��& ]_^ `ba �±��FKURPLXP�R[LGH�FRQFHQWUDWLRQ�LQ�WKH�VROLG�&U c 2 d �FRPSRQHQW�LQ�HTXLOLEULXP�ZLWK�WKH�VDWXUDWLRQ�FRQFHQWUDWLRQ�RI�FKURPLXP�R[LGH�LQ�WKH�VSLQHO��J�FP e d ��&fb^ f g �±��VDWXUDWLRQ�FRQFHQWUDWLRQ�RI�&U c 2 d �LQ�WKH�VSLQHO��J�FP e d ��& ]_^ f g �±�FRQFHQWUDWLRQ�RI�FKURPLXP�R[LGH�LQ�WKH�VSLQHO�LQ�HTXLOLEULXP�ZLWK�WKH�VDWXUDWLRQ�OHYHO�RI�FKURPLXP�R[LGH�LQ�WKH�VODJ��J�FP e d ��&fb^ f�h �±�VDWXUDWLRQ�FRQFHQWUDWLRQ�RI�FKURPLXP�R[LGH�LQ�WKH�VODJ��J�FP e d ��& ]_^ f�h �±�&U c 2 d �FRQWHQW�LQ�WKH�VODJ�DW�HTXLOLEULXP��J��G i �±�HIIHFWLYH��VSLQHO��UHDFWLRQ�OD\HU�WKLFNQHVV��FP��G j �±�HIIHFWLYH�ERXQGDU\�OD\HU�WKLFNQHVV�LQ�WKH�VODJ��FP���
CHAPTER 5 PROJECT GOALS, METHODOLOGY AND ANALYSIS TECHNIQUES
5.1. Project goals The aim of this study is to determine and investigate the degradation phenomena affecting the refractory lining used in the anode furnaces at Cumerio Med. This can be considered as the scientific goal of the thesis, while the industrial goals are (i) find a way to extend the refractory lining lifetime with 7 months, i.e. increase the anode furnaces campaign from 11 to 18 months and (ii) explore the possibility of applying chrome-free bricks in anode furnace linings. Since every 18 months the acid plant of the smelter has to stop for repairs and maintenance, the idea emerged to coordinate this with the stoppage of the anode furnaces, which would generate considerable economic benefits. However, in order to achieve the industrial goals of this thesis, it was first necessary to acquire knowledge about the main refractory wear mechanisms. 5.2. Methodology Laboratory as well as industrial tests can be applied in order to investigate refractory degradation mechanisms. However, both methods have not only benefits but also drawbacks. When studying complex refractory wear phenomena in a real plant environment, it is obvious that laboratory and industrial tests should be combined to obtain the best possible outcome. In the present thesis, the emphasis is put on laboratory testing methods combined with a detailed investigation of industrially worn bricks. The reasons for this choice are the following: (i) the limited availability of literature on laboratory testing of any type of refractory bricks exposed to contact with anode slag and (ii) a lack of capability for industrial testing. In the beginning of this study, a proposal existed for the installation of a new (and larger) anode furnace in early 2004, which would provide the opportunity to use one of the older ones for industrial tests. Unfortunately, the proposal was abandoned, which made the performance of plant trials impossible. Nevertheless, numerous investigations of samples recovered from industrially worn bricks were carried out. The knowledge thus acquired was combined with an assessment of the available literature on refractory degradation phenomena in a copper-making environment. Finally, laboratory experiments were carried out in order to overcome certain restrictions associated with the analysis of the industrially worn bricks. Worn refractory samples were collected from the anode furnaces working lining when the latter were taken out of production for maintenance and relining. The bricks were recovered from characteristic locations while attempting to use the same locations during the three consecutive collections. Owing to some operational difficulties this was not always possible. During the manual recovery of the bricks, their location in the lining was recorded, as well as all available data on the history of that particular lining (e.g. average temperatures, average duration of oxidation and reduction stage, waiting times, partial
102
relinings, any interruptions in normal operational practice …) were collected. From each of the refractory bricks selected for investigation several samples were prepared for microstructural analysis using Scanning Electron Microscopy (SEM) and Electron Probe Micro-Analysis (EPMA). This methodology allowed for the investigation of:
a) the wear mechanisms affecting the (old) direct-bonded magnesia-chromite lining b) the wear mechanisms affecting the (new) mixed type of magnesia-chromite
lining, involving different refractory qualities The laboratory experiments consisted of the so-called ‘static finger tests’. With the help of the latter twelve types of commercially available refractory bricks were tested. Six of them are chrome-free types and the other six are mag-chrome based. The experiments were conducted in the following way: An alumina crucible was filled with electro-refined copper and anode slag, and then heated in a laboratory resistance furnace until a temperature of 1300°C was reached. Four refractory fingers were simultaneously submerged in the molten content of the crucible. The fingers were held in the melt for 24 hours at a constant temperature of 1300°C (± 5°C). A detailed description of the tested refractory materials and the experimental procedure is provided in Chapter 7. 5.3. Analysis techniques 5.3.1. Introduction The majority of the chemical analyses and microstructural characterizations of refractory samples were performed by SEM using a high resolution Philips XL30 FEG microscope equipped with an energy dispersive spectroscopy (EDS) system. Additional compositional and microstructural information was obtained by means of a JEOL JXA-733 electron probe microanalysis system (EPMA), also provided with an EDS system. Scanning electron microscopy (SEM), typically used for the observation of bulk specimens, operates with a fine electron beam illuminating the sample and generating high-energy primary backscattered electrons (BSEs), low-energy secondary electrons (SEs) and characteristic X-rays, which can be used for micro-analysis. SE images are of particular interest when topographic features about the sample are required, whereas BSE images are preferred when information about the different phases present is desired [Jon 2001]. The latter are exclusively used in this thesis and are, therefore, discussed in the following paragraph. BSEs arise when primary beam electrons interact with one or more atomic nuclei, reversing their direction of travel with limited energy loss, resulting in their re-emergence from the sample surface. The number of BSEs produced during this interaction grows with increasing atomic number (Z). The fraction of incident electrons, which reappear as BSEs is the so-called BSE coefficient . The relationship between Z and �has been determined experimentally for the pure elements [Llo 1987]. However, for minerals a
weighted mean BSE coefficient ( ) is required. The mean BSE coefficient is a function of the weighted (average) atomic number ( Z ) of the mineral phase, which can be calculated as:
103
¦¦ ��
)(
)(
AN
AZNZ , (5.1)
where N is the number of atoms of each element with atomic weight A and atomic number Z. Table 5.1 shows the average atomic numbers of some minerals which are encountered in refractory analysis. A variation of atomic number within the specimen therefore provides a variation in the amount of BSEs produced, thus resulting in image ‘Z-contrast’ [Llo 1987 & Gol 1981]. The higher this number, the brighter that particular phase appears on a BSE micrograph. Similar to SEs, the average depth from which the BSEs originate depends on the primary beam energy and the sample density. The best BSE images are obtained from flat specimen surfaces [Jon 2001].
Table 5.1. Average atomic number ( Z ) of some relevant minerals
Mineral Chemical formula Z
Magnesia MgO 10.41
Chromite spinel (MK) MgO.Cr2O3 17.15
Spinel (MA) MgO.Al2O3 10.57
Dicalcium silicate (C2S) 2CaO.SiO2 14.56
Monticellite (CMS) CaO.MgO.SiO2 12.77
Merwinite (C3MS2) 3CaO.MgO.2SiO2 13.71
5.3.2. Scanning electron microscope Philips XL30 FEG The scanning electron microscope Philips XL 30 FEG (Figure 5.1) employs a Schottky based gun design using a point-source cathode of tungsten, which has a surface layer of zirconia (ZrO). The working temperature of the emitter is 1800 K, the tip is always kept clean, flashing is never needed and it takes only a minute to become fully operational for a long period. The technical details of the microscope are presented in Table 5.2. The electron source possesses both low energy spread and low current fluctuations which results in higher effective currents in smaller probes. Particularly for low beam energies the high spatial resolution is the answer for modern materials microscopical investigation. This instrument has the additional advantage of being more reliable for microanalysis of light elements: an EDAX energy dispersive X-ray detector with an ultra thin window is mounted. This allows obtaining analytical information, which is supported by an image created using a solid state backscattered electron detector. In order to attain the optimal settings for both, SEM and analytical work, the microscope is equipped with a multiple objective lens aperture. A CCD camera is mounted to allow the user to control the position of the sample inside the specimen chamber. The SEM and the EDAX system operate within a Microsoft Windows environment – in this configuration they share a common Central Processing Unit and graphical user interface. The EDAX software allows qualitative and standardless quantitative analysis of elements starting from boron on. Software for automated spot analysis, line scans and mapping is available. Data and images can be stored on a hard disk, diskettes or ZIP discs. Printouts, video print output
104
and the use of type 120 negative films on an ultra-high resolution photo monitor offer the user multiple output possibilities [KUL 2006]. More technical information is available at: [email protected].
Figure 5.1. Scanning electron microscope Philips XL 30 FEG
Table 5.2. Technical details of the scanning electron microscope Philips XL 30 FEG
Resolution 2.0 nm at 30 KV; 5.0 nm at 1 KV High voltage continuously variable from 0.2 till
30 KV Specimen stage: X & Y Z Rotation Tilt
50 mm 8 mm + 20 mm inside n x 360° at FWD 10 mm, -15° to +75°
1 x 10-4 Pa (1 x 10-6 mbar) 1 x 10-5 Pa (1 x 10-7 mbar) 5 x 10-7 Pa (1 x 10-9 mbar)
Image storage 8 standard images (702 x 484 pixels - 8 bits) or 2 high definition images (1404 x 968 pixels - 8 bits)
Backscattered electron detector better than 0.1 Z at Z = 30 High resolution photomonitor vertically positioned 7", 2000 lines
monitor Super Ultra thin Window – detection of elements from boron on Resolution 135 eV F.W.H.M. by MnKa
105
5.3.3. JEOL JXA-733 electron probe microanalysis system The JEOL JXA-733 electron probe (Figure 5.2) is a high performance X-ray micro analyser for qualitative and quantitative analysis of a variety of specimens and their composing elements. Capable of handling up to five spectrometers as well as a high resolution scanning electron microscope makes this micro analyser an indispensable instrument in fields of research and industrial engineering. The finely focused electron probe irradiates the specimen generating X-rays, secondary electrons (SEs) and back-scattered electrons (BSEs). X-rays irradiated from the surface of the specimen are analysed on one hand according to their energy and on the other hand according to their wavelength. The specimen can be viewed under direct illumination through an optical microscope installed on the column. Secondary and back-scattered electron images, both compositional and topographic, can be seen on the main control panel (CRT) and photographed using the camera for scanning images [KUL 2006].
Figure 5.2. JEOL JXA-733 electron probe microanalysis system
The JEOL 733 used in this thesis is equipped with an EDS system from Noran (5400 Series), consisting of a beryllium window and a Si (Li) detector. A Semi Quantitative (SQ) procedure was used for quantitative elemental analysis. This routine performs quantitative analysis of X-ray spectra using a library of references stored on disk. The program utilises Multiple Least Squares Analysis and the so-called Phi-Rho-Z (PRZ) algorithm for quantitative correction of matrix effects in electron beam excited X-ray spectra. This theoretical procedure, which corrects for atomic number (Z), absorption (A), and fluorescence (F) effects, differs slightly from the traditional ZAF (‘matrix’) correction method. The PRZ algorithm, which is considered to be more accurate for low-energy X-ray lines [Ame 1997], needs the following inputs: excitation conditions (accelerating voltage, probe current), detector geometry (take-off angle), and intensity ratios (i.e. the ratio of the peak intensity of the unknown to the intensity of the
106
corresponding pure element). The Phi-Rho-Z method treats the atomic number and absorption factors together, and not as separate entities, as is the case in ZAF methods [Phi 1968, Gol 1981]. The technical details of the JEOL JXA-733 electron probe microanalysis system are presented in Table 5.3. More technical information is available at: [email protected].
Table 5.3. Technical details of the JEOL JXA-733 electron probe microanalysis system
Guaranteed resolution 0.70 nm Accelerating voltage 1 - 50 kV in 1 kV steps Probe current range Probe current stability Analysable elements
10-12A - 10-5A ±0.2 %/h at 25 kV, 50nA 5 B - 92 U
Specimen stage: Shift range X-axis 0 - 32 mm
Y-axis 0 - 50 mm Z-axis 10 - 32 mm
Tilt 0° - 60° eccentric Rotation 360° endless WD 11 mm - 31 mm Size specimen 32 mm diameter x 25 mm height EDS X-ray spectrometer: Be-window 10 mm diameter Resolution FWHM 141 Ev
x Oxide selection The SQ program requires that the operator selects a priori the oxides (e.g. FeO, Cr2O3, Al2O3) or elemental compounds (e.g. Fe, Cr, Al) to be measured. Combinations of oxides and elemental compounds (e.g. Fe and MgO) are not possible in the present software configuration. With this electron probe setup, only elements from sodium (Na) onwards can be determined. Since oxygen cannot be measured directly, it has to be calculated indirectly from the measured amount of elemental compounds and the selected oxide stoichiometry [Jon 2001].
x Measurement types
Typically, two types of measurements were performed using both Philips XL30 FEG and JEOL JXA-733: (i) spot analyses of particular positions in the sample and (ii) global analyses of more extensive zones. The latter are useful to acquire average compositions, whereas the former are particularly interesting for the accurate and precise analysis of a distinguishable refractory phase. However, when the phase becomes smaller than a critical size, the measurement is distorted through matrix excitation phenomena. This can be seen in Figure 5.3, which shows the pear-shaped interaction volume of the electrons in a sample. It is obvious that the generation zone of SEs and BSEs is not as deep as that of X-rays [Gol 1981]. During spot analysis, the targeted phase is imaged from a SE or BSE scan, but measured through X-rays which originate from a much larger depth than that of the image forming electrons, subsequently affecting the measurement result. The
107
magnitude of the critical size is dependent on the acceleration potential, sample type, imaging mode (BSE or SE), etc [Jon 2001].
During the present investigation an accelerating voltage of 15-20 kV and an aperture (spot) of 3-4 were used.
Figure 5.3. Schematic of the pear-shaped interaction volume of electrons in a sample generating SEs, BSEs and X-rays [Jon 2001]
108
CHAPTER 6
INVESTIGATION OF INDUSTRIALLY WORN DIRECT-BONDED MAGNESIA-CHROMITE REFRACTORY BRICKS
6.1. Introduction At the time when this thesis was initiated, the anode furnace linings were relatively simple with just one refractory type (traditional direct-bonded magnesia-chromite brick) applied in all areas except the tuyere blocks. This initial setup was far from ideal. The refractory lifetime was limited (both furnaces had to be relined every 11 months, which caused significant financial loss); on-site observations revealed that there were severely worn bricks in some areas and almost intact bricks in other areas. In order to determine and investigate the degradation mechanisms occurring during refractory service, worn bricks were recovered from distinct locations in the anode furnace linings. The bricks were cut into slices and samples were prepared for further examination. The refractories were collected during the relining periods of the anode furnaces in three consecutive years (2003, 2004 and 2005). The results of the investigation performed on the first set of samples prepared from industrially worn bricks recovered from lining N1 are described in this chapter. With the exception of the tuyere blocks, where a special magnesia-chromite-zirconia brick was employed, medium-quality direct-bonded magnesia-chromite refractories (type RADEX H60) were the only material used in the old lining concept of the anode furnaces at Cumerio Med. These refractory bricks suffered from acute wear due to the stringent chemical, thermal and mechanical conditions imposed on them. The chemical composition of the bricks and some of their physical properties are listed in Table 6.1. During the general relining of the furnaces bricks from different locations of the worn lining were collected. A plan of an anode furnace with the collection areas (zones) is presented in Figure 6.1. These particular positions were selected in order to collect bricks from each distinct part of the furnace (roof, bottom, side walls, middle part of the barrel, close to the tuyeres blocks). Bricks from all areas were cut into slices (perpendicular to the refractory hot/cold face) with a thickness of approximately 2 cm. Pictures of slices, representative for the different areas, are shown in Figures 6.2 and 6.3. Table 6.1. Chemical composition & physical properties of refractory type RADEX H60*
MgO [wt%]
Cr2O3 [wt%]
Fe2O3
[wt%] Al2O3 [wt%]
CaO [wt%]
SiO2
[wt%] BD
[g/cm3] CCS
[MPa] AP
[vol%] 58.0 19.0 15.0 6.0 1.5 0.5 3.15 33 17.5
BD – bulk density, CCS – cold crushing strength, AP – apparent porosity *Data provided by refractory supplier
109
1. Barrel wall, opposite tuyere 1 2. Barrel wall, opposite tuyere 2 3. Bottom, opposite the mouth 4. Side wall, below the gas outlet
5. Side wall, below the burner inlet 6. Roof, near the charging mouth T1. Close to Tuyere 1 block T2. Close to Tuyere 2 block
Figure 6.1. Anode furnace plan with collection areas (zones)
The first set of investigated industrially worn samples (collected from lining N1) only consisted of direct-bonded magnesia-chromite refractories type RADEX H60. As can be seen from Table 6.2, there are significant differences in the wear rate of bricks coming from different zones in the lining. For example, bricks recovered from zone 6 were only slightly affected by the corrosive and erosive action of copper/anode slag while bricks collected from a region close to the tuyere blocks were extremely degraded. Zones 1 and 2 suffered from a significant wear while zones 3, 4 and 5 were affected at an intermediate rate. The slice shown in Figure 6.2a is considered representative for zones 1 and 2. The refractory cold face is situated at the left. The copper penetration is clearly visible (right and middle part of the brick). The slice shown in Figure 6.2b is considered representative for zones 4 and 5 since the wear and penetration patterns of these zones were found to be similar. Apart from metallic copper, copper oxide is also visible covering the left part of the brick.
Table 6.2. Summary of lining N1 performance
Lining Collection zone (see Figure 6.1)
Brick type Residual brick thickness (% of original thickness)
Figure 6.2. Slices of bricks recovered from zone 2 (a) and zone 5 (b)
Nearly no metal/oxide infiltration was observed in the slice of the brick recovered from zone 6 (Figure 6.3a). Most of the brick structure remained unchanged. The heaviest infiltration and wear of all areas was observed in the bricks recovered from an area close to the tuyeres zone (Figure 6.3b). The entire refractory (or rather, what was left of it - approximately one fourth of the original thickness) was infiltrated by metallic copper.
a) b) C- cold face, H- hot face
Figure 6.3. Slices of bricks recovered from zone 6 (a) and from the tuyere zone (b)
6.2. Sample preparation Two small specimens, A and B (approximately 2 cm u 2 cm each), were cut with a diamond saw from each slice (see Figure 6.4). Afterwards, the specimens were embedded in low-viscosity resin (Technovit 4004), ground with diamond plates and polished with diamond suspensions. Finally carbon was evaporated on their surface to provide a conducting layer for microscopical examination. The latter was done on a sample surface perpendicular to the brick/melt interface.
5 cm 5 cm
5 cm 5 cm
111
Figure 6.4. Samples cutting procedure
6.3. Results of SEM examination
a) Samples recovered from zone 1 (barrel wall, opposite tuyere 1) and zone 2 (barrel wall, opposite tuyere 2) BSE images of the samples are presented in Figure 6.5. Combining samples from both zones is possible since they have similar features. The samples are heavily infiltrated with metallic copper. The latter has filled most of the grain boundaries together with forsterite (2MgO.SiO2) replacing, in this way, the intergranular secondary chromite (Fig. 6.5a). Large cracks filled with copper and (to a lesser extent) with copper oxide were observed in many samples (Fig. 6.5b).
a) b)
c) d) F - forsterite (2MgO.SiO2), Sil - silicate phase, Sp- in situ formed spinel phase
[(Mg,Fe,Cu)(Fe,Cr,Al)2O4]
Figure 6.5. BSE images of samples recovered from zone 1 and zone 2
A
B
Brick/melt interface
“hot” face
“cold” face
Brick
Melt
Cutting plane
Investigated surfaces
112
Copper oxide is mainly present as cuprous oxide (Cu2O), although a minor amount of cupric oxide (CuO) was also detected at the hot face. Optical microscopy was used to distinguish between the two oxides. Close to the hot face the following phenomenon was observed to have taken place during refractory service: chromite grains (Mg,Fe)(Fe,Cr,Al)2O4 have grown and formed a network of interconnected spinel grains incorporating copper in their composition (Mg,Fe,Cu)(Fe,Cr,Al)2O4. Nearly complete spinelisation of the refractory microstructure has taken place at the hot face due to the incorporation of some periclase into the newly formed kind of spinel (Fig. 6.5c). Small amounts of silicate phase are enclosed in this expanding spinel phase (Fig. 6.5c). At higher magnification (Fig. 6.5d) it can be seen that the silicate phase involves cuprous oxide grains scattered in a (FeOx.Al2O3.SiO2) matrix.
b) Samples recovered from zone 3 (furnace bottom, opposite the mouth) An overview BSE image of a sample from zone 3 is presented in Figure 6.6. This is the refractory zone, which is exposed first to the contact with blister when the latter is poured into the furnace. Subsequently, the wear rate of the bricks in this zone is higher than the average since they were subjected to higher erosion and larger thermal shocks. Cuprous oxide was only detected as an exsolution phase in the infiltrated copper. Highly pronounced spinel (Mg,Cu)(Fe,Cr,Al)2O4 formation was observed at the hot face and is clearly visible on the image. A small amount of forsterite has formed inside the refractory as a result of reaction between magnesia and silica.
Per – periclase, F - forsterite, Sp- in situ formed spinel phase [(Mg,Fe,Cu)(Fe,Cr,Al)2O4]
Figure 6.6. BSE image of a sample recovered from zone 3
1 mm
113
c) Samples recovered from zone 4 (side wall, below the gas outlet) and zone 5 (side wall, below the burner inlet) BSE images of the samples are presented in Figure 6.7. Combining samples from both zones is possible since they have similar features. The wear and penetration rates of these bricks correspond to the average values, which can be explained with their location in the lining (see Fig. 6.1.). Rich-in-oxygen copper has infiltrated the bricks in zone 4 causing the appearance of metallic copper with cuprous oxide inclusions (Fig. 6.7a and 6.7b). On the other hand, in the samples from zone 5 (located below the burner inlet) only cuprous oxide was found, even in the interior of the brick. Some anode slag (CuOx.FeOy) was observed to have penetrated the brick together with copper (Fig. 6.7c). Spinel formation was also highly pronounced in the bricks recovered from these zones. A small amount of fayalite (2FeO.SiO2) crystals embedded in anode slag were detected in the newly formed spinel matrix (Fig. 6.7d).
a) b)
c) d) Per – periclase (MgO), F - forsterite (2MgO.SiO2), Sp- in situ formed spinel phase
[(Mg,Fe,Cu)(Fe,Cr,Al)2O4], Fay - fayalite (2FeO.SiO2), AS – anode slag (CuOx.FeOy)
Figure 6.7. BSE images of samples recovered from zone 4 (a, b) and zone 5 (c, d)
d) Samples recovered from zone 6 (furnace roof, near the charging mouth) BSE images of the samples are presented in Figure 6.8. The refractories from this zone are only occasionally in contact with copper or slag (tilting of the furnace for slag skimming, splashes during oxidation and reduction steps). Therefore, the wear and penetration rate of the bricks collected from this zone is lower than the average level. The
114
spinel growth in zone 6 is also the poorest of all hot face samples (Fig. 6.8a). Metallic copper droplets can be noticed only now and then, closer to the hot face (Fig. 6.8b).
a)
b)
Per – periclase (MgO), F - forsterite (2MgO.SiO2), Chr – primary chromite
Figure 6.8. BSE images of samples recovered from zone 6
e) Samples recovered from an area close to the tuyeres zone
BSE images of the samples are presented in Figure 6.9. The highest temperature in the anode furnace lining is experienced in this zone. It comes as a result of the close proximity of the bricks to the two tuyeres through which air is blown in the first stage and propane in the second (see Chapter 2, section 2.3). The higher temperature (up to 1350°C) leads to higher temperature change (550-600°C), in comparison with the rest of the lining, when the furnace is empty (i.e. cooler) during waiting periods.
a)
b) F – forsterite, Chr – primary chromite, Sp – in situ formed spinel phase (Mg,Fe,Cu)(Fe,Cr,Al)2O4
Figure 6.9. BSE images of samples recovered from an area close to the tuyeres zone
The heaviest infiltration of all was observed in the bricks recovered from this zone. The samples are abundant with cracks (larger than in samples coming from other zones) presumably due to the higher thermal shocks experienced in this particular area. As a result, even some of the primary chromite grains were attacked and infiltrated by oxygen-bearing copper (Fig. 6.9a). The spinel growth in these samples is most intense. A dense (but not continuous) layer of spinel has formed at the refractory hot face (Fig.6.9b).
115
6.4. Discussion The relative amount of the observed phases as well as the wear and penetration rate of bricks recovered from the different collection zones are summarised in Table 6.3. Table 6.3. Summary of the relative amount of observed phases, wear and penetration rate
of refractories from the collection zones
Relative amount of different phases1 Refractory Location
Per Chr F Cu Slag
Wear rate Penetration rate
Zone 1 ** *** ** *** no *** *** Zone 2 ** *** ** *** no *** *** Zone 3 *** ** *** *** no *** *** Zone 4 *** *** ** ** ** ** *** Zone 5 *** *** ** *** ** ** *** Zone 6 *** ** ** * no * * Tuyeres zone ** *** * **** * **** ****
1 The amount of periclase and primary chromite phases shown in the table is not necessarily representative for the different zones. This is a result of the substantial heterogeneity of the bricks as regards to phase distribution. * - low, ** - medium, *** - high, **** - highest Per – periclase (MgO), F - forsterite (2MgO.SiO2), Chr – primary chromite spinel
Infiltration
The examination of samples recovered from different locations of one of the anode furnaces at Cumerio’s Pirdop Smelter revealed that copper infiltration is considerable in all investigated zones, except zone 6 (furnace roof, near the charging mouth). The heaviest penetration and wear occurred in the tuyeres zone, which is a result of the higher temperature, higher level of turbulence and higher oxygen potential (during the oxidation step) in comparison with the other zones. Owing to the heavy penetration, an infiltrated brick has a higher thermal conductivity, higher density and lower porosity than an as-delivered one. Many of the refractory pores and grain boundaries were filled with copper, copper oxide, copper ferrite (anode slag) and forsterite. Nevertheless, a large number of empty pores and cracks were observed, presumably due to the thermal cycling and the final cooling of the furnace for relining. Forsterite Formation
A very limited amount of silicate phase (CMS) was detected in the as-delivered samples (only 0.4 % SiO2 and 1.5 % CaO in as-delivered brick), while large amounts of silicate phases (mainly forsterite) were found in most worn samples. This suggests that a significant quantity of anode slag has infiltrated the brick microstructure. The silica it contains penetrates the refractory, together with copper, copper oxide and iron oxide. Once inside the brick, silica interacts with magnesia (refractory component) to form forsterite following reaction (6.1)
2[MgO]refr + (SiO2)Slag = [2MgO.SiO2]refr, 'G° = -82 kJ, at 1300 °C (6.1)
116
Guggenite formation
Some guggenite (MgCu2O3) might have formed since MgO shows considerable solid solubility for CuO with a mole fraction of about 0.2 at 1000°C [Ass 1996]. However, guggenite decomposes at temperatures below 870°C following reaction (6.2). Therefore, its possible presence was not detected in the investigated samples. 2CuO.MgO Æ MgO + 2CuO, at T < 870°C (6.2) ‘Breathing’
During the oxidation step of the fire-refining process the oxygen partial pressure (PO2) is comparatively high (around 0.2 atm) while in the reduction step its value is about ten times lower. This big difference in PO2 may cause cycling in the oxidation state of iron in the refractory: Fe2+ �)H3+. The latter phenomenon (known in literature as ‘breathing’) has been investigated by Blanpain et al. [Bla 1997]. The research team performed X-ray Photon Spectroscopy (XPS) measurements to show the possibility of changes in the chemical state of iron in spinel phases of magnesia-chromite refractories. Breathing of the bricks is a result of varying oxygen partial pressure and leads to alternating volume changes (expansion/contraction), which in their turn cause micro-cracking (and possibly spalling) of the bricks [Jon 2000]. However, in the course of the present study it was determined that this degradation mechanism has only minor influence on the anode furnace lining. Spinelisation
The infiltrated slag imports copper and iron oxides that diffuse into the primary and secondary chromite spinel (Mg,Fe)(Cr,Fe,Al)2O4 resulting in the growth of these phases. The consequence of the latter phenomenon is a more complex spinel composition with diluted alumina and chromium content, also involving copper: (Mg,Cu,Fe)(Fe,Cr,Al)2O4. A certain amount of the periclase phase is also incorporated in the formation of this network of growing spinel grains. 6.5. Conclusions
The mapping of the initial industrial situation that was made in the beginning of this work revealed that significant problems exist in anode furnace linings. The configuration of the latter included just one refractory type (medium-quality magnesia-chromite brick) applied in all areas (excluding the tuyere blocks). During the relining of the furnaces worn bricks were collected from various locations in the lining for post-mortem assessment. The latter, together with macroscopical observations, allowed obtaining general knowledge about the degradation phenomena affecting anode furnace linings. The investigated refractory material appeared to be rather susceptible to infiltration by the different liquid phases present in the anode furnaces. Once inside the brick, the major constituents of the melt (FeOx, SiO2 and possibly CuO) react with the refractory components forming new phases (forsterite and possibly guggenite). The penetration process is facilitated by the formation of large cracks, which result from the substantial mechanical, chemical and thermal stresses imposed on the lining. Probably, the slight volume expansion associated with forsterite formation inside the refractories, contributes to the cracking and spalling processes.
117
Different degradation mechanisms were found to influence the refractory performance in anode furnaces. Firstly, the infiltrating copper and slag gradually deplete the intergranular secondary chromite by dissolving most of it, partly replacing in this way, the direct-bonded (solid) structure of the bricks with a liquid-bonded one. The latter phenomenon leads to microscopic dissolution of the refractory components, i.e. they are removed by the fluid motion, grain after grain. The thermal cycling reinforces the above-described degradation mechanism. The erosive fluid, the abrasion of solids and the changing brick load (as a result of the furnace rotation) cause significant mechanical stresses on the refractory lining. From this it becomes clear that further work is required in the direction of designing alternative lining configuration that would perform better than the original one. Two options for alternative linings were determined: (i) zoned concept, including different qualities of magnesia-chromite bricks and (ii) zoned and mixed concept, including alternative chrome-free brick types. Unfortunately, it was not possible to perform industrial tests because these would be too risky and costly. Besides, they appeared to have a ‘low learning speed’ (relining of the furnaces is done only once a year, so a very limited number of experiments would be possible). Therefore, it was decided to carry out laboratory tests, which are described in the next chapter.
118
CHAPTER 7
LABORATORY TESTS WITH MAGNESIA-CHROMITE AND CHROME-FREE REFRACTORY BRICKS
7.1. Introduction Since industrial tests would be too risky and costly to carry out and since they have a ‘low learning speed’ (samples can be collected only once a year), it was decided to perform laboratory tests, which are described in the present chapter. These experiments provided the opportunity to test different magnesia-chromite qualities as well as alternative chrome-free refractory materials at typical fire-refining conditions. As a result of the widespread adoption of mag-chrome refractories, nearly all the research in the field of copper production furnace linings over the last three decades has focused on this class of material. However, recent environmental initiatives are providing a driving force for the development of new refractory types. The specific concern lies in the presence of chromium in mag-chrome bricks, which implies the possibility of hexavalent chromium formation (Cr6+) in spent refractories [Mar 1993a]. Chromium has multiple oxidation states (2+, 3+ and 6+). Among these, Cr6+ is the only chromium ion soluble in water. This ion can give rise to detrimental effects on the environment and the food chain because it has a strong oxidizing power and may easily penetrate human tissues [Lee 1997]. Furthermore, Cr6+ is known to be carcinogenic [Lee 1997, Lee 1998]. Used magnesia-chromite refractories are classified as “solid hazardous waste” when they contain over 5 ppm of Cr6+. As a result, regulations treating spent mag-chrome bricks as potentially hazardous waste are being enforced, making their disposal difficult and expensive. This has led to the replacement of mag-chrome bricks in other industries particularly in cement kilns and steelmaking refining vessels [Tsu 1995, Gri 1996]. The success of such initiatives may in turn eventually make mag-chrome refractories more difficult to obtain as well as more difficult to dispose of for copper producers, thus providing further impetus for finding replacement materials [Cri 2000]. In response to this incentive, Schlesinger et al. [Sch 1998] aimed at assessing the viability of chrome-free refractory materials, including spinel-containing (MgAl2O4) ones, as potential replacements for mag-chrome bricks by carrying out static finger tests in copper-containing calcium ferrite slags. The latter are generated in the newer smelting processes, such as the Mitsubishi continuous smelting and converting process. Another set of static [Sch 1997], as well as dynamic tests [Cri 1999], were performed by the same research team using a fayalite type of slag. As a result of these experiments, the researchers established that none of the tested chrome-free alternatives could match the performance of the mag-chrome reference brick. The conclusion was drawn that the real advantage of mag-chrome refractories in copper production furnaces stems from their higher corrosion resistance resulting from the direct-bonding of the periclase and chromia spinel grains. It was also concluded that magnesia-MgAl2O4 spinel refractories with a similar direct-bonding possess the highest potential as replacement materials. The experiments also highlighted the importance of applying different refractory classes according to the specific melts and conditions characteristic for each furnace type.
119
However, our literature survey could not find any data about refractory tests conducted with the type of slag generated during the fire-refining process – the anode slag. This slag differs from the flash furnace and converter slags by its very high basicity – it typically contains above 50 wt% CuOx, 30-35 wt% FeOx, 7-8 wt% SiO2 and minor amounts of As, Sb, Pb. With the intention to simulate the melt that exists in an industrial anode furnace, corrosion tests with magnesia-chromite and chrome-free refractory bricks exposed to copper and anode slag at a typical fire-refining temperature were conducted. These lab scale experiments focus on one of the major degradation causes – refractory corrosion. 7.2. Static finger tests Various methods have been used to simulate the environment a refractory brick is exposed to during commercial service. Each has benefits and limitations and they can be classified as either static or dynamic. In the former no attempt is made to simulate motion of the corroding liquid while in the latter the liquid is made to move relative to the refractory [Lee 2004]. One of the simplest and easy to conduct static methods is the finger test (Figure 7.1). In this experiment one or more cylindrical or square-pillar shaped refractory samples are partially submerged in the corrosive medium. This type of test is normally performed in a resistance or induction furnace. Benefits of the method include: the atmosphere is easily controlled and the slag composition change (associated with rapid saturation with reaction products) can be minimised by using a large amount of slag. An important drawback is that no temperature gradient exists in the refractory samples i.e. they are isothermal, which is not the case in a commercial furnace where a temperature gradient between the hot face and the cold face of the working lining exists. Despite this drawback, static finger tests were used in the present work as they allow to easily obtain important knowledge about refractory penetration and corrosion resistance. However, it must be stressed that the outcome from these tests should be seen as being complementary to more complex laboratory experiments and industrial trials.
Figure 7.1. Static finger test set-up 7.2.1. Materials and experimental procedure Twelve types of commercially available refractory bricks were tested in the present investigation. Six of them are chrome-free types (A – F) and the other six are mag-chrome based (G - L). Table 7.1 presents the composition and selected properties of the tested refractory products. Four of the chrome-free brick types (A, B, D and E) are based on the magnesia-alumina system, which resembles the magnesia-chrome system in terms
sample
furnace
slag
crucible
120
of chemistry and phase equilibria (see Figure 4.9, Chapter 4). Magnesia-alumina bricks have been successfully used as a replacement for mag-chrome in several applications [Dal 1988, Wil 1993]. Types A, B and D are periclase-based refractories with magnesia-alumina (MA) spinel, whereas type E is corundum-based with MA spinel. Two of the chrome-free bricks have a zirconia addition (types D and F) while type C is with added hercynite spinel grain. Type B is typically used for lining cement rotary kilns but can also be used in the steel industry [Guo 2005, Lam 2004]. Type C was developed specifically for application in the burning section of precalciner kilns [Nie 1999]. Type D is mostly applied in the furnaces used in cement and lime production, whereas the main application area of type E is the high-wear lining of steel ladles [Data Ref]. Finally, type F is specifically produced for application in furnaces for waste incineration and regenerators of glass melting tanks [Data Ref]. The mag-chrome refractory types can be subdivided into two classes: class I – conventional, high-fired (at 1650°C-1700°C) direct-bonded bricks, typical of those used in current copper-making furnaces (types G, H and I) and class II – the latest generation of direct-bonded bricks, based on low-in-iron presintered grain, featuring a ceramic bond ( types J, K and L).
Table 7.1. Composition and properties of the tested brick types
The experiments were conducted in the following way: an alumina crucible was filled with 1800 g of electro-refined copper (99.99% Cu) and 500 g of anode slag and then heated in a laboratory resistance furnace (custom made with SiC heating elements) until a temperature of 1300°C was reached. An industrial anode slag provided by Cumerio Med with the following composition was used for the experiments (in wt%): 55 CuOx, 35 FeO, 8 SiO2 and 2 (Pb + Zn + Ni + As + Sb). Four refractory fingers were simultaneously submerged in the molten content of the crucible, as shown in Figure 7.2. Types A, B, C and D were tested in the first crucible, types E, F, G and H in the second crucible and types I, J, K and L in the third one. The dimensions of an as-delivered finger are 2 x 2 x 20 cm. The fingers were held in the melt for 24 hours at a constant temperature of 1300°C (± 5°C). Having in mind the limited duration, the square-pillar shape was chosen for the fingers as it is more prone to corrosion and hot erosion than the cylindrical one.
121
After completion of the tests, the fingers were withdrawn from the crucible and exposed to ambient air, which resulted in a relatively rapid cooling rate.
Figure 7.2. Laboratory set-up (view from above) It must be noted that some oxidation of the copper phase occurred during the experiments because copper and slag were melted together and because the furnace was not sealed. This was done on purpose in order to simulate the actual melt present in an industrial anode furnace. The oxygen content of copper during fire-refining varies between 1 at% O2 (beginning of the process) and 0.1 at% O2 (end of the process). Therefore, whenever in this chapter the word ‘copper’ is mentioned, actually ‘oxidised copper’ is meant. This clarification is necessary since it is difficult for pure copper (99.99% Cu) to penetrate the brick microstructure, because of the large wetting angle to periclase (approximately 140° at 1200°C) [Bar 1981]. However, contaminants (above all, oxygen) can reduce this angle drastically and thus facilitate infiltration. For example, at a temperature of 1200°C, the wetting angle between copper and periclase is reduced to less than 90° by an oxygen content of just a few tenths of a percent [Tik 1978]. The latter creates favourable wetting conditions and copper penetration is easy and rapid. 7.2.2. Sample preparation and analysis technique Three specimens were cut with a diamond saw from each refractory finger (Figure 7.3). Specimen 1 was recovered from the ‘as-delivered’ finger, while specimens 2 and 3 were extracted from the tested finger (from the slag and copper zone, respectively). Afterwards, the specimens were embedded in a low-viscosity resin (Technovit 4004), ground with diamond plates and polished with diamond suspensions. Finally, carbon was evaporated on their surface to provide a conducting layer for further examination. A JEOL JXA-733 microprobe coupled with an Energy Dispersive Spectroscopy (EDS) system was used for semi-quantitative analysis. A high resolution Scanning Electron Microscope (Philips XL-30 FEG), equipped with an EDS detector system with an ultra thin window, was also used for analysing most of the samples and for acquiring high-quality back-scattered electron (BSE) images.
122
1 – specimen recovered from the ‘as-delivered’ finger, 2 – specimen extracted from the tested finger (slag zone), 3 – specimen extracted from the tested finger (copper zone)
Figure 7.3. As-delivered and tested finger
7.2.3. Results of SEM and EPMA examination Due to their poor corrosion resistance, two of the fingers were not tested for the full experimental time (24 h) and are, therefore, not considered for detailed discussion here. The refractory finger type F (magnesia-based brick made of sintered MgO with ZrSiO4 addition) was removed from the crucible after six hours of testing. This was done to prevent the finger from breaking in two, since the area between the copper and slag layer was highly corroded (Figure 7.4). The other refractory type, which proved to be totally inappropriate for the copper/anode slag environment is type E (alumina-based brick made of fused corundum and magnesia-alumina spinel). About half of this finger was dissolved by the melt after only 4 hours of testing (Figure 7.5). This dissolution (together with dissolution of alumina from the crucible wall) somewhat affected the slag composition by increasing its Al2O3 content to ~ 5 wt%.
Slag zone Copper zone
1 2
3
2
1 3
123
Figure 7.4. As-delivered and tested refractory finger – type F
Figure 7.5. As-delivered and tested
refractory finger – type E
a) Magnesia-chromite bricks
According to the classification of the investigated mag-chrome types made earlier, type H was selected from the first class and type K from the second for a detailed presentation. These two types are considered representative for their classes and are of considerable interest for the copper industry. Type H Figure 7.6 presents the microstructure of the “as-delivered” refractory type H. This is a high-fired direct-bonded mag-chrome brick typical of those used in current copper-making furnaces. This refractory material consists of periclase (almost pure MgO), presintered magnesia-chromite grains, primary and secondary chromite spinel (Mg,Fe)(Cr,Al,Fe)2O4 and, to a much lesser extent, calcia-magnesia silicate impurities.
Figure 7.6. Overview image of the microstructure of as-delivered refractory type H
The as-delivered microstructure of this refractory type is characterised by a high proportion of direct bonding between periclase, chromite spinel (primary and secondary) and magnesia-chromite grains, i.e. as Goto and Lee defined it: “a direct attachment without any intermediate silicate film” [Got 1995]. Two types of secondary chromite spinel can be distinguished with respect to their formation mechanism (Figure 7.7). The first type (intergranular chromite spinel) is formed by intergranular precipitation at the periclase grain boundaries. It can be exsolved out of periclase or crystallised from spinel-rich liquid upon cooling, after completion of the firing process. This type of secondary chromite is not only of great significance for the direct bonding but also for the protection against liquid penetration [Jon 2002]. The second type (intragranular chromite spinel) arises from intragranular exsolution precipitation from periclase grains on cooling and thus forms within the grains [Got 1995]. Figure 7.8 shows the post-exposure microstructure of refractory type H (sample from the slag zone). A layer of frozen melt has formed at the refractory hot face (lower part of the image). It is built up of grey angular grains (a complex spinel phase shown in Figure 7.9) and copper oxide. The complex spinel was formed as a result of reaction between slag (Al2O3, FeO, CuOx) and refractory (MgO, Cr2O3) components according to (see Table 7.2 for composition): [MgO]refr + [Cr2O3]refr + (Al2O3)sl + (FeOx)sl + (CuOx)sl = = (Mg,Cu,Fe2+)(Al,Cr,Fe3+)2O4 (7.1) A detailed examination of the sample revealed that copper and copper oxide have completely infiltrated the refractory matrix.
1mm
Chr
Mg-Cr
Per
125
Chr – primary chromite spinel, Inter – intergranular secondary chromite spinel, Intra – intragranular secondary chromite spinel
Figure 7.7. Detailed image of a presintered grain showing the abundance of inter- and
intragranular secondary chromite (as-delivered type H) Hybrid microwave experiments To study the actual high-temperature (1300-1800°C) behaviour of industrially worn magnesia-chromite refractories, Jones and co-workers performed hybrid microwave heating experiments [Jon 2001]. These experiments allowed overcoming certain evaluation difficulties encountered when investigating industrially worn magnesia-chromite samples that were slowly cooled. Compared to conventional heating, hybrid microwave heating clearly showed advantages with respect to warm-up time, energy consumption and quenching possibilities. Especially the latter is of great importance for studying the actual high-temperature microstructure. The results of these tests show that at a typical copper-refining temperature (1300°C) all refractory phases are entirely solid, i.e. the original brick microstructure remains unchanged.
Forsterite – element (at %) (interior) oxide (wt%)
26.1±0.5 51±1
- -
- -
- -
- -
17.1±0.5 49±1
- -
Microprobe analyses performed on the periclase grains near the hot face (Figure 7.9) indicate that a considerable amount of FeO (20 wt%) and CuOx (5 wt%) diffused into it, as compared to periclase grains in the as-delivered brick, which typically contain 98-99 % MgO. These analyses confirmed our expectations based on the phase diagram FeOx-MgO (see Figure 4.10c). The FeO content of periclase gradually decreases when going deeper into the refractory, analysing 11 wt% in the centre of the sample (Figure 7.10), while the amount of CuOx in periclase remains constant (5 wt%) regardless of the grain location. The centre of a primary chromite grain located at the refractory hot face as well as its rim were analysed (Figure 7.9). The results show that a large amount of FeO diffused in the rim reaching 36 wt%, while its quantity in the centre of the grain was only 12 wt%, which corresponds to the average amount of FeO in primary chromite spinel in the as-delivered brick [Pet 2005]. A similar situation is observed for another slag component – Al2O3 – 41 wt% in the rim and 16 wt% in the centre of the grain. Because of the diffusion of slag components in the rim, the amount of Cr2O3 (major ingredient of primary chromite spinel) has decreased to 3 wt%, while in the centre of the grain it analyses 51 wt%. It must be noted that CuOx diffusion did not stop at the rim (4 wt%) but reached the centre of the grain (3 wt% CuOx). The amount of MgO in the hot face primary chromite spinel remains constant (16 wt%) across the whole surface of the grain.
128
Mg-Cr – presintered magnesia-chromite grain, Chr – primary chromite spinel, Per – periclase, F – forsterite, CMS – calcia-magnesia silicate
Figure 7.10. High magnification image of the interior of the tested refractory type H
(slag zone)
Mg-Cr – presintered magnesia-chromite grain, Chr – primary chromite spinel, Per – periclase, Inter – intergranular secondary chromite spinel, Intra – intragranular secondary chromite spinel
Figure 7.11. Detailed image of the microstructure of the tested refractory type H (deeper
in the brick – slag zone)
0.05 mm
Chr
CMS
F
pore CuOx
Mg-Cr
Per
0.2 mm
Chr
Per
Intra
Inter Mg-Cr
129
The intergranular and the intragranular secondary chromite spinels were analysed in the refractory interior (Figure 7.11). As expected, the amount of slag components (CuOx, FeO and Al2O3) in the first type of secondary spinel is higher than in the second one. The explanation of this phenomenon stems from the fact that infiltration along grain boundaries is much easier than diffusion in grains (see Table 7.2). Another remarkable feature is that a forsterite phase (2MgO.SiO2) has formed in the brick as a result of reaction between SiO2 (slag component) and MgO (refractory component): 2[MgO]refr + (SiO2)slag = [2MgO.SiO2]refr (7.2) Analyses performed on the newly formed phase at the hot face (Figure 7.9) show that it contains some dissolved FeO (4 wt%). When going deeper in the brick, the amount of penetrated FeO decreases and, therefore, its amount in forsterite also declines (to 2 wt%). The depleting FeO is being replaced by CaO impurities present intergranularly in the refractory microstructure. Analysis performed on the forsterite phase in the centre of the sample (Figure 7.10) show that the latter contains only 1 wt% FeO and already 4 wt% CaO. The SiO2:MgO ratio in forsterite remains constant in all analysed locations. The copper oxide phase at the hot face (Figure 7.9) contains a certain amount of dissolved slag and refractory components (see Table 7.2). Analyses performed deeper in the brick (Figure 7.10) reveal that apart from FeO, the level of impurities in CuOx gradually diminishes. The copper oxide phase in the centre of the sample (Figure 7.10) contains only 3 wt% FeO and no other species. This seems to suggest that CuOx and FeO penetrate the brick microstructure together. A calcia-magnesia silicate (CMS) phase (impurity in the raw materials for brick making) was detected in the interior of the sample (Figure 7.10 and Table 7.2). Type K Figure 7.12 presents the microstructure of the as-delivered refractory type K. As mentioned earlier, the latter belongs to the latest generation of direct-bonded bricks, based on low-in-iron presintered grain. According to the data sheet of the refractory producer, these bricks were specifically developed for application in the non-ferrous metallurgy and possess a very high corrosion and thermal shock resistance. As can be seen in Figure 7.12, the refractory microstructure is very uniform. A close examination shows that there is a remarkably uniform pore distribution throughout the whole sample. Primary chromite grains are few and small in size. Large presintered magnesia-chromite grains with very fine to fine homogeneous dispersion of secondary chromite can be observed. There is little precipitation of chromite along subgrains and a very good contact between the few large magnesia-chromite grains and the refractory matrix (large in number smaller grains) exists.
Figure 7.12. Overview image of the microstructure of as-delivered refractory type K Figure 7.13 shows the post-exposure microstructure of refractory K – sample from the bottom of the finger (copper zone). It must be noted that this sample was not completely infiltrated (the tiny white dots on the image are infiltrated copper). The limit of copper penetration is indicated in Figure 7.13 with a dashed line and is located at a distance of approximately 3-4 mm from the refractory hot face. The penetration resistance of the other two brick types belonging to this class (J & L) was also as high. The sample recovered from the slag zone is, however, completely infiltrated by CuOx (Figure 7.14). As in the previous mag-chrome sample (type H), a layer of melt has formed at the hot face. The grey angular crystals (1), better visible in the detailed image (Figure 7.15), are built up of the same ingredients as the ones found in the complex spinel phase of type H, only their ratio is different (see Table 7.2). Forsterite (2MgO.SiO2) has also formed in the brick. Analysis performed at the refractory hot face (Figure 7.15) on the newly formed forsterite phase show that (like in type H) it contains a little dissolved FeO (3 wt%). Microprobe analyses performed on periclase grains near the hot face indicate that significant amounts of FeO (15 wt%) and CuOx (4 wt%) have diffused into the grains. Similar to the previous sample (type H), the FeO content gradually falls when going deeper into the refractory, analysing 8 wt% in the centre of the sample, while the amount of CuOx in periclase remains practically constant (4 wt%) regardless of the grain location.
1mm
Mg-Cr Chr
131
Figure 7.13. Overview image of the microstructure of the tested refractory type K (copper zone).
Figure 7.14. Overview image of the microstructure of the tested refractory type K (slag zone)
1mm
Penetration limit
Brick Melt
1mm
132
1 – complex spinel phase, F – forsterite
Figure 7.15. Detailed image of the hot face of the tested refractory type K (slag zone) Analyses carried out at the hot face show that the amount of infiltrated CuOx in intergranular and intragranular secondary chromite phases is negligible (below 1 wt%). The compositions of the two types of secondary chromite are practically the same (see Table 7.2). A calcia-magnesia silicate phase, very close in stoichiometry to CaO.MgO.SiO2 (CMS) was detected in the interior of the sample (Figure 7.16). Its detection confirmed the presence of a certain amount of impurities in the raw materials for brick making.
Figure 7.16. Detailed image of the interior of the tested refractory type K (slag zone)
F
CuOx
1 0.1 mm
1 2
3 Per
0.1 mm
133
b) Chrome-free bricks
Type A Figure 7.17 presents an overview of the microstructure of the pre-exposed refractory type A. This is a burnt magnesia-spinel brick made from fused magnesia (MgO) and spinel with a complex composition (Mg,Fe)(Al,Fe)2O4. Magnesia is present as grains with varying dimensions as well as finely ground particles (matrix). A detailed look at the large periclase grains shows that they are comprised of small ‘subgrains’ with calcia-silicate (CaSiO3) impurities located along their grain boundaries. Discrete spinel grains (Table 7.3), small in number and with varying dimensions (typically between 0.5 and 1 mm) are scattered around the matrix of the brick.
Figure 7.17. Overview image of the microstructure of as-delivered refractory type A An overview of the post-exposure microstructure of refractory A is presented in Figure 7.18. Copper and slag products have completely infiltrated the periclase matrix as well as the spinel grains (Figure 7.19). The infiltrated slag imports copper oxide, iron oxide, silica and alumina (the latter comes from external sources – dissolved from the crucible wall and from the high-alumina finger). FeOx and CuOx diffused into the spinel grains, thus changing their composition (Table 7.3). FeOx and Al2O3 reacted with MgO following reaction (7.3), as a result of which a complex spinel phase was formed at the hot face (see Figure 7.20b and Table 7.3). [MgO]refr + [Al2O3]slag + [FeOx]slag = [(Mg,Fe2+)Al2O4]refr (7.3) Up to 3 wt% of CuOx dissolved in this newly formed phase (see Table 7.3). The infiltrated silica reacts with magnesia following reaction (7.2) to form another new phase – forsterite (Figure 7.20b):
1mm
134
Figure 7.18. Overview image of the post-exposure microstructure of refractory type A (slag zone)
Up to 4 wt% of FeOx can dissolve in the “hot face” forsterite, while in forsterite formed in the interior of the finger, FeOx is replaced by CaO ( up to 3 wt% CaO). In this way the calcia silicate impurities present in the refractory are incorporated in forsterite formation.
Figure 7.19. Detailed image of the post-exposure microstructure of refractory type A (slag zone)
It can be stated that the large magnesia grains have withstood the attack of the liquid phases, apart from some small cracks in the grains that were infiltrated (Figure 7.18). Nevertheless, periclase (even in the interior of the finger) contains a certain amount of diffused iron oxide and copper oxide (4 wt% of each phase – Table 7.3). As the large cracks among periclase grains (Figure 7.18) are not infiltrated, we suggest that they were
1mm
MgO
spinel
0.2 mm
MgO
spinel grain CuOx
135
formed during the rapid cooling of the fingers, when the latter were removed from the crucible and exposed to ambient air.
a) Detailed image of the hot face of the exposed refractory type A (slag zone)
1 – complex spinel phase, F – forsterite
b) High-magnification image of the slag/refractory interface of the exposed refractory type A (slag zone)
Figure 7.20. Detailed images of the microstructure of exposed refractory type A (slag
zone)
0.5 mm
0.1 mm
136
Table 7.3. Compositions of various phases encountered in the tested chrome-free refractory types
Phases Element Oxide
Mg MgO
Fe FeO
Al Al2O3
Cr Cr2O3
Ca CaO
Si SiO2
Cu CuOx
Type A Spinel phase – element (at %) (as-delivered) oxide (wt%)
12.2±0.4 24.8±0.5
8.1±0.3 28.7±0.5
21.3±0.5 47±1
- -
- -
- -
- -
Spinel phase (rim) – element (at %) (interior) oxide (wt%)
9.2±0.3 12.2±0.4
28.1±0.5 65±2
8.8±0.3 13.5±0.4
- -
- -
- -
4.1±0.2 9.2±0.3
Spinel phase (centre) – element (at %) (interior) oxide (wt%)
9.2±0.3 12.8±0.4
21.3±0.5 56±1
14.1±0.4 27.1±0.5
- -
- -
- -
1.1±0.2 4.1±0.2
Complex spinel – element (at %) (hot face) oxide (wt%)
11.9±0.4 20.7±0.5
14.1±0.4 44±1
15.6±0.5 35±1
- -
- -
- -
- -
Periclase – element (at %) (interior) oxide (wt%)
48±1 92±2
1.1±0.2 4.2±0.2
- -
- -
- -
- -
1.2±0.2 4.6±0.2
Type B Complex spinel – element (at %) (hot face) oxide (wt%)
11.1±0.4 20.6±0.5
15.4±0.5 45±1
14.1±0.4 34±1
- -
- -
- -
- -
Forsterite – element (at %) (hot face) oxide (wt%)
26.3±0.5 51±1
- -
- -
- -
- -
17.9±0.5 49±1
- -
Type C Periclase – element (at %) (interior) oxide (wt%)
42±1 74±2
6.1±0.3 18.7±0.5
1.5±0.2 3.7±0.2
- -
- -
- -
2.2±0.2 5.7±0.3
Type D Periclase (centre) – element (at %) (interior) oxide (wt%)
44±1 100±2
- -
- -
- -
- -
- -
- -
Periclase (rim) – element (at %) (interior) oxide (wt%)
Forsterite – element (at %) (interior) oxide (wt%)
26.4±0.5 51±1
- -
- -
- -
- -
17.2±0.5 49±1
- -
Type B Figure 7.21 shows the microstructure of the unexposed refractory type B. This is a burnt magnesia-spinel brick made from sintered magnesia (MgO) and prefabricated spinel (MgAl2O4). Magnesia is in the form of large grains (but smaller than in the previous refractory type) and finely ground phase (matrix), while the spinel phase is present as big grains only (lighter in colour – BSE mode). In Figure 7.22 an overview of the microstructure of the exposed finger is presented. Similarly to the previous sample, copper has easily penetrated into the refractory matrix. Some of the spinel grains were also infiltrated (Figure 7.23a). The large periclase grains, however, were only partially infiltrated thanks to the intragranular presence of non-stoichiometric calcia-magnesia silicate impurities, with a chemical formula very close to 3CaO.MgO.2SiO2. The latter compound has a melting point of 1438°C [Sla 1995], which means that it was entirely solid at the test temperature of 1300°C and therefore acted as a barrier against copper penetration in MgO grains (Figure 7.22 and 8.23b). It is noteworthy that the analyses performed on the sample recovered from the copper zone show no diffusion of copper into periclase and spinel grains. Likewise, analyses carried out on the sample recovered from the slag zone show no diffusion of CuOx and FeOx neither in periclase nor in spinel grains.
137
Figure 7.21. Overview image of the microstructure of as-delivered refractory type B
Figure 7.22. An overview image of the microstructure of the tested refractory type B (copper zone)
MgO spinel
1mm
1mm
MgO
spinel
138
a) Spinel grains were easily infiltrated by the copper
b) Calcia-magnesia silicate hinders copper infiltration along grain boundaries
Figure 7.23. Detailed images of the microstructure of tested refractory B (copper zone) At the hot face of the sample recovered from the slag zone, new phases (complex spinel and forsterite – Figure 7.24 and 7.25, Table 7.3) have formed as a result of reactions between slag and brick components.
MgO
MgO
spinel
0.1 mm
0.2 mm
139
Figure 7.24. Overview image of the microstructure of tested refractory B (slag zone)
Figure 7.25. Detailed image of the microstructure of tested refractory B (slag zone) Type C Figure 7.26 illustates the microstructure of the “as-delivered” refractory type C. This is a burnt magnesia-spinel brick made of sintered magnesia (MgO) and hercynite spinel (FeAl2O4). Magnesia is in the form of large grains and fine matrix, while hercynite is only present as grains with varying dimensions. The brick exhibits medium levels of porosity, density and crushing strength among the tested bricks (Table 7.1). Figure 7.27 shows the result of the exposure to liquid slag of refractory type C. The same reaction that occurred at the hot face of sample A, has taken place on the surface of the present refractory finger, i.e. FeOx and Al2O3 (from the slag) reacted with MgO to form (Mg,Fe)O.Al2O3. The latter phase can be observed as grey crystals in the crust formed at the refractory hot face (Figure 7.27). Forsterite formation has also occurred in this
MgO
1
Hot face
0.5 mm
1
F
Hot face
CuOx
F
0.1 mm
140
refractory type. Similar to the other brick types, forsterite formed at the hot face contains some FeOx (3 wt%), while for forsterite located in the refractory interior, FeOx was replaced by CaO. The periclase matrix is completely infiltrated by copper oxide (white dots on the image): a global analysis of the sample interior shows more than 6 wt% CuOx. The liquid phase has easily penetrated into the hercynite (spinel) grains as well as into the large magnesia grains. Only a small portion of the magnesia grain boundaries remained protected by the calcia silicate impurities present intragranularly thanks to the high melting point of CaO.SiO2 – 1544°C [Sla 1995]. However, copper and iron oxides diffused deeply in periclase grains. Analyses show virtually the same amount of FeOx (18 wt%) and CuOx (5 wt%) at the rim and inside the grains, which suggests that periclase is highly prone to diffusion by these two oxides.
Figure 7.26. Overview image of the microstructure of as-delivered refractory type C
Figure 7.27. An overview image of the microstructure of the tested type C (slag zone)
1mm
1mm
MgO
hercynite
Hot face
141
Type D Figure 7.28 illustrates the microstructure of the “as-delivered” refractory type D. This is a burnt special magnesia brick made from sintered magnesia (MgO), spinel (MgAl2O4) and calcium zirconate (CaZrO3) additive. Periclase grains in this refractory type are not so clearly pronounced as in the above-described samples. Instead, a great amount of magnesia is present in the form of a fine matrix. The CaZrO3 phase (tiny white spots on the image) is evenly distributed throughout the whole surface of the sample. The refractory producer claims that the formation of a second bonding phase by addition of fine powder (calcium zirconate – in this case) improves the chemical resistance of the brick.
Figure 7.28. Overview image of the microstructure of pre-exposed refractory type D Figure 7.29 shows the post-exposure microstructure of refractory D (slag zone). The sample was completely infiltrated by copper oxide (white phase on the image). Neither the periclase matrix nor the spinel grains could withstand the attack of the liquid phase. Only the large periclase grains were not thoroughly penetrated. However, analyses showed that iron oxide (up to 9 wt%) diffused in their rim (lighter in colour than the interior). Figure 7.29 also shows a well-known common weakness of magnesia-based refractories, namely – poor penetration resistance along grain boundaries [Cla 1959]. The presence of a second bonding phase (CaZrO3) in this brick did not change the above-mentioned pattern, which seems logical since this phase is only present in discrete locations; it does not form a continuous network throughout the refractory microstructure (unlike the secondary chromite spinel in mag-chrome bricks).
MgAl2O4
MgO
1mm
142
Figure 7.29. Overview image of the microstructure of tested type D (slag zone) Microprobe analyses reveal that the white phase surrounding the spinel grain shown in Figure 7.30 is CuOx, while the white phase inside the grain is CaZrO3. Apparently the two phases (CuOx and CaZrO3) are very similar in weighted average atomic number since it proved impossible to distinguish them (both phases exhibit the same intensity in BSE mode). The analyses also showed that the tiny white dots in periclase grains (for example, the ones in the upper left corner of the image) are CaZrO3. The grey phase inside and around the spinel grain is actually the complex spinel phase formed at the hot face, which penetrated the refractory microstructure together with copper oxide (see Table 7.3). At high magnification (830x) forsterite phase (F) and CuOx precipitates (present in periclase grains as thin white strips) were also observed in this refractory type (Figure 7.31).
Figure 7.30. Detailed image of the post-exposure microstructure of refractory type D (slag zone)
MgO
MgAl2O4
1mm
MgO
Spinel
CaZrO3
CuOx 1
0.2 mm
1
143
Figure 7.31. High magnification image of the post-exposure microstructure of refractory type D (slag zone)
7.2.4. Wear mechanisms The overall refractory wear, the depth of infiltration and the level of corrosive reactions for the different refractory classes are summarized in Table 7.4.
Table 7.4. Summary of the level of wear, infiltration depth and new phase formation for the tested brick types
* - low, ** - medium, *** - high level Overall wear rate Bearing in mind the limited duration (24 h) and the static nature of the finger tests (no motion of the liquid), a low to very low level of refractory wear could be expected. Indeed, the overall wear rate (i.e. the erosion of the refractory surface) was very low, almost zero, with only two exceptions – types E and F. As mentioned earlier, type F – magnesia brick with zirconia content (13.5 % ZrO2) – and especially type E – alumina-based brick (95 % Al2O3) – appeared to be very prone to dissolution in the liquid copper and anode slag. The latter can be explained by the high basicity of the anode slag (55 wt% CuOx) being in contact with an acidic refractory.
CuOx
MgO
F
pore
F
0.02 mm
144
Penetration resistance Apart from the samples recovered from the copper zone of the latest generation of direct-bonded bricks – types J, K and L – all others were completely infiltrated by copper and slag components (copper oxide, iron oxide, alumina and silica). However, the amount of infiltrated liquid in the chrome-free types was higher than the infiltration level in mag-chrome bricks. An explanation of this phenomenon follows. Considering a single phase periclase brick being infiltrated along its grain boundaries by a liquid phase, the following state of balance arises when equilibrium is reached:
where per-per is the grain boundary surface energy ,� per-liq is the periclase/liquid interfacial energy, and per-per is the dihedral angle of the two-phase boundaries formed at the intersection with the grain boundary. per-per is thus the angle formed in the neck between neighbouring periclase grains by the liquid phase (Figure 7.32).
Figure 7.32. Schematic figure of liquid infiltration between two periclase grains The dihedral angle is a measure of the ability of the liquid to penetrate in between the grains. No penetration occurs when per-per is larger than 120°, whereas low dihedral angles (< 60°) result in strong infiltration. Complete penetration occurs when per-per is zero. In this case:
liqperperper �� � JJ 2 (7.5) The situation for a refractory brick consisting of two solid phases is more complicated, although similar principles apply. In this case there are three different types of dihedral angles. For a magnesia-chromite brick, which consists of periclase and spinel grains, these are: per-per, per-sp and sp-sp. Stephenson and White [Ste 1967] have corroborated that the penetration of a liquid phase between a periclase and a spinel grain is smaller than between two periclase grains and that this behaviour is due to the interfacial energy, which is lower between a periclase and a spinel grain than between two periclase grains. This means that the dihedral angle between two periclase grains ( per-per) is considerably smaller than the dihedral angle between a periclase and a spinel grain ( per-sp). In the chrome-free bricks tested in this study, the secondary phase (spinel) is present as discrete grains, which do not form a continuous network throughout the refractory microstructure. Besides, the amount of spinel phase is very limited: from ~ 4 wt% (type D) to ~ 15 wt% (type B). Therefore, these refractories can be practically considered as
Liquid phase
Periclase
Periclase
145
single-phase (periclase) bricks, which can be easily infiltrated by Cu/CuOx as they do not have a meaningful amount of periclase/spinel contacts. On the other hand, in mag-chrome bricks the secondary chromite spinel phase is evenly distributed throughout the refractory microstructure, enveloping magnesia grains (intergranular spinel) and precipitating inside them (intragranular spinel). This particular distribution results in the formation of a continuous spinel network, which renders the dihedral angle between periclase and chromite spinel ( per-sp) of primary importance. Furthermore, the amount of spinel (primary and secondary) in the tested mag-chrome refractories is considerable: 44-45 wt%. Since per-sp is larger than per-per, Cu/CuOx penetration in mag-chrome bricks is more difficult than in the chrome-free ones and, consequently, the amount of infiltrated liquid in mag-chrome refractories is less. Among all bricks tested, the mag-chrome types J, K and L showed the highest penetration resistance. When discussing the performance of these mag-chrome refractory types, it must be stressed that the samples recovered from the slag zone were thoroughly infiltrated by CuOx, while the ones recovered from the copper zone were only partially infiltrated by Cu (the centres of the samples were free of infiltration). The explanation of this phenomenon stems from the fact that the wetting angle between copper oxide and periclase is much smaller (about 15° at 1200°C) [Bar 1981] than the one between copper and periclase (approximately 140° at 1200°C) [Bar 1981] and it is, therefore, much easier for CuOx to penetrate the brick microstructure. New phase formation In all refractory types, new phases were formed as a result of slag-refractory interactions. The infiltrating slag brought copper oxide, iron oxide, alumina and silica to the brick. Iron oxide and alumina react with magnesia according to reaction (8.3). As a result of this interaction a complex spinel phase is formed at the hot face. Its composition is not constant; it varies in the different refractory types (see Tables II and III). Another slag component (silica) reacts with magnesia following reaction (8.2) to form forsterite. The infiltrated copper does not interact with any refractory components and, therefore, penetrates deeply into the brick. The impact of the formation of these new phases on the refractory performance is difficult to assess. The complex spinel – (Mg,Fe)Al2O4 – was detected only at the hot face while forsterite was analysed throughout the whole refractory. Although both phases can be found at the refractory hot face, they did not manage to protect the brick from infiltration because the latter preceded their formation. To assess the impact of forsterite formation throughout the refractory, estimations can be made with respect to the volume expansion/contraction accompanying this formation. Based on the reaction of forsterite formation: 2 moles [MgO]refr + 1 mole (SiO2)slag = 1 mole [Mg2SiO4]refr (7.6) the volume occupied by 2 moles of MgO was calculated (22.4 cm3) and compared with the volume occupied by the reaction product – 1 mole of Mg2SiO4 (43.4 cm3). The density values of the three compounds are, as follows [g/cm3]: MgO = 3.56-3.68, SiO2 = 2.26 (tridymite)-2.65 (quartz), Mg2SiO4 = 3.22 [Dee 1992]. The comparison shows that a considerable volume expansion (~ 50%) must occur in the brick when magnesia is transformed into forsterite. On the other hand, considering that this expansion takes place
146
at the border of periclase grains by incorporating liquid silica, one may make an alternative calculation by adding the volume occupied by 1 mole of SiO2 (23 –27 cm3) to the volume occupied by 2 moles of MgO (22.4). The latter sum (45.4-49.4 cm3) is already very close to the volume occupied by 1 mole of Mg2SiO4 (43.4 cm3) – even some volume shrinkage (5-10 %) can be expected. This, together with the filling up of some of the pores by the newly formed forsterite, explains why bursting or cracking of the fingers was not observed. 7.3. Conclusions 7.3.1. General conclusions from the finger tests In general, it can be said that during the finger tests the chrome-free refractory bricks performed worse than the mag-chrome ones. Two of the former (types E and F) could not even endure the full duration of the tests while the other four types were easily and thoroughly infiltrated by copper and slag components. The zirconia addition (types D & F) and the two kinds of spinel addition (MgAl2O4 – in types A & B, FeAl2O4 – in type C) did not seem to provide any extra benefit to the brick’ s quality, at least not with regard to Cu and CuOx penetration resistance. Nevertheless, the above-mentioned additives, known to improve the chemical and the thermal shock resistance, respectively, should prove rather useful in industrial practice, where the bricks are exposed to attack by various species and to large temperature fluctuations [Pet 2004, Pet 2005].
It was confirmed that the penetration problem common to periclase-based refractories has not been alleviated by the addition of better bonding materials, such as spinel grain (MgAl2O4 and FeAl2O4) and zirconia addition (ZrSiO4 and CaZrO3). The principal reason is that the added phases are not present in the refractory microstructure as a continuous network unlike the secondary chromite phase in mag-chrome bricks. This explains why the penetration resistance of the latter was reasonable. It should also be noted that the samples recovered from the copper zone of the latest generation of direct-bonded bricks – types J, K and L – showed the highest penetration resistance. There are two main phases present in magnesia-spinel bricks: MgO and MgAl2O4. During the brick firing process some inter-diffusion between these phases occurs. However, the level of interaction is much poorer than the one occuring between MgO and [Mg,Fe2+]O.[Cr,Al,Fe3+]2O3 (chromite spinel) during firing of mag-chrome bricks. The reason is the low solid solubility of alumina in magnesia (4 wt% Al2O3 at 1700°C) with respect to the solubility of FeO and Cr2O3 in magnesia (up to 14 wt% each at 1700°C) [Jon 2002]. Consequently, during cooling of the bricks no precipitation of a network of intergranular secondary MgAl2O4 spinel occurs – it remains in discrete positions instead, which results in worse penetration resistance. The only competitive option available nowadays seems to be a technique known as reinforced spinel bonding, which involves the introduction of fine, crystalline spinel in the magnesia matrix [Guo 2005]. Unfortunately, the latter technique will increase the cost of the bricks significantly. Therefore, it can be concluded that an economically viable chrome-free alternative to mag-chrome bricks for application in copper production furnaces remains elusive.
147
7.3.2. Proposal for zoned lining On the basis of the results of the finger tests it was concluded that the option for a zoned and mixed lining concept, including alternative chrome-free brick types is not viable. Therefore a proposal was made that a zoned type of magnesia-chromite lining should be used in the anode furnaces in order to prolong the refractory service lifetime. The following zoning of the furnaces was proposed:
x Barrel and sidewalls, above the slag line – low penetration and wear rates – the use of
the investigated in Chapter 6 magnesia-chromite brick can be continued; x Barrel and sidewalls, in contact with copper and slag – medium to high penetration
and wear rates (especially pronounced in the slagline) – a high-fired direct-bonded or rebonded brick should be used;
x Tuyere zone – severe penetration and wear rates – a special top-quality (fused-grain) refractory type must be applied.
148
CHAPTER 8
INVESTIGATION OF INDUSTRIALLY WORN REBONDED MAGNESIA-CHROMITE REFRACTORY BRICKS
8.1. Introduction As a result of the detailed post-mortem assessment of an anode furnace lined with a traditional direct-bonded magnesia-chromite bricks, described in Chapter 6, it was found that different wear zones exist in the lining. The results acquired from the finger tests, described in Chapter 7, showed that the investigated chrome-free materials are not suitable for application in anode furnace linings. Therefore, a zoned magnesia-chromite lining was proposed in which several brick qualities compatible with the specific requirements per zone were used. A similar and already detailed lining configuration with exact brick types was proposed by the refractory producer RHI. The proposal was accepted and the new type of lining was installed. The goal of this installation was to extend the overall refractory service lifetime and, thus, increase the campaign of the furnaces. This chapter reports on the performance of the new lining configuration. 8.2. Microstructural investigation of bricks collected from lining N2 When the anode furnaces were taken out of production for the planned relining and maintenance, an inspection of the refractory lining was carried out. Some parts of the latter were found to be in very good condition. Therefore, it was decided that a complete relining of the furnaces was not necessary. On the other hand, those areas where the lining was in a bad state were relined. Bricks from different locations of the replaced lining were collected for detailed examination. A plan of an anode furnace with the collection zones is presented in Figure 8.1. The bricks were cut to slices (perpendicular to the refractory hot/cold face) with a thickness of approximately 2 cm. Pictures of slices and bricks from the different collection zones are shown in Figures 8.2, 8.3 and 8.4. Two specimens (approximately 2 cm u 2 cm each) were cut with a diamond saw from each slice. A schematic diagram of the cutting procedure as well as the sample preparation method are given in Chapter 6, section 6.2. The second set of industrially worn bricks was comprised of the newly employed refractory type (RADEX DB 605-13), which was used for the first time in the anode furnace linings. This is a high-fired rebonded magnesia-chromite brick made of presintered grain and low-in-iron materials. This refractory type is characterised by very low silica content, excellent thermal shock resistance and improved slag penetration resistance with respect to the traditional direct-bonded bricks. The chemical composition of the brick together with some of its physical properties are presented in Table 8.1. The as-delivered microstructure of this refractory was discussed in Chapter 7 (section 7.2.3 – type H).
149
B1 - Sidewall, below gas outlet B2 - Sidewall, below burner inlet T1 - Close to tuyere 1 block T2 - Close to tuyere 2 block
MID - Middle part of barrel wall CM - Barrel wall, near mouth TAP - Barrel wall, near tap hole
Figure 8.1. Anode furnace plan with collection zones
Table 8.1. Chemical composition & physical properties of RADEX DB 605-13*
MgO [wt%]
Cr2O3 [wt%]
Fe2O3
[wt%] Al2O3 [wt%]
CaO [wt%]
SiO2
[wt%] BD
[g/cm3] CCS
[MPa] AP
[vol%] 60.0 19.0 13.5 6.0 1.3 0.2 3.22 70 17.0
BD – bulk density, CCS – cold crushing strength, AP – apparent porosity *Data provided by refractory supplier Since the residual thickness of the lower-quality bricks (H60) was still high enough (application in the low-wear regions) they were not replaced during the partial relining. As can be seen from Table 8.2, the use of the high-fired rebonded bricks in the high-wear zones was successful. This refractory type withstood to a large extent the corrosive and erosive action of the copper/anode slag melt. Similar phases were found in DB 605-13 bricks compared with the phases encountered in H60 bricks recovered from lining N1. The initially determined wear mechanisms affecting H60 type essentially remain the same as well. However, their impact was less detrimental for the higher-quality refractories (DB 605-13) used in the high-wear zones of the new lining compared to the scale at which the medium-quality bricks (H60) were affected.
Charging mouth Burner inlet
Gas outlet
B1 B2
T1 T2
CM
MID
TAP
150
a) b)
Figure 8.2. Slices of bricks recovered from zone B1 (a) and zone B2 (b)
a) b)
Figure 8.3. Slices of bricks recovered from zone CM (a) and zone MID (b)
a) b)
Figure 8.4. Bricks recovered from zone T1 (a – up) & zone TAP (a – down)
and zone T2 (b)
151
Table 8.2. Summary of lining N2 performance
Lining Collection zone (see Figure 8.1)
Brick type Residual brick thickness (% of original thickness)
Overall refractory lifetime
B1 DB 605-13 75-80 B2 DB 605-13 75-80
MID DB 605-13 80-85 TAP DB 605-13 60-65 T1 DB 605-13 55-60
N2 – new lining concept: first trial
T2 DB 605-13 55-60
11 months – complete relining was not needed, only partial
8.2.1. Results of SEM examination Figures 8.5 – 8.10 show BSE overview images (low magnification – 25x) of samples prepared from the industrially worn bricks. Substantial levels of forsterite were observed in the samples coming from zones B1 and B2. Another common feature of these samples is that they were completely infiltrated by copper oxide. However, the images show that the samples recovered from zone B2 contain more forsterite than the samples collected from zone B1. This comes as a result of the different conditions to which these zones are exposed during the fire-refining process. Zone B1 is situated below the gas outlet, while zone B2 is located below the burner inlet. The latter area is exposed to higher temperatures, which cause local overheating of the slag layer. This lowers the slag viscosity and accelerates the kinetics of the reaction of forsterite formation (8.1).
2[MgO]Refr + (SiO2)Slag = [Mg2SiO4]Refr (8.1)
In all investigated samples a dense slag layer was observed along the joint with the adjacent brick. Apparently, slag and copper infiltrated the joints between bricks, causing the formation of a slag layer with copper inclusions. The layer disappears at a distance of 10–11 cm from the hot face. We assume this as the maximum depth of anode slag (and copper) penetration along the joints. A forsterite layer is always present between the frozen slag and the interior of the brick. The layer was formed as a result of reaction (8.1).
152
Per – periclase (MgO), F – forsterite (2MgO.SiO2), Sp – in situ formed spinel phase (Mg,Fe,Cu)(Fe,Cr,Al)2O4
Figure 8.5. Overview image of a sample recovered from zone B1
Per – periclase (MgO), F – forsterite (2MgO.SiO2), Sp – in situ formed spinel phase (Mg,Fe,Cu)(Fe,Cr,Al)2O4
Figure 8.6. Overview image of a sample recovered from zone B2
1 mm
1 mm
153
Per – periclase (MgO), F – forsterite (2MgO.SiO2), Sp – in situ formed spinel phase (Mg,Fe,Cu)(Fe,Cr,Al)2O4
Figure 8.7. Overview image of a sample recovered from zone CM
Per – periclase (MgO), F – forsterite (2MgO.SiO2), Chr – primary chromite spinel (Mg,Fe)(Cr,Al,Fe)2O4
Figure 8.8. Overview image of a sample recovered from zone MID
1 mm
1 mm
154
Per – periclase (MgO), F – forsterite (2MgO.SiO2), Sp – in situ formed spinel phase (Mg,Fe,Cu)(Fe,Cr,Al)2O4
Figure 8.9. Overview image of a sample recovered from zone T1
Per – periclase (MgO), F – forsterite (2MgO.SiO2), Chr – primary chromite spinel (Mg,Fe)(Cr,Al,Fe)2O4
Figure 8.10. Overview image of a sample recovered from zone TAP
1 mm
1 mm
155
8.3. Microstructural investigation of bricks collected from lining N3 After the partial relining the anode furnaces were put back into operation and performed without any problems for one year. Then the furnaces were taken out of operation and their refractory linings were carefully examined. The inspection revealed that the linings were in very good condition; therefore it was decided not to shut down the furnaces for relining. Normal operation was continued while constantly monitoring the linings thickness by means of infrared equipment and other on-site techniques. Proceeding in this way, the furnaces campaign was extended with 5 months. At that time, a need for repairs of some ancillary metallurgical equipment appeared, and therefore the anode furnaces had to be taken out of production as well. A decision was taken to use the non-operational period for maintenance and complete relining of the furnaces. The examination of the worn linings showed that the campaign could still be prolonged by a few more months without taking any risks. During the relining bricks from specific locations were collected for detailed examination. A plan of an anode furnace with the collection zones is presented in Figure 8.11.
B1 – Sidewall, below gas outlet B2 – Sidewall, below burner inlet T1 – Close to tuyere 1 block
T2 – Close to tuyere 2 block MID – Middle part of barrel wall TAP – Barrel wall, near tap hole
Figure 8.11. Anode furnace plan with collection zones This third set of industrially worn samples was also prepared from DB 605-13 bricks. As can be seen from Table 8.3, the excellent performance of the newly employed refractories continued in the next furnace campaign. The results show that even after 16 months of service, the lining was not significantly affected and could be safely used for a few more months. The refractory bricks were cut lengthways to slices with a thickness of approximately 2 cm. Pictures of slices and bricks from the different collection zones are shown in Figures 8.12, 8.13 and 8.14. Two to three specimens (approximately 2 cm u 2 cm each) were cut with a diamond saw from each slice. The exact positions of the extracted specimens are shown in the figures below. A schematic of the cutting procedure as well as the sample preparation method were described earlier (Chapter 6, section 6.2).
B1 B2 T1 T2
MID
TAP
Charging mouth Burner inlet
Gas outlet
156
Table 8.3. Summary of lining N3 performance
Lining Collection zone (see Figure 8.11)
Brick type Residual brick thickness (% of original thickness)
Overall refractory lifetime
B1 DB 605-13 60-65 B2 DB 605-13 60-65
MID DB 605-13 65-70 TAP DB 605-13 50-55 T1 DB 605-13 45-50
N3 – new lining concept: second trial
T2 DB 605-13 45-50
16 months followed by complete relining, but campaign could be still prolonged
a)
b)
H – hot face, C – cold face
Figure 8.12 Bricks recovered from zone B1 (a) and zone B2 (b)
a)
b)
H – hot face, C – cold face
Figure 8.13. Bricks recovered from zone TAP (a) and zone MID (b)
copper penetration limit
slag brick
1 2 3
H C C
C C
H
H H
1 2 1 2
1 2
B1
TAP
MID
B2
157
a)
b)
H – hot face, C – cold face
Figure 8.14. Slices of bricks recovered from zone T1 (a) and zone T2 (b) 8.3.1. Results of SEM examination Figures 8.15, 8.16 and 8.17 show BSE images of the samples. A dense slag layer has formed at the hot face of the bricks collected from zones B1 and B2. Metallic copper is incorporated in the slag layer. The samples are completely infiltrated by copper oxide (white phase on the images). Forsterite was also observed in the bricks collected from zones B1 and B2. As with the samples recovered from these zones during the previous relining, forsterite presence was more pronounced in zone B2 (see Figure 8.16). This confirms the explanation given earlier (see section 8.2.1).
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.15. Overview image of sample 1 recovered from zone B1
copper penetration limit
1 2
H H C C
1 2
3
copper penetration limit
1 mm
T2 T1
Per
slag
Chr
F
C
H
158
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.16. Overview image of a sample 1 recovered from zone B2
Figure 8.17 is a detailed BSE image clearly showing the abundance of forsterite in the sample collected from zone B2. Microprobe analyses performed on periclase grains near the hot face indicate that a considerable amount of FeO (up to 25 wt%) and some CuOx (up to 5 wt%) diffused into the rim of the grains. This diffusion changed the colour of the rims making it lighter. The latter phenomenon was observed in all investigated samples.
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.17. Detailed image of sample 1 recovered from zone B2
1 mm
Per
F
F
F
F
Chr
Chr
Per
slag
C H
H C
159
A BSE image of a sample collected from zone TAP is presented in Figure 8.18. The slag layer on the hot face is thinner than in the previous samples, which can be explained by the refractory’ s location in the lining (see Figure 8.11). Bricks around the tap hole are in contact with slag only during furnace rotation for slag skimming. Nevertheless, a certain amount of forsterite has formed at the refractory hot face. A low level of open porosity is observed in this sample since most of its pores were filled by metallic copper (white phase on the image).
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.18. Overview image of sample 1 recovered from zone TAP
Figure 8.19 shows in detail a piece of slag attached to the hot face of a brick recovered from zone MID. Copper drops with varying dimensions (0.5-0.01 mm) are enveloped in the slag phase. Microprobe analyses show that small amounts of alumina and magnesia have dissolved in the slag with the following elemental composition in at %: 1 Al, 2 Mg, 24 Fe, 73 O.
1 mm
Per
Chr
slag F
H
C
160
Figure 8.19. BSE image of sample 1 (slag crust attached to the refractory hot face) recovered from zone MID
In Figure 8.20 an overview of the hot face of sample 2 from zone MID is presented. A thick slag layer with copper inclusions can be observed. Forsterite is also present in significant amounts. Oxidised copper has penetrated deep into the brick filling most of the pores (Figure 8.21).
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.20. Overview image of sample 2 recovered from zone MID
0.5 mm
Cu
slag
pore
1 mm
slag
Per
Chr
F
F
Cu
H C
H C
161
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.21. Overview image of sample 3 recovered from zone MID
Figures 8.22, 8.23 & 8.24 show images of samples recovered from zone T1. A thin slag layer has formed at the refractory hot face (left part of Figure 8.22). Copper and forsterite inclusions can be observed in the layer. The bricks from this zone were heavily infiltrated by copper and copper oxide not only near the hot face but also in the refractory interior (Figure 8.24). However, silica could not manage to penetrate that deep since forsterite was not identified in this sample (Figure 8.24).
Figure 8.22. Overview image of sample 1 recovered from zone T1
1 mm Chr
Per slag
F
1 mm
Chr
Per C
H C
H
162
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.23. Detailed image of the slag/brick interface of sample 1 recovered from zone T1
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.24. Overview image of sample 2 (refractory interior) recovered from zone T1
Figure 8.25 presents an overview image of sample 2 from zone T2. As in sample 2 from zone T1, no forsterite was found in this sample, which means that silica penetration was not very deep – less than 10 cm from the hot face. A slag layer with copper and forsterite inclusions along the joint with the adjacent brick is visible in the left part of the image.
1 mm
Per
F slag
Cu
Per
Chr
H C
H C
163
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase, F – forsterite
Figure 8.25. Overview image of sample 2 recovered from zone T2
An overview image of sample 3 recovered from zone T2 is presented in Figure 8.26. Clearly visible is the limit of copper oxide infiltration. The distance from the hot face is 23-24 cm.
H – hot face, C – cold face, Chr – primary chromite spinel, Per – periclase
Figure 8.26. Overview image of sample 3 recovered from zone T2
slag
Per
Chr
1 mm
1 mm
F
Per Chr
H
C
C
H
164
As with the bricks recovered during the previous relining (lining N2), a dense slag layer with copper inclusions was observed along the joints with adjacent bricks. Obviously, slag and copper infiltrated the joints between bricks, causing the formation of this layer. The latter disappears at a distance of 10-25 cm from the refractory hot face. The minimum depth of penetration along the joints (10 cm) was measured in bricks recovered from zone B1 while the maximum depth (25 cm) was registered in bricks collected from the tuyeres zone. Another remarkable feature is that a forsterite layer is always present between the slag layer and the interior of the brick. This forsterite layer was formed as a result of reaction (8.1). 8.4. Degradation mechanisms
In the course of the present study 22 bricks in total were collected from distinct locations in the anode furnace linings for microstructural investigation. The first set of bricks (lining N1) included the formerly applied traditional direct-bonded magnesia-chromite refractories while the second and the third sets (linings N2 and N3) were comprised of the newly employed rebonded magnesia-chromite bricks based on presintered grain. The extensive microstructural study revealed that various chemical, thermal and mechanical (physical) degradation mechanisms influence the bricks in service. 8.4.1. Chemical wear mechanisms Copper and slag components infiltration
All industrially worn bricks were heavily infiltrated with metallic copper and copper oxide. Bricks recovered from zones 1, 2, 3, MID and TAP were mostly infiltrated with copper metal while bricks recovered from zones 4, 5, 6, CM, B1, B2, T1 and T2 were predominantly infiltrated with copper oxide. The infiltration limit is located at a distance of around 20 cm from the hot face for most of the investigated bricks. However, this limit for bricks collected near the tuyeres zone reached 25-26 cm from the hot face. The infiltrated copper oxide is mainly present as cuprous oxide (Cu2O), although a minor amount of cupric oxide (CuO) was also analysed mainly at the refractory hot face. The infiltrating anode slag brings copper oxide, iron oxide and silica to the brick. It was noticed that in many samples copper oxide and iron oxide diffused in periclase and primary chromite grains. Another slag component (silica) has reacted with magnesia (from periclase) to form forsterite (see Forsterite Formation for details). The infiltrating liquid phases penetrate the refractory through the open pore network and along the periclase grain boundaries. Once inside the brick they dissolve most of the intergranular secondary chromite spinel, partly replacing in this way, the direct-bonded (solid) structure of the refractory with a liquid-bonded one. The newly-formed, partially liquid-bonded microstructure is more susceptible to hot erosion and abrasion mechanisms resulting in accelerated refractory wear. Contaminants and overheating of the copper, as well as high bath pressure because of great bath depth, facilitate the penetration of metallic copper in the refractories. This process occurs along periclase grain boundaries and through the open pore network. One of the results of the copper penetration is a rise in the thermal conductivity of the refractory material. Owing to this considerable increase, a substantial temperature rise is
165
caused even at large distances from the hot face of the brick. The latter results in an amplification of all corrosive processes and in thermal expansion of the lining for which it may not have been designed. The surface tension, the wetting angle and the density of the infiltrating metal depend on the chemical composition and the temperature. For example, it is difficult for pure copper (99.99% Cu) to penetrate the brick microstructure because of the large wetting angle to periclase (approx. 140° at 1200°C) [Bar 1981]. “Difficult” in this context means that infiltration only occurs in case of great bath heights. On the other hand, the increase in thermal conductivity is in this case particularly severe. The brick’ s thermal conductivity may rise up to six times the original value, depending on the refractory type and its microstructure [Yam 1994]. However, contaminants (above all, oxygen) can reduce the boundary angle dramatically and thus facilitate infiltration. For example, at a temperature of 1200°C, the wetting angle between copper and periclase is reduced to less than 90° by an oxygen content of just a few tenths of a percent [Bar 1981]. The latter creates favourable wetting conditions and the penetration of oxidised copper is rapid and easy (the wetting angle between Cu2O and periclase at 1200°C is about 15°) [Bar 1981]. However, under such conditions the increase in thermal conductivity is lower or in case of strong oxidation may not take place at all, because the highly conductive copper framework (present in the pores) will be interrupted by copper oxide films of poor conductivity. Spinelisation
The infiltrated slag imports iron oxide to the brick. The latter diffuses into the primary (and to some extent into the secondary) chromite spinel (Mg,Fe)[Cr,Fe,Al]2O4. As a result of this diffusion, it is suggested that during refractory service, the primary chromite grains are continuously growing and sintering, thus forming a network of interconnected spinel grains (see Figures 8.5, 8.6 and 8.17). The consequence of this phenomenon is a more complex spinel composition, which is enriched in iron (wt %): 26-35 Cr, 22-27 Fe, 14 Mg, 9 Al, 0.5 Ca. Primary chromite in the as-delivered brick normally contains around 8 wt % Fe. The amount of Cr and Fe in the in-situ formed spinel phase varies strongly (26-35 % Cr, 22-27 % Fe), depending on the amount of diffused iron oxide. The amount of the other spinel components is nearly constant (±1%). The spinelisation process is highly pronounced in the slagline area (zones 4, 5, B1, B2, T1 and T2). As a result of this degradation phenomenon, a difference in the chemical composition ( &� between the hot face and the interior of the brick emerges. This &� together with some densification of the refractory hot face, makes the brick surface more prone to spalling.
Forsterite Formation
Forsterite formation is mostly outspoken in the slagline area (zones 4, 5, B1, B2, T1 and T2). The impact of this reaction on refractory performance was discussed in Chapter 7, section 7.2.4. MgO dissolution into slag
The dissolution of magnesia (from periclase) into the anode slag occurred in a direct way (Chapter 4, section 4.5.4). With the help of EPMA-EDS analyses it was established that up to 3-4 wt% MgO dissolved in the liquid anode slag.
166
8.4.2. Thermal and mechanical wear mechanisms Thermal shock stress, thermal and mechanical fatigue
Temperature changes in case of irregularities and interruptions in normal furnace operation create stresses in the bricks, which can be absorbed only to a limited extent and lead to brick breakage as soon as the microstructural strength limit is exceeded. Thermal fatigue of the lining is due to the temperature changes occurring during consecutive batches (furnace full of molten blister versus empty furnace in waiting periods). The amplitude is typically in the interval 400-500°C. The lining is also subjected to temperature changes according to the cycle of furnace rotation: i.e. the bricks are alternately down under the hot melt and up in the free (cooler) space of the furnace. The final result of thermal fatigue is weakening of the microstructural layers on the hot face. Bricks from all investigated zones are equally exposed to this degradation mechanism. Causes for mechanical fatigue are the same as those of thermal fatigue (e.g. furnace rotation resulting in changing brick load) but this phenomenon reaches also deep zones of the brick in the form of micro-fissures and macro-cracks, as a result of the heavy mechanical stress imposed on the lining (especially pronounced in zones 3 and TAP).
Hot erosion and abrasion
These phenomena are caused by the movement of the solid, liquid and gaseous content of the furnace and create a continuous wear on the hot face of the bricks from all zones except zones CM and 6. The higher the temperature and the speed of movement of materials in the furnace, the more pronounced are these wear mechanisms. The effect of the latter is also accelerated by the chemical corrosion mechanisms described in the previous paragraph. 8.5. Conclusions Based on the conclusions drawn from the post-mortem assessment of industrially worn bricks (described in Chapter 6) and also from the finger tests (described in Chapter 7), a proposal for a zoned magnesia-chromite lining was made. The proposed lining configuration was successfully tested in the anode furnaces. Bricks from various locations of the lining were recovered during two consecutive relinings of the furnaces. Samples were prepared and investigated by means of SEM and EPMA. The microstructural study confirmed the previously acquired knowledge of the degradation mechanisms affecting the bricks in service (during the investigation of the first set of industrially worn samples – lining N1). Due to the installation of the new zoned lining the refractory lifetime was considerably extended and the refractory consumption was reduced. However, it has to be noted that in the longer run any type of magnesia-chromite bricks will become more and more undesirable because of their high cost and environment-unfriendliness. Therefore, a long-term solution lies in the successful implementation of an alternative chrome-free lining.
167
CHAPTER 9
GENERAL CONCLUSIONS AND ACHIEVEMENTS 9.1. Introduction In the first chapters of this thesis it was made clear that the fire-refining of blister copper presents one of the harshest environments to be found in copper making practice because of the alternating changes of oxidation with reduction processes. The main principles of the reactions occurring during the fire-refining process were discussed. The metallurgical shop at Cumerio Med was presented and, more in detail, the applied practice of fire-refining and the lining concept of the anode furnaces were discussed. Due to the complex nature of the refractory/slag/metal interactions in the course of copper making, an extensive review was made of the slag systems and basic refractory materials that are used in the copper smelting, converting and refining furnaces. 9.2. Thesis goals The scientific goal of this study was to increase the knowledge on refractory degradation mechanisms affecting anode furnace linings. The industrial goals were to design an anode furnace lining configuration with an extended refractory lifetime and to investigate the possibility of applying alternative materials (chrome-free bricks) in anode furnace linings. 9.3. Initial industrial situation At the time when this study was initiated, the entire lining (apart from the tuyere blocks) consisted of traditional direct-bonded magnesia-chromite bricks. It was found that these linings suffered from acute wear, especially in the slagline zone, due to the stringent chemical, thermal and mechanical conditions imposed. During the relining of the furnaces worn bricks were collected from various locations in the lining for post-mortem assessment. The latter, together with macroscopical observations, provided general knowledge about the degradation phenomena occurring in the anode furnaces. From this it became clear that the situation was far from ideal. The refractory lifetime was limited; there were severely worn bricks in some areas and almost intact bricks in other zones. Apart from that, the lining consisted entirely of magnesia-chromite bricks. For economic and environmental reasons the preferred option was to (at least partially) replace this with a chrome-free lining. Hence, it had to be investigated if magnesia-chromite refractories applied in anode furnaces can be replaced by chrome-free alternatives or/and a better selection of various magnesia-chromite qualities. 9.4. Investigation of industrially worn samples The microstructural investigation performed on worn bricks recovered from anode furnace linings generated the following conclusions:
168
x The anode furnace lining can be divided in low wear (barrel and sidewalls above the slagline – zones CM and 6), medium wear (barrel and sidewalls in contact with copper – zones 1, 2, 3, MID and TAP), high wear (slagline – zones 4, 5, B1, B2, T1 and T2) and highest wear (tuyere blocks) areas.
x Refractory degradation comes as a consequence of the combined action of
chemical, thermal and mechanical drivers, which interact synergistically with one another, as shown in Figure 9.1.
It was found that various degradation mechanisms influence the refractory performance in anode furnaces (see Figure 9.1). Depending on the region in the lining their relative importance may vary. Firstly, the infiltrating copper, copper oxide and iron oxide continuously deplete the intergranular secondary chromite by dissolving most of it and partly replacing, in this way, the direct-bonded (solid) structure of the bricks with a liquid-bonded one. The latter phenomenon leads to microscopic dissolution of the refractory components, i.e. they are removed by the fluid motion, grain after grain. The thermal cycling reinforces the above-described degradation mechanism. Secondly, a densification of the refractory hot face takes place as a result of spinelisation and forsterite formation. This results in a discontinuous wear of the lining due to cracking and thermo-chemical spalling. The erosive fluid, the abrasion of solids and the changing hydrostatic load on the bricks (as a result of the furnace rotation) cause significant mechanical stresses on the refractory lining. The chemical wear mechanisms reduce the strength of the bricks and add to the susceptibility of the microstructure to further infiltration. This leads to compositional gradients � &�� LQ� WKH�EULFNV��The thermal load DGGV� IXUWKHU� VWUHVVHV� � 7�� WR the refractory material (especially in zones T1 and T2). When these already weakened structures are exposed to mechanical loads as well, this gives rise to both continuous degradation (hot erosion and corrosion) and discontinuous wear (thermo-chemical spalling).
It was also established that the initially determined wear mechanisms (first study of industrially worn bricks) essentially remain the same. However, the impact of the above-described drivers was less detrimental for the higher-quality refractories compared to the scale at which the medium-quality bricks were affected. Due to their excellent corrosion and erosion resistance, the newly employed high-quality bricks can largely withstand the aggressive conditions in the high wear zone. As a result of the fact that high-fired rebonded materials were used in the high-wear zone, the overall wear rate could be limited and brought into accordance with the relatively more rapid wear rate of the traditional direct-bonded bricks applied in the low-wear zones. This enhanced performance resulted in a significant improvement of the overall lining lifetime – from 11 months to 18 months (last relining was in September 2005 with next relining planned for April 2007).
169
7�– difference in temperature (temperature gradient) &�– difference in chemical composition (compositional gradient)
Figure 9.1. Overview diagram of the degradation mechanisms affecting magnesia-
chromite refractory linings in anode furnaces
Chemical impact : - Infiltration - Corrosion - Diffusion in periclase and
Continuous degradation (hot erosion and corrosion)
Discontinuous wear (thermo-chemical spalling)
170
9.5. Laboratory tests The limited availability of literature data on laboratory testing of any type of refractory bricks exposed to anode slag and the lack of capability for industrial testing led to the idea of performing laboratory tests. Static finger tests were chosen to explore the possibility of replacing magnesia-chromite refractories with alternative chrome-free bricks in anode furnace linings. The experiments also provided the opportunity to compare the performance of several qualities of magnesia-chromite materials. In general, it can be said that during the tests the chrome-free refractory bricks performed worse than the mag-chrome ones. Two of the former could not even endure the full duration of the tests while the other four types were easily and thoroughly infiltrated by copper and slag components. The zirconia addition and the two kinds of spinel addition (MgAl2O4 and FeAl2O4) did not seem to provide any extra benefit to the brick’ s quality, at least not with regard to Cu and CuOx penetration resistance. Nevertheless, the above-mentioned additives, known to improve the chemical and the thermal shock resistance, respectively, ought to prove rather useful in industrial practice, where the bricks are exposed to attack by various species and to large temperature fluctuations. As a result of the laboratory testing it was confirmed that the penetration problem common to periclase-based refractories has not been alleviated by the addition of better bonding materials, such as spinel grain (MgAl2O4 and FeAl2O4) and zirconia addition (ZrSiO4 and CaZrO3). The principal reason is that the added phases are not present in the refractory microstructure as a continuous network (unlike the secondary chromite phase in mag-chrome bricks). This explains why the penetration resistance of the latter was reasonable. As a general conclusion from the performed finger tests it can be said that the only competitive option available nowadays seems to be a technique known as reinforced spinel bonding, which involves the introduction of fine, crystalline spinel in the magnesia matrix [Guo 2005]. Unfortunately, the latter technique will increase the cost of the bricks significantly. Therefore, it can be concluded that an economically viable chrome-free alternative to mag-chrome bricks for application in copper production furnaces is still to be developed. 9.6. General conclusions and achievements The general conclusion from this study is that the scientific as well as the industrial goals were largely achieved:
x The degradation mechanisms affecting basic refractory bricks used in anode furnace linings were determined and investigated;
x Zoned magnesia-chromite lining was proposed and successfully implemented; as
a result of this implementation, the service lifetime of the anode furnace linings increased by 7 months (from 11 to 18 months);
x The possibility of applying alternative chrome-free bricks in anode furnace linings
was explored.
171
The scientific novelty of this work is two-fold:
x This is the first systematic study of the degradation mechanisms affecting basic refractory bricks in the industrial environment of anode furnace linings.
x The performed static finger tests with magnesia-chromite and chrome-free
refractories are the first of their kind. They allowed comparing the performance of various brick types exposed to copper and anode slag at a typical fire-refining temperature.
9.7. Guidelines for future research Finally, it should be pointed out that there is always room for improvement and plenty of work can still be done in this broad field. Just a few ideas for future research include:
x Due to their complex and multi-parameter nature, the interactions between refractories, slag and metal are not yet fully understood. These interactions and their influence on the fire-refining process can be further investigated.
x The contamination of spent magnesia-chromite refractories by metallic copper
and copper oxide (CuO), some of which may react with chromium oxide (Cr2O3) to generate the chromate (CuCrO4) labels this refractory type as environment-unfriendly and potentially hazardous for human beings. Therefore, refractory producers and research teams should combine their efforts to develop a viable chrome-free alternative to magnesia-chromite bricks for application in copper smelting, converting and refining furnaces.
x Environmental regulations as well as increased landfill costs and global
competition are driving refractory users and producers to recycle refractory bricks. However, the valorisation process of spent refractories is still in its initial stages of development and therefore gives a lot of opportunities to researchers.
172
LIST OF REFERENCES [Aji 1993] Ajima S., Hayashi M., Nishiyama Y. & Shimizu T., Converting and anode furnace operation of the Mitsubishi process at Naoshima, In Proceedings of the EPD Congress 1993 – Converting, Fire Refining and Casting, ed. by McCain J.D. & Floyd J.M., 1993, TMS-AIME, Warrandale, PA, USA, p. 333 [Aji 1999] Ajima, S., Konda, K., Kanamori, K., Iggarashi, T., Muto, T., Hayashi, S., Copper smelting and refining in Indonesia, , In Proceedings of the COPPER 99 – COBRE 99 International Conference, vol. V – Smelting Operations and Advances, ed. by George D.B., Chen W.J., Mackey P.J. & Weddick, A.J., 10-13 October 1999, Phoenix, AZ, USA, p. 57 [Aks 2003] Aksel C., Warren P. & Riley F., Fracture behaviour of magnesia and magnesia-spinel composites before and after thermal shock, JESC, vol. 3, 2003, p. 45 [All 1994] Allen J. & Stett M. A., In Proceedings of the EPD Congress 1994, ed. by Warren G., 1994, TMS-AIME, Warrendale, PA, p. 541 [Alv 1998] Alvarado R., Achurra G. & Mackay R., Present and future situation of the Teniente process, ed. by Stephens R.L. & Asteljoki J.A., In Proceedings of the Sulfide Smelting Symposium, TMS Annual Meeting, 15-19 February 1998, San Antonio, TX, USA, p.653
[Ame 1997] Amelinckx S., Van Dyck D., Landuyt V. & Van Tendeloo G., Handbook of Microscopy (Methods II), VCH, Weinheim, 1997.
[Ana 2001] Anan K., Wear of refractories in basic oxygen furnaces (BOF), Taikabutsu Overseas, vol. 21, 2001, p. 241
[Ass 1996] Assal J., Hallstedt B. & Gauckler L., Thermodynamic Evaluation of the Mg-Cu-O System, Z. Metallkd., vol. 86, 1996, p. 568
[Ban 2001] Banerjee S., Recent developments in steelmaking refractories, In Proceedings of UNITECR ’01, Mexico, 2001, p. 1033
[Bar 1981] Barthel H., Wear of chromium-magnesite bricks in copper smelting furnaces, Interceram, vol. 30, 1981, p. 250
[Bar 1984] Bartha P., The properties of periclase-spinel brick and its service stresses in rotary cement kilns, Interceram (Special Issue on Refractories), vol. 33, 1984, p. 15
[Ben 1995] Bennett J. P., Kwong K. & Sikich S.W., Bull. Am. Cer. Society, vol. 74, 1995, p. 71
173
[Bis 2002] Biswas A. K. & Davenport W. G., Extractive Metallurgy of Copper, fourth edition – 2002, Pergamon Press
[Bla 1980] Blanco, J. A., Antonioni T. N., Landolt C. A. & Danyliw G. J., Oxy-fuel Smelting at Inco’ s Copper Cliff Smelter, paper presented at the 50th Congress of the Chilean Institute of Mining and Metallurgical Engineers, Santiago, Chile, Nov. 1980 [Bla 1997] Blanpain B., Jones P.T., Vichev R.G. & Wollants P., In Proceedings of the 7th European conference on applications of surface and interfacial analysis, Chichester, 1997, p. 239 [Boo 2000] Boom R., Mills K. C. & Riaz S., Recent trends in research on slags, In Proceedings of the Sixth International Conference on Molten Slags, Fluxes and Salts, Stockholm, Sweden/Helsinki, Finland, 12–17 June, 2000, CD Rom – pdf.110 [Bre 1993] Brewster A. J., Frith M. & Evans D., The application of alumina-graphite products to steel ladles and torpedo ladles, Revue de Metallurgie, Cahiers de Informations Techniques, Paris, vol. 90, 1993, p. 369 [Bru 1998] Brueggerman M. & Caba E., Operation of the Contop Process at the ASARCO El Paso Smelter, In Proceedings of the Sulfide Smelting Symposium, TMS Annual Meeting, 15-19 February 1998, San Antonio, TX, USA, p. 681 [Bus 1995] Bustos, A., Cardoen, M., Janssens, B., High oxygen enrichment at UM-Hoboken converters, In Proceedings of Copper 95 – COBRE 95 International Conference, vol. IV –Pyrometallurgy of Copper, ed. by Chen W.J., Diyaz C., Luraschi A. & Mackey P.J., November 26–29, 1995, Santiago, Chile, p. 255 [Cam 1998] Campos R., Miranda S. & Smith T.J.A., Refinement of smelter operation and development, In: Stephens R.L., Asteljoki J.A. (Eds.), In Proceedings of the Sulfide Smelting Symposium, TMS Annual Meeting, 15–19 February 1998, San Antonio, TX, USA, p.667 [Car 1997] Carr H., Humphris M.J. & Longo A., The smelting of bulk Cu-Ni concentrates at the Inco Copper Cliff smelter, In Proceedings of the Nickel-Cobalt ‘97 International Symposium, vol. 3, August 17-20, 1997, Sudbury, Canada, p. 5 [CDA 2006] Copper Development Association – http://www.copper.org/, 2006
[Cla 1959] Clark C.B. & Mcdowell J.S., Basic brick in copper converters – their mineralogical changes, Journal of Metals, vol. 1, 1959, p. 119 [Coc 1995] Cocquerel M.A., The feasibility of a small copper smelter, In Proceedings of Copper 95 – COBRE 95 International Conference, vol. IV – Pyrometallurgy of Copper, ed. by Chen W.J., Diyaz C., Luraschi A. & Mackey P.J.November 26–29 1995, Santiago, Chile, p. 181 [Cod 2005] Codelco – company website – http://www.codelco.com/, 2005
174
[Con 2005] Contreras J.E., Castillo G.A., Rodriguez E.A., Das T.K. & Guzman A.M., Microstructure and properties of hercynite-magnesia-calcium zirconate refractory mixtures, Materials Characterization, vol. 54, 2005, p. 354 [Con 2006] Conejo A.N., Lule R.G., Lopez F. & Rodriguez R., Recycling MgO-C refractory in electric arc furnaces, Resources, Conservation & Recycling, 2006 [Cri 1999] Crites M. D. & Schlesinger M. E., Corrosion testing of chrome-free refractories for copper production furnaces, In Proceedings of Copper 99 – Cobre 99 International Conference, vol. 6, 1999, p. 187 [Cri 2000] Crites M.D., Karakus M., Schlesinger M.E., Somerville M.A & Sun S., Interaction of chrome-free refractories with copper smelting and converting slags, Canadian Metallurgical Quarterly, vol. 39, 2000, pp. 129
[Cum 2006] Cumerio – company website – http://www.cumerio.com/, 2006 [Dal 1988] Dal Maschio R., Fabbri B. & Fiori C., Industrial Ceramics, 1988, vol. 8, p.121 [Data Ref] Data provided by refractory producer [Dav 1987] Davies J. & McCollum J. M., in Advances in Refractories for the Metallurgical Industries, ed. by Rigaud M. A., 1987, CIM, Montreal, Canada, p. 195
[Dav 1999] Davenport W.G., Jenkins J., Kennedy B., Robinson T., Electrolytic copper refining – 1999 world tankhouse operating data, electrorefining and electrowinning of copper, In Proceedings of the COPPER 99 – COBRE 99 International Conference, ed. by George D.B., Chen W.J., Mackey P.J. & Weddick A.J., vol. 3, 10-13 October 1999, Phoenix, AZ, USA, p.3 [Dee 1992] Deer W. A., Howie R. A. & Zussman J., An Introduction to the Rock-forming Minerals, second edition, Longman Group UK Ltd, 1992 [Des 2006] Desmet D., Masters Thesis, MTM Department, Faculty of Engineering, K.U.Leuven, 2006
[Dou 1985] Douglas E. & Mainwaring P.R., Hydration and pozzolanic activity of non-ferrous slags, Am. Ceramic Soc. Bull., vol. 64, 1985, p. 700
[Eng 1996] Engel R., Marr R. & Pretorius E., Iron and Steelmaker, vol. 23, 1996, p. 66
[Ert 1986] Ertseva L.N., Rolgudin A.M., Ryabko A.G., The mechanism of refractory material wear in the reaction zone of a flash furnace chamber, Refractory materials, vol. 5, 1986, p. 49
[Fen 2002] Feneau C., 2002, Non-ferrous metals (from Ag to Zn), edited by Umicore Corporate Communication, Broekstraat 31, 1000 Brussels, Belgium [Flo 1993] Floyd J. M. & Guorgi G. A., In Proceedings of the EPD Congress 1993 – Converting, Fire Refining and Casting, ed. by J.D.McCain and J.M.Floyd, TMS-AIME, Warrandale, PA, p. 131 [Gad 1984] Gadalla A. M. M. & Mansour N. A. L., Ind. Eng. Chem. Fundam. vol. 23, 1984, p. 440 [Gaj 1987] Gajardo P. A., Advantages of Using Oxy-fuel Burners in Reverberatory Furnaces at Codelco-Chile Chuquicamata Copper Smelter, in Copper ‘87, vol. 4, Pyrometallurgy of Copper, edited by Diaz C., Landolt C. & Lurashi A., p.111 [Ger 1999] Gerasimov A. & Bagarov St., Refractories – a successful partner with metallurgy, In Proceedings of international conference “The efficient use of energy in metallurgy”, 22-24 June 1999, Varna, Bulgaria, p. 21 [Gev 1984] Geveci A. & Elliott J.F., In Proceedings of the 2nd international symposium on metallurgical slags and fluxes, ed. by Fine H.A. & Gaskell D.R., AIME, New York, 1984, p. 547 [Gol 1981] Goldstein J. I., Newbury D. E., Echlin P., Joy D., C. Fiori C. & Lifshin E., Scanning Electron Microscopy and X-Ray Microanalysis, Plenum Press, New York, 1981 [Gon 1989] Goncalves G. E., Mater. Ogniotrwale, vol. 41, 1989, p. 155 [Gor 2003] Gorai B., Jana R. K. & Premchand A., Characteristics and utilisation of copper slags – a review, Resources, Conservation and Recycling, vol. 39, November 2003, p. 299 [Got 1995] Goto K. & Lee W., The ‘direct bond’ in magnesia-chromite and magnesia-spinel refractories, Journal of the American Ceramic Society, vol. 78, 1995, p. 1753 [Got 1997] Goto K., Argent B.B. & Lee W., J. Amer. Cer. Soc., vol. 80, 1997, p. 461 [Gra 1990] Grau Rivas P. In Advances in Refractory Technology, ed. by Fischer R.E., The American Seramic Society, Westerville, OH, 1990, p. 104 [Gre 2003] Gregurek D. & Majcenovic C., Wear mechanisms of basic brick linings in the Non-Ferrous metals industry – Case studies from copper smelting furnaces, RHI Bulletin, vol.1, 2003, p. 17 [Gri 1996] Griffin D.J., Chrome-free generation, World Cement, vol. 27, 1996, p. 88 [Guo 2005] Guo Z., Palco S. & Rigaud M., Reaction characteristics of magnesia-spinel Refractories with cement clinker, Int. J. Appl. Ceram. Technology, vol. 2, 2005, p. 327
176
[Guo 2006a] Guo M., Parada S., Jones P.T., Van Dyck J., Boydens E., Durinck D., Blanpain B. & Wollants P., Degradation mechanisms of magnesia-carbon refractories by high-alumina stainless steel slags under vacuum, accepted for publication in Ceramics International, 2006 [Guo 2006b] Guo M., Jones P.T., Parada S., Boydens E.,Van Dyck J., Blanpain B. & Wollants P., Degradation mechanisms of magnesia-chromite refractories by high-alumina stainless steel slags under vacuum conditions, accepted for publication in the Journal of the European Ceramic Society, 2006 [Hal 2003] Haldar M.K., Tripathi H.S., Das S.K. & Ghosh A., Effect of compositional variation on the synthesis of magnesite-chrome composite refractory, Ceramics International, vol. 30, 2004, p. 911 [Ham 1991] Hammer T., in Proceedings of UNITECR ’91, The German refractories association, Bonn, 1991, p. 73 [Han 1993] Hanniala P;, Kojo I. V. & Kyto M., In Proceedings of the EPD Congress 1993 – Converting, Fire Refining and Casting, ed. by J.D.McCain and J.M.Floyd, TMS-AIME, Warrandale, PA, p.107 [Har 1969] Harris J. D. & Frechette V. D., J. Can. Ceram. Society, vol. 38, 1969, p. 15 [Har 2005] Harvey A., The composition and environmental hazard of copper slags in the context of the Basel Convention, Resources, Conservation and Recycling, vol. 43, March 2005, p. 353 [Has 2006] Hashemi B., Nemati Z.A. & Faghihi-Sani M.A., Effects of resin and graphite content on density and oxidation behavior of MgO-C refractory bricks, Ceramics International, vol. 32, 2006, p. 313 [Her 1996] Hernandez R.F., 1996, The art of producing copper in a Peirce-Smith converter, Journal of Metals (August 1996), p. 39 [Hir 2003] Hirata T., Morimoto T., Ohta S. & Uchida N., Improvement of the corrosion resistance of alumina-chromia ceramic materials in molten slag, JECS, vol. 23, 2003, p. 2089 [Hof 1997] Hoffmann J.E., Bautista R., Ettel V.A., Kudryk V. & Wesely R.J., The Electro-refining and Winning of Copper, In Proceedings of the TMS Annual Meeting 1997, Denver, CO, USA, p. 545 [Iba 1987] Ibarra M.E., In Proceedings of Copper ’87 (Vol. 4 Pyrometallurgy of Copper) ed. by C. Diaz, C. Landolt and A.A. Luraschi, 1987, p. 441 [Iid 1997] Iida O., Hayashi M. & Goto M., Process designs on new smelter projects of the Mitsubishi continuous copper smelting and converting process, In: Proceedings of the Nickel–Cobalt 97 International Symposium, vol. 3, August 17–20, Sudbury, Canada, p. 499
177
[Iid 2000] Iida T., Sakai H., Kita Y. & Shigeno K., ISIJ, vol. 40 (Supplement), 2000, p.110 [Jal 2002] Jalkanen H., Poijarve J. & Pajari H., Sulfide Smelting 2002, Seattle, WA, USA, Feb. 2002, p. 363 [Jan 2005] Janssens E., Masters Thesis, MTM Department, Faculty of Engineering, K.U.Leuven, May 2005 [Jon 2000] Jones P.T., Blanpain B. & Wollants P., Degradation mechanisms of magnesia-chromite refractories in VOD ladles during the production of stainless steel, Ironmaking and Steelmaking, vol. 27, 2000, p. 228 [Jon 2001] Jones P. T., Degradation mechanisms of basic reractory materials during the secondary refining of stainless steel in VOD ladles, PhD Thesis, MTM Department, KULeuven, May 2001 [Jon 2002] Jones P.T., Vleugels J., Volders I., Blanpain B., Van der Biest O. & Wollants P., A study of slag-infiltrated magnesia-chromite refractories using hybrid microwave heating, J. Europen Ceram. Soc., vol. 22, 2002, p. 903 [Kan 1999] Kanno K., Koike N. & Korai Y., Mesophase pitch and phenolic resin blends as binder for magnesia graphite bricks, Carbon, vol. 37, 1999, p. 195 [Kii 1995] Kii T., Hiragushi K., Yasui H. & Mukai K., In Proceedings of the Unified International Technical Conference on Refractories (UNITECR’95), Kyoto, Japan, 1995, p. 379 [Kim 1984] Kimura M.,Yasuda Y. & Nishio H., Development of magnesia spinel bricks for rotary cement kilns in Japan, Interceram (Special Issue on Refractories), vol. 33, 1984, p. 22 [Kob 1982] Kobayashi M., Nishi M., & Miyamoto A., Taikabutsu Overseas, vol. 2, 1982, p. 5 [Koh 2000] Koho V.I., Jokilaakso A., & Hanniala P., Flash smelting and converting furnaces: a 50 year retrospect, Journal of Metals, February 2000, p. 57 [Kof 2006] Koffel M. & Taschler Th., Refractories for the copper and lead industry, World of Metallurgy – ERZMETALL, vol. 59, 2006, p. 133 [Kri 1988] Krivileva S. P., Fridman B. A., Sannikov Y. F. & Shindyapin V. A., Ogneuporui (Refractories), vol. 26, 1988, p. 667 [KUL 2006] Katholieke Universiteit Leuven, MTM Department [Lam 2004] Lampropoulou P. & Katagas C., In Proceedings of the 10th Int. Congress, Thessaloniki, April 2004, p. 97
178
[Lee 1997] Lee Y. & Nassaralla C. L., Minimisation of hexavalent chromium in mag-chrome refractories, Metallurgical and Materials Transactions, part B – Process Metallurgy, 1997, vol. 28, p. 855 [Lee 1998] Lee W. & Moore R. E., J. Amer. Ceram. Society, vol. 81, 1998, p. 1385 [Lee 1999] Lee W. & Zhang S., International Materials Review, vol. 44, 1999, p. 77 [Lee 2004] Lee W. & Zhang S., Direct and indirect slag corrosion of oxide and oxide-C refractories, In proceedings of the 7th International Conference on Molten Slags Fluxes and Salts, Cape town, South Africa, 2004. p. 309 [Lew 1997] Lewowicki S. & Rajczyk J., The usefulness of copper metallurgical slag as a micro-aggregate additive in mortars and concrete mixtures, Proceedings from the International Conference on Solid Waste Technology and Management, Philadelphia, PA (USA), November 1997, p. 16 [Li 1995] Li X., Rigaud M. & Palco S., J. Amer. Cer. Soc., vol. 78, 1995, p. 965 [Lin 1981] Lindkvist G., Interceram, vol. 30, 1981, p. 256 [Liu 2006] Liu J., Guo M., Jones P.T., Verhaeghe F., Blanpain B. & Wollants P., In-Situ observation of the direct and indirect dissolution of MgO particles in CaO-Al2O3-SiO2 based slags, accepted for publication in the Journal of the European Ceramic Society, June 2006 [Llo 1987] Lloyd G. E., Mineralogical Magazine, vol. 51, 1987, p. 3 [Lub 1988a] Lubaba N.C., Rand B. & Brett N.H., Effect of carbon binders on the development of porosity in MgO-C composite refractories, Br. Ceram. Trans. J., vol. 87, 1988, p. 164 [Lub 1988b] Lubaba N.C., Rand B. & Brett N.H., Compaction studies of MgO-Flake graphite mixtures relevant to the fabrication of composite refractory materials, Br. Ceram. Trans. J., vol. 87, 1988, p. 158 [Mac 1997] Macey C.L., Refractory Solutions for High Wear in Cement Kiln Transition Zones, In Proceedings of UNITER’97, Nov. 4-7, 1997. ed. by Stett M.A., The American Ceramic Society, New Orleans, USA, p. 1625 [Mah 1984] Mahant R. P., Mckague A. L., Norman G. E. and Michelutti R. E., Development and operation of the smelting and environmental improvement process: Falconbridge nickel mines – Sadbury operations, CIM Bulletin, vol. 77, 1984, p. 79 [Mak 1992] Makipaa M. & Taskinen P., in 12th Scandinavian Corrosion Congress & Eurocorrosion ’92, ed. by P. J. Tunturi, The Corrosion Society of Finland, p. 601 [Mar 1993a] Marvin C.G., Chrome-bearing hazardous waste, Bulletin of the American Ceramic Society, vol. 62, 1993, p. 66
179
[Mar 1993b] Marvin C.G., Ceram. Eng. Sci. Proc., vol. 14, 1993, p. 468 [Mig 1997] Miglani S. & Uchno J. J., Resin bonded alumina-magnesia-carbon brick for ladle refractories, in Proceedings of UNITECR’97, vol. 1, 1997, p. 193 [Mis 1999] Missen R.W. & Mims C.A., Introduction to Chemical Reaction Engineering and Kinetics, John Wiley & Sons Inc., New York, 1999 [Mob 1996] Mobasher B., Devaguptapu R. & Arino A., Effect of copper slag on the hydration of blended cemetitious mixtures, In Proceedings of the ASCE Materials Engineering Conference – Materials for the New Millennium, 1996, p. 1677 [Moo 1991] Moore B., Frith M. & Evans D., Developments in basic refractories for cement kilns, World Cement, 1991, p. 5 [Mor 1999] Morales P.C. & Mac-Kay R.S., Upgrading of the Teniente technology, In Proceedings of COPPER 99 – COBRE 99 International Conference, vol. V – Smelting Operations and Advances, ed. by George D.B., Chen W.J., Mackey P.J. & Weddick A.J., 10-13 October 1999, Phoenix, AZ, USA, p. 161 [Mos 1999] Moskalyk R.R., Alfantazi A., Tombalakian A.S. & Damir V., Anode effects in electrowinning, Minerals Engineering, vol. 12, 1999, p.65 [Mos 2003] Moskalyk R.R. &. Alfantazi A.M., Review of copper pyrometallurgical practice: today and tomorrow, Minerals Engineering, vol.16, 2003, p. 893 [Mot 2001] Motz H. & Geiseler J., Products of steel slags – an opportunity to save natural resources, Waste Management, vol. 21, 2001, p.285 [Muk 2000] Mukai K., Tao Z., Goto K., Li Z. & Takashima T., In Proceedings of the 6th Conference on molten slags, fluxes and salts, Stockholm/Helsinki, 2000, CD-ROM (145 pdf) [Mur 1984] Murolo P., First work experience with Japanese periclase-spinel bricks in the burning zone of an Italian cement kiln, Interceram (Special Issue on Refractories), vol. 33, 1984, p. 36 [Nak 1999] Nakamura Y., Hirai N., Tsutsui Y., Uchinokura K. & Tamura S. I., Recycling of refractories in the steel industry, Industrial Ceramics, vol. 19, 1999, p. 111 [Nie 1999] Nievoll J. & Buchberger B., Performance of new chrome-free magnesia bricks in precalciner kilns, Veitsch-Radex Rundschau, vol. 1, 1999, p. 57 [Nis 1995] Nishimura D., Technical trends of phenolics for Japanese refractories, Taikabutsu Overseas, vol. 15, 1995, p. 10 [Nol 1999] Nolte, A., Kreymann, R., Optimization of the blast furnace process in a secondary copper smelter, In Proceedings of COPPER 99 – COBRE 99 International Conference, vol. V – Smelting Operations and Advances, ed. by George D.B., Chen W.J., Mackey P.J. & Weddick A.J., 10–13 October 1999, Phoenix, AZ, USA, p. 335
180
[Nor 2006] Noranda Smelting Process – http://www.noranda.com/, 2006 [Oth 2005] Othman A. G. M. & Nour W. M. N., Recycling of spent magnesite and ZAS bricks for the production of new basic refractories, Ceramics International, vol. 31, 2005, p. 1053 [Out 2006] Outokumpu – company website – http://www.outokumpu.com/, 2006 [Pap 1995] Papp J. F., Bull. Am. Cer. Society, vol. 74, 1995, p. 118 [Par 2001] Parsons M.B., Bird D.K., Einaudi M.T. & Alpers C.N., Geochemical and mineralogical controls on trace element release from the Penn Mine base-metal slag dump, Applied Geochemistry, vol. 16, 2001, p. 1567 [Par 2005] Parada S., Guo M., Jones P. T., Blanpain B. & Wollants P., Chemical wear mechanisms of magnesia-chromite and magnesia-carbon refractories exposed to stainless steelmaking slags, In proceedings of UNITECR ’05, 8-11 November 2005, Orlando, Florida, USA, p. 63 [Pet 2004] Petkov V., Jones P. T., Boydens E., Blanpain B & Wollants P., A study of industrially worn refractory bricks recovered from a copper anode furnace lining, In Proceedings of the 47th International Colloquium on Refractories in Aachen, Germany, 2004, p. 125 [Pet 2005] Petkov V., Jones P. T., Boydens E., Dekkers R., Blanpain B. & Wollants P., Optimisation of an industrial anode furnace lining using distinct magnesia-chromite refractory qualities, In Proceedings of UNITECR ’05, ed. by Jeffrey D. Smith, 8-11 November 2005, Orlando, Florida, USA, p. 647 [Phe 2000] Phelps, R.W., The new smelter that complements the Grasberg complex, Engineering and Mining Journal, vol. 6, 2001, p. 78 [Phi 1968] Philibert, Heinrich K. F. J., Ed., NBS Spec. Publ. No. 298, 1968. [Pid 1986] Piderit G. J. & Plaza F. J., Interceram, vol. 35, 1986, p. 26
[Pid 1991] Piderit A.G. & Kronert W., in Proceedings of UNITECR ’91, The German Refractories Association, Bonn, 1991, p. 43 [Pre 1979] Pressley H. & White J., Trans. J. Brit. Cearm. Society, vol. 78, 1979, p. 4 [Pre 1981] Pressley H. & White J., Interceram, vol. 30, 1981, p. 28 [Que 1996] Queneau P.E. & Marcuson S.W, Oxygen pyrometallurgy at Copper Cliff – half a century of progress, Journal of Metals, vol. 48, January 1996, p. 14 [Que 2000] Quested P. & Monaghan B. J., In Proceedings of the Sixth International Conference on Molten Slags, Fluxes and Salts, Stockholm, Sweden/Helsinki, Finland, 12-17 June, 2000, CD Rom – pdf.475
181
[Quo 1988] Quon D.H.H., Wang S.S.B., Aota J., Wheat T.A. & Bell K.E., J. Can. Ceram. Society, vol. 57, 1988, p. 53 [Ram 1995] Ramachandran V., The treatment and minimisation of waste, Journal of Metals, February 1995, p.50 [Ran 1993] Rana I. A., Hanas W. N. & Allee R.E., In Proceedings of the EPD Congress 1993 – Converting, Fire Refining and Casting, ed. by J.D.McCain and J.M.Floyd, TMS-AIME, Warrandale, PA, p.183 [Ren 1981] Renkey A. L., Weaver E. L. & Vernon H. L., Interceram, vol. 30, 1981, p. 271 [Ren 1999] Rentz O., Krippner M., Haahre S. & Schultmann F., Report on best available techniques (BAT) in copper production, French-German Institute for Environmental Research, University of Karlsruhe, Germany, March 1999, p.167 [Res 2000] Resende W.S., Stoll R.M., Justus S.M., Andrade R.M., Longo E., Baldo J.B., Leite E.R., Paskocimas C.A., Soledade L.E.B., Gomes J.E. & Varela J.A., Key features of alumina/magnesia/graphite refractories for steel ladle lining, Journal of the European Ceramic Society, vol. 20, 2000, p. 1419 [Reu 2004] Reuter M., Xioa Y. & Boin U., Recycling and environmental issues of metallurgical slags and salt fluxes, In Proceedings of the 7th International Conference on Molten Slags, Fluxes and Salts, Capetown, South Africa, 24-28 January, 2004, CD ROM – pdf. 050 [RHI 2006] RHI Bulletin – RHI in the non ferrous metals industry, 2006 [Rib 1981] Riboud P.V., Roux Y., Lucas D. & Gaye H., Fachber, Huttenprax, Metallveiter1981, p.85 [Rig 1992] Rigby A. J. & Forrest R. E., in Non-Ferous Pyrometallurgy: trace metals, furnace practices and energy efficiency, CIM, Montreal, Canada, 1992, p. 443 [Rig 1993] Rigby A. J., Wear Mechanisms of Refractory Linings of Converters and Anode Furnaces, In Proceedings of the EPD Congress 1993 – Converting, Fire Refining and Casting, ed. by J.D.McCain and J.M.Floyd, TMS-AIME, Warrandale, PA, p.155 [Rob 1998] Robin J. M., Berthaud Y., Schimitt N., Poirer J. & Themines D., Thermomechanical behaviour of magnesia-carbon refractories, British Ceramic Transactions and Journal, vol. 97, 1998, p. 1 [Rol 1984] Rolgudin A.M., Gorodinski G.I., Ryabko A.G., Analysis of fire-resistant lining service time in the reaction zone of a flash furnace chamber, Non-ferrous metals, vol. 2, 1984, p. 14 [Sar 2002] Sarkar R., Das S. K. & Banerjee G., Effect of addition of Cr2O3 on the properties of reaction sintered MgO-Al2O3 spinels, J. Eur. Ceram. Soc., 2002, vol. 22, p. 1243
182
[Sch 1997] Schlesinger M.E., Karakus M., Crites M.D., Somerville M.A and Sun S., Chrome-Free Refractories for Copper Production Furnaces, In Proceedings of UNITECR ‘97, vol. 3, ed. by Stett M.A., the American Ceramic Society, Westerville, OH, p. 1703 [Sch 1998] Schlesinger M. E., Crites M. D., Sommerville M. A. & Sun S., Assessment of spinel-containing refractories for copper smelting and refining, J. Aust. Cer. Soc., vol. 34, 1998, p. 39 2003, p. 15 [See 1999] Seetharaman S., Sichen D. & Zhang J., Journal of Metals, August 1999, p.38 [Ser 1985] Serry M.A. & Attia M.S., J. Br. Ceram. Society, vol. 84, 1985, p. 142 [She 1999] Sheppard L.M., Minimizing refractory waste, Ceramic Industry, vol. 149, 1999, p. 39 [Shi 1998] Shikano H., Yagi T., Kamude M. & Yamamoto K., Improvement of side wall brick for refining ladle, Taikabutsu Overseas, vol. 8, 1998, p. 28. [Sic 1995] Sichen D., Nilsson R. & Seetharaman S., Steel research, vol. 66, 1995, p.458 [Sis 1978] Sisselman R., Copper: Iran’ s latest desert bonanza, Engineering and Mining Journal, vol. 179, 1978, p.57 [Sla 1995] Slag Atlas, Verein Deutscher Eisenhuttenleute, Dusseldorf, 1995 [Ste 1967] Stephenson I. M. & White J., Factors controlling microstructure and grain growth in two-phase (one solid + one liquid) and in three phase (two solid + one liquid) systems, Transactions of the British Ceramic Society, vol. 66, 1967, pp. 443-483 [Szc 2006] Szczerba J., Pedzich Z., Nikiel M. & Kapuscinska D., Influence of raw materials morphology on properties of magnesia-spinel refractories, JESC, 2006 [Tab 1992] Tabbert W. & Klischat H. J., Magnesia Spinel Bricks for the Cement Industry, In Proceedings of the 2nd International Symposium on Refractories, October 1992, ed. by X. Zhong, The Chinese Metal Society and the Chinese Ceramic Society, Beijing, China, p. 424 [Tan 1995] Tanaka M. & Kitai T., The influence of magnesia particle size distribution on the physical properties of MgO-C bricks, In Proceedings of UNITECR ’95, Japan, 1995, p. 116 [Tas 2000] Taskinen P., MTDATA – advanced program for thermodynamical equilibrium calculations, Internal report, Outokumpu Research Oy, 2000 [Tik 1978] Tikkanen M. H., Behaviour of raw copper melts in contact with refractories, Lecture held on 24/04/1978 in Helsinki, Finland [Tor 1998] Torres W.E., Current Teniente converter practice at the SPCC Ilo smelter, In Proceedings of the Sulfide Smelting Symposium, TMS Annual Meeting, ed. by Stephens R.L. & Asteljoki J.A., San Antonio, TX, USA, 15-19 February 1998, p.633
183
[Tro 1991] Troell P.T. & Michael O.J., Advances in the development and application of magnesia-carbon bricks, In Proceedings of UNITECR ’91, Aachen, Germany, 1991, p. 247 [Tsu 1995] Tsuchinari A., Okamoto H. & Yokoyama Y., Chrome-free refractory for vacuum degassers, In Proceedings of UNITECR ’95, vol. 2, 1995, p. 333 [Ull 1995] Ullmann F., Copper CISTI Document. In: Ullmann’ s Encyclopaedia of Industrial Chemistry, vol. A7, 1995, VCH, p. 471 [Uni 1996] Early history of copper mining and use, University of Nevada, Reno, 1996, http://www.unr.edu/sb204/geology/history.html [Urb 1981] Urbain G., Cambier F., Deletter M. & Anseau M. R., Trans. Brit. Seram. Soc., vol. 80, 1981, p.139 [Ust 1972] Ustyantsev V. M., Marevich V. P. & Lopachak G. G., Inorganic Materials, vol. 8, 1972, p. 518 [Val 1998] Valoref A. O., Recovery and recycling of scrap refractories, World Ceram. Refract., vol. 9, 1998, p. 13 [Var 2002] Vartiainen A. & Kyto M., Olivine slags: the ultimate solution to low copper slags?, Scandinavian Journal of Metallurgy, vol. 31, 2002, p. 298 [Vik 2001] Viklund-White C., Ruotanen K. & Gehor S., Spent ladle refractories characterisation and reuse in new refractory materials, In Proceedings of the International Colloquium on Refractories, Aachen, Germany, 2001, p. 86 [Whi 1970] White J. & Richmond C., Refractories J., vol. 46, 1970, p. 6 [Wil 1993] Wilson D. R., Evans R. M., Wadsworth I. & Cawley J., Properties and applications of sintered magnesium-aluminium spinels, In Proceedings of UNITECR ’93, p. 749 [Won 1990] Wong L. L. & Bradt R.C., Bull. Am. Ceram. Society, vol. 69, 1990, p. 1184 [Xie 2000] Xie D., Tran T. & Jahanshahi S., In Proceedings of the Sixth International Conference on Molten Slags, Fluxes and Salts, Stockholm, Sweden – Helsinki, Finland, 12-17 June 2000, CD Rom – pdf.157 [Xst 2006] Xstrata technology – company website – http://www.xstratatech.com/, 2006 [Yam 1984] Yamaguchi A., Control of oxidation-reduction in MgO-C refractories, Taikabutsu Overseas, vol. 4, 1984, p. 32 [Yam 1994] Yamaguchi K., Ogino F. & Kimura E., In Proceedings of the EPD Congress 1994, ed. by Warren G., TMS-AIME, Warrendale, PA, p. 803
184
[Zai 2004] Zain M. F. M., Islam M. N., Radin S. S. & Yap S. G., Cement-based solidification for the safe disposal of blasted copper slag, Cement and Concrete Composites, vol. 26, October 2004, p. 845 [Zha 2000] Zhang L. & Jahanshahi S., In Proceedings of the Sixth International Conference on Molten Slags, Fluxes and Salts, Stockholm, Sweden – Helsinki, Finland, 12-17 June 2000, CD Rom – pdf.87 [Zha 2001] Zhang S. & Lee W.E., Influence of additives on corrosion resistance and corroded microstructures of MgO-C refractories, JESC, vol. 21, 2001, p. 2393
185
LIST OF PUBLICATIONS Publications in journals: Petkov V., Jones P. T., Boydens E., Blanpain B & Wollants P., Chemical corrosion mechanisms of magnesia-chromite and chrome-free refractory bricks by copper metal and anode slag, Journal of the European Ceramic Society, article in press – available online at doi:10.1016/j.jeurceramsoc.2006.08.020 Publications in conference proceedings: Petkov V., Jones P. T., Boydens E., Blanpain B & Wollants P., A study of industrially worn refractory bricks recovered from a copper anode furnace lining, In Proceedings of the 47th International Colloquium on Refractories, Aachen, Germany, 2004, p. 125 Petkov V., Jones P. T., Boydens E., Dekkers R., Blanpain B. & Wollants P., Optimisation of an industrial anode furnace lining using distinct magnesia-chromite refractory qualities, In Proceedings of UNITECR ’05, ed. by Jeffrey D. Smith, 8-11 November 2005, Orlando, Florida, USA, p. 647
186
CV
Personal information: Name: Veselin Petkov Petkov Date of birth: 19.04.1973 Place of birth: Yambol, Bulgaria Nationality: Bulgarian E-mails: [email protected], [email protected] Professional experience: Period: 05.2000 – 09.2002 Position: research assistant Company/organisation: Department of Non-ferrous Metals and Alloys, Faculty of Metallurgy and Materials Engineering, University of Chemical Technology and Metallurgy, Sofia, Bulgaria Education:
x PhD Period: 09.2002 – 09.2006 Position: doctoral researcher Company/organisation: Department of Metallurgy and Materials Engineering, Faculty of Engineering, Katholieke Universiteit Leuven, Belgium
x University Period: 09.1993 – 07.1998 Speciality/qualification: metallurgical engineer, Master’ s degree in “Metallurgy of non-ferrous metals and alloys” – Department of Non-ferrous Metals and Alloys, Faculty of Metallurgy and Materials Engineering, University of Chemical Technology and Metallurgy, Sofia, Bulgaria
x Secondary Period: 09.1987 – 05.1992 Secondary education in an English language school, Sliven, Bulgaria
