Silicon is a superb material for semiconductor processing. It is highly unlikely that any material will supplant silicon substrates for volume manufacturing of semiconductor products. However, as we replaced conventional gate oxides with HfO2 oxides, and conventional poly gates with metal-gates, there is a possibility that the active transistor channel may be replaced by another material in the next 10-15 years. There are two strong categories of advanced materials being considered for channel replacement, namely, germanium (Ge) and the III-V materials (GaAs, InAs and so on). These materials can be graphically summarized by a plot with the lattice constant (distance between atoms in the crystal) on the X-axis and the bandgap energy on the Y-axis. Note that the majority of these materials (except AlP and GaP) have lattice constants larger than silicon, and many of the materials (InP, GaSb, InSb, InAs, and Ge) have bandgaps smaller than silicon. When materials of a different lattice constant are grown on silicon, the lattices donít match up, creating dislocations and other defects. When materials are used of a smaller bandgap than silicon (Si band-gap~1.1V, Ge ~ 0.75V, InGaAs ~0.5V), it enables good performance at lower operating voltages but creates the possibility for leakage mechanisms at regular or turbo operating voltages.
Germanium is a very deeply researched semiconductor material. (In fact, the first transistors were made from germanium!). The unfortunate situation with germanium is that it has a very poor native oxide (GeO). In contrast, the native oxide on silicon is glass (SiO2) which forms an excellent interface between a well-behaved semiconductor and a well-behaved dielectric. At thin electrical oxide thickness (TOXE), ALL industry/university data show degraded mobility. A radically new oxide invention will be required to enable a Ge/SiGe channel for future technology nodes.
The III-V materials are widely used for optical and very high speed electronic applications. Implementation of III-V materials for semiconductor channels on silicon is extremely challenging for both the fabrication and bandgap reasons detailed earlier. However, significant progress has been made in recent years with InGaAs quantum-well transistors.