Optimizated FinFET Channel Using Rapid Hydrogen Thermal Treatment Technology

Due to the rapid scaling down of device size,the traditional device structure encounters many problems,such as the poor reliability of transistors,short-channel effect,low carrier mobility,breakdown of thin gate oxygen,leakage current and other issues.New device structure became an urgent need for the technology development of integrated circuit industry.Among many candidates of processing technologies,FinFET are currently the best choose for device structure below the 14 nm technology node.FinFET technology can provide lower power consumption,higher performance and smaller chip area.The channel made by self-aligned quadruple patterning still has a rough surface,therefore it is a cucial project aimed at achieving a smooth and perfact surface.Hydrogen thermal treatment technology plays a significant role in reducing surface roughness and rounding micron-scale fin corners via the migration of surface silicon atom.Therefore,a deep understanding of the evolutionary process of nanoscale FinFET channel under the rapid hydrogen thermal treatment is crucial for the development of nanoscale FinFET technology.This thesis innovatively adopts a kinetic Monte Carlo method to investigate the volutionary process of FinFET channel in the rapid hydrogen thermal treatment.In the thesis,the evolution of silicon surface structures under hydrogen thermal treatment was studied both theoretically and experimentally.Experimentally,evolution characteristics of silicon and germanium surface structures were studied systematically under hydrogen thermal treatment.Theoretically,the interaction of hydrogen and silicon is analyzed to explain the migration of silicon atoms.For the first time,a novel KMC model is proposed to study the KMC time evolution of a nanoscale FinFET channel at temperatures above 973 K.The optimized FinFET channel can be achieved by controlling the experimental parameters,and the achievements in our work include:1.In the experimental aspects,the assembly design of rapid-heating system can be used in the ultra-high vacuum,high purity hydrogen,argon,nitrogen and other gases or mixed gas,and heating time can be as accurate as one second,which provided perfect technology for the experimental study of evolution of silicon/germanium surface structures at high temperatures.2.The surface roughness of silicon and germanium can be reduced,yielding atomically smooth surfaces,by optimizing the process of hydrogen thermal treatment.Hydrogen and nitrogen play an important role in the migration of silicon atoms at a high temperature,resulting in the formation self-diffused nanostructures by changing the experimental parameters.Also,surface roughness of germanium was investigated experimentally using hydrogen thermal treatment and deionised water.3.The kinetic Monte Carlo method was used to study surface evolution of silicon under hydrogen thermal treatment.The different activation energies were adopted to analyze the influence of hydrogen pressure on the evolution of surface morphology at high temperatures.It is revealed that the higher hydrogen pressure was more favorable for reducing the roughness and achieving a smooth surface.The reduction in surface roughness is divided into two stages,both exhibiting exponential dependence on the equilibrium time.The research work gives an optimized method of controlling the migration rate of silicon atoms and the improvement of surface roughness under hydrogen thermal treatment.4.A novel model of nanoscale FinFET channel is proposed based on surface diffusion theory of silicon fin structures.The evolutionary characteristics of the fin surface morphology,including line edgeroughness(LER),line width roughness(LWR)and the cross-correlationcoefficient p,are investigated in the diffusion process of silicon fin structures at different temperatures and time.The simulation reveals that the LER and LWR of the nanoscale FinFET channel can be reduced effectively,yielding atomically smooth sidewall surfaces,and the optimized FinFET channel is achieved by optimizing the temperature and time in the hydrogen thermal treatment.The simulated results provide scientific and technical guidance for the development of FinFET technology beyond 10-nm technology node.