New insights into the crystal chemistry of sauconite (Zn-smectite) from the Skorpion zinc deposit (Namibia) via a multi-methodological approach

1 Revision 1 New insights into the crystal chemistry of sauconite (Zn-smectite) from the Skorpion zinc deposit (Namibia) via a multi-methodological approach Emanuela Schingaro1*, Gennaro Ventruti1, Doriana Vinci1, Giuseppina Balassone2, Nicola Mondillo2, Fernando Nieto3, Maria Lacalamita1, Matteo Leoni4,5 1Dipartimento di Scienze della Terra e Geoambientali, Università degli Studi di Bari Aldo Moro, Via Orabona 4, I-70125, Bari, Italy 2Dipartimento di Scienze della Terra dell’Ambiente e delle Risorse, Università “Federico II”, Complesso Universitario Monte S. Angelo, Via Cintia, I-80126, Napoli, Italy 3Departamento de Mineralogía y Petrología, IACT, Universidad de Granada-CSIC, Av. Fuentenueva s/n, 18002, Granada, Spain 4Saudi Aramco Research and Development Center, P.O. Box 5000, 31311, Dhahran, Saudi Arabia 5Dipartimento di Ingegneria Civile, Ambientale e Meccanica, Università di Trento, Via Mesiano, 77, Trento, 38123, Italy *Corresponding author: Emanuela Schingaro, e-mail: [email protected] RUNNING TITLE: New insights into the crystal chemistry of sauconite This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 2 ABSTRACT A multi-methodical characterization of a sauconite (Zn-bearing trioctahedral smectite) specimen from the Skorpion ore deposit (Namibia) was obtained by combining X-ray powder diffraction (XRPD), Cation Exchange Capacity (CEC) analysis, Differential Thermal Analysis (DTA), Thermogravimetry (TG), Fourier Transform Infrared spectroscopy (FTIR) and Transmission Electron Microscopy (TEM-HRTEM-AEM). The X-ray diffraction powder pattern exhibits the typical features of turbostratic stacking disorder with symmetrical basal 00l reflections and longtailed hkbands, confirmed also by TEM observations. Besides sauconite, the sample also contains minor amounts of kaolinite, dioctahedral smectite and quartz. CEC analysis provides a total of Ca (∼69%), Mg (∼26%), Na (∼4%) and K (0.7%) exchangeable cations. Therefore, Zn is located exclusively within the octahedral site of sauconite. TG analysis provides a total mass loss of about 17% in the studied sample. Three endothermic peaks can be observed in the DTA curve, associated to dehydration and dehydroxylation of the material. An exothermic peak at 820°C is also present as a consequence of dissociation and recrystallization phenomena. The infrared spectrum shows the typical Zn3OH stretching signature at 3648 cm-1, whereas, in the OH/H2O stretching region two bands at 3585 and 3440 cm-1 can be associated to stretching vibrations of the inner hydration sphere of the interlayer cations and to absorbed H2O stretching vibration, respectively. Diagnostic bands of kaolinite impurity at ~3698 and 3620 cm-1 are also found, whereas 2:1 dioctahedral layer silicates may contribute to the 3585 and 3620 cm-1 bands. Finally, using the one layer supercell approach implemented in the BGMN software, a satisfactory XRPD profile fitting model for the Skorpion sauconite was obtained. The findings have implications not only for economic geology/recovery of critical metals but, more generally, in the field of the environmental sciences. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 3 Keywords: sauconite, nonsulfide ore deposits, Skorpion (Namibia), CEC, XRPD profile modeling, thermal analysis, FTIR, TEM INTRODUCTION Sauconite, a Zn-bearing trioctahedral smectite, is an ore mineral in nonsulfide Zn-Pb deposits, which typically form by weathering of precursor Zn-sulfide mineralizations under different tectonic contexts and climates, but can be also genetically related to hydrothermal fluids circulation (Large 2001; Hitzman et al. 2003; Boni and Mondillo 2015). Sauconite is commonly associated to other Zn-minerals, like smithsonite, hydrozincite, hemimorphite, willemite and various 10 Åand 7 Åspaced phyllosilicates (Emselle et al. 2005; Mondillo et al. 2015; Buatier et al. 2016; Arfè et al. 2017b; Balassone et al. 2017). With a content up to 50 wt.% Zn oxide, sauconite can be an accessory mineral, like in various occurrences in Belgium, Ireland, Morocco, Peru (i.e. Balassone et al. 2008, 2020; Coppola et al. 2008; Boni et al. 2009a, 2009b; Mondillo et al. 2014, 2015; Buatier et al. 2016; Choulet et al. 2016; Arfè et al. 2017a, 2017b). In some other notable cases e.g., in Yanque, Peru (Mondillo et al. 2014) and Skorpion Namibia (Borg et al. 2003; Kärner 2006; Arfè et al. 2017a; Balassone et al. 2017) it can be an important source of metal. Relatively recent developments in procedures combining leaching, solvent extraction and electrowinning, and their application to Zn smelting, have made sauconite amenable for economic Zn recovery, creating an interest in ore deposits dominated by these Zn-silicates (Gnoinski 2007; de Wet and Singleton 2008; Abkhoshk et al. 2014). However, Zn-clays also represent low-cost materials being synthesized at lower temperature compared to smectites with other transition metals (Carniato et al. 2020 and references therein). Smectite group minerals have a 2:1 layered structure made of negatively charged octahedral and tetrahedral sheets, with water molecules and partly exchangeable cations located in the interlayer space (Brigatti et al. 2006). These features, as well as the small This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 4 particle size (<100 nm), the large specific surface area (>10 m2/g) and the presence of surface functional groups (i.e. ≡Al-OH and ≡Si-OH) are responsible for the uptake of various metal ions (i.e. Churchman et al. 2006; Balderman et al. 2019, and references therein), making such materials suitable for industrial pollution and wastewater treatment, as well as for chemical barrier applications (among the vast literature on these issues, see for instance Churchman et al. 2006; Pascua et al. 2010; Sasaki et al. 2016; Steudel et al. 2017, and references therein; Mercurio et al. 2018, and references therein). For example, Zn can be incorporated as Zn2+ cations in the interlayer space of swelling structures (i.e. Bradbury and Baeyens 1999; Churakov and Dähn 2012). The potential role of clay minerals in relation to the origin of life, for their role of strong adsorbents of polar organic molecules that can facilitate abiogenesis, has been also recently investigated by Zhou et al. 2017, and references therein), who studied in particular the behavior of Zn clay (sauconite). Even though several studies have been published on the chemical characterization of natural sauconite from different deposits (i.e. Ross 1946; Faust 1951; Newman and Brown 1987; Kärner 2006; Coppola et al. 2008; Kaufhold et al. 2015; Mondillo et al. 2015; Buatier et al. 2016; Choulet et al. 2016; Balassone et al. 2017, 2020) and of synthetic sauconite (i.e. Decarreau et al. 1987; Higashi et al. 2002; Yokoyama et al. 2006; Petit et al. 2008; Pascua et al. 2010; Zhou et al. 2017), only few of them (Faust 1951; Coppola et al. 2008; Kaufhold et al. 2015; Choulet et al. 2016) report some structural features of this Zn-smectite. This lack is essentially due to the intrinsic complexity of sauconite, caused by the occurrence of stacking disorder, common in layered minerals (Lanson 2011), as well as by the variable extent of cation substitutions, different coordination and positional disorder of interlayer cations, variable number of water molecules and different interlayer water configurations. Furthermore, to the authors’ knowledge, very few studies exist on the cation exchange behavior of natural sauconite. Specifically, Mitra and Sindhu (1971) investigated the cation exchange capacity of sauconite treated with H-resin and aged in water using the sauconite This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 5 from Arkansas in Ross (1946) as starting material, whereas more recently a multianalytical study (comprising CEC experiments) was carried out by Choulet et al. (2016) on Moroccan sauconite (discussed below). In the present study, a sauconite specimen from the Skorpion ore deposit (Namibia) has been investigated using complementary techniques, i.e. a combination of X-Ray Powder Diffraction (XRPD), thermal analysis, Fourier Transform Infrared (FTIR) as well as Cation Exchange Capacity (CEC) measurements, to provide a sound mineralogical characterization as well as new insights into the species from a structural point of view. The final aim is to shed new light on such Zn trioctahedral smectite, starting from the case of study of a natural Zn-clay, in the context of practical implications not only for the economic geology and ore processing but also, on a broader level, for many fields of materials, environmental and life sciences. MATERIALS AND METHODS Sample occurrence and description The Skorpion ore deposit (Vedanta Ltd.) is located ca. 40 km north of the Orange River and about 15 km north-northwest of Rosh Pinahmine in the southern most Namib Desert (Namibia), and represents the 8th largest Zn mine in the world (Boni and Mondillo 2015). The Zn mineralization is hosted in Neoproterozoic rocks of mixed volcano-sedimentary origin, described in detail in Borg et al. (2003) and Kärner (2006). In the Skorpion ore deposit, sauconite mainly occurs in metasilicoclastic rocks (mainly arkoses), as coatings of intergranular spaces and voids formed through the breakdown or dissolution of detrital feldspar and mica, or also by replacement of earlier deposited hemimorphite and smithsonite (Kärner 2006). Sauconite predominates over the other Znoxidized minerals (e.g. smithsonite, hydrozincite, hemimorphite, tarbuttite). The formation process of the Skorpion Zn-bearing supergene minerals started, probably, at the end of Cretaceous, after the This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 6 exhumation, weathering and oxidation of precursor sulfides (sphalerite, pyrite, minor chalcopyrite and galena), hosted in late Proterozoic felsic metavolcanic and metasilicic clastic rocks (Borg et al. 2003; Kärner 2006). The metal-bearing (Zn-dominated) fluids percolating through the rocks were able to dissolve some of their detrital silicic clastic components, i.e. mainly detrital feldspars and micas as well as the calcite cement, which provided Si and Al, together with alkaline ions, for sauconite formation. The mineral-water reactions of the hydrolysis processes must have buffered the pH of the supergene fluids, since two strong bases were produced, namely NaOH and KOH (Sherman 2001). These bases contributed to the buffering/neutralization of the acid meteoric fluids, which migrated through the Skorpion meta-arkoses. Thus, the supergene fluids became less acidic during the progressive leaching of feldspars from the Late Proterozoic rocks. This neutralization process played a significant role in the formation of the supergene nonsulfide zinc orebody, rising Eh and pH to moderate conditions, most favorable for the precipitation of the supergene zincbearing minerals, among which sauconite. In particular, sauconite formation requires Eh-pH conditions similar to hemimorphite and smithsonite (i.e. neutral to weak alkaline oxidizing conditions), although it can be also stable under weak acid conditions (Kärner 2006, and references therein). Based on a recent study (Arfè et al. 2017a), the stable isotope data of supergene carbonates suggest that the nonsulfide mineral assemblage formed at an average temperature of 17°C, either during the first (Late Cretaceous-Paleocene), or the last humid climatic stage (early-middle Miocene). Starting from very heterogeneous materials and in order to select a sauconite-rich fraction as pure as possible and suitable for a full crystal chemical and structural characterization, the rock sample labelled SK8 was considered the best candidate among the Skorpion sample suite, being particularly rich in sauconite (Balassone et al. 2017). In the bulk sample, Balassone et al. (2017) also found minor amounts of Fe-oxy-hydroxides, quartz and muscovite. For the present study, white This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 7 fragments of the SK8 rock sample were selected and coarsely ground in agate mortar; any possible contaminant discernible at the stereomicroscope scale was carefully removed by hand-picking. Then, the purity of sample was checked by X-ray powder diffraction on a small fraction of powdered sample. Cation exchange capacity The cation exchange capacity was measured with 0.01 M Cu-triethylenetetramine (Cu-TETA) after Meier and Kahr (1999). A detailed description of this method is given in Steudel et al. (2009). About ~100 mg of SK8 powdered sample was divided in two aliquots in order to check for the reproducibility of the analysis. The results are reported in Table 1. X-ray Powder Diffraction Preliminary XRPD screening of selected fractions of sample SK8 was performed by a Seifert MZVI-GE system, with ID3003 generator, equipped with Ni-filtered Cukα radiation generated at 40 kV and 30 mA. The scanning area covered a 3-70° 2θ range with a step scan 0.02° and an exposure time 5 sec/step. Qualitative analysis was carried out with the RayfleX (GE) software package based on the ICDD PDF-2 database. The XRPD patterns for structural investigations of sauconite were collected using a PANalytical Empyrean diffractometer equipped with a 1.8 kW CuKα ceramic X-ray tube, a Real Time Multiple Strip (RTMS) PIXcel3D detector. Analytical conditions were: 40 mA and 40 kV, 2θ range from 4 to 70°, a scanning speed of 0.026°/s, counting time 360 s per step. The incident beam pathway included a 0.125° divergence slit, a 0.25° antiscattering slit and 0.02 rad Soller slits, whereas a Ni filter, 0.02 rad Soller slits and an antiscatter blade (7.5 mm) were mounted in the diffracted beam pathway. Random mount was obtained by side loading the powder in a flat sample holder. Oriented This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 8 mounts were prepared following consolidated procedures (see for instance Moore and Reynolds 1997). The sample heated at 900°C after TGA-DTA analysis (see below) was also investigated by XRPD. X-ray patterns were preliminarily analyzed using the X'Pert High Score 3.0e software which includes the ICSD database, whereas the structural characterization was carried out by means of the Rietveld refinement software BGMN (Bergmann et al. 1998) using the new graphical user interface Profex (Doebelin and Kleeberg 2015). The results of XRPD investigation are reported on Figures 2, 3, 4 and 9. Thermal analysis Simultaneous differential thermal, thermogravimetric and derivative thermogravimetric analysis (DTA/TG/DTG) was carried out using a Toshiba STA7200RV analyzer. The measurements were performed on about 20 mg of SK8 sample placed in an alumina crucible and heated from room temperature up to 900°C at a rate of 10°C/min in a nitrogen flow (60 mL/min). The TG-DTA-DTG curves of the SK8 sauconite are shown in the Figure 5 whereas the observed peaks and the phenomena which may occur at each temperature for the studied and literature sauconite are listed in Table 2. Infrared analysis Fourier transform infrared measurements were acquired using a Nicolet 380 FTIR spectrometer equipped with an EverGlo source, a KBr beamsplitter and a deuterated triglycine sulfate (DTGS) detector. FTIR spectra were collected in transmission mode on pellets of 13 mm discs with a sample-to-dried KBr weight ratio of l:200. The nominal resolution was 4 cm−1 and 128 scans over the range 450–4000 cm−1 were averaged for each sample and background. The complete FTIR spectrum of sauconite is shown in Figure 6. The results of fitting of the OH/H2O stretching region This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 9 are displayed in Figure 7 whereas measured band positions and assignments for the studied and literature sauconite are summarized in Table 3. TEM-HRTEM and AEM Textural vs. structural studies down to nanoscale on sample SK8 were carried out by TEMHRTEM at the Centro de Instrumentacion Cientifica of Granada (CIC, Spain). Copper rings were attached to representative selected areas of thin sections prepared with Canada balsam and ionthinned using a Fischione Model 1050 ion mill, and then carbon coated. Ion milling was performed at 4 kV and ±10°, until the first hole and ±7° during 20 min for final cleaning. The HRTEM analysis was carried out by means of a Titan G2 80 300 TEM with XFEG emission gun, spherical aberration corrector and HAADF detector, working at 300 kV, with a resolution of 0.8 Å in the TEM mode and 2 Å in the STEM mode. EDX spectra for qualitative identification of minerals and chemical maps were obtained using the Super-X system; quantitative X-ray microanalysis performed in an analytical electron microscope (AEM) equipped with an EDS module allowed to obtain elemental composition of nanometric region of mineralogical samples when operating with TEM-HRTEM (Abad et al. 2003). Mineral standards were measured using the same protocol as for the sample evaluation, to obtain K-factors for the transformation of intensity ratios to concentration ratios according to Cliff and Lorimer (1975). RESULTS AND DISCUSSION Chemical features The chemical composition of the studied SK8 sauconite was recently obtained by AEM and ICPOES methods (Balassone et al. 2017). The ICP-OES analyses gave a composition with all the values inside the range of those obtained by AEM, expressed in wt.% oxides, namely: SiO2 37.93, This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 16 anastomosing relationship to each other, with frequent layer terminations at low-angle contacts with the neighbor sub-packet. The chemical mapping of the area of Figure 8e confirms the AEM data (Balassone et al. 2017) and the results reported in the section "Chemical features" above and shows that the specimen is fairly homogeneous at the TEM scale. XRPD profile modeling To date no suitable structural model is available for sauconite, because this clay mineral is found in nature always turbostratically disordered (inset Figure 8a). As already noted in the section XRPD characterization above, the peculiar shape of sauconite diffraction peaks (Figure 2) is indicative of the occurrence of this kind of disorder. In turbostratic stacking disorder, individual layers in the crystal structures are statistically rotated and translated with respect to each other (Drits and Tchoubar 1990), thus breaking the electron density periodicity along the stacking direction. Lattice fringe images and electron diffraction data (Figures 8c and 8a) confirm this microstructural feature with a view in the direct space, which explains the characteristics of the reciprocal space information here obtained. Different methodological approaches can be used to model the powder diffraction pattern of turbostratically-disordered layered structures. The DIFFaX+ code (Leoni et al. 2004) is able to model the pattern of faulted layered materials using the recursive algorithm of Treacy et al. (1991). The latter code was used by Viani et al. (2002) to simulate the diffraction pattern of a Ca-montmorillonite, but their final model of disorder was rather that of a highly faulted structure with random shifts along ± b/3 and – a/3, with 75% of defect density than that of a truly turbostratically disordered strucure, with 100% of defect density. Random translations, which dominate the turbostratic effect, can be simulated by using a supercell containing several layers translated by random amounts relative to each other. This approach was successfully developed by Ufer et al. (2004) and handled in the BGMN software (Bergmann et al. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 17 1998) through a language interpreter. Subsequently it was applied to improve quantitative phase analyses of smectites (Ufer et al. 2008) and to refine some structural parameters (Ufer et al. 2012; Wang et al. 2018). In this work, diffraction pattern of the of turbostratically disordered Skorpion sauconite has been fitted using the one-layer supercell approach developed by Ufer et al. (2004). The starting structure model of sauconite was prepared by modifying the Tsipursky and Drits (1984) de-hydrated K-saturated smectite with monoclinic space group C2 (their model 2). The interlayer space was increased to 15 Å to simulate a bi-hydrated Ca-smectite. The species and occupancies of interlayer, octahedral and tetrahedral cations in the supercell model were fixed from the calculated sauconite formula, considering only Ca as the interlayer cation. The geometry of TOT layer was adapted to that of a trioctahedral one, by varying the coordinates in order to obtain reasonable Zn-O bond distances (Sham and Wu 1999). A rigid-body model of the interlayer cations (Ca) octahedrally coordinated to water molecules was used and the cation polyhedron was allowed to rotate to guarantee that the two water planes be parallel to the ab-plane, as suggested by specific literature (Slade et al. 1985; Beyer and von Reichenbach 2002) The refinement yielded <Ca-Ow> = 2.51(3) Å. A supercell was used, stacking 30 layers along the c0 direction, a size sufficiently large to contain all relevant stacking correlations and to generate smooth hk peaks by superposition of different orders of hkl reflections. An additional broadening parameter which accounts for the dependence of the l index on the peak width was also introduced in the modeling (Ufer et al. 2004). The supercell was filled with just one layer of 15 Å thickness. Then, 00l reflections are generated from the 15 Å subcell and their intensity was modified to account for proper cell filling. The pattern was modeled in the 4-70° 2q range. The best fit model was achieved by also refining lattice parameters (a, b and c), size and microstrain broadening parameters to match the measured sauconite pattern. For sauconite anisotropic preferential orientation was also refined, that in BGMN This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 18 is implemented by using spherical harmonics (Bergmann et al. 2001). The impurity phases, i.e. kaolinite, dioctahedral smectite and quartz were also included in the refinement. The crystal structure data for the impurity phases were taken from: Tsipursky and Drits (1984) cis-vacant model, with d001 modified to ~ 15 Å and disorder model after Ufer et al. 2004 for smectite; Bish and von Dreele (1979) for kaolinite model and Ufer et al. 2015 for kaolinite disorder modeling (their Kga-2 sample); D’Amour et al. 1979 for a-quartz. The chemistry of the diocahedral smectite was constrained to that of beidellite in Balassone et al. 2017. For the impurity phases only cell parameters and peak broadening parameters were refined as well as the phases scale factors. The simulated pattern reproduces the main features present in the experimental pattern (Figure 9). The main misfits are at 2 q = 5.90, 24.5 and 45°. The misfit at 2q = 5.90° corresponding to 001 sauconite reflection is likely due to the not homogenous hydration at the interlayer (Ferrage et al. 2005; Kaufhold et al. 2011; Dazas et al. 2013; Ferrage 2016; Wang et al. 2018; Vinci et al. 2020). The other two misfits have 3.63 and 2.01 Å d-spacings respectively, showing a good match with the 1M muscovite of Soboleva et al. (1975). However, attempts to introduce this phase as well as Kfeldspar into the fitting were unsuccessful. Other contributions to the misfit, such as the occurrence of smectite with mixed 1W/2W interlayer hydration cannot be ruled out. The final fitting converged to Rwp = 5.50 %, Rpb = 7.42 %, R = 5.52 %, Rwp = 7.48 %, Rexp = 2.84 %, c2 = 7.54. The refined values of the lattice constants are: a = 5.3430(5) Å, b = 9.2419(5) Å, c = 15.3696(2) Å. Note that the expected a = b/Ö3 condition for trioctahedral layer (Brindley 1980) is fulfilled within the experimental error. The phase fractions are: sauconite 81.7(7) %, kaolinite 7.8(4) %, dioctahedral smectite 10.4(5) %, and traces of quartz, confirming that sauconite is the major component in the mixture. The overall quality of the fitting can be considered satisfactory, considering that we are dealing with a mixture consisting mostly of disordered phases, where parameter refinement has to proceed with caution, to facilitate convergence and to get stable refinements with structurally This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 19 meaningful parameters. Similar procedures are commonly adopted when the Rietveld technique is applied to such complex patterns also for quantitative phase analyses (see for instance Lutterotti et al. 2010; Ufer et al. 2012, 2015). IMPLICATIONS This paper contributes to the improvement of the knowledge on a class of Zn-bearing minerals, whose properties are of relevance not only to economic geologists, but also to exploration geochemists and environment scientists. The Zn availability has indeed an impact not only at a global industrial level, to develop proper treatments and extraction processes for both primary ore and secondary materials (see, for example, Cole and Sole 2003; Boni et al. 2009a; Choulet et al. 2016), but also in terms of bioavailability considering that Zn is a micronutrient or a toxin to plant and animal life (McPhail et al. 2003). The above issues, among others, benefit from a deeper understanding of the ore mineral types and of the crystal chemical behavior of valuable metals, such as Zn. The employed multi-methodical approach provides some crystal chemical insights of the sauconite structure, based on the octahedral sheet almost entirely occupied by Zn ions and an interlayer region occupied by alkali cations surrounded by two water planes parallel to the octahedral-tetrahedral layers. Consistently, CEC results show that Ca and Mg ions, in small amounts, are hosted in the interlayer sites, whereas Zn ions are located within the octahedral site of sauconite. Therefore, an adsorption mechanism of the Zn2+ cations at the clay surface or fixing of the Zn2+ cations in the interlayer is negligible or lacking (see also Choulet et al. 2016). On considering that the natural sample at hand is not pure and consists of a number of disordered phases, the findings provided here can be considered a further contribution demonstrating that the approach developed in Ufer et al. (2004) can be successfully applied to increasingly challenging real cases, provided that cross-checking with independent techniques is This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 20 also used. In addition, the present work can have implications for a better comprehension of the hydration properties of trioctahedral smectites. The latter topic, related to the assessment of the interlayer cations and the interlayer water organization is widely recognized as crucial in determining the swelling/collapsing behavior in smectite with high impact both in the field of civil engineering and soil science and, more generally, in the environmental sciences, since the smectite hydration properties, often characterized by heterogeneity at different scales, may affect the fate/transfer properties of pollutants in natural media (see for instance Ferrage 2016 and references therein). ACKNOWLEDGMENTS The XRPD laboratory at the Dipartimento di Scienze della Terra and Geoambientali, University of Bari “Aldo Moro”, was funded by Potenziamento Strutturale PONa3_00369 “Laboratorio per lo Sviluppo Integrato delle Scienze e delle TEcnologie dei Materiali Avanzati e per dispositivi innovativi (SISTEMA)”. The XRPD facility at the DiSTAR, University “Federico II” Napoli, is acknowledged. The authors acknowledge Annett Steudel and Katja Emmerich for CEC measurement at the Competence Center for Material Moisture, University of Karlshrue, Germany. GB and NM thank Maria Boni, invaluable scientific guide who promoted the study of the economic geology of worldwide nonsulfide ore deposits, and G. Arfè who firstly characterized the Skorpion samples and for continuous help and fruitful discussions. This work was partly supported by DiSTAR fund 2017 (University of Naples Federico II, Italy) granted to G. Balassone, as well as by the research projects PGC2018-094573-B-100 from the Spanish Government and the Research Group RNM-179 of the Junta de Andalucía granted to F. Nieto. Two anonymous referees are thanked for insightful comments that helped to improve the quality of the manuscript. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 21 REFERENCES CITED Abad, I., Gutiérrez-Alonso, G., Nieto, F., Gertner, I., Becker, A., and Cabero, A. (2003) The structure and the phyllosilicates (chemistry, crystallinity and texture) of Talas Ala-Tau (Tien Shan, Kyrgyz Republic): comparison with more recent subduction complexes. Tectonophysics, 365, 1-4, 103-127. Abkhoshk, E., Jorjani, E., Al-Harahsheh, M.S., Rashchi, F., and Naazeri, M. (2014) Review of the hydrometallurgical processing of non-sulfide zinc ores. Hydrometallurgy, 149, 153-167. Arfè, G., Boni, M., Balassone, G., Mondillo, N., Hinder G., and Joachimski, M. (2017a) New C-O isotopic data on supergene minerals from the Skorpion and Rosh Pinah ore deposits (Namibia): genetic and palaeoclimatic constraints. Journal of African Earth Sciences, 126,148-158. Arfé, G., Mondillo, N., Balassone, G., Boni, M., Cappelletti, P., and Di Palma, T. (2017b) Identification of Zn-bearing micas and clays from the Cristal and Mina Grande zinc deposits (Bongará Province, Amazonas Region, Northern Peru). Minerals, 7(11),214. Arfè, G., Mondillo, N., Boni, M., Balassone, G., Joachimski, M., Mormone, A., and Di Palma, T. (2017c) The karst-hosted Mina Grande nonsulfide zinc deposit, Bongará district (Amazonas region, Peru). Economic Geology, 112, 1089-1110. Balan, E., Saitta, A.M., Mauri, F., and Calas, G. (2001) First-principles modeling of the infrared spectrum of kaolinite. American Mineralogist, 86, 1321-1330. Balassone, G., Rossi, M., Boni, M., Stanley, G., and McDermott, P. (2008) Mineralogical and geochemical characterization of nonsulfide Zn–Pb mineralization at Silvermines and Galmoy (Irish Midlands). Ore Geology Reviews, 33, 168-186. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 22 Balassone, G., Nieto, F., Arfè, G., Boni, M., and Mondillo, N. (2017) Zn-clay minerals in the Skorpion Zn nonsulfide deposit (Namibia): Identification and genetic clues revealed by HRTEM and AEM study. Applied Clay Science, 150, 309-322. Balassone, G., Scognamiglio, V., Nieto, F., Mondillo, N. Boni, M., Cappelletti, P., and Arfè, G. (2020) The nature of Zn-phyllosilicates in the nonsulfide Mina Grande and Cristal zinc deposits (Bongará district, northern Peru): the TEM-HRTEM and AEM perspective. American Mineralogist, in press. Balderman, A., Landler, A., Mittermayr, F., Letofsky-Papst, I., Steind, F., Galan, I., and Dietzel, M. (2019) Removal of heavy metals (Co, Cr, and Zn) during calcium–aluminium–silicate– hydrate and trioctahedral smectite formation. Journal of Materials Science, 54, 9331-9351. Bergmann, J., Friedel, P., and Kleeberg, R. (1998) BGMN - a new fundamental parameter based Rietveld program for laboratory X-ray sources, its use in quantitative analysis and structure investigations. Commission on Powder Diffraction Newsletter, 20, 5-8. Bergmann, J., Monecke, T., and Kleeberg, R. (2001) Alternative algorithm for the correction of preferred orientation in Rietveld analysis. Journal of Applied Crystallography, 34, 16-19. Beyer, J., and von Reichenbach, H. G. (1998) Dehydration and rehydration of vermiculites: IV. Arrangements of interlayer components in the 1.43 nm and 1.38 nm hydrates of Mgvermiculite. Zeitschrift für Physikalische Chemie, 207, 67-82. Bish, D.L., and Von Dreele, R.B. (1989) Rietveld refinement of non-hydrogen atomic positions in kaolinite. Clays and Clay Minerals, 37, 289–296. Bishop, J.L., Pieters, C.M., and Edwards, J.O. (1994) Infrared spectroscopic analyses on the nature of water in montmorillonite. Clays and Clay Minerals, 42(6), 702-716. Boni, M., and Mondillo, N. (2015) The "Calamines" and the "Others": the great family of supergene nonsulfide zinc ores. Ore Geology Reviews, 67, 208-233. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 23 Boni, M., Schmidt, P.R., DeWet, J.R., Singleton, J.D., Balassone, G., and Mondillo, N. (2009a) Mineralogical signature of nonsulfide zinc ores at Accha (Peru): a key for recovery. International Journal of Mineral Processing, 93, 267-277. Boni, M., Balassone, G., Arseneau, V., and Schmidt, P. (2009b) The nonsulfide zinc deposit at Accha (Southern Peru): geological and mineralogical characterization. Economic Geology, 104, 267-289. Borg, G., Kärner, K., Buxton, M., Armstrong, R., and van der Merwe, S.W. (2003) Geology of the Skorpion supergene zinc deposit, southern Namibia. Economic Geology, 98(4), 749-771. Bradbury, M.H., and Baeyens, B. (1999) Modelling the sorption of Zn and Ni on Camontmorillonite. Geochimica and Cosmochimica Acta, 63, 325-336. Bray, H.J., Redfern S.A.T., and Clark, S.M. (1998) The kinetics of dehydration in Camontmorillonite: an in situ X-ray diffraction study. Mineralogical Magazine, 62(5), 647-656. Brigatti, M.F., Galan, E., and Theng, B.K.G. (2006) Structure and mineralogy of clay minerals. In F. Bergaya, B.K.G. Theng, and G. Lagaly, Eds., Handbook of Clay Science, Developments in Clay Science, 1, p. 19-86. Elsevier, Amsterdam. Brigatti, M.F., Malferrari, D., Laurora, A., and Elmi, C. (2011) Structure and mineralogy of layer silicates: recent perspectives and new trends. In M.F. and A. Mottana, Eds., Layered mineral structures and their Application in Advanced Technologies, EMU notes in Mineralogy, 11, p. 1-71. The Mineralogical Society of Great Britain and Ireland. Brindley, G.W. (1980) Order-disorder in clay mineral structures. In G.W. Brindley and G. Brown, Eds., Crystal structures of clay minerals and their X-ray identification, 2, p. 125-195. Mineralogical Society of Great Britain and Ireland. Buatier, M., Choulet, F., Petit, S., Chassagnon, R., and Vennemann, T. (2016) Nature and origin of natural Zn clay minerals from the BouArhous Zn ore deposit: Evidence from electron This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 24 microscopy (SEM-TEM) and stable isotope compositions (H and O). Applied Clay Science, 132, 377-390. Carniato, F., Gatti, G., and Bisio, C. (2020) An overview of the recent synthesis and functionalization methods of saponite clay. New Journal of Chemistry, DOI:10.1039/D0NJ00253D Castellini, E., Malferrari, D., Bernini, F., Brigatti, M.F., Castro, G.F., Medici, L., Mucci, A. and Borsari, M. (2017) Baseline studies of the clay minerals society source clay montmorillonite STx-1b. Clays and Clay Minerals, 65, 4, 220-233. Choulet, F., Buatier, M., Barbanson, L., Guégan, R., and Ennaciri, A. (2016) Zinc-rich clays in supergene non-sulfide zinc deposits. Mineral Deposita, 51, 467-490. Churakov, S.V., and Dähn, R. (2012) Zinc adsorption on clays inferred from atomistic simulations and EXAFS spectroscopy. Environmental Science & Technology, 46, 5713-5719. Churchman, J., Gates, W., Yuan, G., and Theng, B.K. (2006) Clays and clay minerals for pollution control. In F. Bergaya, B.K.G. Theng and G. Lagaly, Eds., Handbook of Clay Science, Developments in Clay Science, 1, p. 625-675. Elsevier, Amsterdam. Cliff, G., and Lorimer, G.W. (1975) The quantitative analysis of thin specimens. Journal of Microscopy, 103, 203-207. Cole, P.M. and Sole, K.C. (2003) Zinc solvent extraction in the process industries. Mineral Processing & Extractive Metallurgy Review, 24, 91-137. Coppola, V., Boni, M., Gilg, H.A., Balassone, G., and Dejonghe, L. (2008) The “Calamine” nonsulfide Zn-Pb deposits of Belgium: petrographical, mineralogical and geochemical characterization. Ore Geology Reviews, 33, 187-210. D'Amour, H., Denner, W., and Schulz, H. (1979) Structure determination of a-quartz up to 68*108 Pa. Acta Crystallographica B, 35, 550-555. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 25 Dazas, B., Lanson, B., Breu, J., Robert, J.L., Pelletier, M., and Ferrage, E. (2013) Smectite fluorination and its impact on interlayer water content and structure: a way to fine tune the hydrophilicity of clay surfaces? Microporous and Mesoporous Materials, 181, 233-247. de Wet, J.R., and Singleton, J.D. (2008) Development of a viable process for the recovery of zinc from oxide ores. Journal of the Southern African Institute of Mining and Metallurgy, 108(5), 253-259. Decarreau, A., Colin, F., Herbillon A., Manceau, A., Nahon, D., Paquet, H., Trauth-Badeaud, D. andTrescases, J.J. (1987) Domain segregation in Ni-Fe-Mg-smectites. Clays and Clay Minerals, 35, 1-10. Doebelin, N., and Kleeberg, R. (2015) Profex: A graphical user interface for the Rietveld refinement program BGMN. Journal of Applied Crystallography, 48, 1573-1580. Drits, V.A., and Zviagina, B.B. (2009) Trans-vacant and cis-vacant 2:1 layer silicates: structural features, identification and occurrence. Clays and Clay Minerals, 57, 405-415. Drits, V.A., and Tchoubar, C. (1990) The modelization method in the determination of the structural characteristics of some layer silicates: internal structure of the layers, nature and distribution of stacking faults. In V.A. Drits and C. Tchoubar, Eds., X-Ray Diffraction by Disordered Lamellar Structures, p. 233-303. Springer-Verlag, Berlin. Drits, V.A., Plançon, A., Sakharov, B.A., Besson, G., Tsipurski, S.I., and Tchoubar, C. (1984) Diffraction effects calculated for structural models of K-saturated montmorillonite containing different types of defects. Clay Minerals, 19, 541-561. Eisenberg, D., and Kauzmann, W. (1969) The structure and properties of water, 296 p. Oxford University Press, New York. Emselle, N., McPhail, D.C., and Welch, S.A. (2005) Reliance, Flinders Ranges: mineralogy, geochemistry and zinc dispersion around a nonsulfide orebody. In C. Roach, Ed., Regolith This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 32 Figure captions Figure 1. Ternary diagram after Choulet et al. (2016). 4Si: Si/4; 3R2+: R2+/3; M+: M+ + 2M2+. Note that 3R2+ here is more general than in Choulet et al. (2016), who set 3R2+ = (Fe + Mg + Zn)/3. Symbols: full black square indicate the SK8 sample from Balassone et al. (2017). Open symbols for literature sauconites (square: SK1 and SK3 samples from Balassone et al. (2017); circle: Choulet et al. (2016); green triangle pointing upward: Kaufhold et al. (2015); triangle pointing downward: Mondillo et al. (2015); diamond: Mondillo et al. (2014); diamond with horizontal bar inside: Boniet et al. (2009); square with horizontal bar inside: Kärner (2006); circle with horizontal bar inside: Faust (1951); red triangle pointing upward with horizontal bar inside: Ross (1946). Figure 2. Random mount XRPD pattern of the sample SK8. The inset showsan enlargement of the 10-65° 2q range. Sau = sauconite. Figure 3. XRPD patterns of the air dried (AD, black line), ethylene glycol treated (EG, red line), heated at 400°C (green line), heated at 550°C (blue line) sauconite. In the inset: enlargement of the 9-40° 2q range with the value of the d-spacing of the sauconite 002 reflection and the position of the kaolinite (K) basal reflection in the EG pattern. Figure 4. XRPD patterns in 4-69° 2q range of sauconite after heating at 900°C. Sau = sauconite; Crist = cristobalite; Wil = willemite. Figure 5. TGA (blue line), DTA (black line) and DTG (red line) curves of SK8 sauconite measured in air. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld 33 Figure 6. Room temperature infrared spectrum of SK8 sauconite in the range 4000-450 cm-1. Figure 7. Results of fitting of the OH stretching region for the SK8 sauconite. Figure 8. (a) TEM textural image of CCP sauconite SK8, with a SAED pattern of the investigated area in the inset; (b) textural image of CCP sauconite locally with a PCP character; (c) lattice fringe image of the area in the yellow frame of (b), showing small disoriented packets, locally vanishing into each other; (d) CCP textural image with the yellow frame corresponding to image (a); (e) chemical mapping the same area as in (d). Figure 9. XRPD modeling of the Skorpion sauconite. (a) Fitting in the 4-70° 2θ range; (b) enlargement of the fitting in the 10-70° 2θ range. Data color legend: black, observed; red, calculated; green, sauconite; cyano, kaolinite; magenta, dioctahedral smectite; brown, quartz; yellow, background; blue, difference. This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld Table 1. Values of the CEC and of the exchangeable cations (cmol+/kg) for the studied sauconite. Aliquot 1 Aliquot 2 Sample weight (g) 0.0432 0.0392 CEC 80.93 80.21 Ca 55.2 Mg 20.7 Na 3.3 K 0.5 Zn 0.5 This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld Table 2. Data (°C) from the DTA curve of the SK8 and literature sauconite. This study Kaufhold et al. (2015) Petit et al. (2008) Higashi et al. (2002) Faust (1951) Phenomena ZnS-1 (<0.5 μm) Synthetic sample c-134 c-80 c-81 c-567 100 120 63-83 90 Sauconite dehydration (absorbed H2O) 178 170 165 158 Sauconite dehydration (absorbed and interlayer H2O) 190 246 289 270 273 261 Sauconite dehydration (interlayer H2O) 305* 353* - 330* Oxydation of organic matter or effect of gibbsite 380** Organic matter decarboxylation 441 Sauconite dehydration (absorbed H2O); dehydroxylation 500** 500 Kaolinite dehydroxylation. Oxydation of organic matter or effect of gibbsite 582 601 569 Decomposition of halloysite 550-700 650 550-615 615 734 728 729 682 Sauconite dehydroxylation 820* 800* 765-781 770* 832* 818* 895* 836* Breakdown of sauconite structure and recrystallization 900* 952-968 964* 978* 940* 1029* Breakdown of sauconite structure and recrystallization Note: *Exothermal peak.**In the 380-500°C range, and, in general at T<600°C, contribution due to dehydroxylation of 1:1 and/or 2:1 trans vacant clay mineral may occur (see text). This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld Table 3. Position (cm-1) and assignment for the bands in the infrared spectrum of the SK8 and literature sauconite. This study Kaufhold et al. (2015) Petit et al. (2008) Yokoyama et al. (2006) Higashi et al. (2002) Band assignment Band cm-1 A 3698 3698 Kaolinite B 3648 3640 3643 3635 3641 ZnZnZn-OH C 3620 3623 AlAl-OH of kaolinite; 2:1 dioctahedral phases 3600-3200 ν3 H2O D 3585 H2O of inner sphere; 2:1 dioctahedral phases E 3551 3558 FeFeZn-OH or MgMgZn-OH or FeMgZn-OH 3530 FeFeZn-OH or MgMgZnOH or FeMgZn-OH or FeFe-OH MgZn-OH F 3440 3442 Absorbed H2O; Gibbsite band 3375 Gibbsite band G 3230 H2O bending overtone 1633 1600 1600 Interlayer H2O bending Note: Assignment for this study in bold (see text) This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld This is the peer-reviewed, final accepted version for American Mineralogist, published by the Mineralogical Society of America. The published version is subject to change. Cite as Authors (Year) Title. American Mineralogist, in press. DOI: https://doi.org/10.2138/am-2020-7460. http://www.minsocam.org/ Always consult and cite the final, published document. See http:/www.minsocam.org or GeoscienceWorld