Modeling dislocation-related reverse bias leakage in GaN p–n diodes. (28th May 2021)

• Record Type: Journal Article
• Authors: Qwah, K S; Robertson, C A; Wu, Y-R; Speck, J S
• Is Part Of: Semiconductor science and technology. Volume 36:Number 7(2021)
• Publication Date: 2021-05-28
• Subjects: GaN -- modeling -- reverse-bias leakage -- dislocations -- p–n diodes; Semiconductors -- Periodicals; 621.38152
• DOI: 10.1088/1361-6641/abfdfc
• Languages: English
• ISSNs: 0268-1242

Abstract

Abstract: Finite element analysis software was used to model and visualize two p-n junction models: one with a single threading dislocation (TD) and a control without one. TDs are modeled as a Gaussian distribution of trap states with a full width at half maximum value of 5 nm localized around the r = 0 line in a cylindrical coordination such that the linear trap state density was 1 trap c −1 -translation; this model allows the cylindrical symmetry of the c -plane GaN crystal orientation to be used to avoid more computationally intensive 3D models. In this work, a vertical p–n diode with typical doping characteristics and an equivalent threading dislocation density of 10 8 c m − 2 was modeled in reverse bias. Our simulations show that the dislocation-mediated leakage mechanism for reverse bias leakage in GaN p–n diodes is the generation of electron–hole pairs via a trap-assisted tunneling mechanism whereby electrons from the valence band use the intermediate trap state to traverse the band gap. This mechanism results in electron–hole pairs that are swept out of the junction by the reverse bias electric field. This behavior results in a measurable leakage current within the model with behavior consistent with experimental values.