Comprehensive study of the thermochemistry of first-row transition metal compounds by spin component scaled MP2 and MP3 methods

The enthalpies of (i) the dissociation reactions of the carbonyl ligand in Cr(CO)(n) [n = 6 (1a), 5 (1b), 4 (1c)], Fe(CO)5 (2a), and Ni(CO)(n) [n = 4 (3a), 3 (3b), 2 (3c)], (ii) the dissociation reactions of the heteroligand L in Cr(CO)(5)L [L = CS (4a), Xe (4b), H-2 (4c), C2H4 (4d), C2F4 (4e)], Cr(CO)(3)L = C6H6 (5a), C6Me6 (5b)], and Fe(CO)(4)L [L = H-2 (6a), C2H4 (6b)], (iii) the deprotonation reactions of Cr(CO)(3)C6H6 (5a) and Fe(CO)(3)C4H6 (8), (iv) the protonation reaction of ferrocene (10), and (v) the hydrogenation reactions of Mn-2(CO)(10) (13) and CO2(CO)(8) (15) were calculated at the DFT/BP86, MP2, MP3, SCS-MP2, and SCS-MP3 levels and compared with the corresponding experimental data. The set of systems studied covers different types of chemical bonding and a broad range of reaction enthalpies (from a few kcal/mol to a few hundred kcal/mol), and it is thus proposed as a general benchmark for quantum chemical methods. It is shown that the erratic behavior of the low-order NIP approaches can be corrected by the newly developed spin component scaled perturbation theory (SCS-MR2, SCS-MP3). Both methods are based on a partitioning of the total correlation energy into contributions from antiparallel-(alphabeta) and parallel-spin (alphaalpha, betabeta) pairs of electrons, followed by a unique scaling procedure of the two correlation energy terms. The calculated thermochemical data at the SCS-MP2 and SCS-MP3 levels are significantly improved as compared to those from their standard counterparts. For the dissociation processes of 1-6 the mean absolute errors are 21.0 kcal/mol (MP2), 22.3 kcal/mol (MP3), 11.6 kcal/mol (SCS-MP2),4.1 kcal/mol (SCS-MP3), and 3.4 kcal/mol (DFT/BP86). It is shown that contrary to standard MP3, the new SCS-MP3 method is able to predict the reaction energetics for wide variety of TM compounds with an accuracy comparable to that from DFT. The SCS-MP3 enthalpy of the hydrogenation reaction of 13 (5.8 kcal/mol) agrees better with the experiment (8.7 +/- 0.3 kcal/mol) than the BP86 value (1.7 kcal/mol). The SCS-MP3 proton affinity of metal-protonated ferrocene 11 (214.8 kcal/mol) and of the agostic form 12 (203.9 kcal/mol) compare well with the experimental values (206-213 kcal/mol) and contrary to MP2 do not exclude the dynamic behavior of protonated ferrocene. It is suggested that for complex chemical systems including transition metals simultanous application of DFT and SCS-MP3 methods may be helpful to increase the reliability of the predictions.