Residual stress in the multilayer Si/Dielectric/Pt/PZT/Pt stack was measured as a function of annealing conditions, sol-gel derived PZT (Lead Zirconate Titanate -52/48) thickness, SiO2 and/or Si3N4 dielectric films thickness. Residual stress in the Si3N4 layer varied from -201 to +1275 MPa and from -430 to + 511 MPa in the Si02 layer. Furnace annealing of the bottom Pt film reduced the stress over rapid thermal annealing (RTA). Stress due to PZT films was the controlling factor for the final stress of the stack. Upon increasing PZT thickness, stress became less tensile for Si3N4 dielectric and more tensile for Si02. The deposition of the top Pt on PZT followed by RTA at 300°C in nitrogen had a minimal effect on the final stress of the stack. The average tensile stress for the Si/Si02 /Pt/PZT/Pt and Si/Si3N4/Pt/PZT/Pt stacks was 140 ± 25 and 476±235 MPa respectively.

The etching behaviour of highly boron doped (20 - 25 mω-cm) silicon wafers in aqueous KOH solutions were investigated in the present study. The etch rate of (100) crystal plane and the fastest etch plane projections (31X) and (41X), where X can be 0,1,2…, were measured as a function of processing conditions. The experimental temperatures of solutions were 65 - 75°C and the concentrations 5 - 40 wt-% KOH. The maximum etch rate of (100) crystal plane was measured to locate between 10 - 15 wt-% KOH and the maximum etch rates of fastest etch plane projections were measured to locate between 0 - 10 wt-% KOH in the whole temperature range. In lower concentrations (5 - 20 wt-% KOH) the fastest etch plane projection was determined to be (31X) crystal plane projection. In higher concentrations (25 - 40 wt-% KOH) it was determined to be (41X) crystal plane projection. The results of the present study indicate that there are differences between etch reactions of different etch planes. The difference may be connected to transferring electrons. This could explain why the different etch planes have the position of etch rate maximums as a function of KOH concentration in different places. Furthermore, this could also explain why fastest etch planes are sensitive to experimental conditions.

In high-density fluorinated plasma processes, the mechanisms that fundamentally limit the etching of silicon are poorly understood. In an effort to improve our understanding of limits to the performance of such systems, the etching of silicon wafers in an inductive coupled plasma reactor, using SF6, has been studied. A systematic empirical investigation has allowed us to define many of the experimental parameters that control the etching rate.

There is little temperature dependence of etching which suggests a diffusion limited process. Systematic variation of parameters controlling the rate of etching: total pressure in the reactor, flow rate, partial pressure of reactive species and the rf power supplied to the discharge enable us to accurately define the performance of the system. Experiments, which segregate the physical and chemical components of the etching process, support the conclusion that etching is dominated by electrically neutral species. These various results are interpreted in terms of accepted models for the reactive chemistry in plasmas containing SF6.

The MEMS industry is placing ever greater demands on etching processes, and there is a need to achieve the high degrees of anisotropy, and critical dimension control, at high etch-rates. The approach outlined allows us to develop effective strategies for evolving improved systems for the high rate plasma etching of silicon.

We have investigated the roughness of a silicon (110)-oriented surface after being etched with 20% TMAH (Tetra-methyl Ammonium Hydroxide). We have used an Atomic Force Microscope to determine the roughness exponent α using three different methods: fractal, power spectrum density and scaling analysis. The value of the parameter α was identified to be close to 1/2. This value is different from the KPZ value, which is 0.4 in 2+1 dimension