Quasiclassical Direct Dynamics Trajectory Simulations of Organometallic Reactions

Homogeneous metal-mediated organometallic reactions represent a very large and diverse reaction class. Density functional theory calculations are now routinely carried out and reported for analyzing organometallic mechanisms and reaction pathways. While density functional theory calculations are extremely powerful to understand the energy and structure of organometallic reactions, there are several assumptions in their use and interpretation to define reaction mechanisms and to analyze reaction selectivity. Almost always it is assumed that potential energy structures calculated with density functional theory adequately describe mechanisms and selectivity within the framework of statistical theories, for example, transition state theory and RRKM theory. However, these static structures and corresponding energy landscapes do not provide atomic motion information during reactions that could reveal nonstatistical intermediates without complete intramolecular vibrational redistribution and nonintrinsic reaction coordinate (non-IRC) pathways. While nonstatistical intermediates and non-IRC reaction pathways are now relatively well established for organic reactions, these dynamic effects have heretofore been highly underexplored in organometallic reactions. Through a series of quasiclassical density functional theory direct dynamics trajectory studies, my group has recently demonstrated that dynamic effects occur in a variety of fundamental organometallic reactions, especially bond activation reactions. For example, in the C-H activation reaction between methane and [Cp*(PMe₃)Ir^III(CH₃)]⁺, while the density functional theory energy landscape showed a two-step oxidative cleavage and reductive coupling mechanism, trajectories revealed a mixture of this two-step mechanism and a dynamic one-step mechanism that skipped the [Cp*(PMe₃)Ir^V(H)(CH₃)₂]⁺ intermediate. This study also showed that despite a methane σ-complex being located on the density functional theory surface before oxidative cleavage and after reductive coupling, this intermediate is always skipped and should not be considered an intermediate during reactive trajectories. For non-IRC reaction pathways, quasiclassical direct dynamics trajectories showed that for the isomerization of [Tp(NO)(PMe₃)W(η²-benzene)] to [Tp(NO)(PMe₃)W(H)(Ph)], there are many dynamic reaction pathway connections due to a relatively flat energy landscape and π coordination is not necessary for C-H bond activation through oxidative cleavage. Trajectories also showed that dynamic effects are important in selectivity for ethylene C-H activation versus π coordination in reaction with Cp(PMe₃)₂Re, and trajectories provide a more quantitative model of selectivity than transition state theory. Quasiclassical trajectories examining Au-catalyzed monoallylic diol cyclizations showed dynamic coupling of several reaction steps that include alkoxylation π bond addition, proton shuttling, and water elimination reaction steps. Overall, these studies highlight the need to use direct dynamics trajectory simulations to consider atomic motion during reactions to understand organometallic reaction mechanisms and selectivity.

Theory and Experiment Demonstrate that Sb(V)-Promoted Methane C-H Activation and Functionalization Outcompete Superacid Protonolysis in Sulfuric Acid

Sb(V) in strong Brønsted acid solvents is traditionally assumed to react with light alkanes through superacid protonolysis, which results in carbocation intermediates, H₂, and carbon oligomerization. In contrast to this general assumption, our density functional theory (DFT) calculations revealed an accessible barrier for C-H activation between methane and Sb(V) in sulfuric acid that could potentially outcompete superacid protonolysis. This prompted us to experimentally examine this reaction in sulfuric acid with oleum, which has never been reported because of presumed superacid reactivity. Reaction of methane at 180 °C for 3 h resulted in very high yields of methyl bisulfate without significant overoxidation. Our DFT calculations show that a C-H activation and Sb-Me bond functionalization mechanism to give methyl bisulfate outcompetes methane protonolysis and many other possible reaction mechanisms, such as electron transfer, proton-coupled electron transfer, and hydride abstraction. Our DFT calculations also explain experimental hydrogen-deuterium exchange studies and the absence of methane carbo-functionalization/oligomerization products. Overall, this work demonstrates that in very strong Brønsted acid solvent, Sb(V) can induce innersphere reaction mechanisms akin to transition metals and outcompete superacid reactivity.

Direct Dynamics Trajectories Reveal Nonstatistical Coordination Intermediates and Demonstrate that σ and π-Coordination Are Not Required for Rhenium(I)-Mediated Ethylene C-H Activation

The C-H activation reaction between Cp(PMe₃)₂Re and ethylene results in kinetic selectivity for the Re-vinyl hydride I over the thermodynamically more stable Cp(PMe₃)₂Re(η²-ethylene) π-complex II. While transition-state and variational transition-state structures were located for individual pathways leading to I and II, DFT and CCSD(T) energies predict a large kinetic selectivity of 10²-10⁴, which is incompatible with the experimental 10:1 ratio. DFT direct quasiclassical trajectories revealed that the transition states do not provide a qualitatively correct reaction mechanism or a quantitatively correct selectivity due to a nonstatistical σ-CH coordination intermediate that precedes the transition states for C-H activation and π coordination. Using metadynamics and quasiclassical direct dynamics, we show that trajectories for the reaction between Cp(PMe₃)₂Re and ethylene result in direct formation of either the Re-vinyl hydride I or the π-complex II. Trajectories leading to the Re-vinyl hydride skip σ-coordination and do not require π-coordination. Consistent with experiments, trajectory selectivity provides a relatively small kinetic selectivity for the Re-vinyl hydride.

Machine learning classification of disrotatory IRC and conrotatory non-IRC trajectory motion for cyclopropyl radical ring opening

Quasiclassical trajectory analysis is now a standard tool to analyze non-minimum energy pathway motion of organic reactions. However, due to the large amount of information associated with trajectories, quantitative analysis of the dynamic origin of reaction selectivity is complex. For the electrocyclic ring opening of cyclopropyl radical, more than 4000 trajectories were run showing that allyl radicals are formed through a mixture of disrotatory intrinsic reaction coordinate (IRC) motion as well as conrotatory non-IRC motion. Geometric, vibrational mode, and atomic velocity transition-state features from these trajectories were used for supervised machine learning analysis with classification algorithms. Accuracy >80% with a random forest model enabled quantitative and qualitative assessment of transition-state trajectory features controlling disrotatory IRC versus conrotatory non-IRC motion. This analysis revealed that there are two key vibrational modes where their directional combination provides prediction of IRC versus non-IRC motion.

Experiments and Direct Dynamics Simulations That Probe η2-Arene/Aryl Hydride Equilibria of Tungsten Benzene Complexes

Key steps in the functionalization of an unactivated arene often involve its dihaptocoordination by a transition metal followed by insertion into the C-H bond. However, rarely are the η²-arene and aryl hydride species in measurable equilibrium. In this study, the benzene/phenyl hydride equilibrium is explored for the {WTp(NO)(PBu₃)} (Bu = n-butyl; Tp = trispyrazoylborate) system as a function of temperature, solvent, ancillary ligand, and arene substituent. Both face-flip and ring-walk isomerizations are identified through spin-saturation exchange measurements, which both appear to operate through scission of a C-H bond. The effect of either an electron-donating or electron-withdrawing substituent is to increase the stability of both arene and aryl hydride isomers. Crystal structures, electrochemical measurements, and extensive NMR data further support these findings. Static density functional theory calculations of the benzene-to-phenyl hydride landscape suggest a single linear sequence for this transformation involving a sigma complex and oxidative cleavage transition state. Static DFT calculations also identified an η²-coordinated benzene complex in which the arene is held more loosely than in the ground state, primarily through dispersion forces. Although a single reaction pathway was identified by static calculations, quasiclassical direct dynamics simulations identified a network of several reaction pathways connecting the η²-benzene and phenyl hydride isomers, due to the relatively flat energy landscape.

Quantum-mechanical transition-state model combined with machine learning provides catalyst design features for selective Cr olefin oligomerization

The use of data science tools to provide the emergence of non-trivial chemical features for catalyst design is an important goal in catalysis science. Additionally, there is currently no general strategy for computational homogeneous, molecular catalyst design. Here, we report the unique combination of an experimentally verified DFT-transition-state model with a random forest machine learning model in a campaign to design new molecular Cr phosphine imine (Cr(P,N)) catalysts for selective ethylene oligomerization, specifically to increase 1-octene selectivity. This involved the calculation of 1-hexene : 1-octene transition-state selectivity for 105 (P,N) ligands and the harvesting of 14 descriptors, which were then used to build a random forest regression model. This model showed the emergence of several key design features, such as Cr-N distance, Cr-α distance, and Cr distance out of pocket, which were then used to rapidly design a new generation of Cr(P,N) catalyst ligands that are predicted to give >95% selectivity for 1-octene.

H/D exchange under mild conditions in arenes and unactivated alkanes with C6D6 and D2O using rigid, electron-rich iridium PCP pincer complexes

The synthesis and characterization of an iridium polyhydride complex (Ir-H4) supported by an electron-rich PCP framework is described. This complex readily loses molecular hydrogen allowing for rapid room temperature hydrogen isotope exchange (HIE) at the hydridic positions and the α-C-H site of the ligand with deuterated solvents such as benzene-d6, toluene-d8 and THF-d8. The removal of 1-2 equivalents of molecular H2 forms unsaturated iridium carbene trihydride (Ir-H3) or monohydride (Ir-H) compounds that are able to create further unsaturation by reversibly transferring a hydride to the ligand carbene carbon. These species are highly active hydrogen isotope exchange (HIE) catalysts using C6D6 or D2O as deuterium sources for the deuteration of a variety of substrates. By modifying conditions to influence the Ir-Hn speciation, deuteration levels can range from near exhaustive to selective only for sterically accessible sites. Preparative level deuterations of select substrates were performed allowing for procurement of >95% deuterated compounds in excellent isolated yields; the catalyst can be regenerated by treatment of residues with H2 and is still active for further reactions.

Machine Learning Analysis of Direct Dynamics Trajectory Outcomes for Thermal Deazetization of 2,3-Diazabicyclo[2.2.1]hept-2-ene

Experimentally, the thermal gas-phase deazetization of 2,3-diazabicyclo[2.2.1]hept-2-ene (1) results in the loss of N₂ and the formation of bicyclo products 3 (exo) and 4 (endo) in a nonstatistical ratio, with preference for the exo product. Here, we report unrestricted M06-2X quasiclassical trajectories initialized from the concerted N₂ ejection transition state that were able to replicate the experimental preference to form 3. We found that the 3:4 ratio results from the relative amounts of very fast (ballistic) exotype trajectories versus trajectories that lead to the 1,3-diradical intermediate 2. These quasiclassical trajectories provided a set of transition-state vibrational, velocity, momenta, and geometric features for the machine learning analysis. A selection of popular supervised classification algorithms (e.g., random forest) provided poor prediction of trajectory outcomes based on only transition-state vibrational quanta and energy features. However, these machine learning models provided more accurate predictions using atomic velocities and atomic positions, attaining ∼70% accuracy using initial conditions and between 85 and 95% accuracy at later reaction time steps. This increased accuracy allowed the feature importance analysis to reveal that, at the later-time analysis, the methylene bridge out-of-plane bending is correlated with trajectory outcomes for the formation of either the exo product or toward the diradical intermediate. Possible reasons for the struggle of machine learning algorithms to classify trajectories based on transition-state features is the heavily overlapping feature values, the finite but very large possible vibrational mode combinations, and the possibility of chaos as trajectories propagate. We examined this chaos by comparing a set of nearly identical trajectories that differed by only a very small scaling of the kinetic energies resulting from the transition-state reaction coordinate.

Synthetic investigations of low-coordinate (N-phosphino-amidinate) nickel chemistry: agostic alkyl complexes and benzene insertion into Ni-H

Treatment of (PN)NiX (X = NHdipp or OtBu; PN = N-phosphinoamidinate ligand) with Me2PhSiH in benzene solvent afforded the crystallographically characterized, antifacial-coordinated, dinuclear species 1, the formation of which corresponds to the hitherto unknown net Ni-H addition of two equivalents of the putative (PN)NiH intermediate across C[double bond, length as m-dash]C units within a single benzene molecule. Computational analysis supports the view of 1 as being comprised of two cationic (PN)NiII fragments ligated by a substituted butadiene dianion μ2-η3:η3-C6H82- bridging group. Also described is the formation and characterization of three-coordinate (PN)Ni(alkyl) complexes stabilized by β-agostic (alkyl = Et, 2; n-Bu, 3; n-hexyl, 4) or γ-agostic (alkyl = neopentyl, 5) interactions, and our efforts to employ 2 and 3 as synthons for the generation of (PN)NiHvia β-hydride elimination. Notably, compound 5 represents both the first crystallographically characterized three-coordinate Ni-alkyl complex featuring a heterobidentate ligation, and the first neutral γ-agostic NiII-alkyl complex.

Semiempirical Peak Fitting Guided by ab Initio Calculations of X-ray Photoelectron Spectroscopy Narrow Scans of Chemisorbed, Fluorinated Silanes

Here, we address the issue of finding correct CF₂/CF₃ area ratios from X-ray photoelectron spectroscopy (XPS) C 1s narrow scans of materials containing -CH₂CH₂(CF₂)_nCF₃ (n = 0, 1, 2, ...) moieties. For this work, we modified silicon wafers with four different fluorosilanes. The smallest had a trifluoropropyl (n = 0) moiety, followed by nonafluorohexyl (n = 3), tridecafluoro (n = 5), and finally, heptadecafluoro (n = 7) moieties. Monolayer deposition of the fluorosilanes was confirmed by spectroscopic ellipsometry, wetting, and XPS. Analysis of the trifluoropropyl (n = 0) surface and a sample of polytetrafluoroethylene provided pure-component XPS spectra for -CF₃ and -(CF₂)_n- moieties, respectively. Initial XPS C 1s peak fitting, which follows the literature precedent, was not entirely adequate. To address this issue, six different fitting approaches with increasing complexity and/or input from the Hartree-Fock theory (HF) were considered. Ultimately, we show that by combining HF results with empirical analyses, we obtain more accurate CF₂/CF₃ area ratios while maintaining high-quality fits.

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

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

Supermetal: SbF5-mediated methane oxidation occurs by C-H activation and isobutane oxidation occurs by hydride transfer

Sb^VF₅ is generally assumed to oxidize methane through a methanium-to-methyl cation mechanism. However, experimentally no H₂ is observed, and the mechanism of methane oxidation has remained unsolved for several decades. To solve this problem, density functional theory calculations with multiple chemical models (mononuclear and dinuclear) were used to examine methane oxidation by Sb^VF₅ in the presence of CO leading to the methyl acylium cation ([CH₃CO]⁺). While there is a low barrier for methane protonation by [Sb^VF₆]^-[H]⁺ (the combination of Sb^VF₅ and HF) to give the [Sb^VF₅]^-[CH₅]⁺ ion pair, H₂ dissociation is a relatively high energy process, even with CO assistance, and so this protonation pathway is reversible. While Sb-mediated hydride transfer has a reasonable barrier, the C-H activation/σ-bond metathesis mechanism with the formation of an Sb^V-Me intermediate is lower in energy. This pathway leads to the acylium cation by functionalization of the Sb^V-Me intermediate with CO and is consistent with no observation of H₂. Because this C-H activation/metal-alkyl functionalization pathway is higher in energy than methane protonation, it is also consistent with the experimentally observed methane hydrogen-to-deuterium exchange. This is the first time that evidence is presented demonstrating that Sb^VF₅ acts beyond a Bronsted superacid and involves C-H activation with an organometallic intermediate. In contrast to methane, due to the much lower carbocation hydride affinity, isobutane significantly favors hydride transfer to give the tert-butyl carbocation with concomitant Sb^V to Sb^III reduction. In this mechanism, the resulting highly acidic Sb^V-H intermediate provides a route to H₂ through protonation of isobutane, which is consistent with experiments and resolves the longstanding enigma of different experimental results for methane versus isobutane.

Synthesis of Structurally Diverse 3-, 4-, 5-, and 6-Membered Heterocycles from Diisopropyl Iminomalonates and Soft C-Nucleophiles

Herein, we present a general synthetic strategy for the preparation of 3-, 4-, 5-, and 6-membered heterocyclic unnatural amino acid derivatives by exploiting facile Mannich-type reactions between readily available N-alkyl- and N-aryl-substituted diisopropyl iminomalonates and a wide range of soft anionic C-nucleophiles without using any catalyst or additive. Fully substituted aziridines were obtained in a single step when enolates of α-bromo esters were employed as nucleophiles. Enantiomerically enriched azetidines, γ-lactones, and tetrahydroquinolines were obtained via a two-step catalytic asymmetric reduction and cyclization sequence from ketone enolate-derived adducts. Finally, highly substituted γ-lactams were prepared in one pot from adducts obtained using acetonitrile-derived carbanions. Overall, this work clearly demonstrates the utility of iminomalonates as highly versatile building blocks for the practical and scalable synthesis of structurally diverse heterocycles.