Seiji Onoue

Topography Simulation of Reactive Ion Etching Combined with Plasma Simulation, Sheath Model, and Surface Reaction Model

We developed an oxide-film reactive ion etching (RIE) topography simulation which consists of a plasma simulation, a sheath model, and a surface reaction model. In the plasma simulation, the plasma parameters were calculated in two dimensions using the particle-in-cell/Monte Carlo collision (PIC/MCC) method. In the surface reaction model, the motion of particles in the etching trench was simulated by the Monte Carlo method. This topography simulation was applied to the etching by a capacitively coupled plasma (CCP). Etching conditions were as follows:gas pressure 5.3–10.6 Pa and RF power 1.1–1.7 kW in the gas mixture of C4F8, CO, O2, and Ar. The calibration method for such simulation parameters as the ion reflection ratio, the etch rate and the polymer etch rate was established based on the experimental results. As a result, the change of the etching profile was reproduced according to the change of the gas pressure and RF power with high accuracy. Furthermore, it was shown that the simulator can predict the profile change corresponding to the process change.

Comparison of CF4 and C4F8 gas etching profiles by multiscale simulation

The CF4 and C4F8 gas etching profiles of oxide films were compared by multiscale simulation that comprises gas reaction, sheath, and surface reaction models. The densities of CF3, CF2, and CF radicals in CF4/Ar or C4F8/Ar gas were measured and compared with those obtained by simulation using the gas reaction model. From this comparison, the electron temperatures were determined to be 2.8 and 4.5 eV for CF4 and C4F8 gases, respectively. In the sheath model, the behavior of ions in a sheath was simulated for sheath lengths calculated from these electron temperatures. In the surface reaction model, we simulated the formation of a polymer and active layers by CF2 radicals, determined the depth of etching resulting from ion bombardment, and obtained the etching profiles of the oxide films. The profile of a contact hole with a depth of 820 nm and an aperture diameter of 160 nm was simulated. The results showed that the photoresist height was approximately 70 nm greater and the bowing diameter was approximately 10 nm smaller in the case of using C4F8 gas than in the case of using CF4 gas. This is because the CF2 density in the C4F8 gas is approximately 30 times higher than that in the CF4 gas and the polymer layer more strongly protects the underlying film. When the etching profiles were simulated with a fixed density of positive ions but with various CF2 density, the bottom diameter was constant but the bowing diameter changed for CF2 densities between 1013 and 1014 cm−3.

Behavior of incorporated nitrogen in plasma-nitrided silicon oxide formed by chemical vapor deposition

The behavior of nitrogen (N) atoms in plasma-nitrided silicon oxide (SiO2) formed by chemical vapor deposition (CVD) was characterized by physical analysis and from electrical properties. The changes in the chemical bonding and distribution of N in plasma-nitrided SiO2 were investigated for different subsequent processes. N–Si3, N–Si2O, and N2 are formed in a SiO2 film by plasma nitridation. N2 molecules diffuse out during annealing at temperatures higher than 900 °C. NH species are generated from N2 molecules and H in the SiO2 film with subsequent oxide deposition using O3 as an oxidant. The capacitance–voltage (C–V) curves of metal–oxide–semiconductor (MOS) capacitors are obtained. The negative shift of the C–V curve is caused by the increase in the density of positive fix charge traps in CVD-SiO2 induced by plasma nitridation. The C–V curve of plasma-nitrided SiO2 subjected to annealing shifts to the positive direction and that subjected to the subsequent oxide deposition shifts markedly to the negative direction. It is clarified that the density of positive charge fixed traps in plasma-nitrided SiO2 films decrease because the amount of N2 molecules is decreased by annealing, and that the density of traps increases because NH species are generated and move to the interface between SiO2 and the Si substrate with the subsequent oxide deposition.

Control of design specific variation in etch-assisted via pattern transfer by means of full-chip simulation

A novel model-based algorithm provides a capability to control full-chip design specific variation in pattern transfer caused by via/contact etch processes. This physics based algorithm is capable to detect and report etch hotspots based on the fab defined thresholds of acceptable variations in critical dimension (CD) of etched shapes for a prospective dry etch process step. It can be used also as a tool for etch process optimization to capture the impact of a variety of patterns presented in a particular design. A realistic set of process parameters employed by the developed model allows using this novel via-contact etch (VCE) EDA tool for the design aware process optimization in addition to the ¿standard¿ process aware design optimization.

Measurements and Simulations of Particles in a Plasma Chemical Vapor Deposition Chamber

In a plasma chemical vapor deposition (CVD) chamber, particles generated during the deposition process adhere to the wafer after the process ends. Since it is important to remove particles efficiently in the discharge step after the deposition process, we determined the optimal conditions by using particle motion simulation. The electrostatic force, ion drag force, and gas drag force produced by neutrals, which particles receive in plasma, are calculated from plasma and gas flow simulations. The particles vibrate in response to the downward ion drag force and the upward electrostatic force, while moving to the wafer edge due to the gas flow. In the discharge step, the baseline condition and another condition (the side power condition) where the electric field is perpendicular to the wafer were compared using 50 particles set at random positions. The number of remaining particles was 15.0 under the baseline condition and 2.8 under the side power condition. In the experimental results corresponding to both conditions, the number was found to be reduced to 23.3% by adjusting the conditions. The simulation results are in good agreement with the experimental results.