Compound Semiconductors for Low-Power p -Channel Field-Effect Transistors
- PDF / 794,022 Bytes
- 7 Pages / 612 x 792 pts (letter) Page_size
- 84 Downloads / 221 Views
Semiconductors for Low-Power p -Channel FieldEffect Transistors
Brian R. Bennett, Mario G. Ancona, and J. Brad Boos Abstract Research in n-channel field-effect transistors based upon III–V compound semiconductors has been very productive over the last 30 years, with successful applications in a variety of high-speed analog circuits. For digital applications, complementary circuits are desirable to minimize static power consumption. Hence, p -channel transistors are also needed. Unfortunately, hole mobilities are generally much lower than electron mobilities for III–V compounds. This article reviews the recent work to enhance hole mobilities in antimonide-based quantum wells. Epitaxial heterostructures have been grown with the channel material in 1–2% compressive strain. The strain modifies the valence band structure, resulting in hole mobilities as high as 1500 cm2/Vs. The next steps toward an ultra-low-power complementary metal oxide semiconductor technology will include development of a compatible insulator technology and integration of n- and p-channel transistors.
Introduction While silicon has long dominated electronics, compound semiconductors that are formed from combinations of Group III and Group V elements (e.g., GaAs) have nevertheless remained of interest because their higher electron mobilities and velocities allow for faster devices of a given size and power. Over the last three decades, the advantages of III–V semiconductor field-effect transistors (FETs) have been exploited in a variety of highperformance, front-end analog applications (e.g., radar systems, cell phones). For digital applications, however, highperformance III–V FET technology analogous to Si complementary metal oxide semiconductor (CMOS) has been difficult to achieve due to the lack of a suitable insulator, higher levels of defects, and the fact that the hole mobilities are generally much lower than the electron mobilities. In this article, we review recent efforts 530
aimed at raising the mobility of holes in III–V materials in the hopes of making a III–V CMOS technology viable. Because III–V semiconductors lack a good insulator, the standard approach for making III–V FETs has been different than for silicon FETs. Growth techniques such as molecular beam epitaxy (MBE) are used to create heterostructures in which the electron transport is restricted to a channel layer by cladding it with wider energy-gap barrier layers. The first such devices employed GaAs channels and AlGaAs barriers,1 and in the years since, the primary method for enhancing performance has been simply to choose channel materials with better electron transport properties. The spectrum of possibilities open to such III–V heterostructure design is best revealed by the plot of the energy bandgap of various semiconductors versus their lattice constant
shown in Figure 1. In general, the history of n-channel III–V transistors has progressed to the right in this diagram because the narrower bandgaps and smaller electron effective masses of those semiconductors offer hi
Data Loading...
