摘要:

Nuclear and Electronic Spins as Dynamic and Kinetic Probes. Annual Meeting of the Deutsche Bunsen-Gesellschaft für Physikalische Chemie, Potsdam, Germany, 9–11 May, 2002

ISSN:1463-9076     

DOI:10.1039/b210208k

出版商:RSC    

年代:2002

数据来源: RSC

摘要:

IntroductionTransition metal NMR is an important analytical probe of the structure and reactivity of inorganic and organometallic compounds.1The key property, the chemical shift, can be computed with reasonable accuracy2using the appropriate tools of modern density functional theory (DFT).3The usual approach, computation of magnetic shieldings for static, isolated molecules in their equilibrium geometries, may be insufficient if comparison with experiments conducted in polar solvents and at ambient temperature is sought. Protic solvents such as water provide not only polar media, but also can interact with solutes by way of hydrogen bonds. Modelling the concomitant effects on structures and properties of these solutes is a considerable challenge for computational chemistry.4A popular way to mimic nature's way of producing observables averaged over large statistical ensembles is to perform classical molecular dynamics (MD) simulations for representative subsets, for instance, for a single molecule together with a limited number of solvent molecules, on potential energy surfaces generated on the fly. Computed magnetic shieldings, averaged for snapshots along MD trajectories, have been used to assess thermal and solvent effects on chemical shifts.5,6We have recently used such an approach to model transition metal chemical shifts in aqueous solution.6To this end,δ(51V) values of several aqueous vanadates and peroxovanadates were obtained by performing Car–Parrinello7MD simulations in small, periodic water boxes (containing of the order of 30 water molecules) and averaging the magnetic shieldings computed for a number of snapshots along the trajectory over a period of 1–2 ps.5The differences between the resulting equilibrium and averaged chemical shift values,δeandδ300 K, respectively, amounted to a few dozen ppm only. This effect is very small compared to the total chemical shift range of several thousands of ppm for the transition metal nucleus.These vanadium complexes bear either one positive or one negative charge. Larger solvent effects are to be expected for systems with higher charges. [Fe(CN)5(NO)]2−(1) and [Fe(CN)6]4−(2) were chosen as promising candidates for this purpose, as both are prominent textbook examples of coordination compounds,8and theδ(57Fe) value of1had only recently been determined experimentally atδ = 2004,9that is, significantly shielded with respect to that of2,δ = 2455. It is of interest to see if the CPMD-based computational protocol outlined above is able to describe, at least qualitatively, these chemical shifts and, especially, the difference between them.Indeed, based on CPMD simulations large thermal and solvent effects onδ(57Fe) are obtained for1and, in particular, for2. It turns out, however, that a considerable discrepancy remains between the simulated and observed chemical shift of2in water. Therefore, a different dynamical approach without periodic boundary conditions has been tested, in which the solvent is described by a classical force field. Such combined QM/MM approaches are frequently applied to study solvent effects, and very promising results are obtained for1and2. The key findings of this latter approach have been communicated separately.10In the present paper, we present a full account of both CP- and QM/MM-based MD simulations. These two very different methods afford remarkably consistent results concerning the chemical shifts of dianionic1.

ISSN:1463-9076     

DOI:10.1039/b202894h

出版商:RSC    

年代:2002

数据来源: RSC