Plasma generator based on non-self-sustained low-pressure arc discharge has been examined as a tool for deposition of highly-adhesive hydrogenated amorphous diamond-like carbon(DLC) films. Since the discharge is stable in wide range of gas pressures and currents, this plasma source makes possible to realize both plasma-immersion ion implantation(PIII) and plasma-immersion ion deposition(PIID) in a unified vacuum cycle. The plasma parameters were measured as functions of discharge current. Discharge and substrate bias voltage parameters have been determined for the PIII and PIID modes. For PIID it has been demonstrated that hard and well-adherent DLC coating are produced at 200-500 eV energies per deposited carbon atom. The growth rates of DLC films in this case are about 200-300 nm/h. It was also shown that short(∼60$\mu\textrm{s}$) high-voltage(> 1kV) substrate bias pulses are the most favorable for achieving high hardness and good adhesion of DLC, as well as for reducing of residual intrinsic stress are.

In order to predict the moldability of an injection molded part, a simulation of filling is needed. Short shot is one of the most frequent troubles encountered during injection molding process. The adjustment of process conditions is the most economic way to troubleshoot the problematic short shot in cost and time since the mold doesn't need to be modified at all. But it is difficult to adjust the process conditions appropriately in no times since it requires an empirical knowledge of injection molding. In this paper, the intelligent CAE system synergistically combines fuzzy-neural network(FNN) for heuristic knowledge with CAE programs for analytical knowledge. To evaluate the intelligent algorithms, a cellular phone flip has been chosen as a finite element model and filling analyses have been performed with a commercial CAE software. As the results, the intelligent CAE system drastically reduces the troubleshooting time of short shot in comparison with the expert's conventional way which is similar to the golden section search algorithm.

The high Q(quality factor) suspended spiral inductors were fabricated on the silicon substrate by 3D surface micromachined process. The integration of 2.4GHz VCO has been performed by fabricating suspended spiral inductor of the top of CMOS VCO circuit. The phase noise of VCO integrated MEMS inductor is 93.5 dBc/Hz at 100kHz of offset frequency.

Spin etcher can process frontside and backside on the wafer, which is used for etching, stripping, cleaning and wafer reclamation. A new generation of spin etchers has been designed to meet 300mm wafer processing. The larger header and higher spin speed make vibration problem a severe problem in developing equipments. This study shows schematic process of solving practical vibration problems, where it is required to analyze the principal ca uses of vibration problem and find out the method of vibration reduction in spin etcher. The vibration under normal operation is measured in time domain and is analyzed in frequency domain. And modal parameters are obtained through modal test. Using the modal parameters from experiments, the model of finite element method is formulated. From diagnosis using many measurements and analyses, it can be shown that main cause of vibration is unbalance of head.

Sludged silicon powders that are generated during silicon ingot slicing process have potential usage as silicon source in fabricating silicon carbide powders by adding carbon. A thermodynamic calculation is performed to consider a plausible formation condition for the silicon carbide powders. A thin silicon oxide layer around silicon powder is sufficient to supply equilibrium oxygen partial pressure at the formation temperature($1400^{\circ}C$) of the silicon carbide in the Si-C-O ternary system. Formation of silicon carbide by using the sludged silicon powders is more efficient than by using silicon oxide powders.

The fluid flow, mass transfer, heat transfer and film thickness variation during the spin coating process are numerically studied. The model is said to be I-dimensional because radial variations in film thickness, concentration and temperature are ignored. The finite difference method is employed to solve the equations that are simplified using the similarity transformation. In early time, the film thinning is due to the radial convective outflow. However that slows during the first seconds of spinning so the film thinning due to evaporation of solvent becomes sole. The time varing film thickness is analyzed according to the wafer spin speed, the various solvent fraction in the coating liquid, and the various solvent vapor fraction in the bulk of the overlying gas during the spin coating is estimated.