Phase equilibrium modelling and zircon-monazite geochronology of HT-UHT granulites from Kambam ultrahigh-temperature belt, south India

ABSTRACT The Madurai block, the largest crustal block in the Precambrian Southern Granulite Terrane of south India, preserves rare assemblages of high- to ultrahigh-temperature metamorphic rocks. These rocks bear exclusive evidence for the high thermal regime prevalent during their formation and can be constrained from accessory phases such as zircon, monazite and garnet. Here, we present the P-T-t evolutionary history of garnet-cordierite-spinel granulites from different localities in the Madurai block. Combined petrography, mineral reaction, geothermobarometry and pseudosection modelling of the samples record HT to near-UHT metamorphic conditions with clockwise P-T trajectories. LA-(MC)-ICPMS U-Pb/Hf isotopic studies on zircons point to Paleoproterozoic high-grade metamorphism in the area with juvenile magmatic signatures. LA-ICPMS/EPMA monazite dating constrains the timing of HT to near-UHT metamorphism at ~580 and ~550 Ma. The heat source required for the Paleoproterozoic high-T event is correlated with the synchronous igneous emplacements reported in the region, whereas the heat source responsible for the Neoproterozoic HT to near-UHT event can be related to the processes associated with crustal thickening and synchronous mafic emplacement. The clockwise P-T trajectories of these granulites are interpreted as the signature of collisional orogeny prevalent in the region during the final stages of Gondwana assembly.


Introduction
High-temperature (HT) to ultrahigh-temperature (UHT) granulitic rocks exposed in many regional granulite terranes across the globe preserve evidence of extreme thermal conditions experienced by the terrane during its evolution. The Southern Granulite Terrane (SGT) of India is known for preserving such rocks with records of extreme P-T conditions up to 1100°C and 12 kbar with both clockwise and anti-clockwise P-T trajectories  and references therein). The SGT is a collage of distinct blocks which are polydeformed and polymetamorphosed with crustal evolution history ranging from early Archaean to late Neoproterozoic time. The spatiotemporal evolution of these blocks has been variably correlated with the presently dispersed continental fragments in former supercontinents such as Columbia, Rodinia and Gondwana. Identification of HT-UHT rocks from SGT and their P-T-t characterization is of paramount importance in demarcating tectonic boundaries and better our understanding on the geodynamic evolution of the terrane. This study presents new occurrence of HT-UHT granulitic rocks from the central part of SGT known as the Kambam UHT belt. Combined petrography, geothermobarometry and phase equilibrium modelling of these granulites were carried out to characterize their P-T evolution. U-Pb/Hf isotopic studies and texturally constrained monazite chemical analysis were undertaken to constra in the timing and nature of metamorphism prevalent in the area. The results are interpreted in the light of the present models proposed for the correlation of SGT with its counterparts in the East Gondwana.
from the Trivandrum and Nagercoil blocks. The Madurai block constitutes different lithological units such as charnockite, hornblende/garnet biotite gneiss, HT-UHT Mg-Al granulites, quartzite, calc-silicate, twopyroxene mafic granulite, along with intrusives such as syenite, granite, anorthosite and carbonatite ( Figure 1). Ghosh et al. (2004) identified a Neoproterozoic shear zone named Karur-Kambam-Painavu-Trichur Shear zone (KKPTSZ) that separates the Madurai block into a dominantly late Archaean crust to the north and Meso-to Neoproterozoic crust to the south. KKPTSZ also roughly overlaps with an isotopic boundary proposed by Plavsa et al. (2012). The eastern flank of KKPTSZ was later renamed as Suruli shear zone (SSZ; Srinivasan and Rajeshdurai 2010). Based on these observations, Brandt et al. (2014) classified the Madurai block into western and eastern domains. Western Madurai domain constitutes massive charnockites and granites, whereas eastern domain dominantly constitutes orthogneisses including migmatitic Hbl-Bt gneisses with rafts of metabasites, charnockites and supracrustal sequence (metapelites, quartzites and calc-silicates). The characteristic feature of Madurai block is the numerous occurrences of UHT metapelites with different mineral assemblages identified along a linear belt coalescing with SSZ, later named the Kambam Ultrahightemperature belt (KUB; Figure 1 inset; Brandt et al. 2014). The first detailed report on UHT assemblages from KUB was from the Kambam valley followed by studies from Palani, Ganguvarpatti, Rajapalayam, Kotameri, Panangad, Vadkampatti localities, all  GSI, 1981). Inset shows Southern Granulite Terrane (modified after Kumar et al. 2017). Madurai block (MB), Trivandrum block (TB), Nagercoil block (NB), Palghat-Cauvery shear system (PCSS), Suruli shear zone (SSZ), Achankovil shear zone (ASZ), Karur-Kambam-Painavu-Trichur Shear zone (KKPTSZ) and Kambam UHT Belt (KUB). estimating extreme P-T conditions up to 1130°C and 12 kbar with clockwise and anti-clockwise P-T trajectories (Dev et al. 2021 and references therein). The timing of UHT metamorphism was first constrained by Santosh et al. (2006) by monazite age peaks at ∼590 Ma from sapphirine bearing rocks in Lachmanapatti and Ganguvarpatti area, followed by zircon and monazite ages from different locations between 593 and 532 Ma Tiwari and Sarkar 2020). These were interpreted as the timing of peak UHT metamorphism during the final stage of Gondwana assembly (Collins et al. 2014).

Field relations and petrography
For the present study, garnet-cordierite-spinel bearing metapelites were collected from two quarries (KT-7B and KT-9B) located within the NE trending metapelite band parallel to SSZ (Figure 1). The metapelite band is surrounded by massive charnockite with intercalation of foliated granite gneiss and are locally intruded by granite veins (Figure 2(a-d)). The metapelites are composed of garnet (~25%), cordierite (~20%), spinel (~10%), biotite (~15%), plagioclase (~10%), alkali feldspar (~10%), sillimanite (~5%) and quartz (1-2%) along with minor amount of Fe-Ti oxides and accessories including zircon and monazite. Garnet is found as large anhedral, medium to coarse-grained porphyroblasts with a diameter of up to 1.5 cm, primarily associated with biotite and plagioclase (Figure 3(a) and Figure 4). Garnet also hosts biotite, K-feldspar and sillimanite inclusions, possibly representing the remnant prograde phases (Figure 3 (a)). Dark greenish-coloured spinel (~80-150 µm in diameter) is found as anhedral grains mostly exsolved with magnetite ( Figure 3(b,c,e)) or as symplectite with cordierite and K-feldspar. Cordierite is found as symplectite and intergrowth, in association with spinel, quartz, biotite, sillimanite and K-feldspar (Figure 3(b,f)). Multiple generations of biotite have been identified from the samples, such as (i) matrix biotite, occurring as coarse flakes, (ii) inclusion biotite, primarily within garnet, and (iii) dark brownish patchy biotite. Sillimanite is observed in two textural associations, as matrix sillimanite and inclusions mostly within garnet. Plagioclase occurs as large irregular grains mostly in contact with garnet and biotite, whereas K-feldspar is present as irregular grains associated with spinel and cordierite. The presence of round quartz grains having cuspate boundaries in feldspar domains suggest the presence of melt during crystallization ( Figure 3(d-f)).

Mineral chemistry
Mineral compositions were analysed using CAMECA SX-100 EPMA at Department of Geology and Geophysics, Indian Institute of Technology, Kharagpur, India. The instrument was operated at an accelerating voltage of 15 kV, 15 nA beam current and 1-2 μm beam diameter. Natural silicate and oxide standards were used for calibration, and raw data were corrected using the ZAF program. EPMA data of different minerals are provided in Supplementary appendix 1.

Spinel
Spinel from the samples are greenish and unzoned with an average composition of Spl 23-30 Her 67-71 Gah 2-9 . They are found as exsolved phase with magnetite. X Mg (X Mg = Mg/Mg + Fe 2+ + Zn) content range between 0.23 and 0.30, whereas ZnO content ranges up to 3.38 wt%.

Cordierite
Cordierite is Mg-rich having X Mg between 0.74 and 0.81.

Biotite
Biotite in the matrix is characterized by high TiO 2 content (up to 5.64 wt %) with X Mg values ranging between 0.47 and 0.64. Biotite inclusions within garnet porphyroblasts (TiO 2 up to 6.13 wt%) have marked compositional variation with respect to other textural occurrences recording considerably high X Mg (0.61-0.73) content.

Sillimanite
Sillimanite from both the samples incorporates similar FeO concentrations (1.06-1.08 wt%), which is typical for high-grade rocks.

Mineral paragenesis
Based on the textural association and reaction textures, the metamorphic evolution of the samples under study has been deciphered. The inclusion of biotite and sillimanite along with K-feldspar in garnet porphyroblast (Figure 3(a)) represents the remnant prograde assemblage, which forms the garnet-bearing peak metamorphic assemblage by biotite dehydration melting (Le Breton and Thompson1988)1980. The reaction is as follows: Post-peak metamorphic condition of the samples is manifested by the formation of spinel and cordierite symplectite by the breakdown of garnet in response to decompression (Harris 1981;Montel et al. 1986;Carson et al. 1997). This can be represented by the following reaction: garnet þ sillimanite ! cordierite þ spinel (Reaction2) This reaction proceeds to the right-hand side in response to continued decompression resulting in the decreased modal abundance of garnet to lower P-T, that involves large volume change. Continued cooling and retrogression to lower P-T conditions results in the development of intergrowth textures involving biotite, quartz, cordierite, sillimanite and K-feldspar ( Figure 3(d,f)). Such textures can be explained by the following reactions (Sorcar et al. 2019):

Geothermobarometry
The high sensitivity of ion exchange reactions and compositional re-equilibration on cooling in high-grade rocks challenges the demarcation of peak metamorphic conditions (Harley 1989(Harley , 2008. In the present study, available mineral assemblages and reaction textures allow the application of multiple geothermobarometers including, garnet-spinel thermobarometry (Nichols et al. 1992), garnetcordierite thermometry (Bhattacharya et al. 1988), Ti-inbiotite thermometry (Henry et al. 2005), garnet-biotite thermometry, GASP barometry (Ganguly et al. 1996), and garnet-cordierite-spinel barometry (Harris 1981). Due to the non-availability of a convincing geothermometer, the maximum temperature of metamorphism cannot be estimated for the studied samples. However, maximum pressure attained by the samples is estimated using the composition of garnet porphyroblast and plagioclase cores by GASP barometry (Ganguly et al. 1996), and this estimates a maximum pressure of 9.5 and 7.5 kbar for KT-7B and KT-9B, respectively. As the post-peak evolution of the sample involves the formation of decompression textures involving cordierite and spinel, the composition of garnet rim and spinel/cordierite in contact with those can be used to estimate P-T conditions for the development of decompression textures. Garnet-spinel thermometry estimates using the composition of garnet rims and associated spinel yield a maximum temperature of 890°C for KT-7B and 840° C for KT-9B, whereas garnet-cordierite thermobarometry on garnet rim and cordierite (in contact with garnet) estimates a maximum P-T condition of 810°C and 725 o C for KT-7B and KT-9B, respectively. These mineral compositions were further used to evaluate the pressure conditions for the development of decompression textures using garnet-cordierite-spinel barometry. This calibration estimates a maximum pressure of 5.5 kbar for KT-7B and 5 kbar for KT-9B. Matrix biotites in samples do not represent the prograde phase and are characterized by low TiO 2 content. Moreover, high TiO 2 biotites found as inclusions with in the garnets cannot be used for constraining the peak metamorphic condition as they are highly susceptible to resetting during retrogression. Hence, garnet-biotite exchange thermometer of Ganguly et al. (1996) cannot yield maximum temperatures. However, temperature estimates using garnetbiotite-plagioclase rim composition can be used to estimate retrograde temperature and pressure by the application of garnet-biotite and GASP thermobarometry. These calibrations estimate retrograde P-T conditions at 550°C/4.5 kbar for KT-7B and 585°C/4 kbar for KT-9B. Ti-in-biotite thermometry estimates a maximum temperature of 674°C for KT-7B and 709°C for KT-9B. A summary of estimated P-T values are given in Table 1.

Phase equilibrium modelling
P-T pseudosections were constructed using Perple_X_6.9.0 in the system Na 2 O-CaO- Figure 4(a,b)). Effective Bulk Composition (EBC; Stüwe 1997) was used for P-T section construction to avoid sample heterogeneity. EBC was calculated by integrating the volume percentage of minerals in the thin section and their mineral chemistry. The selection of solid solutions was carried out with reference to the solution_mo-del_v.6.9.0 of Perple_X for phase diagram computation. Gt(HP) for garnet (Holland and Powell 1998), Sp(HP) for spinel (Holland and Powell 1998), Bio(TCC) for biotite  (Powell and Holland 1999), Ilm(WPH) for ilmenite (Holland and Powell 1998), Melt(HP) for melt (Holland and Powell 2001) and Pl(h) for plagioclase (Newton et al.2014) were chosen as solution models for pseudosection construction. For cordierite, an ideal model with the correction of hfcrd_i stoichiometry by L. Baumgartner was adopted. X(CO 2 )-H 2 O-CO 2 CORK Powell 1991, 1998)  The presence of biotite, sillimanite and K-feldspar inclusions in the garnet porphyroblast (Figure 3(a)) represents the remnant prograde assemblage which is exhibited by the stability field Pl-Bt-Grt-Ilm-Melt-Sill-Kfs-Qtz. The peak metamorphic assemblage is represented by the stability field Pl-Grt-Ilm-Melt-Sill-Kfs-Qtz, with P-T conditions greater than ~800°C and ~6 kbar in both samples. However, due to uncertainties induced by the choice of bulk composition as well as nonapplicability of compositional isopleths together with lack of a promising geothermometer, precise peak metamorphic temperature cannot be constrained (Sorcar et al. 2014; Hand 2015 and references therein). However, the maximum pressure (P max ) attained by the samples can be constrained using the application of GASP barometry (Ganguly et al. 1996) using the composition of garnet and plagioclase cores. This barometry estimates a P max of ~9.5 kbar for KT-7B and ~7.5 kbar for KT-9B, both values falling within the speculated peak field (Figure 4(a,b)). Considering the uncertainties in establishing the peak P-T conditions, studied samples are expected to be metamorphosed at temperatures >800°C with corresponding pressure values as mentioned above. Continued P-T evolution of the samples is demonstrated by the breakdown of garnet in response to decompression forming cordierite and spinel (Reaction 2). Entry of the P-T path into the stability field Pl-Crd-Grt-Ilm-Melt-Sill-Kfs-Qtz demonstrates this reaction. However, the absence of spinel in this field can be due to the non-incorporation of Zn component that can affect the spinel stability line (Montel et al. 1986). The pressure-temperature dependency of spinel stability with respect to Zn incorporation is constrained using the formulation of Nichols et al. (1992), as there is considerable incorporation of ZnO in spinel in the studied samples (up to 3.38 wt%; see Supplementary appendix 1). The results show a shift in spinel-in line to lower P-T conditions in the pseudosection (marked as yellow solid lines in Figure 4(a,b)). This shift of spinel stability towards the lower P-T range incorporates spinel into the speculated field, making it Pl-Crd-Grt-Spl-Ilm-Melt-Sill-Kfs-Qtz. In addition, the presence of spinel in higher temperature condition and its transformation into magnetite in lower temperatures are also noted. This observation is in accordance with the textural assemblage ( Figure 3)where spinel is found in association with magnetite as an exsolved phase. The boundary of spinel-magnetite transition solvus in both samples are located at ~800°C, which is in agreement with the available models proposed for spinel-magnetite transition in natural systems (Sengupta et al. 1999;Wang et al. 2021;Dev et al. 2021 and references therein). The formation of cordierite and spinel in response to decompression by the breakdown of garnet is further corroborated by the increased modal abundance of spinel/ cordierite and decreased modal abundance of garnet towards the speculated field (Supplementary appendix 2). This stage also involves large volume change due to melt-solid back reaction as exhibited by the variation in modal abundance of garnet, cordierite, spinel and melt (Reaction 2). The P-T conditions for the formation of decompression textures is anchored using geothermobarometers by utilizing the composition of garnet rim and cordierite/spinel core compositions. Thermometry based on the garnet-spinel equilibration (Nichols et al. 1992) estimates a maximum temperature of 890°C for KT-7B and 840°C for KT-9B. Using garnetcordierite thermometry (Bhattacharya et al. 1988), a maximum temperature of 810°C for KT-7B and 725° C for KT-9B is estimated, all falling with in the speculated stability field. The pressure values are anchored using garnet-cordierite-spinel barometry (Harris 1981) yielding a maximum pressure of 5.5 and 5 kbar for KT-7B and KT-9B respectively, which is in agreement with the pressure range of inferred stability field. Subsequent cooling to lower P-T condition is marked by the appearance of biotite in the samples as indicated by the entry of P-T path to the stability field Bt-Pl-Crd-Grt-Mt-Ilm-Melt-Sill-Kfs-Qtz (Reaction 3-4). The P-T condition for retrogression and biotite formation can be anchored using garnet-biotite thermometry and GASP barometry, utilizing the composition of garnet/plagioclase rim and biotite core. These results estimate P-T conditions of 550°C/4.5 kbar for KT-7B and 585°C/4 kbar for KT-9B, and this strand of P-T loop represents near-isobaric cooling profile for the studied samples (Figure 4(a,b)).

Accessory mineral geochemistry and geochronology
The samples were crushed and sieved to desirable grain fractions for conventional Wilfley table-magnetic separation techniques. Separated grains were manually handpicked and mounted on a standard epoxy disc of 25 mm diameter. The internal structure of these grains was examined by cathodoluminescence (CL) and backscatter electron (BSE) imaging using TESCAN Vega4 SEM at NCESS. Single spot U-Pb isotope and trace element analysis of zircons and monazites were performed at Isotope Geochemistry Facility (IGF), NCESS using a Teledyne CETAC, Nd:YAG (213 nm) solid-state laser coupled with Agilent 7800 quadrupole ICPMS. The analytical protocol is according to Dev et al. (2021. Data reduction and calculation of ratios and ages were performed offline using Iolite 4.4 (Paton et al. 2011) using the integrated Visual Age Package (Petrus and Kamber 2011). The isotopic ratios and elemental concentrations were processed using Isoplot 4.1 (Ludwig 2008). For U-Pb geochronology, 91500 zircon (Wiedenbeck et al. 1995;206 Pb/ 238 U age = 1062.4 ± 0.8) and Moacyr Monazite (Gonçalves et al. 2016;207 Pb/ 235 U age = 506.44 ± 0.73) were used as primary standards, whereas Plesovice Zircon (Sláma et al. 2008;206 Pb/ 238 U age = 337.13 ± 0.37) and Diamantina Monazite (Gonçalves et al. 2018;206 Pb/ 238 U age age = 495.26 ± 0.45) were monitored for quality control. NIST 610 and NIST 612 (Pearce et al. 1997) were used as primary reference standards for time-drift correction and quality monitoring for in situ trace element determination. 29 Si (IS value = 14.98%) was used as the internal standard for zircon, whereas CaO was used for monazite and garnet REE measurements that were measured using EPMA. REE values were normalized with the chondrite values from Anders and Grevesse (1989). As the samples are quartz-bearing, Ti-in-zircon crystallization temperatures (Ferry and Watson 2007) were calculated using α SiO2 ¼ 1 and α TiO2 ¼ 1. U-Pb ratios, ages, trace element composition of zircon, monazite, garnet and Tiin-zircon temperatures are given in Supplementary appendix 3 and 4.
Monazite grains were analysed for texturally controlled in-situ U-Th-total Pb dating in thin section, using CAMECA SX5 TACTIS electron probe microanalyzer (EPMA) at NCESS. The analytical techniques and operating conditions for the electron microprobe are according to Sorcar et al. (2021). The instrument was operated at a voltage of 20 kV and current 200 nA with a Lanthanum hexaboride filament. X-ray lines were used for major, trace and age measurements are P-Kα, Ca-Kα, Si-Kα, Fe-Kα, Al-Kα, Y-Lα, La-Lα, Ce-Lα, Pr-Lβ, Nd-Lβ, Sm-Lβ, Gd-Lβ, Er-Lα, Eu-Lα, Dy-Lβ, Ho-Lβ, Pb-Mα, Th-Mα and U-Mβ. The age and respective error from the measured values of Th, U and Pb were calculated using the PeakSight software. U-Pb-Th ages and trace element concentrations are given in Supplementary appendix 5.

LA-(MC)-ICPMS zircon U-Pb/Hf isotopes and trace element analysis
Zircons from the samples range between 100 and 200 µm with variable zoning patterns. Mostly they show a dark, planar to oscillatory zoned core wrapped by a sector or oscillatory zoned mantle. This is followed by a very thin CL bright planar outer rim. The mantle occasionally truncates into the core domain or maybe overprinted or modified by lobate curvilinear domains with bright CL (Figure 5(a,b) inset).
Zircons from KT-7B is characterized by variable U (57-416 ppm), Th (30-488 ppm) and Pb (22-802) concentrations with Th/U ratio ranging between 0.17 and 1.40. Sixty-one spot analysis on 43 zircons from the sample defines a discordia on a Wetherill diagram with an upper intercept age of 2474 ± 26 Ma and lower intercept age of 578 ± 81 (MSWD = 4.6) ( Figure 5(a)). The upper intercept ages agree with a concordia age of 2458 ± 15 Ma (MSWD = 2.9) defined by 16 spots. The 207 Pb/ 206 Pb ages of the zircons range between 1235 and 2800 Ma. Zircons from KT-9B is characterized by variable U (78-590 ppm), Th (53-938 ppm) and Pb (135-1065) concentrations with Th/U ratio ranging between 0.14 and 2.14. Sixty-two spot analysis on 35 zircons from the sample defines a discordia on a Wetherill diagram with an upper intercept age of 2481 ± 30 Ma and lower intercept age of 514 ± 200 (MSWD = 5.9) (Figure 5(b)). Sixteen grains define a concordia age of 2456 ± 17 Ma (MSWD = 3.9), which agrees with the upper intercept ages, whereas 207 Pb/ 206 Pb ages range between 1754 and 2668 Ma. Chondrite normalized REE patterns (Figure 5(c,d)) of zircons are characterized by strong positively sloping REE patterns with moderate LREE depletion and strong HREE enrichment. The zircons are also characterized by a prominent positive Ce anomaly and strong negative Eu anomaly with values ranging between Ce/ Ce* = 3.8-300 and Eu/Eu* = 0.10-0.65, respectively. Ti concentration ranges between 2.08 and 16.94 ppm for KT-7B and 1.72-20.58 ppm for KT-9B, respectively estimating crystallization temperatures (Ferry and Watson 2007) between 607 and 809°C with a maximum cluster around ~740°C and ~720°C for both sample sets ( Figure 5(c,d) inset). Lu-Hf isotopic data of zircons from both samples are shown in Figure 6. 176

LA-ICPMS monazite U-Pb and trace element analysis
Monazites from KT-7B and KT-9B are sub-rounded with patchy zoning, exhibiting alternate dark and light domains as revealed in SEM-BSE imaging (Figure 7(a,b) inset). Monazites are characterized by U and Th concentrations between 13634-190315 ppm and 15115-175 ppm. Thirty-two spot analysis yielded a 206 Pb/ 238 U age between 479 and 654 Ma with a prominent peak around ~580 Ma (Figure 7(a) inset). Among these, fourteen grains define a concordia age at 585 ± 7 Ma (MSWD = 3.9) (Figure 8(a)). Monazites from KT-9B shows Th/U ratios between 11 to 415 defined by Th and U concentrations of 13633-182675 ppm and 175-5728 ppm, respectively. Thirty-five analyses on monazite provide a 206 Pb/ 238 U age spread between 596 and 479 Ma with a prominent age peak at ~550 Ma (Figure 7(b) inset). Among these, twenty-three grains define a concordia age at 552 ± 7 Ma (MSWD = 1.7) (Figure 8(b)). Chondrite normalized REE patterns of monazites from both samples (Figure 7(c,d)) are characterized by strong HREE depletion relative to LREE with an overall negatively slopping REE pattern and distinct negative Eu anomaly. A distinct contrast between Y concentration is noticed between the samples where monazites from KT-7B hosts a high-Y concentration (29671-57236 ppm) compared to that of KT-9B (169-11851 ppm).

LA-ICPMS garnet trace element analysis
Garnet from both associations are chemically distinct concerning their REE composition. Garnets from KT-7B are REE rich with ∑REE between 91 and 156 ppm, whereas KT-9B has lower ∑REE with values ranging between 79 and 43 ppm. Similar to ∑REE, Y concentration is also different for both samples, with higher values for KT-7B (509-113 ppm) compared to . Chondrite-normalized REE pattern of the samples (Supplementary appendix 7) show negatively sloping REE pattern with moderate LREE enrichment and near flat to weak HREE depletion. However, core to rim compositional variation is slightly prominent in KT-7B, where depletion in the HREE of the rim (Lu N /Gd N = 0.16-0.27) is noted compared to that of core (Lu N /Gd N = 0.28-0.41). On contrary, HREE pattern of the garnet is similar for the core (Lu N /Gd N = 0.27-0.33) and rim (Lu N /Gd N = 0.33-0.59) domains in KT-9B. A strong negative Eu anomaly is also noticed for both samples with Eu/Eu* ranging between 0.07 and 0.27, respectively.

P-T evolution of metapelites
Phase equilibrium modelling, petrography, mineral chemistry and geothermobarometry of the studied samples provide new evidence for HT to near-UHT metamorphism in the KUB. In the study, the peak metamorphic assemblage is represented by the stabilization of garnet porphyroblast in response to biotite dehydration melting (Reaction 1) defined by the stability field Gt-Pl-Ilm-Melt-Sill-Kfs-Qtz. However, due to the absence of a convincing geothermometer, the peak metamorphic temperature cannot be constrained, although the pressure estimates range up to 9.5 kbar and 7.5 kbar for KT-7B and KT-9B. As this peak field is expected to be formed at a temperature >800°C as inferred from pseudosection modelling, the metapelites under study can also be considered to have attained similar conditions. The subsequent decompression stage involves the formation of cordierite and spinel due to garnet breakdown (Reaction 2). The P-T values tied to this decompression textures are estimated between ~890°C and 810 o C at 5.5 kbar in KT-7B and 840°C to 725°C at 5 kbar for KT-9B (see Table 1). As the upper temperature limit for the development of decompression textures are marked at ~890°C and 840°C, the peak metamorphic temperature for the studied samples naturally must be higher than these. Hence, these samples should have attained high-temperature to near ultrahigh-temperature metamorphic conditions greater than ~850°C. This is supported by other studies reporting UHT metamorphic conditions for Grt-Spl and Crd-Spl metapelites from KUB  and references therein). The near-isobaric cooling profiles subsequent to decompression are expected to be formed at low P-T conditions about ~550°C and ~4.5 kbar. Hence, a clockwise P-T path with an initial near-isothermal decompression followed by near-isobaric cooling is inferred for the samples. Similar P-T profiles with highto ultrahigh-temperature estimates have been previously reported from different parts of Madurai block (Brandt et al. 2011;Clark et al. 2015;Tiwari and Sarkar 2020;Dev et al. 2021 as well as adjacent crustal blocks in SGT (Sorcar et al. 2019;Yu et al. 2019 among many others).

Timing and duration of metamorphism in KUB
U-Pb analysis of zircons from metapelites yield Paleoproterozoic concordant and upper intercept ages at ~2.5 Ga (Figure 5(a,b)). From CL images, it is observed that zircons are characterized by both oscillatory zoned as well as rounded unzoned cores along with overgrowths which is indicative of crystallization from an anatectic melt and are typically found in zircons from high-grade metamorphic rocks (Corfu et al. 2003). Moreover, their high U and low Th concentrations (Supplementary appendix 4), flat to negative sloping REE pattern, enrichment of HREE with respect to LREE, high Hf concentration and positive Ce anomaly ( Figure 5  (c,d)) reflect crystallization from leucosome formed by the partial melting of garnet-bearing metasediments (Harley and Nandakumar 2014). The high Y concentration in zircons further validates this, as the partitioning of Y into zircons takes place during garnet breakdown (Möller et al. 2003). The negative Eu anomaly in zircons denotes formation from a Eu depleted rock, or a melt co-existed with feldspar (Whitehouse and Platt 2003 and references therein). Also, considerably low Th/ U ratios (Rubatto 2002) and Ti-in-zircon temperature values >700°C of these zircons (Figure 5(c,d) inset) emphasize their metamorphic origin. These trace element signatures are similar to that of Paleoproterozoic metamorphic zircons reported from this area , suggesting their formation during the Paleoproterozoic high-T metamorphic event reported from the region. The Paleoproterozoic ages are also consistent with previously published metamorphic zircon ages from Lachmanapatti (~2.4 Ga; Santosh et al. 2006 values (Figure 6). These results are in excellent agreement with the positive ℇHf signatures for Paleoproterozoic zircons reported from MB (Kumar et al. 2017;Dev et al. 2021). Together with the whole-rock Sm-Nd model of rocks from this domain (Plavsa et al. 2012;Tomson et al. 2013), the results can be interpreted as evidence for Archaean crustal component beneath KUB and Madurai block, with new evidence for Paleoproterozoic metamorphism in the area.
The timing of Neoproterozoic HT to near-UHT metamorphism in the samples are represented by age peaks at ~580 and ~550 Ma from monazite chemical ages. These overlap with corresponding weighted mean age populations at 581 ± 10 Ma and 552 ± 8 Ma (Figure 9(c,d)). These prominent age populations are derived from the low-Y zones of matrix monazites. As garnet is inferred to be formed during peak UHT metamorphism, the ages from low-Y zones can be directly correlated with the garnet forming event (UHT event), where garnet could have acted as a sink for Y in granulite facies conditions. EPMA monazite chemical ages also agree with monazite concordia ages obtained by LA-ICPMS dating at 585 ± 7 Ma and 552 ± 7 Ma (Figure 7(a,b)). These ages are similar to the reported timing of UHT metamorphism in KUB and fall within the upper and age lower limits of recorded UHT event between 593 Ma  and 532 Ma (Tiwari and Sarkar 2020). These are also similar to the timing of HT-UHT metamorphism reported in other blocks of SGT (2009a;Harley and Nandakumar 2014;Johnson et al. 2015;Sorcar et al. 2019, among many others), as well as other Gondwanan fragments such as Madagascar and Sri Lanka (Dharmapriya et al. 2021;Tucker et al. 2011 for review), which underlies the fact that this Neoproterozoic UHT event as the most pervasive thermal event throughout East Gondwana. In addition, the polymetamorphosed evolutionary history of the terrane is evidenced by the presence of several metamorphic pulses recorded in monazites in the sample. Analyses of zircon rims/overgrowths in the studied metapelites were restricted owing to their very thin presence. The limited growth of zircon rims can be either low bulk rock Zr concentration or low concentration of little dissolved zircon in the system subsequent to melt loss (Kelsey et al. 2008). Owing to their high analytical error, the zircon lower intercept ages can be roughly correlated with the late Neoproterozoic HT to near-UHT event recorded in the sample.

Metamorphic drivers for HT-UHT metamorphism KUB
As mentioned above, the Paleoproterozoic (~2.5 Ga) high-temperature (HT) event has been interpreted as a pervasive thermal event that can be traced across KUB and Madurai block. Hence, it is important to characterize the heat source responsible for this event.
Madurai block has records of proliferous association of ~2.5 Ga igneous emplacements such as charnockites (~2.4 to 2.5 Ga; Brandt et al. 2014;Raith et al. 2016;Kumar et al. 2017), TTG gneisses (~2.5 Ga; Ghosh et al. 2004), biotite gneisses (2.4-2.6 Ga; Ghosh et al. 2004;Brandt et al. 2014;Raith et al. 2016), granodiorites (2.5 Ga-2.6 Ga; Plavsa et al. 2012), granites and metagranites (2.4-2.5 Ga; Ghosh et al. 2004) and numerous alkaline bodies (2.4-2.6 Ga; Plavsa et al. 2012). The reported hafnium isotopic composition of these zircons indicates their juvenile nature with positive ℇHf values (Kumar et al. 2017), indicating active crust building in the Madurai block during ~2.5 Ga. These results are similar to the Hf isotope signatures of zircons in this study and agrees with the observation of Brandt et al. (2011, who postulated large scale igneous emplacement during ~2.5 Ga as the prominent heat contributor for Paleoproterozoic high-T metamorphism in KUB and Madurai block. As discussed in the foregoing, HT-UHT metamorphic records have been profoundly reported in Madurai block from numerous localities with extreme P-T conditions up to 1130°C and 12 kbar. The timing of these events is roughly bracketed between 593 and 532 Ma with reports of a long-lived and slow-cooled orogenic history (Dev et al. 2022 for review). In this context, the Neoproterozoic ages recorded in monazites from the samples are significant as they fall within the time bracket of HT-UHT metamorphism reported in the area. The spatio-temporal evolution of HT-UHT granulites has been of great interest as the occurrence of large-scale HT-UHT granulites are mainly restricted to collisional tectonic regimes (Brown 2006(Brown , 2007. Previous studies on UHT rocks from SGT has identified different mechanisms as the heat source for regional scale UHT metamorphism, such as (1) HPE enrichment due to crustal thickening , (2) equilibration at the continental root (Brandt et al. 2011) and (3) combination of HPE enrichment and tectonically driven processes . However, KUB has records of largescale mafic emplacements synchronous with UHT metamorphism that can be interpreted as an alternate heat source for HT-UHT granulite formation (Dev et al. 2022 under review). This argument of an external heat source for HT-UHT granulite formation in KUB is in accordance with the results from Trivandrum block, where granulite formation solely by HPE enrichment is considered improbable without the addition of heat from an external source (Nandakumar and Harley 2019).

Zircon-garnet and monazite-garnet equilibrium
Establishing the equilibrium relation between zircon and monazite to petrologically sensitive minerals such as garnet in polydeformed terranes is of great importance as the timing of growth and breakdown of these minerals need not represent the timing of zircon/monazite formation (Clark et al. 2018 and references therein). As the REE partition relationship between zircon and garnet is well constrained in both experimental and natural systems (Whitehouse and Platt 2003;Clark et al. 2009a;Nandakumar 2014, 2016;Taylor et al. 2015;Warren et al. 2019), this can be used as a proxy in linking accessory mineral formation with metamorphic episodes. However, the lack of experimentally constrained partition coefficients for monazite-garnet pairs limits this application. Hence, the available partition coefficient values for equilibrated natural monazite-garnet pairs in comparison with our results can be used to reconstruct their equilibration relationship (Warren et al. 2019;Godet et al. 2020;Dev et al. 2021). As discussed earlier, two major metamorphic pulses have been recorded in the samples, (1) ~2.5 Ga Paleoproterozoic HT event recorded in zircon cores and (2) ~580 to ~550 Ma Neoproterozoic HT to near-UHT events recorded in monazites. As the peak metamorphic assemblage in the sample is defined by the formation and stabilization of garnet porphyroblast, it is essential to relate whether garnet formation was related to the Paleoproterozoic or the Neoproterozoic thermal event. The equilibrium relation between zircon and garnet is tested by plotting log (D Yb /D Gd ) vs log (D Yb ) (Supplementary appendix 7) of individual zircons against the garnet core and rim composition as suggested by Taylor et al. (2017) since these elements are susceptible to metamorphic events (Clark et al. 2018). The values (Supplementary appendix 8) are plotted against the experimentally constrained values from Taylor et al. (2015). The results suggest that zircons from both samples plot away from the experimental values compared with both garnet core and rim domains suggesting its non-equilibrium relationship with garnet. The near-flat HREE pattern of garnets further validates the disequilibrium between zircon and garnet. These relations indicate that garnet formation can be related to the late Neoproterozoic HT to near-UHT event in equilibrium with monazites. To assess the equilibrium relation between monazite and garnet, partitioning coefficient values (Supplementary appendix 8) were calculated for monazite against the garnet core-rim and are compared with the available partition coefficients on natural pairs from literature (Hermann and Rubatto 2003;Buick et al. 2006;Rubatto et al. 2006) as shown in Supplementary appendix 7. The results demonstrate equilibrium formation of garnet and monazite pairs and suggest the final garnet-bearing assemblage as the product of Neoproterozoic HT to near-UHT event recorded in the samples.

Tectonic implications
The timing of metamorphic events recorded in the samples at ~2.5 Ga and ~580-550 Ma represents two distinct thermal events recorded in the area. Traditionally, PCSS has been viewed as a terrane boundary within the SGT separating Archaean crustal blocks to the north and Neoproterozoic crustal blocks to the south based on the age of metamorphism. The ~2.5 Ga HT event recorded in zircon core domains with juvenile magmatic signatures in this study suggests the possible continuation of the Archaean crust south of the PCSS. This record of Archaean crustal continuation with earliest Paleoproterozoic HT event south of PCSS into central Madurai block has significant geodynamic importance, as the absence of the same south of PCSS is presented as the supporting evidence for Neoproterozoic suturing in SGT along PCSS (Collins et al. 2014 and references therein). However, our new results together with increased recognition of Archaean crust south of PCSS with a record for Paleoproterozoic metamorphic overprint (Ghosh et al. 2004;Brandt et al. 2011;Plavsa et al. 2012;Raith et al. 2016;Kumar et al. 2017;Dev et al. 2021 contradicts the status of PCSS as a Neoproterozoic suture zone in SGT. Neoproterozoic HT to near-UHT metamorphic imprints recorded in the samples have significance in the tectonic evolution of SGT as the formation of largescale hot orogens are closely related to crust generation and recycling episodes with a temporal relationship to supercontinent assembly (Brown 2007). Such large-scale hot orogens are primarily found in collisional settings with clockwise P-T evolutionary trajectory and may have evolved over timescales of >50 My (Harley 2004;Beaumont 2011, 2013;Kelsey and Hand 2015;Harley 2016 and references therein). In such collisional orogens HT-UHT rocks can be formed due to collisional slab breakoff and delamination (Harley 2008), magmatic underplating in a thickened crust (Harley 1989), due to high heat flow in the continental back-arc region (Brown 2006(Brown , 2007, equilibration at continental root due to asthenosphere upwelling (Brandt et al. 2011) or due to crustal thickening and HPE enrichment . KUB is known for the preservation of proliferous occurrence of UHT granulites formed during 593-532 Ma (Tiwari and Sarkar 2020;Dev et al. 2021), with maximum P-T conditions ~1130°C and 12 kbar with overall clock-wise P-T trajectory (see Dev et al. 2021 for review). The duration of this UHT orogen is constrained between 36-55 Ma Tiwari and Sarkar 2020;Dev et al. 2021) with an initially slow followed by ultra-slow colling pattern . The spatial distribution of these UHT granulites in KUB is along the transcrustal NW-SE Suruli shear zone. Detailed structural, petrological and geochronological investigations along SSZ has suggested this as a terrane boundary within SGT (Srinivasan and Rajeshdurai 2010;Brandt et al. 2014;Dev et al. 2021). The location of SSZ also overlaps with the proposed structural arc representing a collisional boundary between western and east Madurai blocks with evidence for westward thrusting and dextral shearing (Cenki and Kriegsman 2005). Considering the tectonic significance of HT-UHT granulite formation, a tectonic model has been proposed for the formation of such granulites in KUB and is illustrated in Figure 9. The formation of HT-UHT granulites within KUB can be interpreted as the product of collision between west and east Madurai blocks where subcontinental lithospheric mantle (SCLM) subducted beneath eastern Madurai block and crustal thickening was initiated. The post-collisional extension and slab break following the subduction of SCLM resulted in the upwelling of hot asthenosphere. This interacted with the subducted slab and lithospheric basement, leading to large-scale mafic emplacements (Dev et al. 2022 under review) (Figure 9). These mafic emplacements together with continued crustal thickening must have supplied substantial amount of heat to sustain such a long-lived hot orogen  The timing of this collisional event marks the final stages of Gondwana assembly in the SGT. As the argument of Neoproterozoic suturing along PCSS is debated, a revision to the Neoproterozoic geodynamic model of SGT is required, with KUB and Suruli shear zone at its forefront.

Conclusion
• Mineral reaction, geothermobarometry and phase equilibrium modelling of garnet-cordierite-spinel granulite samples provide new evidence for HT to near-UHT metamorphism in KUB at temperature estimates greater than ~850°C and maximum pressure up to 9.5 kbar and 7.5 kbar (for KT-7B and KT-9B, respectively). The P-T trajectory of these rocks is identified as clockwise with an initial nearisothermal decompression followed by a nearisobaric cooling profile. • Zircon geochronology of samples reveal new evidence for Paleoproterozoic high-T metamorphism at ~2.5 Ga, with juvenile magmatic signatures from Hf isotopes. Monazite U-Pb and U-Pb-Th dating constrain the timing of HT to near-UHT metamorphism at ~580 and ~550 Ma. The timing of this Neoproterozoic metamorphic event is correlated with the final stages of Gondwana supercontinent assembly. • Trace element modelling of zircon-garnet, and monazite-garnet pairs suggests the equilibration relation between monazite and garnet pairs. This relation interprets the growth of garnet in relation with the HT to near-UHT metamorphic events recorded in the samples. • P-T-t evolutionary history of studied metapelites together with the published results from the terrane suggest collisional tectonism along the Suruli shear zone and KUB. Hence, SSZ can be considered as a prominent terrane boundary within the SGT.