Comment on “The Water‐Exchange Mechanism of the [UO₂(OH₂)₅]²⁺ Ion Revisited: The Importance of a Proper Treatment of Electron Correlation” [F. P. Rotzinger Chem. Eur. J., 2007, 13, 800]

Abstract

Rotzinger did not study the uranyl(VI) aqua ion and his statements in reference [1] are therefore not supported by any computations. His claim that quantum-chemical calculations describing the structure of uranyl complexes and their dynamics using wave function based methods at the MP2 levels are flawed due to neglect of static correlation [1] is erroneous as we have pointed out in a previous publication.[2] In a recent paper [3] Rotzinger has found it expedient to base his previous claim on computational evidence and we applaud this. Despite considerable computational efforts his conclusions are still flawed; his own calculations (and some additional ones made by us) demonstrate very clearly that it is entirely proper to use the MP2 method in computational studies of the uranyl(VI) ion.There is no strict division between static and dynamic correlation. An operational definition which is often used is that dynamic correlation is what remains after a MC-SCF calculation, normally in a Complete Active Space (CAS) framework. A "full" CAS comprises Configuration State Functions resulting from distributing the valence electrons in all the bonding and antibonding orbitals in the system. Static correlation is also used to describe a situation with one or a few configurations in the reference space, which interacts strongly with the leading configuration due to the presence of nearly degenerate orbitals in the valence space. This situation can conveniently be referred to as static correlation of the near degeneracy type. Strong atomic coupling, which is not quenched by the ligands, can also give rise to several dominating configurations in the wave function, and this can also be regarded as static correlation. If neither of this occurs, there will normally be only one dominant configuration in the wave function. It is therefore reasonable to make a distinction between valence correlation, when the CAS wave function only has one dominant configuration, and static correlation, where the wave function must be described by several configurations with comparable weight. Another possible way of underlining this distinction is to use two concepts for static correlation namely static correlation of the near-degeneracy type and static correlation of the valence correlation type, but this is in our opinion cumbersome.Second-order Møller-Plesset perturbation theory (MP2) may not be applicable if the reference wave function has more than one dominant configuration. This can occur in systems with nearly degenerate orbitals in the valence space or for open shell systems where strong atomic coupling is important. The possible failure of MP2 in the first case is evident since the difference between orbital energies (or diagonal Fock matrix elements) appears in the denominators in the perturbation expansion.

Reply to the Comment on “The Water‐Exchange Mechanism of the [UO2(OH2)5]2+ Ion Revisited: The Importance of a Proper Treatment of Electron Correlation” [F. P. Rotzinger, Chem. Eur. J. 2007, 13, 800]

“…Precisely this has not been done in the two criticized [UO 2 A C H T U N G T R E N N U N G (OH 2 ) 5 ] 2 + studies. [2,3] In their Comment, [4] VWG present an a posteriori validation of the MP2 method for water-exchange reactions on uranyl(VI) complexes, whereby this approximation is not valid for reactions, which involve variations of the O=U=O geometry.…”

“…[2,3] In the preceding Comment, [4] Vallet, Wahlgren, and Grenthe (VWG) present MP2, MP3, and MP4 energies based on my [1] CAS-SCFA C H T U N G T R E N N U N G (12/11)-PCM geometries (Table 1 of reference [4]). Compared with Hartree-Fock (HF), CAS-SCFA C H T U N G T R E N N U N G (12/11) improves the too short axial U=O bond lengths by 0.056 (Table 4 of reference [4]), whereas the equatorial U À O bonds get slightly worse (elongation by 0.014 ). It should be noted that in most, if not all metal aqua ions, the computed metal-water bond lengths are too long due to the approximations in the solvation models and in the treatment of electron correlation.…”

“…[4] It should be noted that, in contrast to the water-exchange reactions, those in Table 6 in reference [9] involve variations of the U=O bonds (as verified for reaction 1 [10] ). The reaction with a MP2-CCSD(T) difference of~10 kJ mol À1 is likely to exhibit small structural changes in the UO 2 2 + fragments.…”

Investigation of the water exchange mechanism of the Plutonyl(VI) and Uranyl(VI) ions with quantum chemical methods

“…As already seen, ΔG ‡ and ΔG based on DFT are not reliable and DFT does not predict the preference for an associative mechanism of reaction (1) unequivocally. In Actinyl(VI) complexes, static electron correlation is present; its appropriate treatment [4,[59][60][61] is mandatory to obtain reliable results. For reaction (1), the most reliable data were obtained with LC-BOP-LRD or ωB97X geometries and frequencies, and GMC-QDPT2/SO-CI energies.…”

Theoretical studies of UO2(OH)(H2O)n+, UO2(OH)2(H2O)n, NpO2(OH)(H2O)n, and PuO2(OH)(H2O)n+ (n≤21) complexes in aqueous solution

“…These complexes present a considerable challenge [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38] for quantum chemical calculations, due to relativistic effects including the spin-orbit interaction as well as many-electron correlation energy. As demonstrated from the past work [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38] on the coordination complexes with uranyl, there are difficulties in treating these complexes especially in solution. Clark et al 14 investigated the uranyl ion at higher pH conditions and interpreted their results on the basis of a mixture of four and five hydroxide ligands coordinated to the uranyl ion.…”