Sulphates and hematite found in major gold deposits

A notable feature of rocks of Archean age is an absence of oxygenated minerals, such as sulphates and hematite, reflecting a lack of atmospheric oxygen at that time. In the Witwatersrand placers this lack of oxygen lermitted uraninite and pyrite pebbless to concentrate; in later periods these would have quickly weathered. When sulphates and hematite are found in Archean rocks, it denotes unusual conditions. However, they are found in a number of major gold deposits, where they indicate that the ore fluids were relatively oxidizing.

The Geological Survey of Canada and the Derry Laboratory of the University of Ottawa have been investigating this anomaly. Attention was initially directed to this by massive barite within the Hemlo deposit. Sulphates and/or hematite are also present in the major deposits at Kirkland Lake, Ont., and Kalgoorlie in Australia; and have long been known in the McIntyre mine at Timmins, Ont.

There is another guide to oxygenated conditions — the isotopic composition of pyrite — which, in some respects, is more useful than the minerals noted above. This is based on measuring the ratio of the isotope of sulphur with mass 32 to the isotope with mass 34. This ratio is normalized, so that the composition of sulphur in meteorites and the mantle is considered to be zero. If the sulphur in a rock has always been in the reduced state, in the form of sulphide, it will retain this signature.

The great majority of sulphide in Archean rocks and ore deposits is close to this value. However, if fluids are oxidized, containing both sulphide and sulphate, the ratio will change in both molecules. The isotope with mass 34 prefers the sulphate and mass 32 opts for the sulphide. A useful indicator

Pyrite is useful as an indicator because it is ubiquitous in gold deposits. By contrast, sulphate minerals are much less common, since they can more readily be dissolved.

At the Lake Shore and Macassa mines at Kirkland Lake, some sulphate minerals are present, but they amount to much less than 1% of the ore. However, the isotopic composition of pyrite at –10 per mil tells us that there was a large amount of sulphate or sulphur dioxide originally present in the ore-forming fluids, perhaps as much as half the total sulphur.

The data from pyrite suggest that the most strongly oxidized fluids were involved in the formation of gold-telluride ores, of which the most notable examples are at Kirkland Lake and Kalgoorlie.

How does this information aid in understanding or finding gold deposits? First, it helps to identify the origin of the ore fluids. During the Archean there were very few sources for oxygenated fluids, The ocean lacked dissolved sulphate and the atmosphere lacked oxygen, so the fluids were not modified surface waters. Also, metamorphic fluids within the middle and upper crust were reduced, because of the abundant graphite in the sedimentary and volcanic rocks of greenstone belts. Only possible source

It has been argued that the only possible source for these oxidized fluids were oxidized felsic magmas. In more recent time oxydized magmas are known to have produced a variety of mineralization, including porphyry copper and porphyry gold deposits.

The next step in following up this idea was to discover whether oxidized magmas did, indeed, exist near to these deposits.

The oxidation state of a magma is revealed by the presence of certain minerals in intrusions. In reduced magmas, most of the iron enters silicate minerals, while in oxidized magmas much of this iron is instead incorporated into magnetite. Also, sphene only crystallizes from oxidized magmas. Work has confirmed that oxidized felsic magmas were, in fact, present at approximately the same time and place as gold mineralization was being precipitated from oxidized fluids at Hemlo, Matachewan and Kirkland Lake. We obtained a more precise indication of the oxidation status of the magmas by measuring the chemical composition of biotite. This confirmed that these magmas could have produced fluids containing a significant proportion of sulphate.

Work with John Clout at Kalgoorlie also identified the ore fluids as being oxidized, but we could not identify a nearby magmatic source. Porphyry intrusions within the deposit are minor in volume and older than the ore. But on the basis of a number of indicators we hypothesised that the ore fluids came from a magma at depth, composition of which was perhaps similar to that which formed the syenites at Kirkland Lake. Another interesting aspect

This leads to another interesting aspect in the formation of lode gold deposits. The ore fluids were invariably rich in carbon dioxide (CO2), as shown by ubiquitous carbonate alteration. The solubility of CO2is low in magmas and it steadily declines as magma rises and pressure drops. Thus magma could become saturated in CO2 at considerable depth, say 20 km.

This CO2 would then release from the magma as a fluid, taking with it some water and perhaps gold and other constituents. The mineralized fluid would then rise above the magma, being lighter, to form veins in suitably fractured rock several kilometres above. This may be what occurred at Kalgoorlie. Interestingly, this contrasts with porphyry copper deposits, which also formed from magmatic fluid. In the case of the porphyry coppers, the fluid involved was mainly water. The solubility of water in magma is much greater than that of CO2, so it is released only when the magma is within a few kilometres of the surface. Thus porphyry copper deposits lie within or very close to their source intrusions.

Why are hot, oxidized fluids so suitable for forming gold deposits? The answer lies in experiments to determine the solubility of this metal. At high temperature, gold is most soluble in oxidized fluids that also contain a good deal of dissolved sulphur compounds. Magnetic signature

How may this information help in exploration? One of the most practical approaches involves the magnetic signature of the felsic intrusions which were the source of the fluids, or are host to the mineralization. As noted above, oxidized magmas tend to form a good deal of magnetite, such that intrusions are relatively magnetic.

At Hemlo, the Cedar Lake stock next to the deposit forms an aeromagnetic anomaly. However, for reasons given elsewhere, we do not think that the Hemlo ore fluids necessarily came from the same pulse of magma that formed this stock.

In other instances, fluids passing through a partly crystallized intrusion can oxidize the magnetite partly or completely to hematite, which is non-magnetic. At Kirkland Lake and Matachewan, syenite intrusions that host gold ore have a high magnetic susceptibility over most of their extent. Near to ore zones this declines to almost zero as magnetite was entirely converted to hematite. Eion Cameron is with the Geological Survey of Canada.