Metamorphic transistors: Building blocks for hetero-integrated circuits
- PDF / 642,803 Bytes
- 8 Pages / 585 x 783 pts Page_size
- 56 Downloads / 229 Views
oduction Since its invention by Shockley, Bardeen, and Brattain in 1947,1 the transistor has become the fundamental building block of almost all modern electronic circuits and systems. Its importance stems from the fact that an electrical signal applied between one pair of terminals controls the current flowing through another pair of terminals. This allows transistors to function as amplifiers or switches of electrical signals that enable logic operations and power generation/management. Today, Si transistors can be mass produced cheaply on a per-device basis with a high yield and tight specifications using highly automated and highly scalable semiconductor manufacturing processes. Advances in fabrication technology have resulted in a rough doubling of transistor counts per integrated circuit every two years, in accordance with Moore’s Law.2 While Si has proven to be exceptionally suitable for digital processing and logic applications, it loses out to other materials such as the III–V compound semiconductors for radio frequency (RF), high power, and optoelectronic applications.
Benefits of metamorphic transistors Transistors can be made from new materials with relative simplicity if a starting substrate with identical crystal structure and lattice parameters are available for the desired device material. However, if no such native substrate exists, metamorphic
epitaxy is required. This is especially the case when dealing with the III–N material system, where large-area bulk GaN or AlN substrates are not readily available. Additionally, metamorphic epitaxy affords circuit designers access to lattice constants and material properties (e.g., bandgap, mobility) that cannot be readily grown on commercially available binary substrates (e.g., GaAs, InP). A key example is the use of metamorphic buffer layers to access In0.3Ga0.7As and In0.3Al0.7As alloys that have lattice constants between those of GaAs and InP for metamorphic InAlAs/InGaAs highelectron-mobility transistors (HEMTs). Their higher carrier mobility compared to AlGaAs/GaAs HEMTs, and larger channel bandgap than conventional InP/In0.53Ga0.47As HEMTs, potentially offer a superior combination of speed and breakdown for various applications.3 Another potential avenue for performance improvements is nanopatterned metamorphic epitaxy that uses selectively grown III–V materials to create high-mobility channels on a Si platform. Even when appropriate bulk substrates are available, metamorphic techniques allow for the use of cheaper and larger non-native substrates that may offer significant production cost reduction, such as the use of GaAs substrates for InP-based HEMTs and heterojunction bipolar transistors (HBTs).4 GaAs substrates are cheaper, available in larger sizes, and less fragile than InP substrates; consequently, they can reduce per-die device
Kenneth E. Lee, Singapore-Massachusetts Institute of Technology Alliance for Research and Technology, Singapore; [email protected] Eugene A. Fitzgerald, Department of Materials Science and Engineering, Massachus
Data Loading...
