Current‐gain in FETs beyond cut‐off frequency

Abstract

Driven by the fast‐growing demand for high‐frequency applications, electronics engineers are steadily pushing transistor technologies beyond their high‐frequency limits. A key figure of merit to quickly assess the potential of a transistor for high‐frequency applications is given by the cut‐off frequency (fT) that is defined as the frequency at which the short‐circuit current‐gain (h21) drops to unity. However, field‐effect transistors (FETs) can exhibit a current‐gain even at frequencies beyond fT, due to the resonance of the extrinsic reactive contributions leading to a peak in the magnitude of h21. Based on an extensive analysis, this letter aims to demonstrate that, although not all FETs exhibit such a current‐gain peak at frequencies beyond fT, this effect is inherent in any FET devices and its appearance simply depends on the specific amount of the extrinsic reactive contributions, besides to the operating frequency.

Multibias and temperature dependence of the current‐gain peak in GaN HEMT

“…However, it is quite often noticed that at high frequencies, h 21 deviates from its optimal behavior, owing to the occurrence of the current‐gain peak (CGP). The earlier studies have demonstrated that the origin of the CGP is because of the resonance of extrinsic inductances and the intrinsic capacitances . It has been shown that such a resonance leads to a dip in the open‐circuit output impedance magnitude ( Z 22 ), thereby turning into the peak in h 21 .…”

“…As a matter of fact, the determination of the resonant frequency has enabled accomplishment of the challenging task of extracting the intrinsic output capacitance, whose contribution might be rather difficult to distinguish from the extrinsic output capacitance without using optimization algorithms or additional dummy test structures. Furthermore, circuit designers can take advantage from the presence of the peak in h 21 as this phenomenon implies achievement of an increase in the current gain locally around the resonant frequency . The peak in h 21 has been investigated also by using a pole‐zero approach, which has provided simplified mathematical expressions to calculate the location of the peak .…”

“…The CGP has been investigated by means of the second derivative of the magnitude of h 21 against the frequency, enabling the determination of a set of parameters to fully and systematically characterize the CGP . Over the years, the peak in h 21 has been extensively studied under a wide range of working conditions, such as different bias points, device sizes, and high temperatures . To further advance the understanding of this phenomenon, the present study investigates for the first time the peak in h 21 for not only heated but also cooled device.…”

“…The earlier studies have demonstrated that the origin of the CGP is because of the resonance of extrinsic inductances and the intrinsic capacitances. [4][5][6][7][8][9][10] It has been shown that such a resonance leads to a dip in the open-circuit output impedance magnitude (Z 22 ), thereby turning into the peak in h 21 . Therefore, f T should be estimated by extrapolation of the low-frequency h 21 to unity with a − 20 dB/decade slope.…”

“…Furthermore, circuit designers can take advantage from the presence of the peak in h 21 as this phenomenon implies achievement of an increase in the current gain locally around the resonant frequency. 8 The peak in h 21 has been investigated also by using a pole-zero approach, which has provided simplified mathematical expressions to calculate the location of the peak. 10 The CGP has been investigated by means of the second derivative of the magnitude of h 21 against the frequency, enabling the determination of a set of parameters to fully and systematically characterize the CGP.…”