TiC0.5N0.5-Based Cermets with Varied Amounts of Si3N4 Nanopowders Prepared by Spark Plasma Sintering

TiCN-based cermets with varied fractions of Si 3 N 4 nanopowder (0–5wt.%) were prepared by spark plasma sintering. The microstructural and mechanical properties of these cermets were investigated. In general, with increasing addition amount of Si 3 N 4 nanopowder the relative density as well as mechanical properties of the as-prepared TiCN cermets increased first and then decreased. The samples containing 2wt.% Si 3 N 4 nanopowder presented the best performance with the relative density of about 98%, bending strength of 1000MPa, and Vickers microhardness of about 1810HV 10 .


Introduction
TiCN-based cermets have attracted extensive attention as cutting tools, moulds, and various wear-resistant components.They are known to possess higher hardness and better chemical stability in some respects than traditional WCbased hard metals [1][2][3][4][5][6][7].However, their lower toughness still limits their potentially wider application.Given that, most studies have focused on the strengthening and toughening of such kind of materials, in which some tried to produce nanosized or ultra-fine cermets based on the well-known Hall-Patch equation: the smaller TiCN grains the higher cermet hardness and strength [8][9][10], and others added some second phases like metal carbides [11][12][13] or nanoadditives into the cermet matrix [14][15][16].
Spark plasma sintering (SPS) is a kind of fast sintering process, which is known for its relatively low sintering temperature and high heating rate.During sintering, the grain growth could be suppressed, which is in favor of improving the mechanical properties of engineering materials.Consequently, it has been increasingly used in preparing various composites [17][18][19][20][21][22].However, there was no report about TiCN-based cermets with Si 3 N 4 nanopowders prepared by SPS.
Thus, in this study, SPS was used to fabricate TiCN-based cermets with varied fractions of Si 3 N 4 nanopowders.The addition effect of Si 3 N 4 nanopowder on the microstructure and mechanical properties of the as-prepared cermets were investigated, and the strengthening mechanism was also probed.

Sample Preparation.
Commercially available TiC 0.5 N 0.5 , WC, Mo, Ni, and VC powders were used as starting powders, and their specifications were listed in Table 1.In order to explore the effect of Si 3 N 4 addition on the microstructure and mechanical properties of TiCN-based cermets, the added Si 3 N 4 content was varied from 0 to 5 wt.% in increment of 1 wt.%.The cermet compositions designed in this work were 74 wt.% TiC 0.5 N 0.5 + 10 wt.% WC + 10 wt.% Ni + 5 wt.%Mo + 1 wt.%VC with extra addition of different amounts of nanosized Si 3 N 4 powder.
During sample preparation, all the raw powders except Si 3 N 4 nanopowder were firstly mixed together and ballmilled thoroughly in an attrition mill (Model: SY-1, China) for 1.278 × 10 5 s using YG-6 WC balls (ISO: k20) as grinding media and absolute alcohol as dispersion media.Following that, the designed amount of Si 3 N 4 nanopowder was added into the slurry for another 1.8 ks milling.Shortly, the mixed slurry was dried off in a vacuum oven at 338-368 K under a pressure of 0.1 Pa.After that, the powder chunks were pestled and sieved into fine powder, and the powders were sintered in graphite dies by a SPS apparatus (Model: SPS-1050T, Japan).For the sintering, the samples were heated up to 1623 K with a heating rate of 473 K/min under a pressure of about 30 MPa, and the soaking time was 8 min.Finally, the samples were cooled down by furnace cooling, acquiring disks with a final dimension of Φ20 mm × 5 mm.

Materials Characterization.
The cermet disks were first sectioned into small bars with dimension of 2 mm × 3 mm × 10 mm and polished to a mirror surface.Then, using the polished bars, the flexural strength was evaluated on a Zwick/Z005 tester by a three-point bending experiment.
During the testing, a loading rate of 0.5 mm/min was applied.
With the broken bars after the measurement of flexural strength, the bulk densities of the cermet specimens were evaluated by Archimedes method according to international standard (ISO18754).Then the final relative densities were estimated in accordance with the rule of mixtures.The measurement of specimen hardness was taken from the central part of the polished bars.The Vickers hardness test was performed at room temperature on a microhardness tester (Model: VH-5, China) with an indenting load of 10 kgf and a dwelling time of 20 s.Five microindentations were conducted at different locations on the polished cermets, and the reported value here is the average of the measured hardnesses.
The microstructure was examined on both the polished and fracture surfaces of the samples by scanning electron microscopy (SEM, Model: JSM-7410LV, Japan).The content of each element was recorded by an energy dispersive X-ray (EDX) spectroscopy attached to the SEM.The phase compositions of the samples were identified by X-ray diffraction (XRD) on a D8-advance X-ray diffractometer (Cu Ka,  = 1.5418Å) under continuous scanning mode with a speed of 5 ∘ /min.phases were identified in each sample.However, no diffraction peaks of Si 3 N 4 were identified, possibly due to the small addition amount of Si 3 N 4 nanopowder that cannot be detected by XRD in the detection limit, or some reactions happened during sintering that consumed out the added Si 3 N 4 nanopowder.In addition, it can be found that one of the diffraction peaks, that of Ni(111) grain plane, shifts to high diffraction angle with increasing addition of Si 3 N 4 nanopowder.As a result, a new phase, namely, Ni 3 Si (JCPDS: 65-3243) phase can be identified.Also, via further calculation of the lattice constant of Ni (listed in Table 2), it can be found that the value of the so-called Ni lattice constant decreases following the increased addition amount of Si 3 N 4 nanopowder, implying that some of Si atoms are merging into Ni, since it is known that the atomic size of Ni is much larger than that of Si.Taking all these information into account, it is rational to deduce that Si 3 N 4 had reacted with N, generating Ni 3 Si.

Results and Discussion
Typical SEM images in backscattering electron (BSE) mode on polished surface of the as-prepared samples with different amounts of Si 3 N 4 nanopowder are shown in Figure 2. From this figure, it could be observed that there were almost no pores in each sample, revealing the fully dense structures of the samples after sintering.As the addition   amount of Si 3 N 4 nanopowder increased, the size of TiCN grains increased first and then decreased, and the number of small white particles (which was labeled as A) increased.
Besides, the white particles tended to be located at the boundaries between black grains (which was labeled as B) or their triangle boundaries.As is known [1,23], there are typical core-rim microstructure in TiCN-based cermets, and in the SEM images taken under BSE mode, they are most possibly black cores and white rims and/or sometimes white cores and grey rims.And, commonly, the black cores are considered as TiCN, while the white cores are (Ti,W,M)(C,N) (M=Mo, Ni, Ta, and so forth).In this work, however, the typical core-rim structure of TiCN cermets could not be found, which might be attributed to the fast sintering speed of SPS, during which the core-rim structure could not be formed completely and timely.As a result, there are some white particles (labeled as As) and black ones (labeled as Bs) as shown in Figure 2.
As is well-known [1], the heavy metal elements such as W and Mo present a higher brightness.Typical atomic content of each element in Particles As and Bs is summarized in Tables 3 and 4, respectively, from which it could be seen that after increasing addition of Si 3 N 4 nanopowder C and N contents in white particles increased; Ni and Si contents in black ones increased as well; however, Ti and C contents in black ones decreased.Above all, a tentative inference on this result was conducted that more and more Ni and Si would resolve into the hard TiCN phase, following the increasing addition of Si 3 N 4 nanopowder, which is basically consistent with the XRD observation.Figure 3 illustrates the densities of the as-prepared TiCNbased cermets as a function of mass fraction of Si 3 N 4 nanopowder.It reveals that the apparent density decreased with increasing addition amount of Si 3 N 4 nanopowder, while the relative density of the samples increased first and then decreased following the increasing addition amount of Si 3 N 4 nanopowder, reaching its maximum of about 98% when the addition amount was 2.0 wt.%.Due to the lower density of Si 3 N 4 (which is much lower than those of other ingredients in the cermets), the apparent density was reduced with increasing addition amount of Si 3 N 4 nanopowder.But the Relative density (%) increase in relative density of the samples after the addition of small amount of Si 3 N 4 nanopowder was attributed to the fact that the added Si 3 N 4 nanopowder could be of benefit to the sintering, promoting the sample densification during sintering.However, when the addition amount of Si 3 N 4 nanopowder becomes greater (more than 2 wt.%), the sample relative density dropped down on account of two reasons: the first one was that the formation of ceramic Ni 3 Si suppressed the plastic deformation of metal binder phase, and with more and more Ni 3 Si, it became harder for the binder phase to perfectly wet the hard phase grains, resulting in reduced relative density; and another one was that the agglomeration owing to too much of the added nanosized powder was detrimental to the complete compaction of samples, and as a result, the relative density decreased.Figure 4 displays typical second electron (SE) SEM images on the fracture surface of the prepared cermets with different amounts of Si 3 N 4 nanopowder.It can be seen from this figure that the fracture mode of the prepared cermets did not change much even after Si 3 N 4 nanopowder was added, which was still a mixture of intergranular and transgranular fractures.However, when comparing Figure 4(a) with Figure 4(b), it is interesting to see that there were more concave-convex tearing edges in the latter.In other words, when the fracture in such samples occurred, it dissipated more energy, indicating that the bond between the grains in the samples was strengthened.Moreover, when comparing  4(c), there was much more transgranular fracture and something like spalling occurred in the latter, implying that the bond between the particles in such samples became weak.From these phenomena it is reasonable to conclude that the addition of Si 3 N 4 nanopowder with an appropriate amount is of benefit to bind the cermet grains; otherwise, that with too much of Si 3 N 4 nanopowder is of disadvantage.

Mechanical Properties.
Figure 5 presents the Vickers hardness of the as-prepared TiCN-based cermets as a function of mass fraction of Si 3 N 4 nanopowder.It can be seen that the sample hardness increased first and then decreased with increasing addition of Si 3 N 4 nanopowder, reaching its maximum of about 1810 HV 10 when its added amount was 2.0 wt.%.The whole change tendency in hardness was quite the same as that of the sample relative density, because it is well-known that the higher relative density, the higher hardness [24].In the present work, besides the increased relative density, the boosted hardness could be also attributed to the desirable role of moderate amount of the resultant Ni 3 Si in improving the sample hardness.When the addition amount of Si 3 N 4 nanopowder was small (approximately less than 2 wt.%), the white particles as shown in Figure 2 in the cermets became more after the addition of Si 3 N 4 and the content of Ti, C, and N in white particles increased as well, indicating that the content of TiCN in white particles turned to be more.As a result, the whole hardness of the cermets was enhanced.As the addition amount of Si 3 N 4 nanopowder increased (more than 2 wt.%), the hardness, however, turned to go down, not merely because of the decreased relative density as illustrated in Figure 4, but also owing to the increased contents of Ni and Si in the black particles as well as the increased amount of the black particles.Moreover, the hardness of Ni 3 Si is much lower than that of TiCN, which also contributed to the decrease in sample hardness.In addition, the weak binding strength (which can be seen from Figure 4(c)) between the particles also went against the improvement in hardness.For all these reasons, the sample hardness went down when too much of Si 3 N 4 nanopowder was added.
Figure 6 shows the bending strength of the as-prepared cermets as a function of mass fraction of Si 3 N 4 nanopowder.It can be seen that the bending strength of the samples went up first and then fell down, reaching its maximum of about 1000 MPa also when the added amount of Si 3 N 4 nanopowder was 2 wt.%.The possible reason for the increase in cermet bending strength may be attributed to two facts: the increased relative density and the resultant Ni 3 Si distributed in the binder phase during sintering.The as-produced ceramic Ni 3 Si could inhibit the dislocation motion of metal Ni, and as a result, the binder phase was strengthened and the strength of the TiCN-based cermets was enhanced.Besides, according to [25], the added nanosized particles could locate inside the grains and form a kind of transgranular microstructure, which may deflect the cracks during the fracture.Moreover, there would be some dislocation groups located in those microstructures, which could also hinder the crack extension and lead to their deflection, increasing the fracture energy and thus improving the cermet strength.In this study, the addition of Si 3 N 4 nanopowder may also produce transgranular microstructure and more subboundaries as shown in Figure 2(b).Since the coefficient of thermal expansion and modulus of elasticity of the added Si 3 N 4 nanoparticles were Advances in Materials Science and Engineering quite misfit with those of the cermets, the residual stress also came into being quite easily in the subboundaries.As the tip of cracks spreads to these areas, this kind of transgranular microstructure could hold up the crack propagation (see in Figure 2(b)); thus, the sample strength was boosted.Nevertheless, as the addition amount of Si 3 N 4 nanopowder increased, the sample strength decreased as well, which was partially in virtue of the weak bond between the binder phase and hard phase (the spalling as shown in Figure 4(c)).Moreover, the agglomeration of Si 3 N 4 nanopowder could also weaken the bond between the binder phase and hard phase.Consequently, the whole sample strength dropped down.

Conclusions
TiCN-based cermets with varied amounts of Si 3 N 4 nanopowder were fabricated by SPS.No Si 3 N 4 phase was identified by XRD in the detection limit, but with the addition of Si 3 N 4 nanopowder, a new phase of Ni 3 Si was detected.As the addition amount of Si 3 N 4 nanopowder increased, the relative density, hardness, and bending strength of the asprepared samples increased first and then decreased, reaching their maximum of about 98%, 1810 HV 10 , and 1000 MPa, respectively, when the added amount of Si 3 N 4 nanopowder was 2 wt.%.

Figure 1 :
Figure 1: Typical XRD patterns of the as-prepared TiCN-based cermets with different amounts of Si 3 N 4 nanopowder.

Figure 3 :
Figure 3: Apparent and relative densities of the as-prepared TiCNbased cermets as a function of mass fraction of Si 3 N 4 nanopowder.

Figure 4 (
Figure 4(b) with Figure4(c), there was much more transgranular fracture and something like spalling occurred in the latter, implying that the bond between the particles in such samples became weak.From these phenomena it is reasonable to conclude that the addition of Si 3 N 4 nanopowder with an appropriate amount is of benefit to bind the cermet grains; otherwise, that with too much of Si 3 N 4 nanopowder is of disadvantage.

Figure 6 :
Figure 6: Bending strength of the as-prepared TiCN-based cermets as a function of mass fraction of Si 3 N 4 nanopowder.

Table 1 :
Specifications of the initial powders used for the preparation of TiCN-based cermets.

Table 2 :
Lattice constant of Ni in the as-prepared TiCN-based cermets.

Table 3 :
Typical atomic content of each element in Particle A.

Table 4 :
Typical atomic content of each element in Particle B.