Plasma versus ozone photoresist ashing: Temperature effects on process-induced mobile ion contamination

Abstract

Thin gate oxides used in metal-oxide-semiconductor devices are susceptible to mobile ion contamination introduced at critical patterning levels during photoresist stripping. Trace amounts of heavy, alkali metal ions such as sodium (Na+) and potassium (K+) that are found in even the best photoresists are very difficult to oxidize and do not form volatile products that can be easily removed in a typical O2 plasma. These ionic, heavy metals and their byproducts can be driven into critical underlying layers by self-induced bias potentials and positive ion bombardment within an O2 plasma environment. To examine this potential problem, a series of commercially available photoresist ashing systems were selected to study the influence of process-induced mobile ion contamination on sensitive underlying devices. The systems tested included a ‘‘downstream’’ rf plasma asher, a ‘‘downstream’’ microwave plasma system, and a nonplasma ozone ashing system. In this study, photoresist stripping rates and mobile ion (Ni) contamination levels were measured under various processing conditions for 125-, 150-, and 200-mm wafer substrates. Ashing rate measurements are detailed for the ozone ashing system and activation energies are calculated for ozone only and with the addition of nitrous oxide radical products to the ashing gas stream. Measured mobile ion contamination shows a strong temperature dependence for the ‘‘downstream’’ plasma ashing systems tested while the nonplasma, ozone ashing system showed no temperature dependence to process temperatures exceeding 300 °C and it was equal to the wet-chemical controls.