Effect of Sc₂O₃ addition on densification and microstructure of different spinelized magnesium aluminate spinels

Jie LI, Jian-hua LIU, Xiao-jun , Yan-qing LAI

School of Metallurgy and Environment, Central South University, Changsha 410083, China

Received 25 January 2015; accepted 4 June 2015

Abstract: The densification and microstructure of different spinelized magnesium aluminate spinels (MAS) were studied adding Sc₂O₃ as additive. Sintered products were then characterized in terms of densification, phase analysis, quantitative elemental analysis and microstructure. The results show that Sc₂O₃ is found to be beneficial for the densification of MAS. Sc₂O₃ has a more significant effect on the densification of partially spinelized MAS batch than that of fully spinelized MAS batch. At the sintering temperature of 1650 °C, the bulk density of sintered products of partially spinelized powders increases by 0.243 g/cm³ as the Sc₂O₃ content increases from 0 to 4% (mass fraction) and that of fully spinelized powder increases by 0.14 g/cm³. Compared with the sintered samples prepared from the fully spinelized powder, the sintered samples using the partially spinelized powders as raw materials have more compact microstructures.

Key words: magnesium aluminate spinel; Sc₂O₃; densification; microstructure; elemental analysis

1 Introduction

Magnesium aluminate spinel (MAS) features a high melting point (2135 °C), high hardness (16 GPa), low thermal expansion, good chemical stability, and good resistance to thermal spalling and corrosion [1-4]. MAS has been used for various applications, such as refractories [5,6], humidity sensors [7] and transparent ceramic materials [8–11]. MAS has also been utilized as anode materials [12] and man-made ledge materials [13] in aluminum electrolytic cells. Nevertheless, MAS cannot be densified through single-stage firing because its volume expands as high as 8% in the solid-state reaction. Thus, MAS is densified through a double-stage firing process, in which alumina and magnesia are initially calcined at above 1500 °C to produce MAS powders with the desired spinel phase (>80%); the powders are then ground, compacted and sintered at above 1900 °C [14]. The formation temperature of the MAS phase is high in the first stage and the agglomeration of the spinel phase is serious [15]. The calcination temperature of the solid-state reaction is controlled to obtain different spinelization degrees and sintering activities of MAS powders in the first stage, which considerably affects the subsequent sintering process.

The spinelization of the MAS powder is not only related to the formation temperature, but also to the added amount of spinelizer. GANESH et al [16] studied the effect of AlCl₃ on the synthesis and densification of MAS through double-stage firing. The result showed that the spinelization of spinel was approximately 82% without any additive and increased to 100% with the addition of 0.03% AlCl₃ (mole fraction). They also found that AlCl₃ exhibited optimal performance on the densification of partially spinelized MAS powder. These findings indicate that the influence of additive on the sintering behavior varies in different spinelized MAS samples. Over the recent years, researchers have reported the effect of additives on the densification of MAS [17-20]. Adding LiF could enhance the densification of spinel through liquid-phase sintering [17]. The addition of TiO₂ also enhanced the densification of spinel through exsolution of Al₂O₃ and dissolution of TiO₂ [18]. In the case of ZnO addition, the formation of anion vacancy could induce the densification of spinel [19]. Satisfactory densified spinel products were obtained after addition of different rare earth oxides. Dy₂O₃ prevented the exaggerated grain growth and thereby contributed to the densification process [20]. Nevertheless, these studies did not evaluate the effect of sintering additives on the densification of different spinelized MAS samples.

Rare earth oxides can be easily converted to a solid solution with sintered products and then increase the lattice defects [21]. Hence, these oxides can also promote the densification of MAS. In addition, the difference between rare earth cation ionic radius and the main lattice ionic radius directly affects the densification of MAS. The ion radius of Sc³⁺ is the smallest compared with that of the other rare earths, and Sc³⁺ exhibits similar characteristics to Al³⁺. In view of these findings, a systematic study was performed to investigate the influence of Sc₂O₃ addition on the densification and microstructure of different spinelized MAS batches prepared at 1200 and 1400 °C from magnesia and alumina (mole ratio of MgO to Al₂O₃, 1:1). Different spinelized batches were sintered at 1550, 1600 and 1650 °C. The final sintered products were characterized through X-ray fluorescence (XRF), X-ray diffraction (XRD), scanning electron microscopy (SEM), and energy-dispersive analysis of X-ray (EDAX) analyses to determine the role of Sc₂O₃ addition in MAS.

2 Experimental

The physicochemical properties of the starting oxides are listed in Table 1, which shows that the purity values of alumina and magnesia are 97.286% and 99.899%, respectively.

Batch composition was made from the raw materials in such a way that mole ratio of MgO to Al₂O₃ was 1:1. All batches were milled for 3 h using a planetary ball mill. Milled powders were calcined at a heating rate of 3 °C/min and 1200 and 1400 °C for 3 h, respectively. The spinels prepared at 1200 and 1400 °C were termed as LS batch and HS batch, respectively. The degree of conversion to MAS was determined by the conventional acid dissolution method [22].

Sc₂O₃ (99.99%, w(SiO₂)<0.005%, w(CaO)<0.012%, w(Fe₂O₃)<0.05%) was added at 1%, 2% and 4% (mass fraction) into the LS and HS batches, respectively. All batches were individually milled for 1 h. Afterwards, the milled powders were uniaxially pressed at 200 MPa on a hydraulic press (WE–300C) into d20 mm × 15 mm cylindrical samples. The pressed samples were then sintered at 1550, 1600 and 1650 °C for 3 h. Firing was conducted in a program-controlled electric furnace (KSL-1750X) and the heating rate was maintained at 3 °C/min.

The bulk density and apparent porosity of the sintered products were measured using the conventional liquid displacement method, which applied Archimedes’ principle. Phase analysis was conducted via the X-ray diffraction technique. The diffraction patterns of the finely powdered samples were obtained with an X-ray diffractometer using nickel-filtered Cu K_α radiation. The lattice parameter (a) was calculated from the characteristic XRD peaks of the spinel phase using the following equation [23]:

(1)

where a is the lattice parameter; λ is the radiation wavelength; θ is the Bragg’s angle; and h, k and l are the corresponding Miller indices. The microstructures of the sintered products were observed with a scanning electron microscope (SEM, Nova NanoSEM 230) using sputtered gold coating on the polished surface after thermal etching. Quantitative elemental analysis of the selected grain was done by EDAX.

3 Results and discussion

3.1 Characterization of raw materials

The physicochemical properties of the calcined powders prepared at 1200 and 1400 °C are given in Table 2. The XRD patterns (Fig. 1) of the calcined powders indicate the presence of the spinel, periclase and corundum phases for the LS batch, whereas only the spinel phase is observed for the HS batch. The degree of conversion to MAS calcined at 1200 °C is 93.1%. The starting raw materials were converted completely into MAS for the HS batch, which indicated an increase in the conversion degree to MAS with the increase in calcination temperature. Table 2 shows that the true density of the LS batch is the same as that of the HS batch, which is 3.58 g/cm³. However, the average particle size of the HS batch is 65.79 μm, which is significantly larger compared with that of the LS batch (23.76 μm). The difference may be attributed to the higher calcination temperature of MAS, which results in the formation of large agglomerates, thereby inhibiting sintering activity upon subsequent densification operation.

Table 1 Physicochemical properties of magnesia and alumina

Table 2 Physicochemical properties of calcined powders

Fig. 1 XRD patterns of LS (a) and HS (b) batches

3.2 Densification

The variations in bulk density and apparent porosity of the sintered products of the LS and HS batches (termed as S_LS and S_HS, respectively) with sintering temperature and Sc₂O₃ content are shown in Figs. 2 and 3, respectively.

A general trend of increasing bulk density and decreasing apparent porosity was observed with the increase in sintering temperature at the same Sc₂O₃ content. This trend can be explained by the great driving force for sintering at the high sintering temperature. A general trend of increasing bulk density and decreasing apparent porosity was also observed in Figs. 2 and 3 with the increase in Sc₂O₃ content at the same sintering temperature. Sintering at 1650 °C was not sufficient to obtain high densification in the samples without an additive; however, increasing the Sc₂O₃ content to 4% (mass fraction) results in a densification of 96.17%. This result may be attributed to the fact that Sc³⁺ with characteristics similar to Al³⁺ can substitute Al³⁺ in the spinel. Moreover, incorporating Sc³⁺ into the spinel structure causes strain in the lattice (because of ionic size variation), which improves diffusion, mass transfer and densification.

The S_LS samples exhibit better sintering characteristics than the S_HS samples at the same sintering temperature and Sc₂O₃ content. This result may be attributed to the fact that the average grain size of the LS batch (23.76 μm) is smaller than that of the HS batch (65.79 μm). In addition, free magnesia and alumina of the LS batch can react to produce MAS during sintering. The formation of a new spinel can be considered a seed that acts as a core for the grain growth of the spinel.

Fig. 2 Variations in bulk density (a) and apparent porosity (b) of S_LS samples with sintering temperature and Sc₂O₃ content

Fig. 3 Variations in bulk density (a) and apparent porosity (b) of S_HS samples with sintering temperature and Sc₂O₃ content

Upon sintering at 1650 °C, the increasing amplitude of the bulk density of S_LS, with the increase in Sc₂O₃ content is greater than that of S_HS. Moreover, the decreasing amplitude of apparent porosity of S_LS is also larger than that of S_HS. With the increase in Sc₂O₃ content from 0 to 4%, the bulk density of S_HS increases by 0.14 g/cm³, whereas its apparent porosity decreases by 3.6%. However, the bulk density of S_LS increases by 0.243 g/cm³ and its apparent porosity decreases by 7.2%. The higher densification of the S_LS sample sintered at 1650 °C may be largely determined by the combined effects of the newly formed spinel and Sc₂O₃ addition.

3.3 XRD patterns

Figure 4 shows the XRD patterns of the S_LS and S_HS samples with different Sc₂O₃ contents sintered at 1650 °C. Sc₂O₃ contents were 0, 1%, 2% and 4%, respectively. These samples are termed as S_LS0, S_LS1, S_LS2, S_LS4 and S_HS0, S_HS1 S_HS2, S_HS4, respectively. The phase analysis study of all the sintered products, with and without additives, shows only the spinel phase. No periclase and corundum phases are observed in all the samples, which confirms the consumption of Al₂O₃ and MgO in spinel phase formation.

Fig. 4 XRD patterns of S_LS (a) and S_HS (b) samples with different Sc₂O₃ contents sintered at 1650 °C

Figure 5 shows the XRD 2θ scans of S_LS and S_HS with different Sc₂O₃ contents sintered at 1650 °C. The 2θ values of the peak from the (311) plane of the spinel in the sintered products of S_LS and S_HS decrease with the increase in Sc₂O₃ content. These values are 36.98°, 36.90°, 36.90°, 36.84° and 36.96°, 36.82°, 36.82°, 36.72° for batches S_LS0, S_LS1, S_LS2, S_LS4 and S_HS0, S_HS1 S_HS2, S_HS4, respectively.

In order to further investigate the crystal structure of the spinel, the lattice parameters were calculated using Eq. (1). The results are presented in Table 3. The lattice parameter of MAS increases with the increase in Sc₂O₃ content at the same sintering temperature. The lattice parameter of the MAS for S_HS samples increases from 0.806 to 0.811 nm with the increase in Sc₂O₃ content from 0 to 4% (mass fraction). This increase may be attributed to the fact that Sc₂O₃ makes a solid solution with the spinel. The ionic radius of Sc³⁺ is 0.083 nm [24] and that of aluminum ion is 0.05 nm [25]. Given its higher ionic radius, incorporating Sc³⁺ into the spinel structure expands the lattice, which increases lattice parameters.

Fig. 5 XRD 2θ scans of S_LS (a) and S_HS (b) samples with different Sc₂O₃ contents sintered at 1650 °C

Table 3 Lattice parameters (a) of S_LS and S_HS samples with different Sc₂O₃ contents sintered at 1650 °C

3.4 SEM images and EDAX patterns

The SEM images and the EDAX patterns of the polished and thermally etched samples (S_LS0, S_LS4, S_HS0 and S_HS4) are shown in Fig. 6. Figures 6(a₁) and (c₁) show an inhomogeneous grain growth for the S_LS0 and S_HS0 samples. The spinel grain shape is angular with a clear grain profile. The average grain size of S_HS0 sample is considerably smaller than that of S_LS0 sample. Thus, the formation of a new spinel favors grain growth during sintering. Figure 6 shows that the microstructures of S_LS4 and S_HS4 are more compact with a minimal amount of pores, which is responsible for the higher densification of 4% Sc₂O₃-containing samples.

Figure 6 also shows that the major elements of all the sintered products include Mg, Al, and O. In addition, S_LS4 and S_HS4 contain Sc. The quantitative elemental analysis results of various samples by EDAX were converted into oxide contents and presented in Table 4. S_LS0 and S_HS0 have a nearly stoichiometric spinel grain without Sc₂O₃, as shown in Table 4. The Sc₂O₃ additive is mainly presented in the spinel grain, which indicates the solid solubility of Sc₂O₃ in the spinel and verifies that Sc³⁺ ions have replaced Al³⁺ ions from the spinel grains.

Table 4 Chemical compositions of spinel grain in various samples by EDAX

S_LS0, S_LS4, S_HS0 and S_HS4 samples were characterized in terms of chemical compositions by XRF and shown in Table 5. 2.37% Sc₂O₃ measured by XRF is nearly the same as 2.15% Sc₂O₃ measured by EDAX for S_LS4. This finding indicates that Sc₂O₃ is uniformly distributed in the spinel grains of S_LS4. Sc₂O₃ is inhomogeneously distributed in the spinel grains of S_HS4 according to the difference in Sc₂O₃ content between 22.69% (by EDAX) and 5.21% (by XRF). This result may be ascribed to the fact that Sc₂O₃ dissolves firstly in MgO and then precipitates gradually from MgO during the formation of the spinal [26]. Thus, Sc₂O₃ is uniformly distributed throughout the spinel grains of the S_LS sample. This distribution may be the reason why the bulk density of S_LS4 is greater than that of S_HS4.

Table 5 Chemical compositions of S_LS0, S_LS4, S_HS0 and S_HS4 samples by XRF

Fig. 6 SEM images and EDAX patterns of S_LS0 (a₁, a₂), S_LS4 (b₁, b₂), S_HS0 (c₁, c₂) and S_HS4 (d₁, d₂) samples

4 Conclusions

1) Incorporating Sc₂O₃ into magnesium aluminate spinel favors the densification and up to 96.17% of true density can be achieved at 1650 °C. The Sc³⁺ of Sc₂O₃ can substitute the trivalent Al³⁺ of the spinel structure at a high sintering temperature. Incorporating of Sc³⁺ into the spinel structure generates the lattice strain and improves mass transfer and densification.

2) The densification of partially spinelized powder is higher than that of the fully spinelized powder at the same sintering temperature and Sc₂O₃ content.

3) Quantitative elemental analysis by EDAX shows that the Sc₂O₃ additive is mainly present in the spinel grains, which confirms the solid solubility of Sc₂O₃ in the spinel. The combined SEM and XRF analysis results show that Sc₂O₃ is uniformly distributed throughout the spinel grains of the S_LS sample, but inhomogeneously distributed throughout the spinel grains of the S_HS sample.

References

[1] MOHAMMADI F, OTROJ S, NILFORUSHAN M R. Effect of MgCl₂ addition on the sintering behavior of MgAl₂O₄ spinel and formation of nanoparticles [J]. Science of Sintering, 2014, 46(2): 157-168.

[2] KHALIL N M, HASSAN M B, EWAIS M M, SALEH F A. Sintering, mechanical and refractory properties of MA spinel prepared via co-precipitation and sol-gel techniques [J]. Journal of Alloys and Compounds, 2010, 496(1): 600-607.

[3] SARKAR S, SAHOO S. Effect of raw materials on formation and densification of magnesium aluminate spinel [J]. Ceramics International, 2014, 40(10): 16719-16725.

[4] L, MATIAS T, A M, CAMPOS J, PORTUDAL A. MgAl₂O₄ spinel synthesis by combustion and detonation reactions: A thermochemical evaluation [J]. Journal of European Ceramic Society, 2012, 32(12): 3161-3170.

[5] SARKAR S. Refractory applications of magnesium aluminate spinel [J]. Interceram: Refractories Manual, 2010(1): 11-14.

[6] MORITZ K, ANEZIRIS C G, HESKY D. Magnesium aluminate spinel ceramics containing aluminum titanate for refractory applications [J]. Journal of Ceramic Science Technology, 2014, 5(2): 125-130.

[7] KLYM H, INGRAM A, HADZAMAN I, SHPOTYUK O. Evolution of porous structure and free-volume entities in magnesium aluminate spinel ceramics [J]. Ceramics International, 2014, 40(6): 8561-8567.

[8] GOLDSTEIN A. Correlation between MgAl₂O₄ and spinels structure, processing factors and functional properties of transparent parts (Progress review) [J]. Journal of European Ceramic Society, 2012, 32(11): 2869-2886.

[9] RAMISETTY M, SASTRI S, KASHALIKAR U, GOLDMAN L M, NAG N. Transparent polycrystalline cubic spinels: Protect and defend [J]. American Ceramic Society Bulletin, 2013, 92(2): 20-25.

[10] BORRERO L O, ORTIZ A L, GLEDHILL A D, GUIBERTEAU F, MROZ T, GOLDMAN L M, PADTURE N P. Microstructural effects on the sliding wear of transparent magnesium aluminate spinel [J]. Journal of European Ceramic Society, 2012, 32(12): 3143-3149.

[11] LALLEMANT L, PETIT J, LALANNE S, TROMBERT S, VEMHET L, VIROULOAD R. Modeling of the green body drying step to obtain large size transparent magnesium aluminate spinel samples [J]. Journal of European Ceramic Society, 2014, 34(3): 791-799.

[12] ANGAPPAN S, BERCHMANS L J, AUGUSTIN C O. Sintering behavior of MgAl₂O₄—A prospective anode material [J]. Materials Letters, 2004, 58(11): 2283-2289.

[13] Xiao-jun, LAI Yan-qing, ZHANG Chao, LI Jie. A man-made ledge aluminum electrolytic cell: China Patent, 103060848A [P]. 2013-04-24. (in Chinese)

[14] GANESH I. A review on magnesium aluminate (MgAl₂O₄) spinel: Synthesis, processing and applications [J]. International Materials Reviews, 2013, 58(2): 63-112.

[15] GANESH I, OLHERO S M, REBELO A H, FERREIRA J M F. Formation and densification behavior of MgAl₂O₄ spinel: Influence of processing parameters [J]. Journal of American Ceramic Society, 2008, 91(6): 1905-1911.

[16] GANESH I, BHATTACHARJEE S, SAHA B P, JOHNSON R, MAHAJAN Y R. A new sintering aid for magnesium aluminate spinel [J]. Ceramics International, 2001, 27(7): 773-779.

[17] ROZENBURG K, REIMANIS I E, KLEEBE H J, COOK R L. Chemical interaction between LiF and MgAl₂O₄ spinel during sintering [J]. Journal of American Ceramic Society, 2007, 90(7): 2038-2042.

[18] YU J, HIRAGUSHI K. Sintering behavior of spinel with added TiO₂ [J]. Taikabutsu Overseas, 1996, 16(4): 61-63.

[19] GHOSH A, DAS S K, BISWAS J R, TRIPATHI H S, BANERJEE G. The effect of ZnO addition on the densification and properties of magnesium aluminate spinel [J]. Ceramics International, 2000, 26(6): 605-608.

[20] TRIPATHI H S, SINGLA S, GHOSH A. Synthesis and densification behavior of magnesium aluminate spinel: Effect of Dy₂O₃ [J]. Ceramics International, 2009, 35(6): 2541-2544.

[21] SARKAR R, TRIPATHI H S, GHOSH A. Reaction sintering of different spinel compositions in the presence of Y₂O₃ [J]. Materials Letters, 2004, 58(16): 2186-2191.

[22]BOLDIZAROVA E. Mechanochemical synthesis of magnesium aluminate spinel in oxide–hydroxide systems [J]. Journal of Materials Synthesis and Processing, 2000, 8(5-6): 287-293.

[23] CULLITY B D. Elements of X-ray diffraction [M]. 2nd ed. London: Addison-Wesley, 1978: 351.

[24] YIN Zhi-min, PAN Qing-lin, JIANG Feng, LI Hang-guang. Scandium and its alloys [M]. Changsha: Central South University Press, 2007: 8-9. (in Chinese)

[25] SARKAR R, DAS S K, BANERJEE G. Effect of addition of Cr₂O₃ on the properties of reaction sintered MgO–Al₂O₃ spinel [J]. Journal of European Ceramic Society, 2002, 22(8): 1243-1250.

[26] HENRIKSEN A F, KINGERY W D. The solid solubility of Sc₂O₃, Al₂O₃, Cr₂O₃, SiO₂ and ZrO₂ in MgO [J]. Ceramurgia International, 1979, 5(1): 11-17.

Sc₂O₃对不同尖晶石化镁铝尖晶石的致密化和显微组织的影响

李 劼，刘建华，吕晓军，赖延清

中南大学 冶金与环境学院，长沙 410083

摘  要：探讨添加Sc₂O₃对不同尖晶石化镁铝尖晶石的致密化和显微组织的影响。通过致密化研究、物相分析、元素定量分析和显微组织研究对烧结产物的性能进行表征。结果表明：添加Sc₂O₃有利于镁铝尖晶石的致密化，且相对于完全尖晶石化的镁铝尖晶石，Sc₂O₃对促进部分尖晶石化的镁铝尖晶石的烧结作用更显著。在1650 °C烧结，Sc₂O₃添加量从0增加到4% (质量分数) 时，部分尖晶石化原料的烧结产物的体积密度增加了0.243 g/cm³，完全尖晶石化原料的烧结产物的体积密度增加了0.14 g/cm³。与用完全尖晶石化粉末制备的烧结试样相比，以部分尖晶石化粉末为原料的烧结试样样具有更加致密的显微组织。

关键词：镁铝尖晶石；Sc₂O₃；致密化；显微组织；元素分析

(Edited by Wei-ping CHEN)

Foundation item: Project (51374240) supported by the National Natural Science Foundation of China; Project (2012BAE08B02) supported by the National Science and Technology Pillar Program of China

Corresponding author: Xiao-jun ; Tel: +86-731-88830649; E-mail: csuxiaojun@126.com

DOI: 10.1016/S1003-6326(16)64099-2