Content uploaded by Andrew Batchelor
Author content
All content in this area was uploaded by Andrew Batchelor on Oct 13, 2016
Content may be subject to copyright.
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 1 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
PILOT-SCALE INDUSTRIAL MICROWAVE-IR SORTING OF PORPHYRY COPPER ORES
*A.R. Batchelor, R.S. Ferrari-John, J. Katrib, G. Dimitrakis, C. Dodds, S.W. Kingman
Microwave Process Engineering Group
Faculty of Engineering
The University of Nottingham
University Park
Nottingham, United Kingdom, NG7 2RD
(*Corresponding author: andrew.batchelor@nottingham.ac.uk)
ABSTRACT
Sorting technologies have the potential to reduce energy use in comminution by rejecting
uneconomic ore before feed material enters the mill. However, a major challenge for sorting porphyry
copper ores is providing a basis of discrimination between valuable and non-valuable fragments, due to
their complex mineralogies and textural features. Microwave heating in combination with infra-red
detection (MW-IR) has been shown to provide upgraded feedstock for certain ores at bench scale. Selective
heating of copper sulphide minerals compared to the gangue leads to an increased thermal response for
valuable fragments, providing the basis for sorting.
Rapid scale up of MW-IR from bench scale to a 100 t/h continuous pilot plant facility has been
facilitated by the design of microwave choking structures to allow continuous operation at industrially
meaningful throughputs, whilst meeting health and safety and electromagnetic compliance standards, and
the integration of materials handling with electromagnetic systems.
Laboratory batch testing on a pilot scale microwave system has been undertaken to investigate the
thermal response of copper ores under industrial conditions, at microwave energy inputs up to 1kWh/t.
These results provided the operating window and recommended test conditions for the continuous pilot
plant. Sortability curves predicted in the laboratory are in good agreement with separations achieved
during continuous pilot plant operations. The potential microwave sortability of a number of ores has been
studied, allowing informed decisions to be made about potential ore candidates based on mineralogy.
Future development of the technology will be underpinned by the creation and validation of scale up
concepts for 1000 t/h operations. Understanding the value proposition for individual sites/applications will
be central to unlocking the potential economic and environmental benefits of microwave sorting.
KEYWORDS
Microwave, ore, sorting, pilot-scale.
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 2 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
INTRODUCTION
Recent advances in automated sorting technology have seen ore sorting become an increasingly
attractive possibility for the mining industry to help reduce metal specifc processing and waste handling
costs, as well as potentially unlocking previously uneconomic ore resources in the forecasted low-grade
future. However, sensors that have the resolution to discern grade differences in low-grade ores (typically
down to ~0.1 wt%) or have the capacity to process the several thousand tonnes per hour required by the
largest operating mines in an economical manner, particularly on a fragment-by-fragment basis, continue to
be a limiting factor. Microwave treatment followed by infrared thermal imaging (MW-IRT) has been
proposed as an excitation-discrimination technique to facilitate sorting of low-grade ore (Ghosh, Nayak,
Das, & Palit Sagar, 2013; Ghosh, Sharma, Nayak, & Sagar, 2014; John et al., 2015; Sivamohan &
Forssberg, 1991; Van Weert & Kondos, 2007; Van Weert, Kondos, & Gluck, 2009), but until recently has
not been demonstrated at scale.
A bespoke, laboratory-based, high throughput and continuous pilot-scale microwave treatment
system capable of treating up to 100 t/h (instantaneous) of ore in a batch wise fashion was designed and
built at The University of Nottingham. Utilizing the bespoke system, the fragment-by-fragment thermal
response of a variety of porphyry copper ores under different operating conditions was investigated with
the aim of determining the effect on sortability performance across conditions likely to be encountered in
an industrial environment, and to provide recommended settings to the pilot plant operators.
Subsequent to initial laboratory testing, a decision was taken to commission a pilot-scale test
facility at a major porphyry copper mine owned and operated by the project sponsor. The aim of the facility
was to understand and develop know-how surrounding the engineering challenges of microwave sorting at
scale, to compare batch laboratory sortability performance with pilot-scale continuous sorting performance
to validate the testing methodology, and to support project valuation. Pilot plant testing was conducted over
a period of one and a half years, during which time approximately 300 test runs were completed on 11
different ore types with a total of approximately 15,500 tonnes of material processed.
In this paper, we describe the design of the pilot-scale microwave treatment system and the testing
methodology, compare batch laboratory with continuous pilot plant sortability performance, and discuss
potential routes to scale-up for microwave treatment systems in the order of 1,000 t/h.
PILOT-SCALE MICROWAVE TREATMENT SYSTEM
The microwave treatment system was of a belt-based design to enable a direct interface with a
belt-based sorter available off-the-shelf. The belt was made of glass fibre refinforced silicone, with plastic
side skirts and support bars spanning the applicator to prevent sagging of the belt. All internal structures
were microwave transparent (i.e. do not absorb microwave energy) so that they did not interfere with the
electric field distribution inside the applicator or progressively heat over prolonged periods of operation.
The belt was 850 mm wide with an active ore load width of 750 mm. The system was designed to handle
12.7 mm to 76.2 mm (i.e. -3+½”) size fragments, which equated to throughputs ranging from 10-100 t/h
depending on belt speed and ore presentation. The 6:1 top size to bottom size ratio was also chosen as it
approximates the maximum range that industrial sorters can accommodate.
The microwave treatment system is comprised of a one metre long pentagonal applicator (where
heating of the ore load occurs) fed by a variable power 10-100 kW microwave generator split through two
waveguide sections, shown in Figure 1. The resultant microwave treatment energy dose could range from
0.1-10 kWh/t, but was typically tested between 0.5-1.5 kWh/t to keep operating costs low while providing
a suitable temperature rise for discrimination.
The applicator was designed to provide as homogeneous an electric field distribution as possible
across the belt and cumulatively along the direction of travel through which the fragments pass, which
equated to a variation in power density (W/m3) of approximately ±10-15%. A homogeneous field pattern is
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 3 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
essential to ensure that the fragment temperature rise is a function of mineralogy and not predominantly
based on the electric field distribution inside the applicator.
Figure 1 – Microwave treatment system
To facilitate continuous processing with an open microwave system, it was necessary to design
reflective and resistive choking structures that prevent microwave leakage to the environment. The
reflective chokes are corrugated sections that essentially reflect microwave energy back towards the
applicator. The resistive chokes are carbon foam lined boxes on the end of the reflective chokes that absorb
any remaining microwave energy. For a 100 kW input power, the choking structures were designed to
reduce microwave leakage to a level of approximately 0.1 µW, which exceeds both occupational health and
safety, and the more stringent electromagnetic compatability (EMC) regulations. For reference, 0.1 µW is
about one million times less power output than a mobile phone.
METHODOLOGY
Laboratory Apparatus
In the laboratory at the University of Nottingham, between 50 and 150 individual test fragments
were washed, dried at 50 °C, numbered and weighed for each sample. Three size fractions were tested,
namely -76.2+50.8 mm, -50.8+25.4 mm and -25.4+12.7 mm. The test fragments were arrayed on the belt
in either a closely packed monolayer or spaced monolayer with 60% area coverage, mimicking typical
spacings for interrogation on industrial sorters. The closely packed monolayer presentation is illustrated in
Figure 3a.
Ore of the same size and type as the test sample was added to the front and tail of the test sample
array to ensure the test sample could be treated at steady state with the microwave start-up and shut-down
procedure, illustrated in Figure 2. Following microwave treatment, the test fragments were rolled out of the
applicator and positioned under an NEC Avio H2640 thermal imaging camera and thermal images captured
at one frame per second for two minutes, illustrated in Figure 3b. The first thermal image was typically
captured between 10-30 seconds post-treatment depending on the belt speed used. The first thermal image
was also typically used for calculating sortability performance as this was approximately the same thermal
equilibration time likely to be encounterd in the pilot plant.
Direction of Travel
Through Applicator
Waveguide
(Microwave Input)
Reflective
Chokes
Ore Load
Belt
Applicator
Side Skirts
250 mm
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 4 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
The average temperature of each fragment was determined by manually tracing polygons around
fragment boundaries using the Radiometric Complete Online thermal imaging software. A pre-treatment
thermal image was also captured allowing the average temperature rise (T, °C) to be determined.
Figure 2 – Microwave treatment procedure
Following microwave treatments, the test fragments were subsequently crushed, pulverised and
assayed for copper, iron and sulphur content by x-ray fluorescence (XRF). Moisture content was also
determined by loss of weight after placing a one gram sample in an oven at 105°C overnight. Sortability
curves in the form of cumulative mass recovery to accepts versus cumulative copper recovery to accepts
could then be constructed and potential separations based on a T cut-point calculated.
The test conditions investigated included microwave power input (kW), belt speed (m/s), fragment
spacing (kg/m), microwave treatment energy dose (kWh/t) and post-treatment thermal equilibration time.
Fragment orientation (fragment surface presented to the thermal camera) and position in the cavity were
also investigated as a random orientation and position would be experienced in the pilot plant, and it was
necessary to gauge the sortability performance variation introduced by these two factors from laboratory
testing. This was achieved by conducting a bootstrap analysis with 1,000 replicates while randomly
varying the average temperature rise determined from each separate orientation and position test condition
under the same microwave treatment conditions. The 95% confidence intervals could then be plotted to
provide an operating window within which the pilot plant would be expected to operate.
[a] [b]
Figure 3 – Example [a] test fragment layout and [b] thermal image after microwave treatment
4) Thermal Imaging
3) Post-Treatment: Microwaves Off
2) Steady-State Treatment
1) Pre-Treatment: Microwaves On (Ramp Up)
Applicator Chokes Thermal Camera
Conveyor Belt
Test Sample
Fill
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 5 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
Pilot Plant
The pilot plant microwave treatment system had the exact same specification as the laboratory
system, apart from utilizing a slightly different microwave frequency (915 MHz) due to the plant being
located in a region that uses a different industrial, scientific and medical (ISM) band to the UK (896 MHz)
(Meredith, 1998). The major difference was the extensive feed storage and preparation facilities,
integration of a commercially available automated sorter, and the associated materials handling, process
control, sampling, sample preparation and other facilities that would enable continuous processing for
several hours of operation covering a typical working shift.
The pilot plant is located on the host mine site adjacent to the mill. Run-of-mine (ROM) ore is
diverted from the primary crusher and taken by truck to a mobile crushing and screening facility where it is
prepared into the three size classes that were also used in the laboratory. The sized ore is then taken by
truck to the pilot plant stockpile. Ore is transferred to the feed loading bay by a loader where it is first
passed through a sampling station before being conveyed into the main building. Upon entering the
building the ore is washed and partially blow dried with compressed air over a vibrating screen to remove
dust and chips before being stored in an insulated feed conditioning bin to equilibrate the fragment
temperature.
The subsequent microwave treatment and sorting process is illustrated in Figure 4. The ore is fed
by a belt feeder, over a vibrating screen to remove any remaining chips, to the microwave treatment belt at
a pre-determined throughput and immediately imaged under an infrared thermal camera to capture the feed
temperature distribution. The material passes through the microwave cavity at a pre-determined power
level and belt speed to receive the required treatment energy. Upon exiting the cavity, the material passes
over an intermediary belt to accelerate the fragments before moving on to the sorter belt moving at a higher
speed to allow the fragments to spread out to the required spacing for interrogation by the post-microwave
treatment infrared thermal camera and laser scanner for fragment tracking. The fragment thermal
equilibration time was in the order of 10 to 30 seconds depending on the microwave treatment belt speed
employed. In the current arrangement, only the top-facing fragment surface was interrogated by the
thermal camera. The pre-determined sorting decision is implemented and the pneumatic actuators blast the
lowest mass fraction (either accepts or rejects, although for the purposes of illustration rejects (or cold
fragments) have been used) to save energy and minimise work rate. The two product streams are conveyed
though product sampling stations to retention bays from where they may be collected for re-use or
transfered to the ROM stockpile at the mill. The two products were assayed for copper and the sortability
performance determined.
Figure 4 – Microwave treatment and sorting system
Feed Conditioning Bin
Microwave
Cavity
Sorter
Control
Sorting
Belt
Air
Ejectors
Microwave
Treatment
Belt
IRT
Camera
IRT
Camera
Feeder
Acceleration
Belt
Microwave
Generator
Tuner
Waveguide
100kW
Screen
Laser
Scanner
Rejects
(Waste) Accepts
(Ore)
Splitter Plate
Microwave
Treatment
Control
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 6 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
The pilot plant typically ran two types of test. Setup tests were run for a period approximately 30
minutes to determine how the ore responded to treatment at the recommended settings from the laboratory
testing and to make adjustments to the discrimination and separation settings. A longer run of up to 6 hours
was subsequently conducted to assess system and sortability performance across a shift.
Ore Samples
The ore samples used throughout these investigations were sourced from two major porphyry
copper mines, one of which was the pilot plant host mine site owned and operated by the project sponsor.
In the laboratory, 16 ore types of differing lithology were selected, 8 from each mine site. The pilot plant
trials only utilised ore from the host mine site. All of the ores contained chalcopyrite as the dominant
copper sulphide mineral and pyrite as the dominant sulphide gangue mineral, both of which are good
microwave-heating minerals. Other microwave-heating phases included hydrated smectite clays (classified
as montmorillonite) and typically poorly heating iron oxides, such as hematite. The non-microwave-
heating gangue minerals were dominated by quartz, feldspars and micas, but some ores had significant
proportions of pyroxenes, amphiboles, garnets and carbonates.
RESULTS & DISCUSSION
Laboratory Testing
Fragment spacing, belt speed and microwave power input demonstrated no singly discernible
effect on average temperature rise. Instead it was found that microwave treatment energy dose was the
driving force behind the ultimate temperature rise fragments experienced. Thermal equilibration time up to
two mintues post-microwave treatment also demonstrated little difference in T, indicating that sortability
performance would not be detrimentally influenced by longer equilibration times potentially encountered
in an industrial scale system.
Fragment orientation and position in the cavity did yield some variability in average temperature
rise, particularly for the hotter fragments with surface or near-surface mineralisation on one side of the
fragment, but did not greatly change the sortability performance because it did not make a hot fragment
cold, or vice versa. The temperature rise was largely driven by fragment mineralogy and it was found that
the presence of microwave-heating gangue minerals, particularly moisture content introduced
hygroscopically and/or contained in hydrated clays, was the main source of deviation from intrinsic, or
ideal, sortability performance.
Figure 5 illustrates the thermal response versus different component assays for an ore sample with
high pyrite content, -50.8+25.4 mm in size and treated at 1 kWh/t. The greater influence of moisture
content on T for low sulphide content fragments and the influence of high sulphide content on T are
readily seen. Figure 6 illustrates the sortability curve for the same ore type but in the -76.2+50.8 mm size
class and treated at 0.75 kWh/t. The intrinsic sortability is calculated by cumulating mass and copper
recovery from the highest to lowest grade fragment. The MW-IRT sortability is calculated by cumulating
from the hottest to the coldest fragment. Sortability performance may then be calculated from Eq. (1) at
given mass rejections (Tucker, Morrison & Wellwood, 2013):
PXX = ((AXX – 1) / (IXX – 1)) x 100%
(1)
where, PXX is the sortability performance index (i.e. performance of the actual excitation/detection
system against the potential of the intrinsic sortability), IXX is the intrinsic grade concentration factor
(defined as the grade of the richest XX% by mass divided by the feed grade), and AXX is the actual grade
concentration factor (defined as the grade of the hottest XX% by mass divided by the feed grade).
Approximately 5-10% of the ores and fragment sizes tested in the laboratory resulted in a
performance index of >80% across mass rejections of 25%, 50% and 75%. These ores tended to have a low
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 7 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
moisture content, some co-mineralisation of copper and iron sulphides and/or a high copper sulphide
content relative to other microwave-heating gangue minerals. Approximately 55-60% of the samples tested
resulted in a performance index of 40-80%, which typically contained some lesser contribution of the
previously mentioned beneficial mineralogical features. Despite the lower sortability performance at the
arbitrary mass rejections surveyed, it was usually possible to define a thermal cut-point to provide an
attractive separation to yield a rejects fraction with <0.1 %Cu and a high copper recovery to the accepts.
Approximately 30-35% of the samples tested resulted in a performance index of <40%. These ores
typically had a high moisture and iron sulphide content compared to copper.
Figure 5 – Example assay versus average temperature rise
It was further found that test samples comprised of 150 fragments provided a better estimation of
the sortability performance in the pilot plant than test samples comprised of 50 fragments, as expected due
to better representivity of the ore under test. A larger test sample also typically provided a narrower
operating window due to accounting for a wider variety of thermal repsonses and a more contiguous range
of individual fragment copper content values.
0
2
4
6
8
10
12
0.0 0.5 1.0 1.5 2.0 2.5
Average Temperature Rise ( C)
Copper Content (%)
0
2
4
6
8
10
12
0 5 10 15 20 25 30
Average Temperature Rise ( C)
Sulphur Content (%)
0
2
4
6
8
10
12
0.0 0.5 1.0 1.5 2.0 2.5
Average Temperature Rise ( C)
Moisture Content (%)
0
2
4
6
8
10
12
0 5 10 15 20 25 30
Average Temperature Rise ( C)
Iron Content (%)
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 8 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
Figure 6 – Example laboratory sortability curves with corresponding pilot plant runs
Pilot Plant Testing
It can be seen from the example in Figure 6 that the setup and plant runs typically fell within the
operating window defined by the laboratory experiments, with the longer runs usually closer to the
laboratory predicted curve after fine tuning of the discrimination and separation settings.
Figure 7 shows the copper recoveries of all the pilot plant runs, for all ore types and fragment
sizes, compared to their corresponding laboratory experiments. Approximately 42% of the better optimised
pilot plant runs predicted copper recovery to within ±5% and approximately 84% of the runs to within
±10%. These figures were improved to approximately 50% predicted to within ±5% and approximately
90% to within ±10% if the -25.4+12.7 mm size class was omitted, as this size class frequently suffered
from fragment clustering on the sorting belt, which remained unresolved from commissioning activities.
Furthermore, it is fully expected that better predictions would result from larger laboratory sample sizes
during any future testing campaigns.
ROUTES TO SCALE-UP FOR MICROWAVE-IR SORTING SYSTEMS
The primary route to scale-up for microwave systems is to use frequency scaling. As the
microwave frequency reduces, the wavelength increases and larger applicators can be constructed to
provide homogeneous electric field patterns. The ISM frequencies used for industrial heating applications
include 2,450 MHz, 896/915/922 MHz or 433/350 MHz (depending on location around the world). While
896/915/922MHz facilitated the pilot-scale system in these investigations, 433/350 MHz would enable
applicators to handle throughputs in the order of 1,000 t/h. For a dose of approximately 1 kWh/t to provide
the requisite thermal response, 1 MW of microwave power would thus be required. However, a 1 MW
microwave generator for heating applications does not currently exist and this would require significant
development. Nevertheless, it would also be possible to design larger applicators that use multiple 100 kW
generators at 896/915/922 MHz. Careful design of the applicator would be paramount to provide the
homogeneous electric field patterns required for sorting applications.
0
10
20
30
40
50
60
70
80
90
100
010 20 30 40 50 60 70 80 90 100
Cumulative Copper Recovery to Accepts (%)
Cumulative Mass Recovery to Accepts (%)
Lab Intrinsic
Lab MW-IRT
Splitter
95%CI (Orientation)
95%CI (Orientation+Position)
Pilot Plant (Setup)
Pilot Plant (Run)
Blast Hot Blast Cold
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 9 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
Figure 7 – Laboratory predicted copper recovery versus pilot plant copper recovery for a given pilot plant
mass recovery
Furthermore, pilot-scale testing in these investigations identified many drawbacks to using a belt-
based system. Primarily, a more robust materials handling system would be required for any industrial
system as the silicone belt was prone to damage over prolonged usage and more hard-wearing belting
options surveyed were not suitably microwave transparent. Build up of dust in the applicator also caused
problems with intermittent arcing (electrical breakdown) damaging the belt, plastic support structures and
the applicator. A relatively simple and scalable solution would be to move to gravity-fed flow through a
tube, which has the added benefit of a much reduced plant footprint. Many ceramics already currently used
as wear liners are suitably microwave transparent and such a system would effectively isolate any dust
from the applicator space, preventing many of the problems associated with arcing. Utilising the same
pentagonal applicator shape, we have come up with some conceptual designs of what a 100 t/h and 1,000
t/h microwave treatment system might look like, illustrated in Figure 8.
Figure 8 – Concept for 1,000 t/h gravity-fed microwave treatment system
10
20
30
40
50
60
70
80
90
100
10 20 30 40 50 60 70 80 90 100
Pilot Plant Copper Recovery (%)
Laboratory Predicted Copper Recovery (%)
Setup
Run
0%
±5%
±10%
±15%
IMPC 2016: XXVIII International Mineral Processing Conference Proceedings – ISBN: 978-1-926872-29-2
Page 10 of 10
Published by the Canadian Institute of Mining, Metallurgy and Petroleum
CONCLUSIONS
A pilot-scale microwave treatment system capable of treating 10-100t/h of material at 10-100kW,
yielding a microwave treatment energy dose range of 0.1-10kWh/t, was designed, constructed and
commissioned in both a laborated-based environment and a pilot plant.
In summary, it was found that microwave treatment energy dose was the driving force behind the
ultimate temperature rise fragments experienced and that the presence of microwave-heating gangue
minerals was the main source of deviation from intrinsic, or ideal, sortability performance. A microwave
treatment energy dose of approximately 1 kWh/t was typically sufficient to provide a basis for
discrimination.
A method was developed to enable comparison of batch laboratory with continuous pilot plant
sortability performance by defining operating windows based on expected thermal variability from ore
presentation to the microwave treatment system. Laboratory testing on 50 to 150 fragments allowed the
pilot plant sortability performance to be predicted to within ±5-10% for the majority of samples tested.
Many of engineering challenges of microwave-IR sorting at scale were observed and addressed in
the pilot plant, and a potential route to scale-up has been identified in a gravity-fed tube that would allow
scalable design of microwave treatment systems to handle throughputs in the order of 1,000 t/h required by
the mining industry.
ACKNOWLEDGEMENTS
The authors would like to acknowledge Rio Tinto Technology and Innovation along with their
research and industry partners for engagement and collaboration throughout the Copper NuWave™
project. REFERENCES
Ghosh, A., Nayak, B., Das, T. K., & Palit Sagar, S. (2013). A non-invasive technique for sorting
of alumina-rich iron ores. Minerals Engineering, 45, 55-58. doi:
http://dx.doi.org/10.1016/j.mineng.2012.12.004
Ghosh, A., Sharma, A. K., Nayak, B., & Sagar, S. P. (2014). Infrared thermography: An approach
for iron ore gradation. Minerals Engineering, 62, 85-90. doi:
http://dx.doi.org/10.1016/j.mineng.2013.12.002
John, R. S., Batchelor, A. R., Ivanov, D., Udoudo, O. B., Jones, D. A., Dodds, C., & Kingman, S.
W. (2015). Understanding microwave induced sorting of porphyry copper ores. Minerals Engineering, 84,
77-87. doi: http://dx.doi.org/10.1016/j.mineng.2015.10.006
Meredith, R. J. (1998). Engineers' Handbook of Industrial Microwave Processing. London, UK:
The Institution of Electrical Engineers.
Sivamohan, R., & Forssberg, E. (1991). Electronic sorting and other preconcentration methods.
Minerals Engineering, 4(7–11), 797-814. doi: http://dx.doi.org/10.1016/0892-6875(91)90066-5
Tucker, J., Morrison, R., & Wellwood, G. (2013). The development of indices to assess both the
sorting potential of an ore and the performance of any sorting process when treating that ore. Physical
Separation '13 (pp. 1-12), Falmouth, Cornwall, UK.
Van Weert, G., & Kondos, P. (2007). Infrared recognition of high sulphide and carbonaceous
rocks after microwave heating. 39th Annual Meeting of the Canadian Mineral Processors. Ottowa,
Ontario, Canada.
Van Weert, G., Kondos, P., & Gluck, E. (2009). Upgrading molybdenite ores between mine and
mill using microwave/infrared (MW/IR) sorting technology. 41st Annual Meeting of the Canadian Mineral
Processors (pp. 345-363). Ottowa, Ontario, Canada.
