Information technology is one of the key challenges of the 21
century. Beyond any
speculation of the stock markets this technological revolution requires powerful hard- and
software components. This steady development is described e.g. by Moore's law which predicts a
steady doubling of the number of integrated transistors in memory chips every 18-24
months [48]. Mobile communication and information systems undergo a rapid and highly
sophisticated technological change. Both software- and hardware components form a considerable
economic factor, a fact, which changes our understanding of industrial production.
In the recent past, mobile phone systems have undergone an exponential rise of
distribution. At the same time desktop computers have been
provided with a steady growth of computational power, and
saturation effects for the simple scaling of dynamic random access
memories (DRAM) and processors become visible. The next
generation of wireless phone systems, so-called third generation
(3G) cell phones, reach industrial maturity, while an intermediate
generation (2.5 G) of phones is being introduced. Services such
as UMTS (Universal Mobile Telephone Services) will provide mobile
broadband access to the internet data communication, with the
effect, that mobile cell phone communication merges with data
communication features so far known from the desktop computer
world. This encourages the development of mixed signal
technologies due to the increased signal processing of digital
data communication and the increased demands towards the analog
RF transmission. The overall development promotes a merge of the
desktop computer world and the mobile communication world, which
so far have been regarded as separate areas. Being exposed to a
mass market, every function of a cell phone, whether digital or
analog, is subject to an aggressive selection process with
respect to cost and functionality. This represents a new
challenge to III-V semiconductor components which were initially
developed for high prize niche markets.
At the same time, applications in higher frequency bands also develop into markets: The frequency range of 27 GHz-42 GHz has been allocated by national authorities all over the world in order to provide broadband internet access services for the so-called last mile between provider and customer. These services include LMDS (Local Multipoint Distribution Service) and MMDS (Multichannel Multipoint Distribution Service). A specific development issue is based on the linearity requirements needed for the systems and system components such as high power amplifiers. This is caused by the proposed use of specific access techniques such as CDMA (Code Division Multiple Access) techniques to allow for maximum band width. Furthermore, for higher mm-wave frequencies, i.e. at 77 GHz, automotive applications such as collision avoidance radars await their introduction from niche into mass markets.
This work is to deal with the development of High Electron Mobility Transistors (HEMTs) based on III-V semiconductor materials. Originally, HEMTs have attracted great interest in basic research due to their superior transport properties of the underlying III-V heterostructure quantum wells and the low dimensional electron gases. With regard to industrial production HEMTs have gained market shares in mass production from the MESFET in strong competition with the InGaP/GaAs Heterojunction Bipolar Transistor (HBT). Recently, the interest focuses on HEMT receiver circuits due to the excellent low noise properties. In contrast to the well known digital silicon industry, III-V HEMTs are well established with respect to analog RF-applications including for example low loss passive lines and backside substrate processes. These RF-processes are more mature than their rapidly developing analogue Silicon RF-opponents, see e.g. [319]. Being in mass production, III-V wafer size scaling is an issue to ensure cost efficiency in comparison with Silicon based technologies: 6 inch GaAs wafer technologies have been introduced and next generation 8 inch GaAs substrates for MESFETs and HEMTs are under development.
The growing importance of heterostructure devices for information technology has been underlined by the fact that Zh.I. Alferov and H. Kroemer were awarded the Nobel prize in 2000.
The development of device simulation tools for heterostructure devices is desirable since knowledge on the interaction between semiconductor processes and device behavior cannot be easily acquired. With the development of both, processes and device simulation tools, inverse modeling grants improved understanding of systematic and statistical changes. Until recently, III-V technologies had still the fume of single event processes. This has significantly changed, however, with several pseudomorphic HEMT processes being in mass production [301,316] on scaled wafer sizes. The principal functioning of HEMTs has been understood for some time [23,196]. However, the quantification of a large number of effects by device simulation allows, parallel to process development, the design and process control of HEMTs for the specific application needed.
The second chapter of this work gives a detailed overview and comparison of the state-of-the-art of RF devices and means to simulate them. The third chapter introduces the simulator MINIMOS-NT and the extraction of the simulation models and parameters for III-V devices, especially for HEMTs. Chapter 4 deals with the RF quantities and their extraction into compact models. Chapter 5 addresses methods of statistical data analysis suitable for the needs of III-V devices. Additionally, it gives an overview over the basic degradation effects of HEMTs and possible questions arising from the development of mature III-V technologies. Last, the technologies under investigation are described. Chapter 6 deals with device characterization using mainly power and active load-pull measurements up to 40 GHz. Chapter 7 contains several detailed simulation studies performed on GaAs, InP, and GaN based HEMTs. The work concludes with an outlook on further challenges for simulation and device technology.