The Successful Merger of Theoretical Thermochemistry with Fragment-Based Methods in Quantum Chemistry

Abstract

CONSPECTUS: Quantum chemistry and electronic structure theory have proven to be essential tools to the experimental chemist, in terms of both a priori predictions that pave the way for designing new experiments and rationalizing experimental observations a posteriori. Translating the well-established success of electronic structure theory in obtaining the structures and energies of small chemical systems to increasingly larger molecules is an exciting and ongoing central theme of research in quantum chemistry. However, the prohibitive computational scaling of highly accurate ab initio electronic structure methods poses a fundamental challenge to this research endeavor. This scenario necessitates an indirect fragment-based approach wherein a large molecule is divided into small fragments and is subsequently reassembled to compute its energy accurately. In our quest to further reduce the computational expense associated with the fragment-based methods and overall enhance the applicability of electronic structure methods to large molecules, we realized that the broad ideas involved in a different area, theoretical thermochemistry, are transferable to the area of fragment-based methods. This Account focuses on the effective merger of these two disparate frontiers in quantum chemistry and how new concepts inspired by theoretical thermochemistry significantly reduce the total number of electronic structure calculations needed to be performed as part of a fragment-based method without any appreciable loss of accuracy. Throughout, the generalized connectivity based hierarchy (CBH), which we developed to solve a long-standing problem in theoretical thermochemistry, serves as the linchpin in this merger. The accuracy of our method is based on two strong foundations: (a) the apt utilization of systematic and sophisticated error-canceling schemes via CBH that result in an optimal cutting scheme at any given level of fragmentation and (b) the use of a less expensive second layer of electronic structure method to recover all the missing long-range interactions in the parent large molecule. Overall, the work featured here dramatically decreases the computational expense and empowers the execution of very accurate ab initio calculations (gold-standard CCSD(T)) on large molecules and thereby facilitates sophisticated electronic structure applications to a wide range of important chemical problems.

Breaking a bottleneck: Accurate extrapolation to “gold standard” CCSD(T) energies for large open shell organic radicals at reduced computational cost

“…This allowed the use of MP2 energy and CCSD(T) energies of fragments to extrapolate to the CCSD(T) energy of the target molecule . For the sake of brevity, the method is only discussed briefly in the methodology section, while complete details of the method can be found in the previous work, and a recent review . With a test set of 30 varied organic molecules using several basis sets, the mean absolute deviation (MAD) in the extrapolated CCSD(T) energies was found to be only 0.20–0.35 kcal/mol from full CCSD(T) energies.…”

“…[49] For the sake of brevity, the method is only discussed briefly in the methodology section, while complete details of the method can be found in the previous work, [49] and a recent review. [50] With a test set of 30 varied organic molecules using several basis sets, the mean absolute deviation (MAD) in the extrapolated CCSD(T) energies was found to be only 0.20-0.35 kcal/mol from full CCSD(T) energies. For alanine dipeptide and the amino-acids methionine and cysteine, the extrapolation errors were still within 1 kcal/mol, highlighting the good performance of the method.…”

Heats of formation of platonic hydrocarbon cages by means of high‐level thermochemical procedures

“…Wheeler et al introduced the RC-n hierarchy of isodesmic and homodesmotic reactions for hydrocarbons; [95,96] the connectivity-based hierarchy (CBH) of Ramabhadran and Raghavachari represents a generalization of the same concept. [97,98] RC-1 and CBH-0 both correspond to Pople-style isogyric bond separation reactions. For the problem at hand, they work out to:…”

W4‐17: A diverse and high‐confidence dataset of atomization energies for benchmarking high‐level electronic structure methods

“…[1][2][3][4][5][6][7][8][9][10][11] For example, the performance should improve along the sequence: atomization ! In particular, the accuracy of any given approximate method should increase as larger molecular fragments are conserved on the two sides of the reaction, due to an increasing degree of error cancellation between reactants and products.…”

“…In particular, the accuracy of any given approximate method should increase as larger molecular fragments are conserved on the two sides of the reaction, due to an increasing degree of error cancellation between reactants and products. [1][2][3][4][5][6][7][8][9][10][11] For example, the performance should improve along the sequence: atomization ! isogyric !…”