The reduced base-collector capacitance offers significant advantages for microwave performance of HBT. An enhancement of the maximum oscillation frequency by a factor of 2 to 3 has been predicted [ Kroemer 1982, Fonstad 1984 ] over optimized collector-down structures. Moreover, with a suppression of the extrinsic collector capacitance it becomes possible to implement HBT structures with coherent effects in the base [ Grinberg and Luryi (1993), Luryi, Grinberg, and Gorfinkel (1993) ] resulting in a power gain above the conventional cutoff frequencies. The coherent transistor can be designed to have a high-gain peak at any desired frequency, provided the effect is not destroyed by the parasitics. For frequencies below 100 GHz it is possible to use conventional structures where typically the extrinsic (parasitic) collector capacitance is about twice the intrinsic (useful) capacitance. In order to push the peak into a sub-millimeter wave range it is essential to reduce the parasitic capacitance below the value associated with the useful capacitance. The AP process opens a way to substantially reduce the extrinsic capacitance.
An interesting further advantage of an AP HBT is the possibility of accommodating a Schottky collector. An important parasitic resistance in small area devices is due to the metal semiconductor junction. Since the resistance of an ohmic contact scales with its area, at small dimensions it dominates other resistances that scale with the contact perimeter. It therefore makes sense to dispense with the n+ doped semiconductor layer in the collector and terminate the base-collector field region directly by a metal layer. Such an approach has been successfully used in the fabrication of submicron resonant tunneling diodes [ Allen et al. 1993 ]. The high thermal conductivity of a metallic layer and its proximity to the n- base-collector field region, where most of the heat is generated, is another important advantage of a Schottky collector. Implementation of a Schottky collector is fully compatible with the AP process.
Based on the AP HBT technology it is entirely feasible to implement local oscillators and amplifiers that operate at millimeter and even submillimeter wavelengths. One obvious application of such devices would be for satellite communication systems in the atmospheric transmission window of 345 GHz. Another extremely attractive application [ Luryi (1994a) ] is the possibility of fabricating millimeter-wave phased arrays on a silicon chip. A linear array of 20 elements radiating at 300 GHz would be about a centimeter long. The advantage of having transistor oscillators is that the millimeter-wave beam can be electrically steered off broadside by controlling the relative amplitudes of different oscillators, while their relative phases are locked together by the evanescent wave interaction. The point is that most available phase shifters used in centimeter wave phased array systems are bulky elements that cannot be used in on-chip designs. Instead, one can accomplish beam steering by controlling the amplitude of constant-phase array elements [ Costas (1981) ]. As far as we are aware, this idea has not been employed in practical phased-array antenna systems, perhaps because at centimeter wavelengths it is more efficient to control the relative phases of array elements. In the millimeter and submillimeter wavelength range amplitude steering appears to be the only realistic way to implement purely electronic beam steering. Three-terminal devices are ideally suited for this purpose. On-chip focal plane antenna arrays should have important applications as steerable radar systems in avionics, automated manufacturing, and especially in automobile collision avoidance and early warning systems.
Return to the first section [ Active Packaging - Philosophy ]
Return to the second section [ Active Packaging - Introduction ]
Return to the third (preceding) section [ Active Packaging - HBT Process ]
Go to the next section: [ Active Packaging - Conclusion ]
Look at the [ List of References ]
Go to Serge Luryi's [Research Topics] Home Page
Go to [Serge Luryi's] Home Page