Energy Conversion and Management 249 (2021) 114832

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Integration assessment of the hybrid sulphur cycle with a copper

production plant
Ahmad Seyfaee a, *, Mehdi Jafarian a, *, Gkiokchan Moumin b, Dennis Thomey b,
Claudio Corgnale c, Christian Sattler b, Graham J. Nathan a
a
Centre for Energy Technology, School of Mechanical Engineering, University of Adelaide, Australia
b
Institute of Future Fuels, German Aerospace Center (DLR), Germany
c
Greenway Energy, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Copper is the third-most widely-used metal worldwide. However, copper processing is an energy-intensive
Hydrogen production process consuming large quantities of fossil fuels, both as the reducing agent and for energy which contrib­
Sulphur cycle utes significantly to anthropogenic carbon dioxide emissions. The hybrid sulphur cycle combines concentrating
Copper
solar thermal energy with electrolysis to offer strong potential for low-cost green hydrogen production. A pre­
Solar energy
liminary evaluation is reported of the techno-economic potential of this cycle to displace current fossil-based
Carbon dioxide emissions
energy sources in an integrated copper mine and refinery (cradle-to-gate approach) at a remote location in
Australia with an excellent solar resource. The effect of ore composition on the integration of the hybrid sulphur
cycle and copper processing plant is evaluated using models developed in Aspen Plus. The evaluations show that
sizing the hybrid sulphur cycle cycle to meet the oxygen demand of the copper refining process is more techno-
economically viable than sizing the hybrid sulphur cycle to meet the hydrogen required to replace the fossil fuel
demand of the copper processing process. Moreover, it has been found that the integration of the hybrid sulphur
cycle with the copper process plant has the potential to decrease both the carbon dixoide emissions and the
operational expenditure of copper refineries for ores with a sufficiently high sulphide content (~50% mass
fraction).

1. Introduction ground in a crushing unit [5] and then separated, typically via a floa­
tation unit, into sulphide-containing (chalcocite (Cu2S), covellite (CuS)
There is growing interest in possible pathways to cost-effectively and chalcopyrite (CuFeS2)), and oxide-containing (tenorite (CuO) and
decarbonise the production of copper, which is the third-most widely- cuprite (Cu2O)) products. The sulphide ore is processed via a pyromet­
used metal with a current global demand of 18 million metric tonnes per allurgical pathway, in which the heat of combustion of a hydrocarbon
year [1]. Currently, most of the energy required for copper processing is fuel (e.g. liquefied petroleum gas or LPG) in an oxygen-enriched envi­
supplied from fossil fuels. Typically, 3.5–8 tonnes of carbon dioxide ronment is employed to produce industrially pure copper and sulphur
(CO2) are produced per tonne of copper product, depending on the dioxide (SO2) in smelters and fire refiners (the direct to blister method).
composition of the initial raw materials and its refinery process [2]. In Hence, the pyrometallurgical pathway requires, in addition to the fossil
particular the CO2 emissions through the copper refining increase as the fuel, oxygen enriched air, which is typically produced via an air sepa­
copper content of the ore decreases [3]. Hence, as the grade of the ration unit (ASU), e.g. by cryogenic air separation or pressure swing
copper resources continues to fall, it is estimated that the specific CO2 adsorption (PSA). The copper-containing oxide ore stream (Cu2O and
emission will rise to more than 10 tonnes of CO2 per tonne of copper CuO) is treated via the hydrometallurgical pathway, in which sulphuric
(Cu) product by 2033 [1,4]. For this reason, there is a strong need to acid is employed to dissolve copper oxides and other traces of low
identify cost-effective pathways with strong potential to mitigate the content metals, e.g. uranium oxides. The dissolved copper ions are
life-cycle CO2 emissions associated with copper refining. extracted via a solvent extraction process (SE) and deposited using the
In the copper refining process (Fig. 1 (a)), the mined rock is first electrowinning process [6–8]. Typically part or all of the required

* Corresponding authors.
E-mail addresses: Ahmad.Seyfaee@Adelaide.edu.au (A. Seyfaee), Mehdi.Jafarian@Adelaide.edu.au (M. Jafarian).

https://doi.org/10.1016/j.enconman.2021.114832
Received 15 June 2021; Accepted 28 September 2021
Available online 23 October 2021
0196-8904/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 1. Schematic diagrams of (a) copper refining process including pyrometallurgy and hydrometallurgy pathways [5,21,22] (b) HyS cycle for H2 and O2 pro­
duction [23].

sulphuric acid in the hydrometallurgical pathway is produced in the acid source, e.g. concentrated solar thermal energy, and electricity. The HyS
plant, where the SO2 off-gas from the furnaces in the pyrometallurgy cycle has already been demonstrated at technology readiness level (TRL)
pathway (e.g. smelters) is converted into sulphuric acid with the aid of a 5 by the German Aerospace Center, DLR, and has been identified pre­
fuel (e.g. fuel oil) and oxygen from ASU unit. viously as a cycle of highest priority for research and development due
One potential approach to mitigate the CO2 emissions from the to its potential for future realisation at an industrial scale [11] and its
copper processing plants is to replace both the hydrocarbon fuel and simple chemistry [12]. As shown in Fig. 1(b), the HyS cycle comprises
oxygen (which is usually supplied by either air separation units (ASU) or two chemical reactions In the first stage, sulphuric acid is decomposed
pressure swing adsorption (PSA) technology) demand with high-grade catalytically in a highly endothermic reaction at 600–1200 ◦ C, suitable
hydrogen (H2) and oxygen (O2) supplied by water electrolysers, pro­ for the integration of concentrated solar thermal energy [13]. Then, the
vided that their energy requirement is supplied from renewable re­ sulphur dioxide product is electrolysed together with water in the sec­
sources. This has recently been ivestigated at the Atacama Desert in ond reaction at temperatures on the order of 100 ◦ C, requiring a mini­
Chile by Gallardo et al. [9] for alkaline or polymer electrolyte membrane mum theoretical voltage of 0.17 V, which is only about one-seventh of
electrolysers with photovoltaic (PV) cells and concentrated solar ther­ the theoretical value of 1.23 V for conventional water electrolysis
mal plants (CSP) with thermal energy storage (TES). They studied the [14,15]. A practical and realistic efficiency of ~30% (based on the lower
potential for the export of produced H2 in remote areas and showed that heating value of H2) is predicted for the HyS cycle [16]. Key drivers for
the combination of PV and electrolyser could be economically viable if a its development are the much lower requirement for electricity, which
buffer such as hydrogen storage is available. Even more recently the reduces the reliance on the relatively expensive storage of either elec­
techno-economic potential of decarbonising the copper production with tricity or hydrogen for continuous hydrogen production, together with
H2 supplied through retrofitting of water electrolyses to copper pro­ the capacity to utilise low-cost thermal energy storage in the high-
cessing has been explored [10] and it was found that under the current temperature section of the cycle [17]. In other words, one part of the
techno-economic parameters for Germany, the cost of CO2 abatement is electricity demand of the conventional water electrolysis is replaced by a
approximately 326 AuD/t CO2-eq. An alternative method to co-produce thermal energy demand, which is much cheaper to produce and store
hydrogen and oxygen with renewable energy from water is the Hybrid [18,19]. The need for local storage solutions results from the large
Sulphur (HyS) cycle, which is a hybrid between a high-temperature heat number of copper mines worldwide being located in remote areas, e. g.

2
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 2. Schematic representation of the key components of the proposed integration of the copper production plant with the HyS cycle.

Olympic Dam (OD) in Australia, Escondida Copper Mine in Chile, where concentrated solar radiation can be the rotary drum particles receiver
solar energy is abundant, but fuel logistics can be challenging. CentRec, which has been developed and demonstrated by DLR [13,17].
Furthermore, there is potential to identify synergy in the sulphur cycle, Hot particles from the storage are sent to an indirectly heated Bayonet
with SO2 being used in the HyS cycle as the circulating agent between reactor to drive the endothermic splitting of sulphuric acid (H2SO4) into
the high-temperature sulphuric acid (H2SO4) cracking reactor and SO2 and O2 (Fig. 2) [28]. The relatively cold particles from the Bayonet
sulphur dioxide depolarised electrolyser (SDE), denoted in Fig. 1(a) as reactor are then stored in a Cold Reservoir (Fig. 2) prior to re-heating.
“Electrolyser”, while also being a byproduct of the pyrometallurgical The electrolysis part of the HyS cycle is the same as the conventional
pathway. This can offer the potential to lower cost through compatibility HyS cycle described above [29]. Here, it is assumed that its electricity
of construction materials, operation, and the availability of makeup demand is to be supplied from the electrical network. The hydrogen and
streams. A comprehensive review of the HyS cycle, reporting the main oxygen produced by the HyS cycle are used in the copper refining pro­
opportunities and challenges of this technology up to 2020, has been cess in one of two scenarios, as follows:
recently published by Corgnale et al. [20]. 1. The HyS cycle is sized to produce enough H2 to meet the hydro­
Nevertheless, to the best of the authors’ knowledge, the potential carbon (HC) fuel demand of the pyro pathway and sulphuric acid pro­
benefits and challenges of integrating the HyS cycle with the copper duction plant. In this scenario, the O2 product from the HyS cycle is also
refining process has not yet been reported. In addition no data are used in the copper refinery process, reducing the size of the ASU plant.
available of the economic potential of such integration in terms of 2. The HyS cycle is sized to meet the O2 demand of the copper pro­
operational expenditure (OPEX) and capital costs (CAPEX). Therefore, duction process, hence mitigating the dependency on air enrichment
the objective of the current work is to assess the high-level thermody­ units (such as ASU or PSA units). In Scenario 2, the H2 product from the
namic potential and techno-economic feasibility of the integration of the HyS cycle is also used to replace the HC fuel in the pyrometallurgy
HyS cycle and copper processing. Furthermore, the molar ratio of H2 to pathway and acid plant, while the excess hydrogen is given a reference
O2 in water-splitting technology is 2, while copper refineries do not value. In practice, the excess hydrogen could be used to energise some of
necessarily require the same ratio. Therefore, the paper also aims to the mine trucks, to produce electricity or be sold as revenue. It should be
assess different scenarios with the view to find those with the highest noted that the oxygen demand in the copper production process sur­
techno-economic potential for the integration of the HyS cycle and passes that of the fuel. Therefore, if the total oxygen demand is to be
copper refining process. supplied via electrolysis, which produces 8 kg O2/kg H2, it will not be
enough for the copper processing. Therefore, Scenario 1 also requires
2. Description of the proposed processes some oxygen to be supplied from the ASU (shown as dashed arrows in
Fig. 2). However, given sufficient storage, if the HyS cycle is sized to
Fig. 2 presents a schematic diagram of the proposed process supply the total oxygen demand of the copper plant (Scenario 2), it
configuration for the integration of the HyS cycle (Fig. 1(b)) with a avoids the need for an ASU with its associated demands for energy.
copper refining process (Fig. 1(a)). This potential process configuration Fig. 2 also shows a makeup supply of H2SO4, which is drawn as a
comprises two main sections: (i) the HyS cycle and (ii) the copper pro­ dashed arrow. This could potentially be sourced from the outside of the
cessing plant. The configuration of the HyS cycle is based on the DLR’s plant, depending on demand. Since the extraction of metals other than
and Greenway Energy and Savannah River National Laboratory design. copper is not considered in the current work, the amount of the required
In this system, particles are circulated through a high-temperature solar sulphuric acid depends only on the amount of the oxide-containing
receiver when the available concentrated solar radiation is sufficient to copper ores through the leaching and extraction process.
exceed the losses [24] and then sent to a Hot Reservoir for storage
[25–27]. One plausible reactor configuration to heat the particles with

3
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Table 1 plant was varied from 0.55 to 0.75.

The five copper-containing ore compositions assessed in the study. As a general approach in the simulation, the copper process and the
Ore sulphide content (wt%) 10% 25% 50% 75% 90% HyS cycle plant were run simultaneously for each ore composition to
solve (automatically) for the size of the HyS cycle (via using multiple
Cu (wt%) produced via 17% 31% 54% 77% 91%
pyrometallurgy
Calculators and Design Spec units) by altering water input and circu­
lating acid to meet the calculated demand from the copper plant.
CuFeS2 0.0300 0.0400 0.0500 0.0700 0.0800
CuS 0.0984 0.0984 0.0984 0.0984 0.0984
Cu2S 0.1123 0.4087 0.8823 1.3566 1.6357 3.1. Input ore composition
Cu2O 0.9353 0.6288 0.1452 0.1294 0.0779
CuO 0.8241 0.8241 0.8241 0.3455 0.1080
Sum (wt%) 2.0000 2.0000 2.0000 2.0000 2.0000 The input ore was assumed to comprise copper-containing and non-
copper-containing minerals. The copper-containing minerals were cho­
sen from the most common copper ores, namely chalcocite (Cu2S),
3. Materials and methods covellite (CuS), chalcopyrite (CuFeS2), cuprite (Cu2O), and tenorite
(CuO). In all cases, the total mass fraction of copper-containing minerals
Previously reported models of the copper refining process are scarce. (sulphide and oxide ores) and non-copper minerals were 2% and 98%,
Therefore, Aspen Plus V11.0 was used to develop a comprehensive respectively. The five compositions considered here were labelled based
model of this process. Aspen Plus software was used because of its large on the sulphide content of the copper minerals. For example, a
property databanks, its solvers’ capabilities, including their “Goal Seek” composition named as 10% sulphide content indicates that, within the
algorithms, and the ease of modification of the properties and conver­ 2% of the total copper-containing ore, the cumulative concentrations of
gence methods. The model was developed to assess the performance of chalcocite, covellite, and chalcopyrite are 10%, while the rest 90% ac­
the proposed process configuration for the integration of the HyS and counts for the cumulative concentration of cuprite and tenorite
copper refining processes (Fig. 2). It solved the governing steady-state (Table 1). The composition of non-copper containing minerals (98 wt%)
equations of mass and energy simultaneously. The composition of the was constant and included silica (SiO2, 58.2 wt%), alumina (Al2O3, 10
streams were calculated for the crushing unit, concentration unit, and wt%), iron (II) oxide (FeO, 25 wt%), magnesium oxide (MgO, 0.84 wt
pyrometallurgy pathway blocks via the inbuilt “SOLIDS” property %), potassium oxide (K2O, 0.8 wt%), calcium carbonate (CaCO3, 3.2 wt
package. Additionaly, through the “Customize” option in the “Proper­ %) and triuranium octoxide (U3O8, 0.005 wt%).
ties” ribbon, the Barin’s handbook [30] was addressed as the reference A variation in the ore’s sulphide content directly impacts the amount
for the metallurgical thermodynamic calculations. The composition of of copper-containing ores reduced via the pyrometallurgical pathway.
the products leaving the smelter, slag cleaner, and fire refiner reactors in Table 1 reports the fraction of the copper produced via the pyrometal­
the pyrometallurgy pathway, were calculated using the Gibbs energy lurgical route as a function of the inlet ore sulphide content.
minimisation method. Due to the high degree of ionisation and high
concentration of ionic species, in the HyS cycle, sulphuric acid plant and
3.2. Process flowsheet of the hybrid sulphur cycle
hydrometallurgy pathway blocks the ELEC-NRTL property package with
ideal gas assumption was used. The “MIXED” and “CIPSD” packages
Fig. 3 presents the flowsheet of the HYS cycle, developed in Aspen
were chosen as the stream class in the simulations.
Plus. This model has been developed from the process flow diagram
The plant capacity for all simulations was assumed to be 5000 tonnes
reported by Niehoff et al. [16], verified against their simulations and
of ore/h, while the capacity factor of the concentrated solar thermal
further improved to enable: lower temperatures for the sulphuric acid

Fig. 3. Process flow diagram of the developed HyS cycle adapted from earlier work of Niehoff et al. (16).

4
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Table 2 were less than 2% for the same sulphuric acid mass flow (Fig. 3, Stream 1
Comparison of the prediction of the present model of HyS with Niehoff et al.’s and 81). The slight differences are attributable to the difference in the
model (16). databases employed to estimate the thermophysical and thermochem­
Parameter Data from the Data from Deviation ical properties of mixtures.
present model Niehoff et al.’s (%) The efficiency of the pumps and compressors assumed in the model
model was ~0.75 – 0.85, respectively [16], while that of the HyS cycle is ~25%
Sulphuric acid concentration 50 50–75 N/A (in based on the lower heating value of H2 [16,29]. The hydraulic pressure
(wt%) range) losses were ignored. The ELEC-NRTL property package was used for this
Decomposition temperature 850 900–1200 N/A (in
flowsheet. It is worth mentioning that the default ELECT-NRTL property
(◦ C) range)
Hydrogen mass flow for 50 1.38 1.4 1.5% package should be modified for a detailed study on phase separation in
wt% acid concentration the KSEP absorption tower.
(kg/h)
Oxygen mass flow for 50 wt% 8.06 8.1 0.5%
acid concentration (kg/h) 3.3. Copper refining process
Sulphuric acid mass flow (kg/ 98 98 0
h)
Sulphur dioxide depolarised 55% 55% 0 The copper refining process shown in Fig. 2 includes the crushing
electrolyser (SDE) unit, concentration unit, direct to blister smelting unit (Pyrometallurgy
efficiency pathway), acid plant, and hydrometallurgy unit.

3.3.1. Crushing unit process flowsheet

cracking reactor and improved separation in the absorption towers
The process flow diagram (PFD) of the comminution and crushing
Sulphuric acid makeup. This was considered to compensate for the SO2
plant is shown in Fig. 4. The input ore (Stream 1) is screened (SCRN-1) to
loss of the cycle. It should be noted that the HyS process was assumed to
remove particles smaller than 110 mm (Stream 5) from the larger par­
operate continuously. Therefore, the Aspen Plus simulation was carried
ticles (Stream 2, Fig. 4). A small amount of water (less than 1 wt%) is
out in the steady state condition. The full stream results of the model
sprayed into Streams 2 and 10. The mixture (Stream 3) is sent to the first
were compared with the model of Niehoff et al. [16] for the case with the
jaw crusher (JC-1). The product (Stream 4) is mixed with Stream 5 and
circulating sulphuric acid concentration of 50 wt%. However, only the
enters the second screen (SCRN-2), from where the particles of diameter
key indicators are reported in Table 2.
>70 mm are sent to the second jaw crusher, while the undersized par­
As can be seen, the differences between these two models for the
ticles (Stream 7) are mixed with Stream 11 (diameter ≤ 70 mm). Stream
hydrogen (Fig. 3, Stream 56) and oxygen mass flow (Fig. 3, Stream 28)
10 recycles the particles with diameters larger than 70 mm back to the

Fig. 4. Process flow diagram of the crushing unit.

Fig. 5. Process flow diagram of the developed flotation unit. Stream 13, as the final froth, leaves the process with ~45% copper content, while the final tail
containing other minerals, together with most of the oxide-containing copper ores, is directed to the leaching and extraction plant via Stream 8 (the final tail
in Fig. 5).

5
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 6. Process flow diagram of the pyrometallurgy pathway modelled here. The blister from both the smelter and slag cleaner (Streams 12 and 14) is mixed and,
after further purification in the fire refiner, refined to copper of purity >99.5 wt%. The SO2-rich product gas from the smelter (Stream 10) and the fire refiner (Stream
21) is cleaned and sent to the acid plant.

first jaw crusher (JC-1). Stream 13 (diameter ≤ 10 mm) contains the first the particles of Cu ~45 wt% to the pyrometallurgical pathway, which
impact crusher (IC-1) product and is screened via SCRN4. The under­ comprises mostly of smelters, a slag cleaner and fire refiner. First, the
sized product (Stream 14) is then mixed with the product of the second water content of the ore is decreased using a steam dryer at ~T = 200 ◦ C
impact crusher (IC-2), Stream 16 (diameter ≤ 5 mm). The product to final ore moisture (Stream 2, Fig. 6) of ~5 wt%. The dried ores, the
(Stream 17) is sent to the ball mill (BM) and is crushed down to particles flux with 98 wt% silica and the recycled particles (Stream 2, 6 and 7,
with a maximum diameter ≤ 100 μm [31,32]. respectively) are mixed and added to the smelter reaction shaft. An
oxygen-enriched stream (Stream 5) is also injected into the smelter
3.3.2. Concentration unit process flowsheet through tuyers to oxidise the sulphur content of the ore to SO2. The heat
It is assumed that all the mined ores after comminution and crushing duty of the smelter is supplied from the combustion of a fuel (a hydro­
(Stream 18, Fig. 4) are sent to the flotation unit, where the oxide- carbon or hydrogen, Stream 4, Fig. 5) with oxygen (Stream 3, Fig. 5) in
containing copper ores and the sulphide-containing copper ores are the reaction shaft. An ASU unit supplies the required oxygen. The
mostly separated. The sulphide-containing ores are then sent to the products of the combustion in the flue gas are mixed with the gases
pyrometallurgical pathway, while the oxide-containing ores and other released during the metallurgical processes, mainly SO2, and therefore
ores are sent to the hydrometallurgical pathway. The PFD of the flota­ contains a large amount of SO2. Some 8–15% of the inlet particles are
tion plant modelled here is shown in Fig. 5, which includes three tanks also carried with the smelter off-gas (Stream 8, Fig. 6), which are then
for the separation of ores. The fine ore (diameter less than 100 μm) cooled down to ~500 ◦ C in the waste heat boiler (WHB) and are sent to
enters the rougher via Stream 1. Here, a fraction of the oxide and sul­ an electrostatic precipitator (ESP), where the entrained particles are
phide ores are separated from each other and carried via Streams 4 (as separated and recycled back to the smelter (Stream 7, Fig. 6), while the
tails) and 5 (as froth). The froth from the rougher (Froth-1) enters the SO2-rich gas (Stream 10, Fig. 6) is sent to the gas cleaning unit. The ore
cleaner, where the final froth (Stream 13, Fig. 5) with an elemental Cu in the smelter reactor melts and is carried to either the slag or blister
content of ~45% is sent to the smelters [5,7,22,33]. The tail from the stream, based on its density. The slag and blister contain ~20–22 wt%
cleaner (Stream 12, Fig. 5) contains some sulphide-containing ores and and ~95–99 wt% copper, respectively. The slag is sent to the slag
is sent back to the rougher. The tail from rougher (Stream 4, Fig. 5) flows cleaner furnace via Stream 11, Fig. 6, where a layer of coke covers the
to the scavenger, where the final tail (Stream 8, Fig. 5) is sent to the molten mixture (to prevent the oxidation of molten metal with the
hydrometallurgical process. The froth leaving the scavenger (Stream 9, ambient air), and electricity is applied to purify the molten mixture
Fig. 5) is sent back to the rougher – the first separation step. Compressed further. The slag leaving the slag cleaner (Stream 16, Fig. 6) contains
air (P = 1.5 bar) is used to float sulphide-containing copper ores from the ~4–5 wt% Cu and is typically discarded or stored for recycling. The off-
mixture in the tanks. gas from the slag cleaner contains mostly carbon dioxide resulting from
reacting the coke with the oxygen released from the metal ores during
3.3.3. Pyrometallurgy unit process flowsheet (direct to blister method) the reduction process. The present study does not consider the substi­
Moist ore from the concentration unit (Stream 13, Fig. 5) is sent with tution of coke with hydrogen; however, one could consider such

6
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 7. Process flow diagram of the developed acid plant. The sulphuric acid product concentration is ~99 wt%, while the clean gas leaving the plant has less than
10 ppm SO2.

replacement as a subsequent step if the reactor were to operate under a (Stream 23, Fig. 6) to lower the SO2 concentration to ~8 mol% (Stream
neutral atmosphere, such as Argon. 8). Fuel oil (Stream 4) is burnt with an oxygen-rich stream (Stream 5) to
The blister leaving the smelter and the slag cleaner (Stream 12 and supply the required heat. The exothermic conversion of SO2 to sulphur
14) is sent to the fire refiner reactor, where fuel and air are used to in­ trioxide (SO3) occurs in the catalytic bed reactors (the “1st bed” to the
crease the copper purity to more than ~99.5 wt%. The temperature of “4th bed”). It should be noted that the gas mixture temperature after
the smelter and the slag cleaner are ~1300 ◦ C and ~1310 ◦ C, respec­ each catalytic conversion step is decreased to ~400 ◦ C to optimise the
tively [7,33–35]. Here, it is assumed that Stream 23 contains mostly SO2 oxidation reaction [38,39]. The mixture leaving the “3rd bed”
sulphur dioxide (SO2), nitrogen (N2), and oxygen (O2). Stream 23 in (Stream 13 with SO2 < 0.0008 mol%) is cooled and sent to the first
Fig. 6, with a temperature of < ~100 ◦ C, is sent to the acid plant unit, absorption tower (TW-1). Freshwater (Stream 15) and concentrated
where the SO2 is converted to sulphuric acid [7]. sulphuric acid (Stream 28) are fed to the TW-1 to absorb the SO3 product
To account for the smelter duty, while maintaining the smelter using a circulating concentrated sulphuric acid solution.
temperature at ~1310 ◦ C, the Aspen Plus RGIBBS reactor package was The top product of TW-1 is sent to the last catalytic bed (Stream 18,
used to model the combustion of LPG with oxygen-enriched air. The Fig. 7), while the bottom product (Stream 16) is fed to the second ab­
flame generates product gases at 1320 ◦ C and 1 bar after the heat has sorption tower, “TW-2′′ , where the remaining SO3 is absorbed from the
been transferred to the smelter reactor. It is assumed that ~50% of the cooled gas mixture leaving from the last catalytic bed (Stream 20). The
generated heat from combusting LPG with enriched air is used to gas leaving from TW-2 (Stream 21) is cooled to condense and recycle the
maintain the temperature of the streams at 1300 ◦ C throughout the sulphuric acid vapour via stream 24. The mole fractions of SO2 and SO3
pyrometallurgical copper production. Finally, the flow rate of the oxy­ in the final gas leaving the process (Stream 23) are ~5 and ~1 ppm,
gen (Stream 5 in Fig. 6) is matched to ensure that the concentration of respectively. Both TW-1 and TW-2 have five equilibrium stages for the
Cu2S in the smelter is always <0.2 wt% to minimise foaming production of H2SO4 from SO3 and H2O and a pressure of 1.2 bar at their
[5,7,36,37]. The effect of oxygen flow rate on the product composition is top.
shown in Fig. 10. It is worth mentioning that the RGIBBS package with A pump-around flow is assumed for each tower to recycle the liquid
phase and product specification was used to model the smelter, the slag from the bottom to the top stages, although this is not shown in Fig. 7 for
cleaner and the fire refiner products. clarity. The required flow of the pump-around stream depends on the
SO2 content of the pyrometallurgy pathway gases.
3.3.4. Acid plant flowsheet A fraction of the acid product with a ~99 wt% acid content (Stream
Fig. 7 shows the flowsheet of the acid plant that was modelled here, 26, Fig. 7) is sent to the hydrometallurgy process for the leaching pro­
which incorporates two absorption columns and a four-stage catalytic cess, depending on the ore composition. For the assessment here, it is
bed reactor [38]. A stream of air (Stream 1) is preheated and mixed with assumed that the extra sulphuric acid required for this step is shipped to
the SO2-rich stream coming from the pyrometallurgical pathway the site if the sulphuric acid produced in the acid plant is insufficient to

7
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 8. Process flow diagram of the hydrometallurgy pathway evaluated here. The ore is leached in a leaching tank and then fed to the solvent extraction cycle, which
comprises two extraction tanks (ET-1 and ET-2) and two stripping tanks (ST-1 and ST-2). Next, the copper ions are removed in Stripping Tanks ST-1, and ST-2 and the
solution is sent to the electrowinning reactor, which deposits pure copper using electricity.

meet this demand. containing kerosene as the diluent and a ligand (LH) is mixed with it to
extract the copper ions in the form of L2Cu [6,42]. This also lowers the
3.3.5. Hydrometallurgy flowsheet pH of the aqueous solution. The aqueous phase leaving the first
A schematic representation of the hydrometallurgical pathway pro­ extraction tank is transferred to ET-2, where the organic phase absorbs
cess is depicted in Fig. 8. The main sections of the plant are those of acid the remaining copper ions in the aqueous solution. The outlet stream
dissolution, solvent extraction and electrowinning. In the acid dissolu­ from the second extraction stage (Stream 10) contains ~300 ppm Cu2+.
tion (i.e. leaching) section, the Flotation tail, air and the sulphuric acid To purify the solution further, the stripped organic phase leaving ST-2
solution (Streams 1, 2 and 3, respectively) are mixed in the leaching (Stream 15) is sent to the ET-2, where more copper ions are separated.
tanks, where the sulphuric acid dissolves the copper oxide, cupric oxide The organic phase from ET-2 is not saturated with copper ions and hence
and other metallic compounds in the ore, following which the super­ is sent to the ET-1 to further extract Cu2+ ions from the aqueous phase.
natant is sent to T-1. The solvent extraction comprises two extraction The loaded organic phase is next stripped of copper ions in ST-1 and ST-2
tanks (ET1 and ET2) and two stripping tanks (ST-1 and ST-2) [6,40,41]. using a cycling acid solution, after which they are deposited in an
The acidic solution containing the copper ions (Stream 6, Fig. 8) is electrowinning reactor. Since some water is consumed in the electro­
pumped to the first extraction tank (ET-1), where an organic phase winning reactor, a water makeup stream is added to ST-2 (Stream 23,

Fig. 9. The points in the copper production cycle where the outputs from the HyS cycle are used in the copper production plant via scenarios 1 and 2.

8
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

burner. For Scenario 1, where the oxygen product from the HyS cycle
would not be sufficient to cover all the demands, makeup oxygen from
the ASU units are utilised (ASUBKUP1 and ASUBKUP2 streams).
H2-3, OXY-3 and ASUBKUP2 streams in Fig. 9 provided the required
duty for the second burner (BRNR2) located in the acid plant (Furnace in
Fig. 7) and replaced the fuel oil with hydrogen. The “SULPHUR” heater
in Fig. 9 represents the total exchanged heat from the flue gas to the SO2-
rich stream (combined duties of furnace and preheater in Fig. 7). The
flow rate of HYS-H2-3 stream is zero in Scenario 1. However, in Scenario
2, since the HyS cycle is sized to produce the required oxygen for the
process, extra hydrogen would be produced, and the flow rate of the
HYS-H2-3 stream will not be zero. It should be noted that since the HyS
cycle would operate only when enough solar or stored thermal energy is
available, its hydrogen and oxygen production is prone to intermittency
and therefore, capacity factors should be considered (CF: 0.55 to 0.75).
Fig. 10. The effect of oxygen flow rate through the tuyeres on the smelter
product composition at 1300 ◦ C and 1 bar, as calculated with the model.
4. Results and discussion

Fig. 8). Some pH balancing and neutralisation of Stream 10 may then be In Section 4.1, the results specific to the modelled copper plant are
required before being discharged into the environment, which is not discussed. Then in Section 4.2, the results for the integration of the HyS
shown in Fig. 8. cycle with copper refiner are reviewed. Sections 4.3 and 4.4 discuss the
While some further solvent extraction cycles may be employed in potential CO2 reduction and the economics of the integration via both
practice to dissolve small amounts of uranium, gold and/or silver scenarios.
[41,43], these are ignored here. This simplification follows earlier work
[43,44].
4.1. Copper production plant

3.4. Integration of the hybrid sulphur cycle with the copper production Table 3 presents the assumed particle size distribution (PSD) of the
plant input ore stream (Stream 1, Fig. 4), together with the calculated PSD in
each successive stream within the comminution and crushing processes
The HyS cycle (Fig. 3) was reproduced from Niehoff et al. [16], with (Fig. 4). Also shown are the solids flow rate and the calculated specific
an efficiency that is a function of both decomposer temperature and energy for each size reduction stage. A constant grindability of 12 kWh /
circulating sulphuric acid concentration. For a decomposer temperature tonne of solid was assumed for the Bond work index, following earlier
of 850 ◦ C and circulating sulphuric acid concentration of 50 wt%, the works [31,45,46].
HyS cycle efficiency varies between 18 and 25% for cases without and The results from Table 3 are consistent with the expectation that the
with heat recovery, respectively. specific energy for comminution increases with a reduction in the par­
Fig. 9 presents the approach evaluated with which the HyS cycle is ticle diameter. The crushing unit requires ~81.75 MJ/tonne of solids to
integrated with the copper production cycle. The H2 and O2 products reduce the particle size distribution to less than or equal to 100 μm
from the HyS cycle are sent to the copper production cycle via streams without considering the conveying and screening energy.
HYS-H2 and HYS-O2, respectively. A small fraction of the hydrogen Table 4 reports the calculated values of the flow rate and composi­
(Stream H2-1) is sent to the refiner reactor (REFIN) to reduce the oxides. tion of the final froth (feed to the pyrometallurgical pathways) and tail
In addition, ambient air (21 mol% oxygen) is supplied to the reactor via (feed to the hydrometallurgical pathways) in the flotation unit (Fig. 5)
OXY-4. on a dry basis for various values of ore composition. The flow rates and
The oxygen product of the HyS cycle (HYS-O2) contained ~99 mol% the separation ratios of each component in the rougher, scavenger and
oxygen and was diluted with the ambient air (AIR) to get ~70 mol% cleaner units were adjusted for each composition to assure that the
oxygen content in the HYS-O2-1 stream before being fed to the smelter stream leaving the flotation to smelter contains ~45 wt% elemental
via the OXY-1 stream in Fig. 9 (the equivalent of Stream 5 in Fig. 6). copper [7,35].
The smelter duty was also supplied by burning the oxygen-rich Returning to Fig. 6, which presents the pyrometallurgy pathway, it
stream (OXY-2) with the hydrogen stream (H2-2) in the BRNR1 can be seen that the moist ore (final froth from the flotation unit) is dried

Table 3
The particle size distributions, both of the input ore (assumed) and as calculated after each comminution step, together with the values of specific power determined
with the model.
Stream No. (Fig. 4) Particle Size Distribution 1 4 9 13 16 18

0–20 μm 0 0.012 0.020 0.034 0.083 0.045

20–100 μm 0 0.002 0.004 0.059 0.100 0.955
100–300 μm 0 0.002 0.003 0.126 0.172 0
300–1000 μm 0 0.002 0.004 0.299 0.624 0
1000–5000 μm 0 0.005 0.009 0.443 0.022 0
5000–10,000 μm 0.05 0.008 0.019 0.039 0 0
10,000–20,000 μm 0.10 0.064 0.203 0 0 0
20,000–70,000 μm 0.55 0.328 0.735 0 0 0
70,000–110,000 μm 0.10 0.576 0.004 0 0 0
110,000–150,000 μm 0.20 0.001 0 0 0 0
Equipment in simulation NA JC-1 JC-2 IC-1 IC-2 BM
Type of Crushing / Grinding NA Jaw Crusher Jaw Crusher Impact crusher Impact crusher Ball mill
Solid flow rate (tonne /h) 5000 1622 1047 5000 2126 5000
Specific Energy (MJ / tonne solid) NA 0.25 0.68 5.99 13.01 70

9
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Table 4
Calculated values derived with the model of the flow rates and composition of the final froth and final tail, leaving the concentration unit in the flotation circuit for
various values of sulphide content in the ore.
Ore sulphide content (wt %) 10% 25% 50% 75% 90%
Cu (wt %) produced via pyrometallurgy 17% 31% 54% 77% 91%

Outlet type Froth Tail Froth Tail Froth Tail Froth Tail Froth Tail

Mass flow rate (tonne / h) 30 4970 54 4946 93 4907 132 4868 156 4844
Elemental Cu (wt %) 45.03 1.39 45.2 1.10 45.10 0.75 45.20 0.38 45.03 0.15
Component (wt %)
Silica (SiO2) 10.36 58.44 14.33 58.64 17.48 58.93 21.05 59.16 18.03 59.45
Alumina (Al2O3) 1.78 10.05 1.97 10.09 2.86 10.14 3.82 10.17 3.07 10.22
Iron (II) oxide (FeO) 27.13 24.99 23.96 25.01 20.83 25.08 16.37 25.23 20.53 25.14
Magnesium oxide (MgO) 0.00 0.85 0.00 0.85 0.00 0.86 0.00 0.86 0.00 0.87
Potassium oxide (K2O) 0.00 0.80 0.00 0.81 0.00 0.82 0.00 0.82 0.00 0.83
Calcium carbonate (CaCO3) 0.00 3.22 0.00 3.24 0.00 3.26 0.00 3.29 0.00 3.30
Triuranium octoxide (U3O8) 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.01
Chalcopyrite (CuFeS2) 4.97 0.00 3.67 0.00 2.67 0.00 2.64 0.00 2.56 0.00
Covellite (CuS) 16.30 0.00 9.03 0.00 5.26 0.00 3.71 0.00 3.14 0.00
Chalcocite (Cu2S) 18.59 0.00 37.49 0.00 47.17 0.00 51.13 0.00 52.25 0.00
Cuprite (Cu2O) 11.09 0.87 4.13 0.59 0.56 0.14 0.35 0.12 0.18 0.07
Tenorite (CuO) 9.77 0.77 5.42 0.77 3.16 0.78 0.93 0.33 0.25 0.10

Table 5
Simulated values of the contribution to copper recovery from each of the various
stages in the process for a series of values of ore compositions.
Component wt% Sulphide 10% 25% 50% 75% 90%
content

Cu wt% produced via 17% 31% 54% 77% 91%

pyrometallurgy
Cu wt% produced via 83% 69% 46% 23% 9%
hydrometallurgy
Cu wt% content in the slag 21.9% 20.8% 20.6% 20.8% 20.5%
Cu wt% content in the discarded 5.6% 5.4% 5.4% 5.5% 5.4%
solids
Yield (Cu in / Cu out) 89.7% 90.3% 91.2% 92.2% 92.7%

Fig. 12. Effect of copper ore sulphide content on the amount of sulphuric acid
required for leaching per tonne of copper product, together with the corre­
sponding effect on the amount of sulphuric acid produced.

Gibbs reactor at ~1310 ◦ C and 1 bar for the case of sulphide content
equal to 50 wt%. Although the maximum Cu yield can be obtained with
an oxygen flow rate of ~15 tonnes/h, the concentration of Cu2S should
be kept below 0.2 (wt%) to avoid foaming in the smelter [7,47]. Hence
the oxygen flow rate chosen for operation is ~17.42 tonne/h.
Fig. 11 presents the estimated thermal duty of the pyrometallurgy
flowsheet per tonne of copper as a function of sulphide content in the
ore. It can be seen that an increase in the sulphide-containing fraction of
copper ore results in an increase in required thermal duty. These results
are influenced by the requirement to use (chemically neutral) flux,
which prevents auto-thermal conditions from being reached with an
Fig. 11. Effect of copper ore sulphide content on the estimated total thermal
duty of the pyrometallurgical pathway, as calculated with the model. increase in the sulphur content. This maintains the Cu concentration
constant at ~40 wt% in the feed. The effect of an increase in the
exothermic reduction of copper with an increase in the sulphide content
with superheated steam to recover energy [7]. The required mass flow
of the ore was observed in the available heat from the waste heat boiler
rate of steam was estimated to be ~1.1 times more than that of the ore.
(WHB), as the primary heat recovery unit. For ore compositions of 10 to
The semi-dried ore has a temperature of ~130 ◦ C, moisture content of
90 wt% sulphide content, the duty ratio of waste heat boiler (WHB) to
less than 3 wt% and a mass flow rate of ~15% of the smelter input ore
the smelter spanned ~85 to 104%, respectively.
mass flow rate. The flux contains quick lime (CaO, 2 wt%) and silica
The total heat loss in the pyrometallurgical flowsheet was assumed to
(SiO2, 98 wt%). The slag (20–22 wt% Cu) is separated from the matte by
be ~50% of the combusted fuel. In addition, it is assumed that ~10% of
density difference and introduced to the slag cleaner for further purifi­
the heat is carried out from the smelter by the particles in the gaseous
cation. The slag stream leaves the slag cleaner with ~5–7 wt% Cu. The
stream. These particles are sent back to the smelter after passing through
calculated values of Cu wt% in the slag and the discarded solids are
the WHB and electrostatic precipitator (ESP) (Fig. 6). Therefore, only a
shown in Table 5 for the various values of ore compositions and the
fraction of the particles sensible energy can be recovered.
copper yield (i.e. the ratio of Cu at the inlet to the outlet).
Fig. 12 presents the amount of acid generated per tonne of copper
Fig. 10 presents the dependence of the smelter product composition
product (blue circles) as a function of sulphur content (Fig. 7). These
on the oxygen flow rate through the tuyeres (Fig. 6), calculated with the
were calculated assuming that the pH of the effluent stream leaving the

10
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 15. Effect of copper ore sulphur-content on the required electricity in the
Fig. 13. Effect of copper ore sulphide content on the flow rate of circulated HyS cycle for scenarios 1 and 2.
organic phase per tonne of copper product.

Fig. 16. Effect of copper ore sulphur-content and fraction of copper produced
Fig. 14. Effect of copper ore sulphide content on the required solar input and
in the pyro furnaces on the amount of H2 and O2 required to be produced with
the circulating sulphuric acid (in the HyS cycle) for scenarios 1 and 2, as
the HyS cycle for scenarios 1 and 2.
calculated with the model.

scenarios, together with the required flow rate of circulated H2SO4 so­
leaching tanks (Stream 5 in Fig. 8) was ~0.8. It can be seen that this
lution (50 wt%) per tonne of Cu product. It can be seen that an increase
production rate increases with an increase in the ore sulphur content. As
in the sulphur content of the ore leads to an increase in the input solar
mentioned in the section describing the hydrometallurgy flowsheet, the
duty and circulated H2SO4. In addition, since the oxygen demand from
flotation tail containing the copper oxide ores, together with other
the copper production process is higher than the hydrogen demand, the
minerals such as silica, alumina and iron oxide (II), is sent to leaching
size of the HyS cycle for Scenario 2 is greater than for Scenario 1.
tanks where the copper oxide ores are dissolved with the aid of the
Fig. 15 presents the electrical energy required for the SO2 depo­
sulphuric acid solution and air. The amount of acid required can also be
larised electrolyser (SDE) used to recombine SO2, O2 and H2O to the
seen to be a function of the ore composition (red squares). More spe­
sulphuric acid (Fig. 2) as a function of the sulphide content of the ore for
cifically, an increase in the sulphide content of the ore leads to a
both scenarios. It can be seen that the required electrical duty for SDE
decrease in the amount of acid required for the leaching. Hence, for ores
increases with sulphide content up to a maximum value of approxi­
with 10, 25 and 50 wt% sulphide content, an acid makeup stream is
mately 1.2 MWhe/tCu, which corresponds to 115 kJ/mole SO2. The
necessary to fully dissolve the copper oxide ores. However, for cases of
auxiliary power needed to circulate the fluids via pumps and compres­
90 wt% sulphide content, the acid produced in the acid plant is more
sors is also shown. This is approximately one-quarter of the demand of
than sufficient to provide the acid for the leaching tanks. The critical
the SDE.
value of sulphide content for which the acid requirements equal that
The hydrogen and oxygen production rate from the HyS cycle is
produced is approximately 75%.
presented in Fig. 16 as a function of sulphide content in the ore for the
The effect of sulphide content on the circulating organic phase mass
two scenarios. Also shown on the secondary abscissa is the fraction of
flow rate is depicted in Fig. 13 for the organic solution containing 50 wt
the copper ore that is processed via pyrometallurgy (the equivalent of
% ligands and 50 wt% organic diluents (e.g. kerosene), which was used
the 2nd row in Table 4). Point A in the Fig. 16 represents the example of
to extract the dissolved Cu2+ from the effluent stream. The final stream
the case for which the sulphur content of the ore is 60%, for which the
contained ~300 ppm (weight basis) Cu2+ ions. An increase in the ore
corresponding fraction of the ore refined in the pyrometallurgical
sulphide content can be seen to decrease the dependence of the copper
pathway is ~64% if the plant is integrated with the HyS cycle via Sce­
refining on the hydrometallurgical pathway, reducing its load and
nario 2. For this case, the HyS cycle also provides ~0.39 tonnes of ox­
decreasing the requirement for a circulated organic phase to strip the
ygen and ~0.062 tonnes of hydrogen per tonne of Cu. Since the pyro
copper ions.
pathway only requires ~0.039 tonnes of hydrogen, the residual (0.023
tonnes) is the excess hydrogen. The molar purities of the oxygen and
4.2. The hybrid sulphur cycle and its integration with copper refinery hydrogen streams shown in Fig. 16 are also calculated to be 0.99 and
0.96, respectively.
Fig. 14 presents the calculated solar input required to evaporate the Fig. 17 presents the requirements calculated with the model for
sulphuric acid and decompose it to SO2, O2 and water for each of the two

11
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

with the sulphide content of the ore, as does the fraction of copper
produced via the pyrometallurgical pathway. Note that, even for the
case for which the inlet ore comprises only sulphides, some 10–20% of
the copper is still produced via the hydrometallurgical pathway, and
only the rest is produced via the pyrometallurgical pathway. This is
because the uranium content of this ore is removed by leaching with
sulphuric acid, which also dissolves, ~10–20% of the copper that is
recovered via the hydrometallurgical pathway.
Fig. 18 presents the corresponding amounts of hydrogen and oxygen
required by the copper production plant to avoid consuming both hy­
drocarbon fuels and some, or all, of the oxygen demand, which is
conventionally supplied with air separation units (ASU) via scenarios 1
and 2, respectively. Here the acronyms FR, AB, SB and PS correspond to
the fire refiner (Fire refiner in Fig. 6), acid plant burner (Furnace in
Fig. 7), smelter burner (HC/H2 Burner in Fig. 6) and the process smelter
(Stream 5 in Fig. 6), respectively. It is evident that the main consumer of
Fig. 17. Effect of copper ore sulphur-content and the fraction of copper pro­
H2 is the smelter burner. However, for O2, the primary consumers are the
duction in the pyro furnaces on the consumption of oxygen, LPG and fuel oil. smelter burner and the oxygen that is used in the reduction reactions of
The square symbols are the published consumption of these inputs by the BHP copper.
copper production plant at Olympic Dam (OD) Australia (44). The total O2 demand from the copper plant for Scenario 2 is the sum
of the PS, SB, and AB requirements. A comparison between Figs. 17 and
18 shows that the copper production plant would consume slightly less
oxygen if H2 was used as fuel rather than hydrocarbons. This is due to
the different stoichiometric requirements and heating values of the two
fuels. The increase in demand for H2 and O2 with an increase in the
sulphide content of the ore is consistent with previous figures.
Fig. 19a presents the extent to which the load of the ASU is reduced
after integration of the HyS cycle with the copper production plant via
Scenario 1 as a function of the capacity factor of the HyS cycle. Note that
the capacity factor is always less than 100% due to the need to accom­
modate both seasonal and weather-based variability of the solar
resource (Kueh et al., 2015). It can be seen that this fraction varies from
~52% to ~71% for capacity factors of 65% to 85%, respectively. Due to
operational and economic limitations it is not possible to fully avoid the
need for an ASU since it brings the requirement for additional CAPEX in
a new plant and OPEX in a new plant or retrofit. Alternatively, the
Fig. 18. Effect of sulphur-content of the copper ore on the H2 and O2 demands equivalent energy for extra H2 produced ( ~96 mole% purity and
from the pyro furnaces for the case in which these gases are supplied from the LHV=~87 MJ/kg) in Scenario 2 per tonne of copper product is depicted
HyS cycle. Here the terms FR, AB, SB and PS are fire refiner, acid plant burner, in Fig. 19b. The error bars in Fig. 19a and b represent the effect of
smelter burner and process smelter, respectively. variations in the ore sulphide content on the reported parameters.

oxygen, LPG and fuel oil per tonne of Cu for the reference case for which 4.3. The potential in CO2 emissions reduction
fossil fuel is used to provide the energy as a function of sulphide content
of the ore. Also shown on the secondary axis is the fraction of copper that Table 5 presents the assumptions used to calculate the potential for
is produced via the pyrometallurgical pathway and the corresponding the HyS cycle to mitigate CO2 emissions from copper production based
published consumptions of the operating copper plant of BHP’s plant at on the situation relevant to Olympic Dam in South Australia. For this
the Olympic Dam (OD) site in Australia [33,36,44]. Consistent with plant, electricity is presently drawn from the South Australian (SA) grid,
Figs. 11 and 16, the consumption of both fuels and of oxygen increases which already has an estimated 60% penetration of renewable energy

Fig. 19. The dependence on the capacity factor of the HyS plant of A) the demand reduction of the air separation unit after integrating the copper production plant
with the HyS cycle via Scenario 1 as a function of the capacity factor. B) The fractional excess hydrogen produced after integrating the copper production plant with
the HyS cycle via Scenario 2.

12
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 20. The estimated intensity of CO2 emissions from the pyro furnaces and sulphuric acid plants in the copper production plant for the cases before (capacity
factor, CF = 0) and after (CF = 65, 75 and 85) the integration with the HyS cycle. Also shown is the emission from in-situ electrolysers supplied from the South
Australian grid, as estimated using Table 5. The range of values corresponds to the upper and lower bounds of emissions intensity of the SA grid in 2020, while the
copper concentration is set for the conditions that apply in Olympic Dam (where ~80% of Cu is refined through the pyro pathway), Australia.

Table 6 Table 7
Assumptions which were used to evaluate the CO2 emission intensity for data Assumptions were used to estimate the economics of the integration of the
shown in Fig. 20. copper production process with the HyS cycle via scenarios 1 and 2.
Specification Unit Min. Max. Ave. Remarks Specification Value Unit Remarks

e-CO2 emission of tCO2 / 0.09 0.58 0.34 AEMO (SA, 2020) The estimated cost of 4500 Au Both open pit and underground
grid electricity MWhel [48] copper production $/tCu mining [7]
ASU electricity kWhel / – – 200 The value is taken Percentage of the total 12% ± – Both open pit and underground
consumption (70% tO2 from [49,50] cost related to Pyro. and 3% mining [7]
mole oxygen acid plant
purity) Cost component of HC 20 – – This is a fraction of the total
PSA electricity kWhel / – – 500 The value is taken fuel, ASU and electricity 50% operating cost of Pyro. and acid
consumption (70% tO2 from [50] in Pyro. and acid plant plant based on tables 6-8,
mole oxygen Chapter 22 [7]
purity) The estimated cost of 0.68 ± Au$/L The value is used to estimate
Required O2 in Cu tO2 / 0.93 1.15 1.03 Simulation result wholesale LPG 20% Cu production cost at Olympic
production (ASU or tCuPyro (per tonne of refined Dam [52]
HyS) Cu product via The estimated cost of 1.1 ± Au$/L The value is used to estimate
Pyro.) wholesale fuel oil 20% Cu production cost at Olympic
Produced CO2 tCO2 / 7.96 9.96 8.73 Simulation result Dam [52]
because of HC tCuPyro (per tonne of refined The estimated cost of 65 ± Au The value is used to estimate
burning Cu product via wholesale electricity 5% $/MWh Cu production cost at Olympic
Pyro.) Dam [48]
Electricity MWhel / 1.30 1.36 1.33 Simulation result The estimated cost of O2 70 ± Au The value is used to estimate
consumption (HyS tO2 from ASU 5% $/tO2 Cu production cost at Olympic
cycle) Dam [49]
Electricity MWhel / – – 5 The value is taken The estimated cost of H2 5.1 ± Au$/kg [29]
consumption of tO2 from [51] production via the HyS 24%
electrolysers cycle
The HyS cycle production 100 tH2/day [29]
capacity
[37]. Hence, this grid’s CO2 emissions intensity in 2020 was ~0.09–0.58 The capacity factor of the 0.75 – [29]
HyS cycle
tonnes of CO2 per MWhel [48].
Energy storage capacity 13 h [14,29]
Fig. 20 presents the calculated range of CO2 emissions intensity for Cost estimation exponents 0.6 ± – The exponents [53] are used to
the current process compared to a series of technology options used to 0.1 adjust the evaluated cost of
meet the demand for H2 and O2 for the pyro furnaces and the acid units hydrogen for 100 tonnes per
in the copper processing plant for different capacity factors using the day to the estimated hydrogen
demands via scenarios 1 and 2.
assumptions provided in Table 5. Here the ASU + HC and PSA + HC
Note that the required
series refer to the use of hydrocarbon fuels (HC), together with the ox­ hydrogen is taken from the
ygen being produced from ASU from the SA electrical grid or oxygen simulation results (Fig. 16)
produced via pressure swing adsorption method (PSA). The water Revenue from extra 4 AUD Competitive with the CSIRO
electrolysis bar represents the case in which the electricity from the grid hydrogen produced via $/kg 2025 cost for green H2 via PEM
Scenario 2 and AE [54]
(with the assumptions in Table 5) is used to provide the energy for the
oxygen and hydrogen on-site.

13
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

Fig. 21. Estimated cost of hydrogen production in the HyS cycle and its unitlisation in furnaces of a copper plant through Scenario 1, Scenario 2 without H2 revenue
and Scenario 2 with H2 revenue (4 Au$/kg H2) for different fraction of copper product via pyrometallurgy.

The results in Fig. 20 show that the integration of the copper pro­ mine. It can be seen that, for sulphide fractions of the copper-
duction plant with the HyS cycle offers the highest potential to decrease containing ore of 10%, 25%, 50%, and 90% (which corresponds to
the CO2 emissions by more than 50% (blue bars). The CO2 emissions of 17%, 31%, 55%, 77%, and 91% of Cu product via pyrometallurgy), the
the grid significantly impacts the potential decrease in the emission after maximum cost of hydrogen per tonne of pyrometallurgical copper
integrating the copper production plant with the HyS cycle. For product in Au$ should be less than 157, 191, 246, 300 and 332 for
example, with the minimum value of grid CO2 emissions (0.09 tCO2 / underground mines and 132, 160, 205, 250 and 277 for open-pit mines
MWhel), the reduction in CO2 emissions from integrating the copper without any premium for low-carbon products, respectively. The
plant with the HyS cycle is comparable with that from electrolysis. It maximum value of each series in Fig. 21 corresponds to n = 0.5 (scale
should also be noted that both the electrolysis units and the HyS cycle exponent) and 6.3 Au$/ kg H2 for HyS cycle capacity of 100 tonnes per
can further reduce emissions intensity. For the HyS cycle, this would day (row 8 in Table 6), while the corresponding minimum is for n = 0.7
require thermal storage, or chemical storage (sulphuric acid, oxygen and and 3.9 Au$/ kg H2. The second vertical axis in Fig. 21 reports the
hydrogen storage) or a combination of both, whilst for the electrolysis, equivalent values in Au$/ kg H2.
this would require increased penetration of renewable energy into the It is evident from Fig. 21 that an increase in the fraction of the copper
grid. However, the investigation of these options is beyond the scope of produced via pyrometallurgy leads to a reduction in the copper pro­
the present research. duction cost after integration with the HyS cycle. Furthermore, it is
estimated to be economically favourable to employ the HyS cycle via
Scenario 2, where more than 70% of the copper is processed via the
4.4. Preliminary assessment on the economics of integration
pyrometallurgical pathway for underground and open-pit mining
without any carbon price or low-carbon premium.
The economic assessment was based on a previous economic
It should be noted that some data in Fig. 21 are related to plants with
assessment of the HyS cycle [29], which estimated a hydrogen produc­
a capacity of 5000 tonne ore/h, and therefore since the capacity of OD is
tion cost of ~5.1 ± 1.2 Au$/kg H2 (for a capacity of 100 tonnes per day).
about 2.6 times less than the original flowsheet, all the employed sce­
Table 6 lists this, together with the other assumptions that were used in
narios predicted a high cost for the hydrogen replacements. Neverthe­
the economic assessment.
less, Scenario 2, with revenue from the excess H2, could provide a
Fig. 21 presents the estimated equivalent cost of hydrogen per tonne
distinctly lower price than the cost estimated for the conventional
of the pyrometallurgical copper product after integration of the HyS
copper production using HC and ASU at Olympic Dam (the black circle
cycle with a copper plant (i.e. from replacing both the HC fuel and the
in Fig. 21).
ASU demands of the copper plant with O2 and H2 from the solar thermal
and HyS cycle). Data are presented for three cases namely; Scenario 1,
5. Conclusions
Scenario 2 without H2 revenue and Scenario 2 with H2 revenue (4 Au
$/kg H2). This price is competitive with (i.e. slightly below) the esti­
With this model it is estimated that the integration of the renewable
mated production costs of green hydrogen via both PEM and alkaline
HyS cycle into copper production is economically favourable even
electrolysis in Australia by 2025 [54]. This price is also estimated to be
without any carbon price or low-carbon premium for those cases where
economical for heavy trucks, which is a likely application in the mining
more than 70% of the copper is processed via the pyrometallurgical
sector. Also shown is the cost estimation using the model of the current
pathway (i.e., for ores of more than ~50% sulphide) for both under­
operating cost of pyrometallurgical production of copper at Olympic
ground and open-pit mining. This estimate is based on the assumption
Dam (OD) (the black circle). Note that also the three aforementioned
that the excess hydrogen is valued at AUD$4/kg, e.g., for transportation
scenarios tailored for OD estimated capacity (1900 tonne/h) are shown
in the mine, which is typical of the estimated cost for green hydrogen in
in Fig. 21.
2025 [54]. The use of a premium and/or the further development of the
Fig. 21 also presents two sloped dashed lines, which correspond to
HyS cycle to continue to lower its production cost would further increase
the estimated maximum viable cost of hydrogen to justify replacing the
the viability and/or allow its use for ores with a lower sulphide content.
current fuel and air on economic grounds alone in a conventional copper
The viability is also found to be greatest for Scenario 2, in which the
production plant supplied either for an underground or an open-pit

14
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

HyS cycle is sized to meet the oxygen demand of the copper plant rather Mining: Assessing Clean Energy Scenarios. 2012;(1):1–32. Available from: www.
isf.uts.edu.au.
than the hydrogen demand needed to replace the hydrocarbon fuel. The
[5] Sohn HY, Malfliet A, Scheunis L, Jones PT, Blanpain B. Copper Production. Vol. 3,
level of CO2 mitigation from this approach is also greater than would Treatise on Process Metallurgy. 2014. 534–624 p.
occur using electrolysis if fed from a non-renewable electrical grid, such [6] Bacon G, Mihaylov I. Solvent extraction as an enabling technology in the nickel
as that with current emissions intensity. While zero net emissions are industry Coming of age in copper solvent extraction. J South African Inst Min
Metall. 2002;(December):435–44.
potentially possible with either HyS or electrolysis, both require further [7] E. SM, J. KM, G. DW. Extractive metallurgy of Copper [Internet]. Elsevier. 2011.
developments in technology over what is currently available. Hence Available from: http://www.elsevier.com/locate/scp.
further research is justified both to continue the upscaling and further [8] Tow T, Yusup Y, Wei L. Heavy Metal Ion Extraction Using Organic Solvents: An
Application of the Equilibrium Slope Method. Stoichiom Res - Importance Quant
development of the HyS cycle, and to explore further options for sup­ Biomed. 2012;(March).
plying a copper production site with oygen and hydrogen from various [9] Gallardo FI, Monforti Ferrario A, Lamagna M, Bocci E, Astiaso Garcia D, Baeza-
renewable water-splitting routes, including electrolysis. Jeria TE. A Techno-Economic Analysis of solar hydrogen production by electrolysis
in the north of Chile and the case of exportation from Atacama Desert to Japan. Int
J Hydrogen Energy [Internet]. 2020;(xxxx). Available from: https://doi.org/
CRediT authorship contribution statement 10.1016/j.ijhydene.2020.07.050.
[10] Röben FTC, Schöne N, Bau U, Reuter MA, Dahmen M, Bardow A. Decarbonizing
copper production by power-to-hydrogen: A techno-economic analysis. J Clean
Ahmad Seyfaee: Conceptualization, Investigation, Methodology, Prod. 2021;306.
Software, Data curation, Formal analysis, Validation, Visualization, [11] Brown NR, Revankar ST. A review of catalytic sulfur (VI) oxide decomposition
Writing – original draft, Writing - review & editing, Resources, Project experiments. Int J Hydrogen Energy [Internet]. 2012;37(3):2685–98. Available
from: doi: 10.1016/j.ijhydene.2011.10.054.
administration. Mehdi Jafarian: Conceptualization, Investigation,
[12] Hinkley J, Agrafiotis C. Solar thermal energy and its conversion to solar fuels via
Methodology, Software, Data curation, Formal analysis, Validation, thermochemical processes [Internet]. Polygeneration with Polystorage: For
Visualization, Writing – original draft, Writing - review & editing, Re­ Chemical and Energy Hubs. Elsevier Inc.; 2018. 247–286 p. Available from: doi:
10.1016/B978-0-12-813306-4.00009-4.
sources, Project administration, Supervision. Gkiokchan Moumin:
[13] Thomey D, De Oliveira L, Säck JP, Roeb M, Sattler C. Development and test of a
Conceptualization, Investigation, Data curation, Validation, Writing - solar reactor for decomposition of sulphuric acid in thermochemical hydrogen
review & editing. Dennis Thomey: Data curation, Visualization, production. Int J Hydrogen Energy 2012;37(21):16615–22.
Writing - review & editing. Claudio Corgnale: Writing - review & [14] Gorensek MB, Corgnale C, Summers WA. Development of the hybrid sulfur cycle
for use with concentrated solar heat. I. Conceptual design. Int J Hydrogen Energy
editing, Investigation, Visualization, Writing - review & editing. Chris­ 2017;42(33):20939–54.
tian Sattler: Investigation, Writing - review & editing, Supervision. [15] Jung YH, Jeong YH. Development of the once-through hybrid sulfur process for
Graham J. Nathan: Conceptualization, Supervision, Formal analysis, nuclear hydrogen production. Int J Hydrogen Energy [Internet]. 2010;35(22):
12255–67. Available from: doi: 10.1016/j.ijhydene.2010.07.168.
Methodology, Writing – original draft, Writing - review & editing, Su­ [16] Niehoff AG, Botero NB, Acharya A, Thomey D, Roeb M, Sattler C, et al. Process
pervision, Funding acquisition. modelling and heat management of the solar hybrid sulfur cycle. Int J Hydrogen
Energy [Internet]. 2015;40(13):4461–73. Available from: doi: 10.1016/j.
ijhydene.2015.01.168.
Declaration of Competing Interest [17] Bayer Botero N, Thomey D, Guerra Niehoff A, Roeb M, Sattler C, Pitz-Paal R.
Modelling and scaling analysis of a solar reactor for sulphuric acid cracking in a
The authors declare that they have no known competing financial hybrid sulphur cycle process for thermochemical hydrogen production. Int J
Hydrogen Energy [Internet]. 2016;41(19):8008–19. Available from: doi: 10.1016/
interests or personal relationships that could have appeared to influence j.ijhydene.2015.11.088.
the work reported in this paper. [18] Nathan GJ, Jafarian M, Dally BB, Saw WL, Ashman PJ, Hu E, et al. Solar thermal
hybrids for combustion power plant: a growing opportunity. Prog Energy Combust
Sci 2018;64:4–28.
Acknowledgement [19] Bayon A, Bader R, Jafarian M, Fedunik-Hofman L, Sun Y, Hinkley J, et al. Techno-
economic assessment of solid–gas thermochemical energy storage systems for solar
This research was performed as part of the Australian Solar Thermal thermal power applications. Energy [Internet]. 2018;149:473–84. Available from:
doi: 10.1016/j.energy.2017.11.084.
Research Initiative (ASTRI), a project supported by the Australian
[20] Corgnale C, Gorensek MB, Summers WA. Review of sulfuric acid decomposition
Government, through the Australian Renewable Energy Agency processes for sulfur-based thermochemical hydrogen production cycles. Processes.
(ARENA), ARENA 1-SRI002. The views expressed herein are not 2020;8(11):1–22.
[21] Kojo I V., Storch H. Copper production with outokumpu flash smelting: An update.
necessarily the views of the Australian Government, and the Australian
2006 TMS Fall Extr Process Div Sohn Int Symp. 2006;8(June):225–38.
Government does not accept responsibility for any information or advice [22] Wilkomirsky I, Parra R, Parada F, Balladares E. Continuous converting of copper
contained herein. The author would also like to acknowledge the matte to blister copper in a high-intensity molten-layer reactor. JOM 2014;66(9):
financial support of ARENA IEP MI5 program, enabling the paper to be 1687–93.
[23] Lassouane F, Menia S, Khellaf A. An overview of the hybrid sulfur cycle process for
open access. The valuable feedback on the original draft by the anony­ solar hydrogen production. 3rd Int Symp Environ Friendly Energies Appl EFEA
mous reviewers is also gratefully acknowledged. 2014. 2014.
[24] Jafarian M, Arjomandi M, Nathan GJ. Thermodynamic potential of molten copper
oxide for high temperature solar energy storage and oxygen production. Appl
Appendix A. Supplementary data Energy [Internet]. 2017;201:69–83. Available from: doi: 10.1016/j.
apenergy.2017.05.049.
Supplementary data to this article can be found online at https://doi. [25] Jafarian M, Arjomandi M, Nathan GJ. The energetic performance of a novel hybrid
solar thermal & chemical looping combustion plant. Appl Energy [Internet]. 2014;
org/10.1016/j.enconman.2021.114832. 132:74–85. Available from: doi: 10.1016/j.apenergy.2014.06.052.
[26] Jafarian M, Arjomandi M, Nathan GJ. Influence of the type of oxygen carriers on
References the performance of a hybrid solar chemical looping combustion system. Energy
Fuels 2014;28(5):2914–24.
[27] Jafarian M, Arjomandi M, Nathan GJ. A hybrid solar chemical looping combustion
[1] Rötzer N, Schmidt M. Historical, current, and future energy demand from global
system with a high solar share. Appl Energy [Internet]. 2014;126:69–77. Available
copper production and its impact on climate change. Resources 2020;9(4):44.
from: doi: 10.1016/j.apenergy.2014.03.071.
https://doi.org/10.3390/resources9040044.
[28] Corgnale C, Shimpalee S, Gorensek MB, Satjaritanun P, Weidner JW, Summers WA.
[2] Moreno-Leiva S, Haas J, Junne T, Valencia F, Godin H, Kracht W, et al. Renewable
Numerical modeling of a bayonet heat exchanger-based reactor for sulfuric acid
energy in copper production: A review on systems design and methodological
decomposition in thermochemical hydrogen production processes. Int J Hydrogen
approaches. J Clean Prod [Internet]. 2020 Feb;246:118978. Available from: doi:
Energy [Internet]. 2017;42(32):20463–72. Available from: doi: 10.1016/j.
10.1016/j.jclepro.2019.118978.
ijhydene.2017.06.216.
[3] Ekman Nilsson A, Macias Aragonés M, Arroyo Torralvo F, Dunon V, Angel H,
[29] Corgnale C, Summers WA. Solar hydrogen production by the Hybrid Sulfur process.
Komnitsas K, et al. A review of the carbon footprint of Cu and Zn production from
Int J Hydrogen Energy [Internet]. 2011;36(18):11604–19. Available from: doi:
primary and secondary sources. Minerals 2017;7(9):168. https://doi.org/10.3390/
10.1016/j.ijhydene.2011.05.173.
min7090168.
[4] Mudd GM, Memary R, Northey SA, Giurco D, Mohr S, Mason L. CLUSTER
RESEARCH REPORT No. 1.12 Future Greenhouse Gas Emissions from Copper

15
A. Seyfaee et al. Energy Conversion and Management 249 (2021) 114832

[30] Barin I, Platzki G. Thermochemical data of pure substances [Internet]. Third Edit. [42] Girgin I, Resources N. Comparative H 2 SO 4 and H 2 SO 4 + H 2 O 2 leaching of a
Vol. 55. New York; 1997. Available from: https://linkinghub.elsevier.com/ Turkish lateritic nickel ore. Impc 2014;2014:1–10.
retrieve/pii/S0165242796056322. [43] Gri J, Ehrig K. Using existing data to improve the delineation of ore at the Olympic
[31] Ballantyne GR, Powell MS, Tiang M. Proportion of energy attributable to Dam Fe-oxide Cu-U-Au-Ag deposit. 2017;(September).
comminution. Proc 11th Australas Inst Min Metall Mill Oper Conf. 2012;(October): [44] BHP. Existing operation (Olympic Dam Copper refinery). 2009.
25–30. [45] Boisvert JB, Rossi ME, Ehrig K, Deutsch CV. Geometallurgical modeling at olympic
[32] Mudd GM. Prediction of Greenhouse Gas Emissions for the Olympic Dam Mega- dam mine, South Australia. Math Geosci. 2013;45(8):901–25.
Expansion. 2009;(April):1–13. [46] Ballantyne GR, Powell MS. Benchmarking comminution energy consumption for
[33] Alexander DJ, Wigley P. Flotation circuit analysis at WMC Ltd Olympic Dam the processing of copper and gold ores. Miner Eng 2014;65:109–14.
Operation. 8th Mill Oper Conf [Internet]. 2003;(July):41–51. Available from: [47] Crundwell F, Moats M, Robinson T, Ramachandran V, Davenport W. Extractive
https://apps.webofknowledge.com/full_record.do?product=UA&search_ Metallurgy of Nickel and Cobalt. Extractive Metallurgy of Nickel,Cobalt and
mode=GeneralSearch&qid=31&SID=F2T4mXddXK9AcD9TAX2&page=12& Platinim Group Metals. 2011. 489–534 p.
doc=593&cacheurlFromRightClick=no. [48] AEMO. AEMO | Carbon Dioxide Equivalent Intensity Index [Internet]. [cited 2021
[34] Vignes A. Extractive metallurgy 2: metallurgical reaction processes. John Wiley & Mar 9]. Available from: https://aemo.com.au/en/energy-systems/electricity/
Sons; 2013. national-electricity-market-nem/market-operations/settlements-and-payments/
[35] Coursol P, Mackey PJ. Energy consumption in copper sulphide smelting. Mater Sci settlements/carbon-dioxide-equivalent-intensity-index.
2010:649–68. [49] Belaissaouia Bouchra, Moullec Yann Le, Hagi Hayato, Favre Eric. Energy Efficiency
[36] Rich DW. Smelting practice at Olympic Dam. Conf Ser – Australas Inst Min Metall. of Oxygen Enriched Air Production Technologies: Cryogeny vs Membranes. Energy
1993;(April):207–15. Procedia [Internet]. 2014;63:497–503. Available from: www.sciencedirect.com.
[37] Cargill R, Freeman N, Gilbertson R. Metal balance for olympic dam using a [50] Banaszkiewicz T, Chorowski M, Gizicki W. Comparative analysis of cryogenic and
standard industrial software package. Australas Inst Min Metall Publ Ser 2002;8: PTSA technologies for systems of oxygen production. AIP Conf Proc. 2014;1573
149–56. (February 2015):1373–8.
[38] King MJ, Daveport WG, Moats MS. Sulphuric Acid Manufacture Analysis, Control & [51] Moore J. Thermal Hydrogen: An emissions free hydrocarbon economy. Int J
Optimization. 2013. Hydrogen Energy [Internet]. 2017;42(17):12047–63. Available from: https://doi.
[39] Nilay Kumar Sarker TAK. Simulation of the Production of Sulfuric Acid from a org/10.1016/j.ijhydene.2017.03.182.
Sulfur-burning Single-absorption Contact Sulfuric Acid Plant. Int J Emerg Technol [52] Home | Australian Institute of Petroleum [Internet]. [cited 2021 Mar 20]. Available
Innov Res JETIR. 2015;Vol. 2-I(Vol. 2-Issue 6 (July-2015)):2006–11. from: https://www.aip.com.au/.
[40] Cheng CY, Urbani M. The recovery of nickel and cobalt from leach solutions by [53] Gavin T, Ray S. Chemical Engineering Design Principles, Practice and Economics of
solvent extraction. 2005;(January 2005). Plant and Process Design. 2nd ed. 2012.
[41] Macmillan E, Ehrig K, Liebezeit V, Kittler P, Lower C. Use of geometallurgy to [54] Bruce S, Temminghoff M, Hayward J, Schmidt E, Munnings C, Palfreyman D, et al.
predict tailings leach acid consumption at Olympic Dam. GeoMet 2011 - 1st National Hydrogen Roadmap. 2019;116. Available from: www.csiro.au.
AusIMM Int Geometallurgy Conf 2011. 2011;(September):93–102.

16