On the origin of Si nanocrystal formation in a Si suboxide matrix

Abstract

We examined mechanisms underlying Si nanocrystal formation in Si-rich SiO 2 using a combination of quantum mechanical and Monte Carlo ͑MC͒ simulations. We find that this process is mainly driven by suboxide penalty arising from incomplete O coordination, with a minor contribution of strain, and it is primarily controlled by O diffusion rather than excess Si diffusion and agglomeration. The overall behavior of Si cluster growth from our MC simulations based on these fundamental findings agrees well with experiments.

Silicon monoxide – a convenient precursor for large scale synthesis of near infrared emitting monodisperse silicon nanocrystals

“…During the thermal treatment, the broad PXRD peak around 51 degrees gradually diminished and is most likely associated with amorphous Si clusters or a SiO x interface between domains of ncSi and SiO 2 . 18 Note that with increasing reaction time further coarsening 35 or "pseudo-Ostwald ripening" 23 of ncSi is not rare. We observe that longer heating times at the same temperature result in larger size ncSi, revealed by sharper peaks in PXRD and a red-shifted peak in PL spectra (see ESI †).…”

“…Evidently, precursors with higher Si content may be advantageous for obtaining higher yields of ncSi. Indeed, silicon nanoclusters formed in silicon monoxide (SiO) films [16][17][18][19] and other silicon sub-oxide (SiO x , x < 2) films, [20][21][22][23] usually prepared via thermal, plasma, and laser deposition, have been of interest in optical physics for applications as protective layers 24,25 and in silicon based optoelectronics. 26 Notably, SiO is also commercially available as powders in bulk quantities having a structure and properties that appears to be different from thin films.…”

Formation of GaAs nanostructures by droplet epitaxy—Monte Carlo simulation

“…In contrast to ab initio and molecular dynamics techniques, lattice Monte Carlo models have a great advantage through comparability of modeling times and sizes with experimental data [14][15][16][17]. However amorphous and liquid materials require non-lattice Monte Carlo models in combination with molecular dynamic and ab initio calculations [18,19], but non-lattice models need protracted calculating time and deprive the Monte Carlo model of its benefits.…”

“…However amorphous and liquid materials require non-lattice Monte Carlo models in combination with molecular dynamic and ab initio calculations [18,19], but non-lattice models need protracted calculating time and deprive the Monte Carlo model of its benefits. The most perspective is multiscale modeling when energy barriers determined by DFT calculations are used in lattice MC models [16,18]. Recently it has been demonstrated that lattice MC models can be adjusted to imitate the processes in liquid drops during nanowire growth and droplet epitaxy [17,20].…”

Ab initioparameterized valence force field for the structure and energetics of amorphous SiOx(0 ⩽x Lee1, Bondi2, Hwang3 2011Phys. Rev. BSelf Cite

“…[50][51][52] Strain energy (ΔE strain ) arises from lattice distortions involving bond stretching, bond angle distortion, torsion resistance, and non-bonding interactions. The structure, stability, and phonon properties of bulk disordered Si and SiO 2 materials have been successfully studied using a…”