Langmuir probe studies of a transformer-coupled plasma, aluminum etcher

Spatially resolved positive ion densities (n(i)(+)), electron densities (n( e)), electron temperatures (T-e), plasma potentials (V-p), and floating pot entials (Vi) were measured with a scanning Langmuir probe (PMT FastProbe) i n Cl-2 and BCl3/Cl-2, inductively coupled plasmas (Lam Research Alliance, t ransformer-coupled plasma (TCP) metal etcher with a high-flow chamber). Tim e-resolved ion saturation current was measured during etching of Al/TiN met al stacks. Device damage during the metal stack etching was also studied. P ositive ion densities increase nearly linearly with power for all of the ga ses. The maximum plasma density in the reactor is independent of pressure. The density profiles in the plane of the wafer are peaked above the center of the wafer at low pressure and off center at high pressure. Peaking off c enter is enhanced for smaller height-to-radius ratio chamber configurations , varied by changing the TCP window-wafer chuck gap. The n(i)(+): uniformit y across the wafer depends weakly on power, more strongly on feed gases and radio frequency bias, and most strongly on pressure and the TCP window-waf er gap. Within experimental error, T-e is uniform across the reactor at mos t pressures with a slight fall off beyond the wafer edge. At the lowest pre ssure, T-e dips slightly in the center of the reactor. Addition of 28% BCl3 to a Cl-2 plasma causes a 20% decrease in T-e due to a decrease in the eff ective ionization potential of the gas. A small, grounded aluminum electrod e was inserted into the plasma to eliminate perturbations from the Langmuir probe on the plasma, caused by charging and discharging of the insulating walls of the reactor. Such perturbations make apparent T-e, V-f, and V-p, v alues too high, and at least partly explain why T-e's measured with the Lan gmuir probe were higher than these obtained from optical emission spectrosc opy. (C) 1999 American Vacuum Society. [S0734-2101 (99)01902-8].