Efficient Calculation of Accurate Reaction Energies-Assessment of Different Models in Electronic Structure Theory

In this work we analyze the accuracy and the efficiency of different schemes to obtain the complete basis set limit for CCSD(T). It is found that composite schemes using an MP2 increment to reach the basis set limit provide high accuracy combined with high efficiency. In these composite schemes the MP2-F12/cc-pVTZ-F12 method is suitable to compute the MP2 contribution at the basis set limit. We propose to use the def2-TZVP or the TZVPP basis sets at the coupled cluster level in combination with the cc-pVTZ-F12 basis set at the MP2 level to compute reaction energies close to the basis set limit, if high accuracy methods like CCSD(T)(F12*) or 56-extrapolations are no longer feasible due to the computational effort. The standard deviation of CCSD(T)+AMP2/cc-pVTZ-F12/def2-TZVP and CCSD(T)+Delta MP2/cc-pVTZ-F12/TZVPP is found to be only 0.93 and 0.65 kJ/mol for a test set of 51 closed shell reactions. Furthermore, we provide a comprehensive list of different computational strategies to obtain CCSD(T) reaction energies with an efficiency and accuracy measure. Finally we analyze how different choices of the exponent in the correlation factor (gamma) change the results when using explicitly correlated methods. The statistical results in this study are based on a set of 51 reaction energies in the range of 0.7 to 631.5 kJ/mol.