NH and NCl homolytic bond dissociation energies and radical stabilization energies: An assessment of theoretical procedures through comparison with benchmark‐quality W2w data

Abstract

The performance of a large variety of contemporary density functional theory (DFT), double-hybrid DFT, and high-level Gaussian-n (Gn) procedures has been evaluated for the calculation of bond dissociation energies (BDEs) and radical stabilization energies (RSEs) associated with NAX bonds (X ¼ H, Cl). The chosen set of 62 NAX systems (31 NAH and 31 NACl) span a wide range of biologically relevant species. As reference values, we used benchmark-quality W2w data that we recently obtained as part of a systematic thermochemical study of substituent effects in these systems. Of the Gn schemes, the modified G4 procedures (G4-5H and G4(MP2)-6X) perform somewhat better than the corresponding standard G4 procedures for the NAX BDEs of these systems. For the NAH RSEs, G3X, G3X(MP2), G3X(MP2)-RAD, G4-5H, and G4(MP2)-6X emerge as excellent performers, with mean absolute deviations (MADs) from the benchmark W2w values of 0.9-1.4 kJ mol -1 . However, for the NACl RSEs, G4 is the best performer, with an MAD of 1.7 kJ mol -1 . The BDEs of both NAH and NACl bonds represent a challenge for DFT procedures. In particular, only a handful of functionals (namely, B3P86, M05-2X, M06-2X, and ROB2-PLYP) perform well, with MADs 4.5 kJ mol À1 for both bond types. Nearly all of the considered DFT procedures perform signifi-

Stability of the chlorinated derivatives of the DNA/RNA nucleobases, purine and pyrimidine toward radical formation via homolytic CCl bond dissociation

“…By way of contrast, the worst performance is obtained in the case of BH&HLYP (MAD = 43.1 kJ mol −1 ). The notably poor performance of BH&HLYP for the computation of BDEs have been reported previously, for example, in the case of NX (X = H, Cl, and Br) bonds . The double‐hybrid DFT methods offer MADs ranging from 2.6 (PWPB95‐D3) to 14.2 (B2‐PLYP) kJ mol −1 .…”

A dataset of highly accurate homolytic NBr bond dissociation energies obtained by Means of W2 theory

“…The reduced‐order variant of G4, namely G4(MP2), offers the second worst RMSD (6.7 kJ mol −1 ) of all of the considered methods, though the closely related G4(MP2)−6X procedure offers a slight improvement (5.7 kJ mol −1 ). The relatively poor performance of G4(MP2) has also been observed for the calculation of NH and NCl BDEs . The CBS‐QB3 procedure is the worst performing of all the considered thermochemical protocols (RMSD = 9.8 kJ mol −1 ).…”

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.…”

“…[1][2][3][4][5][6][7][8][9][10][11] For example, the performance should improve along the sequence: atomization ! [10][11][12][13][14][15][16][17][18][19][20] For example, the root mean square deviations (RMSDs) for a wide range of dispersion-corrected DFT methods for linear ! isodesmic !…”