Electron trap states at InGaAs/oxide interfaces under inversion through constant Fermi-levelab initiomolecular dynamics

Abstract

We employ constant-Fermi-level ab initio molecular dynamics to investigate defects at the InGaAs/oxide interface upon inversion. We adopt a substoichiometric amorphous model for modelling the structure at the interface and investigate the formation of defect structures upon setting the Fermi-level above the conduction band minimum. The defect formation is detected through both an analysis of the atomic structure and a Wannier-decomposition of the electronic structure. This computer driven approach is able to retrieve In and Ga lone-pair defects and As-As dimer/dangling bond defects, in agreement with previous studies based on physical intuition. In addition, the present simulation reveals hitherto unidentified defect structures consisting of metallic In-In, In-Ga, and Ga-Ga bonds. The defect charge transition levels of such metallic bonds in AlO are then determined through a hybrid functional scheme and found to be consistent with the defect density measured at InGaAs/AlO interfaces. Hence, we conclude that both In and Ga lone pairs dangling bonds and metallic In-In bonds are valid candidate defects for charge trapping at InGaAs/oxide interfaces upon charge carier inversion. This study demonstrates the effectiveness of constant-Fermi-level ab initio molecular dynamics in revealing and identifying defects at InGaAs/oxide interfaces.

Defect agglomeration induces a reduction in radiation damage resistance of In-rich In x Ga1−x N

“…The occurrence of the In-In bond indicates the agglomeration of In elements. The bond lengths of In-In, In-Ga and Ga-Ga in the present work are about 2.32, 2.14 and 2.12 Å, while Bouzid's ab initio calculation [38] shows that the bond lengths of In-In, In-Ga and Ga-Ga are 2.70, 2.56 and 2.43 Å. Both of the results show that the bond length of In-In is larger than In-Ga, while the bond length of In-Ga is larger than Ga-Ga.…”

“…It has been observed in experiments that the In element in InGaN tends to agglomerate during the growth of InGaN materials [37]. In addition, Bouzid and Pasquarello [38] employed constant-Fermi-level ab initio MD to investigate the InGaAs/oxide interface structure and the results showed that metallic In-In, In-Ga and Ga-Ga bonds could be found in InGaAs. In figure 7, the In-In, In-Ga and Ga-Ga partial pair correlation functions of crystalline and amorphized In 0.3 Ga 0.7 N are plotted.…”

Reliability of Buried InGaAs Channel n-MOSFETs With an InP Barrier Layer and Al2O3 Dielectric Under Positive Bias Temperature Instability Stress

“…However, one of the most critical problems that must be solved to realize III-V MOSFETs is the formation of a stable MOS interface with low trap density [7]. Compared with the SiO 2 /Si system, the III-V native oxides negatively affect fermi-level pinning and current drift [8][9][10]. The atomic layer deposited (ALD) Al 2 O 3 dielectric in surface InGaAs channel MOSFETs can achieve a thermally stable interface and large band offsets, as confirmed by the previous research on the dielectric layer of InGaAs MOSFETs [11][12][13].…”

Reaction pathway of oxygen evolution on Pt(1 1 1) revealed through constant Fermi level molecular dynamics

“…In the constant Fermi level scheme, the mass of the fictitious charge degrees of freedom is set to M e = 100 a.u. The fictitious electronic temperature is controlled by a velocity rescaling method and is set to T e = 30 K. Gaussian-smeared occupations with widths of at most 0.10 eV are used to prevent numerical instabilities[27,33,34,26]. In the present dynamics, the inertial mass is set in such a way that the fluctuations of the grand canonical potential occur on a time scale that accompanies the struc-tural rearrangements, to account for smooth transitions upon charge variations in the physical system.…”

“…The constant Fermi level molecular dynamics [27,33,34,26] in this work relies on a description of the electronic structure within density functional theory (DFT) through an optimized rVV10 functional [35,36,37]. In the rVV10 functional used in this work, the b parameter is set to 9.3 to ensure a good description of the water density [37].…”