2023 Volume 118 Issue ANTARCTICA Article ID: 230331
The Napier Complex in East Antarctica has a complex thermal history, including ultra-high-temperature (UHT) metamorphism. Geochronology, trace element, and isotope geochemistry of zircon, apatite, and monazite in three felsic gneisses collected from Harvey Nunatak were studied using a sensitive high-resolution ion microprobe (SHRIMP) for the first time. Most zircons showed nebulous to fir-tree zoning, which is a common feature of zircons in granulite facies rocks, regardless of the core or rim. The U-Pb dating and rare earth element abundance of zircon indicated that zircon crystallization by regional metamorphism continued from 2567-2460 Ma, consistent with the previously proposed timing of UHT metamorphism. The zircon grains contained a large amount of Li (59-668 ppm). Li was incorporated with Cl at the interstitial sites of the zircon structure, and the zircons crystallized in melts with an abundant supply of Li and Cl. Monazite crystallized from apatite after the UHT metamorphism events of 2071 and 1799 Ma. The U-Pb system of apatite was completely disturbed by the crystallization of monazite at 1785 Ma. In addition, the U-Pb systems of apatite and monazite were disturbed at approximately 500 Ma.
Zircon (ZrSiO4) is the most widely used mineral for geochronological and geochemical investigations owing to its favorable properties, such as high physicochemical stability, adequate content of trace elements such as U, rare-earth elements (REE) (e.g., Armstrong-Altrin et al., 2018 and references therein), Li (e.g., Ushikubo et al., 2008) and exhibits good retention of radiogenic Pb (>900 °C; e.g., Cherniak, 2010 and references therein). Trace elements in zircon serve as indicators for estimating the source melt and the environment where it crystallized. Zircon U-Pb geochronology combined with trace-element geochemistry is a powerful tool for investigating geochemistry that have experienced high-temperature environments. However, obtaining geochronological and geochemical information on lower-temperature metamorphism (amphibolite facies or under) from zircon is challenging because of its high durability. Therefore, a suitable approach to obtain geochronological information of the low-temperature metamorphism is U-Pb apatite [Ca5(PO4)3(F, Cl, OH)] geochronology (e.g., Andersen and Hinthorne, 1972), which has lower closure temperature of around 375-600 °C (e.g., Cochrane et al., 2014). Furthermore, apatite is more reactive than zircon, and monazite crystallizes within and around apatite during fluid-rock interactions (e.g., Harlov et al., 2002), which provides an excellent opportunity to obtain geochemical information on low-temperature metamorphism.
The Napier Complex is located in East Antarctica and attracts many scientists as the regional ultra-high-temperature (UHT) metamorphism was first recognized here (Dallwitz, 1968). This region has experienced extremely high temperatures (>1100 °C) based on the mineral assemblage of sapphirine + quartz during the Neoarchean and Paleoproterozoic (Harley, 2016, and references therein). The UHT metamorphism in the Napier Complex has been extensively investigated. However, the thermal activity of post-peak metamorphisms, such as fluid infiltration, has been limited because it is challenging to detect evidence of lower temperature activity using U-Pb zircon geochronology, although several studies have described these post-peak fluid-related episodes (e.g., Mitchell and Harley, 2017; Takehara et al., 2020; Turuani et al., 2022). In a remarkable study, Turuani et al. (2022) reported that the disturbance of radiogenic Pb in monazite is related to the Pb-bearing nanocrystals based on geochemical data and detailed observations of monazite collected from the Zircon point of the Napier Complex. They also reported that the monazites crystallized at ∼ 2440 Ma and experienced two radiogenic Pb redistributions, one at ∼ 1050 Ma and the other at 550 Ma, suggesting that the geochemistry of monazite is the key to revealing post-peak fluid-related episodes in the Napier Complex.
This study is the first report on the geochronology and trace element geochemistry of zircon, apatite, and monazite in Harvey Nunatak in the Napier Complex. Information on UHT metamorphism and the origin of the protolith was obtained by zircon analysis. The coexistence of apatite and monazite provides geochronological and geochemical details of the thermal activity of fluid-related events that affected the gneisses after the end of UHT metamorphism.
The Napier Complex consists of coastal outcrops, islands, and inland mountain ranges in Enderby Land and Kemp Land in East Antarctica (Fig. 1). It is composed of various granulite-facies metamorphic rocks, such as tonalitic-trondhjemitic-granodioritic (TTG) orthogneisses, granitic gneisses, mafic granulites, and paragneisses. These granulite-facies metamorphic rocks were affected by multiple thermal events, including Neoarchean UHT metamorphism. The peak temperature of the UHT metamorphism is estimated to be above 1100 °C (Harley and Motoyoshi, 2000; Hokada, 2001; Ishizuka et al., 2002). U-Pb zircon analysis using microanalyses, such as secondary ion mass spectrometry (SIMS), revealed Eoarchean protolith ages (>3.7 Ga) from Mt. Sones (Williams et al., 1984; Black et al., 1986; Harley and Black, 1997), Gage Ridge (Harley and Black, 1997; Kelly and Harley, 2005; Kusiak et al., 2013a, 2013b), Mt. Jewell, and Budd Peak (Król et al., 2020) in Enderby Land, and Aker Peaks in Kemp Land (Kusiak et al., 2021). Metallic Pb nanospheres were discovered in ancient zircon grains that experienced high-temperature metamorphism from the Napier Complex. The nanospheres led to the heterogeneous distribution of Pb in the zircon grains, which affected the isotopic measurement by microbeam analysis (Kusiak et al., 2015). Because of the effect of Pb redistribution by the metallic Pb nanospheres, the veracity of the ancient ages obtained from the Eoarchean zircons of the Napier Complex is uncertain (Harley et al., 2019). Most zircon ages of the Napier rocks, other than these Eoarchean outcrops, were derived from magmatic precursors with mid- to late-Archean ages of approximately 3300-2600 Ma. For example, protolith ages of approximately 3270 and 3267 Ma were reported from Mount Riiser-Larsen (Hokada et al., 2003), and protolith ages of approximately 2711 and 2726 Ma were reported from the Raggatt Mountains and the southern Scott Mountains (Król et al., 2022). In addition, younger protolith ages of approximately 2539 and 2522 Ma have been reported in the southern Scott Mountains (Król et al., 2022). However, post-UHT cooling and localized intrusion-related fluid incursions and lower-grade metamorphism have been documented from 2380 to 1800 Ma in the Napier Complex using Sm-Nd systematics (Suzuki et al., 2001), U-Pb chemical age obtained from xenotime and zircon (Grew et al., 2001), and U-Pb isotopic age obtained from zircon (Carson et al., 2002).
Harvey Nunatak is located between Mt. Sones and Mt. Renouard in the UHT metamorphic region of the Napier Complex (Fig. 1). This area was explored by the Geological Field Party of the 58th Japanese Antarctic Research Expedition (JARE) during 2016-2017, which is the most inland area reached by a field survey of JARE. In this study, we analyzed three rock samples collected from Harvey Nunatak by the Geological Field Party of the 58th JARE.
Opx-Pl gneiss [170223-2A-08 (HVN08)]The main constituents are orthopyroxene (Opx) and plagioclase (Pl) (Fig. 2a). quartz (Qtz) and clinopyroxene (Cpx) are rarely included locally. Apatite (Ap), zircon (Zrn), and opaque minerals were the accessory phases. The grain size of Opx was variable (0.1-5 mm). Opx occasionally showed thin lamellae of Cpx. Plagioclase has antiperthite lamellae with a diameter of 1-2 mm.
Opx, Qtz, and Pl were the major constituents of the local Grt (Fig. 2b). Zrn and opaque minerals are the accessory phases. Antiperthitic Pl (0.5-2 mm in diameter) occurred along with Qtz (0.5-1 mm in diameter). Opx showed a subhedral, elongated, or irregular shape with a grain size of 0.1-3 mm.
Opx-Pl gneiss [170223-2A-10 (HVN10)]Opx, Pl, and minor amounts of Qtz and Cpx were the main constituent minerals. Ap, Zrn, and opaque minerals are the accessory phases. Elongated Opx grains (up to 2-3 mm long grain diameter) were arranged in parallel to show gneissosity (Figs. 2c and 2d). The Cpx is typically <1 mm in diameter. Plagioclase shows antiperthite lamellae. Qtz occurs locally and is sometimes coarse-grained measuring up to 5 mm.
The rock samples were pulverized using a high-voltage pulse-power fragmentation device (Selfrag Ag, Kerzers, Switzerland) at the National Institute of Polar Research (NIPR, Tokyo, Japan). The pulverized samples were separated using conventional mineral separation techniques, such as heavy liquid separation with methylene iodide and magnetic separation using a Nd magnet. The zircons were randomly mounted together with the reference materials in epoxy resin discs. After curing, the grain mounts were polished along the cross-section through the grains. Images of transmitted light were obtained using an optical microscope, and backscattered electrons (BSE) and cathodoluminescence (CL) were obtained using a scanning electron microscope (SEM) with a Gatan mini-CL detector and a Gatan ChromaCL2 detector was used to select suitable analytical spots.
The order of the analyses, including sample surface preparation of the zircons, was as follows.
The U-Pb isotopic ages and abundances of trace elements (Y, Nb, REE, Hf, Ta, Li, F, Mg, Al, P, Cl, K, Ca, Ti, Mn, and Fe) in the zircons were analyzed using SHRIMP-IIe at NIPR. The oxygen isotope compositions of the zircons were determined using SHRIMP-IIe/AMC at the same positions used for the analyses of U-Pb ages and trace elements. In addition, the Li isotope ratios of the zircons were measured using SHRIMP-IIe. Additional information on sample preparation, including mineral separation and the analytical methods for zircon, is provided as Supplementary Document (available online from https://doi.org/10.2465/jmps.230331).
The analysis of monazite and apatiteU-Pb isotopic dating and REE abundance measurements of monazite and apatite were conducted using SHRIMP-IIe at NIPR. For the monazite U-Pb analysis, an O2− primary ion beam of ∼ 0.7 nA was used to sputter an analytical spot approximately 5 µm in diameter for monazite. Monazite U-Pb isotope data were acquired as described by Williams et al. (1996). USGS 44069 monazite (206Pb/238U age = 424.9 ± 0.8 Ma; Aleinikoff et al., 2006) was used as reference material. The fractionation between 232ThO+ and 238UO+ was corrected by a factor calculated using the correlation between 232ThO+/238UO+ and radiogenic 208Pb/206Pb, as described by Williams et al. (1996). There is common isobaric interference on 204Pb, which leads to an overcorrection of the Pb isotopic ratios. The 204Pb isobaric interference was corrected based on the measured Th content (Stern and Sanborn, 1998).
For the apatite U-Pb analysis, an O2− primary ion beam of ∼ 3 nA was used to sputter an analytical spot approximately 10 µm in diameter. The procedures for the apatite U-Pb and REE analyses followed those of Sano et al. (1999), Horie et al. (2008), Horie et al. (2004), and Gall et al. (2017). AFG2 apatite (206Pb/238U age = 478.69 ± 0.45 Ma; Kennedy et al., 2012) and Durango apatite ([U]: 9 ppm; Trotter and Eggins, 2006) were used as reference materials for the calibration of the 206Pb/238U ratio and U concentration, respectively. The 204Pb+, 206Pb+, 207Pb+, 208Pb+, 232Th+, 238U+, 242(ThO)+, 254(UO)+, 264(ThO2)+, and 270(UO2)+ peaks were measured at a mass resolution of approximately 5600 for Pb and U isotopic analyses. 175(Ca2PO4)+ and 209(PbH)+ were monitored as the derivative ion species from the major elements of apatite and Pb hydride, respectively.
Coriolano Mine monazite and Durango apatite were used as reference materials for REE analyses of monazite and apatite, respectively. Apatite TUBAF #37 (Wudarska et al., 2021) was used as the Sr reference material for apatite. The Coriolano Mine monazite and Durango apatite were crushed, and approximately 10 mg of the fragments were dissolved in 2 M HNO3. Solutions were individually diluted to 15 mL with 0.5 M HNO3, and the REE contents in the solutions were then measured by inductively coupled plasma-mass spectrometry (ICP-MS: Micromass PQ II) at Hiroshima University. Details of the analytical procedures are described in Horie et al. (2004). Additional information on the analytical method for monazite is provided in the Supplementary Material.
170223-2A-08 (HVN08). The zircon grains in sample 170223-2A-08 (HVN08) had an average size of 150 µm and were rounded with newly grown crystal faces. The CL images of typical zircon grains are shown in Figure 3a. Some zircon grains exhibited core-rim structures (Fig. 3a); the cores and rims were characterized by faint to broad zoning with a bright CL response and broad zoning, respectively (Fig. 3a). Two grains (spots 3 and 14 in Fig. 3a) include a bright CL response cores. The other zircon grains with no apparent core-rim structure were characterized by broad zoning with a dark CL response (Fig. 3a).
170223-2A-09 (HVN09). The zircon grains in sample 170223-2A-09 (HVN09) had an average size of 150 µm and were rounded with newly grown crystal faces. The grains contain small mineral inclusions of quartz and feldspar. The BSE and CL images of typical zircon grains are shown in Figure 3b. Most zircon grains exhibited broad zoning with a dark CL response (Fig. 3b). Some grains were characterized by sector zoning and were nebulous to fir-tree zoning (Fig. 3b). Some zircon grains showed dark BSE response domains, indicating that some zircon grains were affected by low-temperature hydrothermal alteration (e.g., Horie et al., 2006). In this study, the altered domain in zircons was defined as the dark BSE response domain, and the unaltered domain in zircons was defined as a domain other than the altered domain. Altered grains were defined as zircon grains containing at least one altered domain, and unaltered grains were defined as zircon grains with no observed altered domains.
170223-2A-10 (HVN10). The zircon grains in sample 170223-2A-10 (HVN10) had an average size of 150 µm and were rounded with newly grown crystal faces. CL images of typical zircon grains are shown in Figure 3c. Some zircon grains in the sample show core-rim structures (Fig. 3c). The cores were characterized by broad-to-fir-tree zoning with a bright CL response, and the rims were characterized by broad zoning (Fig. 3c). Some grains had dark CL response cores.
U-Th-Pb zircon geochronologyIn this study, concordant data were defined as <10% discordance {discordance = [1 − (206Pb/238U age)/(207Pb/206Pb age)] × 100; Song et al., 1996}. The U-Pb data with high Ca content (>100 ppm), high light REE (LREE) pattern, and high discordance (>10%) are also omitted from the discussion and figures, indicating that some zircon grains were affected by low-temperature hydrothermal alteration (e.g., Horie et al., 2006; Takehara et al., 2018). All U-Th-Pb data for the 170223-2A-08 (HVN08), 170223-2A-09 (HVN09), and 170223-2A-10 (HVN10) zircons are shown in Supplementary Table S1 (Tables S1-S4 are available online from https://doi.org/10.2465/jmps.230331).
170223-2A-08 (HVN08). A total of 32 spots on 21 zircon grains in the 170223-2A-08 sample were analyzed, and the results are shown in the Tera-Wasserburg concordia diagram (Fig. 4a) with the rejection of one data point (Spot No. 16.1) because of the high LREE pattern (Table S1). The U-Pb data in the sample ranged from 2639 to 2277 Ma for 207Pb/206Pb age. The U and Th contents of the zircons ranged from 23 to 248 ppm and from 9 to 353 ppm, respectively. The Th/U ratios range from 0.40 to 1.81 (Table S1; Fig. 4j). The bright CL response cores (spot Nos. 3.2 and 14.2) show older ages of 2639 ± 19 Ma and 2580 ± 40 Ma, respectively. The probability density plot of 207Pb/206Pb ages obtained from the cores ranged from 2639 to 2277 Ma and peaked at approximately 2460 Ma (Fig. 4b). A weighted average of 207Pb/206Pb age calculated only on the data obtained from the rim in the CL image is 2463 ± 11 Ma [n = 13, Mean Square Weighted Deviation (MSWD) = 1.4] (Fig. 4c).
170223-2A-09 (HVN09). The altered domains in the zircon grains of the 170223-2A-09 sample were avoided. A total of 38 spots on 37 zircon grains in the sample were analyzed, and the results are shown in the Tera-Wasserburg concordia diagram (Fig. 4d) with the rejection of 5 data points (4.1, 30.1, A1.4, A5.2, and A7.4) showing high Ca content and high LREE patterns (Table S1). The U and Th contents of the zircons ranged from 816 to 4254 ppm and from 70 to 678 ppm, respectively (Table S1; Fig. 4j). The high-U zircons (>2500 ppm) exhibited older 206Pb/238U ages, and the 206Pb/238U ratios of the high-U zircons in the sample were corrected using the method proposed by Williams and Hergt (2000) and White and Ireland (2012). The U-Pb data in the sample revealed a wide age distribution, ranging from 2524 to 2342 Ma in 207Pb/206Pb age (Table S1). Probability density plots of the data obtained from the unaltered domains in the unaltered grains and the unaltered domains in the altered grains are shown in Figures 4e and 4f, with the data ranging from 2524 to 2441 Ma and from 2519 to 2342 Ma, respectively. The Th/U ratios range from 0.04 to 0.26. A weighted average of 207Pb/206Pb age calculated only on the data obtained from the unaltered domain in unaltered grain was 2492.1 ± 9.7 Ma (n = 23, MSWD = 15) with the rejection of 3 data showing HREE-depleted patterns (Fig. 4e), but this weighted mean had no significance as a chronological value due to the high MSWD.
170223-2A-10 (HVN10). A total of 33 spots on 22 zircon grains in the 170223-2A-10 sample were analyzed, and the results are shown in the Tera-Wasserburg concordia diagram with the rejection of 3 data points (17.1, 17.2, and 19.1) showing high Ca contents and high LREE patterns (Table S1). The spots obtained from the cores showing larger 207Pb/206Pb errors (1.1 and 7.1) are shown as y-axis elongated ellipses in the Tera-Wasserburg concordia diagram (Fig. 4g). Pb redistribution may have affected these spots due to the Pb nanospheres (Kusiak et al., 2015). The U-Pb data in the sample range from 2567 to 2415 Ma for 207Pb/206Pb age (Table S1). The U and Th contents of the zircons ranged from 62 to 1084 ppm and from 30 to 268 ppm, respectively. The Th/U ratios range from 0.60 to 1.36 (Table S1 and Fig. 4j). The probability density plot of 207Pb/206Pb ages obtained from the cores ranged from 2567 to 2415 Ma and peaked at approximately 2500 Ma (Fig. 4h). A weighted average of 207Pb/206Pb age calculated only on the data obtained from the rim in the CL image is 2452.8 ± 9.4 Ma (n = 13; MSWD = 1.4) (Fig. 4i).
The contents of trace elements in zirconThe abundances of trace elements (Li, B, F, Mg, Al, P, Cl, K, Ca, Ti, Mn, Fe, Y, Nb, REE, Hf, and Ta) in the 170223-2A-08, 170223-2A-09, and 170223-2A-10 zircons are listed in the Supplementary Materials (Tables S2 and S3). REE abundance patterns normalized by the C1-chondrite values (McDonough and Sun, 1995) of the zircons are shown in Figures 5a-5c. Most of the REE patterns of the zircons are characterized by a large fractionation between LREE (La, Pr, and Nd) and heavy REE (HREE: Tm, Yb, and Lu), positive Ce anomalies, and negative Eu anomalies.
170223-2A-08 (HVN08). No difference was observed between the REE patterns obtained from the cores and rims, except for spot No. 16.1, which was characterized by enrichment of LREE [total light rare-earth elements (ΣLREE): 10.1 ppm] rather than those of the others (average of ΣLREE: 1.45 ppm) (Fig. 5a and Table S3).
The cores in the 170223-2A-08 zircons showed contents of Li, B, F, Mg, Al, P, Cl, K, Ca, Ti, Mn, and Fe (Li: 59.0-355 ppm, B: 0.008-0.22 ppm, F: 35.7-93.7 ppm, Mg: 0.18-56.3 ppm, Al: 1.39-215 ppm, P: 142-238 ppm, Cl: 4.35-27.5 ppm, K: 0.06-88.2 ppm, Ca: 0.19-34.4 ppm, Ti: 21.4-69.9 ppm, Mn: 0.22-9.02 ppm, and Fe: 1.59-1084 ppm). The rims in HVN08 zircons showed contents of Li, B, F, Mg, Al, P, Cl, K, Ca, Ti, Mn, and Fe (Li: 177-355 ppm, B: 0.01-0.11 ppm, F: 37.6-112 ppm, Mg: 2.64-8.75 ppm, Al: 1.36-230 ppm, P: 155-239 ppm, Cl: 9.74-22.5 ppm, K: 0.20-2.20 ppm, Ca: 0.56-5.93 ppm, Ti: 17.7-70.2 ppm, Mn: 0.30-0.83 ppm, and Fe: 10.1-46.2 ppm).
Ti-in-zircon temperatures were calculated for each grain of the HVN08 zircon based on Watson et al. (2006), assuming the activity of TiO2 to be 0.6. The Ti contents of the zircons used in the calculations and the calculated temperatures are listed in Table S2. The activity of TiO2 was assumed to be because this rock contains ilmenite. The temperatures of the cores show an average of 961 °C and range from 867 to 1015 °C. The temperatures of the rims show an average of 944 °C and range from 846 to 1016 °C.
170223-2A-09 (HVN09). The REE patterns obtained from three spots (spot Nos. 8.1, 16.1, and 20.1) in the unaltered zircon grains are characterized by a weak fractionation between middle REE (MREE: Gd, Tb, and Dy) and HREE (HREE/MREE = 0.12-0.55) rather than those of the others (HREE/MREE = 1.11-3.82) (Figs. 5b and 5d; Table S3).
The unaltered zircon grains and the unaltered domains in the grains which contain altered domains show lower and relatively homogeneous contents of B, K, Ca, Ti, and Mn (B: 0.04-2.36 ppm, K: 0.02-16.6 ppm, Ca: 0.36-43.4 ppm, Ti: 19.5-59.0 ppm, and Mn: 0.21-8.77 ppm), and relatively higher contents of Li, F, Mg, Al, P, Cl, and Fe (Li: 311-668 ppm, F: 740-1294 ppm, Mg: 67.8-3777 ppm, Al: 3.70-284 ppm, P: 210-616 ppm, Cl: 210-274 ppm, and Fe: 46.1-511 ppm). A spot of unaltered domains in the altered grains (spot no. A7.4) shows extremely high contents of K, Ca, Mn, and Fe (K: 47.7 ppm, Ca: 497 ppm, Mn: 91.7 ppm, and Fe: 4441 ppm); therefore, these spot data were excluded from the trace element contents.
Ti-in-zircon temperatures were calculated for each grain of the HVN09 zircon based on Watson et al. (2006), assuming the activity of TiO2 to be 0.6. The Ti contents of the zircons used in the calculations and the calculated temperatures are listed in Table S2. The TiO2 activity was assumed to be because this rock contains ilmenite. The temperatures of the unaltered domains in the unaltered grains show an average of 945 °C and range from 865 to 991 °C. The temperatures of the unaltered domains in both unaltered and altered grains show an average of 934 °C and range from 856 to 991 °C.
170223-2A-10 (HVN10). There was no difference between the REE patterns obtained from the cores and rims, except for spot No. 14.2, which was characterized by a lower MREE content (SLREE: 16.6 ppm) and weaker Eu anomalies (Eu/Eu*: 0.466) (Fig. 5c and Table S3).
The cores in the 170223-2A-10 zircons show contents of Li, B, F, Mg, Al, P, Cl, K, Ca, Ti, Mn, and Fe (Li: 86.7-332 ppm, B: 0.007-0.21 ppm, F: 27.3-96.2 ppm, Mg: 0.38-461 ppm, Al: 0.21-317 ppm, P: 102-201 ppm, Cl: 6.17-25.0 ppm, K: 0.04-120 ppm, Ca: 0.55-66.5 ppm, Ti: 26.7-54.1 ppm, Mn: 0.17-19.0 ppm, Fe: 5.55-852 ppm). Data obtained from three cores (spot nos. 8.1, 19.1, and 17.2) were excluded from the range of trace-element contents. The data from spot Nos. 8.1, 19.1, and 17.2 show the U-Pb data with a large error, discordant U-Pb data, and high Ca content (Ca: 589 ppm), respectively. The rims in HVN10 zircons show contents of Li, B, F, Mg, Al, P, Cl, K, Ca, Ti, Mn, and Fe (Li: 96.9-413 ppm, B: 0.007-0.29 ppm, F: 49.4-113 ppm, Mg: 1.49-476 ppm, Al: 10.8-291 ppm, P: 144-208 ppm, Cl: 11.6-26.1 ppm, K: 0.08-37.1 ppm, Ca: 0.32-47.8 ppm, Ti: 29.3-59.3 ppm, Mn: 0.30-10.7 ppm, and Fe: 14.3-377 ppm). A spot on the rims (spot No. 17.1) shows the discordant U-Pb data (Disc: 20%), so the contents of trace elements are excluded from the range of the trace element contents.
The Ti-in-zircon temperatures were calculated for each grain of the HVN10 zircon based on Watson et al. (2006), assuming the activity of TiO2 to be 0.6. The Ti contents of the zircons used in the calculations and the calculated temperatures are listed in Table S2. The TiO2 activity could be assumed due to the ilmenite in this rock. The temperatures of the cores show an average of 931 °C and range from 892 to 980 °C. The temperatures of the rims showed an average of 951 °C and ranged from 902 to 992 °C.
Oxygen isotope ratio in zirconThe oxygen isotope ratios (δ18O) of the 170223-2A-08, 170223-2A-09, and 170223-2A-10 zircons are shown in Figure 6.
170223-2A-08 (HVN08). The δ18O values in the cores of the 170223-2A-08 zircons range from 4.06 to 6.64‰, and the simple average is 5.57 ± 0.32‰. Those in the rims range from 3.96 to 6.71‰, and the simple average is 5.42 ± 0.46‰.
170223-2A-09 (HVN09). Oxygen isotope ratios (δ18O) in the unaltered grains of the 170223-2A-09 zircons range from 4.31 to 5.34‰, and the simple average is 4.93 ± 0.12‰. The δ18O values in the unaltered domains in the altered grains spread from 2.88 to 5.78‰, and the simple average is 4.87 ± 0.91‰.
170223-2A-10 (HVN10). The δ18O values in the cores of the 170223-2A-08 zircons range from 3.27 to 6.06‰, and the simple average is 5.04 ± 0.44‰. Those in the rims range from 4.07 to 5.52‰, and the simple average is 4.73 ± 0.29‰.
Lithium isotope ratio in zirconThe lithium isotope ratios (δ7Li) of the unaltered grains of the 170223-2A-09 zircons are shown in Table S2. The δ7Li values in the zircons show a large variation, ranging from 2.79 to 12.74‰.
Apatite and monazite descriptionApatite grains in the 170223-2A-09 sample are euhedral and range from 20 to 400 µm in size. The backscattered electron (BSE) images of typical apatite grains observed in a thin section are shown in Figure 7. Monazite grains co-exist within and surround the apatite grains and are less than 10-50 µm in size (Fig. 7).
A total of 14 spots on 13 monazite grains in the 170223-2A-09 sample were analyzed, and the results are shown in the Tera-Wasserburg concordia diagram (Fig. 8a; Table S4). The U-Pb data in the sample ranged from 1672 to 2094 Ma in 207Pb/206Pb age, including three spots with discordant U-Pb data (spot Nos. 3.1, 3.2, and 7.5; Fig. 8a). The U and Th contents of the monazites ranged from 289 to 2210 ppm and from 142 to 13750 ppm, respectively (Table S4). The U-Pb monazite data were divided into two groups. The older group (n = 7) consists of concordant data and yields a weighted average 207Pb/206Pb age of 2071 ± 12 Ma (MSWD = 0.86) (Fig. 8b). The younger group (n = 7) showed a discordant data array and yielded an upper intercept at 1799 ± 35 Ma and a lower intercept at 500 ± 180 Ma (MSWD = 0.092) (Fig. 8c).
The REE abundance patterns normalized by the C1-chondrite values (McDonough and Sun, 1995) of the 170223-2A-09 monazites are shown in Figure 9a and Table S5. The REE patterns of the monazites are characterized by the enrichment of LREE with a large fractionation between LREE and HREE and negative Eu anomalies. The degree of fractionation between LREE and HREE differed between the groups (older: 0.000158-0.000537; younger: 0.000017-0.000099).
A total of 18 spots on 6 apatite grains in the 170223-2A-09 sample were analyzed. The apatite grains contained high amounts of common Pb (204Pb/206Pb: 0.00037-0.0092). The data were projected onto a Tera-Wasserburg concordia diagram (Fig. 8d; Table S4). Based on 3-dimensional regression, the corrected U-Pb data yield an upper intercept at 1785 ± 170 Ma and a lower intercept at 553 ± 240 Ma (MSWD = 0.88). The U and Th contents in apatite ranged from 9 to 204 ppm and from 2 to 544 ppm, respectively.
The REE abundance patterns normalized by the C1-chondrite values (McDonough and Sun, 1995) of apatite are shown in Figure 9b and Table S5. The apatites have ‘bell shape’ REE patterns with negative Eu anomalies and the various LREE (ΣLREE: 1403-5182 ppm) and HREE (ΣHREE: 4-71 ppm) contents. The fraction between LREE and MREE varies (MREE/LREE: 0.09-1.16).
The zircon grains in samples 170223-2A-08 and 170223-2A-10 had core-rim structures. The cores had faint, broad, and fir-tree zoning with a bright CL, except for some inherited cores (spot Nos. 3.2, 2639 Ma and 14.2, 2580 Ma for 170223-2A-08). Fir-tree structures are a common feature of zircons in granulite-facies rocks (e.g., Vavra et al., 1996; Corfu et al., 2003). The 207Pb/206Pb ages of the cores ranged from 2525 to 2277 Ma for the 170223-2A-08 sample and from 2567 to 2415 Ma for the 170223-2A-10 sample, suggesting that regional metamorphism occurred after 2567 Ma. The weighted averages of 207Pb/206Pb age obtained from the rims are 2463 ± 11 Ma and 2452.8 ± 9.4 Ma, respectively, indicating that the last crystallization during regional metamorphism occurred around 2460 Ma. This age is consistent with the end of the UHT metamorphism (e.g., Hokada et al., 2004). Takehara et al. (2020) suggested a U-Pb system disturbance after crystallization during regional metamorphism at ∼ 2500 Ma, and ages younger than 2460 Ma are attributed to the U-Pb system disturbance. Therefore, regional metamorphism continued from 2567 to 2460 Ma. However, the paucity of igneous cores in the zircons indicated that most of these zircons recrystallized under high-temperature conditions.
The Th/U ratios of these zircons are 0.40-1.81, with rejections of 2 spots of HVN10 zircon’s core (Fig. 4j). Previous studies suggest that zircon domains with Th/U ratios lower than 0.1 can be interpreted as crystallizing during high-grade metamorphism because non-essential structural constituent cations are purged from the recrystallized structure (Hoskin and Black, 2000), and Th is consumed by the crystallization of Th-rich minerals such as monazite (Williams and Claesson, 1987; Schiøtte et al., 1989; Kinny et al., 1990). The U-Pb data of the monazite in the 170223-2A-09 sample were younger than 2071 Ma (Figs. 8a and 8b). Therefore, moderate Th/U ratios suggest a lack of monazite crystallization during regional metamorphism.
The REE patterns of the zircon grains in samples 170223-2A-08 and 170223-2A-10 were consistent, regardless of age. The LREE-enriched pattern obtained from spot No. 16.1 in the 170223-2A-08 sample is probably derived from the altered domains surrounding the fractures (Takehara et al., 2018). It is possible that the REE pattern of spot No. 14.2 in the 170223-2A-10 sample was derived from different crystallization conditions; however, further studies are necessary.
The Ti-in-zircon temperature, calculated from the Ti content in zircons, reflects the temperature at which the zircon crystallized (e.g., Watson et al., 2006). The minimum temperatures of each sample are about 846 °C for HVN08 zircon, 856 °C for HVN09 zircon, and 865 °C for HVN10 zircon. There is no correlation between the 207Pb/206Pb age and the Ti content (Ti-in-zircon temperature), as shown in Figure 4k, but these Ti-in-zircon temperatures support that the zircons crystallized during UHT and HT metamorphism.
The oxygen isotope ratios were consistent between the core and rim regardless of age. For the 170223-2A-08 zircons, the average δ18O values of the core and the rim are 5.57 ± 0.32 and 5.42 ± 0.46‰, respectively. For the 170223-2A-10 zircons, the average δ18O values of the core and the rim are 5.04 ± 0.44 and 4.73 ± 0.29‰, respectively. These data are consistent with that of zircon in equilibrium with the mantle materials (5.3 ± 0.3%; Valley, 2003); however, it does not indicate that the protoliths of the sample rocks are derived from the mantle because these zircons crystallized during UHT metamorphism. The oxygen isotope ratios of Napier zircons were reported by Kusiak et al. (2013a); The δ18O in zircons of paragneiss (Sample 14178-1) of Mount Sones range from 7.2 to 8.9‰ with an average of 8.1 ± 0.2‰, the δ18O in zircons of sapphirine-bearing paragneiss (Sample 11178-1) of Dallwitz Nunatak range from 6.8 to 8.0‰ with an average of 7.4 ± 0.1‰, and the δ18O in zircons of orthogneiss (Sample 16178-2) of Gage Ridge range from 4.7 to 6.8‰ with an average of 5.8 ± 0.1‰. The δ18O in the zircons from the three Harvey Nunatak samples show lower averages than the paragneisses of Mount Sones and Dallwitz Nunatak. However, the 170223-2A-08 zircons’ cores of Harvey Nunatak show an average of 5.57 ± 0.32‰, which agrees with the average of 5.8 ± 0.1‰ from the orthogneiss of Gage Ridge.
Interpretation of U-Pb data of the 170223-2A-09 sampleThe zircon grains in the 170223-2A-09 sample had no core-rim structure and were characterized by sector zoning and nebulous to fir-tree zoning, which is a common feature of zircon in granulite facies rocks (Fig. 3b). The U-Pb data of the unaltered domains of the 170223-2A-09 zircons are scattered from 2524 to 2342 Ma, and the weighted average is 2493.1 ± 9.7 Ma (MSWD = 15) (Fig. 4e). This weighted average age is statistically meaningless owing to the large MSWD, which indicates a mixture of multiple elements. The largest peak was centered at ∼ 2470 Ma and was consistent with the weighted average ages of the rims of the 170223-2A-08 and 170223-2A-10 samples. Therefore, it is considered that the zircons crystallized by regional metamorphism continued from 2524 to 2342 Ma, similar to the 170223-2A-08 and 170223-2A-10 samples.
The REE patterns of the 170223-2A-09 zircons were similar to those of the 170223-2A-08 and 170223-2A-10 zircons. The LREE-enriched patterns shown in spot Nos. A1.4, A5.2, and A7.4 are probably affected by the altered domains. In addition, the LREE-enriched patterns obtained from spot Nos. 4.1 and 30.1 were probably derived from the altered domains surrounding the fractures (Takehara et al., 2018). The HREE-depleted patterns shown in spot Nos. 8.1 (2524 Ma), 16.1 (2502 Ma), and 20.1 (2522 Ma) are similar to the REE patterns of zircon co-existing with garnet, which were shown in the sample of Mt. Riiser-Larsen reported by Hokada and Harley (2004). Spot No. 8.1 also indicates lower Ti content and lower Ti-in-zircon temperature (about 865 °C) in the data obtained from HVN 09 zircons, which show the average Ti-in-zircon temperature of 934 °C. The garnets localized in the 170223-2A-09 sample indicate that zircon grains were not entirely related to the garnets during regional metamorphism. The REE pattern of spot No. 23.1 (2520 Ma) is similar to that of spot No. 14.2 in the 170223-2A-10 sample. There is a possibility that the REE pattern is derived from different crystallization conditions or inheritance affected by the U-Pb system disturbance; however, further studies are necessary.
The oxygen isotope ratios were consistent regardless of age, suggesting that the oxygen atoms were taken up into the zircons from the environment that was buffered from the minerals or melts, except for the zircon domains in the altered grains. The average δ18O value is 4.93 ± 0.12‰ and slightly lower than that of zircon in equilibrium with the mantle materials (5.3 ± 0.3%; Valley, 2003), but it does not indicate that the protoliths of the sample rocks are of the mantle derivation because these zircons crystallized during UHT metamorphism. The δ18O values in the unaltered domains of the altered grains are widely scattered compared to those of the unaltered grains and are probably attributed to the altered domains surrounding the fractures.
Zircon geochemistry and geological implicationsTrace element abundance and isotope signatures provide an opportunity to deduce the crystallization environment of zircon and its geological setting. Zircon trace element discrimination diagrams between Nb/Yb versus U/Yb and Hf versus U/Yb help classify the origin of the source melts that form zircons into mantle, continental arc, and oceanic arc (Grimes et al., 2015; Schmitt et al., 2018). The discrimination diagram between Nb/Yb and U/Yb reveals that the 170223-2A-08 and 170223-2A-10 zircons fall around the boundary between the mantle-derived melt and continental arc melt, and there is no difference between the cores and rims (Fig. 10). In addition, zircons 170223-2A-08 and 170223-2A-10 fell in the continental arc melt region on the discrimination diagrams between Hf content and U/Yb. It is important to note that Yb and Nb in the zircons may have been fractionated during UHT metamorphism. For example, Yb and Nb are preferentially incorporated into metamorphic garnet and rutile, respectively. Rutile was not observed in 3 sample rocks (170223-2A-08, 170223-2A-09, and 170223-2A-10); however, garnets were observed in sample No. 170223-2A-09. Some spots in the 170223-2A-09 zircons showed HREE-depleted (low HREE/MREE ratio) patterns (Fig. 5), which also showed higher U/Yb and Nb/Yb ratios than other spots (Fig. 10). This is explained by decreased Yb content in the zircons due to fractionation with metamorphic garnets. Therefore, the U/Yb and Nb/Yb plots, particularly those showing the HREE-depleted patterns for the 170223-2A-09 zircons, do not directly represent the geochemical characteristics of the protoliths. If some of the 170223-2A-08 and 170223-2A-10 zircons retained some of the chemical information of the protoliths, it is possible that the protoliths formed from a mixture of continental materials.
Previous studies have suggested that Li+ is important for the charge balance of REE3+ in the zircon structure, based on the correlation of Li content with REE, Y, and P contents (Ushikubo et al., 2008; Bouvier et al., 2012). The ionic radius of Li+ in eight-fold coordination with oxygen is 0.092 nm, which is slightly larger than that of Zr4+ in eight-fold coordination (0.084 nm), but smaller than that of REEs, indicating that Li ions have an appropriate size to substitute for ZrVIII in the zircon structure. REE substitute Zr via xenotime-type substitution (Finch et al., 2001; Hanchar et al., 2001).
| \begin{equation*} (\text{Y} + \text{REE})^{3+} + \text{P}^{5+} = \text{Zr}^{4+} + \text{Si}^{4+} \end{equation*} |
Experiments have shown that zircons grown in a Li-Mo flux can contain more than 100 ppm Li when REEs are present (Hanchar et al., 2001), suggesting that Li concentration in zircon is strongly linked to the REE abundance in zircon according to the following equation:
| \begin{equation*} \text{Li}^{+}_{\text{(interstitial)}} + (\text{Y} + \text{REE})^{3+} = \text{Zr}^{4+} + \text{X}_{\text{(interstitial)}}. \end{equation*} |
If Li and P compensate for the REE, the ratio defining these two coupled substitutions, [(Y + REE)/(Li + P)]atomic, should be 1. However, The (Y + REE)/(Li + P) ratios in this study are clearly low at 0.09-0.41 for 170223-2A-08, 0.07-0.22 for 170223-2A-09, and 0.07-0.45 for 170223-2A-10 zircons. Therefore, this could not be explained by substitution reactions. In contrast, Bouvier et al. (2012) proposed the following reactions unrelated to the substitution of P and REE:
| \begin{align*} &\text{Li}^{+} + 3\text{Li}^{+}_{\text{(interstitial)}} = \text{Zr}^{4+} + 3\text{X}_{\text{(interstitial)}} \\ &4\text{Li}^{+}_{\text{(interstitial)}} + \text{X} = \text{Zr}^{4+} + 4\text{X}_{\text{(interstitial)}}. \end{align*} |
The Li content was correlated with the Cl content (Fig. 11a), which suggests that Li was incorporated with Cl in the interstitial sites of the zircon structure. There was no correlation between the Li and F content, and Cl played an important role in Li incorporation. Therefore, the Li content of the zircons from the 3 samples was mainly dominated by the above reactions that incorporated Li into the interstitial sites. The Li contents in the cores of the 170223-2A-08 and 170223-2A-10 zircons were 59.0-387 and 86.7-413 ppm, respectively, and were higher than those in zircons crystallized from mantle-derived melts (typically <2 ppb; Ushikubo et al., 2008). The high Li contents of the zircons suggest that they crystallized in the melts, as reflected by the breakdown of mica during UHT metamorphism.
The 170223-2A-09 zircons showed higher U (816-4254 ppm; Table S1) and Li (311-668 ppm; Fig. 11b) content than those of the 170223-2A-08 and 170223-2A-10 zircons. As shown in Figure 10, the U/Yb ratios of the 170223-2A-09 zircons were higher than those of the 170223-2A-08 and 170223-2A-10 zircons. The 170223-2A-09 zircons show fractionated Li isotope ratios (δ7Li: −2.79-12.74‰), depicting at least two peaks in the probability density plot (Fig. 11c). The discrimination diagrams and Li isotope ratios themselves suggest the contribution of weathered surface materials (Ushikubo et al., 2008), but the 170223-2A-09 sample contains garnets that must have affected the Yb content in the zircon, especially showing HREE-depleted patterns. Therefore, the 170223-2A-09 zircons do not suggest the contribution of continental surface materials to this sample. In addition, it is based only on the geochemical data obtained from Harvey Nunatak in this study, and the contribution of continental surface material to the protolith of metamorphic rocks on the entire Napier Complex requires additional data. The Cl contents of the 170223-2A-09 zircons were outside the correlation line between the Li and Cl contents. The Mg content of the 170223-2A-09 zircons (67.8-3777 ppm) was higher than those of the 170223-2A-08 (0.18-56.3 ppm) and 170223-2A-10 (0.38-476 ppm) zircons. Chlorine is important for the charge balance in the following Mg2+ substitution: Mg2+ + 2Cl− + 2H2O-Zr4+ + (SiO4)4− (modified after Caruba and Iacconi, 1983).
Thermal activity at post-peak metamorphismThermal activities after UHT metamorphism, such as fluid infiltration, have been reported in the Napier Complex. Apatite provides an opportunity to deduce thermal activity after high-temperature metamorphism owing to its lower closure temperature and higher reactivity. The U-Pb data showed a discordant data array and yielded an upper intercept at 1785 ± 170 Ma and a lower intercept at 553 ± 240 Ma. The upper intercept age of 1785 ± 170 Ma is consistent with the zircon ages of the alteration domain adjacent to the Early Paleozoic pegmatite on Tonagh Island (∼ 1930-1800 Ma; Carson et al., 2002) and the youngest zircon age distribution of felsic orthogneiss collected from the Fyfe Hills (1824 Ma; Horie et al., 2012). The younger age distribution represents a disturbance of the U-Pb system in the zircon grains by local fluid infiltration.
In contrast, the U-Pb data of monazite co-existing with apatite grains were divided into two groups based on age distribution and REE patterns (Fig. 9a). The younger group with the lower-Gd patterns also shows a discordant data array and yields the upper intercept at 1799 ± 35 Ma and the lower intercept at 500 ± 180 Ma and is consistent with the U-Pb apatite ages. The common REE pattern of apatite is characterized by a simple decrease from LREE to HREE, whereas the patterns of the 170223-2A-09 apatite are ‘bell shape’ (Fig. 9b). The ‘bell shape’ REE patterns are derived from the consumption of LREE by the crystallization of monazite. Therefore, the U-Pb system of apatite was completely disturbed by the crystallization of monazite at approximately 1785 Ma. In addition, the U-Pb systems of apatite and monazite were disturbed at approximately 500 Ma.
The younger monazites, which show lower-Gd REE patterns, also show lower Th content (0.01-0.29 wt%) because of the similarity of the ionic radii of [VIII]Th4+ and [VIII]Gd3+. The lower Th content in the monazite can be explained by its hydrothermal origin (e.g., Schandl and Gorton, 2004). Schandl and Gorton (2004) suggested that the low ThO2 content of hydrothermal monazites (<1, or 0.88 wt% as Th content) is distinct from that of igneous one (3 to >5 wt%, or 2.6-4.4 wt% as Th content). The older group of monazites with higher-Gd patterns show higher Th contents (0.21-1.38 wt%), but this group is also of hydrothermal origin rather than igneous based on the Th content. The older group of monazites with higher-Gd patterns yields a weighted average 207Pb/206Pb age of 2071 ± 12 Ma and is consistent with the zircon age peak of felsic orthogneiss collected from Fyfe Hills (2100 Ma; Takehara et al., 2020). The younger age distribution represents a disturbance of the U-Pb system in the zircon grains by local fluid infiltration. Therefore, the monazite grains crystallized in two stages: ∼ 2071 and ∼ 1799 Ma.
The U-Pb ages of monazite from Zircon Point, Napier Complex, were reported by Turuani et al. (2022). Monazite crystallized at ∼ 2440 Ma and was affected by two episodes of radiogenic Pb redistribution. These two episodes occurred at approximately 1050 and 550 Ma and were associated with the crystallization of radiogenic Pb-bearing nanocrystals. In the case of the monazite ages from Harvey Nunatak, there were no data at approximately 1050 Ma, but the lower intercept age of 500 ± 180 Ma in the 170223-2A-09 monazites can be compared to the later episode of Pb mobilization at ∼ 550 Ma.
Geochronological and trace element geochemical analyses were performed on zircon, apatite, and monazite from three felsic gneisses collected from Harvey Nunatak in the Napier Complex. This study is the first report on geochronology in Harvey Nunatak, and the following conclusions can be drawn:
The samples investigated in this study were collected during fieldwork in the 2016-2017 Japanese Antarctic Research Expedition (JARE) supported by the National Institute of Polar Research (NIPR) under MEXT. We thank the members of the 58th JARE, especially Prayath Nantasin, Nugroho Setiawan, Davaa-ochir Dashbaatar, and Yoichi Motoyoshi, for their cooperation in the field. We are also grateful to M. Shigeoka for helping prepare the thin sections. We thank S. Harley and M.A. Kusiak for their careful review and constructive comments and M. Satish Kumar for editorial handling. This study was supported by the National Institute of Polar Research (NIPR) through Project Research No. KP-306, and JSPS KAKENHI Grant Numbers JP21H01182, JP18K05020, and JP17H02976.
Supplementary Document and Supplementary Tables S1-S4 are available online from https://doi.org/10.2465/jmps.230331.
