Research Topics - UT Computational Chemistry Group

Selected Research Topics

GAS-PHASE ACID-BASE EQUILIBRIA

Proton transfer reactions are of fundamental importance in chemistry and biochemistry. Therefore, the knowledge of acidities and basicities of neutral molecules is essential. The gas-phase acidity of Brönsted acids is defined as the Gibbs free energy of acid dissociation reaction. Similarly the gas-phase basicity is defined as the acidity of bases conjugated acid.

The gas-phase acidities of ca. 60 monosubstituted anilines (with acidity span of almost 50 kcal/mol) have been calculated using density functional theory (DFT) at the B3LYP/6-311+G** level. [1] At this relatively simple level of theory, the calculated and available experimental acidities are in reasonable quantitative correlation. Substituent effects on the acidities were dissected separately into those operating in the neutral acid molecule and in its conjugated anion using the isodesmic homodesmotic reactions. All in all, both forms, neutral and anionic, are contributing in combination to make up the gross acidity of anilines. However, the contributions of the anions into the gross substituent effects are much larger than the substituent effects in the neutral anilines. Some of the systems were used in testing a relatively new theoretical model, COSMO-RS (conductor-like screening model for real solvents), using it for the prediction of pKa values in DMSO. The method proved to be rather accurate for showing pKatrends (R2 = 0.980 in DMSO). However, the predicted absolute pKa values were all somewhat lower (rmsd = 2.49 kcal/mol) than the respective experimental values.

We have employed Fourier transform ion cyclotron resonance (FT-ICR) and the Gaussian quantum chemistry composite methods W1 and G2 to experimentally and computationally analyze gas-phase basicities (GB) for a series of weak bases in the basicity region around and below water [2]. The aim of this study was to clarify the long-standing discrepancy between reported GB values for weak bases obtained via high-pressure mass spectrometry (HPMS) and ICR; the ICR scale is observed to be more than 2 times contracted compared to the HPMS scale. The computational results of this work support published HPMS data. This agreement improves with increasing sophistication of the computational method and is excellent at the W1 level. Several equilibria were also re-examined experimentally using FT-ICR. In the experiments with some polyfluorinated weak bases (hexafluoro-2-propanol and nonafluoro-2-methyl-2-propanol), it was found that two protonation processes compete in the gas phase: protonation on oxygen and protonation on fluorine. In these species, protonation on fluorine proceeds faster and is statistically favored over protonation on oxygen but leads to cations that are thermodynamically less stable than oxygen-protonated cations. The process may also lead to the irreversible loss of HF. The rearrangement of fluorine-protonated cations to oxygen-protonated cations is very slow and is further suppressed by the process of HF abstraction. These results at least partially explain the discrepancy between published HPMS data and earlier FT-ICR findings and call for the utmost care in using FT-ICR for gas-phase basicity measurements of heavily fluorinated compounds. The narrower dynamic range of ICR necessitates the measurement of several problematic bases and produces some differences between the ICR results in the present work and the published HPMS data; the wider dynamic range allows HPMS to overcome these difficulties in connecting the ladder.

Absolute (nonrelative) pKa calculations for substituted phenols were carried out in nonaqueous media, demonstrating the predictive power of the integral equation formalism PCM method with a mean unsigned error of 0.6 pK(a) units for DMSO and 0.7 pKa units for MeCN at the B3LYP/6-31+G** level of theory combined with the scaled B3LYP/6-311+G** gas-phase data. [3] At the same time, the correlation between the calculated and experimental pK(a) values yielded the value of the linear regression slope very close to unity for both DMSO and MeCN. Computation of pKa of neutral acids in nonaqueous solutions with a reasonable precision obviously depends on carefully tuned cavities, optimized for nonaqueous solutions. The ability of continuum solvation model to compensate charge escape from the cavity, which is prominent in the case of anions, is also required. And finally, good quality gas-phase data is essential to achieve required pKa precision.

Figure 1. Structure of Na2O-HClO4 complex [4] indicating total proton tranfer from acid to base.

The spontaneous proton transfer reactions are perhaps the most common reactions in solution. In contrast, similar reactions between neutral acids and bases are so far not known in tha gas-phase. We have carried out the computational investigation of interactions of different acid-base pairs regarding the nature and extent of such spontaneous proton transfer. [4] The selected acid-base pairs include the interactions of strong base (K2O) with acids of different strength (HClO4, HCl, and HF), and strong acid (HClO4) with bases ranging from K2O (GB = 322.8 kcal/mol) to H2O (GB = 157.6 kcal/mol). It was shown that spontaneous, unassisted proton transfer can take place in the gas-phase reactions of strong neutral Bronsted acids and bases. The reaction might be barrierless as in case of interactions between strong acids and bases, for example perchloric acid and alkali metal oxides or potassium oxide and halogen hydrides, or involve the encounter complex (hydrogen bonded acid-base cluster), which is separated from ion-pair by the transition state.

DFT B3LYP/6-311+G** calculations were performed to study the proton and lithium cation binding to the acetylacetone, hexafluoroacetylacetone, diacetamide, and hexafluorodiacetamide. [5] It was shown that the most stable Li+ adduct always corresponds to cyclic complex based on the trams, trans-keto form of the base. The product of protonation was found to be similar trans, traps-keto form based cyclic structure in case of diacetamide and hexafluorodiacetamide, while for acetylacetone and hexafluoroacetylacetone the protonation simply involves the addition of proton to (free) carbonyl oxygen in already cyclic enol form of the base with possible rotation of O-H bond.

STUDIES OF SUBSTITUENT EFFECTS OF NMR 17O AND 13C CHEMICAL SHIFTS

The studies of substituent effects, introduced in Tartu by professor Viktor Palm, have long and prosperous history in the Institute of Chemistry of University of Tartu. Vahur Mäemets has participated (together with I. A. Koppel, V. Nummert, and M. Piirsalu from chair of analytical chemistry) in NMR studies, based on 17O and 13C chemical shifts, of the influence of ortho, para, and meta as well as alkyl substituent effects in alkyl tosylates, phenyl esters of ortho-, para,- and meta-substituted benzoic acids, substituted phenyl and alkyl benzoates. [6-8]

THEORETICAL STUDIES OF REACTION MECHANISMS

Computational modeling of chemical reactions has gained wide recognition as it can provide useful information about important reaction intermediates and the role of various molecular species (base, solvent, ligand etc.) in the reaction mixture. The chair of chemical physics have been increasingly active in that direction.

Figure 2. Azaphilic addition intermediates of tetrazines with noncoordinating (1) and coordinating (4) substituents.

The possible reaction pathways between methyllithium and disubstituted 1,2,4,5-tetrazines (bearing methyl, methylthio, phenyl, and 3,5-dimethylpyrazolyl groups) were investigated by means of the density functional theory B3LYP/6-31G* method. [9] Solvation was modeled using the supermolecule approach, adding one tetrahydrofuran molecule to the complexes. Comparison of the calculated energies and structures for the alternate azaphilic and nucleophilic addition pathways showed that the azaphilic addition is kinetically favored over nucleophilic addition, while thermodynamically the nucleophilic addition is usually preferred. The coordination of the tetrazine molecule with methyllithium was found to play a crucial role in the process. These findings provide the first rationale for the experimentally observed unique reactivity of tetrazines toward polar organometallic reagents, suggesting the presence of a kinetically controlled process.

The Sonogashira cross-couplig reaction, consisting of oxidative addition, cis-trans isomerization, transmetalation, and reductive elimination, was computationally modeled using the DFT B3LYP/cc-pVDZ method for reaction between bromobenzene and phenylacetylene. [10] Palladium diphosphane was used as a catalyst, copper(I) bromide as a co-catalyst and trimethylamine as a base. The reaction mechanism was studied both in the gas phase and in dichloromethane solution using PCM method. The complete catalytic cycle is thermodynamically strongly shifted toward products (diphenylacetylene and regenerated palladium catalyst) and is exothermic being in accordance with experimental data. The rate-determining step is the oxidative addition, since the highest point on the Gibbs energy graph of the complete reaction is the transition state of this step. This conclusion is also supported by recent experimental data. The computed energy profile suggests that the transmetalation step is initiated by the dissociation of neutral ligand, while the activation Gibbs energy of this step is 0.1 kcal/mol in the gas phase.

The copper-free Sonogashira cross-coupling reaction consisting of oxidative addition, cis-trans isomerization, deprotonation, and reductive elimination was computationally modeled using the DFT B97D/cc-pVDZ method for reaction between phenyl bromide and phenylacetylene. [11] Tetrakis(triphenylphosphano)palladium was used as a catalyst and sec-butylamine as a base. The reaction mechanism was studied in dichloromethane solution. Oxidative addition proceeds through the biligated pathway, and the catalytically active palladium species is Pd(PPh3)3. Amines, present in the reaction mixture, can inhibit oxidative addition by coordinating to Pd(PPh3)3.

QSPR/QSAR PREDICTIONS OF CHEMICAL, PHYSICAL, AND BIOLOGICAL PROPERTIES

QSPR/QSAR studies have been widely used for relatively quick predictions on practically important chemical, physical, and biological properties of different substances. In our group that direction of research has been led by Kaido Tämm. The successful QSPR/QSAR models developed involve the boiling points of azeotropic mixtures, the penetration capability of cell-penetrating peptides, the log EC50 toxicity values of compounds causing bioluminescent repression of the bacterium strain Pseudomonas isolated from an industrial wastewater, activities of prospective protease inhibitors of the hepatitis C virus, pharmacological permeabilities, blood-brain barrier and human serum albumin binding. [12-17]

GAS-PHASE EQUILLIBRIA INVOLVING CESIUM CATION

Humic substances (HS), including humic and fulvic acids, play a significant role in the fate of metals in soils. The interaction of metal cations with HS occurs predominantly through the ionized (anionic) acidic functions. In the context of the effect of HS on transport of radioactive cesium isotopes in soils, a series of studies of the interaction between the cesium cation and model carboxylic acids was undertaken. [18-23]

The gas-phase cesium cation affinities (CsCAs) and basicities (CsCBs) for 56 simple neutral compounds (mostly aromatic molecules) and 41 anions (carboxylates and phenolates) were calculated using density functional theory (DFT), in the context of the interaction of Cs+with soil organic matter. The B3LYP/def2-TZVP method gives in general CsCAs and CsCBs in a good agreement with experimental data. The strong deviations in case of NO3– and CsSO4–anions need further experimental investigations as the high-level CCSD(T) calculations support B3LYP results. Different cesium cation complexation patterns between Cs+ and the neutral and anionic systems are discussed. As expected, the strongest CsCAs are observed for anions. The corresponding quantities are approximately by 4-5 times higher than for the neutral counterparts, being in the range 90-118 kcal/mol. The weakest cesium cation bonding is observed in the case of unsubstituted aromatic systems(11-15 kcal/mol).

Collision-induced dissociation (CID) of the Cs+ heterodimer adducts of the nitrate anion (NO3–) and a variety of substituted benzoates (XBenz–) [(XBenz–)(Cs+)(NO3–)]– produces essentially nitrate and benzoate ions. A plot of the natural logarithm of their intensity ratio, In(I(NO3–/I(XBenz–)], versus the calculated cesium cation affinity (DFT B3LYP method) of the substituted benzoate ions (equivalent to the enthalpy of heterolytic dissociation of the salt) is reasonably linear. This suggests that the kinetic method can be used as a source of data on the intrinsic interaction between the anionic and the cationic moieties in a salt.

Structure and energetics of the adducts formed between Cs+ and cesium carboxylate salts [Cs+RCOO–] were studied by the kinetic method and density functional theory (DFT). Clusters generated by electrospray ionization mass spectrometry from mixtures of a cesium salt (nitrate, iodide, trifluoroacetate) and carboxylic acids were quantitatively studied by CID. By combining the results of the kinetic method and the energetic data from DFT calculations, a scale of cesium cation affinity, CsCA, was built for 33 cesium carboxylates representing the first scale of cation affinity of molecular salts.

Substituent effects on the formation of cesium cation complexes with a series of 17 benzoic acids (AH), benzoates (A–), and the ion pairs (Cs+A–) are studied by density functional theory (DFT) and mass spectrometry. This study is positioned in the context of the fate of cesium in the environment, with emphasis of the influence of natural organic matter and humic substances. The bond length Cs+-(carboxylic O) in the various adduct geometries are discussed as regards the interaction strength, but quantitative relationships are limited by secondary effects arising mostly from long-distance interactions in systems bearing polar groups in meta-position. Relative cesium cation affinities of [Cs+A–] were experimentally determined by the kinetic method, i.e. by dissociating the required cesium cluster formed by electrospray ionization in a quadrupole ion-trap. Experiments and calculations are in agreement, except for adducts derived from 3- and 4-hydroxybenzoic acids. A change in the localization of the negative charge is proposed as a possible explanation for the divergence.

ZEOLITES

The Li+ cation complexes with one and two CO molecules have been studied computationally. The calculations reveal the conditions when two CO molecules could bind to one Li+ cation in zeolite Li-ZSM-5. In the absence of dicarbonyls ( at low adsorbate coverage or at high temperatures), the IR absorption bands can be assigned only on the basis of the Li+coordination number to the framework oxygen atoms. [24]

NOVEL B-N COMPOUNDS FOR CHEMICAL HYDROGEN STORAGE AND HYDROGEN TRANSFER REACTIONS

Interest in sustainable non-hydrocarbon-based fuels for transportation has grown as the realization that the supply of fossil fuels is limited and the deleterious environmental effects of burning them has come into public focus. As an alternative, the use of hydrogen (H2) has been proposed. H2 has a high energy content per mass unit (120 MJ/kg) vs. that of petroleum (44 MJ/kg) and it can be used to run a fuel cell which increases efficiencies compared to an internal combustion engine, simultaneously eliminating the formation of carbon, sulfur, and nitrogen oxide emissions as well as carbon particulates that are detrimental to the environment. However, H2 has a low energy content per unit volume [0.01 kJ/l at STP (8.4 MJ/l for liquid H2) vs. 32 MJ/l for petroleum]. For transportation applications, a fuel should ideally possess high energy content in a small volume as well as the minimum weight possible in order to maintain overall fuel efficiency. Currently, there are four leading methodologies to store H2: physical means (high pressure tanks), sorbents (nanoporous materials), metal hydrides, and so-called chemical hydrides.

The high gravimetric capacity of B–N materials makes them particularly appealing for hydrogen storage applications. It is important to note that basic research in this area also impacts several other fields such as transfer hydrogenation and B–N polymer synthesis. However, several important practical as well as more basic questions need to be answered.

Combined Fourier transform ion cyclotron resonance spectrometry (FT-ICR) study of the gas-phase protonation of ammonia-borane and sixteen amineboranes R1R2R3N-BH3 (including six compounds synthesized for the first time) has shown that, without exception, the protonation of amineboranes leads to the formation of dihydrogen. [25]

Figure 3. Unexpectedly curved bond paths in protonated MeNH2-BH3 complex.

The structural effects on the experimental energetic thresholds of this reaction were determined experimentally. The most likely intermediate and the observed final species (besides H2) are R1R2R3N-BH4+ and R1R2R3N-BH2+, respectively. Isotopic substitution allowed the reaction mechanism to be ascertained. Computational analyses (MP2/6-311+G(d,p) level) of the thermodynamic stabilities of the R1R2R3N-BH3 adducts, the acidities of the proton sources required for dihydrogen formation, and the structural effects on these processes were performed. It was further found that the family of R1R2R3N-BH4+ ions is characterized by a three-center, two-electron bond between B and a loosely bound H2 molecule. Unexpected features of some R1R2R3N-BH4+ ions were found. This information allowed the properties of amine/boranes most suitable for dihydrogen generation and storage to be determined.

Metal-organic frameworks (MOFs) are promising adsorbents for hydrogen storage. Density functional theory and second-order Moller-Plesset perturbation theory (MP2) were used to calculate the interaction energies between H2 and individual structural elements of the MOF-5 framework. [26] The strongest interaction, ?H = -7.1 kJ/mol, is found for the a-site of the OZn4(O2Ph)6, nodes. We show that dispersion interactions and zero-point vibrational energies must be taken into account. Comparison of calculations done under periodic boundary conditions for the complete structure with those done for finite models cut from the MOF-5 framework shows that the interactions with H2 originate mainly from the local environment around the adsorption site. When used within a Multi-Langmuir model, the MP2 results reproduce measured adsorption isotherms (the predicted amount is 6 wt % at 77 K and 40 bar) if we assume that the H2 molecules preserve their rotational degrees of freedom in the adsorbed state, This allows to discriminate between different isotherms measured for different MOF-5 samples and to reliably predict isotherms for new MOF structures.

We have reported rigorous quantum five-dimensional (5D) calculations of the coupled translation-rotation (T-R) eigenstates of a H2 molecule adsorbed in metal organic framework-5 (MOF-5), a prototypical nanoporous material, which was treated as rigid. [27] The anisotropic interactions between H2 and MOF-5 were represented by the analytical 5D intermolecular potential energy surface (PES) used previously in the simulations of the thermodynamics of hydrogen sorption in this system. The global and local minima on this 5D PES correspond to all of the known binding sites of H2 in MOF-5, three of which, alpha-, beta-, and gamma-sites are located on the inorganic cluster node of the framework, while two of them, the delta- and epsilon-sites, are on the phenylene link. In addition, 2D rotational PESs were calculated ab initio for each of these binding sites, keeping the center of mass of H2 fixed at the respective equilibrium geometries; purely rotational energy levels of H2 on these 2D PESs were computed by means of quantum 2D calculations. On the 5D PES, the three adjacent gamma-sites lie just 1.1 meV above the minimum-energy alpha-site, and are separated from it by a very low barrier. These features allow extensive wave function delocalization of even the lowest translationally excited T-R eigenstates over the alpha- and gamma-sites, presenting significant challenges for both the quantum bound-state calculations and the analysis of the results. Detailed comparison is made with the available experimental data.

Density functional theory is applied to study the heats of absorbtion in metal-organic frameworks with a hybrid functional to which a parametrized damped 1/r6 term has been added to account for dispersion. (B3LYP+D*). [28] This method is used with periodic boundary conditions to get the structures of the adsorption complexes. Dispersion has a substantial share on the calculated adsorption energies (46-77%). For these structures, adsorption energies are also calculated with a hybrid high-level (MP2 with complete basis set extrapolation), low level (B3LYP+D*) method. The MP2 calculations are performed on cluster models. Comparison is made with experimental heats of adsorption. B3LYP+D* underestimates heats of adsorption by about 5 kJ/mol, whereas hybrid MP2:B3LYP+D* slightly overestimates them by about 2 kJ/mol. With MP2:B3LYP+D*, also the mean absolute error is somewhat smaller, 3.8 kJ/mol compared to 5.6 kJ/mol for B3LYP+D*. Both the B3LYP+D* and the hybrid MP2/CBS:B3LYP+D* method predict the same sequence of binding energies for carbon monoxide (Ni > Mg > Zn) and carbon dioxide (Mg > Ni > Zn) adsorption on open metal cation sites in the CPO-27 metal-organic frameworks.