The performance of radio frequency (RF) integrated circuits will strongly influence the versatil-ity and portability of future wireless communication systems. With the ever increasing demands for higher bandwidth and capacity as well as reductions in size weight and cost, the need for more robust and efficient RF circuits is expected to increase. Currently, millimeter-wave monolithic ICs (MMICs) chip sets are under development in the 24-94 GHz range and will represent the platform for the RF components of most wireless systems. With the recent advent of micro-electro-mechanical (MEM) systems, new potentials are being discovered for applications in the RF/millimeter wave ranges.

In order to implement RFICs for future wireless communication systems, several fundamental issues must be resolved. First, a design methodology must be devised and tested; next, a low cost solution for the integration of MEM technology into existing MMIC processes must be developed as well as a reliable packaging scheme; finally, a robust platform of design guidelines, tools and characterization techniques must be made available to insure reliable implementation of these communication systems. More specifically, the effort will focus on:

• Developing a computer-aided design environment for the robust implementation of MMICs and MEMs as well as the generation of reliable design guidelines.
• Implement a robust MEM process for the fabrication of switches, filters and tunable capacitors.
• capacitors.
• Integrate MEM components into a reliable and well-established MMIC process.
• Devise a low-cost and reliable packaging solution for RFICs.

Executions of these tasks will make use of the existing infrastructure and expertise of the Pis in the related areas. These will define several design-build-test cycles which will be optimized through several iterations during the course of the project. Special attention will be devoted in demonstrat-ing the feasibility of these various tasks. Moreover, a testbed prototype will be implemented to validate the proposed flow and assess the hardware advantages of the designed RFICs.

Recent developments in the area of wireless communication systems and micro-electromechanical systems (MEMS) has enabled the networking of distributed transducers in a wireless mode. It is now possible to integrate monolithic microwave integrated circuit (MMIC) front-end modules with MEMS components such as antennas, switches and filters. Our objective is to supply the necessary CAD tools to improve first-pass success and reduce design iterations for such systems. In particular, electromagnetic techniques are used to model various MEMS switch structures and combined with simulation techniques to predict the transient and steady-state response of these components. The goal is to reduce the design cycle from several man-years to one man-week in the successful implementation of these MEMS structures.

There is tremendous demand for increased capacity in high speed communication networks and novel applications in optical control of antenna phased arrays. With clock rates in the GHz range, interconnect considerations and electromagnetic phenomena have moved to the forefront in the design of high-speed computers. Microwave modulation of optoelectronic devices such as semiconductor lasers and modulators plays an important role in determining the high-speed performance of these devices. Continuing development of high speed optical communication systems is contingent upon advances in high speed sources and wavelength conversion devices.

While numerous theoretical models have been developed which predict ways to improve these devices, much experimental work remains in order to verify these models and characterize devices based on new designs. This project proposes the experimental validation of various computational models in inverse scattering, wave propagation, interconnects and optoelectronics. Most of the emphasis will be on the higher frequency range where measurement information is nonexistent. Computational electromagnetics has received in recent years growing interest due to the availability of fast computers and the recent development of fast algorithms. Models that allow the prediction of the most complex problems in scattering and wave propagation have been implemented. Unfortunately, the experimental verification of these models is seriously lagging especially at higher frequencies where measurement accuracy and repeatability are more difficult to achieve. This project will allow us to determine the frequency range of validity of the computational models and generate useful information for high-frequency operation of radiating, optoelectronics and waveguiding systems.

The electromagnetic modeling of packages and interconnects plays a very important role in the design of high-speed digital circuits, and is most efficiently performed by using computer-aided design algorithms. In the past two decades, researchers in the electromagnetic and microwave areas have striven to extend the knowledge of interconnection properties. Their efforts have resulted in models and analytical methods without which the development of reliable design tools would be impossible. Packaging and interconnects nowadays represent a critical area for the design of high-performance digital systems. State-of the-art computational electromagnetic techniques necessitate large processing power and memory requirements.

As the speed of high performance digital circuits increases, the full-wave characteristics of interconnects becomes important. The feasibility of using the finite difference time domain (FDTD) method for interconnect parameter extraction had been demonstrated earlier. The main advantage of the FDTD technique is its ability to model complicated structures and to account for the frequency dependence of the parameters. However, the computational efficiency and memory requirements seriously limit the practicality of FDTD especially for three-dimensional problems.

Recent advances in distributed and parallel computing require one to address the hardware-dependent aspects of these computational issues. The Orion Project takes advantage of the availability of low-cost PC components. Presently, mini-supercomputers can be built at a moderate cost by using fast communication networks. Moreover, the availability of software libraries for distributed computing such as the parallel virtual machine (PVM) and the message passing interface (MPI) have rendered the software development within these environments easier.

Embedded transmission line (ETL) structures have become very commonplace in many high-frequency and high-speed electronic systems; however the analysis and design tools needed for their design are not readily available. The objective of this effort is the modeling, extraction and simulation of three-dimensional complex interconnect structures embedded in multilayer structures. The focus of the work will be on the development of software tools that facilitatae and automate the designerŐs task. The tools will be based on recently developed electromagnetic parameter extraction techniques to determine the electrical parasitic capacitance and inductance coefficients of these structures. These parameters can next be used with efficient simulation algorithms to predict the signal response of ETL structures and provide information about noise immunity and high-speed performance. This will permit the generation of reliable design guidelines as well as a significant reduction in the time to market.

Recent advances in microelectronics have led to considerable reduction in size of components in integrated circuits (ICs). Typical VLSI circuits have dimensions in the submicron range and feature size that can be as low as 0.25 microns. This reduction is a result of several requirements for higher density and shorter interconnection delays. Future state-of-the-art microprocessors will accommodate more than a million transistors in an area as a few hundred squared millimeters.

Along with these trends, several issues related to signal integrity and testing have moved to the forefront. With submicron dimensions, interconnect resistance has become a major bottleneck in circuit performance leading to signal degradation and delays. In addition, measurement and testing in submicron geometries which allows to determine the performance of the structure is a challenging task. Nowadays, the methods employed consist of fabricating special-purpose test vehicles for evaluation, which often require expensive mask processes and complex de-embedding schemes.

This investigation proposes to implement a non-destructive testing methodology for submicron integrated circuits using the recent advances in microelectromechanical systems (MEMS). More specifically, we intend to fabricate and test a micro probe structure that will permit the high-frequency characterization of submicron interconnects and devices in integrated circuits.

The electrical performance of mixed-signal integrated circuits strongly depends on the electromagnetic behavior of the components within the system. Future wireless and personal communication links will be strongly influenced by these considerations. Currently, millimeter-wave monolithic ICs (MMICs) chip sets are under development in the 24-94 GHz range. In recent years, power distribution and parasitic noise control have become critical issues in the design of these MMICs; nowadays, with increased frequencies, interconnect schemes, layout and power distribution have become mainstream design issues. It is now recognized in the CAD community that electromagnetic effects will generally take place at the forefront and will represent the critical limiting factor of MMICs performance. The collaborative effort between Georgia Tech and University of Illinois focuses on developing the technology support for the implementation of low-cost packaging solutions for MMICs. This is to be achieved by harnessing the modeling, simulation, design, fabrication and measurement infrastructure built over the past decade at these two institutions.

ASignal integrity has become a critical area in the design of high-speed communications systems and fast computers. Many research areas have emerged from industry and universities to address issues related to electrical performance. However, the educational infrastructure is seriously lagging. The goal of the National Course in Signal integrity project is to establish a web-based educational platform that will provide the education necessary for aggressive packaging schemes in the area of signal integrity. This is achieved by providing a better understanding of electromagnetics problems and through the use of modeling and simulation tools. With the emergence of visualization tools and on-line simulation packages, access to both qualitative and quantitative answers is immediate.

In addition to the standard components of web-based courses, we have focused our attention on two major components. In the first part of the project, a Movie Creator was implemented in this project; the tool allows the incorporation of taped lectures into a web site with synchronization between the audio and video components.

In the second part of the project, an efficient Perl/Java interface was created that permits the execution of signal integrity modeling and simulation CAD tools in the web server. The study of signal integrity issues is strongly dependent on the ability to simulate and model signal propagation. This task seeks to supply circuit modeling and simulation capabilities using state-of-the-art techniques and analysis tools that were previously developed. A combination of visualization tools and the ability to perform on-line simulations can provide students with immediate access to both qualitative and quantitative answers. Using our newly developed tool, simulation results can be displayed and examined in a web browser shortly after execution. This unique feature of executing CAD software via the web will be a major asset in learning environment.