Metallurgical test work – between a rock and a hard

place
A. Newell1, P.D. Munro2 and K. Fiedler3
1.MAusIMM, Executive Consultant - Processing, RPMGlobal Pty. Ltd., Brisbane, Queensland 4000.
Email: anewell@rpmglobal.com
2.FAusIMM, Principal Consultant, Mineralis Consultants Pty. Ltd., Taringa Queensland 4068.
Email: pmunro@mineralis.com.au
3.MAusIMM, Director, Mineral Processing Consultants Pty. Ltd., Flaxton Queensland 4560.
Email: kelvinfiedler@mineralprocessing.com.au

ABSTRACT
Test work is the foundation upon which the success of projects and subsequent operations is laid, reflected
in Professor Arthur Taggart’s pithy observation: ‘Make your mistakes on the small scale and your profits on
the larger scale’. Indeed, Professor AllanTrench has succinctly noted that ‘Get the metallurgical
characteristics wrong and you don’t have a project’.
With deposits becoming lower grade, more structurally and mineralogically complex, final products requiring
lower levels of impurities and tailings management demanding more attention, test work now assumes even
greater significance in the acceptance and successful execution of a project.
This paper is the first in a series aiming to stimulate thought on the metallurgical test work process by getting
back to the fundamental basics through exploring and providing some guidelines for the metallurgical testing
world, noting preferred practices as well as precautions and mitigations. This covers the key characteristics
of metallurgical test work programs, including what needs to be done, sample selection, quality control and
associated drivers, and most importantly, selecting, assessing and managing test work facilities.

INTRODUCTION
TEST WORK
This paper is the first of a planned series discussing the basis for metallurgical test work and the associated
issues and requirements and issues encountered when doing it.
The papers can simply be thought as fitting into the categories of ‘before’, ‘during’ and ‘after’ for a program of
metallurgical test work with this first one examining the ‘before’ stage.
We have been stimulated to write these papers both as originators of metallurgical test work and consumers
of the product because of our growing dissatisfaction with the standard of much of the metallurgical work we
encounter in our jobs. The papers do not aim to provide a completely detailed check list or ‘how to’ guide for
metallurgical test work but rather discuss key issues that need to be understood and resolved. We believe
that the industry must be more transparent and explicit reporting the results of metallurgical testing as a
precursor to achieving better outcomes.
Frankly, the “fuzzy” framework we currently have for metallurgical testing compares very poorly with the
much more structured environment that geologists operate under when producing a mineral
resource/reserve estimate. An excellent example of this is for the reader to compare the geology and
processing sections of say 10 reports for projects and operating mines prepared under the Canadian
National Instrument 43-101 Standards of Disclosure for Mineral Projects (Ontario Securities Commission
2016), hereafter referred to as ‘NI 43-101’. We have not singled out Canada for any reasons of poor
practices but rather because Canadian companies are required to disclose more technical data than
companies registered on the Australian Securities Exchange (ASX) or its counterparts in South Africa, the
United Kingdom and the United States of America. Also, the NI 43-101 reports follow a structured format for
the technical data.
Observations from comparing the quality of the technical data in geological sections with that in the
metallurgical processing sections of NI 43-101 reports; for the former:

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 • Samples are firmly located in 3-D mineralised space;
• Standards for quality assurance/quality control (QA/QC) in sampling and assaying are firmly
embedded in geological practice;
• Laboratory doing the assays is identified; and
• There is usually a statistical treatment of data.
Whereas for metallurgical test work:
• The concept of what the “sample” actually represents is poorly understood (this is discussed below);
• Not usual to quote precisions and accuracies for test procedures; for example people would be more
cautious using comminution test data if they understood the range of outcomes for a test at an
assigned confidence level; and
• Statistical treatment of data is rare; too many mineral processing engineers are prepared to draw
unwarranted inferences from a set of metallurgical test data without considering the range of outcomes
e.g. is a copper recovery difference of 1% absolute between different conditions really significant in
the context of the errors inherent in metallurgical testing?
An exemplar of QA/QC procedures used in metallurgical testing has been the work done on the
geometallurgical program at Olympic Dam (Ehrig, Liebezeit and Macmillan 2017).
The section ‘Essential Components of the Process Support in the Technical Report’ in the document CIM Best
Practice Guidelines for Mineral Processing (Sub-Committee on Best Practice Guidelines for Mineral
Processing 2011) discusses what should be in the NI 43-101 but not how to produce the information of which
metallurgical testing is the major determinant.
While we are aware that some companies have guidelines for metallurgical test work as mentioned by Colbert
(May 2018, personal communication) the industry does not. A seemingly honourable exception is the uranium
sector (International Atomic Energy Agency 1990).
The Global Mining Guidelines Group has a Working Group on industrial comminution efficiency (2018) and in
future might consider the topic of metallurgical test work.

WHY PROCESSING IS IMPORTANT

While it may seem obvious, the importance of processing, and thus the requirement for test work, needs to be
clearly elucidated:
• Many deposits require processing in order to produce a saleable product;
o Project economics are ‘sensitive’ to processing recoveries and product type, quantity and
quality
• Processing can be a significant, if not major, source of upfront project capital and ongoing operating
costs and thus a ‘critical path to revenue’;
o Processing has a major impact on infrastructure requirements, principally power, water and in
this paper taken to include tailings disposal. The total capital cost of these three items are
becoming an increasingly proportion of the total.
The project should be developed on a sound and competent basis because it can be time consuming and
expensive to address and correct ‘fatal flaws’ encountered upon start-up.
The first few years of operation are the ‘payback period’ when the revenues generated are principally
redirected to repayment of debt capital. Project failure for the providers of equity capital is imminent when
these loans are not repaid when the project has negative revenues.
Capital spent on ‘correcting’ the technical problems, assuming that it can be raised worsens the problem.
While Professor Allan Trench’s wry comment in the Mining News ‘Get the metallurgical characteristics wrong
and you don’t have a project’ is correct, it is also how the metallurgical information is applied as well as
executed.

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
WHY DO TEST WORK
Test work is conducted for a number of reasons:
• Develop a solution to a processing problem (existing operation)
o Includes potentially ‘novel’ or unconventional processing options
• Confirm, optimise or improve an existing flowsheet (existing project or operation)
o May include testing new equipment and process flowsheets, as well as new ore types
• Provide geometallurgical performance predictions in operating mines (existing operation)
o Characterize the process behaviour of ore types for a selected process flowsheet
• Develop a flowsheet and characterise metallurgical response (project evaluation)
o Supporting project studies such Scoping, Pre-Feasibility and Feasibility Studies;
• Generate engineering data for process design and infrastructure requirements (e.g. tailings,
acceptability of water source, power – project design);
• Generating products (project evaluation) for other studies, such as marketing, environmental,
downstream processing
o Extractive metallurgy studies; smelting or hydrometallurgy
o Tailings studies; beach angles for wet tailings disposal, ‘dry stacking’
• Test new equipment (e.g. HPGRs, Up Current Classifiers, etc.) or process flowsheets/circuit
configurations (innovative).
Tailings studies are important as mining projects are becoming increasingly driven by the ‘back end’.
Geotechnical test work can require significant amounts of ‘sample’.
Supporting proposed projects is the most common reason for test work and what the outcomes are used for
provides a foundation for understanding the range and depth of test work requirements.
Studies are used to assess the technical and economic feasibility in stage-gate driven process
(Scoping/Preliminary Economic Assessment, Pre-Feasibility Study and Feasibility Study) of proposed
projects, with the amount and type of test work conducted increasing with the study level, as outlined in the
Appendix.
One of the primary aims of study test work is to demonstrate a feasible and practical way of producing
marketable mineral products in terms of a robust flowsheet that can handle all feed blends presented to the
plant over the Life of Mine (LOM).
Another important study test work outcome is the determination of likely mineral recoveries and product
grades.
Outcomes from metallurgical test work drives study development by providing key technical and engineering
design input that allow:
• Development of mass, water and energy balances (once the mining rate has been determined);
• Determination of Process Design Criteria (PDC);
• Establishment of metal recoveries based on feed grades as well as product grades;
• Selection and sizing of equipment;
• Development of process capital and operating costs
o Consumable requirements including power and water
o Labour schedule
• Comparison of process flowsheets and equipment options (cf. trade-off studies);
• Infrastructure requirements (power, water, tailings storage requirements, etc.); and
• Support marketing studies and logistical requirements.

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
Anyone who reviews the process flowsheets and operating conditions of porphyry copper deposits all over the
world would realize that a generic treatment protocol would be:
• Grinding to 80% passing 150µm;
• Rougher/scavenger flotation time of 30 minutes;
• Regrind rougher flotation concentrate to 80% passing 38µm; and
• Multi-stage cleaning with the cleaner block tailing to final tailing; popular configuration is to use
washed froth cleaning backed up with conventional cells treating the tailing.
While this configuration might be applicable under most circumstances, consider the consequences for a
project whose ore’s comminution behaviour conforms to the Bond Equation if (a) the flotation feed could
have been 80% passing 212µm i.e. 21% of unnecessary primary grinding power installed, or; (b) the
flotation feed had to be 80% passing 106um i.e. throughput reduced by 21%. Debt finance will not be
available where the process plant has been designed on ‘gut feeling’.
The discussion by Lane, Whittering and Jeffery (2012) on managing process risk with the trade-off between
benchmarking versus test work is recommended reading. A test work program should identify risks to
process outcomes which can then be appropriately mitigated. The best economic outcome for the project
may be deciding to increase the capacity of a particular equipment item rather consume both time and
money on additional test work to get an ‘exact’ answer.

TEST WORK SHORTCOMINGS

When conducting due diligences on new projects supported by Feasibility Studies, we have observed that
fatal flaws typically arose from the flowsheet development and the subsequent quality of test work, including
the nature of samples, appropriateness of the test work program and the interpretation of results.
Specific test work concerns are:
• Poor sample selection (this is discussed in more detail below);
• Poor or limited understanding of ore type, ore variability and gangue types;
• Poor or insufficient mineralogical studies;
• Poor or insufficient process characterisation of ore types;
• Insufficient and/or inappropriate testing - poorly conceived test work plan;
• Where an initial separation is performed (e.g. pre-concentration or gravity), the subsequent testwork
(e.g. milling, flotation or leaching) is not conducted on the tailings;
• Poor interpretation of results during testing;
• Poorly established comminution properties;
• Inappropriate separation method(s) and non-optimised conditions;
• Inability to settle on a simple and practical robust flowsheet;
• Site water not tested;
• Poor understanding of separation recovery as a function of head grade and ore type;
• Poor understanding of process losses; and
• Poor characterisation of concentrate/product properties

TEST WORK REQUIREMENTS

Generic test work requirements are listed below. Actual ones for a specific project depend upon several
aspects including test work aims, sample availability, budget amount, etcetera. Bulk minerals (iron ore,
bauxite and coal) and smelting and hydrometallurgical extraction processes have significantly different
needs.
Examples of aspects to be covered:
• Sample

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 o Ore types – classification basis;
o Representative;
o How were they obtained and from where?
o Sufficient quantity available; and
o Appropriate type – reverse circulation drilling chips, diamond drill core, bulk
• Mineralogical studies
o Determines the separation route and thus the flowsheet
• Pre-concentration
o Establish potential
• Comminution
o Dedicated samples
o Parameters for crushing, primary grinding and regrinding
 Bond Mill Work Indices
 JK Drop weight and SMC
• Materials handling studies
o Ore recovery from bins and stockpiles
o Presence of clay and associated minerals
• Separation
o Method
 Gravity (e.g. coarse gold, minerals sands, hematite, tin, tungsten etc.);
 Magnetic (e.g. magnetite, wolframite, ilmenite, etc.);
 Flotation (e.g. sulphide for copper, lead, molybdenum, nickel, zinc; non-sulphide for
graphite. lithium, tin, tungsten etc.);
 Oxidation (e.g. pressure, bacterial, Albion, roasting [‘dead’ or ‘sulphation’]);
 Leaching (e.g. gold, copper, cobalt, nickel, uranium – heap, tank or pressure);
 Pyrometallurgy (e.g. sintering; roasting; smelting); and
 Electrostatic (e.g. zircon).
o Bench scale; continuous (Locked Cycle Test Work (LCT) for flotation or pilot plant) where
required
 Kinetic test work – modelling and vendor process guarantees
o Reproducibility studies – particularly with variable results;
o ‘Site’ water;
o Feed grade-recovery and final concentrate grade-recovery relationships
 Metal recoveries and product grades – as a function of ore type and head grade
 Important for production schedule and financial evaluation
o Understand metallurgical losses;
o Understand ore variability;
o Sufficient work to support flowsheet interpretation and metallurgical response
 Test Work
 Comminution – crushing, milling

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
  Materials handling
 Pre-concentration
 Dewatering (thickening & filtration)
 Ore variability
 Separation product grades and recoveries and product grades – as function of ore
type and head grade
• Product characterisation
o Full assay (penalty and credit elements);
o Sizing; and
o Other characteristics: friability, etc.
• Dewatering
o Concentrates and tailings - settling and filtration properties;
o Concentrates - Transportable Moisture Limit (TML);
o ‘Dry stacking’ of tailings, ensure appropriate range of separation tailings samples are tested
 Typical range of ‘fines’ (<10 microns) quantities, presence of clay

SAMPLES FOR TESTING

A Discourse
Unfortunately too many metallurgical test programs are fatally compromised in achieving their objective from
poor sample selection. A necessary precursor for sample selection is a coherent narrative of the deposit’s
geology that is intelligible to participants such as metallurgists, mineralogists and mining engineers.
Two questions that always asked in any review of work are.
• Where did the samples come from?
• What do the samples represent?
All samples taken for any metallurgical test work should be plotted on the ore body geological model to fix
their spatial positions. This helps determine whether sample coverage of the deposit has been adequate.
Thinking in spatial terms about all samples helps answer these questions.

Ore Types
A prelude to selecting sample types is undertaking what ore types are present and their likely abundances or
proportions.
Two paradigms control processing outcomes:
• Mineralogy controls metallurgy.
• Rock type and associated alteration controls comminution performance i.e. throughput.
An admonition is required about the loose use of terminology such as ‘rock type’, ‘lithology’ and ‘ore type’. For
example geological dictionaries define ‘lithology’ as:
• ‘The character of a rock described in terms of its structure, colour, mineral composition, grain-size and
arrangements of its component parts’ (Thrush 1968)
• ‘All those visible features that in the aggregate impart individuality of the rock’ (ibid)
• ‘Generally relates descriptions of rocks based upon hand-specimens and outcrops rather than
microscopic or chemical features’ (Whitten and Brooks 1972).

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
The metallurgist must not be unduly influenced by the terms ‘rock type’, ‘lithology’ or ‘ore type’ as described
by geologists logging alterations and visible features in the drill core. These may not coincide with a
‘metallurgical ore type’. This will be more properly determined by a mineralogist or perhaps only after the
results of chemical and/or metallurgical tests.
A ‘rock type’, ‘lithology’ or ‘geological ore type’ could contain several different ‘metallurgical ore types’ or
conversely a ‘metallurgical ore type’ can cover several different ‘rock types’.
Examples are:
• The flotation performance of a porphyry copper deposit with an epithermal overprint was driven by the
iron sulphide:copper sulphide ratio which was normally ~2 to 3:1 but reached 10 to 20: 1. It was much
more useful to think of the iron sulphide: copper sulphide ratio within a rock type than to compare the
metallurgical performance between different rock types;
• In a copper deposit geologists defined ‘supergene ore’ as material ‘having visible chalcocite’. This
ranged from (a) traces of chalcocite to (b) where chalcocite was the co-dominant copper mineral. Sub-
dividing ‘supergene ore’ into two separate zones according to the chalcocite content allowed accurate
prediction of metallurgical performance; and
• The flotation performance across three different rock types in the primary zone of a copper deposit
where chalcopyrite was the only copper mineral was identical at the same rougher feed sizing.
Material classified as ‘andesite’ also had chalcopyrite as the only economic sulphide with the same
texture but recovery was lower due to the significantly lower head grade.
Generally, at the Scoping study level, ore types with an abundance of less than 10% are not tested, unless
they will be an early feed to the process plant and/or have a characteristic that would significantly impact the
metallurgical response of blends incorporating them. This may be due the presence of clay, organic carbon,
penalty elements (arsenic, fluorine, mercury etc.), highly reactive sulphides, oxidised minerals, etcetera.
A good understanding of the deposit and the geological domains, namely the lithological units/rock types,
faults, alteration zones and oxidation zones, is required as a ‘first pass’ classification of possible ore types.
For separation test work samples in particular, the typical classification scheme is based on the following
characteristics:
• Degree of oxidation; hypogene (e.g. primary sulphides - chalcopyrite), supergene (e.g. secondary
sulphides – chalcocite) and ‘oxide’ (e.g. malachite, native copper);
o Interestingly, some geologists believe that there is a difference between weathering and
oxidation, so degree of weathering may be a more appropriate terminology
• Presence of in situ metal ions on mineral surfaces;
• Degree of alteration; increasing water content : hematite – hydrated hematite – limonite – goethite;
• Mineral:gangue ratios; e.g. copper sulphides:iron sulphides;
• Mineral:mineral ratios; e.g. galena:chalcopyrite;
• Impurity elements or gangue mineral; arsenic, antimony, mercury, clay, organic carbon, talc, etcetera;
and
• Combination of the above.
The focus must be on rock types modified by alteration when selecting comminution samples. It is
meaningless to have a comminution sample at 0.5% Cu described as ‘chalcopyrite ore type’ when it will be
the other 98.5% of its constituents that will influence its breakage properties; calling it ‘andesite – medium
alteration’ is much more informative.

Sample Selection
The objectives of the test work program must be clearly defined to determine the sample mass and particle
size needed for each section of work.
Once the test work program has been mapped out selecting samples is a critical stage underpinning the
meaningfulness of the program outcomes.

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
Poorly selected samples means ‘wrong’ results, incorrect project decisions and extra project costs, depending
upon when the error is recognised.
Key questions that need to be addressed:
• What are the samples to be used for?
• What are the results to be used for e.g. Studies?
• Test work type - comminution, separation, materials handling, geotechnical; and
• How much sample is required e.g. mass and particle size.
Typically three suites of samples are taken, namely characterization, comminution and separation samples
where the ‘characterization’ samples can be a sub set of ‘separation’ samples. The word ‘representative’ is
often used and means ‘typical of ore body or ore type (e.g. Life of Mine or Resource head grades,
mineralogy, rock types) or process stream’.
However, different minerals and different projects require different samples.
In the early stages of a project, such as for the Scoping Study, the deposit may not be well drilled out
(‘inferred’ status resource) and not fully understood nor defined and a ‘representative sample’ is often a best
interpretation.
Confirm that it is possible to select ore types or samples that will be representative which must involve
discussions with the geologists mindful of the following:
• Characterisation
o Mineralogy - mineral types, associations, size ranges, etcetera;
o Presence of penalty elements and ‘troublesome’ minerals;
o Properties useful in ore type classification
 Physical - hardness, density, fragility, abrasiveness, strength etc.
 Chemical - assays also size fraction-assays
 Other - magnetic, friability, radioactivity, fluorescence etc.
• Comminution
o Ore types – with a focus on domain or host rock types
o Variability - need to test the full range of extremities
• Separation
o Ore types
o Variability
 Head grades (ore type) – feed grade-recovery test work
 General variability
 At least 5 to 10% of deposit (depending upon deposit size)
 Range or extremes of properties
o Composites or blends reflecting the mining schedule first three years of production, later years
of production, etcetera.
• Source of samples
o Hard rock drilling – core recovery an issue?
o Tailings;
o Waste dumps;
o Bulk samples – trenching, trial mining, Bauer Rig, large diameter core drilling;
o Drill core samples : prefer diamond drill but sometimes RC or air core chips

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
  Note that a sample with larger dimensions (>75mm) is required for some tests e.g.
Crushing Work Index, Bond Abrasion Index
o Need to be sure that samples are not contaminated
• How much sample is required
o How many kilograms per test?
o Allow extra for contingency;
o Study stage and/or testing aim
 Scoping (75 to 200kg), PFS (500kg to 2t) or FS (1t to 50t)
o Nature of deposit - particularly number of ore types
 Large iron ore project;
 Porphyry copper with epithermal imprinting and oxidation zones; and
 Small high grade gold deposit.
Obtaining samples can be a challenge; sometimes getting them from the geologists can be like pulling teeth,
since they and corporate management view the drill core both as geological records and proof of drilling.
Dedicated drilling to provide metallurgical samples can resolve these issues and the extra meterage allows
confirmation of geological models.
Sampling a tailings dam is a serious exercise and should be based on historical deposition records to
understand not only the variation in grade but also mineralogy and particle size ranges (which vary for a
number of reasons during the history of an operation).
Sample ‘administration’ should include the following points.
• Organise record of samples. A dedicated person needs to be responsible for collecting samples and
establishing an inventory
o Important to have a detailed record of not only the samples, but also their intervals and sample
ID. Often the sample ID needs to be renamed for testing from drill intervals to more
understandable names
 Plan of sample location;
 Record of sample location co-ordinates and depth/intervals;
 Accurate geological logs are very important. Metallurgists must have geological
definitions clearly noted e.g. definition of secondary mineralisation;
 Observations of colour/smell/moisture; and
 Photographic record.
• It is important how samples are collected, packed as well as the mode of transport, particularly for
overseas samples
• Watch for:
o Contamination due to vegetation matter from near surface samples;
o Contamination due to hydraulic and drilling oil spills;
o Wind borne contaminants;
o Sample age – test work results can vary if drill core has been in core trays for a long time
 Minimise sample oxidation – nitrogen purge + refrigeration
 Sample storage at test work facility (hereafter called ‘laboratory’) – if oxidation an
issue then refrigerate
o Avoid wooden boxes and crates – chemical contamination from tannins, quarantine issues.
• Actual sample record and transportation to laboratory

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 o Ways to manage samples can include the sample stewardship and chain of custody (an
explicit requirement for geologists) – very important that some-one tracks samples all the way
from despatch on site to receipt at the laboratory;
o Sample tracking they can get lost or misplaced;
o Laboratory needs to cross check received samples and reconcile with what was sent
 Photos of samples received
• Site water samples can be very different to laboratory potable water
o Collect typical samples of site water; and
o Analyse site water and prepare ‘synthetic’ site water for test work.
• Importing samples/Customs and Quarantine Clearance e.g. gamma irradiation requirement for
importation into Australia; and
• Paperwork – most laboratories have established procedures and are very good in assisting with
requirements and wording of documents.

Composite Samples
The preferred methodology is to:
• Test individually; and if needed,
• Composite arithmetically or by ore type proportions based on the mine schedule.
Once the metallurgical characteristics of individual zones have been established then our previous
experience on a variety of ores suggests that plant performance should be the mean of the individual ore
zones weighted according to their proportion in the feed.
That is if there are two metallurgical ore types “A” and “B” then the metallurgical result from treating a mixture
of the two will be the weighted mean of the individual results from “A” and “B”.
The exception to this outcome is when there are deleterious chemical effects in one metallurgical ore type
which also affect the other one(s). An example would be a high amount of soluble copper in one
metallurgical ore type which degrades the flotation performance of any others mixed with it.
Hence the preferred route on undertaking the metallurgical evaluation of a deposit is test entities rather than
composites.
We have two objections to the practice of testing composites.
1. All compositing schemes assume a mine plan and associated production schedule thus introducing
the dimension of “time” into what is a purely “spatial concept” if we consider the deposit solely in terms
of its geometry. Since a credible mine plan and production schedule is supposed to be the outcome
of the feasibility study then how can these be accurately known at the start of test work?
2. Compositing invariably diminishes the investigators’ ability to determine the level of variability within
the deposit. We have encountered the absurdity of samples called the ‘variability composite’!
Lunt, Ritchie and Fleay (1997) discuss the use of composites and some finance providers specify using them
to confirm metallurgical performance for the initial years of production.
People preparing a composite must be fully cognisant of what they are doing; too often materials are mixed
together to provide a sufficient mass for a particular test.

Mineralogy
Mineralogy is fundamental to the development of a flowsheet and should underpin every test work program,
especially characterization of the material. It determines the possible mineral extraction techniques,
preparation requirements (i.e. primary grind size) as well as the need for any regrinding, likely metal
recoveries and concentrate grades.
Mineralogical techniques available include:
• Automated MLA/QemSCAN/TESCAN

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 o Sulphides, precious metals;
o Minerals present - economic, gangue;
o Grain size and nature of associations;
o Size ranges especially for liberation;
o Developing mineralogical expertise - reports too often contain a plethora of data but a paucity
of useful processing information because the mineralogist is not attuned to processing needs;
and
o Some difficulty with amorphous/non-crystalline materials.
• XRF
o Non-sulphide/oxide minerals e.g. iron oxides
• Infrared
o Clays, important in leaching operations
• Optical
o Good quality mineragraphy is extremely useful e.g. ‘first pass’ definition of metallurgical ore
types;
o Unfortunately a declining skill set because of the high degree of expertise and knowledge; and
o Too often confused with petrology by geologists.
• Samples
o Head samples, mill feed sample, separation process feed , intermediate products and tailings
samples
• Diagnostic studies
o Nature of precious and penalty element associations;
o Tailings for metal losses;
o Intermediate products for location and ‘recoverability’ of economic mineral; and
o Concentrates for nature of diluents and associations.

SELECTING TEST WORK FACILITIES

A suitable laboratory can be selected once the outline of the likely test work program been developed and
the availability of suitable samples resolved.
The first round of the selection process is understanding the facility’s capabilities, in particular the quality and
experience of the team (manager, metallurgist and technician), the nature of equipment and assaying
capabilities and the current work load.

Due Diligence
Conduct a due diligence on the proposed laboratory with the following focussing questions:
• Discuss the laboratory with colleagues and mentors
o Previous experiences
• Note location and ease of getting samples to the laboratory
o Can they handle overseas samples – importation paperwork, etc.?
• Laboratory capabilities – what can they do, have done and what they can’t do
o Clarity required on work that will be outsourced e.g. mineralogy, solid-liquid separation,
geochemical characterisation tailings, transport properties of concentrates

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 • Laboratory experience (at that site) with the commodity and separation;
• Examples of previous programs successfully conducted at that facility
o Provision of references
• Key personnel and experience
o Particularly technician(s) who will be conducting the separation work
• Amount of your time required to manage the laboratory
o This does not mean a ‘stream of consciousness’ e-mails to the laboratory on your daily
thoughts!
• Availability of facilities i.e. current work schedule;
• Schedule of rates;
• Can they contribute to the program?
o The best laboratories will identify and resolve processing problems of their own volition
• Do they produce a final report
o What is the nature of the final report? Be prepared to pay for a good one.
o Does it just collate results or provide opinions and guidance?

Key Requirements
Important criteria for scoring a laboratory are:
• Experience and competency
o Comminution;
o Gravity;
o Flotation;
o Leaching /hydrometallurgy; and
o Pyrometallurgy.
• Management and supervision
o Managing in your interest – not defending laboratory’s behaviour or deficiencies;
o Dedicated competent staff
 To your project – not several
 Not just recently hired off the street
• Quality of facilities – equipment, drying procedures, etcetera;
• Availability;
• Assaying
o In-house or sent to another analytical lab;
o Equipment and methodology
 Checks and balances (QA/QC)
o Turn-around time.
• Selection of Test Work facilities based on cost (i.e. avoid cheapest quotation)
o For the Test Work program supporting a Feasibility Study, the laboratory has to bring
something more to the table other than bottle rolls or a flotation machine and manual
manipulation of a froth scraper
• Set up a checklist of selection attributes; and

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 • Rank the test work facilities : using a weighting system (e.g. 1 to 5) for each selection attribute, which
is then rated (e.g. 1 to 10) for each test work facility with ranking determined based on the sum of the
weighted ratings.

Trends
In the last decade we have observed that:
• Equipment vendors are taking on EPC (Engineering, Procurement and Construction) /EPCM
(Engineering, Procurement, Construction and Management) contracts to ensure equipment sales, on-
going spares and services as well as occasionally contract operator :
o Generally will advise or do good test work programs since they will be measured – rewarded
with bonuses or punished with corrective costs – by a performance guarantee;
 Main issue is whether ore types and mining schedules have been satisfactorily
identified and developed
 Based on experience, flotation modelling is generally good but can be surprisingly
optimistic and may not include any allowance for scale-up
• EPCs/EPCMs are naturally keen to manage the test work program in order to ensure that they get all
of the studies and most importantly the construction. This ‘one stop shop’ approach is readily accepted
by junior mining companies that do not have any processing capability;
• When EPC/EPCM companies take over the test work programs’ customers an unkind observation on
some of these efforts is that the EPCM company’s main interest is the engineering and construction
hours, not the test work hours; consequently:
o Test work is limited in scope and of an unknown quality;
o May not fully recognise ore types and the mine schedule;
o May not address feed grade-recovery relationships;
o Too often a process flowsheet, equipment types and sizes are selected from previous jobs
(minimum engineering and maximum profits) prior to test work completion. There is a
tendency to ‘gloss over’ test work results that do not conform with these preconceptions; and
o How performance guarantees are prepared and achieved is unknown.

Recurring Problems
Common problems encountered include :
• Failure to take on a dedicated metallurgist early in the study cycle (e.g. exploration/Scoping Study –
company failing);
• Client not wanting to spend the money to allow the metallurgist to sit with the geologist/mining engineer
and acquire a much better appreciation of the idiosyncrasies of the deposit;
• Failure to recognise that ‘think time’ is required by the metallurgist evaluating the results; comparing
with previous test work/plant data and writing the test work report. Need to allocate sufficient budget
for a well analysed and well thought through report;
• Ignoring the laboratory’s capacity to provide a view ‘from its experience’. Many laboratories just want
to do the tests and send out the results. When selecting a laboratory it’s worth considering what it can
add to the technical conversation on the test work as the people shouldn’t be there to just provide a
flotation cell and scrape off the froth;
• Poorly defined purpose and outcomes from test work and what the results will be used for. What are
the required outcomes e.g. data for Process Design Criteria (PDC) and detailed engineering design?
• Incompatible constraints on the test work program – time and budget compared with expected
outcomes;
• Budget does not include any contingency – program predicated on ‘best outcomes’ and when these
don’t happen then what?;

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 • Sample issues – number, mass, preservation e.g. nitrogen purging of bags and/or refrigeration of
samples with sulphides, don’t dry samples for scrubbing or thickening test work;
• Getting samples to the testwork facility…never close to Christmas! and
• Lack of sufficient of ‘think time’ when evaluating results, and comparison to previous test work/plant
data.

PILOT PLANT TEST WORK

For many processes, a high degree of comfort exists with bench scale testing results if properly executed on
suitable samples. This arises because there is considerable process and engineering experience in scaling
up to a production plant.
Pilot plants are not required for the better known commodities such as copper, lead, zinc and gold employing
conventional mineral processing (e.g. crushing, milling, flotation and for gold, agitated tank leaching). Pilot
plant test work is rarely required, as long as good bench scale test work has been done as judged by Locked
Cycle Test (LCT) results and feed grade-recovery relationships.
For the record, the list of processes that normally do not require pilot plant testing includes:
• Comminution: crushing, SAG, rod and ball milling (based on simulation);
• Flotation - sulphide and oxide materials - Locked Cycle Testing (LCT);
• Gravity;
• Magnetic separation
• Electrostatic separation;
• Atmospheric leaching - acid (copper, uranium and nickel), carbonate (uranium) and cyanide (gold)
• In-situ leaching (ISL) - occasionally bench scale Test Work is conducted on drill core samples (poor
sample recovery) however typically a trial well field (cf. demonstration) is required
• Solvent extraction - copper, uranium – less exotic metals;
• Electrowinning- copper and gold;
• Thickening; and
• Filtration.
Pilot plants are used to demonstrate a flowsheet and the associated metallurgical performance for a
representative bulk sample. They are also used to confirm the metallurgical response under continuous
conditions (e.g. impact of recycle streams), including new ore types, the effect of flowsheet changes as well
as generate process data for engineering design.
Where a pilot plant is required for flotation, the best results are typically found based on a flowsheet
consisting of roughing-scavenging and one stage of cleaning and possibly two stages if there is a high mass
recovery and high throughputs. It is not recommended for multiple stages of cleaning, such as molybdenum.
Pilot plants are typically required for so-called ‘novel’ processes where there is nil or limited full scale
experience. Typical applications include:
• ‘Novel’/unconventional processes, usually associated with ‘rarer’ metals and/or new reagents,
particularly hydrometallurgical
o Demonstrating process and recoveries on a continuous basis, as well as managing impurity
build up (need for bleed circuits), scaling issues and unexpected reactions
o Testing materials of construction and materials handling (liquid/solid separations)
• Producing large quantities of marketing samples (e.g. graphite);
• Improving processing confidence; confirm, optimise or improve a bench scale flowsheet on the large
continuous scale – particularly for less common commodities (e.g. graphite);

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 • Generating or confirming process design criteria and engineering requirements, including materials of
construction;
• Demonstrating that the flowsheet will work for continuous operation including the recycle streams –
particularly relevant for hydrometallurgy; and
• Certain unit processes or equipment types
o AG mills;
o Column flotation – confirming carrying capacity and stage recovery;
o Heap leaching – column tests followed by a demonstration heap leach;
o High pressure leaching - pressure oxidation (sulphide destruction), HPAL, carbonate leaching
(regeneration of bicarbonate);
o Solvent extraction - more exotic separations e.g. rare earths, Ta/Nb, Cs/Rb, perhaps U/V;
o Allowing equipment vendor test work on larger scale samples
 HPGR, mainly for sizing and performance guarantee
o Investor requirement.
In the case of ‘novel’ processes, it is expected that a larger scale demonstration plant would be required
upon the successful completion of a pilot plant run. This step would be also necessary to convert a ‘resource
estimate’ into a ‘reserve estimate’.

Potential Pilot Plant Issues

Pilot plant problems include the following:
• So-called ‘bulk samples’ tend to be ‘accessible specimens’;
• Limited experience of personnel working on the pilot plant – have they done it before on the minerals
being processed?
• Insufficient sample to allow steady operation which can have a profound effect
o On one project bench scale test work results were good however a pilot plant run with
insufficient sample gave poorer results. As a result, the investors increased the interest rate
for financing because the pilot plant failed to emulate the bench scale work
• Not operating 24/7;
• Poorly sized equipment – either oversized or undersized
o Pilot plant operation can mask the effect of large circulating loads because of oversized
materials handling equipment such as slurry pumps resulting a shock to the production scale
plant
• No dedicated assaying facilities – very slow turnaround on metallurgical results;
• Milling and classification circuits don’t reflect either the bench scale or plant scale flowsheet and
equipment, and typically generate a different size range although achieving the ‘same’ P80;
• Ignoring the chemical engineering axiom that three residence time periods are required to determine
the effect of a disturbance or process change
o Three residence time periods includes the whole circuit including cleaners
• Using the pilot plant as a research tool to test different circuit configurations and treatment protocols.
The pilot plant’s role is confirmation not investigation; the latter is the domain of bench scale testing.

CONCLUSIONS
The industry needs better understanding and co-operation between its technical disciplines to improve test
work outcomes and thus project decision making,
For example, this means metallurgists working with geologists during the exploration stages to better
understand the nature of the mineralisation and the likely deposit characteristics so that ore types can be

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
identified as well as the probable make-up of representative samples. This may result in the need for
dedicated drilling programs to obtain suitable samples for test work as well as confirming the quality of
geological logging and the need for any drill core sample measurements.
With regard to mining engineers, metallurgists should be involved in the mining studies, providing estimates
to assist in the mining rate (and subsequent plant throughput capacity) and mine scheduling studies. The
mine scheduling needs to be critically reviewed in terms of ore type blends, feed grades and mining rates
and revised to ensure that it can be accommodated by the proposed plant design.
Metallurgists need to be more informative and transparent on what is being done in test work programs
because our efforts do not compare favourably with the more rigorous regime that geologists adhere to in
producing a mineral resource estimate.

ACKNOWLEDGEMENTS
The authors thanks colleagues and collaborators throughout the years who have freely shared their
experiences of the joys, frustrations and disappointments that come with any program of metallurgical testing.

REFERENCES
Ehrig, K, Liebezeit, V. and Macmillan, E., 2017, Metallurgical QA/QC – Who Needs It? The Olympic Dam Experience, in
MetPlant 2017, pp31-44 (The Australasian Institute of Mining and Metallurgy: Carlton)
International Atomic Energy Agency, 1990, Manual on Laboratory Testing for Uranium Processing, Technical Reports
Series No. 313, Vienna.
Lane, G., Whittering, R. and Jeffery, D., 2012, Benchmarking versus Test Work - How Should You Managbe Process
Risk? in Project Evaluation Conference 2012, pp. 49-52 (The Australasian Institute of Mining and Metallurgy:
Carlton)
Lunt, D., Ritchie, I. and Fleay, J., 1997, Metallurgical Process Development and Plant Design, in MINDEV 97 The
International Conference on Mine Project Development, pp. 111-124 (The Australasian Institute of Mining and
Metallurgy: Carlton)
Ontario Securities Commission, 2016, National Instrument 43-101 Standards of Disclosure for Mineral Projects, 05 May
Sub-Committee on Best Practice Guidelines for Mineral Processing, 2011, CIM Best Practice Guidelines for Mineral
Processing, April
The Global Mining Guidelines Group. Available from <http://www.globalminingstandards.org/ > [Accessed 14 May 2018]
Thrush, P.W., 1968, A Dictionary of Mining, Mineral and Related Terms, U.S. Bureau of Mines, p649.
Whitten, D.G.A. and Brooks, J.R.V., 1972, The Penguin Dictionary of Geology, Penguin Books, Harmondsworth, UK,
p273

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
APPENDIX

Study Level Minimum Test Work Requirements

Scoping Study
• Sample : representative, at least dominant ore types (cf. oxidised, primary) and primary diluent
o Quantity - ~ 50 kg with ~20 kg/ore type for separation and characterization; >30kg for
comminution
• Head assay, moisture, ore, bulk density;
• Mineralogy - optical, MLA/QemSCAN/TESCAN on separation feed (degree of liberation), diagnostic
leach (gold ores);
• Size fraction-assay for two crush sizes (pre-concentration potential);
• Grinding studies Bond work indices and/or SMC tests (DWi, SAG Amenability);
• Separation studies
o Rougher-scavenger sighter tests at least two grind sizes, reagent type/addition rate studies,
percent solids, concentrate re-grind;
o Cleaner studies;
o Gravity;
o Leaching - grind size, bottle roll, leaching time, cyanide concentration, etcetera
• Dewatering studies - settling and filtration, concentrates and tailings; and
• Concentrates - full assay and size fraction-assay.

Pre-Feasibility Study
• Sample - representative, ore types, annual mining year samples (years 1 to 3, years 4-5 and later
years), head grade, variability and diluent material
o Quantity - ~ 400 kg with ~60 kg/ore type for separation, characterization and materials
handling; >80kg variability; >200kg for comminution

• Head assay, moisture, bulk density;

• Mineralogy - optical, MLA/QemSCAN/TESCAN on separation feed (degree of liberation), diagnostic

leach (gold ores)
o If pyrite present, need to determine acid forming potential
• Crushing studies - Crushing Work Indices, Unconfined Compressive Strength, Point Load, RQD, –
also primary gangue diluent;

• Test pre-concentration options if appropriate – more than classification; magnetic, colour, SG, X-ray
Transmission (XRT), etcetera;

• Grinding studies - Bond work indices, SMC tests (DWi, SAG Amenability), Abrasion Index, size-assay
(pre-concentration), re-grinding
o Suitable range of samples
o Modeling/simulation of comminution circuit - higher throughputs and/or very hard ores : SAG
mills, ball mills, pebble crushers, crush sizes, recycle streams, etcetera

• Gravity studies;

• Materials handling studies including angle of repose;

• Separation studies :

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018
 o Use site water if appropriate,
o Rougher-scavenger tests - rheology, % solids, optimising primary grind size and concentrate
regrind, reagent type/addition rate studies, kinetic studies;
o Cleaner studies – circuit configuration, cleaner scavengers, etcetera;
o Locked Cycle Tests (LCT), duplication of LCT under best conditions;
o Leaching - optimizing studies, oxygen demand, rheology, etc. (gold ores); ferric leaching
(biological oxidation), fine grind/oxidative leach, etcetera (sulphide ores);
o Size-assay/mineralogy on tailings; and
o Effect of oxidation (sulphide ores).
• Dewatering studies - concentrates and tailings; settling, paste thickener; concentrates - filtration as
well as tailings for ‘dry’ stacking); and

• Concentrates : full assay, size fraction-assay and Transportable Moisture Limit (TML)

Feasibility Study
• Similar range to PFS but more extensive and more detailed, particularly comminution;
• Samples : representative, ore types, annual mining year samples (years 1 to 3, years 4-5 and later
years), head grade, variability and diluent material; and
• May involve a pilot plant.

Other testing requirements:

• Friability/degradation, Drop/Tower tests : hematite ores (lump/fines ratio);
• Product testing : Decrepitation Index (DI), Tumble Index, Sintering properties, Loss on Ignition (LOI),
etcetera (hematite ores);
• Self-heating e.g. fine-grained and pyrrhotite-rich concentrate;
• Concentrate transportation by pipeline :
o Rheological studies (viscosity – percent solids);
o Abrasion; and
o Corrosion.

14th AusIMM Mill Operators' Conference 2018 / Brisbane, Qld, 29–31 August 2018