Recovery of Ultra Fines Using Imhoflot Pneumatic Flotation

– Two Pilot Plant Case Studies Recovering Nickel and Zinc

from Tailings Streams

Michael Battersby1, Richard Mark Battersby1, Stephen Flatman1 Rainer Imhof2, Heinrich
Sprenger2 and Tomas Bragado3

1. Maelgwyn Mineral Services Ltd, Ty Maelgwyn, 1A Gower Road, Cathays, Cardiff,

CF24 4PA, UK. email: [email protected]
2. Maelgwyn Mineral Services Germany, 23 Leharweg, Dorsten, 46282, Germany
3. Rio Narcea Resources, El Real de La Jara, Spain

ABSTRACT

It has been well documented that conventional tank flotation is inefficient in the recovery

of metallic fines and very inefficient in the recovery of ultra fines. In contrast Imhoflot

pneumatic flotation, with the combination of pico bubble generation and order of

magnitude increase in energy utilization in the collection zone over tank flotation has the

ability to recover such ultra fines.

This paper details two separate site based pilot plant investigations into the ultra fine

losses occurring into tailings by tank flotation and the ability of Imhoflot to recover these

ultra fines. The first case study looks at a nickel operation in Europe where pilot plant

work using a three stage Imhoflot G-Cell plant demonstrated its ability to recover

approximately 30% nickel from the plant final tails, predominantly in the minus eleven

micron fraction size range. The second case study investigated losses at a zinc

operation where the liberation size of the zinc mineralisation was deemed to be around

seven microns and this was the target grind size of the operation. A two stage Imhoflot

G-Cell pilot plant achieved a 20% zinc recovery from the final cleaner tailings stream.
 2

KEYWORDS

Froth flotation, pneumatic flotation, ultra-fines beneficiation, nickel sulphide, zinc suphide

INTRODUCTION

Froth flotation became the dominant method for the recovery of metallic minerals from

ores shortly after its initial development now well over a hundred years ago (Lynch

2010). Since its early development the use of agitated tanks and compressed air

became the standard method of effecting the collection of the metallic minerals. Such

“tank” flotation still dominates the market in new beneficiation plants with estimates in

excess of 95% market share; hence such tank flotation is also referred to as

“conventional” flotation.

However over those 100 years numerous technical papers have identified conventional

flotation as been inefficient at both extremities of the particle size range. Figure 1.0

indicates the standard view of almost all tank flotation plants (Pease 2004). In the coarse

sizes (say +150 microns) this may be attributable to liberation issues but also the

inabilities of the froth to carry liberated particles to a concentrate due to the competing

tasks of originally also trying to maintain such particles in suspension and having them

contact with bubbles. In the fine size range (say -30 microns) conventional flotation can

be said to be inefficient and in the ultra fine size range (say -15 microns) extremely

inefficient. Since the early days very few developments have been made to address

these issues. The tanks originally being square are now generally round and, in an

attempt to increase bubble – particle contact for the ultra fines, in some operations much

more energy is inputted into the system by the installation of larger motors. However

putting more energy into the tank system also compromises coarse particle recovery as

mentioned above.
 3

Figure 1.0 Generalized Metal Recovery Verses Size Curve (after Pease 2004)

Historically ore bodies have been mined that are coarse grained and thus give good

recoveries in the tank flotation cell size range. If the ore is more fine grained or highly

disseminated, operations have been made to be economic with lower recoveries by

generally operating at higher throughputs to reduce unit operating costs. It is a

generalization that in recent times the “good” orebodies have largely been depleted and

those remaining are lower grade, highly disseminated ore bodies that require finer and

finer grinding for optimum liberation but yield lower recoveries due to the inherent losses

in the tank flotation process. Consequently much attention has been directed to the area

of fine and ultra fine metals recovery. This is in both areas of liberation with ultra fine

grinding technology and then in the metals subsequent recovery by flotation or other

methods.

There have been attempts to improve recoveries with tank flotation by various

flowsheets such as split streaming into different size fractions before flotation. At the Mt

Keith nickel operation in Western Australia that treats a low grade pentlandite ore the
 4

flotation feed is split into coarse, fine and slimes fractions where flotation takes place in

three separate lines. Recovery must be economic for the operation to continue but

owners reported nickel recovery of only 60% in 1994 the year after start up, increasing to

69% in 2003. Another flowsheet concept is to try and first recover coarse particles by

normal flotation tanks at the start of the flotation process followed by additional tanks

with higher power inputs to try and recover fines. Both these flowsheets will increase

some recovery of fines but at considerable capital cost of the extra equipment.

ALTERNATIVE FLOTATION METHODS

Column Flotation

Commercially developed in the late 1970’s, column flotation offers high selectivity over

tank flotation that can give higher grade concentrates. However, columns, and their

derivatives that include internal or external sparging, like all counter-current flotation

mechanisms suffer from low unit metal recoveries. This generally makes them

unattractive for roughing and main stream flotation recovery applications.

Pneumatic Flotation

Pneumatic flotation started to be commercialized in the early 1980’s and originates out of

research work done by Professors Bahr and Simonis in Germany. If nomenclature is

simplified by “tank” flotation and “column” flotation, pneumatic flotation can be described

as “pipe” flotation.

IMHOFLOT PNEUMATIC FLOTATION

The history and development of Imhoflot pneumatic flotation has been well documented

in many technical papers (Brown 2001, Imhof 2003, Battersby 2005, Sanchez-Pino

2009) and so an overview will only be given here. Maelgwyn Mineral Services (MMS)

has two different types of pneumatic flotation cells. These are the vertically fed (V-Cell)
 5

and tangentially fed (G-Cell). Pneumatic flotation differs from conventional flotation in

that the bubble particle contact takes place outside of the cell itself, within the aerator.

The associated pulp then enters the cell, or a better nomenclature would be the froth

separation tank as the mineralisation of bubbles has now already taken place in the

aerator. Separation of these two vitally important aspects of flotation namely bubble

particle contact and mineral collection has the advantage that both can now be

individually optimised rather than having a trade off situation inherent to conventional

flotation. The method of self aspiration of Imhoflot flotation is based on the well known

venturi principal. However, unlike other pneumatic flotation systems it is not just a simple

venturi. Its patented design has a complex system of nozzles, impingement plates and

gas hold up mechanisms that generate a spectrum of bubble sizes from pico through

nano to milli sizes. Made from advanced ceramic materials the aerators are non blinding

or blocking and do not allow the build-up of scale. There are no moving parts in the

Imhoflot cell where all the energy for mineral collection comes from the feed pump, with

collection/flotation occurring in milliseconds. There is an order of magnitude utilization of

this energy in the collection zone compared with tank flotation.

In Imhoflot pneumatic flotation the residence time of any particle within the aerator is a

function of the feed slurry velocity which is normally kept within the range 15-18m/sec.

This results in a very short residence time (flotation time) in the aerator of a few

milliseconds. However, importantly the majority of the feed pump energy is now focused

into the restricted volume of the aerator. The net effect of this is very high air utilisation

rates up to an order of magnitude higher than conventional flotation (Hay 2009). The

ensuing enhanced kinetics are key to the superior performance of the Imhoflot

pneumatic flotation process. In terms of cell geometry for both cell types the aerator is

common. However the design of the separating zone differs for each cell.
 6

Figure 2.0 Schematic of an Imhoflot V-Cell

Figure 3.0 Schematic of an Imhoflot G-Cell

 7

In the vertical type – Imhoflot V-Cell (Figure 2.0) the separating zone is essentially a

vertical cone. Slurry is injected through nozzles upwards from the base of the cone i.e.

co-current with bubble movement. There is no rotor-stator mechanism as in a

conventional cell as the bubble particle contact has already taken place in the aerator

resulting in a lower energy requirement. Sizing of the Imhoflot V-Cell vertical type is

dictated by the time required for efficient froth separation and corresponds to

approximately two and a half minutes (Imhof, 1993) contributing to an overall very small

footprint.

The Imhoflot G-Cell (Figure 3.0) was developed after the Imhoflot V-Cell. Due to the

bubble-particle contacts happening instantaneously in the aerator, collection is complete

in the pipe system. The cell is just for separation of the mineral rich bubbles from the

tailings gangue. This can be optimized completely independently of the flotation process

and energy is not required to maintain the pulp in suspension or generate more

collisions. The G-Cell was developed that embraces centrifugal froth separation. It

accelerates the separation of fine and ultra fine particles. It has much shorter residence

times of between 30 and 60 seconds, depending on the size of cell (Battersby 2005).

This equates to an even smaller froth separation vessel and consequently plant foot

print. The shorter residence time is thought to reduce gangue entrainment in the froth

phase. Whilst mineral recovery as a function of size varies across different mineral

classes the Imhoflot V-Cell is generally used for normal and coarse particle flotation, with

the Imhoflot G-Cell being favoured for very fine to normal sized particles.

RECOVERY OF ULTRA-FINES

Imhoflot has been installed commercially in plants treating a wide range of minerals such

as gold sulphides, base metals, iron ore, coal, potash, industrial minerals etc. where it is
 8

believed enhanced recovery was obtained in the ultra fines by the nature of Imhoflot.

However no studies have been undertaken confirming this. The work on these two pilot

plant operations therefore specifically related to the recovery of ultra-fine material.

CASE STUDY 1

AGUABLANCA NICKEL MINE, SPAIN

The Aguablanca nickel – copper mine in Spain had identified a problem with high nickel

losses to the tails. Previous QEMSCAN mineralogical analysis had indicated that these

losses were predominantly associated with the very fine fractions. MMS were contracted

to run a three stage Imhoflot G-Cell pilot plant on various streams of the plant to

demonstrate the abilities of Imhoflot to the owners of the mine and to determine if any

improvements could be made to the Aguablanca circuit. This pilot plant work included

treating the final tails of the plant as previous work had indicated the G-Cells ability to

recover metal fines not recovered by conventional tank flotation. However, prior to the

decision to test Imhoflot, Aguablanca had already determined that the losses may have

been attributable to a simple residence time constraint and had therefore purchased new

tank flotation capacity in the form of four, 30m³ Outotec tank cells. After the

commissioning and optimisation of four new tank cells the G-Cell pilot plant was again

run on the new tailings.

BACKGROUND

The Aguablanca nickel – copper mine which is part of the Lundin Mining group is

situated approximately 70 Kilometres north of Seville in Spain. Flotation testwork was

carried out on the mine using a 3x G12 (1.2m diameter) Imhoflot G-Cell pilot plant. The

initial objective of the test work was to investigate whether the G-Cells could increase

nickel recovery to above the current 82%. QEMSCAN analysis of plant tailings, shown
 9

below, indicated significant amounts of liberated nickel sulphides in the minus 9 micron

fraction which were not being recovered by the plants tank cells.

The Aguablanca Mine treats material from the Monaguera nickel- copper sulphide

orebody. The open cast mine produces about 160,000 tpm of run of mine (ROM)

material which is treated by a 260 tph flotation plant (up -rated from the original 190 tph

plant).The head grade is highly variable but is approximately 0.6% nickel, 0.4% copper,

with the sulphide mineralisation being pyrrhotite, chalcopyrite, pyrite and the nickel – iron

sulphides, pentlandite and violerite, and the nickel sulphide millerite. Trace amounts of

nickel also being present in the pyrite and pyrrhotite with the pyrite averaging 1.15 % Ni.

ROM material is fed through a primary jaw crusher producing a d80 of 100mm followed

by a secondary cone crusher with a d80 of 50mm.The grinding circuit is made up of a

SAG mill with pebble cone crusher followed by a ball mill, the cyclone overflow being

80% minus 90 µm. A flowsheet of the flotation plant is given in Figure 5.0 and the G-Cell

pilot plant in Figure 4.0 below. The flotation circuit was originally designed to produce

separate nickel and copper concentrates. These are however now mixed to form a bulk

sulphide concentrate averaging 7% Ni and 6% Cu. Rougher flotation consists of 11x

40m3 Wemco tank cells split into four sections, the first two cells are referred to as pre-

float as they were originally designed to float off talc (MgO) which is a major

contaminant, attracting smelter penalties if above 5% in the final concentrate. These are

now used as copper roughers together with a third cell, which was originally intended as

the only copper rougher. This concentrate is cleaned in 4 x 12 m3 tank cells becoming

part of the final concentrate, the cleaner tailings reporting back to the first nickel cleaner

cell.
 10

Figure 4.0 Imhoflot 3 stage G-12 Pilot Plant Flowsheet

Figure 5.0 Aguablanca Plant Flowsheet
The concentrates of four nickel roughers and four nickel scavengers feed the first, then

second nickel cleaners, the nickel scavenger tails represent the final tails and was the G-

Cell pilot plant feed. The first cleaner tails feeds the scavenger cleaner whose

concentrate is recycled back to the first cleaner, cleaner scavenger tails reports back to

the nickel roughers. The second nickel cleaner tails return to the first cleaner.

Due to the complex and variable mineralogy of the ROM feed together with throughput

changes the reagent suite is varied on a day to day basis but the main collector used is

a Xanthate mix ( 67% SIBX , 33% PAX- 140g/t ) added in stages where required. Aero

3894 (25g/t) is added at the cyclones as a copper promoter together with Aero 412

(10g/t) some of which is added down the rougher circuit. Copper sulphate (33g/t) is

added in the nickel roughers and scavengers as an activator for pyrite, pyrrhotite and the

nickel sulphides. Large amounts of a talc specific depressor mix (70% Depramin, 30%

Finnfix BW,-1000g/t) is added sequentially throughout the circuit, as is a frother mix

(currently 33% Oreprep 591, 33%MIBC, 33% Pine Oil- 120g/t). The rougher circuit is

normally run at pH 8.5, whilst the cleaners are run at up to pH 11.5.

Various mineralogical testwork was performed by Aguablanca on a composite sample of

the current plant tailings including mineral distribution versus size and liberation analysis

by QEMSCAN. The metal distribution according to size results are tabulated in Figure

6.0 below and indicate that approximately 50% of the nickel in the tailings is associated

with ultrafine particles i.e. minus 9 microns. The QEMSCAN testwork results are

summarised in Figure 7.0 which essentially illustrates a plot of the percentage of the

particles for each size fraction against their percentage liberation. For example, for the

+75 micron size fraction approximately 65% of the particles are 10% liberated and the

remaining 35% are 20% liberated. As expected as the size fractions become finer a
 13

higher degree of liberation is apparent but this is limited to approximately 30% liberation

part from the minus 9 micron fraction. The conclusion from all of the above work is that

there are liberated nickel particles which can be recovered but are not recovered by the

existing tank flotation.

Product Weight Assay Distribution

(µm) (%)
Ni Cu Fe STOT(%) Ni Cu Fe STOT
(%) (%) (%)

+ 75 35.48 0.069 0.045 2.85 0.71 19.6 41.0 24.3 25.7

-75+38 21.32 0.097 0.029 5.31 1.73 16.6 15.9 27.2 37.7

-38+9 22.81 0.072 0.018 3.39 0.91 13.2 10.6 18.6 21.2

-9 20.39 0.310 0.062 6.12 0.74 50.6 32.5 30.0 15.4

Head 100.00 0.125 0.039 4.164 0.979 100.00 100.00 100.00 100.00

Assayed Head 0.120 0.039 4.310 1.120

Figure 6.0 Analyses of Aguablanca Tailings Size Fractions

Liberation Ni sulphides
100.00

50.00 +75
%

-75/+38
0.00 -38/+9
<= 10%
<= 20%

-9
<= 30%
<= 40%
<= 50%
<= 60%
<= 70%
<= 80%
<= 90%
< 100%

100%

Liberation %

Figure 7.0 Graphic Showing the Liberation Characteristics of Ni Sulphides Versus

Size
 14

DESCRIPTION OF THE IMHOFLOT G-CELL PILOT PLANT

The G-Cell pilot plant is illustrated in Figures 8.0 and 9.0. Each G-Cell incorporates a

self-aspirating aerator, normally operating between 1.5 and 2.5 bar gauge pressure. This

provides highly dispersed fine bubble froth, with relatively low energy input. The feed

pump feeds the aerator, which discharges to the distributor and in turn the four

tangential separator nozzles. Froth separation takes place as the fluid and froth spin in

the outer bowl of the cell, with flow of froth to the central froth funnel. All feed pumps are

equipped with frequency converters. The speed of the first pump can adjusted according

to the desired flow rate and the pressure at the aerator. The second pump speed is then

automatically controlled by the level of the pulp of the first G-Cell. The third and fourth

pumps are controlled in the same manner. The airflow to the aerators can be adjusted by

a manual valve, which throttles the flow to the self aspirating aeration units. A pressure

gauge in front of the aerator indicates the vacuum. The pulp level in the separation part

of the G-Cell is measured by level probes and controlled by the pulp flow to the next cell.

The froth depth can be varied by means of the level control probes pressure sensors

and a separately controlled movable central cylinder whose top is the froth launder.

Consequently the froth height can be controlled above the fixed pulp surface. The pilot

plant was controlled by means of a laptop connected to a Siemens PLC system with

readings of throughput, pump speeds, aerator pressures, venturi suction vacuums, and

pulp depths being shown. The main variables that can be used by the operator to adjust

performance are, feed pressure and volume, pulp relative density (solids content),

aeration rate, and froth depth.

The pilot plant has a footprint of approximately 5500mm x 2100mm and an installed

height of approximately 3600mm. The diameters of the cells are 1.2m and the plant
 15

requires 3 phase, 400V 40KW power. The plant is designed to fit into a standard 20 foot

container for transportation.

Fig 8.0 Schematic of Imhoflot 3 Stage G 12 Pilot Plant

Figure 9.0 Imhoflot G1.2 Three Cell Pilot Plant at Aguablanca

 16

TESTWORK

The testwork was carried out on pulp piped from the two discharge columns of the last

Wemco nickel scavenger cell. The pulp was gravity fed through two HDPE tubes which

joined together through a Y piece and then passed over a feed box containing a 5mm

tramp screen before being connected directly to the inlet of a Warman 3/2 feed pump. In

addition there was an inline screen situated on the 1st G-Cell feed line to stop tramp

material entering the aerator. This could easily be opened for cleaning.

The pilot plant was designed with a maximum throughput of 35 m3/hr (flowmeter limit)

and a minimum of 20 m3/hr, with feed control through ball valves on the plant feed pipes

in front of the feed box. The speed of the first pump is set at the PLC which then controls

the aerator pressure in the first cell. The venturi suction pressure for each cell is

controlled manually by adjusting the inlet valve next to the aerator. Each inlet had its

own control valve.

Each test run was conducted after a start up procedure during which the plant was

started and flowrates, pump speeds, and pulp levels slowly increased until steady state

is achieved at the required test setting. During test runs the following samples were

taken every 15 minutes; feed, tailings, concentrate from each cell and a composite

concentrate sample taken from the main launder. All sensor readings were recorded,

together with reagent additions, pulp density readings taken using a Marcy scale, and

composite concentrate flowrate. Sampling continued as long as the plant maintained

steady state. One of the major problems encountered during the test work was that the

main plant feed was sometimes highly erratic in both flowrate and percent solids which

made maintaining a steady state in the pilot plant difficult. The averaged results of the

first set of testwork are summarised in Figure 10.0 below.

 17

Feed Concentrate Tailings Recovery Mass

Pull
Ni % Cu % MgO Ni % Cu % MgO Ni % Cu % MgO Ni %
% % % %
0.102 0.037 10.95 1.227 0.206 8.44 0.072 0.032 11.12 29.9 2.6

Figure 10.0 - Averaged Results of the 1st Final Tailings Testwork.

The first set of tailings testwork ran for approximately three weeks on days when the

main plant was in operation and the feed was relatively stable. The testwork was then

halted whilst the Aguablanca mine installed the four new 30 m3 tank cells as final tailings

scavengers. The detailed results of the initial tests have been summarised in Appendix

1.0. Accepting the varying feed conditions and pilot plant test settings the G- Cell pilot

plant was able to recover approximately 30% of the nickel content of the tailings at a

mass pull of 2.6%. This would represent an over 5% increase in overall nickel recovery

for the operation. The corresponding Nickel concentrate grade was 1.2%. Given the G-

Cells well known ability to recover fines which are not recovered by tank flotation in

conjunction with the previously provided tails size versus grade information it was

considered highly probable that a high proportion of the recovered nickel in the G-Cells

is associated with the fines not being recovered by tank flotation. However there was the

possibility that the G-Cell pilot plant was recovering the extra nickel due to a shortage of

residence time in the main plant rather than being a unique feature of Imhoflot. As

previously mentioned Aguablanca mine had already thought that increased residence

time in the current circuit would increase recovery. The new tank cell scavenger cells

increased residence time by 31% which together with a planned move to cut feed from

260tph to 175tph to extend the mine life increased residence time by 95%. This however

did not result in the expected higher recoveries.

 18

Further G-Cell pilot plant testwork on the “new” plant final tails or rather extended

roughing time tailings produced average recoveries of 25 % at relatively low mass pulls

indicative of possible still higher recoveries. Results are summarised below in Figure

11.0 and detailed in Appendix 2.0.

Feed Concentrate Tailings Recovery Mass

Pull
Ni % Cu % MgO Ni % Cu %
MgO Ni % Cu % MgO % %
% % %
0.100 0.029 13.48 0.569 0.127 14.41 0.077 0.027 13.40 25.9 4.9

Figure 11.0 - Averaged Results of Final Tailings Testwork After New Tank Cells

Commissioned.

Samples of pilot plant feed, concentrate and tailings were taken and subjected to

Cyclosizer size classification followed by assay of the fractions. The results of these

tests are shown in Figure 12.0 below, and appear to confirm previous Aguablanca

testwork indicating that a large majority of the nickel present in the final plant tails (pilot

plant feed) is in the minus 9µm fraction. A significant proportion of this material is

recovered by the G-Cells. The results are also shown graphically in Figure 13.0.

Calculated Feed (100%) Concentrate (5%) Tailings (95%)

Size Mass Grade Dist Mass Grade Dist Mass Grade Dist
Fraction % % Ni % % % Ni % % % Ni %
+44μm 16.24 0.04 100.00 0.28 0.541 21.33 15.96 0.04 78.67
+33μm 27.93 0.03 100.00 0.19 0.315 6.81 27.74 0.03 93.19
+23μm 12.45 0.03 100.00 0.29 0.138 10.09 12.16 0.03 89.91
+15μm 5.47 0.03 100.00 0.28 0.091 14.47 5.19 0.03 85.53
+11μm 6.56 0.04 100.00 0.48 0.097 20.34 6.08 0.03 79.66
-11μm 31.35 0.16 100.00 3.49 0.544 38.01 27.87 0.11 61.99
Head 100.00 0.07 100.00 5.00 0.444 30.14 95.00 0.05 69.86

Figure 12.0 Table Indicating Percentage Nickel in Each Size Fraction

 19

Ni Recovery by size fraction

40.00

35.00

30.00

25.00
% Ni Recovery

20.00

15.00

10.00

5.00

0.00
-11µm -15+11µm -23+15µm -33+23µm -44+33µm +44µm
Size Fraction

Figure 13.0 Nickel Recoveries by Size Fraction in Concentrate for G-Cell Pilot
Plant Testwork

In December 2010, during the pilot plant testwork, heavy rain in the area caused a major

pit wall slippage at the mine. This halted production and consequently any further pilot

plant testwork at the operation. The plant is currently on care and maintenance. The

mine owners advise that they plan to start production again at the end of 2012.

One of the issues identified with the ability of pneumatic flotation to recover ultra fines

from the final tails is that they would be lost again if put back into the existing tank

flotation cleaner circuit. The benefit of extra ultra fines recovery by such scavaging would

only be realised if it was cleaned in a bespoke Imhoflot cleaner circuit that could produce

a product that could be blended into the final concentrate. This was one of the proposed

tasks of the pilot plant testwork which was put on hold by the mine closure. However
 20

Imhoflot is also efficient in cleaning applications so it thought that a high grade

concentrate would be produced with acceptable impurity limits.

CASE STUDY 2

To further highlight the abilities of pneumatic flotation in recovering ultra-fines a second

pilot plant study is reported here which details results of treating a cleaner circuit tail that

was a final plant tail.

BACKGROUND

This mine exploits one of the words largest zinc/lead deposits. The treatment plant

produces a bulk concentrate with a required minimum zinc grade of 40%, there are

penalties for excessive iron and silica in the concentrate, and with the silica at a

threshold of 3.7% the most important. The finely disseminated nature of the

mineralisation requires an ultrafine grind of below 7 microns to achieve liberation and to

reduce entrainment of silica. This is achieved using six, 1.1 MW ISA mills. Significant

amounts of zinc are lost to tailings from the cleaner circuit, parts of which are open

circuit, as ultra fines, and the objective of this testwork was to determine whether any of

the cleaner tailings streams could be retreated by the G-Cell pilot plant to produce a final

grade concentrate.

Processing Plant
The Processing plant treats approximately 1.8M tons of high grade ore per year. The

head grade averages 12.6% Zinc, 5.5% Lead, and 57ppm Silver, and the plant produces

about 320,000 tons of bulk flotation concentrate per year at an average 45% zinc. The

crushing circuit product feeds the milling circuit which consists of a 4MW SAG mill in

closed circuit with a 935kW Vertimill. The primary cyclones are set to produce a 75µm

product with the underflow feeding the Vertimill. The rougher flotation circuit consists of 4
 21

x 16m3 tank cells, with the concentrate being fed to cyclones with a set point of 7µm, the

overflow of which feeds Cleaner 3. The cyclone underflow is fed to 6 x 1.1 Mw ISA mills

in closed circuit with another 7µm cyclone pack the overflow of which feeds Cleaner 1.

The concentrate from Cleaner 1 is fed to Cleaner 2, the concentrate from which together

with that from Cleaner 3 feeds Cleaner 4.The tailings from Cleaners 1 to 4 report to final

tailings. Cleaners 5 to 7 are in close circuit with the concentrates reporting to the next

cleaner and tailings reporting to the previous one. Cleaner 7 producing the final

concentrate. The zinc grade of the concentrate averages 45% at 13% moisture with a

minimum smelter requirement of 40% this would allow for a degree of blending with G-

Cell concentrates if necessary. The basic plant flowsheet is given in Fig 14.0 below.

Fig 14.0 Flotation flowsheet

 22

PILOT PLANT AND TESTWORK DESCRIPTION

This pilot plant was a two stage self contained unit consisting of two, G10 (1m diameter)

Imhoflot G-Cells. The pilot plant is shown in Figure 15.0. The units operated in series

with the tailings of the first G-Cell feeding the second G-Cell. The pilot plant is able to

treat up to 23 m3/hr. For this testwork the feed percent solids was very low and therefore

the plant operated at around 1 t/hr during the tests. Testwork was performed upon

several of the cleaner streams but results are only reported for the Cleaner 4 tailings

stream as this stream currently reports to final tailings with the results indicating that a

product suitable for blending into the final concentrate can be produced. Zinc recovery

varied from 9.2% to 20.2% with associated grades of 37.9% and 35.6% zinc. The results

are summarised in Figure 16.0.

Figure 15.0 Imhoflot G10 Two Cell Pilot Plant in Operation

 23

With this extra recovery from a final tails a simple payback exercise indicated a payback

of between 4 and 9 months depending upon the assumptions used. The high upgrade of

zinc and downgrade of silica at 3.7 and 2.7 respectively indicate that at least some of the

zinc is found as liberated material, probably ultra fines, as it would be impossible to

achieve these results from middlings. This would also explain the inability of the current

cells to adequately treat this material. It should also be noted that there were no reagent

additions made during these tests.

Zinc SiO2 SiO2

Zinc Feed Recovery Zinc Grade Feed Grade Mass Pull
Conc
(%) (%) Conc (%) (%) (%) (%)
GA1 10.40 9.2 37.91 12.61 4.69 2.5
GA2 9.91 20.2 35.61 13.67 5.16 5.6

Fig 16.0 Selected Results From Cleaner 4 Tailings Stream

The results from individual cells shown below in Figure 17.0 infer that there is not a

significant drop in zinc grade or rise in silicon grade between cells in either test which

indicates that a third cell could be added to increase recovery without seriously affecting

the combined grade. It is also likely that a third cell would enable the mass pulls to be

better ballanced to optimise the grade recovery relationship.

Cell 1 Cell 2

Test Zinc Zinc SiO2 Zinc Zinc grade SiO2

Recovery grade grade Recovery Conc % grade
% Conc % Conc % % Conc %

GA1 5.7 38.1 4.7 3.5 37.6 4.7

GA2 13.9 36.3 4.9 6.3 34.2 5.7

Fig 17.0 Breakdown of Results on an Individual Cell Basis

 24

CONCLUSIONS

The testwork case studies confirmed the well documented inefficiency of conventional

tank flotation in the recovery of ultrafine particles. In contrast Imhoflot G-Cell pneumatic

flotation at Aguablanca was able to recover up to 30% of these fine particles from

existing plant tailings. It did this from plant tailings that had already been processed by

two different manufacturer’s types of tank flotation cells with an inordinate amount of

excess flotation residence time in the circuit.

Whilst the above recovery improvements alone may provide sufficient justification for the

consideration of G-Cells for new base metal flotation circuits, additional benefits may

accrue for any flotation circuit where the valuable component is poorly liberated when G-

Cells are combined with ultra fine grinding. Further testwork is still required to confirm

this, however given their small footprint, high throughput and other operational benefits

Imhoflot G-Cell technology clearly represents a major breakthrough in flotation plant

circuit design.

References

Lynch, A J., Harbort, G. J., Nelson, M, G., (2010), History of Flotation, The Australasian

Institute of Mining and Metallurgy, ISBN 978 1 921522 25 3

Pease, J. D., Young, M. F., Curry, D., Johnson, N. W., Improving Fines Recovery by

Grinding Finer, MetPlant Conference 6-7 September 2004, Perth, Australia

 25

Brown, J., Imhof, R. M., Lotzien, R., (2001), Self-aspirating Aeration Reactors for

Pneumatic Flotation and other Applications, IX Balkan Mineral Processing Congress 11-

13 September 2001, Istanbul, Turkey

Imhof , R.M., Brown, J.V., (2000). Imhoflot – Evolution of pneumatic flotation. Major

Trends in Development of Sulfide Ores Up-Grading in the 21st Century Conference

Proceedings , 24-28 April 2000, Norilsk, Russia.

Imhof, R.M., Lotzien, R., Sobek, S. (1993). Pneumatic flotation: a reliable procedure for a

correct plant layout. XVIII International Mineral Processing Congress Sydney, Australia

23-28 May 1993 pp 971-978.

Imhof, R.M., Battersby, M.J.G., Brown, J.V., Lotzien, R.M., Kleefeld, J., (2003).

Development of Pneumatic Flotation Incorporating Centrifugal Separation. XXII

International Mineral Processing Congress, 28 October-3 September 2003, Cape Town,

South Africa.

Battersby, M.J.G., Fletcher, M.G., Imhof, R.M., Singh, A.A., Puder, F., (2005). The

Advantages of the Imhoflot G-Cell Pneumatic Flotation Process with Centrifugal Froth

Removal – Two Case Studies. Randol Innovative Metallurgy Forum, 21 – 24 August

2005, Perth, Australia.

Sanchez-Pino, S., Imhof, R.M., Sanchez-Baquedano, S., Rojos-Tapia, F., Fernandez-

Garcia, M., (2003) Pneumatic Flotation Technology – An Interesting Experience in the

Chilean Mining Industry. CIM 35th Canadian Mineral Processors Operators Conference,

21-23, January 2003. Ottawa, Canada

Gonzalez, L.R,. (2002). Pneumatic flotation study as a pre rougher cell at the

molybdenum plant. Technical report. Chuquicamata, Chile

 26

Sanchez-Pino, S., Sanchez-Baquedano, A., Imhof, R.M., (2008). Imhoflot G-Cell

Pneumatic Flotation Technology for Fine Particle Applications. PROCMIN 2008 V

International Mineral Processing Seminar, 22-24 October 2008, Santiago, Chile.

Sanchez-Pino, S., Sanchez-Baquedano, A., Imhof, R.M., Battersby, M.J.G., (2009). Pilot

Plant Testing and the Successful Commercial Installation of Imhoflot G-Cell Pneumatic

Flotation for Molybdenum Concentrate Cleaning at Los Pelambres in Chile. Flotation09

MEI Conference, 9-12 November 2009, Cape Town, South Africa.

Hay, M (2009). Brief evaluation of the Imhoflot G-Cell, Technical report Johannesburg,

South Africa.
Appendix 1.0 Sampling Campaign 1. G-Cell Pilot Plant Flotation of Existing Aguablanca Plant Tailings

RUN Feed Concentrate Tailings Recovery Mass

Ni % pull %
NUMBER solids Ni % Cu % MgO Solids Ni % Cu % MgO% Solids Ni % Cu % MgO% Ni % Pull %
% % % %
RUN 1 36.9 0.109 0.039 11.25 19.4 1.385 0.263 6.66 31.0 0.060 0.024 11.78 47.0 3.7
RUN 2 35.1 0.116 0.036 11.13 15.1 1.167 0.182 8.27 30.7 0.046 0.016 11.37 62.8 6.2
RUN 3 38.4 0.071 0.025 10.20 10.3 1.592 0.232 5.84 36.1 0.061 0.023 10.26 14.6 0.7
RUN 4 37.1 0.068 0.031 11.11 8.8 1.282 0.200 8.64 35.9 0.057 0.032 10.63 16.9 0.9
RUN 5 37.0 0.118 0.042 11.56 20.5 1.392 0.213 8.60 31.5 0.091 0.065 11.51 24.5 2.1
RUN 6 37.2 0.087 0.039 11.49 12.3 1.377 0.262 11.85 33.1 0.066 0.032 11.49 24.4 1.6
RUN 7 33.8 0.085 0.029 10.95 14.2 0.964 0.137 8.24 24.0 0.070 0.025 11.16 19.0 1.7
RUN 8 34.6 0.078 0.042 12.31 12.0 1.082 0.259 11.6 39.8 0.067 0.037 11.45 15.0 1.1
RUN 9 34.9 0.132 0.047 11.68 20.2 1.181 0.263 9.07 31.3 0.104 0.036 11.69 23.3 2.6
RUN 10 35.3 0.074 0.033 11.49 13.1 0.718 0.191 12.82 35.6 0.060 0.031 11.47 20.6 2.1
RUN 11 39.4 0.107 0.032 10.81 13.4 0.983 0.161 8.43 34.9 0.083 0.026 10.93 24.5 2.7
RUN 12 35.3 0.085 0.039 8.46 12.2 1.343 0.237 6.39 34.6 0.064 0.034 11.01 25.9 1.6
RUN 13 36.3 0.094 0.048 11.15 15.0 1.264 0.251 6.73 36.3 0.068 0.042 10.49 29.2 2.2
RUN14 37.3 0.077 0.036 11.40 13.8 1.256 0.257 7.23 37.1 0.053 0.035 10.76 32.5 2.0
RUN 15 39.3 0.077 0.031 10.04 15.1 1.079 0.164 10.23 34.0 0.037 0.020 10.33 53.8 3.8
RUN 16 32.7 0.101 0.032 10.21 16.0 1.268 0.154 6.54 30.9 0.081 0.029 10.64 21.2 1.7
RUN 17 31.2 0.125 0.042 10.19 13.7 1.347 0.242 6.61 28.3 0.093 0.038 10.98 27.5 2.6
RUN 18 34.2 0.122 0.042 10.80 18.5 1.331 0.165 5.82 29.7 0.069 0.035 11.52 45.8 4.2
RUN 19 31.4 0.097 0.033 11.45 23.8 1.236 0.177 9.87 29.5 0.079 0.029 11.11 19.8 1.6
RUN 20 29.7 0.141 0.039 11.27 15.4 1.282 0.149 8.54 27.3 0.085 0.025 11.74 42.5 4.7
RUN 21 35.5 0.136 0.041 11.76 17.9 1.021 0.168 10.00 32.0 0.112 0.033 12.23 19.8 2.6
RUN 22 31.1 0.139 0.040 10.10 19.8 1.438 0.196 7.68 28.9 0.077 0.033 10.00 47.1 4.6
Average 35.2 0.102 0.037 10.95 15.5 1.227 0.206 8.44 32.4 0.072 0.032 11.12 29.9 2.6
 28

Appendix 2.0 Sampling Campaign 2. G-Cell Plant Flotation of Aguablanca Tails After Extension of Plant Roughing Time

RUN No Feed Concentrate Tailings Recovery Mass

Ni % Pull Ni
%
Solids Ni % Cu % MgO % Solids Ni % Cu % MgO % Solids Ni % Cu % MgO % Ni % Pull %
% % %
RUN 1 32.7 0.101 0.032 10.21 16.0 1.268 0.154 6.54 30.9 0.081 0.029 10.64 21.2 1.7
RUN 2 31.2 0.125 0.042 10.19 13.7 1.347 0.242 6.61 28.3 0.093 0.038 10.98 27.5 2.6
RUN 3 34.2 0.122 0.042 10.8 18.5 1.331 0.165 5.82 29.7 0.069 0.035 11.52 45.8 4.2
RUN 4 31.4 0.097 0.033 11.45 23.8 1.236 0.177 9.87 29.5 0.079 0.029 11.11 19.8 1.6
RUN 5 29.7 0.141 0.039 11.27 15.4 1.282 0.149 8.54 27.3 0.085 0.025 11.74 42.5 4.7
RUN 6 35.5 0.136 0.041 11.76 17.9 1.021 0.168 10.0 32.0 0.112 0.033 12.23 19.8 2.6
RUN 7 31.1 0.139 0.040 10.10 19.8 1.438 0.196 7.68 28.9 0.077 0.033 10.00 47.1 4.6
Average 35.2 0.102 0.037 10.95 15.5 1.227 0.206 8.44 32.4 0.072 0.032 11.12 29.9 2.6