Volume 109, Issue 1

This paper summarizes an approach to crystal growth that was published in parts in various articles over the course of 25 years by the present author and his colleagues. Evidence for this approach, which is confirmed in detail by data in the cited publications and in the figures and equations in the supplementary material that accompanies this paper (see the Online Materials Deposit item AM-24-18851. Online Materials are free to all readers. Go online, via the table of contents or article view, and find the tab or link for supplemental materials. ), comes mainly from the shapes of crystal size distributions (CSDs). Such distributions reveal the growth histories of natural minerals and synthetic compounds, histories that can be used to make geological interpretations and to guide industrial syntheses. CSDs have three fundamental shapes: log-normal, asymptotic, and Ostwald. These shapes result from different degrees of supersaturation near the time of nucleation. The first two distribution shapes form according to the Law of Proportionate Efect (LPE) at moderate supersaturation, and the latter rare distribution forms by Ostwald ripening at large supersaturation. Initially, the first two distributions have mean diameters of up to tens of nanometers and grow by surface-limited growth kinetics. The slow step in this reaction is the incorporation of nanoparticles (bits of crystal or adparticles) onto the crystal surface. As the crystals become larger, their demand for nutrients, as calculated by the LPE, increases exponentially. Then the slow step in the reaction changes to the rate of transfer of nutrients to the crystal (supply- or transport-limited growth). Crystal diameters often grow the most during this latter stage, and the initial CSD shapes that originally formed during surface-limited growth are retained and scaled up proportionately. Proportionate growth during the supply-limited stage can be simulated approximately by multiplying the diameter of each crystal in a distribution by a constant. Crystals can also grow by a constant rate law in which a constant length is added to each crystal diameter in the distribution. This rare process causes the original CSD to narrow so that its initial shape is not preserved. The growth law that prevails, either proportionate or constant, is determined by the manner in which nutrients are supplied to the crystal. Supply is by advective flow during proportionate growth, with the nutrient solution moving with respect to the crystals. Constant growth relies on the random difusion of nutrients through a quiescent solution. Proportionate growth is by far the most common growth law, and therefore, nutrient supply by difusion alone during crystal growth is uncommon. Distributions formed by Oswald ripening and those formed by other rare processes are also discussed. During Ostwald ripening, nucleation caused by mixing reactants at large supersaturation forms crystals that are extremely fine and numerous. The larger crystals grow at the expense of the finer, less stable crystals, thereby forming, on completion, the universal steady-state CSD shape predicted by the Lifshitz-Slyozov-Wagner (LSW) theory. This unique CSD shape, as well as other rare shapes, then are scaled up to larger sizes by supply limited proportionate growth.

The mineral zircon is used widely to constrain the age of rocks and the processes that formed them. Although zircon is robust to a range of physical and chemical processes, it may show evidence for rapid re-equilibration that is generally considered to reflect interaction with hydrous fluids. Here, we show that zircon grains that crystallized from melt produced during the catastrophic meltdown of the Chernobyl nuclear reactor exhibit re-equilibration textures that occurred in an environment without free water. The process of re-equilibration involved a melt-mediated interface-coupled dissolutionreprecipitation that took place over a few days to produce textures that are commonly observed in igneous and anatectic systems. Thus, the composition of zircon can be modified even in the absence of hydrous fluids in a short time frame. Through this process, zircon crystals may track the timing of the last silicate melt they interacted with.

The beryl-group mineral johnkoivulaite, Cs(Be 2 B)Mg 2 Si 6 O 18 , was compressed hydrostatically in a diamond-anvil cell up to 10.2 GPa. In situ Raman spectroscopy and X-ray crystallography revealed a P 6/ mcc- to -P 3 c 1 (second-order) phase transition on isothermal compression at the critical transition pressure P c = 4.13 ± 0.07 GPa. The elastic parameters determined for the volume elasticity of the two polymorphs correspond to a Birch-Murnaghan equation of state with K 0 = 148 ± 2 GPa and K ′ = 0 for P < P c and K 0 = 75.5 ± 0.9 GPa with K ′ = 4 for P > P c . The low- P polymorph shows anomalously linear compression behavior, as reported for several other beryl-derived framework structures. The high- P polymorph, which was found to follow a a ′ = a ⋅ 3 , c ′ = c ${{a}^{\prime }}=a\cdot \sqrt{3},{{c}^{\prime }}=c$superstructure according to P 3 c 1, is almost twice as compressible as its low- P form. This is unique for any beryl-derived structure and can be attributed to the high degree of freedom for atomic displacements in the superstructure. The reduced symmetry can also be understood as the effect of the driving mechanism of the transformation. The extra-framework Cs channel components counteract any lateral deformation of the channels parallel to [0001] within the microporous framework and, similar to pezzottaite, are responsible for maintaining the trigonal/hexagonal lattice metrics.

Shock lithification of regolith breccias is a ubiquitous process on the surfaces of airless planetary bodies and may induce thermal effects, including melting on regolith breccia minerals. However, potential thermal effects on lithic and mineral clasts in regolith breccias have seldom been quantitatively constrained. Here, we report two types of micro-textures of armalcolite [(Mg,Fe 2+ )Ti 2 O 5 ] in an Mg-suite lithic clast from lunar regolith breccia meteorite Northwest Africa 8182. One type of armalcolite contains oriented fine-grained ilmenite grains; the other occurs as an aggregate of ilmenite, rutile, spinel, and loveringite. We propose that the two types of micro-textures formed through subsolidus breakdown of armalcolite by different processes. The formation of ilmenite inclusions in armalcolite is related to slow cooling after the solidification of its source rock, whereas the ilmenite-rutile-spinel-loveringite aggregates probably formed during the shock lithification event of NWA 8182. The results indicate that the temperature at the margin of lithic clasts could be raised up to at least 600 °C during strong shock lithification of lunar regolith and has profound thermal effects on the mineralogical and isotopic behaviors of lithic and mineral fragments in lunar regolith breccias.

Several important processes in the petrogenesis of granite are still debated due to a poor understanding of complex interactions between minerals during the melting and melt segregation processes. To promote an improved understanding of the mineral-melt relationships, we present a systematic petrographic and geochemical analysis for melanosome and leucosome samples from the Triassic Jindong migmatite, South China. Petrographic observations and zircon U-Pb geochronology indicate that the Jindong migmatite was formed through water-fluxed melting of the Early Paleozoic gneissic granite (437 ± 2 Ma) during the Triassic (238 ± 1 Ma), with the production of melt dominated by the breakdown of K-feldspar, plagioclase, and quartz. The Jindong leucosomes may be divided into lenticular and net-structured types. Muscovite, plagioclase, and K-feldspar in the net-structured leucosome show higher Rb and much lower Ba and Sr contents than those in the lenticular leucosome. This may be attributed to the elevation of Rb and decreasing Ba and Sr abundances in melts during the segregation process due to early fractional crystallization of K-feldspar and plagioclase. These leucosomes show negative correlation between ε Nd (t) and P 2 O 5 , reflecting increasing dissolution of low-ε Nd (t) apatite during the melting process. The continuous dissolution of apatite caused saturation of monazite and xenotime in melt, resulting in the growth of monazite and xenotime around apatite in the melanosome. This process led to a sharp decrease of Th, Y, and REE with increasing P 2 O 5 in the leucosome samples. This complex interplay of accessory mineral reactions in the source impacts REE geochemistry and Nd isotope ratios of granites. As the granites worldwide exhibit similar compositional and isotopic patterns to the Jindong leucosomes, we suggest that both the melting and melt segregation processes strongly control the granitic melt compositions.

From integrated textural and compositional studies of auriferous and barren pyrite/marcasite in the epithermal Axi gold deposit, China, we have identified a relationship between multiple gold mineralizing events, mafic magma recharge, and fluid-rock reactions. Three generations of pyrite (Py1–3) and four generations of marcasite (Mar1–4) record episodic gold mineralizing events, followed by silver-copper-lead-zinc-cadmium enrichment. The gold mineralizing events are recorded by high concentrations of subnanometer-sized gold in Py1, Py3, and Mar3 (max. = 147, 129, and 34 ppm, med. = 39, 34, and 12 ppm). Based on previous Re-Os age determinations of pyrite and U-Pb zircon ages of the andesitic wallrock, these gold events slightly postdate pulsed mafic magma recharge and represent the incursion of Au-As-S-rich magmatic volatiles into circulating meteoric water. Silver-Cu-Pb-Zn-Cd enrichment in Py2, Mar2, and Mar4 are consistent with quiescent degassing and gradual Ag-Cu-Pb-Zn-Cd enrichment in an evolved felsic magma. Barren Mar1 records the dominance of meteoric water and a limited magmatic fluid contribution. High-Co-Ni-V-Cr-Ti contents in porous cores of Py1 and Mar2 are attributed to wall rock alteration and dissolution-reprecipitation. The results provide convincing evidence that the metal budget (especially for Au, Ag, Cu, Pb, Zn, Sb) of the hydrothermal fluids and sulfides in epithermal systems are controlled by the influx of magmatic fluids and associated magma, whereas the enrichment of certain fluid-immobile elements, such as Co, Ni, V, Cr, and Ti, is caused in part by fluid-rock interaction.

We report the formation of minerals from the tochilinite-valleriite group (TVG) during laboratory serpentinization experiments conducted at 300 and 328 °C. Minerals in the TVG are composed of a mixture of sulfide and hydroxide layers that can contain variable proportions of Fe, Mg, Cu, Ni, and other cations in both layers. Members of this group have been observed as accessory minerals in several serpentinites, and have also been observed in association with serpentine minerals in meteorites. To our knowledge, however, TVG minerals have not previously been identified as reaction products during laboratory simulation of serpentinization. The serpentinization experiments reacted olivine with artificial seawater containing 34 S-labeled sulfate, with a small amount of solid FeS also added to the 300 °C experiment. In both experiments, the predominant reaction products were chrysotile serpentine, brucite, and magnetite. At 300 °C, these major products were accompanied by trace amounts of the Ni-bearing TVG member haapalaite, Ni,Fe-sulfide (likely pentlandite), and anhydrite. At 328 °C, valleriite occurs rather than haapalaite and the accompanying Ni,Fe-sulfide is proportionally more enriched in Ni. Reduction of sulfate by H 2 produced during serpentinization evidently provided a source of reduced S that contributed to formation of the TVG minerals and Ni,Fe-sulfides. The results provide new constraints on the conditions that allow precipitation of tochilinite-valleriite group minerals in natural serpentinites.

We present in situ LA-ICP-MS U-Pb dating of xenotime and monazite in assemblages with native gold and Au (Ag) tellurides from the Xiaoqinling lode gold district in central China. Composite xenotime and monazite grains formed through coupled dissolution-reprecipitation reactions reveal two discrete gold mineralization events. The first gold mineralization event, recorded by monazite (158.6 ± 3.3 Ma, Tera-Wasserburg lower intercept age) and xenotime cores (157.11 ± 0.83 Ma, weighted mean 206 Pb/ 238 U age), is characterized by the mineral assemblage of lingbaoite (AgTe 3 )-sylvanite ([Au,Ag] 2 Te 4 )-stützite (Ag 5 –xTe 3 )/native tellurium-sylvanite-stützite. The second gold mineralization event, recorded in the rims of xenotime (135.46 ± 0.93 Ma, weighted mean 206 Pb/ 238 U age), is characterized by the mineral assemblage of native gold-calaverite (AuTe 2 )-petzite (AuAg 3 Te 2 )-tellurobismuthite (Bi 2 Te 3 ). Our study implies that the large-scale Jurassic mineralization event in eastern China, related to flat subduction of the paleo-Pacific plate beneath the eastern China continent, also caused widespread gold mineralization in the Qinling-Dabie Orogen, in addition to production of its world-class porphyry Mo deposits. The fact that only a few Jurassic gold mineralization ages have been reported before, may be due to the lack of suitable geochronometers to record the earlier Jurassic hydrothermal processes, which have been overprinted by the better-recognized Early Cretaceous gold mineralization event. This study also presents a rare example of xenotime compositional alterations and resetting of U-Pb ages induced by low to moderate salinity carbono-aqueous fluids at low temperatures. The textural relationships between gold minerals in contact with such composite xenotime crystals demonstrate that they could have precipitated before, coeval with, or after the dated domains. Since low to moderate salinity carbonoaqueous fluids are commonly involved in the formation of lode gold deposits, it is crucial to examine xenotime textures and recognize potential alteration textures before carrying out isotopic dating of xenotime collected from these deposits. Without prior compositional and textural characterization, attempts to date such composite crystals could yield mixed dates and meaningless ages.

Amphiboles in the upper mantle (most frequently pargasitic in composition) have recently gained attention due to their role in the water budget and potential control on the rheology and physical discontinuity of layers of the mantle. Although nominally anhydrous minerals are often analyzed with Fourier transform infrared (FTIR) spectroscopy, amphiboles, especially in natural samples, are only rarely the focus of such studies because of their complex structure and variable composition. In mantle xenoliths, amphibole occurs not only interstitially or forming veins, but also as lamellae within orthopyroxene and/or clinopyroxene grains. The genesis of such lamellae is often ambiguous, as they could either be meta-somatic products, or form by exsolution without an external H 2 O source upon decreasing P - T conditions and consequent destabilization of hydrous point defects in the host pyroxene. To constrain the origin of amphibole lamellae in pyroxenes, we studied mantle xenoliths from the Carpathian-Pannonian region (CPR) by applying hyperspectral imaging using an FTIR equipped with focal plane array (FPA) detector. Amphibole lamellae are absent in xenoliths of the central part of the CPR, but appear in those from the marginal localities that represent a well-hydrated supra-subduction mantle environment. Some of the lamella-hosting pyroxene grains are in contact with interstitial amphibole, suggesting that the formation of the lamellae is related to the amphibole-producing metasomatism; however, others have no adjacent amphibole. To determine the origin of the amphibole lamellae in pyroxenes without neighboring amphibole, hyperspectral images were used to give an estimation of their volume proportion (0.8–5.1 vol%) in the pyroxenes. Using these volume proportions, we calculated that a bulk water content of ~330–670 (orthopyroxene) and ~740–1430 (clinopyroxene) wt. ppm is needed to be contained in the host grain to be able to facilitate subsolidus exsolution of the observed amount of amphibole lamellae. These water contents are, however, too high for mantle pyroxenes, even for an aqueous-fluid saturated upper mantle. This suggests that the formation of amphibole lamellae is related to a metasomatic event with fluid input from an external water source (e.g., melt/fluid inclusion or metasomatic agent on the grain boundary).

Two types of turquoise, including homogeneous Cu-rich turquoise and oscillatory zoned turquoiseplanerite series, are recognized in association with quartz-barite veins hosted by Cambrian carbonaceous slates from western Hubei Province of China. Combined fluid inclusion and Raman micro-spectroscopic data reveal that turquoise-bearing barite-quartz veins contain three types of fluid inclusions (pure CH 4 -N 2 , carbonic-aqueous, and aqueous), suggesting the responsible fluid is of carbonic-aqueous composition with low oxygen fugacity. Pressure-corrected homogenization temperatures in quartz and barite show a range from 325 to 485 °C and 186 to 391 °C, respectively. Carbon, oxygen, and hydrogen isotopic data suggest that the mineral-forming fluids have a mixed metamorphic-organic affinity, in which the fluids have δ 18 O and δD values of 15.0 to 18.8‰ and –111 to –93‰, respectively. Generally, the formation of quartz-barite-turquoise veins could be triggered by prior metamorphic devolatilization, followed by the interaction of fluids with country rocks enriched in carbonaceous material, which resulted in the leaching of Cu, Fe, P, and Al from chalcopyrite, pyrite, magnetite, monazite, xenotime, apatite, feldspar, and muscovite in the wall rocks. Decomposition of the organic matter in the carbonaceous slates, caused by regional metamorphism and deformation, could also promote the concentration and transportation of necessary metals for the turquoise. Thus, we propose a new model and suggest that the turquoise gem deposits in western Hubei Province of China belong to the non-magmatic hydrothermal vein type deposit, not the previously proposed supergene origin. The turquoise-forming fluids were characterized by the coexistence of two immiscible fluids of non-magmatic affinity (i.e., moderate to high-temperature and low-salinity aqueous fluid and pure CH 4 -N 2 fluid formed by interaction with carbonaceous slates). The bluish green homogeneous turquoise in the metamorphic quartz-barite-turquoise veins evolved toward the turquoise-planerite solid solution series as the metal-leaching capability of the aqueous fluids decreased.

The origin of the common blue 450 nm (2.8 eV) cathodoluminescence (CL) emission in natural and synthetic quartz has been investigated using a combination of CL microscopy and spectroscopy, electron paramagnetic resonance (EPR) spectroscopy, and trace-element analysis by electron micro-probe analysis as well as inductively coupled plasma-mass spectrometry (ICP-MS). The study shows that the appearance of the ~450 nm emission band can be attributed to two different defects in quartz. First, a transient luminescence can be explained by structural defects in oxygen deficient quartz. The luminescence model implies self-trapped exciton (STE) emission related to oxygen vacancies. This type of CL emission is frequent in high-purity synthetic quartz and natural quartz of hydrothermal origin. Second, in Ti-rich quartz from natural samples (e.g., quartz phenocrysts in rhyolites) and synthetic quartz of Ti-difusion experiments, an additional 450 nm (2.8 eV) emission was detected, which is stable under the electron beam. The intensity of this ~450 nm emission band correlates with the concentration of trace Ti in quartz, and substitutional Ti 4+ at the Si 4+ position was proved by EPR spectroscopy. In quartz crystals with elevated Ti concentrations both intrinsic and extrinsic blue CL emissions at ~450 nm can coexist, hindering a thorough characterization and quantification of the CL signal. A reliable distinction of the two diferent CL emission bands is possible by fitting the peaks of the CL spectra, and the peak width of the 450 nm emission can be used to diferentiate the STE from the Ti 4+ emission. However, the definitive technique is through the observation of CL peak shape change over time at a point by collecting a time series of CL spectra in conjunction with EPR spectroscopy and trace-element analysis of the Ti concentration.

Hinckley index (Hi) can be used to characterize the crystallinity of kaolinite. Stacking order-disorder in kaolinite can considerably affect its crystallinity. However, the correlation between Hi and stacking order-disorder in kaolinite has not been reported thus far. Herein, the correlation between stacking order-disorder in kaolinite and Hi was investigated via experiments (XRD, IR spectroscopy, TG-DSC), molecular simulation, and structure refinement. First, we experimentally discovered that the stacking order-disorder in kaolinite changed the relative position between two adjacent structural layers, mainly afecting the interlayer forces. When the kaolinite layers are orderly stacked, the interlayer force is higher and the stacking lattice energy is lower. The lattice energy of kaolinite in diferent stacking states was simulated and analyzed using first-principles calculation. It was determined that the kaolinite layers are orderly stacked when two kaolinite layers have zero shift and disorderly stacked otherwise. Finally, through structural refinements, we proposed a new crystallinity index based on stacking order-disorder in kaolinite (crystallinity index based on stacking, CIS). CIS was well fitted to Hi (R 2 = 0.986), indicating that kaolinite crystallinity, characterized by Hi, is essentially the ratio of orderly stacking to total stacking (the sum of ordered and disordered stacks). Furthermore, measuring Hi is difficult when kaolinite crystallinity is poor; however, CIS can be used alternatively. This study of the crystallinity of kaolinite will have important significance for its industrial application.

A solid solution of the mineral feiite (Fe 3 TiO 5 ) was recently discovered in a shock-induced melt pocket of the Shergotty martian shergottite. It is particularly interesting for its potential as an indicator of pressure-temperature ( P-T ) and oxygen fugacity in martian crustal and mantle material. To date, complete crystallographic analysis of feiite has not been conducted, as the mineral was previously analyzed by electron backscatter diffraction on micrometer-size grains (Ma et al. 2021). Here we report a convergent crystal-structure model for feiite based on synchrotron single-crystal X-ray diffraction data collected on three grains of feiite synthesized at 12 GPa and 1200 °C. Feiite adopts the CaFe 3 O 5 structure type ( Cmcm , Z = 4), which is composed of two octahedral M 1 and M 2 sites and one trigonal prismatic M 3 site ( M = metal) in a ratio of 1:2:1. The three feiite grains with composition Ti 0.46–0.60 Fe 3.54–3.40 O 5 were best modeled by substituting Ti 4+ into only the octahedral M 2 site, accounting for 30% of this site. Comparisons of the measured average bond lengths in the coordination polyhedra with the optimized Ti 4+ –O, Fe 2+ –O, and Fe 3+ –O bond lengths suggest that ferrous iron occupies the trigonal M 3 site, while iron is mixed valence in the octahedral M 1 and M 2 sites. The Ti 4+ and Fe 3+ content constrained by our crystal-chemical analyses suggests that at least ~30% of the available iron must be ferric (i.e., Fe 3+ /Fe total = 0.3) for the sample synthesized at 12 GPa and 1200 °C and higher P-T conditions may be needed to form the end-member feiite F e 3 2 + T i O 5 $\left(\mathrm{Fe}_{3}^{2+} \mathrm{TiO}_{5}\right)$

The results of investigations into enrichment of precious metals in sphalerite and pyrite from the Maluntou epithermal gold deposit, China, are reported. The obtained data suggest intimate associations of Au- and Ag-bearing nanoparticles with chalcopyrite inclusions in sphalerite and pyrite. The origins of chalcopyrite inclusions involved different hydrothermal processes, including recrystallization-driven phase separation from parent chalcopyrite-sphalerite solid solutions and replacement of pre-existing pyrite in the presence of Cu-bearing fluids. The chalcopyrite blebs/lamellae follow sphalerite {111} planes, which define a shared sulfur layer for both chalcopyrite and sphalerite. This study indicates that mixing and boiling during the evolution of ore-forming fluids for the Maluntou deposit are key processes for the abnormal enrichment of precious metals in sphalerite and pyrite. The chalcopyrite micro/nano inclusions enhanced enrichment of precious metals in sphalerite provides new insights into the controls on the enrichment of precious metals in sulfides.

Synthetic flux-grown end-member gahnite, ZnAl 2 O 4 , and several different colored crystals doped with one or more transition metals including Mn, Ni, Cr, Co, and Fe were studied by electron mi-croprobe methods and UV/Vis/NIR single-crystal optical absorption spectroscopy. The first major objective was to measure and assign the various electronic absorption features. The second was to analyze quantitatively the crystal colors using the experimental spectra and the CIE 1931 color- spacechromaticity diagram. The microprobe results show that the doped gahnites have transition metal concentrations between about 0.001 and 0.1 cations per formula unit. The spectrum of colorless, nominally pure ZnAl 2 O 4 displays no absorption in the visible region. Microprobe analysis of a light-blue gahnite crystal reveals small amounts of Ni and Mn. The UV/Vis/NIR spectrum does not indicate any dd-electronic transitions relating to Mn. All absorption features also cannot be fully interpreted using Tanabe-Sugano diagrams for Ni 2+ in either octahedral or tetrahedral coordination. A series of seven slightly different colored gahnites with differing concentrations of Cr 3+ and most also containing smaller amounts of Ni was investigated. The spectrum of a one pink crystal shows two intense absorption features in the visible region. They are assigned to spin-allowed 4 A 2g → 4 T 2g ( 4 F) and 4 A 2g → 4 T 1g ( 4 F) transitions of VI Cr 3+ . Other spectra display additional weak bands and lines that are most probably spin-forbidden dd-transitions of Ni 2+ . These gahnites with Ni and Cr show varying purple colorations depending on the concentrations of both metals. Two more deeply blue gahnites contain Co 2+ as demonstrated by their UV/Vis spectra but not by microprobe analysis. Two intense absorption features at ~7440 and ~16 850 cm –1 are observed and assigned to the spin-allowed transitions 4 A 2 → 4 T 1 ( 4 F) and 4 A 2 → 4 T 1 ( 4 P) of Co 2+ , respectively. Complex absorption fine structure, caused by spin-orbit and/or vibronic interactions, is also observed. Three diferent gahnites with yellow to orange colorations contain measurable Mn. Their spectra are similar in appearance and display several weak IV Mn 2+ spin-forbidden transitions located above 20 000 cm –1 . The spectra of two green gahnites show several Fe spin-forbidden electronic transitions arising from single, isolated IV Fe 2+ and VI Fe 3+ cations between 10 000 and 25 000 cm –1 . The intensities of some of the VI Fe 3+ -related bands can be increased through exchange-coupled interactions with next nearest IV Fe 2+ neighbors. The colors of various doped gahnites and the end-member galaxite are analyzed using their single-crystal absorption spectra in the visible region. Their dominant wavelength, λ k , and hue saturation, p c , values are given on the CIE 1931 color-space-chromaticity diagram and are discussed. The Hex colors of all crystals are calculated and can be compared to those of the studied crystals.

Regolith-hosted rare earth elements (REEs) deposits received great attention due to the increasing incorporation of REEs in modern technologies. In lateritic Sc deposits and ion-adsorption deposits (IADs), Sc behaves quite differently from REEs: REEs adsorb as outer-sphere complexes on clay surface in IADs, while Sc could enter the lattice of clay minerals in lateritic Sc deposits. The unique behavior of Sc has not been well understood yet. Here, by using first-principles molecular dynamics techniques, we show that the complexation mechanisms of Y 3+ and Sc 3+ on clay edge surfaces are distinctly diferent. Y 3+ preferentially adsorbs on Al(OH) 2 SiO site with its coordination water protonated. Sc 3+ is found to behave similarly to other first-row transition metals (e.g., Ni 2+ ) due to its smaller ionic radius and prefers adsorbing on the vacancy site, from where Sc 3+ can be readily incorporated in the clay lattice. The H 2 O ligands of Sc 3+ get deprotonated upon complexation, providing new binding sites for further enrichment of Sc 3+ . These processes prevent Sc 3+ from being leached during weathering and lead to the formation of Sc-rich clay minerals found in lateritic deposits. Based on these results, it is revealed that the small ionic radius and high affinity to enter the vacancy on edge surfaces make Sc compatible with clay minerals and are the origin of its unique geochemical behavior.

Scandian actinolite evolving to scandio-winchite (up to 5.45 wt% Sc 2 O 3 ) has been found in chloritedominant xenoliths incorporated into marginal portion of a granitic pegmatite. The pegmatite intruded a blackwall schist zone developed around rodingite-type rocks exposed in a serpentinite quarry at Jordanów Śląski near Sobótka, ~30 km south of Wrocław, Lower Silesia, Poland. The amphiboles form irregular overgrowths around cascandite and represent a complex solid-solution series among actinolite and scandio-winchite end-members, with a trace contribution of “scandio-magnesio-hornblende.” Structural studies of a scandian actinolite crystal with composition A [□ 0.995(2) K 0.005(2) ] Σ1 B [Na 0.24(5) Ca 1.73(4) ] Σ1.98(1) C M g 3.74 ( 7 ) F e 0.90 ( 3 ) 2 + M n 0.04 ( 1 ) S c 0.26 ( 3 ) A l 0.05 ( 1 ) Σ 4.99 ( 1 ) T S i 7.98 ( 2 ) A l 0.02 ( 2 ) Σ 8.00 O 22 ( O H ) 2 ${ }^{\mathrm{C}}\left[\mathrm{Mg}_{3.74(7)} \mathrm{Fe}_{0.90(3)}^{2+} \mathrm{Mn}_{0.04(1)} \mathrm{Sc}_{0.26(3)} \mathrm{Al}_{0.05(1)}\right]_{\Sigma 4.99(1)}{ }^{\mathrm{T}}\left[\mathrm{Si}_{7.98(2)} \mathrm{Al}_{0.02(2)}\right]_{\Sigma 8.00} \mathrm{O}_{22}(\mathrm{OH})_{2}$revealed monoclinic C 2/ m structure with unit-cell parameters a = 9.8517(3), b = 18.0881(6), c = 5.28501(18) Å, β = 104.809(4)°, in which scandium is located solely at the C M2 site. Scandian amphiboles are uncommon in geological environments, and invite comments on the origin of the observed Sc enrichment in the amphibole structure. Textural appearance of the chlorite-cascandite-amphibole clusters suggests that the formation of the amphiboles is related to the evolution of the country rocks followed by partial alteration of blackwall schist xenoliths by pegmatite-forming melt.

Nelsonite (Fe-Ti oxide-apatite rock) devoid of silicates offers a rare opportunity to investigate the magma processes for the formation of magmatic Fe-Ti oxide deposits. Both fractional crystallization and silicate liquid immiscibility have been put forward, but the lack of robust evidence has hindered unambiguously distinguishing the role of these two processes in Fe-Ti mineralization. The nelsonite and associated Fe-Ti-P-rich rocks hosted in the Proterozoic Damiao anorthosite complex represent a typical example for studying Fe-Ti ore-forming processes. We recognized a new type of nelsonite (type-I) in the Damiao complex, which is distinct from the two known types of nelsonite (type-II and type-III) from the same complex. The type-I nelsonite is characterized by its coexistence with oxide-apatite gabbronorite and granite in the same dike, and all these rocks have identical emplacement ages (1740 ± 7 Ma), subparallel REE patterns, and major-element compositions lacking intermediate compositions, suggesting derivation from conjugate Fe- and Si-rich melts generated by silicate liquid immiscibility. The large type-II nelsonite bodies form irregular dikes along fractures in anorthosite and constitute the major ore type. The type-III nelsonite occurs as conformable layers or pods within oxide-apatite gabbronorite and pyroxenite, and occupies the end part of the type-II dike. The latter two types of nelsonites formed by extensive fractional crystallization of residual magma with crystal accumulation and subsequent hydrothermal replacement. During residual magma evolution, silicate liquid immiscibility was crucial for Fe-Ti-P enrichment, fractional crystallization was responsible for enhancing oxide-apatite concentrations, and hydrothermal replacement was effective for mobilizing oxide-apatite concentrations. Our newly recognized nelsonite provides an unambiguous, outcrop-scale, field evidence for the operation of silicate liquid immiscibility process. We show that giant magmatic Fe-Ti oxide orebodies can form by a combination of processes involving silicate liquid immiscibility, fractional crystallization and hydrothermal mobilization.

Pre-Cenozoic high-Mg andesites (HMAs) are mostly present in continental interiors, but their genetic relationship with continental lithosphere evolution remains unclear because of uncertainties of their mantle source, magmatic processes, and physicochemical conditions of formation. Early Cretaceous high-Mg dioritic rocks (HMDs, analogs of HMAs) of the Jinling complex in the Luxi area are typical intra-plate intrusions of the eastern North China Craton (NCC) and can be subdivided into two groups (Group-I and -II) on the basis of their petrographic and geochemical features. Group-I HMDs show low SiO 2 contents (52.47–56.10 wt%) and Sr/Y (34.5–39.6) and (La/Yb) N (10.3–13.6) ratios but high contents of MgO (7.86–9.13 wt%), Y (18.3–20.3 ppm), Yb (1.43–1.47 ppm), and compatible elements (Cr = 407–585 ppm; Ni = 117–216 ppm), classifying as sanukitic rocks. Group-II HMDs are characterized by high SiO 2 contents (63.81–64.87 wt%) and Sr/Y (47.1–63.4) and (La/Yb) N (16.1–17.5) ratios with low MgO (2.90–3.08 wt%), Y (0.88–1.04 ppm), Yb (0.88–1.04 ppm), and compatible elements (Cr = 201–213 ppm; Ni = 55–57 ppm) contents, belonging to adakitic rocks. Group-I and Group-II HMDs of the Jinling complex are closely related in spatial and temporal distribution, and all have enriched Sr-Nd isotopic compositions and arc-like trace element patterns with abundant hydrous minerals. Therefore, the Jinling HMDs should share a common source of ancient sub-continental lithospheric mantle that was metasomatized by aqueous fluids derived from the subducted Paleo-Pacific slab. The Jinling HMDs were not formed from interaction between slab-derived melts and mantle-wedge peridotites but were instead derived from partial melting of hydrous mantle peridotites in the continental interior of the eastern NCC. The distinctly different petrography, geochemistry, and mineralogy of the two groups of rocks resulted mainly from differing magmatic processes at crustal depths. Thus, Pre-Cenozoic intra-plate HMAs/HMDs are genetically distinct from Cenozoic HMAs that were mostly present in arc settings and generally represent juvenile crust growth. In a way, Archean tonalitic-trondhjemiticgranodioritic rocks (TTG) and sanukitoids, geochemically similar to HMAs/HMDs, could also be derived from interaction between slab-derived melts and mantle-wedge peridotites in arc settings or partial melting of hydrous mantle peridotites in continental interiors, and thus might not always be related with continental crustal growth and the onset of plate subduction.

Gold in Carlin-type gold ores is commonly hosted in the arsenian pyrite rim, but the formation of arsenian pyrite and its contribution to Au adsorption are poorly understood. Based on our previous NanoSIMS Au mapping, we conducted SEM and HR-TEM analyses to examine the Au deportment and nanoscale texture of individual auriferous arsenian pyrite grains from the giant Carlin-type Lannigou gold deposit, SW China. The results indicate that the arsenian pyrite rim is composed of numerous nanoparticulate pyrite grains (rather than a single crystal), and gold nanoparticles (Au 0 ) occur mainly in sub-rim with the highest Au content, which are porous and have lower degrees of order. We propose that nanoparticulate arsenian pyrite attachment and aggregation is the main mechanism for the arsenian pyrite rim growth, and such mechanism is crucial for the Au efficient enrichment for this giant gold deposit.