Etching Evolution Simulation And Key Parameters Optimization

Itincreasingly requestshigherstandardsof the production process as the rapid development of integrated circuit technology. The etching process isan important part of the integrated circuit production and the process parameters has a great impacton the etching result. The parameters obtained by the conventional method of experiment design has low efficiency and high cost. Through etching evolution simulation and analyzingthe influence on the etching result by different process parameters, we can save a lot of manpower and financial resources. Studying several different algorithms for plasma etching process,we designed the two-dimensional and three-dimensional etching evolution simulation tools, and the results obtainedwerecompared with the corresponding experiment results, verified the effectiveness of the method. Because it is difficult to solvethe related parametersof the simulation process, an optimization methodbased oncomparingthedifferences between the experiment and simulation result is introduced to optimize these parameters, makes the simulation result more accurate. This thesis mainly focus on the following perspectives:1) A new two-dimensional etching evolution method based on a combination of string algorithm and molecular dynamics is proposed to achieve a cross-scale etching evolution simulation. Molecular dynamics method calculating the effect reaches the actual particle etching effect on the surface of the model, is translated into local etch rate return to the string algorithm to update the cross-sectional shape on the next moment. This approach simplifies the relevant parameters, expands the scope of simulation method to achieve a two-dimensional surface etching evolution simulation.2) A three-dimensional etching evolution simulation method based on cellular automata is improved. Through the model compression,the model representionisoptimizated, the process of evolution needthe relatively lowwermemory space; at the same time etching yield is equivalent to particle etching effect on the surface of base material, reduce the difficulty of the simulation implementation. Studying the effects on the etching results under different experimental conditions, verify the effectiveness of the three-dimensional etching evolutionalgorithm.3) The experimental data is proposed to optimize the etching yield parameters. By comparing the actual etching profile and simulation profiles, setting the fitness function translate etch yield parameters solving probleminto an optimization problem. Ordinal optimization method and tabu search algorithm are combined to solve this problem, obtain the simulation result and experimental etching profile mainly similar.