Ore Genesis of the Baishitouwa Quartz–Wolframite Vein-Type Deposit in the Southern Great Xing’an Range W Belt, NE China: Constraints from Wolframite In-Situ Geochronology and Geochemistry Analyses

: The Baishitouwa deposit is a medium-scale quartz–wolframite vein-type deposit in the southern Great Xing’an Range tungsten (W) belt. The W mineralization occurs mainly as veins and dissemina-tion within the mica schist of the Mesoproterozoic Baiyunebo Group. The formation of the deposit can be divided into four stages. The wolframite yielded a lower intercept 206 Pb/ 238 U age of 221.0 ± 3.4 Ma (1 σ , MSWD = 2.0), which records a late Triassic W mineralization event in the Baishitouwa deposit. In combination with previous geochronological data, we suggest that NE China may have an enormous potential for Triassic W mineralization and more attention should be given to the Triassic ore prospecting in the region. This work highlights that the chemical composition of wolframite is controlled by both the crystallochemical parameters and the composition of the primary ore-forming ﬂuid. Trace-element compositions suggest that wolframite (I) was controlled by the substitution mechanism of 4 A (Fe, Mn) 2+ + 8 B W 6+ + B (cid:3) ↔ 3 A M 3+ + A N 4+ + 7 B (Nb, Ta) 5+ + 2 B N 4+ , whereas wolframite (II) was controlled by the substitution mechanism of A (Fe, Mn) 2+ + A (cid:3) + 2 B W 6+ ↔ 2 A M 3+ + 2 B N 4+ . Wolframite (I) contains higher concentrations of Nb, Ta, Sc, and heavy rare earth elements (HREEs), and lower Mn/(Mn + Fe) ratios than wolframite (II). Both wolframite (I) and (II) have similar trace elements and left-dipped REE N patterns, and analogical Nb/Ta ratios. They have similar Y/Ho ratios to Mesozoic highly fractionated W-mineralized granitoids in NE China. These data indicate that the W mineralization at Baishitouwa is genetically related to an underlying highly fractionated granite, and the compositional variation of ﬂuids is likely driven by crystallization of wolframite during the processes of ﬂuid evolution. A change of the ore-forming ﬂuids from an oxidized to a relatively reduced state during the evolution occurred from stage 1 to 2.


Introduction
Most tungsten mineralization is spatially associated with highly evolved granites worldwide and form a large variety of deposit types, including quartz-vein, skarn, greisen, and porphyry-type [1][2][3][4][5][6]. Among these different mineralization styles, quartz-wolframite vein-type deposits provide approximately 44% of the world's known economic tungsten  [37]). (b) Schematic tectonic map of NE China (modified from [38]). (c) Distribution of major tungsten deposits in NE China (modified from [8]). Abbreviations are as follows: NCGB = the northern and central Great Xing'an Range W belt; SGB = the southern Great Xing'an Range W belt; LXZB = the Lesser Xing'an-Zhangguangcai Range W belt.
Accurate isotopic dating of hydrothermal minerals is a vital tool for constraining the timing of hydrothermal activity relative to intrusive magmatism, which is critical for constructing genetic models for hydrothermal deposits [39]. Zircons from spatially related granitic plutons are often used for U-Pb dating to constrain the formation age of W deposits [20,21,40,41]. However, these results are generally built upon a hypothesis that the dated granitic plutons are genetically related to W mineralization, which needs to be proved. Recent development in in situ U-Pb dating of wolframite provides an efficient approach to directly confine the timing of W mineralization and has been widely applied to quartz-wolframite vein-type deposits worldwide [40,[42][43][44][45][46]. Electron probe microanalysis (EPMA) and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) in situ trace-element analysis has been used in many studies of mineral formation environments and mineralization processes [47][48][49][50][51][52][53][54]. Wolframite is the dominant tungsten mineral in quartz-wolframite vein-type deposits, which can incorporate significant amounts of Nb, Ta, rare earth elements (REEs), and other trace elements [55,56]. Numerous studies demonstrated that the trace elements and their variations in wolframite can record the physicochemical conditions and compositions of ore-forming fluids, thereby reflecting their source and evolution [1,[56][57][58][59]. Although some studies have been conducted on the substitution mechanism of trace elements in wolframite, these studies commonly take different generations of wolframite as a whole without considering the possible differences between them [1,42]. The ambiguity has impeded the understanding of the ore-forming process of multi-stage quartz-wolframite vein-type deposits and can be potentially resolved by coupling detailed mineralogical observation and geochemical interpretation as shown in this study.
The Baishitouwa deposit, situated in the SGB, is a typical medium-scale quartzwolframite vein-type deposit (Figures 1c and 2). Previous studies mainly focused on its geological characteristics [60] and petrogenesis [61]. The deposit represents an important Triassic hydrothermal W mineralization event and provides an excellent opportunity to investigate the mineralization mechanism. In this contribution, we identified two generations of wolframite based on extensive field observations and mineralogical thin-section observations. High-quality in situ U-Pb age of hydrothermal wolframite for the Baishitouwa deposit is presented, precisely pinning a Late Triassic W mineralization event in this region. Furthermore, in situ geochemical data of the two generations of wolframite are acquired to determine the substitution mechanism controlling the trace-element compositions of wolframite. The two generations of wolframite geochemistry also provide important information on the source and evolution of the ore-forming fluids. This paper improved our understanding of the regional W mineralization in NE China, as well as brought new insight into the origin and processes of tungsten mineralization of quartz-wolframite vein-type deposits.
The Narenwula-Baishitouwa W ore district, located in the southwestern part of the SGB (Figure 1c), hosts a series of quartz-wolframite vein-type deposits (i.e., Narenwula, Baishitouwa, Shazigou, and Tantoushan) and Sansheng porphyry W-Mo deposit ( Figure 2). The region mainly contains units of Proterozoic-Early Paleozoic, Lower Permian, and Jurassic strata [62]. The Proterozoic-Early Paleozoic strata consists of leptite, granulite, quartzite, slate, phyllite, schist, and marble. The widely exposed Lower Permian and Jurassic strata comprise intermediate-acid volcanic rocks, volcaniclastic rocks, and sedimentary clastic rocks [62]. Four periods of granitic magmatism were identified, including Late Permian, Triassic, Early-Middle Jurassic, and Late Jurassic ( Figure 2). The widespread Late Permian granodiorites and quartz diorites intrude into the Proterozoic-Early Palaeozoic strata. A Triassic granitic pluton is locally exposed in the southwestern part of this district. Early-Middle Jurassic fine-grained and minor biotite granites are mainly exposed in the western part of the district, along with minor occurrences in the eastern part. Late Jurassic monzogranites and granite porphyries are mainly exposed in the central and western parts of this district [62].

Ore Deposit Geology
The Baishitouwa W deposit is situated 16 km northwest of Taipusi Banner at 115 • 07 E, 41 • 56 E (Figure 2). The deposit has estimated reserves of 39,800 t WO 3 with a grade of 0.489% [81], and occurs within the mica schist of the Mesoproterozoic Baiyunebo Group ( Figure 3). Faults in the mining district include NEE-, NW-, and NE-trending faults, with the NE-trending faults being the main ore-controlling structures (Figure 3a,b). An early Cretaceous granite pluton is exposed in the eastern part of the mining district, intruding the upper Jurassic Zhangjiakou Formation [61].
Seven ore bodies (Nos. 1, 2, 3, 4, 5, 6, and 7) are recognized for this deposit, of which No. 2 is the largest. The Nos. 1 and 2 ore bodies are exposed on the surface, and other ore bodies are concealed and controlled by drill holes (Figure 3b The ores are mainly veined, massive, or disseminated ( Figure 4). The main ore minerals are wolframite, pyrite, chalcopyrite, sphalerite, and galena (Figures 4 and 5). The gangue minerals are quartz, mica, sericite, chlorite, epidote, fluorite, and calcite ( Figure 6). Wolframite is the main economic mineral in the Baishitouwa deposit and can be subdivided into two generations. Wolframite (I) commonly occurs as tabular, lath-shaped, columnar, and ir-regular crystals in fine-grained quartz-wolframite veins (Figure 4a,b and Figure 5a,b). Wolframite (II) is disseminated widely in interstices of the quartz grains ( Figure 4c). A small amount of wolframite (II) is distributed in the quartz veins in leaf-like and scale-like forms, and coexists with coarse-grained prismatic quartz (Figure 4d). The wolframite (II) is predominantly euhedral, tabular, and fine-to medium-grained, and coexists with pyrite and chalcopyrite (Figure 5c,d). The pyrite is euhedral to subhedral crystals, with sizes from 0.05 to 3 mm. Fine-and medium-grained pyrite is commonly replaced by sphalerite, galena, and chalcopyrite along the fractures and crystal margins (Figure 5c,d,h,i). The sphalerite is medium-to coarse-grained with an ir-regular shape, and often coexists with galena, pyrite, and chalcopyrite (Figure 5f,h,i). Exsolution textures of sphalerite-chalcopyrite solid solution are common, suggesting that they formed simultaneously (Figure 5f). The chalcopyrite is fine-to medium-grained with ir-regular shape and coexists with sphalerite, galena, and pyrite. The chalcopyrite can be subdivided into two types: the early generation of chalcopyrite is fine-grained and often forms an exsolution texture with sphalerite (Figure 5f), whereas the late generation of chalcopyrite is medium-grained and commonly replaces sphalerite, galena, and pyrite along the fractures and crystal margins (Figure 5f,i). The galena is fine-to medium-grained anhedral crystals and coexists with sphalerite, chalcopyrite, and pyrite (Figure 5g,h).   Different types of alteration occurred in the Baishitouwa deposit, including silicification, sericitization, chloritization, epidotization, and carbonation ( Figure 6). Silicification is observed as veins or as disseminations of quartz in ores or altered mica schist. Sericitization often occurs together with silicification, and is composed of sericite (Figure 6c-f). Chloritization and epidotization mainly consist of chlorite and epidote, respectively (Figure 6d-f). Carbonation is mainly composed of calcite, which represents the low-temperature alteration of the post-mineralization stage (Figure 4f).
Based on morphology, structure, mineralogy, and modes of occurrence, the hydrothermal ore-forming process of the Baishitouwa deposit can be divided into four stages: stage 1 of fine-grained quartz-wolframite (I), stage 2 of coarse-grained quartzwolframite (II)-pyrite-chalcopyrite, stage 3 of quartz-polymetallic sulfides, and stage 4 of quartz-carbonate.
Stage 1 is the dominant W mineralization stage and is characterized by an assemblage of wolframite (I), quartz, and sericite. Abundant tabular crystals of wolframite (I) occur in fine-grained quartz-wolframite veins (Figure 4a,b). Wolframite (I) is medium-to coarse-grained euhedral to subhedral crystals (Figure 5a,b). Sulfides are absent in this stage ( Figure 4b). Stage 2 is characterized by an assemblage of wolframite (II), pyrite, chalcopyrite, quartz, and sericite. The quartz coexisting with wolframite (II) is coarse-grained pectinate or prismatic (Figure 4c,d). Wolframite (II) is disseminated widely in interstices of the quartz grains (Figure 4c), and small amounts of wolframite (II) are distributed in quartz veins in leaf-and scale-like forms (Figure 4d). Minor pyrite and chalcopyrite are present in this stage and intergrown with wolframite (II) (Figures 4d and 5c,d). Stage 3 is characterized by an assemblage of pyrite, chalcopyrite, galena, sphalerite, quartz, sericite, chlorite, and epidote. Abundant sulfides were formed in this stage. Pyrite is the most abundant sulfide, and is featured by medium-to coarse-grained euhedral to subhedral crystals (Figure 5e,h,i). However, the pyrite is commonly replaced by sphalerite, galena, and chalcopyrite along fractures and crystal margins (Figure 5h

Samples
The samples used for in situ U-Pb dating and geochemical analysis during this study were collected from Nos. 1

EPMA
EPMA of the wolframite samples was conducted at the Wuhan Sample Solution Analytical Technology Co., Ltd., Wuhan, China. Major-and trace-elemental composition of the wolframite were measured using a JEOL JXA 8230 EPMA equipped with five wavelengths dispersive X-ray (WDX) spectrometers. The operating conditions used a 15-kV acceleration voltage, a 50 nA beam current, and a 1 µm beam diameter. Standards used for data correction were wolframite for W, hematite for Fe, manganite for Mn, and Nb metal for Nb. Spectral lines, peak times, and off-peak background times used for the WDS analyses were as follows: W (Lα, 10, 5), Fe (Kα, 10, 5), Mn (Kα, 10, 5), and Nb (Lα, 10, 5). The analytical uncertainties were 0.4 wt% for WO 3 , 0.2 wt% for FeO, 0.1 wt% for MnO, and 0.05 wt% for Nb 2 O 5 . The detection limits for all of the analyzed elements were below 0.01 wt%.

LA-SF-ICP-MS U-Pb Dating
The samples were ablated using a GeoLasPro 193 nm ArF excimer laser (CompexPro 102F, Coherent) coupled with a Thermo Scientific Element XR sector field ICP-MS at the State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences in Guiyang, China. A laser frequency of 6 Hz, an energy density of 5 J/cm 2 , and a spot size of 32 µm was used for the analyses. Primary standard materials, YGX [44] and WT [82], were analyzed twice for every 15 analyses of the unknown samples. A detailed description of the analytical conditions and methods can be found in [43,83]. The ICP-MS data were processed offline using the ICPMSDataCal software for calibration, background correction, and floating of the integration signal [84]. No downhole corrections were made for only the first~25 s of ablation data (excluding the initial~2 s) used in the process. Isoplot 4.15 was used to calculate the U-Pb ages and generate Concordia diagrams. Common Pb corrections were employed using a Tera-Wasserburg Concordia or a Tera-Wasserburg Concordia anchored through common Pb [85]. The lower intercept ages were used as the timing of mineral precipitation of wolframite [85][86][87]. Data errors for the isotopic ratios in the following samples are 1 σ.

LA-ICP-MS Analyses
LA-ICP-MS analyses were conducted at the Wuhan Sample Solution Analytical Technology Co., Ltd., Wuhan, China. Laser sampling was performed using a GeoLas2005 laser ablation system and an Agilent 7700e ICP-MS. The instrumental settings and operational procedures used herein were similar to those described in [56,88]. The laser ablation system was equipped with a signal-smoother to compensate for the low-frequency laser pulse [89]. The time-resolved signal included two intervals: the background (20-30 s) and the analytical signal (50 s). No internal standards were used, but multiple external standards [88] were applied to calculate the elemental concentrations. United States Geological Survey (USGS) reference glasses (e.g., BCR-2G, BIR-1G, and BHVO-2G) were used as the calibration standards [90]. The elemental concentrations of the USGS glasses can be found in the GeoReM database (http://georem.mpch-mainz.gwdg.de/ accessed on 1 March 2022). The ICPMSDataCal software [84,88] was used for offline data processing, including analytical and background signal determinations, sensitivity drift calibrations, and elemental concentration calculations.

Wolframite U-Pb Age
The U-Pb isotope data from the wolframite sample BST-17 are listed in Table 1. The uncorrected U-Pb data are plotted on a U-Pb Tera-Wasserburg diagram in Figure 7. Nineteen spot analyses of the wolframite sample had total Pb, Th, and U concentrations varying from 0.29 to 3.89 ppm, from 0.01 to 0.86 ppm, and from 1.65 to 127.46 ppm, respectively. A lower intercept 206 Pb/ 238 U age of 221.0 ± 3.4 Ma (1 σ, MSWD = 2.0) was obtained from the Tera-Wasserburg Concordia diagram ( Figure 7).

In Situ Major and Trace-Element Compositions of Wolframite
The chemical compositions of wolframite from the Baishitouwa deposit determined using EPMA are listed in Table 2. The trace elements in the wolframite measured using LA-ICP-MS are listed in Table 3.
The wolframite had the following major elemental composition: wolframite (I) has higher FeO contents   Wolframite from the Baishitouwa deposit has been characterized by variable traceelemental compositions ( Figure 9). In wolframite (I) samples, concentrations above 10 ppm were observed for six elements (  ). V, Cr, Zn, U, and Yb concentrations were varying from 1 to 10 ppm, whereas the other elemental concentrations were less than 1 ppm. The Hg concentration was mainly below the detection limit ( Table 3). The concentrations of Nb, Ta, Sc, Zr, Hf, Pb, and Eu in wolframite (I) are higher than those in wolframite (II), whereas the concentrations of Ti, Sn, Cr, and Th in wolframite (I) are lower than those in wolframite (II) (Figure 9).     The total REE concentrations of wolframite (I) and wolframite (II) range from 9.75 to 11.95 ppm and 2.59 to 4.83 ppm, respectively. Both wolframite (I) and wolframite (II) are characterized by high heavy REEs (HREEs) concentrations (9.69-11.89 ppm and 2.50-4.70 ppm, respectively), and low light REEs (LREEs) concentrations (0.05-0.11 ppm and 0.03-0.18 ppm, respectively). All REE patterns from wolframite normalized to the upper continental crust (UCC from [91]) show a common preferential enrichment in HREEs relative to LREEs, which are below the detection limit in some samples (Figure 10a). The trace-element patterns normalized to the UCC are used to identify geochemical characteristics of the wolframite crystals [1]. Wolframite grains from both stages show similar trace-element signatures, with relative enrichments in Nb, Sn, U, and Sc, and with relative depletions in Y, Li, Ta, Zn, Th, Ti, Zr, Hf, Mo, Pb, V, and Co ( Figure 10b).

Timing of W Mineralization
Constraining the timing and duration of mineralization events is critical for understanding ore deposit formation, from both academic and economic viewpoints [39]. At the Baishitouwa deposit, according to the intrusive contact relationship between the granite pluton and the upper Jurassic Zhangjiakou Formation (Figure 3a), Dai [61] considered that the granite pluton formed in the Early Cretaceous. In addition, integrated with geochemical data of the granite pluton and W-related granitoids in South China, Dai [61] concluded that the granite pluton was genetically associated with W mineralization in the Baishitouwa deposit. However, this conclusion contradicts our study.
In our study, a lower intercept 206 Pb/ 238 U age of 221.0 ± 3.4 Ma (1 σ, MSWD = 2.0) (Figure 7) was obtained from in situ U-Pb dating of wolframite from the Baishitouwa deposit, which records a Late Triassic W mineralization event in the district. By collecting the published geochronological data of W-related granitoids and associated W deposits in NE China, Xie et al. [5] determined three episodes of W-related magmatism and mineralization events, namely Triassic (240-250 Ma), Early-Middle Jurassic (170-200 Ma), and Late Jurassic-Early Cretaceous . The formation of W mineralization in NE China was initiated in the Early Triassic and peaked in the Late Jurassic-Early Cretaceous [5]. For a long time, researchers have made a comprehensive chronological study of Jurassic and Early Cretaceous granitic magmatism and related W mineralization in this region, whereas the studies of Triassic W mineralization have been in their infancy. Previous geochronologic work only reported two Triassic W deposits, namely the Shazigou W-Mo deposit and Yangjingou W deposit. Peng et al. [31] reported a mean Re-Os age of 243.8 ± 1.6 Ma from five molybdenite samples from the Shazigou W-Mo deposit. Zhao [32] obtained a zircon U-Pb age of 249.4 ± 2.7 Ma and a muscovite 40 Ar-39 Ar age of 230.8 ± 1.2 Ma for the Yangjingou W deposit. It is noteworthy that the Shazigou, Yangjingou, and Baishitouwa deposits are all located along the SXCF, suggesting the formation of these W deposits may be related to the closure of PAO. Xie et al. [5] systematically summarized the tectonic setting of Mesozoic W-mineralized granitoids and associated W deposits, and proposed that Triassic W-mineralized granitoids and associated W deposits formed in a syn-and/or postcollisional setting related to the closure of PAO. According to our wolframite in situ geochronology results, together with previous isotope data, we preliminarily speculate that NE China underwent not only the well-known Jurassic and Early Cretaceous, but also a prevalent Triassic W mineralization event during the Mesozoic period. Consequently, NE China has countless potential for Triassic W mineralization and more attention should be given to the Triassic ore prospecting in the region. The spatial distribution of Triassic W-mineralized granitic intrusions and related W deposits are bound to the closure of PAO, and thus we propose that the Triassic metal exploration should be focused on the two sides of the SXCF.

Factors Controlling Trace-Element Compositions in Wolframite
Wolframite is a typical ABO 4 compound, with two octahedral sites (A and B) [92]. The wolframite structure is very versatile and occurs with multiple isovalent substitutions [1,42,55,[93][94][95][96][97][98]. The incorporation of other elements in the wolframite structure could significantly change the composition of wolframite, which has been the subject of many studies [1,42,56,58]. In this study, wolframite from different stages at Baishitouwa provides an excellent opportunity to understand the factors controlling trace-element compositions in wolframite.
Several correlations between elements in wolframite are discernible in binary diagrams. It is clearly shown in Figure 11a that Nb and Ta concentrations correlate in the wolframite (I), which indicates that Nb and Ta are incorporated together during wolframite (I) crystallization. In contrast, in wolframite (II), there are no significant positive correlations between Nb and Ta contents (Figure 11a). Furthermore, Nb concentrations in wolframite (I) from this study exhibit positive correlations with Sc, V, Y, ∑REE, and HREE (Figure 11b-f).
It is clearly shown in Figure 12a-d that Sc contents in wolframite (I) and wolframite (II) show positive correlations with V, Y, ∑REE, and HREE. Furthermore, similar positive correlations are also observed among V, Y, ∑REE, and HREE (Figure 12e-i). This evidence corroborates that these trivalent cations are incorporated together during wolframite crystallization. The Sc contents in wolframite (I) and (II) correlate positively with Ti, Sn, Zr, and Hf contents (Figure 13a-d). Conformably, these tetravalent cations (Ti 4+ , Zn 4+ , Zr 4+ , Hf 4+ ) also display a positive relationship with V and Y contents in two generations of wolframite samples (Figure 13e-l). These phenomena suggest that these trivalent and tetravalent cations are incorporated through a similar mechanism during the crystallization of wolframite (I) and (II), as expressed by the coupled substitution reaction of Equation (3)    Positive correlations between Ti and other tetravalent cations (Sn 4+ , Zr 4+ , and Hf 4+ ) have been illustrated in Figure 14a-c. Furthermore, the Sn contents also display a positive correlation with Zr and Hf contents in two generations of wolframite (Figure 14d,e). Con-formably, the positive relationship between Zr and Hf contents is also observed in both wolframite (I) and (II) (Figure 14f). The positive correlation among these tetravalent cations reveals that they are incorporated together into the wolframite lattice, as shown by coupled substitution of Equation (4) [1,94,97]. Nb contents in the wolframite (I) show positive correlations with Ti, Sn, Zr, and Hf (Figure 15a-d). Those tetravalent cations also correlate positively with Ta contents in the wolframite (I) (Figure 15e-h). Conformably, it is clearly shown in Figure 15i that (Ti + Sn + Zr + Hf) contents correlate positively with the (Nb + Ta) contents in the wolframite (I). Thus, we infer that these elements are incorporated through a similar mechanism during the crystallization of wolframite (I), as expressed by the coupled substitution reaction of Equation (5) [97,98]. Based on these findings, it can be concluded that the trace-element compositions of the wolframite (I) and (II) from hydrothermal quartz veins in the Baishitouwa deposit are controlled by crystochemical parameters (including ionic radius and charge valence). However, the coupled substitution reactions controlling the trace-element compositions of two generations of wolframite are distinct. Five coupled substitution mechanisms can explain the chemical variations in the wolframite (I), whereas the chemical compositions of the wolframite (II) can be elucidated by two coupled substitution mechanisms. Considering that the substitution mechanisms occurred simultaneously during wolframite formation, and thus the substitution mechanisms controlling trace-element compositions of wolframite (I) and wolframite (II) can be unified as shown in Equations (6) and (7), respectively: The substitution mechanisms controlling trace-element compositions of wolframite in this study are different from those in wolframite from the quartz-wolframite vein-type deposits from the European metallogenic belt and Nanling metallogenic belt [1,42,102], indicating the substitution mechanisms in hydrothermal wolframite from different W deposits are not identical. We preliminarily suggest substitution mechanisms controlling trace-element compositions of wolframite from different stages are different, which should be paid more attention to in future research.
Previous studies revealed that crystallochemical effects and composition of the primary mineralizing fluids are two key controls of wolframite chemistry, with the charge and radius controlled (CHARAC) behavior defined for common igneous rocks, and wolframite by Y/Ho and Zr/Hf ratios [1,58,103]. In the Y/Ho versus Zr/Hf diagram (Figure 16a), almost all two generations of wolframite samples are plotted outside the CHARAC field, indicating that the ionic radius and the charge valence were not the only parameters controlling the mobility of the isovalent trace elements into the fluids precipitating the two generations of wolframite. Hence, the concentrations of elements incorporated during the crystallization of hydrothermal wolframite also reflect the specific chemical composition of the mineralizing fluids [1,58]. Wolframite (I) and (II) from the Baishitouwa deposit distinguish mostly by enrichments in Sc, Ti, Zr, Nb, Sn, and Mg with concentrations ranking from 10 to 10 4 ppm, V, Zn, Y, U, and HREEs with a range of 1 to 10 ppm, and by depletions in Li, Be, B, Cr, Co, Ni, Cu, Pb, Mo, Ag, Cd, Ga, Rb, Sr, Sb, Cs, Ba, Hf, Th, LREEs, etc., in the range of 0.01 to 1 ppm (Figure 9 and Table 3). The high Sc, Ti, Zr, Nb, Sn, and Mg contents in wolframite reveal that these elements are easily incorporated into its crystalline structure, whereas the low contents in other elements (e.g., Cr, Th, Cd, Sr, Pb, Rb, and LREEs) indicate that they are either not incorporated easily into wolframite or present at very low concentrations in the mineralizing fluids owing to low solubility in aqueous solutions. The diagram of ionic radius versus electric charge (Figure 16b) illustrates that the low concentrations of many elements in wolframite can be controlled by crystallochemical parameters (including ionic radius and charge valence). For instance, Rb + , Eu 2+ , Sr 2+ , Pb 2+ , Cd 2+ , Bi 3+ , LREE 3+ , and Th 4+ have too large ionic radii (>0.92 Å) making them incompatible to enter the A-site. In contrast, Cr 6+ has a too low ionic radius (<0.56 Å) to enter the B-site into the wolframite structure, whereas it is possible for Cr 3+ (0.61 Å). Symmetrically, charge-compensating elements such as Cu + , Li + , Zn 2+ , Co 2+ , and Sb 3+ can theoretically enter relatively easily on the A-site, and Ni 3+ , Ga 3+ , Cr 3+ , Co 3+ , and Ge 4+ can theoretically enter relatively easily on the B-site (Figure 16b). However, these elements' contents are significantly low in wolframite from the Baishitouwa deposit, suggesting a control by its concentrations in the mineralizing fluids, and consequently the source of the fluids.  [1]). Dash lines represent the lower and upper limits of 15% relatively to the lattice site radius corresponding to total substitution (Goldschmidt's rule). Ionic radii data are from [92].
In summary, crystallochemical effects and composition of the primary ore-forming fluids are two key controls of wolframite chemistry in the Baishitouwa deposit, and the substitution mechanisms controlling the trace-element compositions of wolframite from two stage quartz-veins are different.

Source of Ore-Forming Fluids and Materials
Previous studies have demonstrated that trace-elemental (such as Nb, Ta, Sc, and REEs) concentrations in wolframite provide clues to the source of the ore-forming fluids and materials [1,[56][57][58]94,104]. The elements Nb, Ta, and REEs are generally enriched in the high-temperature stage of magmatic crystallization differentiation [105,106]. Subsequently, these elements are incorporated into the tungsten-bearing minerals due to the similarity in the characteristics of electronegativity, ion radius, and ion potential [1,42,57,93,94,97]. Zhang [107] concluded that the concentrations of Nb and Ta in wolframite decline with increasing distance from the parent granite. Gan and Chen [99] performed a comparative element analysis of wolframite in quartz vein-type and greisen-type W deposits and found that they have similar concentrations of WO 3 , MnO, and FeO, but that quartz vein-type deposits have lower Nb, Ta, and Sc concentrations and a higher Nb/Ta ratio compared with greisen-type deposits. Xiong et al. [56] compared the concentrations of Nb, Ta, and Sc in two generations of wolframite and found that stage 1 wolframite has higher concentrations of Nb, Ta, and Sc than stage 2 wolframite. Harlaux et al. [1] proposed that progressive increase in Nb and Ta in wolframite with depth indicated that ore-forming fluids and metals were from hidden granite. Zhang et al. [57] found that Nb and Ta concentrations in wolframite from Piaotang increased from stage 2 to stage 1, and thereby inferred that stage 1 wolframite precipitated from high temperature magma-derived fluids. To decipher the origin of the ore-forming fluids and metals in the Baishitouwa deposit, we collected the published geochemical data of Early Mesozoic W-mineralized granitoids in NE China. It is noteworthy that these W-mineralized granitoids have similar Ta (mainly 0.25-3.06 ppm, average 1.17 ppm) and HREEs (mainly 4.75-28.45, average 15.14) concentrations to the early generation of wolframite (Ta = 0.58-1.20 ppm, HREEs = 9.69-11.89 ppm) in the Baishitouwa. In contrast, these W-mineralized granitoids have lower Nb concentrations (2.76-30.70 ppm) than early generation wolframite (Nb = 94.35-165 ppm) in the Baishitouwa, which could explain the tendency for Nb incorporation in wolframite lattice during magmatic-hydrothermal evolution [1,105,106]. Ballouard et al. [108] pointed out that Nb is slightly more mobile than Ta, suggesting that magmatic-hydrothermal processes account for the decrease in the Nb/Ta ratio in peraluminous granites that may be related to Ta, Cs, Nb, Be, Sn, and W mineralization. Xie et al. [5] concluded that Mesozoic W-mineralized granitoids in NE China have lower Nb/Ta ratios than contemporary W-barren granitoids. According to Zhang [109], the concentration of Sc can reach 10-1000 ppm in granitic magma and its residual solution during the period of late crystallization, which suggests that the ore-forming fluid may have originated from deep magma. The Sc concentrations of wolframite (I) and wolframite (II) range from 46.32-57.43 ppm and 21.78-34.24 ppm, respectively, which suggest that the ore-forming fluids of the Baishitouwa deposit were derived from granitic magma. In the LREE-Middle REE (MREE)-HREE diagram [110], wolframite from quartz-wolframite vein-type deposits in South China and French Massif Central all plot near the corner of HREE (Figure 17), and the ore-forming fluids and materials of these deposits have been corroborated to originate from coeval granitoids [1,[56][57][58][59]. The characteristics of the LREE-MREE-HREE in wolframite from the Baishitouwa deposit are similar to those of granitic intrusion-related quartz-wolframite vein-type deposits worldwide, thereby suggesting that the initial ore-forming fluids and materials were of magmatic origin ( Figure 17).  [110]). The data of wolframite from quartz-wolframite vein-type deposits are listed in Supplementary Material Table S1.
Due to the similar ionic radii and valences of 'geochemical twin' elements, these elemental ratios (e.g., Nb/Ta and Y/Ho) tend to remain fairly stable in a given magmatichydrothermal system, allowing their use as a fluid source indicator [103,108,111,112]. If ore-forming fluids originated from magmas only, the wolframite crystallized from such fluids should exhibit a narrow range of Y/Ho ratios [103,111]. Zhang et al. [113] argued that the relatively high Nb (more than 4.0 ppm), Ta (more than 0.01 ppm) and consistent Nb/Ta ratios in scheelite from the Xuefeng Uplift Belt are indicative of a magmatic source. Cao et al. [112] considered that the relatively invariant Y/Ho ratios in both generations of scheelite from the Helukou deposit indicate that they were precipitated from a single source fluid. Cao et al. [112] also noted that Y/Ho ratios of scheelite from the Helukou deposit are consistent with previously published Y/Ho ratios for the Middle Jurassic Guposhan granites, thereby indicating a genetic link between W mineralization and Middle Jurassic Guposhan granites. In the present study, wolframite (I) and (II) from the Baishitouwa deposit have consistent Nb/Ta ratios of 121.76-182.58 and 76.91-158. 27, respectively. Figure 18 shows that most wolframite from the South China and French Massif Central display a relatively narrow range of Y/Ho ratios . The wolframite from the Baishitouwa deposit also shows a narrow range of Y/Ho ratios (6.62-18.97), which is similar to the Y/Ho ratios of granitic intrusion-related quartz-wolframite vein-type deposits worldwide  and the Mesozoic highly fractionated W-mineralized granitoids (5.07-27.97), implying that the W mineralization in the Baishitouwa deposit is genetically related to an underlying highly fractionated granite.  Generally, W mineralization is associated with highly fractionated granites, as corroborated by amounts of cases worldwide [3,5,6,[114][115][116][117][118][119][120][121][122][123]. As previously mentioned, the granite pluton exposed in the eastern part of the district intruded into the Upper Jurassic Zhangjiakou Formation, indicating they may form during the Early Cretaceous [61], whereas a lower intercept 206 Pb/ 238 U age of 221.0 ± 3.4 Ma (1σ, MSWD = 2.0) (Figure 7) obtained from in situ U-Pb dating of wolframite from the Baishitouwa deposit records a Late Triassic W mineralization event in this district. The inconsistent age data indicate that there is no genetic relationship between granite and W mineralization, and the granite records a late postmineralization magmatic event. A huge number of studies have demonstrated that concealed granite is widely distributed in quartz-wolframite vein-type deposits, which provides the source of heat and materials for W mineralization [1,6,40,58,120,[124][125][126]. These W deposits generally do not display granitic intrusion on the surface, or even within the depth controlled by drilling. The ore-forming concealed granite plutons are considered to have better preservation conditions for hydrothermal W deposits than their exposed counterparts, since the deposits are not affected by remarkable surficial weathering or erosion [124].
In summary, we preliminarily conclude that the ore-forming fluids and materials of the Baishitouwa deposit were derived from an underlying highly fractionated granite.

Fluid Evolution Indicated by Compositions of Wolframite
Variations of trace-element compositions in wolframite have been used to decipher the evolutionary history and physicochemical conditions of the ore-forming fluids in many studies [56][57][58]127]. Brugger et al. [127] proposed that the precipitation of tungsten minerals could effectively alter the compositions of fluids. Zhang et al. [57] found that the early wolframite would lower REE, Nb, and Ta in the mineralizing fluids, leading to depletion of these elements in the later ones. Zhang et al. [57] also noted that although both the two generations of wolframite from the Xihuashan and Piaotang deposit have different REE contents and Eu anomalies, they display similar left-dipped REE N patterns, implying that compositional variation of fluids is likely driven by crystallization of wolframite during the processes of fluid evolution. In this study, wolframite (I) contains higher Nb (94.35-165 ppm), Ta (0.58-1.20 ppm), and REEs (9.75-11.95 ppm) concentrations than wolframite (II) (Nb = 7.46-16.86 ppm, Ta = 0.07-0.18 ppm, and REEs = 2.59-4.83 ppm) (Figure 9), and they are characterized by similar left-dipped REE N and trace-element patterns ( Figure 10). We infer that the crystallization of the early generation of wolframite with high Nb, Ta, and REE would lower these elements in the mineralizing fluids. Therefore, a late generation of wolframite precipitating from evolved fluids would contain relatively lower concentrations of these elements.
Nb, Ta, and Sc concentrations in wolframite may be associated with Fe and Mn concentrations and the crystallization temperature, whereas the pH, Eh, and composition of the ore-forming fluid are dependent on the Nb, Ta, and Sc concentrations [128]. Specifically, a low pH and high Eh provide favorable conditions for the enrichment of Nb and Ta in wolframite, whereas a low-pH and low-Eh ore-forming fluid containing Fand/or PO 4 3complexes provides ideal conditions for the enrichment of Sc [56,58,128]. The LA-ICP-MS analysis of wolframite from the Baishitouwa deposit shows that wolframite (I) contains more Nb, Ta, and Sc than does wolframite (II), which suggests that the stage 1 fluid has low pH and high Eh, offering favorable conditions for the enrichment of Nb and Ta. For Sc, we infer that the ore-forming fluid might have originated from or flowed through material abundant in Sc, giving the fluid a high initial Sc concentration, resulting in the abundance of Sc in wolframite (I) despite the high Eh environment.
Eu is a redox-sensitive element and has two ionic valences: Eu 3+ and Eu 2+ [92]. Compared with Eu 2+ (1.17 Å), the ionic radius of Eu 3+ (0.95 Å) is much closer to those of Fe 2+ (0.78 Å) and Mn 2+ (0.83 Å) [92] so that the partition coefficient of Eu 3+ between wolframite and fluids is much higher than that of Eu 2+ . In this study, wolframite (I) has higher Eu concentrations than wolframite (II) (Figure 19a). This finding implies that Eu in the stage 1 fluid occurs mainly as Eu 3+ in a relatively oxidized environment, whereas Eu 2+ was primarily concentrated in the relatively reduced stage 2 fluids. Cr can enter the wolframite lattice as Cr 3+ (0.61 Å), whereas Cr 6+ has a too low ionic radius (<0.56 Å) to enter the B-site into the wolframite structure [1,92]. Therefore, wolframite precipitated from oxidizing fluids tends to contain low Cr concentrations. In addition, Pb concentrations in wolframite can also be a tracer of the redox conditions of the mineralizing fluids. Under oxidizing conditions, Pb 4+ readily enters the wolframite lattice by the substitution mechanism of Equation (5), leading to Pb enrichment [1,92]. In contrast, under reducing conditions, Pb 2+ combines with S 2to form galena. Hence, the redox conditions must be changed during the process of galena precipitation. The variation of oxygen fugacity does not affect the precipitation of wolframite, but galena tends to precipitate under a relatively reducing environment. Therefore, the variation of lead content in wolframite shows changes in physical and chemical conditions of ore-forming fluids in Baishitouwa. In this study, wolframite (I) has higher Cr (0.44-20.12 ppm) and lower Pb (0.01-0.06 ppm) concentrations than wolframite (I) (Cr = 0.10-0.42 ppm, Pb = 0.06-0.45 ppm) (Figure 19b,c). Mineralogical evidence suggests that Pb mainly occurs as galena during the quartz-polymetallic sulfides stage, and that galena had not precipitated during stages 1 and 2. This phenomenon indicates that the ore-forming fluid changed from the oxidation state to the relative reduction state during the evolution from stage 1 to 2, increasing the proportion of Pb 2+ in the oreforming fluids. These Pb 2+ ions eventually precipitated as galena in the quartz-polymetallic sulfides stage. Furthermore, the mineral assemblages can also provide evidence for the change in the metallogenic environment. Stage 1 is the main W mineralization stage and characterized by abundant tabular crystals of wolframite, whereas sulfides are absent in this stage (Figure 4a,b). In contrast, stage 2 is featured with an assemblage of wolframite, pyrite, and chalcopyrite (Figure 4d). The distinct mineral assemblages in stages 1 and 2 implies different redox conditions. In summary, the ore-forming fluids in Baishitouwa changed from an oxidized to a relatively reduced state during the evolution from stage 1 to stage 2.

Conclusions
(1) In situ U-Pb dating of wolframite from the Baishitouwa deposit by LA-SF-ICP-MS yielded a lower intercept 206 Pb/ 238 U age of 221.0 ± 3.4 Ma, which records a late Triassic W mineralization event in the Baishitouwa deposit.
(2) Two generations of wolframite are identified, which have distinctly major-and trace-element characteristics. Wolframite (I) contains higher Nb, Ta, Sc, Zr, Hf, Pb, and Eu concentrations, lower Ti, Sn, Cr, and Th concentrations, and lower Mn/(Mn + Fe) ratios than wolframite (II). Both wolframite (I) and (II) have similar trace elements and LREE-depleted patterns.
(3) Crystallochemical effects and the composition of primary ore-forming fluids are two key controls of wolframite chemistry in the Baishitouwa deposit. The substitution mechanisms controlling the compositions of two generations of wolframite are distinct. Trace-element compositions suggest that wolframite (I) was controlled by the substitution mechanism of 4 A (Fe, Mn) 2+ + 8 B W 6+ + B ↔ 3 A M 3+ + A N 4+ + 7 B (Nb, Ta) 5+ + 2 B N 4+ , whereas wolframite (II) was controlled by the substitution mechanism of A (Fe, Mn) 2+ + A + 2 B W 6+ ↔ 2 A M 3+ + 2 B N 4+ .
(4) In situ LA-ICP-MS trace-element compositions of wolframite indicate that the ore-forming fluids and materials of the Baishitouwa deposit were mainly derived from an underlying highly fractionated granite.
(5) According to the mineralogical observation and various Nb, Ta, Sc, Eu, Cr, and Pb contents in the two generations of wolframite, we proposed that the ore-forming fluids of the Baishitouwa deposit changed from an oxidized to a relatively reduced state during the evolution from stage 1 to stage 2.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/min12050515/s1, Table S1: Chemical composition of wolframite in quartz-wolframite vein-type deposits from South China and French Massif Central; Table S2: Major and trace-element compositions of Mesozoic W-mineralized granitoids in the Xing-Meng Orogenic Belt.
Funding: This research was funded by the National Natural Science Foundation of China, grant number 91962104.