Rare earth elements are critical constituents for modern technologies, and some of their largest natural resource deposits are related to carbonatite systems. However, the mechanisms leading to rare earth element mineralization and the role of magmatic fluids in carbonatite systems remain poorly understood. Here, we present the first in situ characterization of fluids and their trace-element compositions in natural carbonatite systems by studying secondary quartz-hosted fluid inclusions from Oldoinyo Lengai volcano. By comparing our data to other fluids and melts from various carbonatite systems, we constructed a model for fluid-mediated rare earth element transport and mineralization. We show that carbonatite-related fluids are rich in alkali-carbonate + sulfate + chloride and CO2, but poor in H2O, and they can be significant carriers of rare earth elements (>1600 ppm). We argue that fluid CO2 contents are essential to preclude or slow down the interaction with wall rock during migration and that fluid-mediated rare earth element mineralization occurs when partial pressure of CO2 decreases in the fluid (i.e., during degassing).Rare earth elements (REEs) are critical components for many modern technologies, and their demand is expected to increase in the coming decades, leading to renewed interest in carbonatite systems as economically viable REE sources (Wang et al., 2020; Yaxley et al., 2022). The role of fluids in REE mobility (Anenburg et al., 2020; Louvel et al., 2022) and their relationship to REE mineralization (Tucker et al., 2012; Smith and Henderson, 2000) and fenitization (Elliott et al., 2018) have received special attention, as fluids are essential to ore formation (Audétat and Edmonds, 2020; Bain et al., 2020). However, study of fluids coexisting with carbonatite melt (synmagmatic fluids; Yaxley et al., 2022) or analogous fluids that were physically separated from their source (paramagmatic fluids) is challenging, as their composition can easily change following removal. This is reflected in that most of the available data for carbonatite-related fluids (Costanzo et al., 2006; Samson et al., 1995; Walter et al., 2021; Zhang et al., 2021) were actually obtained from externally derived fluids (postmagmatic), which were REE-poor and strikingly different from fluids observed to be in coexistence with melts in carbonatite systems (Guzmics et al., 2019).To better understand the nature of syn- and paramagmatic fluids and their capacity to transport REEs and generate REE deposits, samples were collected at a currently active carbonatite volcano (Oldoinyo Lengai), reducing the possibility of alteration of these fluids by solute precipitation or exchange due to cooling, reaction with country rocks, or mixing with hydrothermal or meteoric fluids. This allowed the characterization of the original composition of magma-derived fluid phases and the sequence of precipitation of various minerals from this fluid.In this study, we obtained the first in situ characterization of fluids in a natural carbonatite system, assessed their capacity to act as fenitizing agents, and evaluated their potential to transport REEs and form REE mineralization. In addition, we compared them to fluids and melts from other carbonatite systems.The fluid inclusions (Fig. 1) were hosted in quartz-rich xenoliths collected near the summit of Oldoinyo Lengai volcano (Fig. S1 in the Supplemental Material1). The xenoliths are composed of metamorphic relics (subparallel-oriented relict feldspar and quartz) hosting abundant secondary fluid inclusions (i.e., entrapped after crystal formation; Fig. 1A) and an igneous groundmass absent of fluid inclusions (Fig. S2E). No melt inclusions were found.The bulk fluid composition (Tables S1 and S2) was determined to be alkali-carbonate + sulfate + chloride–bearing, H2O-poor, and CO2-rich fluid (hereafter referred as alkali-carbonate fluids), determined by Raman imaging and three-dimensional (3-D) focused ion beam–scanning electron microscopy (FIB-SEM) serial sectioning (Fig. 1). At room temperature, the fluid inclusions present were liquid and vapor CO2 (Figs. 1A and 1B) together with nahcolite (NaHCO3), REE-bearing natrite solid solution [(Na2,K2,Ca)CO3], halite (NaCl), sylvite (KCl) (Figs. 1C and 1D), minor thenardite (Na2SO4), and arcanite (K2SO4). Natrite solid solution (natritess) also contained sulfate, as shown by Raman spectra containing SO4 symmetrical stretching bands (Fig. S3).Upon heating, nahcolite dissociation was accompanied by the appearance of OH bands in the CO2-rich fluid phase between 100 °C and 200 °C (Figs. S3 and S4). No Si-bearing phases were detected in the fluid inclusions (Fig. 1), indicating that no detectable fluid-quartz reaction took place in the pressure-temperature (P-T) range from entrapment to ambient conditions. Synmagmatic fluids coexisting with carbonatite melts showed similar CO2-rich, H2O-poor alkali-carbonate compositions to paramagmatic fluids (Fig. 2), with both having lower concentrations of H2O and NaCl relative to postmagmatic fluids. In situ laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) analyses of the fluid inclusions showed high enrichment in Ba, Sr, Th, U, and light (L) REEs (Fig. 3; Table S3) relative to moderate enrichment in Nb, Ta, Zr, Hf, and Ti (Table S3). LREEs are likely held by natritess, which is analogous to burbankite and accepts Sr, Ba, and LREE substitutions. The host quartz is poor in trace elements (Table S4), enabling their determination in the enclosed fluid inclusions.Raman-combined microthermometry showed a homogeneous fluid phase within the inclusions, observed between 600 °C and 700 °C (Fig. S5). With decreasing temperature, the fluid underwent immiscibility into a CO2-H2O fluid phase and a coexisting alkali-carbonate fluid phase (equivalent to molten natritess; Figs. S5 and S6). The CO2-H2O fluid phase went through further immiscibility under 400 °C, forming a CO2-rich and an H2O-rich fluid phase. The alkali-carbonate fluid was consumed by the crystallization of REE-bearing natritess at ~400 °C, whereas H2O was used for nahcolite formation below 200 °C (Fig. S6).Strictly speaking, melts are fluids. However, to conform with the dominant terminology in the earth science literature, we will refer to dense incompressible carbonatite fluids that crystallize to a space-filling mineral assemblage as “carbonatite melts.” Figure 2 shows that the fluid phase has a higher (Na + K)/Ca ratio and is richer in volatiles (CO2 + H2O) than carbonatite melts, and consequently cannot quench into volume-filling assemblages at room temperature. Immiscibility between carbonatite melt and synmagmatic fluids will cease when volatiles are outgassed from fluid-carbonatite melt systems (Berkesi et al., 2020), showing how CO2 plays a key role in maintaining immiscibility between alkali-carbonate fluids and carbonatite melts. This is supported by experimental results in a CO2-free Na2CO3-H2O system (Yuan et al., 2023), suggesting a continuous evolution of a Na2CO3 melt to aqueous fluid without immiscibility at pressures >3 kbar.An important question is: How can Na2CO3-rich compositions of carbonatite fluids (Fig. 2; Guzmics et al., 2019) or melts (Berkesi et al., 2020) form in natural chemically complex carbonatite systems, which include CO2 not contained in carbonate and other important components (K, Ca, Sr, Ba, S, REEs)? Experiments (Weidendorfer et al., 2017) and data from melt inclusions (Guzmics et al., 2011; Berkesi et al., 2023) indicate that high P-T (>0.5 kbar and >1000 °C) carbonatite melts are CaCO3-rich, and their evolution by crystal fractionation progresses toward Na2CO3 + K2CO3 enrichment. However, that enrichment is limited by the natritess-nyerereite [(Na2,K2,Ca)CO3–(Na,K)2Ca(CO3)2] eutectic at ~550–600 °C (Weidendorfer et al., 2017). Accordingly, greater enrichment in alkalis than that allowed by the natritess-nyerereite eutectic is unlikely to be achieved simply by crystal fractionation. Therefore, we suggest that carbonatite melt-fluid immiscibility is an important and dominant process required to produce alkali-carbonate fluid compositions.Previous fluid inclusion studies in carbonatite systems faced difficulties, mostly emerging from precise in situ measurements being unavailable, including: compositional information limited by visual volume estimation of volatile contents (Bühn et al., 2002), questionable representativeness of compositional data calculations based on decrepitate residues (Williams-Jones and Palmer, 2002; Samson et al., 1995), CO2 contents below detection limits (Costanzo et al., 2006), and unidentified solid phases (Andersen, 1986). Additionally, determination of fluid entrapment timing (i.e., distinguishing assemblages formed during host growth or after host growth) has also been infeasible (Morogan and Lindblom, 1995; Zhang et al., 2021). Although H2O-NaCl fluids have never been observed to coexist with carbonatite melts in natural samples, several works have used the H2O-NaCl system to represent synmagmatic fluids (Costanzo et al., 2006; Samson et al., 1995; Walter et al., 2021; Zhang et al., 2021). This might be due to the ubiquitous nature of H2O-NaCl fluids, commonly leading to postmagmatic fluids overprinting synmagmatic compositions (Yaxley et al., 2022). In contrast, an undoubtedly synmagmatic natural fluid revealed a relatively dry (<4 H2O wt%) alkali-carbonate CO2-bearing composition (Guzmics et al., 2019). The composition of the quartz xenocryst-hosted secondary fluid inclusions studied here is similar to that of synmagmatic fluids (Fig. 2), and the fact that the samples were collected at the summit of Oldoinyo Lengai indicates that the studied fluids are paramagmatic.Models that interpret alkali-carbonate fluids to derive from NaCl-H2O fluids do not explain how major components present in alkali-carbonate fluids (CO2, CO32–, SO42–, and REEs) are absent in most NaCl-H2O fluid inclusions (Figs. 2 and 3; Walter et al., 2021). Indeed, alkali-carbonate fluids are capable of precipitating REE-bearing minerals during their evolution (Bühn et al., 2002), as opposed to NaCl-H2O fluids, which have orders of magnitude lower REE concentrations (Fig. 3).Economically viable REE mineralization occurs due to the enrichment of magmatic REE-bearing minerals within the carbonatite body itself (Zaitsev et al., 1998) as late-stage mineralization overprinting the initial magmatic assemblage (Smith et al., 2015), or hosted in fenites surrounding the carbonatite complex (Broom-Fendley et al., 2021; Liu et al., 2019). Fluids play an important role in REE mineralization (Andersen et al., 2017; Wang et al., 2020); however, there is no agreement on the nature of the mineralizing fluids. Consequently, models explaining REE mobility in fluids have considered several anions, with REE-fluorides-chlorides, REE-carbonates, and REE-sulfates being the main complexes studied (Smith et al., 2015). Experimental data indicate that the presence of Na2CO3 (Anenburg et al., 2020) or hydroxyl-carbonate complexes (Louvel et al., 2022) increases REE solubility. The high REE contents in the studied alkali-carbonate fluids (Fig. 3; Fig. S7) are in agreement with experiments, supporting the interpretation that alkali-carbonate and hydroxyl-carbonate content increases REE solubility in fluids in natural carbonatite systems. Considering their high alkali content (Fig. 2; Table S1) and capacity to mobilize REEs (Fig. 3; Table S3), syn- and paramagmatic fluids can play an important role in fenitization (alkali metasomatism) and REE transport/mineralization (Fig. 4).The fluid inclusions (Fig. 2; Tables S1 and S2) showed no signs of reactions with their quartz host, as no Si-bearing step-daughter phases were observed (Fig. 1). The presence of Raman bands of alkali-carbonates, up to 600 °C, within quartz-hosted fluid inclusions (Fig. S5) corroborates the conclusion that no or undetectable amounts of fluid-host reaction happened. A decarbonation reaction between sodium carbonate and silica would be expected to occur above 400 °C (Grynberg, 2012):The absence of Si-bearing step-daughter phases can be explained by the high partial CO2 pressure inside the fluid inclusions. The CO2 in the fluid phase precludes (or significantly slows down) reactions between its Na2CO3 component and siliceous country rocks (Eq. 1). As the fluid inclusions retain the CO2 reaction product in Equation 1, the reaction rapidly becomes self-limiting within the inclusions. In carbonatite systems, decompression leads to exsolution of synmagmatic fluids from melts (Fig. 4A). As fluids migrate to lower P and T (Fig. 4B), they undergo immiscibility and phase separation, forming an alkali-carbonate fluid and a CO2-rich, H2O-poor fluid (Fig. 4C). Continuous separation between these phases (Figs. S4 and S5) and the physical escape of CO2 + H2O fluid from the system (Fig. 4D) reduce the CO2 fugacity, permitting the reaction between the alkali-carbonate fluid and silicate minerals in the country rock (Eq. 1) to freely proceed. This reaction decreases the alkali-carbonate concentration in the carbonatite-derived fluid, which in turn results in reduced REE solubilities and precipitation of REE-bearing minerals. Thus, the processes of fenitization and precipitation of REE-bearing minerals are closely linked (Fig. 4E). Alternatively, if an alkali-carbonate fluid undergoes phase separation and thus CO2 loss within a carbonatite rock (Figs. 4F–4H), the resultant alkali-carbonate fluid will react with calcite (Weidendorfer et al., 2017) and may precipitate a REE-bearing carbonate assemblage within the carbonatite rock itself (Fig. 4I).Syn- and paramagmatic fluids in carbonatite systems appear to have alkali-carbonate + sulfate + chloride–bearing, H2O-poor, and CO2-rich compositions, with high total REE contents (>1600 ppm). These fluids can be distinguished from carbonatite melts by their lower density and inability to quench into volume-filling assemblages, enrichment in volatile components (CO2 + H2O), and higher (Na + K)/Ca ratios. Fluid-mediated REE mineralization and fenitization are linked to fluid-phase CO2 contents, as high CO2 fugacity preserves REE solubility in the fluid phase by precluding reactions between carbonate components of the fluid and siliceous country rocks. Given the importance of CO2 in setting apart carbonatite-derived fluids from other fluid types found in Earth’s crust, we suggest that explanations regarding REE mobility in carbonatite systems should consider the role of alkali-carbonate in syn- or paramagmatic fluids instead of H2O-NaCl fluids. Although similar fluids have been proposed to develop without immiscibility, we suggest that immiscibility is the predominant process that produces alkali-carbonate fluid compositions in CO2-rich systems.We thank the editor M. Norman and the reviewers, M. Steele-MacInnis, H.-R. Fan, and an anonymous reviewer, for constructive comments, and M. Anenburg for discussions that benefited earlier drafts of this work. This research received financial support from the National Research, Development, and Innovation Office of Hungary (NKFIH) (FK-132418 and K-142855), awarded to M. Berkesi and T. Guzmics, respectively.

中文翻译：

---

与碳酸岩系统流体相关的稀土元素迁移和矿化

稀土元素是现代技术的关键组成部分，其一些最大的自然资源矿藏与碳酸岩系统有关。然而，导致稀土元素矿化的机制以及岩浆流体在碳酸岩系统中的作用仍然知之甚少。在这里，我们通过研究来自 Oldoinyo Lengai 火山的次生石英流体包裹体，首次对天然碳酸岩系统中的流体及其微量元素组成进行了原位表征。通过将我们的数据与来自各种碳酸岩系统的其他流体和熔体进行比较，我们构建了流体介导的稀土元素传输和矿化的模型。我们发现，与碳酸岩相关的流体富含碱金属碳酸盐+硫酸盐+氯化物和CO2，但缺乏H2O，它们可以是稀土元素（>1600 ppm）的重要载体。我们认为，流体 CO2 含量对于阻止或减缓运移过程中与围岩的相互作用至关重要，并且当流体中 CO2 分压降低时（即脱气过程中），就会发生流体介导的稀土元素矿化。稀土元素）是许多现代技术的关键组成部分，预计未来几十年其需求将增加，从而导致人们重新对碳酸岩系统作为经济上可行的稀土元素来源产生兴趣（Wang 等人，2020 年；Yaxley 等人，2022 年）。流体在稀土元素流动性中的作用（Anenburg 等人，2020；Louvel 等人，2022）及其与稀土元素矿化（Tucker 等人，2012；Smith 和 Henderson，2000）和铁硝化作用（Elliott 等人，2000）的关系。 2018）受到了特别关注，因为流体对于成矿至关重要（Audétat 和 Edmonds，2020；Bain 等，2020）。然而，对与碳酸盐岩熔体共存的流体（同岩浆流体；Yaxley 等人，2022）或与其来源物理分离的类似流体（顺岩浆流体）的研究具有挑战性，因为它们的成分在去除后很容易发生变化。这反映在大多数碳酸岩相关流体的可用数据（Costanzo 等，2006；Samson 等，1995；Walter 等，2021；Zhang 等，2021）实际上是从外部获取的流体（岩浆后），其稀土元素贫乏，与碳酸盐岩系统中观察到的与熔体共存的流体显着不同（Guzmics等人，2019）。为了更好地了解同岩浆流体和顺岩浆流体的性质及其输送能力稀土元素并生成稀土元素沉积物，样品是在当前活跃的碳酸岩火山 (Oldoinyo Lengai) 采集的，减少了由于冷却、与围岩反应或与热液或大气流体混合而导致的溶质沉淀或交换而改变这些流体的可能性。这使得能够表征岩浆衍生流体相的原始成分以及各种矿物质从该流体中沉淀的顺序。在这项研究中，我们首次对天然碳酸岩系统中的流体进行了原位表征，评估了它们作为铁硝化剂的能力，并评估了它们传输稀土元素和形成稀土矿化的潜力。此外，我们将它们与其他碳酸岩系统的流体和熔体进行了比较。流体包裹体（图1）存在于Oldoinyo Lengai火山山顶附近收集的富含石英的捕虏体中（补充材料1中的图S1）。捕虏体由变质遗迹（近平行取向的残长石和石英）组成，含有丰富的次生流体包裹体（即晶体形成后被捕获；图1A）和不含流体包裹体的火成岩基质（图S2E）。未发现熔体包裹体。整体流体成分（表S1和S2）经测定为含碱碳酸盐+硫酸盐+氯化物、贫H2O、富CO2流体（以下简称碱碳酸盐流体），经测定通过拉曼成像和三维 (3-D) 聚焦离子束扫描电子显微镜 (FIB-SEM) 连续切片（图 1）。在室温下，存在的流体包裹体是液体和蒸气 CO2（图 1A 和 1B）以及苏打石 (NaHCO3)、含稀土元素的钠硝酸盐固溶体 [(Na2,K2,Ca)CO3]、石盐 (NaCl)、钾盐(KCl)（图 1C 和 1D）、少量芒硝 (Na2SO4) 和奥金石 (K2SO4)。钠盐固溶体（Natritess）还含有硫酸盐，拉曼光谱中含有SO4对称伸缩带（图S3）。加热时，在100℃之间，富含CO2的流体相中，苏打石解离伴随着OH带的出现。和 200 °C（图 S3 和 S4）。在流体包裹体中没有检测到含硅相（图1），这表明在从捕获到环境条件的压力-温度（PT）范围内没有发生可检测到的流体-石英反应。与碳酸盐岩熔体共存的同岩浆流体表现出与副岩浆流体类似的富含CO2、缺乏H2O的碱金属碳酸盐成分（图2），两者都具有相对于岩浆后流体较低的H2O和NaCl浓度。对流体包裹体的原位激光烧蚀-电感耦合等离子体-质谱 (LA-ICP-MS) 分析表明，Ba、Sr、Th、U 和轻 (L) 稀土元素高度富集（图 3；表 S3）中度富集 Nb、Ta、Zr、Hf 和 Ti（表 S3）。轻稀土元素很可能由钠石所持有，它类似于伯班石并接受 Sr、Ba 和轻稀土元素替代。基质石英的微量元素含量较低（表 S4），因此能够在封闭的流体包裹体中进行测定。拉曼组合显微测温显示包裹体内存在均匀的流体相，在 600 °C 至 700 °C 之间观察到（图 S5）。随着温度降低，流体不混溶成 CO2-H2O 流体相和共存的碱金属碳酸盐流体相（相当于熔融硝酸盐；图 S5 和 S6）。 CO2-H2O液相在400℃下进一步不混溶，形成富含CO2和富含H2O的流体相。碱金属碳酸盐流体在约 400 °C 时被含稀土硝酸盐的结晶消耗，而 H2O 在 200 °C 以下用于苏苏石的形成（图 S6）。严格来说，熔体是流体。然而，为了与地球科学文献中的主流术语保持一致，我们将结晶成充满空间的矿物组合的致密不可压缩碳酸岩流体称为“碳酸岩熔体”。图 2 显示，与碳酸岩熔体相比，流体相具有更高的 (Na + K)/Ca 比，并且富含挥发物 (CO2 + H2O)，因此无法在室温下淬火成体积填充组合体。当挥发物从流体-碳酸岩熔体系统中脱气时，碳酸岩熔体与同岩浆流体之间的不混溶性将会停止（Berkesi et al., 2020），这表明二氧​​化碳如何在维持碱金属碳酸盐流体与碳酸岩熔体之间的不混溶性方面发挥关键作用。这得到了无 CO2 Na2CO3-H2O 系统中的实验结果的支持（Yuan 等人，2023），表明 Na2CO3 熔体在压力 >3 kbar 下不断演化为水性流体，而没有不混溶性。一个重要的问题是：如何才能实现这一点？天然化学复杂的碳酸岩体系中形成富含Na2CO3的碳酸岩流体成分（图2；Guzmics等，2019）或熔体（Berkesi等，2020），其中包括碳酸盐中不含的CO2和其他重要成分（K 、Ca、Sr、Ba、S、REE）？实验（Weidendorfer 等，2017）和熔体包裹体数据（Guzmics 等，2011；Berkesi 等，2023）表明高 PT（>0.5 kbar 和 >1000 °C）碳酸岩熔体富含 CaCO3，它们通过晶体分馏而演化为 Na2CO3 + K2CO3 富集。然而，这种富集受到约 550–600 °C 的钠盐-尼雷石 [(Na2,K2,Ca)CO3–(Na,K)2Ca(CO3)2] 共晶的限制 (Weidendorfer et al., 2017)。因此，简单地通过晶体分级不可能实现比钠盐-尼雷石共晶所允许的更大的碱富集。因此，我们认为碳酸岩熔体-流体不混溶性是产生碱金属碳酸盐流体组合物所需的重要且占主导地位的过程。以前的碳酸岩系统流体包裹体研究面临困难，主要是由于无法获得精确的原位测量，包括：成分信息有限通过挥发物含量的视觉体积估计（Bühn 等人，2002 年），基于易碎残留物的成分数据计算的代表性存​​疑（Williams-Jones 和 Palmer，2002 年；Samson 等人，1995 年），CO2 含量低于检测限（Costanzo）等，2006）和未鉴定的固相（Andersen，1986）。此外，确定流体截留时间（即区分宿主生长期间或宿主生长后形成的组合）也是不可行的（Morogan 和 Lindblom，1995；Zhang 等，2021）。虽然在天然样品中从未观察到 H2O-NaCl 流体与碳酸岩熔体共存，但一些工作已经使用 H2O-NaCl 系统来代表同岩浆流体（Costanzo 等，2006；Samson 等，1995；Walter 等，2006）。 ，2021；Zhang 等人，2021）。这可能是由于 H2O-NaCl 流体普遍存在，通常导致岩浆后流体覆盖同岩浆成分（Yaxley 等人，2022）。相比之下，毫无疑问的岩浆天然流体显示出相对干燥（<4 H2O wt%）的含碱碳酸盐 CO2 的成分（Guzmics 等，2019）。这里研究的石英异晶次生流体包裹体的成分与同岩浆流体的成分相似（图2），并且样本是在Oldoinyo Lengai山顶采集的事实表明，所研究的流体是顺岩浆流体。将碱金属碳酸盐流体解释为源自 NaCl-H2O 流体并不能解释碱金属碳酸盐流体中存在的主要成分（CO2、CO32-、SO42- 和 REE）为何在大多数 NaCl-H2O 流体包裹体中不存在（图 2 和3；沃尔特等人，2021）。事实上，碱金属碳酸盐流体能够在其演化过程中沉淀出含有 REE 的矿物（Bühn 等，2002），而 NaCl-H2O 流体则相反，后者的 REE 浓度要低几个数量级（图 3）。经济上可行稀土矿化的发生是由于岩浆含稀土矿物在碳酸岩体内富集（Zaitsev 等，1998），因为后期矿化覆盖了最初的岩浆组合（Smith 等，2015），或者托管在周围的有限岩中。碳酸岩复合体（Broom-Fendley 等人，2021；Liu 等人，2019）。流体在稀土矿化中发挥着重要作用（Andersen等，2017；Wang等，2020）；然而，对于矿化流体的性质尚未达成一致。因此，解释流体中 REE 迁移率的模型考虑了多种阴离子，其中 REE-氟化物-氯化物、REE-碳酸盐和 REE-硫酸盐是研究的主要复合物（Smith 等，2015）。实验数据表明，Na2CO3（Anenburg 等人，2020）或羟基碳酸盐复合物（Louvel 等人，2022）的存在会增加 REE 的溶解度。研究的碱金属碳酸盐流体中的高稀土元素含量（图3；图S7）与实验结果一致，支持碱金属碳酸盐和羟基碳酸盐含量增加了天然碳酸岩体系流体中稀土元素溶解度的解释。考虑到它们的高碱含量（图 2；表 S1）和动员 REE 的能力（图 3；表 S3），同岩浆液和准岩浆液可以在铁铁矿化（碱交代作用）和 REE 运输/矿化中发挥重要作用（图 3；表 S3）。 4).流体包裹体（图2；表S1和S2）没有显示出与其石英基质发生反应的迹象，因为没有观察到含硅继子相（图1）。碱金属碳酸盐拉曼带的存在，高达 600 °C，在石英流体包裹体内（图S5）证实了没有发生或无法检测到流体主体反应发生的结论。碳酸钠和二氧化硅之间的脱碳酸反应预计将在 400 °C 以上发生（Grynberg，2012）：含硅继子相的缺失可以通过流体包裹体内的高二氧化碳分压来解释。流体相中的 CO2 阻止（或显着减缓）其 Na2CO3 成分与硅质围岩之间的反应（方程式 1）。由于流体包裹体保留了方程 1 中的 CO2 反应产物，因此反应在包裹体内迅速变得自限性。在碳酸岩系统中，减压导致同岩浆流体从熔体中溶出（图4A）。当流体迁移到较低的 P 和 T 时（图 4B），它们会发生不混溶和相分离，形成碱金属碳酸盐流体和富含 CO2、贫水的流体（图 4C）。这些相之间的连续分离（图 S4 和 S5）以及 CO2 + H2O 流体从系统中物理逸出（图 4D）降低了 CO2 逸度，从而允许碱金属碳酸盐流体与围岩中的硅酸盐矿物之间发生反应（等式1）自由进行。该反应降低了碳酸岩衍生流体中碱金属碳酸盐的浓度，进而导致稀土元素溶解度降低和含稀土矿物的沉淀。因此，铁辉石化和含稀土矿物的沉淀过程是紧密相连的（图4E）。或者，如果碱金属碳酸盐流体发生相分离，从而在碳酸岩中损失二氧化碳（图 4F-4H），则所得碱金属碳酸盐流体将与方解石发生反应（Weidendorfer 等，2017），并可能沉淀出 REE碳酸岩岩石本身内含有碳酸盐组合（图 4I）。碳酸岩系统中的同岩浆流体和准岩浆流体似乎具有含碱金属碳酸盐 + 硫酸盐 + 氯化物、贫 H2O 和富 CO2 的成分，且稀土元素总量较高含量（>1600 ppm）。这些流体与碳酸岩熔体的区别在于其密度较低、无法骤冷成体积填充组合、挥发性成分富集 (CO2 + H2O) 以及较高的 (Na + K)/Ca 比率。流体介导的稀土元素矿化和铁辉石化与流体相二氧化碳含量有关，因为高二氧化碳逸度通过阻止流体碳酸盐成分与硅质围岩之间的反应来保持流体相中稀土元素的溶解度。鉴于 CO2 在区分碳酸岩衍生流体与地壳中发现的其他流体类型方面的重要性，我们建议有关碳酸岩系统中稀土元素迁移性的解释应考虑碱金属碳酸盐在同岩浆流体或顺岩浆流体中的作用，而不是 H2O-NaCl液体。尽管已经提出在不混溶的情况下开发类似的流体，但我们认为不混溶是在富含二氧化碳的系统中产生碱金属碳酸盐流体组合物的主要过程。我们感谢编辑 M. Norman 和审稿人 M. Steele-MacInnis、H.-R。 Fan 和一位匿名审稿人提出了建设性意见，M. Anenburg 提出了有益于本工作早期草稿的讨论。这项研究得到了匈牙利国家研究、发展和创新办公室 (NKFIH) 的财政支持（FK-132418 和 K-142855），分别授予 M. Berkesi 和 T. Guzmics。