The ‘information age’ has seen the significant emergence of wireless applications mainly dedicated to improving the quality of people’s lives (especially disabled and senior citizens), improving health, increasing the efficiency of air and road traffic, and bettering the environment. Over the last twenty years there has been a radical change from the days when communications were exclusively dedicated to military and governmental applications serving relatively few users to the situation today where the trend has been reversed with most applications devoted to the civilian market. The result is a huge increase in the number of customers, together with an increase in the number of applications, which results in an overcrowded frequency spectrum resulting from the continuous rise in the allocated frequencies and a radical change concerning the performance of the electronic modules required.
As is evident, there are various factors impacting on the development of RF components and systems that are being researched and investigated. For general communications systems, for example, there is the continuous movement towards smaller, secure systems that have increased functionality and reduced power consumption. These demands place severe constraints on circuit power dissipation and electromagnetic compatibility, and significantly increase the equipment design complexity, manufacturing costs and system weight.
One factor in particular that is gaining more and more importance is the noise exhibited by electronic modules. Today it is mandatory to realise high frequency receivers with very low noise behaviour. This requirement conflicts with the demand for increased operating frequencies, the reduction of power consumption and the rise in the number of users. Furthermore, the presence of several transmitters operating simultaneously on the same platform with multiple receivers requires very high dynamic range receivers, ultra-clean transmitters and careful attention to the overall electromagnetic compatibility design of the system. This usually requires filtering on both transmitters and receivers to ensure that they do not interfere with each other. Furthermore, the multiplication of the standard requires electronic modules to possess a certain degree of agility in order to optimise communication. The receiver or the transmitter has to switch to the standard or to the operator featuring the more efficient characteristics. There are also similar requirements concerning battery and performance management. In fact, relating to the electromagnetic environment, the performance of the receiver/transmitter can be more or less relaxed in order to save energy, which is a key issue in the case of portable communications. In this case the system has to be smart enough to choose the best configuration by trading off the electrical performance (that is, linearity and noise figure) and the power consumption.
These requirements can be summarised by the statement: “The RF chip has to work properly anywhere, anytime and has to be as cheap as possible.” The task of the system designers then can be tricky, as they have to identify the technology that will meet all of these requirements. To do so, the different ways that are being investigated today include:
- Miniaturisation
- Multifunctionality
- Heterogeneous integration
- Convergence between hardware and software technologies
These four main issues are explored through research at the material, technology, design, modeling and architecture level. With regards to the material level, competition now exists between the conventional approach that has been used in the past (GaAs, InP) and silicon-based materials (SiGe, MODFET, SiGe and HBT). For III-V based-materials, the more promising way seems to be the ‘metamorphic’ HEMT that combines the advantages of InP materials with those offered by GaAs. Furthermore, there is still research into heterogeneous epitaxy to grow III-V-based materials on silicon substrates and the emergence of a new family of nitride-based materials. GaN-based technology is showing very attractive capabilities for power applications in very harsh environments and has recently demonstrated interesting potential in the field of low noise applications, opening the way for very high integrated microwave systems in such conditions.
Concerning the silicon-based approach, SiGe-based technology (MODFET and HBT) is currently exhibiting frequency performance in the 400 GHz range. At the same time CMOS silicon-on-insulator (SOI) is showing significant potential with frequency performance above 100 GHz, thus opening up new paths for very highly integrated receivers featuring both analogue and digital processing potential offering more intelligence. CMOS for high frequencies necessitates research into dielectric materials in order to overcome the very high leakage current associated with reduced dimensions. This is achieved by using ‘high k’ material.
Finally, researchers are investigating an alternative approach that exploits nanomaterials and nanotechnologies. Some interesting results have already been produced that present the potential of carbon nano tubes, nanowires and the DOTFET approach that uses the quantum properties of a germanium-based island.
With regard to passive elements, the situation is becoming more complex as it is difficult to have materials featuring low insertion loss and a high quality factor from the microwave to millimetre-wave range. This is a major issue, particularly for miniaturisation, multifunctionality and heterogeneous integration. Here, we are seeing that III-V-based materials are still offering superior performance with respect to alternative silicon-based technology.
As for materials for passive elements, we are seeing that the multi-layer approach results in an improvement in electromagnetic propagation. The use of polymer technologies (through thick layer) is an interesting alternative, especially as it is easily compatible with any other technology (such as GaAs, GaN, InP and SiGe).
Another solution that is emerging is the exploitation of the micromachining capabilities of semiconductor materials such as silicon, GaAs and InP. In particular, silicon is a very promising candidate as it is very easy to micromachine and all the technological processes are very mature. Additionally, these technologies will be compatible with the integrated circuits (IC) process (including digital ones) that will make it possible to realise high frequency modules featuring a high level of integration. Another attractive advantage of silicon-based technologies relates to its mechanical properties that make it possible to realise mobile regions through electrostatic, magnetic or thermal excitation that will result in devices featuring tuneable behaviour.
All these concepts come under the name of microelectromechanical systems (MEMS). MEMS technology offers the performance advantages of electromechanical components on size scales commensurate with single solid-state components. In many cases, a single MEMS component replaces and outperforms an entire solid-state circuit. In other cases, a judicious association of MEMS components with active devices will result in smart communicating devices. This defines a new concept for microwave and millimetre-wave systems that will combine MEMS technology with integrated circuits. This concept is referred to as MEMSIC and could be achieved in two different ways. One option is MEMS in IC, where all the devices are fabricated on the same chip and the appropriate micromachined process executed to improve the performance of the passive circuit, release the multifunctionality properties and have the antenna surrounding the chip. This approach is risky and probably will be developed in five to 10 years. A less risky approach consists of grafting the functionality on existing integrated circuits that is known as ‘MEMS above IC.’ Here, the system performance is tailored by grafting additional material and devices above the integrated circuits that could also include the antenna.
MEMS technology can also aid in the development of new transceiver architecture based on mechanical resonance properties, which today can be used in the 10 to 100 MHz range (the only limitation being related to the size of the component). However, moving to the ‘nanoscale world,’ specifically nano electromechanical system (NEMS) devices will see a transfer to resonance frequencies in the microwave range. The general belief today is that future telecommunications systems will encompass such devices. Nevertheless, it is important to assess their performance and to investigate the best technological process in terms of cost, reliability, performance and compatibility with integrated circuits. Besides these efforts at both the material and technology level, it is also important to assess the efforts that will be needed in the field of modeling, design, and component and system architecture.
As for the design and modeling, miniaturisation, multifunctionality and heterogeneous integration in relation to an increase in the multi-physic coupling and the multi-scale problems, which are currently not very well covered by commercial tools, it is important to develop research in this area. It is also understood that now we have to take into account all of these phenomena in the design process and it is important to develop a modeling strategy that will integrate these issues.
The requirements for the component architecture are on a different level. First of all, we need to define a component architecture that features high bandwidth characteristics in order to cover high bit rate requirements. Another important consideration is the power amplifier where it is crucial to feature high power-added efficiency and exhibit high reliability. In the field of low level signals and ultra stable signals and clocks, the key challenges are to fabricate low noise amplifiers that are not affected by the matching termination and feature ultra-low noise figures and high bandwidth. For RF and microwave sources, the phase noise is a quantity that determines the performance of the overall transceiver, and efforts are being made to propose architectures that minimize the phase fluctuations. Finally, for all these components, it is important to assess their potential for the reconfigurability that is achieved in different ways — the analogue approach using MEMS technologies and the digital one using the CMOS gates.
System architectures will need to be very compact, very secure and composed of the relevant building blocks in order to have a very robust design phase. This is a major issue in the effort to develop highly complex microwave and millimetre-wave systems featuring reconfigurability, and repair and testability functionalities. Here, the architecture should be revisited as there needs to be high convergence between analogue and digital technologies so as to better exploit the potential of each approach. Today, for instance, there is a tendency to avoid the testing of microwave and millimetre-wave systems by having the test facilities embedded in the chip itself. This is defining a new family of systems called millimetre-wave built in self-test (MWBIST) circuits. Another very promising application is the development of sensor networks deployed ‘everywhere,’ featuring a very high level of autonomy and intercommunication to carry out different missions. In this field, research is being instigated to develop ultra small intelligent systems utilising sensors, actuators, information processing and communication media through ad hoc networks. These ultra small systems will need to feature advanced architectures combining analogue and digital facilities and should have some software embedded.
These small communicating systems are known as ‘smart dust’ and represent a key challenge for microwave and millimetre-wave systems as they will combine hardware and software architectures in order to feature advanced functionalities and intelligence. They will be used in different applications in the industrial, civil and health sectors as well as the defence sector. This kind of RFID network is beginning to be implemented, although there is a strong need for networks featuring a high bit rate and then communicating in the millimetre-wave range.
This will necessitate research efforts at the material level in order to emphasize the heterogeneous integration, miniaturisation and multifunctionality. At the system level, research will be conducted in the design and architecture phase in order to define appropriate partitioning between hardware and software technologies to have intelligence embedded in these smart dusts.
As a final conclusion, future RF systems will result from a two-fold convergence in order to give some intelligence and autonomy to these systems — a convergence between different technologies and materials, and a convergence between hardware and software technologies. They should also be as small as possible to be transparent with respect to potential users that will be surrounded by billions of chips proposing services for different sectors. This will be the wireless revolution for ubiquitous communication that will be a strong economical challenge in the 21st century.
Robert Plana was appointed director of the information and communication department at CNRS in 2005 and is now director of the engineering sciences department. In 2000, he became professor at Paul Sabatier University and Institut Universitaire de France, and started a research team at LAAS-CNRS in the field of micro and nanosystems for RF and millimetre-wave communications. Its main activities are in the technology, design, modeling, test, characterisation and reliability of RF MEMS for low noise and high power millimetre-wave applications, and the development of the MEMS IC concepts for smart microsystems. He has also built the AMICOM European Network of Excellence in this field, which encompasses 25 research groups.
