Advances and Opportunities in Ore Mineralogy

The study of ore minerals is rapidly transforming due to an explosion of new microand nano-analytical technologies. These advanced microbeam techniques can expose the physical and chemical character of ore minerals at ever-better spatial resolution and analytical precision. The insights that can be obtained from ten of today’s most important, or emerging, techniques and methodologies are reviewed: laser-ablation inductively-coupled plasma mass spectrometry; focussed ion beam-scanning electron microscopy; high-angle annular dark field scanning transmission electron microscopy; electron back-scatter diffraction; synchrotron X-ray fluorescence mapping; automated mineral analysis (Quantitative Evaluation of Mineralogy via Scanning Electron Microscopy and Mineral Liberation Analysis); nanoscale secondary ion mass spectrometry; atom probe tomography; radioisotope geochronology using ore minerals; and, non-traditional stable isotopes. Many of these technical advances cut across conceptual boundaries between mineralogy and geochemistry and require an in-depth knowledge of the material that is being analysed. These technological advances are accompanied by changing approaches to ore mineralogy: the increased focus on trace element distributions; the challenges offered by nanoscale characterisation; and the recognition of the critical petrogenetic information in gangue minerals, and, thus the need to for a holistic approach to the characterization of mineral assemblages. Using original examples, with an emphasis on iron oxide-copper-gold deposits, we show how increased analytical capabilities, particularly imaging and chemical mapping at the nanoscale, offer the potential to resolve outstanding questions in ore mineralogy. Broad regional or deposit-scale genetic models can be validated or refuted by careful analysis at the smallest scales of observation. As the volume of information at different scales of observation expands, the level of complexity that is revealed will increase, in turn generating additional research questions. Topics that are likely to be a focus of breakthrough research over the coming decades include, understanding atomic-scale distributions of metals and the role of nanoparticles, as well how minerals adapt, at the lattice-scale, to changing physicochemical conditions. Most importantly, the complementary use of advanced microbeam techniques allows for information of different types and levels of quantification on the same materials to be correlated.

to an explosion of different, complementary in-situ microanalytical capabilities. The rapid expansion of microanalysis, with improved resolution and sensitivity, enables the ore mineralogist to observe, measure and quantify in ways that were not previously possible.
We summarise contemporary research topics from ten analytical methodologies, either well established, or currently emerging, that enable a greater understanding of the structural, physical, and chemical characteristics of many of the more common ore minerals. In most cases, these methods are routinely applied to resolve broader problems in mineralogy, but have proven particularly valuable for the study of ore minerals. We include references to selected publications that have demonstrated both of the types of application for each method and their future potential.

Laser-Ablation Inductively-Coupled Plasma Mass Spectrometry (LA-ICP-MS)
LA-ICP-MS [9,10] has rapidly established itself as a requisite research tool with application across the earth sciences. For ore mineralogists, LA-ICP-MS represents a rapid and relatively inexpensive method for accurate, precise in-situ measurement of concentrations of trace elements in mineral matrices, and to map the distributions of those elements at the grain-scale. This, in turn, enables genetic models for mineral deposits to be refined. Large LA-ICP-MS datasets can form a foundation for understanding trace element incorporation and release from ore minerals, as well as for the development of geochemical vectors of value in mineral exploration. Coupled with automated mineralogy (see below), LA-ICP-MS allows for reliable deportment models to be generated for elements of interest. Moreover, LA-ICP-MS also offers a platform for in-situ geochronology [11], and in-situ chemical characterization of fluid inclusions [12]. Cook et al. (2016) [13] provide an overview of current applications of LA-ICP-MS to the trace element analysis of minerals in hydrothermal ore deposits, also noting research gaps, some of the outstanding issues, and future opportunities. The rapidly expanding capabilities of current LA-ICP-MS instrumentation, including both quadrupole and multi-collector systems, include in-situ determination of stable isotopes, and grain-scale mapping of isotopic signatures [14][15][16][17][18]. Outstanding challenges include inadequately constrained issues of down-hole fractionation, the impact of matrix effects, and concerns about how the behaviour of certain minerals during ablation may induce fractionation (e.g., [19]). To this, we add the longstanding question of how LA-ICP-MS data from fine-grained, heterogeneous, complex ores can best be interpreted, when the sample volumes analysed by LA-ICP-MS contain measured elements that are in solid solution within the host, as well as sub-micron-sized mineral inclusions, which, in many cases, are just a few nanometres across ( Figure 1).  FIB-SEM methods have helped resolve problems that are associated with in-situ extraction of small amounts of sample material that has been well-characterised at the micron-scale, e.g., by EPMA or LA-ICP-MS, from the surface of a polished mount for milling and subsequent analysis by transmission electron microscopy ( [20][21][22][23], and references therein). Direct application of FIB-SEM methods in ore mineralogy is expanding steadily ( [22][23][24][25][26][27][28][29][30][31][32][33], and references therein). The technique has revolutionized mineral characterization by enabling scales of observation to be bridged, i.e., linking observations at the nanometre-scale to features observable at the scale of microns in reflected light or by SEM. In tomographic analysis mode, the FIB-SEM provides three-dimensional visualisation of a sample with resolution well below the micron-scale (e.g., [34]). Figure 2 illustrates an example in which FIB-SEM has been used to visualise and extract thin foils in-situ from iron-oxides to examine and identify nanoscale inclusions using high-angle angular dark field scanning transmission electron microscopy (HAADF STEM). The magnetite illustrated occurs as a magmatic accessory in the host granite (Roxby Downs Granite; RDG) to the Olympic Dam Cu-U-Au-Ag deposit. In general, magmatic magnetite in the RDG is associated with ilmenite, apatite and zircon, as well as minor titanite. The binary inclusions that are shown here were observed at the micron scale only for a sub-population of magnetite grains. Nanoscale investigation shows they are however ubiquitously present throughout the magnetite although they vary in size down to the smallest nanoparticles. The presence of such inclusions in RDG magnetite is taken as further evidence for mixing between granitic and mafic magmas, as was suggested by studies of apatite that contain pyrrhotite-fluorite inclusions [35], relationships among feldspars in the RDG [36], and indeed observations on intrusive rocks from other granitic plutons elsewhere (e.g., [37]). FIB-SEM methods have helped resolve problems that are associated with in-situ extraction of small amounts of sample material that has been well-characterised at the micron-scale, e.g., by EPMA or LA-ICP-MS, from the surface of a polished mount for milling and subsequent analysis by transmission electron microscopy ( [20][21][22][23], and references therein). Direct application of FIB-SEM methods in ore mineralogy is expanding steadily ( [22][23][24][25][26][27][28][29][30][31][32][33], and references therein). The technique has revolutionized mineral characterization by enabling scales of observation to be bridged, i.e., linking observations at the nanometre-scale to features observable at the scale of microns in reflected light or by SEM. In tomographic analysis mode, the FIB-SEM provides three-dimensional visualisation of a sample with resolution well below the micron-scale (e.g., [34]). Figure 2 illustrates an example in which FIB-SEM has been used to visualise and extract thin foils in-situ from iron-oxides to examine and identify nanoscale inclusions using high-angle angular dark field scanning transmission electron microscopy (HAADF STEM). The magnetite illustrated occurs as a magmatic accessory in the host granite (Roxby Downs Granite; RDG) to the Olympic Dam Cu-U-Au-Ag deposit. In general, magmatic magnetite in the RDG is associated with ilmenite, apatite and zircon, as well as minor titanite. The binary inclusions that are shown here were observed at the micron scale only for a sub-population of magnetite grains. Nanoscale investigation shows they are however ubiquitously present throughout the magnetite although they vary in size down to the smallest nanoparticles. The presence of such inclusions in RDG magnetite is taken as further evidence for mixing between granitic and mafic magmas, as was suggested by studies of apatite that contain pyrrhotite-fluorite inclusions [35], relationships among feldspars in the RDG [36], and indeed observations on intrusive rocks from other granitic plutons elsewhere (e.g., [37]).

HAADF STEM Imaging and STEM-EDS Mapping
Transmission electron microscopy has been a routine mineralogical technique dating back to the middle of the 20th Century, but continues to advance at an extraordinary rate [38]. The application of HAADF STEM, is, however, a more recent development. HAADF STEM enhances Z-scale contrast, enabling atomic-scale resolution with applicability to many minerals that are found within ore deposits, e.g., mixed-layer chalcogenide compounds of the tetradymite group [39], as shown in Figure 3.

HAADF STEM Imaging and STEM-EDS Mapping
Transmission electron microscopy has been a routine mineralogical technique dating back to the middle of the 20th Century, but continues to advance at an extraordinary rate [38]. The application of HAADF STEM, is, however, a more recent development. HAADF STEM enhances Z-scale contrast, enabling atomic-scale resolution with applicability to many minerals that are found within ore deposits, e.g., mixed-layer chalcogenide compounds of the tetradymite group [39], as shown in Figure 3.  The atomic-scale resolution possible with contemporary state-of-the-art STEM instrumentation in HAADF STEM mode is particularly well-suited to visualise the structures of complex sulphosalts e.g., Pb-Bi-sulphosalts [23], Zn-bearing smectites [41,42], or disordered intergrowths among mixedlayer compounds, such as the REE-fluorocarbonates [43]. Atomic-scale HAADF STEM imaging is a more direct tool than high-resolution TEM since it provides a better control on the correlation between crystal structures and chemical variation. It is therefore an ideal tool for resolving issues of polysomatism within mineral series, in which the crystal structural motifs or layer stacking change gradually. In combination with energy-dispersive X-ray (EDX) mapping, the HAADF STEM technique is well suited to the identification of nanoparticles, and even the smallest atomic clusters that are present within a mineral. The technique is particularly useful for heavier (U, Au, Te, Bi, Pb, REE etc. [23,43]), rather than lighter atoms. Remarkable insights have, however, been obtained by HAADF STEM into the nanoscale mechanisms responsible for precipitation of siliceous magnetite [44]. Electron energy loss spectroscopy (EELS) on a TEM platform is also an important technique that can be used to show the oxidation state of elements of interest within a nanoparticulate mineral and any variation in such oxidation state at the nanoscale, e.g., [45]. Based on our accumulated experience, HAADF STEM imaging works better on FIB-prepared foils (often with variable thickness and distortions) than HR TEM imaging. Future research topics with potential in ore mineralogy include: (i) trace element incorporation in minerals and their release; (ii) polysomatism in mixed-layer compounds; and, (iii) lattice-scale order/disorder as a tool to constrain equilibrium crystallisation versus mineral replacement.
To illustrate an application of HAADF STEM, we show, in Figure 4, an early hydrothermal association of hyalophane-albite within the core of a magmatic K-feldspar phenocryst (mm-sized), with rapakivi texture [36,46] from granite outside the orebody at Olympic Dam. The sample is considered "fresh", based on whole rock geochemical data analysis and the preservation of magmatic amphibole and biotite. In IOCG deposits, alteration is often defined based on mineral associations that are constrained regionally or at the scale of the deposit (e.g., [47]). This example shows, however, that intensive hydrothermal alteration with the development of calc-silicates and their replacement (e.g., chlorite replacing epidote, as shown here) can occur at the nanoscale, even though magmatic features (e.g., cryptoperthites) are preserved. The assemblage also contains abundant, nm-scale inclusions of hematite and larger grains of sericite epitaxially intergrown with chlorite (also see [46]). The atomic-scale resolution possible with contemporary state-of-the-art STEM instrumentation in HAADF STEM mode is particularly well-suited to visualise the structures of complex sulphosalts e.g., Pb-Bi-sulphosalts [23], Zn-bearing smectites [41,42], or disordered intergrowths among mixed-layer compounds, such as the REE-fluorocarbonates [43]. Atomic-scale HAADF STEM imaging is a more direct tool than high-resolution TEM since it provides a better control on the correlation between crystal structures and chemical variation. It is therefore an ideal tool for resolving issues of polysomatism within mineral series, in which the crystal structural motifs or layer stacking change gradually. In combination with energy-dispersive X-ray (EDX) mapping, the HAADF STEM technique is well suited to the identification of nanoparticles, and even the smallest atomic clusters that are present within a mineral. The technique is particularly useful for heavier (U, Au, Te, Bi, Pb, REE etc. [23,43]), rather than lighter atoms. Remarkable insights have, however, been obtained by HAADF STEM into the nanoscale mechanisms responsible for precipitation of siliceous magnetite [44]. Electron energy loss spectroscopy (EELS) on a TEM platform is also an important technique that can be used to show the oxidation state of elements of interest within a nanoparticulate mineral and any variation in such oxidation state at the nanoscale, e.g., [45]. Based on our accumulated experience, HAADF STEM imaging works better on FIB-prepared foils (often with variable thickness and distortions) than HR TEM imaging. Future research topics with potential in ore mineralogy include: (i) trace element incorporation in minerals and their release; (ii) polysomatism in mixed-layer compounds; and, (iii) lattice-scale order/disorder as a tool to constrain equilibrium crystallisation versus mineral replacement.
To illustrate an application of HAADF STEM, we show, in Figure 4, an early hydrothermal association of hyalophane-albite within the core of a magmatic K-feldspar phenocryst (mm-sized), with rapakivi texture [36,46] from granite outside the orebody at Olympic Dam. The sample is considered "fresh", based on whole rock geochemical data analysis and the preservation of magmatic amphibole and biotite. In IOCG deposits, alteration is often defined based on mineral associations that are constrained regionally or at the scale of the deposit (e.g., [47]). This example shows, however, that intensive hydrothermal alteration with the development of calc-silicates and their replacement (e.g., chlorite replacing epidote, as shown here) can occur at the nanoscale, even though magmatic features (e.g., cryptoperthites) are preserved. The assemblage also contains abundant, nm-scale inclusions of hematite and larger grains of sericite epitaxially intergrown with chlorite (also see [46]). Such associations draw attention to the telescoping character of alteration stages in IOCG deposits, which are otherwise treated only as well-defined spatial-temporal stages. Moreover, these mineral reactions are driven by coupled dissolution-reprecipitation reaction (CDRR), and thus chemical exchange is trapped within areas ranging in scales from a few nm (as shown here) up to several meters. Considering that we show here a representative case for an IOCG deposit with high metal endowment, it may be that these types of reaction are paramount in contributing to the fertility of an IOCG system by confining the alteration within an area that will become the orebody. Such associations draw attention to the telescoping character of alteration stages in IOCG deposits, which are otherwise treated only as well-defined spatial-temporal stages. Moreover, these mineral reactions are driven by coupled dissolution-reprecipitation reaction (CDRR), and thus chemical exchange is trapped within areas ranging in scales from a few nm (as shown here) up to several meters. Considering that we show here a representative case for an IOCG deposit with high metal endowment, it may be that these types of reaction are paramount in contributing to the fertility of an IOCG system by confining the alteration within an area that will become the orebody. In Figure 5, we show HAADF STEM images and element maps of REE-fluorocarbonates from a locality adjacent to the Olympic Dam deposit. These are part of an assemblage containing sericite and molybdenite pseudomorphically replacing magmatic feldspar (plagioclase) in a red-stained Kfeldspar altered granite [36,46]. Bastnäsite [(REE)CO3F] and parisite [Ca(REE)2(CO3)3F2] are recognised as the main REE-species, and nanometre-sized hematite is seen as platelets throughout the parisite [46]. Atomic-scale HAADF STEM imaging was undertaken for the first time on such assemblages, and has shown the presence of a third species, "B2S", with composition between bastnäsite and parisite, as well as the changes between the minerals along directions that are perpendicular to the stacking sequences [43]. The element maps in Figure 5 show the changes in parisite-bastnäsite stacking sequences across a microfracture adjacent to areas where hematite is also present as epitaxial intergrowths with minerals of the bastnäsite-synchysite group. This type of assemblage is important for understanding development of the early stages of alteration associated with IOCG deposits, and particularly the link between REE-enrichment and the initial stages of Femetasomatism. As in the other example above (Figure 4), this reinforces the hypothesis that the earliest hydrothermal Fe-oxide formed at Olympic Dam is not necessarily magnetite, but hematite. This contradicts generic schemes of IOCG alteration invoking a so-called IOA-stage or "magnetiteapatite" assemblage, which is contended as typifying "early mineralisation" [48] at Olympic Dam, or indeed in IOCG and related "IOA" systems elsewhere (e.g., [49]). In Figure 5, we show HAADF STEM images and element maps of REE-fluorocarbonates from a locality adjacent to the Olympic Dam deposit. These are part of an assemblage containing sericite and molybdenite pseudomorphically replacing magmatic feldspar (plagioclase) in a red-stained K-feldspar altered granite [36,46]. Bastnäsite [(REE)CO 3 F] and parisite [Ca(REE) 2 (CO 3 ) 3 F 2 ] are recognised as the main REE-species, and nanometre-sized hematite is seen as platelets throughout the parisite [46]. Atomic-scale HAADF STEM imaging was undertaken for the first time on such assemblages, and has shown the presence of a third species, "B 2 S", with composition between bastnäsite and parisite, as well as the changes between the minerals along directions that are perpendicular to the stacking sequences [43]. The element maps in Figure 5 show the changes in parisite-bastnäsite stacking sequences across a microfracture adjacent to areas where hematite is also present as epitaxial intergrowths with minerals of the bastnäsite-synchysite group. This type of assemblage is important for understanding development of the early stages of alteration associated with IOCG deposits, and particularly the link between REE-enrichment and the initial stages of Fe-metasomatism. As in the other example above (Figure 4), this reinforces the hypothesis that the earliest hydrothermal Fe-oxide formed at Olympic Dam is not necessarily magnetite, but hematite. This contradicts generic schemes of IOCG alteration invoking a so-called IOA-stage or "magnetite-apatite" assemblage, which is contended as typifying "early mineralisation" [48] at Olympic Dam, or indeed in IOCG and related "IOA" systems elsewhere (e.g., [49]).

Electron Back-Scatter Diffraction (EBSD)
Since the technique first emerged in the mid-1990's, EBSD [50][51][52] has found wide application in mineralogy, with particular suitability to more refractory phases, such as zircon and garnet. Among the ore minerals, pyrite has been a major focus. Many studies used EBSD to target mechanisms of sulphide deformation at different conditions in metamorphic environments [53][54][55][56][57][58][59][60]. Other studies [61][62][63] used EBSD to address the growth of colloform and framboidal pyrite. Reddy and Hough (2013) [64] investigated Witwatersrand pyrite, linking trace element distribution patterns to sulphide deformation, and showed that some of the genetic controversies on the Witwatersrand gold ores can be explained with a model of gold mobilization before and after the re-deposition of pyrite. Iron-(Ti)oxides are also refractory minerals in which genetic and deformational history can be preserved. Several EBSD studies have been undertaken in rocks ranging from gabbros [65] and deformed granites [66] to marine sediments [67]. Barbosa and Lagoeiro, 2010 [68] used EBSD to address transformation between magnetite and hematite in samples from the Quadrilátero Ferrífero of Brazil, while Rosière et al. (2013) [69] assessed the interplay between deformation and mineralization in high-grade iron ores hosted by shear zones. Based on synthetic deformation experiments showing the crystallographic textures may be weak or slow to develop, caution has been raised [70] regarding the interpretation of magnetite deformation mechanisms. In recent years, however, researchers have begun to use EBSD to investigate a range of other ore minerals, such as pyrrhotite-pentlandite from komatiite-hosted ores [71], in which trace element concentrations were found to correlate with microstructures, indicating intragrain diffusion.
EBSD has also proven most helpful in examining potential variation in isotopic signature with crystal orientation of iron-oxides [72], rutile [73], or cassiterite [74]. Because electron backscatter patterns are generated from the uppermost few tens of nanometres of the sample surface, preparation and chemical polishing routines need to be optimised. In our experience, this can be problematic for some ore minerals or intergrowths thereof where conventional grinding and polishing causes damage to crystal lattices. Increasingly, EBSD is run on a FIB-SEM platform, offering high-resolution imaging and the potential of three-dimensional tomography (e.g., [75]). FIB-SEM based EBSD also offers the possibility of working on ion-milled surfaces, negating the need for chemical polishing. Investigating single crystals of uraninite by EBSD on a FIB platform, Macmillan et al. [76] addressed how lattice strain could be correlated with changes in chemical zonation patterns. Expanding on this work, the same authors examined mineral replacement between minerals with completely different

Electron Back-Scatter Diffraction (EBSD)
Since the technique first emerged in the mid-1990's, EBSD [50][51][52] has found wide application in mineralogy, with particular suitability to more refractory phases, such as zircon and garnet. Among the ore minerals, pyrite has been a major focus. Many studies used EBSD to target mechanisms of sulphide deformation at different conditions in metamorphic environments [53][54][55][56][57][58][59][60]. Other studies [61][62][63] used EBSD to address the growth of colloform and framboidal pyrite. Reddy and Hough (2013) [64] investigated Witwatersrand pyrite, linking trace element distribution patterns to sulphide deformation, and showed that some of the genetic controversies on the Witwatersrand gold ores can be explained with a model of gold mobilization before and after the re-deposition of pyrite. Iron-(Ti)-oxides are also refractory minerals in which genetic and deformational history can be preserved. Several EBSD studies have been undertaken in rocks ranging from gabbros [65] and deformed granites [66] to marine sediments [67]. Barbosa and Lagoeiro, 2010 [68] used EBSD to address transformation between magnetite and hematite in samples from the Quadrilátero Ferrífero of Brazil, while Rosière et al. (2013) [69] assessed the interplay between deformation and mineralization in high-grade iron ores hosted by shear zones. Based on synthetic deformation experiments showing the crystallographic textures may be weak or slow to develop, caution has been raised [70] regarding the interpretation of magnetite deformation mechanisms. In recent years, however, researchers have begun to use EBSD to investigate a range of other ore minerals, such as pyrrhotite-pentlandite from komatiite-hosted ores [71], in which trace element concentrations were found to correlate with microstructures, indicating intragrain diffusion.
EBSD has also proven most helpful in examining potential variation in isotopic signature with crystal orientation of iron-oxides [72], rutile [73], or cassiterite [74]. Because electron backscatter patterns are generated from the uppermost few tens of nanometres of the sample surface, preparation and chemical polishing routines need to be optimised. In our experience, this can be problematic for some ore minerals or intergrowths thereof where conventional grinding and polishing causes damage to crystal lattices. Increasingly, EBSD is run on a FIB-SEM platform, offering high-resolution imaging and the potential of three-dimensional tomography (e.g., [75]). FIB-SEM based EBSD also offers the possibility of working on ion-milled surfaces, negating the need for chemical polishing. Investigating single crystals of uraninite by EBSD on a FIB platform, Macmillan et al. [76] addressed how lattice strain could be correlated with changes in chemical zonation patterns. Expanding on this work, the same authors examined mineral replacement between minerals with completely different chemistry (uraninite and Cu-Fe-sulphides) via coherent processes coupling dissolution to reprecipitation ratios [77]. The use of the FIB-SEM platform for EBSD analysis can be followed by TEM work on thinned foils subsequently extracted from the same material.
2.1.5. Trace Element X-ray Fluorescence (SXRF) Mapping Using Synchrotron Radiation The resolving power of synchrotron radiation and increased availability of access to synchrotron facilities has generated new avenues for investigation of ore minerals. Megapixel high-definition trace element X-ray fluorescence (SXRF) mapping using the Maia detector [78][79][80][81], for example, enables the generation of maps with the dimensions of a full thin section or drillcore fragment at resolutions of a few µm in only a few hours, complementing EPMA or LA-ICP-MS analysis. Applications of SXRF mapping in ore mineralogy and petrology are given by [82][83][84]. We would also single out the recent use of synchrotron X-ray sources for the determination of the oxidation states of minor elements (~1000 ppm) in common sulphides via extended X-ray absorption fine structure (EXAFS) and X-ray absorption near edge structure (XANES). For example, Cook et al. [85] determined the presence of Cu + in indium-bearing sphalerite, confirming 2Zn 2+ ↔ Cu + + In 3+ coupled substitution. A second study [86] confirmed the presence of Ge 4+ in sphalerite, which is suggested to substitute for (Zn,Fe) 2+ with vacancies introduced to achieve charge balance. Most recently, Bonnet et al. [87] presented evidence for both Ge 2+ and Ge 4+ in sphalerite. Opportunities exist to use such synchrotron source methods to validate the oxidation states of other proposed substituents, e.g., Sn in common sulphides.

Automated Mineralogy
Automated mineralogy techniques on a SEM platform, notably Quantitative Evaluation of Mineralogy via Scanning Electron Microscopy (QEMSCAN) and Mineral Liberation Analysis (MLA), have proven invaluable for the investigation of ore samples ( [88], and references therein). QEMSCAN uses signals from the BSE and EDS detectors to identify each particle based on average atomic mass, and a comparison of the obtained image with a "mineral library" database to search for probable fits for the compound in question. MLA uses grey-scale BSE images as a primary method of mineral identification and is faster, since the dead time that is required for X-ray analysis is minimized. These methods allow for the determination of which minerals are present, and in which proportions, their grain size distributions and associations. MLA data is especially useful where sample representivity is paramount.
Automated mineralogy has thus found widespread application in quantitative geometallurgy (e.g., [89][90][91]) and routine MLA analysis on large numbers of samples is an essential component of validated deposit-scale zonation models, e.g., for Cu-Fe-sulphide species at Olympic Dam [92]. Automated mineralogy is increasingly used for the identification of unconventional or less common ore minerals and to understand the distributions of minor and trace elements. For example, Bachmann et al. (2017) [93] describe new strategies, combining MLA and EPMA, for the systematic detection and identification of small particles containing the critical element indium. Correct interpretation of automated mineralogy data does, however, require a sound mineralogical understanding of the material, and thus should not be considered as a direct substitute for that. MLA may also not necessarily detect phases present as smallest phases, or resolve subtle compositional variations.

Nanoscale Secondary Ion Mass Spectrometry (nanoSIMS)
NanoSIMS [94] offers in-situ visualisation of mass abundances in solid materials with chemical imaging down to the nanoscale. There are currently~40 nanoSIMS instruments worldwide, and the technique is proving extraordinarily valuable for understanding the internal structure and composition of ore minerals. Among ore minerals, pyrite and arsenopyrite have been a major focus in published work, particularly with respect to their capacity for hosting both nanoparticulate and solid solution gold [95][96][97]. Opportunity exists to use nanoSIMS to address other outstanding issues in the mineralogy of compositionally-zoned minerals, trace element partitioning, and in mineral processing. In one innovative application of nanoSIMS, Wacey et al. (2015) [98] have used nanoSIMS mapping of organic carbon and nitrogen in framboidal pyrite, elucidating mechanisms of framboid formation in terms of organic biofilms.
NanoSIMS is proving a frontier technique for tracking the grain-scale distributions of daughter products of uranium decay in ores [99]. This approach has shown both the versatility and extreme sensitivity of the method, even at extremely low concentrations, and points to the future promise of nanoSIMS to constrain element and isotope mobility at the grain-scale, with major implications for improved microbeam geochronology. NanoSIMS has been used to map trace element distributions in minerals that are zoned at the nanoscale, including datable minerals such as zircon or monazite [100,101].

Atom Probe Tomography (APT)
APT has, until recently, been little applied to ore minerals, despite being around for more than a decade and widely used by material scientists and others to map the spatial distribution of atomic species in three dimensions and at the nanoscale [102]. APT couples chemical sensitivity and three-dimensional capabilities with a spatial resolution near the atomic scale to provide unique information about nanoscale structures. The technique has been demonstrated as a valuable tool to complement SIMS and allow for smallest-scale geochronology (e.g., [103]), to understand nanoscale diffusion within zircon [104,105], and for the examination of homogeneity within geochronological reference materials [106].
Two recent papers have shown the potential that this unique method may have in mineralogy. Weber et al. (2016) [107] used APT to demonstrate a hitherto-unrecognised layering of nanoscale pore structures in barites, which carries implications for the observed reactivity of barite and rapid uptake of Ra. In another APT study, Parman et al. (2015) [108] address natural isoferroplatinum (Pt 3 Fe), showing potential order-disorder distribution of trace elements. Both these papers highlight the considerable potential of APT as a breakthrough method that may help to resolve outstanding issues that connect structural features (as observed by TEM) with geochemical features. Introducing APT to a mineralogical audience, Parman (2015) [109] notes that the method "forces one to merge mineralogy and geochemistry", and that APT "should stimulate a range of new discoveries".

Geochronology Using Ore Minerals
The accurate determination of the timing of mineralization has been a long-standing issue in mineral deposit studies. Effective dating is frequently hampered by a mismatch between those minerals that can be accurately and reliably dated, and those which are abundant in ores and can be petrographically constrained as part of, and therefore contemporaneous with mineralization. Molybdenite Re-Os geochronology [110] has proven an invaluable tool to navigate this impasse, and has been widely applied to the many cases where molybdenite is present within the ore assemblage. More recently, other non-traditional minerals have been shown to have utility in geochronology, notably hematite, which, in some ore systems, readily incorporates U and radiogenic Pb into its crystal structure [111][112][113]. These studies have shown that hematite can be accurately and precisely dated by LA-ICP-MS, sensitive high-resolution ion microprobe (SHRIMP), or isotope dilution thermal ionization mass spectrometry (ID-TIMS). Moreover, hematite (and magnetite) appears remarkably robust and do not undergo extensive metamictization. U-Pb dating of iron-oxides by LA-ICP-MS can provide rapid, reliable results, at a sufficient resolution to be of value in reconnaissance studies.
Geochronological data can also be obtained on several other ore minerals, including columbite-tantalite, rutile, thorite and uraninite by LA-ICP-MS [11]. LA-ICP-MS "age" mapping of radiogenic isotopes by rastering across a selected area of a grain using appropriate reference standards in zircon [114], or hematite [115], as demonstrated in Figure 6. Such methods also have potential for the assessment of homogeneity in minerals with zonation patters relative to U and radiogenic Pb. As mentioned above, the ability to investigate minerals and visualise element and isotope distributions at the nanoscale, using LA-ICP-MS mapping, nanoSIMS, APT, and other techniques, with complementary structural information obtained via EBSD, have highlighted the heterogeneity present in natural samples and the possibility of intra-grain diffusion, as well as the need to ensure that reference standards are as homogeneous as possible.
Minerals 2017, 7, 233 10 of 24 at the nanoscale, using LA-ICP-MS mapping, nanoSIMS, APT, and other techniques, with complementary structural information obtained via EBSD, have highlighted the heterogeneity present in natural samples and the possibility of intra-grain diffusion, as well as the need to ensure that reference standards are as homogeneous as possible.

Non-Traditional Stable Isotopes
The growing interest in in-situ analysis of "traditional" stable isotopes notwithstanding (e.g., sulphur [116], development and wider availability of laser-ablation multi-collector (LA-MC)-ICP-MS instrumentation has catalysed a rapid expansion in analysis of non-traditional stable isotopes, including the isotopes of several elements concentrated within ores and measured in ore minerals. These include iron, nickel, copper, zinc, germanium, selenium, and molybdenum. Several other isotopic systems with potential application to understanding the evolution of ore deposits are rapidly emerging. These include thallium, tellurium, and mercury. Table 1 briefly reviews current research on these isotopes in ores and the topics that are being addressed. Table 1. Applications of stable isotope geochemistry in ore mineralogy.
Tl [149] and references therein, [150][151][152][153] Significant isotopic variability in natural samples points to the potential of Tl-isotopes for understanding formation of hydrothermal deposits; more baseline data are needed.

Non-Traditional Stable Isotopes
The growing interest in in-situ analysis of "traditional" stable isotopes notwithstanding (e.g., sulphur [116], development and wider availability of laser-ablation multi-collector (LA-MC)-ICP-MS instrumentation has catalysed a rapid expansion in analysis of non-traditional stable isotopes, including the isotopes of several elements concentrated within ores and measured in ore minerals. These include iron, nickel, copper, zinc, germanium, selenium, and molybdenum. Several other isotopic systems with potential application to understanding the evolution of ore deposits are rapidly emerging. These include thallium, tellurium, and mercury. Table 1 briefly reviews current research on these isotopes in ores and the topics that are being addressed. Table 1. Applications of stable isotope geochemistry in ore mineralogy.
Tl [149] and references therein, [150][151][152][153] Significant isotopic variability in natural samples points to the potential of Tl-isotopes for understanding formation of hydrothermal deposits; more baseline data are needed.

New Approaches and Sub-Disciplines
The recent advances outlined above have, however, not only been brought about by rapid developments in analytical capability, but also by an evolution in conceptual approaches to ore mineralogy.

Trace Elements in Ore Minerals
Very few minerals are ever "pure"; many can incorporate two dozen or more different elements at measurable and quantifiable concentrations-and furthermore, many minerals can be compositionally zoned. Interest in the content and distribution of minor and trace elements in ore minerals has expanded rapidly for several reasons, driven by analytical capabilities that are highlighted above [154]. Firstly, the precise in-situ measurement of trace element concentrations has direct economic implications since it enables the mineralogical distribution of elements that represent either potential economic by-products [155], or harmful elements in an ore potentially decreasing its value, to be accurately pinpointed. This type of knowledge may directly or indirectly impact upon plant design for mineral processing (e.g., [156]). Secondly, accurate information on the mineral site(s) of potentially process deleterious elements provides data that can assist in management and long-term storage of mining or processing waste. Thirdly, trace element analysis can assist in understanding the genetic history of complex ore deposits. The recognition of distinct generations of a given mineral, or of grain-scale compositional zoning, can, for example, help to track the evolution of hydrothermal fluids, or expose evidence for an overprinting event that has implications for metal distributions within the deposit. Such an understanding can, in turn, underpin metallogenetic models, with direct application to model-based mineral exploration. Lastly, spatial patterns of trace element distribution within specific minerals can be potentially applied in exploration vectoring, offering greater sensitivity than whole-rock analysis (see [13] for a brief review).

The Significance of Gangue Minerals-A Holistic Approach to Ore Deposits
The gangue components of an ore mineral assemblage are increasingly recognised as rich sources of information to support models of ore formation and petrogenesis. Some particularly refractory minerals may, in fact, preserve textural and compositional features that reveal deposit evolution better than many ore minerals. In our laboratory, these have ranged from garnets and other calc-silicate minerals in skarns [157,158], through fluorapatite [35,159,160] and REE-fluorocarbonates [43,161], to alkali and plagioclase feldspars [36,46]. A detailed investigation of hematite, which contributes as much as 80 wt % of the ore at Olympic Dam, has shown how this mineral can encode a wealth of valuable trace element information [162].
This approach has led to the recognition that to adequately understand the distribution of those elements, which are particularly valuable for petrogenetic interpretation, notably REE, or the high field strength elements, a holistic approach is needed, in which all of the minerals, major and minor, in the system are studied as potential hosts. Partitioning among coexisting minerals can impact significantly on geochemical signatures, and changes in partitioning behaviour may itself provide insights into the evolution of mineral deposits. Holistic LA-ICP-MS strategies have also revealed how elements of interest (potential by-products and process deleterious impurities) may partition among coexisting hydrothermal base metal sulphides, for example, between galena, sphalerite, chalcopyrite, and tetrahedrite-tennantite [163][164][165][166]. This type of research has also shown how some gangue minerals can host unexpectedly high concentrations of elements of economic interest; concentration of U, W, Sn and Mo in skarn garnets [158] representing one good example. Several opportunities currently exist to confirm whether significant resources of so-called critical metals occur within new resource types, e.g., gallium in epidotites. These represent exciting research areas in which grain-scale analysis by LA-ICP-MS is expected to play a key role.
We have shown [167] how establishing models of trace element distribution in mineralogically complex ore systems need to acknowledge not only the major sulphide and oxide ore minerals and rock-forming silicates, but also relatively minor phases that can sequester some trace elements at significant concentrations. In the example from Fe-Ti-V ores from the Panzhihua layered intrusive complex, China [167], and, as shown as Figure 7, trace elements are distributed between rock-forming silicates (pyroxene, olivine, feldspar), titanomagnetite (the dominant oxide), and sulphides (pyrrhotite and minor chalcopyrite and cobaltpentlandite). Exsolution of ilmenite and spinel from titanomagnetite has, however, sequestered certain trace elements. Importantly, the high field strength elements (Sc, Zr, Hf, Ta, and Nb) are all markedly enriched in ilmenite relative to titanomagnetite, whereas Zn, Ga, and others, occur preferentially in spinel exsolutions within titanomagnetite.
Minerals 2017, 7, 233 12 of 24 significant concentrations. In the example from Fe-Ti-V ores from the Panzhihua layered intrusive complex, China [167], and, as shown as Figure 7, trace elements are distributed between rock-forming silicates (pyroxene, olivine, feldspar), titanomagnetite (the dominant oxide), and sulphides (pyrrhotite and minor chalcopyrite and cobaltpentlandite). Exsolution of ilmenite and spinel from titanomagnetite has, however, sequestered certain trace elements. Importantly, the high field strength elements (Sc, Zr, Hf, Ta, and Nb) are all markedly enriched in ilmenite relative to titanomagnetite, whereas Zn, Ga, and others, occur preferentially in spinel exsolutions within titanomagnetite. Note also relative enrichment of Zn and Ga in the fine-grained exsolved spinel within titanomagnetite. The reader is referred to [167] for additional background information on these samples.

Solid Solutions and Nanoparticles
The capability to explore smaller and smaller volumes of sample material has opened up exciting opportunities to understand the nature and behaviour of ore minerals at the nanoscale. Visualisation of the nanoscale features of minerals allows for the verification of crystal structures, polytypism and/or polysomatism in mixed-layer compounds, etc. Moreover, atomic-scale microscopy can help address fundamental questions on the nature of sold solution and how minerals adapt to incorporation of trace and minor elements. Nanoscale studies have also confirmed the role played by mineral nanoparticles, the stability and reaction behaviour of which are governed by very different sets of thermodynamic rules compared to their larger counterparts [168,169].
The example in Figure 8 shows hematite from the Olympic Dam orebody. This contains not only nanoparticles of uraninite (U-NPs), but also micron-scale oriented lamellae of REE-fluorocarbonates (bastnäsite-parisite group), which are interpreted to have coarsened up during recrystallization of granitophile (U-W-Sn-Mo-bearing)-rich hematite interacting with fluids via CDR reactions [162]. Ti-Mt: titanomagnetite, Ilm-ilmenite, Spn-spinel, Po-pyrrhotite, Pn-pentlandite. Note the relative enrichment of Ti, Sc, Mg, Mn, Zr, Hf, Ta, and Nb in ilmenite relative to titanomagnetite. Note also relative enrichment of Zn and Ga in the fine-grained exsolved spinel within titanomagnetite. The reader is referred to [167] for additional background information on these samples.

Solid Solutions and Nanoparticles
The capability to explore smaller and smaller volumes of sample material has opened up exciting opportunities to understand the nature and behaviour of ore minerals at the nanoscale. Visualisation of the nanoscale features of minerals allows for the verification of crystal structures, polytypism and/or polysomatism in mixed-layer compounds, etc. Moreover, atomic-scale microscopy can help address fundamental questions on the nature of sold solution and how minerals adapt to incorporation of trace and minor elements. Nanoscale studies have also confirmed the role played by mineral nanoparticles, the stability and reaction behaviour of which are governed by very different sets of thermodynamic rules compared to their larger counterparts [168,169].
The example in Figure 8 shows hematite from the Olympic Dam orebody. This contains not only nanoparticles of uraninite (U-NPs), but also micron-scale oriented lamellae of REE-fluorocarbonates (bastnäsite-parisite group), which are interpreted to have coarsened up during recrystallization of granitophile (U-W-Sn-Mo-bearing)-rich hematite interacting with fluids via CDR reactions [162]. Such reactions evolve via transient porosity providing sites for mineral deposition, as in this case U-NPs. The source of U is from hematite, which can incorporate up to hundreds of ppm of these "granitophile" elements [162].
Such reactions evolve via transient porosity providing sites for mineral deposition, as in this case U-NPs. The source of U is from hematite, which can incorporate up to hundreds of ppm of these "granitophile" elements [162].
Accessibility to the nano-to microscale domain has permitted a better understanding of the critical mechanisms involved in fluid-mineral reaction. Coupled dissolution replacement reaction (CDRR) [170][171][172] is recognised as a widespread phenomenon that is responsible for local-scale mineral transformation in hydrothermal ores. In our work on the Olympic Dam deposit, evidence for CDRR is recorded in feldspars [36,46], uraninite [76,77,173], hematite [162], and REEfluorocarbonates [43].

New Sub-Disciplines
In the past two decades, a new sub-discipline, geometallurgy, has emerged, in which geology and mineralogy are combined with extractive metallurgy and geostatistics. Geometallurgy enables the derivation of geologically-based predictive models for minerals processing that can assist plant design and promote efficient exploitation of mineral resources (e.g., [174][175][176][177]). Geometallurgy requires an interdisciplinary approach and widespread application of automated mineralogy, expanding traditional ore mineralogy into new spheres of economic relevance. Geometallurgical approaches are increasingly applied to a wide range of different deposit types. Key geometallurgy publications highlighting the importance of ore textures for mineral processing include [178][179][180].

Discovery of New Minerals
Despite only making up a tiny fraction of the Earth's crust, mineral deposits have proven to be rich sources for the discovery of new mineral species. In fact, if we include samples from abandoned mines and oxidised ores at the surface associated with historical mine workings, mineral deposits account for roughly half of the new minerals that were accepted between 2011 and 2016. This is perhaps unsurprising as mineral deposits are anomalous in elements that are prone to greater mineralogical diversity [181], and even more importantly, will contain various combinations of those elements. Sulphosalts represent a particularly fertile area of ore mineral research, with more than a dozen new sulphosalt species being approved since the publication of the report of the IMA Commission on Ore Mineralogy Sub-Committee on Sulphosalts [182]. It is also reasonable to expect that many of the predicted >1563 yet-to-be-discovered mineral species [183] will be identified in samples from existing or historical mine sites. Localities such as Tsumeb (Namibia), Långban (Sweden), Lengenbach (Switzerland), Shinkolobwe (Democratic Republic of Congo), and Franklin Accessibility to the nano-to microscale domain has permitted a better understanding of the critical mechanisms involved in fluid-mineral reaction. Coupled dissolution replacement reaction (CDRR) [170][171][172] is recognised as a widespread phenomenon that is responsible for local-scale mineral transformation in hydrothermal ores. In our work on the Olympic Dam deposit, evidence for CDRR is recorded in feldspars [36,46], uraninite [76,77,173], hematite [162], and REE-fluorocarbonates [43].

New Sub-Disciplines
In the past two decades, a new sub-discipline, geometallurgy, has emerged, in which geology and mineralogy are combined with extractive metallurgy and geostatistics. Geometallurgy enables the derivation of geologically-based predictive models for minerals processing that can assist plant design and promote efficient exploitation of mineral resources (e.g., [174][175][176][177]). Geometallurgy requires an interdisciplinary approach and widespread application of automated mineralogy, expanding traditional ore mineralogy into new spheres of economic relevance. Geometallurgical approaches are increasingly applied to a wide range of different deposit types. Key geometallurgy publications highlighting the importance of ore textures for mineral processing include [178][179][180].

Discovery of New Minerals
Despite only making up a tiny fraction of the Earth's crust, mineral deposits have proven to be rich sources for the discovery of new mineral species. In fact, if we include samples from abandoned mines and oxidised ores at the surface associated with historical mine workings, mineral deposits account for roughly half of the new minerals that were accepted between 2011 and 2016. This is perhaps unsurprising as mineral deposits are anomalous in elements that are prone to greater mineralogical diversity [181], and even more importantly, will contain various combinations of those elements.
Sulphosalts represent a particularly fertile area of ore mineral research, with more than a dozen new sulphosalt species being approved since the publication of the report of the IMA Commission on Ore Mineralogy Sub-Committee on Sulphosalts [182]. It is also reasonable to expect that many of the predicted >1563 yet-to-be-discovered mineral species [183] will be identified in samples from existing or historical mine sites. Localities such as Tsumeb (Namibia), Långban (Sweden), Lengenbach (Switzerland), Shinkolobwe (Democratic Republic of Congo), and Franklin (NJ, USA) all feature in a listing of the top 10 localities in terms of the number of minerals, for which that locality is registered as the type locality, with 72, 72, 43, 38, and 35 mineral species, respectively [184]. The technological revolution outlined above has accelerated the search for, and characterization of new ore minerals, particularly in cases where the phase in question is only a few microns in size and intergrown with other related phases. For example, FIB-SEM was successfully used to extract tiny samples of the new mineral graţianite for structural study [185]. Hazen et al. (2017) [186] have recently made a convincing case for the acknowledgement, as "anthropogenic minerals", of compounds that originate from human action. The majority of the 208 species put forward as anthropogenic mineral candidates can be attributed to mining-related activities, and are found within ore dumps, in slags and their weathering products, in mine shafts and adits, and on smelter walls.

Challenges-A Short Discussion
Ore mineralogy and ore microscopy remain highly relevant. Underteaching of traditional ore microscopy techniques in the World's higher education institutions is perhaps a reflection of the relative ease of access to advanced microanalytical instrumentation, which offer the advantage of compositional information. The need for innovative, interdisciplinary approaches in the search for mineral resources to guarantee future security of supply means that not only geologists, but also metallurgists and processing engineers, need to appreciate the pivotal role that is played by ore mineralogy in ensuring the sustainable exploitation of limited resources. Mineral characterisation remains a critical part of any study, so that reliable interpretations can be made based on knowledge of exactly what has been analysed.
Descriptions of ore minerals, new occurrences, and information on mineral textures or paragenesis, remain important, as do compilations of compositional or other types of quantitative data. Ores are by essence ephemeral in character, and once mined, are gone forever. Museums and other national, state, or municipal institutions have a vital role to play in archiving particularly interesting or instructive sample material. Likewise, universities need to ensure that teaching and research collections are maintained to provide students with representative material from typical "run of mine" ore. Reference collections should be assembled in collaboration with the mining companies at the time of mining, and maintained as a resource for future research.
Identifying current frontiers in ore mineralogy is a subjective exercise. Obtaining quantitative compositional data on the smallest particles is one obvious frontier that can be crossed with contemporary instrumentation. The increased focus on smaller and smaller scales of observation, and not infrequently on smaller sample suites due to restricted access, or the high cost and time-consuming nature of the more advanced microanalytical techniques, does, however, generate unease about just how representative the investigated samples are. This may, in turn, impact on the reliability of broad interpretations that are based on empirical observation on only the tiniest fractions of an orebody. Further in the future, the widespread use of automated mineralogical techniques and large microanalytical datasets will likely trigger a new revolution in which machine learning can be expected to play an increasing role.
Many of the techniques and approaches discussed here complement one another and can often be used in combination to achieve in-situ characterisation of the same sample at different observational scales ( [23] and references therein). Most methods do, however, have their limitations, in terms of sensitivity, resolution, or less quantifiable factors, such as ill-constrained sources of error, a lack of suitable standard materials, and so on. A comprehensive understanding of mineralogy and geochemistry-and increasingly, a working knowledge of mineral physics and crystallography, is critical.
Re-reading the author's preface to "The Ore Minerals and their Intergrowths" [1], published nearly 50 years ago at the start of the "microprobe era", we wonder what Paul Ramdohr (1890Ramdohr ( -1985 [187] would make of the advances that are summarised here. Some of the points raised, such as the impact of solid solutions and lattice defects on optical properties, are now far better understood and quantified. Ramdohr notes "The variation which a mineral as 'trivial' as pyrite can exhibit is simply baffling", a point that can be readily responded to in the light of a half-century of studies of trace element distributions, deformation mechanisms, and lattice-scale phenomena that have shown pyrite to be among the most complex and intriguing among ore minerals-a host for "invisible gold", stable in an extraordinary range of environments and increasingly viewed as a valuable indicator of ore formation processes. Several hundred new ore minerals have been discovered since the publication of "The Ore Minerals and their Intergrowths", yet the basic principle that ore minerals are best studied in the context of their associations is as true today as it was then. Likewise, only high-quality sample preparation can reveal the intricacies and beauty of ore minerals. The new data that is presented here represent a glimpse into the characterisation of ore minerals that can shed new light on processes of ore formation. Recognition and measurement of minerals, and their characteristics at the nanoscale can explain features seen at the microscale (e.g., petrographic microscope), macroscale (in drillcore or underground), or are mappable at the deposit-scale. Moreover, mechanisms of alteration and mineralization can often only be correctly understood if evidence at the nanoscale is considered. In the case of complex, enigmatic IOCG systems, such as the 10,000 Mt Olympic Dam deposit, observations that can bridge scales of observation from nanometres to hundreds of metres are contributing to a working understanding of how giant ore systems formed and developed.

Conclusions
The diversity of current research on ore minerals is remarkable and only a few topics could be mentioned here without making this text excessive. The breadth of advances made possible by the availability of microanalytical techniques with ever-improving sensitivity and spatial resolution, opens many new approaches to resolve long-standing gaps in knowledge. Changing demand patterns that are driving the very nature of which phases can be defined as ore minerals, a blurring of traditional research domains of mineralogy, geochemistry and even materials science, and the ever-increasing complexity of all minerals as they are studied at smaller and smaller scales, has dramatically changed ore mineralogy during the last 20 years.  (Figures 6 and 7). See text and references therein for additional information on analytical set-up and operating conditions.