Limited ionicity in poor protic ionic liquids: Association Gibbs energies.

Protic ionic liquids (PILs), made from anhydrous mixtures of Bronsted acids HA and bases B (HA + B → BH+ + A-), occasionally suffer from limited ionicity. In cases of "poor" PILs (<10% ionicity, e.g., using carboxylic acids), past simulations have hinted that ion-pair association, more than incomplete proton transfer, is at fault. To improve upon the Fuoss equation for predicting the degree of ion pairing, new electrostatic equations (including induced dipoles) are presented, for ion-pair and other associations that occur in anhydrous amine/carboxylic acid mixtures. The equations present the association Gibbs energies ΔGA (and thus the association constants KA) as functions of three fundamental properties: the acid/base mixing ratio (n = xA/xB), the HA-to-B proton-transfer strength (ΔpKa,ε=78), and the dielectric constant (relative permittivity) of the mixture (ε). Parameter values were obtained from fits to constant-dielectric quantum chemistry data (obtained and presented here). These ΔGA functions were then used to predict ΔGioniz values for the net ion-generating (autoionization) equilibrium in carboxylic acid/amine mixtures: 2B(HA)n⇄B(HA)n-dHB++A(HA)n+d-1 -, where n = xA/xB and d = degree of disproportionation. The agreement with experiment was excellent, demonstrating that these equations could have useful predictive power.

Improving the Yield and Rate of Acid-Catalyzed Deconstruction of Lignin by Mechanochemical Activation.

Lignin is a potential biomass feedstock from plant material, but it is particularly difficult to economically process. Inspired by recent ball-milling results, state-of-the-art quantum mechanochemistry calculations have been performed to isolate and probe the purely mechanochemical stretching effect alone upon acid-catalyzed deconstruction of lignin. Effects upon cleavage of several exemplary simple ethers are examined first, and with low stretching force they all are predicted to cleave substantially faster, allowing for use of milder acids and lower temperatures. Effects upon an experimentally known lignin fragment model (containing the ubiquitous ß-O-4 linkage) are next examined; this first required a mechanism refinement (3-step indirect cleavage, 1-step side reaction) and identification of the rate-limiting step under zero-force (thermal) conditions. Mechanochemical activation using very low stretching forces improves at first only yield, by fully shutting off the ring-closure side reaction. At only somewhat larger forces, in stark contrast, a switch in mechanism is found to occur, from 3-step indirect cleavage to the direct cleavage mechanism of simple ethers, finally strongly enhancing the cleavage rate of lignin. It is concluded that mechanochemical activation of the common ß-O-4 link in lignin would improve the rate of its acidolysis via a mechanism switch past a low force threshold. Relevance to ball-milling experiments is discussed.

Semicontinuum (Cluster-Continuum) Modeling of Acid-Catalyzed Aqueous Reactions: Alkene Hydration.

An effort is made to reduce the errors of continuum solvation models (CSMs) with semicontinuum modeling to achieve 3 kcal mol-1 agreement with experiment for acid-catalysis activation Gibbs energies. First, two underappreciated CSM issues are reviewed: errors in the CSM solvation Gibbs energies grow beyond 5 kcal mol-1 (i) as ions are made smaller and (ii) as water clusters grow larger. Second, the computational reproduction of the known Gibbs energies (ΔrG and Δ‡G) of the paradigmatic reaction ethene + H2O + H3O+ → TS+ → ethanol + H3O+ is attempted. It is argued that, despite the >5 kcal mol-1 solvation errors for ions, it is possible to employ error cancellation strategies to reduce the errors in the reaction and activation Gibbs energies to 3 kcal mol-1 accuracy. A new 3 kcal mol-1 effect due to solvent-molecule "placement" (confinement from 1 M bulk concentration) was isolated and proved useful. Third, computational reproduction of the known entropies (ΔrS and Δ‡S) of the paradigmatic reaction is attempted using Trouton's constant and neglect of solvent structure reorganization effects (which must cancel well for this reaction); this worked well for ΔrS but needs empirical correction of ∼11 cal mol-1 K-1 for Δ‡S due to solvent disorientation when H3O+ is consumed. These entropy estimates allow for enthalpy (ΔrH and Δ‡H) estimation from the Gibbs energy values. Fourth, two recommended options, including A + H3O+·2W → [AHOH2+·2W]‡, are shown to also work well for the activations of propene and isobutene.

The origin of the conductivity maximum in molten salts. III. Zinc halides.

In a continuing effort to master the reasons for conductivity maxima vs temperature in semicovalent molten halides, the structure and some transport properties of molten zinc halide are examined with ab initio molecular dynamics. Molten zinc halides are a special class of molten salts, being extremely viscous near their melting point (with a glassy state below it) and low electrical conductivity, and since they are also known (ZnI2) or predicted (ZnBr2 and ZnCl2) to exhibit conductivity maxima, they would be useful additional cases to probe, in case the reasons for their maxima are unique. Strong attractive forces in ZnX2 result in tight tetrahedral coordination, and the known mixture of edge-sharing vs corner-sharing ZnX4 tetrahedra is observed. In the series zinc chloride → bromide → iodide, (i) the ratio of edge-sharing vs corner-sharing tetrahedra increases, (ii) the diffusion coefficient of Zn2+ increases, and (iii) the diffusion coefficient of the anion stays roughly constant. A discussion of conductivity, with focus on the Walden product W = ηΛe, is presented. With predicted Haven ratios of 1-15 when heated toward their conductivity maxima, the physical chemistry behind molten zinc halide conductivity does not appear to be fundamentally different from other semicovalent molten halides.

Mechanical Activation Drastically Accelerates Amide Bond Hydrolysis, Matching Enzyme Activity.

Amide bonds, which include peptide bonds connecting amino acids in proteins and polypeptides, give proteins and synthetic polyamides their enormous strength. Although proteins and polyamides sustain mechanical force in nature and technology, how forces affect amide and peptide bond stability is still unknown. Using single-molecule force spectroscopy, we discover that forces of only a few hundred pN accelerate amide hydrolysis 109 -fold, an acceleration hitherto only known from proteolytic enzymes. The drastic acceleration at low force precedes a moderate additional acceleration at nN forces. Quantum mechanochemical ab initio calculations explain these experimental results mechanistically and kinetically. Our findings reveal that, in contrast to previous belief, amide stability is strongly force dependent. These calculations provide a fundamental understanding of the role of mechanical activation in amide hydrolysis and point the way to potential applications from the recycling of macromolecular waste to the design of bioengineered proteolytic enzymes.

The origin of the conductivity maximum vs. mixing ratio in pyridine/acetic acid and water/acetic acid.

Explanations are provided for the first time for the historically known locations of electrical conductivity maxima versus mixing ratio (mole fraction of acid, xA) in mixtures of (i) acetic acid with water and (ii) acetic acid with pyridine. To resolve the question for the second system, density-functional-based molecular dynamic simulations were performed, at 1:1, 1:2, 1:3, 1:5, and 1:15 mixing ratios, to gain vital information about speciation. In a zeroth-order picture, the degree of ionization (and hence conductivity) would be maximal at xA = 0.5, but these two examples see this maximum shifted to the left (water/acetic acid, xAmax = 0.06), due to improved ion stability when the effective dielectric constant is high (i.e., water-rich mixtures), or right (pyridine/acetic acid xAmax = 0.83), due to improved acetate stability via "self-solvation" with acetic acid molecules (i.e., acid-rich mixtures) when the dielectric constant is low. A two-parameter equation, with theoretical justification, is shown to reproduce the entire 0 < xA < 1 range of data for electrical conductivity for both systems. Future work will pursue the applicability of these equations to other amine/carboxylic acid mixtures; preliminary fits to a third system (trimethylamine/acetic acid) give curious parameter values.

The Carbocation Rearrangement Mechanism, Clarified.

The role of protonated cyclopropane (PCP(+)) structures in carbocation rearrangement is a decades-old topic that continues to confound. Here, quantum-chemical computations (PBE molecular dynamics, PBE and CCSD optimizations, CCSD(T) energies) are used to resolve the issue. PCP(+) intermediates are neither edge-protonated nor corner-protonated (normally) but possess "closed" structures mesomeric between these two. An updated mechanism for hexyl ion rearrangement is presented and shown to resolve past mysteries from isotope-labeling experiments. A new table of elementary-step barrier heights is provided. The mechanism and barrier heights should be useful in understanding and predicting product distributions in organic reactions, including petroleum modification.

The origin of the conductivity maximum in molten salts. II. SnCl2 and HgBr2.

The phenomenon of electrical conductivity maxima of molten salts versus temperature during orthobaric (closed-vessel) conditions is further examined via ab initio simulations. Previously, in a study of molten BiCl3, a new theory was offered in which the conductivity falloff at high temperatures is due not to traditional ion association, but to a rise in the activation energy for atomic ions hopping from counterion to counterion. Here this theory is further tested on two more inorganic melts which exhibit conductivity maxima: another high-conducting melt (SnCl2, σmax = 2.81 Ω(-1) cm(-1)) and a low-conducting one (HgBr2, σmax = 4.06 × 10(-4) Ω(-1) cm(-1)). First, ab initio molecular dynamics simulations were performed and again appear successful in reproducing the maxima for both these liquids. Second, analysis of the simulated liquid structure (radial distributions, species concentrations) was performed. In the HgBr2 case, a very molecular liquid like water, a clear Grotthuss chain of bromide transfers was observed in simulation when seeding the system with a HgBr(+) cation and HgBr3 (-) anion. The first conclusion is that the hopping mechanism offered for molten BiCl3 is simply the Grotthuss mechanism for conduction, applicable not just to H(+) ions, but also to halide ions in post-transition-metal halide melts. Second, it is conjectured that the conductivity maximum is due to rising activation energy in network-covalent (halide-bridging) melts (BiCl3, SnCl2, PbCl2), but possibly a falling Arrhenius prefactor (collision frequency) for molecular melts (HgBr2).

Challenges in predicting ΔrxnG in solution: the mechanism of ether-catalyzed hydroboration of alkenes.

Ab initio (coupled-cluster and density-functional) calculations of Gibbs reaction energies in solution, with new entropy-of-solvation damping terms, were performed for the ether-catalyzed hydroboration of alkenes. The goal was to test the accuracy of continuum-solvation models for reactions of neutral species in nonaqueous solvents, and the hope was to achieve an accuracy sufficient to address the mechanism in the "Pasto case": B2H6 + alkene in THF solvent. Brown's SN2/SN1 "dissociative" mechanism, of SN2 formation of borane-ether adducts followed by SN1 alkene attack, was at odds with Pasto's original SN2/SN2 hypothesis, and while Brown could prove his mechanism for a variety of cases, he could not perform the experimental test with THF adducts in THF solvent, where the higher THF concentrations might favor an SN2 second step. Two diboranes were tested: B2H6, used by Pasto, and (9BBN)2 (9BBN = 9-borabicyclo[3.3.1]nonane, C8H15B), used by Brown. The new entropy terms resulted in improved accuracy vs traditional techniques (∼2 kcal mol(-1)), but this accuracy was not sufficient to resolve the mechanism in the Pasto case.

The mechanism of permanganate oxidation of sulfides and sulfoxides.

Coupled-cluster (CCSD) and density functional computations are used to investigate historically competing mechanisms for the permanganate oxidation of sulfides and sulfoxides. The calculations all lead to a mechanism of 1,3-dipolar cycloaddition of permanganate, as opposed to historical mechanisms of attack of the sulfur atom by one O or by Mn. Such a mechanism, reminiscent of ozonolysis, may prevail in most permanganate oxidations. The ab initio activation enthalpies are in reasonable agreement with the experimental data; the ab initio activation entropies are not, possibly because of problems with Eyring equation assumptions.

The origin of the conductivity maximum in molten salts. I. Bismuth chloride.

A new theory is presented to explain the conductivity maxima of molten salts (versus temperature and pressure). In the new theory, conductivity is due to ions hopping from counterion to counterion, and its temperature dependence can be explained with an ordinary Arrhenius equation in which the frequency prefactor A (for hopping opportunities) and activation energy E(a) (for hopping) are density dependent. The conductivity maximum is due to competing effects: as density decreases, the frequency of opportunities for hopping increases, but the probability that an opportunity is successfully hopped decreases due to rising E(a) caused by the increased hopping distance. The theory is successfully applied to molten bismuth (III) chloride, and supported by density-functional based molecular dynamics simulations which not only reproduce the conductivity maximum, but disprove the long-standing conjecture that this liquid features an equilibrium between BiCl(3) molecules, and BiCl(2)(+) and BiCl(4)(-) ions that shifts to the left with increasing temperature.

Supra-supra, supra-antara, and stepwise-diradical pathways for an observed 16-electron double-[4 + 4] cycloaddition within metal-templated dialkyne dimers (PtX2)2(µ-R2PCCCCPR2)2.

Quantum chemistry calculations are used to provide insight into the cycloaddition of two dialkyne chains in initially monocyclic organoplatinum dimers of the type (PtX(2))(2)(µ-R(2)PC(4)PR(2))(2), where X = Cl or Me and R = Ph or Me. Previous experimental studies showed that the cycloaddition occurs with {X, R} = {Cl, Ph} but not {Me, Ph}. Two concerted pericyclic paths, a D(2h)-symmetry double-[π4s+π4s] "Hückel path" and a D(2)-symmetry double-[π4s+π4a] "Möbius path", were explored via orbital energy correlation diagrams (OECDs) computed using a singly occupied molecular orbital technique developed earlier. In accord with pericyclic reaction theory, the 16e(-) rearrangement is forbidden along the D(2h) Hückel path; four electrons would need to change their orbital symmetries. The D(2) Möbius path, afforded by the natural twist in the reactant structure which allows the desired Möbius orbital connectivity for a 4n rearrangement, is concluded to be a borderline forbidden pathway. This Möbius path creates avoided crossings in the OECD, which allows consistent orbital populations throughout the reaction, but it does not cause a change in intended orbital correlation, and the predicted activation barrier is rather high (∼50 kcal mol(-1)). The avoided crossings show strong coupling, but a clear HOMO-LUMO gap for the reaction is not produced. A stepwise path is also presented, with evidence of its diradical character.

Photochemistry studied with ab initio orbital-correlation and state-correlation plots: Classic cyclobutene ring opening, and the reaction of N2 with photoexcited O2.

Pericyclic reaction theory arose from ideas presented in 1965, based on orbital-energy correlation diagrams (Woodward and Hoffmann) and state-energy correlation diagrams (Longuet-Higgins and Abrahamson). Here we have used ab initio complete-active-space self-consistent field (CASSCF) calculations to generate such diagrams. First we present diagrams for the classic case of cyclobutene ring opening, to demonstrate agreement between the CASSCF results and the classic diagrams of both Woodward/Hoffmann and Longuet-Higgins/Abrahamson. Then we present diagrams for the more difficult cases of N(2) + photoexcited O(2), to produce either 2 NO or NNO + O. These N(2) + O(2) cases feature significant electron reorganization, for which elementary pencil-and-paper diagrams are less accurate. Finally, the benefits and limitations of such diagrams for predicting photochemistry are briefly discussed.

Semicontinuum Solvation Modeling Improves Predictions of Carbamate Stability in the CO2 + Aqueous Amine Reaction.

Quantum chemistry computations with a semicontinuum (cluster + continuum) solvation model have been used to cure long-standing misprediction of aqueous carbamate anion energies in the industrially important CO2 + aqueous amine reaction. Previous errors of over 10 kcal mol(-1) are revealed. Activation energies were also estimated with semicontinuum modeling, and a refined discussion of the competing hypothetical mechanisms for CO2 + monoethanolamine (MEA) is presented. Further results are also presented to demonstrate that the basicity of an amine (aqueous proton affinity) correlates only with CO2 affinity within an amine class: secondary amines have an extra CO2 affinity that primary amines do not have.

Molecular Dynamics Simulations of Proposed Intermediates in the CO2 + Aqueous Amine Reaction.

Ab initio molecular dynamics simulations of up to 210 ps have been performed on various aqueous intermediates postulated in the CO2 + amine reaction, important for CO2 capture. Observations of spontaneous deprotonation of aqueous carbamate zwitterions R1R2NHCOO(±) by bulk water (instead of additional amine, or via umbrella sampling) are reported apparently for the first time. Carbamic acid structures R1R2NCOOH were observed in some simulations, arising from zwitterions not via classical 1,3-H-shifts but via Grotthuss-style multiple-H(+) transfer pathways that involve bulk H2O and require carbamate anions R1R2NCOO(-) as an intermediate stage along the way. H(+)-bridging complexes, including not only Zundel ions [water·H(+)·water](+) but neutral carbamate complexes [carbamate(-)·H(+)·water], were observed in simulation. These results should assist efforts in improving underlying mechanisms for kinetic modeling.

Computational Study of Tungsten(II)-Catalyzed Rearrangements of Norbornadiene.

Rearrangements of norbornadiene (NBD, C7H8) to various alkylidenes, via a hypothetical 7-coordinate tungsten(II) complex W(CO)3I2(NBD), were studied using density-functional theory computations. An extensive search for intermediates and transition states of rearrangement was made. The theoretical method (basis sets and level of DFT) used was justified by new benchmark studies which compare optimized structural parameters to those from crystal structures of several different tungsten complexes. Transition-metal-catalyzed rearrangements of NBD are not as well-known as those of norbornene and are considerably more complicated than had been thought. This work predicts a large variety of intermediates which may be feasible targets for experimental synthesis. All the rearrangement paths to alkylidenes found here feature high activation energies of over 45 kcal mol(-1), implying that self-initiation for the ring-opening metathesis polymerization of NBD via tungsten(II) complexes must occur via an alternative mechanism.

Improved results for the excited states of nitric oxide, including the B/C avoided crossing.

The potential energy surfaces of ten electronic states of nitric oxide (NO) have been reexamined computationally, with state energies calculated using ab initio multireference methods. Our wave function expansions of 10x10(6) configurations improve upon the results of de Vivie and Peyerimhoff [J. Chem. Phys. 89, 3028 (1988)], who obtained excellent results from expansions of 16 000 configurations in 1988. We present results for the adiabatic properties r(e), B(e), T(e), and omega(e), demonstrating standard errors of 0.012 A, 0.026 cm(-1), 620 cm(-1), and 41 cm(-1), respectively. Vertical excitation energies and oscillator strengths are also presented, as are potential energy surface curves, with special attention to the B/C avoided crossing. The technical issue of state-averaging effects is also discussed.

Competing isomerizations: a combined experimental/theoretical study of phenylpentenone isomerism.

The possible competition of Z/E versus hydrogen-shift isomerization in (E)-5-phenyl-3-penten-2-one (E-1) and (E)-5-phenyl-4-penten-2-one (E-2) was studied, both experimentally and theoretically. Iodine-catalyzed isomerization experiments and computational modeling studies show that the equilibrated system consists predominantly of E-1 and E-2, with E-2 in moderate excess, and with no detectable amounts of the Z (cis) diastereoisomers. Density functional theory (DFT) calculations corroborated the free energy difference (Delta(r) and Delta(r) were -0.7 and -1.1 kcal mol(-1), respectively), and computations of Boltzmann-weighted (1)H NMR spectra were found to be useful in confirming the assignment of the isomers. The relevance of this equilibrium to earlier work on double-bond stabilization is discussed.

Photodissociation dynamics of the NO dimer. I. Theoretical overview of the ultraviolet singlet excited states.

Molecular orbital theory and calculations are used to describe the ultraviolet singlet excited states of NO dimer. Qualitatively, we derive and catalog the dimer states by correlating them with monomer states, and provide illustrative complete active space self-consistent field calculations. Quantitatively, we provide computational estimates of vertical transition energies and absorption intensities with multireference configuration interaction and equations-of-motion coupled-cluster methods, and examine an important avoided crossing between a Rydberg and a valence state along the intermonomer and intramonomer stretching coordinates. The calculations are challenging, due to the high density of electronic states of various types (valence and Rydberg, excimer and charge transfer) in the 6-8 eV region, and the multiconfigurational nature of the ground state. We have identified a bright charge-transfer (charge-resonance) state as responsible for the broadband seen in UV absorption experiments. We also use our results to facilitate the interpretation of UV photodissociation experiments, including the time-resolved 6 eV photodissociation experiments to be presented in the next two papers of this series.

A theoretical comparison of Lewis acid vs bronsted acid catalysis for n-hexane --> propane + propene.

Cracking of an all-trans n-alkane, via idealized Lewis acid and Bronsted acid catalysis, was examined using density functional theory. Optimized geometries and transitions states were determined for catalyst-reactant complexes, using AlCl3 and HCl.AlCl3 as the Lewis and Bronsted acids. For the Lewis acid cycle, hydride-transfer steps are seen to have large barriers in both forward and reverse directions, and an unstable physisorbed carbenium ion (lying 20 kcal mol(-1) above the chemisorbed intermediate) is the launching point for the beta-scission that leads to products. For the Bronsted acid cycle, proton-transfer steps have smaller barriers in both forward and reverse directions, and a semistable physisorbed alkanium ion is the launching point for the alkanium alpha-scission that leads to products. In the idealized Lewis cycle, formation of HCl units (and hence Bronsted acids) was found to be a common side reaction. A recent ionic-liquid catalysis study is mentioned as motivation, although our study is not a computational modeling study; we are more interested in the fundamental differences between Brosnted and Lewis mechanisms rather than merely mimicking a particular system. However, results of exploratory optimizations of various intermediates with Al2Cl7- as the catalyst are presented to provide the first step for future modeling studies on the ionic liquid system.