Blog

IBM Integrates III-V Materials On Silicon

IBM is the first to overcome the lattice mismatch between silicon and III-V materials (including indium, gallium, arsenide and their compounds) by successfully fabricating the complex and versatile structures necessary to manufacture hybrid silicon/III-V transistors.  This includes constrictions along a nanostructure, nanowire cross junctions, and 3D stacked nanowires.  Known as compound semiconductors, they are already in use in high-performance settings required by applications like military radio transceivers.  However, to date, they have not been used in consumer products due to production cost, and defect density to name two factors.

Using template-assisted selective epitaxy (TASE), a method pioneered by IBM of depositing ultra-fast III-V nanowires suitable for transistor channels and other structures on silicon-on-insulator (SOI) substrates, IBM has achieved what semiconductor manufactures around the world have been trying to do for many years.

This is an important step towards making computer chips that allow integrated circuits to continue shrinking in size and cost while increasing in performance.

IBM anticipates this accomplishment to also lead to active photonics on silicon substrates since III-V materials are routinely used to make on-chip lasers and other photonic structures.

IBM has laid a solid foundation to the integration of III-V materials with silicon and intends to pursue more development and optimization in order to achieve the same control over performance in III-V devices as currently exists for silicon.  IBM’s new process is key to integrating the stacked materials on the silicon platform in a cost effective manner.

According to IBM representatives, what sets this work apart from other methods is that the compound semiconductor does not contain detrimental defects due to the TASE process.  This process grows new crystals using metal organic chemical vapor deposition which starts from a small area and evolves into a larger, defect-free crystal.  Because this process is fully compatible with current chip fabrication technology it is also economically viable.

This breakthrough by IBM is expected to allow an extension to Moore’s Law, the famous observation by Gordon Moore that the number of transistors on an integrated circuit double about every two years.  According to Heinz Schmid, an IBM researcher, “The whole semiconductor industry wants to keep Moore’s Law going.  We need better performing transistors as we continue down-scaling, and transistors based on silicon won’t give us improvements anymore.”

On the consumer front this means a continued trend of computer devices having increased speed and bandwidth coupled with reduced power consumption and cost.