Vertical GaN-on-Si membrane power transistors: Efficient power electronics for mass-market applications (VertiGaN)`

Lead Research Organisation: University of Bristol
Department Name: Physics


This project aims to realize transformative vertical gallium nitride-on-silicon (GaN-on-Si) transistors with breakdown voltage in excess of 1200 V. Power electronics is essential in applications including power distribution and transportation, with inefficiency of power electronic systems estimated to account for 20% of global carbon emissions. Furthermore, emerging low-carbon technologies, including electric vehicles and renewable energy generation, require power electronic devices with significant improvements over existing Si based solutions. GaN is a wide bandgap semiconductor alternative to Si, with superior power electronic material properties. Commercially-available lateral GaN transistors show good power performance, but are generally unsuitable for applications >1000 V due to high on resistance and large chip area. Vertical GaN transistors (where current flows into the plane of the chip, rather than along the surface) offer a step-increase in efficiency and power density over Si-based devices currently dominant in power electronics at voltages exceeding 1000 V.

Large-scale commercialisation of vertical GaN devices is currently inhibited by the requirement for expensive and unsustainable GaN bulk substrates. Transfer to sustainable Si substrates as proposed here, with a cost reduction of >1000x, requires management of associated material defects, to be achieved in this work through of implementation of novel device structures and optimisation of material growth processes. Demonstration of vertical GaN on Si transistors with breakdown voltage of >1200 V (i.e. voltage at which device failure occurs), improved from <600 V in previous attempts, will enable exploitation of the outstanding GaN material properties in emerging mass market applications at >1000 V, unlocking new applications and enabling reduced carbon emissions in next-generation power electronic systems including electric vehicles and power distribution.

Breakdown voltage in vertical GaN-on-Si transistors will be increased through improvement of material quality in the active device drift region. The novel structure will use an epitaxially-embedded n+GaN drain contact layer to facilitate a drain-recessed membrane device architecture, eliminating low-quality material from the active device region. In parallel, optimisation of epitaxial growth techniques will produce GaN-on-Si material with increased total thickness and a reduction in both dislocation density and background impurity levels. Drain-recessed GaN-on-Si membrane structures will then be integrated with finFET device topologies, shown to withstand operation voltages >1200 V in GaN-on-GaN, resulting in transistors with enhanced off-state blocking and on-state electron transport characteristics. The development workplan, in close collaboration and with strong support by industry, will enable both a thorough exploration of the underlying physics determining vertical breakdown in GaN-on-Si and improvements in device performance toward that required for large-scale commercialisation. Comprehensive failure analysis via reliability/stability testing and multiphysics modelling will provide further understanding of the GaN-on-Si material system and commercial potential.

Technology demonstrators will be optimally positioned for integration with next-generation manufacturing chains and testing systems, ensuing maximum commercial impact. This will be achieved through regular consultation with the Project Steering Committee, consisting of UK-based manufacturers of power electronic materials, devices and systems, as well as academics and a prominent UK government policy influencer. The use of a Design Kit to promote the benefits of the technology to system designers and manufacturers will ensure maximum uptake and identification of additional application areas, toward achieving wide-scale use of GaN devices and an associated reduction in carbon emissions from inefficiency of power electronics.


10 25 50