A Protocol for Electron Probe Microanalysis (EPMA) of Monazite for Chemical Th-U-Pb Age Dating

Abstract

A protocol for the monazite (LREE,Y,Th,U,Si,Ca)PO₄ in situ Th-U-Pb dating by electron probe microanalyser (EPMA) involves a suitable reference monazite. Ages of several potential reference monazites were determined by TIMS-U-Pb isotope analysis. The EPMA protocol is based on calibration with REE-orthophosphates and a homogeneous Th-rich reference monazite at beam conditions of 20 kV, 50 nA, and 5 µm for best possible matrix matches and avoidance of dead time bias. EPMA measurement of samples and repeated analysis of the reference monazite are performed at beam conditions of 20 kV, 100 nA, and 5 µm. Analysis of Pb and U on a PETL crystal requires YLg-on-PbMa and ThMz-on-UMb interference corrections. Offline re-calibration of the Th calibration on the Th-rich reference monazite, to match its nominal age, is an essential part of the protocol. EPMA-Th-U-Pb data are checked in ThO₂*-PbO coordinates for matching isochrones along regressions forced through zero. Error calculations of monazite age populations are performed by weighted average routines. Depending on the number of analyses and spread in ThO₂*-PbO coordinates, minimum errors <10 Ma are possible and realistic for Paleozoic monazite ages. A test of the protocol was performed on two garnet metapelite samples from the Paleozoic metamorphic Zone of Erbendorf-Vohenstrauß (NE-Bavaria, western Bohemian Massif).

1. Introduction

The phosphate monazite (LREE,Y,Th,U,Si,Ca)PO₄ was first described by Breithaupt (1829) at Freiberg in Saxony [1]. It occurs as an accessory phase in peraluminous granites and in Ca-poor and Al-rich metapsammopelitic rocks. In such micaschists and gneisses, monazite crystallises at metamorphic grades above the upper greenschist facies [2]. Monazite can experience dissolution–reprecipitation and recrystallisation during metamorphism and is also affected by fluid-driven processes [2,3,4,5,6,7]. In consequence, it appears as an incremental time-recorder and may even preserve multiple episodes of metamorphism [8,9,10,11]. This is in marked contrast to the properties of zircon where U-Pb isotope systematics are mostly resilient at metamorphic temperatures up to the granulite facies. Monazite has specific chemical and crystal-physical properties [2]. These characteristics and in situ analytical methods applicable in its preserved textural context make it a suitable target for geochronology [12,13,14]. Among several dating methods, the cost-efficient and rapid in situ Th-U-Pb dating by the electron probe microanalyser (EPMA) requires three important premises: (1) When it crystallises, the monazite incorporates almost no Pb, so that common Pb is negligible compared to the radiogenic Pb resulting from the decay of the abundant Th and minor U [12,15]. (2) Monazite displays extremely low diffusion rates for radiogenic Pb even at high temperatures [16,17]. (3) Monazite has self-annealing properties that prevent the accumulation of radiation damage and thus loss of the radiogenic Pb [18,19]. Since the early developments of the EPMA-Th-U-Pb monazite dating method [15,20,21], numerous protocols for Th-U-Pb and mineral-chemical analysis have been proposed, and their various strengths and disadvantages have been discussed in detail [22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39]. Although numerous pitfalls and problems related to the wavelength-dispersive spectrometry (WDS) analyses of bulk Pb and U at trace element concentrations for monazite Th-U-Pb dating by electron probe microanalysers have been described, this method receives considerable and widespread applications. Even when the resulting monazite ages are afflicted with a considerably larger error when compared to the LA-ICP-MS analyses giving the U-Pb and Th-Pb isotope ratios, the EPMA-Th-U-Pb monazite dating still bears a high potential for the recognition and age resolution of geological processes. A particular advantage of the EPMA compared to the other in situ methods is the analysis of fine-grained monazites < 20 µm, enclosed in other minerals or in veins. This is enhanced by the increasing knowledge of monazite microstructures observed in metapelitic and granitoid rocks and from petrological experiments [9,40,41]. It has been outlined by numerous authors that each electron probe microanalyser instrument requires an individual setup of the age dating measurement routine [22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39]. Therefore, a detailed analytical protocol for EPMA-Th-U-Pb monazite dating and mineral-chemical analyses with a JEOL-JXA-8230 instrument, which is more and more widespread in the audience, is presented in detail here, enclosing the recommendations for reference materials. The protocol was tested by an in situ monazite analysis in polished thin sections of two garnet-bearing micaschists from the Zone of Erbendorf-Vohenstrauß (ZEV), belonging to the Saxothuringian Domain of the Bohemian Massif. The ZEV has been thoroughly investigated as a geological target and site of the German Continental Deep Drilling Program Kontinentale Tiefbohrung (KTB). This gives a suitable frame for an interpretation of the EPMA Th-U-Pb monazite ages.

2. Materials and Methods

2.1. SEM-Automated Mineralogy Methods for Monazite Search

Grain sizes of pegmatite monazite >1 mm are known [2], but igneous and metamorphic monazite mostly ranges below 200 µm and dominantly at 10–100 µm. Thus, the optical microscope with polarised light is not sufficient for search and petrographic documentation of monazite in polished sections. A scanning electron microscope (SEM) is essential. SEM-automated mineralogy (SEM-AM) methods, based on an SEM equipped with image analysis software for interpreting backscattered electron images (BSE), and automated steering of the electron beam for single-spot energy-dispersive spectroscopy (EDS) can be applied [42]. Of special interest in the search for monazite are routines for the detection of rare (sparse) phases. Such routines combine a backscattered electron (BSE) grey colour value trigger and single-spot EDS spectral analysis. If the BSE grey colour value trigger is set at a specific range, then monazite, xenotime, zircon, most sulphides, and Fe- and Fe-Ti-oxides will be detected. Apatite that is also of interest will not be detected as an isolated separate rare phase. Only when apatite occurs in close spatial association with monazite as in corona, satellite, and replacement structures, it will be detected [9]. The method allows to selection of specific monazite grains for a detailed BSE image under optimised contrast and brightness for the resolution of internal grain structures.

2.2. Electron Probe Microanalyser (EPMA) Spectrometer Configurations

The configuration and combination of spectrometer type, crystal placement and assemblage, and X-ray counter types are essential for the accomplishment of the analytical tasks. As an EPMA is commonly used for very diverse and concurrent analytical tasks in several fields of materials and geoscience research, it is apparently problematic to find a configuration, which is optimal for the EPMA-Th-U-Pb monazite dating. Specific configurations for trace element and Th-U-Pb monazite dating were developed, and many analytical issues are presented [26,29,31,32,37,38,39]. As a consequence, a rarely realised configuration has been proposed by the JEOL Company with the installation of a JEOL-JXA-8230 at the Institute of Materials Science at TU Bergakademie Freiberg in Saxony. The details are listed in

Table 1. Spectrometer configuration of a JEOL-JXA-8230 instrument at TU Bergakademie Freiberg/Saxony, Institute of Materials Science, installed in 2019. Crystals marked in yellow are used for EPMA monazite analysis.

JEOL-JXA-8230	Code Spec	Crystal 1	Crystal 2	Crystal 3	Crystal 4	Type Counter
Spectrometer 1	XCE	PETJ	TAP	LDE1	LDE2	Ar-CH4-Flow
Spectrometer 2	XCE-H-type	PETH	LDE1H			Ar-CH4-Flow
Spectrometer 3	XCE	TAP	LDE2			Ar-CH4-Flow
Spectrometer 4	XCE-H-type	LIFH	PETH			Xe capsuled
Spectrometer 5	XCE-L-type	LIFL	PETL			Xe capsuled

. Synchronous measurements of O, N, and C for materials science analyses (e.g., steel alloys) are possible with LDE-type crystals distributed on 3 spectrometers, enclosing an H-type spectrometer with an Ar-CH₄ Flow counter. An analysis of monazite for EPMA-Th-U-Pb dating requires LIFH and PETL crystals with Xe-capsuled counters to avoid Ar-CH₄ spectra gas absorption edges at LREE, Pb, and U measurements as detailed by [26]. The configuration proposed here also allows the analysis of numerous elements in sulphides and silicates without changing the crystal during a single measurement cycle.

2.3. Criteria and Analysis of Reference Monazites

The analysis and selection of appropriate monazites for use as an internal reference material for (1) age controls of the EPMA-Th-U-Pb dating and (2) as matrix-matching reference materials are essential for establishing an analytical protocol. The principal idea is to constrain the age of the reference monazite by other geochronological methods, as with TIMS-U-Pb, SHRIMP-U-Pb, LA-ICP-MS-U-Pb, and LA-ICP-MS-Th-Pb isotope measurements [15,43,44,45,46,47,48,49,50,51]. As it is abundant in S-type peraluminous granites [52], monazite from Carboniferous granites is occasionally used as a reference material [15,53]. A reason is that U-Pb dating of the commonly associated zircon by various isotope methods can provide independent controls of the monazite age. A prerequisite is that inherited zircon or parts of such are not analysed [54]. Also, monazite from these granitoids can be dated by U-Pb and/or Th-Pb isotope analyses with various methods [55,56,57,58,59]. Another problem is that such reference monazites extracted from granites by heavy mineral separation are small, so that usually several individual grains are mounted together on a block. These grains are not the same as were analysed by the isotope measurements. Granites may also contain inherited or partly recrystallised monazite grains in some cases [54]. The statistical definition of the isochrones may be low or additionally hampered by excess or loss of common Pb during post-magmatic hydrothermal processes [46]. It should be possible to check the isochrone of the reference monazite by only a few single analyses during a single session for controls of EPMA-Th-U-Pb monazite dating. This requires that (1) the analyses of the reference monazite in ThO₂*-PbO coordinates should define an isochrone (ThO₂* is ThO₂ + UO₂ equivalents expressed as ThO₂) or (2) are almost homogeneous in PbO and ThO₂ (Figure 1).

However, even if isotope dating of potential reference monazites is successful, due to their variable PbO and ThO₂ along the isochrone, such monazites from granitoid and sometimes pegmatites [47,50] are not eligible for the indispensable coeval control of EPMA-Th-U-Pb age and ThO₂. Usually, for EPMA analysis, the Th is calibrated with a pure Th metal reference, various glass reference materials, or with a Th-bearing mineral such as thorite (ThSiO₄) or thorianite (ThO₂). None of these reference materials is matrix-matching to the phosphate monazite. Also, the Th metal reference material underwent fast degradation by oxidation, especially when it was not permanently stored in the EPMA sample chamber. The oxidation problem is also true for U metal reference material but with much less intensity. This may be handled by keeping the reference materials permanently in the EPMA sample chamber. In consequence, it is further promising to use monazite with a high and homogeneous Th content as a reference. As one can expect ThO₂ contents at 2–8 wt% for metamorphic monazites [2], a minimum ThO₂ content of ~10 wt% would be required for such a reference monazite. As igneous monazites may contain up to 25 wt% ThO₂, this would result in undercalibration.

Several monazites have been checked in this study if they could serve as reference material for both EPMA-Th-U-Pb age and Th controls. The MonAST-TUBAF and MonAST-HIF are multi-grain preparations of monazite, which are integrated in reference material blocks provided by ASTIMEX, and which belong to the reference materials collections of Technische Universität Bergakademie Freiberg (TUBAF) and Helmholtz Institute Freiberg of Resource Technology (HIF). The reference monazites Madmon [44] and Tsaratana are from various pegmatites in Madagascar and were taken from mineral collections. The Tsaratana monazite is from a pegmatite in the Ankazobé District in northern Madagascar. The reference monazite Moss is from a pegmatite near the city of Moss in the Østfold district in SE Norway [60]. The monazite Steenkampskraal is from the apparently first generation of monazite in the REE deposit in South Africa [61]. Monazite B is from the Harvard Mineral Collection Suite for Microanalysis Reference Materials. It is from granitic pegmatites in Topsham with TIMS-U-Pb ages from 275 to 269 Ma, which are related to the Brunswick granite at 278 ± 2 Ma [62]. These potential reference monazites were analysed by EPMA following a protocol presented in [63] and a new protocol as reported below.

Three selected reference monazites from this collection were further analysed by the TIMS-U-Pb method. The reference monazites labelled Madmon, Tsaratana, and Steenkampskraal were washed with 6 N hydrochloric acid and acetone. A grain was transferred with 6 N hydrochloric acid and ²⁰⁵Pb/²³⁵U-spike into small Savillex beakers. After dissolution at 100 °C for 96 h and subsequent evaporation to dryness at ca. 80 °C on a hotplate, 0.2 mL 3 N HCl was added. Chemical separation of lead (Pb) and uranium (U) was performed on 200 μL columns (ion-exchange resin AG 1 × 8, 100–200 mesh). U and Pb isotope ratios were obtained using a Finnigan MAT 261 mass spectrometer in static multicollector mode with simultaneous ion counting of ²⁰⁴Pb [64]. All isotopic ratios were corrected for mass fractionation (1.0 ± 0.3‰/a.m.u.), blank (appx. 1–3 pg), and initial lead using the model lead composition by Stacey and Kramers [65]. Zircon standard GJ-1 [66] and M257 [67] were repeatedly analysed in order to monitor reproducibility and accuracy. It yielded a concordia age at 607 ± 2 Ma (n = 5) and 562.6 ± 0.6 Ma (n = 4), respectively. U–Pb data were calculated using the program PBDAT [68], and isotope ratios were plotted using Isoplot [69], with error ellipses reflecting 2σ uncertainties.

2.4. Test of the EPMA-Th-U-Pb Monazite Dating Protocol

The protocol for EPMA-Th-U-Pb monazite dating reported below was subsequently tested on two samples from the Paleozoic metamorphic complex of the Zone of Erbendorf-Vohenstrauß (ZEV). The studied thin sections ZEV-1349 and ZEV-10PUL are not from recently taken samples but are historical thin section material belonging to a PhD thesis [70]. Both samples are garnet micaschists with kyanite and sillimanite. Glass covers were removed by careful grinding, and then the thin sections were refurbished by polishing with microdiamond-bearing liquid. Refurbished thin sections give a further potential application of the EPMA-Th-U-Pb monazite dating method. The locations of the samples are the pre-KTB exploration drilling site of Püllersreuth (ZEV-10PUL) and a field to the west of the open-air bath of Windischeschenbach (ZEV-1349). Both locations are 1.3 km apart within the northern part of the ZEV and ~5 km to the south of the KTB German Continental Deep Drilling Location. They belong to the ZEV to the west of the Falkenberg granite. Subsequent to the monazite analyses, it appeared continuative to consider the P-T conditions and P-T path recorded by the garnet-bearing assemblage in reference to the monazite stability field. Garnet chemical zonations and mineral-chemical compositions of biotite, muscovite, and plagioclase in the garnet-bearing assemblages were analysed by EPMA JEOL-JXA-8230 at 15 kV acceleration voltage, 20 nA beam current, and 2 µm beam diameter for Si, Al, Ti, Fe, Mn, Mg, Ca, Na, and K. Silicate reference minerals provided by ASTIMEX were used for instrument calibration. The ZAF corrections provided by JEOL were applied.

3. Results

3.1. ThO₂-PbO, Compositions and Ages of Reference Monazites

The monazite Tsaratana displays an isochrone with an EPMA-Th-U-Pb-weighted average age of 531 ± 5 Ma and a concordant TIMS-U-Pb age of 528 ± 1 Ma (Figure 1 and Figure 2;

Table 2. Electron probe analyses of selected reference monazites (Mnz) with homogeneous compositions in ThO₂ and PbO, as shown in Figure 1. Number of analyses in brackets. Data are given in wt%. PbO, UO₂ with interference corrections, and ThO₂ with offline correction. ThO₂* is calculated from Th and U [20]. EPMA-Th-U-Pb monazite ages (in Ma) are weighted averages.

Reference Mnz	P2O5	La2O3	Ce2O3	Pr2O3	Nd2O3	Sm2O3	Gd2O3	Y2O3	SiO2	CaO	ThO2	UO2	PbO	Total	ThSuz	ThO2*	Age		2σ
Steenkamp (134)	28.88	12.79	28.17	2.65	10.46	1.41	1.14	2.10	0.95	1.17	8.32	0.24	0.40	98.70	7.82	9.15	1025	±	2.7
MonAST-HIF (24)	27.86	9.48	25.54	2.81	11.00	3.20	2.67	2.00	1.58	0.87	9.83	0.22	0.21	97.28	9.28	10.56	479	±	3.8
Madmon (25)	25.79	7.84	25.49	3.64	14.92	4.24	1.88	0.92	2.98	0.13	10.92	0.43	0.27	99.44	10.95	12.32	516	±	3.0

and

Table 3. U-Pb-TIMS analyses from selected reference monazites with calculation of U-Pb concordia ages and corresponding EPMA-Th-U-Pb ages in Figure 1. Numbers of EPMA-Th-U-Pb analyses in round brackets.

Sample	207Pb/235U	%	206Pb/238U	%	Rho Error	U-Pb-Concordia and EPMA-Th-U-Pb Ages
Tsara-4252	0.68386	0.59	0.08555	0.35	0.59
Tsara-4251	0.68084	1.21	0.08515	0.64	0.53	528.6 ± 1.1 Ma, MSWD 0.0066, U-Pb-TIMS
Tsara-4251	0.67901	0.72	0.08491	0.47	0.71	531 ± 5 Ma, MSWD 1.4, EPMA-Th-U-Pb (100)
Tsara-4254	0.68049	0.86	0.08504	0.51	0.59
Tsara-4257	0.68394	0.41	0.08558	0.29	0.69
Tsara-4250	0.67273	0.44	0.08442	0.27	0.62
Tsara-4256	0.66944	0.87	0.08408	0.41	0.53
Tsara-4256	0.66726	1.09	0.08379	0.33	0.37
Steen-4407	1.7599	0.41	0.17378	0.29	0.74	1030.8 ± 3.0 Ma, MSWD 1.03, U-Pb-TIMS
Steen-4408	1.7492	0.59	0.17353	0.16	0.32	1015.2 ± 1.7 Ma, MSWD 0.55, U-Pb-TIMS
Steen-4410	1.7548	0.37	0.17293	0.28	0.79
Steen-4409	1.7247	0.41	0.17196	0.31	0.79	1025 ± 3 Ma, MSWD 2.1, EPMA-Th-U-Pb (137)
Steen-4402	1.7249	0.85	0.17126	0.51	0.64
Steen-4403	1.7182	0.27	0.17077	0.22	0.84
Steen-4404	1.7081	0.47	0.16974	0.42	0.89
Madmon-4615	0.66305	0.36	0.08351	0.22	0.68	515.71 ± 1.7 Ma, MSWD 1.6, U-Pb-TIMS
Madmon-4616	0.66068	0.48	0.0834	0.42	0.89	510.4 ± 6.6 Ma, MSWD 1.6, U-Pb-TIMS
Madmon-4612	0.65832	0.52	0.08321	0.49	0.95	496 ± 10 Ma, U-Pb SHRIMP [44]
Madmon-4611	0.65851	0.45	0.08316	0.39	0.89	516 ± 3 Ma, MSWD 1.07, EPMA-Th-U-Pb (28)

). Its variable Th contents along the isochrone are unsuitable for a Th calibration. The same is true for monazite B with an EPMA-Th-U-Pb isochrone-weighted average age at 273 ± 3 Ma, which matches the reported TIMS-U-Pb ages at 275–269 Ma [62]. The monazite grains MonAST-TUBAF in a multi-reference block purchased from Astimex provide an EPMA-Th-U-Pb-weighted average age of 1722 ± 18 Ma with fairly low ThO₂* and Th_SUZ. Data from a large single-grain monazite Moss plots along an isochrone of 879 ± 7 Ma. Monazite grains MonAST-HIF in a multi-reference block purchased from Astimex gives 479 ± 4 Ma at fairly high and homogeneous ThO₂ contents of 9.83 wt%, but the small single grains allowed no further age control by isotope methods. The Steenkampskraal REE mine in RSA provided sufficient monazite grains with an EPMA-Th-U-Pb-weighted average age of 1025 ± 3 Ma at a ThO₂ content of 8.32 wt% (Figure 1 and Figure 2). The TIMS-U-Pb dating of the Steenkampskraal monazite yielded a concordant 1031 ± 3 Ma age (Table 3). Within this suite, the Madmon monazite from a pegmatite at Madagascar appears as the comparably best-suited reference material with a weighted average EPMA-Th-U-Pb age of 516 ± 3 Ma, adjusted by offline re-calibration correction of Th (described below). Several isotope measurements of Madmon yielded ages between 496 ± 9 Ma with SHRIMP-U-Pb [44,46] and 516 ± 2 Ma (TIMS-U-Pb, this study). Its comparably high ThO₂ contents at ~11.0 wt% with a homogeneous spatial element distribution make this monazite also suitable for the calibration of Th (Figure 1). However, other fragments of this large pegmatite monazite may have slightly different ThO₂ contents.

3.2. Setup of EPMA Protocol for Monazite Th-U-Pb Dating

As exemplified in numerous publications dealing with EPMA-Th-U-Pb dating, the analytical setup and protocol need several precedent tests [15,22,26,27,30,34,35,36,37,38,39,45,55,71,72,73,74]. Each instrument requires a specified and individually adapted setup of the protocol.

All the protocols consider the analysis of Pb as a trace element at <1 wt% in many monazites as the most important issue. The analysis of Pb has to be optimised in terms of beam conditions such as acceleration voltage, beam current and beam diameter, choice of peak and background positions as well as measurement time, and choice of spectrometer and crystal in order to minimise the error. Accurate analysis of the trace element Pb in monazite requires an increase in the peak-to-background ratio (P/B). Long counting times can significantly lower the detection limit and increase the peak-to-background ratio [29,75,76]. However, an extensive counting time may cause the sample to be damaged by the beam. Long counting times are also subject to beam stability. Appropriate tests of beam conditions and counting time have been designed by [37]. The own tests and longtime experience yielded 20 kV, 100 nA, 5 µm, and 300 s on Pb peak and 70 s on each positive and negative background. When measurement times for Th and U analyses on the same spectrometer are added, this gives about ~12 min total measurement time. Dry samples were coated 2 times by the evaporation of a carbon thread loaded with 0.18 g/m of C. In most cases, the carbon coating at ~10 nm will stay resistant, even when a reduced beam diameter of 3 µm is applied to small monazite grains. More parameters of the Pb analysis with either PETJ, PETH, or PETL crystals in spectrometers with a capsuled Xe X-ray counter are then set by the given properties in detailed linescans (Figure 3). In a majority of yet-published protocols, the PbMa peak has been chosen, but the alternative PbMb was also proposed [23,77]. At the PbMb peak, the intensities are about 30% lower compared to PbMa. But on PbMb appear no interferences and it allows much lower background offsets, reducing potential errors of the net intensities resulting from linear background approximation of the curved bremsstrahlung spectrum [27]. For the protocol presented here, the PbMa has been chosen, despite the interference by a YLg line. An argument supporting this choice is the observation that the PbMb peak is comparably sharp. This results in an elevated standard deviation of the counting rates, thus, in a higher error, as an analysis of the PbMa. Profound and extensive considerations about the background setting and curvature of the background for measuring PbMa have been presented [26,28,29]. Significant differences in the PbMa and PbMb background situations have been stated among high-Th and low-Th monazites [28,29]. The choice of the positive background position for PbMa analysis should be far beyond the ThMz2 and LaLa1 peak and appears as not problematic (Figure 3a). Several options exist for the negative background position for PbMa analysis. The nearest position in reference to the PbMa peak is a depression of the background between the ThMz and CeLa1 peaks (Figure 3a). An intermediate option is a background depression between the PbMb and LaLb1 peaks, and a far position is a further background position on the negative slope of the LaLb1 peak. This position has been chosen for the protocol presented here (Figure 3a). Another issue is the choice of the peak for the U measurement. Usually, the U is <2 wt% in monazites, in the presence of Th contents of 4–8 wt% for metamorphic monazites and <10 wt% in pegmatite and some igneous monazites [2]. As a consequence, the UMa peak cannot be recommended for U measurement, as it is positioned at the negative slope of a massive ThMb peak (Figure 3b). The UMb peak appears as an alternative. The positive background position for UMb measurement could be a depression between the ThMa and ThMb peaks (far position) or a near background depression at the negative slope of the ThMb peak (Figure 3b).

The measurements of PbMa and UMb are affected by interferences, which need corrections [15,23,25,26,27]. Depending on the choice of the spectrometer and crystal, the positive slope of the ThMz1 peak may affect the PbMa peak position [26]. This is not the case with the PETL crystal proposed here (Figure 4a). Furthermore, the positive slope of a YLg1 peak may interfere with the PbMa peak position, as is the case for PETL at spectrometer 5 of the JEOL-JXA-8230 (Figure 4a). Also, the elongated positive ThMg slope superposes the UMb peak (Figure 4b).

Analysis of the LREE and HREE is also hampered by numerous potential interferences [23,25,26,27]. Some of these interferences can be avoided by analysis with only a single background position and/or consideration of the Lb peaks (

Table 4. Protocol for monazite analysis with JEOL-JXA-8230 electron probe microanalyser, hosted at Institute of Materials Science at TU Bergakademie Freiberg/Saxony in Germany. Counts per second (cts/s) on peak and positive and negative background (BKG+, BKG−) refer to the reference monazite Madmon. JEOL 1σ error is calculated from cts/s rates and compositions. Calibration beam conditions are 20 kV, 50 nA, and 5 µm. Measurement beam conditions are 20 kV, 100 nA, and 5 µm. Reference materials for calibration are from ASTIMEX (AST) and from the Smithsonian Institution (Smi).

JEOL-JXA-8230				EPMA Measurement Specifications								Madmon Reference Monazite
					SCA		Spectrometer Pos. (mm)			Dwell t. (s)		Counting Rate (cts/s)			D. L.	JEOL Error 1σ		Ox.wt%
Ele.	Line	Crystal	Reference	nA	Gain	Volt	Peak	BKG−	BKG+	Pk	BKG	Pk net	BKG−	BKG+	µg/g	(cts%)	(wt%)	Corr.
Th	Ma	5-PETL	Madmon-2	100	32	1734	132.505	3.348	4.000	30	10	6749.8	318.3	252.0	82	0.230	0.025	10.92	Th
U	Mb	5-PETL	Uranium-AST	100	32	1736	118.990	3.495	10.282	80	30	233.8	445.3	318.0	57	1.730	0.008	0.43	U
Pb	Ma	5-PETL	Crocoite-AST	100	32	1770	169.249	13.258	5.353	300	70	64.5	134.9	81.4	36	1.690	0.004	0.27	Pb
Y	La	1-PETJ	YPO4_Smi	100	16	1626	206.188	2.796	2.109	50	20	97.8	93.4	90.2	136	2.470	0.021	0.92	Y
Y	La	2-PETH	YPO4_Smi	100	16	1622	206.555	2.116	1.718	50	20	168.3	495.4	408.7	164	2.898	0.024	0.86	Y
Si	Ka	3-TAP	Diopside_AST	100	8	1710	77.299	0.000	5.042	40	20	2837.8	0.0	421.1	36	0.310	0.009	2.98	Si
Al	Ka	3-TAP	Plagioclase-AST	100	8	1716	90.586	8.290	8.109	40	20	-43.0	566.5	239.2	34	n.a.	n.a.	0.00	Al
Ca	Ka	1-PETJ	Diopside_AST	100	4	1706	107.429	3.903	2.452	50	25	118.3	278.5	240.5	36	2.790	0.004	0.13	Ca
Ca	Ka	2-PETH	Diopside_AST	100	8	1664	107.472	3.185	1.796	50	25	304.3	4614.4	3546.7	43	4.290	0.006	0.14	Ca
P	Ka	1-PETJ	Apatite_AST	100	16	1620	196.795	2.817	4.408	20	10	5208.9	83.7	91.7	108	0.311	0.079	25.79	P
P	Ka	2-PETH	Apatite_AST	100	16	1632	197.123	2.844	3.879	20	10	9871.8	964.3	587.4	178	0.242	0.061	25.41	P
La	La	1-PETJ	LaPO4_Smi	100	32	1586	85.278	0.000	3.009	30	10	2693.3	0.0	629.1	238	0.390	0.032	7.84	La
Ce	La	1-PETJ	CePO4_Smi	100	32	1564	81.936	1.553	1.428	30	10	7914.6	1017.9	865.2	308	0.230	0.061	25.49	Ce
La	La	4-LIFH	LaPO4_Smi	100	32	1716	185.199	0.000	3.441	30	10	2164.8	0.0	225.8	164	0.404	0.031	7.81	La
Ce	La	4-LIFH	CePO4_Smi	100	32	1720	177.933	2.604	3.076	30	10	8519.5	336.5	278.4	160	0.204	0.052	25.44	Ce
Pr	Lb	4-LIFH	PrPO4_Smi	100	32	1708	156.806	0.000	5.041	50	20	880.8	0.0	484.3	198	0.582	0.021	3.64	Pr
Nd	Lb	4-LIFH	NdPO4_Smi	100	32	1698	150.413	6.012	4.877	30	15	3984.6	568.1	483.6	213	0.320	0.047	14.92	Nd
Sm	Lb	4-LIFH	SmPO4_Smi	100	32	1697	138.624	9.497	5.763	50	20	1217.4	865.4	565.7	187	0.590	0.052	4.24	Sm
Gd	Lb	4-LIFH	GdPO4_Smi	100	32	1690	128.008	1.295	1.380	50	20	568.1	846.0	888.3	197	1.228	0.023	1.88	Gd

). The analysis of the HREE should enclose in each case the Gd. Analysis of Eu is hampered by interferences. The analysis of Dy and Er appears suitable despite interferences, as there are low contents in monazites but still markedly above the detection limits. For the other HREE in monazites, the contents are next to the detection limits of the EPMA, and their measurement is not recommended.

Synthetic REE-orthophosphates, obtained from the Smithsonian Institution [78], can be used as reference materials for P, REE, and Y calibration. When carefully checked it turned out that these standards contain <1 wt% but detectable Pb in varying contents as residua from their synthesis in a Pb-flux [79]. REE-orthophosphates crystallised by Pb-free flux methods have been also proposed [80] but are rarely available. REE-fluorides and REE-ultraphosphates appear as alternatives but will cause serious problems (see below). The reference material for uranium analyses was U-metal; crocoite was used for Pb. Diopside was used as a reference for Ca and Si analyses and plagioclase for Al. As Al should not appear in monazite, it can be used for control if the beam fully hits the monazite, or if the beam excitation bulb affects neighbouring phases.

For the calibration of Th, the proposed protocol uses the Madmon reference monazite. This choice is based on the fact that Madmon contains the most ThO₂ (~11 wt%) among the analysed reference monazites and the homogeneous distribution of ThO₂ among the chosen polished segment of a grain. A further argument is that the age of the Madmon is within the expected Paleozoic age range of the samples.

The calibration procedure has been programmed as a separate serial analysis session with beam conditions at 20 kV acceleration voltage, 50 nA beam current, and a beam diameter of 5 µm. When REE-fluorides or REE-ultraphosphates are used for the calibration of REE, a beam current of 50 nA will lead to damage to the reference material. Thus, a lower beam current at 10 nA appears as suitable [63]. As a consequence, the calibration conditions and the sample analysis conditions of the beam current differ considerably (10 nA vs. 100 nA). As the JEOL spectrometers have constant dead time integration factors, this may result in uncertainties and bias. Therefore, the use of beam-resistant REE-orthophosphates and high beam currents (50 nA) for calibration are recommended here despite the potential traces of Pb in that reference material [78,79]. Integration of Pb traces in the REE-orthophosphate reference compositions influences the ZAF corrections, which led to totals above 100 wt% and therefore was not considered.

The measurement sequence in each spectrometer is also a matter of discussion. For measurement of REE with 4PETH crystal, the sequence is Gd–Sm–Nd–Pr–Ce–La or vice versa, which follows the range of the REE La and Lb peaks positions in one direction from increasing to decreasing spectrometer positions, or vice versa. The measurement range of the elements Th, U, and Pb is critical for the calculation of the chemical CHIME ages. When the long required measurement time for Pb is at the beginning of the sequence, this may reduce the risk of decreasing absorbed current on the sample when the carbon coating gets progressively damaged. On the other hand, the activation of X-ray emission from Pb requires some time, so higher counting rates on the PbMa peak position can be expected when Pb measurement occurs at the end of the analytical sequence. For realising an optimal peak-to-background counting rate for the Pb, its analysis at the end of the measurement sequence and sample exposure to the beam <12 min have been adjusted (Table 4).

3.3. Interference and Offline Corrections, Age Calculation, and Error Estimates

A correction of PbMa following the interference with the positive slope of a YLg1 peak may be corrected as has been proposed by [15]. The correction factor is preferably established by analysis of PbMa in several reference materials, which contain Y but no Pb, as Y-garnet, Y-glass, or Y-metal (Figure 4a) and not only by virtual WDS [81]. An interference correction factor C_F for YLg1 on PbMa can be estimated as C_F = PbO wt% in Y-reference material/Y₂O₃ wt% in Y-reference material. The C_F is 0.006 in the studied configuration. In the protocol proposed here, the applied correction for PbO is then PbO_corrYonPb = PbO_measured − (C_F × Y₂O_3measured). In a similar way, the C_F for a correction of the elongated positive ThMg slope on the UMb peak has to be established by measurements of UMb in Th metal and/or ThO₂-bearing reference materials (Figure 4b). The C_F is 0.005 for ThMg-on-UMb in the studied configuration.

Calculation of the EPMA-Th-U-Pb monazite age follows the interference corrections on PbO and UO₂. Usually, the age calculations are performed for each single monazite EPMA analysis as proposed by [15,20,22,43,82]. These age calculations follow several iteration steps. In the first step, a pure ²³²Th-²⁰⁸Pb age is calculated by assuming that all interference-corrected Pb corresponds to ²⁰⁸Pb. The following iterative calculation step calculates the amount of ²⁰⁷Pb and ²⁰⁶Pb resulting from the decay of ²³⁵U and ²³⁸U according to that ²³²Th age. This considers the bulk UO₂ with the fixed isotope ratio of ²³⁵U/²³⁸U (0.00725). In the following calculation step, the ²⁰⁷Pb and ²⁰⁶Pb are subtracted from the bulk Pb, and a new ²³²Th-²⁰⁸Pb age is calculated. Then a new iterative calculation procedure is performed. Usually, the final Th-U-Pb monazite age is reached after 2 or 3 iteration steps [35,83,84]. More specialised age calculations involving Th/Pb and U/Pb coordinates have been introduced by other authors [24,85,86,87].

As reported in several publications, the EPMA measurement of Pb at trace element levels is affected by uncertainties related to the background, of which curvature appears as dependent on the Th contents [29,36]. Also, the calibration of Th for EPMA Th-U-Pb monazite dating is hampered by non-matrix-matching reference materials (e.g., Th metal), which also underwent fast oxidation. This will not affect the low 1σ error of the Th measurement (Table 4) but its values of absolute concentrations. The instrument drift for P, La, Ce, Nd, Sm, and Y, assumed to be linear over time, was corrected against the reference Madmon. Then, the EPMA-Th-U-Pb monazite age calculation proposed here involves a preliminary step with an offline Th correction. Each analytical session involves measurement of the age reference monazite Madmon [46] at the beginning, midterm, and end (10–15 analyses). Each reference monazite analysis provides a single age when calculated according to the method proposed by [15]. As usually, no outliers occur, EPMA-Th-U-Pb arithmetic mean age from all these reference monazite analyses is then calculated. Mostly this EPMA-Th-U-Pb arithmetic mean age is slightly older than the nominal U-Pb age (516 Ma) of the reference monazite established by analysis with SHRIMP and TIMS methods (Table 3). An offline correction of the Th with a corresponding factor (>1.00) will then be used to adjust the Th-U-Pb age to the nominal age of the reference monazite. This offline correction factor is incrementally approached by a stepwise trial increase in the factor, followed by an age recalculation until the nominal reference monazite age is accomplished. Subsequently, this Th offline correction factor is then applied to all ThO₂ analyses within the session, and single analyses sample monazite ages were calculated. Subsequently, the data allow further age determination using the ThO₂*–PbO isochrone method (CHIME), where ThO₂* is the sum of the measured ThO₂ plus ThO₂ equivalent to the measured UO₂ [20,21]. This age is based on the slope of a regression line in ThO₂* vs. PbO coordinates forced through zero. As the calculation of the regression line provides an underestimation of the error, in the third step, the weighted average ages for monazite populations were calculated from the single analyses defining the regression line using Isoplot 3.0 [69]. Such weighted average ages were also calculated from the ages of single analyses when a large monazite grain allowed several analyses, which give a homogeneous distribution of the single ages, and from clusters composed of many small monazite grains. The isochrone ages and the weighted average ages should usually coincide with the error. Furthermore, the monazite ages can be explored by the “zircon age extraction” routine given by [69]. A 1σ error deduced from the counting statistics (JEOL error), and an error ε_Pb = √(Cts/s _PEAK + Cts/s _BKG)/(Cts/s _PEAK − Cts/s _BKG) was propagated to an error in Pb element %. For Pb, the error in element% is ~0.004 (recalculated from the JEOL error) or ~0.001 (recalculated from ε_Pb) for the reference monazite Madmon (~0.25 wt% Pb). We applied an error in Pb element% of 0.004 to all analyses, which propagates for the reference monazite Madmon with ~516 Ma typically to ±16 Ma (2σ), and for Carboniferous monazites (~0.085 wt% Pb) to 30–40 Ma (2σ). The implementation of suitable reference monazite and the corresponding Th offline correction result in the comparability of monazite age data from numerous measurement sessions and from different instruments within project campaigns [88].

3.4. Test of the EPMA Protocol

The SEM-AM investigation revealed some differences between the two test samples from the Zone of Erbendorf-Vohenstrauß (ZEV). Monazite grain sizes at 50 wt% of cumulative grain size classes (MD50) are small at 32 µm (sample ZEV-10PUL) and 25 µm (sample ZEV-1349), as reported in

Table 5. Results of scanning electron microscope automated mineralogy (SEM-AM) analysis of monazite-bearing garnet micaschists from the Zone of Erbendorf-Vohenstrauß, Saxothuringian Zone, NE-Bavaria. MD50 is monazite grain size at a cumulative 50 wt% of all monazite grain size classes. Monazite associations refer to grain boundary contacts to adjacent minerals.

Sample (Thin Section 20 mm × 20 mm)	ZEV-10PUL	ZEV-1349
Monazite grain counts	142	261
Monazite area in µm2	63,932	61,711
Xenotime grain counts	73	111
Zircon grain counts	831	1124
MD50 monazite in µm	32	25
Monazite association with biotite	42.8%	51.4%
Monazite association with muscovite	16.9%	17.0%
Monazite association with quartz	20.2%	2.4%
Monazite association with albite	1.1%	12.7%
Monazite association with K-feldspar	1.2%	0.3%
Monazite association with plagioclase	7.7%	1.3%
Monazite association with garnet	1.2%	0.3%

. Monazite areas in µm² are quite similar, as is the occurrence of zircon and xenotime grains. The monazite associations represent the minerals in contact with the monazite grains. Biotite and muscovite are in both samples the dominant minerals in contact with monazite. Contrasting association relationships are observed with regard to quartz and albite. In sample ZEV-10PUL, the K-feldspar, plagioclase, and garnet are markedly more associated with monazite than in sample ZEV-1349 (Table 5). An application of the EPMA-Th-U-Pb measurements following the presented analytical protocol (Table 4) yielded two distinct age populations at 461 ± 7 Ma and 352 ± 9 Ma (ZEV-1349), and 461 ± 9 Ma and 343 ± 11 Ma (ZEV-10PUL) when isochrones in ThO₂* vs. PbO and weighted average ages [69] are considered (Figure 5; Table S1). Both age populations appear also when the age extraction routine is applied [69]. In detail, when larger monazite grains allowed several single measurements, the older Ordovician ages dominate and the younger Early Carboniferous ages occur sporadically with the older grains. This may indicate a partial recrystallisation of the older monazites within distinct domains [89]. In both samples with similar low modes of garnet, the younger monazites have slightly higher Y₂O₃ contents (Figure 6a) at equivalent cheralite substitution trends in Th+U vs. Ca (Figure 6b). The compositional ranges (e.g., 4–8 wt% ThO₂*) of both monazite age populations signal a crystallisation during metamorphic events [2]. In such a case of a polymetamorphic evolution, the question arises as to which of the two events may be associated with the crystallisation of the garnet-bearing assemblage. Several studies of monazite age populations in polymetamorphic areas revealed that garnet crystallisation at amphibolite-facies conditions may not coincide with a dominant monazite age population [8,9]. The ages of enclosed monazites in garnet in comparison to matrix monazite ages may give a hint in special cases [8,90,91], but in the studied samples, no monazites are enclosed in garnet. For further evaluation, it appears continuative to consider the P-T conditions and P-T path recorded by the garnet-bearing assemblage in reference to the monazite stability field [90]. The almandine-dominated garnets display zonations in spessartine, pyrope, and grossular components (Figure 6c,d). Corresponding core-to-rim trends in XMg-XCa display for sample ZEV-1349 first increasing XMg at decreasing XCa, followed by decreasing XMg at constant XCa. Garnet zonations in sample ZEV-10PUL show decreasing XMg at constant XCa (Figure 6e). The lower garnet XCa and plagioclase anorthite contents in sample ZEV-1349 signal a lower bulk rock Ca content when compared to sample ZEV-10PUL. The garnets display no indications for resorption. P-T conditions of garnet core-to-rim crystallisation were calculated with analyses from coexisting biotite, muscovite, and plagioclase using the cation-exchange and net-transfer thermobarometers garnet–biotite–muscovite–plagioclase–quartz GBMP [92]. The single clockwise P-T segments recorded by the garnet-bearing assemblages in both samples overlap in parts and start at 645 °C/9.2 kbar, followed by decompression–heating to the thermal maximum at 662 °C/7.7 kbar and finished after decompression–cooling at 598 °C/4.3 kbar (Figure 7,

Table 6. Garnet, biotite (Bt), muscovite (Ms), and plagioclase (Pl) mineral chemistry (in mole fractions X) and Si⁴⁺ per formula unit in monazite-bearing garnet micaschists in the Zone of Erbendorf-Vohenstrauß. Temperatures (T) and pressures (P) of P-T evolution recorded by garnet zonations by applications of geothermobarometers [92].

	Garnet			Bt	Ms	Pl	T/P
Sample Garnet	XMg	XCa	XMn	XMg	Si4+	XAn	°C/kbar
ZEV-1349 core	0.111	0.046	0.158	0.432	3.16	0.109	645/9.2
ZEV-1349 mid	0.133	0.019	0.099	0.421	3.16	0.109	653/6.5
ZEV-1349 rim	0.102	0.021	0.137	0.433	3.23	0.147	598/4.3
ZEV-10PUL-core	0.146	0.033	0.102	0.449	3.14	0.143	662/7.7
ZEV-10PUL-mid	0.127	0.035	0.127	0.449	3.14	0.195	637/5.8
ZEV-10PUL-rim	0.095	0.037	0.182	0.449	3.14	0.203	597/4.7

). This segment of P-T evolution during garnet crystallisation can be regarded in reference to the stability field of monazite, as outlined at different bulk rock contents as a function of Ca wt% [93,94,95]. The stability field of monazite is shifted towards lower temperatures with decreasing bulk rock Ca and increasing Al (Figure 7a).

4. Discussion

The intense geochronological and petrological studies on the metabasites and metapelites of the ZEV in the course of the KTB Deep Drilling Program revealed indications of polymetamorphism. An Ordovician (>480 Ma) high-temperature high-pressure event at 10–14 kbar/600–700 °C was followed by partial anatexis and a ~480 Ma pegmatite intrusion at low pressures [96]. A Devonian metamorphism at 6–8 kbar/600–700 °C [97,98] is documented by 390–370 Ma hornblende-K-Ar ages in metabasites [99]. The 370–320 Ma mica-K-Ar ages from paragneisses give evidence of a multiphase cooling history subsequent to the Devonian metamorphism [100,101], followed by the ~315 Ma intrusive emplacement of the Falkenberg granite [102]. Post-Permian K-Ar mica ages signal further ductile to brittle tectonic events during the uplift [101].

It appears that the two monazite age populations, the older one at ~461 Ma and the younger one at ~352/343 Ma, do not coincide with the reported K-Ar ages, interpreted as metamorphic cooling ages [100,101]. The P-T conditions recovered from the garnet-bearing assemblages clearly coincide with the reported P-T conditions of a Devonian metamorphism [97,98], but the monazite ages do not match a Devonian (<417 Ma) but a Carboniferous < 358 Ma) age. As a consequence, these monazite ages cannot be interpreted as dating the Devonian metamorphism and its subsequent cooling. Furthermore, the monazite age populations apparently cannot be directly related to the garnet crystallisation event. Potentially, the P-T conditions recorded by the garnet-bearing assemblage are within the monazite stability field but only for Al-rich and Ca-poor bulk rock compositions. For the more Ca-rich bulk rock compositions, no monazite crystallisation can be expected at the given garnet crystallisation P-T conditions (Figure 7). This leads to an interpretation of the monazite age data in terms of temporally separated monazite and garnet crystallisation events: An early ~461 Ma monazite crystallisation appeared subsequent to an Ordovician high-temperature metamorphic stage [96]. A high-pressure stage of this Ordovician metamorphism is only evident in metabasites and not in paragneisses [96]. Then followed a Devonian ~390–370 Ma garnet crystallisation with no contemporaneous monazite crystallisation. A later Carboniferous 352–343 Ma monazite crystallisation appears as a separate event (Figure 7a). This interpretation is supported by the observation that the ~352–343 Ma monazite age population is minor and partly recrystallised from the dominant ~461 Ma monazite grains. Within the same temporal sequence of crystallisation events, a crystallisation of the Carboniferous 352–343 Ma monazites at higher temperatures may be indicated by their higher Y₂O₃ contents when compared to the Ordovician (~461 Ma) monazites (Figure 6a and Figure 7b).

It appears that the two monazite age populations, the older one at ~461 Ma and the younger one at ~352/343 Ma do not coincide with the reported K-Ar ages, interpreted as metamorphic cooling ages [99,100]. The P-T conditions recovered from the garnet-bearing assemblages clearly coincide with the reported P-T conditions of a Devonian metamorphism [97,98], but the monazite ages do not match a Devonian (<417 Ma) but a Carboniferous (<358 Ma) age. As a consequence, these monazite ages cannot be interpreted as dating the Devonian metamorphism and its subsequent cooling. Furthermore, the monazite age populations apparently cannot be directly related to the garnet crystallisation event. Potentially, the P-T conditions recorded by the garnet-bearing assemblage are within the monazite stability field, but only for Al-rich and Ca-poor bulk rock compositions. For the more Ca-rich bulk rock compositions, no monazite crystallisation can be expected at the given garnet crystallisation P-T conditions (Figure 7). This leads to an interpretation of the monazite age data in terms of temporally separated monazite and garnet crystallisation events: An early ~461 Ma monazite crystallisation appeared subsequent to an Ordovician high-temperature metamorphic stage. A high-pressure stage of this Ordovician metamorphism is only evident in metabasites and not in paragneisses [96], then followed a Devonian ~390–370 Ma garnet crystallisation with no contemporaneous monazite crystallisation. A later Carboniferous 352–343 Ma monazite crystallisation appears as a separate event (Figure 7a). This interpretation is supported by the observation that the ~352–343 Ma monazite age population is minor and partly recrystallised from the dominant ~461 Ma monazite grains. Within the same temporal sequence of crystallisation events, a crystallisation of the Carboniferous 352–343 Ma monazites at higher temperatures may be indicated by their higher Y₂O₃ contents when compared to the Ordovician (~461 Ma) monazites (Figure 6a and Figure 7b).

5. Conclusions

A protocol of electron probe microanalysis for monazite Th-U-Pb dating and mineral chemistry characterisation with a JEOL-JXA 8230 instrument has been developed. This protocol is part of a monazite detection, measurement, and data interpretation methodology:

(1)

Monazite microstructural locations, reaction textures, grain sizes, associations, and locking to other minerals in polished thin sections are documented by SEM automated mineralogy methods and SEM-BSE imaging.

(2)

Subsequent EPMA measurements involve the elemental calibration in a serial analysis routine of REE-orthophosphates and a suitable Th-rich reference monazite (Madmon) with known TIMS-U-Pb age. The choice of reference material for elemental calibration should allow high beam current at 20 kV, 50 nA, and 5 µm, for best possible matrix matches and avoidance of dead time bias.

(3)

For Th calibration and controls of ages, the recommended reference monazite should have high and homogeneous Th contents at corresponding homogeneous Pb contents. These conditions are matched by the pegmatite reference monazite Madmon with a TIMS-U-Pb isotope age at ~516 Ma. Reference monazites with heterogeneous Th and corresponding Pb contents along an isochrone appear to be suitably constricted for age control, but not for the Th element measurement calibration.

(4)

Measurement of samples and repeated analysis of the reference monazite are performed following the proposed protocol (Table 4) at beam conditions of 20 kV acceleration voltage, 100 nA beam current, and 5 µm beam diameter with a total measurement time of ~12 min on carbon-coated samples. Subsequent to the YLg-on-PbMa and ThMz-on-UMb interference corrections, the protocol involves an offline re-calibration of the Th calibration on the reference monazite, so that the mean of the Th-U-Pb ages from the reference monazite matches its nominal age. The Th offline re-calibration is then applied also to the sample analyses. This procedure allows comparability of data in campaign-style studies.

(5)

The EPMA-Th-U-Pb data from single samples should be checked in ThO₂*-PbO coordinates for matching isochrones along regressions forced through zero. Error calculations of the monazite age populations are then performed by weighted average and age extraction routines [69]. Depending on the number of analyses and their spread in ThO₂*-PbO coordinates, minimum errors of <10 Ma are possible and realistic for Paleozoic monazite ages.

As exemplified by a test of the protocol, the interpretation of EPMA-Th-U-Pb monazite ages from garnet-bearing metapelites in terms of geological events should involve a P-T quantification and a P-T path reconstruction. Then P-T data may be considered in reference to the monazite stability field for interpretations on the relative temporal sequence of monazite and garnet crystallisation. However, when more than one monazite age population appears, this may lead to concurrent interpretation models. Compositional layering on the millimeter scale and crystallisation in microstructural domains, as is typical for progressively deformed metapelites, have to be considered for corresponding analytical studies. Due to this mineralogical heterogeneity even within a limited bulk compositional range, single samples can provide an incomplete record. Studies on monazite and garnet in polymetamorphic terrains in consequence require analysis of multiple samples in the frame of regional campaign-style studies.