Gold Exploration using Baro-acoustic Decrepitation
An oral presentation by K. Burlinson at the IMA congress, Budapest, 2010
Using fluid inclusion data to locate mineral deposits
There is a vast database of geological information on hydrothermal systems collected from fluid inclusion studies. Diabolically, this valuable data is completely ignored by the mineral exploration community when they are exploring for hydrothermal mineral deposits. In part, this is due to misinterpretations and wrong conclusions drawn on some early work in the 1950's. At that stage there was almost no understanding of the importance of gases in hydrothermal fluids and no understanding at all on the thermodynamics of gas-water mixtures and how they would affect fluid inclusion decrepitation. This led to completely incorrect conclusions about the decrepitation method and its subsequent disuse in western countries. However the method remained in use in soviet Russia through the 1970's and 1980's and also in China to the present day as an exploration technique.
We are now well aware of the importance and almost ubiquitous presence of gases including CO2 and CH4 in hydrothermal fluids and have a well developed understanding of the thermodynamic behaviour of such fluids, which leads to the realisation that we can use the baro-acoustic decrepitation method as a mineral exploration technique to identify samples with abundant gas-rich fluid inclusions. Such gas rich fluid inclusions are well known to be closely associated with many economic hydrothermal mineral deposits, particularly gold. Using this method we can analyse many samples in a large spatial array and map out contours of fluid inclusion gas contents to define anomalous areas of potential mineralisation.
The baro-acoustic decrepitation method is particularly appropriate as a mineral exploration technique as the analyses are quick and sample preparation is minimal, leading to a low cost per sample for the analyses and a fast turn-around time. The sample needs only to be crushed and sieved and there is no need for the costly and slow preparation of doubly polished thin sections. The entire analysis is performed in about 30 minutes by a computerised instrument which is completely objective and gives consistent and reproducible results. There is no need for any microscope work or the subjectivity that would entail.
The baro-acoustic decrepitation method detects ONLY the pressure pulses from fluid inclusions when they burst open due to their high internal pressures as they are heated. There have been incorrect comments that the method detects crystallographic effects rather than fluid inclusions. But such crystallographic changes in the sample generate shear (S) waves in the solid grains which cannot reach the pressure sensitive detector as shear waves cannot traverse the airpath (a fluid) which separates the sample from the detector. The detector only counts pressure pulses transmitted as P waves caused by release of high pressure gases or steam explosions from superheated water as fluid inclusions decrepitate. (Additional discussion here)
To understand why the baro-acoustic decrepitation method is effective in detecting gas-rich fluid inclusions we need only look at the Pressure vs Temperature graph as fluid inclusions are heated.
On this graph I show the P-T relationships for fluid inclusions of different compositions which all formed at the same hydrothermal PT point of 380 C, 100 MPa. (1000 bars).
When heating the pure water inclusion, its PT graph is shown by the blue and green lines. Initially the pressure does not increase much as the water expands to fill the bubble space. Only after the vapour bubble has gone at the homogenisation temperature of 310 C does the pressure increase substantially. When the pressure reaches some 600 bars at about 350 C the inclusion will decrepitate and this temperature can be used to estimate the formation temperature of the inclusion.
When heating gas-rich inclusions, the PT graph is shown by the red lines. The pressure in such inclusions increases linearly with temperature in accordance with the gas law, PV=nRT. Consequently the inclusions reach internal pressures of some 600 bars at quite low temperatures, as low as 200 C for a pure CO2 filled inclusion. This results in distinctive low temperature decrepitation and is the basis for the use of the baro-acoustic decrepitation method in identifying gas rich hydrothermal fluid systems. (Additional discussion here)
There are numerous accounts in the literature where observers performing microthermometric analyses complain that the fluid inclusions decrepitate before they homogenize and so they were unable to determine their homogenisation temperatures. The reason for this is obvious from the above graph, but this behaviour is considered a nuisance in such work and the exploration significance of this behaviour has been completely ignored. We need to stop focusing on thermodynamic perfection to see the immense potential of this behaviour as an effective exploration technique.
Some typical results from the decrepitation instrument show how easy it is to identify gas-rich fluids.
The peak caused by the presence of gas-rich fluid inclusions is clearly seen near 300 C on the red graph, as expected from the thermodynamics of gas-water mixtures. The blue graph is a sample which had only aqueous inclusions and this lacks decrepitation below 400 C. Both of these graphs show a peak at 570 C, which is a caused by the weakening of the quartz as it undergoes transition from the alpha to beta structure. Fluid inclusions preferentially decrepitate at this temperature while the quartz strength is compromised, which explains the absence of decrepitation above this temperature. The green graph shows the result on a sample that has already been heated previously. It has no response at 570 C, which confirms that the decrepitation method does not detect mere crystallographic effects and only detects fluid inclusions. Details on the quartz phase transition are here.
Some studies have been done where the thermally released gases during decrepitation have been analysed by mass spectrometer and these confirm that the low temperature decrepitation peaks are indeed related to the presence of gas-rich inclusions containing CO2 and CH4.
On this sample, from the Cowra creek goldfield, NSW, Australia, much CO2 is released at very low temperatures, matching the low temperature decrepitation counts observed on this sample. These fluid inclusions are not only gas-rich, but also large, resulting in the substantial amounts of CO2 detected by the mass spectrometer. An abstract of this work is here , and additional decrepitation data is here.
An additional study of a sample from the Pine Creek goldfield in NT, Australia carried out by D. Gaboury at U. Quebec also shows that the low temperature decrepitation events correlate with the release of copious amounts of CO2.
From the above discussion, we see that gas-rich inclusions give rise to prominent low temperature decrepitation peaks. Although CO2 is the most abundant gas in fluid inclusions, the low temperature decrepitation is independent of the type of gas as the P-T properties of all gases are similar and described by the gas law. A discussion showing the similarity of decrepitation of various gases is here. In mineral exploration both CO2 and CH4 are known to be strongly correlated with gold mineralisation and so the identification of low temperature decrepitation peaks can be used in exploration for gold deposits.
The decrepitation method is also used in China, where 2 instruments are known to be in current use. The Chinese instrument uses a completely independent design and analog electronics, in contrast to the computerised and fully digital BGS model 105 instrument.
These are the 2 instrument designs known to be in current use.
The same sample has been analysed in these 2 different instruments giving very similar results and confirming that the technique gives reliable and completely objective results with different instrumentation and operators.
This has shown the theory of the baro-acoustic decrepitation method.
The following data shows real world application of the method in exploration for gold mineralisation.
This mine is on the North Island of New Zealand.
The gold occurs in quartz veins in a swarm.
This is the deposit model for the epithermal quartz veins which are of low sulphidation type.
A suite of samples were provided by a research student from drill core within the mine. The locations and sample numbers are shown on this pseudo section of the mine.
The thick gray lines show the mine workings and the entrance decline.
The locations of samples 1884 and 1885 were incorrectly recorded by the student and the probable location is plotted instead.
These 4 samples from the Moonlight and deep in the Favona vein show very low decrepitation intensity, which is typical of epithermal quartz and also lack any low temperature decrepitation peaks, indicating the absence of gas-rich fluid inclusions. These samples have very low gold grades and are remote from known economic mineralisation. The decrepitation data correctly indicates the low mineralisation content of these samples.
Samples 1884 and 1885 are within the mine workings and have very high gold contents. They have more intense decrepitation and a weak but distinct low temperature decrepitation peak indicating the presence of gas-rich fluid inclusions in these mineralised areas.
The complete results of samples from the Favona mine are shown here.
A rigorous evaluation of the decrepitation curves can be carried out by using de-convolution to fit skewed gaussian curves to the observed data. The individual component gaussian populations can then be precisely measured and their temperatures, widths, heights and skewness values determined. Sample 1885, one of the high gold content samples, plotted with the green curve in the above decrepigram, is shown here as an example of this curve fitting (de-convolution). Here the raw data is shown with the black curve, while the best fit data, representing the sum of each of the 4 component populations, is shown by the red and yellow lines, which are superimposed. Note that there is very good agreement between the raw data and the fitted sum plots and we can be confident of the presence of 4 separate component populations of fluid inclusions in this sample. Note also that the population component plotted in green is centered at the low decrepitation temperature of 340 C, as is characteristic of decrepitation of gas-rich fluid inclusions.
Fitting results for all the data from the Favona project are shown here.
A discussion on curve fitting and the use of skewed gaussian functions is here.
The curve fitting results for all of the samples are shown on this bubble plot, where the circle diameter is proportional to the decrepitation intensity of that component population.
The better mineralised Favona main shoot (red) has more intense decrepitation and small but measurable low temperature, gas-rich inclusion decrepitation. The Favona South shoot (yellow) has less decrepitation and slightly less low temperature gas-rich decrepitation. And the Moonlight area (green) has much less low temperature gas-rich decrepitation. From this it is seen that the best exploration areas are in fact in the shallow areas of the Favona South shoot, which looks more promising that even the deep samples from the Favona main shoot (sample 1879). The data show that the Moonlight area is not a particularly promising exploration target.
Favona SummaryEpithermal deposits are difficult, but not impossible, to explore using baro-acoustic decrepitation. Despite the low decrepitation intensities, this data clearly shows a strong correlation between known gold grades and the presence of low temperature gas-rich decrepitation. Because gold is distributed inhomogeneously within veins, and gold analyses are complicated by the "nugget effect", using decrepitation is a much better way to explore these systems because fluid inclusions are far more abundant and homogeneous within hydrothermal systems and provide a larger and much more reliable target to define the gold potential.
The Brusson gold mine, ItalyThe Brusson gold mine in northern Italy has been known since roman times and exploited up until the mid 20th century. It is particularly interesting because gold occurs within a well defined quartz vein, but the gold abruptly ceases where the vein passes through carbonate host rocks. There have been numerous studies using fluid inclusions and stable isotopes to try and understand this cutoff of the gold mineralisation. However, none of those studies has provided any more than a vague hypothesis to explain this cutoff. The baro-acoustic decrepitation data has provided critical information about the quartz which has led to concise explanation of the gold cutoff in this deposit.
The mine is located in the Alps of northern Italy.
20 samples of quartz were collected at the adit entrances on each of 4 levels in the mine. It was not possible to enter the mine as the workings are no longer safe.
The locations of the 4 adit entrances are plotted here, based on GPS coordinates of the sample sites.
The geological cross section shows that the gold mineralisation occurs within the quartz where the vein is hosted by paragneisses at levels 1, 2 and 3. At level 4, the vein is hosted by carbonate rocks, and there is no gold in the vein at this level.
All 20 samples were analysed and when plotted together it is difficult to determine any pattern related to the mineralisation. But note that low temperature gas-rich decrepitation is intense and very common on many of these samples, correlating with the good gold mineralisation in these samples.
Using a logarithmic type scale, we can accentuate the low decrepitation intensity levels (around 100 counts), which are of most interest when interpreting decrepigrams. This re-scaled plot shows 3 separate types of clearly distinct fluid inclusion populations within the quartz, which are called types 1, 2 and 3. In this plot, the samples are also colour coded based on the mine level they came from and it is immediately obvious that there is no type 3 quartz present at the uppermost mine level. Although it would be nice to have a few more samples, statistical calculations show that the probability of this distribution pattern occurring at random on 20 samples is only about 0.07 (7 %). We can be fairly confident that this pattern in the data is real.
This observation of significant differences in the fluid inclusion populations of the quartz at the different mine levels provides a critically important clue about the reason for the gold cutoff where the vein is hosted by carbonate rocks. This information was not seen in any of the many previous fluid inclusion studies, carried out using microthermometry, because those methods focus too closely on individual fluid inclusions and fail to understand the whole quartz system and the hydrothermal system from which it formed. Sometimes you must stand further back to see the real information pattern within a picture!
Microscope photographs of samples of each of the 3 quartz types are shown below. Quartz types 1 and 2 seem to show the presence of some larger inclusions, but without first having the decrepitation information to classify the samples, it would be exceptionally difficult to distinguish visually between the fluid inclusion populations in these 3 samples. Usually, the differences seen here would be ignored while the search continued for suitably large and visible inclusions to study microthermometrically.
The inclusions in this type 3 quartz sample, H2043, seem to be smaller. But this subtle difference is usually ignored in microthermometric studies.
To quantify the differences between samples we use curve de-convolution into skewed gaussian component populations. (Explanation of why skewed gaussian curves are used)
This sample is from the uppermost mine level and is type 1 quartz.
The data from the de-convolution of the decrepigrams of all the samples in this study is shown here and is summarised in this plot, where the diameter of the circle is proportional to the decrepitation intensity (fluid inclusion abundance) of that component population. Note the absence of type 3 quartz (green) from the uppermost 1710m level. There also seems to be a decrease in both total decrepitation intensity and also low temperature gas-rich decrepitation intensity as the depth within the mine decreases.
Brusson mine model - based on an understanding of the fluid inclusion populations.The Fenilia auriferous quartz vein at Brusson was formed from auriferous fluids coming from the basement, in which the Au is transported as a HS- complex, as shown by L.W. Diamond. This was not, however, a continuous steady flow and the flow was episodic.
During periods of high fluid upflow rates, the effect of the weak fluid inflow from the host rock units was minimal. The vein-flow fluid temperature remained high and there was little chemical change so the gold present remained in solution and passed through the section without being deposited while type 1 and type 2 quartz was deposited.
During periods of low fluid upflow rate, the inflow of fluids from the paragneiss host rock unit mixed with the basement sourced fluid and lowered the temperature of the vein-flow fluid. This resulted in phase separation of the CO2 rich fluid into an almost pure CO2 phase and an aqueous dominated phase. H2S would partition preferentially into the gaseous CO2 phase, depleting the H2S and thus the HS- concentration in the aqueous fluid containing the gold. This caused the deposition of gold because of destabilisation of the Au(HS-)2 complexes when the HS- concentration in the aqueous phase decreased. This also resulted in the deposition of type 3 quartz containing few or no CO2 rich fluid inclusions from this now aqueous, CO2 depleted fluid. At the upper mine level 4, the stratigraphic inflow fluid from the marble and serpentinite has a high pH, possibly greater than 10. Mixing of this fluid with that in the vein increases the pH, which increases the solubility of the Au(HS)-2 complex and also increases the solubility of silica, despite any temperature decrease which terminates the deposition of quartz. Consequently the quartz at this upper level lacks gold and also lacks type 3 quartz deposition. It is this change in pH caused by mixing with high pH stratigraphic inflow fluids that stops the deposition of gold in the vein at this location.
And it is the recognition of changes in types of quartz, based upon the baro-acoustic decrepitation data, which makes it possible to explain this mineralisation cutoff.
A complete discussion of the Brusson data is here.
SummaryWe know that gold deposits are closely associated with gas rich fluid inclusion, but this relationship has not been exploited as an exploration method because it is so difficult to get consistently reproducible and representative CO2 contents of inclusion fluids on a large spatial array of samples. However, this information is available by using the baro-acoustic decrepitation method, and there is a sound theoretical understanding of the decrepitation behaviour of gas-rich inclusions. Although microthermometry can give more precise and intricate details about sub-millimetre scale events in hydrothermal systems, the baro-acoustic decrepitation method is far more appropriate when dealing with large suites of samples and real-world macro-scale features of the quartz and hydrothermal systems of economic interest.
Microthermometry is great if you are trying to understand the the minute details of a sample and reach thermodynamic nirvana.
But if you are interested in exploring for economic mineralisation, baro-acoustic decrepitation is the better method.