Thermal oxidation of silicon is an important process step in MEMS device fabrication. Thicker oxide layers are often used as structural components and can take days or weeks to grow, causing high gas costs, maintenance issues, and a process bottleneck.
Pyrolytic steam, which is generated from hydrogen and oxygen combustion, was the default process, but has serious drawbacks: cost, safety, particles, permitting, reduced growth rate, rapid hydrogen consumption, component breakdown and limited steam flow rates.
Results from data collected over a 24 month period by a MEMS manufacturer supports replacement of pyrolytic torches with RASIRC Steamer technology to reduce process cycle time and enable expansion previously limited by local hydrogen permitting. Data was gathered to determine whether Steamers can meet or exceed pyrolytic torch performance.
The RASIRC Steamer uses de-ionized water as its steam source, eliminating dependence on hydrogen and oxygen. A non-porous hydrophilic membrane selectively allows water vapor to pass. All other molecules are greatly restricted, so contaminants in water such as dissolved gases, ions, total organic compounds (TOC), particles, and metals can be removed in the steam phase.
The MEMS manufacturer improved growth rate by 7% over the growth range from 1 mu m to 3.5 mu m. Over a four month period, wafer uniformity, refractive index, wafer stress, and etch rate were tracked with no significant difference found.
The elimination of hydrogen generated a four-month return on investment (ROI). Mean time between failure (MTBF) was increased from 3 weeks to 32 weeks based on three Steamers operating over eight months.