ABSTRACT
It has been more than three decades since the high-electron-mobility transistor (HEMT) was invented [1], revolutionizing the world of high-frequency electronics. First, it was made on GaAs using an n-AlGaAs/GaAs heterostructure, then on InP with an n-InAlAs/InGaAs heterostructure, later on GaAs using an InAs/AlSb heterostructure and a metamorphic buffer, so-called antimonide-based-compound semiconductor (ABCS), and more recently on GaN with an i-AlGaN/GaN heterostructure. Since their inception, HEMTs have steadily been improved to deliver higher levels of performance in terms of operating high-frequency gain, noise, and power. Today,
InAs HEMTs on a GaAs substrate exhibit the best-balanced highfrequency figures of merit (FOMs) in any transistor technology on any material system [2]: high fT and high fmax simultaneously. As a result, InAs HEMTs are uniquely positioned to become the first true THz semiconductor transistor with both fT and fmax > 1 THz. This book chapter reviews the evolution of III-V HEMTs for the past three decades and discusses steps to be taken to reach the THz regime. 6.1 IntroductionThe high-electron-mobility transistor (HEMT), also known as heterostructure field-effect transistor (HFET) or modulation-doped field-effect transistor (MODFET), was first invented by Mimura et al. at Fujitsu Labs in 1980 [3]. The HEMT was based on the concept of modulation doping, first reported by Dingle et al. at Bell Laboratories in 1978 [4]. In principle, the HEMT is an FET that incorporates a junction between two semiconductor materials with different band gaps, forming a heterostructure. Selective doping (also known as modulation doping) of the wide-band-gap semiconductor creates a two-dimensional electron gas (2DEG) on the narrow band-gap material right at the interface, forming a triangular-shaped potential well. The spatial separation between dopants and the 2DEG greatly enhances the mobility of electrons in the potential well that can significantly exceed the bulk mobility value even at relatively high carrier concentrations. This is exactly why this device was named a “high-electron-mobility transistor.” The first demonstration of modulation doping was observed in the AlGaAs/GaAs material system [3]. This was a consequence of molecular beam epitaxy (MBE), which allowed tight control of atomic-layer growth of various compound semiconductors such as GaAs and AlxGa1-xAs. With the aid of MBE, heterostructures approaching monolayer-level interface abruptness became possible, heralding a new era of “band-gap engineered” devices. This is at the heart of modulation-doped structures. This chapter traces the most significant steps of the evolution of the HEMTs from the high-frequency-operation point of view.
