Abstract:
GaN-On-Silicon (GOS) structures and techniques for accommodating and/or controlling stress/strain incurred during III-N growth on a large diameter silicon substrate. A back-side of a silicon substrate may be processed to adapt substrates of standardized diameters and thicknesses to GOS applications. Bowing and/or warping during high temperature epitaxial growth processes may be mitigated by pre-processing silicon substrate so as to pre-stress the substrate in a manner than counterbalances stress induced by the III-N material and/or improve a substrate's ability to absorb stress. III-N devices fabricated on an engineered GOS substrate may be integrated together with silicon MOS devices fabricated on a separate substrate. Structures employed to improve substrate resilience and/or counterbalance the substrate stress induced by the III-N material may be further employed for interconnecting the III-N and silicon MOS devices of a 3D IC.
Abstract:
Transistors or transistor layers include an InAlN and AlGaN bi-layer capping stack on a 2DEG GaN channel, such as for GaN MOS structures on Si substrates. The GaN channel may be formed in a GaN buffer layer or stack, to compensate for the high crystal structure lattice size and coefficient of thermal expansion mismatch between GaN and Si. The bi-layer capping stack an upper InAlN layer on a lower AlGaN layer to induce charge polarization in the channel, compensate for poor composition uniformity (e.g., of Al), and compensate for rough surface morphology of the bottom surface of the InAlN material. It may lead to a sheet resistance between 250 and 350 ohms/sqr. It may also reduce bowing of the GaN on Si wafers during growth of the layer of InAlN material, and provide a AlGaN setback layer for etching the InAlN layer in the gate region.
Abstract:
A fin over an insulating layer on a substrate having a first crystal orientation is modified to form a surface aligned along a second crystal orientation. A device layer is deposited over the surface of the fin aligned along the second crystal orientation.
Abstract:
Embodiments include high electron mobility transistors (HEMT). In embodiments, a gate electrode is spaced apart by different distances from a source and drain semiconductor region to provide high breakdown voltage and low on-state resistance. In embodiments, self-alignment techniques are applied to form a dielectric liner in trenches and over an intervening mandrel to independently define a gate length, gate-source length, and gate-drain length with a single masking operation. In embodiments, III-N HEMTs include fluorine doped semiconductor barrier layers for threshold voltage tuning and/or enhancement mode operation.
Abstract:
A III-N semiconductor channel is formed on a III-N transition layer formed on a (111) or (110) surface of a silicon template structure, such as a fin sidewall. In embodiments, the silicon fin has a width comparable to the III-N epitaxial film thicknesses for a more compliant seeding layer, permitting lower defect density and/or reduced epitaxial film thickness. In embodiments, a transition layer is GaN and the semiconductor channel comprises Indium (In) to increase a conduction band offset from the silicon fin. In other embodiments, the fin is sacrificial and either removed or oxidized, or otherwise converted into a dielectric structure during transistor fabrication. In certain embodiments employing a sacrificial fin, the III-N transition layer and semiconductor channel is substantially pure GaN, permitting a breakdown voltage higher than would be sustainable in the presence of the silicon fin.