Klippenstein Stephen J, Harding Lawrence B, Ruscic Branko
Chemical Sciences and Engineering Division, Argonne National Laboratory , Argonne, Illinois 60439, United States.
J Phys Chem A. 2017 Sep 7;121(35):6580-6602. doi: 10.1021/acs.jpca.7b05945. Epub 2017 Aug 28.
The fidelity of combustion simulations is strongly dependent on the accuracy of the underlying thermochemical properties for the core combustion species that arise as intermediates and products in the chemical conversion of most fuels. High level theoretical evaluations are coupled with a wide-ranging implementation of the Active Thermochemical Tables (ATcT) approach to obtain well-validated high fidelity predictions for the 0 K heat of formation for a large set of core combustion species. In particular, high level ab initio electronic structure based predictions are obtained for a set of 348 C, N, O, and H containing species, which corresponds to essentially all core combustion species with 34 or fewer electrons. The theoretical analyses incorporate various high level corrections to base CCSD(T)/cc-pVnZ analyses (n = T or Q) using H, CH, HO, and NH as references. Corrections for the complete-basis-set limit, higher-order excitations, anharmonic zero-point energy, core-valence, relativistic, and diagonal Born-Oppenheimer effects are ordered in decreasing importance. Independent ATcT values are presented for a subset of 150 species. The accuracy of the theoretical predictions is explored through (i) examination of the magnitude of the various corrections, (ii) comparisons with other high level calculations, and (iii) through comparison with the ATcT values. The estimated 2σ uncertainties of the three methods devised here, ANL0, ANL0-F12, and ANL1, are in the range of ±1.0-1.5 kJ/mol for single-reference and moderately multireference species, for which the calculated higher order excitations are 5 kJ/mol or less. In addition to providing valuable references for combustion simulations, the subsequent inclusion of the current theoretical results into the ATcT thermochemical network is expected to significantly improve the thermochemical knowledge base for less-well studied species.
燃烧模拟的精度在很大程度上取决于大多数燃料化学转化过程中作为中间体和产物出现的核心燃烧物种的基础热化学性质的准确性。高水平的理论评估与活性热化学表(ATcT)方法的广泛应用相结合,以获得大量核心燃烧物种在0 K时生成热的经过充分验证的高保真预测。特别是,基于高水平从头算电子结构对一组348种含C、N、O和H的物种进行了预测,这基本上涵盖了所有电子数为34个或更少的核心燃烧物种。理论分析使用H、CH、HO和NH作为参考,对基于CCSD(T)/cc-pVnZ分析(n = T或Q)进行了各种高水平校正。对完全基组极限、高阶激发、非谐零点能、芯价、相对论和对角玻恩-奥本海默效应的校正按重要性递减顺序排列。给出了150种物种子集的独立ATcT值。通过(i)检查各种校正的大小,(ii)与其他高水平计算进行比较,以及(iii)与ATcT值进行比较,探索了理论预测值的准确性。这里设计的三种方法ANL0、ANL0-F12和ANL1的估计2σ不确定性对于单参考和适度多参考物种在±1.0-1.5 kJ/mol范围内,对于这些物种,计算出的高阶激发为5 kJ/mol或更低。除了为燃烧模拟提供有价值的参考外,预计随后将当前理论结果纳入ATcT热化学网络将显著改善对研究较少物种的热化学知识库。