Apatite U-Pb Dating and Composition Constraints for Magmatic–Hydrothermal Evolution in the Giant Renli Nb-Ta Deposit, South China

: Apatite is a nearly ubiquitous accessory phase in igneous rocks that crystallizes during the entire magma evolution process and has great implications for geochronology and petrogenesis. Previous studies suggested that Nb-Ta mineralization in the giant Renli deposit was genetically related to Late Jurassic two-mica monzogranite or Early Cretaceous muscovite monzogranite. Moreover, the magmatic–hydrothermal evolution of these two stages is poorly understood. In our study, we conﬁrm that the muscovite monzogranite, biotite monzogranite, and two-mica monzogranite are all spatially associated with Nb-Ta pegmatites. We present new apatite U-Pb ages to constrain the timing of Nb-Ta mineralization and related magmatism. The results show that apatite from the two-mica pegmatite yield a lower intercept age of 130 ± 2 Ma (2 σ ), and apatite grains from two two-mica pegmatite samples yield a lower intercept age of 135 ± 8 Ma (2 σ ) and 134 ± 3 Ma (2 σ ), respectively. Apatite and whole-rock geochemistry suggest the oxidation degree of the Nb-Ta mineralization increases from north (RL-6) to south (RL-16) in the giant Renli deposit. This study demonstrates that a combination of apatite composition and U-Pb ages can be used to constrain the magmatic–hydrothermal evolution of granite and pegmatite-type Nb-Ta deposits.


Introduction
The Nb-Ta deposits represent an important source of critical metals and minerals, and therefore they play an important role in the national defense industry, the aerospace industry, the nuclear industry, super-hard materials industry, and other industries [1,2]. Apatite is a ubiquitous accessory mineral in granite and pegmatite-type Nb-Ta deposits and can carry many trace elements during the entire magmatic-hydrothermal evolution process [3,4]. The occurrence of trace elements (e.g., U, Th, Sr, and REE) in apatite is controlled by melt/phosphate mineral equilibria, the composition of the host rock, and physico-chemical conditions [4][5][6][7]. Therefore, apatite can be used to decipher petrogenesis and constrain magmatic processes. The crystal structure of apatite and the radioactive decay of U and Th make it an ideal mineral for thermochronology and yields crystallization ages of rapidly cooled plutonic rocks using the U-Pb system [8][9][10][11]. LA-ICP-MS (Laser ablation-inductively coupled plasma-mass) apatite U-Pb dating can exhibit weighted mean  [19,27]).

Figure 2.
Geological map of the Renli Nb-Ta deposit (modified after [14]). RL is the abbreviation for Renli.
The magmatic activity in the area was intense, and the outcropping magmatic rocks contain the Neoproterozoic and Yanshanian granitic rocks. The magmatic rocks can be divided into three types: (1) The large-size Yanshanian Mufushan granitoid found in the north, and composed of the Late Jurassic coarse-to medium-grained biotite monzogranite and Early Cretaceous fine-to-medium-grained muscovite monzogranite. Biotite monzogranites appear in a large area of the batholith and are divided into marginal and transitional phases. The muscovite monzogranite is in a rocky area and has a clear boundary with the first stage intrusion, generally sericitization and albitization [30]. They are closely related to rare-metal mineralization [19]. (2) The Neoproterozoic coarse-to-medium-grained biotite monzonitic granite is outcropped in the west (including the Meixian and Sandun plutons).
(3) The Neoproterozoic medium-to-fine-grained two-mica plagioclase granite is developed in the southeast (Figure 2). The biotite monzogranite and two-mica monzogranite are closely spatially related to pegmatites (Figure 3a,b). The biotite monzogranites show a porphyritic structure in hand specimens and consist of quartz (35-45 vol  A total of 926 pegmatite veinlets have been discovered in the Renli Nb-Ta deposit, and 712 and 214 veinlets crosscut into the granite intrusions and the Lengjiaxi group schist, respectively [15,19]. Nb-Ta pegmatite veinlets occur mostly in the Lengjiaxi group adjacent to the contact zone and cut the slate and schist of the Lengjiaxi group and the Mufushan composite granite (Figures 3c,d and 4f). The occurrences of Nb-Ta-rich pegmatite dikes are controlled by the bedded structure of sedimentary rocks and Mesozoic granite [14]. From the northeast to the southwest of the Renli Nb-Ta deposit, the volume of the veins decreases, and the mineralization type become complex (Be → Be + Nb + Ta → Be + Nb + Ta + Li) [13,27].

Samples and Methods
Samples RL-6 and RL-10 were collected from the No. 2 pegmatite, and RL-16 was collected from the No. 3 pegmatite in the Renli deposit ( Figure 2). Other samples were all collected near the central part of deposit. All samples were analyzed for whole-rock geochemistry, LA-ICP MS apatite U-Pb dating, and apatite geochemistry. After crushing the rock samples, apatite grains were separated by standard heavy liquid and magnetic separation techniques, and then handpicked under a binocular microscope, mounted in an epoxy resin, and polished. Before apatite U-Pb dating and analysis of major and trace elements, samples were polished into thin sections (50 µm thickness) for microscopic observation; backscattered electron (BSE) and cathodoluminescence (CL) imaging of the apatite grains was undertaken at the Guangzhou Tuoyan Analytical Technology Co., Ltd., Guangzhou, China.

Whole-Rock Geochemistry Analyses
The whole-rock geochemistry was analyzed at the Guangzhou Tuoyan Analytical Technology Co., Ltd., Guangzhou, China. All the samples from pegmatites were ground to a c. 200 mesh size in an agate mortar. Major element content was determined using an X-ray fluorescence (XRF) spectrometer (XRF-1500) and fused glass discs. Ferric and ferrous iron measurements were performed via wet chemical analyses: acid decomposition and titration with potassium dichromate [15]. The analytical precisions were ≤0.01 wt.% for major elements analyses. An Agilent 7500a system was used to determined trace element concentrations by inductively coupled plasma-mass spectrometry (ICP-MS). The detailed experimental procedure can be found in Xiong et al. [14]. The analytical precisions were ≤5% for the trace element and rare-earth element (REE) analyses.

Major and Trace Element Analyses of Apatite
The chemical compositions of apatite were analyzed at the Guangzhou Tuoyan Analytical Technology Co., Ltd., Guangzhou, China, with a JEOL JXA-8100 Electron Micro Probe Analyzer (EMPA) equipped with four wavelength-dispersive spectrometers (WDS). Before the analysis, the samples were first coated with a thin conductive carbon film. We used an accelerating voltage of 15 kV, a beam current of 20 nA, and a 5 µm spot size. The peak counting time was 10 s for Ca and P, and 20 s for Na, Mg, Si, Fe, Mn, Sr, F, and Cl. The following standards were used: olivine (Fe), rhodonite (Mn), diopside (Mg), celestite (Sr), phlogopite (F), tugtupite (Cl), apatite (Ca, P), and jadeite (Na, Si). Precision was generally better than 5% for element content > 0.5 wt.% and better than 1% for element content > 10 wt.% [4]. LA-ICP-MS analysis of trace elements in apatite was also carried out at the Guangzhou Tuoyan Analytical Technology Co., Ltd., Guangzhou, China, using an Agilent 7500a ICP-MS and Geolas 193 nm laser. The SRM 610 glass standards and the two MAD glasses were repeatedly analyzed after every eight apatite samples. A 33 µm spot size, a 10 Hz repetition rate, and a corresponding energy density of~3 J/cm 2 were used. The Ca measured by EMPA was used as the internal standard to correct the low and high content of trace elements in apatite.

LA-ICP-MS Apatite U-Pb Dating
Apatite U-Pb dating was conducted at the Guangzhou Tuoyan Analytical Technology Co., Ltd., Guangzhou, China, by a laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) instrument using a Resolution 193 nm laser-ablation system coupled to a Thermo iCAP RQ ICP-MS instrument. The Madagascar apatite (MAD) U-Pb ages of 485.0 ± 1.7 Ma were used as the external isotopic calibration reference material [33]. The 206 Pb/ 238 U ratio is 0.0762 and 207 Pb/ 235 U ratio is 0.6013 for the MAD measurements. Each set of eight or nine sample analyses was followed by one measurement of SRM 610 and three measurements of MAD. Before analysis, single apatite grains were mounted in epoxy and polished. Then, they were ablated in a continuous He stream. Before entering a Thermo iCAP RQ ICP-MS, the He stream was mixed with N 2 and Ar downstream. Samples and standards were ablated with a 33 µm laser spot, a repetition rate of 8 Hz, and a 4 J/cm 2 energy density. Furthermore, the 206 Pb/ 238 U ages are reported at an uncertainty level of 2σ [19]. The detailed data reduction process can be found in Chew et al. and Petrus et al. [34,35].

Apatite Texture
The micrograph and backscattered electron (BSE) images of apatite from the Renli Nb-Ta deposit are shown in Figures 4 and 5. The pegmatite consists of quartz, K-feldspar, albite, muscovite, beryl, garnet, tourmaline, and lepidolite and has a figurative structure (Figures 3e-i and 4a,b,d,e). Apatites from pegmatite of the Renli Nb-Ta deposit is subhedral to anhedral and up to 50-100 µm in length (Figure 4g-i). The apatites also show homogeneous texture without obvious fractures and inclusions in BSE images (Figures 4g-i and 5). They usually occur as a single mineral or as an inclusion in other minerals.

Whole-Rock Geochemistry
Major and trace elements composition of pegmatite from the Renli Nb-Ta deposit are listed in Tables 1 and 2. The levels of loss on ignition (LOI) of most samples are less than 1 wt.%, indicating that post-magmatic alteration or weathering is not obvious. The pegmatite has a relatively wide range of chemical compositions in the Renli Nb-Ta deposit. Al 2 O 3 and K 2 O content broadly decrease with increasing SiO 2 ; the content of Na 2 O increases with increasing SiO 2 , whereas Cao, MgO, and Fe 2 O 3 content do not show an obvious correlation with SiO 2 ( Figure 6).     The chondrite-and primitive-mantle-normalized REE [36] and trace element patterns for the studied pegmatite are shown in Figure 7. The average total rare earth element (ΣREE) contents of pegmatite in Renli is 56.63 ppm. The (La/Yb) N ratios range from 1.37 to 24.72, and the Eu/Eu* ratios range from 0.11 to 2.49 in the Renli deposit. The pegmatite from Renli has high LREE/HREE (3.03-43.63). The pegmatites from Renli are enriched in light rare earth elements (LREEs) and depleted in medium (MREEs) and heavy rare earth elements (HREEs) slightly.

Major Elements
The major element composition of apatite from the RL-6, RL-10, and RL-16 samples are summarized in Table 3

Trace Elements
The trace element composition of apatite from the RL-6, RL-10, and RL-16 samples are shown in Table 4

Origin of the Apatite from the Renli Deposit
Many studies have considered that the apatite includes magmatic (mixing, inheritance, and composition variation of the host magma) and hydrothermal apatite [38][39][40][41][42]. Among these, magma mixing and inheritance can lead to sudden changes in the composition of apatite, but the REE in apatite is not easily redistributed [4,43]. RL-6, RL-10, and RL-16 apatite show similar chondrite-normalized REE patterns and a gradual decrease in ΣREE. Therefore, magma mixing and inheritance are unlikely.
The levels of LOI of RL-6, RL-10, and RL-16 are less than 1 wt.%, and they do not have porosity, which indicates that post-magmatic alteration or weathering is not obvious [44]. The unsaturated fluid of REE infiltrates through apatite, and monazite inclusions will be formed in apatite [45,46]. Monazite will extract LREE from the more fractionated melts, resulting in the depletion of these elements in apatite [47]. The contents of LREE in RL-6, RL-10, and RL-16 are enriched, and the LREE/HREE ranges from 3.61 to 24.07 (Figure 8i). Therefore, hydrothermal alteration is unlikely. Moreover, the apatites have a subhedral to anhedral structure and the surface of backscattered electron (BSE) images are homogeneous without fractures (Figure 4g-i). The chondrite-normalized REE patterns of RL-6, RL-10, and RL-16 apatite vary systematically and are similar to those of pegmatite in the Renli deposit (Figure 7), suggesting that is the result of gradually magmatic evolution. Cl in apatite will preferentially combine with LREE, resulting in the separation of LREE and HREE [48,49]. However, the separation degree of LREE and HREE is low, especially in the RL-6 and RL-10 samples (Figure 8). In brief, according to apatite texture and composition of apatite, we confirm that the origin of the RL-6, RL-10, and RL-16 apatite is magmatic.

Time of Nb-Ta Mineralization and Tectonic Environment
Many studies have been conducted on the geochronology of the Renli Nb-Ta deposit. The Zircon U-Pb age obtained by Xiong et al. (2020) from biotite monzogranite was 154 ± 3 Ma, and this represents the crystallization age of biotite monzogranite in the Renli Nb-Ta deposit [14]. The zircon U-Pb age obtained by Xiong et al. (2020) from muscovite monzogranite was 141 ± 2 Ma, and a SHRIMP Zircon U-Pb age obtained by Li et al. (2017) was ca. 137 Ma [13]. Regarding the time of mineralization, a muscovite 40 Ar-39 Ar age range obtained by Li et al. (2017) for the Mufushan complex granite ore-bearing pegmatite was 131-128 Ma, and a coltan U-Pb age obtained by Xiong et al. (2020) was~140 Ma [13,14]. Because of extensive mineralization, many studies confirm that biotite monzogranite and muscovite monzogranite from the Mufushan granite are spatially and genetically related to the Renli Nb-Ta deposit [14]. The new LA-ICP-MS apatite U-Pb ages obtained from the two-mica pegmatite were approximately 130 Ma, which is consistent with the muscovite 40 Ar-39 Ar age reported by Li et al. (2017) [13]. South China was in a lithospheric extension system during the period around 140 Ma [50][51][52]. Therefore, the Renli deposit likely formed in an extensional environment, and the Nb-Ta mineralization of Renli deposit lasted until approximately 130 Ma.
The apatite U-Pb system is characterized by a relatively low closure temperature ranging from 350 to 550 • C, so the system can be easily reset during heating [3]. For the RL-6, RL-10, and RL-16 samples, the LA-ICP-MS U-Pb ages of apatite are almost the same (within the uncertainty range), and the mean value is approximately 130 Ma. The LA-ICP-MS U-Pb age of zircon from the columbite-bearing albite pegmatite is 131 ± 2 Ma, which represents the age of new magmatic-hydrothermal growth and recrystallization in the giant Renli Nb-Ta deposit [19]. The youngest LA-ICP-MS U-Pb age of zircon from columbite-bearing albite pegmatite and the apatite U-Pb age under LA-ICP-MS are similar. Therefore, one possible explanation is that the U-Pb system in apatite crystals is subsequently disturbed throughout the block, which is the result of the inflow of the youngest mantle-related granitic magma and the growth and recrystallization of the new magmatic-hydrothermal solution [3]. This suggests that there was no younger thermal event (above 350 • C) in the Renli area [3].

Oxidation State of Nb-Ta Mineralization
Apatite is an important accessory mineral in various magmatic rocks, and Ca 2+ can be occupied by a number of cations, such as Na + , Fe 2+ , Mn 2+ , Sr 2+ , rare earth elements (REE 2+ / 3+ ), and Y 3+ [4,[53][54][55][56]. The trace elements in apatite depend on the characteristics of the host rock, so the geochemical characteristics of apatite can be used to interpret the mineralogical and geological conditions [57][58][59][60][61][62][63]. Large quantities of elements (Ce, Eu, and Mn) enter apatite, and due to the variable oxidation of these elements, their anomalies in apatite can be used to indicate the redox state of magma [47,64,65].
In this study, the RL-6, RL-10, and RL-16 samples do not have significant Ce anomalies; the Ce anomalies range from 1.02 to 1.22. All the analyzed apatite has strong negative Eu anomalies (Figure 7). Eu anomalies are strongly controlled by feldspar crystallization prior to or simultaneously with apatite crystallization, so it is difficult to explain the oxidation state of the host felsic magma [47,66,67]. The Sr content of apatite has nothing to do with oxidation state [68]. The Sr content of pegmatite is lower than the Sr content of apatite, which also supports the plagioclase crystallization. Previous studies have used Mn anomalies to indicate the redox state of magma, because the radius of Mn 2+ is more similar to that of Ca 2+ ; reduced magmas have a higher content of Mn than oxidized magma [47]. Because the Mn contents of the whole rock are >100 ppm, this effect will be obscured during the fractionation of granitoid magmas [47]. In contrast to Mn, the content of Ga in apatite is independent of the host rock; because the radius of Ga 2+ is more similar to that of Ca 2+ , reduced magmas a have higher content of Ga than oxidized magmas [69]. We found that the Ga content decreases successively from the RL-6 and RL-10 rock to the RL-16 rock, so the degree of oxidation gradually increases. In addition, compared with rocks with a lower oxidation degree, apatite in the rocks with a higher oxidation degree has a lower Y/ΣREE ratio, a higher La/Sm ratio, and a higher Ce/Th ratio [47]. The Y/ΣREE ratio ranges from 0.44 to 0.69 in RL-6 apatite and ranges from 0.05 to 0.46 in RL-16 apatite (Figure 8d). The gradual reduction in the Y/ΣREE ratio in apatite from the Renli pegmatites indicates an increase in the oxidation degree of the Nb-Ta mineralization in the giant Renli deposit. Compare to pegmatite, the apatite has higher content of Y and HREE, suggesting that the apatite is product of water-poor melts [68]. Highly evolved and volatile rich magmas have a non-chondrite Y/Ho ratio [25]. The mostly Y/Ho ratio of RL-6, RL-10, and RL-16 are >34, which indicate that its host pegmatite is formed in a transitional magmatic-hydrothermal system [1]. In brief, the REE behavior of apatite can be used to constrain the oxidation state of Nb-Ta mineralization, and apatite has great potential as an exploration indicator.

Conclusions
(1) The origin of the apatite is magmatic, and the oxidation degree of the Nb-Ta mineralization increases in the giant Renli deposit.

Conflicts of Interest:
The authors declare no conflict of interest.