Overview of wear phenomena in lead processing furnaces

Available online at www.sciencedirect.com

ScienceDirect Journal of the European Ceramic Society 35 (2015) 1683–1698

Review

Overview of wear phenomena in lead processing furnaces D. Gregurek a,∗ , K. Reinharter a , C. Majcenovic a , C. Wenzl b , A. Spanring b a

RHI AG, TC Leoben, Magnesitstrasse 2, Leoben A-8700, Austria b RHI AG, Wienerbergstrasse 9, Vienna A-1100, Austria

Received 24 September 2014; received in revised form 9 December 2014; accepted 10 December 2014 Available online 16 January 2015

Abstract The main wear parameters influencing the refractory lining life in vessels used in the Pb/Zn industry (e.g., QSL reactor, KIVCET furnace, Ausmelt/IsasmelterTM , Kaldo furnace, short rotary furnace) can be subdivided into chemical, thermal and mechanical stresses. In the present work the main wear parameters, such as corrosion by slag attack, high sulfur, soda and iron oxide supply as well as reduction, non-oxidic infiltration and brick damage by hydration, are briefly introduced and discussed. Additionally, the extraordinarily high SiO2 supply caused by the uncontrolled addition of silica sand results in a massive forsterite formation and in a volume expansion (“forsterite bursting”). Increased operation temperatures in the furnace support also microstructural brick degeneration. All these mentioned wear phenomena lead to a severe degradation of the brick microstructure and consequently to a decreased lining life. Therefore, a detailed understanding of the wear mechanisms through “post mortem studies” is an important prerequisite for the refractory producer. © 2014 Elsevier Ltd. All rights reserved. Keywords: Refractories; Lead metallurgy; Wear phenomena

Contents 1. 2. 3. 4.

5.

∗

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of lead production routes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analytical procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Macroscopical overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Wear phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1. Slag attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2. Forsterite bursting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.3. Sulfur corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.4. Iron oxide attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.5. Non-oxidic infiltration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.6. Microstructural changes due to high temperature load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.7. Reduction by varying or low oxygen pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.8. High soda supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.9. Brick damage by hydration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Corresponding author. Tel.: +43 69918705156. E-mail address: [email protected] (D. Gregurek).

http://dx.doi.org/10.1016/j.jeurceramsoc.2014.12.011 0955-2219/© 2014 Elsevier Ltd. All rights reserved.

1684 1684 1686 1687 1687 1687 1687 1689 1690 1693 1693 1695 1695 1695 1696 1697 1697 1697

1684

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

1. Introduction The profitable primary or secondary pyrometallurgical operation depends on many factors such as furnace type, process conditions, lining design, selection of refractory types, etc. In the non-ferrous metals industry, particularly in lead smelting furnaces, magnesia-chromite bricks are the preferred refractory choice due to their high corrosion resistance. Nevertheless, the refractory lining is exposed to complex and mutual wear caused by chemical, thermal, and mechanical stresses.1–4 Therefore the exact understanding of the wear phenomena through post mortem studies is an important prerequisite for the refractory producer, since it provides the basis for both customer recommendations and innovative product development. In addition to post mortem studies, laboratory work and experimental testing carried out in the pilot plant of RHI’s Technology Center Leoben enable the best possible understanding of the brick wear on the pilot scale.5,6 Combining all these facts, the adequate choice of refractory is always essential for a successful furnace campaign. A general overview of wear phenomena in the non-ferrous industry was discussed and introduced in several papers in the past.7–13 Particularly in the lead industry a lot of work was done regarding refractory corrosion testing in different pilotscale and industrial furnaces. For instance, Oprea7 discussed failure mechanisms observed on the magnesia-chromite bricks lined above the slag line of the flash furnace for zinc-lead smelting. In order to explain these findings some laboratory work was done additionally. Prestes et al.8 analyzed wear phenomena on magnesia-chromite bricks from the lead short rotary furnace. Similar to Oprea, in addition to post mortem studies also some experimental work by crucible corrosion testing was carried out. Monshi et al.9 investigated wear of the refractory lining out of the top blowing rotary converter (TBRC), whereas Hoed10 reported on refractory erosion and evidence based on the pilotscale trials in the DC-arc furnace carried out with refractories and lead blast furnace slag. Wei11 reviewed available literature concerning corrosion of refractories in the lead-smelting reactors such as KIVCET furnace and TBRC and evaluated in the laboratory the corrosion behavior of various refractory materials against industrial slags. Finally, Scheunis et al.12 investigated the effect of phase formation during use on chemical corrosion of magnesia-chromite refractories in contact with non-ferrous PbO–SiO2 based slag. Malfliet et al.13 carried out a critical review work on degradation mechanism and use of the refractory linings in the copper production processes. These findings for copper slags are also interesting and relevant for a better understanding of the refractory wear processes in lead metallurgy, as similar slag systems and refractory qualities are used. The special features of the different technologies (i.e., bath/flash smelting, stationary/moveable furnaces, different degree of turbulence, batch/continuous process, etc) and slag systems (FeOx–SiO2 –CaO/soda) also generate different challenges for the refractory furnace lining. This paper gives an overview of the main wear mechanisms affecting the refractory lining from the primary and secondary lead processing furnaces, such as KIVCET furnace, Kaldo converter, QSL reactor, Ausmelt/IsasmelterTM (TSL reactor),

reverberatory furnace and short rotary furnace (SRF). The knowledge about wear mechanisms is based not only on many years of experience through post mortem studies, but also on additional laboratory work and experimental testing in the pilot plant at RHI’s Technology Center Leoben. For a better understanding of the wear phenomena a brief description of the main metallurgical processing routes is given in the following section. 2. Overview of lead production routes Depending on the nature of the input materials various technologies are available for lead production, generally primary and secondary production route can be distinguished. The primary route uses mainly sulfidic lead concentrates, also with the addition of zinc plant residues or battery scrap, whereas the secondary route processes only materials from secondary sources, especially batteries. Nowadays, secondary production volumes already exceed primary lead production.14–16 The traditional primary route is roasting-reduction/smelting (roast-reduction, i.e., sinter plant & blast furnace), however, over the last decades, the direct smelting reduction processes (roast-reaction) have become more important and nowadays are state-of-the-art.15,16 The process paths can also be seen from Fig. 1. Some of the vessels from primary industry are also used in secondary industry (e.g., SRF, TSL).14,18 Generally, the process parameters and technology are chosen according to the input material, i.e. present impurities and required metallurgical work. The conditions range from oxidizing for sulfur removal (roasting) to reducing for smelting and reduction, including slag fuming—sometimes both in one vessel (Fig. 2). The process temperature is generally far higher than the lead liquidus temperature (327 ◦ C), namely around 1000 ◦ C (and even higher for lower PbO levels in the slag), in order to have a liquid and reactive slag that is easy to remove. Additionally, slag chemistry is adjusted in a way to minimize metal overheating. Consequently, the following challenges arise for the refractory18–20 : - Slag chemistry: adjusting slag composition within the system FeOx–SiO2 –CaO and/or choice of other slag systems and additives (soda slag) for achieving low liquidus temperature under consideration of other slag properties (e.g., viscosity, lead solubility, lead fuming) causes chemical attack. - Furnace atmosphere and temperature: especially changing atmosphere (oxidizing/reductive) as well as hot gases lead to increased refractory damage. - Overheated liquid phases (metal and slag) with resulting very low viscosity cause deep refractory infiltration and chemical attack. The following overview gives a short introduction into the lead production processes where the slags investigated in this paper originated from. Other production routes like EAF, Outokumpu Flash, SKS (Shuikoushan), KLS (Kosaka Lead Smelting) and the traditional blast furnace route will not be described in the present work14–16,18,20–24 : A general overview

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

1685

Fig. 1. Phase diagram for Pb–S–O showing smelting paths.17

of the refractory wear mechanism in the discussed furnaces is given in Table 1. - KIVCET is an acronym formed by the Russian words for “oxygen-flash cyclone-electrothermic process”—the process is based on flash smelting of primary and/or secondary materials in the reaction shaft. The molten material flows through a coke layer floating at the bottom of the reaction shaft, causing reduction of species like PbO. A partition wall divides

the SO2 -rich gas atmosphere of the reaction shaft from the second furnace part, namely the electric furnace, where the slag and lead bullion are finally tapped. The process temperature is higher than in the other lead production processes, namely above 1300 ◦ C, the lead being tapped at about 900 ◦ C. Slag chemistry and temperature distribution are essential to avoid excessive accretions in colder furnace parts. - The Kaldo converter – also known as a top-blown rotary converter (TBRC) by other furnace vendors – is a flexible

Fig. 2. Sulfur–Oxygen Potential Diagram for the Pb–S–O-System at 1200 ◦ C with log pO2 on the vertical axis.17

1686

+

± + +

+

+ +

+ +

+ +

+ + + + + + + + ±

+ +

+ + + + + + + + +

+

-

+ + +

+

+ +

+

+ + ±

±

+

+

+ + + ±

+ + +

+

+

± +

-

± + +

-

production method, used both in primary and secondary lead production. The furnace atmosphere can be adjusted as required, furthermore the furnace construction and rotation allow intense mixing and therefore good reaction kinetics. However the rotation causes additional challenges for the refractory lining, both regarding erosion and mechanical stability. The QSL reactor is a horizontal cylindrical vessel that is divided into oxidation and reduction zones for concentrate smelting and slag reduction, respectively. Oxygen and coal are blown into the melt thorough tuyeres in the respective areas. The converter can be turned/rolled for tuyere replacement. Typically, the furnace parts are divided by a partition wall to separate the SO2 -containing gases in the oxidation zone from the reduction gases that also contain the zinc fume. The critical refractory areas are the tuyeres (coal and oxygen) and the partition wall, both requiring optimized design and material selection. The top submerged lance (TSL) reactor is a vertical, refractory-lined vessel with central lance introduced through the top, used in both primary and secondary lead production. The lance injects the fuel and (oxygen-enriched) air into the molten slag layer, causing submerged combustion and high bath turbulence. Oxidizing or reducing atmosphere can be adjusted. The main TSL types are AusmeltTM and IsasmelterTM reactors—one of the main differences, also in refractory attack, are their different splash patterns due to different air swirler technologies. Reverberatory furnaces were used for primary production in former times, but today are found mainly in secondary lead processing in a combination with blast- and electric furnace. The process is continuous: Lead-containing scrap is combined with coke or coal and treated under oxidizing conditions to generate heat from the carbonaceous input, as well as transfer impurities (e.g., Ca, Al, Sb, As, Sn) into the slag and produce a lead bullion with low impurity content. The short rotary furnace (SRF) is a very flexible vessel, used both in primary and secondary metallurgy, that can process a wide variety of input materials. Many operations use a soda slag, some also in combination with a FeOx–SiO2 slag in separate process stages. The soda slag and especially changing slag systems require optimized refractory linings.

+

+ +

+

+

+

+

+

+ + + + + + + +

+

+ + + + +1

+

+

+ + + + +1

+

+ + + +

-

Including overheating. Non-oxidic Infiltration. 1 Depends on Kaldo furnace type. + Standard phenomenon. ± Possible phenomenon.

*

**

M. fatigue Bath Charging Mechanical

+ + + + T. Level* T. Shock Lead** Thermal

+

Slag + Sulfur + FeO RedOx High soda Hydration

+

3. Analytical procedure

Chemical

Endwall Oxidation Zone Reduction Zone Partition Wall Shaft Settler EAF Uptake Upper cone Cylinder Bottom Wall Upper part Hearth Roof Wall Barrel

KIVCET furnace

Table 1 General overview of wear mechanisms in lead processing furnaces.

Kaldo furnace

TSL

Reverberatory furnace

SRF

QSL reactor

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

Generally, every single post mortem study starts with the visual inspection carried out on the brick cross section followed by selection of samples for chemical analysis and mineralogical investigation. The chemical analysis is carried out by using X-ray fluorescence analysis (Bruker S8 TIGER). The mineralogical investigation is performed on polished sections using a reflected light microscope, X-ray diffraction (Bruker D8 ADVANCE), and a scanning electron microscope (SEM) (JEOL JSM-6460) combined with an energy-dispersive and wave-dispersive X-ray analyser.

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

1687

The observed wear phenomena can generally be subdivided into chemical, thermal, and mechanical stresses.3,4 These can appear as a single stress factor, but as a rule combinations of these stresses occur and affect the refractory lining. The most common chemical wear phenomena affecting the performance of magnesia-chromite bricks in the lead processing furnaces discussed in this paper are corrosion by acidic slags and sulfur, iron oxide attack, corrosion by high soda supply, reduction phenomena and brick damage by hydration. Increased temperature as well as non-oxidic infiltration into the brick microstructure are further thermal wear factors.

coloring was caused by salts formed on the brick surface (hexavalent chromium formation due to high alkali supply, Fig. 3e). Sample F (see Fig. 3f) shows crack formation at the hot face due to a phenomena called forsterite bursting. Also in the reverberatory furnace, input and slag chemistry are the main factors, while turbulence or temperature normally do not play an excessive role in this furnaces type: the input is always a challenge in recycling processes, as there generally is a wide variation, which in turns requires corresponding adjustment of the slag chemistry. The latter also includes the use of soda slags which subsequently can cause alkali attack on the refractories. A precise description of the above mentioned wear mechanisms are given in following section.

4.1. Macroscopical overview

4.2. Wear phenomena

The macroscopical overview of the representative magnesiachromite bricks from different lead processing furnaces are shown in Fig. 3. Samples A and B represent magnesia-chromite bricks from the lead SRF. For both samples the immediate brick hot face was covered with thin slag coating (Fig. 3a) respectively reaction zone (Fig. 3b). In both bricks numerous short running cracks formed parallel to the hot face in completely infiltrated and corroded brick microstructure. Sample A was heavily attacked by slag, in the case of sample B it was a special case of high iron oxide attack. In case of sample B the cold face is missing (see Fig. 3b). This means the brick hot face spalled off due to the high thermal shock. As slag chemistry is the main challenge in SRF (i.e., no excessive turbulence or temperature in case of regular furnace operations), slag attack is the main observed wear factor in this case. The spalling caused by thermal shock is supposed to be due to rapid furnace stoppage and cool down (also this is typical, as many SRFs are operated in campaigns, i.e., in operation for several days followed by stop for some days/weekend). Sample C (Fig. 3c) represents a magnesia-chromite brick from the Kaldo converter. Below the slag coating, cracks parallel to the hot face formed in highly infiltrated and corroded brick microstructure. Also in the Kaldo the slag chemistry is the main wear factor: in this furnace type in combination with the rotating movement, slag attack occurs over the whole furnace circumference. Cracks parallel to hot face indicate thermal shock/thermal cycling, which can also be explained by the furnace operation: as it is a batch process, certain temperature cycles occur (e.g. during charging). Sample D from the KIVCET furnace shows deep-reaching infiltration of the brick microstructure with sulfur i.e. sulfur attack (see Fig. 3d). The sulfur comes from the input material, namely sulfidic concentrates. The high infiltration depth can be explained by the higher process temperatures when compared to other lead production processes: in the reaction shaft the sulfur is burned from the sulfidic concentrates, causing high temperatures (approx. 1400 ◦ C) and the formation of an oxidic phase and a SO2 -containing offgas, the latter causing sulfur attack on the furnace lining. Samples E and F represent magnesia-chromite bricks from the reverberatory furnace. In case of sample E yellow brick

4.2.1. Slag attack Similar to the copper industry3,9 the main wear mechanism of the refractory lining in the lead industry is corrosion by acidic slag. The main slag types for both industries can be roughly expressed by the system FeO–CaO–SiO2 (Fig. 4). Nevertheless, in comparison to copper production processes the slag temperatures in the lead industry are lower. Additionally, the production of lead from secondary sources such as battery scrap leads to the formation of different types of slag frequently without or with low content of CaO and SiO2 and/or soda (Na2 CO3 ).2 Such modified slag systems allow metallurgical processing at temperatures even below 1000 ◦ C. Generally the corrosion of the refractories by slag attack manifests itself in three ways:

4. Results and discussion

(a) Dissolution reaction occurring at the immediate brick hot face: The driving force here is the lower activity of the refractory oxides like MgO in the slag. The dissolution process, at least in the closed system, will continue until the liquid slag has reached saturation. However, in practice, the point of saturation is never reached and dissolution continues until the entire refractory has been consumed. (b) Dissolution and chemical reaction within the refractory microstructure: Infiltrating slag will dissolve magnesia especially from the fines and from silicates according to the respective phase equilibria. At service temperature a liquid phase will remain which is then saturated in magnesia. This will not directly contribute to corrosive wear which takes place at the refractory hot face. Nevertheless it will contribute to wear by preparing hot erosion due to a loss of brick bonding. The latter is caused by dissolution of fines in the liquid due to the grain size dependence of the solubility limit and following precipitation on the surface of course particles. This represents so called Ostwald ripening.26 (c) Kinetics of slag infiltration: Kinetics of slag infiltration causing processes mentioned above depends on several parameters like viscosity, pore size distribution and wetting angle. It is assumed that these kinetic considerations can be neglected here as the ratio of the wear rate to the infiltration velocity is much smaller than one. As a consequence, microscopic investigations of used bricks always showed a total

1688

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

Fig. 3. (a)–(f) Macroscopical overview. Selected magnesia-chromite bricks out of lead furnaces. Short rotary furnace (samples A and B). Kaldo converter (sample C). KIVCET furnace (sample D). Reverberatory furnace (samples E and F). Metal-rich coating respectively slag (S) covering the immediate brick hot face. Infiltrated (I) and non-infiltrated brick microstructure (O). Crack formation (arrows). Yellow brick coloring due to hexavalent chromium formation (sample E). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

1689

Fig. 4. Phase diagram CaO–FeO–SiO2 with slag liquidus temperatures.25 Slags for non-ferrous metals production are mainly found in the regions with low liquidus temperatures (indicated area in the phase diagram represents typical range for lead production slags). Under oxidizing conditions not only FeO but also Fe2 O3 is present in the slag, therefore generally the Fe-oxides are summarized as FeOx.

infiltration of the porosity up to the cold end. This moreover means that the invariant point of the respective mineral phase assemblages is below the cold face temperature.

In lead furnaces the low-melting, partly alkali-rich silicatic slag is able to penetrate deeply into the brick microstructure. The silica-rich slag and the reaction products formed by the corrosion of the brick-inherent component can be traced up to the cold face. This is a big difference compared to slag attack and infiltration in copper furnaces, where the infiltration depth is just a few mm from the hot face. Table 2 shows typical chemical compositions of the infiltrated magnesia-chromite bricks from the SRF, reverberatory furnace, and Kaldo furnace. The brick hot face is frequently highly enriched with SiO2 , partly Na2 O, CaO and SO3 . Additionally high lead, tin, antimony and copper content, as well as traces of Ba-oxide were determined. On microscopic scale, several zones can be distinguished at the brick hot face (see Fig. 5a–f):

- Adhering slag layer, frequently covering the immediate brick hot face (Fig. 5a and c); - Below the slag coating: reaction zone showing severe dissolution of the magnesia (Fig. 5a–c). Frequently relics of chromite precipitations after magnesia dissolution can be observed (Fig. 5c); - Adjacent to the reaction zone: infiltrated and corroded brick microstructure (Fig. 5d–f).

In the infiltrated brick microstructure, due to corrosion of the coarse magnesia grains and matrix fines the main reaction products are (Ca)–Mg-silicates, namely monticellite (CaMgSiO4 ), and forsterite (Mg2 SiO4 ), as well as Na–Mg–Fe–Al-silicate (Fig. 5d and f). To minor amounts, (Ca)–Mg–(Fe)-silicate of olivine type and Na–Ca–Ba-silicatic glassy phase can also be detected. Generally the periclase (MgO) is more basic than the second brick component chromite and therefore more susceptible to acidic corrosion than chromite. The corrosion of MgO is based on the reaction between basic oxides (MgO) and acidic oxides (slag), whereas the chromite is not corroded but modified in chemical composition due to diffusion phenomena (enrichment with iron-, zinc,- antimony-, and tin-oxide). Slag also penetrates along the crystals boundaries of magnesia grains. This phenomenon is called intragranular corrosion (Fig. 5e). Thus, especially in case of bricks with high CaO/SiO2 ratio also the interstitial phase of the magnesia, such as for example dicalcium silicate (C2 S), is highly corroded. 4.2.2. Forsterite bursting Forsterite bursting is a wear phenomenon typically observed in the glass industry.1 Especially basic bricks lined in the regenerator chamber are suffering from forsterite bursting due to a high silica carry-over. Although not so frequent as in the glass industry, this wear phenomenon was also observed in nonferrous metal production vessels, namely in lead, copper and cobalt furnaces.27 Remarkable on this wear phenomenon is that an atypical, exceptionally high silica supply (e.g., caused by changes in

1690

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

Table 2 Chemical analysis of representative magnesia-chromite bricks out of short rotary, reverberatory and Kaldo furnaces showing high SiO2 , partly Na2 O, CaO and SO3 content at the brick hot face. Short rotary furnace Na2 O

MgO

Al2 O3

SiO2

SO3 a

6.0

45.0 59.0

5.0 7.0

9.0 0.5

3.0

Reverberatory furnace Na2 O wt.%

MgO

Al2 O3

SiO2

SO3 a

Hot face* Unused**

32.0 60.0

3.0 6.0

8.0 1.5

5.0

1.0 1.1

Al2 O3

SiO2

SO3 a

CaO

2.0 0.6

4.0

2.0 1.4

wt.% Hot face* Unused**

4.0

Kaldo furnace MgO wt.% Hot face* Unused** * ** a b

52.0 58.0

5.0 6.5

Cr2 O3

Fe2 O3 b

PbO

BaO

13.0 21.0

12.0 12.0

3.6

0.3

Cr2 O3

Fe2 O3 b

PbO

CuO

SnO2

BaO

Sb2 O3

9.0 19.0

11.0 12.0

22.8

0.3

1.9

0.4

1.0

Cr2 O3

Fe2 O3 b

PbO

CuO

SnO2

12.0 19.0

10.0 14.0

5.8

0.2

CaO 3.0 0.5

CaO

6.6

Standardless semi-quantitative XRF analysis. Quantitative XRF analysis on ignited sample (1050 ◦ C). Sulfur analyzed as SO3 . Total iron analyzed as Fe2 O3 .

the processing and/or the uncontrolled addition of silica sand) results in a considerable formation of forsterite (Mg2 SiO4 ) upon contact with magnesia. The associated volume expansion causes forsterite bursting and destruction of the brick structure. A brief description of a case study showing forsterite bursting in magnesia-chromite bricks from a reverberatoy furnace is given in the following. According to the chemical analysis the brick hot face is highly enriched with SiO2 (up to 11 wt.%), SO3 and PbO (see Table 3). Additionally, presence of alkalis (Na2 O, K2 O), BaO, as well as Sb2 O3 and SnO2 could be detected. Based on the mineralogical investigations the following microstructural changes at the brick hot face can be summarized (Fig. 6): - Strong degeneration, respectively softening of the brick microstructure caused by infiltration and severe corrosion (partly intragranular) of magnesia. This resulted in formation of a huge amount of idiomorphic forsterite. The second brick component chromite became highly enriched with iron- and tin-oxide. Additionally, PbO-bearing silicatic glassy phase was determined; - Formation of coarse pores respectively pore channels; - Crack formation parallel to the hot face surface. In addition to the mineralogical investigations, phase equilibrium assemblages were calculated using the software FactSage 6.3 in order to determine the dissolution of brick material or the formation of new phases in the slag contact area. For the thermodynamic calculations standard databases in combination with Fact53, FToxid, and FTmisc databases was used. The chemical compositions of slag and refractory used in the calculations are shown in Fig. 7. The calculations were performed at 1450 ◦ C. The thermodynamically stable phases of the

magnesia-chromite brick are magnesia, (Fe,Mg)(Al,Fe,Cr)2 O4 spinel, and forsterite. In a magnesia-chromite brick the amount of forsteritic secondary phase increases when the brick is infiltrated with fayalitic slag (see Fig. 7). The iron oxide of the slag remains mainly in the liquid oxide fraction, however a small amount of the iron oxide of the slag infiltrate forms a solid solution in periclase. Considering the apparent porosity of the original bricks which is around 17 vol% and the apparent porosity of the infiltrated layer between 0 and 5 vol%, the ratio between slag mass and refractory mass that is reacting is estimated to be 12–17% (see Fig. 7). At this ratio the amount of newly formed forsterite is approximately 5–10 vol%. This may be sufficient for destabilizing the grain- and bonding structure of the brick. 4.2.3. Sulfur corrosion Another very common type of chemical attack is corrosion by sulfates. The latter is frequently observed in metallurgical vessels like KIVCET furnace, Kaldo converter, AusmeltTM reactor, reverberatory furnace and QSL reactor where mainly primary ores are charged. For instance, the magnesia-chromite brick from the KIVCET furnace shows up to 8 wt.% SO3 at the hot face (Table 4). An increased SO3 content could also be determined at the cold face. This wear phenomenon is quite similar to the observations in the copper industry, e.g. Peirce Smith converter. The penetration of gaseous SO2 caused by oxidation of sulfidic ore leads to the formation of SO3 , which below temperatures of approximately 1100 ◦ C reacts with the basic oxide of the magnesia-chromite brick leading to the formation of magnesium sulfates (MgSO4 ). The fundamental reactions are shown below3 : MeS + O2 → MeO + SO2 MeS + MeO → Me + SO2

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

1691

Fig. 5. (a)–(f) Magnesia-chromite bricks out of short rotary ((a), (c), (e), (f)), Kaldo (b) and reverberatory (d) furnace. Slag attack. Slag coating ((a) and (c)) covering the immediate brick hot face. Below that reaction zone (R) frequently with relics of primary and secondary chromite precipitations after the corrosion of the magnesia (c). Below the reaction zone infiltrated and corroded brick microstructure (I) ((d) and (f)). The main reaction products include monticellite (CMS), forsterite (Fo) and Na–Mg–Fe–Al-silicate (NaSi). Partly intragranular corrosion of magnesia grain (circle in (e)). Magnesia brick component (MgO). Chromite (Cr). Lead (Pb).

◦

SO2 + 1/2O2 ↔ SO3 (> 760 C) SO3 + MgO → MgSO4 (< 1050 ◦ C) SO3 + CaO → CaSO4 For the corrosion attack the oxidation of SO2 to SO3 is a necessary prerequisite because sulfites, such as for example MgSO3 ,

are not stable at temperatures higher than 600 ◦ C and therefore hardly ever occur under prevailing furnace conditions. Although the oxidation of SO2 to SO3 rapidly decreases above 760 ◦ C, a certain partial pressure of SO3 can be assumed in the temperature range between 760 and 1100 ◦ C that would allow the formation of basic sulfates. The intensity of sulfate corrosion generally depends on many factors such as the amount of supplied SO2 , the surplus of acidic SO2 versus the basic components of the infiltrate (e.g., alkalis

1692

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

Table 3 Chemical analysis of the magnesia-chromite brick out of reverberatory furnace showing extremely high SiO2 content at the brick hot face due to massive silica supply (forsterite bursting). wt.% face*

Hot Unused** * ** a b

Na2 O

MgO

Al2 O3

SiO2

SO3 a

K2 O

CaO

Cr2 O3

Fe2 O3 b

PbO

SnO2

BaO

Sb2 O3

0.8

49.0 63.0

5.0 7.0

11.0 2.0

4.0

0.7

0.5 1.0

11.0 17.0

10.0 10.0

6.3

0.8

1.0

0.4

Standardless semi-quantitative XRF analysis. Quantitative XRF analysis on ignited sample (1050 ◦ C). Sulfur analyzed as SO3 . Total iron analyzed as Fe2 O3 .

and CaO), reaction temperature and time, as well as brick properties such as porosity, bonding strength, kind of bonding, brick composition etc. For instance in case of high alkalis supply the solubility of CaO and MgO in overheated salt melts drastically increases and the corrosion can be much more severe. As can be seen from the phase diagram in Fig. 8 the quantity of MgO soluble in K-sulfate increases with the temperature.

In case of a magnesia-chromite bricks with high CaO/SiO2 ratio also the interstitial phase such as C2 S is massively corroded. The latter results in formation of Ca-sulfate (CaSO4 ) (Fig. 9a). Due to a severe corrosion of the brick bonding phase the initially high CaO/SiO2 ratio is decreased into the stability area of forsterite. Therefore numerous single forsterite crystals could additionally form in such melt (Fig. 9b). This is mainly caused

Fig. 6. (a)–(c) Magnesia-chromite brick from the reverberatory furnace.Severely degenerated brick microstructure with coarse pores and crack formation due to forsterite bursting (a). Infiltrated and corroded brick microstructure ((b) and (c)). Massive forsterite formation (Fo) caused by corrosion of magnesia (MgO). Chemically changed and partly recrystallized chromite (Cr).

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

1693

Table 4 Chemical analysis of representative magnesia-chromite bricks out of KIVCET furnace showing high sulfur supply on the brick hot face. wt.% face*

Hot Data sheet** * ** a b

Na2 O

MgO

Al2 O3

SiO2

SO3 a

K2 O

CaO

Cr2 O3

Fe2 O3 b

PbO

0.6

55.0 58.0

3.0 6.5

0.7 0.6

8.0

0.3

0.9 1.3

7.0 19.0

6.0 14.0

18.0

Standardless semi-quantitative XRF analysis. Quantitative XRF analysis on ignited sample (1050 ◦ C). Sulfur analyzed as SO3 . Total iron analyzed as Fe2 O3 .

by sulfate melt condensating close to the brick hot face. The possible reaction is schematically represented as follows: SO3 + 2Ca2 SiO4 + MgO - > Ca3 MgSi2 O8 + CaSO4 This means that the CaO/SiO2 ratio of the silicate phases is shifted to a lower value, and this similarly proceeds until forsterite is formed. As indicated by the above equation, this reaction takes place only if CaSO4 , which is a component of the sulfate melt and not a solid phase, is penetrating to regions more distant from the hot face. At temperatures above 1100 ◦ C, MgSO4 will dissociate and form fine crystalline MgO at magnesia rims (Fig. 9a and c). This takes place without adequate rebuilding of the original ceramic bonding of the brick.

Fig. 7. Calculated phase assemblage of infiltrated magnesia chromite brick as a function of slag mass to refractory mass ratio (pure brick at coordinate axis, x = 0):27 Mainly periclase and spinels, and minimal forsterite. Infiltrate is forsterite forming. The apparent porosity is approximately 17 vol% (* ). In this case, [wt.%] is identical with [vol%], as the brick and slag have the same density—therefore, the percentages are only indicated as [%]. Brick density 3.18 g/cm3 , refractory composition [wt.%]: 6.9Al2 O3 , 2SiO2 , 10Fe2 O3 , 59MgO, 1CaO, 17.1Cr2 O3 slag composition [wt.%]: 43MgO, 25.3SiO2 , 3.2Al2 O3 , 11.6Fe2 O3 , 1.7Cr2 O3 , 1Na2 O, 1.2BaO, 7PbO, 5SO3 .

4.2.4. Iron oxide attack Either at the interface between the slag coating and the brick hot face or in the first few mm from the hot face the magnesiachromite brick is highly enriched with iron oxide. As can be seen from the chemical analysis, up to 26 wt.% of Fe2 O3 (respectively 23.4 wt.% FeO) can be determined (Table 5). This is most frequently observed in magnesia-chromite bricks from SRF where iron scrap is used as reducing agent in the processing of lead batteries. The high iron oxide supply results in degeneration of the brick microstructure and formation of Mg–Fe-oxide of magnesiawuestite type (see Fig. 9d). 4.2.5. Non-oxidic infiltration In addition to acidic slag also other components like metallic lead, PbS, etc. infiltrate the brick microstructure. Generally, non-oxidic infiltration only densifies the bricks microstructure without any corrosive attack on the brick components. Similar to

Fig. 8. System K2 SO4 –MgO.28

1694

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

Fig. 9. (a)–(f) Magnesia-chromite bricks out of the KIVCET ((a)–(c); (e)–(f)) and short rotary furnace (d). Corrosion of the magnesia (MgO) by sulfur attack ((a)–(c)). Newly formed, Ca-sulfate (CaS, (a)) and idiomorphic periclase crystals at magnesia rims (arrows in (b)–(c)). Idiomorphic forsterite crystals (circle in (b)). Example of microstructural degeneration due to high iron oxide supply (d). Enrichment of the magnesia with iron-oxide and formation of magnesia-wuestite. Microstructural changes due to high temperature load showing periclase crystal growth and lengthening toward the thermal gradient (e). Example of highly reduced brick microstructure (f). Partly bondingless, severely reduced chromite. Chromite (Cr). The chromite core is filled with lead-oxide. Table 5 Chemical analysis of representative magnesia-chromite bricks out of short rotary furnace showing high iron oxide attack at the brick hot face. wt.% face*

Hot Data sheet** * ** a b

Na2 O

MgO

Al2 O3

SiO2

SO3 a

CaO

Cr2 O3

Fe2 O3 b

PbO

CuO

ZnO

2.0

39.0 60.5

4.0 7.0

4.0 1.5

3.0

2.0 0.4

11.0 19.0

26.0 11.6

8.3

0.2

0.4

Standardless semi-quantitative XRF analysis. Quantitative XRF analysis on ignited sample (1050 ◦ C). Sulfur analyzed as SO3 . Total iron analyzed as Fe2 O3 .

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

1695

Fig. 10. Microstructural detail. Line-scan with SEM through reduced chromite. The chromite core is strongly depleted in Cr2 O3 and Fe2 O3 , as well as highly enriched in PbO.

the acidic slag, the degree of infiltration depends on the surface tension, the boundary angle in contact with the refractory oxides, the metal density, the bath height and the size/distribution of the brick pores.3 4.2.6. Microstructural changes due to high temperature load Although the temperatures in lead furnaces are typically quite below the maximum service temperature of the as-delivered refractory materials, the temperature plays an important role in the continuous wear of the bricks. Through interaction with the bath material, the hot strength of the brick can be reduced seriously. The higher degradation rate caused by the reduced viscosity and higher diffusivity associated with an overheated melt clearly demonstrates the effect of temperature. On the microscopic level a very characteristic feature for this wear phenomenon is periclase crystal growth and lengthening toward the thermal gradient (up to approx. 1 mm long crystals, Fig. 9e). The crystal size of the single periclase crystals within the original sintered magnesia is usually up to 140 ␮m.1 However, in highly degenerated brick microstructure the single grains, as well as the classical ceramic microstructure with coarse grains and matrix fines cannot be observed anymore. In addition to this, also the supplied oxide components such as lead-, zinc-, iron- and tin-oxide act as mineralizing agents, thus strongly supporting the periclase crystal growth. Such microstructural changes are quite typical for ferroalloy furnaces where the operating temperature is much higher than in base metal production. So far in lead processing this wear phenomenon was observed in the KIVCET furnace only, possibly due to the high furnace operation temperature. 4.2.7. Reduction by varying or low oxygen pressure Especially in the KIVCET furnace, IsasmeltTM and in the SRF, under reducing conditions, caused by low oxygen partial pressure or by the presence of metals, trivalent iron can be reduced to bivalent and further on to metallic iron. In the

magnesia-chromite brick especially the original chromite and newly formed chromite precipitations are highly sensitive to such phenomena (Fig. 9f). The change in the iron valence results in a volume decrease and causes voids between the single grains. In case of a subsequent oxidation the original brick microstructure will not reform. A frequent repetition of these reactions (Red-Ox) will result in a disintegrated brick microstructure. As can be seen from the SEM line-scan and element distribution shown in Figs. 10 and 11 the chromite core is strongly depleted in Cr2 O3 and Fe2 O3 , as well as highly enriched in PbO. The latter is caused by PbO infiltration in the voids.

4.2.8. High soda supply The high alkali supply in secondary lead production vessels results in a formation of chromium (VI) salts. The unmistakable sign for the hexavalent chromium formation is yellow coloring, which is caused by salts formed on the brick surface (Fig. 3e). By use of the powder X-ray diffraction alkali bearing chromates frequently as a mixtures of K3 NaCr2 O8 and K2 CrO4 (tarapacaite) can be identified.11 Both mentioned salts contain toxic chromium (VI). The chromate formation is caused by the key agents potassium and sodium contained in the furnace feed. These alkalis, along with oxygen in the furnace in contact with magnesiachromite bricks can initiate the following reaction11 : FeCr 2 O4 + 2K2 O + 7/2O2 = 2K2 CrO4 + 1/2Fe2 O3 A similar reaction can also be expected for Na2 O. Degeneration of the refractory, especially magnesia-chromite brick is best detectable at the microscopical level. In case of high alkali supply formation of Na-bearing chromates could be observed at chromite rims. According to thermodynamic calculations by FactSage 6.3 the formation of the latter phases occurs in a temperature range between 900 and 1000 ◦ C.

1696

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

Fig. 11. Microstructural detail. Element distribution with SEM. Reduced chromite showing severe depletion on Cr2 O3 and Fe2 O3 , as well as PbO-enrichment in the chromite.

4.2.9. Brick damage by hydration Although magnesia-chromite bricks provide high refractoriness compared to other refractories, there is still a great concern regarding their vulnerability when exposed to water vapor at low temperatures and subsequent hydration.1 Generally hydration of magnesia-chromite bricks can occur in humid atmospheres preferably at temperatures between 40 and 12 ◦ C.1 It is characterized by the transformation of periclase into brucite (Mg(OH)2 ) according to the reaction: MgO + H2 O → Mg(OH)2

This reaction is associated with an increase in volume of up to 115%. Additionally, the formation of brucite is accompanied by an expansion of the crystal lattice because of the oriented growth of the hexahedral brucite on the cubic periclase—the space between the magnesium ions within the brucite is larger than in the periclase. Due to this reaction the periclase crystals are separated along the natural, crystallographical cleavage planes. The resulting “microcracks” represent new reaction surfaces, which again show a high tendency toward hydration. This leads to a chain reaction which, starting just from a few centers, can finally cause extensive, characteristical crack formation and

Fig. 12. (a)–(b) Example of severe hydration of the magnesia (a). Crack formation along the cleveage planes (arrows) of the MgO. In the crack newly formed brucite (Bc) (b). Chromite (Cr).

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

in extreme cases a sand-like disintegration of the whole brick. Influencing factors having an effect on hydration are humidity, temperature and time. The hydration process can occur during storage, drying after installation or during service. Fig. 12a and b shows a typical hydrated magnesia-chromite brick from an AusmeltTM reactor. The hydration occurred at the brick cold face due to a leakage in the water cooling system. 5. Conclusion Post mortem studies clearly highlight which specific stresses affect the refractory products performance in the lead processing furnaces. Despite of their variety, all wear phenomena described in the present paper lead to significant refractory destruction by degeneration of the brick microstructure, independent of their origin (i.e., production vessel). Different typical wear mechanisms can be observed in the individual metallurgical furnaces, depending on the specific metallurgical process and furnace design. Furthermore post mortem studies demonstrated clearly that as consequence of chemical attack (by acidic slag, high supply on SiO2 , sulfur, soda, iron-oxide) the infiltration and corrosion of the brick’s inherent components led to a softening of the refractory microstructure, loss of flexibility, and brick strength. Such weakened microstructure is highly susceptible to continuous wear by hot erosion. Additionally, due to changed (thermo-)mechanical properties, crack formation, primarily at the interface between the infiltrated and non-infiltrated brick microstructure occurs. Discontinuous wear by spalling is then the final consequence. Non-corrosive infiltration of the brick microstructure by metals or sulfidic components dramatically changes the thermal conductivity of the brick, thus increasing the susceptibility to crack formation and spalling, especially intensified by thermal shocks. Affected by humid atmosphere or leakage in the water cooling system periclase is converted into brucite. This leads to irreversible microstructural defects combined with characteristical crack formation and complete loss of the strength. Also iron oxide reduction by varying or low oxygen pressure will result in permanent defects. Once loosened, the brick microstructure will not reform. This reduces the strength of the brick and increases the susceptibility of the microstructure to further infiltration/corrosion. The precise knowledge of the wear phenomena is an absolutely necessary prerequisite for the refractory producer for providing appropriate brick lining solutions. RHI enhances this refractory optimization process through active collaborations with its customers. The post mortem studies additionally provide an important base for the focused product development, leading to new refractory products for the lead industry (e.g., optimized mineralogical structure). Acknowledgements The authors would like to thank Dr. Markus Kirschen for his valuable contributions (FactSage calculations) to this paper.

1697

References 1. Routschka G. Pocket manual refractory materials. Essen, Germany: VulkanVerlag GmbH; 2008. 2. Köffel M, Taschler T. Refractories for the copper and lead industry. Erzmetall - World of Metallurgy 2006;59:133–42. 3. Barthel H. Wear of chrome magnesite bricks in copper smelting furnaces. Interceram 1981;30:250–5. 4. Rigaud M. Corrosion of refractories and ceramics. In: Winston Revie R, editor. Uhlig’s corrosion handbook. 3rd ed. Hoboken, New Jersey, USA: John Wiley and Sons, Inc.; 2011. 5. Gregurek D, Ressler A, Reiter V, Franzkowiak A, Spanring A, Prietl T. Refractory wear mechanism in the nonferrous metal industry: testing and modeling results. JOM 2013;65:1622–30. 6. Gregurek D, Wenzl C, Reiter V, Studnicka HL, Spanring A. Slag characterization: a necessary tool for modeling and simulating refractory corrosion on a pilot scale. JOM 2014;66:1677–86. 7. Oprea G. Failure mechanism of refractory linings for non-ferrous flash smelting furnaces. In: Proceedings of Tehran international conference on refractories. 2004. 8. Prestes E, Chinelatto ASA, Resende WS. Post mortem analysis of burned magnesia-chromite brick used in short rotary furnace of secondary lead smelting. Ceramica 2009;55:61–6. 9. Monshi A, Akhtari M, Fakhouri M, Tahery MR, Abbaszadeh H. New generation of refractories for the lead industry. In: Proceedings of Tehran international conference on refractories. 2004. 10. Hoed P. An anatomy of furnace refractory erosion: evidence from a pilotscale facility. In: 58th electric furnace conference proceedings. 2000. 11. Wei L. Corrosion of refractories in lead smelting reactors. University of British Columbia; 2000 (MSc thesis). 12. Scheunis L, Mehrjardi AF, Campforts M, Jones PT, Blanpain B, Jak E. The effect of phase formation during use on the chemical corrosion of magnesiachromite refractories in contact with non-ferrous PbO–SiO2 based slag. J Eur Ceram Soc 2014;34:1599–610. 13. Malfliet A, Lotfian S, Scheunis L, Petkov V, Pandelaers L, Jones PT, et al. Degradation mechanism an use of refractory linings in copper production processes: a critical review. J Eur Ceram Soc 2014;34: 849–76. 14. Prengaman RD. Secondary lead production—a survey of smelters and refineries. In: Lead & zinc ’05. Proceedings on the international symposium on lead and zinc processing. 2005. 15. Elvers B, Hawkins S, Russey W. Ullmann’s encyclopedia of industrial chemistry. 5th ed. Hoboken, New Jersey, USA: John Wiley & Sons; 1997. 16. Dittmeyer R, Keim W, Kreysa G, Oberholz A. Winnacker-Küchler: Chemische Technik. 5th ed. Weinheim, Germany: Wiley-VCH Verlag; 2006. 17. Lee YH, Park YM. The experience of lead direct smelting in Korea Zinc’s Onsan Refinery. In: Yazawa international symposium on metallurgical and material process. 2003. 18. Spanring A. Refractory applications in the secondary lead industry. RHI Bull 2010;1(56):62. 19. Yazawa A, Nakazawa S. Comparison between copper, lead, and nickel smelting processes from thermodynamic viewpoints. In: Sulfide smelting’98: current and future practices. 1998. 20. Stephens RL. Advances in primary lead smelting. In: Lead & zinc ’05. Proceedings on the International Symposium on Lead and Zinc Processing. 2005. 21. Sofra J, Hughes R. Ausmelt technology operation at commercial lead smelters. In: Lead & zinc ’05. Proceedings on the international symposium on lead and zinc processing. 2005. 22. Meurer U, Püllenberg R, Griesel J, Developing QSL. Technology to an economicallly superior process. In: Lead & zinc ’05. Proceedings on the international symposium on lead and zinc processing. 2005. 23. Richards GG, Stephens RL, Ashman DW, Kuhn JM. Managmgement of KIVCET bath accretions. In: Lead & zinc ’05. Proceedings on the international symposium on lead and zinc processing. 2005. 24. Errington B, Arthur P, Wang J, Dong Y. The Isa-YMG lead smelting process. In: Lead & zinc ’05. Proceedings on the international symposium on lead and zinc processing. 2005.

1698

D. Gregurek et al. / Journal of the European Ceramic Society 35 (2015) 1683–1698

25. Osborn EF, Muan A. Phase equilibrium diagrams of oxide systems. Columbus, OH, USA: American Ceramic Society with the Edward Orton Jr. Ceramic Foundation; 1960. 26. Brüggmann C. Contribution to the slagging of MgO in secondary metallurgical slags. Refractories Worldforum 2011;3:105–9.

27. Gregurek D, Majcenovic C, Spanring A, Kirschen M. Forsterite bursting in magnesia chromite bricks—two case studies from lead and copper smelting furnaces. RHI Bull 2011;2:49–53. 28. McDaniel CL. Phase diagrams for ceramists. Columbus, OH, USA: American Ceramic Society with the Edward Orton Jr. Ceramic Foundation; 1975.