In recent years Gallium Nitride (GaN) semiconductor devices have matured from the research labs into products. Discrete GaN transistors as well as GaN MMICs are now commercially available from a number of manufacturers. The released commercial offerings are mainly focused on power amplification up to Ku-band. In some applications, notably radar and electronic warfare, the benefits of GaN is not restricted to power amplification. Integration of low noise amplifiers, transmit/receive switches and power amplifiers in a single GaN chip could provide improved system performance. So far multifunction GaN circuits have been fabricated with devices optimized for power amplification. Hence, there is an opportunity to optimize a GaN MMIC process for multifunctional circuits. There is also a need to benchmark commercially available processes in terms of performance for specific or multifunctional applications.
GaN poses a number of challenges, compared to traditional compound semiconductor devices such as GaAs pHEMT and Si LDMOS. In particular the high electric fields, traps and electro-thermal interaction will play an important role in the operation of these devices. The trapping and self-heating introduce long term memory effects that needs to be included in behavioral models used for linearization of wireless transmitters. It is therefore important to characterize these effects to improve these models, and hence enable GaN transistors to be used in applications with very high demands on linearity, such as radio base stations for mobile communication. Furthermore, it is of interest to compare different semiconductor materials in terms of trapping and electro-thermal effects to select the most appropriate technology for a specific task.Objectives
The first objective of the project is to increase the understanding of dispersive effects such as trapping and self-heating using advanced characterization techniques. Gaining knowledge in the time constants in these effects should improve the models, and hence enable better linearization of transistors. Furthermore the characterization of trapping and electro-thermal effects will be used as feed-back to the device processing for evaluating different epi-layer stacks and processing steps. These characterization methods are technology neutral and will be applied not only to Gallium Nitride based devices, but also to other compound semiconductors as well as Silicon.
The secondary objective is to optimize device structures and concepts for specific subsystem functions. In particular, GaN epi-layer stacks will be optimized for multifunction circuits such as monolithically integrated transceiver circuits. Multifunctional circuits will be manufactured in the optimized process for evaluation and benchmarking with commercial MMIC processes. Furthermore, process modules will be evaluated in terms of their effect on circuit and system performance.
This project will also develop models, and model parameter extraction methods, which allows accurate circuit design as well as evaluation of system performance. For the design of circuits the project will focus on methods that can provide quick turnaround application specific models.
The vision with ACC is that the results will help Swedish industry to eventually start to implement GaN HEMT technology in their RF system products, both for telecom power transmitters and defence radar arrays. This will be made by further studies of the delicate components with regard to memory effects etc which still may hinder their use in practical systems. Moreover, benchmarking against established LDMOS is also on the agenda. The functionality of integrated GaN in MMICs is another essential design topic which lends itself for direct exploitation in products. This is closely connected to the model development aided by advanced characterization techniques developed in ACC.