Volume 108, Issue 2

When laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) analyses of pyrite plot below the gold solubility line on a gold vs. arsenic plot and have relatively flat counts on laser ablation time-resolved output graphs, it is often interpreted that the gold is held within the pyrite structure. The study by Ehrig et al. (2023, this issue) shows, using a combination of LA-ICP-MS spot analyses of gold in pyrite, transmission electron microscopy, and electron backscatter diffraction that this is not necessarily the case. Furthermore, they use these same techniques to identify how trace elements, including gold, are remobilized in pyrite during deformation and metamorphism.

Textural and compositional variations in titanite constrain the roles of magma mixing and hydrothermal alteration in two plutons in central Utah: the Jurassic Notch Peak and the Oligocene Little Cottonwood stocks. In the Notch Peak intrusion, magmatic titanite grains usually have oscillatory zones combined with BSE-bright sector zones, in some cases surrounding simple unzoned cores. These grains are frequently overprinted by hydrothermal titanite with low concentrations of high field strength elements (HFSE). Magmatic titanite has an average δ 18 O of 6.0‰ and post-magmatic titanite is 6.2‰, as analyzed by SIMS. Average Zr-in-titanite temperatures are also similar, with 718 °C for magmatic and 711 °C for hydrothermal titanite. These observations indicate simple magmatic growth, followed by hydrothermal alteration by magmatic fluids. Titanite in aplite dikes and sills has lower concentrations of all trace elements except F. Many titanite grains in the aplites have late overgrowths of high-Fe titanite. This high-Fe titanite has δ 18 O of 6‰ and an average Zr-in-titanite temperature of 718 °C and likely precipitated from a last flush of exsolved magmatic water enriched in Cl and Fe. Titanite in the Little Cottonwood stock typically has distinct patchy cores with rounded and embayed ilmenite inclusions. Mafic enclaves have abundant titanite that is similar in texture and δ 18 O (5.1‰) to titanite in the host (δ 18 O = 4.9‰), but it has a slightly higher average Zr-in-titanite temperature (731 vs. 717 °C). The patchy cores in the enclaves have the highest average Zr-in-titanite temperature (759 °C) and distinctive REE patterns. The textural and compositional data indicate that a hotter, more reduced, ilmenite-bearing mafic magma mixed into an oxidized felsic magma, destabilizing existing ilmenite and allowing crystallization of titanite. In the granodiorite and in the enclaves, hydrothermal growth of titanite is evidenced by distinct narrow rims as well as anhedral titanite that grew between sheets of chloritized biotite. Secondary hydrothermal titanite typically has lower concentrations of most HFSE, but is relatively enriched in F, Mg, Mo, and U, and it has higher Nb/Ta and lower Th/U ratios. Post-magmatic titanite also has strikingly different REE patterns than magmatic titanite, including the absence of pronounced Eu anomalies and lower REE abundances. These chemical features are controlled by element solubilities in aqueous fluids. In most cases, hydrothermal titanite has δ 18 O values similar to magmatic titanite, indicating alteration and recrystallization from exsolved magmatic fluids. The involvement of meteoric water with low δ 18 O is evident locally; individual spots have δ 18 O as low as 1.7‰ in the Little Cottonwood stock. Titanite compositions and textures provide important insights into the origins of granitic rocks and can be used to distinguish separate batches of magma, gauge the evolution of magmatic rocks, assess mixing processes, and infer compositions of mixing components. Because titanite also forms hydrothermally, it retains hints about the composition, temperature, and oxygen fugacity of the hydrothermal fluids and reveals details about titanite-forming reactions. However, the Al-in-titanite geobarometer does not yield realistic pressures of crystallization and the use of titanite as a geochronometer is compromised by the development of U-rich hydrothermal titanite.

A new mineral, gismondine-Sr with ordered gismondine framework type [ B 22 1 2 no. 20, Z = 1; a = 14.0256(2) Å, b = 10.45900(10) Å, c = 13.79360(10) Å, V = 2023.44(4) Å 3 ] and the ideal chemical formula Sr 4 (Si 8 Al 8 O 32 )·9H 2 O was discovered in amygdaloidal voids of partly melted gehlenite hornfels at Halamish locality, Hatrurim Basin of the Hatrurim Complex, Negev Desert, Israel. Gehlenite hornfels is mainly composed of gehlenite, wollastonite, and garnet of the grossular-andradite-schorlomite series. In a low-temperature association occur minerals such as thomsonite-Ca, flörkeite, analcime and minerals of the tobermorite supergroup. Gismondine-Sr forms spherulitic aggregates up to 180 μm and, rarely, pseudotetragonal bipyramidal crystals up to 50 μm. Empirical crystal-chemical formula of gismondine-Sr is (Sr 2.02 Ca 1.09 Ba 0.02 K 0.72 Na 0.62 ) Σ4.47 Al 7.91 Si 8.09 O 31.85 ·9H 2 O. It is the strontium analog of gismondine-Ca and the second orthorhombic zeolite with the GIS structure topology. Crystals are transparent to translucent and feature vitreous luster. The mineral exhibits a white color, imperfect cleavage in [101] direction, a brittle tenacity, and uneven fracture. The Mohs hardness was estimated at approximately 4. Gismondine-Sr is biaxial negative, α = 1.488(3), β = 1.492(3), γ = 1.495(3), 2 V obs = 70–80°. The Raman spectrum is characterized by a band at 465 cm −1 , which is also the main band in gismondine-Ca. The structure refinement using SC-XRD ( R 1 = 0.0353) reveals the ordered distribution of framework cations and the disordered arrangement of extraframework cations. The aluminosilicate framework is built by crankshaft chains with 8-membered apertures channels parallel to [101] and [101]. In gismondine-Sr, the 8-membered rings are elliptically deformed and the T-O-T angle of the upward and downward tetrahedra in the double crankshaft chains is smaller compared to that for gismondine-Ca. Consequently, a slight rotation of the double crankshaft chains has been noticed. Similar observations have been made in partially dehydrated and the pressure-modified gismondine-Ca. The present study suggests that, in addition to high-pressure and dehydration, the elliptical deformation of the channels in GIS also arises as a consequence of the extraframework cations and H 2 O content. Thus, the extraframework content influences the aluminosilicate framework leading to the orthorhombic symmetry.

“Invisible gold” refers to gold (Au) occurring either within the lattice of a host sulfide or as discrete nanoparticles (NPs, <100 nm diameter) within a host that are only observable when imaged at very high magnifications. Previous research has regarded the physical form of invisible gold to be partially controlled by the concentration of arsenic (As) in the host sulfide, with stability fields for lattice-bound vs. Au-NPs defined by an empirical Au-As solubility curve. We undertook micrometer- and nanoscale analysis of a representative sample of As-Co-Ni-(Au)-bearing pyrite from Cu-mineralized breccias in the deeper part of the Olympic Dam Cu-U-Au-Ag deposit (South Australia) to define the location and physical form of Au and accompanying elements. Trace element geochemistry and statistical analysis show that >50% of pyrites contain measurable Au and As, and plot below the Au-As solubility curve. Au and As are geochemically associated with Te, Bi, Pb, Ag, and Sn. Primary oscillatory zoning patterns in pyrite defined by As-Co-Ni are reshaped by processes of dissolution-reprecipitation, including new nanoscale growth and rhythmical misorientation structures. Low-angle slip dislocations, twist-wall boundaries and deformation-dipole nanostructures are associated with Te-Bi-Pb-enrichment and host Au-Ag-telluride nanoparticles (NPs). Electrum NPs occur associated with pores coated by Bi-Ag-tellurides or within chalcopyrite particles. Bi-Pb-sulfotellurides, petzite, and sylvanite were identified by atomic-scale scanning transmission electron microscopy. The data support trace element (re)mobilization during pyrite deformation at the brittle to ductile transition (0.5–1 kbar, 300–400 °C) during brecciation. Au-NP formation is decoupled from initial As incorporation in pyrite and instead fingerprints formation of strain-induced, chalcogen-enriched nanoscale structures. Pore-attached NPs suggest scavenging of Au by Bi-bearing melts with higher rates of fluid percolation. Similar scenarios are predictable for pyrite-hosted “invisible Au” in pyrite from other deposits that experienced multiple overprints. Unveiling the cloak of invisibility using contemporary micro- to nano-analytical techniques reveals new layers of complexity with respect to the trace/minor element incorporation in mineral matrices and their subsequent release during overprinting.

Nb-Ta-oxide minerals (NTO) are commonly associated with rare-element pegmatites where they are interpreted to precipitate at magmatic to magmatic-hydrothermal conditions. Although high-temperature experiments show that phosphorus and other fluxing elements (e.g., Li, B, F) can affect the saturation of NTO in pegmatitic systems, it is still uncertain how NTO saturation occurs in natural, flux-rich pegmatitic melts and whether crystallization occurs at multiple stages during magmatic or subsolidus conditions. The lithium-cesium-tantalum (LCT) family P-rich Buranga granitic pegmatite (western Rwanda) is used as a type locality to address this question. NTO mineralization in the Buranga dike occurs in two mineralogical units: in mineralogically complex phosphatic nodules, and in albitized parts. In the phosphatic nodules, Fe-Nb-Ta-rich rutile and columbite-group minerals (CGM) are observed, while in the albitized parts, only CGM is found. Fe-Nb-Ta-rich rutile precipitates at the magmatic stage along with early primary phosphates (i.e., F-rich montebrasite, wyllieite, and fluorapatite). Conversely, CGM mineralization occurs at the magmatic-hydrothermal stage in association with replacement phosphates like bertossaite, after primary minerals (i.e., rutile, wyllieite, rosemaryite, and trolleite) are destabilized due to crystal-melt-fluid interactions. NTO textures and chemical zoning show uneven evolution from core to rim and are related to localized alteration phenomena of the surrounding minerals. This indicates that local processes and element transfers are more important than dike-scale fractionation processes for NTO mineralization in P-rich granitic pegmatites. The restricted availability of Fe and Mn in the system, which is related to the competition between phosphate and oxide minerals, is identified as the main control on the CGM mineralization in the Buranga dike. CGM precipitation is only possible during the magmatic-hydrothermal stage when Fe and Mn are leached from the primary phosphates and remobilized to the Nb-Ta-bearing residual melt by the exsolved fluid.

Interpretation of chemical zoning within igneous minerals is critical to many petrologic studies. Zoning in minerals, however, is commonly observed in thin sections or grain mounts, which are random 2D slices of a 3D system. Use of these 2D sections to infer 3D geometries requires a set of assumptions, often not directly tested, introduces several issues, and results in partial loss of zoning information. Computed X‑ray microtomography (microCT) offers a way to assess 3D zoning in minerals at high resolution. To observe 3D mineral zoning using microCT, however, requires that zoning is observable as differences in X‑ray attenuation. Sanidine, with its affinity for Ba in the crystal lattice, can display large, abrupt variations in Ba that are related to various magma reservoir processes. These changes in Ba also significantly change the X‑ray attenuation coefficient of sanidine, allowing for discrete mineral zones to be mapped in 3D using microCT. Here we utilize microCT to show 3D chemical zoning within natural sanidines from a suite of volcanic eruptions throughout the geologic record. We also show that changes in microCT grayscale in sanidine are largely controlled by changes in Ba. Starting with 3D mineral reconstructions, we simulate thin-section making by generating random 2D slices across a mineral zone to show that slicing orientation alone can drastically change the apparent width and slope of composition transitions between different zones. Furthermore, we find that chemical zoning in sanidine can commonly occur in more complex geometries than the commonly interpreted concentric zoning patterns. Together, these findings have important implications for methodologies that rely on the interpretation of chemical zoning within minerals and align with previously published numerical models that show how chemical gradient geometries are affected by random sectioning during common sample preparation methods (e.g., thin sections and round mounts).

Siderite (FeCO 3 ) and rhodochrosite (MnCO 3 ) are two interesting carbonate minerals, which normally occur in hydrothermal deposits on deep-sea altered oceanic crust. Despite the ubiquity of carbonates in the slab, little is known of the physicochemical behavior of siderite and rhodochrosite at high-pressure ( P ) and high-temperature ( T ) conditions during slab subduction. In this study, we characterized the Raman vibrational spectra of natural siderite and rhodochrosite up to 47 GPa and 1100 K in an externally heated diamond-anvil cell (DAC). Experimental results show that the Raman frequency shifts (ν i ) for siderite and rhodochrosite are a function of both P and T , and the effect of the P - T cross derivative term cannot be neglected, especially at high- P and high- T conditions. Based on the functional relationship of ν i - P - T , the P - T calibrants of siderite and rhodochrosite are developed, respectively. This is significant for studying the water-carbonate interaction at high P - T conditions in a DAC because the undesired change of the experimental system from traditional pressure sensors (e.g., ruby, quartz) in a reaction chamber can be avoided. Like previous studies, we observed a sharp spin transition at ~45 GPa in siderite and a phase transition from MnCO 3 -I to MnCO 3 -II at ~46 GPa for rhodochrosite at room temperature. Furthermore, we determined the isobaric and isothermal equivalents of the mode Grüneisen parameter (γ iT , γ iP ) and the anharmonic parameter ( a i ) for each Raman mode of siderite and rhodochrosite. The δν i /δ P , δν i /δ T , γ iT , γ iP , and a i span a much larger value range for the external lattice modes (T, L) than internal modes (ν 4 , ν 1 ) in both siderite and rhodochrosite. Combining Raman frequency shifts and the first-order Murnaghan equation of state, we also developed a method to calculate the temperature dependence of the bulk modulus (K T ) for siderite and rhodochrosite, respectively.

An important mechanism of carbonate mineral growth is dissolution-reprecipitation, including the transformation of amorphous precursor to crystalline carbonates, and coarsening (ripening) of fine carbonate crystals. However, the mechanistic details of cation exchange associated with carbonate mineral growth via a dissolution-reprecipitation process are still not well understood. In this study, we used Mg isotopes to probe the exchange of Mg between aqueous solutions and norsethite [BaMg(CO 3 ) 2 ] by systematic synthesis experiments. Norsethite is a model double carbonate, with a general formula of AB(CO 3 ) 2 , where A and B stand for two different divalent ions. Formation of norsethite is comprised of three stages, including: (1) precipitation of barium-magnesium (Ba-Mg) amorphous carbonate; (2) transformation of Ba-Mg amorphous carbonate to nano-crystalline norsethite by fast dissolution-reprecipitation; and (3) coarsening (ripening) of nano-norsethite by slow dissolution-reprecipitation. Magnesium isotopes displayed distinct fractionation behaviors in each of the three stages. The Mg isotope fractionation factors (Δ 26 Mg solid-aq ) associated with precipitation of Ba-Mg amorphous carbonate were slightly negative and temperature-dependent, from –0.83‰ at 30 °C to –0.53‰ at 70 °C. During the transformation of Ba-Mg amorphous carbonate to nano-crystalline norsethite, isotopically light Mg isotopes were further enriched in the solid phase, with apparent Δ 26 Mg solid-aq decreasing to –2.12‰ at 30 °C and –1.56‰ at 70 °C. In the ripening stage, norsethite became isotopically heavier, with Δ 26 Mg solid-aq increasing up to –1.95‰ at 30 °C and –1.17‰ at 70 °C. The experimental results show that non-equilibrium isotope fractionation occurred during the transformation of amorphous carbonate to nano-crystalline norsethite (i.e., fast dissolution-reprecipitation). By contrast, the subsequent ripening of the norsethite led to the evolution toward isotopic equilibrium of the system by slower exchange with a longer reaction time (i.e., slow dissolution-reprecipitation). Moreover, our first-principles calculation results indicate that the equilibrium isotope fractionation was approached, but not attained, even after 276 days of recrystallization at temperatures below 70 °C. In short, this study has identified two different types of dissolution-reprecipitation process during the carbonate mineral growth and highlights the importance of understanding formation mechanism and post-depositional history of carbonate in interpreting the isotopic data of carbonate minerals.

The Furong tin deposit (South China) is genetically associated with the multiphase Qitianling batholith that consists of main-phase and minor, but more fractionated, late-phase granites. Several tourmaline and apatite generations are distinguished. Tourmaline (Tur) variants comprise pre-ore Tur-1 as disseminations and nodules in the late-phase granite, pre- to syn-ore Tur-2 as replacements in nodules and as veins crosscutting the late-phase granite and nodules, syn-ore Tur-3 in tin greisens, pre- to syn-ore Tur-4 as veins in the altered main-phase granite, and syn-ore Tur-5 from tin skarns in a distinct Ca-rich environment. Apatite (Ap) generations include accessory Ap-G in the main-phase granite, and Ap-I to Ap-III from three stages related to skarn-type mineralization (garnet-diopside stage-I, pargasite-phlogopite-cassiterite stage-II, and sulfide-rich stage-III). Textural and compositional features suggest that all tourmaline variants are hydrothermal in origin with alkali and schorl to foitite composition and minor extensions to calcic and X-site vacant tourmaline groups, whereas all apatite generations belong to fluorapatite with Ap-G crystallizing from the magma and Ap-I to Ap-III being hydrothermal in origin. The narrow range of tourmaline δ 11 B values (–14.8 to –10.4‰) suggests a single magmatic boron source in the ore-forming fluids. The similar rare earth element patterns and ε Nd(t) values (–8.2 to –5.9 for Ap-G and –8.0 to –7.3 for Ap-I) between magmatic and hydrothermal apatite indicate that the skarn-forming fluids are dominantly derived from granites. The 87 Sr/ 86 Sr ratios of Ap-I to Ap-III (0.70733–0.70795) are similar to the carbonate wall rocks, but distinctly diferent from the more radiogenic granites, indicating Sr exchange with carbonate rocks. Integrating previous H-O isotopic data, the tourmaline and apatite elemental and B-Sr-Nd isotope results suggest that the greisen-type ore formed by interaction of B-, Na-, Li-, Zn-, and Sn-rich magmatic fluids with the late-phase granite in a closed and reduced feldspar-destructive environment, whereas the tin skarns resulted from mixing of magmatic fluids with meteoric water and interaction with the carbonate wall rocks in an open system where oxygen fugacity changed from reduced to oxidized conditions. During fluid-rock interactions and fluid mixing, considerable Ca, Mg, V, Ni, and Sr from the host rocks were introduced into the ore system. Coupled hydrothermal minerals such as tourmaline and apatite have great potential to fingerprint the nature, source, and evolution of fluids in granite-related ore systems.

The crystal structure and chemical composition of boussingaultite from Pécs-Vasas, Mecsek Mountains, South Hungary, were investigated by single-crystal neutron diffraction (at 20 K) along with a series of chemical analytical techniques [i.e., gravimetric determination of sulfates, EDTA titrimetric determination of magnesium, ion selective electrode for F and Cl, indirect gravimetric determination of ammonium as (NH 4 ,Rb,Cs,K) tetraphenylborate, inductively coupled plasma atomic emission spectroscopy for REE and other minor elements, elemental analysis for C, N, and H content, high- T mass loss for H 2 O content]. The concentrations of more than 50 elements were measured. The experimental formula of the boussingaultite is: [(NH 4 ) 1.77 K 0.22 ) Σ1.99 [(Mg 0.95 Mn 0.06 ) Σ1.01 (H 2 O) 5.7 ](SO 4 ) 1.99 . Neutron data analysis confirms that the structure of boussingaultite is built up by isolated Mg(H 2 O) 6 - octahedra, along with isolated NH 4 - and SO 4 -tetrahedra connected by a complex H-bonds network. Mg 2+ is completely solvated by H 2 O molecules in a typical octahedral bonding configuration. All the seven independent oxygen sites in the structure are involved in H-bonds, as donors or as acceptors. The geometry of all the H 2 O molecules, bonded to Mg, is in line with that usually observed in crystalline compounds. The H 2 O molecules show moderate-strong H-bonds, with H···O acceptor and O donor ···O acceptor ranging between 1.72–1.87 and 2.70–2.84 Å, respectively, along with O donor -H···O acceptor angles between 168–178°. The four independent N-H···O bonds show H···O acceptor and N donor ···O acceptor distances ranging between 1.81–2.00 and 2.84–2.98 Å, respectively, with N-H···O angles between 158–176°. All the H-bonds of the H 2 O molecules and of the NH 4 -group involve the oxygen sites of the SO 4 -group as acceptors: the SO 4 -group is, therefore, the “bridging unit” between the NH 4 and the Mg(H 2 O) 6 units, via H-bonds. Our structure refinement proved, unambiguously, that the partial K4 + vs. NH + replacement generates a local disorder. K lies at the N site, and its bonding configuration can be described by a distorted polyhedron with CN = 8. However, the K + vs. NH + replacement implies 4 a change in the configuration of the SO 4 - tetrahedron, through a sort of rotation of the polyhedron. This is the first evidence of the presence of a partial picromerite component in the boussingaultite structure, which gives rise to a local disorder likely due to the significantly different bonding configurations of the two cations. Our refinement also revealed that Mn 2+ replaces Mg 2+ at the Mg site. No evidence of distortion of the octahedron is observed in response to such a replacement, but the fraction of Mn 2+ is modest. An analysis of previous Raman and IR results is provided, and is compared with the experimental results of this study.

Zinc clays are commonly found in oxidized Zn deposits and, even though they rarely represent the main target of the ore exploitation, they can be used as a proxy to restore the genetic conditions during ore-forming processes. This work sheds light on the micro- to nano-mineralogy and on the genesis of Zn-clays in the Kihabe and Nxuu prospects (located in the Aha Hills district, Northern Botswana), through an integrated XRD and TEM study of the mineralized facies occurring in the ore system. The Kihabe and Nxuu ores are hosted in a Neoproterozoic metamorphozed quartzwacke unconformably covered by the recent sedimentary rocks, also containing calcretes, of the Kalahari Group. In the analyzed samples, four distinct mineralogical facies have been recognized: (1) vanadate-calcrete facies, poor of Zn-clays; (2) low Zn-clay facies, characterized mostly by clays showing low Zn concentrations; (3) Zn-clay facies, containing proper Zn clay minerals; and (4) sulfide facies, devoid of Zn-clays. In all the facies detrital dioctahedral mica (muscovite and illite) is interstratified with smectite in the form of random (R0) to short-range ordered (R1) I/S, which locally shows significant Zn concentrations. In the sulfide facies kaolinite overgrowing onto mica packets has been detected. The low Zn-clay facies is dominated by Zn-bearing beidellite, with minor kaolinite and fraipontite. The Zn-clay facies consists mostly of a random (R0) interstratified clay between a 7 Å phase corresponding to fraipontite, and a 2:1 swelling clay component identifiable with the dioctahedral smectite, with minor sauconite. The micro- to nano-scale paragenetic study performed by TEM indicates that the above-mentioned clays formed through a multistage process, eventually ending with the genesis of Zn-bearing phyllosilicates assemblages. The main steps were: (1) alteration of detrital mica and dissolution of feldspar clasts, which led to the formation of epitaxial kaolinite and replacive beidellite; (2) fertilization of barren clays and formation of replacive to epitaxial fraipontite/smectite and of Zn-bearing mica, through input of Zn 2+ deriving from sphalerite or willemite dissolution by mixed meteoric-hydrothermal fluids; (3) formation of low tetrahedral charge sauconite, either in pores or as replacement of K-feldspars under surficial hypersaline conditions, possibly also linked to the establishment of the arid climate in region. These processes provide insights into genesis of Zn-Pb-V ore deposits in northwest Botswana. Furthermore, the identification of a Zn-smectite species having a stevensite-like stoichiometry is valuable for future studies dealing with the systematics of clay minerals.

Finchite (IMA2017-052), Sr(UO 2 ) 2 (V 2 O 8 )·5H 2 O, is the first uranium mineral known to contain essential Sr. The new mineral occurs as yellow-green blades up to ~10 μm in length in surface outcrops of the calcrete-type uranium deposit at Sulfur Springs Draw, Martin County, Texas, U.S.A. Crystals of finchite were subsequently discovered underground in the Pandora mine, La Sal, San Juan County, Utah, U.S.A., as diamond-shaped golden-yellow crystals reaching up to 1 mm. The crystal structure of finchite from both localities was determined using single-crystal X‑ray diffraction and is orthorhombic, Pcan, with a = 10.363(6) Å, b = 8.498(5) Å, c = 16.250(9) Å, V = 1431.0(13) Å 3 , Z = 4 (R 1 = 0.0555) from Sulfur Springs Draw; and a = 10.3898(16), b = 8.5326(14), c = 16.3765(3) Å, V = 1451.8(4) Å 3 , Z = 4 (R 1 = 0.0600) from the Pandora mine. Electron-probe microanalysis provided the empirical formula (Sr 0.88 K 0.17 Ca 0.10 Mg 0.07 Al 0.03 Fe 0.02 ) Σ1.20 (UO 2 ) 2 (V 2.08 O 8 )·5H 2 O for crystals from Sulfur Springs Draw, and (Sr 0.50 Ca 0.28 Ba 0.22 K 0.05 ) Σ0.94 (U 0.99 O 2 ) 2 (V 2.01 O 8 )·5H 2 O for crystals from the Pandora mine, based on 17 O atoms per formula unit. The structure of finchite contains uranyl vanadate sheets based upon the francevillite topology. Finchite is a possible immobilization species for both uranium and the dangerous radionuclide 90 Sr because of the relative insolubility of uranyl vanadate minerals in water.

Magmatic and metasomatic zircon occurs in many alkaline igneous rocks and both are potential economic reservoirs of Zr, and in some places, rare-earth elements. The Baerzhe deposit in China is an example of a system where both types of zircon occur. Previous studies recognized deuteric and variably altered magmatic zircon in a transsolvus miaskitic granite, as well as four types of metasomatic zircon in a transsolvus agpaitic granite. In this study, the relationships among, and origins of, zircon and how these relate to models for rare-metal mineralization are assessed. In situ backscattered electron (BSE) and cathodoluminescence (CL) imaging, Raman spectroscopy (including mapping), and chemistry of zircon from the agpaitic granite were conducted, combined with evaluation of published data on zircon from Baerzhe. Their textural, spectroscopic, and chemical characteristics suggest that the four types of metasomatic zircon in the agpaitic granite were not subjected to metamictization or intense alteration, with trace-element accommodation largely following a xenotime substitution mechanism. The most abundant type of metasomatic zircon in the agpaitic granite occurs in zircon-quartz pseudomorphs and exhibits comparable CL, Raman spectral, and chemical features to rare zircon that has partially replaced elpidite. This confirms that the pseudomorphs formed by complete replacement of elpidite. The pseudomorph zircon occurs in association with snowball quartz that contains inclusions of zircon, aegirine, and albite, and with secondary quartz containing aegirine. This is consistent with their coeval formation during Na metasomatism. The restriction of Na metasomatism to the agpaitic granite indicates that this event and the associated zircon formation resulted from early autometasomatism of the agpaitic phase. REE- and Be-rich zircon that replaced magmatic amphibole crystallized as a result of reaction with a REE- and Be-rich fluid that most likely was responsible for the later REE-Nb-Be mineralization that affected both the miaskitic and agpaitic granites. The miaskitic granite contains deuteric and altered magmatic zircon with diferent chemical characteristics to the four types of metasomatic zircon in the agpaitic granite. This suggests that secondary Zr mineralization in the miaskitic granite formed from diferent fluids to those that metasomatized the agpaitic granite and may also have resulted from autometasomatism. This study reveals a complex picture for the formation of zircon at Baerzhe, the character of which can vary significantly, both temporally and spatially. Such variable chemistry of the various types of zircon resulted not only from their diferent origins (magmatic vs. metasomatic), but also from localized water-rock interaction that involved multiple stages of fluids. Zircon in both the miaskitic and agpaitic phases was mainly the product of autometasomatism that was constrained to their parental granites.