Mechanistic molecular motion of transition-metal mediated β-hydrogen transfer: quasiclassical trajectories reveal dynamically ballistic, dynamically unrelaxed, two step, and concerted mechanisms

Abstract

Quasiclassical direct dynamics reveal new dynamical mechanisms for metal-alkyl to ethylene β-hydrogen transfer.

Catalysis with a Skip: Dynamically Coupled Addition, Proton Transfer, and Elimination During Au- and Pd-Catalyzed Diol Cyclization

“…27,28,29,30,31,32,33,34,35 Recently, we discovered dynamic effects in a few stoichiometric organometallic reactions. For example, in the C-H activation between methane and [Cp*(PMe3)Ir III (CH3)] + , 36 as well as the -hydrogen transfer for [Cp*Rh III (Et)(ethylene)] + , 37 despite a fully characterized intermediate on the DFT-potential energy landscape, direct dynamics simulations revealed that the intermediate is either sometimes or always skipped due to dynamical coupling of multiple reaction steps through dynamic matching or the lack of IVR. We also discovered dynamic effects, specifically dynamical pathway branching, in the reaction between Tp(NO)(PR3)W and benzene 38 as well as Cp(PMe3)2Re and ethylene.…”

“…Previously, we referred to this type of mechanism as dynamically unrelaxed. 37 Third, there is the possibility where proton transfer precedes carbon-oxygen bond formation (Scheme 2b orange dotted arrow III). Similar to this suggestion, Singleton found that a proton can dynamically transfer before carbon-carbon bond cleavage during benzoylacetic acid decarboxylation.…”

“…On the Au surface, intermediate 2-Au is only slightly stabilized compared to TS1-Au and TS2-Au whereas on the Pd surface 2-Pd is 4-6 kcal/mol lower in energy than TS1-Pd and TS2-Pd, depending on the analysis of either the electronic or Gibbs surfaces. In our effort to understand how the qualitative and quantitative shape of a DFT potential energy surface relates to mechanisms that include dynamical motion, 37 we wanted to know if this more stabilized Pd-alkyl intermediate is dynamically skipped akin to the Au reaction or if it is stable enough to result in a genuine intermediate with substantial IVR. Therefore, we initiated and propagated 84 Pd trajectories beginning at TS1-Pd.…”

Computational mechanistic study in organometallic catalysis: Why prediction is still a challenge

“…Some workflows and corrections are becoming accepted for routine cases, that is, catalytic cycles where the solvent is not directly involved, 1,17 and a growing body of knowledge is being developed about what to do for exceptions, such as solvation changing the reaction pathway 32,33 and solvent-mediated/outer-sphere pathways 35 where the use of molecular dynamics is of growing importance. 30,31,[36][37][38] As the system considered becomes more complex, we would also encourage a degree of pragmatism: data that are consistent/deviate systematically may be "good enough" for a problem in hand, 1,34 whereas a full evaluation may be computationally costly and not add further chemical insights. Key to such decisions is often a validation against available experimental data, which, as we discuss below, can be difficult to obtain.…”

“…Furthermore, the treatment of solvent and salt effects, traditionally left out of models for computational simplicity, 1 continues to be explored and improved as well, 1,29–34 with some examples considered below. Some workflows and corrections are becoming accepted for routine cases, that is, catalytic cycles where the solvent is not directly involved, 1,17 and a growing body of knowledge is being developed about what to do for exceptions, such as solvation changing the reaction pathway 32,33 and solvent‐mediated/outer‐sphere pathways 35 where the use of molecular dynamics is of growing importance 30,31,36–38 . As the system considered becomes more complex, we would also encourage a degree of pragmatism: data that are consistent/deviate systematically may be “good enough” for a problem in hand, 1,34 whereas a full evaluation may be computationally costly and not add further chemical insights.…”