In most PIC/MCC simulations of radio frequency capacitively coupled plasmas (CCPs) several simplifications are made: (i) fast neutrals are not traced, (ii) heavy particle induced excitation and ionization are neglected, (iii) secondary electron emission from boundary surfaces due to neutral particle impact is not taken into account, and (iv) the secondary electron emission coefficient is assumed to be constant, i.e. independent of the incident particle energy and the surface conditions. Here we question the validity of these simplifications under conditions typical for plasma processing applications. We study the effects of including fast neutrals and using realistic energy-dependent secondary electron emission coefficients for ions and fast neutrals in simulations of CCPs operated in argon at 13.56 MHz and at neutral gas pressures between 3 Pa and 100 Pa. We find a strong increase of the plasma density and the ion flux to the electrodes under most conditions, if these processes are included realistically in the simulation. The sheath widths are found to be significantly smaller and the simulation is found to diverge at high pressures for high voltage amplitudes in qualitative agreement with experimental findings. By switching individual processes on and off in the simulation we identify their individual effects on the ionization dynamics and plasma parameters. We conclude that fast neutrals and energy-dependent secondary electron emission coefficients must be included in simulations of CCPs in order to yield realistic results.
Download