University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Colorado School of Mines
|
The collector cost was estimated from work done by Dylan Everly. [17] The cost of each
reagent is shown in Table 8.4, along with the required amount per tonne of ore processed based
on the reagents required for the best 10 kg flotation test. Using the estimated prices from Table
9.4 the total reagent cost per year for the flotation circuit is $52,100,000.00 and for the flotation
and gravity circuit it is $45,300,000.00.
Table 9.4: Reagent costs for both economic analyses.
Reagent kg/tonne ore Cost per kg
Collector 3.45 $10.00
Soda Ash 5.14 $1.10
MIBC (frother) 0.01 $2.00
Flotation Only Flotation and Gravity
Hydrochloric Acid (pure) 11.15 6.69 $1.74
The labor costs were estimated from CostMine 2017. It was assumed that for both
processing circuits a mill manager, metallurgist and mechanics were needed along with two
persons in the control room. In the flotation circuit, it was assumed that only six laborers were
needed while in the flotation and gravity circuit nine were needed. The hourly wages and total
cost for each position are shown in Table 9.5.
Table 9.5: Labor cost breakdown for both processing circuits.
Hourly Total Cost Total Cost (Flotation
Flotation Gravity Wages (Flotation) & Gravity)
Mill Manager 1 1 $43.94 $91,200.00 $91,200.00
Metallurgist 1 1 $39.54 $81,300.00 $81,300.00
Mechanic 1 1 $28.01 $58,400.00 $58,400.00
Laborers 6 9 $20.87 $260,300.00 $390,500.00
Control Room
Operator 2 2 $20.87 $86,800.00 $86,800.00
Total $578,000.00 $708,100.00
80
|
Colorado School of Mines
|
9.5 Cash Flow Sheet and Analysis
From these estimated costs for each circuit a cash flow sheet was developed for a 10 year
analysis. The flotation cash flow sheet is shown in Table 9.8 and the flotation and gravity flow
sheet is shown in Table 9.9.
The net present value (NPV), internal rate of return (IRR) and payback period for both
conditions were calculated using the information in Tables 9.8 and 9.9. From these the best
circuit can be determined. These values are shown in Table 9.10, assuming a discount rate of
12%.
Table 9.8: NPV, IRR and payback period for each project.
Flotation Flotation & Gravity
NPV (i*=12%) $2,392,100,000.00 $2,156,100,000.00
IRR 1700% 1500%
Payback Period (years) 0.058 0.069
From these values it can be inferred that the flotation only circuit is more profitable, since
the NPV and IRR are greater than that of the flotation and gravity circuit and the payback period
is less. For a proper conclusion to be made an incremental NPV analysis was done to determine
the best circuit. The incremental analysis was done by subtracting the net cash flow of the
flotation circuit from that of the gravity circuit and calculating an NPV from that. It was found
that the incremental NPV was -$325,000,000.00, meaning that the flotation circuit is the more
profitable choice in the long run. For both projects to be equally profitable the incremental NPV
will be $0.00. Since the price of hydrochloric acid is the driving economic force for this project,
it was varied until the incremental NPV was $0.00. It was found that the price of pure
hydrochloric acid would need to rise to $67.73 per kilogram for both projects to be equally
economical.
82
|
Colorado School of Mines
|
CHAPTER 10: CONCLUSIONS AND FUTURE WORK
This rare earth separations study was done to find a process that incorporated flotation and
gravity separation in an economical way to upgrade the rare earth content of an ore. The
economic driving force for this project was the additional cost of hydrochloric acid consumption
incurred by the presence of calcite.
Flotation fundamentals were studied on a gravity concentrate and a run of mine ore sample.
The fundamentals examined included zeta potential and adsorption. The PZC for each of these
samples were 4.83 and 4.23, for the run of mine ore sample and the gravity concentrate,
respectively. The adsorption density for each collector was measured for each sample. For the
gravity concentrate multilayer adsorption seemed to occur more easily for collectors 2 and 5. It
was found that for collectors 2 and 5 the adsorption onto the run of mine ore sample was less
favorable than adsorption onto the gravity concentrate. It was also determined that the adsorption
of collector 2 was the least favorable of any of the collectors.
Microflotation experiments were conducted using the data from the flotation fundamentals. It
was found that for the gravity concentrated ore the recovery of calcite was decreased, but the
grade of the rare earth bearing minerals was also decreased in some cases. From these
experiments it was determined that further study would focus on flotation followed by gravity
separation.
Bench scale flotation tests were conducted in an effort to find flotation conditions that were
more favorable with the addition of more reagents. Collector 2 proved to have the best results for
flotation with a rare earth oxide grade of 42% and a recovery of 70% while rejecting 90% of the
calcite.
86
|
Colorado School of Mines
|
Of the large scale test work only one test proved to be promising. The same conditions that
produced the best test on the bench scale also produced the best test on the large scale. After a
two minute flotation the rare earth oxide grade was 44.4% and the recovery was 81% while
rejecting 91% of the calcite.
The concentrate from the best large scale flotation test was used for gravity separation on the
ultrafine falcon concentrator. It was found that the falcon could reject another 40% of the calcite
while still maintaining a rare earth oxide stage recovery of 90%.
An economic study was done to determine which process was more economical. Capital and
operating expenses were estimates along with the price of the rare earth products. An
incremental NPV analysis was conducted to compare the flotation circuit to the flotation and
gravity circuit. It was found that the flotation only circuit was more economical and that for the
gravity circuit to be as economical as the flotation circuit the price of pure hydrochloric acid
would need to increase from $1.74/kg to $67.73/kg.
Based on the economic analysis gravity separation using the ultrafine falcon concentrator
after flotation is not as economical as using only flotation. Even though it was able to reject
additional calcite, it could not overcome the additional operating expenses incurred by adding a
gravity circuit.
Recommendations for future work include a study looking into the mechanism by which
collector 2 works. Also, since collector 2 is commercially available it would need to be
determined if it could be produced on a larger scale. Further study can be done on the
optimization of the flotation conditions for each of the collectors that were examined. It could be
worth looking into why the other flotation reagents, besides collectors 2, did not work for large
87
|
Colorado School of Mines
|
ABSTRACT
Mining Negotiation Santa Maria de la Paz, S.A. of C.V. (NEMISA) has operated the
Cobriza and Dolores copper mines in Villa de la Paz, San Luis Potosi, Mexico for more than 150
years. The copper mineral values occur primarily as chalcopyrite and bornite along with a minor
secondary sulfide mineralization. Also, there are ores that contain significant gold and silver
values, these ores are processed through the beneficiation plant with a rate of 8,500 tonnes per
day using a flotation concentrator. Gold and silver values not recovered in the copper flotation
concentrate report to the tailing storage facilities (TSF).
Commonly, flotation tailings contain 0.5-0.7 g/tonne of gold and about 15 g/tonne of
silver respectively. However, different methods have been investigated for recovering the values
of gold and silver currently being lost to the TSF. In addition, NEMISA has performed
metallurgical evaluations for improving the flotation technologies and implementing cyanidation
of the flotation tailings. NEMISA is evaluating the feasibility of installing a cyanidation circuit
to recover these values of gold and silver. However, there are not metallurgical studies for
futures ores to be mined, thus this thesis will explain the practices, processes and metallurgical
results for future ores.
iii
|
Colorado School of Mines
|
Table A.3: Feed particle size analysis for BWI calculation for the Cobriza sample 3.….………86
Table A.4: Particle size analysis F80 average calculation for the Cobriza sample……...………86
Table A.5: Product particle size analysis P80 for the Cobriza sample…………………………..87
Table A.6: Parameters used for the calculation of the BWI for the Cobriza sample……...….….88
Table A.7: Feed particle size analysis for BWI calculation for the Santa Teresa sample……….89
Table A.8: Feed particle size analysis for BWI calculation for the Santa Teresa sample 1…......90
Table A.9: Feed particle size analysis for BWI calculation for the Santa Teresa sample 2……..90
Table A.9: Feed particle size analysis for BWI calculation for the Santa Teresa sample 3…......91
Table A.10: Particle size analysis F80 average calculation for the Cobriza sample………...…..92
Table A.11: Product particle size analysis P80 for the Santa Teresa sample……………………92
Table A.12: Parameters used for the calculation of the BWI for the Santa Teresa sample…...…93
Table B. 1: Rougher stage copper flotation retention study for the Cobriza sample……...….…94
Table B. 2: Rougher stage silver flotation retention study for the Cobriza sample………….....95
Table B. 3: Rougher stage gold flotation retention study for the Cobriza sample……….……..96
Table B. 4: Scavenger stage copper flotation retention study for the Cobriza sample…………97
Table B. 5: Cleaner stage copper flotation retention study for the Cobriza sample……….…..99
Table B. 6: Rougher stage copper flotation retention study for the Santa Teresa sample……..101
Table B. 7: Rougher stage silver flotation retention study for the Santa Teresa sample………102
Table B. 8: Rougher stage gold flotation retention study for the Santa Teresa sample…....….103
Table B. 9: Scavenger stage copper flotation retention study for the Santa Teresa sample......104
Table B.10: Cleaner stage copper flotation retention study for the Santa Teresa sample…….105
Table B.11: Locked cycle flotation copper mass balance for the Santa Teresa sample…….….108
Table B.12: Locked cycle flotation silver mass balance for the Santa Teresa sample………....109
Table B.13: Locked cycle flotation gold mass balance for the Santa Teresa sample……….….110
Table B.14: Locked cycle flotation copper mass balance for the Cobriza sample……………..111
Table B.15: Locked cycle flotation silver mass balance for the Cobriza sample………………112
Table B.16: Locked cycle flotation gold mass balance for the Cobriza sample……....….…….113
Table C.1: Economic Analysis for the leaching plant project…………………….…..……..…114
xii
|
Colorado School of Mines
|
CHAPTER 1: INTRODUCTION
Negociación Minera Santa María de la Paz y Anexas is a mining company, located 200
km to the north of the capital city of San Luis Potosi, it operates one of the main underground
mines in Mexico, with a daily production of 8.500 tonnes per day of copper concentrate at its
beneficiation plant. The copper ore also contains gold and silver values, but only a portion of
these values are recovered into the copper concentrate. The remaining gold and silver values are
discharged to the tailings storage facilities (TSF). Metallurgical assessment on the current
flotation tailings has shown that 70-75% of the gold and 15% of the silver contained in the TSF
can be recovered by conventional cyanidation. NEMISA is interested in studying the feasibility
of implementing a cyanide leaching circuit in order to recover the gold and silver values being
lost to the TSF. Hence, this thesis will show the results of research conducted to investigate
metallurgical processes fore recovering gold and silver values from copper flotation tailings
produced from for ore samples from the Santa Teresa and Cobriza mines, which are the projected
to be the source of future ore to be processed at NEMISA’s copper flotation concentrator.
During this research, mineral processing studies were performed on each composite from the
Santa Teresa and Cobriza mines to obtain flotation tailings to be used for cyanide leaching
studies. These studies included grindability tests, flotation studies, and gold and silver cyanide
leaching as final unit operation to assess the gold recovery from tailings generated during the
flotation studies.
1
|
Colorado School of Mines
|
CHAPTER 2: LITERATURE REVIEW
2.1 Gold Processing
Gold has been a precious metal of interest ever since the Egyptians started to produce
gold under the eleventh dynasty (2133-1991 BC). They were first to quarry gold-bearing rocks.
Gold has been critical for the early development of metallurgy and was typically found in nature
in the native state where there was not need of any chemical or metallurgical knowledge.
However, during the last few centuries different technologies have been developed such as
cyanidation which originally was based on dissolving gold in weak cyanide solutions and then
using zinc dust to precipitate gold from solution (M.D. 2005). The main stages for processing
gold are leaching, solution purification, concentration and recovery of gold. Today, gold is found
associated with complex minerals which has necessitated that more complex hydrometallurgical
be developed to efficiently recover these gold values from these complex ore gangue minerals.
2.2 Gold Pricing and Mine Production
Gold has been one of the main metals of interest in the global economy for many years
due to its scarcity and physical properties, which are required in different industries. Gold has
played a main role in the evolution and development of metallurgy for thousands of years.
Currently, according to the U.S.G.S. the U.S. gold resources are estimated at 33,000 tonnes (U.S.
Geological Survey 2019). The price of gold has been decreasing these last ten years where from
2016 there has been a recovery on the gold price until the year 2020 as shown on Figure 2.1.
Figure 2.1: Gold Price in US dollars per ounce between 2011-2020 (WGC 2020)
2
|
Colorado School of Mines
|
*Units in tonnes (1,000 kg= 32,150.7 troy ounces)
e Estimated, Net Exporter
8,9 Data obtained from the Mineral Commodities Summaries
Figure 2.2: World Mine Production and Reserves (tonnes) (U.S. Geological Survey Report 2019)
The countries with more reserves are Australia that has 9,800 tonnes of gold and Russia
with 5,300 tonnes of gold respectively, as shown in Figure 2.2. The total world gold reserves are
54,000 tonnes; also there is an increase in the mine production from 2017 to 2018 caused by the
gold consumption by the jewelry industry mainly, otherwise, the gold coins and bars decreased
slightly compared to the beginning of 2017. (U.S. Geological Survey 2019).
2.3 Recovering Gold from Tailings
Inefficiencies in some flotation circuits can result in significant losses of gold values to
the tailings. Generally, there are significant amounts of gold that can be recovered from these
tailings in a feasible way. However it is important to perform a metallurgical assessment of each
tailing site to establish the required process parameters and evaluate the feasibility in terms of
metal recovery, gold price, tailings grade and capital costs (Zarate 1987). Hydrometallurgical
processes have been considered as an option for extracting, purifying and recovering the metals
being contained in flotation tailings. The two critical variables to consider in hydrometallurgical
processes are the thermodynamic properties in the system that determine the driving force of the
reaction. The reaction kinetics is also a key concept for the design of a hydrometallurgical
process as it involves physical, chemical and mass transport aspects. In many cases reprocessing
of old tailings can serve to improve an existing environmental impact due to contaminated
ground water, surface water, and unconfined dust emissions.
3
|
Colorado School of Mines
|
As part of a tailings reprocessing project, reprocessed tailings can be placed into a
properly designed and lined TSF to control water contamination. In addition the TSF can be
reclaimed through the use of appropriate revegetation techniques to minimize the impact of
erosion and air-borne dust contamination as shown in Figure 2.3. In some instances the
desulphurization of tailings is appropriate by inclusion of a stage of bulk sulfide flotation to
remove sulfide minerals prior to discharging to the TSF and thereby reduce the potential for acid
rock drainage (Markovic Z 2010).
Figure 2.3: Dry Tailings Storage Facility at NEMISA
2.4 Mineral Processing Studies
The analysis of physical characteristics of an ore can be determined by various
metallurgical processes. For example, raw materials are crushed and ground during processing in
order to liberate the valuable minerals. Once liberation is achieved, the separation of non-
valuable and valuable minerals can be performed base on differences in the physical and
chemical properties. These metallurgical studies are the key element of ore engineering studies;
also they help to assess ore recoverability and designing the best economical option for a
flowsheet.
2.4.1 Sampling
Sampling for metallurgical studies during plant design is critical in order to obtain
representative samples for determination of physical and chemical characteristics. Accuracy is
defined by the selection of the sampling systems, components, preparation, and analysis without
introducing systematic errors (Society for Mining 2002). There are different types of sampling
4
|
Colorado School of Mines
|
for collection of material for future analysis. For example core drilling, reverse circulation
drilling, channel sampling, trench sampling, grab sampling, geochemical and environmental
samples and water samples.
For this research the samples were obtained as drill core intervals from Cobriza and Santa
Teresa which were composited to create the test composites used for the metallurgical studies
presented in this thesis. The geological department at NEMISA took these samples from the
locations within the mine that are planned to be mined in the future. This stage of the research
program is critical because it determines the success of obtaining a sample that really
“represents” each composite. There are different sampling methods used for most metallurgical
or environmental operations. The consideration of probability theory and classical statistics
allows to understand the sources of error when performing a specific sampling method. The
addition of variance depends on the different components in the system, in other words, the total
variance is the sum of the individual variances existing in the system (Wills 2016).
Choosing the most convenient sampling method depends on the nature of the sample and
the future purpose of it. There are characteristics to consider during sampling for example if the
material changes with time or exposure to different environmental conditions, also the material
must be fresh prior to performing any sampling procedure (Taggart 1945).
2.4.1.1 Coning and Quartering
Splitting large lots of materials requires a method like coning and quartering which can
be performed conveniently by using shovels and a tarp. This procedure consists on mixing and
shoveling the material into a uniform conical pile, where the natural segregation in the cone is
radially symmetrical. The cone is after spread uniformly from the center to form a flat disk of
material which is divided into quarters by using boards. The sample is obtained by removing one
pair of opposite quarters and the other pair is used as the sample as or further sub-sampled by
another stage of coning and quartering (Taggart 1945).
2.4.1.2 Grab Sampling
This method involves collecting large or small amounts of sample by using spatulas or
shovels where the material is divided into many samples as many as desired. The material is
mixed and homogenized using a rolling mat which must be flexible. The material is then divided
randomly into samples by grabbing small amounts from the homogenized pile on the cloth. This
5
|
Colorado School of Mines
|
method is known by using the least equipment, on the other hand there are more possibility of
having higher variances in comparison with other methods (Taggart 1945).
2.4.1.3 Chute- Type Riffler Splitter
This method is based on the use of a sample splitter where the material is distributed in
the receptacle and then emptied over by opening the chute. The material passes through the
alternately arranged passages in opposite directions into two collecting containers under the
dividing head outlets (Retsch 2019).
2.4.1.4 Rotatory Sample Splitter
This method is mainly used to obtain representative samples from heterogeneous granular
and powdered materials by using a special device which is a rotatory sample splitter or also
known as spinning splitter. This method minimizes the negative impact and differences in
particles size, shape, specific gravity, and average quality. The rotatory sample has a set of stain
steel containers that rotate at different speeds according to the splitting process requirements.
During this research the chosen sampling method was coning and quartering because the
amount of material was too large to use another alternative method. The samples to split were
approximately 500 kg of crushed ore that resulted from each composite.
2.4.2 Grindability Studies
Grindability testing refers to measure the resistance of ore samples to breakage or
hardness according to their properties. The grindability testing depends on the sampling
requirement and the cost of equipment that is needed to perform these tests. There are different
small-scale methods for performing grindability studies (McKen 2005).
2.4.2.1 Bond Work Index Procedure
The Bond Work Index test calculates how much energy is required to get from an initial
particle size (F ) to a finer required particle size (P ). The determination of the energy
80 80
consumption during grinding was researched by Fred C. Bond who noted that during a grinding
open, closed or complex ore composed of different minerals cycle there can be variations in the
prediction required for grinding. The Bond Work Index procedure is a simulation of dry grinding
in a closed circuit in the Bon ball mill to get 250% circulating load. The work index is a
parameter that represents the resistance of an ore to grinding which represent the energy
(kWh/sht) required to reduce the material of one short ton from a determined particle size (Dejan
T 2017).
6
|
Colorado School of Mines
|
This test is performed using a special mill which is designed to assess the parameters to
calculate the BWI such as revolution per minutes, number of balls, the total specific area of the
grinding charge and size of grinding balls. For this project a SEPOR FC Bond Mill was used
using steel balls with a total weight of 44.5 lb according to the BWI standard procedure (SEPOR,
Operating & Maintenance Manual 2019).
The test is typically performed by having 10 kg of raw material which must be 100%
passing 3.5 mm mesh. The density of the raw material is calculated with 700 cm3 and then it is
used with an arbitrary number of revolutions.
During the procedure after each grinding cycle the material is screened for performing a
size analysis, the mill revolutions are determined using the data of the previous cycles for
obtaining 250% circulating load. This procedure is repeated until the sieve undersize produced
per mill revolution is constant in the last three cycles. Commonly the BWI procedures required 6
to 10 cycles depending on the material, the following equation is used for determined the Bond
Work Index once steady state is achieved (Yap, Sepulveda and Jauregui 1980).
(2.1)
Wi= Bond work index (kWh/t)
Pc= Test sieve mesh size
G= Weight of the test sieve fresh undersize per mill revolution (g/rev)
F80= Sieve mesh size passing 80% of the feed before grinding (um)
P80= Opening of the sieve size passing 80% of the last cycle test sieve undersize product (um)
The use of a laboratory scale Bond mill applies the first order kinetics where there is a
comparison between the parameter P of sieve undersize. The data from every grinding cycle is
80
used to calculate the parameters G and P which are used for calculating Wi as shown on the
80
Bond formula. The accuracy of this method can be evaluated by using data of current
experiments that used standard Bond procedures for the parameter G value using different types
of materials After the BWI is determined for a specific type of ore, it can be used for estimating
the energy required for grinding in a mineral processing circuit. An economical evaluation can
be done by calculating capital (CAPEX) and operational (OPEX) costs. These two economical
7
|
Colorado School of Mines
|
models can help to get more accurate factors (net present value, paybacks, rate of returns) during
performing feasibility studies (Yap, Sepulveda and Jauregui 1980).
2.4.2.2 Bond Low-Energy Impact Test
This procedure measures the hardness of ore samples for crusing. This method requires a
special device that consists of two pendulum hammers mounted on two bicycle wheels that strike
equally and simultaneously on the opposite sides of the ore sample, the height of the pendulum is
lifted until the tested rock breaks measuring the required energy. For this procedure the crusher
work index or impact work index is calculated using the following formula (SGS 2011):
CWI= (2.2)
𝐽
53.29 𝑋 (𝑚𝑚)
J= Energy at which the specimens broke
𝑆.𝐺.
S.G.= Specific gravity of the ore
mm= Thickness of the rock specimen
This test is typically performed on 20 rocks in order to measure the natural dispersion in
the sample, also it allows to use coarse size rocks (2” to 3”).
2.4.2.2.1 SAG Power Index (SPI) Test
This grindability test was developed to model AG/SAG mills as an alternative for
characterizing ore bodies. This method uses a small laboratory mill (30.5 cm diameter) to grind a
2 kg sample from a initial particle size to a product size. The mill is charged with 5 kg of steel
balls, the 2 kg of sample are crushed to 100% minus 1.9 cm 80% minus 1.3 cm and placed in the
mill as well. The required time to achieve 80% minus 10-mesh is called SAG Power index or
SPI. The procedure consists in running the mill with several screening iterations until the sample
is reduced to the particle size 80% minus 10-mesh. During the test there is a parameter called
the P which is the 80% passing size of the material that is finer than 10-mesh at the end of the
64
test. The advantage of this method is the small amount of required sample (2 kg) which can be
easily collected from drill core (Amelunxen 2003).
2.4.2.3 JK Drop-Weight Test
This grindability test is an industry standard for use in characterizing ore under AG/SAG
milling conditions. The minimum amount of required sample to provide enough particles for
testing breakage properties is 100 kg in the size range of -63+ 13.2 mm. The test generates the
appearance function of the ore versus the units of impact. The appearance function can be used
8
|
Colorado School of Mines
|
with the software JKSimMet modeling to predict the ore breakage behavior in comminution
processes (SGS 2011). The device is used for performing this test which consists in a system for
dropping a variable weight and size rock samples from different heights. Also, this procedure the
abrasion of the sample can be determined by using a tumbling test (JKTech 2019).
2.4.2.4 SAG Mill Comminution (SMC) Test
The SMC test is considered as a less expensive version of the JK Drop weight test, the
difference is that it is performed using smaller rock samples (+19/-22 mm) or drill cores. This
test produces a drop-weight index (DWI) in kWh/t and two different parameters A and B. These
parameters care used for modeling purposes using a software such as JKSimMet. The test
required 30 to 40 kg of ore which is a smaller amount in comparison with the KJ Drop Weight
test or JK Rotatory Breakage Test (JKTech 2019).
Drill cores are cut in quarters using a diamond saw and then they are tested similarly to
the standard Drop Weight Test just with the difference that a single size fraction is tested. The
advantage of the SMC test is that it generates the energy versus breakage using a small amount
of sample of a single size fraction. After the test is performed the products are collected and
sized to measure the particle size as part of the energy requirements calculations (SGS 2011).
2.4.2.5 MacPherson Autogenous Grindability Test
The design of power- efficient grinding circuits can be conducting by this test which
determines the MacPherson Correlated Autogenous Work Index (AWI) that can be used in
conjunction with the Ball Mill Work Indices to calculate power requirements for a grinding
process. This method helps to configurate circuits for autogenous grinding (AG) and semi-
autogenous (SAG) circuits. The test is performed continuously for a minimum of sic hours and
until the steady state is reached. Once the test is completed the products are taken to a particle
size analysis, and the charge f the mill is analyzed as well for specific gravity determination.
The the mill power draw, throughput and product size analysis are used to compute the
MacPherson autogenous work index (AWI) (SGS 2011).
2.4.2.6 High Pressure Grinding Roll (HPGR)
This grinding technology has been used for many years and it is considered as an energy-
efficient option for conventional and AG/SAG comminution circuits (SGS 2011). A HPGR
machine consists of a pair of rolls which are mounted in a sturdy frame, the rolls are separated by
a gap and constantly counter rotating. One of the rolls is fixed and the other one can be adjusted
9
|
Colorado School of Mines
|
to produce a gap which depends on the material properties. This gap opening between rolls
determines the grinding pressure and a hydro-pneumatic spring system applies the grinding force
on the rolls (S Rashidi 2017). The HPGR is fed from the top and the material passes through the
gap opening and leaves the machine from the bottom.
The locked-cycle scale test using a scale test HPGR, such as a Bond ball mill grindability
is proposed to be a cheap option to determine grindability requirements for a specific material.
One of the main characteristics of the HPGR is that it has the capability to produce a particle size
distribution with a greater number of fines. The total power for the HPGR system can be
compared to the one required using a conventional circuit formed by a rod and ball mill work
indices considering the Third Theory of comminution (SGS 2011).
The literature survey about the different methods and alternatives to determine the
grinding power requirements for the NEMISA samples indicates that different methods can be
applied, but Bond Mill is the more convenient option for this case. The grinding circuit at
NEMISA consists of wet grinding ball mills which can be simulated with a small-scale Bond
Mill. The amount of material is enough quantity of material (~500kg) that allows to perform the
Bond Mill procedure for each of the two composites. The purpose of simulating according to the
beneficiation plant circuit is to determine results that reflect the NEMISA operation for having a
reliable economic analysis.
2.4.2.7 Primary Grind Size Study
This study helped to analyze the effect of the primary grind size on the recovery of
valuable minerals. During these studies, the particle size determined the quality of grinding, and
stablished the degree of liberation of the minerals. As mentioned before, the particle size analysis
is obtained by passing a known weight of sample material through successively finer sieves and
weigh the amount collected in each fraction. The process of sieving is divided into two stages.
The first one is in the elimination of particles smaller than the screen apertures and the second
one is the separation of the “near-size” particles. The product size is usually quoted in terms of
one point on the cumulative curve which is often the 80% passing size of the product, P .
80
Likewise, the 80% of the passing size of the feed is defined as F . Particle size distributions help
80
to control the material particle size in the case of particle size variations out of specification and
modifications are needed to fix the problem in the grinding process. The two most common
10
|
Colorado School of Mines
|
methods that are used for comminution studies for non-uniform size distributions are Gates-
Gaudin-Schuhmann and the Rosin-Rammler methods (Wills 2016).
For grinding laboratory testing, the grind size P values are commonly between 75 to
80
250 um used for testing but also there are finer and coarser grinds as well. Size-by-size analysis
of the flotation feed, concentrate, and tailings at a certain grind size is performed to provide a
recovery profile and to confirm a grind variability study. The size-by-size analysis is used a
check on the metallurgical balance, a more detailed analysis test is performed to evaluate
recovery and kinetics by size fraction (Thompson, Runge and Dunne 2019). During this research
the primary grind size was determined to be 175 um according to the grinding circuit product at
NEMISA the beneficiation plant. This particle size was confirmed by mineralogical
characterization analysis to verify the copper liberation size contained in minerals such as
chalcopyrite and bornite.
2.4.3 Flotation Studies
Froth flotation is one of the most used mineral separation methods that currently exist in
the mining industry. Froth flotation is constantly being improved and expanded to process
greater tonnes and to treat more complex mineralogy. It started to treat sulfide minerals
containing copper, lead and zinc but also has been developed to include metals such as nickel,
platinum and gold-hosting sulfides. Froth flotation is a separation process which takes
advantages of natural and differences in surface properties of the minerals. For example, some
minerals are hydrophobic which means that the surface of these minerals repel water. This
physical property allows that the mineral particles can attach to air bubbles in order to be
separated through flotation. The flotation process begins from a hydrophilic state where applying
certain chemical conditions in the system to allow a transition to a hydrophobic state. The
important mechanism during flotation is the attachment of the valuable minerals to the bubbles
which are recovered to the concentrate.
The separation is performed using a flotation cell that consists in a cell with an impeller
that rotates mixing the slurry and also injects air to the bottom of the cell. A flotation cell is a
complex system because there are three phases (solids, liquid and gas) after treatment with
reagents the surface properties of the valuable minerals are modified allowing the separation
between a concentrate and tailings respectively (Wills 2016).
11
|
Colorado School of Mines
|
During flotation concentration the activity of a certain mineral surface in relation to
flotation reagents in water depends on the forces modifying the physical properties of the
surface. Figure 2.5 shows the flowsheet of the beneficiation plant at NEMISA, it consists of the
grinding ciruit followed a flotation circuit to produce a final copper concentrate.
Figure 2.4: NEMISA beneficiation plant flowsheet
2.4.3.1 Flotation of Sulfide Ores
It is common to find that the mineralogy of base metals such as copper, zinc, nickel,
cobalt and lead are associated with sulfide minerals. These minerals are commonly treated
through froth flotation where the metals report to the concentrate. Some of the main
characteristics of these minerals are (Ndoro, Lordwell and Witika 2017):
• Sulfide minerals are covalently bonded which result in having a low solubility.
• Some sulfides show natural hydrophobicity which means that there is not need of a collector
to complete flotation.
• Commonly sulfide ore minerals are floated using collectors such as xanthates and dithio-
phophates, on the other hand oxidized minerals do not have the same respond to these
collectors.
Recovering copper from sulfide minerals such as chalcopyrite (CuFeS ), bornite
2
(Cu FeS ), covellite (CuS) and Chalcocite (Cu S) can be achieved by crushing and grinding to a
5 4 2
required particle size that will allow separation of copper from the gangue by froth flotation.
Beneficiation plants typically have series of flotation cells with chemical addition points and re-
12
Tolvade
Paso
QuebradorasTertiaria
de Cono
Quebradorade
Primaria
Stockpile DOLORESCOBRIZA
QuebradoraSecundaria
de Cono
Tolvade Finos
Molino 4
Molino 5
Molino 3
Molino 6
Molino 7
Molino 8
Molino 1
M olino 2(futuro)
Patio de
Concentrado
Tanquede AguaFlot 1raEtapa
Flot 2daEtapa
Flot 3raEtapa
Flot 4taEtapa
Flot
Limpieza
Tanque Espesador 1
Filtro n°1
ConcentradoFiltrado
Flot 5taEtapa
Tanque Espesador 2
Filtro n°2
Presa de Jales
Flot 6taEtapa
Agua de Retorno
Tanque
Espesador 3
|
Colorado School of Mines
|
grind mills designed to maximize recovery of copper in the concentrate. As mention before
mechanical cells injects air into the slurry through a rotating impeller which agitates the slurry
causing that the bubbles disperse throughout the cell. The use of required reagents that are
required to change the surface properties of the valuable minerals to hydrophobic in order to be
floated must be economically analyzed (Crundwell, et al. 2011).
2.4.4 Flotation Reagents
Reagents or chemical compounds are used during froth flotation to change the sulfide
minerals physical properties, of valuable minerals as previously mentioned. There are three
different types of compounds used during froth flotation: Collectors, frothers and suppressors.
2.4.4.1 Collectors
Collectors are used to change the hydrophobicity of the valuable minerals contained in
flotation in order to float them. These kind of chemicals are organic compounds that change the
surface properties of selected minerals to water-repellent by adsorption on their surface. Collectors
can be nonionizing and ionizing compounds; the first ones are insoluble and hydrophobic. This
kind of collectors are used on minerals with hydrophobic properties such as coals and molybdenite
to impulse their floatability.
The ionizing collectors are soluble, the molecule contains a nonpolar hydrocarbon group
and a polar group. These collectors are considered heteropolar, the nonpolar hydrocarbon radical
has water-repellent properties while the polar ones cause that the molecule is soluble. Ionizing
collectors can be classified by the type of ion (cationic or anionic) and by the application for sulfide
or non- sulfide minerals. On the other hand, anionic collectors are divided in two types depending
on the structure if the polar group: sulfhydryl type and oxyhydryl type (Wills 2016).
During the flotation of sulfide minerals, the most used collectors are xanthates, di-
thiophosphates, and the carbamates. The surface adsorption reaction during the flotation process
is through a sulfur atom whereby bonding properties there are modification of the surface
properties by N and O. (Adkins and Pearse 1992). Xanthates are the most important thiol collectors
and they require regulating agents in order to achieve selectivity between sulfide minerals.
This type of collector has a good water solubility and stability in alkaline conditions. Di-
thiophosphates are the second most used thiol collectors and they can be used alone and usually
are used in together with xanthates or other collectors’ alternatives. Also, these collectors are most
stable over a wide range of pH (Wills 2016). During this research a xanthate collector was used
13
|
Colorado School of Mines
|
for floating the copper sulfides minerals during froth flotation as required by the beneficiation
plant process.
2.4.4.2 Frothers
These compounds are specifically for creating a flotation froth according to the sample
mineralogy and physical properties. The general type of these frothers include long chain
carboxylates, sulfonates or sulfates, and amines. For example, thee frothers that are most widely
used in metals flotation are pine oil, aliphatic alcohols, polypropylene glycol, alkyl ethers of
polypropylene glycol, and cresylic acid. The alcohols are composed of five to eight carbon atoms
such as amyl alcohols, methyl isobutyl carbinol (MIBC), and certain heptanols and octanols
(Vidyadhar 2007). The main functions of frothers are: Aid formation and preservation of small
bubbles, reduce bubble rise velocity and aid formation of froth. An increment of the flotation
kinetics can be achieved by reducing the bubble size causing that there is more total surface area
of bubbles that results in a higher rate of collision with particles. Frothers reduce bubble rise
veolocity which increases the residence time of bubbles in the pulp resulting in an increase of the
flotation kinetics.
Frothers are considered a class of surfactants and most of them are heterpolar compounds
having a polar group such as hydroxy and a hydrophonic hydrocarbon chain. This allows that the
surface-active molecules in water have dipoles combined with the polar groups creating a
tendency to force these surface-active molecules to the air phase (Wills 2016). The major
commercial frothers currently are alcohols and polyglycols (Klimpel and Isherwood 1991).
2.4.4.3 Modifying Agents
These agents are used to make it possible to adsorb collectors with according to different
mineral systems. Modifying agents are commonly classified into different groups, for example:
additives, inorganic or inorganic agents. It is important for the flotation process to choose the
most adequate collector, frother and modifying agents to give a maximum ease and control of the
flotation operation, recovery values, and selectivity (Vidyadhar 2007).
Depressants are a type of modifying agent which are used to increase selectivity by
preventing one mineral from flotating while allowing another mineral to float unimpeded. For
example, in order to separate two sulfide ores a depressant can be used to prevent the froth
formation by one ore and allowing the other to come into froth. Depressants are not required in
the NEMISA flotation circuit because of the simple ore mineralogy which is mainly copper
14
|
Colorado School of Mines
|
sulfides. The modifying reagents used during this research were chosen to simulate the
conditions at the beneficiation plant at NEMISA. The collector used is sodium isopropyl
xanthate (SIPX) which is commonly used for copper sulfides ores froth flotation. It requires a pH
above 8 to be effective. Methyl isobutyl carbinol (MIBC) is used as frother which helps to
increase the kinetics and selectivity during the flotation of copper sulfide minerals. The copper
ore from NEMISA does not require any additional modifying agents such as depressants.
2.4.4.4 Locked Cycle Flotation Test
This flotation procedure simulates a continuous circuit by performing repetitive flotation
batches. The main objectives of a locked cycle test are to produce a metallurgical projection for
the sample tests and to assess if the flowsheet and reagent suite is stable. The procedure consists
in performing different flotation cycles, typically the first cycle is followed by a similar batch
tests which have “intermediate” material from previous cycle added to the current cycles
accordingly. The cycles are continued in an iterative manner for a determined number of cycles,
the final concentrate and final tailings are filtered and removed for further analysis. At the end of
the test all the products, final and intermediate, are dried, weighed and taken to elemental
analysis to perform a metallurgical projection (Ounpuu 2011). Commonly, one concentrate and
one tail are collected from each cycle and the intermediates are recycled to the next cycle until
the final one where the intermediates are taken to elemental analysis respectively.
2.4.4.5 Mass Balances: The n-Product Formula
The n-product formula is a simple mass balance technique that used the assays from the
final products to determine the mass balance of the locked cycle test. The procedure uses the
assay of the feed, concentrate and tailing in the familiar formula (Ounpuu 2011):
(2.3)
𝑓−𝑡
T𝐶h=e r(e𝐹m)i 𝑐n −d 𝑡er of the balance is calculated once C (Concentrate mass) is determined. The
weighted average assay for the final 2 to 4 cycles used. An important requirement for using this
technique is that the circuit must have mass conservation. If the circuit does not have mass
conservation, then the n-product formula should not be used (Ounpuu 2011). The locked cycle
15
|
Colorado School of Mines
|
flotation test is an important resource used in grindability determinations, balling and in leaching
tests (Agar 2000).
2.4.5 The Chemistry of Gold Extraction
Gold is one of the most noble metals that makes it very practical in the industry. For
example, gold is not attacked in air or water by either oxygen or sulphur, and its durability under
corrosive conditions have led its use in jewelry and coinage for centuries. Gold is usually found
in nature in the metallic state, and the only gold compounds that exist in nature are AuTe and
2
AuSb . The thermodynamics of gold predicts that the aurous and auric ions cations are not stable
2
in aqueous solutions., these cations can be reduced by water to metallic gold.
There are many gold complexes with different stabilities, also the properties of gold
complexes vary systemically. Gold compounds (II, III) are B-type metal ions, this means that the
stability of their complexes tends to decrease as the electronegativity of the ligand donor
increases.
2.4.5.1 Gold Cyanidation
The majority of gold extraction from ore is performed with the alkaline cyanide leaching
process. The chemical recovery of gold can be defined by two different operations: oxidative
dissolution of gold and reductive precipitation of metallic gold from the solution. Cyanide is one
of the most attractive lixiviants in the current industrial gold leaching process. The gold cyanide
complexes are more stable than any other lixiviant compound as shown previously on Table 2.2.
The alkaline cyanidation process is general accepted as consequence of its simplicity for
recovering gold. Using standard conditions such as ambient temperature and pressure, a dilute
solution of sodium or potassium cyanide can solubilize gold particles with residence times of 24
to 48 hours yielding of 98% to 99% of gold recovery.
The cyanide ion (CN-) is isoelectronic with carbon monoxide (CO) and nitrogen (N )
2
which means that they have the same electronic structure and same number of valence electrons.
As mentioned before gold is a noble metal and chemically fairly inert but it forms chemical
compounds in two oxidation state: aurous (Au+) and auric (Au+3) in aqueous solution (Haque
1992). There are factors that define an efficient dissolution of gold in a cyanide lixiviant such as
particle size, alkalinity control, oxygen consumption, and cyanide concentration.
16
|
Colorado School of Mines
|
applying mild oxidants such as air or oxygen. Also, the intermediate hydrogen peroxide formed
and consumed in cyanidation is considered as a powerful oxidant in the leaching of gold ores.
The solubility region of gold as Au (CN ) extends in acidic and basic regions. There is a risk of
2
generating HCN gas in acidic conditions, because of its toxicity, generally cyanidation is
conducted in the pH range of 10 to 12 (Haque 1992).
During the gold cyanidation, commonly silver and copper are present in the solution
which causes that these metal ions react with the cyanide (CN-) to form complexes as shown on
the equations 2.11 and 2.12.
• Silver cyanide reaction:
4Ag + 8NaCN + 2H O + O → 4NaAg(CN) + 4NaOH (2.12)
2 2 2
In this case, the copper sulfide minerals can form complexes with cyanide, such as Cu (CN)2,
as the following reaction shows:
• Copper cyanide reaction:
4Cu + 8NaCN + 2H O + O → 4NaCu(CN) + 4NaOH (2.13)
2 2 2
The NEMISA samples from both composites Santa Teresa and Cobriza contain
considerable amounts of silver and copper which have to be removed from the pregnant solution
in order to recirculate the cyanide to the leaching process.
2.4.5.2 Alternatives Lixiviants to Cyanide
Cyanide is considered as hazardous compound because of its toxicity; there is
environmental pressure by different groups around the world to ban the industrial use of cyanide.
Recently, research on replacing cyanide as lixiviant has been made and it has shown that there
are other compounds such as thiosulfate, thiourea, halides, sulfide systems, ammonia, bacteria
natural acids, thiocyanate, nitriles and combinations of cyanide with other compounds (Adams
2005). Many of these alternative gold processes are still at early development stages, a key factor
for the commercial success of these alternative lixiviants is the stability of the lixiviant and the
gold complex in solution. Table 2.1 shows the stability constants and standard reduction
potentials for gold complexes. Clearly the cyanide complex is more stable than any of other
alternative reagents with thiosulfate, thiourea and bisulfide several orders of magnitude less
stable.
18
|
Colorado School of Mines
|
The wide range of values of stability constants of the gold complexes results that the
standard reduction for the different gold ligand species vary by almost 2 V. Cyanide is a very
selective lixiviant which reacts with ores containing gold to form a cyanide complex selectively.
The high oxidizing potentials involved with some lixiviants lead to high reagent consumptions
due to reaction with sulfide minerals contained in the ore and the oxidation of the reagent itself.
A point to consider is the adsorption of reagents and/or precipitation of gold onto some gangue
minerals that can affect the overall gold recovery.
2.4.5.3 Sulfidization, Acidification, Recycling and Thickening (SART) Process
Recently, new processes have been developed to treat the effects of complex copper ores
for gold recovery. The presence of soluble cyanide in the process of cyanidation of gold-copper
ores and concentrates increases the cyanide consumption to achieve a sufficient gold recovery. In
order to recover or detox cyanide additional processes must be implemented; one of these
processes is the SART process. SGS Lakefield Group and Teck Corporation developed the
SART process in the 1990s (Littlejohn P, 2013). The benefits of the SART process in the
cyanidation process is that it breaks up the base metal cyanide complexes, precipitates the metals
as high- grade sulfide concentrates and frees the cyanide for recirculation to the leaching process
(Kratochvil D, 2018).
The Sulfidization, Acidification, Recycling and Thickening (SART) Process process is
described by the following sequence of unit operations:
2.4.5.4 Sulfidization and Acidification
During this stage, the cyanide solution is mixed with NaHS as the sulfide source and
H SO to decrease the pH between 4-5 to form Cu S using a precipitator reactor and thickener to
2 4 2
recover this precipitate as a co-product. As shown on equation 2.10, the generated HCN is
mainly in aqueous form but there is HCN volatization in the SART process which depends on
the site conditions (temperature and pressure), operational parameters (pH and cyanide
concentration in the feed solution and design criteria of the SART process (reactors and
mechanical conditions of the equipment). The generated HCN gas is scrubbed with NaOH and
water to avoid any HCN gas liberation to the atmosphere (Estay, Becker , et al. 2011). This
process recovers cyanide and copper from cyanide solutions based on sulfidization reaction as
shown in the following equation 2.10 (Estay, et al. 2020) (Estay, et al. 2020):
19
|
Colorado School of Mines
|
2Cu (CN) -2+ 3H SO + S-2→ CuS + 6HCN + 3SO -2 (2.14)
3 2 4 (s) (aq) 4
This reaction must be carried out at a pH level below 5.5, with the generated precipitate is
recovered using sequential stages of thickener and a filter press.
2.4.5.5 Recycling
The remaining solution containing HCN is neutralized using CaO to form CaSO
4
(Gypsum) and to recycle the cyanide. Figure 2.7 shows that during the first stage the
sulfidization reactor feeds the precipitation slurry generated into a thickener. The underflow of
this reactor is recirculated to the same reactor and the overflow is neutralized in a following
reactor. At the neutralization reactor the pH increases to 10.5 to return the cyanide solution to the
gold cyanidation plant.
2.4.5.6 Thickening
The gypsum is precipitated using a settler and flocculent to separate the solution Ca
(CN) and the gypsum. The solution is recycled to the overall leaching circuit, having all of its
2
cyanide content as soluble Ca(CN) . The soluble Ca(CN)2 is equivalent to free cyanide for the
2
purposes of gold dissolution in the cyanidation process, and the gypsum is precipitated for
disposal. All equipment must be connected to a srubbing system to avoid emissions of HCN
and/or H S to the environment.
2
These three stages of the SART process define the treatment of the cyanide solution after
the stripping of the activated carbon. Figure 2.5 shows the general flowsheet of the SART
process:
Figure 2.5: SART Process Flowsheet (Estay, et al., 2020).
The SART process is essentially suited to the treatment of leach liquors that contain high
concentrations of base metals as weak acid dissociable cyanide complexes. For copper-gold ores
the SART process reduces the impact of the Cu and Ag cyanide complexes during the leaching
process. It recycles the cyanide to diminish the cyanide consumption having operation costs
20
|
Colorado School of Mines
|
CHAPTER 3: EXPERIMENTAL EQUIPMENT & PROCEDURES
This chapter describes the procedures that were followed for experimentation of the
metallurgical program. Experiments were performed in the laboratories of the Kroll Institute for
Extractive Metallurgy (KIEM) at the Colorado School of Mines.
3.1 Sample Preparation
NEMISA provided two containers with the ½ inch drill cores with a total weight of 521.15
kg which 260.27 kg corresponds to Cobriza and 260.88 kg to Santa Teresa respectively as shown
on Figure 3.1.
Figure 3.1: 1/2 drill core samples received at Colorado School of Mines
These samples were previously analyzed at NEMISA’s chemical laboratory as part of the
extraction procedure performed by the geological department having the following compositions
for gold, silver and copper:
Table 3.1: Drill cores elemental analysis before being shipped
This elemental analysis helped to confirm that the content of gold, silver and copper in the
drill cores were representative to proceed with the metallurgical testing and precious metals
22
G
S i
o
l v
l d
e
C o m
C o p p
( F A
r ( F A
p o
e r
/ A
/ A
s i t e
( % )
A ) m
A ) m
g
g
/ k
/ k
g
g
S a n t a T e r
1
0
4
e s a
.0 5
.6 5
8 .7
S a m p l e C o b r i z a S
1 .6 4
2 .9
4 2 .2
a m p l e
|
Colorado School of Mines
|
extraction. Table 3.1 shows that both composites have gold content and Cobriza has higher gold
and copper values than Santa Teresa.
The drill cores were prepared to a particles size of passing 3.5 mm prior to grindability
studies by using a laboratory scale jaw crusher. The opening of the jaw crusher was set to 3.5
mm which is the size required for the Bond Work Index test. The drill cores for each composite
were crushed and organized into buckets before splitting as shown on Figure 3.2.
Figure 3.2: Laboratory scale jaw crusher used to reduce particle size to -3.5 mm
The crushed material was screened by using a US Sieve series No. 6 mesh (3.36 mm
opening). The material retained on the screen was recirculated to the crusher.
3.1.1 Coning and Quartering
As shown in section 2.4.1, coning and quartering is a method for splitting large amounts
of material. For this case, it was required to split approximately 260 kg of crushed material for
each composite. This method usually required a tarp and a shovel in order to perform the
splitting according to the following procedure:
1. The tarp is expanded in a determined area with ventilation and flat surface.
2. The material is poured forming the material into a conical heap by using a shovel upon the
tarp.
3. The heap is divided by a cross and then separated in quarters.
4. Two opposite corners are taken as the other two set aside.
23
|
Colorado School of Mines
|
5. The taken corners are separated into buckets.
6. The remaining material is again poured and mixed following the same procedure until the all
the material is split and separated into buckets.
Figure 3.3: Samples being split by the coning and quartering method
3.2 Material Analytical and Mineralogical Characterization
Samples were then obtained by using a Jones riffle splitter to produce 200 grams of
passing 3.5 mm from each composite material. Gary Wyss of Montana Tech was sent a total of
400 grams for mineral characterization. The following analyses were carried out at the Center
for Advanced Mineral and Metallurgical Processing (CAMP) of Montana Tech by Gary Wyss.
The received samples (-60 mesh) were wet sieved into sized fractions for preparation of
overall mineralogy. The material was separated from the 100 X 200 mesh by particle density
using heavy liquid separation (HSL). The dense fraction was analyzed by using scanning
electron microscope-energy disperse X-ray analysis (SEM-EDS) referred to as Mineral
Liberation Analysis (MLA).
3.2.1 MLA Particle Size Distribution
The received samples (-60 mesh) from both composites had a particle size distribution
P of 160 um as shown on Figure 3.4. The Cobriza material was finer than the Santa Teresa
80
having a P of 60 um for Cobriza and 75 um for Santa Teresa.
50
24
|
Colorado School of Mines
|
Table 3.2: Content by mineral grouping (wt%)
Copper was calculated at 2.9% in the Cobriza sample and it was 1.7% for Santa Teresa,
silver and gold content of the samples was too low for MLA to provide reliable results.
3.2.3 Copper Mineral Distribution
The two most important copper-bearing minerals in the samples were chalcopyrite and
bornite. Bornite is the primary copper-bearing mineral in the Cobriza sample with 53% of copper
distribution composition., 42% of the copper is provided by the chalcopyritre content in the
Cobriza sample. Chalcopyrite provided nearly all the copper in the Santa Teresa sample with
95% of total copper. Bornite provided 1.5 % of the total copper in the Santa Teresa sample
which is considerably low in comparison with the Cobriza sample. The following Table 3.3
shows the copper distribution by mineral for both composites.
Table 3.3: Copper distribution by mineral (%)
3.2.4 Silver Mineral Distribution
The most important silver-bearing minerals in the samples are acanthite (Ag S) and
2
hessite (Ag Te) with considerable matildite content in the Santa Teresa sample. Only thirteen
2
26
|
Colorado School of Mines
|
3.2.7 Copper Mineral Associations
As mentioned before the two predominant minerals in both composites are bornite and
chalcopyrite. In the Cobriza sample these minerals were associated with each other but bornite
was also associated with the silicates, andradite and hedenbergite and to a lesser degree with the
copper-bismuth sulfide, wittichenite. Similarly, chalcopyrite was associated with andradite,
hedenbergite, quartz and a weaker association with calcite.
In the Santa Teresa sample, chalcopyrite was associated with the silicates, andradite,
quartz and had a weaker association with hedenbergite. Also, there were chalcopyrite
associations with calcite and pyrite.
3.2.8 Silver-Gold Mineral Associations
Silver and gold mineral associations were found in the copper sulfides, bornite,
chalcopyrite and wittichenite. There were silver-gold minerals associations found in the gangue
minerals andradite, calcite, siderite, plagioclase and the sulfides. The following Table 3.5 shows
the silver-gold associations by mineral group for both composites:
Table 3.5: Silver-gold mineral associations by mineral group
3.2.9 Mineral Liberation Analysis (MLA) Images
The analysis performed by particle density using heavy liquid separation (HSL) showed
that the silver-gold associations for the Cobriza composite hessite was found primarily with the
copper sulfides, bornite, chalcopyrite and wittichenite. For the Santa Teresa composite the
primary silver minerals were acanthite and matildite. In general, for both composites silver
mineral associations were the greatest with copper sulfides, followed by mild associations with
the sulfides and silicates.
30
|
Colorado School of Mines
|
Figure 3.12: Cobriza particle photomicrographs by HLS: Agate (AgTe) inclusions in
chalcopyrite (CuFeS ).
2
3.2.11 Santa Teresa MLA and Particle Photomicrographs Images
For the Santa Teresa sample, the MLA image on Figure 3.13 shows a small inclusion of
matildite at the boundary between chalcopyrite and an iron silicate (ferrosilite). As shown in the
Table 3.3 on page 40, the Santa Teresa sample has a low composition of bornite in comparison
with the Cobriza sample. This can be confirmed as well in the highlighted particle in the false
color MLA image on Figure 3.13.
According to the MLA image generated from the “sink” fraction by the HLS analysis, the
Figure 3.14 shows a particle with an inclusion of a matildite grain. The matildite and
wollastonite grains arelocked in andradite, also the particle contains mainly hedenbergite. There
are small acanthite inclusions in pyrite as shown on Figure 3.15. this pyrite particle had narrow
seams of acanthite within the pyrite particle. As mentioned before Santa Teresa has a 94.9 % of
chalcopyrite composition, an acanthite grain of about 10 um was attached to chalcopyrite as
shown on Figure 3.16 A. Figure 3.16 B shows a band of acanthite at the phase boundary
between pyrrhotite and chalcopyrite.
33
|
Colorado School of Mines
|
Figure 3.15: Santa Teresa particle photomicrographs by HLS: Acanthite (Ag S) inclusions in
2
pyrite (FeS ).
2
Figure 3.16: Santa Teresa particle photomicrographs by HLS: Acanthite (Ag S) attached to
2
chalcopyrite (CuFeS ) (A) and acanthite at the grain of pyrrhotite (FeS) and chalcopyrite (B).
2
Bornite and chalcopyrite were the main copper-bearing minerals in the Cobriza sample,
each accounting close to the half of the total copper. On the other hand, chancopyrite had 95% of
the total copper. The size distribution P ’s for bornite and chalcopyrite were nearly identical in
80
the Cobriza sample at around 130 um. The chalcopyrite P in the Santa Teresa sample was also
80
around 130 um but it was under 70 um for the bornite and was not a significant contributor to the
copper balance. The HLS study was performed to prepare the ‘sink’ fraction for the silver-gold
minerals analysis. The silver telluride, hessite was mostly associated with the copper minerals,
bornite, chalcopyrite, wittichenite and some silicate and pyrite. The gold-containing phases
35
|
Colorado School of Mines
|
found in the Cobriza sample were rare, but electrum, native gold and petzite were identified. The
silver and silver-gold grains were relatively fine at around 10 um and finer with some up to 20
um. The silver-bearing mineral grains were slightly larger in the Cobriza than in the Santa Teresa
sample.
3.3 Head Elemental Analysis
Head samples from both composites were sent to Hazen Research for elemental analysis
of gold, silver and copper. Table 3.6 shows the content for each element. The gold, silver and
copper concentration are higher in the Cobriza sample than in the Santa Teresa. Table 3.6 shows
the comparison between the drill cores head analysis sent by NEMISA and the one performed at
Hazen laboratories prior to metallurgical testing. The values for the three metals for both
composites are very similar. The gold content in the Santa Teresa sample is lower which can
cause a possible poor gold recovery during the metallurgical studies.
Table 3.6: Head elemental analysis comparison for Santa Teresa and Cobriza samples.
NEMISA head elemental analysis Hazen head elemental analysis
Composite Santa Teresa sample Cobriza sample Santa Teresa sample Cobriza sample
Copper (%) 1.05 1.64 1.04 1.50
Gold (FA/AA) mg/kg 0.65 2.9 0.59 2.63
Silver (FA/AA) mg/kg 48.7 42.2 42.3 44.00
3.4 Comminution Studies
As mentioned in the section 2.4.2 of the chapter 2 on page 15, the Bond Ball Index Test
was performed to measure the resistance of the material to crushing and grinding. The material
from both composites was previously crushed to – 6 US mesh (3.36 mm) as required prior to the
standard Bond test. The liberation particle size was 175 um which is the particle size used in the
grinding circuit at NEMISA.
3.4.1 Bond Ball Index Test
The feed is prepared by crushing to -6 US mesh screen. Eighty percent (80%) of the ore
for each composite should pass 6 mesh but to be retained on a 14 US mesh screen (-6, +14). The
ore samples are screened and packed into a 700 cm3 graduated cylinder, and the weight is placed
36
|
Colorado School of Mines
|
in the mill and ground dry at 250% circulating load. The used equipment is a F.C. Bond Ball
Mill which is a small universal laboratory mill used in calculating the grindability of all ores.
The mill runs at 70 rpm and has a grinding charge of 285 iron balls, ranging in size from 5/8 in to
1 ½ inch in diameter, weighing 20,125 grams and a calculated surface area of 842 sq. inches.
Table 3.7 shows the specifications of the FC Bonds Ball specifications. The following standard
procedure determines the Bond Ball Index according to SEPOR FC BOND MILL:
1. The feed material is prepared by crushing ore to -6 US mesh. Eighty percent (80%) of the ore
to be tested should pass 6 mesh but be retained on a 14 mesh screen (-6, +14). The -14 US
mesh material is screened and weighed to determine the particle size distribution of the feed.
2. The retained 14 mesh particles are placed into a 1000 ml cylinder to determine the
approximate weight of 700 ml of feed. It is important to compact the ore by shaking the
cylinder at 700 ml in volume, then the weight is recorded.
3. Sample splitters may be used to split numerous increments of the feed, and each increment
poured into the cylinder until 700 ml volume is obtained. The Ideal Period Product (IPP) is
equal to the weight of the 700 ml in grams divided by 3.5.
4. A screen analysis is conducted to analyze each size fraction of the 700 ml feed test sample,
the weights are recorded.
5. The FC Bond Balls and the feed are charged in the mill.
6. For the first test the mill revolution counter is set for a specific revolution for the first cycle
(typically 50 for a coarse or 100 for finer grinding). The mill start button has to be pushed
and when the number of revolutions has been reached, the mill stop.
7. The mill must be emptied by using a screen to retain the grinding balls. The balls must be
returned to the mill.
8. A screen analysis is performed on the material and each size fraction is weighted and
recorded. The weight of the undersize -140 US mesh screen or the size fraction being
reduced to in microns.
9. The amount of undersize product present in the test feed is determined.
10. The number of net grams produced per revolution is calculated by dividing the undersize
weight (grams) by the number of revolutions the mill rotated.
37
|
Colorado School of Mines
|
11. The weight which should be ground in the next cycle in order to obtain the desired
circulation load is calculated by IPP minus the amount of product size material present in the
feed.
12. The number of revolutions for the next cycle is determined by dividing the number of the
desired circulating load by the number of grams produced per revolution.
13. Representative feed has to be added to replace the ground product size material. The grinding
balls are placed in the mill and started with the new number of revolutions.
14. This procedure is repeated until the grams per revolution values have equilibrium. A plot of
net grams per revolution vs cycle number should show a upward or downward trend, and
finally a reversal of the trend on the 5th test. If there is not reversal, the test has to be
continued until no significant change occurs in the net grams per revolution.
Table 3.7: FC Bonds Ball specifications (SEPOR, Operating & Maintenance Manual 2019)
38
|
Colorado School of Mines
|
Figure 3.19: Bond Ball Mill discharged after the first cycle
3.4.2 Determination of Bond Work Index Parameters
As mentioned in the section 2.4.2 on page 16 for the Bond Ball Mill procedure, the
parameters to be determined during the experimentation are used for the equation 2.1 to calculate
the BWI (kWh/t). These parameters are:
• F = Sieve mesh size passing 80% of the feed before grinding (um)
80
• G= Weight of the test sieve fresh undersize per mill revolution (g/rev)
• P = Opening of the sieve size passing 80% of the last cycle test sieve undersize product
80
3.4.3 Feed Particle Size Analysis
Particle size analysis was performed for both composites feed prior to the Bond Ball
Index test by using a Ro-tap shaker sieve as shown on Figure 3.18 to determine the parameter F
80
for both composites. Three different feed particle analysis were performed for each composite to
determine the average F of the three different analysis. The feed had to pass 150 US mesh (89
80
um) but to be retained on a 14 US mesh (1400 um). The average F for the Cobriza sample is
80
3836 um, Figure 3.19 shows the logarithmic regression between % passing vs particle size (um),
this graph was used to determine the F for each particle size analysis.
80
The three particle size analysis tests showed in the Figure 3.20 have small variations, but
they showed a consistent particle size range. Likewise, the Santa Teresa sample has a F of 3436
80
um which is a smaller size than the Cobriza sample. Figure 3.21 shows the logarithmic
regression between % passing vs particle size (um) as well, it has a better correlation between the
three different particle size analysis in comparison with the Cobriza sample.
40
|
Colorado School of Mines
|
3.4.4 Product Particle Size Analysis
The cycles needed to reach equilibrium for the Cobriza sample are nine, Table 3.8 shows
the last grams per revolution (g/rev) values for the last three cycles. These values show
equilibrium as result of the consistency of the g/rev results, a final product was obtained from the
last cycle product and a particle size analysis was performed. Likewise, for the Santa Teresa
sample the needed cycles for reaching equilibrium were six. Table 3.9 shows the g/rev values for
the last three cycles during the BWI procedure. During the experimentation it was noticed that
the material from the Cobriza sample showed higher hardness properties than the Santa Teresa
sample.
Table 3.8: Grams of undersize product per revolution (g/rev) for Cobriza sample
Table 3.9: Grams of undersize product per revolution (g/rev) for Santa Teresa sample
In order to use the values shown on Table 3.8 and Table 3.9 ta calculate the BWI using
equation 2.1, these values have been averaged. The G value for Cobriza is 0.80 grams per
revolution and for Santa Teresa is 0.82 grams per revolution, respectively. The P for each
80
composite was determined by a wet particle size analysis on the undersize product. The
undersize of each composite had to pass a 150 US mesh but to be retained on a 400 US mesh.
Logarithmic regression was used for both samples for determining the P as required for the
80
BWI calculation. The P for the Cobriza sample was 85 um, and the P was 108 um according
80 100
to the logarithmic regression. The Santa Teresa sample had a P of 87 um and a P of 111.
80 100
42
C
C
y
y
c
0
c
0
le
.7
l e
. 7
G r
# 7
9
G r a
# 4
8
a
m
m
s
s p
p e
e r r e
r r e
v
v
o
o
lu
l u
t
t
io
C
i o
C
n f o
y c le
0 .8
n f o
y c l e
0 . 8
r t h
# 8
5
r t h
# 5
5
e
e
la s t
l a s
t h
t t
r
h
e
r
e
e
c y
e c
c
y
le s
C y
c l e
C y
c le
0 .7
s
c l e
0 . 8
# 9
5
# 6
3
|
Colorado School of Mines
|
kg of material with 1.2 kg of water to create a pulp with 50% solids prior to be charged into the
rod mill for each charge.
Each cycle of the test consists in charging the rod mill with the 50% pulp sample for a
determined amount of time. For the Cobriza sample the performed cycles lasted the following
times 30, 40, 50, 55, 60, 75, 80 and 90 minutes. Likewise, for the Santa Teresa sample the test
was performed for 20, 30, 40, 45, 60, 70, and 80 minutes. The following procedures shows the
rod mill operation for a cycle:
1. The feed material ( -12 US mesh) is splitted in order to extract a representative sample of 1.2
kg.
2. The feed material is mixed with 1.2 kg of water to create a pulp containing 50% solids and
2.4 kg total weight.
3. The sample and the rods are charged into the mill.
4. The timer is set to the desired time and the charged rod mill is located on the rolling bed as
shown on Figure 3.24.
5. The rod mill is plugged and operating for the desired time as shown on Figure 3.25.
6. The rod mill is unplugged when the time is completed, the rods are removed first and
brushed into a bucket. The material remaining in the mill is poured into the bucket. The mill
walls are brushed clean into the bucket as well.
7. The bucket is located on the floor to let the solids settle.
8. The water is decanted as much as possible, once the solids are settled
9. The remaining pulp in the bucket is poured and washed through a 100 US mesh (150 um)
using the sieve vibrator. The passing 100 US mesh part is separated in another bucket and the
solids are separated in other bucket as well.
10. The passing 100 US mesh pulp is poured through a 325 US mesh (44 um) using the sieve
vibrator. The passing 325 US mesh is separated and dried in the oven.
11. The retained particles of the 100 and 325 US mesh are combined and dried in the oven.
12. The cake as shown on Figure 3.26 of the retained particles of the 100 and 325 US mesh
particles, is broken up.
13. The cake is screened using the Ro-tap sieve shaker and sieves between 45 US mesh (355 um)
and 325 US mesh (44 um).
14. The passing of 325 US mesh is combined with the cake from the wet screening on step 10.
44
|
Colorado School of Mines
|
Figure 3.27 shows the graph with the P from each grinding cycle, this correlation
80
helped to determine the required time for grinding to the required P of 175 um. The time
80
required for the Cobriza sample was 50 min. The Figure 3.28 shows the Santa Teresa grinding
times as well and according to the interpolation in the graph the time required to reach the
desired P was 58 min respectively. These grinding times helped to prepare feed material with
80
liberated minerals such as chalcopyrite and bornite prior to the flotation studies.
3.5.2 Flotation Reagents
According to the mineralogy of the Cobriza and Santa Teresa composites, the separation
process of the contained valuable minerals by froth flotation requires the following reagents:
sodium isopropyl xanthate (SIPX) as collector agent for the sulfide minerals and methyl isobutyl
carbinol (MIBC) as frother. As mentioned before these two reagents are used in the beneficiation
plant at NEMISA. The ore pulp prior to the flotation circuit has a pH approximately of 8.5. For
this reason the separation process of the valuable minerals such as bornite and chalcopyrite for
copper extraction does not require further pH modification. The purpose of these flotation studies
was to create a metallurgical projection for copper, silver and gold recovery. Additionally, the
tailings from the flotation tests were used for the gold cyanide leaching studies for these future
ore samples.
3.5.3 Flotation Retention Times
The flotation retention time studies for rougher, scavenger and cleaner were performed to
determine the flotation required time for each stage prior to the flotation locked cycle testing.
The flotation tests were performed using two different Denver flotation cells with a capacity of 3
and 5 liters as shown on Figure 3.29. The samples were ground by using the rod mill with the
required grinding time for each composite as mentioned in the section 3.5.1. It is recommended
for the collector SIPX and the frother MIBC to maintain a pH above 8 for a better performance.
As the pH of both sample in pulp with 25% (weight) solids had a pH around 8.5, it did not need
any further pH modifiers. The temperature used during the flotation tests was 25 C for each
stage.
The metallurgical laboratory at NEMISA provided the collector and frother
concentrations used for flotation testing at the beneficiation plant and these concentrations were
adjusted according the laboratory flotation cells capacity and the generated 40% solids pulp at
48
|
Colorado School of Mines
|
Colorado School of Mines. The concentration of the collector used during the experimentation
was 25g/tonne which is equivalent to a solution concentration of 7.9 mg/ L for a slurry density of
25%. The frother MIBC used concentration was 45 g/tonne which is equivalent to a solution
concentration of 9.5 mg/L of solution.
3.5.3.1 Rougher Retention Time
The rougher retention time was obtained by charging 1.2 kg of ground feed material with
a P of 175um from each composite into the Denver flotation cell with a 5 L capacity and a
80
slurry with 25% (weight) solids. The following procedure details the retention time test
performed to determine the optimum flotation time:
1. The 1.2 kg feed material is prepared (50% weight solids) ground in the rod mill according to
the time determined by the primary grind size analysis.
2. The ground material (slurry) from the rod mill is filtered using a press filter to remove water.
3. The slurry is generated in the 5 L flotation cell by pouring 3.75 L of water creating a slurry
containing 25% (weight) solids.
4. The flotation cell impeller is positioned in the cell and turned on to 1250 rpm.
5. After 10 minutes of the slurry being agitated in the cell, the pH is measured by using a pH
meter.
6. The collector and the frother are added to the flotation cell. It is recommended that the cell is
agitated with the reagents for 15 minutes.
7. The air valve is opened to start the froth floating in the cell. The timer must be set for the
total flotation time, for this case the time was 4.5 minutes.
8. Using a pan and a palette every 0.5 minutes the froth was swept from the top. Every 0.5
minutes a different pan was used.
9. The air valve was close when the flotation time ended. The removed froth in the pans are
filtered and dried in the oven.
10. The dried samples of each flotation time were sent to Hazen Research for fire assay analysis
for Au and Ag. For Cu the samples were analyzed by Atomic Adsorption Spectrometry
(AAS).
The two samples from the Cobriza and Santa Teresa composites floated rapidly within
the first 3 minutes, for this reason the timer was set to a period of 4 minutes for performing the
previous procedure.
49
|
Colorado School of Mines
|
3.5.4 Locked Cycle Flotation
As mentioned before (section 2.4.3., page 23) the locked cycle testing is a repetitive batch
used to simulate a continuous circuit. A locked cycle test is a series of repetitive batch tests
conducted on a small-scale laboratory equipment, in which the middlings generated in the one
test (nth cycle) are added to the subsequent test (n+1)th cycle.
In order to simulate the continues flotation circuit of the beneficiation plant at NEMISA,
the locked cycle flotation was performed with six cycles. Each cycle consisted in a rougher, a
scavenger and a scavenger stage respectively. The retention times determined in the previous
section (page 68) were used to perform the locked cycle flotation testing. The concentrations of
the flotation reagents used during the locked cycle flotation testing for the collector SIPX was 25
g/tonne and for the frother MIBC was 45 g/tonne respectively. The pH was monitored and
stayed at 8.5, the slurry density for each flotation stage was 25% (weight) solids and the impeller
velocity was 1250 rpm. The following steps shows the procedures to perform a single cycle
during the locked cycle flotation:
1. The rougher stage sample is prepared by having 1.2 kg charge of feed material ground using
the rod mill for the specific grinding time depending on the sample (Cobriza 50 min, Santa
Teresa 58 min). The ground sample had a P of 175 um as determined previously in the
80
primary grind size study. The feed material was filtered to remove water and weighted.
2. The filtered sample was transferred to the flotation 5 liter cell and water was poured to create
a 25% (weight) solids slurry, the impeller was set in the cell with a speed of 1250 rpm for 15
minutes for mixing.
3. The pH was monitored to stay in the range of 8-9.
4. The collector and the frother were added to the cell and conditioned for 15 minutes before
turning on the air valve.
5. The timer was set to the rougher retention time for each sample according to Table 3.11. The
air valve was opened and the frother was removed from the cell during the rougher retention
time.
6. The concentrate was filtered, and the tailings was filtered and dried in the oven.
7. For the cleaner stage, the concentrate is weighted and transferred to the 3 liter flotation cell.
Water was poured to get a slurry of 25% (weight) solids. Flotation parameters were set such
55
|
Colorado School of Mines
|
as pH and impeller speed. Flotation reagents were added to the slurry and conditioned for 15
minutes.
8. The air valve was opened, and the froth was removed during the cleaner retention time
mentioned on Table 3.11.
9. The concentrate was filtered, dried in the oven and labeled as Cleaner Concentrate Cycle #1
(CC1) for fire assay analysis. The tailings were labeled as Cleaner Tailings Cycle #1 (CT1)
were filtered, weighted and returned to the flotation cell for the next cycle.
10. For the scavenger stage, the concentrate was weighted and transferred to the 3 liter flotation
cell. Water was poured to get a slurry of 25% (weight) solids. Flotation parameters were set
such as pH and impeller speed. Flotation reagents were added to the slurry and conditioned
for 15 minutes.
11. The air valve was opened, and the froth was removed during the scavenger retention time
mentioned on Table 3.11.
12. The concentrate was filtered, dried in the oven, labeled as Scavenger Concentrate Cycle #1
(SC1) and returned to the flotation cell for the next cycle. The tailings were labeled as
Scavenger Tailings Cycle #1 (ST1), filtered, weighted and sent for fire assay analysis.
13. The next rougher of the cycle #2 was prepared having a new charge of 1.2 kg of feed
material. Also, the Cleaner Tailings (CT1) and the Scavenger Concentrate (SC1) are
considered middlings and were added to the charge for cycle #2.
14. The same procedure for the rougher cycle #1 was performed for the next 6 cycles.
15. At the end of the 6th cycle the middlings (ST6, SC6), the final Cleaner Concentrate (CC6)
and the final Scavenger Tailings (ST6) were sent to for fire assay analysis.
The elemental analysis for silver and gold in the samples of every cycle for both
composites was performed by Hazen Research. The copper content in the samples was analyze
by atomic analysis spectroscopy (AAS) at Colorado School of Mines.
56
|
Colorado School of Mines
|
flotation test for both composites. Figure 3.36 and Figure 3.37 show the elemental distribution
between the final concentrate and the tailings for both samples. Figure 3.36 shows a closer
distribution between the three elements, with copper the element with the higher distribution %
in the final concentrate for the Cobriza sample. Likewise, Figure 3.37 shows a more separated
elemental distribution, with copper the element with the highest percent as well for the Santa
Teresa sample. In other words, the copper was the element with a higher distribution % to the
final concentrate for both samples.
As mentioned before the Cobriza sample had a higher composition of copper minerals
such as chalcopyrite and bornite in comparison with the Santa Barbara sample. This difference
could be noticed in the recovery for copper as shown on Table 3.14, the recovery of copper for
the Cobriza sample was 88% and for the Santa Teresa was 84% respectively. It was noticed that
mainly the copper was effectively recovered from the samples, but that the gold and silver were
not recovered as efficiently as the copper. Relatively, the silver recovery during the flotation test
was satisfactory for both composites. Gold is the element with the lowest elemental distribution
for both composites, which means that significant amounts of gold will be lost in the tailings.
The final tailings from the locked cycle test were used for gold recovery by cyanide leaching.
3.6 Cyanide Gold Leaching from Copper Flotation Tailings
The final tailings from the locked cycle flotation test were split by using a Jones splitter
as shown on Figure 3.17. Also, the tailings were sent for elemental analysis to determine the
copper, silver and gold concentration prior to the leaching process (Medina and Anderson 2020).
Table 18 shows the results of the elemental analysis on the flotation tailings produced during the
locked cycle test. The Cobriza sample had a higher gold concentration than the Santa Teresa
sample. In contrast, the concentration of copper and silver was higher in the Santa Terresa
composite.
Table 3.15: Elemental analysis of the flotation tailings
59
T a ilin g s H e a d
C o p p e r %
G o ld ( m g / k g )
S ilv e r ( m g / k g )
C o b r iz a
0 .0 6 2 1
0 .5 4 9
3
S a n t a T e r
0 .1 1 7
0 .2 4
7 .8 2
e s a
|
Colorado School of Mines
|
3.6.1 Bottle Roll Test Agitated Cyanidation (BRTAC)
As mentioned before (section 2.4.4., page 32) the gold contained in the copper flotation
tailings can be extracted by using cyanide as lixiviant. The basic principle of the cyanidation
process is that weak alkaline cyanide solutions can have a selective dissolving action on gold and
silver. The dissolution rate of gold is affected by cyanide and oxygen concentrations,
temperature, pH, surface area of gold exposed, agitation/mass transport and ions in solution. The
cyanidation test was performed using HDPE lifter bottles with 3.5-liter volume capacity. A
rolling table was used for agitating the bottles during the cyanidation testing time as shown on
Figure 3.38. The leaching method and the titrimetric method analysis for free cyanide (CN-) used
for this experiment was based on the Montana Tech (Gold Processing Laboratory) procedure.
The parameters recorded during the BRTAC test were pH, dissolved oxygen, free
cyanide concentration, lime (CaO) addition and cyanide addition. The following procedure was
followed to perform the BRTAC test for each composite:
1. A representative sample of 1 kg from the flotation tailings was obtained by using a Jones
splitter. The flotation tailings had a P of 175 um as previously prepared prior to the locked
80
cycle flotation test.
2. The leaching time was 32 hours at 40% solids, pH 11 with lime, and different concentrations
of sodium cyanide (NaCN). The pregnant solution was sampled at 2,6,17,24 and 32 hours.
3. The sodium cyanide solutions were prepared for different concentrations (0.3,0.5,0.75 g/L),
also a make-up solution and titrations solutions.
4. The sample was placed in the leaching bottle. 1.5 L of deionized water was poured in the
leaching bottle to get 40% solids.
5. The leaching bottle was placed for 15 minutes on the rolling bed as shown on Figure 3.38.
6. After rolling the leaching bottle containing the sample, the natural pH is measured and
recorded.
7. Lime was added gradually to adjust the pH to 11 and the lime addition was recorded. The
8. The sodium cyanide addition is calculated with the following formula:
N (3.1)
(10 𝑚𝐿−𝑇𝑖𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑟𝑒𝑎𝑑𝑖𝑛𝑔 (𝑚𝐿) 𝑔
𝑎𝐶𝑁 𝑎𝑑𝑑𝑖𝑡𝑖𝑜𝑛 (𝑔)= ( 10 𝑚𝐿 )( 𝑁𝑎𝐶𝑁 𝑠𝑒𝑡 𝑝𝑜𝑖𝑛𝑡 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 (𝐿))(1.5 (𝐿))
60
|
Colorado School of Mines
|
9. The NaCN is added to the alkaline pulp. The leaching bottle is placed on the rolling bed. The
rotational speed was adjusted to 80 rpm. The time, date and temperature were recorded.
10. The leaching bottle was removed from the rolling bed at the scheduled sampling time and
the solids were allowed to settle for 15 minutes. The stop time was recorded.
11. The pH and DO was measured in the leaching bottle and recorded as shown on Figure 3.40
and Figure 3.41.
12. 40 ml of pregnant solution were extracted from the leaching bottle without extracting solids.
13. 30 ml of the pregnant solution were placed into a sample vial for elemental analysis of
Au,Ag and Cu.
14. The 10 ml of pregnant solution were used for titration for free cyanide according to the
titrimetric method.
15. The silver nitrate (AgNO ) addition was recorded.
3
16. The sodium cyanide addition is calculated by using the equation 3.1 and added to the
leaching bottle.
17. 40 ml of lime were added to maintain the pH and the pulp density to 40% solids.
18. After sampling for the scheduled times, the pregnant solution was placed to another container
for detoxification.
19. The solids (leaching tailings) were washed using a NaOH solution and dried in the oven.
20. The leaching tailings were sent to elemental analysis for Au, Ag and Cu.
3.6.2 Titrimetric Method Analysis for Free Cyanide
The determination of free cyanide concentrations during the BRTAC experiments helped
to monitor NaCN reagent consumptions. The free cyanide in alkaline solutions is titrated by
using silver nitrate (AgNO ) to form cyanide complex, dicyanosilver (I) complex as shown on
3
equation 3.2
Ag + + 2 CN- -→ Ag (CN) - (3.2)
2
At the moment when all the CN- has been complexed and a small excess of Ag has been
added, the excess Ag+ is detected by the silver-sensitive indicator (p-
dimethylaminobenzalrhodanine), which immediately turns from a yellow to a salmon color. The
standard titrant solution had a 0.005 M concentration of silver nitrate (AgNO ), a burette was
3
used for the titration as shown Figure 3.39.
61
|
Colorado School of Mines
|
CHAPTER 4: RESULTS ANALYSIS AND DISCUSSION
The results of the chapter 3 will be discussed on this chapter.
4.1 Head Elemental Analysis and Mineral Characterization
The representative drill cores from the two composites had a considerable amount of
copper, silver and gold as shown on Table 3.6. Since the purpose of this project is to produce
flotation tailings from future ores by having similar operation parameters as the beneficiation
plant at NEMISA. The gold content in the Santa Teresa composite was approximately 4.5 times
lower than the Cobriza composite. As mentioned before NEMISA produces a copper concentrate
which means that the copper is the priority to recover in the concentrate. The copper
concentration in the drill cores were 1.04% for the Santa Teresa and 1.5% for the Cobriza.
Hence, the Cobriza composite has more potential to provide more copper and gold than the Santa
Teresa according to the elemental analysis.
The minerology of both composites was very similar having mainly silicates and
carbonates as gangue minerals. The copper is associated with sulfides minerals such as bornite
and chalcopyrite. The Cobriza composite has bornite and chalcopyrite but the Santa Teresa
sample has mainly chalcopyrite as copper mineral association. The content of silver and gold
minerals such as hessite, matildite, acanthite and petzite for silver, and electrum for gold content
was low in both composites in comparison with the copper minerals. As previously mentioned,
(page 44), the liberation of copper minerals such as chalcopyrite was similar for both composites
at around 75% liberation for particles containing 95% or more of chalcopyrite.
Bornite liberation was 70% in the Cobriza sample and 40% in the Santa Teresa sample
for particles containing 95% or more of bornite. Froth flotation was performed to separate the
sulfides and gold, silver and mainly copper. The grinding circuit of the beneficiation plant at
NEMISA has a P of 175 um prior to flotation. This P was used during the experimentation to
80 80
simulate the processing of these future ores to perform the gold cyanide leaching from the copper
flotation tailings.
4.2 Bond Work Index Results
The BWI results as shown on Table 3.10 of both composites were similar, the Cobriza
sample had 18.27 and the Santa Teresa sample had 17.70 according to the Bond Work Index
scale respectively. The Cobriza sample required 0.45% more energy (Kw-hr/st) to reduce the
67
|
Colorado School of Mines
|
particle size from 3837 um to a P of 175 um in comparison with the Santa Teresa sample which
80
particle size was reduced from 3437 um to a P of 175 um. The different mineralogy between
80
both composites resulted to a different resistance to breakage which can be measure by the BWI
testing. According to Table 3.2, the Cobriza sample has a higher composition of silicates than the
Santa Teresa sample that resulted in a greater resistance to breakage.
Grinding both samples to the desired particle size for the flotation studies was very
consistent, in other words, the determined grinding time for each composite was effective to
achieve a P of 175 um for every flotation feed charge. The determination of the BWI for these
80
future ores is important to perform an economic analysis with a correct OPEX estimation that
supports NEMISA to project the grinding cost for these future ores.
4.3 Flotation Studies Results
The simulation of the continuous flotation circuit at NEMISA was performed by using
the same operational parameters such as particle size (P ), flotation reagents, slurry density, and
80
pH. The results from the locked cycle flotation showed considerable recoveries for Cu, Ag and
Au. The recoveries determined by the locked cycle flotation were higher than the ones reported
at the NEMISA’s operation. Approximately, the copper recovered to the concentrate for both
composites was 87.93% for the Cobriza sample and 83.37% for the Santa Teresa sample
respectively. The silver and gold values not recovered and sent to the scavenger tailings during
the locked cycle flotation test for the Cobriza sample were approximately 7% Ag and 25% Au.
For the Santa Teresa sample they were 12% Ag and 11% Au. The current situation at NEMISA
is that the flotation concentrator recovers 85-92% of the contained copper into flotation
concentrates that contain approximately 23-25% Cu. Only 25-40% Au, and 60-65% Ag are
recovered to the copper concentrate.
The values not recovered are reported to the tailing’s storage facilities (TSF). The
possible variations in the results are due to scale difference between industrial and laboratory
scale equipment. Also, human errors during the experimentation can affect the accuracy of the
results.
As mentioned before (page 69) the particle size is a primary parameter to liberate the
valuable minerals in order to separate them from the gangue minerals. The P achieved for the
80
locked cycle flotation test was 175 um which is the operational particle size in the beneficiation
plant at NEMISA. According to the mineralogical analysis performed on both composites, the
68
|
Colorado School of Mines
|
gold and silver-containing minerals identified in each sample have an approximate grain size
distribution P for the Cobriza over 40 um and for the Santa Teresa was 20 um. If the P is finer
80 80
than 175 um the gold and silver bearing minerals would have a better liberation. Hence, the
concentrate will have an increase of gold and silver recovery due to a better separation of the
gold and silver respectively. However, the OPEX for grinding finer it is considerably high for
improving the recovery of gold and silver values, and NEMISA is focused on recovering the
copper as primary metal product.
4.4 Gold Cyanide Leaching Results
As mentioned in the section 3.5.5 (page 74) cyanide leaching was performed on the
flotation tailings (scavenger tailings). The Table 18 showed the elemental analysis of the tailings
to confirm the content of gold in these tailings from both composites. The NaCN concentration
used for the gold cyanide leaching was 0.5 mg/L. Figure 3.42 and Figure 3.43 showed the
correlation between the free cyanide (mg/L) vs the leaching time for three different NaCN
concentrations (0.3, 0.5, 0.75 mg/L). The free cyanide for the Cobriza sample during the leaching
test had a similar tendency for the three curves on Figure 3.42. The NaCN concentrations 0.5 and
0.75 mg/L had similar free cyanide concentrations at 32 hours which means that all the possible
cyanide complexes were formed. The same tendency was seen for the Santa Teresa sample on
Figure 3.43. Hence, the concentration of 0.5 g/L was used as set point for the gold cyanide
leaching test with a retention time of 32 hours.
Table 4.1: Cyanide leaching results for the Cobriza sample with a 0.5 g/L NaCN concentration
Conditions Assay (mg/kg) Distribution %
NaCN concentration
Time (h) Dissolved Oxygen (mg/L) pH NaCN consumption (g) Lime Consumption (g) Au Ag Cu Au Ag Cu
(mg/L)
0.00 5.55 11.61 0.00 0.32 0.00 0.00 0.00 0.00 0.00 0.00
2.00 5.87 11.60 0.47 0.08 0.11 0.45 85.31 12.00 7.20 10.73
6.00 5.75 11.75 0.69 0.04 0.18 0.76 135.00 18.53 12.16 16.98
0.50
17.00 5.34 11.62 0.64 0.09 0.21 1.24 186.80 22.11 19.84 23.49
24.00 4.29 11.70 0.69 0.04 0.22 1.70 190.00 23.16 27.20 23.90
32.00 3.80 11.71 0.15 0.09 0.23 2.10 198.00 24.21 33.60 24.90
Recovery % 81.79 33.33 35.59 Head (Calc) 100.00 100.00 100.00
69
|
Colorado School of Mines
|
CHAPTER 5: ECONOMIC ANALYSIS
This chapter discusses the economic feasibility of the gold cyanide leaching process,
considering both composites and the leaching experimental results previously obtained. NEMISA
already has a current economic system for selling the copper concentrate; for this reason, this
project showed an independent economic analysis that will be complementary to NEMISA’s
current numbers. The economic analysis was performed by using the CostMine economic model
considering the operational details needed for this project. The considered unit operations for this
economic analysis were the cyanide leaching, carbon adsorption /desorption, and SART process
for cyanide recycling.
The tailings produced by the flotation circuit did not require additional grinding prior to
the cyanidation. The economic analysis was based on 10,000 tonnes of tailings per day. NEMISA
reported that the ratio between tailings and concentrate in their current flotation circuit was 27:1,
which means that they are storing around 8,185 tonnes of tailings per day. NEMISA plans in the
future to increase the flotation circuit throughput to 10,000 tonnes of ore per day to supply the
tailings to the leaching plant for processing. The analysis is focused on the recovering of gold,
having some values of silver extracted during the leaching. In addition, the copper will be
recovered in the SART process, and it is expected to be extracted as a by-product copper sulfide.
Table 5.1: Production costs of gold and silver for a 10,000 tonnes/day production rate
Copper Flotation Tailings Dailiy Production Rate Yearly Production Rate
Tailings Processed 10,000 tons of tailings/day 3,600,000 tons of tailings/year
Au Grade (g/ton) 0.20
Ag Grade (g/ton) 1.80
Au Recovery (%) 85.00
Ag Recovery (%) 28.00
Au Price ($/kg) 54,318.00
Ag Price ($/kg) 481.00
Kg of Au 1.70 612.00
Kg of Ag 5.04 1,814
Bond Work Index Cobriza 18.27
Bond Work Index Santa Teresa 17.7
Working hours 24 hours/day
Working days 360 days/year
Schedule 8,640 hours/year
74
|
Colorado School of Mines
|
The estimated operating costs consider the labor, reagents, and electricity for the
operation of the leaching system for 360 days of operation. The economic analysis estimated 41
total personnel for the leaching circuit considering two shifts (12 hours each shift), as shown in
Table 25.
Table 5.2: Operating costs per year for a production rate of 10,000 tonnes per day (+/- 35%)
(Infomine 2017).
75
O p e r a t in g C o s t s
P e r s o n n e l R e q u ir e m e n t s U S D / y e a r
7 6 3 ,7C o n t r o l R o o m O p e r a t o r s ( 4 )
3 6 8 ,3P r o c e s s M a c h in e r y O p e r a t o r s ( 2 )
7 3 4 ,7L e a c h S y s t e m s O p e r a t o r s ( 4 )
1 ,1 0 2 ,6T a ilin g s S y s t e m O p e r a t o r s ( 6 )
R e a g e n t s M ix e r s ( 2 ) 3 6 7 ,2
2 9 4 ,1M a in t e n a n c e W o r k e r s ( 2 )
7 4 8 ,5M e c h a n ic s ( 4 )
7 9 1 ,5E le c t r ic ia n s ( 4 )
1 ,0 1 2 ,4L a b o r e r s ( 6 )
2 6 8 ,0M e t a llu r g is t ( 2 )
1 9 3 ,6P r o c e s s T e c h n ic ia n ( 2 )
In s t r u m e n t T e c h n ic ia n ( 2 ) 9 6 ,3
1 2 1 ,P 7r o c e s s F o r e m a n ( 1 )
T o t a l L 6 ,a 8b 6o 3r , 0C o s t ( U S D )
U S D / y eR e a g e n t s
S o d iu m C y a n id e ( $ 2 .4 5 / k g ) 8 8 2 ,0
4 3 2 ,0A c t iv a t e d C a r b o n ( $ 1 .2 / k g )
5 4 0 ,0L im e ( $ 0 .1 5 / k g )
1 6 2 ,0C a u s t ic S o d a ( $ 0 .4 5 / k g )
9 9 ,0D ia t o m a c e o u s E a r t h ( $ 0 .5 / k g )
1 5 2 ,6S o d iu m H y d r o s u lf id e ( $ 0 .6 9 5 / k g )
4 5 2 ,9S u lf u r ic A c id ( $ 0 .2 7 5 / k g )
1 6 8 ,8C a lc iu m H y d r o x id e ( $ 0 .1 5 0 / k g )
1 4 ,0F u e l O il ( $ 2 .3 5 / lit e r )
1 4 ,0N a t u r a l G a s ( $ 2 .3 4 / G j)
2 ,9 1 7 ,4T o t a l R e a g e n t s C o s t s ( U S D )
U S D / y eE le c t r ic a l P o w e r C o s t ( $ 0 .1 3 3 1 / k W h )
L e a c h in g E n e r g y
4 ,2 0 9 ,4A g it a t e d C y a n id e L e a c h S y s t e m
C a r b o n a d s o r p t io n / D e s o r p t io n P la n t
1 ,7 4 6 ,0A c t iv a t e d C a r b o n P la n t
C y a n id e R e g e n e r a t io n
1 ,6 2 0 ,0S A R T P la n t
7 ,5 7 5 ,4T o t a l E le c t r ic it y C o s t ( U S D / y e a r )
1 7 ,3 5 5 ,9T o t a l O p e r a t in g C o s t s ( U S D / y e a r )
7
24
3
8
0
7
9
3
0
0
0
0
7
a
0
0
0
0
0
2
2
0
9
3
8
a
2
0
0
2
7
6
36
7
6
6
0
7
5
0
0
0
0
5
r
0
0
0
0
0
2
5
4
4
4
0
r
1
0
0
1
6
|
Colorado School of Mines
|
The processing capacity of the leaching plant is 10,000 tonnes of tailings per day, which
gives a total of 3,600,000 tonnes of tailings processed per year. The recovered gold from the
tailings with this production capacity is 612 kg, and the silver recovered is 1.814 tonnes per year.
Table 5.3: Capital Costs for a capacity of 10,000 tonnes per day (+/-35%)
Capital Costs $USD
Agitated Leach Tanks (8) 1,878,400
Cyanide Leaching Equipment 3 0,000,000
Carbon Columns 59,430
Leach Solution Distribution Pumps 37,180
Leach Solution Make-Up Tank 3 15,290
Carbon Stripping Tank 2 ,070
Strip Solution Storage/Mixing Tank 4 ,580
Make-up Water Pump 35,440
Strip Solution Pump 5 ,490
Electrowinning Cell 63,800
Carbon Storage Tank 8 ,240
Strip Solution Heater 87,000
Regeration Kiln 1 02,700
Dore Furnace 29,800
SART Plant 2,600,000
Installation Labor 9,295,700
Concrete 2,934,300
Piping 7,866,600
Structural Steel 2,414,400
Insulation 2 16,000
Instrumentation 1,170,500
Electrical 2,900,400
Coatings and Sealants 4 55,700
Tailings Facility 1,845,900
Engineering and Design 1 4,339,900
Construction Management 1 1,153,300
Contingency 2 5,000,000
Working Capital 4,953,200
Total Capital Cost (USD) 1 19,775,320
Table 5.3 shows the capital costs for the 10,000 tonnes per day plant, which resulted in a
total of $119,775,320 USD. The working capital and contingency are considered as the
maintenance costs for the plant for a year. The engineering and construction of the leaching plant
are included in the total capital costs. The prices of gold and silver used are $54,318 USD/kg and
76
|
Colorado School of Mines
|
$481 USD/kg, respectively. The total annual revenue for a production of 10,000 tonnes per day is
$ $33,486,979 USD, and a profit of $16,131,003 USD per year. The estimated payback period is
8 years, according to the total capital cost and the annual profit. Also, the maintenance costs of the
TSF will gradually decrease due to tailings processing. The Cobriza and Santa Teresa ores are
estimated to have a mine life of 20 years, so that the IRR and NPV are estimated for this mine life
period.
Figure 5.1: NPV Sensivity Analysis
Figure 5.1 shows the sensitivity analysis of variables that affect the Net Present Value
(NPV) at a discount rate of 10.00 % during the 20 year period during the project, these variables
are capital costs (CAPEX), operating costs (OPEX), revenue and electricity costs. The analysis
shows that the increasing revenue has a positive NPV, on the other hand the decreasing CAPEX
and OPEX, and electricity costs have a positive tendency in the NPV analysis. The NPV of the
project will be affected by all these variables, being the NPV most responsive to capital cost and
revenue variations. The discount rates modify the profitability of the project, at lower discount
rates the project will become more profitable. Considering a 20 year of mine lifetime period for
the project the estimated IRR is 13% for this project.
77
|
Colorado School of Mines
|
CHAPTER 6: CONCLUSIONS AND FUTURE RESEARCH
The main objective of this research was to assess the gold recovery from copper flotation
tailings to be produced from future ores. The mineral processing experiments were performed at
laboratory scale to simulate the beneficiation process of the NEMISA operations. According to
the literature review, different techniques were used to simulate the processing of two different
samples (Cobriza and Santa Teresa) provided by the geology department at NEMISA. The
grindability studies were performed by using the Bond Work Index technique to measure the
necessary energy to achieve a P of 175 um size prior to flotation.
80
The BWI determined for both future ore composites were similar likely between them
due to having similar mineralogical compositions. Both ore composites had a higher BWI than
the current ore reported by NEMISA. The energy requirements for these future ores composites
will be higher for liberating valuable minerals. Froth flotation was performed on the future ore
composites to beneficate the copper, silver, and gold from the gangue minerals and produce the
tailings for the cyanide leaching. The froth flotation results showed that copper was the most
readly recovered metal followed by the silver and gold. Approximately 10% -25% of the gold
reported to the froth flotation tailings at gold concentration of 0.55 mg/kg for the Cobriza
composite and 0.24 mg/kg for the Santa Teresa composite.
These results confirmed that the gold concentrations expected from future ores were
considerable and likely worthy of cyanide leaching to recover additional gold values. The
cyanide leaching experiments showed that 82.00% and 87.5% of the gold contained in the
flotation tailings could be recovered from the Cobriza and Santa Teresa composites, respectively.
The consumption of reagents such as cyanide and lime were economically acceptable for
extracting the gold from the flotation tailings. The recovery of copper and silver was low in
comparison with the gold recovery for both composites, which means they will be lost in the
raffinate (cyanide).
The economic analysis showed a considerable profit of $16,131,003 USD and a payback
period approximately of 8 years. This analysis showed a positive Net Present Value (NPV) at a
discount rate of 10.00% that confirmed the profitability of the project for the 20 years of the
projected ore lifetime. As mentioned in the literature survey, the SART process is a suitable
procedure to remove the sulfides and copper in order to clean and recover the cyanide for
78
|
Colorado School of Mines
|
References
A. Gupta, D. Yan. 2016. "Size Reduction and Energy Requirement." In Mineral Processing
Design and Operations, 71-121.
Abdul Mwanga, Jan Rosenkranz, Pertti Lamberg. 2015. "Testing of Ore Comminution Behavior
in the Geometallurgical Context—A Review." Minerals 276-297.
Adams, Mike D. 2005. Advantages in Gold Ore Processing. Perth, Australia: Elsevier B.V.
Adkins, S J, and M J Pearse. 1992. "The influences of collector chemistry on kinetics and
selectivity in base-metal sulphide flotation." In Mineral Engineering, 295-310.
Agar, G.E. 2000. "Calculation of Locked Cycle Flotation Test Results ." Elsevier 1533-1542.
Amelunxen, Peter. 2003. The application of the SAG power index to ore body hardness
characterization for the design and optimization of autogenous grinding circuits.
Montreal : McGill University.
Aylmore, M G. 2016. "Alternative Lixiviant to Cyanide for Leaching Gold Ores." In Gold Ore
Processing: Project Development and Operations, by M D Adams, 447-460.
Crundwell, Frank K, Michael S Moats, Venkoba Ramachandran, Timothy G Robinson, and
William G Davenport. 2011. "Production of Nickel Concentrate from Ground Sulfide
Ore." In Extractive Metallurgy of Nickel, Cobalt and Platinum Group Metals, by Frank K
Crundwell, Michael S Moats, Venkoba Ramachandran, Timothy G Robinson and
William G Davenport, 21-37. ELSEVIER.
Dejan T, Maja T, Ljubisa A, Vladan M, Milan T. 2017. "A Quick Method For Bond Work Index
Approximate Value Determination." In Physicochemical Problems of Minerals
Processing, 321-332.
Deschenes, G. 2005. "Advances in the cyanidation of gold." In Development in Mineral
Processing , by Mike D Adams, 479-500. Elsevier.
Dominy, S.C. 2010. "Grab Sampling for Underground Gold Mine Grade Control ." The Journal
of The Southern African Institute of Mining and Metallurgy 1-11.
Estay , Humberto, and Pablo Carvajal. 2012. "The SART process experience in the Gedabel
plant." Hydroprocess 2012 1-10.
80
|
Colorado School of Mines
|
ABSTRACT
Flotation of a bastnäsite containing ore with novel collectors was investigated using locked
cycle froth flotation. Previous research into the flotation of bastnäsite ore simply considered single
stage flotation and while this is a good place to start the testing of novel collectors, it is not
representative of the conditions employed in a full-scale flotation plant. This study investigates
both salicylhydroxamic acid (collector 2) and n,2-hydrocyclohexanecarboxamide (collector 5) in
locked cycle flotation and contact angle studies and includes a comparative economic assessment
comparing the two novel collectors listed above to the fatty acid flotation previously used at
Mountain Pass in California.
Locked cycle flotation allows for the simulation of continuous flotation processes using
bench scale flotation equipment. This allows the collectors to be tested in conditions that more
closely match the conditions that would be found in flotation plants in industry. Locked cycle
flotation with collector 2 returned rare earth oxide grades between 58.5% and 66.9% and recovery
between 42.8% and 74.7% while rejecting 78% of the calcite. Locked cycle flotation with collector
5 returned rare earth oxide grades between 13.2% and 13.8% with recoveries between 26.6% and
41.3% while rejecting 9% of the calcite. The rejection of calcite is an important consideration
because it affects the downstream reagent consumption in the leaching step of the rare earth
element processing. Locked cycle flotation showed a large disparity in performance between
collector 2 and collector 5. This disparity was investigated using contact angle studies. Performing
contact angle test work allows for comparisons to be made regarding the applied hydrophobicity
of a collector to the surface of a mineral.
iii
|
Colorado School of Mines
|
CHAPTER 1: INTRODUCTION
1.1 Background and Motivation
Rare earth elements (REE) comprise of fifteen lanthanides with atomic numbers 57 through
71 (lanthanum to lutetium) as well as two group IIIb elements (scandium and yttrium) on the
periodic table. These elements are considered rare because of the challengers associated with both
the discovery of concentrated deposits and the difficulty in selectively separating them from each
other. The general crustal abundance of REE is greater than that of silver and similar to that of
copper and lead [1]. REE are steadily growing in importance because of their heavy involvement
in technological and industrial applications, particularly in areas of green technology and
renewable energy.
REE are currently considered a critical material in the United States. The United States
Department of Energy (DOE) defines criticality in two ways: (A) supply based risk based on
projected market balances, competing energy demands, political, regulatory and social factors, co-
production risks, and producer diversity; and, (B) importance to clean energy based on clean
energy demand and substitutability [2]. At the time of writing, the United States has a 100% import
reliance for rare-earth compounds and metals with the following distribution: China, 80%; Estonia,
6%; France and Japan, 3% each; and other, 8% [3]. Table 1.1 shows the world mine production
and reserves for REE.
1
|
Colorado School of Mines
|
Table 1.1 - World mine production and reserves for REE
From Table 1.1, China is the world’s largest producer of REE by nearly an order of
magnitude. Table 1.1 also shows that the United States has started producing REE once again in
2018. This is because of the reopening of the Mountain Pass Mine in California. The United States
currently produces concentrates enriched in REE and does not possess any REE concentrate
refining capabilities. More information on Mountain Pass can be found in Chapter 2.
1.2 Objectives
Mountain Pass employs froth flotation in order to upgrade the bastnäsite ore from a run of
mine grade of roughly 8% rare earth oxide (REO) to a concentrate with roughly 60% REO. The
major issue with the plant currently is that it has a low recovery of only 60-70%. This means that
as much as 40% of the REO that enter the plant do not get recovered and are sent to the tailings
together with other gangue materials such as calcite and barite. One of the primary objectives of
this research is to investigate the effectiveness of novel collectors, designed by ORNL for
bastnäsite flotation, in a plant simulation flotation flowsheet.
2
|
Colorado School of Mines
|
Although several collectors have been identified as likely functional candidates using
single stage flotation at a micro and bench scale, they have not been tested in a simulated
continuous fashion. It is possible that some collectors might display improved functionality when
used in a multi-stage process. The best method for performing flowsheet simulations at the
laboratory scale is with locked cycle flotation. The method of locked cycle flotation will be further
discussed in Chapter 2.
The other primary objective of this research is to determine the mechanism of adsorption
of these collectors to the bastnäsite surface and the surfaces of prominent gangue minerals such as
calcite. It is known that the collectors that are studied will provide a separation of valuable
bastnäsite from the gangue minerals in the ore but the mechanism through which the collectors
bind to the surface of these minerals is not known. Using contact angle studies, it is possible to
determine how hydrophobic a collector makes the surface of a mineral being studied. This
information can be useful for determining why some collectors perform more strongly than others.
1.3 Funding Source
Funding for this project was provided by the Critical Materials Institute (CMI). The CMI
is a Department of Energy (DOE) research hub that is focused on technologies that make better
use of materials and eliminate the need for materials that are subject to supply disruptions. CMI
helps to assure the supply chains of critical materials through three methods: (a) diversifying and
expanding the availability of these materials throughout their supply chains, (b) reducing wastes
by increasing the efficiency of manufacturing and recycling, and (c) to reduce demand by
identifying substitutes for critical materials. [4]
3
|
Colorado School of Mines
|
CHAPTER 2: LITERATURE REVIEW
2.1 Rare Earth Elements
Rare earth elements (REE) comprise of fifteen lanthanides with atomic numbers 57 through
71 (lanthanum to lutetium) as well as two group IIIb elements (scandium and yttrium) on the
periodic table. These elements may be classified as either light or heavy rare earth elements. Light
REE (LREE) include (La), cerium (Ce), praseodymium (Pr), neodymium (Nd), and samarium
(Sm) and are more common and relatively easier to extract than their heavy counterparts. The
heavy REE (HREE), which include yttrium and elements with atomic numbers from 64 to 71
[europium (Eu), gadolinium (Gd), terbium (Tb), dysprosium (Dy), erbium (Er), thulium (Th),
ytterbium (Yb), and lutetium (Lu)] are less abundant but highly critical in demand and supply
[1][5]. Yttrium is considered to be a HREE because its chemical behavior during separation closely
mimics that of Holmium.
2.1.1 History of the Rare Earths
The discovery of the 17 elements now known as the rare earths began in 1787 and
continued for approximately 160 years until its conclusion in the 1940s. The activity started in
Ytterby, a small village near Stockholm in Sweden. A lieutenant of the Swedish Royal Army, Carl
Axel Arrhenius, found a black mineral in 1787. This mineral had not previously been mentioned
in literature at the time. It was not until 1794 that the mineral was analyzed by the Finnish chemist
Johan Gadolin. Gadolin found that the mineral contained iron and silicate as well as approximately
30% of a material he called a “new earth” (elements were called “earths” until the first decade of
the 19th century). The discovery was confirmed the following year by Swedish chemist Anders
Gustaf Ekeberg. Ekeberg decided to name the “new earth” yttria to honor the town in which it was
4
|
Colorado School of Mines
|
first discovered. He also named the associated mineral gadolinite [6]. The chemical formula for
gadolinite is Y Fe2+Be Si O [7].
2 2 2 10
In 1804, two independent researchers in Sweden and Germany simultaneously reported the
discovery of a new element in a new mineral found in the Bastnäsgrube mine close to
Rydderhyatten, Sweden.The Swedish researchers determined the new earth was an oxide of a new
element. They named the element cerium after the asteroid Ceres that had been discovered only
three years earlier in 1801. The mineral that contained the cerium was named cerite. [6]
Carl Gustaf Mosander, an associate of the researcher that discovered cerium, determined
that both ceria and yttria were complex in nature and contained new elements. The name
lanthanum was suggested for the new element. In Greek, lanthano means “to escape notice”.
Despite this discovery, Mosander believed that the new lanthanum he separated was not pure but
might contain another new element. In 1842, he succeeded in proving his theory by detecting a
new element which he names didymium. The name came from the Greek word didymos, meaning
twins, to acknowledge that it accompanied both cerium and lanthanum in the cerium mineral. After
this discovery, Mosander turned his attention to the mineral gadolinite believing that it contained
additional new elements. In 1843, he reported two additional elements and named them erbium
and terbium. [6]
A Swiss-American chemist, Marc Delafontaine, reported in 1878 that the absorption
spectrum of didymium separated from the mineral samarskite was not fully like the absorption
spectrum of didymium from cerite. He theorized that this meant that didymium was not a single
element. In 1879, the French chemist Paul Emile Lecoq de Boisbaudran disproved Delafontaine’s
report on the spectra but did find a new element in samarskite. He named the new element
samarium after the mineral samarskite in which it was discovered. [6]
5
|
Colorado School of Mines
|
While investigating the erbium fraction of gadolinite, the Swiss chemist Jean Charles
Marignac separated an oxide and salts that were different from erbium in both spectral and
chemical properties. In 1878 he named this new element ytterbium because it stood between
yttrium and erbium in its properties.
In 1879, Lars Frederick Nilson also investigated erbium to confirm the existence of
ytterbium. He confirmed the existence of ytterbium but also discovered a new element. He named
this new element scandium after Scandinavia.
Another Swedish chemist by the name Per Theodor Cleve postulated that there could be
even more elements in the erbium fraction remaining after the separation of ytterbium. He
identified the existence of thulium and holmium in 1879. The name thulium comes from the word
thule, the old name of Scandinavia [8]. Holmium was named after Stockholm, Sweden [6]. In 1886,
Boisbaudran determined that the holmium discovered by Cleve contained another element. He
named this new element dysprosium.
In 1885, Austrian chemist, Carl Auer von Welsbach, began investigations into didymium.
At that point it was widely suggested that didymium might contain multiple elements. Auer applied
fractional crystallization rather than the previously used fractional precipitation to separate
didymium. In 1886, he succeeded in separating two fractions of didymium ammonium nitrated.
He concluded these two fractions belong to different elements. He named them praseodymium and
neodymium. [6]
French chemist Eugene Demarcay announced he separated a new element from samarium
in 1901. He named this new element europium. In 1904, europium was confirmed after French
chemist Georges Urbain separated it from gadolinium. [6]
6
|
Colorado School of Mines
|
In 1905, Auer mentioned that ytterbium likely contained new elements. Two years later he
published his findings that ytterbium consisted of two elements. He named them aldebaranium
and cassiopeium. At nearly the same time, Urbain also determined that ytterbium contained two
elements. He named these two elements neoytterbium and lutetium. Over the course of time,
ytterbium (for neoytterbium) and lutetium survived. The name lutetium was derived from the
ancient Roman name for Paris. [6]
Although all the naturally occurring rare earth had been discovered at this point, the
scientists at the time did not notice this until much later. More and more discoveries of “new
elements” were published. In 1912, the idea of atomic numbers was introduced by van den Broek.
Researchers determined that there must be elements between the atomic numbers of 57-71. All
elements except number 61 had been discovered. As a result of this, many of the reported “new
elements” were disproven. In 1945, researchers at what is now Oak Ridge National Laboratory
provided chemical proof of the element with atomic number 61. It was produced using ion-
exchange chromatography to obtain the element from the products of fission of uranium and of
neutron bombardment of neodymium. This final element was named promethium after
Prometheus. Promethium is not a naturally occurring element. [6]
Table 2.1 shows the chronological discovery of the rare earth elements.
7
|
Colorado School of Mines
|
2.1.2 Applications
REEs have many applications ranging from use in catalysts to use as a polishing agent.
Figure 2.1 below shows a breakdown of REEs by end use. This figure was produced using data
from the annual mineral commodities summary produced by the United States Geologic Survey
(USGS) [9]. The largest use of REEs is in the form of catalysts. REE catalysts that are commonly
used include catalysts for fluid cracking (72%) and automobile catalytic converters (28%), of
which lanthanum oxide and cerium oxide contributed the largest portions [10]. Figure 2.1 shows
the primary categories in which REE are used and Table 2.2 shows the specific applications in
which individual REEs are applied.
2.1.3 Sources
REEs are currently produced all around the world with the largest producer being China.
Until 2016, the United States did not have any domestic production of REEs, but with the
reopening of Mountain Pass Mine in California, the United States has once again begun to produce
REE concentrates. Figure 2.2 shows global REE production by country as a percentage of the total
global production. This figure was also produced using data from the annual mineral commodities
summary produced by the USGS. [9]
The global production of REEs as of 2019 was approximately 210,000 tons. This is the
largest single year of REE production since the USGS started keeping the yearly production values
in 1995. The increasing production of REE shows that the demand for the materials is also on the
increase as more technology begins to use REEs. Figure 2.3 shows the annual global production
in tons of REEs from 1995 to 2019.
9
|
Colorado School of Mines
|
2.2 Rare Earth Minerals
Rare earths do not naturally occur as individual elements or as individual rare earth
compounds, but rather as mixtures in rock formations including carbonates, halides, silicates and
others. More than 200 distinct species of rare earth minerals have been discovered and described.
Approximately half of these have had their crystal structures reported. These discovered mineral
species have been grouped into conventional chemical groups such as halides, carbonates, borates,
etc.
Table 2.5 shows the classification of rare earth minerals by their standard chemical bases.
Of all the minerals which exist, there are only three that are widely used for REE production:
bastnäsite, xenotime, and monazite.
2.2.1 Bastnäsite
Bastnäsite (also spelled Bastnasite or Bastnaesite) is a REE bearing fluorocarbonate
mineral ((Ce, La, Pr)(CO )F),and is the primary source of light rare earth elements (Figure 2.5,
3
[24]). It is closely related to the mineral Parasite ((Ca(Ce,La) (CO ) F ) and Synchysite
2 3 3 2
(CaCe(CO ) F). Bastnäsite was named after the Bastnas Mine, in the Raddarhyttan district in
3 2
Vastermanland, Sweden. It features a hexagonal – ditrigonal dipyramidal crystal system and an
average density of 4.97 g/cm3. Bastnäsite is found in vein deposits, contact metamorphic zones,
and pegmatites and can occur either as veins or disseminated in carbonate/silicate matrix. The
REO content of bastnäsite at Mountain Pass is approximately 75% and contains primarily light
rare earths. The two largest operating Bastnäsite mines are the Mountain Pass mine in California
and the Bayan Obo mine in China [25]. The rare earths produced at Bayan Obo are actually
produced from the tailings of the iron ore processing at that mine because Bayan Obo is also
China’s largest iron mine [25]. Some key properties of cerium Bastnäsite are listed in Table 2.6.
14
|
Colorado School of Mines
|
2.3.2 Mountain Pass
The Mountain Pass deposit sits close to the eastern edge of the Mojave Desert in California
(Figure 2.7). It lies just north of Interstate 15 near Mountain Pass and is about 60 miles southwest
of Las Vegas, Nevada. The Mountain Pass deposit is recognized as the largest known REE deposit
in the United States with current reserves estimated to be greater than 20 million metric tons at a
rare earth oxide grade of approximately 8.9%. [25]
The core of the Mountain Pass igneous complex is made up of a massive carbonatite known
as the Sulphide Queen body. The Sulphide Queen body hosts the bulk of the REE mineral
resources in the district. The carbonatite body has an overall length of 730 m (2,395 ft) and has an
average width of 120 m (394 ft). The typical ore in this body contains 10-15% Bastnäsite (the ore
mineral), 20-25% barite, and 65% calcite or dolomite (or both), plus other minor accessory
minerals [25]. This work is focused on improving the flotation efficiency for Bastnäsite using
novel collectors designed by Oak Ridge National Laboratory and seeks to determine the impact
that these collectors have on the hydrophobicity of mineral surfaces.
Figure 2.7 - Northwest facing view of Mountain Pass district, California, 1997 [25]
18
|
Colorado School of Mines
|
2.4 Froth Flotation
Froth flotation is one of the most widely employed methods of mineral beneficiation in the
industry. Flotation seeks to take advantage of the hydrophobicity (natural or induced) of mineral
surfaces to use gases to separate the valuable and desired minerals from the waste minerals. The
first U.S. patent describing the use of air bubbles to separate and concentrate minerals was awarded
in 1905 to Henry Sulman and Hugh Picard (Patent No. 793,808 [29]). The year 1905 was also the
year in which the Potter process was first introduced to flotation in the mineral industry. The
production of sphalerite at Broken Hill in Australia was the first major commercial application for
froth flotation [30]. After that initial application, froth flotation quickly spread the United States
and around the world and is an essential separation method for the beneficiation of minerals and
coal. The applications of froth flotation continue to be expanded to new industries such as
environmental controls, bitumen extraction from tar sands, and recycling.
2.4.1 Froth Flotation Reagents
Froth flotation is a complex process which requires a wide range of reagents to control the
hydrophobicity of the surfaces of the minerals of interest. Flotation reagents can be broken down
into three main categories: collector, frother, and modifier. Figure 2.8 shows the flotation reagent
triangle. The flotation process is not complete without a careful combination of all three parts.
19
|
Colorado School of Mines
|
Figure 2.8 - Flotation reagent triangle [30]
Collectors are a large group of organic chemical compounds that all differ in chemical
composition and function. The general purpose of a collector is to selectively form a hydrophobic
layer on a given mineral surface and therefore provide conditions for that mineral to be recovered
in the froth product through attachment to an air bubble. According to the ability of collectors to
dissociate in water, they can be divided into distinct groups. [31]
Ionizing collectors are heteropolar organic molecules. Depending on the resulting charge
the collector assumes the character of a cation or anion. Anionic collectors can be further divided
into oxyhydryl and sulfyhydryl collectors based on their solidophilic properties. Cationic
collectors are compounds in which the hydrocarbon radical is protonated. These reagents are
amines from which the primary amines are the most important flotation collectors. The other main
group of collectors is non-ionizing collectors. These are also divided into two groups; the first
group is reagents containing bivalent sulfur and the other contains non-polar hydrocarbon oils. The
collectors used in this research can be found in Chapter 3. [32]
Frothers make up the next major component of froth flotation reagents. They are
compounds that lower the surface tension of water and increase the strength of the bubble films,
therefore facilitating froth formation. The surface tension of the slurry also affects the bubble size
which can affect the recovery of the flotation cell. Frothers come in many different types and are
20
|
Colorado School of Mines
|
classified based on their properties and behaviors in solution. Table 2.9 shows the classification of
frothers. [33]
Table 2.9 - Classification of frothers [33]
Acidic Neutral Basic
Phenols Aliphatic alcohols Pyridine base
Alky sulfates Cyclic alcohols
Alkoxy paraffins
Polypropylene glycol ethers
Polyglycol ethers
Polyglycol glycerol ethers
The final category of flotation reagents is the category of modifier. This is a broad category
of reagents that can serve purposes ranging from pH control to enhancing the selectivity of
collectors. They can be regarded as the most important chemicals in froth flotation because they
control the interaction between individual minerals and collectors. Collectors can be very sensitive
to the conditions of the pulp such as pH and water purity. Modifiers allow this important factor to
be controlled to maximize efficiency of the flotation. Modifying reagents can also be used to
depress or activate certain minerals. Without modifying reagents, it would not be possible to isolate
individual mineral sulfides of lead, zinc, and copper from complex sulfide ores. It is difficult to
classify these reagents into specific groups because their effects can be so varied under different
operating conditions. [34]
2.4.2 Bench Flotation
Bench flotation is the stage of flotation that is usually performed after the initial testing
with microflotation is completed. Microflotation allows for quick scoping tests to be performed
because it requires less collectors and materials than bench flotation. Bench flotation is so named
because a bench top flotation cell is typically used. For this research, the Metso Denver D-12
21
|
Colorado School of Mines
|
Legacy Cell was used. More information about this cell can be found in Chapter 3. Bench flotation
provides a large amount of flexibility for initial scale up testing of different collector and reagent
schemes. After initial performance checks are completed with microflotation, bench flotation
allows testing at an increased scale. Bench flotation equipment such as the Metso Denver D-12
Legacy allow different sizes of flotation cell to be used allowing for the testing of various stages
of flotation. Through locked cycle flotation, bench scale flotation can also be used to simulate a
continuous flotation flowsheet.
2.4.3 Locked-Cycle Flotation
An experimental simulation of a continuous circuit that utilizes repetitive batch tests in a
cyclical manner is universally referred to in the minerals industry as a locked cycle test (LCT). An
LCT is usually begun by doing a complete batch test in the first cycle and then adding the materials
from the first cycle to the appropriate location in the second cycle. Batch tests are continued in this
fashion for an arbitrary number of cycles. The number of cycles must be enough to allow the test
to come to a steady state. A test is determined to be at steady state when the products from each
cycle match that of the previous cycle. A test is cannot be determined to be at steady state until
after the products are weighed and assayed so an arbitrary number of cycles between 6 and 8 is
usually used. The terminal products from each cycle and the intermediate products from the last
cycle are weighed and then subjected to chemical analysis. Usually only one concentrate and one
tailing are collected from each cycle because all the other intermediate materials are passed to the
next cycle. Locked cycle tests are not used only in flotation studies but also in grindability
determinations, balling, and leaching tests [35]. The locked cycle testing flowsheet that was used
for this research is shown in Figure 2.9. It is important to note that this figure only shows the first
three cycles. Additional cycles can be performed in the same manner as previous cycles.
22
|
Colorado School of Mines
|
has strengths and weaknesses in calculation of results and their accuracy usually depends on the
mass conservation of the test.
The n-product formula method is a simple material balance technique that utilizes the
assays from the final products to determine the mass balance. For simple ores with only one
concentrate and tailing, this procedure uses the assay of the feed, concentrate, and tailings using
the familiar formula shown in Equation 2.1. The remainder of the balance is calculated once C (the
concentrate mass) is determined. For locked cycle test balancing, the weighted average assay for
the final 2-4 cycles is used. An important requirement for the n-product formula method is that the
circuit must have mass conservation. If mass conservation is not maintained, then the n-product
formula should not be used as it may return erroneous results.
(2.1)
𝑓 −𝑡
𝐶 = 𝐹 ∗
The SME procedure for calculating the resu𝑐lt−s o𝑡f a locked cycle test is described within the
SME mineral processing handbook [37]. In cases such as the one covered in this research, ores
with a single concentrate and tailing, the concentrate is projected as the average mass and assay of
the concentrate produced in the last several cycles of the test. The tailings are then projected using
a similar method. The feed for the test is then calculated as a sum of those products. This procedure
works well if the test has come to a steady state. If the test has not come to steady state, this
procedure can also produce erroneous results.
The final method for calculating the results of locked cycle tests is the concentrate
production balance method. This procedure is an offset from the SME procedure where the
concentrated is projected in the same way as the SME procedure. The tailings are then calculated
as the difference between the feed and the concentrate. The primary advantage of this method is
that this procedure does not overstate the metallurgy if the test does not have any mass
24
|
Colorado School of Mines
|
conservation. The theory behind the procedure is that the concentrate produced is all the
concentrate produced. Therefore, all other materials must be tailings. This procedure resembles
month end production balances at an operating plant in many ways.
2.5 Bastnäsite Flotation Review
The flotation of Bastnäsite has been receiving increased attention in the last several years.
Most of the studies, however, are focused on the fundamental aspects of the flotation of Bastnäsite
ores. Very few studies exist which focus on flotation at the bench scale or larger. This section will
detail some of the recent fundamental studies which have been performed on the flotation of
Bastnäsite as well as previous studies into the performance of Bastnäsite flotation at the bench
scale.
Previous fundamental research into the flotation of Bastnäsite is shown in Table 2.10.
Research has been primarily focused on the fundamental studies of new collectors and reagents to
increase and optimize the grade and recovery of Bastnäsite from ore. Reducing the amount of
calcite in the flotation products also is an important consideration in flotation research. Reducing
the amount of calcite in the flotation concentrate reduces the consumption of acid in downstream
leaching and purification stages.
2.5.1 Mountain Pass, CA Flotation Process
The flotation process for Bastnäsite at the Mountain Pass mine in California starts with a
two-step comminution process including crushing and grinding. The ore is ground and classified
to a P80 of 325 US Mesh (45 µm) before it is introduced to the flotation circuit. The flotation
circuit, as of 2014, uses elevated temperature fatty acid flotation to recover the Bastnäsite. The
elevated temperature allowed for a higher level of recovery and separation. A schematic of the
flotation process is shown in Figure 2.10. [38]
25
|
Colorado School of Mines
|
Upon entering the flotation circuit, the ore is subjected to four stages of high temperature
(82°C) conditioning as listed below [38]:
1. Soda ash (5lbs/ton) is used to adjust the pH to 9. Soda ash can also act as a gangue
(barite, calcite) depressant
2. Blank stage to allow for further pH adjustment and depressing effect
3. Lignin sulfonate (5 lbs/ton) is added to further depress present gangue minerals
4. Fatty acid solution (0.14 lbs/ton) is added as the collector
After conditioning, the pulp is transferred to the 3 stages (2 tanks/stage) of rougher flotation
at 40% solids. The tailings from this stage are sent to 2 stages (3 tanks/stage) of rougher scavenger
flotation. The rougher concentrate is then moved to a cleaner conditioner before going through 4
stages of cleaner flotation (multiple tanks/stage). This flowsheet produces a concentrate containing
60-70% REO with recoveries ranging from 60-70%. Therefore, the flowsheet leaves significant
room for improvement in REO recoveries.
Figure 2.10 - MolyCorp Mountain Pass REO Flotation Flowsheet 2014 [38]
27
|
Colorado School of Mines
|
2.5.2 Hydroxamate Collectors
Recently, hydroxamate collectors have been receiving the bulk of the attention for the
flotation of rare earth minerals, specifically Bastnäsite. This is due primarily due to their supposed
increased selectivity for rare earths when compared to fatty acid flotation. The increased selectivity
is due to their ability to preferentially form chelates with rare earth metal ions over the alkaline
ions in gangue materials. There is a thermodynamic driving force for the adsorption of
hydroxamates onto rare earths rather than their associated gangue. An example of this driving force
is shown in Table 2.11 by the relatively larger stability constants for the formation of rare earth
hydroxamates compared to the formation of calcium hydroxamate.
Table 2.11 - Stability constants for the formation of metal hydroxamates [46]
Cation Log K Log K Log K
1 2 3
H+ 9.35 -- --
Ca2+ 2.4 -- --
Fe2+ 4.8 3.7 --
La3+ 5.16 4.17 2.55
Ce3+ 5.45 4.34 3.0
Sm3+ 5.96 4.77 3.68
Gd3+ 6.1 4.76 3.07
Dy3+ 6.52 5.39 4.04
Yb3+ 6.61 5.59 4.29
Al3+ 7.95 7.34 6.18
Fe3+ 11.42 0.68 7.23
Like fatty acids, the mechanism of hydroxamate adsorption is chemical in nature because
it adsorbs at a zeta potential where both the hydroxamate species and rare earth mineral are
negative. The free energy of adsorption also becomes more negative as the temperature increases.
This is shown in Figure 2.11. The free energy of adsorption of hydroxamate onto Bastnäsite is
28
|
Colorado School of Mines
|
the surface of bastnäsite and tends to form a multilayer at higher surface coverages while binding
perpendicularly to the surface of calcite. It is possible that this difference contributes to the strong
selectivity of SHA for bastnäsite.
2.6 Contact Angle
The idea of wettability was first described by Thomas Young in 1805 [49]. Wetting of a
solid surface can be quantitatively described from the profile of a liquid droplet and more
specifically from the tangential angle of the liquid-gas-solid interface. The angle of contact, ϴ
determined from this image is known as the Young’s angle or static contact angle. This angle is a
result of the equilibrium between three surface tensions, the liquid surface tension (γ ), the solid
LV
surface tension (γ ), and the liquid-solid interfacial surface tension (γ ) and is expressed as the
SV SL
Young’s equation (Equation 2.2). Figure 2.12 shows a graphic of the components included in
Young’s equation.
(2.2)
𝛾𝑆𝑉 = 𝛾𝐿𝑉 ⋅𝑐𝑜𝑠𝜃 +𝛾𝑆𝐿
Figure 2.12 - Graphic vector representation of parameters in a sessile drop [50]
The measurement of contact angles is a very sensitive measurement. In order to make the
best possible measurement, the surface of the solid should be polished until smooth and the surface
and micro syringe should be cleaned from dirt. In most static contact angle measurements, the
30
|
Colorado School of Mines
|
sessile drop is formed by dispensing the test liquid through a micro syringe onto a horizontal and
smooth test surface. The name sessile drop comes from the word “sessile” meaning attached
directly by a broad base [51]. Figure 2.13 shows a schematic for the formation of a sessile droplet
during a static contact angle measurement.
Figure 2.13 - A schematic of the formation of a sessile drop during static contact angle
measurements [50]
2.7 Simulation
Flotation equipment has been designed in many different sizes and configurations with
flowsheets including varying functions. Flotation equipment can serve as cleaners, scavengers, or
roughers at various positions in flotation plants. Although flotation is widely used in the mineral
processing industry, there is a lack of in-depth knowledge of the principles due to the complicated
nature of three phase flow, the motion of bubbles and particles, and the different surface
interactions involved in the process. The application of computationally assisted modelling in the
mineral industry started in the early 1990s [52]. Since then modelling techniques have rapidly
advanced to provide better understanding into the fluid flows and particle interactions present in
the froth flotation of minerals.
31
|
Colorado School of Mines
|
2.7.1 Principles of Flotation
Much like chemical reactions which involve collisions amongst molecules, froth flotation
is often described as a first order process relating the rate of flotation to the particle concentration
[52]. Following this comparison, the rate of particle flotation can be described as shown in
Equation 2.3, where N is the number of free particles in the pulp phase and the rate constant, k,
p
represents the rate of removal of particles from the pulp.
(2.3)
ⅆ𝑁𝑃
= −𝑘𝑁𝑝
It should also be noted that the firsⅆt-𝑡order rate equation is only valid for a batch process
and does not account for inflows and outflows that would be commonly found in a continuous
flotation system. Integrating Equation 2.3, the recovery defined as the ratio of the number of
particles remaining to the number of initial particles can be expressed as:
(2.4)
−𝑘𝑡
The problem with this equation wh𝑅en= a1pp−lieⅇd to a flotation cell is that the rate constant, k,
does not stay constant during the process. This can be accounted for by expressing the rate constant
as a function of physical parameters of the system. The flotation response can be split into two
primary parts, the ore characteristics and the cell characteristics. Cell characteristics can further be
broken into two parts, the hydrodynamic characteristic and the froth characteristic. The simulation
software used in this research, JKSimFloat, uses the following basic form of the flotation mode to
account for the rate constant:
(2.5)
𝑘 = 𝑃 ⋅𝑆𝑏 ⋅𝑅𝑓
(2.6)
6×𝐽𝑔
𝑆𝑏 =
ⅆ𝑏
32
|
Colorado School of Mines
|
CHAPTER 3: METHODS AND EQUIPMENT
3.1 Minerals
Three pure mineral samples were obtained for contact angle studies. These included
Calcite, Barite, and Bastnäsite crystals. The Calcite materials were sourced from Ward Scientific
and were a product of Mexico. The Barite crystals also were sourced from Ward Scientific. The
pure Bastnäsite crystals originate from Khyber Pass in Pakistan. They were provided to the
Colorado School of Mines by Oak Ridge National Laboratory.
3.2 Bastnäsite Ore
A Bastnäsite ore sample was obtained from Mountain Pass Mine in California. The ore was
jaw crushed, roll crushed, and then wet ground in a rod mill to produce the proper size for flotation
experiments. The goal for the grind size of the Bastnäsite ore for flotation was a P of 45 µm. This
80
matches the grind size used by the Mountain Pass Mine and provides an optimal liberation to
maximize flotation grade and recovery.
To properly calibrate the wet rod mill grind time, a grind curve was created. Various grind
times were selected, and the products were then wet sieved at 325 mesh (44 µm). These sieved
products were then dried, and their masses were measured to determine the percent passing. Table
3.1 shows the data used to generate the grind curve shown in Figure 3.1. The selected grind time
from the grind curve to create a P of 45 µm was 110 minutes.
80
34
|
Colorado School of Mines
|
Table 3.4 - Reagent type, Name, and Chemical Formula
Type Name Chemical Formula
Frother Aerofroth 70
pH Modifier Soda Ash Na CO
2 3
pH Modifier Hydrochloric Acid HCl
Ore Bastnäsite (Ce, La)(CO )F
3
Flux Lithium Borates 66.67% Li B O , 32.83% LiBO , 0.5% LiBr
2 4 7 2
3.4 Froth Flotation
Froth flotation using Bastnäsite ore was performed using two Metso Denver D-12 Legacy
Cells. Figure 3.2 shows an image of the Metso Denver Cell that was used in these flotation
experiments. The ore was ground to a P of 45 µm using the preparation method described in the
80
Bastnäsite Ore section of this chapter. The ore was placed into the flotation cells together with
water at 80°C at volumes and masses appropriate for the respective sizes of cell. The starting slurry
density for each cell was maintained between 30% and 35%. After the ore was added to a cell, the
collector was added, and the mixture was conditioned for 10 minutes. During this time, the pH
was set and maintained using soda ash. Frother (if required) was added to the cell 2 minutes prior
to the end of the conditioning phase.
37
|
Colorado School of Mines
|
Table 3.5 - Froth flotation conditions
Collector 2 Collector 5
Collector Concentration .0075M .0025M
Water Temperature 80° C 80° C
Slurry pH 8.5 9
Continuous pH Adjustment Yes
Flotation Time Various
3.4.1 Locked Cycle Testing
Figure 3.3 shows the flowsheet that was used for the locked cycle flotation tests. The
flowsheet consists of a simple rougher, scavenger, cleaner, recleaner layout. Every cycle begins
with a rougher flotation. Three kilograms of fresh feed are also added to the start of each cycle of
the locked cycle test. These three kilograms were combined with materials from the cleaner tailings
and scavenger concentrates. The cleaner concentrates were floated again in the recleaner stage in
order to further clean the materials and increase the final grade. Recleaner tailings were returned
to the cleaner stage. The tailings from the from the rougher stage are floated again in the scavenger
stage to ensure the largest possible recovery of valuable materials. In this manner, the flowsheet
mimics flowsheets found in industry as closely as possible.
39
|
Colorado School of Mines
|
CHAPTER 4: RECOVERY CURVES
In order to appropriately complete the locked cycle flotation testing for various collector
schemes, it is first important to know the proper retention time for each cell to maximize the
recovery. For this, recovery curves were created for each stage of the locked cycle flotation process
for each of the collectors of interest. The flotation results for collector 2 and collector 5 used to
calculate the rougher curves in this chapter can be found in APPENDIX A.
4.1 Recovery Curves
Cumulative recovery curves are used to calculate the rate at which the recovery of a given
material increases within a flotation cell. To calculate a recovery curve, a normal flotation
experiment is performed. Rather than allowing the froth to continuously be collected into a singular
container, the froth generated by the flotation cell is recovered in separate containers on a timed
basis. For the purposes of the recovery curves that were measured in this work, one minute was
selected as the intervals for which froth would be collected from the flotation cell. Froth was
recovered into a container for one minute at a time after which a new container was placed at the
lip of the cell to capture the next minute of froth. The froth was then filtered by pressure filtering
and assayed for grade and recovery of REE. The recovery versus time for the experiment were
then plotted to give the recovery curve. This process was repeated for each stage of the locked
cycle test as well as for each collector tested.
4.2 Rougher Recovery Curve
4.2.1 Collector 2
Figure 4.1 shows the recovery curve for the rougher stage of the locked cycle circuit when
tested with Collector 2. The recovery of REEs in this stage of flotation begins to slow down
45
|
Colorado School of Mines
|
Table 4.1 - Cumulative mass, REE grade, and cumulative REE recovery for the Collector 2
rougher recovery curve
Cumulative
Cumulative REE Grade
Time (min) REE Recovery
Mass (g) (%)
(%)
1 342.3 28.34 46.35
2 478.4 29.13 66.58
3 501.4 29.78 71.34
4 510.1 29.95 73.00
5 518.2 30.06 74.42
6 524.9 30.20 75.74
4.2.2 Collector 5
Figure 4.3 shows the recovery curve for the rougher stage of the locked cycle circuit when
tested with Collector 5. The flotation response for Collector 5 is much slower than that of Collector
2. Because of this, the flotation time required for the rougher stage of the circuit with Collector 5
is also longer. Ultimately, a flotation time of 6 minutes was selected for the rougher stage of
flotation for Collector 5 flotation. This point was selected because it was a point at which the
flotation process had recovered a significant portion of the available REEs, and the rate was
beginning to decline drastically. Figure 4.4 shows the grade-recovery curve for the collector 5
rougher stage. At the six-minute mark, the grade begins to drop off rapidly further confirming that
this is the optimal operating point. Table 4.2 shows that flotation past 6 minutes, while it did
improve the stage recovery, began to decrease the grade of the concentrate to below that of the
feed grade. Materials that were not recovered in this stage had a second chance to be recovered in
the scavenger stage of the flotation circuit.
47
|
Colorado School of Mines
|
Table 4.2 - Cumulative mass, REE grade, and REE cumulative recovery for the Collector 5
rougher recovery curve
REE
Cumulative REE Grade
Time (min) Cumulative
Mass (g) (%)
Recovery (%)
1 184.3 15.98 15.27
2 521.9 12.82 37.72
3 832.0 11.37 56.00
4 1104.9 9.30 69.17
5 1298.8 9.27 78.49
6 1425.8 8.91 84.36
7 1528.0 8.58 88.90
8 1611.6 6.76 91.84
9 1679.0 6.55 94.13
10 1729.7 5.60 95.60
4.3 Scavenger Recovery Curve
4.3.1 Collector 2
Figure 4.5 shows the recovery curve for the scavenger stage of the locked cycle circuit with
Collector 2. The REE recovery values were very low for this stage of flotation with collector 2 but
so were the mass values that were pulled from this stage. Table 4.3 shows the cumulative mass,
REE grade, and REE cumulative recovery for this stage. This table shows that the mass recovered
by the scavenger stage is very small when compared to the total mass that was fed into the cycle.
Two minutes was selected as the operating point for this stage in the flotation circuit because after
this point, the remaining flotation time only accounted for less than three percent of the total REE
recovery for the system.
49
|
Colorado School of Mines
|
Table 4.4 - Cumulative mass, REE grade, and cumulative REE recovery for the Collector 5
scavenger recovery curve
REE
REE
Cumulative Cumulative
Time Grade
Mass (g) Recovery
(%)
(%)
1 59.3 11.46 15.42
2 199.9 7.44 39.18
3 362.0 5.98 61.17
4 509.0 4.43 75.94
5 602.6 2.98 82.26
6 692.3 2.55 87.45
7 763.9 1.93 90.60
8 825.5 1.79 93.10
4.4 Cleaner Recovery Curve
4.4.1 Collector 2
Figure 4.7 shows the recovery curve for the scavenger stage of the locked cycle circuit with
Collector 2. Table 4.5 shows the cumulative mass, REE grade, and cumulative REE recovery for
the Collector 2 recovery curve. For this stage of the flotation circuit a flotation time of six minutes
was selected. After six minutes of flotation time, the recovery of REEs did not further increase
significantly. Because this is a cleaner stage, it was more important to achieve a high separation
and upgrade ratio than to have the highest possible recovery. Flotation past the six-minute mark
began to show signs of floating gangue materials leading to a decreased grade for the stage.
Materials that were not floated in this stage were returned to the rougher stage of flotation to ensure
they were not lost.
52
|
Colorado School of Mines
|
4.5 Recleaner Recovery Curve
4.5.1 Collector 2
The selected flotation time for this stage was selected to be three minutes. This flotation
time allowed for the grade to be increased as much as possible. Because this was the final stage
for the flotation setup, the primary goal was to increase the grade as much as possible. All the
materials which were not floated during this stage of the flotation process were returned to the
cleaner stage of the flotation process.
4.5.2 Collector 5
Figure 4.9 shows the recovery curve for the recleaner stage of the locked cycle flotation
circuit with Collector 5. This stage reports very little recovery. This is primarily because there was
no additional collector added to the flotation cell during conditioning. The material in this cell
already has been treated by several collector conditioning stages before this point. Table 4.7 shows
the cumulative mass, REE grade, and cumulative recovery for the Collector 5 recleaner recovery
curve. Based on the mass pull and the decreased rate of recovery, a flotation time of 6 minutes was
selected for this stage. At this point the recovery for the cell was maximized with minimal
degradation of the overall grade. Materials that were not recovered in this stage of flotation were
returned to the cleaner to ensure they had a chance to be floated.
55
|
Colorado School of Mines
|
5.2 Collector 2
The locked cycle flotation of collector 2 was performed through the course of 6 cycles.
Each cycle was fed with 3000 grams of freshly ground Bastnäsite ore. The flowsheet for this locked
cycle flotation test can be found in Chapter 2. It included a simple rougher, scavenger, cleaner,
recleaner layout. The test, referred to as LCT-C, returned a REO grade between 58.5% and 66.9%
and a recovery between 42.8% and 74.7%. The ranges for grade and recovery occur because of the
different calculations methods for locked cycle test results. More information about the methods
for results calculations can be found in Chapter 2.
5.2.1 Mass Conservation
The locked cycle test for collector 2 was performed in 6 cycles. Figure 5.1 shows the mass
stability for each of those cycles. The blue line shows the mass conservation for the material for
each cycle. This is calculated by drying and weighing the materials that enter and exit the system
during each cycle. The ratio of the materials leaving the system for that cycle determine the mass
conservation. A similar calculation was performed for the rare earth fraction that enters and leaves
the system as well. That information is shown in the orange line. Table 5.5 shows the values that
were used to make Figure 5.1.
Table 5.5 - Per cycle mass and REE conservation for collector 2 locked cycle test
Cycle Mass (g) Mass % REO %
1 2233.3 74.7 45.1
2 2865.1 95.9 63.3
3 2748.8 92.0 61.2
4 2983.5 99.8 68.1
5 2823.7 94.5 71.5
6 2931.8 98.1 71.5
Total 16586.2
Avg 3-6 96.1 68.1
60
|
Colorado School of Mines
|
Mass Stability
100
)
% 90
(
n
o 80
i
t
a
v
r 70
e
s
n
o
C 60
s
s
a 50
M
40
1 2 3 4 5 6
Cycle Number
Mass % REO %
Figure 5.1 - Mass stability for the locked cycle test of collector 2
Both the mass conservation and the REO conservation for the locked cycle test come to a
consistent value by the end of six cycles. The mass conservation remains consistent at near 100%
of the material being put into the system exiting through the concentrate and tailings. The REE
mass conservation maxes out at approximately 70%. This means that 70% of the REE that enter
the system per cycle exit the cycle through the tailings and concentrate. A possible explanation for
this is that the REE content is not being recovered and is being ‘lost’ to the middlings. In a locked
cycle test, the middlings are not included in the per cycle calculations. The middlings are
consistently passed to the next appropriate spot in the flowsheet. The final middlings are analyzed
at the end of the final cycle and contain the additional REE materials that were ‘lost’ in previous
cycles.
5.2.2 Results
Table 5.6 shows the per cycle results for the recleaner flotation concentrate, scavenger
flotation tailings, and middlings from the locked cycle test. The total in the REO column was
61
|
Colorado School of Mines
|
calculated as the weighted average of all the individual grades. The middlings include the cleaner
flotation tailings, scavenger flotation concentrate, and recleaner flotation tailings. All the
middlings are weighed and assayed after the final cycle, in this case cycle 6. A full breakdown of
the data including grades for REEs, calcium, and barium can be found in APPENDIX B.
Table 5.6 - Per cycle results for the collector 2 locked cycle flotation test
Cycle Mass (g) Mass % REO % REO Dist.
RCFC 1 176.4 1.0 52.8 6.4
RCFC 2 208.9 1.2 53.6 7.7
RCFC 3 170.7 1.0 58.7 6.9
RCFC 4 188.9 1.1 56.6 7.4
RCFC 5 209.9 1.2 57.4 8.3
RCFC 6 185.0 1.0 61.7 7.9
SFT 1 2056.9 11.5 1.2 1.7
SFT 2 2656.2 14.9 2.0 3.6
SFT 3 2578.1 14.5 2.3 4.1
SFT 4 2794.6 15.7 2.5 4.8
SFT 5 2613.8 14.7 2.5 4.5
SFT 6 2746.8 15.4 2.6 5.0
CFT 6 868.9 4.9 33.9 20.3
SFC 6 110.3 0.6 33.6 2.6
RCFT 6 254.3 1.4 50.1 8.8
Total 17819.7 100.0 8.1 100.0
Figure 5.2 shows a grade comparison for REEs between the RCFC and the SFT over the
course of the 6 cycles that were performed in the LCT for collector 2. The figure shows that the
grade of the REE in the concentrates continues to rise over the course of the 6 cycles. The grade
of the tailings also rises during the same period, but this can be expected as the recycling load of
62
|
Colorado School of Mines
|
REE increases. A breakdown of comparisons for each of the four primary REEs that make up the
bastnäsite floated in this work can be found in the 0.
Collector 2 LCT REE Comparison
54.00 3.50
53.00
3.00
)
% 52.00 )
%
(
e 51.00 2.50 (
d e
a d
r G 50.00 2.00 a r
C 49.00
G
F 1.50 T
C 48.00 F
R S
47.00 1.00
46.00
0.50
45.00
44.00 0.00
1 2 3 4 5 6
Cycle Number
REE RCFC REE SFT
Figure 5.2 - REE grade comparison between the recleaner flotation concentrate and the
scavenger flotation tailings from collector 2 locked cycle flotation
Figure 5.3 shows the grade comparison for calcium between the RCFC and the SFT over
the course of the six cycles that were performed in the LCT for collector 2. The figure shows that
the grade of calcium recovered in the RCFC is steady throughout the course of the flotation test.
As the grade of the REE in the RCFC is increasing, there was not a large increase in the calcium
grade indicating that the concentrate recovered from the RCFC was becoming more concentrated
in REE and therefore there was a high level of calcite rejection in the test. Calcite rejection through
flotation allows for lower acid consumption in the downstream processing.
63
|
Colorado School of Mines
|
Table 5.8 - SME Procedure method of calculating locked cycle testing results for collector 2
SME Procedure
Mass (g) Mass % REO % Recovery
Feed 2872.0 100.0 6.2 100.0
REE Con 188.6 6.6 58.5 62.4
Tails 2683.3 93.4 2.5 37.6
Table 5.8 shows the results for the locked cycle test when calculated using the SME
Procedure (see Chapter 2). Using the SME Procedure to calculate the results returned a REO grade
58.5% and a recovery of 62.4%.
Table 5.9 shows the calculation using the Concentrate Production Method and Table 5.10
shows the calculation using the Tailings Production Method. The tailings production method is
simply the inverse of the concentrate production method where the mass of the concentrate is
calculated by subtracting the mass of the tailings from the mass of the feed. These two methods of
calculation show the largest differences in grade and recovery when compared to the other
calculation methods. The concentrate production method returns a recovery of 42.8% and the
tailings production method returns 74.3%. The grade is more similar between the two methods
with the concentrate production method returning a grade of 58.5% and the tailings production
method returning a grade of 66.9%.
Table 5.9 - Concentrate production method of calculating locked cycle testing results for
collector 2
Concentrate Production
Mass (g) Mass % REO % Recovery
Feed 2970.0 100.0 8.7 100.0
REE Con 188.6 6.4 58.5 42.8
Tails 2781.3 93.6 5.3 57.2
65
|
Colorado School of Mines
|
Table 5.10 - Tailing production method of calculating locked cycle testing results for collector 2
Tailing Production
Mass (g) Mass % REO % Recovery
Feed 2970.0 100.0 8.7 100.0
REE Con 286.6 9.7 66.9 74.3
Tails 2683.3 90.3 2.5 25.7
Figure 5.4 shows the points from the four methods of calculation plotted recovery versus
grade. These four points can be used to determine the probable outcome if this reagent scheme and
flowsheet were to be applied to a full-scale flotation plant. The variability in the points on this
figure occur because this test did not include a regrind circuit to recovery middlings. The middlings
materials account for 31.7% of the total rare earth oxides that entered flotation but only account
for 6.9% of the total mass of the test. Additional work regarding the inclusion of a regrind circuit
is recommended.
68.0
67.0
66.0
65.0
64.0
%
,e63.0
d
a62.0
r
G
O61.0
E
R60.0
59.0
58.0
40.0 50.0 60.0 70.0 80.0
REO Recovery, %
N-Product SME Conc Production Tail Production
Figure 5.4 - Recovery vs grade for the locked cycle test of collector 2
66
|
Colorado School of Mines
|
5.3 Collector 5
The locked cycle flotation of collector 5 w0as performed through the course of 6 cycles.
Each cycle was fed with 3000 grams of freshly ground Bastnäsite ore. The flowsheet for this locked
cycle flotation test can be found in Chapter 2. It included a simple rougher, scavenger, cleaner,
recleaner layout. The test, referred to as LCT C5, returned a REO grade between 13.2% and 13.8%
and a recovery between 26.6% and 41.3%. The ranges for grade and recovery occur because of the
different calculations methods for locked cycle test results. More information about the methods
for results calculations can be found in Chapter 2.
5.3.1 Mass Conservation
The locked cycle test for collector 5 was performed in 6 cycles. Figure 5.5 shows the mass
conservation for each one of those cycles and Table 5.11 shows the data that was used to create
that figure. The blue line shows the mass conservation for material in each cycle. The orange line
represents the conservation of rare earth oxides for each cycle. These data points are calculated by
weighing and assaying the materials that enter and exit during each cycle of the test. A ratio of the
mass of materials entering and leaving the system determines the mass stability for that cycle. The
same can be repeated for the calculation of REO that enter and leave the cycle.
Table 5.11 - Per cycle mass and REE conservation for the collector 5 locked cycle test
Cycle Mass (g) Mass % REO %
1 1285.0 43.0 15.3
2 2239.6 74.9 57.3
3 2819.7 94.3 96.3
4 2638.4 88.3 68.8
5 2861.4 95.8 94.7
6 2762.3 92.4 85.8
Total 14606.4
Avg 3-6 92.7 86.4
67
|
Colorado School of Mines
|
Mass Stability
100
)
%
80
(
n
o
i
t a 60
v
r
e
s
n 40
o
C
s
s 20
a
M
0
1 2 3 4 5 6
Cycle Number
Mass % REO %
Figure 5.5 - Mass stability for locked cycle test with collector 5
Both material conservation and REO conservation are relatively stable by the end of 6 six
cycles. The material conservation fluctuates between 90% and 100%. The REO conservation is
more variable in the final cycles but still averages 80% for the final 4 cycles. The remaining 20%
of REO materials may be reporting to the middlings. Middlings are not calculated between
individual cycles of a locked cycle test. It is possible that a portion of the materials from each cycle
reports to the middlings.
5.3.2 Results
Table 5.12 shows the per cycle results for the scavenger tailings, recleaner concentrates
and the middlings recovered and assayed at the end of the locked cycle test. The total in the REO
column was calculated as the weighted average of all the individual grades. The middlings include
the cleaner tailings, scavenger concentrates, and recleaner tailings. A full breakdown of the data
including grades for REEs, calcium, and barium can be found in APPENDIX B.
68
|
Colorado School of Mines
|
Table 5.12 - Per cycle results for collector 5 locked cycle test
Cycle Mass (g) Mass % REO % REO Dist.
RCFC 1 63.6 0.4 20.9 0.9
RCFC 2 223.4 1.3 14.8 2.3
RCFC 3 426.9 2.4 13.1 3.9
RCFC 4 253.6 1.4 13.8 2.4
RCFC 5 527.2 3.0 14.2 5.2
RCFC 6 691.0 3.9 12.4 5.9
SFT 1 1221.4 6.8 1.9 1.6
SFT 2 2016.2 11.3 5.1 7.1
SFT 3 2392.8 13.4 7.2 12.0
SFT 4 2384.8 13.4 5.4 8.9
SFT 5 2334.2 13.1 6.4 10.4
SFT 6 2071.3 11.6 5.7 8.2
CFT 6 1525.7 8.5 13.1 13.8
SFC 6 1069.2 6.0 15.2 11.3
RCFT 6 656.2 3.7 13.5 6.1
Total 17857.5 100.0 8.1 100.0
Figure 5.6 shows a grade comparison for REEs between the RCFC and the SFT over the
course of the 6 cycles that were performed in the LCT for collector 5. The grade of REEs from the
RCFC and SFT move in inverse directions throughout the LCT. After the first several cycles, the
grade becomes stable for both the RCFC and SFT. The figure shows that as the grade of REEs in
the RCFC decreases, the grade in the SFT increases. As materials are not being recovered in the
concentrate, they are instead reporting to the tailings. A breakdown of comparisons for each of the
four primary REEs that make up the bastnäsite floated in this work can be found in the 0.
69
|
Colorado School of Mines
|
Table 5.14 shows the results when the SME Procedure (Chapter 2) is applied to the
calculation of the results for the locked cycle test of collector 5. The SME Procedure returns a
REO grade of 13.2% and a recovery of 30.7%.
Table 5.15 shows the results calculated using the concentrate production method and Table
5.16 shows the results calculated using the tailings production method. As mentioned in Chapter
2, the tailings production method is essentially the same as the concentrated production method
but inverted, using the tailings rather than the concentrates as the basis. These two methods
produce similar grades but a wide range of recovery values. The concentrate method gives a REO
grade of 13.2% while the tailings method gives a REO grade of 13.8%. The recovery values vary
more widely with the concentrate method returning a 26.6% recovery and the tailings method
returning a recovery of 39.8%.
Table 5.15 - Concentrate production method for the calculation of collector 5 locked cycle
testing results
Concentrate Production
Mass (g) Mass % REO % Recovery
Feed 2976.3 100.0 7.9 100.0
REE Con 474.7 15.9 13.2 26.6
Tails 2501.6 84.1 6.9 73.4
Table 5.16 - Tailing production method for the calculation of collector 5 locked cycle testing
results
Tailing Production
Mass (g) Mass % REO % Recovery
Feed 2976.3 100.0 7.9 100.0
REE Con 680.5 22.9 13.8 39.8
Tails 2295.8 77.1 6.2 60.2
72
|
Colorado School of Mines
|
This flowsheet is a simplified version of a flowsheet which may be found in a flotation
plant. Designing the flowsheet in this way allows the results from JKSimFloat to replicate the test
work that was performed through locked cycle testing. The cleaner feed sump was included as a
point in which to reset the slurry density before the cleaner phase. This also closely matches what
was performed in the locked cycle testing circuit. In locked cycle tests, the tailings from the
rougher were directly treated in the scavenger. Materials traveling to the cleaner stage of flotation
are reconditioned with fresh water before being floated. The cleaner feed sump included in this
flowsheet performs the same function.
6.2 Ore Floatability
The first step in beginning to model the flotation flowsheet in JKSimFloat is to determine
the floatability of the ore. The floatability is the primary factor used within JKSimFloat for the
calculation of flotation results. More information about the way that JKSimFloat calculates and
models the results from flotation experiments can be found in Chapter 2.
The floatability of the minerals of interest are different for each collector tested and
therefore must be calculated separately for each collector that is simulated. Calculating the
floatability requires recovery and time data for each stage of flotation. This data is available
through the recovery curves that were performed for each collector and each stage of flotation.
Figures and data from the recovery curves can be found in Chapter 4.
Floatability distribution is the second component of floatability that is required to
successfully simulate the flotation flowsheet in JKSimFloat. The floatability distribution is used
to tell the software what percentage of each material falls in the fast or non-floating floatability
category. To calculate the floatability distribution, the maximum recovery of materials from locked
cycle testing were used as the proportion of materials that fall into the fast-floating section.
75
|
Colorado School of Mines
|
6.2.1 Rate Constant Calculations
Floatability is a function of the flotation time and recovery of the flotation cell. The rate
constant was calculated from first order kinetics derived from Equation 6.1. This equation can be
integrated to calculate the recovery based on the rate constant and time floated. This equation is
represented in Equation 6.2. After rearranging Equation 6.2, Equation 6.3 was used to calculate
the rate constant. This rate constant is then used together with the bubble surface area flux to
calculate the ore floatability. The floatability is calculated for Calcite, Barite, and Bastnäsite for
collector 2 and collector 5, respectively.
(6.1)
ⅆ𝑁𝑃
= −𝑘𝑁𝑝
ⅆ𝑡 (6.2)
−𝑘𝑡
𝑅 = 1−ⅇ
(6.3)
ln(1−𝑅)
𝑘 = −
6.3 Conditions and Assumptions 𝑡
Several assumptions were made for the sake of simplifying the simulation as well as
creating a flowsheet that will more closely match the flowsheet used in locked cycle testing work.
The first assumption that was made is that the materials and water in the flotation cells have already
been adjusted to the appropriate conditions. This assumption is made because the floatability of
the material is directly related to the reagents and conditions applied during the testing. JKSimFloat
also does not simulate the rate of collector attachment in its models. The material used is also
assumed to fall into one of two floatability categories, fast floating and non-floating.
A second assumption that was made when designing the model was that each part of the
flotation process will only include one stage. For example, the Mountain Pass flowsheet (Chapter
2) includes multiple stages of rougher flotation. For simplicity, and to closely match the flotation
76
|
Colorado School of Mines
|
flowsheet used in locked cycle flotation, only one stage was used in each step of the flotation
process.
The flowsheet designed for the modelling with JKSimFloat assumes 100 tph. Assuming
the flotation plant operates for 20 hours per day, this would be approximately in line with the
current operations at Mountain Pass which mine approximately 1800 tonnes of ore per day.
Assuming 2000 tonnes per day of processing also allows the data from JKSimFloat to be used in
the economic models which also assume 2000 tonnes of ore per day are processed through the
plant.
Finally, it is assumed that the floatability values are inherent to the ore that is being floated
and therefore will not change throughout the course of the flowsheet. The floatability value will
be different depending on the collector that is used and is calculated separately for each element
and collector.
6.4 Collector 2 Simulation
6.4.1 Collector 2 Floatability and Floatability Distribution
The graphs for the floatability of barite, calcite, and REE with collector 2 are shown in
Figure 6.2, Figure 6.3, and Figure 6.4, respectively. The rate constant changes through the course
of the flotation period because there is less material available to be floated as the concentration of
materials in the cell decreases. As the flotation continues, the rate constant becomes more stable.
For the calculation of floatability, the rate constant for the final 3 minutes of flotation was
averaged. Table 6.1 shows the rate constant and floatability for the three primary minerals for
collector 2.
77
|
Colorado School of Mines
|
CHAPTER 7: AMICS MINERALOGY
AMICS (Advanced Mineral Identification and Characterization System) software is an
automated mineral classification and identification software package that can be used to determine
prevalent mineral phases as well as liberation of valuable minerals from their respective gangue
materials. Performing mineralogy on the products from locked cycle flotation testing provides
additional information about the performance of a collector relative to the available liberation.
AMICS Analysis was performed by Dr. Paul Miranda at Eagle Engineering.
7.1 Data Correlation
To determine the viability of the data from the AMICS software, the elemental results from
AMICS were correlated against the elemental results from XRF analysis performed at the
Colorado School of Mines. Table 7.1, Table 7.2, and Table 7.3 show the correlation between REE
Minerals, calcium, and barium for the ore, collector 2 locked cycle test, and collector 5 locked
cycle test, respectively. The data for correlation between CSM data and AMICS data for individual
REEs can be found in APPENDIX D.
Table 7.1 - Correlation between CSM XRF elemental results and AMICS elemental results for
the ore used in flotation
REE Minerals (%) Calcium (%) Barium (%)
Sample CSM AMICS CSM AMICS CSM AMICS
Ore 7.36 8.31 12.09 14.47 13.36 14.5
85
|
Colorado School of Mines
|
Table 7.2 - Correlation between CSM XRF elemental analysis and AMICS elemental results for
collector 2 locked cycle flotation products
REE Minerals (%) Calcium (%) Barium (%)
Sample CSM AMICS CSM AMICS CSM AMICS
SFT 6 2.87 3.23 11.29 16.46 14.7 15.02
RCFC 6 53.27 55.47 2.67 2.84 1.74 1.88
CFT 6 29.98 29.72 9.88 11.39 6.21 5.42
SFC 6 28.68 29.13 10.65 12.59 3.53 3.5
RCFT 6 37.47 37.05 5.43 5.46 3.31 3.41
Table 7.3 - Correlation between CSM XRF elemental results and AMICS elemental results for
collector 5 locked cycle flotation products
REE Minerals (%) Calcium (%) Barium (%)
Sample CSM AMICS CSM AMICS CSM AMICS
SFT 6 4.66 4.88 10.95 14.84 12.97 13.34
RCFC 6 10.46 10.4 13.91 16.99 11.6 9.53
CFT 6 10.79 11.17 13.22 15.41 11.31 10.49
SFC 6 12.74 12.64 12.55 13.04 12.75 12.79
RCFT 6 11.44 11.64 13.3 15.46 11.28 9.48
Figures for the correlation between the elemental results from CSM and the results from
AMICS were created using the data shown in the tables above. These figures were created and a
linear trendline was drawn to see the level of correlation between the three sets of interest. Figure
86
|
Colorado School of Mines
|
The figures above show R2 values for the correlation between the data from the AMICS
analysis and the CSM XRF analysis for REE minerals, calcium, and barium. The correlation for
REE Minerals returned the highest value out of the three elements at 0.9984. This means that there
is a very high level of correlation between the XRF data and the AMICS data. A high level of
correlation means that the data from the AMICS analysis closely matches the data from XRF
analysis and can therefore be relied upon as an additional method for analyzing results from locked
cycle testing.
7.2 Modal Mineralogy
Modal mineralogy was calculated for the ore used in flotation, collector 2 locked cycle
flotation products, and collector 5 locked cycle flotation products. According to the analysis the
primary REE mineral present in this material is Bastnäsite. Other REE minerals that were present
to a lesser extent were cerianite, parasite, and allanite. Table 7.4 shows the modal mineralogy for
the ore used in flotation experiments. Table 7.5 and Table 7.6 show the modal mineralogy for the
products of locked cycle tests for collector 2 and collector 5, respectively.
89
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.