Current applications using key mineral phases in igneous and metamorphic geology: perspectives for the future Open Access

Unravelling the composite tectono-magmatic and metamorphic evolution of terrains can be challenging because their multistage histories may have erased past information. One way to interrogate and address these complexities is to investigate geological processes throughout Earth's history that are encapsulated in key minerals such as garnet, zircon, titanite, allanite, rutile, monazite/xenotime and apatite. The mineral phases chosen for this contribution are the most used in petrochronological studies as they are reliable chronometers, and their chemical variability has been used to investigate metamorphic and magmatic crustal processes. As petrochronometers, they can retain information about the petrogenesis of their protoliths, (multiple) pressures (P), temperatures (T) and timing (t) at which they (re)crystallized in magmatic, metamorphic and/or hydrothermal events. Investigations employing these minerals have provided new breakthroughs in Earth Sciences. Scientific findings through petrochronological investigations include (but are not limited to) the study of the oldest minerals on Earth (Valley et al. 2014), the understanding of trace element (TE) partitioning processes between different key mineral phases (e.g. Rubatto 2002) and the study of fluid and magma sources during crust evolution (e.g. Valley 2003; Bouvier et al. 2012; Mulder et al. 2021). Magmatic and/or metamorphic overprinting stages are heterogeneously recorded by these petrochronometers, which preserve snapshots of a rock evolution. The timing, conditions of formation and distinct closure temperatures (i.e. Tc as the temperature of a geochronological system at the time corresponding to its apparent age; Dodson 1973) that characterize minerals and isotope systems (Fig. 1) allow, for example, the U–Pb system in garnet (Mezger et al. 1989a) to retain information at higher temperatures than in rutile, which resets its internal clock as a function of T, grain size and cooling rate. Thus, carrying out multi-mineral studies (Fig. 1) is an incredibly powerful tool to investigate complex magmatic and metamorphic histories (e.g. Cutts et al. 2014; Stearns et al. 2015; Manzotti et al. 2018; Volante et al. 2020c; Fumes et al. 2022; Odlum et al. 2022).

Over the past decades, our ability to interrogate the preserved rock record of crustal metamorphism and magmatism has significantly advanced thanks to continuously improving in situ analytical techniques, measurement methodologies and protocols, as well as availability of reference materials that have been calibrated through worldwide inter-laboratory collaborations. Most commonly used instruments include secondary ion mass spectrometry (SIMS) (e.g. Chamberlain et al. 2010; Ushikubo et al. 2014; Zhou et al. 2016; Chaussidon et al. 2017) and inductively coupled plasma mass spectrometers (single and multi-collector, sector-field, single and triple quadrupoles) associated with a laser ablation system (LA-ICP-MS, LA-MC-ICP-MS or LA-ICP-MS/MS) either used in tandem (split stream) or separately (e.g. Kylander-Clark et al. 2013; Hacker et al. 2015; Simpson et al. 2021). Additionally, improvements in microanalytical techniques, such as (but not limited to) electron backscatter diffraction (EBSD) mapping (e.g. Cavosie et al. 2015), TESCAN integrated mineral analyser (TIMA; e.g. Porter et al. 2020), field emission scanning electron microscope (FESEM; Tacchetto et al. 2022) for qualitative analysis acquisition and field emission gun (FEG) transmission electron microscope (TEM) for atomic lattice resolution imaging and nanoscale microanalysis (Reddy et al. 2020; Tacchetto et al. 2021), have significantly contributed in advancing the micro- to nanoscale petrographic, chemical, physical and structural characterization of sample material. The integration of continuously updated thermodynamic modelling studies of monazite (e.g. Janots et al. 2007; Kelsey et al. 2008; Spear and Pyle 2010; Hoschek 2016), allanite (Spear 2010), zircon (Kelsey et al. 2008; Kelsey and Powell 2011; Yakymchuk and Brown 2014) and apatite (Spear and Pyle 2010; Yakymchuk 2017) allows for the construction of a fundamental basis from which to overcome problems related to qualitative interpretations of accessory minerals within the P–T space, for which rutile (e.g. Fumes et al. 2022; Holtmann et al. 2022; Horton et al. 2022; Vanardois et al. 2022) and titanite (e.g. Kapp et al. 2009; Kirkland et al. 2016, 2020; Apen et al. 2020; Walters et al. 2022) already greatly contribute. The above-mentioned analytical advances combined with more traditional petro-structural (micro)analyses (Volante et al. 2020b) and continuously updated thermodynamic datasets used for modelling pressure (P), temperature (T) and composition (X) are continuously contributing to improve our understanding of crustal processes by allowing a more detailed characterization of mineral phases.

This review aims to showcase snapshots of recent analytical developments such as advancement in understanding the capabilities of LA-ICP-MS/MS and their applications, including new findings in the uptake of water in zircon crystal lattice and its influence on oxygen isotope composition, development and improvement in generating in situ quantitative elemental maps that can be directly linked with petrochronological and textural information using key mineral phases in igneous and metamorphic studies. In particular, we focus on the most employed minerals in the literature that can be used not only as chronometers in a great variety of rock lithologies but also as tracers of fluid–rock interactions at crustal conditions, and they include: garnet, zircon, titanite, rutile, allanite, monazite/xenotime and apatite (Fig. 2). In the following section, each mineral is described including its potential as chronometer, thermobarometer and chemical/isotopic tracer applied to crustal magmatic and metamorphic processes. Ultimately, the Discussion section provides examples of applications that use these petrochronometers in magmatic and metamorphic studies with a focus on the complementary set of information obtained when multi-mineral investigations are used to unravel different steps in the evolution of a volume of rock. Future perspectives on the use of these key mineral phases to investigate metamorphic and igneous processes are also included.

Garnet (Fig. 2a-c; Fig. 3) is a major rock-forming mineral, occurring in metamorphic (felsic and mafic compositions) and felsic peraluminous magmatic rocks (Villaros et al. 2009; Bartoli et al. 2013; Liu et al. 2014; Volante et al. 2020a; R. Li et al. 2021). Garnet is a key metamorphic indicator mineral (Barrow 1893, 1912) and its composition directly relates to the P–T conditions it experienced during growth (the garnet stability field extends from greenschist-facies to high-pressure (HP) and high-temperature (HT) conditions, dependent on rock composition), making it extremely useful as a geothermobarometer. Garnet grains can shield mineral inclusions whose identity or composition can indicate earlier assemblages/conditions experienced by the rock (e.g. St-Onge 1987; Lü et al. 2008; Thomas and Davidson 2012; Pourteau et al. 2019; Schönig et al. 2019; Godet et al. 2022). These mineral inclusions can also preserve or trace earlier foliations, which can be used to infer older foliation trends (e.g. Sayab 2006; Aerden et al. 2013, 2021; Sayab et al. 2015, 2016; Volante et al. 2020b). Like other porphyroblasts, garnet can also infer kinematics via grain rotation (Fig. 3a; Passchier and Simpson 1986; Johnson 1999). Garnet can record a prolonged history, preserving a growth hiatus, where textural, chemical or chronological evidence shows that garnet grew during more than one metamorphic and/or magmatic event (Fig. 3b; e.g. Vance et al. 1998; Cutts et al. 2010; Ortolano et al. 2014; Kulhánek et al. 2021; Massonne and Li 2022).

Garnet is a powerful tool for linking accessory mineral geochronological and/or isotopic data to the evolution of the major mineral assemblage using trace elements. Linking these two data sources together is the cornerstone of petrochronology (Engi et al. 2017). Trace elements such as rare earth elements (REE) partition strongly into garnet (Bea et al. 1994; Hermann and Rubatto 2014), particularly the middle to heavy rare earth elements (M-HREE; Fig. 3c). Other minerals that also readily accommodate M-HREE, such as zircon and monazite, compete with garnet, resulting in modified trace element patterns when the accessory phase grows in the presence or absence of garnet. The use and effectiveness of these trace element partitioning relationships (D_REE) were first presented for zircon and garnet (Fig. 3d; Rubatto 2002). Zircon grown in the presence of garnet has a M-HREE slope of c. 1 effectively flat, and a slightly curved to near flat D_REE pattern (Whitehouse and Platt 2003; Hokada and Harley 2004; Kelly and Harley 2005; Harley and Kelly 2007; Wu et al. 2008a, b; Fornelli et al. 2014, 2018). Since these first studies, trace element partitioning has grown into a widely used tool in petrochronology (Whitehouse and Platt 2003; Hokada and Harley 2004; Baldwin and Brown 2008; Clark et al. 2009; Harley and Nandakumar 2014), and some studies conducted quantitative partitioning experiments to constrain temperature dependent patterns for D_REE (Zrc/Grt) (Rubatto and Hermann 2007; Taylor et al. 2015), building upon empirical studies (Rubatto 2002; Hermann and Rubatto 2003; Whitehouse and Platt 2003; Hokada and Harley 2004; Kelly and Harley 2005; Buick et al. 2006). To address the challenge of handling extensive datasets generated by LA-ICP-MS, Taylor et al. (2017) developed an array plot for trace element partitioning data that can also show additional trends compared to the standard D_REE plot (Fig. 3e). D_REE (Mnz/Grt) is also a useful partitioning system (see Monazite section for more details). Trace element mapping of garnet via LA-ICP-MS is an area that is currently undergoing significant advances (e.g. Chew et al. 2017, 2021; Raimondo et al. 2017; Rubatto et al. 2020). M-HREE, Y and Cr zoning in garnet provides additional information about the growth history and mineral relationships not apparent in major element zonation (Raimondo et al. 2017).

Since the late 1980s, U–Pb and Sm–Nd geochronology (Mearns 1986; Mezger et al. 1989b) used multiple mineral phases to determine the age of whole-rock. The application of these methods led to the discovery of a favourable spread in Sm–Nd ratios between garnet and whole rock (Humphries and Cliff 1982). U–Pb garnet dating was found to be affected by U-rich inclusions such as zircon, monazite and apatite. Thus, for the following 30 years, garnet geochronology has been conducted using Sm–Nd and, subsequently, once instrumentation advanced sufficiently to measure low Hf contents in garnet, the Lu–Hf system was used for garnet dating (Duchêne et al. 1997). Inclusions such as monazite and apatite for the Sm–Nd system were treated with a HCl leaching step method that would remove significant monazite Nd components (Scherer et al. 2000), whereas Lu–Hf garnet geochronology is greatly affected by inclusions such as zircon, with samples containing significant contents of zircon being avoided for garnet whole-rock dating (Scherer et al. 2000). The following studies have compared the two isotopic systems for garnet dating via dissolution (Vervoort 2013), and they have shown that the Lu–Hf system (Tc c. 650–900°C) has slightly higher Tc than Sm–Nd (Tc c. 600–800°C) (Fig. 1; Scherer et al. 2000; Smit et al. 2013; Johnson et al. 2018). Traditionally, both Sm–Nd and Lu–Hf systems via isotope dilution–thermal ionization mass spectrometry (ID-TIMS) or MC-ICP-MS dissolution methods require the separation and purification of garnet prior to dissolution, column separation and measurement that is commonly associated with loss of the textural context of the analysed grains (cf. Pollington and Baxter 2010). Prolonged (e.g. Baxter and Scherer 2013; Bollen et al. 2022) or incremental garnet growth makes a single age from garnet separate, potentially geologically meaningless (e.g. Pollington and Baxter 2010). To overcome these issues, garnet has often been separated based on colour (e.g. Vance et al. 1998) or by using a microdrill to target garnet growth rings in order to determine the growth rate of garnet crystals (Fig. 3f; see Pollington and Baxter 2010, 2011; Dragovic et al. 2012, 2015; Schmidt et al. 2015; Tual et al. 2022). Results of such studies have showcased how effectively garnet can be used for geochronology, the caveat being that these methods are expensive, time consuming (c. 2 months) and the presence of inclusions must be considered and evaluated.

The recent introduction of ICP-MS/MS, which eliminates isotopic interferences for Lu–Hf measurements (i.e. ¹⁷⁶Lu and ¹⁷⁶Yb on ¹⁷⁶Hf), coupled with an increase in machine sensitivity allowing to measure elements at very low concentrations (i.e. part per billion, ppb), has permitted significant advances in garnet geochronology. These analytical advances not only are at the foundation of the development of in situ Lu–Hf measurements in garnet (see Simpson et al. 2021; Brown et al. 2022; Ribeiro et al. 2022; Tamblyn et al. 2022; Simpson et al. 2023), but also resulted in revisiting of the U–Pb system in garnet with in situ analyses now possible using LA-ICP-MS or -SC (single collector) -MS (Seman et al. 2017; Gevedon et al. 2018; Millonig et al. 2020). Since garnet commonly has low U contents, the presence of U-rich inclusions remains an issue for in situ methods. Nonetheless, Uranium content in garnet can easily be tested via LA-ICP-MS spot analysis prior to U–Pb analysis and coupled with detailed textural imaging to identify potential U-rich inclusions. Grossular and andradite garnet tend to have enough U (>1 part per million, ppm), and many studies have used this method as a means of dating skarns related to mineralization (Seman et al. 2017; Gevedon et al. 2018; Wafforn et al. 2018; Duan et al. 2020). More recently, other works have also found almandine-pyrope-rich garnet with lower but enough U (>0.02 ppm) to produce reliable metamorphic ages (Millonig et al. 2020; Cerva-Alves et al. 2021; Schannor et al. 2021; O’Sullivan et al. 2023). Nonetheless, U-rich inclusions such as monazite, apatite and zircon shift data points on a Tera–Wasserburg (TW) diagram to the left, right or along the regression, reflecting older, younger or the same age as garnet, respectively (Fig. 3g; Millonig et al. 2020). Besides textural observations, to screen and filter out inclusion-dominated analyses it is crucial to monitor the masses on the ICP-MS such as Y and Zr for zircon, or light rare element (LREE) for monazite and allanite. The in situ U–Pb garnet dating method (Tc c. 650–900°C) has the advantage of minimal sample preparation, rapid analysis (2 hours) and the possibility of monitoring the signal. The Lu–Hf method is limited to garnet that contains sufficient Lu for measurement and relatively inclusion free, with zircon inclusions significantly affecting the age results (Simpson et al. 2021). This method has larger uncertainties than solution-based Lu–Hf analysis, but numerous advantages. Low Lu or high Hf (i.e. zircon) inclusions can be targeted to produce an isochron from a single garnet grain. The ablated spots (c. 40–120 µm) can target different chemical domains, including separate core and rim isochron ages (Fig. 3h–j; Tamblyn et al. 2022). The ability to produce an isochron from a single garnet grain means that this method can also be used to investigate detrital garnet (see Pereira et al. this volume, in press). An alternative to directly date garnet is targeting datable primary mineral inclusions such as monazite (e.g. Mahan et al. 2006; Cutts et al. 2010; Williams et al. 2017). However, possible preservation of detrital or older inclusions must be evaluated (Martin et al. 2007; Cutts et al. 2009, 2013; Peixoto et al. 2018).

Oxygen isotopes have been applied to whole-rock and mineral samples to trace the sources of metamorphic/metasomatic (e.g. Chamberlain and Conrad 1991; Crowe et al. 2001; Page et al. 2010; Martin et al. 2011, 2014; Vho et al. 2020) and magmatic/mantle fluids (e.g. King and Valley 2001; Lackey et al. 2006; Harris and Vogeli 2010). Initially, oxygen isotopes were applied to garnet to determine pressure estimates, based on the temperature dependent fractionation between co-existing mineral pairs (experimentally or empirically determined; Baxter et al. 2017 and references therein). This method utilizes slow diffusion of oxygen within garnet, thus peak temperature isotope fractionation is retained (Kohn and Valley 1998; Baxter et al. 2017). Improvements in analytical methods (analysis via SIMS) allowed for in situ analysis of oxygen isotopes in garnet to track the origin of metamorphic/metasomatic fluids (e.g. Raimondo et al. 2012, 2017; Russell et al. 2013). For example, garnet can preserve variation of fluid compositions as it grows, particularly in eclogites or skarns (D'Errico et al. 2012; Rubatto and Angiboust 2015). Oxygen in garnet is very similar to the bulk rock δ¹⁸O value, but deviations can occur following temperature variations affecting equilibration factors between minerals (i.e. garnet–quartz), changes in modal proportions and mineral assemblage during metamorphism as well as the presence of externally sourced fluids. Nonetheless, experimentally determined oxygen diffusion rates in garnet show that original isotope compositions can be retained despite partial re-equilibration of major elements (Scicchitano et al. 2021). Other stable isotopes such as Fe or Li may be applied in garnet (Bebout et al. 2014, 2022; An et al. 2017; Gerrits et al. 2019; Penniston-Dorland et al. 2020; Hoover et al. 2021, 2022), with limitations of these largely depending on analytical ability.

Geothermometry in garnet (Ferry and Spear 1978; Baxter et al. 2017) was initially applied via Fe–Mg exchange between garnet and biotite (Ferry and Spear 1978; Perchuk and Lavrent'Eva 1983; Ganguly and Saxena 1984) and geobarometry via the ‘GASP’ (garnet–aluminosilicate–silicate–plagioclase) end-member reaction, which is pressure sensitive due to the large molar volume difference between reactants and products (e.g. Ghent 1976; Newton and Haselton 1981; Holdaway 2001; Caddick and Thompson 2008). Later, large thermodynamic datasets (e.g. Berman 1988; Holland and Powell 1990, 1998, 2011) allowed for P–T pseudosection calculation, which was combined with the use of garnet end-member compositions to constrain the growth path or max P and T conditions (e.g. Spear et al. 1984; Vance and Mahar 1998). A caveat in the use of this method is the isolation of grown garnet (i.e. core domain) from the whole-rock composition used for P–T pseudosection calculation (e.g. Marmo et al. 2002; Evans 2004; Tinkham and Ghent 2005; Cutts et al. 2009, 2010). This approach is particularly useful where garnet growth domains are the result of different orogenic events (e.g. Cutts et al. 2014). Software for modelling evolving bulk compositions during garnet growth along a P–T path was developed to produce more reliable P–T estimates (Theria_G, Gaidies et al. 2008; GRTMod, Lanari et al. 2017). For P–T modelling, garnet compositions (i) may record prograde growth, which is typically defined by bell-shaped Mn enrichment in the garnet core (e.g. Hollister 1966; Atherton 1968; Cygan and Lasaga 1985; Schwandt et al. 1996; Dziggel et al. 2009; Lanari and Engi 2017; Spear 2017; Dempster et al. 2020), or (ii) alternatively, at high temperatures, garnet chemistry may be used to determine the max T conditions experienced by the sample where the prograde growth major element zonation is partially or completely re-equilibrated (Caddick et al. 2010). Also, fluid and/or melt inclusions hosted in peritectic garnet from migmatites have been used to evaluate the starting composition of the anatectic melt and fluid regime during anatexis (e.g. Cesare et al. 2009, 2011; Ferrero et al. 2012, 2021; Carvalho et al. 2019; Borghini et al. 2020). Inclusions such as zircon and quartz in garnet can also be used to produce pressure estimates using Raman barometry (Zhong et al. 2019; Spear and Wolfe 2020). After being trapped in a host mineral during part of the cooling and exhumation history of a rock, inclusions may develop residual pressures due to the differences in thermal expansivity and compressibility between the host and the inclusions (Gonzalez et al. 2019; Zhong et al. 2019, 2020). Within both metamorphic and magmatic studies, however, garnet-based thermobarometry should always be applied with a thorough petrological and microstructural relationships characterization and an appropriate estimate of equilibrium volume of the rock sample. Results that are inconsistent with the observed mineral assemblage should be crosschecked with other comparable tools and should be interpreted with care. Additionally, taking a step back and carefully reanalysing field and textural context of the investigated rocks, as well as reassessing the documentation of sample preparation, analytical procedure and data processing, may also reveal missed information.

The mineral zircon (ZrSiO₄) has traditionally been the most robust accessory phase and widely used geochronological and geochemical tool in petrochronology to investigate crustal formation and evolution, and to untangle complex rock histories that most minerals fail to retain (e.g. Froude et al. 1983; Scherer et al. 2007; Guitreau et al. 2019). For instance, in the Archean Lewisian Gneiss Complex, NW Scotland, zircon is the only accessory mineral that has recorded the several Archean magmatic phases (cores) and the multiple ultra-high temperature (UHT) to HT/medium (M)T overprinting events (rims) (e.g. Whitehouse and Kemp 2010; Taylor et al. 2020; Fischer et al. 2021), whereas others such as monazite and titanite record only parts of the metamorphic events (e.g. Zhu and O'Nions 1999; Goodenough et al. 2013). Its resistance to erosive processes makes zircon a prime choice when studying detrital crystals to unravel past formation of continental crust in deeply eroded areas, or simply ancient terranes (e.g. Iizuka et al. 2005; Dhuime et al. 2012; Næraa et al. 2012; Nordsvan et al. 2018; Gardiner et al. 2019; Kirkland et al. 2021; Tedeschi et al. 2023; Pereira et al. this volume, in press). Over the years, many techniques have been improved to study zircon and retrieve various types of information from it (Table 1).

When zircon crystallizes, it can develop various textures depending on the conditions at which it forms, such as sharp concentric growth zoning and/or marginal resorption if zircon grew in a magmatic environment or more irregular domains of homogeneous zoning cutting discordantly the growth domains in the post-magmatic environment (e.g. Corfu et al. 2003 and references therein). These textures are commonly revealed by routine imaging of cathodoluminescence (CL) or backscattered electrons (BSE) using a scanning electron microscope (SEM). Cathodoluminescence response is mostly due to interaction between the primary electron-beam and zircon lattice through a complex energy transfer reaction that ends with the emission of photons by tetravalent REE, especially Dy (e.g. Nasdala et al. 2003). Visible textures range from broad to fine oscillatory zoning in igneous zircon (Fig. 2d–f; Fig. 4a, b) to large homogeneous to patchy domains in metamorphic conditions (e.g. Corfu et al. 2003). Raman spectroscopy can be used to quantitatively map radiation damage zoning in single zircon crystals (i.e. metamictization; Balan et al. 2001) and, hence, evaluate the effects of radiation damage on the final U/Pb ratio (Anderson et al. 2020) or on any elemental and/or isotopic signature. Igneous zircon most commonly form when the activity of ZrO₂ and SiO₂ is optimal. This is the case in most magmas with SiO₂ concentrations over 57 wt%, although rare crystals in magmas with lower SiO₂ concentrations do exist (e.g. Aranovich et al. 2017; Fischer et al. 2021; Bea et al. 2022). Zircon precipitation from melts was investigated early on by Watson (1979) and Watson and Harrison (1984), and revisited by Boehnke et al. (2013).

The nature of mineral inclusions can provide a first order evidence of phases that co-precipitated with zircon and, thus, give information about the nature/composition of the melt in which it crystallized and the pressure and temperature conditions at which it grew (Delavault et al. 2016; Emo et al. 2018; Antoine et al. 2020). The same can be done with melt inclusions, but these can also be used to estimate partition coefficients for trace elements if they represent bulk-melt composition and their chemistry remained unmodified since entrapment (Thomas et al. 2003; Gudelius et al. 2020).

Zircon incorporates large quantities of trace elements from ppm to thousands of ppm level (e.g. Belousova et al. 2001; Hoskin and Schaltegger 2003), making it an easy target for analysis using electron microprobe (EPMA), ICP-MS, LA-(MC-)ICP-MS (Košler et al. 2005; Chew et al. 2017; Petrus et al. 2017) and secondary ionization mass spectrometers or sensitive high-resolution ion microprobe (SIMS/SHRIMP). The rapid technological improvements using LA-(MC-)ICP-MS allow now the acquisition and extraction of quantitative data from a mapped area, including chemical and isotopic information of accessory phases such as zircon (Petrus et al. 2017; Chew et al. 2021). Experiments and natural observations indicate very slow diffusion for most elements within the zircon lattice (see Lee et al. 1997; Cherniak and Watson 2003 for values). Pristine igneous zircon elemental variability is usually relatively small, making this mineral a poorly sensitive tracer of melt composition, except for sharp variations (i.e. strongly alkaline, felsic and mafic; e.g. Belousova et al. 2002; Grimes et al. 2007; Guitreau et al. 2022). However, when zircon undergoes significant radiation damage, it can be sensitive to thermal events that result in incorporation of measurable quantities of light REEs (e.g. Hoskin and Schaltegger 2003; Bouvier et al. 2012; Pidgeon et al. 2017). This modification of zircon composition is a good proxy for secondary alteration and/or weathering. Along the same line, non-formula elements, such as Ca, K and Al, can enter zircon easily once it becomes porous due to significant radiation damage accumulation (e.g. Holland and Gottfried 1955; Ewing et al. 2003; McCubbin et al. 2016; Pidgeon et al. 2019). Monitoring LREEs and non-formula elements allows for filtering of pristine zircon crystals or domains from altered and weathered ones. In contrast, rare earth element patterns in metamorphic zircon are more useful because equilibrium with other phases has a forcing effect on partition coefficients such that zircon cannot incorporate as much HREE as it would normally do (Fig. 4c). This is well illustrated when zircon forms in equilibrium with garnet (e.g. Kelly and Harley 2005; Blereau 2017), resulting in flattened to steepened heavy REE patterns (Fig. 4c). Whilst extremely sluggish under most conditions of metamorphism (Cherniak and Watson 2003), under extreme metamorphic conditions (>1000°C), the REE content in zircon may also be modified as a function of temperature and time (Watson 1996; Blereau et al. 2022), although others have suggested decoupling of REE from U–Pb above 850°C (Kunz et al. 2018; Jiao et al. 2020b; Durgalakshmi et al. 2021). REEs within zircon are impurities that are expelled at high-T resulting in the removal of internal zoning (Hoskin and Black 2000), with garnet acting as a sink for the released REEs (see Discussion). Diffusional modelling of REE-in-zircon shows that an initially igneous zircon (i.e. an inherited grain with steep M-HREE patterns) within a garnet-bearing metapelite is in disequilibrium with the garnet, and when exposed to metamorphic temperatures, the zircon attempts to reach equilibrium with garnet (i.e. a c. 1:1 D_REE pattern) (Blereau et al. 2022; Fig. 4d). Zircon with a 50 µm radius can be re-equilibrated with the host metamorphic assemblage during both short (1100°C for 1–5 Ma) and extended periods (1050°C for 10–30 Ma or 1000°C for 200 Ma) of UHT metamorphism (Blereau et al. 2022; Fig. 4e). Conversely, unless diffusion is enhanced by fluids or other processes, below 900°C, zircon will largely preserve its pre-metamorphic REE signature, even when metamorphism is prolonged (>100 Ma) (Blereau et al. 2022; Fig. 4e). The change of valence of Ce and Eu to 4+ and 2+, respectively, in igneous zircon can be used to estimate redox conditions through Ce and Eu anomalies (e.g. Trail et al. 2012; Smythe and Brenan 2016; W.T. Li et al. 2021). However, some authors suggest using these proxies with caution because of melt cooling and chemical evolution (e.g. Loader et al. 2022). Europium anomalies can also be used to track feldspar fractionation and to estimate whether zircon crystallized in equilibrium with feldspar in igneous and metamorphic zircon, respectively (Rubatto 2002).

The ratio of Th over U is often used in zircon because it provides information about co-precipitating phases (e.g. Kunz et al. 2018). It can help discriminate igneous (Th/U c. 0.2–0.8) from metamorphic (Th/U < 0.1 or >1) zircon domains/crystals, distinguish between zircon formed under amphibolite-facies and granulite-facies and can trace low-temperature weathering of radiation-damaged zircon (e.g. Vavra et al. 1999; Hoskin and Black 2000; Kirkland et al. 2015; Yakymchuk et al. 2018; Guitreau et al. 2019; Guitreau and Flahaut 2019; Barrote et al. 2020). However, Th/U in zircon is not always a faithful recorder of metamorphic processes (Möller et al. 2003; Harley and Kelly 2007), and it should be used with caution on a case-by-case basis. Typically, U–Pb LA-ICP-MS geochronology routines do not include appropriate internal standard elements, therefore, the U, Th and Pb concentrations are semi-quantitative only.

Zircon is most commonly dated using the U–Pb isotope system (e.g. Schoene 2014) as, for this systematics, it has several advantages compared to other minerals. Zircon U–Pb isotope measurements are done using a great variety of techniques from solution-based such as TIMS for the most precise ages and MC-ICP-MS/ICP-MS to in situ microbeam techniques (i.e. SIMS, SHRIMP, LA-ICP-MS, LA-MC-ICP-MS). One of the main advantages is that Pb is essentially excluded from the zircon lattice because of its size and valence, resulting in virtually all measurable Pb in zircon being radiogenic – produced by the radioactive decay of U and Th isotopes. This means that the parent/daughter ratio measured in pristine zircon combined with U decay constants can be directly converted into an age (Schoene 2014). However, zircon lattice can become damaged by radioactive decay over time, which results in metamictization. This process makes zircon porous to external agents that may incorporate Pb, and other non-formula elements, into the crystal lattice, thus compromising the determined age. Another major benefit of the U–Pb isotope system compared to others is the fact that it contains two isotope systems (i.e. ²³⁸U–²⁰⁶Pb and ²³⁵U–²⁰⁷Pb) with distinct decay constants (e.g. Le Roux and Glendenin 1963; Jaffey et al. 1971), allowing open- v. closed-system evolution to be assessed. For easy visualization, U–Pb data are commonly plotted in ²³⁸U–²⁰⁶Pb and ²³⁵U–²⁰⁷Pb Wetherill concordia diagram (Fig. 4f; Wetherill 1956). When both dates lie on the concordia curve, the date is called concordant and is geologically meaningful since it likely reflects a closed U–Pb system evolution. By contrast, when both dates are different, the measured date is discordant and evaluation on the geological meaning will vary case by case. A third date can also be derived directly from the ²⁰⁷Pb/²⁰⁶Pb ratio that inherently assumes closed-system evolution. This date is the oldest of the three in the case of ‘normally’ discordant data (i.e. datapoint is located below the concordia curve). Unless information regarding the validity of intercepts in concordia diagram is available (e.g. multiple analyses from the same crystal or growth zones), ²⁰⁷Pb/²⁰⁶Pb ages are commonly more precise for zircons >1.5 Ga, whereas ²³⁸U–²⁰⁶Pb ages are used for crystals <1.5 Ga (Spencer et al. 2016).

Depending on when the U–Pb isotope system disturbance occurred, the memory of primary crystallization may or may not be preserved. If U–Pb disturbance is recent, U–Pb data are discordant and distribute along a discordia line that passes through zero (blue points). Since Pb isotopes are not fractionated in detectable proportions in such disturbance, ²⁰⁷Pb/²⁰⁶Pb ratios still provide the primary crystallization age, which corresponds to the discordia upper-intercept (Fig. 4f) and is identical to the purple-filled ellipse which represents a concordant analysis from either the same zircon or the same population that remained unmodified. In contrast, for an old U–Pb disturbance, the datapoints would align along a discordia line that connects a lower-intercept corresponding to the age of U–Pb disturbance and an upper-intercept representing the actual crystallization age of zircon (red-filled ellipses in Fig. 4f). Most cases of discordance correspond to Pb-loss, which graphically corresponds to a migration of datapoints towards the lower end of a discordia line, and less common cases are associated with reverse discordance which reflects Pb accumulation (e.g. Williams et al. 1984; Kusiak et al. 2013). The case in which zircon contains measurable amounts of common-Pb (Pb_c) has not been presented here because, in most cases, it is a sign of zircon post-crystallization modification and/or advanced alteration. Consequently, data are generally discarded when ²⁰⁴Pb is detected, since the Pb_c correction relies on knowledge of the isotopic composition of Pb when it entered the crystal, which is difficult to know in zircon.

The source of zircon parental magma can be assessed using multiple isotope systems, with the most common being Lu–Hf and O isotopes (e.g. Patchett 1983; Valley 2003). Lu–Hf isotopes in zircon are normally measured using MC-ICP-MS, either in solution or laser-ablation mode (e.g. Fisher et al. 2014a, b). The Lu–Hf isotope system is radiogenic and based on the decay of radioactive ¹⁷⁶Lu into ¹⁷⁶Hf with a half-life of c. 36 Ga (Scherer et al. 2001; Söderlund et al. 2004). The measured ¹⁷⁶Hf/¹⁷⁷Hf tracks the time-integrated fractionation of Lu from Hf in a reservoir (Fig. 4g). The principle of this technique is that during partial melting of most mantle and crustal lithologies, Hf and Lu are fractionated from each other, which results in magmas having Lu/Hf lower than that of the melting residue. Over time, melting residues (refractory mantle) develop very radiogenic (elevated) ¹⁷⁶Hf/¹⁷⁷Hf ratios (positive ε_Hf), and crustal lithologies (former magmas) comparatively low radiogenic ratios (negative ε_Hf) (Fig. 4g). For global interpretation, Hf isotope compositions are normalized to the Chondritic Uniform Reservoir (CHUR; Blichert-Toft and Albarède 1997; Bouvier et al. 2008; Iizuka et al. 2015) and transformed into epsilon Hf notation (ε_Hf), with CHUR approximating the bulk silicate Earth composition. For instance, Figure 4g illustrates a magma source produced by partial melting of the mantle at 2600 Ma (green spot in Fig. 4g) and that differentiates into mafic and felsic lithologies, which are both characterized by specific ¹⁷⁶Lu/¹⁷⁷Hf content. These lithologies evolve after crystallization until 1500 Ma, when they are reworked by partial melting forming different granites (blue and red spots in Fig. 4g). Some of the generated melts mix with each other forming hybrid granites (grey spot in Fig. 4g). At 600 Ma, magmas extracted from the mantle (green spot in Fig. 4g) mix with crustal melts forming new hybrid granites (orange spot in Fig. 4g). Note that all crustal lithologies contain zircon, which allows the evolution of these reservoirs to be followed through time. Newly formed zircon domains (grains or overgrowths) within a single rock sample mostly have higher initial ¹⁷⁶Hf/¹⁷⁷Hf than older domains due to incomplete dissolution of detrital or magmatic zircon grains (hosting most of the non-radiogenic Hf). Also, metamorphic zircon incorporates additional radiogenic ¹⁷⁶Hf formed by ¹⁷⁶Lu decay in the rock's matrix between successive zircon growth events (Gerdes and Zeh 2009).

Oxygen isotopes are stable isotopes and the measured ¹⁸O/¹⁶O, and possibly ¹⁷O/¹⁶O when the triple-isotope system is used, tracks isotope fractionations due to magmatic processes and/or sources. These light isotopes do not fractionate much during magmatic processes (i.e. partial melting and fractional crystallization), but differences can be measured (e.g. Valley 2003; Bindeman 2008). Most O isotope studies in zircon use the source-tracing potential of O isotopes, taking advantage of large O isotope fractionation caused by low-temperature alteration of crustal lithologies inducing clay formation and resulting in enrichment of ¹⁸O relative to ¹⁶O. Much like the Lu–Hf isotope system, O isotope ratios are difficult to interpret as numerical values. Therefore, O isotope ratios are normalized relative to the international Vienna standard mean ocean water (VSMOW) and reported in the standard delta (δ¹⁸O) notation. Zircon formed from a reworked clay-rich crustal lithology may have δ¹⁸O up to +12‰ (Valley et al. 2005; Kemp et al. 2007), and mantle zircons have consistent δ¹⁸O of +5.3 ± 0.6‰ (Valley 2003). Zircon δ¹⁸O values below that of the mantle indicate that zircon parental melt interacted with meteoric fluids and/or was altered at high temperatures (Valley 2003 and references therein). Most studies using O isotopes in zircon interpret O isotope variations as evidence for mixtures between mantle- and crustal-derived melts and/or fluids or crustal contaminants (e.g. Kemp et al. 2007; Smithies et al. 2021). Recent studies have demonstrated that the uptake of water into the zircon crystal lattice can significantly modify its oxygen isotopic composition (Pidgeon et al. 2017; Liebmann et al. 2021). Therefore, monitoring of the ¹⁶O¹H/¹⁶O ratio during SIMS oxygen isotope measurements to assess secondary modification of O isotope composition by water addition is recommended (Liebmann et al. 2021).

Recently, new stable isotope systems such as Si and Zr have been applied to zircon. Different techniques such as SIMS, MC-ICP-MS and LA-MC-ICP-MS have been employed to measure Si isotopes in zircon, which have resulted in good precision and closely mimic the natural variability of high-temperature processes (Trail et al. 2018, 2019; Guitreau et al. 2020, 2022). Si isotopes can be used to trace the origin of zircon parental magma due to the various silicon isotope signatures found in different types of igneous rocks (e.g. I, A, S, tonalite–trondhjemite–granodiorite (TTG); Savage et al. 2014; Deng et al. 2019). Si isotope compositions, expressed as permil deviations from a quartz standard (δ²⁹Si and δ³⁰Si), are sensitive to the SiO₂ content, which reflects the degree of polymerization and crystallization temperature (Qin et al. 2016; Trail et al. 2019; Guitreau et al. 2022). This allows for reconstructions of the magma evolution. Changes in Si isotope compositions can occur due to metamorphic alteration and/or zircon recrystallization, which are dependent on the metamorphic grade (Guitreau et al. 2022).

Zr isotopes can be measured using conventional MC-ICP-MS instruments in solution or laser mode and applied to track fractional crystallization processes (Ibañez-Mejia and Tissot 2019; Zhang et al. 2019; Guo et al. 2020; Tian et al. 2020). Recent studies attributed measurable Zr isotope variations to kinetic fractionation in response to chemical gradients rather than equilibrium processes (Chen et al. 2020; Méheut et al. 2021). Zirconium isotope compositions of metamorphic zircon compared to igneous zircon may also result from chemical gradient effects (Zhang et al. 2019). This technique can hence provide insights into magmatic crystallization dynamics.

Zircon is used as a mineral pair thermometer based on incorporation of Ti into zircon at HT coupled with Zr substitution in rutile (Zack et al. 2004a; Watson and Harrison 2005; Watson et al. 2006; Ferry and Watson 2007). This technique requires the presence of both zircon and rutile within the mineral paragenesis, otherwise all temperatures are minimum estimates (Watson and Harrison 2005). Magmatic and metamorphic temperatures can also be overestimated at low pressures (<5 kbar) (Rubatto 2017). Temperature variations within an investigated sample have been interpreted to reflect waves of magmatic pulses within a magma chamber generating thermal and compositional heterogeneities (e.g. Collins et al. 2016; Volante et al. 2020a). Low diffusivity of Ti within zircon was previously measured perpendicular to the crystallographic c-axis (Cherniak and Watson 2007), making this thermometry a widely used tool. In contrast, recent experiments conducted parallel to the c-axis found significant anisotropy in the diffusivity of Ti (Bloch et al. 2022). When extrapolated, the resulting diffusivities are c. 7.5–11 times faster at 950–650°C than the original experiments, indicating that this thermometer can be modified under elevated crustal temperatures, slow cooling and/or small grain sizes (Bloch et al. 2022). Moreover, since this thermometer is dependent upon alpha-quartz and t-TiO₂ activities, software such as Rhyolite-MELTS Gualda et al. (2012) have been proposed to improve temperature estimates’ precision (e.g. Schiller and Finger 2019). The latter tool is useful and intuitive to use, however its caveats (Volante et al. 2020a) and its relatively low precision of about 50°C (Guitreau et al. (2022)) should be considered.

Titanite (Ca[Ti,Al,Fe³⁺]SiO₄[O,OH,F,Cl]) (Fig. 2g-i; Fig. 5) can occur as minor or accessory mineral and it is a particularly efficient chemical reactant with other major mineral phases that contain Ca and Ti, commonly leading to re-crystallization (Frost et al. 2001). Titanite commonly grows during magmatic (e.g. Ca-rich granitic rocks), metamorphic (e.g. calc-silicate, amphibolite) and hydrothermal events (Fig. 5a-c), with each different setting resulting in different titanite REE compositions that can be tracked by trace element fingerprinting (see reviews by Frost et al. 2001; Kohn 2017). Titanite commonly forms in mafic and calc-silicate rocks, and its growth can provide valuable time constraints in cases where other accessory minerals do not develop. Titanite crystallizes over a broad P–T range, including growth after breakdown of rutile, which is commonly stable at higher pressures (>1.4 GPa), ilmenite, Fe–Ti oxide stable at higher temperatures (>650/700°C), magnetite and/or clinozoisite (Frost et al. 2001; Kohn 2017). The combined acquisition of geochronological and geochemical data in texturally controlled titanite has robustly discriminated between distinct metamorphic, hydrothermal and magmatic events, and provided the opportunity to evaluate physical and chemical processes at the micro-scale in titanite-bearing rocks (e.g. Stearns et al. 2015; Garber et al. 2017; Olierook et al. 2019; Cavosie et al. 2022; Walters et al. 2022). However, petrological and chemical complexities can challenge U–Pb and P–T data interpretation. For instance, overgrowths of multiple metamorphic and hydrothermal titanite generations can significantly affect and modify the U–Pb system (e.g. Storey et al. 2006; Marsh and Smye 2017; Kirkland et al. 2018) and trace element compositions (e.g. Gordon et al. 2021), inducing decoupling of titanite U–Pb dates and trace element compositions (e.g. Romer and Rötzler 2001, 2011; Castelli and Rubatto 2002; Bonamici and Blum 2020; Walters et al. 2022).

Texturally, both magmatic and metamorphic (or hydrothermally altered) titanite can exhibit complex compositional zoning (Fig. 5a–c) and associated trace element (TE) patterns related to fluid–rock interaction (e.g. Smith et al. 2009; Garber et al. 2017; Olierook et al. 2019; Walter et al. 2021; Walters et al. 2022). Titanite preferentially incorporates minor and TE including U, high field-strength elements (HFSEs) such as Zr and REEs (e.g. Tiepolo et al. 2002; Lucassen et al. 2012), compared to other accessories (e.g. zircon, monazite). Trace element in titanite can provide information about crystallization pressure and temperature estimates (Hayden et al. 2008), oxygen fugacity (e.g. King et al. 2013; Cao et al. 2015) and fractionation processes (e.g. Piccoli et al. 2000; John et al. 2011). Geochemical and experimental studies on titanite suggest that it preferentially incorporates the medium rare earth element (MREE) in its crystal structure when in equilibrium with melt (e.g. Tiepolo et al. 2002; Prowatke and Klemme 2005; Olin and Wolff 2012), whereas recent petrochronological works show that melt-present and melt-absent metamorphic titanite have different REE patterns strongly depending on the presence and/or abundance of other phases in equilibrium with them (e.g. Garber et al. 2017; Walter et al. 2021; Walters et al. 2022). For instance, different TE uptake in titanite allowed Garber et al. (2017) to discriminate between (i) LREE-enriched (Precambrian) igneous cores, (ii) LREE-depleted and HREE-rich (Caledonian) recrystallized metamorphic rims likely associated with fluid or melt and (iii) (Caledonian) neocrystallized metamorphic titanite reflecting a negative or more positive REE slope based on whether titanite grew in equilibrium with hornblende and garnet or with allanite, apatite, plagioclase and biotite, respectively, during prograde and/or retrograde amphibolite-facies metamorphism (Fig. 5d, e; see also Cioffi et al. 2019; Walter et al. 2021).

Recent studies have highlighted the potential of TE in titanite as petrogenetic discriminator/indicator in magmatic, metamorphic and detrital studies (e.g. Ma et al. 2019; Olierook et al. 2019; Scibiorski and Cawood 2022). For example, Olierook et al. (2019) show that Fe concentrations and Th/U and Th/Pb ratios are systematically higher in magmatic than metamorphic titanite, reflecting useful petrogenetic discriminators when plotted against Al and Zr-in-titanite temperature (T°C), respectively (Fig. 5f–h). Also, normalized LREE/MREE, LREE/HREE and Eu anomalies reflect negative and/or positive REE slope correlations, which are useful to complement the characterization of titanite origin (Olierook et al. 2019). TE-based discrimination can be corroborated with detailed textural investigation of titanite grains, which are commonly characterized by oscillatory or sector zoning when magmatic in origin, with patchy and/or homogeneous zoning in metamorphic grains (e.g. Smith et al. 2009; Garber et al. 2017; Walters and Kohn 2017; Cioffi et al. 2019; Holder and Hacker 2019). Scibiorski and Cawood (2022) show that different titanite host-rock lithologies are reflected in the variation of TE chemistry, with low Zr/Y and high Fe in titanite from felsic host-rocks. This complements Al/Fe v. ΣLREE and U v. Th contents (ppm) used to discriminate magmatic v. metamorphic titanite (Olierook et al. 2019; Scibiorski and Cawood 2022; Fig. 5i) and between metamorphic, recrystallized and inherited igneous titanite (fig. 2 in Scibiorski and Cawood 2022), respectively. Also, systematic integration of titanite microstructural investigations by electron backscatter diffraction (EBSD) with in situ U–Pb petrochronology was proved to be a powerful tool not only to untangle and reconstruct complex deformation histories within crustal-scale high-strain zones but also to constrain deformation mechanism associated with shock and thermal metamorphism (Papapavlou et al. (2017); McGregor et al. 2021).

Titanite is a powerful U–Pb petrochronometer recording different primary and secondary geological processes (Spandler et al. 2016; Ma et al. 2019; Fisher et al. 2020; Barla 2021). However, titanite and other accessories including rutile, apatite and allanite tend to incorporate significant Pb_c (e.g. Kirkland et al. 2017, 2018; Bonamici and Blum 2020), which is reflected in discordant U–Pb ratios (e.g. Marsh and Smye 2017), making U–Pb dates interpretation of these accessory phases challenging (e.g. Olierook et al. 2019; Walters et al. 2022). Figure 5j (modified after Olierook et al. 2019) represents a schematic TW diagram (Tera and Wasserburg 1972) illustrating how titanite U–Pb data (and the other phases that commonly have Pb_c), including Pb_c (see Storey et al. 2006; Kirkland et al. 2017), crystallization and recrystallization mixing lines can be interpreted (e.g. Spencer et al. 2013; Bonamici et al. 2015; Garber et al. 2017; Papapavlou et al. 2017; Holder and Hacker 2019; Mottram et al. 2019; Timms et al. 2020; Gordon et al. 2021). As a result of Pb_c incorporation in the crystals, geochronological data typically form an intercept line (red line in Fig. 5j, ‘igneous titanite ages’) characterized by a ²⁰⁷Pb/²⁰⁶Pb upper intercept and lower concordia intercept (e.g. Spencer et al. 2013; Chew et al. 2014; Garber et al. 2017; Kirkland et al. 2017, 2018, 2020), with the igneous age being calculated using the lower one. However, when such intercept line is not statistically robust, then (i) a weighted mean of uncorrected dates can be calculated, if they are within 2SD uncertainty (e.g. Spencer et al. 2016; Olierook et al. 2019; Barrote et al. 2022b), or (ii) the upper intercept can be used to calculate concordant analyses with negligible Pb_c (Olierook et al. 2019). U–Pb dates of samples affected by overprinting events would fall in a space in the TW diagram defined by a ²⁰⁷Pb/²⁰⁶Pb upper intercept, a first lower concordia intercept recording the first magmatic/metamorphic event and a second lower concordia intercept recording the subsequent event (green triangular shape in Fig. 5j). Titanite is also found to yield non-typical Pb_c values, possibly due to inheritance of radiogenic Pb into the crystal structure (Kirkland et al. 2018; Mottram et al. 2019; Walters et al. 2022).

Additionally, an important factor to consider when dealing with titanite geochronology is Tc (Fig. 1; Dodson 1973). In the past two decades, it has been shown that titanite is much more retentive than previously envisaged (with Tc of c. 600–650°C, Mezger et al. 1991; 600–650°C, Scott and St-Onge 1995, Frost et al. 2001; 650 and 750°C, Cherniak 1993), increasing the temperature threshold for Pb and Zr volume diffusion in titanite as high as c. 750–840°C (c. 740°C, e.g. Schärer et al. 1994, Zhang and Schärer 1996; >750°C, e.g. Kylander-Clark et al. 2008, Kohn and Corrie 2011, Gao et al. 2012, Spencer et al. 2013, Stearns et al. 2015, Kohn 2017; >830°C, e.g. Hartnady et al. 2019, Holder et al. 2019, Kirkland et al. 2020) and challenging the assumption that titanite U–Pb dates commonly reflect cooling ages (Hartnady et al. 2019; Kirkland et al. 2020).

While less explored compared to other accessory minerals such as monazite, Sm–Nd isotope systematics in titanite (e.g. Yang et al. 2008; Amelin 2009; Fisher et al. 2011, 2020; Hammerli et al. 2014; Spandler et al. 2016; Ma et al. 2019; Zhang et al. 2021) has potential as a source tracer for understanding the formation and evolution of the crust (Amelin 2009; Fisher et al. 2020; Zhang et al. 2021). This isotopic system has been demonstrated to have high Tc (850–950°C; Cherniak 1995), surviving HT magmatic, metamorphic and hydrothermal conditions. In situ Sm–Nd isotopic systematics of titanite can be used to trace fluid, melt or crustal/juvenile rock sources (e.g. Lucassen et al. 2011; Hammerli et al. 2014; Spandler et al. 2016; Zhang et al. 2021).

Experimental studies have demonstrated that oxygen diffusion in accessory minerals is slower than in rock-forming minerals (e.g. Fortier and Giletti 1989), but oxygen diffusion in apatite is faster than in titanite, which is in turn faster than in zircon (Bruand et al. 2019). Oxygen isotopes in magmatic and metamorphic titanite have been used as a geochemical indicator to investigate magma petrogenesis as well as metamorphic and hydrothermal overprinting events (e.g. King et al. 2001; Bonamici et al. 2011, 2014, 2015; Bruand et al. 2019). For example, Bonamici et al. (2014) differentiated four generations of titanite with distinct δ¹⁸O values and internal textural zoning. Also, consistent results between δ¹⁸O in titanite and zircon indicate that titanite is a robust accessory mineral preserving the original magmatic δ¹⁸O composition (Bruand et al. 2019).

A thermometer using the Zr content in titanite (Zr-in-titanite) was developed by Hayden et al. (2008), relying on the direct substitution of Zr⁴⁺ for Ti⁴⁺, whereas pressures can be estimated from the net transfer reaction ₂Ca₂Al₃Si₃O₁₂(OH) + TiO₂ + SiO₂ = 3CaAl₂Si₂O₈ + CaTiSiO₅ + H₂O (referred to as TZARS; Kapp et al. 2009), using automated calculations in THERMOCALC (Holland and Powell 2011). The Zr-in-titanite thermometer covers a wide temperature range and is pressure dependent, resulting in large pressure uncertainties and rutile activity in rutile-absent rocks, limiting the reliability of temperature estimates. Therefore, modelling titanite within a mineral paragenesis using software such as THERMOCALC allows for pressure estimates with errors less than c. 0.1 GPa (Kohn et al. 2017).

Rutile (Fig. 2j-l; Fig. 6) is the high-temperature TiO₂ polymorph (Dachille et al. 1968) and a common accessory mineral that occurs in metamorphic (U) high-pressure (UHP) mafic rocks (Fig. 6a; e.g. Zack et al. 2004b; Triebold et al. 2012; Zack and Kooijman 2017; Böhnke et al. 2019), in moderate-pressures and moderate-temepratures to HP and HT metapelitic rocks (Fig. 6b; e.g. Hart et al. 2018; Gonçalves et al. 2019) and in low-temperature/hydrothermal metapelitic rocks (Fig. 6c; e.g. Plavsa et al. 2018; Salama et al. 2018; Agangi et al. 2019; Porter et al. 2020; Schirra and Laurent 2021). In contrast, rutile occurrence in magmatic rocks is limited to HT and dry alkaline, kimberlite or pegmatitic rocks (e.g. Carruzzo et al. 2006; Cerny et al. 2007). At low-T (LT) and low-P (LP) conditions, rutile is no longer stable, and it is replaced by polymorphs anatase (tetragonal) and brookite (orthorhombic), respectively, whereas rutile HP polymorph is TiO₂(II) (Dachille et al. 1968; Jamieson and Olinger 1969). During prograde metamorphism, rutile growth is attributed to breakdown of ilmenite, titanite and biotite (Zack and Kooijman 2017). Rutile popularity as a petrochronometer has exponentially increased due to its multifaceted potential to investigate tectonic processes as single-mineral thermometer (e.g. Zack et al. 2004a; Watson et al. 2006; Tomkins et al. 2007; Luvizotto and Zack 2009), geochronometer (e.g. Mezger et al. 1989a), geospeedometer (Cruz-Uribe et al. 2014; Kohn et al. 2016), as a provenance indicator (e.g. Triebold et al. 2012; Pereira and Storey 2023) and as an isotopic tracer (e.g. Ewing et al. 2011). Additionally, the increasing sensitivity of analytical equipment has enabled dating of low-U phases such as rutile (Luvizotto et al. 2009; Axelsson et al. 2018; Verberne et al. 2019; Moore et al. 2020b).

Rutile primarily incorporates high field strength elements (HFSE; i.e. Nb, Ta, Zr, Hf and Cr; Rudnick et al. 2000; Zack et al. 2002; Schmidt et al. 2009). These can be used (i) to characterize micro-scale processes associated with growth of rutile crystals (e.g. Hart et al. 2018; Verberne et al. 2022a); (ii) as a pressure-proxy, such that trace element budget (i.e. Na and Ta) in the melt rutile crystallized in reflects the depth at which partial melting of the source rock occurred (e.g. Foley et al. 2002; Moyen and Stevens 2006; Meyer et al. 2011; Kooijman et al. 2012); or (iii) as geochemical pathfinders for mineralized rocks (Clark and Williams-Jones 2004; Smythe et al. 2008; Pochon et al. 2017; Plavsa et al. 2018; Agangi et al. 2019; Ballouard et al. 2020; Porter et al. 2020; Sciuba and Beaudoin 2021). While HFSE are compatible in rutile, other trace elements including Sr, Th and REEs are incompatible (Klemme et al. 2005; Meyer et al. 2011). Trace element concentrations between the three TiO₂ polymorphs (rutile, anatase and brookite) systematically differ, leading to erroneous results when applying the Zr-in-rutile thermometer or Cr and Nb discrimination diagrams (see below) to phases other than rutile (Triebold et al. 2012).

Lithology discrimination schemes from Triebold et al. (2012) and Meinhold et al. (2008) using Cr and Nb (ppm) content in rutile as well as Zr/Hf and Nb/Ta ratios are useful to identify the protolith composition (metapelitic v. metabasic). However, care must be taken when applying this technique for source discrimination in provenance studies due to documented unsystematic Nb/Cr ratios for rutile from amphibolite-facies rocks potentially reflecting TE disturbance during retrogression or prolonged HT metamorphism (Meyer et al. 2011; Kooijman et al. 2012). Additionally, the application of principal component analysis (PCA) on rutile from Precambrian UHT and Phanerozoic HP terranes (Hart et al. 2018), as well as on ore-bearing and barren metamorphic and magmatic rocks (Plavsa et al. 2018; Pereira et al. 2019; Porter et al. 2020) indicates that several TE can be used for lithological discrimination (van Schijndel et al. 2021) and to distinguish rutile from mineralized v. barren rocks. For example, the Nb–V–Cr ternary diagram shows that rutile from HP metamafic rocks is commonly V-rich, whereas in metapelites is more Nb-rich (Fig. 6d; Hart et al. 2018). When associated with mineralized systems, rutile contains anomalous concentrations of V, Sn, Sb, W, Ni, Cu, Cr, Ta, Nb and Fe (e.g. Scott and Radford 2007; Plavsa et al. 2018; Porter et al. 2020), where W and Cr variability is used to discriminate its origin from mineralized or barren rocks (Porter et al. 2020). It is also possible to identify rutile deriving from pegmatitic rocks (high Nb, Ta, Sn) and Au-ore, which are enriched in Sb. This can also be done by combining multi-element clustering of PCA analysis (Porter et al. 2020; Sciuba and Beaudoin 2021). However, due to the systematic differences in trace element contents between the three TiO₂ polymorphs, a detailed characterization of the analysed grains is recommended using either EBSD or Raman spectroscopy (Plavsa et al. 2018; Porter et al. 2020; Sciuba and Beaudoin 2021). Multi-variant statistical analysis of rutile TE composition, particularly variations in Sc, REE, Y, Ca, Ba, Th, Zr, U and V, can reflect different metamorphic grades experienced by the country rock (Sciuba and Beaudoin 2021). For example, mafic and ultramafic, lower to middle greenschist-facies country rocks have lower Zr/Th and U/La ratios than intermediate and sedimentary greenschist facies one (Fig. 6e; Sciuba and Beaudoin 2021). A similar trend is observed for relative concentrations of Sc/Y and Zr/Ba, where lower and higher contents reflect lower and higher metamorphic grade, respectively (Fig. 6f; Sciuba and Beaudoin 2021). Also, the identification of localized TE enrichment along twin interface in rutile grains using atom probe tomography is interpreted to occur via volume diffusion during HT metamorphism (Verberne et al. 2022a).

Vry and Baker (2006) calculated the Tc of Pb diffusion in rutile to be between 500 and 540°C, based on natural samples (Fig. 1). For a spherical rutile of 200 µm, experimental data from Cherniak (2000a) predict whole-grain Tc of c. 600°C at an average cooling rate of 2–3°C/Ma, and above c. 640°C for Zr in rutile (2°C/Ma cooling rate; Cherniak et al. 2007; Dohmen et al. 2019). This moderate Tc for rutile implies that it is possible to find crystallization ages in cores of larger rutile crystals, especially if cooling is very fast. The preservation of a crystallization/growth age depends on several factors including max metamorphic temperatures, cooling rates and grain size (Kylander-Clark et al. 2008; Zack and Kooijman 2017; Moore et al. 2020b), and a multi-proxy approach may be needed to distinguish between different metamorphic and deformation stages. For example, Moore et al. (2020b) used rutile Zr thermometry, SHRIMP U–Pb age determination and electron backscatter diffraction (EBSD) microstructural analyses to identified two-stage rutile age populations which were not distinguishable using LA-ICP-MS data alone (Fig. 6g–i). Due to low-U content in rutile (<0.1 ppm; see Zack et al. 2011), it is important to first determine the U content to obtain meaningful metamorphic and/or magmatic ages (Zack et al. 2002, 2004b, 2011). Low-U content can lead to a high common v. radiogenic Pb ratio; therefore, monitoring the ²⁰⁶Pb/²⁰⁸Pb ratio and application of the ²⁰⁸Pb correction method is recommended (Zack et al. 2011). An application of this method allows the investigation of age variations within a single rutile grain, where transects from core to rim give younger ages towards the rim as a result of Pb diffusion during cooling and provide a temperature–time trajectory (Kooijman et al. 2010). In contrast to rutile geochronology obtained with LASS-ICP-MS, recent atom probe investigations indicate evidence for heterogeneous Pb and trace element distribution at the nanoscale (Verberne et al. 2020). Nonetheless, at the microscale (>20 µm), TE variations in a rutile single grain are negligible and concordant U–Pb dating is obtained, indicating that nanoscale defects do not significantly impact the micro-scale analysis (Verberne et al. 2020).

In situ analysis of Hf isotopes in rutile (Sláma et al. 2007; Ewing et al. 2011, 2014) has been applied to trace metasomatic processes in the lithospheric mantle (Choukroun et al. 2005; Aulbach et al. 2008) and recycling of continental material in the mantle (Ewing and Müntener 2018). Despite the relatively low Hf content (<300 ppm Hf), the ¹⁷⁶Hf/¹⁷⁷Hf of rutile can be accurately measured in situ by LA-MC-ICP-MS, provided care is taken with the analytical protocol and data reduction process (Ewing et al. 2011; Ewing and Müntener 2018). Matrix matched standards and a ¹⁷⁶Hf signal intensity above 10 mV are necessary, requiring a large spot size of >160 µm (Yang et al. 2015; Ewing and Müntener 2018). This technique is particularly interesting for rocks that lack zircon but contain rutile, such as mafic lithologies, with low Hf concentrations (Ewing et al. 2011). In contrast, rutile from (U)HT felsic granulites can contain much higher Hf contents, ranging from 20 to 400 ppm (Ewing et al. 2013).

The solubility of ZrO₂ in rutile is strongly temperature-dependent, and Zr-in-rutile has been identified as a useful thermometer (ZiR) when the rutile coexists with the appropriate buffer assemblage, i.e. zircon + quartz (Zack et al. 2004a; Watson et al. 2006; Tomkins et al. 2007; Hofmann et al. 2013). Underestimation of the calculated temperatures occurs when the rutile grows in the absence of zircon and/or in partially reset mineral assemblages (Zack et al. 2004a; Harley 2008). Possible biases include micro-inclusions of zircon (Zack et al. 2004a), prograde relict grains or incomplete equilibration with quartz (high Zr-rutile) or zircon (low-Zr rutile) due to slow diffusion along grain boundaries (see discussions in Taylor-Jones and Powell 2015; Kohn et al. 2016; Kohn 2020). Decoupling between Ti-in-zircon and Zr-in-rutile thermometry during UHT metamorphism is recorded in rutile that occurs as inclusions in zircon (Lei et al. 2020). Diffusion of Zr within rutile and Zr loss are closely related to the distribution of Zr, duration of UHT metamorphism and rutile grain size (Dohmen et al. 2019; Lei et al. 2020). The Zr concentration in rutile (Zr-in-rutile) is temperature-sensitive over a large range of geologically significant temperatures (e.g. Zack et al. 2004a; Ewing et al. 2013; Wawrzenitz et al. 2015; Pape et al. 2016; Böhnke et al. 2019; Clark et al. 2019; Moore et al. 2020b; Adlakha and Hattori 2021; Campomenosi et al. 2021; Bonnet et al. 2022). This relationship has been experimentally calibrated by Watson et al. (2006) and Ferry and Watson (2007), and a significant pressure effect has been calibrated by Tomkins et al. (2007). The Zr-in-rutile thermometer was recently refined and now predicts temperatures up to 40°C lower for T ≤550°C, and systematically higher temperatures for T >800°C (Kohn 2020). With the new calibrations, precisions of ±5°C and accuracy of c. ±15°C may be possible, although a variable rutile composition may lead to larger uncertainties (Kohn 2020). Taylor-Jones and Powell (2015) showed that Zr can leave rutile and move along grain boundaries towards existing distal zircon.

Zr-in-rutile temperatures can record HT events, whereas U–Pb in rutile records cooling ages due to low Tc (Fig. 1). Rutile U–Pb ages likely postdate Zr temperatures following high-grade metamorphism and subsequent simple cooling, although a more complex history of episodic cooling and reheating may lead to more significant decoupling between Zr temperatures and U–Pb ages (e.g. Ewing et al. 2015). Bonnet et al. (2022) show the combined use of Zr-in-rutile and U–Pb ages for rutile occurrence in subduction complexes that may be interpreted as crystallization ages for the units that experienced high-pressure, low-temperature metamorphism, but not for the high-grade units (Fig. 6j–l).

The term ‘allanite’ has been used to designate minerals from the allanite subgroup of the epidote group (Armbruster et al. 2006; Mills et al. 2009), which occurs in magmatic (Fig. 2m-o; Fig. 7a), metamorphic (from greenschist- to granulite-facies) and hydrothermal rocks (Janots et al. 2006, 2009; Rubatto et al. 2011; Airaghi et al. 2019). The allanite subgroup comprises a series of REE minerals with an ideal structural formula of Ca(LREE³⁺)(Al)₂ (Fe²⁺, Fe³⁺)(SiO₄)(Si₂O₇)O(OH). The presence of LREEs as major constituents, the various element substitutions, the broad P and T stability field and the preservation of growth stages, make allanite one of the burgeoning protagonists in petrochronology (e.g. Rubatto et al. 2011; Manzotti et al. 2018; Airaghi et al. 2019). Compositionally, two main factors are crucial for its use in petrochronology: (i) REEs’ sites can be occupied by Th⁴⁺ and U⁴⁺, pivotal elements for geochronology (Gieré and Sorensen 2004), and (ii) REEs are exceptional tracers of geological processes (Hermann 2002; Engi 2017). Allanite often exhibits complex chemical zoning (Fig. 7a; Romer and Xiao 2005; Rubatto et al. 2011; Airaghi et al. 2019), requiring investigation via in situ methods (e.g. Burn 2016; Zhang et al. 2022).

Trace element composition of allanite depends on the interplay between the bulk rock composition and fractionation of these elements during mineral reactions. For instance, a decrease in the LREE in magmatic allanite can be observed in Ca-rich rocks → diorite and granodiorite → granite → syenite (Smye et al. 2014; Engi 2017). The Th/U ratios can be often used as a complementary tool (together with initial Pb_c lead values) to distinguish between magmatic (>100) and metamorphic (<50) grains or domains (Fig. 7b; Gregory et al. 2007, 2012; Di Rosa et al. 2020). REEs have also been used to discriminate distinct allanite growth stages in metamorphic and magmatic rocks (Fig. 7c; Manzotti et al. 2018; Corti et al. 2020). Corti et al. (2020) compared composition and internal structure of allanite crystals from metagranitoids recording different strain rates during HP and LT metamorphism, concluding that the matrix and allanite crystals accommodated plastic and brittle deformation, respectively. Allanite resistance to plastic deformation is a noteworthy characteristic, as evidenced by relicts preserved in the sheared eclogites from Monte Mucrone (Stünitz and Tullis 2001; Cenki-Tok et al. 2011). However, fracturing during brittle deformation can disturb its isotopic system (Burn 2016).

Despite the degree of deformation, the allanite grains exhibit the same sequence of chemical zoning pattern (evident for Ca and Ce), but with different textures and LREE contents, suggesting that deformation facilitates the release of LREEs (Fig. 7a, d and e; Corti et al. 2020). Gregory et al. (2012) used Th/U v. La/Sm and Eu/Eu* v. La/Sm discrimination diagrams to distinguish low- from high-temperature magmatic allanite, revealing useful information about the amount of melt present during allanite growth. However, no geochemical ratios have yet been found to systematically discriminate metamorphic from magmatic allanite (Di Rosa et al. 2020). In migmatitic rocks, allanite incorporates significant amounts of Th relative to melt (e.g. Hermann and Rubatto 2009), and it may exhibit intermediate Pb_c values, between those for magmatic and metamorphic allanite (Gregory et al. 2012).

The favourable composition of allanite allows dating using the ²³²Th/²⁰⁸Pb, ²³⁸U/²⁰⁶Pb and ²³⁵U/²⁰⁷Pb systems. Allanite dating techniques range from single and multi-grain ID-TIMS (e.g. von Blackenburg 1992; Oberli et al. 2004; Smye et al. 2014; López-Moro et al. 2017) to in situ analysis using SHRIMP, SIMS or LA-ICP-MS (e.g. Catlos et al. 2000; Janots et al. 2009; Darling et al. 2012; Regis et al. 2014; Burn et al. 2017; Giuntoli et al. 2018; Liao et al. 2020; Vho et al. 2020). Challenges in the use of allanite reference material for LA-ICP-MS and SHRIMP methods (Burn 2016) include: (i) chemical and isotopic heterogeneities (e.g. Gregory et al. 2007; Boston et al. 2017; Giuntoli et al. 2018); (ii) excess ²⁰⁶Pb in magmatic allanite, which is the most widely used reference material (e.g. BONA, CAP and TARA allanite; Gregory et al. 2007; Burn et al. 2017; Yang et al. 2022); and (iii) use of non-matrix-matched reference materials (e.g. NIST610 glass by McFarlane 2016; Plešovice zircon in Burn et al. 2017). Nonetheless, robust LA-ICP-MS results have been obtained also by using zircon as primary reference material (Darling et al. 2012; Burn et al. 2017; Vho et al. 2020), which has a similar structure to allanite. Rastering (Darling et al. 2012) or a spot analyses routine (Burn 2016) can also be used to minimize matrix sensitivity in LA-ICP-MS analysis. As other accessory minerals, allanite may also incorporate non-radiogenic (²⁰⁴Pb) as well as radiogenic Pb (intermediate nuclei from ²³⁸U decay) affecting dates and uncertainties (Romer and Siegesmund 2003; Darling et al. 2012; Engi 2017). Furthermore, the structure of allanite can complicate geochronological procedures, as matrix matching reference materials is needed and the decay from ²³²Th, ²³⁵U and ²³⁸U causes structural damage to the crystal lattice (Burn 2016; McFarlane 2016). The destruction of the crystalline structure promotes Pb-loss and/or actinide remobilization with the formation of Th- and U-rich mineral phases (e.g. Barth et al. 1994; Smye et al. 2014). Additionally, deformation and interaction with fluids may open isotopic systems and play an important role in mineral re-equilibration, resorption and precipitation (Radulescu et al. 2009; Airaghi et al. 2019; Corti et al. 2020). Thus, syn-kinematic allanite has been used to date deformation processes at upper to middle crustal levels (Cenki-Tok et al. 2011). Investigations of magmatic allanite from an intensely deformed Mesoproterozoic granite in southern Norway indicate that higher mobility of Th than Pb during deformation processes results in the U–Pb system being more reliable than the Th–Pb system (Burn 2016). Nevertheless, allanite crystal structure may protractedly recover by annealing (Karioris et al. 1981), forming preserved (non-metamict) crystals that may yield younger dates (Catlos et al. 2000). More recently, allanite has been used as primary and secondary reference material for epidote dating (Peverelli et al. 2022).

When dating allanite, one of the three approaches discussed by Burn (2016) and Engi (2017) should be utilized to deal with initial or Pb_c correction. (i) Consider Pb_c evolution models. This approach is more often used for magmatic (Barth et al. 1994) than metamorphic (e.g. Radulescu et al. 2009; Rubatto et al. 2011) rocks due to the main issue of using global (silicate Earth or mantle) evolution models. Th, U and Pb contents differ from rock to rock, and their distribution within minerals is heterogeneous. These factors depend on several variables such as local effective bulk composition and fluids availability (Lanari and Engi 2017). (ii) Measuring Pb_c in phases that coexist with allanite (e.g. Cenki-Tok et al. 2014), which can be hampered by the difficulty in interpreting coexisting phases. Finally, (iii) the ‘intercept approach’, in which intercepts from uncorrected TW and ²⁰⁶Pb_c normalized Th–Pb isochron diagrams are used to estimate the initial Pb_c (e.g. Janots and Rubatto 2014; Airaghi et al. 2019). Gregory et al. (2012) indicate that igneous allanite tends to have smaller amounts of non-radiogenic ²⁰⁸Pb than high-grade metamorphic rocks, whereas allanite crystals formed at subsolidus conditions exhibit the highest non-radiogenic ²⁰⁸Pb values.

Allanite enrichment in LREE and Sr allows for both in situ Nd and Sr isotopic-based petrogenetic information to be combined with U–Th–Pb dating, making allanite an important petrogenetic tool (Hoshino et al. 2007). Heterogeneous Pb and Sr isotopic concentrations in allanite are demonstrated to be inherited from precursor minerals involved in allanite-producing metamorphic reactions (Romer and Xiao 2005). Zhang et al. (2022) combined U–Th–Pb dating and Nd isotopes of allanite with U–Pb–Hf analyses of zircon, demonstrating good correlations between the two systems to investigate crustal formation and evolution. Nd isotopes in allanite analysis were used by Su et al. (2021) to unravel the hydrothermal history of an iron oxide copper–gold deposit from ore formation (multi-source) to the post-ore overprinting tectono-thermal events.

Monazite (LREE,Y,Th,Ca,Si)PO₄ is a REE-rich phosphate mineral (Fig. 2p-r) that occurs in a variety of rock compositions and from diagenetic (Evans and Zalasiewicz 1996; Pereira et al. this volume, in press) to granulite-facies conditions (Black et al. 1984). Monazite is common in peraluminous granites (e.g. Montel 1993; Förster 1998), syenite, granitic and quartz veins (e.g. Piechocka et al. 2017) and carbonatitic plutons (e.g. Anenburg et al. 2021; Kamenetsky et al. 2021), and can exhibit various textures (from sector, firtree, oscillatory to lobate; Fig. 8a). However, under certain fluid conditions, in particular alkali-bearing fluids, monazite undergoes coupled dissolution–reprecipitation (CDR) (Vavra and Schaltegger 1999; Harlov and Hetherington 2010; Harlov et al. 2011; Kelly et al. 2012; Taylor et al. 2014, 2016; Bosse and Villa 2019; Weinberg et al. 2020; Salminen et al. 2022). CDR of monazite leaves very distinctive, lobate textures but can also modify U–Pb systematics (Fig. 8b; e.g. Vavra and Schaltegger 1999; Taylor et al. 2014; Blereau et al. 2016; Prent et al. 2020), cause U and Th loss (Williams et al. 2011) or Th gain (Harlov et al. 2011). Monazite is a common accessory mineral in metapelitic rocks experiencing low-amphibolite facies metamorphism (e.g. Rubatto 2002) to high-grade granulites, migmatites and charnockites (Fig. 8b; e.g. Laurent et al. 2018; Dev et al. 2021; Williams et al. 2022), whereas it is less common in HP rocks (Finger and Krenn 2007). Whilst geochronological disruption causes resetting of the oldest growth of the mineral, the reactivity of monazite is useful for tracing or determining the timing of fluid activity, reactivation of metamorphic processes and/or ore genesis. On the other hand, xenotime (Y, HREE)PO₄ is a HREE carrier, which also makes it a critical source of HREEs (Strzelecki et al. 2022). Xenotime occurs in metapelitic, granitic and carbonatitic rocks, and is less abundant in mafic and calc-silicate rocks (Spear and Pyle 2002). Recent phase equilibria modelling studies have demonstrated that xenotime usually occurs at pressures lower than 8 kbar and temperatures lower than 750°C (e.g. Shrestha et al. 2019). In contrast, monazite is found to be stable at higher pressures and temperatures, where water availability strongly controls the monazite P–T stability field and its preservation along the prograde path at much higher P–T conditions than previously envisaged (Larson et al. 2022).

Both monazite and xenotime are REE-rich mineral phases, containing a critical amount of REEs in addition to Y (Schulz 2021 and references therein). Monazite REE patterns are usually characterized by a negative slope, whereas xenotime commonly exhibits a positive REE pattern more like zircon (Fig. 8c). Sr-enrichment in monazite has been interpreted to reflect HP monazite growing in the absence of feldspars (Finger and Krenn 2007). Additionally, a REE signature like HP zircon, with low HREE and absence of a negative Eu anomaly, has been identified in monazite grains from the Kokchetav UHP rocks in Kazakhstan, and UHP rocks in Norway, suggesting this as a geochemical signature for HP monazite (Hermann and Rubatto 2014; Hacker et al. 2015). Furthermore, REEs in monazite have been utilized as geochemical discriminators with U–Pb geochronology to distinguish between magmatic and metamorphic monazite in complex deformed and metamorphosed terranes (e.g. Pe-Piper et al. 2014; Prent et al. 2019; Itano et al. 2020).

Like zircon, monazite has a number of partitioning relationships that can be useful in integrating various data sources to the evolution of a sample's mineral paragenesis, including monazite/melt (Yurimoto et al. 1990; Ward et al. 1992; Bea et al. 1994; Stepanov et al. 2012), monazite/xenotime (Andrehs and Heinrich 1998), monazite/K-feldspar (Villaseca et al. 2003) and monazite/garnet (Hermann and Rubatto 2003). The most applied technique is the partitioning of REEs between monazite and garnet since these minerals frequently occur together (e.g. Buick et al. 2006; Rubatto et al. 2006; Kylander-Clark et al. 2013; Mottram et al. 2014; Taylor et al. 2014; Blereau et al. 2016; Hagen-Peter et al. 2016; Hacker et al. 2019; Warren et al. 2019). Despite monazite being relatively poor in HREEs compared to zircon and garnet, the presence of garnet still impacts the relative concentration of HREEs in monazite. Monazite growing syn- to post-garnet or modified in the presence of garnet (see also Discussion) typically shows a reduction in HREEs and Y compared to monazite grown in the absence of garnet (Hermann and Rubatto 2003; Rubatto et al. 2006). Recent studies have demonstrated more complicated processes associated with HREE partitioning in monazite, where partitioning coefficients between monazite and garnet within investigated metapelitic rocks did not reproduce the expected values (e.g. Larson et al. 2019, 2022; Shrestha et al. 2019). Their temperature dependence is also found to be more relevant than initially envisaged (Hacker et al. 2019; Warren et al. 2019; Jiao et al. 2021). Also, recent work has highlighted by modelling Sm–Eu–Gd partitioning in suprasolidus systems that even though fractionation of Eu by feldspar growth can dominantly control the Eu budget, at equilibrium, other factors such as oxygen fugacity were shown to play an important role (Holder et al. 2020).

Monazite has proved a reliable geochronometer using the U–Pb system (Parrish 1990; Williams et al. 2007). Generally, monazite does not incorporate Pb into its crystal structure but has high Th (often several wt%) and U (several thousand ppm), meaning that it has low Pb_c and high radiogenic Pb. Monazite is also inferred to have a high Tc of up to 900°C (Fig. 1; Cherniak et al. 2004). However, monazite can be quite reactive (via CDR, see above), resulting in resetting of ages in fluid-dominated systems (Seydoux-Guillaume et al. 2002), whereas monazite in dry rocks (i.e. granulites) has been found to preserve detrital ages (e.g. Suzuki and Adachi 1994; Cutts et al. 2013; Guo et al. 2020). The high U and Th contents of monazite allow dating via chemical U–Th–Pb dating using EPMA (e.g. Suzuki and Adachi 1994; Montel et al. 1996, 2018). The advantages of this approach include in situ, non-destructive analysis and a small spot size allowing monazite grains <20 µm to be targeted (Ning et al. 2019; Williams and Jercinovic 2002). The main pitfall is that Pb_c correction is not possible (Williams et al. 2017), involving the assumption that the analysed grain is concordant and contains no Pb_c. The CHIME (chemical Th–U-total Pb isochron method) dating method is like chemical dating but targets multiple spots in an age domain to produce a ‘pseudo-isochron’. This method also uses compositional criteria to determine if the monazite age is concordant, resulting in a more robust method than traditional chemical dating (Suzuki and Kato 2008). Improvements in EPMA sensitivity and the method make this a powerful technique moving forward (Konečný et al. 2018; Montel et al. 2018; Ning et al. 2019). Prior to EPMA dating, many studies utilized ID-TIMS geochronology of monazite (i.e. Smith and Barreiro 1990), and later the LA-ICP-MS approach (Machado and Gauthier 1996; Poitrasson et al. 2000). Due to its high U and Th contents, monazite is extremely amenable to LA-ICP-MS, where small spot sizes (8–12 µm) can be used. Similarly, geochronology of xenotime can be analysed in situ by both SIMS (Cross 2009; Fielding et al. 2017) and LA-ICP-MS (Lawley et al. 2015; Simpson et al. 2021) to collect texturally contextualized isotopic dates. Currently, no matrix-matched reference materials for xenotime are available as no homogeneous natural xenotime has been found. Therefore, first-order matrix corrections have been applied on xenotime reference materials such as z6413 (Stern and Rayner 2003) and MG-1 (Fletcher et al. 2004) for U–Pb geochronology, whereas glass NIST 610 is currently used as primary reference material for in situ Lu–Hf dating (Simpson et al. 2021). Due to the wealth of information provided by REE data, TEs are commonly collected either separately or simultaneously with geochronological data using one or two mass spectrometers (Holder et al. 2013; e.g. Kylander-Clark et al. 2013; Hacker et al. 2015; Volante et al. 2020c; Barrote et al. 2022b). Nanoscale geochronological analysis of monazite (Fougerouse et al. 2020, 2021) and xenotime (Joseph et al. 2021) via atom probe (ATP) is also possible. Analytical development has allowed to investigate diffusion and migration of atoms along monazite grain boundaries and/or crystal defects, and relates e.g. intracrystalline deformation with age resetting due to radiogenic Pb loss (Fougerouse et al. 2021). Great potential to determine ages by using the ATP was shown also on small xenotime crystals (Joseph et al. 2021). In situ studies allow monazite ages to be directly related to mineral textures present in the rock, so the age of deformation can be attributed to metamorphic events in complex terranes (e.g. Smith and Barreiro 1990; Foster et al. 2002; Štípská et al. 2015; Piechocka et al. 2017; Prent et al. 2019, 2020; Jiao et al. 2020a; Volante et al. 2020c).

In granitic rocks, high LREE minerals, like monazite, are great competitors for trace elements, including Sm–Nd and REE, in the host magmas (e.g. Fisher et al. 2017; Hammerli and Kemp 2021). The Sm–Nd isotope tracer system in monazite can complement the more commonly applied Lu–Hf system in zircon (e.g. Fisher et al. 2017; Martin et al. 2020; Mulder et al. 2021; Barrote et al. 2022a; Volante et al. 2022) and whole-rock Sm–Nd and/or Lu–Hf (e.g. Mark 2001; Hoffmann et al. 2011; Caxito et al. 2021), providing insights into the formation, evolution and differentiation of the continental crust (Hammerli and Kemp 2021). Significant contributions have been made to improve in situ Sm–Nd precision using LASS-MC-ICP-MS (e.g. Barrote et al. 2022a) to achieve a comparable level of precision to Lu–Hf in zircon (e.g. Fisher et al. 2011, 2020; Goudie et al. 2014; Spencer et al. 2020). Furthermore, monazite is less prone to weathering than other Sm–Nd-bearing major minerals (e.g. plagioclase), making it a valuable tool to investigate primary isotopic signatures (e.g. Barrote et al. 2022a).

To estimate minimum magmatic temperatures, monazite thermometry relies on whole-rock REE content in peraluminous and metaluminous granitic rocks (Montel 1993; Plank et al. 2009; Stepanov et al. 2012). While the Montel (1993) equation is calibrated only at low pressure, Stepanov et al. (2012) corrected it by extending experiments to higher pressures. Recent studies have presented the importance of applying variable H₂O contents for granitic rocks to determine monazite saturation temperatures, with this approach producing consistent zircon and monazite saturation temperatures across the same granitic samples (Volante et al. 2020a). The Y-HREE fractionation between monazite and xenotime results in an asymmetric miscibility gap in YPO₄-(REE)PO₄. This allows for the Y content of monazite to be used as a geothermometer, which is largely independent of pressure (e.g. Gratz and Heinrich 1997, 1998; Heinrich et al. 1997; Andrehs and Heinrich 1998). This thermometer can only be used when monazite has grown in the presence of xenotime. However, identifying and ascertaining that these two phases grew in equilibrium can be challenging (Pyle and Spear 1999). Like garnet, xenotime also preferentially incorporates HREE. Hence, xenotime is usually consumed during prograde metamorphism by garnet growth, and it commonly reappears during garnet breakdown at post-peak conditions or along the retrograde path (e.g. Hallett and Spear 2015). In this view, compositional variations in monazite (e.g. Regis et al. 2016; Manzotti et al. 2018) have been crucial to assess xenotime saturation within the rock system (e.g. Krenn and Finger 2010). For example, Y-rich monazite domains have been suggested to be plausible targets to use for thermometry (e.g. Viskupic and Hodges 2001; Krenn et al. 2012; Laurent et al. 2018). Y content can also be used to distinguish between different monazite age populations (i.e. Fig. 8d), and can be combined with thermometry to indicate the temperature–time evolution of the sample (e.g. Laurent et al. 2018). Some attempts were made to thermodynamically model monazite and xenotime (e.g. Kelsey et al. 2007, 2008; Spear and Pyle 2010; Shrestha et al. 2019). However, this approach only works with low-Ca bulk-rock compositions and requires the simplification of a complicated system.

Apatite Ca₅(PO₄)₃(F,Cl,OH) is a calcium phosphate (Fig. 2s-u; Fig. 9) that is often present as an accessory mineral in several rock types, including various igneous rocks (ultramafic–felsic) to metamorphic and sedimentary, and even in meteorites (e.g. Harlov 2015; McCubbin and Jones 2015; Webster and Piccoli 2015). Apatite can readily incorporate large amounts of TE (e.g. U, Th, REE, Y, Sr), which replace Ca in the crystal lattice with charge compensation mechanisms (Engi 2017), making apatite important for the trace element budget of a rock. In addition, apatite commonly contains significant amounts of water and/or halogens and volatile species (e.g. F and Cl; Harrison and Watson 1984; Piccoli and Candela 2002), and records the evolution of metasomatic and hydrothermal fluids (e.g. Harlov 2015; Lin et al. 2023). Combined, the geochemical characteristics of apatite make it suitable for many geochemical and petrological applications, with measurements being possible with a large variety of instruments (e.g. SIMS, (LA)-ICP-MS, (LA)-MC-ICP-MS, EPMA), imaging with BSE (Fig. 9a) and TE mapping with LA-ICP-MS (Fig. 2r). Depending on its crystallization conditions, apatite can exhibit various textures visible in CL or BSE images (e.g. Mühlberg et al. 2021) and geochemical signatures that make this mineral a very interesting proxy for rock formation and evolution as well as for chronological information (Chew et al. 2011; Hammerli et al. 2014; Kirkland et al. 2017; Antoine et al. 2020; Fisher et al. 2020; Prent et al. 2020; Paul et al. 2021).

Igneous apatite has reasonably similar total REE contents regardless of the protolith lithology (e.g. Chu et al. 2009), except for apatite in alkali-rich rocks, which are commonly significantly REE-rich (e.g. Zirner et al. 2015). However, relative REE concentrations, and Y and Sr, are variable within different rock-types and particularly effective for discriminating magma types (e.g. Bea 1996; Belousova et al. 2001; Bruand et al. 2020). Negative Eu anomalies are present in apatite from most igneous rocks except ultramafics (e.g. Chakhmouradian et al. 2017; Bruand et al. 2020), and the magnitude of Eu anomalies is typically positively correlated with SiO₂ content. In contrast, Sr content in apatite is negatively correlated with SiO₂ content, with the highest Sr contents being observed in ultramafic rocks (Ihlen et al. 2014) and the lowest Sr contents in felsic rocks (Sha and Chappell 1999). Also, Sr in apatite strongly correlates with Sr content in the corresponding whole-rock (Belousova et al. 2001; Jennings et al. 2011; Bruand et al. 2014). Variations in Sr content in TTGs are interpreted to directly correlate with the melting depth of the TTGs’ source, where a deep source is in equilibrium with garnet and rutile, but no plagioclase (Moyen and Martin 2012). Moreover, apatite chemistry can be used as a tracer of different magmatic petrogenetic processes, reflecting geodynamic changes during crustal evolution (Bruand et al. 2020) such as (La/Yb)_N v. Yb_N discrimination diagram for TTGs (Antoine et al. 2020; Fig. 9b–d).

Metamorphic apatite also exhibits contrasting signatures depending on its metamorphic grade and its textural context (e.g. O'Sullivan and Chew 2020; Prent et al. 2020). Total REE content increases with the metamorphic grade (e.g. El Korh et al. 2009), such that in high-grade rocks it is indistinguishable from that of igneous apatite (e.g. Bingen et al. 1996). In addition, negative Eu anomalies characterize apatite growing in low-grade metamorphic rocks (Henrichs et al. 2018). Also, metamorphic apatite contains higher Sr concentrations in greenschist- and blueschist-facies rocks (e.g. Nishizawa et al. 2005), and lower ones in migmatites and granulite-facies rocks (e.g. Nutman 2007). In felsic igneous and metamorphic rocks, apatite (La/Lu)_N ratio is commonly ≤1, whereas in all other rocks (La/Lu)_N ratio is >1, with the highest values recorded in ultramafic rocks (e.g. O'Reilly and Griffin 2000). A discrimination diagram using Sr/Y v. ΣLREE biplots from various igneous and metamorphic/metasomatic rocks (Fig. 9e) can be used to statistically categorize source lithologies (O'Sullivan et al. 2020). Additional discriminant diagrams based on the enrichment in LREEs and MREEs, Eu anomalies, the tetrad effect, Mn and Sr contents, total REEs and Y in apatite can be used to highlight the variability between different types of mineralization (Decrée et al. 2023). For example, bell-shaped REE patterns defined by MREE enrichment and positive Eu anomalies were found to be a unique indicator of hydrothermal apatite formed in a mineralized, reducing hydrothermal system (e.g. Krneta et al. 2018; Lin et al. 2023).

Apatite has long been the target for geochronological studies (e.g. Schoene and Bowring 2007; Chew et al. 2011; Glorie et al. 2022) either using the U–Th–Pb or Lu–Hf isotope system (e.g. Barfod et al. 2005; Chew et al. 2011; Simpson et al. 2021; Glorie et al. 2023), or Sm–Nd (Fisher et al. 2020). Apatite commonly incorporates significant amounts of Pb_c that vary from limited amounts in igneous apatite to larger amounts in hydrothermal apatite crystals (e.g. Kirkland et al. 2017). Low-grade metamorphic apatite typically provides poor precision for U–Pb dating due to low U concentrations (<5 ppm) and high initial Pb_c contents, while apatite from igneous rocks is typically U-rich (>20 ppm) and has lower Pb_c, leading to more precise U–Pb ages (e.g. Henrichs et al. 2018, 2019; O'Sullivan et al. 2018). Because of its Pb_c content, apatite can often produce U–Pb dates that are strongly discordant (also see Titanite section). This Pb_c presents a particular challenge for young samples that have had little time to accumulate substantial radiogenic Pb (Pb*), or for apatite grains with low concentrations of U. Following to the development of apatite U–Th–Pb age reference materials (e.g. Chew et al. 2011; Thomson et al. 2012; Apen et al. 2022; Lana et al. 2022) and data reduction schemes employing ²⁰⁸Pb-, ²⁰⁷Pb- or ²⁰⁴Pb-based Pb_c corrections to age reference materials and unknowns, it is now possible to routinely date apatite both precisely and accurately (e.g. Andersson et al. 2008, 2022; Chew et al. 2011; Thomson et al. 2012; Antoine et al. 2020; Prent et al. 2020; Glorie et al. 2022). However, some complications in data interpretation can occur in apatite crystals that have undergone metamorphism due to diffusion effects (e.g. Paul et al. 2019). Apatite U–Pb dates may be reset by deformation and fluids, causing the age to represent the last major deformation event or latest dissolution–precipitation and/or chemical exchange (Odlum et al. 2022).

Recent advances including the development of LA-ICP-MS/MS technology allowed resolution of isobaric interferences for Lu–Hf in situ dating of apatite crystals (e.g. Barfod et al. 2003, 2005; Larsson and Söderlund 2005; Simpson et al. 2021; Gillespie et al. 2022; Glorie et al. 2022, this volume, in press) and Sm–Nd (e.g. Hammerli et al. 2014, 2019; Doucelance et al. 2020; Fisher et al. 2020), which is extremely useful to investigate post-metamorphic cooling history and/or compare crystallization ages. Apatite is commonly combined with other mineral phases that exhibit various parent/daughter ratios (i.e. Lu/Hf and Sm/Nd) to obtain a more robust isotope isochron (e.g. Hammerli et al. 2014; Laurent et al. 2017; Simpson et al. 2021). Advantages of the Lu–Hf isotope system in apatite over U–Pb include higher Tc of the former (Fig. 1), resulting in an age closer to the apatite crystallization age (e.g. Henrichs et al. 2019). Comparisons of the two systems can have significant advantages (Fig. 9f; Glorie et al. 2022).

The Sm–Nd isotope system in apatite can provide insights into timing and source of apatite host-rock (e.g. Fisher et al. 2020). Among stable isotopes, oxygen isotopes appear to be proxies for magma sources and to record fluid circulation and metamorphism processes (e.g. Bruand et al. 2019). Chlorine and hydrogen isotopes have also been used as proxies for volatilization/condensation processes (e.g. Potts et al. 2018; Wudarska et al. 2020). Isotope systematics such as ⁸⁷Rb–⁸⁷Sr can be measured using either (LA-)MC-ICP-MS or LA-ICP-MS/MS, TIMS or SIMS, and coupled to age information (Fig. 9g; Emo et al. 2018; Ravindran et al. 2020; Gillespie et al. 2021). The ⁸⁷Sr/⁸⁶Sr ratio of apatite may be used together with K-rich or Rb-bearing minerals such as micas for in situ Rb–Sr geochronology to create a combined isochron (Olierook et al. 2020).

There is no currently developed thermometer for apatite like those for zircon (Ti-in-zircon) or rutile (Zr-in-rutile), but the U–Pb Tc (350–570°C; Fig. 1) can be used as an indirect proxy for tracking metamorphic or magmatic cooling in combination with other chronometers in apatite (Fig. 1; e.g. Cherniak 2000b; Barfod et al. 2003, 2005; Cochrane et al. 2014; Chew and Spikings 2015; Kirkland et al. 2018; Ferreira et al. 2022).

Geochronological and geochemical data retrieved from metamorphic accessory minerals are commonly integrated with information from the major mineral paragenesis, P–T constraints from phase equilibrium diagram calculations and microstructures. Figure 10 reflects an example of two synoptic, theoretical P–T–t evolutions where petrochronological information of metamorphic accessory minerals such as monazite, zircon and garnet is linked to P–T information. The first P–T path (Fig. 10a) represents common clockwise granulite-facies metamorphism of a pelitic protolith hosting inherited zircon (Fig. 10a, i). During prograde metamorphism, garnet growth begins and inherited zircon grains start to recrystallize with increasing temperatures (Fig. 10a ii). This prograde part of the metamorphic evolution is often difficult to accurately constrain due to open-system processes, overprinting and/or exceeding closure temperatures of earlier phases. Zircon has been shown to grow along the prograde path connected to the movement of locally-derived melts (e.g. Harley and Nandakumar 2014; Harley 2016; Weinberg et al. 2020) as well as injected, externally-derived melts (Andersson et al. 2002; Flowerdew et al. 2006; Wu et al. 2007). Once the sample crosses the solidus, partial melting begins. Inherited zircons may be consumed and or dissolved in the melt that becomes enriched in zirconium (Watson and Harrison 1984; Boehnke et al. 2013; Gervasoni et al. 2016), forming embayment and relict core textures in zircon. Monazite might also grow along the prograde and retrograde path (e.g. Rubatto 2002; Shrestha et al. 2019; Larson et al. 2022), but also during (Kelsey et al. 2008; Rubatto et al. 2009; Larson et al. 2022) and/or just after max T conditions are reached (Fig. 10a iii, v, vii; Clark et al. 2014; Shrestha et al. 2019), making it a more suitable time-capsule than zircon by preserving a greater P–T–t window at subsolidus conditions (e.g. Ambrose et al. 2015; Mottram et al. 2015; Hacker et al. 2019; Shrestha et al. 2020; Larson et al. 2022). Recent works modelled the P–T stability fields for monazite and xenotime, with monazite present during prograde growth of garnet, max T(/P) conditions and along the retrograde path during garnet breakdown at both sub- and suprasolidus conditions (Shrestha et al. 2019). In contrast, xenotime appears only at low P–T conditions during the early stages of the prograde path and/or along the retrograde path, when it would start to assimilate most of the Y and HREE, resulting in a decrease in monazite proportions (Shrestha et al. 2019). Prograde garnet sees further growth at suprasolidus conditions, combined with an additional generation of peritectic garnet (Fig. 10a iv) that could be distinguished based on different REE patterns and inclusions (Fig. 3f). Trace element partitioning is mostly utilized for max T(/P) to retrograde portion of the P–T–t history as this information is most likely preserved. For example, Figure 10a shows that garnet growth and metamorphic ‘soccer ball’ zircon (e.g. Vavra and Schaltegger 1999; Blereau et al. 2016; Taylor et al. 2016) grow in the same geochemical system and have a near 1:1 partitioning of M-HREE. How well these partitioning relationships are preserved depends on the conditions of metamorphism and how these affect the equilibration volume of the sample. Typically, this is largest at max T conditions, where rates of diffusion and potential fluids and/or melt increase. Approaching the solidus on the retrograde path, Zr-bearing phases may breakdown, e.g. garnet to cordierite (Fig. 10a vi) resulting in additional zircon growth (e.g. Fraser et al. 1997; Degeling et al. 2001; Wu et al. 2007; Kelsey et al. 2008), from rims to completely new grains (Fig. 10a vii). A subsequent fluid event, e.g. upon crystallization of local melt, can modify monazite and zircon textures, perturbing age and geochemical data (Poitrasson et al. 2000; Seydoux-Guillaume et al. 2012; Kröner et al. 2014; Taylor et al. 2014; Blereau et al. 2016; Prent et al. 2019). In more complex garnet textures, such as the unusual atoll garnet (e.g. Jonnalagadda et al. 2017; Kulhánek et al. 2021; Godet et al. 2022; Massonne and Li 2022), TE content recorded by subsequent growth of concentric garnet rings can also provide useful information about the interchange of TE within the system due to growth and breakdown of other accessory phases (Godet et al. 2022). For example, by using LA-ICP-MS trace element mapping, Godet et al. (2022) reported enrichment in V and Ti in the garnet inner rim compared to the core, which was attributed to rutile breakdown, whereas Cr, Y, LREE and MREE enrichment in the outer rim was interpreted to reflect allanite and monazite breakdown.

The second P–T–t evolution (Fig. 10b) represents a more residual scenario, where the protolith has already been partially melted. In this case, recrystallization and preservation of existing major and accessory minerals are promoted due to the lack and/or reduction in volume of partial melt being produced, which also limits the generation of new assemblages (Fig. 10b, i, ii, iii; Bea and Montero 1999; White and Powell 2002). Monazite and zircon contrast in their behaviour under these conditions. Monazite may grow in larger amounts than zircon, despite being in a melt poor environment, due to higher reactivity (e.g. Högdahl et al. 2012; Rubatto et al. 2013; Morrissey et al. 2016), potentially recording information completely missed by zircon. As in the first scenario (Fig. 10a), relict garnet breaks down during decompression to an intergrowth of plagioclase + orthopyroxene, another micro zircon permitting reaction (Fig. 10b). Local hydration or different amounts of melt loss (Morrissey et al. 2016; Larson et al. 2022) is reflected by a second P–T event (Fig. 10b, v, vi, vii) recorded only by more reactive high-strain microsites. The interpretation of geochronological data obtained from studies of high-grade metamorphic rocks can also be extremely challenging as most terranes record a spread of ages rather than statistical populations (e.g. Whitehouse and Kemp 2010; Farias et al. 2020; Taylor et al. 2020; Finch et al. 2021; Gutieva et al. 2021; Mulder and Cawood 2021; Salminen et al. 2022; Whitehouse et al. 2022). For example, in Figure 10c, the younger age population could represent either a single prolonged metamorphic event or potentially two events, with the second modifying the first and many other possible combinations. Determining the correct age interpretation depends on mineral textures, where the analyses and sample come together with complementary geochemical and isotopic information.

Metamorphic reactions between REE-rich phases such as allanite–monazite–xenotime–apatite are illustrated with those of titanite–rutile on the same theoretical P–T–t–(d) path, though these sets of minerals may occur in different chemical systems (Fig. 11). The timing and sequence of metamorphic reactions (i.e. growth of accessory phases) depend not only on the variation of pressure and temperature conditions, including the residence time of a volume of rock at certain conditions, but also on the bulk-rock composition of the protolith and on strain rate, which strongly influence the reaction sequence during fluid–rock interaction at different crustal levels. Reconstructing the formation and breakdown of these minerals and growth of other major and minor phases in the rock usually requires a detailed petrographic investigation that includes X-ray compositional maps and analysis of Y and REE partitioning between the phases (e.g. Janots et al. 2008; Garber et al. 2017; Manzotti et al. 2018; Airaghi et al. 2019). Additionally, tools such as TE mapping by LA-ICP-MS (e.g. Raimondo et al. 2017; Chew et al. 2021; Sliwinski and Stoll 2021; Godet et al. 2022) allow for greater precision in determining chemical composition effects on growth or break-down of minerals associated with different P–T conditions and fluid-present or absent processes.

During the prograde evolution of a time capsule mineral such as garnet, key inclusions can be incorporated and provide insights regarding the P and T conditions a rock volume experienced during the early stages of the evolution path (Fig. 11). For example, identification of rutile inclusions in garnet core (Fig. 11) and prograde allanite can indicate that rutile grew during the prograde stages towards max P conditions (Boston et al. 2017). In this case, prograde growth of garnet sequestered most of the Mn and HREE, resulting in allanite cores with a relatively HREE-depleted pattern in relation to the rim. Manzotti et al. (2018) investigated rocks in which Y-rich garnet cores enclosed a first generation of rutile and allanite inclusions, whereas Y-poor garnet rims included monazite grains that crystallized aligned parallel to the matrix foliation (Fig. 11). At this stage, monazite may form after breakdown of allanite or other phosphates such as apatite and/or xenotime (Engi 2017; Manzotti et al. 2018), recording peak pressure conditions (e.g. Boston et al. 2017; Manzotti et al. 2018; Volante et al. 2020c; Barrote et al. 2022b; Fumes et al. 2022). Monazite REE patterns may exhibit low Y content together with a steeply negative HREE slope, suggesting that garnet was still stable during monazite growth, whereas pronounced negative Eu anomalies commonly suggest plagioclase stability or oxygen fugacity variations (Fig. 11; e.g. Boston et al. 2017; Holder et al. 2020). At low- to medium-metamorphic grade, allanite and monazite may replace each other, and this process depends on different factors including bulk-rock (e.g. Wing et al. 2003; Janots et al. 2008; Spear 2010) and fluid (Budzyń et al. 2010) compositions, as well as oxygen fugacity (Janots et al. 2011) and pressure and temperature conditions (Janots et al. 2007). Post-metamorphic pressure peak, along the lower-pressure, possibly higher-temperature path, a second generation of allanite and apatite can form at the expenses of monazite grains commonly exhibiting complex coronitic dissolution textures (Fig. 11; e.g. Manzotti et al. 2018). At this stage, in mafic and Ca-rich systems, titanite crystals may grow as a replacement product of rutile, whereas ilmenite would grow in a Ca-poor system and at lower pressures. During retrograde, low-grade hydrothermal metamorphism in pelitic systems, new xenotime and monazite may crystallize following allanite, apatite, residual monazite relicts and garnet breakdown, with xenotime associated with chlorite and minor biotite aggregates replacing garnet porphyroblasts (Fig. 11; Manzotti et al. 2018). Most of the Y released by garnet would be incorporated by xenotime and monazite in minor amounts. New rutile crystals may grow during the retrograde cooling stage at the expense of titanite and as rutile exsolution, growing parallel to biotite cleavage during its replacement by chlorite (Fig. 11). During this late hydrothermal stage, rutile may grow with apatite and record the cooling stages (e.g. Apen et al. 2020).

In this section, we emphasize that a combined multi-mineral approach is a powerful tool to provide constraints on magma sources (e.g. mineralogy, composition), differentiation processes (e.g. mixing, fractional crystallization) and timing of magma formation and crystallization. The first scenario corresponds to simple fractional crystallization of a single magma batch (Fig. 12). The second is a more complex scenario wherein magma is contaminated by a distinct lithology and/or mingles with a different magma. The third scenario refers to Precambrian gneisses, wherein the objective is to recover the initial magmatic signatures despite the occurrence of a metamorphic overprinting event.

To constrain the crystallization age of a magmatic rock, zircon is arguably the easiest mineral to date among those presented here but it is not necessarily observed in all rock types. Consequently, the timing of igneous rock formation can be addressed using other phases such as titanite, rutile, monazite, apatite or allanite, provided that they record a magmatic age or fast cooling. In a scenario where a magmatic intrusion does not experience post-crystallization disturbance and contains most mineral phases presented in this contribution, multi-mineral dating becomes useful as these phases likely record different ages reflecting the magma cooling history (Fig. 12; e.g. Schaltegger et al. 2009; Jayananda et al. 2015). A potential issue is the precision on determined ages, as granite batholiths can cool over relatively short timescales (e.g. Coleman et al. 2004; Schaltegger et al. 2009; Barboni et al. 2013). In the same scenario, Lu–Hf in zircon/apatite, Sm–Nd in monazite/titanite/apatite and Rb–Sr in apatite, combined, allow these source tracing radiogenic isotope systems to be directly compared, and coupled to geochronological data (i.e. U–Th–Pb and Lu–Hf in apatite, and Sm–Nd in monazite/titanite/apatite). Also, comparisons of these isotopic systems at the whole-rock and mineral scale can help constrain source composition or detect small discrepancies that may be important for understanding rock formation (e.g. Guitreau et al. 2012; Fisher et al. 2020; Barrote et al. 2022a; Volante et al. 2022; Zhang et al. 2022).

The second scenario includes a magmatic body that experienced mingling, mixing or contamination. Radiogenic systems allow end-members to be identified due to specific sensitivity to contamination, source characteristics and respective elemental concentrations. For example, Laurent et al. (2017) used Rb–Sr and Sm–Nd in apatite and titanite, respectively, from hybrid granitoids to show that these minerals did not crystallize at the same time and that while apatite formed in both magmas before mingling and after mingling (zonation), titanite crystallized only in the mingled magma. To identify end-member isotopic compositions, another geochemical study investigated zircon, apatite and titanite, which revealed mingling of mantle and crust-derived magmas (Sun et al. 2010). If one of the end-members is crustal derived, zircon xenocrysts can provide further information about crustal diversity (e.g. Villaros et al. 2012; Bea et al. 2021). Also, monazite can inherit Nd isotope heterogeneities from the source while losing the original U–Pb crystallization age (Fisher et al. 2017). Trace element patterns and ratios can provide further insights into magma evolution after and/or before mixing/mingling (e.g. Sun et al. 2010; Laurent et al. 2017). Melt and/or mineral inclusions in accessory minerals can also provide information about the melt source composition (e.g. Bruand et al. 2017; Antoine et al. 2020; Ferrero et al. 2021).

The third scenario considers Precambrian rocks, which often experienced polyphase histories that involved at least one metamorphic episode potentially compromising one or more isotope systems (e.g. Black 1988; Hammerli et al. 2014; Emo et al. 2018; Guitreau et al. 2018; Antoine et al. 2020; Fisher et al. 2020; Hammerli and Kemp 2021). For example, in situ Sm–Nd investigation of apatite, allanite, titanite, xenotime and monazite in Precambrian metasedimentary rocks revealed re-equilibration of heterogeneous Nd isotope signature in apatite over 550°C, after which it retains its Nd isotope signature throughout anatexis (Hammerli et al. 2014). While titanite and allanite equilibrate at LT (350–400°C), REE-rich accessory phases exhibit homogeneous Nd isotopic signatures at HT (600°C) and behave as open systems during partial melting and anatexis (Hammerli et al. 2014).

From the conceptualization of petrochronology (as summarized in Engi et al. 2017; Kohn et al. 2017), significant analytical improvements have occurred in the field of geoscience and have been applied to all key mineral phases that are described in this review. New analytical methods have been developed (e.g. atom probe dating, Fougerouse et al. 2021; in situ Lu–Hf dating using LA-ICP-MS/MS, Simpson et al. 2021, 2022, 2023), previous analytical methods have been continuously updated and improved (e.g. in situ EPMA dating, Montel et al. 2018; in situ U–Pb dating, Millonig et al. 2020), new and updated numerical tools for quantitative petrology have been developed and enhanced, including imaging and modelling tools (e.g. XMapTools, Lanari et al. 2014, 2019; Q–XRMA, Ortolano et al. 2018, Zucali et al. 2021; Bingo-Antidote, Duesterhoeft and Lanari 2020) as well as the use of TE maps (e.g. Raimondo et al. 2017; George et al. 2018; Lanari and Piccoli 2020; Rubatto et al. 2020; Gaidies et al. 2021), and new boundaries have been pushed to investigate melting mechanisms (e.g. Ferrero et al. 2021). In this section, we briefly discuss the exciting and bright future in front of the discipline of petrochronology that has only started to push analytical advances to the limit of conventional methods, allowing for a better understanding of plate- to nanoscale crustal processes.

Garnet is the ultimate petrochronometer (Baxter et al. 2017), with the most exciting advance being the development of in situ Lu–Hf geochronology (Simpson et al. 2021). The use of inclusions as proxies for whole-rock isotopic composition also means that this method may be suitable for detrital garnet studies (Maneiro et al. 2019; Mark et al. 2022). Future work should focus on enhancing accuracy and precision of Lu–Hf and U–Pb measurement in secondary reference material e.g. Aysal et al. 2023 and finding primary reference material (i.e. isotopically homogeneous) for common Hf-bearing phases such as garnet (Simpson et al. 2021). Additionally, incorporation of Mn in the activity–composition relations for amphibole and clinopyroxene in the thermodynamic dataset for mafic systems (Green et al. 2016) would also greatly enhance accuracy for garnet modelling (White et al. 2014). Recent advances using in situ U–Pb dating of garnet have also demonstrated significant potential (e.g. Seman et al. 2017; Millonig et al. 2020; Schannor et al. 2021) and applicability to almandine compositions (e.g. Cerva-Alves et al. 2021; Schannor et al. 2021). Further advances aim to extend the range of metasomatic and metamorphic garnet composition that can be targeted using this technique (Millonig et al. 2022; O'Sullivan et al. 2023). Finally, the continuous development of quantitative elemental and mineral map tools (e.g. XMapTools) along with EPMA and LA-ICP-MS trace element maps (e.g. Chew et al. 2021) will prove an extremely powerful tool to enhance in situ microanalytical investigations and our understanding of complex micro-chemical processes. The Bingo-Antidote add-on to XMapTools performs thermodynamic calculations by comparing modelled and observed mineral assemblages, modes and compositions (Duesterhoeft and Lanari 2020).

Zircon can preserve a realm of information that can be retrieved from images (optical, SEM), mineral inclusions and elemental and isotope data (e.g. Th/U, REE, δ¹⁸O, ε_Hf, δ³⁰Si). Although zircon has long been used in igneous and metamorphic studies, recent technical and methodological advances allow to further extend our use of this mineral. These include analytical improvements in δ³⁰Si and δ⁹⁴Zr data acquisition to enhance and extend the use of these isotopes in zircon studies, and the integration of EBSD data with petrochronological investigation to better understand TE diffusion in zircon crystals under various P–T conditions. Pioneer work on elastic geobarometry for a noncubic host-inclusion system indicates potential of anisotropic quartz-in-zircon elastic model for elastic thermobarometry, and its potential wide applicability to crustal rocks (Gonzalez et al. 2021).

Monazite is an increasingly useful accessory mineral as it can be used to trace the presence of certain fluids (e.g. Weinberg et al. 2020; Salminen et al. 2022), and can potentially grow throughout metamorphism (e.g. Larson et al. 2022). With improvement of analytical and spatial resolution, even if highly zoned, monazite will only become more applicable for unravelling complex and multifaceted processes. However, even with the collection of in situ data, it can still be challenging to tie monazite growth to deformation and fluid-related processes. Further work into monazite behaviour under deformation combined with the use of EBSD may prove a powerful tool. Additionally, the integration of EBSD, crystallographic vorticity axis analysis and petrochronology should be applied to monazite within deformational structures as it has been done for zircon (Brown et al. 2022). More work needs to be done on monazite TE partitioning to better understand temperature dependencies (e.g. Larson et al. 2022). Some attempts have been made to thermodynamically model monazite and xenotime (e.g. Kelsey et al. 2007, 2008; Spear and Pyle 2010; Shrestha et al. 2019), however, this approach only works with low-Ca bulk-rock compositions and requires the simplification of a complicated system. These current limitations represent an extensive platform for improvements in this field.

Titanite is a very reactive mineral compared to other accessory minerals; however, its tendency to incorporate initial Pb requires combining U–Pb dating with trace-element composition, zoning and microstructural information to accurately filter out any additional, unrelated radiogenic Pb (Walters et al. 2022). Analytical improvement in acquisition of Sm–Nd and O isotopic analysis in titanite may be extremely beneficial to trace magmatic sources (e.g. Bonamici et al. 2014; Bruand et al. 2019, 2020) and the origin of fluids playing a major role in crustal-scale shear zones (Gordon et al. 2021; Moser et al. 2022) and ore deposit (e.g. Marfin et al. 2020) genesis. The assessment of ilmenite U–Pb dating by LA-ICP-MS (Thompson et al. 2021) may be a valuable contribution for evaluating residual initial Pb affecting titanite dating. It may also be useful to investigate the timing of replacement of these Ti-rich phases. Advancing and developing the recent application of elastic geothermobarometry of host-inclusion systems (e.g. Mazzucchelli et al. 2021) to titanite grains may provide a complementary non-destructive method to estimate the P and T at which inclusions were trapped in the hosting titanite (Nestola 2021).

Rutile is a crucial mineral to unravel the early history of rock's evolution, including the prograde, UHP, burial stage (e.g. Rezvukhina et al. 2021; Manzotti et al. 2022), to trace melt sources in magmatic systems using Lu–Hf as isotopic tracer (e.g. Ewing et al. 2011) and hydrothermal, metasomatic processes associated with ore deposit genesis (e.g. Agangi et al. 2019; Ballouard et al. 2020; Porter et al. 2020; Sciuba and Beaudoin 2021). Therefore, analytical advances in microanalytical and petrochronological investigations of this mineral phase are essential to contribute to our understanding of orogenic systems, tectonic environments and exploration strategies. A better micro- to nanoscale understanding of trace element mobility in rutile (Kooijman et al. 2012; Ewing et al. 2013; Kohn et al. 2016; Pape et al. 2016; Penniston-Dorland et al. 2018; Smye et al. 2018) may improve U–Pb dating and associated trace element distribution and isotopic information (Verberne et al. 2022a, b), including the investigation of Zr transport through the rock's matrix (Ewing et al. 2013; Kohn et al. 2016). Also, advancing the use of integrated EBSD investigations of rutile grains within different rock-types with LA-ICP-MS trace element mapping may significantly enhance our understanding of trace element partitioning and concentration in this mineral.

Allanite has overcome various challenges as a petrochronometer. Indeed, the highly variable composition of natural allanite complicates petrological interpretations and applications. Thermobarometric, isotopic (e.g. in situ oxygen) and TE-based petrological investigations have to be carried out using microanalytical techniques such as EBSD and/or LA-ICP-MS TE maps to explore these compositional variations. The use of compositional maps to distinguish different generations of allanite has proven to be an essential step in geochronological and petrological studies (e.g. Burn 2016; Airaghi et al. 2019; Corti et al. 2020), and the use of TE maps obtained with LA-ICP-MS can improve classification and understanding of allanite growth history. The chemical distinction between magmatic and metamorphic allanite has been enhanced by machine learning (e.g. Di Rosa et al. 2020), but more data are required to improve this classification. Nd and Sr isotopes are valuable geochemical tools that have been utilized in previous studies. Recent analytical advancements and developments have further strengthened the importance of these isotopic systems as tracers in understanding magmatic and metamorphic processes, especially in investigations of hydrothermal processes. Also, a better comprehension of the stability conditions (pressure, temperature and fluid) of allanite and correlated accessory minerals (e.g. monazite) is required. This would involve experiments, improvements on the thermodynamic dataset and investigation of inclusions (mineral, fluid and melt) in allanite.

Apatite is a useful tool for studying crustal processes by means of a variety of petrochronology methods, continuously enhanced in their accuracy and precision. Glorie et al. (2022, this volume, in press) show the promising apatite Lu–Hf geochronology method using LA ICP-MS/MS, which is more likely to reveal primary apatite growth ages in reworked terranes, due to the higher Tc, than the U–Pb system (Fig. 1). This may be important since Odlum et al. (2022) show that apatite REEs and U–Pb behaviour are decoupled in high-grade gneiss samples, suggesting REEs record higher-temperature processes than U–Pb isotopic systems. LA-ICP-MS trace element mapping is a promising technique to monitor the behaviour of trace elements in accessory mineral phases, and it should be more routinely applied. In apatite, it can be used to distinguish between metamorphic and magmatic grains or identify growth zones (Henrichs et al. 2018, 2019; Chew et al. 2021). Recent techniques such as age-depth profiles and laser ablation split stream make it possible to identify individual apatite generations and can aid with interpreting data from apatite domains with complex thermal histories (Kirkland et al. 2018). New apatite reference material is being characterized for in situ apatite U–Pb petrochronology and Sr–Nd isotope geochemistry (e.g. Apen et al. 2022; Kennedy et al. 2022) to obtain more precise and robust data.

In addition to the key mineral phases discussed in this contribution, we encourage and see potential in integrating other mineral phases in magmatic and metamorphic studies. Multi-mineral studies are paramount to combine major and accessory mineral phases that record different steps of the prograde and retrograde history of a volume of rock with lower temperature petrochronometers, which may retain useful information on the youngest evolution of the system. Critical and crucial advancement in petrochronology in the near future includes in situ investigations, where microanalytical studies are linked to the chemical and chronological information of different minerals and associated fabrics in a rock volume. Extreme potential is foreseen in the development and acquisition of quantitative LA-ICP-MS TE (e.g. George et al. 2018; Muñoz-Montecinos, J. et al. 2023) and geochronological maps (e.g. Chew et al. 2021). In this scenario, by enabling near-simultaneous detection, the time of flight (TOF) detector is a key component of LA-ICP-TOF-MS, a promising technique that captures the complete elemental mass spectrum for each laser pulse (Chew et al. 2021). Generating quantitative chemical and geochronological maps of a whole mineral using LA-ICP-MS can facilitate exploring the relationships between internal textures and trace and major elemental variation over time. This analytical technique will allow to discriminate distinct textural domains based on the ages and major, minor and trace elements, possibly avoiding issues such as mixing domains to help the interpretation of complex mineral textures. The advances made in LA-ICP-MS imaging have made it a critical analytical technique in the geosciences, as it enables the acquisition of high-resolution geochemical data that can help to constrain the timing and nature of igneous, metamorphic, ore-forming, sedimentary and diagenetic processes.

We thank Catherine Mottram and an anonymous reviewer for their constructive and insightful comments that greatly improved the manuscript. We also thank Inês Pereira for helpful comments and for her editorial handling. Martin Engi and Daniela Rubatto are thanked for insightful discussion on allanite. Vitor Barrote is thanked for long insightful discussion throughout the preparation of the manuscript about the pros and cons of using different analytical techniques in petrochronology applied to the mineral phases discussed in this manuscript. M. Tedeschi was supported by the National Council for Scientific and Technological Development – CNPq through a Research Productivity Grant 308733/2021-5. This is contribution no. 591 of the ClerVolc program of the International Research Center for Disaster Sciences and Sustainable Development of the University of Clermont Auvergne.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

SV: conceptualization (lead), formal analysis (lead), project administration (lead), writing – original draft (lead), writing – review & editing (lead); EB: conceptualization (equal), formal analysis (equal), writing – original draft (equal), writing – review & editing (equal); MG: conceptualization (equal), formal analysis (equal), writing – original draft (equal), writing – review & editing (equal); MT: conceptualization (supporting), formal analysis (supporting), writing – original draft (supporting), writing – review & editing (supporting); VVS: conceptualization (supporting), data curation (supporting), writing – original draft (supporting), writing – review & editing (supporting); KC: conceptualization (supporting), data curation (supporting), writing – original draft (supporting), writing – review & editing (supporting).

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.