Introduction

Magnesium aluminate spinel refractories are widely used as important materials for heat equipment that requires long-term resistance to high temperature and chemical corrosion [1, 2]. However, as modern high-temperature industrial production technology proceeds, there is a need for higher requirements including good mechanical properties and especially good high-temperature creep resistance. This results in a shortage of magnesium aluminate spinel. The dissolution of corundum and the secondary spinelization reaction in a saturated spinel structure under a long-term high temperature and load changes the microstructure of the materials [3, 4]. There is then an increasing high-temperature creep of materials [5] and a decrease in the structural stability of the equipment. Therefore, the production safety was reduced.

Researchers have done many works to regulate the grain boundary of spinel. The addition of additives (such as SiO2, Fe2O3, and TiO2) can improve the sintering process and reduce the porosity and the pore size. This can increase the effective area of creep resistance and then improve the creep resistance of the materials at a high temperature. However, the resulting low melting point phase will reduce the high-temperature strength and service life of the materials [6,7,8,9,10,11]. If a high-temperature oxide containing a large-radius metal ion was incorporated in situ formation into the spinel matrix structure at a high temperature, the grain boundary diffusion can be inhibited and the spinel solid solution structure can be stabilized. Then, the breakthrough improvements in the high-temperature creep behavior of spinel refractories can be expected.

It is worth to note that the yttrium aluminum garnet (YAG) has been a good choice because of its high melting point, interesting mechanical properties, good high-temperature chemical stability, and what’s more, a low high-temperature creep rate (lowest of all known oxides 2.5 × 10−9/s at 1700 °C, 100 MPa) [12,13,14,15,16,17,18]. YAG was in situ synthesized in the spinel refractory matrix by using a fused spinel as the raw material and an Y2O3 micropowder as the additive in this work. The effect of Y2O3 addition on the microstructure and sinterability and the creep behavior of the spinel refractories were investigated.

Experimental

Material preparation

Fused spinel (Kaifeng Special Refractory Co., Ltd., China) was used as the raw material. Y2O3 micropowder (purity ≥ 99.9%, Hongde New Technology Development Co., Ltd., China) was used as the additive. The MgCl2 liquor (density is 1.25 g/cm3) was employed as the binder. The chemical compositions (wt%) that resulted from the fused spinel are Al2O3 (75.34), MgO (23.56), SiO2 (0.13), Fe2O3 (0.27), and CaO (0.52). Recipes were prepared by incorporating 0, 1, 2, and 3% Y2O3 micropowders by weight into fused spinel–based compositions that contained 45 wt%, 15 wt%, and 40 wt% with spinel grain sizes of 1–3 mm, 0–1 mm, and ≤ 0.088 mm, respectively. The added Y2O3 micropowders were to replace the spinel with a grain size of ≤ 0.088 mm. The recipes were marked as A, AY1, AY2, and AY3, respectively.

First, all the raw materials were weighed per the desired composition. Then, the additives and the different spinel materials were mixed using a ball mill for 6 h, respectively. Second, this mixture was grounded with fused spinel aggregates using an Eirich sand mixer for 10 min; the MgCl2 liquor was used as the binder. The ground mixture was then kneaded for 12 h and pressed at 150 MPa into molds (230 mm × 114 mm × 65 mm). All of these specimens were dried at 110 °C for 24 h and heated to 1750 °C and held for 6 h. The burned bricks were cut into bars (25 mm × 25 mm × 140 mm), cubes (40 mm × 40 mm × 40 mm), and cylinders with an inner hole (diameter 50 mm, height 50 mm, and inner hole diameter 12 mm) according to the test requirements, respectively.

Characterization

The bulk density and apparent porosity values were carried out in water using the Archimedes principle. The cold modulus of rupture (CMOR) values were measured according to the GB/T 3001-2007 standard by the three-point bending test method using a 100-mm span. The cold crushing strength (CCS) values were measured in accordance with the GB/T 5072-2008 standard. The hot MOR values were measured according to the GB/T 3002-2004 standard using the three-point bending test, which was performed at 1400 °C. The compressive creep rate was measured in accordance with the GB/T 5073-2005 standard, which was performed at 1500 °C for a load of 0.2 MPa. X-ray diffraction (XRD) analysis was performed using a Bruker AXS D8 Advance system (CuKα1, λ = 1.5406 Å), and scanning electron microscopy (SEM) imaging was performed using a JEOL JSM-6610 SEM system.

Results and discussion

Physical properties

The physical properties including the bulk density, apparent porosity, cold modulus of rupture, cold crushing strength, and hot modulus of rupture of the fired specimens are shown in Table 1. There was no obvious difference between the bulk densities of the samples. The apparent porosity decreased from 16.7 to 13.3%, and the CMOR, CCS, and hot modulus of rupture (HMOR) increased from 23.3 MPa, 89 MPa and 5.0 MPa to 25.3 MPa, 112 MPa, and 10.9 MPa, respectively. The values can be divided into two levels of Y2O3 change. From 0 to 2 wt%, the apparent porosity decreased, and the bulk density, CMOR, and CCS increased due to an increase in the formation amount of YAG by the reaction of Y2O3 with a solid solution of Al2O3 in spinel. This accelerated the mass transfer and densification processes. From 2 to 3 wt%, however, these values changed into the opposite. This was because the reaction Y2O3 + Al2O3 → Y3Al5O12 was associated with a volume shrinkage of about 0.3%, and additional new micropores were produced when more YAG was present. This affected the spinel sintering.

Table 1 The physical properties of the samples
Full size table

High-temperature creep resistance

The fitted and predicted creep strain/time curves are shown in Fig. 1, and the curves were fitted according to Eq. (1), which was called the θ-Concept Project by Evans R W and Wilshire B [19]. The fitted data are shown in Table 2. The creep strains of the specimens were all quite small, and the creep rates decreased upon the addition of Y2O3. As shown in Fig. 1, the sample without Y2O3 addition has a rising shrinkage as the time prolongs, while the samples added with Y2O3 have a smaller shrinkage. Their creep strains for 50 h of testing were − 0.38%, − 0.25%, − 0.08%, and − 0.26%, respectively. It also can be seen from the predicted creep strain/time curves that the specimens added with Y2O3 reached stable creep stage earlier than the specimen without Y2O3.

$$ {\varepsilon}_{\mathrm{t}}={\theta}_1\left[1-\exp \left(-{\theta}_2t\right)\right]+{\theta}_3\left[\exp \left({\theta}_4t\right)-1\right] $$
(1)

where εt is creep strain, t is the time, and θ1, θ2, θ3, and θ4 are creep parameters.

Fig. 1
figure 1

The fitted creep curves

Full size image
Table 2 Equation of the fitted curves
Full size table

Generally, the earlier the turning point of deceleration creep appears, the shorter the time of deceleration creep is, the smaller the deformation of the material is, and the longer the service life of the material is [20]. As shown in Fig. 1, the samples added with 2 wt% and 3 wt% Y2O3 reached the stable creep stage at 10 h and 40 h, respectively, whereas, the samples added with 0 wt% and 1 wt% Y2O3 did not get the stable stage even in the duration time of 50 h. The decrease of the compressive creep rate is due to the in situ formation of the second low-creep phase YAG.

Phase composition and microstructure analysis

Figure 2 shows the XRD patterns of the samples added with 0 and 2 wt% Y2O3 (the diffraction peaks of the samples added with different amounts of Y2O3 are almost the same). Spinel is the main phase in the sample without Y2O3; spinel and YAG are the main phases in the sample added with 2 wt% Y2O3, which indicates that YAG can be formed via the reaction of Y2O3 with a solid solution of Al2O3 in spinel.

Fig. 2
figure 2

Results of X-ray patterns of specimens added with 0 and 2 wt% Y2O3

Full size image

Figure 3 shows the microstructure of the samples with different amounts of Y2O3 after sintering and Y element distribution of the sample with 3 wt% Y2O3. There were many pores in the specimen without Y2O3 as shown in Fig. 3a. Instead, a lot of high-temperature continuums (the light-gray and white fields) form uniformly around the spinel grain boundary in the specimen with Y2O3, as shown in Fig. 3b, c. This forms an interlocking structure with the spinel grain, which can stabilize the grain boundary of spinel. In addition, the content of pores obviously declines. In Fig. 3b, the EDAX results indicate that the constituent elements in the light-gray fields were mostly Al, O, and a small amount of Mg, Y, and Ca (Ca came from fused spinel raw materials). These light-gray fields are the precipitation of α–Al2O3 from spinel, as well as a little amount of Mg–Al–Y solid solution [17]. Figure 3 d shows that Y element distributes on the boundary of the spinel grains according to Fig. 3c. Combining the XRD and EDS results, it can be concluded that the white fields are YAG. This indicates that the addition of Y2O3 results in the expulsion of pores and the formation of YAG had a pinning effect on spinel grain boundary, which improves the mechanical properties and high-temperature creep resistance of spinel refractories.

Fig. 3
figure 3

Microstructure of the samples with different amounts of Y2O3 after sintering: a 0 wt%, b 1 wt%, c 3 wt%, and d Y element distribution of the sample with 3 wt% Y2O3

Full size image

The microstructure evolution and interface reaction mechanism

According to the XRD and FESEM results, the matrix structure evolution and interface reaction mechanism of spinel refractories upon Y2O3 addition can be proposed, as shown in Fig. 4. The solid reaction and sintering are all controlled by ion diffusion at a high temperature. The counter-diffusion of Al3+ and Y3+ cations occurs at the interface of the spinel and Y2O3. The added Y3+ is preferentially dissolved into the spinel lattice by substituting Al3+ to form a small amount of the Mg–Al–Y solid solution. When the added Y2O3 content exceeds the limit of solubility, the Y2O3 can react with spinel to extract Al3+ from the spinel structure and form YAG [21,22,23]. During the cooling process, the second low-creep phase YAG precipitates on the spinel grain boundary, forming an interlocking structure that can stabilize the grain boundary of spinel and prevent grain-boundary sliding. This approach improves the high-temperature creep resistance of spinel refractories.

Fig. 4
figure 4

Schematic illustration for the matrix structure evolution and interface reaction mechanism of spinel refractories with the Y2O3 addition

Full size image

Conclusions

A low-creep spinel refractory can be prepared by doping Y2O3 micropowders, the physical properties, high-temperature creep resistance, and microstructure evolution and interface reaction mechanism of the refractories were investigated. The conclusions can be drawn as follows:

  1. 1.

    The mechanical properties, especially the hot modulus of rupture of the spinel refractories were significantly improved by the added 2 wt% Y2O3.

  2. 2.

    A second low-creep phase YAG could be in situ formed with the addition of Y2O3 into a fused spinel. This resulted in a great decrease of the compressive creep strain from 0.36 to 0.08%.

  3. 3.

    The matrix structure evolution and interface reaction mechanism of spinel refractories upon Y2O3 addition were proposed. The MgYxAl2−xO4 solid solution and YAG would be formed in turn with the increasing addition of Y2O3.