Time-dependent etching of GaAs and InP with CCl4 or HCl plasmas: Electrode material and oxidant addition effects

Abstract

In the dry etching of microelectronic devices, a common side effect is the deposition of material which can inhibit etching and catastrophically contaminate devices. However, simultaneous deposition and etching may be necessary to achieve a desired selectivity and anisotropy. Carbon tetrachloride is one plasma reagent which has been used to etch Al, Si, GaAs, and InP and has shown a propensity to deposit polychlorocarbons. Two variables known to affect both plasma etching and deposition rates are electrode material and oxidant concentration. To understand these effects, we utilize optical emission spectroscopy to measure instantaneous CCl₄ plasma etching rates of GaAs and InP as a function of O₂ or Cl₂ concentration with graphite or stainless steel electrodes. For comparison, HCl plasma etching of InP is also examined. Post‐etch surface analysis by Auger electron spectroscopy in conjunction with time‐resolved, etching‐rate measurements enable a separate assessment to be made of the effects of gas‐phase composition and deposition on etching rates. When stainless steel electrodes are used, oxidant addition to the CCl₄ feedstock liberates chlorine and enhances the initial etching rate but does not reduce etch inhibition by plasma‐generated deposits. The use of graphite electrodes reduces deposition but does not eliminate it. If an HCl plasma is used with stainless steel electrodes to etch InP, deposition and etch inhibition by iron and nickel chlorides occur. In general, deposition can be minimized by using graphite electrodes and HCl gas. However, faster etching rates can be achieved by using graphite and CCl₄. In this case oxidant addition should not only enhance etching rates but also reduce polychlorocarbon formation. The use of stainless steel electrodes with chlorine containing plasmas should generally be avoided.