Structural Changes in Copper Slags During Slow Cooling

Abstract

The objects of the study were converter slags from the Balkhash copper plant in their initial state and after heat treatment. Using mineralogical and X-ray phase analysis, scanning electron microscopy (SEM), and electron probe microanalysis (EPMA), it was found that the initial converter slag and its thermally treated samples have identical matrices with almost complete coincidence in mineral and phase compositions. The distinguishing feature is the quantitative ratio of mineral components in the slag mass. Almost all of the iron is oxidized and present in the form of fayalite, magnetite, and magnetite, with other elements (silicon, copper, zinc, and aluminum) incorporated into its lattice. The structure of all slag samples indicates an association of sulfur exclusively with copper. Copper in the slags was identified in both metallic and sulfide forms. Slow cooling of the converter slag after its remelting contributes to the reduction in the sulfide–metal suspension in the volume of the melt and its coarsening. During slow cooling, structural changes occur not only in the main oxide part of the slag but also in the polymetallic globules.

1. Introduction

The task of increasing copper recovery from copper slags is currently quite relevant.

In the research of Kenzhaliyev, B. et al. [1], a method for reducing copper content in copper smelting slags by improving the design of the unit is described. Two improvement options are proposed: electric heating of the slag siphon with graphite electrodes and the depletion process in the dual-zone Vanyukov furnace. In both cases, the copper content in the slags decreased from 1.0% to 0.43% and the magnetite content from 7.95% to 2.5%. In the work of Argyn, A. et al. [2], the results of research on the processing of copper–lead mattes with high-sulfur copper concentrate in a converter to obtain slags suitable for enrichment are presented. The influence of sulfur on the destruction of magnetite during the converting process was established, and new data on the distribution of impurities of non-ferrous and accompanying metals (As, Sb, etc.) were obtained. High recovery rates of non-ferrous metals and impurities into the target products were also established: copper into matte was up to 98%, and lead, zinc, arsenic, and antimony into dust was 87%, 91%, 84%, and 38%, respectively. In the article by Zoldasbay, E.E. et al. [3], it is shown that traditional methods used to calculate autogenous smelting are not entirely accurate and require consideration of the sulfur excess effect on the temperature regime of the process. It was established that during the conversion of copper–lead mattes, there is a wide range of temperature variations—from 1027 to 1300 °C. When combining concentrate with matte, the temperature regime of the process stabilizes, ensuring the optimal SO₂ concentration in the gases and the optimal slag composition. Based on the material balance calculation of copper–lead matte conversion by the existing technology and with the addition of concentrate, the structure of the thermal balance of the conversion process was established. In the article by Yakubova M. et al. [4], the possibility of involving slag and clinker in the pyrometallurgical production of copper at Almalyk Mining and Metallurgical Complex (AO “Almalyk GMK”) to comprehensively extract non-ferrous and precious metals was demonstrated. It was found that the use of the clinker contributed to the reduction in magnetite (the amount of magnetite decreased from 21.9% to 9.8%). As a result of processing the recovered converter slags in the Vanyukov furnace, the copper content in the converter slags was reduced from 2.2–3.5% to 0.58–0.72%, which facilitated their further processing.

Today, one of the most common methods is the slow, controlled cooling of slags before flotation enrichment. In the work of Gao, X. et al. [5], experiments were conducted with varying controlled cooling speeds and different additives. It was discovered that with the decrease in the cooling rate, the particle size of the matte phase enriched with copper increased. The influence of gypsum and carbon additives was found to be insignificant. According to Tshiongo, N. et al. [6], various cooling methods were studied, namely, water quenching, air cooling, and furnace cooling. The effect of cooling speed was investigated. The study noted the influence of slag cooling speed on phase distribution within the slag and the distribution of non-ferrous metals within these phases. Additionally, it was found that rapid slag cooling prevented crystallization and led to the formation of amorphous phases containing non-ferrous metals. According to Tshiongo, N. et al. [6], amorphous slags were more susceptible to leaching in an acidic medium. They also concluded that the total amount of non-ferrous metals leached from slowly cooled slags was significantly lower in comparison to quenched slag samples. It was also found that rapid cooling of the slag prevented crystallization and led to the formation of an amorphous phase that encapsulated non-ferrous metals. Amorphous slags leached better in an acidic medium (HNO₃), with the extraction into the solution in mass percentages: 46% Cu, 95% Co, 85% Zn, 92% Pb, and 79% Fe. The work of Zhang, H. et al. [7] examined the behavior of matte particles in the slag, along with the effect of settling time, system temperature, and the amount of reductant on the macrostructure of the slag, the microstructure of matte particles, and copper losses. The authors believe that during the settling of matte particles in the slag at the experimental temperature, aggregation of these particles occurs. Increasing the settling time and temperature, as well as the addition of anthracite, led to an increase in the degree of matte recovery. In the article by Gazaleeva, G.I. et al. [8], the slow cooling of slag is highlighted as a promising method that leads to improved copper recovery from slags. Industrial trials of a batch of slowly cooled slag at the enrichment plant of JSC “Sredneuralskiy copper smelter” confirmed the effectiveness of the method. As a result, copper recovery in the copper concentrate increased by 15–19%. In a study by Askarov, N.M. et al. [9], the drawbacks of slow-cooled slags are noted, particularly their high abrasiveness and hardness. These properties hinder the exposure of sulfide minerals, affecting the recovery of valuable components. According to Mamonov, S.V. et al. [10], it is established that the temperature regime and cooling rate significantly influence the slow cooling process of dump slag. It is noted that the best structural transformations in the slag occur within the temperature range of 1080 °C to 880 °C at a cooling rate of 5 to 10 °C per hour. It has been shown that slow cooling promotes an increase in the size of sulfide particles and redistributes copper among the mineral forms. Another study by Mamonov, S.V. et al. [11] demonstrated that the technology of slow slag cooling followed by processing increases copper recovery into concentrate by 15–22%. Additionally, grinding equipment productivity increases by 25%, and the degree of copper mineral exposure rises by 25–30%. The possibility of improving the beneficiation performance of dump slags from Vanyukov furnaces and shaft furnaces using ultrafine grinding in bead mills was also demonstrated. In [12], McKerrow et al. provide a detailed description of the stages in the controlled slow cooling process of copper slags. The main steps involve pouring the molten slag into a ladle, waiting for sufficient solidification, discharging the slag from the ladle, and primary crushing. It is also proposed to cool the upper part of the slag in the ladle with water. The authors of this work [12] claim that the optimal time for slow cooling should be no less than 24 h. Tsunazawa, Y. et al. [13] studied the effect of the slow cooling process on the growth of magnetite crystals in the slag. The authors concluded that as the cooling rate decreased, the mass fraction of magnetite increased along with the average grain size of magnetite. Slow cooling in an oxidizing environment and maintaining a certain temperature during this cooling process promoted the growth of magnetite crystals. The authors claimed that the slag, after slow cooling, should be ground to a particle size of less than 20 microns. Only in this case can satisfactory magnetite extraction be achieved through magnetic separation. An article by Guo, Z. et al. [14] describes a technology for improving the recovery of copper and iron-containing phases from copper slag by modifying it. The study explored binary basicity, the dosage of compound additives, modification temperature, cooling rate, and the final temperature of slow cooling. Results showed that under certain conditions, the use of modified slag increased the copper content in the rough copper concentrate from 6.4% to 11.04%. In the work of Mihajlović, A. et al. [15], a technological method of slow slag cooling is described, which is divided into two stages: 24 h air-cooling and 48 h accelerated work cooling. The authors claim that this technology yields the best results in stimulating the growth or coagulation of dispersed copper sulfide and elemental copper particles in the slag, therefore increasing the flotation process efficiency and reducing copper losses. At the same time, it is stated that copper recovery into blister copper from concentrates can reach 98.5%. According to Xue, P. et al. [16], Ausmelt furnace slags also employ slow-cooled slag for copper recovery. Factors affecting the efficiency of copper extraction were studied, such as grinding fineness, the pH value of the flotation medium, various collectors, and the flotation process. It was shown that the granulometric composition of the primary and secondary grinding of the industrial product consisted of 75% for particles smaller than 0.074 mm and 82% for particles smaller than 0.043 mm. The results of closed-circuit experiments with butyl xanthate as a collector showed that the copper content reached 16.11%, the copper recovery rate was 69.90%, and the copper content in the tails was only 0.2%. The publication of Cao, S. et al. [17] indicates that enrichment technology using slag after slow cooling is the most widely applied method for extracting valuable components. CaS was used to study the distribution of elements in copper smelting slag and its sulfuration. As the dosage of CaS increased, the distribution pattern of Fe, Cu, Zn, Pb, Se, Te, and Ag changed, while the distribution of Si, Ca, Al, and Mg changed insignificantly. When the excess CaS ratio was altered, the behavior of As and Bi distribution in the slag changed with the initial mass ratio of Fe/SiO₂ equal to 1.8. However, when the smelting slag had a mass ratio of Fe/SiO₂ of 1.0 and 1.3, an increase in CaS addition resulted in the glassy slag containing more As. Rahimi, E. et al. [18] do not question the potential of using slow-cooled slags. The article examines various methods of controlling the cooling rate. Calculations were carried out for slag cooling in ditch beds with a temperature change from 1200 to 20 °C, a slag layer height in the beds ranging from 1 to 36 cm, and a cooling time from several minutes to 62 h. In the publication of Fernández-Caliani, J.C. et al. [19], copper slags used at the Atlantic Copper Plant (Spain) were studied. The authors concluded that the main cause of copper losses in slags is the mechanical entrapment of matte particles by the slag melt. The highest percentage of matte particles containing copper (2.0–3.5 vol.%) is found in magnetite-rich converter slags. These particles consist of copper sulfide with a metallic copper core. As a result of the mechanical carryover of copper matte by slags from flash furnaces and electric furnaces, copper losses amount to less than 1.5 vol.%. In these slags, the matte particles have a composition close to the stoichiometry of bornite (Cu₅FeS₄). Coursol, P. et al. [20] believe that under optimal conditions, the total copper content in dump slag can be kept within 0.7–0.8 wt.%. Contrary to the previous study, this article argues that the main contributor to copper losses is copper dissolved in the slag (75%), while the mechanical losses of matte particles only account for 25%. The focus of the research of Das, B. et al. [21] was on copper recovery from slag produced by an Indian metallurgical plant, which contains 1.53% Cu, 39.8% Fe, and 34.65% SiO₂. A complete investigation of the various phases present in the slag was carried out using optical microscopy, scanning electron microscopy (SEM), and X-ray diffraction (XRD). It was noted that iron and silica are primarily associated with the fayalite phase, while copper is present in both the oxide and sulfide phases. Therefore, the literature presents conflicting information regarding the proportion of mechanical and chemical copper losses in copper smelting slags. This discrepancy is likely related to the specifics of the smelting technology used. Most authors agree that slow cooling of the slag before flotation extraction of copper sulfides is a necessary operation that promotes the coarsening of sulfide particles in the slag. However, the mechanism of this process has received little attention.

Thus, in publications dedicated to the topic of metallurgical slag processing, particularly slow cooling before enrichment, there is no description of the mechanism for the coarsening of matte inclusions. The novelty of our research lies in the search for the mechanism of sulfide inclusion coarsening during slow slag cooling. The goal is to determine the conditions for the coarsening of sulfide inclusions during slow slag cooling for subsequent flotation enrichment.

2. Materials and Methods

For research on the effect of slow cooling on slag structure, converter slags from the Balkhash Copper Smelter were chosen. Since converter slags contain more copper (3–5%) compared to furnace slags (0.7–0.9%), it was easier to track the behavior of sulfide inclusions during the slow cooling of the slag. A converter slag is heterogeneous, so several samples of a converter slag from the third tap of a single smelting were taken for research. The total weight of the collected samples was 1 kg. Part of the samples, in the form of pieces measuring 15–20 mm, was used for mineralogical and microprobe analyses, while the remaining part of the samples was ground to a size of 74 microns and used for other types of analysis. The composition of the slag is presented in

Table 1. Chemical composition of converter slags.

Element, Compound	Cu	Pb	Zn	Ni	Fe	SiO2	CaO	MgO	Al2O3	S
Content,%	4.97	5.3	3.3	0.05	41.3	20.1	0.30	0.58	4.2	1.08

. The chemical composition was determined by classical quantitative analysis.

The elemental composition of materials was determined with an X-ray fluorescence spectrometer with wave dispersion AXIOS 1 kW (Panalytical, Almelo, The Netherlands); the phase composition of samples was studied on X-ray diffractometer D8 Advance (Bruker AXS GmbH, Karsruhe, Germany).

The slag samples in the form of polished sections were studied by a JEOL JXA 8230 Electron Probe Microanalyzer (JEOL Ltd., Tokyo, Japan). Electron probe examination of the slag was carried out in the following modes: back-scattered electron image (COMPO), energy-dispersive spectroscopy (EDS), and element mapping by means of wavelength-dispersive spectroscopy (WDS). To conduct research using EDS-WDS microanalysis methods, a thin carbon coating film was applied to the polished surface of the samples.

Thermal analysis of slag samples was carried out using the synchronous thermal analysis device STA 449 F3 Jupiter, NETZSCH, Selb, Germany.

Smelting and cooling was carried out in alumina crucibles in the NTS 08/16 Nabertherm GmbH (Nabertherm GmbH, Lilienthal, Germany) furnace by neutral atmosphere pressure (in a nitrogen atmosphere at a pressure of 10⁵ Pa).

3. Results and Discussion

Several samples of converter slag were studied using X-ray phase analysis. The X-ray phase analysis (XRD) of one of the samples showed that the slag mainly consists of two phases—fayalite (2FeO·SiO₂) and magnetite (Figure 1a). Another sample, in addition to slag components, contained copper and lead inclusions in metallic form (Figure 1b).

Mineralogical analysis revealed that the converter slag sample is an alloy based on fayalite, magnetite, and glass. The copper phase is represented by chalcocite and metallic copper.

The metallic copper in the sample is present as finely dispersed inclusions (Figure 2). These inclusions form particles ranging in size from 0.005 to 0.035 mm in cross-section. The metallic copper occurs as free grains and also as particles within chalcocite. Free grains of metallic copper are sometimes observed with a chalcocite rim. The predominant grain size is 0.005–0.01 mm. In the microscope’s field of view at 400·magnification, up to 15 grains of metallic copper can be observed.

The EPMA of the same sample was conducted to identify the phases, revealing a complex composition of the sulfide globules. In the micrograph (Figure 3), a sulfide globule, approximately 0.17 mm in size, composed of chalcocite, is visible. The globule is embedded in a matrix consisting of a silicate slag phase, magnetite, and fayalite. Inside the globule, there are inclusions of metallic copper and lead.

WDS mapping of a section of the studied slag sample was performed (Figure 4). All sulfur present in the slag is associated exclusively with copper, as shown in Figure 4. The presence of iron is only noted in the copper sulfide particles, with a composition close to chalcocite, at a concentration of 1–2%. This indicates that most of the iron has been oxidized to magnetite and fayalite. Unlike the data provided by Das, B. et al. [21], for the slag composition we investigated, the majority of the copper was associated with mechanical inclusions.

To determine the crystallization temperatures of individual slag phases, thermal analysis of the slag sample was conducted. This was carried out using a simultaneous thermal analysis device. Before heating, the furnace chamber was evacuated (with an evacuation level of approximately 92%) and then purged with an inert gas for 5 min. Heating was performed at a rate of 10–15 °C/min in an atmosphere of high-purity argon. Cooling was carried out at a rate of 15 °C/min. The total flow rate of the incoming inert gas was maintained at around 100 mL/min. The results were processed using NETZSCH Proteus (version 5.1.0/24.11.2009) software.

Heating of the slag sample. On the DTA curve (Figure 5a), two intense exothermic effects are observed, with peaks at 728.7 °C and 935.8 °C. An intense endothermic effect was also recorded, with its maximum occurring at 1035.2 °C. Additional effects can be noted on the dDTA curve: a broad exothermic effect with a smooth peak at approximately 253.9 °C, an endothermic effect with an extremum at 575.7 °C, and endothermic effects with extrema at 1008.5 °C and 1052.8 °C. On the DTG curve, an intense minimum is present at 324.1 °C.

The exothermic peaks on the DTA curve represent the crystallization of glass phases. The endothermic effect with an extremum at 575.7 °C on the dDTA curve indicates the enantiotropic polymorphic transformation of quartz. The majority of the sample melts in the region of the endothermic effect with an extremum at 1035.2 °C on the DTA curve.

The process is divided into two stages. The first melting stage is reflected by the endothermic effect with an extremum at 1008.5 °C on the DTA curve. The endothermic effect with an extremum at 1052.8 °C on the dDTA curve represents the third and final stage of melting. Using the software, the melting onset temperature was determined to be 954.7 °C. By 1065 °C, the sample was completely melted.

Cooling of the slag. On the DTA curve obtained during the cooling of the sample (Figure 5b), three exothermic crystallization effects were observed with peaks at 1011.7 °C (dDTA), 954.2 °C, and 910 °C (DTA). Based on the thermal analysis of the slag sample, it can be concluded that the melting (during heating) and crystallization (during cooling) of the sulfide component occur within the temperature range of 900–1000 °C, which corroborates the conclusions made by Mamonov, S.V. et al. [10].

Thus, it can be concluded that slow slag cooling should be conducted down to a temperature of 900 °C. This will facilitate the process of squeezing the liquid (sulfide) phase from the periphery to the center of the crucible in accordance with the temperature gradient, which will contribute to the aggregation of sulfide globules.

To obtain a slag sample after slow cooling, the slag was melted in a laboratory furnace NTS 08/16 Nabertherm. A 200 g slag sample in an alumina crucible was placed in a cold furnace under a graphite lid, and the temperature was increased to 1250 °C at a rate of 10 °C/min. The total weight of the slag and crucible before the experiment was 306.77 g, and after the experiment, it was 302.774 g. The sample was held at this temperature for 30 min. Cooling of the crucible with slag in the furnace down to 900 °C took 2 h. After complete cooling, the crucible was cut, and slag samples were prepared for microprobe analysis. It is worth noting that during the dissection of the crucibles, it was found that part of the matte could segregate into a separate bottom phase due to liquation, some could concentrate in the middle of the slag column, and some could be extruded to the surface of the slag (see Figure 6).

In Figure 7, WDS mapping of the slag sample after slow cooling is shown, demonstrating that although slow cooling of the slag after heat treatment contributes to a reduction in the sulfide–metal suspension in the melt and its coarsening, a significant amount of copper remains in the slag in the form of particles with sizes of 0.01 mm or less.

Figure 8 shows the fine structure of a polymetallic globule (290–360 µm) after slow cooling of the slag. It is almost devoid of iron. The globule is surrounded by slag consisting of aluminosilicates, fayalite, and magnetite.

The globule itself has a heterogeneous structure: inside it, there is a lead droplet surrounded by a shell of an intermetallic compound of the Ni-Cu-As-Sb type, forming a “crown” (see Figure 9). The maximum thickness of the “crown” in the resulting cross-section reaches 15.8 µm. The complex behavior of the lead droplet, which has a diameter of 104 µm and is located inside the polymetallic globule, can be noted.

At the center of the lead droplet, there is a copper sulfide particle measuring 14–37 µm. Additionally, within the copper sulfide globule, several small zinc oxysulphide particles (ranging from 4 to 16 µm) are observed. Most intriguing, from the center of the lead droplet to the periphery of the globule, diverging chains of lead microdroplets (ranging from 1 to 27 µm) are seen throughout the cross-section of the globule. The distances between the microdroplets are very close to the average size of the droplets themselves. These chains are organized into eight distinct clusters. The number of microdroplets in a single cluster can range from 15 to 193. Notably, these chains of lead microdroplets do not extend beyond the boundaries of the globule. It is also important to note that the positioning of the “teeth” of the crown correlates with the beginning of each cluster. Interestingly, part of the crown material is carried away by the microdroplets (see Figure 10—Ni and Sb).

The origin of such a structure can be explained as follows: during the cooling of the slag, the slag-forming phases surrounding the copper sulfide globule solidify first. During this process, the Ni-Cu-As-Sb alloy forms a solid shell around the droplet of metallic lead. Copper sulfide inside the globule begins to crystallize next. As cooling continues and the volume of the slag-forming phases and copper sulfide decreases, the lead droplet undergoes compression. At a certain point, while the copper sulfide remains in a viscous state, the Ni-Cu-As-Sb shell ruptures, and the lead droplet, under pressure, partially penetrates into the globule in a radial pattern throughout its volume. Thus, the Ni-Cu-As-Sb intermetallic shell likely acts as a “shutter” in the formation of lead microdroplets. The speed of such structural changes is likely very high, comparable to an explosion. These are highly dynamic processes of microdroplet formation. The lead does not penetrate beyond the globule because the globule is completely surrounded on all sides by fully crystallized slag-forming phases of fayalite, magnetite, and aluminosilicates.

As a result of the conducted studies, it was found that during the slow cooling of slags, metal redistribution inside chalcocite globules may also occur, likely due to the gradual reduction in slag volume during the cooling process. Traces of high-dynamic processes observed in polymetallic globules indicate varying solidification rates of individual microstructure fragments. Different crystallization temperatures of the oxide and sulfide components of the slag contribute to the extrusion of the sulfide component from the oxide matrix of the slag during slow cooling. For the studied slags, slow cooling must be conducted in the temperature range of 900–1000 °C. The holding time in this range should be at least 2 h. To determine the temperature range for slow cooling of slags with a different composition, thermal studies must also be carried out. Thus, the coarsening of sulfide globules during slow cooling is associated with the difference in melting temperatures of the oxide and sulfide components of the slag. The coarsening of sulfide globules improves the conditions for subsequent flotation enrichment.

4. Conclusions

The conducted studies using mineralogical, X-ray phase, and electron probe microanalysis have shown that the initial converter slag and its thermally treated samples have identical matrices with almost complete matching of mineral and phase compositions. The distinguishing feature is the slight coarsening of sulfide components within the slag mass.

In the investigated converter slag from the Balkhash Copper Smelter, nearly all of the iron is oxidized and is present in the form of fayalite, magnetite, and magnetite, with other elements (silicon, copper, zinc, and aluminum) incorporated into its structure. The structure of all slag samples, as revealed by SEM with EPMA, indicates that sulfur is associated exclusively with copper.

Copper in the slags appears in both metallic and sulfide forms. Lead is found both in copper-bearing grains and within the slag melt, while zinc is present in both silicate and iron oxide melts. In the copper and lead compounds, the presence of arsenic, nickel, and antimony was detected sometimes in significant quantities.

Slow cooling of the converter slag after its thermal treatment contributes to a reduction in the sulfide–metal suspension within the melt due to its coarsening. However, a significant amount of copper remains in the slag in the form of particles 0.01 mm or smaller, which would likely prevent the complete removal of metals during flotation enrichment of the slags.

The different crystallization temperatures of the oxide and sulfide components of the slag contribute to the extrusion of the sulfide component from the oxide matrix of the slag during slow cooling. For the studied slags, slow cooling must be carried out in the temperature range of 900–1000 °C. The holding time in this range should be at least 2 h under crucible smelting conditions. To determine the temperature range for slow cooling of slags, thermal studies must also be conducted for each slag composition. Thus, the coarsening of sulfide globules during slow cooling is associated with the difference in melting temperatures of the oxide and sulfide components of the slag. The coarsening of sulfide globules improves the conditions for subsequent flotation enrichment.