Effects of Aluminium Oxide Content on the Regenerated Magnesia-Calcium Bricks for Cement Rotary Kiln

1. Introduction

Currently, magnesite-chrome bricks have been widely used in cement rotary kiln in China [1]. It has led to serious environmental problems, because the Cr(Ⅲ) is easily oxidized into the toxic Cr(Ⅵ) in magnesite-chrome bricks under natural condition [2], especially spent magnesite-chrome bricks have not been reasonably handle. Therefore, it’s urgently to develop free chrome refractory bricks to substitute for magnesite-chrome bricks in cement rotary kiln [3,4,5].

Magnesia-calcium bricks have been widely used as lining materials in ferrous metallurgy to produce the clean steel such as refining furnace [6], because of their excellent performances [7]. They are also regarded as the substitution of magnesite-chrome bricks for cement rotary kiln [8], because they own the similar operational performances, such as high–temperature resistance, corrosion resistance and good coating adherence, etc. [9,10]. However, the application of magnesia-calcium bricks on cement rotary kiln has been restricted, since they have bad hydration resistance which can result in seriously structural damage [11]. Related researches showed that aluminium oxide (Al₂O₃) could enhance the hydration resistance of magnesia-calcium bricks [12,13,14].

There are more and more spent magnesia-calcium bricks not being reused effectively [15], because of high impurity content such as Al₂O₃ that can reduce the usability of regenerated product, and further leads to a new environmental problem. Besides, there was very rare study on the utilization of spent magnesia-calcium bricks [15,16]. A kind of high hydration resistance regenerated magnesia-calcium brick was prepared basing on the spent magnesia-calcium bricks in previous work [14], without removing the impurities of Al₂O₃ in spent magnesia-calcium bricks. It has not only realized the high reutilization rate of spent magnesia-calcium bricks, but also provided a possible pathway using regenerated magnesia-calcium bricks in the cement rotary kiln to substitute for magnesite-chrome bricks. Nevertheless, high-temperature performances such as high temperature strength and refractoriness need be degraded because of excessive impurity content, particularly the content of Al₂O₃ [17,18].

Therefore, in this work, the regenerated magnesia-calcium bricks with different content of additive Al₂O₃ were prepared respectively. The investigation and comparative of their phases, microstructures, room temperature and hot flexural strength, resistance to cement clinker corrosion and strength of coating adherence were carried out successively. This work can also provide a reference for the replacement of hazardous magnesite-chrome bricks in the cement rotary kiln by the regenerated magnesia-calcium bricks.

2. Materials and Methods

2.1. Main Raw Materials of the Regenerated Magnesia-Calcium Bricks

The main raw materials of the regenerated magnesia-calcium bricks were spent magnesia-calcium bricks, fused magnesia and Al₂O₃ powder additive. The spent magnesia-calcium bricks were from a refining furnace in ferrous industry, the fused magnesia was from a refractory materials plant, and Al₂O₃ powder additive was a kind of analytical reagent. Their chemical compositions were measured by an 1800 X-ray fluorescence spectrometer (XRF) and the results were shown in Table 1.

2.2. Preparation of the Regenerated Magnesia-Calcium Brick Samples

Figure 1 shows the synthetic process of the regenerated magnesia-calcium bricks. The mixture of the regenerated magnesia-calcium brick green body was prepared by using the main raw materials as show in Table 1, and liquid paraffin as the binder firstly. The green body was prepared by 100 MPa pressing the mixture to form a rectangular of 60 mm × 8 mm × 8 mm subsequently. The regenerated magnesia-calcium brick sample was obtained by firing the green body at 1873 K for 2 h under an air atmosphere finally. In this work, three kinds of regenerated magnesia-calcium brick green bodies containing 0 wt.%, 1.5 wt.%, 3.0 wt.% Al₂O₃ powder additive were prepared, their compositions were shown in Table 2, and their corresponding fired samples were marked as A0, A1.5, A3.0 respectively.

2.3. Characterization and Performance Test of the Regenerated Magnesia-Calcium Brick Samples

The phase compositions of the regenerated magnesia-calcium samples (regenerated samples) were identified by a MXP21VAHF X-ray powder diffractometry (XRD) analysis. The microstructures and elements distribution of the regenerated samples were investigated by an MLA 250-FEI Quanta scanning electron microscope (SEM) with an energy dispersive spectrometer (EDS). The room temperature flexural strength and strength of coating adherence of the regenerated samples were determined by a WDW–10E microcomputer controlled electronic universal testing machine. The hot flexural strength of the regenerated samples mechanical was determined by a CMT 5204 temperature-mechanical load coupling testing machine at 1573 K. The resistance to cement clinker corrosion and the coating adherence of the regenerated samples were investigated at 1823 K for 3 h as shown in Figure 2.

3. Results

This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn.

3.1. Phase Compositions of the Regenerated Magnesia-Calcium Bricks

Figure 3(a) shows that the main phase compositions of the sample A0, A1.5 were identical the magnesia (MgO), free calcium oxide (f–CaO), tricalcium silicate (Ca₃SiO₅, C₃S) and tetracalcium aluminoferrite (Ca₂FeAlO₅, C₄AF), while the main phase compositions of the sample A3.0 were the MgO, f–CaO, C₃S, C₄AF and tricalcium aluminate (Ca₃Al₂O₆, C₃A). It shows that a new phase C₃A generated and other phases did not change according to the increase of Al₂O₃ addition in the regenerated samples.

Figure 3(b) shows that the intensity of character diffraction peaks of f–CaO [I_CaO(200)] and MgO [I_MgO(111)] were the two highest peaks in the range of 28–42° and varied significantly. The relative intensities of I_CaO(200)/I_MgO(111) decreased observably along with the increase of increase of Al₂O₃ addition in the regenerated magnesia-calcium bricks as shown in Figure 4. It also indicates that the content of f–CaO decreased along with the increase of Al₂O₃ addition in the regenerated samples, because the phase of MgO was exclusive and stable in the regenerated samples as shown in Figure 3.

Figure 3 shows that the elements Fe and Al in the regenerated samples existed only in the C₄AF and C₃A phases. The standard reaction Gibbs energy of reactions between Fe₂O₃, Al₂O₃ and f–CaO which generate C₄AF and C₃A is shown in Figure 5 [19]. It shows that the phase of C₄AF generates preferentially when the impurity elements of Fe and Al exist concurrently, the phase of C₃A generates when there is excess Al element impurity (the mole ratio of Fe₂O₃ to Al₂O₃ < 1). Besides, preliminary study had proved that the phase of C₂F generates when the mole ratio of Fe₂O₃ to Al₂O₃ >1 [20].

The theoretical content of the C₃S, C₄AF C₂F, C₃A and f–CaO phases (calcium–based phases) in the regenerated samples can be calculated by the Eq. (1) - (7) [21]. The results (as shown in Table 3) indicates that the theoretical content of the C₃A phase increased while that of the f–CaO phase decreased as the increasing of Al₂O₃ addition, and the Fe element of the sample A1.5 and A3.0 all generated the C₄AF phase. These calculation results were identical with the phase compositions analysis shown in Figure 3.

w(C₃S) = 3.80w(SiO₂) (1)

w(C₄AF) = 4.77w(Al₂O₃) (mole ratio of Fe₂O₃ to Al₂O₃ > 1) (2)

w(C₄AF) = 3.04w(Fe₂O₃) (mole ratio of Fe₂O₃ to Al₂O₃ < 1) (3)

w(C₂F) = 1.70[w(Fe₂O₃)-1.57w(Al₂O₃)] (4)

w(C₃A) = 2.65[w(Al₂O₃)-0.64w(Fe₂O₃)] (5)

w(f–CaO) = w(CaO)-1.10w(Al₂O₃)-0.70w(Fe₂O₃)-2.80w(SiO₂) (mole ratio of Fe₂O₃ to Al₂O₃ > 1) (6)

w(f–CaO) = w(CaO)-1.65w(Al₂O₃)-0.35w(Fe₂O₃)-2.80w(SiO₂) (mole ratio of Fe₂O₃ to Al₂O₃ < 1) (7)

w(C₃S), w(C₄AF), w(C₂F), w(C₃A), w(f–CaO)—the mass fraction of the calcium–based phases in the regenerated samples, wt.%;

w(SiO₂), w(Al₂O₃), w(Fe₂O₃), w(CaO)—the mass fraction of the regenerated samples’ composition, wt.%.

3.2. Microstructures of Regenerated Magnesia-Calcium Bricks

Figure 6(a) shows the SEM images of the regenerated samples’ surfaces. The white phases shown in the selected areas 1, 2 and 3 in Figure 6(a) were f–CaO, C₄AF or C₄AF–C₃A, and C₃S phases respectively, while the dark phases shown in Figure 6(a) were all MgO phase. Besides, it indicates that the phase of C₄AF and C₃A showed a tendency of aggregation in the sample A1.5 and A3.0. It is because that the elements of Fe and Al exist in the form of liquid phase at the firing temperature (1873 K), the C₃A phase (melting point of 1808 K [19]) and the C₄AF phase (melting point of 1688 K [22]) successively crystallized from the liquid phases, and sintered together during the furnace cooling process.

Figure 6(b) shows the SEM images of the regenerated samples’ fractures. It shows that the sample A1.5 and A3.0 both present much more compactness than the sample A0, and the calcium–based phases were more continuous along with the increase of Al₂O₃ addition. Table 3 shows that the increase of C₃A content was the main change among that of the calcium–based phases as Al₂O₃ addition increasing. Therefore, it can prove that the generation of C₃A improved the sinterability of the regenerated samples.

Figure 3 shows that there were too few of the C₃A phase to be observed in its character diffraction peak in the sample A1.5. It did not match the theoretical calculation as shown in Table 3. It is because that part of the Al element was dissolved into the C₃S phase, reducing generation of the C₃A phase [20]. The EDS analysis of the C₃S phase in Figure 6(a)–A1.5 is shown in Figure 7. It can be seen a significant solid solution appearing between the Al element and the C₃S phase.

3.3. Room Temperature and Hot Flexural Strength of the Regenerated Magnesia-Calcium Bricks

Figure 8 shows the results of the room temperature and hot (1573 K) flexural strength of the samples A0, A1.5 and A3.0. It indicates that the room temperature flexural strength of the regenerated samples enhanced, oppositely, the hot flexural strength of the regenerated samples reduced with Al₂O₃ addition increasing. It is due to the increasing of the low melting point phases content (C₄AF and C₃A) as shown in Table 3 that improved the sinterability of the regenerated samples as shown in Figure 6, and subsequently enhanced their bulk density as shown in Figure 8. Thus, it results in the opposite phenomenon between the room temperature strength and the hot flexural strength of the regenerated samples with Al₂O₃ addition increasing.

3.4. Resistance to Cement Clinker Corrosion of the Regenerated Magnesia-Calcium Bricks

The SEM images of the sample A0, A1.5 and A3.0 cross-section after corrosion resistance experiment are shown in Figure 9(a)–(c). It indicates that the cement clinker corrosion to the regenerated samples was gradually serious as the increasing of Al₂O₃ addition, and the cement clinker was more prone to corroding the regenerated samples along their calcium–based phases. It is because that the calcium–based phases of the regenerated samples were composed of the C₃S, C₄AF, C₃A and f–CaO phases, among which the phases of C₄AF and C₃A content increased with Al₂O₃ addition increasing. It improved the wettability of the regenerated samples, and thus, the melting cement clinker was much easier to corrode the regenerated samples along the areas of these low melt point phases [red arrow pointing out in Figure(a)–(c)] as they all exist in liquid phase form at 1823 K. Figure 9(d) shows an increasing tendency of the regenerated samples’ coating adherence along with the increase of Al₂O₃ addition. It is because that the reaction areas between the cement clinker and regenerated samples increase with additive Al₂O₃ increased, resulting in their strength of coating adherence enhanced.

The corrosion of the cement clinker to the phase of MgO was carried out by an elements distribution scanning analysis across their contact areas as shown in Figure 10(a). The analysis results indicate that a reaction layer containing element Ca, Si and Mg existed in the interfacial surface of these contact areas. The thickness d of the reaction layer gradually became larger with Al₂O₃ addition increasing as shown in Figure 10(b), which also indicates an aggravating corrosion of the cement clinker to the phase of MgO.

Generally, Al₂O₃ addition had a positive effect on the room temperature performances of the regenerated samples, such as the bulk density and the room temperature flexural strength, on the contrary, it had an adverse effect on the high-temperature performances of the regenerated samples, such as the hot flexural strength and corrosion resistance except of the strength of coating adherence, because of the low melting point phases of C₄AF and C₃A generating. It had been reported that the room temperature flexural strength and the strength of coating adherence of the chromium–free refractories applied on cement rotary kiln were lager than 21.60 MPa and 0.52 MPa [23], in comparison that of sample A0, A1.5, A3.0 were 68.58 MPa, 70.91 MPa, 93.11 MPa and 1.61 MPa, 2.25 MPa, 7.89 MPa accordingly. It shows that sample A0, A1.5, A3.0 could all meet the room temperature flexural strength and the strength of coating adherence. However, the hot flexural strength of sample A0, A1.5, A3.0 was 9.65 MPa, 4.34 MPa, 3.62 MPa respectively, and it shows that the hot flexural strength of sample A0 was only higher than that of the literature report (lager than 5.4 MPa) [24]. Therefore, the sample A0 can be considered as most suitable regenerated sample for cement rotary kiln, and the content of Al₂O₃ in the regenerated magnesia-calcium brick should not be higher than 1.1 wt.% (without extra Al₂O₃ addition).

4. Conclusions

The magnesia-calcium brick samples were prepared by using spent magnesia-calcium brick, fused magnesia and Al₂O₃ powder additive as the main raw materials, liquid paraffin as the binder, and firing at 1873 K for 2 h under an air atmosphere respectively, among which the Al₂O₃ powder addition was 0 wt.%, 1.5 wt.%, 3.0 wt.%. The main phase compositions of the sample A0 and A1.5 were MgO, f–CaO, C₃S, C₄AF, and a new phase C₃A was observed in the sample A3.0 besides MgO, f–CaO, C₃S, C₄AF. There was a solid solution appearance between the Al element and the C₃S phase, resulting in the actual content of the C₃A phase was lower than the theoretical calculation, and it led to the phase of C₃A being not observed in the sample A1.5. The room temperature performances of the regenerated samples such as bulk density and flexural strength enhance along with the increase of Al₂O₃ addition, since the low melting point phases of C₄AF and C₃A content increase which improve the sinterability of the regenerated samples. However, the high-temperature performances of the regenerated samples such as hot flexural strength (1573 K) and corrosion resistance (1823 K) degrade as the addition of Al₂O₃ increasing, because the melting points of the C₄AF and C₃A phases were close to 1573 K, furthermore, they were in liquid phase form at 1823 K that wetted the regenerates sample to enhance the reaction between the cement clinker and the MgO phase, resulting in the increasing of the strength of coating adherence. The sample A0 can be considered as most suitable regenerated sample for cement rotary kiln, as comparing the main performances with other chromium–free refractories for cement rotary kiln, that is, the total content of Al₂O₃ in the regenerated magnesia-calcium brick should be no more than 1.1 wt.%.