Hydrogen Bonding as a Modulator of Aromaticity and Electronic

ARTICLE pubs.acs.org/JPCA

Hydrogen Bonding as a Modulator of Aromaticity and Electronic Structure of Selected ortho-Hydroxybenzaldehyde Derivatives Aneta Jezierska-Mazzarello,†,* Jaroszaw J. Panek,† Halina Szatyzowicz,‡ and Tadeusz Marek Krygowski§ †

University of Wroczaw, Faculty of Chemistry, 14 F. Joliot-Curie, 50-383 Wroczaw, Poland Faculty of Chemistry, Warsaw University of Technology, Noakowskiego 3, 00-664 Warsaw, Poland § Department of Chemistry, Warsaw University, Pasteura 1, 02-093 Warsaw, Poland ‡

bS Supporting Information ABSTRACT: Properties of hydrogen bonds can induce changes in geometric or electronic structure parameters in the vicinity of the bridge. Here, we focused primarily on the influence of intramolecular H-bonding on the molecular properties in selected ortho-hydroxybenzaldehydes, with additional restricted insight into substituent effects. Static models were obtained in the framework of density functional theory at B3LYP/6-311+G(d,p) level. The electronic structure parameters evolution was analyzed on the basis of Atoms In Molecules (AIM) and Natural Bond Orbitals methods. The aromaticity changes related to the variable proton position and presence of substituents were studied using Harmonic Oscillator Model of Aromaticity (HOMA), Nucleus-Independent Chemical Shift (NICS) and AIM-based parameter of Matta and HernandezTrujillo. Finally, CarParrinello molecular dynamics was applied to study variability of the hydrogen bridge dynamics. The interplay between effects of the substitution and variable position of the bridged proton was discussed. It was found that the hydrogen bond energies are ca. 910 kcal/mol, and the bridged proton exhibits some degree of penetration into the acceptor region. The covalent character of the studied hydrogen bond was most observable when the bridged proton reached the middle position between the donor and acceptor regions. The aromaticity indexes showed that the aromaticity of the central phenyl ring is strongly dependent on the bridged proton position. Correlations between these parameters were found and discussed. In the applied time-scale, the analysis of time evolution of geometric parameters showed that the resonance strengthening does not play a crucial role in the studied compounds.

1. INTRODUCTION Compounds with inter- and intramolecular hydrogen bonds have been subjects of many experimental and theoretical studies. The formation of H-bonding is responsible for very diverse molecular features.13 Its presence, especially occurrence of the proton-transfer phenomenon is responsible for many processes at molecular level, e.g., it determines the course of numerous chemical reactions as well as biochemical processes, including enzyme catalysis4 or proton transport.5 Therefore, there is still a need to understand the nature of hydrogen bonding and its properties.6 The effect of substituents has an important influence on chemical and physicochemical properties, among others—on the strength of H-bonding.712 The position of the bridged proton can be modulated by various effects. The most important among them are steric, inductive and resonance ones, because they have a direct or through-bond influence on the exhibited molecular properties.12 The reverse is also true: variations in proton position are able to modulate deeply the geometric and electronic structure of the molecular skeleton. This brings up an important question of how to control the dynamics of the bridged proton and to make it useful to control, e.g., for biological r 2011 American Chemical Society

or industrial processes where the proton transfer phenomena play a crucial role? Another important, and still open issue, is the nature of H-bonding, which, depending on the distance between the donor and acceptor atoms, could even have a dominantly covalent character.13 Quantitative approach to H-bonding properties based on electronic structure is, therefore, of great importance for deeper understanding of interactions at molecular level. This knowledge might be used to design systems with modulated position of the bridged proton, and further—to steer the properties of biologically active compounds as well as of new materials at molecular level.14,15 A proper description of H-bonding features, such as its energy, is significant in the quest for a better insight in the biochemical interactions as well as in the design of new compounds using rational solutions.16,17 After an extensive study, Laurence and Berthelot proposed to measure the strength of hydrogen bond acceptors from the Gibbs energy change for the formation of 1:1 H-bonded complexes formed Received: June 18, 2011 Revised: November 24, 2011 Published: November 30, 2011 460

dx.doi.org/10.1021/jp205730t | J. Phys. Chem. A 2012, 116, 460–475

The Journal of Physical Chemistry A

ARTICLE

Figure 1. Structures and atom numbering scheme (only atoms of interest are numbered) of the investigated compounds: Iortho-hydroxybenzaldehyde, II2-hydroxy-4-nitrobenzaldehyde, III2,4-dihydroxybenzaldehyde, IV2-hydroxy-5-nitrobenzaldehyde, and V2,5-dihydroxybenzaldehyde. Dotted line indicates an intramolecular hydrogen bond.

parameters and further modulate the strength of interaction.26 Therefore, it is necessary to select suitable theoretical approaches to describe such interactions.27 These compounds offer, however, a possibility of fine-tuning of the desired properties by substitution, and the electron conjugation provided by the aromatic ring increases the extent of the substituent influence. The bridge proton position is the chosen primary factor modulating molecular properties in this study, but—as mentioned above— the possibility of introducing substituent effects should not be neglected. Therefore, as model structures for the current study, five aldehyde-type compounds were investigated (see Figure 1). These compounds are substituted at the phenyl ring in order to introduce the resonance and inductive effects. First, the phenyl ring is substituted by a hydroxyl group in the ortho position with respect to the aldehyde moiety, which was essential to provide a strong OH...O intramolecular hydrogen bond in the structure. The properties of OH...O bridges are objects of many studies based on various sets of compounds as well as methodologies.2831 This kind of interaction can have a very shallow potential energy surface (PES), allowing the investigation of proton transfer phenomena and associated effects. Four of the compounds contain additional hydroxyl or nitro groups as substituents in the para position with respect to either the aldehyde or hydroxyl groups in order to introduce an adjective substituent effect as a secondary factor. The main goal of the current study is a multifactor investigation of hydrogen bonding in a series of ortho-hydroxybenzaldehydes, in particular: (i) Proton position in the bridge as a factor modulating structure of the molecular skeleton; (ii) Impact of the bridge proton on the electronic structure in light of the Atoms In Molecules (AIM) theory15 and Natural Bond Orbitals (NBO) method;32

between the hydrogen bond acceptors (bases) and reference hydrogen bond donor.17 They found that the strength of the hydrogen bond acceptor is determined by various variables, e.g., the position of the acceptor in the periodic table, its polarizability and resonance effects introduced by surrounding substituents. The method differs from the previously introduced pKa scale of proton transfer basicity. In the case of aromatic compounds, the substituents of the aromatic ring are able to directly modulate the hydrogen bond strength, i.e., via steric effects (for example in the ortho-hydroxy Schiff and Mannich bases and related compounds),1821 and also indirectly, when the substituent is present in para position. This effect is known as the direct conjugation effect, when the substituent accepts electrons. The counter polar conjugation effect is related to substituents which are electron donors.9 The effect of substituents is widely investigated in terms of intermolecular H-bonding, where the binding energy is influenced by the position of the substituents in the aromatic ring as well as by their kind, e.g., electrophilic or nucleophilic.22,23 The influence of intermolecular H-bonding on aromaticity of the ring, which is additionally variously substituted, was also deeply studied.24 Intramolecular hydrogen bonding in substituted aromatic compounds can also lead, in some cases, to formation of quasi-aromatic rings and coupling between donor and acceptor moieties. In this way, the proton position in the bridge can be coupled effectively with properties of the aromatic ring (especially its degree of aromaticity), and possible tautomerism can influence dominant resonance structures of the aromatic skeleton. However, at this point, we have to mention the modulating influence of resonance effects on the molecular properties, significant in systems containing intramolecular H-bond.25 In complexes where both components contain aromatic rings, it is much more difficult to handle fine-tuning of the H-bonding properties because both steric and inductive effects influence the geometric 461

dx.doi.org/10.1021/jp205730t |J. Phys. Chem. A 2012, 116, 460–475


ARTICLE

(iii) Evolution of the aromaticity of the molecule as a function of the proton position, studied on the basis of Harmonic Oscillator Model of Aromaticity (HOMA),33,34 NucleusIndependent Chemical Shift (NICS),35 and AIM-based parameter of Matta and Hernandez-Trujillo;36,37 (iv) Application of CarParrinello Molecular Dynamics (CPMD)38 to analyze the time evolution of interatomic distances and the structural flexibility of the studied orthohydroxybenzaldehydes. The compounds were chosen in such a way that they do not possess large enough acceptor basicity (pKa) to make the proton transfer probable. Thus, the impact of the substituent on the bridge proton will be conveniently measured “from the donor side only”. The study will provide answers to the following general question: how large modifications in the structure of an ortho-hydroxybenzaldehyde derivative can be achieved by variation of the proton position in the intramolecular hydrogen bridge, and how much these effects are influences by exemplary substituents?

simulations were performed using the Gaussian03 suite of programs.51 The net atomic charges according to the NBO scheme were computed using the NBO Version 3.1 implemented in Gaussian03 suite of programs.52 Next, selected aromaticity indexes were calculated to analyze the aromaticity changes upon the intramolecular H-bond dynamics and the presence of substituents in the phenyl ring: (i) Harmonic Oscillator Model of Aromaticity (HOMA)33,34,53 index was calculated according to the equation: HOMA ¼ 1

α n ðRopt Ri Þ2 n i¼1

∑

ð1Þ

where n indicates the number of bonds taken into the summation and α is an empirical constant (equal to 257.7 for the CC bonds). Ropt indicates an optimized bond length 1.388 Å in the case of CC bonds. The individual bond lengths are depicted by Ri (these are usually experimental or calculated bond lengths).34 (ii) Nucleus-Independent Chemical Shift (NICS)35 was obtained from the NMR calculations at the B3LYP/ 6-311+G(d,p) level using the GIAO formalism.54 The geometric center of the aromatic ring was the reference point for the calculation, and the resulting parameter (negative value of the absolute shielding) will be further denoted as NICS(0) to differentiate it from other possible definitions, in which the reference point is raised above the ring plane. (iii)AIM-derived parameter Θ introduced by Matta and Hernandez-Trujillo.36 The use of delocalization indexes for description of aromaticity was pioneered by Poater et al.55 and resulted in the introduction of new parameters such as the aromatic fluctuation index FLU.56 The Θ used in the current study is calculated according to the formula: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi n c ð2Þ ðδ0 δi Þ2 Θ ¼ 1 n i¼1

2. COMPUTATIONAL METHODOLOGY 2.1. Static DFT Models. Schematic presentation of the investigated compounds with atom numbering scheme is given in Figure 1. The calculations were performed in a framework of density functional theory (DFT).39,40 Geometry minimizations and subsequent harmonic frequency calculations were carried out confirming that the obtained structures correspond with the minima on the potential energy surface (PES). The three parameter functional proposed by Becke41 with correlation energy according to the LeeYangParr formula,42 denoted as B3LYP, and 6-311+G(d,p) triple-ζ split valence basis set43 were employed for this purpose. Next, the reaction path of the bridged hydrogen (H15) atom was scanned (with 0.1 Å increment of the O10H15 distance) and a series of distorted structures were generated, providing also KohnSham determinants for further AIM analysis. The distorted geometries were obtained as a result of the geometry optimization with added constraint of the fixed distance only between O10H15 atoms. The remaining part of the studied structure was left without any constraints. This methodology was successfully applied to study H-bonding properties in Schiff bases44 or ortho-hydroxyaryl ketimines.45 NBO32 and AIM15 analyses were performed to describe the electronic structure changes upon the O10 H15 3 3 3 O13 bridge deformation. Atomic charge evolution was analyzed based on both schemes, while the AIM theory was, additionally, used to investigate changes of properties at the Bond and Ring Critical Points (BCPs and RCPs): the electron density and its Laplacian, as well as the kinetic (GCP), potential (VCP) and total energies (HCP)—parameters determining bonding properties.13,14,46,47 At this point, it is necessary to point out that the AIM formalism should be used at the equilibrium geometry, so that the standard nomenclature related to Bond and Ring Critical Points (BCPs and RCPs) could be applied. In the present study, nonequilibrated structures are investigated and the electronic and topological properties cannot be strictly related to the terminology used in the classical AIM theory. However, we will label the located critical points as BCPs and RCPs following the convention frequently used in numerous studies of nonequilibrium structures where AIM was employed48,49 to facilitate the results discussion. The AIM analysis was carried out using the AIMPAC package50 while the quantum-mechanical

∑

where n = 6 (number of the ring carbon atoms), c is a constant chosen to obtain Θ = 0 for cyclohexane (at our level of theory, c = 2.4888, close to the value of 2.4312 of the original papers36,37 obtained at the HF/6-31G** level). Finally, δi is total delocalization index of an i-th carbon atom, calculated as a sum of its individual delocalization indexes δ(A,B) with the remaining carbon atoms of the aromatic ring. Each delocalization index δ(A,B) is, in turn, formed by overlap integrals of orbitals Sij over the AIM basins of atoms A and B: δðA, BÞ ¼ 4

∑i, j Sij ðAÞSijðBÞ

ð3Þ

The reference value of δ0 is obtained from the structure of benzene, and is 3.0279 (as compared to 3.0170 of the original papers).36,37 Calculation of the atomic overlap matrix elements was carried out using the AIMPAC package,50 while further computations of Θ utilized the AIMDELOC program.57 2.2. Resonance Effect Study. To support our study of the substituents influence on the intramolecular H-bond and aromatic ring properties, the resonance effect introduced by a series of other substituents was investigated (see Scheme 1). HOMA, 462



ARTICLE

Scheme 1. Structures of the Six Additional 2-Hydroxybenzaldehydes Used to Support HOMA and NICS Analyses Concerning the Resonance Effect, Forming Together with the Compounds IV the Extended Set of Studied Moleculesa

Table 1. Relative Energies of the Planar Conformers of III and Va III compound

E

V E+ZPE

E

E+ZPE

closed, Figure 1

0.00

0.00

0.00

0.00

closed, inverted

0.45

0.40

0.10

0.06

0.00 0.62

0.00 0.56

0.00 0.26

0.00 0.25

open, Figure 1 open, inverted

“Open” and “closed” form refer to the absence and presence of the intramolecular hydrogen bond respectively, governed by rotation of the primary OH group (adjacent to the aldehyde moiety). The “inverted” forms refer to the position of the second OH group (at C4 in III and C5 in V): the position used further in the study is shown in Figure 1, while the “inverted” form has the hydrogen rotated 180° around the CO bond axis. All energies in kcal/mol, relative to the conformer of Figure 1. E electronic energy, E+ZPE electronic energy corrected for the vibrational zero point level.

a

a

Dotted line indicates an intramolecular hydrogen bond.

NICS(0), and Θ indexes were calculated for the molecular forms of these additional 2-hydroxybenzaldehydes. The geometry optimization of the studied structures was performed according to the scheme described above, and further calculations of the indexes were carried out in the manner described in the previous section. 2.3. CarParrinello Molecular Dynamics. The dynamical nature of the intramolecular hydrogen bond (O10H15...O13) was investigated using CarParrinello molecular dynamics.38 Energy minimization and molecular dynamics in vacuo were performed in cubic cells of a = 12 Å for I, a = 14.5 Å for II, a = 14 Å for III and IV, and a = 12.5 Å for V. Energy minimization was performed using initial Hessian matrix proposed by Schlegel.58 The Hockney’s scheme59 was applied to remove interactions with periodic images of the cell. The exchange correlation functional proposed by J. P. Perdew, K. Burke, and M. Ernzerhof (PBE)60 coupled with the plane-wave basis set was applied. A kinetic energy cutoff of 100 Ry was used in the current study. The pseudopotentials by N. Troullier and J. L. Martins61 were employed for each type of atom in the studied ortho-hydroxybenzaldehydes. This part of simulations was performed to prepare initial conditions for further CPMD runs. Then, molecular dynamics simulations were performed at 298 K temperature, controlled during the simulations by Nose-Hoover thermostat chains,6264 coupled separately to each degree of freedom. The fictitious electron mass was set to 400 au. The time-step value applied was 3 au. Initially, the systems were equilibrated (the initial 2000 steps were excluded from the trajectory and were not taken into account during the data analyses). The data were collected for 16 ps for I and V and for 15 ps for II, III, and IV. This part of simulations was performed in a framework of the CPMD 3.11.1 program.65 The graphical presentation of results was prepared using the VMD66 and Gnuplot67 programs.

our study, σ+ (OH) = 0.92 and σ (NO2) = 1.27.69 The same picture is shown for gas-phase substituent constants of the same nature, i.e., σ+gas(OH) = 0.38 and σgas(NO2) = 0.18.69 This indicates that each of them is relatively strong in its own category. We note that we are not looking for the strongest possible substituent effects (e.g., in this respect the NEt2 group is a much stronger electron-acceptor with its σ+ (NEt2) = 2.07),69 and we are not aiming at covering the whole range of Taft or Hammett constants, which would be the case if we were trying to carry out quantitative correlation analysis of substituent effects. Our aim is using rather small substituents with opposite signs but comparable absolute magnitudes of the σ constants. Thus, apart from the main issue of the proton position influence, the current study describes differentiation of intramolecular interactions due to the kind and position of the substituent. The molecular properties of a set of ortho-hydroxybenzaldehydes containing OH 3 3 3 O intramolecular hydrogen bond are discussed on the basis of various theoretical approaches. Let us start with discussion results from the conformational behavior of the hydroxyl groups. The five studied structures in their hydrogen-bonded, “closed”, form are presented in Figure 1. However, “open” forms are also possible, but they are less stable—they lack the stabilizing intramolecular hydrogen bond, therefore we will focus our discussion on more stable “closed” forms. There remains, however, the problem of position of the second OH moiety in the compounds III and V (located at the C4 and C5 carbon atoms respectively, using the numbering scheme of Figure 1). Further, we will study the conformers of Figure 1, where the proton of the second OH group is directed toward the aldehyde proton, not toward the first OH group. Right now, we will consider also the “inverted” conformers (rotamers), in which the OH group, not participating in the hydrogen bond, is rotated 180° with respect to the conformers shown in Figure 1. Table 1 presents the relative energies of the “inverted” conformers for III and V with respect to the conformers of Figure 1. It shows that the forms of Figure 1 are always more stable, except for the closed form of III. However, the energy difference is very small and the choice of the conformer does not seem to impact the investigated properties. Therefore, for the sake of uniformity between various structures, we decided to select the slightly higher energetically form of “closed” III shown in Figure 1, for further studies. Additional data potential

3. RESULTS AND DISCUSSION 3.1. Static DFT Models. The general aim of the study, as outlined in the Introduction, is investigation of the interplay between bridge proton position and molecular properties, but additionally the introduction of substituent effects was planned as a secondary aim. The selection of the OH and NO2 groups as substituents in the aromatic ring (see Figure 1) was dictated by their opposite behavior in terms of, e.g., electron-withdrawing/ attracting properties. These may be best presented by values of substituent constants68 σ+ for electron donating groups, and, σ for electron attracting ones. For the functional groups applied in 463



ARTICLE

not extremely high (ca. 910 kcal/mol). The potential energy profiles for compounds III and IV lie lower than those for compound I (unsubstituted phenyl ring) and compound II (paranitro derivative with respect to the aldehyde moiety), which indicates the substituent influence on the H-bonding properties. Interestingly, the curve for V initially follows that of I,II, but later joins the lower-lying profile of III,IV. The lower panel of the Figure 2 shows a more detailed analysis of selected interatomic distances involved in the intramolecular H-bonding and further quasi-ring formation for I. The theoretical model shows interesting events; when the distance between O10O13 is shortest and O10H15 equals 1.28 Å, the C9O13 and C2O10 distances are equalized and their chemical role is exchanged there is a change in the dominant Lewis structure. The H15O13 distance at this point equals 1.165 Å, indicating that the bridged proton reached the acceptor side. Further, the dependence of the donor (O10) acceptor (O13) separation on the donor (O10) proton (H15) distance (Figure 3SI of the Supporting Information) reveals that the minimal distance between donor and acceptor was reached for r(O10H15) = 1.28 Å. This corresponds to the proton position in the middle of the hydrogen bridge in all studied cases (H15 3 3 3 O13 is equal 1.161.165 Å, thus the proton is already caught by the acceptor at this point). This fact of strong coupling between bridge length and the bridged proton position is very characteristic for the strong hydrogen bonding. Next, the relationship between the C1C9 bond length and the O10H15 distance was studied. In all cases we observed (see Figure 4SI of the Supporting Information) the same trend: when the O10H15 distance was elongated, the C1C9 bond length was shortened in accordance with growing domination of the quinonoid Lewis structure. It is well-known that a formation of a hydrogen bond leads to weakening—and elongation—of the donor-proton covalent bond. Experimental correlation between these parameters was observed in the most studied OH 3 3 3 O cases as far back as in 1955.73 More recently, among other studies, Steiner74 proposed an improved version of an exponential relationship between OH and H 3 3 3 O, based on low-temperature neutron diffraction data. We decided to test this valence model of the hydrogen bond in the situation of an intramolecular bridge. The specific fit to the eq 1 of ref 74 which we will use in the current work, (r0 = 0.934 Å, b = 0.388 Å) was derived with exclusion of the water contacts, which are more flexible and introduce larger deviations to the model. The results presented in Figure 3 suggest that the static calculations at the employed DFT level of theory follow the assumed valence model quite well. The deviations can be explained by (a) different equilibrium value of the OH bond length in the DFT framework, (b) intramolecular character of the investigated hydrogen bond, which puts additional constraints on the molecular structure. The last remark is also valid for the experimental data on ortho-hydroxybenzaldehydes,71,72 included in the Figure 3, and another factor leading to deviations is the use of X-ray crystallography in these two studies—this has influence on the positions of the located protons. The next part of the current study is devoted to the electronic structure description. Figure 4 presents changes of the AIM net charges of the atoms involved in the intramolecular H-bond. Figure 4A shows the dependence of the donor O10 net atomic charge on the rO10-H15 elongation. For the equilibrium structures, the value of the O10 net charge in all the studied structures is between 1.114 ÷ 1.124. At the beginning, together with elongation of the O10H15 distance, we observe a sharp

Figure 2. Dependence of energy (upper panel, for all compounds) and selected metric parameters (interatomic distances for the structure I, lower panel) on the donor-proton O10H15 distance. Atom numbering scheme is explained in Figure 1.

energy profiles for the rotation of the second hydroxyl group— are presented in Figure 1SI of the Supporting Information. Interestingly, the rotational barrier is strongly lowered in V, with regards to III. This may be explained by a much stronger resonance interaction of OH group with CHO from position 4 than from position 3 —a well-known weaker substituent effect from the meta rather than from para position.70 However, taking into account the flexibility of the proton, enhanced by the tunneling phenomena, we can assume that in both compounds the rotation of the second, non-hydrogen-bonded, hydroxyl group, is almost free. This adds strength to our freedom of choice of the most similar rotamers for the purpose of our study. Selected computed geometric parameters (describing intramolecular H-bonding) are presented in Tables 15SI of the Supporting Information. Unfortunately, the chemical nature of some of the studied compounds makes them difficult to obtain in a monocrystalline form, leaving us unable to make a comparison of calculated parameters with experimental metric parameters. This was possible only for compounds III and V. As a reference, using the X-ray data for these compounds71,72 we can conclude that the applied level of theory was able to reproduce correctly their molecular structure. Next, to get a more general picture of the bridged (H15) proton dynamics on the basis of DFT, potential energy profiles were calculated for all of the investigated compounds, and they are presented in the upper panel of Figure 2. On the one hand, it is shown that a spontaneous proton transfer is rather not preferable, but on the other hand the energy necessary for moving the proton to the acceptor side is 464



ARTICLE

Figure 3. Dependence of the donor-proton bond length on the protonacceptor distance. Results of calculations and available experimental data in comparison to the model proposed by Steiner.74

increase in the value and subsequent gradual decrease. The substituent influence on the observed changes is well pronounced. Both nitro derivatives II and IV exhibit a smaller decrease in the O10 net charge comparing with ortho-hydroxybenzaldehyde and its analogues (III and V, respectively). Figure 4B shows an atomic charge evolution of the bridged hydrogen (H15) versus elongation of the O10H15 distance. The net charge of H15, for IV when the structures are at minima, is between 0.630 and 0.623. A similar value of the net charge is reached when the proton is at the midpoint between the donor and the acceptor. Upon the elongation of the O10H15 bond, at the beginning the value of the H15 net charges is decreased, but later we observe its increase up to 1.7 Å, and further again the value is decreased. The same tendency is observed for all studied compounds. Concerning the substituents in the phenyl ring, we can draw the same conclusions as described above. Figure 4C shows dependences obtained for the O13 acceptor atom, and notably, at the beginning of the O10H15 bond length elongation, the value of the O13 net charge is increased. In contrast to the previously discussed donor and proton net charges, only when the bond length of O10H15 exceeds ∼1.2 Å do we observe a rapid decrease of the O13 value. At this turning point of r(O10H15) = 1.2 Å, the proton is still dominantly at the donor side, but seemingly it already starts being shared by the O13 acceptor and influences its properties. Further, we proceeded with the analysis of chemical bonding. Figure 5 shows the dependence of the electron density and its Laplacian at selected bond critical points (BCPs) versus the changes in the O10H15 bond length. The electron density and its Laplacian calculated at BCPs are local molecular properties determining the character of relevant chemical bonds—in this case our attention will be focused on the intramolecular hydrogen bonds. A set of criteria to be fulfilled by such a hydrogen bond was proposed by Koch and Popelier.14 The first criterion is associated with the interaction paths joining nuclei with the respective BCPs, therefore it was necessary to check if BCPs and respective connecting paths between O10 H15 and O13 atoms exist. The paths were found in all of the studied structures and for the whole range of the investigated interatomic distances. When the bond length O10H15 was elongated, the electron density was decreased at the O10H15 BCP. An opposite situation was observed for the H15...O13 BCP, so that the two curves of the density dependence formed an

Figure 4. Dependence of the donor oxygen atom (O10, panel A), bridged proton (H15, panel B) and acceptor oxygen atom (O13, panel C) AIM net atomic charges on the donorproton distance for the studied structures.

almost symmetric pair. The electron density Laplacian at selected BCPs also shows an expected tendency. Gradual change from strongly negative (covalent O10H15 bond) to positive Laplacian is matched by the opposite trend for the H15...O13 BCP. It is worth mentioning that the presence of the substituents in the aromatic ring does not influence the pattern of changes of 465



ARTICLE

Figure 5. Dependence of the electron density (upper) and its Laplacian (lower) at selected BCPs on the donorproton distance for the studied structures. Black donor-proton O10H15 BCP, red acceptorproton O13H15 BCP.

Figure 6. Dependence of the electron density (upper) and its Laplacian (lower) at selected covalent BCPs on the donorproton distance for the studied structures. Black C2O10 BCP, red C9O13 BCP.

consistent with changes in the dominant Lewis structure. The electron density Laplacian (see Figure 6, lower part) shows an interesting tendency at BCPs of C2O10 and C9dO13 covalent bonds. The value of the Laplacian at the BCPs initially decreases, and further smoothly grows when the O10H15 bond is elongated. The numerical value of the Laplacian indicates a covalent character of the discussed bonds. Continuing our discussion of electron density topology, we also monitored selected ring critical points (RCPs). The obtained results are presented in Figure 5SI of the Supporting Information, respectively. Let us analyze first the aromatic ring (lower panel). We observe the decrease in the electron density for all studied compounds as a result of the bond length O10H15 elongation. A very similar picture was obtained for the electron density Laplacian. The influence on the substituents is visible. The upper part of the graph presenting the electron density versus O10H15 is dominated by the ortho-hydroxybenzaldehydes substituted by the NO2 group, while in the middle we can observe the data for the unsubstituted I, and the lower part of the graph is occupied by the results obtained for dihydroxybenzaldehydes III and V. This dependence shows different roles of the hydroxy and nitro groups as electron-withdrawing and -donating substituents, respectively. We can conclude that the

electron density and its Laplacian at the bridge BCPs. The criteria for the H-bonding require that the relevant BCPs fulfill further the following conditions: (i) The electron density at the BCP at the hydrogen bridge is in the range of 0.002 ÷ 0.034 au—second Popelier’s criterion; (ii) The electron density Laplacian is between 0.024 ÷ 0.139 au—third Popelier’s criterion.14 The second Popelier’s criterion is fulfilled concerning the electron density at O10H15 and H15...O13 BCPs respectively. The third Popelier’s criterion for the H-bonding is not fully fulfilled in the studied cases. The negative value of the electron density Laplacian indicates a covalent character of the interaction. In our case, evolution of the electron density Laplacian indicates that the intramolecular H-bond has a partially covalent character depending on the position of H15. Further, the electron density at the BCPs of covalent bonds C2O10 and C9dO13 changes due to the elongation of the O10H15 bond length (see Figure 6, upper part). When the O10H15 bond is elongated, we observe an increase in the electron density at the BCP of the C2O10 covalent bond, but the decrease in the electron density at the BCP of the C9dO13 bond. This is 466



ARTICLE

electron density and its Laplacian in the aromatic ring, related to the aromatic character of the ring, are strongly influenced by the position of the H15 bridge proton. Further, lower panel of Figure 5SI of the Supporting Information presents data obtained for a quasi-ring formed by the atoms involved in the intramolecular hydrogen bond (O10H15...O13) with neighboring atoms (C9, C1, and C2). The electron density and its Laplacian show a very similar tendency: when the distance between O10H15 is elongated up to ca. 1.3 Å we observe an increase in both parameters. Further elongation in the bond length results with the value decrease for both parameters. Comparing behavior of the phenyl and quasi-aromatic RCPs, and anticipating the HOMA aromaticity study below, we can conclude that the dependence of the electron density and its Laplacian at the phenyl ring RCP is correlated with the ring aromaticity degree. However, the quasi-aromatic ring is more sensitive to the proton position, which is also connected with the possibility of electronic coupling. This possibility is highest for the proton at the midpoint, and this fact explains positions of the RCP electron density and Laplacian maxima in the Figure 5SI of the Supporting Information. Discussing further the molecular properties related to BCPs, the kinetic electron energy density (GCP), the potential electron energy density (VCP) and the total electron energy density (HCP) are worth analyzing. These parameters give deeper insight into the character of H-bonding present in the studied ortho-hydroxybenzaldehydes. The data obtained for I is presented in Figure 7. We observe that for the both BCPs, the kinetic electron energy density (GCP) has positive values and the values change in a similar way. At BCP related to the O10H15 bond, the potential electron energy density (VCP) and the total electron energy density (HCP) increase due to the bond length elongation. The opposite situation took place in the case of the BCP at H15...O13 bond—both energies decreased while the O10H15 bond length was elongated. Let us pay attention for a moment to changes in the energy density at BCPs for covalent bonds— C2O10 and C9dO13 (see Figure 7, bottom). Changes in the energy components upon elongation of the O10H15 bond length are monotonic. The kinetic electron energy density (GCP) at BCP related to the C2O10 increases, while at BCP of C9dO13 it decreases. The potential electron energy density (VCP) at BCP of C2O10 decreases while it is increased at the BCP of C9dO13. The value of the total electron energy density (HCP) decreases at BCP of C2O10 while its value is increased monotonically at the BCP related to the C9dO13 covalent bond. Interestingly, the crossing point at which the CO bonds exchange their roles is shifted to ca. 1.4 Å, while the “proton at the midpoint” event is at ca. 1.21.3 Å. This shift is especially visible when comparing Figures 5 with 6, and upper vs lower panel of Figure 7. To supplement our discussion of the electronic structure modulation by the proton position and substituents, Natural Bond Orbitals (NBO) method was employed to compute the net charges for the set of studied molecules. This population analysis method has advantages over conventional schemes, e.g., Mulliken analysis, because the orbitals are first transformed into a unique, orthonormal set of Natural Atomic Orbitals, reducing many problems of the Mulliken scheme (negative orbital occupations, large basis set dependence).75 It will be valuable to compare its description of the hydrogen bridge atoms with the results of the AIM. Figure 8 presents the dependence of the O10, H15, and O13 net charges on the H15...O13 bond length

Figure 7. Dependence of the kinetic (G), potential (V), and total (H) energy densities at selected BCPs on the donor-proton distance for the structure I. Upper panel: Black donor-proton O10H15 BCP, red acceptor-proton O13H15 BCP. Lower panel: Blue C2O10 BCP, green C9O13 BCP.

evolution. Figure 8A shows that the changes in the net charge of O10 are modulated not only by the bond length elongation, but also by the presence of substituents in the phenyl ring. The same tendency is observed for all studied cases; first, the net charge grows more negative, but depending on the substituent and the proton position in the hydrogen bridge, finally it starts to become less negative. This is opposite behavior than that of the AIM values (Figure 4A). Figure 8B shows a dependency obtained for the bridged proton, and it is similar to the trend reproduced by the AIM method in the Figure 4B. When the O10H15 and H15...O13 bond lengths are equalized at ca. 1.3 Å, the H15 net charge has the lowest value. The left part of the figure shows that the net charge becomes more positive when the bridged proton comes close to the acceptor atom. Figure 8C shows how the net charge of O13 reacts when the H15...O13 bond length is elongated. When the proton-acceptor distance is shortest, the net charge of O10 is most negative, and when the H15 comes further from O13 and closer to the donor, its net charge becomes less negative, in agreement with the AIM-based Figure 4C. Interestingly, the absolute values of the NBO charges of both oxygen atoms are almost two times smaller than the AIM 467



ARTICLE

Figure 8. Dependence of the donor oxygen atom (O10, panel A), bridged proton (H15, panel B) and acceptor oxygen atom (O13, panel C) NBO net atomic charges on the proton-acceptor distance for the studied structures.

results, but this effect is not as much pronounced for the bridge proton H15. The comparison shows, however, that the qualitative trends are reproduced by the two population analysis methods in a very similar manner. The last point of the analysis of the static models is devoted to the issue of aromaticity evolution upon the bridge proton position and the influence of substituent. There are numerous indexes describing the degree of aromaticity in a given structure. The three used in the current study employ totally different strategies to achieve this goal, and it will be valuable to compare their performance. The Harmonic Oscillator Model of Aromaticity (HOMA)33,34 is based on the changes in the structure of the molecular skeleton and associated deviations from the idealized optimum values. NICS35 is a direct measure of the shielding in a well-defined position in the middle of the aromatic ring, and can be related to the concept of aromatic and antiaromatic ring currents. Finally, the AIM-based parameter Θ reflects electron delocalization on the basis of the atomic overlap integrals36 and is closely related to other proposals employing Bader’s partitioning of molecular space into atomic basins, such as para-delocalization index PDI55 or aromatic fluctuation parameter FLU.56

Figure 9 shows dependence of the three studied aromaticity indexes on the H15...O13 distance for IV. The geometry-based HOMA parameter will be analyzed first. Scheme 1SI of the Supporting Information shows nomenclature of the bonds in the aromatic ring for the HOMA calculations, whereas the relevant bond lengths computed at B3LYP/6-311+G(d,p) level of theory as well as the numerical values of the HOMA index are presented in Tables 6SI-10SI of the Supporting Information. We observe that the aromatic character of the phenyl ring depends strongly on the H15 position. When the bridged proton moves closer to the acceptor, the HOMA index value decreases and is close to 0 for the fully transferred proton. Interestingly, the “proton in the middle” scenario at ca. 1.3 Å is marked by only a small loss of aromaticity, HOMA = 0.8. We can conclude that, when the acceptor has finally attracted the proton, the phenyl ring “loses” its aromatic character due to the electronic structure deformation leading to distortion of geometric parameters. This is a sign of change from a two-structure aromatic resonance description of the molecule in equilibrium to a localized quinoid-like Lewis structure. Such resonance effects are associated with the existence of resonance-assisted hydrogen bonds (RAHB). Our study 468



ARTICLE

Figure 9. Plot of selected aromaticity parameters versus proton-acceptor (H15...O13) distance evolution for IV. Upper left: HOMA index, upper right: AIM-derived aromaticity parameter Θ, lower: NICS(0) parameter (negative value of the absolute shielding at the center of the aromatic ring).

does not show uniform, monotonic influence of the position and kind of substituents in the phenyl ring on the aromaticity described by the HOMA index. This fact will be further expanded and discussed in the next section. Both NICS(0) and Θ also show the loss of aromaticity (see Figure 9), which is slow until the proton reaches the center of the bridge, and then, the loss accelerates when the H15...O13 distance decreases further. It is worth noticing that both of these aromaticity indexes, which are based on electronic structure, discern between the substituents of the aromatic ring better than the HOMA. However, they provide opposite signs of the effect of substituent introduction. While substituted orthohydroxybenzaldehydes IIV are more aromatic than the parent salicylaldehyde I according to the lower NICS(0) values, they are less aromatic according to the parameter Θ. This index also exhibits smaller range of variability. While HOMA falls from 0.93 to ca. 0.2 upon the change of the proton position, the Θ falls only from 0.8 to ca. 0.56. This fact is associated with different scaling of these parameters. The zero value on the scale of HOMA corresponds to the nonaromatic Kekule structure of benzene, while Θ = 0 is achieved for cyclohexane. The delocalization indexes δ(A,B) used during calculation of Θ are indicators of the number of electrons delocalized or shared between the atoms A and B, thus they are related to the concept of bond orders.76

The aromaticity indexes tend to exhibit good intercorrelations,77 not only for “classical” six-membered aromatic rings but also for fulvenes and heptafulvenes,78,79 and the most useful relationships between HOMA, NICS(0), Θ, and electron. density parameters at the aromatic RCP are summarized in Figure 10. Further data in the Supporting Information (Figure 6SI) include significantly nonlinear dependences. The plots of Figure 10 are also in most cases slightly nonlinear, with exception of the NICS(0) vs Laplacian of F(RCP). For a given compound, the correlation coefficient for this particular dependency is r2 = 0.9995. Even if we take into account all of the compounds at once and carry out least-squares linear fitting for the total scatter plot, then we still obtain r2 = 0.850. Such excellent correlation is also found for the NICS(0) vs F(RCP) (Figure 6SI of the Supporting Information), and should not be surprising, since the NMR shielding at a specific point depends primarily on the electron density in the immediate vicinity of that point. Finally, enforcing linear relationship in cases of slight nonlinearity still provides reasonable models; the r2 coefficient for NICS(0) HOMA is 0.933, for HOMA Θ it is 0.942, and for NICS(0) Θ it is 0.881 (with all compounds included in the correlation). The correlation scatter plots show also once more how the substituent effect affects some parameters. The upper right panel of Figure 10, HOMA vs F(RCP), is particularly worth mentioning: the parabola for the parent compound I separates very well the two sets of 469



ARTICLE

Figure 10. Selected correlation charts for the aromaticity and electronic structure parameters of the investigated compounds. Upper left: NICS(0) vs HOMA; upper right: electron density at the aromatic ring critical point vs HOMA; lower left: electron density Laplacian at the aromatic RCP vs NICS(0); lower right: Θ aromaticity index vs electron density at the aromatic RCP.

Hammett-type σp+ and σp constants into inductive and resonance components. One of interesting alternative approaches is the use of gas-phase equilibria, instead of solvent-based data, to derive the σR+ and σR‑ constants.81 In the following text, we will report both types of resonance parameters, but with more emphasis on the R+/R values, for the three main reasons: they are more commonly used than the gas-phase σR+/σR constants, they have been determined for a broader set of substituents (compare Tables V and IX of the ref 69), and, finally, both types of parameters are relatively well correlated. The extended set of substituents comprises not only the OH and NO2 groups used in the previous Section, but also NH2, -F and CN. Possible substitutions at 4- or 5- positions (para with respect to either formyl or hydroxyl group, respectively) lead, apart from the initial set of molecules IV, to the compounds VIXI (see Scheme 1). The substituents cover wide range of resonance capabilities: from strong and moderate π-donors (NH2, OH, F with R+ values of 1.38, 1.25, and 0.52 respectively) to π-acceptors (CN, NO2, with corresponding R values of 0.49 and 0.62).69 The substituent coefficients derived from the gas-phase data are, in the given order of substituents: σR+ = 0.52, 0.38, 0.25; σR = 0.10 and 0.18.69 Thus, investigation of the expended set of compounds IXI will allow estimation of the influence of various magnitudes of resonance effects on the O10H15... O13 intramolecular hydrogen bond and aromaticity of the phenyl ring.

curves: nitro compounds II and IV above, and the dihydroxy derivatives III and V below. This is clear visualization of the coupling between geometric and electronic consequences of substituent effect imposed by electron-donating and electronwithdrawing groups. 3.2. Resonance Effect Study Static DFT Models. Detailed study of the proton position in the intramolecular hydrogen bond as a modifying factor of the molecular properties of the compounds IV was the general aim of the previous section. However, taking into account larger series of substituents, and presenting their modulating properties, will put the results discussed above in a broader perspective. Therefore, the extended set of substituents was taken into account (see Scheme 1) to cover as broad range of the resonance effects as possible, according to the literature values of the resonance parameters.69 Experimental work of Nagaoka et al.80 on the excited-state intramolecular proton transfer in ortho-hydroxybenzaldehyde derivatives shows that there is clear correlation between the fluorescence quantum yield Φf and the Hammett-type YukawaTsuno constant σπ. The yield Φf is, on the other hand, proportional to the energy gap between the S1(π) and S0 states (the adiabatic transition energy). Presence of such a correlation related to the intramolecular hydrogen bond prompted us to carry out computational investigations of the impact of resonance capabilities of the substituents on the salicylaldehyde family. Resonance effect of a given substituent is usually described by using R+ or R parameters,69 which are obtained by factorization of 470



ARTICLE

Figure 11. Charts representing dependence of the selected parameters (donorproton and donoracceptor distance) of the intramolecular hydrogen bond of the extended set of compounds IXI on the resonance constants R+/R of the additional substituents at position 4 (black) or 5 (red), see Scheme 1.

Figure 11 presents dependence of the chosen metric parameters of the O10H15...O13 bridge on the resonance constants of the substituents (extended data are listed in the Supporting Information, Table 11SI). There are several interesting features of these relationships. First of all, the influence of the π-acceptor substituents located para- with respect to the hydrogen bond acceptor, the formyl group, is negligible (compounds II and VIII). If such a substituent is located para- to the hydrogen bond donor, O10-H15 group, the bridge is visibly strengthened (compounds IV and IX). However, a π-donating group acts in the opposite way: it strengthens the bridge if it is located parato the acceptor (compounds III, VI, and VII), and weakens the bridge if para- to the donor (V, IX, X). Remarkably, the strengthening effect is larger than the weakening one for the same substituent. This is visible within a pair of locations for a given substituent, e.g., for III and V; the O10H15 donor bond is elongated by 0.0021 Å for III, and shortened by half of that value, 0.001 Å, in V. The bridge itself is shortened by 0.0064 Å in III, but elongated by only 0.0043 Å in V (all data with respect to the parent unsubstituted salicylaldehyde I). The opposite action of the π-donors is aided by the fact that the bridge-forming groups are of different type themselves: while the hydrogen-bonding donor O10H15 is a π-donor with R+ = 1.25, the acceptor O15 belongs to the π-accepting formyl group with R = 0.70. Therefore, the cases with π-donor and π-acceptor located in para- positions with respect to each other, and one of these substituents is acting as a hydrogen-bonding donor or acceptor, provide optimal conditions for modulation (strengthening) of the intramolecular hydrogen bond. In the case of this study, these

Figure 12. Charts representing dependence of the aromaticity indexes HOMA, NICS(0) and Θ of the extended set of compounds IXI on the resonance constants R+/R of the additional substituents at position 4 (black) or 5 (red), see Scheme 1.

are compounds III, IV, VI, and VII. Our study also shows that the tuning of the intramolecular bridge properties can be efficiently carried out by affecting either the donor or the acceptor group. The study of the extended set of molecules, carried out for the optimal structures only, validates also our choice of the compounds IV for the detailed investigation of the influence of bridge proton position on the molecular properties, performed in the previous Section: the compounds IIV span wide range of substituent effects. However, overall picture is that the substituents were not able to change the behavior of the H15 bridge proton in a qualitative way. Figure 12 describes how the substituents modify aromaticity of the studied compounds (additional data in Table 12SI of the Supporting Information). Here, the relationships are usually nonmonotonic, sometimes even erratic, and this can be explained after careful examination of the physical meaning of the employed 471



ARTICLE

Figure 13. Time evolution of interatomic distances of atoms involved in the intramolecular hydrogen bond (O10H15...O13) for IV.

ring. In such case, the resonance parameters R+/R are not the only decisive factors; the inductive component of the substituent effect is also responsible for the alterations in the electron density within the ring. The remaining aromaticity indexes, HOMA and Θ, show more regular dependencies, which are also similar in both cases. The most important finding is that introduction of a

aromaticity indexes. The most erratic behavior is observed for the NICS(0) parameter (middle panel of Figure 12), which corresponds to the absolute NMR shielding at the center of the phenyl ring. This value is, on the one hand, associated with the presence of the aromatic ring currents, but on the other hand it also depends on the electron density injected or withdrawn from the 472



ARTICLE

π-donor in position 4- (compounds III, VI, VII) leads to the loss of aromaticity, correlated with the resonance strength of the substituent. This fact agrees with our earlier discussion on relative alignment of the π-donor and π-acceptor groups. When they are opposite each other, as in III, VI, VII, the π-electron transfer is at maximum, which results in enhanced presence of the localized quinoid-like Lewis structure, and associated loss of aromatic character. 3.3. CarParrinello Molecular Dynamics. Considering that the potential energy profiles from the static models show the energy scale of ca. 910 kcal/mol for the donoracceptor reaction path, we decided to proceed further with ab initio molecular dynamics models for IV to investigate the time evolution of metric parameters related to the intramolecular hydrogen bond. Figure 13 shows the evolution of O10H15...O13 bond length as a function of time. The bridged proton (H15) is located at the donor (O10) side, in agreement with our design based on the pKa differences between the donor and acceptor groups. In all studied ortho-hydroxybenzaldehydes, the bridged proton does not exhibit even short contacts with the acceptor O13 atom, which, as we expected, is not based on the PES profiles from static DFT. The equilibrium interatomic distance between O10...O13 equals ca. 2.63 Å in all studied cases, therefore we could expect a strong mobility of the bridged proton. However, our previous ab initio molecular dynamics studies of OH...O intramolecular hydrogen bonds showed that this kind of interaction is easily modified by inductive and steric effects31,82,83 and it is very difficult to predict the H-bonding properties and make general conclusions. In the next step, we turned our attention to the dihedral angles of the quasi-ring formation in the studied IV compounds. The obtained results are presented in Figures 7SI-10SI of the Supporting Information. The dihedral angle O13C9C1C2 shows flexibility which results in deviation of ca. 30° from the equilibrium position during the MD run. Such a flexibility of the dihedral angle makes proton transfer or even short contacts of the bridged proton (H15) with an acceptor O13 atom not favorable. Next, the analysis of C1C2O10H15 dihedral angle was performed. The rotation of the dihedral angle is not constrained, which also results with flexibility. The third analyzed dihedral angle contains C2O10H15...O13 atoms. This dihedral angle exhibits even larger fluctuations and flexibility than the discussed above O13C9C1C2 and C1C2O10H15 dihedral angles. A very similar dependency was obtained as well for the O10H15...O13C9 dihedral angle. These results showed that the flexibility and free rotation of the functional groups involved in the intramolecular H-bonding do not allow many processes to occur therefore the H-bridge could be modified via the presence of substituents in the aldehyde moiety, which could introduce additional steric effects. Table 13SI of the Supporting Information presents comparison of selected interatomic distances between the optimized, static structures, and the CPMD results. The static and dynamic approaches yield very similar results for most of the covalent bonds, and the largest elongation of the bond in the CPMD is registered for the hydroxyl group (O10H15). This is caused not only by the large amplitude of motions of the light hydrogen atom, but also by the modulating influence of the hydrogen bond to the O13 oxygen atom.

4. CONCLUSIONS The present study was designed to show the extent of modifications in the structure of ortho-hydroxybenzaldehydes due to

variation of the proton position in the intramolecular hydrogen bridge. These phenomena are additionally influenced by substituents of different sign, but similar magnitude, of electron-donating/withdrawing effect. The studied series of ortho-hydroxybenzaldehydes were selected, on the basis of donoracceptor acidity differences, to make the proton transfer to the acceptor side rather improbable, but to allow the bridged proton to have a reasonable ability to penetrate temporarily into the acceptor region. This selection highlights the substituent effects, because they are not obscured by associated effects of structural changes resulting from the proton transfer. However, our study of enforced proton transfer reflects the impact of substituents on the properties of the intramolecular hydrogen bond. To summarize the analysis of static and dynamic models: (i) Proton potential functions, in agreement with our expectations, suggest that there is no minimum at the acceptor side, but the energy necessary to move the proton is not large, so it is possible for the proton to move into the acceptor region. This fact enabled us to proceed further with the structural analyses on the basis of geometric and electronic structure parameters; (ii) Regardless of the substituent position (para- with respect to either CHO or OH moiety), the nitro- and hydroxy- substituents provide separate series of data with different behavior as described by the AIM and NBO approaches; however, the location of the substituents seems less important than their type. Namely, the nitro compounds II and IV are similar to each other, and the hydroxy derivatives III and V form the other pair of analogues. There are fewer similarities when the compounds are grouped according to the position of the substituent (II and III, IV and V). (iii) The AIM data show smooth transition from hydrogen to covalent bonds in the bridge upon the proton migration to the acceptor; (iv) The HOMA index shows that change in proton position modulates the aromatic properties of the central phenyl ring (up to almost a total loss of aromaticity), and the influence of the substituents is subtle but in line with the classification found in the AIM and NBO studies; (v) The NICS(0) and Θ indexes of aromaticity show trends similar to those of HOMA, and correlations between various parameters are high (in agreement with the literature findings),77 even if the whole set of five compounds is simultaneously taken into account. Correlation scatter plots also reveal qualitative manifestations of the substituent effects; (vi) The study of the resonance effect in an extended set of molecules indicated that the modulation of the hydrogen bonding, although in a limited range, is best achieved by placing groups of opposite π-donor/π-acceptor properties (one of them being the bridge donor or acceptor) in para position; (vii) The CPMD approach revealed a large structural flexibility of the functional groups involved in the intramolecular H-bonding formation, which proves that such effects as resonance strengthening of the hydrogen bond are not strong in the studied compounds.

’ ASSOCIATED CONTENT

bS Supporting Information. Conformational profiles, detailed structural data along enforced proton transfer pathways 473



ARTICLE

(including bond lengths for the HOMA analysis and comparison to previous reports on crystal structures), additional AIM results, correlation scatter plots for aromaticity indexes, and CPMD time evolution of selected parameters. This material is available free of charge via the Internet at http://pubs.acs.org.

(20) Filarowski, A.; Kochel, A.; Koll, A.; Bator, G.; Mukherjee, S. J. Mol. Struct. 2006, 785, 7–13. (21) Filarowski, A.; Koll, A.; Sobczyk, L. Curr. Org. Chem. 2009, 13, 287–298. (22) Ahn, D.-S.; Park, S.-W.; Lee, S.; Kim, B. J. Phys. Chem. A 2003, 107, 131–139. (23) Krygowski, T. M.; Szatyzowicz, H.; Zachara, J. E. J. Chem. Inf. Model. 2005, 45, 652–656. (24) Krygowski, T. M.; Szatyzowicz, H.; Zachara, J. E. J. Chem. Inf. Comput. Sci. 2004, 44, 2077–2082. (25) Sanz, P.; Mo, O.; Ya~ nez, M.; Elguero, J. J. Phys. Chem. A 2007, 111, 3585–3591. (26) Majerz, I.; Koll, A. Pol. J. Chem. 1994, 68, 2109–2119. (27) Rozas, I. Phys. Chem. Chem. Phys. 2007, 9, 2782–2790. (28) Mariam, Y. H.; Musin, R. N. J. Phys. Chem. A 2008, 112, 134–145. (29) Palomar, J.; De Paz, J. L. G.; Catalan, J. J. Phys. Chem. A 2000, 104, 6453–6463. (30) Ríos, M. A.; Rodríguez, J. J. Comput. Chem. 1992, 13, 860–866. (31) Jezierska, A.; Panek, J. J.; Koll, A. Chem. Phys. Chem. 2008, 9, 39–846. (32) Weinhold, F.; Landis, C. R. Valency and Bonding. A Natural Bond Orbital DonorAcceptor Perspective; Cambridge University Press; Cambridge, 2005. (33) Kruszewski, J.; Krygowski, T. M. Tetrahedron Lett. 1972, 13, 3839–3842. (34) Krygowski, T. M.; Cyra nski, M. Chem. Rev. 2001, 101, 1385– 1419. (35) Schleyer, P.v-R.; Maerker, C.; Dransfeld, A.; Jiao, H.; van Eikema Hommes, N. J. R. J. Am. Chem. Soc. 1996, 118, 6317–6318. (36) Matta, C. F.; Hernandez-Trujillo, J. J. Phys. Chem. A 2003, 107, 7496–7504. (37) Matta, C. F.; Hernandez-Trujillo, J. J. Phys. Chem A 2005, 109, 10798. (38) Car, R.; Parrinello, M. Phys. Rev. Lett. 1985, 55, 2471–2474. (39) Hohenberg, P.; Kohn, W. Phys. Rev. 1964, 136, B864–B871. (40) Kohn, W.; Sham, L. J. Phys. Rev. 1965, 140, A1133–A1138. (41) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (42) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789. (43) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650–654. (44) Filarowski, A.; Koll, A.; Gzowiak, T. J. Chem. Soc., Perkin. Trans. 2 2002, 835–842. (45) Filarowski, A.; Koll, A.; Hansen, P. E.; Kluba, A. J. Phys. Chem. A 2008, 112, 3478–3485. (46) Popelier, P. L. A. Atoms in Molecules. An Introduction; PrenticeHall: Harlow, UK, 2000. (47) Bader, R. F. W. Phys. Rev. B 1994, 49, 13348–13355. (48) Galvez, O.; Gomez, P. C.; Pacios, L. F. J. Chem. Phys. 2001, 115, 11166–11184. (49) Pacios, L. F. Struct. Chem. 2005, 16, 223–241. (50) Bader, R. F. W. AIMPAC, Suite of programs for the Theory of Atoms in Molecules; McMaster University: Hamilton, Ontario, Canada, 1991. (51) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.; Vreven, Jr. T.; Kudin, K. N.; Burant, J. C., et al. Gaussian 03, Revision C.02;Gaussian, Inc.: Wallingford CT, 2004. (52) NBO Version 3.1, Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; , Weinhold, F.; . (53) Krygowski, T. M; Cyra nski, M. K.; Czarnocki, Z.; H€afelinger, G.; Katritzky, A. R. Tetrahedron 2000, 56, 1783–1796. (54) Woli nski, K.; Hilton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112, 8251–8260. (55) Poater, J.; Fradera, X.; Duran, M.; Sola, M. Chem.—Eur. J. 2003, 9, 400–406. (56) Matito, E.; Duran, M.; Sola, M. J. Chem. Phys. 2005, 122, 014109.

’ AUTHOR INFORMATION Corresponding Author

*Phone: +48 71 3757 224; fax: +48 71 3282 348; e-mail: [email protected].

’ ACKNOWLEDGMENT We would like to dedicate this work to Professor G€unter H€afelinger (University of T€ubingen, Germany) on the occasion of his 75th birthday, in acknowledgment of his contributions to the studies of the phenomenon of aromaticity. A.J.-M. and J.J.P. gratefully acknowledge the Pozna n Supercomputing and Networking Center and the Academic Computer Center CY FRONET-KRAKOW (Grant Nos.: KBN/SGI/UWrocl/029/ 1998, KBN/SGI/UWrocl/078/2001) for providing computer time and facilities, and thank the Ministry of Science and Higher Education of Poland for supporting this work under the Grant No. N N204 306137. In addition, H.S. and T.M.K. thank the Interdisciplinary Center for Mathematical and Computational Modeling (Warsaw, Poland) for computational facilities, and H.S.the Warsaw University of Technology for financial support. ’ REFERENCES (1) Jeffrey, G. A. An Introduction to Hydrogen Bonding; Oxford University Press: New York , 1997. (2) Hobza, P.; Havlas, Z. Chem. Rev. 2000, 100, 4253–4264. (3) Sobczyk, L.; Grabowski, S. J.; Krygowski, T. M. Chem. Rev. 2005, 105, 3513–3560. (4) Cleland, W. W.; Kreevoy, M. M. Science 1994, 264, 1887–1890. (5) Jeffrey, G. A.; Saenger, W. Hydrogen Bonding in Biological Structures; Springer-Verlag: New York, 1991. (6) Ed.: Grabowski, S. J. Hydrogen BondingNew Insights, (Challenges and Advances in Computational Chemistry and Physics, 3); Springer: Dordrecht, 2006. (7) Jaffe, H. H. Chem. Rev. 1953, 53, 191–261. (8) Exner, O. in Advances in Linear Free Energy Relationships; Chapman, N. B., Shorter, J., Eds; Plenum Press: London, 1972; Chpt. 1. (9) Johnson, C. D. The Hammett Equation; Cambridge University Press: New York, 1973. (10) Shorter, J. in Similarity Models in Organic Chemistry, Biochemistry and Related Fields; Zalewski, R. I., Krygowski, T. M., Shorter, J., Eds; Elsevier: Amsterdam, 1991; Chpt. 2. (11) Exner, O.; Bohm, S. Curr. Org. Chem. 2006, 10, 763–778. (12) Krygowski, T. M.; Ste) pien, B. T. Chem. Rev. 2005, 105, 3482– 3512. (13) Grabowski, S. J.; Sokalski, W. A.; Dyguda, E.; Leszczy nski, J. J. Phys. Chem. B 2006, 110, 6444–6446. (14) Koch, U.; Popelier, P. L. A. J. Phys. Chem. 1995, 99, 9747–9754. (15) Bader, R. F. W. Atoms in Molecules. A Quantum Theory; Clarendon: Oxford, 1990. (16) Panek, J. J.; Jezierska, A.; Mierzwicki, K.; Latajka, Z.; Koll, A. J. Chem. Inf. Model. 2005, 45, 39–48. (17) Laurence, C.; Berthelot, M. Perspect. Drug Discovery Des. 2000, 18, 39–60. (18) Filarowski, A.; Gzowiak, T.; Koll, A. J. Mol. Struct. 1999, 484, 75–89. (19) Rospenk, M. J. Mol. Struct. 1990, 221, 109–114. 474



ARTICLE

(57) Matta, C. AIMDELOC01: Program to Calculate the Localization and Delocalization Indices According to the Quantum Theory of AtomsIn-Molecules (AIM), available from: http://www.chem.utoronto.ca/ ∼cmatta/ (58) Schlegel, H. B. Theor. Chim. Acta 1984, 66, 333–340. (59) Hockney, R. W. Meth. Comput. Phys. 1970, 9, 136–211. (60) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865–3868. (61) Trouiller, N.; Martins, J. S. Phys. Rev. B 1991, 43, 1993–2006. (62) Nose, S. Mol. Phys. 1984, 52, 255–268. (63) Nose, S J. Chem. Phys. 1984, 81, 511–519. (64) Hoover, W. G. Phys. Rev. A 1985, 31, 1695–1697. (65) CPMD Copyright IBM Corp. 19902004, Copyright MPI fuer Festkoerperforschung Stuttgart 19972001. (66) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graph. 1996, 14, 33–38. (67) Gnuplot, Copyright (C) 19861993, 1998, 2004. Williams, T.; Kelley, C.; Copyright (C) 2004 2007. Broeker, H. B.; Campbell, J.; Cunningham, R.; Denholm, D.; Elber, G.; Fearick, R.; Grammes, C.; Hart, L. et al. (68) Hammett, L. P. Physical Organic Chemistry; McGraw-Hill: New York, 1970; Chpt. 11. (69) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165–195. (70) Krygowski, T. M.; Palusiak, M.; Plonka, A.; ZacharaHoreglad, J. E. J. Phys. Org. Chem. 2007, 20, 297–306. (71) Hu, Z.; Hardie, M. J.; Burckel, P.; Pinkerton, A. A.; Erhardt, P. W. J. Chem. Cryst. 1999, 29, 185–191. (72) Kretz, T.; Lerner, H.-W.; Bolte, M. Acta Crystallogr. 2007, E63, o4673. (73) Nakamoto, K.; Margoshes, M.; Rundle, R. E. J. Am. Chem. Soc. 1955, 77, 6480–6486. (74) Steiner, T. J. Phys. Chem. A 1998, 102, 7041–7052. (75) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985, 83, 735–746. (76) Fradera, X.; Austen, M. A.; Bader, R. F. W. J. Phys. Chem. A 1999, 103, 304–314. (77) Palusiak, M.; Krygowski, T. M. Chem.—Eur. J. 2007, 13, 7996– 8006. (78) Ste) pien, B. T.; Cyranski, M. K.; Krygowski, T. M. Chem. Phys. Lett. 2001, 350, 537–542. (79) Ste) pien, B. T.; Krygowski, T. M.; Cyranski, M. K. J. Org. Chem. 2002, 67, 5987–5992. (80) Nagaoka, S.; Nakamura, A.; Nagashima, U. J. Photochem. Photobiol. A 2002, 154, 23–32. (81) Taft, R. W.; Topsom, R. D. Prog. Phys. Org. Chem. 1987, 16, 1. (82) Jezierska, A.; Novic, M.; Panek, J. J. Pol. J. Chem. 2009, 83, 799–819. (83) Jezierska, A.; Panek, J. J.; Mazzarello, R. Theor. Chem. Acc. 2009, 124, 319–330.

475


Hydrogen Bonding as a Modulator of Aromaticity and Electronic

Recommend Documents