a U-Pb thermochronometer or geochronometer

.


Introduction
Apatite, a calcium phosphate mineral with the general formula Ca 5 (PO 4 ) 3 (F,Cl,OH), is prevalent in both magmatic and sedimentary rocks and can be generated and modified during metamorphic processes.This crustal ubiquity, along with its ability to host U, means that apatite has become useful in a range of geochronology and thermochronology applications (Chew and Spikings, 2015;Corfu and Stone, 1998;Kirkland et al., 2017;Mark et al., 2016;Schoene and Bowring, 2007;Weisberg et al., 2018).Thermochronology can be considered to date when a mineral passes through a specific closure temperature for the isotope of interest (Dodson, 1973) and is a consequence of the integrated passage of a mineral through a daughter isotopes partial retention zone.In contrast, geochronology refers to the date at which a specific growth or alteration process may have affected a mineral.Unlike zircon, apatite can be readily found in silica-undersaturated mafic lithologies, however its U content is frequently much lower than that of zircon and apatite usually contains moderate to significant amounts of non-radiogenic (common) Pb.The high ratio of common to radiogenic Pb presents challenges for apatite U-Pb geochronology and calculated dates rarely have the precision of those obtained fromzircon (Chew et al., 2011;Glorie et al., 2017;Krogstad and Walker, 1994;LaFlamme et al., 2017).Nonetheless, apatite is still a very useful U-Pb chronometer as its typical closure temperature of around 375-600 °C (Cochrane et al., 2014;Schoene and Bowring, 2007) potentially can provide additional time-temperature constraints on retrograde or lowtemperature prograde processes.Apatite can accommodate high concentrations of REE (Henrichs et al., 2018), Sr,Mn, Na, U, and Th, and other trace elements, making it potentially powerful in charting the trace element chemistry of the rock system (Fleet and Pan, 1997;Gawęda et al., 2014;Mao et al., 2016;Piccoli and Candela, 2002).So, as apatite is highly susceptible to a range of fluid-induced (metasomatic) changes that occur at variable pressures and temperatures (Harlov, 2015;Krenn and Finger, 2004;Nutman, 2007;Spear and Pyle, 2002), its chemistry linked to age may be powerful intracking fluidrock interaction from the deep crust to near surface (Glorie et al., 2017;Henrichs et al., 2018;Kirkland et al., 2017;Schoene and Bowring, 2007).Metasomatic processes can be considered to affect the U-Pb system in apatite in two broad ways.Metasomatism may causenew growth of apatite, in which case the U-Pb geochronometer will record from the point that Pb stops diffusing out of the grain.Alternatively, metasomatism may cause partial to complete loss of radiogenic-Pb (potentially along with other chemical changes), in which case the U-Pb geochronometer will record either a mixed age or, if Pb loss is complete, the time of metasomatism.

Regional background
The Maniitsoq region, in southern West Greenland, lies immediately north of Nuuk and is a component of the North Atlantic craton, comprising the Mesoarchean Akia terrane (Fig. 1).The region is exceptionally well exposed, and in contrast to most other Archean cratons, which are dominated by low-gradegranite-greenstone associations, it is mainly tonalitic-trondhjemitic-granodioritic (TTG) orthogneiss and tholeiitic amphibolite, metasedimentary rocks, and ultramafic rocks, metamorphosed at amphibolite to granulite facies.The wider Nuuk regionincluding the Akia terranehas been considered a classic example of Archean lateral tectonics and terrane accretion, whereby individual Archean tectono-stratigraphic terranes of different ages were assembled during the Neoarchean (Friend et al., 1996;Friend and Nutman, 2005).It has been proposed that the orthogneiss Fig. 1.Schematic geological map of southern West Greenland (from Szilas et al., 2015) showing the location of samples (yellow filled triangles).Inset shows Landsat image.
precursors of the North Atlantic craton were produced by slab melting of subducted oceanic crust (e.g.Garde, 1997;Winther and Newton, 1991, and references therein).Specifically, Garde et al. (2000) proposed that the AkiaTerrane developed by formation of a c. 3100 Ma magmatic arc over a Paleo-to Mesoarchean crustal core, followed by voluminous emplacement of Mesoarchean TTG gneisses and synchronous low-pressure granulite facies metamorphism and late-tectonic felsic magmatism.However, other models involving extra-terrestrial components have also been proposed for the Akia Terrane (Garde et al., 2014;Keulen et al., 2014;Scherstén and Garde, 2013) whereby a giant bolide impact at c. 3000 Ma was responsible for a prolonged Mesoarchean thermal reworking, and resulted in a range of melting and deformation features across the region.The Akia Terrane has also been regarded as the crustal link between Archean mid-crust, high-gradeorthogneissamphibolite associations exposed in West Greenland, and the lower-gradegranite-greenstone associations of the Superior Province of eastern Canada (Bridgewater and Schiøtte, 1991;Wilton, 1994).
The Akia Terrane comprises three distinct crustal components: a c.3200 Ma dioritic association, a c. 3.0 Ga TTG component (Garde et al., 2000), and a less than c. 2.9 Ga metasedimentary component (Kirkland et al., 2018).The majority of the orthogneisses are tonalite that was the basement for some, or tectonically intercalated with other, supracrustal rocks and associated mafic intrusive complexes.The supracrustal rocks consist largely of heterogeneous, compositionally layered rusty weathering metavolcanic amphibolite, metapelite to metapsammite, and metagabbroic and ultramafic rocks that represent disrupted layered intrusive complexes and overlying siliciclastic components.The supracrustal rocks were originally interpreted as remnants of oceanic crust, into which the precursors of the orthogneisses were intruded, presumably in a convergent plate-tectonic setting (Garde, 1990(Garde, , 1997;;Garde et al., 2000).However, recently many of the supracrustal rocks have been established as components of a package that is now known to be younger than the c. 3.0 Ga TTG association (Kirkland et al., 2018).In this work we use apatite from TTG and leucosome cutting the TTG to help address some of the complex history of overprinting metamorphism and fluid-rock interaction that has affected this region.

U-Pb geochronology
All U-Pb geochronology and geochemistry results are presented in Electronic Appendix Table 1.

Geochronology analytical method
3.1.1.Apatite 3.1.1.1.Spot analysis.Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) data collection for apatite age and trace elements, as single spot ablations, was performed during two sessions at the GeoHistory Facility in the John de Laeter Centre, Curtin University, Perth, Australia.Apatite was separated from samples using standard crushing and heavy liquid techniques.Individual apatite grains (mounted and polished in 1″ epoxy rounds) were ablated using a Resonetics RESOlution M-50A-LR, incorporating a Compex 193 nm excimer laser.The sample cell was flushed with ultrahigh purity He (350 mL min-1) and N 2 (3.8 mL min-1), both of which were passed through an inline Hg trap.Isotopic intensities were measured using an Agilent 7700 s quadrupole ICP-MS, with high purity Ar as the plasma gas (flow rate 1 L min-1).
Apatite U-Pb age determination was performed first, with spot ablations in constant energy mode using a 50 μm spot and a 4 Hz laser repetition rate.Laser energy (measured at the sample surface) was 2 J cm −2 .After two cleaning pulses, each analysis comprised a 30 s background acquisition and 35 s of ablation.A total of 10 isotopes were determined ( 43 Ca, 44 Ca, 201 Hg, 202 Hg, 204 Pb, 206 Pb, 207 Pb, 208 Pb, 232 Th and 238 U).The dwell time on the Pb isotopes was 0.1 s with Ca isotopes measured for 0.03 s and Hg, U and Th measured for 0.0125 s.Laser induced U-Pb fractionation was accounted for during data reduction (Chew et al., 2014;Paton et al., 2010;Paton et al., 2011).McClure Mountain syenite apatite (Thomson et al., 2012) was used as the primary standard as it has moderate but reasonably consistent U and Th contents (~23 ppm and 71 ppm; (Chew and Donelick, 2012)), its thermal history is well known (it is the rock from which the 40 Ar/ 39 Ar hornblende standard MMhb is derived) and the crystallization age (weighted mean 207 Pb/ 235 U date of 523.51 ± 2.09 Ma) and initial Pb isotopic composition ( 206 Pb/ 204 Pb = 17.54 ± 0.24; 207 Pb/ 204 Pb = 15.47 ± 0.04) are known from high precision TIMS analyses (Schoene and Bowring, 2006).MAD Apatite (486.58 ± 0.85 Ma and 474.25 ± 0.41, (Thomson et al., 2012) and Duluth Complex apatite (U-Pb zircon age of 1099.1 ± 0.2 Ma, (Schmitz et al., 2003)) were used as secondary references, and analysed as unknowns.Ages obtained, by free regressions, on the secondary references Duluth Complex apatite (1112 ± 35; MSWD = 0.43; n = 24) and MAD apatite (489 ± 20; MSWD =0.45; n = 24) were within analytical uncertainty of recommended values.
Trace element ablations were placed over the previous U-Pb spot.Acleaning pulse was followed by 30 s of baseline acquisition and 45 s ablations, at a 4 Hz repetition rate using a 60 μm beam and laser energy of 2 J cm-2.Twenty-four isotopes were measured ( 29 Si, 31 P, 43   232 Th and 238 U), with most monitored for 0.02 s.Dwell time on Si and P was 0.01 s.International glass standard NIST 610 was used as the primary reference material for calculation of concentrations and to correct for instrument drift. 43Ca was the internal standard element, assuming 39.36% Ca in apatite.Standard blocks were typically run after 20 unknown analyses.During the timeresolved analysis, contamination resulting from inclusions and compositional zoning were monitored and only the relevant part of the signal was integrated.Secondary trace element standard NIST 612, treated as an unknown, reproduced recommended values to within 2% for all elements except P (5%).The time-resolved mass spectra were reduced using Iolite (Paton et al., 2011, and references therein).All analyses were sited on the centre of apatite grains where a single ablation pit could be located without touching the enclosing epoxy resin.
As a means to assist in grouping analyses into growth or alteration populations, we calculate grain level 207 Pb corrected U/Pb dates and F207% values.We apply an iterative process to determine an appropriate common 207 Pb/ 206 Pb ratio to use in 207 Pb correction using the 207 Pb/ 206 Pb age only as an initial estimate for a model common Pb composition (Stacey and Kramers, 1975).We then use the initial estimated 207 Pb corrected age to revise the common Pb model value to calculate a new 207 Pb corrected 206 Pb/ 238 U age.We repeat this process until ages converge (Thomson et al., 2012). 207Pb corrected ages on an individual spot level are treated as apparent ages only and we prefer to use a more conservative regression approach through the entire population to calculate dates onto which geological interpretation is placed.F207% values are calculated as the difference between the 206 Pb/ 238 U ratio for the 207 Pb corrected age and the measured 206 Pb/ 238 U ratio.Effectively this is the distance along a mixing line from radiogenic to common Pb where higher % represents a greater common component (Kirkland et al., 2017).
3.1.1.2.Depth profile.Apatite age and trace element depth profiles were also measured simultaneously from a single ablation volume using laser ablation split stream ICP-MS in a third session.The two selected apatite grains (mounted and polished in 1″ epoxy rounds) were ablated from the polished c-axis perpendicular apatite surfaces through to the underlying epoxy using the same Resonetics RESOlution M-50A-LR laser system but coupled to a Nu Plasma IImulti-collector inductively coupled plasma mass spectrometer (MC-ICPMS) for age determination and also an Agilent 7700 quadrupole inductively coupled plasma mass spectrometer (Q-ICP-MS) for simultaneous trace element measurement.
Laser system parameters are given in Table1.The sample cellwas flushed with ultrahigh purity He (350 mL min −1 ) and N 2 (3.8 mL min −1 ), both of which were passed through an inline goldbased Hg trap.High purity Ar was used as the carrier gas (flow rate 1 L min −1 ) and the ablated volume was split equally between the two mass spectrometers.
For age determination, the following elements were monitored (detector in brackets: H-Faraday Cup; IC-Ion Counter) over an 0.8 s integration time: 238 U (H10), 232 Th (H8), 208 Pb (IC0), 207 Pb (IC1), 206 Pb (IC2), 204 Pb (IC3), and 202 Hg (IC4).Each unknown apatite was bracketed by ablation of three standard blocks including the McClure Mountain syenite standard apatite (Thomson et al., 2012) and NIST 612 glass both of which were used as primary reference materials.Durango (31.44 ± 0.18 Ma, (McDowell et al., 2005) was included as a secondary standard.The time-resolved mass spectra were subdivided into eight equal time segments of 20 s each (approximate ablation depth of 20 μm for each slice).The McClure Mountain syenite and NIST612 standards were treated similarly.Each segment in the unknowns was processed against the same time segment in the standard grains (e.g., first 20 s segment in the apatite unknown was processed against first 20 s segment for the standards) so that downhole fractionation corrections were always started at the same ablation depth/time in the standard and sample material.In addition, fractionation of 207 Pb/ 206 Pb and 238 U/ 206 Pb was also monitored and corrected with reference to NIST 612, the advantage here over natural apatite being that the glass has a fixed common to radiogenic Pb ratio.Ages were derived for each time segment using Iolite 3.5 (Paton et al., 2011) with the VizualAge_UcomPbine DRS (Chew et al., 2014) and in-house Microsoft Excel macros with 207 Pb corrections calculated as described above except common Pb was defined from the upper intercept of the spot analysis in the same sample (discussed in detail below).Excluding one aberrant step (step 3; 34 ± 1 Ma) 207 corrected ages for the Durango secondary reference reproduced the accepted age within analytical uncertainty (31.6 ± 0.7 Ma; MSWD 2.4).Full instrument setup is given in Table1.
For trace element determination a total of 24 isotopes were measured ( 29 Si, 31 P, 43  Lu, 232 Th and 238 U), with most monitored for 0.02 s.Dwell time on Si and P was 0.01 s.International glass standard NIST 612 was used as the primary reference material for calculation of concentrations and to correct for instrument drift with 43 Ca utilized as the internal standard element, assuming 39.36% Ca in apatite.NIST 614 was also analysed as a secondary standard.Standard blocks bracketed the two apatite unknowns and were treated similarly to samples.As with the age determination, 8 equally spaced time segments were selected over the full length of the ablation period, matching the segments selected for age determination.Each segment in the unknowns was processed against the corresponding time segment in the standard glass.The time-resolved mass spectra were reduced using the Trace Elements data reduction scheme in Iolite 3.5 (Paton et al., 2011 and references therein).Isotopic concentrations for the NIST 614 secondary standard ( 29 Si internal reference isotope) were reproduced to within 7% of the accepted value for most isotopes with considerably better precision on some (e.g., Sr -1.4%, U-3.4%).

Zircon
3.1.2.1.Spot analysis.LA-ICPMSU-Pb analysis of sample 215 zircon grains was also performed at the GeoHistory Facility in the John de Laeter Centre.Individual zircon grains were ablated using the same Resonetics RESOlution M-50A-LR sampling system as was utilized for apatite ablation.Samples were ablated for 30 s at a 7 Hz repetition rate using a 50 μm beam spot and laser energy of 1.5 J/cm 2 .A20 s period of background analysis was acquired prior to ablation.The sample cell was flushed by ultrahigh purity He (350 mL min −1 ) and N 2 (3.8 mL min −1 ).Isotopic intensities were measured using an Agilent 7700 s quadrupole ICPMS, with high-purity Ar as the plasma gas (flow rate 0.98 L min −1 ).The dwell time for all masses was 0.01 s.The primary reference material used for zircon U-Pb dating in this study was the OG-1 zircon standard (3465.4± 0.6 Ma; (Stern et al., 2009)) with the following secondary zircon standards: Plešovice (337.13 ± 0.37 Ma; Sláma et al., 2008), R33 (419.26 ± 0.39 Ma;Black et al., 2004), and GJ-1(601.7 ± 1.4 Ma;Jackson et al., 2004). 206Pb/ 238 U ages calculated for all zircon age standards, treated as unknowns, were within 3% of the accepted values.For example, the mean 206 Pb/ 238 U age determined for Plešovice was 327 ± 10 Ma (2σ, n = 12, MSWD = 1.0), within uncertainty of the recommended value.The time-resolved mass spectra were reduced using the U/Pb Geochronology3 data reduction scheme in Iolite (Paton et al., 2011).No common lead corrections were deemed necessary due to low 204 Pb counts.
SIMS U-Pb analysis of sample 216 zircon grains were performed on Shrimp B in the John de Laeter Centre following analytical protocols matching those discussed in Wingate and Kirkland (2015).Eight analyses of the BR266 standard (Stern, 2001) were obtained during the session, which indicated an external spot-to-spot (reproducibility) uncertainty of 0.56% (1σ), and a 238 U/ 206 Pb* calibration uncertainty of 0.31% (1σ).Calibration uncertainties are included in the errors of 238 U/ 206 Pb* ratios and dates listed in Appendix Table1.Two analyses of the OG-1 standard yielded a 207 Pb/ 206 Pb weighted mean age of 3464 ± 3 Ma (MSWD = 0.96) identical to the accepted value.Common-Pb corrections were applied to all SIMS analyses using contemporaneous isotopic compositions determined according to the Pb evolution model of Stacey and Kramers (1975).1).Apatite in this sample displays homogeneous textures under BSE, but distinct zonation in CL (Fig. 2).
Based on mineral liberation analysis with a Telscan TIMA, apatite grains are up to 208 μm (c-axis parallel), with a mode of b5 μm, although 31% are N40 μm.The modal abundances of minerals, as observed in thin section, are 48% plagioclase (An 81 ), 24% quartz, 20% alkali feldspar (Or 89 ), 7% mica (biotite and muscovite equally present), and 1% accessory minerals such as epidote, apatite, and titanite.Epidote is developed both at grain boundaries, but mostly as discrete grains.Quartz is granoblastic with undulose extinction.Feldspars are generally fresh, although with minor clay minerals within fractures and epidote along grain boundaries when in contact with mica (Fig. 3).U-Pb analyses of apatite define a broad discordia array that scatters significantly with F207% ranging from 5 up to 56%, interpreted as principally reflecting mixing between radiogenic and common Pb.Not only do the data scatter between radiogenic and nonradiogeniccomponents but also, towards the radiogenic end of the mixing array (i.e.lower 207 Pb/ 206 Pb) display increased U/Pb variability at a given Pb/Pb ratio (Fig. 4).This variability in U/Pb ratios is mirrored in the apatite Mn content, which defines a continuum of concentrations from 176 to 385 ppm Mn.A high Mn group (N335 ppm) yields a lower intercept on concordia of c. 2813 Ma (MSWD = 0.44) and an upper intercept 207 Pb/ 206 Pb c ratio of 1.115 ± 0.12.A low Mn group (with an arbitrary b250 ppm Mn cut off) illustrates a younger age with lower intercept of c. 2149 Ma (MSWD = 1.5) and an upper intercept 207 Pb/ 206 Pb c ratio of 1.056 ± 0.09 (Fig. 4).Both upper intercepts are similar in terms of common Pb composition but the radiogenic Pb component is distinctly different with apparent ages that vary by at least N400 Ma.Those analyses with low Mn and younger apparent ages indicate lower La N /Sm N ratios, indicative of more steeply negative LREE profiles.Apatite from sample 205 is, in general, characterized by positive Eu anomalies.
3.2.1.2.216 granitic leucosome in gneiss (zircon and apatite).Sample 216 is from a folded granitic leucosome that intrudes dioritic to tonalitic gneisses.The sample was recovered from a coastal outcrop 2.6 km south-east of Ikerasak and 44 km north-east of Atammik.In thin section, this sample contains about 38 vol% quartz, 30% alkali feldspar (Or 73 ), 22% plagioclase (An 24 ), 9% biotite, and accessory epidote and opaque oxide minerals.A distinct foliation is defined by alignment of biotite.Epidote is rare and only found as isolated grains at the interface between plagioclase and biotite.Feldspars are generally fresh with minimal retrogression.
Zircon crystals in this leucosome are subhedral to euhedral and, in CL images, idiomorphic zoning is ubiquitous and most grains have low CL response (high U)overgrowths up to 30 μm thick.Four analyses (Group M)of homogenous rims yield a weighted mean 207 Pb/ 206 Pb age of 2994 ± 5 Ma (MSWD = 1.3), interpreted as the age of leucosome crystallization.The uranium content of Group M rims is on average c.900 ppm, some c. 700 ppm greater than the cores.Ten analyses (Group X)of idiomorphically zoned zircon yield a weighted mean 207 Pb/ 206 Pb age of 3059 ± 6 Ma (MSWD = 2.4), interpreted as the magmatic crystallization age of an inherited component incorporated into this leucosome.Two analyses (Group P)yield 207 Pb/ 206 Pb ages of2981 ± 4 Ma and 2718 ± 18 Ma (1σ) and are interpreted to reflect ancient radiogenic-Pb loss (Fig. 5; Appendix Table1).
Apatite in this sample is euhedral and displays homogeneous textures in BSE images and can be up to 268 μm in size, with a mode of c. 50 μm.Apatite is also homogeneous in CL images (Fig. 2).U-Pb analyses of apatite broadly define two discordia arrays that scatter between radiogenic and non-radiogenic components.An apparently older array scatters between 3 and 13% F207.This apparently older component  Analyses in this group range from 6 to 37% F207 (Fig. 4).19 analyses lie between these two groupings with both intermediate apparent age and Sr contents.The upper intercept for the older array is imprecisely defined and may not be identical in terms of common Pb composition to the younger array.Furthermore, the radiogenic Pb component is distinctly different between the two arrays with apparent ages that vary by over 500 Ma.Those analyses with low Sr and younger apparent ages correlate with lower total REE content.Furthermore, older apparent ages are associated with higher U content on average, consistent with the lower analytical uncertainties on U/Pb isotopic ratios for these analyses.
3.2.1.3.215 late granite sheet (zircon).Sample 215 is from a structurally late granite sheet that intrudes leucosome-bearing dioritic to tonalitic gneisses.This sample was taken from a few meters away from sample 216.The sheet cuts both the foliation and leucosomes in the host gneiss.The zircon crystals are euhedral with length to width ratios up to 6:1.In CL images the grains show moderate CL response with oscillatory zoning particularly evident in partially resorbed cores.Most grains have a low CL outermost zone.Four analyses are N10% discordant (Group D).These analyses are not considered further.Twenty-two analyses (Group I)yield a weighted mean 207 Pb/ 206 Pb age of 2558 ± 8 Ma (MSWD = 1.8), interpreted as the age of granite crystallization.Nine core or discrete grain analyses (Group X)yield 207 Pb/ 206 Pb ages of 3050-2628 Ma, interpreted to reflect inherited zircon (Fig. 5).Five analyses (Group P)yield 207 Pb/ 206 Pb ages of 2489-2319 Ma and indicate extreme U and Th content.Many of these analyses are on fractures.These analyses are interpreted to have experienced ancient radiogenic-Pb loss.

Apatite depth profiles
In order to better understand the apatite U-Pb geochronology two grains were chosen for depth profiling.The depth profile analyses penetrated through the grain from the polished surface (half grain thickness) into the epoxy and were standardised over the depth interval in which a corresponding reference was ablated. 207Pb corrected ages have been calculated for a sequence of integration steps for both samples.These apparent ages are based on a regression anchored through the upper intercept of the quadrupole single spot ablation regressions, which indicate a common 207 Pb/ 206 Pb ratio of 1.089 for sample 205 and 1.165 for sample 216, which is interpreted as the best estimate of the common Pb composition.
The apparent age profile of sample 205 rises to ages as old as c.2935 Ma before falling towards the rim with a youngest age of c.2277 Ma (Fig. 6).The La N /Sm N ratio for all integrations is N0.5.Concentrations of Mn are N460 ppm and rise towards the rim, which corresponds to the chemical signature in the older of the single spot grain analyses.U content is about 6 ppm until a rim that has N20 ppm.An integration (step 9; Appendix Table1) was selected to just characterize this rim and yields an apparent age of 2277 ± 37 Ma.The core region with uniform U content indicates relatively consistent apparent ages that yield a weighted mean age of 2705 Ma with slight excess dispersion (MSWD = 3).The consistency of these core ages and compositional Fig. 3. Left; Series of photomicrographs of sample 205, tonalitic gneiss.The modal abundances as observed in thin section are 48% plagioclase (An 81 ), 24% quartz, 20% alkali feldspar (Or 89 ), 7% mica (biotite and muscovite equally present), and 1% accessory minerals such as epidote, apatite, and titanite.Right; sample 216, granitic leucosome in dioritic to tonalitic gneisses.Inthin section the modal abundances are about 38% quartz, 30% alkali feldspar, 22% plagioclase, 9% mica (all biotite), and accessory epidote and opaque oxides.Mineral abbreviations: bi = biotite, q = quartz, pl = plagioclase, ksp = K-feldspar.
distinction from the rims is consistent with core cooling through the radiogenic Pb retention temperature for a magmatic crystal across atimescale which did not allow any obvious Pb diffusion profile to develop on the grain scale.The apparent age profile of sample 216 rises to an age of 2031 Ma before falling to ages around 1700 Ma with increasing ablation depth (e.g.towards the rim).Concentrations of Uand Th gradually decrease with ablation depth (towards rim) with the content of both spiking in the last two integrations, reflecting anenriched rim.Sr broadly falls with integration step.La N /Sm N values fall until the last two integrations (Fig. 6).REE patterns are similar in all ablations but the total REE concentration is significantly higher in the last two integrations.The chemical profiles and apparent ages do not directly correspond (e.g.older zones are not necessarily geochemically distinct), consistent with different diffusion rates between trace elements and U and Pb.Nonetheless, based on Th, U,REE concentrations, two distinct zones are apparent: an inner zone corresponding approximately to the core of the grain and a rim enriched in Th, U and REE.The core region shows apparent ages which peak at c. 2000 Ma but decrease away from this high, consistent with radiogenic Pb loss.The rim yields apparent ages of c. 1735 Ma, similar to the apparent ages in the immediately adjacent core region.The broad trend in the core age profile could be considered broadly consistent with some form of disturbance event (thermal, recrystallization, or dissolution driven) causing radiogenic-Pb mobility synchronous with neoblastic apatite rim growth.The higher U content of the rims corresponds with greater radiogenic-Pb in the rims and a position on the concordia diagram that is closer to the radiogenic end of a mixing array.The oldest core apparent age provides a minimum age for original apatite crystallization at N2031 Ma.

Pseudosection methods
Phase equilibria modelling is used to bracket the P-T conditions ofmetamorphism and fluid influx potentially responsible for apatite dissolution-regrowth in sample 205.Modelling was conducted in the NCKFMASHTO system-applicable for mafic to intermediate compositions (Palin et al., 2016)-using the internally consistent thermodynamic database (ds63) of (Holland and Powell, 2011) and the THERMOCALC (v.3.45)software package (Powell et al., 1998).Activity-composition models are from Green et al. (2016).Phases modelled as pure end-members are quartz, rutile, titanite and aqueous fluid (H 2 O).All mineral abbreviations are from (Holland and Powell, 2011) except for titanite (ttn).Bulk rock compositions were determined at ALS Laboratories (Ireland) using their ME-ICP06 analytical package.Samples were digested by lithium metaborate fusion and major element concentrations were determined by ICP-AES.The amount of ferric iron was assumed to be 10% and water was set to be in excess.The measured and modelled compositions are summarized in Table2.

Pseudosection results
A P-T pseudosection for sample 205 and water-saturated conditions is shown in Fig. 7a.At 0.2 to 0.9 GPa, the mineral assemblage at the wet solidus (~650 °C) is predicted to be dominated by plagioclase, quartz and K-feldspar with ~5 mol% (approximately equivalent to vol%) biotite and b 1 mol% rutile and muscovite.These predicted mineral proportions are generally consistent with the observed modes in the sample.At P N 0.9 GPa, the calculated solidus assemblage includes minor amounts of epidote and titanite.Below the solidus, a relatively low-variance field extends from 450 °C at 0.1 GPa up to the solidus at 0.9 GPa that represents the down-temperature breakdown of plagioclase to albite and epidote.Normalized molar proportions of the phase assemblages at 0.4 GPa (Fig. 7b) show the general reaction sequences and mineral proportions as a function of temperature.Plagioclase breakdown to albite is expected to occur from ~550-470 °C, and plagioclase breakdown to epidote and albite from ~520-470 °C.K-feldspar is predicted to break down to muscovite over a similar temperature range.The temperature ranges of these reaction fields increase with increasing pressure.
The presence of grain boundary epidote associated with plagioclase restricts any fluid infiltration in sample 205 to temperatures lower than the set of low-variance fields in Fig. 7a, i.e. to T N 450 °C at 0.1 GPa and T N 600 °C at 0.8 GPa.Regional Neoarchean peak metamorphic conditions are 820-850 °C at 0.8-1.0GPa and represent an apparent thermal gradient of ~1000 °C/GPa (Kirkland et al., 2018).Extrapolation along this gradient from peak conditions results in pressures of ~0.65 GPa at the wet solidus (~650 °C) and at 0.55 GPa at the epidote-in phase boundary (~550 °C).Considering the value of the apparent gradient is a maximum due to thermal relaxation after the peak of metamorphism, the minimum temperature of fluid infiltration into sample 205 to preserve the observed plagioclase-and epidote-bearing assemblage is ~550 °C.However, this does not apply if a separate thermal pulse was associated with fluid infiltration responsible for apatite recrystallization.In this scenario, the temperatures would still need to be N450 °C to preserve the observed mineral assemblage in the presence of a free hydrous fluid phase.

Summary of U-Pb results
Sample 205, from a tonalitic gneiss, reveals distinct rims under CL (Fig. 2).U-Pb analyses define a broad discordia array interpreted to reflect mixing between radiogenic and common components (Fig. 4).The U-Pb data also reveals significant scatter at the radiogenic end of the array.This scatter in U-Pb space correlates with Mn chemistry, with a high Mn group yielding an old age of c. 2813 Ma whereas a lower Mn group yielding a younger age of c. 2149 Ma (Fig. 4).Importantly, both upper intercepts are similar in terms of common Pb composition.Depth profiling of this apatite reveals older apparent ages in the core before younger ages towards a U enriched rim with a best age estimate of c. 2277 Ma.The chemistry, morphology, depth profiling and spot analysis results imply these crystals were originally igneous and whose cores now record a cooling age, with discrete high U rims reflecting later new apatite growth (Fig. 8).The dominant c. 2800 Ma age of high Mn apatite in sample 205 might be considered as reflecting some sort of average time of cooling to the Pb diffusion temperature.However, these apatite grains do not solely record thermally activated volume diffusion of Pb as a dissolution-regrowth process is indicated by a younger rim overgrowth (Fig. 6).The small amount of epidote observed in the sample implies a b 600 °C fluid infiltration event associated with the apatite rims.
In contrast to sample 205, sample 216 from a granitic leucosome is homogeneous in CL images.However, U-Pb analyses define two discordia arrays with potentially distinct common Pb components (Fig. 4).An older array with age of c. 2490 Ma has elevated Sr concentrations whereas a younger c. 1800 Ma array has lower Sr content.Theyounger array is also associated with lower total REE content.The younger apatite array may have sourced older crustal Pb in comparison to the older apatite cores.A depth profile through an apatite grain from sample 216 rises to an apparent age of 2031 Ma before falling to late Paleoproterozoic ages towards the rim (Fig. 6).In the depth profiles, trace element compositions and apparent ages do not directly correspond (e.g.older zones are not necessarily geochemically distinct), although there is a broad similarity, consistent with somewhat different diffusion rates between trace elements and U and Pb (Fig. 6).Specifically, the diffusion closure temperatures of Sr and REEs in apatite aresignificantly higher than Pb (Cherniak, 2000).The core region isinterpreted to reflect radiogenic-Pb loss at the same time as new rim growth.The higher U content of the c. 1800 Ma rims corresponds with a greater portion of radiogenic-Pb.The common Pb content in the younger apatite array implies interaction with a distinct, more radiogenic, crustal source (Fig. 4).This observation cautions that using intra-grain mixing lines to define common Pb corrections must be cognisant of the potential for even individual grains to have different common Pb reservoirs dependant on the growth environment of that apatite domain.

Apatite Pb diffusion or dissolution and new growth
A range of different growth and alteration processes can be identified from the linked geochronology and chemistry signatures in these apatite crystals.It is pertinent to note that the spot analyses need not directly correspond to the age and chemistry of the depth profiles.For example there is generally lower La/Sm ratios in the younger end-member of the individual grain apatite analyses from sample 205, however in the depth profile there is little distinction between core and rim.The Mn concentration trend with age is also apparently opposite in thedepth profile versus individual spot analyses.The depth profile for sample 216 also shows some distinctions from the individual analyses, specifically the Sr depleted rim with young age has enriched total REE relative to the core, while a depletion in total REE concentration is recorded in the low Sr, young individual analyses (Fig. 6).Th and Uare also high in the rim of sample 216 (Fig. 6) while the individual analyses from sample 216 that are young tend to be lower in Th and U(Appendix Table1).Such observations imply that the thin high U rims in both samples were below the individual grain spot's spatial resolution.Nonetheless, there is relatively good age correspondence which may imply alteration processes recorded in cores were synchronous with new growth processes.

Thermally activated Pb diffusion
Two general conditions of thermally activated diffusion can be considered in response to decreasing temperature and a minerals passage through its partial retention zone: A)If the rate of homogenization of composition by diffusion within the mineral fails to keep up with the rate of change of the interface composition then the mineral would not have a single closure temperature, but a closure temperature profile (e.g. a slow cooling situation).B)Conversely, if the mineral passes rapidly through its partial retention zone then a single closure age may be recovered.The c. 2800 Ma age of apatite in sample 205 is interpreted as the time of cooling to the Pb diffusion temperature in some grains given the crystallization age of c. 3000 Ma for tonalitic gneisses in the Akia Terrane (Garde, 2007;Garde et al., 2000) and 820-850 °C, 8-10 kbar, regional metamorphism (Kirkland et al., 2018) which must have occoured prior to or at c. 2800 Ma in this sample.This interpretation of an original magmatic genesis for sample 205 apatite is consistent with its geochemistry; the Sr contents, REE patterns and Eu anomalies of magmatic apatite have been considered to directly relate to the geochemical composition of its parental magma, reflecting magma mixing and compositional heterogeneity in the magma source region.Magmatic apatite REE patterns commonly exhibit negative Eu anomalies, a function of the redox state of the magma and feldspar fractionation.A positive Eu anomaly is not related to fractionation processes but must be a source feature (Chu et al., 2009).Apatite in sample 205 has a distinctive positive Eu feature (Fig. 4) which is also seen in the whole rock composition of this tonalitic gneiss (Fig. 9), consistent with a magmatic source comprising feldspar cumulates.The value of (Sr apatite /Sr whole- rock ) for apatite in sample 205 is b1 which is also consistent with its felsic source (Chu et al., 2009).The depth profile for sample 205 apatite is relatively flat implying a single cooling age for this grain yet the individual spots from multiple grains reveal significant age variation.We interpret these features to imply a short transit time through the partial retention zone (i.e.Condition B), yet different Pb closure temperatures between grains, likely a consequence of grain size (Blackburn et al., 2011).
In contrast to sample 205, apatite core depth profiles from sample 216 do appear to yield a Pb diffusion profile (see also Cochrane et al., 2014).Such profile implies Pb diffusion which could not be accommodated by transport over a finite timescale into the matrix (i.e.Condition A).However, such profile is modified by later compositionally distinct apatite growth.Other examples of U-Pbcore-rim apatite profiles that imply thermally activated Pb diffusion significantly after apatite crystallization have been recognised from the Archean Akilia Island in Greenland (Chew et al., 2014).Whitehouse et al. (2009) and Sano et al. (1999) also published on this Akilia apatite and concluded it likely recorded both resetting of older apatite and new growth in the Proterozoic.

Metamorphic apatite growth
The c. 2500 Ma age of apatite in the leucosome bearing tonalitic gneiss (sample 216) is interpreted to reflect the time of apatite growth during metamorphism, consistent with a comparable zircon crystallization age of 2558 ± 8 Ma for a late cross-cutting granite sheet (sample 215) that intrudes these gneisses.Neither, BSE or CL imaging reveals any significant zonation nor overgrowths in sample 216 apatite, which is generally sub-to euhedral.The apparent lack of clear overgrowths in the imaging can be explained by the magnitude of compositional difference, especially in U, which is much less than in sample 205 apatite which does display rims (the U content in apatite 216 rims is twice thatin cores, whereas apatite 205 rims U is at least five times core concentration).

Apatite dissolution and regrowth
The younger apparent apatite ages of c. 2150 Ma in sample 205, associated both with lower Mn content and lower La N /Sm N ratios, imply the growth of a new metasomatic apatite phase later in the history of this rock.Metamorphic apatites are commonly characterized by depletion of Th and LREE, particularly in metapelitic samples where apatite coexists with monazite (Spear and Pyle, 2002).The lower La N /Sm N ratios in the c. 2150 Ma apatite component, from the discrete grain analysis, therefore could imply competition from monazite during later apatite growth (Henrichs et al., 2018).However, the elevated Th content in the depth profile rims indicates that at least the outermost rims did not grow with monazite, nor has monazite been identifed by mineral liberation analysis in this sample.The upper intercept is similar for both generations of apatite implying a process in which there was no U-rich mineral breakdown in the rock along with new mineral growth, to result in a different common Pb composition.The CL images of sample 205 apatite (Fig. 2) indicate overgrowths, which are not observable in BSE images.The overgrowth may have been generated during a dissolution regrowth process because the inner part appears slightly embayed.To account for the small amount of epidote observed in the sample, dissolution-reprecipitation in response to an infiltrating fluid must have occurred at T b 600 °C (Fig. 7), which is around the closure temperature for Pb in apatite.
The younger U-Pb regression from sample 216 apatite of c. 1800 Ma is interpreted to reflect new apatite growth from a compositionally distinct reservoir (Fig. 8) driven by regional tectonothermal and fluid activity, consistent with c. 1800 Ma phlogopite Ar-Ar ages in ultramafic complexes in the region (Cruz et al., 2016).Those analyses defining the younger regression have, on average, lower Sr, U and total REE than the older apatite population and apparently a different common Pb composition.The mobility of Sr and U is well known in hydrothermal solutions (Cherniak and Ryerson, 1993;Piccoli and Candela, 2002;Sha and Chappell, 1999) and REEs may be incorporated into or removed from apatite via coupled substitution reactions (Hughes and Rakovan, 2015).Additionally, grain depth profiling points to the growth of new U and Th rich apatite during the same process as causing recrystallization of c. 2500 Ma metamorphic apatite.Hence, there appears to be strong temporal coupling between loss of Pb in cores and the growth of REE,U and Th enriched apatite rims.

Apatite grain size and apparent age
Fig. 10 presents probability density plots of 207 Pb corrected apparent age versus grain size for both apatite samples.Radiogenic-Pb loss has been shown in some apatite systems to be solely, or at least dominantly, a function of thermally activated diffusion (Blackburn et al., 2011), a key requirement to use the mineral in thermochronometry modelling.In a situation of thermally activated volume diffusion it would be expected to observe distinct grain size populations that correspond to grain age populations; for example, a larger grain size population would correspond to an older age component, whereas a smaller grain size component may relate to a younger age component given Pb-in-apatite diffusion and spherical geometry characteristics (Cherniak, 2005;Cherniak et al., 1991).The histograms of grain age all indicate bimodal distributions whereas the grain size distributions are multimodal.Moreover, a significant grain size component in these, and likely many apatite geochronology samples, will not be characterizable given current analytical size constraints.Some authors have cautioned that fluid driven elemental mobility may be important in apatite geochronology and thus thermally activated volume diffusion a secondary process affecting U-Pb dates (Bingen et al., 1996;Chen and Simonetti, 2013;Chew et al., 2011;Chew and Spikings, 2015;Cochrane et al., 2014;Corfu and Stone, 1998).For the Greenland samples investigated in this work, fluid driven disturbance and growth processes are implicated on the basis of grain internal textures, petrography, and the relationship between apparent ages and chemistry, where new rim growth appears to be coeval with core U-Pb resetting.

Radiogenic-Pb loss and new apatite growth
The Pb diffusion closure temperature for apatite has been investigated by several workers (Cherniak, 2010;Cherniak et al., 1991;Krogstad and Walker, 1994) and Pb closure temperatures of around c.600 °C for large 100-200 mm apatite grains have been suggested (Krogstad and Walker, 1994).However other studies, comparing dates from different isotopic systems, have implied a somewhat lower closure temperature of c. 400-500 °C for apatite that may more generally apply (Chamberlain and Bowring, 2001).Apatite in both samples are associated with biotite and are located in fractured regions where fluids may have developed.In the current example of Archean apatite from Greenland it is challenging to accept that any of the age-depth profiles reflect solely thermally activated volume diffusion, given that the lower apparent age steps correspond with chemical (and in the case of sample 205) petrographically distinct new rims as well as the relatively low metamorphic temperatures calculated (b 600 °C) during any fluid ingress.The mineral apatite has been considered potentially favourable for investigating post magmatic crystallization thermal histories because of its Pb diffusion temperature, which is lower than many other U-Pb chronometers.Specifically, with knowledge of appropriate diffusion rates and accepting some simplifying assumptions, for example in regard to homogeneity of the diffusion process, it might be considered feasible to invert any age-depth profile in apatite to recover information on the mineral's thermal pathway.However, recrystallization or dissolution-regrowth would render any such thermal history from age-depth profiles suspect, as the patterns would not be governed by Pb diffusion alone.

Conclusions
1. Individual apatite generations (discrete grains or overgrowths) may incorporate different common Pb based on their local (microscopic) environment, which could be important for understanding neoblastic versus cooling ages given the high ratio of common to radiogenic Pb in some apatite grains.
2. Chemical and age zoning can be present in apatite even where apparently difficult to resolve in CL and BSE imaging.This age and chemical zonation need not directly spatially correspond, due to different diffusion rates.
3. New apatite rim growth may be coeval with recrystallization and dissolution of older cores.4.These dissolution/regrowth features in apatite can be formed at temperatures close to the lower temperature limits of what is generally thought to be the closure temperature of Pb diffusion in this mineral.5. To apply Pb diffusion profiles to determine cooling histories, one must not assume but demonstrate that thermally activated volume diffusion is justifiable because our results show that apatite generally may not record simple thermally activated Pb diffusion profiles but rather profiles modified in part or whole by recrystallization or new growth.6. Dissolution and regrowth forming high U rims is prevalent in these Greenland apatite and further illustrates this mineral is highly useful in tracking the timing of b600 °C fluid-rock interaction.
(19 analyses) has Sr concentrations of N224.4 ppm and defines a regression that yields a lower intercept on concordia of c. 2493 Ma and an upper intercept 207 Pb/ 206 Pb c ratio of 0.93 ± 0.27 (MSWD = 0.51).Adistinct younger discordia array (41 analyses) has lower Sr content of b211.1 ppm and yields a lower intercept of c. 1805 Ma (MSWD = 0.37) and an upper intercept 207 Pb/ 206 Pb c ratio of 1.165 ± 0.13.

Fig. 2 .
Fig. 2. Left; backscatter electron images of apatite from the Akia Terrane.Right; cathodoluminescence images of the same apatite.Note presence of clear overgrowths and embayed margin on sample 205 apatite.BSE: Backscatter-electron image, CL: Cathodoluminescene image.

Fig. 5 .
Fig. 5. U-Pb analytical data for zircons from samples 216: Leucosome, and 215: Granite sheet.Yellow squares indicate Group I (magmatic zircon); red squares indicate Group X (inheritance); green squares indicate Group M (zircon dating metamorphism); black squares indicate Group P (radiogenic-Pb loss); grey squares indicate Group D (discordance N10%).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6 .
Fig. 6.Time resolved ablation signal for various chemical, isotopic and derived parameters measured in depth profiling.Left sample 205, right sample 216.Upper, REE plots colour coded for ablation integration step depth.Higher step value indicates a deeper ablation and a position closer to the grain margin.Lower, ablation depth profiles.Values are fractionation corrected at each respective integration interval.Integration intervals shown as black boxes on Th and U signals, bars reflect two standard deviations from the mean value in the integration period.Grey dashed lines indicate the age components as found from the single spot ablations.Steps increase from left (half grain thickness) to right (grain edge).

Fig. 8 .
Fig.8.Inverse Concordia plots for depth profile ablations colours are coded according to integration step with warmer colours indicating a higher step and thus position closer to grain rim.Grey arrows denote interpreted response of U/Pb isotopic system to stated process.

Fig. 7 .
Fig. 7. (a)P-T pseudosection for sample 205 in an H 2 O-saturated system.Quartz predicted to be present in all fields.The stability of epidote is restricted to b650°C at ~0.9 GPa and to b450°C at 0.1 GPa.(b)isobaric (0.4 GPa) proportion-temperature diagram illustrating the main mineralogical changes from 300 to 650°C.At 450-500°C, plagioclase is predicted to completely break down to epidote and albite and K-feldspar is expected to partially breakdown to muscovite.

Table 2
Whole rock compositions for sample 205 used in phase equilibria modelling.