Development and validation of a C5F8/Ar/O2 mixture chemistry model using quantum chemistry methods

Understanding the electron impact processes in c-C5F8 plasmas is of importance to low dielectric constant thin film deposition and etch process development. As much of radicals and excited states in c-C5F8 plasma chemistry are inaccessible by experiment, we used quantum chemistry methods, coupled with a zero-dimensional plasma kinetics model to develop an electron impact cross-section set and its associated plasma chemistry mechanism. The calculations were augmented with quadrupole mass spectrometry and actinometry measurements on a 200mm capacitively coupled plasma source. Predicted etch rates are in good agreement with experimental data examining large substrate RF bias and low pressure. The primary loss process for c-C5F8 is electron impact dissociation into isomers of C5F7 via excitation to the triplet state of c-C5F8. Electron impact dissociation of C5F7 isomers leads finally to the production of C5F5 (an isomer with two conjugate pi bonds) and C5F6 (an isomer with two pi bonds and a folded ring structure). These and other isomers characterized by a plurality of pi bonds and certain ring structures are very stable under electron impact. These ``terminal" species are important from the perspective of polymer deposition. The etch precursor, atomic fluorine, is primarily produced from electron impact dissociation of the feed-gas and its degradation products. CF is produced from dissociation of CF2. CF3 is produced primarily from the walls.