Refractories for the Metallurgy of Copper

Abstract

Copper in the molten state is a rather reactive metal, and it interacts with the refractory lining of the furnaces, but the copper slags and copper matter are much more reactive. The processes in primary metallurgy of copper (purification of copper matte to blister copper) and secondary metallurgy of copper (re-melting of cathode copper and alloying) differ. The processes in primary metallurgy of copper have big differences, as a consequence, the types of furnaces in primary copper metallurgy have considerable peculiarities. However, in reverberatory, flash, electric furnaces and convertors the refractories are more or less similar. In re-melting furnaces, there are principally other silicon carbide-based refractories. A short description of the most common refractories, used in the furnaces and runners, corrosion behavior of refractories in contact with the molten slag and molten copper, corrosion resistance and corrosion tests, are made. There are some words about the service lifetime of the furnaces and repairs with a respect to corrosion testing and structure of refractories, and the possibilities of improving the service lifetime on refractories in the furnaces.

1. Introduction

The annual world production of copper is about 20 million tons, which is a considerable amount. However, this big amount is not that big, compared with the annual production of 64 million tons of aluminum and almost 2000 million tons of steel.

The specific consumption of refractories in copper metallurgy is high due to the very aggressive nature of copper melt and the extremely high aggressive nature of copper matte and slags. The literature on the corrosion of refractories in copper production is limited. There is no information on the values of the crucible static corrosion test and rod dynamic corrosion test of refractories to molten copper matte and molten copper slag. Probably the reason for the lack of this information is rather low total refractory consumption in copper metallurgy, that in 1998 was estimated by Schlesinger [1] as 25 thousand tons (copper production in 1996 was 12,772 million tons [2].

The specialists in steel production and refractories application in the ferrous industry use the term “the specific consumption of refractory per ton of melted metal” since 70–80th of the past century. At that time they calculated the specific consumption of refractories in kilograms per ton, but during the last ten years, they started to calculate the specific consumption in USD per ton [3].

The consumption of refractories plays a certain role in the cost calculations. Untimely shutdowns and unexpected repairs of the furnaces also have certain income in the increase of the cost production of copper.

2. Process stages and the refractories in the copper smelting process

Copper concentrate (produced from ore) is melted in furnaces (reverberatory, flash, electric) to produce matte, that transforms in the convertor to blister copper (Figure 1). In the following oxidation refining complex, the casting anodes for the purpose of electrolysis are produced, and electrolytic copper with a grade of up to 99.9% is melted in the shaft furnaces to produce copper alloys [4].

Figure 1.

Copper smelting process according to [5].

The most used lining refractory material in the metallurgy of copper is chromium-magnesite refractory. Magnesia, magnesia-spinel (MgO*Al₂O₃), and dolomite (MgO*CaO) refractories are also used in copper metallurgy, but generally speaking, the resistance of chromite-magnesia refractories is superior to all other said above materials.

The exception from this baseline of oxide refractories is silicon nitride bonded silicon carbide refractory lining in the shaft furnaces for re-melting of electrolytic copper.

The ratio of Cr₂O₃ to MgO in Cr₂O_3*MgO - chromite-magnesite refractory is a question of investigations, the specific conditions of the service, and to some extent “know-how” of refractory engineering companies.

The main advantage of adding chrome ore to magnesia raw materials is the increase of the resistance to basic slags. Usually, chrome ore contains alumina, and refractory producers do not purify chrome ore from alumina. So, in general, chrome-magnesia bricks contain 20–30% of Cr₂O₃, 25–55% MgO and up to 20% of Al₂O₃ (Table 1).

Composition, %	Producer
	A	B	C	D	E	F	G
MgO	59,2	58,9	61	60	55	56,6	24,2
Cr2O3	19,6	19,8	19,6	12	20	21	29,3
Al2O3	7,2	6,7	5,9			7,5	16,3
SiO2	1,9	0,9	0,8			1,3	2,8
CaO	1,2	0,9	1,2			1,1	2,4
Fe2O3	10,9	12,8	11,4			12,5	22,3
Properties
apparent density, g/sm^3	3,12	3,25	3,09	3	3,09	3,12	—
open porosity, %	17,5	15,5	18	16	18	17,5	14
cold crushing strength, Mpa	39	45	37	50	40	40–50	40–50
bending strength, Mpa	7	7	4			6–7	6–7
hot modulus of rapture, Mpa	6	6
linear change at 1700°C	<+0,1%	<+0,1%	<+0,1%			<+0,1%	—

Table 1.

Chemical composition and properties of chrome-magnesia bricks of various producers.

These refractories have a very high temperature of sintering, the firing temperature for such inert materials may be 1800°C or even more.

Some time ago the main type of chrome-magnesia refractories for the copper industry was so-called silicate bonded materials. Silica from chrome ore reacts with magnesia to form a silicate bond (that is deteriorated by fayalite Fe₂O₃-SiO₂ slag in reverberatory, flash, electric furnaces and converters).

The next step in the technology of refractories for copper furnaces is the so-called direct bonded chrome-magnesia refractories. For such types of refractories more high-purity materials are used to achieve the direct bond between the grains of periclase and chromite and grains of spinel without silicate interlayers between the grains.

Another type of chrome-magnesia material is so-called rebounded refractories. Magnesia and chrome ore are melted in an electric arc furnace to produce fused chrome-magnesia aggregate. The ground powders are shaped and fired to receive bricks with excellent corrosion resistance.

The electrolytic cathode copper sheets are melted in a continuous melting shaft furnace, and the alloys are prepared in a mixer – a small drum horizontal furnace. The approved variant for the refractory lining of the shaft furnaces [5] is silicon nitride bonded silicon carbide refractory or N-SiC. In shaft re-melting furnaces, there is almost no contact of the refractory material with slag, ad relatively limited contact with purified copper. Nitride-bonded silicon carbide refractory offer excellent thermal shock resistance and relatively high oxidation and weak resistance. The properties of the materials are in Table 2.

	Producer
Properties	A	B
SiC, %	72,0	78-75
Si3N4, %	27,8	22-25
α-Si3N4, %	12,0	α/β ratio is from 1:2 to 2:1
β-Si3N4,%	12,4
C free, %	0,15	-
Si free, %	0,21	<0,5
SiO2%	—	Oxides in total <1,2%
A12O3, %	0,23
CaO, %	0,25
Fe2O3 %	0,58 (0,4)
Porosity, %	16,6	14–18
Cold crushing strength, MPa	160	>150
Modulus of rapture, MPa	35	30–40

Table 2.

Chemical composition and properties of Si₃N₄-SiC bricks of various producers.

The conditions in each type of furnace in the copper smelting process differ, but the refractory materials for the furnaces are more or less the same. Barthel summarized the types of transformation of different parts of reverberatory and flash smelting furnaces [5]. The most corrosive agent in the reverberatory furnace is slag, and the most influenced part of the furnace is walls. The picture in the flash smelting furnace is a little bit more complex. The settlers of the furnace are corroded by matte and blister copper very intensively.

In flash smelters, the most corroded zones are the reaction shaft and the slag line. Lower MgO compositions are preferred because magnesia grains react with silicon oxide in the slag. The off-gas shaft and sidewalls are less subjected to slag attack.

In reverberatory furnaces, the situation is approximately the same but the contact of refractories with slag and matte is lower, the consequence is a slightly longer service life.

In fire-refining (anode) furnaces there are almost no silica-containing slags (if compared with flash furnaces and reverberatory furnaces), and refractory engineers often use magnesite refractories. Yet, magnesia-chromite refractories are still popular.

Yet probably the biggest challenge in corrosion resistance and the service lifetime for refractories is in converters, where the matter transforms to slag and blister copper. Slag corrodes the lining, but erosion also has income in the degradation of the lining and the tuyeres.

Magnesia-chromite refractories are used in converters, and the preference is high chromium compositions with high purity materials and low porosity. Fayalite (iron silicate) slag and copper oxide slag strongly deteriorate the refractory lining, especially the borders between grains in silicate bonded magnesia-chromite refractories.

3. Some words on testing of refractories

Refractories should be corrosion-resistant to melts and gases in furnaces and high-temperature devices in copper metallurgy. It is considered [6, 7] that, in general, approximately one-third of refractories are damaged because of poor thermal shock resistance at temperatures sufficiently lower than refractoriness, and two-thirds (2/3) of refractories go out of service due to lack of chemical resistance. It means, that 2/3 of shutdowns of the furnaces take place to the poor corrosion resistance of refractories.

Corrosion phenomena are very complex, and it is necessary to consider different processes and mechanisms of degradation and deterioration.

The corrosion includes the chemical reactions on the border of solid refractory and liquid melts, but it is also necessary to take into account the transport of reagents to the reaction zone and the drift of reactants from the reaction zone. In case of refractories, it is necessary to consider the penetration of aggressive liquids in open permeable pores (Figure 2), the wetting, the diffusion of the aggressive melts (or constituents of the aggressive melts) into the structure of refractory materials, and others.

Figure 2.

Porous refractory in contact with molten liquid (metal or slag).

In metallurgy, the corrosion behavior of refractories is estimated by corrosion tests. Laboratory corrosion tests may be divided into static cup tests and dynamic rod tests. Static and dynamic laboratory corrosion tests are carried out from 2 to 50 hours and provide information on the corrosion resistance of refractory material to liquid metal or slag.

There is no universal standard cup test for corrosion resistance, but metallurgists use the cup test (Figure 2), for comparative investigations [6] for almost a century. The test is simple and provides the information.

In order to make a cup test, a hole about 35–58 mm in diameter and about 40 mm deep is drilled in a brick (Figures 2 and 3), and either the corrosive liquid (metal or slag) is poured into the hole, but more frequently the corrosive agent in the solid state is placed in this hole before testing. The brick with the corrosive agent in a “cup” (hole) and with a lid above is placed in a furnace for 8–24 hours. The cup (hole) in a refractory castable may be made in advance without drilling, and for dry barrier mixtures, a cup with sloped sides is formed with a help of a special shape mold. After cooling, the sample brick with a cup and aggressive liquid is cut into two pieces, and the infiltrated area in the tested brick is measured (in square centimeters). This infiltrated area is a measure of corrosion resistance. The standards for cup corrosion resistance ASTM C768–99 [8] and DIN 51069–2-1972 [9] were withdrawn. However, the method of testing fireclay brick for corrosion resistance to cryolite is in standard ISO 20292 [10], and the standard DIN [11] guidelines for corrosion testing are valid.

Figure 3.

Static corrosion cup test. a – The cross-section of refractory brick with a hole for metal or slag before the test; b – The cross-section of a refractory brick after exposure to reactive metal or slag. The increased surface of the cross-section after exposure is a measure of the corrosion resistance of the refractory.

The dynamic rod test (Figure 4) is more complex from the point of apparatus [12–14], and is used not so often. The dynamic rod tests are produced in big crucibles or vessels, filled with at least one corrosive agent. The dimensions of the tested rods (usually rectangular bars) may vary from “25 mm*25 mm*100 mm” to “10 mm*10 mm*150 mm”. The bars of investigated materials may be rotated in aggressive liquid or be dipped in aggressive liquid and pulled out. There are variants of testing without rotation (Figure 5) [13].

Figure 4.

The results of the corrosion resistance cup test of refractory material to molten aluminum (Mg alloy). a – Conventional refractory fireclay shamotte brick; b – Special refractory with anti-wetting (for molten aluminum) additives (almost undamaged).

Figure 5.

The scheme (a) of dynamic rod test; the scheme of dynamic rod test for SiC refractory (b) in molten cryolite and aluminum [13]; the rods after exposure to molten cryolite and aluminum after 50 hours at 965°C [7].

The term “static corrosion test” means that the corrosive liquid in the cup is in a stationary condition, it does not move against or around the tested refractory, so the system “corrosive liquid–refractory” may reach equilibrium. During a “dynamic” test, there is a movement of the tested sample in the corrosive liquid, there are no stationary conditions between the refractory surface and the corrosive liquid, and the thermodynamic and chemical equilibria between the tested refractory and corrosive liquid will not be reached.

The corrosion resistance in dynamic testing is estimated in the decrease of the weight of the tested materials or the decrease of the volume of the tested rods. Usually, corrosion tests are isothermal. If there are possibilities, it is very good to investigate the structure of the corroded materials (the shape of the corroded grains, the depth of penetration of the liquids in the pores, etc.) with the help of a microscope.

At copper plants sometimes “industrial” testing of corrosion resistance is performed. Such kind of testing, although time-consuming, opens the way to the understanding of the processes, that take place at the service of materials. The investigation of corrosion of Si₃N₄-SiC refractory by the melt of copper and the copper slag was performed in the slag collector [15]. The tested refractory plate was the siphon block (the separator of slag) of the slag collector in the runner of the cathode shaft furnace. This plate (siphon block) is exposed to the permanent flow of copper with a small amount of slag. The slag on the surface of the melt is stopped by a refractory plate (separator). The refractory plate is exposed to the most extensive corrosion wear. However, this refractory plate is a good object for investigation of corrosion resistance, because different parts of this plate are exposed to intensive corrosion by flowing copper and flowing copper slag. General observations had shown that the corrosion of Si₃N₄-SiC refractory by slag is more extensive (the wear is about 3 mm per month), while the wear of Si₃N₄-SiC refractory by the flowing copper is below 1 mm per month.

Microscopic investigations of Si₃N₄-SiC refractory after 6 months of the service in said above conditions had shown 4–5 zones of a different color. There were no changes in the material in direct contact with molten copper without exposure to air. This show on the slow dissolution of silicon carbide and silicon nitride in the flowing melt of copper (physical dissolution in case of permanent removal of reactants). The most severe wear of Si₃N₄-SiC refractory was by copper slag.

Big refractory companies make investigations on corrosion resistance of refractories to copper and copper matte and perform cup testing [14], but publish results without details on the structure of refractories.

So, there is no information on more or less systematic investigations on corrosion resistance of refractories for copper metallurgy with a respect to the structure of refractories.

4. Some words on the structure of the refractory materials for metallurgy

The corrosion tests of refractories are made already for a century. The corrosion tests are performed to estimate if the specific refractories with being resistant to the attack of corrosive liquids (molten metal and slag) with a forecast for a long-term service lifetime of the furnace in a specific metallurgical process.

In real metallurgical processes, many factors affect corrosion [11]. However, in general, they may be divided into two – the chemical inertness of refractory and pore structure.

In Figure 2, it is shown that the liquid may penetrate tunnel permeable pores and open dead-end pores. The possibility of penetration of corrosive liquid in the open pores depends on the wetting of refractory and the pore size (Figure 6). Sometimes liquid cannot penetrate in the small pores but can easily penetrate in the pores with big diameters.

Figure 6.

The scheme of the pore structure of conventional refractory with closed pores, tunnel pores and open dead-end pores [7].

The value of open porosity (Tables 1 and 2) is essential in the estimation of the service of the refractory in corrosive media, but the pore size distribution (Figure 7) is much more important. The value of the average pore size may correlate with the gas permeability, which is also very important in the estimation of the correlation between the laboratory corrosion test, the chemical composition, and the structure of refractory material. The values of gas permeability and pore size distribution may say something on the forecast of the service life of refractory in metallurgical vessels (Figure 8).

Figure 7.

The variants of appearance of the liquid in the open pores of refractory depending on the wetting contact angle [7].

Figure 8.

The variants of the pore size distribution in alumina-silica refractory depend on the processing of refractory. The possibility of intrusion of corrosive liquid in the pores of material 1 (pore size 3,5 μm) is sufficiently lower than in materials 2 (pore size up to 10 μm) and 3 (pore size up to 15–20 μm) [7].

5. Conclusions

1. Copper in the molten state is a rather reactive metal, and it interacts with the refractory lining of the furnaces, but the copper slags and copper matter are much more reactive, which causes a limited-service lifetime of the furnaces in copper metallurgy.

2. The most used lining refractory material in the metallurgy of copper is chromium-magnesite refractory. Magnesia, magnesia-spinel (MgO*Al2O3) and dolomite (MgO*CaO) refractories are also used in copper metallurgy, but generally speaking, the resistance of chromite-magnesia refractories is superior to all other said above materials.

3. The corrosion resistance investigations of refractories for copper metallurgy, coupled with the investigations on pore structure and gas permeability, may give rise to an increase in the service lifetime of the furnaces, that finally may have certain income in the reduction of the cost.