Quantitative Mineralogy of Sub-Micron Gold

Gold deposits displaying mineralogical and textural variability complicate the gold recovery process. For instance, variability in the carbonaceous material found in a gold ore adds complexity when dealing with preg-robbing (Miller, Wan and Diaz. 2005).

Intricate by nature, ores can be explained in terms of the variability of the Au-bearing phases, the granulometry, and the ore’s modal mineralogy, regardless of whether the gold is refractory or free milling (Goodall and Scales. 2007). Each of the three characteristics mentioned here interact to influence gold recovery through a set of different mechanisms (Table 1).

Table 1. Feature of refractoriness in gold ores. (Goodall and Scales. 2007). Source: ZEISS Raw Materials

Modern microscopy techniques accurately quantify the presence of gold and describe its environment. Therefore, continuous monitoring of the ore’s texture and mineralogy leads to a better understanding of its behavior and an evaluation of how mineralogy impacts metallurgical recovery.

Light microscopy (LM) and scanning electron microscopy (SEM)-based automated mineralogy (AM) techniques help to characterize gold ores and estimate recovery. However, even with their widespread application, the information these methods offer is still considered limited to some extent. Some of the main challenges are acquiring information on grain size and the manual nature of the analysis when using the LM method.

There is an inability to account for sub-micron grains in a consistent or quantifiable manner in AM analyses, even though these may make up a considerable proportion of the grade.

This article looks at the developments in automated quantitative mineralogy (AQM) and how they can be applied to overcome the issues typically found when characterizing sub-micron grains.

It is possible to measure the dispersion of non-liberated, submicron (down to a few nanometers) gold phases within a host or hosts, consistently and quantifiably accounting for the presence of nanometer-sized gold and the mineralogical impediments to recovery.

Quantitative chemical analysis facilitates the diagnoses of gold mineralogy, hosting mineralogy, passivation, decomposition, and preg-robbing within a single analysis without introducing complementary chemical analysis.

Methods

During the study, three samples of ground ore were prepared, in accordance with standard metallographic techniques, into three 30 mm stubs labelled Stub 1, Stub 2, and Stub 3.

For each sample, ZEISS Mineralogic was applied to analyze a statistically representative number of particles. This method offers complete chemical, mineralogical, and textural classification and quantification of each sample.

Mineralogic 2D is a petrological analyzer based on an SEM (Figure 1) furnished with several energy dispersive spectrometers (EDS). Using its innovative image processing and analysis abilities, in combination with the full deconvolution and quantification of acquired EDS spectra, Mineralogic 2D delivers a complete set of outputs when the analysis is complete.

Figure 1. ZEISS EVO Scanning Electron Microscope Modular SEM platform delivers high resolution surface information and superior materials contrast. Mineralogic Mining on ZEISS EVO to maximize resource recovery. Image Credit: ZEISS Raw Materials

Classifications takes place in line and considers the sample’s morphological, chemical and physical features.

Since the chemical composition of each analyzed spot is established, Mineralogic 2D can offer an assay and distribution (deportment) of the elements of interest at the end of each analytical run. This removes the necessity to introduce supplementary techniques, such as electron microprobe analysis (EMPA).

ZEISS Mineralogic 2D offers an accurate and immediate AQM analysis solution that assesses the ore’s behavior and possible refractory behavior causes (Table 1).

Results

Results of the analyses of the three stubs show that quartz, pyrite and mica are dominant in the gangue (Table 2).

Table 2. Area %, weight %, average grain size, and average wt % composition of major gangue phases and value bearing phases. Source: ZEISS Natural Resources

Au-bearing phases present are native gold (Au_80–100 ), electrum (Au_20–80Ag_20-80), silver (Ag_80–100), calaverite (AuTe₂), hessite (Ag₂Te), petzite (Ag₃AuTe₂), empressite (AgTe), and sylvanite ((Ag,Au)Te₂). Au distribution in the samples is predominantly gold, calaverite, and petzite (Figure 2), with sylvanite making a slight contribution in stubs 2 and 3.

The average composition and proportion of each major gangue and Au-bearing phase are displayed in Table 2.

The software automatically considers scenarios in which the volume of interaction of the analysis interacts with two or more phases, generating a reported composition that is indicative of the phases sampled, as opposed to that of a sole phase.

Such aberrant compositions are processed automatically throughout the analysis, and classification algorithms are applied to make adjustments. This is a particularly useful feature when, for example, tellurides are <1 μm and contained in another phase.

Figure 2. The majority of the Au present in the 3 samples provided is found in native gold, calaverite, and petzite, with sylvenite making a minor contribution to stubs 2 and 3. Image Credit: ZEISS Raw Materials

The modal mineralogy of the samples can be evaluated to reveal the potential causes of refractoriness. Decomposition is an improbable cause since the sum of phases that may induce it contribute <1% of the composition of any one sample. Equally, the same can be said for preg-robbing as carbonates make up <1% of either sample.

The comparatively low number of grains for Au-bearing phases means that it is not possible to perform a true assessment of the hosting characteristics and another analytical method will be necessary.

Based on the analyses of 2D sections, all AQM analyses attempt to reconstruct a 3D interpretation of the sample. For this stereological approach to be successful, a reasonable number of particles must be analyzed. In this case, the analysis of a large number of particles should offer confidence that the analyses are statistically representative of the modal mineralogy and the average composition of the phases.

With Au-bearing grain numbers being less than 43 per phase per sample, the stereological confidence level for these phases is low and the analyzed mineralogy hosting Au-bearing phases may be deceptive.

Additional analyses must be conducted to validate the granulometry, the hosting mineralogy and possible causes for passivation. A repeat analysis of the sample to cover every particle present is not necessary and will not enhance the modal mineralogy or compositional information.

A bright phase search (BPS) analyses, focused on the Au-minerals, facilitates the evaluation of all the particles of interest in a reasonable time and offers a suitable way to proceed. In its most basic form, the advanced image processing engine in ZEISS Mineralogic is applied during BPS to distinguish between the lightest and heaviest phases.

Only the heaviest phases and their contact associations will be examined. All the Au-bearing grains in a sample can be quickly analyzed and good stereological information acquired, as well as offering comprehensive information relating to compositional and phase association. For brevity, only the results of the BPS of Stub 1 are displayed (Table 3).

Table 3. Average grain size, and average wt% composition of major gangue phases and value bearing phases for the Bright Phase Search (BPS) of Stub 1. Source: ZEISS Natural Resources

A BPS offers a series of advantages compared with conventional analysis. In a conventional AQM analysis, there should be a balance between magnification and the time cost. A general compromise that affects the smallest grain is usually applied in favor of minimizing the time cost.

Where the required information is key to the association and liberation of phases whose dimensions are measured in units of microns, as is the case here, it is also worth taking the smaller grains into account. Accounting for these smaller grains should not negatively impact the grain size measurements (Figure 3) or the overall assessment of the contribution of the phase to the modal mineralogy (Figure 4).

Figure 3. Number of particles per size step for each major Au-bearing phase. Image Credit: ZEISS Natural Resources

Figure 4. Proportional volume, per major gold mineral, for grain sizes between 0.2 to 0.5 μm, 0.5 to 1 μm, 1 to 2 μm, 2 to 10 μm and larger than 10 μm grains. Image Credit: ZEISS Natuiral Resources

The time saving BPS analysis offers can, and is, in part spent enhancing the imaging resolution so that smaller grains can be identified and analyzed (Table 3 and Figure 3). Note that the modal mineralogy of the BPS analysis demonstrates a bias towards particles that are comprised of heavy minerals as it ignores particles that only contain gangue phases.

The BPS analysis does not report modal mineralogy. The average grain sizes of the Au-bearing phases of the BPS analysis significantly vary from those of the traditional analysis for Stub 1 (Table 2), but the increased number of analyses offers greater confidence in the grain size average measurement and of the hosting characteristics of the Au-bearing phases.

Native gold stands out as the main Au-mineral, offering 84 wt% of the Au found in the sample, with the second most prominent phase being calaverite, which makes up 10% (Table 4). Both minerals can be categorized as locked (Table 5), where free surface liberation is concerned, with 66.9% of the gold and 72% of the calaverite being inaccessible for processing.

Table 4. Au distribution across the 4 main phases in Stub1. Native gold far outweighs the other three contributing phases. Gold and calaverite account for 94% of the gold present. Source: ZEISS Natural Resources

Table 5. Free surface liberation. Gold and calaverite, the two main Au contributors to the sample, are either locked (less than 10% of the grain’s perimeter exposed for reaction) or fully liberated (100% of the grain’s perimeter available for reaction). Source: ZEISS Natural Resources

To make the issue more complex, there are two gangue species holding the Au-bearing mineralogy. Native gold has a close association with quartz and a lesser association with pyrite. The tellurides have very little to no association with quartz while having a major if not total association with pyrite Figure 5.

Figure 5. Contact associations of most relevant minerals. Two mechanisms for hosting the Au bearing mineralogy are immediately evident. Native gold has an important association with quartz and a lesser association with pyrite. The tellurides have no or minor association with quartz and major if not total association with pyrite. Image Credit: ZEISS Natural Resources

In this case, refractoriness is the result of the textural setting or lack of liberation and the hosting characteristics. While recovery of native gold could be improved by grinding the samples further to increase liberation, it is probable that the recovery of the tellurides will necessitate baking the host to free the value.

Conclusions

The data Mineralogic Mining provides highlights the potential causes for refractoriness in gold. The fully quantitative nature of the method offers information on the textural setting, gold mineralogy, decomposition, potential passivation, hosting mineralogy, and preg-robbing.

The analysis was improved with the introduction of a bright phase search algorithm. The BPS analysis enhanced confidence in the grain measurements and the validity of the hosting mineralogy and liberation information supplied, facilitating a deeper understanding of the ore and its processing characteristics.

References

1. Goodall, W.R. and Scales, P.J. 2007. An overview of the advantages and disadvantages of the determination of gold mineralogy by automated mineralogy.
2. Miller, J.D., Wan, R.-Y. and Diaz, X. 2005. Preg-robbing gold ores in Developments in Mineral Processing. Vol. 15, 937 – 972. Elsevier

This information has been sourced, reviewed and adapted from materials provided by ZEISS Raw Materials.

For more information on this source, please visit ZEISS Natural Resources.