The Radioactive Rare Metal Mineralization in the World-Class Sn-Nb-Ta-U-Th-REE-Deposit Madeira (Pitinga, Amazonas State, Brazil): With Special Reference to the Complex Alteration of Pyrochlore-Group Minerals

Abstract

This study focuses on the relationship between U and pyrochlore in the world-class Sn-Nb-Ta (U, Th, REE, Li) Madeira deposit within the Pitinga mining district of northern Brazil. The primary U mineralization is of intrusive-type and early magmatic origin, hosted in the peralkaline albite-enriched granite facies of the A-type Madeira granite (~1820 Ma). U-Pb-LREE-enriched pyrochlore is the only primary U ore and is widely and homogeneously dispersed in two albite-enriched granite subfacies: the albite-enriched granite core (AGC) and the albite-enriched granite border (AGB). In both zones, the pyrochlore crystals underwent strong hydrothermal alteration by F-rich, low-temperature aqueous fluids. During this hypogene alteration process, cations such as LREE, Nb, and F were selectively released, while others like Fe and Si were introduced. This led to the successive formation of various secondary pyrochlore varieties and a relative enrichment of U (up to 13.73 wt.% UO₂). The alteration of pyrochlore eventually resulted in the breakdown of its structure, leading to the formation of U-bearing columbite pseudomorphs and the precipitation of U-rich silicates (up to 34.35 wt.% UO₂), galena, and LREE-rich fluorides within pyrochlore vugs. In contrast to the homogeneous distribution of the primary ore mineralization, the secondary pyrochlore mineralization shows striking zonation, being most intense in the AGB and AGC proximal to a massive cryolite deposit. The U mineralization in the Madeira deposit exhibits grades of 328 ppm UO₂, comparable to the main deposits of this type, with significant reserves of up to 52 kt U. However, it is different from those deposits in four key aspects: homogeneous dispersion of mineralization; pyrochlore as the exclusive primary ore mineral; U and Th mineralizations formed at different stages; and intense hydrothermal alteration. These characteristics are attributed to the special conditions imposed by the fluorine-rich nature of the peralkaline magma.

1. Introduction

Intrusive type U deposits are commonly associated with carbonatites and granites, with uraninite, uranothorianite, and uranothorite being the dominant species. These deposits are generally low-grade-large-tonnage deposits with a grade of 20 to 500 ppm U and substantial resources exceeding 100 kt [1]. Pyrochlore supergroup minerals are found in several uranium-bearing ores that are currently being processed, including the Rössing Deposit in Namibia [2], which is currently the world’s 5th largest producer of uranium, with 246,500 t U at a grade of 300 ppm [1]. They are targeting lower-grade deposits due to the significant challenges posed by gangue minerals in these deposits [3]. Only through the simultaneous extraction of metals associated with U can the mining of such deposits be economically viable [4]. Moreover, in recent years, there has been increased interest in gaining a better understanding of the structure and chemical factors that influence U leaching from minerals [5,6].

Uranium dissolution in silicate melts is influenced by the degree of depolymerization of the magma, which is controlled by the melt composition [7]. In alkaline melts, high contents of K, Ca, and Na promote depolymerization, allowing for the solubility of U, Th, Zr, and REE [8]. However, the presence of abundant F suppresses alkalinity by reacting with Al to form AlF₆⁻³, which also depolymerizes the aluminum-silicate tetrahedral chain [8]. The high solubility of U and other high field strength (HFS) elements leads to their continuous and simultaneous enrichment during magma fractionation. U⁴⁺ and Th⁴⁺ exhibit similar geochemical behavior due to their comparable charge and ionic radii of 1.02 Å and 0.97 Å (coordination VIII), respectively, resulting in their incorporation into the same minerals [9]. The high charge and large ionic radii of U and Th prevent them from fitting into most common silicates, leading to their inclusion in complex accessory minerals of U, Th, Zr, Y, REE, Nb, and Ta during late-stage magmatic differentiation [10,11]. Consequently, U and Th are expected to be present in the same minerals in intrusive-type U deposits [12] formed during later paragenesis.

Pyrochlore is a group of Nb–Ta–Ti oxides with the ideal structural formula A₂B₂O₆Z [13,14,15]. The crystal structure of pyrochlore-group minerals is flexible and allows for the incorporation of various elements in the A-[Na, Ca, Mn, Ba, Fe, Sr, Sn, Pb, Sb, Y, REE, Th, U, (□), H₂O] and B-(Nb, Ta, Sb, W, Ti, Si, Zr, Hf, Sn, Fe, Al, V)sites. The Z-site is primarily occupied by F, OH, O, □, H₂O, or large monovalent cations (K, Rb, or Cs) [16]. Primary pyrochlore crystals are enriched in Ca, Na, Nb, Ta, and F. Late-stage pyrochlore, formed through hypogene and supergene alteration of primary pyrochlore, undergo a series of complex substitutions involving A- and B-site cations. The most common composition of late-stage pyrochlore is [(Ba, Sr, REE, Pb, Ca, U, Th)_{Σ << 2}(Nb, Ti, Ta, Zr, Fe³⁺, Si)₂(O, OH)₆(OH, F)_{Σ << 1}H₂O] [16,17].

The Madeira Sn-Nb-Ta (U, Th, REE, F) deposit stands out as a remarkable example of U-Th mineralization of intrusive type [18]. Unlike most U-Th deposits, where U and Th mineralizations occur together, in this deposit, they formed at different stages of magmatic evolution and are concentrated in distinct minerals: early pyrochlore and late thorite, respectively [18,19,20]. Subsequently, both minerals underwent intense hydrothermal alteration, leading to the development of two distinct mineral associations. The Th-enriched one includes Th-Fe-hydroxyfluorides [18], while pyrochlore alteration formed U-bearing columbite [19,20].

The U-bearing columbite in the Madeira deposit is considered by Minuzzi et al. [19] a secondary mineral formed through pyrochlore alteration, while Lenharo [21] and Costi [22] consider it a primary mineral. This aspect holds significant importance for the current study. In addition to fluorine, experimental research by Tang et al. [23] suggests that other factors such as temperature, increase in the A/CNK ratio [molar ratio of Al₂O₃/(CaO + Na₂O + K₂O)], and concentrations of essential compositional components of pyrochlore also influence the preferential formation of pyrochlore over columbite in peralkaline granitic magmas.

This paper focuses on the U mineralization associated with the world-class Madeira Sn-Nb-Ta (U, Th, REE, Li, cryolite) deposit [20,24]. It is noteworthy that the association of these metals in the same peralkaline rock, specifically an albite-enriched granite (AEG) (~1830 Ma) [22], which also hosts a massive cryolite deposit, is unmatched worldwide. The main objectives of this work are as follows:

(1) Assess the U mineralization potential of the Madeira deposit in comparison to other intrusive type U deposits;

(2) Investigate the formation of the primary U ore mineral, pyrochlore, and its relationship with columbite;

(3) Investigate the primary pyrochlore alteration under hypogene conditions;

(4) Evaluate the implications of the study results for the overall evolution of the albite-enriched granite system.

2. Geological Setting

The Pitinga Province is located (Figure 1) in the southern portion of the Guyana Shield [25], in the Tapajos-Parima Tectonic Province [26]. The Pitinga Province is the largest Sn producer in Brazil. The alluvial ore deposits were discovered in 1979 [27] and are almost exhausted. The primary ores are associated with two main tin-bearing granites, the Madeira and Agua Boa A-type granites (Figure 1). Both are part of the ∼1830 Ma Madeira Suite [22]. The Madeira deposit, which has been exploited since 1989, is associated with the Madeira granite (Figure 2). Moreover, a number of small greisens associated with the Agua Boa granite have been intermittently exploited [22].

The volcanic rocks of the Iricoume Group [27] predominate in the Pitinga Province and host the Madeira Granite (Figure 1). They have ²⁰⁷Pb/²⁰⁶Pb zircon ages between 1881 ± 2 and 1890 ± 2 Ma [28]. They comprise mostly effusive and hypabyssal rhyolites, highly welded ignimbrites, ignimbrite tuffs, and surge deposits formed in a subaerial environment with cyclic effusive and explosive activities [29,30,31].

The Madeira granite shown in Figure 1 and Figure 2 contains four facies [20,24,32,33,34]. The early, mostly meta-aluminous porphyritic amphibole-biotite granite (1824 ± 2 Ma) [35] contains plagioclase-mantled K-feldspar mega-crystals, sometimes also showing reverse zonation and thus referred to as the “rapakivi granite”. The amphibole-biotite granite was followed by somewhat younger meta-aluminous biotite granite (1822 ± 2 Ma) [35] which contains its xenoliths. The alkali feldspar hypersolvus porphyritic granite facies (1818 ± 2 Ma) [35] has K-feldspar phenocrysts in a fine- to medium-grained matrix dominantly composed of K-feldspar and quartz. According to Costi [22], the hypersolvus granite (Figure 2) and the AEG were emplaced simultaneously, and then interacted and intruded into the older facies. The age of the AEG is only very roughly constrained at 1822 ± 22 Ma [36] due to the metasomatic alteration of zircons.

The Madeira deposit (Figure 2) corresponds to the AEG. It is an oval-shaped body with an aerial extension of approximately 2 km × 1.3 km at the outcrop. It is divided into a subfacies called an albite-enriched granite core (AGC) and an albite-enriched granite border (AGB) [37]. The AGC is a peralkaline subsolvus granite, porphyritic to seriate in texture, fine- to medium-grained, and composed of quartz, albite, and K-feldspar in approximately equal proportions (25–30%). The accessory minerals are cryolite (4%), polylithionite (4%), green–brown mica (3%), zircon (2%), and riebeckite (2%). Pyrochlore, cassiterite, xenotime, columbite, thorite, magnetite, and galena occur in minor proportions (Horbe et al., 1985). The AGB is peraluminous and presents types of texture and essential mineralogy like the AGC, except for being richer in zircon, for the presence of fluorite instead of cryolite, and for the absence of iron-rich silicate minerals, which have almost completely disappeared due to an autometasomatic process [24,35].

Figure 2. Geological map of the albite-enriched granite (modified from Minuzzi [38]).

Figure 2. Geological map of the albite-enriched granite (modified from Minuzzi [38]).

The ore grade of the disseminated ore (AGC + AGB) stands at 0.17 wt.% Sn (cassiterite), 0.20 wt.% Nb₂O₅, and 0.024 wt.% Ta₂O₅ (both in pyrochlore and columbite) within a total of 164 Mt of rock [39]. The potential by-products of the disseminated ore are F (4.2 wt.% cryolite); Y and HREE [xenotime and gagarinite-(Y)]; Zr and Hf (zircon); Th (0.07 wt.% ThO₂, thorite); and U (0,03 wt.%, pyrochlore) [18,19,39,40,41]. Despite the disseminated character of the AEG mineralizations, there are zones of enrichment associated with the granite in which specific minerals may be considerably abundant:

(1)

Approximately 50-cm thick pods and bands of the pegmatitic AEG (rarely, up to 10 m thick) [42]. They have almost the same minerals as the AGC but with much larger grain sizes. Polylithionite, riebeckite, xenotime, and thorite are much more abundant than in the AGC.

(2)

Border pegmatites (BPEG) that are at the contact between the AGB and the older facies (Figure 2). They are characterized by the increased sizes and amounts of K feldspar; quartz and zircon; advanced alterations of K-feldspar and biotite; and local enrichments in fluorite, polylithionite, thorite, and secondary hematite [43].

(3)

Pegmatite veins which are not mappable occur more commonly in the central, northern, and northwest parts of the AGC, and they have thicknesses ranging from a few centimeters to 2 m. They are heterogeneous and more commonly porphyritic. The phenocrystals may be of quartz, K-feldspar, xenotime, thorite, cryolite, polylithionite, and riebeckite. The matrix is composed of albite, quartz, K-feldspar, polylithionite, cryolite, and riebeckite; the accessory minerals are zircon, cassiterite, pyrochlore, columbite, galena, sphalerite, hematite, gagarinite, and genthelvite [44].

(4)

The massive cryolite deposit (Figure 2) formed by several bodies of hydrothermal massive cryolite intercalated with AGC and hypersolvus granite; these are sub-horizontal, up to 300 m long and 30 m thick, and composed of cryolite crystals (∼87 vol%), quartz, zircon, and feldspar [45].

Zircon and pyrochlore are the only U-bearing minerals identified in the albite-enriched granite (AEG) prior to this study, and the pyrochlore was investigated with a focus on the Nb and Ta mineralization [19,20]. The high grades of U in the AEG have been attributed to the primary U-Pb-pyrochlore (12.2 wt.% UO₂, 29.8 wt.% PbO) [19]. The zircons from the AEG have preferentially incorporated Hf instead of Th (Zr/Hf < 20) [21,41]. The AGC zircons present average concentrations of 1.55 ppm UO₂, 8.24 ppm ThO₂, and a Th/U ratio of 5.31, whereas the AGB zircons present average concentrations of 2.97 ppm UO₂, 6.65 ppm ThO₂, and a Th/U ratio of 2.23 [41]. Xenotime grains from the AEG do not have significant concentrations of Th and U [46].

Costi et al. [24] consider the AEG to be the result of a phase-separation process, or immiscibility, similar to that registered by Thomas et al. [47] in the Variscan Erzgebirge granites, Germany. Bastos Neto et al. [20,36] consider that the AEG magma would have been related to the isotherm rise, which occurred when the mantle fluid ascended further into the crust promoting fenitization-type reactions [48] in rocks previously enriched in Sn, and introduced elements such as F, Nb, Y, REE, and Th in anomalous concentrations. The input of an F-rich fluid took place and generated metasomatism, causing the rock to become fusible. Lenharo [21] and Costi [22] considered that the magma of the AEG evolved towards an extremely Na-, F-enriched residual melt. In accordance with Bastos Neto et al. [20], the extreme fluorine enrichment in the residual melt is unlikely to have been attained, since the F content was buffered by the crystallization of magmatic cryolite [49].

3. Materials and Methods

For this study, we had a collection of more than 500 rock samples and their respective thin sections from the Universidade Federal do Rio Grande do Sul (UFRGS) research group. A total of 70 samples were selected for more detailed studies. To obtain detailed textural data, thin sections were examined by scanning electron microscopy (SEM) with qualitative analysis using an energy-dispersive X-ray detector (Zeiss, model EVO MA10; Manufacturer: Carl Zeiss AG, Oberkochen, Germany) at the Center for Microscopy and Microanalysis in UFRGS.

Electron probe microanalysis (EPMA) was carried out at the EPMA Laboratory of the Universidade de Brasília (UnB) with a JEOL JXA-8230 equipped with five WDS spectrometers for quantitative analyses and one EDS for qualitative analyses. The concentrations of F, Mg, Zn, Al, Si, Hf, Nb, P, Cl, S, Bi, Ti, Mn, Y, Ta, Sn, Ca, Zr, Fe, V, and Rb were determined with an accelerating voltage of 15 kV and 10 nA of sample current, whereas the concentrations of Na, Er, Tm, Yb, Ho, Lu, K, Pb, Dy, Tb, Sm, Gd, Eu, Sr, Th, Pr, Nd, Ce, La, Ba, and U were determined with an accelerating voltage of 20 kV and 50 nA. Each element was analyzed with a beam diameter of 1 μm. The counting times on the peaks were 10s for all elements, and half that time for background counts on both sides of the peaks. Relative analytical errors were 1% for major elements and 5% for minor elements. The crystals of the Wavelength Dispersive X-ray Spectrometers (WDS) are as the following: TAP (Al, Mg, Na, Si, Zn), PETJ (Bi, Cl, Hf, K, Nb, P, Pb, S, Sn, Sr, Ta, Th, Y), PETH (Rb, U, Zr), LIF (Dy, Er, Eu, Gd, Ho, Lu, Mn, Sm, Tb, Ti, Tm, Yb), LIFH (Ba, Ca, Ce, Fe, La, Nd, Pr, V), and LDE1 (F). The following natural and synthetic standards were used: albite (Na), andradite (Fe), apatite (Ca, P), BaSO₄ (Ba), baddeleyite (Zr), Bi₂O₃ (Bi), forsterite (Mg), HfO₂ (Hf), LiNbO₃ (Nb), LiTaO₃ (Ta), microcline (Al, K, Si), MnTiO₃ (Mn, Ti), PbS (Pb), pyrite (S), RbSi (Rb), SnO₂ (Sn), SrSO₄ (Sr), ThO₂ (Th), topaz (F), UO₂ (U), vanadinite (Cl, Pb, V), YFe₂O₁₂ (Y), ZnS (Zn), and synthetic REE-bearing glasses. Interference corrections were applied in all cases of peak overlap. Galena calibration: Pb (Mα) and S (Kα) were determined with an accelerating voltage of 20 kV and current of 50 nA and 20 nA, respectively, using the PETJ crystal for both elements, and with PbS (Pb) and pyrite (S) as standards.

Chemical data of the AEG and associated pegmatites were revised in order to de-fine the potential of each subfacies for U. Most of the whole-rock geochemical data (268 analyses) were obtained by the UFRGS research group and are available in Bastos Neto et al. [20,39], Minuzzi et al. [19,38,45], Pires [40], Stolnik [42], and Lengler [43]. The samples were collected from drill cores and fresh outcrops, and the analyses were performed at Actlabs (Canada). Major elements were determined by ICP-AES, the minor and trace elements by ICP-MS, and the F by ISE. The database was completed with data published by other research groups, which may be accessed in Lenharo [21], Costi [22], and Costi et al. [24,34].

4. Results

4.1. Mineralogy and Petrography

4.1.1. Pyrochlore

Pyrochlore from the AGC and AGB occurs as individual crystals dispersed within a matrix of quartz, albite, and orthoclase (Figure 3A,B). It can also be included in quartz (Figure 3C), polylithionite, and zircon (Figure 3D), or surrounded by recrystallized quartz (Figure 3E). In the AGC, it can be surrounded by hydrothermal cryolite (Figure 3F,G), while in the AGB, it can be surrounded by hydrothermal fluorite (Figure 3D,H,I). The crystal sizes range from 0.1 to 0.9 mm, similar to other minerals in the matrix. The grains are typically partially rounded, but pseudo-cubic crystals can also be observed. Under natural light, they appear to be dark yellow in color (Figure 3C,F). These rounded grains result from the in situ alteration of pyrochlore by hydrothermal fluids, leading to the formation of columbite. The associated columbite grains in both granite facies are opaque (Figure 3D,E,G–I). Even the well-preserved pyrochlore grains show incipient alteration along their edges and internal microfractures, mainly occurring in the AGC (Figure 3C,F). Advanced alteration is more prevalent in the AGB and central part of the AGC, where only remnants of the original pyrochlore can be observed (Figure 3D,H,I).

Twinning of pyrochlore crystals (Figure 3E) and intergrowth with zircon (Figure 3D) are observed in several cases. When included within quartz, the pyrochlore retains its well-formed euhedral crystal shape (Figure 3C). However, the contact between pyrochlore and the matrix minerals exhibits slight reactivity and undulation (Figure 3A). The contact between pyrochlore and hydrothermal cryolite in the AGC (Figure 3G) and hydrothermal fluorite in the AGB (Figure 3H,I) show even more corrosive features. In such cases, pyrochlore and columbite grains become highly rounded. A similar feature is observed at the interfaces between hydrothermal cryolite and fluorite with zircon crystals. Hydrothermal alteration of pyrochlore leads to the formation of columbite and iron oxide, which exsolve along the grain edges (Figure 3E,H). In a few instances from the AGB and central AGC, sphalerite is also associated (Figure 3I).

Due to the observed petrographic features, pyrochlore in both the AGC and AGB is considered a primary mineral that crystallized during the early magmatic stage. Subsequent hydrothermal events rich in fluorine affected and altered the pyrochlore throughout the AEG, with greater intensity in the AGB and central portion of the AGC. Columbite formation occurred during the early hydrothermal stage but was later corroded by fluids, leading to the formation of cryolite in the AGC and fluorite in the AGB. Analysis using X-ray dispersive energy spectroscopy (EDS) on numerous pyrochlore grains allowed for the identification of alteration products from the early and late hydrothermal stages. The following representative cases illustrate these differences.

Well-preserved pyrochlore crystals appear homogeneous with light gray tones in backscattered electron images (Figure 4A) and primarily consist of U-Pb-LREE-rich pyrochlore. Along the grain borders, microfractures and cavities are surrounded by white U-LREE-Pb-rich pyrochlore (Figure 4B). This white coloration is attributed to gradual enrichment in Pb because the concentrations of other elements do not vary significantly (Figure 4C). The grain showing incipient alteration (Figure 4D) consists of U-LREE-Pb-rich pyrochlore, displaying lower Pb concentration in its central portions (min. 7.5 wt.% PbO, light gray) and higher Pb concentrations along the grain border and microfractures (max. 14.5 wt.% PbO, white). Within the same grain, Fe-U-Pb-rich pyrochlore is surrounded by columbite (Figure 4E), exhibiting an irregular shape and a composition depleted in LREE-Nb-Ta-F but enriched in U-Pb-Fe-Si compared to U-LREE-Pb-rich pyrochlore (Figure 4F). Thus, even in the best-preserved pyrochlore grains, there is evidence of alteration, likely resulting from hydration and significant leaching in magmatic pyrochlore. As the degree of alteration increases, compositional heterogeneity becomes more pronounced, primarily attributed to the early hydrothermal process. This process gave rise to secondary phases enriched in Pb found at the grain borders and along microfractures, and to the formation of columbite. Consequently, the most extensively altered pyrochlore remnants are included within columbite. Grains exhibiting advanced alteration no longer contain primary pyrochlore but instead show remnants of hydrothermal pyrochlore, along with abundant columbite and/or iron oxide, as well as other secondary minerals.

4.1.2. Columbite

Different varieties of columbite formed in conjunction with hydrothermal pyrochlore. The predominant phase is Mn-Fe-rich columbite, present in both the AGC and the AGB. This phase initially formed at the borders of the pyrochlore grains, and the fluid responsible for its crystallization advanced through cleavages and other areas of weakness, corroding and filling microfractures and cavities within the crystal (Figure 4). Advanced stages of alteration are more prevalent in the AGB and the central portion of the AGC.

In the AGB, columbite within the same grain may exhibit heterogeneous composition. Mn-Fe-rich columbite is the predominant variety, with some portions enriched in elements inherited from pyrochlore, primarily U (Figure 5). The grain converted into columbite in Figure 5A consists predominantly of Mn-Fe-rich columbite, with subordinately Mn-U-Fe-rich columbite. The latter appears as irregular-shaped masses disseminated throughout the grain and has lower concentrations of Fe, Mn, and Nb, but higher concentrations of U and Si compared to the former (Figure 5B). Similarly, the grain in Figure 5C is composed of Mn-Fe-rich columbite, which surrounds irregular-shaped masses gradually richer in U. These masses consist of U-Mn-Fe-rich columbite and a Fe-U-Nb-rich phase, with the latter exhibiting significantly higher concentrations of U and Si (Figure 5D). The grain in Figure 5E is composed of Mn-U-Fe-rich columbite with anomalously high Si (Figure 5F), which surrounds irregular masses of Fe-U-Si-Nb-rich phase.

Columbite is commonly surrounded by iron oxide and quartz, and in the AGC, by hydrothermal cryolite, or in the AGB, by hydrothermal fluorite. This results in a rounded or irregular shape of the previously prismatic pyrochlore crystals (Figure 5). These secondary minerals also fill microfractures and voids in columbite and surround fragments of columbite.

4.1.3. Miscellaneous Products of Pyrochlore Alteration

In both the AGC and AGB, siliceous phases composed of U-Th-Zr-Y-REE and F are associated with columbite and pyrochlore grains showing advanced alteration. These silicates likely represent intermediate phases within the coffinite–thorite–zircon–xenotime mineral system, potentially comprising Zr-Y-HREE-rich coffinite and thorite grains with variable enrichments in Nb, F, and P. They are observed (i) as irregular-shaped masses surrounded by columbite, exhibiting reactive contact with it, such as the Th-U-rich silicate in Figure 6A; (ii) filling cavities within pyrochlore and columbite (Figure 6B); and (iii) included in the matrix adjacent to columbite. The textural relationships suggest that these silicates formed simultaneously with columbite and were significantly affected and corroded by late-stage hydrothermal fluids that precipitated cryolite (AGC), fluorite (AGB), iron oxide, and quartz.

LREE-rich fluorides are frequently associated with pyrochlore grains exhibiting incipient alteration (AGC) as well as with intensely altered grains (AGB). They can be enriched in Th, Y, and Ca and occur (i) arranged along the edges of the pyrochlore grains, in contact with or surrounded by columbite, displaying reactive contact with columbite, iron oxide, and quartz (Figure 4B and Figure 6C); and (ii) disseminated within pyrochlore and columbite grains as rounded and irregular-shaped masses in reactive contact with the associated minerals. In contexts (i) and (ii), the LREE-rich fluoride encompasses columbite fragments, and vice versa, suggesting simultaneous formation of these minerals. The LREE-rich fluoride underwent alteration by a hydrothermal fluid that created cavities subsequently filled by quartz and iron oxide.

Galena, associated with pyrochlore alteration, is found in the AGC only in the central part, near to the massive cryolite deposit. It appears as rounded crystals included in columbite, exhibiting an abrupt contact with columbite. In the AGB, galena is more common and occurs within or near grains almost totally altered into columbite, where it is included in iron oxide, fluorite, or other minerals. The contacts with all minerals are abrupt and irregular (Figure 5A). Less frequently, the following secondary phases have been observed associated with columbite: (i) Mn-Fe-Zn-rich sulfo-silicate (likely genthelvite) in the AGC and AGB, surrounding columbite with a corrosive contact (Figure 6D); (ii) sphalerite in the AGB and central part of the AGC, surrounding columbite grains and showing reactive contact with it (Figure 6A); (iii) Y-HREE-rich phosphate (probably xenotime) in the AGC and AGB, occurring as inclusions in columbite (Figure 6E) and in the matrix minerals with a dissolution-like appearance; (iv) LREE-rich phosphate (Figure 6F, probably monazite) in the AGB, situated between columbite grains and the surrounding iron silicate, with reactive contact with these minerals and included in the matrix; and (v) native Bi and Bi sulfide, in the AGC and AGB, measuring up to 5 μm, and occurring as inclusions in the matrix minerals surrounding pyrochlore, columbite, zircon, and thorite grains (Figure 6D).

4.1.4. Late Hydrothermal Alterations

The hydrothermal fluid, which gradually became enriched in F and Si, partially corroded the minerals formed during the early hydrothermal stage, such as columbite and U-rich silicates. Among these minerals, columbite was particularly affected by the late hydrothermal fluid, while secondary pyrochlore showed greater resistance. The borders of the pyrochlore grains made of columbite underwent significant dissolution, and the leached Fe from columbite may have been the primary source for the formation of surrounding iron oxide. This iron oxide fills microfractures and cavities within the columbite and pyrochlore grains (Figure 5A).

Following the crystallization of cryolite and fluorite, the remaining hydrothermal fluid, which was predominantly siliceous, caused intense hydraulic fracturing in both the pyrochlore grains of the AGC and the AGB (Figure 4). These fractures affected magmatic and hydrothermal pyrochlore, columbite, and other early secondary minerals, and were subsequently filled with hydrothermal quartz. Associated with these fractures are cavities that are also filled with quartz (Figure 5E). Hydrothermal quartz is also observed in reactive contact with iron oxide (Figure 5A) and fluorite (Figure 5B), causing disaggregation of these minerals and filling the resulting cavities.

4.1.5. Pegmatites

We conducted analyses (including petrography, EDS, EPMA, and whole-rock geochemistry) on a few dozen pegmatite samples, specifically examining pegmatite veins within the AGC, pegmatitic AGC, and border pegmatites. The pyrochlores found in these pegmatites are clearly inherited from the AGC and AGB, with no other primary U-rich minerals observed. However, due to the limited size and scope of the pegmatites, we will not provide detailed information about these pyrochlores and their alterations in this study. Instead, a separate article dedicated to the alteration of pyrochlore will cover these aspects extensively.

4.2. Mineral Composition

4.2.1. Pyrochlore

The representative compositions of pyrochlore are provided in

Table 1. EPMA data (in wt.%) for (1) U-Pb-LREE-rich pyrochlore, (2, 3) U-LREE-Pb-rich pyrochlore, (4) LREE-U-Pb-rich pyrochlore, (5) LREE-Pb-U-rich pyrochlore, (6, 7, 8) Fe-U-Pb-rich pyrochlore, (9) Pb-Fe-U-rich pyrochlore, (10, 11) Fe-U-rich pyrochlore, (12) Fe-Mn-U-rich pyrochlore.

Facies	Core Albite-Enriched Granite								Boder Albite-Enriched Granite
Crystal	(1)	(2)	(3)	(4) 1	(5) 1	(6)	(7) 1	(8)	(9)	(10)	(11)	(12)
Nb2O5	46.20	50.83	48.01	40.10	45.81	27.55	32.10	21.69	53.78	31.93	52.23	39.26
Ta2O5	06.13	06.15	06.25	02.47	03.72	05.99	02.43	04.56	04.99	01.69	04.75	02.66
SiO2	00.21	00.21	00.19	00.42	00.18	03.19	00.89	08.31	01.62	13.82	01.35	00.16
SnO2	01.16	02.72	02.31	00.72	02.60	01.11	00.90	00.00	00.00	b.d.l.	00.00	00.08
TiO2	01.05	00.47	00.50	00.92	b.d.l.	b.d.l.	01.17	00.81	01.43	01.45	00.90	02.61
UO2	02.81	02.24	03.08	06.97	06.54	04.06	07.39	13.51	08.38	12.64	08.56	13.73
ThO2	01.77	02.24	01.64	00.49	00.29	01.21	00.10	00.00	00.75	00.84	00.45	00.92
Y2O3	01.00	00.71	00.78	00.13	00.25	b.d.l.	00.00	00.00	00.11	00.25	00.30	00.16
La2O3	01.04	00.63	01.12	00.57	00.75	00.04	00.08	00.00	00.06	00.06	00.12	00.00
Ce2O3	03.43	02.19	03.77	02.38	03.15	00.58	00.93	00.33	00.85	00.62	01.19	00.47
Pr2O3	00.39	00.27	00.43	00.26	00.38	b.d.l.	00.00	00.00	00.06	00.18	00.15	00.06
Nd2O3	01.60	01.12	01.75	00.74	00.92	00.27	00.35	00.15	00.52	00.49	00.65	00.24
Sm2O3	00.56	00.43	00.35	00.13	00.21	b.d.l.	00.11	00.00	00.27	00.24	00.37	00.00
Eu2O3	b.d.l.	b.d.l.	00.00	00.00	00.04	00.00	00.00	00.00	b.d.l.	00.05	00.00	00.09
Gd2O3	b.d.l.	00.20	00.00	00.00	00.00	b.d.l.	00.00	00.00	00.06	00.24	00.00	00.00
Dy2O3	00.46	00.13	00.35	b.d.l.	00.00	b.d.l.	00.00	00.00	00.08	00.59	00.10	00.00
Ho2O3	b.d.l.	00.00	00.00	b.d.l.	b.d.l.	00.21	00.00	00.14	b.d.l.	00.20	00.00	00.00
Er2O3	00.18	00.13	00.11	00.05	00.21	00.00	00.00	00.00	00.10	00.39	00.00	00.00
Tm2O3	00.07	00.16	00.17	00.07	00.18	00.12	00.00	00.00	00.14	00.04	00.14	00.00
Yb2O3	00.20	00.08	00.00	00.08	00.05	00.08	00.00	00.00	00.23	00.25	00.06	00.00
Lu2O3	b.d.l.	b.d.l.	00.00	00.00	00.10	b.d.l.	00.00	00.00	00.12	00.16	00.17	00.00
FeO 2	00.69	00.28	00.16	01.70	00.10	02.71	02.93	03.10	05.92	03.24	04.30	04.62
CaO	01.42	01.73	01.30	01.04	03.01	b.d.l.	00.20	00.00	00.30	00.34	01.83	00.00
MnO	00.11	00.19	00.22	00.23	b.d.l.	00.00	00.12	00.09	01.20	00.49	00.66	05.98
PbO	07.23	14.51	07.52	13.98	05.56	30.69	17.23	25.94	02.85	00.02	02.19	00.64
Na2O	00.76	00.20	00.30	00.18	00.70	00.06	00.61	00.20	00.49	00.31	00.42	00.00
F	02.73	02.96	02.73	00.85	02.79	00.40	00.23	00.20	00.45	00.21	02.50	00.00
F=O2	−01.15	−01.24	−01.15	−00.36	−01.17	−00.17	−00.10	−00.08	−00.19	−00.09	−01.05	−00.00
Total 3	80.08	89.57	83.03	74.18	76.41	78.12	67.67	78.93	83.94	69.34	82.33	71.69
Structural formula based on a sum of 2 a.p.f.u. in the [6]B site
U4+	0.052	0.038	0.055	0.154	0.127	0.105	0.190	0.301	0.131	0.189	0.141	0.296
Th4+	0.034	0.039	0.030	0.011	0.006	0.032	0.003		0.012	0.013	0.008	0.020
Y3+	0.044	0.029	0.033	0.007	0.012				0.004	0.009	0.012	0.008
La3+	0.032	0.018	0.033	0.021	0.024	0.002	0.003		0.002	0.002	0.003
Ce3+	0.105	0.061	0.111	0.087	0.101	0.025	0.039	0.012	0.022	0.015	0.032	0.017
Pr3+	0.012	0.007	0.012	0.010	0.012				0.001	0.005	0.004	0.002
Nd3+	0.048	0.030	0.050	0.026	0.029	0.011	0.014	0.005	0.013	0.012	0.017	0.008
Sm3+	0.016	0.011	0.010	0.005	0.006		0.004		0.007	0.006	0.010	0.003
Eu3+					0.001					0.001
Gd3+		0.005							0.002	0.005	0.002
Dy3+	0.012	0.003	0.009						0.002	0.013
Ho3+						0.008		0.004		0.004
Er3+	0.005	0.003	0.003	0.002	0.006				0.002	0.008	0.003
Tm3+	0.002	0.004	0.004	0.002	0.005				0.003	0.001	0.001
Yb3+	0.005	0.002		0.002	0.001	0.003			0.005	0.005	0.004
Lu3+					0.003				0.003	0.003
Pb2+	0.162	0.297	0.163	0.373	0.131	0.958	0.537	0.700	0.054		0.044	0.017
Fe2+	0.048	0.018	0.011	0.141	0.008	0.263	0.283	0.260	0.349	0.182	0.267	0.375
Mn2+	0.008	0.013	0.015	0.020			0.011	0.007	0.072	0.028	0.042	0.491
Ca2+	0.127	0.141	0.112	0.111	0.282		0.025		0.023	0.024	0.146
Na+	0.123	0.030	0.046	0.035	0.119	0.013	0.138	0.038	0.067	0.041	0.060
Ʃ [8]A	0.835	0.750	0.698	1.006	0.872	1.423	1.248	1.329	0.774	0.565	0.797	1.238
Nb5+	1.739	1.747	1.744	1.795	1.805	1.441	1.677	0.982	1.714	0.968	1.753	1.721
Ta5+	0.139	0.127	0.137	0.067	0.088	0.189	0.076	0.124	0.096	0.031	0.096	0.070
Si4+	0.017	0.016	0.016	0.042	0.016	0.370	0.104	0.833	0.115	0.929	0.100	0.016
Sn4+	0.039	0.083	0.074	0.028	0.091		0.041					0.003
Ti4+	0.066	0.027	0.030	0.069			0.102	0.061	0.076	0.073	0.050	0.190
Ʃ [6]B	2.000	2.000	2.000	2.000	2.000	2.000	2.000	2.000	2.000	2.000	2.000	2.000
O2−	4.878	4.672	4.667	5.328	4.983	5.789	5.559	5.349	4.643	3.580	4.771	5.938
OH−	1.122	1.328	1.333	0.672	1.017	0.211	0.441	0.651	1.357	2.420	1.229	0.062
ƩX	6.000	6.000	6.000	6.000	6.000	6.000	6.000	6.000	6.000	6.000	6.000	6.000
F−	0.721	0.713	0.695	0.268	0.771	0.148	0.085	0.062	0.101	0.046	0.588
OH−	0.279	0.287	0.305	0.732	0.229	0.852	0.915	0.938	0.899	0.954	0.412	1.000
ƩY	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000

¹ Center of the core albite-enriched granite; ² Total Fe as FeO; ³ Calculated. Abbreviations: b.d.l. = below detection limit.

. The structural formula of pyrochlore (A_2-mB₂X_6-wY_1-n. pH₂O; m = 0–1.7, w = 0–0.7, n = 0–1, p = 0–2) was calculated based on the assumptions of Atencio et al. [14] and Ercit et al. [50]. These assumptions include: (i) charge balance in the crystal structure; (ii) full occupancy of octahedral B-sites by Nb⁵⁺, Ta⁵⁺, Ti⁴⁺, Si⁴⁺, and Sn⁴⁺ (i.e., site VI = 2); (iii) presence of a vacancy in the cubically coordinated A-site, which can be occupied by M⁴⁺ (U, Th), M³⁺ (Y, REE), M²⁺ (Pb, Fe, Ca, Mn), and Na⁺ (i.e., site VIII = 2-□); (iv) substitution of oxygen in the X-site by F⁻ and OH⁻; and (v) occupancy of the Y-site by F⁻, OH⁻, H₂O, and O²⁻. The OH content was calculated by considering the total cationic charges at sites A and B. In addition to these elements, concentrations of P, V, Zr, Hf, Al, Bi, Mg, Zn, Sr, Ba, K, Rb, Cl, and S were also analyzed and found to be present in the pyrochlore samples in concentrations ranging from hundreds to thousands of ppm. These elements were not included in the totals of the analyses and structural calculations.

The systematically low totals observed in the pyrochlore analyses can be attributed to several factors, including strong hydration of the mineral, metamictization (amorphization due to radiation damage), and the presence of voids. Niobium is the dominant substituent at the B-site in all pyrochlore crystals, with Nb₂O₅ concentrations ranging from 21.69 to 53.78 wt.%. Based on the relative proportions of Nb, Ta, and Ti in the B-site, all the samples belong to the pyrochlore group (Figure 7), as defined by Hogarth [51]. Silica (SiO₂) is clearly substituting for Nb in the B-site, with concentrations ranging from 0.16 to 13.82 wt.%. There is a strong negative correlation (−0.97) between the amount of silica and niobium content (Figure 8A). While Hogarth [51] considered silicon to be present as an impurity in pyrochlore, it has been reported by Lumpkin and Mariano [52] and by Johan and Johan [53] that high SiO₂ contents can be incorporated into the B-site of the pyrochlore lattice, reaching up to 7.9 wt.% and 10.12 wt.%, respectively.

In Hogarth’s [51] pyrochlore classification scheme, the individual varieties within the subgroup are defined by the A-site cations. In the AEG, various pyrochlore varieties can be found, including U-Pb-LREE-rich, U-LREE-Pb-rich, LREE-Pb-U-rich pyrochlore, Fe-U-Pb-rich, and Fe-U-rich pyrochlores (Table 1), as well as others with less common proportions of A-site cations such as U, Pb, Fe, LREE, Ca, Th, Na, and Mn. The diverse compositions of pyrochlore reflect the intense hydrothermal alteration that affected the original magmatic pyrochlore. In this work, the classification scheme proposed by Atencio et al. [14] was not utilized. Instead, the terminology and descriptive terms introduced by Hogarth [51] were employed, as they were more suitable for the SEM analysis conducted in this investigation. The highly altered and hydrated nature of the pyrochlore crystals necessitated the use of more descriptive terms to maintain coherence and continuity with the SEM-based observations.

The pyrochlore samples analyzed in this study exhibit a range of U contents, with concentrations varying from 2.24 to 13.73 wt.% UO₂. Pyrochlore crystals in the AGB and the central portion of the AGC generally show higher U enrichment (average ~7.75 wt.% UO₂) compared to the rest of the AGC (average ~5.05 wt.% UO₂). Among the pyrochlore species analyzed, the Fe-U-Pb-rich pyrochlore demonstrates the highest U content (4.06 to 13.51 wt.% UO₂) and is also associated with significant concentrations of Pb (17.23 to 30.69 wt.% PbO₂). The maximum observed Th content in pyrochlore is 2.24 wt.% ThO₂, and it occurs in the U-LREE-Pb-rich species. In the U-rich species, Th content is generally lower than 1 wt.% ThO₂ or absent. Fluorine is the dominant anion in the Y-site of the LREE-enriched pyrochlore species, with concentrations of up to 2.96 wt.% F, while OH is dominant in the Fe-U-Pb-enriched species. According to Johan and Johan [54], the high U concentrations in the A-site of defect pyrochlore (A²⁺ □ B₂⁵⁺ O₆ □) can lead to the appearance of significant M⁴⁺ in the B-site, which supports the presence of Si, Ti, and Sn in the B-structural site of the albite-enriched granite pyrochlore. The high vacancies in the A-site can be explained by a hypothetical end-member U⁴⁺ □ B₂⁴⁺ O₆ □. Another compatible substitution scheme is 2Ca²⁺ + 2(Nb, Ta)⁵⁺ or Na⁺REE³⁺ + 2(Nb, Ta)⁵⁺ ↔ (Pb, Fe)²⁺U⁴⁺ + 2(Si, Ti)⁴⁺. This is supported by the positive correlation (0.76) observed between U + Pb + Fe and Si concentrations (Figure 8B).

The central portions of the less altered grains in the AGC can be considered as relict varieties of primary pyrochlore, which are relatively rich in LREE, as observed in Table 1. These crystals are (1) U-Pb-LREE-rich pyrochlore, (2, 3) U-LREE-Pb-rich pyrochlore, (4) LREE-U-Pb-rich pyrochlore, and (5) LREE-Pb-U-rich pyrochlore. Some grains of U-LREE-Pb-rich pyrochlore (Table 1, analyzes 2, 3) also exhibit a pattern where the central portions have lower Pb concentration (min. 7.5 wt.% PbO), while higher Pb concentrations are observed along the border and microfractures (max. 14.5 wt.% PbO). In this case, there is an inverse correlation between Pb and LREE (−0.93), and the U content does not show significant variation.

The alteration of U-Pb-LREE-rich pyrochlore in both AGC and AGB involves the loss of LREE and Nb, resulting in a progressive enrichment of Pb, U, Fe, and Si. This overall trend is shown in Figure 8C, although it should be noted that the alteration process is not a continuous progression. Through the detailed SEM study, corroborated by EPMA analyses, it was possible to compare samples with different degrees of pyrochlore alteration. This comparison reveals that the alteration process involved preferential leaching of specific cations in successive stages. It is important to note that the enrichments observed were not solely relative; there was also the incorporation of elements by the pyrochlore. Figure 9 summarizes the main exchanges that occurred during the alteration process. The first stage of alteration involved the leaching of LREE (Figure 8D). In the second stage, the loss of LREE was accompanied by losses of Nb and F, while Fe and Si were incorporated, leading to a significant relative enrichment in U and Pb (Figure 8E,F). This stage resulted in the formation of the most common variety of pyrochlore in the AGC, the Fe-U-Pb-rich pyrochlore (Table 1, analysis 6, 7, 8). In the third stage, losses of Pb and Fe began, along with continued losses of Nb and F. This gave rise to the varieties Pb-Fe-U-rich pyrochlore, Fe-U-rich-pyrochlore, and Fe-Mn-U-rich pyrochlore (Table 1, analysis 9, 10, 11, 12), which are the richest in U and are commonly found in the AGB and the central zone of the AGC. Previous studies by Minuzzi et al. [38] described continuous losses of Pb and Fe since the initial stage of alteration, which they attribute to the relative enrichment in U. However, these authors did not recognize and analyze the pyrochlores corresponding to the first two stages described in this study.

4.2.2. Columbite

As the pyrochlore alteration progresses, a significant leaching of Pb occurs from its structure, leading to the collapse of the pyrochlore phase and the formation of columbite and other hydrothermal phases.

Table 2. EPMA data (in wt.%) for (1, 2) Mn-Fe-rich columbite, (3, 4) U-Mn-Fe-rich columbite, (5) U-Fe-Mn-rich columbite, and (6, 7, 8) Mn-U-Fe-rich columbite.

Facies	AGC	AGB	AGB	AGB	AGB	AGB	AGC 1	AGB
Crystal	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Nb2O5	66.74	68.93	65.61	66.67	63.63	51.29	40.31	57.12
Ta2O5	03.32	05.87	05.72	05.17	04.21	05.10	03.17	04.64
SiO2	00.51	00.15	00.57	00.18	01.70	04.60	15.80	00.68
SnO2	b.d.l.	b.d.l.	b.d.l.	01.72	00.20	b.d.l.	00.84	b.d.l.
TiO2	02.57	01.26	02.36	01.63	03.20	01.14	02.74	02.74
UO2	01.15	00.61	03.64	03.48	05.19	06.34	06.72	08.82
ThO2	b.d.l.	00.05	00.18	00.32	00.44	01.48	00.56	00.78
Y2O3	00.12	b.d.l.	00.07	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.22
La2O3	b.d.l.	00.06	b.d.l.	b.d.l.	b.d.l.	00.13	b.d.l.	b.d.l.
Ce2O3	00.05	00.07	00.18	00.13	00.33	00.42	00.11	00.42
Pr2O3	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.
Nd2O3	b.d.l.	b.d.l.	00.05	00.05	00.10	00.14	00.03	00.29
Sm2O3	00.17	00.10	00.18	00.14	b.d.l.	00.25	b.d.l.	00.28
Eu2O3	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.04	b.d.l.	00.12
Gd2O3	00.08	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.15
Dy2O3	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.40
Ho2O3	b.d.l.	b.d.l.	00.15	b.d.l.	00.13	00.10	b.d.l.	00.15
Er2O3	b.d.l.	00.07	00.16	00.18	00.09	00.10	b.d.l.	00.38
Tm2O3	b.d.l.	b.d.l.	00.16	b.d.l.	00.09	00.13	00.06	00.11
Yb2O3	b.d.l.	00.05	00.15	00.07	00.07	00.10	00.06	00.31
Lu2O3	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.07
FeO 2	15.33	11.79	16.13	12.27	08.37	10.08	07.69	13.85
CaO	00.40	00.37	b.d.l.	00.25	00.28	00.99	b.d.l.	b.d.l.
MnO	06.70	08.72	04.92	07.28	10.05	02.25	04.39	04.99
PbO	00.81	b.d.l.	00.06	00.13	00.38	02.83	03.78	00.36
Na2O	b.d.l.	00.03	00.04	00.07	b.d.l.	00.76	b.d.l.	00.11
F	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.44	b.d.l.	b.d.l.
F = O2	−00.00	−00.00	−00.00	−00.00	−00.00	−00.18	−00.00	−00.00
Total 3	97.89	98.03	99.74	99.73	98.45	88.50	86.15	95.71
Structural formula based on 3 cations and 6 oxygens
Fe2+	0.725	0.571	0.771	0.598	0.408	0.551	0.385	0.708
Mn2+	0.321	0.428	0.238	0.359	0.496	0.125	0.223	0.258
Ʃ [8]A	1.046	0.999	1.009	0.957	0.903	0.676	0.607	0.966
Nb5+	1.704	1.803	1.693	1.754	1.674	1.515	1.089	1.578
Ta5+	0.051	0.093	0.089	0.082	0.067	0.091	0.052	0.077
Si4+	0.029	0.008	0.032	0.010	0.099	0.301	0.946	0.042
Sn4+				0.040	0.005		0.020
Ti4+	0.109	0.055	0.101	0.071	0.140	0.056	0.123	0.126
U4+	0.014	0.008	0.046	0.045	0.067	0.092	0.089	0.120
Th4+		0.001	0.002	0.004	0.006	0.022	0.008	0.011
Y3+	0.004		0.002					0.007
La3+		0.001				0.003
Ce3+	0.001	0.002	0.004	0.003	0.007	0.010	0.002	0.009
Pr3+
Nd3+			0.001	0.001	0.002	0.003	0.001	0.006
Sm3+	0.003	0.002	0.004	0.003		0.006		0.006
Eu3+						0.001		0.002
Gd3+	0.002							0.003
Dy3+								0.008
Ho3+			0.003		0.002	0.002		0.003
Er3+		0.001	0.003	0.003	0.002	0.002		0.007
Tm3+			0.003		0.002	0.003	0.001	0.002
Yb3+		0.001	0.003	0.001	0.001	0.002	0.001	0.006
Lu3+								0.001
Pb2+	0.012		0.001	0.002	0.006	0.050	0.061	0.006
Ca2+	0.025	0.023		0.016	0.017	0.069
Na+		0.004	0.004	0.007		0.096		0.013
Ʃ [8]B	1.954	2.001	1.991	2.043	2.097	2.324	2.393	2.034
Mn/(Mn + Fe)	0.307	0.428	0.236	0.375	0.549	0.184	0.367	0.267
Ta/(Ta + Nb)	0.029	0.049	0.050	0.045	0.038	0.056	0.045	0.047

¹ Center of the core albite-enriched granite; ² Total Fe as FeO; ³ Calculated. Abbreviations: AGC = albite-enriched granite core, AGB = albite-enriched granite border, b.d.l. = below detection limit.

presents examples of columbite compositions, showcasing variations in Fe, Mn, and U contents. The observed compositions include (a) Mn-Fe-rich columbite, which is the most common composition found in both AGC and AGB (crystals 1, 2); (b) U-Mn-Fe-rich columbite, characterized by ~3 wt.% UO₂ (crystals 3, 4); (c) U-Fe-Mn-rich columbite (up to 10.05 wt.% MnO₂, crystal 5), observed only in the AGB and the central zone of AGC; and (d) Mn-U-Fe-rich columbite, typically exhibiting high U content (~6 wt.% UO₂, crystals 6, 7, 8). The U-enriched columbite species are commonly associated with highly or completely altered pyrochlore grains. The presence of U in columbite is often, but not always, associated with Si, which ranges from 0.15 to 15.8 wt.% SiO₂. In the general formula AB₂O₆, both U and Si occupy the B-site, substituting Nb and Ta, along with Pb and Ti (Figure 10A). The systematic excess in the B-site and the vacancy in the A-site are associated with high Si contents (correlation of 0.84, Figure 10B), suggesting a coupled substitution mechanism involving both the A- and B-sites in the columbite crystal structure, i.e., (Fe, Mn)²⁺ + 2(Nb, Ta)⁵⁺ → □_A + 3(Si, U, Th, Ti, Sn)⁴⁺. The contents of REE₂O₃ (0.26–2.68 wt.%) in the columbite compositions are also unusually high compared to general studies of columbite composition [55,56]. The A-site in the columbite structure is occupied by Fe and Mn, which can substitute for each other (Figure 10C). All columbite crystals exhibit enrichment in Fe (from 7.69 to 16.13 wt.% FeO), and the majority of them are classified as columbite-(Fe) [57], with Mn/(Mn + Fe) atomic ratios ranging from 0.184 to 0.549 (Figure 10D).

4.2.3. Other Products of Pyrochlore Alteration

Secondary minerals associated with columbite formation exhibit enrichment in U and often have a non-stoichiometric multivariate composition (

Table 3. EPMA data (in wt.%) for the following secondary minerals: (1) Pb-Fe-U-Nb-rich phase; (2) REE-Mn-Fe-U-Nb-rich phase; (3) Fe-U-Nb-rich phase; (4) Fe-U-Si-Nb-rich phase; (5) Th-U-Si-Nb-rich phase; (6) Th-U-rich silicate; (7) REE-Y-U-rich silicate; (8) U-Th-rich silicate; (9) U-Pb-Th-Zr-rich silicate; (10) U-LREE-rich fluoride; (11) Th-LREE-rich fluoride.

Facies	AGB	AGC	AGB				AGC
Crystal	(1)	(2) 1	(3)	(4)	(5)	(6)	(7)	(8) 1	(9)	(10) 1	(11)
Nb2O5	53.78	49.58	46.52	16.23	13.48	03.54	00.40	01.13	06.52	00.09	02.49
Ta2O5	04.99	03.03	04.97	01.98	00.33	b.d.l.	b.d.l.	b.d.l.	01.41	b.d.l.	b.d.l.
P2O5	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.43	03.52	01.04	01.05	00.07	00.10
SiO2	01.62	02.62	05.42	10.19	07.75	14.02	14.13	11.53	14.39	00.04	00.07
SnO2	b.d.l.	0.79	b.d.l.	b.d.l.	b.d.l.	b.d.l.	b.d.l.	00.23	b.d.l.	b.d.l.	b.d.l.
TiO2	01.43	02.37	03.71	01.97	01.80	00.41	b.d.l.	00.15	01.18	b.d.l.	00.22
UO2	08.38	13.35	22.94	19.91	42.38	34.35	21.21	17.34	04.48	03.81	00.65
ThO2	00.75	05.19	01.08	01.48	08.23	10.00	04.70	30.39	11.82	03.13	11.61
ZrO2	b.d.l.	b.d.l.	b.d.l.	00.61	0.848	00.68	00.18	02.52	13.32	b.d.l.	b.d.l.
Y2O3	00.11	00.21	00.18	00.33	b.d.l.	01.07	10.27	03.51	01.89	b.d.l.	01.72
La2O3	00.06	00.02	00.06	00.04	00.08	00.05	b.d.l.	b.d.l.	b.d.l.	09.08	05.82
Ce2O3	00.85	00.44	00.47	00.22	00.85	00.68	00.02	00.18	00.23	26.33	17.03
Pr2O3	00.06	00.08	00.05	b.d.l.	00.15	00.19	b.d.l.	00.09	b.d.l.	03.55	02.22
Nd2O3	00.52	00.40	00.20	00.05	00.51	00.63	00.10	00.25	00.27	08.89	08.34
Sm2O3	00.27	00.18	00.15	00.06	00.19	00.21	b.d.l.	00.63	00.24	00.87	02.94
Eu2O3	b.d.l.	00.11	00.05	b.d.l.	b.d.l.	00.08	00.08	b.d.l.	b.d.l.	00.37	00.40
Gd2O3	00.06	00.35	00.10	b.d.l.	00.07	00.08	00.50	00.77	00.29	b.d.l.	00.38
Dy2O3	00.08	00.72	00.12	b.d.l.	b.d.l.	b.d.l.	02.43	00.23	00.71	b.d.l.	00.69
Ho2O3	b.d.l.	b.d.l.	b.d.l.	00.10	b.d.l.	00.24	00.49	01.00	00.14	00.13	00.28
Er2O3	00.10	00.54	00.19	00.11	00.12	00.21	02.06	01.64	00.44	00.30	00.26
Tm2O3	00.14	00.09	b.d.l.	00.04	b.d.l.	b.d.l.	00.26	00.30	00.11	b.d.l.	b.d.l.
Yb2O3	00.23	00.39	00.14	00.22	00.25	00.48	02.09	00.85	00.28	00.08	00.12
Lu2O3	00.12	00.11	b.d.l.	b.d.l.	b.d.l.	00.13	00.57	b.d.l.	00.14	b.d.l.	00.11
FeO 2	05.92	12.09	10.28	05.63	00.49	00.17	00.80	00.25	01.47	00.12	01.50
CaO	00.30	b.d.l.	00.28	b.d.l.	00.23	b.d.l.	00.62	00.34	00.65	b.d.l.	02.11
MnO	01.20	05.47	02.32	00.25	00.23	b.d.l.	00.13	b.d.l.	00.38	00.12	00.48
PbO	02.85	01.30	b.d.l.	00.58	00.08	b.d.l.	00.09	01.23	10.70	00.55	00.29
Na2O	00.49	00.14	00.44	00.26	b.d.l.	b.d.l.	00.02	00.04	00.17	00.12	00.24
F	00.45	00.49	b.d.l.	b.d.l.	01.24	02.97	03.92	04.54	02.61	17.31	10.75
F = O2	−00.19	−00.21	−00.00	−00.00	−00.52	−01.25	−01.65	−01.91	−01.10
Total 3	83.94	101.88	99.22	61.66	80.13	69.43	66.94	81.10	73.99	74.96	70.82

¹ Center of the core albite-enriched granite; ² Total Fe as FeO; ³ Calculated. Abbreviations: AGC = albite-enriched granite core, AGB = albite-enriched granite border, b.d.l. = below detection limit.

). In the AGB, a Pb-Fe-U-Nb-rich hydrothermal phase commonly observed displays Nb content up to 53 wt.% Nb₂O₅ (Table 3, crystal 1). However, the mineral species’ nature remains unclear, as its composition is intermediate between pyrochlore and columbite, and the stoichiometry resembles that of either a U-rich columbite or a highly vacant U-rich pyrochlore (Table 3, crystals 1, 2). Another hydrothermal phase, Si-Fe-U-Nb-rich with higher U (22.94 wt.% UO₂, Table 3, crystal 3) could potentially be anoxi-petcheskite [U⁴⁺(Fe³⁺_2/3 □_1/3)(Nb,Ta)₂O₇(O,OH)] [58] if Si occupies the U structural site. The calculated structural formula, assuming (Nb + Ta) = 2 a.p.f.u. is (U⁴⁺_0.41Si_0.43Th_0.02Y_0.01REE_0.03Ca_0.02Na_0.07)_0.99(Fe³⁺_0.62Mn_0.16□_0.22)_0.78(Nb_1.67Ta_0.11Ti_0.22)₂O₇(O_0.6OH_0.4). Petcheskite occurrences are typically associated with pyrochlore supergroup minerals, but none have Si contents. Petscheckite found in the Hagendorf–Süd pegmatite in Germany [59] was discovered included in columbite, while in the Antsakoa I pegmatite in Madagascar [58], columbite and petscheckite form a primary diataxial intergrowth. Heating experiments on the liandratite–petschekite series [58] at 1000 °C reveal that hydroxy-petscheckitere acts towards a uraniferous pyrochlore composition. Hence, it is reasonable to suggest that the inverse reaction may have occurred during pyrochlore alteration in the AEG, involving continuous hydration of U-enriched pyrochlore with a Fe-enriched fluid. Furthermore, the Fe-U-Si-Nb-rich hydrothermal phase (Figure 5E) is enriched in Si (10.19 wt.% SiO₂, Table 3, crystal 4) and likely represents an intermediate phase between U-enriched pyrochlore and U-enriched silicate. Similarly, the Th-U-Si-Nb-rich hydrothermal phase (Figure 6A) contains up to 42.38 wt.% UO₂ and 8.23 wt.% ThO₂ (Table 3, crystal 5).

Another pathway for the alteration of the aforementioned pyrochlore, caused by pervasive supergene alteration near-surface (vadose zone) and extending down to the phreatic hydraulic zone close to the fresh granitic host rocks, should also be considered [60,61]. The hydration of U-enriched pyrochlore with an Fe-enriched fluid and an A-site deficiency may, in places, also mark a transition into supergene alteration, which might result in Nb-enriched Fe oxide-hydroxides or goethite at shallower depths [61]. This alteration can be recognized down to 200 m in drill holes and is due to the reducing hydraulic conditions at that depth. It is often overlooked because of the absence of “yellow uranium minerals”, which typically form under near-ambient, strongly oxidizing conditions near the surface and are known as marker minerals typical of U ore deposits [62].

Pyrochlore alteration typically yields U-rich silicates and LREE-rich fluorides, along with columbite. Uranium-rich silicates occur in highly or completely altered pyrochlore grains, making them more common in the AGB and the central zone of the AGC. Uranium content ranges from 4.48 to 34.35 wt.% UO₂, along with variable concentrations of Th, Y, REE, and Zr (Table 3, crystals 6-9), consistent with intermediate compositions in the coffinite-thorite-xenotime-zircon mineral system. However, the general formula ABX₄ (A = U, Th, Y, REE, Pb, Fe, Mn, Ca, Na; B = Si, Ti, Sn, P, Nb, Ta; X = O, F, OH) calculated to yield X = 4 reveals a systematic deficit in both A and B-sites [A_1-□B_1-□X)₄], with □_A = 0.01–0.39 and □_B = 0.01–0.42, likely due to high Nb⁵⁺ contents in the B-site (up to 6.52 wt.% Nb₂O₅, Table 3, crystal 9). Incorporation of Nb associated with high F amounts (2.97 to 4.54 wt% F, Table 3) limits the occurrence of OH in the structure, although significant amounts of molecular H₂O should occur given the low totals of all secondary minerals in this mineral system. LREE-rich fluorides are often found alongside pyrochlore grains with incipient alteration (AGC), as well as with intensely altered grains (AGB), and can be enriched in U (up to 3.81 wt.% UO₂), Th, Y, and Ca (Table 3, crystals 10, 11).

4.3. Geochemical Distribution of Uranium in the Albite-Enriched Granite and Pegmatites

The average concentrations of U and Th, as well as the Th/U ratios, in the albite-enriched granite (AEG) subfacies and the associated pegmatites are presented in

Table 4. U and Th contents and Th/U ratios in the albite-enriched granite core (AGC), albite-enriched border (AGB), and pegmatites (PEG). Number of analyses in parentheses (Data from Hadlich et al. [18]).

		AGC	AGB	PEG
UO2 (ppm)	Min.	40.00	34.00	20.00
	Max.	1610.00	796.00	1180.00
	Avg.	322 (111)	345 (54)	553
ThO2 (ppm)	Min.	70.00	36.10	1080.00
	Max.	2388.00	2419.00	18,400.00
	Avg.	800 (113)	696 (71)	5127 (98)
Th/U	Min.	0.29	0.13	3.30
	Max.	30.40	8.94	389.50
	Avg.	3.82 (110)	1.85 (53)	19.85 (64)

. The average U concentration in the AGC is 322 ppm UO₂, with values reaching as high as 1600 ppm UO₂. The AGB exhibits a slightly higher average U content of 345 ppm UO₂, with a maximum of 796 ppm UO₂. The pegmatite veins associated with the AGC display the highest U concentrations, averaging 553 ppm UO₂. Regarding Th, the average content in the AGC is 800 ppm ThO₂, while in the AGB, it is 696 ppm ThO₂, with a maximum content of 1.8 wt.% ThO₂ observed in pegmatites. The combined AEG (AGC + AGB) has average U and Th concentrations of 329 ppm and 760 ppm, respectively. Consequently, the U content in the Pitinga mine exceeds the average U concentration in granites/rhyolites (4.5 ppm) by a factor of over 73 [8]. The average Th/U ratio ranges from 1.85 (AGB) to 3.82 (AGC), both of which are lower than the world average Th/U ratio for acid igneous rocks (5.6). However, rocks that have undergone significant alteration with post-magmatic mobilization typically exhibit Th/U ratios below 3 [63].

The maps depicting U, Th, Nb, and Zr concentrations in the AEG (Figure 11) reveal distinct patterns. The highest U values are observed in the northern, northeastern, and central regions of the granite body. The areas with the highest U grades coincide with the regions exhibiting elevated Nb concentrations. However, the distribution of U does not correspond to that of Zr, as Zr contents remain relatively consistent throughout the AEG. These findings suggest that the U mineralization within the AEG is predominantly associated with pyrochlore and its alteration products. While zircon is abundant, it exhibits low U grades, averaging 1.5 ppm UO₂ in the AGC and 3.0 ppm UO₂ in the AGB [41]. Notably, xenotime grains from the AEG do not display significant U concentrations [46], and thorite exhibits an average U content of 0.35 wt.% UO₂ [18].

5. Discussion

5.1. Primary Pyrochlore: Formation, U-Enrichment, and Distribution in the AEG

Uranium-bearing deposits typically encompass various uranium minerals, which vary depending on the ore genesis. The primary minerals uraninite (UO₂) and coffinite (USiO₄), both reduced U⁴⁺ minerals, are the most common. These minerals form during magma crystallization and are often found in association with feldspar and quartz. On the other hand, the majority of exploitable uranium minerals are considered secondary uranium minerals. They include pitchblende (U₃O₈), carnotite [K₂(UO₂)₂₍VO₄)₂ xH₂O], autunite [Ca(UO₂)₂(PO₄)₂xH₂O], and uranophane [Ca(UO₂)₂(HSiO₄)₂xH₂O]. Additionally, there are less common refractory uranium minerals such as brannerite (UTi₂O₆), davidite [(La,Ce,Ca)(Y,U)(Ti,Fe³⁺)₂₀O₃₈], and betafite [(Ca,U)₂(Nb,Ti,Ta)₂O₇] [3].

According to Bea [64], the nature, composition, and associations of the primary assemblage of accessory minerals rich in U, Th, REE, and Y vary according to the Al enrichment of the rock. In peraluminous granites, the primary assemblage includes monazite, xenotime, apatite, zircon, Th-orthosilicates, uraninite, and betafite-pyrochlore. Meta-aluminous granites are associated with allanite, sphene, apatite, zircon, monazite, and Th-orthosilicates. Peralkaline granites exhibit an assemblage of arscinite, fergusonite, samarskite, bastnaesite, fluocerite, allanite, sphene, zircon, monazite, xenotime, and Th-orthosilicates. The size and density of these accessory minerals, rich in U, Th, REE and Y, are too small to settle by gravity in the magmatic chamber. As a result, these minerals form at the beginning of magmatic crystallization and remain suspended in the magmatic melt until they are included into the crystallization of some major mineral [64].

The characteristics and concepts described above do not apply to the Pitinga deposit, as both the primary U ore mineral and the secondary paragenesis differ significantly from the aforementioned descriptions. Th/U ratio averages of 1.85 in the AGB and 3.82 in the AGC [18] attest to the high availability of U in the earlier stages of magma evolution. The zircon abundance in the AGB (5 vol.%, Th/U_avg = 2.23) relative to AGC (2 vol.%, Th/U_avg = 5.33) [39,41] suggests that the enrichment in U of the AGB is more likely due to primary magmatic crystallization rather than to post-magmatic processes, as suggested by Killeen [63] for rocks with Th/U ratio below 3 [18]. Fluorine-bearing complexes transported Sn and HFS elements throughout the melt, leading to the dispersed nature of cassiterite and U-Pb-pyrochlore mineralization during the early magmatic stage [20]. However, the extreme enrichment of F in the residual melt [21] was prevented due to the buffering effect of magmatic cryolite crystallization [49]. This crystallization process hindered the formation of zones with higher concentrations of ore.

The early occurrence of pyrochlore in F-rich magmas, which led to the formation of granites containing disseminated cryolite, has been documented in the albite arfvedsonite granite of the Ririwai Complex [65]. The preferential crystallization of pyrochlore over columbite in this context was attributed to the high fluorine content in the system [66]. It is important to note that the early formation of pyrochlore does not solely rely on an extremely high concentration of fluorine in the melt. The solubility product (Ksp) of pyrochlore is only weakly influenced by the fluorine content when concentrations exceed 1 wt.% [23].

In peralkaline granitic melts with A/CNK < 1, the Ksp values of pyrochlore are lower than those of columbite. Conversely, in peraluminous melts with A/CNK > 1, the Ksp values of pyrochlore are higher than those of columbite. In sub aluminous melts, the Ksp values of pyrochlore and columbite are nearly the same [23]. Tang et al. [23] proposed three specific controls on pyrochlore crystallization during the evolution of peralkaline magma. (1) The Ksp of pyrochlore decreases significantly with decreasing temperature; in order for magmas to exist at low temperatures, they must be highly fluxed, which may explain the common occurrence of Nb mineralization in F-rich granites. (2) Increase in the A/CNK ratio of the melt, related to processes such as fractionation, assimilation, and alkali diffusion active during magma evolution. (3) Pyrochlore crystallizes when the concentrations of the essential components (ESCs) that compose pyrochlore reach the solubility product. At the AEG, the F richness, as well as the lower temperature of the magma, were the key factors determining the crystallization of primary pyrochlore instead of columbite.

The primary pyrochlore has ~2% F, which corresponds to an OH site occupation of ~30%. This F content may seem relatively low considering the richness of F in the magma and the high concentration of U in the mineral, given the affinity between these two elements. The interaction of columbite and uraninite with a fluid–magma system, consisting of a melt of Li–F-granite and fluoride fluid at 750 °C and P = 2300 bar [67], leads to the formation of zonal pyrochlores with considerably distinct uranium and fluorine contents. The fluorine-rich pyrochlores (7–12 wt.% F) preserve the Nb/U ratio of the initial columbite (15–30). Conversely, uranium-bearing pyrochlores contain 2–4 times lower amounts of fluorine, above 2 at.% U, and the Nb/U molar ratio decreases to 5–15. The trends in Ca, U, and F concentrations suggest the influence of temperature on the reactions involving the exchange of Ca²⁺ and U⁴⁺ cations. The concentrations of U and F in the U-Pb-LREE-rich pyrochlore in the present study are similar to those found in the uranium-bearing pyrochlore from the experiment. Therefore, the enrichment of uranium in the U-Pb-LREE-rich pyrochlore within the magma likely accompanied the loss of F and Ca.

The greater stability of HREE in F-bearing complexes [41] contributed to the preferential incorporation of LREE into pyrochlore. Zircon crystallization in the early magmatic stage was greatly inhibited due to high F content and alkalinity [68]. As the crystallization of hydrous Na-bearing silicates commenced, the reduction in alkalinity enabled intensified zircon crystallization, accompanied by the formation of xenotime and thorite. Consequently, zircon originated from a magma that was previously depleted in U, Nb, Ta, and LREE [18].

5.2. Primary Pyrochlore Hydrothermal Alteration and Its Products

All pyrochlore crystals in the AGC and AGB were affected by hydrothermal alteration caused by F-rich aqueous fluids that formed the massive cryolite deposit. In the most altered grains, only remnants of hydrothermal pyrochlore included in columbite are observed, along with other products resulting from pyrochlore alteration. Extensive research has been conducted on the transformation of pyrochlore through hydrothermal and weathering processes, with a predominant focus on carbonatite occurrences. While the original composition of pyrochlore influences the specific variety formed, the weathering-induced transformation of secondary pyrochlores tends to follow well-defined sequences at each locality [69]. At Mount Weld, pyrochlore alteration is characterized by a gradual leaching of Ca and Na, with partial replacement by varying proportions of Sr and Ce [70]. In the Nb-deposit of Catalão I, the pyrochlore is considered secondary and, with increasing weathering, it undergoes enrichment in Ba and an increase in vacancies due to the loss of Ca and Na [71]. Lumpink and Ewing [17] observed hydrothermal alteration of ‘uranpyrochlore’ during the later stages of granitic pegmatite evolution. This alteration process was characterized by a decrease in Na and F, accompanied by an increase in Ca and vacancies in the A- and Y-site, represented by the coupled substitutions ^A□^Y☐ → ^ACa^YO, ^ANa^YF → ^ACa^YO, and ^ANa^YOH → ^ACa^YO. Exchange reactions between pyrochlore and fluid indicate that this alteration occurred at ~450–650 °C and 2–4 kbar. The fluid-phase composition was characterized by relatively low a_Na+, high a_Ca2+, and high pH. In the present study, the first cations to be leached from pyrochlore were the LREE. In the second stage, the loss of LREE was accompanied by the losses of Nb and F, and by the incorporation of Fe and Si, with a great relative enrichment in U and Pb. In the third stage, losses of Pb and Fe began, while Nb and F losses continued. This resulted in the formation of various pyrochlore varieties including Pb-Fe-U-rich pyrochlore, Fe-U-rich-pyrochlore, and Fe-Mn-U-rich pyrochlore, which exhibit the highest U content (up to 13.82 wt.% UO₂). Thus, a selective release of different cations occurred throughout the alteration process. The occurrence of this mechanism is also supported by experimental studies investigating the alteration of uranium-containing pyrochlore supergroup minerals under hydrothermal conditions (T = 100–300 °C), which demonstrated incongruent dissolution behavior, with varying release rates for different elements [72,73,74].

To maintain charge balance, the increase in U concentration at Pitinga also led to an increased number of vacancies at the A-site in the pyrochlore structure. The presence of vacancies at the A-site has been attributed to selective leaching of cations during hydrothermal processes [54,75,76] and to the presence of uranium and other radioactive elements, as they can produce amorphization of the structure [77]. In the case of the analyses presented here, it is believed that both mechanisms contribute to the observed variations in the number of A-site vacancies. The preferential loss of Na and F and the corresponding increase in A-site vacancy can also be observed in the pyrochlore group minerals found in the A-type granitic rocks of the Katugin complex-ore deposit, which contains Nb, Ta, Y, REE, U, Th, Zr, and cryolite [78]. Within this deposit, three main types of pyrochlore have been identified: (i) primary magmatic pyrochlore, characterized by high concentrations of Na, REE, and F (with minor amounts of Ca, U, Th, and Pb); this type crystallized during the late magmatic stage, when the presence of Fe in the melt hindered the crystallization of columbite; (ii) secondary post-magmatic pyrochlore, which follows cracks or replaces primary pyrochlore in grain rims; it exhibits similar composition to the early phase, but with lower concentrations of Na and F and less complete occupancy of the A- and Y-sites; (iii) secondary hydrothermal pyrochlore, formed through late-stage hydrothermal alteration; this type shows a wide range of element variations and contains minor amounts of K, Ba, Pb, Fe, and U (up to 5.6 wt.%), as well as significant Si concentrations (up to 9.2 wt.%); notably, it exhibits low Na and F concentrations.

In the AEG, the incorporation of Si and Fe during a specific stage of pyrochlore alteration was found to be significant. According to Johan and Johan [54] the high concentration of U in the A-site of defective pyrochlore (A²⁺ □ B₂⁵⁺ O₆ □) could explain the presence of notable amounts of M⁴⁺ in the B-site, suggesting a hypothetical end-member composition of U⁴⁺ □ B₂⁴⁺ O₆ □. This model provides an explanation for the occurrence of Si, as well as Ti and Sn, in the B-structural site of the pyrochlore from the AEG, along with the high vacancy content in the A-site. Another compatible substitution scheme involves the exchange of 2Ca²⁺ + 2(Nb, Ta)⁵⁺ or Na⁺REE³⁺ + 2(Nb, Ta)⁵⁺ ↔ (Pb, Fe)²⁺U⁴⁺ + 2(Si, Ti)⁴⁺. This substitution scheme is supported by the positive correlation (0.76) observed between U + Pb + Fe and Si concentrations (Figure 8B). The incorporation of Fe and Si, along with Sr and Ba, from the fluids was also documented in the Miaoya complex [79], where the ultimate in situ replacement was represented by secondary ferrocolumbite, along with uraninite and Nb-bearing rutile.

A notable experiment conducted by Geisler et al. [80,81] on pyrochlore alteration yielded results inconsistent with a solid-state diffusion mechanism. Natural pyrochlore was treated in a solution containing 1M HCl and 1M CaCl₂ at 175 °C, selectively removing Ca and Na from the pyrochlore. This process resulted in a rim of depleted composition while retaining the crystal structure [80]. The rapid reaction rate at moderate temperatures, the observation of a sharp nanometer-scale reaction interface through transmission electron microscopy [81], and the incorporation of ¹⁸O from an enriched fluid into the pyrochlore structure support the notion of a pseudomorph reaction. This reaction involves the dissolution of the pyrochlore parent and simultaneous reprecipitation of a defect pyrochlore at a moving reaction interface. In fact, in several cases, the alteration of pyrochlore under hydrothermal conditions led to its recrystallization [73,74].

The alteration of pyrochlore in the AEG culminates in the breakdown of the pyrochlore structure and formation of columbite, as suggested by Minuzzi et al. [19]. The AEG Mn-Fe-rich columbite is considered a secondary pseudomorph phase. The reaction of pyrochlore with the hydrothermal fluid caused the complete or partial removal of Na, Ca, REE, Pb, U, and Si and the incorporation of Fe and Mn. The M³⁺ cations (Y, REE) are supposedly incorporated in the B-site along with anomalously high concentrations of M⁴⁺ cations (mostly U, Si, and Ti) and minor M²⁺ cations (Pb, Ca), resulting in vacancies in the A-site, as in the scheme (Fe, Mn)²⁺ + 2(Nb, Ta)⁵⁺ → □_A + 3(Si, U, Th, Ti, Sn)⁴⁺. Hydrothermal fluid reactions may result in the replacement of pyrochlore by columbite-(Fe) [82,83] following the reaction: H⁺ + Fe²⁺ + (CaNaNb₂O₆F)_(S) = FeNb₂O₆ + Ca²⁺ + Na⁺ + HF. In order to facilitate the removal of Na, Ca, and F, and the influx of Fe, these exchange reactions should take place at low pH, low Ca and Na activities, and relatively elevated activity of Fe. The replacement of ‘uranpyrochlore’ by ‘ferrocolumbite’, rather than by lueshite (NaNbO3) or fersmite (CaNb2O6) in the Miaoya carbonatite [79] also indicates a moderate to high Fe²⁺, but low Na⁺ and Ca²⁺, environment, which is corroborated by phase diagrams of pyrochlore in the system of Na-Ca-Fe-Nb-O-H [17,84].

Columbite is a typical product of the hydrothermal alteration of pyrochlore in many carbonatites, syenites, and alkali granites during the later stages of alteration. As examples, Uher et al. [85] found columbite-(Fe) forming rare irregular intergrowths with Nb-Ta-rich rutile in the Prasivá granitic pegmatites, Slovakia. Doroshkevich et al. [86] described columbite + quartz replacing pyrochlore in the Amba Dongar carbonatite complex, Gujarat, India. Columbite-(Fe) in syenogranites and related greisen from the reduced A-type Desemborque Pluton [87] are mainly associated with hydrothermal origin during the post-magmatic stage of crystallization. The columbite-1 is characterized by zoned crystals, which record two hydrothermal stages of crystallization: early Nb-rich core, and later Ta-rich rims. In contrast, columbite-2 is defined by irregular crystals with patchy textures, and its formation is related to disequilibrium processes driven by fluid-induced hydrothermal alterations involving the partial replacement of fluorite and/or cassiterite at the final post-magmatic stage. The chemical contrasts among the columbite types are related to disequilibrium crystallization processes (columbite-1) and to hydrothermal alterations during the post-magmatic evolution (columbite-2).

During fluid-assisted alteration of the AEG pyrochlore, the released compounds, either leached or the remains of coupled dissolution-reprecipitation processes, also resulted in the formation of phases with non-stoichiometric intermediate compositions between pyrochlore and columbite, suggesting the involvement of an additional process in the redistribution of uranium. The alteration of pyrochlore commonly leads to the formation of an amorphous layer and/or various crystalline phases on pyrochlore surfaces. These phases often form micro- and nanoparticles deposited on the pyrochlore surface along grain boundaries, pores, and fractures, forming secondary veins. In most cases, pyrochlore alteration leads to the development of porosity, such as micro-cracks at grain boundaries between the original material, and secondary phases, which further facilitates the migration of fluids [88,89]. In our study, we have found secondary minerals associated with columbite that likely precipitated within the opened cavities in the pyrochlore structure. These minerals exhibit intermediate compositions between U-enriched pyrochlore and U-enriched silicate, as well as intermediate compositions within the coffinite–thorite–xenotime–zircon mineral system. Additionally, galena and LREE-rich fluorides were observed. These findings support the notion that these secondary minerals incorporated leached U, Pb, and LREE from pyrochlore, as well as compounds derived from a hydrothermal fluid that was previously enriched in HFSE (Zr, Th, Y, HREE) and S. If the formation of petscheckite or its hydrated forms did occur, it was not a significant process in the AEG. The uranium incorporated in primary pyrochlore was relatively enriched in secondary pyrochlore, until the breakage of pyrochlore occurred, leading to the distribution of U throughout the secondary minerals, with preferential incorporation into Si-rich phases. The subsequent precipitation of iron oxide (hematite) shows the high Fe activity in the hydrothermal fluids during columbite formation.

Significant concentrations of U, Nb, and Ti are observed in Fe-rich veins on the surface of pyrochlore grains, indicating their mobilization and migration from the pyrochlore, and that the Fe-rich environment provides favorable conditions for the immobilization of these elements [89]. While lixiviation of U and other A-site cations from pyrochlore is driven by gradients in chemical potential [17,84], the presence of Nb and Ti in secondary phases implies the decomposition of the relatively stable B(Nb, Ta, Ti, Zr, Fe³⁺, Si)₂X(O, OH)₆ framework of metamict pyrochlore during alteration [52]. It is concluded that reduced forms of actinide species can be immobilized as AcO_{2 + x} immediately at the surface of various waste forms during alteration under reducing or mildly oxidizing conditions in geological repositories.

5.3. Typology and Importance of U Mineralization

The main U transfer mechanisms from the mantle to the crust are fractional crystallization and partial melting [7]. Due to its incompatible nature, U becomes highly enriched in magmatic fluids in the late stages of differentiation [10]. Deposits related to magmatic fractionation may occur through extreme fractional crystallization, mostly of peralkaline magmas, as well as by partial melting of U-enriched supracrustal rocks [8]. The extreme fractional crystallization of peralkaline and syenite magmas may lead to the formation of very large and low-grade U and Th resources, such as the Kvanefjeld deposit at Ilimaussaq, Greenland [90]. Other occurrences of this type are Poços de Caldas, Brazil [91]; Bokan Mountain, Alaska [92]; Lovozero Massif, Russia [93]; and the Kaffo Valley, Nigeria [94].

The U mineralization of the Pitinga mine can be classified among the Intrusive Plutonic Deposit of Peralkaline Complexes associated with magmatic differentiation processes [1]. This categorization is equivalent to the type of deposit Fractional Crystallization of Magmas [7], in which mineralization is more efficient in peralkaline fusions. When the peralkaline fusion crystallizes, U-Th-Zr-REE-Nb oxide, phosphate, and silicate complexes are formed, which are very refractory. The mineralization of U-Th-REE-Y-Zr-Nb in granitoids with associated hydrothermal processes, with greater or lesser enrichment of U, occurs in several geological contexts (Table 5), as in the Rössing Deposit [2], the Kvanefjeld deposit [95], the Bokan Mountain deposit [96], and the Ghurayyah deposit [97]. Other examples with less U mineralization are the Beauvoir albite granite in France [98,99]; the Ririwai Complex in Nigeria [100,101,102]; and the Erzgebirge Li-mica granites in Germany [103,104].

The granite-hosted Rössing Deposit (Table 5) is among the ten largest world producers of uranium [105]. The main primary uranium mineral is magmatic uraninite, and approximately 5% of the uranium reserves occur in high-Nb + Ti betafite. Secondary uranium mineralization, due to hydrothermal or surficial weathering, takes the form of uranophane, beta-uranophane, gummite, torbernite/meta-torbernite, carnotite, meta-hawaiweeite, and thorogummite [2,106,107]. The crystallization of uraninite was related to the boiling of the magma and unmixing of an H₂O-CO₂-NaCl brine. The low oxygen fugacity allowed the uranium to be present in the tetravalent state, preventing it from being lost with the solution during the boiling of the magma [107].

Table 5. Uranium mineralization in the Madeira deposit and comparison with major uranium rock-hosted deposits in the world. The table was compiled using the following sources: [2,8,21,35,92,95,96,97,106,107,108,109,110,111,112].

Deposit and Location	Tectonic Setting	Ore Age (Ma)	Deposit Type	Host Rock/Structure (other Associated Rocks)	Ore Minerals (Minerals of Potential Interest)	Economic Parameters
Madeira, Pitinga, North Brazil	Guianas Shield, Amazonas craton	1.822–1.794	Alkaline rock-hosted	Albite granite (alkali-feldspar granite and amphibole-biotite-granite)	Pyrochlore, columbite cassiterite (thorite, xenotime, cryolite)	164 Mt at 328 ppm UO2 (52 kt U) a
Rössing, Namibia	Fengcheng Mamatic Massif, Damara Orogen	510 ± 3–429 ± 17	Alkaline rock-hosted	Pegmatitic leucogranite-alaskite (biotite-amphibole gneiss, amphibole-biotite schist)	Uraninite, pitchblende, betafite, beta-uranophane, gummite (monazite, zircon, apatite)	246.500 tU at 300 ppm UO2 b
Kvanefjeld, Ilimassauq, South Greenland	Eastern Gardar intracratonic rifting Province	1.280–1.140	Alkaline rock-hosted	Nepheline syenite and lujavrite (alkali granite, pulaskite, and nauajite)	Steenstrupine, monazite, eudialyte (pyrochlore, thorite, rinkite)	673 Mt at 248 ppm U2O3 (184 kt U) c
Bokan Mountain, Southeast Alaska	Alexander terrane, western Canadian Cordillera	151 ± 5	Alkaline rock-hosted	Aegirine granite, veins and shear zones (riebeckite granite and aegirine syenite)	U-rich thorite, uraninite, U-rich thorianite, coffinite, allanite	562 kt at 0.15–0.33 wt% UO2 (635 t U) d
Ghurayyah, Hijaz region, Saudi Arabia	Northwestern Arabian Shield	620–530	Alkaline rock-hosted	Leucocratic microgranite	Uraninite (monazite, thorite, pyrochlore, columbite, cassiterite, xenotime)	440 Mt at 117 ppm UO2 (635 t U) e
Morro do Ferro, Brazil	Poços de Caldas plateau	83–64	Carbonatite-hosted	Lateritic profile (magnetite dyke and syenitic rocks)	Uranothorite (fluorcarbonates)	100 t U; 110–120 ppm UO2 f

^a [18]; ^b [1]; ^c U total resources and grade from Energy Transition Minerals Ltd. [113], cut-off at 150 ppm U₃O₈; ^d Potential resources and grades from the United States Geological Survey-USGS [114]; ^e Preliminary estimates of tonnage and grade from Drysdall et al. [97]; ^f Grades and tonnage after Gentile and Figueiredo Filho [115].

Table 5. Uranium mineralization in the Madeira deposit and comparison with major uranium rock-hosted deposits in the world. The table was compiled using the following sources: [2,8,21,35,92,95,96,97,106,107,108,109,110,111,112].

Deposit and Location	Tectonic Setting	Ore Age (Ma)	Deposit Type	Host Rock/Structure (other Associated Rocks)	Ore Minerals (Minerals of Potential Interest)	Economic Parameters
Madeira, Pitinga, North Brazil	Guianas Shield, Amazonas craton	1.822–1.794	Alkaline rock-hosted	Albite granite (alkali-feldspar granite and amphibole-biotite-granite)	Pyrochlore, columbite cassiterite (thorite, xenotime, cryolite)	164 Mt at 328 ppm UO2 (52 kt U) a
Rössing, Namibia	Fengcheng Mamatic Massif, Damara Orogen	510 ± 3–429 ± 17	Alkaline rock-hosted	Pegmatitic leucogranite-alaskite (biotite-amphibole gneiss, amphibole-biotite schist)	Uraninite, pitchblende, betafite, beta-uranophane, gummite (monazite, zircon, apatite)	246.500 tU at 300 ppm UO2 b
Kvanefjeld, Ilimassauq, South Greenland	Eastern Gardar intracratonic rifting Province	1.280–1.140	Alkaline rock-hosted	Nepheline syenite and lujavrite (alkali granite, pulaskite, and nauajite)	Steenstrupine, monazite, eudialyte (pyrochlore, thorite, rinkite)	673 Mt at 248 ppm U2O3 (184 kt U) c
Bokan Mountain, Southeast Alaska	Alexander terrane, western Canadian Cordillera	151 ± 5	Alkaline rock-hosted	Aegirine granite, veins and shear zones (riebeckite granite and aegirine syenite)	U-rich thorite, uraninite, U-rich thorianite, coffinite, allanite	562 kt at 0.15–0.33 wt% UO2 (635 t U) d
Ghurayyah, Hijaz region, Saudi Arabia	Northwestern Arabian Shield	620–530	Alkaline rock-hosted	Leucocratic microgranite	Uraninite (monazite, thorite, pyrochlore, columbite, cassiterite, xenotime)	440 Mt at 117 ppm UO2 (635 t U) e
Morro do Ferro, Brazil	Poços de Caldas plateau	83–64	Carbonatite-hosted	Lateritic profile (magnetite dyke and syenitic rocks)	Uranothorite (fluorcarbonates)	100 t U; 110–120 ppm UO2 f

^a [18]; ^b [1]; ^c U total resources and grade from Energy Transition Minerals Ltd. [113], cut-off at 150 ppm U₃O₈; ^d Potential resources and grades from the United States Geological Survey-USGS [114]; ^e Preliminary estimates of tonnage and grade from Drysdall et al. [97]; ^f Grades and tonnage after Gentile and Figueiredo Filho [115].

The Kvanefjeld deposit (Table 5) is associated with the nepheline syenite and with the highly differentiated lujavrite enriched in U, as well as in Nb, Th, Zr, Be, Li, F, and REE. The highest U concentrations occur in the upper part and at the contact of the lujavrite with enclosing altered volcanic rocks. The main U-Th minerals are eudialyte, rinkite, monazite, lovozerite, steenstrupine, thorite, and pyrochlore [95]. In the Bokan Mountain [1], the U-Th mineralization is associated with a desilicified and albitized part of the pluton, forming plunging pipe-like bodies along the contact with the aegirine granite or occurring as pods in an echelon NW-striking shear zone. The main ore minerals are U-rich thorite, uraninite, and U-rich thorianite with sulfides disseminated in nearly pure albite.

In the Ririwai Kaffo Valey ring complex, Nigeria [101,102,116], the albite-enriched granite is identical to that from Pitinga (except for the absence of a massive cryolite deposit). However, the U-Th mineralization is metasomatic, hosted in a biotite granite that has undergone extensive post-magmatic metasomatism to produce an albitized, microclinized, and greisenized rock that carries late coffinite, thorite, and xenotime. In the Ghurayyah deposit [97], Saudi Arabia, the rare metal mineralization is disseminated in peralkaline microgranites, and the main ore minerals are uraninite, thorite, monazite, pyrochlore, samarskite, aeschinite, cassiterite, columbite–tantalite, and xenotime [109].

The Madeira deposit with 164 Mt at 328 ppm UO₂ (52 kU) is comparable in grades and reserves to the deposits above (Table 5). However, this deposit is in stark contrast to those deposits in five aspects: (1) the uranium mineralization is homogeneously dispersed in the AEG (with a grade of 328 ppm UO₂); (2) the whole U-Th paragenesis is simple; (3) there is only one primary U ore mineral (U-Pb-LREE-enriched pyrochlore); (4) the U and Th mineralizations are divided into different minerals formed in distinct stages of magma evolution (early U-Pb-LREE-enriched pyrochlore and late thorite); and (5) both mineralizations were affected by intense hydrothermal alterations related to F-rich hydrothermal fluids. As discussed above, these characteristics are related to special conditions imposed by the fluorine-rich fluids on the evolution of the magma and, consequently, on the evolution of the Th/U ratio and the paragenesis.

Despite the homogeneous distribution of the primary ore mineral, the U mineralization exhibits zonation on the deposit scale related to the degree of hydrothermal alteration of the pyrochlore. The alteration is more intense in the AGB and in the central zone of the AGC (closer to the massive cryolite deposit). In these regions, the more common pyrochlore varieties are Pb-Fe-U-rich pyrochlore, Fe-U-rich-pyrochlore, and Fe-Mn-U-rich pyrochlore (the richest in U, up to 13.82 wt.% UO₂). Uranium-enriched varieties of columbite are also more abundant in these areas, along with intermediate compositions in the coffinite–thorite–xenotime–zircon mineral system and galena. In the other parts of the AGC, Fe-U-Pb-rich pyrochlore predominates. In both the AGC and AGB, the Mn-Fe-columbite is the more common species, and secondary LREE-rich fluorides are also present.

6. Conclusions

The U mineralization in the Madeira Sn-Nb-Ta (U, Th, REE, cryolite) world-class deposit is disseminated in the AEG facies (AGC+ AGB), as well as in pegmatite veins within the AGC, where pyrochlore is inherited. The primary ore mineral of U is exclusively early magmatic U-Pb-LREE-enriched pyrochlore. The U mineralization is homogeneously dispersed due to transportation of fluorine-bearing complexes that carried HFS elements throughout the melt, and the buffering of F content in the magma prevented the formation of zones with higher enrichment. The peralkaline magma, F richness, and the low magma temperature conditioned the crystallization of pyrochlore instead of columbite. The enrichment of U in pyrochlore in the magma resulted in the loss of Ca and F. In the late stage of magmatic evolution, as zircon crystallization became more intense and accompanied by xenotime and thorite, the magma was previously depleted in U, Nb, Ta, and LREE. Therefore, the U and Th mineralizations in the Madeira deposit were formed at different stages, associated with different primary minerals, and have distinct secondary minerals.

All pyrochlore crystals in AGC and AGB underwent hydrothermal alteration caused by F-rich aqueous fluids. The alteration process selectively released different cations, leading to the successive formation of various secondary pyrochlore varieties and the relative enrichment of U. The Fe-Mn-U-rich pyrochlore, which contains the highest U content, can reach up to 13.82 wt.% UO₂. The alteration of pyrochlore culminates in the breakdown of the pyrochlore structure and the formation of U-bearing columbite. The most intense alteration occurs in the central part of the AGC, close to the massive cryolite deposit, and in the AGB, where the secondary pyrochlores richer in U and the U-bearing columbite are more abundant.

The U mineralization in the Madeira deposit is classified as an intrusive deposit type according to IAEA [12]. It exhibits grades (328 ppm UO₂) comparable to the main deposits of this type and significant reserves (52 kt U). However, it is in stark contrast to those deposits in four key aspects: homogeneous dispersion of mineralization; pyrochlore as the exclusive primary ore mineral; U and Th mineralizations formed at different stages; and being affected by intense hydrothermal alterations. These characteristics are attributed to the special conditions imposed by the fluorine-rich nature of the peralkaline magma.