(173b) Importance of Simultaneous Reduction of Gas and Surface Mechanisms in Capturing Dominant Kinetic Features

Abstract

Chemical changes in many heterogeneous systems involve a strong coupling of gas and surface reactions. Detailed chemistries are used to account for the changes in both these phases during reactive flow simulations in CFD. Such comprehensive mechanisms usually require large simulation time and memory requirements. Hence, mechanism reduction is performed. Primary reduction techniques are usually applied independently on gas phase and surface mechanisms; often, only the former is done since the number of gas phase species is equal to the number of species conservation equations to be solved. In this work, we show that applying the reduction techniques for gas and surface mechanisms independently can lead to loss of important reaction paths during the reduction procedure, when strong coupling between gas and surface species exist. A reaction path that is redundant when gas phase mechanism is decoupled from the surface mechanism could become dominant on coupling. Hence, the reduction of both these chemistries should be performed simultaneously to have an accurate and consistent reduced mechanism. A case study of GaAs deposition mechanism with 11 gas phase reactions and 25 surface mechanisms [1] will be used to demonstrate this, using a two-step iterative reduction procedure that employs Directed Relation Graph (DRG) [2] followed by Principal Component Analysis (PCA) [3]. While most mechanism reduction techniques focus on idealized reactors (batch, plug flow, etc.), we incorporate CFD simulation error to identify appropriate reduced-order model. Important reaction paths from both decoupled and combined reduction approaches will be studied. The performance of the reduced mechanisms from both the approaches will be used in CFD to predict the deposition rate of GaAs on the wafer. It will be showed that the reduced mechanism from the combined approach matches well with the results from the full model in predicting the deposition rate.

References

1. Mazumder and S. A. Lowry, “The importance of predicting rate-limited growth for modeling MOCVD reactors,” J. Cryst. Growth, vol. 224, pp. 165–174, 2001.
2. Lu and C. K. Law, "A directed relation graph method for mechanism reduction," Proceedings of the Combustion Institute, vol. 30, pp. 1333-41, 2005.
3. Vajda, P. Valko and T. Turanyi, "Principal Component Analysis for Kinetic Models," International Journal of Chemical Kinetics, vol. 17, pp. 55-81, 1985.