LA-ICP-MS and TEM constraints on the magmatic and post-magmatic processes recorded by the zircon-xenotime intergrowth in pegmatite (Pi ł awa G ´ orna, G ´ ory

A strongly altered zircon-xenotime intergrowth in pegmatite from the Pi


Introduction
Zircon (ZrSiO 4 ) is unambiguously the most frequently used mineral in the geochronology of igneous and metamorphic processes, and in provenance studies (Harley and Kelly, 2007;Nemchin et al., 2013).The chemical stability and limited incorporation of common Pb during zircon formation make it a powerful tool for U-Pb geochronology; however, the age record can be significantly disturbed by metamictization or fluid-mediated alteration processes (Villa and Hanchar, 2017).
Metamictization occurs in zircon through α decay of U and Th (or Pu), resulting in the destruction of the crystal lattice due to the recoil of the heavy daughter isotope when injecting a lighter alpha particle into the structure, leading to local displacement of atoms (Ewing et al., 2003;Geisler et al., 2007).The number of these recoil events destroying the zircon structure depends on the concentration of alpha emitters, their half-life and the duration of this process.The results are centres of displaced atoms and a local distortion of the structure, leading to a local decrease in density (Trachenko et al., 2002).Altered zircon from granite may be characterized by a matrix of Ca-rich metamict zircon with embedded islands of crystalline zircon, which indicates that the recoil events assimilate and overlap gradually, creating the texture of an amorphous matrix (Geisler et al., 2003a(Geisler et al., , 2003b(Geisler et al., , 2003c(Geisler et al., , 2007)).This certainly enables element transport in the less dense regions.The accumulation of crystal structure damage from irradiation is opposed by annealing effects, thus repairing the distorted crystal structure (Nasdala et al., 2001).
In addition to zircon, xenotime-(Y) (henceforth xenotime) is used as a geochronometer for a wide range of igneous, metamorphic and sedimentary processes (Rasmussen, 2005;Hetherington et al., 2008).Xenotime is a robust and chemically stable mineral with a high closure temperature of >900 • C (Cherniak, 2006;Cherniak, 2010).Its high resistance to fluid-induced alteration processes has been demonstrated by experimental studies replicating P-T conditions of the middle to upper crust (Budzyń et al., 2017;Budzyń and Kozub-Budzyń, 2015;Budzyń and Sláma, 2019;Hetherington et al., 2010).Nevertheless, the application of xenotime in geochronology is limited, mostly due to its low abundance in the Earth's crust.Further limitations are related to fluid-induced alteration that may disturb the U-Th-Pb system in xenotime, which has been demonstrated for metasediments (Rasmussen et al., 2011), pegmatites (Budzyń et al., 2018;Švecová et al., 2016) and in laboratory experiments (Budzyń and Sláma, 2019).This work investigates the intergrowth of zircon and xenotime (Fig. 1) from pegmatite from the Góry Sowie Block (NE Bohemian Massif, SW Poland) on micrometre and nanometre scales, which was previously studied by Budzyń et al. (2018).This strongly altered zircon- F. Tramm et al. xenotime intergrowth was selected for the study because it represents a model system that allows us to investigate different processes, which also affect the U-Pb system, leading to significant disturbance of the age record.The complex history recorded by the zircon-xenotime intergrowth includes three main events (Fig. 1d, e; Budzyń et al., 2018) The zircon-xenotime intergrowth contains important features to expand our knowledge of the alteration processes that disturb the geochronological record in both minerals.The altered xenotime with patchy zoning and porosity, together with zircon containing exceptionally high U contents, is an ideal natural laboratory to investigate a complex history.Transmission electron microscopy (TEM) imaging and analysis to explore nanoscale interactions in both minerals provide new implications to understand alteration processes on a submicron level.In addition, laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) trace element analysis completes the geochemical context of TEM submicron observations.
The pegmatite from the Piława Górna quarry (50 • 42 ′ 12.3 ′′ N 16 • 43 ′ 58.6 ′′ E; Fig. 2), classified as a NYF + LCT ((Nb, Y, F) + (Li, Cs, Ta)) pegmatite, demonstrates various zones from a fine-grained border and coarse-grained wall to intermediate zones and quartz core, which contain different mineral assemblages and geochemical signatures (Pieczka et al., 2012(Pieczka et al., , 2014;;Szuszkiewicz et al., 2013).The investigated pegmatite sample consists of large crystals of microcline, quartz, plagioclase and biotite, with accessory zircon, xenotime, fluorapatite and Y-rich allanite-(Ce).The studied case includes a ca. 2 mm-sized zircon, having a euhedral shape, intergrown with ca.0.5 × 1 mm xenotime.The entire xenotime grain is patchy zoned with developed porosity within the volumetric range of the zircon and is partially replaced by two fluorapatite grains on both sides (Fig. 1a, c).The two secondary fluorapatite grains form a sharp boundary with the zircon and an irregular transition zone with the replaced xenotime.The zircon demonstrates a distinct patchy zoned core with domains rich in microinclusions, such as uraninite and (U, Th)-rich phases.The zircon rim displays irregular patchy zoning and a high microporosity filled with (U, Th)-rich phases.For a more detailed sample description see Budzyń et al. (2018).

LA-ICP-MS analysis of trace elements
A Thermo Scientific Element 2 high-resolution sector field ICP-MS coupled to a 193 nm ArF excimer laser (Teledyne Cetac Analyte Excite laser) at the Institute of Geology of the Czech Academy of Sciences, Prague, Czech Republic, was used to acquire the zircon and xenotime trace element data.The laser was fired at a frequency of 5 Hz and fluence of 3.5 J/cm 2 with 16 μm laser spot size.The carrier gas was flushed through the two-volume ablation cell at a flow rate of 0.9 L/min He and mixed with 0.7 L/min Ar and 0.005 L/min N prior to introduction into the ICP.An in-house glass signal homogenizer (design of Tunheng and Hirata, 2004) was used for mixing all gases and aerosols, resulting in a smooth, spike-free signal.The signal was tuned for maximum sensitivity while keeping a low oxide level, commonly below 0.1%.Typical acquisition consisted of 15 s of blank measurement followed by measurement of TE signals from the ablated materials for another 25 s with 30 s of wash time between each analysis.This ablation pattern was repeated on two analytical spots located next to each other with the first ablation being used for collection of data at the low-mass resolution mode (m/Δm = 300: 89 Y, 91 Zr, 93   49 Ti in zircon).The measurement sequence consisted of repeated blocks of 2 analyses of NIST SRM612, one analysis of the BCR-2 standard, one analysis of 91,500 zircon reference material (Wiedenbeck et al., 1995) and 10 unknowns.
Trace element data were calibrated against the NIST SRM612 glass using Zr as an internal standard for zircon analyses and Y as an internal standard for xenotime analyses together with concentration values in the investigated zircon and xenotime obtained by using EPMA measurements.
The minimum detection limit was calculated by multiplying the variance of the background by 3.25 for individual elements, and the data are listed in the data table.The time-resolved signal data were processed using Glitter software (Van Achterbergh et al., 2001).The precision of the analyses (1 RSD) ranges between 5 and 15% for most elements.The accuracy was monitored by using a homogenized basalt reference material BCR-2 (USGS) and zircon 91,500 (Wiedenbeck et al., 1995).

Transmission electron microscopy
Ten electron transparent TEM foils were prepared using focused ion beam (FIB) milling.TEM was conducted on these foils to evaluate the structural properties of zircon and xenotime from the zircon-xenotime intergrowth in dedicated regions to characterise the eventual effects of various alteration processes.The FIB and TEM analyses were performed at the German Research Centre of Geoscience (GeoForschungsZentrum GFZ, Potsdam).For this purpose, several areas of 20 × 2 μm were selected and first protected with a 1.5 μm layer of Pt to protect the designated areas from FIB sputtering with the Ga-ion beam.An acceleration voltage of 30 kV and beam current of 45 nA were set for the lateral sputtering of the raw foils.The raw foils were placed on a halfmoon copper grid.Here, they were thinned to ca. 150 nm thickness using a setting of 30 kV and beam currents from 0.79 nA to 80 pA.Final polishing occurred at 5 kV and 41 pA.The foil used for TEM analysis had a size and thickness of 20 × 10 × 0.15 μm.For detailed information on the sample preparation procedure see Wirth (2004Wirth ( , 2009)).
TEM observations were performed using an FEI Tecnai G2 X-Twin TEM using a Schottky field emitter as the electron source.The TEM was equipped with a Fishione high-angle annular dark-field (HAADF) detector, an EDAX X-ray analyser and a Gatan electron energy-loss spectrometer (EELS).Diffraction patterns were calculated from highresolution electron microscopy images (HREM) as Fast Fourier Transforms (FFT).Phase identification was based on crystallographic parameters, such as Miller indices (hkl), angles between adjacent planes and the d hkl spacing.For unambiguous phase identification, the error of the angles had to be ≤1 • .The angular error in the FFT measurements was <0.5.
HAADF images were usually acquired as Z-contrast images at a camera length of 75 mm.In some cases, Z-contrast + diffraction contrast images were acquired at a camera length of 330 mm.In HAADF Zcontrast images, bright objects have a chemical composition dominated by high atomic numbers (e.g., U, Th or Pb) or higher density.In bright field (BF) images, this contrast is reversed and the same phases appear dark because the electrons are strongly absorbed by the high atomic number phases or the lesser density.

Geochemical characteristics of xenotime
Chondrite-normalized data of xenotime reveal high homogeneity in heavy rare earth elements (HREE) and high variations in the light rare earth elements (LREE) contents (Fig. 3a).The variation in LREE gradually decreases with decreasing ionic radius, from La to Eu.The characteristics of Ta vs. Nb (Fig. 3b) and Zr vs. Hf (Fig. 3c), which are geochemically closely related elements, show a strong positive linear correlation.Zirconium vs. LREE show correlations of Zr with La and Ce (Fig. 3d).The plot of (Th + U) vs. (Ca + Si) shows a cluster of data concentrated at ca. 8000 ppm Th + U and ca.1800-3100 ppm Ca + Si (Fig. 3e).Some data points plot along the line (Th + U)/(Ca + Si) = 1, which may indicate partial substitution of the components coffinitethorite and/or cheralite (cf., Förster, 1998;Spear and Pyle, 2002).Another substitution mechanism may be considered for the zircon component (Spear and Pyle, 2002).A plot of Si vs. Zr reveals two groups of data, (i) scattered below ca.1000 ppm Zr and above 1500 ppm Si, and (ii) data correlating with increasing Zr and Si concentrations, which may indicate partial substitution of the zircon component (Fig. 3f).Concentration of the nonformula Al increases with increasing Fe content roughly suggesting exponential correlation (Fig. 3g).

Geochemical characteristics of zircon
Chondrite-normalized zircon data from the three characteristic domains (core, rim and outermost rim) reveal distinct patterns (Fig. 4).The LREE pattern of the zircon core displays an elevated plateau followed by a negative Eu anomaly (Fig. 4a).The nearly plateau pattern is opposing the typical pattern of igneous zircons showing increasing slope for LREE with characteristic strong positive Ce anomaly (Hoskin and Schaltegger, 2003, and references therein) and resembles the LREE-enriched pattern of hydrothermal type-II zircons (Cavosie et al., 2006;Hoskin, 2005).The pattern of HREE displays a slightly increasing slope from Gd to Lu.The elevated LREE plateau in the rim of zircon is more developed, with chondrite-normalized LREE values above 1000, compared with the core data (Fig. 4b).The outermost rim has significantly lower concentrations of REE than the core and rim, with chondrite-normalized pattern showing increasing slope from La to Lu with a negative Eu anomaly (Fig. 4c).The investigated zircon represents a rare and exotic specimen containing exceptionally high contents of REE compared with a wide range of compositions reported for zircon (e.g., El-Bialy and Ali, 2013;Nardi et al., 2013;Bell et al., 2016Bell et al., , 2019)).
The element distribution of Ti vs. U for the core data shows a distinct group at exceptionally high contents ranging from ca. 19,000 to 106,000 ppm U, whereas the rim data are broadly scattered from ca. 2500 to 26,600 ppm U (Fig. 5a).The element distribution of P vs. (HREE + Y) shows elevated contents of HREE + Y for most data for core (ca.9300-34,500 ppm) and rim (ca.5800-48,300 ppm) populations, whereas data of the outermost rim is concentrated at lower HREE + Y contents (ca.2000-4100 ppm; Fig. 5b).The plot of (Nb + Ta) vs. Y reveals positive trends in the core and rim populations (Fig. 5c).The positive trend in the core is nearly parallel to the Y/(Nb + Ta) = 100 line, whereas the rim data follow a parallel trend above the Y/(Nb + Ta) = 10 line.The outermost rim has lower concentrations of ca.700-3100  ppm Y and ca.28-130 ppm Nb + Ta (Supplementary Table S2) plotting along the line Y/(Nb + Ta) = 10.
The investigated zircon bears an overabundance of La N and Ce N up to 3 magnitudes compared with igneous zircon characteristics of Hf vs. La N and Ce N (Fig. 6a, c).Data points of all domains show a broad scattering of Hf vs. Th/U, with broader scattering of data for the rim and outermost rim compared with the core (Fig. 6b).However, the Hf vs. U plot shows two distinct groups of data for the core and rim measurements (Fig. 6d).The core data plot at elevated contents of U (ca. 19,000-106,000 ppm) and lower contents of Hf (ca.14,000-18,000 ppm), whereas the rim data generally concentrate at lower U (ca. 2500-26,600 ppm) and higher Hf contents (ca.17,000-21,700 ppm).
Representative EDX spectra for all relevant phases in this and the following sections are presented in Supplementary Figs.S2-5.
The bright domains in HAADF mode (Fig. 9a) show a continuation of fine patchy zoning of bright and dark domains at the nanoscale with a higher abundance of bright domains.An array of micropores separates coarser patchy zoning from finer patchy zoning with dispersed nanoinclusions.These nanoinclusions of Th-and U-rich oxides, possibly thorianite and uraninite, are embedded in Ca-and F-rich phosphate, possibly fluorapatite (Supplementary Fig. S3c, d).Galena formed as a single crystal as displayed in the BF and HREM images and was included in the fluorapatite submicron inclusion (Fig. 9b, d-f).HREM imaging revealed that fluorapatite is amorphous near the contact with galena (Fig. 9e), whereas its crystallinity increases with increasing distance from the galena inclusion (lattice fringes in fluorapatite, Fig. 9f).Despite its rare abundance, the presence of galena inclusions documents potentially significant mobilization of Pb in zircon.
In HAADF mode, the dark patchy domains of the core are dominated by coarser patchy zoning (Fig. 9g) than the fine zoning in the bright domains (Fig. 9a).In both domains, EDX spectra indicate lower concentrations of U, and possibly Th, in the bright nanopatches compared with the dark nanopatches, which contain better defined spectral peaks of U (Supplementary Fig. S3a, b).The atomic number of elements, their concentration and density of materials influence the contrast of HAADF images.Uranium and Th are highly radioactive elements that destroy the crystal integrity of zircon by irradiation damage.Therefore, bright nanopatches (in HAADF imaging) with lower U and Th contents retain their crystal integrity more than dark patches with higher U and Th contents.Hence, the bright patches represent stronger electron scattering and thus brighter contrast in HAADF, whereas the dark patches display lower crystal integrity and density of the material, which reduces the electron scattering intensity.BF imaging emphasizes this variation in crystal integrity with a reverse contrast compared to HAADF.In BF mode, bright cloudy areas indicate a lower degree of crystallinity and a lower density of material (less electron absorption; Fig. 9h).Furthermore, diffraction patterns from bright and dark patches in HAADF display diffuse diffraction rings, which indicate the metamict state (Fig. 9i, j).Inclusions in the dark domain include predominantly Th-and U-oxides, possibly thorianite and uraninite.However, their unambiguous identification remains difficult due to their small size of only a few nanometres and interference with the surrounding zircon.One larger inclusion of uraninite is surrounded by a dark rim of amorphous zircon (Fig. 9k).The uraninite contains C (Supplementary Fig. S3f), possibly due to partial substitution of blatonite.The diffraction contrast of the BF image indicates a single crystal (Fig. 9l).However, diffraction patterns show broad diffuse scattering intensity, which is indicative of amorphous phases (Fig. 9m).
The inclusion-rich domain is dominated by a significantly higher abundance of micro-and nanoporosity compared with the other domains in the core, i.e., the bright and dark domains (Fig. 9n-q).The micropores are mostly filled with U-rich zircon (Fig. 9n, o), interpreted as the secondary zircon (Zrn2) based on EDX spectra compared with the primary zircon (Supplementary Fig. S3a, b vs. Fig.S3h).These inclusions of secondary zircon are accompanied or cut by veins of coffinite and surrounded by a dark rim of amorphous primary zircon (Fig. 9n, o).BF imaging revealed a matrix of crystalline secondary zircon penetrated by veins of less crystalline coffinite (Fig. 9o).The contact of the secondary zircon with the amorphous primary zircon is diffuse and frayed.The nanopores are either filled with clusters of small flakes of uraninite (Fig. 9p) or uraninite cubes embedded in coffinite (Fig. 9q).Furthermore, uraninite shows minor signals of C in the EDX spectra, which may suggest substitution of blatonite component (Supplementary Fig. S3i, j).All inclusions that contain significant amounts of U and Th commonly have a rim of amorphous zircon, which is related to radiation damage due to alpha decay of U and Th.

Zircon rim
The rim of zircon (foils F08, F09 and F10; Fig. 7) is the domain with the highest porosity, containing micro-and nanopores filled with the broadest variation of secondary phases ranging from Fe-rich silicate accompanied by monazite (Fig. 10b), Fe-and Pb-rich oxides (Fig. 10c,  d), uraninite (Fig. 10e) and uraninite-thorianite solid solution (Fig. 10k, l) accompanied by xenotime (Fig. 10k).Patchy zoning of the zircon rim is irregular, and the bright and dark domains in HAADF imaging display sharp transitions (Fig. 10a).Fe-oxides appear as folded plates embedded in Fe-rich silicate (Fig. 10c).Occasionally, zircon displays textures of either connected veins in a darker cloudy texture (dashed line in Fig. 10b) or as cloudy patches of primary zircon located in the wall material, potentially pathways for fluids (Fig. 10c).Aside from Fe-rich phases, uraninite is the dominant phase forming microinclusions in the porous rim.An example displays a polycrystalline core with an irregular rim surrounded by a thick rim (ca.150 nm) of amorphous zircon (Fig. 10f), which also occurs in similar U-rich inclusions in the inclusion-rich domain in the metamict core.Aligned HREM imaging revealed amorphization of the zircon rim surrounding the U-oxide (Fig. 10g), followed by a gradual increase in crystallinity demonstrated by a sharper diffraction pattern with decreasing diffuse scattering  intensity from the material further away (Fig. 10h-j).The zircon rim also comprises a domain of a homogeneous, highly crystalline zircon forming a thin rim surrounding the dominating, porous part of the rim (foil F10; Fig. 7).

The zircon-xenotime interface region
The phase boundary between xenotime and zircon is sharp and clear, highlighted by the crystalline xenotime and the metamict, patchy zoned zircon (foils F03 and F04; Fig. 7).Microporosity is more frequent in the zircon than in the xenotime (Fig. 11a).Inclusions representative of the zircon-xenotime interface region include small nanocrystals of uraninite, which form spheres (Fig. 11c) and cubic crystals embedded in a matrix of Y-rich Ca phosphate, interpreted as Y-bearing fluorapatite (Fig. 11d).The EDX data revealed compositional changes only within a few nanometres along a ca.400 nm line scan across a distinct geochemical phase transition (Fig. 11e, g-l).HAADF imaging reflects the sharp boundary between the xenotime and zircon (Fig. 11f).Additionally, BF imaging revealed a fine rim (ca. 100 nm) of radiation damaged xenotime embedding nanocrystals close to the zirconxenotime phase boundary.The zircon-xenotime intersection shows a decrease in Y content (Fig. 11g) and a significant increase in Si and Zr, as well as Ca and Th (Fig. 11i-l).A micropore in the primary xenotime, several microns away from the phase boundary, contains a sphere of fibrous Fe-oxides surrounded by oscillatory zoned Pb-and Fe-rich silicates, which are embedded in secondary zircon (Fig. 11n-y).The fibrous core contains nanocrystals of Pb-bearing Fe-oxides embedded in a matrix of Fe-silicate, which demonstrate their polycrystallinity in HREM images (Fig. 11p, q).In summary, all domains of zircon and xenotime investigated at the microscale and nanoscale demonstrate characteristic textural and compositional features summarized in Table 1.

LA-ICP-MS trace element characteristics
The investigated xenotime originated from Paleoproterozoic pegmatite (ca.2.1 Ga) and was altered during the formation of younger pegmatite in the Late Devonian (ca.370 Ma, Budzyń et al., 2018).The small variation in the chondrite-normalized HREE patterns (Fig. 3a) suggests the high chemical stability of the xenotime regarding its major element composition.Variations in trace and nonformula elements can provide information on alteration processes and fluid composition.Substitution mechanisms in xenotime include components of zircon due to their isostructural nature (Spear and Pyle, 2002), which is partially reflected by a weak correlation of Zr and Si in the investigated xenotime (Fig. 3f).However, the Si content in xenotime is also related to the coffinite-thorite substitutions (Fig. 3e; cf.Förster, 1998Förster, , 2006;;Spear and Pyle, 2002) and possibly reflects the presence of nanoinclusions of Fesilicate (Fig. 8b).The positive scattering of Al vs. Fe resembles an exponential correlation, which may be related to either (i) incorporation of Al during the formation of the xenotime or (ii) the transport of Al into the xenotime structure via fluid-supported diffusion and homogenization (Fig. 3g).Iron is one of the dominant elements in the secondary inclusions of the investigated xenotime and thus may serve as a key element for the presence of fluids.Furthermore, pegmatitic environments are strongly influenced by hydrous melts, with fluids enabling temperature-related diffusion (London and Morgan, 2012;Simmons and Webber, 2008).
To summarize, the presence of secondary inclusions may interfere with the true geochemical signatures of the investigated zircon and xenotime.The secondary submicron inclusions were emplaced in microand nanopores when precipitated from fluids, however, it cannot be excluded that some were incorporated during the growth of their host mineral.

Alteration processes at a submicron scale
Fluid-mediated alteration of minerals can result in the formation of complex textures and structural changes of the primary mineral, which is replaced by secondary material.There are two well-described alteration mechanisms, (i) diffusion-reaction and (ii) coupled dissolutionreprecipitation processes, that affect minerals under a broad range of temperature conditions (Geisler et al., 2007, and references therein).Alteration products of both processes display similar features and characteristics, which include developed porosity and secondary phases; however, their mechanisms differ significantly.Diffusion-reaction is a temperature-driven alteration process mediated by a hydrous medium (most likely H + ) resulting in nanoporosity providing fast exchange reactions between the reaction front and fluid and a gradual structural variation towards the reaction front (Geisler et al., 2003a(Geisler et al., ,c, 2004(Geisler et al., , 2007 ) ).In cases of metamict zircon, these reactions are represented by gradual structural recovery (Geisler et al., 2007).Furthermore, incompatible elements enter the solution and precipitate as phase inclusions (Geisler et al., 2007).Fluid-mediated coupled dissolution-reprecipitation reactions result in a replacement of the primary material by secondary phase(s) in a solid-state (Putnis, 2002(Putnis, , 2009;;Putnis and Austrheim, 2012).A chemically sharp boundary at the reaction front further penetrating inward is indicative of the breakup of elemental bonds (dissolution) and coupled formation (precipitation) of new bonds (Geisler et al., 2007;Putnis, 2002Putnis, , 2009;;Putnis and Austrheim, 2012).Nano-to microporosity is the result of the loss of volume in the crystalline secondary material (Geisler et al., 2007;Putnis and Austrheim, 2012).
Here, coupled dissolution-precipitation processes resulted in significant alteration of the xenotime, which is indicated by the presence of microscale patchy zoning and the nano-to microporosity filled with secondary inclusions, which are dominated by Fe-oxides, Fe-silicate, Pb phosphate, uraninite and coffinite-thorite (Figs.1c, 8).In general, xenotime remains mostly crystalline, even under high doses of selfirradiation, due to the counteracting alpha-annealing effects of He + ions, as an experimental study has recently shown (Rafiuddin et al., 2020).The effects preventing amorphization of xenotime due to selfirradiation may be similar or the same as the effects of monazite annealing, which are well described in experimental studies (Budzyń et al., 2021;Nasdala et al., 2020;Seydoux-Guillaume et al., 2002, 2018).
Because of the high resistance of xenotime against metamictization, the strongly damaged crystal structure in the xenotime close to inclusions of uraninite and coffinite-thorite solid solution was unexpected (Fig. 8i-m).The metamictization of xenotime observed in HREM imaging (Fig. 8l) indicates a higher rate of radiation damage caused by the decay of U and Th in the coffinite-thorite and uraninite inclusions than the annealing capacity of xenotime could repair.The metamict zircon core has been previously interpreted as affected by diffusion-reaction alteration, which possibly resulted in partial annealing of the crystal structure (Budzyń et al., 2018).Indeed, the fine matrix of bright and dark patchy domains in HAADF imaging may display features of partial annealing by diffusion-reaction processes.Damage of the crystal structure is the result of accumulating recoil events during the decay of alpha particle emitting elements (e.g., Th, U and daughter isotopes), which progressively disrupt the crystal structure of zircon (Geisler et al., 2007).The contrasting variations in zoning of the zircon core are the results of slightly higher U and Th contents, which lead to a lower crystal integrity and lower density of material in the dark patches (Fig. 9a, g).Thus, metamictization occurred at different degrees in these domains.Additionally, annealing of the radiationdamaged zircon starts at low temperature conditions of ca.200 • C (Geisler et al., 2007;Nasdala et al., 2010).It is possible that annealing initially supported by hot fluids repaired the primary zircon by gradually reconnecting the crystalline islands of zircon.Simultaneously, the fluid-activated elements accumulated in the matrix in the metamict zircon, ultimately leading to the formation of secondary zircon highly enriched in U. We interpret the patchy zoning of the zircon, which continues at submicron levels, as the result of metamictization and annealing effects predominantly supported by temperature-driven The aggregate is embedded in secondary zircon (Zrn3) forming an amorphous rim ca.70 nm around the aggregate.The white arrow represents the EDX line measurement presented at (r-y).(p) HREM image of the fibrous Fe-oxide revealing randomly oriented islands of crystal patterns indicating polycrystallinity.(q) The diffraction pattern shows multiple diffraction centres typical for a polycrystalline mineral surface.diffusion-reaction processes.The TEM observations demonstrate that the lower resolution of commonly used microanalytical techniques prevents the determination of detailed characteristics of the small patches in the zircon core, reflected in BSE imaging as patchy zoning (Figs.1a, 7), with respect to their compositional and textural heterogeneities.However, the hydrous medium enabling diffusion-reaction processes probably only allows limited element transport (Geisler et al., 2007).The formation of galena in nanopores (Fig. 9b) was supplied by Pb released from the altered zircon, whereas S must have been added by the fluid, which also supplied Ca and P for the crystallization of the secondary fluorapatite inclusions.These locally limited features might be associated with coupled dissolution-reprecipitation processes rather than diffusion-reaction alone.
The rim of amorphous zircon surrounding the uraninite inclusions in nanopores of the dark patchy domain (Fig. 9i, k, m) is a typical feature demonstrating that the presence of uraninite additionally enhanced metamictization of the zircon.In the inclusion-rich domain of the core, U-rich zircons formed as aggregates of bright microinclusions, which can be recognized in BSE imaging (Fig. 7).However, nanoscale observations revealed that the nanoinclusions of U-rich zircon are accompanied by coffinite or uraninite (Fig. 9n-q).This indicates the limited diffusion of the components released from the altered zircon via diffusion-reaction processes.
The porous rim of the zircon has a significant amount of connected micropores filled with uraninite or Fe-silicate.The predominance of the latter indicates that Fe was supplied by fluids from outside of the zircon.The network of micropores occasionally adjacent to veins of cloudy textured zircon (Fig. 10b) may be connected to differences in density or composition, thus serving as potential pathways for fluid penetration enabling rapid chemical exchange between zircon and the fluids as a result of coupled dissolution-reprecipitation processes.Typical features of coupled dissolution-reprecipitation processes and the gradual growth of pores are represented in a nanopore, which contains Fe-oxides and partially disconnected cloudy patches of zircon at the wall of the pore Fig. 12. Geochemical data of the zircon distributed in (a, b) discrimination fields of provenance (Grimes et al., 2007).(c-e) Distinction into zircons of magmatic or hydrothermal origin compared with zircons from the Boggy Plain Zoned Pluton (Hoskin, 2005).(f) Discrimination of magmatic and altered zircons (after Bouvier et al., 2012).
(Fig. 10c).The outermost rim of zircon displays a homogeneous crystalline structure (Foil 10, Fig. 7) and follows the almost euhedral prismatic shape of the zircon.Geochemical data of the outermost rim generally display a different composition compared with the core, and the rim is highly enriched in nonformula elements (Figs. 5,6,12,13).Coupled dissolution-reprecipitation reactions can result in the partial to complete replacement of the primary phase with a secondary phase of the same structure but different compositions with regard to nonformula and trace elements (Geisler et al., 2007;Putnis and Austrheim, 2012).Furthermore, the replaced phases often maintain their original shape.These results suggest that the origin of the outermost rim is the result of partial replacement due to coupled dissolution-reprecipitation processes.
The zircon-xenotime intersection highlights the alteration characteristics of both phases.It includes boundaries and phase transitions between xenotime and zircon in a concave pit (Fig. 11e).The zircon, accompanied by small cubic flakes of unknown material, displays a diffuse matrix accompanied by very small nanopatches of dark contrasting crystal islands in BF imaging (Fig. 11f).This indicates its secondary nature as an alteration product compared with the smooth patchy zoned texture of primary zircon in the core (Figs. 9g,11f).However, this would imply the precipitation of secondary zircon at the cost of the dissolution of primary xenotime.This interpretation is supported by the observed narrow pathway of zircon dispersing into the xenotime (Fig. 11e, f), which may serve as a channel for fluids penetrating the primary structure.
F. Tramm et al. consists of alternating layers of Pb-bearing Fe-silicate at its rim and needles of Pb-bearing Fe-oxides at its core (Fig. 11n).The walls of the micropore display frayed sharp edges (Fig. 11m, n).However, a clear interpretation is obscured due to loss of material during foil preparation.The layered Pb-bearing Fe-silicate is a product of alteration, with Fe and Pb probably precipitating from solution alternating in composition according to the fluid composition, serving as a trap for mobilized Pb.The rim probably formed once the fluid was saturated with Si and Pb, entrapping the conglomerate of Fe-oxides connecting the sphere of inclusions with the secondary zircon at the wall of the pore.

Constraints of the magmatic or hydrothermal records in zircon
The geochemical characteristics of zircon provide information regarding the geological setting and host rock in which zircon formed.A U/Yb ratio dependent on Hf and Y is used for differentiation into continental or oceanic provenances (Grimes et al., 2007(Grimes et al., , 2015, ;, ;El-Bialy and Ali, 2013).The Hf vs. U/Yb and Y vs. U/Yb data indicate the continental origin of the investigated zircon (Fig. 12a, b).
The genesis of pegmatites has not been completely resolved until now.The recent consensus is that pegmatites form from a granitic melt with a strong influence of fluids either floating buoyantly on the melt or in fluxes containing network suppressors and diffusion emphasizing elements, such as cations of rare alkalis, B, H, K and REE, or F and P (Simmons and Webber, 2008;London and Morgan, 2012 and references therein).Therefore, contradictory information on hydrothermal geochemical characteristics opposing igneous structures and the appearance of zircons is possibly unique to their pegmatitic provenance.The unresolved mystery of the genesis of pegmatites reflects the problematic differentiation of the origin of zircon formed in these environments.
Numerous previous studies have presented various approaches constraining zircon origin (magmatic or hydrothermal) or its alteration processes, which can be used to decipher the complex history of the zircon studied here (Bell et al., 2019;El-Bialy and Ali, 2013;Hoskin and Schaltegger, 2003).Hafnium is an important indicator of magmatic evolution for zircons in igneous systems (Claiborne et al., 2006;Hoskin and Schaltegger, 2003), as HfSiO 4 and ZrSiO 4 form solid solutions due to the close cation radii of Hf 4+ = 0.83 Å and Zr 4+ = 0.84 Å (Shannon, 1976).The overabundances of Hf vs. La N and Hf vs. Ce N (Fig. 6a, c) indicate enrichments far beyond those typical for igneous zircons.A high U content in the investigated zircon (up to ca. 10 wt%) also prevents application of the Th/U ratio (Fig. 6b, d).Classification of zircon origin according to fields using Ce/Ce*, (Sm/La) N, La (Fig. 12c, d), La N and Pr N (Fig. 12e) indicates that the characteristics of the investigated zircon are of hydrothermal origin, whereas only Ce/Ce* and (Sm/La) N show elevated contents.The U vs. Ca plot (Fig. 12f) shows that the investigated zircon is compositionally altered compared with fields from the data of Bouvier et al. (2012).A recent approach using the LREE index (LREE-I = Dy/Nd + Dy/Sm) classifies zircon into magmatic (LREE-I > 60) and altered (LREE-I < 10) zircon (Bell et al., 2016).According to this classification, only a fraction of data from the core falls into the unaltered zircon category of magmatic origin (Fig. 13).Indeed, alteration of the zircon core is evident from BSE and TEM observations (Figs. 1a,7,9).However, the consistency of the geochemical characteristics in the core data indicate a common origin and alteration for this dataset.On the other hand, clear indications for altered zircon are represented by all rim data plotting mostly close to LREE = 10 and below (Fig. 13).This break in trends between core and rim data reflects the different alteration processes of dominant diffusion-reaction processes in the core and coupled dissolution-reprecipitation processes in the rim, which are consistent with TEM observations discussed above (Figs. 9,10).This classification combined with TEM structural observations and the compositional characteristics of the core and rim demonstrate how different alteration processes may completely overwrite the geochemical fingerprints of primary crystallization of the zircon and may be attributed to diffusion-reaction and coupled dissolution-reprecipitation processes.

Fluid composition and element transport in the zircon-xenotime intergrowth
Fluid composition is one of the primary factors that controls the stability of phases, the potential precipitation of secondary phases and the formation of inclusions during the rock-fluid reaction.From the characteristic textures of zircon and xenotime at the nanoscale, observations of micro-and nanopores filled with secondary inclusions and geochemically relevant compositional changes of nonformula elements U, Ca, Fe, P and Pb, it is evident that fluids served as the dominant medium for element transport in the zircon and xenotime system.Using LREE-I as an indication for fluid transport-dependent enrichment in the altered rim of the zircon for elements such as Fe and Ca would imply transportation into the zircon, whereas P and Pb have been mobilized by fluids partially caught in pores in the altered rim (Fig. 14).TEM observations of secondary phases and inclusions in pores of the altered xenotime demonstrate additional pathways for element transport between xenotime and zircon and the environment: (i) The metamict core underwent minor element transport with a fluid composition including Ca.The porosity was mostly limited to nanopores (Fig. 9a, g), in which P, Pb, U and Th were captured.Predominant nanoporosity and limited element transport in the core indicate that fluids altered the core dominantly via the diffusion-reaction processes.Coupled dissolution-reprecipitation processes were limited to connected microporosity serving as pathways for fluid penetration.The high contents of Fe in the core may have been incorporated into the zircon during its formation.Although secondary Fe-bearing phases have not been documented in the core, element transport of Fe from the fluid to the core cannot be entirely excluded.(ii) The zircon rim was altered via coupled dissolutionreprecipitation reactions induced by Ca-and Fe-bearing fluids, which resulted in the development of microporosity.The concentration of Fe in the core is more than one magnitude lower than that in the rim (Fig. 13d), indicating that most of the Fe supplied by the fluid has been captured in the rim.(iii) The zircon-xenotime phase boundary was most likely influenced by fluid flow leading to major element transport that was effectively supported by coupled dissolution-precipitation processes.
The result was the formation of foreign inclusions (e.g., Fesilicate; Fig. 11o) and secondary zircon at the cost of primary xenotime directly at the phase boundary (Fig. 11a,e).
Aside from the element transport observed in inclusions, zircon and xenotime in this study, there are indications for F-bearing fluids related to fluorapatite formation.Fluorapatite formed on two sides of the xenotime and directly intertwined between the zircon and xenotime phase boundaries (Fig. 1a; Table S3), serving as a sink for Ca and F from the fluid.This interpretation is supported by the documentation of fluorapatite inclusions in the zircon core (Fig. 9b) and zircon-xenotime interface (Fig. 11d).

Geochronological implications
Interpretation of geochronological data of altered material includes (i) the consideration of substitution mechanisms that result in broad compositional changes inside the structure of the mineral and (ii) interfering inclusions that potentially accumulate geochronologically important elements, such as U, Th and Pb.Interfering microinclusions are usually recognized during measurements with LA-ICP-MS.However, TEM observations reveal nanoinclusions, which will influence the geochronological record unrecognized by microscopic analyses due to their small size.An example of these inclusions are nanoinclusions of galena, incorporating major amounts of Pb, and uraninite or thorite.
Bright patchy domains and dark patchy domains (Fig. 9a, g) at the nanoscale in HAADF images represent various degrees of a metamict state in the core of the investigated zircon (Fig. 9i, j).This demonstrates how only slight variations in the U and Th contents can result in different degrees of crystallinity.TEM revealed that the bright and dark domains are relatively scarce of nanoinclusions, which accumulate in connectedpore systems (Fig. 9a, g).Therefore, discordant age data in these domains are predominantly the result of nanoscale heterogeneity in U and Th contents and variations in the degree of metamictization.The inclusion-rich domain (Fig. 9n) is dominated by secondary inclusions of U-rich zircon, coffinite and uraninite, which resulted in a stronger accumulation of U, Th and Pb.These domains are capable of locally disturbing the geochronological record to a greater extent because differentiation of inclusions from the surrounding zircon will be difficult during measurements in these inclusion-rich domains using microscopic methods.Furthermore, the predominant microporosity in the altered rim of the zircon, filled with inclusions of Fe-oxides, Fe-silicate and uraninite, demonstrates substantial element transport of the geochronologically important elements U, Th and Pb.Nanoinclusions accumulating these elements are not detectable at the microscale and might be difficult to avoid using a large spot size of several microns.Therefore, interpreting the ages recorded in highly heterogeneous domains with a high abundance of nanoinclusions has to be treated with great caution and might not be suited to constraining the age of the zircon without nanoscale analyses, such as TEM observations.
The altered xenotime is patchy zoned at the microscale and shows a greater extent of microporosity closer to the neighbouring zircon, but bears a homogeneous crystallinity and composition regarding nanoscale observations of foils in domains of lower porosity further away from the zircon.The limited presence of microinclusions would be easily detected during measurement with microscopic methods; however, nanoinclusions of Pb-phosphate, uraninite and thorite might be problematic if their presence is significant within the area of interest.Furthermore, the zircon and xenotime are patchy zoned, indicating broad alteration of the entire intergrowth.Patchy zoning has been recognized as a result of alterations, which induce partial to complete resetting of the U -Pb and Th-Pb age record due to mobilization of these elements (Budzyń et al., 2021;Williams et al., 2011).To summarize, understanding zircon and xenotime mineralogy at the submicron-scale is important to avoid misinterpretations in km-scale geochronological reconstructions.

Conclusions
TEM investigations of the zircon-xenotime intergrowth demonstrated characteristic and complex features of different alteration processes, such as: (i) Diffusion-reaction processes in the metamict core of the zircon generally promoting nanoporosity and partial repair of the internal structure of the zircon core, albeit showing limited element transport locally and accumulating aggregates of larger microinclusions.(ii) Coupled dissolution-precipitation processes resulting in major porosity and element mobilization of U, Th and Pb outward, whereas nonformula elements such as Ca and Fe were transported inward in the fluids.
The TEM observations are in good agreement with the geochemical characteristics of the investigated zircon and xenotime, and demonstrate their importance in identifying nanoinclusions and their abundance, which potentially interfere with the geochronological interpretation of these materials undetected using microscopic methods.This study emphasizes the importance of TEM investigation in addition to microanalytical methods to complete the geochemical and geochronological evaluation of altered zircon or xenotime, particularly in complex alkalirich environments.

Declaration of Competing Interest
None.
: (i) Formation of the zircon-xenotime intergrowth ca.2.09 Ga in Paleoproterozoic pegmatite.(ii) Incorporation of the zircon-xenotime intergrowth as a restite into the newly formed Late Devonian pegmatite at ca. 370 Ma; diffusion-reaction processes induced re-equilibration in the zircon core; coupled dissolution-reprecipitation processes affected the zircon rim and the xenotime, which resulted in disturbance of the U -Pb system and partial replacement of xenotime by fluorapatite.(iii) Termination of the fluid-mediated, low-temperature alteration ca.280 Ma, which involved crystallization of uraninite inclusions in the structurally damaged transition between the core and rim of the zircon.

F
.Tramm et al.

Fig. 4 .
Fig. 4. Comparison of chondrite-normalized trace element patterns in (a) the zircon core, (b) the zircon rim and (c) the outermost rim.C1 chondrite composition after McDonough and Sun (1995).

Fig. 5 .
Fig. 5. Summary of LA-ICP-MS trace element data for the zircon.See text for details.

Fig. 6 .
Fig. 6.Trace element characteristics presenting constraints on the provenance of the investigated zircon.The fields of igneous zircon after Hoskin and Schaltegger (2003).

F
.Tramm et al.

Fig. 8 .
Fig. 8. TEM results for the xenotime in foils F01 and F02.(a, b) HAADF images demonstrating micropore filled with Fe-silicate.(c) Dislocation veins occasionally pinned to nanopores.(d) A shifted crystal lattice along dislocations (dotted line) resulting in a shift of diffraction contrast.(e) Nanopore enclosing Pb 3 (PO 4 ) 2 .(f-h) BF and HREM images demonstrating unicrystallinity of the Pb phase and FFT patterns of the crystallographic parameters.(i) Nanopore enclosing inclusions of coffinitethorite solid solution and uraninite.(j) The rim of the nanopore demonstrates aggregates of nanocrystals of xenotime.(k) Xenotime nanocrystals (dashed line) in the poorly crystalline xenotime matrix.(l, m) HREM and FFT images of the primary xenotime with a damaged crystal structure surrounding the pore.Abbreviations: Cof-Thrcoffinite-thorite solid solution, Urnuraninite, Xtmxenotime.

Fig. 9 .
Fig. 9. TEM results for the metamict core of the zircon, which consists of three domains.Bright patchy domain: (a) overview of the foil revealing two areas of patchy zoning divided by a group of micropores.(b-d) Inclusion of galena surrounded by secondary fluorapatite.(e) HREM image of the transition between galena and amorphous fluorapatite.(f) HREM image of the crystalline fluorapatite ca. 10 nm from the transition with the galena inclusion.The dark patchy domain: (g) overview of coarse patchy zoning displaying bright and dark domains.(h) BF image showing differences in crystallinity and density of the patchy zones; here, the contrast is inverted compared with HAADF.(i, j) Diffraction patterns of the patchy domains of zircon.(k) Nanoinclusion of uraninite, which is common in the dark patchy domain.(l) BF image of the uraninite that indicates unicrystallinity, yet (m) diffuse diffraction rings indicate a low degree of crystallinity.The inclusion-rich domain: (n) HAADF overview of the inclusion-rich domain.(o) BF image of a microinclusion of U-rich zircon (Zrn2) cut by veins of coffinite.(p) Aggregate of flakes of Pbbearing uraninite.(q) Cubic inclusions of uraninite embedded in coffinite.Abbreviations: Apfluorapatite, Cofcoffinite, Gngalena, Urnuraninite, and Zrn zircon.

Fig. 10 .
Fig. 10.TEM results for the rim of the zircon.(a) HAADF overview documenting patchy zoned zircon with a high microporosity filled predominantly by Fe-silicate or uraninite.(b) A micropore enclosing Fe-silicate and monazite, with the channel of patchy zoned zircon attached to the pore (dashed lines).(c) Fe-oxides forming folded layers embedded in Fe-silicate in a nanopore.(d) Nanocrystals of Fe-oxides in the core of the folded material.(e) The uraninite inclusion in a micropore.(f) Amorphous zircon surrounding the nanopore with the uraninite.(g-j) HREM images of zircon demonstrating increasing crystallinity subsequently with increasing distance from the uraninite inclusion.(k) A rare case of a micropore containing inclusions of xenotime and uraninitethorianite solid solution.Galena also formed in a nanopore close to the xenotime inclusion.(l) HREM image of the highly absorsive oxide reveals clouds of shifted crystal orientation indicating polycrystallinity.Abbreviations: Gngalena, Mnzmonazite, Urnuraninite, Urn-Thouraninite-thorianite solid solution, Xtmxenotime, and Zrnzircon.

Fig. 11 .
Fig. 11.TEM results for the zircon-xenotime interface.(a) HAADF overview showing a sharp phase boundary between the xenotime and the patchy zoned zircon (F03).(b) domain located at the phase boundary demonstrating alteration pits in the primary xenotime, partially filled with secondary zircon (Zrn2).(c) Nanocrystals of uraninite in xenotime close to the xenotime boundary and (d) nanocrystals of uraninite embedded in the fluorapatite inclusion in the zircon, which is several micrometres from the Zrn-Xtm boundary.(e) The alteration pit shows the reaction front with the secondary zircon adjacent to the primary xenotime.The white arrow represents the EDX line measurement presented in (g-l).(f) BF image showing a ca. 100 nm thick rim of altered xenotime containing fine nanocrystals (dark areas).The inclusions in the secondary zircon (Zrn2) form a pathway penetrating into the primary xenotime (white dotted arrow).(m) Overview of the zircon-xenotime phase boundary (F04) with a micropore in xenotime close to the xenotime-zircon boundary.(n) HAADF image showing the micropore enclosing secondary zircon (Zrn2) and an aggregate containing a fibrous core of Fe-oxides surrounded by an oscillatory zoned rim with varying contents of a Pb-bearing Fe-silicate.(o)The aggregate is embedded in secondary zircon (Zrn3) forming an amorphous rim ca.70 nm around the aggregate.The white arrow represents the EDX line measurement presented at (r-y).(p) HREM image of the fibrous Fe-oxide revealing randomly oriented islands of crystal patterns indicating polycrystallinity.(q) The diffraction pattern shows multiple diffraction centres typical for a polycrystalline mineral surface.

Fig. 14 .
Fig. 14.Sketch of the investigated zircon-xenotime intergrowth with a summary of the element transport during fluid-mediated alteration processes.

Table 1
Summary of textural features in the zircon and xenotime determined by TEM.