The deportment of gold in plant feed, product or tailing samples is the primary driver behind how we can best recover that gold or in the case of tailings, why it was lost.  As the gold price climbs, recovery of every available ounce of gold becomes more desirable and consequently developing a fundamental understanding of gold deportment should be the first action of anyone involved in gold processing.

While it is of fundamental importance, the analysis of gold deportment poses some unique challenges that if not addressed can result in a lot of wasted effort for little gain.  These challenges are not insurmountable but do require the appropriate methods and expertise to be used.  While I have looked in the past at ways to simplify the process, the reality is that gold deportment studies still require experts in the field for them to be completed well.

Before getting to any actual analysis the most important thing to consider in a gold deportment program is sampling.  Gold by its very nature occurs in low concentrations as discrete grains.  This means that a sufficient mass of sample is required to gain representative results.  The key consideration in a gold deportment study should be the determination of the ideal sample mass.  With a basic understanding of the sample genesis and grade this can be achieved using sampling theories, such as Gy’s theory of sampling.  It is not unusual that the optimum sample mass could be in the order of 10’s or 100’s of kilograms and this should be considered when looking at appropriate methods.

Examination of the methods available for gold deportment analysis is a topic that I will cover in future posts.  The key consideration for any method or technique should be that it provides representative analysis for each potential gold form and carrier.  The form of gold relates to whether it is visible, sub-micron or occurs as a gold mineral.  The carrier relates more to the host for gold.  A good method should take all the types of form and carrier into account so that key populations are not missed.

Overall, if you are undertaking a gold deportment program the most important thing to think about is that all stages are comprehensive as possible.  If the sample collected is of insufficient mass or not representative of the bulk material then even the best gold deportment methods will not give good results.  Conversely, if the method used does not account for all possible gold populations then there is the chance that key populations will be missed.

If you have any questions about setting up or reviewing gold deportment programs then feel free to contact us and we can try to help.

Today I wanted to share some more thoughts on the usefulness of ore typing to mining and mineral processing operations. I have recently been working on a the beginnings of a geometallurgical program for an open pit operation where the whole processing chain from mining, through processing and to smelting is included. This has given me a unique opportunity to view ore typing from the perspective that although the mill is the direct customer of the mines ore types, consideration must be made for the effect that changing ore type can have on products fed to a smelter.

Examining this further I wanted to talk about some of the considerations that should be made when defining ore types on a mineralogical basis. The use of ore type definitions can provide vital information to a milling operation to ensure that variations in feed material do not adversely affect performance. Ore typing can be as simple as grade variations or definition of geological textures but I have been looking at taking it a step further and using fundamental mneralogical parameters as the basis.

It is now possible to use automated mineralogical techniques, like MLA or QEMSCAN, to quantify a huge number of process specific parameters within an ore sample and use these to group ores by type. This typing can be based on process specific parameters and elemental or mineralogical signatures can be used to differentiate these ore types at the blast hole and then track them through the process. The commerciLisation of infra-red logging technology, like GeoSCAN, will make this process even easier.

The key consideration in all this is which mineralogical parameters are best used to ensure that relevant information is fed to the mill. Mineralogical studies produce vast volumes of data for each sample and it is important to sift through and define what is really important for that specific application. For mineralogical ore typing I have found that definition of the key ore minerals, along with impurity minerals that may affect concentrate quality or general processing is the best starting point. For these minerals the most important parameter becomes the mean grain size and contextual setting of the mineral grains. Using this fundamental information it is possible to provide an ore type definition that is readily applicable for processing. From there it is simply a matter of identifying a suitable signature for that ore type that is easily measured and beginning to look at the process behaviour to draw correlations that can be used in predictive process control.

I realize this is only a very broad and brief introduction into the world of using mineralogical ore typing. I hope it helps you think on some of the possibilities having this type of information early in the mining cycle can bring. If you have any questions please don’t hesitate to contact us at MinAssist or leave a comment and stimulate some discussion.

This year we will be pushing ahead with development of this blog and other online services that will be aimed at enhancing the usability and appeal of mineralogy to the greater mining industry and beyond.  Over the last few months I have been more heavily involved in projects focussed on geometallurgy and will seek to explore some of the key aspects of this field through a number of targeted articles through the year.  The link between mineralogy and ore behaviour is increasingly gaining acceptance as a useful parameters for mineral processors and MinAssist will be continually developing programs through which understanding of ore types and characteristics early in the mining cycle can be used as a predictive tool to enhance overall recovery and minimise metal losses.  This is a very exciting field and I look forward to exploring the possibilities further.

Over the next few months you should expect to see a series of articles, hopefully more frequent than last year, that will explore all the aspects of using mineralogy.  The focus will be on new techniques in mineralogy and geometallurgy as core areas with articles covering the following areas:

1. First impressions and review of FEI’s iDiscover v4.3 package for analysis of QEMSCAN mineralogy data.
2. An overview of how I see the application of mineralogy in developing effective geometallurgy programs, bringing maximum benefits to mining and mineral processing operators.
3. More tips and tricks for using mineralogy in commodity specific applications.  This will include an exploration of where we have come in analysis of precious metals ores for operators to get maximum benefit from record high gold prices.
4. General observations on developments in the use of mineralogy, new areas we see mineralogy becoming more useful and ways to get maximum benefit for investment in this type of analysis.

Overall, I am looking forward to a great year and sharing with you my experiences, ideas and opinions.  Keep an eye on the MinAssist website for updated reference material and don’t hesitate to contact me with any questions of comments throughout the year.

To establish the theoretical grade-recovery curve for a material a mineralogical liberation study should be undertaken using tools such as QEMSCAN or MLA.  Using these tools in combination with standard flotation tests can quickly and efficiently define the efficiency of a process and where potential improvements may be made.

The following example shows the key aspects of using the theoretical grade-recovery curve for a copper flotation circuit where chalcopyrite is the major copper bearing mineral.  The material in this example was ground to P80 of 125 micron in a closed ball mill circuit prior to flotation.

Figure 1 shows the theoretical grade-recovery curve determined from the liberation characteristics of this material at this grind size.  It should be noted that the modal mineralogical distribution of this material contained 2% w/w chalcopyrite, which can be seen on the curve as the grade value at 100% recovery.  We can also see the proportion of fully liberated chalcopyrite particles was 30% w/w based on the recovery demonstrated for 100% chalcopyrite grade.

Process data for this material showed that a concentrate containing 80% w/w chalcopyrite at a recovery of 60% was achieved.  When compared with the theoretical grade-recovery chart it can be seen that this was significantly under the maximum expected recovery for this grade concentrate and hence an opportunity for flotation optimisation was present.

In this case optimisation of a series of key flotation parameters for this material resulted in an increase in recovery to 71% chalcopyrite while still maintaining a final concentrate grade of 80% w/w chalcopyrite.

Figure 2 shows that this correlated to a more efficient process with the actual plant recovery much closer to the theoretical grade recovery curve defined for that grind size.  A key thing to remember is that the actual achievable recovery can never fall on the right hand side of the theoretical grade-recovery curve as this would contradict the fundamental parameters of the material.

In our example the operators wished to increase recovery further even after optimisation of the flotation circuit.  To achieve this, the basic theoretical grade-recovery curve needed to be moved and to achieve this the liberation of chalcopyrite had to be increased.

The theoretical grade-recovery curve for chalcopyrite when the grind size was reduced to a P80 of 106 micron can be seen in figure 3.  It can be seen that the proportion of fully liberated chalcopyrite increased to 50% w/w and the whole curve was pushed right, giving greater opportunities for recovery improvements to be made.

The increased recovery at finer grind size can be seen in figure 4 to rise from 71% chalcopyrite to 85% chalcopyrite in a concentrate with 80% chalcopyrite grade.  This was a significant recovery increase and although the improved recovery was greater than the original theoretical grade-recovery curve suggested possible a change in the liberation characteristics allowed the operator to see the benefit. 

Finally, we should remember that the theoretical grade-recovery curve is generally defined for the value minerals in an ore and not based on the final element/metal to be recovered.  From the example here if we looked at Cu recovery the theoretical grade recovery curve would correspond to figure 5. 

In this case the maximum grade of Cu is restricted by the host mineral, hence to chalcopyrite the Cu grade cannot exceed 34.63% in the concentrate, even assuming 100% recovery of chalcopyrite.  This is a simply a function of the chemical composition of chalcopyrite with 34.63% Cu, 30.42% Fe and 34.94% S.  The grade at 100% chalcopyrite recovery therefore corresponded to the actual Cu head grade (assuming all Cu occurred with chalcopyrite) and was 0.69% Cu. 

Overall, the theoretical grade-recovery curve can be a very powerful tool for metallurgists in defining whether target recoveries and grades are feasible for the grind size of their operation and more importantly where recovery increases may be possible through optimisation.  Used well a simple mineralogical study to define liberation characteristics can offer a fast and easy way to define where efforts in plant optimisation should be directed for maximum benefit.

The use of process mineralogy in comprehensive ore characterisation has always been a benchmark application.  The depth of understanding that mineralogy can give us into the fundamental parameters of an ore can be invaluable in devising the most effective method for winning value from it.  The advent of new mineralogical technologies, such as the e-beam based QEMSCAN and MLA, has made this even more accessible, opening up possibilities for comprehensive characterisation of ores at any stage in the mine life cycle.

In my experience with process mineralogy, characterisation of materials can have the greatest impact on process improvement.  This is not restricted to development of new projects but is equally applicable in optimisation of all areas from grade control and estimation to minimising losses to tailings.

In essence ore characterisation can be applied to the following areas in the most traditional sense.

1. Definition of ore types in conjunction with geological setting as an indicator of mineralisation for exploration projects.
2. Ore typing and material definition for development of flowsheet options in new projects.
3. A tool for more effective grade estimation, control and variability analysis during mining.
4. A diagnosis tool for identification of process inefficiencies and fundamental optimisation of metal recovery.
5. An environmental tool for evaluation of waste rock and tailings for storage or clean-up.

The areas through which detailed ore characterisation can be used cover the entire mine life cycle.  It forms the benchmark for all the uses of mineralogy in processing or mining operations because of the fundamental parameters it examines.

Addressing ore characterisation is about more than simply looking at the mineralogy of a representative sample.  A good study will require an understanding of geology, mineralogy and metallurgy.  Integrating all of these disciplines will allow the drivers for ore formation to be established and how they will affect final processing.  Using this approach can be a very powerful tool as the first stage in any evaluation project and it addresses the fundamentals of the problem and draws together all aspects.

Generally, by utilising existing knowledge and smart use of automated mineralogical systems it is possible to generate comprehensive ore characterisation programs that are cost effective and valuable.  Much improved turn-around times can also be achieved, meaning that if characterisation has not been previously performed it is possible to integrate it into a troubleshooting program.

Overall, by using ore characterisation in a smart way there are significant cost savings to be made in many other areas of evaluation.  Effectively, it is possible to target specific areas of the ore behaviour once they are identified, eliminating the need to unfocused metallurgical testwork.  Doing this allows metallurgical work to focus on validating solutions rather than diagnosing problems.

If you have any thoughts or questions about ore characterisation please let us know.  This is a very broad overview and raises more questions than it answers but will provide the springboard for future posts examining each area of characterisation in more depth.

For those readers familiar with setup of QEMSCAN test programs a standard test is preliminary analysis of the bulk sample by XRD.  In combination with elemental analysis this is a vital step, especially for unknown samples.  It develops a level of understanding about the material that can focus the QEMSCAN analysis and help give sufficient information for important processes such as SIP development.

It may appear that the information generated by XRD is made redundant by more detailed information generated by QEMSCAN further down the project workflow.  It is true that an accurate mineralogical balance can easily be generated from QEMSCAN mineralogy on sized fractions if the mass distribution is known.  This will give an idea of the bulk mineralogy but will be subject to any errors that may have been introduced by statistical factors or errors in the mass distribution.  The bulk mineralogy generated by XRD will provide a validation tool for the accuracy of QEMSCAN mineralogy by back-calculation.

Simple validation of data is only one aspect that XRD brings to QEMSCAN programs.  Of greater importance is the accurate definition of major minerals that can be used in Species Identification Protocol (SIP) development.  For those unfamiliar with how QEMSCAN identifies minerals, the SIP is a master list and rules that assigns X-Ray spectra generated for each pixel to a mineral group.  The rules work on a hierarchy and can allow for broad or narrow elemental distributions within the specta depending on whether mineral groups or specific minerals require definition.  Generic SIP’s are generally available based on key minerals for various commodities; however, these need refinement for each new ore type to capture unique minerals.  The use of XRD to quantify the bulk mineralogy of a head sample can greatly assist with this refinement process as key minerals can be targeted and assigned in the SIP.  

What this topic also raises is the fundamental differences between XRD and e-beam systems, such as QEMSCAN.  Both techniques can define bulk mineralogy and have strengths and weaknesses in this area.  Fundamentally, the techniques differ in the manner data is generated.  XRD measures the unique diffraction pattern of X-Rays passing mineral crystals.  It is generally used as a bulk method with the average diffractogram generated for a pressed powder pellet of material.  In contrast, QEMSCAN examines discreet mineral particles in section with an electron beam and determines the mineral composition from elemental composition measured by characteristic X-Rays generated by the material.  Where XRD is a bulk technique, QEMSCAN is a microscopic technique.

Generating data in these ways means XRD and QEMSCAN can be used very effectively as complementary techniques.  QEMSCAN is very powerful in determining associations of specific minerals through imaging and generating loads of very useful data relating to these areas.  QEMSCAN can however be sensitive to statistical errors caused by the relatively small number of particles that can be analysed.  XRD cannot give the mapping capabilities of QEMSCAN but removes many of the statistical difficulties.  Hence, if used together they provide more robust and useful data.

This is just one area of QEMSCAN program setup that should be understood.  I will periodically investigate other areas that I believe are important but if you have any specific questions or areas that you think could benefit from clarification feel free to post a comment and I will look into them.

As operating costs increase and feed grades decrease it is important to maintain tight control of your process and ensure maximum recovery is maintained at all times.  Just a few days of losses in recovery caused by changes in ore type or processing conditions can have a large effect on total metal recovery and profit.  Mineralogical audits can be used to minimize these losses by monitoring the fundamental behavior of feed ore on a regular basis.

Smart use of mineralogical auditing can develop mineralogical and ore type benchmarks in a detailed initial audit that can be used in process control and monitoring.  A detailed initial audit of geological ore types can be used to identify key mineralogical parameters that may be used in routine and cost effective analysis.  Automated mineralogical systems like QEMSCAN are ideally suited to this because of the depth of information on everything from modal mineralogy to locking/liberation characteristics and mineral associations.

The advantage that automated mineralogy brings to auditing is the regularity of analysis that can be performed.  Traditional intervals for audits are yearly or quarterly.  However, the speed and cost effective nature of systems like QEMSCAN allow better mineralogical data to be collected at monthly or quarterly intervals.  This provides greater options for use of mineralogy in monitoring of key process streams and opens the possibility to use auditing in control for aspects such as maintenance scheduling.

Whole process audits on targeted streams provide information on all aspects of the operation.  It is possible to draw on other mineralogical information from ore characterization, tailings evaluation and risk reduction studies to build a comprehensive picture that is consistently developed upon.  Subtle changes in ore type can be picked up quickly and key process parameters tweaked to compensate for the new ore type.

Mineralogical audits can equally be applied to specific unit operations within the process.  This can allow more detailed information to be gathered on areas of the process that have been identified as critical.  This can be especially effective for flotation or oxidation processes where process parameters can be very sensitive to small changes in mineralogy.

Overall, regular mineralogical auditing can provide targeted information for definition of inefficiencies over unit operations or the whole process.  Smart use of auditing can establish trends in key process streams, define the effect of changing ore type and evaluate the effectiveness of ore blending from the mine.  MinAssist can help with development, implementation and then monitoring of these programs through the virtual mineralogist program. 

I encourage you to give us your thoughts on using process auditing.  Larger mining companies use these techniques very effectively but the advent of automated mineralogy has meant that there is no reason why smaller operations cannot effectively implement very beneficial programs.

Smart use of mineralogy in defining ore types and then monitoring their behavior through the process can help in numerous ways to improve the efficiency of control and key processes such as ore blending. 

The fundamental nature of mineralogical information means that the risk associated with processing and ore variability can be monitored and the effects predicted ahead of time.

The key for operators to get a good handle on the risk posed by ore variability is to take an approach that addresses the whole process life cycle.  Mineralogy focused just on the process can make a big difference in troubleshooting and optimization.  However, to really reap the benefits of technical risk reduction consideration needs to be made for what is being fed to the process and how that will affect recovery.

If a mineralogical monitoring program can be expanded to include feed material then the operation is on the front foot, changing the focus from reactive troubleshooting to predictive system control.

Knowing how a particular ore type will behave through the process before operations such as blending take place helps to ensure more consistent blends based on a larger range of factors than just grade are fed to the plant.  This can be readily achieved by undertaking an initial ore characterization program to define ore types based on mineralogy and correlate their behavior in the process.  Once detailed mineralogy is completed a more focused and cost effective monitoring program can be established that will provide the required data for effective risk control at an acceptable cost and turn-around time.

By using a long term approach to mineralogy as a tool to reduce the risk associated with ore variability and unknowns in the process.

There is some extra cost associated with using mineralogy as a monitoring tool over standard chemical assays but the benefit of gaining so much valuable information and reduction in recovery losses quickly account for it.

The advantages mineralogical monitoring can bring in risk reduction are something that we don’t often consciously think about but can make a big difference when ore variability is a factor and tight control of recovery is required.  I won’t address the specifics of how to set up a monitoring program that is suitable for this method of risk reduction but you can find more information in the GoldAssist Monitoring package.  I will also be presenting a paper at Mineral Processing Plant Design 2009 in Arizona later this year that addresses using mineralogy in technical risk reduction for roject evaluation that will give some more insight.

Mineralogy in monitoring and risk reduction is a tool that has really only found application among the big miners but could make a profound difference to any operation that would like better control of their systems.  Have a think about the advantages that being able to accurately predict changes in ore type could bring to your operation and feel free to throw up any questions or comments.

Of course, working with tailings is never easy.  Minor or trace concentrations of valuable elements or minerals means that sample collection and integrity become one of the major concerns in any project relating to tailings.  My expertise lies specifically in gold analysis and often the process of selecting and preparing an appropriate sample is more work than undertaking the analysis.

The rewards for getting sampling right can be significant.  Using your process tailings as a diagnostic tool gives insights into everything from what is being lost to why it is being lost and whether it will cause problems in the tailings dam.

Some examples of what a well designed tailings evaluation study can provide using systems such as QEMSCAN include:

1. The distribution of valuable elements with host minerals.  Identifying the major mineral host/s for unrecovered value.
2. The grain size of key minerals.  Can be used as a diagnostic tool in identifying inefficient process areas.
3. The locking/liberation characteristics of key minerals.  Can be used as a diagnostic tool to define root cause for losses.
4. The presence and nature of environmentally sensitive minerals such as sulphides or heavy metals.

As a diagnostic tool a well designed mineralogical study of your tailings can act as a fast and efficient means of first stage process auditing or ongoing process monitoring.  This is even more powerful when combined with mineralogical investigation of the feed (ore characterization).  This analysis of input and output streams can immediately show core inefficiencies in the process that can be targeted by more focused stream or unit operation audits.

Overall, I believe that mineralogical tailings evaluation should be the first step for anyone wanting to get a handle on the fundamentals in their process.  The use of automated systems like QEMSCAN or MLA makes this practical now.  While good sampling is still the key, these systems allow good statistics to be achieved and reliable data to be generated.

If you haven’t looked hard at what is in your tailings before I encourage you to think about the benefits a single study can bring.  Over and over you see historic tailings dumps with recoverable value still there.  It is much better to get the process right first time and make sure that you are getting maximum recovery from your process.

As a rule, from my experience the mineralogy of an ore is given lip service in process design and is generally considered the realm of the geologists.  Key ore minerals are usually identified but a deep understanding of the minor ore minerals and gangue is rarely obtained through feasibility, piloting and design.  Invariably, once commissioning is completed and the plant is operational the name plate recoveries are not achieved and a process of optimisation and general troubleshooting is undertaken.  It is at this point that smart project metallurgists get hold of a process mineralogist to help them dig into the fundamentals of why recovery is low or grade targets are not being achieved.

When done to support these optimisation or troubleshooting projects process mineralogy can often quickly and efficiently identify the fundamental cause of problems, hence allowing metallurgical testwork to focus on identifying and validating potential solutions.  When used early in the process, mineralogy can effectively focus the program and allow faster progression to results, minimising guesswork and ensuring that improved recoveries are achieved in the shortest possible time.

Where is it best used?

Process mineralogy can find uses in virtually all optimization and troubleshooting studies where ore processing is necessary.  Traditionally, this focuses around recovery or grade improvements in flotation or leaching.  However, there are numerous innovative ways to use mineralogy in everything from ore sorting or comminution to process contaminant identification.   Mineralogy can even be used in tackling problems in water management, through understanding of de-waterability or slurry properties, or environmental management.

These diverse applications also extend to commodities.  Copper, Base Metal Sulphides, Nickel, Industrial Minerals and precious metals have been the traditional domains for mineralogical studies but increased applications are being seen in Fe Ore, uranium and even extending to coal.

What this tells us is that process mineralogy can and should be the first thought in establishing optimization or troubleshooting programs in virtually any situation.  This can then be further extended for integration in ongoing monitoring of aspects such as feed variability or unit operation performance to reduce the risk of problems arising again.

Where does mineralogy fit in a study?

To use mineralogy most effectively it should be integrated right at the beginning.  Using mineralogy at this first step can help define the root cause of problems or inefficiencies and really focus a study.  This is especially important if the problem is a complete mystery but even when you have a good idea of what is going on mineralogy can be used to quickly validate those thoughts and give a good basis for justifying further expenditure.

Designed well, a mineralogical study can provide solutions that then must only be validated by metallurgical testwork.  This takes a lot of guesswork out of the process, eliminating wasted resources and allowing faster, more efficient completion of the whole project.

How can it be done?

The key to use of mineralogy in these types of studies is to utilize efficient integration with the metallurgical component of the testwork.  Numerous methodologies have been presented on how to achieve this and it is beyond the scope of this overview to explore all of them.  However, the key theme to remember is that mineralogy can be used to guide what areas should be focused on in metallurgical work.

The emergence of automated mineralogy as an industry, predominantly with QEMSCAN and MLA, has greatly enhanced the usability of mineralogy in metallurgical studies.  With these systems it is possible to get statistically relevant information quickly and for important areas that it was previously very difficult to deal with.  Huge amounts of data are produced in these systems but the most widely used are analysis for liberation/locking characteristics and mineral associations.  An understanding of these fundamentals, along with accurate mineralogy, can immediately show us things like why specific minerals are not being recovered, what other minerals they generally occur with and what needs to be done to liberate them.

In my opinion automated mineralogy, simply by making mineralogy more accessible, has removed most of the hurdles faced by metallurgists in its use.   The bottlenecks and ambiguity associated with manual optical mineralogy no longer exist.  Metallurgists can now generate mineralogical data tailored specifically to their projects and presented in a fashion that is clear and understandable.

Why go to the extra effort?

If I haven’t made this point enough so far, the reason for using mineralogy in optimization and troubleshooting is simply to streamline the process, giving better results both faster and cheaper.  In times like today, where every dollar is hard to come by, spending a little more up-front on a mineralogical study of material can save time and money later in the project.  For these types of project turn-around time is a key factor as every day recovery is not improved or sub-standard concentrates are being produced directly affects the profits at the end of the month.  If by using mineralogy as the first step in the project you take some of the guesswork out of the metallurgy and only have to validate the solution then you have immediately saved money and produced a faster result.

I have been working in this area for a number of years now and almost every project I have worked on an understanding of the mineralogy has pointed to the solution.  Often the mineralogy was done as an afterthought and used to explain why various approaches didn’t work.  I believe that by doing it last you are wasting an opportunity that could be gained by getting a direct understanding first up.

I encourage you to think about where you use mineralogy in optimization or troubleshooting and then where you could have used it.  I would love to hear your comments on whether it helped solve a problem and if you have any questions please post them or contact me directly.