Silicon nitride (Si3N4) is an extensively used material in the automotive, aerospace, and semiconductor industries. However, its widespread use is in contrast to the scarce availability of reliable interatomic potentials that can be employed to study various aspects of this material on an atomistic scale, particularly its amorphous phase. In this work, we developed a machine learning interatomic potential, using an efficient active learning technique, combined with the Gaussian approximation potential (GAP) method. Our strategy is based on using an inexpensive empirical potential to generate an initial dataset of atomic configurations, for which energies and forces were recalculated with density functional theory (DFT); thereafter, a GAP was trained on these data and an iterative re-training algorithm was used to improve it by learning on-the-fly. When compared to DFT, our potential yielded a mean absolute error of 8 meV/atom in energy calculations for a variety of liquid and amorphous structures and a speed-up of molecular dynamics simulations by 3–4 orders of magnitude, while achieving a first-rate agreement with experimental results. Our potential is publicly available in an open-access repository.
Silicon nitride films are widely used as the charge storage layer of charge trap flash (CTF) devices due to their high charge trap densities. The nature of the charge trapping sites in these materials responsible for the memory effect in CTF devices is still unclear. Most prominently, the Si dangling bond or K-center has been identified as an amphoteric trap center. Nevertheless, experiments have shown that these dangling bonds only make up a small portion of the total density of electrical active defects, motivating the search for other charge trapping sites. Here, we use a machine-learned force field to create model structures of amorphous SiN by simulating a melt-and-quench procedure with a molecular dynamics algorithm. Subsequently, we employ density functional theory in conjunction with a hybrid functional to investigate the structural properties and electronic states of our model structures. We show that electrons and holes can localize near over- and under-coordinated atoms, thereby introducing defect states in the band gap after structural relaxation. We analyze these trapping sites within a nonradiative multi-phonon model by calculating relaxation energies and thermodynamic charge transition levels. The resulting defect parameters are used to model the potential energy curves of the defect systems in different charge states and to extract the classical energy barrier for charge transfer. The high energy barriers for charge emission compared to the vanishing barriers for charge capture at the defect sites show that intrinsic electron traps can contribute to the memory effect in charge trap flash devices.
Silicon nitride is commonly used as the charge storage layer of nonvolatile charge trap flash (CTF) memory devices. Previous theoretical investigations have shown a large variety of possible trapping centers in which might contribute to the memory effect, including vacancies, dangling bonds, and overcoordinated atoms. So far, the high H concentration detected in silicon nitride thin films, resulting in a large number of passivated Si–H and N–H bonds, has not been considered in depth in theoretical studies regarding the charge trapping properties. In this work, we show that this passivation of dangling bonds with H gives rise to the formation of hole and electron (bi)polarons at fully coordinated sites and hole traps at Si–H bonds. We employ density functional theory combined with a hybrid functional to demonstrate that both holes and electrons can localize at intrinsic, fully coordinated sites in the amorphous :H network. We classify precursor configurations of the polaron states by analyzing bond lengths, partial charges, maximally localized Wannier functions, bond orders, and electronic states introduced in the band gap upon charge trapping. All trapping sites are characterized according to their thermodynamic properties, such as charge transition level (CTL) and relaxation energy on a statistical level. The CTLs of the (bi)polarons are in the energetic vicinity of the band edges of a Si substrate, showing that these intrinsic sites can trap charges from Si under applied voltages and contribute to the memory effect.
10 October 2024 in The Journal of Chemical Physics
Looking back at seven decades of highly extensive application in the semiconductor industry, silicon and its native oxide SiO2 are still at the heart of several technological developments. Recently, the fabrication of ultra-thin oxide layers has become essential for keeping up with trends in the down-scaling of nanoelectronic devices and for the realization of novel device technologies. With this comes a need for better understanding of the atomic configuration at the Si/SiO2 interface. Classical force fields offer flexible application and relatively low computational costs, however, suffer from limited accuracy. Ab initio methods give much better results but are extremely costly. Machine learning force fields (MLFF) offer the possibility to combine the benefits of both worlds. We train a MLFF for the simulation of the dry thermal oxidation process of a Si substrate. The training data are generated by density functional theory calculations. The obtained structures are in line with ab initio simulations and with experimental observations. Compared to a classical force field, the most recent reactive force field, the resulting configurations are vastly improved. Our potential is publicly available in an open-access repository.