This project aims to improve and develop electromagnetic (EM) analysis tools to meet the demand of the computer chip industry. The modeling of EM effects in computer integrated circuits (IC) has in recent years become increasingly important, due to the increased transistor density and switching clock rate of CPU?s in a computer. For this reason, EM effect is a challenge identified by the International Technology Roadmap for Semiconductors (ITRS) as an area of priority. But the exorbitant computational cost and complexity of such EM modeling problems have precluded their precise solution so far. Commercial simulation tools for multi-function broadband ICs, which trade accuracy for efficiency, cannot fully meet the stringent demands for broad bandwidth and complexity of next generation applications. However, fast algorithms for EM simulations have potentials when stabilized with appropriate techniques for broadband applications to capture both circuit physics and wave physics. Fast, efficient, and highly stable algorithms will be developed for seeking broadband, multi-scale solutions of IC problems. The solutions will be integrated with existing Electronics Design and Automation (EDA) tools, and co-design will be studied at chip and package levels. The work will be connected to real-world problems and models will be validated with measurements.
The potential impact of fast and efficient modeling technique for electronic ICs can alter how computers are designed. It will remove bottlenecks caused by EM effects due to increased transistor density and switching clock rates, and allow the accurate virtual prototyping of circuit design over a broad frequency range. It will also encourage IC designers to use more microwave engineering paradigms in future IC designs. Hence, it will expand the design space of IC and circuit designers, increase their repertoire of toolboxes, and enrich new possibilities for future IC designs. Due to the lack of high-quality computer-aided design (CAD) design tools that incorporate EM effects efficiently and accurately, IC designers face bottlenecks due to signal and power integrity issues in 3D IC designs. By precise and efficient electromagnetic modeling, CAD IC characterization will be improved and the design barriers faced by IC designers will be pushed back.
In addition, this project will train students, at various levels ranging from undergraduates to graduates, to be well versed in electromagnetic physics, applied mathematics, computer science, and measurements (with a keen focus on the science of IC simulation and design). Students schooled in this cross-disciplinary field generally adapt easily to adjacent areas of research in academia and industry. Hence, this project will train badly needed human resource in computational science and engineering and adjacent fields for the advancement of high tech. EM physics, valid over a vast length scale, is fundamental to many electrical engineering technologies. Fast, broadband computational electromagnetics (CEM) algorithms for complex structures are applicable to a large variety of other applications including micro- and nano-technologies, ranging from meta-material modeling, to nano-optics and nano-lithography. It can help in the design of super-resolution lithography, and improve the modeling of EM effects in N/MEMS, sensors and actuators, interconnects in computers at the package level, as well as at the board level. It will enable the modeling of small, complex, smart, and reconfigurable antennas, RF integrated circuits, as well as greatly impacting terahertz modeling, biotech, and homeland security technology development.
