1 COPPER HEAP LEACH TESTING, INTERPRETATION AND SCALE UP Graeme Miller and Tim Newton ALTA Copper Hydrometallurgy Forum, QLD, 1999. CONTENTS 1. ABSTRACT 2 2. INTRODUCTION 3 3. TOTAL COPPER VS RECOVERABLE COPPER 4 4. TEST PROGRAMME OVERVIEW 5 5. COLUMN TEST DESIGN 9 6. PILOT LEACH HEAPS 13 7. CALCULATION AND PRESENTATION OF RESULTS 14 8. INTERPRETATION OF RESULTS 16 9. SCALE UP 19 10. CONCLUSIONS 21 11. ACKNOWLEDGEMENTS 22 12. NOMENCLATURE 23 13. REFERENCES 24
27
Embed
Copper Heap Leaching Testing, Interpretation and Scale Up
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
COPPER HEAP LEACH TESTING, INTERPRETATION AND SCALE UP
Graeme Miller and Tim Newton
ALTA Copper Hydrometallurgy Forum, QLD, 1999.
CONTENTS
1. ABSTRACT 2
2. INTRODUCTION 3
3. TOTAL COPPER VS RECOVERABLE COPPER 4
4. TEST PROGRAMME OVERVIEW 5
5. COLUMN TEST DESIGN 9
6. PILOT LEACH HEAPS 13
7. CALCULATION AND PRESENTATION OF RESULTS 14
8. INTERPRETATION OF RESULTS 16
9. SCALE UP 19
10. CONCLUSIONS 21
11. ACKNOWLEDGEMENTS 22
12. NOMENCLATURE 23
13. REFERENCES 24
2
1. ABSTRACT
The testing of copper heap leach systems has undergone radical revision in the last five years
in an effort to obtain better interpretation and modelling of field results. A number of workers
have stressed the importance of establishing a project specific data base of different
characterisation and leaching responses which enables proper planning of the mining and
placement sequence.
Testing methods have been refined to allow information to be gathered on both the early
leaching and that carried out in mature leach solutions with high concentrations of dissolved
gangue mineral ions. Both are important in establishing the leaching rate and the acid
consumption rate. Further testing methods have been developed to more accurately define the
different leaching regimes of oxide and sulphide dominated minerals. Use of appropriate
bacterial cultures is needed to properly leach the secondary sulphides.
Interpretation of the leaching tests has been greatly assisted recently by the availability of
excellent data on large-scale field leaching trails from a number of Australian copper operations.
This has allowed the development of two distinct models that allow the oxide and sulphide
mineral leaching to be interpreted more reasonably. The method of interpretation takes into
account not only the copper extraction kinetics but also the acid consumption dynamics.
Sufficient test work needs to be planned to ensure that the acid consumption response to
alterations in acid concentration can be assessed.
The scale up from columns to field expectations still relies on the application of some factors to
account for the non-ideal conditions encountered in the field. Some previous work has been
reassessed to provide an idea of the effect of this on the leaching parameters. The copper
extraction and acid consumption models have been used as the basis of scale up to predict the
field response to the design conditions. With sufficient test work knowledge the effects of
altering the acid concentration during the leach cycle can be assessed and applied to the
economic modelling of the project. These new tools also allow the full assessment of operating
parameters such as heap height, head grade, material bulk density and irrigation rate before
decisions on the selected values are finalised.
3
2. INTRODUCTION
To ensure the success of copper heap leaching projects, the metallurgical characteristics of the
orebody must be well defined. An extensive test programme should be initiated in the early
stages of development to determine the effect on the project’s feasibility, and to provide
information needed for the plant design. Where a bankable feasibility document is required, the
results must stand up to the scrutiny of third parties.
Orebody/reserve data, basic mineralogy and lithology should be known prior to planning the
leach test program. Representative samples are then required for each ore category. The
testing program can then proceed in stages, starting with more detailed mineralogy, size
distribution, elemental distribution, total and soluble copper distribution and preliminary
assessment of the leach response in small scale agitated tests. Subsequent stages can then
be planned on the basis of these results.
A series of column leach tests form the most important part of the testing programme. The two
key results to be ascertained are copper recovery and acid consumption. Operating parameters
such as leach cycle time, crush size, lift height, and acid addition methods can be optimised. A
program will often consist of a set of small column tests, to get a rough idea of the metallurgical
characteristics, followed by more detailed testing to optimise the conditions. A set of tall column
tests can then be conducted to confirm the results under conditions appropriate to a commercial
leach heap. Physical tests and water tests are also conducted to provide additional design
information.
For realistic predictions of full-scale production, column tests should closely model the
anticipated plant conditions. A set of laboratory procedures has been developed for this
purpose, and this paper includes a discussion of the technical aspects that must be considered
to ensure such resemblance. Suggestions for the presentation of results are also included,
along with a discussion of scaling up methods.
4
3. TOTAL COPPER VS RECOVERABLE COPPER
Total copper assays do not normally indicate the material amenable to recovery by heap
leaching. The solubilities of the different oxide and sulphide minerals in acid/ferric solutions
vary greatly (Parkinson and Bhappu, 1995). Assay procedures have been developed so that
the copper grade can be expressed as `recoverable’ copper.
For carbonate/chrysocolla (`oxide’) leaching, a simple acid leach can give a reliable result. For
other oxide minerals, sulphide ores and mixed oxide/sulphide ores, which will be bio-leached, a
more complex method is required. The determination of the soluble components, and hence
the proportion recoverable by leaching, is called a diagnostic leach or sequential copper
analysis (Hiskey, 1997). This testing technique is based on oxide minerals being, in general
readily acid soluble, whereas bio-leachable (secondary sulphide) minerals are generally cyanide
soluble. In certain instances an acetic acid digest is also used to distinguish carbonate minerals
from chrysocolla and other oxides.
The basic procedure is to conduct a sequential cold acid, cyanide and aqua regia digestion to
establish the level of soluble copper. The sum of cold acid and cyanide soluble models the
long-term ferric/acid soluble species. The sequential method is important as species such as
cuprite and native copper are not totally CN soluble and need to be acid leach beforehand. The
CN test alone is also very difficult to get reproducible results from those species that have a
high acid solubility.
Leach recovery is best expressed as a percentage of soluble, or recoverable copper. In
defining the reserves it is also more meaningful to use recoverable than total copper. This gives
a better basis for economic analysis and engineering design.
5
4. TEST PROGRAMME OVERVIEW
There is no one ideal test programme, every deposit is different and the programme will depend
upon the availability of sample and time. However, adopting a standard approach will promote
the most effective use of resources and generate the most reliable results. The programme
outlined is intended to give the best results in the shortest time, and make efficient use of the
samples. Factors affecting the test programme include:
• Size and grade of the ore body.
• Lithology (number of potential ore types).
• Mineralogy (Copper and gangue).
• Feasibility and cost of obtaining sample (May need to compromise between statistically accurate samples size and practically obtainable sample size).
• Project schedule.
• Financing requirements.
• Previous data available.
4.1 Preliminary Characterisation
After the ore types have been designated, a composite sample of each should be prepared.
Fine RC chips are adequate at this stage, although drill core may be preferable, and is
necessary in later stages. Multiple element analysis should be done to identify any components
potentially harmful to bio leaching or SX-EW, such as arsenic, manganese, molybdenum,
fluoride, silver and chloride.
Mineralogical examinations are very helpful for planning the column tests and interpreting the
results (Baum 1996). They should identify copper and gangue minerals, grain size,
intergrowths, associations etc. Later comparison of column leached residues against head
samples can support the leaching data, indicating the actual reactions occurring and the nature
of the copper in tails.
Small-scale sulphuric acid bottle roll tests are used to indicate the leach response of oxide
minerals. Similar tests using an acidified ferric solution can estimate the bio-leachable
components, including the secondary sulphides. Acid consumption and copper recovery are
determined by these tests, although acid use is generally much higher than in column tests.
Further bottle rolls, or shake flask tests using an inoculum obtained from site, or other projects,
can give a better idea of the bio-leachable component, and warn of any adverse effects the ore
may have on the bacteria.
6
The leach conditions in a bottle roll test are very different from a commercial heap, so the
results have limited use. They can be considered as determinations of the maximum target
copper extraction and acid consumption. They are useful for giving quick estimates, and for
identifying the metallurgically significant rock types so that subsequent stages can be planned.
Cure tests can be performed to estimate how much acid should be added during agglomeration.
Acid consuming ores should be sufficiently cured so that the pH of the initial effluent is low
enough to continue leaching, but not so low as to waste acid. The tests involve curing ore
samples with various acid doses under conditions similar to an agglomerator. The range of
doses should be based on the maximum acid consumption determined by the bottle roll tests
(eg 25, 50, 75 & 100% of the maximum consumption).
4.2 Size Distribution
When drill core or bulk samples become available, it is possible to look at the crushing
characteristics and the resultant size distribution. The presence of clay or other fine material
may create problems for heap permeability and aeration. The distribution of soluble copper
across the range of particle sizes should also be understood.
4.3 Water Tests
A detailed site water analysis is an important aspect of the metallurgical assessment. The effect
of impurities on all stages of the process must be considered, particularly bio leaching and SX.
Where practical it is best to use site water for leach tests.
4.4 Column Tests - Programme Conditions
Heap leaching is a slow process, and column tests can take many months to complete. This
factor, along with scarcity of samples, tends to have a major effect on the planning of column
test programmes. The mineralogy is also a deciding factor, with different approaches taken for
sulphide and oxide minerals. Whether the minerals are finely disseminated or present in
fractures and veins, the presence of acid consuming gangue, the level of fine clay material, all
influence the test conditions selected. A typical programme consists of several stages, such as:
Phase I: Small columns (typically 2 m x 100 mm) using 12.5 mm or 25 mm crushed ore, to
estimate major leaching criteria.
Phase II: Small columns run under a variety of conditions, to optimise parameters such as
crush size and acid addition.
Phase III: Tall columns (typically 6 m x 150 mm) using the optimum leach method predicted
by phase II tests, to confirm leach response under realistic conditions.
7
4.5 Optimisation
The parameters to be optimised tend to depend on whether the ore is oxide or sulphide
dominated. With oxide ore, the acid addition method can be critical to the acid consumption
(Miller et al, 1997). With sulphide ore the emphasis is normally on creating favourable
conditions for bacterial activity.
• Crush size: Single particle leach kinetics are improved by crushing finer, but physical
considerations such as permeability and aeration are favoured by a coarser crush. The
optimum size will be a compromise between these factors and the cost of crushing. Typical
sizes to test are p80s of 12.5 mm, 25 mm, and 50 mm. (Tertiary, secondary and primary).
Factors affecting the results include the mineral assemblage (veins/fractures or fine
dissemination), the importance of aeration, porosity of the host rock, and the copper grade.
• Acid addition method: In cases of high acid consumption, minimising the acid used can be
critical to the project’s economics. The pH should be kept low enough to continue copper
leaching, but not so low as to cause avoidable gangue consumption. A suite of tests using a
range of target feed acid levels can be conducted to see which results in the lowest overall
consumption without loss of copper recovery. The amount added in pre-treatment can also
be optimised. In the case of sulphide ores the leach/SX/EW process can be nett acid
generating, in which case external additions of H2SO4 should be minimised.
• Irrigation rate: This is usually 6 L/m2/hr, but can be increased to around 10 L/m2/hr for high
acid consuming oxides. Provided that sulphide ore is fully wetted, increasing the rate will
probably serve only to dilute the pregnant solution. In the event of aeration or short circuiting
problems, an intermittent application system may improve the leaching rate.
• Aeration: Simulating the air flow patterns in a leach heap is not easy to do in the laboratory.
To test the importance of aeration to a particular rock type there is an option of sealing the
column discharge from the atmosphere, or allowing air ingress at the base (Readett and
Miller 1997). Air sparging the feed solution helps ensure the bacterial activity in a bio-leach.
Forced aeration of sulphide ore is another option to consider.
• Grade effect on recovery: Throughout the life of a mine the grade of ore supplied to a heap
will probably vary, so it may be advisable to test the effect of different grades on leaching.
• Ore height: A trade-off exists between leach cycle time, which is reduced by lower heights,
and the cost of stacking and irrigation, which is higher for lower lifts. The choice is
particularly critical for sulphide ores, where poor aeration conditions at depth may control the
recovery. Column tests should ideally be done at the anticipated heap height, but shorter
columns can give adequate results if the chemical conditions, especially effluent pH, are
similar to the commercial operation (Dreier, 1995). A typical height is 6 m.
8
4.6 Physical Tests
The following physical characteristics should be determined from column leach tests:
• Size distribution
• Slump = (Initial height - final height)/initial height x 100%.
• Agglomeration moisture requirement.
• Total moisture under active leach.
• Drained moisture.
• Drainage rate.
• Maximum percolation rate.
Additional testing should be undertaken to determine the following ore properties:
• Specific gravity.
• Bulk density.
• Moisture content.
• Angle of repose.
• Abrasion index.
• Compressive strength.
• Crushing work index.
If the ore contains considerable fine clayey material, permeability could be a major issue. In this
case additional percolation tests are called for.
9
5. COLUMN TEST DESIGN
Column tests should be designed to simulate the conditions in a commercial leach heap as
closely as possible. This is not as easy as it sounds; particular attention should be paid to the
following:
• Water source.
• Initial solution (mature raffinate).
• Bacterial inoculum.
• Solution recycling.
• Temperature.
• Loading method.
• Acid and copper levels in column feed solutions.
• Aeration.
• Agglomeration.
• Sampling.
Open circuit tests have a limited application in preliminary tests on oxide ores. In all other cases
it is recommended to use a closed circuit with solvent extraction of copper from PLS and recycle
of raffinate. Closed circuit tests have advantages including:
• Concentration of ionic species builds up to equilibrium levels, giving more realistic estimates
of acid consumption and copper extraction.
• Bacterial populations can adapt to the conditions of the particular ore.
• Copper cations in the PLS are replaced by H+, giving a feed raffinate which replicates an SX
plant.
The disadvantages are that they are more time consuming, require extra equipment, and in the
early stages can produce highly acidic raffinates from the high grade PLS solutions.
5.1 Head and Tail Assay
To provide ongoing recovery estimates and give confidence in metallurgical accountability, an
accurate head assay should be determined for each column test. The recommended way to do
this is by size fraction (Keane, 1995). Head assays are done at each size fraction, along with a
composite assay to compare against the weighted average of the fraction assays. Ideally the
10
head assay sample should be equal in weight to the column charge, but if the quantity of
sample is limited a reasonable compromise can be made.
After termination of the leach, the entire residue should be weighed, dried, screened, weighed
again, crushed and prepared for assay. As with the head sample, assaying by size fraction is
recommended for accurate results, and to enable copper recovery to be calculated at each size
fraction.
Wet screen analysis on the fine material is recommended for both head and tail samples.
This is a labour intensive procedure, but has numerous advantages:
• The size distribution of the sample can be examined to ensure it reasonably approximates a plant situation.
• A high level of metallurgical accountability improves confidence in the results.
• A reliable head assay enables recovery to be followed as the test proceeds, and interpretation of data can be done long before the final residue assays are available.
• Information regarding copper recovery by particle size is essential in optimising the crush size.
• The different sizing techniques (wet/dry) provide an insight into the ultra fine particle characteristics.
5.2 Inoculation and Agglomeration
Laboratory scale agglomeration can be conveniently done in a rotating drum such as a plastic
cement mixer. The agglomeration time should be similar to that of a commercial operation,
which is typically about 1 minute. Sufficient moisture should be added to dampen all of the ore,
but not leave a shiny surface. The amount of water required depends mainly on the fines
content, and should be determined by trial and error before hand. The acid cure, polymer and
inoculum should also be added during agglomeration if required.
Biological leach tests should preferable use solutions and bacteria from the specific site.
Alternative sources are similar operations, old workings and previous tests. An artificial raffinate
can be used, provided the full suite of chemical species and a sufficient bacterial population is
present. Leaching tests should always be conducted under the chemical and bacterial
conditions anticipated for the industrial scale operation, or the results will be misleading.
In the presence of swelling clays the physical benefits of agglomeration are particularly
important. Enough liquid must be added at this stage to satisfy the clays’ water of hydration and
prevent subsequent blinding of the ore.
11
5.3 Solution Chemistry
In a commercial operation, freshly stacked ore will usually be a small proportion of the total ore
under leach. The rich solutions will be mixed with weaker solutions, thus stabilising both the
PLS copper grade and the raffinate acid concentration. This is not the case with column tests,
as a high PLS copper grade will produce a strongly acidic raffinate. There is therefore a need to
remove some acid from the solution inventory to avoid exaggerating the leach results.
In practice this means splitting the leach feed solution after the first SX cycle, and diluting it
down to a reasonable acid strength. The remaining strong solution can then be gradually added
back to the system. This has the disadvantage of diluting the other ions as well, unless dilution
is with the un-acidified initial solution. An alternative and preferred method is to partially
neutralise the solution to a target acid strength, using sodium hydroxide. The advantage of this
is that it avoids dilution of all the other ionic species, which may affect the leach. At all stages
the solutions must be carefully accounted for, to get an accurate acid consumption figure.
5.4 Solution Sampling
In closed circuit column tests, the sample volume should be kept to a minimum or it will interfere
with the leach accounting, and can complicate the acid consumption calculation. In the initial
stages daily sampling of PLS is recommended, but once conditions stabilise it is only necessary
to sample every few days. For convenience this can coincide with the batch SX cycle. Samples
(both feed and discharge) should be analysed for copper, free acid, iron (ferrous and ferric), pH
and Eh.
Late in the leach cycle a PLS sample should be subject to the same detailed characterisation
(ICP analysis) as the initial leach feed. A bacteria count may be advisable at this stage.
5.5 Temperature
In a commercial operation it is generally not possible to alter the temperature in the heap.
Laboratory tests should therefore be run as close as possible to the temperatures expected on
site. This temperature will be a function of the climatic conditions, and of the chemical reactions
occurring. Higher temperatures will exaggerate the leach kinetics, or may kill the bacteria in a
bio-leaching system.
5.6 Polymer Agglomeration
Ores with a high fines content create a challenge in terms of heap permeability. To deal with
this situation it is possible to add a polyacrylamide flocculant to the agglomerating solution. The
polymer should be compatible with the chemical conditions in the leach solutions, and should be
well mixed before addition.
12
5.7 Aeration
Aeration within a full sized heap is particularly difficult to simulate in the laboratory, as oxygen
supply can often be the rate controlling factor. This is an inherent problem with column tests. A
leach heap is open to the atmosphere on all four sizes, and effectively sealed at the base. A
column test, however, is usually closed at the sides and open at the base. Depending on the
aims of a test, it may be desirable to either restrict or enhance the aeration. If excessive
aeration in a column could give exaggerated results, the base can be sealed from the
atmosphere by simply looping the discharge tube to create an air trap. To test the effect of
forced aeration, a low pressure air line can be connected near the base of the column.
In practice, leach feed solutions get well aerated while passing through the ponds, plant and
irrigation sprays. To replicate this in the laboratory, solutions should be left open to the
atmosphere, and possibly air sparged.
13
6. PILOT LEACH HEAPS
The operation of pilot scale heap leach - SX - EW plants has been undertaken at several sites.
It is generally difficult to justify the expense of such a plant in terms of the metallurgical
information generated. The data can serve to confuse, rather than clarify the leaching
characteristics because of the inherent difficulties. The difficulties include getting an accurate
head sample, tolerance of metallurgical accounting and inflexibility regarding leach parameters.
A pilot plant can, however, be a very useful tool for staff training and transfer of knowledge in
areas unfamiliar with the technology.
14
7. CALCULATION AND PRESENTATION OF RESULTS
7.1 Calculations.
A single column test can generate a very large volume of data. To facilitate interpretation of this
data, a standard reporting format is recommended. Following this approach will ensure that the
correct measurements are taken, and that the data can be readily understood and processed by
others. Particular attention should be paid to the following points:
• Final copper recovery and extraction curve based on calculated head, not assay head.
• Acid consumption as nett and gross.
• Daily measurement of volume fed and drained from the column.
In addition to leaching results, reports should contain head and tail size and copper distribution,
a summary of physical characteristics, and a log of batch SX and solution make-up. Producing
a summary sheet for each column is also recommended.
Principle calculations are done as follows:
• Copper recovery % = 100 x (Cumulative Cu in PLS - Cumulative Cu in Feed)/ Cu in
Head.
• Copper recovery % = 100 x (Cu in head - Cu in tail)/Cu in head.
Cu in PLS is the summation of the product of daily PLS volume (L) and assay (g/L). Cu in feed
is done the same way. Cu in head can be either assay head or calculated head. During the
course of a test the assay head is used to provide ongoing recovery estimates. At the end of
the test the calculated head is determined by adding the residual copper (tail) to the total
solution copper (PLS- feed).
The metallurgical accountability can be determined as calculated head / assay head. If the
head sample was properly split, and all assays were done accurately, the accountability should
be close to 100%.
There are two ways to calculate acid consumption. The easiest and most accurate way is to do
a mass balance over the closed cycle system:
• Acid consumption = Acid added to system - Acid removed from system - Acid inventory.
Acid added includes agglomeration and curing. Acid removed includes samples, and any
solution splitting and neutralisation. All these terms should include the acid equivalent of copper
(=1.54 x Cu grade). During the test, acid inventory can be estimated from the approximate
system volume and concentration. After the test, it is measured accurately once the final
drained solution has been extracted.
15
The second method is a mass balance on the column only, and includes a credit for the acid
recovered by SX.
• Acid consumption = Acid in column feed - Acid in column discharge - 1.54 x Cu
extracted.
These two methods are for nett consumption and they should, in theory, give the same result.
The first method relies on weighed quantities of sulphuric acid, while the second uses the free
acid titration results. Results from the second method can be normalised at the end of a test to
conform with the first method. Acid consumption is usually expressed in kilograms per tonne of
ore, and can also be expressed as kilograms per kilogram of copper extracted.
7.2 Graphs
The following plots can enhance understanding of the leach response:
• Copper recovery over time (% of total and recoverable copper).
• Copper recovery over leachate flux (kL feed per tonne of ore) (% of total and recoverable copper) – Figure 1.
• Acid consumption over time (nett & gross) – Figure 2.