Multicomponent Thermodynamic Databases for Complex Non-ferrous Pyrometallurgical Processes

Abstract

The pyrometallurgical production and recycling of non-ferrous metals involves the use of complex feed stocks, having a wide range of chemical compositions from sources that include mineral sulphide concentrates, high value obsolete materials and process wastes. The commercial viabilities of these operations hinge on the ability to extract value from these materials. Increasingly, modern computer-based tools are used to describe and predict process outcomes, including mass and heat balances, the partitioning of elements and phase equilibria . At the heart of these predictive tools are thermodynamic databases that describe the fundamental chemical properties of a system and all the components present. A comprehensive research program has been established to develop an accurate, self-consistent thermodynamic database describing all gas-slag -matte-metal-speiss-solid phases in the system Cu₂O-PbO-ZnO-Al₂O₃-CaO -MgO-FeO-Fe₂O₃-SiO₂-S-(As-Bi-Sb-Sn-Ag-Au ). The database can be used in conjunction with the FactSage computer platform. The accuracy of the database and its application to industrial practice is demonstrated.

Introduction

The non-ferrous pyrometallurgy industry is providing increasing volumes of primary metal and is recycling an ever-widening variety of metals. These developments bring with them a number of challenges including issues associated with source materials scarcity, the increasing compositional and structural complexity of modern devices, environmental impact and sustainability .

The rapid increases in computer power, the ability to collect and analyse large volumes of data, the ability to measure change and control equipment in industrial plant, has created opportunities to further improve the performance of existing operations. The implementation of practices that take advantage of accurate process models would enable improved process stability, process and feed optimization , improved campaign planning, and ultimately the widespread implementation of feed-forward process control systems in pyrometallurgical processes. These actions have the potential to increase process throughput, metal recoveries and to enable the efficient treatment of complex but profitable feed sources. The research program outlined in this paper, on the development of the robust thermodynamic databases and process models, is driven by the need to respond to the above challenges and with the aim of taking advantage of these new opportunities.

The availability of accurate thermodynamics databases is essential for the development of accurate predictive models and their use in the optimisation of pyrometallurgical processes. These databases are the foundations of the process models—and determine the quality of the predictive outcomes. A variety of fundamental data are required to develop these databases including, but not limited to, information on solidus, liquidus , phase equilibria , solid and liquid solubilities, distribution coefficients, thermodynamic activities, vapor pressures, enthalpy functions in multi-component, multi-phase systems.

The research program is planned to enable the development of databases that cover the whole range of compositions and process conditions in industrial operations and technologies used in copper , lead and zinc sulfide smelting (see Table 1). The databases includes gas, slag , matte, metal, speiss and solid phases. The major chemical components of slag (molten oxide) phase are described by the Cu₂O-PbO-ZnO-FeO-Fe₂O₃-SiO₂-S chemical system; Al₂O₃-CaO -MgO appear principally in smelting as contaminants or fluxes in the feed and As-Bi-Sb-Sn-Ag-Au are minor elements that also partition between all phases present in the systems. Major components in matte database include Cu-Fe–Pb-Zn-O-S, and in metal Cu-Fe–Pb-Zn alloys. Common solid solutions include spinels, melilites, zincite, in addition to stoichiometric compounds encountered in these systems (see Table 2).

Table 1

Overview of common phases and composition ranges observed in some pyrometallurgical Cu, Pb and Zn smelting and recycling processes

		Cu	Pb	Zn	S	Fe	SiO₂	Al₂O₃	CaO	MgO	As	Sn	Sb	Bi	Ag	Au
		wt%	wt%	wt%	wt%	wt%	wt%	wt%	wt%	wt%	wt%	wt%	wt%	wt%	ppm	ppm
Cu smelting	Gas/dust	15–25	0.1–2	0.1–2	5–15	15–25	5–10	1–3	0.1–2	0.1–2	1–3		0.05–0.15	0.2–1.1
	Slag	0.5–1			0.5–2	30–45	30–45	2–5	1–3	0.3–3
	Matte	50–70	0.1–5	0–2	20–26	10–25	0–1	0	0	0	0–0.5		0–0.5	0–0.1	0–3000	0–30
Cu direct-to-blister Cu converting	Slag	15–25	0.1–5	0.05–2	0.1–1	30–40	15–30	2–5	1–3	0.3–3
	Fayalite Slag	2–5	0–1	0–0.5	0.2–1.5	35–50	30–40	0–0.5	0–0.5	0–0.5
	Copper blow slag	20–45	0–1	0–0.5	0–0.3	30–40	10–20	0–0.1	0–0.1	0–0.1
	Ca-ferrite slag	15–25	0.1–1.0	0–0.1	0–0.3	35–45	0.1–1.5	0–0.5	15–20	0–0.1
	Matte	50–80	0.1–5	0–2	19–25	0.5–10	0–0.1	0	0	0	0–0.5		0–0.5	0–0.1	0–3000	0–30
	Blister	98.0–99.7	0–0.01	0	0.01–1	0–0.01	0	0	0	0	0–0.03				0–5000	0–40
Pb Sintering	Pb sinter	0.3–1.5	35–60	4–10	1–43	8–14	5–13	0.5–2.5	2–11	0.3–2	0.1–0.5		0.1–1	0.01–0.1
Pb smelting	Pb smelting slag	0.3–1.5	35–55	3–7	0–1	5–15	20–40	0.5–2	2–8	0.1–1
Pb Blast Furnace	Pb reduction slag	0.1–0.7	1–3	10–20	1–3	20–30	20–25	2–5	10–20	0.5–2	0–0.1
	Pb bullion	0–4	94–98	0–0.1	0–1	0–0.01					0–1
	Matte	10–40	10–30	3–13	5–15	3–25					0.5–2
Zn fuming	Slag	0.01–0.3	0.01–2	0–20	0.4–0.7	20–25	25–30	5–8	13–20	3–7
Cu and Pb refining	Speiss	20–30	30–50	0–5	3–10	0–3					5–12	0–3	1–3	0–0.4	0–1000	0–30
Common solid solutions	Spinels	0–0.1	0–0.5	0–25	0	55–70		0–10	0–0.1	0–5
Common solid solutions	Melilite	0	0–5	1–20	0	20–30	20–25	1–5	20–25	2–5

Table 2

Solution phases important for sulphide smelting of Cu, Pb, Zn and the thermodynamic models used to describe these in the database. For details see Refs. [1–4]

Liquid Slag :
(Al⁺³, Ca⁺², Mg⁺², Si⁺⁴,Cu⁺¹, Fe⁺², Fe⁺³, Pb²⁺, Zn²⁺, Sn²⁺, Sb⁺³, As⁺³, Bi⁺³, Ag⁺¹, Au ⁺¹,)(O⁻², S⁻²), Modified Quasichemical Formalism (MQF) in Quadruplet Approximation Spinel: [Cu⁺², Fe⁺², Fe⁺³, Al⁺³, Mg⁺², Zn⁺²]^tetr[Cu⁺², Fe⁺², Fe⁺³, Al⁺³, Ca⁺², Mg⁺², Zn⁺², Vacancy⁰]₂^octO₄, Compound Energy Formalism (CEF) Monoxide: (FeO, FeO_1.5, CuO, AlO_1.5, CaO , MgO), Bragg-Williams model (B-W) Olivine: [Fe²⁺, Ca²⁺, Mg²⁺, Zn²⁺]^M2[Fe²⁺, Ca²⁺, Mg²⁺, Zn²⁺]^M1SiO₄, CEF Dicalcium silicates: (Ca₂SiO₄, Fe₂SiO₄, Mg₂SiO₄, Pb₂SiO₄, Zn₂SiO₄), B-W Wollastonite: (CaSiO₃, FeSiO₃, MgSiO₃, ZnSiO₃), B-W Melilite: [Ca²⁺, Pb²⁺]₂[Fe²⁺, Fe³⁺, Al³⁺, Zn²⁺][Fe³⁺, Al³⁺, Si⁴⁺] ₂O₇, CEF Willemite: [Zn²⁺, Fe²⁺, Mg²⁺][Zn²⁺, Fe²⁺, Mg²⁺]SiO₄, CEF Zincite: (FeO, ZnO, MgO), B-W Corundum: (FeO_1.5, AlO_1.5), B-W Mullite: [Al⁺³, Fe⁺³]₂[Al⁺³, Si⁺⁴, Fe⁺³][O⁻², Vacancy]₅, CEF Calcium ferro-aluminates Ca(Al, Fe)₂O₄, Ca(Al, Fe)O₇, Ca(Al, Fe)₁₂O₁₉ Pyroxenes: [Fe²⁺, Ca²⁺, Mg²⁺]^M2[Fe²⁺, Fe³⁺, Mg²⁺, Al³⁺]^M1[Fe³⁺, Al³⁺, Si⁴⁺]^BSi^AO₆, CEF Liquid metal/matte/speiss: (Cu^I, Cu^II, Fe^II, Fe^III, Pb^II,As^III, Zn^II, Sn^II, Sb^III, Bi^III, Ag^I, Au^I, O^II,S^II), MQF in Pair Approximation Digenite-bornite: (Cu₂S, FeS, PbS, ZnS, Vacancy₂S) ‘Cu₃As’, (Cu, As), MQF in Pair Approximation fcc and bcc solid alloys: (Co , Ni, Mn, Cu, Fe, Pb, O, S, Zn, As, Sb, Ag, Au ), B-W Ideal gas:> 100 species, including N₂, CO , CO ₂, S₂, SO₂, H₂O, AsO, AsS, As₄O₆, PbO, PbS Stoichiometric compounds: > 150, including SiO₂, FeS₂, CaSO₄, CuFeO₂, Ca₃Al₂O₆, S, ZnS, Sb₂O₃

Liquid Slag :

(Al⁺³, Ca⁺², Mg⁺², Si⁺⁴,Cu⁺¹, Fe⁺², Fe⁺³, Pb²⁺, Zn²⁺, Sn²⁺, Sb⁺³, As⁺³, Bi⁺³, Ag⁺¹, Au ⁺¹,)(O⁻², S⁻²),

Modified Quasichemical Formalism (MQF) in Quadruplet Approximation

Spinel: [Cu⁺², Fe⁺², Fe⁺³, Al⁺³, Mg⁺², Zn⁺²]^tetr[Cu⁺², Fe⁺², Fe⁺³, Al⁺³, Ca⁺², Mg⁺², Zn⁺², Vacancy⁰]₂^octO₄,

Compound Energy Formalism (CEF)

Monoxide: (FeO, FeO_1.5, CuO, AlO_1.5, CaO , MgO), Bragg-Williams model (B-W)

Olivine: [Fe²⁺, Ca²⁺, Mg²⁺, Zn²⁺]^M2[Fe²⁺, Ca²⁺, Mg²⁺, Zn²⁺]^M1SiO₄, CEF

Dicalcium silicates: (Ca₂SiO₄, Fe₂SiO₄, Mg₂SiO₄, Pb₂SiO₄, Zn₂SiO₄), B-W

Wollastonite: (CaSiO₃, FeSiO₃, MgSiO₃, ZnSiO₃), B-W

Melilite: [Ca²⁺, Pb²⁺]₂[Fe²⁺, Fe³⁺, Al³⁺, Zn²⁺][Fe³⁺, Al³⁺, Si⁴⁺] ₂O₇, CEF

Willemite: [Zn²⁺, Fe²⁺, Mg²⁺][Zn²⁺, Fe²⁺, Mg²⁺]SiO₄, CEF

Zincite: (FeO, ZnO, MgO), B-W

Corundum: (FeO_1.5, AlO_1.5), B-W

Mullite: [Al⁺³, Fe⁺³]₂[Al⁺³, Si⁺⁴, Fe⁺³][O⁻², Vacancy]₅, CEF

Calcium ferro-aluminates Ca(Al, Fe)₂O₄, Ca(Al, Fe)O₇, Ca(Al, Fe)₁₂O₁₉

Pyroxenes: [Fe²⁺, Ca²⁺, Mg²⁺]^M2[Fe²⁺, Fe³⁺, Mg²⁺, Al³⁺]^M1[Fe³⁺, Al³⁺, Si⁴⁺]^BSi^AO₆, CEF

Liquid metal/matte/speiss:

(Cu^I, Cu^II, Fe^II, Fe^III, Pb^II,As^III, Zn^II, Sn^II, Sb^III, Bi^III, Ag^I, Au^I, O^II,S^II), MQF in Pair Approximation

Digenite-bornite: (Cu₂S, FeS, PbS, ZnS, Vacancy₂S)

‘Cu₃As’, (Cu, As), MQF in Pair Approximation

fcc and bcc solid alloys: (Co , Ni, Mn, Cu, Fe, Pb, O, S, Zn, As, Sb, Ag, Au ), B-W

Ideal gas:> 100 species, including N₂, CO , CO ₂, S₂, SO₂, H₂O, AsO, AsS, As₄O₆, PbO, PbS

Stoichiometric compounds: > 150, including SiO₂, FeS₂, CaSO₄, CuFeO₂, Ca₃Al₂O₆, S, ZnS, Sb₂O₃

The databases that are constructed contain fundamental descriptions of the chemical behavior of the systems and are independent of the technology that is used. This means that the databases can be used at the heart of predictive models of pyrometallurgical processes with additional parameters that take into account factors related to furnace design, construction and operation. The database can be used to predict the outcomes of copper , lead and zinc sulfide smelting processes, including copper smelting ; copper converting ; copper refining in anode furnaces; lead sintering and smelting ; lead reduction ; zinc and lead fuming; and copper and lead refining to extract precious metals . The databases can be used to describe reactions in suspended and bath smelting , blast furnaces; batch and continuous processes.

Outline of the Overall Program

In order to develop the databases for the complete range of conditions in copper , lead and zinc pyrometallurgy , an integrated research program of thermodynamic modelling and experimental measurements of phase equilibria is being undertaken.

The extensive program for lead consortium companies involves a number of focussed projects,

1.
Slag /metal phase equilibria in the PbO-ZnO-CaO -FeO-Fe₂O₃-SiO₂-Cu_xO-Al₂O₃-MgO slag —Pb-Cu-Fe-Zn metal alloy.
2.
Matte formation conditions within the gas/PbO-ZnO-CaO -FeO-Fe₂O₃-SiO₂-Cu_xO-Al₂O₃-MgO slag and the Pb-Zn-Fe-O-S-Cu matte/alloy systems.
3.
Pb refining systems equilibria within the Pb-Cu-S-As-Sb-Sn-Fe matte/metal/speiss system.
4.
Elemental Distributions between slag , matte and metal of minor elements , including Ag, Au , As, Bi, Sn, Sb and Zn.
5.
Improvement of the thermodynamic database of oxide systems: Development of the thermodynamic database of the PbO-ZnO-CaO -FeO-Fe₂O₃-SiO₂-Cu_xO-Al₂O₃-MgO slag in Pb-Cu-Fe-Zn metal alloy systems.
6.
Improvement of thermodynamic database of sulphur-containing systems: Development of the matte/metal/speiss thermodynamic database (focus on low temperatures) and the incorporation of minor elements .

These projects cover a wide range of chemical compositions and conditions (temperature , oxygen and sulphur partial pressures) relevant to the whole range of key lead smelting , refining and recycling systems.

For copper consortium companies the scope of the work includes complete experimental revision of the thermochemistry of the base system “Cu₂O”-FeO-Fe₂O₃-SiO₂-S with the Al₂O₃, CaO and MgO slagging components and As-Pb-Zn-Sn-Sb-Bi-Ag-Au other minor elements . The experiments involve determining equilibria between gas/slag /matte/blister/solid (tridymite, spinel) phases as functions of temperature , P(O₂), P(SO₂)/P(S₂), slag Fe/SiO₂ (and equilibria with tridymite or spinel). This development of thermodynamic database working with the FactSage software [5] for the above system provides direct support of the copper smelting industry sponsors.

To investigate the distribution of minor elements between phases, it is essential to initially accurately characterise the base system, and then systematically investigate the effect of all of the key operating parameters on the thermochemistry of all phases within the selected range of chemical compositions. Two types of experiments are performed:

Open experiments with P(O₂) and P(SO₂) in the gas/slag /matte and gas/slag /metal systems controlled by the CO /CO ₂/SO₂ gas mixtures, and
Closed experiments undertaken in sealed ampoules for the slag /matte/metal system.

The overall program therefore contains the following key directions:

1.
Base system with slagging components—initial description at Matte Grades between 50 and 80% Cu
Gas/Slag /Matte,
Gas/Slag /Metal [Cu-Fe-O-S-Si] x [Temperature ] x [P(O₂)/P(SO₂)] x [Fe/SiO₂—Tridymite/Spinel] x [Al, Ca, Mg].
Slag /Matte/Metal.
2.
Distribution of Minor Elements As, Zn, Pb, Sn, Sb, Bi, Ag, Au
Gas/Slag /Matte,
Gas/Slag /Metal [Cu-Fe-O-S-Si] x [Temperature ] x [P(O₂)/P(SO₂)] x [Fe/SiO₂—Tridymite/Spinel] x [Al, Ca, Mg]
Slag /Matte/Metal.

The techniques developed during this program for the first time enable the systematic accurate measurements of this kind to be undertaken, and these measurements provide an important foundation for the development of the thermodynamic database as well as for the overall quantitative description of the thermochemistry of copper smelting

The total number of experiments needed to completely and quantitatively characterise the whole chemical system as functions of key operational parameters is very large. The experimental needs are therefore carefully and critically reviewed. An overall summary of all required experiments is prepared and continuously revised to enable systematic analysis and selection of optimum research plan to support the development of the thermodynamic database , to close the gaps where no data is available and to resolve discrepancies.

An example of the systematic approach undertaken for the planning of the copper consortium experimental program is given in the following paragraph. A list of experiments needed to characterise the whole chemical system as functions of the key operational parameters is selected, as illustrated in Table 3:

Table 3

An example of the systematic approach taken to experimental study and modelling for copper database development

../images/468727_1_En_68_Chapter/468727_1_En_68_Tab3_HTML.gif

Columns 2 through 8 indicate the matrix of key parameters selected for characterisation
Columns 9 and 10 indicate current status of system description

Each line in Table 3 represents a series of experiments. For example,

line 1.1 represents gas/slag /matte equilibria on a tridymite substrate for a range of matte grades from 55 to 80 wt%Cu at base case conditions of 1200°C and P(SO₂) = 0.25 atm.
line 1.2 presents the effect of temperature in the above system
line 1.3—the effect of P(SO₂) in the above system, as so on for all key parameters.

In this way, the whole list of the experiments needed to quantitatively characterise the system as functions of the key parameters is being carefully analysed, and most critical experiments are selected and performed to support the development of the thermodynamic database using FactSage package [5].

The next stage of the program is to combine and integrate the copper and lead databases. Figure 1 illustrates the number of binary, ternary and higher order systems that must be accurately described. As the number of components are added to the system so the number of chemical interactions and sub-systems that must be described increases.

../images/468727_1_En_68_Chapter/468727_1_En_68_Fig1_HTML.gif — Fig. 1
Binary, ternary and higher order sub-systems in the system Cu₂O-PbO-ZnO-Al₂O₃-CaO -MgO-FeO-Fe₂O₃-SiO₂-S-(As-Bi-Sb-Sn-Ag-Au )

Demonstrating Capabilities

Complete Description of the Copper-Matte-Slag -Gas System Cu₂O-Al₂O₃-CaO-MgO-FeO-Fe₂O₃-SiO₂-S

The databases provide detailed descriptions of a range of effects with high accuracy. The agreements between experimental data and databases are demonstrated in Fig. 2a–g for slag /matte systems at 1200 °C. The sensitivity of the systems is exemplified by the effect of oxygen pressure on copper matte grade is shown in Fig. 2a. The extent of non-stoichiometry of matte and the relatively small influence of temperature on this factor is shown in Fig. 2b. The Fe³⁺/Fe_total ratio in slag in equilibrium with mattes of varying grades and the influences of relatively small concentrations of MgO are given in Fig. 2c. The compositions of the slag liquidus as a function of matte grade for the limiting conditions of silica and spinel saturation are given in Fig. 2d. The effect of alumina in slag on the dissolved sulphur in slag is provided in Fig. 2e. The predicted and experimentally determined concentrations of copper dissolved in slag are shown in Fig. 2f, illustrating the complex non-linear behavior with changing matte grade. Figure 2g shows the concentration of dissolved oxygen in matte as a function of matte grade. Please note that the effect of P(SO₂), temperature , MgO, CaO and Al₂O₃ in slag is available now for all 7 diagrams in Fig. 2, but not shown due to limitations of the present article.

../images/468727_1_En_68_Chapter/468727_1_En_68_Fig2_HTML.gif — Fig. 2
Calculated lines and experimental data [6–15] obtained for gas/matte/slag equilibria in the Cu-Fe-O-S-Si system + Al₂O₃, CaO , MgO

Resolving Complexities in Database Development

As indicated, in constructing the optimised databases for these multi- component systems, it is necessary to obtain not just information on the phase equilibria but also to ensure that the parameters selected are consistent with other thermodynamic properties. An example of this is shown in Figs. 3a–d. The binary phase diagram for the PbO–SiO₂ system at low temperatures has been investigated and characterised in numerous studies. The liquidus in the silica primary phase fields has remained uncertain until recent research due to the difficulties associated with high vapour pressures of lead species at high temperatures in this system. Many combinations of enthalpy and entropy contributions can be selected to describe the shape of the liquidus . However these parameters must also be able to describe other thermodynamic data in related systems, such as activities of the components and elemental distributions between phases. These properties should be consistent with experimental data in the Pb–Fe-Si-O and Cu-Pb–Fe-O-S systems as shown in Fig. 3b–d. The key point to note from this example is that information from different types of thermodynamic properties is required and must be described in order to obtain accurate parameters for the database; phase equilibrium data on their own are not sufficient to identify the unique parameters required to unambiguously describe the system.

../images/468727_1_En_68_Chapter/468727_1_En_68_Fig3_HTML.gif — Fig. 3
Examples of the different types of information required to select modelling parameters that are valid across the whole range of compositions and process conditions for the lead -matte-slag -gas system [16, 17]

To effectively manage the complexity of the 16 component system Cu₂O-PbO-ZnO-Al₂O₃-CaO -MgO-FeO-Fe₂O₃-SiO₂-S-(As-Bi-Sb-Sn-Ag-Au ) the database development has been organised by splitting the system into two parallel, yet integrated tasks. One involves the experimental investigation and database development of the Cu-Fe-O-Pb-Zn-Ca–Si + (Ca–Mg) systems, that is the base slag /metal systems [18]. The second task is the optimisation of the slag /matte/speiss equilibria and minor element distributions between these phases.

Applications to Industrial Systems

Fluxing Diagrams

Having prepared the updated and optimised thermodynamic databases they can be used in conjunction with the FactSage computer platform to predict specific process outcomes.

Figure 4a shows how the liquidus surface in the Fe-SiO₂ system at P(SO₂) = 0.5 atm., 60%Cu matte grade changes with the presence of impurity elements Al₂O₃, CaO , MgO in the slag . It can be clearly seen that the liquidus on this pseudo-binary section moves significantly to lower Fe/SiO₂ ratio in both the spinel and silica primary phase fields with the presence of these impurities. These effects have significant implications for practice since they demonstrate the need to adjust the fluxing requirements of the process depending on the impurity levels in the slag ; potential savings on flux additions and minimising potential operational difficulties through operating at conditions that are below the liquidus and potentially creating tapping difficulties due to high %solids in slag . Figure 4b shows the sensitivity of the system to changing temperature ; here the % solids can be estimated as a function of Fe/SiO₂ ratio and temperature for a given set of operating conditions in copper converting .

../images/468727_1_En_68_Chapter/468727_1_En_68_Fig4_HTML.gif — Fig. 4
a Projection of the slag liquidus as a function of Fe/SiO₂ ratio for P(SO₂) = 0.5 atm. 60%Cu matte grade. b % solids in slag as a function of Fe/SiO₂ and temperature for P(SO₂) = 0.15 atm. 78%Cu matte grade

Pseudo-ternary Sections

The major components in lead blast furnace slags are CaO -FeO-SiO₂-ZnO; Al₂O₃ and MgO are also routinely present in the slags at low concentrations. Using the thermodynamic database the liquidus surface of the slag is predicted and shown in Fig. 5a, b) for a fixed CaO /SiO₂ wt. ratio and fixed Zn/Fe wt ratio respectively. The predictions make it possible to identify the sensitivity of the system to changes in slag composition. From Fig. 5a it can be seen that increasing the Zn/Fe ratio is limited by the zincite, (Zn,Fe)O, solid solution primary phase field, the liquidus temperature of which rises sharply with increasing zinc concentration. The optimum total flux (CaO +SiO₂) requirement for the CaO /SiO₂ of 0.72 would appear to be at approximately (CaO + SiO₂)/(CaO + FeO + SiO₂ + ZnO) = 0.45 by wt. In Fig. 5b) it can be seen that for Zn/Fe = 0.8 there appears to be a local minimum in liquidus temperature at CaO /SiO₂ wt. ratio between approximately 0.6–0.7.

../images/468727_1_En_68_Chapter/468727_1_En_68_Fig5_HTML.gif — Fig. 5
Predicted liquidus surfaces of the CaO -FeO-SiO₂-ZnO-Al₂O₃-MgO slags at lead metal saturation for a CaO /SiO₂ wt. ratio = 0.72, b Zn/Fe wt. ratio = 0.8

Process Simulation

The thermodynamic databases and computer software that are now available can be applied to not only the determination of equilibria particular cases but also to the prediction and simulation of dynamic and multi-stage metallurgical processes. The databases now include critical information on a wide range of elements present in complex industrial process systems. Increasing the number of elements present in the system increases the computation time required to undertake the Gibbs Free Energy minimisation calculation routines particularly with the formation of complex solutions. This is potential limitation to the application of this technology to practice so it is worthwhile examining these issues and how they might be addressed.

By way of illustration we consider the simulation or modelling of a batch process. As in any thermodynamic calculation the first consideration is definition of the system boundaries. If this is a closed system, with no material of energy transfer taking place, then the outcomes of the reaction can be calculated in a single stage. In practice, this is rarely the case, material and energy is introduced or removed throughout the reaction; how can we deal with this and what are the implications of different calculation strategies?

Take the example of the Peirce-Smith copper converting process. In this case there is an oxygen containing gas blast into molten matte phase, sulphur containing product gas that is removed from the system, slag phase generation and periodic removal , and enthalpy losses associated with product removal and heat losses throughout the process. An example of a detailed calculation incorporating all of the stages undertaken in typical Peirce Smith converter was reported in [19]. Figure 6 summarized the process inputs and outputs of the calculations of the present study. These values were obtained by considering the system as a series of linked equilibrium batch processes; the time interval between each iteration was relatively short (5 min); gas was removed at each iteration. Each stage ended with the slag removal ; fluxes, reverts, copper scrap were added several times during one stage.

../images/468727_1_En_68_Chapter/468727_1_En_68_Fig6_HTML.gif — Fig. 6
Example of the process simulation of a typical Peirce-Smith converter cycle, consisting of two slag blow stages and one copper blow stage

This example demonstrates the capability of the current thermodynamic databases and computer modelling systems. In the present article, calculations were performed for the Cu-Fe-O-S-Si-(Pb-Zn-As-Bi-Sb-Sn-Ag-Au )-C-N-H chemical system, which is of interest for industrial practice. What becomes clear is that, i) the greater the number of chemical components the longer it takes to complete the calculation, and ii) the accuracy of the calculations depend on the time or increment steps considered. The smaller the steps between iterations the greater the accuracy but the longer the computational time. The time response is an important consideration if we are to introduce these tools into industrial practice. We need to select the calculation approach that is appropriate for the application. For example, an on-line “logistics” model to optimise ladle movement and charging cycles requires a rapid response, which is more important than accuracy of the predictions below a given level of uncertainty. On the other hand, accurate predictions may be required to maximise minor element recovery or impurity removal .

The sensitivity of the predictions to the number of iterations or calculation steps is illustrated in Fig. 7a, b). It can be seen that there are significant differences between the single step calculation and many step cases for the predicted mass of slag produced in the copper blow stages and the final sulphur in blister. Detailed analysis shows that there will be less slag and lower sulphur in the blister as the number of calculation steps is increased. Since the gas phase is continuously removed from the system, there will be also differences in the total masses of volatile metal species, such as lead and arsenic , reporting to the gas phase depending on the calculation strategy. A single calculation provides only the overall enthalpy change in the system, however it is desirable to be able to predict variations in temperature of the reactor during processing. Higher reaction temperature leads to increased attack on refractories and reduced lining life; decreases in temperature can result in the formation of solids in the slag with consequent increases in slag viscosity . From a process control perspective further detailed information is necessary to understand and avoid these potential operating difficulties.

../images/468727_1_En_68_Chapter/468727_1_En_68_Fig7_HTML.gif — Fig. 7
The effect of the number of calculation steps on the predicted outcomes of a typical Peirce-Smith converter cycle, consisting of two slag blow stages and one copper blow stage, a mass of copper blow slag , and b the concentration of sulphur in final blister copper

It can be clearly seen from these examples that the use of accurate thermodynamic databases coupled with predictive power of computer platforms provides opportunities for optimization and improvement of metallurgical processes. Specific applications include:

Optimising processing conditions, e.g. fluxing, temperature , oxygen to feed ratio and fuel utilisation.
Increasing process stability and throughput.
Improving metal recovery , through reduced losses of dissolved and entrained metal in slags.
Removing unwanted impurity elements, increasing and optimising complex feed intake.
Improving vessel integrity/lining life by reducing operating temperatures, tuning slag and refractory compositions, and stabilising operating conditions.

These predictive capabilities are already being used in industry. See an example of case study of performance of ISASMELT reactor with different types of converters [20]. As confidence in the use of these tools is increased and predictions verified by plant practice , the opportunities for the introduction of the next phase of implementation will arise—the introduction of feed forward control of pyrometallurgical reactors.

Summary

An integrated research program of thermodynamic modelling and experimental measurements of phase equilibria is being undertaken to develop accurate and internally-consistent thermodynamic databases that can be used to describe the chemical equilibria in copper , lead and zinc smelting , metal production and recycling systems. The databases contain fundamental thermodynamic descriptions of the gas-slag -matte-metal-speiss-solids phases Cu₂O-PbO-ZnO-Al₂O₃-CaO -MgO-FeO-Fe₂O₃-SiO₂-S-(As-Bi-Sb-Sn-Ag-Au ). The program has provided valuable new fundamental information on phase equilibria and thermodynamic properties of complex multi-component, multi-phase systems.

Acknowledgements

The authors would like to thank Australian Research Council Linkage program, Altonorte Glencore, Atlantic Copper , Aurubis, BHP Billiton Olympic Dam Operation, Kazzinc Glencore, Nyrstar, PASAR Glencore, Outotec (Espoo and Melbourne), Anglo-American Platinum , and Umicore for the financial and technical support.

References

1.
Shishin D, Jak E, Decterov SA (2015) Thermodynamic assessment and database for the Cu–Fe–O–S system. CALPHAD 50:144–160Crossref
2.
Hidayat T, Shishin D, Decterov SA, Jak E (2016) Critical thermodynamic re-evaluation and re-optimization of the CaO-FeO-Fe₂O₃-SiO₂ system. CALPHAD 56:58–71Crossref
3.
Hidayat T, Shishin D, Decterov S, Jak E (2017) Critical assessment and thermodynamic modeling of the Cu-Fe-O-Si system. CALPHAD 58:101–114Crossref
4.
Shishin D, Prostakova V, Jak E, Decterov S (2016) Critical assessment and thermodynamic modeling of the Al–Fe–O system. Metall Mater Trans B 47:397–424Crossref
5.
Bale CW, Belisle E, Chartrand P, Decterov SA, Eriksson G, Gheribi AE, Hack K, Jung IH, Kang YB, Melancon J, Pelton AD, Petersen S, Robelin C, Sangster J, Spencer P, Van Ende MA (2016) FactSage thermochemical software and databases, 2010-2016. CALPHAD 54:35–53Crossref
6.
Fallah-Mehrjardi A, Hidayat T, Hayes PC, Jak E (2017) Experimental investigation of gas/slag/matte/tridymite Equilibria in the Cu-Fe-O-S-Si system in controlled gas atmospheres: Experimental Results at T = 1473 K [1200 °C] and P(SO₂) = 0.25 atm. Metall Mater Trans B 48:3017–3026Crossref
7.
Fallah-Mehrjardi A, Hayes PC, Jak E (2018) Experimental investigation of gas/slag/matte/tridymite equilibria in the Cu-Fe-O-S-Si-Ca system in controlled gas atmospheres: Experimental results at T = 1473 K [1200 °C] and P(SO₂) = 0.25 atm, submitted to Metall Mater Trans B
8.
Fallah-Mehrjardi A, Hidayat T, Hayes PC, Jak E (2018) Experimental investigation of gas/slag/matte/tridymite equilibria in the Cu-Fe-O-S-Si system in controlled gas atmospheres: experimental results at T = 1473, 1523, 1573 K [1200, 1250, 1300°C] and P(SO₂) = 0.25, submitted to Metall Mater Trans B
9.
Fallah-Mehrjardi A, Hayes PC, Jak E (2017) Experimental investigation of gas/slag/matte/tridymite equilibria in the Cu-Fe-O-S-Si-Al system in controlled gas atmospheres: experimental results at T = 1473 K [1200°C] and P(SO₂) = 0.25 atm, unpublished
10.
Hidayat T, Fallah-Mehrjardi A, Hayes PC, Jak E (2018) Experimental investigation of gas/slag/matte/spinel equilibria in the Cu-Fe-O-S-Si system at T = 1200 °C and P(SO₂) = 0.25 atm. Metall Mater Trans B, ahead of print
11.
Tavera FJ, Bedolla E (1990) Distribution of copper, sulfur, oxygen and minor elements between silica-saturated slag, matte and copper—experimental measurements. Intl J Miner Process 29:289–309Crossref
12.
Takeda Y (1997) Oxygen potential measurement of iron silicate slag-copper-matte system. In: Proceedings international conference on molten slags, fluxes salts, iron and steel society Warrendale, PA, pp 735–743
13.
Kuxmann U, Bor FY (1965) Studies on the solubility of oxygen in copper mattes under ferric oxide slags saturated with silica. Erzmetall 18:441–450
14.
Korakas N (1964) Etude thermodynamic de l’équilibre entre scories ferro-siliceuses et mattes de cuivre. Application aux problèmes posés par la formation de magnetite lors du traitement des minerais sulfurés de cuivre. PhD thesis Univirsité de Liège
15.
Hidayat T, Fallah-Mehrjardi A, Hayes PC, Jak E (2018) Experimental investigation of gas/slag/matte/spinel equilibria in the Cu-Fe-O-S-Si system at T = 1250 °C and P(SO₂) = 0.25 atm. Metall Mater Trans B, ahead of print
16.
Shevchenko M, Jak E (2018) Experimental phase equilibria studies of the PbO-SiO₂ system. J Amer Ceram Soc 101:458–471Crossref
17.
Hollitt MJ, Willis GM, Floyd JM (1984) Thermodynamics of the silica-saturated Pb-Fe-O-SiO₂ system at 1200 °C. In: 2nd international conferenceon metallurgical slags and fluxes, pp 497–516
18.
Shevchenko M, Hayes PC, Jak E (2018) Development of a thermodynamic database for the multicomponent PbO-“Cu₂O”-FeO-Fe₂O₃-ZnO-CaO-SiO₂ system for pyrometallurgical smelting and recycling. In: Peter Hayes Symp on Pyrometallurgy, TMS/CIM, Ottawa
19.
Shishin D, Hidayat T, Decterov S, Jak E (2016) Thermodynamic modelling of liquid slag-matte-metal equilibria applied to the simulation of the Peirce-Smith converter. In: 10th international conference on molten slags, fluxes and salts, Seattle, USA
20.
Nikolic S, Shishin D, Hayes PC, Jak E (2018) Case study on the application of research to operations—calcium ferrite slags. Extraction 2018, Ottawa, Canada