INVESTIGATION INTO THE THERMAL UPGRADING OF

NICKELIFEROUS LATERITE ORE

By

Filipe Manuel Rodrigues

A thesis submitted to the Robert M. Buchan Department of Mining

In conformity with the requirements for

The degree of Master of Applied Science

Queen’s University

Kingston, Ontario, Canada

(December, 2013)

Copyright © Filipe Manuel Rodrigues, 2013

 Abstract

Nickeliferous laterite ores are currently processed using complex energy intensive

flowsheets. Limited mineral upgrading can be achieved by low-cost mineral processing as the

nickel is not found as a separable mineral phase but finely disseminated throughout the host

goethite mineral. Whole ore extraction processes are required which result in intrinsically higher

capital and operating costs. Market pressure has provided incentives to develop alternative

upgrading techniques that can produce a nickel concentrate and reduce the material input to

downstream processing facilities. Thermal upgrading through a selective reduction mechanism to

produce a ferronickel concentrate has been studied extensively and has shown promising

potential. In this research, a two stage selective reduction of nickeliferous laterite ore was

investigated at 600oC and 1000 – 1100oC with varying coal and sulphur additions.

Experiments showed that the limonite ore could be selectively reduced using a coal

additive to a ferronickel and wustite phase. A combination of XRD and bromine/methanol

diagnostic leach tests confirmed the presence of metallic nickel and iron in the calcine. Higher

degrees of metallization corresponded with higher sulphur additions and growth zone

temperatures. Sulphur was added to improve particle growth through the establishment of a Fe-O-

S liquid phase, which was found to improve Ni recovery from 13.8% to 75.8% over the range of

0 – 4 wt% S.

Ferronickel particles ranging in size from 20 – 60 microns were shown to be present but

highly dispersed throughout the upgraded ore. Particle growth improved with higher growth zone

temperatures and longer retention times. Magnetic separation of the calcine showed maximum

upgrading of grades to 3 – 4 wt% nickel with recoveries ranging from 83.7 – 93.2%. Partial

oxidation of wustite particles to magnetite caused the particles to be magnetic and resulted in

recovery of unwanted iron oxides. The presence of iron oxide fines was believed to allow for

ii
rapid oxidation of wustite phase and also produce slimes that hindered physical separation of the

upgraded ore.

iii
 Acknowledgements

I would like to thank my supervisor Dr. John Peacey for his mentorship, support and patience

throughout my graduate studies. Under his guidance I have matured both professionally and

personally.

I would like to thank in particular Martitza Bailey for her guidance and assistance in and out of

the laboratory. Thank you to Dr. Chris Harris for his insightful comments and suggestions. The

individuals at KPM are thanked for the time and resources they provided. To me fellow graduate

students I thank you for your ongoing support.

Finally, I would like to thank my family and friends for their continuous support and

encouragement. You continue to be my inspiration and driving force behind any endeavor I

choose to face.

Funding for this research was provided by NSERC and Xstrata.

iv
 Table of Contents

Abstract ...................................................................................................................................... ii
Acknowledgements .................................................................................................................... iv
List of Figures........................................................................................................................... vii
List of Tables ............................................................................................................................. ix
Chapter 1 : Introduction .............................................................................................................. 1
1.1 General Overview ............................................................................................................. 1
1.2 Nickel from Laterites ......................................................................................................... 1
1.3 Current Processing Options ............................................................................................... 4
1.4 INCO Patent for the Thermal Upgrading of Lateritic Ores ................................................. 5
Chapter 2 : Literature Review ..................................................................................................... 8
2.1 Nickel Laterite Deposits .................................................................................................... 8
2.2 Beneficiation Difficulties with Laterite Ores ...................................................................... 9
2.3 Current Processing Methods of Nickel Laterites .............................................................. 10
2.3.1 Pyrometallurgical Processes ...................................................................................... 11
2.3.1.1 RKEF Process .................................................................................................... 11
2.3.1.2 Nippon Yakin Oheyama Process ........................................................................ 12
2.3.1.3 Nickel Pig Iron................................................................................................... 13
2.3.2 Hydrometallurgical Processes ................................................................................... 13
2.3.2.1 Caron Process .................................................................................................... 14
2.3.2.2 HPAL Process ................................................................................................... 14
2.3.2.3 Atmospheric Leaching (AL) and Heap Leaching (HL) ....................................... 15
2.4 Examples of Innovative Upgrading Techniques ............................................................... 17
2.4.1 Thermal Upgrading................................................................................................... 17
2.4.2 Selective Sulphidation .............................................................................................. 20
2.4.3 Segregation Process .................................................................................................. 21
2.5 Reduction of Nickel Laterites .......................................................................................... 23
2.5.1 Solid State Reduction Mechanism ............................................................................. 23
2.5.2 Reducibility of Nickeliferous Laterite Ores ............................................................... 25
2.6 Ferronickel Particle Growth ............................................................................................. 32
2.6.1 Effect of Temperature ............................................................................................... 33
2.6.2 Effect of Sulphur ...................................................................................................... 34
v
Chapter 3 : Methods and Materials ............................................................................................ 36
3.1 Introduction..................................................................................................................... 36
3.2 Ore Characterization........................................................................................................ 36
3.3 Ore Preparation and Pellet Formation .............................................................................. 38
3.4 Stationary Bed Furnace Experiments ............................................................................... 40
3.5 Rotary Kiln Experiments ................................................................................................. 41
3.6 Magnetic Separation Tests ............................................................................................... 43
3.7 Analytical Techniques ..................................................................................................... 45
Chapter 4 : Results and Discussion ............................................................................................ 48
4.1 Preliminary Analysis of Thermal Upgrading Requirements .............................................. 48
4.2 Degree of Reduction ........................................................................................................ 48
4.3 Bench-Scale Testing of the Thermal Upgrading Process .................................................. 53
4.3.1 Thermal Upgrading Process Test Work using Stationary Bed Furnace....................... 54
4.3.2 Thermal Upgrading Process Test Work using Rotary Kiln ........................................ 64
4.4 Thermogravimetric and Differential Thermal Analysis .................................................... 66
4.5 Ferronickel Particle Growth ............................................................................................. 75
Chapter 5 : Conclusions and Future Work ................................................................................. 91
5.1 Conclusions ..................................................................................................................... 91
5.2 Future Work .................................................................................................................... 94
Bibliography ............................................................................................................................. 96
Appendix A - US Patent 5,178,666 ......................................................................................... 102
Appendix B - Examples of Stoichiometric Reduction Calculations .......................................... 108
Appendix C – Raw Data Tables .............................................................................................. 110
Appendix D - TGA/DTA Curves for Coal and Sulphur Additives ............................................ 119

vi
 List of Figures

Figure 1: World's land based nickel resources and primary nickel production (Dalvi, Bacon, &
Osborne, 2004) ........................................................................................................................... 2
Figure 2: World nickel laterite resources (Dalvi, Bacon, & Osborne, 2004) ................................. 2
Figure 3: Fraction of nickel production by ore type (Mudd, 2010) ............................................... 3
Figure 4: Flowsheet diagram of U.S. Pat. No. 5,178,666 pilot plant process (Diaz, et al., 1993) ... 6
Figure 5: Range of laterite profiles (Dalvi, Bacon, & Osborne, 2004) .......................................... 9
Figure 6: Simplified process flow diagrams for current commercial production methods for the
extraction of nickel from laterite ores ........................................................................................ 11
Figure 7: Fe-Ni-O stability diagram for goethite containing 1.2% Ni, reduction isopleths labelled
with percent values join points of constant degree of Ni metallization (Hallet, 1997) ................. 27
Figure 8: Phase Equilibria for the reduction of nickel and iron oxides with carbon (Canterford,
1975) ........................................................................................................................................ 28
Figure 9: Diagram of metal agglomeration process via suspension model (Watanabe, Ono, Arai,
& Matsumori, 1987).................................................................................................................. 33
Figure 10: SEM images of Fe-Ni and Fe-S phases: (2) and (A) Fe-S, (1) and (B) Fe-Ni (Zhu, et
al., 2012) and (Li, Shi, Rao, Jiang, & Zhang, 2012) ................................................................... 35
Figure 11: X-ray spectrum of limonite ore ................................................................................. 37
Figure 12: Pressed 3g limonite pellet ......................................................................................... 40
Figure 13: Experimental tube furnace setup diagram ................................................................. 41
Figure 14: Schematic diagram of rotary kiln setup ..................................................................... 42
Figure 15: Quartz capsule containing 300g of pelletized feed at 6% coal, 4% S for rotary kiln
experiment ................................................................................................................................ 43
Figure 16: Magnetic separators used in experiments: (left) HIWMS and (right) Davis tube ........ 44
Figure 17: XRD pattern of upgraded ore.................................................................................... 49
Figure 18: XRD pattern of upgraded ore using Co tube ............................................................. 50
Figure 19: Effect of magnetic field strength of HIWMS on the mass fraction of magnetics
extracted ................................................................................................................................... 57
Figure 20: Effect of coal addition on recovery and grade of nickel in magnetic concentrate, 10 g
samples with 4% sulphur........................................................................................................... 62
Figure 21: Effect of sulphur addition on recovery and grade of nickel in magnetic concentrate, 10
g samples with 6% coal ............................................................................................................. 64
Figure 22: TGA/DTA curves of limonite ore up to 1200oC under N2 atmosphere ....................... 67

vii
Figure 23: Comparison of mass loss curves for blends of limonite ore with reagents, roasted up to
1000oC at 20oC/min under N2 atmosphere, blends of ore - coal - sulphur at 2:1:1 weight ratio
respectively............................................................................................................................... 68
Figure 24: TGA/DTA curves for blend of 2:1:1 weight ratio of ore - coal - sulphur, roasted up to
1000oC under N2 atmosphere..................................................................................................... 69
Figure 25: TGA/DTA curves for upgrading limonite ore with coal and sulphur using simulated
U.S. Pat. No. 5,178,666 temperature profile .............................................................................. 71
Figure 26: TGA mass loss curves comparing growth zone temperatures for 3g limonite pellet with
6% coal, 4% sulphur ................................................................................................................. 73
Figure 27 TGA mass loss curves comparing growth zone retention times for 3g limonite pellet
with 6% coal, 4% sulphur ......................................................................................................... 74
Figure 28: Backscattered image of ferronickel particle produced in stationary tube furnace ....... 76
Figure 29: EDS scans of FeNi particle and surrounding oxide gangue shown in Figure 28 ......... 77
Figure 30: Backscattered image of pellet surface upgraded at 1000 oC growth zone temperature. 78
Figure 31: Backscattered image of ferronickel pellet upgraded in rotary kiln with corresponding
EDS scan, particle displays signs of liquid phase sintering ........................................................ 80
Figure 32: SEM images of pellets upgraded at growth zone temperatures of 1000, 1100, 1200 and
1300oC...................................................................................................................................... 82
Figure 33: Backscattered electron image of sectioned pellet upgraded at 1300 oC growth zone
temperature, large continuous FeNi particles are present throughout the sample ........................ 84
Figure 34: SEM images at 500x magnification of thermally upgraded pellets with various hold
times at 1000oC with 6% coal, 4% S.......................................................................................... 86
Figure 35: SEM image of FeNi particles at 5000x magnification, produced at various hold times
at 1000oC with 6% coal, 4% S ................................................................................................... 87
Figure 36: SEM image of pellet upgraded at growth temperature of 1200 oC; (A) FeNi particle,
(B) iron silicate, (C) iron oxide matrix ....................................................................................... 89
Figure 37: EDS scans of Figure 36 phases; (A) FeNi particle, (B) iron silicate, (C) iron oxide
matrix ....................................................................................................................................... 90

viii
 List of Tables

Table 1: Current processing options for laterite ore...................................................................... 4

Table 2: Thermal upgrading patents .......................................................................................... 18
Table 3: Example 1 bench test results from U.S. Pat. No. 5,178,666 (Diaz, et al., 1993) ............ 20
Table 4: Experimental test work results for the segregation process ........................................... 22
Table 5: Head grade analysis of limonite ore samples and INCO patent samples ....................... 37
Table 6: Estimated mineral composition of limonite ore ............................................................ 38
Table 7: Chemical analysis of coal and INCO patent bituminous coal ........................................ 39
Table 8: Degree of metallization of upgraded ore determined by bromine/methanol diagnostic
leach ......................................................................................................................................... 51
Table 9: Preliminary upgrading test work with stationary bed furnace setup and Davis tube
magnetic separator, TF-1 to TF-4 pre-calcined ore, TF-5 to TF-8 raw ore .................................. 55
Table 10: Thermally upgraded 3g pellets with 6% coal at a thermal upgrading temperature of
1000oC and separated using HIWMS ........................................................................................ 58
Table 11: Multiple 15g samples with 6% Coal, 4% S at thermal upgrading temperature of 1000 oC
separated using HIWMS ........................................................................................................... 60
Table 12: Magnetic separation results of thermally upgraded ore using rotary kiln setup, charge
feed of 300g at 6% coal, 4% sulphur ......................................................................................... 65

ix
 Chapter 1: Introduction

1.1 General Overview

The purpose of this thesis was to investigate the thermal upgrading of a limonite ore.

Nickeliferous or limonite ores contain economical values of nickel finely disseminated

throughout the host mineral goethite (FeOOH), a hydrated ferric iron oxide. Due to the solid

substitution of nickel within the goethite matrix, traditional beneficiation processes are unable to

upgrade laterite ores. The whole limonite ore must therefore be treated by expensive

hydrometallurgical methods to extract the nickel and produce a saleable nickel product, either as

Ni hydroxide, sulphide or metallic Ni. Since the discovery of sulphide Ni deposits are becoming

more difficult, attention is shifting to processing the more abundant but difficult-to-process

lateritic Ni ores. Increasing supply of nickel from laterite deposits has provided motivation to

develop an alternative upgrading technique. The following sections will overview the demand for

nickel from laterite projects, current processing methods and an alternative process proposed by

INCO in U.S. Pat. 5,178,666 on the thermal upgrading of laterite ores.

1.2 Nickel from Laterites

According to Dalvi et al (2004) and Mudd (2010) , 70% of land based nickel resources

are contained within laterite deposits with the remaining 30% present in sulphide deposits (Figure

1). Two thirds of the laterite fraction are composed of the nickeliferous laterite formations

(limonite, nontronite/smecitite) while the remaining third is comprised of magnesium silicate

laterites (saprolite, garnierite). The breakdown of reported laterite resources throughout the world

is shown in Figure 2. As of 2004 world nickel laterite resources were 1.26 Mt averaging at a

1.28% Ni grade and containing 161 Mt of nickel (Dalvi, Bacon, & Osborne, 2004).

1
Figure 1: World's land based nickel resources and primary nickel production (Dalvi, Bacon, &
Osborne, 2004)

Figure 2: World nickel laterite resources (Dalvi, Bacon, & Osborne, 2004)
2
 Due to the relative simplicity of upgrading sulphide ores by flotation to produce a nickel

sulphide concentrate that can be processed using established metallurgical methods, the majority

of nickel production has been derived from sulphide ores. Many sulphide mines are reaching the

end of production life as grades are falling, reserves are dwindling and rising operating costs are

limiting mine activity (Kuck P. H., 2012). The replenishment rate of sulphide reserves has lagged

behind the depletion rate with new discoveries being found in non-traditional mining districts that

are more remote, technically complex and thus harder to finance (Dalvi, Bacon, & Osborne,

2004; Kuck P. H., 2013). Demand for nickel is increasing at 4% per year mainly due to

expansion in the stainless steel capacity of China (Dalvi, Bacon, & Osborne, 2004). The new

demand is expected to be supplied from new laterite production. Figure 3 shows the fraction of

global nickel production from laterite ores has been increasing since 1950 and as of 2007 nickel

production from laterite ores (856.5 kt) surpassed nickel production from sulphide ores (817.2 kt)

(Mudd, 2010). New Greenfield laterite projects will be required to meet the increase in demand.

Figure 3: Fraction of nickel production by ore type (Mudd, 2010)

3
 Greenfield laterite projects are difficult to finance as typical capital costs range from 10 –

15 $/lb Ni for smelters and 12 – 16 $/lb Ni for High Pressure Acid Leach (HPAL) (Dalvi, Bacon,

& Osborne, 2004). These projects must be associated with massive deposits to reach economies

of scale in order to be developed. To reduce capital and operating costs of Greenfield projects,

new technologies should be developed to allow smaller projects to become economical.

1.3 Current Processing Options

The main difference between processing of laterite and sulphide ores is the ability of

sulphide ores to be upgraded through traditional beneficiation processes to a concentrate (10 -

26% Ni) while only limited upgrading (by factor < 2) can be applied to laterite ores. Limited

upgrading results in larger tonnages of feed material to be processed and produces larger volumes

of waste slag or tailings material (Dalvi, Bacon, & Osborne, 2004). Three commercial processing

options are currently utilized and are shown in Table 10.

Table 1: Current processing options for laterite ore

Process Feed Ore Recoveries Description

RKEF Saprolite Ni 90 – 95% Drying/dehydroxylation at 250oC followed
by reduction or sulphidation at 800 – 900oC
Co 50%
in rotary kilns. Electric arc smelting (up to
1600oC) separates slag from either
ferronickel or nickel matte phase
Caron Limonite/Saprolite Ni 80 - 85 % Selective reduction of Ni/Co values to the
blends metallic state at 750oC followed by
Co 40 – 45 %
ammoniacal leach to recover the Ni/Co.
Further refining produces either a Ni/Co
carbonate or oxide sinter
HPAL Limonite Ni > 90 % Pressure acid leach (at 250oC and 40 MPa) in
autoclaves with sulphuric acid followed by
Co > 90%
CCD and purification

4
 These processes along with other niche flowsheets are described in further details within

this thesis. Nickel production from laterites is more energy intensive and associated with higher

levels of greenhouse gas emissions when compared to sulphide ores (Mudd, 2009). Laterite

projects are further complicated by higher capital costs for infrastructure and equipment

associated with the remoteness of the ore bodies and complexity of the extractions methods. As

the nickel industry trends towards higher production volumes from laterite ores, a need for new

technologies becomes apparent to address the metallurgical and environmental problems

associated with laterite processing, in order to ensure the viability of the industry (Mudd, 2010).

For new laterite technologies to be successful they need to focus on reducing the capital costs of

the project and produce an intermediate product similar to a sulphide concentrate that can be

easily transported to and treated by existing smelter/refinery facilities (Neudorf & Huggins,

2004). The present work investigated low temperature thermal upgrading of laterite ores, an

emerging new technology for processing laterite ore.

1.4 INCO Patent for the Thermal Upgrading of Lateritic Ores

U.S. patent 5,178,666 filed by INCO outlined a novel process for the thermal upgrading

of limonite and limonite/saprolite blends to a ferronickel concentrate that is amenable to further

refining via smelting or leaching techniques. The design calls for the selective reduction of

nickel, cobalt and some iron, up to a maximum temperature of 1050oC, to form a ferro-nickel

alloy while the bulk of the iron content is reduced to the non-magnetic wustite phase. Magnetic

separation can then separate the ferro-nickel alloy from the wustite gangue. The patent outlined

the process through a series of bench scale experiments that determined the optimal operating

parameters which were then implemented in a 500kg/d pilot plant. The process flow sheet for the

pilot plant can be seen in Figure 4. To the author’s knowledge this patent was never developed at
5
the commercial scale. Compared to other thermal upgrading patents, which focus solely on

higher grade saprolite, this patent focuses on upgrading low grade limonite ores with minimal

reagent additions, lower processing temperatures and shorter resident times. The process,

according to the results given in the patent, was successful in producing high nickel recoveries to

a concentrate grading 8 – 10% Ni.

Figure 4: Flowsheet diagram of U.S. Pat. No. 5,178,666 pilot plant process (Diaz, et al., 1993)

The pilot plant process begins with mixing of the feed ore with 4 – 6% bituminous coal

and 2 – 4% elemental sulphur. The mixed feed is then pelletized to improve solid-solid reduction

reactions and reduce losses through dust formation. The pellets are charged to a counter-

currently fired rotary kiln held under neutral or silently reducing conditions. The pellets undergo

drying, dehydroxylation and selective reduction of nickel and cobalt values to the metallic states

while the bulk of the iron content is reduced to the wustite phase. The pellets are then held in a

metallic growth zone set at 1050oC for one hour to allow for the liquid phase migration of

6
metallics which agglomerate into ferronickel particles. The temperature throughout the kiln is

maintained below the point at which the laterite ore becomes “sticky” and forms sintered

accretions on the furnace walls. A reducing atmosphere is maintained within the growth zone

through the partial combustion of natural gas at 62 – 63% aeration. The discharge from the kiln is

rapidly cooled in an oxygen deficient environment to prevent the oxidation of the wustite phase to

magnetite. Possible wustite disproportionation reaction may occur if a high temperature is

maintained to improve kinetics of the reaction. Wustite is a non-magnetic iron oxide phase. The

cooled pellets are ground and magnetically separated to form a ferronickel concentrate and

tailings made up of wustite and magnesium silicates. Recoveries and grades of nickel achieved

ranged from 80 – 94% and 8.5 – 9.5wt% respectively.

The scope of this work focused on the thermal upgrading via selective reduction of a

limonite ore using the experimental bench scale test work performed in the patent as a starting

point. A copy of the patent can be found in Appendix A of this thesis. The bench scale results in

terms of recovery, grade and gangue rejection indicate that the patent was a viable upgrading

technique for limonite ores. Therefore the goal of this thesis was:

 To study the role of temperature, retention time and sulphur addition on the

promotion and growth of ferronickel particles to improve Ni recovery and grade of

the concentrate from the specific limonite ore used

 To improve the upgrading process through the implementation of various process

modifications

7
 Chapter 2: Literature Review

A literature review was conducted to formulate an understanding of nickel laterite

deposits and the current processing difficulties that arise due to the nature of the ore. Physical

upgrading and current whole ore treatments processes were reviewed to illustrate the challenges

facing nickel laterites processing and the need for new techniques. Examples of innovative

upgrading processes currently being investigated such as thermall upgrading, selective

suphidation and the segregation process were illustrated through a review of published journal

articles and patents. Studies on the selective reduction of nickel laterites in terms of mechanism,

reducibility and ferronickel particle growth were also reviewed. The information gathered from

the literature review was used a basis for the experimental work performed in this study.

2.1 Nickel Laterite Deposits

Nickel laterite deposits occur in tropical to sub-tropical regions where in the past

prolonged weathering of ultramafic peridotite rocks containing ferro-magnesium silicate minerals

(olivine and pyroxene) occurred. The process of weathering or laterization produces profiles with

great variability in thickness, grades, chemistry and ore mineralogy (Dalvi, Bacon, & Osborne,

2004). The range of layered laterite formations along with average metal assays for each layer can

be seen in Figure 5. Boldt (1967) and Alcock (1988) describe the laterization process that

produces the characteristic zoned profile of a laterite deposit. The host peridodtite rock is

dissolved by CO2 enriched groundwater and organic acids from decomposing vegetation which

solubilizes the contained magnesium, iron, nickel and silica. While still near the surface, the

dissolved iron rapidly oxidizes and precipitates out as goethite and hematite. The remaining

dissolved material continues to percolate down the profile until the acidic conditions are

neutralized by the host rock precipitating out the nickel, silica and some of the magnesium as
8
hydrated silicates forming the saprolite layer. Varying weathering conditions may result in

substantial quantities of nickel remaining in the upper region forming the limonite layer, enriched

with iron and depleted of magnesium and silica. Limonite layers also contain economic values of

cobalt and chromium. As a product of weathering, the laterite ore resembles a soft clay material

found 10 - 40m below the surface and can therefore be mined using earth moving techniques,

such as power shovels and draglines (Betteridge, 1977; Moskalyk & Alfantazi, 2002).

Figure 5: Range of laterite profiles (Dalvi, Bacon, & Osborne, 2004)

2.2 Beneficiation Difficulties with Laterite Ores

With nickel laterite ores, the contained nickel is not present in a separate mineral grain

but is finely disseminated throughout the host material. The nickel and cobalt values are

substituted in for iron in the goethite matrix of limonite ores. Low grades and the variable

distribution of Ni throughout the ore body limit the effectiveness of physical beneficiation

processes (De-quing, Yu, Hapugoda, Vining, & Jian, 2012). Onodera et al (1987) conducted a

9
comprehensive study to determine the effectiveness of upgrading laterite ores using standard

mineral beneficiation techniques. The ores were subjected to a range of upgrading techniques

including sink-and-float, magnetic, electrostatic separation and froth flotation test. The test work

found that laterite ores were not amenable to any of the upgrading techniques. Current

commercial applications of laterite upgrading is limited to selective screening of less weathered

fragments which contain less nickel, rejecting a small fraction of the gangue material (Habashi,

1997). Current metallurgical processing options must treat whole ore feed resulting in elevated

energy, reagent, and transportations costs while also producing significant quantities of waste

material.

2.3 Current Processing Methods of Nickel Laterites

Reviews of the current processing methods for laterite ores have been conducted by Dalvi et al

(2004), Habashi (1997), Crundwell et al (2011), Canterford (1975) and Simons (1988).

Metallurgical processing of laterites can be divided into pyrometallurgical and hydrometallurgical

treatments. The development of laterite projects implementing these processes are affected by

low grades with limited upgrading ability, complex processing technology, geographically remote

and large infrastructure requirements such as; power plants, ports, roads and tailings/slag

containment facilities. As a result laterite projects must be “mega projects” to achieve economies

of scale (Neudorf & Huggins, 2004; Moskalyk & Alfantazi, 2002; King, 2005). Figure 6 below

illustrates simplified flow sheets for the currently practiced commercial processes. Further

description of the processes and associated drawbacks are provided in the sections to follow.

10
 Figure 6: Simplified process flow diagrams for current commercial production methods for the
extraction of nickel from laterite ores

2.3.1 Pyrometallurgical Processes

The higher grade saprolite zone, containing nickel bearing magnesium silicate minerals

such as serpentine and garnierite are treated using pyrometallurgical methods. Limonite ore is

considered an overburden and is rejected as feed material due to lower nickel grades and

significantly higher iron content.

2.3.1.1 RKEF Process

Moisture and chemically bound water ranging from 30 – 45% by weight contained within

the ore must first be removed. Direct fired rotary kilns, typically running at 250 oC, reduce the

11
contained water content to 15 – 20%. Counter-currently fired rotary kilns running at 800 – 900 oC

then dehydroxylate and calcine/reduce or sulphidize the ore, pre-conditioning the material before

smelting in an electric arc furnace up to 1600oC. Smelting produces either ferronickel through

selective reduction of oxide phase or nickel matte via sulphidizing conditions. Of the laterite

smelting operations, 90% produce ferronickel while 10% produce nickel matte (Crundwell,

Moats, Ramachandran, Robinson, & Davenport, 2011). Furthering processing of

ferronickel/matte produces marketable nickel products and the overall recovery of nickel and

cobalt in the RKEF process is 90 – 95% and 50% respectively. Though recoveries are high,

energy requirements for the RKEF process are extensive since both hydrocarbons and electrical

energy are used. The high throughput volume because of the treatment of whole ore adds to

energy and material costs.

2.3.1.2 Nippon Yakin Oheyama Process

Watanabe et al. (1987) describes the Nippon Yakin Oheyama Process for the direct

reduction of garnierite ore for the production of FeNi in a rotary kiln. Silicate ore (2.3- 2.6 %

Ni, 12 – 15% Fe) along with anthracite, coke breeze and limestone are ground and briquetted

together. The briquettes are fed to a coal fired countercurrent rotary kiln with a temperature

gradient of 700 – 1300 oC. Through the kiln, the charge is dried, dehydrated, reduced and

smelted in a semi-fused state to form coalesced particles of spongy ferro-nickel called luppen.

The discharge is quenched in water and the luppen particles, ranging in size from 2 – 3 mm and

with a grade of 22% Ni, 0.45% Co, are separated from the slag through a grinding, screening,

jigging and magnetic separation circuit. Initial nickel recovery was only 80% due to the high

content of gangue material in the feed that is difficult to process using only a rotary kiln. Nickel

recovery was increased to 95% with improvements to flue dust recycling, pretreatment
12
technology and segregation of luppen product. The Nippon Yakin Oheyama Process flowhsheet

exhibits similar properties to the purposed thermal upgrading patent for limonite ores studied in

this work. The two processes may be compared to assess the feasibility of the proposed patent.

2.3.1.3 Nickel Pig Iron

Nickel Pig Iron (NPI) production from laterite ores in China began in 2006/2007 due to

extremely elevated LME nickel prices (Cartman, 2012). Modified iron blast furnaces and electric

arc furnaces are used to produce a low grade (10 – 15% Ni) ferro-nickel product, which is used as

an alternative to LME nickel for stainless steel production. NPI production follows traditional

iron ore blast furnace practices, which require minimal investment in modifications, and thus is

cheaper than bringing Greenfield laterite projects online. Limonite/transitional laterites are used

as feed for NPI production as saprolite ore feed produces large slag volumes. China is the largest

importer of nickel for stainless steel production and as of 2010 NPI accounts for 30% of China’s

primary nickel consumption (Cartman, 2012). NPI is considered “dirty nickel” as production

results in large volumes of slag, high energy consumption, pollutes the environment and produces

low quality products (Youping, Yusheng, & Zhaoyi, 2008). Thus NPI production only becomes

economical during times of elevated nickel prices. Improvement to existing operations, and a

planned increase of 600kt of new NPI production capacity means that China’s NPI industry will

affect the global nickel markets for years to come.

2.3.2 Hydrometallurgical Processes

Hydrometallurgical processes have been established to treat the goethite rich limonite

layer of laterite deposits. The entrapment of nickel within the magnesium silicate matrix present

in the saprolite layer, along with the presence of acid consuming minerals, limits the ability of

13
hydrometallurgical process to treat saprolite ore. Hydrometallurgical treatments focuses on the

recovery of nickel and cobalt, predominantly from limonite ores.

2.3.2.1 Caron Process

The Caron Process is a combined pyro/hydro flowsheet developed to treat limonite and

limonite/saprolite blended feeds. Feed ore, dried to 5% moisture content, is selectively reduced at

750oC to convert nickel and cobalt oxides to the metallic state while leaving the iron content in

the magnetite phase. The ore is cooled under non-oxidizing conditions and leached using an

ammoniacal ammonium carbonate lixiviant, which selectively recovers the metallic components.

The pregnant solution is stripped of ammonia by boiling and the Ni/Co values are precipitated as

basic carbonates. Carbonates can be sold as an intermediate product or calcined at 900 oC to

produce nickel oxide sinter (90% Ni 10% O). Typically, the recovery of nickel is low (75%) and

decreases with increasing saprolite content due to Ni/Co values contained within the silicate

matrix which are difficult to reduce at the designed temperature. The energy intensive front end

and the extensive reagents required for the hydrometallurgical stages limits the cost effectiveness

of the Caron process.

2.3.2.2 HPAL Process

An extensive review by Whittington and Muir (2000) outlines the pressure acid leach

process describing how the operations at Moa Bay and the three Western Australian HPAL

projects are affected by variations in mineralogy, leaching process chemistry, operating

conditions, residue properties and scaling. The HPAL Process was developed as an alternative to

the Caron process, eliminating the energy intensive drying and reduction roast steps (50% of

energy consumption) required to remove the high moisture content contained within the limonite
14
ore (Habashi, 1997). The feed ore is restricted to limonite with low magnesia and alumina

content to limit the acid consumption by these minerals. The process begins with mixing the ore

with water, then pressure leaching the slurry with sulphuric acid at a temperature and pressure of

250oC and 40 MPa, respectively. Ambient leach conditions cannot be used as total dissolution of

the ore occurs. Pressure acid leach, at temperatures greater than 250oC selectively dissolve Ni/Co

oxides as stable sulphates, while iron sulphates destabilize and precipitate out as hematite. After

the leaching stage, solid/liquid separation recovers the pregnant solution, which then undergoes

further refinement, such as solvent extraction, to separate out the nickel and cobalt. Downstream

unit operations vary between the different HPAL operations, producing a variety of marketable

products;

 Mixed sulphide precipitate (Moa Bay, Cuba)

 Mixed hydroxide precipitate (Ravensthrope, Australia)

 Nickel carbonate precipitate (Cawse, Australia)

 Direct solvent extraction (Goro, New Caledonia)

The established HPAL Process achieves > 90% nickel recoveries, supplying 100 000 tonnes

of Ni and 10 000 tonnes of Co annually (Crundwell, Moats, Ramachandran, Robinson, &

Davenport, 2011). Costly capital expenditure on construction material, reagent costs,

commissioning difficulties and restrictions on feed material have limited the effectiveness of the

HPAL flowsheet.

2.3.2.3 Atmospheric Leaching (AL) and Heap Leaching (HL)

Leaching under ambient pressures provides advantages over the commercially practiced

HPAL Process in terms of lower capital and operating costs. McDonald and Whittington (2008)

reviewed the atmospheric leaching of laterite ores and provide an extensive comparison between
15
AL and HPAL processes. Atmospheric leaching avoids the need for expensive capital associated

with installing autoclaves and reduces operating cost due to lowered energy consumption and

maintenance costs. Most atmospheric leach processes operate at temperatures near the boiling

point of the lixiviant, which is typically sulphuric acid. Investigations into utilizing alternative

leaching liquors including, hydrochloric acid, organic acids and bio-technologies, have been

conducted but have not been implemented on the pilot plant level (McDonald & Whittington,

2008). Canterford (1978) and Agatzini-Leonardou and Zafiratos, (2004) found that atmospheric

leaching was non-selective as similar leaching behavior was exhibited by Ni, Co, Mg and Fe.

The AL process produced highly contaminated pregnant liquor and resulted in high acid

consumption. For AL to become economical a cheap source of sulphuric acid must be present

and improved downstream extraction methods need to be developed for the removal of nickel

from the pregnant solution.

Heap leaching (HL) processes have been successfully applied commercially for the

treatment of gold and copper ores. Heap leaching offers the most cost effective extraction

method for the recovery of nickel from laterite ores (McDonald & Whittington, 2008).

Successful heap leach operations require adequate irrigation, good permeability and extensive

leaching times. High acid consumption, potential breakdown of the aggregated mineral structure

and poor selectivity of nickel are the drawbacks that limit the commercial application of heap

leaching. Currently the Yunnan Ting Group, CVRD Inco and Mirabela Nickel are examining

heap leaching processes and European Nickel PLC conducted a trial operation at the Caldag

project which produced 20,400 tpa of nickel and 1200 tpa of nickel and cobalt respectively as a

mixed hydroxide product in 2009 (McDonald & Whittington, 2008).

16
2.4 Examples of Innovative Upgrading Techniques

2.4.1 Thermal Upgrading

Thermal upgrading is a purposed technique to treat limonite ore in order to make the ore

more amenable to physical beneficiation processes and was the focus of this study. Thermal

upgrading refers to the reduction of the contained nickel and cobalt values to the metallic state,

which can then be removed from the non-metallic gangue material. Varying operating

conditions, such as temperature, retention time, reducing atmosphere and reagent additions aim to

enhance the growth of metallic particles to sizes which are susceptible to physical upgrading such

as magnetic separation or flotation. Numerous patents have been filed that describe thermal

upgrading processes, as shown in Table 2 below. These patents claim to achieve effective nickel

recoveries and grade, but one process outlined in patent No. 5,178,666 seems promising as it uses

limited reagents additions, lower temperatures, shorter resident times and a cheaper reductant in

the form of bituminous coal. To the authors knowledge none of these patents have been used at

the commercial scale.

17
 Table 2: Thermal upgrading patents
Head Residence Ni
Temperature Reactor Ni Grade
Patent # Ore type Grade Reductant Additives Time Recovery Reference
(oC) Type (wt%)
(wt%) (min) (%)
Ni 1.7
Thumm and
Transitional Fe 34.3 13% Fuel Oil/ 4%
3,388,870 8% NaSO4 950 - 1150 120 Rotary Kiln 80 – 90 8 – 15 Heitmann
Saprolite SiO2 16.9 Lignite/ 1:1 CO:CO2
1968
MgO 10.4
Ni 0.77
Fe 40.2
3 – 7% Tube Crama and
4,490,174 Limonite Si 7.5 70: 30 CO:CO2 950 – 1050 30-300 92 -94 4– 5
FeS/NaSO4 Furnace Baas 1984
Mg 0.4
Al 4.5
Tube
3,656,935 N/A Ni 0.5% 3 – 11% Fuel oil 2 – 4% NaCl 950 30-60 23 – 70 % 15 – 26 Iwasaki 1972
Furnace
Ni 2.71 Beggs,
Transitional Fe 19.65 Rotary Bunge and
3,503,735 3% Coal None 980, 1400 5 - 10 90% 15 - 35
Saprolite SiO2 38.56 Hearth Vedensky
MgO 11.52 1970
Ni 1.63
1% S
Transitional Fe 33.8 Rotary Bell and
4,049,444 6% Fuel Oil 1300 20 80 – 92 9 – 15
Saprolite SiO2 16.4 Hearth Sridhar 1977
2% CaCO3
MgO 12.0
Ni 0.70 1.5 – 7.5%
Fe2 O3 32.0 NaCl Electric Nestoridies
4,002,463 Transitional 2 – 5% Coke 1050 60 – 90 78 12
SiO2 47.1 0.1 - 0.5% Furnace 1977
MgO 4.9 gypsum
Ni 0.70
Limonite Fe2 O3 32.0 Diaz, et al.
5,178,666 4 - 6% Coal 2 - 4% Sulfur 600, 1000 - 1100 30 - 60 Rotary Kiln 80 - 94 8.5 - 9.5
Saprolite SiO2 47.1 1993
MgO 4.9

18
 The experiments performed in the Inco patent on limonite ore mixed with 4 – 6%

bituminous coal and 2 – 4% sulphur, produced a ferronickel concentrate with acceptable nickel

grades and recoveries as shown in Table 3 below. The low weight fraction of magnetic

concentrate produced illustrated that the process was efficient at rendering the contained nickel

into a separable phase. The gangue, comprised mostly of iron oxides, was reduced to wustite and

separated from the ferronickel. Reduction of iron oxides proceeds along the following pathway:

Goethite → Hematite→ Magnetite→ Wustite→ Metallic Iron

Careful control of the reduction potential during the experiments enabled the production

of a wustite phase and limited the formation of metallic iron. The low levels of metallic iron in

the concentrate were illustrated by the low iron to nickel ratio. Metallic iron dilutes the FeNi

concentrate, reducing the nickel grade and increasing the mass of concentrate produced.

Magnetite can also dilute the concentrate and is produced by either the under reduction of the iron

oxide phase or through the oxidation or disproportionation of the wustite phase.

Wustite is a metastable phase that becomes unstable in a neutral atmosphere below 570 oC

(Pattek-Janczyk & Miczko, 1990). Wustite undergoes a disproportionation reaction forming

magnetite and metallic iron as seen in Equation 1 below.

𝟒𝑭𝒆𝑶 = 𝑭𝒆𝟑 𝑶𝟒 + 𝑭𝒆 1

This disproportionation reaction has been found to occur at temperatures ranging from

300 – 570oC using Mossbauer spectrometry (Broussard, 1969; Mori, 1999; Pattek-Janczyk &

Miczko, 1990) . Although thermodynamically possible, the kinetics of this reaction are slow and

require retention times >60 min at elevated temperatures to proceed.

19
 Table 3: Example 1 bench test results from U.S. Pat. No. 5,178,666 (Diaz, et al., 1993)

Magnetic Fractions
% % Ni Grade Ni Recovery Fe/NI wt%
Ore Type S Coal wt % % % Ratio
A 4 6 13.1% 9.8% 88.0% 5.2
A 4 6 12.2% 11.4% 92.0% 4.8
A 4 6 11.2% 11.0% 87.0% 4.4
A 4 6 12.7% 11.3% 95.0% N/A
A 2 6 15.6% 9.1% 91.0 % 4.6
A 2 6 14.3% 10.1% 94.0% 4.6
A 2 5 14.4% 10.7% 87.0% 3.2
A 2 5 14.0% 10.6% 88.0% 3.0
B 4 4 14.0% 9.2% 88.0% 3.8

2.4.2 Selective Sulphidation

Harris (2011) and Harris et al. (2011) studied low temperature sulphidation of laterite

ores and were able to demonstrate that the nickel oxide within the ore could be selectively

sulphidized to a nickel-iron sulphide. The nickel-iron sulphide could be concentrated via

flotation, producing an intermediate product at lower energy consumption levels and a relatively

simpler flowsheet compared to existing processes. Nickel sulphides form at lower sulphur

potentials than iron sulphides, and it is believed that NiO could then be selectively sulphidized to

mono and disulphides. Sulphidation was attempted at temperatures between 450 – 1100 oC, with

sulphur additions ranging from 25 – 1000 kg S/tonne ore. Selectivity was strongly dependent on

temperature and sulphur additions. The highest degree of nickel sulphidation was achieved with

100 kg S/tonne ore at temperatures greater than 500 oC but the Fe-Ni-S phase formed was

submicron in size. When temperatures were elevated to 1050 – 1100 oC, the sulphide particles

grew via a liquid phase sintering mechanism in an oxygen rich matte to a d80 of 14 μm. The

sulphide growth temperature was similar to the metallic growth zone temperature established in

20
the INCO patent. Flotation studies of the nickel-iron sulphide produced a concentrate grading

4–5% nickel, with 50% recovery on a sulphide basis, 35-45% overall nickel recovery. Low

recovery was attributed to the fine sulphide particles produced and entrainment of magnetite

fines. Proof of concept was shown for the selective sulphidation of nickel in limonite ores, but

further study is required to develop the process.

2.4.3 Segregation Process

The segregation process selectively reduces Ni, Co and some Fe to the metallic state via

gaseous chlorides. The metallic component can then be separated from the gangue material by

magnetic, flotation or leaching techniques. Various studies have been conducted on the

mechanism and process of segregation roasting of laterite ore shown in the work by Iwasaki et al.

(1966), Liu et al. (2010), Kwatara et al. (2011) and Reznik et al. (2003). Experimental work has

predominantly been conducted on upgrading saprolite ore. The process involves mixing laterite

ores with a chloridizing agent (CaCl2, NaCl, or MgCl2) and a solid reducing agent (coke or coal)

then roasting the mixture between temperatures of 900 – 1000oC. The process mechanism

proceeds through three steps:

Step 1: Hydrogen chloride gas produced by the hydrolysis of the salt

𝑺𝒊𝑶𝟐 (𝒔) + 𝑪𝒂𝑪𝒍𝟐 (𝒔) + 𝑯𝟐 𝑶(𝒈) = 𝑪𝒂𝑺𝒊𝑶𝟑 (𝒔) + 𝟐𝑯𝑪𝒍(𝒈) 2

Step 2: HCl reacts with nickel oxide to produce nickel chloride gas

𝑵𝒊𝑶(𝒔) + 𝟐𝑯𝑪𝒍(𝒈) = 𝑵𝒊𝑪𝒍𝟐 (𝒈) + 𝑯𝟐 𝑶(𝒈) 3

Step 3: Hydrogen reduction of nickel chloride to metallic nickel on the surface of the
carbonaceous reductant
𝑵𝒊𝑪𝒍𝟐 (𝒈) + 𝑯𝟐 (𝒈) = 𝑵𝒊(𝒔) + 𝟐𝑯𝑪𝒍(𝒈) 4

21
 Cobalt and iron undergo similar reactions during the process though selectivity is

achieved as NiCl2 is preferentially reduced over FeCl2. The range of recoveries and grades of Ni

achieved by the various experimental works are illustrated in Table 4 below.

Table 4: Experimental test work results for the segregation process

Recovery Concentrate Roasting Chloridizing Reference

(%) Grades (wt %) Temperature (oC) Agent
85 - 90 30 950 CaCl2 (Iwasaki, Takahasi, &
Kahata, 1966)
91.5 5.25 1000 MgCl2 (Liu, et al., 2010)

82 – 90 8 - 10 1000 CaCl2 (Reznik, Tarasov, &

Mayorov, 2003)

33 - 85 2.1 - 15 900 - 1000 CaCl2 (Kwatara, et al., 2011)

Segregation roasting test work illustrates an attractive process for the recovery of nickel from

laterite ores that utilizes lower reaction temperatures and a relatively simpler flowsheet. The

drawbacks limiting the implementation of the process are:

- Material selection and reactor design are limited when using chloride based reagents

- Unconsumed carbon present after roasting impacts downstream operations

- Small size of metallic particles produced require fine grinding

- Salts reagents are required in larger volumes which are difficult to store and transport

- Difficulties associated with roasting salts (sintering of charge, residue material)

No commercial plants have been commissioned implementing segregation roasting

technology for nickel extraction and limited pilot plant studies have been conducted. The

22
MINPRO-PAMCO nickel segregation process was piloted from 1981 – 1983. The 1 tonne/hr

plant treated saprolite ore mixed with CaCl2 and coke at 950oC. The facility operated for 200

days and produced a 55-60% Ni concentrate via magnetic separation at a recovery of 90%

(Ericson, Svensson, & Ishii, 1987).

2.5 Reduction of Nickel Laterites

2.5.1 Solid State Reduction Mechanism

The reduction mechanism for limonite ore reduced with a solid carbon additive has been

studied by various authors (Zevgolis, Zografidis, & Halikia, 2010; Wang, Yang, Tian, Li, & Sun,

1997; Antola, Holappa, & Paschen, 1995). Solid state reduction at the contact sites between

carbon and limonite ore is believed to be the origin of the reduction process. The reduction

process is continued via gaseous intermediates produced through the solid state reduction

mechanism and the Boudourd reaction. Zevgolis et al (2010) summarized the solid state reduction

of the laterite ore using a coal reductant into three simplified categories; iron oxide reduction,

nickel oxide reduction and carbon gasification. The solid state reduction mechanism of the iron

oxide component of the laterite ore is expected to follow a three stage reaction mechanism.

Stage 1: Reduction of hematite to magnetite

𝟑𝑭𝒆𝟐 𝑶𝟑 + 𝑪𝑶 → 𝟐𝑭𝒆𝟑 𝑶𝟒 + 𝑪𝑶𝟐 5

Stage 2: Reduction of magnetite to wustite

𝐹𝑒3 𝑂4 + 𝐶𝑂 → 3𝐹𝑒𝑂 + 𝐶𝑂2 6
Stage 3: Reduction of wustite to metallic iron
𝑭𝒆𝑶 + 𝑪𝑶 → 𝑭𝒆 + 𝑪𝑶𝟐 7

Nickel oxide reduction occurs in a single reaction.

𝑵𝒊𝑶 + 𝑪𝑶 → 𝑵𝒊 + 𝑪𝑶𝟐 8

23
Carbon gasification occurs via the Boudouard reaction.

𝑪 + 𝑪𝑶𝟐 → 𝟐𝑪𝑶 9

The solid-state reduction mechanism is believed to be carried out by gaseous

intermediates produced by the Boudouard reaction that maintain the reducing atmosphere. The

initial CO present in the early stages of the reaction is produced through the direct contact of the

carbon particles and ore according to the following reactions:

𝟑𝑭𝟐 𝑶𝟑 + 𝑪 → 𝟐𝑭𝟑 𝑶𝟒 + 𝑪𝑶 10

𝑵𝒊𝑶 + 𝑪 → 𝑵𝒊 + 𝑪𝑶 11

In concurrence with the formation of CO, volatile matter is produced from heating coal

which contains hydrocarbons. Above 700oC Wang et al (1997) stated that the hydrocarbons

began to crack into H2 and C. The CO, H2 and C produced can act as reductants. Therefore true

solid state direct reduction via gaseous intermediates using a coal reductant occurs according to

the following reactions:

3𝐹𝑒2 𝑂3 + 𝐶𝑂⁄𝐻2 → 2𝐹𝑒3 𝑂4 + 𝐶𝑂2 ⁄𝐻2 𝑂 12

𝐹𝑒3 𝑂4 + 𝐶𝑂⁄𝐻2 → 𝐹𝑒𝑂 + 𝐶𝑂2 ⁄𝐻2 𝑂 13

𝑁𝑖𝑂 + 𝐶𝑂⁄𝐻2 → 𝑁𝑖 + 𝐶𝑂2 ⁄𝐻2 𝑂 14

𝐶𝑂2 ⁄𝐻2 𝑂 + 𝐶 → 2𝐶𝑂 ⁄𝐶𝑂 + 𝐻2 15

Antola et al (1995) assumed that upon dehydroxylation of goethite, the oxide mineral

phase forms trevorite, NiO·Fe2O3, which undergoes the following reduction reactions:

3𝑁𝑖𝑂 · 𝐹𝑒2 𝑂3 + 𝐶𝑂 → 3𝑁𝑖𝑂 + 2𝐹𝑒3 𝑂4 + 𝐶𝑂2 16

𝑁𝑖𝑂 + 2𝐹𝑒3 𝑂4 + 𝐶𝑂 → 𝑁𝑖 + 2𝐹𝑒3 𝑂4 + 𝐶𝑂2 17

𝐍𝐢 + 𝐅𝐞𝟑 𝐎𝟒 + 𝐂𝐎 → 𝐍𝐢 + 𝟑𝐅𝐞𝐎 + 𝐂𝐎𝟐 18

24
 Therefore selective reduction of nickel over iron could be achieved if the reduction

process could conclude after Equation 15, thus minimizing further reduction of iron oxide to the

metallic phase.

Nickel substitution into serpentine minerals may be present in the limonite formation and

therefore the reduction mechanism of these phases should be considered. Decomposition of the

silicate mineral will occur between 500 – 800oC according to the following reaction:

(𝐍𝐢, 𝐅𝐞, 𝐌𝐠)𝟑 𝐒𝐢𝟐 𝐎𝟓(𝐎𝐇)𝟓 → 𝟑(𝐍𝐢, 𝐅𝐞, 𝐌𝐠)𝟐 𝐒𝐢𝐎𝟒 + 𝟒𝐇𝟐 𝐎 + 𝐒𝐢𝐎𝟐 19

The decomposed material forms an amorphous solid where NiO is present and relatively

easy to reduce. At temperatures > 800oC the amorphous material recrystallizes out as olivine (as

forsterite - Mg2SiO4 or fayalite - Fe2SiO4), entrapping the NiO in a phase that is difficult to reduce

(Li & Coley, 2000). Fayalite also forms through the reduction of magnetite and free silica

forming the stable iron silicate as shown in equation below.

𝟐𝐅𝐞𝟑 𝐎𝟒 + 𝟑𝐒𝐢𝐎𝟐 + 𝟐𝐂𝐎 → 𝟑(𝟐𝐅𝐞𝐎 ∙ 𝐒𝐢𝐎𝟐 ) + 𝟐𝐂𝐎𝟐 20

Reduction of the recrystallized fayalite can occur, but requires higher temperatures and stronger

reducing conditions (Harris, Peacey, & Pickles, 2009). The reduction mechanism for fayalite is

shown below:

(𝐍𝐢, 𝐅𝐞)𝟐 𝐒𝐢𝐎𝟒 + 𝟐𝐂𝐎 → 𝟐(𝐍𝐢, 𝐅𝐞) + 𝐒𝐢𝐎𝟐 + 𝟐𝐂𝐎𝟐 21

2.5.2 Reducibility of Nickeliferous Laterite Ores

Extensive studies have been conducted on the solid state reducibility of nickeliferous

laterite ores using gaseous and solid reductants. Hallett (1997) preformed a thermodynamic

analysis of the solid-state reduction of nickel from goethite, producing various predominance area

25
diagrams that outlined the limits of stability for the various mineral phases as a function of

temperature and reducing conditions. Limonite, composed primarily of goethite, FeOOH, was

dehydroxylated when exposed to heat and could be reduced under mild conditions to magnetite,

Fe3O4, and nickel bearing trevorite, NiO·Fe2O3. Magnetite and trevorite from a spinal type solid

solution with the formula (Fe,Ni)O· Fe2O3. Varying the temperature and gas composition of the

reducing environment can alter the mixed oxide phase according to the stability diagram

presented in Figure 7. Examinations of these predominance area diagrams illustrate that iron has a

greater affinity for oxygen than nickel, and thus selective reduction of nickel oxide to the metallic

phase should be achievable.

These findings agreed with Canterford (1975) review of the metallurgical treatments of

nickeliferous laterites which found that the process of selective reduction of nickel and cobalt

without the reduction of iron as a critical aspect in the study of extractive metallurgy of

nickeliferous laterites. Thermodynamic data for the reduction potential of iron and nickel oxides

shown in Figure 8 indicates the potential to selectively reduce NiO to the metallic state while

minimizing the metallization of iron. Minimal metallization of iron is required to achieve

economically acceptable reaction rates, therefore careful control of atmosphere conditions and

temperature are required to establish a suitable environment. Practical application of the selective

reduction process has been restricted as the optimal temperatures, retention time and composition

of the reductant vary significantly over the range of ores treated. Selective reduction is ore

specific, limiting the application of the process to a wide variety of mined laterite ores.

26
 (Fe,Ni)

95%

9O%

(Fe,Ni)O +
(Fe,Ni) 5O%

(Fe,Ni)O·Fe2O3
+ (Fe,Ni) 1O%
O%
(Fe,Ni)O
)O

95%

9O%
(Fe,Ni)O·Fe2O3

5O%
10%
O%

Figure 7: Fe-Ni-O stability diagram for goethite containing 1.2% Ni, reduction isopleths labelled with
percent values join points of constant degree of Ni metallization (Hallet, 1997)

27
Figure 8: Phase Equilibria for the reduction of nickel and iron oxides with carbon (Canterford, 1975)

28
 Even though, thermodynamically NiO can be reduced at a lower potential compared to

iron oxides, Zhu et al. (2008) and Wantanbe et al. (1987) found that nickel and iron oxides in

laterite ores are reduced with coal simultaneously despite the differences in the free energy of

formations between these oxides. An array of published work has illustrated various degrees of

success to selectively reduce and upgrade limonite ore at the experimental bench scale level. De

Graaf (1979) showed that selective reduction of limonite to the magnetite/wustite phase boundary

was possible using a CO/CO2 and H2/H2O reducing atmosphere. The presence of water vapour in

the reducing atmosphere was found to have a deleterious effect on the extraction of nickel.

Using a gaseous reductant in a tube furnace and fluidized bed apparatus, Antola et al

(1995) were able to produce a high grade (70% - 80% Ni) FeNi product by selective reduction in

the solid state. Pure metallic nickel could not be produced from selective reduction as nickel was

always found in a nickel-iron alloy phase. Low partial pressure of reducing gas with relatively

short experiment times provided the best results. Changes within the temperature range of 700-

1000 oC were found to have little effect on the reduction selectivity although reduction proceeded

faster at higher temperatures. Reduction occurred 5-7 times faster in the fluidized bed compared

to the tube furnace. Although selective reduction of nickel was achieved, the formed ferronickel

phase could not be separated from the iron oxide phase by magnetic separation.

Zevgolis et al (2010) concluded that the reduction degrees of iron and nickel oxides

reduced between 700 – 900oC could not exceed 33% and 76% respectively in Greek laterites even

though adequate amounts of carbon remained in the calcine. The reduction mechanism finished

after 20 – 40 minutes and was prevented from reaching completion due to the formation of iron

silicates such as fayalite (2FeO·SiO2) and forsterite (Mg2SiO4), which coated the oxide grains,

impeding further reduction.

29
 Valix and Cheung (2002) studied phase transformations of laterite ores under reducing

conditions and temperatures up to 800oC. When limonite ores were reduced, the mineral phases

formed were magnetite, taenite and fayalite. Taenite is a nickel rich iron-nickel alloy mineral.

XRD patterns showed the presence of taenite becomes more dominant with a temperature

increase from 650oC to 750oC, but further heating resulted in the disappearance of taenite. Upon

cooling the reduced limonite from 800 oC to 25oC in air, no evident changes occurred in the

phases suggesting the stability of the reduced phases formed at higher temperatures. Valix and

Cheung (2002) also conducted leach tests using an ammoniacal leach and achieved an optimum

recovery of nickel of 80% when the limonite ore was reduced at 600oC.

Kawahara et al (1988) investigated the reducibility of laterite ores using hydrogen gas

reductant over a temperature range of 400 to 1000 oC. Reducibility was based on the degree of

reduction of iron, nickel and cobalt to metallic form determined using a bromine-methanol

diagnostic leach followed by atomic absorption spectrophotometry. Degree of nickel reduction

for limonite samples varied from 19% to 70% and all reduction concluded after 40 minutes.

Reducibility of nickel increased with increasing temperature and iron content within the ore.

Olivine formation due to the magnesia and silica content resulted in ion substitution of Ni 2+ with

Mg2+ into the silicate matrix. The activity of nickel contained within the silicate solid solution is

low, explaining the poor reducibility via gaseous reductants.

The effect of pre-calcination of the limonite ore before reduction roasting in a CO/CO 2

atmosphere was studied by O`Connor et al (2006) . The function of pre-calcination was to drive

the dehydroxylation reaction of goethite, which is the removal of the structural OH- group or loss

of crystalline bound water. Calcination opened the goethite mineral structure, increasing the

porosity and surface area of the ore, exposing the reduced nickel phase to recovery via an

30
ammoniacal leach solution. Under reducing conditions, goethite was dehydroxylated to hematite,

and then reduced to magnetite at 500oC. The formation of an iron nickel alloy taenite (Ni3Fe) and

fayalite simultaneously occurred at this temperature. Pre-calcination produced recrystallized

hematite, which was less susceptible to reduction and thus resulted in reduced ferro-nickel

formation. Utigard and Bergman (1992) found the pretreatment of saprolite ore in air or nitrogen

before reduction in an H2/CO2 atmosphere had no effect on the formation of metallics. Under the

same conditions, limonite ore was easier to reduce to the metallic state compared to saprolite

ores.

Chueng and Valix (2008) found that a prolonged reduction period resulted in “over

reduction” of limonite resulting in decreased nickel recovery. Lower reduction potential (1:2.5

versus 1:1 CO:CO2) prevented the recrystallization of the dehydroxlated serpentine mineral to

olivine and pyroxene. Recrystallization of the silicate phase can incorporate the nickel oxide,

preventing reduction and consequently the degree of nickel separation from the gangue material.

Reoxidation of the reduced wustite phase to the higher oxidized iron oxides impedes the

effectiveness of magnetic separation to remove the gangue material. Chander and Sharma (1981)

found that cooling in an inert atmosphere prevented re-oxidation and retained the highest portions

of wustite within the reduced limonite ore. Reoxidation was limited to the surface layers of the

sample. Quenching the samples caused extensive re-oxidation of the hot sample by either the

water or air during the transfer from the reaction tube.

31
2.6 Ferronickel Particle Growth

The size of ferro-nickel particles produced during the thermal upgrading process must be

adequate in size in order to be susceptible to magnetic separation. A goal of the INCO patent was

to provide a metallic growth zone where controlling the reduction atmosphere, temperature,

retention time and sulphur addition provided the ideal conditions for metallic particle growth. A

review of published work on the factors and mechanisms surrounding growth of ferro-nickel

particles during the reduction roasting of laterite ores was conducted.

Wantanabe et al. (1987) proposed that ferro-nickel particles agglomerate in a semi-fused

state via a suspension system model as shown in Figure 9 below. The model suggests that

agglomeration of metallic particles is strongly influenced by interfacial interaction between the

metal particles and the slag phase. Additions to the charge feed such as sulphur, lime and alumina

along with iron oxide present in the slag help to reduce surface tension and viscosity of the slag

phase allowing for the movement and coalescing of metal particles. Excessive residual carbon

remaining after the reduction stage was found to increase viscosity and thus reduce agglomeration

of ferro-nickel particles.

32
 Figure 9: Diagram of metal agglomeration process via suspension model (Watanabe, Ono, Arai, &
Matsumori, 1987)

2.6.1 Effect of Temperature

Huang et al. (2011) found the growth of ferronickel particles, within a laterite-coal

composite pellet, was greatly influenced by the roasting temperature. Scanning Electron

Microscope (SEM) images found ferronickel particles ranging in size of 5 – 10 μm when reduced

at 1200oC, increasing to the range of 15 – 20 μm at 1250oC and 20 – 150 μm at 1300oC. With a

reducing temperature of 1350oC, ferronickel particles agglomerated forming 3 – 8 mm granules

similar to luppen formed in the Nippon Yakin Oheyama process, which contributed to the

increased carburization rate causing the starting melting temperature to decrease. Images also

detected networks of voids formed from the gasification of solid carbon. Theses voids provided

locations for the diffusion and agglomeration of the reduced ferronickel particles.

33
2.6.2 Effect of Sulphur
Diaz et al. (1993) claimed that metallic nickel, cobalt and iron congregate into ferro-

nickel particles through a Fe-S-O liquid phase. Sulphur additions are required to establish this

phase and promote the growth of ferro-nickel particles. Numerous studies (Li et al. 2012; Zhu et

al. 2012; Valix and Cheung, 2002; Jiang et al. 2013) found that the addition of sulphur promoted

the growth of FeNi particles during the selective reduction of saprolite ores. Limited published

work was found that investigated the effect of sulphur on the promotion of particle growth during

the reduction roast of limonite ore.

Encapsulated sulphur released S2 gas when heated within the briquettes which reacts with

reduced iron to form troilite FeS. Troilite forms a Fe-FeS low melting point eutectic (985oC)

which facilitated the agglomeration of the FeNi particles while also enriching the nickel phase by

supressing the complete metallization of iron. Metallic iron is preferentially sulphidized,

reducing the iron content within the ferro-nickel alloy (Kukura, Stevens, & Auck, 1981). The

release of gaseous sulphur opens the mineral structure, allowing for water to be removed during

dehydroxylation and increases the surface area for reduction between the ore and reducing gases.

The fissures/voids left behind after sulphur volatilization provides nucleation zones for FeNi

particle growth. Sulphur also reduced the surface tension of the FeNi particles, causing the metal

particles to agglomerate. SEM images, below in Figure 10, show ferronickel particles enwrapped

in FeS and it is believed the FeNi phase precipitated out of the FeS phase. The addition of sulphur

increased the FeNi particle size from 5-10 μm to > 50 μm.

34
 Figure 10: SEM images of Fe-Ni and Fe-S phases: (2) and (A) Fe-S, (1) and (B) Fe-Ni (Zhu, et al.,
2012) and (Li, Shi, Rao, Jiang, & Zhang, 2012)

Compared to the INCO patent which uses sulphur additions of 2 – 4 wt%, Bell and

Sridhar (1977) found that additions of 1wt% sulphur promoted the concentration of nickel into

the ferronickel particles. Sulphur contents higher than 1 wt% were found to adversely affect the

recovery of ferronickel particles by magnetic separation. Jiang et al. (2013) found that sulphur

additions over 2 wt% in a reduction roast process of saprolite ore reduced the recovery of nickel

from 82% to 72%.

Lu at al. (2013) found that the addition of Na2SO4 induced the formation of a Fe-S solid

solution and improved mass transfer and gaseous diffusion throughout the laterite ore. The low

melting point Fe – S solid solution enwrapped and then facilitated the aggregation of the

ferronickel particles. The addition of Na2SO4 provides a source of sulphur while also liberating

nickel and iron from silicate formations. Growth of ferronickel particles with the addition of

Na2SO4 reached an average size of 20 μm with the maximum particle sizes being 40 μm.

35
 Chapter 3: Methods and Materials

3.1 Introduction
Experimental test work was conducted to investigate the thermal upgrading of nickel

containing limonite ore. The goal of the experimental work was to thermally upgrade limonite in

a two stage reduction roast at relatively low temperatures and with minimal reagent additions.

The experiments conducted used the conditions, procedure and results established in Example 1

of the U.S. Pat. No. 5,178,666 as a starting point.

Bench scale experiments were conducted using a stationary bed furnace and rotary kiln

setups, where the effects of temperature, retention time and sulphur addition on the thermal

upgrading of limonite ore were examined. Magnetic separation tests were utilized to assess the

ability of the thermally upgraded material to be concentrated into a ferronickel concentrate.

Samples obtained were exposed to X-ray diffraction, SEM microscopy and thermogravimetric

differential thermal analysis to fully characterize the thermally upgraded product.

3.2 Ore Characterization

Limonite ore provided by Xstrata from the Ivory Coast, was used for all experiments

conducted in this study. The head grade on the limonite ore used in this study along with

limonitic ore utilized in the INCO patent experiments are shown and compared in Table 5.

36
 Table 5: Head grade analysis of limonite ore samples and INCO patent samples

Component Xstrata Inco Patent

Ore (wt %) Ore
(wt %)
Ni 1.38 1.34
Co 0.35 0.02
Fe 45.70 46.70
SiO2 2.15 4.51
MgO 0.22 1.46
Al 2.92 2.46

The content of both ores were similar and should allow for favorable comparisons when

attempting to treat the Xstrata ore using the experimental conditions outlined in the INCO patent.

Both ores exhibited high values of iron (>45 %) with low concentrations of SiO2 and MgO which

is typical of limonite formations. X-ray analysis was conducted, shown in Figure 11, identifying

goethite as the major mineral phase present in the limonite ore.

G G
G G—Goethite

G
G
G

GG

Figure 11: X-ray spectrum of limonite ore

37
 Using the Mineralogy Iterations feature in HSC Chemistry, the head grade analysis of the

limonite ore was used to estimate the mineral composition of the ore shown in Table 6. X-ray

analysis identified goethite as the dominant mineral phase, therefore all iron was assumed to be

present as goethite. Contained nickel was assumed to present as nickel oxide, NiO. The

estimated mineral composition was used in theoretical reduction calculations of the limonite ore.

Table 6: Estimated mineral composition of limonite ore

Component Wt %
NiO 2.06
FeOOH 85.48
Mg3Si2O5(OH)4 0.98
SiO2 4.98
Al2O3 6.49
Other <0.01
Total 100.00

3.3 Ore Preparation and Pellet Formation

The as-received ore was dried at 100oC in an oven for one week to remove the free

moisture content of the ore. A steel rolling pin was used to crush 2kg of the ore to a size that

could be adequately segregated into representative samples via an open bin riffler splitter. Riffle

splitting produced 8 representative samples of approximately 180 g each. The particle size of the

ore in each individual sample was reduced using a ring pulverizer. The ore was pulverized for

fifteen seconds then mechanically screened at 100 mesh. Oversized material was re-pulverized

for an additional five seconds then re-screened until 100% of the sample was – 100 mesh. To

prevent absorption of water into the ore and to maintain a uniform composition, the screened

samples were stored in a 65oC oven until being used in the experiments.

38
 In the test work, limonite ore was mixed with requisite amounts of a solid carbon

reductant and sulphur bearing agent. A coal reductant, similar to the bituminous coal used in the

patent, was used in all the experiments. A comparison of chemical analysis of the coal additive

used in this study with the bituminous coal used in the patent is shown in Table 7.

Table 7: Chemical analysis of coal and INCO patent bituminous coal

Component Coal Bituminous

Coal
Total Carbon 67.7 73.0
Fixed Carbon 37.2 51.6
Volatiles 41.2 36.6
Moisture 10.7 1.2
Ash 10.9 7.9
Sulphur 0.4 2.3

Reagent grade elemental sulphur was used in all the experimental work. To reduce dust

formation, improve solid-solid reactions and promote diffusion of volatiles through the ore, the

reagents and ore feed were briquetted for all experiments. Limonite, coal and sulphur additions

weighting a total of 3g were thoroughly mixed until the sample was homogenous, then briquetted

into 1.3 cm diameter cylindrical pellets at a pressure of 8000 psi. The resulting pellet height was

approximately 1.0 cm.

39
 Figure 12: Pressed 3g limonite pellet

3.4 Stationary Bed Furnace Experiments

A sealed quartz tube placed in a horizontal tube furnace was used as a reaction chamber

for the two stage thermal upgrading of the pellets. The schematic diagram of the furnace setup

used in all bench scale experimental work is shown in Figure 13. The experimental method used

was aimed at reproducing the method performed in Example 1 of U.S. Pat. No. 5,178,666. The

pellets were placed in a quartz sample carrier lined with magnesium oxide powder to prevent

sintering of the pellet with the carrier at elevated temperatures (≥1000oC). The furnace was then

sealed and purged with high purity nitrogen gas at a flow rate of 300 mL/min. The gas flow rate

was maintained throughout the entire experiment. The sample carrier maintained the pellet

within the cool zone of the furnace while the atmosphere was purged and the hot zone ramped up

to the reduction temperature of 600oC. The sample was then introduced to the hot zone and

allowed to react for 60 min. The pellet was held with the hot zone undergoing non-isothermal

heating at 40oC/min to raise the temperature to 1000 – 1100oC to promote metallic growth.

Temperature control overshoot at 1000 – 1100oC was limited to ± 2oC. The pellets were held at
40
the growth zone temperature for 60 min then retracted to the cool zone and cooled for 40 min.

The cooled pellets ranged in temperature from 60 – 80oC. Separation and analytical testing was

performed immediately after removing the reduced pellets from the furnace to minimize re-

oxidation.

Figure 13: Experimental tube furnace setup diagram

3.5 Rotary Kiln Experiments

A rotary kiln experimental setup shown in Figure 14 was utilized to investigate the effect

of rotational forces on the growth of ferronickel particles. Particle size, recovery and grade of

ferronickel produced, while using the rotary kiln setup, were compared to experiments conducted

in the stationary tube furnace. Rotary kiln experiments were conducted off-campus at Kingston

Process Metallurgy.

41
 Figure 14: Schematic diagram of rotary kiln setup

Two furnaces were installed around the kiln to establish a reduction zone (600 oC) and a

growth zone (1000oC). A pelletized feed of 300 g was placed inside a quartz capsule (Figure 15)

and inserted into the kiln. The capsule was held in the cooling zone in front of the furnaces while

the furnaces ramped up to the set point temperature. The capsule was then pushed into position

and held in the reduction zone and then the growth zone for 60 min respectively. After

completion of the two stage roasting procedure, the capsule was pulled back into the cool zone

and allowed to cool until the material could be safely handled. Nitrogen was metered into the

kiln at 1.5 L/min to maintain a neutral atmosphere throughout the experiment. The rotation speed

of kiln was constant throughout the experiment at 2 rpm.

42
 Figure 15: Quartz capsule containing 300g of pelletized feed at 6% coal, 4% S for rotary kiln
experiment

3.6 Magnetic Separation Tests

Magnetic separation tests were conducted on the thermally upgraded pellets to test the

viability of upgrading the limonite material to a ferronickel concentrate. Magnetic separation

tests were performed using a Davis tube and High Intensity Wet Magnetic Separator (HIWMS),

shown in Figure 16. Prior to separation, liberation of the ferronickel particles was performed

using a ring pulverizer. The samples were pulverized for 20 s then mechanically screened at 200

mesh. Oversized material was re-pulverized for an additional 5 s then re-screened. Water was

added to the – 200 mesh material prior to magnetic separation.

43
 Figure 16: Magnetic separators used in experiments: (left) HIWMS and (right) Davis tube

Initial magnetic separation test work was performed using a Davis tube as the device was

used extensively in literature and also in U.S. Pat. No. 5,178,666. The Davis tube treats sample

sizes of 10 g or less at 4800 Gauss at various flow rates and rotation speeds. Separation

difficulties were encountered as most of the material reported to the concentrate, with little tailing

being produced. Similar problems were encountered by Iwasaki et al. (1966) who identified, via

X-ray analysis of the roasted material, that the presence of a small amount of magnetite in wustite

was sufficient to prevent efficient magnetic separation. To provide a cleaner separation of

ferronickel particles, magnetic separation studies were conducted using an Outokumpu

Technology’s CARPCO WHIMS Model 3x4L. The WHIMS separates magnetic particles

through a static cell containing a media of 1/4” soft iron spheres at magnetic flux density ranging

from 0 – 21.6 kiloguass. The matrix media provides a collection surface more suited for

collecting smaller particles and minimizes entrapment of non-magnetic material.

44
3.7 Analytical Techniques

Analysis of the raw materials and thermally upgraded samples was carried out by a

variety of analytical techniques. Full characterization of the samples was performed to determine

material quantity, chemical composition and mineral phases present.

Chemical composition of the samples was determined using wet analysis techniques

where dissolution of the material was conducted using aqua regia or a bromine methanol

diagnostic leach. Total digestion of the samples was performed using aqua regia, while 5%

bromine in 95% methanol solution selectively dissolved metallics from the oxide media. The

diagnostic leach test followed a procedure adapted from Harris (2011) who utilized a variation of

the diagnostic leach procedure published by Young (1974) to selectively remove sulphides from

oxide materials. Dissolved solutions were analyzed for nickel and iron using a Thermo Scientific

iCE 3000 Series Atomic Absorption Spectroscopy (AA) unit. Replicate samples were also

analyzed using a PerkinElmer 2100DV unit to perform Inductively Coupled Plasma Optical

Emission Spectroscopy (ICP-OES).

The bromine methanol diagnostic leach required the thermal upgraded briquettes to be

ground into powder, then 0.3g of sample was added to 50 mL of the 5% bromine / 95% methanol

solution. The solution was heated to 60oC under reflux for 20 min. Boiling chips were added to

prevent violent evaporation of the solution. The solution was then cooled in a water bath and

vacuum filtered using Whatman No.2 glass fiber micro filters. Methanol was used to wash all

glassware. Twenty five milliliters of hydrochloric acid was then added to the filtered solution and

then the sample was boiled on a hot plate until 20 mL of solution remained. Boiling the solutions

allows the hydrochloric acid to convert the nickel and iron bromides into chlorides and remove

45
excess methanol. The remaining solution was diluted with deionized water to less than 20 ppm

and analyzed for nickel and iron via AA or ICP.

Powder X-ray diffraction (XRD) was used to characterize the mineralogy of raw the

materials and thermally upgraded samples. An Xpert Pro Phillips powder diffractometer with Cu

kα radiation linked with the commercial software X-pert High Score was utilized to analyze

samples. A cobalt radiation tube was also used on iron rich samples as the tube decreases

fluorescence emitted by iron, therefore reducing background noise and enabling detection of

minor phases. Peak identification was performed by comparing detected peaks with a database

on known spectrums. The database is maintained by the International Centre for Diffraction Data.

Scanning electron microscopy was performed using a MLA 650 FED Environmental

SEM. Thermally upgraded samples were prepared in powder form or the pellets were sectioned

vertically along the middle of the pellet. The metallic presence in the upgraded samples allowed

the SEM to operate at low vacuum and voltage, eliminating the need to pre-coat the samples with

carbon. Back scattered electron imaging was the primary mode used to analyze the samples.

Variations in brightness corresponded to differences in densities of the mineral phases present.

Altering the brightness scale distinguished the ferronickel particles from the surrounding oxide

matrix. Energy dispersive X-ray spectroscopy (EDS) helped verify the presence and chemical

composition of the mineral phases detected.

Thermogravimetric and Differential Thermal Analysis (TGA/DTA) was conducted using

a Netzsch STA 449 F3 Jupiter. Loaded samples were placed in alumina crucibles which were

mounted onto a carrier connected to a balance below. Mass changes as a function of time and

temperature, were monitored as the sample was heated at specific heating rates and gas

atmospheres. A neutral atmosphere was maintained by flowing 50 – 150 mL/min of nitrogen

46
through the equipment. The unit performed differential thermal analysis by comparing the

sample temperature with a blank reference sample; differences in temperature indicated

exothermic or endothermic reactions and identified phase changes. Powdered samples (25 – 30

mg) were analyzed in 0.3 mL alumina crucible while larger 5 mL crucibles were utilized to

analyze pellet samples (3000 mg). Only thermogravimetric analysis could be performed on the 5

mL crucibles as the size of the crucible prevented the mounting of a reference crucible on the

sample carrier. The buoyancy of the sample mass caused by the flow of gas was accounted for by

performing correction tests prior to each run.

47
 Chapter 4: Results and Discussion

4.1 Preliminary Analysis of Thermal Upgrading Requirements

The aim of the work performed in this thesis was to conduct thermal upgrading of a

limonite ore starting with the conditions claimed to produce a FeNi concentrate by the Inco

patent. The process conditions could then be altered to meet the requirements of the limonite ore

tested in order to achieve suitable upgrading results. Further study was conducted to investigate

the growth and morphology of the ferronickel particles.

The ranges of coal and sulphur additions used in the patent were defined specifically for

the limonite ore tested. Reagents additions could be altered based on preliminary reduction

potential calculations to accommodate variations in the limonite ore used in this study. By

assuming both gaseous and solid state reduction mechanisms, the range of coal addition was

estimated assuming complete metallization of Ni and the reduction of all iron oxides to the FeO.

Coal additions were estimated to be 3.7 – 7.1%, which is the range of coal additions required to

establish a selective reduction environment. The sample calculation for the required coal

additions can be found in the Appendix A. The estimated coal addition range agrees with the

values published in the patent of 4 – 6 wt%.

4.2 Degree of Reduction

Precise control over the reducing atmosphere was required to selectively reduce the

goethite phase to wustite while also promoting the formation of metallic nickel and some metallic

iron. To ascertain the presence of wustite and metallic nickel within the upgraded ore, X-ray

diffraction analysis was conducted on powdered upgraded pellets. Multiple scans were

conducted, one sample upgraded with 6% coal is shown in Figure 17 below. Wustite was shown

48
to be present within the upgraded ore along with ferronickel. A separate metallic nickel phase

was not detected. Metallic nickel was found only in a ferronickel alloy phase, which was

expected as work by Antola et al. (1995) concluded similar findings. The XRD analysis

illustrated that the reducing potential established through additions of 6% coal was sufficient to

reduce the iron oxide phase to wustite while limiting the formation of metallic iron. Given that

the XRD results are only qualitative, the extent of conversion could not be determined.

W
W—Wustite W
W
FeNi— Ferronickel

FeNi

FeNi

Figure 17: XRD pattern of upgraded ore

The high iron content of the limonite ore produced significant background noise when

performing XRD analysis using a Cu kα radiation tube. To reduce the background noise, a

secondary scan of the upgraded samples utilizing a cobalt radiation tube was conducted and is

shown in Figure 18 below. The scan detected the presence of wustite and metallic iron, which was

shown earlier to be associated with metallic nickel. The cobalt scan identified minor peaks of

fayalite at multiple 2θ positions. Fayalite formation is acceptable for the thermal upgrading

49
process as the iron silicate is non-magnetic and should report to the tailings during magnetic

separation. Some nickel incorporation into the iron silicate matrix could occur, rendering the

nickel inert to magnetic separation and thus reducing recovery values. Fayalite may also coat

oxide grains and impede the reduction process (Zevgolis, Zografidis, & Halikia, 2010). The

reduction of fayalite to unwanted metallic iron could also occur, although stronger reducing

potentials and higher temperatures are required (Harris, Peacey, & Pickles, 2009). Fayalite

formation is expected to be minimal when thermally upgrading limonite ore due to the limited

silica content contained within the ore. Re-oxidation of metallic nickel and incorporation into the

wustite phase despite the reducing conditions may also occur as was noted by Hallet (1997). No

magnetite was detected in any of the XRD scans of the upgraded ore. The presence of magnetite

could indicate the destabilization of the wustite phase either through oxidation or

disproportionation reactions. The quantity of magnetite formed may be below the detection limit

of the XRD but still sufficient enough to magnetize some wustite particles.

W
W—Wustite
W Fe—Metallic Iron
Fe
W f—Fayalite

W
f f f W Fe
f f Fe
f W

Figure 18: XRD pattern of upgraded ore using Co tube

50
 The amount of nickel and iron metallization was determined using 5% bromine/ 95%

methanol diagnostic leach test as outlined by Young (1974) and Sotka et al. (2004). The

bromine/methanol solution selectively dissolved metallic phases while leaving the oxide phases

insoluble. Leach tests were performed on upgraded pellets with sulphur additions of 2 or 4 wt%

and subjected to growth zone temperatures of 1000oC or 1100oC. The results from the diagnostic

leach tests are shown in Table 8 below. For all tests, the degree of nickel metallization was

greater than that of iron, confirming thermodynamic predictions that NiO reduces at lower

reduction potentials than iron oxide. Complete reduction of nickel oxide was not achieved as the

highest degree of nickel metallization was 81.8%. The degree of nickel and iron reduction

obtained exceeded the results of Zevgolis et al. (2010) where Ni and Fe reductions degrees did

not exceed 33% and 76% respectively. . Higher degrees of nickel metallization occurred in

tandem with higher degrees of iron metallization, suggesting a similar reduction mechanism. Poor

selectivity was achieved as Fe metallization degrees were greater than 40%. Increases in sulphur

additions and growth zone temperatures increased the degree of metallization of both nickel and

iron.

Table 8: Degree of metallization of upgraded ore determined by bromine/methanol diagnostic leach

Sample # Coal Sulphur Growth Zone Metallics Ni Fe Total

(wt %) (wt%) Temperature (wt%) Metallization Metallization Metallization
(oC) (%) (%) (%)

TF-17 6% 2% 1000 19.9% 68.7% 41.5% 42.3%

TF-18 6% 4% 1000 21.1% 72.7% 43.9% 44.7%
TF-19 6% 2% 1100 22.7% 78.4% 47.4% 48.3%
TF-20 6% 4% 1100 23.7% 81.8% 49.4% 50.3%

51
 The weight percent of metallics contained within the upgraded pellets ranged from 19.9%

- 23.7%. Under optimal separation conditions only metallics would be recovered to the

concentrate with minimal contamination due to non-metallic magnetic phases. Assuming 100%

recovery of metallics, the ferronickel concentrate produced from the samples in Table 8 would be

approximately 4.8% Ni with a Fe/Ni ratio of 20. The high content of metallic iron dilutes the

concentrate, driving down Ni grade and increasing the Fe/Ni ratio. Assuming perfect separation

conditions, the FeNi concentrate that could be produced from thermally upgrading this limonite

ore would fall short in terms of grade and iron/nickel ratio when compared to the patent results.

Stoichiometric reduction calculations, giving the additions of 6% coal through a solid

state/gaseous intermediate reduction mechanism found that complete metallization of NiO should

occur. Thermodynamically, NiO should preferentially reduce before iron oxide reduction occurs.

As experimental results showed, 68.7 – 81.8% metallization of NiO was achieved; suggesting

reduction of the contained nickel was being restricted. Stoichiometric calculations assumed NiO

was present only in the goethite phase, which may not accurately reflect the composition of the

limonite ore as some nickel could be substituted into the silicate phase. Reduction of silicate ore

requires stronger reducing potentials compared to the tested conditions. Complete metallization

may also be restricted due to the highly dispersive nature of Ni within the ore, limiting the

interaction between NiO and the reductant.

Based on stoichiometry, the degree of iron metallization was calculated to be about 55%

assuming reduction via a solid-state carbon reductant. The actual reduction mechanism most

likely proceeded through a mix of solid state and gaseous mechanisms. Only reduction through a

solid carbon reductant mechanism was used in the theoretical calculation in order to simplify the

process. Experimental results of 41.5% - 49.4% degree of iron metallization are lower than the

52
theoretical predications of 55% metallization. The lower degree of iron metallization suggests

underutilization of the carbon reductant, possibly due to poor contact interaction between the

reductant and the ore. High iron metallization with a 6% coal additions has been shown to occur

experimentally (Diaz, et al., 1993; Zhu, et al., 2012). Lower coal additions may be required to

reduce the amount of metallic iron produced, thus reducing the dilution of the ferronickel

concentrate. Although smaller coal additions might reduce the recoverable metallic nickel

produced.

The percent of metallics provide a bench mark for comparing the experiments conducted

in this work with the work in U.S. Pat. 5,178,666. The lower range of magnetic fraction (11.2% -

15.6%) achieved in the patent illustrates the process was able to produce recoverable FeNi while

minimizing amount of iron reduction to the metallic state. In this study, the reducing condition

applied was effective in reducing oxide material to the desired states, but the degree of selectivity

of nickel oxide reduction over iron oxide reduction was limited. Significantly more metallic iron

was produced which increased the magnetic fraction recovered (19 – 24%). The limitation may

be a result of the close association of the nickel and iron oxides within the goethite matrix.

4.3 Bench-Scale Testing of the Thermal Upgrading Process

Bench scale experimental work was initially conducted using the conditions outlined in

the INCO patent in order to thermally upgrade a limonite ore. Attempts were made to modify the

process to improve recovery and grade of the ferronickel concentrate. Modifications to the

process included furnace type, pre-calcination of ore, grinding method and type of magnetic

separator utilized.

53
4.3.1 Thermal Upgrading Process Test Work using Stationary Bed Furnace
A series of tests were conducted as shown in Table 9, to study the effect of sulphur

addition, particle growth zone temperature and the difference between reducing raw ore versus

pre-calcined ore. The upgraded samples was cooled under a N2 atmosphere in the water cooled

section of the furnace for 40 minutes reaching temperatures ranging from 60 – 80oC. The cooled

pellets were then ground with a motor and pestle, slurried and magnetically separated using a

Davis tube set a 4800 Gauss. The concentrate produced was digested with aqua regia, then

analyzed using AA spectroscopy. Complete digestion of the upgraded samples could not be

achieved with aqua regia. The undissolved residue remaining after digestion was believed to be

mainly composed of iron silicates such as fayalite. Sulphur additions and particle growth zone

temperatures were varied between 2 - 4% and 1000 – 1100oC respectively, which were the ranges

specified in the patent. O’Connor et al. (2006) showed that pre-calcination of limonite ore

opened the goethite mineral structure as the dehydroxylation reaction occurred, and allowed for

rapid interaction between the nickel species and gaseous reductant. To study the effects of pre-

calcination, limonite ore was roasted at 600oC in air for 4 hours to ensure that complete

dehydroxylation of the goethite gangue to hematite was achieved. The calcined ore was then

briquetted with the required additions of coal and sulphur then upgraded under the same

conditions as were applied to the raw ore sample tests. Reduction of pre-calcined ore resulted in

higher degrees of nickel metallization when compared to the reduction of raw ore. The addition

of a pre-calcination step prior to mixing the ore with the coal and sulphur additives could improve

the reduction mechanism, and limit the loss of additives before reaching the reducing and growth

zones temperatures.

54
 Table 9: Preliminary upgrading test work with stationary bed furnace setup and Davis tube
magnetic separator, TF-1 to TF-4 pre-calcined ore, TF-5 to TF-8 raw ore

Magnetic Fractions

Thermal
Upgrading
Temperature Ni Grade Ni Recovery Fe/Ne
o
Test C %S % Coal wt % % % wt% Ratio
TF-1 1000 2 6 76.10% 1.34% 74.08% 62.24
TF-2 1000 4 6 87.70% 1.35% 85.78% 65.14
TF-3 1100 2 6 88.16% 1.33% 85.20% 64.67
TF-4 1100 4 6 86.75% 1.42% 89.46% 62.62
TF-5 1000 2 6 89.74% 1.26% 81.80% 62.82
TF-6 1000 4 6 85.39% 1.30% 80.66% 66.87
TF-7 1100 2 6 84.72% 1.31% 80.41% 68.47
Tf-8 1100 4 6 81.25% 1.39% 82.05% 63.32

No definitive trend could be concluded in terms of the effect of sulphur additions and

growth zone temperature on the quality of magnetic concentrate produced. A slight improvement

in nickel recovery was noted with the runs using pre-calcined ore when compared to the tests

conducted using raw ore. Overall the test results compare poorly against the results produced in

the INCO patent. The poor grade values can be attributed to the dilution of the concentrate with

metallic or oxidized iron as shown by the Fe/Ni ratio (INCO Patent Fe/Ni ratio typically < 5). The

bulk of the feed material (typically >80%) was recovered in the magnetic concentrate meaning

limited gangue rejection was achieved. An acceptable range for the weight percent magnetic

fraction is 11 – 16%.

The XRD analysis shown in Figure 17 illustrates the reducing conditions were sufficient

to reduce the iron oxide gangue to the wustite/metallic iron phase boundary. The excessive

retention of iron to the magnetic concentrate could be due to the destabilization of the wustite

phase to magnetite via oxidation or disproportionation. The pellets were cooled in a neutral
55
atmosphere of ultra-high purity nitrogen to minimize oxidation of wustite. Oxidation could occur

rapidly in subsequent liberation and separation stages due to highly reactive iron oxide fines.

Disproportionation of wustite could have occurred, although rapid cooling of the sample should

have limited the kinetics of the reaction as shown by Broussard (1969) and Diaz et al. (1993).

Mori (1999) performed a kinetic study of the disproportionation reaction and suggested the

process exhibited creep phenomenon characteristics as a result of having a wide and flat

distribution of effective thermal activation energy. Removing the thermal activation energy

quickly should limit the kinetics of the disproportionation reaction, therefore stabilizing the

wustite phase. In the Inco patent, the wustite phase was stabilized by maintaining a

neutral/reducing atmosphere throughout the entire process and by rapid cooling of the upgraded

pellet from 1000/1100oC down to room temperature in 30 - 60 minutes. Longer cooling times (12

– 16 hrs) resulted in more ferric iron in the form of magnetite to be produced (Diaz, et al., 1993).

In this study a cooling time of 40 min was used which allowed the temperature of the pellets to be

cooled to 60 – 80 oC.

Finite magnetite formation on the wustite particles could be enough to magnetize the

whole particle. These magnetized wustite particles could then be recovered and further dilute the

concentrate. Alternatively, over reduction producing excess metallic iron could have contributed

to the high proportion of magnetic material. Quantitative analysis to determine the distribution of

iron into the magnetite, wustite and metallic phases could help to determine the degree of

reduction and subsequently the required reductant addition when variations to the feed ore occur.

The poor separation result may have also been attributed to the characteristics of the

Davis tube magnetic separator. The high magnetic field strength (4800 Gauss) applied by the

Davis tube resulted in clumps of magnetic material that would entrap the non-magnetic gangue

56
thus preventing separation. Multiple attempts to improve separation by altering the parameters of

the Davis tube (agitation and flow rate) were conducted with minimal effect. The Davis tube was

exchanged for a High Intensity Wet Magnetic Separator (HIWMS) which provided more control

over the magnetic field strength and allowed for larger sample sizes (≥50 g) to be processed. The

HIWMS utilized a collection box filled with iron spheres to capture the magnetic particles. The

magnetically charged iron spheres provided collection points for the particles to attach and

separate from the slurry. The small ferronickel particles produced during thermal upgrading are

believed to be more susceptible to separation using the HIWMS compared to the Davis tube.

Preliminary screening tests were performed using the HIWMS to determine the appropriate

strength of the magnetic field to be applied to separate the upgraded pellets. The results from the

preliminary screening tests are shown in Figure 19 below.

100.00%
90.00%
80.00%
% Magnetics (wt %)

70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
0 0.5 1 1.5 2 2.5 3
Field Strength (amps)

Figure 19: Effect of magnetic field strength of HIWMS on the mass fraction of magnetics extracted

57
 A magnetic field strength corresponding to an applied current of 0.5 amps was found to

produce a concentrate with a weight fraction of approximately 20%. The HIWMS was able to

achieve significantly more gangue rejection compared to the Davis tube and thus produce a leaner

concentrate. Further test work utilizing the HIWMS was conducted at 0.5 amps. Two sets of

tests were performed using the INCO patent conditions and then separated using the HIWMS

producing the results as shown in Table 10 below.

Table 10: Thermally upgraded 3g pellets with 6% coal at a thermal upgrading temperature of
1000oC and separated using HIWMS

Experiment # Sulphur wt% Ni Grade (wt %) Ni Fe/Ni

(wt %) Magnetic Recovery wt% ratio
(%)
STA-1 2% 33.3% 0.9% 18.2% 83.3
STA-2 4% 25.6% 1.5% 22.4% 48.5
STA-3 2% 8.5% 1.3% 6.3% 57.8
STA-4 4% 13.3% 3.2% 23.1% 25.2

A significant drop in nickel recovery was noted when using the HIWMS compared to the earlier

tests conducted using the Davis tube. Loss of material during handling and the initial small

sample size were believed to be the attributing factors influencing the lower recovery values.

Additional test work was to be conducted at larger sample sizes. A marked improvement was

noted in the weight fraction of magnetic concentrate produced using the HIWMS. Specifically,

test runs STA-3 and STA-4 achieved concentrate weight fractions of 8.52% and 13.30%

respectively which resembled the values obtained in the patent experiments. In both sets of

experiments the higher sulphur additions of 4% resulted in increases to both nickel grades and

recoveries, while also lowering the Fe/Ni ratio found in the concentrate. Further study of the

effect of sulphur on nickel recovery and grade was performed.

58
 The results displayed in Table 10 show an improvement in the attempt to thermally

upgrade the limonite ore. Significant differences in grade and recovery exist between the results

from this study and the results published in the patent. The dilution of the nickel in the

concentrate by iron is believed to be the main reason for the difference in Ni grades between the

two experimental studies. In the patent, a higher grade (11 – 15 wt% Ni) ferronickel concentrate

was produced because successful rejection of the iron oxide gangue in the form of wustite was

achieved. In the current study, a wustite phase was shown to be present in the thermally upgraded

ore, but it is believed that the degradation of this phase to a magnetic material, mostly likely

magnetite, occurred prior to separation. The formation of highly magnetic magnetite, or metallic

iron, on the surface of a wustite particle is believed to be enough to magnetize the entire particle.

Stabilizing the wustite phase has proven difficult as it requires precise control of the reducing

atmosphere and temperature during heating and cooling of the calcine.

To reduce the effect of material loss during handling, 15g sample sizes were thermally

upgraded then magnetically separated using the HIWMS. The 15 g samples were thoroughly

mixed then pressed into multiple pellets and charged to the stationary tube furnace. A set of three

multiple runs were performed and the separation results are shown in Table 11 below. The

separation recovery was determined based on mass balance around the magnetic separation stage.

The nickel head grade in the upgraded pellets was 1.38%, which was equal to the head grade of

the original ore. Nickel grade was expected to increase after reduction due to the removal of

water, yet the results of the assay indicated no change.

59
 Table 11: Multiple 15g samples with 6% Coal, 4% S at thermal upgrading temperature of 1000oC
separated using HIWMS

Magnetic Separation
Ni Grade Fe/Ni
Experiment Fraction Recovery
(wt %) ratio
(wt %) (%)
TF-25 27.7% 3.8% 83.7% 29.2
TF-26 28.8% 4.0% 93.2% 25.0
TF-27 31.0% 3.8% 93.1% 27.3

Using a larger sample size, recovery values of 83.7% to 93.2% were achieved. The

experimental grade and Fe/Ni ratios obtained from these tests compare favorability to the

theoretical values (4.8 wt% Ni and 20 Fe/Ni ratio) predicted from the reduction studies earlier

when 100% recovery of metallics was assumed. Lower grade values with corresponding higher

Fe/Ni ratios indicates dilution of the concentrate by iron and not unwanted silicate material.

Magnetic separation seems to be effective in recovering the ferronickel from the gangue material.

The nickel grade was upgraded from 1.38% Ni to 3.8 – 4.0%, indicating that the thermal

upgrading process could selectively separate the FeNi from the wustite phase. The concentrate

grade produced was diluted with iron as illustrated by the high Fe/Ni ratio and weight magnetic

fraction. The results displayed in Table 11 reflect the best results achieved in this study as

compared to the findings of Experiment 1 in the U.S. Pat. 5,178,666. To further improve results,

the focus should be on converting and stabilizing more of the iron oxide gangue as wustite in

order for it to be rejected during the separation process.

The effect of coal addition at 2, 4 and 6 wt% on nickel grade and recovery was studied

and is shown in Figure 20 below. Altering coal addition amounts would vary the reducing

potential within the furnace, and would primarily impact the degree of iron oxide reduction. At 2

wt% coal, limited grade upgrading and lower recoveries were achieved, which was attributed to

60
the lower reducing potential established. The dilution of the magnetic concentrate was believed to

be due to magnetite formed at the weaker reducing condition. Lower reducing conditions would

produce less FeNi reducing the amount of recoverable metallic material. As the coal addition was

increased to 4 wt%, the formation of more wustite allowed for greater iron rejection producing a

cleaner FeNi concentrate. The reduced Ni recovery was unexpected as recovery should increase

in the presence of a stronger reducing environment. The highest recovery was achieved when a 6

wt% coal addition was added to the feed. At this addition the reduction potential was believed to

be at the wustite/metallic iron boundary. The increase in formation of metallic iron allowed for

improved FeNi particle growth but also resulted in the dilution of the magnetic concentrate. A

coal addition of 6 wt% provided the highest recovery of Ni at approximately 75% with some

upgrading in the Ni grade values (from 1.38% to 1.96% Ni) compared to lower additions. No

coal addition value provided ideal separating conditions as lower coal addition lead to greater

reduction selectivity but poorer recovery and vice versa for higher additions.

61
 3.00% 100.0%
Head Grade 1.38% Ni
90.0%
2.50%
80.0%

70.0%
2.00%

Ni Recovery (%)
Ni Grade (wt%)

60.0%

1.50% 50.0%

40.0%
1.00%
30.0%

20.0%
0.50%
10.0%

0.00% 0.0%
0.0% 2.0% 4.0% 6.0% 8.0%
Coal Addition (wt%)

Ni Grade Ni Recovery

Figure 20: Effect of coal addition on recovery and grade of nickel in magnetic concentrate, 10 g
samples with 4% sulphur

The effect of sulphur additions between 0 – 4% on the separation of the thermally

upgraded pellets was studied. Sulphur is believed to promote the growth of ferronickel particles

allowing for effective separation of the metallic particles from the surrounding iron oxide gangue.

The recoveries and grades obtained from the experiments testing the range of sulphur additions

are shown in Figure 21 below. Improved nickel recovery was noted to begin at 2% sulphur

addition, increasing up to 4% which agrees with the suggested sulphur additions range outlined in

the patent. No improvement in nickel recovery was noted at sulphur additions under 2%, which

does not agree with the findings of Bell and Sridhar (1977) and Jiang et al. (2013) who suggested

62
that sulphur additions above 2% would hinder recovery of nickel by magnetic separation. The

effect of sulphur on Ni concentrate grade was less prevalent, with a minimal changed noted over

the tested range. The nearly constant nickel concentrate grade corresponding with an increase in

recovery could indicate that at higher sulphur additions the recovery of iron increases, resulting in

dilution of the concentrate. Nickel grade remained nearly constant over the range of sulphur

additions tested which was not expected. Kukura et al. (1981) and Li et al. (2012) found that

sulphur preferentially sulphidizes metallic iron reducing the content within the ferronickel alloy.

Improved mobility established with higher sulphur additions could allow the agglomeration of

more metallic iron particles onto the ferronickel particles, increasing the metallic Fe/Ni ratio.

Sulphur may be acting as a reducing agent for the reduction of hematite/magnetite beginning at

approximately 400oC. Sulphur reduction of the higher order iron oxide could free up carbon to

reduce more FeO. The combined reducing potential of carbon and higher additions of sulphur

may result in higher degrees of metallization of iron.

63
 2.50% 80.0%

Head Grade 1.38% 70.0%

2.00%
60.0%
Ni Grade (wt%)

Recovery (%)
1.50% 50.0%

40.0%
1.00% 30.0%

20.0%
0.50%
10.0%

0.00% 0.0%
0% 1% 2% 3% 4%
Sulphur (wt%)

Grade Recovery

Figure 21: Effect of sulphur addition on recovery and grade of nickel in magnetic concentrate, 10 g
samples with 6% coal

4.3.2 Thermal Upgrading Process Test Work using Rotary Kiln

In an attempt to try to improve particle growth of the ferronickel particles, a rotary kiln

experimental setup was utilized in place of the stationary tube furnace. The rotational force

applied on the purposed Fe-S-O liquid phase due to the rotation of the furnace should increase the

mobility of the ferronickel particles allowing them to congregate into larger particles. A single

charge of 300 g of pressed pellets was upgraded using the rotary kiln setup. Multiple 50 g grab

samples were taken from the upgraded ore and separated using the HIWMS and the results are

shown in Table 12 below. The results obtained using the rotary kilns are similar to the results

achieved using the stationary tube furnace with 15 g samples sizes shown in Table 11. The

rotational effects may be limited by the nature of the pelletized charge feed. The pelletized feed

64
remained intact during the experiment, limiting the mobility of the liquid phase. Pelletizing the

feed was required to improve reaction kinetics, and limit dust formation but may hinder particle

growth.

Table 12: Magnetic separation results of thermally upgraded ore using rotary kiln setup, charge feed
of 300g at 6% coal, 4% sulphur

Experiment # wt% Overall Separation Fe/Ni wt%

Magnetic Ni Grade Recovery ratio
(wt %) (%)

RK-1-3 23.5% 4.0% 60.5% 20.8

RK-1-4 24.9% 3.4% 60.7% 22.9
RK-1-5 27.8% 3.9% 80.3% 21.5

Recovery values from the rotary kiln experiments compared favorably with metallization

degrees, indicating recovery of nickel in the metallic phase and not in unwanted oxide or silicate

phases. Lower nickel recovery values obtained using the rotary kiln setup compared to the

stationary bed setup may be a result of material loss during the thermal upgrading process. As the

charge feed to the rotary kiln was placed in a quartz capsule that rotated, it was impractical to line

the capsule with magnesium oxide powder, as was done with the quartz sample carrier used in the

stationary tube setup. Without the lining, the charge sintered with the quartz material when

temperature values were greater than 1000oC. The sintered calcine could not be recovered and

resulted in loss of material. Therefore, limited testing was performed using the rotary kiln setup.

Future work could utilize a refectory capsule to hold the charge feed within the kiln.

65
4.4 Thermogravimetric and Differential Thermal Analysis
Figure 22 illustrates the TGA/DTA curves for the limonite ore roasted up to 1000oC

under a nitrogen atmosphere. Two mass loss stages were identified over the given temperature

range. The first mass loss stage occurred due to the evaporation of free moisture which occurred

between 36 – 223oC and resulted in a gradual loss of 1.42% of the initial mass. The second mass

loss stage occurred between 223 - 347oC and represents the removal of crystalline water by the

dehydroxylation of goethite into hematite. The dehydroxylation reaction resulted in a mass loss

of 9.19% and the overall mass loss was 12.66%. An endothermic peak at 296oC corresponded

with the decomposition of goethite matrix. The peak temperature compared favorability with

values obtained by O’Connor et al. (2006) and was indicative of poorly crystalline goethite

structure. The crystal lattice of the goethite phase is expected to be highly substituted. The formed

hematite is stable under a neutral atmosphere as the mass loss was minimal at higher

temperatures. Some residual crystalline water may have remained within the ore up to

temperatures of 700 – 750oC and may account for the gradual mass loss that continued at elevated

temperatures (Harris, Peacey, & Pickles, 2009).

66
 100 1.60E+00

98 1.40E+00

1.20E+00
96
1.00E+00
Mass Loss (%)

DTA (μV/mg)
94
8.00E-01
92
6.00E-01
90
4.00E-01
88
2.00E-01

86 0.00E+00

84 -2.00E-01
0 200 400 600 800 1000
Temperature (oC)

Mass Lost DTA

Figure 22: TGA/DTA curves of limonite ore up to 1200oC under N2 atmosphere

67
 100

95

90
Mass Loss (%)

85

80

75

70
0 100 200 300 400 500 600 700 800 900 1000
Temperature (oC)

Limonite Limonite + Coal Limonite + Coal + Sulphur

Figure 23: Comparison of mass loss curves for blends of limonite ore with reagents, roasted up to
1000oC at 20oC/min under N2 atmosphere, blends of ore - coal - sulphur at 2:1:1 weight ratio
respectively

Figure 23 shows the mass loss curves for the reactions of blends of limonite with coal and

sulphur additives roasted up to 1000oC under a nitrogen environment. The limonite curve

compares favorably to the mass loss curve in Figure 22, illustrating consistency of the limonite

ore and reproducibility of the TGA/DTA experimental test work. The limonite – coal blend

resulted in a smaller mass loss (≈ 5.0%) during the dehydroxylation stage when compared to the

mass loss curve for limonite. As equal amounts of limonite were used in each blend, the presence

of coal seemed to inhibit the complete dehydroxylation of the contained goethite. Both curves

68
exhibited a total mass loss of approximately 13% after reaching 1000 oC. Complete

decomposition of the goethite in the limonite – coal blend may still have occurred although at a

diminished rate. The limonite – coal – sulphur blend exhibited the largest mass loss during the

dehydroxylation stage. The excess mass loss was attributed to the rapid gasification of sulphur

which was found to occur between 218 – 400oC. An additional mass loss stage was detected at

500oC which did not occur in the previous samples with limonite and limonite – coal blends. The

mass loss curve for the limonite – coal – sulphur blend with corresponding DTA curve is shown

in Figure 24 below.

100 9.00E-01

8.00E-01
95
7.00E-01

6.00E-01
90
Mass Loss (%)

DTA (μV/mg)
5.00E-01

85 4.00E-01

3.00E-01
80
2.00E-01

1.00E-01
75
0.00E+00

70 -1.00E-01
0 100 200 300 400 500 600 700 800 900 1000
Temperature (oC)

Mass Loss DTA

Figure 24: TGA/DTA curves for blend of 2:1:1 weight ratio of ore - coal - sulphur, roasted up to
1000oC under N2 atmosphere

69
 An endothermic peak at onset temperature of 489.6oC corresponded with a mass loss of

3.12%. As the new mass loss stage occurred only when sulphur was added, therefore the

involved reaction was assumed to be dependent on the presence of sulphur. Sulphur could be

acting as a reductant to reduce hematite to magnetite during this mass loss segment. The

corresponding release of SO2 gas would be indicative of this reduction reaction occurring.

Figure 25 shows the TGA/DTA curves developed using the patent temperature profile and

a blend of limonite, coal and sulphur powders. The temperature profile did not accurately

replicate the profile used in the patent experiments as non-isothermal heating was implemented to

bring the sample up to the required heating sections. Mass loss occurred only during non-

isothermal heating sections, as no mass loss was evident during both isothermal stages. The DTA

curve revealed an exothermic peak with an onset temperature of 908.1oC during the cooling stage.

No mass loss was observed to correspond with the exothermic peak, which could indicated a

possible solidification temperature of a liquid phase present in the sample. The identification of a

phase change would support the proposal by Diaz et al. (1993) that the metal values migrate

within the pellets by way of a Fe-S-O liquid phase. It should be noted that no prior evidence was

found during the heating stages, which would have indicated melting or partial melting of the

sample.

70
 100 2.00E+00

95 1.50E+00

90 1.00E+00
Mass Loss (%)

DTA (μV/mg)
85 5.00E-01

80 0.00E+00

75 -5.00E-01

70 -1.00E+00

65 -1.50E+00
0 20 40 60 80 100 120 140 160 180 200
Time (min)

Mass Loss DTA

Time Range (min) Temperature Profile

0 – 30 Ramp up to 600oC at 20oC/min
30 – 90 Hold at 600oC
90 – 110 Ramp up to 1000oC at 20oC/min
110 – 170 Hold at 1000oC
170 – 200 Cooling at 30oC/min

Figure 25: TGA/DTA curves for upgrading limonite ore with coal and sulphur using simulated U.S.
Pat. No. 5,178,666 temperature profile

71
 Compressed limonite pellets with 6% coal and 4% sulphur were thermally upgraded

using the TGA/DTA apparatus in order to monitor the mass change in the pellets throughout the

temperature profile. Abrupt mass changes reflect various chemical and physical changes

occurring in the pellets during the thermal upgrading process. Figure 26 shows the effect of

metallic growth zone temperature between 1000 – 1300 oC on mass loss in the pellets. Each test

behaved similarly during the initial reduction stage at 600 oC and began to vary in mass loss

during the metallic growth zone stage which was expected. Elevated growth zone temperatures

(≈1300oC) were tested in an attempt to form large ferronickel particles (2 – 3mm) similar to the

particles produced by Huang et al. (2011) and the ``luppen`` product formed in the Nippon Yakin

Oheyama Process. Total mass loss increased with increasing growth zone temperature reaching

a plateau value of 32% for temperature ≥1200oC. At temperatures ≥1200oC the upgraded pellets

began to fuse to the alumina crucibles.

The effect of retention time in the metallic growth zone was compared at 1000 oC

between 30 – 120 minutes as shown in Figure 27 below. Retention time appeared to have minimal

effect on the mass loss in the pellets. No significant variation in the mass loss curves was noted

over the range of hold times tested.

72
 100

95

90
Mass Loss (%)

85

80

75

70

65

60
0 50 100 150 200 250
Time (Min)

1000 1100 1200 1300

Temperature Profile
Ramp up to 600oC at 50oC/min
Hold at 600oC for 60 min
Ramp up to growth zone temperature at 20oC/min
Hold at growth zone temperature for 60 min
Cooling at 25oC/min

Figure 26: TGA mass loss curves comparing growth zone temperatures for 3g limonite pellet with
6% coal, 4% sulphur

73
 100

95

90
Mass Loss (%)

85

80

75

70

65

60
0 50 100 150 200 250
Time (min)

60 mim 120 min 30 min

Temperature Profile
Ramp up to 600oC at 50oC/min
Hold at 600oC for 60 min
Ramp up to 1000oC at 20oC/min
Hold at 1000oC for desired retention time

Figure 27 TGA mass loss curves comparing growth zone retention times for 3g limonite pellet with
6% coal, 4% sulphur

74
4.5 Ferronickel Particle Growth
The effect of maintaining the limonite pellets within a growth zone on the agglomeration

and growth of ferronickel particles is essential to understanding the effectiveness of the INCO

patent. The patent performed the reduction stage at 600 oC for 60 min then established the

metallic growth zone at 1000 – 1100 oC with a residence time of 60 min. This temperature profile

enabled the growth of the ferronickel particles to a size which could be liberated and magnetically

separated efficiently. The patent did not provide a description of the ferronickel particles

produced in the experiments performed. Backscattered electron imaging coupled with EDS

scans of thermally upgraded pellets was used to identify the presence, size and composition of the

ferronickel particles. A thorough analysis of the SEM images was performed to insure the proper

identification of ferronickel particles. Rhamdhani et al (2009) studied the microstructure of

reduced laterite ore using SEM and EMPA analysis, and encountered difficulties distinguishing

between the bright images of magnetite and ferronickel from back scattered electron images.

High resolution (200 000 x magnification) secondary electron images found ferronickel nuclei

ranging from 15-20 nm on the surface and pores of magnetite particles.

Images of ferronickel particles were produced from backscattering imaging using an

SEM as shown in Figure 28 below. The dense metallic ferronickel particle was identified by the

brighter contrast when compared to the surrounding oxide matrix. The particle identified was

approximately 50µm in size. The particle does not accurately represent the average ferronickel

particles produced as the image was selectively chosen because of the clarity and size of the

particle displayed. The ferronickel particles produced throughout the pellets ranged in size from

sub-micron to 50 µm in size. A spot point EDS scan, shown in Figure 29, of the surface of the

particle indicates the presence of iron and nickel. No oxygen was detected by the EDS scan which

75
indicated that the particle was comprised of metallic species. Iron oxide phase with minor traces

of Si, Al, Mg, Mn, Cr and C were detected when the material surrounding the ferronickel particle

was scanned. Distinguishing the iron oxide phase as either wustite or magnetite could not be

performed. No nickel was detected by the EDS scan in the surrounding material. The grain size

of the iron oxide material was sub-micron in size, which could lead to the formation of slimes

when slurried and separated. Slimes are difficult to separate using any physical methods and their

presence could explain the poor separation results achieved in this study.

Figure 28: Backscattered image of ferronickel particle produced in stationary tube furnace

76
 FeNi Particle

Oxide Gangue

Figure 29: EDS scans of FeNi particle and surrounding oxide gangue shown in Figure 28

Scanning of the pellet surface at a magnification of 200x provides a wider image as

shown in Figure 30. Ferronickel particles ranging in sizes are seen throughout the entire pellet.

Residual particles of carbon can be seen on the surface along with voids imbedded into the pellet

77
that were produced by gasification of carbon and sulphur additives. The ferronickel particles are

highly dispersed throughout the pellet.

Figure 30: Backscattered image of pellet surface upgraded at 1000oC growth zone temperature

SEM imaging of thermally upgraded pellets via the rotary kiln setup was conducted to

examine the effect of gravitational forces on the agglomeration and growth of the ferronickel

particles. The applied rotation was believed to improve metallic mobility within the semi-liquid

phase allowing the congregation and agglomeration of the ferronickel particles. Figure 31

illustrates a ferronickel identified within a pellet upgraded using the rotary kiln setup. The

particle appeared to have formed through the agglomeration of smaller metallic particles of

78
similar composition, unlike the particle shown in Figure 28. The abundant presence of similar

sized ferronickel particles collecting at distinct locations could suggest that gravitational forces

applied affected the migration of the particles. The distinct outlines of the smaller ferronickel

particles remained intact, suggesting full integration into a single particle did not occur. Instead

ferronickel particle growth appeared to proceed through the semi-fusion of smaller particles as

was suggested by Watanabe et al. (1987). The Fe-S-O liquid phase as suggested by Diaz et al.

(1993), could allow for Liquid Phase Sintering (LPS) of FeNi particles. LPS, as described by

Geramn et al. (2009), is a variant form of the sintering cycle where liquid improves particle

growth as compared to solid-state sintering. The liquid wets the surface of the particles and

creates strong capillary forces that pulls the grains together rapidly increasing the density of the

pellet. The rearranged grains are now connected via liquid bridges where solid solubility within

the liquid improves transportation rates leading to coarsening and densification.

The particle shown in Figure 31 displays signs of LPS as the shapes of the grains are

similar to ones which are predicted to form when the liquid phase sinter mechanism has occurred.

LPS provided evidence of a liquid phase present during the thermal growth zone that drives FeNi

particle growth.

Ferronickel particle size was found to be similar when both experimental setups were

utilized. The high degree of dispersion of ferronickel particles throughout the pellet may limit the

probability of the particles coming in contact. Limited particle contact will restrict the

agglomeration and growth of the ferronickel phase. The EDS scan identified iron and nickel as

the main components of the particles, which compared favorably to previously identified

ferronickel particles.

79
Figure 31: Backscattered image of ferronickel pellet upgraded in rotary kiln with corresponding
EDS scan, particle displays signs of liquid phase sintering

80
 The effect of growth zone temperature on ferronickel particle growth was examined in

the range of 1000 – 1300 oC and SEM images of the upgrade pellets are shown in Figure 32

below. To obtain reproducible results at these elevated temperatures, pellets were placed in 5

mL alumina crucibles and subjected to the thermally upgrading heating profile while within the

TGA/DTA unit.

81
 1000 Co 1100 Co

1200 Co 1300 Co

Figure 32: SEM images of pellets upgraded at growth zone temperatures of 1000, 1100, 1200 and
1300oC

The images in Figure 32 show that as the growth zone temperature increased the presence

and size of the ferronickel particles increased. At 1000 oC the ferronickel particles were

predominantly located around the edges of the voids. The stronger reducing environment in the

area surrounding the voids provides the ideal conditions for ferronickel reduction and growth.

82
When reacted at 1100 oC the ferronickel particles were enlarged and the porosity of the

surrounding matrix had been reduced. The densification of pellets with increasing temperature

agrees with the earlier suggestion of liquid phase sintering. The trend continued as higher growth

zone temperatures are applied. Particle growth was shown to be affected by growth zone

temperature as particle size ranges from <20 microns at 1000 oC to >100 microns at 1300 oC as

shown in Figure 33 below. The grain size of the iron oxide phase has grown also with higher

temperatures but most particles remained <10 microns in size. At temperatures ≥ 1200 oC the

pellets began to fuse to the alumina crucibles and were difficult to remove. The densification of

the pellets with higher temperatures led to difficulties when attempting to section the pellets for

analysis.

83
 Figure 33: Backscattered electron image of sectioned pellet upgraded at 1300oC growth zone
temperature, large continuous FeNi particles are present throughout the sample

To study the effect of retention time on the growth of ferronickel particles a similar

approach using the TGA/DTA unit to upgrade 3 g pellets to produce samples for SEM analysis

was performed. The pellets were held at the growth zone temperature of 1000oC while the

retention time was varied from 30 – 120 minutes. Images at 500x magnification of the interior of

the pellets are shown in Figure 34 below. The images provided a general overview of the

84
consistency of the upgraded pellets. The fracturing of the compressed pellets was noticed to

increase with longer retention times. The bright spots associated with the presences of metallic

particles were difficult to identify when the retention time was 30 minutes. The metallic particles

appeared larger and more prominent throughout the samples at longer retention times. The effect

of retention time on the growth of metallic particles was found to be minimal when compared to

the effect of temperature. Figure 35 provides images of individual ferronickel particles produced at

the various hold times. The largest particles (approximately 20 µm) were found when the

retention time in the growth zone was 60 min. Ferronickel particles were submicron in size when

a retention time of 30 min was applied.

85
 30 min 60 min

120 min

Figure 34: SEM images at 500x magnification of thermally upgraded pellets with various hold times
at 1000oC with 6% coal, 4% S

86
 30 min 60 min

120 min

Figure 35: SEM image of FeNi particles at 5000x magnification, produced at various hold times at
1000oC with 6% coal, 4% S

87
 Figure 36 below shows three distinct phases present in the upgraded pellet, and are

identified using EDS analysis show in Figure 37. The phases were believed to be present in all

upgraded pellets at all growth zone temperatures and retention time variations. The image in

Figure 36 was chosen, as the distinct phases were clearly visible and easily identifiable. Label

(A) identifies an agglomerated ferronickel particle highly fused as the outlines of the smaller

particles were becoming less distinct. Position (B) illustrates a homogenous iron silicate phase,

mostly likely fayalite. The surrounding iron oxide matrix material (C) was believed to be

comprised of wustite or magnetite, as these two phases could not be distinguished in the SEM

image. The ferronickel particles were distinct separate phases that should be susceptible to

liberation techniques. Poor separation results may have been a result of the fine grained iron

oxide phase formed in the upgraded ore. Even if the iron gangue was rendered as wustite, the fine

grained structure could have formed slimes when slurried, limiting the effectiveness of physical

separation.

88
 A

Figure 36: SEM image of pellet upgraded at growth temperature of 1200oC; (A) FeNi particle, (B)
iron silicate, (C) iron oxide matrix

89
 (A)

(B)

(C)

Figure 37: EDS scans of Figure 36 phases; (A) FeNi particle, (B) iron silicate, (C) iron oxide matrix

90
 Chapter 5: Conclusions and Future Work

5.1 Conclusions
The thermal upgrading of nickeliferous laterite ore has been studied extensively, and

numerous patented processes have been filed. The most promising results were reported by Inco

in its U.S. Pat. 5,178,666, which established a novel process for the thermal upgrading via

selective reduction of limonite and limonite/saprolite blends to a ferro-nickel concentrate using

bench-scale and pilot plant experiments. In this study experimental work was performed to

thermally upgrade a limonite ore using the conditions established in the patent as a starting point.

Preliminary thermal upgrading experiments indicated that coal additions of 6 wt% were

sufficient at reducing the nickel and iron oxides to metallic nickel and wustite respectively. XRD

analysis identified phases of ferro-nickel, wustite and fayalite present in the upgraded pellets. No

magnetite was detected in the upgraded ore. Confirmation of a ferro-nickel and wustite phases

agrees with the objectives of the patent to segregate the ore into a metallic phase and a non-

magnetic oxide phase.

Bromine/methanol diagnostic leach tests confirmed that nickel oxide reduction occurred

at lower reduction potentials when compared to iron oxide reduction. The highest degree of

nickel metallization achieved was 81.8%. Metallization of nickel and iron increased with higher

sulphur additions and growth zone temperatures. Further metallization of nickel was prevented

due to nickel incorporation into iron silicates such as fayalite which restricted reduction.

Nonmetallic nickel could not be separated using magnetic separation thus reducing recovery and

grade values of the ferronickel concentrate.

91
 Pre-calcination of the ore in air at 600 oC prior to thermal upgrading was attempted to

improve reduction characteristics of the ore by opening the mineral structure of goethite,

therefore improving the contact area available between the ore and reducing atmosphere.

Experimental work showed minimal improvement in terms of recovery and grade of the

ferronickel concentrate produced. Sulphur additions beginning at 2 wt%, were found to improve

Ni recovery as additions between 0– 4 wt% increased recoveries from 13.8% to 75.8%.

The limonite ore displayed a typical two stage mass loss curve representing removal of

moisture and the dehydroxylation of goethite when heated under a nitrogen atmosphere. The

peak decomposition temperature of the goethite matrix was 296 oC, indicating poor crystallinity.

A blend of limonite with coal resulted in a smaller mass loss from 9.2% to 5% during the

dehydroxylation stage. The addition of sulphur to the blend resulted in significantly higher mass

loss during the dehydroxylation stage, followed by an additional mass loss stage. Gasification of

sulphur occurring between 218 – 400 oC may explain the increase mass loss during the

dehydroxylation stage.

When a two stage reduction/particle growth temperature profile was applied to pressed

limonite pellets, the TGA/DTA curves illustrated an exothermic peak at 908.1 oC during the

cooling stage with no corresponding mass loss. The peak could indicate the solidification of the

proposed Fe-O-S liquid phase which facilitates the growth of the ferronickel particles. Prior to the

cooling stage no indications of melting were observed during the TGA/DTA profile.

Ferronickel particles ≤60 μm were identified in the reduced pellets using SEM imaging

coupled with EDS scans. The particles were highly dispersed throughout the pellet. FeNi particles

were found to be conglomerates of smaller semi-fused particles, which suggested a similar

growth pattern to the model suggested by Watanabe et al. (1987). FeNi particles showed signs of
92
Liquid Phase Sintering, providing evidence towards the formation of a liquid phase. Increasing

the growth zone temperatures from 1000 – 1300 oC improved FeNi particle growth with particle

size increasing from < 20μm to approximately 100 μm over the temperature range. Temperatures

above 1200 oC led to the sintering of the limonite pellets to the alumina crucibles which reflects

the temperature where the ore becomes “sticky” and sinters to refractory lining. Longer retention

times were found to improve particle growth with the largest particles being produced after 60

min, which agreed with the optimum conditions set forth in the patent. The presence of sub-

micron iron oxide gangue was noted and believed to result in slime formation. Physical

separation of slimes can be difficult and could provide a reason for the poor separation results

obtained in this study.

Experimental magnetic separation tests produced lower nickel recovery and grade values

when compared to the results in the patent. High Fe/Ni ratios (>62.2) in the concentrate indicated

significant dilution due to iron bearing species. Utilizing larger sample sizes and a High Intensity

Wet Magnetic Separator over a Davis tube improved upgrading results. Recovery values ranged

from 83.7 – 93.2% with grades between 3.8 – 4.0 wt% Ni. These recovery and grade values

represent the best results obtained during this study. Limited testing was conducted using a rotary

kiln setup and produced similar grade values (3.4 – 4.0 wt %) but lower recovery values (61.6 –

80.4 %).

The experiments performed in this study were not able to thermally upgrade a limonite

ore to produce a high grade FeNi concentrate. The lack of iron rejection during magnetic

separation led to the dilution of the FeNi concentrate. Poor separation results were attributed to;

the destabilization of the wustite phase by either oxidation/disproportionation, slime formation

due to iron oxide fines and limited FeNi particle growth reducing liberation efficiency from iron

93
oxide gangue. The Inco patent, despite achieving good results that should have led to an

economic process, was never developed to the commercial scale suggesting that the results in the

patent are difficult to reproduce.

5.2 Future Work

The experimental work of this study focused on the selective reduction of limonite ore

using solid reagent additions reacted as a pelletized feed under stationary conditions in a bench

scale tube furnace. Results agreed with literature illustrating that selective reduction followed by

magnetic separation could be performed on limonite ore to produce a ferronickel concentrate.

However, the need to pelletize the feed material, low grades and recovery, and long retention

times limit the application of this process to bench scale experimentation. Further development of

the process flowsheet should be an area of research investigated.

The use of solid reactants (coal & sulphur) limits the reaction kinetics of the selective

reduction process due to initial solid-solid reactions. Utilizing a fluidized bed could improve solid

particle contact efficiency during initial reduction reactions while also providing larger surface

area for gas-solid interactions during the reduction via gaseous intermediates. Improved solid-

solid interactions would avoid the need to pelletize the feed material. Work performed by Antola

et al. (1995) and Orth and Kerstiens (2004) using fluidized bed for the reduction of laterite ore

showed improved heat/mass transfer, temperature control and lower retention over conventional

furnace setups. Reacting the laterite ore in suspended fluidized bed could limit the wear on the

refractory lining within the furnace, thus allowing the use of higher growth temperatures.

Fluidized bed reduction could alter the particle growth mechanism compared to the static bed

experiments, as the particles would be in a constant state of motion. The role of sulphur in a

94
fluidized bed would have to be re-evaluated as liquid formation could lead to the destabilization

of the fluidized bed.

Sulphur, required to promote ferronickel particle growth, was removed from the calcine

before being exposed to the growth zone conditions. Sulphur removal from the system was

believed to occur through volatilization as S2 or through the reduction of hematite as SO2. The

residual sulphur might not meet the compositional requirements to establish a Fe-S-O liquid

phase to allow ferronickel particle migration. Future work should focus on reducing the unwanted

loss of sulphur prior to utilization. A new experimental setup where sulphur is not contained in

the initial feed but introduced into the calcine during the growth zone, could improve sulphur

utilization during FeNi particle growth. Sulphur bearing species such as pyrite, which

decomposes to release S2 at the growth zone temperatures, may provide a means of stabilizing

sulphur until it is required.

An essential component of the thermal upgrading process was the bulk reduction of the

iron oxide gangue to non-magnetic wustite and stabilizing that phase in order to separate out the

magnetic ferronickel particles. Wustite can destabilize to magnetite or metallic iron through

oxidation and disproportionation reactions. Avoiding these reactions could improve process

performance by minimizing the dilution of the ferronickel concentrate by unwanted magnetic

material. Future work using Mossbauer Spectroscopy could elucidate the ratio of ferrous/ferric

ions present in the calcine. Understanding phase transitions of the iron oxide gangue could help in

attempts to stabilize the wustite phase.

95
 Bibliography

Agatzini-Leonardou, S., & Zafiratos, I. G. (2004). Beneficiation of a Greek serpentinic

nickeliferous ore Part II. Hydrometallurgy, 267-275.
Alcock, R. A. (1988). the Character and Occurrence of Primary Resources Available to the
Nickel Industry. In G. P. Tyroler, & C. A. Landolt, Extractive Metallurgy of Nickel &
Cobalt (pp. 67-89). Mississauga, Ontario: The Metallurgical Society .
Antola, O., Holappa, L., & Paschen, P. (1995). Nickel Ore Reduction by Hydrogen and Carbon
Monoxide Containing Gases . Mineral Processing and Extractive Metallurgy Review,
169-179.
Beggs, D., Bunge, F. H., & Vedensky, D. N. (1970, March 31). Patent No. 3,503,735. United
States.
Bell, M. C., & Sridhar, R. (1977, September 20). Patent No. 4,049,444. United States.
Betteridge, W. (1977). Nickel and its Alloys. Estover, Plymouth: MacDonald and Evans Ltd .
Boldt, J. R. (1967). The Winning of Nickel. Toronto: Longmans Canada Lt .
Broussard, L. (1969). The Disproportionation of Wustite. The Journal of Physical Chemistry ,
1848-1854.
Canterford, J. H. (1975). The Treatment of Nickeliferous Laterites. Minerals Science Engineering
, 3-16.
Canterford, J. H. (1978). Leaching of some Austrialian nickeliferous laterites with sulphuric acid
at atmoshpheric pressure. Proceedings of the Australasian Institute of Mining and
Metallurgy, 19-26.
Cartman, R. (2012). Nickel pig iron - A long term solution? 3rd Euronickel Conference .
Helsinki: Hatch .
Chander, S., & Sharma, V. N. (1981). Reduction roasting/Ammonia Leaching of Nickeliferous
Laterites. Hydrometallurgy , 315-327.
Cheung, W. H., & Valix, M. (2008). Selective Reduction Roasting of Limonite Ores: Effect of
Over-reduction. Proceedings of XXIV International Mineral Processing Congress, (pp.
2951-2957). Beijing.

96
Crama, W. J., & Baas, A. H. (1984, December 25). Patent No. 4,490,174. United States.
Crundwell, F. K., Moats, M. S., Ramachandran, V., Robinson, T. G., & Davenport, W. G. (2011).
Extractive Metallurgy of Nickel, Cobalt and Platinum-Group Metals. Oxford: Elsvier.
Dalvi, A., Bacon, W., & Osborne, R. (2004). The Past and the Future of Nickel Laterites.
Toronto: PDAC 2004 International Convention .
De Graaf, J. (1979). The Treatment of Lateritc Nickel Ores - A Further Study of the Caron
Process and Other Possible Improvements Part I. Effect of Reduction Conditions.
Hydrometallurgy , 47-65.
De-quing, Z., Yu, C., Hapugoda, S., Vining, K., & Jian, P. (2012). Mineralogy and crystal
chemistry of a low grade nickel laterite ore. Transactions of Nonferrous Metals Society of
China, 907-9016.
Diaz, C. M., Vahed, A., Shi, D., Doyle, C. D., Warner, A. E., & MacVicar, D. J. (1993, January
12). Patent No. 5,178,666. United States.
Ericson, A. S., Svensson, J., & Ishii, K. (1987). The MINPRO-PAMCO Nickel Segregation
Process. International Journal of Mineral Processing , 223-236.
German, R. M., Suri, P., & Park, S. J. (2009). Review: Liquid phase sintering. Journal of
Material Science , 1-39.
Habashi, F. (1997). Handbook of Extractive Metallurgy. In Vol.2. Primary metals, secondary
metals, lights metals (pp. 715-794). Weinheim : Wiley-vch.
Hallet, C. (1997). A Thermodynamic Analysis of the Solid State Reduction of Nickel from
Laterite Minerals . Proceedings of Nickel-Cobalt 97 International Symposium
Pyrometallurgical Fundamentals and Process Development (pp. 299-312). Sudbury,
Canada: C.A. Levac, R.A. Berryman (Eds.).
Harris, C. H. (2011). The Selective Sulphidation and Physical Upgrading of Nickel from a
Nickeliferous Lateritic Ore. Kingston, Ontario : Queen's Univeristy .
Harris, C. T., Peacey, J. G., & Pickles, C. A. (2009). Theraml Upgrading of Nickeliferous
Laterites - A Review. The 48th Annual Conference of Metalurgists . Sudbury: The
Metallurgy and Materials Society .
Harris, C. T., Peacey, J. G., & Pickles, C. A. (2011). Selectives sulphidation of a nickeliferous
lateritic ore. Minerals Engineering, 651-660.

97
Huang, D., Zhang, J., Mao, R., & Cao, M. (2011). Thermal behaviors and growth of reduced
ferronickel particles in carbon-laterite composites. Rare Metals, 681-687.
Iwasaki, I. (1972, April 18). Patent No. 3,656,935. United States.
Iwasaki, I., Takahasi, Y., & Kahata, H. (1966). Extraction of nickel from iron laterties and
oxidized nickel ores by a segregation process. Metallurgical Transactions, 308-320.
Jiang, M., Sun, T., Liu, Z., Kou, J., Liu, N., & Zhang, S. (2013). Mechanism of sodlium sulfate in
promoting selective reduction of nickel laterite ore during reduction roasting process.
International Journal of Mineral Processing .
Kawahara, M., Toguri, J., & Bergman, R. (1988). Reducibility of Laterite Ores. Metallurgical
Transactions , 181-186.
King, M. (2005). Nickel Laterite Technology - Finally a New Dawn? JOM, 35-39.
Kuck, P. H. (2012). 2010 Minerals Yearbook - Nickel (Advance Release). Reston, Virginia,
United States: U.S. Geological Survey.
Kuck, P. H. (2013). Nickel. In U. G. Survey, Mineral commodity summaries 2013: U.S.
Geological Survey (pp. 108-109). Reston,Virgina: U.S. Geological Survey.
Kukura, M. W., Stevens, L. G., & Auck, Y. T. (1981). Development of the UOP Process for
Oxide Silicate Ores of Nickel and Cobalt. ATB Metallurgie, 139-150.
Kwatara, M., Tayabally, J., Peek, E., & Schonewille, R. (2011). Segregation Roasting of a
Saprolitic Ore - An Experimental Investigation. TMS 2011 140th Annual Meeting &
Exhibition . San Diego: The Minerals, Metals and Materials Society .
Li, B., Wang, H., & Wei, Y. (2011). The reduction of nickel from low-grade nickel laterite ore
using a solid-state deoxidisation method. Minerals Engineering, 1556-1562.
Li, G., Shi, T., Rao, M., Jiang, T., & Zhang, Y. (2012). Beneficition of nickeliferous laterite by
reduction roasting in the presence of sodium sulfate. Minerals Engineering, 19-26.
Li, S., & Coley, K. S. (2000). Kinetics and Mechanism of Reduction Laterite Ore High in
Serpentine. The 39th Annual Conference of Metallurgists (pp. 179-192). Ottawa: CIM.
Liu, W.-r., Li, X.-h., Hu, Q.-y., Wang, Z.-x., Gu, K.-z., Li, J.-h., & Zhang, L.-x. (2010).
Pretreatment study on chloridizing segregation and magnetic separation of low-grade
nickel laterites. Trans. Nonferrous Metallurgical Society of Chinna, 82-86.
Lu, J., Lui, S., Shangguan, J., Du, W., Pan, F., & Yang, S. (2013). The effect of sodium sulphate
on the hydrogen reduction process of nickel laterite ore. Minerals Engineering, 154-164.
98
McDonald, R. G., & Whittington, B. I. (2008). Atmospheric acid leaching of nickel laterites
review Part I. Sulphuric acid technologies . Hydrometallurgy, 35-55.
McDonald, R. G., & Whittington, B. I. (2008). Atmospheric acid leaching of nickel laterites
review. Hydrometallurgy, 56-69.
Mori, K. (1999). Oxidation and Disproportionation of Wustite Studied by Mossbauer
Spectroscopy. Japanese Journal of Applied Physics, L189-L191.
Moskalyk, R. R., & Alfantazi, A. M. (2002). Nickel laterite processing and electrowinning.
Minerals Engineering, 593-605.
Mudd, G. M. (2009). Nickel Sulfide Versus Laterite: The Hard Sustainability Challenge Remains.
48th Annual Conference of Metallurgists (pp. 23-32). Sudbury, Ontario: Canadian
Metallurgical Society .
Mudd, G. M. (2010). Global trends and environmental issues in nickel mining: Sulfides versus
laterites. Ore Geology Reviews, 9-26.
Nestoridis, A. (1977, January 11). Patent No. 4,002,463. United States.
Neudorf, D. A., & Huggins, D. A. (2004). An Alternative Nickel Laterite Project Development
Model. International Laterite Nickel Symposium (pp. 63-76). Charlotte : The Minerals,
Metals & Materials Society .
O`Connor, F., Cheung, W., & Valix, M. (2006). Reduction roasting of limonite ores: effect of
dehydroxylation. International Journal of Mineral Processing , 88-99.
Onodera, J., Inoue, T., & Imaizumi, T. (1987). Attempts at the Beneficiation of Lateritic Nickel
Ore. International Journal of Mineral Processing, 19, 25-42.
Orth, A., & Kerstiens, B. (2004). Calcination and Reduction of Laterite Nickel Ores .
International Laterite Nickel Symposium (pp. 493 - 502). Charlotte, North Carolina: The
Minerals, Metals & Materials Society .
Pattek-Janczyk, A., & Miczko, B. (1990). Disproportionation Phenomena of Wustite Phase in the
Model Iron Catalyst for Ammonia Sythesis studied by Mossbauer Spectroscopy. Solid
State Ionics, 171-178.
Reznik, L. D., Tarasov, A. V., & Mayorov, A. D. (2003). Mechanism and technology of
segregation roasting of oxide nickel ores with subsequent calcine flotation and
concentrating leaching. EPD Congress 2003 (pp. 539-551). San Diego: The Minerals,
Metals & Materials Society .
99
Rhamdhani, M. A., Hayes, P. C., & Jak, E. (2009). Nickel laterite Part 1 - microstructure and
phase charaterisations during reduction roasting and leaching . Mineral Processing and
Extractive Metallurgy , 129-145.
Simons, C. (1988). The Production of Nickel: Extractive Metallurgy - Past, Present, and Future.
117th TMS Annual Meeting (pp. 91-134). Phoenix, Arizona: The Metallurgical Society,
Inc.
Sotka, P., Lamberg, P., & Lahtinen, M. (2004, December). Non-recoverable metals - do you
know how much you have? . Output, pp. 1-9 .
Thumm, W., & Heitmann, G. (1968, June 18). Patent No. 3,388,870. United States.
Utigard, T., & Bergman, R. A. (1992). Gaseous Reduction of Laterite Ore. Metallurgical
Transactions B , 271--275.
Valix, M., & Cheung, W. H. (2002). Effect of sulfur on the mineral phase of laterite ores at high
temperature reduction . Minerals Engineering, 523-530.
Valix, M., & Cheung, W. H. (2002). Study of phase transformation of laterite ores at high
temperature. Minerals Engineering, 607-612.
Wang, Q., Yang, Z., Tian, J., Li, W., & Sun, J. (1997). Mechanisms of reduction in iron ore-coal
composite pellet. Ironmaking and Steelmaking, 457-460.
Watanabe, T., Ono, S., Arai, H., & Matsumori, T. (1987). Direct Reduction of Garnierite Ore for
Production of Ferro-Nickel with a Rotary Kiln at Nippon Yakin Kogyo Co. Ltd.,
Oheyama Works. International Journal of Mineral Processing , 173-187.
Whittington, B. I., & Muir, D. (2000). Pressure Acid Leaching of Nickel Laterites: A Review.
Mineral Processing and Extractive Metallurgy Review, 527-600.
Young, R. S. (1974). Chemical Phase Analysis. New York: John Wiley & Sons.
Youping, Z., Yusheng, Z., & Zhaoyi, L. (2008). Technical analysis of producing low Ni pig iron
with laterite in a blast furnace. Baosteel Technical Research , 36-40.
Zevgolis, E., Zografidis, C., & Halikia, I. (2010). The reducibility of the Greek nickeliferous
laterites: a review . Mineral Processing and Extractive Metallurgy, 9-17.
Zhu, D. Q., Cui, Y., Pan, J., Li, Q. H., & Qiu, G. Z. (2008). Investigation on Manufacture of
Ferronickel from Low Grade Nickel Laterite by Prereducing-magnetic Separation-
smelting Process. Proceeding of XXIV Internationl Mineral Processing Congress, (pp.
1622-1633). Beijing.
100
Zhu, D. Q., Cui, Y., Vining, K., Hapugoda, S., Pan, J., & Zheng, G. I. (2012). Upgrading low
nickel content laterite ore using selective reduction followed by magnetic separation.
International Journal of Mineral Processing, 1-7.

101
Appendix A - US Patent 5,178,666

102
103
104
105
106
107
 Appendix B - Examples of Stoichiometric Reduction Calculations

Assuming reduction through gaseous intermediates

The Boudouard equation forms CO reductant from solid
𝐶(𝑠) + 𝐶𝑂2(𝑔) → 2𝐶𝑂(𝑔) carbon additions

𝑁𝑖𝑂(𝑠) + 𝐶𝑂(𝑔) → 𝑁𝑖(𝑠) + 𝐶𝑂2(𝑔) Nickel was assumed to be present as NiO and reduces to
metallic nickel at a 1:1 molar ratio with CO

2𝐹𝑒𝑂𝑂𝐻(𝑠) → 𝐹𝑒2 𝑂3 + 𝐻2 𝑂(𝑔) Goethite dehydroxylates to hematite which then is reduced

(𝑠)
to magnetite and subsequently wustite. Overall reduction
3𝐹𝑒2 𝑂3(𝑠) + 𝐶𝑂(𝑔) → 2𝐹𝑒3 𝑂4(𝑠) + 𝐶𝑂2(𝑔) reactions shows a molar ratio of 2:1 for Fe:CO
𝐹𝑒3 𝑂4(𝑠) + 𝐶𝑂(𝑔) → 3𝐹𝑒𝑂(𝑠) + 𝐶𝑂2(𝑔)

2 𝐹𝑒𝑂𝑂𝐻(𝑠) + 𝐶𝑂(𝑔) → 2𝐹𝑒𝑂(𝑠) + 𝐶𝑂2(𝑔) + 𝐻2 𝑂(𝑔)

Assuming 100 g sample of limonite ore with grade of 1.38% Ni and 45.70% Fe
Moles of Ni in ore 0.023512 Mol
(100g X 0.0138 / 58.6939 g/mol)
Moles of Fe in ore 0.818336 Mol
(100g X 0.457 / 55.845 g/mol)

Moles of CO required for Ni reduction 0.023512 Mol

(Molar ratio of CO:Ni =1)
Moles of CO required for Fe reduction 0.409168 Mol
(Molar ratio of CO:Fe =0.5)
Total Moles of CO required for reduction 0.43268 Mol

Moles of Carbon required 0.21634 Mol

(Molar ratio of C:CO = 0.5)
Mass of Carbon required to produce CO 2.598244 g
(0.21634 mol X 12.01 g/mol)
Mass of Coal required based on %Carbon value in Coal 3.839013 g
(2.598244 g / 0.677)

108
Estimated wt% Coal Addition 3.70 %

Assuming reduction through solid state reduction mechanism

Nickel was assumed to be present as NiO and reduces to

𝑁𝑖𝑂(𝑠) + 𝐶(𝑠) → 𝑁𝑖(𝑠) + 𝐶𝑂(𝑔)
metallic nickel at a 1:1 molar ratio with carbon

Goethite dehydroxylates to hematite which then is reduced

2𝐹𝑒𝑂𝑂𝐻(𝑠) → 𝐹𝑒2 𝑂3 + 𝐻2 𝑂(𝑔)
(𝑠)
to magnetite and subsequently wustite. Overall reduction
3𝐹𝑒2 𝑂3(𝑠) + 𝐶(𝑠) → 2𝐹𝑒3 𝑂4(𝑠) + 𝐶𝑂(𝑔) reactions shows a molar ratio of 2:1 for Fe:C
𝐹𝑒3 𝑂4(𝑠) + 𝐶(𝑠) → 3𝐹𝑒𝑂(𝑠) + 𝐶𝑂(𝑔)

2 𝐹𝑒𝑂𝑂𝐻(𝑠) + 𝐶(𝑠) → 2𝐹𝑒𝑂(𝑠) + 𝐶𝑂(𝑔) + 𝐻2 𝑂(𝑔)

Assuming 100 g sample of limonite ore with grade of 1.38% Ni and 45.70% Fe
Moles of Ni in ore 0.023512 Mol
(100g X 0.0138 / 58.6939 g/mol)
Moles of Fe in ore 0.818336 Mol
(100g X 0.457 / 55.845 g/mol)

Moles of C required for Ni reduction 0.023512 Mol

(Molar ratio of C:Ni =1)
Moles of C required for Fe reduction 0.409168 Mol
(Molar ratio of C:Fe =0.5)
Total Moles of C required for reduction 0.43268 Mol

Mass of Carbon required 5.196487 g

(0.43268 mol X 12.01 g/mol)
Mass of Coal required based on %Carbon value in Coal 7.678025 g
(5.196487 g / 0.677)
Estimated wt% Coal Addition 7.13 %

109
 Appendix C – Raw Data Tables

Reduction Studies

Design Design Design Total Pellet

Experiment Limonite Limonite Coal Coal Coal Sulphur Sulphur Sulphur Mass Mass Temperature
# (g) (g) (wt%) (g) (g) (wt%) (g) (g) (g) (g) (oC)
TF-1 2.7600 2.7640 6.00% 0.1800 0.1850 2% 0.06 0.0723 3.02 2.4037 1000
TF-2 2.7000 2.7023 6.00% 0.1800 0.2150 4% 0.12 0.1204 3.04 3.0315 1000
TF-3 2.7600 2.7610 6.00% 0.1800 0.1840 2% 0.06 0.0606 3.01 2.9925 1100
TF-4 2.7000 2.7024 6.00% 0.1800 0.1810 4% 0.12 0.1208 3.00 2.9813 1100
TF-5 2.7600 2.7633 6.00% 0.1800 0.1835 2% 0.06 0.0607 3.0075 3.005 1000
TF-6 2.7000 2.7044 6.00% 0.1800 0.1805 4% 0.12 0.1214 3.0063 3.0053 1000
TF-7 2.7600 2.7640 6.00% 0.1800 0.1816 2% 0.06 0.0608 3.0064 2.9933 1100
TF-8 2.7000 2.7027 6.00% 0.1800 0.1826 4% 0.12 0.1212 3.0065 2.9957 1100

Mass of
Non Mass of Mass of Mass of
Mags Mags filter + filter + Mass Mass
Pellet Mass of Filter Filter Non Mag Mag of Non of
Experiment Mass Grind Paper Paper Sample Sample Mags Mags
# (g) (g) (g) (g) (g) (g) (g) (g)
TF-1 2.4037 2.0500 2.62 2.54 3.0300 4.1 0.41 1.56
TF-2 3.0315 2.5200 2.62 2.58 2.8300 4.79 0.21 2.21
TF-3 2.9925 2.4500 2.62 2.62 2.7600 4.78 0.14 2.16
TF-4 2.9813 2.3400 2.63 2.56 2.8200 4.59 0.19 2.03
TF-5 3.005 2.3400 2.63 2.66 2.8000 4.76 0.17 2.10
TF-6 3.0053 2.1900 2.6 2.66 2.8600 4.53 0.26 1.87
TF-7 2.9933 2.1600 N/A 2.64 N/A 4.47 N/A 1.83
TF-8 2.9957 2.0800 2.59 2.65 2.9200 4.34 0.33 1.69

110
 Magnetic Fractions

Ni Fe
Mass Reading Fe Reading Ni Grade Ni Grade Fe
Experiment # (mg) (mg/L) (mg/L) (wt%) Recovery (wt%) Recovery
TF-1 311 16.7113 4.1604 1.34% 54.94% 83.61% 103.258%
TF-2 310 16.7373 4.3609 1.35% 79.99% 87.92% 157.339%
TF-3 316.5 16.8845 4.3677 1.33% 75.61% 86.25% 147.649%
TF-4 318.5 18.1306 4.5411 1.42% 77.47% 89.11% 146.474%
TF-5 308.6 15.5269 3.9014 1.26% 69.27% 79.01% 131.395%
TF-6 302 15.7471 4.2118 1.30% 65.32% 87.16% 131.885%
TF-7 299.7 15.7013 4.3005 1.31% 62.84% 89.68% 129.930%
TF-8 307.5 17.1404 4.3416 1.39% 63.14% 88.24% 120.742%

Non Magnetic Fraction

Ni Fe Ni
Experiment Reading Reading Grade Ni Fe Grade Fe
# Mass (mg) (mg/L) (mg/L) (wt%) Recovery (wt%) Recovery
TF-1 303.2 6.8852 3.2481 0.57% 6.1023% 66.95% 21.733%
TF-2 117.6 0.6425 1.0923 0.14% 0.7692% 58.05% 9.872%
TF-3 35.2 0.2675 0.2689 0.19% 0.6981% 47.75% 5.298%
TF-4 77.2 0.1762 0.8478 0.06% 0.2907% 68.64% 10.560%
TF-5 105.9 1.8844 1.0964 0.44% 1.9832% 64.71% 8.711%
TF-6 84.4 0.1957 1.2867 0.06% 0.4038% 95.28% 20.045%
TF-7 N/A N/A N/A N/A N/A N/A N/A
TF-8 81.4 N/A 1.2917 N/A N/A 99.18% 26.498%

111
 Degree of Metallization Studies

Design Coal Design Design Total Pellet Growth Zone

Experiment Limonite Limonite Coal Coal Sulphur Sulphur Sulphur Mass Mass Temperature
# (g) (g) (wt%) (g) (g) (wt%) (g) (g) (g) (g) (oC)
TF-17 2.7600 2.7691 6.00% 0.1800 0.1817 2% 0.06 0.0623 3.0131 2.9939 1000
TF-18 2.7000 2.7000 6.00% 0.1800 0.1820 4% 0.12 0.1223 3.0043 2.9856 1000
TF-19 2.7600 2.7665 6.00% 0.1800 0.1838 2% 0.06 0.067 3.0173 2.994 1100
TF-20 2.7000 2.7053 6.00% 0.1800 0.1816 4% 0.12 0.1273 3.0142 2.9905 1100

Design Coal Design Design Total Pellet Growth Zone

Experiment Limonite Limonite Coal Coal Sulphur Sulphur Sulphur Mass Mass Temperature
# (g) (g) (wt%) (g) (g) (wt%) (g) (g) (g) (g) (oC)
TF-17 2.7600 2.7691 6.00% 0.1800 0.1817 2% 0.06 0.0623 3.0131 2.9939 1000
TF-18 2.7000 2.7000 6.00% 0.1800 0.1820 4% 0.12 0.1223 3.0043 2.9856 1000
TF-19 2.7600 2.7665 6.00% 0.1800 0.1838 2% 0.06 0.067 3.0173 2.994 1100
TF-20 2.7000 2.7053 6.00% 0.1800 0.1816 4% 0.12 0.1273 3.0142 2.9905 1100

Process Replication Studies – Tube Furnace Setup

Design Coal Design Design Total Pellet Growth Zone

Experiment Limonite Limonite Coal Coal Sulphur Sulphur Sulphur Mass Mass Temperature
# (g) (g) (wt%) (g) (g) (wt%) (g) (g) (g) (g) (oC)
STA-1 2.7600 2.7643 6.00% 0.1800 0.1827 2% 0.06 0.0628 3.0098 2.9938 1000
STA-2 2.7000 2.7004 6.00% 0.1800 0.1855 4% 0.12 0.1229 3.0088 3.0017 1000
STA-3 2.7600 2.7604 6.00% 0.1800 0.1836 2% 0.06 0.0675 3.0115 2.9886 1000
STA-4 2.7000 2.7004 6.00% 0.1800 0.1815 4% 0.12 0.1262 3.0081 2.994 1000

112
 Non - Mass Loss
Magnetic In
Mass Mass of Magnetic Fraction Separation wt%
Experiment # of Pellet (g) Grind (g) Fraction (g) (g) (g) Magnetic
STA-1 2.13 2.19 0.73 1.23 0.23 33.33%
STA-2 2.1717 2.11 0.54 1.37 0.2 25.59%
STA-3 2.3176 2.23 0.19 1.9 0.14 8.52%
STA-4 2.1572 2.03 0.27 1.67 0.09 13.30%

Ni Reading Fe Reading Ni Grade Ni Fe Grade

Experiment # Mass (mg) (mg/L) (mg/L) (wt%) Recovery (wt%)
STA-1
Magnetic 301.9 1.1514 9.5923 0.95% 18.2458% 79.43%
Non-magnetic 318.2 1.18147 9.7619 0.93% 76.70%
STA-2
Magnetic 310.5 1.9183 9.2996 1.54% 22.3811% 74.88%
Non-magnetic 302.1 1.0436 11.555 0.86% 95.62%
STA-3
Magnetic 93.9 4.73 2.7339 1.26% 6.2811% 72.79%
Non-magnetic 314.2 16.0356 9.8198 1.28% 78.13%
STA-4
Magnetic 225 28.6898 7.2159 3.19% 23.0963% 80.18%
Non-magnetic 301.2 11.1915 9.2426 0.93% 76.71%

113
 Process Replication Studies – Tube Furnace Setup with larger sample and using HIWMS

Design Coal Design Design Total Pellet Growth Zone

Experiment Limonite Limonite Coal Coal Sulphur Sulphur Sulphur Mass Mass Temperature
# (g) (g) (wt%) (g) (g) (wt%) (g) (g) (g) (g) (oC)
TF-25 13.5000 13.5000 6.00% 0.9000 0.9259 4% 0.6 0.6015 15.0274 15.0141 1000
TF-26 13.5000 13.5000 6.00% 0.9000 0.9102 4% 0.6 0.6047 15.0149 14.9899 1000
TF-27 13.5000 13.5400 6.00% 0.9000 0.9015 4% 0.6 0.605 15.0465 15.0323 1000
TF-28 9.4000 9.4100 6.00% 0.6000 0.6058 0% 0 0 10.0158 10.0097 1000
TF-29 9.3000 9.3000 6.00% 0.6000 0.6044 1% 0.1 0.1008 10.0052 9.9759 1000
TF-30 9.2000 9.2000 6.00% 0.6000 0.6060 2% 0.2 0.2067 10.0127 10.018 1000
TF-31 9.1000 9.1000 6.00% 0.6000 0.6056 3% 0.3 0.3003 10.0059 9.9472 1000
TF-32 9.0000 9.0000 6.00% 0.6000 0.6085 4% 0.4 0.4054 10.0139 9.9907 1000

Experiment Mass Mass Magnetic Non - Mass Loss In wt% Ni Recovery Fe/Ni wt
# of of Fraction Magnetic Separation Magnetic Grade (%) ratio
Pellets Grind (g) Fraction (g) (wt%)
(g) (g) (g)
TF-25 10.6 9.66 2.94 6.13 1.53 27.74% 3.79% 83.66% 29.1981
TF-26 10.64 9.66 3.06 6.67 0.91 28.76% 4.06% 93.19% 25.0057
TF-27 10.6 9.66 3.29 7.05 0.26 31.04% 3.77% 93.06% 27.2772
TF-28 7.76 7.52 1.15 8.55 -1.94 14.82% 1.42% 15.71% 56.9001
TF-29 7.71 7.47 1.07 7.33 -0.69 13.88% 1.33% 13.77% 52.1165
TF-30 7.44 7.07 1.54 5.9 0 20.70% 1.55% 24.53% 53.4015
TF-31 7.22 6.82 2.77 4.42 0.03 38.37% 1.56% 46.00% 50.9229
TF-32 6.99 6.62 3.53 3.6 -0.14 50.50% 1.96% 75.76% 46.4406

114
TF-25

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Pass 1 2.23 300.7 48.9043 1361.096 4.07% 113.16% 0.090669 0.09066893
Pass 2 +3 0.71 314.8 36.9887 1299.698 2.94% 103.22% 0.020856 0.111525011
Non Mags 6.13 315.9 3.4864 1028.706 0.28% 81.41% 0.016913

TF-26

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Pass 1 2.13 310.7 53.8489 1241.964 4.33% 99.93% 0.09229 0.092290117
Pass 2 +3 0.93 302.1 41.4968 1270.524 3.43% 105.14% 0.031936 0.124226581
Non Mags 6.67 331.7 3.166 1153.937 0.24% 86.97% 0.015916

TF-27

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Pass 1 2.27 314.9 51.4243 1282.849 4.08% 101.85% 0.092675 0.092674786
Pass 2 +3 1.02 320.5 39.4395 1347.293 3.08% 105.09% 0.031379 0.124054108
Non Mags 7.05 306 1.1522 1158.097 0.09% 94.62% 0.006636

TF-28

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Mags 1.15 336.2 19.0657 1084.841 1.42% 80.67% 0.016304 0.016303952
Non mags 8.55 322.8 15.8717 941.6821 1.23% 72.93% 0.105098 0.121402337
TF-29

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Mags 1.07 318.7 16.9128 881.4354 1.33% 69.14% 0.014196 0.014195714
Non mags 7.33 304.8 13.8717 1009.179 1.14% 82.77% 0.083399 0.097594304

115
 TF-30

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Mags 1.54 326.4 20.2899 1083.511 1.55% 82.99% 0.023933 0.023932633
Non mags 5.9 312.2 13.5045 1048.712 1.08% 83.98% 0.063802 0.087735124
TF-31

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Mags 2.77 318.3 19.8984 1013.284 1.56% 79.59% 0.043291 0.043291367
Non mags 4.42 311.3 7.0235 954.157 0.56% 76.63% 0.024931 0.068222197
TF-32

Mass Digested Ni Fe Fe
Fraction Mass Reading Reading Ni Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Mags 3.53 317.1 24.8688 1154.921 1.96% 91.05% 0.069211 0.06921071
Non mags 3.6 308.4 0.4688 873.9908 0.04% 70.85% 0.001368 0.070578803

Process Replication Studies – Rotary Kiln Setup

Estimated
Design Coal Design Design Total Total Growth Zone
Experiment Limonite Limonite Coal Coal Sulphur Sulphur Sulphur Mass Mass Temperature
# (g) (g) (wt%) (g) (g) (wt%) (g) (g) (g) (g) (oC)
RK-1 270 270.12 6 18 18.01 4 12 12.07 300.2 300.13 1000
RK-2 270 270.11 6 18 18.01 4 12 12.04 300.16 300.07 1000

Estimated
Design Coal Design Design Total Total Growth Zone
Experiment Limonite Limonite Coal Coal Sulphur Sulphur Sulphur Mass Mass Temperature
# (g) (g) (wt%) (g) (g) (wt%) (g) (g) (g) (g) (oC)
RK-1 270 270.12 6 18 18.01 4 12 12.07 300.2 300.13 1000
RK-2 270 270.11 6 18 18.01 4 12 12.04 300.16 300.07 1000

116
Run RK-1-1

Mass of
Field Magnetic Digested Ni Fe Ni Fe
Strength Fraction Mass Reading Reading Grade Grade Total Ni Cumulative
(amps) (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
0.2 0.24 140 25.0528 474.7279 4.47% 84.77% 0.010737 0.0107369
0.6 2.89 316 43.2086 950.114 3.42% 75.17% 0.098792 0.1095287
1 4.05 312.6 40.1234 936.7421 3.21% 74.92% 0.129958 0.239487
1.5 5.05 324.6 32.765 1068.656 2.52% 82.31% 0.127436 0.3669232
2 4.94 302.8 17.0864 948.2281 1.41% 78.29% 0.069689 0.4366118
2.5 3.9 325.8 13.7835 1089.679 1.06% 83.62% 0.041249 0.4778608
4 2.8 300.1 10.2745 1059.105 0.86% 88.23% 0.023966 0.5018266
0.501827

RK-1-3 at 1 amp

Mass Digested Ni Fe Ni Fe
Fraction Mass Reading Reading Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Head Grade 2.34 317.9 19.9972 999.79 1.57% 78.62% 0.036799
Pass1 2.2 337.4 62.0735 1165.79 4.60% 86.38% 0.101187 0.1011868
Pass 2 & 3 1 344.4 38.725 1089.99 2.81% 79.12% 0.02811 0.1292973
Non Mags 10.46 348 5.3094 1069.1 0.38% 76.80% 0.039897 0.1691941

RK-1-5 at 1 amp

Mass Digested Ni Fe Ni Fe
Fraction Mass Reading Reading Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Pass 1 2.83 307 52.901 1044.507 4.31% 85.06% 0.121914 0.1219135
Pass 2 0.77 306.5 43.3431 1123.186 3.54% 91.61% 0.027222 0.1491356
Pass 3 0.39 227.2 23.0169 704.2278 2.53% 77.49% 0.009877 0.159013
Non Mags 10.79 311.5 2.8844 1016.005 0.23% 81.54% 0.024978 0.183991

117
RK-1-4 at 1 amp

Mass Digested Ni Fe Ni Fe
Fraction Mass Reading Reading Grade Grade Total Ni Cumulative
Sample (g) (mg) (mg/L) (mg/L) (wt%) (wt%) (g) Ni (g)
Head Grade 2.17 323.5 18.1332 1084.149 1.40% 83.78% 0.030409
Pass1 2.34 307 47.287 957.205 3.85% 77.95% 0.090107 0.0901071
Pass 2 & 3 1.04 343.1 33.9607 1102.614 2.47% 80.34% 0.025735 0.1158425
Non Mags 9.93 329.2 2.0444 1027.978 0.16% 78.07% 0.015417 0.1312593

118
 Appendix D - TGA/DTA Curves for Coal and Sulphur Additives

100 1.80E+00
90 1.60E+00

80 1.40E+00
1.20E+00
70
Mass Loss (%)

DTA (μV/mg)
1.00E+00
60
8.00E-01
50
6.00E-01
40
4.00E-01
30
2.00E-01
20 0.00E+00
10 -2.00E-01
0 -4.00E-01
0 200 400 600 800 1000
Temperature (oC)

Mass Loss DTA

Figure A 1: TGA/DTA curves for sulphur up to 1000oC under N2 atmosphere

119
 100 1.60E+00

1.40E+00
80 1.20E+00

1.00E+00
Mass loss (%)

DTA (μV/mg)
60
8.00E-01

6.00E-01
40
4.00E-01

20 2.00E-01

0.00E+00

0 -2.00E-01
0 200 400 600 800 1000
Temperature (oC)

Mass Loss DTA

Figure A 2: TGA/DTA curves for coal up to 1000oC under N2 atmosphere

120