Abstract
This study presents a detailed methodology to verify and correct the chemical consistency of any reaction model based on collision limit violations (CLV) analyses, thermochemical data calculations, and reaction rates reffiting. As a study case, an a priori well validated reaction model representative of silane chemistry is selected to highlight the benefit of this methodology. In the final mechanism, 30 new thermochemical data obtained using a high theory composite method (G4//B3LYP/6-311++G(3df,2p)) are corrected, as well as 13 reaction rates initially above the collision limits. In addition to the suppression of unphysically high reaction rates, the collision limit violation analyses allow to identify and correct important inconsistencies of the initial reaction model such as species mislabeling or discontinuities in some thermodynamic data, which are not easily detectable without an extensive review of each submodels. The novelties of the present work include (i) the discussion of the CLV uncertainty, (ii) the use of this uncertainty to identify CLVs, and (iii) the establishment of the relationship between high CLV values and species mislabeling. As a results of this study case on a specific mechanism, this work also provides a more consistent reaction model for silane chemistry, relevant for both high temperature combustion and flame relevant conditions, as well as a large number of ab initio calculation results, relevant to the SixHyOz chemical system. Based on the present analyses, the predictions of the final reaction model are mainly improved for silane pyrolysis, while the benefits are more limited for silane oxidation.
Original language | English (US) |
---|---|
Pages (from-to) | 346-359 |
Number of pages | 14 |
Journal | Accepted by Combustion and Flame |
Volume | 217 |
DOIs | |
State | Published - 2020 |
Bibliographical note
KAUST Repository Item: Exported on 2020-10-01Acknowledgements: The research reported in the present paper was supported by funding from King Abdullah University of Science and Technology (KAUST). RM was supported by the 1000 Young Talent of China program and a Matching program from Tsinghua University. The authors acknowledge Kiran Yalamanchi (KAUST) and Scott Shaw (KAUST) for providing access to the KAUST CloudFlame and for their help in the usage of the collision limit violation tool. The authors acknowledge Guillaume Blanquart (Caltech) and Binod Giri (KAUST) for fruitful discussions regarding the G4 thermochemical data calculations.