Toshiyuki Yokosuka

The development of computational chemistry approach to predict the viscosity of lubricants

Sophisticated machinery requires the development of new lubricants that realize friction-less contact, perfect transmittance of torque, and so on. However, experimental approaches have needed long time to find new lubricants. Therefore, the prediction of lubricant properties by computational chemistry should be crucial to accelerate the design of lubricants. This study treats precise evaluation of viscosity employing molecular dynamics simulation and prediction based on neural network.

Quantum Chemical Molecular Dynamics Studies on the Chemical Mechanical Polishing Process of Cu Surface

The dynamic behavior of the oxidation reaction of the Cu surface during the Cu chemical-mechanical polishing (CMP) process was investigated by a novel tight-binding quantum chemical molecular dynamics method. We confirmed that our tight-binding quantum chemical molecular dynamics method with first-principles parameterization can calculate the structures, and electronic states of various molecules and solids related to the Cu-CMP process as accurately as the density functional calculations, while the CPU time of the new method is around 5,000 times faster than that of the first-principles molecular dynamics calculations. We employed hydrogen peroxide solution as a slurry and the Cu surface as a substrate to simulate the Cu-CMP process by using our accelerated quantum chemical molecular dynamics method. Three types of models were constructed to analyze the effect of the pH of the slurry and Miller plane of the Cu surface on the dynamic behaviors of the oxidation process of the Cu surface. We indicate that the pH of the slurry strongly affects the oxidation process of Cu surface. Moreover, we clarified that the oxidation mechanism depends on the Miller plane of the Cu surfaces.

Computational methods for the design of zeolitic materials

Recently we developed an accelerated quantum chemical molecular dynamics program “colors”, which is approximately 5000 times faster than the first principle quantum chemical molecular dynamics calculation. In the present study we have applied our Colors program to the investigation of the structure and dynamics of big periodic models for crystobalite (96 atoms) and H-mordenite (145 atoms). Our results show clearly the proton migration in-between the two oxygen atoms named O16 and O18 in H-mordenite model which are more energetically preferable than other two oxygen atoms named O41 and O19 surrounding the A1 atom. The change of the proton charge during the simulation has been observed.

A Theoretical Study on the Realistic Low Concentration Doping in Silicon Semiconductors by Accelerated Quantum Chemical Molecular Dynamics Method

We present a theoretical study of the structural and electronic properties of a realistic low concentration (<0.5%) doping model in silicon semiconductor. The density of states was calculated using our newly developed accelerated quantum chemical molecular dynamics method, based on our original tight-binding theory. Using this approach, the band structures of large-size silicon model including n-type and p-type dopants were successfully simulated. The results are in good agreement with the experimental results. Furthermore, we also performed quantum chemical molecular dynamics simulation of the dopants in silicon and observed the change of the dopant levels during the simulation. These results clearly suggest that our original quantum chemical molecular dynamics program is a very effective tool for not only the band structure of a realistically low concentration dopant model but also for the electronic states dynamics of silicon semiconductors.

Ab Initio Calculation of F Atom Desorption in Tungsten Chemical Vapor Deposition Process Using WF6 and H2

Chemical vapor deposition (CVD) of tungsten is widely used in semiconductor devices. In this process, although many atoms are expected to adsorb on a surface during the deposition, the behavior of these atoms has not been understood well. In this study, we have first investigated the reduction step of F adsorbates by H adsorbates on a W surface by density functional theory. The calculated reaction heat (36.5 kcal/mol) is in good agreement with the experimentally estimated value.

23.1: Invited Paper: Electrical and Mechanical Properties of MgO Protecting Layer

This paper focuses on the secondary electron emission coefficient (γ) and the degradation dynamics of the MgO protecting layer. It is demonstrated that the γ could be improved by 50 % at the maximum using an ion implantation technique. We discuss that the γ improvement is not based on the same principle as reported in the thermally equilibrium compound. As for the degradation dynamic, we point out that the desorption of MgO clusters from the MgO surface is enhanced in the surface normal direction by applying a bias voltage to the substrate. From a computer simulation, it is confirmed for the first time that the desorption of MgO constituent, in other words, the surface degradation, is certainly occurring under the electric field condition.