C_0.3N_0.7Ti-SiC Toughed Silicon Nitride Hybrids with Non-Oxide Additives Ti₃SiC₂

Abstract

In situ grown C_0.3N_0.7Ti and SiC, which derived from non-oxide additives Ti₃SiC₂, are proposed to densify silicon nitride (Si₃N₄) ceramics with enhanced mechanical performance via hot-press sintering. Remarkable increase of density from 79.20% to 95.48% could be achieved for Si₃N₄ ceramics with 5 vol.% Ti₃SiC₂ when sintered at 1600 °C. As expected, higher sintering temperature 1700 °C could further promote densification of Si₃N₄ ceramics filled with Ti₃SiC₂. The capillarity of decomposed Si from Ti₃SiC₂, and in situ reaction between nonstoichiometric TiC_x and Si₃N₄ were believed to be responsible for densification of Si₃N₄ ceramics. An obvious enhancement of flexural strength and fracture toughness for Si₃N₄ with x vol.% Ti₃SiC₂ (x = 1~20) ceramics was observed. The maximum flexural strength of 795 MPa for Si₃N₄ composites with 5 vol.% Ti₃SiC₂ and maximum fracture toughness of 6.97 MPa·m^1/2 for Si₃N₄ composites with 20 vol.% Ti₃SiC₂ are achieved via hot-press sintering at 1700 °C. Pull out of elongated Si₃N₄ grains, crack bridging, crack branching and crack deflection were demonstrated to dominate enhance fracture toughness of Si₃N₄ composites.

1. Introduction

Although several alternatives of structural ceramics have been proposed, silicon nitride (Si₃N₄)-based ceramics remain competitive due to their superior properties, involving high strength and hardness at elevated temperatures, high resistance to oxidation and chemical attack, low coefficient of tribological friction and thermal expansion, and low dielectric permittivity, etc. [1,2,3,4,5,6,7,8,9,10]. As important multifunctional materials, Si₃N₄ ceramics have found wide range of successful application towards gas turbine engine components [10,11,12,13,14], cutting tools [11,15], radomes [2], and even integrated circuit [16,17], optical devices [18,19], etc.

However, due to the high degree of covalent bonding, Si₃N₄-based ceramics are very difficult to densify through the solid-state sintering process. Therefore, effective approaches to ensure rapid consolidation and high mechanical performance of Si₃N₄-based ceramics are actively being explored, including gas pressure sintering (GPS) [11], hot-pressing sintering (HPS) [20,21,22,23,24,25,26], hot isostatic pressing sintering (HIP) [27], spark plasma sintering (SPS) [26,28,29], and microwave sintering [1,30], etc. However, considering the requirement of high gas pressures for gas pressure sintering and extra current devices for SPS with a significantly higher furnace costs, HPS allows the dense and complex-shaped parts with medium cost [2]. Previous considerable efforts have demonstrated that fully dense Si₃N₄ ceramics with superior strength could be achieved through liquid phase sintering by addition of rare-earth oxides to promote mass transport and accelerate the rate of α − β transformation, most notably the rare-earth oxides involving Y₂O₃ [4,31,32]. A combination of various rare-earth oxides and other metallic oxides, such as Y₂O₃, La₂O₃, Nd₂O₃, Sm₂O₃, Yb₂O₃, Lu₂O₃, Al₂O₃, and MgO, also are effective sintering aids to densify Si₃N₄ [25,30,32,33,34,35].

Nevertheless, these oxides additives crystallized to intergranular glassy phase in the cooling stage [31], which deteriorate the high-temperature performance of the ceramics such as creep and high-temperature strength due to the relative low eutectic temperature [32,36]. As a result of the early interest in hot-pressed Si₃N₄ ceramics as a high-temperature gas turbine material, attention was directed to high-temperature strength and creep resistance. Therefore, it is quite essential to explore novel heat-resistant sintering aids for high-performance Si₃N₄ ceramics from new view point.

The last two decades have been witness to the dramatic development on MAX phase cermets with the hexagonal symmetry due to their unique combination of characteristics of both ceramics and metals (M is an early transition metal, A is a group A element, X is either carbon and/or nitrogen), especially the layered ternary carbide titanium aluminum carbide (Ti₃AlC₂) and titanium silicon carbide (Ti₃SiC₂) [37,38,39]. The crystal structure of these MAX cermets can be described by alternately stacking of TiC₆ and Al/Si atomic planes. The unique combination of excellent properties of Ti₃AlC₂ or Ti₃SiC₂, including high melting point, high hardness, high elastic modulus, good thermal and electrical conductivity, and considerable chemical stability, make them to be fascinating candidates for various application. Moreover, elemental metal powder-derived MAX materials have demonstrated to be effective reinforcement in TiB₂ [40,41,42], Al₂O₃ [43,44,45] composites with enhanced mechanical properties by in situ reaction. More recently, as explicated in our previous work [46,47,48,49,50], titanium aluminum carbide (Ti₃AlC₂) was chosen as an effective sintering aid to effectively densify B₄C ceramics with enhanced sintering ability and mechanical performance simultaneously. High hardness and toughness values of 28.5 GPa and 7.02 MPa·m^1/2 respectively were achieved for B₄C composites sintered with 20 vol.% Ti₃AlC₂ at 1900 °C. The mechanisms of the enhanced sinterability of high-performance ceramics in previous works [1,10,46,47,48,49,50] could be classified into two aspects: Firstly, the decomposed metals from Ti₃SiC₂ or Ti₃AlC₂ at high-temperature can form liquid phase which promote sintering effectively. Secondly, in situ reaction sintering between matrix and titanium carbon compound would also promote densification and mechanical performance. The main competitive advantage of MAX aids is considered to be formation of reaction bonding between Si₃N₄ matrix and aids, rather than intergranular glassy phase. Motivated by such an idea [46,47,48,49,50], these non-oxides cermets are highly expected to play a multifunctional role in densification and enhancement of mechanical properties of Si₃N₄ ceramics.

It is also noteworthy that Al decomposed from Ti₃AlC₂ and residual O originated from raw Si₃N₄ powders would be dissolved into Si₃N₄ grains during high-temperature sintering procedure, which is harmful to the purity and thermal performance of Si₃N₄-based ceramics. As Y. Zhou et al. illustrated [51], a tendency of decreasing fracture toughness with increasing Al dopant could be observed. Moreover, even the 0.4 wt.% concentration of Al would lead to a drastically reduce of thermal conductivity by 36.9% (from 91.9 to 58.0 W·m⁻¹·K⁻¹) for Si₃N₄ ceramics. Therefore in this work, Ti₃SiC₂ were introduced to densify Si₃N₄ ceramics in order to demonstrate that it provided any advantages over the rare-earth oxide system. Besides, the effect of Ti₃SiC₂ volume fraction on the microstructure, hardness, flexural strength and fracture toughness was also studied. It is believed that Ti₃SiC₂ or other members of MAX family would lead to new scientific and technological data providing new insight into functionalization of Si₃N₄ ceramics.

2. Experimental Procedure

2.1. Preparation of Samples

Commercially available α-Si₃N₄ powder (purity > 93%, d₅₀ = 0.7 μm, Jinshenghao New Materials Co. Ltd., Anyang, China) was used as a starting material. As a novel sintering aid, Ti₃SiC₂ powders (d₅₀ = 5 μm, purity > 98%) were kindly provided by Forsman Scientific Co., Ltd., Beijing, China. To investigate the effect of Ti₃SiC₂ content on the mechanical properties, experiments were conducted with various amounts of Ti₃SiC₂ powders (1 to 20 vol.%) embedded in α-Si₃N₄ powders. To ensure the homogeneity of the mixed powders, α-Si₃N₄ with x vol.% Ti₃SiC₂ powders (x = 1~20) were wet ball-milled for 10 h by using ethanol as ball-milling media. The substance was dried at 80 °C, and sieved with a filter with a mesh size of 63 μm, then placed in a graphite die coated with BN powder to avoid reaction between the powder and graphite die. Hot-press (HP) sintering was performed on vacuum hot press sintering furnace (ZT-63-21Y, Shanghai Chenhua Technology Co., Ltd., Shanghai, China) with ramp of 10 °C /min 1600 °C and 1700 °C for 90 min in flowing nitrogen under 30 MPa uniaxial pressure during the whole cycle. After natural cooling to room temperature inside furnace, samples were polished and ultrasonic cleaned before characterization. For comparison, α-Si₃N₄ powders with 2 wt.% Alumina (Al₂O₃, AR, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) and 5 wt.% yttria (Y₂O₃, AR, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) were hot-pressed at the same sintering condition.

2.2. Characterizations

Before microstructure and mechanical performance characterizations, all hot-pressed samples with diameter of 50 mm were cut into bars and cuboids with help of inside diameter slicer. To reduce surface roughness of samples and guarantee sufficient experimental precision, the abrasive SiC papers with grit size of P320, P600, P1200 and P2000 were used in chronological order during the polishing process. Final manual polishing was carried out with polishing cloth containing alumina suspension with particle size 0.3 μm. The bulk density of each sample was determined according to the Archimedes principle in distilled water. X-ray Diffraction (XRD) patterns were recorded on X’pert PRO (PANalytical B.V., Almelo, Netherlands). Phase identification and quantitative analysis were performed on MDI Jade software (version 6.0, MDI, Livermore, CA, USA) according to Rietveld method. The microstructures of polished surfaces and fracture surfaces were observed using scanning electron microscopy (SEM, Nova NanoSEM 230, FEI Company, Hillsboro, OR, USA) with an energy dispersive X-ray (EDX) analyzer. The Vickers hardness was performed on micro hardness tester (VTD 512, Beijing Weiwei Technology Co., LTD, Beijing, China) under load of 9.8 N with a dwell time of 10 s, and determined by the Vickers diamond indentation method using the following equation:

$$H_V=0.102\frac{F}{S}=0.102\frac{2F\sin \raisebox{1ex}{$136$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}{d^2}=0.1891\frac{F}{d^2}$$

(1)

where P is the indentation load on the polished surface and d is the average diagonal length of the Vickers indentation. For accuracy, 11 Vickers indentations on each specimen were applied. After indentation, the microstructures were immediately observed by optical microscopy (ECLISPE LV150N, Tokyo, Japan). As a simple way of estimating toughness, indentation techniques were applied from observed corner cracks and calculated Vickers hardness using the Anstis equation:

$$K_{IC}=0.016{\left(\raisebox{1ex}{$E$}\!\left/ \!\raisebox{-1ex}{$H_V$}\right.\right)}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}\left(\raisebox{1ex}{$P$}\!\left/ \!\raisebox{-1ex}{$c^{\raisebox{1ex}{$3$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}$}\right.\right)$$

(2)

where E is the Young’s modulus and c is the half-length of cracks formed by the indentation. According to the international standard of test method for flexural strength of fine monolithic ceramics at room temperature which described in ISO 14704-2000, three-point flexural strength of specimens with size of 3 mm × 4 mm × 36 mm was performed on the mechanical testing machine (Instron 3369, INSTRON Corporation, Norwood, MA, USA) at a cross head speed of 0.5 mm/min.

3. Results and Discussion

Figure 1 illustrates the density of Ti₃SiC₂ filled Si₃N₄ ceramics as a function of Ti₃SiC₂ volume fraction. For Si₃N₄ ceramics which HP sintered at 1600 °C without aids, the density is only 2.58 g·cm⁻³. Partial densification may be attributed to the residual SiO₂ liquid phase during firing at high-temperature which always present on Si₃N₄ powder particles. A remarkable increase to 3.11 g·cm⁻³ could be observed for Si₃N₄ ceramics filled with only 5 vol.% Ti₃SiC₂ when sintered at the same temperature. The enhanced density is even more noticeable than the Si₃N₄ ceramics with 7 wt.% Y₂O₃-Al₂O₃ aids. These observed results demonstrate Ti₃SiC₂ to be a effective sintering aid to densify Si₃N₄ ceramics. However, further increase in Ti₃SiC₂ content dose not bring any appreciable consolidation.

As expected, higher sintering temperature 1700 °C could further promote densification of Si₃N₄ ceramics filled with Ti₃SiC₂. Besides, experimental points are inclined to distribute on a straight line with coefficient of determination (R²) above 0.99. To investigate the composition evolution after sintering, XRD patterns of Si₃N₄ ceramics filled with different volume fraction of Ti₃SiC₂ sintered at 1600 and 1700 °C, as well as 7 wt.% (Al₂O₃-Y₂O₃) densified Si₃N₄ ceramics are shown in Figure 2. As seen in Figure 2a, both α and β phase of Si₃N₄ could be detected when sintering temperature is 1600 °C, which suggests only a partial transformation of α phase to the more stable β phase. In contrast, when further improving sintering temperature to 1700 °C, all diffraction peaks of α-Si₃N₄ phase disappear (see in Figure 2b). This completely transformation of α to β-Si₃N₄ phase is believed to be essential to the enhancement of densification and mechanical performance [1,52].

Another important feature should be noted here is that the characteristic diffraction peaks of the raw Ti₃SiC₂ powder nearly disappear completely after sintering. This could be ascribed to the fact that Ti₃SiC₂ powder is thermal stable up to ~800 °C, and the following reaction can be responsible for the decomposition of Ti₃SiC₂ [53]:

$$3T{i}_{3}Si{C}_2\stackrel{~800°C to 1400°C}{\to }4Ti{C}_x+T{i}_{5}S{i}_3{C}_y+(6-4x-y)C$$

(3)

where the value of x ranges from 0.6 to 0.8 and y ≤ 1. Besides, the TiC_x phase appears to result in more rapid deterioration of the Ti₃SiC₂ phase. Also noted that the decomposition usually accomplished with decomposition of Ti₃SiC₂ to form nonstoichiometric TiC_x and gaseous Si, as demonstrated previously [54]:

$$T{i}_{3}Si{C}_2\stackrel{~1300°C}{\to }3Ti{C}_{\raisebox{1ex}{$2$}\!\left/ \!\raisebox{-1ex}{$3$}\right.}+Si$$

(4)

The Si is believed to be act as lubricating phase between Si₃N₄ grains to promote densification of Si₃N₄ ceramics through capillarity. Meanwhile, nitriding of TiC_x which originated from interatomic diffusion of C and N [10] during high-temperature sintering process would lead to the formation of new phase C_0.3N_0.7Ti. In addition, the residual Si would further react with nitrogen to form Si₃N₄ [55,56,57]. On the other hand, further heating during insulation stage will result in the likely loss of gaseous silicon. Therefore, it is reasonable to assume that the decomposition products of Ti₃SiC₂ would further react with Si₃N₄ through diffusion of C and N according to the following reactions:

$$T{i}_{5}S{i}_3{C}_y\to T{i}_5{C}_y+3Si$$

(5)

$$Ti{C}_x+S{i}_3{N}_4\to (3.333x-10)C_{0.3}{N}_{0.7}Ti+3SiC+(5.5-1.167x)N_2$$

(6)

$$T{i}_5{C}_y+(4.5-y)C+S{i}_3{N}_4\to 5C_{0.3}{N}_{0.7}Ti+3SiC+0.25N_2$$

(7)

$$3C+S{i}_3{N}_4\to 3SiC+2N_2$$

(8)

$$C+Si\to SiC$$

(9)

or

$$\begin{array}{c}Ti{C}_x+T{i}_{5}S{i}_3{C}_y+(9.45-x-y)C+1.55S{i}_3{N}_4\\ \to 6C_{0.3}{N}_{0.7}Ti+7.65SiC+N_2\end{array}$$

(10)

It is reasonable to claim that the in situ reaction is responsible for the additional characteristic diffraction peaks of C_0.3N_0.7Ti and SiC in XRD patterns.

Furthermore, detailed refinement parameters by means of Rietveld method are summarized in

Table 1. Refinement parameters of Ti₃SiC₂ doped Si₃N₄ ceramics.

Si3N4−x vol.% Ti3SiC2	Weight Fraction			R Factor of Rietveld	Theoretical Density (g·cm−3)	Densification
	β-Si3N4	C0.3N0.7Ti	SiC
x = 1	97.3%	0.9%	1.8%	4.76%	3.21	99.26%
x = 3	97.1%	0.9%	2.1%	4.55%	3.21	99.89%
x = 5	92.8%	4.1%	3.1%	3.46%	3.25	99.92%
x = 10	87.3%	7.6%	5.1%	3.42%	3.30	99.73%
x = 15	82.3%	11.7%	6.0%	3.47%	3.35	99.98%
x = 20	78.4%	14.6%	7.0%	4.07%	3.39	99.79%

. Detailed refinement parameters, including weight fraction and R factor, of Ti₃SiC₂ doped Si₃N₄ ceramics are illustrated in Figures S1–S6 (see in Supplementary Materials). Obviously, the weight fraction of both C_0.3N_0.7Ti (Fm-3m, PDF# 42-1448) and moissanite-3C SiC (F-43m, PDF# 29-1129) are inclined to increase with Ti₃SiC₂ content, which in turn confirms the in situ densification sintering mechanism discussed above. In addition, theoretical density could be derived with help of Rietveld method. Results have shown that nearly full densification for Ti₃SiC₂ doped Si₃N₄ ceramics sintered at 1700 °C could be achieved.

Figure 3 shows the micromorphology of polished surface of Si₃N₄ with different volume fractions of Ti₃SiC₂ sintered at 1700 °C. Due to the lack of sufficient sintering aids, lots of pores could be observed, and grain growth of β-Si₃N₄ is not complete for monolithic Si₃N₄ ceramic (see Figure 3a). However, the microstructures of Ti₃SiC₂-Si₃N₄ ceramics (see Figure 3b–g) exhibit much more close-grain structure and consist of randomly oriented elongated Si₃N₄ grains which is accordant with XRD results in Figure 2. The average diameters of grains present slight increasing trend from 0.68 to 0.98 μm by quantitative image analysis as the amount of Ti₃SiC₂ increased. Besides, the bright contrasted phase which uniformly embedded in Si₃N₄ matrix could be observed and are inclined to aggregate especially when Ti₃SiC₂ content exceeds 15 vol.%. Furthermore, as shown in

Table 2. EDS chemical analysis (at.%) at different positions in Figure 3g.

Location	Si	N	O	Ti	C	Possible Phases
Spot A	52.21	35.43	2.04	1.29	9.03	Si3N4, SiC
Spot B	5.10	27.18	-	50.13	17.59	C0.3N0.7Ti, SiC

, energy dispersive spectrometer (EDS) at spot A in Figure 3e suggests dominant phase of Si₃N₄ and SiC, which is associated with reaction described by Equation (10). Additional O element may be originated from surface of raw α-Si₃N₄ powders. Meanwhile, the bright region at spot B is proved to be enriched by Ti according to the EDS results in Table 2. Combined with the results of XRD analysis, it is reasonable to claim that the dispersive bright regions consist of C_0.3N_0.7Ti and SiC, which are believed to affect the mechanical performance of reaction bonded Si₃N₄ ceramics.

The mechanical properties, including Vickers hardness, flexural strength, and fracture toughness, of dense Si₃N₄ ceramics with different Ti₃SiC₂ content sintered at 1700 °C are illustrated in Figure 4. Clearly, the Vickers hardness of Si₃N₄ ceramics has been upgraded after modification of Ti₃SiC₂, and presents slight increase compared with that of Si₃N₄ ceramics containing conventional oxides aids. Besides, an obvious enhancement of flexural strength and fracture toughness could be observed. A maximum flexural strength of 795 MPa could be achieved for 5 vol.% Ti₃SiC₂ doped Si₃N₄ composites, which is almost twice that of 7 wt.% (Y₂O₃-Al₂O₃)-Si₃N₄ ceramics prepared at the same condition. This enhancement of flexural strength could be attributed to the C_0.3N_0.7Ti and SiC which originated from reaction bonding between Ti₃SiC₂ and Si₃N₄ [10]. However, further increment of Ti₃SiC₂ content reduces the flexural strength of Si₃N₄ ceramics which may be ascribed to the enhanced residual stresses around grain boundary [58,59]. Please note that this residual stress is believed to result in microcracks and intergranular fracture mode which will be discussed later. Moreover, the fracture toughness of Si₃N₄ composites is also effectively boosted after Ti₃SiC₂ decoration, and reaches maximum value of 6.97 MPa·m^1/2 for 20 vol.% Ti₃SiC₂-Si₃N₄ ceramics which is 37% higher than that of 7 wt.% (Y₂O₃-Al₂O₃)-Si₃N₄ ceramics.

Figure 5 illustrates the typical optical micrographs of the Vickers hardness indents and the induced cracks of Si₃N₄ ceramics with different Ti₃SiC₂ contents, as well as 7 wt.% (Y₂O₃-Al₂O₃). Clearly, the polished surfaces of Ti₃SiC₂ doped Si₃N₄ ceramics become much smoother than the monolithic Si₃N₄ ceramic which HP sintered at 1700 °C, corresponding to the enhancement of densification. Besides, it can be seen that the area of indentation presents no obvious change for Ti₃SiC₂ doped Si₃N₄ ceramics, which is consistent with the stable Vickers hardness. However, the cracks obviously become shorter especially when the Ti₃SiC₂ contents exceed 10 vol.%, which is responsible for the enhancement of fracture toughness.

To illustrate the fracture behaviors and activated toughening mechanisms, micromorphology and crack paths are investigated on cross-sectional fracture surfaces and polished surfaces, respectively. Comparison of typical fracture surfaces between Si₃N₄ doped with Al₂O₃-Y₂O₃ and Ti₃SiC₂ is illustrated in Figure 6. As can be seen from Figure 6a, a small number of pores occur in the Si₃N₄-7 wt.% (Al₂O₃-Y₂O₃) composites, which is harmful for the mechanical performance. In contrast, the Si₃N₄-Ti₃SiC₂ specimen presents a much more close-grain fracture surface owning to the higher density. As marked by red arrows in Figure 6b, large amounts of dimples corresponding to the transgranular fracture could be observed. In addition, this fracture mode is considered to make a dominant contribution to the superior flexural strength of Si₃N₄-Ti₃SiC₂ composites [10]. Besides, as marked by yellow arrows, lots of interface debonding between the Si₃N₄ grains and grain boundary phase could be observed. This intergranular fracture mode may result from the pullout of elongated β-Si₃N₄ grains, which is believed to make a contribution to the enhancement of overall fracture toughness [60,61,62].

Another mechanism of the enhanced fracture toughness of Si₃N₄-Ti₃SiC₂ composites could be ascribed to the crack branching, deflection, and grain bridging by in situ derived C_0.3N_0.7Ti and SiC grains embedded in Si₃N₄ matrix, which illustrated in Figure 7. Due to the superior hardness of C_0.3N_0.7Ti and the thermal mismatch between Si₃N₄ and C_0.3N_0.7Ti, there exists residual stress around C_0.3N_0.7Ti grains during cooling, giving rise to the microcracks inside composites. When subjected to the external mechanical stress, these microcracks tend to be activated and the propagation path of cracks tends to be split by C_0.3N_0.7Ti hard-phase and deflected along the interface. Such mechanisms consumed more fracture energy during the crack propagation which leads to crack arrest [63,64,65,66,67].

A comparison of mechanical properties of Si₃N₄-based ceramics obtained in the present work and selected previous works with conventional oxide aids is shown in

Table 3. Selected results on mechanical properties of Si₃N₄ ceramics by pressure-assisted sintering.

Composition	Sintering Conditions	Vickers Hardness (GPa)	Flexural Strength (MPa)	Fracture Toughness (MPa·m1/2)	Ref.
α-Si3N4 + 4 wt.% Al2O3 + 6 wt.% Y2O3	Hot isostatic pressing at 1700 °C, 20 MPa, 3 h	16.4	730	6.5	[27]
α-Si3N4 + 5 wt.% Al2O3 + 5 wt.% Y2O3	Hot press at 1800 °C, 30 MPa, 1.5 h	16.1	-	5.2	[20]
α-Si3N4 + 4 wt.% Al2O3 + 6 wt.% Y2O3	Hot press at 1700 °C, 50 MPa, 1.5 h	17.01	-	-	[21]
α-Si3N4 + 30 vol.% β-Si3N4 whiskers + 5 wt.% Al2O3 + 5 wt.% Y2O3 + 5 wt.% CeO2	Hot press at 1700 °C, 30 MPa, 30 min	19.0	794	8.6	[22]
α-Si3N4+4 wt.% Al2O3+4 wt.% Y2O3+15 vol.% SiC whiskers	Hot press at 1800 °C, 30 MPa, 30 min	16	680	6.1	[68]
α-Si3N4 + 5 wt.% Al2O3 + 4 wt.% Y2O3 + 3 wt.% TiC	Gas pressure sintering at 1750 °C, 2 MPa	16.4	475	7.6	[11]
α-Si3N4 + 5 vol.% Ti3SiC2	Hot-pressed at 1700 °C for 90 min	19.78	795	4.88	This work
α-Si3N4 + 20 vol.% Ti3SiC2	Hot-pressed at 1700 °C for 90 min	20.11	549	6.58	This work

. Clearly, the Vickers, flexural strength, and toughness of Ti₃SiC₂ doped Si₃N₄ ceramics present the same level or even better compared with Si₃N₄ ceramics sintered with oxide aids. Moreover, due to the superior mechanical and thermal properties of in situ formed C_0.3N_0.7Ti and SiC, Si₃N₄ ceramics obtained in this work are believed to have a significant competitive advantage and to promote the development of Si₃N₄-based ceramics at high temperatures.

4. Conclusions

In summary, we proposed non-oxide Ti₃SiC₂ (one of typical MAX cermets) as a novel sintering aid to densify Si₃N₄ ceramics with enhanced mechanical properties. A remarkable relative density increment of 20.5% (from 2.58 to 3.11 g·cm⁻³) could be observed for 1600 °C hot-press sintered Si₃N₄ ceramics doped with only 5 vol.% Ti₃SiC₂ compared with Si₃N₄ ceramics without aids. Further increase in sintering temperature to 1700 °C brought appreciable consolidation of nearly full dense Ti₃SiC₂-Si₃N₄ ceramics. XRD and EDS investigations demonstrated the formation of C_0.3N_0.7Ti and SiC which resulted from in situ reaction between Ti₃SiC₂ and Si₃N₄ through diffusion of C and N. The Vickers hardness of Ti₃SiC₂ doped Si₃N₄ ceramics increased slight compared with that of Si₃N₄ ceramics containing conventional oxides aids. Nevertheless, an obvious enhancement of flexural strength and fracture toughness could be observed. A maximum flexural strength of 795 MPa could be obtained for 5 vol.% Ti₃SiC₂ doped Si₃N₄ composites. Moreover, the fracture toughness of Ti₃SiC₂ densified Si₃N₄ composites exhibited a remarkable increase with increasing in volume fraction, and reached maximum value of 6.97 MPa·m^1/2 for 20 vol.% Ti₃SiC₂-Si₃N₄ ceramics. Pull out of elongated Si₃N₄ grains, crack bridging and deflection were demonstrated to promote fracture toughness of Ti₃SiC₂ densified Si₃N₄ composites. With these successes, MAX phase densified Si₃N₄ ceramics with enhanced strength and toughness will be necessary to meet demands of potential future markets for advanced ceramics. Further efforts are encouraged to be devoted to thermal properties investigations of MAX enabled Si₃N₄ composites.