The Development of a New Ion Versus Log(Ig) Plot to Characterize Depletion-Mode High Electron Mobility Transistors with the Insertion of a Very Thin Evaporated Al Layer to AlGaN/GaN High Electron Mobility Transistors as an Example

Abstract

Previously, we showed that the Id_on vs. log Id_off plot can be used to characterize silicon-based enhancement-mode MOS transistors for both p-channel and n-channel [1]. AlGaN/GaN high electron mobility transistors (HEMTs) have attracted world-wide interest for microwave power transistor applications. There are two kinds of AlGaN/GaN HEMTs: depletion-mode and enhancement-mode AlGaN/GaN HEMTs. In general, depletion-mode AlGaN/GaN HEMTs are more common. For depletion-mode devices with the source at 0 V, even though the gate voltage may be close to zero, the gate leakage current increases with the increase of the drain voltage and can be significant because of a large drain voltage. Thus the gate leakage current may limit the maximum drain voltage and thus the RF power output of the device. For depletion-mode high electron mobility transistors (HEMTs), it can be easily imagined that both the on current (Ion) and the gate leakage current (Ig) will be increased by using a low work function metal as gate metal. Thus the drain on current (Id_on) and the gate leakage current (Ig) alone will not be a good criterion to judge the quality of depletion-mode HEMTs. In this paper, we would like to propose the application of a new Id_on vs. log Ig plot to characterize depletion-mode HEMTs. In the literature, a lot of various approaches have been proposed to cut down the gate leakage current. Recently, Nanjo et al. proposed to use a thin Al layer inserted between AlGaN and the Schottky metal gate to reduce the gate leakage current [2]. The advantage of this approach was subsequently confirmed by Selvaraj and Egawa [3]; however, they emphasized on the increase of the drain current instead of the reduction in the gate leakage current. It is natural to wonder how Al insertion can ever increase the drain current. Hence, the physics behind this approach is not yet clear. In this paper, we will also propose our theory regarding the physics of Al gate for AlGaN/GaN HEMT’s. With the help of this new plot, the superiority of the sample with Al/Ni/Au gate compared to the control sample with Ni/Au gate can be easily seen because the on current is higher for the same gate leakage current.As shown in Fig. 1, it can be easily shown that the Al insertion significantly improves the device for the same gate leakage current. According to theory, the Schottky barrier height φb is given by φb = φm - χs, where φm and χs are the metal work function and the electron affinity of the semiconductor. It is known that the work function of Al is significantly lower than that of Ni. It can be easily predicted that AlGaN/GaN HEMT with Al gate should have much larger gate leakage current compared to that with Ni gate. In fact, this was experimentally observed by us if the Al was evaporated rapidly during sample deposition. Our explanation of the advantage of Al insertion was that there was residual water vapor present in the vacuum chamber during electron beam evaporation [4]. According to Kordos et al. [5], the electron mobility of the 2-D electron gas at the AlGaN/GaN interface can be improved by a thin layer of Al2O3 on top of AlGaN/GaN. This experimental observation was also reported by Liu et al. [6]. It can be easily imagined that the presence of a thick layer of Al2O3 on top of AlGaN/GaN will degrade the performance of AlGaN/GaN HEMT because the distance between the metal gate and the AlGaN/GaN interface becomes bigger. However, when the thickness of the Al2O3 layer is small enough, the improvement of the electron mobility will be more significant than the degradation due to the increase of the distance between the metal gate and the AlGaN/GaN interface, resulting in the improved performance of the AlGaN/GaN HEMT.