Absence of evidence for Palaeoproterozoic eclogite-facies metamorphism in East Antarctica: no record of subduction orogenesis during Nuna development

The cratonic elements of proto-Australia, East Antarctica, and Laurentia constitute the nucleus of the Palaeo-Mesoproterozoic supercontinent Nuna, with the eastern margin of the Mawson Continent (South Australia and East Antarctica) positioned adjacent to the western margin of Laurentia. Such reconstructions of Nuna fundamentally rely on palaeomagnetic and geological evidence. In the geological record, eclogite-facies rocks are irrefutable indicators of subduction and collisional orogenesis, yet occurrences of eclogites in the ancient Earth (> 1.5 Ga) are rare. Models for Palaeoproterozoic amalgamation between Australia, East Antarctica, and Laurentia are based in part on an interpretation that eclogite-facies metamorphism and, therefore, collisional orogenesis, occurred in the Nimrod Complex of the central Transantarctic Mountains at c. 1.7 Ga. However, new zircon petrochronological data from relict eclogite preserved in the Nimrod Complex indicate that high-pressure metamorphism did not occur in the Palaeoproterozoic, but instead occurred during early Palaeozoic Ross orogenesis along the active convergent margin of East Gondwana. Relict c. 1.7 Ga zircons from the eclogites have trace-element characteristics reflecting the original igneous precursor, thereby casting doubt on evidence for a Palaeoproterozoic convergent plate boundary along the current eastern margin of the Mawson Continent. Therefore, rather than a Palaeoproterozoic (c. 1.7 Ga) history involving subduction-related continental collision, a pattern of crustal shortening, magmatism, and high thermal gradient metamorphism connected cratons in Australia, East Antarctica, and western Laurentia at that time, leading eventually to amalgamation of Nuna at c. 1.6 Ga.


Significance of the Miller Range relict eclogite
Eclogite-facies rocks are rare in the Palaeoproterozoic geological record 19,29 , but are critical in delineating the existence of palaeosubduction systems. Previous geochronology indicated that the Nimrod Complex in the Miller Range, located in the modern Transantarctic Mountains, records one such example of Palaeoproterozoic eclogitefacies metamorphism 13 . The Nimrod Complex contains a diverse assemblage of amphibolite-to granulite-facies quartzofeldspathic, migmatitic, mafic and granitic gneisses, that enclose boudinaged mafic blocks preserving relict eclogite-facies mineral assemblages 17,23 (Fig. 2). Although the Nimrod Complex rocks have been pervasively overprinted by the Neoproterozoic to early Paleozoic Ross Orogeny 30,31 , the existing geochronology highlights a record of Mesoarchean to Palaeoproterozoic magmatism and metamorphism 13,23,32,33 . SHRIMP U-Pb zircon ages of c. 1720 Ma from a relict eclogite 13 (sample 90-131A) were interpreted to reflect Palaeoproterozoic eclogitefacies metamorphism and orogenesis related to crustal thickening driven by plate convergence and/or collision (Nimrod Orogeny). However, the U-Pb age data were not accompanied by zircon trace-element compositions that link the zircon ages explicitly to the high-pressure mineral assemblage. Applying in-situ petrochronological techniques to zircons from a second relict eclogite sample (90-130D) located in the same area as sample 90-131A (Gerard Bluffs, Miller Range; Fig. 2 www.nature.com/scientificreports/ Whereas zircon ages from eclogite 90-131A are primarily Palaeoproterozoic, with only a small proportion showing slight isotopic resetting toward Cambrian ages 13 , the zircons in eclogite 90-130D are primarily Cambrian-aged, with a less dominant proportion giving Palaeoproterozoic ages which are highly discordant and typically significantly older than c. 1720 Ma 20 . Considering these contrasting age patterns between two mineralogically similar eclogitic samples from a similar location, and because the zircons from eclogite 90-131A lack trace-element data, it is necessary to obtain trace-element information from the Paleoproterozoic-aged zircons to establish whether: (1) the mafic eclogites of the Miller Range record two high-pressure events separated by ~ 1200 Myr, or (2) high-pressure metamorphism in the Miller Range was solely a consequence of active convergent margin processes along the Cambro-Ordovician East Gondwana margin 20 .

Zircon petrochronology
To test the previously proposed hypothesis that the Palaeoproterozoic zircons from sample 90-131A formed during high-pressure metamorphism along a plate boundary adjacent to western Laurentia, U, Pb, Th and trace-element isotopic concentrations were measured from twenty-three zircons using laser-ablation inductivelycoupled mass spectrometry (LA-ICP-MS). Zircon U-Pb ages and trace-element concentrations are given in Supplementary Table S1. Zircons were separated using traditional mineral separation techniques as outlined by Goodge et al. 13 Zircon grains are approximately 100-150 µm in length and have equant to semi-prismatic morphologies (Fig. 3). The zircons exhibit features typically observed in both magmatic or metamorphic zircon, showing concentric growth zoning patterns, sector zoning and less commonly, oscillatory zoning (Fig. 3). A common feature in many of the zircon grains are thin (~ 5-10 µm wide) bright-CL rims which were too small to target using laser ablation (Fig. 3). However, previous SHRIMP analysis of similar zircons gave rim ages between 535 and 480 Ma 23,33 .
Sixty U-Pb analyses define a tight discordant array on a Wetherill concordia plot, with 11 analyses > 5% discordant (Fig. 4a). A regression line through all analyses yields a constrained upper-intercept age of 1744 ± 20 Ma (MSWD = 1.05) and a poorly constrained lower-intercept age of 577 ± 120 Ma, confirming that this sample records both the Nimrod and Ross orogenies. On a chondrite-normalised trace-element plot, all analyses have similar trace-element signatures, showing pronounced positively sloping heavy rare earth element (HREE) trends (Fig. 4b). There appears to be no correlation between trace-element concentration and internal zoning (Fig. 3b). Th/U ratios are between 0.23-0.48, consistent with values for magmatic zircon. Sm concentrations (in ppm) are below detection for most of the analyses; however, where Sm was detected, Eu anomalies (Eu*) are weakly negative (Supplementary Table S1). The magnitudes of the Eu anomalies (Eu/Eu*) range from 0.21 to 0.55 and are consistent with low-pressure zircon formation 35 (Fig. 4b; Supplementary Table S1).

Discussion
Palaeoproterozoic or Cambrian high-pressure metamorphism in the Miller Range? The new U-Pb age results from zircons in eclogite 90-131A are consistent with existing SHRIMP U-Pb data, yielding a Palaeoproterozoic upper-intercept age of 1744 ± 20 Ma, which is within uncertainty of the earlier c. 1723 ± 29 Ma intercept age 13 . Both datasets show Pb-loss trends projecting to Cambrian-age lower intercepts, attesting to overprinting during Ross-aged metamorphism. Goodge et al. 13 interpreted the Palaeoproterozoic zircons from 90-131A to have formed during eclogite-facies metamorphism given that the zircons came from a rock containing an eclogite-facies mineral assemblage and the zoning features and morphologies pointed toward a possible metamorphic origin. The zircons were interpreted to have recrystallised texturally and structurally from igneous precursors. Moderate Th/U signatures (0.25-0.40), characteristic of igneous zircon, were thought to be partially retained from the igneous relicts.
However, our study shows that the trace-element compositions of these zircons are dissimilar to metamorphic zircons formed at high-pressure conditions, which typically have depleted chondrite-normalised HREE concentrations that reflect the presence of garnet 34,[37][38][39] . Rather, the comparatively enriched HREE concentrations of these zircons are typical for zircons formed in a magmatic environment or during metamorphism at conditions which do not stabilize garnet 37,40,41 . Th/U ratios between 0.23-0.48 are in the range expected for igneous zircons and agree with those determined from SHRIMP analysis 13 , although this is not uniquely diagnostic of a magmatic paragenesis given the known variability in Th/U values in metamorphic zircon 40 , particularly in mafic rocks. Among those zircon analyses for which the Eu anomaly could be calculated, Eu anomalies are only slightly negative ( Fig. 4b; Supplementary Table S1). The moderate Eu/Eu* values between 0.21-0.55 are typical of zircon in equilibrium with either magmatic or metamorphic plagioclase, the latter of which has a lower trace-element budget than magmatic plagioclase 34,35 . The textural morphologies of many of the zircons are consistent with a metamorphic origin (i.e., equant to ovoid, faceted 'soccer-ball' shapes), yet internal zoning features such as sector zones are consistent with both a magmatic and a high-grade metamorphic origin, and the presence of oscillatory zoning in some of the grains is suggestive of magmatic growth 40,42,43 (Fig. 3). By re-examining the zircons in eclogite 90-131A, we can more confidently conclude from the new trace-element data that the zircons did not form in the presence of garnet at c. 1.7 Ga. Therefore, there is no evidence for eclogite-facies metamorphism at that time. The zircons are either magmatic in origin or formed during low-pressure metamorphism.
Contrary to the earlier interpretation of Paleoproterozoic-aged eclogite-facies metamorphism in this sample 13 , we suggest that the zircons are primary igneous grains that have partially to completely retained their traceelement and isotopic characteristics through Cambro-Ordovician eclogite-facies metamorphism (Fig. 5). The primary morphologies and internal structures of the precursor igneous zircons are not known due to the effects of Ross-aged metamorphism and deformation, which likely promoted the development of new internal structures. Furthermore, it is conceivable that the ~ 5-10 µm wide-bright-CL rims on the igneous cores crystallised during Cambrian-aged high-pressure metamorphism, in agreement with previous SHRIMP ages of similar low-Th  20 . The differences in zircon ages and trace-element compositions between eclogite samples 90-131A (this study) and 90-130D 20 can be reconciled with an interpretation that the Cambrian-aged neoblastic zircon grains in eclogite 90-130D are analogous to, and coeval with, the thin bright-CL rims characterising the igneous zircons from 90-131A (Fig. 5). It should be noted that no zircon grains having a www.nature.com/scientificreports/ c. 1.7 Ga age were found in sample 90-130D. This may be a consequence of the different procedures undertaken in determining the zircon ages (e.g., grain-mounted vs. in-situ), whereby the larger c. 1.7 Ga igneous relicts obtained by mineral separation were not found in polished thin-section, and the smaller zircons (in-situ) were not recovered through traditional mineral separation techniques. An alternative possibility is that the c. 1.7 Ga zircons from 90-131A were inherited from the Palaeoproterozoic gneisses enclosing the mafic eclogitic boudins, which is supported by extensively documented c. 1.7 Ga zircon ages from Nimrod rocks in the same area 23 . Given the absence of c. 1.7 Ga zircon ages in eclogite sample 90-130D 20 , it may be possible that c. 1.7 Ga zircon inheritance did not contribute to the zircon age dataset for this sample. A third possibility is that the eclogite precursors were petrologically similar mafic igneous rocks, possibly emplaced as dikes, but having different emplacement ages (90-130D yields a very poorly defined discordant array projecting to an upper-intercept at c. 2.2 Ga whereas 90-131A yields an upper-intercept age at c. 1.7 Ga). Of these three possibilities, the first is the simplest explanation given the data available, but we cannot discount the other explanations. The results from this study, and other zircon age and trace-element patterns confirming an early Cambrian age for the Miller Range eclogite 20 , demonstrate that high-pressure metamorphism occurred as a consequence of plate convergence along the active East Gondwana margin during the early Palaeozoic Ross Orogeny. Therefore, an interpretation of c. 1.7 Ga subduction and high-pressure metamorphism along the eastern margin of the Mawson Continent is not supported by current evidence.   [24][25][26][27][28] . Therefore, despite the lack of metamorphic evidence for Palaeoproterozoic subduction and/or collisional orogenesis in most of East Antarctica, South Australia, and western Laurentia 10 (i.e., the absence of high-pressure metamorphic rocks signifying tectonic convergence and collision), metamorphism and magmatism in these regions is interpreted to have occurred as either a direct or indirect response to collision-subduction 11,15,23,44 (Fig. 1). However, the new results from zircon petrochronology indicate that the Palaeoproterozoic (c. 1720 Ma) Nimrod Orogeny in the central Transantarctic Mountains did not involve high-pressure metamorphism and subduction orogenesis. Instead, the Palaeoproterozoic record between 1740-1690 Ma is characterised by extensive magmatism and high thermal gradient metamorphism in Australia, East Antarctica, and western Laurentia 23-28,33 (Fig. 1)-a record that is incompatible with subduction-related continental collision (e.g., Himalayan-or Caledonide-style collision). Without petrologic evidence for Palaeoproterozoic-aged subduction and collisional orogenesis between Australia/East Antarctica and western Laurentia, the thermal regimes recorded within these regions are quite similar, suggesting that a single coherent tectonothermal framework dominated by low-pressure metamorphism and crustal magmatism, and involving local tectonic deformation, may have existed at this time 23-28,33 . Relation to Nuna assembly. Based on the new data presented here and elsewhere 20 , a lack of evidence for c. 1.7 Ga eclogite-facies metamorphism in East Antarctica signifies that there was not a subduction-collisional regime in operation that was associated with Palaeoproterozoic-aged Nuna assembly between proto-Australia, the Mawson Continent, and western Laurentia. The lack of evidence for Palaeoproterozoic-aged high-pressure metamorphism in Australia, East Antarctica, and North America may simply reflect that eclogite-facies rocks in the respective metamorphic records were not preserved. Poor preservation of high-pressure rocks is not uncommon due to the mechanics of subduction and collision, as well as the potential for petrologic overprinting 19,45,46 , but the absence of Paleoproterozoic eclogites throughout this extensive region of Nuna raises question about the viability of tectonic models calling for subduction-collisional orogenesis among these cratonic elements.
Palaeogeographic and tectonic models of Nuna assembly are broadly defined as having occurred in one of three periods: (1) pre-1700 Ma, (2) c. 1800-1650 Ma, and (3) at c. 1650-1550 Ma. The c. 1.7 Ga events in proto-Australia, the Mawson Continent, and western Laurentia-notably widespread and not encompassing high-pressure metamorphism-help to inform Nuna assembly models. In the first case, the petrological and geochronological connections at c. 1.7 Ga within these key cratonic elements may indicate common crustal processes operating in an already-amalgamated Nuna. The shared patterns of low-pressure metamorphism and magmatism 23-28,33 detailed earlier may have occurred within a broadly formed supercontinent. This model is www.nature.com/scientificreports/ supported by correlations between interpreted c. 1.85 Ga orogenies in northern Australia and north-west Laurentia (e.g., Barramundi and Trans-Hudson orogenies) 5,47 , and similarities between detrital zircons from rocks of the Mawson Continent and the Mojave Province 12 . However, there is no evidence for major inter-cratonic collision in these regions. Other models suggest progressive Nuna assembly by crustal thickening and accretionary processes active between approximately 1800-1600 Ma 3,11,15,44 . For example, Betts et al. 15 correlated interior basins in proto-Australia and Laurentia, and propose that accretionary processes along the margins of Australia and southwestern Laurentia between 1800 and 1600 Ma extended from the margin of proto-Australia, through the Mawson continent, to Laurentia. As outlined earlier, these models are consistent with some palaeomagnetic data 3,48 and evidence for Barrovian-style crustal thickening, thrust-style shortening, and accretionary imbrication from the magmatic, structural, and metamorphic records in Australia 10,15,25,49 , East Antarctica 23,24,28,33,50 , and North America 12,16,26,27,44,51,52 . Importantly, the evidence for Palaeoproterozoic moderate-pressure Barrovianstyle metamorphism and crustal thickening does not imply the operation of collisional orogenesis as seen in the Himalayas, Alps, and Caledonides, all of which contain high-and ultrahigh-pressure eclogites [53][54][55] . However, these processes may signify the operation of broad-scale low-pressure, moderate-to high-temperature crustal interactions between proto-Australia, East Antarctica, and western Laurentia, potentially related to the early development of Nuna in the Palaeoproterozoic 5,48 . This interpretation may be explained by a setting involving a comparatively passive connection between the Mawson Continent and western Laurentia at this time, whereby a narrow marine basin may have existed between the cratonic fragments (see Kirscher et al. 9 and references therein). Notably, the c. 1.7 Ga events recorded in proto-Australia, the Mawson Continent, and western Laurentia may be related to the ongoing processes described by these models more broadly, but they are not an indicator of subduction-collision at this time. Lastly, some models suggest that the cratonic elements within Nuna were amalgamated in a single collisional event during the Mesoproterozoic 9 , as suggested on the basis of petrological and geochronological investigations of garnet-bearing rocks from northern Australia 6,7,56 . Collision and assembly between proto-Australia and western Laurentia at c. 1.6 Ga 57 is also suggested by Mesoproterozoic-aged (c. Ultimately, the shared record of c. 1.7 Ga low-pressure metamorphism and magmatism in proto-Australia, the Mawson Continent, and western Laurentia, as well as from isotopically-similar 2.01-1.85 Ga igneous rocks, suggests a long-standing association that is inconsistent with initial Nuna assembly at 1.6 Ga. Our preferred interpretation for the assembly of Nuna follows that of Kirscher et al. 9,48 . Geological and palaeomagnetic data indicate that assembly was a protracted, multi-stage process involving: (1) the establishment as early as c. 1.8 Ga of a semi-stable Paleoproterozoic connection between proto-Australia/East Antarctica and Laurentia, possibly separated by a narrow marine basin or epicontinental sea, elements of which interacted through shortening and crustal thickening processes (see above), and (2) final reorganisation at c. 1.6 Ga characterised by crustal thickening and high-temperature processes. Despite a lack of evidence for high-and ultrahigh-pressure metamorphism during Paleoproterozoic time, the crustal record from the Nimrod Complex and correlative areas in Australia and North America indicates active metamorphism and magmatism at c. 1.7 Ga that likely reflects an early stage in the history of Nuna. Although there is no evidence in the Nimrod Complex of the activity seen elsewhere at 1.6 Ga, glacial igneous and volcanic clasts sampled from central East Antarctica and Terre Adélie (1.57 and 1.60 Ga, respectively 61,63 ) may be an expression of magmatic activity during final Nuna consolidation.

Conclusions
An interpretation of a two-stage assembly process for Nuna (or at least a protracted assembly 48 ) is supported by Palaeoproterozoic and Mesoproterozoic palaeomagnetic data 4,48 and evidence from the geological records spanning 1.7-1.6 Ga in Australia, East Antarctica, and Laurentia (e.g., 5,6,23,58,60 ). Importantly, such a model is consistent with the results from this study demonstrating the lack of evidence for subduction-related orogenesis between East Antarctica and Laurentia in the Palaeoproterozoic. Our results also highlight the importance of trace-element geochemistry to distinguish different mineral growth processes involved in polyphase formation of crustal rocks. Coupling geochronology and trace-element analysis enabled us to distinguish the disparate igneous and metamorphic stages related to early Nuna supercontinent development and much later Gondwana plate boundary convergence.

Methods
Laser-ablation inductively-coupled mass spectrometry (LA-ICP-MS) analyses were performed at the University of Adelaide using a RESOlution LR 193 nm Excimer laser and an Agilent 7900 × ICP-MS. Zircon grains from sample 90-131A were separated using traditional mineral separation techniques (heavy-liquid and magnetic www.nature.com/scientificreports/ procedures), hand-picked, and mounted in an epoxy disk as outlined by Goodge et al. 13 . Mounted zircon grains were imaged using back-scatter electron (BSE) and cathodoluminescence (CL) methods at the University of Adelaide with an FEI Quanta MLA-600 scanning electron microscope. The ICP-MS obtained isotopic concentrations from targeted domains in individual zircon grains. Measured concentrations for both unknowns and standards are given in Supplementary Table S1. Zircons were ablated using a 19 μm spot-size. Analyses were conducted with an operating energy of 30 mJ, power output of ~ 2 J/cm 2 , 30% attenuation and a total acquisition time of 60 s, encompassing 30 s of background measurement and 30 s of ablation. The primary reference material used to correct zircon isotopic concentrations was GJ-  64,65 , and the age for 91,500 is within 1% of the published age 64 . Therefore, the results from the primary and secondary reference materials are considered reliable. Thus, the propagated uncertainty associated with the upper-intercept age from 90-131A zircon (1744 ± 20 Ma) is of reasonable magnitude. The use of a comparatively young primary reference material (GJ-1) to monitor the accuracy of the U-Pb ratios of 90-131A zircons was considered a reasonable approach despite the possibility of larger 207 Pb/ 206 Pb uncertainties (relative to 207 Pb/ 206 Pb uncertainties from an older zircon standard) given the aforementioned reliability of the age of GJ-1, which was tested by determining the ages of the secondary reference materials. Trace-element concentrations were calibrated to the synthetic glass standard, NIST-610 66 . Time-resolved mass spectra were corrected for mass bias and elemental fractionation using Iolite 67 . The data reduction scheme U_Pb_Geochronology4 was used for isotopic concentrations and Trace_Elements_IS was used for trace-element concentrations. Si was used as the internal reference element for trace-element data reduction (Si = 15.32 wt% assuming a stoichiometry of 32.77 wt% SiO 2 ). www.nature.com/scientificreports/