Influence of Arsenic on the Chemical Wear of Magnesia-Chromite Refractories in Copper Smelting Furnaces

Abstract

Arsenic can be found in different sulfidic copper concentrates and residues, which for several years have commonly been used in copper metallurgy, as the “arsenic-free” resources are getting rare. Due to the high toxicity the removal of arsenic in the copper smelting process is a very important topic. The typical chemical slag and sulfur attack on the refractory material is getting enhanced by the presence of arsenic. This work deals with post-mortem investigations of a magnesia-chromite brick and castable used in copper smelting furnaces showing an additional and increased chemical attack by arsenic. The knowledge on the wear behaviour is not only based on a detailed chemical and mineralogical characterization, but also on thermochemical FactSage^TM calculations, which are carried out on provided post-mortem samples. Post-mortem investigations on used refractory materials represent an important prerequisite for the product development, as well as for special engineered lining concepts to support our customers.

Introduction

Due to the high demand of copper products the copper metallurgy has to deal with more complex copper containing ores. This means ores with a lower amount of Cu and a higher amount of minor elements. This paper will deal with the presence of arsenic (As) in the copper smelting process and its consequences to refractory degradation. Arsenic is problematic due to its toxicity. Generally arsenic can be found in mainly sulphidic copper/iron concentrates such as arsenopyrite (FeAsS), enargite (Cu₃AsS₄), lautite (CuAsS) an tennantite ((Cu,Fe)₁₂As₄S₁₃). The arsenic concentration in these compounds is varying between 19 and 50 wt%. Typically copper concentrates must not exceed a certain As-limit to be able to treat them without blending.

Those increasing amounts of minor elements, not only As, but also Sb, Sn and Ni will not only have metallurgical consequences but also impact the interaction with refractory materials used in copper smelting furnaces. In order to sustain the products quality (metal, slag, acid) the fractional distribution of these elements during the smelting process is a necessary prerequisite for a smelter to be able to deal with the increasing complexity of feed materials. Factors affecting the partitioning of minor elements in primary copper smelting from industrial plant data have been collected in an extensive study by Larouche [1]. Four major factors have been determined:

Process temperature
Oxygen enrichment in the blow
Final matte grade
Type of reactor.

For As, it was concluded that bath smelting reactors have higher evaporation rates than flash smelting reactors mainly due to lower O₂-enrichment and hence higher off-gas-volumes while the final matte grade does not seem to have a major effect on the partitioning to the gas phase. If As, Ni, Sn and Sb are associated with Cu, Pb and/or Fe a so called speiss phase can be formed in the smelting processes of copper and lead concentrates. This speiss phase has to be understood as a generic name to summarize intermetallic compounds of these minor elements from different processes in rather broad compositions. All of these have two characteristics in common: On the one hand these speiss phases can be used for minor element control in a smelter and on the other hand a subsequent process step is required to make use of the valuable metal content from these phases or to safely dispose of them.

Usually all furnaces used in copper production are lined with magnesia-chromite bricks or castables and alumina-chromia bricks, which have to withstand chemical, thermal and mechanical stresses [2].

Influence and Removal of Arsenic in Copper Smelting Furnace

Arsenic compounds are often associated with copper and their removal during copper smelting is not straight forward. Depending on operating conditions and the type of reactor the partitioning of As can be controlled in order to meet the requirements on product quality. Arsenic is unintentional in combination with metal as it lowers the product purity and quality and additionally influences the viscosity, the surface tension as well as the liquidus temperature of matte and copper metal in a negative way. Also the electrical conductivity decreases and the electrical resistivity increases due to arsenic impurity [3].

The most common solution today is blending, which means mixing of “clean” and “dirty” concentrates to ensure a certain As-level in the smelter feed. For high As-bearing copper concentrates both, hydro- and pyrometallurgical processes have been developed and are described in literature [4–12]. The most common used option for arsenic removal is the roasting process. Arsenic is transferred into gas phases, e.g. As-trioxide (As₂O₃) or As-trisulfide (As₂S₃) [4, 5] and collected in the off-gas. Which phases mainly will form depends on the oxygen and sulfur partial pressure [12].

For example roasting of high-As bearing concentrates: At 600 °C and defined oxidizing atmosphere decomposition of enargite (Cu₃AsS₄) under formation of Cu-sulfide, gaseous As₂S₃ and further on As₂O₃ takes place [10]:

$\begin{aligned} & 2{\text{Cu}}_{3} {\text{AsS}}_{{4({\text{s}})}} \to {\text{As}}_{2} {\text{S}}_{{3({\text{g}})}} + 3{\text{Cu}}_{2} {\text{S}}_{{({\text{s}})}} + {\text{S}}_{{2({\text{g}})}} \\ & {\text{As}}_{2} {\text{S}}_{{3({\text{g}})}} + 4.5{\text{O}}_{{2({\text{g}})}} \to {\text{As}}_{2} {\text{O}}_{{3({\text{g}})}} + 3{\text{SO}}_{{2({\text{g}})}} \\ \end{aligned}$

As-trioxide is extremely volatile and sublimes at 465 °C [13]. This means that the off-gas has to be cooled down to sufficiently low temperatures to ensure that most of the arsenic is condensed and can be collected. In an oxygen atmosphere it is likely that As-trioxide can oxidize to higher oxides like As₂O₅, which is less volatile and form stable non-volatile arsenates with other metallic oxides [11]. Nevertheless, minor amounts of arsenic will always remain in the slag or matte. This is highly depending on the furnace type and process conditions (Table 1).

Table 1

Distribution of arsenic in different product phases of industrial smelting processes for recovery of copper [12] (modified)

Process	Feed material	As distribution in products (in wt%)
Process	Feed material	Flue dust	Slag	Matte
Flash smelting	Cu-concentrate	75–85	2–20	rest
Electric furnace smelting	Cu-concentrate	10–25	50–70	rest
Reverbatory smelting	Cu-concentrate (As in feed > 0.2%)	50–70	10–20	rest
Reverbatory smelting	Cu-concentrate (As in feed < 0.2%)	5–35	15–55	rest
Pierce smith converter	Cu-matte from reverbatory furnace	75–90	5–25	rest

Post-mortem Investigation

For a better understanding of the wear parameters of used refractory materials in the non-ferrous metal industry post mortem studies are performed at RHI Magnesita’s Technology Center Leoben, Austria. The detailed information gained out of such studies are necessary for development of new, for the process adapted, refractory products, to improve the furnace process.

Selected examples for the post-mortem studies are discussed in the following. The studies were carried out on one magnesia-chromite brick and one castable sample out of a copper smelting furnace. In general, arsenic attack on those refractory products is not the main wear mechanism, but represents an important additional factor.

Magnesia-Chromite Brick

The received brick out of the bath area of a copper smelting furnace has a residual thickness between 160 and 170 mm after eight months of operation. The original brick thickness is 450 mm. The immediate brick hot face is rough and covered with a thin slag coating. On the cross section a thick reaction zone and several cracks running parallel to the bricks hot face can be recognized (Fig. 1a). The lower part of the refractory (approx. 40 mm from the cold end) broke off while removal from the furnace. On the fracture surface a yellowish-reddish coating is visible, which is indicative of arsenic bearing phases (Fig. 1b).

../images/468727_1_En_20_Chapter/468727_1_En_20_Fig1_HTML.gif — Fig. 1
Cross sectional view. Magnesia-chromite brick out of a copper smelting furnace. a The immediate brick hot face is covered with a thin slag coating (S). In addition a thick reaction zone up to 10 mm is visible (R). Cracks running parallel to the hot face are observed (arrows). The lower part of the brick (approx. 40 mm from cold end) was completely broken. b On the fracture surface a yellowish-reddish coating is visible. Sampling for polished sections (rectangle)

Chemical Analysis

Chemical analyses of the deeply infiltrated magnesia-chromite brick were carried out by X-ray fluorescence analysis (Bruker S8 Tiger). At the brick hot face within the area 0–15 mm an extremely high SiO₂- and Fe-oxide content was determined. Additionally, slightly higher amounts of CaO and CuO were detected. The middle part of the refractory (85–100 mm) is enriched with SO₃, whereas at the cold end of the brick (140–155 mm) a high content of sulfur and arsenic (up to 1.4% As₂O₃ by wt.) was determined. The arsenic content at the hot face and the middle part of the refractory is about 0.3 wt% (Table 2).

Table 2

Chemical analyses (semi-quantitatively, without calibrating standard) of the magnesia-chromite brick out of a copper smelting furnace (wt%)

Sample	MgO	Al₂O₃	SiO₂	SO₃^a	CaO	TiO₂	Cr₂O₃	Fe₂O₃^b	CuO	As₂O₃^c
Hot face (0–15 mm)	41	6	9	0,5	2	0.3	12	28	1	0.3
Middle part (85–100 mm)	59	5	1	2	1	0.2	19	12	–	0.3
Cold end (140–155 mm)	64	6	1	1	1	0.1	16	11	–	1.4

^aSulfur calculated as SO₃

^bTotal iron calculated as Fe₂O₃

^cTotal arsenic calculated as As₂O₃

Mineralogical Investigation

For the mineralogical investigation polished sections from the hot face and the fracture surface of the brick were prepared (Fig. 1). The mineralogical investigation is carried out with reflected light microscope (NIKON Eclipse LV150 N) and scanning electron microscope (JEOL JSM-6460).

At the immediate brick hot face corrosion of the magnesia component under formation of forsterite took place. Additionally, a phosphor- and sulfur containing Mg-V-Ca-As-oxide was observed (Fig. 2a). At the top of the fracture surface near the cold side a 3 mm thick layer consisting of As-oxide and As-sulfide was determined. Below this layer a thin metallic arsenic layer (thickness of 0.03 mm) was formed. The brick microstructure below the As-containing layer is infiltrated with As-sulfide (up to 1 mm). Due to corrosion of the magnesia components by SiO₂- and As₂O₃-supply As-Mg-sulfate, Ca–Mg-silicate of type monticellite (CaMgSiO₄) and minor merwinite (Ca₃Mg(SiO₄)₂) were formed as main reaction products (Fig. 2b, c).

../images/468727_1_En_20_Chapter/468727_1_En_20_Fig2_HTML.gif — Fig. 2
Photomicrographs of the hot face and the fracture surface near the cold end. a Hot face. Magnesia component (1). Chromite (2). Monticellite (3). Phosphor- and sulfur containing Mg-V-Ca-As-oxide (4). b Fracture surface near cold end. Infiltrated and corroded brick microstructure. As-sulfide and As-oxide containing layer (1). As_met (2). Mix analysis containing sulfur-rich Mg–Al-Cr–Fe- and As-oxide (3). Infiltration of As-sulfide (4) forming As-Mg-sulfate (5). Monticellite (6). c 3 mm below fracture surface. Magnesia component (1). As-Mg-sulfate (2). Monticellite (3)

Magnesia-Chromite Castable

The magnesia-chromite castable out of the upper area of a copper smelting furnace (6 weeks of operation) is completely degenerated (from hot face to the cold end). The residual thickness is approx. 170 mm. A crumbly surface is visible on the hot face. The sample is completely infiltrated (Fig. 3).

../images/468727_1_En_20_Chapter/468727_1_En_20_Fig3_HTML.gif — Fig. 3
Cross sectional view. Complete infiltration of the sample. A crumbly surface is visible on the hot face. Samples for chemical analysis (A) (B). Sampling for polished Sect. (1)

Chemical Analysis

Chemical analysis was carried out on the hot and additionally on the cold end of the sample. The hot face is highly enriched with PbO, arsenic, CuO, SiO₂, Sb₂O₃ and ZnO. The cold end is also enriched with lead, arsenic and copper. The arsenic content is about 5 wt% (Table 3).

Table 3

Chemical analyses (semi-quantitatively, without calibrating standard) of the magnesia-chromite castable out of a copper smelting furnace (wt%)

Sample	MgO	Al₂O₃	SiO₂	SO₃^a	CaO	Cr₂O₃	Fe₂O₃^b	ZnO	CuO	PbO	As₂O₃^c	Sb₂O₃
Hot face (0–15 mm)	24	3	5	0.7	0.3	12	9.5	1.0	7.5	28.5	5.0	1.3
Cold end (160–170 mm)	22	3	3	0.2	1	14	10	–	2	39	4.6	0.4

^aSulfur calculated as SO₃

^bTotal iron calculated as Fe₂O₃

^cTotal arsenic calculated as As₂O₃

Mineralogical Investigation

Microscopically the castable is completely degenerated. A thick reaction layer is visible at the hot side. Mainly corrosion of magnesia and chromite can be detected (Fig. 4a). The wear can be described by two mechanisms: Corrosion of mainly the magnesia component due to SiO₂ supply under formation of forsterite (Mg₂SiO₄) took place. The chromite component got strongly corroded and is enriched with Cu-Zn-Ni-Fe-Sb- and Sn-oxide at the rims. Additionally the castable became infiltrated up to the cold end by Pb-arsenate of type Pb₄As₂O₉ and Pb₈As₂O₁₃ (Fig. 4b) and corrosion of the magnesia component took place. Cu- oxide and Cu_met can also be detected up to the cold end.

../images/468727_1_En_20_Chapter/468727_1_En_20_Fig4_HTML.gif — Fig. 4
a Immediate hot face. Reaction layer—Corroded magnesia component (1). Recrystallized chromite enriched with Cu-Zn-Ni-Fe-Sb- and Sn -oxide at the rims (2) b Corroded magnesia (1) and chromite (2) component. Chromite highly enriched with Cu-Zn-Ni-Fe-Sb- and Sn- oxide (3). Formation of a forsterite rich layer (4). Additionally forsterite crystals can be found in the infiltrate. Cu-oxide (5). Pb₄As₂O₉ and Pb₈As₂O₁₃ (6)

FactSage^TM Calculations

Arsenic corrosion mechanism can be presumed as comparable to the well-known sulfur corrosion [14]. In an oxidizing atmosphere (for example while matte blowing in a converter) As-sulfide will oxidize forming As-oxide and sulfur-oxide. For instance in contact with the refractory material As-trioxide As₂O₃ will react with the most basic components in the brick, in this case MgO and CaO, under formation of Mg/Ca-arsenate.

$\begin{array}{*{20}l} {2{\text{As}}_{2} {\text{S}}_{3} + 9{\text{O}}_{2} \to 2\,{\text{As}}_{2} {\text{O}}_{3} + 6{\text{SO}}_{2} \left( {800\,^\circ {\text{C}}} \right)} \hfill & {{\text{SO}}_{2} + 1/2{\text{O}}_{2} \leftrightarrow {\text{SO}}_{3} \left( { > \,760\,^\circ {\text{C}}} \right)} \hfill \\ {3{\text{MgO}} + {\text{As}}_{2} {\text{O}}_{3} + {\text{O}}_{2} \to {\text{Mg}}_{3} \left( {{\text{AsO}}_{4} } \right)_{2} \left( {800\,^\circ {\text{C}}} \right)} \hfill & {{\text{SO}}_{3} + {\text{MgO}} \to {\text{MgSO}}_{4} \left( { < \,1050\,^\circ {\text{C}}} \right)} \hfill \\ {3{\text{CaO}} + {\text{As}}_{2} {\text{O}}_{ 3} + {\text{O}}_{2} \to {\text{Ca}}_{3} \left( {{\text{AsO}}_{4} } \right)_{2} \left( {800\,^\circ {\text{C}}} \right)} \hfill & {{\text{SO}}_{3} + {\text{CaO}} \to {\text{CaSO}}_{4} } \hfill \\ \end{array}$

As described in the second post-mortem study As-oxide can react as As-pentoxide As₂O₅ with volatilized lead (PbO) in the off-gas. These phases become liquid >640 °C depending on the As₂O_5/PbO-ratio (Fig. 5).

../images/468727_1_En_20_Chapter/468727_1_En_20_Fig5_HTML.gif — Fig. 5
PbO-As₂O₅ phase diagram [15]

With help of the scanning electron microscope the PbO-As₂O₅ phases Pb₈As₂O₁₃ and Pb₄As₂O₉ have been detected. Following reactions can be assumed:

$\begin{aligned} & 8{\text{PbO}} + {\text{As}}_{2} {\text{O}}_{5} \to {\text{Pb}}_{8} {\text{As}}_{2} {\text{O}}_{13} \\ & 4{\text{PbO}} + {\text{As}}_{2} {\text{O}}_{5} \to {\text{Pb}}_{4} {\text{As}}_{2} {\text{O}}_{9} \\ \end{aligned}$

In addition to the mineralogical investigation FactSage^TM calculations with data bases FToxide and FTmisc were carried out to study the theoretical corrosion of the brick components as well as phase formations. The calculation results were compared with the microscopic investigation results. As-phases react with the MgO component, forming Mg₃(AsO₄)₂ and hence represent an additional corrosion factor for the refractory bricks. Figure 6 shows several magnesia-chromite bricks with different chemical composition reacting with an As-containing slag. Brick F forms the lowest amount of solid phase. Due to database limitations, an additional reaction with sulfur and formation of a mixed Mg-As-sulfate as mineralogically detected in the post mortem sample cannot be seen in the calculations.

../images/468727_1_En_20_Chapter/468727_1_En_20_Fig6_HTML.gif — Fig. 6
Formation of As-containing solid phases (at a temperature of 1200 °C and pO₂ = 1*10⁻⁹) in different MgCr-refractories (bricks A–F), i.e. newly formed reaction product of As-containing process phases with brick phases

Conclusion

As the arsenic content in copper ores has become an increasingly important factor over the last years it is essential to know the influences of arsenic in the copper smelting process and its contribution to refractory wear. Especially post-mortem studies are important for industrial application and specific customer requirements.

The investigated magnesia-chromite brick out of a copper smelting furnace shows mainly continuous wear by hot erosion and is severely attacked by Fe-silicate slag (fayalite type Fe₂SiO₄). Enrichment of arsenic caused severe corrosion of mainly the magnesia component forming Mg-As-sulphate as main reaction product. Arsenic attack on the refractory material is not the main chemical attack, but represents an important additional factor influencing lining wear and lifetime.

In the second post-mortem study on the magnesia-chromite castable arsenic reacted with lead under formation of Pb-arsenate of type Pb₈As₂O₁₃ and Pb₄As₂O₉. Two corrosion mechanisms were observed: Chemical attack by SiO₂ on the magnesia under formation of forsterite. Chromite was also highly attacked by Sb-Fe- and Sn-oxide. Additionally the castable got deeply infiltrated and corroded by Pb-arsenate.

The observed attack on the magnesia-chromite brick and castable by As compounds in the zone of liquid metal and slag (for the brick) and off-gas area (for the castable) may have been caused by insufficient As volatilization due to not optimized process conditions.

FactSage^TM calculations, based on available data, show that As will mainly attack the basic component of the refractory under formation of e.g. magnesium-arsenate Mg₃(AsO₄)₂. Arsenic can also react with other components in the off-gas e.g. lead, forming lead-arsenic oxide phases, which can infiltrate and corrode the microstructure.

Post-mortem investigations on used refractory materials represent an important prerequisite for the product development, as well as for special engineered lining concepts to support our customers.

References

1.
Larouche P (2001) Minor elements in copper smelting and electrorefining
2.
Routschka G (2011) Refractory materials, 4th edn. Vulkan-Verlag, Essen, p 505
3.
http://www.conductivity-app.org/single-article/cu-overview, European Copper Institute, Accessed 26 Feb 2018
4.
Hoffmann JE (1993) Remediating copper smelter dusts: the arsenic problem. JOM:30–31
5.
Piret NL (1999) The removal and safe disposal of arsenic in copper processing. JOM:16–17
6.
Imris I, Klenovcanova A (2003) Removal of arsenic, antimony and mercury from copper concentrates by roasting process. In: Proceedings of copper 2003, Santiago, Vol IV, pp 125–139
7.
Dalewki F (1999) Removing arsenic from copper smelter gases. JOM:24–26
8.
Lindkvist G, Holmström A (1983) Roasting of complex concentrates with high arsenic content. Adv Sulfide Smelt:451–472
9.
Alvear GR et al (2008) Copper isasmelt—dealing with impurities. In: Proceedings of sohn international symposium advanced processing of metals and materials, Vol. 8, International symposium of sulfide smelting, TMS, 2006, pp 673–685
10.
Wilkomirsky I, Parra R, Parada F, Balladares E (2013) Physico-chemistry and kinetics mechanisms of partial roasting of high-arsenic copper concentrates. In: Proceeding of copper 2013, Chile, pp 539–552
11.
Weisenberg IJ, Bakshi PS (1979) Arsenic distribution and control in copper smelters. J Metals:38–44
12.
Grund SC, Hanusch K, Wolf HU (2008) Ullmann’s encyclopedia of industrial chemistry. Arsen Arsen Compd
13.
Holleman AF, Wiberg N (2007) Lehrbuch der anorganischen chemie, Auflage. Walter de Gruyter & Co, Berlin, p 102
14.
Gregurek D, Reinharter K, Majcenovic C, Wenzl C, Spanring A (2015) Overview of wear phenomena in the lead processing furnaces. J Eur Ceram Soc 35(6):1683–1698Crossref
15.
Firoozi S (2005) Thermodynamics and mechanisms of lead softening: 16–24

Influence of Arsenic on the Chemical Wear of Magnesia-Chromite Refractories in Copper Smelting Furnaces

Abstract

Keywords

Introduction

Influence and Removal of Arsenic in Copper Smelting Furnace

Post-mortem Investigation

Magnesia-Chromite Brick

Chemical Analysis

Mineralogical Investigation

Magnesia-Chromite Castable

Chemical Analysis

Mineralogical Investigation

FactSageTM Calculations

Conclusion

FactSage^TM Calculations