Beyond Silicon's Elemental Logic

Almost since the first silicon MOSFET was invented, engineers have sought to construct versions using gallium arsenide or other III-V semiconductors, which would be able to operate at considerably higher speeds.

The main roadblock has been in finding a suitable material to use as a gate insulator. But in recent years, considerable progress has been made.

One technique uses molecular-beam epitaxy to deposit a gallium oxide-gadolinium oxide insulator on a III-V substrate. Another successful method is to deposit an aluminum oxide gate insulator using atomic-layer deposition, which is less technically demanding than molecular-beam epitaxy.

These and other approaches are bringing the day closer when engineers will be able to integrate millions of III-V MOSFETs into microprocessors or other digital ICs. By combining these transistors (which use electrons as charge carriers) with others made of germanium (which use “holes” as charge carriers), chip manufacturers should be able to build CMOS ICs that operate several times faster than those built from silicon.