Abstract
This paper reviews the state of emerging transistor technologies capable of terahertz
amplification, as well as the state of transistor modeling as required in terahertz electronic
circuit research. Commercial terahertz radar sensors of today are being built using bulky and
expensive technologies such as Schottky diode detectors and lasers, as well as using some emerging
detection methods. Meanwhile, a considerable amount of research effort has recently been invested
in process development and modeling of transistor technologies capable of amplifying in the
terahertz band. Indium phosphide (InP) transistors have been able to reach maximum oscillation
frequency (fmax) values of over 1 THz for around a decade already, while silicon-germanium bipolar
complementary metal-oxide semiconductor (BiCMOS) compatible heterojunction bipolar transistors have
only recently crossed the fmax = 0.7 THz mark. While it seems that the InP technology could be the
ultimate terahertz technology, according to the fmax and related metrics, the BiCMOS technology has
the added advantage of lower cost and supporting a wider set of integrated component types. BiCMOS
can thus be seen as an enabling factor for re-engineering of complete terahertz radar systems, for
the first time fabricated as miniaturized monolithic integrated circuits. Rapid commercial
deployment of monolithic terahertz radar chips, furthermore, depends on the accuracy of transistor
modeling at these frequencies. Considerations such as fabrication and modeling of passives and
antennas, as well as packaging of complete systems, are closely related to the two main
contributions of this paper and are also reviewed here. Finally, this paper probes active terahertz
circuits that have already been reported and that have the potential to be deployed in a
re-engineered terahertz radar sensor system
and attempts to predict future directions in re-engineering of monolithic radar sensors.