Avalanche multiplication phenomena in ambipolar TMDC field-effect transistors

Abstract

Two-dimensional (2D) materials are considered as one of the most prominent candidates for next-generation semiconductor technology. Due to their high mobility, tunable band gap and structural uniformity preserved down to single atom thickness, 2D materials are under intensive research in a wide range of fields from transistors, photodetectors, sensors to neuromorphic devices. Taking advantage of their potential in the field of optoelectronics, there has recently been growing number of reports on implementing avalanche multiplication in these materials. Previous studies have mainly focused on optimizing material selection and device architecture, and as a result realized highly sensitive avalanche photodetectors and low subthreshold-swing avalanche transistors. However, most of the investigations are based on the theory for avalanche multiplication in conventional three-dimensional (3D) materials, and fundamental analysis on avalanche phenomena in 2D materials is required to further enhance device performance and establish novel architectures. A fundamental question is the comparison between ionization rates of electrons and holes. In most of the conventional 3D semiconductors, it has been thoroughly examined that the ionization rates are different for these two charge carriers, and their ratio also varies in different materials. However, until now there has been very little research on the comparison of these quantities in 2D materials. In order to study the ionization rates of both charge carriers in a single 2D material, it must be tunable to allow the flow of either charge carrier, i.e., it must be ambipolar. Although ambipolar 2D materials have the advantage of facile carrier type tuning through electrostatic gating, simultaneously allowing both carrier types in a single channel poses an inherent difficulty in analyzing their individual contributions to avalanche multiplication. In ambipolar field-effect transistors (FETs), two different phenomena of ambipolar transport and avalanche multiplication can occur, and both exhibit secondary rise of output current at high lateral voltage. In my thesis study, I proposed the method of channel length modulation to distinguish these two phenomena, and successfully analyzed the properties of electron- and hole-initiated multiplication in ambipolar WSe2 FETs. First, I investigated ambipolar transport in WSe2 FETs, which occurs in devices with longer channel lengths. Ambipolar transport is a phenomenon where electrons and holes flow through the channel simultaneously and each contribute to the total current. It is induced when the drain voltage surpasses the gate voltage, so that the electrostatic gating is locally reversed in part of the channel near the drain electrode. To confirm the occurrence of ambipolar transport in WSe2 FETs, I conducted high voltage sweep measurements in the devices under various gate bias conditions. The critical voltage of the secondary current shifted in accordance with the modulation of gate voltage, which is direct evidence of the ambipolar transport model. I also discuss the working mechanism of the phenomenon through charge carrier distribution and band structure description of the system. Next, I analyzed avalanche multiplication in WSe2 FETs, which occurs in devices with shorter channel lengths. Here, the lateral electric field is much higher than that in long-channel devices, so that impact ionization can begin before the drain voltage reaches the critical voltage for ambipolar transport. The measurements in devices with various channel lengths clearly show the correlation of breakdown voltage with channel length, which yields the intrinsic breakdown field of the material in agreement with the avalanche model. Band structure description is also utilized to discuss the contrasting results from the ambipolar transport characteristics. As the principal motive of the research, I extracted and compared the ionization rates of electrons and holes through the analysis of multiplication factors and breakdown voltages. Ionization rates of hole were found to be higher than electrons, in agreement with the result of lower breakdown voltage of holes than electrons. Finally, I suggest a method of selectively engineering the two competing phenomena of ambipolar transport and avalanche multiplication through careful choice of the channel and dielectric materials. The essential accomplishment of my thesis study was devising a simple and robust method of channel length modulation to separate and thoroughly differentiate between two seemingly alike phenomena of ambipolar transport and avalanche multiplication. With the commercialization of 2D materials approaching and the interest in the fields of both phenomena still expanding, this research will foster the development of high-performance devices and novel architectures utilizing ambipolar transport and avalanche multiplication.