Numerical Study of RTN Amplitude in SiNWFETs

• Main Authors: Su, Chong-Zih, 蘇重齊
• Other Authors: Watanabe, Hiroshi
• Format: Others
• Language: en_US
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/3fyxb6

碩士 === 國立交通大學 === 電信工程研究所 === 107 === 　　In recent years, the miniaturization of transistors has progressed according to the device scaling rule and then the geometry of the most advanced transistor reaches in the level of nano-meters. The device structure of transistor also has substantially changed from the conventional MOSFET. The plain substrate is replaced with silicon FIN to form FINFET. If the device scaling progresses further, the silicon FIN may be replaced with silicon nanowire to form SiNWFET soon. This trend decisively enhances two-dimensional confinement of electrons’ wave in channel, which is perpendicular to the direction of electrons’ transport from source to drain. On the other hand, the charge of a trapped electron in gate dielectric may significantly influence potential profile inside channel under the influence of two-dimensional confinement. If a trap is so close to the channel in SiNWFET, how may it modulate drain current flowing through channel? Or, how may it influence the amplitude of random telegraph noise (RTN) in SiNWFETs? 　　The goal of the present thesis is to numerically investigate how the RTN amplitude varies with the change in the location of trapped charge in SiNWFET. We use a homemade three-dimensional device simulator to simulate a high-K gate SiNWFET comprising a silicon nanowire having 3.2 nm on a side in the cross-section. Note that the mobility enhancement occurs at SOI thickness = 4.5 nm (thicker than 3.2 nm) due to the subband effect [1], which is a one-dimensional confinement of electron’s wave. Let us suppose, although the dimension is different, that two-dimensional subband effect is indispensable in the modeling of this SiNWFET. First of all, we briefly review physical models of RTN amplitude and the density gradient method for modeling the subband effect in the device simulation. Next, we expand the density gradient method to the two dimensional confinement and then perform three-dimensional device simulation of SiNFETs to obtain I-V characteristics and related profiles of potential, carriers, and current densities inside the nanowire. At last, by visually analyzing these profiles together with the simulated I-V characteristics, we can conclude that the RTN amplitude decreases when the carrier density increases in the vicinity of trapped charge along the core of nanowire in the strong inversion region. It is because the wave function is swept out from the boundary inside the nanowire to gather nearby the core of the nanowire. We can also conclude that if a trap locates in the high-K dielectric film above the middle of channel then the RTN amplitude is the maximum in the weak inversion region.