Analysis of a radiofrequency plasma reactor for etching

Microelectronics are a core component of modern technology. Production of microelectronics relies on plasma processing for etching fine features in semiconductors. While the plasma etch process is relatively mature, as microelectronic structures become smaller, plasma properties that lead to non-uniformities in the etch must be explored and addressed. To this end, we couple experimental and modeling efforts to characterize a radiofrequency argon plasma reactor operating in the 10 MHz regime with input power on the order of 100 W and 14 cm diameter electrodes. Experimental plasma diagnostics include Langmuir probe measurements, laser-induced fluorescence, optical tomography, and a retarding field analyzer to measure the ion energy distribution function (IEDF). The modeling effort consists of a one-dimensional PIC/DSMC model that mimics the experimental configuration. Cross sections for electron-neutral collisions include elastic, excitation, and ionization, as well as elastic and charge exchange collisions for ions. Furthermore, we analyze the IEDF reaching the electrode surface as a function of pressure and input power. The outlook for this work is to train and use machine learning models using measured and simulated data to predict optimized plasma parameters for semiconductor etch and other relevant scenarios.