A study was performed to determine some of the causes of the edge-to-center "bullseye" clearing pattern, in which the etch rate decreases monotonically from the wafer periphery to its center, observed when certain films are etched in a parallel-plate reactive ion etching system. It was found that gradients in local reactant concentrations, with a higher number density present near the edge of the substrate than over the center, are largely responsible for the observed nonuniformity. The concentration gradients appear to come about as a consequence of two reactant generation phenomena and one loss mechanism. The reactant generation phenomena are: (i) the "hot spot," which the edge of the wafer may represent to the plasma, causing enhanced reactant generation there, and (ii) differences in capacitive coupling from the RF generator to the plasma between the wafer's location on the cathode and the rest of the cathode. Stronger coupling surrounding the wafer leads to increased production of reactant around the wafer compared to the area directly above it. The reactant loss mechanism is controlled by the relative etch rates of the wafer vs. the cathode material, the using cathodes which etch at an appreciable rate in the plasma was found to promote improved uniformity. Ion bombardment was found to be uniform across the surface of the wafer, and, consequently, etch processes which are strongly bombardment dependent were found to be uniform.Nonuniform clearing of films etched in plasma and reactive ion etching systems is a major problem in current integrated circuit patterning technology. In particular, when polysilicon, aluminum, or silicon nitride, as well as many other films, are etched in a parallel-plate system, they often display an edge-to-center "bullseye" clearing pattern in which the etch rate decreases monotonically from the periphery of the wafer to its center. This clearing pattern causes problems in selectivity and dimensional control. The present work was done to determine the relative importance of several possible causes of the bullseye effect.Five possible causes were investigated: (i) local reactant concentration gradients near the wafer, (ii) temperature effects (possible thermal gradients across the wafer and their effect on etch rate), (iii) a possible gradient in the degree of ion bombardment across the wafer, (iv) the effects of the abrupt step represented by the edge of the wafer on the local discharge intensity and chemistry, and (v) the effect of the series capacitance represented by the wafer on the coupling of the RF to the plasma. ApparatusA plasma reactor, modified to give field-assisted etching in a pseudo-parallel-plate configuration (Fig. 1), was used in all of the experiments. The modifications are as follows: an aluminum plate ("dark shield" in Fig. 1), whose length (28.1 cm) is equal to the length of the cylindrical reactor, was welded to the inner wall of the chamber, effectively bisecting it. The discharge is confined to the upper half of the reactor. The chamber diameter (and dark ...

Polymer films or residues formed during aluminum etching in carbon te~rachloride plasmas were investigated using Auger, x-ray photoelectron, and Fourier transform infrared spectroscopy. The Amultaneous etching and deposition process occurring in the CC14 plasma resulted in a chlorocarbon-based film containing aluminum. Due to residual gases in the vacuum chamber and to post-deposition air exposure, the films also contained hydrogen and oxygen. The resulting organic/inorganic film matrix generates insight into previously observed behavior of these polymer residues.ABSTRACT InP and GaAs etching in chlorine plasmas at 0.3 Torr follows an Arrhenius dependence on substrate temperature. Apparent activation energies, Ea, of 34.5 -2.8 and 10.5 + 0.7 kcaYmol, respectively, were determined from both optical emission of product species, and step height or weight change measurements. For InP, Ea equals the heat of vaporization of InClose, and the absolute etch rate (7/~m/min at 250~ is in reasonable agreement with the predicted vaporization rate of InC13. Hence, volatilization of InCl~s) is most likely the rate-controlling step for etching InP. Sputter Auger analysis shows that etching in a chlorine plasma leaves multiple layer coverage of InC13 on InP (removable by washing with deionized water), and submonolayer levels of chlorine on GaAs. Both surfaces are rich in the group III element. The etched surface morphologies of InP and GaAs are strongly dependent on temperature, exhibiting a rough-to-smooth texture transition above -250 ~ and -120~ respectively.Plasma etching of III-V compound semiconductor materials has potential application for device processing. Several have already been demonstrated, includ-

No abstract

A process to planarize thick (>4000 A) polysilicon deposited into high aspect ratio openings using spin-on glass (SOG) as the planarizing material is described. The process uses NF3 + CHF3 + O2 in a commercially available batch reactive ion etching system, and has been optimized to etch polysilicon and SOG at the same rate. The mechanism of the etch is explored through the use of a physical model.