Corrosion Resistance of MgO and Cr2O3-Based Refractory Raw Materials to PbO-Rich Cu Slag Determined by Hot-Stage Microscopy and Pellet Corrosion Test

Chemical resistance of commercial refractory raw materials against Cu slag is critical to consider them as candidates for the production of refractories used in Cu metallurgy. In this study, we show the comparative results for the corrosion resistance of four commercial refractory raw materials—magnesia chromite co-clinkers FMC 45 and FMC 57, PAK, and fused spinel SP AM 70—against aggressive, low-melting PbO-rich Cu slag (Z1) determined by hot-stage microscopy (up to 1450 °C) and pellet test (1100 and 1400 °C). Samples were characterized after the pellet test by XRD, SEM/EDS, and examination of their physicochemical properties to explore the corrosion reactions and then assess comparatively their chemical resistance. Since many works have focused on corrosion resistance of refractory products, the individual refractory raw materials have not been investigated so far. In this work, we show that magnesia chromite co-clinker FMC 45 exhibits the most beneficial properties considering its application in the production of refractories for the Cu industry. Forsterite (Mg2SiO4) and güggenite (Cu2MgO3) solid solutions constitute corrosion products in FMC 45, and its mixture with slag shows moderate dimensional stability at high temperatures. On the other hand, the fused spinel SP AM 70 is the least resistant to PbO-rich Cu slag (Z1); it starts to sinter at 970 °C, followed by a fast 8%-shrinkage caused by the formation of güggenite solid solution in significant amounts.


Introduction
The lifetime of refractory lining is one of the most critical factors during high-temperature processes. The repairs of the lining are both time-consuming and expensive, so high-quality refractories are desired and consistently developed. Refractory materials used in the Cu industry are not exposed to extremely high temperatures. However, they have to resist major thermal shocks. Moreover, they are subjected to chemical interactions with the surrounding environment, mechanical wear due to the movement of the stove charge, induced mechanical stresses in the linings, and hot erosion [1]. Refractory linings should withstand the influence of Cu, Cu matte, Cu slag, and SO 2 . Melted Cu does not wet the ceramics; thus, it slightly affects the lifetime of the refractory. However, Cu slag addition of spinel (MA) did not improve the corrosion resistance of MgO-based bricks. Consistently, magnesia-chromite refractories exhibited the best performance. Cr-free solutions for the Cu industry are continuously under investigation, e.g., refractories from the system MgO-Al 2 O 3 -SnO 2 [33,34], Si 3 N 4 -SiC [35], or MgO doped with ZrO 2 nanoparticles [36]. The resistance of MgO to liquid CuO x was examined in [37].
The influence of PbO-rich non-ferrous slag [38][39][40][41] and PbO-rich ferrous slag [38,41], or PbO-CuO x -Cu mixture [2], on the corrosion resistance of MgO-Cr refractories is well recognized in the literature. In our recent work, we concluded that the higher the concentration of PbO in the lead-rich copper slags, the higher the degradation level of magnesia-chromite refractory [42]. However, the chemical interactions between crucial refractory raw materials with slag enriched in PbO and CuO x are poorly investigated so far. Therefore, this work aimed to study the chemical interactions between commercial Cr-containing (magnesiachromite co-clinkers FMC 45 and FMC57 as well as chromium ore PAK) and Cr-free refractory raw material (fused magnesia-aluminate spinel SP AM 70) against commercial PbO-rich Cu slag (Z1) and identify the most corrosion-resistant one.

Materials and Methods
The experiment consisted of three main stages: the chemical and structural characterization of four refractory raw materials and Cu slag, preparation of the samples for corrosion tests (refractory raw materials-Cu slag mixtures), and corrosion tests of mixtures by hot-stage microscopy (HSM) and pellet test (PT).

Characterization of Refractory Raw Materials and Cu Slag
Refractory raw materials used in the experiment were the following: • Magnesia-chromite co-clinkers, FMC 45 and FMC 57 A semi-product for the production of magnesia-chromite co-clinkers is magnesia clinker, which is fabricated by burning magnesia carbonates at 1500-2000 • C and typically contains 95.0-99.8 wt.% MgO. Then, by burning the mixture of MgO clinker and chromite ore in the rotary kiln, or by fusion in an electric arc furnace, the magnesia-chromite coclinkers are produced [6,43]. In addition to MgO and Cr 2 O 3 , the secondary components in magnesia-chromite co-clinkers are CaO, SiO 2 , and Fe 2 O 3 in a small percentage of contents. Details are provided in Table 2.  [43].
Listed raw materials are commonly used during the fabrication of magnesia-chromite refractories applied as linings in slag zones of heating devices for the production of copper.
The corrosive agent used in the experiment was industrial PbO-rich Cu slag, obtained after the first stage of the converting process. The examined raw materials and slag were characterized in terms of their chemical and phase composition on samples with grain size <63 µm, applying X-ray diffraction (XRD) and X-ray fluorescence (XRF) techniques. The diffraction patterns were collected by applying Philips PANalytical X'Pert-Pro diffractometer at room temperature, in the 2θ range of 5-90 • , using a goniometer of 240 mm in diameter and Cu-Kα radiation. Chemical composition was measured using a PANalytical WDXRF Axios mAX spectrometer.
Samples of raw materials and slag used for chemical analysis (by XRF) and phase analysis (by XRD) as well as for both corrosion tests applied in the present study were powdered to below 63 µm. In the first step, bigger pieces of raw materials and slag (above 3 cm) were crushed using a laboratory jaw crusher. Then, they were subjected to a milling process using a centrifugal ball mill with ceramic grinders for 4 h. After this step, each component was sieved using a sieve with a mesh size of 63 µm, and oversized grains were manually milled using agate mortar. Eventually, the powders were mixed for 1 h.
The high-temperature behavior of the slag was tested by hot-stage microscopy (HSM) test to determine its sintering, melting, and flow temperatures and in situ dimensional changes, from room temperature to 1000 • C with a heating rate of 10 • C/min applying Carl Zeiss MH02 microscope. Sintering temperature was determined as corresponding to 2% shrinkage of linear sample dimensions. HSM method was described in detail in works [53,54].

Hot-Stage Microscopy
The powdered mixtures of raw materials and slag in mass ratio 1:1 were homogenized for 1 h and shaped into 3 mm cuboidal samples. Measurements were conducted using Carl Zeiss MH02 microscope with a heating rate of 10 • C/min to 1500 • C. Characteristic temperatures-sintering, melting, and flow-were determined based on in situ changes of the sample cross-section in microscopic view recorded by the camera during heating.

Pellet Test
The pellet test was conducted to identify and characterize the quasi-equilibrium reactions between the powdered mixture of refractory raw materials and Cu slag. Each refractory raw material was separately mixed and homogenized with Cu slag, ground below 63 µm, in the ball mill for 2 h in the mass ratio of 3:1 raw material to slag. The mixtures were uniaxially pressed under 65 MPa in cylindrical steel molds of 20 mm in diameter. The shaped samples were heat-treated in an air atmosphere at 1100 and 1400 • C in the laboratory electric furnace with a heating rate of 5 • C, soaked 3 h at maximum temperature, and finally cooled freely with the furnace. The temperature of 1100 • C was selected to evaluate the changes in phase composition of the samples before reaching a working temperature of copper converter (1300 • C), which can affect the performance of samples' components.
After the pellet test, the samples were subjected to examination of their bulk density and open porosity using a vacuum impregnator, based on the Archimedes' principle method. The XRD measurements for identification of the phase composition changes in corroded samples were conducted using Philips PANalytical X'Pert-Pro diffractometer at room temperature with 240 mm goniometer diameter applying Cu-Kα radiation, in the 2θ range of 5-90 • . The microstructural analysis of the corroded samples was performed by a scanning electron microscope Nova Nano SEM 200 equipped with an energy-dispersive spectrometer (EDS) to measure the chemical composition in microareas. The samples for microscopic observations were prepared by embedding the material in the epoxy resin, polishing by standard ceramic technique, and coating the surface with a carbon layer.  [43]. The impurities in this Cr-free raw material existed mainly as CaO, SiO 2 , and V 2 O 5 , equaling 0.9% in total.

Characterization of Raw Materials
The low content of SiO 2 and CaO in both magnesia-chromite co-clinkers <1.5% indicates their high suitability for refractory applications. On the other hand, in PAK the total content of Cr 2 O 3 and Al 2 O 3 was over 60% [55], and SiO 2 was slightly higher than recommended for high-temperature applications (<3.5 wt.%) [13]. In addition, SP AM 70 was characterized by a low amount of impurities. Thus, all test commercial raw materials were initially proposed as suitable for application in the Cu industry [43] In the PAK sample, the increased background is ascribed to the raised iron content among all tested samples ( Table 2). This phenomenon was also observed elsewhere [56,57].   Table 3. PbO was the main component of the test slag, reaching about 40%. Due to the low melting point of PbO (897 • C) [7], it is highly corrosive towards refractory components and, additionally, reduces the viscosity of liquid slag [2]. Other oxides in the test slag were Fe 2 O 3 , CuO, and SiO 2 , which existed in relatively high and comparable amounts of 18%, 16%, and 14%, respectively. The slag was contaminated by CaO and As 2 O 3 in comparable amounts (3%), as well as by Al 2 O 3 and ZnO (1.5%). In addition, low amounts of other oxides were detected such as Co 3 O 4 (1.18%), Na 2 O (0.97%), K 2 O (0.29%), NiO (0.31%), SnO 2 (0.12%), and Cr 2 O 3 (0.27%). Typically, Cu slags contain high levels of SiO 2 and Fe x O y [58,59]; hence, the properties of the selected slag in this work were predicted to be distinct. Compared to already investigated PbO-rich slags [38,41], the present slag is distinguished by a high level of aggressive Cu oxide.    Figure 3 presents the relative change in the linear dimension of the Cu slag cuboid sample as heated during the hot-stage microscopy test. During heating up in the range of 350-650 • C, the sample slightly expanded, as seen by the rising curve. Then, after the sharp descending stage, a second expansion in the narrow temperature range of 800-850 • C was observed. Finally, the sample started to sinter at 888 • C, followed by melting at 918 • C and full flow at 969 • C. The slag melted at a significantly lower temperature compared to typical fayalite-based slags (1420 • C in air [60] and above 1420 • C [61]) due to high Pb and Cu content.   Table 4. The highest sintering temperatures were observed for the raw materials containing high levels of Cr 2 O 3 -both magnesia-chromite co-clinkers and PAK, with the highest sintering point registered for FMC 45 of 1440 • C containing 25% Cr 2 O 3 . Such a high sintering temperature can be an effect of optimal content ratio Cr 2 O 3 /MgO (Table 2). Pakistani chromite ore PAK and magnesia-chromite co-clinker FMC 57 showed comparable sintering points of 1285 • C and 1260 • C, respectively. PAK is characterized by the highest Cr 2 O 3 content of about 50% and, simultaneously, by the lowest MgO amount of 17% among all tested raw materials.  The presented curves ( Figure 4) show the expansion for all test raw materials. The most significant linear expansion of 4% was observed for the mixture of FMC 57 + slag at about 950 • C. This resulted from the highest level of MgO, which possesses a high thermal expansion coefficient (α MgO = 15.60 × 10 −6 1/K [62]). FMC 45 + slag reached a maximum expansion of about 3% at 1100 • C; for PAK, the maximum expansion was registered at 800 • C as 2.5%.

Hot-Stage Microscopy Test of Cu Slag
Conversely, no-chrome raw material-fused spinel SP AM 70-showed steady dimensions up to 850 • C, followed by very sharp shrinkage from 850 • C up to 1250 • C, reaching 7%. The dimensions were maintained up to 1400 • C, above which the sample started to shrink again. Additionally, SP AM 70 revealed the lowest sintering point of 970 • C and, through this test, was confirmed to be the least resistant to PbO-rich Cu slag. For this material, only shrinkage was observed. In contrast to the fused spinel (SP AM 70), PAK expanded much more during heating, which is typical for Cr-containing spinels. Notably, FeCr 2 O 4 is characterized by a high thermal expansion coefficient of 12.38 × 10 −6 1/K (close to the thermal expansion coefficient of MgO) when compared to the thermal expansion coefficient of spinel sensu stricto of 8.83 × 10 −6 1/K (all coefficients were given for temperature range 20-1200 • C) [62]. None of the test mixtures started to flow when reaching the limit temperature of the test of 1500 • C.

Pellet Test
Phase Composition of the Raw Material-Slag Mixtures after the Pellet Test Figure 5 shows that both magnesia-chromite co-clinkers-FMC 45 and FMC 57-showed the same phase composition after corrosion by pellet test at 1100 and 1400 • C. The new phases, which appeared at a lower temperature of 1100 • C, were forsterite, Mg 2 SiO 4 , and spinel solid solution, identified as a separate spinel phase probably due to different stoichiometry compared to the original spinel, as its reflexes are located just next to the original spinel ones. This phase disappeared at 1400 • C. In PAK, for corrosion at 1100 • C, two new phases formed, fayalite (Fe 2 SiO 4 ) and lead silicate (Pb 2 SiO 4 ), while the former disappeared when heating the mixture at 1400 • C. The clear phase identification of solution phases by XRD is very limited, as the method does not identify the compounds directly but only their crystalline structure by the position of the peaks. The replacement of one element by another in a solid solution may shift the lattice parameters and consequently the position of the peaks.
Corrosion of fused spinel SP AM-slag mixture at 1100 • C showed a more significant number of corrosion products: forsterite (Mg 2 SiO 4 ), PbO, and CuO. Copper (II) oxidetenorite-was detected (instead of Cu 2 O) due to oxidizing atmosphere, as the investigation was done after cooling. According to the CuO-Cu 2 O-MgO phase diagram [63], below 1021 • C, the phase transformation Cu 2 O→CuO occurs. However, heating at a higher temperature of 1400 • C resulted in the CuO transformation into güggenite (Cu 2 MgO 3 ) due to a direct reaction of tenorite with magnesia. Moreover, PbO disappeared during heating at 1400 • C as a result of its dissolution in other phases at increased temperatures.
In summary, all of the samples underwent active corrosion without the formation of any corrosion-protecting interfaces. At 1400 • C, forsterite (Mg 2 SiO 4 ) was identified in FMC 45, FMC 57, and SP AM 70 (which possess above 20% of MgO), while fayalite (Fe 2 SiO 4 ) was detected in PAK, which contains a significant iron content (16% of Fe 2 O 3 ). Nevertheless, those two compounds create a continuous solid solution, called olivine, as per the diagram Mg 2 SiO 4 -Fe 2 SiO 4 [64]. Endmembers detected by XRD are in fact solid solutions where Mg and Fa can replace each other more or less arbitrarily or be substituted by other elements of similar geometry such as Ca, Zn, or Ni. There are works [65] showing that even large-size Pb can substitute for Mg in oxide structure despite the great difference in cation radii (r Pb2+ = 119 pm, r Mg2+ = 72 pm); however, very scarce information on the system MgO-PbO exists in literature (and a low number of cards are available in crystallographic databases). In the discussed system Mg 2 SiO 4 -Fe 2 SiO 4 , the increased amount of Fe leads to a continuously descending melting point of the solid solution from 1890 • C for forsterite to 1205 • C for fayalite. Table 5 shows summarized phases detected using XRD method.  Figure 6. The greatest density was detected for chromite ore PAK, being 3.3 g/cm 3 and 4.5 g/cm 3 for 1100 and 1400 • C, respectively. This behavior agreed with the low porosity of 25.6% for 1100 • C and 18.0% for 1400 • C. Out of the mixtures with magnesia raw materials, the purer FMC 57 showed a lower density at 1100 • C of 2.3 g/cm 3 than FMC 45. However, after heating at 1400 • C, it increased about 70% to 4.0 g/cm 3 , which significantly exceeded the density of FMC 45 of 3.4 g/cm 3 . Nevertheless, FMC 45 + slag reached the lowest porosity of 11.1% at 1400 • C, while for FMC 57 + slag, it equaled 20.9%. The slag mixture with spinel SP AM 70 exhibited the lowest density at 1100 and 1400 • C of 2.1 g/cm 3 and 3.2 g/cm 3 , respectively. This behavior was accompanied by the extremely high porosity of 76% at lower temperatures and 31.7% at higher test temperatures. Such porosity can be associated with a high MgO diffusion rate from the spinel into slag at 1100 • C, leaving pores, which preceded güggenite formation, confirmed by XRD at 1400 • C (Figure 5h).
The results show the positive influence of Cr 2 O 3 content on the high densification of the material linked with higher compactness of the sample represented by low open porosity. Moreover, the impurities in magnesia also lead to greater densification of the sample. This trend may influence lower penetration of materials by liquid slags present at operation conditions.

Microstructure Analysis of the Raw Material-Slag Mixtures after the Pellet Test
SEM microphotographs of FMC 45 corroded by Cu slag at 1100 and 1400 • C, together with EDS chemical analysis, are presented in Figure 7 and Table 6. Chemical analysis verified XRD results revealing magnesia and complex spinel solid solution in the samples after corrosion (Figure 7b, p. 1 and 3, respectively). For samples treated at 1400 • C, güggenite solid solution formed between magnesia and spinel grains (Figure 7b, p. 2), as a result of a reaction between magnesia from refractory and copper oxide from the slag. MgO grains at 1400 • C (Figure 7b) contained other dissolved elements such as Si (0.2%), Pb (0.8%), Cu (6.1%), and Fe (8.4%). Points 2 and 3 in Figure 7 represent average chemical compositions of magnesia grains containing PbO inclusions at 1100 • C, represented by dispersed small lightest-color microareas due to high molar weight. Scarce information on the system MgO-PbO exists in literature; thus, PbO solubility in MgO needs further study. Chen et al. [66] examined the MgO-PbO-SiO 2 system in the temperature range 700-1400 • C and found no phases containing only PbO and MgO, but they found three stable phases in the system PbO-SiO 2 (PbSiO 3 , Pb 2 SiO 4 , and Pb 4 SiO 6 ) and one ternary compound (Pb 8 Mg(Si 2 O 7 ) 3 ). Najem et al. [65] recently doped MgO nanoparticles with Pb 2+ , but the maximum lead amount doped was low, from 0.03 to 0.05 at.%. X-ray diffractometry confirmed that only the cubic phase of space group Fm3m was generated, characteristic of MgO. The increased Pb 2+ fraction doped in MgO caused the most intensive reflex (200) shift towards the lower 2θ values, and the increase in lattice parameter and crystallinity was observed. Moreover, the intensity of the most substantial peak decreased, attributed to the difference in the ionic radius between Pb 2+ (r Pb2+ 119 pm) and Mg 2+ (r Mg2+ = 72 pm).  Table 6. EDS analysis in microareas of mixture FMC 45 + slag after corrosion, in the points marked in Figure 7, together with corresponding phases.   Spinel solid solutions were found to accept various stoichiometries and numerous ions into their structure (XRD showed secondary spinel as a separate phase). At 1400 • C, grains of spinel solid solution, as shown by point 3 in Figure 7b, were enriched in Mg (16%) and Fe (21%) with a relatively low content of Cr (11%), compared to other locations presented by point 4 where they were enriched with Cr (17%) and impoverished concerning Fe (16%). In addition, it can be seen that spinel solid solution contained low substitutions of Pb at the level of 1%. Similar to the binary system MgO-PbO, the solubility of Pb 2+ in spinel solid solution in chromite ore was not found in the literature. Forsterite was not observed in SEM images, although it was identified by XRD, probably due to its low quantity in the sample.
The sample FMC 45 after corrosion at 1400 • C was visibly more densified compared to 1100 • C, with almost no slag phase in between the grains, which confirms the results showing its greater density and lower open porosity ( Figure 6).
In FMC 57 for both 1100 and 1400 • C ( Figure 8 and Table 7), magnesia and complex spinel solid solution were detected, which confirms XRD results of the corroded mixtures (Figure 5c,d). Pb from slag diffused and dissolved into the structure of complex spinel solid solution in low amounts at 1100 • C of 1% in p. 2 ( Figure 8a) and 0.4% in p. 2 (Figure 8b) at 1400 • C.  Table 7. EDS analysis in microareas of mixture FMC 57 + slag after corrosion, in the points marked in Figure 8, together with corresponding phases.   Among main slag components, MgO accepted Cu more than Pb-greater amounts of Cu dissolved at 1100 • C (5.5%: p. 1; 3.4%: p. 3; Figure 8a) than at 1400 • C (0.7%, p. 4, Figure 8a). Slightly higher amounts of Cu dissolved in spinel solid solution (2.7%-1100 • C, 0.7%-1400 • C).
Spinel solid solution grains detected in the material contained different amounts of constituent ions, e.g., at 1400 • C spinel grains mainly differed in the content of Cr and Fe (points 2,3 in Figure 8b). After treatment at 1400 • C, the slag phase was no longer visible in SEM images, compared to 1100 • C, at which point slag was distributed at grain boundaries.
The main phase detected in PAK by SEM/EDS ( Figure 9 and Table 8) was Cr-rich complex spinel solid solution, the only phase component in the original PAK (XRD- Figure 1c). This phase was found to accept about 5% of Cu for every measured microarea (points 1,3 in Figure 9a; points 3,4 in Figure 9b). Large grains of chromite spinel solid solution of above 50 µm showed high corrosion resistance against Pb, as lead was not detected in them at 1100 • C (point 1, Figure 9a) and only in slight amounts of 1.7% at 1400 • C (point 4, Figure 9b).  Table 8. EDS analysis in microareas of mixture PAK + slag after corrosion, in the points marked in Figure 9, together with corresponding phases.   The new phases-corrosion products in the mixture PAK + slag-confirm XRD results: lead silicate Pb 2 SiO 4 (light grey areas; 1100 • C, 1400 • C: point 2 in Figure 9a,b) and olivine [Mg,Fe] 2 SiO 4 (dark grey area; 1400 • C: point 1, Figure 9b). The amount of secondary Pb 2 SiO 4 was significant in SEM images. Pb 2 SiO 4 dissolved comparable, low amounts of Mg 2+ and Cr 3+ of about 2-3%. 2.5% of arsenic ions were registered only in Cr-spinel solid solution (point 1 in Figure 9a; point 3,4 in Figure 9b).
SEM/EDS analysis of the test mixture of fused spinel SP AM 70 + slag ( Figure 10, Table 9) revealed güggenite as corrosion product at 1400 • C (also identified by XRD- Figure 5h). Güggenite is a non-stoichiometric phase that is stable over a wide range of compositions, as shown by phase equilibria of the MgO-Cu 2 O-CuO system in [63]. Güggenite did not accept any Pb. It dissolved very slight amounts of Cr of 0.2% and Al of 0.4% and an increased amount of Fe of 4.2%. This phase appeared in SEM images as light-grey microareas. It shall be mentioned that there is scarce information in the literature about the properties of this phase. The slag phase, visible as light-grey microareas (p. 2, Figure 10a), was enriched with Mg (23%), which comes from the dissolution of spinel from refractory by aggressive slag. Si was not detected to greatly infiltrate the spinel phase, as maximum Si content in spinel grains was 0.3% at 1100 • C, and it increased to 5.1% at 1400 • C.  Table 9. EDS analysis in microareas of mixture PAK + slag after corrosion, in the points marked in Figure 10, together with corresponding phases. Spinel phase MgAl 2 O 4 showed relatively high resistance to Pb-rich copper slag at high temperatures, as only about 0.7% of Pb and 0.6% of Cu were detected in the spinel grains at 1100 • C, while at 1400 • C only Cu was observed at the level of 1.3%, and no Pb was detected. Spinel accepted above 5% of Fe at 1400 • C (and 0.7% at 1100 • C), and this behavior is characteristic for spinel structure, which tolerates ions with radii 40-80 pm (ferrous ions: r Fe2+,IV = 61.5 pm, r Fe2+,VI = 74 pm; ferric ions: r Fe3+,IV = 49 pm, r Fe3+,VI = 64.5 pm) [67][68][69][70][71]. Arsenic ions were not registered in any analyzed region.

Discussion
The study presents comparative results of chemical resistance of four commercial raw materials-two kinds of magnesia-chromite co-clinkers (FMC 45 and FMC 57), chromite ore PAK, and fused spinel SP AM 70-commonly used for the production of refractories. So far outside this study, only refractory products have been investigated in terms of their chemical resistance to copper slags. Table 10 summarizes and compares the results obtained in this work with results obtained for refractory products in other works.
The slag used in the experiment was PbO-rich slag containing about 40% PbO, 18% Fe 2 O 3 , 16% CuO, and 14% SiO 2 , derived from the Cu converting process. The oxides existed in the slag in the form of four main phases-Pb 2 SiO 4 , Fe 3 O 4 , Fe 2 SiO 4 , and Cu 2 O-which are aggressive towards refractory material as they possess low melting points: Pb 2 SiO 4 at 747 • C [7] Fe 2 SiO 4 at 1220 • C [72], and Cu 2 O at 1230 • C [73].
The presence of these low-melting-point components affected the high-temperature behavior of the raw materials tested by hot-stage microscopy (HSM), especially fused spinel SP AM 70, which started to sinter at 970 • C followed by drastic shrinkage reaching 7% at 1400 • C. This test showed the low resistance of this kind of raw material to PbO-rich Cu slag and confirmed that it should not be used as the main refractory component for non-ferrous metallurgy where similar slags occur.
The opposite behavior was found for Pakistani chromite ore PAK, which showed high resistance to PbO-rich slag and the best dimensional stability up to 1200 • C (the most crucial temperature range for the copper industry). A similar high-temperature behavior during the HSM test was observed for FMC 45 and 57, although they expanded more. The highest sintering point of 1440 • C, determined as 2% linear shrinkage of the test sample, was registered for the mixture FMC 45 + slag.
Comparing both HSM and pellet corrosion tests, FMC 45 is the most corrosion-resistant against PbO-rich slag among all test raw materials. The new phases that appeared after corrosion in FMC 45 were forsterite (Mg 2 SiO 4 ) due to reaction between MgO from refractory and SiO 2 from slag (acc. to Reaction (1)) and spinel solid solution having different stoichiometry than the original, which disappeared at 1400 • C. Here, FMC 45 revealed extremely low open porosity (11%). Forsterite formed in relatively low amounts in FMC 45 and FMC 57, as it was not observed during SEM/EDS analysis (Figure 7a,b and Figure 8a,b), and this was confirmed by low intensity of reflexes in XRD patterns (Figure 5a,b).
Although forsterite is a high-temperature phase, melting congruently at 1890 • C [6], thus improving thermal resistance of the material, and possessing good chemical stability and excellent insulation properties, it is simultaneously characterized by a relatively low density (3.21 g/cm 3 ). This is lower compared to the rest of the predominant phasesρ MgO = 3.59 g/cm 3 , ρ (Fe,Mg)Cr2O4 = 4.70 g/cm 3 -resulting in the major loosening of the microstructure and causing, if it occurs in excess, so-called "forsterite bursting", detrimental for the performance of refractories. The microstructure disintegration was previously reported for massive formation of forsterite as corrosion product [2,14,22,39,40,67,74,75].
Aggressive components of the slag such as Cu and Pb dissolved in the MgO grains of FMC 45 and FMC 57. For instance, 2% of Cu dissolved in MgO grains of FMC 45 at 1100 • C, increasing to 6.1% at 1400 • C. For FMC 57, 5.5% of Cu was detected in MgO at 1100 • C. As per the MgO-CuO x phase diagram [63], MgO can dissolve up to 21% of Cu ions at 700 • C (2), maintained until 1048 • C, then decreases to 10% and 5% at 1200 • C and 1400 • C, respectively, which proves the observed results.
MgO refractory + CuO x,slag → (Cu,Mg)O solid solution (2) Together with Cu, Fe diffused into periclase grains in significant amounts of about 8% (point 1, Figure 7b). This phenomenon is commonly observed [2][3][4]14,25] due to the similar geometry of ions in both oxides (r Mg2+ = 72 pm, r Fe2+ = 77 pm) [76]. In general, Pb diffused more willingly into Cr-spinel solid solution grains than MgO grains, which was previously observed in our recent work [42] for interactions between MgO-Cr refractory and PbO-rich slags and in [2] for interactions between MgO-Cr refractory and Cu-Cu x O-PbO melt. The study on the MgO-PbO system, as well as solubility of Pb 2+ in Cr-spinel solid solution, requires future study due to a gap in the literature, despite its significance in the industry. Si tends to diffuse mainly to MgO grains, with its greatest level of 5.7% at 1100 • C, but it remained unreacted (dissolved in magnesia). In contrast, Mg diffused into slag, which was observed in FMC57 (26% Mg, p. 4, Figure 8a) or SP AM 70 (24% Mg, p. 2, Figure 10a). FMC 57 and PAK, although possessing various phase compositions, showed similar high-temperature behavior in the HSM test, with sintering points for both raw materials above 1250 • C, followed by a 6% shrinkage up to the limit temperature of measurement of 1500 • C. Large grains of Cr-spinel solid solution in PAK, especially those enriched with Cr and Fe, showed excellent resistance to Pb infiltration. In contrast, the small ones were surrounded by the corrosion product-low melting phase of non-stoichiometric (as confirmed by SEM/EDS) lead silicate, Pb 2 SiO 4 (T m = 747 • C [7]), appeared as light-grey microareas. Pb 2 SiO 4 constituted the main phase component of the original slag identified by XRD; it was generated in slag as a result of a reaction between PbO (slag component) and Cr-spinel solid solution accepted more Cu than Pb, as the average amount of Cu in spinel grains was 5%, while the content of Pb was about 1.7%. Overall, the amount of Cu accepted by the phases of PAK was doubled when compared to FMC 45. The second component of the slag-fayalite Fe 2 SiO 4 -accepted Mg from PAK and formed olivine solid solution [Mg , Fe] 2 SiO 4 , which is commonly found as a corrosion product in MgO-containing refractories when fayalite-based Cu slags accompany the process [14,74]. With the increased content of MgO in the slag or refractory, part of Fe 2+ (r Fe2+ = 74 pm) in Fe 2 SiO 4 is replaced by Mg 2+ of smaller ionic radius (r Mg2+ = 66 pm), which leads to a decrease in lattice parameter reflected by the overall right shift of XRD reflexes [77]. As long as the MgO is more than FeO in olivine, the melting temperature of solid solution is high [64], and the material is safe for refractory application up to 1350 • C, typical for Cu production. PAK showed relatively high density after corrosion at 1400 • C of 4.70 g/cm 3 , which is the effect of high Cr 2 O 3 content (ρ Cr2O3 = 5.22 g/cm 3 ). In contrast to the shrinkage of SP AM 70 during heating, PAK-chromium-containing raw material-expands much during heating, which may result from changing Fe 2+ /Fe 3+ ratio during heating in the air-a partial reduction of iron oxides in (Mg,Fe)(Fe,Cr,Al) 2 O 4 up to 1200 • C, followed by their re-oxidation [68].
Fused spinel SP AM 70 showed the lowest corrosion resistance to PbO-rich slag among all tested refractory raw materials. This study showed that the spinel phase of low Cr 2 O 3 content is poorly resistant to Cu slag. SP AM 70 mixed with slag started to sinter as at 970 • C, which is especially unfavorable due to the fast and significant shrinkage of the material afterward. Such behavior resulted from the massive formation of a new corrosion product-güggenite. Güggenite formed in significant amounts in SP AM 70, as reflected by the high intensity of reflexes in XRD (Figure 5h), and this new phase was also observed by SEM/EDS (Figure 10b). As shown in phase equilibria of the MgO-Cu 2 O-CuO system in [63], güggenite is a non-stoichiometric phase that is stable over a wide range of compositions from CuMgO 2 to Cu 3 MgO 4 . Güggenite was found to exist in three crystallographic forms, namely, Güggenite A (orthorhombic, Pmmn (59)), güggenite B (orthorhombic, space group I(0)), and güggenite X (monoclinic). Güggenite A (67% CuO, 33% MgO) was obtained [63] in air at 1000 • C, and it was found to transform into güggenite X above 1050 • C. Güggenite B (75% CuO, 25% MgO) was obtained above 1050 • C. Güggenite can exist in equilibria with CuO (tenorite) and Cu 2 O (cuprite); with tenorite (monoclinic) at temperatures below 1000 • C; and with cuprite (cubic) above 1021 • C. This phase was lately shown in MgObased material corroded by CuO from matte [77]. In our work, the CuO x -MgO phase that appeared is characterized by the lower Cu/Mg ratio of 0.7 for FMC45 (p. 2, Figure 7b) and 0.2 for PAK (p. 2, Figure 10b), while theoretically Cu/Mg in güggenite can vary from 1 (for CuMgO 2 ) to 3 (for Cu 3 MgO 4 ). However, in [63] where authors investigated compositions with the general formula Mg 1-x Cu x O (x = 0 ÷ l.0), güggenite A appeared from x = 0.17 (950 • C), which corresponds to Cu/Mg = 0.2. This confirms the phase observed in the present work of low Cu/Mg ratio is in fact güggenite. Thus, güggenite can occur in a wider composition range, depending on temperature.
As can be seen in the magnified image (Figure 10b), the güggenite phase is likely to separate CuO of needle morphology. The needles-like small crystals were previously found in tenorite by several authors [77]. In fact, the presence of CuO needles in the neighborhood of güggenite can result from the fact that güggenite underwent decomposition during free cooling, as it decomposes at 1062 • C. Similarly, very scarce information is available on the system PbO-MgO, which requires further study.
MgO refractory + 2CuO slag → Cu 2 MgO 3solid solution (4) Cu 2 MgO 3 was previously found [67] during corrosion of MgO-Cr refractory by slag in the air atmosphere. Güggenite formed during cooling of the sample below 1100 • C; however, its physicochemical properties are poorly recognized and require future studies. According to the ternary system CuO-Cu 2 O-MgO [63], Cu 2 MgO 3 is stable below 1062 • C. This means that the presence of güggenite is unfavorable in ab refractory, which operates in temperature range 900-1100 • C, as its decomposition above 1062 • C causes spalling of the refractory structure, e.g., during refractory reheating [67]. Moreover, forsterite formed in SP AM 70, but its amount was relatively low, as confirmed by the slight intensity of XRD reflexes characteristic for this phase. Despite corrosion resistance of spinel solid solution grains ( Figure 1, Table 9; [31]), discussed phenomena of formation of new phases (forsterite, güggenite) contributed to the overall poor resistance of SP AM 70 to Cu slag. In addition, in [19] authors observed the liquid Cu-slag penetrated throughout the interfaces in a MgO-MgAl 2 O 4 -based refractory.
In the present work, low amounts of As were detected (~2%), which dissolved mostly in MgO and spinel grains. As assumed by Reinharter et al. [78], As 2 O 3 tend to react mainly with basic components of refractories.
Simplified corrosion resistance mechanism was given in Figure 11. Figure 11. The main mechanisms of the corrosion interaction between refractory raw materials and slag Z1 at 1100 • C and 1400 • C.

Conclusions
• Four commercial refractory raw materials commonly used in refractories for Cu metallurgy-two magnesia-chromite co-clinkers (FMC 45 and FMC 57), a chromite ore (PAK), and a fused spinel (SP AM 70)-were comparatively investigated against PbO-rich Cu slag by hot-stage microscopy and pellet corrosion test. • Test slag, containing the high levels of PbO of 39% and CuO x of 16%, was characterized by a low melting point of 969 • C, determined by hot-stage microscopy test.

•
From the results of both pellet and hot-stage microscopy corrosion tests, the most beneficial behavior was determined for FMC 45. It exhibited relatively stable dimensions during heating with 2% shrinkage, which corresponded to sample sintering at 1440 • C, being the highest sintering point among all tested raw materials. The only corrosion product was forsterite, which formed in slight amounts.
• Fused spinel (SP AM 70) was the least resistant to PbO-rich slag, as it started to sinter as first at 970 • C, followed by a fast and high 8% shrinkage of the material afterward. Güggenite solid solution formed due to the reaction between CuO x from slag and MgO from refractory. This phase is potentially detrimental, as it decomposes above 1062 • C, leading to the spalling of the material during reheating. Moreover, forsterite (Mg 2 SiO 4 ) formed during corrosion can be harmful in larger amounts due to its significant volume.

•
Despite the different phase composition, PAK and FMC 57 showed comparable corrosion resistance and high-temperature behavior; thus, they constitute the most promising prospective raw materials for refractories dedicated to non-ferrous metallurgy where aggressive Pb-Cu-O slags occur.