3.1 Phase analysis and microstructure
As shown in Table 2, all specimens had high relative densities (≥ 99%) when pressureless solid-state sintered at 2100°C for 2 h in flowing Ar gas using 0.7 wt% B4C and 2.5 wt% C as sintering aids. The XRD patterns and SiC polytype contents in the sintered specimens are shown in Fig. 1 and Table 2, respectively. The XRD patterns revealed α-SiC (6H, 4H, 15R polytypes), and traces of B4C and graphite in the sintered specimens. β-SiC (3C) was not detected in the sintered specimens because of its complete conversion to α-phase caused by B-doping in the SiC lattice [38] and the high sintering temperature [44]. Residual C left after the carbothermal reduction of native SiO2 was transformed to graphite at the sintering temperature [5]. As shown in Table 2, 4H-SiC content decreased with an increase in initial α-phase content due to a slower 6 H to 4H-SiC phase transformation in α-SiC powders compared to β-SiC powders [7, 24–26]. The 3C to 6H to 4H phase transformation was accelerated in the β-SiC powder due to its high thermodynamic instability at the sintering temperature and high Al content (refer to Table 2) [45–47].
Typical microstructures of PSS SiC ceramics with varying initial α-phase content are shown in Fig. 2. The SEM micrographs exhibited a coarse-grained microstructure with elongated α-SiC platelet grains formed in situ via an accelerated 3C to 6H to 4H phase transformation in A0 and a gradual transformation to fine-grained microstructure with increasing initial α-phase content. The average aspect ratio of SiC grains decreased from 7.4 ± 1.3 in A0 to 2.9 ± 0.4 in A100. The fine-grained microstructure was responsible for the increased density (99.1→99.8%) with increasing initial α-phase content (refer to Table 2). It is well documented that equiaxed grains can easily accommodate each other and annihilate the residual pores, a phenomenon known as grain-shape-accommodation [29]. On the contrary, platelet grains grown preferentially along the c-axis hinder the elimination of inter-grain pores, due to interlocked elongated grains, resulting in a lower relative density.
3.2 Electrical and thermal properties
The electrical resistivity and carrier density of PSS SiC ceramics with varying initial α-phase content are shown in Fig. 3. The electrical resistivity monotonically decreased (2.6 × 107 → 1.5 × 106 Ω‧cm) with increasing initial α-phase content presumably due to (1) a decreased bandgap energy with increasing 6H-, 15R- and decreasing 4H-SiC contents (refer to Table 2) and (2) the low impurity content of the α-SiC powders (Table 1). The high content of Fe and V impurities in β-SiC powders resulted in the donor-acceptor compensation and trapping of charge carriers (refer to Table 1) [48], which is also indicated by an increase in carrier density (1.1 × 1010 → 2.8 × 1013 cm− 3) with increasing initial α-phase content (refer to Fig. 3). The bandgap energies of 6H, 15R, and 4H, are 3.10, 2.98, and 3.29 eV, respectively [49, 50]. Various metallic and non-metallic impurities are known to affect the electrical properties of SiC ceramics [48, 51]. Impurities like Al, B, Ga, and Sc are p-type dopants, whereas N, P, and Fe are n-type dopants. Both n- and p-type dopants increase the charge carrier density and thus decrease the electrical resistivity of SiC ceramics. However, in the case of codoping of n- and p-type dopants, an acceptor-donor compensation takes place causing a decrease in the charge carrier density [52, 53]. On the other hand, impurities like Be, O, V tend to create a deep level within the bandgap trapping the charge carriers causing an increased electrical resistivity [53].
As shown in Table 3, the electrical resistivity of the solid-state sintered SiC ceramics could be tuned in the range of 101–1011 Ω‧cm by changing the composition and sintering atmosphere [14, 17, 21, 54]. For example, PSS SiC ceramics with high electrical resistivities of 108–1011 Ω∙cm were fabricated by sintering in an N2 atmosphere at ≥ 2275°C and the high electrical resistivity was attributed to the formation of an electrically insulating C-B-N intergranular phase at grain boundaries [54]. On the other hand, the electrical resistivity of PSS SiC ceramics with 1 wt% B4C could be decreased (106→102 Ω‧cm) by four orders of magnitude by adding 2 wt% electrically conductive graphene [21]. The ceramics in this study exhibited intermediate electrical resistivities (106–107 Ω∙cm) comparable to those of PSS SiC ceramics sintered with 1 wt% B4C at 2050°C in Ar gas [21]. As shown in Table 3, the reports on the electrical resistivity of PSS SiC ceramic sintered with β-SiC powders are rarely found, presumably due to their higher cost and lower stability at high sintering temperatures compared to α-SiC powders. This study revealed that the high thermodynamic instability of β-SiC powders can be used to tune the electrical properties of PSS SiC ceramics to some extent via a controlled 3C to 6H to 4H phase transformation.
Table 3
Starting powder polytype, processing conditions, relative density, electrical resistivity, and thermal conductivity of solid-state sintered SiC ceramics.
Specimen | Batch composition (wt%) | Theoretical density (g/cm3) | Relative density (%) | SiC polytype content (%) |
6H | 4H | 15R |
A0 | 96.80% β-SiC + 2.5% C + 0.7% B4C | 3.174 | 99.1 | 75.3 | 24.7 | 0.0 |
A20 | 19.36% α-SiC + 77.44% β-SiC + 2.5% C + 0.7% B4C | 3.174 | 99.0 | 76.8 | 22.4 | 0.8 |
A40 | 38.72% α-SiC + 58.08% β-SiC + 2.5% C + 0.7% B4C | 3.174 | 99.6 | 78.6 | 19.9 | 1.5 |
A60 | 58.08% α-SiC + 38.72% β-SiC + 2.5% C + 0.7% B4C | 3.174 | 99.7 | 82.3 | 15.4 | 2.3 |
A80 | 77.4% α-SiC + 19.36% β-SiC + 2.5% C + 0.7% B4C | 3.174 | 99.7 | 84.8 | 12.3 | 2.9 |
A100 | 96.80% α-SiC + 2.5% C + 0.7% B4C | 3.174 | 99.8 | 85.5 | 9.8 | 4.7 |
Staring powder polytype, processing conditions, relative density, electrical resistivity, and thermal conductivity of solid-state sintered SiC ceramics. |
The thermal conductivity of PSS SiC ceramics with varying initial α-phase content is shown in Fig. 4. The thermal conductivity monotonically increased from 117.1 Wm− 1K− 1 in A0 to 154.2 Wm− 1K− 1 in A100 due to (1) an increased 6H-SiC content, which has a higher intrinsic thermal conductivity compared to that of 4H-SiC [38] and (2) the low soluble impurity content of the α-SiC starting powders (refer to Table 1). Soluble impurities, such as Fe and V [55, 56], act as phonon scattering sites in the SiC lattice and deteriorate the thermal conductivity. As shown in Table 3, the reported thermal conductivity of solid-state sintered SiC ceramics varies from 107 to 192 Wm− 1K− 1 depending on the composition and processing conditions. Li et al. [12] reported a high thermal conductivity of 192.2 Wm− 1K− 1 attributed to an optimum content of sintering aids (0.4 wt% B4C and 3 wt% C) derived from polymeric precursors. However, a low thermal conductivity of 107 Wm− 1K− 1 was reported for SiC-5 wt% graphene composite attributed to an increased interfacial thermal resistance at SiC-graphene heterogeneous interfaces [14]. As shown in Table 3, the literature is limited to the thermal conductivity of PSS SiC ceramics started with α-SiC powders. This study revealed that β-SiC powders are detrimental to thermal conductivity because they promote 6H to 4H phase transformation, leading to an increase of 4H phase, which has a low intrinsic thermal conductivity compared to 6H phase.
Based on the above discussion, the thermal conductivity of PSS SiC ceramics can be improved by (1) minimizing the content of sintering additives and secondary phases, (2) using high purity starting powders, and (3) avoiding or minimizing 6H to 4H phase transformation during sintering.
3.3 Mechanical properties
The mechanical properties of PSS SiC ceramics with varying initial α-phase content are shown in Table 4. The flexural strength (σf) increased from 289.5 MPa in A0 to 335.1 MPa in A100 attributed to a decreased defect size with decreasing grain size [11, 13, 24, 57]. However, the fracture toughness (KIC) and hardness were almost constant at 2.2–2.3 MPa‧m1/2 and approximately 28 GPa, respectively, with the variation in initial α-SiC content. As shown in Fig. 5, the crack path did not exhibit any active crack-microstructure interaction due to strong interfacial bonding between the SiC grains, leading to a predominant transgranular fracture in all specimens (refer to Fig. 6). In the absence of active crack-microstructure interaction, the fracture toughness was insensitive to changes in microstructure. These results are consistent with the findings of Magnani et al. [11] (refer to Table 5). It was demonstrated that fine-grained PSS SiC ceramics fabricated via two-step sintering exhibited 39% higher flexural strength compared to the coarse-grained ceramics fabricated via a conventional sintering process [11]. However, the fracture toughness and hardness did not exhibit any significant change [11]. As shown in Table 4, the calculated critical flaw sizes (c, 26–35 µm) were larger than the average grain sizes implying that the fracture had originated from processing defects. The critical flaw sizes were calculated using the equation KIC = 1.35 σf c1/2, assuming semielliptical surface flaws [24]. Figure 7 shows the typical fracture origins of PSS SiC ceramics which were mostly residual pores near the surface. The critical flaw sizes of A0 and A100 observed in Fig. 7 are comparable to the sizes calculated (Table 4). A decreased critical flaw size with increasing initial α-phase content indicates a superior sinterability of α-SiC compared to β-SiC powders, which is also indicated by a decreased scatter in flexural strength values with the addition of α-SiC powders (refer to Table 4).
Table 4
Flexural strength, fracture toughness, hardness, and calculated flaw size of pressureless solid-state sintered SiC ceramics.
Reference | Starting powder | Processing conditions | Relative density (%)/ sintering aid (wt%) | Electrical resistivity (Ω‧cm) | Thermal conductivity (Wm− 1K− 1) |
Munro et al. [4] | α-SiC | 2500°C | 98/B + C | - | 114 |
Li et al. [14] | α-SiC | 2130°C/1 h/Ar | 96.7–99.2/1% B4C + 3.2% C + 0–5% graphene | | 5.6×101 – 5.0×102 | 107–145 |
Li et al. [12] | α-SiC | 2150°C/1 h→ 1950°C/4 h/Ar | 99.4/0.4% B + 0.4% C | | - | 192.2 |
Taki et al. [17] | α-SiC | SPS: 2100°C/ 50 MPa/5 min/ vacuum or N2 | 97/1–5 mol% B | 103–104 | 132–185 |
Cai et al. [21] | - | 2050°C/2 h/Ar | 97.7/1% B4C | > 106 | |
| 96.7/1% B4C + 2% graphene | 1.9×102 | - |
Boecker et al. [54] | α-SiC | 2300°C/2 h/N2 | 99.3/1% B4C + 1% C | 108–1011 | - |
This study | α-SiC | 2100°C/2 h/Ar | > 99/0.7% B4C + 2.5% C | 1.5×106 | 154.2 |
β-SiC | 2.6×107 | 117.1 |
Flexural strength, fracture toughness, hardness, and calculated flaw size of pressureless solid-state sintered SiC ceramics. |
Table 5
Starting powder polytype, processing conditions, relative density, sintering aid composition, and mechanical properties of pressureless solid-state sintered SiC ceramics. ISB, IF, and SENB denote indentation strength in bending, indentation fracture, and single edge notched beam, respectively. * denotes the flexural strength measured using a three-point bending method.
Specimen | Flexural strength (MPa) | Fracture toughness (MPa‧m1/2) | Hardness (GPa) | Calculated critical flaw size (µm) |
A0 | 289.5 ± 45.6 | 2.3 ± 0.1 | 28.1 ± 1.2 | 35 |
A20 | 300.3 ± 42.7 | 2.2 ± 0.1 | 28.8 ± 2.8 | 29 |
A40 | 308.8 ± 35.9 | 2.2 ± 0.1 | 28.2 ± 2.8 | 28 |
A60 | 316.7 ± 42.2 | 2.2 ± 0.1 | 28.0 ± 3.0 | 26 |
A80 | 330.7 ± 50.4 | 2.3 ± 0.1 | 27.8 ± 1.4 | 27 |
A100 | 335.1 ± 23.9 | 2.3 ± 0.1 | 28.4 ± 2.6 | 26 |
Starting powder polytype, processing conditions, relative density, sintering aid composition, and mechanical properties of pressureless solid-state sintered SiC ceramics. ISB, IF, and SENB denote indentation strength in bending, indentation fracture, and single edge notched beam, respectively. * denotes the flexural strength measured using a three-point bending method. |
As shown in Table 5, the reported flexural strength, fracture toughness, and hardness of PSS SiC ceramics vary from 312–556 MPa, 2.5–5.7 MPa‧m1/2, and 21.2–30.2 GPa, respectively [6, 10, 11, 13, 15, 16]. A wide distribution of flexural strength is due to the difference in composition, processing conditions and measurement methods among the various works. According to Biswas’s work [9] on the flexural strength of hot-pressed SiC ceramics sintered with Al–B–C additives, the flexural strength measured using the three-point bending method was approximately 150 MPa higher than that obtained using the four-point bending method. Thus, the flexural strength values (289–335 MPa) of the present specimens measured by four-point bending are estimated to be superior to those (312–435 MPa, three-point bending) of PSS SiC ceramics with 0.5–1 wt% B4C and 3.1–4.7 wt% C [10, 15, 16]. Magnani et al. [11, 13] reported high flexural strengths of 530–556 MPa for two-step sintered SiC ceramics with fine-grained microstructure compared to those (320–467 MPa) of PSS SiC ceramics with coarse-grained microstructures reported in other works [3, 4, 6, 15, 16]. A fine-grained microstructure is beneficial in improving flexural strength because of the reduced critical flaw size [11, 13]. As shown in Table 5, the previous works were limited to the flexural strength of PSS SiC ceramics sintered with α-SiC powders. This study revealed that β-SiC powders promote abnormal growth of platelet grains via an in situ 3C to 6H to 4H phase transformation, leading to increased flaw size and reduced flexural strength. A direct comparison of fracture toughness values in the literature is difficult because the fracture toughness (2.5–4.6 MPa‧m1/2) was measured using various methods such as indentation strength in bending (ISB), single edge notched beam (SENB), and indentation fracture (IF). According to Rudnayova et al.’s work [58] on the fracture toughness of hot pressed Si3N4 ceramics measured using IF, ISB, and SENB methods, the fracture toughness varied by 3 MPa‧m1/2 depending on the method and indentation load. Nonetheless, the literature [11, 15] revealed that the low fracture toughness of PSS SiC ceramics is attributed to a strong interfacial bonding between the SiC grains, leading to a poor crack-microstructure interaction. Li et al.’s work [16] on PSS SiC ceramics with 0–5 wt% graphene suggests that the fracture toughness could be improved by adding graphene to promote crack deflection at SiC-graphene interfaces. The newly developed PSS SiC ceramics exhibited high hardness (28.4 GPa) comparable to the highest reported values (27.3–30.2 GPa).
Based on the above discussion, high flexural strength in PSS SiC ceramics can be obtained by (1) using α-SiC starting powders to ensure a fine-grained microstructure with reduced flaw size, (2) optimizing the sintering cycle to avoid undesirable grain growth, and (3) avoiding an in situ 6H to 4H phase transformation causing abnormal grain growth. The fracture toughness of PSS SiC ceramics can be improved by adding secondary phases such as graphene, which results in weak heterogeneous interfaces and promotes crack deflection along the grain boundaries [16, 19].