Articles you may be interested inModeling of fluorine-based high-density plasma etching of anisotropic silicon trenches with oxygen sidewall passivation J. Appl. Phys. 94, 6311 (2003); 10.1063/1.1621713Kinetics and crystal orientation dependence in high aspect ratio silicon dry etching High density fluorocarbon etching of silicon in an inductively coupled plasma: Mechanism of etching through a thick steady state fluorocarbon layer Surface kinetics of ion-enhanced chlorine plasma etching in the low ion energy regime was studied by utilizing Ar ϩ for ion bombardment and Cl and Cl 2 as reactants. The argon ion and chlorine atom ͑molecular͒ fluxes were controlled independently over more than an order of magnitude and at flux levels within an order of magnitude of that typically used in high density plasma processes. The ion-enhanced etching yield was characterized as a function of Ar ion energy, ion flux, neutral-to-ion flux ratio, and the ion incident angle. Possible reaction pathways are proposed and reduced into a two-parameter model which is useable in a profile simulator. The etching yield increases with the increase of flux ratio but gradually saturates at higher flux ratios as the ion flux limits the etching yield. The ion energy dependence was found to scale linearly with ͑ E ion 1/2 ϪE th 1/2 ͒, where the threshold energy E th is found to be 16 eV. The etching yield of Cl is found to be similar to that of Cl 2 at flux ratios below 10, but 4-5 times higher than that of Cl 2 at higher flux ratios. The sticking coefficient of Cl 2 is similar to that of Cl for smaller neutral-to-ion flux ratios where the ion bombardment produces highly reactive surface silicon sites for both atomic and molecular chlorine. At high neutral-to-ion flux ratios, however, the dissociative adsorption of molecular chlorine is slower than the adsorption of atomic chlorine on the highly chlorinated surfaces. The angular dependence of ion-enhanced etching yield was also measured. The etching yield was reduced by approximately 35% when an ion impingement angle was changed from normal impingement to 60°off-normal.

The effect of low level, between 0.001 and 10%, O2 addition on C12 based plasma etching of polysilicon in a parallelplate etcher was investigated. Three strong effects on the etching rate of unmasked polysilicon were observed. Low level O2 addition, 0.1-2%, enhanced the etching rate up to 4 times that of the pristine system. High level 02 addition, >-6%, stopped the etching altogether. Pure C12 plasma etching rates were enhanced by previous runs with 02 addition; a hysteresis effect. The hysteresis was reduced by subsequent system plasma exposure. Running a CF4 plasma after 02 addition runs cleaned the system and returned it to its pristine state. The stopping of etching at high O2 levels was attributed to plasma oxidation of the polysilicon surface; since C12 plasmas are very selective with respect to oxide, the etching is stopped by the surface oxide. Both the low level 02 etching rate enhancement and the hysteresis effect can be attributed to the deposition of an oxychloride film on the etcher surfaces. The oxychloride film, formed by reaction between the etching products and the added O=, passivates the interior etcher surfaces against C1 surface recombination, and, therefore, reduces the major C1 loss mechanism. The gas-phase C1 concentration increases, which in turn increases the polysilicon etching rate. These results demonstrate the importance of etcher surface condition and of the etching product deposition. Photoresist masked polysilicon films etched in the same configuration showed similar 02 addition dependencies as unmasked polysilicon films. The main difference is that the 02 addition effect thresholds are higher with the photoresist mask due to the consumption of the added 02 by photoresist etching. Unmasked polysilicon film etched in a commercial etcher, Applied Materials Precision 5000, under magnetically enhanced reactive ion etching configuration, exhibited the same trends as in the parallel-plate research etcher. This indicates that results obtained from the parallel-plate research etcher can be generalized to other configurations.

The etching of polysilicon has been characterized as a function of feed gas impurity levels, to determine specifications for gas purity in plasma etching processes. CF4, CF2C12, CF3Br, and C12 feed gases were blended with 10 to 100,000 ppm impurity gases. H2, 02, and N2 impurities were added to simulate air and water contamination; C2H4 was added to simulate hydrocarbon impurities; and CHF3 and CF3C1 were added to test the effect of feed gas purification. Typically, changes in discharge properties (ion bombardment flux, discharge electrical impedance, and plasma-induced emission) correlated well with changes in the etching rate. In all but one case, the impurity levels below 1000 ppm had no apparent effect on discharge properties. In C12 discharges, H2 + 02 or 02 addition was found to alter the etching characteristics such that after impurity addition ceased, approximately lh was needed for the system to return to the initial etching rate of pure C12. This hysteresis is associated with the presence of photoresist and has been tentatively attributed to polymeric deposition on the chamber surfaces.

Radio-frequency (RF) plasmas are widely used for materials processing in the manufacturing of integrated circuits, e.g., plasma etching and plasma-enhanced deposi-