Qihai Shu

Globally magmatic hydrothermal Sn deposits are typically associated with partial melting of reduced, Sn-enriched metasedimentary protoliths. The newly recognized Southern Great Xing’an Range Sn metallogenic belt in northeastern China, however, is characterized with Sn-mineralized intrusions commonly being A- or I-type granites derived dominantly from juvenile crust. How Sn is enriched in these magmas remains unclear. Haliheiba is a typical Sn deposit in this region that is associated with a biotite granite stock containing mafic microgranular enclaves (MMEs). In this study, we aimed to clarify the role of magmatic processes in the Sn deposit formation. Laser ablation ICP-MS U–Pb dating on magmatic zircons from the biotite granite and MMEs yields identical ages of ~142 Ma, consistent with the regional Early Cretaceous Sn mineralization event. The texture and composition zoning patterns are abundant in zircon, plagioclase, and amphibole, and can be attributed to the mafic and felsic magma mixing within a shallow felsic magma chamber. Amphibole and biotite thermobarometers indicate an estimated depth of ~13 km of the magma chamber, while the deep mafic magma resided at ~26 km with a temperature of ~871 °C. The calculated oxygen fugacities from zircon geochemistry for the biotite granite and the MMEs are FMQ + 0.30 and FMQ + 0.11, respectively, indicative of a relatively reduced condition similar to many other Sn-related intrusions. The amphibole Fe3+/(Fe2+ + Fe3+) molar ratios of the biotite granite and the MMEs range from 0.19 to 0.28, also supporting a relatively low magmatic ƒO2 nature. The reduced magmas could prevent Sn from oxidizing to Sn4+ to partition into Ti- or Fe-bearing minerals, so that Sn remains incompatible during magma crystallization. The high-temperature (~871 °C) mafic magma injection could lead to decomposition of biotite and amphibole from both the previously crystallized Sn-bearing phases and the country rocks, releasing a significant amount of Sn into the mixed melt, which has elevated the Sn contents from ~49 to ~102 ppm in the melt, as revealed by the amphibole Sn concentration and Monte Carlo modeling. The current study highlights the importance of the injected mafic magma, the reduced nature of the parental magmas, and the interaction with Sn-rich country rocks in facilitating the Sn mineralization at Haliheiba, and suggests that these factors may also be critical in the formation of the other Sn deposits in the Great Xing’an Range.

峪耳崖是冀东地区最具代表性的金矿床之一,前人对其成矿物质来源、成矿时代和流体包裹体方面均有研究,但矿床成因仍存在争议。本次研究采用激光剥蚀电感耦合等离子体质谱(LA-ICP-MS)对矿床载金黄铁矿开展原位微区测试分析,并基于全球4544组黄铁矿地球化学数据,利用随机森林和支持向量机等机器学习算法,构建岩浆热液型与造山型金矿床黄铁矿地球化学数据的分类模型,系统探讨峪耳崖金矿床的成因。LA-ICP-MS分析结果表明,峪耳崖黄铁矿以富集Co、Au、Ag、Cu、As、Pb、Te元素为特征,微量元素组成与典型岩浆热液型金矿床一致。传统Co-Ni及Ni-Co-As等元素判别图解存在多解性,难以有效判别黄铁矿成因类型。本研究构建的机器学习模型经十折交叉验证分类准确率达93%,显著优于传统判别手段。结合矿区已有的矿物学、年代学证据以及模型预测结果,证实峪耳崖金矿床属岩浆热液成因。本次研究表明机器学习方法在整合高维黄铁矿地球化学数据、判别矿床成因及重建成矿过程方面的有效性,为复杂成矿系统的解析提供了新思路。研究进一步深化了对峪耳崖金矿床成因的认识,也为区域岩浆热液型金矿的勘查提供了重要支撑。

The identification of the genetic type of a mineral deposit, based on the compositional characteristics of specific minerals, has long been a focus of interest for economic geologists and mining companies. Traditional binary plots, due to the limitation of their dimensions, fail to encompass the whole element information, potentially introducing bias to the discrimination results. This is particularly the case in classifying Zn-Pb deposits. The current study employs four widely used machine learning algorithms (random forest, extreme gradient boosting, support vector machine, and multi-layer perceptron) to train 4908 sets of element data for sphalerite compiled from five distinct Zn-Pb deposit types (VMS, SEDEX, MVT, skarn, and epithermal). The data are then visualized and interpreted through principal component analysis and t-distributed stochastic neighbor embedding, which indicate that reducing sphalerite element data to a two-dimensional projection leads to the loss of significant feature information, hindering the ability to effectively distinguish the genesis of the deposit. The machine learning results show that all four models have macro F1-scores above 0.95 on the test set, demonstrating robustness and excellent generalization ability, which reflects the reliability of using sphalerite geochemistry to distinguish Zn-Pb mineralization types. The SHAP value analysis highlights the key role of Mn, Fe, Ge, Cd, and Co concentrations in facilitating the differentiation of deposit types through machine learning algorithms. Our models show an accuracy rate of 83% in predicting the combined results on an external, independent data set. The models have also been applied to classify three Zn-Pb deposits of unknown types, and the results are consistent with geological observations. The models’ parameters have been further exported and programmed into an Excel macro program and a user-friendly software application, which can be accessed via https://sdeakii.github.io/machine-learning.

内蒙古浩布高铅锌矿床位于大兴安岭南段黄岗-甘珠尔庙成矿带, 该带是我国北方重要的铅锌多金属成矿带之一。目前, 浩布高矿床闪锌矿微量元素赋存机制尚不清晰, 矿床成因类型亦存在一定的争议。本研究采用激光剥蚀电感耦合等离子体质谱仪(LA-ICP-MS)原位微量元素分析手段, 结合机器学习方法, 对浩布高闪锌矿微量元素组成特征、赋存机制以及矿床成因类型进行了探讨。LA-ICP-MS分析结果表明, 浩布高矿床中闪锌矿以富集Fe、Mn、Co、Cu、Se、Ag、Cd、In、Sn等元素, 贫Ni、Ga、Ge、As、Mo、Sb、Au、Tl、Pb、Bi等元素为特征。其中, Fe、Mn、In等元素主要以类质同象的方式替代Zn赋存于闪锌矿中, Cu、Ag、Sn等元素含量变化范围较大, 可能部分以微粒包裹体的形式存在。In与Cu含量具有一定的正相关, 推测In在闪锌矿中富集机理可能为Cu⁺+In³⁺↔2Zn²⁺。Cd与Fe含量呈一定的正相关性, 而与Zn含量呈负相关性, 这可能暗示在闪锌矿中Cd主要以类质同象的方式置换Zn元素而非Fe元素。通过穷举闪锌矿微量元素图解发现, 即使得分最高的Co/Ag-Mn图解依然存在较大范围的重叠区域, 因此不能简单地用闪锌矿微量元素二元图解来判别矿床类型。通过测试四种机器学习中经典的不同核函数的支持向量机算法, 最终得出准确率为91.5%的高斯内核支持向量机闪锌矿成矿类型分类器, 可以用于多种矿床类型闪锌矿的研究。浩布高矿床中闪锌矿微量元素组成明显不同于密西西比河谷型(MVT)、沉积喷流型(SEDEX)、火山块状硫化物型(VMS)和浅成低温热液型铅锌矿床, 而与矽卡岩型接近。综合已有矿物学、年代学以及本研究获得的闪锌矿地球化学证据表明, 浩布高应是一个典型的矽卡岩型铅锌矿床。

There are more than 80 porphyry (or skarn) Mo deposits in northeastern China with Jurassic or Cretaceous ages. These are thought to have formed mainly in a continental arc setting related to the subduction of the Paleo-Pacific oceanic plate in the Jurassic and subsequent slab rollback in the early Cretaceous. The Jurassic Daheishan porphyry Mo deposit is one of the largest Mo deposits in NE China, which contains 1·09 Mt Mo with an average Mo grade of 0·07 %. To better understand the factors that could have controlled Mo mineralization at Daheishan, and potentially in other similar porphyry Mo deposits in NE China, the geochemical and isotopic compositions of the ore-related granite porphyry and biotite granodiorite, and the magmatic accessory minerals apatite, titanite and zircon from the Daheishan intrusions, were investigated so as to evaluate the potential roles that magma oxidation states, water contents, sulfur and metal concentrations could have played in the formation of the deposit. Magmatic apatite and titanite from the causative intrusions show similar ε_Nd(t) values from –1·1 to 1·4, corresponding to T_DM2 ages ranging from 1040 to 840 Ma, which could be accounted for by a mixing model through the interaction of mantle-derived basaltic melts with the Precambrian lower crust. The Ce and Eu anomalies of the magmatic accessory minerals have been used as proxies for magma redox state, and the results suggest that the ore-forming magmas are highly oxidized, with an estimated ΔFMQ range of +1·8 to +4·1 (+2·7 on average). This is also consistent with the high whole-rock Fe₂O₃/FeO ratios (1·3–26·4). The Daheishan intrusions display negligible Eu anomalies (Eu/Eu* = 0·7–1·1) and have relatively high Sr/Y ratios (40–94) with adakitic signatures; they also have relatively high Sr/Y ratios in apatite and titanite. These suggest that the fractionation of amphibole rather than plagioclase is dominant during the crystallization of the ore-related magmas, which further indicates a high magmatic water content (e.g. >5 wt%). The magmatic sulfur concentrations were calculated using available partitioning models for apatite from granitoids, and the results (9–125 ppm) are indistinguishable from those for other mineralized, subeconomic and barren intrusions. Furthermore, Monte Carlo modelling has been conducted to simulate the magmatic processes associated with the formation of the Daheishan Mo deposit, and the result reveals that a magma volume of ∼280 km³ with ∼10 ppm Mo was required to form the Mo ores containing 1·09 Mt Mo in Daheishan. The present study suggests that a relatively large volume of parental magmas with high oxygen fugacities and high water contents is essential for the generation of a giant porphyry Mo deposit such as Daheishan, whereas a specific magma composition (e.g. with unusually high Mo and/or S concentrations) might be less critical.

Northeastern China is one of the richest Mo-mineralized regions in China, with 95 Mesozoic Mo-bearing deposits and a total metal resource of >12.2 Mt Mo. The reasons behind the large-scale Mesozoic Mo mineralization in NE China remain unclear, and whether or not there was any regional-scale pre-enrichment of the source region (e.g., a Mo-rich lower crust) is still a matter of debate. In this study, whole-rock Nd and zircon Hf isotope compositions of the intrusions related to the Mo mineralization have been compiled. The results show that the isotopic compositions are highly heterogeneous among these deposits with different ages and Mo tonnages, indicating that Mo-related magmas could be derived from either ancient lower crust or relatively juvenile lower crust or via mixing of mantle-derived magmas with varying proportions of crustal melts. This suggests that different magmas, independently from their sources, can produce Mo mineralization in NE China, and, therefore, that there was probably not an unusually Mo rich basement underlying NE China. In addition, Monte Carlo simulations have been carried out to explore the magmatic processes potentially associated with the formation of magmatic-hydrothermal (porphyry- or skarn-type) Mo deposits. The results reveal that a variably large magma volume (e.g., >150 km³) was required to provide enough Mo metal to form the large (>0.1 Mt Mo) deposits. In NE China, the Jurassic Mo deposits are much more abundant and larger than the Triassic and Cretaceous deposits, which could be attributed to the specific Jurassic tectonic regime. In the Jurassic, the subduction of the Paleo-Pacific oceanic plate led to regional compression, which favors accumulation of larger amounts of magmas at depth, ultimately resulting in larger Mo deposit formation. In this study, we highlight the importance of magma volume, rather than magma source, in the formation of the Mesozoic Mo deposits in NE China.