Jin Cho from the University Of Glasgow, UK, talks to Electronics Letters about his paper ‘The Impact of Stress in ICP-CVD SiNx Passivation Films on the Leakage Current in AlGaN/GaN HEMTs’, page 947. Jin Cho My current research interest includes development of high power, high breakdown voltage, and low switching energy loss in metal oxide semiconductor high electron mobility transistors (MOS-HEMT) using gallium nitride (GaN) on silicon (Si). Compared to Si power devices, wide band gap power devices such as GaN transistors offer great performance improvements due to their significantly better material properties. GaN has a higher bandgap, breakdown field, and electron mobility compared to Si, which leads to greater performance in power electronics. Since improving the performance of power converters will lead to saving energy consumption of electrical machines in factories, I am particularly interested in GaN power transistor design, fabrication, and characterisation for power converter applications Improvement of the on- and off-switching speed and minimised switching energy loss are essential parameters for all of power converter applications. To do this, we need to minimise the two main leakage currents in a GaN transistor, namely the off-state drain-to-source, and the gate leakage currents. These leakage currents are affected significantly by the different types of passivation material and/or deposition technique. Passivation is a vital fabrication step for all of multi-layer thin film devices. Silicon nitride (SiNx) has been broadly used as a passivation film material and shown to improve device performance. However, several papers have been reported leakage currents of GaN transistors also being increased by SiNx passivation film. To address this problem, we investigated the characteristics of SiNx passivation films and their impact on leakage currents in GaN transistors. To mitigate leakage currents, several methods have been reported in the literature such as aluminium oxide (Al2O3) passivation film deposited by atomic layer deposition, wet chemical, or plasma surface treatment before passivation and rapid thermal annealing after passivation. In this Letter, we proposed to reduce leakage currents by exploiting the physical stress in the SiNx passivation scheme. Most thin films tend to be in either tensile or compressive stress. Indeed, a few papers have been investigating stress of SiNx in the passiviation film of GaN HEMT. To date, all reports on the impact of stressed SiNx films have been restricted to passivation layers deposited by plasma enhanced chemical vapour deposition (PECVD) techniques with a maximum compressive stress of 150 MPa. Up to this level of stress, no change or improvement in the transistor leakage currents was seen. In our work, we realised leakage current mitigation under both tensile and compressive stress in the range of −1622 to +440 MPa for room temperature deposited inductively coupled plasma chemical vapour deposition (ICPCVD) SiNx passivation films. A significant reduction in off-state drain-to-source and gate leakage currents was observed for the optimally stressed films. The key parameter to reduced leakage currents is to realise the high stress level of SiNx deposition. Properties of SiNx film show that the leakage current can be changed by variation of the stress level. Due to two kinds of plasma power, ICP-CVD can make very high tensile/compressive type of SiNx. Off-state drain-to-source and gate leakage current was reduced by up to four orders of magnitude when compared to unpassivated devices. However, high compressive stressed SiNx films were shown to be randomly cracked and had delamination issues across the surface. It was discovered that this situation could be overcome by first depositing SiNx films using low or medium stressed process, followed by the higher stress films. In both the short and longer terms, high stressed SiNx film can be directly used as passivation for all of GaN based power or radio frequency (RF) HEMTs to mitigate their unwanted leakage currents. Further optimisation work for this approach would be useful. We are using the developed passivation technique in the realisation of new GaN power device that we are developing. The work includes assessing the switching characteristics of these devices. I have studied GaN based transistors since 2010. I see the key challenges for this technology being to increase cost effectiveness and reliability. In my opinion, over the next ten years, the future challenges will be how to improve device performances and reliability under extreme environments.
Read full abstract