Mauricio
Maldonado-Domínguez
a,
Rafael
Arcos-Ramos
a,
Margarita
Romero
a,
Blas
Flores-Pérez
a,
Norberto
Farfán
*a,
Rosa
Santillan
b,
Pascal G.
Lacroix
c and
Isabelle
Malfant
c
aFacultad de Química, Departamento de Química Orgánica, Universidad Nacional Autónoma de México, 04510 México, D.F., México. E-mail: norberto.farfan@gmail.com; Fax: +52 55 56223722; Tel: +52 55 56223899 ext. 44443
bDepartamento de Química, Centro de Investigación y de Estudios Avanzados del IPN, México, D.F. Apdo. Postal 14-740 07000, México
cLaboratoire de Chimie de Coordination du CNRS, 205 Route de Narbonne, F-31077, Toulouse, France
First published on 21st October 2013
Transmission of electronic information through amide bonds may be, under appropriate conditions, effectively achieved. In this work, a family of explicitly designed donor–(amide bridge)–acceptor architectures was synthesized. NMR studies and UV-vis absorption solvatochromism support that cross-conjugation leads to measurable polarization across push–pull, amide-bridged molecules. Computational analysis of structural parameters and frontier molecular orbitals shows the contribution of an additional, dienoid amide canonical structure to intramolecular electron delocalization, as the electron donor–acceptor strength of the substituents increases. Within the context of nonlinear optics and molecular materials, computational comparison between amide-bridged molecules and those containing typical linkers shows that there is a compromise between nonlinear optical response, ease of synthesis and chemical inertness, making the systems studied herein interesting alternatives for such applications.
Intramolecular electron delocalization (IED) is intimately involved in photophysical (light absorption and emission,5 nonlinear optical (NLO) response6) and photochemical (light-driven isomerizations,7 photorelease of small molecules,8 electron transfer reactions9) phenomena of academic, technological and commercial interest. Although π electron delocalization is the most broadly studied and exploited form of IED,10 electrons are also known to delocalize over σ bonds.11 Today, it is recognized and widely accepted that IED occurs through a synergistic coupling of phenomena, such as resonance, hyperconjugation,12 superexchange (SE)13 and cross-conjugation mechanisms,14 the hydrogen-bond-assisted electron delocalization being a very interesting manifestation of SE.15
In materials science, the study, design and synthesis of molecules exhibiting electron delocalization is an area of great interest, especially in the fields of optoelectronics,16 nonlinear optics, light harvesting for solar cells,17 along with more traditional applications, the dye industry being the most representative of them. A subclass of delocalized systems of particular interest is that involving a donor and an acceptor asymmetrically located in a bridging molecule, usually a conjugated one, the so-called “push–pull” systems;18 these architectures, ushered by important charge transfer (CT) processes and potential NLO capabilities, present efficient IED because π electrons facilitate overlap between the donor and acceptor functions leading to a predetermined, spatially-biased delocalization. Among the features which may be desirable within a molecular push–pull device are rigidity (unless conformational switching properties are sought for), unidirectionality, effective electronic communication between both ends of the molecule and, of course, ease of synthesis. Although many effects (superexchange, hyperconjugation) and non-π-delocalized molecular backbones (oligopeptides,19 saturated bridges,20 spirolinked polyenes21) have proven to be useful for electron delocalization to occur, the study of alternative, readily accessible bridges for IED is a seldom studied topic, classically conjugated compounds being the focus of most of the research done within this field.
Synthetically, when a molecule with push–pull electron delocalization and large dipole moment is targeted, the first line of strategies for its design and synthesis are arene,7 alkene,22 alkyne,23 azo24 and imine25 bridges even though, sometimes, they imply long and tedious synthetic routes26 and purification procedures. A molecular motif with interesting potential is the amide function. Its low chemical reactivity and restricted rotation account for a robust structural unit, while its electronic properties (C–N bond order greater than 1) render it an interesting channel for the transmission of electronic information via cross-conjugation. The amide group is also capable of forming strong intramolecular H-bonds,27 known to behave as active electron-reorganizing bridges.28 The amide moiety is usually invoked when conformational freedom29 or chemical reactivity are to be reduced; it is also relevant for the synthesis of polymers,30 dendrimers31 and peptides.32 Up to now, no studies have been carried out for the application of the electronic features of the amide group to molecular materials. Herein we propose a model for the study of the capabilities of the amide function to act as a bridge for electron delocalization in donor–acceptor substituted molecules and its potential application in molecular materials.
The amide-bridged push–pull systems 3a–i were synthesized from different substituted anilines 2a–i and 7-diethylaminocoumarin-3-carboxylic acid351 as illustrated in Scheme 2. Compound 1 was obtained by modifying a known method.36 The amide derivatives were purified by crystallization; their structures were confirmed by 1H and 13C NMR spectroscopy, High-Resolution Mass Spectrometry (ESI-TOF), reflectance IR and elemental analysis.
The common intermediate 1 was synthesized from the reaction between Meldrum's acid and 4-(diethylamino)-salicylaldehyde with good yields (ca. 95%). The amide-bridged push–pull systems were obtained from the reaction between 1 and the respective substituted aniline 2a–i using EEDQ (N-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline) in CH2Cl2 with moderate yields (51–74% see ESI† for the general procedure). The FTIR-ATR spectroscopy of all compounds 3a–i showed the characteristic bands of the amide group, N–H (3350–3100 cm−1 and ∼1540 cm−1) and CO (∼1660 cm−1). HRMS (High Resolution Mass Spectrometry) using the Electrospray-Time of Flight technique confirmed the molecular ions. The signals in the 1H-NMR spectra at ca. 8.7, 7.4, 7.1, 6.6 and 6.5 ppm corroborate the presence of the coumarin core for all derivatives. Likewise, the 13C-NMR spectra showed the characteristic signals for the same system. The appearance of the signals between 10.8–11.1 ppm and 161.2–161.5 ppm in the 1H-NMR and 13C-NMR spectra, respectively, confirmed the presence of the amide bridge for all push–pull systems. Subtle changes are to be expected since delocalization through cross-conjugated systems is known to be relatively weak.
Fig. 1 ORTEP diagram of compounds (a) 3b, (b) 3f, (c) 3h and (d) 3i; thermal ellipsoids drawn at 50% probability level for all atoms other than H. |
All the compounds crystallized in the triclinic space group P with two molecules per unit cell. The amide bond has the trans conformation due to the existence of one intramolecular H-bond (N–H⋯O, see Table S2, ESI†); this interaction provides certain planarity to the system (torsion angles between the averaged mean planes of the aniline ring-amide bridge-coumarin ring: 3.92° for 3b, 1.44° for 3f, 0.74° for 3h and 2.76° for 3i and interplanar angles between the same rings: 10.29° for 3b, 0.05° for 3f, 3.07° for 3h and 12.75° for 3i (see Fig. 2)), which is favourable for the effective communication between the donor and acceptor.
Fig. 2 Interplanar angles for (a) 3b, (b) 3f, (c) 3h and (d) 3i. The observed quasiplanarity allows effective communication between the donor and the acceptor. |
Compound 3b shows intermolecular interactions including π–π stacking (face to face) of the coumarin core and the EWG ring (average interplanar distance 3.88 Å), intermolecular hydrogen bond between the EWG ring and the amide carbonyl (C–H⋯O, 2.663 Å), and intermolecular hydrogen bond between the aromatic fused rings of the coumarin core (C–H⋯O, 2.715 Å); these interactions produce 1D channels with the diethylamino groups oriented at opposite ends. Compound 3f shows two bifurcated intermolecular hydrogen bonds, one of them between the coumarin core and amide carbonyl (C–H⋯O, 2.695 Å), and the other one between the aniline ring and the coumarinic carbonyl (C–H⋯O, 2.713 Å). An unusual hydrogen bond ring R22 (10) between two adjacent methyl esters produce 1D layers with the diethylamino groups embraced at the other end (see Fig. S2, ESI†).
Compound 3h displays bifurcated hydrogen bonds between coumarins (C–H⋯O coumarinic carbonyl, 2.543 and 2.697 Å) and hydrogen bonds between the amide group and the aniline ring (C–H⋯O, 2.636 Å). It is interesting that the different orientation of the ethyl groups produces a bi-layer structural motif; the diethylamino methyls join the layers. Similarly to 3f, compound 3i shows intermolecular π–π stacking (face to face, interplanar average 3.639 Å), two hydrogen bonds between coumarins (C–H⋯O coumarinic carbonyl, 2.632 and 2.658 Å) and a hydrogen bond between the amide group and the aniline ring (C–H⋯O, 2.676 Å); it is worth noting that the nitro group does not participate in any interaction (Fig. 3).
These results evidence that, in the solid state, the quasiplanarity envisioned for the explored compounds is present, supporting the hypothesis that both ends of the push–pull amides may establish electronic communication.
Compound | Amide group | Diethylamino group | ||
---|---|---|---|---|
1H-NMR (NH) | 13C-NMR (CO) | 1H-NMR (CH2) | 13C-NMR (CH2) | |
3a | 10.86 | 161.09 | 3.46 | 45.14 |
3b | 10.90 | 161.19 | 3.46 | 45.17 |
3c | 10.97 | 161.16 | 3.46 | 45.17 |
3d | 10.90 | 161.14 | 3.46 | 45.16 |
3e | 10.90 | 161.15 | 3.46 | 45.17 |
3f | 11.10 | 161.39 | 3.46 | 45.19 |
3g | 11.09 | 161.50 | 3.47 | 45.20 |
3h | 11.18 | 161.62 | 3.48 | 45.24 |
3i | 11.32 | 161.71 | 3.49 | 45.26 |
As expected, the chemical shift of carbon and hydrogen atoms within the amide group correlated with the strength of the EWG under analysis. As an important result, the study showed also a good linear correlation between the 13C chemical shifts of the methylene units within the diethylamino group and the Hammett constant of the involved EWG (Fig. 4). It is noteworthy that the effect is only evident in the methylene 13C chemical shift, being highly diluted in 1H NMR and practically lost in both nuclei at the methyl termini, most likely because those groups, in addition to being less involved in inductive effects, possess the largest conformational freedom and are far more affected by the surrounding medium. These results evidence the effective communication between both ends of a push–pull system bridged by an amide spacer.
Fig. 4 Correlation between the Hammett constant (σpara) and 13C-NMR chemical shifts for the (a) 7-diethylamino group (methylene) and (b) amide carbonylic carbon. |
The absorption spectra of the family of compounds 3a–i display clear solvent polarity dependence (Table 2). An increase in the electron affinity of the EWG leads to an important bathochromic effect, with a maximum red shift of 31 nm, when comparing compound 3a in cyclohexane to molecule 3i in DMSO. These results suggest that charge separation is present and it is more important in the excited state compared to the polarization in the ground state; this implies an overall, solvent-dependent, narrowing of the π → π* excitation energy gap. Such an effect accounts for the observed red shift in absorption.
Compound | R | λ max c-hexane | λ max AcOEt | λ max DMSO |
---|---|---|---|---|
3a | H | 414 | 422 | 433 |
3b | F | 415 | 422 | 433 |
3c | Cl | 417 | 424 | 434 |
3d | Br | 418 | 424 | 433 |
3e | C2H | 419 | 426 | 437 |
3f | CO2Me | 419 | 427 | 438 |
3g | CF3 | 419 | 427 | 436 |
3h | CN | 421 | 429 | 440 |
3i | NO2 | 422 | 433 | 445 |
This dependence is reflected in the linear correlation between λmax and the normalized polarity index (PI) of the studied solvents (Fig. 5).38 Also, the EWG force linearly correlates with λmax red shift (Fig. 6), demonstrating that lower-energy absorption occurs due to push–pull-induced delocalization of the system, i.e., longer wavelength of the associated unidimensional potential-well, leading to a correspondingly higher degree of charge separation in the excited state as EWG force increases.
The double-zeta valence 6-31+G* was chosen as the basis set throughout this work. IR vibrational frequencies were computed for all the molecules at the same level of theory. All the frequencies obtained were positive and real, confirming that the stationary points correspond to stable molecular geometries. The analysis of structural parameters was made using Mercury.42
According to Linus Pauling's theory of resonance, amide bonds possess typically three canonical forms (Scheme 3A–C). In a push–pull amide, involvement of this bridge in IED would imply electron reorganization. Describing this with resonant structures leads to the appearance of an additional, dienoid form (Scheme 3D).
Structurally, a shortening of the EDG-C and N-EWG bonds should be appreciated, with the concomitant lengthening of the central C–N bond. To confirm this hypothesis, our first studies focused on the unperturbed N-phenylcoumarin-3-carboxamide molecule, introducing the push–pull system in a stepwise fashion (Scheme 4). As will be discussed, it is evident that the dienoid form becomes more important as the push–pull effect increases.
After the addition of a nitro group, the “pull system” is formed. Subtle deviations in bond lengths evidence the contribution of the dienoid canonical form; a similar effect may also be observed with the introduction of the diethylamino group in the “push system”. Communication between both units is clear in the push–pull molecule, where the net influence of the EWG and the EDG is greater than the sum of individual changes produced by each group (Table 3).
Bond 1 | Bond 2 | Bond 3 | Bond 4 | Bond 5 | |
---|---|---|---|---|---|
Unpertubed | 1.513 | 1.357 | 1.408 | 1.017 | 1.875 |
Pull system | 1.509 | 1.364 | 1.398 | 1.019 | 1.858 |
Push system | 1.507 | 1.360 | 1.406 | 1.018 | 1.874 |
Push–pull | 1.502 | 1.368 | 1.395 | 1.019 | 1.855 |
To study the IED in detail, the effect of different EWG keeping the EDG constant was evaluated. The optimized structures for our family of compounds exhibit the behaviour condensed in Table S3 (ESI†). For bond assignment, see Scheme 5. Furthermore, the variation in bond lengths linearly correlates with the Hammett constant, σpara, of the corresponding EWG (see Tables S3 and S4, ESI†).43,44
Assuming that the dienoid resonant form contributes to the net structure, it would imply an important degree of charge separation; thus, the optimized molecules were further refined, imposing a polar medium of a given dielectric constant. The Self-Consistent Reaction Field method (SCRF) with the Polarizable Continuum Model (PCM) was employed as implemented in Gaussian 09, selecting MeCN as a solvent due to its large dielectric constant (ε = 37.5).45 The results condensed in Fig. 7 contrast in vacuo values vs. structural parameters in MeCN. It is noteworthy that a highly polar medium does, in fact, favour charge separation leading to a more pronounced dienoid character of the amide bridge. It is also interesting that a linear correlation is, in general, higher in the polar medium, especially for bond 1 (R2 = 0.8949 in vacuo vs. 0.9469 in MeCN). This observation is attributed, as it will be later discussed, to a LUMO/LUMO + 1 stabilization and redistribution as solvent polarity increases.
The simultaneous electron delocalization through H-bond becomes evident when analyzing the N–H⋯O bond shortening with increasing EWG strength (Fig. 7d). Electron reorganization via the SE mechanism through this fragment linearly correlates with the stress exerted by the push–pull system, especially in polar media.
It is interesting to observe the frontier molecular orbitals of compound 3f. The ester group has a well-defined electron accepting character (σpara = 0.45), but even under such push–pull stress, the LUMO does not display the substituted anilide fragment as an important acceptor. Further investigation of higher-energy MOs shows that LUMO + 1 comprises the expected topology. The anilide moiety is, again, part of the principal donor, residing in both HOMO and HOMO − 1. Electron redistribution towards EWG occurs only between HOMO − 1 and LUMO + 1, being a transition of the type n → π*. In this system, that electron promotion involves only the anilide portion of the molecule. Up to this point, no net electron donation from the Et2N group to EWG is evident. The amide bond behaves mostly as an isolating bridge for useful electron transmission.
For molecule 3h where LUMO and LUMO + 1 are relatively low in energy, the envisioned D → (amide bond) → A transmission is finally observable, though rather weak, the IED towards the amidic carbonyl group being more pronounced (see Fig. S21, ESI†).
The most marked push–pull structure studied here, compound 3i, becomes very interesting in this context. The 4-nitroanilide portion clearly increases its acceptor character to the extent of populating the LUMO. The HOMO involves, principally, the electron-rich coumarin as expected. Furthermore the HOMO, being more localized in polar media, increases its energy thus narrowing even more the HOMO – LUMO gap (261 meV in vacuo vs. 245 meV in MeCN).
This analysis shows three interesting results: (1) the dienoid canonical form, although evidently involved in the a priori resonance description, is practically nonexistent until the push–pull stress is large enough, i.e., compound 3i in vacuum; (2) under dissociating conditions, i.e., highly polar medium, the dienoid structure appears as the LUMO + 1 for the full set of molecules, but is energetic enough to be excluded from the principal transition in most cases, except for 3h and particularly 3i where its population effectively migrates to the LUMO and (3) the appearance of such delocalizing, dienoid LUMO + 1 imprints a weak but nonzero EWG-dependent attractive character to the anilide fragment, explaining the higher correlations observed in MeCN vs. vacuum.
Since the most common linkers in such molecules are alkene, alkyne, imine and azo bridges, the first hyperpolarizability, β, of compounds analogous to 3i (Scheme 7, molecules 4a–d) was computed under the DFT/wB97X-D framework. β, a third-rank tensor, is related to second-order nonlinear phenomena such as the Kerr effect, parametric up and down conversion, among others. These values allow a direct, numerical comparison between 3i and push–pull molecules bridged by typical conjugated moieties within the context of one of their main applications in materials science.
A comparison of the values of βHRS (ref. 46) obtained for the five compounds under study (Table 4) shows that the response of 3i is expected to be roughly half of that displayed by molecules bridged by double (alkene, imine, azo) or triple (alkyne) bonds. This represents an interesting result since, in some chemical contexts, it is synthetically simpler to introduce an amide group than another linker, where heavy metal-mediated synthesis (compounds 4b and 4c, for example)47 or handling of unstable intermediates (4a) is needed.
Compound | β HRS |
---|---|
3i | 27.3 |
4a | 63.1 |
4b | 48.3 |
4c | 51.0 |
4d | 42.1 |
Its low chemical reactivity makes an amide, also, a safe alternative to bridges such as the imine double bond, highly susceptible to hydrolysis. Considering this, the amide bridge offers an interesting alternative to common bridges, with the compromise between smaller, yet comparable (within the same order of magnitude), nonlinear optical response, synthetic feasibility and chemical stability. In large-scale processes or long-term applications, this compromise may tilt towards ease of synthesis and inertness.
Computational studies also show the involvement of both structural motifs. Linear correlation with the Hammett constants of the studied EWG supports these results, while frontier orbital analysis shows the contribution of an additional, dienoid, canonical form of the amide-bridged systems as push–pull stress increases. UV-vis positive solvatochromism and DFT computations of polarity-dependent structural parameters suggest that a photoinduced charge-separation process may occur through amide bonds under appropriate conditions.
In the context of the potential application of amide bridges in molecular materials, there is a compromise between nonlinear optical response (smaller, yet significant, for amide bridges) and ease of synthesis and inertness, the amide group being an interesting alternative when considering these aspects.
Footnote |
† Electronic supplementary information (ESI) available: DFT studies for computed structures 3a–i including frontier molecular orbitals, 1H and 13C NMR spectra for compounds 3a–i, crystal data for compounds 3b, 3f, 3h and 3i. CCDC 957820–957823. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c3nj01176c |
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014 |