baro-acoustic decrepitation study at the drake mineral field nsw, deconvolution and curve fitting to gaussian populations

Using Baro-acoustic decrepitation data in exploration at the Drake mineral field, NSW and the application of curve fitting to the decrepitation data

Melissa Gibbes and Malcolm Clark, Southern Cross University, Lismore NSW

Abstract

The Drake mineral field comprises fissure lodes, stockworks and disseminations of epithermal to mesothermal Au-Ag mineralisation, hosted within the Permian Drake volcanics. There are base metal sulphides and pyrite associated with the main mineralisation comprised of silver sulphosalts and native metal alloys. To better understand this complex hydrothermal system a suite of samples were analysed using baro-acoustic decrepitation of fluid inclusions to understand the hydrothermal fluid. The decrepigrams are complex and so deconvolution fitting of multiple gaussian distributions was used to distinguish individual fluid events. Initially, symmetrical gaussian distributions were used, but we show that using skewed gaussian distributions is preferable and provides better deconvolution and understanding of the fluid systems.

Introduction

The discovery of valuable mineral deposits is a long process, taking into account many aspects including: the structure of the geology, the grade of the ore, mineral processing procedures, the ongoing costs of operation, environmental impacts, plus predictions of future metal prices. Mining feasibility studies are long and costly and can be limited by a lack of geochemical, geophysical, geological and remote sensing exploration techniques available for a given location (Bateman and Jensen 1981). The use of decrepitation techniques to derive information from fluid inclusions in the mineral samples can help to assess suitable locations for mining exploration, improving the economic viability of the venture and potentially minimizing environmental disturbance ( Bateman and Jensen 1981, Craig and Vaughan 1981, Pomârleanu & Mârza 2002)

The study of fluid inclusions can provide valuable data on the nature of the ore fluid (Peach 1949, Roedder & Bodnar 1980, Edwards & Atkinson 1986, Burlinson 2006, 2007a, 2007b). Parameters that can be obtained include the temperature of formation, salinity and chemical composition of the mineral fluid. Upon reheating, the process that forms the fluid inclusions is reversed; this is utilized in the decrepitation study of inclusions (Edwards & Atkinson 1986).

Fluid inclusions are small to microscopic quantities of liquid that are trapped within a mineral as it crystallises (Peach 1949, Roedder & Bodnar 1980).Although the inclusions are usually homogeneous when trapped, once cooled to ambient temperature, they tend to separate into a multiphase inclusion, most commonly vapor and liquid (Roedder & Bodnar 1980). Sometimes during this process, daughter minerals, commonly halite, are formed as the fluid contracts and increases in concentration.

Gold deposits are frequently associated with CO₂ rich inclusions, which give a distinctive low temperature decrepitation peak (Partamines & Poutiainen 2001, Mavrogenes et al 1995, Burlinson 2006, 2007a, 2007b). This low decrepitation temperature is caused by the commonly high internal pressures of CO₂ rich fluid inclusions (Mavrogenes et al 1995). Vapour rich inclusions also suggest that the fluid was at boiling point at some stage during its accent to its current position (Edwards & Atkinson 1986). Research has shown that areas that are determined to be of epithermal origin are highly likely to have gold and silver mineralisation at some depth below the surface (Mavrogenes et al 1995).

Drake Geology

The Late Permian Drake Volcanics has widespread epithermal Gold and Silver mineralisation and has been mined intermittently for Gold, Silver, Copper, Lead and Zinc since the 19^th century (Snowden 1987, Bottomer 1986, Clark et al 2001).

The Drake Mineral Field is located within the area of a New England convergent plate boundary leading in the north to calc-alkaline volcanism (Bottomer 1986, Leitch 1975, Cawood 1982). This overlies the Razorback Creek Mudstone which then overlies the Gilgurry Mudstone (Beeson and Borton, 2005). The geology is described on the Drake 1:100,000 map by Thompson, J, 1976. The mineralisation of the Mineral Field is attributed to three styles:

Discordant fissure veins,
Stratabound stockworks and,
Stratabound disseminations.

Most mineral assemblages in the area are made up of a combination of these three styles (Beeson and Borton, 2005). The tectonic relationship of the Drake Volcanics is not fully understood as detailed published studies of this region have been insufficient (Bottomer 1986).

Figure 1. Location of Drake within Australia displaying geological features. (Source: Modified after Thompson 1976).

geology map

Mineralisation and Alteration

The mineralisation and alteration of the Drake Volcanics has been multistage, most accounts suggest that the depositions occurred over two episodes (Bottomer 1986, Perkins 1997, Smith 1989). The Mt Carrington site has large areas that have undergone pervasive alteration, including intense silica – sericite – pyrite alteration in the central Strauss pit and North Kylo areas (Beeson and Borton, 2005, Smith 1989).

This Study

The study site is located within the Mt Carrington mine site, situated in the Drake mining area in the Upper Clarence Catchment of Northern NSW, Australia (Figure 1). The Drake mining area is located within the New England Fold Belt (Lin et al 2003).

A total of 33 samples were chipped out of insitu quartz veins at three locations (Figure 2) on September 18^th 2007. Four samples were collected from North Kylo (S28˚ 54. 449’, E152˚ 22. 331’), 5 samples were from Strauss Pit (S28˚ 54. 579’, E152˚ 22. 382’), and 2 samples from Guy Bell (S28˚ 54. 761’, E152˚ 22. 441’). At each sampling point three replicates were taken from veins occurring within a 1m² area (Table 1).

Figure 2. Locations of sampling sites: North Kylo, Strauss pit and Guy Bell sites: (Source: adapted from Clark et al)

pit map

Table 1. Sample matrix showing sampling sites and replicate numbers.

Sample Site	Sample Location #	Replicate 1 Sample Number	Replicate 2 Sample Number	Replicate 3 Sample Number
North Kylo	1	1	2	3
North Kylo	2	4	5	6
North Kylo	3	7	8	9
North Kylo	4	10	11	12
Strauss Pit	5	13	14	15
Strauss Pit	6	16	17	18
Strauss Pit	7	19	20	21
Strauss Pit	8	22	23	24
Strauss Pit	9	25	26	27
Guy Bell	10	28	29	30
Guy Bell	11	31	32	33

The samples were weighed and photographed after collection. The weight of the samples varied from 13.7g to 775.9g. A visual assessment determined variation in the width of the quartz veins to be around 1 to 10cm across. The chemical composition of the samples also showed variation, from semi-translucent to milky grey and white, sometimes greenish quartz, jasper was very prominent is a few of the samples. Other minerals present include galena, haematite, and some of the samples showed signs of oxidation. The samples were analysed by Burlinson Geochemical Services in the Northern Territory on the BGS model 105 decrepitometer, as per methods used in Burlinson 1988 and described on this website.

De-convolution modeling

The data was curve fitted and de-convoluted using Plot® 0.997 software. Estimates of the position, width and height of Gaussian peaks were made and entered, the software then iteratively fits these Gaussian curves. Fitting is assumed complete when further iterations provide no further improvement of the quality of fit by minimising the RSD of the fitted envelope curve to the experimental data. An RSD of <5% was used as a benchmark for a good fit. Chi-squared is also kept to a minimum, but the overall effectiveness of chi-squared is more limited as it depends on the degrees of freedom in the sample: the more Gaussian curves to fit, the higher the degrees of freedom. The assumption used is that the distributions applicable would be symmetrical about the point. However, this assumption is shown to be invalid and skewed-Gaussian distributions should be applied.

After de-convolution was completed, the plots and data were analysed for trends, similarities and differences between sample site curve fits. The temperature of each Gaussian sub-population was recorded and plotted as a frequency distribution at 50˚C intervals: this was done for all samples combined, and for individual sites in the study area.

The results show multiple Gaussian distributions of temperature populations, which occur under curves that range from quite narrow (Figure 3) to very broad (Figure 4)

Figure 3. Sample 28 exhibits narrow distribution.

fig 3

Figure 4. Sample 18, exhibits broad distribution.

fig 4

De-convolution

Samples 10, 11 & 12 were fitted using both symmetrical and asymmetrical distributions (Figures 5- 10). These samples show what is present in most of the other samples, a platy-kurtic population curve leading into a lepto-kurtic population at a slightly higher temperature. This is prominent at just below 400˚C in sample 10 (Figure 5), when refitted allowing for a skewed distribution (Figure 6) the two Gaussian sub-populations become one skewed Gaussian sub-population. This trend is also easy to see in sample 11 (Figure 7) at 300˚C, and again just above 500˚C in sample 12 (Figure 9). This refitting simplified the curve fit, resulting in fewer population distributions within each sample.

Figure 5. Sample 10, symmetrical distribution.

fig 5

Figure 6. Sample 10, asymmetrical distribution.

fig 6

Figure 7. Sample 11, symmetrical distribution.

fig 7

Figure 8. Sample 11, asymmetrical distribution.

fig 8

Figure 9. Sample 12, symmetrical distribution.

fig 9

Figure 10. Sample 12, asymmetrical distribution.

fig 10

Jasper Rich Veins

Six samples were visually rich in jasper (Appendix 1.2) all displaying high temperature distributions with the low temperature ranges mostly absent (Figures 11-16). The curve shape of each of these samples is very similar, featuring one high intensity temperature peak at around 450˚C.

Figure 11. Sample 13, rich in Jasper.

fig11

Figure 12. Sample 15, rich in Jasper.

fig 12

Figure 13. Sample 22, rich in Jasper.

fig 13

Figure 14. Sample 14, rich in Jasper.

fig 14

Figure 15. Sample 19, rich in Jasper.

fig 15

Figure 16. Sample 23, rich in Jasper.

fig 16

Discussion

Data Complexity.

Multiple distributions due to temperature variation within a single event.

From the initial analysis of the raw data it appeared that each sample consisted of multiple distributions of data. Then each data set underwent curve fitting, using symmetrical Gaussian distributions. This revealed that in fact there were multiple distributions occurring within each sample, and that the data was quite complex.

Although the emplacement for a particular peak may have occurred in one event, there is temperature variation between the pulses of fluid over the duration of the emplacement.

As the vein opens, fluid coming up from below is hotter than the surrounding rock, which chills the fluid. An equilibration occurs between the fluid and the surrounding rock, resulting in what may be detected as a separate distribution beneath a larger curve.

Symmetrical or Asymmetrical “skewed” distribution?

Upon initial analysis of the raw data it was clear that the Decrepigram was displaying multiple populations within each sample, as such, multiple symmetrical Gaussian curves were chosen as the most appropriate method of statistical analysis. The resulting data plots from this method were interpreted over a period of time, but it soon became apparent that there was something not quite right about the shape and fit of the curves. Possibly due to the complexity of the software, and limited processing time, it took some time before the realisation that the distribution of the data would be better represented by a skewed curve.

This realisation was brought about by the common occurrence of a sequence of curve shapes. A platy-kurtic population curve leading into a lepto-kurtic population at a slightly higher temperature was determined to be a good candidate for a single distribution, which is skewed toward a ceiling before dropping away.

Samples 10, 11 and 12 were refitted using skewed distributions (Figures 5-10), the curve shape indicated a slow increase in decrepitation followed by a faster increase to a maximum, where the decrepitation counts fell suddenly. The re-fitted curve provided very similar fit parameters in terms of error distribution, but a much simpler fit with fewer distributions underneath the curve.

Variation due to wall thickness of inclusions.

Due to the variation in wall thickness of the samples an asymmetric distribution makes more sense. Temperature of decrepitation is likely to be affected by the position of the inclusion within the mineral. Inclusions that are surrounded by a thin outer wall mineral are likely to decrepitate earlier, even if the temperature has not yet reached the temperature of emplacement, as the thinner the mineral wall the weaker it invariably is. Inclusions that are embedded deeper within the mineral are likely to decrepitate closer to the actual temperature of mineralisation.

Another factor influencing the early decrepitation of fluid inclusions is the shape of the inclusion itself. Angular inclusions are known to stress at the corners which causes premature decrepitation. Due to the size of most inclusions, tedious preparation and microscope work would be required to determine if this was the case, this would require the traditional methods of ore microscopy be used.

A variation in the wall thickness of the inclusion would explain the tendency toward an asymmetrical skewed distribution. The frequency of the decrepitation pops creeps up as the inclusions with thin walls decrepitate, then the frequency of the decrepitation ceilings as the true temperature is reached, then drops off sharply at the end of the event.

Variation due to multiple deposition events.

The results indicate multiple mineralisation events, both in the physical appearance and in the multiplicity of distributions within the data. This is consistent with the theories of previous research (Bottomer 1986, Perkins 1997, Smith 1989). The composition of the fluid is not necessarily changing but the multiple distributions may be represented by different opening regimes. Further studies could clarify this issue.

Jasper rich Veins

The veins which were embedded within rocks with high levels of jasper generally show high temperature peaks only, low temperature peaks are mostly absent. This indicates first phase mineralisation, as jasper is iron contaminated quartz, formed from high temperature oxidising fluids. The types of veins exhibited here are skewed toward the mesothermal system, due to their high temperature range. These samples are not representative of the full range of temperatures present within the area, but do provide a good representation of the high temperature range.

Variation in distribution

The results show that there is a high level of variation in the distribution of temperature ranges (Figures 3 &4). The jasper rich veins exhibit temperature ranges that are spread over the high end of the temperature range, with one main peak, whereas other samples have temperature distributions spread over the high and low temperature ranges, some displaying sharp individual peaks, and others displaying very wide distributions that are not as well defined.

Variations in decrepitation intensity.

There is a noticeable variation of the intensity (frequency) of the decrepitation peaks. Although this study does not focus on decrepitation intensity, this is a point worthy of noting. This variation can be attributed to variations in the number of inclusions present in the sample and their size distribution, and/or the influence of the alpha-beta phase transition.

Conclusions

Mineralized quartz samples from the Drake area give complex decrepigrams with moderate intensities which indicates that the mineralisation formed at upper mesothermal to deep epithermal conditions from fluids of very variable composition during the mineralising event. Although some previous fluid inclusion studies using microthermometry were unable to obtain satisfactory data due to the small size of the fluid inclusions, baro-acoustic decrepitation is able to produce consistent and meaningful results on these samples. We used de-convolution to resolve the complex decrepigrams into individual, overlapping gaussian distributions of inclusions which are thought to represent major fluid composition events. There were as many as 11 gaussian distributions identified in the deconvolution of some of the decrepigrams (sample 11, symmetrical distributions), which may indicate the presence of highly varying fluid conditions and complex zonation of the mineralisation. The deconvolution was improved by using skewed gaussian components rather than symmetrical gaussian components. Using skewed distributions, the fit for sample 11 reduced to 8 components instead of 11. We suggest that the skew is related to the physical environment of the inclusions within the host mineral, with inclusions near grain surfaces or fractures being able to decrepitate at lower pressures than is normal for other inclusions from the same fluid event.

References

Beeson B, Borton D. 2005. The Mt Carrington epithermal Gold-Silver-Zinc system and the host Drake volcanics. Drake Resources. Company report.
Bottomer L. R. 1986. Epithermal silver-gold mineralization in the Drake area, Northeastern New South Wales. Australian Journal of Earth Sciences. Vol. 33, Issue 4, 457 – 473.
Bateman A. M. and Jensen M. L. 1981. Economic Mineral deposits. Revised Printing. John Wiley & Sons, Inc.
Burlinson K. 2007a. Acoustic Decrepitation as a means of rapidly determining CO₂ (and other gas) contents in fluid inclusions and its use in exploration, with examples from gold mine in the Shandong and Hebei provinces in China. Acta Petrologica Sinica.Vol. 01.
Burlinson K. 2007b. An updated understanding of Acoustic emission decrepitation. European Current Research on Fluid Inclusion (ECROFI-XIX). University of Bern Switzerland, 17-20 July, 2007. Abstract volume, p. 1.
Burlinson K. 2006. Hydrothermal fluid chemistry in exploration. Acoustic Decrepitation as a method of locating potentially auriferous quartz systems formed from CO₂ rich fluids. Australian Earth Sciences Convention, Melbourne
Burlinson K. 1988. An instrument for fluid decrepitometry and examples of its application. Bulletin de Minĕralogie. 111: 267-278.
Cawood P. A. 1982. Tectonic reconstruction of the New England Fold Belt in the Early Permian: An example of development at an oblique-slip margin. New England Geology, University of New England, Armidale.
Clark M. W., Walsh S. R, Smith J. V. 2001. The Distribution of heavy metals in an abandoned mining area; a case study of Strauss Pit, the Drake mining area, Australia: implications for the environmental management of mine sites. Environmental Geology. 40 (6) March 2001. Springer-Verlag.
Craig J R, Vaughan D J, 1981. Ore Microscopy and Ore Petrography. Wiley Interscience Publication
Edwards R P and Atkinson K. 1986. Ore Deposit Geology. Chapman and Hall Ltd.
Leitch E. C. 1975. Plate tectonic interpretation of the Palaeozoic history of the New England Fold Belt. Geological Society Am.Bull. 86, 141-144.
Lacerda L. D. 1997. Global emissions from gold and silver mining. (Abstract only). Water, Air, & Soil Pollution. Springer Netherlands. 97. 3-4/July.
Lin C, Maddox G, Bateman E, Clark M, McConchie D. 2003. Acidity and Major Cations in Surface Soils of a Sulfidic Mine Site, Australia: Implications for Mine Site Rehabilitation. Environmental Sciences, 10, 3 (2003) 165 – 173.
Mavrogenes J. A, Bodnar R. J, Graney J. R, McQueen K. G, Burlinson K. 1995. Comparison of Decrepitation, microthermometric and compositional characteristics of fluid inclusions in barren and auriferous mesothermal quartz veins of the Cowra Creek Gold District, New South Wales. Journal of Geochemical Exploration. 54. 167-175
Partamines S. A. G, Poutiainen M. A. J. 2001. Application of acoustic fluid inclusion decrepitometry to gold exploration in Finland. Geochemistry: Exploration, Environment, Analysis. 1. 109-117.
Peach P. A. 1949. A Decrepitation Geothermometer. American Mineralogist. Vol 34. May-June. 413-421.
Perkins C. 1997. Geological and chemical comparison of two epithermal precious metal systems: the Late Permian Red Rock deposit, Drake Volcanics, New South Wales, Australia, and the Miocene Gooseberry Deposit, Kate Peak Volcanics, Nevada,USA. Unpublished PHD thesis. University of New England, Armidale.
Pomârleanu V, Mârza I. 2002. Implications resulting from the study of fluid inclusions in the shear zone-type gold mineralization from Bozovici (Banat) and Someşul Rece (Gilâ U Mountains). Studia Universitatais Babeş-Bolyai, Geologia, XLVII, 1, 2002, 105-116.
Roedder E, Bodnar R. J. 1980. Geological pressure determinations from fluid inclusion studies. Annual Review. Earth Planet. Science 1980. 8 : 263 – 301.
Smith S. G. 1989. Geology and geochemistry of Permian epithermal deposits at Drake: a study of a precious metal-base metal hydrothermal system. BSc Hons thesis. University of New England, Armidale (unpublished).
Snowden H. H. 1987. Drake: Stories of Life and Times in an Old Mining Town.
Thompson J. 1976. Geology of the Drake 1:100 000 sheet. Geol. Surv. N.S.W. Sydney