close

Вход

Забыли?

вход по аккаунту

?

Microwave assisted communition of sulfide ore

код для вставкиСкачать
MICROWAVE ASSISTED COMMUNITION OF SULFIDE ORE
By
Matthew D. Andriese
A THESIS
Submitted in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
In Materials Science and Engineering
MICHIGAN TECHNOLOGICAL UNIVERSITY
2015
© 2015 Matthew D. Andriese
UMI Number: 1590453
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMI 1590453
Published by ProQuest LLC (2015). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
This thesis has been approved in partial fulfillment of the requirements for the Degree of
MASTER OF SCIENCE in Materials Science and Engineering.
Department of Materials Science and Engineering
Thesis Advisor: Jiann-Yang Hwang
Committee Member: Xiaodi Huang
Committee Member: Timothy C. Eisele
Department Chair: Mark R. Plichta
To k.m.h. & the ski house crew
Table of Contents
Table of Contents ............................................................................................................... iv
List of Figures ................................................................................................................... vii
List of Tables ..................................................................................................................... xi
Abstract ............................................................................................................................. xii
Chapter 1: Introduction ...................................................................................................... 1
1.1
Material Bonding and Structure ........................................................................... 2
1.2 Mineral Processing.................................................................................................... 3
1.3 Communition Theory ................................................................................................ 4
1.4 Mechanical Breakage of Rock .................................................................................. 5
1.5 Petrology of the Ore Body ........................................................................................ 7
1.6 Silicate Minerals ..................................................................................................... 10
1.6.1 Peridotite Rock................................................................................................. 10
1.6.2 Olivine.............................................................................................................. 11
1.6.3 Pyroxene .......................................................................................................... 12
1.6.4 Incongruent Melting of Pyroxene .................................................................... 13
1.7 Sulfide Minerals ...................................................................................................... 14
1.7.1 Pyrrhotite.......................................................................................................... 17
1.7.2 Chalcopyrite ..................................................................................................... 21
1.7.3 Pentlandite........................................................................................................ 22
1.7.4 Metal-Sulfide Phase Equilibria ........................................................................ 23
1.8 Spinel Minerals ....................................................................................................... 26
1.9 Platinum Group Metals ........................................................................................... 29
1.10 Microwave Processing of Materials...................................................................... 30
1.10.1 Introduction .................................................................................................... 30
1.10.2 Microwave-Material Physics ......................................................................... 33
1.10.2 Microwave-Material Heating ......................................................................... 39
1.10.3 Microwave-Mineral Processing ..................................................................... 47
Chapter 2: Experimental Procedure ................................................................................. 57
2.1 Ore Particle Characterization .................................................................................. 57
2.1.1 Imaging of Samples ......................................................................................... 57
iv
2.1.2 Magnetic Susceptibility Measurements ........................................................... 57
2.2 Crushing and Grinding Experiments ...................................................................... 57
2.2.1 Obtaining Ore Particle Samples for Grinding Experiments ............................ 57
2.2.2 Microwave Heating of Ore Particles................................................................ 58
2.2.3 Preliminary Grinding Experiments of Ore Particles ........................................ 59
2.2.4 Ball Mill Grindability Experiments of Ore Particles ....................................... 59
2.2.5 Roll Crushing Experiments .............................................................................. 60
2.3 Heavy Liquid Separation ........................................................................................ 61
2.4 X-Ray Diffraction (XRD) ....................................................................................... 62
2.5 Inductively Coupled Plasma (ICP) ......................................................................... 62
Chapter 3: Ore Particle Characterization ......................................................................... 63
3.1 Imaging of Ore Samples ......................................................................................... 63
3.1.1 Mineral Phase Assemblages ............................................................................ 63
3.1.2 Silicate Minerals .............................................................................................. 64
3.1.3 Metallic Minerals ............................................................................................. 64
3.2 Magnetism in Ore Particles ..................................................................................... 69
3.2.1 Curie Temperature Measurement .................................................................... 69
3.2.2 Magnetism in Nickel Pyrrhotite ....................................................................... 70
3.3 Microwave Exposure of Ore Particles .................................................................... 72
3.3.1 Observations During Microwave Exposure ..................................................... 72
3.3.2 Observations After Microwave Exposure ........................................................ 72
3.3.3 Cracking in Ore Particles ................................................................................. 75
3.3.4 Microwave Induced Arcing ............................................................................. 80
Chapter 4: Crushing and Grinding Experiments ............................................................... 91
4.1 Experimental Background ...................................................................................... 91
4.2 Preliminary Ball Milling Experiments .................................................................... 93
4.2.1 Effect of Increasing Ball Mill Revolutions ...................................................... 93
4.2.2 Effect of Microwave Treatment Time on Particle Size Reduction .................. 94
4.3 Ball Mill Grindability Experiments ........................................................................ 98
4.3.1 Ball Mill Grindability of Jaw Crushed Material .............................................. 98
4.3.2 Ball Mill Grindability of Gyratory Crushed Material ...................................... 99
4.3.3 Sample Weight Loss ...................................................................................... 100
4.3.4 Particle Size Distribution of Ball Milled (±100) Material ............................. 101
4.3.5 XRD and ICP of Ball Milled (-100) sink material ......................................... 103
v
4.4 Roll Crusher Experiments ..................................................................................... 106
4.4.1 Particle Size Distribution Roll Crushed Particles (-4+6) ............................... 106
4.4.2 XRD and ICP of Roll Crushed Particles (-65+100) Sink Material ................ 108
4.4.3 XRD and ICP of Roll Crushed Particles (-100+200) Sink Material .............. 110
4.4.4 XRD and ICP of Roll Crushed Particles (-200+325) Sink Material .............. 113
Chapter 5: Conclusion.................................................................................................... 116
References ....................................................................................................................... 117
Appendix 1: Images ........................................................................................................ 131
Appendix 2: EDS of Silicate Minerals............................................................................ 143
Appendix 3: EDS of Metallic Minerals .......................................................................... 148
Appendix 4: Ball Milling Procedure ............................................................................... 156
Appendix 5: XRD Raw Data .......................................................................................... 162
vi
List of Figures
Figure 1.1 Combination of bonding types in various minerals, metals and elements. ....... 2
Figure 1.2 Map of Michigan showing the location of Kennecott’s Eagle Mine................. 8
Figure 1.3 Illustration of low-temperature phase relations in the Fe-S system suggesting a
composition gradient existing among superstructures formed between end-members due
to a strong driving force in the formation of magnetic 4M-type Po (Fe7S8)..................... 20
Figure 1.4 High-temperature fractionation crystallization of metal-sulfide liquid and the
resultant formation of low-temperature mineral assemblages .......................................... 26
Figure 1.5 The sum of magnetic Fe moments seated in octahedral and tetrahedral
positions of the inverse spinel mineral magnetite (Fe3O4)................................................ 28
Figure 1.6 The dielectric properties, permittivity () and permeability (µ), of matter. .... 34
Figure 1.7 Random alignments of magnetic dipoles upon reaching Curie temperature
(Tc). ................................................................................................................................... 42
Figure 2.1 Experimental parameters that affect microwave absorption in samples. ........ 59
Figure 3.1 BSE image of ore particles (Jaw -8+12 mesh). Mineral phases are pyroxene
(Px), pentlandite (Pnt), pyrrhotite (Po), chalcopyrite (Ccp) and ferrospinel (Mag). ........ 64
Figure 3.2 BSE image of globular ferrospinel (Mag) inclusions in chalcopyrite (Ccp). .. 65
Figure 3.3 BSE image of cracking in a Ti-ferrospinel (Mag) grain. Two pyrrhotite (Po)
grains with different compostions of are located in the image by arrows. ....................... 66
Figure 3.4 BSE image of pentlandite (Pnt) and Ni-rich pyrrhotite (Ni-Po) grains exsolved
from pyrrhotite (Po). ......................................................................................................... 67
Figure 3.5 BSE image of galena (Gn) inclusions in pyrrhotite (Po). ................................ 68
Figure 3.6 BSE image of sphalerite (Sl) inclusions near a chalcopyrite (Ccp) grain. ..... 68
Figure 3.7 BSE image of michenerite (Mich) grains dissemenated in pyrrhotite (Po)..... 69
Figure 3.8 Curie temperature (TC) plots of metallic ore samples with magnetic moment in
electromagnetic units (0.001A∙m²) plotted as a function of temperature (ºC). ................. 70
vii
Figure 3.9 XRD pattern of the Curie temperature sample plotted as intensity (arbitrary
units) vs. 2θ (degrees). Labeled peaks are metallic-sulfides pyrrhotite (Po), chalcopyrite
(Ccp), and pentlandite (Pnt). ............................................................................................. 71
Figure 3.10 Electrical arcing (plasma) during MW exposure of sulfide ore particles. ..... 72
Figure 3.11 Pieces of pure chalcopyrite (Ccp) in the MW oven cavity after bursting apart
during irradiation. ............................................................................................................. 73
Figure 3.12 BSE image of a particle surface heavily oxidized after microwave exposure
in open atmosphere. .......................................................................................................... 74
Figure 3.13 Optical and BSE image of melt-flux formed during 120s of microwave
exposure in open atmosphere. ........................................................................................... 74
Figure 3.14 Optical images of sulfide mineral grains before (a) and after (b) microwave
exposure (2.45GHz, 1000W) in air for 10s. Scale bar not shown. ................................... 76
Figure 3.15 BSE image of cracking in silicate host rock and metallic sulfide minerals.
The circled portion of (a) can be viewed at higher magnification in (b). ......................... 77
Figure 3.16 BSE image of cracking in a chromite grain (Mag) enclosed by pyrrhotite
(Po). ................................................................................................................................... 78
Figure 3.17 BSE image of cracking between metallic bearing minerals ulvöspinel (Mag),
pyrrhotite (Po), pentlandite (Pnt), and chalcopyrite (Ccp). .............................................. 79
Figure 3.18 BSE image of cracking along a pyrrhotite (Po)-pyroxene (Px) grain
boundary. .......................................................................................................................... 79
Figure 3.19 Oxidation and arcing of pure chalcopyrite (Ccp) by MW energy. ................ 80
Figure 3.20 BSE image of a heavily oxidized particle surface during MW exposure...... 81
Figure 3.21 EDS of the iron oxide scale (Fe 94 O 6) at. %. ............................................. 81
Figure 3.22 BSE image of melt-flux produced during electric arcing on metallic minerals
........................................................................................................................................... 82
Figure 3.23 EDS of metallic veins indicating partial dissolution of sulfide into silicate . 83
Figure 3.24 BSE olivine crystals nucleated in the melt-flux material .............................. 84
Figure 3.25 Illustration showing electrochemical diffusion Fe into melt-flux solution ... 84
Figure 3.26 XRD pattern of the melt-flux solution (-150 mesh). ..................................... 85
viii
Figure 3.27 EDS spectra of the olivine (Ol) crystal interior. ............................................ 86
Figure 3.28 EDS spectra of the olivine (Ol) crystal edge. ................................................ 86
Figure 3.29 EDS spectra of clinopyroxene containing iron sulfide (Cpx+Po). ................ 87
Figure 3.30 EDS spectra of metal-sulfide liquid (sul-mt). ................................................ 87
Figure 3.31 BSE image of a metal-sulfide particle surronded by silicate melt (si-mt).
Close observation shows ferrospinel (Mag) exsolved in areas of monosulfide solution
(mss). ................................................................................................................................. 88
Figure 3.32 BSE image of metal-sulfide droplets dispersed in the melt-flux solution. .... 89
Figure 3.33 SEI image of gold (Au) flecks formed on a metal-sulfide particle. .............. 90
Figure 3.34 SEI image of metal-sulfide particles formed on the surface of a sample. ..... 90
Figure 4.1 Particle size distribution for microwave 60s and as-received (-6+8) material
ball milled for 100, 300, 500 and 1000 revolutions. ......................................................... 94
Figure 4.2 Particle size distribution for (-6+8) material ball milled 100 revolutions. ...... 96
Figure 4.3 Particle size distribution for (-6+8) material ball milled 300 revolutions. ...... 96
Figure 4.4 Particle size distribution for (-6+8) material ball milled 500 revolutions. ..... 97
Figure 4.5 Cumulative percent passing of oversize material (+100 mesh) from ball mill
grindability experiments using gyratory (-6+8) particles. .............................................. 102
Figure 4.6 Cumulative percent passing of undersize material (-100 mesh) from ball mill
grindability experiments using gyratory (-6+8) particles. .............................................. 103
Figure 4.7 XRD plots of intensity (arbitrary units) vs. 2θ (degrees) of ball milled
undersize (-100 mesh) sink material. .............................................................................. 104
Figure 4.8 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s
and as-received ball milled (-100 mesh) sink material. .................................................. 105
Figure 4.9 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s and
as-received ball milled (-100 mesh) sink material. ......................................................... 105
Figure 4.10 Particle size distribution for microwave and as-received gyratory crushed ore
particles (-4+6) passed through the roll crusher. ............................................................ 107
Figure 4.11 XRD pattern of MW 60s and as-received rolled (-65+100) sink material. . 108
ix
Figure 4.12 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s
and as-received rolled (-65 +100) sink material. ............................................................ 109
Figure 4.13 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s
and as-received rolled (-65 +100) sink material. ............................................................ 110
Figure 4.14 XRD pattern of the MW 60s and as-received rolled (-100+200) sink material.
......................................................................................................................................... 111
Figure 4.15 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s
and as-received rolled (-100+200) sink material. ........................................................... 112
Figure 4.16 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s
and as-received rolled (-100+200) sink material. ........................................................... 113
Figure 4.17 XRD pattern of MW 60s and as-received rolled (-200+325) sink material. 114
Figure 4.18 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s
and as-received rolled (-200+325) sink material. ........................................................... 115
Figure 4.19 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s
and as-received rolled (-200+325) sink material. ........................................................... 115
x
List of Tables
Table 1.1 Ideal site occupancy of metal cations in the M1, M2, T sites of pyroxene. ....... 12
Table 1.2 Coordination number, orbital hybridization, and electron spin state for Fe, Cu,
and Ni transition metal cations in sulfides. ....................................................................... 16
Table 2.1 Mineral compositions and specific gravities of common minerals in ore. ....... 61
Table 4.1 Ball mill grindability (GBP) and work index (wi) results for microwave and asreceived jaw crushed (-6+10) and (-8+12) material. ........................................................ 98
Table 4.2 Ball mill grindability (GBP) and work index (wi) results for microwave and asreceived gyratory crushed (-6+8) and (-8+12) material. ................................................. 100
Table 4.3 The average mass of particle samples before and after microwave heating. .. 101
xi
Abstract
The goal of these experiments was to induce cracking in ore particles using MW energy
for improved communition behavior of ore material by thermal expansion of MW
absorbing phases to increase metallic mineral liberation. The material used for experiments
was from a metallic sulfide ore deposit located in Michigan’s Upper Peninsula composed
of peridotite (olivine + pyroxene) rock with Fe-Cu-Ni sulfide minerals pyrrhotite,
chalcopyrite, and pentlandite containing minor concentrations of precious metals (PGM’s)
and inclusions of Fe-Cr-Ti spinel oxide minerals disseminated throughout the
mineralization. Rapid heating in ore particles was mainly attributed to the high absorption
capabilities of magnetic phase’s pyrrhotite, Ni-pyrrhotite, and Fe-Cr-Ti oxide minerals.
MW exposure of ore particle caused differential thermal expansion of constituent phases
producing macro and microscopic cracks in ore particles. The production of sulfur gas and
subsequent weight loss in samples was also reported. Imaging of samples using scanning
electron microscope (SEM) in back scatter electron mode (BSE) shows fracturing
originating at metallic mineral phases and continuing throughout the host rock matrix. Ball
milling experiments showed increased grindability of MW treated ore particles and an
overall decrease in the work index compared with as-received material. MW treated ore
particles passed through a roll crusher showed an increase in metallic bearing minerals
reporting to coarser size fractions then as-received material.
In the presence of MW field, it was observed that sulfide ore particles exposed to open
atmosphere can electrically arc during sulfide oxidation generating plasma. The plasma
reaction is highly exothermic creating high local temperatures that melts constituent
phases. The formation of a flux-type solution comprised of metal sulfides droplets
dispersed in silicate melt was observed. Multiple flecks of gold were observed to be
nucleated on metal sulfide droplets as indicated by electron dispersive spectroscopy (EDS)
during SEM beam probing.
xii
Chapter 1: Introduction
Communition, or size reduction, of ore material is accomplished by crushing coarse-size
rock and grinding it into finer-size particles for liberation of economically valuable
minerals. This the most energy intensive step of mineral beneficiation, so finding methods
in reducing the amount of energy input can potentially result in a large reduction in overall
costs. Unfortunately from a technical standpoint, the mineral processing industry is largely
archaic relying on “tried and tested” methods that leaves little-to-no room for new
technologies. Recently, microwaves (MW) have been employed by researchers in the
mineral processing industry for heating of various materials with an overall goal of
improving metallic yield of the final product.
Sulfide minerals are the main source for production of non-ferrous metals. Generally
referred as chalcogenides, these mineral compounds exhibit an array of electrical and
magnetic properties that can be utilized during mineral separation processes. The unique
dielectric properties also enable sulfides to strongly absorb MW energy. Heat produced by
microwave energy (irradiation) can result in physical and chemical alterations of phases.
The thermal stresses generated within particles and subsequent cracking can be used to
assist with communition experiments. Ideally, strong absorbers of MW energy (metallic
minerals) will be liberated from host rock (gangue) due to differences in their thermal
expansion properties. Potentially, MW induced cracking and embrittlement of sulfide ore
particles can be used to aid with communition efforts in liberating metallic minerals.
Due to the large scale operations of the mining industry, new extractive metallurgy
techniques have been precluded by the longstanding history of the mining industry. Even
proper interpretation of an ore mineralization would allow for better extraction techniques
to be used on metal-rich zones of a deposit. Though, reduction of energy is always the main
objective for most industrial applications, environmental factors and sustainable processes
are becoming increasingly important considerations for future mining technologies [1], [2].
1
1.1 Material Bonding and Structure
Inorganic matter atomically bonds as one of three types: ionic, covalent or metallic.
However, most mineral compounds are a combination of bonding types and exhibit
variations in the degree of bonding types between atoms distributed in the crystal structure.
In matter, this is known as real bonding which refers to each bond type’s partial
contribution to the chemical structure (Figure 1.1). Thus, the type and degree of bonding
in a crystal structure dictates the physical and chemical properties of a material.
Figure 1.1 Combination of bonding types in various minerals, metals and elements.
All mineral compounds can be classified as ceramic materials [3]. Ceramics are formed
with heat and pressure by combining a: (1) metal-nonmetal, (2) metal-metalloid (3)
metalloid-nonmetal, and (4) nonmetal-nonmetal. As a product of their chemical
2
environment, natural minerals tend to employ a high amount of atomic substitution that
allows for a multitude of compositions, crystal structures and defect concentrations.
Ordering is a free energy term used to describe the degree of randomness in a
thermodynamic system. With minerals, varying compositions between end members of a
solution, including impurity atoms, creates short and long range ordering of atoms that
compose the structure. With the likeliness that atoms will only occupy certain atomic sites,
the overall charge distribution of a crystal is specific to composition.
Differences in the electronegativity of any two elements that compose a mineral or
compound define its % ionic character. An ionic bond formed between two atoms having
a large difference in size between atomic radii creates a directional bond that is effectively
covalent because electrons are likely to reside in certain parts of the orbital as predicted by
the quantum energy distribution.
Many transition metal sulfides behave as a metal due to the strong covalent nature of the
sulfur structure having relatively weak metal-sulfur bonding and strong metal-metal
interactions. Ferro-spinel oxides form complex electronic and magnetic structures from
overlapping atomic orbitals (super-exchange) of adjacent metal atom giving these minerals
strong ionic character. Silicate minerals display variations between covalent and ionic
bonding types depending on Si-O chain formations and isomorphic chemistry.
1.2 Mineral Processing
Mineral processing is concerned with the economic treatment of ore material for the
recovery of valuable material (beneficiation). The two main objectives to accomplish this
are liberation and separation. Liberation means to free minerals from the host rock while
separation is the task of concentrating valuable material from waste material. Valuable
material is then treated with hydro/pyro-metallurgical methods for further refinement.
3
As the technical age depletes the world’s commodity metal resources by gross human
consumption, current mining involves the extraction of a large volume of rock that may
contain only a small percentage of valuable material. The mineable concentration of
valuable material designates the grade of an ore type. Thus, low grade ores have high
energy requirements to economically produce a concentrate for metal extraction [4].
The first step of mineral beneficiation is communition; the successive size reduction of ore
material. This involves a reduction in size until the sought after mineral(s) and host rock
matrix theoretically occur as two separate particles [5]. This is accomplished by stage
crushing coarse rocks to a size suitable for grinding material to a fine size. Most often
hardness and abrasion are the two physical characteristics taken into consideration during
installation of grinding equipment used in the communition circuit.
By far, the biggest energy expenditure during beneficiation is communition. The reason for
high energy input is that mechanical breakage of ore is an extremely energy intensive yet
inefficient process. Communition leads to 95% of energy wasted during beneficiation,
mainly due to the generation of heat during collisions, with the energy spent during
grinding to create new surfaces being less than 1% efficient [6]. It has been shown that
energy input increases disproportionately with decreasing particle sizes. Thus, grinding
leads to the greatest energy expenditure during communition so it to a minimum is best
depending on the desired product size. Also, any excess grinding of material leads to large
energy expenditures plus increased wear on equipment.
1.3 Communition Theory
Communition theory is used to describe relationship between the energy inputs needed to
produce a final size product from an initial feed size. Fred C. Bond (1961) developed the
most widely accepted theory by asserting that it is a rough measure of the power consumed
in reducing 80% of the feed size to pass 80% of a given product size [7]. From this theory,
Bond formulated the equation in units of kilowatt-hour per short ton (kW·h):
4
BWI
=
10Wi
10Wi
−
P
F
Where BWI the Bond work index, Wi is the work index of the tested material, P is the
diameter in microns in which 80% of product material passes and F is the diameter in
microns in which 80% of the feed material passes. The Bond equation is valid using either
a rod or ball mill but laboratory experiments are limited to only testing material passing a
6 mesh sieve (>3.327mm). BWI data for numerous ore and refractory materials can be
found in literature which are used to gauge a materials resistance to crushing and grinding
[8]. It should be noted that BWI is not an absolute quantitative value because the source of
initial material, laboratory equipment and experimental procedures will vary.
Many authors express dislike with the timely length to complete ball mill grindability
experiments. Attempts have been made to create a shorten version of ball milling
experiments for determining the work index using just two grinding cycles as opposed to
5-20 cycles usually needed to complete F. Bonds original procedure [9]. Single particle
drop weight testing has also been subject of study for determining the energy-size reduction
relationship used for estimating the work index of a material [10], [11]. From statistical
approach, this is an incorrect way of examining a population and should only be used when
materials are limited because time for completing experiments should always be available.
It is important that sample assays accurately represent the bulk ore material. The design,
operation, and success of a mineral processing operation relies on sampling locations and
correct evaluation of material. Since size reduction is such an inefficient process, it is
necessary to optimize operating conditions by understanding parameters that influence the
performance of communition equipment.
1.4 Mechanical Breakage of Rock
The configuration of atoms is determined by the type(s) of physical and chemical bonds
present that give a material its overall bulk mechanical properties, among many other
5
things. A minor stress applied to the material will lead only to elastic deformation, which
is small, reversible inter-atomic bond displacement. The bonding between atoms is broken
when the applied stress exceeds a critical value during tensile or compressive loading. In
crystalline materials, this critical value for tensile strength can be as little as 1/10 of its
compressive strength.
For any system under an applied stress, the force per unit area is equivalent to the energy
per unit volume. When a material is subjected to loading conditions, stress by compression
raises the chemical potential and stress by tension lowers it. When bonds are broken as a
result of intense atomic vibrations, the freed atoms tend to move from areas of high
chemical potential to areas of low chemical potential. For an atom under loading
conditions, this can be written as:
µ= µ0 − Ωσ
Where μ is the chemical potential (energy per atom), μ0 is the standard chemical potential
in a stress-free state, Ω is the atomic volume (volume per atom), σ is the localized stress on
the atom (energy per volume).
Rocks are a composite material composed of mineral grains of different morphology and
size. Under applied loading, stress distribution throughout a rock depends on the
mechanical properties of each individual phase. Some rocks deform elastically when
submitted to stresses under moderate temperature and pressure conditions but above a
certain threshold, the internal stresses produce cracks within the rock matrix [12].
Therefore, the mechanical properties of a rock depend on the interactions of individual
phases, but more importantly, the population of cracks and imperfections in the matrix.
The mechanical breakage of rocks or ore particles is accomplished by compression, impact,
chipping, and abrasion mechanisms. For crushing, compression is the main breakage
mechanism. For grinding, impact in the breakage mechanism. Chipping and abrasion
breakage mechanism result during all stages size reduction.
6
Griffith (1920) showed that materials fail by crack propagation when the strain energy is
greater than the energy needed to produce a new surface. Under an applied stress, the
stability criterion for minimizing the total bulk free energy (UT) of a rock is given as:
U T =− W + U e + U s
Where -W represents a decrease in the potential energy due to applied loading, Ue is the
stored strain energy, and Us is the total surface energy of the crack.
When atomic bonds fail by fracture, some of the crystals stored potential energy is
transformed into surface energy that is associated with newly exposed atoms on the surface.
Any increase in the potential energy of atoms will increase the chemical reactivity of the
newly formed surface that will often display irregular topography including surface defects
introduced by fracture [13]. This modification of the physio-chemical properties of the
newly liberated surface can now be utilized to improve hydro/pyro metallurgical processes.
1.5 Petrology of the Ore Body
Michigan’s Upper Peninsula is historically known for native copper but a magnetic
anomaly investigated in the late 1970’s located a massive Fe-Cu-Ni sulfide mineral deposit.
Michigan’s DNR performed the geological survey finding the deposit is ultra-mafic in
origin forming during the Precambrian era as a sulfur-rich magma conduit in the Baraga
dikes [14]. The Eagle sulfide mineralization was formed as an immiscible sulfide solution
in parental magma with fractional crystallization of sulfide liquids upon cooling. The
mineralization is composed of host rock peridotite containing massive Cu and Ni-bearing
iron sulfides with appreciable amounts of ferrospinel oxides and trace platinum group
metals (PGM’s). Now an economic prospect for metal recovery, the ore body is located in
the northwest corner of Marquette County (46°44ʹ47ʺN 87°52ʹ50ʺW) in an area known as
the Yellow Dog Plains (Figure 1.2).
7
Figure 1.2 Map of Michigan showing the location of Kennecott’s Eagle Mine.
Mining efforts began in 1994 with exploratory drilling by Kennecott Exploration (KEX) at
a prospective site near a sacred rock of local Indian tribes named Eagle rock [15]. By 2003,
mineralization was discovered and Kennecott’s “Eagle” project was launched. In April
2004, the project was handed over from exploratory activities to Kennecott Mineralization
Company (KMC) marking the potential for a mine development project. In December
2005, a consensus was reached on the new mining statute, “Michigan’s nonferrous metallic
mining regulations Part 632” [16]. In 2012, a mine infrastructure has been built including
a clear-cut site, access roads, storage buildings, lined tailing pits, and a water-treatment
facility despite a large local opposition to nonferrous “sulfide” mining. The estimated total
reserves at the Eagle site is ≈ 4 million tonnes containing 3.57 % Ni, 2.9 % Cu, and 0.10%
Co, 0.28 ppm Au, 0.73 ppm Pt, and 0.47 ppm Pd but also other minor amounts of metals
8
commonly associated with sulfide minerals [17]. When fully developed, the ore deposit
would be the only operational nickel mine in the United States.
The sulfide mineralization was formed in a dynamic magma conduit of the ~1.1 Ga
Midcontinent Rift System hosted in mafic-ultramafic intrusive rocks associated with the
Marquette-Baraga dike swarm in northern Michigan [18], [19]. The types of sulfide
mineral textures recognized in the deposit are: (1) disseminated sulfides in olivine-rich
rocks; (2) rocks with semi-massive sulfides located above and below the massive sulfide
zone; (3) massive sulfides; and (4) sulfide veins in sedimentary rock. The massive sulfides
are mainly composed of minerals pyrrhotite (65-80%), chalcopyrite (5-15%), and
pentlandite (10-25%) with PGM fractional zoning resulting from the cooling conditions of
monosulfide solution (MSS).
Petrographic and electron microprobe analysis from Michigan’s D.N.R.’s geological
survey shows that the peridotite contains 40 to 50 % olivine (Ol) and 20 to 30 % pyroxene
(Px), with clinopyroxene (Cpx) generally more abundant then orthopyroxene (Opx). The
Ol is Mg-rich containing 80 to 81 mole % forsterite (Fo80). The Px consists of 10 to 15
percent enstatite (En78Wo04Fs18), and diopsidic augite (En47Wo42Fs11). The plagioclase is
labradorite (An57-65) varying from 5 to 10 %. Major oxide minerals present are ilmenite
(Ilm) and magnetite (Mag) occurring from 4 to 6 %. Major sulfide minerals are chiefly
pyrrhotite (Po), pentlandite (Pnt), and chalcopyrite (Ccp) occurring from 1 to 2 %. Trace
amounts of sphalerite (Spl) and galena (Gn) are known to be present. The concentration of
Ni in pyrrhotite does not exceed 0.82 wt. % having an average Ni content of 0.52 wt. %.
Pentlandite in upper semi-massive zones tends to have higher Co content. It was found
85% of PGM mineralization occurs in sulfide veins with only 15% present in massive
sulfides occurring as sub-micron sized (Pt,Pd)-(Te,Bi,Ar,Sb) chalcogenide minerals.
It is inferred that the massive, metal enriched zones within the interior of a sulfide ore
deposit containing PGM’s should be processed differently than bulk material. All material
extracted from an ore body is considered as equal in value so high grade material is mixed
9
in with low grade material during processing. There is no regard or consideration to the
potential loss (or gain) of valuable metals in subsequent processing steps. It makes cents to
process metal enriched zones of a fractionated ore body with better methods to maximize
recovery of PGM’s during beneficiation.
There has been an opposition to the Eagle mine since the initial stages of development due
to the unmerited effects that sulfide mining can have on the environment [20], [21], [22].
The collective of individuals against development of the mine is called “Save the Wild
U.P.” [23]. The Keweenaw Bay Indian Community (KBIC) claims sulfide mining infringes
on their indigenous rights and land, asserting that Eagle Rock is sacred to them. Metallic
sulfide mining also threatens the habitat of coaster brook trout which are rare according
Michigan’s D.N.R. The Salmon Trout River, located in the vicinity of the mining
exploration, forms a tributary with Lake Superior and is the only place on U.S. mainland
where native coaster brook trout are known to spawn [24].
1.6 Silicate Minerals
Silicate minerals compose the majority of Earth’s crust forming a wide range of
compositions and structures [25]. Being extensively studied for many geological purposes,
vast amounts of thermal dynamic data is available on silicate systems, [26], [27], [28], [29].
These minerals form complex structures and chemistries of bridged oxide layers with high
cationic substitution. Silicate minerals form the parent rock in most geological systems and
are often a good indicator of perspective minerals during exploration. Silicate rock
comprises the majority of “gangue” material that is discarded as waste to tailings piles.
1.6.1 Peridotite Rock
The upper earth’s mantle is composed of mostly of peridotite rock; an ultramafic, course
grained, igneous rock with a high Mg content. It is dense and green in color from the Ol
content with a high hardness reflecting its Px content. It is composed of primary silicate
minerals, olivine (Mg,Fe)2SiO4 and pyroxene (MgFeCaSi2O6), with accessory minerals
10
plagioclase
(NaCaAlSi3O8),
amphibole
(Ca2(Mg,Fe)5Si8O22(OH)2)
and
chlorite
((Fe,Mg5)Al)(AlSi3)O10(OH)8, but also containing secondary minerals such as serpentine
(Mg3[Si2O5](OH)4). Serpentinization is a low temperature hydrothermal weathering
processes of exposed olivine in peridotite rock shown by the replacement reaction [30]:
2(Fe,Mg)2SiO4 + H2O + 2H+ → Mg3[Si2O5](OH)4 + Mg2+ + Fe2+
1.6.2 Olivine
The Mg-Fe silicate mineral olivine (Ol) is one of the most common terrestrial minerals. It
forms a continuous solid-solution between end-members forsterite (Mg2SiO4) and fayalite
(Fe2SiO4). The composition (Mg,Fe)2SiO4 is commonly expressed by a molar percentage
of the end-members forsterite (Fo) and fayalite (Fa) i.e. (Fo80Fa20). It has been
demonstrated that ferric (Fe3+) iron does not substitute into Ol but prefers B-site occupancy
in spinel (MgFe2O4). Olivine is unstable at ambient temperatures and subject to weathering.
The Fo-Fa system forms as simple binary solution having approximately 700°C difference
between melting temperature of the two end-members; Fo melts at 1890°C and Fa melts at
1205°C [31]. The nature of bonding type between end-members is the explanation for such
a large difference in melting temperature with Fo exhibiting a high degree of covalent
bonding and Fa exhibiting more ionic type bonding. Once a critical temperature is reached,
melting is a continuous phase transition from Olsolid→Olliquid. This phase transition is not
accompanied by large changes in the initial composition so Ol crystals and solution coexist
in equilibrium. This process is known congruent melting occurring in both binary and
eutectic phase systems.
During Ol dissolution, the Ca concentration gradient is small showing uphill diffusion
which indicates that Ca is rejected from Ol crystal during melting. Also, studies of nickel
(Ni) partitioning in Ol’s, with Fe-Ni sulfides, show that ferrous iron (Fe2+) strongly prefers
Ol, while Ni strongly prefers the sulfide solution, all being a strong function of temperature
11
and oxygen fugacity [32]. Dissolution of copper (Cu) into silicates is not as easily explained
because increasing the Fe concentration in silicate melt tends to increase Cu valency [33].
1.6.3 Pyroxene
Pyroxene (Px) is one of the most abundant rock-forming minerals in earth’s crust
displaying a variety of crystal chemistries due to a high degree of ionic substitution with
structural distortions and polymorphism extensive among compositions [34], [35]. The
structure is based on single SiO3 chains with linked SiO4 tetrahedra forming either
monoclinic (Cpx) or orthorhombic (Opx) crystal systems. In general, it can be written as
(M1M2T2O6) where M1 refers to cations occupying regular octahedral sites and M2 refers
to cations occupying distorted octahedral positions, and T corresponding to tetrahedral
coordination. The crystal structure can be summed up by the ideal site-occupancy of
cations positioned among the M1, M2, and T atomic sites. The data in Table 1 displays the
order of cation site occupation in Px but real site occupancy will often vary from ideal [36].
Table 1.1 Ideal site occupancy of metal cations in the M1, M2, T sites of pyroxene.
T
Si4+
Al3+
Fe3+
M1
M2
Al3+
Fe3+
Ti4+
Cr3+
Zn2+
Ti3+
Fe2+
Mn2+
Mg2+
Fe2+
Mn2+
Ca2+
Na+
Pyroxenes provide a source of metal cations during formation of spinel oxide minerals.
Spinels are referred to as the “garbage can” of minerals, capable of containing a variety of
cationic substitution but pyroxenes can be considered the “trash dump”. This attribute gives
12
pyroxenes high hardness create high energy requirements for beneficiation of spinel
minerals giving them a classification as refractory ore material.
1.6.4 Incongruent Melting of Pyroxene
Silicate solutions have been extensively studied for many geological and geochemical
purposes [37]. Here incongruent melting of pyroxenes, chemical diffusion of cations in
melts, and crystal growth are discussed.
In pyroxenes, metal cation substitution acts as a silicate chain network modifier effectively
creating negatively charged non-bridging oxygen atoms. The addition of metal oxides
disrupts bridging between Si-O-Si chains due to the addition of an extra oxygen ion.
Si–O–Si + MgO → 2(Si–O–)+ Mg2+
Si–O–Si + CaO → 2(Si–O–)+ Ca2+
Si–O–Si + TiO2 → 2(Si–O–)+ Ti4+
For pyroxenes containing various cationic species with different oxidation states, charge
balance is maintained by,
2M3+ ↔ M2+ + M4+
Solid state diffusion of cations in pyroxenes plays a critical role in the mass transport of
many geochemical processes [38]. The pyroxene crystal structure changes with
temperature (and pressure) that activates chemical diffusion and exchange of major cations.
Depending on the cation species present between bridged oxygen atoms, M-O bonds
expand at different rates during heating and cooling [39]. In alkali oxide glasses, the
modifier cation’s activation energy for diffusion decreases as the modifier cation
concentration increases [40]. Lowering the activation energy promotes diffusion in
channels of the network structure and increased isomorphism at elevated temperatures
which is shown experimentally by various cation diffusion mechanisms within the
pyroxene structure [41].
13
Pyroxene minerals display a wide range of compositional variation and melting
temperatures [42], [43]. Incongruent melting (Solid 1→Solid 2 + Liquid) of pure diopside
crystals (CaMgSi2O6) in the temperature (Tm) range of 1330– 1400 ºC was much below
what was previously reported by researchers [32]. This type of melting takes place
according to the dissociation reaction,
CaMgSi2O6 → Mg1-xSi2O6(solid) + SiO2(melt) + xCa2+(melt)
The free energy-composition diagram for the two pyroxene minerals enstatite (Mg2Si2O6)
and diopside (CaMgSi2O6) shows the existence of several polymorphs [44]. Pure diopside
(Di) can only occur in the monoclinic (DiCpx) form but enstatite (En) can exist as either a
monoclinic (EnCpx) or orthorhombic (EnOpx) polymorph. At Di-rich compositions, mole
fraction (XDi > 8.5), only DiCpx exists. At En-rich compositions, mole fraction (XDi < 1.5),
the only phase present is EnOpx. At intermediate compositions (8.5 > XDi > 1.5) all available
polymorphic forms (DiOpx, EnCpx, EnOpx) exist with the amount of each poly-type present
dependent on the composition of the melt. Orthorhombic diopside (DiOpx) is
thermodynamically unstable and does not exist.
Incongruent melting of pyroxene introduces chemical heterogeneity and diffusion of
atomic species in solution [45]. Crystal growth and exsolution is achieved in pyroxenes by
nucleation and growth or by spinodal decomposition but sometimes occurring by both
mechanisms [46]. In either case, melting temperature (Tm) and concentration of cations in
solution (activity) dominates crystal growth and diffusivity with ionic radii and charge of
a chemical species does not have much influence on the kinetics [47]. Therefore,
dissolution of Px into a melt is much faster than the diffusive exchange of solid to melt
allowing En crystals (Mg1-xSi2O6) to retain their original composition.
1.7 Sulfide Minerals
Generally known as chalcogenides, the physical and chemical properties of sulfide
minerals are the most diverse and complex in comparison with all other mineral groups
[48]. The origin of their properties comes from the strong covalent nature of close-packed
14
sulfur anions, having moderate electronegativity, which permits various cationic
coordination’s and hybridization of atomic orbitals. The nature of bonding among ions,
and the resulting interactions between their atomic orbitals, reflects the variety of
stoichiometries, densities, chemical stabilities, oxidation states, coordination numbers,
crystal chemistries, dielectrics, and magnetism displayed by sulfide minerals [49].
The chemical bonding and electronic structure formed between ions in sulfide minerals
results in a variety of electrical and magnetic properties [50]. The complex exchange
interactions of electrons in atomic orbitals give sulfides electrical properties that range
from insulators to semiconductors and metals while also exhibiting all types of magnetic
properties. Most sulfides with technological or industrial importance exhibit electrical
properties of a semiconductor or metal.
The electronic polarizabilty (i.e. displacement of the electron cloud relative to the nucleus)
is greater for the S2- anion (10.2 × 10-24 cm3) then the O2- anion (3.88 × 10-24 cm3). The
moderate electronegativity of sulfur (2.58), in comparison with oxygen (3.44), allows for
weaker metal-anion electron interactions with respect to bond-type. The electronic
polarizability and electronegativity values can be used for describing anion-anion
interactions in the structure and the amount of covalence between anion-cation bonds.
In crystalline materials, cation coordination in the structure is dependent on size, oxidation
state, and electron configuration of the species. The cation coordination number is defined
using the hybridization of orbital concept. For transition metals sulfides, the hybridization
of orbitals explains the coordination of 4-fold tetrahedral cations to sp3, 4-fold squareplanar cations to dsp2, 6-fold octahedral cations to d2sp3, and 5-fold square-pyramidal to
dsp3 [51]. Transition metal sulfides can be further categorized by their electrical and
magnetic properties resulting from strong interactions of d-orbital electrons [52]. A
summary of hybridization types for common transition metal cations that occur in many
common sulfides is shown below in Table 1.2.
15
Table 1.2 Coordination number, orbital hybridization, and electron spin state for Fe, Cu,
and Ni transition metal cations in sulfides.
CN
Hybridization
Cation Type
4
sp3
Cu2+, FeH2+
4 (sq)
dsp2
FeH3+ , NiH2+ , NiH3+ , Cu2+
5
dsp3
NiL2+
6
d2sp3
FeL2+, FeL3+, NiL4+
Many poly-types of sulfides exist having a variety of crystal chemistries and structural
arrangements that have attracted scientific interest for technological applications but their
industrial importance as an ore material is the main reason for extensive studies of these
minerals [53]. Metallic-bearing sulfides are the most important of all ore types because
they supply the raw material for world production of non-ferrous metals. Fe-Cu-Ni-S ore
deposits are the most economically important due to high concentrations of metals.
In geochemical systems, the bulk reaction kinetics determines a mineralization’s
composition but can be altered by its surrounding geological environment [54]. The overall
composition, over a geological timescale, will result in local redistribution of metals by
mineral exsolution and recrystallization occurring at temperatures and pressures different
from the initial conditions of formation. Also, hydrothermal processes can produce
compositional changes by supergene oxidation and/or mineral precipitation [55]. Proper
interpretation of ore genesis informs a mining operation of the best strategies for extraction.
The unmerited environment effects of sulfide mining creates large public opposition to
these type of operations. This includes many topics on environmental geochemistry of
sulfide-oxidation but the biggest concern is sulfuric acid produced by water leaching
through mine tailings; this is also known as acid mine drainage (AMD) [56]. Sulfuric acid
(H2SO4) is harmful when it enters into the environment because it sharply increasing the
16
pH of the local hydrological cycle. Prevention of AMD is accomplished by containing
runoff water and proper construction of tailing ponds [57]. Environmental retribution is
made possible with sustainable mining efforts during the life of the operation along with
proper mine remediation after resources are spent.
Metallic sulfides (MS) are chemically unstable and dissociate upon oxidation into metalsulfate (MSO4) and/or metal-oxide (MO) reaction products even under the slightest
oxidizing conditions. Formation of metal-sulfide to metal-oxide can be represented by [58]:
MS(s) + O2(g) → MO(s) + SO2(g)
The temperature at which the reaction MS→MO proceeds varies between metal-sulfide
compounds. During heating, cations to diffuse toward the oxide interface and the highly
exothermic formation of SO2 gas produces mass loss of the reactant(s) [59]. The creation
of an oxide layer will affect the physical and chemical properties of a particle surface that
can either aid or hinder metallurgical processes [60], [61].
Metal sulfide solution systems have received considerable attention but prove problematic
in studies due to extensive solid solutions, miscibility gaps, non-quenchabable phases, slow
reaction kinetics and metastability. Many phase transformations are reported in sulfide
minerals upon cooling such as polymorphic, order-disorder, inverse peritectic
(incongruent), dissociation reactions, magnetic transitions, etc. These are likely due to nonideal solution behavior from strong metal-metal interactions but also with sulfur fugacity
having a significant effect on the stability of phase fields.
1.7.1 Pyrrhotite
Aside from pyrite, pyrrhotite (Po) is the most common iron sulfide on Earth created under
a variety of geological conditions [62]. It forms an array of non-stoichiometric (Fe1-xS) iron
deficient structures between end-members triolite (FeS) and pyrite (FeS2). Natural Po
displays a range of compositional variations (0 < x < 0.125) resulting in the formation of
superstructures with slightly different metal to sulfur ratios. All compositional variations
17
of Po exhibit ferrimagnetic properties that contribute to the natural remnant magnetization
of Earth’s crust which provides geophysical signals that make Po a good indicator of metalbearing sulfide deposits [63]. Though Po is the primary mineral associated with Cu-Ni ore
deposits, it has zero economic value and is rejected to mine tailings during processing [64].
Po superstructures are based on the hexagonal Ni-As structure with ordering of cation
vacancies in alternate layers down the C-axis. The list of Po superstructures that form
below 100 ºC are: Fe7S8 (4C), Fe9S10 (5C), Fe10S11 (11C) and Fe11S12 (6C) denoted by an
integral number about the C-axis with respect to the Ni-As (6/m 2/m 2/m) 1C unit cell. Of
these polytypes, the most significant is the pseudo-monoclinic (F2/d) 4M-type Po (Fe0.875S)
that displays strong ferrimagnetic properties. When written, the chemical formula for
magnetic 4M-Po is Fe23+Fe52+S82- but for hexagonal NC-Po is Fe23+Fe72+S102- showing that
NC-Po contains more antiferromagnetically coupled iron [65].
Magnetism in 4M-Po is due to long range ordering of vacancies down the c-axis that creates
a net magnetic dipole moment by aligned Fe atoms producing ferrimagnetism and an
effective magnetic moment of 5.24 μB. This is accomplished by different (A, B, C, D)
alternating metal deficient layers coupled with fully occupied (F) layers in an AFBFCFD
sequence [66]. Cation vacancies alternate sites every other row which orders them in a way
that achieves maximum separation between one another and strongly influences magnetic
properties. Thus, the local charge balance of the crystal is maintained by vacancies.
All types of Po superstructures share similar XRD reflections with the most intense
occurring at 2θ ≈ 34° and 44° corresponding to the [004] and [522] planes respectively.
The composition of Po, or the x in Fe1-xS, can be determined by the d102 lattice spacing.
Though XRD can be used to distinguish hexagonal and monoclinic polytypes of Po, it can
be shown geometrically that monoclinic type Po is a derivative of the hexagonal structure.
Po incommensurate polytypes are poorly characterized by XRD methods because of the
defect chemistry associated with vacancy concentration. The long-range ordering of
18
vacancies in iron deficient layers makes XRD somewhat ambiguous because x-ray
radiation scatters from atoms, not vacancies. Further pulverizing samples disrupts
crystallinity and long range ordering of superstructures. Additionally, no experimental
research has been performed regarding structural changes brought by mechanical stress
that induces strain on the Po crystal lattice.
Many of the low temperature (< 350 ºC) phase relations of Po are unclear with portions of
solvus lines marked with question marks and dashed lines [67].Non-equilibrium cooling
conditions, creates complex non-stoichiometric superstructures and introduces metastable
phases [68], [69]. A proper interpretation of Fe-S phase relations is the basis for
understanding iron-sulfide systems with additional components.
At iron-rich compositions, Po undergoes the Morin transition (atomic spin-rotation)
accompanied by the α-transition (super structure transformation) at Tα= 140 ºC. At 223 ºC,
the γ-transition is observed with a rise in the magnetic susceptibility due to the structural
transition of NC to NA polytypes. At temperatures in the ≈ 315 ºC range, 4M-type Po under
goes the β-transformation to a 1C polytype + pyrite (FeS2) with the binary sub-solidus
phase (1C + pyrite) now exhibiting paramagnetic behavior. Naturally occuring 4M-Po is a
highly conductive material exhibiting p-type semiconductor electrical properties with
resistivity values in the range of 10-6 to 10-1 Ω·m [70]. For non-stoichiometric Po, electrical
conduction in the solid occurs by electron hole (h-) hop involving Fe3+ cations and
interacting metal vacancies (VFe).
Fe3+ + VFe2- → Fe2+ + h+
At compositions near 46 at. % Fe, it is likely there is a strong driving force for magnetic
4M-type Po to occur with NC-type subsequently forming around it. The presence of
vacancies leads to oxidation of ferrous (Fe2+) iron to ferric (Fe3+) iron which produces
stronger ferrimagnetism. Increasing the iron content (Fe7+xS8) fills vacancies that leads to
more complex stacking of metal deficient layers and destroys long-range ordering of the
ferrimagnetic structure. Approaching compositions near trolite (FeS), all metal sites
become filled with alternating Fe and S layers in hexagonal symmetry along the <001>
19
direction. Figure 1.3 qualitatively illustrates a composition gradient among lowtemperature Po superstructures.
Figure 1.3 Illustration of low-temperature phase relations in the Fe-S system suggesting a
composition gradient existing among superstructures formed between end-members due to
a strong driving force in the formation of magnetic 4M-type Po (Fe7S8)
Upon heating, Po assemblages undergo a variety of crystallographic transformations and
magnetic transitions [71], [72]. Differential thermal analysis (DTA) studies show unit cell
structural transitions, marked by exothermic events, that have a large enthalpy change
associated with disordering of vacancies at the Curie temperature [73]. Thermal agitation
causes vacancy motion and disordering leads to structural transformations [74], [75]. The
pyrolytic decomposition of Po to Fe-oxide and sulfur dioxide gas is a highly exothermic
reaction and can be shown by the equation [76],
2Fe1-xS +
5
O 2 → (1-x)Fe 2O 3 + 2SO 2
2
During mineral processing of iron sulfide ores, Po is separated from the concentrate and
rejected to mine tailings as waste. Since not all Po exhibits strong ferromagnetic properties,
generally two steps are involved in the separation process. First, 4M-Po is removed by
magnetic separation then the remaining H-type Po is removed by froth flotation. Therefore,
surface chemistry of Po under aqueous conditions has been of great interest for improving
flotation performance [77], [78], [79].
20
1.7.2 Chalcopyrite
The ternary iron copper sulfide mineral chalcopyrite (CuFeS2) is the main source of
metallic copper provided from ore resources. It is a yellow in color, often misinterpreted
as fool’s gold (pyrite). Chalcopyrite (Ccp) readily alters when subjected to weathering
(redox) conditions and exhibits a variety of colors thereafter coining the term peacock ore.
The crystal structure of Ccp (I42d) is based on the sphalerite (ZnS) structure but is a double
stacked cubic close packed structure with tetragonal unit cell of lengths a1 = a2 = 2c [80].
Metal cations are in 4-fold coordination with sulfur atoms that form in a cubic close pack
structure with tetrahedral sites shared equally among Cu and Fe atoms. The long-range
ordering of Cu and Fe atoms creates an extensive solid solution with the distribution of
cations having varying degrees of ordering from completely ordered to completely random
[81]. These compositional variations change the partitioning of Cu and Fe on tetrahedral
sites affecting the short and long range charge distribution of the crystal.
The chemical formula for Ccp can be written (Cu,Fe)S2 indicating a probability that either
metal species will occupy 50% of the available tetrahedral sites. This would suggest a
chemical formula of Cu2+Fe2+S22- but there is disagreement in the nature of oxidation states
among metal cations. More recent studies show that the chemical formula is likely closer
to Cu+Fe3+S22-, indicating an appreciable amount of covalency within the crystal structure.
Ccp is antiferromagnetically ordered up to a Néel temperature of 823 K (550ºC). The
antiferromagnetically ordered structure arises from opposed spins of the two Fe cations
along the c-axis that are tetrahedrally bonded to a common sulfur atom giving rise to
indirect exchange between the intermediate sulfur anion [82]. Neutron diffraction studies
show a magnetic moment of zero for Cu and 3.85μB for Fe which strongly supports this
configuration [83].
Natural samples of Ccp show n-type properties with band gap values of 0.33 eV, 0.5 eV
and 0.6 eV. Additional amounts of Cu contributes extra electrons to the conduction band
21
while additional Fe provides holes in the valence band for Cu+ electrons to hop into [84].
Thus, Ccp is an n-type or p-type semiconductor depending on the composition and defect
chemistry of the crystal structure. This is due impart to the nature of Fe and Cu 3d-orbital
configurations and electron delocalization that affects electrical properties [85]. Above
tempertures of 350ºC, Ccp exhibits intrinsic semiconductor behavior.
During heating in the temperature range of 350 ºC and 450 ºC, Ccp oxidizes marked by
strong exothermic effects during DTA studies [86]. Upon oxidation, the magnetic
properties of the Ccp surface are enhanced due to the formation of maghemite (γ-Fe2O3),
[87]. The magnetic susceptibility increases with the volume of maghemite formed on the
surface during oxidation that can be useful for magnetic separation [88]. Partial oxidation
starts at 350ºC, and is finished by 850ºC, represented by the combined reactions:
2CuFeS2 + 15/2O2 → 2CuSO4 + γ-Fe2O3 + 2SO2
2CuFeS2 + 13/2O2 → 2CuO + γ-Fe2O3 + 4SO2
1.7.3 Pentlandite
The mineral pentlandite (Pnt) is a major source of Ni and invariably associated with Po in
sulfide mineral assemblages. It’s ideal chemical composition, (Fe,Ni)9S8, suggests a
“metal-rich” sulfide mineral forming a slightly distorted cubic close pack structure of 32
sulfur anions having cubic clusters of metal atoms with short metal-metal distances [89].
The Pnt structure has four formula units of (Fe,Ni)9S8 per unit cell with 36 metal atoms
total; 32 in tetrahedral sites and the remaining 4 in octahedral sites (Fm3m). The unit cell
parameters (interatomic distances) are sensitive to the type of metal cation, coordination
number, electron spin state, and valance of atomic species. The idealized metal-to-sulfur
ratio is 9:8 but the exact ratio is not known and varies in natural samples [90]. Pnt is shown
to have low resistivity and exhibits Pauli-paramagnetism. When Pnt is heated to even
moderately low temperatures (150º-200ºC), the crystal structure undergoes irreversible
thermal expansion associated with disordering of Ni and Fe [91].
22
The thermochemistry of the Fe-Ni-S system has been extensively studied using
experimental data from ternary phase equilibria and molar Gibbs energy expressions to
understand the kinetics of solution during heating and cooling [92], [93], [94]. At
temperatures exceeding 1200 °C, Fe-Ni-S liquid is present with the Ni-Fe rich portion of
the system solid taenite (FeNi), a γ-FCC alloy commonly found in meteorites. At
temperatures T ≈ 610 °C, a complete solid solution of (Fe,Ni)1-xS (mss) is formed having
the Ni-As structure [95]. Upon cooling, decomposition of mss results in exsolution of Pnt
from nickeliferrous-pyrrhotite (Ni-Po) downward to temperatures of 300 °C [96]. During
exsolution, Ni is rejected out of mss solution and concentrates in Pnt with Fe-Ni-S
mineralizations usually having less than 1 at. % Ni contained in Po [97]. Mineral grain
textures of Pnt are dependent on the exsolution conditions during cooling. Three textural
varieties of Pnt are common to natural Ni assemblages: 1) massive or blocky, 2) partial or
rimmed, and 3) stringers or flame lamellar [98].
The secondary mineral violarite (Vo), with the ideal chemical formuala FeNi2S4, is a
thiospinel mineral that sometimes forms by exsolution during cooling of mss. Most
commonly, Vo is formed during weathering of Pnt by dissolution and reprecipitation in the
supergene zone of a sulfide mineralization. The pyrolytic decomposition of Vo occurs at
420ºC and is a highly exothermic reaction so Vo can be used to kick-start smelting [99].
1.7.4 Metal-Sulfide Phase Equilibria
The phase relations of high temperature metal-sulfide solutions have been subject to
numerous studies for mineralogical, geophysical, and mining exploratory purposes [67].
The Fe-Cu-Ni-S mineral deposits are especially attractive for their precious metal (PGM)
content [100]. Metal-sulfide mineralizations solidify by fractionation crystallization of
metal-sulfide liquid resulting in compositional zoning of metals. An accurate interpretation
of the bulk fractionated composition of a Fe-Cu-Ni-S mineralization could improve metal
yield during beneficiation by utilizing the best extraction technique(s) for metal rich zones
of an ore body [101]. The “wetness” of bulk mineralization refers to the amount of
23
entrained metal-sulfide liquid remaining in the melt after solidification which is used to
measure the concentrations of PGM’s.
As magma, oversaturation of metal-sulfide liquid (sul-mt) is relieved by sulfur
volatilization and degassing of the melt [105]. Concentrations of dissolved sulfur in
deposits shows little dependence on melt composition but a strong dependence on
temperature and pressure [106]. Due to low sulfur solubility, sul-mt is immiscible in silicate
melts (si-mt). This has led to an assumption that sul-mt interacts with its geochemical
environment in a way that scavenges metals from the surrounding magma. This is
supported by high partitioning of Fe, Ni, Cu, and PGM’s occurring in sulfide
mineralizations. Formation of sulfur bearing accessory minerals requires removing sulfur
from large volumes of melt; a process highly dependent on sulfur and oxygen fugacities.
Thus, the rate of sulfur diffusion within the melt is controlled by the kinetics of crystal
growth indicated by sulfide mineral textures.
At high temperature, Cu-Ni-Fe-S is a continuous metal-sulfide liquid [107]. It is thought
that (Fe,Ni1-xS) monosulfide solution (mss) and (Cu,Fe)1-xS intermediate sulfide solution
(iss) are related by a miscibility gap but mineral texture relations show these sulfide phase
fields solidify at different temperatures starting with early crystallization of mss followed
by iss [108]. Since Fe occurs at the highest concentrations in these systems, it has the
greatest chemical activity making it the dominate component during solidification. At
temperatures near ≈ 1190 ºC, mss initially solidifies to magnetic Ni-rich pyrrhotite (Po).
The mss exsolves pentlandite (Pnt) down to 300 ºC during which Ni is rejected from Po.
The existence of the miscibility gap between Ni-Fe-S and Cu-Fe-S solutions suggests the
Cu-Fe-S solution begins to solidify at temperatures of 1129 ºC. First with the formation of
high digenite (Cu,Fe)9S5, then iss is subsequently formed at 960 ºC. At around 600 ºC, the
phase field completely extends from bornite (Cu5FeS4) to Po (Fe1-xS) with chalcopyrite
(CuFeS2) formed at intermediate iss compositions between these end-members. PGM’s are
chemically insoluble in mss and remain in iss until final solidification.
24
By using the phase relations of various metal-sulfide systems and mineral textures, a route
for fractionation crystallization has been proposed [109]. At temperatures >1200 ºC, the
Cu-Fe-Ni-S system is complete liquid metal-sulfide solution (sul-mt). Upon cooling to
1100 ºC, a portion of the sul-mt crystallizes forming mss 1 + liquid 1 that results in
fractionation of mss 1 → disseminated mineralization 1 comprised of Po + Pnt. Cooling to
1000 ºC, liquid 1 decomposes to mss 2 + iss 1 + liquid 2 resulting in fractionation of mss 2
+ iss 1 → semi-massive mineralization 2 comprised of Po, Pnt, and Ccp. Further cooling
to 900- 800 ºC, liquid 2 yields mss 3 + iss 2 + liquid 3 resulting in fractionation of mss 3
+ iss 2 → massive mineralization 3 that contains all major sulfide minerals (Po, Pnt, Ccp)
and precious metals (PGM’s). Final solidification of liquid 3 → massive mineralization 4
results in miscellaneous accessory sulfide minerals and residual formations. A flow
diagram of high-temperature metal-sulfide liquid fractionation crystallization and the
resulting formations of low-temperture mineral assembaleges is shown in Figure 1.4.
A bulk mineral deposit is subject to compositional changes by natural geological processes
such as partial melting, exsolution, recrystallization, and deformation. These occur at
temperatures and pressures different than the original emplacement conditions that results
in an increase in the mobility of minerals allowing for local redistribution (transport) of
metals [102]. Also, minerals on the surface exposed to the hydrological cycle are subject
to weathering processes which also play a role in the redistribution of minerals and metals
[103]. These processes, spanned over a geological time-scale, can significantly alter the
bulk composition of a mineralization [104].
25
Figure 1.4 High-temperature fractionation crystallization of metal-sulfide liquid and the
resultant formation of low-temperature mineral assemblages
1.8 Spinel Minerals
Spinel minerals are a chemically diverse group of oxide minerals forming in the cubic
crystal structure. Spinel minerals are referred to as the “garbage can” of minerals, being
able to accommodate a wide variety of cationic substitution and form an extensive solid
solution series between end-members. The cationic distribution of transition metals in
spinel minerals permits a wide range of electrical and magnetic properties that have
prompted extensive mineralogical and technological studies [110].
Naturally occurring ferrospinel minerals are common to extrusive and intrusive igneous
rock being the main contributor to rock magnetism in earth’s mantle [113], [114]. Pure end
26
members that form a solid solution series are rare as natural minerals and are sub-divided
on the basis of the dominant M2+ and M3+ metal cations [115]. Spinel-ferrites, which
contain Fe3+ as the main component, are commercially important in the production of
magnetic materials for electronic device applications.
Magnetite (Fe3O4) is one of the most studied minerals due to its abundance in nature,
magnetism, high conductivity, and noted phase transitions [111]. It is also a major ore
material in the production of iron and steel. Chromite (FeCr2O4) is the world’s source of
elemental chromium but is also used in furnace brick material (refractory) due to its high
melting temperature, moderate thermal expansion properties, and neutral chemical
behavior [112]. Both these minerals often occur as finely disperesd mineral grains so they
are regarded as refractory ore requiring high energy for liberation and refining.
In the spinel structure, there are 56 atoms per unit cell in which 32 oxygen anions stack as
the cubic close pack structure (Fd3m). The 24 metal cations are in either tetrahedral or
octahedral coordination with oxygen anions among the 64 and 32 respective sites. Two
structural types of spinel occur, normal and inverse, distinguished by the distribution of
cations in tetrahedral (A sites) and octahedral (B sites) coordination. Metal cations have A
or B site preference to maintain local charge balance with the distribution of cations to sites
determined by the interplay of atoms during chemical and magnetic ordering [116].
The mineral spinel (MgAl2O4) has one formula unit that can be written as A82+B163+O32. In
normal spinel minerals, 8 M2+ cations occupy A sites and 16 M3+ cations occupy B sites.
These can be sub-divided into normal spinel structures having one divalent (M2+) cation
existing with two trivalent (M3+) cations (M2+M23+O4) or one quadvalent (M4+) cation
existing with two divalent cations (M4+M22+O4). In inverse spinel minerals, 8 M3+ cations
occupy A sites with the remaining 8 M2+ and 8 M3+ cations randomly distributed in B sites.
To illustrate site occupancy with inverse spinel, the chemical formula for Fe3O4 can also
be written as (Fe3+)A(Fe2+ Fe3+)B(O2+)4 although the ideal chemical formula will deviate
with temperature and oxygen fugacity.
27
Magnetite (Mag) is a ferrimagnetic mineral due to anti-parallel coupling of atomic
moments. The anti-parallel coupling of ferric (Fe3+) iron in A and B sites leads to ferrous
(Fe2+) iron in octahedral position only contributing to the effective magnetic moment which
is shown in Figure 1.5 [117].
(Fe3+)tetra
→
-5μB
(Fe3+
←
+5μB
Fe2+)octa
←
+4μB
Figure 1.5 The sum of magnetic Fe moments seated in octahedral and tetrahedral positions
of the inverse spinel mineral magnetite (Fe3O4)
The conductive properties of Mag are also due to the presence ferrous (Fe2+) iron that
permits n-type semiconduction. Thermally activated small polaron conduction occurs by
electron hopping so electrical properties are determined by the concentration of donor
(Fe2+) and acceptor (Fe3+) ions [119]. The electron motion between Fe2+ and Fe3+ cations in
octahedral coordination does not change the local charge balance. This translates to
electron-hole hopping between the 3d6 and 3d5 orbitals requiring only a relatively small
activation energy (0.065eV) [120].
The semi-conducting properties of defect spinel ferrites are mainly from the cation
deficient structure and production of vacancies [121]. This is caused by substitution of
metals with different valences (extrinsic defect) or by a change in oxidation state of the
metal ion (intrinsic defect). These defects can be illustrated by substitution of Ti for Fe in
ulvöspinel (Fe2TiO4) shown by 2Fe3+↔ Fe2++ Ti4+. To preserve local charge balance in the
crystal structure, metal-cation vacancies (VM) are created by mass transfer of oxygen. This
can be shown in Kroger-Vink notation as: 2Fe2+ + ½O2 ↔ 2Fe3+ + VFe + O2-. Further
thermal excitation leads to structural rearrangement in spinel-ferrites by Fe3+ hopping from
A-sites to B-sites accompanied by a Fe-vacancy (VFe) shown by: Fe3+(tet) ↔ Fe3+(oct) + VFe.
This is made possible by Fe cations adopting new electron spin configurations [122].
28
1.9 Platinum Group Metals
Platinum group metals are invariably associated with sulfide mineral deposits making the
discovery of new mineralizations considerably attractive [123], [124]. But low
concentrations of these metals enclosed within indistinct zones of a mineral deposit makes
recovery of PGM’s increasingly difficult. PGM’s are most common in sulfide deposits but
can also be found in layered strata of vanadium enriched chromites [128].
The Bushveld Complex, located in South Africa, reportedly contains ≈ 80 % of the world’s
PGM reserves [125], [126]. The mafic sulfides at the Bushveld Complex were formed by
repeated injection of sulfide magma into a sub-volcanic, shallow-level chamber. The large
volume of magma made cooling conditions, and crystallization, an extremely slow process
that exsolved PGM’s from concentration of a few ppb and enriching them to a few ppm.
Many other PGM resources are found to be layered intrusions also resulting from slow
cooling processes [127].
The flotation of sulfide minerals is significant to proper recovery of PGM’s for
metallurgical refinement. In Ni-Cu sulfide mining operations, pyrrhotite (Po) is rejected by
flotation as waste to mine tailings. Therefore, the texture of the Po-PGM occurrence
determines the overall flotation efficiency of the cell [129]. The Merensky reef of the
Bushveld Complex is a location where flotation of Po is significant for recovery of PGM’s.
29
1.10 Microwave Processing of Materials
1.10.1 Introduction
Microwaves were developed in the early 1940’s for military communications and were
observed to generate thermal energy, then known as radio-frequency heating [130]. Current
microwave (MW) technology spans a variety of fields from communications to metallurgy
but due to this very broad application, MW applications are still being developed and
improved upon [131]. More recently, complex theories of microwave-material interactions
combined with improved experimental characterization have furthered the development of
MW heating for material processing technologies used in commercial applications.
The MW band is the most versatile form of electromagnetic energy in terms of its
commercial applications. The ability to transmit MW energy at high power levels over
considerable distances gives it a variety functions over other forms of radiation in the
electromagnetic spectrum. There are six diverse fields where MW technology is used:
communications, material processing, science research, medical fields, food processing
and national defense.
In the electromagnetic spectrum, MW energy is the frequency range between 300 MHz
and 300 GHz corresponding to wavelengths between 1m and 1mm. This frequency range
is sub-divide into three bands: the ultra high frequency band (UHF: 300 MHz – 3 GHz),
the super high frequency band (SHF: 3 GHz – 30 GHz), and the extremely high frequency
band (EHF: 30 GHz – 300GHz). Since MW’s are largely used for telecommunications,
restrictions have been imposed on the frequencies used for heating applications. Currently,
the most common frequencies used for heating are 915 MHz and 2.45 GHz corresponding
to wavelengths of 33.5 cm and 12.2 cm, respectively.
For material processing requiring thermal energy, MW’s has been used for a variety of
applications. Some of these include drying, synthesis, chemical reactions, physical
modifications, sintering, de-binding, annealing, and joining. Though currently not well
understood in terms of thermodynamics, using MW’s as a means of generating thermal
30
energy in materials will become more significant in the future. Some advantages of using
MW heating for material processing applications include [132]:
•
quick startup time
•
rapid temperature increase
•
volumetric and selective heating
•
unique product formation
•
energy efficient
•
readily automated equipment
•
excellent safety standards
•
environmentally safe
The basis of MW heating is transmission and absorption of energy through the dielectric
response of electronic or magnetic dipoles that re-orientate themselves in the presence of
the applied electromagnetic field. The inability of dipoles to stay in phase with the rapidly
alternating field causes them to lag resulting in dissipation of energy in the form of heat, a
process known as irradiation. MW heating is also referred to as hysteresis heating, or joule
heating, but is a combination of both. During MW processing, absorption of MW energy
into a material causes frictional heating which can activate physical and/or chemical
reactions [133].
There are three main microwave heating mechanisms: electrical conduction, dielectric loss,
and magnetic loss. The effect of electrical conduction originates from an increased
concentration of charged particles (ions, electrons, vacancies). Electronic polarization
torque induced on electronic dipoles under an applied EM field. Magnetic loss is the
absorption of MW energy into a material with non-zero magnetic susceptibility. Thus, an
electrically conductive material containing magnetic dipoles i.e. (Fe3O4) can absorb MW
energy from separated E and B fields, or simultaneously from both fields.
Materials can be classified in accordance to their response to MW energy as conductors
(reflective), insulators (transparent), or dielectric (absorbing) [134]. When referring to their
31
dielectric loss factor, materials have high-loss, low-loss, or zero loss in a MW field. High
loss materials absorb MW energy and convert it into heat. Low loss materials transmit MW
energy without substantial absorption into the medium. Zero loss materials do not interact
with MW energy because wave propagation is redirected upon reaching the medium, or
travels through the medium with no loss of energy. Most often, materials contain different
phases, each with unique dielectric properties, so these materials are called mixed or
composite type.
MW heating of water (H2O) is essentially different than that of most other materials but is
used as an example of a molecule with a polarization mechanism. Unlike metals or
ceramics, an H2O molecule responds to MW field because it is composed of hydrogen
bonds that are directional and are inherently polarized. The electrons that compose bonds
attenuate MW energy which exerts torque on molecules that forces them to oscillate rapidly
creating friction between one another. This response to MW energy is termed attenuation
analogous to a radio antenna receiving a broad band signal. Similarly, other materials heat
by MW energy during polarization of dipole mechanisms functioning at various molecular
levels and attenuate MW energy.
It has been experimentally shown that many minerals and compounds will effectively heat
when exposed to MW energy [135]. This can be attributed to the semiconducting crystal
structure of many mineral types when exposed to an applied electromagnetic field, and the
frictional heating (irradiation) caused during atomic vibrations. [136]. The presence of
magnetic dipoles within the crystal structure will also significantly contribute to MW
heating. Thus, the material properties that affect the transmission and absorption of MW
energy are electrical conductivity, permittivity and permeability.
MW energy is a form of non-ionizing electromagnetic radiation that effectively heats
matter by transmission and absorption of energy. The particles (electrons, atoms,
molecules) that compose matter become excited from attenuation of MW energy with
heating being the effect of frictional losses as particles collide with one another.
32
Microwave-material polarization mechanisms are atomic in scale with each mechanism
having maximum response in a certain MW frequency band [137]. The dielectric response
of polarization mechaims can be traced back to the quantum structure of atomic particles
that compose matter [138].
1.10.2 Microwave-Material Physics
From physics, the acceleration of charge creates electromagnetic (EM) energy that
propagates as a wave in space capable of interacting with its surroundings. As the wave
experiences a change in medium, it can be fully transmitted, reflected, or absorbed. Very
often though, some energy is absorbed by the medium that converts it into a different form
of energy.
The term attenuation is used to describe energy loss into a medium; for matter, the amount
of EM energy absorption is determined by the dielectric properties. A common example
being when light interacts with matter the partial absorption of various wavelengths by
electrons is perceived as color to the human eye [139]. In general, EM attenuation by matter
depends on the intensity (frequency and power) of EM energy and the dielectric properties
(permittivity and permeability) of matter.
In the electromagnetic spectrum, all forms of energy have a characteristic frequency as it
propagates in space [140]. The Maxwell equations are physical laws used to describe timevarying electromagnetic fields, shown in classic, time-dependent form,
∇ × � = −
�

� = � +
∇ × 
� = 
∇ ∙ 
∇ ∙ � = 0
(Faraday’s Law)
�


(Ampere’s Law)
(Guass’s Law)
(No Magnetic Charges)
33
Where, � is the electric field intensity (V/m), � the magnetic flux density (Wb/m2),  is
� is the electric flux
the magnetic field intensity (A/m), � is the current density (A/m2), 
density (C/m2), and  is the charge density (C/m), The constitutive relations correlates the
between the electromagnetic field intensities and their corresponding auxiliary fields with
the dielectric properties of a medium given by,
 = 
 = 
Where,  is permittivity (F/m) and  is the permeability (A/m) of the medium. All forms
of matter have dielectric properties permittivity (±ε) and permeability (±μ) which
determine the dielectric response to an externally applied EM field (Figure 1.6). Ultimately,
dielectric properties are established by the energy of particles that compose the type of
matter and their quantum interactions with electromagnetic radiation.
Figure 1.6 The dielectric properties, permittivity () and permeability (µ), of matter.
For real matter, the dielectric properties are denoted by the complex dielectric constants
for permittivity () and permeability (µ) shown respectively by,
 = 0  = 0 (′ − ′′ )
34
 = 0  = 0 (′ − ′′ )
Where, 0 is the permittivity of free space (8.854 × 10-12 F/m), 0 is the permeability of
free space (1.256 × 10-12 H/m),  is the imaginary unit equal to √−1,  is the relative
permittivity, ′ is the relative real permittivity, ′′ is the relative imaginary permittivity, 
is the relative permeability ′ is the relative real permeability, and ′′ is the relative
imaginary permeability.
For a material in the presence of an applied EM field, the physical meaning for the real and
imaginary parts is as follows. The real part of the dielectric constant is a materials ability
to store energy and the latter imaginary part is the materials ability to dissipate energy. For
permittivity (), ′ is materials ability to store electrical energy and ′′ is the dielectric
loss of electrical energy to the material. Correspondingly, for permeability (), ′ is the
material’s ability to store magnetic energy and ′′ is the dielectric loss of magnetic energy
to the material. For the complex dielectric constants, the real part is related to the frequency
shift while the imaginary part is related to attenuation.
The loss tangent in the ratio of imaginary part to real part of the complex dielectric
constants used to describe a material’s ability to attenuate EM energy [141]. The electric
and magnetic loss tangents are shown by,
tan  =
′′
′
tan  =
′′
′
The relative permittivity ( ) relates angular frequency () and relaxation time () to the
applied electric field (E), shown by the Debye equations for the real and imaginary parts
of permittivity,
′ = 1 +
[ (0) − 1]
1 + ()2
′′ =
[ (0) − 1]
1 + ()2
The ability for electromagnetic energy to propagate through a medium is known as
impedance. It is a ratio between the electromagnetic field components in the frequency
35
domain. For circuit analysis, the electrical impedance () is the complex ratio measuring
the opposition of current () when voltage () is applied,
 =


Analogous of electrical impedance in circuits, this can be applied to waves traveling
through a medium. Impedance is defined as a ratio of the relative strengths of the electric
and magnetic fields. For a material, the intrinsic impedance is given by the square root of
the ratio between complex permeability and permittivity. For materials having a
conductivity equal to zero ( = 0), this is given by,
 = �


For an electromagnetic plane wave traveling in free space, impedance (0 ) is equal to a
value of 120π Ω. As many materials have a conductivity greater than zero ( > 0), the
equation for impedance is now take the form,

 = �
 + 
By using a techniques that measure , the MW absorption properties of a material can be
optimized by measuring the reflection loss of energy from the surface and calculating the
transmission of energy into the material. Some  measurement techniques for a certain
band width in the MW frequency range are an Auto Balance LCR (10 Hz – 10 MHz), Gain
Phase Analyzer (100 Hz – 40 MHz), Current-Voltage Probe (1 MHz- 1.86 GHz), and a
Network Analyzer (45 MHz – 110 GHz).
In dielectric materials, an applied electromagnetic field puts charge carriers in motion
which is known as the dielectric response. For bound charge carriers, they are localized to
a site and only able to move within a finite volume. On the other hand, free charge carriers
are able to move about a material relatively unrestricted. In terms of dielectric response,
electronic displacement is the motion of bound charge carriers and conduction is the motion
of free charge carriers.
36
The dielectric response mechanisms in a material are electronic polarization, atomic
polarization, dipole polarization, interfacial polarization and ionic conduction. These are
interrelated at different quantum levels by micro-to-macroscopic dielectric quantities.
Under applied electromagnetic field, the electronic ( ) displacement is,
 = 0  
The relative permittivity ( ) is related to the electric susceptibility ( ) by,
 = 1 + 
With the electric susceptibility ( ) given by,

0
 =
Where  is number of molecules, 0 is the permittivity of free space, and  is the
electronic polarizability. The electronic susceptibility ( ) and polarization (P) are related
by,
 =  0 
The polarization (P) vector defined as,
 = 
with  being the average dipole moment per molecule,
 =  
And the electronic polarizability is given as,
 2  2
 = () =

Here,  is number of electrons orbiting the nucleus of the atom,  is the charge of an
electron, x is the distance from the center of the charge to the nucleus, and β is constant. In
summary, the bulk electronic displacement (D) is related to the microscopic electronic
polarizability by,
37
 = 0   = 0 (1 +  ) = 0  + 
The dielectric response relates the electromagnetic frequency to the motion of electronic
dipole moments. When listed in descending order of largest to smallest, the various
dielectric polarization mechanisms that function at discrete quantum energy levels are:
Relative permittivity (ε)
↓
Dielectric susceptibility (χ)
↓
Macroscopic polarizability (P)
↓
Microscopic polarizability (α)
The four basic types of electronic polarization mechanisms in materials: electronic, ionic,
dipolar and interfacial are dielectrically responsive in electromagnetic frequency range of
1015-1010 Hz. Electronic polarization is distortion of the electronic charge distribution
relative to the nucleus of an atom (1015 Hz) i.e. orbital electrons. Ionic polarization is
displacement of ions from their relative position of equilibrium in the presence of an
electric field i.e. atomic site hop (1013 Hz). Dipolar polarization is torque on permanent
dipoles (1011 Hz) i.e. magnetic dipoles. Interfacial polarization is due to macroscopic
displacement of charge carriers (1010 Hz) i.e. grain boundaries. Total polarization is the
sum of electronic, atomic, dipolar, and interfacial polarization mechanisms [142],
 =  +  +  + 
Note the scale of polarization mechanism to the corresponding electromagnetic frequency
that it will attenuate. The various polarization mechanisms function near resonant
frequencies by short-range motion of charge carriers [143]. Each polarization mechanism
has a characteristic resonance frequency (fr) where it will become increasingly responsive
to EM energy until reaching maximum attenuation. Typically, multiple fr signals are
observed within a relatively close bandwidth with only one corresponding to the maximum
attenuation [144]. At MW frequencies used for heating, dipolar and interfacial polarization
mechanisms the most important for generating thermal energy.
38
1.10.2 Microwave-Material Heating
Microwave heating involves the transmission and absorption of energy into a material by
molecular dipoles. MW energy induces torque on dipoles that causes reorientation but due
to the inherent resistance of dipoles, they are unable to follow hysteresis loop and “lag”
with the rapidly alternating electromagnetic field which causes heating during frictional
losses. The following is a thermodynamic interpretation of MW heating.
MW heating is a highly energetic electro-physical process that increases the overall internal
energy (U) of a material or system. Internal energy is expressed as,
 =  − 
When MW energy is absorbed by various dielectric polarization mechanisms, there is a
large increase in the vibration modes of atomic particles resulting in atomic collisions and
the production of phonons. This influx of phonons results in a non-equilibrium increase of
internal energy. The incremental change in internal energy () with temperature () is
shown by the molar heat capacity term at constant volume,
So the change in thermal entropy is,
 = �

�
 
∆ = �



Thus, MW energy causes lattice vibrations that produce phonons which increases ∆ of a
material [145]. Einstein proposed a model to describe  in which a solid is composed of a
finite amount of atoms, each having 3 degrees of freedom. With these assumptions, the
molar heat capacity at constant volume can be expressed as,
 2
 ⁄
 = 3  � � � ⁄
�
(

− 1)2
In which the polarization term () is defined as,
39
Where, 
ℎ

is Avogadro’s number,  is Boltzman constant, ℎ is Planck’s constant, and 
 =
is the oscillation frequency of the material. It can be seen that maximum polarization will
occur at the resonance frequency resulting in the greatest increase of ∆ for the solid.
MW energy induces torque on molecular dipole moments within the material but due to
restricted rotational motion, dipoles are unable to follow hysteresis of the rapidly
alternating EM field. Due to phase lag occurring between reoriented dipole moments and
the field, known as dielectric relaxation, a portion of energy is absorbed and converted into
heat.
Heat generated by MW irradiation is the effect of energy absorbed into a medium which is
also known as power loss. For a plane wave incident to the surface of a medium, the power
flow is (volts per meter) × (ampres per meter). Conversion of energy to a power flux
(watts per square meter) is approximated by the Poynting Vector Theorem. When
integrated across the surface of a volume this becomes,
 =
1
�  ∗ ×  ∗ ∙ 
2
Where the term  ∗ ×  ∗ is the Poynting vector and * denotes the complex conjugates of
the  and  fields. By using the Divergence Theorem, Maxwell’s equations, and dielectric
properties, the real portion of the Poynting vector can be obtained, given as,
1
 = � 2( "  ∙  ∗ +  "  ∙  ∗ )
2
Where, 2 is the angular frequency,  " is the imaginary part of the complex permittivity,
and  " is the imaginary part of the complex permeability. Assuming uniform
electromagnetic field within a finite volume, the following equation for power absorption
density ( ) is,
 = �0 " ||2 + 0 " ||2 �
40
Where,  is the power absorption density (W/m³),  is the frequency (Hz),  is the
magnitude of electric field strength (V/m), and  is the magnitude of magnetic field
strength (A/m). It can be seen MW power absorption depends mainly on the frequency of
wave, a materials dielectric properties, and the internal electromagnetic field strength
within the material volume or working load. This is an oversimplification with other factors
such as load composition, volume, geometry contributing to the amount of heating as well
as the design of the MW heating system [146].
The heating rate of a material is proportional to the amount energy absorbed into the
material. Using the derived equation for power density, the incremental change in
temperature with time can be expressed by the following,
 �0 " ||2 + 0 " ||2 �
=


Where,  is material density and  is the molar heat capacity at constant volume. Note
this equation does not account for changes in the dielectric properties during heating.
Penetration depth ( ) is a used to describe the decay of MW power as it is transmitted
into a material. It is defined as the distance from the surface to the interior of a medium
where power drops to a value of e-1. Here, the equation for  considers both permittivity
and permeability of a material where λi is the incident EM wavelength [147].
 =
λ
2√2
�′′ ′′
−
′ ′
1
−
+ [(′ ′ )2 + (′′ ′′ )2 + (′ ′′ )2 + (′′ ′ )2 ]2 �
1
2
From the above equation, it can be seen that the largest  are at low frequencies becoming
increasingly smaller at higher frequencies. For MW heating applications, having a large
depth  will result in less power loss occurring within a medium [148]. At microwave
frequencies,  can be on the order of meters to millimeters depending on the incident
wavelength (frequency), dielectric properties of the medium, but is also largely dependent
on operating conditions and set-up of the MW system. For most MW heating applications,
domestic systems operate at 2.45GHz and industrial systems operate at 915 GHz.
41
Thermal runaway is a MW heating phenomenon where an exponential temperature rise is
observed upon reaching a critical temperature. This is because the rate of energy input far
exceeds the rate of energy loss into the medium. Physically, thermal runaway is understood
as unrestricted dipole motion in the presence of a MW field resulting in a very rapid,
uncontrolled heating (Figure 1.6). Decompostion and melting has been reported in Fe/Soda
Lime glasses due to extremly high tempertures generated as a result of thermal runaway.
Thus, controlling the power distribution within the load is necessary to avoid an event like
thermal runaway.
Figure 1.7 Random alignments of magnetic dipoles upon reaching Curie temperature (Tc).
Experimentally, thermal runaway is observed in ferromagnetic materials after exceeding
the Curie temperature (TCurie). For ferrimagnetic materials this critical temperature is
known as the Neel Temperature (TNeel) but has not been reported in literature as being
associated with the event of thermal runaway. Thus, TNeel might be considered in
prospective MW heating processes, most notably with magnetic minerals.
Material defects will act as physical oscillation mechanisms by attenuation of MW energy
[149]. The origin of the dielectric loss tangent is the anharmonic term of the crystals
potential energy, including defects, which increase the loss tangent (tan δ) [150]. These
include lattice defects such as grain boundaries, voids, dislocations, point defects, and
substitutional ions. Defects are distributed in a way that breaks the periodic arrangement
42
of charges over the long-range order of the crystal, with disordered charge defects have the
most influence on tan δ. For ceramic materials, tan δ has a linear dependence on frequency,
so as the concentration of defects increases so does tan δ.
The rate kinetics of chemical and physical reactions can be improved with microwave field
exposure but the reasons fully not understood. The accelerated reaction rate during MW
exposure was explained using numerous theories such as superheating by MW irradiation
that does not reflect standard reaction conditions, reduction in the activation energy from
non-thermal MW effects, improved mass transport from large temperature gradients, and
enhanced diffusion of ions/molecules that couple with MW energy.
Non-thermal MW effects have reported by numerous authors which include enhanced
chemical reactions, alternate reaction paths, and lowering of activation energy [151].
Although many theories have been postulated for the reason these MW effects exist but the
origin of these effects still remain unclear [152]. One such effect is the MW pondermotive
force in which MW-excitement of ionic currents provides an additional driving force for
diffusion and ion exchange other the thermal or chemical gradients. The efficiency of the
pondermotive force on mass transport depends on the frequency of radiation and the ionic
conductivity of the crystal.
When chemical bonds become broken by thermal agitation, there is an increase in the
concentration of ions that are able to interact with the electromagnetic field which can be
illustrated by the expression for electrochemical potential [153]:
 = ∇( +   +  )
Where the ionic force (Fi) produced by the potential gradient ( ∇ ) is the sum of the
chemical (μi) electro (qiø), and magneto (miφ) potential terms. For a material subjected to
electromagnetic fields, an electrostatic force (ø) is exerted on an electrically charged ion
(qi) and a magnetostatic force (φ) is exerted on an ion with a magnetic moment (mi).
43
Chemical bonds restrict the motion of atoms but as thermal vibrations in a material become
rapid between atoms, but when bonds are excessively strained from thermal vibrations,
they break. An atom free from bondage becomes a ion, now having an increase in free
energy, will diffuse by local charge imbalance. As bonds become increasingly broken,
more ions are available for atomic migration. The flux of charged ions in response to an
applied electromagnetic field can be shown as [154].
 =
 2
∇

It can be seen by the above equation that thermal, chemical, and electro-potential gradients
all contribute to the total ionic diffusion. This also demonstrates that the electro-potential
gradient is a strong driving force for mass transfer of a system [155]. Thus, MW energy
activates diffusional processes by molecular dipole coupling causing atomic collisions that
producing internal friction and heat, also known as irradiation [156].
Some bulk properties that affect the MW heating characteristics are sample composition,
particle morphology, and volume fraction of phases. So for a multi-component material
containing different phases, variations in the dielectric properties of phases will affect the
overall amount of MW absorption [157]. A susceptor is a material that readily absorbs MW
energy and can be used as an additive to heat MW transparent phases.
MW heating equipment consists of three main components: the source, the transmission
line, and the applicator [158]. The source generates MW radiation by acceleration of
electrons in the presence of an external magnetic field and oscillation of the
electromagnetic field that they produce. The source is coupled with a transmission line,
such as a wave guide, coaxial cable or strip line, where MW energy is delivered to the
applicator where it is transferred into the sample load.
MW energy is generated by a device called a magnetron, which is called a cross-field
device because, during operation, it employs both electric and magnetic fields produced
44
orthogonal to each other so the fields “cross”. The device works by accelerating electrons
emitted from a hot cathode to an anode with a high potential difference. In the presence of
an external magnetic field, a force is induced on the electrons making them travel in a spiral
motion. The collective electrons are subjected to oscillations which produces
electromagnetic energy at the microwave frequency. The frequency emitted is dependent
on the size of resonant cavities in the magnetron.
A mode is the E and H pattern of a wave propagating in the z-direction. Modes of
propagation can be Transverse Electric (TE), Transverse Magnetic (TM), Transverse
Electric Magnetic (TEM) and Hybrid (HE). With TE mode, there is zero electric field
intensity in the direction of wave propagation (Ez = 0, Hz ≠ 0). With TM mode, there is
zero magnetic field intensity in the direction of wave propagation (Ez ≠ 0, Hz = 0). With
TEM mode, there is both zero electric and magnetic field intensity in the direction of wave
propagation (Ez = 0, Hz = 0). For HE mode, there are non-zero values for the E and H field
values (Ez ≠ 0, Hz ≠ 0). For a MW heating system, modes describe the distribution of the
electric and magnetic field components that vary cyclically within the applicator.
An applicator is a type of cavity resonator used for oscillations of electromagnetic energy
[159]. There are single mode applicators and multi-mode applicators distinguished by
discrete Eigen values (modes) that the cavity resonator operates at. For MW heating
systems, the applicator transfers MW energy to a material via transmission line with the
cavity often having a rectangular or cylindrical geometry. The applicator is designed so it
functions at the desired resonant mode(s) accomplished by solving Maxwell’s equations
with the proper boundary conditions for the given cavity dimensions and geometry [160].
For a typical multi-mode, rectangular MW applicator, the resonant frequency mode
equation in Cartesian coordinates is,

 2
 2
 2
=  �� � + � � + � �
2 �
2 �
2 ̂
45
Where,  is the resonant frequency in TE or TM mode, c is the speed of light, , , 
are the half-sinusoidal variations in the standing wave pattern, , , and  are the
dimensions of the applicator, directed along the respective x, y, and z-axes.
For a multi-mode applicator, the MW energy has multiple electromagnetic wave patterns.
For a multi-mode applicator to function properly, the dimensions must be several half
wavelengths long in two orthogonal directions in order to meet the boundary conditions
that satisfy Maxwell’s equations. For single mode MW applicators, the applicator width is
= λ/2. Minimum MW energy occurs at the nodes or at integer values of and λ the maximum
MW energy occurs at the anti-nodes or at integer values of λ/2. At 2.45 GHz, the first antinode corresponding to maximum energy occurs at λ/2 = 6.12 cm.
There are pros and cons to the two different MW applicators. The advantages to a multimode MW heating system uniform MW distribution and large batch sizes while the
disadvantages are low MW intensity, difficulty heating low-loss materials. The advantages
of a single-mode MW heating system is high MW intensity in the node area and separation
of E-H fields while the disadvantages are non-uniform MW distribution and limited sample
size. Each type of applicator is dependent on the process for which it is used. Generally,
the design of the MW applicator is specific to the heating application so issues that would
need to be addressed are power efficiency, power distribution, and process conditions.
When microwaves heat a material, energy is transferred into the material causing an inverse
temperature gradient within the body of a material. The interior of a material is generally
hotter than the exterior because heat is radiated from the surface and lost the surrounding
environment. Due to inverse temperature effects, quantitative measurements of
temperature are at best inaccurate and may not reflect the actual amount of heat evolved
within a material.
During MW heating, precise temperature measurements are difficult to make which are
needed for understanding the processing conditions. There are two practical ways of doing
46
this: using a thermocouple or infrared pyrometer. When using a thermocouple, it is difficult
to take continuous measurements because the MW-field interacts with the device so, for
experiments, temperature is usually measured after heating when power is turned off. To
avoid the MW field interactions, a sheathed type-K thermocouple is used for continuous
temperature measurement by periodically inserting it thru the top of the applicator to the
sample load. Infrared pyrometers have the advantage of taking temperature measurements
continuously during an experiment but only measure heat radiated from the surface of the
load and give no information of the heat produced within the interior of the sample, where
temperatures are highest during MW heating.
1.10.3 Microwave-Mineral Processing
MW heating of materials is gaining increasing interest for various scientific and industrial
applications [161], [162]. The ability to transmit and absorb MW energy into a medium
offers enhanced heat and mass transfer mechanisms allowing for innovations in hightemperature material processing [163]. Studies of MW heating show promising results for
many types of applications but fundamental data of material properties that determine the
MW interactions [164], [165]. The ongoing research of MW heating applications will need
to address many technical issues before industrial scale-up of this technology.
At high tempertures, many minerals and metals undergo various types of physical and
chemical transformations which alter the dielectric properties which influence the amount
of MW absorption. Thus, high-temperature MW processing for metallurgical applications
can be affected by the thermodynamic stability of phases and subsequent formation of new
phases. Therefore, it will be essential for process control to understand changes in the
dielectric properties of metal compounds that result from phase transformations and
dielectric transitions during high temperature MW processing [166].
In extractive metallurgy, MW technology has been investigated for improving metal
recovery at various steps of beneficiation circuit [167]. MW heating for application in
47
mineral processing has been studied for pretreatment of ore material for improved
communition behavior, flotation properties, leaching kinetics, and smelting performance,
[168], [169]. MW heating of reagent-grade chemical and mineral compounds has been
investigated to determine heating rates power levels [170]. MW energy has also been used
as an environmental management tool for waste remediation, which is associated with any
metallurgical process [171].
Due to high energy costs, it is always of interest to find methods that increase communition
efficiency for reducing costs of mineral recovery. The U.S. Department of Energy has
invested a great deal of resources into its Industrial Technologies Programme, Mining
Industry of the Future, to plan ways of increasing energy efficiency for certain sectors of
the mining industry. Improvements to mineral processing technologies would provide
benefits regarding energy, environment, productivity, health and safety, resource base
[172]. In theory, microwave heating can contribute to all these sectors of industry.
For industrial use of MW technologies in mineral processing, it was suggested that research
should be directed at minimizing the energy input needed to achieve the desired heating
effects, optimizing MW frequencies used to achieve maximum heating efficiencies, data
concerning material dielectric properties varying with temperature, and equipment scaleup able to handle material at a tonnage per hour rate.
During mineral beneficiation, after ore is mined, the next step is communition which is the
size reduction of ore material accomplished by crushing and grinding. Initially, course
sized material is stage crushed to sizes amendable for subsequent grinding which reduces
material to fine particle size(s). Crushing and grinding consumes 50-70% of the energy
required for mineral beneficiation making it the most energy intensive step of the process
[173]. Therefore, finding methods that decrease the energy input of communition, even
slightly, can substantially reduce production costs.
48
The communition behavior of ore material can be altered by producing thermo-mechanical
stress within particles. Also known as thermally assisted fracturing, it has been used for
centuries to weaken host rock for help liberating valuable minerals. Today, using any type
of conventional heating for thermally assisted communition is extremely uneconomical and
cannot been implemented in large scale mining efforts. Therefore, MW pretreatment of
various ore types as a communition aid has been the subject of numerous studies in effort
to remedy the high energy costs of thermal pretreatment when using conventional types of
heating [174], [175].
During exposure to MW energy, the differential expansion of susceptible phases causes
thermal-mechanical stress within the host rock matrix. The stress is relieved by cracking
amongst mineral grains occurring as transgranular and intergranular fractures [176]. The
fractures created within ore particles by MW irradiation increasing the material’s
grindability during milling which ultimately decreases the Bond work index [177]. Further,
the liberation of minerals by MW heating has good implications of increasing the grade of
an ore concentrate by aiding with various physical and chemical seperation techniques
commonly used at processing facilties.
Short-pulsed MW treatment of ore particles was shown to cause breakage but at a fraction
of the energy used during traditional MW heating [178]. Irradiated minerals were reported
to show realtivly good breakage charactersitics during short-pulsed MW treatment
compared to samples MW heated continuously for long treatment times concluding that
continuous MW heating is an unneassarry waste of energy when thermally assisted
breakage is the objective. Computer modeling and simulations studying the effect of
particle breakage provided insight into the mechaisms of fracture [179]. It is now reconized
that short exposure time using pulsed, high power MW energy is the best treatment type
for thermally assited communition of ore material.
MW heating has been used to alter the surfaces of ore material to improve the flotation
performance of minerals [180]. This works by oxidizing the surface to an oxide phase that
49
allows collectors to perform more effectively [181]. Also, an increase in the surface area
from cracks intiated during MW heated particles allows for better surface adhesion of
floculants. A huge driving factor for improved flotation performance is that MW energy
can improve the kinetics of the flotation cell.
MW treatment of coal has received considerable attention with studies involving heating
rate, desulfurization, and grindability. The grade of coal determines its industrial purpose
ranked by the content of water, organic volatiles, sulfur, and ash minerals which are
indicators of its environmental impact when incinerated. Prior to combustion, raw coal
requires grinding to attain particle sizes that facilitate complete combustion of product so
MW pretreatment is used to aid with the communition [182].
Sulfur in coal is present in two types of forms: organic and pyritic. These cause
environmental damage upon combustion with the formation of SOx gases. The mineral
pyrite (FeS2) is part of a family of iron sulfides that readily absorbs MW energy.
Differences between the relative dielectric properties ( and ) of pyrite and coal allows
for greater absorption of MW energy and faster heating of pyrite then the surrounding coal
matrix. Some of the ash minerals absorb MW energy too, likely attributed to their small
grain size and aluminum (Al) content. Also, since water molecules are inherently polar,
they readily absorb MW energy which creates internal pressure that weakens the coal
matrix. Thus, MW heating and weakening of the coal matrix was reportedly made possible
by the inherent moisture content and thermo-expansion of lossy minerals, mainly that of
pyrite [183].
When pyrite is exposed to MW energy, it partly converts to pyrrhotite (Fe1-xS), a
ferrimagnetic mineral and/or the surface converts to an iron oxide (Fe1-xO) with either
phase change enhancing magnetic susceptibility. MW exposure in open atmosphere causes
pyrite (FeS2) to oxidize rendering pyrrhotite Fe(1-x)S, triolite (FeS), α- hematite (Fe2O3) and
γ-hematite (Fe2O3) [184]. These newly formed phases on the pyrite surface makes it
possible to remove using low-intensity mangetic seperation [185].
50
Increasing the MW power wattage allows for higher tempertures attained during heating.
Sulfide oxidation at the surface-air interface is an exothermic reaction producing heat
during the reaction. Pyrite particles showed a sharp increase in the magnetic recovery
product which was attributed to more magnetic phase formed on the particle and realigment of atoms resulting in a more ordered structure of the lattice [186]. Decompostion
of pyrite to more magnetic iron sulfide phases follows the reaction [187].
FeS2 → Fe1-xS → FeS
After MW pretreatment, low intensity magnetic separation can be used to remove iron
sulfide from coal for desulfurization purposes [188]. Other studies have shown that the
Bond Work Index is reduced by MW pretreatment of coal from thermal-mechanical
stresses generated by pyrite and superheated water within the coal matrix. Also, the oxide
phases are brittle under an applied load [189]. Thus, as long as carbonaceous material is
used as a major industrial fuel source, MW pretreatment of coal could potentially
demonstrate significant environmental benefits.
Advances in MW steelmaking have required proper characterization of the dielectric
properties for all raw materials used in the process. This technology is highly sought after
for its high efficiency in which reduction of ore material is quickly achieved by irradiation
without significant amounts of energy lost as heat as with blast furnaces used in today’s
steel making industry. Proponents of MW steelmaking claim it’s a purer, greener, cheaper
product. One major environmental incentive for MW reduction of iron is a decrease in CO2
production compared with conventional process, shown below by the thermo-chemical
equations,
Conventional
2Fe3 O4 + 9C + 5O2 → 6Fe + 9CO2
Microwave
2Fe3 O4 + 4C → 6Fe + 4CO2
During reduction of iron oxide by conventional process, additional carbon (C) is needed to
support the reaction 2C + O2 = 2CO, which supplies exothermic heat and produces CO
reducing gas. Since magnetite and carbon are strong absorbers of MW energy, both
51
materials are supplied with enough thermal energy for reduction to occur without the need
for additional carbon to react with oxygen in the surrounding environment [190].
Carbon is a good absorber of MW energy because it can be easily polarized due to the
directional nature of covalent bonds that link atoms. Carbon for 2-D and 3-D periodic
arrangements that have high surface area to volume ratios from voids between carbon
structures and layers which supplies free-moving surface electrons (non-bulk) to the
conduction band making it an excellent conductor of electricity. The MW field creates
numerous particle collisions per volume, and along with excellent thermal conductivity,
allows for excellent MW heating properties. Due to a high concentration of large voids,
there is excellent transmission of MW energy which allows for high MW penetration depth
(Dp) into carbonaceous matter. Thus, reduction of various metal oxides using carbon in the
form of coal has been the subject of studies not only due shear abundance but also for its
quality MW absorbing properties [190].
One difficulty of the MW reduction of iron is that various iron oxide phases subsequently
form during the reduction process. As oxygen is driven off by reducing agent (not shown),
a different chemical species with unique dielectric properties, and MW absorption
capabilities, will subsequently forms:
Fe2O3 → Fe3O4 → FeO → FeO1-x → Fe
As reduction of iron progresses, MW susceptibility of an oxide phase changes, therefore,
so does the quality of MW absorption and effectiveness of heating. Therefore, other heating
methods might be employed to complete the reduction of iron. The quality of MW
absorption with iron oxide phases during reduction might be remedied with proper
dielectric characterization iron oxide phases at high temperature.
Hydrometallurgy will become more significant in the future for its environmental,
economical and technical attributes [191]. It produces less SO2 gas emissions than with
conventional pyrometallurgical methods and metals can be recovered directly from
solution. The problems of hydro-met processes are long processing times for high recovery
52
of metals, difficulties in solid-liquid separation, and the effect impurities have on refining.
Thus, MW-assisted leaching has been employed as a pre-treatment of ore material to
initiate cracking for increased rock permeability and wetting of chemical leachant on newly
formed surfaces or MW energy is used to directly irradiate the leaching system to cause
both heating and agitation of the media. Many different materials have been subject to
studies of MW assisted leaching conditions for metal extraction of Ni, Cu, Co, Zn, Pb, and
Mn [192], [193], [194], [195].
MW-assisted hydrometallurgy has been investigated for prevention of industrial pollutants,
enhanced precipitation of metals from solution, and improving metal yield after leaching.
MWs are reported to improve the rate kinetics of reactions resulting in faster leaching times
compared to conventional methods. Improved reaction kinetics were explained by: cracks
initiated during MW pre-treatment of material that increase the wetability, the existence of
non-thermal MW effects that reduce activation energy, a super-heating effect that makes
the temperature no longer representative of the reaction conditions, large thermal gradients
that assists with mass transport between the solid-liquid interface, and diffusion of ions in
solution interacting with the MW-field. Thus, a better understanding of the effect MW’s
have on a specific leaching system might allow for implementation of this technology at
an industrial scale.
Gold (Au) is often difficult to recover because it is finely disseminated within host rock
therefore MW pretreatment of refractory gold ores has been investigated. Generally, gold
is type-classified occurring as free milling gold or refractory gold [196]. There are two
types of refractory gold: (1) locked gold ore and (2) reactive gangue ore. With locked gold
ore, the gold is either in solid solution or enclosed by metal-sulfide minerals, mainly pyrite
(FeS2) and arsonopyrite (AsFeS2). Reactive gangue ore can be further classified as pregrobbing carbonaceous type that absorbs gold from leach solution and/or ores containing
minerals that consume large quantities of leachant. An ore is classified as double refractory
if is composed of metal sulfide minerals and carbonaceous material.
53
Most commonly, extraction of gold is accomplished by leaching with sodium cyanide
solution [197]. Pretreatment of refractory ore gold ore makes it is amendable to leaching
but increases the energy input and cost. Methods such as high temperature oxidation, high
pressure leaching or biochemical oxidation have been investigated for improving gold
recovery but are energy intensive, time consuming, or impractical. MW energy is a good
candidate for pre-treatment of gold ore because gold-bearing metal-sulfide minerals are
highly reactive to MW energy [198].
Pyrite and arsenopyrite heated quickly by MW irradiation and reach temperatures for the
oxidation reaction to occur with the emission of sulfur gas and weight loss samples. If
As2O3, SO2 completely volatilize, the remaining product is Fe2O3. Subsequent cyanide
leaching of this product recovered 98% of the gold [199]. When this same product was
mixed sodium hydroxide and exposed to MW energy, 99% of the gold was recovered.
The mineral chalcopyrite (CuFeS2) is an important ore of copper and has been subject to
many MW heating studies from alter is magnetic properties to floatation to leaching of
material for extraction of copper. The chemical activity of ferrous and ferric iron increases
during MW exposure [200]. This is an interesting finding because the most important
oxidant during smelting of chalcopyrite is ferric iron. MW enhanced roasting copper
sulfide concentrate in the presence CaCO3 followed by leaching of the calcine product
showed increased metal recovery when compared to non-treated samples [201]. MW
assisted oxygenated leaching of mineral slurries consisting of copper concentrates mixed
in NaCl-HCl solutions have been investigated for the extraction of copper [202].
For highly electrically conductive materials, such as elemental metals, there is very little
MW penetration depth because much of the energy is reflected at the surface by eddy
currents. For this reason, MW heating systems are designed with smooth metal surfaces
that reflect radiation so it can be transmitted into the load-material for heating. Though
bulk metal surfaces reflect MW energy, powder metal compacts will heat with MW energy
because the penetration depth of MW energy is sufficient compared to the size of individual
54
particles. This creates high power density within individual particles that allows for both
solid and liquid state sintering of powder metal compacts.
The bulk density of a powder compact is low compared to the theoretical density because
the material volume contains a large void fraction. These voids are filled with air which is
an excellent medium for transmitting electromagnetic radiation. Thus, MW energy can be
transmitted throughout the bulk material compact with relatively uniform power density.
MW heating of material concentrates in the form of powder concentrates has been used in
a variety of metallurgical processes such as sintering and smelting [203]. MW sintering of
ceramic alloys and metal-ceramic composite materials has shown that high quality sintered
components can be made with a fully automated industrial process [204]. The ability to
attain high temperatures within short time duration creates unique microstructure features
unlike those of conventional heating. This is likely due to the non-equilibrium
thermodynamics of MW heating during which diffusive properties are enhanced,
accompanied by a rapid temperature increase, creates phases that cannot chemically
equilibrate as during conventional heating.
At standard tempertures and pressures many materials are poor MW absorbers but
laboratory heating experiments have shown that certain materials “activate” at elevated
tempertures and readily absorb MW energy due to changes in their dielectric properties.
This has been shown with alumina (Al2O3) where samples are conventionally heated prior
to MW exposure.
A susceptor can be used to aid with heating a material that has poor MW absorption
properties. Silicon Carbide (SiC) is an example of chemically stable susceptor at high
temperatures that is used to jump start MW heating reactions. Magnetite is also a good MW
susceptor material but is more chemically reactive its environment.
55
MW plasma(s) can be created during irradiation of highly reactive materials generating
very high local temperatures. The reason for plasma generation is not fully understood and
experiments are often terminated when this phenomenon is observed. For a gaseous
plasma, the dielectric constant is approximated by the equation [37],

 = 0 �1 − �

Where,  is the plasma permittivity, 0 is the permittivity of free space,  is the plasma
frequency, and  is the angular frequency of the incident radiation.
The creation of plasma has been used for material processing applications. MW plasma
processing has been used for chemical vapor deposition and similar processes involving
layering gaseous metals. Adversely, during MW smelting of rare earth metals, gaseous
plasma formed over the top of the melt hindering the reduction process.
56
Chapter 2: Experimental Procedure
2.1 Ore Particle Characterization
2.1.1 Imaging of Samples
To image ore particles, samples were mounted in epoxy, polished to 0.06μm finish, and
examined under optical and scanning electron microscope (SEM). A Nikon polarized
microscope with rotating stage was used to observe birefringence of minerals on the surface
of the sample. SEM samples were carbon-coated for a conductive surface and viewed with
25 kV accelerating voltage using a Jeol 6400 SEM (Appendix 1.1). Secondary electron
image (SEI) mode was used to view topological features of unpolished specimens, such as
oxidation products and melting. Then, backscattered electron mode (BSE) in collaboration
with electron dispersive spectroscopy (EDS) was used for phase identification.
2.1.2 Magnetic Susceptibility Measurements
Magnetic susceptibility experiments were performed by the Metallurgy Department of the
Indian Institute of Technology (IITK) Kanpur, India. The equipment (Appendix 1.2) made
available is normally used for Curie temperature (TC) measurements of metals and
ceramics able to test small samples (<2g). To obtain samples, large ore particles were
crushed into smaller pieces and particles with metallic appearance were hand-picked for
testing. Using conventional heating, magnetic susceptibility measurements were taken in
5 °C increments up to 400 °C as indicated by a thermal couple in contact with the sample.
Once a rapid decrease in the magnetic susceptibility of samples was observed experiments
were terminated. The samples used for TC measurements was examined with x-ray powder
diffraction (XRD) to probe for phases present.
2.2 Crushing and Grinding Experiments
2.2.1 Obtaining Ore Particle Samples for Grinding Experiments
The material used for crushing and grinding experiments was obtained by stage crushing
core samples extracted from the Eagle metal-sulfide deposit. The core samples, taken from
57
various locations in the deposit, were initially jaw crushed with the jaws spaced ≈ ½” apart.
The crushed output material was choke feed through the jaw crusher again in an attempt to
increase the amount of size reduction in larger size pieces. The total amount of jaw crushed
material was approximately 500lbs. The jaw crushed samples were size classified
successively from course size sieves (2”, 5/8”, 1/2”) to medium size sieves (7/16”, 3/8”,
5/16”) to remove any large particles. The remaining material was sieved using the 6, 8, 10,
and 12 Tyler mesh sieves to obtain particles passing 6 mesh. For clarity, size fractions are
represented by passing (-) and retained (+) material.
Any remaining oversize material (> 5/16”) was put through the laboratory gyratory crusher
for further size reduction. The gyratory crushed material was then classified using 4, 6, 8,
and 12 Tyler mesh sieves that produced gyratory crushed samples in size ranges of (-4+6),
(-6+8), and (-8+12). Particles in all size ranges were run through a particle splitter a dozen
times and recombined to ensure homogeneity of the samples.
2.2.2 Microwave Heating of Ore Particles
A conventional, multi-mode, 1000 watt (2.45 GHz) microwave with a 1.1 ft³ cavity was
used for irradiation of samples. Ore particles of various size ranges were massed then
exposed to microwave energy for various times. Observations of the ore particles were
made during and after MW exposure. For jaw (-6+10) and jaw (-8+12) samples,
approximately (~500) gram samples directly on the glass tray were exposed for 60s and
30s, respectively (Appendix 1.3.). After MW exposure, particles are allowed to cool before
any other treatment is performed.
For gyratory crushed samples, approximately (~100) gram samples contained in an
alumina crucible were placed on a square piece of insulating ceramic located in the center
of the rotating glass tray (Appendix 1.4.). Samples were exposed to MW energy for 30s,
and also 60s, to obtain +3000 grams of material that was used for ball milling. The gyratory
samples were massed before and after MW heating to determine weight loss due to
oxidation. Particle samples are allowed to cool after treatment (Appendix 1.5).
58
MW power (1000 W) and frequency (2.45 GHz) are fixed during experiments so treatment
time, sample volume, and particle size are the three parameters that affect the amount of
MW absorption and resulting heating properties of a bulk particle sample.
Figure 2.1 Experimental parameters that affect microwave absorption in samples.
2.2.3 Preliminary Grinding Experiments of Ore Particles
Using the F. Bond procedure for grinding, 1000 grams of material was ball milled dry for
100 revolutions, dumped and screened with multiple sieve sizes. Particle size classification
was performed using a Ro-Tap to shake the stack of sieves. The material retained on each
sieve was massed and recorded. Before returning the partly ground material back to the
ball mill for further grinding, the charge was balanced back to 1000 grams with new feed
material to balance the grinding charge for material lost to the ball mill and/or sieves.
2.2.4 Ball Mill Grindability Experiments of Ore Particles
The laboratory ball mill (Appendix 1.6) was fitted with 285 balls of the approximate size
following Bond’s procedure for ball mill grindability experiments. As specified by Bond,
all material acceptable for grinding experiments has to pass the 6 mesh sieve size
59
(3.327mm). Then, 1000 grams (≈ 700 cm3) of dry material was ball milled, dumped and
size classified. The material retained (oversize) and passing (undersize) on the test sieve
was massed and recorded. Undersized material was removed from the charge and fresh
feed was added to balance the charge back to 1000 grams. The charge material mass,
undersize material mass, and ball-mill revolutions were used to calculate the new number
of revolutions used for the next grinding cycle. This procedure was repeated until the target
circulating load (250 %) was obtained and reversed direction. The last undersize material
was size classified using a sieve with smaller aperture openings then the one chosen for
grinding experiments to find 80% passing size of the ball milled material. For
determination of the work index (kW·h per short ton ore), the following equation was used:
Wi =
44.5
 10
10 
Pi 0.23Gbp 0.82 
−

 P
F80 
 80
Where Pi is the test sieve size in microns, GBP is the grindability of the ore, P is the size
in microns of 80 percent of the last cycle sieve undersize product passes, and F is the size
in microns of 80 percent of the new feed passes. See Appendix 4 for a description of
performing ball mill grindability experiments.
Particle size distribution of the oversize (+100) and undersize (-100) material was
measured after completion ball milling experiments. The last oversize material (≈ 700 g)
was size classified using the 6, 10, 20, 35, 48, and 65 Tyler mesh sieve sizes. The undersize
material (≈ 1000 g) was massed and size classified using the 150, 200, 270, 325, and 500
Tyler mesh sieve sizes.
2.2.5 Roll Crushing Experiments
According to F. Bonds procedure for measurement of the work index in ball mill
grindability experiments, particles unable to pass the 6 mesh sieve are considered too large
for testing. So, roll crushing and particle size classification was performed on as-received
and MW treated (-4+6) gyratory crushed material. This was accomplished by exposing the
(-4+6) particles to MW energy. Then, particle samples were passed thru a roll crusher and
60
size classified using the 20, 42, 65, 100, 200, and 325 Tyler mesh sieves. Material retained
on each sieve was massed and recorded.
2.3 Heavy Liquid Separation
Density separation was performed using the heavy liquid methelyne iodide (CH2I2) of 99+
% purity supplied by Acros Organics. The specific gravity of CH2I2 is 3.32. To start, 10
grams of material from grinding experiments is massed and put into a 50 ml centrifuge
glass tube. Then the tube is filled ~ 2/3 of the way to the top with CH2I2 and the contents is
mixed with a glass stir rod. The mixture of heavy liquid and solid material is centrifuged
for 30 min. at 1500 rpm using an International Clinical Centrifuge Model CL (Appendix
1.7). After allowing the centrifuge to halt, the float and sink are poured through separate
pieces of #1 (11μm) filter paper. The filter paper is washed with acetone to remove excess
CH2I2. The filter paper and filtrate are allowed to dry in an oven overnight to remove
acetone (CH3)2CO. All density separation experiments were contained in a fume hood and
a hot plate was used the regenerate the heavy liquid for reuse by removing the acetone.
Minerals common in the ore, composition, and specific gravities are given in Table 2.1.
Table 2.1 Mineral compositions and specific gravities of common minerals in ore.
Mineral
Pyrrhotite
Chalcopyrite
Pentlandite
Spinel
Pyroxene
Composition
Fe1-xS
CuFeS2
(Fe,Ni)9S8
X2+Y23+O4
Mg2Si2O6
Fe2Si2O6
CaMg2Si2O6
CaFe2Si2O6
Specific Gravity
4.58-4.65
4.1-4.3
4.6-5.0
3.7-5.0
3.20
4.10
3.28
3.66
Olivine
Mg2SiO4
3.23
61
Fe2SiO4
4.40
2.4 X-Ray Diffraction (XRD)
X-ray diffraction (XRD) experiments were performed using a Scintag XDS 2000 powder
diffractometer using Cu-Kα radiation (λ=1.5418Å) (Appendix 1.8). The X-ray tube has a
1mm divergence slit and a 2 mm scatter slit with the detector having a 0.5mm scatter slit
and 0.3 receiving slit. All samples were mounted in a rectangular aluminum holder and
examined with a continuous scan in the range of 10-75° to obtain diffraction patterns for
qualitative phase identification. JCPDF win computer software was used for phase
identification of the peaks obtained in diffraction patterns (JCPDS-International Centre for
Diffraction Data v. 2.2; 2001).
2.5 Inductively Coupled Plasma (ICP)
Inductively coupled plasma (I.C.P.) was performed on select samples obtained in ball
milling and roll crushing experiments. Using the fusion digestion method, samples were
milled to at least -100 mesh (147μm) and broke down in 10 % hydrochloric acid solution.
The digested samples were further diluted for I.C.P. analysis. Inductively coupled plasma
(I.C.P.) was performed on as-received samples and MW 60s samples to obtain the
analytical chemistry of heavy liquid sink material.
62
Chapter 3: Ore Particle Characterization
3.1 Imaging of Ore Samples
The geological survey was used as a guide for mineral phase identification in specimens.
Metallic bearing mineral grains within the ore were optically identified by color and
birefringence using a petrographic microscope. The optical petrographic microscope is
difficult for the examiner and is only effective in identifying phases with a skilled user. For
this reason, the petrographic microscope was used in combination with the SEM
microscope to identify mineral phases present within mounted polished samples.
Secondary electron imaging (SEI) was used to observe mineral grain morphology and
texture. Backscatter electron (BSE) mode was used in combination with electron dispersive
spectroscopy (EDS) to qualitatively identify phases contained in the ore particle samples.
3.1.1 Mineral Phase Assemblages
The mineralization consists of peridotite host rock having intergrowths of Fe-Cu-Ni
bearing sulfide minerals and accessory Fe-Cr-Ti spinel oxide minerals. Peridotite is
composed of Mg-rich olivine (Fo), and pyroxene (Px) that consists of diopside (Ca-rich)
and enstatite (Al-rich). Sulfide minerals occur as intergrowths of pyrrhotite (Po),
chalcopyrite (Ccp) and pentlandite (Pnt). Ferro-spinel (Spl) oxide minerals span a Fe-TiCr solid solution series (Mag-Chr-Usl) occurring as euhedral inclusions in both silicate and
sulfide minerals. Metallic-bearing minerals were the main focus of the study, though other
minerals were identified within ore particle samples. An assemblage of minerals can be
viewed in Figure 3.1.
63
Figure 3.1 BSE image of ore particles (Jaw -8+12 mesh). Mineral phases are pyroxene
(Px), pentlandite (Pnt), pyrrhotite (Po), chalcopyrite (Ccp) and ferrospinel (Mag).
3.1.2 Silicate Minerals
EDS of the peridotite rock shows it is composed of forsterite (Fo) (Appendix 2.1-2.2)
occurring with pyroxene silicates diopside (Ca-rich) and enstatite (Al-rich). The diopside
is a clinopyroxene (Cpx) having varying calcium (Ca) compositions (Appendix 2.3-2.4).
The enstatite (En) is an orthopyroxene having variable iron (Fe) and magnesium (Mg)
compositions (Appendix 2.5 -2.6). EDS detected titanium (Ti) occurring in Ca-silicates
near (Spl) mineral inclusions (Appendix 2.7). Other silicate minerals qualitatively
identified by EDS were quartz, feldspar and chlorite-group minerals containing sodium
(Na), potassium (K) and chlorine (Cl) (Appendix 2.8-2.10).
3.1.3 Metallic Minerals
Ferro-spinel (Spl) oxide minerals are a common occurrence in pyroxene (Px) silicates.
These are found as sub-micron inclusions of magnetite (Fe3O4), ulvöspinel (Fe2TiO4) and
chromite (FeCr2O4) with composition indicated by EDS (Appendix 3.1-3.3). The round,
globular morphology of magnetite (Mag) inclusions are seen dispersed throughout a large
grain of chalcopyrite (Ccp) shown in Figure 3.2.
64
Figure 3.2 BSE image of globular ferrospinel (Mag) inclusions in chalcopyrite (Ccp).
EDS shows Spl minerals contain mainly iron (Fe), chromium (Cr), and titanium (Ti), along
with small amounts of magnesium (Mg), aluminum (Al), and silicon (Si) (Appendix 3.43.5). A Spl inclusion containing low amounts of manganese (Mn) was also detected in a
sample (Appendix 3.6). Low energy peaks are Al, Mg, Si, and Mn while high energy peaks
are Fe, Cr, Ti, and O. The charge based chemical formula for this isomorphic phase might
be written as (Fe,Mg,Mn2+),(Al,Cr,Fe3+)2O4-(Fe,Mg,Mn2+)2(Ti,Si4+)O4 which shows
various cation substitutions. For additional BSE images of Spl inclusions, see Appendix
1.9-1.11.
The BSE image in Figure 3.3 shows a crack running through a large Spl grain caused by
exposure to MW energy. The dark grey hue of the Spl grain is thought to cause an
incomplete signal of electrons to the BSE detector which may indicate it is a strongly
magnetic phase. Partially exsolved grains of Spl within Px also can be seen in the image.
Two distinct grains of Po are pointed out by arrows in the image of Figure 3.3. EDS spectra
of both areas results with the phases having two distinct compositions. The Po grain, having
a darker contrast, has a composition of Fe 47 S 53 at. % (Appendix 3.7) and the remaining
65
grain of lighter contrasting area has a composition of Fe 34 S 66 at % (Appendix 3.8). The
iron-rich Po is in the compositional range of monoclinic 4M-type with the remaining area
being iron deficient hexagonal NC-type Po. As mentioned, all compositional variations of
Po exhibit ferrimagnetism to some extent with 4M-type having the strongest magnetic
properties amoung them. Small amounts of Ccp is present along edges of Po grains
distinguised by its bright contrast and EDS analysis showing the presents of copper (Cu)
within the phase (Appendix 3.9).
Figure 3.3 BSE image of cracking in a Ti-ferrospinel (Mag) grain. Two pyrrhotite (Po)
grains with different compostions of are located in the image by arrows.
The BSE image in Figure 3.4 shows Ni-rich phases of pentlandite (Pnt) and nickel-rich
pyrrhotite (Ni-Po) within pyrrhotite (Po). EDS of both phases shows the composition of
Pnt (Fe 18 Ni 20 S 62) at. % (Appendix 3.10) and Ni-Po (Fe 19 Ni 3 S 78) at. % (Appendix
3.11). The image in Figure 3.4 (a) shows the blocky morphology of metal-rich Pnt grains
and Ni-Po dissemenated in Po. The black circled area in Figure 3.4 (a) is shown at higher
magnification in (b) having flame-type Pnt that strems from a grain of Ni-Po also having
blocky morphology. It is possible that Ni is retained in Po if there are poor exsolution
conditions of Pnt. Additional images of Ni-Po are found in Appendicies 1.12-1.15.
66
Figure 3.4 BSE image of pentlandite (Pnt) and Ni-rich pyrrhotite (Ni-Po) grains exsolved
from pyrrhotite (Po).
The Ni-Po is not believed to be the secondary mineral violarite (Vo) with composition
FeNi2S4. Core samples are not taken from the supergene zone where alteration of Pnt to
Vo takes place. Vo is reported as a non-magnetic thiospinel mineral mineral. Though it is
not correct to evaluate the magnetic properties of a mineral simply from the hue of a BSE
image, the Ni-Po appears greyish in color just as Spl minerals do.
The two common sulfide minerals found by BSE imaging were identified by EDS as galena
(PbS) having composition of (Pb 57 S 43) at. % (Appendix 3.12) and sphalerite (ZnFeS)
having composition (Zn 24 Fe 5 S 71) at. % (Appendix 3.13). The small grains of galena
(Gn) are found disseminated in Po along with large Mag inclusions (Figure 3.5). The grains
of Gn appear to brightly contrast with its surroundings (from the presence of Pb) allowing
it to be highly distinguishable with BSE imaging and later identified by EDS (Appendix
1.16). Sphalerite (Sl) grains are found occurring aside quartz (Qtz) in Figure 3.6.
The BSE image in Figure 3.7 shows grains of the telluride mineral michenerite (Mich), a
palladium bismuth tellurium (Pd-Bi-Te) solid solution compound. The grains are small
(10-50μm) but are revealed in the BSE image of a sample with high metallic content. The
composition of Mich is (Pd 34 Bi 32 Te 34) at. % as indicated by EDS (Appendix 3.14).
67
Mich grains are dispersed throughout the Po with a collection of grains appearing near
inclusions of Magnetite (Mag). The blocky morphology of Pnt demonstrates good
exsolution conditions which might indicate geochemical enrichment of Mich grains.
Figure 3.5 BSE image of galena (Gn) inclusions in pyrrhotite (Po).
Figure 3.6 BSE image of sphalerite (Sl) inclusions near a chalcopyrite (Ccp) grain.
68
Figure 3.7 BSE image of michenerite (Mich) grains dissemenated in pyrrhotite (Po).
3.2 Magnetism in Ore Particles
3.2.1 Curie Temperature Measurement
Curie temperature (TC) measurements were performed on samples with metallic
appearance. Magnetic moment measurements were taken every 5°C until saturation was
reached. The samples shown in Figure 3.8 have magnetic moment measurements measured
in electromagnetic units (0.001 A∙m²) and are plotted as a function of temperature (°C).
The overall strength of the magnetic moments measured between Curie sample 1 and 2 are
not equal but the plot shapes are in agreement. At roughly 325°C, the magnetic moment
rapidly decreases indicating TC has been reached.
69
Figure 3.8 Curie temperature (TC) plots of metallic ore samples with magnetic moment in
electromagnetic units (0.001A∙m²) plotted as a function of temperature (ºC).
The rapid decline in magnetization is within close approximation of the known Curie
temperature for 4C-type Po (TC = 325°C). The small percentage of Ni contained in Po may
have effect on its magnetism and the TC measurement. It is thought that Ni-rich pyrrhotite
(Ni-Po) may contribute to the bulk magnetic properties of ore particles.
3.2.2 Magnetism in Nickel Pyrrhotite
XRD of Curie temperature samples produced the pattern in Figure 3.9. The raw scan shows
diffraction lines belonging to magnetic 4M-type Po (Appendix 5.1). Diffraction lines fit
well with Po planes having high D-spacing but peaks with low D-spacing are shifted right.
This may indicate that the Ni present in Po affects lattice spacing by strain relaxation. The
mineral Ni-Po is poorly characterized by XRD methods and is assumed to have similar
70
diffraction lines as 4M-type Po.
Figure 3.9 XRD pattern of the Curie temperature sample plotted as intensity (arbitrary
units) vs. 2θ (degrees). Labeled peaks are metallic-sulfides pyrrhotite (Po), chalcopyrite
(Ccp), and pentlandite (Pnt).
EDS of Ni-Po shows a metal-deficient composition comparable to Hex-type Po (Appendix
3.8 and 3.11). The origin of magnetism in Ni-Po is unclear and unreported, but it is believed
that Ni site occupation in the lattice produces a net magnetic moment due to a high Fevacancy concentration in the structure. Assuming Ni behaves like an elemental metal (Ni2+)
preferring long-range atomic ordering, clustering of Fe vacancies (VFe) around Ni atoms
would produce a net magnetic moment in Ni-Po. Another possible theory for magnetism
in Ni-Po is the small portion of nickel atoms contained in the Po structure effectively
oxidize the surrounding Fe atoms to ferric (3+) iron, a strongly a ferromagnetic cation.
If the origin of magnetism in Ni-Po occurs by the clustering of vacancies around Ni atoms
as suggested, then a non-stoichiometric structure containing a high concentration of
71
vacancies will have strain relaxation energy associated with the crystal lattice. Strain
relaxation of the lattice is on the order of a few angstroms and would only be detectible in
planes with low d-spacing. Though the atomic radii of Ni is slightly larger than Fe, the shift
in peak position to a lower d-spacing (λ=2dsinθ) could be the effect of a high vacancy
concentration that creates strain relaxation in the Po crystal lattice.
3.3 Microwave Exposure of Ore Particles
3.3.1 Observations During Microwave Exposure
The rapid, volumetric heating of samples exposed to microwave (MW) energy is thought
to be achieved mainly by strong B-field coupling of magnetic phases in ore particles. Often
during MW exposure, surfaces of particles produced sparks within the first few seconds of
irradiation (Figure 3.10). Bright white and orange sparks were emitted throughout the
duration MW exposure accompanied by white or black smoke and the production sulfur
gas during oxidation of sulfide minerals on the surface. The sparking observed was plasma
generated by MW induced electrical arcing during oxidation metallic minerals.
Figure 3.10 Electrical arcing (plasma) during MW exposure of sulfide ore particles.
3.3.2 Observations After Microwave Exposure
After exposure to MW energy, some particle samples are hot to the touch and smell slightly
72
of sulfur gas. Samples exposed for longer durations of time (> 60s) can glow orange, red,
white, and purple. It was observed that pure chalcopyrite (Ccp) samples exposed to MW
energy reacted violently. During some experiments, Ccp particles burst apart into smaller
pieces from large thermal stresses generated during irradiation (Figure 3.11).
Figure 3.11 Pieces of pure chalcopyrite (Ccp) in the MW oven cavity after bursting apart
during irradiation.
As previously reported, surfaces of metallic bearing minerals exposed to MW energy in
open atmosphere sometimes became heavily oxidized (Figure 3.12). Weight loss was
reported in samples exposed to MW energy in open atmosphere but no weight loss was
observed in samples exposed to MW energy in argon gas. Weight (mass) loss is
characteristic of sulfide oxidation.
Heavily altered, decomposed particle surfaces were observed in samples exposed to MW
energy in both open atmosphere and argon gas. Particle surfaces that reach very high local
temperatures produced a black, glassy flux-type product formed by localized melting of
phases. The images in Figure 3.13 are of unpolished surfaces displaying an incipient melt
product formed on a single ore particle exposed to MW energy for 120 seconds.
73
Figure 3.12 BSE image of a particle surface heavily oxidized after microwave exposure in
open atmosphere.
Figure 3.13 Optical and BSE image of melt-flux formed during 120s of microwave
exposure in open atmosphere.
74
3.3.3 Cracking in Ore Particles
Thermal stresses generated by differential thermal expansion properties of various lossytype phases results in cracking within samples. Some cracking appears to propagate along
grain boundaries of metallic-silicate and metallic-metallic interfaces. Intergranular and
transgranular cracking among mineral assemblages is thought to be attributed to rapid MW
absorption of magnetic phases.
Slow thermal diffusivity of heat from lossy phases to neighboring mineral grains produces
internal stresses in ore particles which is relieved by producing intergranular cracks. Small,
micron sized cracking observed in metallic grains is due high thermal-mechanical stress
developed during irridiation. Larger, sub-micron sized transgranular cracks in ore particles
are thought to be the result of thermal shock during heating and rapid cooling in air.
The polished section in Figure 3.14 shows a mineral assemblage before and after MW
exposure in air for 10s. The images do not have a scale bar but large macroscopic cracks
are visible to the naked eye. Before heating, the sample has blocky, cream colored grains
of nickel pyrrhotite (Ni-Po) surronding a greyish grain of pyrrhotite (Po) (Figure 3.14 a).
After short MW exposure (10 s) (as to not heavily alter the surface), Po grains are slightly
oxidized seen by their reddish appearance (Figure 3.14 b). The before and after images in
Figure 3.14 illutrates that Po is highly reactive to MW exposure but the Ni-Po grains do
not appear to be oxidized on the surface of this sample.
In Figure 3.14 b, a large crack runs across the field of view and another crack runs through
the top portion of the Po grain. The black circled portion of the Po grain in Figure 3.14 b
is shown at higher magnification in Figure 3.14 c. Intergranular cracking is seen throughout
the redish Po grain with cracks appearing to propagate radially from the Ni-Po inclusion
within the Po grain.
75
Figure 3.14 Optical images of sulfide mineral grains before (a) and after (b) microwave
exposure (2.45GHz, 1000W) in air for 10s. Scale bar not shown.
The backscatter electron (BSE) images in Figure 3.15 show transgranualar cracking
originating from Po grains after MW exposure. Large transgranular cracks propagate
between metallic grains through pyroxene (Px) in Figure 3.15 a. The circled portion of
Figure 3.15 (a) is viewed at higher magnification in Figure 3.15 b. It appears Ni-Po grains
76
readily absorb MW energy causing cracking within the Po grain. If Ni-Po is highly
magnetic, it would absorb MW energy quickly and possibly expand faster then the
surronding Po grain causing internal stress and cracking.
Figure 3.15 BSE image of cracking in silicate host rock and metallic sulfide minerals. The
circled portion of (a) can be viewed at higher magnification in (b).
The mineral phases imaged in Figure 3.16 a are cracked and heavily altered from MW
exposure. The highly absoptive magentic Spl grain (circled) is liberated from its
surrondings seen by cracking along the grain boundary but also formed cracks within itself.
The appearance and composition of the Spl inclusion before heating is unknown but EDS
shows it to be a Cr-Ti-Fe bearing oxide. It is thought that high tempertures generated by
MW irridiation altered the grains initial composition because it now contains oxide phases
different from the original composition of Spl inclusions when comparing EDS data. The
small oxide precipitates shown in Figure 3.16 b might be from partial dissolution of Cr-TiFe occurring within the grain during MW irradiation.
77
Figure 3.16 BSE image of cracking in a chromite grain (Mag) enclosed by pyrrhotite (Po).
The BSE image in Figure 3.17 a shows cracks propagating radially from an assemblage of
metallic-bearing minerals. The black circled portion in Figure 3.17 a is viewed at 200× in
Figure 3.17 b. Again, it is seen that Po grains produce cracks that propagate through the Px
host rock. The Ti-bearing ulvöspinel (Usp) grain shows intergranular cracking which is a
good indication of rapid heating during MW exposure.
The BSE image in Figure 3.18 is a jaw crushed particle (-8+12 mesh) exposed to MW
energy and also one size range of particles used for grinding experiments discussed in
Chapter 4. Cracking around the silicate/sulfide grain boundary is ideal for MW induced
cracking of ore particle samples. The metallic bearing minerals remain intact with little or
no chemical alteration but are liberated from the silicate host rock. Additionally, the
successful liberation of magnetic Spl grains from neighboring host rock using MW energy
has implications for processing refractory ore material that contains these types of minerals.
78
Figure 3.17 BSE image of cracking between metallic bearing minerals ulvöspinel (Mag),
pyrrhotite (Po), pentlandite (Pnt), and chalcopyrite (Ccp).
Figure 3.18 BSE image of cracking along a pyrrhotite (Po)-pyroxene (Px) grain boundary.
79
3.3.4 Microwave Induced Arcing
The phenomenon of MW induced electric arcing was observed on surfaces of particle
samples. Upon exposure to MW energy, metallic phases become highly conductivity with
the increasing temperature and the electric arc is produced during oxidation of sulfide
minerals exposed to open atmosphere. The electric arc is the result of a highly exothermic
sulfide oxidation reaction under MW energy that ionizes the gases produced by oxidation.
The electric arc, or “cold” plasma generation, chemically alters phases due to high localized
temperatures in the proximity of the oxidation reaction induce phase changes. In some
circumstances, the highly concentrated temperature gradient from the plasma locally melts
phases creating a multi-phase solution or flux-type material.
The image in Figure 3.19 is of two chalcopyrite (Ccp) particles put into contact and exposed
to MW energy for 30s. The (smaller) particle on the left reached oxidation temperatures
first, likely from a higher power density within the particle volume. It appears the reaction
is physically driven by MW energy inferred by the displacement of oxidation product to
the left of the particle. MW energy is thought to drive the oxidation reaction of metallic
sulfides exposed in open atmosphere, possibly a pondermotive effect.
Figure 3.19 Oxidation and arcing of pure chalcopyrite (Ccp) by MW energy.
80
As previously shown, the oxidation reaction under microwave energy is extremely
exothermic creating high local temperatures that decompose particle surfaces. The surface
in Figure 3.20 (a) is heavily oxidized during the decomposition of sulfide minerals in open
atomosphere. The image in Figure 3.20 (b) shows formation of oxide scale. EDS of the
scale in Figure 3.21 indicates the presence of iron oxide.
Figure 3.20 BSE image of a heavily oxidized particle surface during MW exposure.
Figure 3.21 EDS of the iron oxide scale (Fe 94 O 6) at. %.
81
The SEM image in Figure 3.22 (a) shows a polished cross-section two fused ore particles
from MW exposure in open atmosphere. Away from the fused section, a large amount of
cracking is seen in the interior of particles from thermal stresses produced during irridation.
Compositionally unaltered phases within samples are pyrrhotite (Po), chalcopyrite (Ccp),
spinel (Spl), nickel-pyrrhotite (Ni-Po), olivine (Ol) and Ca-rich clinopyroxene (Cpx). The
circled area in Figure 3.22 (a) can be viewed at higher magnification in Figure 3.22 (b).
The image in Figure 3.22 (b) is a portion of the sample where decomposition of phases has
taken place. Electric arcing of metallic minerals created a melt flux consisting of silicates,
iron sulfides and spinel minerals. The decomposition and melting of leaves large voids
around sulfide mineral grains when phases dissolve into solution. Sulfide minerals form a
metal-sulfide liquid seen as veins within a portion of partially melted Cpx. SEM beam
probing the veins in multiple areas shows a composite phase of sulfides and silicates.
Figure 3.22 BSE image of melt-flux produced during electric arcing on metallic minerals
The EDS spectra in Figure 3.23 is the composition of veins showing them to be partially
melted silicate and metal-sulfide liquid. The large grey globules are a multi-phase solution
mainly consisting of the ferrospinel (Spl) mineral chromite. The veins are seen in Spl
globules throughout the BSE images in Figure 3.22. Enclosed in the large grain of Mag is
82
Ni-Po which is the source of nickel during EDS of metal-sulfide liquid. Magnetic
ferrospinels were suspected to enhance heating by rapid MW absorption upon exposure.
Figure 3.23 EDS of metallic veins indicating partial dissolution of sulfide into silicate
High localized temperatures produced by MW energy caused incipient (early) melting of
peridotite rock and dissolution of sulfide minerals that together formed a melt-flux solution.
The BSE images in Figure 3.24 shows fully decomposed peridotite that melted
incongruently into a two phase silicate melt (si-mt) composed Mg-rich olivine crystals and
Ca-rich pyroxene (Cpx) liquid that contains a portion of sulfur as indicated by EDS. The
white circled area of Figure 3.24 (a) can be viewed at 1000× in Figure 3.24 (b). It appears
olivine crystals grow progressivly larger toward the center of the melt. The Cpx liquid is
thought to be a solution Ca-rich silicate melt that contains partially dissolved iron sulfide
(Cpx+Po). Liquid metal-sulfide (sul-mt) droplets are seen dispersed throughout the
Cpx+Po portion of si-mt.
83
Figure 3.24 BSE olivine crystals nucleated in the melt-flux material
An x-ray diffraction (XRD) pattern of flux material is viewed in Figure 3.25. The pattern
is noisey from a high silicate content and shows the flux is mainly composed of Ol and Px
but peaks of Ccp and Po are detected which are presumably signals from sul-mt droplets.
These metal sulfide peak signals more likely belong to the high temperture phases of Curich intermidiate sulfide solution (iss) and Ni-rich monosulfide solution (mss).
EDS taken on the interior portion of crystals in the melt (Figure 3.26) shows it is a Mg-rich
olivine (Ol). The crystal edges have intermidiate compositions of Ol, or enstatite (En) in
equalibria with Cpx, with Fe and Mg occuring at relativaly the same atomic percent (Figure
3.27). It was hypothesized that Ol crystals are Mg-rich from the possible effect of MW
activated electrochemical diffusion that drives Fe out of Ol crystals into Cpx during
irradiation of melt-flux solution. This is demostrated by the illustration:
Ol
En
Cpx+Po
Mg
▼ Fe→
Fe, Al,Ti,Ca,S
Figure 3.25 Illustration showing electrochemical diffusion Fe into melt-flux solution
84
EDS of Cpx melt in Figure 3.27 shows it contains titanium (Ti) and is depleted of
magnesium (Mg). The Ti content now found in Cpx could be the result of ion exchange
and atomic migration during irradiation having originally belonged to ferrospinel. This
might further suggest that MW energy exerts a strong driving force for mass transfer of
ions in the melt by electrochemical diffusion of metal cations in solution.
High local tempertures decompose and melt sulfide minerals Po, Ni-Po and Ccp that form
sul-mt with a relative composition shown by EDS in Figure 3.28. The metal-sulfide
droplets dispersed throughout the flux are immisible in si-mt. Multiple EDS analyses taken
of the Cpx+Po portion of the melt showed that it did not contain Ni or Cu but did indicate
sulfur (S) was present. The EDS spectra before and after melting might suggest that iron
sulfide partially dissolves into Cpx during dissolution of phases.
Figure 3.26 XRD pattern of the melt-flux solution (-150 mesh).
85
Figure 3.27 EDS spectra of the olivine (Ol) crystal interior.
Figure 3.28 EDS spectra of the olivine (Ol) crystal edge.
86
Figure 3.29 EDS spectra of clinopyroxene containing iron sulfide (Cpx+Po).
Figure 3.30 EDS spectra of metal-sulfide liquid (sul-mt).
87
In Figure 3.30, a sul-mt droplet is entrained in si-mt. EDS shows the droplet is a two phase
metal-sulfide solution of Ni-rich monosulfide solution (mss) and Cu-rich intermidiate
sulfide solution (iss). Close inspection of the mss shows it contains areas of exsolved
ferrospinel (Mag) with dendritic features similar to Widmanstätten’s pattern in Fe-Ni
meteorites. Since Mag and mss are magnetic phases, it is hypothesized that Mag initiated
the formation of mss from sul-mt by magnetic devitrification.
Figure 3.31 BSE image of a metal-sulfide particle surronded by silicate melt (si-mt). Close
observation shows ferrospinel (Mag) exsolved in areas of monosulfide solution (mss).
The BSE images in Figure 3.31 illustrates a possible macroscopic effect of electrochemical
diffusion with the sul-mt droplets suspended si-mt solution. It is thought that strong MW
absorption by E-H field coupling drives sul-mt droplets through the si-mt causing mass
transfer in the melt. Extreme local temperatures produced in this sample permitted fine
dispersion of sul-mt droplets and is thought to be the effect of high sulfur saturation in the
si-mt. Ferrospinel minerals were not discovered in any portion of this sample demonstrating
decomposition of sulfide minerals during MW exposure.
88
Figure 3.32 BSE image of metal-sulfide droplets dispersed in the melt-flux solution.
The BSE image in Figure 3.32 is of an unpolished sample with metal-sulfide droplets that
“flowered” on the surface. The gold flecks found here formed on metal-sulfide droplets
that nucleated on the sample surface during decomposition of phases. Elemental gold (Au)
is known to occur in Cu-Ni sulfide deposits and is found in trace amounts only being
detected by precise analytical chemistry techniques. EDS confirms the presence of gold
(Au) which can be found in Appendix 3.49. Other elements detected in this EDS spectra
are due to rastering of the SEM beam and the inability to precisely probe the sub-micron
fleck.
The secondary electron image (SEI) in Figure 3.33 shows a line metallic sulfide particles
that formed from a crack on the surface of a sample. The displacment of sulfide minerals
onto the surface of the sample further supports the MW electrochemical effect previously
discussed. MW heating generated high local temperatures that produced sul-mt which
readily attenuates MW energy and is electromotively driven out of the crack onto the
surface. Upon cooling, sul-mt segregates into line of mss and iss phases having distinct
grain morphologies. Although these metal sulfides were heated in open atmosphere, EDS
showed no indication of oxide product(s) formed on these particles.
89
Figure 3.33 SEI image of gold (Au) flecks formed on a metal-sulfide particle.
Figure 3.34 SEI image of metal-sulfide particles formed on the surface of a sample.
90
Chapter 4: Crushing and Grinding Experiments
4.1 Experimental Background
Laboratory-scale work index experiments are time consuming and require significant
amounts of material for completion when using F. Bonds grinding procedures. Numerous
sieve sizes cannot be tested during a single work index experiment because the undersize
material of the test sieve is removed and replaced with fresh feed after each grinding cycle.
Further, it takes more ball mill revolutions to test and obtain material for increasingly finer
particle sizes. Completion of grinding experiments is achieved when the circulating load
(the ratio of oversize to undersize material) reaches 250% and reverses direction of either
increasing or decreasing. A detailed step-by-step ball milling procedure for determining
the work index (Wi) can be found in Appendix 4.
Understanding the breakage characteristics of particles during grinding is important when
performing a comparative study of a material’s communition performance. This entails
choosing a test sieve that displays differences in cumulative percent passing values for the
different data sets. Proper determination the test sieve greatly reduces laboratory
expenditures such as material, equipment wear, time and energy. Therefore, preliminary
grinding experiments were performed in an attempt to better understand the size
classification characteristics of communized material before and after MW pre-treatment.
During MW heating, particle size affects the power absorption density and overall heating
of bulk particle samples. Large, heterogeneous particles contain varying amounts of MW
lossy and non-lossy phases. The amount of air between particles (packing density) also
affects MW heating properties of particle samples. Small volumes of material heat quickly
because the MW energy density (power per unit volume) within the cavity is greater than
for large volumes of material that are able to absorb more MW energy. Thus, the overall
heating in a bulk particle sample will depend on the particle size, volume of material, and
the time duration samples are irradiated.
91
In these experiments, MW energy was used in an attempt to improve communition
performance of sulfide ore material by inducing cracks in particles due to differential
thermal expansion of constituent phases. The possibility of an oxide layer formed on the
exterior of particles may also improve grindablity of ore particles being that it is more
brittle and breaks away from the existing sulfide interior. Further, under zero applied MW
field, MW heated particles cooled quickly in air, and additional cracking might be caused
by thermal shock due to the residual stresses formed in particles.
The x-ray diffraction (XRD) work performed on samples proved somewhat problematic
for completely accurate phase identification in ball milled (-100) material. Silicates
introduce many peaks into an XRD pattern from low symmetry crystal structures making
it nearly impossible to fully identify all the peaks produced in patterns. It was not necessary
to identify all the XRD peaks that appeared in patterns because metallic-bearing minerals
were the main focus during characterization of ball milled material.
A comparison of raw XRD patterns is included in Appendix 5 to demonstrate the relative
differences in the intensity of peaks. The individual XRD patterns for as-received and MW
60s material show diffraction lines of metallic-bearing minerals pyrrhotite (Po),
chalcopyrite (Ccp), pentlandite (Pnt) and spinel (Spl). The most intense peaks for Po occur
at 2θ ≈ 34° and 44° corresponding to the [004] and [522] planes respectively [1]. The 522
reflection was used as the main indication Po present in patterns because the 004 reflection
was not easily de-convoluted from silicate peaks. Any Po peak signals also support the
presence of Ccp and Pnt in samples.
Metal oxides of Fe, Cu, and Ni were imposed on XRD patterns and were seemingly not
present. If sulfide minerals oxidized on the surface of a particle during MW heating, the
amount of oxide product created would be relatively small in volume compared to the bulk
particle. A large amount of material was produced during communition experiments so
oxide phases that formed during heating fell under the limits of XRD detection.
92
Heavy liquid separation was employed to investigate thermally induced liberation of
metallic-minerals from host rock. The sink material of centrifuged samples was examined
with XRD for qualitative phase identification (Appendix 5). It is understood that XRD
sample texturing issues are reduced with decreasing particle size but material was not
further pulverized after size classification to preserve intergrowths of sulfide minerals and
also not to mechanically alter phases. Therefore, it is possible that silicate minerals can
report to sink material if un-liberated from metallic bearing minerals and/or have a density
greater than the heavy liquid used for density separation. Inductively coupled plasma
(I.C.P) results also supported the presence of silicate minerals in sink material.
It should be noted that many minerals exhibit slight compositional variations that affects
their overall density. Silicate minerals display a wide range of densities making heavy
liquid separation of pyroxene silicates from metallic minerals ineffective when using
methelyne iodide (CH2I2). Mg-Fe-Ca-rich pyroxenes (diopside) reported to sink material
because its specific gravity is greater than that of the CH2I2 (ρ=3.25). Consequently, these
heavy silicates “mask” various peak signals during XRD experiments by introducing
numerous peaks that overlapped metallic peaks. More dense heavy liquids are available
but are highly toxic (Cleric’s solution) and use was formidable for our purpose here.
4.2 Preliminary Ball Milling Experiments
4.2.1 Effect of Increasing Ball Mill Revolutions
The plots in Figure 4.1 show the particle size distribution for MW 60s and as-received
material (-6+8) ball milled for 100, 300, 500 and 1000 revolutions. The results show there
is over a 25% increase in the amount fine material (≤ 74μm) produced when increasing ball
mill revolutions from 100 to 1000. It is intuitive that increasing the amount of ball mill
revolutions increases the amount of cumulative percentage passing material. Here,
differences in the plots of Figure 4.1 show that MW 60s material produced more
cumulative percent passing material then as-received material for the same amount of ball
mill revolutions.
93
Upon closer inspection of the plots in Figure 4.1, a difference in the linearity of sections in
the range of 208-74μm (65 to 200 mesh) is seen as ball mill revolutions are increased. At
100 revolutions, the 208-74μm section of the two plots have similar slopes but as the
amount of grinding revolutions is increased, the slope of this section becomes steeper for
as-received material. This indicates an increased amount material produced in this size
range for MW 60s material suggesting finer particle sizes of ball milled product.
Ball Milled Microwave Treated and As-received Particles -6+8
Cumulative % Passing
100.0
As Rec'vd
10.0
MW 60s
1.0
10
100
Particle Size (μm)
1000
10000
Figure 4.1 Particle size distribution for microwave 60s and as-received (-6+8) material ball
milled for 100, 300, 500 and 1000 revolutions.
4.2.2 Effect of Microwave Treatment Time on Particle Size Reduction
Ore particle samples exposed to MW energy for shorter durations of time consume less
energy but may not attain high enough temperatures to induce thermal stresses that produce
cracks in ore particles. To illustrate the effect of MW treatment time on the ball mill
grindability, the data sets in Figures 4.2-4.4 also include MW 30s material.
94
In Figure 4.2, 50% of all grinded material is able to pass the original retained feed size after
ball milling for 100 revolutions. The MW 30s material does not show any differences in
cumulative percent passing material when compared with as-received. Both plots follow
similar behavior and are basically super-imposed on each other. The MW 60s plot follows
a similar trend as the other plots but with an increased amount of passing material.
A significant effect in the overall amount of passing material is seen for ball milling 300
revs (Figure 4.3). The as-received and MW 30s plots appear have slight differences in the
200 mesh (74μm) sieve though there is very little difference (≈ 1%) in the amount of
cumulative percent passing material at larger particle sizes between the two plots.
Therefore, during grinding experiments in which 300 revolutions was needed to achieve
the target circulating load (250 %) the 200 mesh (74μm) test sieve should be used when
comparing MW 30s and as-received material in order to show improved grindability. For
MW 60s material, the separation between plots demonstrates that any test sieve size can be
used to show an improvement in grindability if approximately 300 ball mill revolutions
were needed to produce the target circulating load.
Upon increasing the ball mill revolutions to 500, nearly 80% of the ball milled material is
able to pass the original feed size (Figure 4.4). All of the plots display similar behavior and
the amount of cumulative percent passing material significantly drops off after 65 mesh
(208μm). There is not much linearity for as-received material in the particle size range
between 208-74μm. The linearity of plots in this size range improves for microwave treated
material showing consistent particle size distribution for MW 60s material that passes 65
Tyler mesh (208μm).
95
Particles -6+8 Ball Milled 100 revolutions
Cumulative % Passing
100.0
10.0
As-Rec'vd
MW 30s
MW 60s
1.0
10
100
1000
10000
Particle Size (μm)
Figure 4.2 Particle size distribution for (-6+8) material ball milled 100 revolutions.
Particles -6+8 Ball Milled 300 revolutions
Cumulative % Passing
100.0
As-Rec'vd
MW 30s
MW 60s
10.0
10
100
1000
10000
Particle Size (μm)
Figure 4.3 Particle size distribution for (-6+8) material ball milled 300 revolutions.
96
Particles -6+8 Ball Milled 500 revolutions
Cumulative % Passing
100.0
As-Rec'vd
MW 30s
MW 60s
10.0
10
100
Particle Size (μm)
1000
10000
Figure 4.4 Particle size distribution for (-6+8) material ball milled 500 revolutions.
It may be important to understand the particle size classification characteristics after ball
mill grinding. Using sieves with larger aperture sizes will reduce the amount of grinding
cycles needed to test a particle size but might not show any difference in the grindability
during a comparative study. It should be mentioned it takes more ball mill revolutions to
obtain the target circulating load with finer particle sizes. When using the plots to aid with
choosing a test sieve for a comparative grindability study, material should show significant
differences in the cumulative percent passing material for a given sieve size but the test
sieve used for experiments should have aperture openings large enough to obtain the target
circulating load without requiring excessive laboratory expenditures.
97
4.3 Ball Mill Grindability Experiments
The ball mill grindability (GBP) had the greatest effect on work index (wi) calculations.
The test sieve size (Pi), 80% feed passing material (F80) and 80% undersize passing
material (P80) did not vary much when comparing ball milled material of the same origin.
The amount of material needed for the completion of a single grindability experiment
varied between 3-5kg. This is a fairly large amount of material needed to do a comparative
study using a specific size fraction of particles. As mentioned, decreasing the test sieve size
requires more grinding cycles to achieve the target circulating load during measurement of
the work index (kW·hr/ton) and consequently more material expenditures.
4.3.1 Ball Mill Grindability of Jaw Crushed Material
For initial ball mill grinding experiments, two separate sample lots of jaw crushed particles
in the Tyler mesh ranges of (-6+10) and (-8+12) were obtained. Ore particles were spread
evenly (single layer) on the glass turntable in ≈ 500g batches and exposed to MW energy.
Grinding experiments were completed using F. Bonds procedure and the test sieves used
for this material were 35 and 65 Tyler mesh. The results of the ball mill grindability
experiments are reported in Table 4.1.
Table 4.1 Ball mill grindability (GBP) and work index (wi) results for microwave and asreceived jaw crushed (-6+10) and (-8+12) material.
Feed Material
Jaw (-6+10)
As Rec’vd
MW 60s
Test Sieve
GBP, grams
undersize per
mill revolution
35 mesh (417µm)
35 mesh (417µm)
1.83
1.97
17.16
17
Jaw (-8+12)
As Rec’vd
MW 30s
65 mesh (208µm)
65 mesh (208µm)
1.57
1.58
16.1
15.78
98
Work Index,
kW·h per ton
ore
The MW material shows an increase in the grindability (GBP), and a subsequent increase
in the work index, when compared with as-received material. Particles spread evenly on
the turntable made heating more homogenous during MW exposure. Though heating of
samples was uniform, it was later realized that significant amounts of heat radiated into the
glass turntable by conduction. Thermal loss to the surroundings is thought to be the reason
for minor differences in GBP between MW treated and as-received jaw crushed material.
4.3.2 Ball Mill Grindability of Gyratory Crushed Material
The results obtained by grinding experiments in Section 4.3.1 motivated the experimental
set-up to be modified in order to obtain greater MW heating of ore particle samples. Now,
samples were contained in an alumina crucible to increase the amount of heat transfer
between neighboring ore particles. The sample mass exposed to MW energy was decreased
to approximately 100g which allows for higher energy absorption within particle samples.
The crucible was placed on a small piece of insulating ceramic in the cavity center to reduce
the amount of heat transfer to the glass turntable. Modification of experiments achieved
more heating in samples but produced appreciable amounts of sulfur gas.
The 100 Tyler mesh (147μm) was the test sieve chosen for grinding experiments of
gyratory crushed material. This was to test in the sub-micron particle range where the
sections of plots from preliminary grinding experiments (Section 4.1) showed significant
differences in passing material. The 100 mesh sieve was also chosen because it is a
commonly known sieve size in mineral processing.
The results displayed in Table 4.2 shows an increase in the grindability (GBP) of gyratory
crushed MW treated material. There is also a decrease in the work index values for gyratory
crushed samples as MW treatment time increases. When comparing the two size fractions,
the gyratory (-6+8) material has a lower GBP then samples of the gyratory (-8+12) size
fraction. This could likely be the result of more abrasive type size reduction taking place
by collisions of balls with larger size ore particles. There is only a slight increase in
grindability of the gyratory (-8+12) microwave 30s sample; a similar result is shown by
99
jaw crushed particles (-8+12). This is thought to be partially attributed to (-8+12) particles
reduced to sizes during grinding that are now able to fit interstitially in spaces between
balls and becoming increasingly difficult to reduce in size.
Table 4.2 Ball mill grindability (GBP) and work index (wi) results for microwave and asreceived gyratory crushed (-6+8) and (-8+12) material.
Feed Material
Gyra (-6+8)
As Rec’vd
MW 30s
MW 60s
Gyra (-8+12)
As Rec’vd
MW 30s
Test Sieve
Size
(100 mesh)
GBP, grams
undersize per mill
revolution
Work Index,
kW·h per ton
ore
147μm
147μm
147μm
0.79
0.82
0.86
21.31
20.52
19.87
147μm
147μm
0.97
0.98
18.16
18.12
The work index values tabulated here for all material cannot be directly compared because
the test sieve size changed. Nonetheless, the GBP of MW treated particles increased
resulting in a decrease in the work index observed in samples. Particle size reduction in the
ball mill is largely caused by abrasion to the exterior of particles, so the MW induced
cracking in ore particles is not completely took advantage of during grinding.
4.3.3 Sample Weight Loss
Sample weight loss is the result of oxidation of sulfide minerals from the creation of sulfur
gas during heating and sulfur lost to the atmosphere. Samples were massed to
approximately 100 ± 1g and then exposing them to MW energy for various times. The
average mass loss for individually heated particle samples is reported in Table 4.3.
100
Table 4.3 The average mass of particle samples before and after microwave heating.
Sample
Gyro (-4+6) MW 30s
Gyro (-6+8) MW 30s
Gyro (-8+12) MW 30s
Gyro (-10+20) MW 30s
Gyro (-6+8) MW 60s
Before (g)
100.04
100.07
100.02
100.02
100.06
After (g)
99.96
99.95
99.93
99.89
99.70
wt loss (g)
0.09
0.12
0.09
0.13
0.32
After observing gyratory (-6+8) MW 60s particle samples lost nearly 3 times the mass per
sample then the same size particles heated for 30s, all experiments thereafter exposed
particle samples to MW energy for 30s to reduce the production of sulfur gas. MW
irradiation for longer durations of time increases mass loss in samples by increased
oxidation of sulfide minerals exposed to open atmosphere during heating. Particle samples
heated in argon atmosphere reported mass loss in samples, possibly by redox reactions
occurring between neighboring mineral phases.
The packing density of particle during MW exposure can affect heating in samples. The
combination of more surface area exposed to the atmosphere and enhanced heat transfer
between particles is thought to have resulted in the greatest weight loss in samples. The
overall power density for small particles may not be very high due to a low volume fraction
of MW absorbing phases in individual particles. In contrast, the MW penetration depth
may be inadequate for large particles. Thus, the best particle size should achieve maximum
penetration depth. The heating in samples is mainly due to the metallic mineral content of
ore particles that creates high power absorption density (W/m3) per individual particle.
4.3.4 Particle Size Distribution of Ball Milled (±100) Material
The remaining oversize and undersize material from grindability experiments was size
classified to gain understanding the particle size distribution of ball milled material. After
removing the last undersize material used for finding 80% passing material (P80), ≈ 700g
of oversized material was retained after each run of the ball mill. The oversized material
produced during ball milling experiments was size classified using the Tyler sieves 6 mesh
(3.327mm), 10 mesh (1.651mm), 20 mesh (850μm), 35 mesh (417μm), 48 mesh (295μm)
101
65 mesh (208μm). The test sieve used during grinding experiments was the Tyler 100 mesh
(147μm) that had retained oversize material (+100) and passing undersize material (-100).
The particle size distributions for +100 material is seen in Figure 4.5 with all plots
following the same general behavior. The as-received and MW 30s plots show no
differences in the cumulative percent passing material but it is shown that greater size
reduction took place with the MW 60s material.
Particle Size Distribution of Ball Milled -6+8 Oversized Material
Cumalutive % passing
100.00
As-Rec'vd
MW 30s
10.00
1.00
100
MW 60s
1000
10000
Particle Size(μm)
Figure 4.5 Cumulative percent passing of oversize material (+100 mesh) from ball mill
grindability experiments using gyratory (-6+8) particles.
To generate the plots in Figure 4.6, 1000g of undersize material produced during
grindability experiments was sized classified using the Tyler sieves 100 mesh (147µm),
150 mesh (104µm), 200 mesh (74µm), 270 mesh (53µm), 325 mesh (45µm), and 500 mesh
(25µm). The undersize size plots in Figure 4.6 all follow similar trends until material
passing the 270 mesh (53μm). At this particle size, the as-received material passing
material begins to drop off while MW treated material remains with the original line plot
until then it drops off at 325 mesh (45μm). The plot shows greater than 30% of MW treated
material passes the 325 mesh while approximately only 23% of as-received passes this
mesh size. The creation of fine sub-micron sized particles (<53μm) is thought to be the
102
reason for the increased grindability of MW material during measurement of the work
index. The plots show that size reduction in the ball mill is accomplished by abrasive wear
of balls on particle surfaces resulting in the production of very fine size material.
Particle Size Distribution of Ball Milled -6+8 Undersized
Material
Cumulutive % Passing
100.0
10.0
As-Rec'vd
MW 30s
MW 60s
1.0
10
100
1000
0.1
Particle Size(μm)
Figure 4.6 Cumulative percent passing of undersize material (-100 mesh) from ball mill
grindability experiments using gyratory (-6+8) particles.
4.3.5 XRD and ICP of Ball Milled (-100) sink material
The XRD patterns in Figure 4.7 is of undersize (-100) ball milled material obtained during
grinding experiments. The oxidation of sulfide minerals on the surface of ore particles was
a frequent occurrence during irradiation but newly formed oxide phases could not be
identified by XRD analysis. This is likely due to a relatively low volume fraction of oxide
phases formed on the surfaces of particles. It was not well investigated exactly what phase
changes occurred during heating. A detailed analysis of newly formed phases would give
insight about the temperatures various minerals reached during MW exposure.
The XRD patterns in Figure 4.7 show numerous peaks because samples contained many
different minerals along with a high volume of silicates. Silicate minerals are known to
103
produce sub-micron sized material during communition and, here, greatly contribute to
noise in XRD patterns making peaks convoluted, difficult to resolve, and consequently
interfere with metallic peak identification.
Figure 4.7 XRD plots of intensity (arbitrary units) vs. 2θ (degrees) of ball milled undersize
(-100 mesh) sink material.
Inductively coupled plasma (I.C.P.) was used to analyze the elemental composition of asreceived and MW 60s ball milled (-100) samples. MW treated ore particles show an
increase in copper (Cu) and nickel (Ni) content (Figure 4.8), thought to be from successful
liberation of sulfide minerals during MW exposure. The contamination of heavy silicates
in sink samples is confirmed by the results in Figure 4.9. The increase of silicates in MW
treated sink samples in further confirmed by an increase in the grindability which is thought
to increase the fine size silicates resulting in a greater amount of Ca-rich pyroxene (Cpx).
The liberation of ferrospinel (Mag) minerals enclosed by silicates may be indicated by an
increase in Cr and a slight increase in the Mn content of the MW 60s sample.
104
Figure 4.8 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s and
as-received ball milled (-100 mesh) sink material.
Figure 4.9 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s and
as-received ball milled (-100 mesh) sink material.
105
4.4 Roll Crusher Experiments
It was previously shown that increasing the ball mill revolutions during grinding yields a
greater amounts of fine size material. When a large amount of fines are created, compared
to that of the starting feed size material, it is assumed that size reduction is taking place
largely by abrasion. To take advantage of MW induced cracking in ore particles, and avoid
size reduction by abrasion, gyratory crushed particles in the range (-4+6) were feed through
the roll crusher.
After heavy liquid separation, it is observed in XRD patterns of MW treated sink material
that there is less silicate peak signals and an overall reduction in the amount of noise
(Appendix 5). This is thought to be the result of less silicate minerals reporting to MW
treated material suggesting that metallic bearing minerals were liberated to some degree
from the host rock. The reduction of peak broadening in XRD patterns of MW treated
material is reduced which is further an indication of liberated or partially liberated metallic
minerals. The XRD patterns of as-received sink material have many peaks of varying
intensities that introduce significant background noise. The broadening of metallic peaks
in as-received sink material might further indicate that metallic-bearing minerals report to
the sink fraction with silicate minerals still attached.
For roll crushing experiments, the heavy liquid sink material of size fractions (-65+100),
(-100+200), and (-200+325) was examined corresponding to (208-147μm), (147-74μm),
and (74-43μm) particle sizes respectively. Material passing -325 mesh (<43μm) was
assumed to contain significant amounts of silicate minerals that would be unfavorable in
the analysis of XRD patterns.
4.4.1 Particle Size Distribution Roll Crushed Particles (-4+6)
Particles in the (-4+6) size range were fed through the roll crusher and size classified using
Tyler sieves 20 mesh (850µm) , 42 mesh (355µm) , 65 mesh (212µm), 100 mesh (150µm),
200 mesh (75µm), and 325 mesh (45µm). The particle size distribution of roll crushed MW
treated and as-received ore particles are plotted in Figure 4.10. Now, it is shown that MW
106
treated material generates less fine sized particles then as-received material. This is a
different effect as seen in the particle size distribution plots of ball mill grindability
experiments. Here, the plots demonstrate that MW heating produced thermal stresses in
samples for particles to fracture more easy when passed through the roll crusher. The asreceived particles show a higher amount of cumulative percent passing material; a result
of size reduction taking place due to abrasion and the creation of fine size material.
Roll Crushed Particles (-4+6)
Cumulative % Passing
100.00
As Rec'vd
MW 30s
MW 60s
10.00
1.00
10
100
1000
Particle Size (μm)
Figure 4.10 Particle size distribution for microwave and as-received gyratory crushed ore
particles (-4+6) passed through the roll crusher.
These plots show that increasing MW treatment time from 30s to 60s did not have any
significant effect on the breakage characteristics of ore particles. Since MW power density
was the same for both samples, increasing MW exposure time from 30s to 60s did not
generate significantly more thermal stresses to weaken particles even if 60s samples heated
to an overall higher temperature. For this experimental set-up, MW treatment time was
reduced by half and still achieved a comparable amount of particle breakage.
107
4.4.2 XRD and ICP of Roll Crushed Particles (-65+100) Sink Material
The XRD patterns in Figures 4.11 are of roll crushed (-65+100) material that reported to
the sink fraction during heavy liquid separation. The MW 60s pattern shows an increase in
sulfide bearing minerals reporting to this size fraction. This most intense Po signal (2θ ≈
44°) is seen in both patterns but has greater intensity in MW 60s material. The presence of
an intense Po peak further supports identification of Ccp and Pnt peaks because these
minerals are intergrown among each other. Intergrowths of sulfide minerals reporting to
large size fractions may prove with successful “piggybacking” of crushed ore material by
magnetic separation. Raw XRD patterns of roll (-65+100) sink material with imposed
diffraction lines of metallic-bearing minerals can be found in Appendix 5.2 and 5.3.
Figure 4.11 XRD pattern of MW 60s and as-received rolled (-65+100) sink material.
Silicate peak signals are seen in both XRD patterns resulting from Ca-rich pyroxenes
reporting to sink material. The varying intensity of peaks could be partially attributed to xray texturing issues in samples. XRD patterns show that many silicate peaks present in the
108
as-received material are now absent from the MW 60s material. Indications of olivine
reporting to sink material are likely from silicates un-liberated from metallic minerals.
An increase in sulfide minerals reporting to MW 60s (-65+100) sink material is confirmed
by the I.C.P. results in Figure 4.12. It is shown that more elemental Cu and Ni reported to
this size fraction which indicates the liberation or partial liberation of sulfide minerals
exposed to MW energy. Theoretically at this size fraction, ferrospinels (Mag) are still
enclosed in their parent silicates (SEM imaging) and do not show any indication of
increased liberation.
The higher iron content in the silicate element I.C.P analysis for MW 60s (-65+100) sink
material (Figure 4.13) is attributed to an increase in the amount of iron sulfide minerals
reporting to this size fraction. It further shows that more silicate elements present in the asreceived sink material. Raw XRD patterns of roll (-65+100) sink material with imposed
diffraction lines of metallic-bearing minerals are found in Appendix 5.2 and 5.3.
Figure 4.12 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s and
as-received rolled (-65 +100) sink material.
109
Figure 4.13 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s and
as-received rolled (-65 +100) sink material.
4.4.3 XRD and ICP of Roll Crushed Particles (-100+200) Sink Material
The XRD patterns in Figures 4.14 are of roll crushed (-100+200) material that reported to
the sink fraction during heavy liquid separation. Various metallic peak signals are observed
in both patterns but peaks are well defined, and more intense, for MW 60s material
indicating higher metallic mineral content. Less background noise from silicates is also
observed in MW treated material. The key indicator of Po is the peak at 2θ ≈ 44° with the
most intense peak signal (2θ ≈ 29°) observed in the pattern belonging to a Ccp reflection,
possibly from a small contribution of Pnt that broadens the peak (Ccp and Pnt share this
reflection). The Ccp reflection is much more intense in the MW 60s pattern. Liberated
metallic sulfide minerals reporting to MW 60s material have sharp, well defined peaks and
less background noise in patterns. Indications of Mag minerals also appear. The raw XRD
patterns of roll (-100+200) sink material with imposed diffraction lines of metallic-bearing
minerals are found in Appendix 5.2 and 5.3.
110
Figure 4.14 XRD pattern of the MW 60s and as-received rolled (-100+200) sink material.
An increase in the Cu and Ni content of MW 60s material is shown by I.C.P. results in
Figure 4.15. Indications of Mag minerals in the MW 60s XRD pattern of (-100+200) sink
material are supported by the substantial presence of transition metals in the I.C.P. analysis.
Theoretically, Mag inclusions are partially liberated from silicate host rock at this size
fraction.
The I.C.P. results of the (-100+200) size fraction in Figure 4.16 showed the least amount
of silicate elements present then in the other samples studied. This is the reason for the
increase on iron content for MW 60s material in the I.C.P. analysis of silicate elements.
The results shown here indicate the MW 60s (-100+200) size fraction has a high metallic
mineral content and might prove significant with further invstigations.
111
Figure 4.15 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s and
as-received rolled (-100+200) sink material.
112
Figure 4.16 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s and
as-received rolled (-100+200) sink material.
4.4.4 XRD and ICP of Roll Crushed Particles (-200+325) Sink Material
The XRD patterns in Figures 4.17 are of roll crushed (-200+300) material that reported to
the sink fraction during heavy liquid separation. XRD patterns of the (-200+325) size
fraction ishow many silicate peak signals and weak indications from metallic bearing
minerals. The excessive noise and clustering of peaks is caused by a high amount of silicate
material in samples and it was expected that silicates would report to this size fractions. In
as-received material, the most intense Po peak signal (θ=44º) is strong accompanied by a
low intensity Ccp peak signal that appears faintly in MW 60s material. Raw XRD patterns
of roll (-200+325) sink material with imposed diffraction lines of the metallic-bearing
minerals can be found in Appendix 5.2 and 5.3.
113
Figure 4.17 XRD pattern of MW 60s and as-received rolled (-200+325) sink material.
The I.C.P. results in Figure 4.18 support the decrease in iron sulfide minerals that reported
to the (-200+325) size fraction for MW 60s sink material. It is thought that liberated sulfide
minerals reported to larger size fractions for MW 60s material which might explain the
nearly absent Po peak signal. Intergrowths of sulfide minerals (Po-Ccp-Pnt) may have been
liberated to larger size fractions decreasing the Ni-content of MW 60s (-200+325) sink
material. The higher in Cu for MW 60s material might be explained by the creation of
appreciable amounts of Cu-oxide that reports to this sub-micron size fraction. MW 60s sink
material also shows to have higher silicate content for this size fraction (Figure 4.19).
114
Figure 4.18 Inductively coupled plasma (I.C.P.) of metallic elements in microwave 60s and
as-received rolled (-200+325) sink material.
Figure 4.19 Inductively coupled plasma (I.C.P.) of silicate elements in microwave 60s and
as-received rolled (-200+325) sink material.
115
Chapter 5: Conclusion
Microwave (MW) exposure of ore particle caused differential thermal expansion of
constituent phases producing macro and microscopic cracks in ore particles. The
production of sulfur gas and subsequent weight loss in samples was also reported.
Fracturing originates at metallic mineral phases and continue throughout the host rock
matrix. Ball milling experiments show an increased grindability (GBP) of MW treated ore
particles resulting in an overall decrease in the work index (wi) compared with as-received
material. MW treated ore particles passed through a roll crusher showed an increase in
metallic bearing minerals reporting to coarser size fractions then as-received material and
less fine material produced.
Sulfide ore particles exposed to MW energy in open atmosphere can electrically arc during
sulfide oxidation generating plasma. The plasma reaction is highly exothermic and
generates high local temperatures that melt constituent phases into a flux-type solution.
The flux-type solution is a mixture of metal sulfide droplets (sul-mt) dispersed in silicate
melt (si-mt) that also contains partially dissolved sulfur. The droplets are a two phase
solution composed of Ni-rich monosulfide solution (mss) and Cu-rich intermediate sulfide
solution (iss).
116
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
R. Goodland, "Responsible Mining: The key to Profitable Resource
Development," Univ. of Vermont, 2012.
S. Dudka and D. C. Adriano, "Environmental Impacts of Metal Ore Mining and
Processing: A Review," Journal of Environmental Quality, vol. 26, no. 3, pp. 590602, 1997.
M. W. Barsoum, Fundamentals of Ceramics, Taylor & Francis, 2003.
A. Gupta and D. Yan, Mineral Processing Design and Operations: An
Introduction, Elsevier, 2006.
M. Rhodes, Introduction to Particle Technology, Chicheste: Wiley, 1998.
B. A. Wills, Mineral Processing Technology, Elevier, 1979.
F. C. Bond, "Crushing and Grinding Calculations: Part 1," 1959.
F. C. Bond, "Crushing and Grinding Calculations: Part 2," 1959.
R. Ahmadi and S. Shahavari, "Procedure for Determination of Ball Bond Work
Index in the Commercial Operations," Minerals Engineering, vol. 22, no. 104-106,
2009.
L. Tavares, "Energy Absorbed in Breakage Of Single Particles in Drop Weight
Testing," Minerals Engineering, vol. 12, pp. 43-50, 1999.
K. S. Free and M. K. McCarter, "Evaluation of a New Method for Work Index
Estimation using Single Particle Impact Tests," in SME Annual Meeting, 2004.
Y. Guéguen and V. Palciauskas, Introduction to the Physics of Rocks, Princeton
University Press, 1994.
K. M. Rosso and D. J. Vaughn, "Reactivity of Sulfide Mineral Surfaces;," in
Sulfide Minerology and Geochemistry, vol. 61, Geochemical Society and
Mineralogical Society of America, pp. 557-600.
J. S. Klasner, D. W. Snider, W. F. Cannon and J. F. Slack, "The Yellow Dog
Peridotite and a Lower Keweenawan Age in the Northern Peninsula of Michigan,
Report of Investigation 24," Michigan D.N.R., Lansing, 1979.
M. E. &. G. A. contact: Jon Cherry (cherryj@kennecott.com), "Kennecott Eagle
Minerals Communities Relations Plan 2007-2011," Kennecott Eagle Minerals
Company.
S. E. Chester, "Michigan's Nonferrous Metallic Mining Regulations Part 632,"
Michigan Department of Enviromental Quality Geological Survey Division,
Lansing, 2006.
D. Rossell and S. Coombes, "The geology of the Eagle nickel-copper deposit
Michigan, USA: Kennecott Exploration report to Kennecott Minerals Company,"
2005.
X. Ding, E. Ripley and C. Li, "PGE geochemistry of the Eagle Ni-Cu-(PGE)
deposit Upper Michigan: constraints on ore genesis in a dynamic magma conduit,"
Miner Deposita.
117
[19] G. L. LaBerge, Geology of the Lake Superior Region, Geoscience Press Inc, 1994.
[20] Foth and V. Dyke, "The Kennecott Flambeau Mine Process Documents," 1989.
[Online]. Available:
http://uwdc.library.wisc.edu/collections/EcoNatRes/Kennecott. [Accessed 19
March 2013].
[21] "Water Legacy," Copyright© 2009. [Online]. Available:
http://waterlegacy.org/flambeau_mine.
[22] "Friends of the Boundary Waters Wilderness," Minnesota Enviromental Fund,
[Online]. Available: http://www.friends-bwca.org/issues/sulfide-mining/.
[23] "Save The Wild UP," Copyright © 2007 - 2012 Save the Wild U.P.. [Online].
Available: http://savethewildup.org/.
[24] C. Huckins, "Metallic Sulfide Mining: A Threat to Coaster Brook Trout," March
1996. [Online]. Available: www.nwf.org/greatlakes.
[25] H.-R. Wenk and A. Bulakh, Minerals: Their Constitution and Origin, Cambridge
University Press, 2004.
[26] J. J. Papike, Pyroxenes, Vols. Reviews in Mineralogy, Vol. 7, C. T. Prewitt, Ed.,
Mineralogical Society of America, 1980.
[27] J. G. E. Brown, "Orthosilicates," in Olivine and Silicate Spinels, vol. 5, P. H.
Ribbe, Ed., Mineralogical Society of America, 1980.
[28] A. C. Lasaga and R. Cygan, "Electronic and Ionic Polarizabilities of Silicate
Minerals," Americna Minerologist, vol. 67, pp. 328-334, 1982.
[29] E. H. Oelkers and J. Schott, Thermodynamics and Kinetics of Water-Rock
Interactions, Vols. Reviews in Mineralogy and Geochemistry Vol. 70,,
Mineralogical & Geochemical Society of America, 2009.
[30] A. Putnis, "Mineral Replacement Reactions," Thermodynamics and Kinetics of
Water-Rock Interactions.
[31] R. Powell, Equalibrium Thermodynamics in Petrology: An Introduction, Harper
and Row, 1978.
[32] Y. Lang, "Dissolution in molten silicates: Effects of solid solution," Geochimica et
Cosmochimica Acta, vol. 64, no. 9, pp. 1617-1627, 2000.
[33] A. Holzheid and K. Lodders, "Solubility of copper in silicate melts as a function of
oxygen and sulfur fugacities, temperature, and silicate composition," Geochimica
et Cosmochimica Acta, vol. 65, no. 12, pp. 1933-1951, 2001.
[34] N. Morimoto, "Nomenclature of Pyroxenes," vol. 27, pp. 143-156, 1989.
[35] E. Huang, C. Chen, T. Huang, E. Lin and J. Xu, "Ramen spectroscopic
charateristics of Mg-Fe-Ca pyroxenes," American Minerologist, vol. 85, pp. 473479, 2000.
[36] M. Cameron and J. Papike, Crystal Chemistry of Silicate Pyroxenes, C. Prewitt,
Ed., Mineralogical Society of America.
[37] A. Lasaga and R. Kirkpatrick, "Kinetics of Geothermal Processes," Reviews in
Mineralogy, vol. 8, 1981.
118
[38] D. J. Cherniak and A. Dimanov, "Diffusion in Pyroxene, Mica, and Amphibole,"
in Diffusion in Minerals and Melts, vol. 72, Y. Zhang and D. Cherniak, Eds.,
Mineralogical Society of America & Geochemical Society, 2010, pp. 641-690.
[39] R. Lange, J. J. D. Yoreo and A. Navrotsky, "Scanning calorimetric measurement
of heat capacity during incongruent melting of diopside," American Mineralogist,
vol. 76, pp. 904-912, 1991.
[40] R. Balluffi, S. M. Allen and W. C. Carter, Kinetics of Materials, Wiley, 2005.
[41] Y. Zhang, H. Ni and Y. Chen, "Diffusion Data in Silicate Melts," in Diffusion in
Minerals and Melts, vol. 72, Mineralogical Society of America Geochemical
Society, 2008, pp. 311-408.
[42] J. Glover, "Thermodynamics of Pyroxenes," Reviews in Minerology, vol. 7.
[43] N. Doukhan, J. Doukahn, J. Ingrin and P. R. O. Jaoul, "Early partial melting in
pyroxenes," American Mineralogist, vol. 78, pp. 1246-1256, 1993.
[44] P. Buseck, J. G.L. Nord and D. Veblen, "Subsolidus Phenomena in Pyroxenes," in
Pyroxenes, vol. 7, C. Prewitt, Ed., Mineralogical Society of America, 1982.
[45] C. Lesher, "Self-diffusion in Silicate Melts: Theory, Observations and
Applications to Magmatic Systems," in Diffusion in Minerals and Melts, vol. 72,
Mineralogical Society of America Geochemical Society, 2008, pp. 269-309.
[46] J. G. E. Brown, "Olivine and Silicate Spinels," in Orthosilicates, vol. 5, P. Ribbe,
Ed., Mineralogical Society of America, 1980.
[47] Y. Z. a. C. Walker, "Diffusive crystal dissolution," in Contributions to Mineral
Petrology, vol. 102, 1989, pp. 492-513.
[48] P. Barton, "Sulfide Petrology," in Sulfide Mineralogy, vol. 1, P. Ribbe, Ed.,
Mineroligical Society of America, 1974, pp. B1-B11.
[49] D. J. Vaughn, "Introduction and Overview, D. Vaughn," in Sulfide Mineralogy
and Geochemistry, vol. 61, Mineralogical Society of America, 2006, pp. 1-5.
[50] K. M. Rosso and D. J. Vaughn, "Chemical Bonding in Sulfide Minerals," in
Reviews in Sulfide Mineralogy & Geochemistry, vol. 61, Geochemical Society,
Mineralogical Society of America, 2006, pp. 231-259.
[51] E. Makovich, "Crystal Structures of Sulfides and other Chalcogenides," in Sulfide
Minerology and Geochemistry, vol. 61, D. J. Vaughn, Ed., Geochemical Society
and Mineralogical Society of America, 2006, pp. 7-125.
[52] C. I. Pearce, R. A. Pattrick and D. J. Vaughan, "Electrical and Magnetic Properties
of Sulfides," in Sulfide Minerology and Geochemistry, vol. Reviews in
Mineralogy and Geochemistry 61, Geochemical Society and Mineralogical
Society of America, 2006, pp. 127-172.
[53] K. M. Rosso and D. J. Vaughn, "Sulfide Mineral Surfaces," in Sulfide Minerology
and Geochemistry, vol. 61, D. J. Vaughn, Ed., Geochemical Society and
Mineralogical Society of America, 2006, pp. 505-550.
[54] H. Wang, "Kinetic Studies of Some Solid-State Reactions of Metal Sulfides,"
Australia, 2005.
119
[55] M. H. Reed and J. Palandri, "Sulfide Mineral Precipitation from Hydrothermal
Fluids," in Sulfide Minerology and Geochemistry , vol. 61, Geochemical Society
and Mineralogical Society of America, 2006, pp. 609-630.
[56] C. Alpers and D. Blowes, "Enviromental Geochemistry of Sulfide Oxidation," in
ACS Symposium series 550, Washington D.C, 1994.
[57] B. Dold, "Basic Concepts of Environmental Geochemistry of Sulfide MineWaste," Latinoamericano de Metalogenia, 2005.
[58] S. R. Shatynski, "Oxidation of Metals," in The Thermochemistry of Transition
Metal Sulfides, vol. 11, 1977.
[59] J. Dunn, "The Oxidation of Sulphide Minerals," Thermochimica Acta, vol. 300,
pp. 127-139, 1997.
[60] K. H. Rao and I. Chernyshova, "Challenges in Sulfide Mineral Processing," The
Open Mineral Processing Journal, vol. 4, pp. 7-13, 2011.
[61] R. Smart, J. Amarantidis, W. Skinner, C. A. Prestidge, L. L. Vanier and S. Grano,
"Surface Analytical Studies of Oxidation and Collector Adsorption in Sulfide
Mineral Flotation," Scanning Microscope, vol. 12, no. 4, pp. 553-583, 1998.
[62] J. Craig and S. Scott, "Sulfide phase equilibria,," in Sulfide Mineralogy, vol. 1, P.
Ribbe, Ed., Mineralogical Society of America Short Course Notes, 1974, pp. CS
1-110.
[63] M. Dekkers, " Magnetic Properties of Natural Pyrrotite Part 1: Behaviour of Initial
Susceptibility and Saturaation Magnetization Related Rock Magnetic Parameters
in a Grain-Size Dependent Framework," Physics of the Earth and Planetary
Interiors, vol. 52, pp. 376-393, 1988.
[64] A. Pratt, H. Nesbitt and J. Mycroft, "The increased reactivity of pyrrhotite and
magnetite phases in sulphide mine tailings," Journal of Geochemical Exploration,
vol. 56, pp. pp.1-11, 1996.
[65] M. Becker, "The Minerology and Crystallography of Pyrrhotite From Selected
Nickel and PGE Ore Deposits and its Effect on Flotation Performance," 2009.
[66] A. Putnis, Introduction to Mineral Sciences, Cambridge University Press, 1992.
[67] R. O. Sacks and D. S. Ebel, "Thermochemistry of Sulfide Mineral Solutions," in
Sulfide Minerology and Geochemistry, vol. 61, Geochemical Society and
Mineralogical Society of America, 2006, pp. 265-349.
[68] J. Lambert, G. Simkovich and J. Walker, "The kinetics and mechanisms of the
pyrite- pyrrhotite transformation," Metallurgical and Materials Transactions B,
vol. 29B, pp. 385-396, 1998.
[69] P. Budka and F. Milillo, " The inverse peritectic phsse transformation in the Fe-S
system: Evidence for the remelting of Troilite during cooling," Meteoritics: The
Journal of the Meteoritical Society, vol. 24, no. 4, pp. 1-7, 1986.
[70] A. Theodossiou, "Measurement of the Hall effect and resistivity in pyrrhotite,"
Physical Review 137, pp. 1321-1326, 1965.
[71] F. Li, "Studies of structures and phase transitions in pyrrhotite," 1997.
120
[72] E. Selivanov, R. Gulyaeva and A. Vershinin, "Thermal Expansion and Phase
Transformations of Natural Pyrrotite," Inorganic Materials, vol. 44, no. 4, pp.
438-442, 2008.
[73] F. Xu and A. Navrotsky, "Enthalpies of formation of pyrrhotite Fe1-0.125xS (0 ≤
x ≤ 1) solid solutions," American Mineralogist, vol. 95, pp. 717-723, 2010.
[74] J. Kondoro, "Mössbauer study of vacancies in natural pyrrhotite," Journal of
Alloys nad Compounds 289, pp. 36-41, 1999.
[75] H. Nam, K. S. Baek and H. J. Kim, "Crystallographic and magnetic properties of
iron-deficient iron sulfide,," Journal of the Korean Physical Society, vol. 25, no. 5,
pp. 447-449, 1992.
[76] N. Blezile, Y.-W. Chen, M.-F. Cai and Y. Li, "A Review on Pyrrhottie
Oxidation," Journal of Geochemical Exploration, vol. 84, pp. 65-76, 2004.
[77] Y. L. Mikhlin, "Disordered surface layers of metal sulfides and their reactivity,"
Glass Physics and Chemistry, vol. 33, no. 4, pp. 402-410, 2007.
[78] Y. L. Mikhlin, Y. V. Tomashevich, G. Pashkov, A. A. V. Okotrub, I. Asanov and
L. Mazalov, "Electronic structure of the non-equalibrium iron-deficient layer of
hexagonal pyrrhotite," Applied Surface Science 125, pp. 73-84, 1998.
[79] M. Wolthers, L. Charlet, P. Vanderlinde, D. Rickard and C. Vanderweijden,
"Surface chemistry of disordered mackinawite (FeS)," Geochimica et
Cosmochimica Acta, vol. 69, no. 14, pp. 3469-3481, 2005.
[80] R. J. Goble, "The relationship between crystal structure, bonding and cell
dimension in the copper sulfides," Canadian Minerologist, vol. 23, pp. 61-76,
1985.
[81] D. J. Vaughn and J. A. Tossell, "The chemical bond and the properties of sulfide
minerals. I. Zn, Fe, and Cu in tetrahedral and triangular coordination with sulfur,"
Canadian Minerologist, vol. 18, pp. 157-163, 1980.
[82] G. Donnay, L. Corliss, J. D. H. Donnay, N. Elliot and J. M. Hastings, "Symmetry
of Magnetic Structures: Magnetic Structure of Chacopyrite,," Physical Review,
vol. 112, no. 6, 1958.
[83] A. Rais, A. M. Gimelseed and A. Al-Rawas, "Magnetic properties of natural
chalcopyrite at low temperatures,," Materials Letters, vol. 46, pp. 349-353, 2000.
[84] T. Teranishi and K. Sato, "Optical, electrical, and magnetic properties of
chalcopyrite, CuFeS2," Journal de Physique, pp. 149-153.
[85] A. Zunger, "New insights on chalcopyrites from solid-state theoy," Thin Solid
Films, no. 515, pp. 6160-6162, 2007.
[86] C. Sahyoun, S. Kingman and N. Rowson, "The Effect of Heat Treatment on
Chalcopyrite," Physical Separation in Science and Engineering, vol. 12, no. 1, pp.
23-30, 2002.
[87] P. Baláž and K. Tkáčová, "Changes in magnetic susceptibility of chalcopyrite
induced by vibration grinding,," Magnetic and Electrical Seperation, vol. 3, pp.
113-115, 1992.
121
[88] M. Lovás, I. Murová, A. Mockovciakoá, N. Rowson and Š. Jakabský,
"Intensification of magnetic separation and leaching of Cu-ores by microwave
radiation," Seperation and Purification Technology, vol. 31, pp. 291-299, 2003.
[89] V. Bakakin and V. Drebushchak, "On the Unusual Arrangement of Metal Atoms
in Pentlandite," Journal of Structural Chemistry, vol. 39, no. 5, 1998.
[90] J. F. Riley, "The pentlandite group (Fe, Ni, Co)9S8: New data and an appraisal of
structure-composition relationships," Mineralogical Magazine, vol. 41, pp. 345349, 1977.
[91] V. Rajamani and C. Prewitt, "Thermal Expansion of the Pentlandite Structure,"
American Mineralogist, vol. 60, pp. 39-48, 1975.
[92] P. Waldner and W. Sitte, "Thermodynamic modeling of Fe-Ni pentlandite,"
Journal of Physics and Chemistry of Solids, vol. 69, pp. 923-927, 2007.
[93] B. Etschmann, A. Pring, A. Putnis, B. A. Grguric and A. Studer, "A Kinetic Study
of the Exsolution of Pentlandite (Ni,Fe)9S8 From the Monosulfide Solid Solution
(Fe,Ni)S," American Minerologist, vol. 89, pp. 39-50, 2004.
[94] H. Wang, A. Pring, Y. Ngothai and B. O’Neill, "A low-temperature kinetic study
of the exsolution of pentlandite from the monosulfide solid solution using a
refined Avrami method," Geochimica et Cosmochimica Acta, vol. 69, no. 2, pp.
415-425, 2005.
[95] V. Drebushchak, T. Kravchenko and V. Pavlyuchenko, "Synthesis of pure
pentlandite in bulk," Journal of Crystal Growth, no. 193, pp. 728-731, 1998.
[96] J. Craig, "Pyrite-Pentlandite Assembaleges and Other Low Temperature Relations
in the Fe-Ni-S system," American Journal of Science, Vols. 273-A, pp. 496-510,
1973.
[97] H. Wang, "Isothermal kinetics of the pentlandite exsolution from mss/pyrrhotite
using model-free method," Tsinghua Science and Technology, vol. 11, no. 3, pp.
368-373, 2006.
[98] C. Francis, M. Fleet and J. C. K. Misra, "Orientation of Exsolved Pentlandite in
Natural and Synthetic Nickeliferous Pyrrhotite," The American Minerologist, vol.
61, pp. 913-920, 1976.
[99] A. Chamberlain and J. G. Dunn, "The pyrolytic decomposition of synthetic
violarites of various stoichiometry," Thermochimica Acta, no. 318, pp. 101-113,
2006.
[100] A. Naldrett and J. M. Duke, "Platinum Metals in Magmatic Sulfide Ore," Science,
vol. 208, no. 4451, pp. 1417-1424, 1980.
[101] D. Ebel and A. Naldrett, "Crystallization of sulfide liquids and the interpretation of
ore composition," Canadian Journal of Earth Science, vol. 34, pp. 352-365, 1997.
[102] C. Ballhaus, M. Tredoux and A. Späth, "Phase Relations in the Fe-Ni-Cu-PGE-S
System at Magmatic Temperature and Application to Massive Sulphide Ores of
the Sudbury Igneous Complex," Journal of Petrology, vol. 42, no. 10, pp. 19111926, 2001.
122
[103] C. Ballhaus, C. Bockrath, C. Wohlgemuth, V. Laurenz and J. Berndt,
"Fractionation of the noble metals by physical processes," Contributions in
Mineral Petrology, vol. 152, pp. 667-684, 2006.
[104] F. Parat, F. Holtz and M. J. Streck, "Sulfur-bearing Magmatic Accessory
Minerals," in Sulfur in Magmas and Melts: Its Importance for Natural and
Technical Processes, vol. 73, Mineralogical Society of America, 2011, pp. 285314.
[105] A. Naldrett, "World-class Ni-Cu-PGE deposits: key factors in their genesis,"
Mineralium Deposita, vol. 34, pp. 227-240, 1999.
[106] J. E. Mungall, "Crystallization and magmatic sulfides: An empirical model and
application to Sudbury ores," Geochimica et Cosmochimica Acta, no. 71, pp.
2809-2819, 2007.
[107] M. E. Fleet, "Phase Equilibria at HighTemperatures," in Sulfide Minerology and
Geochemistry, vol. 61, Reviews in Mineralogy and Geochemistry, Geochemical
Society and Mineralogical Society of America, 2006, pp. 365-405.
[108] V. Raghavan, "Section II: Phase Diagram Evolution Cu-Fe-Ni-S (Copper-IronNickel-Sulfur)," Journal of Phase Equilibria and Diffusion, vol. 25, no. 5, 2004.
[109] A. Simon and E. M. Ripley, "The Role of Magmatic Sulfur in the Formation of
Ore Deposits," in Sulfur in Magmas and Melts: Its Importance for Natural and
Technical Processes, vol. 73, Minerological Society, pp. 513-578.
[110] D. Lindsley, "Oxide Minerals: Petrologic and Magnetic Significance," Reviews in
Minerology, vol. 25, 1991.
[111] G. Waychunas, "Crystal Chemistry of Oxides and Oxyhyroxides," in Reviews in
Minerology, vol. 25, 1991, pp. 11-68.
[112] N. McEwan, T. Courtney, R. Parry and P. Knupfer, "Chromite-A Cost-Effective
Refractory Raw Material for Refractories in Various Metallurgical Applications,"
Southern African Pyrometallurgy, 2011.
[113] R. Sack and M. Ghiorso, "Chromite as a Petrographic Indicator," Reviews in
Minerology, vol. 25, pp. 323-353, 1991.
[114] W. Deer, R. Howe, J. Zussman and ., An Introduction to the Rock-Forming
Minerals, Longman, 1992, pp. 425-433.
[115] R. Pattrick, G. V. D. Laan, C. Henderson, P. Kuiper, E. Dudzik and D. Vaughn,
"Cation site occupancy in spinel ferrites studies by X-ray magnetic circular
dichroism: developing a method for mineralogists," Eur. J. Mineral, vol. 14, pp.
1095-1102, 2002.
[116] B. Burton, "The Interplay of Chemical and Magnetic Order," in Reviews in
Minerology, vol. 25, 1991, pp. 303-319.
[117] S. Banerjee, "Magnetic Properties of Fe-Ti Oxides," in Reviews in Minerology,
vol. 25, 1991, pp. 107-127.
[118] L. Ramajo, A. Cristóbal, P. Botta, J. P. López, M. Reboredo and M. Castro,
"Dielectric and magnetic response of Fe3O4/epoxy composites," Composites: Part
A, vol. 40, pp. 388-393, 2009.
123
[119] B. Wood, J. Nell and A. Woodland, "Macroscopic and Microscopic
Thermodynamic Properties of Oxides," in Reviews in Minerology, vol. 25, 1991,
pp. 265-302.
[120] L. V. Uitert, "Dielectric Properties of and Conductivity in Ferrites," in
Proceedings of the IRE, 1956.
[121] E. Wolska, "Defect Structures in Spinel Ferrites," Defect and Diffusion Forum ,
Vols. 134-135, pp. 89-100, 1996.
[122] D. Bruce, D. O’Hare and R. Walton, Functional Oxides, vol. Inorganic Material
Science, Wiley, 2010.
[123] C. L. Seabrook, "Platinum-Group Minerals in the Raglan Ni-Cu-(PGE) Sulfide
Deposit, Cape Smith, Quebec, Cananda," The Canadian Minerologist, vol. 42, pp.
485-497, 2004.
[124] C. Y. Wang and M. Zhou, "Genesis of the Permian Baimazhai Magamti Ni-Cu(PGE) Sulfide Deposit, Yunnan, SW China,," Miner Deposita, vol. 41, pp. 771783, 2006.
[125] R. Cawthorn, "The Platinum and Palladium Resources of the Bushveld Complex,"
South African Journal of Science, no. 95, 1999.
[126] W. Maier and S. Barnes, "Platinum-Group elements in Silicate Rocks of the
Lower, Critical and Main Zones at Union Section, Western Bushveld Complex,"
Journal of Petrology, vol. 40, no. 11, pp. 1647-1671, 1999.
[127] J. E. Hawley and L. G. Berry, "The Canadian Mineralogist," Michenerite and
Froodite, Palladium Bismuthide Minerals, vol. 6, no. 2, pp. 200-209, 1958.
[128] C. Morrissey, H. Prichard, P. Potts, J. Bowles and S. Cribbs, "Exploration for
Platinum," Geo-Platinum, vol. 87, pp. 1-12, 1988.
[129] J. Miller, J. Li, J. Davidtz and F. Vos, "A Review of Pyrrhotite Floatation
Chemistry in the Processing of PGM Ores," Minerals Engineering, vol. 18, pp.
855-865, 2005.
[130] R. J. Meredith, Engineers Handbook of Industrial Microwave Heating, London:
Institution of Electrical Engineers, 1998.
[131] S. Das, A. K. Mukhopadhyay, S. Datta and D. Basu, "Prospects of Microwave
Processing: An Overview," Bull. Mater. Sci., vol. 31, no. 7, pp. 943-956,, 2008.
[132] K. Haque, "Microwave Energy for Mineral Treatment Process- A Brief Review,"
International Journal of Mineral Processsing, vol. 57, pp. 1-24, 1999.
[133] E. Thostenson and T. W. Chou, "Microwave Processing: Fundamentals and
Applications," Composites: Part A, vol. 30, pp. 1055-1071, 1999.
[134] D. E. Clark, D. C. Folz and J. K. West, "Processing materials with microwave
energy," Material Science and Engineering A, no. 287, pp. 153-158, 2000.
[135] L. Chenpung, X. Yousheng and H. Yixin, "Heating Rate of Minerals and
Compounds in Microwave Field," Transaction of NFsoc, vol. 6, no. 1, 1996.
[136] C. Sahyoun, S. Kingman and N. A. Rowson, "The Effect of Heat Treatment on
Chalcopyrite," Physical Seperation in Science and Engineering, vol. 12, no. 1, pp.
23-30, 2003.
124
[137] C. Lehmann, Interaction of Radiation with Solids and Elementary Defect
Production, North-Holland Publishing Company, 1977.
[138] M. McChesney, Thermodynamics of Electrical Processes, Wiley Interscience,
1971.
[139] A. S. Marfunin, Physics of Minerals and Inorganic Materials, Springer-Verlag,
1979.
[140] U. S. Inan and A. S. Inan, Electromagnetic Waves, Prentice Hall, 2000.
[141] D. C. Dube, R. Zurmuhlen, A. Bell and N. Setter, "Dielectric Measurment on
High-Q Ceramics in the Microwave Region," American Ceramic Society, vol. 80,
no. 5, pp. 1095-1100, 2007.
[142] E. I. Parkhomenko, Electrical Properties of Rocks, Plenum Press, 1967.
[143] R. Chaffin, Microwave Semiconductor Devices: Fundamentals and Radiation
Effects, John Wiley & Sons Inc., 1973.
[144] J. Sheen, "Study of microwave dielectric properties measurments by various
resonance techniques," Measurement, no. 37, pp. 123-130, 2005.
[145] S. Kieffer, "Thermodynamics and Lattice Vibrations of Minerals: 1. Mineral Heat
Capacities and Their Relationships to Simple Lattice Vibration Models," Reviews
of Geophysics and Space Physics, vol. 17, no. 1, pp. 1-20, 1979.
[146] Y. Kobayashi and M. Katoh, "Microwave Measurment of Dielectric Properties of
Low-Loss Materials by the Dielectric Rod Resonator Method," IEEE Transactions
on Microwave Theory and Techniques, vol. 33, no. 7, pp. 586-592, 1985.
[147] Z. Peng, J. Hwang, J. Mouris, R. Hutcheon and X. Huang, "Microwave
Penetration Depth in Materials with Non-zero Magnetic Susceptibility," ISIJ
International, vol. 50, no. 11, pp. 1590-1596, 2010.
[148] Z. Peng, J.-Y. Hwang, M. Andriese, W. Bell, X. Huang and X. Wang, "Numerical
Simulation of Heat Transfer during Microwave Heating of Magnetite," ISIJ
International , vol. 51, no. 6, pp. 884-888, 2011.
[149] F. Agullo-Lopez, C. Catlow and P. Townsend, Point Defects in Materials,
Academic Press, 1988.
[150] H. Tamura, "Microwave dielectric losses caused by lattice defects," Journal of the
European Ceramic Society, vol. 26, pp. 1775-1780, 2006.
[151] M. Kanno, K. Nakamura, E. Kanai, K. Hoki, H. Kono and M. Tanaka,
"Theoretical Verification of Nonthermal Microwave Effects on Intermolecular
Reactions," The Journal of Physical Chemistry, vol. 116, pp. 2177-2183, 2012.
[152] N. Yoshikawa, "Recent Studies on Fundamentals and Application of Microwave
Processing of Materials," in Advances in Induction and Microwave Heating in
Mineral and Organic Materials, InTech, 2011, pp. 1-26.
[153] R. W. Balluffi, S. M. Allen and W. C. Carter, Kinetics of Materials, WileyInterscience, 2005.
[154] Y. Zhang, "Diffusion in Minerals and Melts: Theoretical Background," in
Diffusion in Minerals and Melts, Reviews in Mineralogy and Geochemistry, vol.
72, Mineralogical Society of America, 2010, pp. 5-59.
125
[155] S. A. Freeman, J. H. Booske and R. F. Cooper, "Microwave Field Enhancment of
Charge Transport in Sodium Chloride," Physical Review Letters, vol. 74, no. 11,
pp. 2042-2045, 1995.
[156] M. McChesney, Thermodynamics of Electrical Process, Wiley-Interscience, 1971.
[157] Y. Zheng, S. Wang, J. Feng, Z. Ouyang and X. Li, "Measurment of the complex
permittivity of dry rocks and minerals: application of polythene dilution method
and Lichtenecker’s mixture formulae," Geophysics Journal International, vol.
163, pp. 1195-1202, 2005.
[158] M. Vollmer, "Physics of the microwave oven," Physics Education, vol. 39, no. 1,
pp. 74-81, 2004.
[159] Y. V. Bykov, K. I. Rybakov and V. E. Semenov, "High-temperture microwave
processing of materials," Journal of Physics D: Applied Physics, vol. 34, pp. R55R75, 2001.
[160] K. S. Yee, "Numerical Solution of Initial Boundary Value Problems Involving
Maxwell’s Equations in Isotropic Media," IEEE Transactions on Antennas and
Propagation, vol. 14, no. 3, pp. 302-307, 1966.
[161] D. E. Clark, J. G. P. Binner and D. A. Lewis, "Microwaves: Theory and
Application in Materials Processing V," Ceramic Transactions, vol. 111, 2000.
[162] National Materials Advisory Board Commission on Engineering and Technical
Systems, Microwave Processing of Materials, Washington, D.C.: National
Academy Press Publication NMAB-473, 1994.
[163] C. A. Pickles, "Microwaves in extrative metallurgy: Part 2 - Review of
applications," Minerals Engineering, vol. 22, pp. 1111-1118, 2009.
[164] V. Tikhonov, O. Polyakova, G. Goltsman, A. Dzardanov and D. Boyarskiy,
"Determination of Dielectric Properties of Ore Minerals In The Microwave Band,"
Radiophysics and Quantum Electronics, vol. 51, no. 12, 2008.
[165] J. de los Santos, D. Garcia and J. Eiras, "Dielectric Characterization of Materials
at Microwave Frequency Range," Materials Research, vol. 6, no. 1, pp. 97-101,
2002.
[166] Z. Svindrych, A. Youssef and Z. Janu, "Link between Magnetic and Dielectric
Properties in Magnetite," Acta Physica Polonica A, vol. 118, no. 5, pp. 940-941,
2010.
[167] S. W. Kingman, K. Jackson, A. Cumbane, S. M. Bradshaw, N. A. Rowson and R.
Greenwood, "Recent developments in microwave-assisted comminution,"
International Journal of Mineral Processing, vol. 74, pp. 71-83, 2004.
[168] C. A. Pickles, "Microwaves in extrative metallurgy: Part 1 - Review of
fundamentals," Minerals Engineering, vol. 22, pp. 1102-1111, 2009.
[169] D. E. Clark and W. H. Sutton, "Microwave Processing of Materials," Annual
Review in Material Science, vol. 26, pp. 299-331, 1996.
[170] S. McGill, J. Walkiewicz and A. Clark, "Microwave Heating of Chemicals and
Minerals," Bureau of Mines, 1990.
126
[171] D. A. Jones, T. P. Lelyveld, S. D. Mavrofidis, S. W. Kingman and N. J. Miles,
"Microwave heating applications in enviromental engineering—a review,"
Resources, Conservation, and Recycling, vol. 34, pp. 75-90, 2002.
[172] S. W. Kingman, "Recent developments in microwave processing of minerals,"
International Materials Review, vol. 51, no. 1, pp. 1-12, 2006.
[173] S. M. Javad Koleini and K. Barani, "Microwave Heating Applications in Mineral
Processing," in The Development and Application of Microwave Heating, InTech,
2012, pp. 79-104.
[174] G. M. Liubakka, T. L. Sullivan, D. E. Lamphere, G. J. Corradi and K. J. Dykhuis,
"Separator and Crusher of Minerals with Microwave Energy and Method
Thereof". USA Patent 0263482A1, 21 Oct 2010.
[175] E. Kaya, "Comminution behavior of microwave heated two sulphide copper ores,"
Indian Journal of Chemical Technology, vol. 17, pp. 455-461, 2010.
[176] S. W. Kingman, W. Vorster and N. A. Rowson, "The influence of minerology on
microwave assisted grinding," Minerals Engineering, vol. 13, no. 3, pp. 313-327,
2000.
[177] J. W. Walkiewicz, A. E. Raddatz and S. L. McGill, "Microwave-Assisted
Grinding," Reno Research Center U. S. Bureau of Mines, 1989.
[178] J. B. Salsman, R. L. Williamson, W. K. Tolley and D. A. Rice, "Short-pulse
Microwave Treatment of Disseminated Sulfide Ores," Minerals Engineering, vol.
9, no. 1, pp. 43-54, 1996.
[179] D. Jones, S. Kingman, D. Whittles and I. Lowndes, "Understanding Microwave
Assisted Breakage," Minerals Engineering , vol. 18, pp. 659-669, 2005.
[180] R. Henda, G. R. Hermas and M. R. Islam, "Microwave Enhanced Recovery of
Nickel-Copper Ore: Communition and Floatability Aspects," International
Microwave Power Institute, vol. 40, no. 1, pp. 7-16, 2006.
[181] M. Becker, J. de Villiers and D. Bradshaw, "The flotation of magnetic and nonmagnetic pyrrhotite from selected nickel ore deposits," Minerals Engineering, vol.
7, no. 2, 2007.
[182] S. Marland, B. Han, A. Merchant and N. Rowson, "The effect of microwave
radiation on coal grindability," Fuel, vol. 11, no. 79, pp. 1283-1288, 2000.
[183] J. Lytle, N. Choi and K. Prisbrey, "Influence of preheating on grindability of
coal," International Journal of Mineral Processing, no. 36, pp. 107-112, 1992.
[184] K. Waters, N. Rowson, R. Greenwood and A. Williams, "Characterising the effect
of microwave radiation on the magnetic properties of pyrite," Seperation and
Purification Technology, vol. 56, pp. 9-17, 2007.
[185] S. Marland, B. Han, N. Rowson and A. Merchant, "Microwave Embrittlement and
Desulphurisation of Coal," Acta Material Science, vol. 3, pp. 351-355, 1998.
[186] T. Usla and U. Atalay, "Microwave Heating Characteristics and Magnetic
Seperation Enhancement of Pyrite," in International Mining Congress and
Exhibition of Turkey, 2003.
127
[187] S. H. Weng and J. Wang, "Mossbauer analysis of the microwave desulphurisation
of raw coal," Journal of Applied Physics, vol. 73, no. 9, pp. 239-243, 1993.
[188] D. D. Bluhm, G. E. Fanslow and S. O. Nelson, "Enhanced Magnetic Seperation of
Pyrite from Coal After Microwave Heating," IEEE Transactions on Magnetics,
vol. 22, no. 6, 1986.
[189] W. Vorster, N. A. Rowson and S. W. Kingman, "The effect of microwave
radiation upon the processing of Neves Corvo copper ore," International Journal
of Mineral Processing, vol. 63, pp. 29-44, 2001.
[190] N. Standish and H. Worner, "Microwave Application in the Reduction of Metal
Oxides with Carbon," International Microwave Power Institute, vol. 25, no. 3, pp.
177-180, 1990.
[191] M. Al-Harahsheh and S. W. Kingman, "Microwave-assisted leaching—a review,"
Hydrometallurgy, vol. 73, pp. 189- 203, 2004.
[192] C. A. Pickles, "Microwave heating behaviour of nickeliferous limonitic laterite
ores," Minerals Engineering, vol. 17, pp. 775-784, 2004.
[193] Y.-f. Chang, X.-j. Zhai, Y. Fu, L.-z. Ma, B.-c. Li and T.-a. Zhang, "Phase
Transformation in reductive roasting of laterite ore with microwave heating,"
Transaction of Nonferrous Metals Society of China, vol. 18, pp. 969-973, 2008.
[194] D. K. Xia and C. A. Pickles, "Microwave Caustic Leaching of Electric Arc
Furnace Dust," Minerals Engineering, vol. 13, no. 1, pp. 79-94, 2000.
[195] X. Sun, J. Y. Hwang and X. Huang, "Micorwave Processing of Eletric Arc
Furnace Dust," JOM, vol. 60, no. 10, pp. 26-30, 2008.
[196] J. Vaughn, "The process minerology of gold: The classification of ore types.,"
Journal of Metals, vol. 56, no. 3, pp. 46-48, 2004.
[197] F. Habashi, Principles of Hydrometallurgy, vol. 2, New York: Gorden & Breach,
1970.
[198] V. D. Lunin, A. V. Narseev, N. I. Barahnev and E. V. Ratnikov, "Model of
Process of Microwave Action on Refractory Gold Concentrate," Journal of Mining
Science, vol. 33, no. 4, pp. 373-377, 1997.
[199] S. Kingman and N. Rowson, "Microwave Treatment of Minerals-A Review,"
Minerals Engineering, vol. 11, no. 11, pp. 1081-1087, 1998.
[200] R. Kelly and N. Rowson, "Microwave reduction of oxidised illmenite
concentrates," Minerals Engineering, vol. 8, no. 11, pp. 1427-1438, 1995.
[201] Y. Hua, C. Cai and Y. Cui, "Microwave-enhanced roasting of copper sulfide
concentrate in the presence of CaCO3," Seperation and Purification Technology,
vol. 50, pp. 22-29, 2006.
[202] J.-Y. Hwang, S. Shi, Z. Xu and X. Huang, "Oxygenated Leaching of Copper
Sulfide Mineral under Microwave-Hydrothermal Conditions," Journal of Minerals
& Materials Characterization & Engineering, vol. 1, no. 2, pp. 111-119, 2002.
[203] E. T. Thostenson and T. W. Chou, "Microwave processsing: fundamentals and
applications," Composites: Part A, vol. 30, pp. 1055-1071, 1999.
128
[204] D. K. Agrawal, "Microwave processing of ceramics," Current Opinion in Solid
State & Materials Science, vol. 3, pp. 480- 485, 1998.
[205] C. Clauser and E. Huenges, "Thermal Conductivity of Rocks and Minerals," Rock
Physics and Phase Relations, 1995.
[206] J. A. Tyburczy and D. Fisler, "Electrical Properties of Minerals and Melts,"
Mineral Physics and Crystallography, 1995.
[207] J. Jakubovics, Magnetism and Magnetic Materials, The Bath Press, 1987.
[208] D. H. Tarling and F. Hrouda, The Magnetic Anisotropy of Rocks, Chapman &
Hall, 1993.
[209] D. A. Robinson, "Calculation of the Dielectric Properties of Temperate and
Tropical Soil Minerals from Ion Polarizabilities using the Clausius-Mosotti
Equation," Soil Science Society of America Journal, vol. 68, pp. 1780-1785, 2004.
[210] R. D. Shannon and G. R. Rossman, "Dielectric constants of silicate garnets and the
oxide additivity rule," American Minerologist, vol. 77, pp. 94-100, 1992.
[211] A. C. Lasaga and R. T. Cygan, "Electronic and ionic polarizabilities of silicate
minerals," American Minerologist, vol. 67, pp. 328-334, 1982.
[212] H. Yixan and L. Chunpeng, "Heating Rate of Minerals and Compounds in
Microwave Field," Transactions of NFsoc, vol. 6, no. 1, pp. 35-40, 1996.
[213] G. Wang, P. Radziszewski and J. Ouellet, "Particle modeling simulation of
thermal effects on ore breakage," Computational Materials Science, vol. 43, pp.
892-901, 2008.
[214] Y. Hua, C. Cai and Y. Cai, "Microwave-enhanced roasting of copper sulfide
concentrate in the presence of CaCO3," Seperation and Purification Technology,
vol. 50, pp. 22-29, 2006.
[215] C. H. Hsieh, H. Jain and E. I. Kamitsos, "Correlation between dielectric constant
and chemical structure of sodium silicate glasses," Journal of Applied Physics,
vol. 80, no. 3, pp. 1704-1712, 1996.
[216] A. K. Shukla, A. Mondal and A. Upadhyaya, "Science of Sintering," Numerical
Modeling of Microwave Heating, vol. 42, pp. 99-124, 2010.
[217] P. Bergese, "Specific heat, polarizability and heat conducttion in microwave
heating systems: A nonequalibrium thermodynamic point of view," Acta
Materialia, vol. 54, pp. 1843-1849, 2006.
[218] D. Stuerga and P. Gaillard, "Microwave heating as a new way to induce localized
enhancements of reaction rate: Non-isothermal and heterogeneous kinetics,"
Tetrahedron, vol. 52, no. 15, pp. 5505-5510, 1996.
[219] E. Pert, Y. Carmel, A. Birnboim, T. Olorunyolemi, D. Gershon, J. Calame, I. K.
Lloyd and O. C. Wilson Jr., "Temperature Measurments During Microwave
Processing: The Significance of Thermocouple Effects," American Ceramic
Society, vol. 84, no. 9, pp. 1981-1986, 2001.
[220] A. K. Shukla, A. Mondal and A. Upadhyaya, "Numerical Modeling of Microwave
Heating," Science of Sintering, vol. 42, pp. 99-124, 2010.
129
[221] Z. Peng, J. Y. Hwang, C. L. Park, B. G. Kim, M. Andriese and X. Wang,
"Microwave Permittivity, Permeability, and Absorption Capability of Ferric
Oxide," ISIJ International, vol. 52, no. 9, pp. 1535-1538, 2012.
[222] Z. Peng, J. Y. Hwang, C. L. Park, B. G. Kim, M. Andriese and X. Wang,
"Microwave Permittivity, Permeability, and Absorption Capability of Ferric
Oxide," ISIJ International, vol. 52, no. 9, p. 1535–1538, 2012.
[223] X. Sun, J. Y. Hwang and X. Huang, "the Micorwave Processing of Eletric Arc
Furnace Dust," JOM, vol. 60, no. 10, pp. 26-30, 2008.
[224] G. M. Liubakka, T. L. Sullivan, D. E. Lamphere, G. J. Corradi and K. J. Dykhuis,
"System and Method for Recovering Minerals". USA Patent 0263483A1, 21 Oct
2010.
[225] Y. Hua, C. Cai and Y. Cai, "Microwave-enhanced roasting of copper sulfide
concentrate in the presence of CaCO3," Seperation and Purification Technology,
vol. 50, pp. 22-29, 2006.
[226] Marsh, John, R. Wroe and A. J. Wroe, "REDUCTION PROCESSING OF
METAL-CONTAINING ORES IN THE PRESENCE OF MICROWAVE AND
RF ENERGY". US Patent 0324440, 2009.
[227] Y. Hua, C. Cai and Y. Cui, "Microwave-enhanced roasting of copper sulfide
concentrate in the presence pf CaCO3," Seperation and Purification Technology,
vol. 50, pp. 22-29, 2006.
[228] P. A. Olubambi, J. H. Potgieter, J. Y. Hwang and S. Ndlovu, "Influence of
microwave heating on the processing and dissolution behaviour of low-grade
complex sulphide ores," Hydrometallurgy, vol. 89, pp. 127-135, 2007.
130
Appendix 1: Images
Appendix 1.1 Joel 6400 scanning electron microscope (SEM)
Appendix 1.2 Water cooled electromagnetics for Curie temperature measurements
131
Appendix 1.3 Laboratory setup for treatment of jaw crushed ore particles placed directly
on glass tray (turntable). (1000 W, 2.45 GHz)
Appendix 1.4 Laboratory microwave setup for treatment of gyratory crushed ore particles
in an alumina crucible. (1000 W, 2.45 GHz)
132
Appendix 1.5 Cooling microwave treated ore particles on ceramic plate
Appendix 1.6 Institute of Material Processing (IMP) Laboratory Ball Mill
133
Appendix 1.7 Laboratory Centrifuge manufactured International Clinical Centrifuge
Model CL
Appendix 1.8 Scintag 2000 X-ray diffractometer (XRD). λCuKα =1.5418
134
Appendix 1.9 Magnetite (Mag) and olivine (Ol) enclosed in pyroxene (Px).
Appendix 1.10 Round inclusions of magnetite (Mag) disseminated in pyrrhotite (Po).
135
Appendix 1.11 BSE image blocky inclusions of magnetite (Mag).
Appendix 1.12 BSE image of intergrowths of sulfide minerals pyrrhotite (Po), nickelpyrrhotite (Ni-Po), and pentlandite (Pnt).
136
Appendix 1.13 Optical image of iron-nickel sulfides pyrrhotite (Po) and nickel-pyrrhotite
(Ni-Po) containing a spinel (Spl) inclusion.
Appendix 1.14 BSE image iron-nickel sulfides pyrrhotite (Po), nickel-pyrrhotite (Ni-Po),
and pentlandtite (Pnt) containing some quartz (Qtz).
137
Appendix 1.15 Microwave heating to 400°C in Argon for 5 minutes.
Appendix 1.16 Microwave heating to 800°C in argon for 5 minutes.
138
Appendix 1.17 Microwave heating to 800°C in air for 5 minutes.
Appendix 1.18 SEI image of the silicate flux formed during MW heating in air to 970°C.
139
Appendix 1.19 BSE image of forsterite crystals and sulfide melt droplets dispersed
throughout silicate melt.
Appendix 1.20 BSE image of large forsterite crystals and sulfide melt droplets dispersed
throughout silicate melt.
140
Appendix 1.21 BSE image of olivine (Ol) crystals with melt sulfide (MSS) droplets
dispersed in the silicate melt (Cpx + Po).
Appendix 1.22 BSE image of a fused sample portion.
141
Appendix 1.23 BSE image of a sulfide-silicate melt interface.
Appendix 1.24 BSE image of a sulfide-silicate melt interface and crystal formation.
142
Appendix 2: EDS of Silicate Minerals
Appendix 2.25 Forsterite (Fo) olivine (Mg 20 Fe 2 Si 36 O 42) at. %
Appendix 2.26 Forsterite (Fo) olivine (Mg 32 Fe 2 Si 32 O 34) at. %
143
Appendix 2.27 Diopside (Di) pyroxene (Mg 9 Fe 2 Al 1 Ca 7 Si 20 O 60) at. %
Appendix 2.28 Diopside (Di) pyroxene (Mg 14 Fe 4 Al 1 Ca 1 Si 21 O 60) at. %
144
Appendix 2.29 Enstatite (En) pyroxene (Mg 17 Fe 13 Al 14 Si 21 O 34) at. %
Appendix 2.30 Enstatite (En) pyroxene (Mg 14 Fe 17 Al 15 Si 21 O 34) at. %
145
Appendix 2.31 Ca-Ti pyroxene (Mg 1 Fe 5 Al 3 Ca 24 Ti 10 Si 33 O 62) at. %
Appendix 2.32 Quartz (Si 55 O 45) at. %
146
Appendix 2.33 Feldspar (Mg 9 Fe 25 Al 6 K 3 7 Ti 5 Si 17 O 34) at. %
Appendix 2.34 Chlorite-group mineral. Relative compositions not shown.
147
Appendix 3: EDS of Metallic Minerals
Appendix 3.35 Magnetite (Mag) (Fe 83 O 17) at. %
Appendix 3.36 Chromite (Chr) (Fe 19 Cr 18 Al 15 Mg 9 O 37) at. %
148
Appendix 3.37 Ulvöspinel (Usp) (Fe 36 Ti 31 O 33) at. %
Appendix 3.38 Spinel (Spl) (Fe 33 Cr 28 Ti 3 Al 11 Si 3 O 22) at. %
149
Appendix 3.39 Spinel (Spl) (Fe 24 Cr 20 Ti 8 Mg 3 Al 5 O 32) at. %
Appendix 3.40 Spinel (Spl) containing manganese (Mn)
150
Appendix 3.41 4M-type pyrrhotite (Po) (Fe 47 S 53) at.%
Appendix 3.42 NC-type pyrrhotite (Po) (Fe 34 S 66) at. %
151
Appendix 3.43 Chalcopyrite (Ccp) (Fe 18 Cu 15 S 67) at. %
Appendix 3.44 Pentlandite (Pnt) (Fe 18 Ni 20 S 62) at. %
152
Appendix 3.45 Ni-pyrrhotite (Ni-Po) (Fe 19 Ni 3 S 78) at.%
Appendix 3.46 Galena (Gn) (Pb 57 S 43) at. %.
153
Appendix 3.47 Sphalerite (Sl) (Zn 24 Fe 5 S 71) at. %
Appendix 3.48 Michenerite (Mich) (Pd 34 Bi 32 Te 34) at. %
154
Appendix 3.49 Gold (Au) found in a metal-sulfide melt
155
Appendix 4: Ball Milling Procedure
Ball Milling Experimental Method for Determining Material
Grindability and Work Index
Matthew Andriese
Michigan Technological University
Material Science and Engineering
Institute of Material Processing
Introduction:
Measurement of the work index in ball milling experiments is the industry standard for
determining the energy it takes to reduce particles to a given size. It is has been well
established for many years and numerous papers have been written that include work index
measurement. Most authors tend to express dislike to the timely length of the milling
experiments. Authors from the Iran Institute of Technology have even attempted to create
a shorten version for performing these type of experiments. From statistical approach this
is an incorrect way of examining bulk particle samples and should only be used when
material is limited as time should always be available.
Fred Bond’s paper (Bond, F. C., Crushing and Grinding Calculations: Part 1; British
Chemical Engineering, 1959) briefly outlines the procedure for performing work index
experiments. The main objective during the experiment is to reach a 250% ratio of oversize
to undersize material. Thus the finer mesh used during experiments the more revolutions
of the ball mill it will take to achieve this ratio.
It is important to know that material is lost to the numerous components of the experiment
including the surfaces of the balls, pans during exchange of material, and sieves. Since
material is lost during each grinding cycle, keeping material losses to a minimum is the
key to reliable experiments. The laboratory perfectionist should know loss of material is
inevitable. As always the key to achieving quality results during experiments is being
consistent. The goal of this paper was to clarify some of the ambiguity of the F. Bond paper
so anybody can perform these experiments.
156
Setting-up the ball mill
The mill should be the 12” × 12” × 12” laboratory ball mill. Count and size the balls
according the F. Bonds paper the grinding charge should consist of 285 iron balls weighing
20,125 grams total. The balls are as follows: 43 1.45” balls, 67 1.17” balls, 10 1” balls, 71
0.75” balls, and 94 0.61” balls with a calculated surface area of 842 in2. In practice balls
wear or break creating a variation in size so this is thought to be a rough guideline for
getting started with these experiments. After the balls are counted and sized, they are put
back into the mill. Put 0.5-1 kg of sand into the mill, cover it, and run for 500 revolutions.
Dump the mill and the exteriors of balls will be cleaned of foreign material. The mill can
be arbitrarily re-run with the material being investigated as to “contaminate” the mill and
balls. If this step is not taken, the output material of the work index experiment, especially
fines, will remain in the ball mill and the material massed will be low (>5g).
Appendix 4.50 MTU Institute of Mineral Processing Laboratory Ball Mill for Grinding
Running the ball mill
Material should be poured through a particle splitter to insure a homogenous bulk sample.
To start grinding experiment, put 1000 grams (700cc) of material passing 6 Tyler mesh
material into the ball mill. Place the door and seal (gasket) correctly on by tightening the
bolts by incrementally going back and forth between the two bolts (cross-tighten). Don’t
over tighten the bolts! Plug in the ball mill into electrical outlet and set the revolution
counter back to zero, turn the mill on, and run for 100 revs. It is highly recommended to
WEAR SOUND PROTECTION!!
Dumping the Mill
157
The door is removed, brushed and tapped back into the mill. The content of the mill is
slowly dumped to minimize airborne particles and so the balls don’t rapidly waterfall down
the grate into the hopper. It is easy to lose the smallest balls during the dump as they tend
to bounce out of the hopper when it becomes full. It is best to dump out the balls in steps
to minimize the material that goes into the hopper. Use a brush to tap the grate of excess
material into the collection pan, as well as, brush any loose material from the bottom
portion of the grate. Then, dump ½ the balls back into the mill, shake the hopper, then
dump the other ½ of the balls back into the mill carefully not to pour material that made it
into the hopper back into the mill. Return the hopper. It is good to then tap the bottom
portion of the grate with a wrench or some hard object to jar and remove any hidden
material.
Recovering the Material
Remove the collection pan full of material from underneath the mill and unload it onto
another pan or container. Use a brush to help with this and tap the back of collection pan.
Vibration is very useful in removing material from the collection pan. Now transfer the
material into the sieve and pan setup chosen for measurement of the work index.
Sieving the Ball Milled Material (Pi)
Picking the correct test mesh size is important for producing a reliable result. As stated by
Bond, start with material passing Tyler 6 mesh. But using fine mesh sizes and thus creating
enough small particles in the ball mill to obtain the needed ratio might take many
revolutions depending on the breakage characteristics of the testing material. The ground
material should be classified using the Roe-Tap machine for at least 10min or longer to
separate the ball milled material. The sieve and pan is removed from the Roe-Tap machine
to mass the retained and passing material. In Bonds paper this is called oversize and
undersize respectively. Some oversize material will be trapped in the sieve apertures and
is recoverable by using a horse hair brush on the back of the sieve. This step also helps
keep the apertures of the sieve open and free from blinding during future cycles. The size
of the sieve, in microns, used for testing is denoted by Pi. An example of the table used for
recording data in the laboratory notebook is shown the below table:
158
Appendix 4.51 Example of sample data in a ball mill grindability experiment
Material:
Gyro -6+8
cycle
P1
P2
P3
P4
P5
P6
P7
P8
100 mesh
147 μm
147 μm
147 μm
147 μm
147 μm
147 μm
147 μm
147 μm
undersize
58.78
326.19
314.74
295.12
292.29
286.22
285.42
287.61
oversize
941.85
673.06
685.28
704.6
707.56
713.28
714.13
712.65
total
1000.63
999.25
1000.02
999.72
999.85
999.5
999.55
1000.26
Circ.
load
1602.3
206.3
217.7
238.8
242.1
249.2
250.2
247.8
revs
100
486
425
386
374
365
364
364
Once the both the oversize and undersize material has been massed, the undersize material
is removed and fresh feed material is added to bring the weight back to the original charge.
Removal of the undersize material is the reason why multiple sieves cannot be used when
performing these experiments. Once the charge has been balanced back to its original
weight, it is put back into the ball mill for further grinding.
Calculation of Revolutions for Subsequent Grinding Steps (GDP)
The goal of the experiment is to obtain a 250% mass ratio of oversize to undersize material,
or the circulating load. The number of revolutions needed to obtain this ratio is recalculated
for each consecutive grinding cycle using the new data obtained from the previous cycle.
The simple calculation for determining the amount of revolutions needed for the next cycle
is as follows:
Mass total
# Revolutionsi
×
=
# Revolutions* (1)
3.5
Mass undersize
Where Mtotal is the total mass of material after sieving (oversize + undersize), Revi is the
amount of revolutions for the previous cycle (this will be 100 for the first cycle), and
Mundersize is the mass of undersize material produced by # Revi , and # Rev* is the number
of revolutions for the next grinding cycle. The grinding cycles are continued until the net
grams of sieve undersize produced per mill revolution reaches equilibrium (250%) then
reverses direction of increasing or decreasing [1].
159
Grindability (GDP) calculation
The ball mill grindability, denoted by GDP, is the average of the last three cycles of net
grams of undersize material per ball mill revolution.
Finding (F80) from Feed Material
Finding the amount of 80 per cent passing feed material (F80) is necessary if the starting
material has a wide range of particle sizes. In some cases the range of particle sizes is
narrow making this step rather arbitrary. For example if the starting material passes six
mesh but retained on 8 mesh (-6+8) then 100 percent of the material passes 6 mesh but 0
per cent passes 8. The 80% of feed passing the 6 mesh can be extrapolated graphically from
the particle size vs. percent passing but in contrary this may not predict the 80 percent
passing feed particle size very accurately. Therefore, for a relatively narrow range of
particle sizes, the F80 is assumed to be the particle size, in microns, in which 100% passes.
This may not be true for testing material with a wide range of particle sizes. It is then
necessary to find F80 using multiple sieve sizes.
F80 can be obtained by starting with 1kg of material that accurately represents the sample.
The representative material is then classified using multiple sieves of varying sizes, large
to small aperture openings. The stack of sieves is Roe-Tapped for 10min or more and the
retained material on each sieve is massed. The cumulative percent passing material is then
calculated and graphed on a log-log plot, particle size vs. cumulative percent passing, to
obtain the size in microns of the 80 percent passing feed material (F80).
Finding (P80) from Last Undersize Material
Once the 250% oversize to undersize mass ratio is obtained, the last undersize material is
sieved to find the size, in microns, in which 80 per cent of the material passes denoted P.
Because the passing material is usually not exactly 80 per cent, it is necessary sieve the last
undersize material multiple times using decreasing mesh sizes. Then extrapolate
graphically the particle size in which 80% passes.
Work index (Wi) calculation
When variables Pi , GDP, F80, P80 have been experimentally determined, the equation
below is used to find the work index denoted Wi (kW·h per ton ore),
Wi =
44.5
 10
10 
Pi 0.23Gbp 0.82 
−

 P
F
80
80 

160
Quick Summary
1.) Start by correctly setting up the mill with the 285 iron balls as stated by F. Bond
2.) Mass 700 cm3 of passing 6 mesh material in a graduated cylinder
3.) Grind the starting material 100 revolutions in the mill
4.) Dump the material and transfer it into the sieve/ pan that will be used for the screen
analysis
5.) Roe-Tap the sieve/ pan for 10 minutes or more
6.) Mass the retained (oversize) and passing (undersize) material
7.) Calculate the amount of revolutions needed for the next grinding cycle
8.) Remove the undersize material and add fresh feed material to bring the weight back
to that of the original charge
9.) Ball mill material and repeat steps 1-8 until the circulating load of 250% is reached
and reverses direction of increasing or decreasing
10.) Find the sieve size in which 80% of the last undersize material passes.
11.) Calculate the GDP and the Work index
161
Appendix 5: XRD Raw Data
Appendix 5.52 XRD pattern of the Curie Temperture sample plotted as intensity (CPS)
vs. lattice plane D-spacing (Å).
162
Appendix 5.53 XRD pattern plotted as intensity (CPS) vs. 2θ (Deg.) of roll crushed asreceived (-65+100) sink material with imposed diffraction lines of pyrrhotite (red),
chalcopyrite (blue), pentlandite (green) and ferrospinel (black)
Appendix 5.54 XRD pattern plotted as intensity (CPS) vs. 2θ (Deg.) of roll crushed MW
60s (-65+100) sink material with imposed diffraction lines of pyrrhotite (red), chalcopyrite
(blue), pentlandite (green) and ferrospinel (black)
163
Appendix 5.55 XRD pattern plotted as intensity (CPS) vs. 2θ (Deg.) of roll crushed asreceived (-100+200) sink material with imposed diffraction lines of pyrrhotite (red),
chalcopyrite (blue), pentlandite (green) and ferrospinel (black)
Appendix 5.56 XRD pattern plotted as intensity (CPS) vs. 2θ (Deg.) of roll crushed MW
60s (-100+200) sink material with imposed diffraction lines of pyrrhotite (red),
chalcopyrite (blue), pentlandite (green) and ferrospinel (black)
164
Appendix 5.57 XRD pattern plotted as intensity (CPS) vs. 2θ (Deg.) of roll crushed asreceived (-200+325) sink material with imposed diffraction lines of pyrrhotite (red),
chalcopyrite (blue), pentlandite (green) and ferrospinel (black)
Appendix 5.58 XRD pattern plotted as intensity (CPS) vs. 2θ (Deg.) of roll crushed MW
60s (-200+325) sink material with imposed diffraction lines of pyrrhotite (red),
chalcopyrite (blue), pentlandite (green) and ferrospinel (black)
165
Документ
Категория
Без категории
Просмотров
0
Размер файла
18 073 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа