Investigators:

David Rutledge, California Institute of Technology

Zoya Popovic', University of Colorado at Boulder

Robert York, University of California at Santa Barbara

Mark Rodwell, University of California at Santa Barbara

Michael P. DeLisio, University of Hawaii at Manoa

Mehran Matloubian, Hughes Research Laboratories

Huan-Chun Yen, TRW

Quasi-optical power combining promises to significantly improve the power and frequency range of solid-state devices into areas now dominated by vacuum tubes and machined waveguide systems. By combining power from large numbers of devices in free space, we should reduce the loss and complexity associated with traditional feed systems. We envision a new generation of radar and communications transmitters that are smaller, lighter, more reliable, and less expensive than current systems. Quasi-optics is entering an exciting phase. Monolithic millimeter-wave quasi-optical amplifiers were announced at the International Microwave Symposium, and monolithic submillimeter-wave quasi-optical oscillators were announced at the Device Research Conference. At the same time, they have also defined a new set of problems to be solved for practical realization of the technology. These include thermal management and associated device and circuit design issues, high-power and high-efficiency circuit design, improvements in array designs for bandwidth and increased functionality, and extending the technology to higher frequencies.

Our team members and industrial affiliates have each demonstrated state-of-the-art results in either quasi-optical power combining or in the fabrication of high-frequency solid-state devices. They have also successfully collaborated on previous quasi-optical power combining projects. This cooperation is vital, because progress in the quasi-optical field has most often come from an innovative idea from one research group, combined with sophisticated fabrication in another. Our eight research topics are chosen to dramatically improve the state-of-the art in quasi-optical power combining over a range of frequencies from 35 GHz to 140 GHz, and to investigate fundamental issues that apply generally.

94 and 140-GHz cascade-cell pHEMT amplifiers

High-efficiency quasi-optical amplifiers

35-GHz flip-chip HEMT amplifier

FDTD/SPICE modelling with thermal effects

60-GHz folded-slot amplifier

94-GHz transferred-substrate HBT amplifier

Quasi-optical TR module

Quasi-Optical Power Combining

In this program, we propose new concepts, and we propose to demonstrate previous concepts at higher frequencies than before. In many ways, the higher the frequency, the more attractive quasi-optics becomes. At higher frequencies the power limitations of solid-state devices become more acute. Waveguide circuits become more difficult to construct as the sizes shrink. Microstrip substrates suffer from increasing substrate-mode and loss problems at higher frequencies. On the other hand, quasi-optical circuits function like waveguide circuits, except that we need not construct the walls, because symmetry planes provide the proper boundary conditions. The period of a quasi-optical amplifier must shrink with the wavelength, and this usually allows quasi-optical circuits to make more efficient use of a semiconductor wafer.

Monolithic quasi-optical pHEMT amplifier grids have demonstrated gain at frequencies up to 60 GHz. With a cascade-based unit cell, it should be possible to extend this range to 94 and 140 GHz, where there are important atmospheric windows for missile seekers and for imaging. In the past, many quasi-optical concepts have been demonstrated at X-band with hybrid circuits. Recent developments in flip-chip hybrid circuits at frequencies up to 77 GHz suggest that millimeter-wave hybrid quasi-optical amplifiers are now an attractive possibility. Quasi-optical amplifiers based on HBT's have been quite successful at 40 GHz, but low gain had encouraged workers to shift to pHEMT's. Now, new developments in high-speed transferred-substrate HBT's indicate that a 94-GHz HBT grid could be successful. High-efficiency switch-mode amplifiers have been demonstrated at X-band, and future increases in transistor gain should allow us to apply these ideas at millimeter-wave frequencies. Quasi-optical oscillators have been demonstrated at millimeter-wave frequencies, but advances in diode fabrication and stabilizing circuits make submillimeter quasi-optical oscillators possible. The development of new quasi-optical components makes it possible to integrate these to demonstrate a quasi-optical system, the TR module.

The transition from demonstrations of single devices to industrial production and incorporation into radar and communications systems will require sophisticated computer models that include the radiating elements, array edge effects, non-linear device behavior, stability analysis, and thermal effects. We believe that software based on finite-difference-time-domain (FDTD) modeling of electro-magnetic effects, combined with time-domain non-linear circuit simulation (SPICE), is an attractive approach, and that close relations between the groups will allow the software that is developed to be shared and continuously improved.

Quasi-optics is fundamentally interdisciplinary, involving antennas, arrays, device and circuit design, modelling, and millimeter-wave measurements. It takes considerable time to develop the necessary intellectual and capital infrastructure to support such work. For these reasons, our team includes only members have previously demonstrated important results in quasi-optics. These groups are already well equipped, so there are minimal equipment requests in this program, and our team members can make an immediate impact if this program is awarded. Since the primary goal of quasi-optical research is to develop high-power millimeter-wave components, the groups must also have access to technology at the frontier of transistor development. All of our team members have intimate links with leading industries in this area. The industrial affiliates bring the best available device technology to bear on this program, and will also lend expertise in millimeter-wave systems applications.