Martin Wells

Mineralogy and crystal chemistry of “garnierites” in the Goro lateritic nickel deposit, New Caledonia

The mineralogy and crystal chemistry of “garnierites” in saprolitic ore from the Goro lateritic nickel deposit, New Caledonia, was investigated using optical and scanning electron microscopy, X-ray diffraction (XRD) and electron-microprobe analyses. These conspicuously, green-coloured phases occur either as sub-mm to cm-sized veins or as macroscopic (sub-cm sized) “booklets”. Veins comprised ~ 10 A (2:1) talc-like minerals identified as species of the Ni-kerolite/pimelite Ni for Mg solid solution series, with Ni contents ranging from 10 to 24 % NiO. Mineral nomenclature, defined by the inverse relationship between Ni and Mg content, varied continuously over the scale of several hundred micrometres. Pimelite, defined as containing > 1.5 Ni per formula unit (p.f.u.), was the main 10 A phase for one vein core whereas at the vein edges where the Ni content decreased to < 1.5 Ni p.f.u. (and Mg increased) kerolite was identified. “Booklets” comprised the ~ 7 A (1:1) serpentine-like phase, nepouite (Mg0.67Ni1.59Fe0.053+) (Si2.17Al0.22)O5(OH)4 with Ni contents averaging 30 % NiO and occurred as accordion-like structures supported in an undifferentiated matrix of mixed ~ 7 A and ~ 10 A phases.

THERMAL AND MINERAL PROPERTIES OF Al-, Cr-, Mn-, Ni- AND Ti-SUBSTITUTED GOETHITE

Mineralogical and thermal characteristics of synthetic Al-, Cr-, Mn-, Ni- and Ti-bearing goethites, synthesized via alkaline hydrolysis of metal-ferrihydrite gels, were investigated by powder X-ray diffraction and differential thermal analysis. Shifts in unit-cell dimensions were consistent with size of substituent metal ions and confirmed the incorporation of Al3+, Cr3+, Mn3+, Ni2+ and Ti4+ in the goethite structure. A weight loss of 6.2 wt.% for goethite containing 12.2 mol.% Ti, being significantly less than for stoichiometric goethite, is consistent with the replacement of Fe by Ti in the goethite structure coupled with the substitution of O2− ions for OH− (i.e. proton loss). These data provide the first confirmation of the direct replacement of Fe by Ti within goethite. Formation of multiple dehydroxylation endotherms for goethite containing 4.5 mol.% Al, 15.3 mol.% Mn and 12.2 mol.% Ti was not attributed to the decomposition of surface OH groups or related simply to the crystallinity of precursor goethite (‘high-a’ vs. ‘low-a’) as defined by the magnitude of a. Instead, endotherm doublet formation was associated with weight loss due to the dehydroxylation of goethite remaining after initial phase transformation to protohematite and to the evolution of OH− associated with the rapid increase in crystallite size of protohematite directed primarily along the a direction. Development of the first endotherm is due to initial dehydroxylation and transformation to protohematite. With continued heating of well ordered goethite or goethite containing moderate to high levels of substituent cations, domain growth along the a direction is delayed or inhibited to a critical point that provides enough thermal energy to enable goethite transformation to proceed to completion and for proto-hematite domain growth to occur. This results in the formation of a second endotherm. For less well ordered goethite and/or goethite containing only low levels of foreign metal cations, protohematite domain growth is not inhibited and proceeds continuously with heating to give only a single endotherm.

PROPERTIES AND ACID DISSOLUTION OF METAL-SUBSTITUTED HEMATITES

The dissolution in 1 M HC1 of Al-, Mn-, and Ni-substituted hematites and the influence of metal substitution on dissolution rate and kinetics of dissolution were investigated. The inhomogeneous dissolution of most of the hematites investigated was well described by the Avrami-Erofe’ev rate equation, kt = √[-ln(l − α)], where k is the dissolution rate in time, t, and α is the Fe dissolved. Dissolution of Al-substituted hematite occurred mostly by edge attack and hole formation normal to (001), with the rate of dissolution, k, directly related to surface area (SA). Dissolution of rhombohedral Mn- and Ni-bearing hematites occurred at domain boundaries, crystal edges, and corners with k unrelated to SA. The morphology of Mn- and Ni-substituted hematites changed during dissolution with clover-leaf-like forms developing as dissolution proceeded, whereas the original plate-like morphology of Al-bearing hematite was generally retained. Acid attack of platy and rhomboidal hematite is influenced by the direct (e.g., metaloxygen bond energy, hematite crystallinity) and indirect (e.g., crystal size and shape) affects associated with incorporation of foreign ions within hematite.

Rapid dehydroxylation of nickeliferous goethite in lateritic nickel ore: X-ray diffraction and TEM investigation

A method for extracting Ni and other metals from lateritic ores by means of shock heating has been investigated. Shock heating releases some of the metal from its goethitic host. Even though the transformation of pure goethite to hematite is known to occur via intermediate hydroxylated phases, the effect of other metals such as Ni substituting for Fe in goethites on this thermal transformation to hematite is unknown. The purpose of this study was to fill this gap, with the hope that the results will lead to more energy-efficient extraction methods and/or a better understanding of Fe geochemistry in thermally activated soils. X-ray diffraction, transmission electron microscopy with EDS, and thermal analysis were used to investigate mineralogical changes in nickeliferous goethites from five oxide-type lateritic nickel ore deposits that had been subjected to shock heating at temperatures in the range 220–800°C. Acicular, nano-sized goethite was the main constituent of the samples with minor to trace amounts of quartz, talc, kaolinite, chromite, maghemite, and Mn oxides. Goethite was partially dehydroxylated to OH-hematite at 340–400°C and had completely altered to well ordered hematite at 800°C. The OH-hematite was characterized by broad XRD peaks for reflections associated with the Fe sublattice. The goethite unit-cell a and b lengths remained almost constant with increasing preheating temperature up to 300°C, while the size of the c axis dimension contracted. The neoformed hematite crystals were larger than the precursor goethite crystals due to development, by sintering and surface diffusion, of regularly ordered hematite domains. The increase (1.5–2.6 fold) in surface area with increasing heating temperature (up to 340–400°C) reflected the development of slit-shaped micropores (∼300°C), which further developed into elliptically shaped micropores (∼400°C) in OH-hematite. With increased heating temperature, well ordered hematite formed with only a few micropores remaining. Such results may contribute to the development of more efficient procedures for extracting Ni from lateritic nickel ores, as the rate of dissolution of goethite in acid in ‘heap and pressure’ leach facilities will be enhanced by the increases in surface area and microporosity. The results may also provide valuable information on the probable effects of natural heating on pedogenic Fe oxides.

Nickel distribution and speciation in rapidly dehydroxylated goethite in oxide-type lateritic nickel ores: XAS and TEM spectroscopic (EELS and EFTEM) investigation

The distribution of Ni in four lateritic Ni-goethites that were rapidly dehydroxylated to form hematite by shock heating at 340/400°C and 800°C for 30 min was investigated using synchrotron X-ray diffraction (SXRD), TEM spectroscopy (EELS and EFTEM) and synchrotron X-ray absorption spectroscopy (XAS). The Ni K-edge EXAFS results for non-heated samples showed three distinct Ni–Fe shells, including two edge-sharing (R Ni–Fe ∼ 3.01 and R Ni–Fe ∼ 3.22 Å) and a double corner-sharing (R Ni–Fe ∼ 3.52 Å) complex for most of the samples. These interatomic distances are indicative of Ni substituting for Fe in goethite, which has resulted in an expansion in the goethite structure along the a-axis direction and a contraction along the b-axis direction. Ravensthorpe Ni goethite was considerably different from the other goethites, with two Ni–O interatomic lengths (R Ni–O ∼ 2.04 and 2.46 Å), an edge-sharing (R Ni–Fe ∼3.04 Å) and a corner-sharing (R Ni–Fe ∼ 3.56 Å) complex. The R Ni–O ∼ 2.46 Å bond length is not indicative of Ni substituting for Fe in goethite, nor is it associated with single or multiple scattering events in NiO, Ni(OH)2 or Ni substituting for Fe in Fe oxides. The corresponding Ni K-edge EXAFS results for 340/400°C and 800°C heated samples (i.e. hematite) were very similar. Four distinct metal neighbours correspond to Fe/Ni in face-sharing (R Ni–Fe ∼ 2.87–2.91 Å) and three different corner-sharing complexes (R Ni–Fe ∼ 3.37– 3.41 Å, R Ni–Fe ∼ 3.62– 3.64 Å and R Ni–Fe ∼ 3.92– 4.09 Å) represent Ni substituting for Fe in hematite. The fourth shell is indicative of an inner sphere surface complex. EFTEM maps for Ni in goethite are consistent with the formation of a surface complex as they provide evidence for clustering of Ni on the surface of neoformed hematite crystals. There was no evidence from EXAFS or SXRD supporting the formation of discrete Ni phases (e.g. NiO) as a result of shock heating. Therefore, for Ni-goethites subjected to shock heating at 800°C (i.e. high-temperature dehydroxylation to hematite), most of the Ni is retained in the structures of the neoformed hematites, whereas some of the Ni migrates to the surface of the neoformed hematite where it forms a surface complex. During acid dissolution (e.g. heap leaching) of oxide-type lateritic Ni ores, Ni on the hematite surface is more accessible to acid solutions; therefore, these results may provide a basis for more efficient extraction methods for Ni in oxide-type lateritic Ni ores, as well as providing information on the possible redistribution of Ni in heated goethite-rich soils.

Quantification of Ni laterite mineralogy and composition: a new approach

Quantification of the chemical and mineralogical composition of Ni-bearing laterites is challenged by their complex nature, with the distribution of Ni and other metals controlled by a number of phases. Anew approach to predict the Ni (and Fe, Mg) composition of Ni laterites more reliably was tested using samples from drill holes from the Siberia North and Highway deposits. These deposits, selected as examples of (siliceous) oxide-style lateritic deposits, form part of the Kalgoorlie Nickel Project (KNP) in the north-eastern Yilgarn, Western Australia. Diffuse reflectance spectra over the 380–2500 nm wavelength range, measured using the CSIRO HyChips™ automated scanning system, were integrated with a multi-element dataset for 1 m-composite pulp samples using partial least-squares regression (PLSR) analysis. PLSR cross-validation models were developed to predict Ni, MgO and FeO, contents in samples with bulk composition limits of >0.08 wt% Ni and <6 wt% MgO. Calibration and validation models showed a strong, linear trend for characterising FeO and, to a lesser extent, Ni content, whereas reflectance spectroscopic-chemical models for predicting MgO content were less reliable. This reflected the prevailing mineralogy and associated element distribution within the oxide laterite profiles. For example, Fe and Ni distribution is controlled by relatively few minerals—predominantly goethite and, to a lesser extent, the Fe3+-bearing smectite, nontronite. Magnesium is associated with a range of Mg-bearing minerals, some of which are compositionally complex and include serpentine, talc, chlorite, smectite and carbonates (e.g. dolomite and magnesite). The MgO abundance alone does not reflect the varied mineralogy of samples with relatively high MgO contents, which can be spectrally distinct. Hence, samples with similar Mg contents, although comprising Mg-bearing minerals with distinct reflectance spectra, were difficult to model (e.g. smectite vs Mg carbonates). The findings of this study demonstrate that characterisation of Ni laterite chemistry may be achieved as part of routine logging of either drill cuttings or diamond drill core using PLSR analysis, but a detailed understanding of the mineral–element association is essential for reliable predictive analysis.

Reflectance spectroscopic characterisation of mineral alteration footprints associated with sediment-hosted gold mineralisation at Mt Olympus (Ashburton Basin, Western Australia)

ABSTRACT Hydrothermal ore deposits are typically characterised by footprints of zoned mineral assemblages that extend far beyond the size of the orebody. Understanding the mineral assemblages and spatial extent of these hydrothermal footprints is crucial for successful exploration, but is commonly hindered by the impact of regolith processes on the Earths surface. Hyperspectral drill core (HyLogger™-3) data were used to characterise alteration mineralogy at the Mt Olympus gold deposit located 35 km southeast of Paraburdoo along the Nanjilgardy Fault within the northern margin of the Ashburton Basin in Western Australia. Mineralogy interpreted from hyperspectral data over the visible to shortwave (400–2500 nm) and thermal (6000–14500 nm) infrared wavelength ranges was validated with X-ray diffraction and geochemical analyses. Spaceborne multispectral (ASTER) and airborne geophysical (airborne electromagnetic, AEM) data were evaluated for mapping mineral footprints at the surface and sub-surface. At the deposit scale, mineral alteration patterns were identified by comparing the most abundant mineral groups detected in the HyLogger data against lithology logging and gold assays. Potential hydrothermal alteration phases included Na/K-alunite, kaolin phases (kaolinite, dickite), pyrophyllite, white mica, chlorite and quartz, representing low-T alteration of earlier greenschist metamorphosed sediments. The respective zoned mineral footprints varied depending on the type of sedimentary host rock. Siltstones were mainly characterised by widespread white-mica alteration with proximal kaolinite alteration or quartz veining. Sandstones showed (1) distal white mica, intermediate dickite, and proximal alunite + kaolinite or (2) widespread white-mica alteration with associated intervals of kaolinite. In both, sandstones and siltstones, chlorite was distal to gold mineralisation. Conglomerates showed distal kaolinite/dickite and proximal white-mica/dickite alteration. Three-dimensional visualisation of the gold distribution and spatially associated alteration patterns around Mt Olympus revealed three distinct categories: (1) several irregular, poddy, SE-plunging zones of >0.5 ppm gold intersected by the Zoe Fault; (2) sulfate alteration proximal to mineralisation, particularly on the northern side of the Mt Olympus open pit; and (3) varying AlIVAlVISiIV–1(Mg,Fe)VI–1 composition of white micas with proximity to gold mineralisation. Chlorite that developed during regional metamorphic or later hydrothermal alteration occurs distal to gold mineralisation. ASTER mineral mapping products, such as the MgOH Group Content used to map chlorite (±white mica) assemblages, showed evidence of correlation to mapped, local structural features and unknown structural or lithological contacts as indicated by inversion modelling of AEM data.

Hyperspectral Imaging of Iron Ores

Hyperspectral sensing is based on reflectance spectroscopy that studies light as a function of wavelength reflected or scattered from a solid surface. In reflectance spectroscopy, the most common iron oxides minerals in iron ores (hematite and goethite) have broad absorptions between 380 and 1000 nm whereas waste materials such as kaolinite, gibbsite, montmorillonite and riebeckite show narrow absorption features between 1000 and 2500 nm.