The main thrust of our research is to develop the genetic models that are needed to ensure the future successful exploration for economic deposits of metallic minerals.  Many of these deposits form as a result of metal transport by fluids, mainly hydrothermal, and the subsequent precipitation of these metals in response to changes in physico-chemical conditions.  In some cases, the fluids are of magmatic origin and there is often an intimate relationship between fluids and magmatic processes of metal concentration.  The primary objectives of our group are thus to determine the nature of such fluids (and melts), identify the factors which control their ability to dissolve high concentrations of metals, and establish the mechanisms which form ore deposits.  In order to meet these objectives, we combine studies of natural systems, which can be used to reconstruct the environments of metal transport and deposition, with analytical, experimental and theoretical studies designed to evaluate the behaviour of the metals in fluids and magmas, and model the processes of ore formation.

Some recent contributions of our research have demonstrated that:  1) aqueous vapour can play a major role in metal transport in porphyry/epithermal ore-forming systems; 2) hydrocarbons can transport metals in concentrations sufficient to form ore deposits, and may have been contributed to the formation of MVT deposits and metal-rich black shales; 3) colloidal processes may be essential for the formation of ultra-high grade gold deposits, and not all Archean gold deposits are products of the widely accepted orogenic model; 4) the REE and strategic metals are concentrated to economic levels by a combination of magmatic (liquid immiscibility) and hydrothermal processes, and that the highly sought after heavy REE are effectively fractionated from the light REE by their lower solubility in hydrothermal fluids; 5) battery metals such as Co reach their highest concentrations as a result of redox dependent hydrothermal processes, and that aqueous fluids are essential to the formation of Li-rich pegmatites; 6) contrary to conventional wisdom, U can be transported in concentrations sufficient to form ore deposits by reduced hydrothermal fluids; and 7) the hydrothermal transport of W and Mo in granitic systems is dominated by fluorine- and sodium-bearing species, respectively, instead of simple tungstate and molybdate species.


Vapour Transport of Metals - Experiments, Volcanoes and Epithermal/Porphyry Deposits

For more than 100 years, aqueous liquids have been thought to be the main agents of metal transport for magmatic hydrothermal ore deposits.  Despite compelling evidence from analyses of fluid inclusions that aqueous vapours can transport metals in concentrations sufficient to form porphyry and epithermal ore deposits, until our studies, experimentalists had been unable to duplicate these concentrations in the laboratory.  We have succeeded in experimentally reproducing these concentrations for Au, Ag, Cu and Mo, and developed new thermodynamic models of metal transport in the vapour phase that satisfactorily explain the enrichment and distribution of these metals in natural systems.  Specifically, we have shown that these and other metals are concentrated in the vapour through the formation of hydrated species and that their solubility increases exponentially with hydration number and in turn with H2O fugacity.  The resulting model envisages the exsolution of supercritical vapour-like fluids from magma that transport metals, and evolve through condensation of a limited amount of brine to become a true vapour, which dominates by mass, and becomes the main ore fluid for the subsequent formation of porphyry and epithermal ore deposits.  This model represents a radical departure from previous genetic models in which the vapour is considered to be a product of boiling and participates in ore formation only to the extent that it carries heat and acidic components from the system.

Our experimental studies have been complemented by work on active volcanoes, notably Merapi and Kawah Ijen in Indonesia, where we have measured metal concentrations in volcanic gases comparable to those determined experimentally.  In the case of gold and copper, these studies also help explain the evidence in some high sulfidation epithermal systems that alteration and mineralization were both products of condensed acidic vapours.


Metals, Hydrocarbons and Black Shales

A fluid which has not received serious attention as a possible ore transport medium is petroleum.  Traditionally, the role of liquid hydrocarbons in hydrothermal systems has been restricted to that of a reductant.  Recent field-based studies, including our own, have documented a close connection between hydrocarbons and the concentrations of metals, e.g., Zn, Pb, Hg, U, Ni, V and the PGE.  Experiments in our laboratory have provided the first direct evidence that petroleum and its analogues can dissolve metals in the high concentrations required of an ore fluid.  Models developed using these data have shown that the genesis of MVT Zn-Pb and PGE-rich black shale deposits can be satisfactorily explained by petroleum transport of the corresponding metals.  Demonstration that, in some cases, petroleum may be a viable alternative to liquid water as an ore fluid should encourage researchers to re-evaluate accepted ore deposit models for environments where liquid hydrocarbons are known to be important.


Gold Transport and Deposition

Although it has been long assumed that all hydrothermal gold deposits form as a result of the transportation of the metal to the site of deposition as aqueous species, this is problematic for bonanza-type vein deposits for which the dissolved concentrations of gold are likely to be insufficient (or the required volumes of fluid unreasonably high) to form such deposits.  Recently, however, we potentially resolved this problem by providing compelling evidence that the ultra-high grade mineralization at the Brucejack gold mine (BC) was the result of colloidal processes.  Specifically, we showed that the gold occurs as discrete 1 to 5 nm particles and as much larger (micron-scale) flocculated aggregates, and that these colloids formed as a result of supersaturation of the fluid in gold in response to vigourous boiling; flocculation may have occurred due to mixing of the boiled fluid with seawater.  By enabling the mechanical concentration of the nano-particulate and flocculated gold colloids through processes such as seismic pumping, colloidal transport and deposition may explain the formation of very high grade gold mineralization in a variety of geological settings.  In principle, colloids could also play an important role in the genesis of deposits involving other metals (e.g., Ag), but further field and experimental studies (in progress) will be necessary to test this idea.

Although most Archean mesothermal/orogenic gold deposits are thought to form from fluids derived largely from metamorphic dewatering reactions, some large and important deposits are a product of fluids with a significant magmatic component.  Building on our previous studies of the world-class Hemlo (Ontario) deposit, we recently investigated disseminated gold mineralization at Canadian Malartic (Quebec), a very large (>20 Moz), low grade gold deposit.  As this deposit is partly hosted by porphyritic intrusions and is associated with well-developed potassic alteration, it may represent an example of the still poorly understood intrusion-related class of deposits, or be an unusual variant of the more typical mesothermal/orogenic class of deposits, in which a significant proportion of the mineralizing fluid was orthomagmatic in origin.  We have reconstructed the hydrothermal and metamorphic history of the Malartic area using textural, modal and mineral-chemical features of pyrite and pyrrhotite, and in the process have provided a method for vectoring toward gold mineralization in the district.  Similarly, we have developed methods using mica compositions and lithogeochemistry to evaluate the hydrothermal footprint of alteration at Canadian Malartic, and to decipher the relative effects of subsequent metamorphism.

On the other side of the world, the Witwatersrand constitutes the greatest, and most controversial, Archean gold-(uranium) deposit.  There has been a broad consensus that the Witwatersrand is a product of hydrothermally modified paleoplacer deposition, although there have been recent suggestions that these deposits are the result of direct chemical precipitation from relatively reduced ocean or fluvial waters.  Our studies of the Carbon Leader Reef suggest that an additional process, namely mobilisation by hydrocarbon liquids, may play an important role in the concentration of U.  Over 95% of the U and ~40% of the Au in the Reef is intimately associated with pyrobitumen, and we have demonstrated that the U occurs in the form of nanocrystals that flocculated to form masses up to 50 μm in diameter.  We have developed a model in which detrital uraninite was dissolved by liquid hydrocarbons and transported to the Reef, where it re-precipitated due to interaction of hydrocarbons with hydrothermal fluids.  Hydrothermally remobilized placer gold is interpreted to have precipitated by reduction around oil droplets, evidence of which is provided by spherical cavities observed in secondary gold grains.


Critical Metal Deposits – Magmatic and Hydrothermal Processes

The need to address the problem of global warming and to find green energy solutions have created an impetus for exploring for strategic metals, including the REE.  As modern exploration depends on robust genetic models for ore deposits, we have conducted extensive field, laboratory and theoretical studies designed to improve our understanding of the processes that concentrate these metals to economic levels.  Many of these strategic metals (e.g., the REE including Sc, and Nb, but not the battery metals Co-Ni and Li which are discussed below) reach their highest concentrations in alkaline igneous rocks and carbonatites.  We have therefore focused much of our recent research on deposits hosted by these rocks (e.g., peralkaline granites, layered alkaline mafic complexes, carbonatites).  Although magmatic processes are important in enriching strategic metals, formation of their ores is, in most cases, highly dependent on hydrothermal processes. 

In the absence of information on the speciation of most of the critical metals in hydrothermal fluids, our laboratory has pioneered experimental solubility studies designed to identify the dominant species of these metals in high temperature aqueous fluids, determine their stability, and generate thermodynamic data that can be used to model their transport and deposition.  These studies have shown in the case of the REE, that main ligand for transport is the chloride ion; fluoride forms stronger complexes but does not play a role in transport because of the associated nature of HF, and the limited availability of the fluoride ion.  They have also shown that LREE complexes are considerably more stable than those of the HREE.  An important consequence of this discovery is that the LREE are more mobile than the HREE in hydrothermal fluids, thereby explaining the commonly observed zonation of the REE many deposits.  This also has exploration implications because the La:Lu ratio can be used as a vector to REE (and U) mineralization.  In contrast to the behaviour of the REE, our experimental studies of Zr, Nb and Ta solubility have emphasized the importance of mixed hydroxy-fluoride complexes in their transport.  They have also shown that zircon has retrograde solubility and that Nb is significantly more mobile than Ta, which explains the preferential mobilization of Nb that is observed in hydrothermally altered Nb-Ta pegmatite deposits.  Until recently, it had been assumed that W and Mo are transported exclusively as tungstic and molybdic acid, respectively, and their dissociated products.  Our work on W has shown that this metal forms very strong complexes with fluoride ions, and for most deposits, these complexes dominated W transport.  This finding indicates that the common association of W mineralization with F-bearing minerals (such as fluorite and topaz) is not a coincidence but reflects the need for an F-rich fluid to ensure the formation of W deposits.  In the case of Mo, our work shows that the dominant complexes in most mineralizing fluids are Na-bearing molybdate species; this explains the observation that the ore fluid in some, deeper porphyry Mo deposits is typically of very high salinity.

Most critical metal deposits involve the interplay of magmatic and hydrothermal processes.  An outstanding example of this is the peralkaline granite at Strange Lake, QC, which hosts a world-class REE-Nb resource, and is the best-studied deposit of this type.  The REE and Nb are concentrated in pegmatites, and in part this concentration is due to a combination of low degrees of partial melting and extreme fractional crystallization.  An important contribution of our research has been the discovery of a new process for REE enrichment, namely fluoride-silicate liquid immiscibility.  In a manner analogous to Ni enrichment by sulfide liquids, the REE are partitioned preferentially into the fluoride liquid by a factor of approximately 100, and the resulting magma is focused into the core of pegmatite bodies.  A feature of REE-Nb pegmatites is that they saturate relatively early with a hydrothermal fluid that is in principle capable of scavenging the REE (e.g., from the crystallized fluoride melt).  Few previous studies, however, have analysed the compositions of REE in fluid inclusions and, until recently, none have attempted to trace their path during the subsequent evolution of the system.  In our work on Strange Lake, we have reconstructed this path and shown that the REE are transported from the pegmatites into the adjacent granites, and moreover, that they are fractionated in the process leading to a deposit in which the HREE are concentrated proximal to the pegmatites and the LREE distal to them.  This distribution is predicted by our experimental solubility and speciation studies discussed above.

Our studies of the Nechalacho layered igneous complex (NWT), Canada’s only REE producer, has also led to new understanding of REE-Nb behaviour in peralkaline rocks.  There we have shown that end-stage, top down, bottom up fractional crystallization of an alkaline basic magma can concentrate the REE and Nb to economic levels.  Significantly, and in contrast to other alkaline layered complexes, the magma saturated with a hydrothermal fluid that obliterated most of the primary mineralogy, including eudialyte and zircon, which are the main magmatic REE-bearing phases.  This hydrothermal alteration led to the concentration of the REE in fergusonite (HREE.NbO4), which is now the main HREE ore mineral in the deposit, along with the main LREE mineral, monazite.  The alteration was particularly important for the liberation of the HREE from zircon, which is a highly refractory phase.

One of the outstanding questions pertaining to the genesis of carbonatite-hosted deposits concerns the relative roles of magmatic versus hydrothermal processes in concentrating the REE.  Many previous studies have concluded that the REE mineralization is dominantly magmatic in origin.  By contrast, our work on the Wicheeda (BC) and Eldor (QC) carbonatites have demonstrated clearly that the mineralization is almost exclusively hydrothermal, consistent with the observation that carbonatitic magmas can dissolve high proportions of aqueous fluids and will exsolve them in shallow crustal settings.  These fluids ultimately concentrate the REE as bastnaesite (LREE.FCO3) and monazite.  Significantly, the LREE:HREE ratio within these minerals varies within a deposit, and in the case of the Eldor carbonatite (Ashram deposit), this ratio is greatest proximal to the inferred site of fluid exsolution (a breccia), and decreases away from it.  This distribution of the REE is explained by both the greater stability of LREE chloride complexes relative to those of the HREE, and the lower solubility of HREE end members of bastnaesite and monazite.  The Lofdal HREE deposit (Namibia) provides an extreme example of carbonatite-associated mineralization in which the intrusion is not the host but simply the source of the fluids that deposited the REE.  It is also an example of the extreme fractionation of the REE in which the low pH buffering capacity of the host quartzofeldspathic gneisses ensured that only the relatively insoluble HREE precipitated (as xenotime).   Other carbonatites that we are working on include the St. Honoré and Oka intrusions which are currently and/or have been exploited, respectively, for Nb.

Scandium is a critical metal that is expected to be of increasing importance because of its use in Al-Sc alloys that are finding application in the aerospace and automotive industries to improve energy efficiencies and address the problem of global warming.  Our recent work on Sc includes the first comprehensive study of the behaviour of Sc in nature and the processes that concentrate it to potentially economic levels.  Such concentration occurs by the introduction of Sc into Fe-enriched clinopyroxene, either during fractional crystallization of alkaline magmas of intermediate to ultramafic composition or metasomatically, and that the weathering of such rocks can lead to the formation of Sc-rich laterites that are also exploitable.  Building on this work, we are also currently investigating the controls and distribution of Sc at Crater Lake, Quebec; this research shows that Sc is concentrated preferentially in a pyroxene ferrosyenite, and that as predicted, clinopyroxene is the dominant host for this metal.  The field-based studies are being complemented by experimental research of the solubility and speciation of Sc in hydrothermal fluids.  This work has shown that fluoride complexes are potentially important in the hydrothermal transport of Sc, which imply that this mechanism may have been responsible for the fluorite-associated, clinopyroxene-hosted Sc mineralization at Bayan Obo, China.

Iron oxide-apatite deposits represent a potentially significant source of REE mineralization that is best exemplified by the Kwyjibo deposit, Québec.  Although iron oxide-apatite deposits have been the subject of many previous studies, their genesis is hotly debated, with opinion divided over whether they are entirely magmatic or hydrothermal, and in the case of the former, whether they are the products of silicate-oxide liquid immiscibility or some type of fractionation.  Evidence from the Kwyjibo magnetite-apatite-REE deposit indicates that the REE were initially concentrated in apatite of magmatic origin, and then subsequently remobilized and enriched to higher grade by hydrothermal fluids.  The evidence also suggest that the magnetite-apatite rocks were the product of fractional crystallization of early magnetite and apatite from an A-type granitic magma. 

Energy Metals for a Greener Future

Concerns about global warming have placed sharp focus on the need to find new resources for the electrification of the world economy.  These include both battery metals for energy storage and transportation, and metals for the direct production of electricity.  Modern battery technology requires metals such as Li, Co and Ni, which are currently in short supply.  We are addressing this need with research projects on pegmatite-hosted Li and Li-enriched oilfield brines.  Pegmatites are the main source of hard-rock lithium.  Our studies have identified the key role that aqueous fluid saturation plays in their genesis, and supports the Jahns-Burnham model.  The other major sources of Li are brine of meteoric/evaporitic origin, and Li-bearing clay deposits.  An interesting and poorly understood source of brine is that associated with oilfields.  This source is important because it represents a potential value-added byproduct of hydrocarbon extraction.  Our research has identified element correlations with Li which strongly suggest that the detrital component of black shales (the source rocks for hydrocarbons) is the source of the Li, and that the exploration for brines rich in Li should be guided by an evaluation of the major element chemistry of the black shales.

We have launched a comprehensive, multi-pronged approach to understand the processes that concentrate cobalt to economic levels.  This research was initiated with detailed review and analysis of metallogeny of Co, and the factors that control its enrichment in magmatic, hydrothermal and lateritic environments.  A major contribution of this research was the demonstration that Co is optimally transported in highly oxidized hydrothermal fluids as chloride complexes, and deposited either in response to reduction or pH neutralization depending on temperature.  One of the impediments to modelling hydrothermal Co ore formation has been the lack of thermodynamic data for the key ore minerals.  We have recently conducted a series of calorimetric and solubility experiments that have provided these data for carrolite (CuCo2S4), and have initiated a study of sediment-hosted Co deposits in the African Copperbelt where carrolite is the main ore mineral.  We are also investigating the genesis of Alaskan-type ultramafic-hosted, magmatic Ni-Co mineralization, and its potential remobilization by hydrothermal activity during serpentinization.  An important conclusion of our research is that Co is much more mobile in aqueous fluids than Ni, thereby explaining the common occurrence of major sediment-hosted hydrothermal Co deposits and the paucity of hydrothermal Ni deposits.  The other important conclusion is that ultramafic rocks are the source of the Co that forms both sediment-hosted and vein-type hydrothermal cobalt deposits.

The two metals that are important or potentially important for energy production are U and Th, respectively.  Our research has focused on the behaviour of these metals in hydrothermal fluids.  Until recently, it was assumed that U can only be transported in the 6+ state, and consequently models for the genesis of various types of U deposits were based on this assumption.  We have shown experimentally however, that U can also be transported in ore-forming concentrations in the 4+ state.  Indeed, at temperatures above 350°C, reduced U4+ species dominate uranium transport in the saline fluids thought to be responsible for the formation of some U deposits.  In contrast to uranium, Th has been thought to be immobile in hydrothermal fluids.  Our research has shown this to be untrue.  Indeed, our experiments indicate that Th is extremely soluble in sulfate-bearing aqueous fluids because of the formation of highly stable Th-sulfate complexes.  This finding has considerable implications for the exploration of Th deposits, as well as for radioactive waste disposal.

© Hydrothermal Geochemistry Group - Earth & Planetary Sciences - McGill University   /   issues: