Compound semiconductors on silicon epitaxy


To calibrate performance achieved, we will compare with structures grown on conventional substrates. This will also aid development of the necessary fabrication process modules (WP 2). The focus will be to establish compatible epitaxial growth processes for various materials and device structures grown on Si.


We will combine approaches to minimise defect production at the Si to CS interface, the use of nanopillars, and subsequent growth coalescence, and the use of defect filter layers to reduce the number of defects reaching the active region. We will examine the relative benefits of quantum dots for carrier localisation and nanopillar active regions, to exclude defects, as mechanisms to cope with those defects that do reach the active region. This will build on our international lead and other EPSRC support. To integrate different CS (and ultimately also Si functionality) on the same substrate we will fabricate deep trenches in SOI (Si on Insulator), grow a Si buffer to bury contaminants and smooth the surface, and then selectively grow CS materials on the smooth Si in the trenches.

III-Sb (CF, UCL): Growth of III-Sb are a particular expertise of Prof Huffaker. Challenges will include the development of very high uniformity InSb for application to RF electronics and solid state qubits and the development and transfer of InAs / GaSb based Type II strained layer superlattice to growth on GaAs substrates for Grand Challenge 2.

III-As (UCL, CF): The focus will be on developing uniformity and reproducibility, including between multiple areas of selective growth over large substrates, and approaches for defect minimisation shown to work with other materials, such as nanowires for long wavelength (red) emitters for GC1.

III-N (Sheffield, CF): To overcome problems with polar (c-plane) GaN, including long radiative recombination lifetime (10 ns) that limits modulation bandwidth in communications and leads to poor efficiency especially at longer wavelengths, we will use semi- or non -polar GaN. We will develop semi-polar (11-22) GaN, which can accommodate more indium atoms than both polar and non-polar GaN, for longer wavelength emission and facilitate “normally off” devices for RF switching power amplifiers and power convertors.