The pKBHX Hydrogen-Bond Basicity Scale: From Molecules to Anions.

The pKBHX (logarithm of complexation constant K of 4-fluorophenol with bases) hydrogen-bond basicity scale of neutral hydrogen-bond acceptors (HBAs) is extended to anionic HBAs. The scale is constructed for 26 anions through (i) the infrared measurement of K on NBu4+X- ion pairs in CCl4, (ii) the estimation of K from linear free energy relationships between measured K values and literature K values for various phenols in polar solvents, and (iii) the computation of K at the density functional theory level in CCl4. The scale extends on a 9.4 pK unit range from fluoride to tetraphenylborate. Considering a number of anions as organic functions substituted with unipolar substituents, their pKBHX values can be related to the Hammett-Taft substituent constants σ. Unipolar substituents (O- and S-) obey the same pKBHX versus σ relationships as dipolar ionic (N-N+R3) and dipolar (OH, CF3, NR2, or OR) ones for the nitrile, carbonyl, nitroso, nitro, sulfonyl, and phosphoryl functions. Like dipolar substituents, unipolar substituents at carbon and nitrogen operate by field-inductive and resonance effects, whereas substituents at sulfur and phosphorus operate only by the field-inductive effect.

Theoretical, Semiempirical, and Experimental Solvatochromic Comparison Methods for the Construction of the α1 Scale of Hydrogen-Bond Donation of Solvents.

Today, the hydrogen bonding donation (HBD) ability parameter of new solvents, α, is generally determined either by the Kamlet-Taft solvatochromic comparison of two probes, Reichardt betaine dye B(30) and 4-nitroanisole, or by the measurement of a single probe (e.g., solvatochromism of an iron coordination complex). This work highlights the shortcomings of these probes and recommends three replacement methods: (a) the theoretical comparison of the experimental and PCM-TD-DFT calculated transition energies ET(30) of B(30), (b) the semiempirical comparison of the experimental and McRae calculated ET(30), and, (c) for ionic liquids, the experimental comparison of ET(30) and ET(33) lying on the lower basicity of the betaine dye B(33) compared to B(30). These methods yield a new HBD parameter, α1, for 101 molecular solvents and 30 ionic liquids. The novelty is emblematic for water, with α1 = 1.54 instead of α (Kamlet-Taft) = 1.17. The solvent parameter α1 is not equivalent to the solute hydrogen-bond acidity parameter α2H, partly because of the self-association of HBD solvents.

Hydrogen-Bond Acceptance of Solvents: A 19F Solvatomagnetic ß1 Database to Replace Solvatochromic and Solvatovibrational Scales.

A variety of physicochemical properties and several hydrogen-bond donors have been used to define methods and to build scales aiming at measuring the hydrogen-bond acceptance of solvents. There is a great deal of confusion in these scales and methods. Solvatochromic, solvatocalorimetric, solvatovibrational, and 19F solvatomagnetic comparison methods are critically reviewed. Only two methods, the solvatomagnetic and the solvatocalorimetric ones, are able to yield reliable solvent hydrogen-bond acceptance scales. The solvatomagnetic ß1 scale defined from the 19F chemical shift of 4-fluorophenol is extended to many solvents including ionic liquids and green solvents. The results for about 240 hydrogen-bond acceptor solvents are organized in a numerical ß1 database. The comparison of ß1 with solvatochromic scales highlights their shortcomings, in particular for the important class of amphiprotic solvents. Therefore, the use of the 19F solvatomagnetic comparison method and of the solvatomagnetic ß1 scale is recommended in solvent effect studies.

Exploring the Solvatochromism of Betaine†30 with Ab Initio Tools: From Accurate Gas-Phase Calculations to Implicit and Explicit Solvation Models.

Betaine 30 is known for the extraordinary solvatochromism of its visible absorption band that goes from λ=882 nm in tetrachloromethane to λ=453 nm in water (Δλ=-429 nm). This large blueshift partly originates from a dramatic decrease of the dipole moment upon excitation. Despite several decades of research, experimental works still disagree on the exact value of the excess dipole moment, the orientation of the dipole moment of the excited-state, the role and amplitude of the change of the polarisability upon excitation as well as on the gas-phase excitation energy. In this work, we present an in-depth theoretical investigation. First, we carefully tested several levels of theory on the model system and next calculated the electric properties of betaine 30 at the CC2 level. Our best estimates are Δµ=-7 D for the excess dipole moment, that is, a significant decrease but no change of direction, a Δα value of -120 a.u. and a gas-phase vertical excitation energy of 1.127 eV. The implicit solvation models are able to reproduce the experimental trends, with large correlation coefficients for non-hydrogen-bond-donating solvents, the smallest root-mean-square deviation error being reached with the vertical excitation model (VEM). The explicit effective fragment potential method combined with time-dependent density functional theory (TD-DFT) in a QM/MM framework provides accurate estimates for hydrogen-bond-donating solvents, whereas the addition of a dispersion correction is needed to restore the correct solvatochromic direction in tetrachloromethane.

Solvatochromic Shifts in UV-Vis Absorption Spectra: The Challenging Case of 4-Nitropyridine N-Oxide.

4-Nitropyridine N-oxide is a well-known molecular probe for which the experimental UV/vis absorption spectrum has been measured in a large number of solvents. Previous measurements and their analyses suggest a dominant role of the solvent hydrogen-bond donation (HBD) capability in the solvatochromic shifts measured for the absorption spectra. Herein, we analyze these solvatochromic effects using a series of complementary approaches, including empirical solvent parameters, high-level calculation of the excited-state dipole and polarizability, several flavors of the polarizable continuum model, as well as dynamics using an effective fragment potential (EFP) description of the solvent molecules. First, applying a recently proposed set of solvent parameters, we show the importance of dispersion interactions for non-HBD solvents. This statement confronts advanced coupled-cluster and multireference calculations of dipole moments and polarizabilities of both the ground and excited states in gas phase. We further address the pros and cons of implicit solvent models combined to time-dependent density functional theory (TD-DFT) in describing the solvents effects for all (HBD and non-HBD) media, the simplest linear-response approach turning out to be the most adequate. Finally, we show that the explicit TD-DFT/EFP2 models work correctly for HBD molecules and allow for restoration of the main experimental trends.

A database of dispersion-induction DI, electrostatic ES, and hydrogen bonding α1 and ß1 solvent parameters and some applications to the multiparameter correlation analysis of solvent effects.

For about 300 solvents, we propose a database of new solvent parameters describing empirically solute/solvent interactions: DI for dispersion and induction, ES for electrostatic interactions between permanent multipoles, α1 for solute Lewis base/solvent Lewis acid interactions, and ß1 for solute hydrogen-bond donor/solvent hydrogen-bond acceptor interactions. The main advantage over previous parametrizations is the easiness of extension of this database to newly designed solvents, since only three probes, the betaine dye 30, 4-fluorophenol, and 4-fluoroanisole are required. These parameters can be entered into the linear solvation energy relationship A = A0 + di(DI) + eES + aα1 + bß1 to predict a large number of varied physicochemical properties A and to rationalize the multiple intermolecular forces at the origin of solvent effects through a simple examination of the sign and magnitude of regression coefficients di, e, a, and b. Such a rationalization is illustrated for conformational and tautomeric equilibria and is supported by quantum-mechanical calculations.

Solvatomagnetic Comparison Method: A Proper Quantification of Solvent Hydrogen-Bond Basicity.

The hydrogen-bond-acceptor basicity of an important class of solvents, the amphiprotic solvents (water, alcohols, primary and secondary amides, and carboxylic acids), has not yet been properly parametrized. In this work, the first scale of solvent hydrogen-bond basicity applicable to amphiprotic solvents is established by means of a new method that compares the 19F NMR chemical shifts of 4-fluorophenol and 4-fluoroanisole in hydrogen-bond-acceptor solvents. This so-called solvatomagnetic comparison method is free of the shortcomings of the solvatochromic comparison method used so far and is easier to carry out than the pure base calorimetric method. The validity of the new scale is assessed by good linear correlations with spectroscopic, thermodynamic, and kinetic solute properties depending on the solvent hydrogen-bond basicity. In such correlation analysis of solvent effects on physicochemical properties, solvent and solute hydrogen-bond basicity scales must not be mixed, since it is shown here that solute and solvent scales are not equivalent. A comprehensive collection of parameters quantifying the hydrogen-bond basicity is presented for 168 solvents.

Determination of a solvent hydrogen-bond acidity scale by means of the solvatochromism of pyridinium-N-phenolate betaine dye 30 and PCM-TD-DFT calculations.

Empirical parameters of solvents describing their hydrogen-bond (HB) acidity (e.g., the Kamlet-Taft α parameter) are often difficult to determine for new solvents because they are not directly related to a single definition process. Here, we propose a simple method based on one probe, the betaine dye 30, and one reference process, the solvatochromism of this dye, measured by its first electronic transition energy, ET(30). These ET(30) values are calculated within the time-dependent density functional theory framework, using a polarizable continuum solvent model (PCM). The part of ET(30) values that is not included in the PCM calculation is taken as the HB component of the measured ET(30) values, allowing us to deduce a solvent HB acidity parameter α1. The validity of this simple model is assessed by good linear correlations between α1 and a variety of solute properties mainly depending on the solvent's HB acidity. The quality of fit observed with α1 is at least comparable with that obtained by previous solvent HB acidity scales. The simplicity of our method is illustrated by the determination of α1 and of its companion, the electrostatic solvent parameter ES, for some new green solvents derived from glycerol.

Can quantum-mechanical calculations yield reasonable estimates of hydrogen-bonding acceptor strength? The case of hydrogen-bonded complexes of methanol.

The thermodynamics and some vibrational properties of hydrogen-bonded complexes of methanol with 23 hydrogen-bond acceptors (HBAs) have been determined in CCl(4) by FTIR spectrometry. The experimental sample contains carbon, nitrogen, oxygen, sulfur, fluorine, and chlorine organic bases and covers an energetic range of 13 kJ mol(-1) in the basicity scale (-ΔG), 22 kJ mol(-1) in the affinity scale (-ΔH), and 400 cm(-1) in the spectroscopic scale (Δν((OH))) (from benzene to trimethylphosphane oxide and amines). The experimental results in CCl(4) are compared to those computed in the gas phase at various levels of theory. Ninety five percent of the variance of the red shift and 89% of the variance of the intensification of the OH stretching upon hydrogen bonding are explained by gas-phase B3LYP/6-31+G(d,p) calculations. However, this level does not satisfactorily explain the thermodynamic properties. Only 68% of the variance of the methanol affinity (-ΔH) is taken into account. MP2/aug-cc-pVTZ//B3LYP/6-31+G(d,p) affinity calculations raise the explanation to 77% for all HBAs and to 93% when three outliers (Me(2)SO, Me(3)PO, and tetrahydrothiophene) are excluded. Discrepancies are analyzed in terms of experimental errors, calculation approximations, and solvation.

The diiodine basicity scale: toward a general halogen-bond basicity scale.

The new diiodine basicity scale pK(BI2) is quasi-orthogonal to most known Lewis basicity scales (hydrogen-bond, dative-bond and cation basicity scales). The diiodine basicity falls in the sequence N>P≈Se>S>I≈O>Br>Cl>F for the iodine-bond acceptor atomic site and SbO≈NO≈AsO>SeO>PO>SO>C=O>-O->SO(2) or PS≫-S->C=S≫N=C=S for the functionality of oxygen or sulfur bases. Substituent effects are quantified through linear free energy relationships, which allow the calculation of individual complexation constants for each site of polybases and thus the classification of aromatic ethers as carbon π bases and of aromatic amines, thioethers and selenoethers as N, S and Se bases, respectively. The pK(BI2) values of nBu(3)N(+)-N(-)C≡N, 2-aminopyridine and 1,10-phenanthroline reveal a superbasic nitrile, a hydrogen-bond-assisted iodine bond and a two-centre iodine bond, respectively. The diiodine basicity scale is a general inorganic but family-dependent organic halogen-bond basicity scale because organic halogen-bond donors such as IC≡N and ICF(3) have a stronger electrostatic character than I(2). The family independence can be restored by the addition of an electrostatic parameter, either the experimental pK(BHX) hydrogen-bond basicity scale or the computed minimum electrostatic potential.

Influence of the structure of polyfluorinated alcohols on Brønsted acidity/hydrogen-bond donor ability and consequences on the promoter effect.

The influence of substituents on the properties of tri- and hexafluorinated alcohols derived from 2,2,2-trifluoroethanol (TFE) and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) was examined. Measurements of specific solvent-solute interactions revealed that H-bond donation (HBD) of fluorinated alcohols is sensitive to the steric hindrance of the OH group, whereas their Brønsted acidity is dependent only on the number of fluorine atoms. For hexafluorinated alcohols (HFAs), their association with amines characterized by X-ray diffraction showed that the balance between HBD and acidity is influenced by their structure. Moreover, the ability of HFAs to donate H-bonds is exerted in synclinal (sc), synperiplanar (sp), and also antiperiplanar (ap) conformations along the C-O bond. Comparison of the effects of fluorinated alcohols as promoting solvents in three reactions is reported. The positive correlation between rate constants and H-bonding donation ability for sulfide oxidation and imino Diels-Alder reaction brings to light the role of this property, while acidity might have a minor influence. In the third reaction, epoxide opening by piperidine, none of these properties can clearly be put forward at this stage.

An enthalpic scale of hydrogen-bond basicity. 4. Carbon pi bases, oxygen bases, and miscellaneous second-row, third-row, and fourth-row bases and a survey of the 4-fluorophenol affinity scale.

The thermodynamics of the O-H...B hydrogen bond (HB) has been determined in CCl(4) by FTIR spectrometry for a wide variety of carbon pi bases, oxygen bases, and miscellaneous first- to fourth-row bases, using 4-fluorophenol as a reference hydrogen-bond donor (HBD). After inclusion of previously studied nitrogen, sulfur, and halogen bases, this 4-fluorophenol affinity scale contains 314 varied organic bases and ranges over 40 kJ mol(-1). The 4-fluorophenol affinity scale in CCl(4) is shown to be applicable to most HBDs in most media, provided a small family dependence is taken into account. The HB affinity orders are quantitatively established according to the atomic acceptor site or to its bearing functional group. A comprehensive survey of the influence of substituents on these affinity orders is then achieved, considering electronic and steric effects, as well as effects of vinylogy or iminology. Iminology is found to be more efficient than vinylogy for transmitting resonance effects. Steric effects are shown to be less important in HB affinity than in HB basicity since they mainly act on the HB entropy. The spatial proximity of two acceptor sites can favor complexation through three-center hydrogen bonds, leading to superhydrogen-bond bases on the affinity scale.

B3LYP and MP2 calculations of the enthalpies of hydrogen-bonded complexes of methanol with neutral bases and anions: comparison with experimental data.

To perfect a method for building a theoretical hydrogen-bond basicity scale, the enthalpy of hydrogen bonding between methanol and thirteen neutral and anionic bases (MeOH, MeNH2, Me2NH, Et2NH, Me3N, Et3N, Br-, CN-, SH-, Cl-, HCOO-, MeO-, F-) was calculated by DFT and ab initio methods. The theoretical results were compared to selected experimental ones. It appears that B3LYP/6-31+G(d,p) calculations are satisfactory for optimizing the geometry of complexes and giving a general order of basicity. However, they are deficient for reproducing the large effect of alkyl groups on the hydrogen-bond basicity of amines. This deficiency is explained by intermolecular perturbation theory calculations, which show that the alkylation of nitrogen dramatically increases the dispersion energy component not taken into account by the B3LYP functional. Of the methods considered, only MP2/aug-cc-pVTZ calculations are capable of reproducing the binding enthalpy within the experimental error for the first-row acceptor atoms N, O, and F, and of accounting for dispersion effects created by alkylation at the hydrogen-bond acceptor site.

An enthalpic scale of hydrogen-bond basicity. 3. Ammonia, primary, secondary, and tertiary amines.

A reliable enthalpic scale of hydrogen-bond acceptor strength (basicity) is built for aliphatic amines by means of a new infrared method, from the temperature variation of hydrogen-bond equilibrium constants. Enthalpies of hydrogen bonding to a reference hydrogen-bond acceptor, 4-fluorophenol, have been determined in CCl4 and/or C2Cl4 for ammonia and 68 primary, secondary, and tertiary amines. The scale spans from -23.8 kJ mol(-1) for i-Pr2NCH(Et)2 to -39.4 kJ mol(-1) for Et3N. This large variation is mainly explained by the basicity-enhancing electronic effects of alkyl groups, which can be overcompensated by dramatic basicity-decreasing steric effects. Relationships between DeltaH degrees and the change in electronic energy or the infrared shift of the OH stretching upon hydrogen bonding are studied and found useful in the prediction of the hydrogen bond enthalpies of amines with several hydrogen-bond acceptor sites. A careful statistical analysis of the enthalpy-entropy relationship shows an isoentropic tendency. The entropies of 65% of hydrogen-bonding reactions between aliphatic amines and 4-fluorophenol have a mean value of -55.1 +/- 4.2 J K(-1) mol(-1). Amines excluded from the isoentropic set are mainly severely hindered ones. The hydrogen-bond enthalpic scale can be useful in measuring the electrostatic character of Lewis bases.

The nicotinic pharmacophore: thermodynamics of the hydrogen-bonding complexation of nicotine, nornicotine, and models.

The thermodynamics of the hydrogen-bonding complexation of the acetylcholine agonists nicotine and nornicotine and of model pyridines, pyrrolidines, and N-methylpyrrolidines has been measured in CCl(4) by FTIR spectrometry toward a reference hydrogen-bond donor, 4-fluorophenol. Various methods are devised for measuring separately the hydrogen-bond acceptor strength of each nitrogen of nicotine and nornicotine: variation of the stoichiometry of complexation; correlations with electrostatic potentials on nitrogens and with substituent constants in the series of 3-substituted pyridines, 2-substituted pyrrolidines, and 2-substituted N-methylpyrrolidines; and linear free energy relationships between 4-fluorophenol and hydrogen fluoride hydrogen-bonded complexes. It is consistently found that nicotine and nornicotine have two active hydrogen-bond acceptor sites, the pyridine and pyrrolidine nitrogens, and that ca. 90% (for nicotine) and 80% (for nornicotine) of the 1:1 hydrogen-bonded complexes are formed to the pyridine nitrogen, although the pyrrolidine nitrogen is the first protonation site of nicotine and nornicotine in water. The low hydrogen-bond basicity of the pyrrolidine nitrogen in nicotine is mainly explained by the inductive electron-withdrawing and steric effects of the 2-(3-pyridyl) substituent. The partition of the Gibbs energy of the isomerism of complexation (AH...Nsp(2) <==> AH...Nsp(3)) into enthalpic and entropic contributions shows that the selectivity in favor of the pyridine nitrogen is driven by entropy. It is important to recognize the bifunctionality of nicotine in hydrogen bonding for understanding its lipophilicity and molecular recognition in non protonic media. When monoprotonated on their sp(3) nitrogen, nicotine and nornicotine keep, through their sp(2) nitrogen, a significant hydrogen-bond basicity which is greater than that of the ester group of acetylcholine.

Halogen-bond geometry: a crystallographic database investigation of dihalogen complexes.

X-ray crystal structures of 141 halogen-bonded complexes Y-X.B formed between homo- and heteronuclear dihalogens Cl(2), Br(2), I(2), IBr and ICl with O, S, Se, N, P and As Lewis bases show remarkable and constant geometrical features. The metrics of the halogen bond found in the gas phase for simple complexes [Legon (1999a). Angew Chem. Int. Ed. Eng. 38, 2686-2714] is supported (i). in the solid state, (ii). for new Lewis acids (I(2) and IBr), (iii). for new basic centers (Se, As and =N-) and (iv). for more complicated bases. The Y-X...B arrangement is more linear than the corresponding Y-H...B hydrogen bond and the axis of the Y-X molecule lies in the plane of the B lone pair(s), with a preference for the putative lone-pair direction within that plane. However, exceptions to this lone-pair rule are found for sterically hindered thiocarbonyl and selenocarbonyl bases. A bond-order model of the halogen bond correctly predicts the observed correlation between the shortening of the X...B distance and the lengthening, deltad(Y-X), of the Y-X bond. The expectation that the solid-state geometric parameters d(X...B) and deltad(Y-X) reflect the strength of the interaction is supported by their significant relationships with the solution thermodynamic parameters of Lewis acidity and basicity strength, such as the Gibbs energy of 1:1 complexation of Lewis bases with diiodine. This analysis of halogen-bonded complexes in the solid state reinforces the similarities already known to exist between hydrogen and halogen bonding.

Site of protonation of nicotine and nornicotine in the gas phase: pyridine or pyrrolidine nitrogen?

The gas-phase basicities (GBs) of nornicotine, nicotine, and model pyrrolidines have been measured by FT-ICR. These experimental GBs are compared with those calculated (for the two sites of protonation in the case of nicotine and nornicotine) at the B3LYP/6-311+G(3df,2p)//B3LYP/6-31G(d,p) level, or those estimated from substituent effects on the GBs of 2-substituted pyrrolidines, 2-substituted N-methylpyrrolidines, and 3-substituted pyridines. It is found that, in contrast to the Nsp(3) protonation in water, in the gas phase nornicotine is protonated on the pyridine nitrogen, because the effects of an intramolecular CH.Nsp(3) hydrogen bond and of the polarizability of the 3-(pyrrolidin-2-yl) substituent add up on the Nsp(2) basicity, while the polarizability effect of the 2-(3-pyridyl) substituent on the Nsp(3) basicity is canceled by its field/inductive electron-withdrawing effect. The same structural effects operate on the Nsp(3) and Nsp(2) basicities of nicotine, but here, the polarizability effect of the methyl group puts the pyrrolidine nitrogen basicity very close to that of pyridine. Consequently, protonated nicotine is a mixture of the Nsp(3) and Nsp(2) protonated forms.