Several Important Issues On The Low-Frequency Noise Nano-MOSFETs

In nano-MOSFETs, the channel length and gate oxide thickness are very small, whereas the electric fields both in the tranverse and longitudinal directions become very large. As a result, severe quantum mechanical effects and different types of variation phenomena occur. Togther with low frequency noise, those effects, such as random dopant fluctuations, line edge roughness, quantization effect of the inversion layer and depletion effect of the poly-silicon gate, lead to significant fluctuations in both threshold votages and terminal currents. The investigation of the low-frequency noise under the influence of above-mentioned effects in nano-MOSFETs possesses necessitously realistic meanings.In this thesis, on the basis of the existing MOSFETs' low-frequency noise research results, some important noise sources in nano-MOSFETs, such as RTS noise and back-ground 1/f noise in drain and gate currents, were investigated in detail using a comprehensive methods of theoretical modeling, experimental testing and noise signals time-domain analyzing, in which quantum mechanical effects and other process fluctuations were taken into account. And the following results were acquired.1. An analytical model for RTS noise amplitudes including quantization effect, poly-silicon depletion effect and image charge effect was presented. The results of is new model fit the experimental results of RTS noise in nano-MOSFETs better than those existing models.2. A percolation model for RTS noise amplitudes in nano-MOSFETs was built. Based on the numerical solutions of coupled 2D Schrodinger and Poisson equations, this new model can simulate the dependences of RTS noise amplitudesΔID/ID on devices geometries, bias conditons and trap locations. The simulated results agree well with those of the time-consuming numerical schemes.3. An analytical model for the power spetrum density of 1/f noise in gate current was derived for wide-channel nano-MOSFETs. Based on the barrier hight modulation mechanism, this model sucessfully predicted the variation of the power spetrum density of gate current 1/f noise with the bias conditions. Using this model, the distributions of the traps density in gate oxide was extracted, which agrees well with the nitrogen atoms distribution in a typical DPN process.4. An analytical amplitude model based on Schottkey effect for gate current RTS noise in nano-MOSFETs was brought forward. This model can reproduce the dependences of gate current RTS noise amplitudes on bias conditions. Because the relative amplitudes of gate current RTS noise are quite larger than those of the drain currents, a Coulomb attractive trap were characterized using the gate current RTS noise under room temperature for the first time.5. Noise Scattering Pattern (NSP) method was introduced to detect RTS noise components in the time trains of drain currents. Using NSP method, RTS noise amplitudes and the relative magnitudes of time constantsτc andτe can be obtained.6. Through an independent study of the background 1/f noise, it was found that this type of noise originated from the mobility fluctuations induced by the different scattering mechanisms the carriers encountered during transportation. Background 1/f noise and RTS noise have different physical mechanisms. The former abides by the Hooge's mobility flutucation hypothesis, while the latter complies with McWhorter's number fluctuaton assumption.Though some results were achieved in this work, researches on the low-frequency noise in nano-MOSFETs still need to be push forward, there are more and more issues need to be addressed with the down-scaling of Si nano-MOSFETs and the emergence of innovative nano-structures.