No abstract

Abstract-We have carried out pulsed measurements of the kink effect in InAlAs/InGaAs HEMT's on InP with nanosecond resolution. Our measurements show that the kink's characteristic time constant is strongly dependent on VDS, i.e., it drops by more than three decades, from 100 s down to 50 ns, between low and high values of V DS : This suggests that the kink should not be operational for frequencies in the microwave and millimeter wave regimes. We also find that the kink turn-on dynamics correlate with impact ionization. In particular, the inverse of the kink's characteristic time constant follows a classical impact ionization behavior.

No abstract

No abstract

A b s t r a c tWe present a new model for the dynamic behavior of the kink effect in InAlAs/InGaAs HEMTs. The model suggests that the kink is due to a threshold voltage shift which arises due to hole pileup in the extrinsic source and an ensuing charging of the surface. The model is incorporated in a simple equivalent circuit, which explains well the DC characteristics of the kink, its time evolution in the nanosecond range, as well as its dependence on illumination. I n t r o d u c t i o nThere is continuing interest in the origin and consequences of the kink effect in InAlAs/InGaAs HEMTs. The kink is associated with reduced gain and excess noise at high frequencies. While there is mounting evidence that the kink might be associated with impact ionization and a resulting hole pile-up somewhere in the device, traps are generally invoked to explain the light sensitivity and dynamic behavior of the kink . In this work, we have performed all these measurements as well as light sensitivity measurements, all on the same set of devices. We show that all our experimental results can be explained by an impact-ionization based model where hole pile-up at the source shifts the threshold voltage of the device. This has allowed us to develop the first dynamic physical model and a complete equivalent circuit for the kink in InAlAs/InGaAs HEMTs.[I, 21. M o d e lWe and other workers [3,6] have previously suggested how the kink may arise due to an accumulation of i.i. generated holes in the extrinsic source and an ensuing reduction of the source resistance (Rs). However, Rs reduction is not sufficient to explain the kink effect, especially in high performance devices and close to threshold, where there is little voltage drop on the extrinsic source. Some other effect must be present.The mechanism we suggest is outlined in Fig. 1. An ideal device with a perfect insulating buffer is considered; the device is biased in the saturation regime. To understand the mechanism, we consider how the kink evolves as a function of time if impact ionization is instantaneously "turned on" at t = 0. When impact ionization is first turned on, holes are generated in the high-field drain region of the channel. While some of these holes escape through the gate, many of them flow back through the channel of the intrinsic device into the extrinsic source, where they accumulate in the channel [3,6]. These holes reach quasi-equilibrium within the channel at the source in about a recombination lifetime (Fig. lb).Reaching,epuilibrium with the surface takes a longer time. The bending of the hole quasi-Fermi level in the insulator requires a small hole current to the surface, which slowly changes the surface charge and raises the potential of the channel [6]. This change in channel potential does result in a reduction in source resistance [6], but also results in a change of the intrinsic device's threshold voltage (Fig. 2). To first order, 1(1) kBTI,,{PO +P'where po is the "pre-kink" hole concentration in the region next to the source, and p' is the excess...