Molecular dynamics simulation of vacancy cluster formation in β- and α-Si3N4
Graphical abstract
Introduction
Silicon nitride (Si3N4) possesses excellent high-temperature stability, high thermal shock resistance, good creep behavior, and good corrosion and oxidation resistance [1], [2], [3]. It is very well known that lattice defects, particularly Schottky defects, play an important role in determining the overall properties of Si3N4 ceramics [4], [5]. The Schottky defects are either caused by the introduction of oxygen as an impurity or thermal annealing that results in the formation of simultaneous Si and N vacancies [6]. In the case of oxygen impurity, vacancies are created to maintain the overall charge neutrality of the lattice [7]. Although α-Si3N41 initially was assumed to be established by oxygen impurity, but later research showed that nearly pure α-Si3N4 can be synthesized [7]. For α-Si3N4, previous researches showed the presence of vacancy clusters (as large as 30 nm) or dislocation loops [8]. Such clusters or loops are not yet detected in β-Si3N4, seemingly due to the unstrained lattice of β-Si3N4 (while α-Si3N4 is strained) impeding Schottky defects formation. Moreover, it is assumed that Si and N mobilities are higher in β-Si3N4 than in α-Si3N4, so vacancies are assumed to get vanished in the later during high-temperature sintering [9]. This reasoning is based on the diffusion data of Kijima and co-workers [10], which are controversial, since, in spite of all similarities of β and α modifications, the activation energies of N diffusion in the former and latter are measured as 778 and 235 kJ/mol, respectively. Wang et al. [11] calculated the amount of Schottky defects in α-Si3N4 sintered at 1973 K to be 0.08 at%. This calculation is only based on a large cluster of vacancies (~25 nm in diameter), ignoring the smaller ones. Accordingly, using , in which , , and are diffusion coefficient of N, diffusion coefficient of vacancy (estimated as 2 × 10−18 m2/s), and volumetric vacant fraction of Si3N4, respectively, Wang et al. [11] calculated the in α-Si3N4 as 6 × 10−21 m2/s. This is stated to be close to that obtained by extrapolation of Kajima’s data at 2000 K (10−19 m2/s) [10], although being two orders of magnitude different. Additionally, the extrapolation of Kajima’s data seems to be considerably (three orders of magnitude) miscalculated, since extrapolation of Kajima’s work yields diffusion coefficient of N as ~10−22 m2/s at 2000 K. The recent more accurate measurement by Schmidt [12] demonstrated that and in both modifications are in the same orders of magnitude and in α-Si3N4 is ~6 × 10−19 m2/s at ~2000 K. Putting this in the formulae that Wang used, and considering as 2 × 10−18 m2/s, the is estimated as ~8 at%. Considering these discrepancies, in the present study, we have used molecular dynamics simulation (MD) to investigate the tolerance of both modification of silicon nitride to contain varying content of Si and N vacancies (0–25.6 at%) as well as their tendency to form cluster when heat-treated for long time (114 ns).
Section snippets
Simulation procedure
MD simulation was carried out by QuantumATK (classical forcefield) package [13], with Virtual NanoLab (VNL) for visualization. The forcefield used in this study is Marian-Gastreich two-body potential [14], which is used quite frequently for simulation of silicon nitride ceramics [15], [16], [17], [18]. Furthermore, it was compared recently [19] with other existing forcefields for silicon nitride ceramics and was shown to be the best for reproducing properties of amorphous and crystalline
Results
Fig. 1 shows the lattice of different systems of β-Si3N4 with various contents of defects. Except in 0 and 25.6 systems, in the rest, for better illustration of the cluster vacancy formation, Si atoms with a coordination number of 4 and N atoms with a coordination number of 3 are deleted4. Therefore, the atoms left in these figures are either under-coordinated or over-coordinated. These atoms form
Conclusion
The current results clearly indicate that both α and β are capable of containing a considerable content of Schottky defects (up to 12.8 at%) without lattice destruction, confirming the recalculated data of Wang and co-workers (~8 at%). In both structures, these vacancies form considerably large-size vacancy clusters, although in β the cluster sizes are larger than in α. Introduction of more vacancies (25.6 at%) causes lattice destruction and complete amorphization in β, while in α such vacancy
CRediT authorship contribution statement
Esmaeil Adabifiroozjaei: Conceptualization, Funding acquisition, Data curation, Investigation, Software, Validation, Formal analysis, Writing - original draft, Writing - review & editing. S.S. Mofarah: Visualization, Writing - review & editing, Investigation. H. Ma: Software, Visualization, Writing - review & editing. Y. Jiang: Visualization, Writing - review & editing. M. Hussein N. Assadi: Visualization, Formal analysis, Writing - review & editing. T.S. Suzuki: Funding acquisition, Project
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgment
E.A. acknowledges the financial support (JSPS KAKENHI Grant Number: 18F18064) provided by the Japan Society for the Promotion of Science (JSPS).
Conflict of interest
The authors declare no competing interests.
Data availability statement
All relevant data are available upon request from corresponding authors ([email protected]).
References (31)
- et al.
Effect of temperature on the pre-creep mechanical properties of silicon nitride
J. Eur. Ceram. Soc.
(2009) - et al.
Creep of silicon nitride
Prog. Mater Sci.
(2004) - et al.
Effect of lattice defects on the thermal conductivity of β-Si3N4
J. Eur. Ceram. Soc.
(2003) - et al.
Structural nano-defects in α-silicon nitride
J. Eur. Ceram. Soc.
(1996) - et al.
Effect of the Si content on the structure, mechanical and tribological properties of CrN/AlSiN thin films
Surf. Coat. Technol.
(2013) - et al.
Modeling amorphous silicon nitride: a comparative study of empirical potentials
Comput. Mater. Sci.
(2018) - et al.
Molecular dynamics simulations on the effect of nanovoid on shock-induced phase transition in uranium nitride
Phys. Lett. A
(2019) Silicon nitride and related materials
J. Am. Ceram. Soc.
(2000)- et al.
Ionic interdiffusion as interaction mechanism between Al and Si3N4
J. Am. Ceram. Soc.
(2019) - et al.
Origin of dislocation loops in α-silicon nitride
J. Mater. Res.
(1996)
Fine structural features in α-silicon nitride powder particles and their implications
J. Am. Ceram. Soc.
Silicon nitride crystal structure and observations of lattice defects
J. Mater. Sci.
Dislocation loops in a-silicon nitride
J. Am. Ceram. Soc.
Nitrogen self-diffusion in silicon nitride
J. Chem. Phys.
Simultaneous diffusion of Si and N in silicon nitride
Phys. Rev. B
Cited by (2)
Microstructure, phase compositions and mechanical properties of slip cast sintered SiC/Si<inf>3</inf>N<inf>4</inf> composites
2021, Ceramics InternationalCitation Excerpt :Overall, these compounds are employed to improve thermal and mechanical properties [6–10]. The main ceramic compounds include C/C, Al/SiC, and Si3N4 or SiC reinforced with continuous carbon fiber composites [11–16]. A proper approach to produce silicon nitride/silicon carbide composites is the nitridation of silicon and silicon carbide in controlled atmosphere furnaces [13].
Molecular dynamics simulations of brittle to ductile transition in failure mechanism of silicon nitride nanoporous membranes
2020, Materials Today CommunicationsCitation Excerpt :For example, Shargh and Abdolrahim [29] captured that structural changes in N–Si–N bond angles of nanomembranes lead to variations of Young’s modulus in the elastic regime. Most recently, formation of vacancy clusters in silicon nitride upon thermal annealing was detected via MD simulations [33]. Although silicon nitride nanostructures are brittle materials inherently [29,30,34], it is expected that the distribution of nanopores controls the plastic deformation mechanisms and change the failure behavior from brittle to ductile.