Effects of cross field diffusion in a low pressure high density oxygen/silane plasma

Abstract

A low pressure high density oxygen/silane radio frequency (13.56 MHz) plasma coupled in a helicon reactor used for silicon dioxide deposition is characterized by using an energy selective mass spectrometer situated at the wall of the processing chamber: measurements of positive and negative ion energy distribution functions and mass spectra (1⩽amu⩽150) are obtained for various flow-rate ratios (R=O2/SiH4=1 to 10 but constant total flow rate of 30 sccm), and for a constant radio frequency power and magnetic field of 800 W and 70 G, respectively. Plasma potentials between 35 (R=10) and 60 V (R=1) are measured depending on the silane and oxygen flows showing charging of the silica-covered diffusion chamber wall. The magnitude of the wall charging most likely depends on the effective capacitance formed by the silica layer (which results from months of deposition) and on the inbalance between the positively and negatively charged particles which impinge onto the sidewalls at the discharge initiation until equilibrium of fluxes is reached. This inbalance is a result of the magnetic field configuration generated by the four coils surrounding the reactor and of the subsequent cross-field diffusion of the positively and negatively charged particles to the sidewalls. Maximum wall charging is observed when the silane flow is maximum (R<2), a situation where polymerization is observed in the mass spectrum of the positive ions, and where a minimum density of negative ions (O−, OH−, and H−) are detected close to the walls. Although the polymerization appeared to be a primary candidate for the increase of the plasma potential to ∼60 V, its presence does not significantly change the total positive ion density at equilibrium but is accompanied by a dramatic decrease (two orders of magnitude for the O− ions) in the negative ion density close to the walls when R is increased from 1 to 10. The change in the degree of electronegativity close to the walls affects the global plasma equilibrium and appears as an indirect factor in the magnitude of the wall charging.