needed to establish reference frameworks that would correlate initial damage state with most plausible treatment conditions.
NU
Secondly, we demonstrate that physical mismatch between treated and pristine standardunknown-pairs results in up to 8% age difference compared to known TIMS values (here for
MA
pristine Plešovice using TA Temora2 as the primary reference standard), attesting for the existence of LA-ICP-MS matrix effects. However, despite their antiquity, matrix effects are
ED
negligible for our unknown cores, which experienced natural annealing at a time close to the crystallization age of the used reference standard Temora-2 characterized by similar U
PT
content. Therefore, matrix effects are reduced through natural or experimental TA-induced
CE
structural recovery. Standard material should be treated in the same way as unknowns, and standards should be selected carefully to comply with both U content and the age of
AC
crystallization, or the timing of the most recent natural annealing experienced by the unknowns. We propose to establish a Raman spectroscopic database for commonly used reference zircons in order to readily select physically matching reference standards for LAICP-MS data reduction of well-characterized unknowns. Furthermore, we demonstrate that reverse discordance in our zircons is not an analytical artifact, but a genuine U-Pb systematics characteristic, highlighting the impact of intra-grain Pb redistribution on the µm-scale of common LA-ICP-MS analytical spot sizes. We show
36
ACCEPTED MANUSCRIPT that some excess radiogenic Pb within our cores could not be removed through CA treatment, and records primary U-Pb systematics. In combination with supporting validation of lower Discordia intercepts from Raman radiation damage ages, the reverse discordant analyses, which are not removed during CA, represent more reliable primary U-Pb systematics than the normal discordant analyses from high-U ablations.
SC RI PT
Finally, we observe an increasing coupled incorporation of non-formula elements (e.g. Ca, Fe) and LREE (e.g. La) with increasing U content and structural damage. This suggests that high LREE signatures in zircon are at least partially the result of hydrothermal alteration, and not of primary magmatic origin.
NU
In synthesis, we recommend a relatively time-efficient LA-ICP-MS U-Pb zircon workflow, in which a subset of pristine grains are characterized along with TA and CA grains by means of
MA
CL imaging and Raman spectroscopy. During LA-ICP-MS analyses, co-measurement of nonformula elements is recommended. Zircon standard material should be treated under the same
PT
CE
5. Acknowledgements
ED
conditions as unknowns.
The Geological Survey of Western Australia, in particular A. H. Hickman, is acknowledged
AC
for logistical support of the fieldwork, during which the sample DW-370-1-14 was collected. D. Wiemer acknowledges fruitful input during the U-Th-Pb LA-ICP-MS network workshop at the Goldschmidt Conference, Prague, 2015. H. Cathey (CARF) is acknowledged for assistance with electron microprobe analyses, L. Rintoul (QUT) for assistance with Raman spectroscopy, C. East for assistance with VP-SEM-CL, and K. H. Moromizato for assistance with LA-ICP-MS analyses. M. DeBruyn (CARF) performed thermal annealing, and D. McAuley (CARF) assisted mount preparation and polishing. G. Fraser (Geoscience
37
ACCEPTED MANUSCRIPT Australia) kindly supplied OG1 standard material. The authors thank two anonymous reviewers for improving the final paper.
6. References
SC RI PT
Allen, C.M., and Campbell, I.H., 2012: Identification of a matrix-induced systematic error in LA-ICP-MS 206Pb/238U dating of zircon. Chemical Geology, 332-333, 157-165 Balan, E., Mauri, F., Pickard, C.J., Farnan, I., and Calas, G., 2003: The aperiodic states of zircon: an ab initio molecular dynamics study. American Mineralogist, 88, 1769-1777
NU
Bell, E.A., Harrison, T.M., McCulloch, M.T., and Young, E.Y., 2011: Early Archaean crustal evolution of the Jack Hills Zircon source terrane inferred from Lu-Hf,
207
Pb/206Pb, and
MA
δ18O systematics of Jack Hills zircon. Geochimica et Cosmochimica Acta, 75, 4816-4829 Belousova, E.A., Griffin, W.L., O’Reilly, S.Y., and Fisher, N.I., 2002: Igneous zircon: trace
ED
element composition as am indicator of source rock type. Contrib Mineral Petrol, 143, 602-622
PT
Belousova, E.A., Kostitsyn, Y.A., Griffin, W.L., Begg, G.C., O’Reilly, S.Y., and Pearson,
CE
N.J., 2010: The growth of the continental crust: Constraints from zircon Hf-isotope data. Lithos, 119, 457-466
AC
Black, L.P., Kamo, S.L., Allen, C.M., Davis, D.W., Aleinikoff, J.N., Valley, J.W., Mundil, R., Campbell, I.H., Korsch, R.J., Williams, I.S., and Foudoulis, C., 2004: Improved 206Pb/238U microprobe geochronology by the monitoring of a trace-element-related matrix effect; SHRIMP, ID-TIMS, ELA-ICP-MS and oxygen isotope documentation for a series of zircon standards. Chemical Geology, 205, 115-140
38
ACCEPTED MANUSCRIPT Bizzarro, M., Connelly, J.N., Thrane, K., and Borg, L.E., 2012: Excess hafnium-176 in meteorites and the early Earth zircon record. Geochem. Geophys. Geosyst., 13, Q03002, doi:10.1029/2011GC004003 Campbell, I.H., and Allen, C.M., 2008: Formation of supercontinents linked to increases in atmospheric oxygen. Nature Geoscience, 1, 554-558
SC RI PT
Carson, C.J., Ague, J.J., Grove, M., Coath, C.D., and Harrison, T.M., 2002: U-Pb behavior of zircon during upper-amphibolite facies infiltration in the Nappier Complex, east Antarctica. Earth Planet. Sci. Lett., 199, 287-310
Cawood, P.A., Hawkesworth, C.J., and Dhuime, B., 2013: The continental record and the
NU
generation of continental crust. GSA Bulletin, 125, 1/2, 14-32
Chen, F., Siebel, W., and Satir, M., 2002: Zircon U-Pb and Pb-isotope fractionation during
MA
stepwise HF acid leaching and geochronological implications. Chemical Geology, 191, 155-164
ED
Cherniak, D.J., and Watson, E.B., 2000: Pb diffusion in zircon. Chemical Geology, 172, 5-24 Compston, W., and Pidgeon, R.T., 1986: Jack Hills, evidence of more very old detrital
PT
zircons in Western Australia. Nature, 321, 766-769
CE
Compston, W., Williams, I.S., Kirschvink, J.L., Zhang, Z., and Guogan, M.A., 1992: Zircon U-Pb ages for the Early Cambrian time-scale. Journal of the Geological Society, London,
AC
149, 171-184
Condie, K.C., and Aster, R.C., 2009: Zircon Age Episodicity and Growth of Continental Crust. Earth and Space Science News, 90, 41, 364-365 Condie, K.C., 2014: Growth of continental crust: a balance between preservation and recycling. Mineralogical Magazine, 78, 3, 623-637
39
ACCEPTED MANUSCRIPT Crowley, Q.G., Heron, K., Riggs, N., Kamber, B., Chew, D., McConnell, B., and Benn, K., 2014: Chemical abrasion Applied to LA-ICP-MS U-Pb Zircon Geochronology. Minerals, 4, 503-518 Davis, D.W., and Krogh, T.E., 2000: Preferential dissolution of
234
U and radiogenic Pb from
α-recoil-damaged lattice sites in zircon: implications for thermal histories and Pb isotopic
SC RI PT
fractionation in the near surface environment. Chemical Geology, 172, 41-58 Davis, D.W., Williams, I.S., and Krogh, T.E., 2003: Historical Development of Zircon Geochronology. Reviews in Mineralogy and Geochemistry, 53, 1, 145-181 Dawson, P, Hargreave, M.M., and Wilkinson, G.R., 1971: The vibrational spectrum of zircon
NU
(ZrSiO4). J. Phys. C: Solid St. Phys., 4, 240-256
Devanathan, R., Corrales, L.R., Weber, W.J., Chartier, A., and Meis, C., 2006: Molecular
MA
dynamics simulation of energetic uranium recoil damage in zircon. Mol Simul, 32, 10691077
ED
Froude, D.O., Ireland, T.R., Kinny, P.D., Williams, I.S., Compston, W., Williams, I.R., and Myers, J.S., 1983: Ion microprobe identification of 4,100-4,200 Myr-old terrestrial
PT
zircons. Nature, 304, 616-618
CE
Gaft, M., 1992: Application of thermal treatment of zircon for the interpretation of luminescence centers. Journal of Thermal Analysis, 38, 10, 2281-2290
AC
Geisler, T., 2002: Isothermal annealing of partially metamict zircon: evidence for a threestage recovery process. Phys. Chem. Minerals, 29, 420-429 Geisler, T., Pidgeon, R.T., Van Bronswijk, W., and Pleysier, R., 2001a: Kinetics of thermal recovery and recrystallization of partially metamict zircon: a Raman spectroscopic study. Eur. J. Mineral., 13, 1163-1176
40
ACCEPTED MANUSCRIPT Geisler, T., Ulonska, M., Schleicher, H., Pidgeon, R.T., and van Bronswijk, W., 2001b: Leaching and differential recrystallization of metamict zircon under experimental hydrothermal conditions. Contrib Mineral Petrol, 141, 53-65 Geisler, T., Pidgeon, R.T., van Bronswijk, W., and Kurtz, R., 2002: Transport of uranium,
Chemical Geology, 191, 141-154
SC RI PT
thorium, and lead in metamict zircon under low-temperature hydrothermal conditions.
Geisler, T., and Pidgeon, R.T., 2002: Raman scattering from metamict zircon: comments on “Metamictization of natural zircon: accumulation versus thermal annealing of radioactivity-induced damage” by Nasdala et al. 2001 (Contrib. Mineral. Petrol., 141, 125-
NU
144). Contrib. Mineral. Petrol., 143, 750-755
Geisler, T., Zhang, M., and Salje, E.K.H., 2003a: Recrystallization of almost fully amorphous
Nuclear Materials, 320, 280-291
MA
zircon under hydrothermal conditions: An infrared spectroscopic study. Journal of
ED
Geisler, T., Trachenko, K., Ríos, S., Dove, M.T., and Salje, E.K.H., 2003b: Impact of selfirradiation damage on the aqueous durability of zircon (ZrSiO4): implications for its
PT
suitability as a nuclear waste form. J. Phys.: Condens. Matter, 15, L597-L605
Experimental
CE
Geisler, T., Pidgeon, R.T., Kurtz, R., van Bronswijk, W., and Schleicher, H., 2003c: hydrothermal
alteration
of
partially
metamict
zircon.
American
AC
Mineralogist, 88, 1496-1513 Götze, J., and Kempe, U., 2009: Physical principles of cathodoluminescence and its applications in geosciences. In Gucsik, A. (ed.), Cathodoluminescence and its application in the planetary sciences. Springer, Berlin Heidelberg New York, 1-22 Hay, D.C., and Dempster, T.J., 2009: Zircon Behaviour during Low-temperature Metamorphism. Journal of Petrology, 50, 4, 571-589
41
ACCEPTED MANUSCRIPT Holland, H.D., and Gottfried, D., 1955: The effect of nuclear radiation on the structure of zircon. Acta Crystallographica, 8, 291-300 Jochum, K.P., Weis, U., Stoll, B., Kuzmin, D., Yang, Q., Raczek, I., Jacob, D.E., Stracke, A., Birbaum, K., Frick, D.A., Günther, D., and Enzweiler, J., 2011: Determination of Reference Values for NIST SRM 610-617 Glasses Following ISO Guidelines.
SC RI PT
Geostandards and Geoanalytical Research, 35 (4), 397-429
Kempe, U., Gruner, T., Nasdala, L., and Wolf, D., 2000: Relevance of cathodoluminescence for interpretation of U-Pb zircon ages (with an example of application to a study of zircon from the Saxonian Granulite Complex, Germany). In M. Pagel, V. Barbin, and D.
NU
Ohnenstetter (eds.), Cathodoluminescence in Geosciences, Springer, Berlin, 415-455 Košler, J., Wiedenbeck, M., Wirth, R., Hovorka, J., Sylvester, P., and Mikova, J., 2005:
MA
Chemical and phase composition of particles produced by laser ablation of silicate glass and zircon: Implications for elemental fractionation during ICP-MS analysis. J. Anal. At.
ED
Spectrom, 20, 5, 402-409
Krogh, T.E., 1982: Improved accuracy of U-Pb zircon ages by the creation of more
PT
concordant systems using an air abrasion technique. Geochimica et Cosmochimica Acta,
CE
46, 4, 637-649
Krogh, T.E., 1994: Identification of concordant zircons using etch techniques. Abstr 8th Intl
AC
Conf Geochron Cosmochron Iso Geol, US Geol Surv Circular, 1107, 180 Krogh, T.E., and Davis, G.L., 1974: Alteration in zircons with discordant U-Pb ages. Carnegie Inst Washington Yrbk, 73, 560-567 Krogh, T.E., and Davis, G.L., 1975: Alteration in zircons and differential dissolution of altered and metamict zircon. Carnegie Inst Washington Yrbk, 74, 619-623
42
ACCEPTED MANUSCRIPT Kusiak, M.A., Dunkley, D.J., Wirth, R., Whitehouse, M.J., Wilde, S.A., and Marquardt, K., 2015: Metallic lead nanospheres discovered in ancient zircons. PNAS, 112 (16), 49584963 Lee, J.K.W., and Tromp, J., 1995: Self-induced fracture generation in zircon. Journal of Geophysical Research, 100, B9, 17,753-17,770
SC RI PT
Lenz, C., and Nasdala, L., 2015: A photoluminescence study of REE3+ emissions in radiation-damaged zircon. American Mineralogist, 100, 1123-1133
Marillo-Sialer, E., Woodhead, J., Hergt, J., Greig, A., Guillong, M., Gleadow, A., Evans, N., and Paton, C., 2014: The zircon ‘matrix effect’: evidence for an ablation rate control on
NU
the accuracy of U-Pb age determinations by LA-ICP-MS. J. Anal. At. Spectrom., 29, 981989
MA
Marillo-Sialer, E., Woodhead, J., Hanchar, J.M., Reddy, S.M., Greig, A., Hergt, J., and Kohn, B., 2016: An investigation of the laser-induced zircon ‘matrix effect’. Chemical Geology,
ED
438, 11-24
Marsellos, A.E., and Garver, J. I., 2010: Radiation damage and uranium concentration in
PT
zircon as assessed by Raman spectroscopy and neutron irradiation. American
CE
Mineralogist, 95, 1192-1201
Mattinson, J.M., 1994: A study of complex discordance in zircons using step-wise dissolution
AC
techniques. Contrib. Mineral. Petrol., 116, 117-129 Mattinson, J.M., 2005: Zircon U-Pb chemical abrasion (“CA-TIMS”) method: Combined annealing and multi-step partial dissolution analysis for improved precision and accuracy of zircon ages. Chemical Geology, 220, 1-2, 47-66 Mattinson, J.M., Graubard, C.M., Parkinson, D.L., and McClelland, W.C., 1996: U-Pb reverse discordance in zircon: the role of fine-scale oscillatory zoning and sub-micron
43
ACCEPTED MANUSCRIPT transport of Pb. In: A.R. Basu, and S.R. Hart (eds.), Earth Processes: Reading the Isotopic Code, American Geophysical Union, Geophysical Monograph, 95, 355-370 Meldrum, A., Boatner, L.A., Weber, W.J., and Ewing, R.C., 1998: Radiation damage in zircon and monazite. Geochimica et Cosmochimica Acta, 62, 14, 2509-2520
evaluation. J. Metamorphic Geol., 15, 127-140
SC RI PT
Mezger, L., and Krogstad, E.J., 1997: Interpretation of discordant U-Pb zircon ages: An
Murakami, T., Chakoumakos, B.C., Ewing, R.C., Lumpkin, G.R., and Weber, W.J., 1991: Alpha-decay event damage in zircon. American Mineralogist, 76, 1510-1532 Nasdala, L., Irmer, G., and Wolf, D., 1995: The degree of metamictization in zircon: a
NU
Raman spectroscopic study. Eur. J. Mineral., 7, 471-478
Nasdala, L., Wenzel, M., Vavra, G., Irmer, G., Wenzel, T., and Kober, B., 2001:
MA
Metamictization of natural zircon: accumulation versus thermal annealing of radioactivityinduced damage. Contrib. Mineral. Petrol., 141, 125-144
ED
Nasdala, L., Lengauer, C.L., Hanchar, J.M., Kronz, A., Wirth, R., Blanc, P., Kennedy, A.K., and Seydoux-Guillaume, A.-M., 2002: Annealing radiation damage and the recovery of
PT
cathodoluminescence. Chemical Geology, 191, 1-3, 121-140
CE
Nasdala, L., Zhang, M., Kempe, U., Panczer, G., Gaft, M., Andrut, M., and Plötze, M., 2003:
427-467
AC
Spectroscopic methods applied to zircon. Reviews in Mineralogy and Geochemistry, 53, 1,
Nemchin, A.A., and Pidgeon, R.T., 1999: U-Pb ages on titanate and apatite from the Darling Range granite: implications for Late Archaean history of the southwestern Yilgarn Craton. Prec. Res., 96, 1-2, 125-139 Palenik, C.S., Nasdala, L., and Ewing, R.C., 2003: Radiation damage in zircon. American Mineralogist, 88, 770-781
44
ACCEPTED MANUSCRIPT Paton, C., Woodhead, J.D., Hellstrom, J.C., Hergt, J.M., Greig, A., and Maas, R., 2010: Improved laser ablation U-Pb zircon geochronology through robust downhole fractionation correction. Geochemistry, Geophysics, Geosystems, 11, Q0AA06, 1-36, doi: 10.1029/2009GC002618 Peterman, E.M., Reddy, S.M., Saxey, D.W., Snoeyenbos, D.R., Rickard, W.D.A.,
SC RI PT
Fougerouse, D., and Kylander-Clark, A.R.C., 2016: Nanogeochronology of discordant zircon measured by atom probe microscopy of Pb-enriched dislocation loops. Sci. Adv., 2, e1601318, doi: 10.1126/sciadv.1601318
Pidgeon, R.T., 2014: Zircon radiation damage ages. Chemical Geology, 367, 13-22
NU
Pidgeon, R.T., Nasdala, L., and Todt, W., 1998: Determination of radiation damage ages on parts of zircon grains by Raman microprobe: implications for annealing history and U-Pb
MA
stability. Mineralogical Magazine, 62A, 1174-1175
Ríos, S., Salje, E.K.H., Zhang, M., and Ewing, R.C., 2000: Amorphization in zircon:
ED
evidence for direct impact damage. J. Phys.: Condens. Matter, 12, 2401-2412 Romer, R.L., 2003: Alpha-recoil in U-Pb geochronology: effective sample size matters.
PT
Contrib Mineral Petrol, 145 (4), 481-491
CE
Salje, E.K.H., Chrosch, J., and Ewing, R.C., 1999: Is “metamictization” of zircon a phase transition? American Mineralogist, 84, 1107-1116
AC
Schaltegger, U., Schmitt, A.K., and Horstwood, M.S.A., 2015: U-Th-Pb zircon geochronology by ID-TIMS, SIMS, and laser ablation ICP-MS: Recipes, interpretations, and opportunities. Chemical Geology, 402, 89-110 Sláma, J., Košler, J., Condon, D.J., Crowley, J.L., Gerdes, A., Hanchar, J.M., Horstwood, M.S.A., Morris, G.A., Nasdala, L., Norberg, N., Schaltegger, U., Schoene, B., Tubrett, M.N., and Whitehouse, M.J., 2008: Plešovice zircon – A new natural reference material for U-Pb and Hf isotopic microanalysis. Chemical Geology, 249, 1-35
45
ACCEPTED MANUSCRIPT Solari, L.A., Ortega-Obregón, C., and Bernal, J.P., 2015: U-Pb zircon geochronology by LAICPMS combined with thermal annealing: Achievements in precision and accuracy on dating standard and unknown samples. Chemical Geology, 414, 109-123 Steiger, R.H., and Jäger, E., 1977: Subcommission on geochronology: convention on the use of decay constants in geo- and cosmochronology. Earth Planet. Sci. Lett., 36, 359-362
SC RI PT
Stern, R.A., Bodorkos, S., Kamo, S.L., Hickman, A.H., and Corfu, F., 2009: Measurement of SIMS Instrumental Mass Fractionation of Pb Isotopes During Zircon Dating. Geostandards and Geoanalytical Research, 33, 2, 145-168
Trachenko, K.O., Dove, M.T., and Salje, E.K.H., 2001: Atomistic modeling of radiation
NU
damage in zircon. J. Phys.: Condens Matter, 13, 9, 1947-1959
Trachenko, K., Dove, M.T., and Salje, E.K.H., 2003: Large swelling and percolation in
MA
irradiated zircon. J. Phys.: Condens. Matter, 15, L1-L7 Trachenko, K., Dove, M.T., Geisler, T., Todorov, I., and Smith, B., 2004: Radiation damage
ED
effects and percolation theory. J. Phys.: Condens. Matter, 16, S2623-S2627 Tsuchiya, Y., Nishido, H., Kayama, M., and Noumi, Y., 2013: Cathodoluminescence of
PT
radiation-induced zircon. AGU Fall Meeting abstract 2013AGUFMMR31A2298T
CE
Utsunomiya, S., Palenik, C.S., Valley, J.W., Cavosie, A.J., Wilde, S.A., and Ewing, R.C., 2004: Nanoscale occurrence of Pb in an Archaean zircon. Geochimica et Cosmochimica
AC
Acta, 68, 22, 4679-4686
Valley, J.W., Peck, W.H., King, E.M., and Wilde, S.A., 2002: A cool early Earth. Geology, 30, 4, 351-354 Valley, J.W., Cavosie, A.J., Ushikubo, T., Reinhard, D.A., Lawrence, D.F., Larson, D.J., Clifton, P.H., Kelly, T.F., Wilde, S.A., Moser, D.E., and Spicuzza, M.J., 2014: Hadean age for a post-magma-ocean zircon confirmed by atom-probe tomography. Nature Geoscience, 7, 219-223
46
ACCEPTED MANUSCRIPT Weber, W.J. 1993: Alpha-Decay-Induced Amorphization in Complex Silicate Structures. Journal of the American Ceramic Society, 76, 7, 1729-1738 Weber, W.J., Ewing, R.C., and Wang, L.-M., 1994: The radiation-induced crystalline-toamorphous transition in zircon. J. Mater. Res., 9, 3, 688-698 Wiemer, D., Schrank, C.E., Murphy, D.T., and Hickman, A.H., 2016: Lithostratigraphy and
Western Australia. Prec. Res., 282, 121-138
SC RI PT
structure of the early Archaean Doolena Gap greenstone belt, East Pilbara Terrane,
Wilde, S.A., Valley, J.W., Peck, W.H., and Graham, C.M., 2001: Evidence from detrital zircons for the existence of continental crust and oceans on the Earth 4.4 Gyr ago. Nature,
NU
409, 175-178
Williams, I.S., Compston, W., Black, L.P., Ireland, T.R., and Foster, J.J., 1984: Unsupported
MA
radiogenic Pb in zircon: a cause of anomalously high Pb-Pb, U-Pb and Th-Pb ages. Contrib Mineral Petrol, 88, 322-327
ED
Zhang, M., Salje, E.K.H., Capitani, G.C., Leroux, H., Clark, A.M., Schlüter, J., and Ewing, R.C., 2000: Annealing of α-decay damage in zircon: a Raman spectroscopic study. J.
PT
Phys.: Condens. Matter, 12, 3131-3148
CE
Zhao, K.-D., Jiang, S.-Y., Ling, H.-F., and Palmer, M.R., 2014: Reliability of LA-ICP-MS UPb dating of zircons with high U concentrations: A case study from the U-bearing
AC
Douzhashan Granite in South China. Chemical Geology, 389, 110-121
Figure captions
47
ACCEPTED MANUSCRIPT Fig. 1: Selected VP-SEM CL images showing characteristic features of the distinct zircon domains; a) part of a pristine grain displaying very faint oscillatory zoning towards the outer cores; CL intensity of the alt-1 domain is hardly distinguished from the core, while domain alt-2 shows lower CL intensity; The rim domain shows the brightest CL and features radial fractures; b) annealed grain showing good contrast between the domains; fracture networks
SC RI PT
of bright CL extend from the alt-1 domain into adjacent domains and fractures aligned perpendicular to the concentric zoning; c) detail of annealed grain showing dark “CL holes” (arrow a) in bright CL alt-1 domain and dark CL fractures terminating in alt-1 with gradual CL increase at terminations (arrow b); d) detail of annealed grain displaying the convolute
NU
heterogeneous CL intensity distribution within alt-1 domain; e) detail of annealed grain featuring spongy textured zones (arrows d and c) within the dark CL alt-2 domain, and the
MA
inward curvature boundary of the rim domain.
ED
Fig. 2: Selected SEM-CL, integral EMPA spectral CL and BSE images (rows, from left to right), comparing a pristine (pr01, uppermost row), a thermally annealed (TA94, center row),
CE
PT
and a chemically abraded (CA-II 17, lowermost row) grain; refer to text for description.
Fig. 3: Representative CL spectra comparing each of the four domains from a pristine grain
AC
(pr01) with an annealed grain (TA94); Note the dominant yellow-green broad band (A) in the pristine cores compared to the dominant blue broad band (B) in the annealed domains, and the appearance of superimposed narrow REE3+ emission lines in the alt-1, alt-2 and to a minor extent in the core domains, upon annealing; refer to text for further explanation.
Fig. 4: a) Representative EMPA x-ray mineral maps of selected elements of an annealed zircon, highlighting the incorporation of non-formula Ca and Fe in the alt-1 and alt-2
48
ACCEPTED MANUSCRIPT domains; b) semi-quantitative BSE and CL intensity variations and selected elemental concentrations corresponding to the mineral profile indicated in the integral spectral CL image above.
Fig. 5: a) Representative Raman spectra comparing each domain from a pristine with an
SC RI PT
annealed zircon, highlighting the increased intensity and sharpening of the dominant peaks at ~970 and ~1000 cm-1; the v3 SiO4 peaks at ~1000 cm-1 are shown in detail for a pristine (b) and an annealed (c) grain; Note the increasing frequency/wavenumber upon annealing; a1,
NU
alt-1; a2, alt-2; refer to text for description.
Fig. 6: Raman frequency versus line-widths plot showing the radiation damage trend and the
MA
annealing trend of Raman response to the structural disorder (after Geisler et al., 2001b); the line-width is given as the full-width at half-maximum (FWHM); note that some of the
ED
pristine data domains plot above the radiation damage trend, indicating that natural annealing
PT
occurred; refer to text for description.
CE
Fig. 7: a) Th versus U plot, showing cores at common magmatic Th/U values, while alt-1, alt-2 and rim domains plot at relatively low Th/U; b) amorphous fraction p as a function of
AC
alpha-dose D; the critical value of D~8*1015 a/mg is indicated, demonstrating advanced metamictization of the alt-1 and alt-2 domains; c) non-formula Ca as a function of the amorphous fraction p; note the drastic increase in Ca above an amorphous fraction of p~0.8; the latter is interpreted as the second percolation point after Geisler et al. (2001b); d) La versus Ca concentration plot, suggesting coupled incorporation of non-formula Ca and LREE during alteration.
49
ACCEPTED MANUSCRIPT Fig. 8: Raman FWHM (v3 SiO4 peak) versus alpha-dose plot showing the position of pristine zircon domains; using the total 3.5 Ga alpha-doses, data plot below the calibration curve for Sri Lankan zircon (Palenik et al., 2003), indicating the occurrence of natural annealing; recalculation of Raman equivalent alpha doses (squares) yield consistent radiation damage ages
Fig. 9: Concordia
206
Pb/238U versus
SC RI PT
of ~0.5 Ga for all domains.
207
Pb/235U (black curve) plot showing the effects of
annealing on the U-Pb systems of Temora2 and Plešovice standards compared to their known TIMS values using annealed Temora2 as the primary reference for LA-ICP-MS data
NU
reduction; inlet shows the effects of mismatching annealed Temora2 for reduction of pristine OG1 standard data; note the introduction of reverse discordance; refer to text for further
MA
description.
AC
CE
PT
ED
Fig. 10: U-Pb Concordia diagrams (a-f); refer to text for description.
50
ACCEPTED MANUSCRIPT
AC
Fig. 1; 2 columns
CE
PT
ED
MA
NU
SC RI PT
Figures
51
CE
PT
ED
MA
NU
SC RI PT
ACCEPTED MANUSCRIPT
AC
Fig. 2; 1.5 to 2 columns
52
NU
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Fig. 3; 2 columns
53
ED
MA
NU
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
PT
Fig. 4; 2 columns
54
PT
ED
MA
NU
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
Fig. 5; 2 columns
55
NU
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
Fig. 6; 1.5 to 2 columns
56
ED
MA
NU
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
PT
Fig. 7; 2 columns
57
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
NU
Fig. 8; 1 column
58
MA
NU
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Fig. 9; 1.5 columns
59
AC
CE
PT
ED
MA
NU
SC RI PT
ACCEPTED MANUSCRIPT
Fig. 10; 2 columns
60
ACCEPTED MANUSCRIPT Tables
Table 1: Asymmetric Si-O stretching mode v3 at ca. 1000 cm-1 grain spot position width 1008.15 1007.64 1007.33 1006.94 1006.53 1005.77 1005.62 1005.25 1005.13 1004.85
10.14 11.20 7.73 8.37 16.38 16.35 19.79 17.39 17.00 10.96
pr2rim pr3core pr2core pr4rim pr3rim pr2alt11 pr1core pr2alt2 pr3alt2 pr4core pr2alt1
1004.04 1001.95 1001.55 1001.35 1000.64 998.90 998.62 998.34 998.10 997.67 996.43
9.54 12.75 12.65 10.77 15.92 21.05 17.54 21.96 25.59 14.73 19.54
CE
PT
ED
MA
NU
SC RI PT
ca1-2alt1 TA50rim 94alt1 94rim ca1-2alt2 TA50core 94alt2 ca1-2core TA50alt1 94core
AC
Annealed CAI02 TA50 TA94 TA94 CAI02 TA50 TA94 CAI02 TA50 TA94 Pristine pr04 pr01 pr04 pr72 pr01 pr04 pr07 pr04 pr01 pr72 pr04
61
ACCEPTED MANUSCRIPT
-1
cm
12.75 12.65 17.54 14.73 9.54 10.77 21.05 19.54 25.59
a-mg *10 0.81 0.80 1.23 0.97 0.57 0.66 1.63 1.45 2.31
-1
a-mg *10
15
0.65 0.70 2.11 2.43 0.82 5.10 7.30 7.30 51.17
RadAGE
U
Th
Ma
ppm
ppm
527 526 512 506 518 488 496 494 470
33 36 108 121 44 257 392 392 2720
11 13 36 65 2 113 9 9 246
CE
PT
ED
MA
NU
01-2 04-3 07-1 72-1 04-1 72-2 04-4 04-4 01-4
15
AC
core core core core rim rim alt-1 alt-1 alt-2
-1
SC RI PT
Table 2: Calculation of radiation ages for pristine zircon Domain Spot FWHM Deq Dcalc
62
ACCEPTED MANUSCRIPT
Table 3: LA-ICP-MS U-Pb data for zircon standards compared to accepted TIMS values Temora annealed ± pristine ± annealed
±
Plešovice pristine 15 337.1 8.88 0.0537 0.0527 1.0196 0.3937 0.3766 1.0455 0.0532 0.0008 0.0153 365 28.1 356 18.4 292 -45.1
0.0010
0.0059
4.2 7.8
AC
CE
PT
ED
MA
NU
SC RI PT
N 13 12 14 TIMS age [Ma]* 416.9 416.9 337.1 Calc. alpha dose [a/mg 1015] 4.41 4.37 8.68 Accepted 206Pb/238U 0.0668 0.0668 0.0537 Measured 206Pb/238U 0.0603 0.0006 0.0618 0.0005 0.0499 0.0003 Fractionation factor 206Pb/238U 1.1079 1.0817 1.0750 Accepted 207Pb/235U 0.5077 0.5077 0.3937 Measured 207Pb/235U 0.4560 0.0090 0.4690 0.0120 0.3607 0.0030 Fractionation factor 207Pb/235U 1.1134 1.0825 1.0916 Accepted 207Pb/206Pb 0.0551 0.0551 0.0532 Measured 207Pb/206Pb 0.0548 0.0012 0.0551 0.0014 0.0523 0.0008 0.0518 Fractionation factor 207Pb/206Pb 1.0062 0.0220 1.0007 0.0254 1.0172 0.0150 1.0270 Calc. 206Pb/238U age [Ma]** 417 427 347 Delta age ref std 9.9 10.1 Calc. 207Pb/235U age [Ma]** 417 5.9 427 8.4 343 2.5 Delta age ref std 9.7 5.7 Calc. 207Pb/206Pb age [Ma]** 418 8.0 431 14.5 314 8.4 Delta age standard 14.1 -23.1 +these data are not drift corrected and no down hole correction is applied, for demonstration purposes *Black et al. (2004); **calculated ages using the fractonation factors derived from annealed Temora
±
63
SC RI PT
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA
NU
Graphical abstract
64