Thermal shock resistance of porous ZrB 2 –SiC ceramic after oxidation

Abstract

Thermal shock water quenching tests on porous ZrB₂–SiC (ZS) ceramics with 8.7, 17.9 and 27.3% porosity were conducted after these specimens were oxidised in static air at a constant temperature of 1200°C for 30 min. The effect of oxidation on their thermal shock resistance was investigated in detail. The oxidation improved the material's thermal shock fracture resistance, but reduced their thermal shock damage resistance (TSDR). The thermal shock critical temperature difference (ΔT_c) of the oxidised porous ZS ceramics decreased as the porosity increased, which was the opposite trend to that observed in the case of unoxidised porous ZS ceramics. ZS ceramics with higher porosity had higher TSDR with or without oxidation treatment. Thickness of the oxidation layer, degree of oxide layer coverage and internal oxidation were observed to affect the thermal shock behaviour of the porous ZS ceramics.

Introduction

Porous ceramics with the excellent properties of ceramics and porous materials have many applications, such as supports for catalytic reactions, high-temperature thermal insulation and filtration [1]. Recently, pores have been introduced into ceramics to tailor their physical properties and to improve their thermal shock resistance [2-6]. Scholars have already tried to introduce pores into some ultra-high-temperature ceramics such as SiC [2], Si₃N₄ [3], ZrB₂ [5] and ZrB₂–SiC (ZS) [6], which significantly improved their thermal shock resistance. However, under high-temperature conditions necessary for many applications, oxidation of the material is inevitable. The oxidation greatly influences the performance of these porous ceramics [79]. Many scholars have investigated the oxidation and pre-oxidation of dense and porous ceramics to evaluate the effect on thermal shock resistance of these ceramics with variable porosities [10].

ZS is a promising ceramic for a variety of high-temperature structural applications. In this research, the thermal shock behaviour, including thermal shock fracture resistance (TSFR) and thermal shock damage resistance (TSDR), of the oxidised porous ZS ceramics with varying porosities is investigated. These results provide insights into the design, preparation and application of non-oxide porous ceramics for use in actual oxygen-containing high-temperature environments.

Experimental procedures

Porous ZrB₂–20vol.-%SiC ceramic samples with different porosities were prepared by the partial hot-pressing method: hot pressing a powder mixture to a definite dimension/height by designing the stroke of punch and the porosity being easily adjusted by the powder amount, whose details were described elsewhere [6]. Ceramics with different porosities (8.7, 17.9 and 27.3%) were designated as ZSA, ZSB and ZSC, respectively. The bulk density and total porosity was measured by the Archimedes technique. The open porosity was measured according to the international standard (ISO5017).

The 3 × 4 × 36 mm specimens with the tensile surface perpendicular to the hot-pressing direction were cut by means of electric discharge machining from the same billet, and then polished to 5 μm using diamond abrasives, which were used for the three-point flexural strength test, oxidation experiment and thermal shock test. The edges of all specimens were chamfered to minimise the effect of stress concentration due to machining flaws. The specimens of ZSA, ZSB and ZSC after the oxidation experiment were designated as ZSA_o, ZSB_o and ZSC_o, respectively. The oxidation experiment was carried out in static air in a Muffle furnace (GXL-15-25, RiXin High Temperature Technology Co. Ltd, HeFei, China). The specimens were supported on a zirconia crucible and heated at the rate of 25°C min⁻¹ from room temperature to 1200°C. The holding time for the target temperature was 30 min followed by furnace cooling to room temperature. The thermal shock test was performed by quenching the specimens into a water bath (25°C) from different high temperature (325, 425, 475, 500, 525, 625 or 725°C). The measured thermal shock critical temperature difference (ΔT_c) value is defined as the temperature difference corresponding to 70% of the room temperature strength. Flexural strength and residual flexural strength were determined using three-point bending tests (ASTMC1161). After the thermal shock and flexural strength tests, the relationship curve between residual flexural strength and quenching temperature difference was drawn for thermal shock behaviour analysis. The value of strength retention ratio (SRR) was calculated at a certain quenching temperature difference from which each material's strength began to decrease sharply. At least five specimens were tested for each experimental condition. Microstructures were observed by scanning electron microscopy (SEM, FEI Quanta 200F, U.S.A.) equipped with energy dispersive spectroscopy (EDS, EDAX Inc., U.S.A.) for chemical analysis. The chemical composition of oxidation layer was characterised using X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific, U.S.A.) with monochromatic Al Ka radiation.

Results and discussion

Figure 1 presents the morphology of the surface and cross-section of ZSA, ZSB and ZSC ceramics. No obvious difference was observed in their pore structures: the average grain size and pore size being almost identical. There only existed porosity difference among them, which might lead to difference in their oxidation resistance. Figure 2 presents the morphology of the surface and cross-section of the ZSA_o, ZSB_o and ZSC_o. Many areas of these specimens’ surfaces were observed to be covered with oxidation products after oxidation treatment and the corresponding EDS spectra of the surface of the ZSB_o is shown in Figure 3. In previous investigations, the oxidation products were reported to be SiO₂, B₂O₃ and ZrO₂ for the dense ZS ceramic (almost without pores) [8]. EDS analyses herein are consistent with previous observations [8] and also confirm that the formed glass phase was mainly composed of SiO₂ and B₂O₃ (see location 1 in Figure 3). The existence of B–O bond was confirmed by XPS further and not presented here. The porosity difference among the ZSA, ZSB and ZSC did not affect the composition but the amount of the oxidation products therefore affected the surface coverage degree and the thickness of the oxidation layers formed on the specimens’ surfaces. The complete surface coverage was observed for ZSA_o. The increased porosity led to a decrease in the degree of oxide layer coverage and the degree followed the sequence ZSA_o > ZSB_o > ZSC_o.

Figure 1

SEM micrographs of the surface and the cross-section of porous ZS ceramics before oxidation: (a, b) ZSA; (c, d) ZSB and (e, f) ZSC.

Figure 2

SEM micrographs of the surface and the cross-section of porous ceramics after oxidation at 1200°C for 30 min: (a, b) ZSA_o; (c, d) ZSB_o; (e, f) ZSC_o.

Figure 3

SEM micrograph and EDS analysis of the surface of ZSB_o.

Table 1 lists the properties of porous ZS ceramics before and after oxidation. The oxidation layer thickness values of the ZSA_o, ZSB_o and ZSC_o in the table were average of the test results obtained from the SEM determination. The thickness of the oxidation layer was determined by the material microstructure (porosity) and the diffusion of oxygen and gaseous products. On the one hand, the pores on the material surface or within the material could provide channels for the diffusion; on the other hand, once a dense oxidation layer generated on material surface, oxidation would not continue to develop inside [11]. As shown in Figure 2(a,c), after 30 min exposure to the air in the furnace under 1200°C, the surfaces of both the ZSA and ZSB produced relatively denser oxidation layers compared with that of the ZSC and no oxygen could diffuse inward through the dense oxidation layers, which might lead to their similar oxidation layer thickness. The oxidation was more serious when the porosity increased to 27.3% and the influence of porosity on the oxidation became more pronounced compared with that of the ZS with 8.7 and 17.9% porosity.

Table 1

Properties of porous ZS ceramics before and after oxidation.

	ZSA	ZSAo	ZSB	ZSBo	ZSC	ZSCo
Oxidation layer thickness (μm)	–	23 ± 0.2	–	25 ± 0.2	–	35 ± 0.3
Strength (MPa)	478.7 ± 55.7	644.9 ± 25.5	303.9 ± 48.7	427.9 ± 20.2	207.6 ± 49.3	289.7 ± 22.0
Bulk density (g cm−3)	5.03	–	4.53	–	4.01	–
Total porosity (%)	8.7	–	17.9	–	27.3	–
Open porosity (%)	1.6	–	16.9	–	24.7	–
Measured ΔTc (°C)	333.6	456.4	358.0	458.3	361.2	423.3
SRR (%)	6.46	15.11	10.72	23.66	30.42	35.06

The flexural strength of porous ZS before and after oxidation is tabled in Table 1. The flexural strength of ZAS, ZSB and ZSC ceramics were all improved by the oxidation owing to surface flaws healing as did dense ZS. It has been reported in some literatures [8-11] that for ZS-based ceramics with high density, the oxidation products produced during oxidation can heal the defects on the sample surface and increase the radius of curvature at the crack tip, thus increasing the strength. The ZS ceramics investigated here was in the low-density state, namely in the porous state. Because the pores of the partially sintered ZS were intergranular, existing among grains, the pore corner was necessarily narrow and sharp as well as acted as a stress concentrator and thus the pore also can be treated as a defect or a generalised crack. Therefore, the improvement in strength in the porous ZS ceramics was the same as that of the dense ceramic: because the flaws on the specimen surface were sealed by oxide products (ZrO₂, B₂O₃ or SiO₂).

The strength of porous material had been reported to decrease exponentially with the increase of porosity [6]. In this investigation, porous ZS ceramics with low porosity had high strength either before or after oxidation. These three batches of ceramics’ initial strength before and after thermal shock depended on the combination of material's porosity as well as the degree of oxide layer coverage and thickness of oxidation layer, which both were materially affected by the porosity for a fixed oxidation condition. The measured total porosity and open porosity of ZSA, ZSB and ZSC ceramics were 8.7 and 1.6%, 17.9 and 16.9% as well as 27.3 and 24.7%, respectively. The existence of the open pores might make a big influence on their oxidation behaviours.

Figure 4 displays the oxidation status of the centre (1.5 mm depth from sample surface) within ZSA_o, ZSB_o and ZSC_o. It could be observed that the oxidation occurred deeply into the centre of these materials. The areas of oxidation occurring were indicated by arrows and ellipses. As shown in Figure 4, more porosity led to more serious oxidation. However, even for the ZSC_o with the most oxidation, oxidation had just occurred on the outer surface of some grains only.

Figure 4

SEM micrographs of the central section on the fractured surfaces after oxidation: (a) ZSA_o; (b) ZSB_o and (c) ZSC_o. Red arrows and ellipses indicate the areas of oxidation occurred. Blue arrows indicate the defects produced in the process of oxidation.

Figure 5 depicts the thermal shock behaviours of ZSA_o, ZSB_o and ZSC_o. The ΔT_c, representing the temperature difference at which cracks begin to initiate, was used to evaluate their TSFR. Improvement in ΔT_c after oxidation was clearly observed in Figure 5(a) for each batch. The measured ΔT_c values of the ZSA_o, ZSB_o and ZSC_o were 456.4, 458.3 and 423.3°C, respectively, which were much higher than those for unoxidised samples (333.6, 358.0 and 361.2°C, respectively). Herein we employed Equation (1) [12] for a comprehensive analysis of the ΔT_c. (1) Here, R″ is a thermal shock parameter often used to characterise the resistance of a ceramic to thermal shock crack initiation. And σ_f is the flexural strength; ν is the Poisson's ratio; E is the Young's modulus; α is the thermal expansion coefficient; B is a shape factor dependent on specimen geometry; k is the thermal conductivity; A is the characteristic heat-transfer length; and h is the surface heat-transfer coefficient. For specimens of porous ZS prepared for thermal shock tests with fixed geometry and dimension (3 × 4 × 36 mm), A and B were constant, therefore, the change of ΔT_c was only affected by the material's properties. ΔT_c is proportional to the parameter σ_f and inversely proportional to the surface heat-transfer coefficient h. The oxidation helped to seal defects to increase the σ_f and played the role of thermal insulation to reduce the h in the case of either porous or dense ceramics. Taken together, the oxidised specimens required a greater thermal shock load compared with that required for the un-oxidised specimens to fracture, which is reflected in the higher value of ΔT_c.

Figure 5

(a) Thermal shock behaviour of porous ZS ceramics after oxidation; (b) thermal shock SRR of the original and the oxidised specimens.

In previous investigations [6], we have discussed the influence of porosity on ΔT_c using the same ZS ceramics studied here and concluded that for these un-oxidised ZS materials, the increased porosity led to a slight increase in ΔT_c. The ΔT_c followed the sequence ZSA < ZSB < ZSC. However, as seen in Figure 5(a), the opposite trend was observed after they were oxidised: the order in this case was ZSA_o ≈ ZSB_o > ZSC_o. In the case of the ZSC_o, as presented in Table 1, the strength increase was minimal because of its thickest oxidation layer. Moreover, the surface of the sample was least covered with, whereas the ZSA and ZSB were equally covered with the oxidation layer (Figure 2(b,d,f). The poor oxide layer coverage would lead to increased heat transfer. Accordingly, the ZSC_o with the highest porosity, which had lowest strength and the poorest heat-transfer impedance, had the lowest ΔT_c value. The oxidation layer thickness and the covering on the sample surfaces were the two main factors that affected the material's strength and surface heat transfer property [9], thereby the material's ΔT_c value.

Oxidation might also change the TSDR of the porous ZS ceramics, which could be evaluated by the SRR, the ratio of a material's strength after thermal shock relative to its original strength at a given temperature difference. Oxidation-produced internal defects are reported to be conducive to increase in crack resistance [13-16]. Some defects produced during the oxidation inside the samples were shown in Figure 4 by blue arrows. The oxidation inside the porous ZS ceramic would result in an increased SRR. However, oxides produced on the specimen surface, during the oxidation process, healed the surface defects and cracks thereby improving the material's initial strength, which would reduce the SRR.

Figure 5 also presents the TSDR of the unoxidised and oxidised porous ZS samples. The temperature difference for the SRR calculation was 475, 475 and 450°C, respectively, for ZSA_o, ZSB_o and ZSC_o, according to the experimental data collection points in the ellipse in Figure 5(a). The SRR of oxidised ceramics were lower than that of the corresponding unoxidised ceramics because of the surface and internal oxidation, as shown in Figure 5(b). The surface oxidation effect on the SRR overcame the competing influence of the internal oxidation on the porous ZS ceramics since the surface oxidation could reduce SRR of ZS ceramics but internal oxidation could increase the SRR [13]. For both the ZSA and ZSB ceramics, big differences in SRR before and after oxidation were observed. However, for the ZSC ceramic, the SRR maintained a relatively high level, even after oxidation, which might be due to its more internal oxidation than that of the ZSA or ZSB ceramic, thus leading to a certain enhancement on its crack resistance. Though the oxidation affected the TSDR of the ZSA, ZSB and ZSC ceramics, it did not change the trend with respect to porosity. ZS ceramics with higher porosity had higher TSDR after oxidation. The above experimental results were discussed from the point of view of material's microstructures, and mechanical and thermal properties. However, the micro-mechanism, such as the influence of the viscoelastic bridge [17] and crack growth [18] on the thermal shock behaviour of porous ceramics needs further investigation.

Conclusions

The effect of oxidation (at 1200°C for 30 min) on the thermal shock performance of porous ZS ceramics with 8.7, 17.9 and 27.3% porosity was investigated. The oxidation not only occurred on the surface but also slightly within the material. The combined effects of the changes of the material surface, internal microstructures and the oxidation layer thickness all determined the thermal shock performance of these porous ceramics after oxidation. The ΔT_c values of porous ZS ceramics with different porosities after oxidation were always higher than those unoxidised ZS ceramics. The oxidised porous ZS ceramics with lower porosity exhibited higher ΔT_c values. However, the oxidation reduced the TSDR of the porous ZS ceramics. The ZS ceramics with higher porosity had higher TSDR with or without oxidation.